scieee Open visual document viewer

Lipidomic profiling of patient-specific iPSC-derived hepatocyte-like cells

Kiamehr, Mostafa,Viiri, Leena,Vihervaara, Terhi,Koistinen, Kaisa,Hilvo, Mika,Ekroos, Kim,Käkelä, Reijo,Aalto-Setälä, Katriina

Full text

RESOURCE ARTICLE Lipidomic p o iling o pa ien -speci ic iPSC-de i ed hepa ocy e-like cells Mos a a Kiameh 1, * ,‡ , Leena E. Vii i 1, *, Te hi Vihe aa a 2 , Kaisa M. Kois inen 2 , Mika Hil o 2 , Kim Ek oos 2 , Reijo Ka kela  3 and Ka iina Aal o-Se a la  1,4 ABSTRACT Hepa ocy e-like cells (HLCs) di e en ia ed om human induced plu ipo en s em cells (iPSCs) o e an al e na i e model o p ima y human hepa ocy es o s udy lipid abe a ions. Howe e , he de ailed lipid p o ile o HLCs is ye unknown. In he cu en s udy, unc ional HLCs we e di e en ia ed om iPSCs gene a ed om de mal ib oblas s o h ee indi iduals by a h ee-s ep p o ocol h ough he de ini i e endode m (DE) s age. In pa allel, de ailed lipidomic analyses as well as gene exp ession p o iling o a se o lipid- me abolism- ela ed genes we e pe o med du ing he en i e di e en ia ion p ocess om iPSCs o HLCs. Addi ionally, a y acid (FA) composi ion o he cell cul u e media a di e en s ages was de e mined. Ou esul s show ha majo al e a ions in he molecula species o lipids occu ing du ing DE and ea ly hepa ic di e en ia ion s ages mainly mi o he quali y and quan i y o he FAs supplied in cul u e medium a each s age. Polyunsa u a ed phospholipids and sphingolipids wi h a e y long FA we e p oduced in he cells a a la e s age o di e en ia ion. This wo k unco e s he p e iously unknown lipid composi ion o iPSC-HLCs and i s al e a ions du ing he di e en ia ion in conjunc ion wi h he exp ession o key lipid- associa ed genes. Toge he wi h biochemical, unc ional and gene exp ession measu emen s, he lipidomic analyses allowed us o imp o e ou unde s anding o he conce ed in luence o he exogenous me aboli e supply and cellula biosyn hesis essen ial o iPSC-HLC di e en ia ion and unc ion. Impo an ly, he s udy desc ibes in de ail a cell model ha can be applied in explo ing, o example, he lipid me abolism in ol ed in he de elopmen o a y li e disease o a he oscle osis. KEY WORDS: Induced plu ipo en s em cell, iPSC, Hepa ocy e-like cell, HLC, Di e en ia ion, Cell model, Lipidomics, Fa y acid INTRODUCTION The li e is he main me abolic and syn he ic o gan in he human body, ca ying ou mo e han 500 di e en unc ions. I is mainly composed o hepa ocy es, which cons i u e app oxima ely 60% o he cells in he li e and possess many impo an unc ions. Hepa ocy es p oduce he majo i y o ci cula ing plasma p o eins, including anspo e s (such as albumin and lipop o eins), p o ease inhibi o s (α1-an i ypsin, an i h ombin and α2-mac oglobulin), blood coagula ion ac o s, and modula o s o immune complexes and in lamma ion (complemen C3, C- eac i e p o ein). Hepa ocy es also con ol he homeos asis o ene gy/ uel molecules such as glucose/glycogen and a y acids (FAs) as well as o he essen ial compounds o lipid me abolism such as choles e ol and bile acids. Addi ionally, li e has a cen al ole in lipid me abolism as i is he majo si e o he gene a ion o plasma lipop o eins (Godoy e al., 2013). The use o hepa ocy es as in i o models o explo e di e en aspec s o li e unc ion and me abolism has escala ed in ecen yea s bu p ima y human hepa ocy es (PHHs), he key in i o cell ype in ol ed in e.g. choles e ol me abolism, a e sca ce because hey a e ob ained om o gan dono s. Fu he mo e, when in cul u e he PHHs quickly dedi e en ia e and lose hei li e unc ions, hus making hem imp ac ical o modelling he li e in i o (Elau e al., 2006). Human induced plu ipo en s em cell (iPSC)-de i ed hepa ocy es p o ide a good al e na i e o PHHs because iPSCs can be easily ep og ammed om de mal ib oblas s and hen di e en ia ed in o hepa ocy e-like cells (HLCs), which unc ionally esemble PHHs (Hu and Li, 2015). iPSC-HLCs can ecapi ula e me abolic a ia ions obse ed in he popula ion and ha e p o ed o be po en in bo h sho - and long- e m d ug sc eening and in in es iga ing hepa o oxici y o de eloping no el he apeu ics (Holmg en e al., 2014; Medine e al., 2013; Szkolnicka e al., 2014). In addi ion, hey ha e been u ilised o s udying e al li e exposu e o ha m ul subs ances (Lucendo-Villa in e al., 2017) and in iden i ying noncoding mic o-RNAs egula ing human li e damage (Szkolnicka e al., 2016; Yang e al., 2016). Fu he mo e, HLCs ha e been success ully used in de eloping in i o models o s udying hepa ic diseases such as sys emic amyloidosis (Leung e al., 2013), li e -s age mala ia (Ng e al., 2015) and hepa i is C i al in ec ion (Zhou e al., 2014). iPSC-HLCs could also o e a good model o in es iga ing basic mechanisms o e.g. lipid me abolism as well as i s dys egula ion ela ed o di e en diseases such as a y li e disease o a he oscle osis. Lipids a e a highly di e se class o biological molecules wi h c ucial oles in cellula ene gy s o age (mainly in he o m o iacylglyce ols), s uc u e (e.g. key componen s o plasma and nuclea memb anes, endoplasmic e iculum and Golgi appa a us, and a icking esicles like endosomes and lysosomes) and signalling (as ligands ha ac i a e signal ansduc ion pa hways as well as media o s o signalling pa hways) (Van Mee e al., 2008). Fu he mo e, de no o FA syn hesis in he cells has a signi ican impac on he acquisi ion and main enance o cellula plu ipo ency h ough inc eased mi ochond ial ission (Wang e al., 2017). Mammalian cells exp ess ens o housands o di e en lipid species and use hund eds o p o eins o syn hesise, me abolise and Recei ed 30 May 2017; Accep ed 10 July 2017 1 Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e, 33520, Finland. 2 Zo a Biosciences, Espoo, 02150, Finland. 3 Depa men o Biosciences, Uni e si y o Helsinki, Helsinki, 00014, Finland. 4 Hea Hospi al, Tampe e Uni e si y Hospi al, Tampe e, 33520, Finland. *These au ho s con ibu ed equally o his wo k ‡ Au ho o co espondence ([email p o ec ed]) M.K., 0000-0003-1894-3237 This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License (h p://c ea i ecommons.o g/licenses/by/3.0), which pe mi s un es ic ed use, dis ibu ion and ep oduc ion in any medium p o ided ha he o iginal wo k is p ope ly a ibu ed. 1141 © 2017. Published by The Company o Biologis s L d | Disease Models & Mechanisms (2017) 10, 1141-1153 doi:10.1242/dmm.030841 Disease Models & Mechanisms anspo hem (Mu o e al., 2014). Mo eo e , lipid de ec s a e cen al o he pa hogenesis o many common diseases such as non- alcoholic a y li e disease (Ruhanen e al., 2017; Younossi e al., 2016) o a he oscle osis (Meikle e al., 2011; S übige e al., 2012). Howe e , because he cellula lipidome is highly complex and e y dynamic, he s udy o lipids and iden i ying he p ecise unde lying de ec s has p e iously been hampe ed by analy ical limi a ions. This has been esol ed by he eme gence o ad anced lipidomic echnologies. Lipidomics aims o p ecisely de ine and quan i a e he molecula p o iles o lipids p esen in a cell, o ganism o issue (Wa son, 2006; Wenk, 2005), and p o ides p ecise quan i a i e snapsho s o he lipidomes comp ising hund eds o di e en molecules (Llo en e e al., 2013; Sales e al., 2016). This echnological ad ancemen in conjunc ion wi h compu a ional echnologies has made his ield a p omising a ea o biomedical esea ch (Ek oos, 2012). In his s udy, we p oduced iPSCs om pa ien -de i ed de mal ib oblas s and di e en ia ed hem in o HLCs. In o al, we quan i ied 165 molecula species o lipids, and a e he i s o epo a comp ehensi e lipidomic p o ile o cells du ing he en i e di e en ia ion p ocess om iPSCs h ough he de ini i e endode m (DE) s age o HLCs. We assess he occu ing lipidomic al e a ions in ela ion o he a iable supply o di e en FAs measu ed in he cell cul u e media a each s age o he di e en ia ion. Thus, we do no o e look he comp ehensi e e ec s o he essen ial exogenous me aboli es he cells acqui e om he cul u e medium. Addi ionally, we p esen biochemical and unc ional measu emen s and exp ession o sphingolipid (SL) me abolism- ela ed genes du ing HLC di e en ia ion. RESULTS Cha ac e isa ion o he iPSC lines The h ee iPSC lines used in his s udy –UTA.10100.EURCAs, UTA.11104.EURCAs and UTA.11304.EURCCs –we e cha ac e ised in de ail o hei plu ipo ency. All h ee iPSC lines exp essed cha ac e is ic ma ke s o plu ipo ency a he p o ein (Fig. S1A) and gene (Fig. S1B) le el, and he i ally ans e ed exogenous plu ipo ency genes we e silenced (Fig. S1C). The plu ipo ency o iPSC lines was p o en in i o by emb yoid body (EB) o ma ion and pe o ming PCR o show he p esence o all h ee ge m laye s (Fig. S1D). The ka yo ypes o all h ee iPSC lines we e no mal (Fig. S2). Di e en ia ion o iPSCs o HLCs h ough a DE s age We used a p e iously desc ibed h ee-s ep p o ocol (Hay e al., 2008) o p oduce HLCs om human iPSCs (Fig. 1A). Du ing he di e en ia ion p ocess, he cells unde wen mo phological changes: iPSCs g adually los hei ypical ound and dense mo phology and a e mig a ion o med spiky-shaped DE cells. Du ing he second week o di e en ia ion, a he hepa oblas s age, cells g adually inc eased in size and, as hey ma u ed du ing he hi d week, o med polygonal HLCs wi h dis inc canalicula ed bo de s (Fig. 1B). The immunocy ochemical s aining a day 5 showed ha cells exp essed only low le els o plu ipo ency ma ke OCT3/4 bu we e s ongly posi i e o he DE ma ke SOX17 (Fig. S3A). Fu he mo e, low cy ome y analyses showed ha 71-96% o he cells we e posi i e o he DE ma ke CXCR4 a day 5 o di e en ia ion (Fig. S3B). These esul s indica ed ha he iPSCs had e icien ly di e en ia ed in o DE du ing he endode mal induc ion s ep o he di e en ia ion p o ocol. Consis en wi h he mo phological changes obse ed du ing he di e en ia ion p ocess (Fig. 1B), changes in gene exp ession we e de ec ed (Fig. S3C). The exp ession o OCT3/4 was clea ly down egula ed by day 5, whe eas he DE ma ke SOX17 was highly exp essed a his s age. FOXA2, a ansc ip ion ac o in ol ed in li e me abolism (Wol um e al., 2004), had a con inuous exp ession pa e n s a ing a ound day 5 and con inuing un il he end o di e en ia ion. As he cells di e en ia ed u he om DE owa ds HLCs, hey s a ed exp essing AFP (a ound day 10, which peaked a ound day 15-20 depending on he cell line). Exp ession o ALB, he gene o he mos abundan li e p o ein, inc eased s ongly a ound day 15 and peaked a day 20-25 (Fig. S3C). Immunocy ochemical s aining u he con i med he exp ession o hepa ic p o eins a day 20 o di e en ia ion (Fig. 1C,D). Low-densi y lipop o ein ecep o (LDL-R), asialoglycop o ein ecep o (ASGR), alpha e op o ein (AFP) as well as albumin (ALB) we e all exp essed in HLCs (Fig. 1C,D). HLCs s ained posi i e o ALB in 13, 18 and 19% o he cells di e en ia ed om UTA.10100, UTA.11104 and UTA.11304, espec i ely. Bo h AFP and ASGR s ained posi i e in mo e han 90% o HLCs in all h ee cell lines. Func ionali y assessmen o ma u e HLCs To assess he unc ionali y o he iPSC-HLCs, we pe o med se e al expe imen s. Because li e is he mos impo an o gan o LDL ca abolism and LDL-R ac i i y, he abili y o HLCs o up ake LDL om he cul u e medium was e alua ed. As shown in Fig. 1E, iPSC- HLCs we e able o e icien ly up ake LDL. The le el o sec e ed albumin was measu ed om he cell cul u e medium a days 14, 17, 20 and 23 o di e en ia ion, and all he iPSC-HLCs syn hesised and eleased albumin in o he cul u e medium wi h maximum sec e ion a day 20 (Fig. 1F). Apolipop o ein B (APOB) and apolipop o ein A-I (APOA1) gene exp ession le els o he iPSC-HLCs du ing he di e en ia ion we e also measu ed as a su oga e o es ima ing e y low-densi y lipop o ein (VLDL) and high-densi y lipop o ein (HDL) p oduc ion le els o he cells. APOB is he main p o ein componen o VLDL, and he p oduc o he APOA1 gene is he main p o ein componen o HDLs. The exp essions o bo h APOB and APOA1 genes ose a ound day 10, eached hei peak a day 15 and hen lowe ed owa ds he le els o exp ession in PHHs (Fig. 1G,H). The inc easing exp ession o hese li e -speci ic genes u he demons a ed ha he iPSCs we e di e en ia ing and ma u ing owa ds unc ional hepa ocy es. Lipidomic p o iling o iPSC-HLCs We pe o med lipidomic p o iling o he cells du ing he en i e di e en ia ion p ocess om iPSCs o HLCs, and obse ed changes in cellula lipid con en and composi ion. O e all, mo e han 160 molecula species o lipids including choles e yl es e s (CEs), diacylglyce ols (DAGs), phospholipids (PLs) and SLs we e de ec ed du ing he cou se o di e en ia ion (Table S5). The cellula con en s o lipid classes (pe p o ein) we e cons an ly al e ed du ing he di e en ia ion. In he beginning, he p o ein- no malised concen a ions o he majo i y o lipid classes dec eased om day 0 o day 6, which mos ly mi o s he low supply o lipids o FAs om he cul u e media a his s age (Fig. 2A,B). Howe e , clea inc eases we e obse ed in h ee classes –sphingomyelin (SM), CE and phospha idylinosi ol (PI) –independen o he lowe o unde ec able le els o hose lipids in medium B. This was ollowed by a d as ic inc ease in all lipid classes om day 6 o day 12, which clea ly e lec ed he abundan lipid and FA in medium C (supplied o he cells om ∼day 7 o day 12). Thus, he amoun o FAs and lipids a ailable in he media a ec s he lipid 1142 RESOURCE ARTICLE Disease Models & Mechanisms (2017) 10, 1141-1153 doi:10.1242/dmm.030841 Disease Models & Mechanisms composi ion and con en o he cells. The measu ed concen a ions o medium lipids and FAs we e nicely in line wi h each o he (Fig. 2A,B). Du ing he hepa ic ma u a ion, om ∼day 12 onwa ds, changes in he lipid le els we e mo e sub le, in ag eemen wi h he cons an FA le el o medium D, which was gi en o he cells du ing he hepa ic ma u a ion phase and con ained on a e age 32.0 nmol FA/ml o medium (Fig. 2A). P incipal componen analysis (PCA) showed ha he cellula lipid composi ion o iPSCs and DE s age Fig. 1. See nex page o legend. 1143 RESOURCE ARTICLE Disease Models & Mechanisms (2017) 10, 1141-1153 doi:10.1242/dmm.030841 Disease Models & Mechanisms cells (days 0 and 6, espec i ely) clea ly di e s om he la e ime poin s ep esen ing hepa oblas s (day 12) and HLCs (day 16-28) (Fig. 2C). A simila PCA plo pa e n was de ec ed o he FA composi ion o cell cul u e media (Fig. 2D). To isualise he al e a ions occu ing in he cellula lipidome du ing he iPSC o HLCs di e en ia ion in mo e de ail, a hea map was compu ed using day 0 as he e e ence and including all he de ec ed molecula lipids (Fig. S4A). Because he bigges al e a ions in he cellula lipid con en and composi ion happened be o e day 12 as obse ed in he PCA (Fig. 2C) and appea ed mos ly o e lec he FA supply om he media, we compu ed ano he hea map using day 12 as he e e ence (Fig. S4B). F om his hea map, he lipid emodelling du ing he hepa ocy e ma u a ion phase became mo e e iden as he luc ua ion om days 0 o 6 was excluded. We u he es ic ed he hea map o include only molecula lipids, he concen a ion o which changed s a is ically signi ican ly be ween days 12 and 28 (P<0.05). In he esul ing hea map, 41 molecula lipids di ided in o wo dis inc clus e s isible on he hea map dend og am (Fig. 2E). The uppe clus e consis s o mainly SLs wi h e y-long-chain FAs (C24-26), le els o which inc eased du ing he ma u a ion. Lipids ha dec eased a e day 12 a e clus e ed in he lowe pa o he hea map consis ing o se e al CE species as well as lipids con aining FA 18:2 [linoleic acid (LA)]. Simila phenomena we e obse ed in ano he hea map d awn sepa a ely o he 45 FAs p esen in speci ic lipid classes in he cells (Fig. S5). We de ec ed 16 di e en CE species, he mos abundan du ing he iPSC o HLC di e en ia ion being CE 18:1, CE 16:0 and CE 18:0 (Fig. 3A, lines), concu ing wi h he highe le el o FAs 16:0, 18:0 and 18:1 a ailable in he medium C (Fig. 3A, columns). In e es ingly, CE 18:0 concen a ion was low in medium C whe eas CE 18:2 was de ec ed he highes (Fig. 3A, inse ). This CE 18:2 loading did no cause compa able accumula ion o CE 18:2 in he cells. The majo i y o CEs peaked a day 12, wi h he excep ion o CE 20:1 (peak day 16) and CE 20:3, CE 22:5 and CE22:6 (peak day 24). The dec ease o se e al CEs and he inc ease o CE 22:6 a e clea ly isible in he hea maps, illus a ing he s a is ically signi ican lipid and FA changes be ween days 12 and 28 (Fig. 2E, Fig. S5). I is o no e ha only he las medium (medium D, gi en o he cells om ∼day 13 onwa ds) p o ided he cells abundan ly wi h FA 22:6 (Fig. 3A). Du ing hepa ic di e en ia ion and ma u a ion, we de ec ed an o e all inc ease in he FA chain leng h o SLs (Fig. 3B). The C16-18 species p edomina ed (64%) in he beginning o di e en ia ion bu hei ela i e abundance g adually dec eased ( o 44%) du ing he hepa ocy e ma u a ion p ocess. A day 6, a empo a y inc ease in he p opo ion o he C16-18 species was de ec ed. In e es ingly, as he di e en ia ion p og essed, he SL chain leng h p o ile app oached ha o PHHs, in which he e y-long-chain species (C20-26) p edomina e (Fig. 3B). Close examina ion o he PL molecula species e ealed ha he la ges changes occu ed in polyunsa u a ed a y acid (PUFA)- con aining lipid species, whe eas lipids comp ised o sa u a ed o monounsa u a ed FAs (SFA o MUFA, espec i ely) we e mo e cons an h oughou he en i e di e en ia ion (Fig. 3C, Table S5). The polyunsa u a ed lipid species we e essen ially absen a he beginning o di e en ia ionand inc eased d as ically un il day 12 o 16, e lec ing he inc easing concen a ion o n-3 and n-6 PUFAs in he media du ing di e en ia ion, as shown in Fig. 3C o he phospha idylcholines (PCs). On he o he hand, he PCs con aining FA 22:5 and 22:6 appea ed la e du ing he di e en ia ion and a a lowe concen a ion, consis en wi h hei low medium concen a ion as well as a la e and (as compa ed o n-6 PUFAs) low le el supply o n-3 PUFAs om he medium (Fig. 3C). Lipids wi h di e en PUFA cons i uen s mos ly peaked a ound day 12, a e which hey s eadily dec eased ollowing he pa e n o medium FA 18:2 and 18:3 le els, which we e a he highes in medium C (du ing ∼days 7-13) and hen d opped o a lowe le el o he es o he di e en ia ion (Fig. 4). The molecula species con aining FA 20:3 and FA 20:4 peaked a ound day 12, e lec ing he media ich in hei C18 p ecu so s, and hen hei con en s emained a he ai ly cons an le el due o he la e mode a e supply o FA 20:3 and 20:4 hemsel es (Fig. 4). The n-3 PUFA p ecu so , FA 18:3, and he successo s on he me abolic pa hway, FA 20:5 and 22:5, we e no able o e icien ly aise he le els o he 22:6-con aining species, which peaked only du ing days 16-24, i.e. a e swi ching o 22:6- ich medium (Fig. 4, lowe ow). Gene–lipid in e ac ion Le els o se e al ce amides wi h e y-long-chain FAs, e.g. Ce d18:1/26:1 and Ce 18:1/26:0, inc eased s a is ically signi ican ly du ing he hepa ic ma u a ion phase (Fig. 2E). Ce amide syn hesis in ol es six di e en ce amide syn hases (Ce Ss): Ce S1-Ce S6. The CERS1 gene was he only highly exp essed ce amidase gene a he iPSC s age and i g adually educed o he exp ession le els de ec ed in PHHs (Fig. 5A). The exp ession o he CERS2 gene, by con as , slowly inc eased du ing hepa ocy e di e en ia ion, eaching he exp ession le el in PHHs by day 20 o di e en ia ion (Fig. 5A). A s a is ically signi ican co ela ion was obse ed be ween CERS2 exp ession and he e y-long-chain ce amides Ce d18:1/24:0 ( =0.83, P<0.0001), Ce d18:1/23:0 ( =0.78, P<0.001) and Ce d18:1/26:0 ( =0.73, P<0.001) (Fig. 5B). A simila gene exp ession pa e n o CERS2 was obse ed o CERS3.CERS4 exp ession a ied whe eas ha o CERS6 g adually inc eased du ing he di e en ia ion, and CERS5 was ubiqui ously exp essed h oughou he di e en ia ion (Fig. S6A). Ce amides a e deg aded o sphingosine and ee FAs by ce amidases encoded by dis inc genes such as ASAH1 and ASAH2. The exp ession o ASAH1 inc eased du ing he di e en ia ion (Fig. S6B). ASAH2 exp ession peaked s ongly a day 15 and hen g adually educed owa ds he le els de ec ed in PHHs. ASAH2b exp ession was cons an h oughou he di e en ia ion and a he same le el as in PHHs, only peaking sligh ly a day 15 (Fig. S6B). Lac osylce amide (LacCe ) and glucosyl/galac osylce amide (Glc/GalCe ) a e membe s o he glycosphingolipid (GSL) amily and he UDP-glucose:ce amide glucosyl ans e ase (UGCG) gene Fig. 1. Hepa ic di e en ia ion o iPSCs h ough he de ini i e endode m (DE) s age. (A) A schema ic ep esen a ion o he hepa ic di e en ia ion p o ocol (modi ied om Hay e al., 2008) and cell cul u e media used a each s age. See ex o de ails o he media. (B) Phase-con as images showing sequen ial mo phological changes om iPSCs (day 0) o DE (day 5) and inally hepa ocy e-like cells (HLCs) (day 20) du ing di e en ia ion. (C,D) Immunocy ochemis y o he iPSC-HLCs di e en ia ed om h ee pa ien lines (UTA.10100.EURCAs, UTA.11104.EURCAs and UTA.11304.EURCCs) a day 20 showing he exp ession o (C) LDL ecep o (LDL-R) and asialoglycop o ein ecep o (ASGR), as well as (D) α- e op o ein (AFP) and albumin (ALB). Nuclei a e s ained wi h DAPI. (E) The iPSC-HLCs we e able o up ake LDL a day 20 o di e en ia ion and (F) p oduce and sec e e albumin du ing di e en ia ion. Values a e no malised pe 1 million cells pe 24 h. Ba s ep esen means±s.d. o h ee biological eplica es. Gene exp ession le els o (G) apolipop o ein B (ApoB) and (H) apolipop o ein AI (ApoAI) a di e en ime poin s du ing he iPSC o HLC di e en ia ion no malised o he housekeeping gene GAPDH, and exp essed ela i e o day 0. Each sample was un in iplica e and ba s ep esen means±s.d. o h ee s udied cell lines. iPSC, induced plu ipo en s em cell; KO-DMEM, KnockOu Dulbecco’s modi ied Eagle medium; DMSO, dime hyl sul oxide; HGF, hepa ocy e g ow h ac o ; OSM, oncos a in M; PHH, p ima y human hepa ocy e. Scale ba s: 200 μm. 1144 RESOURCE ARTICLE Disease Models & Mechanisms (2017) 10, 1141-1153 doi:10.1242/dmm.030841 Disease Models & Mechanisms encodes he enzyme ha ca alyses he i s glycosyla ion s ep in GSL biosyn hesis. The e was a s a is ically signi ican inc ease in ou di e en GSLs, including LacCe d18:1/26:0 and Glc/GalCe d18:1/26:1, du ing he hepa ic ma u a ion phase (Fig. 2E). Compa ibly, UGCG exp ession inc eased du ing he iPSC-HLC di e en ia ion (Fig. 5A), and we obse ed a s a is ically signi ican co ela ion be ween UGCG exp ession and he o al le el o Glc/ GalCe ( =0.54, P<0.05), pa icula ly wi h Glc/GalCe d18:1/26:0 ( =0.84, P<0.0001) (Fig. S6C). An almos ou old inc ease in o al SM le el was de ec ed om day 0 o day 28 o he iPSC-HLC di e en ia ion (Fig. 2A). Speci ically, an inc ease in he longe (C23-26) SM species was obse ed (Fig. S7A). SM is me abolised om ce amides by SM syn hases (SMS1 and SMS2). The exp ession o he co esponding Fig. 2. Changes in he lipidome du ing iPSC o hepa ocy e-like cell (HLC) di e en ia ion. (A) Lines desc ibe he o al concen a ion o di e en lipid classes de ec ed in he cells a se en ime poin s du ing he iPSC o HLC di e en ia ion. Ba s ep esen he o al FA concen a ion (nmol FA/ml o medium) o di e en media used du ing hepa ic di e en ia ion. Values a he op o ba s and he op line show he o al concen a ions o FAs in he medium and PC in he cells, espec i ely. (B) To al concen a ion (µM) o di e en lipid classes in cul u e medium A, B, C and D. (C) P incipal componen analysis (PCA) plo showing he sepa a ion o samples based on lipid p o iles. PCA analysis o he 165 molecula lipids showed ha he wo i s p incipal componen s cons i u e 63.2% o he a iance. (D) PCA analysis o he FA con en o cell cul u e media used a di e en s ages o hepa ic di e en ia ion. (E) Lipidomic hea map showing old inc ease o molecula lipid species a day 16, 20, 24 and 28 o iPSC o HLC di e en ia ion as compa ed o day 12. Each ho izon al ow ep esen s a molecula lipid and each e ical column ep esen s an indi idual ime poin o he di e en ia ion. Lipid abundance a ios a e colou ed acco ding o he old changes and he colou key indica es he magni ude o log2 old change. Da a shown a e he lipids o which concen a ion di e ed s a is ically signi ican ly (P<0.05) be ween day 12 and day 28. iPSC, induced plu ipo en s em cell; CE, choles e yl es e ; Ce , ce amide; DAG, diacylglyce ol; Gb3, globo iaosylce amide; Glc/GalCe , glucosyl/galac osylce amide; LacCe , lac osylce amide; LPC, lysophospha idylcholine; LPE, lysophospha idyle hanoamine; PC, phospha idylcholine; PC O, alkyl-linked phospha idylcholine; PE, phospha idyle hanolamine; PE O, alkyl-linked phospha idyle hanolamine; PI, phospha idylinosi ol; PS, glyce ophosphose ine; SM, sphingomyelin. 1145 RESOURCE ARTICLE Disease Models & Mechanisms (2017) 10, 1141-1153 doi:10.1242/dmm.030841 Disease Models & Mechanisms genes, SGMS1 and SGMS2, inc eased du ing di e en ia ion, eaching he exp ession le els o PHHs by day 10 (Fig. 5A). A s a is ically signi ican co ela ion was de ec ed be ween SGMS2 exp ession and le els o e y-long-chain SMs (Fig. 5C) such as d18:1/24:0 ( =0.82, P<0.0001), SM d18:1/24:1 ( =0.78, P<0.001) and SM d18:1/26:1 ( =0.67, P<0.01). Le els o hese e y-long-chain SMs inc eased independen ly o he exogenous lipid supplies du ing he hepa oblas phase, whe eas he sho e -chain SMs such as SM d18:1/18:0 and SM d18:1/18:1 dec eased a he hepa ic ma u a ion s age (Fig. S7A,B). Se e al di e en sphingomyelinase (Smase) enzymes deg ade SMs. The SMPD1 gene encodes a lysosomal acid Smase, he exp ession o which inc eased du ing he hepa ic di e en ia ion (Fig. 5A) along wi h he exp ession o ano he sphingomyelinase gene, SMPD3 (Fig. S7C). The exp ession o wo o he Smase genes, SMPD2 and SMPD4, was ai ly cons an h ough he di e en ia ion and simila o ha in PHHs (Fig. S7C). DISCUSSION In his s udy, we success ully p oduced unc ional HLCs om iPSCs and desc ibe a comp ehensi e lipidomic inge p in o hese cells in conjunc ion wi h biochemical and unc ional measu emen s. We quan i a i ely moni o ed a o al o 165 molecula species o lipids du ing he cou se o hepa ic di e en ia ion, as well as measu ing he concen a ions o di e en lipids and FAs in he cell cul u e media a each s age o di e en ia ion. The analyses o lipids and FAs in he media we e complemen a y in cha ac e ising he supplies o he cells. A e commi men o he cells o he hepa ic lineage (a ound day 12), p o ein-no malised concen a ions o mos s udied lipid classes inc eased. These included unc ionally impo an speci ic PLs and SLs bu , as expec ed, he main building blocks o bulk memb ane, PC and PE, showed less a ia ion. A he same ime, he le els o s o age lipids we e ele a ed, indica ed by he inc eased CE o als. DAGs we e also analysed because hey a e an impo an in e media y lipid class be ween he s o age and s uc u al lipids, and hei ele a ed le els a day 12 indica e enhanced lipid me abolism (complex iacylglyce ol molecula species we e no add essed in his s udy). These lipidome al e a ions mi o he la ge inc ease in medium FA le els and imp o ed supply o PUFAs o he cells. Simul aneously, he cells g ew in size and app oached hepa ocy e cell mo phology. Along wi h he in luence o Fig. 3. Changes in choles e yl es e (CE), polyunsa u a ed a y acid (PUFA)-con aining phospha idylcholine (PC) species and he chain leng h o sphingolipids (SLs) du ing he iPSC o hepa ocy e- like cell (HLC) di e en ia ion. (A) Lines ep esen he concen a ion (pmol/µg o al p o ein) o all de ec ed CE species a di e en ime poin s o he iPSC o HLC di e en ia ion. Concen a ions a e p esen ed on a loga i hmic scale. Ba s ep esen concen a ion o all di e en FAs (nmol FA/ml o medium) de ec ed in di e en cell cul u e media used du ing hepa ic di e en ia ion. The inse on he lowe igh shows he concen a ion (µM) o CE species de ec ed in cul u e medium A, B, C and D. No CE species we e de ec ed in medium B. (B) The dis ibu ion o SLs wi h di e en FA chain leng hs a di e en ime poin s shows ha , du ing hepa ic di e en ia ion and ma u a ion, he ela i e abundance o C16-18 species g adually dec ease, whe eas C23-24 species inc ease. A he end o he hepa ic ma u a ion phase, he dis ibu ion closely esembles ha o p ima y human hepa ocy es (Hu8209 and Hu8210). All SLs de ec ed we e included in he calcula ions. (C) Lines ep esen he concen a ions (pmol/ µg o al p o ein) o selec ed PC species a di e en ime poin s. The le els o sa u a ed/monounsa u a ed FA (in blue)-con aining species emain a he cons an , whe eas le els o PUFA-con aining lipid species (20:4 in ed; 22:5/ 22:6 in o ange) eme ge/inc ease conside ably a day 6 and 12. Ba s ep esen he o al n-3 FA, n-6 FA and FA 22:6 concen a ion (nmol FA/ml o medium) in cell cul u e media a each s age o di e en ia ion. The inse on he lowe igh shows he concen a ion (µM) o PC species de ec ed in cul u e medium A, B, C and D. PCs 18:0/22:5 and 18:0/22:6 we e no de ec ed in he media. Values a he op line show he concen a ion o PC 16:0/18:1 in he cells du ing di e en ia ion. 1146 RESOURCE ARTICLE Disease Models & Mechanisms (2017) 10, 1141-1153 doi:10.1242/dmm.030841 Disease Models & Mechanisms exogenous supplies, we also ound endogenous cellula esponses: speci ic SL-me abolism- ela ed genes co ela ed wi h dis inc SL species le els, which showed ha , du ing he di e en ia ion, changes in gene exp ession le els we e e lec ed in he le els o speci ic lipid species a a gi en ime poin . Choles e ol biosyn hesis is one o he impo an unc ions o hepa ic cells. In e es ingly, he h ee mos abundan s o age o ms o CEs de ec ed du ing he di e en ia ion o iPSCs o HLCs, i.e. CE 18:1, CE 16:0 and CE 18:0, ha e p e iously been shown o be among he mos abundan CEs in he human li e (Nes el and Couzens, 1966). Despi e he le el o CE 18:2 being e y high in media C and D, when analysing cellula CE le els, his species was no among he highes species bu he essen ial 18:2n-6 [linoleic acid (LA)] was appa en ly inco po a ed in o s uc u al PLs, p iming hese PUFAs o signalling. This gi es con idence ha lipid me abolism o his cell model esembles ha o genuine hepa ocy es. Mos CE species we e a hei highes le els a day 12 and hen, along wi h dec easing medium CE and FA le els, CEs in he cells dec eased du ing hepa ocy e ma u a ion. Compa ed o medium A and B, medium C (p o ided o he cells om ∼day7 o ∼day 13) con ained a e y high o al concen a ion o FAs, exceeding he su icien amoun o a cell’s consump ion. This is also consis en wi h ou obse a ion o lipid d ople (LD) o ma ion du ing he second week o di e en ia ion ( esul s no shown), sugges ing ha he excess amoun o FAs p o ided in medium C a e s o ed as LDs in he cells. LDs a e he main ese oi o neu al lipids in he cells and can be used o me abolism, memb ane syn hesis (PLs and choles e ol) and s e oid syn hesis (Ma in and Pa on, 2006). SLs o m a class o lipids de ined by hei C18 amino-alcohol backbones, which a e syn hesised in he ER om non-SL p ecu so s. Modi ica ion o his basic s uc u e gi es ise o he as amily o SLs such as ce amides, SMs and GSLs, which a e s uc u al componen s o biological memb anes and bioac i e molecules pa icipa ing in di e se cellula ac i i ies such as cell di ision, di e en ia ion, gene exp ession and apop osis (Gaul e al., 2010). SLs also pa icipa e in cell signalling and modula e in lamma ion (Hannun and Obeid, 2008; Maceyka and Spiegel, 2014). Fu he mo e, inc easing e idence shows ha SLs con ibu e o he pa hogenesis o me abolic diseases including a he oscle osis (Bo odzicz e al., 2015; Haus e al., 2009; Ho nemann and Wo gall, 2013) and ha SL me abolism is a ec ed by dyslipidemia (Holland and Summe s, 2008). Because he li e is hea ily in ol ed in lipid me abolism, hepa ocy es o e a good cell model o s udying he basic mechanisms o lipid me abolism and i s dys egula ion. The p ecu so o all SLs is ce amide, which p ima ily consis s o a sphingoid long-chain base (sphingosine) and one FA chain, he leng h o which ypically anges om C14 o C26, he ce amide species wi h FAs 16:0, 24:0 and 24:1 being p edominan in mos mammalian issues. Ce amide syn hesis is a complex p ocess and o ches a ed by six mammalian Ce Ss, each o which p oduces ce amides wi h dis inc FA chain leng hs (Cingolani e al., 2016; Le y and Fu e man, 2010; Pa k e al., 2014). The exp ession pa e n o Ce S is cell speci ic, which is e lec ed in he di e en SL acyl chain composi ion in a gi en issue (Le y and Fu e man, 2010). Ce S2, he dominan Ce S iso o m ound in he li e , u ilises C20- C26 acyl-CoA species as subs a e and is one o he majo Ce Ss esponsible o he syn hesis o long-chain ce amide species Fig. 4. The le els o polyunsa u a ed a y acid (PUFA)-con aining lipids du ing iPSC o HLC di e en ia ion. Lines in he g aphs show concen a ions (pmol/µg o al p o ein; loga i hmic scale on he le ) o lipids con aining he gi en PUFA moie y a di e en ime poin s. Each line ep esen s he lipid species wi h he speci ic PUFA in a pa icula lipid class; line colou e e s o lipid class (colou key depic ed on he igh ). The ba s ep esen he concen a ion o FA 18:2, 20:3, 20:4, 18:3, 20.5, 22:5 and 22:6 (nmol FA/ml o medium, scale on he igh ) in he cell cul u e media du ing hepa ic di e en ia ion. CE, choles e yl es e ; PC, phospha idylcholine; PC O/PC P, e he -linked PC; PE, phospha idyle hanolamine; PE O/PE P, e he -linked PE; PS, phospha idylse ine; FA, a y acid. 1147 RESOURCE ARTICLE Disease Models & Mechanisms (2017) 10, 1141-1153 doi:10.1242/dmm.030841 Disease Models & Mechanisms (C20-C26) (La iad e al., 2008). In ou s udy, he concen a ion o se e al e y-long-chain (C22 o C26) SLs inc eased s a is ically signi ican ly be ween days 12 and 28, which indica es an inc ease in Ce S2 ac i i y as he cells ma u ed owa ds unc ional HLCs. This was suppo ed by a clea and s a is ically signi ican inc ease o CERS2 gene exp ession du ing iPSC-HLC di e en ia ion. Fu he mo e, a s a is ically signi ican posi i e co ela ion be ween CERS2 exp ession and e y-long-chain ce amides (e.g. Ce d18:1/ 24:0) was ound. The e y-long-chain ce amides p oduced by CERS2 a e essen ial o li e unc ion (Pa k e al., 2014); hus, he inc ease in hei le els ac s as e idence o hepa ocy e unc ionali y. Admi edly, we ha e no measu ed CERS2 enzyme ac i i ies bu only he le el o CERS2 gene exp ession. S ill, bo h he inc eased CERS2 gene exp ession as well as he inc ease in long-chain ce amides and o he long-chain SLs showed ha he s epwise hepa ic di e en ia ion p o ocol p oduced iPSC-HLCs wi h a lipid pheno ype esembling ha o PHHs. Ce amide can be deg aded o sphingosine and ee FA by ce amidases (Mao and Obeid, 2008). Acid ce amidase (encoded by he ASAH1 gene) mos e icien ly hyd olyses ce amide wi h medium-chain FA componen s (C12 o C14), whe eas he neu al ce amidase (ASAH2) p e e s long-chain o e y-long-chain componen s (C16 o >C24) (Gaul e al., 2010). In he cu en s udy, exp ession o ASAH1 and ASAH2b was a he cons an , whe eas, concu en wi h he inc ease in e y-long-chain ce amide, he exp ession o ASAH2 peaked a day 15 and hen g adually dec eased owa ds le els in PHHs. These indings sugges ha he e is a balance be ween ce amide p oduc ion by Ce S and deg ada ion by ce amidases du ing di e en ia ion, which could pa ly explain why ce amide le els s ay cons an owa ds he end o he di e en ia ion. The C26 FAs we e no de ec ed in any o he cul u e media and he C26 species o ce amides and SLs (syn hesised om hei p ecu so ) we e s ill ound in he cells, wi h inc easing amoun s owa ds he end o he di e en ia ion p ocess. Thus, he le els and molecula species composi ion o ce amide and he successo SLs seemed o be endogenously egula ed. To al SM le els inc eased du ing iPSC-HLC di e en ia ion. We showed ha he co esponding genes o SM syn hases –SGMS1 and SGMS2 –we e exp essed in he iPSC-HLCs h oughou he di e en ia ion a le els app oaching hose in PHHs. I has p e iously been shown ha bo h SGMS1 and SGMS2 posi i ely co ela e wi h le els o cellula SM (Li e al., 2007). Ou s udy suppo s his because he exp ession o SGMS1 and SGMS2 co ela ed well wi h se e al molecula SM species, especially he long-chain ones (C23 o C26). Fu he mo e, du ing he DE di e en ia ion s age, a sha p inc ease in o al le el o SM was obse ed, whe eas PC and ce amide dec eased. SM syn hases use ce amide and PC as subs a es o p oduce SM (Liu e al., 2009), o e ing an explana ion o he ansien dec eases o ce amide and PC in he beginning o he hepa ic di e en ia ion. Howe e , his s ill equi es u he s udies owing o he Fig. 5. Exp ession le els o sphingolipid (SL) me abolism- ela ed genes du ing hepa ic di e en ia ion and in p ima y human hepa ocy es (PHHs). (A) The exp ession o CERS1, CERS2,UGCG,SMPD1,SGMS1 and SGMS2 du ing hepa ic di e en ia ion om iPSCs a day 0, 5, 10, 15, 20 and 25 as well as in PHHs. The exp ession o each gene was no malised o he endogenous con ol gene, GAPDH, and p esen ed ela i e o day 0 (iPSCs). Each sample was un in iplica e and ba s ep esen means±s.d. o h ee s udied cell lines o wo PHH dono s. **P≤0.01; ***P≤0.001. (B) The exp ession o CERS2 co ela es posi i ely wi h Ce d18:1/24:0 (Spea man =0.83, P<0.0001), Ce d18:1/23:0 ( =0.78, P<0.001) and Ce d18:1/26:0 ( =0.73, P<0.001). (C) SGMS2 exp ession co ela es posi i ely wi h le els o SM d18:1/24:0 ( =0.85, P<0.0001), SM d18:1/ 24:1 ( =0.78, P<0.001) and d18:1/26:1 ( =0.67, P<0.01). Ce , ce amide; CERS, ce amide syn hase; SGMS, sphingomyelin syn hase; SMPD, sphingomyelin phosphodies e ase; UGCG, UDP-glucose ce amide glucosyl ans e ase; SM, sphingomyelin. 1148 RESOURCE ARTICLE Disease Models & Mechanisms (2017) 10, 1141-1153 doi:10.1242/dmm.030841 Disease Models & Mechanisms e y complex and dynamic na u e o he cellula lipidome, and s ong in luence o he cul u e media. Because SM is he mos abundan SL in human cells, i s coo dina ed b eakdown is an essen ial pa o memb ane homeos asis. I occu s by he Smase amily and esul s in he p oduc ion o ce amide and ee phosphocholine. The SMPD1 gene encodes a lysosomal acid Smase, whe eas SMPD2, SMPD3 and SMPD4 a e neu al Smases localised o di e en cellula compa men s (Kim e al., 2008). SMPD2 is loca ed in ER, and SMPD4inbo hERandGolgi.Theexp essiono SMPD1 inc eased du ing he hepa ic di e en ia ion as ea ly as a DE s age, sugges ing ha lysosomal deg ada ion o SM migh be c ucial o di e en ia ion, ma u a ion o unc ional s uc u e o he HLCs because ce amide, as one o he p oduc s o SM ca abolism, is conside ed c ucial o he abo e-men ioned i al cellula p ocesses, as a e i s de i a i es (Delgado e al., 2006; Gaul e al., 2010). Pha macological inhibi ion o SM de no o syn hesis dec eases no only SM le els bu also ce amide and sphingosine (Liu e al., 2009). This clea ly emphasises ha he me abolic con e sions among he SLs in he cell a e igh ly in e connec ed. In line wi h his, we obse ed a simila gene exp ession pa e n be ween SMPD1,ASAH1 and UGCG du ing he en i e hepa ic di e en ia ion. GSLs a e complex ca bohyd a e-con aining SLs and cha ac e is ic componen s o plasma memb anes, esiding speci ically in he memb ane mic odomains called lipid a s. GSL me abolism and composi ion a e al e ed du ing he p oli e a ion and di e en ia ion o a ious ypes o cells (Iwabuchi e al., 2010). LacCe is a p ecu so in he biosyn hesis o complex GSLs, and known o ac i a e a signal ansduc ion pa hway leading o cell p oli e a ion. In ou s udy, LacCe d18:1/ 26:0 as well as h ee di e en e y-long-chain Glc/GalCe inc eased du ing hepa ic ma u a ion. The e was also a end o inc eased exp ession o UGCG ( esponsible o glycosyla ion) du ing he iPSC-HLC di e en ia ion, co ela ing wi h he le els o Glc/ GalCe . This sugges s ha ac i e UGCG enzyme unc ion is i al in hepa ic ma u a ion because i s GSL p oduc s a e essen ial playe s in cell g ow h, de elopmen and di e en ia ion, and media e cell adhesion and modula e signal ansduc ion (Gaul e al., 2010). We de ec ed an o e all inc ease in he FA chain leng hs in he s udied SL classes du ing he di e en ia ion. This migh sugges induc ion o memb ane la e al he e ogenei y wi h mo e lipid a s inco po a ing SLs wi h longe chain leng hs as he cells di e en ia e owa ds HLCs. The lipid a s a e ega ded as small ( om 10 o 200 nm) specialised egions o plasma memb ane en iched in choles e ol and SLs wi h e y-long and sa u a ed acyl chains. The lipid a s a e mo e o de ed and sligh ly hicke lipid domains han he su ounding bulk memb ane, which con ains unsa u a ed PL molecules. In line wi h his, we ound signi ican inc ease in e y- long and sa u a ed SLs (24:0 and 26:0) (Fig. S5), as well as high le els o unsa u a ed PLs p o ided by medium C om day 7 (Fig. 3C). Signalling and cell-adhesion molecules a e localised in lipid a s, implica ing ha hese domains may o m pla o ms o signal ansduc ion and cell adhesion. Complex SLs a e needed in he lipid a s o example o cell–cell con ac and e icien memb ane a icking (Wassall e al., 2004). Howe e , besides choles e ol and SLs, he eme gence o he a domains om bulk memb ane also in ol es unsa u a ed PLs. En ichmen o n-3 PUFA in he plasma memb ane al e s he la e al o ganisa ion o lipids. Polyunsa u a ed PLs, especially hose wi h highly unsa u a ed n-3 PUFAs, a e s e ically incompa ible wi h choles e ol and lipids wi h long sa u a ed acyl chains, and hus he inco po a ion o hese PUFAs in o he memb ane PLs o ces he a lipids ou o he bulk memb ane, inducing a o ma ion (Wassall and S illwell, 2009). Howe e , supplemen ing cells in excess wi h highly unsa u a ed FAs, consequen ly inse ed in o he memb anes, may dis u b a in eg i y (Tu k and Chapkin, 2013). By egula ing he sizes and p ope ies o he a e sus non- a domains, he n-3 PUFAs egula e a -media ed downs eam signalling, ansc ip ional ac i a ion and cy okine sec e ion (Tu k and Chapkin, 2013). We de ec ed a s ong inc ease in PUFA- con aining PL species du ing hepa ic di e en ia ion. In addi ion o d i ing a o ma ion, he polyunsa u a ed PLs a ec bulk memb ane p ope ies, such as luidi y, lexibili y and pe meabili y (Rawicz e al., 2008). The changing o hese p ope ies a ec s memb ane esicle o ma ion and he eby lipid and p o ein a icking, which needs o be e icien as he cells g ow, di e en ia e and make mo e memb anes (Van Mee e al., 2008). Ve y ecen ly, Ghini and co-wo ke s showed ha supplemen ing cul u e medium wi h 22:6n-3 [docosahexaenoic acid (DHA)] a ec ed bo h he lipidome and me abolome o HepG2 hepa oma cells. Owing o he supplemen , he o al con en s o choles e ol, SFAs and MUFAs dec eased, whe eas he PUFA and TAG con en s o he cells inc eased (Ghini e al., 2017). In ou s udy, 22:6n-3 was p esen a a e y low le el in media A, B and C, and p esen mainly in medium D, used du ing he la e hepa ic ma u a ion. The p ecu so s o 22:6n-3 (e.g. 18:3n-3), gi en o he di e en ia ing cells a an ea lie phase, did no immedia ely gi e ise o 22:6n-3, which is likely owing o he equi ed mul is ep syn he ic pa hway. Compa ed o he syn hesis o 20:4n-6 om i s p ecu so 18:2n-6, he me abolism o 22:6n-3 om 18:3n-3 equi es wo addi ional chain elonga ions and a desa u a ion in he ER ollowed by a pe oxisomal chain-sho ening s ep, he p ocess hus equi ing mo e ime and being less e icien han he syn hesis o 20:4n-6 (Tigis u-Sahle e al., 2017). The HLCs, howe e , we e able o p oduce some 22:6n- 3 om i s p ecu so s, as can be seen om he ising le els o e.g. 22:6-con aining PC-P (plasmalogen) du ing hepa ic ma u a ion. Inco po a ing 22:6n-3 in o PS is c ucial o ac i a e p o ein kinase C pa hways (Ai es e al., 2007; Gio gione e al., 1995) and, appa en ly, his ype o signalling is ac i ely ec ui ed only a la e s ages o he di e en ia ion. PC is he majo componen o euka yo ic cell memb anes and he majo PL componen o all plasma lipop o ein classes (Cole e al., 2012), and is cu en ly he only known PL class o be equi ed o lipop o ein assembly and sec e ion (Cas o-Gómez e al., 2015). PC molecules con ain a ange o FA chains wi h a ying leng hs and double-bond posi ions (Yamashi a e al., 1997). We saw a mo e d as ic inc ease (a ∼day 12) o PC species con aining FA 20:4, he p ecu so o which is an essen ial FA (18:2n-6), han PCs con aining o he PUFAs. This can pa ly be explained by he concu en s ong ise o n-6 PUFAs in he medium. When he cells ge enough 18:2n- 6 om he medium, hey a e able o syn hesise all he o he membe s o he n-6 PUFA amily (Russo, 2009). Howe e , cellula emodelling p ocesses p e e p oducing PC species wi h 20:4n-6 in he sn-2 posi ion o he molecule: such PC species a e he p e e ed subs a es o cy osolic phospholipase A2 ype IV (PLA2IV), clea ing 20:4n-6 o syn hesis o eicosanoids, which modula e immune esponses, cell g ow h and di e en ia ion (As udillo e al., 2012; Fujishima e al., 1999). Appa en ly, he capaci y o he di e en ia ing cells o p oduce hese lipid media o s a ises along wi h he appea ance o 20:4n-6-con aining PC species. I is especially in e es ing ha , in ou cells, APOB exp ession peaked and LD we e obse ed a 2 weeks, soon a e he ma u ing HLCs eached hei highes le els o 20:4n-6-con aining PC. I was ecen ly e ealed ha de ec s in p o eins such as Lpca 3 and Tm6s 2, needed o e icien inco po a ion o 20:4n-6 in o ER 1149 RESOURCE ARTICLE Disease Models & Mechanisms (2017) 10, 1141-1153 doi:10.1242/dmm.030841 Disease Models & Mechanisms