Host immune genetic variations influence the risk of developing acute myeloid leukaemia : results from the NuCLEAR consortium

Sánchez-Maldonado e al. Blood Cance Jou nal (2020) 10:75 h ps://doi.o g/10.1038/s41408-020-00341-y Blood Cance Jou nal ARTICLE Open Access Hos immune gene ic a ia ions influence he isk o de eloping acu e myeloid leukaemia: esul s om he NuCLEAR conso ium J. M. Sánchez-Maldonado 1,2,3 ,D.Campa 4 , J. Sp inge 5 , J. Badiola 2,3 ,Y.Niazi 6,7,8 , A. Moñiz-Díez 1,2,3 , F. He nández-Mohedo 2,3 ,P.González-Sie a 2,3 ,R.Te Ho s 9 ,A.Macauda 4,10 ,S.B ezina 11 ,C.Cunha 12,13 ,M.Lackne 14 , M. A. López-Ne o 15 , L. Fianchi 16 ,L.Pagano 17 , E. López-Fe nández 2,3 ,L.Po enza 17 ,M.Luppi 17 ,L.Mo a alla 2,3 , J. J. Rod íguez-Se illa 18 ,J.E.Fonseca 19,20 ,M.To mo 21 ,C.Solano 21 ,E.Cla e o 2 , A. Rome o 2 ,Y.Li 9,22 , C. Lass-Flö l 14 , H. Einsele 5 ,L.Vazquez 23 ,J.Loe fle 5 ,K.Hemminki 6,24,25 ,A.Ca alho 12,13 ,M.G.Ne ea 9,26 ,A.Gsu 11 ,C.Dumon e 27 , F. Canzian 10 ,A.Fö s i 6,7,8 ,M.Ju ado 1,2,3 and J. Sainz 1,2,3,28 Abs ac The pu pose o his s udy was o conduc a wo-s age case con ol associa ion s udy including 654 acu e myeloid leukaemia (AML) pa ien s and 3477 con ols asce ained h ough he NuCLEAR conso ium o e alua e he e ec o 27 immune- ela ed single nucleo ide polymo phisms (SNPs) on AML isk. In a pooled analysis o coho s udies, we ound ha ca ie s o he IL13 s1295686A/A geno ype had an inc eased isk o AML (P Co =0.0144) whe eas ca ie s o he VEGFA s25648T allele had a dec eased isk o de eloping he disease (P Co =0.00086). In addi ion, we ound an associa ion o he IL8 s2227307 SNP wi h a dec eased isk o de eloping AML ha emained ma ginally significan a e mul iple es ing (P Co =0.072). Func ional expe imen s sugges ed ha he e ec o he IL13 s1295686 SNP on AML isk migh be explained by i s ole in egula ing IL1Ra sec e ion ha modula es AML blas p oli e a ion. Likewise, he p o ec i e e ec o he IL8 s2227307 SNP migh be media ed by TLR2-media ed immune esponses ha a ec AML blas iabili y, p oli e a ion and chemo esis ance. Despi e he po en ial in e es o hese esul s, addi ional unc ional s udies a e s ill wa an ed o un a el he mechanisms by which hese a ian s modula e he isk o AML. These findings sugges ed ha IL13,VEGFA and IL8 SNPs play a ole in modula ing AML isk. In oduc ion Acu e Myeloid Leukaemia (AML) is a common hae- ma ological malignancy cha ac e ised by he clonal ans o ma ion o haema opoie ic p ecu so s ha al e no mal hema opoie ic cell g ow h and di e en ia ion 1 . Epidemiological s udies sugges ed ha AML onse can be igge ed by mul iple ac o s including age, sex, li es yle, exposu e o chemicals and a numbe o blood and con- geni al diso de s 2 . Howe e , he biological mechanisms unde lying AML ae iology emain la gely elusi e. E en hough cy ogene ic analysis ha e allowed he s a ifica ion o AML pa ien s in o a ou able, in e media e and un a- ou able classes and has imp o ed ou abili y o p edic clonal e olu ion and disease p og ession 3 , many AML pa ien s (~ 45%) ha e a no mal ka yo ype, which sugges s ha addi ional gene ic al e a ions a e needed o de elop he disease. Sequencing s udies iden ified genes equen ly mu a ed in AML, some o which p edic poo p ognosis (NPM1 w /FLT3-ITD high ,RUNX1,ASXL1 and TP53) 4 . © The Au ho (s) 2020 Open Access This a icle is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional License, which pe mi s use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any medium o o ma , as long as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o heC ea i eCommons license, and indica e i changes we e made. 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 a c edi line o he ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons license and you in ended use is no pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will need o ob ain pe mission di ec ly om he copy igh holde . To iew a copy o his license, isi h p://c ea i ecommons.o g/licenses/by/4.0/. Co espondence: J Sainz ([email p o ec ed]) 1 Genomic Oncology A ea, GENYO, Cen e o Genomics and Oncological Resea ch: Pfize / Uni e si y o G anada / Andalusian Regional Go e nmen , PTS G anada, G anada, Spain 2 Hema ology depa men , Vi gen de las Nie es Uni e si y Hospi al, G anada, Spain Full lis o au ho in o ma ion is a ailable a he end o he a icle These au ho s con ibu ed equally: M. Ju ado, J. Sainz Blood Cance Jou nal 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,; Fu he mo e, i is inc easingly e iden ha hos immuni y migh also be implica ed in AML isk and su i al 5 . AML blas s ac i a e immunosupp essi e mechanisms o e ade he immune sys em whe eas immune esponse changes induced by he gu mic obio a can also influence he an i- leukaemic e ec s o immune cells 6 . In addi ion, he e fi- cacy o allogeneic s em cell ansplan a ion (SCT) in e adica ing AML is linked o he appea ance o he g a - e sus-leukaemia e ec , media ed by he ecogni ion o majo his ocompa ibili y an igens p esen in malignan blas s by T cells 7 . Likewise, he disappea ance o hese ci cula ing T cells ecognising AML o he loss o cos i- mula o y (CD28/CD80, ICAM-1/CD11a) o inhibi o y in e ac ions (PD-1/PDL-1) e en ually leads o elapse 8 and he in usion o dono -de i ed CD8+memo y T cells induces emission in pa ien s who elapsed ollowing allogeneic SCT 9 . Conside ing ha a ound wo- hi ds o AML pa ien s elapse wi hin he fi s 18 mon hs a e fi s -line he apy, clinical ials a e ying o assess he e ficacy o immuno he apies in AML and o un a el he in e play be ween he immune sys em and AML blas s. Conside ing he aspec s de ailed abo e, he pu pose o his s udy was o conduc a wo-s age case con ol asso- cia ion s udy including 654 AML pa ien s and 3477 con ols asce ained h ough he NuCLEAR conso ium o e alua e whe he 27 single nucleo ide polymo phisms (SNPs) wi hin he IL4,IL8,IL8RB (CXCR2),IL12A,IL12B, IL13,IFNG,IFNGR2,CCR5,MIF and VEGFA loci influ- ence he isk o de eloping AML. We also decided o in es iga e he co ela ion o selec ed SNPs wi h se um s e oid ho mone le els and hei ole in modula ing immune esponses a e s imula ion o whole blood, pe iphe al mononuclea cells (PBMCs) and mac ophages wi h lipopolysaccha ide (LPS), phy ohemagglu inin (PHA), Pam3Cys and CpG. Ma e ial, subjec s and me hods S udy design and s udy popula ions We conduc ed a wo-s age gene ic associa ion s udy o assess whe he 27 unc ional single nucleo ide poly- mo phisms (SNPs) wi hin hos immuni y- ela ed genes could influence AML isk. The disco e y popula ion consis ed o 2027 Eu opean subjec s (338 AML pa ien s and 1689 heal hy con ols). AML pa ien s we e diagnosed by expe ienced clinicians and asce ained h ough he iN e na ional Conso ium o LEukaemiA Resea ch (NuCLEAR; Table 1). A se o AML pa ien s we e ec ui ed om 2 Spanish medical ins i u ions (Vi gen de las Nie es Uni e si y Hospi al, G anada and Hospi al o Salamanca, Salamanca), he Uni e si y o Wü zbu g (Wü zbu g, Ge many) and he Uni e si y o Innsb uck (Innsb uck, Aus ia) 10 . Heal hy con ols included 667 Spanish blood dono s om he REPAIR conso ium 11 , 1000 Ge man con ols came om he Heinz-Nixdo Recall (HNR) s udy 12 and 22 dono s o allogeneic s em cell ansplan a ion om he Medical Uni e si y o Innsb uck (Innsb uck, Aus ia). In acco dance wi h he Decla a ion o Helsinki, all s udy pa icipan s p o ided hei w i en in o med consen o pa icipa e in he s udy and he e hical commi ees o all pa icipa ing cen es and hospi als app o ed he s udy. DNA ex ac ion, SNP selec ion c i e ia and geno yping Genomic DNA om all indi iduals was ex ac ed om sali a o blood samples using he O agen®-DNA Sel - Collec ion ki (O agene) o he Maxwell®16 Blood DNA Pu ifica ion ki (P omega) acco ding o manu ac u e ’s ins uc ions. SNP selec ion c i e ia we e based on p e ious associa ions wi h haema ological malignancies (AML, ALL, CML, CLL and non-Hodgkin lymphomas) o solid umou s and clinical ela ed pa ame e s (g a e sus hos disease, whole blood leucocy e coun s, an h opome ic measu es, e c.) bu also acco ding o hei unc ionali y in Haplo eg (h ps://pubs.b oadins i u e.o g/mammals/haplo eg/ haplo eg.php), Regulome (h ps://www. egulomedb.o g/ Table 1 Demog aphic and clinical cha ac e is ics o AML pa ien s and heal hy con ols. Demog aphic cha ac e is ics Disco e y Popula ion (n=2027) 338 AML cases and 1689 heal hy con ols Replica ion Popula ion (n=2104) 316 AML cases and 1788 heal hy con ols O e all Popula ion (n=4131) 654 AML cases and 3477 heal hy con ols AML cases Age (yea s) 55.19 ± 15.12 56.91 ± 17.25 56.02 ± 16.20 Sex a io (male/ emale) 1.13 (179/159) 1.29 (178/138) 1.20 (356/297) Coun y o o igin Spain 257 97 354 Ge many 26 74 100 I aly –145 145 Aus ia 55 –55 P esen a ion de no o (n, %) 324 (95.86) 285 (90.19) 609 (93.12) Seconda y (n, %) 14 (04.14) 31 (09.81) 45 (06.88) Heal hy con ols Age (yea s)* 56.10 ± 9.55 42.76 ± 11.76 49.57 ± 11.33 Sex a io (male/ emale) 1.07 (871/818) 0.91 (848/937) 0.98 (1719/1754) Coun y o o igin Spain 667 507 1174 Ge many 1000 1087 2087 I aly –194 194 Aus ia 22 –22 Da a a e means ± s anda d de ia ion o pe cen age (%). A se o 99 pa ien s (39 and 61 om he disco e y and eplica ion coho s, espec i ely) could no be classified acco ding o he FAB classifica ion. AML acu e myeloid leukaemia *Age was no a ailable in a se o Ge man con ols included in he disco e y (n=1000) and eplica ion coho s (n=1068). Sánchez-Maldonado e al. Blood Cance Jou nal (2020) 10:75 Page 2 o 11 Blood Cance Jou nal Table 2 Lis o selec ed ma ke s. Gene name Gene symbol_SNP dbSNP s# Risk allele Repo ed associa ions wi h haema ological malignancies, solid umou s, pa ien su i al and di e en clinical pa ame e s (GVHD, blood cell coun s, BMI, e c.) Re s. In e leukin 4 (IL4) IL4_-1098 s2243248 G Associa ed wi h inc eased isk o T-cell lymphomas and HBV eac i a ion in i uximab- ea ed pa ien s wi h non-Hodgkin lymphoma (NHL) 1,2 IL4_IVS2-1443 s2243268 C Associa ed wi h IL4 le els in whole blood and o he issues. Maps among p omo e his one ma ks in bone ma ow de i ed mesenchymal s em cells. 3 In e leukin 8 (IL8) IL8_-251 s4073 A Associa ed wi h NHL isk and p edic o o su i al in ollicula lymphoma. Associa ed wi h IL8 a bo h ansc ip ional and ansla ional le els and wi h inc eased ansmig a ion o p ima y neu ophils. Maps among p omo e and enhance his one ma ks in mul iple p ima y immune cells, hema opoie ic s em cells and bone ma ow de i ed cul u ed mesenchymal s em cells. Regulome sco e 2b. 4–7 IL8_ IVS1 +230 (+396) s2227307 G Associa ed wi h IL8 a bo h ansc ip ional and ansla ional le els and wi h inc eased ansmig a ion o p ima y neu ophils. Associa ed wi h ollicula lymphoma pa ien su i al 6,7 CXC-Chemokine ecep o 2 (IL8RB) CXCR2_Ex3-1010 s1126580 G Associa ed wi h CXCR2 and CXCR1 le els in whole blood (GTEx). Associa ed wi h sho e su i al in di use la ge B-cell lymphoma and suscep ibili y o bile duc cance 8,9 In e leukin 12 alpha (IL12A) IL12A_Ex7 +277 s568408 A Binding mo i s o TFE and SIX5 In e leukin 12 be a (IL12B) IL12B_Ex8 +159 (+1188) s3212227 C Associa ed wi h isk o solid umou s and su i al o ollicula lymphoma pa ien s 5,10 In e leukin 13 (IL13) IL13_-1069 s1800925 T Associa ed wi h suscep ibili y o glioma, glioblas oma mul i o me and CRC and an inc eased isk o leukopenia in me as a ic enal cell ca cinoma pa ien s. Regulome sco e 2b 11–13 IL13_Ex4 +98 s20541 T Associa ed wi h suscep ibili y o mul iple cance s including NHL, CRC and glioma. Associa ed wi h adia ion-induced oxici y ollowing ea men o non-small cell lung cance . Regulome sco e 3a 11,14–17 IL13_IVS3-24 s1295686 A HBV eac i a ion in i uximab- ea ed pa ien s wi h NHL. Regulome sco e 3a 2 In e e on gamma (IFN-γ) IFNG_-1615 s2069705 C Cy ogene ic and molecula esponse wi h Ima inib in CML pa ien s 18 IFNG_IVS3 +284 (+2109) s1861494 C Cy ogene ic and molecula esponse wi h Ima inib in CML pa ien s 18 In e e on gamma ecep o 2 (IFN-γR2) IFNGR2_Ex7-128 s1059293 T Associa ed wi h B eas cance isk. Regulome sco e 1 19 IFNGR2_Ex2-16 s9808753 G Associa ed wi h IFNGR2 le els in whole blood (GETx) and isk o NHL 17 C-C chemokine ecep o ype 5 (CCR5) CCR5_IVS1 +246 s1799987 G Associa ed wi h CCR2 le els in whole blood (GTEx) and wi h a mo e a ou able MRD s a us in child en wi h B-p ecu so acu e lymphoblas ic leukaemia (ALL). Regulome sco e 3a 20 CCR5_IVS1 +151 s2734648 T Associa ed wi h CCR1, CCR2 and CCR5 le els in whole blood (GTEx). Regulome sco e 3a 21,22 Mac ophage mig a ion inhibi o y ac o (MIF) MIF_-173 s755622 G Sánchez-Maldonado e al. Blood Cance Jou nal (2020) 10:75 Page 3 o 11 Blood Cance Jou nal Table 2 con inued Gene name Gene symbol_SNP dbSNP s# Risk allele Repo ed associa ions wi h haema ological malignancies, solid umou s, pa ien su i al and di e en clinical pa ame e s (GVHD, blood cell coun s, BMI, e c.) Re s. Associa ed wi h MIF and MIF-AS1 le els in whole blood (GTEx). Associa ed wi h solid and non- solid umou s such as childhood ALL. Maps nea mul iple p omo e and enhance his one ma ks in mul iple issues and all immune cell ypes and hema opoie ic s em cells. Vascula Endo helial G ow h Fac o alpha (VEGFα) VEGFA_-2578 s699947 A Associa ed wi h disease p og ession in ch onic myeloid leukaemia (CML) and an inc eased isk o hy oid cance and me as asis in men. Associa ed wi h su i al in ad anced-s age non-small-cell lung cance . Regulome sco e 2b 23–25 VEGFA_-7 s25648 T Associa ed wi h p ognosis in AML and CLL pa ien s. Associa ed wi h he isk o de eloping acu e GVHD a e allogeneic-s em cell ansplan a ion. Associa ed wi h he isk o de eloping solid umou s such as bladde cance and su i al o pa ien s wi h enal cell ca cinoma. Maps nea mul iple p omo e and enhance his one ma ks in mul iple issues and all immune cell ypes and hema opoie ic s em cells. 26–29 VEGFA_IVS2 +1378 s3024994 T Associa ed wi h a educed isk o bladde cance . Mul iple p omo e his one ma ks in immune cells and hema opoie ic s em cells. 26 VEGFA_IVS7-919 s3025035 T Associa ed wi h ecu ence o hepa ocellula ca cinoma a e ansplan a ion and su i al o pa ien s wi h non-small cell lung cance . Regulome sco e 3a 30,31 VEGFA_6112 s2146323 A Al e s a binding si e o P53. Regulome sco e 2b VEGFA_IVS-99 s3024997 A Associa ed wi h VEGFA mRNA exp ession in human monocy es 32 VEGFA_IVS7 +763 s3025030 C Maps nea enhanced his one ma ks in 9 issues VEGFA_5530 s998584 T Associa ed wi h whole blood leukocy e coun , adiponec in, HDL choles e ol and iglyce ides le els. Associa ed wi h BMI and wais ci cum e ence 33,34 VEGFA_5958bp 3’o STP s6899540 C Maps nea enhanced his one ma ks in mul iple immune cell ypes including p ima y monocy es, p ima y B cells, NK cells, neu ophils, hema opoie ic s em cells and bone ma ow de i ed mesenchymal s em cells. Al e s binding o 7 mo i s (AP-1, BCL, Nkx2, Pax5…). VEGFA_6119bp 3’o STP s6900017 T Maps nea enhanced his one ma ks in mul iple immune cell ypes including p ima y monocy es, p ima y B cells, hema opoie ic s em cells and bone ma ow de i ed mesenchymal s em cells. Al e s binding o 11 mo i s (AP-1, p300, HDAC2, NFAT,…). eQTL da a we e ga he ed om he GTEx po al (h ps://g expo al.o g/home/) and Wes a e al. 45 . Regulome sco e 1 (Eq l +TF binding/DNase peak), 2b (TF binding+any mo i +DNase Foo p in +DNase peak) and 3a (TF binding+any mo i +DNase peak) we e conside ed as selec ion c i e ia. Re e ences a e included as Supplemen a y Ma e ial. SNP single nucleo ide polymo phisms, Allo-SCT allogeneic s em cell ansplan a ion, OR odds a io, CI confidence in e al, NHL non-Hodgkin lymphoma, CML ch onic myeloid leukaemia, CRC colo ec al cance , AIDS acqui ed immune deficiency synd ome. Sánchez-Maldonado e al. Blood Cance Jou nal (2020) 10:75 Page 4 o 11 Blood Cance Jou nal egulome-sea ch/), Blood eQTL b owse s (h ps://gene- ne wo k.nl/bloodeq lb owse / and h ps://g expo al.o g/ home/index.h ml), and linkage disequilib ium alues (Table 2and Supplemen a y Fig. 1). Geno yping was pe o med using KASP®p obes (LGC Genomics, Hoddesdon, UK) acco ding o p e iously epo ed p o ocols 13 . Fo quali y con ol, ∼5% o DNA samples we e andomly included as duplica es and conco dance be ween duplica e samples was ≥99.0%. AML cases and con ols we e andomly dis ibu ed in 384-well pla es and he pe son doing geno yping expe imen s did no know how AML cases and con ols we e dis ibu ed. S a is ical analysis De ia ion om Ha dy-Weinbe g Equilib ium (HWE) was es ed in he con ols by chi-squa e (χ 2 ). Logis ic eg ession adjus ed o sex and coun y o o igin was used o assess he associa ions o he SNPs wi h AML isk assuming log-addi i e, dominan and ecessi e models. Acco ding he M e me hod 14 , 24 o 27 SNPs we e inde- penden and, consequen ly, he s udy-wide significan h eshold was se o 0.0007 (0.05/24SNPs/3models). S a- is ical powe was calcula ed using Quan o ( .12.4) assuming a log-addi i e model o inhe i ance. Replica ion coho Fo eplica ion pu poses, he mos ele an findings (P< 0.05) we e eplica ed in a coho o 2104 subjec s (316 AML cases and 1788 heal hy con ols). AML cases we e ec ui ed om an independen Spanish medical ins i u- ion (Hospi al Gene al o Valencia, Valencia, Spain), om he Uni e si y Hospi al o Wü zbu g (Ge many) and om wo I alian medical ins i u ions (Uni e si à Ca olica del Sac o Cuo e, Rome and Uni e si y o Modena and Reggio Emilia, AOU Policlinico, Modena) be ween 2015 and 2017. Fi e hund ed and se en Spanish con ols we e blood dono s ec ui ed om he Blood T ans usion Cen e (CRTS, G anada-Alme ía), 194 I alian con ols om he REPAIR conso ium, 1068 Ge man con ols om a second and independen se o he Heinz-Nixdo Recall (HNR) s udy (Uni e si y Hospi al o Essen) and 19 dono s o allogeneic s em cell ansplan a ion om he Uni e si y o Wü zbu g (Ge many). The e hical com- mi ees o hese cen es app o ed he s udy. Func ional analysis o he hos immune- ela ed a ian s In o de o de e mine he biological unc ion o he mos ele an SNPs, cy okine p oduc ion in esponse o s imula ion was measu ed in he 500 Func ional Geno- mics coho om he Human Func ional Genomics P o- jec (HFGP; h p://www.human unc ionalgenomics.o g/). The A nhem-Nijmegen E hical Commi ee app o ed he s udy (42561.091.12) and biological specimens we e collec ed a e in o med consen was ob ained. We in es iga ed whe he any SNP was co ela ed wi h cy o- kine le els (IFNγ, IL1Ra, IL1β, IL6, IL8, IL10, TNFα, IL17, and IL22) a e s imula ion o pe iphe al blood mono- nuclea cells (PBMCs), whole blood o monocy e-de i ed mac ophages om 408 heal hy subjec s wi h LPS (1 o 100 ng/ml), PHA (10 μg/ml), Pam3Cys (10 μg/ml), and CpG (100 ng/ml). A e log ans o ma ion, linea eg ession analyses adjus ed o age and sex we e used o de e mine he co ela ion o selec ed SNPs wi h cy okine exp ession quan i a i e ai loci (cQTLs). All analyses we e pe o med using R so wa e (www. -p ojec .o g/). In o de o accoun o mul iple compa isons, we used a significan h eshold o 0.00006, i.e., he quo ien o 0.05/ (24 independen SNPs × 9 cy okines × 4 cell s imulan s). De ailed p o ocols o PBMCs isola ion, mac ophage di e en ia ion and s imula ion assays ha e been epo ed elsewhe e 15–17 . B iefly, PBMCs we e washed wice in saline and suspended in medium (RPMI 1640) supple- men ed wi h gen amicin (10 mg/ml), L-glu amine (10 mM) and py u a e (10 mM). PBMC s imula ions we e pe o med wi h 5×10 5 cells/well in ound-bo om 96- wells pla es (G eine ) o 24 h in he p esence o 10% human pool se um a 37 °C and 5% CO 2 . Supe na an s we e collec ed and s o ed in −20 °C un il used o ELISA. LPS (100 ng/ml), PHA (10μg/ml) and Pam3Cys (10 μg/ ml) and CpG (100 ng/ml) we e used as s imula o s o 24 o 48 h. Whole blood s imula ion expe imen s we e conduc ed using 100 μl o hepa in blood ha was added o a 48 well pla e and subsequen ly s imula ed wi h 400 μl o LPS and PHA (final olume 500ul) o 48 h a 37 °C and 5% CO 2 . Supe na an s we e collec ed and s o ed in −20 °C un il used o ELISA. Concen a ions o human TNFα, IFNγ, IL1β, IL1RA, IL6, IL8, IL10, IL17, and IL22 we e de e mined using specific comme cial ELISA ki s (PeliKine Compac , Ams e dam, o R&D Sys ems), in acco dance wi h he manu ac u e ’s ins uc ions. Co ela ion be ween s e oid ho mone le els and immuno egula o y SNPs Gi en he impac o s e oid ho mones in modula ing immune esponses, we also e alua ed he co ela ion o SNPs wi h se um le els o 7 s e oid ho mones (and os- enedione, co isol, 11-deoxy-co isol, 17-hyd oxy p o- ges e one, p oges e one, es os e one and 25 hyd oxy i amin D3) in a subse o subjec s wi hou ho monal eplacemen he apy o o al con acep i es (n=280). Comple e p o ocol de ails ha e been epo ed else- whe e 17 . S e oid ho mones we e analysed by liquid ch oma og aphy andem–mass spec ome y (LC–MS) a e p o ein p ecipi a ion and solid-phase ex ac ion as desc ibed in Te Ho s e al. 17 (see also Supplemen a y Ma e ial). Ho mone le els and geno yping da a we e a ailable o a o al o 406 subjec s. A e log ans o - ma ion, co ela ion be ween SNPs and se um s e oid Sánchez-Maldonado e al. Blood Cance Jou nal (2020) 10:75 Page 5 o 11 Blood Cance Jou nal ho mone le els was e alua ed using linea eg ession adjus ed o age and sex in R (h p://www. -p ojec .o g/). Significance h esholds we e se o 0.0003 (0.05/24 inde- penden SNPs/7 ho mones). Resul s This s udy was conduc ed in a disco e y popula ion comp ised o 338 AML pa ien s and 1689 heal hy con- ols. AML pa ien s had a simila age han con ols (55.19 ±15.12 s. 56.91±17.25) and showed a sligh ly inc eased male/ emale a io compa ed o heal hy con ols (1.13 [179/159] s. 1.07 [871/818]. Nine y fi e pe cen o he pa ien s had de no o AML whe eas he emaining 5% p esen ed seconda y disease e ol ing om a p eceding dysplasia (Table 1). The associa ion analysis o he disco e y popula ion e ealed ha 11 immuno egula o y SNPs we e associa ed wi h AML isk (P< 0.05; Table 3). We ound ha ca ie s o he IFNGR2 s1059293T allele o he IL4 s2243248G/G , IL13 s20541T/T ,IL13 s1295686A/A and VEGFA s998584T/T gen- o ypes showed an inc eased isk o de eloping he disease (OR Dom =1.51, P=0.0074; OR Rec =4.33, P=0.012; OR Rec =1.98, P=0.028; OR Rec =2.16, P=0.012; and OR Rec =1.40, P=0.034). In addi ion, we obse ed ha each copy o he IL4 s2243268C allele was associa ed wi h a 1.31- old inc eased isk o AML (OR Add =1.31, P=0.042). On he o he hand, we ound ha ca ie s o he IL8 s2227307G and VEGFA s25648T alleles had a significan ly dec eased isk o AML (OR Dom =0.70, P=0.012 and OR Dom =0.42, P=0.00002) whe eas each copy o he IL8 s4073A ,CCR5 s1799987G ,CCR5 s2734648T alleles was asso- cia ed wi h ~ 20–25% dec eased isk o AML (OR Add = 0.81, P=0.020; OR Add =0.82, P=0.043 and OR Add =0.75, P=0.0044). E en hough only he associa ion o he VEGFA s25648 SNP wi h a dec eased isk o de eloping AML emained significan a e co ec ion o mul iple es ing in he disco e y coho (P Co =0.0014), we ound ha he associa ion o IL8 s2227307 and IL13 s1295686 wi h AML isk was confi med in he eplica ion popula ion (OR Dom =0.74, P=0.040 and OR Dom =2.24, P=0.0051, espec i ely; Table 3). The pooled analysis including 4131 subjec s (654 AML cases and 3477 con ols) confi med ha ca ie s o he IL13 s1295686 geno ype had a significan ly inc eased isk o AML (OR Rec =2.18, P=0.0002, P Co =0.0144) whe eas ca ie s o he IL8 s2227307G allele had a dec eased isk o de eloping he disease ha emained ma ginally significan a e co ec ion o mul iple es ing (OR Dom =0.72, P= 0.0010, P Co =0.072). In e es ingly, al hough i was no s a is ically significan in he eplica ion popula ion likely due o he ela i ely limi ed powe , he pooled analysis also e ealed a s ong associa ion o he VEGFA s25648T allele wi h a dec eased isk o AML ha la gely su passed he s ingen s udy-wide significan h eshold (OR Dom =0.60, P=0.0000012, P Co =0.00086; Table 3). In an e o o de e mine he unc ional ele ance o hese polymo phisms, we pe o med in i o s imula ion expe imen s in a la ge coho o heal hy dono s o in es- iga e whe he IL8,IL13 and VEGFA SNPs could co ela e wi h le els o IFNγ,IL1Ra,IL1β, IL6, IL8, IL10, TNFα, IL17, and IL22 a e s imula ion o PBMCs, whole blood o monocy e-de i ed mac ophages wi h LPS, PHA, Pam3Cys, and CpG. These expe imen al s udies e ealed ha ca ie s o he IL8 s2227307T allele had inc eased le els o IL1βa e he s imula ion o PBMCs wi h Pam3Cys (P=0.00058; Fig. 1a). Al hough his associa ion did no su i e mul iple es ing co ec ion, hese esul s sugges ed ha his a ian migh ha e an impac on AML isk h ough he modula- ion o TLR2-immune esponses. In suppo o a unc ional ole o he IL8 s2227307 SNP in AML, i has been also epo ed ha his SNP ep esen s an eQTL o PF4V (Fig. 1b), a locus in ol ed in chemokine-media ed immune esponses. In e es ingly, al hough i nei he eached s a- is ical significance a e mul iple es ing co ec ion, we also ound a nega i e co ela ion be ween he IL13 s1295686A allele and le els o IL1Ra a e s imula ion o PBMCs wi h LPS (P=0.002; Fig. 1c), which sugges ed ha he IL13 locus migh play a ole in he pa hogenesis o AML likely h ough he modula ion o IL1Ra-media ed immune esponses. No co ela ion be ween selec ed SNPs and se um s e oid ho mone le els was ound sugges ing ha he unc ional e ec o hese ma ke s on he immune esponses was no media ed by s e oid ho mones. Discussion AML has been he objec o in es iga ions ha ha e demons a ed ha hos immuni y con ibu es o disease suscep ibili y. This s udy epo s o he fi s ime an associa ion o he IL13 s1295686 ,IL8 s2227307, and VEG- FA s25648 polymo phisms wi h AML isk. The associa ion o he IL13 and VEGFA SNPs wi h AML isk emained significan a e mul iple es ing co ec ion, whe eas he associa ion o IL8 s2227307 was no significan bu close o he mul iple es ing significance h eshold. These esul s sugges ed ha he IL13,VEGFA and IL8 loci migh be suscep ibili y ma ke s o AML. The IL13 gene is loca ed on ch omosome 5q31 and encodes o IL13, an immuno egula o y cy okine wi h pleio opic unc ions. Se e al SNPs ( s20541, s18000925 and s1295686) wi hin his gene ha e been consis en ly associa ed, a GWAS le el, wi h immune- ela ed dis- eases 18,19 and haema ological malignancies 20 . In his wo-s age case con ol associa ion s udy we ound a consis en and s a is ically significan associa ion o he IL13 s1295686A/A geno ype wi h an inc eased isk o de eloping AML ha sugges ed a ole o his locus in he pa hogenesis o he disease. Mechanis ically, we obse ed a nega i e co ela ion be ween he IL13 s1295686A allele and IL1Ra le els a e s imula ion o PBMCs wi h LPS Sánchez-Maldonado e al. Blood Cance Jou nal (2020) 10:75 Page 6 o 11 Blood Cance Jou nal Table 3 Associa ion o immuno egula o y SNPs and isk o de eloping acu e myeloid leukaemia. Gene name dbSNP s# Gene symbol_SNP Risk allele Disco e y Popula ion (n=2027) 338 AML cases and 1689 heal hy con ols P Co Replica ion Popula ion (n=2104) 316 AML cases and 1788 heal hy con ols O e all Popula ion (n=4131) 654 AML cases and 3477 heal hy con ols P Co OR (95% CI) a POR (95% CI) a POR (95% CI) a P IL4 s2243248 IL4_-1098 G4.33 (1.37–13.7) b 0.012 0.864 1.75 (0.31–9.67) b 0.52 3.09 (1.26–7.58) b 0.014 1.000 IL4 s2243268 IL4_IVS2-1443 C1.31 (1.01–1.69) 0.042 1.000 0.87 (0.66–1.15) 0.32 1.09 (0.90–1.31) 0.39 1.000 IL8 s4073 IL8_-251 A0.81 (0.67–0.97) 0.020 1.000 0.96 (0.76–1.20) 0.70 0.89 (0.78–1.01) 0.072 1.000 IL8 s2227307 IL8_ IVS1 +230 (+396) G0.70 (0.53–0.92) c 0.012 0.864 0.74 (0.56–0.99) c 0.040 0.72 (0.59–0.87) c 0.0010 0.072 IL8RB s1126580 CXCR2_Ex3-1010 A 0.82 (0.68–1.00) 0.044 1.000 1.05 (0.87–1.27) 0.61 0.95 (0.83–1.09) 0.49 1.000 IL12A s568408 IL12A_Ex7 +277 A 2.48 (1.00–6.15) b 0.050 1.000 1.07 (0.35–3.32) b 0.90 1.66 (0.84–3.27) b 0.14 1.000 IL12B s3212227 IL12B_Ex8 +159 (+1188) C 0.99 (0.79–1.25) 0.96 1.000 IL13 s1800925 IL13_-1069 T 1.01 (0.80–1.28) 0.93 1.000 IL13 s20541 IL13_Ex4 +98 T1.98 (1.08–3.65) b 0.028 1.000 1.75 (0.90–3.39) b 0.10 1.89 (1.21–2.94) b 0.0048 0.346 IL13 s1295686 IL13_IVS3-24 A2.16 (1.19–3.93) b 0.012 0.864 2.24 (1.27–3.93) b 0.0051 2.18 (1.45–3.26) b 0.0002 0.0144 IFNG s2069705 INFG_-1615 C 1.10 (0.90–1.35) 0.34 1.000 IFNG s1861494 INFG_IVS3 +284 (+2109) C 1.19 (0.96–1.47) 0.12 1.000 IFNGR2 s1059293 INFGR2_Ex7-128 T1.51 (1.11–2.05) c 0.0074 0.533 0.90 (0.67–1.21) c 0.48 1.16 (0.94–1.43) c 0.16 1.000 IFNGR2 s9808753 INFGR2_Ex2-16 G 1.05 (0.79–1.41) 0.73 1.000 CCR5 s1799987 CCR5_IVS1 +246 G0.82 (0.67–0.99) 0.043 1.000 0.98 (0.81–1.19) 0.85 0.90 (0.79–1.03) 0.13 1.000 CCR5 s2734648 CCR5_IVS1 +151 T0.75 (0.61–0.92) 0.0044 0.317 1.12 (0.93–1.36) 0.24 0.93 (0.81–1.06) 0.27 1.000 MIF s755622 MIF_-173 G 0.88 (0.67–1.14) 0.32 1.000 VEGFA s699947 VEGFA_-2578 A 1.05 (0.87–1.27) 0.58 1.000 VEGFA s25648 VEGFA_-7 T0.42 (0.29–0.62) c 0.00002 0.0014 0.79 (0.58–1.06) c 0.12 0.60 (0.47–0.75) c 0.000012 0.00086 VEGFA s3024994 VEGFA_IVS2 +1378 T 0.87 (0.59–1.29) 0.49 1.000 VEGFA s3025035 VEGFA_IVS7-919 T 1.08 (0.79–1.48) 0.62 1.000 VEGFA s2146323 VEGFA_6112 A 1.01 (0.82–1.23) 0.95 1.000 VEGFA s3024997 VEGFA_IVS-99 A 1.01 (0.83–1.24) 0.91 1.000 VEGFA s3025030 VEGFA_IVS7 +763 C 0.95 (0.72–1.26) 0.72 1.000 VEGFA s998584 VEGFA_5530 T1.40 (1.03–1.89) b 0.034 1.000 1.08 (0.80–1.47) b 0.61 1.24 (1.00–1.54) b 0.048 1.000 VEGFA s6899540 VEGFA_5958bp 3’o STP C 1.01 (0.77–1.32) 0.93 1.000 VEGFA s6900017 VEGFA_6119bp 3’o STP T 0.94 (0.68–1.30) 0.72 1.000 Associa ion es ima es we e adjus ed o sex and coun y o o igin. P< 0.05 in bold. Co ec ed P- alue was calcula ed by mul iplying he unadjus ed P- alue by he numbe o es s pe o med (n=72, 24 SNPs by 3 inhe i ance models es ed). SNP single nucleo ide polymo phisms, OR odds a io, CI confidence in e al. a Es ima es we e calcula ed acco ding o an addi i e model o inhe i ance. b Es ima es we e calcula ed acco ding o a ecessi e model o inhe i ance. c Es ima es we e calcula ed acco ding o a dominan model o inhe i ance. Sánchez-Maldonado e al. Blood Cance Jou nal (2020) 10:75 Page 7 o 11 Blood Cance Jou nal (P=0.002; Fig. 1c). Al hough his associa ion did no emain significan a e co ec ion o mul iple es ing, his finding suppo ed ou gene ic esul s sugges ing a ole o he IL13 s1295686 SNP in he pa hogenesis o AML. Conside ing ou esul s bu also hose om an ea ly epo ha demons a ed ha IL1Ra le els a e dec eased in AML pa ien s compa ed o con ols 21 , we hypo hesise ha he e ec o he IL13 s1295686A allele on AML isk migh be explained by i s ole in inhibi ing IL1Ra sec e- ion, likely h ough he inhibi ion o IL1Ra sec e ion om ei he AML blas s o heal hy cells. In line wi h his a gumen , i has been consis en ly epo ed ha IL1Ra inhibi s AML blas p oli e a ion 22 and ha i is associa ed wi h he immunosupp essi e e ec o he mesenchymal s em cells (MSCs) in he bone ma ow ha accoun s o mac ophage pola isa ion ( owa d he M2 pheno ype) and B cell di e en ia ion and su i al 23 . Al hough a his poin i is emp ing o specula e ha he IL13 s1295686A allele, which co ela es wi h lowe le els o IL1Ra sec e ion, migh ep esen a bioma ke wi h a po en ial benefi in AML by an agonising IL1 e ec s on blas p oli e a ion and blocking inflamma ion, we belie e ha addi ional unc ional expe imen s a e s ill wa an ed o explain he exac mechanism by which he IL13 s1295686 a ian influence he isk o AML. Ano he in e es ing finding o his s udy was he con- sis en associa ion o he IL8 s2227307T allele wi h a dec eased isk o de eloping AML. Al hough he associa- ion o he IL8 s2227307 SNP wi h AML isk emained only ma ginally significan a e mul iple es ing co ec ion, his finding sugges ed ha he IL8 locus migh play a ole in he pa hogenesis o AML. The IL8 gene is loca ed on ch omosome 4q12-q21 and encodes o IL8, a chemokine mainly p oduced by mac ophages and epi helial cells. P e ious s udies ha e sugges ed ha he blocking IL8- CXCR2 pa hway migh ha e a he apeu ic po en ial in a a ie y o umou s 24–27 including AML and myelodys- plas ic synd omes (MDS) 28 . Howe e , he ole o IL8 in AML is s ill sca ce. A ecen s udy has demons a ed ha IL8 and i s ecep o a e significan ly o e exp essed in Fig. 1 Func ional impac o he IL8 s2227307 SNP on immune esponses. Co ela ion be ween he IL8 s2227307 SNP and IL1βle els a e s imula ion o PBMCs (n=408) wi h Pam3Cys (10μg/ml) (a)o PF4V exp ession in pe iphe al blood (b) and co ela ion be ween he IL13 s1295686 SNP wi h IL1Ra le els a e s imula ion o PBMCs wi h LPS (100ng/ml) (c). Gene exp ession plo om he GTEx po al; h ps://g expo al.o g/home/index.h ml). Sánchez-Maldonado e al. Blood Cance Jou nal (2020) 10:75 Page 8 o 11 Blood Cance Jou nal AML and MDS pa ien s 28 and ha he exp ession o hese molecules also co ela es wi h poo ou comes. In addi ion, i has been epo ed ha he IL8-CXCR2 axis is highly exp essed in hema opoie ic s em cells and p ogeni o compa men s in compa ison wi h heal hy con ols 28 and ha his pa hway plays a key ole in he egula ion o cance s em cell unc ion 29–31 and mesenchymal s em cell-induced T cell p oli e a ion. In addi ion, Schinke e al. (2015) ha e expe imen ally demons a ed ha he inhibi ion o CXCR2 leads o dec eased iabili y and clonogenic capaci y o p i- ma y cells om AML pa ien s, which poin ed owa ds he use o IL8-CXCR2 pa hway as no el he apeu ic a ge 28 .In line wi h ou gene ic da a and he no ion o a ole o he IL8 locus in he pa hogenesis o AML, we ound ha ca ie s o he IL8 s2227307T allele had inc eased le els o IL1βa e he s imula ion o PBMCs wi h Pam3Cys (P=0.00058; Fig. 1a). These esul s sugges ed ha he p o ec i e e ec o he IL8 s2227307 SNP on AML isk migh be media ed by TLR2- induced immune esponses ha a e ini ially egula ing IL1β sec e ionand,subsequen ly,IL8p oduc ioninawide ange o pa hological condi ions 32–35 . Gi en ha he co ela ion o he IL8 s2227307 SNP wi h inc eased le els o IL1βdid no each he significance h eshold a e co ec ion o mul i- ple es ing, we need o in e p e hese esul s wi h cau ion. None heless, i wo h men ioning ha hey we e in ag ee- men wi h p e ious s udies showing ha TLRs a e exp essed in mul iple AML cell lines and p ima y AML samples 36 and ha s imula ion o TLR2 in no mal hema- opoie ic cells led o di e en ia ion and p oli e a ion o hema opoie ic s em cells and myeloid p ogeni o cells. Fu he mo e, ano he s udy p oposed a TLR2-binding cell- pene a ing pep ide as a p omising candida e o a ge ed d ug de elopmen in AML 37 .Inaddi ion o hesefindings, IL8 s2227307 hasbeenalso epo ed obeaneQTL o PF4V (Fig. 1b), a locus in ol ed in chemokine-media ed immune esponses. These esul s sugges ha he IL8 s2227307 poly- mo phism migh also influence he isk o AML h ough chemo axis s imula ion in he mic oen i onmen o he bone ma ow (BM). In line wi h his no ion, i has been demons a ed ha IL8 is a hypoxia- egula ed cy okine ha p omo es mig a ion in mesenchymal s omal cells in he BM 38 and ha bo h endogenous and hypoxia-induced p oduc ion o IL8 was highe in AML cases compa ed o con ols and was p ognos ically un a ou able 38 .Amo e ecen s udy has also sugges ed ha IL8 blockade migh be used as new he apeu ic s a egy o AML, as i p e en s ac i a ed endo helial cell media ed p oli e a ion and chemo esis ance 39 . Finally, e en hough we did no find any unc ional e ec o he VEGFA s25648 SNP o modula e immune esponses, ou gene ic findings a e in line wi h p e ious s udies epo ing an inc eased ascula i y and VEGFA le els in AML pa ien s, and a specific VEGFA-dependen ascula mo phology in he leukemic BM 40 . In addi ion, i has been epo ed ha VEGFA le els a e an independen p ognos ic ac o 41 and ha hey modula e he appea ance o g a e sus hos disease a e SCT 42 . Based on he cu en e idence, we hypo hesize ha he VEGFA s25648 SNP migh influence he isk o de eloping AML h ough changes in BM ascula i y and mo phology and mig a ion o human leukemia cells. One o he majo s eng hs o ou s udy is he inclusion o wo la ge popula ions. In he combined analysis, we had 80% powe o de ec an odds a io o 1.33 (α=0.0007) o a SNP wi h a equency o 0.25, which unde lined he easi- bili y o he s udy design. Ano he impo an s eng h o his s udy is he de elopmen o cy okine s imula ion expe imen s and he measu emen o se en se um s e oid ho mones in a la ge coho o heal hy subjec s, which allowed us o in es iga e he unc ional ole o he mos ele an ma ke s in modula ing immune esponses bu also in de e mining se ological s e oid ho mone le els. A d awback is he mul icen ic na u e o his s udy ha placed ine i able limi a ions such as he impossibili y o uni o mly collec cy ogene ic and mu a ion p ofiles o a significan se o pa ien s. Ano he limi a ion was ha age was unknown o a subse o Ge man con ols. Howe e , gi en ha selec ed SNPs ha e no been linked o su i al in AML, we hink ha age is no a modi ying ac o ha could sig- nifican ly influence he esul s. In conclusion, we iden ified o he fi s ime IL8,IL13, and VEGFA SNPs as suscep ibili y bioma ke s o AML and p o ided new insigh s abou he possible ole o hese loci in modula ing inna e and adap i e immune espon- ses, and he eby becoming po en ially clinical a ge s o enhancemen o he an ileukemic e ec s o immune cells. Func ional da a used in his p ojec ha e been me icu- lously ca alogued and a chi ed in he BBMRI-NL da a in as uc u e (h ps://h gp. bbm i.nl/) using he MOL- GENIS open sou ce pla o m o scien ific da a 43 . This allows flexible da a que ying and download, including su ficien ly ich me ada a and in e aces o machine p ocessing (R s a is ics, REST API) and using FAIR p inciples o op imise Findabili y, Accessibili y, In e - ope abili y and Reusabili y 44 . Acknowledgemen s We a e deeply g a e ul o he s udy pa icipan s. We also hank As ella Pha ma Inc. and M s. Consuelo González Mo eno (AML su i o ) o suppo ing his wo k. This s udy was suppo ed by g an s PI12/02688 and PI17/02276 om Fondo de In es igaciones Sani a ias (Ins i u o de Salud Ca los III, Mad id, Spain), he No he n Po ugal Regional Ope a ional P og amme (NORTE 2020), unde he Po ugal 2020 Pa ne ship Ag eemen , h ough he Eu opean Regional De elopmen Fund (FEDER) (NORTE-01-0145-FEDER-000013), and by Fundação pa a a Ciência e Tecnologia (FCT) (CEECIND/04601/2017 o C.C, and CEECIND/03628/2017 o A.C.). This s udy was also suppo ed by As ellas Pha ma Inc. and by a dona ion o Consuelo González Mo eno, an acu e myeloid leukemia su i o . Au ho de ails 1 Genomic Oncology A ea, GENYO, Cen e o Genomics and Oncological Resea ch: Pfize / Uni e si y o G anada / Andalusian Regional Go e nmen , PTS G anada, G anada, Spain. 2 Hema ology depa men , Vi gen de las Nie es Sánchez-Maldonado e al. Blood Cance Jou nal (2020) 10:75 Page 9 o 11 Blood Cance Jou nal