Introduction

The FTO gene, which encodes 2-oxoglutarate-dependent nucleic acid demethylase [1], has been reported as an obesity-associated gene [2, 3] and is expressed in many tissues, including the cerebral cortex, hypothalamus, pituitary, and muscle [4]. The FTO gene was the first gene to be cloned after the identification of a fused toe (Ft) mutation resulting in a 1.6 Mb deletion in mouse chromosome 8 [5, 6]; however, its function remains unknown. FTO mRNA in the arcuate nucleus of the hypothalamus is regulated by dietary restrictions, suggesting its involvement in energy homeostasis [7]. In recent years, the FTO gene has been attracting attention in the field of orthopedics as a risk gene for osteoarthritis (OA) [8,9,10,11], and a genome-wide association study by Zeggini et al. reported that the C allele of an FTO SNP (rs8044769) was strongly associated with OA in females [12]. Their excellent discoveries that have powerful influence on the subsequent studies have undoubtedly provided a great understanding of OA. However, whereas many studies conclude that the rs8044769 polymorphism affects OA through obesity [12, 13], some believe that this hypothesis is not sufficiently proven because of the difficulty in confirming the causal relationship between obesity and OA [14]. The mandibular condyle, which develops due to the same intracartilaginous ossification as the knee joint, has a structure that hangs from the skull. There are few studies on the relationship between body mass index (BMI) and TMJOA, and the absence of this relationship was reported by Winburn et al. [15]. Furthermore, due to the weak relationship between occlusal force and BMI [16], TMJ is minimally affected by body weight.

Degenerative lesions mainly indicate OA in the hard tissue of the joint, which develops in all joints of the human body, including the TMJ in the orofacial region. TMJOA is a sub-pathological condition of TMD and is considered the end stage of the disease [17, 18]. Its clinical symptoms include at least one joint noise (mainly crepitus), jaw dyskinesia, and pain in the TMJ. In addition, it has been reported that TMJOA clinical symptoms significantly reduce the quality of life [19], and many refractory patients do not obtain the desired therapeutic effects even if they receive appropriate treatment. Nonetheless, the pathophysiology of TMJOA is highly complex, and its definitive etiology has not yet been established.

Investigation of FTO in the mandibular condyle allows for an understanding of the effects of FTO SNPs on OA while adjusting the effect of body weight on joints to the fullest extent. We hypothesized that the FTO gene (rs8044769) is involved in the development of OA in the mandibular condyle, regardless of the body mass index (BMI). Therefore, in this study, we aimed to elucidate the effects of rs8044769 allele on mandibular condyle OA using multivariate analysis after adjusting for the effects of age, sex, BMI, and temporomandibular joint (TMJ) disc abnormalities.

Methods

Subjects

One hundred and sixty-two Japanese patients (average age 57.0 ± 12.2 years; 32 males, 130 females) who visited the Osaka University Dental Hospital with symptoms of temporomandibular disorders (TMDs) and underwent MRI examination were analyzed in this study (Fig. 1). Patients under 30 years of age, those without MRI scans, and those presenting metal and motion artifacts on MRI scans were excluded from the study. Of the 162 patients, 104 were consecutive patients who visited the hospital between January 2019 and March 2020 (48 patients with TMJOA, 56 patients without TMJOA). Samples from the remaining 58 patients were extracted from the MR database of 678 patients who visited the hospital between 2015 and 2018 and underwent TMJ MRI examination. Of these 58 patients, 31 had severe degenerative bone changes in the mandibular condyle, and 27 had normal TMJ with no anatomical abnormalities, such as disc displacement and degenerative bone change. Written informed consent to participate in the study was obtained from all subjects. All procedures performed in studies involving human participants were in accordance with the 1964 Helsinki declaration and its later amendments or comparable ethical standards and conformed with the ethical standards of the institutional and/or national research committees. The present research complied with the STROBE statement. The study procedures have been approved by the Ethical Review Board of Osaka University (H30-E11).

Fig. 1
figure 1

Recruitment scheme. TMJOA, temporomandibular joint osteoarthritis

Full size image

MRI was obtained by utilizing a 1.5 T MR imaging scanner (SignaHDxt 1.5 T; GE Healthcare) equipped with a TMJ surface coil. All participants underwent imaging in axial, sagittal, and coronal planes utilizing fast spin-echo sequences. Axial T1-weighted images were obtained as the localizer as follows: TR = 150, TE = 4.2.The sagittal planes were oriented perpendicular to the long axis of the condylar head. Sagittal and coronal proton density-weighted images at the mouth closing were oriented as follows: TR = 2500, TE = 20, ETL = 8, and NEX = 2. Sagittal and coronal fat-suppressed T2-weighted images at the mouth closing were acquired as follows: TR = 2000, TE = 85, ETL = 16, and NEX = 3. Proton density-weighted images at the mouth closing and opening were acquired to evaluate disc displacement and the presence of reduction as follows: TR = 800, TE = 24, ETL = 4, and NEX = 2. The other parameters were as follows: 22.2 × 10 cm field of view, 256 × 160 matrix size, 3-mm section thickness.

Height and weight self-reported by the patients at the time of MRI were recorded. BMI was calculated by dividing the patient’s body weight by the square of their height.

Diagnosis of OA and internal derangement of the TMJ

Two oral and maxillofacial radiologists performed diagnosis of OA and internal derangement of the TMJ upon MRI examination. The presence or absence of the following was recorded as degenerative bone changes in the mandibular condyle from the proton density-weighted images of the sagittal and coronal planes and T2-weighted images: osteophyte (marginal hypertrophy with sclerosing borders and growth of bone tissue from the bone surface, Fig. 2A), erosion (loss of continuity of cortical bone, Fig. 2B), subchondral cyst (high signal areas for cysts under the joint surface in T2-weighted images, which are different from normal bone marrow, Fig. 2C), and atrophic deformity (its sagittal plane has a thin rod-like appearance, and its coronal plane has a mandibular condyle that is less than half the size of the normal mandibular condyle inferred from the mandibular fossa, Fig. 2D). The diagnostic criteria for OA were at least one imaging finding of osteophytes, erosion, atrophic deformity, and subchondral cysts. Generalized sclerosis was excluded from OA because of the difficulty of accurate diagnosis by MRI [20]. All 324 mandibular condyles were classified into the TMJs with OA and the TMJs without OA. The positional abnormalities of the articular disc were classified as normal, sideways displacement, partial anterior displacement, partial anterior and sideways displacement, complete anterior displacement, or posterior displacement [21,22,23].

  1. 1.

    Normal: In a closed-mouth position, the top of the condyle was located underneath the intermediate thinnest portion of the disc, and the disc had a biconcave shape in all sections of sagittal imaging slices.

  2. 2.

    Sideways displacement: In the coronal imaging plane, the disc was laterally displaced over the medial or lateral poles.

  3. 3.

    Partial anterior displacement: Disc was anteriorly displaced in the lateral or medial part of the disc, whereas the top of the condyle was positioned underneath the intermediate zone of the disc in the contralateral portion.

  4. 4.

    Partial anterior and sideways displacement: Combination of partial anterior displacement and sideways displacement.

  5. 5.

    Complete anterior displacement: The posterior band was located anterior to the anterior surface of the condylar head in all sagittal slices.

  6. 6.

    Posterior displacement: In sagittal slices, the disc is posterior to the condyle, and enlargement of the posterior band appears.

Fig. 2
figure 2

Examples of osteoarthritis on MRI. A Osteophyte. B Erosion. C Subchondral cyst. D Atrophic deformity

Full size image

Dynamic articular disc abnormalities were classified into the following three groups [24]: no dynamic disc abnormality (Fig. 3A), in which articular disc displacement was not observed when the mouth was closed; articular disc displacement with reduction (DDwR, Fig. 3B), in which the articular disc was displaced when the mouth was closed, but the mandibular condyle captured the articular disc, and the relationship between the mandibular condyle and mandibular fossa appeared normal when the mouth was open; and articular disc displacement without reduction (DDwoR, Fig. 3C), in which the articular disc was displaced when the mouth was closed, and the posterior thickened part of the articular disc appeared to be located anterior to the anterior surface of the mandibular condyle in at least one sagittal plane when the mouth was open (alternatively, it had the appearance of residual sideways displacement when the mouth was open).

Fig. 3
figure 3

Dynamic disc abnormality in TMJ. A No dynamic disc abnormality. B Disc displacement with reduction. C Disc displacement without reduction

Full size image

The Kappa coefficient was used to analyze the inter-rater reliability for the presence of dynamic disc displacement and OA. When a diagnosis differed between radiologists, a discussion for the definitive diagnosis was held, and a final decision was made by mutual agreement.

Genotyping

DNA samples were collected non-invasively from the buccal mucosa of the subjects using FLOQSwabs (COPAN Diagnostics Inc.). Each of the patients was asked to obtain two buccal swabs from bilateral sides of the cheek. The examiner scraped the inside of cheek with the swabs firmly more than six times. Then, the swabs were stored in the freezer at − 20 °C. The obtained DNA samples were used to analyze the FTO gene (rs8044769). Genotyping was performed blindly by laboratory personnel (Takara Bio Inc.), and its accuracy was validated in advance using samples from another 22 cases, which showed a degree of agreement of 100%.

  • (dependent variable: the presence of OA) and the predictors

  • (covariates: age, sex, dynamic disc state [the presence of woR])

Statistical analysis

The polymorphisms (rs8044769) were categorized into CC, CT, or TT. The age of the patients on the day of MRI was used in the statistical analysis. The descriptive statistics for the TMJs with OA and the TMJs without OA were evaluated using Fisher’s exact test and the Mann–Whitney U test. The BMIs of patients harboring different alleles of FTO polymorphisms were compared using the Kruskal–Wallis test. Furthermore, binomial logistic regression analysis was performed with the dependent variable being the presence or absence of OA, and the explanatory variables were sex, age, BMI, the rs8044769 genotypes (CC, CT, or TT), the presence of positional disc abnormalities, and the presence of disc displacement without reduction. Binomial logistic regression analysis was performed for each mandibular condyle to consider the effect of the laterality of the mandibular condyle. Model fitness was evaluated by using the Hosmer–Lemeshow test. The data were analyzed using the Statistical Package for Social Science (SPSS) version 23.0.

Results

Inter-rater reliability for the presence of dynamic disc displacement and OA in 314 TMJs was excellent (κ = 0.92 and 0.82, respectively). Sixty-two osteophytes, 22 erosions, 38 subchondral cysts, and 47 atrophic deformities were diagnosed by MRI from the 324 TMJs. Consequently, the numbers of TMJs with OA and TMJs without OA were 113 and 211, respectively. The results of the descriptive statistics for the TMJs with OA and the TMJs without OA are shown in Table 1. The median age and BMI of the TMJs with OA were 59 years (interquartile range [IQR] 51–66) and 21.4 kg/m2 (IQR 18.8–23.7), respectively, whereas those of the TMJs without OA were 58 years (IQR 47–66) and 21.4 kg/m2 (IQR 19.1–23.7), respectively, showing no statistically significant difference between them. In addition, the TMJs with OA and the TMJs without OA showed no statistically significant differences in the distribution of sex. A Fisher’s exact test showed a statistically significant difference in the allele frequency of FTO gene polymorphisms between the TMJs with OA and the TMJs without OA (P = 0.011). In addition, there was a statistically significant difference in the distribution of positional disc abnormalities and dynamic disc abnormalities between the TMJs with OA and the TMJs without OA (P < 0.001).

Table 1 Descriptive statistics for TMJs with OA and TMJs without OA. OA osteoarthritis, IQR interquartile range, AR adjusted residual, DDwR disc displacement with reduction, DDwoR disc displacement without reduction
Full size table

The median BMIs of patients in the CC, CT, and TT allele groups were 21.8 (IQR 19.1–24.1), 21.4 (IQR 19.1–22.8), and 19.6 (IQR 18.1–22.0) kg/m2, respectively, showing no significant differences among the groups (P = 0.116).

Furthermore, logistic regression analysis indicated that the CC allele of rs8044769 is a risk factor for OA, in contrast with the TT allele (Table 2; odds ratio = 2.765). Moreover, the presence of disc displacement without reduction was significantly associated with OA. The Hosmer–Lemeshow test showed the goodness of fit of the model adjusted for sex, age, BMI, and dynamic disc abnormality (χ2 = 6.266, df = 8, P = 0.617).

Table 2 The result of binary logistic analysis for associations between FTO polymorphism (rs8044769) and OA (n = 324TMJs). Hosmer–Lemeshow test: χ2 = 6.266, df = 8, P = 0.617. OA osteoarthritis, BMI body mass index, PDA positional disc abnormality, DDwR disc displacement with reduction, DDwoR disc displacement without reduction, OR odds ratio
Full size table

Discussion

The involvement of gene polymorphisms in the development of OA in the knee and hip joints has been reported, and the discovery of risk-associated genes has attracted broad attention as they expand the possibilities of new treatments for OA [8, 25]. One such gene is the FTO gene, which has been reported to be associated with obesity [2, 26]. Although the structure of FTO has been elucidated, its exact function remains unknown [27]. A genome wide association study reported that rs8044769 in the FTO gene is associated with OA in the knee and hip joints [12]. However, Panoutsopoulou et al., who adjusted for the effects of BMI in a meta-analysis, reported that variations in FTO SNPs (rs8044769) were not associated with the presence of OA in the knee and hip joints [13]. In addition, Dai et al. provided evidence that rs8044769 in the FTO gene is not associated with either the risk of OA or BMI in the Chinese population [28]. However, traditional studies cannot determine whether obesity causes OA or whether weight gain is caused by reduced activity due to OA. Therefore, in the present study, we investigated the effects of the FTO gene polymorphism rs8044769 on OA in the TMJ, which is minimally affected by BMI. Logistic regression analysis showed no significant association between BMI and the presence of OA, and it suggested that the rs8044769 variant is significantly associated with TMJOA. However, results of an in vivo mouse model suggested that the development of pathological changes in TMJ could be caused by the excessive compressive mechanical force and high-fat diet-induced obesity [29]. In addition, it has been reported that a quick eating rate is positively associated with excess body weight [30]. Furthermore, several chewing features such as chewing pace and time are associated with obesity in young adolescents, and they might affect the development of the TMJ [31]. Thus, further investigation will be required to reveal the relationship among jaw mastication, obesity, and TMJOA.

Internal derangement has been categorized as normal → DDwR → DDwoR → OA. However, we support the recent protocol by Schiffman et al. which assesses the disc displacement and degenerative bone change in the condyle independently [32]. Therefore, in the present study, each TMJ was diagnosed, e.g., DDwR with OA or DDwoR without OA. Not a few patients have asymmetrical TMJs, for example, DDwoR with OA on the right side and DDwR without OA on the left side, and if the statistical analysis is calculated in units of patients, it is difficult to evaluate the association between disc abnormalities and OA. Hence, the present logistic regression analysis was performed in units of TMJs. Consequently, in this study, an association between rs8044769 and OA was observed even after adjusting for the effects of dynamic articular disc abnormalities, which have been reported to be strongly associated with TMJOA [33, 34]. Thus, the FTO gene might enhance bone metabolism dysfunction and promote the development of OA, triggered by disc displacement with reduction. These results indicate that the FTO gene could be involved in bone metabolism, as shown by abnormal finger development due to chromosomal duplication in the FTO locus region [35]. However, the present logistic regression analysis performed in units of TMJs may result in overestimation in the incidence rate for OA, and the action of rs8044769 might be more likely to occur in the TMJ than in other joints. In addition, the position of the disc was evaluated in a subjective way by two operators, but objective measurements to define the TMJ disc position were not performed. Yet, the operators had excellent agreement.

Early-onset OA is associated with a strong familial representation of disease. In contrast, the common late onset OA is caused by additive effects of genetics and environmental factors [28]. Since the FTO gene is likely to be involved in late onset OA, subjects under 30 years of age were excluded to minimize the possibility of early onset OA. However, the time of onset of OA cannot be determined from the results of this cross-sectional study, and patients with a history of early onset OA might have been included in the present study. Moreover, the prevalence and severity of OA vary according to race and ethnicity, and geographical and environmental factors also have a significant impact. Thus, further research is required to investigate the potential mechanisms underlying OA. The major barrier to genetic studies of OA is the need to obtain a large number of individuals diagnosed with OA with standardized imaging examinations. The availability of TMJ MRI and CT scans poses a major obstacle, and few studies on TMJOA and gene polymorphisms have been reported [36]. In particular, data on normal TMJ in the elderly are invaluable as a control, but their extraction is challenging. Our research group possesses a large MR database of TMJ data, from which sufficient control individuals’ data could be extracted for this study. On the other hand, several patients with normal disc without OA were extracted from the MR data set, and eventually, only one TMJ has a normal disc with OA among all the subjects, and the odds ratio of the presence of DDwoR in the binominal logistic regression analysis was significantly higher than those of the previous reports [34, 37]. In the present study, the categorical data of the dynamic disc abnormalities were input as explanatory variables to adjust for their effects on OA, and thus, the odds ratio of the presence of DDwoR should not be referred to as a risk assessment parameter.

The present cross-sectional study supported the evidence that genetic risk factors might be associated with the development of TMJOA in the Japanese population without involvement from the effect of BMI. However, large-scale longitudinal trials in more diverse and generalized populations will be needed to reveal details of such genetic involvement. In the future, the identification of genetic risk factors that influence the therapeutic outcome in OA is expected to lead to preoperative diagnostic techniques for the selection and provision of an optimal treatment method suitable for the patient’s genetic predisposition. The findings of this study also suggest the mechanism underlying the development of TMJOA and its evolutionary significance. In the future, these are expected to contribute to the understanding of degenerative diseases of the skeletal system caused by genomic variation, their treatment, and genome-based drug discovery for skeletal regeneration.

Conclusions

We concluded that rs8044769 in the FTO gene could be involved in the development of TMJOA. Within the limitations of this retrospective study, the present study will allow for a deeper understanding of the pathophysiology of the TMJOA and supports the use of genetic risk factors for the preoperative examination and evaluation of therapeutic outcomes in this condition.