Protective effect of Caftaric Acid against hepatic cold ischemia/reperfusion injury in the liver Authors: Ying Zhu, Hao Li, Minghao Li, Wenzheng Ruan, Mingjie Ding, Guohua You, Lele Zhang, Wenzhi Guo Published: 2026 ABSTRACT Hepatic cold ischemia/reperfusion (I/R) injury persists during the perioperative period of liver transplantation (LT), exerting a profound impact on graft survival and being closely associated with the occurrence of early allograft dysfunction (EAD). Currently, preventive strategies against cold I/R injury remain limited. This study aimed to investigate the protective effects and underlying molecular mechanisms of caftaric acid (CA), a metabolite specifically elevated in cold-adapted hibernating animals, against hepatic cold I/R injury. The results demonstrated that CA significantly reduced the area of hepatic necrosis, attenuated liver injury, inhibited hepatocyte apoptosis, and promoted the restoration of redox homeostasis. RNA-seq analysis suggested that the MAPK pathway, particularly the activation of JNK1, plays a pivotal role in the process of cold I/R injury. Through integrated experiments including drug-target screening, molecular docking, and surface plasmon resonance (SPR), this study confirmed that CA directly binds to and interacts with JNK1, thereby regulating its phosphorylation function. Further investigations revealed that CA modulates the MAPK cascade (JNK1/FOS) to alleviate oxidative stress, mitochondrial dysfunction, and cellular apoptosis during hepatic cold I/R injury. In summary, these findings suggest that CA holds promise as a potential therapeutic strategy for ameliorating hepatic cold I/R injury, with significant clinical implications.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13062-026-00826-z. Introduction Liver transplantation (LT) is the ultimate and only curative treatment for patients with end-stage liver disease [1, 2]. However, donor scarcity and stringent inclusion criteria have resulted in persistently high mortality rates among transplant waitlist candidates [3]. Furthermore, cold ischemia/reperfusion (I/R) injury is the primary cause of early allograft dysfunction (EAD), considerably limiting the expansion of donor criteria and the utilization of marginal grafts [4]. For a long time, cold I/R injury has been a significant challenge during the perioperative period of LT [5]. Cold ischemia time (CIT) of > 6 h is an independent risk factor for postoperative graft dysfunction [6, 7]. Prolonged CIT leads to ATP depletion and lactate accumulation, whereas the sudden increase in oxygenation during reperfusion exacerbates oxidative stress via a burst of reactive oxygen species (ROS) production, thereby aggravating I/R injury and increasing the risk of EAD [8, 9]. Various preservation strategies are currently being adopted across clinical centers to address cold-induced I/R injury, including normothermic machine perfusion (NMP), which has been shown to assess and improve organ utilization rates as well as post-LT outcomes [10, 11]. Moreover, multiple studies have shown that the presence of flavin mononucleotide in the perfusate during hypothermic oxygenated perfusion (HOPE) is associated with post-transplant graft nonfunction. HOPE can alleviate mitochondrial damage and reduce the incidence of postoperative biliary complications [12–14]. Similarly, advancements in ischemia-free liver transplantation offer a novel approach to avoiding hepatic I/R injury [15, 16]. Nonetheless, owing to limitations such as application in large-scale populations, conventional static cold storage (SCS) remains the preferred method for organ preservation in most transplant centers. The University of Wisconsin (UW) solution is the most commonly used preservation medium in LT, but it often requires that the CIT of donor livers be limited to 6–10 h [6, 7]. In recent years, increasing attention has been paid to the protective effects of traditional Chinese herbal medicines and natural compounds against various types of injury [17–20]. Caftaric acid (CA) has been shown to alleviate organ damage [21], and its protective mechanisms are attributed to anti-inflammatory, antioxidative stress, free radical scavenging, and antiapoptotic effects [22–24]. Currently, the role of CA in hepatic cold I/R injury remains unknown. Hence, this study aimed to investigate the pharmacological effects of CA on hepatic cold I/R injury and elucidate its potential molecular mechanisms. Materials and methods Criteria for donors and recipients We enrolled patients who underwent LT between February 2023 and June 2024 and met the following criteria: Donor inclusion and exclusion criteria Inclusion criteria Age: ≥3 years and ≤ 75 years.Donor liver source: Donation after citizen’s death or donation after brain death.Donor liver quality: Preoperative biopsy showing macrovesicular steatosis ≤ 30%.Donor free from severe infectious diseases (e.g., HIV, active hepatitis B/hepatitis C, etc.).Donor liver procurement and preservation procedures comply with standard operating protocols. Age: ≥3 years and ≤ 75 years. Age: ≥3 years and ≤ 75 years. Donor liver source: Donation after citizen’s death or donation after brain death. Donor liver source: Donation after citizen’s death or donation after brain death. Donor liver quality: Preoperative biopsy showing macrovesicular steatosis ≤ 30%. Donor liver quality: Preoperative biopsy showing macrovesicular steatosis ≤ 30%. Donor free from severe infectious diseases (e.g., HIV, active hepatitis B/hepatitis C, etc.). Donor free from severe infectious diseases (e.g., HIV, active hepatitis B/hepatitis C, etc.). Donor liver procurement and preservation procedures comply with standard operating protocols. Donor liver procurement and preservation procedures comply with standard operating protocols. Exclusion criteria Donor age < 3 years or > 75 years (extreme values may affect outcomes).Donor liver steatosis > 30% (severe steatosis may lead to graft dysfunction).Donor with a history of malignancy.Donor with uncontrolled infection or sepsis.Donor liver with anatomical abnormalities or severe damage precluding transplantation. Donor age < 3 years or > 75 years (extreme values may affect outcomes). Donor age < 3 years or > 75 years (extreme values may affect outcomes). Donor liver steatosis > 30% (severe steatosis may lead to graft dysfunction). Donor liver steatosis > 30% (severe steatosis may lead to graft dysfunction). Donor with a history of malignancy. Donor with a history of malignancy. Donor with uncontrolled infection or sepsis. Donor with uncontrolled infection or sepsis. Donor liver with anatomical abnormalities or severe damage precluding transplantation. Donor liver with anatomical abnormalities or severe damage precluding transplantation. Recipient inclusion and exclusion criteria Inclusion criteria First-time liver transplant recipient.No early vascular complications (e.g., hepatic artery thrombosis, portal vein thrombosis, etc.).Donor frozen liver tissue specimen of adequate quality for histopathological analysis (e.g., RNA extraction, immunohistochemistry, etc.).Patient or legal guardian signs informed consent, agreeing to clinical data collection and sample analysis.Successful transplant surgery with postoperative survival ≥ 7 days (to allow assessment of early allograft dysfunction, EAD). First-time liver transplant recipient. First-time liver transplant recipient. No early vascular complications (e.g., hepatic artery thrombosis, portal vein thrombosis, etc.). No early vascular complications (e.g., hepatic artery thrombosis, portal vein thrombosis, etc.). Donor frozen liver tissue specimen of adequate quality for histopathological analysis (e.g., RNA extraction, immunohistochemistry, etc.). Donor frozen liver tissue specimen of adequate quality for histopathological analysis (e.g., RNA extraction, immunohistochemistry, etc.). Patient or legal guardian signs informed consent, agreeing to clinical data collection and sample analysis. Patient or legal guardian signs informed consent, agreeing to clinical data collection and sample analysis. Successful transplant surgery with postoperative survival ≥ 7 days (to allow assessment of early allograft dysfunction, EAD). Successful transplant surgery with postoperative survival ≥ 7 days (to allow assessment of early allograft dysfunction, EAD). Exclusion criteria Retransplantation or combined organ transplantation (e.g., combined liver-kidney or liver-heart transplantation).Early postoperative vascular complications (e.g., hepatic artery thrombosis, portal vein thrombosis, or biliary anastomotic leakage, etc.).Donor liver specimen of inadequate quality (e.g., severe hemolysis, RNA degradation, or poor fixation).Patient with severe uncontrolled systemic infection or active malignancy before transplantation.Patient death within 7 days after surgery.Incomplete clinical data (e.g., missing key liver function parameters or cold ischemia time records).Refusal to sign informed consent or inability to comply with follow-up. Retransplantation or combined organ transplantation (e.g., combined liver-kidney or liver-heart transplantation). Retransplantation or combined organ transplantation (e.g., combined liver-kidney or liver-heart transplantation). Early postoperative vascular complications (e.g., hepatic artery thrombosis, portal vein thrombosis, or biliary anastomotic leakage, etc.). Early postoperative vascular complications (e.g., hepatic artery thrombosis, portal vein thrombosis, or biliary anastomotic leakage, etc.). Donor liver specimen of inadequate quality (e.g., severe hemolysis, RNA degradation, or poor fixation). Donor liver specimen of inadequate quality (e.g., severe hemolysis, RNA degradation, or poor fixation). Patient with severe uncontrolled systemic infection or active malignancy before transplantation. Patient with severe uncontrolled systemic infection or active malignancy before transplantation. Patient death within 7 days after surgery. Patient death within 7 days after surgery. Incomplete clinical data (e.g., missing key liver function parameters or cold ischemia time records). Incomplete clinical data (e.g., missing key liver function parameters or cold ischemia time records). Refusal to sign informed consent or inability to comply with follow-up. Refusal to sign informed consent or inability to comply with follow-up. This study was approved by the Ethics Committee of the First Affiliated Hospital of Zhengzhou University (No.2024-KY-0841-001). Animals and the normothermic machine perfusion (NMP) model Male Sprague–Dawley rats, aged 6–8 weeks, were acquired from HFK Bio Technology Co., Ltd. (Beijing, China). All animal experiments were approved by the Animal Care Committee of the First Affiliated Hospital of Zhengzhou University. The methods for rat liver isolation and subsequent NMP have been detailed elsewhere [25]. Cell culture and SCS and rewarming model in vitro Rat primary hepatocytes and the human normal hepatocyte cell line THLE-2 (American Type Culture Collection) were used for the experiments. The cells were cultured in DMEM/F12 medium (Solarbio, Beijing, China) and BEGM Kit Medium (Lonza/Clonetics, CC-3170), supplemented with dexamethasone (40 ng/mL, Beyotime, China), ITS-G (1×, BasaIMedia, China), and fetal bovine serum (10%, KeyGen Biotech, Nanjing, China). The cell experiments included the following groups: control, static cold storage (SCS), SCS + CA, SCS + rewarming (SR), and SR + CA. To simulate LT-like injury, the cells were cultured in UW solution in a 4 °C refrigerator, followed by normal culturing for 2 h. CA was added during the SCS period, and the remaining treatments were consistent with those performed in the SR group. During temperature transitions, a gradual acclimation approach was adopted to avoid thermal shock to the cells. Before transferring cells from 37 °C to 4 °C, they were first equilibrated at room temperature for at least 15 min. Before transferring cells back from 4 °C to 37 °C, the pre-chilled maintenance medium was replaced, and the cells were equilibrated at room temperature for at least 15 min before subsequent operations. Isolation of primary rat hepatocytes Based on previous research [26], rat primary hepatocytes were isolated. Briefly, following isoflurane anesthesia, the hepatic lobes of rats were ligated, followed by perfusion with Hank’s Balanced Salt Solution (HBSS; GE, USA) supplemented with 0.5 mM EGTA. Subsequently, the liver was perfused with 0.5 mg/mL collagenase type IV (Sigma-Aldrich, USA) for 7–10 min for digestion. The livers were harvested and prepared into single-cell suspensions. Following washing and centrifugation, primary hepatocytes were isolated using a 90% Percoll solution (Huayun, China). After 8 h of adherence, the medium was replaced and incubated overnight. Measurement of hepatocyte injury The perfusate and cell supernatant were collected, and the activities of lactate dehydrogenase (LDH), aspartate aminotransferase (AST), and alanine aminotransferase (ALT) were measured using the corresponding kits (C010-2-1, A020-2-2, and C009-2-1; Jiancheng, Nanjing, China). Hematoxylin-eosin (H&E) and TUNEL staining The liver samples were fixed overnight, rinsed with phosphate-buffered saline, dehydrated in a series of gradient ethanol, embedded in paraffin, and sectioned. HE (Solarbio, Beijing, China) staining was then performed on the liver sections, followed by microscopic observation. The paraffin-embedded rat liver sections were subjected to TUNEL staining using an in situ cell death detection kit (G1504-50T Servicebio, Wuhan, China), which was subsequently examined by confocal microscopy (LSM980, Zeiss, Germany). Enzyme-linked immunosorbent assay (ELISA) According to the manufacturer’s protocol, the perfusion solution was harvested from the liver, and the levels of tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), and IL-6 in the perfusate were determined by using an ELISA kit (Servicebio, Wuhan, China). Cell viability analysis The CCK-8 kit (B5350B; Biosharp, Hefei, China) was used for the cell viability assay. After seeding the primary hepatocytes in a 96-well plate, the cells were treated with a gradient of CA concentrations and incubated for 24 h. CCK-8 reagent was then added, followed by measurement of absorbance using a microplate reader (Synergy H1, BioTek, USA). Oxidative stress analysis After snap-freezing in liquid nitrogen, the liver sections were used for dihydroethidium (DHE) quantification. DHE was measured by using a DHE detection kit (Beyotime, Nanjing, China) in adherence to the manufacturer’s protocol. ROS assessment in cells was conducted immediately following co-incubation with MitoNeoD (563761, MedKoo Biosciences, NC, USA) at a concentration of 2.5 µM for 15 min at 37 °C. Mitochondrial ROS were evaluated by using the MitoNeoD probe through confocal microscopy (LSM980, Zeiss, Germany) [27]. The levels of reduced glutathione (GSH) and oxidized glutathione (GSSG) were determined by using a GSH/GSSG detection kit (Cat. No. S0053, Beyotime, China). The cells were collected, washed, centrifuged, and then incubated with Protein Removal Reagent M; after repeated freeze-thawing, the samples were assigned to two groups: one sample group was subjected to the determination of native GSSG, and the other to the determination of total GSSG. A GSSG standard solution was prepared to plot a standard curve, and colorimetric determination was performed by using a multi-mode microplate reader (Synergy H1, BioTek). Flow cytometry The Annexin V FITC/PI apoptosis kit (KeyGen Biotech, Nanjing, China) and MitoNeoD were used to detect the hepatocyte apoptosis rate and the MitoNeoD positive rate, respectively. Briefly, the cells subjected to different treatments were digested with trypsin without EDTA (T1350, Solarbio, Beijing, China), stained with respective probes, and then analyzed by using the BD FACS Canto II flow cytometer (BD Biosciences, CA, USA). Determination of JNK1 kinase activity JNK1 kinase activity was determined using a JNK1 Kinase Assay Kit (GMS50164.1, Genmed Scientifics). Briefly, adherent THLE‑2 cells were washed with the cleaning solution provided in the kit and gently harvested by cell scraper to avoid trypsin digestion. Cells were centrifuged at 300 g for 5 min at 4 °C. After discarding the supernatant, cell lysis buffer was added for 30 min of lysis on ice. The lysate was centrifuged at 16,000 g for 5 min, and the supernatant was collected. Protein concentration was quantified via the Bradford method. The absorbance values of blank control, total kinase activity and non‑specific activity were detected by a microplate reader under corresponding conditions, and the specific activity of JNK1 kinase was calculated according to the manufacturer’s formula. Cellular thermal shift assay (CETSA) THLE2 cells were seeded in 10 cm culture dishes and cultured until reaching 90% confluence. The cells were washed three times with pre‑cold PBS. Subsequently, cells were collected by scraping and centrifuged at 1000 × g for 5 min at 4 °C to obtain cell pellets. The cell pellets were resuspended in cell lysis buffer (P0013, Beyotime) supplemented with 1% protease inhibitor cocktail and lysed on ice for 10 min. After lysis, the samples were centrifuged at 12,000 × g for 15 min at 4 °C, and the supernatants were collected. The lysate supernatants were incubated with 50 µmol/L CA or vehicle control (DMSO) at 20 °C for 2 h. Each supernatant was aliquoted at 60 µL per microtube, followed by gradient temperature treatment using a PCR thermal cycler (Thermo Fisher Scientific): heating at 48–60 °C for 5 min, and then cooling to 25 °C for 3 min. RNA sequencing (RNA-Seq) and analysis Total RNA was isolated as per the Trizol protocol. Subsequently, a cDNA library was prepared. The fragment sizes were detected by using the Qsep400 High-throughput Biological Fragment Analyzer. Based on the Illumina sequencing parameters (i.e., effective concentration and target data output), the libraries were pooled to generate 150-bp paired-end reads. After filtering raw data with fastp, an HISAT index was constructed, and the clean reads were aligned with the reference genome. Gene alignment was quantified based on feature counts, followed by FPKM calculation for each gene using its length. DESeq2 facilitated differential expression analysis between the groups, with Benjamini and Hochberg-adjusted P-values. The significant differences were determined by adjusted P-value < 0.05 and log2 (fold change) > 2. Real-time quantitative PCR Animal tissues and cells were used for total RNA isolation with TRIzol (Invitrogen). A NanoDrop 2000 (Thermo Fisher Scientific, Madison, WI, USA) was used to determine the RNA concentration and purity. After synthesizing cDNA from 2 µg of RNA, RT-qPCR was performed on the QuantStudio™5 Real-Time PCR System (Thermo Fisher Scientific) with specific primers. Western blotting (WB) analysis The liver tissues and cells served as sources for protein isolation. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was used to fractionate the proteins, followed by electrotransfer onto a polyvinylidene fluoride (PVDF) membrane. Following blocking, the blots were incubated with primary antibodies overnight. Thereafter, secondary antibodies (1:5000; goat anti-rabbit or goat anti-mouse; Abclonal, Wuhan, China) were used to probe the membranes. Enhanced chemiluminescence (ECL) reagents (Biosharp) enabled the visualization of protein bands, and ImageJ software was applied to quantify the protein expression. Table 1Relationship between cold ischemia time and donor liver characteristicsCharacteristics of donor liversCIT< 6 h (n = 138)6–8 h (n = 94)> 8 h (n = 34)P valueAge (years)43.30 ± 15.7149.74 ± 13.8848.32 ± 17.410.006Female proportion (%)19.6%12.8%14.7%0.374Height (cm)169.48 ± 11.53170.62 ± 11.14167.64 ± 15.310.450Weight (kg)69.04 ± 15.7172.38 ± 14.7469.26 ± 17.650.259steatosis proportion (%)34.1%34.0%35.3%0.990 Table 1Relationship between cold ischemia time and donor liver characteristicsCharacteristics of donor liversCIT< 6 h (n = 138)6–8 h (n = 94)> 8 h (n = 34)P valueAge (years)43.30 ± 15.7149.74 ± 13.8848.32 ± 17.410.006Female proportion (%)19.6%12.8%14.7%0.374Height (cm)169.48 ± 11.53170.62 ± 11.14167.64 ± 15.310.450Weight (kg)69.04 ± 15.7172.38 ± 14.7469.26 ± 17.650.259steatosis proportion (%)34.1%34.0%35.3%0.990 Relationship between cold ischemia time and donor liver characteristics Table 2Relationship between Cold Ischemia Time and Recipient Liver CharacteristicsCharacteristics of Recipient liversCIT< 6 h (n = 138)6–8 h (n = 94)> 8 h (n = 34)P valueAge (years)48.18 ± 14.4446.39 ± 13.4147.32 ± 10.760.618Female proportion (%)25.4%23.4%14.7%0.420ALT peak (U/L)964.59 ± 1801.321107.98 ± 1126.991689.94 ± 1859.930.063AST peak (U/L)735.01 ± 979.98913.26 ± 970.962521.65 ± 5982.98< 0.001TBIL(µmol/L)93.35 ± 75.6295.34 ± 68.43118.08 ± 106.700.244ALB(U/L)109.84 ± 70.49107.17 ± 64.91107.75 ± 70.630.955PAB (U/L)146.39 ± 98.76135.68 ± 82.41156.53 ± 87.140.474GGT (U/L)80.86 ± 82.0684.81 ± 103.6396.39 ± 67.400.657ALP (U/L)78.37 ± 77.3880.14 ± 76.6076.99 ± 60.120.973INR1.24 ± 0.271.26 ± 0.271.28 ± 0.260.700EAD (%)18.1%25.5%47.1%0.002 Table 2Relationship between Cold Ischemia Time and Recipient Liver CharacteristicsCharacteristics of Recipient liversCIT< 6 h (n = 138)6–8 h (n = 94)> 8 h (n = 34)P valueAge (years)48.18 ± 14.4446.39 ± 13.4147.32 ± 10.760.618Female proportion (%)25.4%23.4%14.7%0.420ALT peak (U/L)964.59 ± 1801.321107.98 ± 1126.991689.94 ± 1859.930.063AST peak (U/L)735.01 ± 979.98913.26 ± 970.962521.65 ± 5982.98< 0.001TBIL(µmol/L)93.35 ± 75.6295.34 ± 68.43118.08 ± 106.700.244ALB(U/L)109.84 ± 70.49107.17 ± 64.91107.75 ± 70.630.955PAB (U/L)146.39 ± 98.76135.68 ± 82.41156.53 ± 87.140.474GGT (U/L)80.86 ± 82.0684.81 ± 103.6396.39 ± 67.400.657ALP (U/L)78.37 ± 77.3880.14 ± 76.6076.99 ± 60.120.973INR1.24 ± 0.271.26 ± 0.271.28 ± 0.260.700EAD (%)18.1%25.5%47.1%0.002 Relationship between Cold Ischemia Time and Recipient Liver Characteristics Table 3Putative molecular targets of Caftaric AcidFitnessHypoIDTarget IDMatched Ligand Sites2.414zhiKIF11D(8)N(10)R(11)2.3284n0sMK01A(5)D(7)R(11)2.3271yi4PIM1D(6)D(7)R(11)2.2692a5uPK3CGA(5)A(1)R(11)2.2623v3vJNK1A(3)A(5)D(7) R(11)2.2561byqHS90AD(7)N(9)R(11)2.2224nxwMID51D(7)N(10)R(11)2.1594qx5KGP1A(3)A(2)R(11)2.1531hwlHMDHN(10)N(9)R(11)2.1462ax4PAPS2D(8)N(10)R(11) Table 3Putative molecular targets of Caftaric AcidFitnessHypoIDTarget IDMatched Ligand Sites2.414zhiKIF11D(8)N(10)R(11)2.3284n0sMK01A(5)D(7)R(11)2.3271yi4PIM1D(6)D(7)R(11)2.2692a5uPK3CGA(5)A(1)R(11)2.2623v3vJNK1A(3)A(5)D(7) R(11)2.2561byqHS90AD(7)N(9)R(11)2.2224nxwMID51D(7)N(10)R(11)2.1594qx5KGP1A(3)A(2)R(11)2.1531hwlHMDHN(10)N(9)R(11)2.1462ax4PAPS2D(8)N(10)R(11) Putative molecular targets of Caftaric Acid Molecular docking The target compound CA was screened against the pharmacophore library previously reported by Aurélien et al. [28]. The CA was downloaded from the PubChem database and subjected to protonation and structural optimization by using the LigPrep module in Schrodinger 2021-1. Target screening was performed using the Phase module [29] of the Schrodinger 2021-1 software package. After the screening, Fitness scoring was conducted based on the screening outcomes. Molecular docking simulations were performed using AutoDock Vina 1.2.3 software [30]. Furthermore, ADFRsuite 1.0 was employed to convert all processed receptor proteins and small molecules into the PDBQT format, as required for docking in AutoDock Vina 1.2.3 [31]. The protein structures were retrieved from the PDB and subjected to visual analysis using PyMol 2.5.4 software. CA’s van der Waals force, salt bridge interaction, hydrogen bonds, and docking score with two proteins in the mitogen-activated protein kinase (MAPK) signaling pathway (Table S3). Molecular dynamics simulation AMBER 22 software was used to conduct all-atom molecular dynamics simulations, using the small molecule–protein complex acting as the starting structure. Before initiating the simulations, CA charges were computed via the Antechamber module and Hartree–Fock (HF) SCF/6-31G* method within Gaussian 09 software. Next, the small molecule and protein were parameterized with the GAFF2 small molecule force field and the ff14SB protein force field [32, 33], respectively. Finally, topology and parameter files were generated for the simulations. Furthermore, binding free energies between proteins and ligands across all systems were calculated by using the MM/GBSA method [34, 35] (Table S4). Alanine scanning mutagenesis The binding free energy between JNK1 and Caftaric Acid and its per‑residue energy decomposition were calculated using the MM/GBSA method [58–60]. The calculations were based on molecular dynamics trajectories from 90 to 100 ns after equilibration. The binding free energy (ΔGbind) was derived from the difference in free energies among the complex, receptor, and ligand, with the energy components expressed as follows:1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\eqalign{ & \Delta Gbind\, = \,\Delta Einternal\, + \,\Delta EvDW\, \cr & + \,\Delta Eelec\, + \,\Delta GGB\, + \,\Delta GSA \cr}$$\end{document} In Formula (2), ΔEinternal represents the internal energy (including bond, angle, and dihedral terms), ΔEvDW and ΔEelec denote the van der Waals and electrostatic interactions, respectively. The solvation free energy is divided into a polar component (ΔGGB) and a non-polar component (ΔGSA): ΔGGB was calculated using the generalized Born model (parameter igb = 2), while ΔGSA was estimated as 0.0072 × SASA. Entropic contributions (-TΔS) were not included due to computational cost and accuracy limitations. Subsequently, alanine mutation was performed individually for the top 10 energy-contributing residues, and the change in binding free energy (ΔΔG) after mutation was calculated. Surface plasmon resonance Surface plasmon resonance (SPR) analysis was performed with proteins immobilized via amine coupling. All experiments were conducted using running buffer (1×PBS, pH 7.4), with the interaction analysis buffer containing 5% DMSO. The procedure was as follows: after system preparation, a CM5 sensor chip was loaded into the instrument. First, an EDC/NHS mixture was injected into channel 2 at a flow rate of 10 µL/min to activate the carboxyl groups on the chip surface. The target protein was then diluted to 50 µg/mL with sodium acetate buffer and injected at the same flow rate into channel 2 for coupling, generating a binding curve. Ethanolamine was subsequently injected to block unreacted activated sites. To account for background signals, channel 1 served as the reference channel and underwent a “mock coupling” procedure identical to that of channel 2 except that protein‑free acetate buffer was used instead of protein solution. Statistical analysis Statistical analyses were performed using GraphPad Prism (version 9.5.1, GraphPad Software, San Diego, USA). All data are presented as the mean ± standard deviation (SD). Differences between the groups were analyzed by the student’s t-test or one‐way analysis of variance. P < 0.05 was considered to indicate statistical significance. Results Caftaric Acid alleviates liver cold ischemia/reperfusion injury induced by extended cold ischemia time Based on the data presented in Tables 1 and 2, when donor liver data showed no significant differences, recipient analysis indicated that peak AST levels within seven days after surgery rose with longer cold ischemia time. Moreover, prolonged cold ischemia time significantly increased the risk of early allograft dysfunction (EAD), highlighting the clinical relevance of this study for reducing hepatic cold ischemia/reperfusion injury. Our previous studies have shown that beneficial metabolites upregulated during hepatic cold adaptation in hibernators confer protection against cold I/R injury in nonhibernators [27, 36]. Here, CA was a unique metabolite in cold-stored hamster (hibernator) livers compared with rat livers (Figure S1A). We first evaluated the cytotoxicity of CA. By assessing the viability of rat primary hepatocytes incubated with varying concentrations of CA, 50 µM was determined as the optimal concentration and subsequently used in the following experiments (Figure S1B). HE staining revealed no significant pathological changes in any of the examined tissues (Figure S1C). Serum biochemical analysis further demonstrated that all evaluated markers of cardiac (CK, LDH1), hepatic (AST, ALT), and renal (BUN, CREA) function remained within normal physiological ranges (Figure S1D). These results indicate that no obvious signs of toxicity were observed in vital organs under the tested concentration conditions. To investigate the potential protective effect of CA in cold I/R injury, an isolated perfused rat liver model was applied in this study (Fig. 1A). Following an extended CIT and subsequent NMP, perfusates from CA-supplemented livers exhibited significantly lower levels of ALT, AST, and LDH than those from the vehicle control group (Fig. 1B). Furthermore, reduced portal vein resistance indicated improved vascular function and recovery in CA-treated livers (Fig. 1C). In addition, histological analysis demonstrated substantial sinusoidal congestion and hepatocyte injury upon reperfusion in control livers, which were considerably attenuated by CA supplementation (Fig. 1D). Moreover, TUNEL and DHE staining showed substantially reduced apoptosis and mitochondrial ROS in CA-treated livers compared with controls (Figs. 1E–H). Western blot analysis of apoptosis (Bcl-2, Bax, and cleaved caspase-3) and antioxidant response (GCLC, GCLM, and HO-1) markers validated these findings (Figs. 1I–J). Also, CA supplementation reduced the secretion of proinflammatory cytokines (IL-1β, IL-6, and TNF-α) (Fig. 1K). Fig. 1CA mitigates hepatic cold I/R injury and mitochondrial ROS in CIT extended rat livers. (A) Schematic diagram illustrating the static cold storage (SCS) and reperfusion procedures for isolated rat livers. (B-H) After 48 h of SCS followed by 2 h of reperfusion, analyses were performed for the following: AST, ALT and LDH in perfusate samples (B, n = 8), portal venous resistance (PVR) (C), hematoxylin-eosin (HE) staining of liver sections (D), TUNEL positive cells (E, F, n = 8), and Dihydroethidium Staining (DHE) (G, H, n = 8). (I-J) Western blot analysis (I) and quantitative levels (J, n = 4) of antioxidant-related proteins (including HO-1, GCLC, and GCLM) and apoptosis-related proteins (Bax, Bcl2, and cleaved Caspase-3) under the treatment conditions. (K) Levels of proinflammatory cytokines (TNF-α, IL-1β, and IL-6) in the perfusate (n = 8). Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one-way analysis of variance CA mitigates hepatic cold I/R injury and mitochondrial ROS in CIT extended rat livers. (A) Schematic diagram illustrating the static cold storage (SCS) and reperfusion procedures for isolated rat livers. (B-H) After 48 h of SCS followed by 2 h of reperfusion, analyses were performed for the following: AST, ALT and LDH in perfusate samples (B, n = 8), portal venous resistance (PVR) (C), hematoxylin-eosin (HE) staining of liver sections (D), TUNEL positive cells (E, F, n = 8), and Dihydroethidium Staining (DHE) (G, H, n = 8). (I-J) Western blot analysis (I) and quantitative levels (J, n = 4) of antioxidant-related proteins (including HO-1, GCLC, and GCLM) and apoptosis-related proteins (Bax, Bcl2, and cleaved Caspase-3) under the treatment conditions. (K) Levels of proinflammatory cytokines (TNF-α, IL-1β, and IL-6) in the perfusate (n = 8). Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one-way analysis of variance Fig. 2CA effectively alleviates cellular structural injury and apoptosis during cold ischemia/reperfusion. (A) Timeline of CA supplementation, static cold storage (SCS), and cell sampling after 2 h of rewarming. (B) Morphological changes in primary rat hepatocytes under different treatment conditions. (C) Levels of AST, ALT, and LDH in the culture supernatant of primary rat hepatocytes after SCS and rewarming (SR), with or without CA supplementation. (D-E) Proportion of PI positive cells in primary rat hepatocytes and THLE2 hepatocytes after SR under specified conditions. (F-G) Flow cytometry analysis of apoptosis in primary rat hepatocytes and THLE2 hepatocytes after treatment. (H-I) Western blot analysis of the expression levels of apoptosis related proteins (Bax, Bcl-2, and cleaved caspase-3) in primary rat hepatocytes and THLE2 hepatocytes following treatment. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one‐way analysis of variance CA effectively alleviates cellular structural injury and apoptosis during cold ischemia/reperfusion. (A) Timeline of CA supplementation, static cold storage (SCS), and cell sampling after 2 h of rewarming. (B) Morphological changes in primary rat hepatocytes under different treatment conditions. (C) Levels of AST, ALT, and LDH in the culture supernatant of primary rat hepatocytes after SCS and rewarming (SR), with or without CA supplementation. (D-E) Proportion of PI positive cells in primary rat hepatocytes and THLE2 hepatocytes after SR under specified conditions. (F-G) Flow cytometry analysis of apoptosis in primary rat hepatocytes and THLE2 hepatocytes after treatment. (H-I) Western blot analysis of the expression levels of apoptosis related proteins (Bax, Bcl-2, and cleaved caspase-3) in primary rat hepatocytes and THLE2 hepatocytes following treatment. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one‐way analysis of variance Fig. 3CA restores redox homeostasis in hepatocytes following LT-like stress. (A) MitoNeoD staining in primary rat hepatocytes. (B) MitoNeoD staining in THLE2 hepatocytes. (C) Flow cytometric analysis of MitoNeoD-positive primary rat hepatocytes after treatment. (D) Flow cytometric analysis of MitoNeoD-positive THLE2 hepatocytes after treatment. (E–F) Quantitative analysis of GSH levels and GSH/GSSG ratios in primary rat hepatocytes and THLE2 hepatocytes under indicated conditions. (G–H) Western blot analysis of antioxidant‑related proteins (including HO‑1, GCLC, and GCLM) in primary rat hepatocytes and THLE2 hepatocytes. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one‐way analysis of variance CA restores redox homeostasis in hepatocytes following LT-like stress. (A) MitoNeoD staining in primary rat hepatocytes. (B) MitoNeoD staining in THLE2 hepatocytes. (C) Flow cytometric analysis of MitoNeoD-positive primary rat hepatocytes after treatment. (D) Flow cytometric analysis of MitoNeoD-positive THLE2 hepatocytes after treatment. (E–F) Quantitative analysis of GSH levels and GSH/GSSG ratios in primary rat hepatocytes and THLE2 hepatocytes under indicated conditions. (G–H) Western blot analysis of antioxidant‑related proteins (including HO‑1, GCLC, and GCLM) in primary rat hepatocytes and THLE2 hepatocytes. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one‐way analysis of variance Fig. 4JNK1 may serve as a key molecular target through which CA exerts its effects within the MAPK pathway. (A) Schematic diagram of RNA-seq sampling. (B) Volcano plot of differentially expressed genes (DEGs) induced by SR. (C-D) KEGG and GO enrichment analyses of DEGs induced by SR. (E) GSEA plot showing enrichment of gene sets in the MAPK pathway during SR. (F-G) Western blot analysis of phosphorylated JNK1/JNK1, phosphorylated ERK/ERK, phosphorylated p38/p38 phosphorylated MKK4/MKK4 and phosphorylated MKK7/MKK7 under corresponding treatment conditions. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by one-way analysis of variance JNK1 may serve as a key molecular target through which CA exerts its effects within the MAPK pathway. (A) Schematic diagram of RNA-seq sampling. (B) Volcano plot of differentially expressed genes (DEGs) induced by SR. (C-D) KEGG and GO enrichment analyses of DEGs induced by SR. (E) GSEA plot showing enrichment of gene sets in the MAPK pathway during SR. (F-G) Western blot analysis of phosphorylated JNK1/JNK1, phosphorylated ERK/ERK, phosphorylated p38/p38 phosphorylated MKK4/MKK4 and phosphorylated MKK7/MKK7 under corresponding treatment conditions. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by one-way analysis of variance Fig. 5CA specifically binds to and regulates JNK1. (A) Schematic visualization of molecular docking between CA and JNK1. (B) Surface plasmon resonance (SPR) analysis of CA-JNK1 interaction. (C) Conformational changes of CA and JNK1 over simulation time. (D) Variations in RMSD, RoG, and number of hydrogen bonds for the CA–JNK1 complex over simulation time, along with the RMSF fluctuation plot of the protein during simulation. (E) Top 10 amino acid residues contributing most to binding energy. (F) Alanine scanning mutagenesis of the top 10 binding energy-contributing residues. (G) Protein thermal shift assay for characterizing the interaction of recombinant JNK1 or JNK1 LEU168 with CA (n = 3). (H-I) Determination of JNK1 kinase activity to verify the inhibitory effect of CA on it (n = 3). Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one‐way analysis of variance CA specifically binds to and regulates JNK1. (A) Schematic visualization of molecular docking between CA and JNK1. (B) Surface plasmon resonance (SPR) analysis of CA-JNK1 interaction. (C) Conformational changes of CA and JNK1 over simulation time. (D) Variations in RMSD, RoG, and number of hydrogen bonds for the CA–JNK1 complex over simulation time, along with the RMSF fluctuation plot of the protein during simulation. (E) Top 10 amino acid residues contributing most to binding energy. (F) Alanine scanning mutagenesis of the top 10 binding energy-contributing residues. (G) Protein thermal shift assay for characterizing the interaction of recombinant JNK1 or JNK1 LEU168 with CA (n = 3). (H-I) Determination of JNK1 kinase activity to verify the inhibitory effect of CA on it (n = 3). Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one‐way analysis of variance Fig. 6Activation of JNK1 exacerbates hepatic cold ischemia/reperfusion injury in rats with prolonged CIT. (A) HE staining under conditions of JNK1 activation. (B-C) TUNEL and DHE staining of the liver after reperfusion. (D-E) Western blot analysis revealing the expression levels of phosphorylated JNK1/JNK1, HO-1, GCLC, GCLM, Bax, Bcl-2, and cleaved caspase-3. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one‐way analysis of variance Activation of JNK1 exacerbates hepatic cold ischemia/reperfusion injury in rats with prolonged CIT. (A) HE staining under conditions of JNK1 activation. (B-C) TUNEL and DHE staining of the liver after reperfusion. (D-E) Western blot analysis revealing the expression levels of phosphorylated JNK1/JNK1, HO-1, GCLC, GCLM, Bax, Bcl-2, and cleaved caspase-3. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one‐way analysis of variance Fig. 7Activation of JNK1 significantly exacerbates hepatocyte injury following LT-like stress. (A) Hepatocyte enzyme levels in the supernatant of THLE2 cells under corresponding conditions. (B-C) Flow cytometric analysis of apoptosis rate in THLE2 hepatocytes. (D) Proportion of MitoNeoD positive THLE2 hepatocytes. (E) Levels of GSH and GSH/GSSG ratio under indicated experimental conditions. (F) Western blot analysis revealing the expression levels of phosphorylated JNK1/JNK1, HO-1, GCLC, GCLM, Bax, Bcl-2, and cleaved caspase-3. (G) Verify JNK1 knockdown via Western blot. (H) Determination of liver injury under JNK1 knockdown conditions. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one‐way analysis of variance Activation of JNK1 significantly exacerbates hepatocyte injury following LT-like stress. (A) Hepatocyte enzyme levels in the supernatant of THLE2 cells under corresponding conditions. (B-C) Flow cytometric analysis of apoptosis rate in THLE2 hepatocytes. (D) Proportion of MitoNeoD positive THLE2 hepatocytes. (E) Levels of GSH and GSH/GSSG ratio under indicated experimental conditions. (F) Western blot analysis revealing the expression levels of phosphorylated JNK1/JNK1, HO-1, GCLC, GCLM, Bax, Bcl-2, and cleaved caspase-3. (G) Verify JNK1 knockdown via Western blot. (H) Determination of liver injury under JNK1 knockdown conditions. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one‐way analysis of variance Fig. 8CA modulates hepatic cold ischemia/reperfusion injury through the JNK1/FOS pathway. (A) Schematic diagram of primary rat hepatocyte sampling. (B) Top 20 differentially expressed genes (DEGs) induced by CA. (C) Expression levels of the top 20 CA-induced DEGs in THLE2 cells. (D) Western blot analysis of FOS overexpression. (E) Levels of ALT, AST, and LDH in THLE2 hepatocytes under corresponding treatment conditions. (F) PI staining. (G) MitoNeoD staining. (H) Western blot analysis revealing the expression levels of phosphorylated c-FOS, HO-1, GCLC, GCLM, Bax, Bcl-2, and cleaved caspase-3. (I) Western blot analysis of phosphorylated FOS/FOS in rat liver and THLE2 hepatocytes under JNK1 activated conditions. (J) Intracellular GSH levels in THLE2 hepatocytes. (K-L) Changes in FOS phosphorylation levels under JNK1 knockdown (K, n = 3) and overexpression (L, n = 3) conditions. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one‐way analysis of variance CA modulates hepatic cold ischemia/reperfusion injury through the JNK1/FOS pathway. (A) Schematic diagram of primary rat hepatocyte sampling. (B) Top 20 differentially expressed genes (DEGs) induced by CA. (C) Expression levels of the top 20 CA-induced DEGs in THLE2 cells. (D) Western blot analysis of FOS overexpression. (E) Levels of ALT, AST, and LDH in THLE2 hepatocytes under corresponding treatment conditions. (F) PI staining. (G) MitoNeoD staining. (H) Western blot analysis revealing the expression levels of phosphorylated c-FOS, HO-1, GCLC, GCLM, Bax, Bcl-2, and cleaved caspase-3. (I) Western blot analysis of phosphorylated FOS/FOS in rat liver and THLE2 hepatocytes under JNK1 activated conditions. (J) Intracellular GSH levels in THLE2 hepatocytes. (K-L) Changes in FOS phosphorylation levels under JNK1 knockdown (K, n = 3) and overexpression (L, n = 3) conditions. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t-test or one‐way analysis of variance Fig. 9JNK1/FOS expression is elevated in liver tissues following liver transplantation and predicts unfavorable prognosis. (A-B) Protein expression levels of p-JNK1/JNK1 and p-FOS/FOS in liver tissues before and after transplantation. (C-D) Protein expression levels of p-JNK1/JNK1 and p-FOS/FOS in liver tissues from patients who developed EAD and IGF after transplantation. (E) ROC curve analysis was performed to evaluate the ability of intrahepatic JNK1 and FOS mRNA levels to discriminate patients who developed early allograft dysfunction (EAD) from those with immediate graft function (IGF) after liver transplantation. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t‐test or one‐way analysis of variance JNK1/FOS expression is elevated in liver tissues following liver transplantation and predicts unfavorable prognosis. (A-B) Protein expression levels of p-JNK1/JNK1 and p-FOS/FOS in liver tissues before and after transplantation. (C-D) Protein expression levels of p-JNK1/JNK1 and p-FOS/FOS in liver tissues from patients who developed EAD and IGF after transplantation. (E) ROC curve analysis was performed to evaluate the ability of intrahepatic JNK1 and FOS mRNA levels to discriminate patients who developed early allograft dysfunction (EAD) from those with immediate graft function (IGF) after liver transplantation. Results are presented as mean ± SD. ns, not significant, *P < 0.033, **P < 0.002, ***P < 0.001 by the student’s t‐test or one‐way analysis of variance Caftaric Acid alleviates hepatocyte injury following LT-like stress Primary rat hepatocytes and THLE-2 cells were used as in vitro models for assessing the protective effects of CA against LT-like injury (Fig. 2A). Morphologically, primary hepatocytes subjected to SCS and Rewarm (SR) exhibited significant swelling and nuclear pyknosis. However, these pathological changes were markedly ameliorated following CA treatment (Fig. 2B). Moreover, hepatocytes supplemented with CA demonstrated decreased levels of AST, LDH, and ALT after SR (Fig. 2C). Regarding cell death, PI staining revealed that SR treatment led to extensive late necrosis, whereas CA pretreatment significantly reduced the proportion of PI positive cells (Fig. 2D-E). To quantitatively evaluate the effect of CA on SR induced apoptosis at the cellular level, a high sensitivity flow cytometry assay using Annexin V-FITC and PI double staining was performed. In SR models established with both rat primary hepatocytes and the human THLE2, quantitative analysis consistently showed that compared with the normal control group, the total apoptosis rate was significantly increased in the SR group, indicating that cold stress followed by rewarming successfully induced substantial apoptotic cell death. However, in cells pretreated with CA, the total apoptosis rate was significantly lower than that in the SR group in both cellular models (Figs. 2F-G). To elucidate the underlying molecular mechanism, we examined key apoptotic pathway proteins by Western blot. The results demonstrated that CA markedly upregulated the expression of the anti-apoptotic protein Bcl-2, while down regulating the levels of the pro-apoptotic protein Bax and the cleaved caspase-3 (Figures. 2H-I). Together, these results indicate that CA inhibits SR-induced hepatocyte apoptosis, thereby alleviating cold ischemia/reperfusion injury. Caftaric acid normalizes the redox homeostasis induced by LT-like stress Further studies revealed that the protective effect of CA is closely associated with its systemic regulation of cellular redox homeostasis. A sharp increase in ROS levels was detected by the MitoNeoD fluorescent probe in both rat primary hepatocytes and THLE2 cells after SR treatment, as revealed by intensified fluorescence signals. In contrast, CA pretreatment markedly suppressed this ROS burst, significantly reducing the fluorescence intensity (Figs. 3A-B). Quantitative flow cytometric analysis of the MitoNeoD positive rate further confirmed that the CA-treated group exhibited a significantly lower proportion of MitoNeoD positive cells than the SR group, statistically verifying the ability of CA to suppress mitochondrial oxidative stress (Figs. 3C-D). As a critical intracellular antioxidant, the level of reduced glutathione (GSH) directly reflects the cellular antioxidant reserve. This study demonstrated that SR treatment led to a sharp depletion of intracellular GSH in both rat primary hepatocytes and THLE2 cells, whereas CA intervention significantly promoted the restoration of GSH, allowing its levels to return closer to normal more rapidly after rewarming (Figs. 3E-F). To elucidate the molecular basis of CA-enhanced antioxidant defense, we examined the expression of related enzymatic proteins. Western blot analysis indicated that CA significantly upregulated the expression of both GCLC and GCLM-the rate-limiting enzymes for GSH synthesis, while also inducing the upregulation of HO-1 expression (Figs. 3G-H). Collectively, these results demonstrate that CA restores redox homeostasis in hepatocytes following LT-like stress by inhibiting mitochondrial ROS generation, promoting GSH biosynthesis, and inducing the expression of key antioxidant enzymes. JNK1 mediates the protective role of Caftaric Acid in hepatic cold ischemia/reperfusion injury Based on transcriptomic analysis, the study first found that SR treatment significantly altered the gene expression profile (Fig. 4A), identifying a total of 2,956 upregulated and 534 downregulated genes (Fig. 4B). Enrichment analysis further showed that the differentially expressed genes were primarily concentrated in pathways related to inflammatory response and apoptotic processes, with the MAPK signaling pathway being markedly activated (Figs. 4C-D). Gene set enrichment analysis further confirmed that the MAPK pathway remained activated following SR (Fig. 4E). Subsequently, we conducted heatmap analysis (Figure S2A) on the top 15 DEGs enriched in the MAPK pathway and further validated their expression in THLE2 cells. The expression patterns of these DEGs were well consistent with transcriptomic profiles (Figure S2B), solidifying the reliability of transcriptome data. By further examining the phosphorylation levels of key proteins in the MAPK pathway, we found that SR treatment significantly upregulated the phosphorylation of JNK1, ERK, MKK4, and MKK7, indicating that SR effectively activates the MAPK signaling pathway. Notably, CA intervention specifically inhibited SR-induced JNK1 phosphorylation, but exerted no significant effect on the phosphorylation of ERK, MKK4, or MKK7 (Figs. 4F-G). The overactivation of this pathway may exacerbate inflammation and apoptosis, suggesting it as a potential key target for intervention. To further investigate the underlying mechanism, and based on the aforementioned omics findings, we focused on the MAPK pathway as a key research target and performed target screening using a pharmacophore library. The results indicated that among MAPK pathway-related proteins, CA exhibited the highest binding potential for JNK1 and MAPK1 (Table 3). Molecular docking analysis revealed that CA displayed a higher binding preference for JNK1, with a binding energy of -8.187 kcal/mol (Fig. 5A). Subsequently, subsequent surface plasmon resonance experiments demonstrated a strong binding affinity between CA and JNK1, with a KD value of 2.76 × 10− 6 M (Fig. 5B). Further 100 ns all-atom molecular dynamics simulations were performed, and the results showed that the CA molecule remained stably bound within the active binding pocket of JNK1 at four representative sampling time points (Fig. 5C), intuitively confirming its favorable initial binding affinity for JNK1. Simulations showed that the RMSD of the JNK1-CA complex fluctuated around 0.3 nm, the Rg remained stable at 2.18 nm, the number of hydrogen bonds stayed constant, and the overall RMSF was < 0.5 nm (Fig. 5D). These results indicate that the complex maintained a stable conformation with a rigid binding site, supporting stable binding between CA and JNK1. To identify the key amino acid residues mediating the interaction between CA and JNK1, based on binding energy contribution analysis, LEU-168 was identified as the key residue contributing most to ligand binding. Additionally, residues such as GLN-37 and VAL-158 also made significant contributions to the total binding energy (Fig. 5E). By performing alanine mutagenesis on the above sites, we found that mutation of LEU-168 induced the most significant change in the binding energy between CA and JNK1 (Fig. 5F). Therefore, wild-type JNK1 and the JNK1 LEU168 mutant were expressed and purified for cellular thermal shift assay (CETSA) (Fig. 5G). CA markedly enhanced the thermal stability of wild-type JNK1, with a melting temperature (Tm) shift of 4.31℃. In contrast, CA failed to alter the thermal stability of the JNK1 LEU-168 mutant after incubation. JNK1 kinase activity assays further verified that CA significantly inhibited the enzymatic activity of wild-type JNK1, while exerting no obvious effect on the LEU-168 mutant (Figs. 5H-I). Taken together, these results indicate that CA binds to and inhibits JNK1 activity, while it does not significantly modulate other upstream modulators of the MAPK pathway. JNK1 activation aggravates hepatic cold I/R injury under prolonged CIT in rats The JNK/c-Jun/c-Fos pathway has been reported to play a pivotal role in spinal cord injury [37]. The JNK agonist anisomycin [38] was used to investigate the association between the JNK1 and CA following cold I/R injury. HE staining revealed that livers in the model group, subjected to SCS followed by subsequent NMP, exhibited typical morphological features of cold I/R injury. In stark contrast, pathological alterations in liver tissue were significantly ameliorated in the CA-treated group. However, when anisomycin was administered in addition to CA treatment, the severity of histological damage markedly increased, characterized by the reappearance of extensive hepatocyte edema and sinusoidal congestion (Fig. 6A). Similarly, TUNEL and DHE staining confirmed that the reductions in cellular apoptosis and ROS levels achieved with CA treatment were abolished by anisomycin (Figs. 6B-C). Western blot analysis demonstrated that cold I/R induced phosphorylation of JNK1 was suppressed by CA. Nevertheless, this inhibitory effect of CA on JNK1 was counteracted upon anisomycin treatment, leading to a recurrence of aggravated liver injury and disruption of redox homeostasis (Figs. 6D-E). Activation of JNK1 aggravates hepatocyte injury under LT-like stress To further verify at the cellular level whether the protective effect of CA specifically depends on the inhibition of JNK1, this study employed a pharmacological antagonism assay. Using a SR model established in THLE2 cells, we examined whether anisomycin could reverse the effects of CA. First, regarding cellular injury markers, CA pretreatment significantly reduced the levels of AST, LDH, and ALT in the culture supernatant following SR stress. However, when anisomycin was co-administered with CA, the levels of these three enzymes showed a significant rebound (Fig. 7A). This result directly indicates that artificial activation of the JNK1 pathway can antagonize the protective effect of CA in mitigating liver injury, suggesting that CA is closely associated with the inhibition of JNK1 activity. Second, at the level of cell fate and oxidative stress, quantitative flow cytometry provided more in-depth evidence. CA effectively reduced the total apoptosis rate induced by SR, but co-treatment with anisomycin completely reversed this effect, restoring the apoptosis rate (Figs. 7B-C). Meanwhile, specific detection of mitochondrial superoxide using the MitoNeoD probe revealed that the inhibitory effect of CA on the SR induced ROS burst was counteracted by anisomycin (Fig. 7D). These data demonstrate that the protective effects of CA in regulating apoptosis and redox balance cannot be achieved when JNK1 is forcibly activated. As the most important intracellular antioxidant, changes in the level of GSH further support the above conclusions. SR treatment led to severe depletion of intracellular GSH, which was partially restored by CA intervention. However, anisomycin not only exacerbated the GSH depletion caused by SR alone but also completely blocked the estorative effect of CA on GSH levels (Fig. 7E). Western blot analysis similarly showed that cold I/R induced phosphorylation of JNK1 could be suppressed by CA, whereas this inhibition was counteracted by anisomycin, leading to a recurrence of aggravated liver injury and disruption of redox homeostasis (Figs. 7F).To rule out the off-target effects of anisomycin, we specifically constructed a JNK1 knockdown cell line (Fig. 7G). The results showed that the protective effect of CA against SR-induced liver injury was abolished after JNK1 knockdown(Fig. 7H). Consistent with the above findings, JNK1 silencing alone mimicked the anti-apoptotic and antioxidant effects of CA. Moreover, under JNK1 knockdown conditions, CA failed to further enhance the protective effects against cellular injury (Figures S3A-B). Taken together, these results suggest that CA exerts its protective effects specifically through the inhibition of JNK1 activity. Caftaric Acid modulates hepatic cold ischemia/reperfusion injury via the JNK1/FOS pathway Transcriptomic profiles of rat primary hepatocytes during SR, with or without CA intervention, were systematically compared to analyze the regulatory effects of CA at the gene level (Fig. 8A). Analysis of transcriptomic sequencing data showed that CA treatment altered the expression patterns of a cohort of key genes (Fig. 8B). To further identify the most significantly affected downstream gene of CA, we constructed an SR model again in THLE2 cells and analyzed the top 20 differentially expressed genes from the primary hepatocyte transcriptome. Specifically, in a THLE2 cell model of SR injury, subsequent qPCR analysis demonstrated significant upregulation of FOS, which acts downstream of the MAPK pathway. Notably, CA pretreatment significantly suppressed this SR induced elevation in FOS expression (Fig. 8C; S4A). Combined with our earlier finding that CA directly targets and inhibits JNK1 kinase activity, we propose a new scientific hypothesis: CA may exert its hepatoprotective effects by inhibiting the JNK1/FOS signaling axis, thereby intervening in downstream injury events. To directly verify the functional role of FOS in SR induced injury, we successfully established a stable FOS-overexpressing cell line in THLE2 cells and confirmed the stable high expression of FOS protein by Western blot (Fig. 8D). Subsequent functional assays strongly substantiated the pathogenic role of FOS. Under SR stress, compared with the control group, FOS-overexpressing cells showed a significant increase in the release of LDH, ALT, and LDH into the culture supernatant, directly reflecting more severe disruption of plasma membrane integrity. This enzymatic leakage was effectively suppressed by co-treatment with CA (Fig. 8E), preliminarily indicating that FOS overexpression exacerbates SR injury and that CA can counteract this effect. To further investigate the cytological mechanisms underlying this aggravated injury, we examined cell death and oxidative stress levels. PI staining revealed that FOS overexpression significantly increased the proportion of late apoptotic cells after SR. Concurrently, detection using the MitoNeoD probe showed a burst increase in ROS (Figs. 8F-G). Quantitative flow cytometric analysis provided precise data supporting these morphological observations (Figures S4B-C). Western blot analysis validated the impact of FOS overexpression on the expression of apoptosis and antioxidant-related proteins at the protein level and further demonstrated that CA could reverse these changes (Fig. 8H, Figures S4D). Notably, in both rat liver and THLE2 cells with activated JNK1, the phosphorylation level of FOS was further elevated (Fig. 8I, Figures S4E).Furthermore, GSH depletion was more rapid and severe in FOS-overexpressing cells, whereas CA significantly maintained intracellular GSH homeostasis (Fig. 8J). To further verify whether CA-mediated FOS inhibition depends on the specific regulation of JNK1, we constructed THLE2 hepatocyte cell lines with JNK1 knockdown and overexpression, respectively (Fig. 7G, Figure S4F). The results showed that FOS phosphorylation levels were significantly reduced upon JNK1 knockdown, while JNK1 overexpression led to a marked increase in FOS phosphorylation levels (Figs. 8K-L). T The JNK1/FOS regulatory axis may serve as an indicator for post-LT EAD Finally, we validated the JNK1/FOS signaling axis using clinical human liver transplantation samples collected at our center. Western blot analysis revealed that the protein expression levels of p-JNK1 and p-FOS were significantly upregulated in post-transplant liver tissues compared with pre-transplant samples, indicating that this signaling pathway is activated in the early post-transplant period (Fig. 9A-B). Further analysis of liver tissues from patients who developed EAD showed that the expression levels of both p-JNK1 and p-FOS were markedly higher in the EAD group than in the immediate graft function (IGF) group (Fig. 9C-D). Meanwhile, we used receiver operating characteristic (ROC) curves to evaluate the ability of intrahepatic JNK1 and FOS mRNA levels to discriminate between early allograft dysfunction (EAD) and immediate graft function (IGF) after liver transplantation. As shown in Fig. 9E, the area under the curve (AUC) for JNK1 was 0.8168, and for FOS, the AUC was 0.8549. The results indicate that JNK1/FOS exhibited moderate diagnostic value for predicting EAD. These results demonstrate that elevated p-JNK1 and p-FOS are closely associated with post-transplant liver injury and the occurrence of EAD, suggesting their potential as injury-related biomarkers that could aid in the early identification and intervention of postoperative EAD risk. Discussion Hepatic cold I/R injury is an inevitable stage of LT and is closely associated with graft injury and postoperative adverse outcomes [39]. Previous studies have shown that certain metabolites in hibernators enable organ protection against cold I/R injury during periodic torpor, during which the body temperature drops from 37 °C to 4 °C [27, 36, 40]. CA, a unique metabolite found in Syrian hamsters, can effectively protect hepatocytes from cold I/R injury. To this end, an isolated perfused rat liver model and a cold I/R model with prolonged CIT were established in this study. The supplementation of CA in UW and perfusion solutions significantly alleviated LT-like injury and promoted the restoration of redox homeostasis. Furthermore, CA protected the liver from cold I/R injury via the JNK1/FOS pathway. Finally, the predictive role of this axis in postoperative EAD was confirmed in liver tissues after LT. SCS has long been a common method for the preservation and transportation of transplanted organs; nevertheless, tissue and organ viability decline rapidly with prolonged CIT [41–44]. For example, the ideal CIT for donor hearts is < 4 h [45], and that for donor livers generally does not exceed 6–10 h [6, 7]. In contrast, the CIT for donor kidneys can be appropriately extended [46]. A recent study identified that cold ischemia stimulation activates ferroptosis-related signaling pathways in human lung tissues, thereby promoting injury after lung transplantation [47]. Therefore, intervention during SCS is a crucial measure for alleviating hepatic cold I/R injury. Supplementation of apoptosis-related siRNA during SCS of the liver has been shown to considerably reduce postoperative hepatic enzyme levels and proinflammatory cytokine levels [48]. Another study reported that the cold tolerance of hibernating animals depends on their effective utilization of α-tocopherol to resist lipid peroxidation and ferroptosis [49]. Furthermore, Wang et al. noted that a protocol combining crystalloid cardioplegia with UW preservation solution reduced the activation of NF-κB–related inflammatory pathways. Moreover, this method decreased cardiomyocyte apoptosis, ROS production, and mitochondrial damage; extended the cold ischemia tolerance of donor hearts; and improved post-transplant cardiac function [50]. In previous studies, isolated rat livers had been found to exhibit superior cold tolerance compared to human livers [51, 52]According to Bruinsma et al., prolonged CIT of > 48 h in rat livers substantially impairs LT outcomes [51]. Therefore, 48 h was selected as the intervention time point in this study to verify the effect of CA in alleviating cold I/R injury. As expected, CA effectively reduced the area of hepatic necrosis and improved liver function in rat livers subjected to prolonged CIT. In addition, although various clinical centers have adopted multiple preservation strategies to address cold ischemia/reperfusion injury—including NMP, HOPE, and ischemia-free liver transplantation, which have been shown to evaluate and enhance organ utilization and improve prognostic outcomes in liver transplantation—whether CA remains stable and exerts comparable protective effects under these conditions warrants further investigation. JNK1, a conserved component of the MAPK signaling pathway, is a key regulator of several cellular events [53, 54]. Numerous studies have demonstrated that the activation of the JNK pathway significantly influences the induction of apoptosis and the generation of oxidative stress. Specifically, in primary mouse embryonic fibroblasts, deficiency of JNK1 protects these cells from apoptosis induced by ultraviolet stimulation [55]. In contrast, the excessive activation of JNK1 mediates oxidative stress induced by menadione and alcohol, ultimately leading to hepatocyte death [56, 57]. Yin et al. demonstrated that during liver I/R, supplementation of nicotinamide N-methyltransferase/N1-methylnicotinamide can regulate hepatic injury via the ANGPT2/JNK signaling pathway [58]. Interestingly, the findings from the present study suggest that CA can inhibit the activation of its downstream pathways by directly interacting with JNK1 and that this effect is achieved by suppressing the phosphorylation of JNK1. We further performed JNK pathway activation rescue experiments with anisomycin, and results showed JNK activation significantly abolished CA-mediated protective effects. Although anisomycin, a classic JNK agonist, broadly activates JNK family proteins without absolute isoform specificity, it is a well-established tool for JNK pathway activation and functional rescue [38, 59]. However, anisomycin has been used to specifically activate JNK1, thereby promoting the phosphorylation of ZDHHC14 at the Thr440 site and inhibiting its palmitoylation modification [60]. Importantly, we did not rely solely on anisomycin; combined with JNK1 knockdown, overexpression, LEU168 site mutation and kinase activity detection, we confirmed CA directly and specifically targets JNK1, excluding anisomycin’s potential off-target effects and non-specific stress interference. SPR is widely used to study natural product-target interactions [61], with micromolar KD values accepted as evidence of effective binding. For example, reported KD values in similar studies ranged from 5.59 × 10− 6 M to 323.4 µM [62–65]-far higher than our 2.76 µM-yet were still validated by functional assays. Thus, our KD, combined with multi-layered functional evidence, supports in vivo inhibitory efficacy. FOS is a canonical transcription factor that directly binds to the promoters of downstream target genes and widely regulates the transcription of genes involved in apoptosis, redox homeostasis and oxidative stress. Previous studies have shown that the downregulation of FOS inhibits the transcription of USP53, thereby promoting the ubiquitination and degradation of the key necroptosis protein MLKL and the core pyroptosis protein GSDMD [66]. This cascade ultimately alleviates necroptosis and pyroptosis in damaged cells. Additionally, aberrant upregulation of FOS has been proven to aggravate neuronal apoptosis and mitochondrial damage during myopia progression [67]. In the progression of intervertebral disc degeneration, oxidative stress and excessive apoptosis of nucleus pulposus cells are critical pathogenic factors. Inhibition of FOS can effectively relieve oxidative stress and cell apoptosis, thereby delaying intervertebral disc degeneration [68]. Consistent with these reported findings, our functional experiments confirmed that FOS overexpression markedly exacerbates hepatocellular oxidative stress and apoptosis under cold stress, and ultimately aggravates liver cold injury. Due to the lack of reliable CA detection methods, ethical constraints, and limited resources, direct clinical validation of CA in liver transplant recipients remains challenging. Therefore, we only the indicative role of the JNK1/FOS pathway in clinical prognosis was verified, without assessing CA’s specific protective effects on clinical outcomes. This study focused on elucidating CA’s protective mechanism in experimental models. Comprehensive pharmacokinetic and dose-response analyses were not performed, which is a major limitation. We acknowledge that entropy was not included in the MM/GBSA calculations, which may affect the quantitative interpretation of binding free energy and represents a limitation of this study. Collectively, the findings from this study establish that CA can significantly ameliorate hepatocellular injury and mitochondrial ROS imbalance during cold I/R. The underlying mechanism is likely to involve improvements in mitochondrial redox homeostasis and the inhibition of hepatocellular apoptosis by suppressing the JNK1/FOS signaling axis. Electronic Supplementary Material Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1