Blood hsa-miR-122-5p and hsa-miR-885-5p levels associate with fatty liver and related lipoprotein metabolism : The Young Finns Study

1 Scien i ic RepoR s | 6:38262 | DOI: 10.1038/s ep38262 www.na u e.com/scien i ic epo s Blood hsa-miR-122-5p and hsa- miR-885-5p le els associa e wi h a y li e and ela ed lipop o ein me abolism—The Young Finns S udy Emma Rai oha ju1, Ilkka Seppälä1, Leo-Pekka Lyy ikäinen1, Jo ma Viika i2, Mika Ala-Ko pela3,4,5, Pasi Soininen3,4, An i J. Kangas3, Melanie Waldenbe ge 6, No man Klopp7,8, Thomas Illig6,7,8, Jaana Lei iskä9, B i -Ma ie Loo9, Niku Oksala1,10, Mika Kähönen11, Nina Hu i-Kähönen12, Reijo Laaksonen1, Olli Rai aka i13,14 & Te ho Leh imäki1 Mic oRNAs a e in ol ed in disease de elopmen and may be u ilized as bioma ke s. We in es iga ed he associa ion o blood miRNA le els and a) a y li e (FL), b) lipop o ein and lipid pa hways in ol ed in li e lipid accumula ion and c) le els o p edic ed mRNA a ge s in gene al popula ion based coho . Blood mic oRNA p o iling (TaqMan OpenA ay), genome-wide gene exp ession a ays and nuclea magne ic esonance me abolomics we e pe o med o Young Finns S udy pa icipan s aged 34–49 yea s (n = 871). Li e a s a us was assessed ul asonog aphically. Le els o hsa-miR-122-5p and -885-5p we e up- egula ed in indi iduals wi h FL ( old change (FC) = 1.55, p = 1.36 * 10−14 and FC = 1.25, p = 4.86 * 10−4, espec i ely). In eg ession model adjus ed wi h age, sex and BMI, hsa-miR- 122-5p and -885-5p p edic ed FL (OR = 2.07, p = 1.29 * 10−8 and OR = 1.41, p = 0.002, espec i ely). Toge he hsa-miR-122-5p and -885-5p sligh ly imp o ed he de ec ion o FL beyond es ablished isk ac o s. These miRNAs may be associa ed wi h FL o ma ion h ough he egula ion o lipop o ein me abolism as hsa-miR-122-5p le els associa ed wi h small VLDL, IDL, and la ge LDL lipop o ein subclass componen s, while hsa-miR-885-5p le els associa ed in e sely wi h XL HDL choles e ol le els. Hsa-miR-885-5p le els co ela ed in e sely wi h oxys e ol-binding p o ein 2 (OSBPL2) exp ession ( = −0.143, p = 1.00 * 10−4) and supp essing he exp ession o his lipid ecep o and s e ol anspo e could link hsa-miR-885-5p wi h HDL choles e ol le els. 1Depa men o Clinical Chemis y, Pi kanmaa Hospi al Dis ic , Fimlab Labo a o ies, and Uni e si y o Tampe e, School o Medicine, Tampe e, Finland. 2Di ision o Medicine Tu ku Uni e si y Hospi al and Depa men o Medicine, Uni e si y o Tu ku, Tu ku, Finland. 3Compu a ional Medicine, Facul y o Medicine, Uni e si y o Oulu and Biocen e Oulu, Oulu, Finland. 4NMR Me abolomics Labo a o y, School o Pha macy, Uni e si y o Eas e n Finland, Kuopio, Finland. 5Compu a ional Medicine, School o Social and Communi y Medicine and he Medical Resea ch Council In eg a i e Epidemiology Uni , Uni e si y o B is ol, B is ol, UK. 6Resea ch Uni o Molecula Epidemiology, Helmhol z Zen um, Ge man Resea ch Cen e o En i onmen al Heal h, Munich, Ge many. 7Hanno e Uni ied Biobank, Hanno e Medical School, Hanno e , Ge many. 8Ins i u e o Human Gene ics, Hanno e Medical School, Hano e , Ge many. 9Depa men o Heal h, Na ional Ins i u e o Heal h and Wel a e, Helsinki and Tu ku, Finland. 10Di ision o Vascula Su ge y, Depa men o Su ge y, Tampe e Uni e si y Hospi al, Tampe e, Finland. 11Depa men o Clinical Physiology, Tampe e Uni e si y Hospi al, and School o Medicine, Uni e si y o Tampe e, Tampe e, Finland. 12Depa men o Pedia ics, Uni e si y o Tampe e and Tampe e Uni e si y Hospi al, Tampe e, Finland. 13Resea 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. 14Depa men o Clinical Physiology and Nuclea Medicine, Uni e si y o Tu ku and Tu ku Uni e si y Hospi al, Tu ku, Finland. Co espondence and eques s o ma e ials should be add essed o E.R. (email: [email p o ec ed]) Recei ed: 08 July 2016 Accep ed: 08 No embe 2016 Published: 05 Decembe 2016 OPEN www.na u e.com/scien i ic epo s/ 2 Scien i ic RepoR s | 6:38262 | DOI: 10.1038/s ep38262 Fa y li e (FL) is cha ac e ized by he accumula ion o iacylglyce ol- ich lipid d ople s in hepa ocy es (li e a > 5–10% o li e weigh ). The p ima y causes leading o hepa ocellula lipid accumula ion a e no ye well unde s ood, bu hey a e hough o include al e a ions in he hepa ic lipid up ake, syn hesis, deg ada ion, and sec e ion1. FL is he i s s age o a con inuum om benign, simple s ea osis o s ea ohepa i is, ib osis, ci hosis, and e en hepa ocellula ca cinoma2. Non-alcoholic FL disease (NAFLD) is de ined as FL wi hou e idence o excessi e alcohol in ake o o he causes o s ea osis2 and i has his ological and pa hological simila i ies wi h alco- holic li e disease (ALD)3. NAFLD is common in Wes e n and Asian popula ions wi h a p e alence o 20–30% in he gene al popula ion and 70–90% among he obese and diabe ics4. I is also equen ly associa ed wi h obesi y, ype 2 diabe es, and hype lipidemia4. Indi iduals wi h ALD ha e been shown o be as me abolically unheal hy as subjec s wi h NAFLD, and i is highly possible ha ALD and NAFLD coexis 5. In ou popula ion based Young Finns S udy (YFS) sample, we ha e ecen ly obse ed ha alcohol consump ion was signi ican ly associa ed wi h he p e alence o FL only in no mal-weigh indi iduals, whe eas in o e weigh o obese subjec s he mos sig- ni ican co ela es o FL we e body mass index (BMI) and iglyce ides6. Majo i y o he indi iduals wi h FL in YFS (ages 34–49 yea s) a e men (74.1%). This is o be expec ed, as e en hough o e all p e alence o NAFLD has been shown o be simila be ween he sexes, men each hei peak p e alence o NAFLD in hei o ies, whils in women he p e alence o NAFLD is s ill inc easing and eaches i s peak in hei six h decade7. Al hough women a e known o de elop ALD wi h lesse alcohol consump ion, men in YFS consume conside able mo e alcohol han he women8. Mic oRNAs (miRNAs, miRs) a e small non-coding RNAs ha p ima ily egula e gene exp ession by bind- ing o a ge mRNAs and in e e ing wi h hei ansla ion9. Mic oRNAs a e ansc ibed om DNA o p ima y ansc ip s, which a e hen clea ed o sho e hai pin s uc u es, expo ed o cy oplasm, and u he p ocessed o o m ma u e one-s anded miRNAs. In mos cases, miRNAs ep ess hei a ge s by in e ac ion wi h he 3′ UTR o he a ge mRNA, inducing a de ec able change in he mRNA le els10. Mic oRNAs can also be anspo ed be ween cells and issues ia ci cula ion. Memb ane- ee miRNAs s abilized by p o eins11,12 and miRNAs associ- a ed wi h lipid esicles can be ound in blood. Ci cula ing miRNAs ha e also been ound in high-densi y lipop o- eins (HDL)13,14. Ci cula ing miRNAs ha e been shown o pa icipa e in cell- o-cell communica ion15, po en ially con ibu ing o disease p og ession. Humans a e p edic ed o ha e 2,588 ma u e mic oRNAs (mi Base, elease 21) ha can con ol se e al genes, and indi idual mRNAs can be bound by se e al miRNAs—miRNAs can hus es ablish wide egula o y ne wo ks a ec ing se e al me abolic p ocesses. The de elopmen o FL has been shown o al e he miRNA p o iles in he li e 16,17, se um18–20, and exosomes21. Mic oRNAs ha e also been shown o a ec he egula ion o choles e ol me abolism22, li e in lam- ma ion23, hepa ic ib osis24, and he de elopmen o hepa ocellula ca cinoma25, all associa ed wi h NAFLD and ALD26,27. E en a se um miRNA exp ession panel consis ing o hsa-miR-122-5p, -1290, -27b-3p, and -192-5p has been sugges ed o clinical NAFLD diagnos ics18. S udies ha e shown inc eased le els o ci cula o y miR-122 in li e diseases wi h di e en e iologies18,20,23,28–31, indica ing ha i may be a po en ial ma ke o li e inju y. In addi ion, o example he ci cula o y miRNAs ela ed o in lamma ion, such as miR-21, -34a, -45128 and -155 ha e been associa ed o NAFLD ( e iewed in e s 32 and 33). S ill, many o he s udies o miRNAs in FL disease ha e been pe o med wi h animal models. So a , miRNA p o iling in humans has been done in small selec ed g oups, and he la ge s udies ha e ocused only on ew speci ic p eselec ed miRNAs18–21,23,28,30,34. These s udies ha e also been pe o med in non-popula ion based coho s, and hus canno be di ec ly applied o gene al popula ion. The e o e, wide miRNA p o iling analyses wi h an adequa e sample size and popula ion based sampling a e needed. The YFS is well-sui ed o his p o iling, as he s udy pa icipan s ha e been ex ensi ely cha ac e ized and he popula ion has a e y low p e alence o i al hepa i is: acco ding o he Finnish Red C oss, he p e alence o hepa i is B and C was 0.03% and 0.05%, espec i ely, among Finnish blood dono s5. The aims o he cu en s udy we e o (i) compa e blood miRNA le els be ween indi iduals wi h o wi hou FL in a la ge popula ion-based s udy coho , (ii) in es iga e he associa ion o disco e ed miRNAs and li e enzyme le els and explo e he p edic i e abili y o hese miRNAs o de ec FL o e es ablished isk ac o s, (iii) es he hypo hesis ha hese miRNAs migh be in ol ed in he egula ion o lipop o ein and lipid pa hways connec ed o li e lipid accumula ion by s udying he associa ion be ween dys egula ed miRNAs and nuclea magne ic es- onance (NMR) me abolomics and o he cha ac e is ic p e iously associa ed wi h a me abolic dys unc ion, (i ) u he explo e he mode o unc ion o he miRNAs by analyzing he connec ions o hese FL associa ed miRNAs wi h hei p edic ed mRNA a ge s using ansc ip omics analyses om he same samples. The low and summa y o he s udy is desc ibed in Supplemen a yFigu e1. Ma e ials and Me hods The Young Finns S udy. YFS is a mul icen e ollow-up s udy on ca dio ascula isk om childhood o adul hood in Finland. The YFS was launched in 1980, when 3,596 child en and adolescen s (3–18 yea s old) pa - icipa ed in he baseline s udy35. The subjec s we e andomly selec ed om Finnish na ional egis y among he chosen age g oups and he i e s udy dis ic s. The ea e , he subjec s ha e been ollowed up (in 1986, 2001, 2007 and 2011) wi h se e al examina ions including comp ehensi e isk ac o assessmen s. The 30-yea ollow-up was pe o med in 2011, wi h 2,063 adul s, aged 34–49 yea s, pa icipa ing in he s udy. The examina ions included physical measu emen s, blood es s, and ques ionnai es. Pa icipan s in he ollow-up s udies ha e been ound o be mo e o en women and olde han hose who d opped ou , bu no signi ican di e ences in isk ac o s ha e been ound35. The p esen s udy has been app o ed by he 1s e hical commi ee o he Hospi al Dis ic o Sou hwes Finland on Sep embe 21s , 2010 and by local e hical commi ees. All s udy subjec s ga e an in o med consen and he s udy was conduc ed acco ding o he p inciples o he Decla a ion o Helsinki. YFS samples o miRNA analysis (n = 992) we e selec ed independen ly o m li e s a us om indi iduals ha ing he mos com- p ehensi e da a on isk ac o , me abonomic, ansc ip omics, genome wide geno yping and o he pheno yping www.na u e.com/scien i ic epo s/ 3 Scien i ic RepoR s | 6:38262 | DOI: 10.1038/s ep38262 in he ollow-up s udies pe o med in 1986, 2001, 2007 and 2011. A e quali y con ol, s udy popula ion com- p ises o 871 indi iduals wi h success ul miRNA p o iling (demog aphics in Table1). Clinical and biochemical measu emen s. Weigh and heigh we e measu ed and BMI calcula ed as weigh (kg)/(heigh (m))2. Wais ci cum e ence was measu ed o he nea es 0.1 cm. Blood p essu e was measu ed wi h a andom ze o sphygmomanome e . Venous blood samples we e d awn om he igh an ecubi al ein a e an o e nigh as . Fo blood coun analysis, whole blood was an icoagula ed wi h EDTA. Blood cell pa ame e s we e measu ed by low cy ome - ic pa icle coun ing (cells) and pho ome y (Hb) using Sysmex XE- 5000 and XT-2000i analyze s (Sysmex Co po a ion) wi h eagen s p o ided by he manu ac u e (Cellpack and Sul olyse ). Fo he biochemical measu emen s, se um was sepa a ed, aliquo ed and s o ed a − 70 °C un il analysis. Se um alanine amino ans e ase (ALT), aspa a e amino ans e ase (AST), gamma-glu amyl ans e ase (GT), glucose, choles e ol, and iglyce ide concen a ions we e measu ed wi h ALT, AST, GT, Glucose, Choles e ol, and T iglyce ides Sys em Reagen , (Beckman Coul e Biomedical). Apolipop o ein A1 (ApoA1), apolipop o ein B (ApoB), and C- eac i e p o ein (CRP) we e de e mined immuno u bidime ically (ApoA1 and B assay eagen , O ion Diagnos ica and CRP La ex eagen , Beckman Coul e Biomedical). The se um iglyce ide concen a- ion was assayed using he enzyma ic glyce ol kinase–glyce ol phospha e oxidase me hod (Beckman Coul e Biomedical). Se um o al choles e ol le els we e measu ed by he enzyma ic choles e ol es e ase–choles e ol oxi- dase me hod (Beckman Coul e Biomedical). The same eagen was used o es ima ing HDL choles e ol le els a e he p ecipi a ion o low-densi y lipop o ein (LDL) and e y low-densi y lipop o ein (VLDL) wi h dex an sul a e- Mg2+. Se um glucose concen a ions we e de e mined by he enzyma ic hexokinase me hod (Beckman Coul e Biomedical). All he abo e-men ioned assays we e pe o med on an AU400 ins umen (AU400, Olympus). Glyca ed hemoglobin (HbA1c) ac ion in whole blood was measu ed by an Abbo A chi ec ci8200 analyze (Abbo Labo a o ies). The concen a ion o o al hemoglobin was i s de e mined colo ime ically, a e which he concen a ion o HbA1c was measu ed immuno u bidime ically using he mic opa icle agglu ina ion inhi- bi ion me hod (Fishe Diagnos ics). These wo concen a ions we e used o calcula e he HbA1c pe cen age. All he abo e-men ioned me hods besides ALT, AST, and GT quan i ica ions we e acc edi ed by he Finnish Acc edi a ion Se ice (FINAS). NMR me abolomics. A high- h oughpu se um NMR me abolomics pla o m was used o absolu e quan- i ica ion o se um lipids and me aboli es, including lipop o ein subclass dis ibu ions, a y acids, and a ious small molecules such as amino acids and glycolysis p ecu so s36,37 (Supplemen a yTable1). The analyzed 14 lipo- p o ein subclasses we e de ined based on pa icle size. These de ailed lipop o ein subclass measu emen s, oge he wi h s anda d o al lipids and apolipop o eins, p o ide a good basis o s udying he lipid- and lipop o ein- ela ed me abolic pa hways. The NMR-based me abolic p o iling has p e iously been used in a ious epidemiological All subjec s NL FLAll FLMFLC Numbe o subjec s 871 724 147 119 28 Age, yea s 42.4 (4.8) 42.1 (4.9) 43.6 (4.4) 43.5 (4.5) 44.3 (4.0) Men, (%) 45.4 39.5 74.1 75.6 67.9 To al choles e ol, mmol/l 5.1 (0.9) 5.1 (0.9) 5.4 (1.0) 5.5 (1.0) 5.1 (1.1) HDL choles e ol, mmol/l 1.3 (0.3) 1.4 (0.3) 1.2 (0.3) 1.2 (0.3) 1.1 (0.3) LDL choles e ol, mmol/l 3.3 (0.8) 3.2 (0.8) 3.4 (0.9) 3.5 (1.0) 3.1 (0.9) T iglyce ides, mmol/l 1.2 (0.7) 1.1 (0.6) 1.8 (1.0) 1.8 (1.0) 1.8 (0.9) Type 2 diabe es, (%) 1.7 1.2 10.9 9.2 17.9 Blood glucose, mmol/l 5.4 (0.9) 5.3 (0.8) 5.8 (1.0) 5.7 (1.0) 6.0 (1.1) HbA1c, % 5.5 (0.4) 5.4 (0.4) 5.7 (0.6) 5.7 (0.5) 6.0 (0.7) HbA1c, mmol/mol 36.5 (4.6) 36.0 (3.9) 39.1 (6.5) 38.5 (6.0) 41.8 (7.9) Hype ension, (%) 8.2 5.4 21.8 17.6 39.3 Sys olic BP, mmHg 119.1 (13.9) 117.3 (13.3) 127.7 (13.1) 127.0 (13.2) 130.8 (12.3) Dias olic BP, mmHg 75.1 (10.5) 73.6 (9.9) 82.1 (10.4) 81.6 (10.8) 84.2 (8.2) Body mass index, kg/m226.4 (4.9) 25.6 (4.3) 30.6 (5.8) 29.2 (4.0) 36.5 (8.2) Alanine amino ans e ase, U/l 16.9 (12.2) 14.3 (8.0) 29.6 (19.5) 28.0 (17.7) 36.4 (25.2) Aspa a e amino ans e ase, U/l 22.2 (8.5) 21.0 (7.2) 28.2 (11.4) 27.4 (10.5) 31.3 (14.6) γ -glu amyl ans e ase, U/l 30.6 (28.3) 25.6 (46.0) 55.6 (46.0) 54.8 (49.4) 58.8 (27.8) Fa y li e index 3.8 (10.5) 1.9 (5.6) 13.3 (19.7) 8.5 (12.0) 33.3 (31.0) Table 1. Demog aphics o he subjec s o Young Finns S udy popula ion wi h success ul miRNA p o iling. Con inuous a iables a e p esen ed by means wi h s anda d de ia ions in pa en heses. Abb e ia ions: NL = no mal li e , FL = a y li e , FLM = mild a y li e , FLC = clea ly inden i ied a y li e , FLAll = all subjec s wi h a y li e , HDL = high-densi y lipop o ein, LDL = low-densi y lipop o ein, HbA1c = glycosyla ed hemoglobin, BP = blood p essu e. www.na u e.com/scien i ic epo s/ 4 Scien i ic RepoR s | 6:38262 | DOI: 10.1038/s ep38262 and gene ic s udies37, and de ails o he expe imen a ion ha e been desc ibed36–38. Da a was a ailable om all he indi iduals wi h success ul miRNA p o iling. Li e ul asonog aphy. Ul asound imaging and e alua ion o hepa ic a in hepa ic ul asound scan images ( aken in 2011) o he li e was pe o med using a alida ed p o ocol39 and Sequoia 512 ul asound main ames (Acuson) wi h 4.0 MHz adul abdominal ansduce s. The diagnos ic e alua ion o hepa ic s ea osis was pe - o med isually using a s anda d sys em by a highly ained sonog aphe acco ding o li e - o-kidney con as , pa enchymal b igh ness, deep beam a enua ion, and b igh essel walls40. Acco ding o hese c i e ia, he p es- ence o hepa ic s ea osis (= a y li e , FLAll) was assessed and he pa icipan s we e u he classi ied in o subjec s wi h clea ly iden i iably a y li e (FLC) o mild a y li e (FLM), and no mal li e (NL). RNA isola ion and quali y con ol. Whole blood (2.5 ml) was collec ed in o PaXgene Blood RNA Tubes (P eAnaly ix). The ubes we e in e ed 8–10 imes hen s o ed a oom empe a u e o a leas 2 hou s. The ubes we e ozen (− 80 °C) and hawed o e nigh be o e RNA isola ion (bo h miRNA and o al RNA) wi h a PAXgene Blood mic oRNA Ki (Qiagen) including he DNase Se using he QiaCube. The concen a ions and pu i y o he RNA samples we e e alua ed spec opho ome ically (BioPho ome , Eppendo ). The RNA isola ion p ocess was alida ed by analyzing he in eg i y o se e al RNAs wi h he RNA 6000 Nano Chip Ki (Agilen ). The p esence o he small RNA ac ion was con i med by he Agilen Small RNA Ki (Agilen ). Mic oRNA exp ession p o iling. Mic oRNA exp ession p o iling was pe o med wi h he TaqMan® OpenA ay® Mic oRNA Panel (Applied Biosys ems) con aining 758 mic oRNAs. B ie ly, 100 ng o RNA was used o un bo h A and B pools o Megaplex (Applied biosys ems) p eampli ica ion o cDNA syn hesis. In he OpenA ay Sample Loading Pla e, 22.5 μ l o each p eampli ied pool was mixed 1:1 wi h TaqMan OpenA ay Real-Time PCR Mas e Mix. Mic oRNA panels we e loaded using he AccuFill Sys em and un wi h he Quan S udio 12 K Flex (Applied Biosys ems). P ima y da a analysis was pe o med wi h Exp ession Sui e So wa e e sion 1.0.1. As ecommended by he manu ac u es o he miRNA panels, RNU6, RNU44, and RNU48 we e used as housekeeping small RNAs. Assays wi h Ampli ica ion sco e > 1 and Cq Con idence > 0.7 we e accep ed. Nine y- i e samples we e excluded due o a low numbe o miRNAs exp essed (≤ 200 miRNAs pe sample), and in u he analysis, 243 miRNAs ha we e exp essed in a leas 2/3 o he samples we e included (numbe o miRNAs p esen in de ec able le els in majo i y o he blood samples is well in line wi h p e ious simila analysis o blood issue41). The RNA quali y and unc- ionali y o he TaqMan OpenA ay mic oRNA exp ession panels ha e been alida ed p e iously42. A e quali y con ol and emo al o ou lie miRNAs, p o iling was success ul on 871 samples. To co ec o ba ch e ec s, he p incipal componen analysis was pe o med o he miRNA exp ession da a. The da a was adjus ed o 10 o he i s 20 p incipal componen s om he p incipal componen analysis. Genome-wide exp ession analysis ( ansc ip omics). The exp ession le els we e analyzed wi h an Illumina HumanHT-12 e sion 4 Exp ession BeadChip (Illumina Inc.). U ilizing he same RNA sam- ple o bo h mRNA and miRNA exp ession p o iling, 200 ng o RNA was e e se- ansc ibed in o cDNA and bio in-UTP-labeled using he Illumina To alP ep RNA Ampli ica ion Ki (Ambion); 1,500 ng o cDNA was hen hyb idized o he Illumina HumanHT-12 4 Exp ession BeadChip. The BeadChips we e scanned wi h he Illumina iScan sys em. Raw illumina p obe da a was expo ed om Genomes udio and analyzed in R (h p:// www. -p ojec .o g/) using he Bioconduc o (h p://www.bioconduc o .o g/) packages. The exp ession da a was p ocessed using nonpa ame ic backg ound co ec ion, ollowed by quan ile no maliza ion wi h con ol and exp ession p obes, using he neqc unc ion in he limma package and log2 ans o ma ion. Da a p ocessing desc ibed in mo e de ailed in e . 43. The exp ession analysis was success ul in 743 o he 871 samples wi h a miRNA exp ession p o ile. S a is ical analysis. Mic oRNA exp essions we e compa ed o e indi iduals wi h NL, FLM and FLC (using one-way ANOVA o no mally dis ibu ed miRNAs). In o de o ake accoun o he mul iple es ing Bon e oni co ec ed p- alues (pc- alue) we e calcula ed and pc- alue < 0.05 (= p < 0.00021) was conside ed signi ican . Fo dys egula ed miRNAs old changes (FCs) we e calcula ed o each indi idual sample in compa ison o he a e age o all indi iduals wi h NL. FL index was calcula ed as p e iously desc ibed44 and he le els o disco e ed miRNAs we e co ela ed wi h his index using Spea man’s ank-o de co ela ion and independen associa ion was e alu- a ed wi h linea eg ession model adjus ed wi h age and sex. The independen p edic o s o a y li e s a us (FLAll s. NL and FLC s. NL) we e esea ched using s epwise Akaike in o ma ion c i e ion (AIC) logis ic eg ession. Th ee di e en models we e used as ollows: Model 1: FL s a us (FLAll/FLC s. NL) p edic ed wi h disco e ed miRNAs only, (one by one o ced in o he model); Model 2: age, sex and BMI added among explana o y a iables o model 1; and ully adjus ed Model 3 including all explan- a o y a iables known o be associa ed wi h FL in YFS s udy6 i.e., age, sex, BMI, alcohol consump ion, wais ci - cum e ence, age, apoB le els, iglyce ides, insulin le els, sys olic blood p essu e, smoking, and physical ac i i y index, (excluding li e enzymes due o high co ela ion wi h s udied miRNAs). Only independen p edic o s o li e s a us emained in he inal model. Simila analysis we e done sepa a ely o no mal weigh (BMI < 25) and o e weigh o obese indi iduals (BMI > 25) as he explana o y pa ame e s associa ed o FL ha e been shown o di e in hese subpopula ions6 and also sepa a ely o men and women and indi iduals wi h o wi hou excess o alcohol consump ion (> 1.67 s anda d d inks pe day; co esponding o 20 g o pu e e hanol). In s a i ied analy- sis, only models p edic ing FLAll s NL we e possible, due o low numbe o indi iduals wi h FLC. The sensi i i y and speci ici y o he disco e ed miRNAs o de ec indi iduals wi h FLAll o FLC was ana- lyzed by he ecei e ope a o cu e (ROC) analysis and a eas unde cu e (AUC) we e compa ed be ween miR- NAs and he li e enzymes ALT, AST and GT. To e alua e he inc emen al p edic i e alue o miRNA le els www.na u e.com/scien i ic epo s/ 5 Scien i ic RepoR s | 6:38262 | DOI: 10.1038/s ep38262 in compa ison o pa ame e s known o be associa ed wi h FL in YFS pa icipan s6 (li e enzymes, sex, BMI, wais ci cum e ence, age, apoB le els, iglyce ides, insulin le els, sys olic blood p essu e, smoking, alcohol con- sump ion, and physical ac i i y index), a con inuous ne classi ica ion imp o emen (NRI) was calcula ed using eclassi ica ion unc ion in he P edic ABEL R package. Only pa ame e s wi h independen associa ion o li e s a us we e included in he model (ALT, GT, wais ci cum e ence, insulin le els and sys olic blood p essu e). All con inuous a iables we e in e se-no malized. A p- alue < 0.05 o con inuous NRI was conside ed signi ican . The co ela ion be ween miRNAs and me aboli e le els as well as physiological ea u es p e iously associa ed wi h me abolic dys unc ion (lis o me aboli es and o he cha ac e is ics shown in Supplemen a yTable1) was analyzed using Spea man ank-o de co ela ion. The independen p edic o s alue o he FL associa ed miRNAs was analyzed by applying he s epwise AIC eg ession model including FL associa ed miRNAs (one by one o ced in o model) and sex, age, BMI, and li e s a us as explana o y a iables when analyzing li e enzymes, and also wi h ALT, AST and GT when analyzing o he pheno ypes. All con inuous a iables we e in e se-no malized, and indi iduals wi h ALT, AST o GT le els o e he Finnish e e ence anges we e disca ded om he analysis when o he pheno ypes we e used as an dependen a iable in he model. The p edic ed mRNA a ge exp essions we e included in he analysis i hey we e ecognized as a a ge o he miRNA o in e es by a leas wo a ge p edic ion p og ams in miRGa o .3.045. Spea man’s ank-o de co ela- ions be ween he FCs o he miRNAs o in e es and he exp ession hei a ge s mRNA in ansc ip omics analysis wi h a p < 0.05 a e epo ed. An independen associa ion was assessed wi h a same model as wi h me aboli es. The pa hways en iched by he down- egula ed (p < 0.05) p edic ed a ge o hese mic oRNAs we e disco e ed by compu ing o e laps in a molecula signa u e da abase (h p://www.b oadins i u e.o g/gsea/msigdb/anno a e.jsp). Resul s Di e ences in blood le els o hsa-miR-122-5p and hsa-miR-885-5p be ween indi iduals wi h and wi hou FL. One-way ANOVA o e li e a s a us g oups (NL, FLM and FLC) showed a signi ican inc eas- ing end o he le els o hsa-miR-122-5p (assay numbe 002245, success ully p o iled o m 703 indi iduals, p = 7.89 * 10−18, Bon e oni co ec ed pc = 1.92 * 10−15) and hsa-miR-885-5p (assay numbe 002296, success- ully p o iled o m 868 indi iduals, p = 1.06 * 10−06, pc = 2.58 * 10−04). In compa ison o indi iduals wi h NL, he exp ession o hsa-miR-122-5p was up- egula ed in indi iduals wi h ei he FLM (FC = 1.45, p = 5.52 * 10−11) o FLC (FC = 1.97, p = 4.17 * 10−11) (Fig.1A). Simila ly, he exp ession o hsa-miR-885-5p was up- egula ed in indi id- uals wi h FLM (FC = 1.18, p = 3.80 * 10−4) and in hose wi h FLC (FC = 1.55, p = 4.30 * 10−5) when compa ed o hose wi h NL (Fig.1B). Bo h, hsa-miR-122-5p and -885-5p we e also up- egula ed when compa ing indi iduals wi h FLAll o indi iduals wi h NL (FC = 1.55, p = 1.36 * 10−14 and FC = 1.25, p = 4.86 * 10−4, espec i ely). These miRNAs we e also associa ed wi h a FL, when analyzing sepa a ely in indi iduals wi h and wi hou excess alcohol consump ion (Supplemen a yFigu e2, One-way ANOVA o e li e a s a us g oups; hsa-miR-122-5p p = 0.0001 and p = 3.10 * 10−12, espec i ely and hsa-miR-885-5p p = 0.005 and p = 0.007, espec i ely) o ep esen ALD and NAFLD. The e we e sex*miRNA in e ac ion wi h hsa-miR-122-5p and hsa-miR-885-5 in espec o p esence o FLAll (p = 2.69 * 10−4 and 0.046, espec i ely). In men he associa ion o hsa-miR-122-5p le els wi h FL s a us (NL s. FLM s. FLC) was in pa allel wi h women bu s a is ically s onge (p = 1.72 * 10−12 o end) han in women (p = 0.001). In simila analysis o hsa-miR-885-5p le els, he di e ences o e li e s a us g oups was seen only in men (p = 0.001 o men and p = 0.121 o women). Hsa-miR-122-5p and 885-5p co ela ed wi h FL index ( = 0.281, p = 3.54 * 10−14 and = 0.105, p = 0.002, espec i ely). In linea eg ession model including age, Figu e 1. Blood le els o hsa-miR-122-5p (A) and hsa-miR-885-5p (B) in indi iduals wi hou a y li e (NL), o wi h mild (FLM) o clea ly inden i ied a y li e (FLC). Abb e ia ions: ANOVA = Analysis o a iance, SE = s anda d e o . www.na u e.com/scien i ic epo s/ 6 Scien i ic RepoR s | 6:38262 | DOI: 10.1038/s ep38262 sex and miRNA, hsa-miR-122-5p emained a signi ican independen p edic o o FL index (p = 4.16 * 10−11, β = 0.227, 95% CI = 0.161–0.294 SD change in FL index pe one SD inc ease o miRNA le els). Hsa-miR-122-5p and -885-5p as p edic o s o a y li e . In s epwise logis ic eg ession model adjus ed wi h age, sex and BMI hsa-miR-122-5p signi ican ly p edic ed he exis ence o FLAll (OR = 2.07, 95% CI = 1.62–2.68, p = 1.29 * 10−8) and he e ec o hsa-miR-122-5p emained signi ican also in he ully adjus ed model (OR = 1.78 95% CI = 1.35–2.38, p = 6.84 * 10−5) (Table2). In s a i ied analysis, wi h no mal weigh indi iduals (BMI ≤ 25) o o e weigh and obese indi iduals (BMI > 25) he co esponding ORs we e 3.1 (95% CI = 1.54–7.0, p = 0.003) and 1.57 (95% CI = 1.17–2.13, p = 0.003), espec i ely (supplemen a yTable2). Simila ORs could also be seen when analyzing women and men sepa a ely (OR = 2.02, 95% CI = 1.20–3.56, p = 0.010 and OR = 1.77, 95% CI = 1.26–2.51, p = 0.001, espec i ely) and in mode a e (< 20 g alcohol pe day) and excess alcohol use s (≥ 20 g alcohol pe day) (OR = 1.75, 95% CI = 1.28–2.44, p = 0.001 and OR = 2.03, 95% CI = 1.02– 4.59, p = 0.059, espec i ely) (Supplemen a yTable2). Also hsa-miR-885-5p signi ican ly p edic ed he exis ence o FLAll in s epwise logis ic eg ession model adjus ed wi h age, sex and BMI (OR = 1.41 95%CI = 1.13–1.77, p = 0.002). In he ully adjus ed model (Model 3) hsa-miR-885-5p p edic ed FLAll when all o he isk ac o s we e added o he model (including alcohol consump- ion), bu he addi ion o insulin le els abolished he signi icance o he p edic i e alue o his miRNA (OR = 1.23 95%CI = 0.94–1.60, p = 0.131) (Table2). In he sex, weigh and alcohol usage s a i ied analysis, hsa-miR-885-5p was an independen p edic o o FLAll in he model including age, BMI and sex (no included in he sex s a i ied analysis) bu no in he ully adjus ed model, wi h he excep ion o he g oup o subjec s wi h excess alcohol usage whe e he esul s emained signi ican (OR = 2.08, 95%CI = 1.18–3.94, p = 0.016) (Supplemen a yTable2). Associa ions o hsa-miR-122-5p and -855-5p wi h se um li e enzymes. Hsa-miR-122-5p and 855-5p bo h co ela ed wi h he le els o ALT, AST and GT (pc < 3.05 * 10−6 o all). Also in s epwise linea eg es- sion models including age, sex, BMI and li e s a us, he indi idual FL associa ed miRNAs (hsa-miR-122-5p and -885-5p one by one) p edic ed signi ican ly all li e enzyme le els (Table3). Imp o emen o a y li e p edic ion by hsa-miR-122-5p and -855-5p o e adi ional a y li e isk ac o s. When compa ing he u ili y o FL associa ed miRNAs in iden i ying indi iduals FLC o wi h FLAll o commonly u ilized li e enzyme le els (ALT, AST and GT), he AUC o he hsa-miR-122-5p le els om he was compa able o ALT and GT (Fig.2A–C). Mo eo e , hsa-miR-122-5p ou pe o med AST when de ec ing indi iduals wi h FLC (Fig.2AandC). In his ROC analysis when compa ing AUCs o hsa-miR- 885-5p le els o hose o li e enzymes in de ec ing FLAll all li e enzymes pe o med be e han hsa-miR-885-5p (Fig.2BandC). Toge he hsa-miR-122-5p and -885-5p sligh ly imp o ed he de ec ion o FLC beyond es ablished isk ac o s alone (con inuous NRI = 1.36, 95% CI = 1.06–1.67, p = 2.8 * 10−18) (Table4). Also including hese miRNAs o he model p edic ing indi iduals wi h FLAll showed mino , bu signi ican NRI (NRI = 0.278, 95%CI = 0.077–0.480, p = 0.0067) (Table4). In sex s a i ied analysis, hsa-miR-122-5p and -885-5p sligh ly imp o ed he de ec ion o FLAll beyond es ablished isk ac o s alone in bo h men and women sepa a ely (con inuous NRI = 0.445, 95% CI = 0.068–0.821, p = 0.021 o women and con inuous NRI = 0.383, 95% CI = 0.142–0.624, p = 0.0019, o men) bu no in he weigh s a i ied analysis (Supplemen a yTable3). Ou come MODEL 1 MODEL 2 MODEL 3 miR-122 miR-885 miR-122 miR-885 miR-122 miR-885 NL s. FLAll n 703 871 701 868 632 781 p- alue 1.26 * 10−14 3.91 * 10−61.29 * 10−80.002 6.84 * 10−50.131 OR 2.44 1.55 2.07 1.41 1.78 1.23 95% CI 1.96–3.09 1.29–1.86 1.62–2.68 1.13–1.77 1.35–2.38 0.94–1.60 NL s. FLC n 600 752 598 749 545 681 p- alue 5.59 * 10−91.33 * 10−41.93 * 10−69.83 * 10−42.48 * 10−30.353 OR 4.61 2.14 6.67 2.47 3.64 1.37 95% CI 2.83–7.96 1.46–3.19 3.28–15.97 1.46–4.33 1.69–9.29 0.71–2.73 Table 2. Logis ic eg ession models (1–3) p edic ing a y li e (FLAll o FLC) wi h hsa-miR-122-5p o -885- 5p and known isk ac o s and bioma ke s o FL. MODEL 1: *S epwise logis ic eg ession model p edic ing FL (ei he FLAll o FLC) wi h miR-122-5p o miR-885-5p (one by one o ced in o model). MODEL 2: Model 1+ age, sex and BMI. MODEL 3: Model 2+ alcohol consump ion, wais ci cum e ence, apolipop o ein B le els, iglyce ides, insulin le els, sys olic blood p essu e, smoking, and physical ac i i y index. This model in ol es all he explana o y a iables ha ha e been p e iously associa ed wi h li e s a us in Young Finns S udy [see e . 6] excluding li e enzymes due o high co ela ion wi h s udied miRNAs. Abb e ia ions: NL = No mal li e , FL = a y li e , FLAll = all subjec s wi h a y li e , FLC = clea ly iden i ied a y li e , miR-122 = hsa-miR- 122-5p, miR-885 = hsa-miR-885-5p. www.na u e.com/scien i ic epo s/ 7 Scien i ic RepoR s | 6:38262 | DOI: 10.1038/s ep38262 Hsa-miR-122-5p and -885-5p as p edic o s o se um lipop o ein subclass componen s in ol ed in lipid and lipop o ein pa hways. In me abonomics analysis, hsa-miR-122-5p and -885-5p le els co e- la ed wi h se e al me aboli es and physiological ea u es associa ed wi h me abolic dys unc ion (da a no shown), bu independen p edic ion alue in linea eg ession model including sex, age, BMI, li e s a us, ALT, AST, GT and miRNA (miRNAs o ced in he model one by one) exis ed only when p edic ing size and componen s o lipop o ein subclasses and apolipop o ein le els (Fig.3, Supplemen a yTable4). Hsa-miR-122-5p (n = 668) p edic ed small VLDL, IDL, and la ge LDL pa icle concen a ions, hei lipid componen and choles e ol con- cen a ions and apoB le els (Fig.3). Hsa-miR-885-5p (n = 835) p edic ed signi ican ly e y la ge HDL subclass ee choles e ol, choles e ol es e s, o al choles e ol, phospholipid and o al lipid concen a ions (Fig.3. and Supplemen a yTable4). The hsa-miR-122-5p also signi ican ly p edic ed immuno u bidime ically measu ed apoB le els (β = 0.092, 95%CI = 0.021–0.162, p = 0.011). Nei he hsa-miR-122-5p no -885-5p le els co ela ed signi ican ly wi h leucocy e coun (da a no shown). Spea man co ela ion Linea eg ession model* n p pc n p ß (95% CI) Hsa-miR-122-5p ALT 703 5.01 * 10−33 1.22 * 10−30 0.43 701 1.07 * 10−25 0.319 (0.262–0.376) AST 703 4.49 * 10−26 1.09 * 10−23 0.384 701 3.43 * 10−22 0.321 (0.258–0.384) GT 703 5.43 * 10−17 1.32 * 10−14 0.309 701 2.10 * 10−70.163 (0.102–0.224) Hsa-miR-885-5p ALT 868 3.78 * 10−19 9.18 * 10−17 0.297 868 9.33 * 10−14 0.377 (0.149–0.252) AST 868 1.71 * 10−16 4.16 * 10−14 0.275 868 4.19 * 10−12 0.205 (0.148–0.262) GT 868 1.26 * 10−08 3.05 * 10−60.192 868 4.97 * 10−40.096 (0.042–0.150) Table 3. Co ela ion and adjus ed linea eg essions be ween hsa-miR-122-5p and -885-5p and se um li e enzyme le els. *S a is ical model: S epwise linea eg ession model p edic ing li e enzyme le els (one by one) wi h miR-122-5p o miR-885-5p (one by one o ced in o model), age, sex, BMI, and li e s a us. Abb e ia ions: ALT = Alanine amino ans e ase, AST = Aspa a e amino ans e ase, GT = gamma-glu amyl ans e ase, = co ela ion es ima e, pc = Bon e oni co ec ed p- alue. Be as (β ) indica es he s anda d de ia ion (SD) change o he li e enzyme le els pe inc ease o one SD o miRNA le els. Figu e 2. Recei e ope a ing cha ac e is i cu e (ROC) analysis o he clinical u ili y (sepeci ici y and sensi i i y) o blood le els o hsa-miR-122-5p and -885-5p in compa ison o he se um le els o li e enzymes when p edic ing indi iduals wi h FLC (A and C) o FLAll (B and C). Abb e ia ions: ALT = alanine amino ans e ase, AST = aspa a e amino ans e ase, GT = gamma-glu amyl ans e ase, FLAll = all subjec s wi h a y li e , FLC = clea ly iden i ied a y li e . www.na u e.com/scien i ic epo s/ 8 Scien i ic RepoR s | 6:38262 | DOI: 10.1038/s ep38262 Hsa-miR-122-5p and -855-5p as p edic o s o hei in silico p edic ed a ge mRNA le els. Co ela ion and adjus ed (age, sex, BMI, li e s a us and li e enzyme le els) linea eg essions be ween hsa-miR- 122-5p and -885-5p and in silico p edic ed mRNA a ge exp ession le els om ansc ip omics analyses a e shown in Table5. Hsa-miR-885-5p le els co ela ed signi ican ly wi h he le els o 48 p edic ed mRNA a - ge s (Supplemen a yTable5). The co ela ions wi h AHNAK nucleop o ein (AHNAK), ophinin (TRO) and oxys e ol binding p o ein like 2 (OSBPL2) we e signi ican a e mul iple es ing co ec ion (pc < 0.05 i.e. p < 0.00021). In linea eg ession adjus ed o age, sex, BMI, li e s a us, ALT, AST and GT, hsa-miR-885-5p p e- dic ed in e sely (indica ing a possible di ec miRNA-induced a ge gene exp ession educ ion) wi h GABA(A) ecep o -associa ed p o ein (GABARAP), OSBPL2, a ylsul a ase A (ARSA), and adial spoke 3 homolog (Chlamydomonas) (RSPH3) (Table5), OSBPL2 being he only mRNA wi h s a is ically signi ican nega i e Spea man co ela ion wi h hsa-miR-885-5p le els (a e mul iple es ing co ec ion), o which le els we e inde- penden ly p edic ed by hsa-miR-885-5p le els in he eg ession model. The majo i y o he 77 co ela ions (Spea man co ela ion p < 0.05) be ween hsa-miR-122-5p le els and i s p edic ed a ge mRNA exp essions le els we e di ec (Supplemen a yTable6) and none we e signi ican a e mul iple es ing co ec ion (pc < 0.05 i.e. p < 0.00021). In linea eg ession model adjus ed wi h age, sex, BMI, li e s a us, ALT, AST and GT, hsa-miR-122-5p in e sely p edic ed euka yo ic ansla ion ini ia ion ac o 1A, he X-linked (EIF1AX) and zinc inge CCCH- ype con aining 10 (ZC3H10) exp essions, indica ing hei possible di ec miRNA-induced a ge gene exp ession educ ion (Table5). In pa hway analysis, miRNA a ge s wi h neg- a i e associa ion o hsa-miR-122-5p and/o -855-5p we e no en iched in any speci ic pa hways. Discussion We ound ha ele a ed blood le els o hsa-miR-122-5p and -885-5p we e associa ed wi h ul asonog aphically de ec ed FL in indi iduals wi h and wi hou excess alcohol consump ion in a la ge, popula ion-based coho o adul s. These miRNAs highly co ela ed wi h se um li e enzyme le els and FL index. In he p edic ion o FLC, he le els o hsa-miR-122-5p we e compa able o mos commonly used li e enzyme le els and bo h hsa-miR-122-5p and -885-5p oge he imp o ed isk s a i ica ion beyond es ablished isk ac o s. In he me abolomic analysis, hsa-miR-122-5p le els we e ound o independen ly p edic apoB, small VLD, IDL, and la ge LDL componen le els, while hsa-miR-885-5p le els we e in e sely associa ed wi h e y la ge HDL o al lipid, phospholipid and choles e ol con en . Finally, he le els o hsa-miR-885-5p co ela ed in e sely wi h i s p edic ed a ge mRNA OSBPL2 exp ession, p esen ing a possible link be ween hsa-miR-885-5p and bo h HDL le els and i lipid con en . The associa ion o hsa-miR-122-5p blood le els wi h FL, FL index and li e enzyme le els independen ly o age, sex, and BMI was shown he e in a la ge clinically asymp oma ic popula ion. Le els o his miRNA emained as a signi ican p edic o o FL when adjus ing wi h known adi ional isk ac o s o FL and he p edic i e alue could be seen in no mal weigh and o e weigh indi iduals independen ly. This con i ms he p e ious esul whe e ele a ed miR-122 se um/plasma le els in indi iduals wi h NAFLD in compa ison o hose wi hou has been de ec ed small case-con ol p o iling s udies (n < 50)18,20,34 and wo la ge exp ession analyses wi h selec ed miRNAs (n~400 each; 4–5 miRNAs quan i ied)18,28. Ou s udy also ex ends he da a applicabili y o gene al pop- ula ion. Hsa-miR-122-5p is almos solely exp essed in he li e , and i is he mos exp essed miRNA in he li e , cons i u ing app oxima ely 70% o li e miRNA exp ession46. Du ing he p og ession o NAFLD, li e miR-122 le els ha e been shown o be down- egula ed in mice47,48 and humans16. The e ha e also been con lic ing esul s on he e ec s o alcohol consump ion on miR-122 li e exp ession49,50, bu i seems ce ain ha he de elopmen o ALD also a ec s miR-122 exp ession in he human li e 17. E en modes amoun o alcohol consump ion has been show o ele a e se um miR-122 le els51. Pa hological s a es in he li e [NAFDL18,28, ALD23 ci hosis52, can- ce 29, oxici y53 and in ec ion31] ha e all been shown o be associa ed wi h al e ed se um/plasma le els o miR-122. In suppo o ou esul s associa ing hsa-miR-122-5p le els wi h li e enzyme le els, he se um le els o miR-122 ha e been p e iously shown o be associa ed wi h ALT and o al choles e ol in hepa i is C pa ien s ecei ing a Model/mic oRNAs AUC NRI (95% CI) p- alue Model 1 (p edic ing FLC s. NL) 0.983 Re e ence Re e ence + hsa-miR-122-5p 0.986 0.963 (0.570–1.356) 1.6 * 10−6 + hsa-miR-885-5p 0.983 0.139 (− 0.296–0.575) 0.53 + hsa-miR-122-5p and hsa-miR-885-5p 0.987 1.363 (1.057–1.670) 2.8 * 10−18 Model 2 (p edic ing FLAll s. NL) 0.897 Re e ence Re e ence + hsa-miR-122-5p 0.898 0.125 (− 0.080–0.329) 0.23 + hsa-miR-885-5p 0.897 0.055 (− 0.1515–0.262) 1.00 + hsa-miR-122-5p and hsa-miR-885-5p 0.898 0.278 (0.077–0.480) 0.0067 Table 4. Con inuous ne eclassi ica ion imp o emen (NRI) o isk s a i ica ion o a y li e a e adding hsa-miR-122-5p and hsa-885-5p o he di e en base models* including con en ional isk ac o s. *S a is ical model: S epwise models 1 and 2 consis o all a iables associa ed wi h FL in he Young Finns S udy [see e . 6]. Va iables wi h independen and signi ican p edic ion alue we e: Model 1: Alanine amino ans e ase (ALT), gamma-glu amyl ans e ase (GT), wais ci cum e ence, insulin le els p edic ing indi iduals wi h FLC s. NL (NL n = 524, FLC n = 21). Model 2: ALT, GT, wais ci cum e ence, insulin le els, and sys olic blood p essu e p edic ing indi iduals wi h FLAll s NL (NL n = 524, FLAll n = 108). Abb e ia ions: AUC = a ea unde cu e, NL = no mal li e , FL = a y li e , FLAll = all subjec s wi h FL, FLC = clea ly iden i ied a y li e . www.na u e.com/scien i ic epo s/ 9 Scien i ic RepoR s | 6:38262 | DOI: 10.1038/s ep38262 nucleic acid-based miR-122 inhibi o 54. P iola e al. also showed signi ican co ela ion be ween miR-122 and ALT, AST and GT and hei esul s indica e ha miR-122 may e en egula e ALT ac i i y34. The inc eased le els o hsa-miR-885-5p we e associa ed wi h FL. Le els o his miRNA emained an inde- penden p edic o o FL when adjus ing wi h age, sex, BMI and mos o he isk ac o s o FL, including alcohol consump ion, bu he signi icance was abolished when insulin le els we e added o p edic ing model. P e ious s udies ha e indica ed ha hsa-miR-885-5p exp ession is ela i ely li e issue en iched55 bu also pe iphe al blood mononuclea cells ha e been hough o exp ess his miRNA56, indica ing possible unc ion in li e and in blood s eam. In line wi h ou esul s, up- egula ion o plasma miR-885-5p le els has been epo ed in pa ace a- mol induced li e oxici y53. In con as o a p e ious inding o inc eased le els o miR-885-5p in li e ci hosis and hepa ocellula ca cinoma in a small case con ol s udy55, we obse ed an independen and di ec associa ion be ween miR-885-5p and ALT, AST and GT concen a ions in ou wide popula ion-based sample. In addi ion, Figu e 3. Associa ions o hsa-miR-122-5p and -885-5p wi h lipop o ein subclass pa icle size and lipid componen s as well apolipo o eins A1 and B. S anda d de ia ion (SD) inc emen in me aboli e measu e and 95% con idence in e als (95% CI) pe one SD change o miRNA a e p esen ed. S a is ically signi ican esul s a e highligh ed wi h a s a . Associa ions ha e been analyzed wi h a s epwise linea eg ession model adjus ed wi h age, sex, BMI, li e s a us, alanine amino ans e ase (ALT), aspa a e amino ans e ase (AST) and gamma- glu amyl ans e ase (GT) and indi iduals wi h ALT, AST o GT le els o e he Finnish e e ence anges ha e been emo ed om he analysis. Abb e ia ions: XL = ex a la ge, L = la ge, M = medium, S = small, XS = ex a small, HDL = high-densi y lipop o ein, LDL = low-densi y lipop o ein, IDL = in e media e-densi y lipop o ein, VLDL = e y-low-densi y lipop o ein.