ORIGINAL ARTICLE 原文

The E3 ubiquitin ligase NEDD4-1 protects against acetaminophen-induced liver injury by targeting VDAC1 for degradation
E3 泛素连接酶 NEDD4-1 通过靶向 VDAC1 降解来防止对乙酰氨基酚诱导的肝损伤

2.1. Animals and treatment
2.1. 动物和治疗

Experimental animal protocols were approved by the Ethics Committee for Animal Care and Use of Jilin University (permit number: SY202008007; Changchun, China). The animals were treated humanely according to Laboratory animal-the guidelines for ethical review of animal welfare (GB/T 35,892–2018, China). Wild-type (WT) C57BL/6 N mice were purchased from Cyagen Biosciences (Suzhou, China). Mice with the floxed Nedd4-1 alleles (Flox; on a C57BL/6 N background) were also purchased from Cyagen Biosciences. Hepatocyte-specific Nedd4-1 knockout (HepKO; on a C57BL/6 N background) mice were generated by mating Flox mice with albumin-Cre mice (Cyagen Biosciences). Although it is now increasingly recognized that crossing Flox mice with albumin-Cre mice will not generate pure hepatocyte specific knockout mice since albumin is expressed in bipotent fetal liver progenitor cells that give rise to both hepatocytes and biliary cells²⁸, albumin-Cre mice is still widely used and acceptable for generating hepatocyte specific knockout mice today29, 30, 31. All experiments described in this report used male mice aged 4–8 weeks. Mice were housed in a temperature (23 ± 2 °C) and humidity (40 ± 5%) controlled specific pathogen-free facility with a 12 h/12 h dark/light photocycle. Food and water were provided ad libitum.
实验动物实验方案由吉林大学动物护理与使用伦理委员会批准（许可证号：SY202008007;长春，中国）。根据《实验动物-动物福利伦理审查指南》（GB/T 35,892–2018，中国）对动物进行人道对待。野生型（WT）C57BL/6 N小鼠购自Cyagen Biosciences（中国苏州）。具有floxed Nedd4-1等位基因（Flox;在C57BL / 6 N背景上）的小鼠也从Cyagen Biosciences购买。通过将Flox小鼠与白蛋白-Cre小鼠交配产生肝细胞特异性Nedd4-1敲除（HepKO;在C57BL / 6 N背景上）小鼠（Cyagen Biosciences）。尽管现在人们越来越认识到，将Flox小鼠与白蛋白-Cre小鼠杂交不会产生纯肝细胞特异性敲除小鼠，因为白蛋白在产生肝细胞和胆细胞 ²⁸ 的双能胎儿肝祖细胞中表达，白蛋白-Cre小鼠今天仍然被广泛使用和接受用于产生肝细胞特异性敲除小鼠29， 30， 31.本报告中描述的所有实验都使用4-8周龄的雄性小鼠。将小鼠饲养在温度（23±2°C）和湿度（40±5%）控制的无特定病原体设施中，具有12小时/ 12小时的暗/光光循环。食物和水是随意提供的。

For modeling AILI mouse models, mice were fasted overnight to reduce hepatic GSH levels before APAP injection. APAP (IA0030; Solarbio, Beijing, China) was dissolved in warm saline (0.9% NaCl; 55 °C), cooled to 37 °C, and injected intraperitoneally at a dose of 300 or 600 mg/kg. Animals were sacrificed 6, 12, or 24 h after APAP administration. Liver and serum samples at each indicated time point were collected for further experiments.
为了模拟AILI小鼠模型，小鼠在APAP注射前禁食过夜以降低肝脏GSH水平。APAP （IA0030;Solarbio，Beijing，China）溶于温盐水（0.9%NaCl;55°C）中，冷却至37°C，并以300或600mg/kg的剂量腹膜内注射。在APAP给药后6、12或24小时处死动物。收集每个指定时间点的肝脏和血清样本以进行进一步实验。

2.2. Analysis of serum parameters
2.2. 血清参数分析

Serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), lactate dehydrogenase (LDH) activities, and high mobility group box 1 (HMGB1) concentration were measured using the ALT activity assay kit (MAK052; Sigma–Aldrich, St. Louis, MO, USA), AST activity assay kit (MAK055; Sigma–Aldrich), LDH activity assay kit (MAK066; Sigma–Aldrich), and mouse HMGB1 ELISA kit (NBP2-62767; Novus Biologicals, Littleton, CO, USA) according to the manufacturer's protocols, respectively.
使用 ALT 活性测定试剂盒 （MAK052） 测定血清丙氨酸氨基转移酶 （ALT）、天冬氨酸氨基转移酶 （AST）、乳酸脱氢酶 （LDH） 活性和高迁移率组框 1 （HMGB1） 浓度;Sigma–Aldrich， St. Louis， MO， USA）、AST 活性测定试剂盒 （MAK055;Sigma-Aldrich）、LDH活性测定试剂盒（MAK066;Sigma-Aldrich）和小鼠HMGB1 ELISA试剂盒（NBP2-62767;Novus Biologicals，Littleton，CO，USA）分别根据制造商的协议。

2.3. Hematoxylin and eosin (H&E) and immunohistochemical (IHC) staining
2.3. 苏木精和伊红（H&E）和免疫组化（IHC）染色

H&E and IHC staining were performed in paraffin-embedded liver sections. Briefly, liver samples were fixed in 4% paraformaldehyde overnight, dehydrated, cleared, and embedded in paraffin. The waxes were sectioned serially at 3–5 μm thickness. After deparaffinization and rehydration, standard H&E staining was carried out to visualize the pattern in necrotic areas of the liver. The VDAC1 expression profiles in mice liver samples were determined by incubating the liver sections with anti-VDAC1 primary antibodies. Histological features were observed and captured under a light microscope (Olympus Corporation, Tokyo, Japan). The details of antibodies are listed in Supporting Information Table S1.
H&E和IHC染色在石蜡包埋的肝脏切片中进行。简而言之，将肝脏样品固定在4%多聚甲醛中过夜，脱水，澄清并包埋在石蜡中。蜡以 3-5 μm 的厚度连续切片。脱蜡和复液后，进行标准H&E染色以可视化肝脏坏死区域的模式。通过将肝切片与抗 VDAC1 一抗孵育来测定小鼠肝脏样品中的 VDAC1 表达谱。在光学显微镜下观察和捕获组织学特征（奥林巴斯公司，日本东京）。抗体的详细信息列在支持信息表 S1 中。

2.4. Protein extraction and western blotting

Total protein was extracted from liver tissue or cells, and prepared for Western blotting according to manufacturer's instructions (P0013K; Beyotime Institute of Biotechnology, Shanghai, China). Mitochondrial and cytosolic fractions used for Western blotting were isolated from fresh liver tissue or hepatocytes using a mitochondria isolation kit (MP-007; Invent Biotechnologies, Eden Prairie, MN, USA) according to manufacturer's instructions. Protein abundance of each molecular was determined by Western blotting as previously described³². The details of antibodies are listed in Table S1.

2.5. RNA extraction and quantitative real-time polymerase chain reaction

Protocol for quantitative real-time polymerase chain reaction was the same as described previously^32,33. Briefly, total RNA was extracted using RNAiso Plus (D9108, TaKaRa Biotechnology, Dalian, China) according to manufacturer's instructions. Complementary DNA was reverse-transcribed from 1 μg total RNA using a reverse transcription kit (RR047A, TaKaRa Biotechnology). Relative mRNA expression of target genes was detected using the TB Green Premix Ex Taq II (RR82LR, TaKaRa Biotechnology) on the 7500 Real-Time PCR System (Applied Biosystems, Bedford, MA, USA). The relative expression of target genes was normalized to reference gene (Actb) using 2^−ΔΔCT method³⁴. Primer sequences are listed in Supporting Information Table S2.

2.6. Cell isolation, culture, and treatment

Mouse primary hepatocytes were isolated from six- to 8-week-old male mice (C57BL/6 N background) as previously described^35,36, with modification. Briefly, mice were anesthetized and perfused in situ with D-HBSS (PB180322; Procell Life Science & Technology, Wuhan, China) containing 5 mmol/L EDTA and then HBSS (PB130824; Procell Life Science & Technology) containing 0.5 mg/mL collagenase type IV (17104019; Gibco, Grand Island, NY, USA) via the portal vein. The cell suspension was filtered through a 70-μm cell strainer (431,751; Corning, NY, USA) and centrifuged at 50 × g for 5 min at 4 °C. After that, the cell pellets were resuspended and mounted over a cushion of 25% Percoll (17089101; Cytiva Sweden AB, Uppsala, Sweden) to eliminate non-parenchymal cells. The obtained hepatocytes were washed with DMEM (SH30021.01; HyClone, Logan, UT, USA) and then cultured with William's E medium (PM151213; Procell Life Science & Technology) containing 10% fetal bovine serum (FBS; FB15015; Clark Bioscience, Richmond, VA, USA) in a 37 °C incubator with 5% CO₂ and humidified atmosphere. Endothelial cells, cholangiocytes and stromal cells were isolated based on bilio-vascular tree isolation procedure as previously described³⁷. In situ perfusion of the liver was performed as described above. The digested liver was acquired and placed in pre-chilled DMEM. Liver parenchyma was mechanically detached and discarded. The remaining bilio-vascular tree was minced and digested in DMEM containing 0.075 mg/mL collagenase P (11213865001; Roche, Indianapolis, IN, USA), 0.02 mg/mL DNase I (D8071; Solarbio, Beijing, China), 3% FBS, 1 mg/mL bovine serum albumin, 1% HEPES and 1% Penicillin–Streptomycin at 37 °C. The digested bilio-vascular fragments were collected on top of a 40-μm cell strainer (352340; Corning) and digested in 0.05% Trypsin–EDTA containing 0.02 mg/mL DNase I at 37 °C. Endothelial cells and cholangiocytes were isolated through a magnetic selection method using CD31 microbeads (130-097-418; Miltenyi Biotec, San Diego, CA, USA) and EpCAM microbeads (130-105-958; Miltenyi Biotec), respectively. Enriched stromal cell fraction was collected after excluding endothelial cells and cholangiocytes. 293 T cell lines were purchased from the Cell Bank of Type Cultural Collection of Chinese Academy of Sciences (Shanghai, China), and cultured in DMEM (SH40007.01; HyClone) supplemented with 10% FBS at a 37 °C incubator with 5% CO₂.

To detect the effect of APAP on NEDD4-1 in vitro, primary hepatocytes were treated with different concentrations (0, 5, 10, or 20 mmol/L) of APAP for 12 h, or 10 mmol/L APAP for different times (0, 3, 6, or 12 h). To mimic APAP-induced hepatotoxicity in vitro, primary hepatocytes were treated with 10 mmol/L APAP for 12 h. To overexpress NEDD4-1 or knockdown VDAC1 in vitro, hepatocytes were infected with adenovirus expressing mouse Nedd4-1 gene sequence (Ad-NEDD4-1) or adenovirus expressing short hairpin RNA (shRNA) targeting mouse Vdac1 gene (Ad-shVDAC1) at a multiplicity of infection of 50 for 36–48 h, respectively. To inhibit the lysosomal degradation pathway of proteins, hepatocytes were treated with lysosome inhibitor chloroquine (CQ; 50 μmol/L) for 6 h before harvest. To inhibit the proteasomal degradation pathway of proteins, hepatocytes were treated with proteasome inhibitor MG132 (10 μmol/L) for 10 h before harvest. To inhibit c-Jun N-terminal kinase (JNK) signaling, hepatocytes were pretreated with 20 μmol/L SP600125 2 h before APAP treatment. For coimmunoprecipitation assays, 293 T cells were transfected with the indicated plasmids for 24–48 h using Lipofectamine 3000 (L3000015; Invitrogen, Carlsbad, CA, USA).

2.7. Immunofluorescence assay

For immunofluorescence, cells were fixed with 4% paraformaldehyde at room temperature for 15 min, washed twice with PBS, and then permeabilized with 0.1% Triton X-100 for 1 h. After washing with PBS twice, cells were incubated with primary antibodies diluted in PBS containing 5% bovine serum albumin overnight at 4 °C. Cells were then washed three times for 5 min with Tris-buffered saline–Tween 20, and incubated with corresponding fluorescent secondary antibodies. Subsequently, cells were washed twice with Tris-buffered saline–Tween 20, and then stained with DAPI. Cells were observed by a laser scanning confocal microscopy (Olympus FluoView 1200; Olympus Corporation). Fluorescence line profile and co-localization analyses were performed with Fiji software³⁸. The details of antibodies are listed in Table S1.

2.8. RNA sequencing (RNA-seq) and data processing

For the RNA-seq analysis, total RNA was extracted from liver samples of APAP-treated Flox or APAP-treated HepKO mice. Libraries were constructed from total RNA and paired-end sequencing was performed on an Illumina Novaseq 6000 (Illumina, San Diego, CA, USA). Raw sequencing data were quality-controlled with Cutadapt and FastQC program. Reads were matched to reference genome sequences (mm10/GRCm38) using HISAT2 (version 2.1.0). Fragment per kilobase of transcript per million mapped reads mapped value was estimated by StringTie (version 1.3.4 d). Differential gene expression analysis was performed using DEseq2 (version 1.2.10) according to the following two criteria: (1) folding change ≥2 and (2) false discovery rate <0.05. The hierarchical clustering analysis was performed to analyze global sample distribution profiles by constructing a clustering tree based on data from RNA-seq by R function hclust. Gene set enrichment analysis (GSEA) was performed using the R package ReactomePA (version 1.30.0) with default parameters. Gene sets with P values < 0.05 were considered statistically significant. RNA-seq data generated in this study have been deposited in the Gene Expression Omnibus public database under accession number GSE199353.

2.9. Measurement of viable and dead cells

Calcein-AM/propidium iodide (PI) double staining kit (40747ES76; Yeasen Biotechnology, Shanghai, China) was utilized for simultaneous fluorescence staining of viable and dead cells. This kit contains Calcein-AM and PI solutions, which stain viable and dead cells, respectively. After treatment as indicated in results, cells were stained according to manufacturer's instruction. Firstly, washing cells with 1 × assay buffer three times to remove residual esterase activity. After that, cells were incubated with the working solution (2 μmol/L Calcein-AM and 4.5 μmol/L PI in 1 × assay buffer) for 20 min at 37 °C, protected from light. Following incubation, cells were washed gently with 1 × assay buffer three times, and then visualized by a laser scanning confocal microscopy (Olympus FluoView 1200).

2.10. Adeno-associated virus 8 (AAV8) injection

The AAV8 delivery system was constructed by Vigene Bioscience (Jinan, China). For NEDD4-1 overexpression, mice were injected with AAV8-TBG (a genetic delivery system containing hepatocyte-specific promoter TBG and exhibiting high specificity and transfection efficiency in hepatocytes) carrying the full length mouse Nedd4-1 gene sequence (AAV8-NEDD4-1). For VDAC1 knockdown, mice were injected with AAV8 vectors bearing shRNA targeting mouse Vdac1 gene (AAV8-shVDAC1). Mice were injected in the lateral tail vein with 200 μL of virus containing 2.5 × 10¹¹ AAV8 vector genomes. The shVdac1 sequences are listed in Supporting Information Table S3.

2.11. Transmission electron microscopy

Primary hepatocytes were fixed in 2.5% glutaraldehyde overnight at room temperature. Then, cells were post-fixed in potassium ferrocyanide–osmium tetroxide for 1 h and dehydrated in a bath with increasing concentrations of ethanol. After embedding and slicing, cells were staining with uranyl acetate and lead citrate. Sections were observed under a H-7650 transmission electron microscope (Hitachi Limited, Tokyo, Japan).

2.12. Detection of liver reactive oxygen species (ROS) content and ATP content

The content of ROS in liver tissue was measured using the tissue ROS test kit (HR8821; BioRab, Beijing, China) according to manufacturer's instructions and normalized to protein concentration. Briefly, fresh liver tissue was homogenized in cold buffer and centrifuged (100 × g) for 5 min at 4 °C. The supernatant was used to analyze ROS content. Add 200 μL of supernatant and 2 μL of dihydroethidium probe to a 96-well plate, and mix them well. Plate was incubated at 37 °C for 30 min in dark, and then read on a Molecular Devices SpectraMax Gemini EM (Molecular Devices, Sunnyvale, CA, USA) at λ_exc = 510 nm and λ_emi = 610 nm.

The content of ATP in liver tissue was measured using the ATP colorimetric/fluorometric assay kit (K354-100; BioVision, Milpitas, CA, USA) according to manufacturer's instructions and normalized to protein concentration.

2.13. Seahorse assay

Oxygen consumption rate (OCR) was measured using the Seahorse XFe 24 Analyzer (Agilent Technologies, Santa Clara, CA, USA). 3 × 10⁴ hepatocytes per well were plated as a monolayer culture in Seahorse cell plates. To perform the mitochondrial stress test, 2 μmol/L oligomycin, 1.5 μmol/L FCCP, and 1 μmol/L antimycin A + 1 μmol/L rotenone were added as indicated.

2.14. Measurement of intracellular mitochondrial membrane potential and mitochondrial ROS (mtROS)

Intracellular mitochondrial membrane potential and mtROS production were visualized using the JC-1 mitochondrial membrane potential assay kit (C0008; Applygen Technologies, Beijing, China) or the MitoSOX Red Mitochondrial Superoxide Indicator (40778ES50; Yeasen Biotechnology) according to manufacturer's instruction, respectively. To measure intracellular mitochondrial membrane potential, cells were incubated with 10 μg/mL of JC-1 working solution for 20 min at 37 °C, protected from light. To measure intracellular mtROS, cells were incubated with 4.5 μmol/L of MitoSOX working solution for 10 min at 37 °C, protected from light. After incubation, cells were washed gently three times with warm buffer, and then visualized by a laser scanning confocal microscopy (Olympus FluoView 1200).

2.15. VDAC1 cross-linking assay

To detect the oligomeric forms of VDAC1, chemical cross-linking of cells was performed as described previously^39,40. Briefly, hepatocytes were washed twice with PBS, harvested, and incubated with 0.5 mmol/L EGS (21565, ThermoScientific, Waltham, MA, USA) in PBS (pH 7.4) at 30 °C for 20 min. The pellet was lysed in NP-40 lysis buffer by sonication on ice. Samples were subjected to SDS-PAGE and immunoblotting using anti-VDAC1 antibody.

2.16. Plasmid construction

WT and truncated fragments of mouse Nedd4-1 were amplified from mouse complementary DNA and cloned into the pcDNA3.1-HA vector. WT fragments of mouse Vdac1 were amplified from mouse complementary DNA and cloned into the pEnCMV-3 × Flag vector. Mutant fragment of mouse NEDD4-1 (NEDD4-1^C854S) and PPTY deletion fragment of mouse VDAC1 (VDAC1^△PPTY) were generated using Q5 Site-Directed Mutagenesis Kit (E0554S; New England BioLabs, Ipswich, MA, USA). Each construction was verified by sequencing (Comate Bioscience, Changchun, China).

2.17. Coimmunoprecipitation (Co-IP) assay

293 T cells were lysed in ice-cold NP40 lysis buffer supplemented with a protease inhibitor cocktail, and centrifuged (12,000 × g) at 4 °C for 10 min. The cell lysates were used for immunoprecipitation with Pierce Crosslink Immunoprecipitation Kit (26,147, Thermo Scientific), anti-Flag M2 affinity gel (A2220, Sigma–Aldrich), or anti-HA affinity gel (E6779, Sigma–Aldrich). The obtained beads were washed five times with lysis buffer, and the immunoprecipitates were obtained by boiling in 2 × loading buffer at 95 °C for 5 min, or eluted by 100 μg/mL HA peptide (I2149; Sigma–Aldrich). Finally, the immunoprecipitates and whole cell lysates were subjected to immunoblotting using the indicated primary antibodies and corresponding secondary antibodies. The details of antibodies are listed in Table S1.

2.18. Measurement of cell viability

Cell viability was detected using cell counting kit-8 (C0038; Beyotime Institute of Biotechnology) according to manufacturer's instructions.

2.19. Statistical analysis

Data were generated from multiple repeats of different biological experiments and were analyzed using the appropriate statistical analysis methods. Statistical significance of Kaplan–Meier survival curves was analyzed by log-rank test. The correlation between two parameters was analyzed by the Pearson correlation analysis. Statistical significance of the rest of data was calculated using the two-tailed unpaired Student t test, or ANOVA with Bonferroni correction. Results are presented as the mean ± standard error of mean (SEM). P value < 0.05 is considered statistically significant. Statistics were performed using GraphPad Prism software (Version 6; GraphPad Software Inc., La Jolla, CA, USA) and Statistical Package for the Social Sciences (SPSS) software (Version 19.0; SPSS Incorporated, Chicago, IL, USA).

3.1. Hepatocyte NEDD4-1 expression is dramatically downregulated in AILI and negatively correlates with liver injury severity

To explore the potential role of NEDD4-1 in AILI, six- to 8-week-old male C57BL/6 N mice were administered either a nonlethal (300 mg/kg) or lethal (600 mg/kg) dose of APAP over different durations by intraperitoneal injection. Administration of APAP resulted in hepatic injury, as indicated by highly increased serum ALT and AST activities (Supporting Information Fig. S1A and S1B). Typical features of liver injury and centrilobular hepatocellular necrosis were also observed by H&E staining in APAP-treated mice (Fig. S1C).

The protein (Fig. 1A and B) and mRNA (Fig. 1C) abundances of NEDD4-1 were markedly reduced in time- and dose-dependent manners in the livers of APAP-challenged mice. In addition, hepatic Nedd4-1 mRNA expression was negatively correlated with serum ALT and AST activities, as determined by Pearson correlation analysis (Fig. 1D and E). Due to the cellular heterogeneity of liver tissue, we analyzed a publicly available single-cell RNA-seq dataset derived from normal human livers (GSE192742; https://www.livercellatlas.org/index.php)⁴¹ and found that NEDD4-1 was predominantly expressed in hepatocytes and, to a lesser extent, in cholangiocytes, endothelial cells, and stromal cells (Fig. S1D). These four types of cells were thus isolated from the livers of mice administered saline or APAP. The NEDD4-1 protein abundance was lower in hepatocytes but not in other cell types from APAP-treated mice (Fig. S1E), suggesting that APAP overdose selectively decreases the abundance of NEDD4-1 in hepatocytes. Similarly, in vitro, APAP treatment downregulated NEDD4-1 expression in WT hepatocytes in time- and dose-dependent manners (Fig. 1F and G; Fig. S1F). The decrease of NEDD4-1 abundance in APAP-treated mouse primary hepatocytes was also confirmed by immunofluorescence analyses (Fig. 1H). These results suggest a potential regulatory role of NEDD4-1 in APAP-induced hepatotoxicity.

1. Download : Download high-res image (761KB)
2. Download : Download full-size image

Figure 1. Hepatocyte NEDD4-1 expression is dramatically downregulated in AILI and negatively correlates with liver injury severity. (A, B) Representative Western blots of indicated proteins in livers of mice. (C) The mRNA expression of Nedd4-1 in livers of mice. (D, E) Pearson correlation analyses showing the correlations between Nedd4-1 mRNA level and ALT activity (D), or AST activity (E); n = 38. (F, G) Representative Western blots of indicated proteins in primary hepatocytes treated with 10 mmol/L APAP for different times (3, six or 12 h), or different concentrations (5, 10 or 20 mmol/L) of APAP for 12 h. (H) Representative confocal images of primary hepatocytes treated with vehicle or 10 mmol/L APAP for 12 h. Scar bar, 10 μm. Experiments A–E were performed on mice treated with saline, or 300 mg/kg APAP for 6, 12 or 24 h, or 600 mg/kg APAP for 24 h; n = 6–9 per group. Data are presented as the mean ± SEM; ∗P < 0.05.

3.2. Hepatocyte NEDD4-1 deficiency exacerbates APAP-induced liver injury and necrosis

To understand the role of NEDD4-1 in the liver and the mechanism by which its downregulation contributes to AILI pathogenesis, we generated mice with hepatocyte-specific Nedd4-1 knockout (hereafter referred to as HepKO mice) (Supporting Information Fig. S2A). Western blot analysis confirmed the specific knockout of NEDD4-1 in the livers of HepKO mice (Fig. S2B). Unchallenged HepKO mice were healthy and fertile without noticeable gross phenotypes and demonstrated no significant differences in body weight (Fig. S2C) and the liver weight/body weight ratio (Fig. S2D) compared with those of their Flox counterparts. However, under challenge with 300 mg/kg APAP, NEDD4-1 loss led to more severe liver injury, as indicated by the higher serum ALT (Fig. 2A), AST (Fig. 2B), and LDH (Fig. 2C) activities. In addition, hepatocyte NEDD4-1 deficiency evidently exacerbated the APAP-induced increase in the serum level of HMGB1 (Fig. 2D), a nuclear protein released extracellularly from necrotic cells42, 43, 44. H&E staining (Fig. 2E) and evaluation of gross liver appearance (Fig. S2E) further confirmed the enhancement of APAP-induced liver injury and necrosis in APAP-treated HepKO mice. In particular, HepKO mice had a much lower survival rate than Flox mice after treatment with a lethal dose (600 mg/kg) of APAP (Fig. 2F). In addition, identical protein abundances of CYP2E1 (Fig. S2F) and GSH levels (Fig. S2G) were observed in the livers of APAP-treated Flox and HepKO mice, indicating unaffected APAP metabolism. Taken together, these results indicate that hepatocyte-specific NEDD4-1 deficiency exacerbates APAP-induced liver injury and necrosis.

1. Download : Download high-res image (1MB)
2. Download : Download full-size image

Figure 2. Hepatocyte-specific deficiency of NEDD4-1 exacerbates APAP induced hepatotoxicity in vivo and in vitro. (A–D) Serum ALT (A), AST (B), LDH (C) activities, and HMGB1 levels (D) of mice. (E) Representative H&E-stained images (left) and necrotic area quantification (right) of mouse liver sections. Scale bar, 200 μm. (F) Survival rate of Flox (n = 18) or HepKO (n = 20) mice treated with 600 mg/kg APAP for 72 h. (G, H) RNA-seq was conducted on total RNA isolated from livers of Flox or HepKO mice treated with 300 mg/kg APAP for 24 h; n = 3 per group. Volcano plot (G) shows the differentially expressed genes (1201 genes were up-regulated and 736 genes were down-regulated in APAP-treated HepKO mice). GSEA (H) shows the gene sets that the differentially expressed genes significantly enriched. (I) Representative images (left) and quantification (right) of Calcein-AM (green) and PI (red) double-stained primary hepatocytes. Scale bar, 200 μm. (J, K) LDH release (J) and HMGB1 level (K) in the culture medium of primary hepatocytes. Experiments A–E were performed on Flox or HepKO mice treated with saline or 300 mg/kg APAP for 24 h; n = 5–8 per group. Experiments I–K were performed on primary hepatocytes isolated from Flox or HepKO mice, with vehicle or APAP (10 mmol/L for 12 h) treatment. Data are presented as the mean ± SEM; ∗P < 0.05, ∗∗P < 0.01.

To gain better insight into the pathological alterations initiated by hepatocyte NEDD4-1 deficiency in the AILI setting, we performed RNA-seq using liver samples from APAP-treated Flox and HepKO mice. Unsupervised hierarchical clustering showed that liver samples from these two groups were clearly separated (Supporting Information Fig. S3A), with 1201 upregulated and 736 downregulated genes (Fig. 2G, Supporting Information Table S4) in the APAP-treated HepKO group. GSEA revealed that the upregulated genes were associated with the chemokine, inflammasome, DNA damage, tumor necrosis factor signaling, and cell death pathways (Fig. 2H, red; Fig. S3B). In addition, mitochondrial function-related genes associated with fatty acid oxidation and mitochondrial translation were downregulated (Fig. 2H, gray; Fig. S3B). These RNA-seq data further demonstrate the negative effects of NEDD4-1 deficiency on AILI.

For further in vitro experiments, hepatocytes were isolated from HepKO and Flox mice (Supporting Information Fig. S4) and were then exposed to APAP. NEDD4-1 deletion in hepatocytes markedly exacerbated APAP-induced hepatocyte death, as determined by Calcein-AM/PI double staining (Fig. 2I), an LDH release assay (Fig. 2J) and measurement of the HMGB1 concentration in the culture medium (Fig. 2K). Taken together, these in vivo and in vitro results reveal that loss of hepatocyte NEDD4-1 exacerbates APAP-induced liver injury and necrosis.

3.3. Overexpression of hepatocyte NEDD4-1 alleviates APAP-induced liver injury and necrosis

To further confirm the function of hepatocyte NEDD4-1 in AILI, we generated mice with hepatocyte NEDD4-1 overexpression using AAV8 delivering hepatocyte-specific promoter TBG and full-length Nedd4-1 gene sequence (AAV8-NEDD4-1), and examined their response to APAP treatment (Supporting Information Fig. S5A). Western blot analysis confirmed the overexpression of NEDD4-1 in the livers of AAV8-NEDD4-1-injected mice (Fig. S5B). The similar levels of CYP2E1 protein abundance (Fig. S5C) and GSH depletion (Fig. S5D) indicated that APAP metabolism was not influenced by AAV8-NEDD4-1 injection. Compared to AAV8-Control-injected mice, AAV8-NEDD4-1-injected mice showed alleviation of liver injury and necrosis in response to APAP challenge, as indicated by the decreased necrotic area (Fig. 3A); decreased serum ALT (Fig. 3B), AST (Fig. 3C) and LDH (Fig. 3D) activities; and reduced serum HMGB1 level (Fig. 3E). Moreover, AAV8-mediated hepatocyte NEDD4-1 overexpression was sufficient to increase the mouse survival rate upon administration of a lethal dose (600 mg/kg) of APAP (Fig. 3F).

1. Download : Download high-res image (1MB)
2. Download : Download full-size image

Figure 3. Hepatocyte NEDD4-1 overexpression alleviates APAP-induced hepatotoxicity in vivo and in vitro. (A) Representative H&E-stained images (left) and necrotic area quantification (right) of mouse liver sections. Scale bar, 200 μm. (B–E) Serum ALT (B), AST (C), LDH (D) activities, and HMGB1 level (E) of mice. (F) Survival rate of AAV8-Control (n = 15) or AAV8-NEDD4-1 (n = 13) infected mice, with 600 mg/kg APAP treated for 72 h. (G) Representative images (left) and quantification (right) of Calcein-AM (green) and PI (red) double-stained primary hepatocytes. Scale bar, 200 μm. (H, I) LDH release (H) and HMGB1 level (I) in the culture medium of primary hepatocytes. Experiments A–E were performed on AAV8-Control or AAV8-NEDD4-1 infected mice, with saline or 300 mg/kg APAP treated for 24 h, n = 5–8 per group. Experiments G–I were performed on primary hepatocytes transfected with Ad-Vector or Ad-NEDD4-1, with vehicle or APAP (10 mmol/L for 12 h) treatment. Data are presented as the mean ± SEM; ∗P < 0.05, ∗∗P < 0.01.

For in vitro experiments, primary hepatocytes obtained from WT mice were transfected with adenovirus containing the NEDD4-1 plasmid or control vector (Fig. S5E) and then exposed to APAP. Consistent with the in vivo results, APAP-induced cell death was also alleviated by NEDD4-1 overexpression in primary hepatocytes, as indicated by the decreased cell death rate (PI positive cells %; Fig. 3G), level of LDH release (Fig. 3H), and concentration of HMGB1 in the culture medium (Fig. 3I). Based on the combined in vivo and in vitro results, we conclude that hepatocyte NEDD4-1 overexpression ameliorates APAP-induced liver injury and necrosis.

3.4. NEDD4-1 mediates APAP-induced hepatocyte mitochondrial damage

Our RNA-seq data demonstrate an enhanced defect in mitochondrial function in APAP-treated HepKO mice. Therefore, we investigated the influence of NEDD4-1 on mitochondrial homeostasis in the setting of AILI by determining the ATP and ROS contents, mitochondrial membrane potential, mitochondrial membrane permeabilization, etc. APAP-induced hepatic ATP depletion (Fig. 4A), ROS overproduction (Fig. 4B), and cytosolic leakage of mitochondrial intermembrane proteins (apoptosis-inducing factor [AIF], endonuclease G [Endo G] and cytochrome c [Cyt C]; Fig. 4C) were aggravated by hepatocyte NEDD4-1 knockout in vivo, indicating that hepatocyte NEDD4-1 deficiency exacerbated APAP-induced mitochondrial damage. Conversely, overexpression of hepatocyte NEDD4-1 by AAV8-NEDD4-1 injection caused a beneficial improvement in APAP-induced mitochondrial damage (Supporting Information Fig. S6A–S6C).

1. Download : Download high-res image (2MB)
2. Download : Download full-size image

Figure 4. Hepatocyte-specific deficiency of NEDD4-1 exacerbates APAP induced mitochondrial damage in vivo and in vitro. (A, B) ATP (A) and ROS (B) contents in livers of mice. (C) Representative Western blots of indicated proteins in the cytosolic and mitochondrial fractions of liver lysates from mice. (D) Representative transmission electron microscopy micrographs of primary hepatocytes. Scale bar, 2 μm. (E) Representative Western blots of indicated proteins in the cytosolic and mitochondrial fractions of cell lysates. (F) Seahorse assays show the OCR and corresponding statistical results in primary hepatocytes. (G) Representative JC-1-stained images (left) and quantification of mitochondria depolarization (right) in primary hepatocytes. Scale bar, 100 μm. (H) Representative MitoSOX Red-stained images (right) and quantification (left) in primary hepatocytes. Scale bar, 200 μm. Experiments A–C were performed on Flox or HepKO mice treated with saline or 300 mg/kg APAP for 24 h; n = six per group. Experiments D–H were performed on primary hepatocytes isolated from Flox or HepKO mice, with vehicle or APAP (10 mmol/L for 12 h) treatment. Data are presented as the mean ± SEM; ∗P < 0.05, ∗∗P < 0.01.

In vitro, transmission electron microscopy revealed mitochondrial swelling and damage in APAP-treated Flox hepatocytes, which was even worse in APAP-treated HepKO hepatocytes (Fig. 4D). In addition, upon APAP treatment, HepKO hepatocytes displayed greater increases in the cytosolic release of mitochondrial intermembrane proteins (AIF, Endo G and Cyt C) than Flox hepatocytes (Fig. 4E). OCR measurements obtained by a Seahorse assay demonstrated that the impairment of basal respiration, ATP production and maximal respiration induced by APAP exposure were further aggravated by hepatocyte NEDD4-1 ablation, revealing a more severe defect in mitochondrial function (Fig. 4F). JC-1 staining, which is used to visualize the mitochondrial membrane potential by detecting fluorescence, indicated that NEDD4-1 ablation in primary hepatocytes aggravated the APAP-induced collapse of the mitochondrial membrane potential (Fig. 4G). Meanwhile, MitoSOX staining was performed to evaluate mtROS production in vitro and showed that knockout of NEDD4-1 in hepatocytes increased APAP-induced overproduction of mtROS (Fig. 4H). In contrast, transfection of NEDD4-1 adenovirus into WT primary hepatocytes significantly mitigated APAP-induced mitochondrial damage, as evidenced by the Seahorse assay (Fig. S6D), JC-1 staining (Fig. S6E), and MitoSOX staining (Fig. S6F).

In summary, these gain- and loss-of-function studies in vivo and in vitro show that hepatocyte NEDD4-1 at least partially mediates APAP-induced hepatocyte mitochondrial damage.

3.5. Knockdown of VDAC1 mitigates the enhanced AILI events caused by hepatocyte NEDD4-1 deficiency

The striking effect of NEDD4-1 on mitochondrial intermembrane protein leakage in the setting of AILI prompted us to investigate the influence of NEDD4-1 on proteins related to mitochondrial membrane permeabilization. Intriguingly, under APAP treatment, hepatocyte NEDD4-1 ablation increased the mitochondrial translocation of phosphorylated JNK (P-JNK) and the protein abundance of mitochondrial VDAC1, but had a negligible influence on the protein abundance of mitochondrial bcl-2-associated X (Bax) and cyclophilin D (CYPD) (Fig. 5A). In contrast, NEDD4-1 overexpression resulted in the opposite trends in the protein levels of mitochondrial P-JNK and VDAC1 in APAP-treated primary hepatocytes (Supporting Information Fig. S7A).

1. Download : Download high-res image (1MB)
2. Download : Download full-size image

Figure 5. Knockdown of VDAC1 mitigates the enhanced AILI events caused by hepatocyte-specific NEDD4-1 deficiency in vivo. (A) Representative Western blots of indicated proteins in the cytosolic and mitochondrial fractions of liver lysates from Flox or HepKO mice treated with 300 mg/kg APAP for 24 h. (B) Representative Western blots of indicated proteins in the cytosolic and mitochondrial fractions of liver lysates from mice treated with saline or 300 mg/kg APAP for 6, 12 or 24 h. (C) Representative IHC staining (left) and quantification (right) in mice treated with saline or 300 mg/kg APAP for 24 h; n = 3 per group. Scar bar, 200 μm. (D) The hepatic mRNA expressions of Vdac1 in mice treated with saline or 300 mg/kg APAP for 24 h; n = 8 per group. (E) Representative H&E-stained images (left) and necrotic area quantification (right) of mouse liver sections. Scale bar, 200 μm. (F–I) Serum ALT (F), AST (G), LDH (H) activities, and HMGB1 levels (I) of mice (J) Survival rate of AAV8-shControl (n = 11) or AAV8-shVDAC1 (n = 11) infected HepKO mice in response to a lethal dose of APAP (600 mg/kg) for 72 h. (K, L) ATP (K) and ROS (L) contents in the livers of mice. (M) Representative Western blots of indicated proteins in the cytosolic and mitochondrial fractions of liver lysates from AAV8-shControl or AAV8-shVDAC1 infected HepKO mice, with 300 mg/kg APAP treated for 24 h. Experiments E–I, K and L were performed on AAV8-shControl or AAV8-shVDAC1 infected Flox or HepKO mice, with 300 mg/kg APAP treated for 24 h; n = 5–6 per group. Data are presented as the mean ± SEM; ∗P < 0.05, ∗∗P < 0.01, n. s., not significant.

Since NEDD4-1 markedly influenced the VDAC1 protein abundance in the setting of AILI, we sought to determine whether VDAC1 participates in the regulation of NEDD4-1 in AILI pathogenesis. First, the expression profile of hepatic VDAC1 in the setting of AILI was determined. Results showed that the VDAC1 protein abundance was markedly increased in the livers of APAP-challenged mice (Fig. 5B and C) and in APAP-treated primary hepatocytes (Fig. S7B), although VDAC1 mRNA expression did not differ significantly (Fig. 5D), suggesting that the increase in the VDAC1 protein abundance upon APAP treatment was not due to changes in transcription.

To further elucidate the role of VDAC1 in NEDD4-1-regulated AILI, VDAC1 was silenced in Flox and HepKO mice via tail vein injection of AAV8-shVDAC1, and the mice were then exposed to APAP (Supporting Information Fig. S8A). The silencing efficiency of AAV8-shVDAC1 in HepKO mice was validated by Western blotting (Fig. S8B). AAV8-shVDAC1 injection did not affect the CYP2E1 protein abundance (Fig. S8C) and GSH depletion (Fig. S8D), indicating unaffected APAP metabolism. Nevertheless, AAV8-shVDAC1 injection significantly alleviated liver injury and necrosis both in APAP-treated Flox mice and in APAP-treated HepKO mice compared with their corresponding controls, as indicated by H&E staining (Fig. 5E); serum ALT (Fig. 5F), AST (Fig. 5G), LDH activities (Fig. 5H); HMGB1 levels (Fig. 5I); and gross liver appearance (Fig. S8E). Consistent with these findings, the reduction in the survival rate was ameliorated by VDAC1 knockdown (Fig. 5J). Moreover, hepatic mitochondrial damage was alleviated by injection of AAV8-shVDAC1 (Fig. 5K and L). Notably, the mitochondrial translocation of P-JNK induced by NEDD4-1 deficiency under APAP exposure was reduced by VDAC1 knockdown, suggesting that VDAC1 may function cooperatively with JNK (Fig. 5M). Collectively, these results show that knockdown of VDAC1 effectively alleviates APAP-induced mitochondrial damage, liver injury and necrosis. More importantly, the enhanced AILI events caused by hepatocyte NEDD4-1 deficiency were also weakened by VDAC1 knockdown in vivo. These in vivo phenomena were also evaluated in vitro in HepKO hepatocytes treated with APAP and Ad-shVDAC1 (Fig. S8F). Under APAP exposure, the cell death rate (PI positive cells %; Fig. 6A), LDH release (Fig. 6B) and HMGB1 concentration in the culture medium (Fig. 6C) were lower in the Ad-shVDAC1 group than in the Ad-shControl group. Furthermore, mitochondrial damage enhanced by hepatocyte NEDD4-1 knockout was also mitigated by VDAC1 knockdown under APAP treatment in vitro, as determined by a Seahorse assay (Fig. 6D), JC-1 staining (Fig. 6E), and MitoSOX staining (Fig. 6F). These data indicate that knockdown of VDAC1 protects against the enhancement of APAP-induced pathological events resulting from hepatocyte NEDD4-1 deficiency in vitro.

1. Download : Download high-res image (1MB)
2. Download : Download full-size image

Figure 6. Knockdown of VDAC1 mitigates the enhanced APAP-induced pathological events triggered by hepatocyte-specific NEDD4-1 deficiency in vitro. (A) Representative images (left) and quantification (right) of Calcein-AM (green) and PI (red) double-stained primary hepatocytes. Scale bar, 200 μm. (B, C) LDH release (B) and HMGB1 level (C) in the cell culture medium. (D) Seahorse assays showed the OCR (top) and corresponding statistical results (bottom) in primary hepatocytes. (E) Representative JC-1-stained images (left) and quantification of mitochondria depolarization (right) in primary hepatocytes. Scale bar, 100 μm. (F) Representative MitoSOX Red-stained images (left) and quantification (right) in primary hepatocytes. Scale bar, 200 μm. Experiments A–F were performed on HepKO hepatocytes transfected with Ad-shControl or Ad-shVDAC1, with 10 mmol/L APAP treated for 12 h. Data are presented as the mean ± SEM; ∗P < 0.05, ∗∗P < 0.01, n. s., not significant.

As a recent study revealed that VDAC1 oligomerization mediates APAP toxicity⁴⁵, we further investigated whether NEDD4-1 affects VDAC1 oligomerization in APAP toxicity. In vitro experiments showed that NEDD4-1 knockout promoted APAP-induced VDAC1 oligomerization (Supporting Information Fig. S9A), which was reduced by NEDD4-1 overexpression (Fig. S9B). These results indicate that NEDD4-1 regulates VDAC1 oligomerization in APAP toxicity.

3.6. Identification of VDAC1 as a functional ubiquitination substrate of NEDD4-1

Since the increase in the VDAC1 protein abundance upon APAP treatment was not due to transcriptional changes, we rationally deduced that in the setting of AILI, VDAC1 is regulated through proteasomal degradation via the E3 ubiquitin ligase NEDD4-1. As expected, NEDD4-1 overexpression significantly decreased the protein abundance of VDAC1, and this effect was blocked by the proteasome inhibitor MG132 but not by the lysosome inhibitor CQ, emphasizing that NEDD4-1 regulates the proteasomal degradation of VDAC1 (Fig. 7A).

1. Download : Download high-res image (1MB)
2. Download : Download full-size image

Figure 7. VDAC1 is a ubiquitination substrate of NEDD4-1. (A) Representative Western blots of indicated proteins in the mitochondrial fractions of primary hepatocytes transfected with Ad-Vector or Ad-NEDD4-1 (left) or treated with vehicle or 10 mmol/L APAP for 12 h (right), with DMSO or MG132 or CQ treatment. (B, C) Representative Western blots of indicated proteins after IP with anti-HA affinity gel (B) or anti-Flag affinity gel (C), and Western blots of indicated proteins in the cell lysates of 293 T cells transfected the indicated plasmids, with MG132 treatment. (D) Representative confocal images (left) and fluorescence line profile (right) of primary hepatocytes stained with NEDD4-1 (red), VDAC1 (green), and DAPI (blue). Scale bar, 30 μm. (E) Schematic (left) of mouse WT and various truncated forms of NEDD4-1 and representative Western blots (right) for Co-IP analyses in 293 T cells transfected with the indicated plasmids, with MG132 treatment. (F) Schematic (top) of mouse VDAC1 and representative Western blots (bottom) for Co-IP analyses in 293 T cells transfected with the indicated plasmids, with MG132 treatment. (G) Representative Western blots of ubiquitinated or total VDAC1 protein after IP with anti-VDAC1 antibody, and Western blots of indicated proteins in the cell lysates of 293 T cells transfected the indicated plasmids, with MG132 treatment. (H–J) Representative Western blots of indicated proteins after IP with anti-Flag affinity gel, and Western blots of indicated proteins in the cell lysates of 293 T cells transfected the indicated plasmids, with MG132 treatment. (K) Representative Western blots of K48-Ub-linkaged or total VDAC1 protein after IP with anti-VDAC1 antibody, and Western blots of indicated proteins in the cell lysates of 293 T cells transfected the indicated plasmids, with MG132 treatment.

The interaction of exogenous NEDD4-1 with VDAC1 was demonstrated by reciprocal Co-IP (Fig. 7B and C). Consistently, intracellular colocalization of endogenous NEDD4-1 and VDAC1 was further observed in primary hepatocytes via an immunofluorescence colocalization assay (Fig. 7D). Subsequently, molecular mapping assays using WT and truncated fragments of NEDD4-1 revealed that the WW domain of NEDD4-1 is responsible for its interaction with VDAC1 (Fig. 7E). Moreover, missing PPTY motif in VDAC1 abolished its interaction with NEDD4-1 (Fig. 7F). Taken together, these results clarify that NEDD4-1 binds to the PPTY motif of VDAC1 through its WW domain.

Next, we investigated whether NEDD4-1 promotes the ubiquitination of VDAC1. As shown in the ubiquitination assays, NEDD4-1 significantly facilitated the ubiquitination of VDAC1 (Fig. 7G and H), but the NEDD4-1^C854S mutant lacking ubiquitin ligase activity did not (Fig. 7H). These results indicate that NEDD4-1-catalyzed ubiquitination of VDAC1 depends on its E3 ubiquitin ligase activity. Screening for potential lysine ubiquitination types revealed that both WT and K48O (ubiquitin with the intact Lys48 residue only) ubiquitin but not K6O, K11O, K27O, K33O, or K63O (Fig. 7I) could be linked to VDAC1 by NEDD4-1. Moreover, NEDD4-1 failed to link K48R ubiquitin (ubiquitin with mutation of Lys48) to VDAC1 (Fig. 7J). Investigation using a K48 linkage-specific anti-ubiquitin antibody further confirmed that NEDD4-1 facilitated K48-linked ubiquitination of VDAC1 (Fig. 7K). In summary, the present data indicate that NEDD4-1 directly interacts with VDAC1 and predominantly promotes K48-linked ubiquitination of VDAC1, eventually resulting in its degradation.

Accordingly, we propose a signaling mechanism in which VDAC1 is targeted as a degradation substrate of NEDD4-1 to mediate hepatocyte mitochondrial damage and necrosis during AILI (Fig. 8).

1. Download : Download high-res image (257KB)
2. Download : Download full-size image

Figure 8. Schematic representation modeling that the E3 ubiquitin ligase NEDD4-1 suppresses AILI in mouse by targeting VDAC1 for degradation.