Association Analysis of Single Nucleotide Polymorphisms in C1QTNF6,RAC2,and an Intergenic Region at 14q32.2 with Graves' Disease in Chinese Han Population

Abstract

Background:

Variation within the C1QTNF6 gene at 22q12.3, the RAC2 gene at 22q13.1, and an intergenic region at 14q32.2 were found to be associated with risk to Graves' disease (GD) in a recent study. We aimed to validate these associations with GD in an independent sample set of Han Chinese population.

Methods:

We investigated these associations by genotyping the most significantly associated single nucleotide polymorphisms (SNPs) located in these three regions. Rs1456988 within the intergenic region at 14q32.2, rs229527 within C1QTNF6 at 22q12.3, and rs2284038 within RAC2 at 22q13.1 were selected for genotyping. These three SNPs were genotyped using a case-control study that included 2382 GD patients and 3092 unrelated healthy controls from Northern Han Chinese ancestry. The genotyping was performed using TaqMan assays on the ABI7900 platform.

Results:

We found both the rs229527 allele within C1QTNF6 (odds ratio [OR] = 1.23, confidence interval [95% CI]: 1.12-1.33, p_Allelic = 4.60 × 10⁻⁶) and the rs2284038 allele within RAC2 (OR = 1.10, 95% CI: 1.01-0.19, p_Allelic = 3.00 × 10⁻²) showed significant associations with GD susceptibility. However, rs1456988 located in 14q32.2 (OR = 1.08, 95% CI: 0.99-1.16, p_Allelic = 7.01 × 10⁻²) showed no association. Analysis of models of inheritance suggested that both the dominant and recessive models showed significant associations for rs229527 (OR = 1.24, 95% CI: 1.13-1.38, p_Dominant = 9.90 × 10⁻⁵; OR = 1.49, 95% CI: 1.19-1.86, p_Recessive = 3.90 × 10⁻⁴), with the dominant model being preferred. For rs2284038, the recessive model was preferred (OR = 1.18, 95% CI: 1.00-1.40, p_Recessive = 4.76 × 10⁻²), whereas analysis of dominant model showed no association (OR = 1.10, 95% CI: 0.98-1.22, p_Dominant = 0.10).

Conclusions:

Our findings confirmed that chromosome 22q12.3 and 22q13.1 variants are associated with GD in an independent Han Chinese population; however, 14q32.2 showed no association with GD.

Introduction

Graves' disease (GD) is one of the most common forms of the autoimmune thyroid diseases (AITDs) and was described as a syndrome consisting of hyperthyroidism, goiter, and orbitopathy. Hyperthyroidism is the most common manifestation of GD caused by TSH-receptor autoantibodies (TRAb) binding to and activating the thyrotropin receptor (TSHR) on the follicular cells of the thyroid (Michalek et al., 2009). Population-based surveys showed the prevalence of clinical GD was up to 2-3% in China (Zhang et al., 2002). The adult Europeans were affected by GD with 0.5-2% approximately (Aoki et al., 2007). About 1% of the population in the United States was affected by AITDs and the prevalence of AITDs may be similar in Japan (Ban et al., 2013).

The precise pathogenesis of GD was not clarified; extensive researches gave abundant evidences that the tendency to develop GD might be genetically inherited. Twin and family studies indicated that genetic factors made up about 79% of the risk for development of GD (Brix et al., 2001). Up to now, several risk loci were found in repeated associations with GD in various ethnic populations, including the immune-related genes at 1p13 (PTPN22), 1q23.1 (FCRL3), 2q33 (CTLA-4), 5q32 (SCGB3A2), 6p21 (MHC), 20q12 (CD40), Xq21.1 (GPR174), and a thyroid-specific gene at 14q31 (TSHR) (Chu et al., 2011, 2013; Zhao et al., 2013). A recent genome-wide association study (GWAS) in Chinese population revealed two novel risk loci for GD, 4p14 (GDCG4p14) and 6q27 (RNASET2-FGFR1OP-CCR6), which were further confirmed in Polish and UK Caucasians (Chu et al., 2011; Cooper et al., 2012; Szymanski et al., 2012). All these risk loci partially explained the heritability of GD, meanwhile, indicating the existence of other unknown genetic factors.

A more recent three-stage GWAS study revealed five novel susceptibility loci, including C1QTNF6 (22q12.3), RAC2 (22q13.1), SLAMF6 (1q23.2), ABO (9q34.2), GPR174-ITM2A (Xq-21.1), and an intergenic region at 14q32.2 (Zhao et al., 2013). The single nucleotide polymorphism (SNP) with the most significant association to GD at 22q12.3 is rs229527, which is a nucleotide transversion (139T→G) in the second exon of C1QTNF6 that causes the amino acid substitution V21G. C1QTNF6 encoded C1q/TNF-related protein 6 (CTRP6), which was implicated in the regulation of immune-related disease development (Murayama et al., 2015). Rs2284038 is the most significantly associated SNP at 22q13.1, which is located in the second intron of RAC2. Human RAC2 belongs to the Ras superfamily and was involved in the pathogenesis in autoimmune diseases by regulating the activities of T cell, B cell, and neutrophil (Croker et al., 2002; Crispin et al., 2007; Sironi et al., 2011; Shelef et al., 2013). Rs1456988 was located within the largest human miRNA cluster at 14q32, known as the C14MC (Chr 14 mega cluster), which contains 56 miRNAs (Abuhatzira et al., 2015). Several miRNAs in the 14q32 cluster are predicted to target thyroid peroxidase (TPO) and thyroglobulin (TG), which are the major autoantigens in GD (Abuhatzira et al., 2015). Of note, variation within C1QTNF6 at 22q12.3, RAC2 at 22q13.1, and an intergenic region at 14q32.2 was found in association with other autoimmune disease in previous studies (Zhao et al., 2013).

In this study, we aimed to validate associations of three newly identified susceptibility loci with GD in North Chinese Han population. We carried out a case-control study to test the association of rs1456988 located on chromosome 14q32.2, rs229527 at C1QTNF6, and rs2284038 at RAC2 with GD in a Han Chinese sample set recruited from Weifang City of Shandong province, China.

Subjects and Methods

Subjects

We carried out a case-control study in a sample set consisting of 2382 Chinese Han GD patients and 3092 unrelated, sex- and age-matched healthy controls. Patients with GD were recruited from Department of Endocrinology, Weifang People's Hospital of Shandong province in China. Ethnically and geographically matched healthy Chinese Han volunteers were recruited as controls. All subject participants signed the informed written consent form with Institutional Review Board approval and this study was approved by the Ethics Committee of Weifang People's Hospital (No. 2012018).

GD was diagnosed based on clinical symptoms and documented biochemical confirmation of hyperthyroidism combined with diffuse goiter, positive TSHR antibody tests, Graves' ophthalmopathy, or diffusely increased 131I (iodine-131) uptake in the thyroid gland (Chu et al., 2011). Control subjects were recruited from the same geographic region as cases. All controls enrolled in this study have no personal and family history of thyroid disorders or any other autoimmune diseases.

Genotyping

A total of 2 mL venous blood was collected from each participant and genomic DNA was extracted from peripheral blood cells using the FlexiGene DNA Kit (Qiagen Hilden, Germany) according to the manufacturer's directions. DNA concentration and purity were measured by Nanodrop2000 (Thermo Fisher Scientific, Inc.). SNP genotyping was implemented using TaqMan SNP Genotyping Assays as previously described (Chu et al., 2011). In this study, rs1456988, rs229527, and rs2284038 were selected for genotyping, and the SNP Genotyping Assays were provided by Applied Biosystems (C___8425898_20 for rs1456988, C___3289757_20 for rs229527, and C___2493934_20 for rs2284038, respectively). Assays were performed according to the specifications on ABI 7900 Sequence Detection System (Applied Biosystems). The data completion rate of rs1456988, rs229527, and rs2284038 was 97.1%, 99.1%, and 98.9%, respectively.

Statistical methods

Statistical analyses were performed using Plink. Three genetic models, including the allelic, dominant, and recessive model, together with a genotypic association test (2df test) were used to analyze the association for each SNP. The odds ratio (OR) and 95% confidence intervals (95% CIs) were calculated according to Woolf's method (Woolf, 1955). A p-value <0.05 was considered significant. Hardy-Weinberg equilibrium were tested for each locus in both case and control groups.

An online genetics power calculator was used to estimate the genetics power for each SNP in this study with the sample size of 2382 cases and 3092 controls, and the disease prevalence of 1% (Zhang et al., 2002). Assuming a multiplicative genetic model and per-allele copy OR of 1.08 for rs1456988, 1.23 for rs229527, and 1.10 for rs2284038, our study had a power of 52%, 95%, and 51% at p < 0.05 to detect a risk allele with a population frequency of 0.54, 0.73, and 0.65 respectively.

Results

The demographic information of 2382 GD cases and 3092 controls was showed in Table 1. The gender ratio and the mean age of GD patients and controls exhibited no significant differences (p > 0.05). The distribution of genotype frequencies of three SNPs in all subject participants was in Hardy-Weinberg equilibrium (p > 0.05). The comparison of genotypic and allelic distribution between cases and controls are shown in Tables 2 and 3, respectively.

Table 1.

Demographic Information for the Samples

	Cases	Controls
Number of subjects	2382	3092
Female: Male	3.02: 1	2.96: 1
Average age at enrollment	40.3 ± 14.4	42.6 ± 11.7
Age range at enrollment	4-81	20-88

Table 2.

The Risk Allele of Three Single Nucleotide Polymorphisms in Case-Control and Association Results

					Risk allele frequency
Chr.	SNP	Chr. position ^a	Annotated gene	Risk allele	Control	Case	p	OR (95% CI)
14	rs1456988	98488007	Intergenic	G	0.54	0.55	7.01 × 10⁻²	1.08 (0.99-1.16)
22	rs229527	37581485	C1QTNF6	A	0.73	0.77	4.60 × 10⁻⁶	1.23 (1.12-1.33)
22	rs2284038	37635055	RAC2	A	0.65	0.67	3.00 × 10⁻²	1.10 (1.01-1.19)

GRCh38 was used as the reference assembly.

Chr, chromosome; CI, confidence interval; OR, odds ratio; SNP, single nucleotide polymorphism.

Table 3.

The Genotype Distributions of Three Single Nucleotide Polymorphisms in Case-Control and Association Results

					Genotype distribution N (%)		Genotypic	Dominant	Recessive
Chr.	SNP	Chr. position ^a	Annotated gene	Genotype	Case	Control	p	p value, OR (95% CI)
14	rs1456988	98488007	Intergenic	T/T	458 (19.5)	639 (21.6)	0.15	0.24, 0.93 (0.83-1.05)	6.53 × 10⁻², 0.88 (0.77-1.01)
				T/G	1183 (50.4)	1475 (49.8)
				G/G	704 (30.1)	845 (28.6)
22	rs229527	37581485	C1QTNF6	A/A	1393 (58.8)	1634 (53.6)	2.33 × 10⁻⁵	9.90 × 10⁻⁵, 1.24 (1.13-1.38)	3.90 × 10⁻⁴, 1.49 (1.19-1.86)
				C/A	847 (35.8)	1179 (38.6)
				C/C	127 (5.4)	238 (7.8)
22	rs2284038	37635055	RAC2	A/A	1057 (44.9)	1300 (42.6)	8.09 × 10⁻²	0.10, 1.10 (0.98-1.22)	4.76 × 10⁻², 1.18 (1.00-1.40)
				G/A	1040 (44.1)	1360 (44.6)
				G/G	259 (11.0)	389 (12.8)

GRCh38 was used as the reference assembly.

Rs229527 in C1QTNF6 region located at 22q12.3 showed significant association with GD (OR = 1.23, 95% CI: 1.12-1.33; p_Allelic = 4.60 × 10⁻⁶; Table 2). The frequency of the major allele A of rs229527 was significantly higher in GD cases (77%; Table 2) than in controls (73%; Table 2). A significant difference was observed in the genotype frequencies of rs229527 between patients with GD (C/C, 5.40%; C/A, 35.8%; and A/A, 58.8%) and healthy controls (C/C, 7.80%; C/A, 38.6%; and A/A, 53.6%, respectively) (p_Genotypic = 2.33 × 10⁻⁵). Analysis of model of inheritance revealed that the risk allele A of rs229527 was associated with GD susceptibility both under the recessive model (OR = 1.49, 95% CI: 1.19-1.86; p_Recessive = 3.90 × 10⁻⁴) and the dominant model (OR = 1.24, 95% CI: 1.13-1.38; p_Dominant = 9.90 × 10⁻⁵). The dominant model should be preferred.

Rs2284038 in RAC2 region located at 22q13.1 (OR = 1.10, 95% CI: 1.01-1.19; p_Allelic = 3.00 × 10⁻²; Table 2) showed significant association with GD. The major allele A of rs2284038 was associated with GD risk (OR = 1.10, 95% CI = 1.01-1.19; p_Allelic = 3.00 × 10⁻²). The frequency of A allele was 67% in GD patients and 65% in controls. The distribution of rs2284038 genotype in GD patients (G/G, 26.4%; A/G, 50.3%; and A/A, 26.4%) was not significantly different from those in controls (G/G, 11.0%; A/G, 44.1%; and A/A, 44.9%, respectively) (p_Genotypic = 8.09 × 10⁻²). Analysis of model of inheritance showed that rs2284038 was associated with GD following a recessive model (OR = 1.18, 95% CI: 1.00-1.40; p_Recessive = 4.80 × 10⁻²), whereas no significant association was found following a dominant model (OR = 1.10, 95% CI: 0.98-1.22; p_Dominant = 0.10).

No association was observed between rs1456988 in the intergenic region located at 14q32.2 and GD using any genetic model (OR = 1.08, 95% CI: 0.99-1.16; p_Allelic = 7.01 × 10⁻²; Tables 2 and 3). The frequency of the major allele G of rs1456988 was slightly higher in GD cases (55%; Table 2) than in controls (54%; Table 2).

Discussion

Several studies reported that genetic polymorphisms of C1QTNF6, RAC2, and an intergenic region at 14q32.2 conferred significant risk for GD and other autoimmune diseases (Olsson et al., 2007; Cooper et al., 2008; Julia et al., 2008; Wallace et al., 2009; Jin et al., 2010; Sironi et al., 2011; Tang et al., 2013; Abuhatzira et al., 2015). In the current study, we genotyped SNPs with top risk in these three regions in a Chinese Han case-control sample collection from Weifang City of Shandong Province, which is a coastal city located at North China. Our results revealed that rs229527 at C1QTNF6 showed a significant association with GD and rs2284038 at RAC2 showed a slightly significant association with GD. However, rs1456988 located in 14q32.2 showed no association with GD susceptibility in our sample set.

Rs229527 in exon 2 of C1QTNF6 showed an association with GD in this Chinese Han population. C1QTNF6 encoded C1q/TNF-related protein 6 (CTRP6), which was implicated in the development of immune-related disease (Murayama et al., 2015). C1qtnf6^−/− mice were highly susceptible to rheumatoid arthritis for the enhanced complement activation (Murayama et al., 2015). Moreover, aged C1qtnf6^−/− mice were reported to produce higher levels of autoantibodies presented by the increased C3a and C5a, which were known to promote IgG production (Ricklin et al., 2010). The influence of CTRP6 on complement activation indicated that it might be involved in the development of RA as well as on other autoimmune and inflammatory diseases (Murayama et al., 2015). In addition, C1QTNF6 was found to induce the expression of interleukin-10 (IL-10), which was an anti-inflammatory cytokine, and modulated inflammatory signaling pathway (Kim et al., 2010). These findings made CTRP6 as a novel target for pharmacologically therapeutic drugs in immune-related diseases.

Of note, variation in the C1QTNF6 gene region was associated with type 1 diabetes mellitus (Cooper et al., 2008) and rheumatoid arthritis (Julia et al., 2008). SNP rs229541 at C1QTNF6 was an established risk locus to islet autoimmunity and type 1 diabetes (Frederiksen et al., 2013). It was also reported to be associated with risk for enterovirus infection in healthy children (Witso et al., 2015). Enterovirus infection could activate strong innate immune responses and was suggested as triggers of type 1 diabetes (Lind et al., 2012). Furthermore, a recent GWAS identified that rs229527 at C1QTNF6 was associated with vitiligo in Caucasians, and SNP rs2051582, resided between C1QTNF6 and IL2RB, was associated with susceptibility to vitiligo in Chinese (Jin et al., 2010; Tang et al., 2013). These findings suggested this chromosome band was a shared susceptibility locus of autoimmune diseases.

Rs2284038 is located in intron 2 region of RAC2 mapped on 22q13.1. Rs2284038 showed a slightly significant association with GD in these data. Chromosome band 22q13.1 was reported as a susceptibility locus for systemic sclerosis and autoimmune hepatitis type 1 (Zhou et al., 2003; de Boer et al., 2014). Multiple lines of evidences supported significant associations of RAC2 SNPs with rheumatoid arthritis, Crohn's disease, and multiple sclerosis (Olsson et al., 2007; Sironi et al., 2011; Muise et al., 2012). In addition, RAC2 mutation (D57N) was found to be a causal genetic factor for human neutrophil immunodeficiency syndrome (Ambruso et al., 2000). All these findings indicated that RAC2 was an important genetic determinant for human autoimmune disorders.

Human RAC2 (Ras-related C3 botulinum toxin substrate 2) is an important member of the Ras superfamily of GTP binding proteins. The expression of Rac2 is restricted to the hematopoietic lineage and played a major role in immune and inflammatory responses by regulating several processes (Sironi et al., 2011). First, RAC2 participates in the regulation of T cell proliferation and differentiation, T cell cytoskeletal restructuring, T cell activation, T cell receptor signaling, T cell migration, and T cell apoptosis (Yu et al., 2001; Pernis, 2009). Deficiency in the proper regulation of T cell functions underlined the development of autoimmune disorders (Crispin et al., 2007). Second, Rac2 was essential for B cell development and signaling. Study from Rac2-deficient mice suggested that Rac2 was an essential element in regulating B lymphocyte functions and maintaining B lymphocyte populations in vivo (Croker et al., 2002). In addition, recent findings suggested that Rac2 guided neutrophil recruitment, which could reverse migration away from sites of inflammation, and hence was involved in the pathogenesis of autoimmune diseases (Shelef et al., 2013).

Rs1456988 in an intergenic region at 14q32.2 showed no association with GD risk in these data. Except for association with GD susceptibility, the 14q32.2 region was also revealed as a susceptibility locus for type 1 diabetes (Barrett et al., 2009; Wallace et al., 2009). The 14q32 miRNA cluster neighboring rs1456988 contains 56 miRNAs, several of which were predicted to target TPO and TG, which are thyroid-specific genes (Abuhatzira et al., 2015). These findings gave a clue that the related miRNA might regulate the expression of TPO and TG. Therefore, the 14q32 miRNA cluster might be a potential functional candidate locus for GD. The risk allele frequencies of rs1456988 were 0.55 in cases and 0.54 in controls in these data, which was in the same effect direction as in previous study (0.56 in cases and 0.53 in controls, respectively) (Zhao et al., 2013). We estimated the power for rs1456988 in this study with the sample size of 2382 cases and 3092 controls, and the disease prevalence of 1% (Zhang et al., 2002). Assuming a multiplicative genetic model and per-allele copy OR of 1.08 for rs1456988, our study had only a power of 52% to detect a risk allele with a population frequency of 0.54 at p < 0.05. Therefore, the possible explanation of the inconsistency between this study and the previous report might be that the sample size was not large enough to detect the modest effect of rs1456988.

In summary, we replicated the associations of both rs229527 and rs2284038 with GD susceptibility in an independent North Han Chinese population in China. No association was observed between rs1456988 located in 14q32.2 and GD susceptibility in our sample set. These results further confirmed association of SNPs in C1QTNF6 and RAC2 region with GD. Further comprehensive studies are needed to evaluate these variants at the susceptibility loci of GD in individuals from different geographic and ethnic populations. Identification of GD susceptibility genes is vital in understanding the genetic architecture for disease predisposition.

Footnotes

Acknowledgments

We thank all subjects for participating in this study. This work was supported by National Natural Science Foundation of China (31271343, 31471190, 31671317, and 81272307) and Shanghai Science and Technology Commission of Shanghai Municipality (5411953400). Study design: W.H. and X.C.; Data analysis: X.C.; Article preparation: X.C. and Z.X.; Experiment: X.Z., M.S., F.L., Q.H. J.Z., and Z.W.; Clinical samples collection: L.L., L.P., and H.L.

Author Disclosure Statement

No competing financial interests exist.

References

Abuhatzira

, Xu

, Tahhan

, et al. (2015) Multiple microRNAs within the 14q32 cluster target the mRNAs of major type 1 diabetes autoantigens IA-2, IA-2beta, and GAD65. FASEB J, 29:4374-4383.

Ambruso

, Knall

, Abell

, et al. (2000) Human neutrophil immunodeficiency syndrome is associated with an inhibitory Rac2 mutation. Proc Natl Acad Sci U S A, 97:4654-4659.

Aoki

, Belin

, Clickner

, et al. (2007) Serum TSH and total T4 in the United States population and their association with participant characteristics: National Health and Nutrition Examination Survey (NHANES 1999-2002). Thyroid, 17:1211-1223.

Ban

, Tozaki

, Taniyama

(2013) The replication of the association of the rs9355610 within 6p27 with Graves' disease. Autoimmunity, 46:395-398.

Barrett

, Clayton

, Concannon

, et al. (2009) Genome-wide association study and meta-analysis find that over 40 loci affect risk of type 1 diabetes. Nat Genet, 41:703-707.

Brix

, Kyvik

, Christensen

, et al. (2001) Evidence for a major role of heredity in Graves' disease: a population-based study of two Danish twin cohorts. J Clin Endocrinol Metab, 86:930-934.

Chu

, Pan

, Zhao

, et al. (2011) A genome-wide association study identifies two new risk loci for Graves' disease. Nat Genet, 43:897-901.

Chu

, Shen

, Xie

, et al. (2013) An X chromosome-wide association analysis identifies variants in GPR174 as a risk factor for Graves' disease. J Med Genet, 50:479-485.

Cooper

, Simmonds

, Walker

, et al. (2012) Seven newly identified loci for autoimmune thyroid disease. Hum Mol Genet, 21:5202-5208.

10.

Cooper

, Smyth

, Smiles

, et al. (2008) Meta-analysis of genome-wide association study data identifies additional type 1 diabetes risk loci. Nat Genet, 40:1399-1401.

11.

Crispin

, Kyttaris

, Juang

, et al. (2007) Systemic lupus erythematosus: new molecular targets. Ann Rheum Dis, 66 Suppl 3:iii65-iii69.

12.

Croker

, Tarlinton

, Cluse

, et al. (2002) The Rac2 guanosine triphosphatase regulates B lymphocyte antigen receptor responses and chemotaxis and is required for establishment of B-1a and marginal zone B lymphocytes. J Immunol, 168:3376-3386.

13.

de Boer

, van Gerven

, Zwiers

, et al. (2014) Genome-wide association study identifies variants associated with autoimmune hepatitis type 1. Gastroenterology, 147:443-452.e445.

14.

Frederiksen

, Steck

, Kroehl

, et al. (2013) Evidence of stage- and age-related heterogeneity of non-HLA SNPs and risk of islet autoimmunity and type 1 diabetes: the diabetes autoimmunity study in the young. Clin Dev Immunol, 2013:417657.

15.

Jin

, Birlea

, Fain

, et al. (2010) Variant of TYR and autoimmunity susceptibility loci in generalized vitiligo. N Engl J Med, 362:1686-1697.

16.

Julia

, Ballina

, Canete

, et al. (2008) Genome-wide association study of rheumatoid arthritis in the Spanish population: KLF12 as a risk locus for rheumatoid arthritis susceptibility. Arthritis Rheum, 58:2275-2286.

17.

Kim

, Lee

, Park

, et al. (2010) C1qTNF-related protein-6 increases the expression of interleukin-10 in macrophages. Mol Cells, 30:59-64.

18.

Lind

, Huhn

, Flodstrom-Tullberg

(2012) Immunology in the clinic review series; focus on type 1 diabetes and viruses: the innate immune response to enteroviruses and its possible role in regulating type 1 diabetes. Clin Exp Immunol, 168:30-38.

19.

Michalek

, Morshed

, Latif

, et al. (2009) TSH receptor autoantibodies. Autoimmun Rev, 9:113-116.

20.

Muise

, Xu

, Guo

, et al. (2012) NADPH oxidase complex and IBD candidate gene studies: identification of a rare variant in NCF2 that results in reduced binding to RAC2. Gut, 61:1028-1035.

21.

Murayama

, Kakuta

, Inoue

, et al. (2015) CTRP6 is an endogenous complement regulator that can effectively treat induced arthritis. Nat Commun, 6:8483.

22.

Olsson

, Lindqvist

, Kallberg

, et al. (2007) A case-control study of rheumatoid arthritis identifies an associated single nucleotide polymorphism in the NCF4 gene, supporting a role for the NADPH-oxidase complex in autoimmunity. Arthritis Res Ther, 9:R98.

23.

Pernis

(2009) Rho GTPase-mediated pathways in mature CD4+ T cells. Autoimmun Rev, 8:199-203.

24.

Ricklin

, Hajishengallis

, Yang

, et al. (2010) Complement: a key system for immune surveillance and homeostasis. Nat Immunol, 11:785-797.

25.

Shelef

, Tauzin

, Huttenlocher

(2013) Neutrophil migration: moving from zebrafish models to human autoimmunity. Immunol Rev, 256:269-281.

26.

Sironi

, Guerini

, Agliardi

, et al. (2011) An evolutionary analysis of RAC2 identifies haplotypes associated with human autoimmune diseases. Mol Biol Evol, 28:3319-3329.

27.

Szymanski

, Bednarczuk

, Krajewski

, et al. (2012) The replication of the association of the rs6832151 within chromosomal band 4p14 with Graves' disease in a Polish Caucasian population. Tissue Antigens, 79:380-383.

28.

Tang

, Zhang

, Hu

, et al. (2013) Association analyses identify three susceptibility Loci for vitiligo in the Chinese Han population. J Invest Dermatol, 133:403-410.

29.

Wallace

, Smyth

, Maisuria-Armer

, et al. (2009) The imprinted DLK1-MEG3 gene region on chromosome 14q32.2 alters susceptibility to type 1 diabetes. Nature Genetics, 42:68-71.

30.

Witso

, Cinek

, Tapia

, et al. (2015) Genetic determinants of enterovirus infections: polymorphisms in type 1 diabetes and innate immune genes in the MIDIA study. Viral Immunol, 28:556-563.

31.

Woolf

(1955) On estimating the relation between blood group and disease. Ann Hum Genet, 19:251-253.

32.

, Leitenberg

, Li

, et al. (2001) Deficiency of small GTPase Rac2 affects T cell activation. J Exp Med, 194:915-926.

33.

Zhang

, Lin

, Fang

, et al. (2002) The effect of salt iodization for 10 years on the prevalences of endemic goiter and hyperthyroidism. Chin J Endocrinol Metab, 18:342-344.

34.

Zhao

, Xue

, Liu

, et al. (2013) Robust evidence for five new Graves' disease risk loci from a staged genome-wide association analysis. Hum Mol Genet, 22:3347-3362.

35.

Zhou

, Tan

, Wang

, et al. (2003) Genome-wide association study for regions of systemic sclerosis susceptibility in a Choctaw Indian population with high disease prevalence. Arthritis Rheum, 48:2585-2592.