3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) and dimethyl sulfoxide (DMSO) were purchased from Sigma Aldrich (USA). Gly (98% purity) was purchased from Nanjing Corsas Medical Technology (Nanjing, China), Cis was purchased from Qilu Pharmaceutical (Jinan, China) Co., and arginine was purchased from Lablead Biotech, Co., Ltd. (Beijing, China). An aspartate aminotransferase (AST/GOT) test kit (C010-2-1), alanine aminotransferase (ALT/GPT) assay kit (C009-2-1), malondialdehyde (MDA) assay kit, catalase (CAT) assay kit, glutathione peroxidase (GSH-PX) assay kit and lactate dehydrogenase (LDH) assay kits were purchased from Nanjing Jiancheng Institute of Biological Engineering (Nanjing, China). Intracellular iron colorimetric assay kits were purchased from Wuhan Elabscience Biotechnology Co. Acridine orange (AO) assay kit, mitochondrial membrane potential (JC-1) assay kit, C11 BODIPY 581/591 lipid peroxidation fluorescent probe, mitochondrial reactive oxygen species (Mito-ROS) probe and intracellular ferroOrange fluorescent probe were purchased from Sigma‒Aldrich, Abcam, Invitrogen and Beiren Chemical Technology (Beijing) Ltd. BECN1, β-actin, FTH1, GPX4 and xCT antibodies were purchased from Cell Signaling Technology (Danvers, MA, US). LC3 antibody was purchased from Sigma‒Aldrich (L8918). LAMP1 antibody was purchased from SANTA CRUZ BIOTECHCHNOLOGY, INC. A glutamate release kit was purchased from Thermo Fisher Scientific (A12221, MA, US).

Forty specific pathogen-free (SPF) 8-week-old male C57BL/6J mice (weighing 20–25 g) were purchased from Beijing Bochiyuan Biotechnology Co., Ltd. (Beijing, China). All animals were housed in a light-controlled room (12 h light/dark cycle) with an ambient temperature of 25°C and free access to drinking water and standard food. All animals were humanely cared for according to institutional guidelines. This experiment was approved by the Animal Ethics Committee of North China University Of Science And Technology (2021-SY-043).

The animals (n = 40) were randomly divided into four groups (n = 10 for each group): the control group, Cis group, Cis + Gly-Arg low-dose treatment group (L-Gly-Arg, 100 mg/kg once every day for 7 days) and Cis + Gly-Arg high-dose treatment group (H-Gly-Arg, 200 mg/kg once every day for 7 days). The pharmacological Cis-induced liver injury model was constructed by intraperitoneal injection of Cis 20 mg/kg once in the Cis group 72 h before execution. The control group was injected with saline only. For the L-Gly-Arg and H-Gly-Arg groups, 100 mg/kg and 200 mg/kg Gly-Arg were given once every day for 7 days before the induction of acute liver injury models. Blood and liver tissue samples were collected for the subsequent studies.

Hepatocytes were isolated from mice by two-step collagenase perfusion as previously described (Storey et al., 2012; Fukuoka et al., 2017; Matsumura et al., 2019). First, after perfusion of Hanks’ balanced salt solution (HBSS, Sigma‒Aldrich) containing ethylene glycol tetraacetic acid without Ca or Mg was perfused at a rate of 3–4 mL/min through the portal vein. When the color of the liver was observed to turn beige or light brown, the liver was perfused with .3 mg/mL collagenase (type IV; Sigma Aldrich) in HBSS by the same route and at the same rate until the liver structure was significantly degraded. The isolated cells were suspended in high-glucose DMEM (Cyclone Utah, United States) containing 10% fetal bovine serum (FBS). Cells were then filtered through 100 sieve wells and purified by centrifugation (500 x g, 5 min) at 4°C. Hepatocyte viability was assessed by a trypan blue exclusion assay. Cells were incubated overnight at 37°C in a CO₂ incubator before being used for experiments.

Cells (3 × 10³ cells/well) were seeded in 96-well plates and incubated overnight at 37°C and treated with 1 µM Fe²⁺ inhibitor (Fer-1), 10 µM deferiprone (DFO), 10 µM 3-methyladenine (3-MA), 10 µM chloroquine (CQ), 10 µM necrostatin-1 (Nec-1), 10 µM Z-VAD-FMK (ZVAD) and 25, 50, 100, 150, or 200 µM Gly-Arg. Different concentrations of Cis were added for 24 h, followed by the addition of 5 mg/mL MTT solution (20 µL/well), and then the cells were incubated at 37°C in the dark for 4 h. Finally, the supernatant was removed, and 150 µL DMSO was added. Cell viability was measured at OD490 nm by a Multiskan™ microplate reader (Thermo Fisher Scientific, Inc., USA). IC50 values were calculated using GraphPad Prism 8.4 software (GraphPad software, Inc., USA).

Four micromolar thick tissue sections were dewaxed and treated with xylene. The sections were then sequentially stained with hematoxylin (3 min) and eosin (45 s). After washing with ddH₂O, sections were dehydrated with a concentration gradient of ethanol, treated with xylene, sealed with coverslips, and observed by light microscopy (Olympus, Japan).

Serum glutamate aminotransferase (ALT), glutamate aminotransferase (AST) and lactate dehydrogenase (LDH) were measured using commercial reagent kits (Nanjing Built) according to the manufacturer’s instructions.

The differently treated hepatocytes were fixed in 4% paraformaldehyde. The cells were then incubated overnight at 4°C with the following primary antibodies: anti-LAMP1 (Santa Cruz, sc-19992, 1:100), anti-FTH1 (CST, 4393, 1:100), anti-4-HNE antibody (Bioss Antibodise, bs-6313R; 1:100), anti-xCT (CST, 1269; 1:100), and anti-BECN1 (Proteintech, 66665-1). Subsequently, the cells were washed with PBS and incubated with FITC- and TRITC-coupled secondary antibodies and then examined by fluorescence microscopy (Keyence, Osaka, Japan). For liver tissue immunofluorescence, 4 µm thick frozen sections were prepared from the different experimental groups. The sections were washed with PBS and immersed in 3% BSA to block the nonspecific background. Sections were then incubated with rabbit polyclonal anti-4-HNE, -LAMP1, -FTH1, -XCT or -BECN1 primary antibodies at 4°C overnight. After washing 3 times with PBS, secondary antibodies were incubated for 2 h at room temperature. Sections were rerinsed with PBS, sealed with coverslips and observed under a fluorescence microscope (Keyence, Osaka, Japan).

Four-micrometer-thick paraffin-embedded liver tissue sections were heated to 60°C for 40 min and then subjected to xylene dewaxing and gradient ethanol hydration. The sections were blocked with 3% hydrogen peroxide and subjected to antigen repair. Next, the slides were washed with PBS 3 times for 5 min each, incubated with CD68 antibody (1:200) (CST, 97778; 1:100) overnight at 4°C, and then incubated with HRP-coupled secondary antibody for 30 min at room temperature. After washing with PBS, sections were treated with DAB solution for 20–35 s to generate peroxidase reaction products. After hematoxylin restaining, gradient ethanol dehydration, xylene transparency, and sealing, the sections were observed under a light microscope (Keyence, Osaka, Japan).

The liver tissue was cut into 1 mm³ sections, and the cells were lysed with RIPA buffer. After centrifugation at 12,000 rpm for 20 min, the supernatant was collected. The protein concentration in the supernatant was determined using the Pierce BCA Protein Assay Kit (23225, Thermo Fisher). Equal amounts of proteins (20 μg/lane) were separated by 10% SDS‒PAGE and then transferred to PVDF membranes. The PVDF membranes were then incubated at 4°C with the following primary antibodies: anti-LC3 (CST, 4599; 1:1000), anti-BECN1 (CST, 3495; 1:1000), anti-FTH1 (CST, 4393; 1:1000), anti-GPX4 (CST, 52455; 1:1000), anti-xCT (CST, 12691; 1.1000), anti-cleaved (c)-caspase3 (CST, 9664, 1:1000) and anti-β-actin (CST, 4970; 1:1000). PVDF membranes were washed with TBST buffer (137 mM NaCl, 2.7 mM KCl, 16.5 mM Tris, pH 7.4, containing 0.1% Tween-20) and incubated with secondary antibodies at room temperature for 2 h. After washing again with TBST, protein bands in the blots were detected using an ECL chemiluminescence system (GE Healthcare, US). β-actin was used as an internal control.

Primary hepatocytes were analyzed for MMP using the JC-1 staining method. In brief, primary hepatocytes were incubated with a JC-1 probe at a final concentration of 20 μM for 10 min at 37°C in the dark. Mitochondrial reactive oxygen species (mito-ROS) were measured using a final concentration of 5 μM MitoSOX™ red mitochondrial superoxide indicator (MitoSOX™, 5 μM, Invitrogen. M36008) at 37°C for 10 min in the dark. Cells were washed with PBS and immediately observed by fluorescence microscopy (Keyence, Osaka, Japan).

The LMP of hepatocytes was determined by acridine orange (AO) staining. Hepatocytes were incubated with AO (4 μg/mL; Sigma‒Aldrich, A6014) for 20 min at 37°C in the dark. The stained cells were washed twice with 3% FCS/PBS and immediately observed by fluorescence microscopy (Keyence, Osaka, Japan).

Primary hepatocytes were seeded at a density of 1 × 10⁵ cells in six-well plates at 37°C in a 5% CO₂ incubator overnight, and then cells were transfected with mRFP-GFP-LC3 adenovirus (Hanbio Biotechnology, Shanghai, China) at a multiplicity of infection (MOI) of 30 for 24 h, followed by Cis and Gly-Arg for another 24 h. The average number of GFP-LC3 and RFP-LC3 puncta per cell was counted under a fluorescence microscope (Keyence, Osaka, Japan). When red and green fluorescence were combined, yellow puncta represented autophagosomes, and red puncta represented autolysosomes (AL).

After conventional sampling, double fixation, dehydration, immersion, embedding, sectioning (1 mm thickness), and double staining with uranyl acetate and lead citrate, the liver tissue was observed by TEM as previously described (Cai et al., 2022).

Lipid ROS was measured using the C11 BODIPY 581/591 probe (RM02821, ABclonal, US). Isolated primary hepatocytes were seeded in 6-well plates and pretreated with 1 μM Fer-1, 10 μM DFO, and 100 μM Gly-Arg for 2 h as described above and then incubated with 10 μM BODIPY 581/591 C115 for 1 h in the dark after 24 h of Cis treatment. Following three washes in PBS, the slides were observed under a fluorescence microscope (Keyence, Osaka, Japan).

In brief, primary hepatocytes were pretreated with Fer-1, DFO, and Gly-Arg for 2 h, treated with Cis for 24 h, incubated with a final concentration of 1 μM FerroOrange (Dojindo, F374) at 37°C in a 5% CO₂ incubator in the dark for 30 min and immediately observed by fluorescence microscopy (Keyence, Osaka, Japan).

Glutamate release from primary hepatocytes into the extracellular medium was detected using the Amplex Red Glutamate Release Assay Kit (Thermo Fisher Scientific, A12221) according to the manufacturer’s instructions.

MDA, CAT and GSH-PX were measured using the Micro Malondialdehyde (MDA) Assay Kit, Catalase (CAT) Assay Kit, and Glutathione Peroxidase (GSH-PX) Assay Kit (Nanjing Jiancheng, Nanjing, China), respectively, according to the manufacturer’s requirements.

All data analyses were performed using GraphPad Prism version 8.4.2 software. Data are expressed as the mean ± standard deviation. Comparisons between two groups were made using the t-test. One-way analysis of variance (ANOVA) was used for studies with more than two groups. p values <0.05 were considered to indicate statistically significant differences.

In the livers of mice in the Cis group, pathological damage (such as variable morphology and size of hepatocytes) appeared, inflammatory cell infiltration was significantly increased, and serum ALT, AST and LDH levels were elevated. In comparison, the elevated ALT, AST and LDH levels in the Gly-Arg group were recovered, and histological changes were reversed, with the H-Gly-Arg group showing better effects than the L-Gly-Arg group (Figures 1A–F). The MTT assay revealed that Cis treatment decreased the viability of primary hepatocytes in a concentration- and time-dependent manner (Figures 1G, H). We further found that different concentrations of Gly-Arg could reverse the decrease in cell survival caused by Cis treatment, especially at 100 μM (Figure 1I). Next, we investigated the pathway Cis might induce primary hepatocyte death through. It was found that the addition of Fer-1 (a lipid peroxidation scavenger), DFO (an Fe²⁺ chelator), 3-MA (an autophagy inhibitor) and ZVAD (an apoptosis inhibitor) could reverse the Cis-induced decrease in cell viability to some extent relative to Cis treatment alone (Supplementary Figure S1). However, the addition of CQ and nec-1 (necrosis inhibitor) did not restore the Cis-induced decrease in cell viability of primary hepatocytes (Supplementary Figure S1). Therefore, we speculated that Cis treatment induces hepatocyte injury through apoptosis, iron death and autophagy pathways. Next, we investigated the specific mechanism underlying Gly-Arg amelioration of Cis-induced liver injury.

In the livers of Cis group mice, the level of an apoptosis marker, c-caspase 3, was elevated, but Gly-Arg did not significantly decrease the level of c-caspase 3, indicating that Gly-Arg did not alleviate Cis-induced hepatocyte apoptosis (Supplementary Figure S2). Next, we examined whether Gly-Arg alleviated Cis-induced liver injury by inhibiting autophagy. Western blot assays revealed that the expression levels of LC3II/LC3I and BECN1 were elevated after 24 h of Cis treatment in primary hepatocytes compared with those of the control group, suggesting that autophagy increased in primary hepatocytes after Cis treatment (Figure 2A). However, after pretreatment with Gly-Arg and an autophagy inhibitor (3-MA) for 2 h, LC3II/LC3I and BECN1 expression levels were decreased in primary hepatocytes compared with those of the Cis group (Figure 2B). The fusion of autophagic vesicles and lysosomes has an important role in autophagic flux, and sustained autophagic flux leads to autophagic death (Mizushima et al., 2010). Transfection of mRFP-GFP-LC3 adenovirus demonstrated that Cis treatment enhanced the formation of yellow puncta and red puncta, suggesting that it activated autophagic flux in primary hepatocytes (Figure 2C). In contrast, Gly-Arg significantly reduced Cis-induced yellow puncta and red puncta, indicating that Gly-Arg blocked autophagic flux in primary hepatocytes (Figure 2C). Moreover, LC3-II/LC3-I expression was significantly higher in the Cis-induced acute liver injury mouse model than that of the control group (Figure 2D). L-Gly-Arg and H-Gly-Arg significantly reduced the ratio of LC3-II/LC3-I in mouse liver tissues compared with the Cis-induced liver injury mouse model (Figure 2E). Therefore, we hypothesized that Gly-Arg could inhibit Cis-induced excessive autophagy activation.

Next, we investigated whether Gly-Arg rescued Cis-induced ferroptosis in primary hepatocytes. Compared to the control group, Fe²⁺ was significantly increased in primary hepatocytes after Cis treatment; however, the levels of Fe²⁺ were significantly lower in hepatocytes cotreated with Fer-1, DFO and Gly-Arg (Figures 3A, B). Moreover, the levels of lipid ROS in the primary hepatocytes were reduced after cotreatment with Fer-1, DFO, Gly-Arg and Cis compared with the Cis treatment alone (Figures 3C, D). In addition, the production of 4-hydroxynonenal (4-HNE), one of the major products of lipid peroxidation, was significantly reduced after pretreatment with Fer-1, DFO, and Gly-Arg in contrast to Cis treatment alone (Figures 3E, F). Additionally, the levels of MDA and Fe²⁺ were increased (Figures 3G, H), and CAT levels were decreased (Figure 3I), in the livers of mice in the Cis-group whereas advanced intervention with Gly-Arg alleviated these indices. IF staining showed that the accumulation of the lipid peroxidation product 4-HNE was increased in the Cis-group, whereas L-Gly-Arg and H-Gly-Arg were able to downregulate the accumulation of 4-HNE (Figure 3J). These results suggest that Gly-Arg rescues pharmacological Cis-induced liver injury by inhibiting ferroptosis.

Mitochondrial damage is also an important marker of ferroptosis (Wu et al., 2021). Therefore, we measured MMP in primary hepatocytes using JC-1 staining. Figure 4A shows that Cis treatment decreased JC-1 dimers and increased JC-1 monomers in primary hepatocytes, indicating that Cis treatment decreased MMP and disrupted mitochondrial function. In contrast, Fer-1, DFO, Gly-Arg and Cis cotreatment reversed these results (4A, 4B). Additionally, Fer-1, DFO, and Gly-Arg reduced the level of mito-ROS in primary hepatocytes compared to that of the Cis-group (4A, 4C). In addition, in the livers of control mice, TEM revealed a clear structure of mitochondrial cristae and few AL (Figure 4D). In the livers of the Cis group, mitochondrial cristae were reduced or disappeared, and the number of ALs was higher than the control group. In contrast, the Gly-Arg group had mildly smaller mitochondria, increased mitochondrial cristae, and less AL than the Cis-group (Figure 4D). Therefore, Gly-Arg resisted Cis-induced ferroptosis by alleviating mitochondrial damage.

During ferroptosis, the lysosomal permeability (LMP) of hepatocytes increases (Skouta et al., 2014). When the LMP increases, AO leaks from the lysosome into the cytoplasm, as evidenced by increased green fluorescence and diminished red fluorescence. We noticed that in Cis-treated hepatocytes, green fluorescence increased while red fluorescence decreased, indicating that Cis treatment enhanced LMP (Figures 5A, B). However, after Gly-Arg pretreatment, LMP was alleviated (Figures 5A, B). Free iron is required for lipid peroxidation and ferroptosis, but most of the intracellular iron remains bound to ferritin (Stockwell et al., 2017). During autophagy, ferritin is phagocytosed by lysosomes and then undergoes enzymatic degradation, resulting in the overproduction of free iron (Stockwell et al., 2017). Therefore, we investigated whether lysosomes and ferritin colocalize to elucidate the uptake of ferritin by lysosomes in hepatocytes. IF staining results showed considerable colocalization of FTH1 (green) and lysosome-associated membrane protein 1 (LAMP1) (red) in Cis-treated primary hepatocytes, indicating translocation of ferritin into lysosomes (Figures 5C, D). In contrast, Gly-Arg significantly disrupted the colocalization of the two (Figures 5C, D). In primary hepatocytes, the expression of FTH1 decreased after Cis treatment, while Gly-Arg reversed this decrease (Figure 5E). However, FTH1 expression in liver tissues of mice in the Cis group was opposite to that of primary hepatocytes cultured in vitro, which may be related to feedback regulation (Figure 5F). L-Gly-Arg and H-Gly-Arg caused a sustained increase in FTH1 expression compared to the Cis group (Figure 5G). Overall, these results suggest that Gly-Arg antagonizes Cis-induced ferroptosis by inhibiting lysosomal-mediated ferritinophagy.

The BECN1-xCT complex plays a key role in ferroptosis (Lee et al., 2022; Tan et al., 2022). In the current study, we found that Gly-Arg was able to reverse the Cis-induced increase in BECN1 expression. Therefore, we investigated whether Gly-Arg attenuates ferroptosis by regulating BECN1-mediated changes in the expression and activity of xCT and GPX4. After Cis treatment, the expression of GPX4 and xCT decreased in primary hepatocytes, and Gly-Arg was able to rescue this change (Figure 6A). In addition, we examined GSH and glutamate release in hepatocytes. As we expected, Cis reduced the release of GSH and glutamate in primary hepatocytes, while Gly-Arg was able to alleviate this change (Figures 6B, C). IF staining showed that colocalization between BECN1 and xCT occurred mainly in the cytoplasm of Cis-treated primary hepatocytes (Figures 6D, E), while Gly-Arg pretreatment reduced their colocalization. Similarly, GSH was downregulated in the livers of Cis group mice, whereas L-Gly-Arg and H-Gly-Arg reversed the Cis-induced decrease in GSH (Figure 6F). IF results showed considerable colocalization of xCT (green) and BECN1 (red) in the liver tissue of Cis group mice (Figure 6G). In contrast, xCT-BECN1 colocalization was reduced in both the L-Gly-Arg and H-Gly-Arg groups, and the effect of the high dose was better than that of the low dose (Figure 6G). Meanwhile, in mouse liver tissues, Cis caused upregulation of BECN1 expression and downregulation of xCT and GPX4 expression relative to the control group (Figure 6H). Moreover, L-Gly-Arg and H-Gly-Arg were able to reverse this Cis-induced upregulation of BECN1 expression and downregulation of xCT and GPX4 expression (Figure 6I). Thus, Gly-Arg can block Cis-induced ferroptosis by inhibiting the formation of the BECN1-xCT complex.

We used Tat-beclin 1, an agonist of BECN1, to determine whether it could reverse the protective effect of Gly-Arg. In mouse liver primary cells, Tat-beclin 1 upregulated BECN1 expression and inhibited xCT and GPX4 expression compared to Cis treatment (Figures 7A, B). However, Cis, Gly-Arg and Tat-beclin 1 cotreatment upregulated BECN1 expression and downregulated xCT and GPX4 expression relative to the Cis and Gly-Arg cotreatment group (Figures 7A, B). In addition, Cis, Gly-Arg and Tat-beclin 1 cotreatment showed an increase in LMP relative to the Cis and Gly-Arg cotreatment group (Figure 7C). In addition, Tat-beclin 1 reversed the Gly-Arg-induced increase in MMP (Figure 7D). Moreover, Tat-beclin 1 also reversed the alleviation of Cis-induced Fe²⁺ accumulation by Gly-Arg (Figure 7E). This suggests that BECN1 plays a key role in the protective effect of Gly-Arg against pharmacological Cis-induced liver injury.