Correlation of HLA-DQ and TNF-α gene polymorphisms with ocular myasthenia gravis combined with thyroid-associated ophthalmopathy

The present study aims to explore the correlation of human leucocyte antigen (HLA)-DQ and tumour necrosis factor (TNF)-α gene polymorphisms with ocular myasthenia gravis (OMG) combined with thyroid-associated ophthalmopathy (TAO). From March 2009 to March 2015, 56 OMG patients complicated with TAO (OMG + TAO group), 134 patients diagnosed with OMG only (OMG group) and 236 healthy individuals (control group) were enrolled in the present study. PCR-sequence specific primer (PCR-SSP) was used for HLA-DQ genotyping and PCR-restriction fragment length polymorphism (PCR-RFLP) for TNF-α genotyping. ELISA kit was applied to detect acetylcholine receptor antibody (AchRAb) level and chemiluminescence immunoassay (CLIA) to measure thyroid-associated antibody (T-Ab) level. Logistic regression analysis was carried out to analyse the risk factors for OMG combined with TAO. DQA1*0103 showed lower frequency in the OMG group than in the control group. DQA1*0301 showed increased and DQB1*0601 showed decreased frequency in the OMG + TAO group. DQB1*0501 showed higher frequency in the OMG and OMG + TAO groups than in the control group. Patients carrying TNF-α -863C > A (CA + AA) might confront with greater risks of OMG combined with TAO. Frequency of DQA1*0103/*0301 and DQB1*0501/*0601, and TNF-α -863C > A, -238G > A and -308G > A were associated with the levels of AchRAb and T-Ab. TNF-α -863C > A (CA + AA) and high level of T-Ab were risk factors for OMG combined with TAO. Our results demonstrate that TNF-α -863 polymorphism is possibly correlated with the risk of OMG combined with TAO.


Introduction
Myasthenia gravis (MG) is an autoimmune disease resulting from an immunological response against the acetylcholine receptor (AChR) at the neuromuscular junction, leaving the voluntary muscles weak [1]. With an overall prevalence of 12 cases per 100000, ocular MG (OMG) shares the very same process with MG and affects almost half of MG patients, but its symptoms are restricted to the extraocular and levator palpebrae muscles, which take the form of diplopia and ptosis [2]. OMG can lead to visual disabilities and disrupt daily activities, and approximately 44% OMG patients are reported to develop generalized MG, which even poses a threat to survival [3]. Therefore, early diagnosis and treatment are fundamentally important. Graves' disease (GD) is a common autoimmune thyroid disorder and thyroid-associated ophthalmopathy (TAO) can affect 30-50% of GD patients and is characterized as orbital inflammation and expansion of fat and extraocular muscles [4]. The concurrence of GD and MG, though with relatively rare incidences, is particularly popular in the Asian women with age under 50 years [5,6]. Currently, there is no standard method for the treatment of OMG combined with TAO and the use of thymectomy and rituximab is quite controversial [7,8]. Anti-AChR antibodies are found to impair neuromuscular transmission and induce exacerbating receptor condition, causing increasingly reduced AChR availability and ensuing OMG [9]. Therefore, a new way to deal with OMG is very significant on genetic molecular basis.
Human leucocyte antigen (HLA) genes are vital in modulating the immune response to viral pathogens [10]. HLA class II, especially highly polymorphic HLA-DR and -DQ alleles, is crucial to human immune responses due to their functions in stemming the predisposition or protecting human bodies from certain diseases [11]. Previous studies have reported that HLA-DQ gene polymorphisms are closely associated with the development of MG [12,13]. Tumour necrosis factor (TNF)-α is a prototypical and pro-inflammatory cytokine, which can signal cell survival, proliferation and activation or even death via its type 1 receptor (TNFR1) [14]. Genetic polymorphisms in the promoter region of TNF-α are found to be involved in regulating levels and related to a large number of inflammatory and malignant diseases such as lung cancer and TAO [15,16]. However, the specific mechanisms behind these molecules still remain elusive. 'See-saw' relationship has been discussed very early between MG and hyperthyroidism [17,18]. While a previous study demonstrates a reverse 'see-saw' relationship between MG and GD on the basis of their clinical and immunological features and also suggests that HLA-DQ3 may play a pathogenic role in the concomitance of the two diseases [19]. Although accumulating support has been given to the reverse one, scholars at home and abroad have not reached a unified understanding of the mechanism of the concurrence of OMG and GD, which may be involved in various aspects of immunity and abnormal genes. And as the HLA-DQ gene has been shown as a genetic marker for resistance to autoimmune thyroid diseases [20] and TNF-α receptor blockers are involved in MG [21], we hypothesized that HLA-DQ and TNF-α gene polymorphisms may be a promising genetic target of OMG combined with TAO. Consequently, the present study is carried out to explore the correlation of HLA-DQ and TNF-α gene polymorphisms with OMG combined with TAO and related antibodies, with hope to lay molecular basis for early diagnosis of the disease.

Ethics statement
The present study was approved by the Ethics Committee of Tianjin Geriatrics Institute, Tianjin Medical University General Hospital and all participants signed an informed consent.

Study subjects
From March 2009 to March 2015, 190 patients (82 males and 108 females, mean age: 33.59 + − 7.84 years) with OMG receiving treatment in Tianjin Geriatrics Institute, Tianjin Medical University General Hospital were selected in the present study, including 56 patients complicated with TAO (the OMG + TAO group) and 134 patients diagnosed with OMG only (the OMG group). Inclusion criteria for OMG patients: (i) patients who were consistent with Chinese Medical Syndrome Differentiation Standards; (ii) patients who were in conformity to Western Diagnosis of OMG [22]; (iii) patients who were diagnosed as extraocular muscle involvement only with no other muscles involvement, no electrophysiology and no evidence of progression to other muscles. Exclusion criteria for OMG patients: (i) patients with affected muscles in parts other than in ocular part; (ii) patients diagnosed with other severe systemic diseases; (iii) patients with severe skin disease, mental disorder or other diseases involving heart, brain, liver and kidney; (iv) women under pregnancy and lactation and those who can seek self-relief out of psychological and mental concerns. Inclusion criteria for OMG patients combined with TAO: (i) patients who were consistent with inclusion criteria for OMG patients; (ii) patients who were in conformity to criteria for Western Diagnosis of TAO [23]. Exclusion criteria for OMG patients combined with TAO: (i) patients who were consistent with exclusion criteria for OMG patients; (ii) patients with thyroid crisis; (iii) patients with severe mental disorder; (iv) patients who were diagnosed with other severe systemic diseases. Concurrently, 236 healthy individuals (94 males and 142 females, mean age: 32.61 + − 8.23 years) who had no relationship with the included OMG patients and who received physical examinations in Tianjin Geriatrics Institute, Tianjin Medical University General Hospital were selected as the control group. There was no obvious difference among the three groups in the baseline data (P>0.05).

HLA-DQ genotyping
PCR-sequence specific primer (PCR-SSP) was employed to detect HLA-DQ gene polymorphism. Peripheral venous blood (10 ml) was collected from all the subjects, 5 ml of which was placed in an EDTA anticoagulation test tube, the rest of which was placed in common tubes for antibody detection. Whole blood genomic DNA kit (Promega Corp., Madison, WI, U.S.A.) was used to extract whole blood genomic DNA. Sequence specific primers (SSP) of DQA1 and B1 were designed consulting Bunce et al. [24] and internal reference primers (IRP) (forward (F): 5 -TGC CAA GTG

TNF-α genotyping
PCR-restriction fragment length polymorphism (PCR-RFLP) was used to detect TNF-α gene polymorphism and Primer Premier 5.0 software was employed to design PCR primers, which were synthesized by Shanghai Sangon Biological Engineering Technology & Services Co., Ltd. (Shanghai, China) ( Table 1), of which, TNF-863 site entailed mismatched base PCR, namely, a mismatched base (A→C) was introduced at the second site of 3 tail of downstream primers so as to mimic artificially a taiI enzyme restriction site. Reaction system for PCR was 20 μl: 2 μl of DNA genome, 1 μl of primers of downstream and upstream each, 2.5 μl of 10 buffer, 1.5 μl of MgCl 2 (25 mM), 2 μl of dNTP, 0.2 μl of Taq enzyme (TIANGEN Biotechnology Co. Ltd., Beijing, China) and 9.8 μl of ddH 2 O. Reaction conditions for PCR were: 30 cycles of pre-denaturation at 94 • C for 2 min, denaturation at 94 • C for 30 s, annealing at 55 • C for 45 s and extension at 72 • C for 45 s, ending with 10-min extension at 72 • C. PCR products received enzyme digestion at 37 • C overnight and were identified by 3% agarose gel electrophoresis to locate TNF-α promoter region,

Enzyme-linked immunosorbent assay (ELISA)
Peripheral blood without anticoagulation (4 ml) was collected and placed at room temperature for 30 min, followed by centrifugation (3000 rev/min) for 15 min to obtain the serum. ELISA kit (R S R Ltd., Cardiff, U.K.) was used to detect the absorbance (A) value of AChR antibody (AChRAb) (450 nm) and thus its concentrations were calculated. According to the standard for positive AChRAb, standard curve is drawn in conformity with A value of standards and corresponding concentrations are found in line with A value of samples. When the concentration reached over 0.2 nmol/l, the antibody was positively expressed [26]. Chemiluminescence immunoassay (CLIA) was employed to assess the serum levels of thyroglobulin antibody (TgAb), thyrotrophin receptor antibody (TRAb) and thyroid peroxidase antibody (TPOAb). Normal reference range: TgAb, 0-115 IU/ml; TRAb, 0-15 IU/ml; TPOAb, 0-34 IU/ml; over the upper limit proved to be positive [27]. If one of the three antibodies was positive, expression of thyroid-associated antibody (T-Ab) was considered as positive.

Statistical analysis
Statistical software of SPSS 20.0 was employed for statistical analysis, with measurement data presented as mean + − S.D. The Student's t test was employed for the comparison among groups, whereas ANOVA was used for comparisons among the three groups. Enumeration data were presented as a percentage or ratio and examined with Chi-square. *, comparison between the control and OMG + TAO groups; † , comparison between the control and OMG groups; ‡ , comparison between the OMG + TAO and OMG groups; Control, the healthy individuals.
Relative risks of genotype were presented as odds ratio (OR) or 95% confidence interval (95% CI). Group representation of the samples was assessed using the Hardy-Weinberg equilibrium test. Logistic regression analysis was conducted for risk factors of OMG patients combined with TAO. P value presented two-tailed probability and P<0.05 was considered as statistically significant.

Correlation of HLA-DQ gene polymorphism with OMG and OMG combined with TAO
As shown in Table 2

Correlation of TNF-α gene polymorphism with OMG and OMG combined with TAO
Goodness of Fit was employed for Hardy-Weinberg equilibrium test on the locus in the promoter region of TNF-α in the control, OMG and OMG + TAO groups, namely, -238G > A, -308G > A, -1031C > T, -857C > T and -863C > A. The results showed that among the three groups, the actual values of the five genotypes of HLA-DQ and their allele distributions were equal to the expected values, which conformed to Hardy-Weinberg equilibrium, suggesting that the samples were representative.
At the locus -863C > A in the TNF-α promoter region, the OMG + TAO group showed notably higher frequencies of CA + AA genotype and A allele than the control and OMG groups (all P<0.05). No significant difference was found in the frequency of -863C > A genotype between the control and OMG groups (P>0.05). In addition, there were no significant differences in the frequency of the genotypes and alleles of -238G > A, -308G > A, -857C > T and -1031T > C in the TNF-α promoter regions among the control, OMG and OMG + TAO groups (all P>0.05) ( Table 3).

Relationship between HLA-DQ and TNF-α gene polymorphisms and AChRAb
The positive rate of AChRAb in patients with HLA-DQ DQA1*0301 and DQB1*0501 was notably higher in the OMG + TAO group than in the OMG group (both P<0.05); and the positive rate of AChRAb in patients with DQA1*0103 and DQB1*0601 was remarkably lower in the OMG + TAO group than in the OMG group (both P<0.05) ( Table 4). The positive rate of AChRAb in the patients carrying CA + AA of -863C > A in the TNF-α promoter region was notably higher in the OMG + TAO group than those in the OMG group (P<0.05) ( Table 5).

Relationship between HLA-DQ and TNF-α gene polymorphisms and T-Ab
There was significant difference in AchRAb and T-Ab among the three groups. The OMG + TAO group exhibited higher positive rates of AchRAb and T-Ab than the OMG and the control groups (all P<0.05). The positive rate of T-Ab in patients with HLA-DQ DQA1*0301 and DQB1*0501 was remarkably higher in the OMG + TAO group than in the OMG group (both P<0.05). However, the positive rate of T-Ab in patients with DQA1*0301 and DQB1*0601 was remarkably lower in the OMG + TAO group than in the OMG group (both P<0.05) ( Table 6). The positive rate of T-Ab in patients carrying CA + AA of -863C > A in the TNF-α promoter region was notably higher in the OMG + TAO group than the patients in the OMG group (P<0.05); the positive rate of T-Ab in patients carrying CA + AA of -238G > A and -308G > A were notably higher in the OMG + TAO group than in the OMG group (both P<0.05) ( Table 7).

Logistic regression analysis for risk factors of OMG patients combined with TAO
Logistic regression analysis was carried out with whether OMG is complicated with TAO or not as a dependent variable, and the locus -863C > A, -238G > A and -308G > A in the TNF-α promoter region and DQA1*0103 and *0301, DQB1*0501 and *0601 and expressions of AChRAb and T-Ab as independent variables. The results revealed that TNF-α -863C > A (CA + AA) genotype and high expression of T-Ab were risk factors for OMG patients combined with TAO ( Table 8).

Discussion
Reported as one of the main reasons for discomfort and decreased quality of life, OMG combined with TAO requires immediate attention [28]. However, treatments for OMG and TAO do not seem to have any achievements substantially during the past few decades [2,4]. Nonetheless, previously, genetic polymorphism has been reported to be associated with MG, giving us inspiration to look at it from genetic perspectives [12].
It was revealed in the study that HLA-DQ DQB1*0501 and TNF-α -863C > A showed higher frequencies in the OMG and OMG + TAO groups than in the control group, suggesting that gene polymorphisms of HLA-DQ and TNF-α probably increased the risks of OMG combined with TAO. Many factors are found associated with MG,  including MHC (whose epithet in human beings is referred to as HLA), AchR, Ig, T-cell antigen receptor (TCR) and interleukin (IL)-1 genes [29]. HLA II molecules are responsible for presenting pathogenic epitopes to CD4+ T cells, which are important in the manifestation and perpetuation of MG by producing high-affinitive Abs [13]. How CD4+ T cells are selected in the thymus is regulated by HLA class II molecules, whose polymorphism at this locus is  very likely to give rise to susceptibility of OMG coexisting with TAO. Many previous studies have demonstrated the association between MG and HLA-DQ polymorphism [12,30], and patients with thyroid disease are prone to suffer from MG [6]. As TAO and OMG are related to thyroid, we assume that there must be similar pathogenesis between these two diseases. Therefore, HLA-DQ polymorphism may be involved in both TAO and OMG, which however, could be supported by few previous studies and further studies are still needed to confirm the result. TNF-α genes are situated in the region of HLA class III (250-kb centromeric of the HLA-B and 850-kb telomeric of the class II HLA-DR genes in humans) and were found to be involved in the incidence and progression of some contagious and  autoimmune diseases [31]. TNF-α enhanced the expression of HLA class I and adhesion molecules on thyrocytes, and increased expression of the molecules on thyrocytes would promote the autoimmune responses related to the pathogenesis of TAO, besides the polymorphic allele of TNF-α gene at position 308 (G to A, -308A) or 238 (G to A, -238A) was found to be involved in severity of and susceptibility to some autoimmune and infectious diseases [32]. Skeie et al. [33] have also found that presence of TNF-α alleles correlates with the onset of MG [33]. Additionally, Yan et al. [16] have demonstrated that allele A of TNF-α at position -863 is probably linked to TAO, in male patients in particular, which is completely consistent with our findings. We also found that the frequency distributions of DQA1*0301, DQB1*0501 and TNF-α -863C > A (CA + AA), -238G > A (GG) and -308G > A (GG) genotypes are associated with the positive rates of AchRAb and T-Ab in the OMG + TAO group. Furthermore, the logistic regression analysis confirmed that the CA + AA genotype of TNF-α -863C > A and high expression of T-Ab are risk factors for OMG combined with TAO. TNF-α -863C > A is associated with TNF-α secretion [34], and TNF-α may directly destroy AChR or promote the differentiation and growth of B cells resulting in an enhanced production of AChRAb [35], stimulating high expressions of AchRAb. Therefore, they are proceeding to cause increasingly reduced AChR availability and ensuing OMG by impairing neuromuscular transmission and inducing exacerbating receptor condition [9]. Additionally, it has long been recognized that positive AchRAb is synonymous with the onset of OMG [36]. Furthermore, as to HLA-DQ polymorphism, a result consistent with ours has also been reported by Jonsdottir et al. [37] that HLA-DQ is associated with positive T-Ab, further supporting our findings. However, despite the discovery of correlation between HLA-DQ and positive T-Ab in patients with TAO-combined OMG, we failed to figure out the mechanism behind it, which awaits further studies to elucidate it.
To summarize, we can come to the conclusion that TNF-α -863 polymorphism is likely to be associated with OMG combined with TAO, promising to lay a molecular basis for early diagnosis of the disease. However, the present study also has its own limitations. Due to the limited sample size and DQA1 and DQB1 gene polymorphism, which have multiple alleles, the present study showed significant difference in the AChRAb between the TAO and OMG groups with no obvious difference in distribution of HLA-DQ alleles between the OMG + TAO group and the OMG group, which may be further confirmed in the future and may be a research direction. And more similar studies with larger sample size are needed to be launched with prudence so as to verify our results and further elucidate possible mechanisms. T-Ab, thyroid-associated antibody; TAO, thyroid-associated ophthalmopathy; TBE, tris-borate-EDTA; TgAb, thyroglobulin antibody; TNF, tumour necrosis factor; TRAb, thyrotrophin receptor antibody; TPOAb, thyroid peroxidase antibody; 95% CI, 95% confidence interval.