Molecular changes in peripheral blood involving osteoarthritic joint remodelling

Abstract Biomarkers of temporomandibular joint (TMJ) osteoarthritis (OA) remain unknown. The objective was to detect whether molecular biomarkers from peripheral blood leucocytes (PBLs) engage in TMJ OA lesions. Thirty‐four six‐week‐old Sprague Dawley rats were used. The top upregulated gene ontology categories and gene‐fold changes in PBLs were detected by a microarray analysis comparing rats that received 20‐week unilateral anterior crossbite (UAC) treatment with age‐matched controls (n = 4). Twenty weeks of UAC treatment had been reported to induce TMJ OA‐like lesions. The other twenty‐four rats were randomly placed in the UAC and control groups at 12‐ and 20‐week time points (n = 6). The mRNA expression levels of the selected biomarkers derived from the microarray analysis and their protein expression in the alveolar bone and TMJ were detected. The microarray analysis indicated that the three most highly involved genes in PBLs were Egr1, Ephx1 and Il10, which were confirmed by real‐time PCR detection. The increased protein expression levels of the three detected molecules were demonstrated in cartilage and subchondral bone (P < 0.05), and increased levels of EPHX1 were reported in discs (P < 0.05); however, increased levels were not present in the alveolar bone. Immunohistochemistry revealed the increased distribution of EGR1‐positive, EXPH1‐positive and IL10‐positive cells predominantly in the osteochondral interface, with EXPH1 also present in TMJ discs. In conclusion, the increased mRNA expression of Egr1, Ephx1 and Il10 in PBLs may serve as potential biomarkers for developed osteoarthritic lesions relating to osteochondral interface hardness changes induced by dental biomechanical stimulation.


| BACKG ROU N D
Osteoarthritis (OA) is a widespread joint problem that is closely associated with biomechanical factors. 1 The temporomandibular joint (TMJ), which is biomechanically associated with dental occlusions, is often insulted by osteoarthritis. 2 The diagnosis of OA is often based on osseous changes on TMJ images, such as those revealed by X-ray and computerised tomographic (CT) scanning. 3,4 Blood cells flow all around the body and respond to internal and external signals; therefore, changes in TMJs should be reflected in the cells in the peripheral blood. 5 Whether the OA changes in the TMJs induced by dental occlusion can be indicated by biological markers in peripheral blood leucocytes (PBLs) is an area of interest.
Currently, the microarray technique has advanced to be able to detect the relative expression levels of potentially interesting genes in a single experiment. 5,6 This technology has been used in studies on pain-related diseases including oro-facial pain 7 and in the biological activities of chondrocytes, such as those observed to respond to an in vitro stimulator and in comparisons of samples from healthy and osteoarthritic joints. 8,9 Thus, there is a possibility of gene detection in PBLs for the TMJ OA lesions; however, this possibility remains unexplored.
Animal studies are generally accepted for human disease investigations. Many studies have been performed using animal models, including mice and rats, to alter dental occlusions to study TMJ OA. [10][11][12] Based on a gene array analysis, Appleton et al indicated that the gene expression changes in a surgically developed rat OA model were similar to those found in human OA. 13 Recently, we developed a prosthetic unilateral anterior crossbite (UAC) rodent model in which typical OA-like lesions were observed in TMJs, such as chondrocyte death, cartilage matrix loss, osteochondral interface stiffening and abnormal reparative bone turnover. [14][15][16][17][18] In the present study, we first detected molecular changes in the PBLs of UAC rats. We then described the significance of the molecular changes disclosed in the PBLs of the UAC rats in the related tissues, including the alveolar bone, TMJ disc, mandibular condylar cartilage and subchondral bone. The purpose of this study was to "fingerprint" the molecular changes in PBLs that may be linked to severe OA lesions in TMJs; these findings will bring new insights regarding clinical biomarker detection and the evaluation of TMJ OA lesions.

| MATERIAL S AND ME THODS
The animal studies were approved by the Laboratory Animal Care and Welfare Committee, School of Stomatology, Fourth Military Medical University for animal research (2015-033).

| Animals and UAC operation
As we previously reported, there were slight gross morphological changes in the TMJ condyle heads of UAC-treated Sprague Dawley (SD) rats at 12 weeks and deformations such as twisted shapes at 20 weeks. Significant mineralisation was exhibited at 12 and 20 weeks after UAC operation, which was displayed as osteochondral interface stiffening. 16,19,20 As a result, we used the 20-week group for the PBL mRNA microarray detections and the 12-and 20-week groups to detect the TMJ and alveolar tissues for verification.
There was an 8-weeks gap between each time points.
Thirty-two 6-week-old female SD rats (150-160 g) were included in this study. Eight of the rats were used for the gene array analysis and randomly assigned to the sham-operated control (Con_Gene) or UAC_Gene groups (n = 4). These eight rats were sacrificed at the end of 20 weeks after the UAC operation. The remaining 24 rats were randomly assigned to four groups, that is, two sham-operated groups (CON) and two UAC groups (n = 6) that were sacrificed at the end of 12 and 20 weeks after the UAC operation, respectively.
UAC was created as previously described by our laboratory. [14][15][16] Briefly, a straight metal tube (ShinvaAnde, Shandong, China) and incisors. The rats in the control group underwent similar procedures but without bonding of the metal tubes. All rats were housed at 23°C with 30-40% relative humidity and normal 12 hours light and 12 hours dark cycles. They were fed with standard cylindrically shaped pressed food pellets.

| Blood sampling and processing for gene analysis
An expression analysis with the mRNA microarrays was performed on PBLs from the Con_Gene and UAC_Gene groups as an initial screening tool and to guide the subsequent gene expression studies. The rats were euthanised with an intraperitoneal injection of 1% pentobarbital (3 mL/kg body weight). Immediately, approximately 5 mL of whole blood was collected from the heart of each rat in each group. The other 24 PBL samples from the two CON and two UAC groups were lysed in TRIpure reagent (Roche, Basel, Switzerland) for real-time PCR detection.

| Real-time PCR
To verify the RNA sequencing data of the Con_Gene and UAC_Gene groups, real-time PCR was performed on the PBL RNA from the two CON and two UAC groups. Prime Script RT Master Mix Perfect All data were normalised to the GADPH internal standard and calculated using the 2ΔΔ-CT method. Three observations were performed for each sample, and the mean value was used to represent the expression level of the respective sample.

| Alveolar bone and TMJ samples
There were no differences in the histomorphology or molecular properties of the two sides according to our previous study. 14 The mandibular alveolar bone, TMJ discs, condyle cartilage and subchondral bone were rapidly removed in each UAC and CON group. The right side of the mandibular alveolar bones and TMJ tissues were used for protein preparation (n = 6). On the left side, the TMJ blocks were prepared for immunohistochemistry (IHC) staining as previously reported (n = 6). 16

| Histomorphometric measurements
Haematoxylin and eosin (HE) staining was performed as we previously reported. 15 Images were obtained with a Leica DFC490 system.

| Statistical analysis
The statistical analysis was performed using a two-way ANOVA test of variance after confirming the normality and homoscedasticity of the data sets in the two CON and UAC groups. When significant differences were identified, post hoc Student-Newman-Keuls tests were used to identify between-group differences. For all analyses, α = 0.05.

| Microarray detection of Rat PBLs
A total of 18 702 probes were tested. There were 96 upregulated and 78 downregulated genes in the PBLs of the UAC_Gene group, which was a change of more than 2-fold compared to that of the Con_Gene group ( Figure 1A). Analysis by the GO database showed that the top six GO categories were significantly upregulated in the UAC_Gene group, including the chemokine-mediated signalling pathway, cellular response to organic substances, cellular response to chemical stimuli, response to activity, cellular response to lipids and the cytokine-mediated signalling pathway. Of these top six upregulated GO categories ( Figure 1B), with a P value <0.05 and fold changes >2, three genes related to bone and cartilage were selected. The three genes were Egr1, Ephx1 and Il10, which were involved in three of the top six upregulated GO categories, including cellular response to organic substances, cellular response to chemical stimuli and cellular response to lipids. The expression in the GeneChip of these three genes increased, respectively ( Figure 1C-E). These upregulations were confirmed by a realtime PCR assay for PBLs from UAC rats at 12 and 20 weeks. Real-time PCR data indicated that the mRNA expression level of Egr1 had the earliest fold change in the UAC group at 12 weeks, whereas all three genes were increased at 20 weeks ( Figure 2, P < 0.05).

| Protein expression detection for the three selected molecules
The expression levels of the three gene-encoded proteins were increased in the TMJ tissues. The protein expression levels of EGR1, EPHX1 and IL10 were higher in the condylar cartilage and the subchondral bone of the UAC group at both time points (P < 0.05, except for EGR1 at 12 weeks in the subchondral bone and IL10 at 12 weeks in subchondral bone and 20 weeks in the cartilage), whereas that of EPHX1 was higher in the discs at both 12 and 20 weeks (P < 0.05). No such differences were noticed in alveolar bone tissues (Figure 3).

| D ISCUSS I ON
The present study was designed to identify some target genes in PBLs involved in TMJ OA-like lesions induced by UAC, which was demonstrated to induce observable changes at the macro-, microand molecular levels at 20 weeks of operation. 16 First, a microarray analysis of samples from 20-week UAC treatment rats and controls was detected. Candidate genes and proteins were selected when the expression fold differences between the PBLs of the UAC and control groups were over two times those of the top six upregulated GO categories. The three selected genes, Egr1, Ephx1 and Il10, were verified by comparing real-time PCR assay results between the UAC and control rats. Using Western blotting methods, the upregulation of EGR1, EPHX1 and IL10 was identified in the mandibular condylar cartilage and subchondral bone and that of EPHX1 was identified in the TMJ disc, all of which are tissues from TMJ; no upregulation was noted in the alveolar bone. None of the three molecules were expressed in the jaw muscle tissues detected by Western blotting assay (data not shown). Further analysis indicated that the increasing levels of EGR1 and EPHX1 in the UAC group occurred at 12 weeks in the mandibular condylar cartilage or subchondral bone.
Increased EPHX1 expression also occurred in the discs at both 12 and 20 weeks, and an increase in IL10 protein expression was identified in the subchondral bone at 12 weeks but in the cartilage at F I G U R E 2 Real-time PCR detection in the PBL mRNA focusing on the three selected upregulated genes in the top six upregulated GO categories, that is, Egr1, Ephx1 and Il10 (n = 6). 12W = 12 weeks, 20W = 20 weeks; CON = control group; UAC = unilateral anterior crossbite group; *P < 0.05; **P < 0.01 20 weeks. All three molecules were frequently detected at the osteochondral interface, as revealed by the IHC assay. Therefore, it is primarily the changes in the deep zone cartilage and the subchondral plate bone that contribute to the increased mRNA expression levels of Egr1, Ephx1 and Il10 in PBLs. These results agree with our recent report that the osteochondral interface could be a new diagnostic and therapeutic target. 16 Biomarkers related to OA lesions are attractive. 21 EGR1 is a transcription factor for growth and differentiation and serves as a regulator in several biological processes. 22 EGR1 is involved in responding F I G U R E 3 Comparison between the UAC and age-matched control groups for the protein expression levels of EGR1, EPHX1 and IL10 in the alveolar bone, TMJ disc, condylar cartilage and subchondral bone using a Western blot assay (n = 6). 12W = 12 weeks, 20W = 20 weeks; CON = control group; UAC = unilateral anterior crossbite group; *P < 0.05; **P < 0.01 to mechanical stress in bone cells 23,24 and wound-healing processes 25 such as the healing of vessels, cartilages and bones. 22,26 It can be detected in the callus in multiple tissues, such as lung injuries, endothelial wounds, vascularised tissue and bone fractures. 25 The depletion of Egr1 in mice affects the structure of cortical bones. 25 There is decreased bone mass in the limbs of adult Egr1 −/− mice. 27 The expression of EGR1 increases after osteogenic differentiation with osteogenic differentiation medium. 22 Currently, the EGR1 protein expression level was significantly higher in the condylar cartilage and subchondral bone of UAC rats than in the controls, but it was not elevated in the alveolar bone or TMJ discs. In addi- Although animal models were generally accepted for human disease investigation, humans are different from animals. The present outcomes must be verified by clinical data before being finalised.
Considering the present indications that there are time-and multitissue effects on the detected biomarkers, it is a challenge to enrol patients at different disease stages of OA without any other osteocartilage problems, such as osteoporosis and periodontal problems, which are widespread in elderly people.

| CON CLUS ION
We used a microarray analysis to focus on the upregulation of three molecules in the PBLs of rats with dentally induced TMJ OA-like lesions, including Egr1, which is related to tissue wound-healing processes, Ephx1, which is related to inflammatory responses and tissue destruction, and Il10, which plays a role in preventing the destruc-