Decompression Effects on Suppurative and Sclerosing Osteomyelitis in the Jaw

Background: Osteomyelitis (OM) of the jaw is usually caused by a chronic odontogenic infection. Decompression is the release the intraluminal pressure in the cystic cavity allowing gradual bone growth from the periphery. The aim of this study was to analyze the effectiveness of decompression in an OM jaw model. Methods: A 4-mm-diameter defect was made on mandibles of fourteen Sprague-Dawley rats and inoculated with S. aureus (20 μl of 1x10 7 CFU/ml) injection. Two weeks later, four groups were made as non-treatment (C1), only curettage (C2), curettage and decompression (E1), and curettage and decompression with normal saline irrigation (E2). After four weeks, each group was analyzed. Results: Most micro-CT parameters of C1 and C2 were signicantly lower, and bone mineral density with bone volume was enhanced in E2. E1 and E2 groups in histology showed prominent bone healing with a signicantly high number of osteocytes, E2 had the weakest expression of IL-6 compared to that of C1. TNF-α and OPN were expressed strongly in the E1. Conclusion: Decompression drains induced advanced bone healing compared to that of curettage alone in an OM jaw model. Therefore, it could be recommended to use decompressive drain for the enhancement of jaw OM management.


Introduction
Osteomyelitis (OM) of the jaw is an in ammatory process that starts in the medullary space of the bone and progresses to cortical bone, the Haversian system, periosteum, and overlying soft tissue. This is usually caused by micro-organism infection into the bone tissues due to a trauma or odontogenic infection 1 . The gram positive pathogen Staphylococcus aureus (S. aureus) is the most common OM causative agent in both children and adults 2 . Other reasons include steroids, chemotherapeutic drugs, and bisphosphonates, which have been linked to jaw OM 3 .
The treatment of jaw OM in the literature is classi ed as surgical and non-surgical, while the aim differs depending whether or not bacterial infection is apparent 4 . The universally acknowledged and effectual treatment is a combination of antibiotic therapy and surgery consisting of sequestrectomy, saucerization, decortication, and closed-wound suction irrigation 5 . The surgical therapy approach has three main goals, which includes decompression and drainage of intramedullary pressure and subperiosteal abscesses caused by the osteomyelitic effect, surgical treatment of infected tissue and removal of infectious foci, and grafting healthy bone tissue into the infected area 6 .
By de nition, decompression is a technique that creates a small opening in the cystic wall for drainage that releases intraluminal pressure that causes cystic reduction and permits gradual bone growth from the periphery. This technique has been associated with minimal surgical morbidity, changes in the malignancy environment, prevention of reestablishment of intraluminal pressure, decrease or inhibition of the expression of interleukin (IL) 1α and IL-6 inside odontogenic cysts, maintenance of pulp vitality, preservation of anatomical structures, prevention of pathologic fracture, thickening of the cystic lining, and relatively low risk of recurrence 7 .
Surgical drains are used for decompression effects, to eliminate pooled blood, serum and edema reduction, exudate management, and dead space reduction of the surgical wound by drawing the separated surfaces together 8 . These interventions are classi ed as open and closed with the closed one subdivided as passive or active. Active negative pressure drainage, also known as vacuum-assisted wound closure or negative-pressure wound therapy is a popular method for wound care including limb wounds, soft-tissue defects, chronic OM, osteofascial compartment syndrome, amputation, and replantation. Past research results have shown the effects of using negative pressure wound therapy in the head and neck region that include decreased healing time, less pain, and full drainage effects 9 .
Despite the increasing number of studies of decompression, there are no reported studies of decompressive effects using drains in the management of jaw OM.
The hypothesis of this study was that decompression is more effective with regard to bone healing for the treatment of jaw OM compared with conventional surgical treatment alone. The purpose of this study was to investigate the effectiveness of decompression using a drain compared to management without drainage in a rat model of S. aureus-induced OM using micro-computed tomography (micro-CT) and histopathological analysis.  10 . The animal experiment was conducted at the Institute for Experimental Animals, College of Medicine, SNU, in a laboratory infection room classi ed as for high risk infection studies or infectious studies that use experimental animals (Animal Biosafety Level 2: ABL 2). All animals were maintained in an individually ventilated 12hour light/dark cycle cage system with the temperature ranging from 20-26°C (23 ± 3°C), and were provided rodent food and water ad libitum.

Materials And
The bacterial strain used in our study was S. aureus, the most common causative pathogen for jaw OM 11 . We used a 2°C to 8°C freeze-dried S. aureus subsp. Aureus (ATCC® 29213, Manassas, VA, USA) and a Wichita designated clinical isolate that was provided by the Korean Culture Center of Microorganisms (KCCM®, Seoul, Korea).
The suspended sample containing the S. aureus strain was then inoculated and spread with the spread method into a tryptic soy agar (TSA, BD Difco™, New Jersey, USA) plate medium using a sterilized inoculation loop and cultured in an incubator for 24 hours at 37°C. After incubation, a visible colony of S. aureus formed. To determine bacterial density, we used the direct method of plate count technique (PCT) and the indirect method of turbidometry 12 . The number of bacterial inoculation was determined by PCT, in which the number of colonies formed on the plate medium is proportional to the live bacteria contained in the sample, and the dilution ratio and the number of colonies are calculated by stepwise dilutions.
In the turbidity measurement, as the concentration of bacteria increases, the turbidity (absorbance) increases proportionally, therefore in order to measure turbidity as the actual number of bacteria, a correlation must be obtained. This can be obtained by measuring the number of bacteria with the direct plate count technique in parallel. The bacterial colony was harvested and was washed two times with 1 x PBS by vortexing and by centrifuge. The suspended S. aureus solution was transferred to a new glass cuvette containing 1xPBS and was adjusted to an optical density (OD) of 0.8 using a UV/VIS spectrophotometer (Spectrophotometer, PerkinElmer®, MA, USA) at 600 nm with a clear PBS solution as a control (Fig. 1A). For the study, the The infection with S. aureus was performed using a local inoculation route by injecting the bacterial suspension through the created defect 12 . The inoculation procedure was performed under general anesthesia using 90 mg/kg ketamine (50 mg/ml) (ketamine hydrochloride®; Yuhan Co., Seoul, Korea) + 10 mg/kg xylazine (23.32 mg/ml) (Rompun®; Bayer Korea, Ansan, Korea) that was administered intraperitoneally. The preparations for the surgical procedure including the skin preparation, disinfection, and draping were all performed according to standard protocols (Fig. 1B). An approximately 12 mm fullthickness longitudinal extra-oral incision was made parallel to the inferior border of the right and left side of rat mandibles. Adequate subcutaneous (Fig. 1C), deep fascial and periosteal dissections were performed followed by retraction with forceps (Fig. 1D). Using a low-speed hand piece with 1.2 mm diameter round bur, a bilateral circular 4 mm defect was created in the rat mandible ( Fig. 1E) with copious irrigation. The defect was made from the buccal side, inferior to the incisor tooth root, posterior to the second molar, and at the attachment site of the super cial masseter muscle. Considering the anatomy of the rat and the objective of the study, a circular 4 mm defect is a generally accepted mandibular bone defect 13 .

Grouping and experimental design
The animals were randomly divided into control (non-decompression groups, C1 and C2) and experimental groups (decompression groups, E1 and E2) ( Table 1). The C1 group (n = 3) served as the control group, who only received wound closure. The C2 control group (n = 4) received conventional surgical curettage for jaw OM. The experimental groups were further classi ed into two subgroups, E1 group (n = 3), which received removal of pus and necrotic bone tissue and curettage, followed by introducing the tube drain and E2 group (n = 4), which received removal of pus and necrotic bone tissue and curettage, drain insertion and irrigation with normal saline every week.
Blood samples were collected from the tail vein pre-infection, 1-week post-infection, 2-weeks postinfection/the start of the treatment, 1-week after treatment, and 4-weeks after the treatment, and rat weights were checked.
Surgical treatment was performed under general anesthesia as previously described. Two mm inner diameter and three mm outer diameter silicone tubes that were approximately 2 cm in length (Daihan scienti c Co. Ltd., Wonju, Korea) were used as a drain. The length of the tube was adjusted to each animal according to the post-curettage conditions and these were sutured in place using 4 − 0 silk sutures. To keep the tubes intact and in place, we used a plastic collar to prevent scratching and accidental displacement of the draining tubes. The animals that died during the experimental trial were recorded for weight loss and clinical symptoms of OM of the jaw.
Upon the completion of the six-week experimental trial, the animals were euthanized by CO 2 inhalation.
The mandibles of the rats were immediately harvested and carefully isolated.

The analysis of bone healing with micro-CT
The rat mandible specimens were subjected to high-resolution micro-CT scanning Skyscan 1172® (Bruker, Kontich, Belgium). The scanning parameters of the source were adjusted to an Al lter of 0.5 mm, source voltage of 70 kV, source current of 141 µA, and 360° rotations at 0.4° rotation steps. This resulted in images that were 496 pixels in width and 900 pixels in height.

The histological and immunohistological analysis of OM healing
The samples from each group were trimmed and decalci ed with 0.5 M ethylene diamine tetra-acetic acid EDTA® (pH 8.0) (Biosesang, Seongnam, Korea) solution for ten days, dehydrated with 70% ethanol, and embedded into para n. The 4 µm thick slides were then washed with xylene for approximately 10 minutes and were stained with hematoxylin and eosin (H&E) and Masson's trichrome (MT). The histological slides were then scanned with a 3D Scan Panoramic Histech scanner® (3D HistechKft. Budapest, Hungary) and examined using CaseViewer® (version: 2.0, 3DHISTECH, Budapest, Hungary).
For quantitative analysis, the number of osteocytes and Haversian canals within the regenerated bone tissues of the defect area were counted. The 4mm circular defect area was determined from the micro-CT 3D images. An area of interest using a xed rectangular form of 350 x 300 µm within the initial defect area was established by the histological imaging program CaseViewer® (version: 2.0, 3DHISTECH, Budapest, Hungary) in all of the specimens at a magni cation of 20x (Fig. 3).

Statistical analysis
Means and standard deviations for bone healing parameters were obtained. The data normal distribution was tested by Shapiro-Wilk test and showed homogeneity. The differences between groups were tested by ANOVA followed by Tukey-Kramer multiple comparison tests. Statistical analyses were done using SPSS 25.0® (SPSS Software Company, Chicago, USA). P < 0.05 were considered statistically signi cant.

Establishment of an S. aureus-infected jaw osteomyelitis rat model
The pathogen dose of 20 µl of 10 7 CFU/ml was effective in creating jaw OM in the rat model and a repeatable animal model was established. After two weeks of infection, all groups showed visible clinical manifestation of infectious jaw OM including: skin redness, swelling, purulent discharge and alopecia. Six animals died after two weeks of infection due to OM complication. Consequently, after six weeks, the specimens of the remaining animals were collected and were further analyzed. All of the animals were in good health before the infection. The clinical ndings showed common characteristics of the jaw osteomyelitis. The establishment of S. aureus-infected jaw osteomyelitis rat model were con rmed by the following parameters: clinical ndings, blood test, micro-CT bone architecture ndings, and histological analysis.

Clinical evaluation with blood test
After the infection with S. aureus all animals from the control and the experimental groups showed weight loss. Signi cant weight loss in all groups was observed with an average of -31.57 ± 21.98 g at one week after the infection that was recovered after two weeks (p < 0.05). There was no statistical signi cance in weight loss between the groups at 1 week after infection (Table 2). The weight changes are depicted as mean ± standard deviation (SD).
The neutrophil count was signi cantly increased after the infection in all groups. There were no statistically signi cant differences in neutrophil count between the groups at one week and four weeks after treatment ( Table 3). The white blood cell count (WBC) was signi cantly increased in all groups after infection and was recovered to the normal range at four weeks after treatment. No signi cant difference was observed between the groups at one week after treatment and four weeks after treatment (Table 4). Serum levels of alkaline phosphatase (ALP) were also measured and analyzed. At one week after infection, ALP levels were signi cantly increased in all groups and a signi cant reduction was observed at one week and two weeks after treatment. No signi cant differences were found between groups (Table 5). The changes are depicted as mean ± standard deviation (SD). The changes are depicted as mean ± standard deviation (SD). The changes are depicted as mean ± standard deviation (SD).

Micro-CT results of bone healing
From the 3D images, more bone healing was observed in the E1 and E2 groups, where the initial bone defect was replaced by new bone tissue. The E2 group had the most compact bone formation compared to the other groups. The C1 group, which received no treatment, showed bone destruction that continuously spread from the initial defect affecting a wider area. The common characteristics of osteomyelitis of the jaw such as the bone necrosis and sequestrum formation, a segment of necrotic bone that is separated from the viable bone by a granulation tissue and bone resorption was observed from the control groups.
The BMD results were signi cantly different between the groups. The BMD in the C1 group was signi cantly lower compared to that of the E2 group, with a mean difference of -0.45 g/cm 3 (p < 0.05) ( Table 6) (Fig. 4A).  The data is depicted as mean ± standard deviation (SD). *The osteocytes found in the E2 group were signi cantly higher than in the C1 and C2 groups, p < 0.05).
The BV was highest among E1 and E2 groups. The E2 group showed signi cantly higher results compared to that of the C1 group (p < 0.05) with a mean value of 0.73 ± 0.08 mm 3 (p < 0.05) (Fig. 4B). The BV/VOI parameter was signi cantly higher in the E2 group, with an average value of 75.70 ± 14.32 % compared to that of the C1 group (Fig. 4C). The BS, BS/VOI, Tb.N, Tb.Sp parameters were signi cantly different between the control and experimental groups ( Table 6). Most bone healing parameters were lower in the control groups compared to the experimental groups (Figs. 5A-C). However, the Tb.Sp in the C1 group was signi cantly higher than that in the experimental groups (Fig. 5D). The Tb.Th was not signi cantly different between the groups.

Histological and immunohistochemical results of OM healing
The morphological changes in bone healing were macroscopically observed in the H&E and MT stained slides at 4-weeks in the defect area.
The C1 group showed high grade in ammatory in ltration consisting of neutrophils, eosinophils, and macrophages around the bacterial colonies. Signs of infective osteomyelitis including bone necrosis, bone resorption and destruction, with no bone healing were observed. In MT staining, the C1 group were stained with thick blue color, indicating old bone, while no new bone formation stained with bright blue color were observed (Figs. 6A1-A6). Histological ndings for the C2 group showed bone healing with osteoblastic cell lining in the parenchymal tissue found at the center of the defect area with evidence of in ammatory in ltrates (Figs. 6B1-B6).
The histology features of the E1 group included scattered lymphocytic in ammatory in ltrates and loose marrow brosis. In the MT stain, new bone formation stained with bright blue color and new blood vessels were observed (Figs. 6C1-C6). The E2 group showed active bone remodeling with the thickest and most compact new bone formation being in the defect area compared to that of other groups. Increased osteophytic bone formation was observed. Furthermore, an increased number of Haversian canals with osteoblast rimming and new blood vessel formation stained with thick red color by the MT stain were seen (Figs. 6D1-D6).
We counted the number of osteocytes and Haversian canals in the ROI for quantitative analysis. The results showed that the E2 groups had a statistically signi cant greater osteocyte count compared to the control groups (p < 0.05) ( Table 7).
In order to con rm the in ammatory, angiogenic, and osteogenic properties in the control and experimental groups, IHC staining was performed. The expression of in ammation-related antibody IL-6 in the E2 group was weak (score "2": 1-25% of cells positive), compared with that of the C1 group ("5": 76-100% cells stained). TGF-β1 expression was markedly high in the E1 group, while the C1 group showed no expression ("1": none). The TNF-α antibody stained strongly in the C1 and C2 groups compared to that in the other groups ("5": 76-100% cells stained).

Discussion
The standard regimen for jaw OM treatment has been well established in the scienti c literature, which consists of early diagnosis, elimination of infection source, establishment of surgical drainage, bacteriological identi cation, antibiotic therapy, surgical treatment, supportive treatment, and lastly reconstruction 16 . There are a variety of conventional treatment options that include antibiotics, nonsteroidal anti-in ammatory drugs, and hyperbaric oxygen therapy that is used as an adjunctive treatment, but the disease still has a high recurrence rate and life-threatening complications 1,17,18 . Therefore, it is important to establish an effective treatment method for jaw OM.
In the oral and maxillofacial surgery eld, evidence suggests that active and passive decompression promotes osteogenesis, stimulates angiogenesis, creates short-term hypoxia, and exerts forces capable of removing the cystic lining and its contents as a de nitive treatment for odontogenic cysts of the jaw 7 .
According to an observational study by Lin et al., negative pressure drainage for oral and maxillofacial surgery patients found that its application can hasten wound healing, decrease hospital stay length, improve patient comfort, and decrease clinician workload 19 . Also, an animal study by Zhang et al. found that intermittent negative pressure reduced bone healing time and enhanced bone regeneration by enhancing the expression of VEGF and bone morphogenetic protein 2 20 . Nakahashi et al. reported that closed continuous irrigation-suction treatment was effective not only for chronic suppurative jaw OM but also for the more intractable sclerosing type of OM after following complete removal of the affected tissues 21 . Based on our current animal study and previous reports, decompression effects with drainage have several important therapeutic effects including reduced treatment time, enhanced bone regeneration, and reduced complication and recurrence rates.
The signi cance of this study is that it demonstrates the effectiveness of decompression using a drain in jaw OM, which had signi cant bone healing effects according to micro-CT, histology, and IHC analyses.
For developing new therapeutic methods for OM of the jaw, it is important to establish standard operative protocols for animal modeling 22 . To our knowledge, there are no scienti c data in the literature on decompression effects using drain in jaw OM animal model. Also, the current S. aureus-infected jaw osteomyelitis rat model has not been previously described in the literature before.
The mechanisms of decompression and drainage in bone healing can be explained by uid removal and alteration of the wound environment to be conducive to healing. Excess uid build-up is regarded as one of the major factors that compromise healing, partly owing to the compressive pressure that it exerts on local cells and surrounding tissue. Each cell has an intrinsic tension exerted by the interactions between its cytoskeleton and the extracellular matrix, that induce a proliferative response. If the uid pressure is elevated in the interstitium, the proliferative response diminishes due to dampened intrinsic tension buildup. The uid from the extracellular space is thought to communicate with the wound surface.
Applying decompression and drainage to this area permits uid removal from the extracellular space 23 .
The removal of postoperative uid allows decompression of the microvasculature that permits tissue perfusion by reducing pressure and enhancing blood circulation to the area. It will also remove the toxins, in ammatory exudate, and pathogenic bacteria from the operative site, which is considered to be an important element in the wound healing process.
Micro-CT analysis tools provided highly reproducible methodology with qualitative and quantitative assessment of bone microstructure. Micro-CT is considered to be a gold standard method for assessing mineral density, bone morphology, and bone micro-architecture under various pathologic conditions 24,25 .
We report our methodology according to the guidelines for assessment of bone microstructure in rodents using micro-CT 26 and results by following morphometric indices that can determine new bone formation 27 . The most informative parameters which show the course of bone healing are BV, BV/VOI, BS/BV, and BMD 25,27 . The BV/VOI in the E2 group showed a mean value of 11.44 ± 2.68 %, indicating the highest amount of bone regeneration that was signi cantly higher than that seen in the control groups.
The Tb.N was the highest in the E2 group at 3.18 ± 1.101/mm, while the Tb/Sp was lowest among all the other groups at 0.13 ± 0.06 mm, indicating more bone closure. The volume of mineralized bone tissue was the highest in the E2 group at 0.73 ± 0.16 mm 3 , which was signi cantly higher than that seen in the control groups. Bone healing is characterized by a gradual increase in BV, BMD, and a decrease in BS/BV, and the volume of pore space. The E1 and E2 groups had the highest values for BV and BMD, and a decreased value for BS/BV. The 3D evaluation showed more rapid bone healing was detected in the E1 and E2 groups compared to that of the positive control group C2 (Fig. 8). Micro-CT analysis showed more favorable outcomes in bone healing after decompression treatment.
In terms of bone healing and anti-in ammation, the experimental groups showed more effective results that included rapid bone healing, blood vessel formation, and reduced in ammation. The E2 group also had the highest osteocyte count. An increase in osteocytes plays a pivotal role in regulating bone turnover, which also enhances osteogenesis of stem cells, suggesting an important role in tissue regeneration 28 . The histological results suggested the most optimal bony healing was seen in the experimental groups, which is in accordance with the micro-CT analysis.
VEGF-A is thought to play an important role in regulating angiogenesis as well as bone development and regeneration. Angiogenesis and osteogenesis are two intimately connected processes that must be closely coupled to permit physiological bone function. In fact, alterations in vascular growth can alter the physiological bone healing process, which may lead to osteoporosis, osteonecrosis, and non-union fractures 29 . According to a previous clinical study, decompression had a direct in uence on the microvascular circulation, and enhanced VEGF protein in the rst day following surgical treatment, thereby activating osteogenesis-related proteins such as OPN and ALP was decreased on the second day 30 . In our study we demonstrated that VEGF was activated and reduced in the E2 group, leading to enhanced and accelerated bone healing compared to the E1 group, while bone healing was evident, but much slower.
In the bone remodeling phase, TNF-α and other pro-in ammatory cytokines are thought to play an important role in bone healing 31 . IL-6 is a well-known cytokine that stimulate osteoclast differentiation and bone resorption indirectly depending on the context of release 32 . TNF-α is a widely known key player in the pathogenesis of osteomyelitis. Yokota et al. reported that the combination of IL-6 and TNF-α differentiated tartrate-resistant acid phosphatase-positive multinucleated cells, that resembles boneresorbing cells. There osteoclast-like cells are easily inhibited by extracellular signal-regulated kinase unlike the osteoclasts. Therefore, these cells can be called as "in ammatory osteoclast like cells" 33 . In IHC staining, the pro-in ammatory antibodies IL-6 and TNF-α were strongly stained in the C1 group, indicating an in ammatory reaction to the bacteria, while the groups which received surgical treatment were weakly stained. From the Micro-CT and histology analysis, C1 and C2 groups exhibited the weakest bone healing with strong in ammatory in ltration, which could be explained by the high expression of IL-6 found in C1, and high expression of TNF-α in the C1 and C2 groups. The high expression of IL-6 and TNF-α found in these groups had active bone resorptive effects, which suppressed the osteogenesis.
TGF-β is an important cytokine that balances bone formation and bone resorption, mineral storage, hematopoietic cell generation and osteoimmunology 34,35 . In the immune system, TGF-β modulates in ammatory responses by regulation of chemotaxis, activation and survival of macrophages, natural killer cells, mast cells, granulocytes and dendritic cells 35 . Especially, high concentration of TGF-β1 reduces the activation of osteoclasts, while low concentrations promotes osteoclast maturation 36 . The TGF-β1 was expressed in high expression in the E1 group. Previous studies suggest that TGF-β1 prevents TNF-α-induced bone destruction by suppressing effecter T cell function 35 . TGF-β1 also play an important role in vasculogenesis, induced by VEGF-mediated apoptosis 37 . The high expression of TGF-β1 found in the E1 group indicate active angiogenic function of decompression as well reduced activation of osteoclasts and bone resorption. Compared to E1 group, the E2 group showed well-formed cortical bone with less marrow bone. This could also explain the weak stain of angiogenesis and osteogenesis related antibodies, since adequate and compact bone healing was already established compared to that of other groups.
The results were in accordance with our hypothesis that decompression using a drain had signi cant therapeutic effects on bone regeneration for jaw OM. When decompression was applied to the curettage treatment, enhanced wound and bone healing were achieved. The decompression effects on the healing process were much enhanced with weekly normal saline irrigation.
Therefore, it could be recommended to implement decompression using a drain as a treatment in combination with surgical treatment to allow accelerated bone healing.

Declarations
Compliance with Ethical Standards -This study had no funding.
-There are no con icts of interest in this article -Ethical approval: The experimental protocols were approved by the Seoul National University (SNU) Institutional Animal Care and Use Committee (SNU-121123-12-11) and Institutional Biosafety Committee of SNU (SNUIBC-R121226-1-6). The experiment was in accordance with the "Recommendations for handling of Laboratory Animals for Biomedical Research" and complied with the Committee on Safety and ethical Handling Regulations for Laboratory Experiments at SNU. Animal studies were conducted following the ARRIVE guidelines and are in accordance with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Data availability -The datasets generated during and/or analyzed by the authors during this study are available from the corresponding author on reasonable request.