Complement gene variants in relation to autoantibodies to beta cell specific antigens and type 1 diabetes in the TEDDY Study

1 Scien i ic RepoR s | 6:27887 | DOI: 10.1038/s ep27887 www.na u e.com/scien i ic epo s Complemen gene a ian s in ela ion o au oan ibodies o be a cell speci ic an igens and ype 1 diabe es in he TEDDY S udy Ca ina Tö n1, Xiang Liu2, William Hagopian3, Åke Le nma k1, Olli Simell4, Ma ian Rewe s5, Ane e-G Ziegle 6, Desmond Scha z7, Beena Akolka 8, Suna Onengu -Gumuscu9, Wei-Min Chen9, Jo ma Toppa i4,10, Juha Mykkänen4,10, Jo ma Ilonen10, S ephen S Rich9, Jin-Xiong She11, Ashok Sha ma11, And ea S eck5, Je ey K ische 2 & The TEDDY S udy G oup† A o al o 15 SNPs wi hin complemen genes and p esen on he ImmunoChip we e analyzed in The En i onmen al De e minan s o Diabe es in he Young (TEDDY) s udy. A o al o 5474 subjec s we e ollowed om h ee mon hs o age un il isle au oimmuni y (IA: n = 413) and he subsequen onse o ype 1 diabe es (n = 115) o a median o 73 mon hs (IQR 54–91). Th ee SNPs wi hin ITGAM we e nominally associa ed (p < 0.05) wi h IA: s1143678 [Haza d a io; HR 0.80; 95% CI 0.66–0.98; p = 0.032], s1143683 [HR 0.80; 95% CI 0.65–0.98; p = 0.030] and s4597342 [HR 1.16; 95% CI 1.01–1.32; p = 0.041]. When ype 1 diabe es was he ou come, in DR3/4 subjec s, he e was nominal signi icance o wo SNPs: s17615 in CD21 [HR 1.52; 95% CI 1.05–2.20; p = 0.025] and s4844573 in C4BPA [HR 0.63; 95% CI 0.43–0.92; p = 0.017]. Among DR4/4 subjec s, s2230199 in C3 was signi ican ly associa ed [HR 3.20; 95% CI 1.75–5.85; p = 0.0002, unco ec ed] a signi icance ha wi hs ood Bon e oni co ec ion since i was less han 0.000833 (0.05/60) in he HLA-speci ic analyses. SNPs wi hin he complemen genes may con ibu e o IA, he i s s ep o ype 1 diabe es, wi h a leas one SNP in C3 signi ican ly associa ed wi h clinically diagnosed ype 1 diabe es. Cu en ly, au oan ibodies a e he bes ma ke s o sc eening pu poses in la ge popula ions o an au oimmune a ack di ec ed owa ds he panc ea ic isle be a cells. Al hough he a ack is hough o be p ima ily ca ied ou by cy o oxic T cells ha ecognize au oan igen pep ides on be a-cell HLA Class I p o eins i canno be excluded ha o he componen s o he inna e immune sys em also con ibu es. Complemen ac o s a e p o eins in ol ed in he inna e immune sys em assis ing in opsonizing o eign an igens o inc ease he immune a ack owa ds bac e ia and i uses. O he unc ions o he complemen sys em include chemo axis whe e mac ophages and neu ophils a e a ac ed o he si e o in ec ion, cell lysis which is a up u e o o eign cells and agglu ina ion whe e pa hogens a e clus e ed oge he . Complemen ac o s a e also linked o he adap i e immune sys em since an ibodies o m complexes wi h complemen s. Gene ic a ian s o genes encoding complemen ac o s ha e shown associa ions wi h au oimmune diseases, in pa icula o ype 1 diabe es, sys emic lupus e y hema osus (SLE) and au oimmune macula degene a ion (MAD). In ype 1 diabe es, s udies ha e shown an inc eased deposi ion o C4 in panc e- a a1, as well as inc eased le els o plasma p o ein inhibi o (C1-inhibi o )2. Deposi ion o C5b-9 ( he memb ane 1Depa men o Clinical Sciences, Lund Uni e si y/CRC, Malmö, Sweden. 2Heal h In o ma ics Ins i u e, Mo sani College o Medicine, Uni e si y o Sou h Flo ida, Tampa, FL, USA. 3Paci ic No hwes Diabe es Resea ch Ins i u e, Sea le, WA, USA. 4Depa men o Pedia ics, Tu ku Uni e si y Hospi al, Tu ku, Finland. 5Ba ba a Da is Cen e o Childhood Diabe es, Uni e si y o Colo ado, Au o a, CO, USA. 6Ins i u e o Diabe es Resea ch, Helmhol z Zen um, München, and Klinikum ech s de Isa , Technische Uni e si ä München, and Fo sche g uppe Diabe es e. V., Neuhe be g, Ge many. 7Depa men o Pedia ics, Uni e si y o Flo ida, Gaines ille, FL, USA. 8Na ional Ins i u es o Diabe es & Diges i e & Kidney Diso de s, Be hesda, MD, USA. 9Cen e o Public Heal h Genomic, Uni e si y o Vi ginia, Cha lo es ille, VA, USA. 10Depa men s o Physiology and Pedia ics, Uni e si y o Tu ku, Tu ku, Finland. 11Cen e o Bio echnology and Genomic Medicine, Medical College o Geo gia, Augus a Uni e si y, Augus a, GA, USA. †A comp ehensi e lis o conso ium membe s appea s a he end o he pape . Co espondence and eques s o ma e ials should be add essed o C.T. (email: Ca ina.T[email p o ec ed]) Recei ed: 20 Janua y 2016 Accep ed: 06 May 2016 Published: 16 June 2016 OPEN www.na u e.com/scien i ic epo s/ 2 Scien i ic RepoR s | 6:27887 | DOI: 10.1038/s ep27887 a ack complex (MAC) o e minal complemen complex (TCC)) has been iden i ied in neu ons o hose wi h ype 1 diabe es who la e died om ke oacidosis3. Inc eased inna e immune eac i i y has been de ec ed be o e diabe es-associa ed se ocon e sion in child en wi h high- isk HLA geno ypes, o example an inc eased exp es- sion o C2 and C4 binding p o ein alpha-chain (C4BPA) one o wo yea s p io o diagnosis o ype 1 diabe es4. Human se um may be oxic o be a-cells by ac i a ing he al e na i e complemen pa hway5. Fou genes encoding complemen p o eins ( ac o B, C2, C4A and C4B) a e loca ed on ch omosome 6 in he human majo his ocompa ibili y complex (MHC) ha also con ains he ype 1 diabe es associa ed HLA-genes. The complemen genes (CFB, C2, C4A, C4B) a e loca ed elome ic be ween he HLA class II and class I egions. Gi en he ex ensi e linkage disequilib ium in he MHC, hese complemen genes a e inhe i ed as a ixed se (a complo ype) and ha e been included in he HLA clus e as an ex ended HLA haplo ype. A leas 15 di e - en complo ypes ha e been iden i ied wi h equency g ea e han 1% wi hin Caucasian popula ions6. Ce ain haplo ypes wi hin he HLA- complemen sys em a e associa ed wi h ype 1 diabe es, pa icula ly hose ha include HLA-DR3-DQ2. The ex ended haplo ypes o HLA and complemen (HLA-A1-B8-B S- C4AQ0-DR3, HLAB18-B F1-C4A3-C4BQ0-DR3-DQ2 and HLA-A2-CW3-BW62-B S-C4A3-DR4-DQ8) a e inc eased among ype 1 diabe es cases compa ed wi h con ols7. These h ee haplo ypes ha e been eplica ed many imes as well as h ee addi ional haplo ypes exhibi ing associa ion wi h ype 1 diabe es. The B7-B S-DR2 haplo ype has nega i e associa ion wi h ype 1 diabe es (mo e common in con ols han in cases), while i e a e isk haplo ypes o ype 1 diabe es, mo e common in pa ien s han in con ols (B8-B S-DR3, B8-B S-DR4, B15-B S-DR4, B18-B F1-DR3 and B40-B S-DR4)8. Whe he hese complemen gene a ian s a e hemsel es a ec ing suscep ibili y o ype 1 diabe es o a e in linkage disequlib ium wi h o he isk con ibu ing genes (including ce ain HLA class I alleles) emains un esol ed9–11. The e a e h ee di e en pa hways ha ac i a e complemen : he classical pa hway (immunoglobulin bound o an igen), he mannose binding lec in pa hway and he al e na i e pa hway. The ini ial enzyme in he classical pa hway is C1 (C1q/C1 /C1s). C1q can bind o IgG molecules leading o ansac i a ion o C1q. C1s in u n clea es he p o eins C4 and C2 in o C4a and C2b. The o ma ion o C4b2a ac s as he C3 con e ase ac i a ing enzyme in he classical pa hway. The mannose binding lec in pa hway (mannose binding o lec in/ icolin bound o ca bohyd a e on pa hogen), s a s wi h ecogni ion ia mannose-binding lec in (MBL), icolin 1–3 and CL 11, he p ocess con inues wi h in ol emen o mannose-binding lec in associa ed se ine p o eases (MASP-1, MASP-2 and MASP3) leading o clea age and ac i a ion o C4, C2 and C3. The al e na i e pa hway is ini ia ed by p esence o pa hogen leading o a spon aneous ac i a ion o C3. The h ee pa hways esul s in he o ma ion o C3 con e - ases ha clea e C3 o he C3b and C3a o ms. The ea e C5 is ac i a ed and oge he wi h C6–9 o m he C5b9 (MAC)12,13. I has been shown ha se e al Ig classes (IgG, IgM and IgA) can be in ol ed in he h ee pa hways13. App oxima ely, 50 SNPs wi hin complemen genes ha e been associa ed wi h di e en au oimmune diseases. On he ImmunoChip, a cus om geno yping a ay based upon signi ican genome-wide associa ion wi h au o- immune loci wi h au oimmune diseases, he e a e 17 complemen SNPs ha can be in e oga ed in he TEDDY s udy. The En i onmen al De e minan s o Diabe es in he Young (TEDDY) is an in e na ional p ospec i e s udy o newbo n child en wi h an inc eased gene ic isk o ype 1 diabe es. Child en om he gene al popula ion we e en olled in o TEDDY based on hei inc eased gene ic isk o ype 1 diabe es based upon any o he ou high- isk HLA-geno ypes (DR3/4, DR4/4, DR4/8 o DR3/3). The goal o TEDDY is o iden i y en i onmen al ac o s ha p ecipi a e isle au oimmuni y (IA) and subsequen ype 1 diabe es in gene ically p edisposed chil- d en. Type 1 diabe es de elops in a leas wo phases. A hallma k o he i s phase, is an unknown e en o agen ha may igge he immune sys em o p oduce au oan ibodies agains ei he insulin, GAD65 o IA-2, alone o in combina ions as an exp ession o IA. In his i s subclinical phase, he e is a sho coming o insulin p oduc- ion. In he second phase, inc eased blood glucose le els can be de ec ed i an o al glucose ole ance es (OGTT) is pe o med14. This second phase ends wi h an in il a ion o cy o oxic T- cells in o he isle s o Lange hans’15. Pe sis en con i med IA, de ined as posi i i y o an au oan ibody on wo consecu i e isi s and con i med in a second labo a o y, is he p ima y ou come o TEDDY. The seconda y ou come in TEDDY is he diagnosis o ype 1 diabe es as de ined by he Ame ican Diabe es Associa ion16. I is well documen ed ha he HLA- egion con ib- u es he g ea es p opo ion o gene ic isk (~50%) o ype 1 diabe es17. Se e al o he genes wi hin he immune sys em also con ibu e o he gene ic suscep ibili y18,19 in combina ion wi h en i onmen al ac o s. In he p esen s udy, he p ima y aim was o de e mine whe he SNPs wi hin complemen ac o s we e asso- cia ed wi h IA, o ype 1 diabe es, o bo h, in TEDDY child en. Resul s As o 31 Decembe 2014, 7.5% (413/5474) child en de eloped any pe sis en con i med IA (M/F = 1.43; 243/170). A o al o 2.1% (115/5474) o high isk child en had p og essed o ype 1 diabe es (M/F = 1.30; 65/50); how- e e 5 o he 115 who con e ed o ype 1 diabe es did no ha e IA. The median ollow-up ime was 73 mon hs (IQR 54–91). The median age a se ocon e sion o pe sis en IA was 27.9 (IQR: 15.4–47.8) mon hs and he median age a onse o ype 1 diabe es was 51.4 (IQR: 29.2–67.5) mon hs. Tables1 and 2 summa izes demo- g aphic cha ac e is ics o he subjec s who de eloped IA and hose who de eloped ype 1 diabe es. The h ee HLA-geno ypes including DR4 all con e ed inc eased haza d a ios using he ime o de elop IA as compa ed wi h DR3/3; DR4/8 [HR = 1.57; 95% CI 1.08–2.28; p = 0.018], DR4/4 [HR = 1.74; 95% CI 1.22–2.47; p = 0.002] and in pa icula DR3/4 [HR = 2.37; 95% CI 1.75–3.22; p < 0.001]. The high- isk geno ype HLA-DR3/4 con- e ed an inc eased haza d a io o ime o de elop ype 1 diabe es as compa ed wi h DR3/3 [HR = 2.28; 95% CI 1.30–4.00; p = 0.004]. Haza d a ios using he ime o IA in ela ion o SNPs in complemen genes. P opo ional haza d modeling iden i ied h ee SNPs as nominally associa ed (p < 0.05, p io o Bon e oni co ec ion) wi h IA in he www.na u e.com/scien i ic epo s/ 3 Scien i ic RepoR s | 6:27887 | DOI: 10.1038/s ep27887 p esence o high- isk HLA geno ypes in TEDDY pa icipan s. Two SNPs in in eg in alpha M (ITGAM) we e mo e common in con ols han cases (p o ec i e): s1143678 [HR = 0.80; 95% CI 0.66–0.98; p = 0.032] and s1143683 [HR = 0.80; 95% CI 0.66–0.98; p = 0.030]. SNP s4597342 in ITGAM was associa ed wi h an inc eased haza d a io using ime o IA [HR = 1.16; 95% CI 1.01–1.32; p = 0.041] (Table3). A e Bon e oni co ec ion, none o he 15 SNPs emained as signi ican . The e was no signi ican in e ac ion be ween any o he complemen SNPs on IA and HLA- isk ca ego ies. Haza d a io using ime o ype 1 diabe es in ela ion o complemen SNPs. None o he 15 SNPs es ed was associa ed wi h ime- o- ype 1 diabe es in he en i e coho (Table4). The e we e wo signi ican in e ac ions o SNPs wi h HLA geno ype on ype 1 diabe es; s2230199 in C3 (p = 0.0125) and s4844573 in C4BPA (p = 0.0275). When he coho was s a i ied based on HLA- isk- ca ego y, h ee SNPs achie ed nominal signi icance in he ime- o-e en analysis. Among DR3/4 subjec s SNPs: s17615 in CD21 (C3d- ecep o ) [HR = 1.52; 95% CI 1.05–2.20; p = 0.025: n = 2204] and s4844573 in C4BPA [HR = 0.63; 95% CI 0.43–0.92; p = 0.017: n = 2204] eached p < 0.05 (Fig.1A). Among DR4/4 subjec s SNP s2230199 in C3 [HR = 3.20; 95% CI 1.75–5.85]; p = 0.0002; n = 1085] achie ed Bon e oni-co ec ed signi icance since i was less han 0.000833 (0.05/60) in he HLA-speci ic analyses (Fig.1B). The e we e no signi ican associa ions o any o he 15 SNPs wi h espec o ype 1 diabe es among DR4/8 o DR3/3 subjec s (Fig.1C,D). Subjec s who de eloped IA Subjec s who did no de elop IA Subjec s who de eloped T1D Subjec s who did no de elop T1D Numbe o subjec s (n) 413 5061 115 5359 Age a i s IA, T1D o he mos ecen isi (mon hs) Median 27.9 69.9 51.4 72.9 IQR 15.4–47.8 49.7–89.5 29.2–67.5 54.7–91.1 Coun y n (%) Finland 125 (30.3) 1327 (26.2) 41 (35.7) 1411 (26.3) Ge many 16 (3.9) 235 (4.6) 6 (5.2) 245 (4.6) Sweden 159 (38.5) 1746 (34.5) 39 (33.9) 1866 (34.8) US 113 (27.3) 1753 (34.7) 29 (25.2) 1837 (34.3) High- isk HLA-DR-DQ geno ype n (%) DR3/4 218 (52.8) 1986 (39.2) 65 (56.5) 2139 (39.9) DR4/4 78 (18.9) 1008 (19.9) 19 (16.5) 1067 (19.9) DR4/8 66 (16.0) 900 (17.8) 16 (13.9) 950 (17.7) DR3/3 51 (12.3) 1167 (23.1) 15 (13.1) 1203 (22.5) Gende n (%) Female 170 (41.2) 2487 (49.1) 50 (43.5) 2607 (48.7) Table 1. Cha ac e is ics o subjec s by he s a us o isle au oimmuni y (IA) and ype 1 diabe es (T1D) in The En i onmen al De e minan s o Diabe es in he Young (TEDDY) s udy. IA Type 1 diabe es HR (95% CI) pHR (95% CI) p Coun y (Re e ence = US) Finland 1.31 (1.01–1.69) 0.045 1.61 (0.99–2.62) 0.054 Ge many 1.17 (0.70–1.98) 0.552 1.71 (0.71–4.13) 0.231 Sweden 1.23 (0.97–1.57) 0.088 1.08 (0.66–1.75) 0.765 High- isk HLA-DR-DQ geno ype (Re e ence = DR3/3) DR3/4 2.37 (1.75–3.22) < 0.001 2.28 (1.30–4.00) 0.004 DR4/4 1.74 (1.22–2.47) 0.002 1.38 (0.70–2.72) 0.353 DR4/8 1.57 (1.08–2.28) 0.018 1.17 (0.57–2.40) 0.669 Gende (Re e ence = Male) Female 0.72 (0.59–0.87) < 0.001 0.80 (0.56–1.16) 0.243 Table 2. Adjus ed haza d a ios (HR) and p- alues om he Cox p opo ional HR models in he analysis o ime- o he de elopmen o isle au oimmuni y (IA) and he analysis o ime o he de elopmen o ype 1 diabe es espec i ely. Each o he wo models had co a ia es o coun y o esidence, HLA-DR-DQ geno ype and gende . www.na u e.com/scien i ic epo s/ 4 Scien i ic RepoR s | 6:27887 | DOI: 10.1038/s ep27887 Discussion In he p esen s udy we we e no able o demons a e a obus associa ion be ween IA and SNPs wi hin he com- plemen genes. Ne e heless, we ound a signi ican associa ion wi h ype 1 diabe es ime- o-e en wi h one SNP ( s2230199) in he C3 gene, bu only among DR4/4 subjec s. The e was a s ong associa ion o C3 s2230199 among HLA-DR4/4 ca ie s, he geno ype con e ing he second highes isk o ype 1 diabe es. An in ec ion could p ecipi a e ype 1 diabe es in hose wi h gene ic p edisposi ion (based upon s2230199 in C3) since he exp ession le els o C3 could be di e en and hese subjec would no be able o igo ously igh he pa hogen. C3 is a key playe in he o ma ion o MAC om all h ee pa hways in he ac i a ion o complemen s. An in ec ion wi hin he isle s o Lange hans’ sho ly be o e he onse o ype 1 diabe es could explain he ind- ing o cy o oxic T- cells, which a e pa o he inna e immune esponse, a he ime o diagnosis. The C3 s2230199 SNP is a unc ional polymo phism, whe e he G allele causes an amino acid subs i u ion whe e a ginine is sub- s i u ed wi h glycine (R102G)20. The glycine a ian o C3 mo es as e in elec opho esis and has been shown o con e inc eased adhesion o monocy es21. The R102G a ian has shown a s ong associa ion wi h age- ela ed macula degene a ion (AMD)20,22,23. One ea u e indica ing di e en C3 le els in hose wi h ype 1 diabe es is ha child en wi h ype 1 diabe es ha e impai ed clo lysis due o inc eased inco po a ion o C3 in o clo s he eby p olonging lysis ime24. The e was a s ong co ela ion be ween he h ee SNPs in ITGAM ( s1143678 s s1143679, s = 0.85, p < 0.0001, s1143678 s s1143683, s = 0.99, p < 0.0001 and s1143679 s s1143683, s = 0.85, p < 0.0001) con- sis en wi h s ong linkage disequilib ium in his egion. All h ee SNPs ha e p e iously been shown o impai Mac-1-media ed neu ophil phagocy osis and phagocy osis media ed by Fcγ ecep o s as well as o impai he i m adhesion o neu ophils in heal hy subjec s25. SNP s17615 in CD21 was associa ed wi h an inc eased haz- a d a io using ime o ype 1 diabe es among HLA-DR3/4 ca ie s, while hose wi h he s17615 mino allele ha e been shown o ha e a p o ec ion om SLE. SNP s17615 ( oge he wi h s1048971 and s4308977) ha e a biological unc ion ha dec ease he splicing o he al e na i ely spliced exon 11 con e ing inc eased amoun s o he sho CR2 iso o m26. ITGAM, also deno ed CD11b-in eg in is an adhesion molecule exp essed on se e al cells wi hin he inna e immune sys em such as g anulocy es, monocy es, mac ophages and na u al kille cells. Mo eo e , ITGAM is in ol ed in he complemen sys em due o i s capaci y o bind inac i e complemen com- ponen 3b (iC3b)27. One a ian wi hin ITGAM ( s1143679, A-allele) con e s an amino acid subs i u ion whe e a ginine is subs i u ed wi h his idine R77H. This subs i u ion causes a unc ional dys unc ion o ITGAM lead- ing o impai ed adhesion and educed phagocy osis o complemen opsonized an igens. In e es ingly, he R77H a ian o ITGAM is highly associa ed wi h sys emic lupus e y hema osus (SLE)28. CD21 is also known as he C3d- ecep o o he Eps ein-Ba i us ecep o 29. The C3d- ecep o is exp essed on B-cells and has been shown Ch SNP Gene o in e es Mino allele MAF HR (95% CI) IA posi i e subjec s (n = 413) s subjec s wi h no IA (n = 5061) p 1 s665691 C1qC G 0.44829 0.98 (0.85–1.12) 0.7450 1 s17615 CD21 A 0.29459 1.02 (0.88–1.19) 0.7638 1 s1048971 CD21 A 0.34597 1.02 (0.88–1.18) 0.8277 1 s4308977 CD21 C 0.29010 1.02 (0.88–1.19) 0.7540 1 s4844573 C4BPA C 0.37678 0.93 (0.81–1.07) 0.3332 6 s9332739 C2 C 0.06065 1.04 (0.79–1.38) 0.7733 6 s2857009 C4 C 0.40608 0.98 (0.82–1.18) 0.8539 9 s10818488 C5 A 0.45661 1.11 (0.98–1.27) 0.1129 9 s7029078 C5 G 0.37450 1.03 (0.90–1.18) 0.6711 16 s1143678 ITGAM T 0.14770 0.80 (0.66–0.98) 0.0324 16 s1143683 ITGAM T 0.14773 0.80 (0.65–0.98) 0.0301 16 s1143679 ITGAM A 0.11538 0.82 (0.66–1.01) 0.0667 16 s4597342 ITGAM T 0.33638 1.16 (1.01–1.32) 0.0406 19 s2230199 C3 C 0.19865 0.94 (0.79–1.12) 0.4870 19 s7951 C3 A 0.08916 0.90 (0.70–1.15) 0.3918 Table 3. P ima y s a is ical analysis o isle au oimmuni y (IA). A o al o 15 SNPs on he ImmunoChip we e in e oga ed in The En i onmen al De e minan s o Diabe es in he Young (TEDDY) s udy in 413 subjec s posi i e o any IA e sus 5061 au oan ibody nega i e subjec s. Subjec s eligible o TEDDY ca ied any o he high isk HLA-DR-DQ geno ypes. The mino allele equency (MAF) was calcula ed om all subjec s (n = 5474). P opo ional haza d models we e adjus ed o HLA geno ypes, sex and coun y o esidence and popula ion s a i ica ion. The i s wo p incipal componen s esul ing om a p incipal componen s analysis o he US pa icipan s we e included in he model o adjus o popula ion s a i ica ion. Fi s deg ee ela i es we e excluded om he analyses. Haza d a ios (HR) and 95% con idence in e als (95% CI) we e es ima ed in his ime- o-e en analysis. A obus a iance es ima e was used o accoun o he dependence wi hin amilies in he models. The ac o s indica ing nominal signi ican isk o p o ec ion a e indica ed in bold. Fo a Chi-squa e alue o emain signi ican a e Bon e oni co ec ion o mul iple compa isons o 15 SNPs i mus be less han 0.00333. Abb e ia ions: CD21: C3D- ecep o , complemen ecep o ype 2, EBV- ecep o ITGAM: In eg in alpha M = MAC1 o complemen ecep o 3, C4BPA: C4 binding p o ein alpha-chain. www.na u e.com/scien i ic epo s/ 5 Scien i ic RepoR s | 6:27887 | DOI: 10.1038/s ep27887 o inc ease he B-cell ecep o (BCR) signaling as well as o enhance he p ocessing o an igens30. C4BPA has an inhibi o y e ec wi hin he complemen sys em31. The majo s eng h o he p esen s udy is ha we ha e child en ollowed om bi h who, in some cases, de el- oped IA and in some cases p og essed o ype 1 diabe es, he eby allowing us o s udy bo h disease ini ia ion and p og ession. Addi ional s eng hs a e ha we only ha e included subjec s ha ha e high- isk HLA geno ypes, he eby se ing an equal baseline o he majo gene ic suscep ibili y. The s udy design wi h epea ed sampling (mo e o en in he oddle s and mo e dispe sed a e ou yea s o age) allowed us o use p opo ional haza d a io modeling o iden i y haza d a ios o ime o end-poin s (IA and ype 1 diabe es) in his coho consis ing o h ee Eu opean popula ions and a USA popula ion ( h ee cen e s combined). This s udy design inc ease he powe compa ed wi h c oss-sec ional s udies o cases and con ols. A cu en limi a ion on he s udy, is he ela i ely low numbe o pa icipan s ha ha e eached he end-poin s (IA and ype 1 diabe es). The Bon e oni co ec ion migh be oo conse a i e i some SNPs a e in linkage disequilib ium as is mos likely he case o he ou SNPs wi hin ITGAM, he h ee SNPs in CD21 he wo SNPs in C3 and he wo SNPs in C5. Ne e heless, among DR4/4 subjec s SNP s2230199 in C3 achie ed Bon e oni-co ec ed signi icance. As he p edisposi ion o ype 1 diabe es occu ed only in hose ca ying HLA-DR4/4, he p oposed appa en con ibu ion is necessa y and equi es con i ma ion in unc ional s udies. As ou TEDDY coho ge s olde mo e child en will p oceed o IA and ype 1 diabe es, and we will ha e mo e powe in ou de e mina ion o associa ion o complemen ac o gene a ian s and haza d a ios using ime o IA and ype 1 diabe es. In summa y, al hough complemen gene SNPs a e no obus ly associa ed wi h IA o ype 1 diabe es, we concluded ha a ian s in he complemen genes do ha e some appa en con ibu ion o IA and ype 1 diabe es. Howe e , disc imina ion o complemen pa hways e ec s equi es u he esea ch. Subjec s and Me hods Newbo n child en we e sc eened o high- isk HLA geno ypes and en olled in o he TEDDY s udy om 1 Sep embe 2004 un il 28 Feb ua y 2010. A o al o 424 788 newbo ns in Finland, Sweden, Ge many and he Uni ed S a es (Colo ado, Geo gia and Washing on) we e sc eened as p e iously desc ibed32,33. The HLA high- isk geno ypes o subjec s om he gene al popula ion (GP) we e as ollows: DR3/4, DR4/4, DR4/8 and DR3/3. An addi ional six geno ypes we e used o i s deg ee ela i es (FDRs) o a subjec wi h ype 1 diabe es: DR4/4, DR4/1, DR4/13, DR4/4, DR4/9, DR3/9 (Table5). DR4 sub yping was pe o med o exclude child en om he GP wi h DRB1* 04:03. DQB1* 03:04 also quali ied o inclusion in o he TEDDY s udy. Sub yping was no done o Ch SNP Gene o in e es Mino allele MAF HR (95% CI) Subjec s wi h T1D (n = 115) s subjec s wi hou T1D (n = 5359) p 1 s665691 C1qC G 0.44829 1.10 (0.86–1.41) 0.4540 1 s17615 CD21 A 0.29459 1.24 (0.94–1.65) 0.1296 1 s1048971 CD21 A 0.34597 1.14 (0.87–1.50) 0.3333 1 s4308977 CD21 C 0.29010 1.23 (0.93–1.62) 0.1499 1 s4844573 C4BPA C 0.37678 0.82 (0.62–1.08) 0.1635 6 s9332739 C2 C 0.06065 1.08 (0.64–1.80) 0.7848 6 s2857009 C4 C 0.40608 1.02 (0.73–1.44) 0.8984 9 s10818488 C5 A 0.45661 1.15 (0.90–1.46) 0.2693 9 s7029078 C5 G 0.37450 1.12 (0.86–1.47) 0.3954 16 s1143678 ITGAM T 0.14770 0.84 (0.59–1.20) 0.3447 16 s1143683 ITGAM T 0.14773 0.84 (0.58–1.20) 0.3363 16 s1143679 ITGAM A 0.11538 0.66 (0.43–1.04) 0.0709 16 s4597342 ITGAM T 0.33638 1.05 (0.81–1.37) 0.6935 19 s2230199 C3 C 0.19865 1.30 (0.95–1.79) 0.1019 19 s7951 C3 A 0.08916 0.90 (0.56–1.44) 0.6591 Table 4. P ima y s a is ical analysis o ype 1 diabe es. A o al o 15 SNPs on he ImmunoChip we e in e oga ed in The En i onmen al De e minan s o Diabe es in he Young (TEDDY) s udy in 115 subjec s ha had de eloped ype 1 diabe es e sus all o he subjec s (n = 5359). Subjec s eligible o TEDDY ca ied any o he high isk HLA-DR-DQ geno ypes. The mino allele equency (MAF) was calcula ed om all subjec s (n = 5474). P opo ional haza d models we e adjus ed o HLA geno ypes, sex and coun y o esidence and popula ion s a i ica ion. The i s wo p incipal componen s esul ing om a p incipal componen s analysis o he US pa icipan s we e included in he model o adjus o popula ion s a i ica ion. Fi s deg ee ela i es we e excluded om he analyses. Haza d a ios (HR) and 95% con idence in e als (95% CI) we e es ima ed in his ime- o-e en analysis. A obus a iance es ima e was used o accoun o he dependence wi hin amilies in he models. Fo a Chi-squa e alue o emain signi ican a e Bon e oni co ec ion o mul iple compa isons o 15 SNPs i mus be less han 0.00333. Abb e ia ions: CD21: C3D- ecep o , complemen ecep o ype 2, EBV- ecep o , ITGAM: In eg in alpha M = MAC1 o complemen ecep o 3, C4BPA: C4 binding p o ein alpha chain. www.na u e.com/scien i ic epo s/ 6 Scien i ic RepoR s | 6:27887 | DOI: 10.1038/s ep27887 dis inguish BQB1* 02:0X and DQA1* 03:0X sub ypes. In he DQB1* 05:01 haplo ype, DR10 had o be excluded, only DR1 was eligible. A o al o 8667 newbo n child en we e en olled based on hese c i e ia. The ini ial blood sample o HLA- sc eening was ob ained ei he as co d blood in he ma e ni y clinic o as d y blood spo (DBS) on day h ee o ou . I he child ca ied any o he high- isk geno ypes, he amily was con ac ed by a s udy nu se and in i ed o pa icipa e in he 15 yea ollow-up s udy wi h blood sampling o analysis o au oan ibodies (GADA, IA-2A and mIAA) e e y 3 mon hs be ween 3 and 48 mon hs o age and e e y six mon hs he ea e . HLA- geno ypes we e con i med in a second blood sample a he 9- mon h isi . This blood sample was also used o he ImmunoChip SNP- analyses. This s udy was pe o med acco ding o he p inciples exp essed in he Decla a ion o Helsinki. Signed in o med consen s we e ob ained o all s udy pa icipan s om hei pa en s o p ima y ca e ake s, sepa a ely, o gene ic sc eening and pa icipa ion in he ollow-up. The s udy was app o ed by local Ins i u ional Re iew Boa ds and is moni o ed by an ex e nal e alu- a ion commi ee o med by he Na ional Ins i u es o Heal h. The local Ins i u ional Re iew Boa ds we e as ol- lows: Colo ado Mul iple Ins i u ional Re iew Boa d (Colo ado), E hics Commi ee o Sou hwes Finland Hospi al Dis ic (Finland), Uni e si y o Flo ida Heal h Cen e Ins i u ional Re iew Boa d, Medical College o Geo gia Human Assu ance Commi ee (2004–2010)/Geo gia Heal h Sciences Uni e si y Human Assu ance Commi ee (2011–2012)/Geo gia Regen s Uni e si y Ins i u ional Re iew Boa d (2013–p esen ) (Geo gia), Baye ischen Landesä z ekamme E hics Commi ee (Ge many), Regional E hics Boa d in Lund, Sec ion 2 (2004–2012)/Lund Uni e si y Commi ee o Con inuing E hical Re iew (2013–p esen ) (Sweden), Washing on S a e Ins i u ional Re iew Boa d (2004–2012)/Wes e n Ins i u ional Re iew Boa d (2013–p esen ) Washing on. Figu e 1. The igu e illus a es a summa y o 15 SNPs wi hin genes in he complemen sys em and p esen on he ImmunoChip ha we e in e oga ed in The En i onmen al De e minan s o Diabe es in he Young (TEDDY) s udy in 115 subjec s ha had de eloped ype 1 diabe es e sus all o he subjec s (n = 5359). Subjec s eligible o TEDDY ca ied any o he ou high isk HLA-DR-DQ geno ypes (DR3/4, 4/4, 4/8 o 3/3). The uppe 95% con idence in e al (CI) is indica ed by he uppe ba o he box, he lowe . 95% CI is indica ed by he lowe ba o he box and he haza d a io (HR) is indica ed by a solid ba in he middle o he box. Th ee haza d a ios we e di e en om 1.0 bu only s2230199 in C3 (p = 0.0002) among he DR4/4 ca ie s emained signi ican a e Bon e oni co ec ion since i was less han 0.000833 (0.05/60) in he HLA speci ic analyses. The explici alues o he HR (95% CI) o ype 1 diabe es in he HLA s a i ied analyses can be ound in an online Appendix. Panel A HRs and 95% CIs in 65 subjec s wi h ype 1 diabe es and 2139 subjec s wi hou ype 1 diabe es ca ying he HLA-DR3/4-geno ype (n = 2204). Panel B HRs and 95% CIs in 19 subjec s wi h ype 1 diabe es and 1067 subjec s wi hou ype 1 diabe es ca ying he HLA DR4/4-geno ype (n = 1086). Panel C HRs and 95% CIs in 16 subjec s wi h ype 1 diabe es and 950 subjec s wi hou ype 1 diabe es ca ying he HLA- DR4/8-geno ype (n = 966). Panel D HRs and 95% CIs in 15 subjec s wi h ype 1 diabe es and 1203 subjec s wi hou ype 1 diabe es ca ying he HLA- DR3/3- geno ype (n = 1218). www.na u e.com/scien i ic epo s/ 7 Scien i ic RepoR s | 6:27887 | DOI: 10.1038/s ep27887 S udy ou come – Isle au oimmuni y (IA) and ype 1 diabe es. The p ima y ou come was he ind- ing o pe sis en con i med IA assessed e e y 3 mon hs be ween 3 and 48 mon hs o age and e e y 6 mon hs he ea e . IA was con i med i iden i ied in bo h Re e ence Labo a o ies. Pe sis en au oimmuni y was de ined by he inding o con i med IA (any o GADA, IA-2A o mIAA) on wo o mo e consecu i e isi s. Da e o pe - sis en con i med IA was de ined as he d aw da e o he i s o wo consecu i e samples o a which he child was con i med as “posi i e” o a speci ic au oan ibody o any au oan ibody. Because child en can be bo n wi h ma e nal isle au oan ibodies, posi i e esul s due o ma e nal ansmission ia placen a we e excluded when de ining he child’s IA s a us. In o de o disc imina e ma e nal au oan ibodies om IA in he child, he IA s a us o he mo he was assessed when he child was 9 mon hs i IA was de ec ed a 3 o 6 mon hs o age. The child’s IA s a us was de ined based on bo h ma e nal and child IA. I ma e nal au oan ibodies we e iden i ied, he child was no conside ed pe sis en ly IA posi i e unless he child had a nega i e sample p io o he i s posi i e sample. All samples wi h a posi i e esul o IA and 5% o nega i e samples we e e-analyzed o con i ma ion in bo h Re e ence Labo a o ies. In he U. S., all samples we e assayed a he Ba ba a Da is Cen e o Childhood Diabe es a he Uni e si y o Colo ado, Den e ; in Eu ope all samples we e assayed a he Uni e si y o B is ol, he U. K. All he au oan ibodies we e measu ed using adioimmuno-binding assays34–37. The seconda y ou come o he s udy was he diagnosis o ype 1 diabe es as de ined by guidelines om he Ame ican Diabe es Associa ion16. HLA- yping. HLA- geno ype sc eening was pe o med using ei he a DBS punch o a small olume whole blood lysa e (WBL)33,38. Following PCR ampli ica ion o exon 2 o he HLA class II gene (DRB1, DQA1 o DQB1), alleles we e iden i ied ei he by di ec sequencing, oligonucleo ide p obe hyb idiza ion, o o he geno yping ech- niques. Typing o ce i y speci ic DR-DQ haplo ypes we e speci ied o each clinical cen e . HLA geno ypes in eligible subjec s we e con i med, using he 9-mon h sample, by he cen al HLA Re e ence Labo a o y a Roche Molecula Sys ems, Oakland, CA. Single Nucleo ide Polymo phisms (SNPs). SNP analysis was pe o med by he Cen e o Public Heal h Genomics a Uni e si y o Vi ginia, using he Illumina ImmunoChip. The ImmunoChip is a cus om a ay o geno yping o SNPs selec ed om egions o he human genome i mly associa ed wi h au oimmune diseases. The inal selec ion o SNPs con aining ~186 000 SNPs in 186 egions, o 12 au oimmune diseases was decided by he ImmunoChip Conso ium. The 9- mon h sample was used o SNP geno yping a e DNA ex ac ion done by Roche Molecula Sys ems (Pleasan on, CA). Quali y con ol (QC) s eps o assu e high quali y o he epo ed SNPs comp ised he exclusion o subjec s due o low call a e (> 5% SNPs missing) and disco dance wi h epo ed sex and p io geno yping. Secondly, SNPs we e emo ed om analysis due o low call a e (< 95%), Ha dy-Weinbe g equilib ium (HWE) p- alue < 10−6 (excep o ch omosome 6 due o HLA eligibili y equi e- men s) as well as being monomo phic o an inse ion-dele ion. A o al o 17 SNPs esiding in loci wi hin genes coding o complemen ac o s we e p esen on he ImmunoChip. O hese 17 SNPs, wo did no pass QC ( s1061170 in CFH and s2274567 in CD35). In he TEDDY s udy, we ha e p e iously pe o med a con i ma ion s udy o 41 SNPs associa ed wi h ype 1 diabe es and isk o au oan ibody posi i i y39. Code in TEDDY Geno ypes Abb e ia ion Gene al popula ion ADR3-DQA1* 05:01-DQB1* 02:01/ DR4-DQA1* 03:0X-DQB1* 03:02§DR3/4 Yes BDR4-DQA1* 03:0X-DQB1* 03:02§/ DR4-DQA1* 03:0X-DQB1* 03:02§DR4/4 Yes CDR4-DQA1* 03:0X-DQB1* 03:02§/ DR8-DQA1* 04:01-DQB1* 04:02 DR4/8 Yes DDR3-DQA1* 05:01-DQB1* 02:01/ DR3-DQA1* 05:01-DQB1* 02:01 DR3/3 Yes EDR4-DQA1* 03:0X-DQB1* 03:02§/ DR4-DQA1* 03:0X-DQB1* 02:01 DR4/4 No FDR4-DQA1* 03:0X-DQB1* 03:02§/ DR1#-DQA1* 01:01-DQB1* 05:01 DR4/1 No GDR4-DQA1* 03:0X-DQB1* 03:02§/ DR13-DQA1* 01:02-DQB1* 06:04 DR4/13 No HDR4-DQA1* 03:0X-DQB1* 03:02/ DR4-DQA1* 03:01-DQB1* 03:04 DR4/4 No IDR4-DQA1* 03:0X-DQB1* 03:02§/ DR9-DQA1* 03:01-DQB1* 03:03 DR4/9 No JDR3-DQA1* 05:01-DQB1* 02:01/ DR9-DQA1* 03:01-DQB1* 03:03 DR3/9 No Table 5. Display o high isk HLA-geno ypes cons i u ing he c i e ia o eligibili y o i s deg ee ela i es and child en om he gene al popula ion in o The En i onmen al De e minan s o Diabe es in he Young (TEDDY). §DR4 sub yping was pe o med o exclude child en om he gene al popula ion wi h DRB1* 04:03. DQB1* 03:04 also quali ied o inclusion in o he En i onmen al De e minan s in he Young. Sub yping was no done o dis inguish DQB1* 02:0X and DQA1* 03:0X sub ypes. #In his DQB1* 05:01 haplo ype, DR10 had o be excluded, only DR1 was eligible. The as e isk (* ) indica e ha he alignmen is unknown a any poin . www.na u e.com/scien i ic epo s/ 8 Scien i ic RepoR s | 6:27887 | DOI: 10.1038/s ep27887 S udy popula ion. To allow mo e gene alizable in e p e a ions o he gene ic associa ions, we included non-Hispanic Whi es om he US si es and all pa icipan s om he Eu opean si es. Only pa icipan s iden- i ied wi hou a ype 1 diabe es -FDR we e included. A e applying hese c i e ia, a o al o 5474 pa icipan s (M/F = 1.06; 2817/2657) wi h da a on 15 SNPs emained in he s udy. S a is ical analyses. Two ou comes we e analyzed: he ime o pe sis en con i med IA and he ime o ype 1 diabe es. The ime o pe sis en con i med IA was de ined as he age when he i s o con i med posi i e samples was aken, and he igh censo ed ime when he las nega i e sample was aken. The ime o ype 1 dia- be es was he child’s age a he ime o diagnosis o ype 1 diabe es. The p ima y analysis was o in es iga e he associa ion o he SNPs wi hin he complemen genes wi h he isk o de elopmen o pe sis en con i med IA/ ype 1 diabe es. The mino allele equency (MAF) was de e mined in he selec ed coho (n = 5474) and used h oughou he analyses. This s a egy in e s ha he isk is associa ed wi h he mino allele, e en hough he ce ain allele ha has he lowe equency may a y ac oss popula ions. Cox p opo ional haza d model was used o he analyses, adjus ing o HLA, sex and coun y o esidence. In addi ion, he i s wo p incipal compo- nen s (PC) o he ImmunoChip da a we e used as co a ia es in he Cox model o adjus o po en ial popula ion s a i ica ion. The axes o he p incipal componen s we e es ima ed based on he TEDDY US Caucasian subjec s and p inciple componen s o all TEDDY Subjec s we e calcula ed by p ojec ing hei da a on o he es ima ed axes. The di e ences be ween coun ies (especially Finland indi iduals ha ypically ha e di e en PCs) we e adjus ed by he coun y co a ia e, and o he popula ion subs uc u e (such as no he n s sou he n Eu opean o o igin) was adjus ed by PC1 and PC2. We ha e e-calcula ed he da a also wi hou co ec ion o he i s wo PCs and he only ma ked change was ha he p- alue o s4597342 in he IA analysis inc eased om 0.0406 o 0.0519. The e o e, we p esen he da a wi h adjus men o he i s wo PCs h oughou his publica ion. A o al o 4850 amilies pa icipa ed wi h only one child, 297 amilies wi h 2 child en, and 10 amilies wi h 3 child en. A obus a iance es ima e was used o accoun o he dependence wi hin amilies in he Cox p opo ional haza d model40. The seconda y analysis was o s udy he en a i e in e ac ion be ween non-HLA SNPs and HLA-geno ype and also coun y-speci ic associa ion o he SNPs wi h he ime o de elopmen o pe sis en con i med IA/ ype 1 diabe es. The s eng h and di ec ions o associa ions we e deno ed by haza d a ios (HR) wi h 95% con idence in e als (95% CI). P- alues less han 0.05 we e conside ed o indica e nominal s a is ical signi icance. Signi icance a e Bon e oni co ec ion was achie ed when p- alues we e less han 0.00333 (0.05/15) in he en i e coho and less han 0.000833 (0.05/60) in he HLA-speci ic analyses. S a is ical analyses we e pe o med using he S a is ical Analy ical So wa e (Ve sion 9.3, SAS Ins i u e, Ca y, NC). Quali y con ol analyses we e pe o med using PLINK (h p://pngu.mgh.ha a d.edu/pu cell/plink)41. The PCA was pe o med using EIGENSTRAT so wa e42. Re e ences 1. Rowe, P. e al. Inc eased complemen ac i a ion in human ype 1 diabe es panc ea a. Diabe es Ca e 36, 3815–3817 (2013). 2. Zhang, Q. e al. Se um p o eomics e eals sys emic dys egula ion o inna e immuni y in ype 1 diabe es. J Exp Med. 210, 191–203 (2013). 3. Ho man, W. H., Cud ici, C. D., Za anskaia, E. & Rus, H. Complemen ac i a ion in diabe ic ke oacidosis b ains. Exp Mol Pa hol. 80, 283–288 (2006). 4. Kallionpaa, H. e al. Inna e immune ac i i y is de ec ed p io o se ocon e sion in child en wi h HLA-con e ed ype 1 diabe es suscep ibili y. Diabe es 63, 2402–2414 (2014). 5. Idahl, L. A., Sehlin, J., Taljedal, I. B. & Tho nell, L. E. Cy o oxic ac i a ion o complemen by mouse panc ea ic isle cells. Diabe es 29, 636–642 (1980). 6. Alpe , C. A., Awdeh, Z. L. & Yunis, E. J. Complo ypes and ex ended haplo ypes in labo a o y medicine. Complemen In lamm. 6, 8–18 (1989). 7. McCluskey, J. e al. HLA and complemen allo ypes in Type 1 (insulin-dependen ) diabe es. Diabe ologia 24, 162–165 (1983). 8. Rich, S. S., Wei kamp, L. R. & Ba bosa, J. Gene ic he e ogenei y o insulin-dependen ( ype I) diabe es melli us: e idence om a s udy o ex ended haplo ypes. Am J Hum Gene . 36, 1015–1023 (1984). 9. Za a a i, P. e al. Condi ional linkage disequilib ium analysis o a complex disease supe locus, IDDM1 in he HLA egion, e eals he p esence o independen modi ying gene e ec s in luencing he ype 1 diabe es isk encoded by he majo HLA-DQB1, -DRB1 disease loci. Hum Mol Gene . 10, 881–889 (2001). 10. Nejen se , S. e al. Localiza ion o ype 1 diabe es suscep ibili y o he MHC class I genes HLA-B and HLA-A. Na u e 450, 887–892 (2007). 11. Noble, J. A. e al. HLA class I and gene ic suscep ibili y o ype 1 diabe es: esul s om he Type 1 Diabe es Gene ics Conso ium. Diabe es 59, 2972–2979 (2010). 12. Chen, M., Daha, M. R. & Kallenbe g, C. G. The complemen sys em in sys emic au oimmune disease. J Au oimmun. 34, J276–286 (2010). 13. Hole s, V. M. Complemen and i s ecep o s: new insigh s in o human disease. Ann Re Immunol. 32, 433–459 (2014). 14. Sosenko, J. M. e al. A isk sco e o ype 1 diabe es de i ed om au oan ibody-posi i e pa icipan s in he Diabe es P e en ion T ial-Type 1. Diabe es Ca e 31, 528–533 (2008). 15. Gep s, W. & Lecomp e, P. M. The panc ea ic isle s in diabe es. Am J Med. 70, 105–115 (1981). 16. Ame ican Diabe es Associa ion. Diagnosis and classi ica ion o diabe es melli us. Diabe es Ca e. 33 Suppl 1, S62–69 (2010). 17. Lambe , A. P. e al. Absolu e isk o childhood-onse ype 1 diabe es de ined by human leukocy e an igen class II geno ype: a popula ion-based s udy in he Uni ed Kingdom. J Clin Endoc inol Me ab. 89, 4037–4043 (2004). 18. Ba e , J. C. e al. Genome-wide associa ion s udy and me a-analysis ind ha o e 40 loci a ec isk o ype 1 diabe es. Na Gene . 41, 703–707 (2009). 19. Coope , J. D. e al. Con i ma ion o no el ype 1 diabe es isk loci in amilies. Diabe ologia 55, 996–1000 (2012). 20. Seddon, J. M. Complemen ac o 3 a ia ion is associa ed wi h AMD isk. Re ina Today 37–39 (2008). 21. A ilommi, H. Capaci y o complemen c3 pheno ypes o bind on o mononuclea cells in man. Na u e 251, 740–741 (1974). 22. Ya es, J. R. e al. Complemen C3 a ian and he isk o age- ela ed macula degene a ion. N Engl J Med. 357, 553–561 (2007). 23. Zhang, M. X. e al. Associa ion be ween a unc ional gene ic polymo phism ( s2230199) and age- ela ed macula degene a ion isk: a me a-analysis. Gene Mol Res. 14, 12567–12576 (2015). www.na u e.com/scien i ic epo s/ 9 Scien i ic RepoR s | 6:27887 | DOI: 10.1038/s ep27887 24. Hess, K. e al. A no el mechanism o hypo ib inolysis in diabe es: he ole o complemen C3. Diabe ologia 55, 1103–1113 (2012). 25. Zhou, Y. e al. Mul iple lupus-associa ed ITGAM a ian s al e Mac-1 unc ions on neu ophils. A h i is Rheum. 65, 2907–2916 (2013). 26. Douglas, K. B. e al. Complemen ecep o 2 polymo phisms associa ed wi h sys emic lupus e y hema osus modula e al e na i e splicing. Genes Immun. 10, 457–469 (2009). 27. Solo jo , D. A., Plusko a, E. & Plow, E. F. Dis inc oles o he alpha and be a subuni s in he unc ions o in eg in alphaMbe a2. J Biol Chem. 280, 1336–1345 (2005). 28. Fage holm, S. C., MacPhe son, M., James, M. J., Se ie -Guy, C. & Lau, C. S. The CD11b-in eg in (ITGAM) and sys emic lupus e y hema osus. Lupus. 22, 657–663 (2013). 29. Gau e, A. e al. Nuclea localiza ion o he Eps ein-Ba i us/C3d ecep o (CR2) in he human Bu ki B lymphoma cell, Raji. Mol Immunol. 29, 1113–1120 (1992). 30. Che uku i, A., Cheng, P. C. & Pie ce, S. K. The ole o he CD19/CD21 complex in B cell p ocessing and p esen a ion o complemen - agged an igens. J Immunol. 167, 163–172 (2001). 31. Sjobe g, A. P., T ouw, L. A. & Blom, A. M. Complemen ac i a ion and inhibi ion: a delica e balance. T ends Immunol. 30, 83–90 (2009). 32. The TEDDY S udy G oup. The En i onmen al De e minan s o Diabe es in he Young (TEDDY) s udy: s udy design. Pedia Diabe es 8, 286–298 (2007). 33. Hagopian, W. A. e al. The En i onmen al De e minan s o Diabe es in he Young (TEDDY): gene ic c i e ia and in e na ional diabe es isk sc eening o 421 000 in an s. Pedia Diabe es 12, 733–743 (2011). 34. To n, C., Muelle , P. W., Schlosse , M., Boni acio, E. & Bingley, P. J. Diabe es An ibody S anda diza ion P og am: e alua ion o assays o au oan ibodies o glu amic acid deca boxylase and isle an igen-2. Diabe ologia 51, 846–852 (2008). 35. Babaya, N. e al. Compa ison o insulin au oan ibody: polye hylene glycol and mic o-IAA 1-day and 7-day assays. Diabe es Me ab Res Re . 25, 665–670 (2009). 36. Boni acio, E. e al. Ha moniza ion o glu amic acid deca boxylase and isle an igen-2 au oan ibody assays o Na ional Ins i u e o Diabe es and Diges i e and Kidney Diseases Conso ia. J Clin Endoc inol Me ab. 95, 3360–3367 (2010). 37. Schlosse , M. e al. Diabe es An ibody S anda diza ion P og am: e alua ion o assays o insulin au oan ibodies. Diabe ologia 53, 2611–2620 (2010). 38. Dan onio, P. e al. P o iciency es ing o human leukocy e an igen-DR and human leukocy e an igen-DQ gene ic isk assessmen o ype 1 diabe es using d ied blood spo s. J Diabe es Sci Technol. 4, 929–941 (2010). 39. To n, C. e al. Role o Type 1 diabe es associa ed SNPs on isk o au oan ibody posi i i y in he TEDDY S udy. Diabe es 64, 1818–1829 (2015). 40. Wei, L. J., Lin, D. Y. & Weiss eld, L. Reg ession analysis o mul i a ia e incomple e ailu e ime da a by modeling ma ginal dis ibu ion. J Am S a Assoc. 84, 1065–1073 (1989). 41. Pu cell, S. e al. PLINK: a ool se o whole-genome associa ion and popula ion-based linkage analyses. Am J Hum Gene . 81, 559–575 (2007). 42. P ice, A. L. e al. P incipal componen s analysis co ec s o s a i ica ion in genome-wide associa ion s udies. Na Gene . 38, 904–909 (2006). Acknowledgemen s Funded by U01 DK63829, U01 DK63861, U01 DK63821, U01 DK63865, U01 DK63863, U01 DK63836, U01 DK63790, UC4 DK63829, UC4 DK63861, UC4 DK63821, UC4 DK63865, UC4 DK63863, UC4 DK63836, UC4 DK95300, and UC4 DK100238, and Con ac No. HHSN267200700014C om he Na ional Ins i u e o Diabe es and Diges i e and Kidney Diseases (NIDDK), Na ional Ins i u e o Alle gy and In ec ious Diseases (NIAID), Na ional Ins i u e o Child Heal h and Human De elopmen (NICHD), Na ional Ins i u e o En i onmen al Heal h Sciences (NIEHS), Ju enile Diabe es Resea ch Founda ion (JDRF), and Cen e s o Disease Con ol and P e en ion (CDC). This wo k suppo ed in pa by he NIH/NCATS Clinical and T ansla ional Science Awa ds o he Uni e si y o Flo ida (UL1 TR000064) and he Uni e si y o Colo ado (UL1 TR001082). Au ho Con ibu ions C.T. w o e he main manusc ip ex . X.L., S.O.-G. and W.-M.C. analyzed da a. S.S.R. e iewed and e ised he manusc ip . J.K. ensu e ha o iginal da a upon which he submission is based is p ese ed and e ie able o eanalysis, app o ing da a p esen a ion as ep esen a i e o he o iginal da a, o eseeing and minimizing obs acles o sha ing da a, ma e ials, algo i hms o eagen s in his wo k. The TEDDY S udy G oup join ly supe ised he wo k. Addi ional In o ma ion Supplemen a y in o ma ion accompanies his pape a h p://www.na u e.com/s ep Compe ing inancial in e es s: The au ho s decla e no compe ing inancial in e es s. How o ci e his a icle: Tö n, C. e al. Complemen gene a ian s in ela ion o au oan ibodies o be a cell speci ic an igens and ype 1 diabe es in he TEDDY S udy. Sci. Rep. 6, 27887; doi: 10.1038/s ep27887 (2016). This wo k is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License. The images o o he hi d pa y ma e ial in his a icle a e included in he a icle’s C ea i e Commons license, unless indica ed o he wise in he c edi line; i he ma e ial is no included unde he C ea i e Commons license, use s will need o ob ain pe mission om he license holde o ep oduce he ma e ial. To iew a copy o his license, isi h p://c ea i ecommons.o g/licenses/by/4.0/