A whole herb of A. roxburghii was purchased from Yongan City Huangnijia Co., Ltd. (Sanming, China) and identified by Prof. Jianping Wang at the School of Pharmacy, Huazhong University of Science and Technology. A voucher specimen (No. TJ20180901) has been deposited at the Hubei Key Laboratory of Natural Medicinal Chemistry and Resource Evaluation. KD was prepared using a previously reported method (Ming et al., 2021). The purity of KD (>98%) was analyzed using a high-pressure liquid chromatograph equipped with an evaporative light scattering detector, and the identity was confirmed by comparing its nuclear magnetic resonance data with those from a previous study (Xiang et al., 2016). Silymarin was purchased from Sigma Chemical Co. (St. Louis, MO, United States). The Lieber–DeCarli ethanol liquid diet was purchased from Trophic Animal Feed High-tech Co., Ltd. (Jiangsu, China). 3-Nitrotyrosine (3-NT; ab61392), 4-hydroxynonenal (4-HNE; ab46545), catalase (ab76110), aldehyde dehydrogenase 2 (ALDH2; ab108306), glyceraldehyde 3-phosphate dehydrogenase (GAPDH; ab181602), cluster of differentiation 3 (CD3; ab215212), Bcl-XL (ab32370), Bax (ab32503), STRAD (ab192879) and phospho (p)-mTOR (ab109268) antibodies were purchased from Abcam (Cambridge, United Kingdom). AMPKα (5832), p-AMPKα (2535), Beclin-1 (3495), p-Beclin-1 (84966S), BIP (3177), eIF2α (5324), p-eIF2α (3398), F4/80 (70076), JNK1 (3708), p-SAPK/JNK (4668), LC3A/B (12741), Raptor (2280), p-Raptor (2083), sequestosome-1 (SQSTM1)/p62 (5114), ULK1 (8054), p-ULK1 (5869), mTOR (2983), goat anti-rabbit (5151), and goat anti-mouse (5257) antibodies were purchased from Cell Signaling Technology (Beverly, MA, United States). Alcohol dehydrogenase 1B (ADH1B; 66939-1-Ig), LKB1 (10746-1-AP) and cytochrome P450 2E1 (CYP2E1; 19937-1-AP) antibodies were purchased from Proteintech (Rosemount, IL, United States). The CHOP (WL00880) antibody was purchased from Wanleibio (Liaoning, China). The mCherry-EGFP-LC3 adenovirus was purchased from Hanbio Co. Ltd. (Hanbio, Shanghai, China).

The mouse hepatocyte cell line, AML12, was cultured in Dulbecco’s modified Eagle’s medium/F12 (Zhong Qiao Xin Zhou, China) supplemented with 10% fetal bovine serum, 1% penicillin–streptomycin, 10 μg/ml insulin, 5.5 μg/ml transferrin, 5 ng/ml selenium, and 40 ng/ml dexamethasone. Cells were cultured in a humidified atmosphere containing 5% CO₂ at 37°C.

AML12 cells at the third passage were seeded into 96-well plates at a density of 1 × 10⁴ cells per well. After 24 h of incubation, the cells were washed with phosphate-buffered saline (PBS) and then cultured with different concentrations of KD (0, 5, 10, 20, 40, and 80 μM) or ethanol (0, 50, 100, 200, 300, and 400 mM). After incubation for another 24 h, the culture media were replaced with a culture medium containing a 10% Cell Counting Kit-8 solution. After incubation for 2 h, the absorbance was measured at 450 nm (Varioskan LUX spectrophotometer; Thermo Fisher Scientific, Inc.).

AML12 cells at the fifth passage were seeded at a density of 2 × 10⁵ cell per well into a 6-well plate with slides. On the next day, the medium was changed, and the cells were incubated with KD (0, 10, 20, and 40 μM) for 1 h, followed by incubation with or without 200 mM ethanol for 24 h, after which medium containing KD or ethanol was replaced. After another 24 h, the cells were washed with PBS, fixed with paraformaldehyde, and then stained with the BODIPY staining solution (B&P Biotech, Hangzhou, China) according to the manufacturer’s instructions. 4′,6-Diamidino-2-phenylindole (Servicebio, Wuhan, China) was used to counterstain the nuclei. Images were observed under a fluorescence microscope (Olympus, Tokyo, Japan). The ImageJ software was used for quantitative analysis of lipids.

The protocol of the animal study was reviewed and approved by the Institutional Animal Care and Use Committee of the Tongji Medical College, Huazhong University of Science and Technology. The Lieber–DeCarli ethanol liquid diet was used according to the manufacturer’s instructions and was combined with carbon tetrachloride (CCl₄) to induce alcoholic liver injury in mice (Lamas-Paz et al., 2018). Eight-week-old male C57BL/6J mice (21–23 g) were purchased from SPF Biotechnology Co., Ltd. (Beijing, China) and were raised in a specific pathogen-free environment. Silymarin was chosen as the positive control (Federico et al., 2017; Gillessen and Schmidt, 2020). All mice were first fed a control (ethanol-free) liquid diet for 5 days. Subsequently, the mice were randomly divided into the following five groups: control, model, KD-20 mg/kg, KD-40 mg/kg, and silymarin-80 mg/kg. The control group was administered the control liquid feed, while the other groups were administered the ethanol liquid feed, and the ethanol concentration was gradually increased to 4% (w/v) in the first week and then maintained at the same level. On week 5, the mice in all groups, except the control group, were intraperitoneally injected with 5% CCl₄ in olive oil (2 ml/kg) twice a week and continued in subsequent weeks. The mice were sacrificed 5 weeks after the first injection of CCl₄ to obtain tissues and blood.

Samples were lysed with a buffer containing 4% sodium dodecyl sulfate, 100 mM Tris-HCl, and 1 mM dithiothreitol (pH 7.6) to extract proteins, and the bicinchoninic acid method was used for protein quantification. An appropriate amount of protein from each sample was prepared using the filter-aided sample preparation method for trypsin digestion (Wisniewski et al., 2009). An Empore™ solid phase extraction cartridge (C18, standard density, bed internal diameter: 7 mm, volume: 3 ml; Sigma) was used to desalt the peptides. The peptides were concentrated by vacuum centrifugation and then reconstituted with 40 μL of 0.1% (v/v) formic acid. Peptides (100 μg) from each sample were labeled according to the manufacturer’s instructions (Thermo Fisher Scientific) and then fractionated using a high pH reversed-phase peptide fractionation kit (Thermo Fisher Scientific). Liquid chromatography–tandem mass spectrometry was performed on a Q Exactive mass spectrometer (Thermo Fisher Scientific) coupled with an EASY-nLC liquid chromatograph (Thermo Fisher Scientific) to analyze the peptides. Protein identification and quantitative analyses were performed by searching the mass spectrometric raw data of each sample using the MASCOT search engine version 2.2 (Matrix Science, London, United Kingdom) integrated into the Proteome Discoverer 1.4 software. Finally, bioinformatics analysis was performed as previously described (Ashburner et al., 2000; Ross et al., 2004; Gotz et al., 2008).

Serum levels of tumor necrosis factor-alpha (TNF-α) and interleukin-6 (IL-6) were measured using an enzyme-linked immunosorbent assay kits (Neobioscience) according to the manufacturer’s instructions. Serum levels of alanine aminotransferase (ALT) and aspartate transaminase (AST), as well as triglycerides (TG), reduced glutathione/oxidized glutathione (GSH/GSSG) and nicotinamide adenine dinucleotide (NAD⁺)/reduced NAD⁺ (NAD⁺/NADH) ratios of liver homogenate, were determined according to the instructions of each kit obtained from Nanjing Jiancheng Bioengineering Institute. Activities of ADH, catalase, ALDH, and superoxide dismutase (SOD) in liver homogenates were also measured according to the instructions of each kit obtained from Nanjing Jiancheng Bioengineering Institute.

Liver tissue was fixed with 10% formalin and embedded in paraffin. The paraffinized tissue was cut into 5-μm sections and subjected to hematoxylin–eosin (H&E) and Masson’s trichrome staining.

The levels of 3-NT, 4-HNE, F4/80, and CD3 in the liver were determined by immunohistochemical staining. Paraffin sections of the liver tissue were prepared using high temperature and pressure and then incubated with 3% hydrogen peroxide for 20 min to inactivate endogenous peroxidase. After incubation with 10% serum for 20 min to block nonspecific binding, the sections were separately incubated with the 3-NT, 4-HNE, F4/80, and CD3 primary antibodies in a working solution, followed by incubation with the corresponding secondary antibody to develop color. After the slides were rinsed with Tris-buffered saline, 3,3′-diaminobenzidine was added, and the sections were observed under a microscope. Subsequently, the slides were rinsed with tap water, counterstained with hematoxylin, dehydrated with gradient ethanol, and finally loaded with a neutral adhesive for microscopic examination.

A TUNEL kit (Roche, Shanghai, China) was used to detect apoptosis in liver tissue. Paraffin sections were deparaffinized and hydrated. A mixture of 3% bovine serum albumin and 20% normal calf serum was used to block nonspecific reactions. The TUNEL working solution was added dropwise, and the sections were incubated in the dark for 60 min at 37°C. 4′,6-Diamidino-2-phenylindole was used to counterstain the nuclei, and images were observed under a fluorescence microscope.

The liver tissue was removed from storage at −80°C and lysed in a buffer containing protease and phosphatase inhibitors. The lysate was then centrifuged at 12,000 × g for 10 min, and the protein concentration was quantified in the supernatant using the bicinchoninic acid method. After boiling with a sample loading buffer for 5 min, approximately 20–50 µg of total protein was separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis. The separated proteins were transferred onto nitrocellulose membranes, and the membranes were blocked with a fast blocking solution, followed by incubation with primary antibodies overnight at 4°C. After washing with Tris-buffered saline with 0.1% Tween 20, the membranes were incubated with secondary antibodies for 30–60 min at room temperature. A dual-color near-infrared fluorescence imaging system was used to detect protein bands. The ImageJ software was used for quantitative analysis of proteins.

1 × 10⁶ AML12 cells were seeded in a 6-well plate. The next day, the cells were treated with 100 mM ethanol for 0, 1, 4, 8, 12, and 24 h. After the cells were collected, western blotting was performed to detect the expression levels of LC3A/B-I, LC3A/B-II and SQSTM1.

The AML12 cells were seeded into 6-well plates at a density of 1 × 10⁶ cells/well for 24 h. Then, 40 μM KD with or without 50 μM chloroquine (CQ) were added for 1 h according to the grouping, and cells were treated with 100 mM ethanol for another 8 h. The expression levels of LC3A/B-I, LC3A/B-II and SQSTM1 were further detected as described above.

The AML12 cells were seeded in a confocal culture dish with a glass bottom (diameter of 15 mm; 1.5 × 10⁵ cells/dish, and transfection was carried out on the next day. After 39 h of transfection, 40 μM KD or 50 μM CQ was added for 1 h, followed by addition of 100 mM ethanol for 8 h. After being washed with PBS, the cells were fixed with 4% paraformaldehyde for 10 min, and then washed with PBS again. The cells were immersed in PBS for observation by confocal fluorescence (Nikon, Tokyo, Japan).

Data are presented as the mean ± standard error of the mean (SEM) and were analyzed with Dunnett’s t-test or one-way analysis of variance using the GraphPad Prism software (GraphPad Software, Inc., San Diego, CA, United States). Statistical significance was set at p < 0.05.

AML12 normal hepatocytes were selected to analyze the effect of KD on ethanol-induced lipid accumulation. The cytotoxicity of KD and ethanol to AML12 cells was assayed to determine the optimal concentrations for subsequent experiments. As shown in Figure 1B, KD showed no cytotoxic effect on AML12 cells at 80 μM, while ethanol caused liver cell damage in a concentration-dependent manner in a range of 100–400 mM. Based on the above results, 200 mM ethanol was chosen for further treatment. BODIPY staining was used to localize lipids, which showed that 200 mM ethanol induced obvious lipid accumulation (green fluorescence intensity in Figure 1C). However, after pretreatment with different concentrations of KD, the lipid accumulation levels significantly decreased, and the lipid accumulation status was comparable to that of normal cells (Figure 1C). The BODIPY staining results showed that KD reduced ethanol-induced lipid accumulation in a concentration-dependent manner in a range from 20 to 40 μM (Figure 1D).

A mouse model was established and pathologically analyzed to explore whether KD alleviates alcoholic liver injury. The detailed modeling method is illustrated in Figure 2A. The results of H&E staining showed ballooning degeneration and inflammatory cell infiltration in the model group. After KD treatment, ballooning degeneration was significantly reduced, and inflammation was alleviated (Figure 2B). Masson’s trichrome staining showed that the number of blue collagen fibers dramatically increased in the model group and decreased after KD treatment, indicating the therapeutic effect of KD against liver tissue fibrosis (Figure 2B). The liver index of mice also changed after ethanol and KD treatment. KD reduced the increase in liver to body weight ratio caused by ethanol (Figure 2C) The essential indicators of hepatocellular damage were further assessed, including ALT, AST and TG. As shown in Figures 2D–F, the ALT, AST and TG levels were obviously elevated in the model group compared with those in the control group. KD at 20 and 40 mg/kg significantly (p < 0.05) decreased the levels of these biomarkers. These findings demonstrated that KD alleviated alcoholic liver injury in mice, and its effects were comparable to those of silymarin.

To explore the mechanism of action of KD in alleviating alcoholic liver injury, proteomics was used for screening. The proteomic volcano plots and heatmap show that protein expression significantly differed between the model and control groups, with 836 proteins downregulated and 690 proteins upregulated in the model group. Compared with their expression in the model group, 165 proteins were downregulated and 75 proteins were upregulated in the KD group (Figure 3A). Cluster analysis of differentially expressed proteins showed that the protein expression levels in the liver tissue after KD treatment were significantly different from that in model group (Figure 3B). Gene Ontology (GO) analysis of the differentially expressed proteins between the KD and model groups showed that the enriched biological process terms were mainly associated with acute inflammatory response, extracellular structure organization, acute-phase response and cell killing (Figure 3C). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis showed that differentially expressed proteins between the KD and model groups (Figure 3D), in combination with the pathogenesis of ALD, were mainly involved in the ER, autophagy, and AMPK signaling pathways.

The metabolism of ethanol in vivo is an important physiological and pathological process in ALD development. After entering the body, ethanol is converted into acetaldehyde by ADH, CYP2E1 and catalase. Acetaldehyde is then converted into nontoxic acetic acid by ALDH (Seitz et al., 2018). Analysis of enzymes of ethanol metabolism in the liver of mice showed that KD inhibited the ethanol-induced increase in ADH activity, enhanced the increase in catalase activity, and reversed the decrease in ALDH activity (Figures 4A–C). Western blotting revealed that KD reversed the ethanol-induced decrease in ADH1B (the main subunit of ADH) expression and the increase in CYP2E1 expression but had no effect on the expression of catalase and ALDH2 (the main subunit of ALDH) (Figures 4D,E).

ROS, another byproduct of ethanol metabolism, can cause oxidative stress, together with acetaldehyde (Farfan Labonne et al., 2009; Seitz et al., 2018). Analysis of the oxidative stress indicators SOD, GSH/GSSG, and NAD⁺/NADH showed that KD improved the antioxidant status in mice (Figures 5A–C). In the presence of ROS, reactive nitrogen species are derived from NO and nitrify tyrosine or protein tyrosine residues to produce 3-NT. The 3-NT level in the diseased state is significantly higher than the normal 3-NT level (Li Y. et al., 2019). 4-HNE is an aldehyde product of lipid peroxidation and is considered a key mediator of oxidative stress-caused cell death (Ma et al., 2020; Hyatt et al., 2021). The ethanol-induced increases in the 3-NT and 4-HNE levels were reversed by KD treatment (Figures 5D,E). These results indicated that KD reduced the ethanol-induced oxidative stress.

ROS and acetaldehyde cause not only oxidative stress but also ER stress by damaging proteins and DNA (Ji, 2012; Ji, 2014; Ganne-Carrié and Nahon, 2019). KEGG analysis of proteomics data also showed the association of differentially expressed proteins with ER processes (Figure 3D). Western blotting showed that KD reduced the levels of the key ER stress protein BIP and the transcription factor CHOP, as well as the phosphorylation of eIF2α, thereby alleviating the translation block. JNK, a key kinase in ER stress, was also suppressed (Figures 5F,G). These results suggested that KD inhibited the persistent ER stress caused by long-term ethanol consumption.

Oxidative stress and ER stress can trigger inflammation and cell apoptosis. Our results showed that KD reversed the liver inflammation-associated increases in the levels of serum TNF-α and IL-6 (Figures 6A,B). Immunohistochemistry staining of liver tissue sections showed that the ethanol-induced increases in the F4/80 and CD3 levels were reversed by KD (Figure 6C). In addition, CHOP and JNK activation, which can trigger cell apoptosis, was inhibited, while the expression of the antiapoptotic factor Bcl-XL was increased and that of the proapoptotic factor Bax was decreased by KD treatment (Figures 6D,E). The TUNEL assay also confirmed that the ethanol-induced apoptosis of hepatocytes decreased after KD administration (Figure 6F). Thus, the results showed that KD inhibited the ethanol-induced liver inflammation and cell apoptosis. Together, KD alleviates oxidative stress and ER stress by affecting ethanol metabolism in the liver, thereby reducing inflammation and cell apoptosis, which in turn relieves alcoholic liver injury.

AMPK, a central regulator of eukaryotic cell and organism metabolism, is a key protein involved in a variety of signaling pathways. Long-term drinking inhibits the activity of AMPK, thereby disrupting metabolic processes (Gao and Bataller, 2011; Seitz et al., 2018). Our results showed that KD increased the protein levels of STRAD and LKB1 (Figures 7A,B), which were the upstream of AMPK, thereby promoting the phosphorylation of AMPK.

Autophagy can remove damaged proteins, organelles, and lipids from cells and maintain intracellular balance. The increases in ROS and acetaldehyde levels caused by long-term alcohol drinking can destroy proteins, cause lipid accumulation, and inhibit liver autophagy, thereby aggravating the alcohol-induced liver injury (Wang et al., 2015; Allaire et al., 2019). Our results showed that KD treatment led to the phosphorylation of Raptor by activated AMPK, thereby directly inhibiting the phosphorylation of mTOR (Total mTOR strip is shown in Supplementary Figure S1). MTOR is an autophagy inhibitory protein, and the inhibition of mTOR by AMPK can activate autophagy (Allaire et al., 2019). In addition, the activated AMPK in the KD group could directly activate autophagy via phosphorylation of ULK1 at Ser555. KD treatment increased the phosphorylation level of Beclin-1, a key autophagy protein, and that of LC3A/B-II and decreased the level of the autophagy substrate SQSTM1 (Turco et al., 2019), indicating enhanced autophagy (Galluzzi and Green, 2019). The results suggested that KD reversed the ethanol-induced inhibition of the AMPK signaling pathway and autophagy (Figure 7A,B). In summary, in mice with alcoholic liver injury, KD treatment led to the upregulation of STRAD/LKB1, resulting in the activation of AMPK-dependent protective autophagy; which was also shown in Supplementary Figure S2 that the protective effect of KD was abrogated by autophagy inhibition. Silymarin also enhanced protective autophagy by promoting AMPK phosphorylation but had no influence on STRAD and LKB1 expression.

In order to determine the role of KD in autophagy of alcoholic liver injury, we exposed AML12 cells to 100 mM ethanol for 0, 1, 4, 8, 12, and 24 h. The results of multiple experiments demonstrated that the phenomenon of autophagy was a dynamic process and ethanol inhibited LC3A/B-II expression at 8 h with higher stability (Figure 8A), we thus chose 8 h as the end point to study the changes of autophagic flux.

Cells treat with 100 mM ethanol for 8 h showed reduced expression of LC3A/B-II and increased expression of SQSTM1, indicating that autophagy was inhibited. This change was reversed after KD treatment, which means that KD could activated the suppressed autophagy induced by 100 mM ethanol. LC3A/B-II and SQSTM1 both increased after adding KD and autophagy inhibitors CQ, revealing that KD had no effect on autophagy inhibited by CQ (Figure 8B).

For further proof, the AML12 cells were transfected with mCherry-EGFP-LC3 adenovirus. Because the EGFP fluorescence is quenched in the acidic environment of the lysosome, the EGFP and mCherry fluorescence express at the same time, which means the yellow dots after the merge indicate the autophagosomes. Fluorescent dots that do not express EGFP but express mCherry indicate autolysosomes (Liu B. et al., 2020). It was found that 100 mM ethanol reduced the number of autophagosomes and autolysosomes. KD treatment increased the number of autophagosomes and autolysosomes. Treatment with CQ inhibited the binding of autophagosomes and lysosomes, and further caused the accumulation of autophagosomes. After the cells treated with KD and CQ at the same time, the number of autophagosomes was increased, meanwhile the increase ratio of autolysosomes was less than that of KD alone. In summary, the data suggested that KD increases the production of autophagosomes, and could partially restore the autolysosome function inhibited by 100 mM ethanol, thereby enhancing autophagic flux (Figure 8C).