Identification of Novel Genetic Loci Associated with Thyroid Peroxidase Antibodies and Clinical Thyroid Disease

Autoimmune thyroid diseases (AITD) are common, affecting 2-5% of the general population. Individuals with positive thyroid peroxidase antibodies (TPOAbs) have an increased risk of autoimmune hypothyroidism (Hashimoto's thyroiditis), as well as autoimmune hyperthyroidism (Graves' disease). As the possible causative genes of TPOAbs and AITD remain largely unknown, we performed GWAS meta-analyses in 18,297 individuals for TPOAb-positivity (1769 TPOAb-positives and 16,528 TPOAb-negatives) and in 12,353 individuals for TPOAb serum levels, with replication in 8,990 individuals. Significant associations (P<5×10−8) were detected at TPO-rs11675434, ATXN2-rs653178, and BACH2-rs10944479 for TPOAb-positivity, and at TPO-rs11675434, MAGI3-rs1230666, and KALRN-rs2010099 for TPOAb levels. Individual and combined effects (genetic risk scores) of these variants on (subclinical) hypo- and hyperthyroidism, goiter and thyroid cancer were studied. Individuals with a high genetic risk score had, besides an increased risk of TPOAb-positivity (OR: 2.18, 95% CI 1.68–2.81, P = 8.1×10−8), a higher risk of increased thyroid-stimulating hormone levels (OR: 1.51, 95% CI 1.26–1.82, P = 2.9×10−6), as well as a decreased risk of goiter (OR: 0.77, 95% CI 0.66–0.89, P = 6.5×10−4). The MAGI3 and BACH2 variants were associated with an increased risk of hyperthyroidism, which was replicated in an independent cohort of patients with Graves' disease (OR: 1.37, 95% CI 1.22–1.54, P = 1.2×10−7 and OR: 1.25, 95% CI 1.12–1.39, P = 6.2×10−5). The MAGI3 variant was also associated with an increased risk of hypothyroidism (OR: 1.57, 95% CI 1.18–2.10, P = 1.9×10−3). This first GWAS meta-analysis for TPOAbs identified five newly associated loci, three of which were also associated with clinical thyroid disease. With these markers we identified a large subgroup in the general population with a substantially increased risk of TPOAbs. The results provide insight into why individuals with thyroid autoimmunity do or do not eventually develop thyroid disease, and these markers may therefore predict which TPOAb-positives are particularly at risk of developing clinical thyroid dysfunction.


Introduction
Autoimmune thyroid disease (AITD), including Hashimoto's thyroiditis and Graves' disease, is one of the most common autoimmune diseases, affecting 2-5% of the general population [1,2,3].Thyroid dysfunction has been associated with osteoporosis, depression, atrial fibrillation, heart failure, metabolic syndrome, and mortality [4,5,6,7,8,9,10,11].High serum antibodies against the enzyme thyroid peroxidase (TPO), which is located in the thyroid and plays a key role in thyroid hormone synthesis, are present in 90% of patients with Hashimoto's thyroiditis [12,13], the most frequent cause of hypothyroidism and goiter.Although TPO antibodies (TPOAbs) are a useful clinical marker for the detection of early AITD, it remains controversial if these antibodies play a causative role in the pathogenesis of Hashimoto's thyroiditis [14,15,16].
Interestingly, TPOAb-positive persons also have an increased risk of developing autoimmune hyperthyroidism (Graves' disease) [17,18], which is caused by stimulating antibodies against the thyroid stimulating hormone (TSH) receptor [19].Numerous studies have shown that Graves' hyperthyroidism and Hashimoto's thyroiditis show co-inheritance [17,20,21].Finally, thyroid autoimmunity is the most common autoimmune disorder in women of childbearing age, and TPOAb-positive women have an increased risk of developing pregnancy complications such as miscarriage and pre-term delivery [17,18,22,23,24,25,26].
The prevalence of TPOAb-positivity in the general population ranges from 5-24%, but it is currently unknown why these people develop TPOAbs, nor is it known why not all individuals with thyroid autoimmunity develop clinical thyroid disease [27,28].It is estimated that around 70% of the susceptibility to develop thyroid autoantibodies is due to genetic factors [29].In this context it is remarkable to note that little is known about the genetic factors that determine TPOAb-positivity and the risk of AITD.
We therefore performed a genome wide association study (GWAS) meta-analysis for TPOAbs in the general population in 18,297 individuals from 11 populations.Newly identified genetic variants were studied in relation to subclinical and overt hypo-and hyperthyroidism, goiter, thyroid autoimmunity during pregnancy and thyroid cancer risk.

Results
Characteristics of the studied populations are shown in Table 1 and the Supplementary Material S1.Heritability estimates in the family-based cohorts SardiNIA, TwinsUK and Val Borbera were, respectively, 0.65, 0.66, and 0.54 for TPOAb-positivity, and 0.43, 0.66, and 0.30 for TPOAb levels.

Loci associated with TPOAb-positivity and TPOAb levels
See Table 1 and Supplementary Figure S1 for TPOAb measurements and Supplementary Table S1 for genotyping procedures.In most autoimmune diseases, both the presence and the level of autoantibodies are relevant for the disease onset [18,30,31].Furthermore, different pathophysiological processes may be involved in the initiation and severity of the autoimmune response.We therefore performed a GWAS on TPOAb-positivity (including 1769 TPOAb-positives and 16,528 TPOAb-negatives), as well as a GWAS on continuous TPOAb levels (including 12,353 individuals) in stage 1.See Supplementary Figures S2 and S3 for QQ (quantile-quantile) and Manhattan plots.
Table S4 shows the stage 1 TPOAb-positivity and TPOAb level meta-analyses results for GWAS significant SNPs reported in previous GWAS on thyroid related phenotypes.

Associations with hypo-and hyperthyroidism
The associations between the 5 GWAS significant SNPs and the risk of abnormal thyroid function tests are shown in Table 4. MAGI3-rs1230666 was associated with an increased risk of overt hypothyroidism and increased TSH levels below the Bonferroni threshold (i.e., P = 0.05/5 = 0.01).Borderline significant signals were observed at BACH2-rs10944479 with a higher risk of increased TSH levels as well as overt hyperthyroidism (P = 0.011 and P = 0.012), and at the KALRN-rs2010099 SNP with a lower risk of decreased TSH levels (P = 0.010).
Furthermore, a higher genetic risk score was associated with a higher risk of increased TSH levels (Supplementary Table S5).No effects of the genetic risk score on the risk of overt hypothyroidism, hyperthyroidism or decreased TSH levels were observed.

Associations with goiter
Individuals with a high genetic risk score had a 30.4% risk of sonographically-proven goiter, compared to 35.2% in subjects with a low score (P = 6.5610 24 ) (Table 5).None of the individual SNPs was significantly associated with goiter risk.

Thyroid autoimmunity during pregnancy
As autoimmunity significantly changes during pregnancy [25], we additionally studied these effects in an independent pregnant population.Pregnant women with a high genetic risk score had a 2.4 times increased risk of TPOAb-positivity compared to women with a low score (10.3% vs 4.8%, P = 0.03).These women did not have a higher risk of increased TSH levels.However, a borderline significant signal with a lower risk of increased TSH levels was observed at ATXN2-rs653178 (OR, 0.54 [95% CI, 0.34-0.87],P = 0.012).
b) Thyroid cancer.Supplementary Table S6 shows the associations of the 5 GWAS significant SNPs with thyroid cancer.No statistically significant associations were detected, but a borderline significant signal with an increased risk of thyroid cancer was observed at ATXN2-rs653178 (OR, 1.32 [95% CI, 1.02-1.70],P = 0.03).

Pathway analyses
Ingenuity Pathway Analyses (IPA; Ingenuity Systems, Ca, USA) and GRAIL analyses [32] were performed to identify potential pathways involved in AITD, the results of which are shown in Supplementary Tables S7 and S8, and Figure S6.The identified top pathways involved cell death, survival, movement, and OX40 signalling.

Discussion
This is the first GWAS meta-analysis investigating the genetics of TPOAbs in the normal population in up to 18,297 individuals from 11 populations with replication in up to 8,990 individuals from 5 populations.We identified 5 GWAS significant loci associated with TPOAb-positivity and/or levels.
The most significant hit for both TPOAb-positivity and TPOAb levels was located near the TPO gene itself.TPO is a membranebound protein located on the apical membranes of the thyroid follicular cell, catalyzing key reactions in thyroid hormone synthesis [33].Mutations in TPO have been found in patients with congenital hypothyroidism [34,35].Although TPOAbs are

Author Summary
Individuals with thyroid peroxidase antibodies (TPOAbs) have an increased risk of autoimmune thyroid diseases (AITD), which are common in the general population and associated with increased cardiovascular, metabolic and psychiatric morbidity and mortality.As the causative genes of TPOAbs and AITD remain largely unknown, we performed a genome-wide scan for TPOAbs in 18,297 individuals, with replication in 8,990 individuals.Significant associations were detected with variants at TPO, ATXN2, BACH2, MAGI3, and KALRN.Individuals carrying multiple risk variants also had a higher risk of increased thyroidstimulating hormone levels (including subclinical and overt hypothyroidism), and a decreased risk of goiter.The MAGI3 and BACH2 variants were associated with an increased risk of hyperthyroidism, and the MAGI3 variant was also associated with an increased risk of hypothyroidism.This first genome-wide scan for TPOAbs identified five newly associated loci, three of which were also associated with clinical thyroid disease.With these markers we identified a large subgroup in the general population with a substantially increased risk of TPOAbs.These results provide insight into why individuals with thyroid autoimmunity do or do not eventually develop thyroid disease, and these markers may therefore predict which individuals are particularly at risk of developing clinical thyroid dysfunction.valid clinical biomarkers of AITD, they are generally considered to be secondary to the thyroid damage inflicted by T-cells.The FOXE1 gene has been previously associated with hypothyroidism [36,37] and is known to regulate transcription of TPO [38].In this context it is interesting to note that we did not find any associations of the variant near TPO with hypothyroidism.Most genes that have been associated with AITD (predominantly Graves' disease) by candidate gene and GWAS studies so far are located in the HLA class I and II regions, or in genes involved in T-cell (i.e., CTLA-4, PTPN22) or other autoimmune responses [28,39].Until now, the TPO gene itself had not been associated with AITD, except in one recent candidate gene analysis in a small cohort (n = 188) without replication [40].A variant near TPO (rs11694732), which is in LD with rs11675434 (r2 = 0.97 in HapMap2), has previously been associated with TSH levels by Gudmundsson et al [41].However, various other GWAS on serum TSH and FT4 levels have not found any significant associations in or near this locus, including a recent similar sized GWAS by Porcu et al [42].
Three of the other four loci identified here are located in or are in linkage disequilibrium (LD) with genes previously associated with other autoimmune diseases.Rs1230666 is located in intron 9 of MAGI3, encoding a protein that modulates activity of AKT/PKB.AKT/PKB is expressed in the thyroid and regulates apoptosis [43], which seems to play an important role in the development of AITD [44,45].In addition, rs1230666 is in LD with rs2476601 (r2 = 0.70 in HapMap2), a variant causing a R620W substitution in PTPN22.PTPN22 is a lymphoid-specific intracellular phosphatase involved in the T-cell receptor signaling pathway.Variations in PTPN22, and specifically R620W, are associated with various autoimmune disorders including type 1 diabetes, rheumatoid arthritis, systemic lupus erythematosus and Graves' disease [46,47,48,49].The associations of the MAGI3 locus with TPOAb-positivity and Graves' disease may therefore also be explained by linkage with disease-associated variants in PTPN22 [50].Of note, the association signal at rs2476601 is one order weaker than that of the top variant rs1230666.
The BACH2 locus has been implicated in the susceptibility to several autoimmune diseases, including celiac disease, type 1 diabetes, vitiligo, Crohn's disease, and multiple sclerosis [46,51,52,53,54].A recent candidate gene analysis associated the BACH2 locus with an increased risk of AITD, including Hashimoto's thyroiditis and Graves' disease [55].However, the associations were not significant when Hashimoto's thyroiditis and Graves' disease were studied separately.BACH2 is specifically expressed in early stages of B-cell differentiation and represses different immunoglobulin genes [56].Interestingly, BACH2 can bind to the co-repressor SMRT (silencing mediator of retinoid and thyroid receptor), which may suggest a more direct effect on thyroid hormone secretion and action as well.
Polymorphisms in ATXN2 have been associated with multiple neurodegenerative diseases, including spinocerebellar ataxia and Parkinson's disease [57,58,59].Different epidemiological studies have associated thyroid dysfunction with cerebellar ataxia [60,61].Furthermore, the identified SNP in ATXN2 has been previously associated with renal function, serum urate levels and blood pressure [62,63,64].However, this SNP is in high LD with rs3184504 (r2 = 0.873), a variant causing a Trp262Arg substitution of SH2B adaptor protein 3 (SH2B3).SH2B3 encodes the adaptor protein LNK, a key negative regulator of cytokine signaling playing a critical role in hematopoiesis.This variant is associated with susceptibility to several autoimmune diseases, including celiac disease, type 1 diabetes, vitiligo, and rheumatoid arthritis [46,51,53,65], suggesting more relevance for TPOAb levels than ATXN2.This is supported by a recent study which showed that variants in LD with SH2B3, BACH2, and PTPN22 are associated with TPOAb levels in patients with type 1 diabetes [66].
Whereas the above four loci are located in genes involved in the immune response or the autoantigen, the KALRN (Kalirin) gene encodes a multi-domain guanine nucleotide exchange factor for GTP-binding proteins of the Rho family.The relation of KALRN with levels of TPOAbs is unclear.This gene has recently been found to be associated with megakaryopoiesis and platelet formation [67], which may suggest a function in the immune system [68].We furthermore performed pathway analyses on the stage 1 TPOAb-positivity and TPOAb level lead SNPs, and identified the cell death, survival and movement pathway as an important pathway for TPOAbs.This finding is supported by previous studies, which show an important role for apoptosis in the development of AITD [44,45].Another top pathway involved was the OX40 signalling pathway, and it is of interest to note that OX40 is a T-cell activator promoting the survival of CD4+ T-cells at sites of inflammation [69].
Our results have potential clinical relevance for several reasons.Genetic risk scores based on these novel common (risk allele frequencies: 9-40%) TPOAb-associated SNPs enabled us to identify a large subgroup in the general population with a two-fold increased risk of TPOAb-positivity (10.4% vs 5.4%).These individuals also have a higher risk of increased TSH levels and a lower risk of goiter, suggesting an advanced stage of destruction of the thyroid due to autoimmune processes.Furthermore, pregnant women with high genetic risk scores had a 2.4 times increased risk of TPOAb-positivity during pregnancy.In this context it is interesting to note that TPOAb-positive pregnant women have an increased risk of miscarriages and preterm births independent of thyroid function [70].
Associations with thyroid disease were also found on an individual SNP level.The MAGI3 SNP was associated with a substantially increased risk of hypothyroidism, and the BACH2 SNP showed a borderline significant association (P = 0.011) with a higher risk of increased TSH levels, which includes subjects with subclinical and overt hypothyroidism.Furthermore, both loci were significantly associated with an increased risk of Graves' hyperthyroidism in an independent population.To predict which patients with first or second degree relatives with documented Hashimoto's or Graves' disease will develop clinical thyroid disease, a clinical algorithm has been developed (i.e., the THEA score) [18].Future studies should analyze if these genetic markers increase the sensitivity of the THEA score.Graves' hyperthyroidism and Hashimoto's thyroiditis co-segregate in families and subjects with TPOAbs have an increased risk of both diseases [17,18,20,21,22,26].The current study provides insight into this phenomenon by showing that specific loci associated with TPOAbs and (subclinical) hypothyroidism, i.e.MAGI3 and BACH2, are also associated with Graves' hyperthyroidism in an independent case-control study.
The prevalence of TPOAb-positivity in the general population is high (5-24%), but it is currently unknown why part of the individuals with thyroid autoimmunity develop clinical thyroid disease whereas others do not [27,28].In this context it is interesting to note that the TPOAb-associated SNPs located in TPO and ATXN2 were not associated with clinical thyroid disease.This suggests that the TPOAbs in these individuals may be of less clinical relevance, providing insight into why TPOAb-positive individuals do or do not eventually develop clinical thyroid disease.
Our study has some limitations.The validity of the results is restricted to individuals from populations of European ancestry.Future GWASs in populations from non-European descent will be required to determine to which extent our results can be generalized to other ethnic groups.Secondly, we did not perform conditional analyses to further identify secondary association signals within the identified loci, nor did we perform functional studies for the identified variants.Further research is therefore needed to unravel the exact biological mechanism behind the observed associations.The fact that various TPOAb assays were used across the participating cohorts could lead to bias.We Table 4. Newly identified TPOAb associated loci and the risk of thyroid disease in stage 1 and 2 populations.therefore used TPOAb-positivity cut-off values as provided by the respective assay manufacturer, instead of using one fixed cut-off value.This is also of clinical importance as in clinical practice most institutions rely on the TPOAb-positivity cut-off as provided by the assay manufacturer.Furthermore, we did not detect heterogeneity in our results, supporting the fact that results obtained with different assays can be combined across cohorts using the z-score based meta-analysis.Finally, as AITD coincides with other autoimmune diseases, our results could be driven by indirect associations with other autoimmune diseases.However, AITD is the most common autoimmune disease in the general population.We furthermore show that carriage of multiple risk alleles is associated with an increased risk of thyroid dysfunction, which underlines the clinical importance of our findings.
In conclusion, this first GWAS for TPOAbs identified five newly associated loci, three of which were also associated with clinical thyroid disease.Furthermore, we show that carriage of multiple risk variants is not only associated with a substantial increased risk of TPOAb-positivity, but also with a higher risk of increased TSH levels (including subclinical and overt hypothyroidism) and a lower risk of goiter.These genetic markers not only help to identify large groups in the general population with an increased risk of TPOAbpositivity, but may also predict which TPOAb-positive persons are particularly at risk of developing clinical thyroid disease.

Study cohorts
For the TPOAb GWAS stage 1 and 2 analyses, and the hypothyroidism, hyperthyroidism and goiter analyses, individuals were recruited from 16 independent community-based and family studies.For the Graves' disease analyses, cases were recruited from the United Kingdom Graves' disease cohort and controls from the British 1958 Birth Cohort.Thyroid cancer cases and controls were recruited from the Nijmegen and Ohio thyroid cancer cohorts.A detailed description of the original cohorts contributing samples is provided in Table 1 and in the Supplementary Material.All participants provided written informed consent and protocols were approved by the institutional review boards or research ethics committees at the respective institutions, and conducted according to the Declaration of Helsinki.

Phenotype definitions
Serum TPOAb levels were determined with a range of assays.TPOAb-positives were defined as subjects with TPOAb levels above the assay-specific TPOAb-positivity cut-off, as defined by the manufacturer (Table 1).Serum TSH and free thyroxine (FT4) levels were determined using a range of assays (Table 1).Assayspecific TSH and FT4 reference ranges were used, as provided by the manufacturer (Table 1).Overt hypothyroidism was defined as a high TSH (i.e., a TSH level above the TSH reference range) and a low FT4.Increased TSH was defined as a high TSH, including persons with overt hypothyroidism or subclinical hypothyroidism (i.e., high TSH with a normal FT4).Overt hyperthyroidism was defined as a low TSH and a high FT4.Decreased TSH was defined as a low TSH, including persons with subclinical or overt hyperthyroidism.
The diagnosis of goiter is described in the Supplementary Material, and the diagnosis of Graves' disease and thyroid cancer in the respective cohorts have been described previously [41].

Genotyping
Samples were genotyped with a range of GWAS genotyping arrays (Supplementary Table S1).Sample and SNP quality control procedures were undertaken within each study.For each GWAS, over 2.5 million SNPs were imputed using CEU samples from Phase 2 of the International HapMap project (www.hapmap.org).Genotyping procedures in the stage 2, Graves' disease and thyroid cancer populations are described in the Supplementary Material.

Association analyses
The heritabilities of TPOAb-positivity and serum TPOAb levels were estimated, as described in the Supplementary Material.
In stage 1, we performed a GWAS on TPOAb-positivity as well as a GWAS on continuous TPOAb levels.Persons taking thyroid medication were excluded.Each SNP was tested for association with TPOAb-positivity using logistic regression analyses, adjusting for age and sex.For cohorts with family structure, we approximated the probability of being affected with a linear mixed model adjusting for age and sex.The produced model was used to predict the expected proportion of ''risk'' (effective) alleles in cases and controls, hence giving the means to estimate odds ratios.Only unrelated individuals were considered for the SardiNIA cohort.For the GWAS of continuous TPOAb levels, samples with a TPOAb level lower than the minimum TPOAb assay detection limit (Table 1) were excluded.TPOAb levels were natural log-transformed, and sex-specific, age adjusted standardized residuals were calculated.Each SNP was tested for association with these TPOAb level residuals using linear regression analyses (additive model),

Funding:
The Asklepios Study was supported by a Fonds voor Wetenschappelijk Onderzoek-Vlaanderen FWO research grant G.0427.03 and G.0838.10N (Asklepios Study).The 1994-5 Busselton Health Survey was funded by Healthway, Western Australia.The Busselton Health Studies are supported by the National Health and Medical Research Council of Australia and the Great Wine Estates Auctions.The CHS research reported in this article was supported by NHLBI contracts HHSN268201200036C, N01HC85239, N01HC55222, N01HC85079, N01HC85080, N01HC85081, N01HC85082, N01HC85083, N01HC85086; and NHLBI grants HL080295, HL087652, HL105756 with additional contribution from the National Institute of Neurological Disorders and Stroke (NINDS).Additional support was provided through AG023629 from the National Institute on Aging (NIA).DNA handling and genotyping at Cedars-Sinai Medical Center was supported in part by the National Center for Research Resources, grant UL1RR033176, and is now at the National Center for Advancing Translational Sciences, CTSI grant UL1TR000124; in addition to the National Institute of Diabetes and Digestive and Kidney Disease grant DK063491 to the Southern California Diabetes Endocrinology Research Center.Additional funding was provided by the Cedars-Sinai Board of Governors' Chair in Medical Genetics (JIR).The CARLA Study was founded by a grant from the Deutsche Forschungsgemeinschaft as part of the Collaborative Research Center 598 ''Heart failure in the elderly -cellular mechanisms and therapy'' at the Medical Faculty of the Martin-Luther-University Halle-Wittenberg, by a grant of the Wilhelm-Roux Programme of the Martin-Luther-University Halle-Wittenberg; by the Ministry of Education and Cultural Affairs of Saxony-Anhalt, and by the Federal Employment Office.The Exeter Family Study of Childhood Health (EFSOCH) was supported by South West NHS Research and Development, Exeter NHS Research and Development, the Darlington Trust, and the Peninsula NIHR Clinical Research Facility at the University of Exeter.Genotyping of EFSOCH DNA samples was supported by the Endocrine Research Fund.ATH and BMS are employed as core members of the Peninsula NIHR Clinical Research Facility.RMF is funded by a Sir Henry Wellcome Postdoctoral Fellowship (Wellcome Trust grant: 085541/Z/ 08/Z).The Health2006 Study is funded by grants from The Velux Foundation; The Danish Medical Research Council, Danish Agency for Science, Technology and Innovation; The Aase and Ejner Danielsens Foundation; ALK-Abello ´A/S (Hørsholm, Denmark), Timber Merchant Vilhelm Bangs Foundation, MEKOS Laboratories (Denmark), the Health Insurance Foundation, and Research Centre for Prevention and Health, the Capital Region of Denmark.Helsinki Birth Cohort Study has been supported by grants from the Academy of Finland, the Finnish Diabetes Research Society, Finnish Society for Cardiovascular Research, Folkha ¨lsan Research Foundation, Novo Nordisk Foundation, Finska La ¨karesa ¨llskapet, Signe and Ane Gyllenberg Foundation, University of Helsinki, European Science Foundation (EUROSTRESS), Ministry of Education, Ahokas Foundation, Emil Aaltonen Foundation, Juho Vainio Foundation, and Wellcome Trust (grant number WT089062).This work was supported by KORA, which is a research platform initiated and financed by the Helmholtz Center Munich, German Research Center for Environmental Health, by the German Federal Ministry of Education and Research and by the State of Bavaria.The work of KORA is supported by the German Federal Ministry of Education and Research (BMBF) in the context of the German National Genome Research Network (NGFN-2 and NGFN-plus).The present research was supported within the Munich Center of Health Sciences (MC Health) as part of LMUinnovativ.Thyroid examinations in KORA-F4 were supported by Sanofi-Aventis in the framework of the Papillon Initiative.Collection and genotyping of the NBS samples was funded in part by the European Commission (POLYGENE: LSHC-CT-2005-018827) and a research investment grant of the Radboud University Nijmegen Medical Centre.This work was sponsored by the National Computing Facilities Foundation (NCF) for the use of supercomputer facilities, with financial support from the NWO.The Thyroid Cancer Program (P.I.Matthew Ringel) at the Ohio State University is supported by grants P30 CA16058 and P01 CA124570 from the National Cancer Institute, USA.The generation and management of GWAS genotype data for the Rotterdam Study is supported by the Netherlands Organisation of Scientific Research NWO Investments (no.175.010.2005.011,911-03-012).This study is funded by the Research Institute for Diseases in the Elderly (014-93-015; RIDE2), the Netherlands Genomics Initiative (NGI)/Netherlands Organisation for Scientific Research (NWO) project no.050-060-810.The Rotterdam Study is funded by Erasmus Medical Center and Erasmus University, Rotterdam, Netherlands Organization for the Health Research and Development (ZonMw), the Research Institute for Diseases in the Elderly (RIDE), the Ministry of Education, Culture and Science, the Ministry for Health, Welfare and Sports, the European Commission (DG XII), and the Municipality of Rotterdam.The SardiNIA study is supported by the Intramural Research Program of the National Institute on Aging (NIA), National Institutes of Health (NIH).The SardiNIA (''Progenia'') team was supported by Contract NO1-AG-1-2109 from the NIA; the efforts of SS were supported in part by contract 263-MA-410953 from the NIA to the University of Michigan and by research grant HG002651.SHIP is part of the Community Medicine Research net of the University of Greifswald, Germany, which is funded by the Federal Ministry of Education and Research (grants no.01ZZ9603, 01ZZ0103, and 01ZZ0403), the Ministry of Cultural Affairs as well as the Social Ministry of the Federal State of Mecklenburg-West Pomerania, and the network 'Greifswald Approach to Individualized Medicine (GANI_MED)' funded by the Federal Ministry of Education and Research (grant 03IS2061A).Genome-wide data have been supported by the Federal Ministry of Education and Research (grant no.03ZIK012) and a joint grant from Siemens Healthcare, Erlangen, Germany and the Federal State of Mecklenburg-West Pomerania.Data analyses were further supported by the German Research Foundation (DFG Vo 955/10-2; SPP 1629: THYROID TRANS ACT WA 1328/5-1) and the Federal Ministry of Nutrition, Agriculture and Consumer's Safety (BMELV 07 HS 003).SHIP-Trend is part of the Community Medicine Research net of the University of Greifswald, Germany, which is funded by the Federal Ministry of Education and Research (grants no.01ZZ9603, 01ZZ0103, and 01ZZ0403), the Ministry of Cultural Affairs as well as the Social Ministry of the Federal State of Mecklenburg -West Pomerania.Thyroid-related examinations have been funded by the Federal Ministry of Nutrition, Agriculture and Consumer's Safety (BMELV 07 HS 003) and the German Research Foundation (DFG Vo 955/10-1; SPP 1629: THYROID TRANS ACT WA 1328/5-1

Figure 1 .
Figure 1.Genome wide association studies meta-analyses: Loci associated with TPOAb-positivity (a-c) and TPOAb levels (d-f) on a genome-wide level of significance.Regional association plots of the genome-wide significant loci associated with TPOAb positivity (a-c) and TPOAb levels (d-f).The y-axis on the left indicates the -log 10 P value for the association with TPOAb -positivity (a-c) or TPOAb levels (d-f).SNPs are plotted on the x-axis according to their chromosomal position against the association with the phenotype on the y-axis.The most significant stage 1 SNP is indicated in purple.The combined stage 1 and 2 result of this SNP is indicated in yellow.The SNPs surrounding the most significant SNP are color-coded to reflect their LD with this SNP.Symbols reflect functional genomic annotation, as indicated in the legend.The blue y-axes on the right of each plot indicate the estimated recombination rates (based on HapMap Phase II); the bottom of each panel shows the respective annotated genes at the locus and their transcriptional direction.Mb, megabases.doi:10.1371/journal.pgen.1004123.g001 ). Genome-wide data have been supported by the Federal Ministry of Education and Research (grant no.03ZIK012) and a joint grant from Siemens Healthcare, Erlangen, Germany and the Federal State of Mecklenburg, West Pomerania.Whole-body MR imaging was supported by a joint grant from Siemens Healthcare, Erlangen, Germany and the Federal State of Mecklenburg West Pomerania.TwinsUK received funding from the Wellcome Trust; the Chronic Disease Research Foundation; the European Community's Seventh Framework Program grant agreement (FP7/2007-2013); ENGAGE project grant agreement (HEALTH-F4-2007-201413); the Department of Health via the National Institute for Health Research (NIHR) Comprehensive Biomedical Research Centre award to Guy's & St Thomas' NHS Foundation Trust in partnership with King's College London; the Canadian Institutes of Health Research, Canadian Foundation for Innovation, Fonds de la Recherche en Sante ´Que ´bec, Ministe `re du De ´veloppement E ´conomique, de l9Innovation et de l9Exportation Que ´bec and the Lady Davis Institute of the Jewish General Hospital; the Australian National Health and Medical Research Council (Project Grants 1010494, 1031422) and the Sir Charles Gairdner Hospital Research Fund.Val Borbera was supported by funds from Compagnia di San Paolo, Torino, Italy; Fondazione Cariplo, Italy and Ministry of Health, Ricerca Finalizzata 2008.The UK Graves' disease cohort was funded by the Wellcome Trust grant 068181.The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Table 1 .
Population characteristics and serum TPOAb, TSH, and FT4 level measurements specifications.

Table 2 .
Newly identified loci associated with TPOAb-positivity and/or serum TPOAb levels reaching genome wide significance.
a Risk allele frequency: Weighted mean frequency of the risk allele across all included cohorts.b Adjusted for age and gender c Expressed in sd of natural logarithm transformed serum TPOAb level, adjusted for age and gender doi:10.1371/journal.pgen.1004123.t002

Table 3 .
Genetic risk score and the risk of TPOAb-positivity.

Table 5 .
Newly identified TPOAb associated loci, genetic risk scores and the risk of goiter.Adjusted for age, gender, and body surface area.ATXN2-rs653178 is in high LD with SH2B3-rs3184504.MAGI3-rs1230666 is in high LD with PTPN22-rs24756601.doi:10.1371/journal.pgen.1004123.t005Table S7 Top IPA associated networks for the Stage 1 TPOAbpositivity and TPOAb level lead SNPs.(DOCX) Table S8 Top IPA associated canonical pathways for the Stage 1 TPOAb-positivity and TPOAb level lead SNPs.(DOCX) a