Bisphosphonate-associated osteonecrosis of the jaw: a key role of inflammation?

(1)

Article

Reference

Bisphosphonate-associated osteonecrosis of the jaw: a key role of inflammation?

LESCLOUS, Philippe, et al.

Abstract

Osteonecrosis of the jaw (ONJ) can be associated with nitrogen-containing bisphosphonates (NBPs) therapy. Various mechanisms of NBP-associated ONJ have been proposed and there is currently no consensus of the underlying pathogenesis. The detailed medical and dental histories of 30 ONJ patients treated with NBPs for malignant diseases (24) or osteoporosis (6) were analyzed. The necrotic bone was resected and analyzed histologically after demineralization. In 10 patients the perinecrotic bone was also resected and processed without demineralization. Alveolar bone samples from 5 healthy patients were used as controls. In 14 ONJ patients, serial technetium-99m-methylene diphosphonate scintigraphic scans were also available and confronted to the other data. Strong radionuclide uptake was detected in some patients several months before clinical diagnosis of ONJ. The medullary spaces of the necrotic bone were filled with bacterial aggregates. In the perinecrotic bone, the bacteria-free bone marrow characteristically showed an inflammatory reaction. The number of medullary inflammatory cells taken as an index of inflammation [...]

LESCLOUS, Philippe, et al . Bisphosphonate-associated osteonecrosis of the jaw: a key role of inflammation? Bone , 2009, vol. 45, no. 5, p. 843-852

DOI : 10.1016/j.bone.2009.07.011 PMID : 19631301

Available at:

http://archive-ouverte.unige.ch/unige:79930

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

Bisphosphonate-associated osteonecrosis of the jaw: A key role of in ﬂ ammation?☆

Philippe Lesclous

^a,b,

⁎ , Semaan Abi Najm

^b

, Jean-Pierre Carrel

^b

, Brigitte Baroukh

^a

, Tommaso Lombardi

^b

, Jean-Pierre Willi

^c

, René Rizzoli

^d

, Jean-Louis Saffar

^a

, Jacky Samson

^b

aLaboratoire Réparation et Remodelages Oro-Faciaux, EA2496, Université Paris Descartes, Faculté de Chirurgie Dentaire, France

bDivision of Stomatology and Oral Surgery, Faculty of Medicine, University of Geneva, Geneva, Switzerland

cDivision of Nuclear Medicine, University Hospitals of Geneva, Geneva, Switzerland

dDivision of Bone Diseases, World Health Organization Collaborating Center for Osteoporosis Prevention, Department of Rehabilitation and Geriatrics, University Hospitals of Geneva, Geneva, Switzerland

a b s t r a c t a r t i c l e i n f o

Article history:

Received 20 January 2009 Revised 5 May 2009 Accepted 7 July 2009 Available online 22 July 2009 Edited by: R. Baron Keywords:

Aminobisphosphonates Osteonecrosis Inﬂammation Macrophage/polykarions Bone scintigraphy

Osteonecrosis of the jaw (ONJ) can be associated with nitrogen-containing bisphosphonates (NBPs) therapy.

Various mechanisms of NBP-associated ONJ have been proposed and there is currently no consensus of the underlying pathogenesis. The detailed medical and dental histories of 30 ONJ patients treated with NBPs for malignant diseases (24) or osteoporosis (6) were analyzed. The necrotic bone was resected and analyzed histologically after demineralization. In 10 patients the perinecrotic bone was also resected and processed without demineralization. Alveolar bone samples from 5 healthy patients were used as controls. In 14 ONJ patients, serial technetium-99m-methylene diphosphonate scintigraphic scans were also available and confronted to the other data. Strong radionuclide uptake was detected in some patients several months before clinical diagnosis of ONJ. The medullary spaces of the necrotic bone were filled with bacterial aggregates. In the perinecrotic bone, the bacteria-free bone marrow characteristically showed an inflammatory reaction. The number of medullary inflammatory cells taken as an index of inflammation allowed us to discriminate two inflammation grades in the ONJ samples. Low-grade inflammation, characterized by marrowfibrosis and low inflammatory cells infiltration, increased numbers of TRAP⁺mono- and multineacleated cells was seen in patients with bone exposureb2 cm². High-grade inflammation, associated with larger lesions, showed amounts of tartrate-resistant acid phosphatase⁺/calcitonin receptor⁻ mono- and multinucleated cells, osteocyte apoptosis, hypervascularization and high inflammatory cell infiltration. The clinical extent of ONJ was statistically linked to the numbers of inflammatory cell. Taken together these data suggest that bone necrosis precedes clinical onset and is an inflammation-associated process. We hypothesize that from an initial focus, bone damage spreads centrifugally, both deeper into the jaw and towards the mucosa before the oral bone exposure and the clinical diagnosis of ONJ.

Introduction

Postmenopausal osteoporosis, Paget's disease, and metastatic bone diseases are characterized by high bone turnover. One of the most efﬁcient ways to treat these affections is the use of bisphosphonates.

Among them nitrogen-containing bisphosphonates (NBPs) are the most potent in inhibiting bone resorption[1]. Osteonecrosis of the jaw (ONJ) associated with NBPs therapy is a recently described detri- mental effectﬁrst reported in 2003[2]. ONJ is characterized by an area of exposed bone in the maxillofacial region that does not heal within 8 weeks, in a patient who has no metastasis or radiation therapy in the craniofacial region [3]. High repeated doses of NBPs administered

intravenously to cancer patients, are most frequently associated with ONJ[4–6]. In contrast, patients receiving oral NBPs for osteoporosis are less prone to ONJ[7].

Various mechanisms of NBP-associated ONJ have been proposed. It is generally admitted that this adverse effect targets the jaws because they concentrate NBPs owing to their high bone turnover compared with other anatomical sites [8], and bisphosphonate toxicity on osteoclasts[9]. It has also been proposed that NBPs may have an antiangiogenic effect, reducing bloodﬂow and thus inducing bone cell necrosis and apoptosis [10,11]. One alternative explanation is an

“outside-in”process in which mucosal damage provides oral bacteria with access to the underlying bone, leading to bone infection and necrosis[12]. NBP soft tissue toxicity would thus be the trigger leading to necrosis of the mucosa and bone[13–15]. Genetic variations may also constitute a risk factor; it was recently shown that ONJ is associated with gene polymorphisms in multiple myeloma patients [16,17]. Moreover, the primary disease is probably a predisposing factor as frequency of ONJ is 2 fold higher in multiple myeloma than in breast

☆ Source of support: This work was supported in part by the Société Francophone de Médecine Buccale et Chirurgie Buccale.

⁎Corresponding author. Faculté de Chirurgie Dentaire Paris Descartes, 1 Rue Maurice Arnoux, 92120 Montrouge, France.

E-mail address:philippe.lesclous@univ-paris5.fr(P. Lesclous).

doi:10.1016/j.bone.2009.07.011

Contents lists available atScienceDirect

Bone

j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / b o n e

(3)

(4)

cancers despite similar NBPs treatments[6]. It is likely that ONJ results from a conjunction of environmental, metabolic and genetic factors.

The role of NBPs in the process was suggested in dogs receiving clinically relevant daily oral alendronate treatment for 3 years that developed mandible bone matrix necrosis without oral exposure[18].

As there is currently no consensus on its underlying mechanism, ONJ task forces have called for more clinical and basic research [3,7,19,20]. The aim of the present study was to examine clinical and histopathological features of patients with ONJ in an attempt to better understand the pathogenesis of this disorder.

Patients and methods

Patients

Thirty bisphosphonate-treated patients with persistent oral bone exposure, referred by dental practitioners or oncologists to the Department of Stomatology and Oral Surgery of the University Hospitals of Geneva (Switzerland), were enrolled in this study. The patients were anonymously included in the study that complied with the Declaration of Helsinki. Each patient's detailed medical and dental histories were analyzed. Criteria for diagnosis of bisphosphonate- related osteonecrosis of the jaw, according to the ASBMR[3], were:

− ongoing intravenous or oral bisphosphonate therapy,

− no history of maxillofacial radiation therapy,

− oral exposure of necrotic bone at the maxilla or the mandible for 8 weeks without healing, and

− exclusion by histological examination of bone metastases of a primary malignant disease and conﬁrmation of bone necrosis.

Before the initial consultation in the department, the patients had been diagnosed with chronic refractory osteomyelitis and treated with repeated courses of antibiotics and analgesics. Antecedents of local infectious conditions, especially periodontal status, were carefully assessed with past clinical and X-ray data provided by the dental practitioners. An orthopantomogram (OPG) was systematically done to rule out local etiologies. Serial technetium-99m-methylene diphosphonate scintigraphic scans (⁹⁹Tc^m-MDP) done prior ONJ diagnosis were available for 14 patients treated for bone metastases of solid tumors (Fig. 1).

Local ONJ treatment strategies

Three treatment strategies based on the size of bone exposure were used. Patients with mild ONJ (bone exposureb1 cm²) and patients who refused invasive procedures were treated conservatively with systemic antibiotics and local antiseptics, and analgesics if necessary. Patients with moderate ONJ (1 cm²Nbone exposureb2 cm²) had minimal invasive surgery consisting of decortication andﬂattening of rough necrotic bone,ﬁbrin glue coverage, closure with mucosal tissue, and medical treatment as described above. Patients with severe ONJ (bone exposureN2 cm²) had a more extensive procedure, including bone

resection and curettage of the surrounding bone, and local and medical treatments as described above.

All the patients were closely monitored, every 3 or 4 months the ﬁrst year and then every 6 months. An OPG was performed every 6 months. Outcome (extension, improvement or healing of mucosal and bone tissues) was evaluated clinically and radiographically.

Mucosal healing was deﬁned as complete mucosal closure, mucosal improvement as a decrease in the area of exposed bone, and mucosal spreading as an increase of this area.

Processing of biopsy specimens

Necrotic bone was resected in 26 patients and was examined after formaldehyde fixation, formic acid 10% demineralization, paraffin embedding, sectioning, hematoxylin–eosin or PAS stainings. In 10 patients, the perinecrotic bone was also sampled, it wasfixed in 70%

ethanol, dehydrated and embedded without previous demineralization in methylmethacrylate (Merck, Darmstadt, Germany). Bone samples taken from 5 healthy alendronate-treated patients during dental implant surgery were processed in the same way and used for control.

Series of 4-μm-thick sections were stained with toluidine blue (pH 4.3) or hematoxylin–eosin or processed for tartrate-resistant acid phosphatase (TRAP) revelation and immunohistochemistry. TRAP was detected by using hexazotized pararosanilin (Sigma, St Louis, MO) and naphthol ASTR phosphate (Sigma) to reveal osteoclasts; non-osteoclastic acid phosphatase was inhibited by adding 100 mMol/L L(+)-tartaric acid (Sigma) to the substrate solution. A mouse monoclonal antibody against human calcitonin receptor (CTR; MCA2191, AbD Serotec, Düsseldorf, Germany), a rabbit polyclonal antibody against caspase-3 to reveal apoptotic cells (12-01-16362; Biocarta, San Diego, CA), and a rabbit polyclonal antibody against von Willebrand factor (F3520, Sigma) to reveal vascular structures, were used. The sections were incubated with the primary antibodies (CTR 1:25; caspase-3 1:500; von Willebrand factor 1:400, at room temperature in a moist chamber overnight) and then with secondary biotinylated antibodies (horse anti-mouse or goat anti-rabbit IgG, both from Vector; 1:200, 90 min at room temperature).

After treatment with 3% hydrogen peroxide for 10 min and avidin–biotin peroxidase complex (ABC Vectastain Kit, Vector, Burlingame, CA) for 1 h, the chromogen 3,3′-diaminobenzidine tetrahydrochloride (Sigma) was added. Negative controls were prepared by omitting the primary Table 1

Distribution of ONJ severity according to the primary pathology and potential co- morbidity factors.

Primary pathology Co-morbidity factors

Myeloma Solid cancer Osteoporosis C+C Db To

Mild 1 0 1 1 1 1

Moderate 2 1 3 3 2 2

Severe 7 13 2 11 2 2

10 14 6

C+C: current chemotherapy + cortisteroids; Db: diabetes; To: active smoker.

Fig. 1.History of ONJ onset and development. Patient 1: ONJ-free patient (A and B). A 78-year-old woman with breast cancer and temporal bone metastases receiving IV zoledronic acid (4 mg/month for 11 months) showed spots of radionuclide uptake in the anterior part of the maxilla and in the left mandible corresponding to recent tooth extractions (arrows) and to an infectious dental focus (arrowheads). Patient 2 (C to F). A 60-year-old woman with breast cancer and bone metastases receiving IV zoledronic acid (4 mg/month for 20 months). In September 2005, two left lower molars (C) became painful and mobile, with no radiographic evidence of an infectious process; they were extracted. In March 2006, the sockets of the two molars were still open, and no signs of bone healing were discernible on the X-ray (D). Clinically, necrotic bone was exposed to the oral environment and extended to a large part of the left mandible (E). The patient died one month later. Successive⁹⁹Tc^m-MDP scintigraphies (F) show the progression of radionuclide uptake in the mandible: in April 2004, two spots of radionuclide uptake (arrows) were visible; the uptake progressed backwards on the left side and the spots merged; after the dental extractions, most of the mandible was involved. Patient 3 (G to M). A 45-year-old man with multiple myeloma treated with IV pamidronate (30 mg/month for 78 months), then zoledronic acid (4 mg/month for 8 months). In April 1999 (G), after 2 years of pamidronate therapy, there was no evidence of jaw involvement. Five years later (H), the patient self-extracted a very mobile and painful second left lower molar. Oral bone exposure occurred. At the same time, CT scan (I) showed, besides the empty molar socket (arrowhead), interruption of oral cortex continuity in the right lower molar region (arrows), with no clinical manifestations. One month later, scintigraphy showed two distinct spots of radionuclide uptake in the mandible (J), the left (Le) one was large and corresponded to the oral bone exposure that occurred after molar self-extraction, whereas the smaller right (Ri) one (arrows), superimposable on the CT image, occurred in a clinically unaffected zone. In June 2005, another molar was lost (K) and the necrotic bone exposure extended to this zone (L). The last scintigraphy showed the growth of the left spot and the persistence of the right spot (M); the latter did not become clinically manifest, as the patient died two months later.

(5)

Table 2

Characteristics of the patients.

The patients are classiﬁed according to ONJ outcome: mucosal healing (light shade), mucosal improvement (medium shade) and deterioration (dark shade). MM: Multiple Myeloma;

Pm Ost: Postmenopausal osteoporosis; Cort Ost: Corticosteroid Osteoporosis;⁎the corticoid treatment was discontinued more than 1 year before ONJ onset; NBP: Nitrogen- containing bisphosphonate; Pam: Pamidronate; Zol: Zoledronic acid; Alen: Alendronate; Iban: Ibandronate; IV: Intravenous; PO: Per Os.

(6)

antibody, replacing the primary antibody with rabbit non-immune serum at the same dilution, and using an irrelevant secondary antibody.

Morphometry was performed in the non-demineralized samples.

Statistical analysis

Data were compared using non-parametric tests (Kruskal–Wallis followed, if signiﬁcant, by group comparisons with the Mann– WhitneyUtest). The association of clinical extent of ONJ with the numbers of inﬂammatory cells was assessed using the Wilcoxon test.

Differences were considered signiﬁcant atPb0.05. Data are expressed as means ± SEM.

Results

Characteristics of the patients (Tables 1 and 2)

The distribution of the patients according to the initial pathology and potential co-morbidity factors is shown inTable 1. The cancer patients had previously received various cancer treatments but no maxillofacial radiation therapy. The female:male sex ratio was 2:1. The mean age was 69.9 ± 2.4 years (females: 68.9 ± 2.9; males: 71.3 ± 4.8). The control patients (4 women and 1 man) were treated for osteoporosis. The mean age was 73.8 ± 2.6 years (females: 72.2 ± 2.7;

male: 80); among these patients 1 was active smoker and another one diabetic(Table 2).

Bisphosphonate regimens (Table 2)

The cancer patients received intravenous pamidronate or zoledronic acid alone, or sequentially, pamidronate followed by zoledronic acid;

one patient received ibandronate after one year of zoledronic acid.

Osteoporosis patients received oral alendronate; in one patient, oral alendronate was replaced after 4.5 years by intravenous pamidronate.

The average time between NBP initiation and clinical diagnosis of ONJ was shorter with zoledronic acid (median: 19 months, range 5–25) than with pamidronate (median: 32 months, range 13–60), pamidronate/

zoledronic acid (median: 37.5 months, range: 23–87) and alendronate (median: 72 months, range: 44–120). Owing to the strong and long- lasting accumulation of bisphosphonates in bone tissue, and because of the different therapeutic protocols, the cumulative dose of bisphosphonates is probably more relevant than the treatment duration. The mean cumulative doses at the time of ONJ diagnosis were 67.2± 15.9 mg for the zoledronic acid, 3252.8± 229.9 mg for the pamidronate,1347.3 ± 249.5 mg+ 72.0± 14.4 mg for the pamidronate/zoledronic acid regi- men, and 20608.0 ± 2414.6 mg for the alendronate treatment. In this cohort, no ONJ was recorded under the cumulative dose of 20 mg zoledronic acid, 1170 mg pamidronate and 12320 mg alendronate.

Clinical aspect of osteonecrosis (Table 3)

The mandible was affected in 54% of cases, the maxilla in 43%, and both sites simultaneously in 3%. Most lesions (84%) affected the

premolar–molar regions, i.e. where the forces of mastication are heavier[21]. ONJ was recorded in 87% of cases in sites of dentoalveolar surgery (mainly tooth extraction). The interval between surgery and ONJ diagnosis ranged from 2 to 6 months. In 13 patients, teeth were extracted for abnormal and painful mobility with no clinical or radiographic signs of dental or periodontal infection; in these cases a generalized subtle widening of the periodontal ligament was discernible by radiography. These signs do not ﬁt in with typical periodontitis features, so we hypothesized that pain and tooth mobility were early manifestations of an already present ONJ. The other dental extractions were due to infections. The remaining 13% of patients were edentulous and did not wear dentures(Table 3).

In 5 of the 14 patients who had⁹⁹Tc^m-MDP scintigraphies, focal jaw uptake of bone-seeking isotopes preceded clinical onset and oral exposure of necrotic bone by several months to more than a year (Fig. 1). In the 9 other patients, jaw radionuclide uptake occurred after the extractions. Other causes of radionuclide uptake, including local metastasis, were excluded. An additional series of scintigraphies from 14 age and sex-matched ONJ-free patients was used as control;

two of them showed a focal uptake in the jaws, corresponding to infectious dental foci or recent extraction sites (Figs. 1A–B).

Follow-up (Table 4)

All the patients continued to receive NBPs during follow-up. Owing to the strong and long-lasting accumulation of bisphosphonates in bone tissue and to the beneﬁts of the treatment, the medical team decided to not interrupt the NBP treatment. In fact, no evidence suggests that discontinuation of the bisphosphonate therapy will promote ONJ resolution[3] (Table 4).

Four patients (13%) were conservatively treated (two patients with mild ONJ and two patients with severe ONJ refusing more invasive procedures); the six patients (20%) with moderate ONJ had minimal invasive surgery. The other 20 patients with severe ONJ (70%) had a more extensive surgical procedure. In the group subjected to conservative approach, mucosal healing and bone stabilization were reached in 2 osteoporotic patients with alendronate alone whereas mucosal improvement and bone spreading were recorded in the 2 patients affected by severe ONJ with pamidronate and zoledronate sequentially. In the 6 patients subjected to minimal invasive surgery, 2 osteoporotics with alendronate alone showed mucosal healing and bone improvement; mucosal closure did not occur and bone condition deteriorated in the 4 others whatever the NBP treatment was. In the 20 remaining patients subjected to a more extensive procedure, all types of outcomes were recorded. No mucosal closure or bone stabilization occurred in the patients receiving zoledronic acid alone.

Histopathologicalﬁndings Necrotic bone

In the 26 demineralized samples, necrotic bone showed empty osteocyte lacunae, the medullary spaces wereﬁlled with bacterial Table 3

Distribution of ONJ sites and precipitating events.

Number Region Surgical procedure Spontaneous onset

Associated manifestations

Mandible 16 Pm.M:

14 I.C.:

2

Dental extraction:

12 Implant surgery:

1

3 Extraoral

ﬁstula: 3

Maxilla 13 Pm.M:

11 I.C.:

2

Dental extraction:

12 Implant surgery:

1

Sinusitis: 8

Maxilla + Mandible

1 Pm.M:

1 I.C.: 1

Dental extraction: 1 1

Pm: premolar; M: molar; I: incisor; C: cuspid.

Table 4

Outcome of ONJ according to the local treatment procedure.

Number Follow-up (months)

Clinical issue Radiological issue Conservative

approach

4 20.5 ± 2.1 Mucosal healing: 2 Improvement: 2 Spreading: 0

Improvement: 0 Stabilization: 2 Spreading: 2 Minimal invasive

surgery

6 21.8 ± 2.5 Mucosal healing: 2 Improvement: 3 Spreading: 1

Improvement: 2 Stabilization: 0 Spreading: 4 Bone resection 21 22.6 ± 1.7 Mucosal healing: 8

Improvement: 7 Spreading: 6

Improvement: 6 Stabilization: 2 Spreading: 13

(7)

Fig. 3.Variations in the multinucleated cell phenotype and numbers according to the severity of inflammation. Revelation of TRAP (A, C, E) and calcitonin receptors (CTR, in B, D, F). In a control sample a single TRAP⁺multinucleated cell in contact with the bone surface is visible (A); this cell is also CTR⁺(B) and can be identified as an osteoclast. In a sample with low-grade inflammation, TRAP⁺cells are more numerous (C) but only some of them are also CTR⁺(D); only one of these osteoclasts (arrows in C and D) is in contact with the bone surface. In a high-grade sample, mono- and multinucleated TRAP⁺cells are very numerous (E) within the marrow space, but only one of them is CTR⁺(F). The TRAP⁺/CTR⁻ multinucleated cells are likely macrophage polykarions. Bar = 50μm.

Fig. 2.ONJ-related inflammatory cell infiltration in the marrow spaces. A. Schematic distribution of the control (○) and ONJ patients (■) according to the number of medullary inflammatory cells. The ONJ cohort appeared to be heterogenous and 2 distinct subgroups could be evidenced: a low-grade (LG) group and a high-grade (HG) group, which were different when statistically tested. B. Control sample. C. Low-grade inflammation sample. D. High-grade inflammation sample. The inflammatory cells are stained in dark with the hematoxylin–eosin technique. B to D, same magnification, bar = 100μm.

(8)

aggregates (cocci andﬁlaments), and, deeper, by polymorphonuclear leukocytes.Actinomycescolonies were frequent andCandida albicans was occasionally found.

Perinecrotic bone

In the 10 non-demineralized samples, the main feature of perinecrotic bone was an inflammatory reaction in the marrow spaces, with marrowfibrosis, inflammatory cell infiltration, development of the vascular network, and osteocyte apoptosis. Inflammation was evaluated by counting the inflammatory cells. The comparison between the control and ONJ patient samples showed a marked difference (pb0.01;

29.9 ± 2.9 vs 183.6 ± 82.8 cells/mm², respectively). The large intragroup variation in the perinecrotic bone samples prompted us to visualize their distribution (Fig. 2); we could then discriminate two ONJ patient subgroups, one including 4 samples ranging from 78.2 to 124.4 cells/mm²and another one including 6 samples ranging from 177.2 to 329.9 cells/mm². When these groups were statistically compared, they differed significantly (pb0.01). From these data, two grades of inflammation were distinguished (Fig. 2). In the low-grade group, the bone surface activities appeared to be slightly modified compared with samples from the controls. TRAP⁺/CTR⁺osteoclasts, in contact or not with the bone surface, were frequent (Fig. 3). The peripheral bone lamellae appeared to be less mineralized than the core of the trabeculae. Osteocytes were seldom caspase-3-positive (Fig. 4) and empty osteocyte lacunae were rare. The medullary spaces showed some fibrosis. Small von Willebrand⁺vessels were distributed throughout the marrow (Fig. 5).

In the high-grade inﬂammation group, bone formation was no longer visible. TRAP⁺multinucleated cells in contact with the bone surface were more numerous than in the low-grade samples, and most of them were CTR⁻(Fig. 3). Most osteocyte lacunae were empty and the remaining osteocytes were generally apoptotic (Fig. 4). The medullary spaces wereﬁbrotic. Mono- and multinucleated TRAP⁺/

CTR⁻cells were numerous (Fig. 3). The vascular network was strongly increased, and the vessels were large and dilated (Fig. 5).

Low-grade inﬂammation was mainly seen in patients treated with oral NBPs while high-grade inﬂammation was most frequent in patients treated with intravenous NBPs.

Morphometric data (Table 5)

The control, low-grade and high-grade specimen groups as described above were significantly different for all the parameters, except for the number of CTR⁺cells that remained similar. A close association between the clinical extent of ONJ (score 1: lesionb2 cm²; score 2: lesionsN2 cm²) and the numbers of inflammatory cells (taken as an index of inflammation severity) was found (Wilcoxon test:z=−2.8, p = 0.01)(Table 5).

Discussion

ONJ has recently emerged as a serious potential adverse effect of NBP therapy. For the ﬁrst time, we present data pointing to an association with local inﬂammation. We also propose that bone necrosis and exposure are preceded by early asymptomatic and infectious-free bone involvement.

Thefirst noteworthy result of this study is the presence in the perinecrotic bone of bone marrow inflammation whose severity is associated with the clinical extent of ONJ. The higher bone marrow inflammation the more severe the ONJ. The tissue underlying necrotic bone showed two grades of inflammation. In thefirst situation, the inflammatory reaction was mild, and its hallmarks were bone marrow fibrosis and moderate inflammatory cell infiltration without vascular change. The observed stimulation of osteoclast recruitment despite ongoing bisphosphonate treatment may result from this inflammatory state; accordingly, Endo et al. reported that NBP injections in mice increased marrow macrophage and osteoclast numbers[22]. Other

Fig. 5.Changes in the vasculature. In the control (A) and low-grade inﬂammation (B) samples, a few von Willebrand factor⁺vessels are scattered throughout the bone marrow.

In contrast, in the high-grade inﬂammation sample (C), the vessels are abundant, large and dilated.

Fig. 4.Osteocyte apoptosis increases with the severity of the inﬂammatory process. In the control (A) and low-grade inﬂammation (B) samples, few osteocytes are caspase 3⁺. In the high-grade sample (C), the number of apoptotic osteocytes is strongly increased. Bar = 50μm.

(9)

patients had severe inflammation of the bone marrow with strong inflammatory cell infiltration. Another striking feature was osteocyte apoptosis and empty osteocyte lacunae increasing with the degree of inflammation. Osteocyte apoptosis might then be a consequence of the intense inflammatory reaction. Although NBPs are thought to protect osteocytes from apoptosis, they may be toxic beyond a threshold cumulative dose, as this preventive effect was shown either in vitro or with low doses of NBPs [23–24]. Interestingly, only osteocytes underwent apoptosis. Another characteristic was an increased number of TRAP⁺multinucleated cells within the marrow spaces. Marie et al. reported such accumulation of TRAP⁺ multinucleated cells, not in contact with the bone surface, in mice treated with NBPs [25]. TRAP expression is not specific to the osteoclast phenotype, as other monocyte-derived cells, such as activated and inflammatory macrophages, also express TRAP in bone marrow[26].

Macrophages also can fuse to form giant cells able to digest large components or to resorb extracellularly after strongly adhering to the substrate in chronic inflammatory sites; as for osteoclasts, multi- nucleation amplifies their resorptive ability [27]. These cells share with osteoclasts numerous functional markers and calcitonin receptor expression distinguishes osteoclasts from macrophage polykarions [28]. In our samples, most of the multinucleated cells did not express this marker. These TRAP⁺/CTR⁻polykarions probably differentiate in response to strong marrow inflammation and/or as a consequence of NBP treatment itself. Unlike TRAP⁺/CTR⁺osteoclasts, these TRAP⁺/ CTR⁻polykarions may not take up NBPs and would thus be less sensitive to their cellular toxicity. In fact, macrophages can internalize NBPs depending on their endocytic ability and on the soluble pool of NBP available[29], which is probably low in the medullary spaces. We suspect that the polykaryons are involved in clearing necrotic tissues.

The strong inﬂammation was accompanied by prominent development of medullary vessels. This argues against the assumption that

NBPs, and particularly zoledronic acid, reduce bone angiogenesis [10,11]. Our observation is in keeping with the pronounced bone- seeking-isotope uptake, pointing to the persistence of an effective blood supply to the affected site. Accordingly, no differences in genes associated with angiogenesis were reported in multiple myeloma patients with NBPs related ONJ[17]. Thus, defective local circulation is unlikely to be responsible for the necrosis, contrary to previous deductions[19,30].

The site speciﬁcity of ONJ could be attributed to the embryological origin of bone marrow stromal cells (BMSC), indeed Stefanik et al.

showed that therapeutically relevant dose of pamidronate promoted in vitro higher RANKL gene expression and osteoclast differentiation by mandible BMSC than by iliac crest BMSC[31]. Moreover, the rate of bisphosphonate binding and incorporation by bone is directly related to the rate of local bone turnover. Alveolar bone and edentulous ridges exhibit high turnover[32,33]and probably incorporate more NBP than other skeletal sites[13]. By drastically reducing bone turnover, NBPs lead to dose-dependent accumulation of microdamages within the bone matrix[34,35], thereby affecting the mechanical properties of bone and reducing its energy-absorption capacity[36]. ONJ may follow microdamage accumulation resulting from mastication loading as microdamage accumulation occurs early during NBP treatment[37].

Osteocytes are sensors of matrix microdamages and control the targeted remodeling by entering apoptosis [38]. As resorption is inhibited by NBPs, the signals emitted by apoptotic osteocytes may trigger an inflammatory process that can be detected in the form of focal radionuclide uptake. NBPs themselves might also induce the inflammatory reaction. Indeed, a transient early acute-phase reaction [39]mediated by stimulated Vγ9Vδ2 T cells occurs after thefirst NBP administration, but not after subsequent infusions[40]. NBPs may act on cells other than Vγ9Vδ2 T cells, through sustained inhibition of farnesyl diphosphate synthase, a key enzyme in the mevalonate pathway[41]. Mevalonate inhibition would induce the production and release of IFNγby monocytes and macrophages[42], followed by TNF- α and IL-1β [41,43]. NBPs also induce leukocyte adherence to endothelial cells and transendothelial migration of mononuclear cells[44], mediated by increased ICAM-1 expression by monocytes through TNF-α stimulation; this occurs preferentially following intravenous administration[43]. An intriguing observation was that the number of TRAP⁺/CTR⁺osteoclasts remained unchanged between the control and ONJ samples, suggesting that the osteoclast differentiation process was not impaired as suggested by Stefanik et al.[31].

However our static analysis provides no indication on osteoclast activity; according to the current literature osteoclast activity in the NBP-treated patients was probably profoundly altered[1].

In these conditions, we hypothesize that ONJ is primarily an aseptic process. Many patients presented with pain and discomfort due to mobile teeth that were extracted despite the lack of clinical or radiological evidence of an infectious process. In our series, some

99Tc^m-MDP scintigraphic follow-up of malignant disease showed radionuclide uptake, and thus bone involvement, well before clinical manifestation onset, showing that early localized bone damage Table 5

Variations in the morphometric data in the perinecrotic bone samples.

Control samples LG samples HG samples

N= 5 N= 4 N= 6

Inﬂammatory cells (N/mm²) 29.9 ± 2.9 103.6 ± 11.9⁎⁎ 236.9 ± 24.0⁎^,⁎⁎⁎

Mononucleated TRAP⁺cells (N/mm²)

3.8 ± 0.4 5.5 ± 0.4⁎⁎ 8.1 ± 0.8⁎^,⁎⁎⁎

Multinucleated TRAP⁺cells (N/mm²)

1.4 ± 0.2 2.7 ± 0.2⁎⁎ 6.0 ± 0.2⁎^,⁎⁎⁎

CTR⁺cells (N/mm²) 1.2 ± 0.2 1.1 ± 0.1 1.0 ± 0.2 Osteocytes (N/mm²) 233.1 ± 19.7 176.4 ± 16.8⁎⁎ 98.2 ± 7.8⁎^,⁎⁎⁎

Apoptotic osteocytes (N/mm²) 2.4 ± 0.2 5.3 ± 0.3⁎⁎ 11.0 ± 1.3⁎^,⁎⁎⁎

Empty osteocyte lacunae (%) 15.8 ± 0.9 19. 1 ± 0.5⁎⁎ 24. 7 ± 1.1⁎^,⁎⁎⁎

Vessels (N/mm²) 26.8 ± 2.4 40.6 ± 2.7⁎⁎ 94.7 ± 2.2⁎^,⁎⁎⁎

Mean vessel area (μm²) 118.9 ± 5.2 245.5 ± 7.5⁎⁎ 660.4 ± 17.8⁎^,⁎⁎⁎

Two grades of inﬂammation severity (LG: low grade; HG: high grade) were determined after the numbers of inﬂammatory cells. Except osteocyte data, the parameters were counted in the medullary spaces.

⁎P= 0.006.

⁎⁎ P= 0.01 vs the control group.

⁎⁎⁎P= 0.01 vs the LG group.

Fig. 6.Proposed chronological sequence of events underlying ONJ during NBP therapy. A focus of inﬂammation of unclear origin forms within the bone (visible by radionuclide uptake on scintigraphy) and progresses aseptically to necrosis over a period of several months. Necrosis spreads and reaches the oral mucosa, destroying it and leading to bone exposure.

Oral bacteria invade the exposed necrotic bone, aggravating the situation. ONJ is generally diagnosed several weeks after the mucosal breach. The process (and its diagnosis) can be hastened by surgical procedures indicated by unexplained clinical manifestations (pain, tooth mobility, etc.).

(10)

preceded oral bone exposure. Recently, Raje et al. reported in clinically unaffected sides of the jaws of ONJ patients imaging abnormalities similar to the affected sides[17], a feature shown here inFig. 1(patient 3). Thisﬁts with Allen and Burr's proposal that bone necrosis precedes clinical manifestation of ONJ[18]. Thus, the unexplained manifestations leading to dental extraction would be theﬁrst clinical sign of underlying bone necrosis. Interestingly, cases of bisphosphonate- related ONJ without bone exposure were recently reported[45,46].

However, a subclinical infectious origin cannot be totally excluded even if, when mucosal closure was achieved, to protect the bone from the septic oral environment, bone necrosis frequently continued to spread. The early bone radionuclide uptake strongly suggests that bone, and not the mucosa, is the primary target of the NBP-related insult, implying that the observed breakdown of the oral mucosa is secondary to osteonecrosis. Locally high bone concentrations of NBPs would be toxic for adjacent soft tissues, resulting in mucosal disruption [13]. As NBPs inhibit keratinocyte growth by inhibiting the mevalonate pathway[47], healing of the mucosal breach would be delayed[15], thus facilitating secondary colonization of the necrotic bone by the oralﬂora.

This study conﬁrms that ONJ is drug-, dose- and route of administration-dependent [6]. Patients receiving repeated high doses of NBPs were at an increased risk of ONJ. Zoledronic acid is more frequently implicated in ONJ, possibly because of its strong antiresorptive potential[48]. Use of the intravenous route probably results in rapid tissue accumulation[49,50]and is more deleterious than the oral route. It is noteworthy in this respect that no cases of ONJ occurred in a series of 3105 osteoporotic patients treated for 3 years with 5 mg of zoledronic acid once a year, a dose ten-fold lower than that used to treat metastatic bone disease (4 mg monthly) [51].

Outcome after local treatment was related to the dose and the drug, patients treated with zoledronic acid at a high cumulative dose having the poorest outcome.

In conclusion, ourfindings in a small patient cohort suggest that an initial focus of damaged bone, which takes up radionuclides, spreads centrifugally both deeper into the jaw and towards the mucosa. The whole process appears to be associated with an inflammatory reaction. A tentative sequence of events is proposed inFig. 6. However the exact role of inflammation in this sequence is still not clear; it may be revelatory of other not yet disclosed phenomena. As NBPs are highly beneficial for cancer patients with bone involvement and osteoporotic patients, early markers of bisphosphonate-associated ONJ would be highly useful.

Acknowledgments

The authors thank Annie Llorens (laboratory technician) and the practitioners for their kind assistance.

References

[1] Rogers MJ, Gordon S, Benford HL, Coxon FP, Luckman SP, Monkkonen J, et al.

Cellular and molecular mechanisms of action of bisphosphonates. Cancer 2000;88:

2961–78.

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

[3] Khosla S, Burr D, Cauley J, Dempster DW, Ebeling PR, Felsenberg D, et al.

Bisphosphonate-associated osteonecrosis of the jaw: report of a task force of the AmericanSociety forBone andMineral Research. J Bone Miner Res 2007;22:

1479–91.

[4] Durie BGM, Katz M, Crowley J. Osteonecrosis of the jaw and bisphophonates.

N Engl J Med 2005;353:99–100.

[5] Bamias A, Kastritis E, Bamia C, Moulopoulos LA, Melakopoulos I, Bozas G, et al.

Osteonecrosis of the jaw in cancer after treatment with bisphophonates: incidence and risk factors. J Clin Oncol 2006;34:8580–5.

[6] Hoff AO, Toth BB, Altundag K, Johnson MM, Warneke CL, Hu M, et al. Frequency and risk factors associated with osteonecrosis of the jaw in cancer patients treated with intravenous bisphosphonates. J Bone Miner Res 2008;23:826–36.

[7] Rizzoli R, Burlet N, Cahall D, Delmas PD, Eriksen EF, Felsenberg D, et al.

Osteonecrosis of the jaw and bisphosphonate treatment for osteoporosis. Bone 2008;42:841–7.

[8] Baron R, Saffar JL, Duﬂot-Vignery A. Alveolar bone remodelling in the rat: normal status and effects of PTX and PTH on remodelling sequence and the osteoclastic pool. Calcif Tissue Res 1977;22 Suppl:502–4.

[9] Russel RG, Rogers MJ, Frith JC, Luckman SP, Coxon FP, Benford HL, et al. The pharmacology of bisphosphonates and new insights into their mechanisms of action. J Bone Miner Res 1999;14(Suppl 2):53–65.

[10] Fournier P, Boissier F, Filleur S, Guglielmi J, Cabon F, Colombel M, et al.

Bisphosphonates inhibit angiogenesis in vitro and testosterone-stimulated vascular regrowth in the ventral prostate in castrated rats. Cancer Res 2002;62:

6538–44.

[11] Santini D, Vincenzi B, Avvisati G, Dicuonzo G, Battistoni F, Gavasci M, et al.

Pamidronate induces modiﬁcations of circulating angiogenic factors in cancer patients. Clin Cancer Res 2002;8:1080–4.

[12] Hansen T, Kunkel M, Weber A, Kirkpatrick CJ. Osteonecrosis of the jaws in patients treated with bisphosphonates: histomorphologic analysis in comparison with infected osteoradionecrosis. J Oral Pathol Med 2006;35:155–60.

[13] Reid IR, Bolland MJ, Grey AB. Is bisphosphonate-associated osteonecrosis of the jaw caused by soft tissue toxicity? Bone 2007;41:318–20.

[14] Scheper MA, Badros A, Chaisuparat R, Cullen KJ, Meiller TF. Effect of zoledronic acid on oralﬁbroblasts and epithelial cells: a potential mechanism of bisphosphonate- associated osteonecrosis. Br J Haematol 2009;144:667–76.

[15] Landesberg R, Cozin M, Cremers S, Woo V, Kousteni S, Sinha S, et al. Inhibition of oral mucosal cell wound healing by bisphosphonates. J Oral Maxillofac Surg 2008;66:839–47.

[16] Sarasquete ME, García-Sanz R, Marín L, Alcoceba M, Chillón MC, Balanzategui A, et al. Bisphosphonate-related osteonecrosis of the jaw is associated with polymorphisms of the cytochrome P450 CYP2C8 in multiple myeloma: a genome-wide single nucleotide polymorphism analysis. Blood 2008;112:2709–12.

[17] Raje N, Woo SB, Hande K, Yap JT, Richardson PG, Vallet S, et al. Clinical, radiographic, and biochemical characterization of multiple myeloma patients with osteonecrosis of the jaw. Clin Cancer Res 2008;14:2387–95.

[18] Allen MR, Burr DB. Mandible matrix necrosis in beagle dogs after 3 years of daily oral bisphosphonate treatment. J Oral Maxillofac Surg 2008;66:987–94.

[19] American Association of Oral and Maxillofacial Surgeons. Position paper on bisphosphonate-related osteonecrosis of the jaws. J Oral Maxillofac Surg 2007;65:

369–76.

[20] Tubiana-Hulin M, Spielmann M, Roux C, Campone M, Zelek L, Gligorov J, et al.

Physiopathology and management of osteonecrosis of the jaws related to bisphosphonate therapy for malignant bone lesions. A French expert panel analysis. Crit Rev Oncol/Hematol 2008.do:10.1016/j.critrevonc.2008.10.009.

[21] Johnsen SE, Svensson KG, Trulson M. Forces applied by anterior and posterior teeth and roles of periodontal afferents during hold-and-split tasks in human subject.

Exp Brain Res 2007;178:126–34.

[22] Endo Y, Nakamura M, Kikuchi T, Shinoda H, Takeda Y, Nitta Y, et al. Aminoalkyl- bisphosphonates, potent inhibitors of bone resorption, induced a prolonged stimulation of histamine synthesis and increase macrophages, granulocytes and osteoclasts in vivo. Calcif Tissue Int 1993;52:248–54.

[23] Plotkin LI, Manolagas SC, Bellido T. Dissociation of the pro-apoptotic effects of bisphosphonates on osteoclasts from their anti-apoptotic effects on osteoblasts/

osteocytes with novel analogs. Bone 2006;39:443–52.

[24] Follet H, Li J, Phipps RJ, Hui S, Condon K, Burr DB. Risedronate and alendronate suppress osteocyte apoptosis following cyclic fatigue loading. Bone 2007;40:1172–7.

[25] Marie P, Hott M, Garba MT. Inhibition of bone matrix apposition by (3-amino-1- hydroxypropylidene)-1,1-bisphosphonate (AHPrBP) in the mouse. Bone 1985;6:

193–9.

[26] Janckila AJ, Slone SP, Lear SC, Martin A, Yam LT. Tartrate-resistant acid phosphatase as an immunohistochemical marker for inﬂammatory macrophages. Am J Clin Pathol 2007;127:556–66.

[27] Vignery A. Osteoclasts and giant cells: macrophage–macrophage fusion mechanism.

Int J Exp Pathol 2000;81:291–304.

[28] Quinn JMW, Morﬁs M, Lam MHC, Elliott J, Kartsogiannis V, Williams ED, et al.

Calcitonin receptor antibodies in the identiﬁcation of osteoclasts. Bone 1999;25:

1–8.

[29] Coxon FP, Thompson K, Roelofs AJ, Ebetino FH, Rogers MJ. Visualizing mineral binding and uptake of bisphosphonate by osteoclasts and non-resorbing cells.

Bone 2008;42:848–60.

[30] Woo SB, Hellstein JW, Kalmar JR. Systematic review: bisphosphonates and osteonecrosis of the jaws. Ann Intern Med 2006;144:753–61.

[31] Stefanik D, Sarin J, Lam T, Levin L, Leboy PS, Akintoye SO. Disparate osteogenic response of mandible and iliac crest bone marrow stromal cells to pamidronate.

Oral Dis 2008;14:465–71.

[32] von Wowern N. Variations in structure within the trabecular bone of the mandible.

Scand J Dent Res 1977;85:613–22.

[33] Saffar JL, Lasfargues JJ, Cherruau M. Alveolar bone and the alveolar process: the socket is never stable. Periodontol 2000;1997(13):76–90.

[34] Mashiba T, Turner CH, Hirano T, Forwood MR, Johnston CC, Burr DB. Effects of suppressed bone turnover by bisphosphonates on microdamage accumulation and biochemical properties in clinically relevant skeletal sites in beagles. Bone 2001;28:524–31.

[35] Allen MR, Iwata K, Phipps R, Burr DB. Alterations in canine vertebral bone turnover, microdamage accumulation, and biochemical properties following 1-year treatment with clinical treatment doses of risedronate or alendronate. Bone 2006;39:

872–9.

(11)

[36] Allen MR, Burr DB. Mineralization, microdamage, and matrix: how bisphosphonates inﬂuence material properties of bone? IBMS BoneKEY-Osteovision 2007;4:49–60. http://www.bonekey-ibms.org/cgi/content/full/ibmske;4/2/49 doi:10.1138/20060248.

[37] Allen MR, Burr DB. Three years of alendronate treatment results in similar levels of vertebral microdamage as after one year of treatment. J Bone Miner Res 2007;22:

1759–65.

[38] Verbogt O, Gibson GJ, Schaffer MB. Loss of osteocyte integrity in association with microdamage and bone remodelling after fatigue in vivo. J Bone Miner Res 2000;15:60–7.

[39] Santini D, Fratto ME, Vincenzi B, La Cesa A, Dianzani C, Tonini G. Bisphosphonate effects in cancer and inﬂammatory diseases: in vitro and in vivo modulation of cytokine activities. Bio Drugs 2004;18:269–78.

[40] Adami S, Bhalla AK, Dorizzi R, Montesani F, Rosini S, Salvagno G, et al. The acute- phase response after bisphosphonate administration. Calcif Tissue Int 1987;41:

326–31.

[41] Deng X, Yu Z, Funayama H, Yamaguchi K, Sasano K, Sugawara S, et al. Histidine decarboxylase-stimulating and inﬂammatory effects of alendronate in mice:

Involvement of mevalonate pathway, TNFα, macrophages and T-cells. Int Immunopharmacol 2007;7:152–61.

[42] Richards PJ, Amos N, Williams AS, Williams BD. Pro-inﬂammatory effects of the aminobisphosphonate ibandronate in vitro and in vivo. Rheumatology 1999;38:

984–91.

[43] Pietschmann P, Stohlawetz P, Brosch S, Steiner G, Smolen JS, Peterlik M. The effect of alendronate on cytokine production, adhesion molecule expression and transendothelial migration of human peripheral blood mononuclear cells. Calcif Tissue Int 1998;63:325–30.

[44] Zysk SP, Dürr HR, Gebhard HH, Schmitt-Sody M, Refior HJ, Messmer K, et al. Effects of ibandronate on inflammation in mouse antigen-induced arthritis. Inflamm Res 2003;52:221–6.

[45] Junquera L, Gallego L. Nonexposed bisphosphonate-related osteonecrosis of the jaws: another clinical variant? J Oral Maxillofac Surg 2008;66:1516–7.

[46] Mawardi H, Treister N, Richardson P, Anderson K, Munshi N, Faiella RA, et al. Sinus tracts: an early sign of bisphosphonate-associated osteonecrosis of the jaws? J Oral Maxillofac Surg 2009;67:593–601.

[47] Reszka AA, Halasy-Nagy J, Rodan GA. Nitrogen-bisphosphonates block retinoblas- toma phosphorylation and cell growth by inhibiting the cholesterol biosynthetic pathway in a keratinocyte model for esophageal irritation. Mol Pharmacol 2001;59:193–202.

[48] Dimopoulos MA, Kastritis E, Anagnostopoulos A, Melekapopoulos I, Gika D, Moulopoulos LA, et al. Osteonecrosis of the jaw in patients with multiple myeloma treated with bisphosphonates: evidence of increased risk after treatment with zoledronic acid. Haematologica 2007;91:968–71.

[49] Diel IJ, Fogelman I, Al-Nawas B, Hoffmeister B, Migliorati C, Gligorov J, et al.

Pathophysiology, risk factor and management of bisphosphonate-associated osteonecrosis of the jaw: is there a diverse relationship of amino- and non- aminobisphosphonates? Crit Rev Oncol Hematol 2007;64:198–207.

[50] Migliorati CA, Siegel MA, Elting LS. Bisphosphonate-associated osteonecrosis: a long-term complication of bisphosphonate treatment. Lancet Oncol 2006;7:

508–14.

[51] Grbic JT, Landesberg R, Lin SQ, Mesenbrink P, Reid IR, Leung PC, et al. Incidence of osteonecrosis of the jaw in women with postmenopausal osteoporosis in the health outcomes and reduced incidence with zoledronic acid once yearly pivotal fracture trial. JADA 2008;139:32–40.