© 2017 Nature America, Inc., part of Springer Nature. All rights reserved. 12-HETE 12-HETE production is a promising strategy for preventing and treating IR-induced liver damage. HETE–GPR31 axis that functionally determines hepatic IR procession. We have also provided proof of concept that blocking this model. Collectively, this study has revealed previously metabolic uncharacterized involving reprogramming an ALOX12–12- following liver resection. Most strikingly, blocking 12-HETE accumulation effectively attenuated all pathologies of hepatic IR in pigs. Furthermore, we established a nonhuman primate hepatic IR model that closely clinical recapitulates liver dysfunction damage. Notably, blocking 12-HETE production inhibits IR-induced liver dysfunction, inflammation and cell death in mice and accumulation and that 12-HETE then directly binds to GPR31, triggering an inflammatory response that exacerbates liver hepatic IR process. We found that ALOX12 was markedly upregulated in hepatocytes during ischemia to promote 12-HETE receptor 31 acid (12-HETE)–G-protein-coupled (GPR31) hydroxyeicosatetraenoic signaling axis as a key determinant of the targets. Here, using integrative ‘omics’ analysis, we identified an (ALOX12)–12- arachidonate 12-lipoxygenase (IR) Hepatic ischemia–reperfusion injury is a common clinical issue lacking effective therapy and validated pharmacological Pennsylvania, USA. Pennsylvania, Hepatic Surgery, Xijing Hospital, Fourth Military Medical University, Xi’an, China. Zhongnan Hospital of Wuhan University, Wuhan, China. of Gastroenterology, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China. of Organ Transplantation, Ministry of Health, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China. University, Wuhan, China. School, Wuhan University, Wuhan, China. 1 systemic unbiased an used we insight, this achieve To process required. is pathogenic this underlying mechanisms fundamental the clinic the in unsatisfactory are benefits achieved the date to although properties, antiapoptotic to led and has anti-inflammatory direct with agents concept develop to reper efforts This intense and response. damage cell inflammatory fusion-induced ischemia-induced of process biphasic a needed. urgently are damage liver IR-induced resec liver tion of volume low a but occlusion inflow of duration long a had who patients—those young of subset a only to limited are ing outcome improve to is strategy only the promising IR injury. preconditioning Ischemic for of the treatment hepatic interventions approved pharmacological dysfunction liver and rejection organ of risk increased for cause major a represents and failure graft early shock orrhagic liver hem as during occurrence an such as well as resection, and surgeries, transplantation liver of consequence unavoidable an is IR Zan Huang Jun-Yong Wang Yu Liu Xiao-Jing Zhang mediator of hepatic injury ischemia–reperfusion An ALOX12–12-HETE–GPR31 signaling axis is a key nature medicine nature Received 19 September; accepted 1 November; published online 11 December 2017; Department Department of Cardiology, Renmin Hospital of Wuhan University, Wuhan, China. The current pathogenic paradigm for hepatic IR injury involves involves injury IR hepatic for paradigm pathogenic current The 5 . Thus, effective pharmacological treatments for patients with with patients for treatments pharmacological effective Thus, . 2 , 5 , , Li-Jun Shen 5 1

, 2 12 . IR-related tissue injury accounts for up to 10% of 10% to up for accounts injury tissue IR-related .

4 , , Zhi-Gang She These These authors contributed equally to this work. should Correspondence be addressed to H.L. (

; however, the beneficial effects of precondition of effects beneficial the however, ; 2 1– advance online publication online advance , 4 6 , , Song Tian 4 Division Division of and Cardiothoracic Vascular Surgery, Key Laboratory of Organ Transplantation, Ministry of Education and Key Laboratory , 12 , , Xu Cheng 1 , 1 2 , 6– , , Peng Zhang 8 . Thus, a deeper understanding of of understanding deeper a Thus, . 4 Medical Medical Research Institute, School of Medicine, Wuhan University, Wuhan, China. 1 1 – , 3 2 4 . Unfortunately, there are no no are there Unfortunately, . , , Xue-Yong Zhu & Hongliang Li 4 , 12 , , Zhen-Zhen Yan 2– 9 Department Department of Radiology, Zhongnan Hospital of Wuhan University, Wuhan, China. 4 , 8 , Pi-Xiao , WangPi-Xiao

1 1 , 2 – , , Yan Zhang 4 , 8 5 , 12 - - - -

2 11 , , Jing Fang Institute Institute of Model Animals of Wuhan University, Wuhan, China. 1 Department Department of Emergency Medicine, Thomas Jefferson University, Philadelphia, – (rhesus macaque). Collectively, our data reveal that lipid metabolic metabolic lipid that reveal data our Collectively, macaque). (rhesus and pig mouse, including species, different across genetic or means pharmacological substantially ameliorated IR injury through manner. production Notably,dependent 12-HETE blocking damage via induction of a robust response inflammatory in a GPR31- liver IR-induced promoted markedly signaling ALOX12–12-HETE We that found unknown. is injury IR hepatic in ALOX12 of role the platelet regulate to aggregation, cell capacity migration and cancer its atherosclerosis, cell proliferation of in because cancer participation and its thrombosis on focused mainly hepatocytes ALOX12 including types, cell various in stage. ischemia the at liver in enriched greatly most were pathway metabolic (AA) acid arachidonic the in 12-HETE metabolite its and ALOX12 that found IR underwent surgery without vasculature occlusion that (sham group). In mice particular, of we that to compared as stage ischemic the during IR injury hepatic in profile the of livers that mice experienced so, a we pronounced identified of reprogramming the lipid metabolic investigative approach employing integrative omics analyses. In doing 4 d , , Rufang Liao ALOX12 is highly expressed in platelets and is extensively expressed o i : 1 1 0 , 2 . 1 , , Rui-Feng Tian 0 3 6 8 , 12 / n m , Xiaozhan , WangXiaozhan . 4 4 9 5 , , Yan-Xiao Ji 1 [email protected] 1 , 2 , Lin , WangLin 8 Medical Medical Science Research Center, 5 2 College College of Life Sciences, Wuhan 1 – , 4 2 , , , Weijun Wang 8 ). ,

10 10 , Xin-Liang Ma, Xin-Liang 9 . Previous studies of of studies Previous . Department Department of s e l c i t r a 3 Macaca mulatta Macaca Basic Basic Medical 7 10 Division Division , 1 7 1 , , Zhen- . . However,

11 , 

© 2017 Nature America, Inc., part of Springer Nature. All rights reserved. IR model and the human subjects who underwent liver resection resection liver underwent who subjects human the and model IR hepatic mouse the both in baseline at than phases reperfusion and at much ischemic the higher was also ALOX12 the level protein sion, compared to baseline ( greater ALOX12 significantly baseline found We at ischemia). to individual exposure same before the (i.e., from tissue in that with samples these in expression compared and resection liver to due IR hepatic ALOX12 maximal with injury, Alox12 IR hepatic of process entire the throughout 1f Fig. ( group sham the to comparison in ischemia during expression of two other ALOX family members (ALOX5 and ALOX15) mRNA in the or protein no alteration was such and there noblotting, immu and qPCR by confirmed was expression ALOX12 of increase ( set data teomics and mRNA level, as determined by analysis of differentially expressed gene, of version AA to 12-HETE the AA pathway, ( data DGE the with relation AA pathway also stood out in the proteomics data set with a high cor tistical significance ( changed most pathway the on the as basis of the metabolism number of AA changed genes and highlighted sta further stage ischemic the at stage. changed that substantially most features among the ming as occurs early as the stage, ischemic with lipid reprogram metabolic robust transcriptome reprogramming in response to hepatic IR injury and ( reperfusion after h 24 and 12 6, at particularly reperfusion, during observed were repair DNA and death cell response, immune of genes and transduction, pathways signal involved in inflammatory ( group sham the to comparison in changed dramatically as identified were metabolism lipid in involved genes which in profile, expression gene unique highly a elicited reperfusion without Notably,ischemia ( clusters six into categorized were another one to related by visualized was IR k during reprogramming transcriptome The ( manner time-dependent a in process IR samples from six subgroups were clearly during separated the hepatic t and clustering hierarchical unsupervised performed we data, DGE ( durations for different reperfusion from sue 60 after samples mice obtained bymin followed of ischemia tis liver on analyses proteomics and (DGE) expression gene digital alterations initiated by hepatic IR at the molecular level, rap we performed develops and process idly dynamic highly a is injury IR Hepatic Upregulation of ALOX12 in the ischemic stage RESULTS damage. liver IR-induced treat to target interventional therapeutic early an as axis this of translation clinical potential light hepatic IR progression, a previously unappreciateddetermines concept, and axis high ALOX12–12-HETE the involving reprogramming A  ( -distributed stochastic neighbor embedding analyses, and we found that Fig. 1 Fig. 1h Fig. -means clustering analysis, in which the expression trends most most trends expression the which in analysis, clustering -means

To explore the human relevance of these findings, we measured measured we findings, these of relevance human the explore To during genes expressed differentially most the of analysis Deeper s e l c i t r 12 Supplementary Fig. 1c Alox12 , 1 b 3 mRNA expression at 6 h after reperfusion ( reperfusion at 6 h after mRNA expression , and . To gain better insight into the full spectrum of pathological pathological of spectrum full the into To . insight better gain g – mRNA expression in the ischemia and reperfusion stages as stages mRNA and in expression reperfusion the ischemia mRNA expression in liver tissues from individuals with with individuals from tissues liver in expression mRNA ). The induction of ALOX12 expression was sustained sustained was expression ALOX12 of induction The ). k and and was one of the three most upregulated proteins in our pro Supplementary Fig. 1c Supplementary Fig. 1h Fig. Supplementary Alox12 Fig. 1 Fig. Fig. 1 Fig. 1 e , , encoding a lipoxygenase that targets the con and c 14 and g ). These multivariate analyses indicate that ). In line with the increased mRNA expres , Fig. 1 Fig. 1 Supplementary Fig. 1e Fig. Supplementary 5 , was the most upregulated gene at gene the , upregulated most was the Supplementary Supplementary Fig. 1d d ). In contrast, selective enrichment Fig. 1 Fig. ). Among the factors involved in in involved factors the Among ). ). Supplementary Fig. 1a Fig. Supplementary a ). Using these mouse liver ). Usingmouse these Fig. 1 Fig. Supplementary Supplementary ). This dramatic dramatic ). This ). ). Notably, the f ). Fig. 1 Fig. Fig. 1 Fig. , b b ). ). ). ). b ------

and earliest changed species among the tested AA metabolites ( ( Supplementary increased Fig. 2b robustly as not were 20-HETE, and 15- 5-, e.g., HETEs, other of levels the whereas insult, IR hepatic after mice of serum and tissue hepatic the both in content increased most markedly the showed 12-HETE enzymes, P450 cytochrome or genases lipoxy by catalyzed is formation whose metabolites AA major the ( 12-HETE to AA of conversion for activity specific with lipoxygenase a is ALOX12 above, noted As ALOX12 expression during ischemia–reperfusion 12-HETE is the primary metabolite affected by changes in HETE accumulation during hepatic IR injury, injury, IR ( hepatic during accumulation HETE ( cells control to compared as lysates cell in content 12-HETE increased pression whereas inhibited, greatly was production 12-HETE we genetically knocked down ( expression protein ALOX12 in increase dramatic ( by hypoxia challenged hepatocytes and primary livers ischemic both in ALOX12 of upregulation the with sync in temporally was HETE and expression or NF- in serum levels of liver enzymes, tissue necrosis, inflammatory marker relativedifferencenoted thatsignificant no shouldbe controlto mice ( surgery controls IR hepaticresponse to in comparedresponseas NTG to inflammatory and death cell function, of expression ( reperfusion after at h 6 in blunted of the livers of NF- activation differences, these with accordance of livers the into infiltration cell inflammatory less by as well as mediators, tory accompanied by lower expression levels of genes encoding inflamma ( chemokines and cytokines inflammatory Alox12 ( regulators apoptotic staining of expression morphological the and both in differences marked by evidenced by liver the in ameliorated greatly were death cell and necrosis tissue IR-induced finding, this with accordance In ( controls WT to compared as IR hepatic to subjected mice in decreased significantly were levels aminotransferase Serum reperfusion. 6-h by followed ischemia liver 60-min a to mice ALOX12–12-HETE axis in hepatic IR injury, we subjected To obtain direct evidence supporting a causative role for changes in the ischemia–reperfusion-induced liver injury Manipulation of ( resection liver during surgery IR hepatic to to subjected compared individuals in when baseline IR during greater significantly were secretion ( controls in liver and serum relative to the corresponding nontransgenic (NTG) mice, whereas control (WT) wild-type to compared as serum and liver both in tent deficiency 2d Fig. an Alox12 i. 2b Fig. To obtain more evidence supporting the role of ALOX12 in 12- in ALOX12 of role the supporting evidence more obtain To In contrast with with contrast In Alox12 upeetr Fg 2b Fig. Supplementary deficiency significantly lowered the serum concentrations of concentrations serum the lowered significantly deficiency , -KO) mice and mice with hepatocyte-specific expression of of expression hepatocyte-specific with mice and mice -KO) e , ). In accordance with the data obtained obtained data the with accordance In ). c Fig. 2h Fig. transgene ( transgene . n utrd rmr hptcts hpxa nue a induced hypoxia hepatocytes, primary cultured In ). Fig. 2 Fig. in vivo in Alox12 Fig. 2d Fig. Alox12 Alox12 advance online publication online advance j , κ ). i Alox12 correlated with significantly lower 12-HETE con 12-HETE lower significantly with correlated ). Finally, we found that 12-HETE production and production 12-HETE that Finally, ). we found B signaling was observed in Alox12 -KO mice at 6 h after reperfusion ( reperfusion after h 6 at mice -KO elicited a remarkable exacerbation of liver dys liver of exacerbation remarkable a elicited – Alox12 -HTG mice exhibited greater 12-HETE -HTG contentgreater mice exhibited 12-HETE g , Alox12 c ). ). Indeed, 12-HETE was the most dramatically Fig. 3 Fig. expression affects the degree of deficiency, the hepatocyte-specific over hepatocyte-specific the deficiency, -HTG) -HTG) were ( generated Alox12 , -KO mice -KO as to mice compared controls WT c h . utemr, h ices i 12- in increase the Furthermore, ). ). Supplementary Fig. 2a Fig. Supplementary via short hairpin RNA (shRNA), Supplementary Fig. 3 Fig. Supplementary Fig. 3 Fig. Alox12 Fig. 3b Fig.

Alox12 nature medicine nature e in vitro in Alox12 ). This effect was was effect This ). Fig. 2d Fig. -KO mice or mice κ Supplementary Supplementary – B signaling was B signaling Alox12 d knockout, as as knockout, Fig. 3f Fig. ). Moreover, ). Fig. 2 Fig. Alox12 Alox12 -knockout -knockout , ALOX12 ALOX12 , ). Among Among ). , e Fig. 3 Fig. ). When When ). overex Fig. 2 , a g and and -KO ). In ). -KO ). It ). a ). ). a ------

© 2017 Nature America, Inc., part of Springer Nature. All rights reserved. two-tailed Student’s individuals subjected to hepatic IR surgery ( mice for each group with 24 images for each mouse. Scale bar, 20 immunohistological staining showing ALOX12 expression profiles in the livers of mice after hepatic ischemia for 60 min or at 6 h after reperfusion. IR surgery. B, baseline; I, ischemia; R, reperfusion. GAPDH served as a loading control. for each band. ( levels in the livers of mice subjected to 60-min ischemia followed by reperfusion for the indicated periods. of ischemia (isch) or at 3, 6, 12 and 24 h after reperfusion (rep). group for proteomics analysis. ( the DGE. In analysis with Pearson correlation coefficient. ( indicate area. Correlation tracks are shown inside the circular layout as follows: red bars represent positive correlation ( correlations with a −log ( differentially expressed genes between ischemia and sham groups. The ten most significantly enriched pathways ( IR injury) using the for each period of hepatic IR injury determined by the normalized average gene expression level of biological replicates ( of hepatic IR injury. Mice without blood blockage served as controls (sham group). ( Figure 1 nature medicine nature statistical plots, data are shown as the mean d ) Circos plot showing the correlations between the proteomics and DGE data. Histograms with a blue border at the outer edge of the circle represent ACTB ) in the livers of individuals subjected to hepatic IR surgery due to liver resection.

r values larger than 0.50; green bars represent negative correlations, and the light green area indicates an ALOX12 is dramatically upregulated during hepatic IR injury. ( a – e i , ) Representative western blot (of n = 5 mice in each group for DGE; Z t

-score transformation method. ( -test ( c d a f j 10 advance online publication online advance Pathway Cytokine−cytokine receptorinteraction of hepaticIR Time points

P Relative Alox12 expression ALOX12 value, and only correlations with

f (normalized to Actb) ) or one-way ANOVA with Tamhane’s T2 (M) 10 15 Sham group 0 5 Digital geneexpression(DGE JAK−STAT signalingpathwa Steroid hormonebiosynthesis Arachidonic acidmetabolism Sham Sham f Ischemic ) Relative Hematopoietic celllineage stage Linoleic acidmetabolism –1 IL-17 signalingpathwa

Isch 1 TNF signalingpathway * 0

RNA samples Retinol metabolism h ** Rep 3 ** Potential determinant(s)ofIRproces 3 Tight junction Pathway P * Alox12 The overviewofhepaticIRinjur <0.05 <–0. h >0.5 RNA andproteinextractio and purificationfromliver Isch 1h n

Rep 6 ** * 5 Reperfusion stage:tissueandserum ± Bioinformatics analysi = 6 subjects in each group with 24 images for each subject). Scale bar, 20 6 e s.e.m. a ( ) ) Heat maps generated using the mRNA expression of members in the AA metabolism pathway detected by h collection ateachtimepoint r

Rep 12 =0.58; Arachidonic acidmetabolism mRNA expression levels (normalized to that of n y y = 5 western blots) showing ALOX12 protein expression in the livers of individuals subjected to hepatic

n h

c Rep 24 = 15 mice in each group with samples from 5 mice pooled for proteomics analysis; P ) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of the identified 12 value=8.00 3.

Rep 6 Hypergeometric samples Protein

h 3 0 s n

P y h Proteomic .5 g < 0.05 (by correlation analysis with Pearson correlation coefficient) extend to the light blue s

n Relative ALOX12 expression × = 5 mice in each group. ( 4. IR group (normalized to ACTB) 10 4 0 s 0 1 2 3 4 µ –5 Z Negativ -score m. ( ) control –log .5 Gene numbe Reperfusion Ischemi Baseline a post hoc * k 24 (h 10 ) Schematic of the experimental strategy used to identify the potential determinants 2. 2. 3. 3. 10.0 15.0 5. e ) Representative images of ALOX12 expression profiles in the liver sections of 0 5 0 5 ( 0 5. P * value) 0 ) n.s. a r b Rep 24 Rep 12 e Cluster no k analysis ( ALOX12 Rep 6 Rep 3 Isch 1 Z Sham -score Sham b ) Six Baseline h h h h h h ALOX12 ALOX12 Low GAPDH GAPDH i C . n Isch 1 = 19 individuals in each group. ( k n C 1 g g -means clusters showing distinct and specific expression trends = 19 individuals in each group. ( ) The relative mRNA expression of ) was used. * h R B Sham C 2 Case Rep 3 β B I -actin-encoding gene Ischemia High 1 h 3 Isch 1 Rep 6 P R h Case < 0.05, ** h n C4 B I ep 3 = 6 mice in each group; Rep 12 Reperfusio C 2 r > 0), and bars located in the light red area R R h P h ep 6 r Rep 24 < 0.05 by Fisher’s exact test) are shown. value lower than −0.50 by correlation ase n P I R h < 0.01. n.s., not significant. In all C5 3 Actb n R h ep 12 = 5 mice in each period of hepatic µ Pla2g12b Cyp2b9 Cyp2c2 Cyp4a3 Pla2g4a Cyp4a1 Plb1 Cyp2j Ptgi Pla2g12a Cyp2c3 Cyp4a3 Cyp2c4 Cyp2c3 Cyp2c5 Cbr3 Cyp2c3 Gpx3 Cyp2c5 Ggt5 Cyp2c7 Hpgd Cyp4a12b Pla2g Gpx8 Ptges2 Cyp2c6 Cyp2e1 Ptgs Ephx Cyp4f18 Cbr1 Cyp4a12a Cyp2j Lta4 Ptges3 Pla2g16 Cyp4f13 Cyp2j Gpx7 Ltc4 Cyp4f14 Ptgd Gpx1 Gpx6 Tbxas1 Cyp2c2 Fam213b Ggt1 Pla2g4b Cyp2j13 Cyp2u1 Gpx2 Cyp2j Ptgs Cyp2c6 Pla2g10 Cyp2b1 Pla2g Pla2g2d Pla2g Cbr2 Ptge Cyp4a1 Cyp2c5 Alox Alox12 Alox15 Alox12e Alox j m. For statistical analysis, a s ) Representative images of h s 8 5 1 2 s s Negative 2 s 9 5 6 8 6 3 5 control ) in the livers of mice at 1 h after R B 9 1 0 9 2 0 7 0 8 4 0 8 3 6 4 5 0 h hR Case ALOX12 ) ALOX12 protein expression (log Relative gene ep 24 I 2 expressio C foldchange) 6 4 2 –2 –1 0 1 h n n (normalized to that

= 3 western blots s e l c i t r a n = 3 pools in each n = 4  © 2017 Nature America, Inc., part of Springer Nature. All rights reserved. n examination; 12-HETE liver for groups sham in mice Alox12 of (right) serum in as treated hepatocytes primary of lysates cell the following t mean the as shown hypoxia. of min 60 and 30 10, after hepatocytes primary of lysates cell the in content 12-HETE ( ischemia hepatic after points time indicated at mice of livers the in group, process. IR hepatic the during points time indicated the at mice of livers the in enzymes P450 cytochrome or lipoxygenases by 2 Figure by IR on hepatic signaling the of ALOX12–12-HETE effect for pathological responsible mechanisms underlying the explored next We induction of a burst of inflammation 12-HETE promotes hepatic ischemia–reperfusion injury via ent with a previous report with A  (M) -test. ( -test. = 17) and serum (right; (right; serum and = 17)

s e l c i t r post hoc post Alox12 n -HTG and NTG control mice in the sham and 6-h reperfusion groups. In In groups. reperfusion 6-h and sham the in mice control NTG and -HTG d = 15 mice in other groups. ( groups. other in mice = 15

, e ALOX12 directly promotes 12-HETE accumulation in hepatic IR injury. ( IR hepatic in accumulation 12-HETE promotes directly ALOX12 Alox12 i f c a ) Representative western blot (of two western blots for each band) showing ALOX12 protein expression in cell lysates of primary hepatocytes hepatocytes primary of lysates cell in expression protein ALOX12 showing band) each for blots western two (of blot western ) Representative 12-HETE content in 12-HETE content in cell Sham analysis was used. * used. was analysis liver (fold change) lysate (fold change) ALOX12 protein overexpression at the baseline condition, which is consist 20 40 60 0 1 2 3

0 expression (fold change) Alox12 knockdown ( knockdown 0 1 2 3 sh sh 1 0 ± NTG Normoxia Alox12 GFP Isch 1 Sham s.e.m. * s.e.m. ** * -KO and WT controls at 6 h after reperfusion or in the sham group. ( group. sham the in or reperfusion 6 h after at controls WT and -KO ** ** h 2 0 n * HTG = 19) of human individuals subjected to hepatic IR injury. For statistical analysis in in analysis statistical injury. For IR hepatic to subjected individuals human of = 19) 12-HETE ALOX12 1 Live 6 ( P d r Fig. < 0.05, ** < 0.05, ) or ) or 3 0 Hypoxia Rep 3 Hypoxia (min) Rep 6 ** P < 0.05, ** < 0.05, * * Alox12 3 b h and h 4 0 ) Time-dependent changes in the profiles of relative ALOX12 protein expression, 12-HETE content and AA content content AA and content 12-HETE expression, protein ALOX12 relative of profiles the in changes ) Time-dependent ** overexpression ( overexpression P Supplementary Fig. 3 < 0.01, **** < 0.01, Rep 6 5 0 g 12-HETE content in 12-HETE content in cell P serum (fold change) < 0.01, *** < 0.01, lysate (fold change) h 10 20 30 10 15 20 6 0 0 0 5 d n ( ** ** GFP = 6 mice in rep 6 h groups for liver 12-HETE examination. ( examination. 12-HETE liver for 6 h groups rep in = 6 mice Alox12 Sham f Rep 12 Normoxia ) and ) and 0 * P ** 0 1 2 3 < 0.0001 compared to ischemia at 0 min ( 0 min at ischemia to compared < 0.0001

e

* **

S P ) ) by adenovirus under the condition of normoxia or hypoxia for 1 h. ( 1 h. for hypoxia or normoxia of condition the under adenovirus ) by change (fold

h * erum

< 0.001. In all statistical plots, data are shown as mean mean as shown are data plots, statistical all In < 0.001. e ( content 12-HETE g ). For For ). ). Rep 24 Rep 6 Hypoxia n * * = 6 mice in each time point). ( point). time each in = 6 mice d h h – g - , , 11,12-EE 8,9-EE 18-HETE 19-HETE 14,15-EE 20-HETE 5-HETE 11,12-DHET 14,15-DHET 15-HETE 8,9-DHET 16-HETE 9-HETE 11-HETE 12-HETE n = 4 independent experiments. ( experiments. = 4 independent of correlation of genes in the modules with expression of of expression with modules the in genes of correlation of modules with different colors. We then established the overall degrees these labeled and expression coordinated of degree high a with ules numerous gene we genes mod across all samples, identified between pression correlation (WGCNA)network analysis pairwise to identify in shown surgery IR hepatic mining the transcriptome data set of livers from mice to subjected the j h h 12-HETE content in 12-HETE content in and and T

a liver (fold change) liver (fold change) T T ) Heat map showing the contents of major AA metabolites catalyzed catalyzed metabolites AA major of contents the showing map ) Heat ALOX1 ALOX1 10 20 30 GAPD GAPDH Relative metabolite 0 1 2 3 4 (log e d 0 i , , abundanc 2 foldchange n KO WT H 2 2 = 5 mice in each group for detection of serum 12-HETE; 12-HETE; serum of detection for group each in = 5 mice –3 –2 –1 0 1 2 3 Sham i ** ) 12-HETE content in the livers (left) and serum (right) of (right) serum and (left) livers the in content ) 12-HETE e * Baselin sh GFP * Live ) advance online publication online advance GFP * Live n r = 3 independent experiments. For For experiments. = 3 independent e Normoxia b n.s. Normoxia r ALOX12 protein expression Rep 6 c

b (fold change) ) Relative protein expression level of ALOX12 and and ALOX12 of level expression protein ) Relative ) or hypoxia at 0 min ( 0 min at hypoxia ) or * 10 0 2 4 6 8 sh h Alox12 1 0 Alox12 Ischemia h Figure 1 Figure ) 12-HETE content in the livers (left) and and (left) livers the in content ) 12-HETE * 2 0

f 12-HETE content in AA 12-HETE ALOX1

– 12-HETE content in

j serum (fold change) , one-way ANOVA with Tamhane’s with ANOVA T2 , one-way j serum (fold change) Ischemia (min ) 12-HETE content in the livers (left; (left; livers the in content ) 12-HETE 3 0 0 1 2 3 4 * * 0 2 4 6 2 ± sh Reperfusio GFP a s.e.m. 4 0 GFP . Using weighted gene coex gene weighted Using . Sham Hypoxi * Hypoxi * * 5 0 ) ** * c S * * n ) by two-tailed Student’s Student’s two-tailed ) by n erum f * S = 10 mice in the sham sham the in mice = 10 a , a erum g

6 0 sh Alox12 *** ) 12-HETE content in content ) 12-HETE n.s. Alox12 nature medicine nature ** * * b and and 0 Rep 6 0 2 4 6 8 10

*

) change (fold

h c , data are are , data content AA and 12-HETE Alox12 n = 5 or or - - © 2017 Nature America, Inc., part of Springer Nature. All rights reserved. Tnf group; each for mice and mice WT of livers (blue). DAPI with labeled were CD11b of staining immunofluorescent of expression the to normalized was expression injury. Gene IR hepatic 6 h after at and before and and mice WT of serum the in content CCL2 and (TNF) groups. indicated the in mice of livers the in (c-CASP3) caspase-3 cleaved and BAX BCL2, of 20 (40×); staining H&E reperfusion. 6 h after ( (TUNEL) labeling nick-end dUTP transferase deoxynucleotidyl reperfusion. 6 h after at or baseline) (at group sham the in controls WT or (KO) mice deficient 3 Figure nature medicine nature mean the as shown are data plots, statistical all In 0.0001. e g a in in AST (U/L) ALT (U/L) n TNF content in serum 10,000 15,000 20,000 25,000 = 6 mice in other groups. ( groups. other in = 6 mice 10,000 15,000 20,000 25,000 f 5,000 5,000

) or with Tamhane’s with ) or (M) T2 (ng/ml) 120

120 240 20 40 60 60 Alox12 0

KO WT 0 0 KO, rep6 WT, rep6 KO, sham WT, sham KO Sham Sham knockout inhibits IR-induced liver dysfunction, cell death and inflammation. ( inflammation. and death cell dysfunction, liver IR-induced inhibits knockout n.s. n.s. n.s. **** **** ****

n n Alox12 WT = 3 western blots for each band. For statistical analysis, one-way ANOVA with Bonferroni’s Bonferroni’s with ANOVA one-way analysis, statistical For band. each for blots = 3 western advance online publication online advance h h = 4 mice per group at each time point with 24 images for each mouse. Scale bar, 200 Scale mouse. each for images 24 with point time each at group per = 4 mice CD11b µ m for TUNEL images. NA, necrotic area. ( area. necrotic NA, images. TUNEL m for Rep 6 -KO mice in the sham group and at 6 h after reperfusion. p-, phosphorylated. GAPDH served as a loading control. control. a loading as served GAPDH phosphorylated. p-, reperfusion. 6 h after at and group sham the in mice -KO **** n **** **** = 4 mice in each group with 24 images for each mouse. Scale bar, 20 Scale mouse. each for images 24 with group each in = 4 mice f ) The relative mRNA expression levels of cytokines and chemokines in the livers of WT mice and and mice WT of livers the in chemokines and cytokines of levels expression mRNA relative ) The Rep 6 h post hoc post + (left) and Ly6G and (left) h b analysis (for (for analysis CCL2 content in serum KO WT (pg/ml) c

2,000 4,000 6,000 8,000 KO WT Sham 0 Sham Alox12 + (right) inflammatory cells (red) in liver sections from mice in the indicated groups. The nuclei nuclei The groups. indicated the in mice from sections liver in (red) cells inflammatory (right) a 10

, CCL2 in in , CCL2 × Sham n.s. ± Ly6G s.e.m. -KO mice in the sham group and at 6 h after reperfusion. reperfusion. 6 h after at and group sham the in mice -KO c *** ) staining of liver sections from WT mice and and mice WT from sections liver of ) staining d Rep 6 e Rep 6 ) Representative western blot (of (of blot western ) Representative and and Sham Rep 6 h h Il6 *** and and h Cxcl2 40 × c-CASP3 d h f GAPDH in in GAPDH

p-I BCL2 p-p65 BAX I f κ κ p65

) was used. * used. ) was Relative mRNA expression B B α α

n (normalized to Actb) n Actb = 10 mice in each group. ( group. each in mice = 10 = 6 mice for each group. ( group. each for = 6 mice a 120 ) ALT and AST levels in the serum of of serum the in ) ALT levels AST and 10 15 20 45 70 95 0 5 . . n WT WT NA = 4 mice per group. ( group. per = 4 mice n.s. n = 3 western blots) showing protein expression expression protein showing blots) = 3 western *** µ WT, sham KO, sham P m. ( m. Tnf < 0.05, ** < 0.05, Sham Sham h NA ** 10 Alox12 ) Activation of NF- of ) Activation µ NA NA × m for H&E staining (10×); 50 50 (10×); staining H&E m for

post hoc post KO KO -KO mice in the sham group or at or group sham the in mice -KO P n.s. < 0.01, *** < 0.01, WT, rep6 KO, rep6 n NA NA = 4 mice in sham WT group group WT sham in = 4 mice * Rep 6 e b Il6 analysis (for TNF in in TNF (for analysis ) Tumor necrosis factor factor ) Tumor necrosis g , c ) Representative images of images ) Representative ) H&E ( ) H&E h * h h WT WT κ B signaling in the the in B signaling Alox12 P s e l c i t r a < 0.001, **** < 0.001, b Rep 6 Rep 6 Alox12 ) and terminal terminal ) and n.s. 40 * -KO mice mice -KO × Cxcl2 h h - KO KO µ * n e m for m for = 6 and and P <  © 2017 Nature America, Inc., part of Springer Nature. All rights reserved. kinase C (PKC)–JNK signaling, whereas silencing of the other three three other the of silencing whereas signaling, (PKC)–JNK C kinase hepatocyte cell line abolished activation 12-HETE-induced of protein insult. IR hepatic to response tify the receptor isotype responsible for 12-HETE-elicited signaling in Therefore, GPR31 and GPR120, GPR40, we GPR75 to screened iden HETEs for receptors be to reported been also have GPR75, and be to acids known fatty long-chain are to responsive GPR120) as known (also FFAR4 and GPR40) as (FFAR1)–FFAR4), 1 receptor acid fatty FFAR1known free (also (i.e., metabolites regulation, tory and as GPCRs can receptors serve for numerous lipid ( signaling (GPCR) receptor ing genes were clearly enriched for terms related to G-protein-coupled ( hepatic IR terms injury to in enriched 12-HETE-mediated determine 5a Fig. Supplementary ( conditions hypoxia and normoxia under 12-HETE to genes that leading overlapped of changed in the response highlighted ( network coexpression gene ( qPCR by confirmed and (GSEA) analysis enrichment set gene by revealed as processes, inflammation-related in implicated mostly were change) fold of basis the on determined (as 12-HETE by affected most genes and pathways pres cellular the Remarkably, the 12-HETE. of absence or in ence hypoxia primary by in challenge or normoxia analyses under RNA-seq hepatocytes additional performed we mation, inflam promotes 12-HETE how of understanding better a Togain reperfusion injury GPR31 is responsible for 12-HETE-mediated hepatic ischemia– in vitro hepatocytes in primary induction of on hypoxic injury inflammation inhibitor ALOX12 specific a ML355, by its primary product, 12-HETE. Blocking 12-HETE production with mediated is ALOX12 of action proinflammatory the that notion the gene ( treatment vehicle inflammatory to compared as and expression phosphorylation ERK and p38 JNK, p65, greater in resulted 12-HETE with hepatocytes primary of treatment ( reperfusion at after h 6 in confirmed further was signals inflammatory key of activation the on ALOX12 of effect ing ( challenge hypoxia upon hepatocytes primary in readouts these on effect opposite the to led ( genes inflammatory of expression representative enhanced with correlated which tocytes, in the yellow cluster, when p38 and ERK, the major downstream mediators of pathways enriched of the phosphorylation NF-κB subunit p65 and MAPK subunits JNK, IR. hepatic during regulation inflammatory in 12-HETE ( mediators proinflammatory of upregulation the with concomitantly or before occurred production 12-HETE elevated the that showed profile ral ( mediators inflammatory ing encod genes for and pathways signaling inflammatory multiple in levels, and members of the yellow module were enriched significantly with correlation positive of degree highest the showed module yellow the Specifically, levels. 12-HETE A  GPCRs had no detectable impact on 12-HETE function ( Fig. Fig. 5

GPCR-mediated signaling has long been implicated in inflamma in implicated been long has signaling GPCR-mediated increased significantly observed we conclusion, this Supporting s e l c i t r b ( ). ). Notably, the genes neighboring most to connected the lead Fig. 4e Fig. Fig. 5 Fig. Supplementary Fig. 4a Fig. Supplementary 1 8 . Among four well-established ‘lipid-sensing’ GPCRs GPCRs ‘lipid-sensing’ well-established four Among . – a g and and ). Supplementary Fig. 4c Fig. Supplementary Supplementary Fig. 5a Fig. Supplementary ) were used for gene ontology (GO) analysis analysis (GO) ontology gene for used were ) Supplementary Fig. 4g Fig. Supplementary Supplementary Fig. 4e Fig. Supplementary Alox12 Alox12 Fig. 5 Fig. Fig. 4 Fig. Fig. 4a Fig. GPR31 was overexpressed in primary hepa -KO and and -KO b , a Alox12 b ). 19– ), the heaviest 5% weight edges edges weight 5% heaviest the ), ), suggesting a potential role for a potential ), suggesting 1 , knockdown in the human L02 L02 human the in knockdown b 7 2 , largely abolished the impact impact the abolished largely , ). Examination of the tempo the of Examination ). 1 . Two other GPCRs, GPR31 GPR31 GPCRs, Twoother . expression and 12-HETE 12-HETE and expression Alox12 , d , Fig. 4c Fig. b ). ). , ). In the context of a of context the In ). h Alox12 -HTG mice mice -HTG , ). Moreover,). direct f ). The exacerbat The ). , d ), supporting supporting ), Fig. Fig. 5 knockdown knockdown Fig. 5 Fig. c in vivo in ). ). The a 22 and and , 2 3 ------.

KO mice ( rvosy eotd LX2 niios M35 n CDC and ML355 two inhibitors, using ALOX12 production reported 12-HETE previously injury, blocking IR whether hepatic on examined 12-HETE we of effect critical the Considering reperfusion injury in mice and pigs Blocking 12-HETE production inhibits hepatic ischemia– ( reperfusion 6 h after by inhibited cantly and signifi were consistently induction and inflammatory death cell 5f Fig. global generated next we injury, liver 12-HETE-mediated ( 5c Fig. ( RNA-seq by demonstrated as genes, restored the effects of 12-HETE on inflammation-related pathways and Fig. 5c firmed by RNA-seq and GSEA analyses ( of effect effective in inhibiting 12-HETE production and suppressing liver liver suppressing damage afterand hepatic IR insult,production but ML355 12-HETE exhibited a strongerinhibiting capacityin effective could suppress these pathologies. Both ALOX12 inhibitors were highly time course profiles of serum alanine transaminase (ALT), aspartate (ALT), aspartate transaminase alanine serum of profiles course time and values peak The period. ischemic the for min 30–120 of range a with macaque, rhesus the primate, nonhuman a in model IR hepatic a established we Thus, humans. resembles closely more that model animal an of use requires still clinic the to translation potential a of demonstration surgery, IR by induced damage liver treating for egy conserved strat promising a a and injury indicate IR hepatic for injury mechanism IR underlying hepatic of models pig and mouse both in treatment ML355 of effects therapeutic major the Although hepatic ischemia–reperfusion ML355 inhibits liver damage in a nonhuman primate model of ( point time examined each at response inflammatory and death cell ML355 treatment led to a considerable inhibition of liver dysfunction, pared to vehicle-treated controls ( com as group ML355-treated the in insult IR hepatic after duction 7a Fig. organs pig and human and anatomy between biology in size, similarities the cal model for disease therapy, surgical especially operation, because of Compared to the mouse, the pig provides a more appropriate preclini application. for clinical therapies effective in developing a roadblock lar level, phenotypic rodents differences between and humans remain an invaluable contribution to understanding of at disease the molecu ( treatment all ML355 by were attenuated transduction, signaling and infiltration cell inflammatory production, chemokine and cytokine including responses, matory group vehicle-treated at the to death ( compared cell as point lower time and tested levels each aminotransferase serum reduced significantly the by evidenced as period, IR hepatic entire the over ( dose was damage dependent and consistent with liver the inhibition of 12-HETE production against ML355 of effect protective the that ( CDC than Supplementary Fig. 5d Supplementary Supplementary Fig. 6g Supplementary Supplementary Fig. 7c Fig. Supplementary Supplementary Fig. 6d Fig. Supplementary To obtain Although small experimental animals, particularly mice, have made ). ). Notably, GPR31 rescue in L02 GPR31-deficient cells globally ) and exposed them to hepatic IR surgery. Liver dysfunction, dysfunction, Liver surgery. IR hepatic to them exposed and ) 2 ), we observed a reduction in liver and serum 12-HETE pro 12-HETE serum and liver in reduction a observed we ), ) and further confirmed by western blotting and qPCR qPCR and blotting western by confirmed further and ) 6 . In a porcine model of hepatic IR injury ( injury IR hepatic of model porcine a In . GPR31 Gpr31b in vivo upeetr Fg 6a Fig. Supplementary advance online publication online advance inactivation on 12-HETE function was further con further was function 12-HETE on inactivation corresponds to corresponds evidence supporting the role of GPR31 in ALOX12– Gpr31b Fig. 5f Fig. – , – e – i h ). ). ). Remarkably, hepatic IR–induced inflam f ). ablation as compared to WT controls at controls WT to compared as ablation ). The protective effects of ML355 lasted lasted ML355 of effects protective The ). – h and and Supplementary Supplementary Fig. 7b GPR31 Supplementary Fig. 6j Fig. Supplementary Supplementary Fig. 5g Fig. Supplementary Fig. 5d Fig. – Fig. 5d c in humans; ). Furthermore, we found found we Furthermore, ). , e , e and and

and nature medicine nature Supplementary Supplementary Supplementary Supplementary Supplementary Supplementary Supplementary ). ). Moreover, – m , Gpr31b h ). 17 ). , 24 , 2 5 ------,

© 2017 Nature America, Inc., part of Springer Nature. All rights reserved. one-way ANOVA with Bonferroni’s normalized to that of normoxia (Nor) or hypoxia (Hyp) conditions. For western blot analysis, GAPDH served as a loading control. In the qPCR assay, mRNA expression of ofgenes p65, was JNK, p38 and ERK ( with 12-HETE for the indicated periods. JNK, p38 and ERK protein expression ( are indicated by thicker lines and darker colors. ( genes related to the pathways listed in inflammatory pathways ( ( and the coexpression modules (represented by the colored boxes at the left of the table) determined based on WGCNA using the DGE data in Figure 4 ( ards stand as used were clinic the in resection liver to subjected viduals transaminase (AST) and alkaline phosphatase (ALP) observed in indi nature medicine nature * n P = 6 groups with 5 mice in each). Right, KEGG analysis results for genes in the yellow coexpression module that were significantly enriched in the indicated < 0.05, ** Time (min) c a GAPDH Fig. 6 Fig. p-ERK p-JN p-p3 p-p6

ERK JN 12-HETE promotes hepatic IR injury by inducing a burst of inflammation. ( p38 p65 K K 8 5 Alox12 –0.17 –0.36 –0.13 –0.59 –0.24 –0.69 –0.33 –0.84 –0.45 a 0.04 0.40 0.12 0.84 0.76 0.46 0.74 0.09 0.47 0.69 0.10 0.14 0.48 P ). Liver damage in the monkeys was further evaluated evaluated further was monkeys the in damage Liver ). < 0.01, *** 1 0

Actb GAPDH f p-ERK p-JNK p-p3 p-p6 advance online publication online advance 12-HETE ERK JN P p38 p65 −0.39 −0.54 –0.18 –0.04 –0.10 –0.66 –0.66 –0.79 –0.94 –0.52 –0.02 0.14 0.55 0.04 0.78 0.36 0.25 0.63 0.11 0.62 0.81 0.44 < 0.05 by Fisher’s exact test). ( . 2 0 K 8 5 n P f = 4 independent experiments for < 0.001, **** ) and relative mRNA expression of cytokines and chemokines ( 12-HETE Vehicl post hoc 4 0 Normoxia e –log –1 –0. 0 0.5 1 a c . The correlation was visualized in a user-defined manner with line width and color saturation; stronger correlations ) and qPCR results showing relative mRNA expression levels for n 10 5 2.13 2.21 2.42 2.43 2.49 2.52 2.74 2.86 3.19 3.63 4.21 5.01 7.46 = 3 independent experiments with 2 replicates each. ( 6 0

ML355 analysis (for ( P P value) < 0.0001. In all statistical plots, data are shown as mean c , JAK−STAT signalingpathway PI3K−AKT signalingpathway T RAS signalingpathway NF TGF Apoptosis Cytosolic DNA−sensingpathwa NOD-like receptorsignalingpathway Cytokine−cytokine receptorinteractio MAPK signalingpathway TNF signalingpathway IL-17 signalingpathway 0 d H ) Representative western blots ( 17 celldifferentiation - Vehicl κ - B signalingpathway β signalingpathwa Ccl2 d Hypoxi e b Relative mRNA expression ) Correlations between the expression of

in (normalized to Actb) e a ML355 0 2 4 6 8 d ; and for n = 3 independent experiments with two replicates each for * y * Il 12-HETE 1 Vehicl * 6 e and y - - e g n h ) or with Tamhane’s T2 (M) in the monkeys, whereas a 120-min ischemic period induced a liver liver a induced period ischemic 120-min a whereas monkeys, the in IR injury hepatic was to short too substantial induce period ischemic 30-min a that Wefound (CT). tomography computed perfusion by **** **** n a = 3 western blots for each band) showing phosphorylated and total p65, *** Hsp90ab Tn ) Left, Pearson correlation coefficient between g Hsp90aa b

Relative mRNA expression f (normalized to Actb) 12-HETE 6 12-HETE 3 Fgf2 g Fgf1 0 1 2 3 1 ) in primary hepatocytes treated with ML355 (10 Cxcl5 1 8 Il1 **** e *** Csf **** 5 – g Il1 Ccl2 Alox12 Il6 3 h h ) 12-HETE content ( ML355-No Vehicle-Nor * Il23a b **** 12-HET Il6 ± Il6 s.e.m. Csf or the level of 12-HETE and the expression of core , post hoc 2 Tnf Li r Fgf2 e f E

**** 12-HETE content in cell and Lt (fold change) 1 0. 0. 1. 1. 2. 2. Tn a Csf 0 5 0 5 0 5 analysis (for **** f Ccl2 Nfkb1 1 Tnfaip ML355-Hyp Vehicle-Hyp Alox12 e Nfkbi ( ), phosphorylated and total levels ML35 Vehicle d Bcl2l1 3 ) in primary hepatocytes treated Normoxia f Nfkbi a and **** 5 Ccl2 Tn Tnf b Ng Tgfbr2 f g ** **** Osmr . For statistical analysis, Alox12 f Correlation with12-HET 0. 0. Birc3 and Correlation with Rela 5 5

Ccl2 Edge width s e l c i t r a Il6 Correlatio or 12-HETE Figure 1 in Hypoxia µ **** M) under d ) was used.

Alox12 n a

1. 1. 0. 0. 0. 0. 0. 1.

E 0 0 5 6 7 8 9 0

 © 2017 Nature America, Inc., part of Springer Nature. All rights reserved. each group) in the livers of mice in the indicated groups. For statistical analysis, one-way ANOVA with Tamhane’s T2 (M) for each group with 24 images for each mouse. Scale bar, 20 staining (green) of liver sections at 6 h after reperfusion of mice with in the serum of WT mice and shared leading genes between two comparisons. Data were statistically analyzed by GSEA based on three independent experiments. ( NES, normalized enrichment score. ( GPR31 as a loading control. ( in most significantly enriched pathways ( data shown in contributor to 12-HETE function. The data used in this GO analysis were from the leading genes highlighted in the red areas of the circles in the normoxiaRNA-seq and hypoxia. Analysis included data from four independent experiments. ( indicates the number of overlapped leading-edge genes (the core genes that account for their gene set’s enrichment signal) induced by 12-HETE under of RNA-seq results from primary hepatocytes treated with 12-HETE (100 nM) under normoxia or hypoxia conditions. The red area in each numbered circle Figure 5 A  Student’s

GPR31 s e l c i t r -rescue (

GPR31 is responsible for 12-HETE-mediated hepatic IR injury. ( -, t -test ( GPR40 a and GPR31 h ) was used. * -, c a e f Supplementary Figure 5a GPR75 d 12-HET ALT (U/L) Ranked list metric Enrichment score (FDR) (100 nM) –log GPR12 GAPDH

) Dot blot representing pairwise GSEA comparison of RNA-seq data among 12-HETE-treated hepatocytes (WT control, 12-HETE hypoxia positive GPR75 GPR40 GPR31 -res) groups;

(Signal2Noise) (ES) p-PKC p-JNK 15,000 20,000 25,000 10,000 −0. −0. −0. 0 1 2 3 4 5,000 0. 10 PK JN −4 −2 10 2 0 1 3 0 2 2 1 0 WT, sham 50 C E K 0 0 0 5 0 Normalized ES:−1.2456 Nominal (positively correlated) - or GPR31 Gpr31b 12-HETE normoxiapositive sh – P Rank inorderedgenelist GF < 0.05, ** ,000 GPR120 9 -KO 12-HETE P P n.s. 10 + value: 0.003 **** n e -KO mice in the sham group or at 6 h after reperfusion. sh 10,000 = 3 in each group). Dot color indicates that the gene signature set is enriched in the group with corresponding font color. Gpr31b ) Left, GSEA showing the merged gene sets of genes in the pathways shown in 8 (positively correlated) P – GPR31 < 0.05 by Fisher’s exact test; edge number > 100) are shown. ( WT 12-HETE 3 -knockdown cells treated with 12-HETE or vehicle control for 1 h. +

-KO, sham 7 15,000 **** P sh < 0.01, *** 5 – with the heaviest 5% of edge weights ( GPR40 2 3 + 1 4 6 sh 0. 0. 0. 0.1 – −2 AST (U/L) WT, rep6 0 GPR75 10 2 0 3 4 0 2 2 1 9 8 7 6 5 4 3 10,000 15,000 20,000 25,000 5,000 0 Overlap ofleading-edgegenes Natural killercellmediatedcytotoxicity Toll-like receptorsignalingpathway NF- Chemokine signalingpathway Cytosolic DNA−sensingpathway IL-17 signalingpathway RIG-l-like receptorsignalingpathwa NOD-like receptorsignalingpathwa TNF signalingpathway Cytokine−cytokine receptorinteractio 14 + (positively correlated) GPR31 P 70 0 0 5 < 0.001, **** κ Rank inorderedgenelist sh B signalingpathway 5,000 h – GPR12 -res 12-HETE µ m. ( 10 n.s. + 10,000 Gpr31b **** 0 Normalized ES:1.2220 Nominal (positively correlated) GPR31 h 15 ) Relative mRNA expression of inflammatory cytokines and chemokines ( Gpr31b d vs -KO 12-HETE -KO, rep6 P 15,000

value:0.013 GPR31 GPR31 GPR31 20 a P **** ) The major pathways contributing to 12-HETE function determined by GSEA

< 0.0001. In all statistical plots, data are shown as mean Toll-like receptor signaling pathway vs knockout or WT controls. The nuclei were labeled with DAPI (blue).

NF- >25 Cytokine−cytokine receptor interaction -KO 12-HETE -KO 12-HETE -res 12-HETE WT 12-HETE

h κ

B signaling pathway y y Positive regulationoftumornecrosisfactorproductio n 49 v b s G G NOD-like receptor signaling pathway b

) Schematic of GO analysis showing that GPCR signaling is a major Defense responsetoGram-negativebacterium n G PR PR vs GFP12-HETE PR31 = 4 samples from independent experiments for each group). The 15 leading genesandtheirneighbor RIG-I-like receptorTNF signaling pathway Regulation ofactincytoskeletonorganizatio 3 3 1 h 1 1 -K -res 12 Network edgesbetween

Relative mRNA expression -K O 99 n O 12-HETE Positive regulationofIL-6productio (normalized to Actb) G-protein-coupled receptorsignalin = 10 mice in each group. ( 12-HET .. Positive regulationofphagocytosi 0. 0. 1. 1. 2. - HETE 0 5 0 5 0 Cytosolic DNA−sensing pathway . IL-17 signaling pathway A Integrin-mediated signaling E advance online publication online advance ** Tnf 40 Natural killer cell mediated cytotoxicity Lo Chemokine signaling pathway Innate immunerespons Immune systemproces * B cellreceptorsignalin Inflammatory response WT, rep6 w Neutrophil chemotaxi Correlatio c ) Levels of phosphorylated and total PKC and JNK Cell adhesion Il1b Phagocytosi * g n h s Chemotaxis n = 3 independent experiments. GAPDH served Gpr31b-KO WT High 2 Gpr31 Edge weight–basedsortin Sham .. g n n e g n s s s s ** Number ofnetworkedges Figure 5 Il6 b . -KO, rep6 25 intensit g = FDR 0 <0.2 <0.0 Color <0.5 >0.5 ) Representative images of TUNEL post hoc Heaviest 5%ofedge 3 50 5 5 –log Edge number 20 25 y GO enrichment Ccl2 d 7 0 ** sample h 5 0 . Right, Venn diagram showing red 25 Enriched in: 10 analysi weight

00 ( 50 P 30 value) IR 6 sample 75 analysis ( blue s g 0 s 35 h Cxcl Dot size f = NES ** >1.2 ) ALT and AST content >1.0 >1.5 <1.0 ±

2 s.e.m. 5 nature medicine nature f ) or a two-tailed GPR31 n = 4 mice in n = 4 mice -KO and © 2017 Nature America, Inc., part of Springer Nature. All rights reserved. bar, 1 mm. bar,( 1 mm. surgery. MI 7 d at after or group sham the in control vehicle or ML355 with ( perfusion. HPP, portal hepatic perfusion; monkey. HAP, arterial hepatic reperfusion. after 2 weeks and 1 week h, 48 Bonferroni’s with ANOVA one-way reperfusion. h after 48 at groups vehicle-treated and ML355- in monkeys of lobes control and lobes hepatic IR-challenged the in mediators inflammatory of bar, 200 monkey. Scale reperfusion. h of 48 after groups vehicle-treated and ML355- in sections liver monkey on staining ** * reperfusion. of periods indicated the by followed ischemia hepatic of min 60 after vehicle with or weight) body kg per mg (3 ML355 with ( (M) * w, weeks. ( 6 Figure nature medicine nature n a = 8 mice in each group with 7 consecutive sections for each mouse. Scale bar, 5 mm. In all statistical plots, data are shown as mean mean as shown are data plots, statistical all bar,In 5 mm. Scale mouse. each for sections 7 consecutive with group each in = 8 mice , , P n < 0.01, *** < 0.01, = 38 individuals) and of monkeys subjected to 60-min ischemia followed by periods of reperfusion of varying duration ( duration varying of reperfusion of periods by followed ischemia 60-min to subjected monkeys of and individuals) = 38 a ) and with Bonferroni’s Bonferroni’s with ) and

ML355 inhibits liver damage in a monkey hepatic IR model. ( model. IR hepatic a monkey in damage liver inhibits ML355 P h n < 0.05, ** < 0.05, g f d b a ) Representative images of 2,3,5-triphenyl-2 of images ) Representative = 4 monkeys in the vehicle-treated group and and group vehicle-treated the in = 4 monkeys ALT in serum (U/L)

ALT in serum (U/L) Monkey Huma P 10 20 30 40 50

Baseline 20 40 60 80 < 0.001 compared to corresponding time points of vehicle group by two-tailed Student’s Student’s two-tailed by group vehicle of points time corresponding to compared < 0.001 0 0 0 0 0 0 0 0 0 Isch 0.5 0 0 Baseline ML355 Vehicle HAP HPP H&E n

Isch 1 h Rep 1 Rep 1 advance online publication online advance

NA h **** P Rep 3 d µ Rep 2 ****

Sham h NA Vehicle < 0.01, *** < 0.01, m for H&E staining (10×); 20 20 (10×); staining H&E m for Rep 6 Vehicl h **** d Rep 12 Rep 3 h ** Rep 24 **** post hoc post h d

**** Rep 4 e

Rep 36 **** NA h Rep 48 NA *** Rep 5 d h * Rep 72 post hoc post *** Rep 48h h d **** Rep 1 * Rep 6 h

P Rep 2 analysis ( analysis NA w Rep 7 d **

< 0.001, **** < 0.001, Rep 3 NA

ML355 w MI d Rep 4 Rep 8 w * analysis. ( analysis. w d n M = 5 monkeys at each time point for each group. Each image was generated from 39 scans in each each in scans 39 from generated was image Each group. each for point time each at = 5 monkeys

L355 AST in serum (U/L) AST in serum (U/L) 10 20 30 40 50 20 40 60 80 b Baseline

0 0 0 0 0 0 0 0 0 0 0 ). ( ). Isch 0.5 Baseline

× ×

40 10 Isch 1

c h Rep 1 Rep 1 ) Serum ALT and AST levels ( ALT levels ) Serum AST and f h ) Representative perfusion CT images of livers from monkeys treated with vehicle or ML355 at ML355 or vehicle with treated monkeys from livers of images CT perfusion ) Representative P Rep 3 d Rep 2

< 0.0001 compared to the corresponding baselines by one-way ANOVA with Tamhane’s with ANOVA T2 one-way by baselines corresponding the to compared < 0.0001 h µ *

Rep 6 ****

m for H&E staining (40×). ( (40×). staining H&E m for h

* d Rep 12 Rep 3 H

h *** Rep 24 * -tetrazolium chloride (TTC) staining of serial brain sections of mice at 24 h after stroke. stroke. h after 24 at mice of sections brain serial of staining (TTC) chloride -tetrazolium

n h Rep 4 d Rep 36 * **** = 5 monkeys in the ML355-treated group. * group. ML355-treated the in = 5 monkeys h

Rep 48 * d Rep 5 h Rep 72 h

Vehicl d Rep 1 Rep 6 h Rep 2 d w Rep 7 e Rep 3 e w a Relative mRNA expression Rep 4 d

, Rep 8

g w

b (normalized to 18S rRNA) ) Representative images of H&E staining of heart sections from mice treated treated mice from sections heart of staining H&E of images ) Representative Rep 1week ) Serum ALT, AST and ALP levels of human subjects after liver resection resection liver after subjects ALT, human of ) Serum levels ALP and AST w d 2 3 4 0 1 n h

= 4 mice for each group with 2 consecutive sections for each mouse. Scale Scale mouse. each for sections 2 consecutive with group each for = 4 mice ALP in serum (U/L) ML355 Vehicle ALP in serum (U/L) n.s. 10 15 20

Baseline 20 40 60 80 50 0 0 0 0 ** Isch 0.5 0 0 0 0 0 Baseline Control, ML35 Control, vehicle IL1B

n Isch 1 h = 5 in vehicle group, group, vehicle = 5 in Rep 1 Rep 1 e M h ** ) Relative mRNA expression (normalized to that of 18S rRNA) rRNA) 18S of that to (normalized expression mRNA ) Relative

L355 Rep 3 d h Rep 2 Rep 6 h Rep 12 d Rep 3 h

5 Rep 24 h d Rep 36 Rep 4 h

n.s. Rep 48 d h Rep 5 Rep 72 * h TNF d Rep 48h,ML355 Rep 48h,vehicle Rep 1

n Rep 6 h

= 5 monkeys per group with 24 images for each each for images 24 with group per = 5 monkeys Rep 2 w d Rep 7

** Rep 3 w Rep 4 d w Rep 8

w n d P Vehicle = 6 in ML355 group) in monkeys treated treated monkeys in group) ML355 = 6 in t < 0.05, ** < 0.05, -test. ( -test. c

n.s. AST in serum (U/L) ALT in serum (U/L) ** 20 40 60 80 Baseline Baseline 20 40 60 80 CCL2 0 0 0 0 0 0 0 0 0 Isch 0.5 Isch 0.5 0 Rep 2weeks d Isch 1 Isch 1 ) Representative images of H&E H&E of images ) Representative h h ** Rep 1 Rep 1 h h * Rep 3 P Rep 3 ** ML35 Vehicl

h h **

< 0.01, *** < 0.01, Rep 6 Rep 6 h h ** Rep 12

Rep 12 5 e ** * h * h Rep 24 Rep 24 ** * b h * h Rep 36

Rep 36 * ML355 , , n.s. h * h * n Rep 48 Rep 48 * CXCL5 *

= 5 in each group). group). each = 5 in h h Rep 72 Rep 72 h h Rep 1 Rep 1 h h ** Rep 2 Rep 2 s e l c i t r a P w w

* Rep 3 < 0.001 by < 0.001 Rep 3 w w ± Rep 4 Rep 4 w

s.e.m. w w P w < 0.05, < 0.05,

 © 2017 Nature America, Inc., part of Springer Nature. All rights reserved. our conclusion that the pathogenesis of hepatic IR injury is marked marked is injury IR hepatic of pathogenesis the that conclusion our reprogramming metabolic lipid exacerbate can in turn which pathways, transduction mation, oxidative stress and apoptosis by regulating relevant signaling inflam induce can profiles lipid of derangement metabolic the that previously reported been has It phase. ischemic the during pathway reprogramming of lipid metabolism in the liver and notably in the AA robust and early we specific, have to primates, uncovered nonhuman mice from ranging species multiple across demonstration functional and approach omics integrative an on based investigation systemic unbiased an using Here, uncertain. are approach this of benefits cal and have the clini in limited, models rodent been effects therapeutic stage reperfusion the at death cell and inflammation for development have on therapeutic inhibiting focused been directly response is the main culprit in liver destruction inflammatory reperfusion-induced the whereas liver, the by erated phase reperfusion the during inflammation and phase ischemic the dam in age cell involving events, separate apparently two on based is IR need clinical unmet an is injury for IR attenuating and cause Preventing surgery. major transplant liver of a failure the is and injury liver severe to lead can IR Hepatic DISCUSSION pathologies. IR-related to or organs different in ischemia among injury IR hepatic to specific be might 12-HETE blocking of effect protective the that suggesting ( or stroke (MI) infarction dial in of of the setting effect was ML355 no myocar observed significant by ( injury protection liver IR–induced hepatic long-term against ML355 and systematic found and examination 8i Fig. Supplementary 8e Fig. ( cohorts human in and mice in observed to was what similar insult, IR hepatic following expression ALOX12 liver in setting. clinical the in disease the resembles that injury IR hepatic In short, by using a 60-min ischemia, we developed a monkey model of normal levels in the 60-min ischemia model ( of were the restored parameters monkeys to the CT perfusion fusion, ( surgery IR hepatic after increase compensatory a showed perfusion arterial hepatic the and then slightly increased at 1 and 2 weeks after reperfusion, whereas ischemia 60-min following reperfusion after h 48 at reduced was greatly perfusion portal hepatic the that showed clearly CT perfusion ( resection liver to subjected individuals human in observed those matched closely reperfusion by followed liver the of lobe lateral left the of ischemia time course profiles of serum levels of ALT, AST and ALP after 60-min ing injury ( damage response that was too severe to be repaired by 4 weeks follow A 1 second a to leads that reprograming metabolic lipid by primarily the activation of inflammatory signaling pathways ( pathways signaling inflammatory of activation the inhibited largely and chemokines, and cytokines of release and tion produc the suppressed death, cell and dysfunction liver IR-induced blocked significantly ML355 with production 12-HETE inhibiting reperfusion after h ( 12 at supplementation ML355 to due increase secondary a with stage, fusion reper the during declined gradually then and stage ischemic the as early as level highest its reached ML355 of concentration serum the 0 In this nonhuman primate model, we observed a significant increase

s e l c i t r 28 , , f 2 ). After ML355 administration ( administration ML355 After ). 9 . It is also believed that ischemic insult can be readily tol readily be can insult ischemic that believed also is It . Supplementary Fig. Supplementary 8a Fig. 6a Fig. 33 , 3 4 Supplementary Fig. 8d Fig. Supplementary . These findings support our results here and and here results our support findings These . , b – and and l ). Moreover, we performed perfusion CT CT perfusion performed we Moreover, ). Supplementary Fig. 8a Fig. Supplementary Fig. 6g Fig. 2 7 Supplementary Fig. 8h Fig. Supplementary . The current paradigm of hepatic hepatic of paradigm current The . – c ). ). Notably, the peak values in the , h and Supplementary Fig. 8g Fig. Supplementary ). At 4 weeks after reper after At weeks ). 4 Supplementary Fig. 8d Supplementary Fig. 9 Fig. Supplementary 3 0 . . Thus, recent efforts Fig. 6 Fig. Supplementary Supplementary ). Furthermore, Furthermore, ). 31 Fig. 6c Fig. , f 3 ). However, ). 2 . However, . ). Notably, ). – e and and ), ), ), ). ------

effective as a therapeutic target in hepatic IR injury and not in heart heart in not and injury IR hepatic in target therapeutic a as effective HETE appears therefore, to is 12-HETE be tissue specific; specifically of 12- signature,function the is a common mulation,ischemic which accu succinate unlike Furthermore, response. inflammatory the of sustained throughout the entire process is of IRaxis and serves ALOX12–12-HETE as a the key driverin change dramatic the stress, dative promote oxito stage ischemic the during succinate of accumulation transient the to contrast in However,injury. IR of pathogenesis the reperfusion during change for(ROS) oxygen ofreactive induction species mitochondrial metabolic critical a as ischemia during accumulation succinate fied identi A previous report reperfusion. during mation and injury liver liver of 12-HETE, which acts as a key initiator of the subsequent ischemic inflamthe in accumulation and hepatocytes in ALOX12induction intervention. therapeutic of targeted mechanism for a molecular but as inflammation, might the serves also basis reperfusion-induced for precursor necessary a is ischemia that suggesting mechanism lar cellu unknown previously a provides only not insight This damage. inflammation liver and death cell subsequent avoid to uncontrollable stage reperfusion the of during source the off cut effectively and directly can stage ischemic the of beginning the at reprogramming profile lipid the targeting Consequently, inflammation. on effect ary liver while calling for the development of more specific and effective effective and specific more of development the for calling while liver as a key metabolic signaling axis in the pathogenesis of IR injury in the conclusion, our data delineate an ALOX12–12-HETE–GPR31 pathway hepatic IR injury but also that for other liver inflammatory diseases. In for therapy benefit may only not axis this targeting response, matory Furthermore, as 12-HETE–GPR31 signaling promotes a potent inflam gability drug their increases binding small-molecule for pockets molecular types cell all almost of and relevance their relatively in distribution extensive pathological Administration (FDA)-approved drugs are targeted to GPCRs because members promising. The human genome more encodes than 800 GPCR family injury liver of tion induces a massive loss of hepatocytes that exacerbates the severity leukocyte recruitment and activation proinflammatory cytokines and chemokines, as well as the subsequent downstream NF- as acts a for and activating ligand 12-HETE GPR31, occurs 12-HETE of upregulation process, IR hepatic the during that, found we Here IR in is hepatic of largely unknown. ticipation the superfamily GPCR role in cancer progression and KRAS activation limited attention, with only a few recent reports regarding its potential very received has role functional its receptor, 12-HETE a as covered dis was GPR31 Since IR process. hepatic the during ates its function and injury for prognosis after liver transplantation or resection. 12-HETE content might act as crucial biomarkers in evaluation of liver not explored here, the upregulated ALOX12 expression and increased preferred to targeting a pathway in the later reperfusion stage. Although the ALOX12–12-HETE axis during the early stage of ischemia is likely injury,blocking IR hepatic initiates stage ischemic the during HETE tion. Furthermore, considering that the dramatic accumulation of 12- adding to of the importance fortransla finding this clinical potential macaques, and pigs including animals, large across conserved injury were IR liver of model mouse a in induction ALOX12–12-HETE inhibiting by produced effects beneficial The (stroke). brain or (MI) The early changes in metabolism during the ischemic stage involve The early ischemic during the changes in metabolism The GPCR member GPR31 directly recognizes 12-HETE and medi 43 , 4 4 0 4 . Notably, approximately 20–30% of US Food and Drug Drug and Food US of 20–30% approximately Notably, . . Thus, GRP31 is a potential drug target in its own right. right. own its in target drug potential a is GRP31 Thus, . 2 8 advance online publication online advance . We believe that this observation is translationally translationally is observation this that believe We . κ B and MAPK cascades to promote the production of 3 5 41 , highlighting the importance of metabolism in in metabolism of importance the highlighting , , 4 2 . Furthermore, the fact that GPCRs possess possess GPCRs that fact the Furthermore, . 22 , 38 , 3 9 . . This excessive inflamma 36

, 3 7 nature medicine nature . The potential par ------© 2017 Nature America, Inc., part of Springer Nature. All rights reserved. 12. 11. 10. 9. 8. 7. 6. 5. 4. 3. 2. 1. claims jurisdictional in published maps and affiliations. institutional r Reprints and permissions information is available online at at online available is information permissions and Reprints The authors declare no competing interests.financial experiments, designed wrote the manuscript the andstudy. supervised Z.H. and Z.-G.S. helped design the project and edited the manuscript; and H.L. monkey surgeries; L.W. clinical humancollected liver and samples;serum X.-L.M., performed staining experiments; R.-F.T. assisted in the performance of pig and animal hepatic IR models; X.-Y.Z. performed western blot experiments; Y.Z. liquid chromatography–mass experiments; S.T. spectrometry established the CT performed perfusion experiments; J.-Y.W. performed ultra-high-performance Y.-X.J. performed omics analyses and provided important advice for this study; R.L. analyzed data; L.-J.S. analyzed data and organized figures; P.Z., P.-X.W. and and edited the manuscript; W.W. and Z.-Y.L. performed biological experiments and data and wrote the manuscript; X.W. performed animal experiments, analyzed data X.-J.Z., X.C., Z.-Z.Y. and J.F. anddesigned performed the experiments, analyzed the Foundation (no. 91639304; H.L. and Z.-G.S.). Z.-G.S.) and the Key Collaborative Project of the National Natural Science the (Z.-G.S.)), National Natural FoundationScience of China (no. 81770053; Development Program (no. 2013YQ030923-05 (H.L.) and 2016YFF0101504 (no. 2014BAI02B01 and 2015BAI08B01; H.L.), the National Key Research and and 81630011; H.L.), the National andScience Technology Support Project H.L.), the Key Project of the National Natural FoundationScience (no. 81330005 from the National FundScience for Distinguished Young Scholars (no. 81425005; metabolites in the liver, and serum cell lysates. This work was supported by grants We thank Shanghai Metabolome Institute–Wuhan for their help in examining AA o Note: Any Supplementary Information and Source Data files are available in the t the in available are references, and codes accession statements of including Methods, and data availability any associated M injury.IR for therapy hepatic in used be potentially could that inhibitors GPR31 and/or ALOX12 nature medicine nature COMPETING FINANCIAL INTERESTS COMPETING AUTH A e h n c p

l ethods e i r

kno n Powell, W.S. & Rokach, J. Biosynthesis, biological effects, and receptors of of Un, K. receptors and effects, biological Biosynthesis, J. Rokach, novel & potential W.S. a Powell, as 12-lipoxygenase Targeting M. Holinstat, & B.E. Tourdot, new pathophysiology; in lipoxygenases of role The T. Okuyama, & R. Mashima, and ischemia Hepatic J. Gracia-Sancho, & M.B. Jiménez-Castro, C., Peralta, to reperfusion and mechanism ischemia Liver J.W. reperfusion—from Kupiec-Weglinski, & R.W. Busuttil, and Y., Zhai, Ischemia T. Eckle, & H.K. Eltzschig, P.A. Clavien, J.H. Wang, Y. Uchida, Vollmar, B. & Menger, M.D. The hepatic microcirculation: mechanistic contributions J.W. Kupiec-Weglinski, & R.W. Busuttil, J.C., Hong, H., Petrowsky, Y., Zhai, ultrasound-responsive and mannose-modified lipoplexes inhibits acute hepatic acute inhibits inflammation. lipoplexes mannose-modified and ultrasound-responsive acid. arachidonic (oxo-ETEs) from derived acids oxoeicosatetraenoic and (HETEs) acids hydroxyeicosatetraenoic therapy. antiplatelet perspectives. future and insights (2013). reperfusion injury: effects on the liver sinusoidal milieu. inflammation. tissue immune-mediated innate-adaptive of mechanisms into insights new injury: translation. Surg. undergoing major liver resection with versus without ischemic preconditioning. mice. of livers in injury 139 manner.TLR4-dependent a in mice in injury ischemia/reperfusion repair.(2009). and injury liver in targets therapeutic and Hepatol. Gastroenterol. Rev. bedside. to bench transplantation—from liver in injury Ischaemia–reperfusion

i p n e t

a s v O p / e , 2195–2206 (2010). 2195–2206 , i r n

wl R R C e s 238 d i r et al. o e . n x ed ON .

, 843–850, discussion 851–852 (2003). 851–852 discussion 843–850, , o h Efficient suppression of murine intracellular adhesion molecule-1 using f Nat. Med. Nat. t al. et t t al. et

m gm t TRIBUTI t al. et h Am. J. Transplant.J. Am. Hepatology l e . . Publisher’s with neutral Nature remains regard to note: Springer

p en uohg sprse aedpnet shma n reperfusion and ischemia age-dependent suppresses Autophagy a -el muolbln ui- dtrie svrt o liver of severity determines mucin-3 immunoglobulin T-cell popcie admzd td i 10 osctv patients consecutive 100 in study randomized prospective A p

e ts Trends Pharmacol. Sci. TrendsPharmacol. r advance online publication online advance . 17 ON Gastroenterology

, 1391–1401 (2011). 1391–1401 , 56 S

10 , 259–269 (2012). 259–269 , , 79–89 (2013). 79–89 , Biochim. Biophys. Acta Biophys. Biochim.

Redox Biol. Redox 11 , 1563–1569 (2011). 1563–1569 ,

141

38

, 2188–2199 (2011). 2188–2199 , 6 , 297–310 (2015). 297–310 , , 1006–1015 (2017). 1006–1015 , hso. Rev. Physiol. J. Hepatol.

1851 h t t o p , 340–355 (2015). 340–355 , :

n / / l w

i

89 n w Gastroenterology 59 w e 1269–1339 , , 1094–1106

. v n a e t r u s

r i

e

o . c n Ann. Nat. o

m o f /

44. 43. 42. 41. 40. 39. 38. 37. 36. 35. 23. 22. 21. 20. 19. 18. 17. 16. 15. 14. 13. 34. 33. 32. 31. 30. 29. 28. 27. 26. 25. 24.

in, J. Xiang, ligands of pathways and kinetics Binding R. Seifert, & H.J. Wittmann, Strasser,A., Rask-Andersen, M., Almén, M.S. & Schiöth, H.B. Trends in the exploitation of novel molecular GPCR illuminate ligands How B.L. Roth, & R.C. Stevens, D., Wacker, R.C. Stevens, p38 of Role L.H. Toledo-Pereyra, & F.A. Rivera-Chavez, A.H., Toledo, L.A., King, NF- of years 30 D. Baltimore, & M.J. Lenardo, Q., Zhang, K.V.Honn, Fehrenbacher, N. E.T. Chouchani, Garcia, V.Garcia, Y.Guo, appetitecontrol,insulininvolvedsensorin acid fatty a FFA4(GPR120)D.S.as Im, J. Lu, T.W. FFA4/ Tobin,A.B. & R. Prihandoko, Schwartz, B.D., Hudson, E., & Alvarez-Curto, G., Milligan, S.A. Hjorth, O., Rudenko, M., Trauelsen, A.S., Husted, D.K. Luci, in ADP to sensitivity platelet Increased C.D. Funk, & L.F. Brass, E.N., Johnson, P.,Songia,B., Porro, Tremoli,Squellerio,I., V.Cavalca, & E. Analysis, physiological K. Tang, A. Moles, Shimano, H. & Sato, R. SREBP-regulated lipid metabolism: convergent physiology convergent metabolism: lipid SREBP-regulated R. Sato, & H. Shimano, against Corcoran, S.E. & O’Neill, strategies L.A. HIF1 Protective P.A. Clavien, & R. Graf, H., Rudiger, N., Selzner, ischemia liver of mechanisms Molecular B.R. Davidson, & B.J. Fuller, response G., Datta, immune sterile The M. Heger, & T.M. Gulik, van R.F., Golen, van Ju, C., Colgan, S.P. & Eltzschig, H.K. Hypoxia-inducible factors as molecular targets in signals death cell Decoding G. Kroemer, ischaemic & O. Kepp, L., of Galluzzi, C., Brenner, evaluation An B.R. Davidson, & B.J. Fuller, F.P., Robertson, transplantation the solve pig the Will D.H. Sachs, & B. Gollackner, D.K., Cooper, A.M. Gregus, D. Luci, GPCRs. GPCRs. to targets. drug pharmacology. function. and structure GPCR human (2009). 763–770 reperfusion. and ischemia liver in JNK and pathobiology. human to relevance progression. cancer acritical prostate plays in role and cancer in prostate up-regulated significantly is receptor, acid KRAS. of association ROS. mitochondrial through injury n (2017). hypertension. trigger and function vascular affect to (2011). receptor for 12-( (2017). sensitivityand inflammation regulation. GPR40. receptor acid (2017). 809–821 opportunities. therapeutic and pharmacology GPR120: metabolites. of (2017). signaling GPCR-mediated 12-lipoxygenase. of (2014). inhibitors selective potent as and derivatives 4-((2-hydroxy-3-methoxybenzyl)amino)benzenesulfonamide (1998). 3105 12-lipoxygenase. platelet-type lacking mice product. 12-lipoxygenase platelet-derived neglected a 12-HETE: of significance clinical and partner. a seeks 60 mouse. the in injury liver chronic and acute in recruitment neutrophil dvret pathophysiology. divergent — Invest. Clin. J. liver. the of injury ischemic 19 reperfusion injury: insights from transgenic knockout models. (2012). ischemia/reperfusion. hepatic during diseases. liver for inflammation. liver transplantation. and preconditioning as a method of reducing ischaemia reperfusion injury in liver surgery backlog? J. FASEB critical role for spinal eLOX3 hepoxilin synthase activity in inflammatory hyperalgesia. 2010). Health, of Institutes in lipoxygenase. r e , 782–791 (2014). 782–791 , , 1683–1698 (2013). 1683–1698 , n d o et al. et . 2 et al. et Trends Pharmacol. Sci. TrendsPharmacol. 0 et al. et J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. Life Technol.Biomed.Analyt. Chromatogr.B J.

Annu. Rev. Med. Rev. Annu. t al. et t al. et 27 1 et al. et Trends Pharmacol. Sci. TrendsPharmacol. et al. et Structural basis for the cooperative allosteric activation of the free fatty free the of activation allosteric cooperative the for basis Structural 7 t al. et , 1939–1949 (2013). 1939–1949 , . t al. et Identification of the orphan G protein–coupled receptor GPR31 as a as GPR31 receptor protein–coupled G orphan the of Identification Nat. Rev. Drug Discov. Drug Rev. Nat. 9 et al. et Discovery of ML355, a potent and selective inhibitor of human 12- human of inhibitor selective and potent a ML355, of Discovery

et al. et Cell 1 ovrec o ecsni ad nern ilg: 12-lipoxygenase biology: integrin and eicosanoid of Convergence 126 20-HETE signals through G-protein-coupled receptor GPR75 (Gq) GPR75 receptor G-protein-coupled through signals 20-HETE ucsfl taeis o eemn hg-eouin tutrs of structures high-resolution determine to strategies Successful S 12-HETER1/GPR31, a high-affinity 12( high-affinity a 12-HETER1/GPR31, (2017). Probe Reports from the NIH Molecular Libraries Program Libraries Molecular NIH the from Reports Probe Mol. Cancer Mol. t al. et )-hydroxyeicosatetraenoic acid. TR/109CC- sgaig ewr i ncsay for necessary is network signaling TLR2/S100A9/CXCL-2 A et al. J. Mol. Med. (Berl.) Med. Mol. J.

Systematic analysis of rat 12/15-lipoxygenase enzymes reveals enzymes 12/15-lipoxygenase rat of analysis Systematic Synthesis and structure–activity relationship studies of of studies relationship structure–activity and Synthesis J. Hepatol. J. 170 , 3699–3707 (2016). 3699–3707 , The GPCR Network: a large-scale collaboration to determine determine to collaboration large-scale a Network: GPCR The J. Clin. Med. Clin. J. J. Cell Biol. Cell J. shei acmlto o sciae otos reperfusion controls succinate of accumulation Ischaemic The G protein–coupled receptor GPR31 promotes membrane Nat. Struct. Mol. Biol. Mol. Struct. Nat. , 414–427 (2017). 414–427 ,

53 Gastroenterology

14 , 133–147 (2002). 133–147 ,

59

a. e. Endocrinol. Rev. Nat. 37 , 111 (2015). 111 ,

, 583–594 (2013). 583–594 ,

Cell α 216 6 , 1055–1069 (2016). 1055–1069 , 38 Nature and metabolic reprogramming in inflammation.

, E69 (2017). E69 , FASEB J. FASEB 10 , 717–732 (2017). 717–732 ,

Mol.Aspects Med. Nat. Rev. Drug Discov. Drug Rev. Nat. , 2329–2338 (2017). 2329–2338 , 168 94 yoie rwh atr Rev. Factor Growth Cytokine , 579–590 (2011). 579–590 ,

, 613–627 (2016). 613–627 , 515 , 37–57 (2017). 37–57 , Proc. Natl. Acad. Sci. USA Sci. Acad. Natl. Proc.

. eaoiir Pnra. Surg. Pancreat. Hepatobiliary J.

125 24 30 , 431–435 (2014). 431–435 , J. Biol. Chem. , 570–577 (2017). 570–577 , , 2360–2369 (2016). 2360–2369 , . e. Chem. Med. J. , 917–936 (2003). 917–936 , el Metab. Cell Circ. Res. Circ.

Trends Pharmacol. Sci. TrendsPharmacol. S0098-2997(17)30064-X S h )-hydroxyeicosatetraenoic t t s e l c i t r a p World J. Gastroenterol. :

/ κ 964

/ 286 d 12 : bosmn of blossoming a B:

x

120 . , 25–34 (2013). 25–34 , d , 26–40 (2014). 26–40 ,

, 33832–33840 25 o 57 i . , 1776–1788 , o , 777–796 777–796 ,

r , 495–506 495–506 ,

23 g J. Hepatol. J. 95 / (National 1 , 69–84 69–84 , 0 , 3100– , . 1 0

16 38 3 11 8 , , /

© 2017 Nature America, Inc., part of Springer Nature. All rights reserved. subsequent GO analysis. the for selected were signal enrichment set gene the to most contributed that than 0.25 were considered statistically significant, and the leading-edge subsets nominal with sets Gene type. tion with the ‘Signal2Noise’ metric to generate a ranked list and a ‘gene set’ permuta gene set, and GSEA was implemented on the Java GSEA (version 2.2.4) platform the number of the same genes selected assigned to a given pathway. annotation,respectively, KEGG and where using calculated the following formula significantly enriched. The hypergeometric Only pathways with a corresponding and the annotation of all genes in the selected genome was used as the reference. Perlscript, Fisher’sin-house our using with test performed was exact analysis pathways was obtained from the KEGG database. A KEGG pathway enrichment analyses. Kyoto Encyclopedia of Genes and Genomes pathway and gene set enrichment specific in pathways KEGG enriched 4.9 package, yielding six mation method and then imported into the multiple experiment viewer (MeV) sion levels of replicates were using biological the normalized expres gene average The patterns. expression gene stage-specific the explore maps. 2D and 1D in data expression gene the of profile the visualize better to employed also ‘Rtsne’ was package R The algorithm. UPGMA) mean, arithmetic with method group pair (unweighted distance average unweighted an using samples of tree phylogenetic a struct analysis. Clustering protein quantification. Unique peptides with a 1% peptide false discovery rate (FDR) were reserved for able ofmodifications with oxidationmethionine and of with acetyl. N termini TMT with and 6-plex carbamidomethylation;mini with of cysteines and vari ter N peptide and oflysines modification fixed cleavages; missed 2 digestion; trypsin Da; 0.1 tolerance fragment p.p.m.; 20 tolerance precursor follows: as were integratedserver Mascot the in used parameters main The 1.4). (version the UniProt mouse protein using sequences the Proteome Discoverer software tem. The raw data obtained from the proteomics analysis were searched sys against HF Exactive Q Scientific Fisher Thermo the using experiments MS/MS) and used to conduct liquid chromatography–tandem (LC– mass spectrometry each stage of hepatic IR injury were with labeled six tandem mass tags (TMTs) spectrometry. mass and preparation Protein corresponding (ii) and 2 than larger change fold a (i) including criteria, selection two satisfy to had comparisons pairwise the from defined genes expressed Differentially algorithm. Benjamini–Hochberg and groups, sample different the between Student’s method (RPKM) reads mapped million per kilobase per reads the using (ref. 1,000 to 0 from ranged size insert the mismatches; two no more than allowing 2.21), (version (SOAP2) 2 package analysis oligonucleotide short using mm10 UCSC from mRNA sequences to reference mapped then were million reads clean All 10 reads. clean approximately in resulting data, raw the from filtered first were scores quality sequencing low showing bases the of half or bases non-ATCG 5% than more sequences, adaptor Illumina with Reads bp. a 49 of with length 4000 read HiSeq Illumina an using sequenced were bp) 150 of size insert expression. gene Digital ONLINE METHODS nature medicine nature Each Each known KEGG pathway or integratedspecific gene list was as defined a N and The annotation information regarding genes involved in biological biological in involved genes regarding information annotation The t hypergeome -test algorithm was used to identify differentially expressed genes genes expressed differentially identify to used was algorithm -test n are the number of all genes and the number of selected genes with A hierarchical clustering analysis was performed to con was performed analysis clustering A hierarchical 4 5 tric ). Normalized gene expression values were calculated calculated were values expression gene Normalized ). k -mean -mean clusters expression with profiles. specific The For the DGE assay, single-end libraries (with an an (with libraries single-end assay, DGE the For Z- sc or e P = values less than 0.05 and FDR values less less values FDR and 0.05 than less values M P k value less than 0.05 were considered as and n -means clusters were then identified. then were clusters -means   M N k Z P m -means clustering was applied to to applied was clustering -means -scores of these enrichments were   values were corrected using the the using corrected were values are the number of all genes and genes ofall number are the   n m The liver tissue samples from samples tissue liver The 1 1 − P − values less than 0.05. than less values M N ) (   M N    − N n Z − -score -score transfor − 1 1    4 6 . A . ------replicates were averaged as WGCNA input. A soft power threshold of 22 was was 22 of threshold power soft WGCNA A input. as averaged were replicates network was constructed using WGCNA (version 1.51) construction. network coexpression Weighted gene Institute Genomics (BGI). Beijing the at performed were analyses RNA-seq and DGE The gene. detected each for value FPKM the estimate to applied was 2.2.1) (version cufflinks Finally, reads mapped unique the of alignments the store to files BAM produce to respectively, using HISAT2 (version 2.1.0) genomes, (hg38) human and (mm10) mouse the to aligned were L02 line cell liver human and hepatocytes mouse cultured primary the from reads clean reads, low-quality removing bp.After 150 of length read a and bp 250 of size insert an with 4000 HiSeq Illumina using hepatocytes in conducted was seq preparation library RNA-seq 0.69) (version Circos and the coefficients showing layout correlation circular A R. in ‘cor.test’ function the using coefficient correlation Pearson the calculate to applied was abundance protein or sion expres mRNA the of change fold the and levels, protein and mRNA the at groups sham and ischemia between profiles distinct the of conservation estimate the to used was coefficient correlation Pearson the pathway, each For pathways. KEGG to assigned were data abundance protein proteomics RPKM) and (nonzero data expression mRNA the both by detected genes 5,067 Circos. using visualization data and analysis omics Integrative and Gpr31b-P2 were in the deletion to the used verify presence of multiple highly homologous transposons. The primers Gpr31b specific target sites shown in procedures to similar to used those generate the The product. deletion the identify to used were Alox12-P2 and Alox12-P1 primers The tails. from isolated DNA total byassay using PCR and confirmed were identified mice generated of the Alox12 mRNA encoding the Cas9 nuclease and the short guide RNA (sgRNA) targeting (ssOligonucleotide) primers were annealed and cloned into pUC57 vectors. The site in the conserved exon 5 was selected, and two single-strand oligonucleotide out allele for Generation of genetically modified mice. ment were avoided in our study. Academy of Sciences and the National Institutes of Health. Abuse and maltreat National the by published Animals UseLaboratory andof Care the for Guide Animals of Wuhan University. Animals received humane care according to the and the Institutional Animal Care and Use Committee of the Institute of Model by the Animal Care and Use Committee of Renmin HospitalAnimals and treatment.of Wuhan University processes. logical bio indicated the to contributed weights) edge of 5% heaviest the in (those neighbors connected most the whether determine to performed was analysis weight to An enrichment terms. process biological to corresponding according and were assigned sorted were edges network, coexpression constructed the in genes leading the of 0.02) > weight edge with (those edges and genes neighboring the integrating After function. biological and expression gene ( tions condi hypoxia and normoxia under treatment 12-HETE following analysis RNA-seq the from genes leading-edge highlighted of weights edge network on with a were the of basis evaluated GO performed the analysis coexpression significantly associated with the measured traits. that eigengene,were modules correlation a to identify analysisperformed was (power, 22; minModuleSize, 30; mergeCutHeight, 0.25). Based on each module function ‘blockwiseModules’ with the default settings and modified parameters ‘pickSoftThreshold’ operation. wereusing Coexpression modules the detected functions the from obtained index fitting model regression linear the of basis selected to ensure the network satisfied a topology ( scale-free Correlations between leading genes and 12-HETE function in hepatic injury Fig. 5 Fig. were injected into zygotes from a C57BL/6 background. The genotypes a Alox12 ). A heavier edge weight indicates a closer relationship between between relationship a closer indicates weight edge A heavier ). ( Alox12 4 7 . All animal protocols used in this study were approved -KO), a Supplementary Table 1 P values of the pathways was generated using using generated was pathways the of values , sequencing and analysis. and sequencing Streptococcus pyogenes Gpr31b To generate mice with a global knock 4 8 . SAMtools (version 1.4.1) was used -KO mice were generated using using generated were mice -KO Alox12 were chosen because of the 5 The gene coexpression coexpression gene The 0 Gpr31b . All available biological -KO mice, and the two Cas9 (SpCas9) target doi: Paired-end RNA- Paired-end 10.1038/nm.4451 R gene. 2 > 0.90) on the A total of of total A - P1 P1 4 9 - - - - - . © 2017 Nature America, Inc., part of Springer Nature. All rights reserved. the corresponding secondary antibodies for 1 h at room temperature. Signals temperature.Signals room at h 1 for antibodies secondary corresponding the with incubation by followed was which °C, 4 at overnight antibodies primary membranes were blocked with 5% skim milk and incubated with the indicated to PVDF membranes (IPVH00010; Merck Millipore, Darmstadt, Germany). The protein samples were loaded and separated on SDS–PAGE gels and transferred BCA Protein Assay Kit (catalog no. 23225; Thermo Fisher Scientific). The total inhibitor tablets (catalog no. 4906837001; Roche) and quantified using a Pierce phosphatase and Roche) 04693132001; no. (catalog tablets protease cocktail inhibitor with EDTA) mM 1 SDS, 0.1% deoxycholate, sodium 0.5% NP-40, 1% NaCl, mM Tris-HCl,150 mM RIPA7.4, (50 using pH samples buffer lysis as previously described Westernantibodies. and analysis blot in are listed study this in used pairs primer The sion levels of the target genes were to of normalized those (95 °C for 10 s, 60 °C for 10 s and 72 °C for 20 s; The 45 mRNA cycles). expres 1 Green SYBR Master 480 Mix following no.Roche) a (catalog standard procedure 04887352001; LightCycler by quantified were products amplification PCR The manufacturer’sinstructions. the following Roche) 4896866001; no. (catalog Kit Synthesis cDNA Strand First Transcriptor the using cDNA into Madison, WI, USA) and gel electrophoresis. 2 fied and checked for quality using a Nanodrop 2000 (Thermo Fisher Scientific, no.quanti was RNA isolated The USA). (catalog IN, Indianapolis, Roche, Reagent 11667165001; Isolation TriPure using cells cultured or samples sue assay. qPCR in provided is regarding individuals these information Clinical Board. Review Fourth University the Medical of Military Hospital Xijing the and Board Review University Wuhan of Hospital approved were Renmin by the samples human involving procedures per were studies the formed according to the and principles outlined by the Declaration of Helsinki. All objective, experimental the achieve to only used were samples All families. their or donors Informed all by samples. signed liver were forms the consent as points time same the at obtained were reperfusion ples before but ischemia sam serum The group). (reperfusion reperfusion after after and group) (ischemia group), (baseline points: ischemia time three before at individuals the from tissue liver cyst-free and tumor- from collected were samples Liver cysts. hepatic or carcinoma hepatocellular to owing surgery resection liver underwent who patients from obtained were samples. Human weight. A blank solution was used as the vehicle control. 30 min before at ischemia vein and tail at 6 the h via after reperfusionmice into at a injected dose was of 3 5:10:20:65) mg per kg Solutol:PEG400:water; body USA) or (ab141560; Abcam,CDC Cambridge, MA, USA) in solution (DMSO: NJ,MonmouthJunction,(HY-12341; water. MCE, and ML355 food to access with cycle) dark 12-h and light 12-h °C, (20–26 environment free served as sham controls. Mice were maintained in a standard specific-pathogen- Mice that underwent the same surgical procedure without vasculature occlusion were euthanized to collect liver samples and serum for subsequent examination. mice reperfusion, after point time At indicated removedeach for reperfusion. by bleaching of the ischemic liver lobes. After 60 min of ischemia, the clamp was North Vancouver, BC, Canada). A successful ischemia operation was indicated Tools, Science (Fine clamp microvascular atraumatic an using occluded then performed to expose the liver. The blood supply to the left and median lobes was described previously as injury IR hepatic warm to 70% were g) subjected (weight, 24–27 model. ischemia–reperfusion hepatic Mouse in are listed generation for mouse used Alox12 the identify to used primers The of mice. genotype the confirm to performed were sequencing and assay PCR A zygotes. into injected and linearized was plasmid obtained The promoter. hepatocyte-specific a of control the under of sequence coding full-length the doi: Alox12 10.1038/nm.4451 insertion were Alb-seq-F and Alb-seq-R. The primers and target sites sites target and primers The Alb-seq-R. and Alb-seq-F were insertion 51 -HTG mice were generated using homologous recombination, and and recombination, homologous using generated were mice -HTG , 5 2 For the qPCR assay, total mRNA was extracted from liver tis liver from extracted was mRNA total assay, qPCR the For . Briefly, after the mice were anesthetized, midline laparotomy was The human serum and liver samples used in this study study this in used samples liver and serum human The 5 3 . . In brief, total protein was isolated from tissue or cell Supplementary Table 2 Supplementary Alox12 Western were blot performed analyses Supplementary Table 1 Supplementary was cloned into an albumin vector vector albumin an into cloned was µ g g RNA was reverse-transcribed Male mice aged 8–10 weeks weeks 8–10 aged mice Male Supplementary Table 3 Supplementary . ACTB or 18S rRNA. . ad libitum ad . - - - - -

of 1:1,000, and antibodies weresecondary used at a 1:5,000 dilution. Nagoya, Japan). Unless specified, all primary antibodies were used no.at M185-3L; (catalog MBL from obtained was the antibody Flag The USA). dilution ME, 115-035-003), which were purchased from the Jackson Laboratory (Bar Harbor, no. (catalog (H+L) and goat 111-035-003) anti-mouse-IgG no. (catalog (H+L) goatanti-rabbit-IgG Peroxidase studyincluded AffiniPure this in used bodies anti Secondary Biotechnology. Cruz Santa from obtained were (sc-390752) and CYP4A (ab3573) were obtained from Abcam. Antibodies against GPR120 (ab75581) GPR75 (ab211049), GPR40 (ab75579), GPR31 against Antibodies Biotechnology. Cruz Santa from purchased was (sc-133085) ALOX15 against Antibody USA). MO, Louis, St. (SAB2100109; Sigma and USA) TX, Dallas, Antibodies against ALOX12 were obtained from Santa Cruz (sc-365194, 1:200; no. (catalog no. no. and GAPDH (catalog (catalog 2832) CPLA2 2118). 3289), no. (catalog p-PKC 9664), no. (catalog PKC 2055), no. (catalog ALOX5 2058), (catalog no. (catalog no. BCL2 9252), BAX3498), (catalog no. c-CASP32772), ERK1–ERK2 (catalog no. 4695), p-JNK1–JNK2 (catalog no. 4668), JNK1–JNK2 4370), no. (catalog p-p38 p-ERK1–ERK2 9212), no. (catalog p38 4764), 4511), no. (catalog no. (catalog p65 3033), no. (catalog p-p65 4814), no. (catalog p-I USA): MA, (Beverly,Technology Signaling following proteins againstthe Cell fromwere purchased Antibodies facturers. blots are shown in (Bio-Rad). GAPDH was used as a loading control. Uncropped blots for western Bio-Rad, Hercules, CA, USA) and captured (170-5061; reagents by a (ECL) ChemiDoc MP chemiluminescence Imaging enhanced Systemby visualized were of 0.005% formic acid in water and a mobile phase B of 0.005% formic acid in formicacid 0.005%of B mobilephasewater a and in formic acid 0.005%of 45 °C. Metabolites were isolated using a biphasic system with a mobile phase A at maintained wastemperature column the and examination, molecule small An Agilent ZORBAX Eclipse Plus column (2.1 × 100 mm, 1.8 sample preparation. Samples were stored at −80 °C until analyses. supernatant was collected and dried following the procedure 12,000 forat liver centrifuging and After serum ice. on sonication by 600 in times three freeze-thawed and under dried N 30 s and centrifuged at 12,000 150 samples, serum and dried under nitrogen gas (N methanol aqueous 15% a solution with extracted supernatantThe further was 600 in homogenized and tube For liver samples, 50 mg of liver tissue was added to a polypropylene extraction (UPLC–MS; 1290-6470 UHPLC–MS/MS system) as previously described ined using ultra-high-performance liquid chromatography–mass spectrometry liver, serum and cell lysates, as wellSample preparation asfor UPLC–MS theanalysis. ML355 concentration in serum, was exam captured under a light microscope (Olympus, Tokyo, Japan). Zhongshan Biotech, Beijing, China). The histological images were observed and Wuhan, China) and immunoreactive cells were visualized using DAB (ZLI-9032; with HRP-conjugated antibodies (A21020, 1:500 dilution;secondary Abbkine, incubated were sections the °C, 4 at antibodies primary the with incubation no. 551459, 1:25 dilution; BD Biosciences, San Jose, CA, USA). After overnight Abcam), CD68 (MCA1957; AbD Serotec, Raleigh, NC, USA) and LY6G (catalog tions of pigs was quantified using primary antibodies against CD11b (ab75476; antibodies, respectively. The infiltration of inflammatory cells into the liver sec primary Sigma) dilution; 1:50 rabbitanti-humanandALOX12 (SAB2100109, tions with mouse anti-human-ALOX12 1:50 dilution; (sc-365194, Santa Cruz) sec liver the incubating by determined was samples humanliver and mice in of mice subjected to the cerebral IR procedure. The ALOX12 expression profile ously described previ as performed was staining TTC The orheart. liver ofthe areas necrotic rehydration, standard out patternstaining the H&E was carried to in visualize The waxes were sectioned serially at 5- chemical staining were performed in paraffin-embedded liver or heart sections. staining. immunohistochemical and Histological All antibodies used were purchased from and validated by commercial manu For the UPLC–MS examination, samples were dissolved in pure acetonitrile. 2 gas. For cultured cell samples, the cells were countedcollected, 5 4 on brain sections (1 mm thick) to visualize the infract volume Supplementary Supplementary Figures 10 µ l cold methanol was added to 50 50 to added was methanol cold l g at 4 °C for 10 min. Supernatant was collected and µ 2 l methanol containing 6 6 containing methanol l ) ) using a Termovap Sample Concentrator. For µ m thickness. After deparaffinization and µ l of methanol. HETEs were extracted extracted were HETEs methanol. of l The content of AA metabolites in – 22 κ B . g α for 10 min, the separated separated the min, 10 for H&E and immunohisto and H&E (catalog no. 9246), I 9246), no. (catalog µ nature medicine nature l serum, vortexed for vortexed serum, l µ l BHT (0.10 mM). mM). (0.10 BHT l µ m) was used for 55– κ B 5 α 7 ------.

© 2017 Nature America, Inc., part of Springer Nature. All rights reserved. according to the manufacturer’s instructions. Tarrytown,NY, (Siemens, USA) analyzer System Chemistry 2400 ADVIA an assayed by measuring the concentrations of ALT, AST and ALP in serum using examination. function Liver in was 5 human for site target The screening. blasticidin and puromycin after the hepatocytes. with L02 Homozygous human into cotransfected plasmid and pST1374-N-NLS-Flag-Linker-Cas9 vector pGL3-U6-sgRNA-PGK-puromycin described viously pre as cells L02 in system SpCas9 dual-vector a using constructed was line GPR31 virus cabinet. a in biosafety were performed involving operation procedures All China). (Shanghai, Ltd. Co., Biotechnology mouse with fied by western blot analyses. The adenoviruses used to overexpress or interfere veri and puromycin screening after obtained were construct each expressing 10 with cells host was infect to used virus and harvested replication-defective The pMD2.G. and psPAX2 plasmids aging pack and lentivirus constructed the with cotransfected were cells HEK293T production, lentivirus the For control. the as used was GFP targeting shRNA described previously as vectors lentiviral constructed we expression, of knockdown stable with cells transfected infection. and production adenovirus and Lentivirus used for plasmid construction are listed in primers The sequencing. through verified were plasmids vector. All lentiviral constructed by inserting an shRNA targeting one of these genes into the pLKO.1 The knockdown plasmids for human were transcription, reverse and sequences cloned the into pcDNA5 the vector. (ORF) of the human gene ing protocols described in previous studies Plasmid construction. function. and tion produc 12-HETE on effects inhibitory its determine to medium cell the to (10 ML355 nM. 100 of dose a at USA) MI, were cells Ann Cayman, Arbor,(34570; treated with in 12-HETE hepatocytes, to mimic by followed reoxygenation of periods, 1% for concentration indicated oxygen CO 5% with incubator an in °C 37 at collagen tail rat with coated plates in FBS 10% with supplemented Invitrogen) 11965-092; no. (catalog DMEM in cultured were hepatocytes Primary Biosciences). BD Falcon, 352350; no. (catalog 70- a strainer through portal filtered and the minced through excised, then was liver Scientific) The vein. Fisher Thermo 17104-019; no. (catalog type IV collagenase 0.5% containing solution a administering by digested was previously described as method perfusion collagenase the in study.used this contamination was observed in the cells. No commonly mycoplasma misidentifiedNo cells were penicillin-streptomycin. 1% and Scientific), Fisher Thermo and China) maintained in DMEM containing 10% FBS no.(catalog 10099141; from the Type Culture Collection of the Chinese Academy of hepatocytes. primary and Sciences lines Cell (Shanghai, MassHunter software (version B.08.00) was used for the quantitative analysis. sure 30 psi, sheath gas temperature 350 °C and sheath gas flow 11 l/min. Agilent desolvation temperature 300 °C, desolvation gas kV,3.5 voltageflow capillary follows: as rate set wereparametersionization The mode. 10 ml/h, nebulizer pres multiple-reactionnegativemonitoring in detected were molecules Small min. acetonitrile. The injection volume was 3 nature medicine nature Supplementary Supplementary Table 4 Primary hepatocytes were isolated from male mice aged 6–8 weeks using using weeks 6–8 aged mice male from isolated were hepatocytes Primary ′ -TCCACACCCTGACCCGGAAC-3 koku cl ln generation. line cell -knockout in vivo Alox12 2 . The primary hepatocytes were challenged with hypoxia at an an at hypoxia with challenged were hepatocytes primary The . GPR31 hepatic IR injury. To examine role the functional of 12-HETE 6 0 in the primary hepatocytes were purchased from Hanbio from purchased were hepatocytes primary the in . sgRNA targeting human human targeting sgRNA . -knockout cells were obtained from a monoclonal line line monoclonal a from obtained were cells -knockout All plasmids used in this study were constructed follow GPR31 . The liver function of animals in this study was was study this in animals of function liver The was amplified from cDNA obtained by mRNA GPR31 HEK293T and L02 cells were purchased cells and L02 HEK293T µ 5 l, and the running rate was 0.6 ml per The human human The 9 Supplementary Supplementary Table 4 ′ . The full-length open reading frame GPR31 . . The sequencing primers are listed µ , GPR40 M; HY-12341, MCE) was added added was MCE) HY-12341, M; µ GPR31 g/ml polybrene. Cells stably stably Cells polybrene. g/ml , , GPR40 , GPR75 was cloned into the the into cloned was GPR31 To generate stably stably generate To , , 5 GPR75 8 and . Briefly, the liver liver the . Briefly, -knockout cell cell -knockout GPR120 or . GPR120 µ GPR31 5 m cell cell m 9 were . An An . ------

96T 96T for TNF, SEA087Mu 96T for and CCL2 SEA371Mu 96T for CXCL10; pig: CCL2 and CXCL10) in were by serum determined ELISA (mouse: SEA133Mu IL-6, (TNF), factor necrosis (tumor chemokines and cytokines inflammatory mediators. inflammatory of concentration Serum Images were obtained under a fluorescence microscope (Olympus DX51). according to the manufacturer’s instructions. The nuclei were labeled ApopTagwith Plus DAPI. antibody.ary to staining TUNEL evaluatewas performed liver using injury an (A-11011, 1:200 dilution; Thermo Fisher was Scientific) applied as the second anti-rabbit-IgG Goat overnight. °C 4 at antibodies primary Biosciences) BD dilution; Abcam) or rabbit anti-mouse-Ly6G (catalog no. 551459, 1:25 dilution; goat serum and then incubated with rabbit anti-mouse-CD11b (ab75476, 1:100 tration of inflammatory cells, mouse liver sections were first blocked with 10% of was The 5 thickness liverthe sections in paraffin. tion in 4% neutral paraformaldehyde for more than 48 h, followed by immunofluorescenceembedding and TUNEL staining assays were prepared through fixa staining. TUNEL and Immunofluorescence Esaote Inc.) with a 15-MHz probe, as reported by previous studies MyLabultrasoundusing30CVa (Biosoundsurgery after MI miceoffunction turer’s instructions. Echocardiographic assay was alsoby performedADVIA 2400to Chemistryexamine Systemheart analyzer (Siemens) according serum toconcentrations the manufacof creatine kinase (CK) and lactate dehydrogenase (LDH) Heartfunctionassay. sonal fruits, and had free access to water. The room lighting was maintained maintained was lighting room water. The to access free had and fruits, sonal in standard primate cages, fed a chowspecial for nonhuman primates and sea ties for at least 28 d before the operation. The monkeys were individually housed passed the local quarantine test. Monkeys were acclimated to andtheir health general good werein monkeys housing All wereincluded. kg 4–6 weighing facili Technology Co. Ltd. (Suizhou, China). Only male monkeys aged 4–6 years and this study were purchased from Hubei Topgene Laboratory Animal Science and model. ischemia–reperfusion hepatic Monkey recorded and are shown in were pigs of signs vital The conditions. sterile under performed was solution. Surgery vehicle of volume equal an received pigs control The reperfusion. after h 6 vein jugular internal the via delivered was dose same the and vein, portal the through ischemia before min 30 5:10:20:65) Solutol:PEG400:water; euthanized and liver samples were forcollected subsequent examination. and drug administration after the operation. After 48 h of reperfusion, pigs were central vein catheter was not removed fromcollection neck tothe allow blood the but closed, then was wall liver. abdominal the The to inflow restore the to of cannulas all ischemia, wereperiod withdrawn and the clamps were released reduction of microperfusion was considered adequate ischemia. After a 120-min Perimed, Ardmore, PA, US) before and at 10 min after clamping, and a 70–80% examined by a laser Doppler perfusion monitoring unit (PeriFlux System 5000; the bypass cannula was kept below 20 cm H flow was bypassed portal to the right jugular external vein, and the pressure in the congestion, To splanchnic avoid heparinization. after vein portal the and artery hepatic the clamping simultaneously by achieved was ischemia warm hepatic Total hilum. liver the expose to performed was laparotomy Midline at 1.5–3% after atracurium intravascular injection (0.15 mg per kg body weight). Ruicheng, Shangdong, China). Anesthesia was maintained(Kelong, atropine weight by body inhaled kg isoflurane per mg 0.02 of administration by followed II (0.15 ml per kg body weight; Nanjing Agricultural University, study.Nanjing, For the China),full hepatic IR surgery, pigs were first this sedated with Shumianning in used were kg) 18–30 (weight, months 3 aged Wuhan,China) Jiaxing, model. ischemia–reperfusion hepatic Pig instructions. Cloud-Clone Corp. (Wuhan, China) and used according to the manufacturer’s for IL-6 and SEA087Si 96T for CCL2). All commercial kits were purchased from SEA079Po 96T for IL-6 and SEA087Po 96T for SEA079Si 96T monkey: CCL2; ML355 was administered at a dose of 0.4 mg per kg body weight (DMSO: weight body kg per mg 0.4 of dose a at administered was ML355 InSitu Apoptosis Fluorescein Detection Kit (S7111; Millipore) Millipore) (S7111; Kit Detection Apoptosis Fluorescein The heart function of mice was evaluatedfunctionwasheartofmiceThe by measuring Supplementary Supplementary Table 5 2 O. Microcirculatory perfusion was ae oetc is (Wanqian pigs domestic Male The liver sections used for the the for used sections liver The All rhesus monkeys used in in used monkeys rhesus All . µ m. m. To examine infil the h cnetain of concentration The doi: 10.1038/nm.4451 61 , 6 2 . ------

© 2017 Nature America, Inc., part of Springer Nature. All rights reserved. using SPSS (version 21.0), and the data are expressed as the mean analysis.Statistical vein. A blank solution was injected as vehicle into control mice. kg weightbody at 30 min before and surgery after 6 h of reperfusion in the tail per mg 3 at ML355 with injected were mice stroke,on ML355 of function the To model. MI the in evaluate continuous d fora day 7 per weight body kg per monitored by Doppler analysis (Periflux System 5010; Perimed, Sweden). was surgery during flow blood cerebral regional The period. indicated the for flow blood restore to withdrawn was filament the ischemia, of min 45 After the internal carotid and artery blocked the origin of the middle cerebral artery. arteries were exposed. A 6-0 silicon-coated monofilament was then inserted into by 2.5–3% isoflurane in oxygen, the left common, internal and external carotid dle cerebral artery according to previous studies ligation. without LAD using a 7-0 silk suture. Mice in the sham group the underwent same procedure opened, and was was encircled ligated artery (LAD) descending the proximal coronary left pericardium the After space. intercostal fourth or third the at per kg body weight, intraperitoneally) and then subjected to a left thoracotomy previously as reported artery coronary left the ligating by established was model MI Mouse myocardial infarction model liver perfusion and the hepatic perfusion index were generated. andperfusion maps of hepatic arterial perfusion, hepatic portal perfusion, total transferred to the Philips IntelliSpace Portal workstation (version 06.0.3.12200), a scan interval of 3 s were performed. The perfusion CT data were collected and nously injected and was followed by ten noncontrast scans. A total of 39 scans at rate of 2 ml/s. The perfusion scan began when the contrast solution was intrave material at a dose of 2 ml per kg body weight, followed by a 15-ml 11-2; salineJiang flush Su atHeng a Rui Medicine Co., Ltd., Lianyungang, China) as the contrast examination was then performed using an iodixanol injection (catalogscans, 66; slice no.thickness, 5mm; rotation 92339- time, 0.5 s; delay, 2s). Aperfusion CT theentire abdomen was performed at 27effective mAs and 100 kV(number of were acquired in two orthogonal directions, a noncontrast helical scan covering line was established and connected to the contrastanesthetizedand fixed on material.a scanning table in a supine position. AfterAn intravenous localization scans MedicalSolutions Inc.).beforescanning,USA,monkeyminAtwasCTthe 30 using perfusion CT imaging using a Siemens SOMATOM Definition4 weeks64 (Siemensafter inducing hepatic IR injury, liver injuryPerfusion in the computedmonkeys wastomography evaluated imaging. and are shown in kg body weight Tolfedine injected intramuscularly. Vital signs were monitored lobes after 48 h of reperfusion. Postoperative pain was alleviated with 0.1 ml per median and lateral left the from collected weresamples Liver saphenousvein. blood blockage through the portal vein and at 6 h after reperfusion through the before min 30 administered was 5:10:20:65) DMSO:Solutol:PEG400:water; of inhibition on hepatic IR injury, ML355 (3 mg weightkg per body in a solution ALOX12of To closed. effect was the abdomen evaluate the reperfusion, After reperfusion. with restored and ischemia during 20% than less to reduced was Ischemia and reperfusion were considered successful when the microperfusion warm ischemia (30, 60 of or 120 period min), a the After clamp tightened. was was released clamp to the allow before reperfusion. tissues liver the damaging prevent to taken was Care heparinization. after clamp right-angle atraumatic an using occluded was lobe median the and lobe lateral left the mentbetween liga liver, the andthe expose to performed then laparotomyMedian was ess. proc surgery entire the during isoflurane inhaled 1–1.5% by maintained was Anesthesia relaxation. muscle induce to injection intravascular weight) body kg per mg (0.15 atracurium 0.02 an by followed and atropine, weight II body kg per Shumianning mg weight body kg per ml 0.1 of injection muscular tained at 22– 26 °C. on a 12-h dark–12-h light automatic cycle, and the room temperature was main doi: Mice in the ML355 group ML355 Mice in the vein at tail werein the ML355 with 3 mginjected The mouse stroke model was established by transiently occluding the left mid intra an with anesthetized were monkeys the surgery, IR hepatic the For 10.1038/nm.4451 6 3 . . Briefly, mice were anesthetized using sodium pentobarbital (50 mg Supplementary Supplementary Table 6 All statistical analyses used in this study were performed werestudy this in used analyses statistical All , stroke model and treatment. . At 48 h, 1 week, 2 weeks, 3 weeks and 6 4 . After mice were anesthetized ± s.e.m. unless The mouse ------64. 63. 62. 61. 60. 59. 58. 57. identifiers identifiers database the Archivewith Read Sequence the submittedto havebeen data ics included within the article and its additional files. The DGE and transcriptom availability. Data is available in the Life Sciences Reporting Summary. ments were performed in triplicate unless specified in the figure legends. were blinded to animal genotype and grouping information. All surgery groups or the vehicle and ML355 treatment groups. The experimenters genotypeandsimilarbaselinevalueswere randomly sham andtheassigned to collection and thus was excluded from statistical analysis. Animals with the same the group and other scholars and were fixed in a prospective manner. One mouse in set of animals were determined according to previous studies performed by our ment is provided in the corresponding figure legends. The sample sizes for each 0.05 were considered significant. Detailed statistical information for each experi similar among the compared groups. Statistical differences with performed only when the global null hypothesis was rejected. The variance was analysis when there was heteroscedasticity in the data. Pairwisea normal distribution comparisons and homogeneity of variance or usingwere Tamhane’s T2 (M) among more than two groups using a Bonferroni two-tailed Student’s otherwise indicated. Statistical differences between two groups were calculated by 56. 55. 54. 53. 52. 51. 50. 49. 48. 47. 46. 45. upon reasonable request. identifier proteomicsthe via data with the set repository identification (PRIDE) partner Consortium ProteomeXchange the to deposited been have data proteomics

u, S. Guo, Bao, M.W. Y.X. Ji, X. Jiang, O. Shalem, G.N. Zhao, P.X.Wang, Q.F. Zhu, aod, . Edn M, ola, . Sae, . Vle, .. Comprehensive D.A. Volmer, & D. Spaner, A., Koulman, M., Eiden, M., Masoodi, chromatography–tandem Liquid J.J. Moreno, & O. Jáuregui, R., Martin-Venegas, L. Li, P. Sun, J. Hu, X. Wang, Langfelder, P. & Horvath, S. WGCNA: an R package for weighted correlation network H. Li, low with aligner spliced fast a HISAT: S.L. Salzberg, & B. Langmead, D., Kim, M. Krzywinski, Mortazavi, A., Williams, B.A., McCue, K., Schaeffer, R. L. & Wold,Li, B. Mapping and quantifying J. Neurosci. J. ventricularleft remodellingdysfunction. and signalling. TAK1-dependent via hypertrophy hypertrophy. cardiac pathological of regulator Science lysosomal the Tlr4. of targeting degradation by monkeys and mice in disease liver fatty non-alcoholic (2017). 449 liquid high–performance primates. nonhuman and mice in steatohepatitis nonalcoholic ultra with coupled labeling spectrometry. chromatography/mass probe isotope stable 82 networks. regulatory complex in lipids bioactive of analysis lipidomics Sci. Life Technol.Biomed. Analyt. B Chromatogr. J. models. cell in compounds related and eicosanoids of analysis spectrometry mass rat. deficient 3 injury. injury. reperfusions Tak1suppression. analysis. 25 requirements. memory Res. Genome mammaliantranscriptomes byRNA-Seq. Bioinformatics Alox12 , 8176–8185 (2010). 8176–8185 , (2009). 2078–2079 , et al. et J. Hepatol. J. t al. et P -HTG/rep 6 h (at 6 h after reperfusion)

t al. et S t al. et t al. et 343 X t al. et BMC Bioinformatics BMC t al. et R t al. et t al. et et al. t al. et D P Attenuation of cerebral ischemic injury in interferon regulatory factor regulatory interferon in injury ischemic cerebral of Attenuation t al. et et al. et et al. et

h Sqec AinetMp omt n SAMtools. and format Alignment/Map Sequence The 0 , 84–87 (2014). 84–87 , 35 1

0 h uiutn 3 iae RF eaebts ahlgcl cardiac pathological exacerbates TRAF6 ligase E3 ubiquitin The 19 OP: n mrvd lrfs to fr hr ra alignment. read short for tool ultrafast improved an SOAP2:

Mindin deficiency protects the liver against ischemia/reperfusion against liver the protects deficiency Mindin 1 agtn TA3 inln poet aant eai ischemia/ hepatic against protects signaling TRAF3 Targeting Cardioprotective role of growth/differentiation factor 1 in post-infarction 25 nottn cnes erpoeto aant shmc stroke. ischemic against neuroprotection confers M Oncostatin 7 up4 rtcs gis hptc shmarpruin nuy via injury ischemia–reperfusion hepatic against protects Dusp14 uo ncoi fco rcpo–soitd atr i a positive a is 3 factor receptor–associated factor necrosis Tumor , 12047–12062 (2015). 12047–12062 , Life Sciences Reporting Summary Life Sciences nlss f yohoe 40 eaoie o aahdnc cd by acid arachidonic of metabolites P450 cytochrome of Analysis 7 t al. et The data sets supporting the conclusions of this article are article this of conclusions the supporting sets data The J. Neurochem. J. 8 , 1639–1645 (2009). 1639–1645 , Genome-scale CRISPR–Cas9 knockout screening in human cells. human in screening knockout CRISPR–Cas9 Genome-scale agtn CS8 n FD-ie ppoi rgltr ameliorates regulator apoptosis FADD-like and CASP8 Targeting 5 Tmbim1 is a multivesicular body regulator that protects against protects that regulator body multivesicular a is Tmbim1 t

, 1966–1967 (2009). 1966–1967 , 2 9 -test, and one-way ANOVA was performed for comparisons 63 J. Hepatol. J. 7 4 J. Hepatol. J. . Plasmids are available from the corresponding author corresponding the from available are Plasmids . Nat. Med. Nat. , , ics a ifrain etei fr oprtv genomics. comparative for aesthetic information an Circos: , 1198–1211 (2015). 1198–1211 , S Nat. Methods Nat. R P 1 1

S0168-8278(17)32275-4 (2017). S0168-8278(17)32275-4 7 9

6 h , 559 (2008). 559 , 64 23 t 6 t 5 p , 146–159 (2016). 146–159 , Further information on experimental design , 742–752 (2017). 742–752 , : and and / J. Chromatogr. A Chromatogr. J.

/ d 12 x Nat.Methods . d , 357–360 (2015). 357–360 , S o i R . J. Pathol.J. o P Nat. Commun. Nat. r g 1 Hypertension / 1 1 7 0 post hoc 6 .

group died before the sample 1 6

5 1 7

964 , 621–628, (2008). 236 1 1410 . The mass spectrometry spectrometry mass The . . 1 / j , 41–49 (2014). 41–49 , analysis if the data had n , 360–372(2015)., nature medicine nature

c

, 154–163 (2015). 154–163 , 66 7 . 1 , 11267 (2016). 11267 , Nat. Med. Nat. 3 , 356–367 (2015). 356–367 , 4 P 4 values less than in vitro 8 (2015). Bioinformatics Anal. Chem. Anal.

23 experi , 439– , - - -

nature research | life sciences reporting summary

Corresponding author(s): Hongliang Li Initial submission Revised version Final submission Life Sciences Reporting Summary Nature Research wishes to improve the reproducibility of the work that we publish. This form is intended for publication with all accepted life science papers and provides structure for consistency and transparency in reporting. Every life science submission will use this form; some list items might not apply to an individual manuscript, but all fields must be completed for clarity. For further information on the points included in this form, see Reporting Life Sciences Research. For further information on Nature Research policies, including our data availability policy, see Authors & Referees and the Editorial Policy Checklist.

` Experimental design 1. Sample size Describe how sample size was determined. The sample sizes of each set of animals were determined according to previous studies performed by our group and other scholars and were fixed in a prospective manner. 2. Data exclusions Describe any data exclusions. One mouse in Alox12-HTG/Rep 6h group died before the sample collection and thus were excluded for statistical analysis 3. Replication Describe whether the experimental findings were All in vitro experiments were performed in triplicate unless specified in the figure reliably reproduced. legends. The detailed replication of each experiments has been provided in Figure Legend. All attempts at replication were successful for all experiments. 4. Randomization Describe how samples/organisms/participants were Animals with the same genotype and similar baseline values were randomly allocated into experimental groups. assigned to the sham and surgery groups or the vehicle and ML355 groups. 5. Blinding Describe whether the investigators were blinded to The experimenters were blinded to the animal genotype and grouping information. group allocation during data collection and/or analysis. Note: all studies involving animals and/or human research participants must disclose whether blinding and randomization were used.

6. Statistical parameters For all figures and tables that use statistical methods, confirm that the following items are present in relevant figure legends (or in the Methods section if additional space is needed).

n/a Confirmed

The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement (animals, litters, cultures, etc.) A description of how samples were collected, noting whether measurements were taken from distinct samples or whether the same sample was measured repeatedly A statement indicating how many times each experiment was replicated The statistical test(s) used and whether they are one- or two-sided (note: only common tests should be described solely by name; more complex techniques should be described in the Methods section) A description of any assumptions or corrections, such as an adjustment for multiple comparisons

The test results (e.g. P values) given as exact values whenever possible and with confidence intervals noted June 2017 A clear description of statistics including central tendency (e.g. median, mean) and variation (e.g. standard deviation, interquartile range) Clearly defined error bars

See the web collection on statistics for biologists for further resources and guidance.

1 Nature Medicine: doi:10.1038/nm.4451 ` Software nature research | life sciences reporting summary Policy information about availability of computer code 7. Software Describe the software used to analyze the data in this Agilent MassHunter software (version B.08.00) was used for the quantitative study. analysis of UPLC-MS. Philips IntelliSpace Portal workstation (version 06.0.3.12200) was used to analyze the perfusion CT data. All clean reads were mapped to reference sequences using SOAP2 (version 2.21) in DGE analysis. Proteome Discoverer software (version 1.4) was used to search against the Uniport mouse protein sequences in proteomics analysis. MeV (version 4.9) was used for K-means clustering analysis. GSEA was implemented in the java GSEA (version 2.2.4). A circular layout showing the correlation coefficients and P values of the pathways was generated using Circos (version 0.69). Clean reads were aligned by HISAT2 (version 2.1.0) in RNA-seq analysis. SAMtools (version 1.4.1) was used to produce BAM files in RNA-seq analysis. Cufflinks (version 2.2.1) was applied to estimate the gene expression in RNA-seq analysis. The gene co-expression network was constructed using WGCNA (version 1.51). All statistical analyses used in this study were performed using SPSS (Version 21.0).

For manuscripts utilizing custom algorithms or software that are central to the paper but not yet described in the published literature, software must be made available to editors and reviewers upon request. We strongly encourage code deposition in a community repository (e.g. GitHub). Nature Methods guidance for providing algorithms and software for publication provides further information on this topic.

` Materials and reagents Policy information about availability of materials 8. Materials availability Indicate whether there are restrictions on availability of Agilent MassHunter software was used for the quantitative analysis.The data sets unique materials or if these materials are only available supporting the conclusions of this article are included within the article and its for distribution by a for-profit company. additional files. The DGE and transcriptomics data have been submitted to NCBI Sequence Read Archive with the database identifier of SPR117594, SPR117665 and SPR117667. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD007827. Plasmids are available upon reasonable request. A Life Sciences Reporting Summary for this paper is available. 9. Antibodies Describe the antibodies used and how they were validated The detailed information of all antibodies used in our present study has been for use in the system under study (i.e. assay and species). provided in the "Western blot analysis and antibodies" section in ONLINE METHODS. Antibodies against the following proteins were purchased from Cell Signaling Technology (Beverly, Massachusetts, USA): p-Ikba (#9246), Ikba (#4814), p-p65 (#3033), p65 (#4764), p-p38 (#4511), p38 (#9212), p-ERK1–ERK2 (#4370), ERK1–ERK2 (#4695), p-JNK1–JNK2 (#4668), JNK1–JNK2 (#9252), BCL2 (#3498), BAX (#2772), c-caspase3 (#9664), p-PKC (#2055), PKC (#2058), ALOX5 (#3289), CPLA2 (#2832) and GAPDH (#2118). Antibodies against ALOX12 were obtained from Santa Cruz (sc-365194; 1:200; Dallas, TX, USA) and Sigma (SAB2100109; St.Louis, MO, USA). ALOX15 (sc-133085) antibody was purchased from Santa Cruz Biotechnology. Antibodies against GPR31 (ab75579), GPR40 (ab211049), GPR75 (ab75581) and CYP4A (ab3573) were obtained from Abcam. GPR120 (sc-390752) was obtained from Santa Cruz Biotechnology. Secondary antibodies used in this study included Peroxidase AffiniPure goat anti-rabbit-IgG (H+L) (#111-035-003) and goat anti- mouse-IgG (H+L) (#115-035-003), which were purchased from the Jackson Laboratory (Bar Harbar, ME, USA). The Flag antibody was obtained from medical June 2017 and biological laboratories (MBL, Nagoya, Japan). Unless specified, all primary antibodies were used at the dilution of 1:1000, and secondary antibodies were used at a 1:5,000 dilution.

2 Nature Medicine: doi:10.1038/nm.4451 10. Eukaryotic cell lines nature research | life sciences reporting summary a. State the source of each eukaryotic cell line used. The HEK293T and L02 cells were purchased from the Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China).

b. Describe the method of cell line authentication used. The HEK293T and L02 cell lines were verified by short tandem-repeat DNA profiling before the study.

c. Report whether the cell lines were tested for No mycoplasma contamination was observed in the cells. mycoplasma contamination.

d. If any of the cell lines used are listed in the database No commonly misidentified cells were used in this study. of commonly misidentified cell lines maintained by ICLAC, provide a scientific rationale for their use.

` Animals and human research participants Policy information about studies involving animals; when reporting animal research, follow the ARRIVE guidelines 11. Description of research animals Provide details on animals and/or animal-derived The detailed information about animals and animal-derived materials used in this materials used in the study. study has been provided in our ONLINE METHODS section. Male mice aged 8-10 weeks (24-27 g) were subjected to a 70% warm hepatic IR injury. Mice were maintained in a standard SPF environment with ad libitum access to food and water. ML355 (HY-12341; MCE, Monmouth Junction, NJ, USA) or CDC (ab141560; Abcam, Cambridge, MA, USA) in the solution (DMSO : Solutol : PEG400 : water; 5/10/20/65 v/v/v/v) was injected into mice via the tail vein at 30 min before ischemia and at 6 h after reperfusion at a dose of 3 mg/kg. A blank solution was used as the vehicle control. Primary hepatocytes were isolated from male mice aged 6-8 weeks using collagenase perfusion method. All animal protocols used in this study were approved by the Animal Care and Use Committee of Renmin Hospital of Wuhan University and Institutional Animal Care and Use Committee of the Institute of Model Animal of Wuhan University. Animals received humane cares according to Guide for the Care and Use of Laboratory Animals published by the National Academy of Sciences and the National Institutes of Health. Abuse and maltreatment were avoided in our study.

Policy information about studies involving human research participants 12. Description of human research participants Describe the covariate-relevant population The descriptions of characteristics of human samples have been provided in the characteristics of the human research participants. "Human samples" section and in Supplementary Table 2. The human serum and liver samples used in this study were obtained from patients who underwent liver resection surgery due to hepatocellular carcinoma or hepatic cyst. Liver samples were collected from normal liver tissues of individuals at 3 time points, including before ischemia (baseline group), after ischemia but before reperfusion (ischemia group), and after reperfusion (reperfusion group). The serum samples were obtained at the same time points as the liver samples. Informed consent forms were signed by all donors or their families. All these samples were used only to achieve experimental objective, and the studies were performed according to the principles outlined by the Declaration of Helsinki. All procedures involving human samples were approved by the Renmin Hospital of Wuhan University Review Board and Xijing hospital of The Fourth Military Medical University Review Board. June 2017

3 Nature Medicine: doi:10.1038/nm.4451