Bisphosphonate-related osteonecrosis of the jaws (BRONJ) occurs subsequent to intravenous and oral bisphosphonate exposure in a small subset of patients. The identification of the pathophysiologic mechanisms has not been fully elucidated. Evidence of concurrent bacterial colonization at sites of bone necrosis, previous reports of neutrophil-related complications in some patients taking some bisphosphonates, along with perturbed neutrophil function in bisphosphonate-treated mice, suggest an innate immune role in the development of BRONJ. This study investigated neutrophil function in BRONJ patients to determine if neutrophil functional defects may serve as a potential biomarker for BRONJ susceptibility. Two populations were studied: patients with BRONJ and those beginning intravenous pamidronate. Healthy control patients were used for comparison. Twenty-three patients with BRONJ and five patients who were beginning pamidronate therapy provided neutrophil samples from the mouth (oral rinses) and from blood. Neutrophils from the population of patients with BRONJ and from those post-pamidronate treatment showed lower reactive-oxygen species production and impaired chemotaxis relative to controls. These data suggest that a compromise in neutrophil function may be a potential biomarker for BRONJ susceptibility.
Neutrophils are the most numerous and critical cellular elements of the innate immune system. They are important surveyors in the oral cavity and protect against bacterial and fungal invasion, helping to maintain a healthy equilibrium by recognizing, phagocytosing and killing microorganisms. Some patients who are taking bisphosphonate medications may have neutrophils that have become altered by the drug. The neutrophils in bisphosphonate-exposed mice have demonstrated dampened chemotaxis and reactive oxygen species (ROS)-killing. This alteration in neutrophil function may help to explain why a subset of patients on bisphosphonate medications have chronic soft-tissue defects that may develop into bisphosphonate-related osteonecrosis of the jaws (BRONJ).
Neutrophils are able to hone in on foreign targets via cell-surface receptors, including those for leukotriene B4 and CXC, leading to ligation, phagocytosis and oxidative responses. Other cell-surface receptors, such as the N -formyl peptide (formyl-Met-Leu-Phe; fMLP) receptor, function as chemotactic receptors by binding bacterial N -formylated peptides, which are potent pro-inflammatory mediators. Upon target identification and phagocytosis, the enzyme-filled lysosome of the neutrophil fuses with the phagosome, leading to pathogen destruction. Part of the killing mechanism involves activation of the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase complex, which yields a toxic respiratory burst of hydrogen peroxide (H 2 O 2 ), the superoxide anion O 2− and nitric oxide (NO).
In samples from BRONJ sites, bacterial colonization and the presence of Actinomyces species are common. While factors such as microtrauma, alteration in bone turnover and soft and hard tissue toxicity may contribute in the pathogenesis of BRONJ, bacterial challenge and an inadequate immune response are important players among the host of influences in BRONJ development. In order to test this hypothesis, oral and peripheral blood neutrophils were collected from patients with previous bisphosphonate exposure [either intravenous (IV) or by mouth (PO)] and BRONJ and from patients without BRONJ who had just initiated an IV pamidronate regimen. Oral neutrophils were sampled since this technique has proven useful in monitoring neutrophil mobilization to peripheral sites and functionality after transplantation of hematopoietic stem cells. These neutrophils were tested for their ability to form ROS and their chemotactic ability in response to a stimulus. Additionally, sampling oral neutrophils in parallel to peripheral blood may help demonstrate whether the neutrophil issue is local. The data presented here support the idea that neutrophil function may be an important predictor and/or biomarker for BRONJ susceptibility.
Materials and methods
Study population 1: patients with BRONJ
Eighteen patients (11 females and seven males; mean age 71.2 years, range 57–90 years) with a diagnosis of BRONJ, attending regular follow-up appointments at a hospital oral and maxillofacial surgery clinic, were enrolled in this study between December 2010 and August 2012 for ROS assay testing. An additional five patients were identified as candidates for the study but declined to participate. A medical history was collected for all participants, and clinical histories were reviewed for staging of BRONJ. The severity was staged according to the classification system outlined in the American Association of Oral and Maxillofacial Surgeons (AAOMS) 2009 BRONJ Position Paper. As part of an initial small pilot study at Sunnybrook Hospital (carried out by CF as part of her MSc. thesis) five patients with a diagnosis of BRONJ who were similarly attending another hospital dentistry clinic were also enrolled to participate in this study for neutrophil chemotaxis testing. Three patients with a diagnosis of osteoradionecrosis (ORN) with no bisphosphonate exposure were enrolled as controls for a background of chronic inflammation. All candidates were asked for their consent to participate in the study. Patients unable to provide informed consent were excluded.
Blood and oral rinse samples were collected from subjects, with an additional 22 volunteers serving as healthy controls for the ROS assays and five for chemotaxis testing. The healthy control subjects had no prior exposure to bisphosphonate treatment. The healthy controls were not age-matched; neutrophil functions including phagocytosis, ROS generation, and superoxide dismutase (SOD) activity are not significantly affected by ageing, as shown by Niwa et al.
Study population 2: patients on intravenous bisphosphonate therapy
Patients referred to a separate hospital dentistry clinic for pre-pamidronate dental evaluation were recruited to participate in the study. Five patients (all males) met the inclusion criteria of a recent diagnosis of multiple myeloma, no previous bisphosphonate treatment or history of radiation to the jaw and consent to participate. Patients who were unable to provide informed consent were excluded. Patients were enrolled between November 2011 and July 2012 (mean age 65.2 years, range 53–83 years). Four patients provided pre-pamidronate saliva samples and four provided post-pamidronate saliva samples, one of which was processed but was discarded due to experimental error. Two patients provided pre-pamidronate blood samples, which reflected their concurrent medications (notably dexamethasone and bortezomib chemotherapy), and four patients provided post-pamidronate blood samples. Control samples were processed in concert with the patient samples.
Blood sample neutrophil preparation
Blood samples were collected and immediately transported to the laboratory for quantification of ROS production or evaluation of neutrophil chemotaxis. The samples were coded and blinded to the evaluator. A blood sample from a healthy control volunteer was collected each day that samples were collected from a patient with BRONJ. In order to isolate neutrophils, the 10-ml blood samples were drawn into a citrate-containing vacutainer (Sodium Citrate Vacutainer; Becton Dickinson, Rutherford, NJ, USA). Neutrophils were isolated using a one-step neutrophil isolation solution (1-Step Polymorphs; Accurate Chemical and Science Corp., NY, USA). A 4.0-ml aliquot of blood was layered carefully over 4.0 ml of this solution (polymorphoprep: sodium metrizoate, 13.8% (weight/volume (wt/vol)) and dextran 500, 8.0% (wt/vol)), and the mixture was centrifuged at 1600 rpm for 30 min at 4 °C. The lower two pink translucent bands containing neutrophils were collected and washed in Hanks’ balanced salt solution (HBSS; University of Toronto Media Preparation Services, Toronto, ON, Canada) and centrifuged at 2500 rpm for 5 min. Finally, a wash in 1 ml distilled water at 2500 rpm for 5 min served to lyse the remaining erythrocytes and concentrate the neutrophil pellet. The pellet was resuspended in 1 ml HBSS, and neutrophils were counted with a hemocytometer. This method yielded neutrophils with >95% cell viability as determined by haematoxylin and eosin staining.
Oral rinse neutrophil collection and isolation
Oral neutrophils were collected using a rinse of 3 ml of sterile 0.9% normal saline. The patients were instructed to rinse with this solution for 30 s and expectorate into a collection tube. This was repeated three times, with 1-min intervals separating the rinses. The samples were processed via sequential filtration starting with a 40 μm nylon net filter (Millipore, Canada), then a 20 μm and finally a 10 μm filter. The collected cells were then washed and centrifuged for 10 min at 3500 rpm. They were re-suspended in 0.5 ml of HBSS, and neutrophils were counted with a hemocytometer as with the peripheral blood neutrophils.
Neutrophil stimulation and superoxide formation
A neutrophil suspension at 1 × 10 7 /ml concentration in phosphate buffered saline (PBS; University of Toronto Media Preparation Services) with 10 mM d -glucose (University of Toronto Media Preparation Services) was made and kept on ice. Aliquots of 0.1 ml (1 × 10 6 cells) of neutrophil stock were combined with 1.76 ml of PiCM-G buffer (University of Toronto Media Preparation Services) and 10 μl of equine horse ferricytochrome c (0.1 mM final; Sigma–Aldrich Chemical Co.) in the sample cuvettes. The reference cuvette was prepared in the same fashion, and the PiCM-G buffer was replaced with 10 μl (total 100 μg) of a 5 mg/ml concentration of SOD (bovine erythrocyte, Sigma–Aldrich Chemical Co.). The cuvettes were then incubated at 37 °C for 10 min on a shaker prior to neutrophil stimulation. For each sample, one cuvette was stimulated with 10 μg/ml phorbol myristate acetate (PMA) and another with 10 μM fMLP (Sigma–Aldrich Chemical Co.). The stimuli were added simultaneously to the cuvettes (10 μl) and the time of the addition was noted. The rate of superoxide formation was quantified using a spectrophotometer set to 550 nm with a head-on photomultiplier tube; the absorbance of the reduced cytochrome c was measured at 10 min and then at 30 min when the rate of increased absorbance tapered off substantially.
Blood neutrophil chemotaxis
To assess activation and detect a chemotactic defect, isolated neutrophils were placed into the upper well of a trans-well chamber, separated from the bottom chamber – containing vehicle alone or 10 −6 M fMLP and a glass cover slip – by a membrane with 3-μm pores (Costar ® 3-μm pore membrane; Corning Inc., Lowell, MA, USA). The trans-well plates were placed at 37 °C for 1 h and cells migrating through the membrane and onto the bottom cover slip were washed once and quantified. Chemotaxis was assayed by counting the mean number of cells per field over six fields.
To measure the ROS formation capacity of the neutrophils, the spectrophotometric value was measured after the neutrophils were exposed to a stimulus (either fMLP or PMA) and compared to the value of neutrophils exposed to no stimulus (sham). By dividing the ROS yield after stimulation by the sham control, a fold change in oxidase production is generated. This was the number used for comparison between groups. Statistical significance was assessed using the Student’s t -test. A P -value of <0.05 was considered to be significant. Data are expressed as mean ± standard error of the mean.
Study population 1: patients with BRONJ
Twenty-three patients were candidates for the study at a hospital-based oral and maxillofacial surgery clinic. Eighteen patients were recruited for the study with a mean age of 71.2 years ( Table 1 ) and five patients declined to participate.
|Age, years a||Gender b||Stage c||Drug type and duration d||Last BP dose e||Concomitant risk factors f|
|57||F||0||Pamidronate IV, 1 year||3 months||Previous smoking (quit); steroid|
|59||M||2||Risedronate PO, 10 years; alendronic acid PO, 1 year||9 months||Previous smoking (quit)|
|60||F||0||Risedronate PO, 5 years||6 months||–|
|60||F||1||Alendronic acid PO, 6 years||1.8 years||Steroid|
|61||F||2||Alendronic acid PO, 6 years||8 months||Steroid|
|63||M||1||Pamidronate IV, 7 years||6 years||–|
|67||M||2||Zoledronic acid IV, 4 years||Ongoing||Diabetes mellitus|
|70||F||2||Alendronic acid PO, 5.5 years||6 months||Steroid|
|71||M||2||Pamidronate IV, 4 years||2.8 years||Diabetes mellitus|
|74||F||1||Risedronate PO, 5 years||1.7 years||Diabetes mellitus|
|74||F||0||Alendronic acid PO, 10 years||1 year||–|
|75||F||3||Risedronate PO, 6.5 years||1 month||Previous smoking|
|77||F||3||Alendronic acid PO, 5.5 years||1.8 years||–|
|77||M||1||Zoledronic acid IV 1.5 years||10 months||Steroid|
|78||F||2||Pamidronate IV, 8 months; zoledronic acid IV||1 week||Steroid|
|84||F||0||Alendronic acid PO, 3 years||6.8 years||–|
|85||M||0||Alendronic acid PO, 3 years||4.2 years||Diabetes mellitus; steroid|
|90||M||2||Pamidronate IV, 4 years||2 years||–|
At a separate hospital dentistry clinic, blood samples were taken from five patients diagnosed with BRONJ, three patients with diagnosed ORN and three control patients.
Peripheral blood neutrophils: respiratory burst
The samples from patients with a diagnosis of BRONJ showed lower levels of oxygen radicals after stimulation with 10 6 M fMLP or when stimulated with 10 5 M PMA in standard PBS in comparison to the respective bisphosphonate-naïve control group ( Fig. 1 ). In comparison to controls, ROS generation was an average of 47% and 34% lower for the BRONJ samples stimulated with fMLP and PMA, respectively. This difference was statistically significant for the PMA-stimulated samples ( P < 0.05). Two severities of BRONJ existed in the patient pool sampled: mild–moderate (stages 0–1) and moderate–severe (stages 2–3). There was no statistically significant difference in ROS generation between the mild–moderate and the moderate–severe groups ( Fig. 2 ).
Oral neutrophils: respiratory burst
The average ROS yield from the BRONJ oral neutrophils was less than that of the healthy bisphosphonate-naïve oral neutrophils when tested with both fMLP and PMA ( Fig. 3 ). These findings echo the blood neutrophil results ( Figs 1 and 2 ), and were statistically significant for the neutrophils stimulated with PMA ( P < 0.05). No statistically significant relationship existed between the ROS yield and the severity of BRONJ (fMLP P = 0.1342, PMA P = 0.2111) ( Fig. 4 ).
Peripheral blood neutrophils: chemotaxis
Peripheral blood neutrophils from patients with BRONJ, ORN and healthy controls were isolated and their ability to migrate through a porous membrane was measured using fMLP as a chemoattractant in a Boyden chamber. Significantly fewer numbers of neutrophils from BRONJ patients were observed to migrate towards the fMLP stimulus compared with control neutrophils and ORN neutrophils ( P < 0.05) ( Fig. 5 ).