The Expression of Netrin-1 in the Osteoarthritic Temporomandibular Joint


 Subchondral bone degeneration is the main pathological change during temporomandibular joint (TMJ) osteoarthritis (OA) development. Netrin-1, an axon-guiding factor, might play roles in OA development and pain. The purpose of this study was to investigate the expression of Netrin-1 in TMJ OA and its possible role in the progression of TMJ OA and pain. The content of Netrin-1 in the synovial fluid of temporomandibular joint disorders (TMDs) patients was detected by ELISA. After injection of MIA into the TMJ animal model, hematoxylin & eosin (H&E) staining, toluidine blue (TB) staining and micro-CT were performed to detect the change of TMJ. Netrin-1 expression in the condyle was detected by immunohistochemical (IHC) staining. The pain experience was examined by the von Frey test in animal model. The concentration of Netrin-1 increased in the synovial fluid of TMJ OA patients with the VAS scores increased. In the TMJ OA model, the structural disorder of subchondral bone was the most obvious at the 2nd week. In addition, the head withdrawal threshold (HWT) was significantly decreased in the TMJ OA model. Netrin-1 expression increased in the subchondral bone at the 2nd week after MIA injection in the mouse TMJ. Cultivate raw264.7 cells and induce them to differentiate into osteoclasts. Netrin-1 and its receptor Unc5B are detected to be expressed on the osteoclast membrane. Thus, Netrin-1 might be a regulator during degeneration and pain in the process of TMJ OA.


Introduction
Temporomandibular joint (TMJ) osteoarthritis (OA) is a common disease characterized by joint clicking, restriction of mouth opening and pain and has a huge impact on the patients' daily life. Although the underlying mechanism of TMJ OA is not clear, subchondral bone degeneration is the main pathological change during TMJ OA development. 1 In the process of subchondral bone degeneration, on the one hand, the generation of blood vessels causes severe cartilage degradation. On the other hand, the release of various in ammatory factors and the induction of sensitization factors cause peripheral sensitization and central sensitization of TMJ nerves, thus occurring pain.
Activated osteoclasts play a role in subchondral bone degeneration by decreasing bone density and disordering subchondral bone structure. 2 Netrin-1, an axon-guiding protein with chemical attraction and capable of promoting axon elongation, is a secreted protein with strong chemotaxis for axon orientation, cell migration, morphogenesis, and angiogenesis. 3 Netrin-1 has been found to participate in osteoclast differentiation. Osteoclast precursor cells increased the secretion of Netrin-1 and promoted osteoclast differentiation by autocrine or paracrine pathways. 4 The combination of Netrin-1 and its receptor Unc5b stimulates the formation of RANKL-induced osteoclasts in bone marrow macrophages (BMMs). 5 In addition, bronectin leucine-rich transmembrane protein 2 (Flrt2) regulated osteoclast polynucleation by interfering with the Netrin-1-Unc5b interaction. 6 At the same time, Netrin-1 also plays roles in angiogenesis and in ammation by acting as a potent vascular mitogen that can promote the proliferation, migration and adhesion of vascular endothelial cells and smooth muscle cells. 7 In addition, due to the axonal guiding function of Netrin-1, it can induce the growth of nerve axons in the subchondral cartilage bone, thus leading to the excitation of the dorsal root ganglion and hypersensitization of pain in OA mice. 8 Netrin-1 is also highly expressed in the model of osteolysis in ammation induced by wear particles. 9 Furthermore, monoclonal antibody blocking of Netrin-1 and its receptor has been shown to prevent bone destruction and to reduce the severity of arthritis. 9 In terms of TMJ OA, the role of Netrin-1 is not clear. In this study, the expression of Netrin-1 in TMJ OA was revealed, and its possible role in the progression of TMJ OA was studied.

Results
The content of Netrin-1 increased in the synovial uid of human temporomandibular joint osteoarthritis and related to the degree of the temporomandibular joint pain. 24 patients with TMJ articular disk displacement (ADD), 16 patients with TMJ OA were enrolled in this study. The demographic and clinical characteristics of patients were collected (Table 1). Netrin-1 was detected in the synovial uid of TMD patients. The Netrin-1 concentrations were 643.83 ± 97.24 pg/mL in the ADD group, 869.98 ± 100.58 pg/mL in the OA group (Fig. 1A). The VAS scores of the TMJ OA patients were higher (Fig. 1B). And there was a positive linear correlation between the concentration of Netrin-1 and the VAS score in the OA group (n = 16) (Fig. 1C). Bone destruction was observed in the mouse TMJ OA model.
After MIA injection, it was found that compared with the control group, the cartilage thickness of the mice in the TMJ OA group became thinner from the rst week. At the second week, both thinning and thickening of the chondrocyte layer appeared at the same time. There were obvious clusters of chondrocytes. At the third week, the undifferentiated mesenchymal layer was obviously thickened and the cartilage layer structure was disordered. The 4th week, the cartilage thickness was still thinner than that of the control group. And the brosis of the cartilage layer appeared ( Fig. 2A). At this time, the mast cell layer became extremely thin, which indicated that the condylar cartilage layer of the mouse became thin and structurally disordered after the injection of MIA ( Fig. 2A, B).
Meanwhile, the percentage of proteoglycan in the cartilage matrix also decreased signi cantly, indicating the destruction of the cartilage matrix. The chondrocytes in the control group were evenly arranged, and the cartilage was rich in proteoglycan. In the TMJ OA group, the chondrocytes were not aligned in the second and third weeks. The percentage of aggrecan area decreased (Fig. 2C, 2D(a)). It was found that the Mankin score was very high from the rst week until the fourth week in the TMJ OA group, and reached the highest in the second week, indicating that the most serious pathological changes occurred at this time ( Fig. 2D(b)).
The results of Micro-CT showed that the condyles of TMJ OA group were severely damaged and bone remodeling occurred. From the second week of MIA injection, the morphology of the condyle has been signi cantly altered, and the anterior slope of the condyle appears damaged and the surface is rough and uneven (Fig. 2E). Bone white lines are blurred, bone marrow cavity appears hollow, bone destruction was severe at the second and third weeks, trabecular bone sparse and reduced, and condyle surface morphology was remodeled, and bone hyperplasia and resorption can be observed at the same time ( Fig. 2F). At the same time, bone analysis showed that the bone volume fraction (BV/TV) and trabecular bone number (Tb.N) of the TMJ OA group were signi cantly lower than those of the control group at the second week (Fig. 2G).
Netrin-1 expressed on the membrane of the osteoclasts in a mouse TMJ OA model and was associated with TMJ OA pain.
After MIA injection, the head withdrawal threshold was signi cantly decreased at the 1st, 2nd, 3rd and 4th weeks. The head withdrawal threshold decreased signi cantly on the rst day, decreased to the minimum on the seventh day, and then gradually increased. However, there were still marked gaps between the OA group and the control group at the 2nd, 3rd and 4th weeks (Fig. 3). Netrin-1 was strongly expressed in the subchondral bone and no expression in the chondrocyte layer (Fig. 4A). Furthermore, the expression intensity was the highest at the 2nd week (Fig. 4B). Cytology experiment showed that Netrin-1 and its receptor Unc5B expressed on the membrane of the osteoclasts while the other receptor DCC showed no expression (Fig. 4C, 4D).

Discussion
In this study, Netrin-1 was revealed in TMD patients' synovial uids. The content of Netrin-1 in the synovial uid was higher in TMJ OA than in TMJ ADD. Furthermore, the content of Netrin-1 was related to the VAS score. Therefore, we can speculate that as the degree of pain in patients with OA increases, the concentration of Netrin-1 in their synovial uid also elevated. Previous studies also found the expression of Netrin-1 in synovial uid of OA and Rheumatoid arthritis (RA). 10 In addition, it indicated the inhibition of osteoclast multinucleation of Netrin-1 in RA, which can play bone-protective functions in bonedestructive diseases. However, another study showed that Netrin-1 regulated osteoclast differentiation through autocrine/paracrine. 4 In this study, the expression of Netrin-1 is increased in the synovial uid of TMJ OA patients who showed apparent bone destruction of their condyles, which may indicate that Netrin-1 may be more involved in osteoclast differentiation. Netrin-1 in the synovial uid might be mainly secreted from synovial macrophages. Resident synovial macrophages provide a protective barrier for the joint 11 , and recent studies have identi ed macrophage-derived Netrin-1. Macrophage-derived Netrin-1 is critical for neuroangiogenesis in endometriosis 12 and plays a role in vascular diseases 13 . The upregulated expression of Netrin-1 is likely caused by synovial in ammation, which destroys the macrophage barrier in the lining layer of the synovial membrane, and then synovial macrophages secrete Netrin-1, which causes subsequent development of in ammation.
Although the mechanism of TMJ OA is unknown, in ammation combined with angiogenesis, including the articular disc, condylar cartilage and subchondral bone with degenerative changes, is an important pathological feature of OA. 14 Netrin-1 acts as an angiogenic factor, 7 is regulated by infection and in ammatory cytokines and is a novel regulator of vascular endothelial function. 15 Netrin-1 promoted angiogenesis through its receptors Unc5B and DCC. 16 Netrin-1 and the receptor Unc5B play crucial roles in axonal development and angiogenesis. 17 During in ammation, the expression of Netrin-1 is downregulated through the inhibition of the receptor Unc5B, thus blocking its function of inhibiting leukocyte aggregation and leading to leukocyte in ltration. 18 The DCC-dependent ERK1/2-eNOS feedforward mechanism was observed to be the method by which Netrin-1 induces angiogenesis. 19 Netrin-1 and its receptor Unc5B are thought to be novel targets for the treatment of in ammatory arthritis. 9 In this study, the expression of Netrin-1 on the membrane of osteoclasts were found and Unc5B expressed on the membrane of osteoclasts.
Additionally, in the mouse TMJ OA model, the results of the von Frey test showed that the head withdrawal threshold (HWT) of the TMJ OA group was signi cantly lower than that of the control group.
In the TMJ OA model, the mice were more sensitive to pain at 1-2 weeks, which indicated that they may suffered more in the early stage of OA. Currently, the speci c mechanism of TMJ pain is not clear, but most take the view that peripheral sensitization and central sensitization play indispensable roles. 20,21 Studies have shown that peripheral Netrin-1 reduces pro-in ammatory substances and stabilizes the blood-brain barrier, restricting immune cells from entering the central nervous system and playing a role in neuroin ammation. 22 Netrin-1 produced by osteoclast differentiation may induce the growth of sensory nerve axons in subchondral bone. 8 Netrin-1 may induce the growth of sensory axons in the subchondral bone of the TMJ, which are sensitized to in ammatory factors that cause pain.
The destruction of subchondral bone is accompanied by vascular proliferation, and its vascularization may develop deeply in the cartilage layer, damaging the integrity of the cartilage structure and thus promoting the process of TMJ OA. In the early stage of TMJ OA, bone destruction, enlarged bone marrow cavity and decreased bone density, along with cartilage vascular invasion into the deep layer and enhanced activity of osteoclasts, could be revealed. In the late stage of TMJ OA, slow bone formation with inferior mineralization, osteosclerosis and osteophyte formation were the main features. [23][24][25] Early increased bone remodeling is considered to be an indispensable condition for OA. 26 In this study, the pathological structural remodeling process of cartilage and subchondral bone was revealed in a mouse TMJ OA model, and the results were similar to those of a previous study. 27 In this TMJ OA model, typical pathological changes include cartilage layer thinning, structural disorder, destructive absorption of subchondral bone and destruction of condyle morphology. The results of Micro-CT also con rmed the changes in bone density.
In recent years, an increasing number of studies have shown that axon guiding factors and neurotrophic proteins play increasingly important roles in bone metabolism. 28 The neuronal axonal guiding factor family mainly includes four proteins, namely, Slits, Ephrins, Semaphorins and Netrins. Semaphorins participate in bone metabolism, regulate bone homeostasis, and function in both bone absorption and bone formation. 29 The effect works by regulating sensory nerves and in turn inducing bone mass. 30 The bidirectional signaling of the Eph receptor family and the associated ephrin ligands exert functions in osteoblast-osteoclast communication, regulating bone homeostasis. 31 Parathyroid hormone (PTH) induces ephrinB2 in osteoblasts and enhances osteoblastic bone formation. In addition, ephrinA2 reverse signaling into osteoclasts enhances osteoclastogenesis. 32,33 Slit binding with its receptor Robo1/2 inhibits osteoblast differentiation and function. In this study, Netrin-1 was an extracellular expressed protein in the TMJ OA model; more speci cally, it is secreted and is present in the subchondral bone lacunae or bone marrow cavities. The content of Netrin-1 expression also changed with the development of TMJ OA. The expression of Netrin-1 increased in the rst week, was the highest in the second week, and then decreased. The different expression levels of Netrin-1 in lacunae and bone marrow cavities might be closely related to subchondral bone destruction and remodeling. Both osteoclast precursors and osteoclasts may secrete Netrin-1 and affect their differentiation, and the binding of Netrin-1 to its receptor Unc5b may mediate the molecular pathway of osteoclast differentiation. Ntn-1 −/− knockout mice have signi cantly fewer osteoclasts and increased cortical and trabecular bone density compared to wild-type mice. Osteoclast precursor cells can increase the expression of Netrin-1 and its receptor UNC5b and regulate the differentiation of osteoclasts by changing the cytoskeleton, which increases the regulator of the Rho-GEF subfamily (LARG) and repulsive guidance molecule (RGMa) association with Unc5b. 34 Netrin-1 might be a regulator during bone degeneration and pain in the process of TMJ OA. However, the speci c role of Netrin-1 in the progression of TMJ OA needs further study to clarify.

Materials And Methods
Synovial uid collection and enzyme-linked immunosorbent assay (ELISA).
Synovial uids were obtained from 40 temporomandibular disorder patients, and Visual Analog Scales (VSA) scores were recorded according to their own temporomandibular joint pain. The 40 patients were divided into two groups based on their clinical symptoms and imaging ndings. More speci cally, they differed from the position of the TMJ discs and the destruction of the condyle which was much clear through the Cone beam Computer Tomography (CBCT). The two groups respectively are the articular disk displacement (ADD) group (n = 24) and the osteoarthritis (OA) group (n = 16). The method to collect synovial uid was described previously. 35,36 A totals of 2 ml of diluted synovial uid was extracted after normal saline was injected into the TMJ cavity with a 5 ml syringe containing 2ml of normal saline, which was irrigated three times. Informed consent was obtained from the patients, and the protocol was approved by the Ethics Committee, School and Hospital of Stomatology, Wuhan University. The content of Netrin-1 in the synovial uid was measured by a human ELISA kit (HM11241; Bioswamp, Wuhan, China). The method was performed according to the manufacturer's protocol.
Mouse TMJ OA model. 40 eight-week-old male C57BL/6 mice were divided into the control group and the MIA group. After intraperitoneal injection of pentobarbital sodium, the preauricular area of the mice was palpated to feel the zygomatic arch; under this was the TMJ cavity. After the preauricular area and zygomatic arch were disinfected with 70% ethyl alcohol, the needle was injected into the upper front of the junction of the zygomatic arch and the temporal bone with a 1-ml syringe, and the needle was slowly injected backward and downward so that it could be judged that the needle had been inserted into the joint cavity. 23 For the MIA group, the mouse was given a unilateral TMJ injection of 20 µL of MIA with a concentration of 1 mg/50 µL. 27,37 The same amount of saline was injected into the unilateral temporomandibular joint cavity in the control group. Finally, euthanize the mice by intraperitoneal injection of an overdose of sodium pentobarbital solution. The protocol was approved by the Ethics Committee, School and Hospital of Stomatology, Wuhan University.
Von Frey test.
The Von Frey test was used to measure the pain experienced by the mice. 38 The hard-plastic tip was used to stimulate the midpoint of the connection between the eyes and ears of the mouse's head and face. In the process of the experimental operation, the mechanical stimulation intensity increased, and in turn, mouse behavior was observed at the same time. When reactions such as rubbing the mouth or scratching the head were observed, the stimulus intensity (g) was recorded. The operation was repeated ve times every 30 s, and the head withdrawal threshold was recorded. Then, the average value was taken as the TMJ pain threshold.
Every 10 mice which consist of 5 from the control group and 5 from the MIA group were euthanized 1, 2, 3, and 4 weeks after the injection. All condylar specimens were immediately xed with 4% paraformaldehyde for 24 hours. Take four specimens from each group for micro-CT detection after ushing overnight. And then, all the specimens were decalci ed in 10% EDTA for two months, dehydrated and embedded in para n. All 40 specimens were sliced and performed histological staining immunohistochemistry staining.
The condyles were collected and xed with paraformaldehyde for 24 h. After ushing overnight, the condyles were analyzed by micro-CT (SkyScan1176, Germany). The camera pixel size was 12.59 µm at 50 kV and 500 µA. The image pixel size was 9 µm. Three-dimensional reconstruction of the temporomandibular joint condylar head was performed to observe the bone changes, and BV/TV, Tb.Th and Tb.N values were also analyzed to compare the destruction of bone structures.
Histological staining. were measured as previously reported. 39 The Mankin score assesses the integrity of the cartilage surface shown by HE staining, cellularity and the condition of cartilage matrix shown by toluidine blue staining.