June 2016
Volume 57, Issue 7
Open Access
Genetics  |   June 2016
Association of Polymorphisms in MACRO Domain Containing 2 With Thyroid-Associated Orbitopathy
Author Affiliations & Notes
  • Jwu Jin Khong
    Melbourne Clinical School-Western Campus Department of Medicine, University of Melbourne, Sunshine Hospital, St. Albans, Victoria, Australia
    Orbital, Plastics and Lacrimal Unit, The Royal Victorian Eye and Ear Hospital, Victoria, Australia
    Department of Surgery, University of Melbourne, Victoria, Australia
  • Kathryn P. Burdon
    Menzies Institute for Medical Research, University of Tasmania, Hobart, Tasmania, Australia
  • Yi Lu
    Statistical Genetics, QIMR Berghofer Medical Research Institute, Herston, Queensland, Australia
  • Lefta Leonardos
    Department of Ophthalmology, Flinders University of South Australia, Bedford Park, South Australia, Australia
  • Kate J. Laurie
    Department of Ophthalmology, Flinders University of South Australia, Bedford Park, South Australia, Australia
  • John P. Walsh
    Department of Endocrinology and Diabetes, Sir Charles Gairdner Hospital, Nedlands, Western Australia, Australia
    School of Medicine and Pharmacology, The University of Western Australia, Crawley, Western Australia, Australia
  • Adam D. Gajdatsy
    Centre for Ophthalmology and Visual Sciences, University of Western Australia, Western Australia, Australia
  • Peter R. Ebeling
    Department of Medicine, School of Clinical Sciences, Monash University, Clayton, Victoria, Australia
  • Alan A. McNab
    Orbital, Plastics and Lacrimal Unit, The Royal Victorian Eye and Ear Hospital, Victoria, Australia
    Centre for Eye Research Australia, University of Melbourne, East Melbourne, Victoria, Australia
  • Thomas G. Hardy
    Orbital, Plastics and Lacrimal Unit, The Royal Victorian Eye and Ear Hospital, Victoria, Australia
    Department of Surgery, University of Melbourne, Victoria, Australia
  • Richard J. Stawell
    Centre for Eye Research Australia, University of Melbourne, East Melbourne, Victoria, Australia
  • Garry J. Davis
    South Australian Institute of Ophthalmology, University of Adelaide, South Australia, Australia
  • Dinesh Selva
    South Australian Institute of Ophthalmology, University of Adelaide, South Australia, Australia
  • Angelo Tsirbas
    Australian School of Advanced Medicine, Macquarie University, Sydney, Australia
  • Grant W. Montgomery
    Molecular Epidemiology, QIMR Berghofer Medical Research Institute, Herston, Queensland, Australia
  • Stuart Macgregor
    Statistical Genetics, QIMR Berghofer Medical Research Institute, Herston, Queensland, Australia
  • Jamie E. Craig
    Department of Ophthalmology, Flinders University of South Australia, Bedford Park, South Australia, Australia
  • Correspondence: Jwu Jin Khong, Melbourne Clinical School-Western Campus, Department of Medicine, University of Melbourne, 176 Furlong Road, St. Albans, VIC 3021, Australia;jwujinkhong@gmail.com
Investigative Ophthalmology & Visual Science June 2016, Vol.57, 3129-3137. doi:10.1167/iovs.15-18797
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Jwu Jin Khong, Kathryn P. Burdon, Yi Lu, Lefta Leonardos, Kate J. Laurie, John P. Walsh, Adam D. Gajdatsy, Peter R. Ebeling, Alan A. McNab, Thomas G. Hardy, Richard J. Stawell, Garry J. Davis, Dinesh Selva, Angelo Tsirbas, Grant W. Montgomery, Stuart Macgregor, Jamie E. Craig; Association of Polymorphisms in MACRO Domain Containing 2 With Thyroid-Associated Orbitopathy. Invest. Ophthalmol. Vis. Sci. 2016;57(7):3129-3137. doi: 10.1167/iovs.15-18797.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: Thyroid-associated orbitopathy (TO) is an autoimmune-mediated orbital inflammation that can lead to disfigurement and blindness. Multiple genetic loci have been associated with Graves' disease, but the genetic basis for TO is largely unknown. This study aimed to identify loci associated with TO in individuals with Graves' disease, using a genome-wide association scan (GWAS) for the first time to our knowledge in TO.

Methods: Genome-wide association scan was performed on pooled DNA from an Australian Caucasian discovery cohort of 265 participants with Graves' disease and TO (cases) and 147 patients with Graves' disease without TO (controls). Top-ranked single nucleotide polymorphisms (SNPs) then were genotyped in individual DNA samples from the discovery cohort, and two replication cohorts totaling 584 cases and 367 controls.

Results: In the GWAS of pooled DNA samples, several SNPs showed suggestive association with TO at genome-wide P ≤ 10−6; rs953128 located on chr10q21.1, rs2867161 on chr7q11.22, rs13360861 on chr5q12.3, rs7636326 on chr3q26.2, rs10266576 on chr 7q11.22, rs60457622 on chr3q23, and rs6110809 on chr20p12.1. However, the only SNP consistently associated with TO on individual genotyping in the discovery and replication cohorts was rs6110809, located within MACROD2 on chromosome 20p12.1. On combined analysis of discovery and replication cohorts, the minor A allele of rs6110809 was more frequent in TO than in Graves' disease controls without TO (P = 4.35 × 10−5; odds ratio [OR] = 1.77; 95% confidence interval [CI], 1.35–2.32) after adjusting for age, sex, duration of Graves' disease, and smoking.

Conclusions: In patients with Graves' disease, a common genetic variant in MACROD2 may increase susceptibility for thyroid-associated orbitopathy. This association now requires confirmation in additional independent cohorts.

Thyroid-associated orbitopathy (TO) is an autoimmune-mediated orbital inflammation affecting 25% to 50% of patients with Graves' disease (GD).1 The overall age-adjusted incidence rate of TO is 16 cases per 100,000 per year for females and 2.9 cases per 100,000 population per year for males. The average age at diagnosis is 44.7 years.2 The onset of TO occurs mostly within 18 months of diagnosis of GD,3 and the close temporal relationship suggests the two diseases share a common autoantigen. 
Genetic susceptibility for GD is complex and involves multiple genetic loci regulating immune system and thyroid-specific genes. The HLA class II alleles, CTLA4, PTPN22, CD40 polymorphisms were identified early on by candidate gene association studies as genetic risk factors for GD, even though the overall contributions from each of these genetic variants are modest.47 The introduction of high throughput microarray technology, in combination with large cohorts of cases and controls, has seen genome-wide association studies (GWAS) revolutionize the study of the genetics of complex traits, and has allowed detection of association of common genetic variants with many diseases.8 The earliest GWAS studies of autoimmune thyroid disease reported associations with HLA class I and II regions, and TSHR and FCRL3 genes.9 Subsequent GWAS studies replicated association of FCRL3, HLA loci, TSHR, CTLA4, PTPN22, IL2RA, and discovered new susceptibility loci at SENP1, SLAMF6, RNASET2-FGFR1OP-CCR6, an intergenic region at 4p14, MMEL1, LPP, BACH2, PRICKLE1, and ITGAM.1013 
Although genetic risk factors for GD are becoming well defined, limited progress has been made with TO, largely due to small sample sizes limiting power to detect modest genetic effects and failure to replicate genetic findings. Candidate gene studies, predominantly of genes and polymorphisms associated with GD or genes involved in immune response and inflammation, have been investigated for a contribution to TO. While some studies showed CTLA4, HLA I, and II alleles, and IL-23R rs10889677 and rs2201841 SNPs were associated with TO compared to GD controls, other studies have not supported these associations.1421 TNF-α alleles were associated significantly with TO in a Japanese but not in a Polish study.22,23 The adipogenesis-related gene PPARγ Pro12Ala polymorphism decreased the risk of TO in one study but not in another, but both studies found the Pro12Ala variant possibly reduces the severity and activity of TO.24,25 Using tag single nucleotide polymorphisms (SNP), adipocyte-related immediate early genes showed increased risk for TO with genetic variants in CYR61 (rs1378227), ZFP36 (rs1057745, rs11083522), and SCD (rs1393491) with odds between 1.3 and 1.4 compared to GD, but the findings must be replicated.26 Similarly, using tag SNPs, SNPs in IL-1β did not differ between TO and non-TO in GD in one study but C allele of rs1143634 increased the risk in TO and GD compared to normal controls,27 whereas another identified rs1800587 polymorphism at IL-1α and rs16944 polymorphism at IL-1β significantly associated with GD and TO.28 PTPN12 (rs1468682, rs4729535, and rs17467232), functionally related to PTPN22, was associated with TO compared to non-TO in GD with odds at 1.4, and 2 of the SNPS (rs1468682 and rs4729535) showed interactions with TSHR rs2268458 SNP; the findings again require replication.29 However, TSHR rs2268458 SNPs, previously found to be associated with GD, were not associated with TO.18 
A pooled GWAS study is a cost-effective screening strategy for identifying gene variants in multiple diseases. Pooled DNA strategy has successfully identified known susceptibility loci CFH, ARMS2/HTRA1 in age-related macular degeneration, LOXL1 in pseudoexfoliation syndrome, and discovered new a susceptibility locus HGF in keratoconus.30,31 We aim to use a GWAS to discover genomic loci associated with TO. We report the first attempts to our knowledge at using a genome-wide association scan for discovering genetic variants in TO, and an association of variants in the MACROD2 gene suggesting it as a new TO susceptibility locus. 
Methods
The genetic study of TO was approved by human research ethics committees at multiple recruitment sites; the Royal Victorian Eye and Ear Hospital, Sunshine and Footscray Hospitals, the Royal Melbourne Hospital, and the Alfred Hospital in Melbourne, Victoria, Australia, and Flinders Medical Centre, and the Royal Adelaide Hospital in Adelaide, South Australia. The genetic research was conducted in accordance with the Declaration of Helsinki. 
Australian Thyroid-Associated Orbitopathy Research (ATOR) Cohorts
Definition of Graves' Disease and TO.
Graves' disease was defined by the presence of hyperthyroidism based on thyroid function test, including T3, T4, and TSH, and either elevated thyrotropin receptor (TSHR) antibodies or diffuse uptake on technetium-99m pertechnetate thyroid scan. Cases are defined by the presence of TO clinically and controls by the absence of TO. 
Thyroid-associated orbitopathy was defined by the presence of symptoms of TO and at least one sign of TO, for example, lid retraction. Early symptoms of TO were detected using the Vancouver Orbitopathy Rule symptom questionnaire, including red eyes, lids swelling, eye protrusion and stare, and blurred vision.32 Thyroid-associated orbitopathy status was examined and classified using the Vision, Inflammation, Strabismus, and Appearance (VISA) classification.33 Ophthalmologic measurements included the following: visual acuity, pupil response, color vision, inflammatory index score, extraocular movement, strabismus, lid measurements including palpebral aperture, marginal reflex distance, lid retraction, and Hertel exophthalmometry. The status of TO was determined by an ophthalmologist. 
ATOR Discovery Cohort.
Thyroid-associated orbitopathy cases (n = 265) and GD without TO controls (n = 147) of European descent were recruited from endocrine and eye hospital outpatient clinics and private practices of endocrinologists and ophthalmologists from Victoria and South Australia. Whole blood samples in EDTA were stored at 4°C. 
Replication Cohorts.
A further 319 TO cases and 220 GD controls were recruited from Victoria and Western Australia as replication cohorts. These comprised 231 TO cases and 142 GD controls recruited from Victoria using the same protocol as the ATOR discovery cohort, and 88 TO cases and 78 GD controls of European descent recruited from endocrine and ophthalmology clinics in Western Australia. Thyroid-associated orbitopathy was defined clinically by the presence of eye symptoms and signs with severity classified according to the European group on Graves' orbitopathy (EUGOGO) recommendations (by coauthor JW)34 or VISA classification (by coauthor AG).33 DNA samples (50 ng/μL) from venous blood were used for individual genotyping. 
A total of 282 normal controls, consisting of healthy hospital volunteers and retirement village residents over 50 years of European descent from South Australia, were used for additional individual genotyping for comparison with the TO and GD cohorts (with and without TO), respectively. The normal control cohort was not actively screened for GD, but no participant reported having this condition. 
DNA Pooling for GWAS Study in the Discovery Cohort.
Venous blood aliquots from the discovery cohort were stored at 4°C before construction of 2 comparative blood pools (pool number 1, 154 TO cases versus 102 controls; pool number 2, 116 TO cases versus 40 controls), totaling 270 cases and 142 controls. In between the construction of the blood pools and individual genotyping, the phenotypes of 5 individuals were updated (5 from cases to controls); these small numbers do not materially affect the overall blood pool results. Protease and lysis buffer were added to 100 μL of thawed whole blood from each sample, vortexed, and incubated, and then samples were combined in one tube for each pool. DNA was extracted from the pooled blood using the QIA-amp DNA blood maxi kit (Qiagen, Valencia, CA, USA) according to the manufacturer's protocol. For the remaining discovery samples, equimolar DNA pools were generated in duplicate after stepwise dilutions of genomic DNA extracted using QIA-amp DNA blood maxi kit to 75 ng/μL as previously described.30,31 The DNA was quantitated using Fluroskan Ascent (ThermoFisher Scientific, Waltham, MA, USA) together with double-stranded DNA quantitation assay using Picogreen reagents. (Invitrogen, Carlsbad, CA, USA). A total of 451 ng of DNA from each sample was added to the DNA pool. 
Pooled Genotyping and Analysis
Genome-wide genotyping was conducted by hybridization to HumanOmni5-Quad (Illumina, San Diego, CA, USA) beadchip microarray containing approximately 4.3 million markers, at QIMR Berghofer Medical Research Institute, according to standard protocol for case and control pools on two independent arrays for each pool. Two case pools and two control pools were generated and assigned into two case–control pool comparisons. Each pool was replicated 2 to 4 times to reduce measurement error. For the purpose of this analysis, we retained SNPs with minor allele frequency (MAF) over 1% from the reference panel of EUR samples in the 1000 Genome Project (excluded 1.8 M SNPs). Thus, the number of SNPs was reduced to 2.5 million. The output of the raw red and green bead scores from Illumina Beadstation was used for the pooled data analysis. The data cleaning and quality control procedures in array-based pooling have been described previously.3537 Briefly, SNPs with more than 10% negative scores on each array were excluded, as well as the SNPs with the sum of mean red and green scores lower than 1200 across each array to ensure calibration was performed on a precleaned dataset. A normalization/correction factor (corr) was calculated to make the mean value of the pooling allele frequency (PAF) 0.5 across all SNPs on each strands of the array. The PAF then was estimated based on the raw red intensities and the corrected green intensities for all the SNPs [PAF = red/(red + green/corr)].36 Further quality control included filtering nonautosomal SNPs and any SNPs with a significant variance difference between cases and controls. 
After data cleaning, 2.2 M SNPs were left for association testing in the ATOR discovery cohort. A linear model was used to test for allelic association in the two case–control pool comparisons. A test statistic, which corrects for pooling error, was used to rank SNPs, with P values based on a χ2 distribution with 1 degree of freedom. The application of the test statistic is based upon contrasting case and control pools, the effect of unequal amplification of alleles is minimal as such effects cancel out.35 Specifically, a final set of autosomal SNPs was determined meeting the following criteria: (1) have sufficient number of probes (i.e., over 1/2 of expected number of probes for MAF 1% to 5%, or over 1/3 of expected number of probes for SNPs >5%; and (2) have no large differences (>0.3) between PAF and reference allele frequency from EUR samples in the 1000 Genome Project. We then performed a meta-analysis weighted by inverse variance, combining results from two case–control pool comparisons. The top-ranked SNPs in the meta-analysis then were counter checked for proxies (linkage disequilibrium r2 > 0.5 with the index SNPs in the EUR populations). To be considered a valid result, if proxy SNPs were available, they also should show evidence of association. A genome-wide significance threshold of 5 × 10−8 was set to account for multiple testing. 
Validation Genotyping and Analysis
Top-ranked SNPs with P value threshold ≤ 1 × 10−6 indicative of association, or corresponding proxy if failed genotyping, were selected for individual genotyping in the discovery and replication cohorts, to validate pooled genotyping results as well as to replicate the associations in independent samples. In total, 7 SNPs were genotyped on Sequenom iPLEX Gold assay using Sequenom MassARRAY Analyzer in 2 batches, the ATOR discovery cohort and Western Australia replication cohort were genotyped at the Australian Genome Research Facility (Brisbane, Australia), and the ATOR replication cohort at GeneWorks Pty Ltd. (Adelaide, Australia). The allelic association analysis was performed in PLINK,38 using a χ2 test and performed separately for the ATOR pooled discovery cohort for validation, the replication cohorts, and for combined discovery and replication cohorts. The odds ratio (OR) and P values were adjusted for age of disease onset, duration of GD, and sex using logistic analysis. Single nucleotide polymorphisms with P ≤ 1 × 10−4 were considered validated in validation genotyping. In the replication cohort, P < 0.05 was considered associated. In the secondary analysis, we further adjusted the individual genotyping results for smoking status (smoker, ex-smoker, and never smoked) in addition to sex, age, and duration of GD for the discovery, replication, and combined cohorts, bearing in mind there are missing data points for smoking in 10.9% of the replication cohorts. Further individual genotyping for the top ranking SNPS were compared between TO and normal controls; and between GD and normal control to differentiate whether the validated SNPs were associated with TO and GD as a whole or the risk variant is specific for TO in GD. As smoking status data were not available from the healthy controls, the genotyping results for these genotyping analyses were adjusted for sex and age only. 
Following individual genotyping in discovery and replication cohorts revealing rs6110809 in MACROD2 as the locus of interest, further SNPs residing in MACROD2 with nominal P ≤ 5 × 10−4 were extracted from pooled GWAS top ranking data. They were assessed for linkage disequilibrium with rs6110809, to determine if multiple independent risk alleles were potentially present within MACROD2. Genotype data limited to a 100 kb region of MACROD2 capturing all 4 additional SNPs were downloaded from International Hapmap Project (available in the public domain at http://hapmap.ncbi.nlm.nih.gov/cgi-perl/gbrowse/hapmap24_B36/#search). Genotype data were analyzed in HaploView (Version 1.0, Broad Institute) for pairwise comparisons of marker SNPs with minor allele frequency > 0.01. 
Results
The age at disease onset, duration of GD, sex, and smoking status distributions of the three cohorts of participants are described in Table 1. Thyroid-associated orbitopathy cases were slightly older compared to GD without TO at diagnosis and the duration of GD in TO cases was longer. The proportions of males to females were similar between the groups and the proportion of smokers was greater in cases compared to controls for ATOR cohorts. The age and sex distribution of the normal controls in comparison with TO as a whole and GD as a whole are shown in Table 2
Table 1
 
Clinical Characteristics of GWAS Discovery and Replication Cohorts
Table 1
 
Clinical Characteristics of GWAS Discovery and Replication Cohorts
Table 2
 
Demographics of Overall Cohort With GD, TO Compared to Normal Healthy Controls
Table 2
 
Demographics of Overall Cohort With GD, TO Compared to Normal Healthy Controls
Pooled DNA was used to identify genetic variants associated with TO. None of the top ranked SNPs reached the genome-wide significance threshold of 5 × 10−8 (Supplementary Table S1), nevertheless multiple top ranked SNPs approached genome-wide significance with P ≤ 10−6 and in particular with rs2867161 on chr7q11.22 (P = 5.94 × 10−8) and rs60457622 on chr3q23 (P = 9.28 × 10−8) almost reaching significance (Table 3). 
Table 3
 
Top Ranking Pooled Genome-Wide Association Study and Validation Genotyping in ATOR Discovery Cohort
Table 3
 
Top Ranking Pooled Genome-Wide Association Study and Validation Genotyping in ATOR Discovery Cohort
The seven top ranked SNPs from pooled GWAS were individually genotyped in the ATOR discovery cohort. rs6110809 was validated with P < 1 × 10−4, the minor allele is associated with TO risk, showing higher frequency in cases compared to controls (Table 3). After further adjustment for smoking, rs6110809 was associated with TO with OR at 2.66 (95% confidence interval [CI], 1.69–4.18; P = 2.45 × 10−5) in the discovery cohort. 
All 7 SNPs also were assessed in the replication cohorts. The only SNP consistently associated with TO in replication cohorts was rs6110809, located within an intron of MACROD2 on chromosome 20p12.1 with an age, duration, sex-adjusted OR 1.42 (95% CI, 1.04–1.95; P = 0.03); after further adjustment for smoking, the OR was 1.35 (95% CI, 0.96–1.91; P = 0.08). Overall, for the combined discovery and replication cohorts, the MACROD2 rs6110809 was associated with TO compared to GD controls without TO (OR = 1.81; 95% CI, 1.40–2.34; P = 6.78 × 10−6; Table 4). After adjusting for age, duration of GD, sex, and smoking, the minor A allele was more frequent in TO than in GD without TO with OR 1.77 (95% CI, 1.34–2.32; P = 4.35 × 10−5). 
Table 4
 
Replication Individual Genotyping of Top Ranking SNPs From Discovery GWAS in Replication Cohorts and in Combined Discovery and Replication Cohorts
Table 4
 
Replication Individual Genotyping of Top Ranking SNPs From Discovery GWAS in Replication Cohorts and in Combined Discovery and Replication Cohorts
Additional individual genotyping analyses were undertaken to determine if the top ranking SNPs from pooled GWAS in TO versus non-TO in GD were associated with TO or GD as a whole or the risk variant was specific to TO in GD. We found that none of the seven top ranking SNPS was associated with TO compared to normal individuals or GD compared to normal individuals (Supplementary Table S2). Hence, SNP at MACROD2 locus was an association specific to risk of TO in GD. 
Four additional highly ranked SNPs at the MACROD2 gene locus with P < 10−4 extracted from the discovery pooled GWAS data also were assessed for their linkage disequilibrium to rs6110809 in the Caucasian HapMap population. They are rs978767 (P = 1.8 × 10−4), rs175805 (P = 3.05 × 10−7), rs6135575 (P = 4.32 × 10−4), and rs761684 (P = 3.2 × 10−4). Two SNPs were highly correlated with rs6110809 (rs6135575 and rs761684) and two SNPs showed more moderate linkage disequilibrium with rs6110809 (rs978767 and rs175805; Figure). This suggests the top ranking MACROD2 SNPs may represent two separate loci within MACROD2 associated with TO, although further genotyping and analysis are required to confirm this. 
Figure
 
Haploview linkage disequilibrium plot for 5 highly ranked SNPs within MACROD2 in HapMap CEU population data.
Figure
 
Haploview linkage disequilibrium plot for 5 highly ranked SNPs within MACROD2 in HapMap CEU population data.
Discussion
This GWAS discovered a positive association of SNP on chromosome 20 within the MACROD2 gene with TO with replication in an independent cohort. The genetic variant appeared specific to the risk for TO in patients with GD and not for the risk of GD itself, potentially implicating MACROD2 as a new genetic susceptibility locus for TO. Multiple genetic loci on chromosome 10, 7, 5, and 3 trended towards positive association with TO but the associations dropped out of significance in subsequent individual genotyping. 
To our knowledge, this is the first study to characterize the genetic risk factors for TO using a GWAS approach, where associated loci were discovered a priori and confirmation of these findings in additional independent cohorts was performed. While none of the SNPs reached genome-wide significance, there was evidence of association of MACROD2 with TO first reported by this study. The evidence of association with MACROD2 stays true overall with adjustment for potential confounders age, duration of GD, sex, and smoking in the combined dataset, bearing in mind that the smoking data are absent in 10.9% of the replication cohorts, and 6.6% in the combined cohorts dataset, which might affect the final estimation of OR and its significance level when smoking was taken into consideration. Hence, we have presented the data in this study for confounder adjustments with and without smoking and found P values slightly increased after smoking adjustment. A further limitation of this study is the use of pooled DNA, which, although highly economical, reduces the ability to correct for confounders in the GWAS stage, or to assess population substructure, which may lead to false-positive findings. While we and others find limited evidence for genes with major effects affecting risk for TO,18,39 the number of cases studied in this study still is relatively small, limiting the power and certainty of the association found. Much larger samples will be required to confirm our findings and to prove genetic associations in TO. Nevertheless, the identification of genetic variants, even of modest effects, will improve our understanding of pathogenesis of TO and identify pathways for therapeutic considerations. In the context of these provisions, we will discuss the results of interest from this study that are potentially relevant to the genetic risk of TO. 
The OR for rs6110809 A allele in predicting risk for TO ranges between 2.77 for the pooled discovery cohort and 1.42 in the replication cohorts, with an OR of 1.81 for the extended cohort when all cohorts were combined after adjusting for age at disease onset, duration of GD, and sex. The OR is notably smaller in the replication cohorts, which is consistent with “the winner's curse” where locus-specific size estimate bias upwards in GWAS.40 Hence, if the association of rs6110809 is true, the estimated conferred risk is more realistic from the replication samples, where the estimated effect size also is more in line with unbiased estimates of ORs discovered by GWAS studies of approximately 1.1 to 1.3.41 
The MACROD2 gene encodes a protein with a highly conserved macrodomain binding module in the N-terminal domain that interacts with mono-ADP-ribosylated proteins.42,43 There are no existing data as to whether MACROD2 is expressed in the orbital adipocytes or myofibroblasts, and further studies are required to examine this. However, MACROD2 is expressed strongly in the brain, and is moderately to strongly expressed in cuboidal epithelium of the lens and the inner nuclear layer of the retina.42 The association of MACROD2 with eye diseases has not been previously reported to our knowledge, but genetic variants in MACROD2 have been reported to be associated with neuropsychiatric disorders. De novo deletion of exon 5 within MACROD2 was identified in Kabuki syndrome, where the index case suffered from mental retardation and had ophthalmic features including ptosis, hypermetropic astigmatism, alternating strabismus, and blue sclera.42 MACROD2 allelic variants and common copy number gain has been linked to several diseases.41,4446 Recent GWAS identified SNPs within MACROD2 as a risk variant in autism spectrum disorder,41 and possibly a protective factor for MRI-defined brain infarcts.44 A candidate gene study investigating autistic-like traits also reported an association between rs4141463 SNP in intron 5 of MACROD2.45 Overexpression of MACROD2 mediates estrogen-independent growth and tamoxifen resistance in breast cancer, suggesting the expressed gene confers advantage to cellular survival.46 
The MACROD2 protein binds to mono-ribosylated proteins and its catalytic domain functions as O-acetyl-ADP-ribose deacetylases and hydrolases, which reverses mono-ADP ribosylation,43,47 thus underlying the key role of MACROD2 as an eraser of mono-ADP-ribosylation.48 The function of macrodomain molecules in general is tightly linked to ADP-ribosylation, which modulate functions of protein by posttranscriptional modification. The addition of ADP-ribose to proteins affects diverse cellular processes including DNA repair, chromatin remodeling, gene transcription, lipid metabolism, and apoptotic processes.4850 To understand the role of MACROD2 in disease, we must understand the function of intracellular ADP-ribosyltransferase Cholera toxin-like (ARTD), which inserts ADP-ribose monomers to protein that MACROD2 erases. ARTD10 regulates cell signaling, represses nuclear factor-κβ by mono-ribosylation of NF-κβ essential modulator (NEMO), where NF-κβ signaling pathway promotes inflammation, innate immune response, cellular proliferation, and cell survival.48,5053 
Overexpression of ARTD10 promotes apoptosis in HeLa cells whereas ARTD10 knockdown increases cell survival upon DNA damage caused by DNA damaging agents.54 ARTD10 is inducible by inflammatory and immunogenic stimuli, hence suggesting its role in innate immunity.48 ARTD10 also inhibits glycogen synthase kinase 3β (GSK3β), a key enzyme that modulates WNT signaling via β-catenin phosphorylation, which regulate apoptosis, immunity, and development of neurodegenerative diseases.43,48,55,56 Hence, inhibition of GSK3β and NF-κβ signaling by ARTD10 may explain enhanced apoptosis and MACROD2, on the other hand, counteracts the actions of ARTD10. 
The role of MACROD2 in modulating DNA repair is yet unknown, but it is recruited to poly-ADP-ribose synthesized by ARTD1 in response to DNA damage.50 MACROD2 removes the most proximal ADP ribose linked to the target amino acid after partial breakdown of poly-ADP-ribosylation by poly-ADPr-glycohydrolase to fully reverse ARTD1 regulatory modification.57 This functional link of MACROD2 to ARTD1 becomes more noteworthy in light of the findings that carriers of G allele of ARTD1 G1672A polymorphism are at risk of GD and a positive association of ARTD1 C410T polymorphism with TO confers an increased risk of TO by 1.7 times.58 
MACROD2 and ARTD10 regulation of apoptosis and NF-κβ signaling may be relevant to the mechanism of immune-mediated orbital inflammation in TO. Gene set enrichment analysis of molecular signatures of TO revealed dysregulation of epigenetic signatures, T cell activation, Th1 differentiation, defensin pathway, apoptosis among a few in active TO.59 Dysregulation of WNT signaling also was evident in TO from microarray studies.60,61 Orbital inflammation in TO is mediated by communication of orbital fibroblast and fibrocytes with immune cells via CD40-CD40 ligand. This is inducible by IFN-γ, evidenced by nuclear translocation of NF-κβ and increase secretion of pro-inflammatory cytokine IL-6 and IL-8 upon triggering of CD40 bearing orbital fibroblast by CD40 ligand.62,63 CD40 ligand significantly increases intercellular adhesion molecule-1 (ICAM-1) protein in a dose- and time-dependent manner, where ICAM-1 has a role in initiating and sustaining inflammatory immune response.64 The signaling pathways of increased ICAM-1 involve mitogen activated protein kinases and NF-κβ; CD40 ligand at 100 ng/mL induced a moderate activation of NF-κβ, which increases in a time-dependent manner.64 In addition, polymorphisms of del/ins of NFκβ1 gene may be related to the development of TO, further supporting the role of NF-κβ in TO pathogenesis.58 Overexpression of MACROD2 is expected to enhance NF-κβ signaling and increase cell survival, cell proliferation, inflammation and immune response, mechanisms all relevant to the pathogenesis of TO. The functional study of MACROD2 is beyond the scope of this study; this study provided grounds for further candidate gene study and future functional studies looking at expression levels of MACROD2 in orbital fibroblasts culture, and its effects on IL-6, IL-8, ICAM-1, NF-κβ expression levels in vitro. 
In conclusion, this exploratory GWAS study of TO suggests that in patients with GD, a common variant at the MACROD2 locus increases susceptibility to TO. This novel finding should prompt further replication studies to confirm its role in the pathogenesis of TO. 
Acknowledgments
The authors thank Simon Forehan, Mark Stein, Shane Hamblin, Rosemary Wong, Duncan Topliss, Peter Colman, Spiros Fourlanos, Cherie Chiang, Robert Thompson, and Matthew Doogue for private patient sources; Emmanuelle Souzeau for assistance in mail-out blood and data collection; and Stacey Rylander and Chamika De Silva for patients recruitment. 
Supported by Ophthalmic Research Institute of Australia, National Health and Medical Research Council of Australia, Australian Research Council Fellowship (SM), National Health, and Medical Research Council of Australia Fellowship (GM), Australian Postgraduate Award (JJK). 
Disclosure: J.J. Khong, None; K.P. Burdon, None; Y. Lu, None; L. Leonardos, None; K.J. Laurie, None; J.P. Walsh, None; A.D. Gajdatsy, None; P.R. Ebeling, None; A.A. McNab, None; T.G. Hardy, None; R.J. Stawell, None; G.J. Davis, None; D. Selva, None; A. Tsirbas, None; G.W. Montgomery, None; S. Macgregor, None; J.E. Craig, None 
References
Kendall-Taylor P, Perros P. Clinical presentation of thyroid associated orbitopathy. Thyroid. 1998; 8: 427–428.
Bartley GB. The epidemiologic characteristics and clinical course of ophthalmopathy associated with autoimmune thyroid disease in Olmsted County Minnesota. Trans Am Ophthalmol Soc. 1994; 92: 477–588.
Kendler DL, Lippa J, Rootman J. The initial clinical characteristics of Graves' orbitopathy vary with age and sex. Arch Ophthalmol. 1993; 111: 197–201.
Kavvoura FK, Akamizu T, Awata T, et al. Cytotoxic T-lymphocyte associated antigen 4 gene polymorphisms and autoimmune thyroid disease: a meta-analysis. J Clin Endocrinol Metab. 2007; 92: 3162–3170.
Velaga MR, Wilson V, Jennings CE, et al. The codon 620 tryptophan allele of the lymphoid tyrosine phosphatase (LYP) gene is a major determinant of Graves' disease. J Clin Endocrinol Metab. 2004; 89: 5862–5865.
Allahabadia A, Heward JM, Nithiyananthan R, et al. MHC class II region, CTLA4 gene, and ophthalmopathy in patients with Graves' disease. Lancet. 2001; 358: 984–985.
Tomer Y, Concepcion E, Greenberg DA. A, C/T single-nucleotide polymorphism in the region of the CD40 gene is associated with Graves' disease. Thyroid. 2002; 12: 1129–1135.
Pearson TA, Manolio TA. How to interpret a genome-wide association study. JAMA. 2008; 299: 1335–1344.
Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature. 2007; 447: 661–678.
Simmonds MJ. GWAS in autoimmune thyroid disease: redefining our understanding of pathogenesis. Nat Rev Endocrinol. 2013; 9: 277–287.
Nakabayashi K, Tajima A, Yamamoto K, et al. Identification of independent risk loci for Graves' disease within the MHC in the Japanese population. J Hum Genet. 2011; 56: 772–778.
Chu X, Pan CM, Zhao SX, et al. A genome-wide association study identifies two new risk loci for Graves' disease. Nat Genet. 2011; 43: 897–901.
Cooper JD, Simmonds MJ, Walker NM, et al. Seven newly identified loci for autoimmune thyroid disease. Hum Mol Genet. 2012; 21: 5202–5208.
Vaidya B, Imrie H, Perros P, et al. Cytotoxic T lymphocyte antigen-4 (CTLA-4) gene polymorphism confers susceptibility to thyroid associated orbitopathy. Lancet. 1999; 354: 743–744.
Vaidya B, Oakes EJ, Imrie H, et al. CTLA4 gene and Graves' disease: association of Graves' disease with the CTLA4 exon 1 and intron 1 polymorphisms, but not with the promoter polymorphism. Clin Endocrinol (Oxf). 2003; 58: 732–735.
Daroszewski J, Pawlak E, Karabon L, et al. Soluble CTLA-4 receptor an immunological marker of Graves' disease and severity of ophthalmopathy is associated with CTLA-4 Jo31 and CT60 gene polymorphisms. Eur J Endocrinol. 2009; 161: 787–793.
Han S, Zhang S, Zhang W, et al. CTLA4 polymorphisms and ophthalmopathy in Graves' disease patients: association study and meta-analysis. Hum Immunol. 2006; 67: 618–626.
Yin X, Latif R, Bahn R, Davies TF. Genetic profiling in Graves' disease: further evidence for lack of a distinct genetic contribution to Graves' ophthalmopathy. Thyroid. 2012; 22: 730–736.
Liu YH, Chen YJ, Wu HH, Wang TY, Tsai FJ. Single nucleotide polymorphisms at the PRR3 ABCF1, and GNL1 genes in the HLA class I region are associated with Graves' ophthalmopathy in a gender-dependent manner. Ophthalmology. 2014; 121: 2033–2039.
Bednarczuk T, Gopinath B, Ploski R, Wall JR. Susceptibility genes in Graves' ophthalmopathy: searching for a needle in a haystack? Clin Endocrinol (Oxf). 2007; 67: 3–19.
Huber AK, Jacobson EM, Jazdzewski K, Concepcion ES, Tomer Y. Interleukin (IL)-23 receptor is a major susceptibility gene for Graves' ophthalmopathy: the IL-23/T-helper 17 axis extends to thyroid autoimmunity. J Clin Endocrinol Metab. 2008; 93: 1077–1081.
Bednarczuk T, Hiromatsu Y, Seki N, et al. Association of tumor necrosis factor and human leukocyte antigen DRB1 alleles with Graves' ophthalmopathy. Hum Immunol. 2004; 65: 632–639.
Kamizono S, Hiromatsu Y, Seki N, et al. A polymorphism of the 5′ flanking region of tumour necrosis factor alpha gene is associated with thyroid-associated ophthalmopathy in Japanese. Clin Endocrinol (Oxf). 2000; 52: 759–764.
Alevizaki M, Mantzou E, Cimponeriu A, Saltiki K, Philippou G, Wiersinga W. The Pro 12 Ala PPAR gamma gene polymorphism: possible modifier of the activity and severity of thyroid-associated orbitopathy (TAO). Clin Endocrinol (Oxf). 2009; 70: 464–468.
Pawlak-Adamska E, Daroszewski J, Bolanowski M, et al. PPARg2 Ala(1)(2) variant protects against Graves' orbitopathy and modulates the course of the disease. Immunogenetics. 2013; 65: 493–500.
Planck T, Shahida B, Sjogren M, Groop L, Hallengren B, Lantz M. Association of BTG2 CYR61, ZFP36, and SCD gene polymorphisms with Graves' disease and ophthalmopathy. Thyroid. 2014; 24: 1156–1161.
Liu YH, Chen RH, Wu HH, et al. Association of interleukin-1beta (IL1B) polymorphisms with Graves' ophthalmopathy in Taiwan Chinese patients. Invest Ophthalmol Vis Sci. 2010; 51: 6238–6246.
Liu N, Li X, Liu C, Zhao Y, Cui B, Ning G. The association of interleukin-1alpha and interleukin-1beta polymorphisms with the risk of Graves' disease in a case-control study and meta-analysis. Hum Immunol. 2010; 71: 397–401.
Syed AA, Simmonds MJ, Brand OJ, Franklyn JA, Gough SC, Heward JM. Preliminary evidence for interaction of PTPN12 polymorphism with TSHR genotype and association with Graves' ophthalmopathy. Clin Endocrinol (Oxf). 2007; 67: 663–667.
Craig JE, Hewitt AW, McMellon AE, et al. Rapid inexpensive genome-wide association using pooled whole blood. Genome Res. 2009; 19: 2075–2080.
Burdon KP, Macgregor S, Bykhovskaya Y, et al. Association of polymorphisms in the hepatocyte growth factor gene promoter with keratoconus. Invest Ophthalmol Vis Sci. 2011; 52: 8514–8519.
Mohaseb K, Linder M, Rootman J, et al. Development and validation of a patient symptom questionnaire to facilitate early diagnosis of thyroid-associated orbitopathy in Graves' disease. Orbit. 2008; 27: 419–425.
Dolman PJ, Rootman J. VISA classification for Graves' orbitopathy. Ophthal Plast Reconstr Surg. 2006; 22: 319–324.
Wiersinga WM, Perros P, Kahaly GJ, et al. Clinical assessment of patients with Graves' orbitopathy: the European Group on Graves' Orbitopathy recommendations to generalists, specialists and clinical researchers. Eur J Endocrinol. 2006; 155: 387–389.
Macgregor S, Zhao ZZ, Henders A, Nicholas MG, Montgomery GW, Visscher PM. Highly cost-efficient genome-wide association studies using DNA pools and dense SNP arrays. Nucleic Acids Res. 2008; 36: e35.
Lu Y, Dimasi DP, Hysi PG, et al. Common genetic variants near the brittle cornea syndrome locus ZNF469 influence the blinding disease risk factor central corneal thickness. PLoS Genet. 2010; 6: e1000947.
Brown KM, Macgregor S, Montgomery GW, et al. Common sequence variants on 20q11.22 confer melanoma susceptibility. Nat Genet. 2008; 40: 838–840.
Purcell S, Neale B, Todd-Brown K, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007; 81: 559–575.
Villanueva R, Inzerillo AM, Tomer Y, et al. Limited genetic susceptibility to severe Graves' ophthalmopathy: no role for CTLA-4 but evidence for an environmental etiology. Thyroid. 2000; 10: 791–798.
Goring HH, Terwilliger JD, Blangero J. Large upward bias in estimation of locus-specific effects from genomewide scans. Am J Hum Genet. 2001; 69: 1357–1369.
Anney R, Klei L, Pinto D, et al. A genome-wide scan for common alleles affecting risk for autism. Hum Mol Genet. 2010; 19: 4072–4082.
Maas NM, Van de Putte T, Melotte C, et al. The C20orf133 gene is disrupted in a patient with Kabuki syndrome. J Med Genet. 2007; 44: 562–569.
Rosenthal F, Feijs KL, Frugier E, et al. Macrodomain-containing proteins are new mono-ADP-ribosylhydrolases. Nat Struct Mol Biol. 2013; 20: 502–507.
Debette S, Bis JC, Fornage M, et al. Genome-wide association studies of MRI-defined brain infarcts: meta-analysis from the CHARGE Consortium. Stroke. 2010; 41: 210–217.
Jones RM, Cadby G, Blangero J, Abraham LJ, Whitehouse AJ, Moses EK. MACROD2 gene associated with autistic-like traits in a general population sample. Psychiatr Genet. 2014; 24: 241–248.
Mohseni M, Cidado J, Croessmann S, et al. MACROD2 overexpression mediates estrogen independent growth and tamoxifen resistance in breast cancers. Proc Natl Acad Sci U S A. 2014; 111: 17606–17611.
Chen D, Vollmar M, Rossi MN, et al. Identification of macrodomain proteins as novel O-acetyl-ADP-ribose deacetylases. J Biol Chem. 2011; 286: 13261–13271.
Kaufmann M, Feijs KL, Luscher B. Function and regulation of the mono-ADP-ribosyltransferase ARTD10. Curr Top Microbiol Immunol. 2015; 384: 167–188.
Erener S, Mirsaidi A, Hesse M, et al. ARTD1 deletion causes increased hepatic lipid accumulation in mice fed a high-fat diet and impairs adipocyte function and differentiation. FASEB J. 2012; 26: 2631–2638.
Feijs KL, Forst AH, Verheugd P, Luscher B. Macrodomain-containing proteins: regulating new intracellular functions of mono(ADP-ribosyl)ation. Nat Rev Mol Cell Biol. 2013; 14: 443–451.
Burnett RM, Craven KE, Krishnamurthy P, et al. Organ-specific adaptive signaling pathway activation in metastatic breast cancer cells. Oncotarget. 2015; 6: 12682–12696.
Li M, Carpio DF, Zheng Y, et al. An essential role of the NF-kappa B/Toll-like receptor pathway in induction of inflammatory and tissue-repair gene expression by necrotic cells. J Immunol. 2001; 166: 7128–7135.
Naumann M. Nuclear factor-kappa B activation and innate immune response in microbial pathogen infection. Biochem Pharmacol. 2000; 60: 1109–1114.
Herzog N, Hartkamp JD, Verheugd P, et al. Caspase-dependent cleavage of the mono-ADP-ribosyltransferase ARTD10 interferes with its pro-apoptotic function. FEBS J. 2013; 280: 1330–1343.
Golpich M, Amini E, Hemmati F, et al. Glycogen synthase kinase-3 beta (GSK-3beta) signaling: Implications for Parkinson's disease. Pharmacol Res. 2015; 97: 16–26.
Beurel E, Grieco SF, Jope RS. Glycogen synthase kinase-3 (GSK3): regulation actions, and diseases. Pharmacol Ther. 2015; 148: 114–131.
Jankevicius G, Hassler M, Golia B, et al. A family of macrodomain proteins reverses cellular mono-ADP-ribosylation. Nat Struct Mol Biol. 2013; 20: 508–514.
Niyazoglu M, Baykara O, Koc A, et al. Association of PARP-1, NF-kappaB, NF-kappaBIA and IL-6, IL-1beta and TNF-alpha with Graves disease and Graves ophthalmopathy. Gene. 2014; 547: 226–232.
Khong JJ, Wang LY, Smyth GK, et al. Differential gene expression profiling of orbital adipose tissue in thyroid-associated orbitopathy. Invest Ophthalmol Vis Sci. 2015; 56: 6438–6447.
Ezra DG, Krell J, Rose GE, Bailly M, Stebbing J, Castellano L. Transcriptome-level microarray expression profiling implicates IGF-1 and Wnt signalling dysregulation in the pathogenesis of thyroid-associated orbitopathy. J Clin Pathol. 2012; 65: 608–613.
Kumar S, Leontovich A, Coenen MJ, Bahn RS. Gene expression profiling of orbital adipose tissue from patients with Graves' ophthalmopathy: a potential role for secreted frizzled-related protein-1 in orbital adipogenesis. J Clin Endocrinol Metab. 2005; 90: 4730–4735.
Sempowski GD, Rozenblit J, Smith TJ, Phipps RP. Human orbital fibroblasts are activated through CD40 to induce proinflammatory cytokine production. Am J Physiol. 1998; 274: C707–C714.
Gillespie EF, Raychaudhuri N, Papageorgiou KI, et al. Interleukin-6 production in CD40-engaged fibrocytes in thyroid-associated ophthalmopathy: involvement of Akt and NF-kappaB. Invest Ophthalmol Vis Sci. 2012; 53: 7746–7753.
Zhao LQ, Wei RL, Cheng JW, Cai JP, Li Y. The expression of intercellular adhesion molecule-1 induced by CD40-CD40L ligand signaling in orbital fibroblasts in patients with Graves' ophthalmopathy. Invest Ophthalmol Vis Sci. 2010; 51: 4652–4660.
Figure
 
Haploview linkage disequilibrium plot for 5 highly ranked SNPs within MACROD2 in HapMap CEU population data.
Figure
 
Haploview linkage disequilibrium plot for 5 highly ranked SNPs within MACROD2 in HapMap CEU population data.
Table 1
 
Clinical Characteristics of GWAS Discovery and Replication Cohorts
Table 1
 
Clinical Characteristics of GWAS Discovery and Replication Cohorts
Table 2
 
Demographics of Overall Cohort With GD, TO Compared to Normal Healthy Controls
Table 2
 
Demographics of Overall Cohort With GD, TO Compared to Normal Healthy Controls
Table 3
 
Top Ranking Pooled Genome-Wide Association Study and Validation Genotyping in ATOR Discovery Cohort
Table 3
 
Top Ranking Pooled Genome-Wide Association Study and Validation Genotyping in ATOR Discovery Cohort
Table 4
 
Replication Individual Genotyping of Top Ranking SNPs From Discovery GWAS in Replication Cohorts and in Combined Discovery and Replication Cohorts
Table 4
 
Replication Individual Genotyping of Top Ranking SNPs From Discovery GWAS in Replication Cohorts and in Combined Discovery and Replication Cohorts
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×