Human Molecular Genetics, 2021, Vol. 30, No. 5 393–409 doi: 10.1093/hmg/ddab023 Advance Access Publication Date: 30 January 2021 Association Studies Article A S S O C I AT I O N S T U D I E S A R T I C L E Genome-wide association study of circulating interleukin 6 levels identifies novel loci Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 Tarunveer S. Ahluwalia1,2,*,† , Bram P. Prins3 , Mohammadreza Abdollahi3 , Nicola J. Armstrong4 , Stella Aslibekyan5 , Lisa Bain6 , Barbara Jefferis7 , Jens Baumert8 , Marian Beekman9 , Yoav Ben-Shlomo10 , Joshua C. Bis11 , Braxton D. Mitchell12 , Eco de Geus13,14 , Graciela E. Delgado15 , Diana Marek16 , Joel Eriksson17 , Eero Kajantie18,19 , Stavroula Kanoni20 , John P. Kemp21,22 , Chen Lu23 , Riccardo E. Marioni24,25 , Stela McLachlan26 , Yuri Milaneschi27 , Ilja M. Nolte3 , Alexandros M. Petrelis28 , Eleonora Porcu29 , Maria Sabater-Lleal30,31 , Elnaz Naderi3 , Ilkka Seppälä32 , Tina Shah33 , Gaurav Singhal34 , Marie Standl8 , Alexander Teumer35,‡ , Anbupalam Thalamuthu36 , Elisabeth Thiering8,37 , Stella Trompet38,39 , Christie M. Ballantyne40 , Emelia J. Benjamin41,42 , Juan P. Casas43 , Catherine Toben34 , George Dedoussis44 , Joris Deelen9,45 , Peter Durda46 , Jorgen Engmann33 , Mary F. Feitosa47 , Harald Grallert8,48 , Ann Hammarstedt49 , Sarah E. Harris25,50 , Georg Homuth51 , Jouke-Jan Hottenga13,14 , Sirpa Jalkanen52,53 , Yalda Jamshidi54 , Magdalene C. Jawahar34 , Tine Jess55 , Mika Kivimaki56 , Marcus E. Kleber15 , Jari Lahti57,58 , Yongmei Liu59 , Pedro Marques-Vidal60,61 , Dan Mellström17 , Simon P. Mooijaart39 , Martina Müller-Nurasyid62,63,¶ , Brenda Penninx27 , Joana A. Revez6,|| , Peter Rossing1,64 , Katri Räikkönen58 , Naveed Sattar65 , Hubert Scharnagl66 , Bengt Sennblad30,67 , Angela Silveira30 , Beate St Pourcain22,68,69 , Nicholas J. Timpson22 , Julian Trollor36,70 , CHARGE Inflammation Working Group3 , Jenny van Dongen13,14 , Diana Van Heemst40 , Sophie Visvikis-Siest28 , Peter Vollenweider60,61 , Uwe Völker52 , Melanie Waldenberger8 , Gonneke Willemsen13,14 , Delilah Zabaneh71 , Richard W. Morris72 , Donna K. Arnett73 , Bernhard T. Baune74,75,76 , † Tarunveer S. Ahluwalia, http://orcid.org/0000-0002-7464-3354 ‡ Alexander Teumer, http://orcid.org/0000-0002-8309-094X ¶ Martina Müller-Nurasyid, http://orcid.org/0000-0003-3793-5910 || Joana A. Revez, http://orcid.org/0000-0003-3204-5396 Received: December 2, 2020. Revised: December 2, 2020. Accepted: January 4, 2021 © The Author(s) 2021. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 393 394 Human Molecular Genetics, 2021, Vol. 30, No. 5 Dorret I. Boomsma13,14 , Yen-Pei C. Chang12 , Ian J. Deary25,50 , Panos Deloukas20,77 , Johan G. Eriksson78,79 , David M. Evans21,22 , Manuel A. Ferreira6 , Tom Gaunt80,81 , Vilmundur Gudnason82,83 , Anders Hamsten30 , Joachim Heinrich8,84,85 , Aroon Hingorani33 , Steve E. Humphries33 , J. Wouter Jukema39,86 , Wolfgang Koenig87,88,89 , Meena Kumari56,90 , Zoltan Kutalik16,91 , Deborah A. Lawlor80,81 , Terho Lehtimäki32 , Winfried März15,66,92 , Karen A. Mather36,93 , Silvia Naitza29 , Matthias Nauck94,95 , Claes Ohlsson17 , Jackie F. Price26 , Olli Raitakari96,97,98 , Ken Rice99 , Perminder S. Sachdev36,100 , Eline Slagboom9,45 , Thorkild I.A. Sørensen101,102 , Tim Spector103 , Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 David Stacey104 , Maria G. Stathopoulou28 , Toshiko Tanaka105 , S. Goya Wannamethee7 , Peter Whincup106 , Jerome I. Rotter107 , Abbas Dehghan108 , Eric Boerwinkle109,110 , Bruce M. Psaty11,111 , Harold Snieder3 and Behrooz Z. Alizadeh3,* 1 Steno Diabetes Center Copenhagen, Gentofte DK2820, Denmark, 2 Department of Biology, The Bioinformatics Center, University of Copenhagen, Copenhagen DK2200, Denmark, 3 Department of Epidemiology, University of Groningen, University Medical Center Groningen, Groningen 9700 RB, The Netherlands, 4 Mathematics and Statistics, Murdoch University, Perth 6150, Australia, 5 Department of Epidemiology, University of Alabama at Birmingham School of Public Health, Birmingham, Alabama 35233, USA, 6 QIMR Berghofer Medical Research Institute, Brisbane 4006, Australia, 7 Department of Primary Care & Population Health, UCL Institute of Epidemiology & Health Care, University College London, London NW3 2PF, UK, 8 Institute of Epidemiology, Helmholtz Zentrum München—German Research Center for Environmental Health, Neuherberg 85764, Germany, 9 Department of Biomedical Data Sciences, Section of Molecular Epidemiology, Leiden University Medical Center, Leiden 2300 RC, The Netherlands, 10 Population Health Sciences, University of Bristol, Bristol BS8 2PS, UK, 11 Cardiovascular Health Research Unit, Department of Medicine, University of Washington, Seattle, WA 98101, USA, 12 Department of Medicine, University of Maryland School of Medicine, Baltimore, MD 21202, USA, 13 Department of Biological Psychology, Behaviour and Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam 1081 BT, The Netherlands, 14 Amsterdam Public Health Research Institute, Amsterdam University Medical Center, Amsterdam 1105 AZ, The Netherlands, 15 Vth Department of Medicine (Nephrology, Hypertensiology, Rheumatology, Endocrinology, Diabetology), Medical Faculty Mannheim, University of Heidelberg, Mannheim 68167, Germany, 16 SIB Swiss Institute of Bioinformatics, Lausanne 1015, Switzerland, 17 Department of Internal Medicine and Clinical Nutrition, Sahlgrenska Academy, Centre for Bone and Arthritis Research (CBAR), University of Gothenburg, Gothenburg 41345, Sweden, 18 Chronic Disease Prevention Unit, National Institute for Health and Welfare, PO Box 30, Helsinki 00271, Finland, 19 Hospital for Children and Adolescents, Helsinki University Central Hospital and University of Helsinki, Helsinki 00014, Finland, 20 William Harvey Research Institute, Barts & the London Medical School, Queen Mary University of London, London EC1M 6BQ, UK, 21 The University of Queensland Diamantina Institute, The University of Queensland, Woolloongabba, Queensland 4102, Australia, 22 MRC Integrative Epidemiology Unit, Population Health Sciences, University of Bristol, Bristol BS8 2BN, UK, 23 Department of Biostatistics, Boston University School of Public Health, Boston, MA 02118, USA, 24 Centre for Genomic and Experimental Medicine, Institute of Genetics and Molecular Medicine, University of Edinburgh, Edinburgh EH4 2XU, UK, 25 Centre for Cognitive Ageing and Cognitive Epidemiology, University of Edinburgh, Edinburgh EH8 9JZ, UK, 26 Usher Institute, University of Edinburgh, Edinburgh EH8 9AG, UK, 27 Department of Psychiatry, Amsterdam UMC, Vrije Universiteit, Amsterdam 1081 HJ, The Netherlands, 28 Université de Lorraine, Inserm, IGE-PCV, Nancy 54000, France, 29 Istituto di Ricerca Genetica e Biomedica, Consiglio Nazionale delle Ricerche, Monserrato (CA) 09042, Italy, 30 Cardiovascular Medicine, Department of Medicine Solna, Center for Molecular Medicine, Karolinska Institutet, Stockholm 17176, Sweden, 31 Unit of Genomics of Complex Diseases, Institut d’Investigació Biomèdica Sant Pau (IIB-Sant Pau), Barcelona 08041, Spain, 32 Department of Clinical Chemistry, Fimlab Human Molecular Genetics, 2021, Vol. 30, No. 5 395 Laboratories, and Finnish Cardiovascular Research Center—Tampere, Faculty of Medicine and Health Technology, Tampere University, Tampere 33520, Finland, 33 Institute of Cardiovascular Science, University College London, London WC1E 6BT, UK, 34 Discipline of Psychiatry, Adelaide Medical School, University of Adelaide, Adelaide 5005, Australia, 35 Institute for Community Medicine, University Medicine Greifswald, Greifswald 17475, Germany, 36 Centre for Healthy Brain Ageing, School of Psychiatry, University of New South Wales, Sydney 2052, Australia, 37 Division of Metabolic Diseases and Nutritional Medicine, Ludwig-Maximilians-University of Munich, Dr. von Hauner Children’s Hospital, Munich 80337, Germany, 38 Department of Cardiology, Leiden University Medical Center, Leiden 2300 RC, The Netherlands, 39 Section of Gerontology and Geriatrics, Department of Internal Medicine, Leiden University Medical Center, Leiden 2333 ZA, The Netherlands, 40 Baylor College of Medicine, Houston, TX 77030, USA, 41 National Heart, Lung, and Blood Institute’s and Boston University’s Framingham Heart Study, Framingham, MA 01702, USA, 42 Section of Cardiovascular Medicine and Preventive Medicine, Department of Medicine, Boston University School of Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 Medicine, Boston, MA 02118, USA, 43 Massachusetts Veterans Epidemiology Research and Information Center (MAVERIC), VA Boston Healthcare System, Boston, MA 02130, USA, 44 Department of Nutrition-Dietetics, Harokopio University, Athens 17671, Greece, 45 Max Planck Institute for Biology of Ageing, Cologne 50931, Germany, 46 Department of Pathology and Laboratory Medicine, Larner College of Medicine, University of Vermont, Burlington, VT 05405, USA, 47 Division of Statistical Genomics, Department of Genetics, Washington University School of Medicine, St. Louis, MO 63110-1093, USA, 48 German Center for Diabetes Research (DZD), Neuherberg 85764, Germany, 49 The Lundberg Laboratory for Diabetes Research, Department of Molecular and Clinical Medicine, Sahlgrenska Academy at the University of Gothenburg, Gothenburg SE-41345, Sweden, 50 Department of Psychology, University of Edinburgh, Edinburgh EH8 9JZ, UK, 51 Interfaculty Institute for Genetics and Functional Genomics, University Medicine Greifswald, Greifswald 17475, Germany, 52 MediCity Research Laboratory, University of Turku, Turku 20520, Finland, 53 Department of Medical Microbiology and Immunology, University of Turku, Turku 20520, Finland, 54 Genetics Research Centre, Molecular and Clinical Sciences Institute, St George’s University of London, London SW17 0RE, UK, 55 Department of Epidemiology Research, Statens Serum Institute, Copenhagen DK2300, Denmark, 56 Department of Epidemiology & Public Health, UCL Institute of Epidemiology & Health Care, University College London, London WC1E 7HB, UK, 57 Turku Institute for Advanced Studies, University of Turku, Turku 20014, Finland, 58 Department of Psychology and Logopedics, University of Helsinki, Helsinki 00014, Finland, 59 Department of Epidemiology and Prevention, Wake Forest School of Medicine, Winston-Salem, NC 27157, USA, 60 Department of Internal Medicine, Lausanne University Hospital (CHUV), Lausanne 1011, Switzerland, 61 University of Lausanne, Lausanne 1011, Switzerland, 62 IBE, Faculty of Medicine, Ludwig Maximilians University (LMU) Munich, Munich 81377, Germany, 63 Institute of Medical Biostatistics, Epidemiology and Informatics (IMBEI), University Medical Center, Johhanes Gutenberg University, Mainz 55101, Germany, 64 Department of Clinical Medicine, University of Copenhagen, Copenhagen DK2200, Denmark, 65 BHF Glasgow Cardiovascular Research Centre, Faculty of Medicine, Glasgow G12 8TA, UK, 66 Clinical Institute of Medical and Chemical Laboratory Diagnostics, Medical University of Graz, Graz 8036, Austria, 67 Department of Cell and Molecular Biology, National Bioinformatics Infrastructure Sweden, Science for Life Laboratory, Uppsala University, Uppsala 75124, Sweden, 68 Max Planck Institute for Psycholinguistics, Nijmegen XD 6525, The Netherlands, 69 Donders Institute for Brain, Cognition and Behaviour, Radboud University, Nijmegen 6525 AJ, The Netherlands, 70 Department of Developmental Disability Neuropsychiatry, School of Psychiatry, University of New South Wales, Sydney 2031, Australia, 71 Department of Genetics, Environment and Evolution, University College London Genetics Institute, London WC1E 6BT, UK, 72 Department of Population Health Sciences, Bristol Medical School, University of Bristol, Bristol BS8 1UD, UK, 73 Dean’s Office, College of Public Health, University of Kentucky, Lexington, KY 40536, USA, 74 Department of Psychiatry, Melbourne Medical School, The University of Melbourne, Parkville 3000, Australia, 75 Department of Psychiatry and Psychotherapy, University of Muenster, Muenster 48149, Germany, 76 The Florey Institute of Neuroscience and Mental Health, The University of Melbourne, Parkville 3000, Australia, 77 Centre for Genomic Health, Queen Mary University of London, London EC1M 6BQ, UK, 78 National Institute for Health and Welfare, University of Helsinki, Helsinki 00014, Finland, 79 Department of General Practice and Primary Health Care, University of Helsinki, Helsinki 00014, Finland, 80 MRC Integrative Epidemiology Unit at the University of Bristol, Bristol BS6 2BN, UK, 81 Population Health Science, Bristol 396 Human Molecular Genetics, 2021, Vol. 30, No. 5 Medical School, University of Bristol, Bristol BS8 2BN, UK, 82 Icelandic Heart Association, Kópavogur 201, Iceland, 83 Faculty of Medicine, University of Iceland, Reykjavik 101, Iceland, 84 Institute and Clinic for Occupational, Social and Environmental Medicine, University Hospital, LMU Munich, Munich 81377, Germany, 85 Allergy and Lung Health Unit, Melbourne School of Population and Global Health, The University of Melbourne, Melbourne 3010, Australia, 86 Durrer Center for Cardiogenetic Research, Amsterdam 1105 AZ, The Netherlands, 87 Deutsches Herzzentrum München, Technische Universität München, Munich 80636, Germany, 88 DZHK (German Centre for Cardiovascular Research), partner site Munich Heart Alliance, Munich 80336, Germany, 89 Institute of Epidemiology and Medical Biometry, University of Ulm, Ulm 89081, Germany, 90 Institute for Social and Economic Research, University of Essex, Colchester CO4 3SQ, Germany, 91 University Center for Primary Care and Public Health, University of Lausanne, Lausanne 1010, Switzerland, 92 SYNLAB Academy, SYNALB Holding Deutschland GmbH, Mannheim 68163, Germany, 93 Neuroscience Research Australia, Sydney 2031, Australia, 94 Institute of Clinical Chemistry and Laboratory Medicine, University Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 Medicine Greifswald, Greifswald 17475, Germany, 95 DZHK (German Center for Cardiovascular Research), partner site Greifswald, Greifswald 17475, Germany, 96 Centre for Population Health Research, University of Turku, Turku University Hospital, Turku 20520, Finland, 97 Research Centre of Applied and Preventive Cardiovascular Medicine, University of Turku, Turku 20520, Finland, 98 Department of Clinical Physiology and Nuclear Medicine, Turku University Hospital, Turku 20014, Finland, 99 Department of Biostatistics, University of Washington, Seattle, WA 98195, USA, 100 Neuropsychiatric Institute, Prince of Wales Hospital, Sydney 2031, Australia, 101 Novo Nordisk Foundation Center For Basic Metabolic Research, Section of Metabolic Genetics, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen DK2200, Denmark, 102 Department of Public Health, Section on Epidemiology, University of Copenhagen, Copenhagen DK1014, Denmark, 103 Department of Twin Research and Genetic Epidemiology, King’s College London, London SE1 7EH, UK, 104 MRC/BHF Cardiovascular Epidemiology Unit, Department of Public Health and Primary Care, University of Cambridge, Cambridge CB1 8RN, UK, 105 Longitudinal Study Section, Translational Gerontology Branch, National Institute on Aging, Baltimore, MD 21224, USA, 106 Population Health Research Institute, St George’s, University of London, London SW17 0RE, UK, 107 Department of Pediatrics, The Institute for Translational Genomics and Population Sciences, The Lundquist Institute at Harbor-UCLA Medical Center, Torrance, CA 90502, USA, 108 Department of Epidemiology, Erasmus MC, Rotterdam 3000 CA, The Netherlands, 109 Human Genetics Center, School of Public Health, University of Texas Health Science Center at Houston, Houston, TX 77030, USA, 110 Human Genome Sequencing Center, Baylor College of Medicine, Houston, TX 77030, USA and 11 1 Departments of Epidemiology and Health Services, University of Washington, Seattle, WA 98101, USA *To whom correspondence should be addressed at: Steno Diabetes Center Copenhagen, Gentofte, and Department of Biology, The Bioinformatics Center, University of Copenhagen, Copenhagen, Denmark. Email: tarun.veer.singh.ahluwalia@regionh.dk (Tarunveer S. Ahluwalia); Department of Epidemiology, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands. Email: b.z.alizadeh@umcg.nl (Behrooz Z. Alizadeh) Abstract Interleukin 6 (IL-6) is a multifunctional cytokine with both pro- and anti-inf lammatory properties with a heritability estimate of up to 61%. The circulating levels of IL-6 in blood have been associated with an increased risk of complex disease pathogenesis. We conducted a two-staged, discovery and replication meta genome-wide association study (GWAS) of circulating serum IL-6 levels comprising up to 67 428 (ndiscovery = 52 654 and nreplication = 14 774) individuals of European ancestry. The inverse variance fixed effects based discovery meta-analysis, followed by replication led to the identification of two independent loci, IL1F10/IL1RN rs6734238 on chromosome (Chr) 2q14, (Pcombined = 1.8 × 10−11 ), HLA-DRB1/DRB5 rs660895 on Chr6p21 (Pcombined = 1.5 × 10−10 ) in the combined meta-analyses of all samples. We also replicated the IL6R rs4537545 locus on Chr1q21 (Pcombined = 1.2 × 10−122 ). Our study identifies novel loci for circulating IL-6 levels uncovering new immunological and inf lammatory pathways that may inf luence IL-6 pathobiology. Human Molecular Genetics, 2021, Vol. 30, No. 5 397 Introduction ABO, the results of which were included in the discovery meta- Interleukin 6 (IL-6) is a multifunctional cytokine, which is analyses. The overall genomic control inflation factor (λGC after involved in a wide range of immunomodulatory processes, from correction) at the discovery stage meta-analysis was 1.0. cellular migration and adhesion to proliferation and maturation We identified 94 variants that were genome-wide signif- (1,2). Interleukins are involved in immune cell differentiation icantly (Pdiscovery < 5.0 × 10−8 ; Supplementary Material, Table and activation (3). IL-6 is synthesized by a variety of different S1) associated with IL-6 levels, representing two independent immune cells such as monocytes (4), B cells (5) and T cells (6) and genetic loci on chromosomes 1q21 and 6p21. Two common SNPs also non-immune cells such as epithelial and smooth muscle (rs4537545 and rs660895), one per locus, Chr. 1q21 (IL6R), and Chr. cells (7), adipocytes (8), endothelial cells (9) and osteoblasts (10). 6p21 (HLA-DRB1/HLA-DRB5), showed the most significant associ- Several factors have been implicated in circulating IL-6 ation with IL-6 levels (index SNPs) and the third SNP (rs6734238) levels. We have previously demonstrated that IL-6 levels mapped on Chr. 2q14 (IL1F10/IL1RN) locus showed suggestive decrease with age in children and increase with age in adults (5.0 × 10−8 < Pdiscovery < 1.0 × 10−5 ) association in addition to (11). Also, increased levels of IL-6 have been observed in various 5 other loci (LHFPL3, LZTS1, GPC5/GPC6, USP32/APPBP2, STAU1; diseases, not surprisingly in autoimmune diseases such as Supplementary Material, Table S2). Manhattan and QQ plots have been depicted in Figures 1A and 1B. Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 rheumatoid arthritis (12) and systemic juvenile idiopathic arthritis (13), but also cardio-metabolic diseases like type 2 The minor alleles of IL6R rs4537545∗ T (β = 0.091; Pdiscovery diabetes (14), heart failure, coronary heart disease (15) and = 8.39 × 10−85 ), IL1F10/IL1RN rs6734238∗ G (β = 0.025; Pdiscovery atherosclerosis (16), as well in cancers (17), atopic dermatitis = 1.45 × 10−7 ) and HLA-DRB1/5 rs660895∗ G (β = 0.036; Pdiscovery (18) and psychological disorders like depression (19). Due to = 1.80 × 10−9 ) associated with increased circulating IL-6 levels its implications in the pathogenesis of different disorders, Il- (Table 1). Two additional genome-wide significant SNPs in 6 has been used as an appropriate choice for drug targeting the IL1R locus, rs11265618 (β = 0.047; Pdiscovery = 1.21 × 10−15 ) and used as a monitoring biomarker of disease progression and rs10796927 (β = 0.034; Pdiscovery = 1.24 × 10−11 ), in low LD and response to treatments (20). The most illustrious IL-6 (r2 < 0.25) with the lead SNP rs4537545 were carried forward inhibitor is tocilizumab (21), a monoclonal antibody binding for replication, and later conditional analysis as they seemed the IL-6 receptor, which is already in use for treating patients potential candidates as independent signals. with allergic asthma (22), and immune system disorders like Overall, 12 SNPs spanning over 9 loci at a Pdiscovery < 1 × 10−5 in rheumatoid arthritis (23) and systemic juvenile idiopathic the discovery GWAS meta-analyses were selected for the repli- arthritis (24), with high efficacy with some initial benefits cation stage (Supplementary Material, Table S2). This included towards respiratory illnesses like COVID-19 (25). the three index SNPs, two additional SNPs from the 1q21 locus IL-6 baseline levels are heritable with estimates from twin (GWS but in low LD, r2 < 0.25 with index SNP) plus an addi- studies ranging between 15 and 61% (26–29). However, efforts to tional set of seven statistically suggestive SNPs with a P-value of identify genetic variants associated with levels of IL-6 consti- 5 × 10−8 < P < 1 × 10−5 in the discovery meta-analyses (either in tuted relatively small-scale GWAS (30–33) or sequencing-based low LD, r2 < 0.25 with the index SNP or independent loci). Addi- candidate gene association studies (34). To date, variants in the tionally, 3 SNPs as negative controls and 3 SNPs in LD (r2 > 0.25) IL-6 receptor gene (IL-6R) and the gene encoding histo-blood with the Chr.1 index SNP, to control for possible genotyping group ABO system transferase (ABO) have been identified as errors of index SNP across replication cohorts, were also added statistically significant for an association to IL6-levels. Also, the to the replication list, yielding 18 SNPs for replication stage genetic risk score constructed of IL-6 variants identified in the (Supplementary Material, Table S3). study by Shah and colleagues explained up to 2% of the variation Three loci including Chr.1q21 IL6R, Chr.6p21 HLA-DRB1/5 in IL-6 levels (33), leaving a substantial part of its heritability and Chr.2q14 ILF10/IL1RN replicated at Preplication < 0.05, reaching unexplained. These seemingly sparse results and limited find- GWS; 1q21 rs4537545, Pcombined = 1.20 × 10−122 ; 6p21 rs660895, ings could be due to limitations in the study power caused by low Pcombined = 1.55 × 10−10 ; and 2q14 rs6734238, Pcombined = 1.84 × 10−11 sample size or a great inter-individual variability of IL-6 levels. in the combined meta-analyses (Table 1 and Supplementary One may speculate a substantial increase in the study size by Material, Table S3). Locus zoom plots available in Figures 1C, increasing the number of participants, which would very likely 1D, and 1E. The two additional signals at Chr.1q21 IL6R lead to the identification of additional variants explaining IL-6 locus were replicated at Preplication = 1.7 × 10−4 for rs11265618 levels (35–37). and P = 0.03 for rs10796927, reaching Pcombined = 2.5 × 10−9 and The current study is the (till date) largest meta GWAS study Pcombined = 4.1 × 10−13 , respectively (Supplementary Material, including 67 428 individuals of European ancestry to identify Table S3). The conditional analysis confirmed that rs11265618 genetic variants explaining the levels of circulating IL-6 and to and rs10796927 SNPs were not independent from (Supplemen- understand underlying genetic mechanisms implicated in the tary Material, Table S4) but were driven by the index rs4537545 pathophysiology of this cytokine. SNP. In both, discovery and replication association analyses, the effect directionality was generally consistent across individ- ual studies for GWS variants, while there was some evidence Results of borderline heterogeneity in one of the two novel loci (I2 (P A total of 52 654 individuals of European descent from 26 cohorts value) < 0.05) during the discovery and combined meta-analysis were included in the discovery GWAS meta-analysis with up to (Table 1, Fig. 2). The imputation quality scores (r2 ) for the GWS 2 454 025 autosomal SNPs passing quality control. Four cohorts (index) SNPs for each cohort (discovery and replication) are avail- (ALSPAC, MONICA/KORA, NTR and SardiNIA) identified genome- able in Supplementary Material, Table S5. The other seven SNPs wide significant associations in the ABO region, whereas none of that showed suggestive association in the discovery stage, and the other 22 cohorts did, either individually or combined. These expectedly the negative control SNPs did not reach GWS in the cohorts conditioned their results on their relevant top-SNP in combined meta-analyses (Supplementary Material, Table S3). 398 Human Molecular Genetics, 2021, Vol. 30, No. 5 The three GWAS index SNPs when combined explained <0.01 (72) Index SNPs that reached P < 5 × 10–8 in the combined analysis from each locus are reported here. Sample sizes: discovery cohorts, n = 52 654; replication cohorts, n = 14 774; combined, n = 67 428. The effect sizes (β) in the discovery phase, given for the effect allele. EAF: effect allele frequency; effect sizes and standard error (SE) values are based on natural log transformed IL6 (pg/ml) levels. Phet: combined meta-analysis heterogeneity P value; I2 : heterogeneity 0.03 (32) 0.08 (26) Phet (I2 ) approximately 1.06% of the variance in circulating levels of IL- 6 using data from the NESDA cohort. The phenotypic variance explained by all the common variants was estimated to be 4.45% using the SumVg method (38). HLA-DRB5/DRB1 Nearest Genes IL1F10/IL1RN Replication of other known/suggestive loci for IL-6 IL6R was the only IL6 known locus that we replicate at GWS. IL6R IL1RN and HLA-DRB1, our primary findings have been reported as suggestive loci (1 × 10−6 < P < 1 × 10−4 ) by Shah et al. while some known/suggestive IL6 loci (ABO, BUD13, TRIB3 and SEZ6L) did not Annotation replicate (Pdiscovery > 0.05) in the current study. Intergenic Intergenic Intronic Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 SNP functionality We looked up SNPs in LD with the index SNPs from the 1.20 × 10−122 immunologically associated loci including IL-6R, rs4537545, 1q21; 1.84 × 10−11 1.55 × 10−10 IL1F10, IL1RN, rs6734238, 2q14, intergenic; and HLA-DRB1/DRB5, Pcombined rs660895, 6p21, intergenic. The search for functional/missense variants in high LD (r2 > 0.8) with the lead SNPs led to the identification of only one nonsynonymous rs2228145 SNP in LD Table 1. Novel and replicated loci associated with circulating IL-6 levels at P < 5.0 × 10−8 in the combined GWAS meta analyses (r2 = 0.95) with the rs4537545 index SNP from the IL6R locus. We 7.88 × 10−37 3.24 × 10−5 3.38 × 10−2 used the Combined Annotation-Dependent Depletion (CADD) Preplication database to identify the functionality, i.e. deleterious, disease causal, pathogenicity, of rs2228145 in IL6R. CADD is an integrative annotation based on multiple genomic features scored into a single metric (39). IL6R missense the rs2228145 variant has a 8.39 × 10−85 CADD score of 15.98 (https://cadd.gs.washington.edu). 1.45 × 10−7 1.80 × 10−9 Pdiscovery Associations with other traits and gene expression data Genome-wide significant associations between IL6-associated top SNPs and other traits, and gene expression, data were mined 0.025 (0.005) 0.036 (0.006) 0.091 (0.005) using the Pheno Scanner v2 database (accessed, October 2020). Beta (SE) GWAS-based IL1F10/IL1RN rs6734238∗ G allele has been asso- ciated with increased levels of serum C-reactive protein (CRP) and decreased fibrinogen levels, and blood cell traits in recent GWAS reports (40,41) (PMID:27863252; Supplementary Material, 0.42 0.19 0.39 EAF Table S6). HLA-DRB1/DRB5 rs660895∗ G allele is associated with increased risk of rheumatoid arthritis (RA) in Europeans and Asians (42), Effect/Other allele IgA nephropathy in Asians (43), while the decreased risk of ulcerative colitis and inflammatory bowel disease (IBD) (44). IL-6R rs4537545∗ T allele has been associated with increased circulating CRP levels (45), a decreased risk of RA (42) in mixed ancestries, while an increased risk of diabetes and asthma from G/A G/A T/C the UK Biobank Neale’s lab rapid GWAS (See Web Resources; Supplementary Material, Table S6). IL6R rs4537545T∗ allele is also associated with C-reactive protein, allergic disease, rheumatoid 113 841 030 154 418 879 32 577 380 BP (Hg19) arthritis and coronary artery disease (Supplementary Material, Table S6). Gene expression. IL1F10/IL1RN rs6734238 is associated with IL1F10/IL1RN expression levels in the skin, peripheral blood Lead SNP (rsID) and whole blood (P < 5.0 × 10−8 ; Supplementary Material, Table Replicated Known Locus rs6734238 rs4537545 S7). HLA-DRB1/DRB5 rs660895 has been associated with HLA- rs660895 DRB1/DRB5/DRB6/DQB1/DQB2 expression levels in multiple tissues including peripheral blood, whole blood, monocytes, adipose tissue, thyroid, tibial artery, coronary artery, heart, lung, Novel Loci measure. brain, colon, skeletal muscle, tibial nerve, skin and lymphoblas- 2q14 6p21 1q21 toid cell lines (P < 5.0 × 10−8 ; Supplementary Material, Table S7). Chr IL6R rs4537545 SNP is also associated with IL6R expression levels Human Molecular Genetics, 2021, Vol. 30, No. 5 399 in peripheral and whole blood (P < 5.0 × 10−8 ; Supplementary domain containing E (SHE), and Tudor domain containing 10 Material, Table S7). (TDRD10), but have been ruled out to play a role at this locus (33). At the identified chromosome 2 locus the lead SNP, rs6734238, is intergenic and has also been associated with circulating Power estimates CRP and fibrinogen levels (40,41,62). The nearest genes to Based on power calculator and assumptions mentioned under this locus are the interleukin 1 family member 10 (IL1F10, methods section, the estimated power for the 2 novel index distance = 7.6 kb, currently known as IL-38) and interleukin 1 SNPs was 98.3% rs6734238 (effect allele frequency, EAF: 0.42), and receptor antagonist (IL1RN, distance = 34.4 kb). IL1F10/IL-38 and 76.9% rs660895 (EAF: 0.19), respectively. IL1RN variants (rs6759676 and rs4251961) in partial LD with the lead SNP (r2 LD :0.10 and 0.61) have been recently reported to be protective against the development of insulin resistance (63). Discussion This further supports the molecular mechanisms behind IL-6- We performed the largest (to date) GWAS meta-analysis for mediated insulin secretion via glucagon-like peptide 1 (GLP-1) circulating IL-6 levels, which includes 66 341 individuals of Euro- (64) contributing to type 2 diabetes (T2D) pathophysiology. For pean ancestry. We identified three loci associated with levels of IL-6 specifically, it has been found that synthesis increases when Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 circulating of IL-6 in the general population amongst which two dendritic cells are stimulated by bacterial lipopolysaccharides are novel (Chr6p21, and Chr2q14), located in/nearby genes (HLA- (LPS) in the presence of IL1F10 (65). IL-1RN is another member DRB1 and IL1RN/IL-38) with inflammatory roles explaining up to of the interleukin 1 cytokine family, with suggestive evidence 1.06% variance. for involvement in determining IL-6 levels in the blood. One The strongest associated SNP, interleukin 6 receptor (IL-6R) study found significant associations of IL-1RN rs4251961 with rs4537545 at the 1q21 locus, is in high LD (r2 = 0.95) with a plasma CRP and IL-6 levels, albeit not independently replicated missense IL-6R SNP rs2228145 (D358A) that results in an amino and not genome-wide significant (P = 1 × 10−4 and P = 0.004) acid substitution at position 358 (Asp → Ala) on the extracellular (66). Our lead SNP was not in high LD (r2 < 0.8) with variants domain of IL-6R and a high CADD score suggesting that the in either neighboring genes and therefore in conjunction with variant is pathogenic or functional or deleterious (among top its intergenic position, identifying a causal variant in this locus 10% variants of the genome). The missense SNP is known to remains non-trivial. impair the responsiveness of cells targeted by IL-6 (46) by reduc- The 6p21 rs660895, which was identified, resides within the ing IL-6R expression on cell surfaces (47), and increasing levels HLA region, which forms one of the most complex genomic of soluble IL-6R in individuals homozygous for this mutation regions to study due to its large LD blocks and sequence diver- (48,49). Recently it has been demonstrated that increased levels sity. This region has some population substructure in Euro- of sIL-6R induced by this variant can be explained by ectodomain peans, which may have influenced the results; however, (1) shedding off IL-6R, a mechanism in which membrane-associated each cohort population substructure adjustment was applied, proteins are rapidly converted into soluble effectors whereby followed by genomic correction for overall discovery stage meta- simultaneously cell surface expression of the same protein is analyses. Thus, we reduced the chances that the population reduced (50). Increased levels of sIL-6R may act as a counter- substructure may have had on this locus. The nearest genes balance to limit exaggerated IL-6 signaling and may explain the to the index SNP, HLA-DRB1 (distance = 19.8 kb) and HLA-DQA1 protective effect of the 358A allele for various cardiovascular (distance = 27.8 kb) are both histocompatibility complex genes diseases including coronary artery disease (CAD) (51–53), atrial encoding proteins that form cell surface complexes for certain fibrillation (54), lung function in asthmatics (55) and abdominal immune system cells helping in antigen presentation to trigger aortic aneurysm (56) as well as RA (57). However, in contrast an immune response. It is noteworthy that variations at this with this finding, the IL-6-sIL-6R complex itself is capable of locus code for antigen-presenting complexes (APCs), which have transducing IL-6 signaling to non-IL-6R expressing cells, known been previously associated with diseases having a dysfunctional as trans-signaling (58), and it is this mechanism, as opposed to immune system; while we report for the first time that there classic signaling, that is linked to chronic inflammatory disor- exists also a strong association of this locus with circulating ders including IBD and RA (59). Blocking IL-6 signaling cascades cytokine levels. Therefore, the association of this locus with can be achieved by using an IL-6R specific inhibitor in the form the disease may corroborate through its effect through IL6 lev- of a monoclonal antibody, tocilizumab, which is a widely used els. One high-LD SNP (rs9272422, r2 = 0.82 with our index SNP, therapy in the treatment of RA. Several variants in IL-6R, includ- rs660895) residing in the promoter region of HLA-DQA1 support ing rs2228145, may assist in the prediction of patient response to this hypothesis and has been identified previously for systemic tocilizumab in RA (60). The rs4537545∗ T allele that is associated lupus erythematosus (67) and ulcerative colitis (68). rs660895∗ G with IL6 levels is known to associate with increased circulating allele is associated with increased risk of RA in Europeans and CRP levels (61) and a decreased risk of RA (42) in studies com- Asians (42), IgA nephropathy in Asians (43), while the decreased prising mixed ancestries. Moreover, this SNP has been associated risk of ulcerative colitis and IBD. (44) with IL6R expression in peripheral blood, skin, brain and adipose Various studies aimed to identify genetic variation underly- tissue (Supplementary Material, Table S7). The causal involve- ing levels of IL-6 (22–26) have found genome-wide significant ment of IL-6 levels in disease remains to be elucidated, but a associations in the IL-6R and ABO genes. The study performed recent study using a Mendelian randomisation (MR) approach by Shah and colleagues (33) found suggestive evidence (non- did demonstrate that by using this SNP as instrumental variable, GWS; P = 3.8 × 10−6 , respectively) for additional loci, including modelling the effects of tocilizumab, that IL-6R signalling has ABO, BUD13, TRIB3 and SEZ6L, none of which replicate in the a causal effect on CAD (52). On the other hand, pleiotropic current study (Pdiscovery > 0.05) indicating that these might be nature of the IL-6R locus, influencing IL-6, CRP and fibrinogen false-positive findings due to low sample size (∼7800) or loci with levels, prohibits instrumental variable analysis and attribution sex-specific effects (associations based on women dominant of causality to one particular intermediate. Finally, several other population) or due to technical shortcomings with measurement genes encompass the 1q21 locus, including Src homology 2 assay (ABO locus). 400 Human Molecular Genetics, 2021, Vol. 30, No. 5 It is surprising that even with increased statistical power confounders as we had limited access to summary level data. (ndiscovery = 52 654; nreplication = 14 774) in the current study (com- In conclusion, we identify two novel common genetic variants pared to the previous IL6 GWAS) (33), we could identify three associated with circulating IL-6 levels that may influence the genetic loci (1q21, 2q14 and 6p21) accounting for ∼1% of the pathophysiology of complex cardio-metabolic, psychiatric and genetic variance for circulating IL-6 levels. According to the cur- immunological traits, among individuals of European ancestry. rent estimates, the heritability levels for IL6 levels range between This is a step further towards unravelling new biological path- 15 and 61%, suggesting that an enormous increase in sample ways and potential therapeutic targets that can be developed for sizes would be required to identify additional variants explain- the IL-6-related disorders, while suggesting looking deeper into ing this remaining heritability. Multiple explanations for this so- the genome for coding variants (rare and common) having larger called missing heritability phenomenon have been proposed in individual effects (Figs 1 and 2). the past, which can be sought in rare or low frequency coding variants as observed for a similar metabolic quantitative trait by us (69) or can be explained by non-additive effects, which may Material and Methods cause inflated estimates of heritability. Plausible evidence for Discovery stage other sources of unexplained heritability that have been found Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 are epigenetic changes, and haplotypes of common SNPs. Study populations. The overall study design (Supplementary Collectively, our results provided additional insights into the Material, Fig. S1) involved the discovery cohorts with 53 893 biology of circulating IL-6. We identify new loci, limited by com- individuals. After overlapping individuals with available geno- mon variants in the Hap Map Reference panel. Albeit this is com- type and phenotype data, the discovery stage included 52 654 parable to the 1000 genomes reference panel (70) but narrower individuals from 26 cohorts of European ancestry listed under compared to some newly available panels that show greater vari- Supplementary Material, Table S8, described in Supplementary ant coverage in numbers and frequency range. Future studies are Text S1 and study summary characteristics in Supplementary recommended to aim for identification of additional common Material, Table S9. Only population-based samples or healthy but also rare variants, by firstly using richer imputation panels, controls from case–control studies were included in the final such as UK10K project or the Haplotype Reference Consortium, analyses. a strategy that holds great promise, and secondly by making use of genetically isolated populations. Thirdly, we would like Serum IL-6 measurements. Each study typically collected to stress the importance of phenotype harmonization. As we venous blood samples stored below −80◦ C until the time identified genome-wide variants in the ABO locus, in four stud- of measurement using various types of immunoassays and ies participating in the discovery, but not in the remaining 22 expressed as pg/ml as presented in Supplementary Material, cohorts, there is a strong indication that this locus may be assay Table S10. The trait transformation and phenotype data quality specific. However, a proper demonstration of this hypothesis control (QC) were presented by Supplementary Material, Text would require further testing, including repeating the GWAS in S3 (Supplementary Material, Text S3.1 and S3.2). In brief, ABO-positive cohorts using a different IL-6 assay. Indeed it is participating cohorts have checked for the percentage of emerging that the ABO locus has pleiotropic effects on many missingness in IL6 measurements and evaluated for indices different traits and diseases (71), which would suggest a more of QC (Supplementary Material, Text S3.2), yielded the final thorough analysis before disregarding this signal. Also, conven- number of participants with available validated IL6 levels, of tionally increasing sample sizes without correction for popula- whom those with available genotype data were included in the tion substructures may raise heterogeneity within populations study as characterized in Supplementary Material, Table S9 and (72), likely concealing the SNPs that affect particular subgroups. in Supplementary Material, Text S1 and S2. Future specific studies should counter the widely held assump- tion of unconditional risk alleles of complex traits and focus on the importance of studying more homogenous subgroups Genotyping and imputation. Each participating cohort per- to, for example, investigate the age-dependent effect of genetic formed genome-wide genotyping using a variety of genotyping variants (73,74). Here, while further exploring the pleiotropic platforms and applied a predefined quality control (QC) of effect of IL-6-related variants, we identified phenotypes differ- genotype data (Supplementary Material, Table S11) followed entially regulated by diverse variants in the 1q21 locus. Bio- by performing imputation of non-genotyped genetic variants, logic systems are dynamic complex networks and are evolv- on the backbone of haplotypes inferred from the Hapmap ing through lifespan and investigating the interrelationships Phase II reference panel (NCBI Build 36), and using statistical existing between phenotypes as well as between genetic vari- software such as IMPUTE (75), MACH, Minimac (76) or BIMBAM ations and phenotypic variations has the potential for uncov- (77) (Supplementary Material, Table S11). Each cohort was ering the complex mechanisms. This is the case here for IL-6 recommended a set of general SNP quality filters including and tailored methodologies should be devoted to the study of MAF < 0.01; Hardy Wienberg Equilibrium (HWE) P ≤ 10−6 ; such traits, hopefully resulting in clinically significant break- imputation quality r2 ≤ 0.3; and genotyping call rate <0.95 throughs. Future collaborative efforts therefore should strive to (Supplementary Material, Fig. S1). Once we received summary use well-calibrated assays, z-standardized protocols for sample results from each participating study, we ran a series of QC handling, and processing (75), though this will be difficult to checks. Firstly, these included a set standard checks, including achieve in practice. Lastly, we have attempted to perform formal the imputation quality filters (basis the imputation program association-based causal analysis to identify the likely causal used and/or r2 imputation < 0.3 were excluded), and then checks for loci, using the DEPICT approach; unfortunately, instead with only genomic inflation (quantile–quantile or QQ plots). We adapted 2 novel GWS findings, our analyses were underpowered and thus filter thresholds per cohort to reduce any observed deviation not included. We also mine the gene expression and eQTL data from the null while missing SNP loss due to the QC process. for the identified SNPs using established databases; however, we Finally, ∼2.45 million (2 454 025) common SNPs were part of the were unable control for random co-localization signals or other discovery meta-analysis. Human Molecular Genetics, 2021, Vol. 30, No. 5 401 Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 Figure 1. Manhattan, QQ and LocusZoom plots of the discovery GWAS meta-analyses. (A) Manhattan plot showing the association of SNPs with IL-6. Loci coloured in red or blue, three in total, represent those for which the lead SNPs reached genome-wide significance (P = 5 × 10−8 ). Horizontal axis: relative genomic position of variants on the genome, vertical axis: −log10 P-value of each SNP; (B) Quantile–quantile plot for P-values obtained from the meta-analysis. The horizontal and vertical axes represents the expected distribution of −log10 (P-values) under the null hypothesis of no association, whereas the vertical axis shows the observed −log10 (P- values). The blue dashed line represents the null, and λGC value represents the genomic inflation factor lambda. Each data point represents the observed versus the expected P-value of a variant included in the association analyses; (C–E) Regional association plots for each of the three genome-wide significant loci, 1q21, 2q14 and 6p21, respectively. Pairwise LD (r2 ) with the lead SNP is indicated following a color-coded scale. Horizontal axis: relative genomic position of variants within the locus, vertical axis: −log10 P-value of each SNP. 402 Human Molecular Genetics, 2021, Vol. 30, No. 5 Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 Figure 2. Combined discovery and replication forest plots for the GWAS Index SNPs. Forrest plots for (A) IL6R rs4537545 (chr. 1q21), (B) IL1RN rs6734238 (chr. 2q14), (C) HLA-DRB5 rs660895 (chr. 6p21) with discovery, replication and combined effect estimates, 95% CI and weights based on the fixed effects inverse variance meta-analyses. Statistical methods additive genetic effects, accounting for imputation uncertainty GWAS analysis. Each study conducted an independent GWAS while adjusting for age, sex, population substructure (through analysis between SNPs and natural log-transformed values of study-specific principal components) and/or study-specific site, serum IL-6 levels following a predefined analysis plan (Sup- when necessary. GWAS summary result obtained from each plementary Material, Methods S4). Association analyses were cohort underwent a series of QC checks using the QCGWAS pack- conducted using linear regression model, or linear mixed effect age in R (78) (Supplementary Material, Text S3 and Supplemen- models to account for familial correlation when warranted, with tary Material, Fig. S1). Being aware of the potential false-positive Human Molecular Genetics, 2021, Vol. 30, No. 5 403 Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 Figure 2. Continued. association in the ABO region on chromosome 9 (28,30), while in the final discovery meta-analysis (Supplementary Material, using an R&D systems high-sensitivity assay kit to measure IL6 Text S3.3). levels (R&D systems, Minneapolis, MN, USA), four (out of 22) discovery cohorts that observed genome-wide significant results Discovery GWAS meta-analyses. Individual GWAS results from in the ABO locus were asked to rerun the GWAS analysis condi- 26 European studies were meta-analyzed using the inverse vari- tional on the top ABO SNP (i.e. rs8176704) before including them ance weighted, fixed-effects method as implemented in GWAMA 404 Human Molecular Genetics, 2021, Vol. 30, No. 5 Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 Figure 2. Continued. while applying the double genomic control (GC) correction for LD with the index SNP (r2 < 0.2) and (2) available in at least 50% population stratification, i.e. first to each study individually and of study cohorts. subsequently also to the pooled results after meta-analysis (79). Regional association plots for the discovered loci were gen- Replication and combined meta-analysis erated through the LocusZoom (78) tool. We used the SNAP tool (80) to perform the pairwise LD checks (HapMap release 22 data) Study population, phenotyping and QC. The overall study and to verify low LD with secondary signals. All SNPs selected (Supplementary Material, Fig. S1) comprised 15 785 individuals for the replication stage had to fulfill the following criteria: (1) for replication. After removing individuals with missing data, having an association Pdiscovery ≤ 1 × 10−5 and being in very low the replication analyses were performed using a combination Human Molecular Genetics, 2021, Vol. 30, No. 5 405 of in silico and de novo genotyping in 14 774 individuals from 12 variants was estimated using the observed Z-statistics from the cohorts of European ancestry as described in Supplementary discovery analyses for a subset of pruned SNPs. Following the Material, Text S2. Similar QC (Supplementary Material, Text S3 original method (SumVg) (38), we pruned the imputed (based on and Supplementary Material, Table S11) and statistical checks the 1000G Phase1 Integrated Release, Version 3, 2012.04.30 reference were made as in the discovery stage. panel) genotypes of the NESDA cohort using PLINK v1.07 (84), by Venous blood samples (serum or plasma) were collected removing correlated SNPs at r2 > 0.25 within a 100-SNP sliding and stored at −80◦ C. Serum/plasma IL-6 levels (pg/ml) were window, and with a step size of 25 SNPs per forwarding slide. measured using various immunoassay methods described in This resulted in a pruned set of 163 459 SNPs. Supplementary Material, Table S10. Each cohort tested the selected SNPs using the same statistical model as for the SNP mapping and functionality. We searched for variants in high discovery association analyses (Supplementary Material, Fig. LD (r2 > 0.8) within a 1 Mb region on either side of the lead S1). Effect size estimates of all replication SNPs from each SNPs using 1000 Genomes sequence data (Phase1 Integrated replication study were compared with the effect size estimates Release, Version 3, 2012.04.30), utilizing tools available in Liftover from the discovery meta-analyses. When effect sizes from (85), VCFtools (86) and clumping in PLINK (84). We subsequently individual cohorts did not align, we excluded these cohorts from annotated these variants using ANNOVAR (87) with the RefSeq Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 the replication meta-analyses (ncohorts = 3). To account for the (88) database for variant function and genic residence or dis- inter-study assay differences insensitivity, we combined results tance. We used the Combined Annotation-Dependent Depletion across the replication studies using a fixed effect sample size (CADD) database to identify the functionality, i.e. deleterious, weighted Z-score meta-analysis as implemented in the METAL disease causal, pathogenicity, for the index SNPs. package (https://genome.sph.umich.edu/wiki/METAL) (81). The summary GWAS meta-analyses result from the discovery Associations with other traits and gene expression data. Pheno and replication stages were then used to perform a combined Scanner v2 (89) data mining tool was used (Access date Octo- (discovery + replication) GWAS meta-analysis using a sample ber 2020) to identify existing GWS (at P < 5 × 10−8 ) associations size weighted Z-score method. Test for heterogeneity was also between IL6 identified SNPs and other traits, and gene expres- performed as part of the meta-analysis package using METAL sion/eQTLs data. where I2 statistic denoting the percentage of variation across studies was estimated (I2 = 100% × (Q − df)/Q) where Q is the Chi- Square statistic. Significance for heterogeneity was denoted by Power calculation the heterogeneity (or HetP ) P values. Variants that were signif- We used GWAs power estimator (see Web Resources) by assum- icant in the replication meta-analysis at P < 0.05 and had an ing a relative risk of 1.10 (or an effect estimate of 0.10), given overall Pcombined < 5 × 10−8 in the combined meta-analysis were N = 66 000, alpha (P-value) = 5 × 10−8 (also GCTA power calculator, considered statistically GWS. SNPs within the range of 1 Mb Supplementary Text S3.5). (or 106 bases) on either side of the most significant (i.e. index) SNP (with LD, r2 > 0.25) were considered part of the same locus, whereas those in low LD (r2 < 0.25) were tested if they were Supplementary Material independent using conditional analysis. Supplementary Material is available at HMG online. GWAS sum- mary statistics data is available upon request to corresponding Conditional analysis. We performed an approximate joint con- authors T.S.A. and/or B.Z.A. ditional analysis to identify distinct signals in a specific chro- mosomal region as implemented in GCTA (82) using high-quality genome-wide genotyped/imputed reference data from two stud- Web Resources ies (NEtherlands Study of Depression and Anxiety (NESDA) from the Netherlands and/or Genetics of Obesity in Young Adults QCGWAS, https://cran.r-project.org/web/packages/QCGWAS/i (GOYA) from Denmark) to estimate linkage disequilibrium (LD) ndex.html (83) between SNPs. GWAMA, http://www.well.ox.ac.uk/gwama/ Conditional analysis for identification of independent signals METAL, http://csg.sph.umich.edu//abecasis/metal/ was performed on GWS SNPs ( ±1 Mb to the index SNP and GCTA, http://www.complextraitgenomics.com/software/gcta/ having low LD, r2 < 0.2 with the index SNP) using summary statis- LocusZoom, http://csg.sph.umich.edu/locuszoom/ tics from the discovery GWAS meta-analysis data (—COJO option 1000 Genomes, ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/re in GCTA) after confirming the GWS loci from the combined lease/20110521/ meta-analysis. PLINK, http://pngu.mgh.harvard.edu/purcell/plink/ VCFtools, http://vcftools.sourceforge.net/ Heritability estimates. We approximated the variance explained ANNOVAR, http://www.openbioinformatics.org/annovar/ by all distinct lead SNPs from the meta-analysis using the follow- PhenoScanner, http://www.phenoscanner.medschl.cam.ac.u ing formula: k/phenoscanner Power calculations: http://csg.sph.umich.edu/abecasis/cats/ n βi2 · 2 · EAFi · (1 − EAFi )    gas_power_calculator/index.html σ 2 residuals ln (IL6) i=1 The UK Biobank Neale’s lab rapid GWAS: (http://www.neale where EAF is the effect allele frequency, βi the effect size of lab.is/blog/2017/7/19/rapid-gwas-of-thousands-of-phenotypes- the individual variants, and n is the total number of lead vari- for-337000-samples-in-the-uk-biobank). ants. The current formula may overestimate the variance to a small extent as some level of SNP correlation was existent (LD r2 < 0.25). The variance of the residuals of ln (IL-6) was calculated Authors Contributions using data from the NESDA cohort (n = 2517). The total common The guarantors of the study are T.S.A., B.P. and B.Z.A., from SNP heritability of serum IL-6 levels explained by all GWAS conception and design to conduct of the study and acquisition of 406 Human Molecular Genetics, 2021, Vol. 30, No. 5 data, data analysis and interpretation of data. T.S.A. and B.P. have production and decreases OPG expression by osteoblasts. written the first draft of the manuscript. All co-authors have Osteoarthr. Cartil., 17, 473–481. provided important intellectual input and contributed consid- 11. Haddy, N., Sass, C., Maumus, S., Marie, B., Droesch, S., Siest, erably to the analyses and interpretation of the data. All authors G., Lambert, D. and Visvikis, S. (2005) Biological variations, guarantee that the accuracy and integrity of any part of the work genetic polymorphisms and familial resemblance of TNF- have been appropriately investigated and resolved and all have alpha and IL-6 concentrations: STANISLAS cohort. Eur. J. approved the final version of the manuscript. BP had full access Hum. Genet., 13, 109–117. to the data and the corresponding authors T.S.A. and B.Z.A. had 12. Madhok, R., Crilly, A., Watson, J. and Capell, H.A. (1993) Serum the final responsibility for the decision to submit for publication. interleukin 6 levels in rheumatoid arthritis: correlations with clinical and laboratory indices of disease activity. Ann. Rheum. Dis., 52, 232–234. Acknowledgments 13. de Benedetti, F., Massa, M., Robbioni, P., Ravelli, A., Burgio, We acknowledge the participants of this study and the funding G.R. and Martini, A. (1991) Correlation of serum interleukin- resources that have contributed to achieving this effort. Detailed 6 levels with joint involvement and thrombocytosis in sys- acknowledgments can be found in Supplementary Material, temic juvenile rheumatoid arthritis. Arthritis Rheum., 34, Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 Table S12. 1158–1163. Conflict of Interest. There is no conflict of interest from any of 14. Ahluwalia, T.S., Kilpelainen, T.O., Singh, S. and Rossing, P. the participating co-authors. T.S.A. is a shareholder with Novo (2019) Editorial: novel biomarkers for type 2 diabetes. Front Nordisk A/S and Zealand Pharma A/S. Endocrinol (Lausanne), 10, 649. 15. Cesari, M., Penninx, B.W., Newman, A.B., Kritchevsky, S.B., Nicklas, B.J., Sutton-Tyrrell, K., Rubin, S.M., Ding, J., Simonsick, E.M., Harris, T.B. and Pahor, M. (2003) Funding Inflammatory markers and onset of cardiovascular Novo Nordisk Foundation (NNF18OC0052457 to T.S.A.). Detailed events: results from the health ABC study. Circulation, 108, Funding statements from participating study cohorts can be 2317–2322. found in Supplementary Material, Table S12. 16. Qu, D., Liu, J., Lau, C.W. and Huang, Y. (2014) IL-6 in dia- betes and cardiovascular complications. Br. J. Pharmacol., 171, 3595–3603. References 17. Mauer, J., Denson, J.L. and Bruning, J.C. (2015) Versatile func- 1. Kishimoto, T. (2010) IL-6: from its discovery to clinical appli- tions for IL-6 in metabolism and cancer. Trends Immunol., 36, cations. Int. Immunol., 22, 347–352. 92–101. 2. Nishimoto, N. and Kishimoto, T. (2006) Interleukin 6: from 18. Mucha, S., Baurecht, H., Novak, N., Rodríguez, E., Bej, S., bench to bedside. Nat. Clin. Pract. Rheumatol., 2, 619–626. Mayr, G., Emmert, H., Stölzl, D., Gerdes, S., Jung, E.S. et al. 3. Brocker, C., Thompson, D., Matsumoto, A., Nebert, D.W. and (2020) Protein-coding variants contribute to the risk of atopic Vasiliou, V. (2010) Evolutionary divergence and functions of dermatitis and skin-specific gene expression. J. Allergy Clin. the human interleukin (IL) gene family. Hum. Genomics, 5, Immunol., 145, 1208–1218. 30–55. 19. Zhang, C., Wu, Z., Zhao, G., Wang, F. and Fang, Y. (2016) Iden- 4. Gelinas, L., Falkenham, A., Oxner, A., Sopel, M. and Legare, J.F. tification of IL6 as a susceptibility gene for major depressive (2011) Highly purified human peripheral blood monocytes disorder. Sci. Rep., 6, 31264. produce IL-6 but not TNFalpha in response to angiotensin II. 20. Calabrese, L.H. and Rose-John, S. (2014) IL-6 biology: impli- J. Renin-Angiotensin-Aldosterone Syst., 12, 295–303. cations for clinical targeting in rheumatic disease. Nat. Rev. 5. Kitani, A., Hara, M., Hirose, T., Harigai, M., Suzuki, K., Rheumatol., 10, 720–727. Kawakami, M., Kawaguchi, Y., Hidaka, T., Kawagoe, M. and 21. Scheinecker, C., Smolen, J., Yasothan, U., Stoll, J. and Kirk- Nakamura, H. (1992) Autostimulatory effects of IL-6 on patrick, P. (2009) Tocilizumab. Nat. Rev. Drug Discov., 8, excessive B cell differentiation in patients with systemic 273–274. lupus erythematosus: analysis of IL-6 production and IL-6R 22. Revez, J.A., Bain, L.M., Watson, R.M., Towers, M., Collins, T., expression. Clin. Exp. Immunol., 88, 75–83. Killian, K.J., O’Byrne, P.M., Gauvreau, G.M., Upham, J.W. and 6. Li, T. and He, S. (2006) Induction of IL-6 release from human Ferreira, M.A. (2019) Effects of interleukin-6 receptor block- T cells by PAR-1 and PAR-2 agonists. Immunol. Cell Biol., 84, ade on allergen-induced airway responses in mild asthmat- 461–466. ics. Clin. Transl. Immunol., 8, e1044. 7. Ng, E.K., Panesar, N., Longo, W.E., Shapiro, M.J., Kaminski, 23. Yazici, Y., Curtis, J.R., Ince, A., Baraf, H., Malamet, R.L., Teng, D.L., Tolman, K.C. and Mazuski, J.E. (2003) Human intestinal L.L. and Kavanaugh, A. (2012) Efficacy of tocilizumab in epithelial and smooth muscle cells are potent producers of patients with moderate to severe active rheumatoid arthritis IL-6. Mediat. Inflamm., 12, 3–8. and a previous inadequate response to disease-modifying 8. Fain, J.N. (2006) Release of interleukins and other inflam- antirheumatic drugs: the ROSE study. Ann. Rheum. Dis., 71, matory cytokines by human adipose tissue is enhanced in 198–205. obesity and primarily due to the nonfat cells. Vitam. Horm., 24. Yokota, S., Imagawa, T., Mori, M., Miyamae, T., Aihara, Y., 74, 443–477. Takei, S., Iwata, N., Umebayashi, H., Murata, T., Miyoshi, 9. Podor, T.J., Jirik, F.R., Loskutoff, D.J., Carson, D.A. and Lotz, M. et al. (2008) Efficacy and safety of tocilizumab in M. (1989) Human endothelial cells produce IL-6. Lack of patients with systemic-onset juvenile idiopathic arthritis: responses to exogenous IL-6. Ann. N. Y. Acad. Sci., 557, 374–385 a randomised, double-blind, placebo-controlled, withdrawal discussion 386-377. phase III trial. Lancet, 371, 998–1006. 10. Sanchez, C., Gabay, O., Salvat, C., Henrotin, Y.E. and Beren- 25. Gupta, S., Wang, W., Hayek, S.S., Chan, L., Mathews, K.S., baum, F. (2009) Mechanical loading highly increases IL-6 Melamed, M.L., Brenner, S.K., Leonberg-Yoo, A., Schenck, E.J., Human Molecular Genetics, 2021, Vol. 30, No. 5 407 Radbel, J. et al. (2021) Association between early treatment in a genome-wide association study. Genet. Epidemiol., 35, with tocilizumab and mortality among critically ill patients 447–456. with COVID-19. JAMA Intern. Med., 181, 41–51. 39. Rentzsch, P., Witten, D., Cooper, G.M., Shendure, J. and 26. Worns, M.A., Victor, A., Galle, P.R. and Hohler, T. (2006) Kircher, M. (2019) CADD: predicting the deleteriousness of Genetic and environmental contributions to plasma C- variants throughout the human genome. Nucleic Acids Res., reactive protein and interleukin-6 levels–a study in twins. 47, D886–D894. Genes Immun., 7, 600–605. 40. Sabater-Lleal, M., Huang, J., Chasman, D., Naitza, S., Dehghan, 27. Sas, A.A., Jamshidi, Y., Zheng, D., Wu, T., Korf, J., Alizadeh, A., Johnson, A.D., Teumer, A., Reiner, A.P., Folkersen, L., B.Z., Spector, T.D. and Snieder, H. (2012) The age-dependency Basu, S. et al. (2013) Multiethnic meta-analysis of genome- of genetic and environmental influences on serum cytokine wide association studies in >100 000 subjects identifies 23 levels: a twin study. Cytokine, 60, 108–113. fibrinogen-associated loci but no strong evidence of a causal 28. Neijts, M., van Dongen, J., Kluft, C., Boomsma, D.I., Willemsen, association between circulating fibrinogen and cardiovascu- G. and de Geus, E.J. (2013) Genetic architecture of the pro- lar disease. Circulation, 128, 1310–1324. inflammatory state in an extended twin-family design. Twin 41. Dehghan, A., Dupuis, J., Barbalic, M., Bis, J.C., Eiriksdottir, G., Res. Hum. Genet., 16, 931–940. Lu, C., Pellikka, N., Wallaschofski, H., Kettunen, J., Henneman, Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 29. Amaral, W.Z., Krueger, R.F., Ryff, C.D. and Coe, C.L. (2015) P. et al. (2011) Meta-analysis of genome-wide association Genetic and environmental determinants of population studies in >80 000 subjects identifies multiple loci for C- variation in interleukin-6, its soluble receptor and C-reactive reactive protein levels. Circulation, 123, 731–738. protein: insights from identical and fraternal twins. Brain 42. Okada, Y., Wu, D., Trynka, G., Raj, T., Terao, C., Ikari, K., Behav. Immun., 49, 171–181. Kochi, Y., Ohmura, K., Suzuki, A., Yoshida, S. et al. (2014) 30. Melzer, D., Perry, J.R., Hernandez, D., Corsi, A.M., Stevens, Genetics of rheumatoid arthritis contributes to biology and K., Rafferty, I., Lauretani, F., Murray, A., Gibbs, J.R., Paolisso, drug discovery. Nature, 506, 376–381. G. et al. (2008) A genome-wide association study identi- 43. Yu, X.Q., Li, M., Zhang, H., Low, H.Q., Wei, X., Wang, J.Q., fies protein quantitative trait loci (pQTLs). PLoS Genet., 4, Sun, L.D., Sim, K.S., Li, Y., Foo, J.N. et al. (2011) A genome- e1000072. wide association study in Han Chinese identifies multi- 31. Naitza, S., Porcu, E., Steri, M., Taub, D.D., Mulas, A., Xiao, X., ple susceptibility loci for IgA nephropathy. Nat. Genet., 44, Strait, J., Dei, M., Lai, S., Busonero, F. et al. (2012) A genome- 178–182. wide association scan on the levels of markers of inflam- 44. Liu, J.Z., van Sommeren, S., Huang, H., Ng, S.C., Alberts, R., mation in Sardinians reveals associations that underpin its Takahashi, A., Ripke, S., Lee, J.C., Jostins, L., Shah, T. et al. complex regulation. PLoS Genet., 8, e1002480. (2015) Association analyses identify 38 susceptibility loci for 32. Comuzzie, A.G., Cole, S.A., Laston, S.L., Voruganti, V.S., Haack, inflammatory bowel disease and highlight shared genetic K., Gibbs, R.A. and Butte, N.F. (2012) Novel genetic loci iden- risk across populations. Nat. Genet., 47, 979–986. tified for the pathophysiology of childhood obesity in the 45. Elliott, P., Chambers, J.C., Zhang, W., Clarke, R., Hopewell, Hispanic population. PLoS One, 7, e51954. J.C., Peden, J.F., Erdmann, J., Braund, P., Engert, J.C., Bennett, 33. Shah, T., Zabaneh, D., Gaunt, T., Swerdlow, D.I., Shah, S., Tal- D. et al. (2009) Genetic loci associated with C-reactive pro- mud, P.J., Day, I.N., Whittaker, J., Holmes, M.V., Sofat, R. et al. tein levels and risk of coronary heart disease. JAMA, 302, (2013) Gene-centric analysis identifies variants associated 37–48. with interleukin-6 levels and shared pathways with other 46. Ferreira, R.C., Freitag, D.F., Cutler, A.J., Howson, J.M., Rainbow, inflammation markers. Circ. Cardiovasc. Genet., 6, 163–170. D.B., Smyth, D.J., Kaptoge, S., Clarke, P., Boreham, C., Coulson, 34. Sidore, C., Busonero, F., Maschio, A., Porcu, E., Naitza, S., R.M. et al. (2013) Functional IL6R 358Ala allele impairs clas- Zoledziewska, M., Mulas, A., Pistis, G., Steri, M., Danjou, sical IL-6 receptor signaling and influences risk of diverse F. et al. (2015) Genome sequencing elucidates Sardinian inflammatory diseases. PLoS Genet., 9, e1003444. genetic architecture and augments association analyses 47. Stone, K., Woods, E., Szmania, S.M., Stephens, O.W., Garg, for lipid and blood inflammatory markers. Nat. Genet., 47, T.K., Barlogie, B., Shaughnessy, J.D., Jr., Hall, B., Reddy, M., 1272–1281. Hoering, A. et al. (2013) Interleukin-6 receptor polymorphism 35. Denny, J.C., Ritchie, M.D., Basford, M.A., Pulley, J.M., Bas- is prevalent in HIV-negative Castleman disease and is asso- tarache, L., Brown-Gentry, K., Wang, D., Masys, D.R., Roden, ciated with increased soluble interleukin-6 receptor levels. D.M. and Crawford, D.C. (2010) PheWAS: demonstrating the PLoS One, 8, e54610. feasibility of a phenome-wide scan to discover gene-disease 48. Rafiq, S., Frayling, T.M., Murray, A., Hurst, A., Stevens, associations. Bioinformatics, 26, 1205–1210. K., Weedon, M.N., Henley, W., Ferrucci, L., Bandinelli, S., 36. Pendergrass, S.A., Brown-Gentry, K., Dudek, S.M., Torsten- Corsi, A.M. et al. (2007) A common variant of the inter- son, E.S., Ambite, J.L., Avery, C.L., Buyske, S., Cai, C., Fesin- leukin 6 receptor (IL-6r) gene increases IL-6r and IL-6 lev- meyer, M.D., Haiman, C. et al. (2011) The use of phenome- els, without other inflammatory effects. Genes Immun., 8, wide association studies (Phe WAS) for exploration of novel 552–559. genotype-phenotype relationships and pleiotropy discovery. 49. Galicia, J.C., Tai, H., Komatsu, Y., Shimada, Y., Akazawa, K. and Genet. Epidemiol., 35, 410–422. Yoshie, H. (2004) Polymorphisms in the IL-6 receptor (IL-6R) 37. Pendergrass, S.A., Brown-Gentry, K., Dudek, S., Frase, A., gene: strong evidence that serum levels of soluble IL-6R are Torstenson, E.S., Goodloe, R., Ambite, J.L., Avery, C.L., Buyske, genetically influenced. Genes Immun., 5, 513–516. S., Bůžková, P. et al. (2013) Phenome-wide association study 50. Hayashida, K., Bartlett, A.H., Chen, Y. and Park, P.W. (2010) (Phe WAS) for detection of pleiotropy within the popula- Molecular and cellular mechanisms of ectodomain shed- tion architecture using genomics and epidemiology (PAGE) ding. Anat Rec (Hoboken), 293, 925–937. network. PLoS Genet., 9, e1003087. 51. Collaboration, I.R.G.C.E.R.F, Sarwar, N., Butterworth, A.S., 38. So, H.C., Li, M. and Sham, P.C. (2011) Uncovering the total Freitag, D.F., Gregson, J., Willeit, P., Gorman, D.N., Gao, P., heritability explained by all true susceptibility variants Saleheen, D., Rendon, A. et al. (2012) Interleukin-6 receptor 408 Human Molecular Genetics, 2021, Vol. 30, No. 5 pathways in coronary heart disease: a collaborative meta- 64. Ellingsgaard, H., Hauselmann, I., Schuler, B., Habib, A.M., Bag- analysis of 82 studies. Lancet, 379, 1205–1213. gio, L.L., Meier, D.T., Eppler, E., Bouzakri, K., Wueest, S., Muller, 52. Interleukin-6 Receptor Mendelian Randomisation Analysis, Y.D. et al. (2011) Interleukin-6 enhances insulin secretion C, Swerdlow, D.I., Holmes, M.V., Kuchenbaecker, K.B., Eng- by increasing glucagon-like peptide-1 secretion from L cells mann, J.E., Shah, T., Sofat, R., Guo, Y., Chung, C., Peasey, A. and alpha cells. Nat. Med., 17, 1481–1489. et al. (2012) The interleukin-6 receptor as a target for preven- 65. van de Veerdonk, F.L., Stoeckman, A.K., Wu, G., Boeckermann, tion of coronary heart disease: a mendelian randomisation A.N., Azam, T., Netea, M.G., Joosten, L.A., van der Meer, J.W., analysis. Lancet, 379, 1214–1224. Hao, R., Kalabokis, V. et al. (2012) IL-38 binds to the IL-36 53. Consortium, C.A.D., Deloukas, P., Kanoni, S., Willenborg, C., receptor and has biological effects on immune cells similar Farrall, M., Assimes, T.L., Thompson, J.R., Ingelsson, E., Sale- to IL-36 receptor antagonist. Proc. Natl. Acad. Sci. USA., 109, heen, D., Erdmann, J. et al. (2013) Large-scale association 3001–3005. analysis identifies new risk loci for coronary artery disease. 66. Reiner, A.P., Wurfel, M.M., Lange, L.A., Carlson, C.S., Nord, Nat. Genet., 45, 25–33. A.S., Carty, C.L., Rieder, M.J., Desmarais, C., Jenny, N.S., Iribar- 54. Schnabel, R.B., Kerr, K.F., Lubitz, S.A., Alkylbekova, E.L., Mar- ren, C. et al. (2008) Polymorphisms of the IL1-receptor antag- cus, G.M., Sinner, M.F., Magnani, J.W., Wolf, P.A., Deo, R., onist gene (IL1RN) are associated with multiple markers of Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 Lloyd-Jones, D.M. et al. (2011) Large-scale candidate gene systemic inflammation. Arterioscler. Thromb. Vasc. Biol., 28, analysis in whites and African Americans identifies IL6R 1407–1412. polymorphism in relation to atrial fibrillation: the National 67. Hom, G., Graham, R.R., Modrek, B., Taylor, K.E., Ortmann, W., Heart, Lung, and Blood Institute’s candidate Gene Asso- Garnier, S., Lee, A.T., Chung, S.A., Ferreira, R.C., Pant, P.V. et al. ciation resource (CARe) project. Circ. Cardiovasc. Genet., 4, (2008) Association of systemic lupus erythematosus with 557–564. C8orf13-BLK and ITGAM-ITGAX. N. Engl. J. Med., 358, 900–909. 55. Hawkins, G.A., Robinson, M.B., Hastie, A.T., Li, X., Li, H., Moore, 68. Jostins, L., Ripke, S., Weersma, R.K., Duerr, R.H., McGovern, W.C., Howard, T.D., Busse, W.W., Erzurum, S.C., Wenzel, S.E. D.P., Hui, K.Y., Lee, J.C., Schumm, L.P., Sharma, Y., Anderson, et al. (2012) The IL6R variation asp (358) ala is a potential C.A. et al. (2012) Host-microbe interactions have shaped the modifier of lung function in subjects with asthma. J. Allergy genetic architecture of inflammatory bowel disease. Nature, Clin. Immunol., 130, 510, e511–515. 491, 119–124. 56. Harrison, S.C., Smith, A.J., Jones, G.T., Swerdlow, D.I., Ram- 69. Ahluwalia, T.S., Schulz, C.A., Waage, J., Skaaby, T., Sandholm, puri, R., Bown, M.J., Aneurysm, C., Folkersen, L., Baas, N., van Zuydam, N., Charmet, R., Bork-Jensen, J., Almgren, P., A.F., de Borst, G.J. et al. (2013) Interleukin-6 receptor path- Thuesen, B.H. et al. (2019) A novel rare CUBN variant and ways in abdominal aortic aneurysm. Eur. Heart J., 34, three additional genes identified in Europeans with and 3707–3716. without diabetes: results from an exome-wide association 57. Eyre, S., Bowes, J., Diogo, D., Lee, A., Barton, A., Martin, P., study of albuminuria. Diabetologia, 62, 292–305. Zhernakova, A., Stahl, E., Viatte, S., McAllister, K. et al. (2012) 70. Apelqvist, J. (2012) Diagnostics and treatment of the diabetic High-density genetic mapping identifies new susceptibility foot. Endocrine, 41, 384–397. loci for rheumatoid arthritis. Nat. Genet., 44, 1336–1340. 71. Pickrell, J.K., Berisa, T., Liu, J.Z., Segurel, L., Tung, J.Y. and 58. Scheller, J., Ohnesorge, N. and Rose-John, S. (2006) Hinds, D.A. (2016) Detection and interpretation of shared Interleukin-6 trans-signalling in chronic inflammation genetic influences on 42 human traits. Nat. Genet., 48, and cancer. Scand. J. Immunol., 63, 321–329. 709–717. 59. Chalaris, A., Schmidt-Arras, D., Yamamoto, K. and Rose-John, 72. Mac Rae, C.A. and Vasan, R.S. (2011) Next-generation S. (2012) Interleukin-6 trans-signaling and colonic cancer genome-wide association studies: time to focus on pheno- associated with inflammatory bowel disease. Dig. Dis., 30, type? Circ. Cardiovasc. Genet., 4, 334–336. 492–499. 73. Dumitrescu, L., Carty, C.L., Franceschini, N., Hindorff, L.A., 60. Enevold, C., Baslund, B., Linde, L., Josephsen, N.L., Tarp, Cole, S.A., Buzkova, P., Schumacher, F.R., Eaton, C.B., Goodloe, U., Lindegaard, H., Jacobsen, S. and Nielsen, C.H. (2014) R.J., Duggan, D.J. et al. (2013) Post-genome-wide association Interleukin-6-receptor polymorphisms rs12083537, study challenges for lipid traits: describing age as a modifier rs2228145, and rs4329505 as predictors of response of gene-lipid associations in the population architecture to tocilizumab in rheumatoid arthritis. Pharmacogenet. using genomics and epidemiology (PAGE) study. Ann. Hum. Genomics, 24, 401–405. Genet., 77, 416–425. 61. Ligthart, S., Vaez, A., Vosa, U., Stathopoulou, M.G., de Vries, 74. Kulminski, A.M., Culminskaya, I., Arbeev, K.G., Arbeeva, L., P.S., Prins, B.P., Van der Most, P.J., Tanaka, T., Naderi, E., Rose, Ukraintseva, S.V., Stallard, E., Wu, D. and Yashin, A.I. (2015) L.M. et al. (2018) Genome analyses of >200,000 individuals Birth cohort, age, and sex strongly modulate effects of lipid identify 58 loci for chronic inflammation and highlight path- risk alleles identified in genome-wide association studies. ways that link inflammation and complex disorders. Am. J. PLoS One, 10, e0136319. Hum. Genet., 103, 691–706. 75. Marchini, J., Howie, B., Myers, S., McVean, G. and Don- 62. Reiner, A.P., Beleza, S., Franceschini, N., Auer, P.L., Robinson, nelly, P. (2007) A new multipoint method for genome-wide J.G., Kooperberg, C., Peters, U. and Tang, H. (2012) Genome- association studies by imputation of genotypes. Nat. Genet., wide association and population genetic analysis of 39, 906–913. C-reactive protein in African American and Hispanic 76. Howie, B., Fuchsberger, C., Stephens, M., Marchini, J. and American women. Am. J. Hum. Genet., 91, 502–512. Abecasis, G.R. (2012) Fast and accurate genotype imputation 63. Herder, C., Nuotio, M.L., Shah, S., Blankenberg, S., Brunner, in genome-wide association studies through pre-phasing. E.J., Carstensen, M., Gieger, C., Grallert, H., Jula, A., Kaho- Nat. Genet., 44, 955–959. nen, M. et al. (2014) Genetic determinants of circulating 77. Servin, B. and Stephens, M. (2007) Imputation-based analysis interleukin-1 receptor antagonist levels and their associa- of association studies: candidate regions and quantitative tion with glycemic traits. Diabetes, 63, 4343–4359. traits. PLoS Genet., 3, e114. Human Molecular Genetics, 2021, Vol. 30, No. 5 409 78. van der Most, P.J., Vaez, A., Prins, B.P., Munoz, M.L., Snieder, 84. Purcell, S., Neale, B., Todd-Brown, K., Thomas, L., Ferreira, H., Alizadeh, B.Z. and Nolte, I.M. (2014) QCGWAS: a flexible M.A., Bender, D., Maller, J., Sklar, P., de Bakker, P.I., Daly, M.J. R package for automated quality control of genome-wide et al. (2007) PLINK: a tool set for whole-genome association association results. Bioinformatics, 30, 1185–1186. and population-based linkage analyses. Am. J. Hum. Genet., 79. Magi, R. and Morris, A.P. (2010) GWAMA: software for 81, 559–575. genome-wide association meta-analysis. BMC Bioinformatics, 85. Hinrichs, A.S., Karolchik, D., Baertsch, R., Barber, G.P., Bejer- 11, 288. ano, G., Clawson, H., Diekhans, M., Furey, T.S., Harte, R.A., 80. Pei, Y.F., Zhang, L., Li, J. and Deng, H.W. (2010) Analyses and Hsu, F. et al. (2006) The UCSC genome browser database: comparison of imputation-based association methods. PLoS update 2006. Nucleic Acids Res., 34, D590–D598. One, 5, e10827. 86. Danecek, P., Auton, A., Abecasis, G., Albers, C.A., Banks, 81. Willer, C.J., Li, Y. and Abecasis, G.R. (2010) METAL: fast and E., DePristo, M.A., Handsaker, R.E., Lunter, G., Marth, G.T., efficient meta-analysis of genomewide association scans. Sherry, S.T. et al. (2011) The variant call format and VCFtools. Bioinformatics, 26, 2190–2191. Bioinformatics, 27, 2156–2158. 82. Yang, J., Benyamin, B., McEvoy, B.P., Gordon, S., Henders, 87. Wang, K., Li, M. and Hakonarson, H. (2010) ANNOVAR: func- A.K., Nyholt, D.R., Madden, P.A., Heath, A.C., Martin, N.G., tional annotation of genetic variants from high-throughput Downloaded from https://academic.oup.com/hmg/article/30/5/393/6124523 by guest on 15 March 2023 Montgomery, G.W. et al. (2010) Common SNPs explain a large sequencing data. Nucleic Acids Res., 38, e164. proportion of the heritability for human height. Nat. Genet., 88. Pruitt, K.D., Tatusova, T. and Maglott, D.R. (2005) NCBI refer- 42, 565–569. ence sequence (RefSeq): a curated non-redundant sequence 83. Penninx, B.W., Beekman, A.T., Smit, J.H., Zitman, F.G., database of genomes, transcripts and proteins. Nucleic Acids Nolen, W.A., Spinhoven, P., Cuijpers, P., De Jong, P.J., Van Res., 33, D501–D504. Marwijk, H.W., Assendelft, W.J. et al. (2008) The Nether- 89. Staley, J.R., Blackshaw, J., Kamat, M.A., Ellis, S., Surendran, lands study of depression and anxiety (NESDA): rationale, P., Sun, B.B., Paul, D.S., Freitag, D., Burgess, S., Danesh, J. objectives and methods. Int. J. Methods Psychiatr. Res., 17, et al. (2016) Pheno scanner: a database of human genotype- 121–140. phenotype associations. Bioinformatics, 32, 3207–3209.