Human hepatocellular carcinoma cell lines (HepG2, Huh7, and PLC/PRF/5) purchased from ATCC (Manassas, VA, United States) were cultured in the Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum, penicillin (100 IU/mL), and streptomycin (100 mg/mL) [all obtained from Gibco (Grand Island, NY, United States)] and placed in an incubator at 37 °C with 5% CO₂ and 95% humidity.

HSCs LX-2 cells (FH0108, Fuheng Biology, Shanghai, China) were cultivated in LX-2 complete medium (FH-LX-2, Fuheng Biology) at 37 °C, with 5% CO₂ in the air and a humidity of 95% (Xie et al., 2019).

LX-2 cells were treated with 0 mM, 50 mM, 100 mM and 150 mM ethanol for 24 h. Changes in cell viability were assessed via the CCK-8 assay. The expression levels of α-smooth muscle actin (α-SMA) and collagen type I alpha 1 chain (CoL1A1) were determined using Western blot.

si-XIST and miR-663a inhibitors and their corresponding negative controls (NCs) (GenePharma, Shanghai, China) were introduced into LX-2 cells using Lipofectamine®2000 (Invitrogen, Carlsbad, CA, United States). All operations were performed as per the guidelines. Cells were gathered after 48-h transfection for subsequent experiments. The sequences were as follows: si-XIST: sense: 5′-GCAAAUGAAAGCUACCAAU-3’; antisense: 5′-AUUGGUAGCUUUCAUUUGC-3’; the miR-663a inhibitor: 5′-UCCGCCCCGCGGCGCCCUGGCG-3’. Sequences of si-NC and inhi-NC were nonsense RNA sequences with no homology to known gene sequences.

LX-2 cells were grouped as follows (detailed in Supplementary Table 1): (1) the Con group (cultured normally); (2) the LX-2 group (induced with ethanol for 24 h); (3) the LX-2 + si-NC group (induced with ethanol for 24 h and then transfected with si-NC for 48 h); (4) the LX-2 + si-XIST group (24 h ethanol induction and then 48 h transfection with si-XIST) (Xie et al., 2019); (5) the LX-2 + si-XIST + Li group [24 h ethanol induction, 48 h transfection with si-XIST, followed by treatment with 1 μM Li (a ferroptosis inhibitor; SML1414, Sigma-Aldrich, St Louis, MO, United States) for 48 h (Li D. et al., 2023)]; (6) the LX-2 + si-XIST + Vehicle group [24 h ethanol induction, 48 h transfection with si-XIST, followed by treatment with the same amount of solvent as Li (50% PEG300 and 50% saline) for 48 h]; (7) the LX-2 + si-XIST + inhi-NC group (induced with ethanol for 24 h and then co-transfected with si-XIST and 10 nM inhibitor NC for 48 h); (8) the LX-2 + si-XIST + miR-663a inhi group (induced with ethanol for 24 h and then co-transfected with si-XIST and 10 nM miR-663a inhibitor for 48 h). Each experiment was independently repeated 3 times.

Total RNA was extracted using TRIzol reagent (Invitrogen) and then reverse transcribed into complementary DNA using the PrimeScript RT reagent kit (Takara, Dalian, Liaoning, China). TaqMan primers and probes for the detection were all obtained from Takara. qPCR was carried out on the ABI PRISM 7900 sequence detection system of SYBR Green II (Takara). The program of PCR consisted of pre-denaturation at 95 °C for 5 min, and 40 cycles of denaturation at 95 °C for 15 s, annealing at 60 °C for 20 s, and extension at 72 °C for 35 s. With β-actin utilized as an internal reference gene for lncRNA-XIST and GPX4, and U6 as an internal reference gene for miR-663, respectively, data were normalized using the 2^−ΔΔCt method (Livak and Schmittgen, 2001). The primer sequences are listed in Table 1. All primers were synthesized by Sangon Biotechnology (Shanghai, China).

LX-2 cells were seeded onto 96-well plates at a density of 5 × 10³ cells per well. The cell viability of each group at 0, 24, 48 and 96 h was evaluated using the CCK-8 kit (CA1210, Solarbio, Beijing, China). Briefly, cells were added with 100 μL/well CCK-8 working solution and then incubated at 37 °C for 2 h. The optical density (OD) at 450 nm was assessed using a microplate reader (Thermo Fisher Scientific, Waltham, MA, United States).

LX-2 cells were seeded onto 6-well plates (500 cells per well) and resuspended for 2 weeks. Upon the visible formation of colonies to the unaided eye, the 6-well plates were taken. After the removal of the culture medium, cells were fixed with 2 mL/well methanol for 30 min, followed by the discarding of the methanol. Afterwards, cells were stained with 2 mL/well of 0.1% crystal violet (548-62-9, MACKLIN, Shanghai, China) for 3 min, and then observed and counted using a microscope (Olympus, Tokyo, Japan).

Lipid peroxidation (LPO) levels in LX-2 cells were detected using the BODIPY 581/591 C11 fluorescent probe (HY-D1301, MCE, Monmouth Junction, NJ, United States). Briefly, LX-2 cell coverslips were prepared and treated with 5 μM BODIPY 581/591 C11 working solution to ensure the coverage of all cells. Next, cell coverslips were fostered for 30 min at room temperature in the dark, washed with phosphate buffered saline, and sealed with an anti-fluorescence quencher. Finally, cells were observed and photographed under a fluorescence microscope (Olympus). The BODIPY 581/591 C11, in the absence of LPO, exhibited orange-red fluorescence, whereas the presence of LPO resulted in the manifestation of green fluorescence in BODIPY 581/591 C11. The LPO level in cells could be evaluated by analyzing the intensity of green fluorescence/red fluorescence.

A total of 42 male C57BL/6 mice (6–8 weeks old, weighing 20 ± 2 g, Guangdong Shuanglin Biopharmaceutical Co., Ltd., Zhanjiang, Guangdong, China) were maintained in a specific pathogen-free facility under standard conditions of 24 °C ± 2 °C, 50% ± 10% humidity, and a 12 h light-dark cycle.

After adaptive feeding for 1 week, mice were injected intraperitoneally with 10 mL/kg olive oil (#56–23-5, MACKLIN) containing 10% carbon tetrachloride (CCl₄) twice a week for 8 consecutive weeks to induce HF, thereby establishing an HF mouse model (Yuan et al., 2022). Olive oil was utilized as a solvent for CCl₄, and the solvent control group received an intraperitoneal injection of an equivalent volume of olive oil.

A total of 42 mice were arranged into 7 groups (six mice per group) as below: (1) the Normal group (normal control mice without any treatment); (2) the Vehicle control group (solvent control group; mice injected with equivalent volume of olive oil intraperitoneally); (3) the Model group (HF mice); (4) the short hairpin (sh)-NC and sh-XIST groups [sh-NC or sh-XIST lentivirus (pGLVU6/GFP vector; GenePharma) was dissolved in 0.5 mL of saline at a titer of 1 × 10⁹ TU/mL, and then injected into HF mice via tail vein once a week for 4 consecutive weeks] (Xie et al., 2019); (5) the sh-XIST + anta NC and sh-XIST + anta miR groups [equal amounts of sh-XIST and antagonist NC or antagonist miR-663a (5 μg per mouse) were dissolved in 1.5 mL saline, and then injected into HF mice via tail vein once a week for 4 consecutive weeks (Chai et al., 2020)].

The day after intervention, mice were weighed and 0.3 mL blood was taken from the orbit for enzyme-linked immunosorbent assay (ELISA). Next, liver tissues were collected, photographed, and weighed after sacrifice of mice with 150 mg/kg of 0.5% pentobarbital sodium (P3761, Sigma-Aldrich). The liver weight ratio (liver weight/weight × 100%) was calculated and normalized with the Normal group. There were six mice in each group. The right lobe of the liver of each mouse was utilized for immunohistochemistry, hematoxylin-eosin (H&E) staining and Masson staining, and the left lobe of the liver for Western blot and kit assays.

H&E staining and Masson staining were applied for histopathological examination of mouse liver tissues. Liver tissue samples were immersed in 10% formalin solution, dehydrated in gradient ethanol, embedded in paraffin, and subsequently sectioned into slices with a thickness of 4 μm. The paraffin sections were routinely dewaxed and hydrated, and later subjected to staining with the H&E staining kit (60524ES60, Yeasen Company, Shanghai, China) and Masson staining kit (60532ES58, Yeasen Company), respectively. The experiment was conducted in strict accordance with the kit instructions. The samples were then observed and photographed under an optical microscope (Olympus). Collagen deposition in mouse liver tissues stained with Masson’s trichrome was quantified and analyzed using Image Pro Plus 6.0 (Media Cybernetics, Silver Spring, MD, United States). Collagen deposition volume fraction (CVF) = collagen fiber area/total area of liver tissue sections * 100%. There were six mice in each group. Three discontinuous sections were obtained from each mouse, and five non-overlapping fields of view were randomly selected from each section for statistical analysis. The results were expressed as the mean value.

Serum levels of ALT, AST and HYP, as well as HYP levels in liver tissue supernatant, were determined using ALT kit (C009-2-1, JianCheng Bioengineering Institute, Nanjing, Jiangsu, China), AST kit (C010-2-1, JianCheng Bioengineering Institute) and HYP kit (A030-2-1, JianCheng Bioengineering Institute). The specific experimental procedures were conducted following the manufacturer’s instructions.

The collected tissues and LX-2 cells were added with lysis buffer (AR0107, Boster, Wuhan, Hubei, China) to extract proteins. After the determination of concentration using the bicinchoninic acid kit (AR1189, Boster), protein samples were added with an appropriate amount of loading buffer and heated in a boiling water bath for 5 min for protein denaturation. The denatured protein samples were added onto the loading wells and separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis. Subsequently, samples were electrotransferred onto the polyvinylidene fluoride membranes, which were then blocked in 3% bovine serum albumin (AR0184, Boster) for 2 h, and later cultivated with the following primary antibodies: anti-α-SMA (1:1000, #19245, CST, Beverly, MA, United States), anti-CoL1A1 (1:1000, #39952, CST) and anti-GPX4 (1:1000, ab125066, Abcam, Cambridge, United Kingdom) overnight at 4 °C, with glyceraldehyde-3-phosphate dehydrogenase (1:2500, ab9485, Abcam) as an internal reference. Thereafter, membranes were incubated with goat anti-rabbit immunoglobulin G (IgG) H&L [horseradish peroxidase (HRP)] (1:2000, ab6721, Abcam) at room temperature in the dark for 1 h, followed by development with enhanced chemiluminescence working solution (AR1191, Boster). Grayscale quantification of each group of bands in Western blot images was performed using Image Pro Plus 6.0 (Media Cybernetics). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as an internal reference protein, and the results were depicted as the ratio of relative GAPDH. The experiment was conducted thrice.

Fe²⁺, MDA, and GSH levels were assayed utilizing the Fe detection kit (ab83366, Abcam), MDA detection kit (ab118970, Abcam), and reduced GSH detection kit (BC1175, Solarbio), respectively. Specific experimental procedures were conducted in light of the manufacturer’s manuals.

Cytotoxicity was quantified using the LDH kit (A020-2-2, Jiancheng Bioengineering Institute). After cells were centrifuged at 600 g and 4 °C for 5 min, the supernatant was collected. The kit working solution was added into the 96-well plates sequentially. Later, cells were incubated at 37 °C for 30 min, with the OD value measured at 450 nm. The specific experimental steps were carried out according to the instructions.

The binding sites between lncRNA-XIST and miR-663a were predicted using the Starbase database (https://rnasysu.com/encori/), and the binding sites between miR-663a and GPX4 were predicted utilizing the TargetScan database (http://www.targetscan.org/vert_72/).

Wild-type (WT) and mutant (MUT) plasmids were synthesized and inserted into the lncRNA-XIST 3′-untranslated region (3′UTR) (lncRNA-XIST-WT/MUT) or GPX4 3′UTR (GPX4-WT/MUT) using the pMIR-reporter vector (KL-ZL-1015-01, Ke Lei Biological Technology, Shanghai, China), respectively. LX-2 cells in the logarithmic phase were seeded onto 96-well plates. Upon reaching 70% cell confluence, cells were co-transfected with 100 ng of lncRNA-XIST WT and MUT plasmids (lncRNA-XIST-WT/MUT) or GPX4 WT and MUT plasmids (GPX4-WT/MUT), along with mimics NC or miR-663a mimics. Following a 48 h transfection period, the relative activity of luciferase was evaluated using the luciferase detection kit (HY K1013, MCE) to evaluate whether miR-663a had a targeted binding relationship with the 3′UTR of lncRNA-XIST or GPX4, respectively.

Liver tissues were paraffin-embedded and repaired under high temperature and high pressure for 2 min, followed by blockade with goat serum (SL038, Solarbio) for 20 min at room temperature. After overnight incubation with anti-α-SMA (1:500, #19245, CST) or anti-CoL1A1 (1:500, #39952, CST) at 4 °C, the sections were incubated with goat anti-rabbit antibody IgG H&L (HRP) (1:1000, ab214050, Abcam) for 60 min at 37 °C. Subsequently, the sections were subjected to diaminobenzidine staining (DA1016, Solarbio), hematoxylin staining (H8070, Solarbio), and photography under a microscope (Olympus). Quantitative analysis of immunohistochemistry was conducted using Image Pro Plus 6.0 (Media Cybernetics).

Data were statistically analyzed and graphed using GraphPad Prism 8.01 (GraphPad Software, San Diego, CA, United States). Six mice were assigned into each group in animal experiments (n = 6), and cellular experiments were repeated thrice independently (n = 3). Normal distribution of continuous variables was tested using the Kolmogorov-Smirnov test. Normally distributed data were expressed as mean ± standard deviation (SD). Data comparisons between two groups were conducted using an independent sample t-test; data comparisons among multiple groups were performed using one-way analysis of variance (ANOVA), followed by Tukey’s multiple comparisons test for post-hoc analysis. Statistical significance was defined as P < 0.05. The actual difference between the two groups of data was evaluated dependent on the effect size (Cohen’s d). The formula was as follows: Cohen‘s d = (mean difference between the two groups)/(combined standard deviation of the two groups), with d > 0.8 regarded as a large effect size, which is indicative of an extremely significant difference between the two groups of data.

LX-2 cells were treated with 0 mM, 50 mM, 100 mM, and 150 mM ethanol for 24 h or with 100 mM ethanol for 0 h, 12 h, 24 h, and 48 h, respectively, to induce HSC activation. CCK-8 assay results illustrated that ethanol enhanced the cell viability of LX-2 cells dose- and time-dependently, and the LX-2 cell viability reached its peak and was stable upon ethanol treatment at a concentration of 100 mM for 24 h (Figure 1A, all P < 0.001). The expression levels of HSC activation markers α-SMA and CoL1A1 (Xie et al., 2019) were further examined. It was found that ethanol promoted α-SMA and CoL1A1 expression dose- and time-dependently, and α-SMA and CoL1A1 levels were the highest and remained stable under ethanol treatment at 100 mM for 24 h (Figure 1B, all P < 0.05). The above results suggested that 100 mM and 24 h were the optimal concentration and time of ethanol treatment, respectively. As reflected by RT-qPCR results, lncRNA-XIST expression increased with the increase of ethanol treatment time and remained stable at 24-48 h (Figure 1C, all P < 0.001). In addition, the differential expression of lncRNA-XIST in activated LX-2 cells and several hepatocellular carcinoma cell lines that might be linked with fibrosis (HepG2, Huh7, and PLC/PRF/5) was analyzed, with the results showing that the lncRNA-XIST level in hepatocellular carcinoma cells was further upregulated (Figure 1D, all P < 0.001). These results revealed that lncRNA-XIST might increase with HF progression.

Subsequently, LX-2 cells were treated with 100 mM ethanol for 24 h to induce their activation according to the experimental conditions determined in the pre-experiment. Colony formation assay further revealed that ethanol evidently increased cell proliferation (P < 0.001) (Figure 1G). Moreover, after transfection with si-XIST, LX-2 cells displayed decreased lncRNA-XIST mRNA level. The corresponding alterations in cellular function further suggested effective transfection, as evidenced by reduced cell viability and α-SMA and CoL1A1 expression levels (P < 0.05), and diminished cell proliferation (P < 0.001) (Figures 1E–H). It is documented that promotion of ferroptosis represses HSC activation and mitigates HF (Kuo et al., 2020). The LPO assay with BODIPY 581/591 C11 fluorescent probe found no prominent change in the ratio of orange-red and green fluorescence in the LX-2 cells after ethanol treatment, indicating that the cells did not exhibit substantial LPO (Figure 1I). Additional results revealed no significant changes in Fe²⁺, MDA, and GSH levels in cells after ethanol treatment (all P > 0.001) (Figure 1J). After transfection with si-XIST, the green fluorescence in LX-2 cells was intensified, indicating the occurrence of significant LPO (Figure 1I). Additionally, Fe²⁺ and MDA levels were increased (P < 0.01), and GSH level was reduced (P < 0.001) (Figure 1J). LDH assay results illustrated no substantial change in cell death after ethanol induction (P > 0.05), which was increased after transfection with si-XIST (P < 0.001) (Figure 1K). The results collectively indicate that ethanol induces lncRNA-XIST expression, and knockdown of lncRNA-XIST promotes ferroptosis to inhibit HSC activation.

LX-2 cells were induced using ethanol for 24 h and then transfected with si-XIST while being treated with 0, 0.5, 1, and 2 μM Li for 48 h. CCK-8 assay indicated that Li dose-dependently restored LX-2 cell viability, with the Li treatment at 1 μM yielding the highest and stable cell viability (Figure 2A, all P < 0.05). Li dose-dependently downregulated Fe²⁺ and MDA levels and upregulated the GSH level, with the 1 μM concentration indicating the most pronounced effects (Figure 2B, all P < 0.05).

Subsequently, LX-2 cells transfected with si-XIST were treated with 1 μM Li. Cells in the LX-2 + si-XIST + Li group showed elevated LX-2 cell viability and α-SMA and CoL1A1 expression levels (P < 0.05), enhanced proliferation (P < 0.05), diminished green fluorescence, decreased Fe²⁺ and MDA levels (P < 0.05), raised GSH level (P < 0.05), and reduced cell death (P < 0.05) (Figures 2C–H). The aforementioned results suggest that inhibition of ferroptosis partially abrogates the effect of lncRNA-XIST knockdown on HSC activation.

LX-2 cells were treated with 100 mM ethanol for 0 h, 12 h, 24 h and 48 h. Changes in mRNA expression of GPX4 were detected by RT-qPCR, and it was observed that GPX4 mRNA was increased with the increase of ethanol treatment time, which tended to be stable at 24-48 h (Figure 3A, all P < 0.001). RT-qPCR and Western blot were conducted to determine GPX4 expression in different groups, with the results displaying that ethanol treatment dramatically raised GPX4 expression in cells (P < 0.001) (Figures 3B,C), which was reduced by lncRNA-XIST silencing (P < 0.01) (Figures 3B,C). Starbase database analysis manifested that miR-663a was a downstream target miRNA of lncRNA-XIST, and TargetScan database prediction showed that miR-663a had target sites binding to GPX4 (Figures 3D,F). Furthermore, dual-luciferase reporter gene assay verified the presence of the binding relations of miR-663a with lncRNA-XIST and GPX4, respectively (Figures 3E,G). RT-qPCR data revealed that miR-663a levels reduced with the increment of ethanol treatment time and stabilized at 24–48 h. Furthermore, ethanol decreased miR-663a level in cells, whereas silencing of lncRNA-XIST partly restored miR-663a level (P < 0.01) (Figure 3H). These results indicate that lncRNA-XIST acts as a ceRNA for miR-663a to regulate GPX4 expression.

Cells in the LX-2 + si-XIST + miR-663a inhi group delineated decreased miR-663a level (P < 0.01), raised GPX4 level (P < 0.001), repressed LPO, reduced Fe²⁺ and MDA levels (P < 0.001), elevated GSH level (P < 0.001), diminished cell death (P < 0.01), increased cell viability and α-SMA and CoL1A1 expression levels (P < 0.01), and enhanced cell proliferation (P < 0.01) (Figures 4A–H). These data indicate that lncRNA-XIST promotes HSC activation largely through a miR-663a/GPX4-dependent suppression of ferroptosis.