Confirmation of Association of Chromosome 5q31–33 with United Kingdom Caucasian Graves' Disease

Abstract

Background:

Previous genome-wide microsatellite screening in Graves' disease (GD) has suggested several regions of linkage to disease. Although replication has been inconsistent, some regions such as chromosome 5q31–33 have been associated with several Oriental GD patient cohorts. Recently, two studies have reported association of single-nucleotide polymorphism (SNP) rs31480 in interleukin 3 (IL-3) and the rs1368408 and SNP75 (−623∼−622 AG/-T) SNPs in secretoglobulin family 3a member 2 (SCGB3A2) with GD and suggested that this may account for linkage to the 5q31–33 region in Oriental GD datasets. We sought to confirm this association in a large Caucasian U.K. GD cohort.

Methods:

The rs31480 SNP was shown to tag all known common variations in IL-3 and the rs1368408 SNP was shown to tag all common variations in SCGB3A2. The SCGB3A2 SNP75 was found to be rare in the U.K. Caucasian population and, therefore, was not screened. We genotyped rs31480 and rs1368408 and performed a case–control association study in 2504 GD cases and 2688 controls from the U.K.

Results:

Association between the SCGB3A2 rs1368408 SNP and GD was detected (p = 0.007, odds ratio = 1.18, 95% confidence intervals = 1.05–1.33). No association between the IL-3 rs31804 SNP and U.K. Caucasian GD patients was observed.

Conclusions:

These data suggest that chromosome 5q31–q33 contains a susceptibility locus for Caucasian GD patients as well as Oriental GD patients. Although association was detected between SCGB3A2 and U.K. Caucasian GD subjects, the size of effect was smaller than that seen in the Oriental population (odds ratio = 1.28–1.73). Fine mapping within this region will be required to determine the exact location of the etiological variants present within this region for both Caucasian and Oriental GD patients.

Introduction

S everal genome-wide screens employing microsatellite markers have been undertaken in Graves' disease (GD). The majority of reported regions of linkage have failed to replicate between studies (1 –6). Some regions did, however, map to either known susceptibility loci, including, for example, linkage on chromosome 6 mapping to the human leukocyte antigen (HLA) region that has been consistently associated with GD (1,7) and linkage on chromosome 2 mapping to the CTLA-4 region, or to novel genes, for example, linkage on chromosome 20 mapping to CD40 (8).

Chromosome 5q31–q33, which encodes a large variety of cytokines, inflammatory mediators, and other immune modulators, represents a region of linkage, which has generated much interest. Linkage of microsatellite marker D5S436 located within chromosome 5q31–33 was first detected in a Japanese GD genome-wide screen (5). Linkage was later replicated in a Han Chinese GD dataset along with additional more strongly associated microsatellite markers including D5S2090 (6) and in a second Japanese GD cohort to an extended 5q12–q33 region (4). Although no Caucasian GD genome-wide microsatellite screen has reported linkage to 5q31–33 (1 –3,9), some evidence for linkage to a separate region of chromosome 5, 5q11.2–5q14.3, has been previously detected (9). Several genes within 5q31–33 have been proposed to explain the association of this region with GD, including beta-2-adrenergic receptor (ADRB2) (10), interleukin 12b (IL-12b) (11,12), interferon regulatory factor 1 (13), IL-13 (13 –15), and IL-4 (13), although replication of these loci has proved difficult. More recently, two independent studies attempted to narrow down association within this region and have identified two distinct GD susceptibility loci approximately 15 Mb apart, which could potentially explain linkage to GD in this region.

The first study used 51 Tag single-nucleotide polymorphisms (SNPs) to screen 19 genes, chosen because of their potential as candidate loci for GD within the 5q31–33 region, in 428 Han Chinese GD and 690 control subjects (16). The rs40401 SNP (and the rs31480 SNP, which was in complete linkage disequilibrium [LD]) located within IL-3 (p = 0.002, odds ratio [OR] = 1.51) showed the strongest association, even after the addition of a further seven Tag SNPs within the surrounding haplotype block (16). Association of rs40401 was further replicated in an independent dataset composed of 468 Han Chinese GD and 471 control subjects (p = 0.0001, OR = 1.76) (16). A second study investigated a total of 179 SNPs within a 3-Mb region containing 25 genes surrounding microsatellite marker D5S2090 in 384 Han Chinese GD and 382 controls (17). Four SNPs were associated with GD, with rs1368408 being the most associated one, located in secretoglobulin family 3a member 2 (SCGB3A2) (17). Further screening of 122 SNPs in a 1-Mb region surrounding rs1368408 in 541 Han Chinese GD and 478 controls revealed 20 SNPs associated with GD, with logistic regression revealing the SCGB3A2 rs1368408 (p = 4.11 × 10⁻⁸, OR = 1.73) and SNP75 (-623∼-622 AG/-T) (p = 1.37 × 10⁻⁸, OR = 2.17) as the most important SNPs in this region. Association of rs1368408 (p = 1.43 × 10⁻⁶, OR = 1.28) and SNP75 (p = 7.62 × 10⁻⁵, OR = 1.32) was replicated in another much larger dataset of 2811 Han Chinese GD and 2807 controls, with rs1368408 (p = 0.001, OR = 1.47) also showing association in a further independent dataset of 545 Han Chinese GD and 603 control subjects from Shanghai, although they failed to detect association of SNP75 (p = 0.1348). Haplotypes containing rs1368408 and rs6882292 (another SNP in SCGB3A2) or rs1368408 and SNP75 were also shown to be associated with GD (p = 0.0007 and p = 0.019, respectively), with functional analysis revealing that presence of these haplotypes correlated with lower SCGB3A2 promoter efficiency, leading to lower SCGB3A2 expression (17). The aim of our study was, therefore, to determine if IL-3 and SCGB3A2 also play a role in GD susceptibility in the U.K. Caucasian population.

Materials and Methods

Subjects

A dataset consisting of 2504 unrelated white Caucasian patients with GD from the U.K. National Autoimmune Thyroid Disease Collection was screened. Patients were recruited from specialist clinics in Birmingham, Bournemouth, Cambridge, Cardiff, Exeter, Leeds, Sheffield, and Newcastle in the United Kingdom. Clinical definition of GD was as previously described (18). GD was defined by the presence of biochemical hyperthyroidism with (a) a diffuse goiter on a radionuclide or sonographic scan, (b) Graves' ophthalmopathy (no signs or symptoms; only signs, no symptoms; signs only; proptosis; eye muscle involvement; corneal involvement; sight visual acuity reduction [NOSPECS] classification score ≥2) (19), (c) positive autoantibodies to the thyroid stimulating hormone receptor (TSHR), (d) diffuse goiter on physical examination and positive antibodies to thyroglobulin or thyroid peroxidase, or (e) confirmation of a lymphocytic infiltrate in thyroid histology. A total of 2688 geographically matched white Caucasian control subjects were also obtained from the 1958 British Birth Cohort as previously stated (20). All patients gave informed written consent and the project was approved by the local ethics committee.

Tag SNP selection and genotyping

To screen the IL-3 promotor rs31480 SNP (a proxy for rs40401) and any other common variation within IL-3, genotyping data were downloaded from the International Haplotype Mapping Project website (www.hapmap.org). SNP rs31480 was found to tag all known common polymorphisms (major allele frequency ≥0.05) within IL-3 in the Caucasian (CEPH [Utah residents with ancestry from Northern and Western Europe]) population present within this region (Phase II, Build 36). To screen the SCGB3A2 rs1368408 and SNP75 SNPs, genotyping data were downloaded, which showed that rs1368408 tagged all known common variations within SCGB3A2 in the U.K. Caucasian population. Both the rs1368408 and rs31480 Tag SNPs were purchased from Applied Biosystems, Warrington, United Kingdom, as prevalidated assays on demand. SCGB3A2 SNP75 was not present on Hapmap and no prevalidated assay on demand was available from Applied Biosystems, and so a custom Taqman assay was designed. To genotype SNP75, forward 5′ GTCAAATCCAAGATTACACAAACAGTGAA 3′ and reverse 5′ GTCTTCTCTGAGCCTTTGGATAGAATAAT 3′ primers were designed with the probe 5′ CAGGTTTTTCAAAAGACACT 3′ labeled with the VIC fluorophore to detect the −623 AG and the 5′ CAGGTTTTTCAAATACACT 3′ probe labeled with the FAM fluorophore to detect −622-T (Applied Biosystems). All genotyping was performed using Taqman^® genotyping technology on an ABI7900HT (Applied Biosystems) and all genotyping plots were independently verified by two investigators to prevent sample miscalling. When the SCGB3A2 SNP75 was run on the first 768 cases and 768 controls of our dataset, all samples had two copies of the −623 AG, with the exception of one case that had one copy of the −623 AG and one copy of the −622-T, suggesting that we would not have the power to detect such a rare variant within our dataset (major allele frequency <1%).

Clinical phenotype correlations

Association analysis of SNP genotype with specific clinical phenotype was performed as described previously (21).

Statistical and haplotype analyses

We had greater than 80% power to exclude an OR = 1.15 for rs31480 and an OR = 1.17 for rs1368408 at α = 0.05. Analysis of case–control data was performed using the χ ² test within the MINTAB statistical package (MINITAB Release 15, © 1972–2003, Minitab, State College, PA) and p < 0.05 was considered significant. OR with 95% confidence intervals were calculated by the method of Woolf with Haldane's modification for small numbers, where appropriate (22). Power and Hardy–Weinberg equilibrium calculations were performed using formula written within Excel (Microsoft^® Office Excel, © 2003; Microsoft, Redmond, WA) based on established statistical methodology (23,24).

Results

Genotyping results

Both SNPs were in Hardy–Weinberg equilibrium in both the cases and controls. The SCGB3A2 rs1368408 SNP was shown to be associated within our dataset (p = 0.007, OR = 1.18, 95% confidence intervals = 1.05–1.33; Table 1). No association of the IL-3 rs31480 SNP was detected with GD (p = 0.295).

Table 1.

Association of the SCGB3A2 rs1368408 and IL-3 rs31480 Single-Nucleotide Polymorphisms in a United Kingdom Caucasian Graves' Disease Dataset

Gene	SNP	Allele/genotype	Graves' disease, n (%)	Controls, n (%)	Chi-square	p-Value	OR	95% CI
SCGB3A2	rs1368408	G	4270 (87.3)	4569 (89.0)	7.310	0.007	1.18	1.05–1.33
		A	622 (12.7)	563 (11.0)
		GG	1870 (76.4)	2029 (79.1)	9.519	0.009
		GA	530 (21.7)	511 (19.9)
		AA	46 (1.9)	26 (1.0)
IL-3	rs31480	C	3786 (78.9)	3974 (78.1)	1.097	0.295	—	—
		T	1010 (21.1)	1116 (21.9)
		CC	1499 (62.5)	1541 (60.6)	2.654	0.265
		CT	788 (32.9)	892 (35.0)
		TT	111 (4.6)	112 (4.4)

All SNPs are reported 5′–3′.

SNP, single-nucleotide polymorphism; OR, odds ratio; 95% CI, 95% confidence interval.

Clinical phenotype associations

Tests of correlation were performed with both the SCGB3A2 rs1368408 and IL-3 rs31480 SNPs with GD subphenotype, including severity of ophthalmopathy as determined by the NOSPECS classification (NOSPECS <2 vs. NOSPECS ≥2), age of onset of GD (<31 years vs. ≥31 years), biochemical severity as indicated by serum-free T₄ concentration at time of diagnosis (serum-free T₄ concentrations of <40 or ≥40 pmol/L), the presence or absence of palpable goiter, and autoantibody-positive subjects compared with autoantibody-negative subjects (for TSHR, thyroglobulin, or thyroid peroxidase antibodies). For rs1368408, a weak allelic association was detected with lower levels of free T₄ (p = 0.037), although association with individual genotypes was not detected (p = 0.085). No further associations were observed between the rs1368408 and rs31480 SNPs and any other of the subphenotypes investigated (data available upon request).

Discussion

Although linkage to chromosome 5q31–33 had been found in the Oriental population, no Caucasian GD microsatellite genome scan has ever reported linkage to 5q31–33 (1 –3,9). Weak support for a role for chromosome 5q31–q33 in the development of GD in Caucasian patients has come from the Wellcome Trust Case Control Consortium (WTCCC). The WTCCC performed a 15,000 nsSNP genome screen in which an SNP within ADRB2, rs1042714 (p = 0.04), which is located approximately 948 kb away from SCGB3A2, and a further SNP in RAD50, RAD50-327 (p = 0.03), which is approximately 516 kb away from IL-3, were weakly associated with GD (20). In this study we have replicated the association of SCGB3A2 rs1368408, located within chromosome 5q31–33, which has been previously shown to be associated with GD in a Japanese GD cohort.

Although SCGB3A2 may not seem like an obvious candidate for GD, insights about its role within the immune response suggest a possible explanation for its association with GD. SCGB3A2 encodes a secretary protein that binds to the macrophage scavenger receptor with collagenous structure (MARCO). MARCO is a scavenger receptor present on a subpopulation of macrophages and is believed to be involved in binding lipopolysaccharides and gram-positive and gram-negative bacteria (25). Knockout mouse studies showed that MARCO-deficient mice demonstrated an impaired ability to clear bacteria from their lungs compared with wild-type mice, suggesting a key role for MARCO in host defense against bacterial infections (26). Although MARCO has been shown to be expressed in the spleen, lymph nodes, and thymus (17,27), low levels of MARCO and SCGB3A2 have been more recently detected within the thyroid (17). In lung tissue it has been proposed that SCGB3A2 expressed on the surface of epithelial cells of the bronchioles interacts with MARCO to enable functional interaction between macrophages and lung epithelial tissue to aid the removal of bacteria and other pathogens from the lung (27). It is appealing to hypothesize that, in thyroid, MARCO and SCGB3A2 could interact in a similar manner. Interestingly, the presence of the rs1368408 allele A in a haplotype containing the SCGB3A2 SNP75 allele T or rs6882292 allele A was shown to reduce SCGB3A2 mRNA levels in human thyroid tissue (17). As bacteria are proposed to be one environmental trigger for GD (7,28), any disruption or variations within this interaction could alter how bacteria and other pathogens are recognized by the immune system and the response triggered, which could possibly cross react to cause thyroid autoimmunity. SCGB3A2 is also believed to be a target of homeodomain transcription factor T/EBP (TTF-1), which regulates expression of thyroid-specific genes, including thyroglobulin, thyroid peroxidase, and the TSHR (29 –32). Disruption within this control system could also contribute to the onset of GD, although further work is required to determine the exact role by which SCGB3A2 contributes to GD (17).

We failed to replicate association of the IL-3 rs31480 SNP within our Caucasian GD cohort. This can be due to several potential explanations. The original associations in the Oriental populations could have been a false positive or the size of effect overinflated because of a first time effect (33). This is unlikely as the original study replicated association of IL-3 with GD in two small independent Oriental GD datasets. Our dataset had >80% power to exclude an OR ≥1.15 for rs31480, and although we cannot rule out a minor effect for this SNP in our dataset, we can confidently exclude the same level of association in the Caucasian GD population as was seen in the Oriental GD population (OR = 1.51–1.76) (16). GD is a complex disease caused by both environmental and genetic factors, with well-established differences in presentation between patients including, for example, variation in degrees of thyroid eye disease, autoantibody titers, and appearance of goiter. Polymorphism within IL-3 could be linked to a specific clinical subphenotype or environmental trigger that is more prevalent in GD onset and progression in Oriental patients compared with Caucasian patients. However, our study revealed no major association with disease subphenotype.

It is of note that the magnitude of the association of SCGB3A2 rs1368408 seen in our GD population (OR = 1.18) was lower than that seen in the Oriental cohorts (OR = 1.28–1.73) (17). This could be explained by the well-reported, first time effect phenomenon (33). Equally, the differences observed between the Caucasian and Oriental populations could represent ethnic or geographical differences in genetic architecture. Levels of SNP diversity, allele frequencies, and LD patterns are known to vary significantly between Oriental and Caucasian populations. This suggests that although an etiological variant could be localized to a given set of SNPs in one population, diversity within the other population combined with differences in LD patterns could mean that the location of the etiological variant within the other population could be linked to a different set of SNPs located either elsewhere in the same gene or within the surrounding region. For example, association of the TSHR with GD in U.K. Caucasians has been refined to a probable 40-kb region within intron 1, whereas within the Japanese GD population the largest association is located within intron 7 (34,35).

Our study tagged all known common variations present within SCGB3A2 and IL-3 according to Caucasian Hapmap data (Build 36). However, within the Oriental populations many more SNPs were detected within these genes, suggesting that Hapmap data for this region at the time of project inception may have been limited. We may have missed other common variants within this region, which could be part of another haplotype not tagged by rs1368408.

Conclusion

Although association of SCGB3A2 rs1368408 has been found with GD in U.K. Caucasians, further SNP coverage surrounding SCGB3A2 is essential to localize the etiological variant within this region. This next step is, however, at present hampered by the lack of known SNP data within this region for Caucasians, making it difficult to determine the boundaries of LD. Data from the soon to be published 1000-genome project are likely to provide much more detailed information on the number and location of SNPs and the haplotype structure for Caucasians (www.1000genomes.org/page.php). This will help define the boundaries of LD within this region where the etiological variant is likely to reside. To further help narrow down the likely location of the etiological variant within this region, fine mapping of this region could be performed in Oriental GD populations in which these effects were first detected. Once fine mapping around these genes has been performed and etiological variants identified, LD block structure could then be investigated to determine the differences within this region between the Oriental and Caucasian populations. Our study confirms association of a region previously linked to GD and provides further support to proceed onto fine-mapping studies.

Footnotes

Acknowledgments

The authors would like to thank patients, nurses, and doctors for recruiting into the Autoimmune Thyroid Disease National Collection, and the Wellcome Trust for funding. The authors acknowledge use of DNA from the British 1958 Birth Cohort Collection, funded by the Medical Research Council grant G0000934 and the Wellcome Trust grant 068545/Z/02.

Disclosure Statement

None of the authors have any conflicts of interest or any competing financial interest to declare.

References

Tomer

, Ban

, Concepcion

, Barbesino

, Villanueva

, Greenberg

, Davies

. 2003. Common and unique susceptibility loci in Graves and Hashimoto diseases: results of whole-genome screening in a data set of 102 multiplex families. Am J Hum Genet, 73:736–747.

Tomer

, Barbesino

, Greenberg

, Concepcion

, Davies

. 1999. Mapping the major susceptibility loci for familial Graves' and Hashimoto's diseases: evidence for genetic heterogeneity and gene interactions. J Clin Endocrinol Metab, 84:4656–4664.

Taylor

, Gough

, Hunt

, Brix

, Chatterjee

, Connell

, Franklyn

, Hegedus

, Robinson

, Wiersinga

, Wass

, Zabaneh

, Mackay

, Weetman

. 2006. A genome-wide screen in 1119 relative pairs with autoimmune thyroid disease. J Clin Endocrinol Metab, 91:646–653.

Akamizu

, Hiratani

, Ikegami

, Rich

, Bowden

. 2003. Association study of autoimmune thyroid disease at 5q23-q33 in Japanese patients. J Hum Genet, 48:236–242.

Sakai

, Shirasawa

, Ishikawa

, Ito

, Tamai

, Kuma

, Akamizu

, Tanimura

, Furugaki

, Yamamoto

, Sasazuki

. 2001. Identification of susceptibility loci for autoimmune thyroid disease to 5q31-q33 and Hashimoto's thyroiditis to 8q23-q24 by multipoint affected sib-pair linkage analysis in Japanese. Hum Mol Genet, 10:1379–1386.

Jin

, Teng

, Ben

, Xiong

, Zhang

, Xu

, Shugart

, Jin

, Chen

, Huang

. 2003. Genome-wide scan of Graves' disease: evidence for linkage on chromosome 5q31 in Chinese Han pedigrees. J Clin Endocrinol Metab, 88:1798–1803.

Gough

SCL

, Simmonds

. 2007. The HLA region and autoimmune disease: Associations and mechanisms of action. Curr Genomics, 8:453–465.

Jacobson

, Tomer

. 2007. The CD40, CTLA-4, thyroglobulin, TSH receptor, and PTPN22 gene quintet and its contribution to thyroid autoimmunity: back to the future. J Autoimmun, 28:85–98.

Allen

, Hsueh

, Sabra

, Pollin

, Ladenson

, Silver

, Mitchell

, Shuldiner

. 2003. A genome-wide scan for autoimmune thyroiditis in the Old Order Amish: replication of genetic linkage on chromosome 5q11.2-q14.3. J Clin Endocrinol Metab, 88:1292–1296.

10.

Chu

, Dong

, Shen

, Sun

, Dong

, Wang

, Zhang

, Hua

, Xu

, Huang

. 2009. Polymorphisms in the ADRB2 gene and Graves disease: a case-control study and a meta-analysis of available evidence. BMC Med Genet, 10:26.

11.

Ikeda

, Yoshida

, Noguchi

, Asaba

, Nishioka

, Takao

, Hashimoto

. 2004. Lack of association between IL-12B gene polymorphism and autoimmune thyroid disease in Japanese patients. Endocr J, 51:609–613.

12.

Hiromatsu

, Fukutani

, Ichimura

, Mukai

, Kaku

, Miyake

, Yamada

. 2006. Interleukin-12B gene polymorphism does not confer susceptibility to Graves' ophthalmopathy in Japanese population. Endocr J, 53:753–759.

13.

Yang

, Lingling

, Ying

, Yushu

, Zhongyan

, Wei

, Weiping

. 2005. Association study between the IL4, IL13, IRF1 and UGRP1 genes in chromosomal 5q31 region and Chinese Graves' disease. J Hum Genet, 50:574–582.

14.

Simmonds

, Heward

, Franklyn

, Gough

. 2005. IL-13 and chromosome 5q31-q33: problems of identifying association within regions of linkage to Graves' disease. Clin Endocrinol (Oxf), 63:695–697.

15.

Hiromatsu

, Fukutani

, Ichimura

, Mukai

, Kaku

, Nakayama

, Miyake

, Shoji

, Koda

, Bednarczuk

. 2005. Interleukin-13 gene polymorphisms confer the susceptibility of Japanese populations to Graves' disease. J Clin Endocrinol Metab, 90:296–301.

16.

Chu

, Dong

, Lei

, Sun

, Wang

, Dong

, Shen

, Wang

, Zhang

, Yang

, Li

, Yuan

, Wang

, Song

, Jin

, Xiong

, Huang

. 2009. Polymorphisms in the interleukin 3 gene show strong association with susceptibility to Graves' disease in Chinese population. Genes Immun, 10:260–266.

17.

Song

, Liang

, Shi

, Zhao

, Liu

, Zhao

, Peng

, Gao

, Tao

, Pan

, Shao

, Cheng

, Wang

, Yuan

, Xu

, Han

, Huang

, Chu

, Chen

, Sheng

, Li

, Su

, Gao

, Jia

, Jin

, Chen

. 2009. Functional SNPs in the SCGB3A2 promoter are associated with susceptibility to Graves' disease. Hum Mol Genet, 18:1156–1170.

18.

Manji

, Carr-Smith

, Boelaert

, Allahabadia

, Armitage

, Chatterjee

, Lazarus

, Pearce

, Vaidya

, Gough

, Franklyn

. 2006. Influences of age, gender, smoking, and family history on autoimmune thyroid disease phenotype. J Clin Endocrinol Metab, 91:4873–4880.

19.

Werner

. 1977. Modification of the classification of the eye changes of Graves' disease: recommendations of the Ad Hoc Committee of the American Thyroid Association. J Clin Endocrinol Metab, 44:203–204.

20.

WTCCC, TACS 2007 Association study of 14,500 nsSNPs in four common diseases identifies variants involved in autoimmunity. Nat Genet, 39:1329–1337.

21.

Simmonds

, Heward

, Carr-Smith

, Foxall

, Franklyn

, Gough

. 2006. Contribution of single nucleotide polymorphisms within FCRL3 and MAP3K7IP2 to the pathogenesis of Graves' disease. J Clin Endocrinol Metab, 91:1056–1061.

22.

Mathews

. 1984. Statistical aspects of immunogenetic association with disease. Simons

, Tait

. Detection of Immune-Associated Genetic Markers of Human Disease. Churchill Livingstone: London, 106–136.

23.

Risch

, Merikangas

. 1996. The future of genetic studies of complex human diseases. Science, 273:1516–1517.

24.

Wigginton

, Cutler

, Abecasis

. 2005. A note on exact tests of Hardy-Weinberg equilibrium. Am J Hum Genet, 76:887–893.

25.

Areschoug

, Gordon

. 2009. Scavenger receptors: role in innate immunity and microbial pathogenesis. Cell Microbiol, 11:1160–1169.

26.

Arredouani

, Yang

, Ning

, Qin

, Soininen

, Tryggvason

, Kobzik

. 2004. The scavenger receptor MARCO is required for lung defense against pneumococcal pneumonia and inhaled particles. J Exp Med, 200:267–272.

27.

Bin

, Nielson

, Liu

, Mason

, Shu

. 2003. Identification of uteroglobin-related protein 1 and macrophage scavenger receptor with collagenous structure as a lung-specific ligand-receptor pair. J Immunol, 171:924–930.

28.

Tomer

, Davies

. 1995. Infections and autoimmune endocrine disease. Bailliere's Clin Endocrinol Metab, 9:47–70.

29.

Shimura

, Okajima

, Ikuyama

, Shimura

, Kimura

, Saji

, Kohn

. 1994. Thyroid-specific expression and cyclic adenosine 3',5'-monophosphate autoregulation of the thyrotropin receptor gene involves thyroid transcription factor-1. Mol Endocrinol (Baltim), 8:1049–1069.

30.

Kikkawa

, Gonzalez

, Kimura

. 1990. Characterization of a thyroid-specific enhancer located 5.5 kilobase pairs upstream of the human thyroid peroxidase gene. Mol Cell Biol, 10:6216–6224.

31.

Francis-Lang

, Price

, Polycarpou-Schwarz

, Di Lauro

. 1992. Cell-type-specific expression of the rat thyroperoxidase promoter indicates common mechanisms for thyroid-specific gene expression. Mol Cell Biol, 12:576–588.

32.

Civitareale

, Lonigro

, Sinclair

, Di Lauro

. 1989. A thyroid-specific nuclear protein essential for tissue-specific expression of the thyroglobulin promoter. EMBO J, 8:2537–2542.

33.

Lohmueller

, Pearce

, Pike

, Lander

, Hirschhorn

. 2003. Meta-analysis of genetic association studies supports a contribution of common variants to susceptibility to common disease. Nat Genet, 33:177–182.

34.

Hiratani

, Bowden

, Ikegami

, Shirasawa

, Shimizu

, Iwatani

, Akamizu

. 2005. Multiple SNPs in intron 7 of thyrotropin receptor are associated with Graves' disease. J Clin Endocrinol Metab, 90:2898–2903.

35.

Brand

, Barrett

, Simmonds

, Newby

, McCabe

, Bruce

, Kysela

, Carr-Smith

, Brix

, Hunt

, Wiersinga

, Hegedus

, Connell

, Wass

, Franklyn

, Weetman

, Heward

, Gough

SCL

. 2009. Association of the thyroid stimulating hormone receptor gene (TSHR) with Graves' disease (GD) Hum Mol Genet, 18:1704–1713.