Serum concentrations of TNF-α and its soluble receptors in Graves’ disease

Inclusion criteria

The diagnosis of Graves’ disease was based on the WHO criteria (17). Graves’ disease patients often have one or more of the following characteristics:

Diagnosis

Among these, signs of hyperthyroidism, diffuse goitre of thyroid, decrease in serum TSH and increase of thyroid hormone are essential diagnostic conditions.

Exclusion criteria

Another 72 euthyroid healthy controls (HC) (47 females and 25 males, mean age 39 ± 9 years) without clinical evidence or a family history of any autoimmune diseases were selected to participate in this study after recording histories, physical examination and confirmation of the normal FT3, FT4, TSH and negative thyroid autoantibodies.

All the participants were Chinese. Informed consent was obtained from all participants. The study was approved by the institution review board of Nantong City No 1 People’s Hospital and Second Affiliated Hospital of Nantong University and complied with the Declaration of Helsinki.

Serum TNF-α, sTNF-R1 and sTNF-R2 levels in GD and HC groups

As shown in Table 2, serum TNF-α levels were 39.0 (6.18–121) pg/mL for GD patients and 23.6 (4.99–72.0) pg/mL for the controls, but there were no significant differences between these two groups (P = 0.25). The GD group (33.6 (19.8–52.1) pg/mL) exhibited a significantly higher serum concentration of sTNF-R1 than that of the healthy group (25.8 (17.3–39.0) pg/mL) (P = 0.037). Similarly, the concentration of sTNF-R2 was greater in GD patients (24.6 (13.8–38.7) pg/mL) than in control subjects (14.0 (8.17–29.9) pg/mL) (P= 0.001). The detailed distributions are shown in Fig. 1.

View Full Size

Serum TNF-α (A), sTNF-R1 (B) and sTNF-R2 (C) levels in GD patients and healthy controls.

Citation: Endocrine Connections 9, 7; 10.1530/EC-20-0162

• Download Figure
• Download figure as PowerPoint slide

According to the median values for FT3, GD patients were further divided into a high FT3 GD group and a low FT3 GD group

Based on the median of values for FT3 (17.7 pmol/L) levels, GD patients were further divided into a high FT3 GD group (FT3 > 17.7 pmol/L) and a low FT3 GD group (FT3 < 17.7 pmol/L). The clinical parameters for each group are shown in Table 3. There were no significant differences in age (P = 0.585), gender (P = 0.293) distribution, BMI (P = 0.274), SBP (P = 0.674), DBP (P = 0.706) or course of disease (P = 0.793) between the high FT3 GD group and the low FT3 GD group. Compared with the low FT3 GD group, the thyroid volume (P = 0.002), thyroid antibody TRAb (P = 0.007) level, and immune index CD16 + 56 (P = 0.048) increased significantly, while values of the immune index IgM (P = 0.046) decreased significantly in the high FT3 GD group. The serum levels of sTNF-R1 (P = 0.008) in patients in the high FT3 GD group were significantly higher than those in the low FT3 GD group, but there was no significant difference for levels of TNF-α (P = 0.150) and sTNF-R2 (P = 0.126).

Correlations between serum TNF-α and its soluble receptor levels and related thyroid parameters in GD patients

There was strong correlation between the serum levels of sTNF-R1 and sTNF-R2 (Fig. 2). When the sTNF-R1 and sTNF-R2 levels were positively correlated with indices of thyroid function such as FT3 (P = 0.001, P = 0.012, respectively, Fig. 3B and C) and FT4 (P = 0.024, P = 0.028, respectively, Fig. 3E and F) within the patient group, TNF-α was negatively correlated with TSH (P = 0.024, Fig. 3G). It was also found that TRAb level was positively correlated with levels of TNF-α, sTNF-R1 and sTNF-R2 (P = 0.024, P = 0.010, P = 0.030, respectively, Fig. 4G, H, I and Table 4). In addition, we did not find any correlations among levels of TPOAb, TgAb and TNF-α and its soluble receptors (Fig. 4A, B, C, D, E, F and Table 4). Similarly, the inflammatory cytokines we measured were not correlated with either Thyroid VolT or duration.

View Full Size

Correlations between serum levels of TNF-α and sTNF-R1 (A), TNF-α and sTNF-R2 (B), sTNF-R1 and sTNF-R2 (C) in GD group.

Citation: Endocrine Connections 9, 7; 10.1530/EC-20-0162

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Correlations between serum levels of FT3 (A–C), FT4 (D–F), TSH (G–I) and TNF-α, sTNFRs in GD group.

Citation: Endocrine Connections 9, 7; 10.1530/EC-20-0162

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Correlations between serum levels of TPOAb (A–C), TgAb (D–F), TRAb (G–I) and TNF-α, sTNFRs in GD group.

Citation: Endocrine Connections 9, 7; 10.1530/EC-20-0162

• Download Figure
• Download figure as PowerPoint slide

Correlations between serum TNF-α and its soluble receptor levels and related immune parameters in GD patients

We also studied the relationship between TNF-α, sTNF-R1, sTNF-R2 and related immune indicators (Table 5). Relevant immune indicators such as complement C4 (C4) and the percentage of CD4, CD8, and CD19 showed differences between GD group and HC group (P = 0.004 for C4 and CD19, P = 0.001 for CD4 and P = 0.002 for CD8, respectively, Table 1). However, absolute counts for CD3, CD8 and CD16 + 56 showed differences between the two groups (P = 0.026 for CD3, P < 0.001 for CD8 and P = 0.001 for CD16 + 56, respectively, Table 1). Only serum levels of sTNF-R1 showed a significant positive correlation with C3 (P = 0.045). No correlation was found between these immune parameters and TNF-α or sTNF-R2 (Table 5).

Multiple linear regression analysis of serum sTNF-R1 and sTNF-R2 in GD patients

To analyse whether sex, age, SBP, DBP, BMI, duration, thyroid volume, FT4, TSH, TRAb, TPOAb, TgAb, sTNF-R1 or sTNF-R2 had significant effects on FT3, we performed multiple linear regression analysis. As is shown in Table 6, the serum levels of sTNF-R1 (P < 0.001) and FT4 (P = 0.023) may play an important role in the serum level of FT3.

Analysis of serum TNF-α level in GD patients

Investigation of the role of cytokines in the onset and course of autoimmune thyroid diseases has been ongoing for decades. A pro-inflammatory cytokine, TNF-α, is engaged in the induction and maintenance of immune responses. It is thought to play a critical role in autoimmune thyroid diseases, including Graves’ disease.

As it has been reported that TNF-α is usually short-lived in the peripheral blood, differing changes in serum levels of TNF-α have been reported in patients with hyperthyroidism. Chopra et al. documented serum levels of TNF-α in hyperthyroidism and hypothyroidism for the first time (10). Consistent with our findings, they found no abnormality in serum levels of TNF-α with these diseases. Siddiqi et al. found that serum TNF-α was undetectable in both patients and controls as well (12). It has been argued that mononuclear cells critical to autoimmune thyroid disease may be different from those critical for the production of TNF-α (10). In contrast, Celik et al. reported high median serum TNF-α levels in Graves’ disease patients, which could not revert to normal after 6 weeks of antithyroid medication. Elevated levels of TNF-α in the serum of GD patients may be affect to a systemic response to disease or to an intrathyroidal antigen–antibody reciprocity. Increase of cytokines in the thyroid gland may lead to the production of other cytokines by autoreactive cells, thus perpetuating the autoimmune reaction (13). From thyroid tissues of longstanding, relapsing Graves’ disease patients, Paschke et al. could not detect any TNF-α and suggested that TNF-α may be distinctly lower due to the chronic phase of the disease (19). It has been argued that tissue levels are not reflected in serum concentrations (10, 20). Aysen et al. were surprised to find that serum TNF-α levels were slightly lower in patients with hyperthyroidism. All the patients were given antithyroid medication until they became euthyroid. After the treatment, TNF-α returned to normal levels, comparable with healthy controls (14). Similarly, a reduced TNF response after repeated stimulation has been demonstrated in vitro (7).

The results of this study showed that there was no significant difference in serum TNF-α between GD patients and healthy controls. Possible explanations are as follows: (1) Serum TNF-α may be less stable than the soluble receptors (21, 22). Its concentration varies at different stages of the disease. (2) TNF-α is produced at lower levels in affected tissues and degrades after its production. (3) TNF-α is a secretory product of fat tissue (23). The serum level of TNF-α in the GD group did not increase significantly, which may be related to the significant decline in BMI due to hyperthyroidism. (4) It is also acknowledged that Graves’ disease is a common autoimmune disorder with a genetic predisposition. Owing to the biological effect of TNF-α on the thyroid gland and its gene location, TNF-α may be able to influence an individual’s susceptibility to GD (24). Gu LQ reported that both the G alleles of TNF-alpha −238 and +419 SNPs conferred a higher risk of GD than did A alleles. No significant difference in −308 allelic frequency was observed. Recently, Tu Y reported that, after ethnicity stratification, significant genetic GD susceptibility was detected in the European population in all genetic models. In contrast, no significant association could be detected in the Asian population (24). TNF-α gene polymorphism may play an important role as well. The influence of gene polymorphism on the occurrence and development of GD deserves further in-depth study. (5) In this research, the people we recruited were all Chinese; different races and regions may lead to different results. The duration in patients varied from 1 week to 10 years. The discrepancies among reports may result from differences in methodology, patient selection and chronicity of the autoimmune disease.

It was interesting to find that TNF-α was significantly negatively correlated only with TSH. TSH is produced by the anterior pituitary, so it serves as a sensitive index reflecting the function of the hypothalamus–pituitary–thyroid axis. TSH is the most sensitive and valuable objective index reflecting thyroid function. However, the TSH levels of GD patients in this study were generally low and imprecise; approximately 92% of them were distributed at 0.01, 0.02 and 0.03 mIu/L. Further investigation is needed to confirm the possible relationships and related mechanisms. These controversial results further suggest that it is necessary to measure the levels of both types of sTNFRs to assess TNF-α levels as a whole.

Analysis of serum sTNF-R1 and sTNF-R2 levels in GD patients

Our findings were consistent with those of previous studies demonstrating elevated levels of sTNF-R1 and sTNF-R2 in GD patients. These two soluble receptors were significantly correlated with the parameters of thyroid function. As sTNFRs are induced by TNF-α, their elevated concentrations in serum may reflect the activity of TNF-α even if TNF-α itself is not detected to be significantly increased. Indeed, there is considerable evidence suggesting that sTNF-R1 and sTNF-R2 are more reliable markers of TNF-α and inflammatory activity (25).

Only a few studies have included the measurements of serum levels of soluble cytokine receptors such as sTNF-R1 (2, 26, 27, 28, 29) and sTNF-R2 (30). Reuter et al. showed a significant correlation between elevated sTNF-R1 levels and increased levels of FT4, FT3, TRAb and thyroidal Tc-99m uptake in 39 hyperthyroid patients (29). Similar findings were reported by Díez et al. who also found increased sTNF-R1 levels in parallel with increased TNF-α serum levels in 18 Graves’ patients. However, they did not find any correlation between TNF-α or sTNF-R1 levels and FT4 and T3 levels in their hyperthyroid patients (28).

The level of FT3 is the most sensitive and valuable index reflecting thyroid function and is not affected by thyroglobulin. According to the median value of FT3, the GD patients were further divided into a high FT3 GD group and a low FT3 GD group. The serum levels of sTNF-R1 in patients in the high FT3 GD group were significantly greater than those in the low FT3 GD group, but there was no significant difference between TNF-α and sTNF-R2. Multiple linear regression analysis suggests that serum levels of sTNF-R1 and FT4 may play an important role in the serum level of FT3. FT4 is an important indicator in routine clinical diagnosis and can monitor the inhibition treatment of thyroid function. Our findings show that serum sTNF-R1 was significantly correlated with thyroid function, which is consistent with the interaction between different signal pathways of sTNF-R1 and sTNF-R2 stimulated by TNF-α, as reported by Petrus (31). The main function of sTNF-R1 in the signal pathway is to promote inflammation and apoptosis, while the main function of sTNF-R2 is to promote survival. Therefore, sTNF-R1 may play a more important role in the development of GD, but the underlying mechanism needs to be further explored.

We also found that TRAb level was positively correlated with TNF-α, sTNF-R1 and sTNF-R2. TRAb is a specific biomarker for GD (32, 33). It can be used in the study of immunology and pathogenesis of GD. It plays an important role in the differential diagnosis between GD and other thyroid diseases. TNF-R2 is a member of the costimulatory TNFRs. TNF-R2 transduces signals through TNF receptor-associated factors (TRAFs). TRAFs are critical for activating the canonical NF-κB, noncanonical NF-κB, and MAPK pathways. Engagement of TNF with TNF-R1 leads to the release of silencer of death domain protein (SODD) and the formation of a receptor-proximal complex, including some important adaptor proteins. These adaptor proteins, in turn, recruit other key pathway-specific enzymes. These enzymes then become activated and initiate downstream events leading to apoptosis, NF-κB activation, and JNK activation (34). This results in an imbalance of humoural and cellular immune responses leading to the appearance of autoantibodies. sTNFRs may play an important role in the occurrence and development of Graves’ disease.

Analysis of serum TNF-α, sTNFRs and immune indices in GD patients

A large number of studies show that the aetiology of GD is mainly on the basis of polygenetic inheritance; stress factors (such as mental stimulation) induce the occurrence of the autoimmune response. Antibodies are produced to bind with the TSH receptor, being the surface receptor of thyroid cells, stimulating thyroid cells to produce a high level of thyroid hormone, inducing immune disorder, resulting in infiltration by a large number of T lymphocytes, especially CD4 + T-cells, causing an imbalance in the proportion of helper T-cells (Th cells) and regulatory T-cells (Treg cells) and an imbalance of Th1 and Th2 cells, leading to abnormal expression of many cytokines (35). Professor Zhang Yanyun found that, in GD patients, the number of important Treg cells decreased significantly, the polarisation of dendritic cells shifted, and the proportion of its subtype plasma-like dendritic cells increased significantly, which can cause the increase of IFN-α secretion, leading to Treg cell apoptosis and GD immune imbalance. The mechanism and key regulatory targets of immune imbalance in GD are revealed, which provides an important basis for new immune intervention and treatment strategies (36).

CD4 + T-cells play an important role in selecting high-affinity B-lymphocytes to differentiate into memory cells or plasma cells and maintaining humoural immune response. As shown in this study, immune indices such as C4, CD4, CD8, CD19 and CD16 + 56 changed significantly in the patient group. The TNF system regulates immune function by inducing the expression of MHC class II molecules, stimulating the differentiation of monocyte/macrophage precursors, activating cells such as natural killer cells, and inducing the synthesis of cytokines. As a result, the autoimmune function of patients may be stimulated or aggravated and the composition of lymphocyte subsets may change. Compared to normal controls, C4 was significantly elevated in Graves’ patients. However, only serum levels of sTNF-R1 showed a significant positive correlation with C3. This indicates that sTNF-R1 may have a closer relation with the autoimmune mechanisms of Graves’ disease. The inflammatory mediators in hyperthyroid patients may induce C3 produced by hepatocytes, epithelial cells and fibroblasts (37). The reason for these contradictory observations may be the differential inducers of C3 and C4. It has been reported that the synthesis of C3 is dependent on IL-1, IL-2 and TNF-α, whereas the production of C4 is dependent on IL-6 and TNF-α cytokines (38). Accordingly, different cytokine profiles in hyperthyroid patients may lead to differences in C3 production. The detection of thyroid autoantibodies and lymphocyte subsets in GD patients may have practical significance for the guidance of clinical treatment, the evaluation of disease activity and the assessment of the prognosis of this disease.

Analysis of research limitations

It must be addressed that there are several limitations to our study. First, it was a cross-sectional design with small sample sizes. This study could not explain the cause–effect connection between increased serum sTNFRs and varying parameters of thyroid and immune function in Graves’ disease patients. Second, our research was conducted with Chinese participants, and the generalisability of our results needs to be evaluated. Third, due to the small sample sizes, we did not further group the patients according to the course or severity of Graves’ disease to study correlations with the factors we measured. Therefore, further studies should be conducted to validate the results of our research and handle these limitations.

In summary, our results suggest that Graves’ disease may be associated with an inflammatory profile. However, the mechanism is still unclear. Additional studies need to be done to determine if increased sTNFRs reflect a state or trait marker and to determine the functional significance of altered peripheral cytokine levels.