NMR-based Metabolomics and Fatty Acid Profiles to Unravel Biomarkers in Preclinical Animal Models of Compulsive Behavior

NMR-based Me abolomics and Fa y Acid P ofiles o Un a el Bioma ke s in P eclinical Animal Models o Compulsi e Beha io Ana C. Ab eu, § San iago Mo a, § Ana Isabel T is án, Elena Ma ín-González, Angeles P ados-Pa do, Ma ga i a Mo eno,*and Ignacio Fe nández* Ci e This: J. P o eome Res. 2022, 21, 612−622 Read Online ACCESS Me ics & Mo e A icle Recommenda ions * sıSuppo ing In o ma ion ABSTRACT: Compulsi i y is a key mani es a ion o inhibi o y con ol defici and a ca dinal symp om o psychopa hological condi ions such as obsessi e- compulsi e and a en ion-defici hype ac i i y diso de s, in which me abolic al e a ions ha e aised a en ion as pu a i e bioma ke s o ea ly iden ifica ion. The p esen s udy assessed he me abolic p ofile in a p eclinical model o a compulsi e pheno ype o a s. We used he schedule-induced polydipsia (SIP) me hod o classi y male Wis a a s in o high d inke s (HDs) o low d inke s (LDs) acco ding o hei compulsi e d inking a e de eloped by exposu e o a fixed- ime 60 s (FT-60) schedule o ein o cemen wi h wa e a ailable ad libi um du ing 20 sessions. Be o e and a e SIP, blood samples we e collec ed o subsequen se um analysis by nuclea magne ic esonance spec oscopy coupled o mul i a ia e analysis. Al hough no diffe ences exis ed in he p e-SIP se , he compulsi e d inking beha io induced ema kable me abolic al e a ions: HD a s selec ed by SIP exhibi ed a hype lipidemic, hypoglycemic, and hype glu amine gic p ofile compa ed wi h hei low-compulsi e coun e pa s. In e es ingly, hese al e a ions we e no a ibu able o he me e exposu e o ewa d pelle s because a con ol expe imen did no show diffe ences be ween HDs and LDs a e 20 sessions o pelle consump ion wi hou in e mi en ein o cemen . Ou esul s shed ligh owa d he implica ion o die a y and me abolic ac o s unde pinning he ulne abili y o compulsi e beha io s. KEYWORDS: compulsi e beha io , schedule-induced polydipsia, bioma ke s, NMR, me abolomics 1. INTRODUCTION Compulsi i y is defined as a pe se e a ion o a esponse ha is i esis ible, inapp op ia e, and una oidable despi e i s nega i e consequences. 1 I is he co e ea u e obse ed in obsessi e- compulsi e diso de (OCD), al hough i is p esen in o he neu opsychopa hological condi ions such as schizoph enia, au ism, a en ion-defici hype ac i i y diso de , and addic- ion. 2−4 These a e conside ed as impulsi e compulsi e spec um diso de s, wi h a high p e alence (1−3%) in Wes e n coun ies and an app oxima e economic cos o $5 billion pe yea acco ding o he Wo ld Heal h O ganiza ion. 5 Howe e , li le is known abou he ela ion be ween me abolic ac o s and inhibi o y con ol defici . Resea ch on he possible iden ifica ion o me abolic bioma ke s unde lying inhibi o y con ol defici and he effec o die on he ne ous sys em and beha io could help scien is s and physicians o imp o e hei knowledge abou new mechanisms o p e en ion and ea men in psychopa ho- logical diso de s. Me abolomics has been widely applied in biomedicine o p o ide a p ecise analysis o small molecules (<1500 Da) associa ed wi h human me abolism. Unlike DNA, RNA, o p o eins, me aboli es can accu a ely eflec he mos di ec me abolic changes in ou body unde a ce ain condi ion in a sho ime pe iod and a e hus good indica o s o he onse and p og ession o human diseases. 6 So, he analysis o me aboli es ep esen s a sensi i e measu e o biological s a us in heal h o unde disease. 7 Nuclea magne ic esonance (NMR) spec os- copy offe s he unique p ospec o holis ically sc een se e al me aboli es wi h a non-a-p io i selec ion in di e se ma ices as biological fluids and e en in issue biopsies. 8 NMR-based me abolomics analysis coupled o mul i a ia e s a is ical echniques ha e been inco po a ed in o ag icul u al cul u e p og ams and clinical disease esea ch o iden i y unique me aboli e bioma ke s in he ques o nu i ional o o gano- lep ic ad an ages o in a specific human disease, espec i ely. When he ocus is placed on he iden ifica ion o disease- associa ed bioma ke s, his could allow us o (1) p edic and diagnose diseases and hei s ages, (2) p o ide insigh s in o unde lying pa hways in he pa hogenesis and p og ession o he Special Issue: Me abolomics Resea ch Recei ed: No embe 3, 2021 Published: Feb ua y 10, 2022 A iclepubs.acs.o g/jp © 2022 The Au ho s. Published by Ame ican Chemical Socie y 612 h ps://doi.o g/10.1021/acs.jp o eome.1c00857 J. P o eome Res. 2022, 21, 612−622 Downloaded ia UNIV DE ALMERIA on Oc obe 21, 2025 a 08:53:41 (UTC). See h ps://pubs.acs.o g/sha ingguidelines o op ions on how o legi ima ely sha e published a icles. diseases, and (3) aid in disease ea men by assessing he efficacy and mechanism o ac ion o he apeu ic solu ions. 6,9 The use o animal models wi h highe ansla ional powe , associa ed wi h neu obeha io al endopheno ypes based on bioma ke s, has enabled us o b idge he gap be ween (dys) unc ional neu onal ci cui s and psychological cons uc s wi h bigge p edic i e s eng h han he cu en psychia ic nosology. 10 Howe e , o he bes o ou knowledge, only a ew s udies ha e in es iga ed me abolomic bioma ke s in animal models. In a ecen s udy, Pe ez-Fe nandez e al. 11 ound a hype lipidemic and hypoglycemic p ofile in animals exposed o Chlo py i os, which also showed beha io al al e a ions, ha is, impai men s in he eac ion o social no el y in he C awley social es . Schedule-induced polydipsia (SIP) is one o he mos well- es ablished p eclinical models o he s udy o neu opsychopa- hological diso de s p esen ing compulsi e beha io , such as OCD, schizoph enia, and alcohol abuse. (Fo a e iew, see Mo eno and Flo es). 12 The SIP p ocedu e is cha ac e ized by he de elopmen o an adjunc i e beha io o excessi e and pe sis en d inking, which is non egula o y and does no ely on physiological demands, in ood-dep i ed animals exposed o in e mi en ood- ein o cemen schedules in ope an chambe s wi h wa e a ailable ad libi um. 13,14 Howe e , he e a e impo an indi idual diffe ences in he de elopmen o adjunc i e d inking a e 15−20 sessions in SIP. Ra s can be sepa a ed in o wo g oups acco ding o hei a es o d inking, one wi h high o compulsi e d inking (HD a s) and a second g oup wi h low d inking o no SIP acquisi ion (LD a s). This phenomenon esembles he key ea u es o human compulsi i y and OCD, hus allowing us o iden i y ulne able popula ions based on hei endopheno ype ( o a e iew, see Mo eno and Flo es) 12 and en ailing an ou s anding chance o s udying compulsi e beha io unde labo a o y condi ions. In his sense, he p omising ole o me abolism as a pu a i e con ibu ing ac o conce ning ulne abili y o inhibi o y con ol defici necessi a es u he esea ch; specifically, he use o bioma ke s o he ea ly iden ifica ion o such condi ions c ea es he con ex o a be e unde s anding o he p oblem and o ea ly de ec ion and, in he u u e, ea ly in e en ion. The p esen s udy aims o add e idence in ha di ec ion wi h he aid o SIP. 2. MATERIALS AND METHODS 2.1. Subjec s Fo y male Wis a a s (En igo, Spain) we e used in his s udy (40 o he SIP expe imen , 20 o hem o he con ol expe imen ) and a i ed a he labo a o y weighing 200−250 g. They we e housed in a ou a s pe cage (50 ×35 ×20 cm) dis ibu ion a a empe a u e o 22 ±1°C wi h a 12:12 h ligh − da k cycle wi h ligh s offa 08:00 h. They also had en i onmen al en ichmen consis ing o PVC pipe ubes and wooden blocks and ood and wa e p o ided ad libi um. Be o e SIP, he animals’bodyweigh s we e g adually educed o 85% o hei ee- eeding baseline le el h ough con olled eeding and daily weighing and hen main ained h oughou he expe imen . Food, consis ing o lab chow, was p o ided daily ∼30 min a e each expe imen al session. All o he es ing occu ed be ween 9:00 am and 2:00 pm. All p ocedu es we e in acco dance wi h he Spanish Royal Dec ee 53/2013 on he p o ec ion o expe imen al animals and he Eu opean Di ec i e 2010/63/EU and app o ed by he Animal Resea ch Commi ee o he Uni e si y o Alme ia. We decla e ha he esea ch shows commi men o he 3Rs p inciple ( eplacemen , educ ion, efinemen ). 2.2. Expe imen al Design Once all animals eached 85% bodyweigh compa ed wi h hei baseline, blood samples (p e-SIP) we e aken (see as ollows) on he day be o e he s a o he SIP p ocedu e. Then, a e SIP, pos -SIP samples we e collec ed. A e 1 mon h o washou (based on p e ious publica ions; see Mo a e al.), 15 hal o he animals we e selec ed o a con ol es conce ning he die a y impac on hei me abolism. Blood samples we e collec ed be o e (p e-pelle ) and a e (pos -pelle ) exposu e o he es pelle s wi h wa e a ailable ad libi um.Figu e 1 shows he expe imen al imeline o ele an e en s. 2.3. Blood Sampling Animals we e anes he ized using isoflu ane, and blood samples (1 mL) we e collec ed om he la e al ail ein in 1.5 mL au ocued plas ic ubes be ween 9:00 am and 12:00 pm (da k cycle). The samples we e allowed o s and o 10 min be o e cen i uging (Sigma 3-18KS, Ge many) a 3000 pm (800g) o 10 min a 23 °C, a e which se um was collec ed in o duplica e 0.5 mL au ocued plas ic ubes and s o ed a −80 °C un il assay. 2.4. SIP P ocedu e Ra s we e es ed in 12 ope an SIP chambe s (32 ×25 ×34 cm) (MED Associa es, S . Albans, VT). A p e ious desc ip ion o he appa a us can be ound in Mo eno e al. 16 P og amming and da a eco ding we e pe o med wi h he aid o a compu e and comme cial so wa e Med PC (Cibe ec SA, Spain). P io o SIP, wo baseline wa e inges ion es s on successi e days we e pe o med, whe e he amoun (in mL) o wa e consumed by each animal du ing a pe iod o ime o 60 min wi h ee access o 60 pelle s (Noyes 45 mg dus less ewa d pelle s; TSE Sys ems, Ge many) was measu ed. A e 1 day o habi ua ion o he chambe s session, a s unde wen 60 min daily sessions o a fixed- ime 60 s (FT-60s) schedule o ood pelle deli e y, whe e bo les con aining eshwa e ad libi um we e placed in he wall opposi e o he pelle dispense . Measu es eco ded we e: (1) o al amoun o wa e (in mL) consumed, (2) o al numbe o licks o he bo le, and (3) o al numbe o ood magazine en ies. A e 19 daily sessions o SIP acquisi ion, animals we e selec ed in wo g oups, high and low d inke s (HDs and LDs, n= Figu e 1. Expe imen al p ocedu e illus a ed in a imeline. A e habi ua ion o he lab, animals we e di ided in o high d inke s (HDs) and low d inke s (LDs) by schedule-induced polydipsia (SIP)be o e (p e-SIP) and a e (pos -SIP) blood samples we e aken. A e SIP, animals emainedundis u bed o 1 mon h. Then, an addi ional expe imen ega ding exposu e o ewa d pelle s ook place o assess i s die a y impac on me abolomic analyses; again, blood samples we e collec ed be o e (p e-pelle ) and a e (pos -pelle ) exposu e. Jou nal o P o eome Resea ch pubs.acs.o g/jp A icle h ps://doi.o g/10.1021/acs.jp o eome.1c00857 J. P o eome Res. 2022, 21, 612−622 613 20 in each g oup), acco ding o i hei d inking a es du ing SIP (a e age o wa e in ake on he las fi e sessions) we e abo e o below he g oup median, espec i ely. 2.5. Exposu e o Die Pelle s Hal o he animals o each g oup (HD and LD, n= 10 in each g oup) we e exposed o 19 consecu i e exposu e sessions simila o SIP o u he assess any die a y impac o he ewa d pelle s. The animals ecei ed he same numbe o pelle s as du ing he SIP p ocedu e (60 pelle s), bu in his case, he pelle s we e p esen ed unde mass eeding condi ions wi hou a ood- ein o cemen ime schedule; wa e was a ailable ad libi um.As in SIP, he consump ion o all pelle s was assu ed by he expe imen e s a e each session. 2.6. Sample P epa a ion o NMR Fo NMR expe imen s, 150 μL o a blood se um was mixed wi h 350 μLo D 2O con aining 0.9% NaCl and he sodium sal o 3-( ime hylsilyl)p opionic-2,2,3,3-d4acid (TSP) a 0.01% (w/w). The esul ing mix u e was o exed and cen i uged o 5 min a 13 500 pm, and 500 μL o supe na an s was ans e ed in o o en-d ied 5 mm NMR ubes. 2.7. NMR Expe imen s Acquisi ion o 1H NMR spec a o se um samples was conduc ed as desc ibed by Pe ez-Fe nandez e al. 11 wi h some modifica- ions. Measu emen s we e ca ied ou on a B uke A ance III 600 spec ome e ope a ing a 600.13 MHz, equipped wi h a 5 mm QCI quad uple esonance pulse field g adien c yop obe and a SampleJe au osample , a 293 ±0.1 K and wi hou o a ion. The wa e -supp essed Ca −Pu cell−Meibom−Gill (CPMG) pulse sequence was applied wi h a o al spin echo delay o 100 ms (τ−180°−τ, 400 μs−37 μs−400 μs) o a enua e b oad signals om p o ein signals. The spec ome e ansmi e was locked o D2O equency. Acquisi ion pa ame e s we e se as ollows: NS = 60, DS = 16, size o fid= 32K, spec al wid h = 22.0 ppm, acquisi ion ime = 1.24 s, elaxa ion delay = 3 s, numbe o loops = 120, line b oadening = 0.3 Hz, ecei e gain = 203. Spec a we e au oma ically phased, baseline-co ec ed, and calib a ed o TSP signal a 0.0 ppm. Acquisi ion and p ocessing o NMR spec a we e ca ied ou by he TOPSPIN so wa e ( e sion 3.6.2). Me aboli e assignmen s we e pe o med hanks o in o ma ion on scala couplings ex ac ed om 1H−1HCOSY,1H−1HTOCSY,1H−13C HSQC, and 1H−13C HMBC spec a, which we e eco ded using s anda d B uke sequences, and wi h he help o he Chenomx da abase (Chenomx, Edmon on, Canada), public NMR da abases (HMDB), and li e a u e. 17−19 Quan ifica ion o me aboli es was achie ed h ough he in eg a ed alues o he ela ed peak a eas o nono e lapped signals in ela ion o he inne s anda d (TSP). 2.8. Quan ifica ion o Fa y Acids A e NMR acquisi ion, se um samples we e eeze-d ied o 72 h. Then, he a y acid con en and p ofile in se um samples and also in he pelle s we e de e mined by gas ch oma og aphy (Agilen Technologies 6890 N Se ies Gas Ch oma og aph, San a Cla a, CA) a e di ec anses e ifica ion, as desc ibed by Rod iguez-Ruiz e al. 20 2.9. S a is ical Da a Analysis SIP acquisi ion da a we e analyzed using wo-way epea ed- measu es analysis o a iance (ANOVA) wi h be ween-subjec ac o (g oup: HD and LD) and wi hin-subjec ac o (session: 19 sessions). Pos hoc compa isons we e pe o med using he Bon e oni co ec ion. The s a is ical significance was se a p< 0.05, and he effec size was epo ed when app op ia e: Pa ial η2 alues a e epo ed and conside ed as small (0.01), medium (0.06), o la ge (0.14) ollowing Cohen 21 ecommenda ions. All analyses we e ca ied ou using S a is ica so wa e (S a so , e sion 6.0). Wi h espec o chemome ics analyses o 1H NMR spec al da a, AMIX 3.9.15 (B uke BioSpin) so wa e was used o bucke ing NMR spec a using wo ypes o bucke ing p ocesses: (1) egula bucke ing employing a bucke size o 0.04 ppm and (2) a iable bucke ing o NMR peaks assigned o specific me aboli es ( o uni a ia e s a is ical analyses). In bo h cases, no maliza ion was achie ed by scaling he in ensi y o indi idual peaks o he o al in ensi y eco ded in he egion om δH0.2 o 10.0 ppm, excep o he egion o δH5.2 o 4.74 ppm con aining he esidual signal o H2O, which was emo ed. NMR egula bucke ed da a was in es iga ed by means o p incipal componen analysis (PCA) in explo a o y s udies and by pa ial leas -squa es disc iminan analysis (PLS-DA) o o hogonal pa ial leas -squa es disc iminan analysis (OPLS-DA) o de e mine he exis ence o diffe ences be ween expe imen al g oups and o iden i y he me abolic ea u es esponsible o hei disc imina ion. This mul i a ia e da a analysis was pe o med using SIMCA-P so wa e ( . 17.0, Ume ics). The esul s o he c oss- alida ion o PLS-DA and OPLS-DA models a e gi en by means o cumula i e R2and Q2 alues, whe e Q2 alues o >0.5 we e conside ed indica i e o a good p edic i e model. Models we e also alida ed by being subsequen ly subjec ed o pe mu a ion es s (a o al o 100), and he new coefficien s R2and Q2gene a ed om he pe mu a ion es we e Figu e 2. Means (±SEM) o (A) wa e in ake and (B) numbe o licks in FT-60s ac oss 19 sessions o SIP. S a is ical analyses indica ed significan diffe ences be ween low d inke s (LDs, n= 20) and high d inke s (HDs, n= 20; ** p< 0.01) in bo h wa e in ake and o al licks. Significan diffe ences om session 1 we e ound in HDs om session 6 in wa e in ake (## p< 0.01) and om session 8 (### p< 0.001) onwa d. Jou nal o P o eome Resea ch pubs.acs.o g/jp A icle h ps://doi.o g/10.1021/acs.jp o eome.1c00857 J. P o eome Res. 2022, 21, 612−622 614 compa ed wi h hose om he eal model. I in e cep R2and Q2 alues om he pe mu a ion es we e significan ly smalle han Q2o he eal model, hen he model was ega ded as p edic able. The s a is ical significance o he es ima ed p edic i e powe o PLS-DA and OPLS-DA models was u he assessed wi h an ANOVA es o he c oss- alida ed esiduals (c -ANOVA). Models wi h p alues o <0.05 we e conside ed o ha e a good p edic ion. Impo an loadings (spec al egions) o he disc imina ion obse ed om he p edic i e models we e selec ed by gene a ing he a iable impo ance in p ojec ion (VIP) plo . Loadings wi h VIP sco es o >1 we e conside ed ele an o he gene a ed PLS- DA and OPLS-DA models. ANOVA (analysis o a iance) analyses ollowed by leas significan diffe ence (LSD) pos hoc es s we e employed o de e mine he significance o diffe ences o he me aboli e a ios be ween g oups; p alues o <0.05 we e conside ed s a is ically significan . Finally, me abolic changes wi h alse disco e y a e (FDR)-adjus ed p alues (q alues) o <0.05 we e conside ed. 3. RESULTS AND DISCUSSION 3.1. SIP Acquisi ion LD and HD beha io is clea ly e idenced by no only he wa e in ake bu also he numbe o licks, as p e iously men ioned. Figu e 2 shows SIP acquisi ion and main enance du ing 19 sessions. One-way epea ed ANOVA measu es e ealed diffe - ences in he SIP acquisi ion conce ning wa e in ake, as shown by he in e ac ion be ween sessions and g oup (in e ac ion SIP session ×g oup effec : (F18, 666 = 9.146, p< 0.001, pa ial η2= 0.198). This effec was also confi med by he significan in e ac ion in o al licks (in e ac ion SIP session ×g oup effec : F18, 576 = 8.445, p< 0.001, pa ial η2= 0.208). Apos hoc compa ison e ealed ha SIP induced diffe en d inking beha io s ac oss he 19 sessions in high and low d inke s: LD and HD animals exhibi ed ema kable diffe ences in wa e in ake om session 6 (p< 0.01) onwa d. Mo eo e , when compa ed wi h session 1, he HD g oup significan ly inc eased i s wa e consump ion om session 4 (p< 0.01) onwa d. A simila pa e n was ound conce ning o al licks, whe e LD and HD g oups diffe ed om session 8 (p< 0.01) onwa d, and HD showed an inc eased numbe o licks om session 4 (p< 0.001) onwa d compa ed wi h session 1. No significan diffe ences we e ound be ween LD and HD animals conce ning o al magazine en ies. 3.2. Assignmen o Me aboli es in Blood Se um De ec ed by 1H NMR Spec oscopy The assignmen o each me aboli e p esen in se um samples was achie ed and is illus a ed in Figu e 3.Table S1 p o ides ull in o ma ion on chemical shi s, mul iplici y, and coupling cons an s o each me aboli e o o each me aboli e- ype compound. The assigned me aboli es belong mos ly o he classes o amino acids ( aline, isoleucine, leucine, alanine, lysine, glu ama e, glu amine, aspa a e, y osine, and phenylalanine) and de i a i es (e.g., 3-hyd oxybu y a e, c ea ine), o ganic acids (ace ic acid, ace oace ic acid, ci ic acid, and py u ic acid), ca bohyd a es (α-andβ-glucose), choline (Cho)-based compounds, which a e essen ial componen s o cellula memb anes, polyols (e.g., glyce ol), and a y acids. T ime hyl- amine N-oxide is an osmoly e used by he body o coun e ac he effec s o inc eased u ea concen a ion ha accumula es du ing kidney ailu e and was also ound in he spec a. Ke one bodies like 3-hyd oxybu y a e, ace a e, and ace oace a e, gene ally induced by as ing, 17 could also be ound in a se um samples. Figu e 3. Sub egions o a ypical 1H NMR spec um (600 MHz) o a blood se um sample aken o a a om p e-SIP g oup. Nume a ion: 1: LDL/ VLDL, 2: a y acids (FA), 3: leucine, 4: isoleucine, 5: aline, 6: 3-hyd oxybu y a e (3HB), 7: E OH, 8: lac a e, 9: alanine, 10: lysine, 11: ace a e, 12: glu ama e, 13: glu amine, 14: ace oace a e, 15: py u a e, 16: ci a e, 17: PUFA, 18: aspa a e, 19: c ea ine, 20: choline, 21: ime hylamine N-oxide (TMAO), 22: glyce ol, 23: glucose, 24: UFA, 25: py imidine, 26: uma a e, 27: y osine, 28: his idine, 29: phenylalanine, 30: o ma e. Jou nal o P o eome Resea ch pubs.acs.o g/jp A icle h ps://doi.o g/10.1021/acs.jp o eome.1c00857 J. P o eome Res. 2022, 21, 612−622 615 Figu e 4. (A) OPLS-DA sco e plo ob ained o 1H NMR da a o samples aken p io o (n= 40) and a e SIP (n= 38). (B) S-plo showing he mos significan me aboli es o disc imina ion (me aboli es wi h VIP > 1 alues a e colo ed): Bo om le me aboli es we e significan ly dec eased (alanine: loadings 1.46, 1.50; lac a e: loadings 1.34, 4.14; UFA: loadings 5.34, 2.06, 2.18, 2.26, 0.94), whe eas hose loca ed a he op igh we e inc eased pos - SIP (choline: loading 3.22; ci a e: loadings 2.54, 2.70; ace a e: loading 1.94; ace oace a e: loading 2.22; glucose: loadings 3.34, 3.38, 3.42, 3.46, 3.54, 3.66, 3.78, 3.82, 5.26; e hanol: loadings 1.18, 3.66; and LDL/VLDL/sa u a ed a y acids: loadings 0.86, 1.26, 1.30, 2.14). Pa e o scaling was ca ied ou . R2X= 0.639, Q2= 0.764, p(c -ANOVA) = 7.0 ×10−18. The model was alida ed by a pe mu a ion es (wi h 100 pe mu a ions). Figu e 5. OPLS-DA (A) sco e and (B) con ibu ion plo s ob ained o 1H NMR da a o pos -SIP samples ha we e classified as LD (n= 20) and HD (n= 18). Posi i e and nega i e ba s in he con ibu ion plo explain he spec al egions con aining me aboli es ha disc imina e o HD and LD samples, espec i ely. These include glu amine (loadings 2.14, 2.46), a y acids/LDL/VLDL (loadings 0.86, 0.90, 0.94, 1.26, 1.30, 1.34, 2.02, 2.06, 2.26, 2.30), 3-hyd oxybu y a e (HB, loadings 2.32, 2.40, 4.16), choline (loading 3.22), and lac a e (1.34, 4.10), which inc eased in HD d inke s, and e hanol (loadings 1.18, 3.66), alanine (loading 1.50), glyce ol (loadings 3.62, 3.66), and glucose (loadings 5.26, 4.66, 3.90, 3.86, 3.74, 3.46, 3.26), which dec eased. Only he samples om e y low d inke s (<10 mL) and e y high d inke s (>20 mL) we e conside ed. Pa e o scaling was ca ied ou . R2X= 0.68, Q2= 0.58, p(c -ANOVA) = 0.032. Jou nal o P o eome Resea ch pubs.acs.o g/jp A icle h ps://doi.o g/10.1021/acs.jp o eome.1c00857 J. P o eome Res. 2022, 21, 612−622 616 3.3. Chemome ics Analyses o 1H NMR Spec al Da a Mul ipa ame ic s a is ical ools we e applied o NMR da a in he analyses o a s se um samples o e alua e changes in he me abolic p ofiles in he p esence o OCD. Blood samples we e aken a ou diffe en imes du ing he expe imen : p e- and pos -SIP and p e- and pos -pelle (as indica ed in Figu e 1). An explo a o y analysis o NMR da a was fi s achie ed by means o PCA ha clus e ed simila samples oge he based on he inpu da a (Figu e S1). PCA is use ul o e ealing he majo ends in he 1H NMR da a and he possible analy ical and biological con ounde a iables. On he basis o he PCA esul s, a sligh di ision o samples be ween p e- and pos -SIP g oups (Figu e S1a) and be ween p e- and pos -pelle g oups (Figu e S1b)is e iden , independen o he obsessi e-compulsi e beha io o he a s. 3.3.1. Diffe en ia ion o P e-SIP and Pos -SIP Se um Samples. To imp o e he disc imina ion obse ed be ween p e-SIP and pos -SIP da a, we applied an OPLS-DA model o he 1H NMR da a (Figu e 4A). The chosen p ep ocessing me hod con ains he o hogonal signal co ec ion (OSC), which Figu e 6. Box plo s showing he a e age, median qua iles, and ex emes o specific peak in eg a ion alues ( ela i e o he TSP peak in eg al and no malized o o al spec a in ensi y) o 3-hyd oxybu y a e, glu amine, lac a e, choline, LDL/VLDL, and FA le els, excep o UFA, which inc eased pos -SIP o HD a s, and glyce ol, alanine, and glucose, which dec eased pos -SIP o HD a s when compa ed wi h LD a s. All o hese me aboli es p esen ed p alues o <0.05 using ANOVA analyses ollowed by he leas significan diffe ence (LSD) pos hoc es s in pos -SIP. Jou nal o P o eome Resea ch pubs.acs.o g/jp A icle h ps://doi.o g/10.1021/acs.jp o eome.1c00857 J. P o eome Res. 2022, 21, 612−622 617 allows one o elimina e unnecessa y in o ma ion. In he OSC p ocedu e, he X ma ix was co ec ed by a sub ac ion o a ia ion o hogonal o he y(con aining he classes o each sample, in his case, p e-SIP e sus pos -SIP) ec o calib a ion. The co esponding S-plo (Figu e 4B) e ealed he mos ele an me aboli es o he disc imina ion (wi h VIP alues >1). Me aboli es o he bo om le we e significan ly dec eased (alanine, lac a e, unsa u a ed a y acids), whe eas hose loca ed a he op igh we e inc eased pos -SIP (choline, ci a e, ace a e, ace oace a e, e hanol, glucose, and LDL/VLDL/sa u a ed a y acids). The esul s indica e ha SIP induced changes in lipid- me abolism- ela ed molecules ( a y acids and choles e ol le els, choline, ace a e, ace oace a e, glu amine) and ene gy me abo- lism ( ela ed o he ica boxylic acid (TCA) cycle). Ace a e is o med in he body by he me abolism o ce ain subs ances, in pa icula , in he li e in he oxida ion o lipids. I is a p ecu so o ace yl-CoA, which is used by cells o he syn hesis o a y acids and choles e ol. To u he in es iga e i a diffe en ial effec is induced by he pa adigm in HD and LD g oups, we added his a iable o he model in he ollowing analysis (de ailed in Sec ion 3.3.2). 3.3.2. Diffe en ia ion be ween LD and HD G oups in P e- and Pos -SIP Se um Samples. A alid disc imina ion be ween LD and HD d inke s was ound only o pos -SIP samples a e he applica ion o an OPLS-DA model (Figu e 5). This means ha i was no possible o obse e me abolic changes a p io i o he SIP p ocedu e, so i would no be possible o “p edic ” he beha io o each a based on se um me abolic p ofiles. To ob ain a alid model, we emo ed se um samples o a s showing a d inking olume be ween 10 and 20 mL om he model. So, his model ied o maximize diffe ences be ween e y low d inke s (<10 mL) and e y high d inke s (>20 mL). The blood se um o pos -SIP HD a s specifically showed inc eased low-densi y lipop o ein (LDL)/ e y low-densi y lipop o ein (VLDL), a y acid (excep unsa u a ed a y acids), 3-hyd oxybu y a e, glu amine, choline, and lac a e le els, accompanied by a dec ease in glyce ol, glucose, alanine, and e hanol le els when compa ed wi h pos -SIP LD a s. Figu e 6 shows he box-and-whiske plo s o no malized in eg a ion alues o peaks om hese al e ed me aboli es ( ela i e o TSP signal in eg al) wi h an indica ion o he in eg a ion anges, median qua iles, and ex emes. These esul s sugges ha lipid-me abolism- ela ed molecules (including o al choles e ol (TC) con en , glu amine, choline, and glyce ol) migh be associa ed wi h a compulsi e beha io pheno ype du ing SIP. To u he suppo he obse ed ends, we conduc ed he iden ifica ion and quan ifica ion o a y acid p ofiles in se um samples by gas ch oma og aphy wi h flame ioniza ion de ec ion (GC-FID) (Figu e 7). The polyunsa u a ed a achidonic acid (C20:4n6) was ound o be he majo a y acid in he samples ollowed by he sa u a ed palmi ic acid (C16:0) and by linoleic acid (C18:2n6) and oleic acid (C18:1n9). HD a s showed an inc ease in he sa u a ed palmi ic acid (C16:0), he monounsa u a ed oleic acid (C18:1n9), and he polyunsa- u a ed linoleic acid (C18:2n6) compa ed wi h LD a s (p< 0.05) and a dec ease in a achidonic acid, he majo polyunsa u a ed a y acid (PUFA) in se um samples. As shown in Figu e 7, s a is ically significan p e-SIP diffe ences be ween LDs and HDs exis and emain cons an pos -SIP. These esul s ag ee well wi h NMR esul s ha de ec ed an inc ease in a y acid/LDL/VLDL con en in HD a s, excep o UFA. In e es ingly, GC-FID analysis showed ha he o al PUFA con en was significan ly highe in LD a s compa ed wi h HD a s al eady in p e-SIP samples (p< 0.05). These esul s a e in iguing because some p eclinical s udies ha e also ound a simila ou pu in compulsi e animals: Inc eased TC, VLDL, and LDL le els we e ound in dogs exhibi ing ail chasing (a s e eo yped beha io p oposed as a alid p eclinical model o s udying OCD in animals) 22 in compa ison wi h con ol subjec s, 23,24 whe eas dec eased glucose se um le els we e ound in highly compulsi e a s selec ed by SIP compa ed wi h noncompulsi e animals. 25 In human s udies, OCD pa ien s ha e been shown o exhibi highe high-densi y lipop o ein (HDL) se um le els han heal hy con ols. 26 Mo eo e , B ennan e al. 27 pe o med a c i ical e iew o 1H magne ic esonance imaging s udies on OCD in b ain issues and concluded ha changes in glu ama e/ glu amine le els we e common among pa ien s e sus heal hy indi iduals, al hough some inconsis encies ha e been un a eled. Fo ins ance, some s udies epo ed a dec ease whe eas o he s epo ed an inc ease in he glu ama e-glu amine (Glx) le el. Also, choline was e e ed o be a c ucial bioma ke in OCD, and mos s udies showed an inc ease in his me aboli e in OCD indi iduals. Simila ly, o he neu opsychia ic condi ions ela ed o inhibi o y con ol ha e been s udied ega ding me abolic changes. Schwa z e al. 28 epo ed significan al e a ions o b ain issue ee a y acids and phospha idylcholine le els in subjec s wi h schizoph enia and bipola diso de using a high- h oughpu mass spec ome y app oach (UPLC−MS) and sugges ed ha lipid abno mali ies may be an in insic ea u e o bo h schizoph enia and bipola diso de . A maca e al. 29 obse ed dec eased se um choles e ol and lep in le els in bipola diso de pa ien s, whe eas Ozbulu e al. 30 ound dec eased se um gh elin and inc eased TC le els in eu hymic pa ien s unde li hium ea men when compa ed wi h con ols. In dep ession, Kaddu ah-Daouk and K ishnan 31 epo ed ha a y acids, glyce ol, and γ-aminobu y ic acid (GABA) we e al e ed in cu en ly dep essed pa ien s when compa ed wi h con ols. Also, an inc ease in he concen a ion o he ke one 3- hyd oxybu y ic acid was ound in emi ed pa ien s ela i e o dep essed pa ien s. Fu he mo e, Nakaza o e al. 32 ound a ela ionship be ween inc eased glu amine se um le els and compulsi e-like beha io , as assessed by o al and pe se e a i e e o s in se -shi ing asks in subjec s eco e ed om ano exia ne osa, which also is adop ed in he impulsi e-compulsi e spec um. Figu e 7. Fa y acid quan ifica ion by GC-FID. Ba s wi h (*) and (**) e ealed a significan inc ease and dec ease (p< 0.05 by uni a ia e es s), espec i ely, in he lipid con en in he se um o HD a s compa ed wi h LD a s. Jou nal o P o eome Resea ch pubs.acs.o g/jp A icle h ps://doi.o g/10.1021/acs.jp o eome.1c00857 J. P o eome Res. 2022, 21, 612−622 618 I could be a gued, howe e , ha hese me abolic effec s migh ha e been induced by die due o he me e exposu e o ewa d pelle s du ing SIP sessions. Indeed, obese indi iduals ha e been epo ed o exhibi poo e beha io al inhibi ion 18 and a s onge a en ional bias owa d ood 33,34 compa ed wi h heal hy subjec s. Mo eo e , impulsi e beha io seems o be co ela ed wi h a highe consump ion o as ood 35 and a highe body mass index (BMI), 36 whe eas in p eclinical models, exposu e o highly pala able, a - ich die s has been shown o induce compulsi e ood-seeking beha io 37−39 and o affec ma ble bu ying beha io . 40−42 Thus o explo e ha possibili y, we planned and ca ied ou an addi ional expe imen whe e a coho o animals we e subjec ed o 19 sessions o exposu e o ewa d pelle s, as desc ibed in he Me hods sec ion. 3.3.3. Diffe en ia ion be ween P e-Pelle and Pos - Pelle Se um Samples. A PLS-DA model was gene a ed o in es iga e he me abolic diffe ences be ween se um samples collec ed be o e and a e he con ol die expe imen . A alid disc imina ion be ween p e- and pos -pelle g oups was ound in he sco e plo (Figu e 8A), meaning ha he ewa d pelle s had effec s on he se um me abolic composi ion. The loading plo (Figu e 8B) highligh s he mos significan a iables o he model by desc ibing he influence and ela ion among he a iables in he model plane. The e o e, i is possible o conclude ha he peaks in spec al zones should be significan ly inc eased (uppe igh loadings) and dec eased (bo om le loadings) in he blood se um due o he exposu e o ewa d pelle s in his con ol expe imen (VIP alues >1) and also due o a physiological me abolic e olu ion o e ime. An inc ease in c ea ine, ci a e, glucose, lysine, glu amine/glu ama e, ace a e, and PUFAs was de ec ed in he blood se um a e such exposu e. The spec al zones in he bo om le o he loading plo e e ing o he me aboli es ha dec ease in pos -pelle se um include choline and o he spec al egions ha mos ly co espond o noise and we e no ele an o analyses, so hese we e no conside ed. As in he p e ious se o expe imen s, he lipid p ofile o he pelle s was in es iga ed and quan ified by GC-FID (Table S2). The ewa d pelle s du ing he SIP expe imen showed a 2.3- old inc ease in a y acid con en , including on PUFAs. Because he model was no able o disc imina e be ween HDs and LDs ei he p e- o pos -exposu e o ewa d pelle s, i is plausible o assume ha such al e a ions affec ed bo h g oups equally, hus poin ing o he idea ha he SIP pa adigm, and no he die , is esponsible o he me abolic al e a ions obse ed in he ulne able subjec s, he HD g oup. The e a e se e al unde lying mechanisms ha should be in es iga ed in he u u e ega ding he me abolic al e a ions obse ed in he compulsi e-d inking HD g oup selec ed by SIP, o example, he possible ole o an al e ed asop essin, a ho mone implica ed in he me abolic synd ome, ha d i es a p oduc ion as a mechanism o s o ing me abolic wa e . 43 Fu he mo e, he ele ance o gu mic obio a dys egula ion in diffe en neu opsychopa hological diso - de s 44−46 should be conside ed in he me abolic al e a ions obse ed in compulsi e HD a s. In his sense, in ou labo a o y, we ha e demons a ed ha compulsi e HD a s showed a lowe bac e ial di e si y han LD a s, i espec i e o he die . 47 Howe e , he same s udy also demons a ed ha he adminis a ion o a yp ophan-deple ed die educed bac e ial e enness and showed a highly unc ionally o ganized communi y in he compulsi e HD a s selec ed by SIP. This poin s owa d a bac e ial communi y ha is agile o ex e nal changes due o he dominance o a low numbe o species in compulsi e HD a s compa ed wi h noncompulsi e LD a s. The d ama ic effec ound in he p esen wo k adds e idence no only o he s ess ul p ope ies o in e mi en ein o cemen in SIP inducing compulsi e beha io bu also o se e e physiological effec s, such as inc eased co icos e one le els, 25 inc eased amygdaloid and dec eased hippocampal olume, 15 and inc eased dend i ic spinal densi y in he do sal s ia um. 48 In his sense, he a o emen ioned finding o inc eased glu amine le els in HD in he p esen s udy is no su p ising gi en i s majo ole in he b ain as he p ecu so o glu ama e, which is a key ac o ega ding neu oplas ici y 49,50 and one o he al e ed mechanisms 51 and pu a i e he apeu ic a ge s 52 in compulsi e d inking in SIP. 4. CONCLUSIONS The p esen s udy has in es iga ed he me abolomic p ofile by means o NMR o high- and low-compulsi e a s selec ed by SIP Figu e 8. (A) PLS-DA sco e plo ob ained o 1H NMR da a o samples be o e (n= 20) and a e (n= 20) changing he die o ewa d pelle s (p e- and pos -pelle , espec i ely). (B) PLS-DA loading plo e ealing he mos significan me aboli es o disc imina ion (wi h VIP alues >1), which co espond o c ea ine (loadings 3.06, 3.94), glucose (loadings 5.26, 3.94, 3.78, 3.74, 3.56, 3.50, 3.26), PUFA (loading 2.78), lysine (loadings 1.50, 1.74), ace a e (loading 1.94), o ma e (loading 8.46), and glu amine/glu ama e (loadings 2.14, 2.46, 2.36), which inc eased in pos -pelle samples, and choline (loadings 3.22), which dec eased. Uni - a iance scaling was ca ied ou . R2X= 0.65, Q2= 0.90, p(c -ANOVA) = 6.5 ×10−9. The model was alida ed by a pe mu a ion es (wi h 100 pe mu a ions). Jou nal o P o eome Resea ch pubs.acs.o g/jp A icle h ps://doi.o g/10.1021/acs.jp o eome.1c00857 J. P o eome Res. 2022, 21, 612−622 619 as a po en ial ool o iden i ying c i ical bioma ke s in ulne able subjec s. We ound ha al hough SIP i sel induced a change in he me abolomic p ofile, i affec ed he HD animals diffe en ly, which showed a hype lipidemic, hypoglycemic, and hypoglu amine gic p ofile, compa ed wi h he LD animals, in line wi h he li e a u e ega ding bo h p eclinical models and human pa ien s in OCD and ela ed diso de s. Mo eo e , me e exposu e o ewa d pelle s did no esul in a alid model o p edic ing he pheno ypic p ofile based on me abolomics, hus leading us o disca d an effec due o die alone. Ou da a add significan e idence o a c ucial opic in basic neu oscience and po en ially clinical fields; howe e , u u e s udies a e ongoing in ou labo a o ies o unsc amble his complex phenomenon and o u he cha ac e ize he pu a i e ole o me abolism in he igge ing and ea ly iden ifica ion o la en ulne abili ies. In addi ion, he pu a i e impac o a high- a die on inhibi o y con ol defici and he po en ial use o me abolomic bioma ke s in he ea ly diagnose o compulsi e spec um diso de s a e en isaged as well. ■ASSOCIATED CONTENT * sıSuppo ing In o ma ion The Suppo ing In o ma ion is a ailable ee o cha ge a h ps://pubs.acs.o g/doi/10.1021/acs.jp o eome.1c00857. Table S1. Peak assignmen o me aboli es iden ified. Table S2. GC-FID a y acid p ofiles. Figu e S1. PCA sco e plo s ob ained om 1H NMR da a (PDF) ■AUTHOR INFORMATION Co esponding Au ho s Ma ga i a Mo eno −Depa men o Psychology and Heal h Resea ch Cen e CEINSA, Uni e si y o Alme ía, 04120 Alme ía, Spain; Email: [email p o ec ed] Ignacio Fe nández −Depa men o Chemis y and Physics, Resea ch Cen e CIAIMBITAL, Uni e si y o Alme ía, 04120 Alme ía, Spain; o cid.o g/0000-0001-8355-580X; Phone: +34 950214465; Email: [email p o ec ed] Au ho s Ana C. Ab eu −Depa men o Chemis y and Physics, Resea ch Cen e CIAIMBITAL, Uni e si y o Alme ía, 04120 Alme ía, Spain San iago Mo a −Depa men o Psychology and Heal h Resea ch Cen e CEINSA, Uni e si y o Alme ía, 04120 Alme ía, Spain; P esen Add ess: San iago Mo a - Depa men o Neu oscience and Facul y o Heal h and Medical Sciences, Uni e si y o Copenhagen, 2200 Copenhagen, Denma k Ana Isabel T is án −Depa men o Chemis y and Physics, Resea ch Cen e CIAIMBITAL, Uni e si y o Alme ía, 04120 Alme ía, Spain Elena Ma ín-González −Depa men o Psychology and Heal h Resea ch Cen e CEINSA, Uni e si y o Alme ía, 04120 Alme ía, Spain A  ngeles P ados-Pa do −Depa men o Psychology and Heal h Resea ch Cen e CEINSA, Uni e si y o Alme ía, 04120 Alme ía, Spain Comple e con ac in o ma ion is a ailable a : h ps://pubs.acs.o g/10.1021/acs.jp o eome.1c00857 Au ho Con ibu ions § A.C.A. and S.M. con ibu ed equally. Funding This esea ch has been unded by he S a e Resea ch Agency o he Spanish Minis y o Science, Inno a ion and Uni e si ies (CTQ2017-84334-R), Jun a de Andalucia (102C2000004, UAL-FEDER 2020-AGR-B1781, UAL-FEDER 2020-CTS- D2068, P20_01041, and CV20-78799), and he Gobie no de Espana MCIN/AEI/10.13039/501100011033/Union Eu opea “Nex Gene a ionEU”/PRTR (PDC2021-121248-I00 and PLEC2021-007774) and MCIN/AEI/10.13039/ 501100011033/FEDER “Unamane adehace Eu opa” (PGC2018-099117-B-C21). A.C.A. hanks he Uni e si y o Alme ia o he Hipa ia schola ship. We hank P o . Ma ia del Ca men Ce on Ga cia (Uni e si y o Alme ia) o hei help ul assis ance ega ding GC-FID measu emen s. No es The au ho s decla e no compe ing financial in e es . ■REFERENCES (1) Robbins, T. W.; C ocke , M. J. Chap e 3.8 - Role o Cen al Se o onin in Impulsi i y and Compulsi i y: Compa a i e S udies in Expe imen al Animals and Humans. In Handbook o Beha io al Neu oscience;Mu lle , C. P., Jacobs, B. L., Eds.; Else ie , 2010; pp 415−427. (2) Ame ican Psychia ic Associa ion. Desk Re e ence o he Diagnos ic C i e ia om DSM-5; Ame ican Psychia ic Publishing: Washing on, DC, 2013. (3) Hollande , E.; Kim, S.; Khanna, S.; Pallan i, S. Obsessi e- Compulsi e Diso de and Obsessi e-Compulsi e Spec um Diso de s: Diagnos ic and Dimensional Issues. CNS Spec ums 2007,12,5−13. (4) Robbins, T. W.; Vaghi, M. M.; Banca, P. Obsessi e-Compulsi e Diso de : Puzzles and P ospec s. Neu on 2019,102,27−47. (5) WHO. Global Heal h Obse a o y (GHO) Da a; Wo ld Heal h O ganiza ion, 2018. (6) Song, Z.; Wang, H.; Yin, X.; Deng, P.; Jiang, W. Applica ion o NMR me abolomics o sea ch o human disease bioma ke s in blood. Clin. Chem. Lab. Med. 2019,57, 417−441. (7) Nagana Gowda, G. A.; Ra e y, D. Bioma ke Disco e y and T ansla ion in Me abolomics. Cu . Me abolomics 2013,1, 227−240. (8) Collino, S.; Ma in, F.-P. J.; Rezzi, S. Clinical me abolomics pa es he way owa ds u u e heal hca e s a egies. B . J. Clin. Pha macol. 2013,75, 619−629. (9) Mayeux, R. Bioma ke s: po en ial uses and limi a ions. Neu oRx 2004,1, 182−188. (10) Robbins, T. W., Sahakian, B. J. Cu en Topics in Beha io al Neu osciences. In T ansla ional Neu opsychopha macology; Sp inge In e na ional Publishing: Cham, Swi ze land, 2016; Vol. 28. (11) Pe ez-Fe nandez, C.; Mo ales-Na as, M.; Aguile a-Sáez, L. M.; Ab eu, A. C.; Gua dia-Esco e, L.; Fe nández, I.; Ga ido-Cá denas, J. A.; Colomina, M. T.; Giménez, E.; Sánchez-San ed, F. Medium and long- e m e ec s o low doses o Chlo py i os du ing he pos na al, p eweaning de elopmen al s age on sociabili y, dominance, gu mic obio a and plasma me aboli es. En i on. Res. 2020,184, 109341. (12) Mo eno, M.; Flo es, P. Schedule-induced polydipsia as a model o compulsi e beha io : neu opha macological and neu oendoc ine bases. Psychopha macology 2012,219, 647−659. (13) Falk, J. L. P oduc ion o polydipsia in no mal a s by an in e mi en ood schedule. Science 1961,133, 195−196. (14) Falk, J. L. The na u e and de e minan s o adjunc i e beha io . Physiol. Beha . 1971,6, 577−588. (15) Mo a, S.; Me chán, A.; Azna , S.; Flo es, P.; Mo eno, M. Inc eased amygdala and dec eased hippocampus olume a e schedule-induced polydipsia in high d inke compulsi e a s. Beha . B ain Res. 2020,390, 112592. Jou nal o P o eome Resea ch pubs.acs.o g/jp A icle h ps://doi.o g/10.1021/acs.jp o eome.1c00857 J. P o eome Res. 2022, 21, 612−622 620