Membrane lipid modifications and therapeutic effects mediated by hydroxydocosahexaenoic acid on Alzheimer's disease

Accep ed Manusc ip Memb ane lipid modi ica ions and he apeu ic e ec s media ed by hyd oxydocosahexaenoic acid on Alzheime 's disease Manuel To es, Saman ha L. P ice, Ma ia A. Fiol-deRoque, Amaia Ma cilla-E xenike, Hasna Ahyayauch, Gwendolyn Ba celó-Coblijn, Sil ia Te és, Loukia Ka sou i, Ma ga i a O dinas, Da id J. López, Mai ane Iba gu en, Félix M. Goñi, Xa ie Busque s, Ja ie Vi o ica, Magdalena Sas e, Pablo V. Esc ibá DOI: 10.1016/j.bbamem.2013.12.016 To appea in: BBA - Biomemb anes Recei ed da e: 26 Sep embe 2013 Recei ed in e ised o m: 16 Decembe 2013 Accep ed da e: 18 Decembe 2013 Please ci e his a icle as: Manuel To es, Saman ha L. P ice, Ma ia A. Fiol-deRoque, Amaia Ma cilla- E xenike, Hasna Ahyayauch, Gwendolyn Ba celó-Coblijn, Sil ia Te és, Loukia Ka sou i, Ma ga i a O dinas, Da id J. López, Mai ane Iba gu en, Félix M. Goñi, Xa ie Busque s, Ja ie Vi o ica, Magdalena Sas e, Pablo V. Esc ibá. Memb ane lipid modi ica ions and he apeu ic e ec s media ed by hyd oxydocosahexaenoic acid on Alzheime 's disease. BBA - Biomemb anes (2014), doi: 10.1016/j.bbamem.2013.12.016 This is a PDF ile o an unedi ed manusc ip ha has been accep ed o publica ion. As a se ice o ou cus ome s we a e p o iding his ea ly e sion o he manusc ip . The manusc ip will unde go copyedi ing, ypese ing, and e iew o he esul ing p oo be o e i is published in i s inal o m. Please no e ha du ing he p oduc ion p ocess e o s may be disco e ed which could a ec he con en , and all legal disclaime s ha apply o he jou nal pe ain. This is he accep ed manusc ip o he a icle ha appea ed in inal o m in Biochimica e Biophysica Ac a - Biomemb anes 1838(6): 1680-1692 (2014), which has been published in inal o m a h ps://doi.o g/10.1016/j.bbamem.2013.12.016. © 2013 Else ie unde CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/) ACCEPTED MANUSCRIPT 1 Memb ane lipid modi ica ions and he apeu ic e ec s media ed by hyd oxydocosahexaenoic acid on Alzheime 's disease Manuel To esa, *, Saman ha L. P iceb, Ma ia A. Fiol-deRoquea, Amaia Ma cilla-E xenikea, Hasna Ahyayauchc, Gwendolyn Ba celó-Coblijna, Sil ia Te ésa, Loukia Ka sou ib, Ma ga i a O dinasa, Da id J. Lópeza, Mai ane Iba gu ena, Félix M. Goñic, Xa ie Busque sa, Ja ie Vi o icad, Magdalena Sas eb, *, Pablo V. Esc ibáa, * a Labo a o y o Molecula Cell Biomedicine, Uni e si y o he Balea ic Islands, Palma de Mallo ca, Spain b Di ision o B ain Sciences, Impe ial College London, London, Uni ed Kingdom c Biophysics Uni (CSIC, UPV/EHU) and Depa men o Biochemis y and Molecula Biology, Uni e si y o he Basque Coun y, Bilbao, Spain d IBIS Se ille Biomedical Resea ch Ins i u e, Vi gen del Rocio Uni e si y Hospi al, CSIC — Uni e si y o Se ille, and CIBERNED, Se ille, Spain *To whom co espondence should be add essed: Labo a o y o Molecula Cell Biomedicine, Depa men o Biology, Uni e si y o he Balea ic Islands, C a. Valldemossa km. 7.5, 07122 Palma de Mallo ca, Spain. Tel.: + 34 97117 3331; ax: + 34 97117 3184. Di ision o B ain Sciences, Impe ial College London, Hamme smi h Hospi al, Du Cane Road, W12 0NN London, Uni ed Kingdom. Tel.: +44 2075946673; ax: +44 2075946548. E-mail add esses: manuel. o [email protected] (M. To es), m.sas e@impe ial.ac.uk (M. Sas e), pablo.[email p o ec ed] (P.V. Esc ibá). ACCEPTED MANUSCRIPT 2 Abs ac Alzheime 's disease (AD) is a neu odegene a i e pa hology wi h ele an unme he apeu ic needs. Bo h na u al aging and AD ha e been associa ed wi h a signi ican decline in he omega-3 polyunsa u a ed a y acid docosahexaenoic acid (DHA), and acco dingly, adminis a ion o DHA has been p oposed as a possible ea men o his pa hology. Howe e , ecen clinical ials in mild- o-mode a ely a ec ed pa ien s ha e been inconclusi e ega ding he eal e icacy o DHA in hal ing his disease. He e, we show ha he no el hyd oxyl-de i a i e o DHA (2-hyd oxydocosahexaenoic acid — OHDHA) has a s ong he apeu ic po en ial o ea AD. We demons a e ha OHDHA adminis a ion inc eases DHA le els in he b ain o a ansgenic mouse model o AD (5xFAD), as well as hose o phospha idyle hanolamine (PE) species ha ca y long polyunsa u a ed a y acids (PUFAs). In 5xFAD mice, adminis a ion o OHDHA induced lipid modi ica ions ha we e pa alleled wi h a educ ion in amyloid-β (Αβ) accumula ion and ull eco e y o cogni i e sco es. OHDHA adminis a ion also educed Aβ le els in cellula models o AD, in associa ion wi h al e a ions in he subcellula dis ibu ion o sec e ases and educed Aβ-induced au p o ein phospho yla ion as well. Fu he mo e, OHDHA enhanced he su i al o neu on-like di e en ia ed cells exposed o di e en insul s, such as oligome ic Aβ and NMDA-media ed neu o oxici y. These esul s we e suppo ed by model memb ane s udies in which inco po a ion o OHDHA in o lipid- a -like esicles was shown o educe he binding a ini y o oligome ic and ib illa Aβ o memb anes. Finally, he OHDHA concen a ions used he e did no p oduce ele an oxici y in zeb a ish emb yos in i o. In conclusion, we demons a e he plei opic e ec s o OHDHA ha migh p o e bene icial o ea AD, which sugges s ha an ups eam e en , p obably he modula ion o he memb ane lipid composi ion and s uc u e, in luences cellula homeos asis e e sing he neu odegene a i e p ocess. This A icle is Pa o a Special Issue En i led: Memb ane S uc u e and Func ion: Rele ance in he Cell's Physiology, Pa hology and The apy. Keywo ds Lipid; Memb ane; Alzheime ’s disease; DHA; Lipid a s; Amyloid-; Tau phospho yla ion ACCEPTED MANUSCRIPT 3 1. In oduc ion Alzheime 's disease is a neu odegene a i e diso de ha p oduces se e e cogni i e impai men as i p og esses. This pa hology is he main neu ological cause o demen ia and i is su e ed by 36 million people wo ldwide, elde ly adul s in mos cases (Wo ld Alzheime Repo 2011). Un o una ely, he e a e s ill no e ec i e ea men s ha mi iga e he neu ological de ici s associa ed wi h AD. Cu en ly, hese pa ien s may be ea ed wi h wo classes o app o ed d ugs ha amelio a e he symp oms o AD, ace ylcholines e ase inhibi o s and NMDA ecep o an agonis s, al hough hei clinical e icacy is conside ed o be e y limi ed [1]. O he p omising he apeu ic app oaches ha e been p oposed o AD, such as s a ins and non-s e oidal an i-in lamma o y d ugs, al hough hey ha e ye o o e conclusi e esul s in clinical ials [2], [3], [4]. DHA (22:6 n-3) is he mos abundan omega-3 PUFA in he b ain and i is igh ly in ol ed in he unc ioning o he cen al ne ous sys em (CNS) [5], pa icula ly in neu ogenesis, synap ogenesis and synap ic ansmission [6], [7]. This a y acid is ob ained h ough he die and i s de iciency is associa ed wi h age- ela ed cogni i e decline and wi h neu odegene a i e diseases, such as AD [8], [9]. In ecen yea s, PUFAs like DHA ha e gained much a en ion due o p omising esul s ha sugges hey may be use ul o ea AD. In his sense, se e al s udies ha e demons a ed ha o al in ake o DHA o ish oil educes AD-associa ed b ain pa hology, o ins ance, imp o ing cogni i e de ici s, p o ec ing agains synap ic degene a ion and lowe ing Aβ le els in ansgenic AD mouse models [10], [11], [12], [13]. Mo eo e , hese esul s a e suppo ed by epidemiological s udies indica ing an in e se ela ionship be ween DHA in ake and AD incidence, which co ela e high DHA le els wi h educed isk o cogni i e dys unc ion [14], [15]. Howe e , di ec adminis a ion o DHA in clinical ials only showed imp o ed cogni ion o a small subg oup o pa ien s wi h e y mild cogni i e dys unc ion and he e was no clea e ec in mos pa ien s [16], [17], e en hough DHA adminis a ion imp o es he physiological, bu no pa hological, age- ela ed cogni i e decline [18]. In his con ex , he e would appea o be a link be ween AD and lipid al e a ions in neu onal memb anes, especially diminished DHA le els. The e o e, molecules ha a e e ec i e in es o ing DHA and no malizing he memb ane lipid composi ion could cons i u e he apeu ic ools o ea AD. In he p esen wo k, we show ha OHDHA egula es memb ane lipid composi ion and s uc u e, cell signaling and, addi ionally, i imp o es cogni i e sco es in animal models o AD [19], he eby ep esen ing a no el he apeu ic candida e o he ea men o AD. This DHA de i a i e bea s a hyd oxyl g oup on he α- ca bon ha impedes i s β-oxida ion and inc eases i s hal -li e in lipid memb anes. In e es ingly, na u al DHA hyd oxyl de i a es a e also p oduced in he b ain, such as neu op o ec in D1 (NPD1), and like DHA, NPD1 is also s ongly diminished in he b ain o AD pa ien s [20]. The biological unc ion o his molecule has been ela ed o mul iple neu op o ec i e e ec s, such as an ioxidan , an i-in lamma o y and an i-apop o ic oles. NPD1 also down egula es Aβ pep ide p oduc ion by modula ing β- and α-sec e ase ac i i ies, and i a o s neu onal su i al agains Aβ oxici y [20], [21]. In he p esen wo k we ound ha OHDHA adminis a ion leads o en ichmen o memb anes in long PUFAs, which is associa ed wi h neu onal su i al and neu o egene a ion, so ha OHDHA mimics he e ec s o NPD1. These changes in memb ane lipid composi ion could acili a e he main enance o a unc ional cell memb ane s uc u e ha , in he case o AD, may e e se neu ons om a pa hological o a heal hy condi ion [22]. Speci ically, we show ha by down egula ing Aβ gene a ion and Aβ-induced au p o ein hype phospho yla ion, OHDHA p omo es neu op o ec ion, and cell su i al agains di e en known AD-associa ed insul s. In addi ion, his compound can also induce neu on s em cell p oli e a ion ia molecula and cellula mechanisms ha emain la gely unknown [19] and pa ien s. Thus, OHDHA-induced neu on su i al and p oli e a ion should lead o imp o ed cogni ion in AD models [19] and pa ien s. In conclusion, OHDHA is p esen ed he e as a no el he apeu ic candida e o he ea men o he AD- ela ed neu odegene a ion. 2. Ma e ial and me hods 2.1. T ansgenic mice and ea men s A double ansgenic PS1/APP mouse model was used in his wo k (5xFAD; line Tg6799) ha ha bo s i e human mu a ions associa ed o amilial AD: he Swedish (K670N/M671L), Flo ida (I716V) and London (V717I) mu a ions in APP (amyloid p ecu so p o ein); and a human mu a ed PS1 (p esenilin 1) ha bo ing he M146L and L286V clinical mu a ions. Bo h hese ansgenes a e exp essed unde con ol o he Thy-1 p omo e . These mice display cogni i e decline om 4 mon hs o age [23]. These ansgenic 5xFAD and wild ype (WT) mice we e pu chased om Jackson Labo a o ies (USA), and hey we e main ained on a B6/SJL hyb id gene ic backg ound (C57BL/6 x SJL) by c ossing he e ozygous ansgenic mice wi h B6/SJL WT (F1) b eede s. All animals we e housed a a con olled empe a u e and ACCEPTED MANUSCRIPT 4 humidi y (22 ± 2 °C; 70% humidi y) on a 12 h–12 h ligh –da k cycle, and hey we e p o ided a s anda d labo a o y die ad libi um (Panlab A03; Ba celona, Spain). WT and ansgenic 5xFAD male mice we e o ally adminis e ed OHDHA (Lipopha ma The apeu ics; Palma de Mallo ca, Spain) dissol ed in 5% e hanol a a dose o 15 mg/kg·day o he ehicle solu ion alone (5% e hanol; 15 ml/kg·day). These ea men s s a ed when he mice eached 3 mon hs o age (dosed 5 days/week) and hey we e con inued un il he mice eached 7 mon hs o age. Du ing he las mon h o ea men , all he animals we e submi ed o a hypocalo ic die necessa y o pe o m he selec ed beha io al spa ial lea ning and memo y es ( ood c a ing es in a adial a m maze) [24]. The esul s conce ning he adial a m maze es ha e been epo ed p e iously [19], and a summa y able (con aining mo e ele an indings and o al numbe o animals ha we e used o he es ) has been also included in he discussion sec ion o he p esen wo k (Table 4). Following he beha io al es , he mice we e kep on no mal die (and ea men ) o an addi ional week, a e which hey we e eu hanized, and hei b ain was emo ed immedia ely and dissec ed down he midline on a cold su ace. Ha ing emo ed he ce ebellum, each ce ebellum- ee hemib ain was ozen in liquid ni ogen and s o ed a − 80 °C. A o al numbe o 9 animals we e used in his wo k: 3 ehicle- ea ed WT, 3 ehicle- ea ed 5xFAD and 3 OHDHA- ea ed 5xFAD mice. All he p o ocols employed we e app o ed by he Bioe hical Commi ee o he Uni e si y o he Balea ic Islands, and hey a e in ag eemen wi h na ional and in e na ional egula ions on animal wel a e. 2.2. Lipid ex ac ion and de e mina ion o choles e ol con en One hemi-b ain om each animal was homogenized in a guanidine-sal bu e (RLT Bu e ; Qiagen) a 1:20 (w: ) using a blade homogenize (Poly on PT3100). The samples we e hen incuba ed a oom empe a u e o 10 min and cen i uged o 5 min (10,000 g, 4 °C). The esul ing supe na an was eco e ed, aliquo ed and s o ed a − 80 °C. Lipids we e ex ac ed using Bligh and Dye 's me hod [25]. The eco e ed o ganic phase was s o ed unde a N2 a mosphe e a − 80 °C. The choles e ol (Cho) con en was de e mined as desc ibed p e iously [26]. B ie ly, lipid ex ac s we e e apo a ed unde a gon low o a leas 2 h and hen esuspended in isop opanol. To al cho was measu ed in an aliquo (10 μl) using an enzyma ic colo ime ic ki (Biosys ems; Ba celona, Spain) and de e mined h ough he appea ance o a p oduc abso bing a 500 nm. 2.3. Sphingolipid and phospholipid lipidomic analysis Lipidomic s udies we e pe o med as desc ibed p e iously [27], [28]. Liquid c oma og aphy/mass spec ome y (LC/MS) analysis was pe o med on a Wa e s Aqui y UPLC sys em connec ed o a Wa e s LCT P emie o hogonal accele a ed ime o ligh mass spec ome e (Wa e s), ope a ed in posi i e elec osp ay ioniza ion mode. Full scan spec a om 50 o 1500 Da we e acqui ed and indi idual spec a we e summed o p o ide da a poin s each 0.2 s. Mass accu acy and ep oducibili y we e main ained by using an independen e e ence sp ay ia LockSp ay in e e ence. The analy ical column was a 100 × 2.1 mm inne diame e , 1.7 μm C8 Acqui y UPLC b idged e hylene hyb id (Wa e s), he mos a ed a 30 °C. The wo mobile phases bo h con ained 5 mM ammonium o ma e: phase A, MeOH/H2O/HCOOH (74:25:1, / ); and phase B: MeOH/HCOOH (99:1, / ). Quan i ica ion was ca ied ou on 50 mDa windows using he ex ac ed ion ch oma og am o each compound, and he linea dynamic ange was de e mined by injec ing s anda d mix u es. Posi i e iden i ica ion o compounds was based on accu a e mass measu emen , wi h an e o o < 5 ppm, and LC e en ion imes compa ed o ha o s anda ds (± 2%). Fo lysosphingolipid quan i ica ion, ex ac s we e analyzed by LC/MS/MS wi h a sys em consis ing o a Wa e s Alliance 2690 LC pump equipped wi h an au osample and connec ed o a Qua o LC iple- quad upole mass spec ome e (Mic omass, Manches e , UK). Sepa a ion was achie ed on a Pu osphe STAR-RP-18 column (125 × 2 mm, 5 μm; Me ck) using he same mobile phases as desc ibed abo e wi h a low a e o 0.3 ml min− 1. The g adien used was: 0.0 min, 50% B; 2 min, 50% B; 7 min, 100% B; 17 min, 100% B; 19 min, 50% B; and 26 min, 50% B. MS/MS de ec ion was pe o med wi h an elec osp ay in e ace ope a ing in he posi i e ion mode acqui ing he ollowing selec ed eac ion moni o ing ansi ions: C17 d-e y h o-dihyd osphingosine-1-phospha e, 368–252 Da, collision ene gy 18 eV; and S1P, 380–264 Da, collision ene gy 16 eV. 2.4. Cell cul u e and ea men s 2.4.1. Cell lines Mouse neu oblas oma N2a cells s ably ans ec ed wi h human Swedish-mu a ed APP, also called N2aSw (kindly p o ided by Gopal Thinaka an; Uni e si y o Chicago, USA), we e main ained in a 1:1 ( : ) ACCEPTED MANUSCRIPT 5 mix u e o Dulbecco's Modi ied Eagle Medium (DMEM) and OPTI-MEM (In i ogen) supplemen ed wi h 5% e al bo ine se um (FBS; Sigma), penicillin/s ep omycin (PAA) and G-418 ( inal concen a ion, 0.2 mg/ml). N2aSw cells we e ans ec ed wi h PS1-GFP (kindly p o ided by D Ch is oph Kae he ; Leibniz Ins i u e o Age Resea ch; Ge many) o BACE1 (kindly p o ided by Jochen Wal e ; Uni e si y o Bonn; Ge many) cDNAs using he calcium phospha e me hod [29]. Human neu oblas oma SH-SY5Y cells we e main ained in DMEM Hams F12 (In i ogen) supplemen ed wi h 10% FBS (Sigma), penicillin/s ep omycin (PAA), non-essen ial aminoacids (Sigma) and 2 mM L- glu amine (Sigma). Di e en ia ion o hese cells o a neu on-like pheno ype was ca ied ou as desc ibed p e iously [30]. B ie ly, cells we e pla ed in poly-L-Lysine p e-coa ed dishes and 24 h la e , he medium was eplaced wi h esh medium supplemen ed wi h 10 μM e inoic acid (Sigma). The cells we e hen incuba ed in he da k o 5 days and he medium was eplaced wi h medium wi hou se um and supplemen ed wi h 50 ng/ml o human b ain-de i ed neu o ophic ac o (hBDNF; Alomone Labs; Tel A i , Is ael). Finally, he cells we e incuba ed o 6 days o comple e di e en ia ion. Emb yonal mul ipo en s em P19 cells we e cul u ed in α-Minimum Essen ial Medium (αMEM; Sigma) con aining 10% FBS (Sigma) and penicillin/s ep omycin (PAA). Di e en ia ion o P19 cells o a neu on- like pheno ype was ca ied ou as desc ibed p e iously [31]. B ie ly, di e en ia ion was induced by 300 nM e inoic acid o 48 h, ollowed by sub-cul u ing 1:4 in he p esence o 300 nM e inoic acid o ano he 48 h. Incuba ion wi h e inoic acid was pe o med in he da k. Cell clus e s we e hen seeded 1:4 in 6-well pla es o an addi ional 24 h o comple e he di e en ia ion p ocess. MDCK cells, s ably exp essing he human APP-GFP usion p o ein we e cul u ed in DMEM supplemen ed wi h 5% FBS (Sigma) and penicillin/s ep omycin (PAA). All cell lines we e incuba ed in a 5% CO2 a mosphe e a 37 °C. 2.4.2. Hyd oxyla ed lipids, Aβ and NMDA ea men s. S ock solu ions con aining 100 mM OHDHA in DMSO we e used o cell ea men s. DMSO was dilu ed o be below 0.1% in he medium. N2aSw cells we e incuba ed in he p esence o absence o 5, 10 o 50 μM OHDHA o 24 h. Neu on-like di e en ia ed SH-SY5Y cells we e exposed o oligome ic Aβ (5 μM) o 24 h in he p esence o absence o OHDHA (5 and 10 μM). Neu on-like di e en ia ed P19 cells we e ea ed o 24 h wi h OHDHA (10 μM), and hen o 30 min wi h NMDA (10 mM, Sigma) in esh media con aining glycine (530 μM, Sigma) and calcium (10 mM, Sigma). Finally, APP-GFP MDCK cells we e ea ed o 72 h wi h OHDHA a 1, 3, 10, 15, 20, 25, 30, 40, 50, 60 and 70 μM. 2.5. P o ein isola ion and Wes e n blo ing Hemib ains om WT and 5xFAD mice we e homogenized in a guanidine- hiocyana e-based bu e (RLT bu e , Qiagen) as desc ibed abo e (Sec ion 2.2) and he p o ein was isola ed using he All P ep DNA/RNA/p o ein isola ion ki (Qiagen) ollowing he manu ac u e 's ins uc ions. The p o ein pelle was esuspended in Laemmli's SDS-PAGE loading bu e and incuba ed o e nigh a oom empe a u e p io o use. Al e na i ely, he p o ein was ex ac ed om SH-SY5Y cells as desc ibed p e iously [32]. In he case o N2aSw cells, soluble αAPP and Aβ pep ide le els we e de e mined in ha es ed cell media [33]. The Aβ pep ide was immunop ecipi a ed om he medium o e nigh a 4 °C using he 2964 an ibody [34] and P o ein-A sepha ose beads (In i ogen). The beads we e hen eco e ed by cen i uga ion (500 g, 5 min, 4 °C), washed h ee imes and dilu ed in Laemmli's bu e . P o ein samples (5–20 μg) we e esol ed on 16% ( o b ain samples) o 4–12% ( o cell cul u e samples) polyac ylamide gels, using T is- icine o T is-glycine elec opho esis bu e . The p o eins we e hen ans e ed o ni ocellulose memb anes (GE, Ame sha m) ha we e subsequen ly blocked wi h 5% (w: ) non- a d y milk in 0.1% ( : ) Tween-20 PBS. These memb anes we e hen p obed o e nigh a 4 °C wi h he co esponding p ima y an ibody: an i-Aβ1-16 (clone 6E10, 1:4000; Signe Labs.), an i- Se 202/Th 205-PHF- au (clone AT8, 1:1000; Pie ce), an i-phospho-Se 202- au (clone CP13; 1:1000; kindly p o ided by D . Pe e Da ies, Albe Eins ein College o Medicine, USA), an i- o al- au (Tau46; 1:2000; Pie ce), an i-α- ubulin (1:5000; Sigma). An ibody binding was de ec ed wi h he IRDye-800CW- labeled an i-mouse IgG o b ain samples (1:5000; LI-COR Inc.) o ho se adish pe oxidase-conjuga ed an i-mouse IgG o cell samples (1:2000; GE Heal hca e). Memb ane luo escence was de ec ed using an Odyssey In a ed Imaging Sys em (LI-COR Inc.) and memb ane chemiluminescence was de ec ed by ECL (GE, Ame sham) in an au oma ed SRX 101A de elope (Konica). The in ensi y o he bands was quan i ied by densi ome y using he Image J so wa e (Na ional Ins i u es o Heal h) and no malized o α- ubulin. ACCEPTED MANUSCRIPT 6 2.6. RNA isola ion, e e se ansc ip ion and eal ime RT-PCR RNA isola ion, e e se ansc ip ion (RT) and quan i a i e eal- ime PCR we e pe o med acco ding o he Minimum In o ma ion o publica ion o Quan i a i e eal- ime PCR Expe imen s (MIQE) guidelines [35]. A e homogeniza ion in RLT-bu e (see Sec ion 2.2), o al RNA was isola ed om he b ain using he All P ep DNA/RNA/p o ein isola ion ki (Qiagen) acco ding o he manu ac u e 's ins uc ions. Residual DNA was emo ed by incuba ion wi h DNase (Qiagen) p io o pe o ming e e se ansc ip ion. RT was pe o med wi h he High Capaci y cDNA A chi e ki (Applied Biosys ems) on 4 μg o he o al RNA empla e, ollowing he manu ac u e 's ecommenda ions [36]. Human and mouse APP gene p oduc s we e ampli ied om 5xFAD b ain cDNA samples using speci ic p ime s: human/mouse APP695 o wa d (5′-CATCATGGTGTGGTGGAG-3′), human APP695 e e se (5′-GCGATAATGAGTAAATCATAAAAC-3′) and mouse APP695 e e se (5′- GGGTGAGTAAATAAACGGAA-3′). Fo each assay, a s anda d cu e was i s de ined using inc easing amoun s o cDNA and in all cases, he slope o hese cu es indica ed he op imal ampli ica ion e iciency (slope 3.2-3.4). GAPDH was used as a con ol housekeeping gene and his ampli ica ion was de ec ed by using comme cial Taqman p obes (Mm99999915_g1; Applied Biosys ems) acco ding o he manu ac u e 's ins uc ions. GAPDH was ampli ied in pa allel wi h he a ge gene. The esul s we e exp essed using he compa a i e C me hod, as desc ibed elsewhe e [36], [37]. As a con ol condi ion, we selec ed mouse APP exp ession in WT mice. In consequence, he exp ession o bo h mouse and human APP genes, o all mouse s ains and ea men s, was e e enced o mouse APP exp ession le els obse ed in WT mice. 2.7. Immunocy ochemis y Mouse neu oblas oma N2aSw cells we e ixed and pe meabilized in 100% me hanol a − 20 °C o 20 min, he cells we e hen ehyd a ed in PBS o 10 min and blocked in 1% BSA in PBS o 20 min a oom empe a u e. Subsequen ly, he cells we e incuba ed wi h he p ima y an ibody dilu ed in 1% BSA in PBS o e nigh : mouse an i-APP (clone 6E10, 1:200; Signe Labs.), abbi an i-BACE-1 (1:50; Cell Signaling), abbi an i-APP-C- e minal (R1(57); kindly p o ided by D P. Meh a, New Yo k S a e Ins i u e o Basic Resea ch in De elopmen al Disabili ies, USA); mouse an i-Lamp-1 (1:200; Abcam) and/o abbi an i-LC3 (1:200; Abcam). The cells we e hen washed 3 imes wi h PBS, and incuba ed wi h Alexa 488 and 594-conjuga ed seconda y an ibodies (In i ogen) a oom empe a u e in he da k. A e 3 washes wi h PBS, he co e slips we e moun ed on o slides using Vec ashield moun ing solu ion con aining DAPI (Vec o Labs) and images we e cap u ed on a Leica LAS AF SP5 con ocal mic oscope. N2aSw cells ans ec ed wi h PS1-GFP we e also pe meabilized, ehyd a ed and analyzed as abo e. MDCK cells exp essing APP-GFP we e ixed wi h 4% pa a o maldehyde (Sigma) o 30 min and moun ed on o slides using Vec ashield moun ing solu ion (Vec o Labs). Images we e cap u ed by con ocal mic oscopy (Bec on Dickinson) and he numbe o luo escen APP-con aining esicles pe cell was calcula ed wi h A oVision so wa e (Bec on Dickinson). 2.8. Amyloid-β42 pep ide p epa a ion Monome ic, oligome ic and ib illa Aβ we e p epa ed as desc ibed p e iously [38]. An Αβ42 s ock solu ion was p epa ed by dissol ing he pep ide a 1 mg/ml in hexa luo o-2-p opanol (HFIP; Sigma) o ende he Aβ monome ic. HFIP was emo ed unde acuum in a SpeedVac (Millipo e) and he pep ide ilm was s o ed desicca ed a − 20 °C. Fo he Aβ42 monome p epa a ion, he pep ide ilm was esuspended in DMSO jus p io o use a a concen a ion o 5 mM and i was sonica ed o 10 min in a ba h- ype ul asound de ice. Oligome ic Aβ was gene a ed by dilu ing he monome ic p epa a ion o 100 μM in 150 mM NaCl, 1 mM EDTA, 10 mM T is–HCl [pH 7.4], which was incuba ed o 24 h a 4 °C. Finally, ib illa Aβ was gene a ed by adding 10 mM HCl o he ini ial solu ion o monome ic Aβ o ob ain 100 μM Aβ [pH 2.0], and his solu ion was hen incuba ed o 48 h a 37 °C. 2.9. Aβ binding assays o lipid esicles 2.9.1. P epa a ion o lipid esicles La ge unilamella esicles (LUVs) we e gene a ed as desc ibed p e iously [39]. LUVs composed o sphingomyelin (SM) and Cho (1:1 mol a io) o SM:Cho:OHDHA (1:1:0.1 mol a io) we e ex uded in 10 mM Hepes, 1 mM EDTA, 100 mM NaCl [pH 7.4] using an ex ude wi h 200 nm il e s. The lipid composi ion o LUVs was quan i ied as desc ibed p e iously [40], showing ha he inal composi ion did no di e signi ican ly om he ini ial lipid mix u e. ACCEPTED MANUSCRIPT 7 2.9.2. Iso he mal Ti a ion Calo ime y (ITC) ITC measu emen s we e pe o med in a VP-ITC Mic o-calo ime e (Mic oCal, Inc., No hamp on, USA) as desc ibed p e iously [41], [42]. B ie ly, he expe imen s we e se -up wi h 23 μM o Aβ pep ide (monome , oligome o ib ils) in he cell a 37 °C (cell olume: 1.4 ml) and 35 mM o LUV in he sy inge. Thi y injec ions o 10 μl we e adminis e ed a an in e al be ween injec ions o 10 min. All he he modynamic pa ame e s we e calcula ed using Mic oCal O igin so wa e. The binding cons an Ka (Kd = 1 / Ka) and en halpy (ΔH) we e ob ained om he i ing o ITC iso he ms, and he Gibbs ee ene gy (ΔG) and en opy (ΔS) o binding we e de e mined om he exp ession: 𝛥𝐺 =𝛥𝐻 − 𝛵𝛥𝑆 = −𝑅𝑇𝑙𝑛𝐾𝑎 whe e R is he gas cons an and T is empe a u e. 2.10. Cell iabili y (MTT) Cell iabili y was de e mined using he MTT (me hyl- hiazolyl diphenyl e azolium b omide) me hod, as desc ibed p e iously [32]. The MTT eagen (Sigma) was dilu ed o a inal concen a ion o 0.5 mg/ml in PBS and added o he cell cul u e o 2 h. Mi ochond ial dehyd ogenases in iable cells educed he e azolium sal , yielding wa e insoluble colo ed o mazan c ys als. The MTT eagen was hen emo ed and he o mazan c ys als solubilized by adding one olume o DMSO o 5 min, and a e gen le shaking, he abso bance o he solu ion was de e mined a 570 nm. 2.11. Zeb a ish emb yo oxici y E alua ion o OHDHA oxici y in zeb a ish emb yos was pe o med as desc ibed p e iously [43] and ollowing he OECD (O ganiza ion o Economic Co-ope a ion and De elopmen ) D a Guidelines. B ie ly, i e concen a ions o OHDHA we e added o di e en anks o wa e : 1, 3, 10, 30 and 100 mg/l. The OHDHA was p ima ily dissol ed in DMSO o ob ain he wo king s ocks, keeping he inal concen a ion o DMSO always below 0.1% in he oxici y assays. Each concen a ion o OHDHA was es ed in 20 indi idual emb yos o 24, 48 and 72 h a 25 °C. A e incuba ion, he pe cen age o emb yos showing emb yo mo ali y, sub-le hal e ec s, e a ogenic e ec s and iabili y was de e mined. 2.12. S a is ical analysis Da a we e exp essed as he mean ± SEM. Compa ison be ween wo g oups o da a was pe o med using a wo- ailed - es . Fo compa ison be ween se e al g oups, we used one-way ANOVA ollowed by Tukey's pos hoc mul iple compa isons (S a g aphics plus 3.1). The le el o signi icance was se a 95% o con idence (p < 0.05). 3. Resul s 3.1. E ec s o OHDHA on memb ane lipid composi ion in he b ains o 5xFAD mice The memb ane lipid species in he b ains o WT, 5xFAD and OHDHA- ea ed 5xFAD mice we e analyzed by lipidomic analysis (Fig. 1, Table 1 and Fig. S1). The mos abundan lipid species de ec ed we e he di e en classes o PE (see Fig. S1), wi h all PE species oge he cons i u ing app oxima ely hal o he o al memb ane lipids in he b ain o WT mice (47.77 ± 4.26%). A signi ican inc ease in o al PE le els in he b ain o 5xFAD mice was obse ed a e ea men wi h OHDHA (47.26 ± 6.11% inc ease wi h espec o un ea ed 5xFAD mice), and hese OHDHA- ea ed mice also showed highe PE le els compa ed o WT mice (43.80 ± 5.96%; Fig. 1A). In addi ion, he ele a ed PE le els in he b ain o OHDHA- ea ed mice we e due o a gene al inc ease in all analyzed PE species (Fig. S1). T ea men wi h OHDHA also inc eased he concen a ion o PUFA-con aining phospholipids (con aining 5 o mo e double bonds, e.g. DHA o EPA-con aining phospholipids) compa ed o WT and 5xFAD un ea ed mice, whe eas i had a smalle impac on sa u a ed a y acid-con aining phospholipids (Figs. 1B and C). The highe le el o PUFA -con aining phospholipids in he b ain o 5xFAD mice was mainly due o he inc ease in PE species p o oked by OHDHA (see Table 1). In his sense, he mos impo an inc ease among diacyl-PE subspecies was obse ed in PE 40:6 (an 85.2 ± 8.7% inc ease wi h espec o he WT), while o he polyunsa u a ed diacyl-PE subspecies also showed impo an ele a ions in he b ain o 5xFAD mice a e OHDHA ea men (40:5, 38:5, 38:4, 36:5 and 36:4; Table 1). This esul is compa ible wi h he ele a ed le els o DHA in lipid memb anes a e OHDHA ea men since diacyl-PE 40:6 may be cons i u ed p incipally by 22:6 (DHA) and 18:0 a y acids. Indeed, ou lipidomic analysis ACCEPTED MANUSCRIPT 8 also e ealed ha lyso-PE 18:0 is he mos abundan subspecie among all he lyso-PE lipids (no shown), such ha he diacyl-PE 40:6 de ec ed mus ca y DHA a posi ion C2 o glyce ol. In he same con ex , an analysis o he hyd oca bon chain leng h o Table 1-showed diacyl-PE subspecies e ealed ha long (40C) and medium (36–38C)-chain- a y acid-con aining PEs inc eased mo e s ongly in he b ain o 5xFAD mice a e OHDHA ea men han sho -chain (32–34C) a y-acid-con aining PEs (a 50.59 ± 8.04% and 40.47 ± 7.04% inc ease, espec i ely, compa ed o he WT). Again, no signi ican di e ences we e obse ed when compa ing WT and 5xFAD con ol mice (Table 1). All hese da a demons a e ha OHDHA p oduces an ele a ion in he PE phospholipids in he b ain o 5xFAD mice ha p e e en ially con ain long/middle polyunsa u a ed acyl chains (40:6, 40:5, 38:5, 38:4, 36:5 and 36:4). Finally, a signi ican inc ease in he le els o phospha idylinosi ol (PI) in he b ain o 5xFAD mice was also associa ed wi h OHDHA ea men (Fig. 1E). By con as , no signi ican changes we e obse ed in o he phospholipid species, such as phospha idylcholine (PC) o phospha idylse ine (PS: Figs. 1D and F), no we e he e signi ican di e ences in o he lipid classes, such as choles e ol, sphingolipids (sphingomyelin and dihyd osphingomyelin) and ce amides (ce amide, hexosyl-ce amides and lac osyl- ce amide) (Figs. 1G, H and I, espec i ely). Howe e , he highe le els o PE in he b ain o OHDHA- ea ed mice induced modes gene al dec eases in o he lipids, such as Cho, diacyl-PC and SM (see Fig. S1). In conclusion, changes in lipid memb ane composi ion induced by OHDHA may in luence he lipid memb ane s uc u e, p o oking he o ma ion o liquid diso de ed p one s uc u es [44]. 3.2. OHDHA ea men s induce a educ ion in Aβ le els and au p o ein phospho yla ion in ansgenic 5xFAD mice and cellula models o AD The p oduc ion and accumula ion o Aβ is one o he mos s udied neu opa hological ea u es o AD. The accumula ion o Aβ co ela es s ongly wi h synap ic degene a ion, and wi h memo y and lea ning de ici s in he 5xFAD model [23], [45], [46]. Indeed, se ious cogni i e de ici s we e epo ed p e iously in hese mice when cogni i e impai men was analyzed using he adial a m maze es , as e lec ed by he inc ease in he ime spen in comple ing he es and he numbe o e o s compa ed o WT mice [19]. On he o he hand, enzymes implica ed in he amyloidogenic ou e (β- and γ-sec e ases) a e bo h in eg al memb ane p o eins ha a e highly egula ed by he memb ane lipid en i onmen [47], [48]. In his con ex , we hypo hesized ha OHDHA-media ed memb ane lipid modi ica ions migh in luence Aβ gene a ion/accumula ion in he b ain o 5xFAD mice. Thus, we in es iga ed he e ec o OHDHA on Aβ p oduc ion in 5xFAD b ain samples. As expec ed, we only iden i ied monome ic Aβ pep ide in 5xFAD b ain samples (Fig. 2A, uppe panel), in which he e was a modes , ye signi ican down egula ion o Aβ le els ollowing OHDHA ea men (Fig. 2A, middle panel). In addi ion, human APP ansgene exp ession emained unmodi ied by OHDHA in hese samples (see Fig. S2), demons a ing ha Aβ le els a e no a ec ed by changes in human APP gene exp ession. In pa allel wi h he educ ions in Aβ, OHDHA- ea ed 5xFAD mice also displayed a signi ican eco e y o lea ning and memo y capabili ies compa ed o he 5xFAD con ol mice, e iden h ough he educ ion in he ime spen and wo king/ e e ence memo y e o s commi ed when comple ing he beha io al es (p e iously epo ed by [19], see discussion below). Thus, hese esul s sugges ha he modi ica ions o lipids in b ain memb anes a e associa ed wi h a dec ease in Aβ le els and imp o ed cogni i e sco es in his AD model. The e ec o DHA (and p obably OHDHA) on Aβ p oduc ion is s ongly in luenced by die a y lipid in ake. Indeed, his e ec migh be unde es ima ed due o he p esence o speci ic lipids in he die ha s imula e amyloidogenesis [49]. The e o e, we u he s udied his issue in a cellula model o neu oblas oma cells s ably exp essing he Swedish-mu an o m o APP (N2aSw). In his con ex , he Aβ p oduced and sec e ed in o he cell cul u e medium was immunop ecipi a ed and analyzed by immunoblo ing. Cell cul u es we e ea ed wi h 5 and 10 μM o OHDHA, which led o a d as ic educ ion in Aβ in he medium o e 24 h (Fig. 2B). We also de e mined he soluble APPα (sAPPα) le els in he cul u e medium (Fig. 2B) and he ull leng h APP ( lAPP) in cell ex ac s (no shown) using he 6E10 an ibody. Unlike p e ious epo s [12], we ound no ema kable changes in ei he sAPPα o lAPP, sugges ing ha he non-amyloidogenic ou e is appa en ly una ec ed by OHDHA and ha he changes obse ed in Aβ a e no due o he down egula ion o human APP exp ession. These da a con i med ou p e ious esul s in he b ain o OHDHA- ea ed 5xFAD mice. On he o he hand, au phospho yla ion was es ed in he b ain o 5xFAD mice a Se 202/Th 205 (AT8 epi ope) and Se 202 (CP13 epi ope). In e es ingly, AT8 an ibody e ie ed none o e y low signal when using hese b ain samples (no shown). Howe e , when using CP13 an ibody ou expe imen s showed a clea band o ca. 64 kDa, especially in 5xFAD mice (Fig. 2A, uppe panel). Indeed, au phospho yla ion ACCEPTED MANUSCRIPT 15 [22] P.V. Esc ibá, Memb ane-lipid he apy: a new app oach in molecula medicine, T ends Mol. Med. 12 (2006) 34–43. [23] H. Oakley, S.L. Cole, S. Logan, E. Maus, P. Shao, J. C a , A. Guilloze -Bongaa s, M. Ohno, J. Dis e ho , L. Van Eldik, R. Be y, R. Vassa , In aneu onal be a-amyloid agg ega es, neu odegene a ion, and neu on loss in ansgenic mice wi h i e amilial Alzheime 's disease mu a ions: po en ial ac o s in amyloid plaque o ma ion, J. Neu osci. 26 (2006) 10129–10140. [24] B.A. Wi sching, R.J. Beninge , K. Jhamandas, R.J. Boegman, S.R. El-De awy, Di e en ial e ec s o scopolamine on wo king and e e ence memo y o a s in he adial maze, Pha macol. Biochem. Beha . 20 (1984) 659–662. [25] E.G. Bligh, W.J. Dye , A apid me hod o o al lipid ex ac ion and pu i ica ion, Can. J. Biochem. Physiol. 37 (1959) 911–917. [26] J.M. Jiménez-López, M.P. Ca asco, C. Ma co, J.L. Sego ia, Hexadecylphosphocholine dis up s choles e ol homeos asis and induces he accumula ion o ee choles e ol in HepG2 umou cells, Biochem. Pha macol. 71 (2006) 1114–1121. [27] D. Canals, D.Mo meneo, G. Fab iàs, A. Lleba ia, J. Casas, A. Delgado, Syn hesis and biological p ope ies o Pachas issamine (jaspine B) and dias e eoisome ic jaspines, Bioo g. Med. Chem. 17 (2009) 235–241. [28] A. Ga an o, N.A. Mandal, M. Egido-Gabás, G. Ma any, G. Fab iàs, R.E. Ande son, J. Casas, R. Gonzàlez-Dua e, Speci ic sphingolipid con en dec ease in Ce kl knockdown mouse e inas, Exp. Eye Res. 110 (2013) 96–106. [29] R.E. Kings on, C.A. Chen, H. Okayama, Cu en P o ocols in Immunology, JohnWiley & Sons, Inc., 2001 [30] A. Jämsä, K. Hasslund, R.F. Cowbu n, A. Bäcks öm,M. Vasänge, The e inoic acid and b ain- de i ed neu o ophic ac o di e en ia ed SH-SY5Y cell line as a model o Alzheime 's disease-like au phospho yla ion, Biochem. Biophys. Res. Commun. 319 (2004) 993–1000. [31] R.M. Gill, R. Slack, M. Kiess, P.A. Hamel, Regula ion o exp ession and ac i i y o dis inc pRB, E2F, D- ype cyclin, and CKI amily membe s du ing e minal di e en ia ion o P19 cells, Exp. Cell Res. 244 (1998) 157–170. [32] A. Ma cilla-E xenike, M.L. Ma ín, M.A. Nogue a-Sal à, J.M. Ga cía-Ve dugo, M. So iano- Na a o, I. Dey, P.V. Esc ibá, X. Busque s, 2-Hyd oxyoleic acid induces ER s ess and au ophagy in a ious human glioma cell lines, PLoS One 7 (2012) e48235. [33] M. Sas e, I. Dewach e , G.E. Land e h,T.M. Willson, T. Klockge he , F. an Leu en, M.T. Heneka, Nons e oidal an i-in lamma o y d ugs and pe oxisome p oli e a o -ac i a ed ecep o -gamma agonis s modula e immunos imula ed p ocessing o amyloid p ecu so p o ein h ough egula ion o be a- sec e ase, J.Neu osci. 23 (2003) 9796–9804. [34] L. Ka sou i, C. Pa , N. Bogdano ic, M. Willem, M. Sas e, PPARγ co-ac i a o -1α (PGC-1α) educes amyloid-β gene a ion h ough a PPARγ-dependen mechanism, J. Alzheime s Dis. 25 (2011) 151–162. [35] S. Taylo , M. Wakem, G. Dijkman, M. Alsa aj, M. Nguyen, A p ac ical app oach o RT-qPCR- Publishing da a ha con o m o he MIQE guidelines, Me hods 50 (2010) S1–S5. [36] B. Ramos, D. Baglie o-Va gas, J.C. del Rio, I. Mo eno-Gonzalez, C. San a-Ma ia, S. Jimenez, C. Caballe o, J.F. Lopez-Tellez, Z.U. Khan, D. Ruano, A. Gu ie ez, J. Vi o ica, Ea ly neu opa hology o soma os a in/NPY GABAe gic cells in he hippocampus o a PS1xAPP ansgenic model o Alzheime 's disease, Neu obiol. Aging 27 (2006) 1658–1672. [37] M. To es, S. Jiménez, R. Sánchez-Va o, V. Na a o, L. T ujillo-Es ada, E. Sánchez-Mejias, I. Ca mona, J.C. Da ila, M. Vizue e, A. Gu ié ez, J. Vi o ica, De ec i e lysosomal p o eolysis and axonal anspo a e ea ly pa hogenic e en s ha wo sen wi h age leading o inc eased APP me abolism and synap ic Abe a in ansgenic APP/PS1 hippocampus, Mol. Neu odegene . 7 (2012) 59. [38] M. Gobbi, F. Re, M. Cano i, M. Beeg, M. G ego i, S. Sesana, S. Sonnino, D. B ogioli, C. Musican i, P. Gasco, M. Salmona, M.E. Masse ini, Lipid-based nanopa icles wi h high binding a ini y o amyloid- be a1-42 pep ide, Bioma e ials 31 (2010) 6519–6529. [39] L.D. Maye , M.J. Hope, P.R. Cullis, Vesicles o a iable sizes p oduced by a apid ex usion p ocedu e, Biochim. Biophys. Ac a 858 (1986) 161–168. [40] M.B. Ruiz-A güello, G. Basáñez, F.M. Goñi, A. Alonso, Di e en e ec s o enzyme-gene a ed ce amides and diacylglyce ols in phospholipid memb ane usion and leakage, J. Biol. Chem. 271 (1996) 26616–26621. [41] C. A nulphi, J. So ,M. Ga cía-Pacios, J.-L.R. A ondo, A. Alonso, F.M.Goñi, T i on X-100 pa i ioning in o sphingomyelin bilaye s a subsolubilizing de e gen concen a ions: e ec o lipid phase and a compa ison wi h dipalmi oylphospha idylcholine, Biophys. J. 93 (2007) 3504–3514. ACCEPTED MANUSCRIPT 16 [42] H. Ahyayauch, M. Raab, J.V. Bus o, N. And aka, J.-L.R. A ondo, M. Masse ini, I. T a oska, F.M. Goñi, Binding o β-amyloid (1–42) pep ide o nega i ely cha ged phospholipid memb anes in he liquid- o de ed s a e: modeling and expe imen al s udies, Biophys. J. 103 (2012) 453–463. [43] L. Yang, N.Y. Ho, R. Alshu , J. Leg adi, C. Weiss, M. Reischl, R. Miku , U. Liebel, F. Mülle , U. S ähle, Zeb a ish emb yos as models o emb yo oxic and e a ological e ec s o chemicals, Rep od. Toxicol. 28 (2009) 245–253. [44] M. Iba gu en, D.J. López, J.A. Encina , J.M. González-Ros, X. Busque s, P.V. Esc ibá, Pa i ioning o liquid-o de ed/liquid-diso de ed memb ane mic odomains induced by he luidi ying e ec o 2- hyd oxyla ed a y acid de i a i es, Biochim. Biophys. Ac a 1828 (2013) 2553–2563. [45] C.Y. Shao, S.S.Mi a,H.B.R. Sai ,T.C. Sack o , E.M. Sigu dsson, Pos synap ic degene a ion as e ealed by PSD-95 educ ion occu s a e ad anced Aβ and au pa hology in ansgenic mouse models o Alzheime 's disease, Ac a Neu opa hol. 122 (2011) 285–292. [46] S. Jawha , A. T awicka, C. Jenneckens, T.A. Baye , O. Wi hs, Mo o de ici s, neu on loss, and educed anxie y coinciding wi h axonal degene a ion and in aneu onal Aβ agg ega ion in he 5XFAD mouse model o Alzheime 's disease, Neu obiol. Aging 33 (2012) e129–140(196). [47] P. Osenkowski,W. Ye, R. Wang, M.S.Wol e, D.J. Selkoe, Di ec and po en egula ion o gamma- sec e ase by i s lipid mic oen i onmen , J. Biol. Chem. 283 (2008) 22529–22540. [48] L. Kal odo a, N. Kahya, P. Schwille, R. Ehehal , P. Ve kade, D. D echsel, K. Simons, Lipids as modula o s o p o eoly ic ac i i y o BACE: in ol emen o choles e ol, glycosphingolipids, and anionic phospholipids in i o, J. Biol. Chem. 280 (2005) 36815–36823. [49] Z. Am ul, M. Kee , L. Wang, P. Me i ield, D. Wes away, R.F. Rozmahel, DHA supplemen ed in pep amen die o e s no ad an age in pa hways o amyloidosis: is i ime o e alua e composi e lipid die ? PLoS One 6 (2011) e24094. [50] B. De S oope , Aph-1, Pen-2, and Nicas in wi h P esenilin gene a e an ac i e gamma-Sec e ase complex, Neu on 38 (2003) 9–12. [51] S.L. Cole, R. Vassa , BACE1 s uc u e and unc ion in heal h and Alzheime 's disease, Cu . Alzheime Res. 5 (2008) 100–120. [52] C. Haass, D.J. Selkoe, Soluble p o ein oligome s in neu odegene a ion: lessons om he Alzheime 's amyloid be a-pep ide, Na . Re . Mol. Cell Biol. 8 (2007) 101–112. [53] R. Williamson, A. Usa di, D.P. Hange , B.H. Ande on, Memb ane-bound β-amyloid oligome s a e ec ui ed in o lipid a s by a yn-dependen mechanism, FASEB J. 22 (2008) 1552–1559. [54] T. Kawa abayashi, M. Shoji, L.H. Younkin, L. Wen-Lang, D.W. Dickson, T. Mu akami, E. Ma suba a, K. Abe, K.H. Ashe, S.G. Younkin, Dime ic amyloid β p o ein apidly accumula es in lipid a s ollowed by apolipop o ein E and phospho yla ed au accumula ion in he Tg2576 mouse model o Alzheime 's disease, J. Neu osci. 24 (2004) 3801–3809. [55] B. De S oope , R. Vassa , T. Golde, The sec e ases: enzymes wi h he apeu ic po en ial in Alzheime disease, Na . Re . Neu ol. 6 (2010) 99–107. [56] C. Gao, C. Hölsche , Y. Liu, L. Li, GSK3: a key a ge o he de elopmen o no el ea men s o ype 2 diabe es melli us and Alzheime disease, Re . Neu osci. 23 (2012) 1–11. [57] B.D. Le y, Resol ins and p o ec ins: na u al pha macopho es o esolu ion biology, P os aglandins Leuko . Essen . Fa y Acids 82 (2010) 327–332. [58] K.H. Weyland , C.-Y. Chiu, B. Gomolka, S.F. Waech e , B. Wiedenmann, Omega-3 a y acids and hei lipid media o s: owa ds an unde s anding o esol in and p o ec in o ma ion, P os aglandins O he Lipid Media . 97 (2012) 73–82. [59] M. Söde be g, C. Edlund, K. K is ensson, G. Dallne , Fa y acid composi ion o b ain phospholipids in aging and in Alzheime 's disease, Lipids 26 (1991) 421–425. [60] Z. Guan, Y. Wang, N.J. Cai ns, P.L. Lan os, G. Dallne , P.J. Sindela , Dec ease and s uc u al modi ica ions o phospha idyle hanolamine plasmalogen in he b ain wi h Alzheime disease, J. Neu opa hol. Exp. Neu ol. 58 (1999) 740–747. [61] M.R. P asad, M.A. Lo ell, M. Ya in, H. Dhillon, W.R. Ma kesbe y, Regional memb ane phospholipid al e a ions in Alzheime 's disease, Neu ochem. Res. 23 (1998) 81–88. [62] J.T. Wood, J.S. Williams, L. Panda ina han, D.R. Jane o, C.J. Lammi-Kee e, A. Mak iyannis, Die a y docosahexaenoic acid supplemen a ion al e s selec physiological endocannabinoid-sys em me aboli es in b ain and plasma, J. Lipid Res. 51 (2010) 1416–1423. [63] I. O es ad, S. Hassani, G.I. Bo ge, A. Kohle , G. Vog , T. Hyö yläinen, M. O ešič, K.W. B ønne , K.B. Hol en, S.M. Ul en, M.C.W. Myh s ad, Fish oil supplemen a ion al e s he plasma lipidomic p o ile and inc eases long-chain PUFAs o phospholipids and iglyce ides in heal hy subjec s, PLoS One 7 (2012) e42550. ACCEPTED MANUSCRIPT 17 [64] M. G ønn, E. Ch is ensen, T.A. Hag e, B.O. Ch is ophe sen, Pe oxisomal e ocon e sion o docosahexaenoic acid (22:6(n-3)) o eicosapen aenoic acid (20:5(n-3)) s udied in isola ed a li e cells, Biochim. Biophys. Ac a 1081 (1991) 85–91. [65] R.B. Chan, T.G. Oli ei a, E.P. Co es, L.S. Honig, K.E. Du , S.A. Small, M.R. Wenk, G. Shui, G. Di Paolo, Compa a i e lipidomic analysis o mouse and human b ain wi h Alzheime disease, J. Biol. Chem. 287 (2012) 2678–2688. [66] A.A. Spec o , M.A. Yo ek, Memb ane lipid composi ion and cellula unc ion, J. Lipid Res. 26 (1985) 1015–1035. [67] Y. Tajima, M. Ishikawa, K.Maekawa,M.Mu ayama, Y. Senoo, T. Nishimaki-Mogami,H. Nakanishi, K. Ikeda, M. A i a, R. Taguchi, A. Okuno, R. Mikawa, S. Niida, O. Takikawa, Y. Sai o, Lipidomic analysis o b ain issues and plasma in amousemodel exp essingmu a ed human amyloid p ecu so p o ein/ au o Alzheime 's disease, Lipids Heal h Dis. 12 (2013) 68. [68] C. Sahlin, F.E. Pe e sson, L.N.G. Nilsson, L. Lann el , A.-S. Johansson, Docosahexaenoic acid s imula es non-amyloidogenic APP p ocessing esul ing in educed Abe a le els in cellula models o Alzheime 's disease, Eu . J. Neu osci. 26 (2007) 882–889. [69] O. Holmes, S. Pa u i, W. Ye, M.S. Wol e, D.J. Selkoe, E ec s o memb ane lipids on he ac i i y and p ocessi i y o pu i ied γ-sec e ase, Biochemis y 51 (2012) 3565–3575. [70] S.R. Shaikh, Biophysical and biochemical mechanisms by which die a y N-3 polyunsa u a ed a y acids om ish oil dis up memb ane lipid a s, J. Nu . Biochem. 23 (2012) 101–105. [71] K.S. Ve i el, G. Thinaka an, Memb ane a s in Alzheime 's disease be a-amyloid p oduc ion, Biochim. Biophys. Ac a 1801 (2010) 860–867. [72] H. Schulze, T. Kol e , K. Sandho , P inciples o lysosomal memb ane deg ada ion: cellula opology and biochemis y o lysosomal lipid deg ada ion, Biochim. Biophys. Ac a 1793 (2009) 674–683. [73] R.A. Nixon, J.Wegiel, A. Kuma ,W.H. Yu, C. Pe e ho , A. Ca aldo, A.M. Cue o, Ex ensi e in ol emen o au ophagy in Alzheime disease: an immuno-elec on mic oscopy s udy, J. Neu opa hol. Exp. Neu ol. 64 (2005) 113–122. [74] R. Sanchez-Va o, L. T ujillo-Es ada, E. Sanchez-Mejias, M. To es, D. Baglie o-Va gas, I. Mo eno-Gonzalez, V. De Cas o, S. Jimenez, D. Ruano, M. Vizue e, J.C. Da ila, J.M. Ga cia-Ve dugo, A.J. Jimenez, J. Vi o ica, A. Gu ie ez, Abno mal accumula ion o au ophagic esicles co ela es wi h axonal and synap ic pa hology in young Alzheime 's mice hippocampus, Ac a Neu opa hol. 123 (2012) 53–70. [75] V. Schae e , I. La eni , S. Ozcelik, M. Tolnay, D.T.Winkle ,M. Goede , S imula ion o au ophagy educes neu odegene a ion in a mouse model o human auopa hy, B ain 135 (2012) 2169–2177. [76] J.V. Rushwo h, H.H. G i i hs, N.T. Wa , N.M. Hoope , P ion p o ein-media ed oxici y o amyloid-β oligome s equi es lipid a s and he ansmemb ane LRP1, J. Biol. Chem. 288 (2013) 8935– 8951. [77] M.L. Giu ida, F. Ca aci, B. Pigna a o, S. Ca aldo, P. De Bona, V. B uno, G.Molina o, G. Pappala do, A. Messina, A. Palmigiano, D. Ga ozzo, F. Nicole i, E. Rizza elli, A. Copani, Be a-amyloid monome s a e neu op o ec i e, J. Neu osci. 29 (2009) 10582–10587. [78] R. Resende, C. Pe ei a, P. Agos inho, A.P. Viei a, J.O.Mal a, C.R. Oli ei a, Suscep ibili y o hippocampal neu ons o Abe a pep ide oxici y is associa ed wi h pe u ba ion o Ca2+ homeos asis, B ain Res. 1143 (2007) 11–21. [79] L. Texidó, M. Ma ín-Sa ué, E. Albe di, C. Solsona, C. Ma u e, Amyloid β pep ide oligome s di ec ly ac i a e NMDA ecep o s, Cell Calcium 49 (2011) 184–190 ACCEPTED MANUSCRIPT 18 Table 1. LC/MS cha ac e iza ion o diacyl-PE acyl chains in he b ain o OHDHA- ea ed 5xFAD mice Diacyl-PE subspecies WT 5xFAD 5xFAD + OHDHA Meana SEMa Meana SEMa % changeb Meana SEMa % changeb Sho acyl chains 32:0 112 3 105 3 − 6.3 ± 3.0 NS 134 3 19.6 ± 3.0 ** 32:1 168 4 205 39 22.2 ± 22.9 NS 240 13 42.5 ± 7.7 NS 34:1 2197 148 1817 71 − 17.3 ± 3.2 NS 3030 188 37.9 ± 8.5 ** 34:2 326 25 372 71 14.4 ± 21.7 NS 402 17 23.5 ± 5.4 NS Sub o al 2803 154 2500 85 − 10.7 ± 3.0 NS 3805 216 35.7 ± 7.7 ** Medium acyl chains 36:1 3862 477 3332 185 − 13.7 ± 4.8 NS 5426 204 40.5 ± 5.3 ** 36:2 3108 195 2584 98 − 16.9 ± 3.2 NS 4319 210 39.0 ± 6.8 ** 36:3 274 30 254 27 − 7.2 ± 9.9 NS 303 22 10.4 ± 8.2 NS 36:4 3184 454 3416 465 7.3 ± 14.6 NS 5390 352 69.3 ± 11.0 * 36:5 201 16 318 69 58.8 ± 34.6 NS 295 11 46.9 ± 5.5 * 38:1 588 12 441 50 − 24.9 ± 8.5 NS 713 53 21.4 ± 9.0 * 38:2 547 18 389 6 − 28.8 ± 1.0 * 593 57 8.5 ± 10.4 NS 38:4 16,804 2171 16,670 1969 − 0.8 ± 11.7 NS 24,716 1072 47.1 ± 6.4 * 38:5 2799 135 2903 143 3.7 ± 5.1 NS 4579 90 63.6 ± 3.2 *** 38:6 13,444 2330 12,244 821 − 8.9 ± 6.1 NS 16,532 1412 23.0 ± 10.5 NS 38:7 244 30 340 100 38.9 ± 40.8 NS 425 19 73.7 ± 7.6 NS Sub o al 45,054 5749 42,891 3716 − 4.8 ± 8.2 NS 63,290 3171 40.4 ± 7.0 * Long acyl chains 40:1 174 13 106 17 − 38.9 ± 9.7 NS 160 23 − 8.0 ± 13.0 NS 40:3 2768 369 2764 315 − 0.2 ± 11.4 NS 4121 257 48.9 ± 9.3 NS 40:5 26,517 3064 26,194 2062 − 1.2 ± 7.8 NS 38,945 2170 46.9 ± 8.2 * 40:6 3289 896 3901 785 18.6 ± 23.9 NS 6089 286 85.2 ± 8.7 * Sub o al 32,747 4196 32,964 3137 0.7 ± 9.6 NS 49,314 2633 50.5 ± 8.0 * To al 80,605 10,036 78,356 6833 − 2.8 ± 8.5 NS 116,411 10,334 44.4 ± 7.4 * Mul iple s a is ical analysis was pe o med wi h one-way ANOVA & Tukey's pos hoc es . NS: no signi ican , * p < 0.05, ** p < 0.01 and *** p < 0.001. a The measu e o diacyl-PE subspecies (mean ± SEM) shown as pmol o lipid/mg o p o ein. b The ela i e change in 5xFAD o OHDHA- ea ed 5xFAD b ains ela i e o he mean WT alue is shown as he mean ± SEM. ACCEPTED MANUSCRIPT 19 Table 2. OHDHA inco po a ion in o lipid a -like esicles inhibi s binding o Aβ oligome s and ib ils. Emp y Cell SM/Cho/OHDHA a SM/Cho a Emp y Cell (1:1:0.1 mol a io) (1:1 mol a io) A. Monome Kd (μM)b 58 ± 7.7 No in e ac ion ΔH (Kcal/mol)b − 3.31 ± 0.24 de ec ed ΔS (cal/mol.°C)b 8.7 ± 0.5 – ΔG (Kcal/mol)b − 6.01 ± 0.40 – B. Oligome Kd (μM)b 21.6 ± 2.8 *** 2.7 ± 0.9 ΔH (Kcal/mol)b − 1.06 ± 0.98 ** − 2.11 ± 0.09 ΔS (cal/mol.°C)b 17.9 ± 3.2 18.6 ± 2.5 ΔG (Kcal/mol)b − 6.61 ± 1.10 − 7.80 ± 0.40 C. Fib ils Kd (μM)b 775 ± 52 *** 4.76 ± 0.11 ΔH (Kcal/mol)b − 2.77 ± 0.24 *** − 0.87 ± 0.06 ΔS (cal/mol.°C)b 5.3 ± 0.9 ** 21.6 ± 1.2 ΔG (Kcal/mol)b − 4.41 ± 1.00 * − 7.56 ± 1.50 The da a a e exp essed as he mean ± SEM. S a is ical analysis was pe o med by wo- ailed - es : * p < 0.05, ** p < 0.01 and *** p < 0.001. a La ge unilamella esicles composed o SM and Cho, plus/minus OHDHA, we e used o iso he mal calo ime ic i a ion, oge he wi h di e en o ms o Aβ42 pep ide: (A) monome ic, (B) oligome ic and (C) ib illa (n = 3). b The pa ame e s measu ed in he assay: Kd (dissocia ion cons an ), ΔH (en halpy change), ΔS (en opy change) and ΔG (Gibbs ee ene gy change) we e ob ained as desc ibed in he Sec ion 2.9. ACCEPTED MANUSCRIPT 20 Table 3. E alua ion o OHDHA oxici y in zeb a ish emb yos. OHDHA (mg/ml) Incuba ion (h) 24 48 72 A. Emb yo mo ali y 1 0 0 0 3 0 0 0 10 0 0 0 30 0 0 10 100 20 20 35 B. Emb yo iabili y 1 – – 100 3 – – 80 10 – – 80 30 – – 94 100 – – 46 C. Suble hal e ec s 1 0 0 0 3 0 0 0 10 0 5 5 30 0 10 0 100 0 56 77 D. Te a ogenic e ec s 1 0 0 5 3 0 0 70 10 5 5 50 30 10 25 100 100 100 66 38 The da a a e exp essed as he pe cen age o emb yos ha showed any o he pa ame e s s udied: (A) emb yo mo ali y, (B) emb yo iabili y, (C) sub-le hal e ec s: lack o spon aneous mo emen s, pigmen a ion de ici s and appa i ion o edema o clo s in in e nal s uc u es, and (D) e a ogenic e ec s: de o ma ion o in e nal s uc u es, scoliosis and gene al g ow h delay. ACCEPTED MANUSCRIPT 21 Table 4. Imp o ed cogni ion and neu ogenesis in OHDHA- ea ed-5xFAD mice. Emp y Cell WT 5xFAD 5xFAD + OHDHA mean ± SEM mean ± SEM % changea mean ± SEM % changeb Time (s) c 75.69 ± 4.77 124.4 ± 12.38 64.35 ± 9.95 ** 105.4 ± 10.34 − 15.27 ± 9.81 NS WME (numbe o e o s) c 2.32 ± 0.32 3.95 ± 0.47 70.25 ± 11.89 * 2.21 ± 0.40 − 44.05 ± 18.09 § RME (numbe o e o s) c 5.88 ± 0.39 8.31 ± 0.64 41.32 ± 7.70 * 6.07 ± 0.73 − 26.94 ± 12.02 § Phospho his one-H3 posi i e neu ons (cells/mm3)d 7431 ± 1918 1970 ± 767.1 − 73.48 ± 38.93 * 5374 ± 939.7 272.79 ± 17.47 § The s a is ical analysis was pe o med wi h a wo- ailed - es : * p < 0.05 and ** p < 0.01 a e compa ison be ween WT and 5xFAD; NS: no signi ican and § p < 0.05 compa ing OHDHA- ea ed- 5xFAD and 5xFAD con ol mice. Summa y o da a ex ac ed om [19]. a Pe cen age change in 5xFAD con ol mice was exp essed as he mean ± SEM ela i e o he WT as a e e ence. b The pe cen age change in 5xFAD + OHDHA was exp essed as he mean ± SEM ela i e o 5xFAD mice. c Time (s), and he numbe o wo king and e e ence memo y e o s (WME and RME, espec i ely) we e de e mined du ing he adial a m maze es (WT, n = 10; 5xFAD, n = 11 and 5xFAD + OHDHA, n = 12). d Neu onal p oli e a ion in he den a e gy us (hippocampus) was de e mined by quan i ying he numbe o phospho-his one H3-posi i e neu ons in he g anulla cell laye o immunolabeled b ain sec ions (WT, n = 4; 5xFAD n = 7 and 5xFAD + OHDHA, n = 7). ACCEPTED MANUSCRIPT 22 Figu e legends Figu e 1. E ec o OHDHA on memb ane lipid composi ion in he b ain o 5xFAD mice. Ba diag ams show he ela i e change in lipids in OHDHA- ea ed (da k g ay ba s) and con ol 5xFAD mice (ligh g ay ba s) compa ed o he WT (conside ed as 100%, illed ba s). LC/MS was used o analyze he le els o : (A) phospha idyle hanolamine (diacyl-PE, lyso-PE and hei plasmalogens), (B) polyunsa u a ed phospholipids (con aining 5 o mo e double bonds), (C) sa u a ed phospholipids, (D) phospha idylcholine (diacyl-PC, lyso-PC and hei plasmalogens), (E) phospha idylinosi ol (diacyl- PI), (F) phospha idylse ine (diacyl-PS and lyso-PS), (H) sphingolipids (sphingomyelin and dihyd osphingomyelin) and (I) ce amides (ce amide, lac osyl-ce amide and hexosyl-ce amides); and (G) choles e ol le els we e de e mined by an enzyma ic colo ime ic me hod. These da a showed a signi ican inc ease in PE, PI and PUFA-con aining phospholipids in OHDHA- ea ed 5xFAD mice compa ed wi h 5xFAD and WT con ol mice (panels A, B and E, espec i ely). No ema kable di e ences be ween 5xFAD and WT samples we e obse ed o any o he lipids s udied. Ba s show he mean alue om 3 animals ± SEM and he s a is ical analysis was pe o med by one-way ANOVA and Tukey's pos hoc mul iple compa ison: *, p < 0.05. Figu e 2. OHDHA diminishes he β-amyloid b ain pa hology and au hype phospho yla ion. (A) Rep esen a i e wes e n blo s showing monome ic Aβ, phospho- au (Se -202; CP13 epi ope) and o al au (Tau46) in 5xFAD b ain samples (same mice used in Fig. 1) wi h ubulin-α used as a loading con ol (uppe panel). Wes e n blo quan i ica ion showed a signi ican dec ease in Aβ in he b ain o 5xFAD mice ha ecei ed OHDHA (middle panel). In he same way, quan i ica ion o phospho and o al au e ealed a signi ican inc ease in phospho- au le els in 5xFAD ha was majo ly p e en ed in OHDHA- ea ed 5xFAD mice, as compa ed wi h WT, whe eas o al- au le els did no show any signi ican di e ences. Consequen ly, he phosho/ o al au a io was showed signi ican ly inc eased in 5xFAD bu no in OHHDA- ea ed 5xFAD mice, as compa ed wi h WT (lowe panel) (ba s show he mean alue om 3 animals ± SEM). (B) Rep esen a i e wes e n blo o Aβ and sAPPα in N2aSw cell cul u e media (uppe panel). Wes e n blo quan i ica ion e ealed ha exposing N2aSw cells o OHDHA (10 μM) o 24 h d as ically educed Aβ p oduc ion wi hou al e ing sAPPα le els (n = 3) (lowe panel). (C) Rep esen a i e wes e n blo o au hype phospho yla ion in neu on-like di e en ia ed SH-SY5Y cells (uppe panel). Wes e n blo quan i ica ion showed a s ong induc ion o au phospho yla ion a he AT8 epi ope a e Aβ s imula ion, whe eas addi ion o OHDHA (5 μM) pa ly educed Aβ-media ed au phospho yla ion and i was comple ely abolished in he p esence o OHDHA (10 μM; n = 3) (lowe panel). #, p < 0.05 acco ding o wo- ailed - es be ween wo g oups; *, p < 0.05 acco ding o mul iple s a is ical analysis by one-way ANOVA and Tukey's pos hoc es . Figu e 3. Subcellula dis ibu ion o APP, PS1 and BACE1 in N2aSw (APPsw) and MDCK (APP-GFP) cell cul u es ea ed wi h OHDHA. (A) Con ocal images showing N2a cells s ably exp essing Swedish APP. Cells we e immunolabeled wi h he 6E10 an ibody o de ec APP (panels A1– A2), wi h an an i-BACE1 an ibody (panels A5–A6) o ans ec ed wi h a PS1-GFP ec o (panels A3– A4). Immuno luo escence assays showed APP, BACE-1 and PS1 clus e in cell esicles/o ganelles ollowing OHDHA ea men (panels A2, A4 and A6, espec i ely) as compa ed o con ol cells (panels A1, A3 and A5, espec i ely). (B) Con ocal images o MDCK cells s ably exp essing APP-GFP and exposed o inc easing concen a ions o OHDHA (B1: 10 μM; B2: 15 μM; B3: 20 μM; B4: 25 μM). Vesicula accumula ion o APP-GFP con i med p e ious esul s wi h N2aSw cells (panel A2). APP clus e ing was e iden in he p esence o OHDHA (10 μM) and he numbe o GFP-posi i e esicles apidly inc eased a highe concen a ions un il sa u a ion. Fluo escen esicle quan i ica ion showed an EC50 alue o 17.8 μM (B5). Scale ba s: 10 μm (A1–A6) and 20 μm (B1–B4). Figu e 4. Fig. 4. OHDHA p o ec ed cells wi h a neu on-like pheno ype agains an Aβ o a NMDA- media ed oxic insul . (A) SH-SY5Y cells we e di e en ia ed in o neu on-like cells and ea ed wi h oligome ic Aβ pep ide (5 μM) o 24 h in he p esence o absence o OHDHA (5 and 10 μM). The MTT assay was used o quan i y he numbe o iable cells/well. Oligome ic Aβ educed cell iabili y bu he numbe o iable cells was eco e ed upon OHDHA ea men o he un ea ed con ol alues (5 μM OHDHA) o e en highe (10 μM OHDHA; n = 3). (B) P19 emb yonic s em cells we e di e en ia ed in o neu on-like cells and exposed o NMDA (in he p esence o 10 mM Ca2 +). Concomi an OHDHA ea men educed NMDA/Ca2 +-induced cell dea h and e en inc eased he numbe o iable cells abo e ha o he un ea ed con ols (n = 3). #, p < 0.05 acco ding o wo- ailed - es be ween wo g oups; *, p < 0.05 compa ed o Aβ- o NMDA- ea ed con ols, acco ding o mul iple s a is ical analysis by one- way ANOVA and Tukey's pos hoc es . ACCEPTED MANUSCRIPT 23 Figu e 5. Pos ula ed mechanism o ac ion o OHDHA. Summa y o he p oposed molecula mechanisms unde lying he e ec s o OHDHA on neu onal memb anes. OHDHA en iches b ain memb anes in PE, especially hose ca ying DHA and o he long PUFAs. These lipid changes may in luence he s uc u e o he cell memb ane, leading o he o ma ion o liquid-diso de ed-p one s uc u es, and hey migh e e se he cellula signaling associa ed wi h AD by: (i) down egula ing amyloidogenic p ocessing and Aβ- dependen au p o ein hype phospho yla ion; and (ii) dec easing neu on ulne abili y o ex acellula oxic agen s such as oligome ic Aβ and NMDA/Ca2 +-media ed exci o oxici y. In addi ion, OHDHA also induced p oli e a ion o bo h neu on-like cells in i o and neu on s em cells in he mouse b ain. Toge he , his e idence suppo s a neu op o ec i e and neu o egene a i e ole o OHDHA ha may be associa ed wi h he imp o ed cogni i e capabili ies seen in he 5xFAD mouse model. Figu e S1. Rela i e abundance o lipid species in he b ain o WT, 5XFAD and OHDHA- ea ed 5xFAD mice. The da a a e exp essed as he pe cen age o one lipid species ela i e o he o al lipids s udied in he mouse b ain. The PE amily (diacyl-PE, lysoPE and hei alkyl/alkenyl-plasmalogens) is he mos abundan , ollowed by Cho, diacyl-PC and SM. OHDHA- ea ed 5xFAD mice showed highe le els o all PE species while o he species such as Cho, diacyl-PC and SM we e sligh ly diminished as a consequence o he high PE le els compa ed o WT and 5xFAD con ol mice. Ba s show he mean alue om 3 animals ± SEM. Figu e S2. Mouse and human APP exp ession in he b ain o WT and 5xFAD mice. The gene ic exp ession o mouse (mAPP) and human APP (hAPP) was de e mined by eal ime PCR using speci ic p ime s. Exp ession o hAPP was exclusi ely de ec ed in 5xFAD mice and no ema kable di e ences we e ound be ween OHDHA- ea ed and 5xFAD con ol mice. Addi ionally, exp ession o he human gene was highe han ha o mAPP in 5xFAD mice. Ba s show he mean alue om 3 animals ± SEM: *, p < 0.05 acco ding o mul iple s a is ical analysis by one-way ANOVA and Tukey's pos hoc es . Figu e S3. Cellula co-localiza ion o APP and LC3/Lamp1 in N2aSw (APPsw) cell cul u es ea ed wi h OHDHA. Con ocal images showing N2a cells s ably exp essing Swedish APP. Double immuno luo escence labeling using mouse an i-APP (clone 6E10; panel A1) and abbi an i-LC3 (au ophagic esicle ma ke ; panel A2), o well, abbi an i-APP C- e minal (R1(57); panel B1) and mouse an i-Lamp1 (lysosomal ma ke ; panel B2), and con ocal lase mic oscopy demons a ed weak co- localiza ion o bo h APP/LC3 (panel A3) o APP/Lamp1 (panel B3) in he same APP-posi i e esicles accumula ed a e OHDHA ea men (10 μM, 24 h). Scale ba s: 10 μm. Figu e S4. Weigh o con ol and OHDHA- ea ed mice. The weigh o he animals was moni o ed h ee imes a week o he whole pe iod o ea men wi h OHDHA (o ehicle alone, con ols). All mice inc eased hei weigh du ing he ea men un il hey s a ed he hypocalo ic die necessa y o pe o m he adial a m maze es (4 h mon h). The da a a e p esen ed as he pe cen age change ela i e o he i s day o ea men . Each poin shows he mean alue om 3 animals ± SEM. ACCEPTED MANUSCRIPT 24 Figu e 1 ACCEPTED MANUSCRIPT 31 Figu e S3 ACCEPTED MANUSCRIPT 32 Figu e S4 ACCEPTED MANUSCRIPT 33 G aphical abs ac ACCEPTED MANUSCRIPT 34 Highligh s  OHDHA inc eases he le el o PE ca ying DHA and o he polyunsa u a ed a y acids.  OHDHA educes Aβ le els and Aβ-induced au phospho yla ion.  OHDHA p o ec s neu on-like cells agains oligome ic-Aβ and NMDA/Ca2+-induced oxici y.  OHDHA educes a ini y o ib illa /oligome ic Aβ o lipid- a memb anes.  OHDHA showed no ema kable oxici y a he apeu ic doses.