Meta-analysis reveals significant association between FOXP3 polymorphisms and susceptibility to Graves' disease.

OBJECTIVE: This meta-analysis aimed to determine the associations between the rs3761547, rs3761548, and rs3761549 single-nucleotide polymorphisms (SNPs) of the forkhead box P3 (FOXP3) gene and susceptibility to Graves' disease (GD).
METHODS: Case-control studies with information on the associations between the rs3761547, rs3761548, and rs3761549 FOXP3 SNPs and GD published before 01 May 2020 were identified in the PubMed, Embase, Web of Science, and China National Knowledge Infrastructure databases. Data from the studies were analyzed using RevMan version 5.3.
RESULTS: Seven independent case-control studies including 4051 GD patients and 4569 controls were included in the meta-analysis. The overall pooled analysis indicated that FOXP3/rs3761548 and FOXP3/rs3761549 polymorphisms were significantly associated with GD susceptibility (rs3761548: A vs. C, odds ratio [OR] = 1.32, 95% confidence interval [CI] 1.05-1.67; rs3761549: TT vs. CC, OR = 1.98, 95%CI 1.49-2.65; (TT + TC) vs. CC, OR = 1.44, 95%CI 1.11-1.88). In contrast, the FOXP3/rs3761547 polymorphism was not associated with GD susceptibility. Subgroup analysis according to ethnicity showed that rs3761548 was associated with GD in Asians but not in Caucasians, whereas rs3761549 was associated in both Asians and Caucasians.
CONCLUSION: This meta-analysis demonstrated that FOXP3/rs3761548 and FOXP3/rs3761549 SNPs were significantly associated with susceptibility to GD, at least in Asian populations.

Introduction

Graves’ disease (GD), also known as toxic diffuse goiter, is a common autoimmune thyroid disease (AITD) and the most common cause of hyperthyroidism, accounting for more than 80% of cases.[1] GD is more common in women (male-to-female ratio of 1:8) and usually occurs between the ages of 20 and 40 years.[2] The prevalence of GD in China is about 1.1% to 1.4%.[3] GD is typically characterized by the unique association of thyrotoxicosis, goiter, ophthalmopathy, and the presence of circulating thyrotropin receptor antibody; however, the exact etiology of GD remains unclear. Numerous recent studies have investigated the roles of epigenetic, environmental, and immunological factors in the pathogenesis of GD, and have suggested that interactions among all these factors play a significant role in the pathogenesis of GD.[4]

Regulatory T cells (Tregs) play a pivotal role in suppression of the immune response and the development of immune tolerance. Tregs are important factors in the pathogenesis of multiple human autoimmune diseases, such as multiple sclerosis, systemic lupus erythematosus, rheumatoid arthritis, and type 1 diabetes.[5,6] In addition, some studies have described a correlation between Tregs and GD.[5,7] Compared with healthy controls, GD patients show decreased levels of Tregs, thus emphasizing their significant role in the pathogenesis of GD.[8]

Forkhead box P3 (FOXP3) is an important regulatory factor for the differentiation of T cells into natural Tregs and is a key molecule controlling Treg development and function.[9-15] The locus of FOXP3 on human chromosome Xp11.23 encodes a protein with 431 amino acids.[16-19] Lack of FOXP3 impairs the immunosuppressant action of Tregs.[16,20] FOXP3 regulates T-cell activation and functions as a transcriptional repressor to downregulate cytokine production in T cells.[18,21] Five possible polymorphisms in the FOXP3 gene may change the expression levels of the protein and thus impair its function, thereby damaging the suppressive ability of Tregs and leading to autoimmune diseases.[22-27] The known functions of FOXP3 suggest that it might be a candidate susceptibility gene for autoimmune diseases.[9,18,28] FOXP3 has also been associated with AITDs.[9,29] Some single-nucleotide polymorphisms (SNPs) of the FOXP3 gene, including −2383 C/T (rs3761549), −3279 C/A (rs3761548), −924 A/G (rs2232365), −1383C/T (rs2232364), and −3499A/G (rs3761547), may influence its expression.[16,24] The SNPs rs3761547, rs3761548, and rs3761549 have been the most commonly tested polymorphisms in previous genetic association studies.

Decreased expression of FOXP3 can damage the function of Tregs and may result in an increase in autoreactive T cell activity, leading to destruction of the thyroid gland in GD patients.[9,16] Although some studies have investigated the relationship between FOXP3 variants (rs3761547, rs3761548, and rs3761549) and the risk of GD in diverse populations,[5,9,16,22,30,31] the results have been inconclusive. To reach a firmer conclusion, we therefore collected data from previous relevant studies and conducted a meta-analysis to examine the possible associations between FOXP3 polymorphisms and susceptibility to GD.

Methods

Search strategies

We searched the PubMed, Web of Science, Embase, and the China National Knowledge Infrastructure databases for case–control studies examining the relationship between FOXP3 and GD susceptibility, published before 1 May 2020, with no language restrictions. The keywords were: “forkhead box P3 or FOXP3” AND “Graves’ disease or Graves disease or Basedow disease.”

Inclusion and exclusion criteria

Studies were required to meet the following inclusion criteria: (1) studies examining the association between FOXP3/rs3761547/rs3761548/rs3761549 polymorphisms and GD susceptibility; (2) case–control studies; (3) studies concentrating on humans; and (4) detailed genotype data could be obtained to calculate odds ratios (ORs) and 95% confidence intervals (CIs). The exclusion criteria were: (1) not case–control study; (2) duplicate publication; (3) animal study; (4) no original data on allele or genotype frequencies; and (5) article type was a review, letter, case report, meta-analysis, or commentary.

The study did not require ethics committee approval because it was an analysis of previously published studies.

Quality assessment

The qualities of the included studies were evaluated according to the Newcastle–Ottawa assessment scale (NOS). The evaluation criteria included case selection, comparability, exposure, data, and genetic testing method. Each study was awarded a star for each item that was evaluated satisfactorily according to the NOS, up to a maximum of nine stars.

Data extraction

Two reviewers independently extracted data from each eligible article, including the first author, date of publication, original country, race, genotyping method, numbers of cases and controls, frequencies of genotypes and alleles, and P-value for the Hardy–Weinberg equilibrium (HWE). Any disagreements were resolved by another reviewer.

Statistical analysis

The meta-analysis examined each SNP according to five genetic comparison models: allele, homozygote, heterozygous, dominant, and recessive. The HWE was assessed for each study using the χ2 test in healthy subjects. ORs with 95%CIs were calculated using the χ2 test to evaluate the strength of the association between FOXP3 polymorphisms (rs3761547, rs3761548, and rs3761549) and GD risk. The statistical significance of the ORs was analyzed by the Z-test, with Pz < 0.05 considered a significant association. Heterogeneity was investigated using the Q-test and I2 statistics. I2 < 50% indicated a low degree of heterogeneity, in which case a fixed-effects model was used to calculate the ORs and 95%CIs, otherwise, a random-effect model was used. Potential publication bias was evaluated by funnel plots. Sensitivity analysis was conducted by removing each individual study to evaluate the robustness of the results. HWE statistical analyses were undertaken using Review Manager (RevMan) Version 5.3 (Cochrane Collaboration, 2014, Copenhagen, Denmark).

Results

Study characteristics

A total of 245 studies were retrieved through an initial search, of which 86 were removed as duplicates and 114 were removed as irrelevant after reading the titles and abstract. We read the full texts of the remaining 18 studies, and finally included seven eligible case–control studies in this meta-analysis. A completed PRISMA flow chart is provided as Supplementary Figure 1. The characteristics of the included studies are summarized in Table 1. The seven included studies involved 4051 GD patients and 4569 controls. Among these, five studies involving 1086 cases and 1157 controls examined rs3761547, six studies involving 1166 cases and 1442 controls examined rs3761548, and seven studies involving 1799 cases and 1970 controls examined rs3761549. For FOXP3/rs3761547, four studies were conducted in Asian populations and one in a Caucasian population; for FOXP3/rs3761548, five studies were conducted in Asian populations and one in Caucasians; and for FOXP3/rs3761549, five studies were conducted in Asian populations and two in Caucasian populations. We subsequently intended to include studies on other FOXP3 loci and Hashimoto’s thyroiditis, as a common AITD, but these studies were not included here because of their small numbers. In terms of NOS evaluation, all articles were awarded more than six stars, indicating high methodological quality.

Table 1.

Characteristics of studies included in the meta-analysis.

Study ID	Year	Ethnicity	Genotyping methods	Cases					Controls					P HWE	NOS stars
				CC	CA	AA	C	A	CC	CA	AA	C	A
rs3761548
Inoue et al.[31]	2010	Asian	PCR-RFLP	64	27	7	155	41	58	8	5	124	18	<0.05	8
Bossowski et al.[9]	2014	Caucasian	TaqMan	40	43	24	123	91	26	29	19	81	67	0.07	8
Zheng et al.[22]	2015	Asian	PCR-RFLP	196	92	20	484	132	225	70	11	520	92	0.07	8
Yu et al.[5]	2017	Asian	PCR-RFLP	331	166	37	828	240	466	132	32	1064	196	<0.05	8
Fathima et al.[30]	2019	Asian	PCR-RFLP	10	53	17	73	87	17	207	61	241	329	<0.05	7
Shehjar et al.[16]	2018	Asian	PCR-RFLP	29	101	5	159	111	69	69	12	207	93	0.36	7
rs3761549
Owen et al.[32]	2006	Caucasian	MALDI-TOF-MS	_	_	_	977	142	_	_	_	682	114	_	7
Inoue et al.[31]	2010	Asian	PCR-RFLP	56	24	7	136	38	38	19	5	95	29	0.25	8
Bossowski et al.[9]	2014	Caucasian	TaqMan	78	30	1	186	32	64	11	0	139	11	0.49	8
Zheng et al.[22]	2015	Asian	PCR-RFLP	180	102	26	462	154	188	98	20	474	138	0.15	8
Yu et al.[5]	2017	Asian	PCR-RFLP	304	171	59	779	289	391	202	37	984	276	0.12	8
Fathima et al.[30]	2019	Asian	PCR-RFLP	24	46	10	94	66	132	115	38	379	191	0.11	7
Shehjar et al.[16]	2018	Asian	PCR-RFLP	68	37	30	173	97	102	40	8	244	56	0.14	7
rs3761547
Inoue et al.[31]	2010	Asian	PCR-RFLP	69	33	7	171	47	46	20	4	112	28	0.37	8
Bossowski et al.[9]	2014	Caucasian	TaqMan	80	25	1	185	27	64	10	0	138	10	0.53	8
Zheng et al.[22]	2015	Asian	PCR-RFLP	190	95	23	475	141	195	92	19	482	130	0.08	8
Yu et al.[5]	2017	Asian	PCR-RFLP	315	171	48	801	267	391	189	50	971	289	<0.05	8
Shehjar et al.[16]	2018	Asian	PCR-RFLP	85	41	9	211	59	98	43	9	239	61	0.16	7

HWE, Hardy–Weinberg equilibrium; NOS, Newcastle–Ottawa scale; PCR, polymerase chain reaction; RFLP, restriction fragment length polymorphism; MALDI-TOF-MS, matrix assisted laser desorption ionization-time of flight mass spectrometry.

Quantitative synthesis

FOXP3/rs3761548 polymorphism

Six case–control studies involving a total of 1166 GD cases and 1442 controls assessed the relationship between the FOXP3/rs3761548 polymorphism and GD susceptibility. The meta-analysis demonstrated that the rs3761548 polymorphism was significantly associated with GD (A vs. C: OR = 1.32, 95%CI = 1.05–1.67, Pz = 0.02; AC vs. CC: OR = 1.58, 95%CI = 1.01–2.45, Pz = 0.04; (AA+AC) vs. CC: OR = 1.51, 95%CI = 1.02–2.25, Pz = 0.04) (Table 2). Subgroup analysis according to ethnicity indicated that this polymorphism was significantly associated with GD in Asians (A vs. C: OR = 1.41, 95%CI = 1.13–1.77, Pz = 0.003; AC vs. CC: OR = 1.72, 95% CI = 1.06–2.79, Pz = 0.03; (AA + AC) vs. CC: OR = 1.66, 95%CI = 1.10–2.53, Pz = 0.02) but not in Caucasians. A forest plot of the relationship between the rs3761548 polymorphism and the risk of GD is shown in Figure 1.

Table 2.

Meta-analysis of the association between the rs3761548 polymorphism and Graves’ disease susceptibility.

Polymorphism	Comparison model	Subgroup	No. of studies	Sample size (cases/controls)	Test of association				Test of heterogeneity
					OR	95% CI	P Z	Effect model	I2	P H
rs3761548	Allele comparison (A vs. C)	Overall	6	2524/3032	1.32	1.05–1.67	0.02	Random	64%	0.02
		Asian	5	2310/2884	1.41	1.13–1.77	0.003	Random	57%	0.05
		Caucasian	1	214/148	0.89	0.59–1.36	0.60	_	_	_
	Homozygote comparison (AA vs. CC)	Overall	6	780/1001	1.26	0.92–1.72	0.15	Fixed	39%	0.15
		Asian	5	716/956	1.37	0.97–1.93	0.07	Fixed	41%	0.15
		Caucasian	1	64/45	0.82	0.38–1.79	0.62	_	_	_
	Heterozygous comparison (AC vs. CC)	Overall	6	1152/1376	1.58	1.01–2.45	0.04	Random	77%	0.0005
		Asian	5	1069/1321	1.72	1.06–2.79	0.03	Random	79%	0.0007
		Caucasian	1	83/55	0.96	0.49–1.91	0.92	_	_	_
	Dominant model (AA+AC vs. CC)	Overall	6	1262/1516	1.51	1.02–2.25	0.04	Random	75%	0.001
		Asian	5	1155/1442	1.66	1.10–2.53	0.02	Random	75%	0.003
		Caucasian	1	107/74	0.91	0.49–1.68	0.76	_	_	_
	Recessive model (AA vs. AC+CC)	Overall	6	1262/1516	1.11	0.84–1.47	0.45	Fixed	20%	0.28
		Asian	5	1155/1442	1.18	0.87–1.60	0.29	Fixed	27%	0.24
		Caucasian	1	107/74	0.84	0.42–1.6	0.61	_	_	_

OR, odds ratio; CI, confidence interval.

Figure 1.

Forest plot of the association between FOXP3/rs3761548 polymorphism and risk of Graves’ disease in the dominant model (AA + AC vs. CC).

CI, confidence interval; M-H, Mantel–Haenszel.

FOXP3/rs3761549 polymorphism

Seven studies including 1799 GD patients and 1970 controls were examined to evaluate the association between the rs3761549 SNP and GD susceptibility. There was a significant relationship between GD and FOXP3/rs3761549 genotype (T vs. C: OR = 1.32, 95%CI = 1.03–1.70, Pz = 0.03; TT vs. CC: OR = 1.98, 95%CI = 1.49–2.65, Pz<0.00001; TC vs. CC: OR = 1.23, 95%CI = 1.03–1.45, Pz = 0.02; (TT+TC) vs. CC: OR = 1.44, 95%CI = 1.11–1.88, Pz = 0.006; TT vs. (TC+CC): OR = 1.71, 95%CI = 1.04–2.81, Pz = 0.03) (Table 3). According to subgroup analysis stratified by ethnicity, we found a significant association between the FOXP3/rs3761549 polymorphism and GD risk in both Asians (T vs. C: OR = 1.38, 95%CI = 1.07–1.78, Pz = 0.01; TT vs. CC: OR = 1.92, 95%CI = 1.17–3.15, Pz = 0.01; (TT+TC) vs. CC: OR = 1.38, 95%CI = 1.05–1.79, Pz = 0.02) and Caucasians (TT+TC vs. CC: OR = 2.31, 95%CI = 1.08–4.96, Pz = 0.03).

Table 3.

Meta-analysis of the association between the rs3761549 polymorphism and Graves’ disease susceptibility.

Polymorphism	Comparison model	Subgroup	No. of studies	Sample size (cases/controls)	Test of association				Test of heterogeneity
					OR	95% Cl	P Z	Effect model	I2	P H
rs3761549	Allele comparison (T vs. C)	Overall	7	3625/3812	1.32	1.03–1.70	0.03	Random	75%	0.0006
		Asian	5	2288/2866	1.38	1.07–1.78	0.01	Random	69%	0.01
		Caucasian	2	1337/946	1.29	0.53–3.14	0.57	Random	82%	0.02
	Homozygote comparison (TT vs. CC)	Overall	6	843/1023	1.98	1.49–2.65	<0.00001	Fixed	47%	0.09
		Asian	5	764/959	1.92	1.17–3.15	0.01	Random	57%	0.05
		Caucasian	1	79/64	2.46	0.10–61.54	0.58	_	_	_
	Heterozygous comparison (TC vs. CC)	Overall	6	1120/1400	1.23	1.03–1.45	0.02	Fixed	45%	0.11
		Asian	5	1012/1325	1.18	1.00–1.41	0.06	Fixed	39%	0.16
		Caucasian	1	108/75	2.24	1.04–4.81	0.04	_	_	_
	Dominant model (TT+TC vs. CC)	Overall	6	1253/1508	1.44	1.11–1.88	0.006	Random	52%	0.06
		Asian	5	1144/1433	1.38	1.05–1.79	0.02	Random	53%	0.07
		Caucasian	1	109/75	2.31	1.08–4.96	0.03	_	_	_
	Recessive model (TT vs. TC+CC)	Overall	6	1253/1508	1.71	1.04–2.81	0.03	Random	56%	0.05
		Asian	5	1144/1433	1.70	1.00–2.88	0.05	Random	65%	0.02
		Caucasian	1	109/75	2.09	0.08–51.94	0.65	_	_	_

OR, odds ratio; CI, confidence interval.

FOXP3/rs3761547 polymorphism

We finally investigated the relationship between the FOXP3/rs3761547 polymorphism and GD susceptibility, but found no significant associations in either the overall or stratified analysis.

Sensitivity analysis and publication bias

We performed sensitivity analysis to assess the influence of individual studies on the pooled results. Sequential omission of individual studies from the pooled analysis had no effect on the overall pooled results, indicating that the results of the analysis were statistically robust. We evaluated possible publication bias by funnel plots, which showed no obvious indication of publication bias.

Discussion

In the present study, we evaluated the associations between FOXP3/rs3761547/rs3761548/rs3761549 polymorphisms and GD susceptibility based on seven eligible case–control studies including 4051 GD patients and 4569 controls. We showed that the rs3761548 and rs3761549 SNPs contributed to GD susceptibility. Furthermore, subgroup analysis according to ethnicity revealed that FOXP3/rs3761548 was associated with GD in Asians but not in Caucasians, while FOXP3/rs3761549 was associated in both Asians and Caucasians. However, there was no significant association between FOXP3/rs3761547 and GD.

FOXP3 is mainly expressed by CD4+ and CD25+ Tregs, and normal FOXP3 expression has been shown to be important for maintaining the inhibitory function of Tregs.[13] Genetic variations in the FOXP3 gene, leading to reduced expression, may thus promote the pathogenesis of GD by weakening the inhibitory function of Tregs and promoting an autoimmune response. The association between FOXP3 polymorphisms and GD susceptibility has attracted recent attention. FOXP3 can affect the differentiation of Tregs and is significantly associated with susceptibility to GD.[5,9,16,22,30,31] For instance, Fathima et al.[30] analyzed the correlation between FOXP3/rs3761548/rs3761549 polymorphisms and GD in an Indian population and found that presence of the rs3761549 T allele predisposed patients to GD. In addition, a research team from India investigated the association between FOXP3 promoter SNPs (rs3761547, rs3761548, and rs3761549) and GD in a Kashmiri population, and found significant differences between affected individuals and controls with respect to the genotype and allele frequencies of rs3761548 and rs3761549, but no significant association between rs3761547 and GD.[16] Zheng et al.[22] studied the possible associations of FOXP3 polymorphisms (rs3761547, rs3761548, rs3761549, and rs2280883) with GD in a Chinese Han population, based on 308 GD patients and 306 healthy controls, and found that the frequencies of the AA/CA genotype of rs3761548 and the CC genotype of rs2280883 were linked to an increased risk of GD. In addition, the AA/CA genotype of rs3761548 was more frequent in female than in male GD patients. With regard to rs3761548, GD patients with higher thyroid-stimulating hormone levels or lower thyrotropin receptor antibody levels were more likely to carry the A allele. Another case–control study including 534 Chinese Han patients with GD and 630 healthy controls showed that heterozygote and minor allele (rs3761548, rs3761549, and rs2280883) frequencies were significantly higher in GD patients than in healthy volunteers. The results suggest that FOXP3/rs3761548/rs3761549/rs2280883 polymorphisms are associated with GD susceptibility in the Chinese Han population.[5]

Inoue et al.[31] genotyped FOXP3 polymorphisms (rs3761547, rs3761548, rs3761549, and rs2280883) in a Japanese population including 65 patients with intractable GD, 44 patients with GD in remission, and 71 healthy subjects. They showed that the CA genotype of rs3761548 was more frequent in patients with GD in remission than in patients with intractable GD, and that the AA genotype of rs3761548, which is related to defective transcription of FOXP3, was absent in patients with GD in remission. Bossowski et al.[9] investigated the role of FOXP3 SNPs (rs3761547, rs3761548, and rs3761549) in GD susceptibility in a Polish study of 145 GD patients and 161 healthy subjects, and showed that rs3761549 G/A and rs3761547 T/C were more frequent in female GD patients compared with control females. In conclusion, these results suggest that the FOXP3 rs3761549 G/A variant may contribute to the development of GD in female patients. All of the above findings support a role for FOXP3 polymorphisms in GD susceptibility.

In contrast to the above results however, Owen et al.[32] investigated the association between FOXP3 polymorphisms (rs3761549, rs2280883, rs2232365, rs2294021, rs6609857, (GT)n, and (TC)n) and GD in a Caucasian population from the northeast of England (633 GD and 528 controls) and found no link between FOXP3 polymorphisms and susceptibility to GD in the UK population.

The current meta-analysis differs from two previous meta-analyses examining the association between the FOXP3/rs3761548 variant and autoimmune diseases performed by Lee et al.[18] and He et al.[33] Both these previous studies found that the FOXP3/rs3761548 SNP was associated with susceptibility to autoimmune diseases in Asian populations. However, their meta-analyses only included a few studies and a small number of GD patients, and they did not carry out an independent analysis of the relationship between FOXP3/rs3761548 polymorphism and GD susceptibility. Their results could therefore not fully explain the relationship between the FOXP3/rs3761548 polymorphism and susceptibility to GD. The present meta-analysis had several advantages, including being the first meta-analysis to focus on the relationship between the three FOXP3 SNPs (rs3761547, rs3761548, and rs3761549) and the risk of GD. Additionally, compared with former meta-analyses of autoimmune diseases,[18,33] the current analysis included more studies of patients with GD and more FOXP3 loci, as well as carrying out supplementary tests including subgroup and sensitivity analyses. Moreover, the results for rs3761548 and rs3761547 did not change significantly, even after excluding studies with a HWE P-value <0.05, indicating that our meta-analysis results were reliable.

This study had several limitations. First, the sample size in our meta-analysis was relatively small due to a shortage of original studies, and only two studies in the subgroup analyses investigated the genetic effects of FOXP3 polymorphisms in Caucasian populations. Further studies with larger sample sizes and additional ethnic populations are therefore needed to verify the present findings. In addition, raw data such as information on sex, lifestyle, clinical factors, and environmental exposure could not be obtained from all of the included studies. Furthermore, the current analysis did not contain enough data to analyze gene–environment and gene–gene interactions.

Conclusion

The current meta-analysis found significant associations between the FOXP3/rs3761548 and FOXP3/rs3761549 SNPs and susceptibility to GD in Asian populations.

Click here for additional data file.

Supplemental material, sj-pdf-1-imr-10.1177_03000605211004199 for Meta-analysis reveals significant association between FOXP3 polymorphisms and susceptibility to Graves’ disease by Guiqin Tan, Xin Wang, Guangbing Zheng, Juan Du, Fangyu Zhou, Zhongzhi Liang, Wenwen Wei and Hongsong Yu in Journal of International Medical Research