Enterovirus-associated changes in blood transcriptomic profiles of children with genetic susceptibility to type 1 diabetes

ARTICLE En e o i us-associa ed changes in blood ansc ip omic p o iles o child en wi h gene ic suscep ibili y o ype 1 diabe es Niina Lie zen 1 &Le T. T. An 1 &Ma ia K. Jaakkola 1,2 &Henna Kallionpää 1 & Sami Oika inen 3,4 &Juha Mykkänen 5,6 &Mikael Knip 7,8,9,10 &Rii a Veijola 11,12 & Jo ma Ilonen 13,14 &Jo ma Toppa i 5,15 &Heikki Hyö y 3,4 &Rii a Lahesmaa 1 & Lau a L. Elo 1 Recei ed: 8 June 2017 /Accep ed: 24 Augus 2017 /Published online: 8 No embe 2017 #The Au ho (s) 2017. This a icle is an open access publica ion Abs ac Aims/hypo hesis En e o i us in ec ions ha e been associa ed wi h he de elopmen o ype 1 diabe es in mul iple s udies, bu li le is known abou en e o i us-induced esponses in child en a isk o de eloping ype 1 diabe es. Ou aim was o use genome-wide ansc ip omics da a o cha ac e ise en e o i us-associa ed changes in whole-blood samples om child en wi h gene ic suscep ibili y o ype 1 diabe es. Me hods Longi udinal whole-blood samples (356 samples in o al) collec ed om 28 pai s o child en a inc eased isk o de eloping ype 1 diabe es we e sc eened o he p esence o en e o i us RNA. Se en o hese samples we e de ec ed as en e o i us-posi i e, each o hem collec ed om a di e en child, and ansc ip omics da a om hese child en we e analysed o unde s and he indi idual-le el esponses associ- a ed wi h en e o i us in ec ions. T ansc ip clus e s wi h peaking o d opping exp ession a he ime o en e o i us posi i i y we e selec ed as he en e o i us-associa ed signals. Resul s S ong signs o ac i a ion o an in e e on esponse we e de ec ed in ou child en a en e o i us posi i i y, while Niina Lie zen and Le T. T. An con ibu ed equally o he s udy. Rii a Lahesmaa and Lau a L. Elo a e join senio au ho s. Elec onic supplemen a y ma e ial The online e sion o his a icle (h ps://doi.o g/10.1007/s00125-017-4460-7) con ains pee - e iewed bu unedi ed supplemen a y ma e ial, which is a ailable o au ho ised use s. *Niina Lie zen niina.lie zen@u u. i *Lau a L. Elo lau a.elo@u u. i 1 Tu ku Cen e o Bio echnology, Uni e si y o Tu ku and Åbo Akademi Uni e si y, Tykis öka u 6, FI-20520 Tu ku, Finland 2 Depa men o Ma hema ics and S a is ics, Uni e si y o Tu ku, Tu ku, Finland 3 Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e, Finland 4 Fimlab Labo a o ies, Pi kanmaa Hospi al Dis ic , Tampe e, Finland 5 Depa men o Pedia ics, Tu ku Uni e si y Hospi al, Tu ku, Finland 6 Resea ch Cen e o Applied and P e en i e Ca dio ascula Medicine, Uni e si y o Tu ku, Tu ku, Finland 7 Child en’s Hospi al, Uni e si y o Helsinki and Helsinki Uni e si y Hospi al, Helsinki, Finland 8 Resea ch P og ams Uni , Diabe es and Obesi y, Uni e si y o Helsinki, Helsinki, Finland 9 Folkhälsan Resea ch Cen e , Helsinki, Finland 10 Tampe e Cen e o Child Heal h Resea ch, Tampe e Uni e si y Hospi al, Tampe e, Finland 11 Depa men o Pedia ics, PEDEGO Resea ch Uni , Uni e si y o Oulu, Oulu, Finland 12 Depa men o Child en and Adolescen s, Medical Resea ch Cen e , Oulu Uni e si y Hospi al, Oulu, Finland 13 Immunogene ics Labo a o y, Ins i u e o Biomedicine, Uni e si y o Tu ku, Tu ku, Finland 14 Depa men o Clinical Mic obiology, Tu ku Uni e si y Hospi al, Tu ku, Finland 15 Depa men o Physiology, Ins i u e o Biomedicine, Uni e si y o Tu ku, Tu ku, Finland Diabe ologia (2018) 61:381–388 DOI 10.1007/s00125-017-4460-7 ansc ip omic changes in he o he h ee child en indica ed ac i a ion o adap i e immune esponses. Addi ionally, a la ge p opo ion o he en e o i us-associa ed changes we e speci ic o indi iduals. An en e o i us-induced signa u e was buil using 339 genes peaking a en e o i us posi i i y in ou o he child en, and 77 o hese genes we e also up egula ed in human pe iphe al blood mononuclea cells in ec ed in i o wi h di e en en e o i uses. These genes sepa a ed he ou en e o i us-posi i e samples clea ly om he emaining 352 blood samples analysed. Conclusions/in e p e a ion We ha e, o he i s ime, iden i- ied en e o i us-associa ed ansc ip omic p o iles in whole- blood samples om child en wi h gene ic suscep ibili y o ype 1 diabe es. Ou esul s p o ide a s a ing poin o unde - s anding he indi idual esponses o en e o i us in ec ions in blood and hei po en ial connec ion o he de elopmen o ype 1 diabe es. Da a a ailabili y The da ase s analysed du ing he cu en s udy a e included in his published a icle and i s supple- men a y in o ma ion iles (www.b k. i/ esea ch/ compu a ional-biomedicine/1234-2) o a e a ailable om he Gene Exp ession Omnibus (GEO) eposi o y (accession GSE30211). Keywo ds Clinical immunology .En e o i us .Human . Mic oa ay .P edic ion and p e en ion o ype 1 diabe es Abb e ia ions DAVID Da abase o Anno a ionVisualiza ion and In eg a ed Disco e y DIPP Type 1 Diabe es P edic ion and P e en ion FDR False-disco e y a e IPA Ingenui y Pa hway Analysis PBMC Pe iphe al blood mononuclea cell RMA Robus mul ia ay a e age UPC Uni e sal exP ession Codes In oduc ion En e o i uses a e among he mos common i uses causing in ec ions in humans. They a e single-s anded RNA i uses ha eplica e ypically in he in es ine, bu can occasionally sp ead also o blood and ce ain in e nal o gans. Al hough en e o i us in ec ions a e mos ly asymp oma ic o cause only mild symp oms, hey can also cause se e e illnesses such as meningi is, myoca di is and hand- oo -and-mou h disease. Se e al s udies ha e associa ed i al in ec ions, especially human en e o i us in ec ions, wi h he de elopmen o ype 1 diabe es [1–4]. En e o i uses ha e a clea opism o panc e- a ic be a cells [5], and low-g ade en e o i us in ec ion has been de ec ed in panc ea ic isle s o li ing indi iduals wi h ecen ly diagnosed ype 1 diabe es [6]. P ospec i e s udies ha e also ound signs o en e o i us in ec ions mo e common- ly in child en who la e de elop ype 1 diabe es au oan ibodies o clinical ype 1 diabe es han in con ol child en [1,2,7]. The p esence o en e o i uses is no , howe e , hough o di- ec ly esul in an inc eased isk o ype 1 diabe es. The ou - comes o in ec ion likely depend on complex ela ionships be ween he hos and he i us: o example, he gene ic back- g ound and indi idual p ope ies o he hos [8–11], he iming o in ec ions [12] and he ype o en e o i us in ading he hos [13,14]. Cu en ly, only limi ed da a a e a ailable ega ding in i o en e o i us esponses in child en a isk o de eloping ype 1 diabe es, and mos s udies s ill ely on in i o in ec ion models. The e o e, be e unde s anding o he indi idual- le el esponses o en e o i us in ec ion is equi ed o gain insigh s in o he a iable ou comes o hese in ec ions. In his s udy, we pe o med, o he i s ime, genome-wide ansc ip omic analysis o en e o i us-associa ed changes in child en wi h gene ic suscep ibili y o ype 1 diabe es. We analysed mic oa ay da a om 44 longi udinally collec ed whole-blood samples [15] om se en child en who we e en e o i us-posi i e in one o he ollow-up samples. Ou aim was o unde s and he indi idual-le el ansc ip omic changes associa ed wi h en e o i us in ec ions and o cha ac- e ise he common ea u es o en e o i us esponses in young child en. Me hods S udy pa icipan s and sample selec ion The mic oa ay da a used in his s udy a e pa o he da ase published by Kallionpää e al [15] (GEO accession GSE30211) co e ing 356 PAXGene whole-blood RNA samples measu ed using he A yme ix Human Genome U219 A ay (A yme ix, San a Cla a, CA, USA). The samples we e collec ed om 28 pai s o child en pa icipa ing in he Finnish Type 1 Diabe es P edic ion and P e en ion (DIPP) s udy in Tu ku, Finland. All child en in he DIPP s udy ca y HLA- con e ed gene ic isk o ype 1 diabe es, and hey ha e been obse ed om bi h a egula in e als [16]. All child en had w i en pa en al consen and he E hics Commi ee o Tu ku Uni e si y Hospi al had g an ed app o al o he DIPP s udy. The s udy was ca ied ou in acco dance wi h he p inciples o he Decla a ion o Helsinki. The p esence o en e o i us RNA was s udied using quan i- a i e RT-PCR as desc ibed p e iously [17] om he same 356 RNA samples used o he mic oa ay analyses. The RT-PCR was ca ied ou in h ee pa allel eac ions. I all h ee eac ions ga e a posi i e esul , he sample was classi ied as s ongly en e o i us-posi i e; i only one o he eac ions was posi i e, he sample was classi ied as weakly en e o i us-posi i e. In o al, se en samples we e de ec ed as en e o i us-posi i e, each 382 Diabe ologia (2018) 61:381–388 collec ed om a di e en child. Mic oa ay da a o all 44 sam- ples om hese en e o i us-posi i e child en we e selec ed o u he analyses (elec onic supplemen a y ma e ial [ESM] Table 1). Mic oa ay da a p ocessing and clus e ing The mic oa ay da a we e p e-p ocessed using he obus mul ia ay a e age (RMA) me hod implemen ed in he Bioconduc o package a y e sion 1.44.0 (h p://bioconduc o .o g/packages/ elease/ bioc/h ml/a y.h ml), and log 2 - ans o med. The Uni e sal exP ession Codes (UPC) me hod o he Bioconduc o pack- age SCAN.UPC e sion 2.12.1 (h ps://bioconduc o .o g/ packages/ elease/bioc/h ml/SCAN.UPC.h ml)[18] was used o il e ou p obe se s wi h low exp ession (UPC < 0.5 in all he samples). Fo each p obe se , he RMA-no malised exp ession alues we e ans o med in o zsco es based on hei child-speci ic mean and s anda d de ia ion o e he i us-nega i e samples. The zsco e p o iles we e clus e ed sepa a ely o each child using he k-means algo i hm, wi h Pea son co ela ion and k = 10. Fo each child, he clus e s wi h he highes and lowes cen oid alues a he ime o en e o i us posi i i y we e iden i ied ( e e ed o as peaking and d opping clus e s, espec i ely). Func ional da a analysis Func ional classi ica ion o he da a was pe o med using he Da abase o Anno a ion, Visualiza ion and In eg a ed Disco e y (DAVID; h ps://da id.nci c .go /; accessed No embe o Decembe 2016) [19] and Ingenui y Pa hway Analysis (IPA; Qiagen Bioin o ma ics, Aa hus, Denma k, www.qiagenbioin o ma ics.com/p oduc s/ingenui y- pa hway-analysis/; accessed No embe o Decembe 2016) ools. Gene on ology classes wi h DAVID alse-disco e y a e (FDR) < 0.05 and IPA pa hways wi h p alue < 0.001 we e conside ed signi ican ly en iched. The In e e ome 2.01 da abase (www.in e e ome.o g; accessed No embe o Decembe 2016) [20] was used o s udy he p esence o human in e e on- egula ed genes. Mic oa ay da a om human PBMCs in ec ed in i o The ansc ip omics da a om child en a isk o ype 1 diabe- es we e compa ed wi h mic oa ay da a (HumanHT-12 V3.0 BeadChip, Illumina, San Diego, CA, USA) om pe iphe al blood mononuclea cells (PBMCs) in ec ed wi h en e o i us in i o [14], including h ee eplica e samples o PBMCs in- ec ed wi h ATCC s ain o echo i us 9 o wild- ype Coxsackie B1 i us s ains CDC10802 and CDC10796 o 48 h, and unin ec ed con ol PBMCs (see Da ase 1 published on h ps://www.b k. i/1234-2/). The da a we e p e-p ocessed using he a iance-s abilising no malisa ion o he Bioconduc o lumi package e sion 2.18.0 (h ps://bioconduc o .o g/packages/ elease/bioc/h ml/lumi.h ml)[21]. The UPC me hod was used o il e ou p obes wi h low exp ession un il eaching he numbe o p obes equal o he da ase om he en e o i us- posi i e child en. Di e en ial exp ession was de e mined using he Bioconduc o ROTS [22] package e sion 1.1.1 (h ps:// bioconduc o .o g/packages/ elease/bioc/h ml/ROTS.h ml)and cu -o alues FDR < 0.05 and old change > 1.5. To enable compa ison be ween he Illumina and A yme ix pla o ms, he p obes and p obe se s we e mapped o genes using IPA (Qiagen; accessed No embe o Decembe 2016). Resul s En e o i us RNA was de ec ed in se en o 356 whole-blood RNA samples, wi h i e s ongly en e o i us-posi i e and wo weakly en e o i us-posi i e samples each aken om a di e - en child. To cha ac e ise en e o i us-associa ed changes in whole- blood ansc ip ome, we s udied longi udinal gene exp ession p o iles o hese en e o i us-posi i e child en by di iding all p obe se s child-speci ically in o en clus e s (ESM Fig. 1;see also Da ase 2 published on h ps://www.b k. i/1234-2/). Fo each child, he clus e s wi h he highes and lowes cen oid alue a he ime o en e o i us posi i i y we e selec ed as he en e o i us-associa ed signals (Fig. 1a–g). Fo ou s ongly en e o i us-posi i e child en, he o e laps be ween peaking and d opping clus e s we e highe (a e age o e laps o 46% and 37%, espec i ely) han hose be ween he o he child en (a e age o e laps o 8%) (Fig. 1h, i). In o al, 593 p obe se s mapping o 339 dis inc genes we e de ec ed in he peaking clus e s o all ou child en. This se was de ined as he en e o i us-induced signa u e. Howe e , app oxima ely 20% o he p obe se s in each o hese peaking clus e s we e child-speci ic, indica ing he p esence o indi idual di e ences in en e o i us esponses. The o he h ee child en had lowe o e laps wi h each o he and wi h all o he child en. Genes in ol ed in an i i al immune esponses and espe- cially in e e on signalling we e signi ican ly en iched in bo h he peaking clus e s o he ou s ongly en e o i us-posi i e child en and he en e o i us-induced signa u e (Fig. 2a; see also Da ase 2 published on h ps://www.b k. i/1234-2/). Also, child-speci ic exp ession p o iles o wo in e e on signalling genes, STAT2 and MX1, showed clea peaks in he ou chil- d en a en e o i us posi i i y (Fig. 2b–h). Fu he mo e, 74% o he signa u e genes we e egula ed by in e e ons based on he In e e ome da abase. The B cell ecep o signalling pa h- way was among he mos signi ican ly en iched pa hways in he peaking clus e s o he o he h ee child en, po en ially e lec ing he ac i a ion o adap i e immune esponses. Based on he d opping clus e s, en e o i us in ec ion was also associa ed wi h down egula ion o genes in ol ed in mRNA p ocessing, ansc ip ion o ansla ion in six o he se en chil- d en (see Da ase 3 published on h ps://www.b k. i/1234-2/). En e o i us RNA is de ec able in blood o only a ew days du ing he acu e phase o in ec ion. To es ima e he iming o Diabe ologia (2018) 61:381–388 383 in ec ion ela i e o sample collec ion, we compa ed ou peaking and d opping clus e s wi h whole-blood ansc ip ional changes du ing acu e and eco e y phases a e in luenza i us in ec ion, as epo edbyZhaie al[23] (see Da ase 2 published on h ps:// www.b k. i/1234-2/). O he 25 op genes up egula ed du ing he acu e phase o in luenza in ec ion [23], 23 o e lapped ou en e o i us-induced signa u e. Also, all se en na u al kille (NK) cell ac i a ion signa u e genes associa ed wi h he acu e phase o in luenza in ec ion [23] peaked in mo e han one o he en e o i us-posi i e child en. O e laps wi h he op up- and down egula ed genes speci ic o he eco e y phase a e in lu- enza i us in ec ion [23] we e low o all he child en. We also compa ed ou esul s wi h en e o i us-induced esponses in human PBMCs in ec ed in i o wi h h ee di e - en en e o i uses [14]. The genes up egula ed in he in i o in ec ions we e en iched wi h hose associa ed wi h he de ence esponse o i us and he ype I in e e on signalling pa hway, simila ly o he genes in ou en e o i us-induced signa u e. O he genes up egula ed by any o he en e o i uses, 70% we e p esen in a leas one o he peaking clus e s. O e all, 77 genes p esen in he en e o i us-induced signa u e we e up egula ed in all h ee in i o in ec ions (ESM Fig. 2a). O hese genes, 73 we e in e e on- egula ed based on he In e e ome da abase. Al hough only app oxima ely 50% o he genes down egula ed in he in i o in ec ions we e p esen in any o he d opping clus e s o en e o i us-posi i e child en, genes associa ed wi h ansla ion we e en iched in bo h da ase s. As en e o i uses a e known o in ec panc ea ic isle s [5] and ha e been ound in panc eases o indi iduals wi h ype 1 diabe es mo e o en han in non-diabe ic con ol g oups [24, 25], we also compa ed ou esul s wi h en e o i us-induced esponses in human panc ea ic isle s in ec ed in i o wi h en- e o i uses [10,26]. App oxima ely hal o he en e o i us- induced genes in human panc ea ic isle s [10,26]we ealso p esen in ou en e o i us-induced blood ansc ip omic signa- u e in child en a isk o de eloping ype 1 diabe es (ESM Fig. 2b), while he o e laps wi h he genes p esen in he peaking clus e s o he o he h ee ype 1 diabe es isk child en we e low (ESM Fig. 2b). In o al, he e we e 64 en e o i us- induced genes common o he in i o in ec ed panc ea ic isle s 2314 2219 3425 2325 2508 1318 2572 2515 3844 4590 2360 1532 1258 2417 S ong 1 S ong 2 S ong 3 S ong 4 S ong 5 Weak 1 Weak 2 S ong 1 S ong 2 S ong 3 S ong 4 S ong 5 Weak 1 Weak 2 Clus e cen oid alue Clus e cen oid alue Clus e cen oid alue Clus e cen oid alue Clus e cen oid alue Clus e cen oid alue Clus e cen oid alue 1 2 0 -1 -2 1 2 0 -1 -2 1 2 0 -1 -2 Time om EV+ (mon hs) Time om EV+ (mon hs) Time om EV+ (mon hs) Time om EV+ (mon hs) Time om EV+ (mon hs) Time om EV+ (mon hs) Time om EV+ (mon hs) a 1 2 0 -1 -2 d 1 2 0 -1 -2 e 5 1 2 0 -1 -2 ● ● ● ● 1 2 0 -1 -2 g hi S ong 1 S ong 2 S ong 3 S ong 4 S ong 5 Weak 1 Weak 2 S ong 1 S ong 2 S ong 3 S ong 4 S ong 5 Weak 1 Weak 2 b 0 10203040 0 5 10 15 20 25 30 0 5 10 15 20 25 0 10 15 -20 -15 -10 -5 0 -5 0 5 10 -10 0 10 20 c Fig. 1 (a–g) A e age exp ession p o iles o clus e s peaking o d opping a en e o i us posi i i y o he se en en e o i us-posi i e child en: S ong 1 (a), S ong 2 (b), S ong 3 (c), S ong 4 (d), S ong 5 (e), Weak 1 ( )andWeak2(g). Red, peaking clus e s; blue, d opping clus e s. EV+, en e o i us-posi i e. (h,i) O e lapping p obe se s be ween he peaking and d opping clus e s. The black a eas indica e he p opo ion o o e lapping p obe se s ela i e o he child/clus e no ed a he op o he column. The boxes highligh ed in he ou lined ame show he peaking (h) and d opping (i) clus e s o he ou s ongly en e o i us- posi i e child en wi h he mos simila en e o i us-associa ed changes. The o al numbe s o p obe se s in each clus e a e p esen ed in he diagonal. The i e child en wi h s ongly en e o i us-posi i e blood samples a e deno ed as S ong 1–S ong 5. The wo child en wi h weakly en e o i us-posi i e blood samples a e deno ed as Weak 1 and Weak 2 384 Diabe ologia (2018) 61:381–388 [10,26] and ou en e o i us-induced blood ansc ip omic sig- na u e, all o which we e associa ed wi h an i i al in e e on esponses, including IFIH1,IRF7,MX1,STAT1 and STAT2. Finally, we es ed whe he he 339 en e o i us-induced sig- na u e genes o he 77 genes also up egula ed in all h ee in i o in ec ions o human PBMCs could di e en ia e be- ween he en e o i us-posi i e and en e o i us-nega i e blood samples in he ull ansc ip omics da ase by Kallionpää e al [15]. The ou en e o i us-posi i e samples wi h clea signs o in e e on esponse ac i a ion we e clea ly sepa a ed om all o he samples using ei he o he gene se s, while he o he h ee en e o i us-posi i e samples we e no sepa a ed om he en e o i us-nega i e samples (Fig. 3a, ESM Fig. 3a). To ensu e ha he obse ed sepa a ion was no due o he use o he same ou en e o i us-posi i e child en in de ining he signa u e, a simila analysis was pe o med using he genes up egula ed in he in i o in ec ed PBMCs by any o he h ee en e o i uses. He e also, he ou s ongly en e o i us-posi i e samples we e 6 8 10 12 −5 6 8 10 12 −10 6 8 10 12 6 8 10 12 6 8 10 12 6 8 10 12 −20 −10 0 10203040 0 5 10 01020 0 5 10 20 30 0 5 10 15 20 25 0 5 10 15 0 6 8 10 12 15 25 Time om EV+ (mon hs) Log2 exp ession Time om EV+ ( mon hs ) Log2 exp ession Time om EV+ (mon hs) Log2 exp ession Time om EV+ (mon hs) Log2 exp ession Time om EV+ (mon hs) Log2 exp ession Time om EV+ ( mon hs ) Log2 exp ession Time om EV+ ( mon hs ) Log2 exp ession bcde gh a Nucleus Cy oplasm IFN IFN R NF-KB BCL-2 BAX BAK IRF1 TAP1 IFITM3 IFITM1 IFITM2 IFITM1 IFITM2 IFIT3 IFI35 G1P3 IRF9 DRIP150 IRF9 PSMB8 MX1 G1P2 IFI35 IFIT1 OAS1 ISRE Glucoco icoid IRF9 PSMB8 p65 JAK1 TC-PTP TC-PTP STAT1 STAT1 STAT2 SOCS1 P STAT1 PSTAT1 P STAT2 P P STAT1 PIAS1 P STAT1 PSTAT2 P STAT1 P PP TC-PTP JAK2 JAK1 JAK1 JAK2 TYK2 R IFNA R1 / IFNA R2 GAS γ IFN γ IFN γ IFN γ IFN γ αβ RR α α β β STAT1 Fig. 2 (a) Genes in peaking clus e s mapping o he in e e on signalling pa hway based on he IPA ool. Red, genes p esen in a leas ou peaking clus e s o he s ongly en e o i us-posi i e child en; pink, genes p esen in a leas one peaking clus e . (b–h) Child-speci ic exp ession p o iles o wo genes o he in e e on signalling pa hway, MX1 (g ey) and STAT2 (black). IFNγis also known as IFNG;IFNα/βis also known as IFNA1/ B1;TC-PTP is also known as PTPN2;NF-κB p65 is also known as RELA;BCL-2 is also known as BCL2;BAK is also known as BAK1; DRIP150 is also known as MED14;G1P2 is also known as ISG15; G1P3 is also known as IFI6. EV+, en e o i us-posi i e; GAS, IFNG- ac i a ed sequence; ISRE, in e e on-s imula ed egula o y elemen . The i e child en wi h s ongly en e o i us-posi i e blood samples a e deno - ed as S ong 1–S ong 5 (b– ). The wo child en wi h weakly en e o i us- posi i e blood samples a e deno ed as Weak 1 and Weak 2 (g,h) Diabe ologia (2018) 61:381–388 385 clea ly sepa a ed om he en e o i us-nega i e samples (ESM Fig. 3b). Up egula ion o in e e on- egula ed genes du ing en e o- i us in ec ion is one o he conspicuous ea u es in his s udy, and ac i a ion o in e e on signalling has been obse ed in he blood o child en who ha e de eloped diabe es- ela ed au oan ibodies o clinical ype 1 diabe es be o e he i s de- ec ion o au oan ibodies [15,27]. Howe e , he exp ession o he 339 en e o i us-induced signa u e genes (ESM Figs 3a, c and 4) o he 77 genes also up egula ed wi h in i o en e o- i us in ec ion (Fig. 3a, b) did no show ma ked di e ences be ween au oan ibody-nega i e and au oan ibody-posi i e child en be o e o a e se ocon e sion based on he longi u- dinal blood ansc ip omics da a om child en a isk o ype 1 diabe es epo ed by Kallionpää e al [15] o Fe ei a e al [27]. Mo eo e , ewe han 50% o he genes up egula ed in child en who ha e de eloped diabe es- ela ed au oan ibodies o clinical ype 1 diabe es in he wo a o emen ioned s udies [15,27] o e lapped wi h he en e o i us-induced signa u e (Fig. 3c, ESM Fig. 3c). Discussion In he cu en s udy, we ha e iden i ied en e o i us-associa ed ansc ip omic p o iles in whole-blood samples om se en chil- d en wi h gene ic suscep ibili y o ype 1 diabe es and cha ac e ised hei indi idual esponses o en e o i us in ec ions. In e e on esponse is a cen al pa o he inna e an i i al immune esponse, and se e al en e o i us s ains induce a p o ound in e e on esponse in human blood cells [14,28]. We de ec ed clea signs o in e e on esponse ac i a ion in ou s ongly en e o i us-posi i e child en. En e o i us- associa ed changes in hese child en esembled p e iously e- po ed di e ences occu ing du ing he acu e phase o i us in ec ion [23], indica ing ha hese samples we e collec ed du ing he acu e phase o in ec ion cha ac e ised by high i us load. In wo child en wi h only weakly en e o i us-posi i e blood samples and one child wi h a s ongly en e o i us- posi i e blood sample, no s ong signs o in e e on esponse ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● Aab− ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● Pos Se o ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● P eSe o 19 Ve i ied DIPP en e o i us up egula ed genes Kallionpää e al up egula ed genes Fe ei a e al IFN g enes Sum o z sco es 150 100 50 0 -50 a b c Aab−Pos Se oP eSe o Sum o z sco es ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 1500 ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ●● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ●● ● ●● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● 1000 500 0 3143 19 22 33 151  Fig. 3 Exp ession o he 77 genes p esen in ou en e o i us-induced signa u e and up egula ed in all h ee in i o en e o i us in ec ions in mic oa ay da a published by: (a) Kallionpää e al [15]; and (b) Fe ei a e al [27]. Sums o child-speci ic zsco es o e he 77 genes we e calcula ed o each o he 356 whole blood samples by Kallionpää e al [15] (GEO: GSE30211) and he 454 PBMC samples by Fe ei a e al [27] (A ay Exp ess: E-MTAB-1724), as desc ibed in Me hods, using he published p e-p ocessed da ase s and sample in o ma ion based on pe sonal communica ions wi h Fe ei a e al. All p obes (a)o he highes -in ensi y exons mapping o genes (b) o e lapping wi h he 77 genes we e summed. (a) Black, s ongly en e o i us-posi i e blood samples; whi e, weakly en e o i us-posi i e blood samples; g ey, en e o i us-nega i e blood samples. (a,b) P eSe o, samples collec ed om be o e se ocon e sion om child en wi h au oan ibody posi i i y o ype 1 diabe es (in a,n=22;inb,n= 65); Pos Se o, samples collec ed a e se ocon e sion om child en wi h au oan ibody posi i i y o ype 1 diabe es (in a,n=169;bn=84),Aab − , samples collec ed om au oan ibody-nega i e child en (in a,n=165;inb,n=305).(c)Venn diag am showing he o e laps be ween he 77 genes p esen in he en e o i us-induced signa u e and up egula ed in all h ee in i o en e o i us in ec ions; genes up egula ed be o e o a e se ocon e sion in au oan ibody-posi i e child en based on he esul s by Kallionpää e al [15] and he 225 in e e on-inducible genes de ec ed by Fe ei a e al [27] 386 Diabe ologia (2018) 61:381–388 we e de ec ed, bu changes implying he ac i a ion o adap i e immune esponses we e obse ed. En e o i us-associa ed down egula ion o ansc ip ion, ansla ion o mRNA p ocessing-associa ed genes was obse ed in six child en, al- hough he indi idual p obes and genes mapping o hese ca - ego ies a ied be ween indi iduals. Up egula ion o in e e on esponse genes and down egula ion o ansla ion-associa ed genes we e also de ec ed in human PBMCs in ec ed in i o wi h h ee di e en en e o i uses. Finally, up egula ion o genes associa ed wi h in e e on esponses was he common ea u e be ween en e o i us-induced blood ansc ip omic changes in ou child en a isk o de eloping ype 1 diabe es and in i o en e o i us-in ec ed human panc ea ic isle s [10, 26], c ea ing a link be ween he i us–hos in e play in blood and in he panc eas. We buil an en e o i us-induced signa u e co e ing 339 genes p esen in he peaking clus e s o he ou child en wi h clea indica ions o in e e on esponse ac i a ion, and a mo e selec i e signa u e o 77 genes addi ionally up egula ed in human PBMCs in ec ed in i o wi h h ee di e en en e o i- uses. Bo h signa u es sepa a ed he ou s ongly en e o i us- posi i e samples om he o he samples in he ull mic oa ay da ase published by Kallionpää e al [15]. The en e o i us-associa ed signa u e showed only mode - a e o e lap wi h he up egula ed genes in Kallionpää e al [15] and Fe ei a e al [27], and could no di e en ia e be ween child en who de eloped ype 1 diabe es au oan ibodies o clinical ype 1 diabe es and au oan ibody-nega i e child en in hose s udies (ESM Fig. 3). Al hough ac i a ion o in e e - on signalling has been shown o p ecede he de elopmen o au oimmuni y in child en a isk o ype 1 diabe es, ou esul s indica e di e ences be ween en e o i us-associa ed and ype 1 diabe es-associa ed in e e on signals. The ou child en wi h clea signs o in e e on esponse ac i a ion included wo pe sis en ly au oan ibody-nega i e child en, one child who la e became posi i e o mul iple ype 1 diabe es au oan ibodies and one au oan ibody-posi i e child who la e de eloped clinical ype 1 diabe es. Wi h he limi ed numbe o child en a ailable o he cu en s udy, and he signi ican amoun o he e ogenei y in en e o i us-associa ed changes be ween he child en, i is no possible o d aw con- clusions ega ding connec ions be ween en e o i us in ec ions and ype 1 diabe es. The e a e se e al ac o s ha can explain he obse ed he e ogenei y in he en e o i us-associa ed esponses. Fi s , ea lie in i o s udies ha e shown ha he magni ude o in e e on esponse induc ion in PBMCs a ies signi ican ly be ween di e en en e o i uses [14]. Second, he apid kine - ics o an i i al immune esponses can be a sou ce o signi i- can he e ogenei y when cha ac e ising en e o i us-associa ed blood ansc ip omic changes in ollow-up s udies wi h long sampling in e als, al hough en e o i us RNA can be de ec ed in blood o only a ew days du ing he acu e phase o in ec ion. Thi d, hos esponses o acu e in ec ions caused by di e en i uses can be simila , and some imes he di e - gences be ween i uses a e be e explained by he di e en magni udes o he e ec han by he ac ual genes esponding o in ec ion [23,29]. Al hough ou en e o i us-induced signa- u e has a high o e lap wi h in i o en e o i us-induced changes in human PBMCs and panc ea ic isle s, we canno conclude ha hese changes a e uniquely obse ed a e in ec- ion wi h en e o i uses. Finally, al hough en e o i us in ec- ions a e o en asymp oma ic, clinical symp oms we e epo - ed o i e o he se en child en less han a week be o e he collec ion o he en e o i us-posi i e blood samples. Th ee child en we e su e ing om e e a ound he ime o en e o i us-posi i e sample collec ion, including one child also su e ing om conjunc i i is, and common cold-like symp oms we e epo ed o wo child en. In e es ingly, he h ee child en wi h e e a ound he ime o en e o i us- posi i e sample collec ion we e s ongly en e o i us-posi i e based on quan i a i e RT-PCR and had clea signs o in e e - on esponse ac i a ion associa ed wi h he en e o i us- posi i e blood sample. Despi e he limi a ions o he cu en s udy, i p o ides a s a ing poin o unde s anding he indi idual esponses o en e o i us in ec ions in i o, and how hese esponses a e e lec ed in he mRNA exp ession p o iles in whole blood. Fu he longi udinal s udies wi h la ge coho s, sho e sam- pling in e als and be e knowledge o he ac ual i us s ains in ec ing he indi iduals will p o ide deepe insigh s in o he associa ions be ween en e o i us in ec ions and ype 1 diabe es. Acknowledgemen s The au ho s a e g a e ul o he DIPP amilies o hei pa icipa ion and he s a o he DIPP s udy o wo king wi h he amilies and ob aining he samples o he s udy. We hank O. Simell, he hono a y p incipal in es iga o o he DIPP s udy, o his wo k. We also hank R.C. Fe ei a (Uni e si y o Camb idge, Camb idge, UK) and co-au ho s o sha ing addi ional me ada a o hei samples (E-MTAB- 1724 [27]) wi h us. Da a a ailabili y The da ase s analysed du ing he cu en s udy a e included in his published a icle and i s supplemen a y in o ma ion iles (www.b k. i/ esea ch/compu a ional-biomedicine/1234-2)o a e a ailable om he GEO eposi o y (accession GSE30211). Funding This wo k was inancially suppo ed by he Eu opean Resea ch Council (ERC) (decision numbe 677943), JDRF (g an s 17- 2013-533 and 2-2013-32), he Academy o Finland (Cen e o Excellence in Molecula Sys ems Immunology and Physiology Resea ch, 2012– 2017, decision numbe 250114 and g an s 287423, 288671, 292482, 292335, 294337, 296801 and 304995), he Eu opean Union’s Ho izon 2020 esea ch and inno a ion p og amme (decision numbe 675395), Tekes, he Finnish Funding Agency o Inno a ion (1877/31/2016), he Sig id Jusélius Founda ion, he Y jö Johansson Founda ion, he Finnish Diabe es Resea ch Founda ion, he Reino Lah ika i Founda ion, he Eu opean Commission (Pe sis en Vi us In ec ion in Diabe es Ne wo k Diabe ologia (2018) 61:381–388 387 [PEVNET] F ame P og amme 7, con ac numbe 261441) and he Paulo Founda ion. Duali y o in e es The au ho s decla e ha he e is no duali y o in e - es associa ed wi h his manusc ip . Con ibu ion s a emen LTTA and LLE planned he da a analyses. LTTA was esponsible o analysing he da a and pa icipa ed in w i ing he manusc ip and p epa ing he igu es. NL pa icipa ed in planning he da a analyses, in e p e ed he esul s and pa icipa ed in w i ing he man- usc ip and p epa ing he igu es. HK pa icipa ed in planning he da a analyses and in e p e ing he esul s. MKJ pa icipa ed in analysing he da a and p epa ing he igu es. JM, JT, JI, MK and RV p o ided and in e p e ed he clinical in o ma ion o he s udy child en. HH and SO we e esponsible o he i us analysis wi hin he s udy, p o ided he in i o in ec ion da a and con ibu ed o he ini ia ion and design o he s udy. RL and LLE ini ia ed and designed he s udy, supe ised he s udy and pa icipa ed in in e p e a ion o he esul s and w i ing he manusc ip . All au ho s edi ed/ e ised and app o ed he inal e sion o he manu- sc ip . LLE is he gua an o o his wo k. Open Access This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0 In e na ional License (h p:// c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided you gi e app o- p ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he C ea i e Commons license, and indica e i changes we e made. Re e ences 1. Yeung WC, Rawlinson WD, C aig ME (2011) En e o i us in ec- ion and ype 1 diabe es melli us: sys ema ic e iew and me a- analysis o obse a ional molecula s udies. BMJ 342:d35 2. Lai inen OH, Honkanen H, Pakkanen O e al (2014) Coxsackie i us B1 is associa ed wi h induc ion o be a-cell au oimmuni y ha po - ends ype 1 diabe es. Diabe es 63:446–455 3. Beye lein A, Donnachie E, Je gens S, Ziegle AG (2016) In ec ions in ea ly li e and de elopmen o ype 1 diabe es. JAMA 315:1899–1901 4. Hyo y H (2016) Vi uses in ype 1 diabe es. Pedia Diabe es 17(Suppl 22):56–64 5. Ylipaas o P, Klingel K, Lindbe g AM e al (2004) En e o i us in- ec ion in human panc ea ic isle cells, isle opism in i o and ecep o in ol emen in cul u ed isle be a cells. Diabe ologia 47: 225–239 6. K og old L, Edwin B, Buanes T e al (2015) De ec ion o a low- g ade en e o i al in ec ion in he isle s o Lange hans o li ing pa- ien s newly diagnosed wi h ype 1 diabe es. Diabe es 64:1682–1687 7. Honkanen H, Oika inen S, Nu minen N e al (2017) De ec ion o en e o i uses in s ools p ecedes isle au oimmuni y by se e al mon hs: Possible e idence o slowly ope a ing mechanisms in i us-induced au oimmuni y. Diabe ologia 60:424–431 8. San in I, Eizi ik DL (2013) Candida e genes o ype 1 diabe es modula e panc ea ic isle in lamma ion and be a-cell apop osis. Diabe es Obes Me ab 15(Suppl 3):71–81 9. Ash on MP, Eugs e A, Wal he D e al (2016) Incomple e immune esponse o coxsackie B i uses associa es wi h ea ly au oimmuni y agains insulin. Sci Rep 6:32899 10. Domsgen E, Lind K, Kong L e al (2016) An IFIH1 gene polymo - phism associa ed wi h isk o au oimmuni y egula es canonical an i i al de ence pa hways in coxsackie i us in ec ed human pan- c ea ic isle s. Sci Rep 6:39378 11. Schul e BM, Gielen PR, Ke s-Rebel ED e al (2016) En e o i us exposu e uniquely disc imina es ype 1 diabe es pa ien s wi h a homozygous om a he e ozygous melanoma di e en ia ion- associa ed p o ein 5/in e e on induced wi h helicase C domain 1 A946T geno ype. Vi al Immunol 29:389–397 12. Oika inen S, Ma iskainen M, Tau iainen S e al (2011) En e o i us RNA in blood is linked o he de elopmen o ype 1 diabe es. Diabe es 60:276–279 13. Oika inen S, Tau iainen S, Hobe D e al (2014) Vi us an ibody su ey in di e en eu opean popula ions indica es isk associa ion be ween coxsackie i us B1 and ype 1 diabe es. Diabe es 63:655– 662 14. Hamalainen S, Nu minen N, Ahl o s H e al (2014) Coxsackie i us B1 e eals s ain speci ic di e ences in plasmacy oid dend i ic cell media ed immunogenici y. J Med Vi ol 86:1412–1420 15. Kallionpää H, Elo LL, Laajala E e al (2014) 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 16. Kupila A, Muona P, Simell T e al (2001) Feasibili y o gene ic and immunological p edic ion o ype I diabe es in a popula ion-based bi h coho . Diabe ologia 44:290–297 17. Honkanen H, Oika inen S, Pakkanen O e al (2013) Human en e o- i us 71 s ains in he backg ound popula ion and in hospi al pa- ien s in inland. J Clin Vi ol 56:348–353 18. Piccolo SR, Wi he s MR, F ancis OE, Bild AH, Johnson WE (2013) Mul ipla o m single-sample es ima es o ansc ip ional ac- i a ion. P oc Na l Acad Sci U S A 110:17778–17783 19. Huang da W, She man BT, Lempicki RA (2009) Sys ema ic and in eg a i e analysis o la ge gene lis s using DAVID bioin o ma ics esou ces. Na P o oc 4:44–57 20. Rusino a I, Fo s e S, Yu S e al (2013) In e e ome 2.0: An up- da ed da abase o anno a ed in e e on- egula ed genes. Nucleic Acids Res 41:D1040–D1046 21. Du P, Kibbe WA, Lin SM (2008) Lumi: a pipeline o p ocessing illumina mic oa ay. Bioin o ma ics 24:1547–1548 22. Suomi T, Seyednas ollah F, Jaakkola MK, Faux T, Elo LL (2017) ROTS: an R package o ep oducibili y-op imized s a is ical es - ing. PLoS Compu Biol 13:e1005562 23. Zhai Y, F anco LM, A ma RL e al (2015) Hos ansc ip ional esponse o in luenza and o he acu e espi a o y i al in ec- ions—a p ospec i e coho s udy. PLoS Pa hog 11:e1004869 24. Richa dson SJ, Willcox A, Bone AJ, Foulis AK, Mo gan NG (2009) The p e alence o en e o i al capsid p o ein p1 immuno- s aining in panc ea ic isle s in human ype 1 diabe es. Diabe ologia 52:1143–1151 25. Richa dson SJ, Lee e P, Bone AJ, Foulis AK, Mo gan NG (2013) Exp ession o he en e o i al capsid p o ein VP1 in he isle cells o pa ien s wi h ype 1 diabe es is associa ed wi h induc ion o p o ein kinase R and down egula ion o mcl-1. Diabe ologia 56:185–193 26. Ylipaas o P, Ku lu B, Rasilainen S e al (2005) Global p o iling o coxsackie i us- and cy okine-induced gene exp ession in human panc ea ic isle s. Diabe ologia 48:1510–1522 27. Fe ei a RC, Guo H, Coulson RM e al (2014) A ype I in e e on ansc ip ional signa u e p ecedes au oimmuni y in child en gene - ically a isk o ype 1 diabe es. Diabe es 63:2538–2550 28. Alidjinou EK, Sane F, Engelmann I, Hobe D (2013) Se um- dependen enhancemen o coxsackie i us B4-induced p oduc ion o IFNalpha, IL-6 and TNFalpha by pe iphe al blood mononuclea cells. J Mol Biol 425:5020–5031 29. Hu X, Yu J, C osby SD, S o ch GA (2013) Gene exp ession p o iles in eb ile child en wi h de ined i al and bac e ial in ec ion. P oc Na l Acad Sci U S A 110:12792–12797 388 Diabe ologia (2018) 61:381–388