GSH standard was purchased from Shandong Jincheng Biopharmaceutical Co., Ltd. (Zibo, Shangdong, China). The standards of CYS, CYS-GLY, GLU, GLY, captopril (CAP), N-ethylmaleimide (NEM), and ¹³C-glutamine were purchased from Sigma-Aldrich Co. (St. Louis, MO, United States). Test kits for blood ammonia, glutathione disulfide (GSSG), GSH, Malondialdehyde (MDA), Superoxide dismutase (SOD), ROS, and BCA protein were purchased from Shanghai Beyotime Biotechnology Research Institute, China (Shanghai, China). ALT and aspartate AST assay kits were purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). Interleukin-6 (IL-6), interleukin-10 (IL-10), tumor necrosis factor-alpha (TNF-α), interleukin-1β (IL-1β) assay kits were purchased from Shanghai Excell Biological Technology Co., Ltd. (Shanghai, China). Primers of NOX4, 3-phosphate Glyceraldehyde dehydrogenase (GADPH), Ddit3, and ATF4 were purchased from Abcam Co., Ltd. (Cambridge, UK). NF-κB p65 antibodies were purchased from Cell Signaling Technology (Danvers, MA, United States). RNAiso Plus, Prime Script RT Master Mix, and SYBR Premix Ex Taq (Perfect Real Time) were purchased from Takara Biotechnology Co., Ltd. (Shiga, Japan). NOX4, Ddit3, and ATF4 antibody were purchased from Abcam Ltd. (Cambridge, UK). NF-κB antibody was purchased from Cell Signaling Technology (Danvers, MA, United States).

Male Sprague-Dawley (SD) rats (aged 8–9 weeks, 200–220 g) were purchased from the Sipper-BK laboratory animal CO., Ltd. (Shanghai, China) (approval number: SCXK (HU) 2013–0016), by guidelines of National Institutes of Health Guide for the Care and Use of Laboratory Animals. The rats were housed under controlled conditions (25°C, 55%–60% humidity, and 12 h light/dark cycle) with free access to laboratory food and water. All studies were conducted strictly with animal care regulations and guidelines and were approved by the China Pharmaceutical University Animal Care and Use Committee. The rats were habituated to the facilities for a week before the experiment.

To construct an AHE rat model, we injected TAA intraperitoneally into the rats at a dose of 250 mg/kg for three consecutive days. After AHE modeling, rats in the GSH-dosed group were intragastrically administrated with GSH at a dose of 250 mg/kg for two consecutive days. Besides, an individual experiment also investigated the therapeutic effect of intravenous administration of GSH. After AHE modeling, rats in the GSH-dosed group were intravenously administered with GSH at 10 mg/kg for two consecutive days. All rats were euthanized after 2 days of GSH treatment, then blood and tissues were collected.

The behavioral changes observed in rats were categorized into different levels to assess their neurological status. Here are the descriptions of each level:

Level 0: Rats exhibit normal or regular physical activity, display a startled response to sound stimuli, and demonstrate the presence of the righting reflex, indicating normal neurological function.

Level 1: Rats display mild drowsiness characterized by reduced spontaneous motor activity. However, they still exhibit a normal regular response to stimuli and retain the righting reflex.

Level 2: Rats show a lack of spontaneous motor activity but may exhibit some movements, reactions, and righting reflexes when aroused. Despite reduced activity, they maintain some level of responsiveness.

Level 3: Rats remain stationary and show reluctance to move. Their reactions to stimuli are significantly reduced, and they may demonstrate flipping reflexes when manipulated.

Level 4: Rats remain stationary with no observable reaction to stimuli. They do not display the righting reflex, indicating severe neurological impairment.

The same part of liver tissue slices from different groups were fixed in 4% Paraformaldehyde and then stained with Picro-Sirius red solution according to standard technique. The tissue sections were washed with acidified water and counterstained with Carazzi’s hematoxylin.

The levels of blood ammonia, ALT, AST in plasma, the levels of intrahepatic IL-6, IL-10, and the levels of intracerebral SOD, ROS, IL-6, IL-10, TNF-α were determined using the corresponding commercial biochemical kits following the manufacturer’s instructions. The results were calibrated using protein concentration.

Our previous study has established an LC-MS/MS method for the quantitative analysis of GSH, CYS, and CYS-GLY in the biological matrix (Chen et al., 2020). It used NEM derivatization according to established protocol (Giustarini et al., 2013). Briefly, 10 mg fresh tissues were homogenized with 100 μL buffer solution (6 mg/mL Tris, 0.2 mg/mL serine, 1.24 mg/mL boric acid, 4 μg/mL acivicin, and 7.76 mg/mL NEM). After adding 10 µL of internal standard solution (CAP, 5 μg/mL) into 50 µL of tissue homogenate, derivation by NEM was lasted under the light proof environment for 1 h, followed by adding 200 μL of methanol to precipitate proteins. After vortex-mixing for 5 min and centrifuging at 30,000 g for 10 min at 4°C, 50 μL supernatant was collected and analyzed using the LC-MS/MS system (SCIEX 6500, MA, United States) in positive ionization mode. The injection value was 5 μL and chromatographic separation was performed on a Sepax Bio-ODS SP column (5 μm, 4.6 mm × 150 mm). Mobile phase A was H₂O with 2 mM ammonium formate and 0.1% formic acid. Mobile phase B was methanol. The gradient elution was set as follows: 10% B at 0 min and held for 0.2 min, increased to 50% B at 0.4 min and held for 3.8 min, then decreased to 10% B at 4.5 min followed by 5 min for equilibration. The source conditions were set as follows: ion source gas 1, 60 psi; ion source gas 2, 50 psi; collision gas, 6 psi; curtain gas, 10 psi; ion-spray voltage, 4,500 V; source temperature, 550°C. The details of the optimized multiple reaction monitoring (MRM) parameters were summarized in Supplementary Table S1.

Quantitative analysis of GLU, GLY, glutamine, and taurine in tissues was performed according to the following procedure. Fresh tissue (10 mg) was homogenized with 100 μL of H₂O, followed by adding 400 μL of methanol solution containing internal standard (¹³C-glutamine, 15 μg/mL). After centrifugation at 40,000 g for 10 min, 300 μL supernatant was evaporated and reconstituted with 150 μL H₂O. Then the GLU, GLY, glutamine and taurine were analyzed using an LC-Q-TOF-MS system (SCIEX 5600, MA, United States) in negative ionization mode, and chromatographic separation was performed on a Waters Bridge Amide column (3.5 μm, 4.6 × 100 mm). Mobile phase A was H₂O with 5 mM ammonium acetate and 5% acetonitrile, and pH was adjusted to 9 with aqueous ammonia. Mobile phase B was acetonitrile. The gradient elution was as follows: 85% B at 0 min and held for 3 min, decreased to 30% B at 6 min, decreased to 2% B at 15 min and maintain for 3 min, then increased to 85% B at 19 min followed by balancing for 5 min. The MS parameters were set as follows: nebulizer gas, 50 psi; heater gas, 60 psi; curtain gas, 30 psi; declustering potential, 70 V; ion-spray voltage, −4.5 kV; source temperature 500°C. The mass scanning ranges of MS¹ and MS² were set at 100–1,000 and 50–1,000, respectively. CE value was set at 30 V with a spread of 20 V. The data were processed by MultiQuant 2.0 Software (Sciex, Concord, Ontario, Canada).

As to MDA, thiobarbitoric acid assay is the most commonly used method for MDA determination (Janero and Burghardt, 1988), and we used the MDA Assay Kit purchased from Shanghai Beyotime Biotechnology Research Institute (Cat. No. S0131S). Based on the kit instruction, firstly, tissues of rat brain cortex were homogenized with PBS (the proportion of tissue weight to PBS is 10%). After centrifugation at 12,000 g for 10 min, 100uL of the supernatant were collected as test samples. Next, added 100 µL of MDA standards (1, 2, 5, 10, 20, 50 µM) or test samples to 1.5 mL tubes, then added 200 µL of MDA working solution (containing 0.0925% thiobarbitoric acid), heating in a boiling water bath for 15 min. The water bath was cooled to room temperature, followed by centrifugation at 1,000 g for 10 min 200 μL of the supernatant were transferred to a 96-well plate, and absorbance was measured at 532 nm using a spectrophotometer. Finally, calculated the concentration of MDA according to the standard curve.

As to GSSG, we used the GSSG Assay Kit purchased from Shanghai Beyotime Biotechnology Research Institute (Cat. No. S0053), the theory of which was based on the established enzymatic recycling method (Rahman et al., 2007). According to the kit instruction, specifically, tissues of rat cortex were frozen with liquid nitrogen and then ground into powder. For every 10 mg of ground tissue powder, 100 uL of protein removal reagent M solution was added and homogenized. After being placed at 4°C for 10 min, homogenate was centrifuged at 12,000 g for 10 min, and 100 uL of supernatant were collected as test samples. Next, added 100 µL of GSSG standards (0.5, 1, 2, 5, 10, 15 μM) or test samples to 1.5 mL tubes, then added 20 μL of GSH Removal Buffer, and vortex-mixed immediately. Next, added 4 μL of GSH Removal Reagent to every 100 μL of sample, and vortex-mixed immediately. After incubating at 25°C for 60 min, measured absorbance at 412 nm and calculated the concentration of GSSG according to the standard curve.

Tissues of liver and brain were sectioned at a 10 μm thickness by using a Leica CM1950 cryostat (Leica Microsystem Ltd.) and mounted onto glass slides. Targeted MSI experiments were carried out on a DESI-MSI platform equipped with a DESI XS source and a tandem quadrupole mass spectrometers (Waters, Milford, MA, United States). The spray solvent was acetonitrile/water/formic acid (98:2:0.01, v/v/v, 2 μL/min). The MRM parameters for each analyte were summarized in Supplementary Table S2.

Our previous study has established a LC-MS/MS method for the quantitative analysis of amino acids in biological matrix (Shen et al., 2021). Ascorbic acid (20 mM) was added to 30 mg of the rat striatum. The mixtures were homogenized and sonicated in an ice bath. Then, 500 μL of ice-cold acetonitrile containing 200 ng of 2,5-dihydroxy benzoic acid was added to the homogenate. After drying under vacuum, 50 μL of borate buffer and 50 μL of benzoyl chloride were added for derivatization. The chromatographic separation was performed on an XBridge^® Amide column (3.5 µm × 4.6 mm × 100 mm, Waters). The mobile phase consisted of solvent A (0.2% formic acid and 5.0 mM ammonium formate in water) and solvent B (acetonitrile). The MS was operated using the LC-MS/MS system (SCIEX 6500, MA, United States) in positive mode. The MS parameters were set as follows: ion source gas1, 55; ion source gas 2, 50; curtain gas, 20; CAD gas, medium; ion spray voltage, 5.0 kV; source temperature, 550 °C. Supplementary Table S3 summarized the MRM monitoring conditions for each analyte.

Total RNA was isolated from rat cortex with RNAiso plus and Q-PCR analyses were performed using SYBR Premix Ex Taq following the manufacturer’s protocols. The primers used were summarized in Supplementary Table S4. The Q-PCR cycle conditions were at 95°C for 30 s, followed by 40 cycles at 95°C for 5 s and at 60°C for 30 s. Then the amplification specificity was evaluated with melting curve analysis.

Protein levels of NADPH Oxidase 4, NF-κB p65, Ddit3, and ATF4 in rat cortex were measured by Western blot analysis. Equal loading of proteins was verified by immunoblotting of GAPDH. Gel-Pro analyzer was used for the semi-quantitative analysis of the proteins obtained from gel imaging system.

All the results were shown with mean ± SD of 6–8 mice per group. Data were analyzed by using two-tailed t-test, or one-way ANOVA test with Student-Newman-Keuls test for comparison of two groups. Values of p < 0.05 were considered statistically significant.

Previous research has unequivocally validated the pivotal role of GSH in conferring significant protection against liver and brain injuries. To further enhance the clinical utility of GSH, we investigated the therapeutic effect of GSH on AHE in rats via both oral and intravenous administration modes. Firstly, we investigated the hepatoprotective effect of exogenous GSH on AHE rats. In this process, TAA-induced AHE rats were orally administered GSH or physiological saline for two consecutive days, followed by euthanasia under anesthesia 24 h after the final administration. Behavioral assessments revealed significantly elevated scores in AHE rats compared to the control group, with oral GSH administration markedly ameliorating the scores (Figure 1A). Furthermore, oral GSH administration effectively mitigated the substantial rise in blood ammonia concentration caused by TAA (Figure 1B). AHE modeling led to notable hepatic injury, characterized primarily by a significant elevation in serum ALT and AST levels, which were dramatically reduced by oral GSH administration (Figures 1C,D). Subsequently, Hematoxylin and eosin (H&E) staining was used to estimate the liver histopathological changes in different groups. As shown in Figure 1E, AHE rats exhibited severe inflammatory cell infiltration in the liver accompanied by tissue hemorrhage, which was effectively alleviated by oral GSH administration. AHE modeling significantly increased pro-inflammatory factor IL-1β protein levels and downregulated anti-inflammatory factor IL-10 in rat liver (Figure 1F, G). Oral GSH administration extensively modulated intrahepatic inflammatory factors in AHE rats to the normal controls. Therefore, exogenous GSH could exert significant hepatoprotective effects in AHE rats.

According to prior reports, amino acid profiles dysregulation is impilicated in the onset and progression of AHE, with the BCAAs/AAAs ratio serving as a clinical prognostic indicator (Holecek, 2015; Hadjihambi et al., 2018). To investigate the influence of oral GSH on AHE, concentrations of AAAs (phenylalanine (Phe), tyrosine (Tyr), and tryptophan (Trp)) and BCAAs (leucine (Leu) and valine (Val)) in rat plasma were quantitatively analyzed using the LC-MS/MS assay previously established (Shen et al., 2021). As shown in Figure 1H, I, AHE modeling led to a noticeable elevation in AAAs and a significant reduction in BCAAs. Oral GSH administration restored AAAs and BCAAs in AHE rat plasma to conventional levels, thereby maintaining amino acid homeostasis.

The therapeutic effect of intravenous GSH injection in AHE was also investigated in rats. Notably, intravenous GSH administration at a dose of 10 mg/kg significantly improved the behavioral scores of AHE rats (Figure 1J). Additionally, intravenous GSH administration effectively mitigated the notable increase in plasma ammonia concentration caused by TAA (Figure 1K). AHE modeling resulted in a significant elevation in serum ALT and AST levels. However, intravenous GSH administration of GSH did not significantly reduce the levels of ALT and AST in the serum of AHE rats (Figure 1L, M). Thus, the hepatoprotective effect of intravenous administration of 10 mg/kg of GSH on AHE rats was considerably lower than that of oral administration of 250 mg/kg of GSH.

After determining the hepatoprotective effect of GSH, we further investigated its cerebral protective effects on AHE rats. The protein and mRNA levels of intracerebral inflammatory factors (IL-6, TNF-α, and IL-10) were determined using Elisa and Q-PCR assays. In this process, we compared the alterations of inflammatory factors across different brain regions (hippocampus, striatum, and cortex) and identified a strong correlation between inflammatory factors in the cortex and the onset/treatment of AHE. Notably, AHE modeling significantly upregulated the pro-inflammatory factors IL-6 and TNF-α in the rat cortex, while down-regulating the anti-inflammatory factor IL-10. Oral GSH administration effectively modulated intracerebral inflammatory factors in AHE rats towards the conventional levels (Figure 2A–F). Additionally, cortical tissues from different groups were subjected to TUNEL staining. AHE model rats showed a substantial number of stained condensed nuclei, indicating significant cell apoptosis. GSH administration notably reduced the number of stained condensed nuclei and attenuated cell apoptosis (Figure 2G). Intracerebral glutamine, a toxic substance, was quantified using LC-MS, revealing a marked accumulation in AHE-induced rats. Conversely, oral GSH administration significantly alleviated glutamine accumulation (Figure 2H). Given that GS catalyzes the synthesis of glutamine from ammonia, we further investigated the effect of GSH on GS enzyme activity. Evidently, AHE modeling markedly upregulated GS expression, while GSH administration significantly reversed the upregulation (Figure 2I). Moreover, we assessed the concentration of taurine, which maintains sodium ion balance in the cortex of conventional AHE model and GSH-treated rats. The results demonstrated that GSH administration significantly reversed the reduction in intracerebral taurine exposure caused by AHE modeling (Figure 2J). Therefore, exogenous GSH exerted significant cerebral protective effects in AHE rats.

In general, GSH could be hydrolyzed to cysteine-glutathione (CYS-GLY), glutamate (GLU), CYS, and GLY in vivo (Chen et al., 2020). To further elucidate the pharmacological substances contributing to the therapeutic effects on AHE, the exposure of GSH-derived ingredients (GSH, CYS, GLU, GLY, and CYS-GLY) in target tissues was determined using LC-MS/MS and LC-Q-TOF/MS systems. As shown in Figure 3A–D, AHE modeling led to a considerable reduction in GSH and CYS exposure in the rat liver and brain, whereas oral GSH administration significantly increased intracellular GSH and CYS levels. Furthermore, the concentrations of GSH and CYS across different brain regions, including cortex, hippocampus, and striatum, were quantitatively analyzed to further determine the specific action sites of GSH. As shown in Figure 3E, F, we observed significantly lower exposure of GSH and CYS in the cortex of AHE rats compared to controls, with oral GSH administration of GSH effectively elevating their levels in this region. However, neither AHE induction nor GSH administration had a discernible effect on GSH and CYS exposure in the hippocampus and striatum of rats (Supplementary Figure S1G–J). DESI-MSI analysis further supported our findings, demonstrating that GSH levels decreased in the liver and cerebral cortex due to AHE modeling but were restored by oral GSH administration. Additionally, the MSI results of CYS, glutamine, and taurine were consistent with previous data (Figure 3G, H). We also investigated the influence of AHE modeling and GSH administration on the targeted exposure of CYS-GLY, GLY, and GLU. Our results indicated no significant differences in intrahepatic CYS-GLY, GLY, and GLU levels among control, AHE, and GSH-treated rats (Supplementary Figure S1A–C). Similarly, neither AHE induction nor GSH administration had a pronounced effect on the intracerebral exposure of CYS-GLY and GLY, although there was some regulatory effect on GLU (Supplementary Figure S1D–F). Following 2 days of intravenous GSH administration, intrahepatic GSH and CYS levels increased to a certain extent, although there was no significant difference between the GSH-treated and AHE rats (Supplementary Figure S2A, B). Similarly, intracortical GSH and CYS levels were significantly lower in AHE rats than in controls, with intravenous GSH administration effectively increasing their content in the cortex (Supplementary Figure S2C, D). However, intravenous GSH administration of GSH had no obvious effect on GSH and CYS exposure in the hippocampus and striatum of rats (Supplementary Figure S2E–H).

To further delineate the action sites of GSH, we studied its distribution in other tissues of AHE rats. Concentrations of GSH in various rat tissues (heart, kidneys, duodenum, jejunum, ileum, and colon) were quantitatively analyzed at 2.5, 8, and 24 h after administration. As shown in Figure 4, AHE modeling had no obvious effect on GSH concentrations in the heart, kidney, and intestines, while oral GSH administration significantly enhanced the exposure of GSH in the intestine of AHE rats. Particularly, GSH exposure in the ileum increased substantially after oral GSH administration. Similarly, oral GSH administration also greatly elevated GSH exposure in rat colon. Conversely, oral GSH administration had no significant influence on GSH levels in the heart and kidney of AHE rats. Intratissue concentrations of GSH in AHE rats after intravenous GSH administration were also measured, revealing no significant effect on GSH exposure in heart, kidney, and intestinal segments (Supplementary Figure S3).

Oxidative stress is a ubiquitous phenomenon in the body and its exacerbation is associated with several diseases (Simicic et al., 2022). However, the intricate interplay between oxidative stress regulation and the pathophysiology/treatment of AHE remains largely unexplored. Therefore, we endeavored to elucidate the molecular mechanism underlying GHS treatment for AHE by examining its effects on cortical oxidative stress in rats. Firstly, the ROS level in rat cortex was detected by DCFH-DA fluorescent probes. The results revealed a marked increase in ROS following AHE modeling, which were effectively restored to baseline levels upon GSH administration (Figure 5A). Moreover, the GSSG/GSH ratio has consistently been used to indicate oxidative stress in cells and tissues (Skowrońska and Albrecht, 2013). Herein, the concentrations of GSSG and GSH in rat cortex were measured, and the results demonstrated that AHE modeling led to a noticeable increase of GSSG/GSH ratio, which was substantially reduced following GSH administration (Figure 5B). Similarly, GSH administration mitigated the increase in MDA levels caused by AHE modeling (Figure 5C). Additionally, SOD, a crucial antioxidant enzyme responsible for scavenging superoxide anion free radicals, was found to be downregulated in the AHE rat cortex. GSH administration effectively reversed this downregulation of SOD (Figure 5D).

NOX4, a major source of ROS, has been identified in brain damage. However, its role in regulating the development and treatment of AHE remains elusive. Thus, the protein expression of NOX4 in control, AHE, and GSH-treated groups were measured. The results demonstrated a significant upregulation of NOX4 expression following AHE induction, which was reversed upon GSH administration (Figure 5E, F). Furthermore, NF-κB, which modulates gene expression in various cellular processes including immune response, cell proliferation, and stress response, has been implicated in inflammation and oxidative stress modulation (Skowrońska and Albrecht, 2013). Notably, the protein expression of NF-κB P65 in the cortex of rats showed a significant upregulation post-AHE modeling, which was effectively reversed by GSH administration (Figures 5E, G).

Microglia, the foremost active defense cells in the CNS, usually maintain a quiescent state. Herein, we performed immunofluorescence staining on microglia in rat cortex to observe their morphological changes. As shown in Figure 6A, the microglia in the rat cortex became active after AHE modeling, characterized by increased density, rounded shape, and diminished branching. Conversely, in the cortex of rats administered with GSH, microglial density decreased, and the number of branches increased, indicating a significant inhibition of microglial activation by GSH. Microglial activation can be broadly categorized into classical (M1: pro-inflammatory) or alternative (M2: anti-inflammatory) phenotypes. Generally, M1 activation state is associated with the release of pro-inflammatory cytokines, with genes such as iNOS, TNF-α, and IL-1β being key markers. Subsequently, we measured the expression of M1 microglia-related genes in the rat cortex of the control, AHE, and GSH-treated groups. The results demonstrated that iNOS, TNF-α, and IL-1β levels were significantly elevated after AHE modeling and greatly reduced by GSH administration (Figure 6B). Notably, iNOS exhibited the highest degree of regulation following both AHE modeling and GSH administration. To further elucidate the role of iNOS in AHE treatment, we employed the iNOS inhibitor N(ω)-nitro-L-arginine methyl ester (L-NAME) and observed its impact on cell viability in a NH₄Cl-simulated human malignant cell line U251. L-NAME could significantly enhance the cell viability in a dose-dependent manner (Figure 6C). Additionally, we measured the NO conctent in the rat cortex of different groups. It was found that GSH administration significantly reduced the abnormal elevation of NO induced by AHE modeling (Figure 6D).

Since apoptosis by the ER stress pathway could be enhanced by the iNOS, we measured ER stress-related genes (ATF4, IRE1α and ATF6) to further verify the crucial role of iNOS in AHE occurrence and treatment. As shown in Figures 6E–G, GSH treatment significantly reversed the upregulation of ATF4 caused by AHE modeling, while there was no obvious effect on the expression of IRE1α and ATF6. We then investigated the effects of AHE modeling and GSH administration on the target proteins (Ddit3 and Xbp1) of ATF4. Clearly, GSH treatment significantly reversed the upregulation of Ddit3 caused by AHE modeling, with no notable effect on Xbp1 mRNA expression (Figures 6H–K). Therefore, GSH administration might mitigate AHE by regulating ER stress through the iNOS/ATF4/Ddit3 pathway.

M2 activation state expresses anti-inflammatory cytokines, arginase-1 (Arg1), transforming growth factor β1 (TGFβ1), CD206 and Chitinase-3-like-3 (YM-1 in rodents) (Skowrońska and Albrecht, 2013). We investigated the influence of AHE modeling and GSH administration on the expression of Arg1, CD206 and YM-1 in the rat cortex. However, our results indicated no significant difference in the levels of M2 microglia-related genes among the groups, suggesting that neither AHE modeling nor GSH administration induced alternative activation of microglia (Figure 6L).