293T-Nrf2-luc cell line was developed via lentiviral stable transfection of pGMLV-Nrf2-luc. pGMLV-Nrf2-luc is a kind of plasmid that designed by Genomeditech (Shanghai, China), which harbors Nrf2 promoter-luciferase reporter gene construct driven by multiple dioxin response elements. After that, 293T-Nrf2-luc cell line was obtained through antibiotic screening.

Both of 293T-Nrf2-luc cell lines and HepG2 cell lines were grown in DMEM/high glucose medium (Hyclone, United States) containing 10% fetal bovine serum (Hyclone, United States) and 1% penicillin-streptomycin solution (Hyclone, United States). The culture condition was constant temperature of 37°C, a humidified atmosphere of 95% air and 5% CO₂. The fresh media were replaced every 2 days.

293T-Nrf2-luc cell lines were first seeded in 96-well plates at a density of 8 × 10³ cells per well and grown overnight. Baicalin and iso-dihydromyricetin were prepared in our laboratory (Xiong et al., 2021; Zhao et al., 2021). Propofol (Fresenius Kabi, China), midazolam (Enhua, China), sufentanil (Humanwell, China), remifentanil (Humanwell, China), dexmedetomidine (Enhua, China), ketamine (Hengrui, China), dopamine (Fenghe, China),norepinephrine (Fenghe, China), epinephrine (Fenghe, China), diazepam (Xinyi, China), estazolam (Zhongxin, China), alprazolam (Enhua, China), rocuronium (Xianju, China), wogonin, liquiritin, anhydroicaritin, silybin A-SILYBIN-A, isoginkgetin, wogonoside, baicalein, scutellarein, scutellarin, hyperoside, liquiritigenin, glutamate, myricetin, dihydromyricetin, procyanidin (They were all purchased from Chengdu Purfield Biotechnology Co., Ltd. China), Baicalin and iso-dihydromyricetin were added at the final concentration of 10 μM, while sulforaphane (Sigma-Aldrich, United States) was used as a positive control (Russo et al., 2018). The negative control group was treated with the vehicle DMSO alone. The final concentration of DMSO in all experiments was below 1% (V/V).

Following 24 h incubation, luciferase activity was measured by the Steady-Glo^® Reagent (Promega, United States) according to the manufacturer’s directions. Luciferase activity was measured on Spectramax M3 (Molecular devices, United States) with luciferase production as readout. Data was shown as the percentage of fold induction of the test compound against the vehicle control group.

HepG2 cells were seeded in 96-well culture plates (1.0  ×  10⁴ cells/well) and grown overnight. To verify the effects of midazolam on HepG2 viability, cells were treated with midazolam at different concentrations (0, 1.25, 2.5, 5, 10, 20, 40 μM) for 24 h. The cells were incubated with CCK-8 (DOJINDO Laboratories, Japan) solution (20 μl/well) and cultured at 37°C for another 1 h. The absorbance of the dissolved solutions was detected at 450 nm on Spectramax M3.

In order to verify the protective effect of midazolam on CCl₄-induced HepG2 cells activity, different concentrations of CCl₄ (0, 0.5, 1, 2, 4, 6%) were added to culture for 24 h, and HepG2 cells activity was detected by CCK-8 to determine the IC₅₀ value of CCl₄. After 24 h of intervention with different concentrations of midazolam, IC₅₀ concentration of CCl₄ was added for another 24 h of intervention, and finally CCK-8 was performed to detect cell activity.

The antioxidant effect of midazolam was tested in adult C57BL/6 mice (male, 6–8 weeks, 18–22 g) provided by Shanghai Slake Experimental Animal Co., Ltd. All mice were fed in the laboratory at a temperature of about 22°C and a relative humidity of 50%, with a light-dark cycle for 12 h. A 1-week acclimatization period was adopted before the experiment, during which the mice also had free access to food and water. All animal experiments were approved by the animal ethics committee of Yueyang Hospital of integrated traditional Chinese and Western Medicine Affiliated to Shanghai University of traditional Chinese medicine (ethics No.: YYLAC-2020-0801).

Next, 32 mice were randomly divided into the following groups (n = 8 in each group): control, CCl₄ group (CCl₄, Sinopharm Chemical Reagent, China), CCl₄ plus midazolam 4 mg/kg/day (CCl₄ + M4 group); CCl₄ plus midazolam 8 mg/kg/day (CCl₄ + M8 group). In the CCl₄ group, mice were intraperitoneally (i.p.) injected with 0.5% (v/v) CCl₄ (10 ml/kg, dissolved in corn oil). In the CCl₄ + M4 group or CCl₄ + M8 group, midazolam was respectively i. p. injected at 4 or 8 mg/kg for 3 days, and 0.5% CCl₄ was i. p. injected at 2 h after the last dose of midazolam. Mice were sacrificed 24 h after CCl4 injection, and blood and liver samples were collected. A part of the liver was immediately fixed in 10% formaldehyde for histopathological observation, while the rest was frozen in liquid nitrogen and stored at −80°C for other experiments.

In order to explore the effect of midazolam on CCl₄-induced lethality, 40 mice were used and randomly divided into the CCl₄ group and CCl₄ + midazolam group (n = 20). Midazolam was i. p. injected at 8 mg/kg for 3 days, and 50% CCl₄ (2.6 ml/kg, dissolved in corn oil) were i.p. injected at 2 h after the last dose of midazolam. The mortality of mice in each group was recorded up to 36 h after CCl₄ administration.

The liver of each mouse (n = 8) was fixed in 10% neutral buffered formalin for 48 h and embedded in paraffin. Then 4 μm thick sections were cut and stained with hematoxylin and eosin (H&E) according to standard procedure. The samples were observed and images were obtained using an optical microscope (Olympus, Japan).

Blood samples were centrifuged at 3000 × g for 10 min at 4°C and the serum was collected to measure the activity of aspartate aminotransferase (AST) and alanine transaminase (ALT), which were measured by kits (Nanjing Jiancheng Bioengineering Institute, China). Liver tissue homogenization was performed using tissue homogenizer (FSH-2A, XFK, China). The homogenates were centrifuged (1,000 × g, 15 min) at 4°C, and the obtained supernatant was used for subsequent measurement of MDA and SOD by kits (Nanjing Jiancheng Bioengineering Institute, China). The levels of TNF-α and IL-6 in the liver tissues were measured using ELISA kits according to the manufacturer’s instructions (Elabscience, China).

Total proteins were isolated from the liver tissues and cells using protein extraction kit (Beyotime Biotechnology, China). The cell nucleus proteins were extracted using Nuclear and Cytoplasmic Protein Extraction Kit (Beyotime Biotechnology, China). Protein concentrations were measured using BCA protein assay kit (Beyotime Biotechnology, China). Equal amounts of protein from each sample were subjected to 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and the separated proteins were transferred to PVDF membranes. Then, the membranes were blocked with 1 × TBST containing 5% nonfat milk for 2 h at room temperature and continued to incubate with primary antibodies against mice Nrf2 (1:1000, Abcam, United Kingdom), HO-1 (1:1000, Abcam, United Kingdom), β-actin (1:1000, Beyotime Biotechnology, China) overnight at 4°C. Subsequently, the membranes were washed three times with TBST and incubated with secondary antibody at room temperature for 2 h. Blots were visualized using a hypersensitive chemiluminescence kit (Biotechnology, China). The results were quantified by densitometry using ImageJ software.

The 3D structure of Kelch-like ECH-associated protein 1(Keap1) was downloaded from the Protein Data Bank (PDB Code:1ZGK) (Beamer et al., 2005). AutoDockTools 1.5.6 was applied to pre-process of coordinate files of the Keap1 and ligand by adding polar hydrogens, calculating Kollman charges and assigning AD4 atomic type. Docking site of Keap1 was situated at the kelch-like site (KS) as well as the functional site of Keap1 described by Beamer et al. (Beamer et al., 2005). The grid box was placed at KS, and the docking simulations was conducted by AutoDock Vina (1.1.2). Further analyses of the docking result was conducted by Discovery Studio Visualizer (BIOVIA Discovery Studio 2019; Dassault Systèmes, San Diego, United States).

The data were analyzed with GraphPad Prism 7.0 (San Diego, CA, United States) and presented as the means ± standard deviation (S.D.). One-way ANOVA analysis was used for differences between groups. The survival rate was analyzed by the log-rank test p-values less than 0.05 were considered statistically significant.

Using the Nrf2 luciferease reporter construct, the active effect of Nrf2 by 30 kinds of therapeutic agents (10 μM, final concentration) were evaluated. The results showed that midazolam up-regulated the activity of Nrf2 more than 2-fold compared with the negative control group (p < 0.01) (Figure 1).

The expression of Nrf2 activity in 293T-Nrf2-luc cells increased with increasing midazolam concentration, and was dose-dependent (p < 0.01) (Figure 2A). To determine the effect of midazolam on cell viability, cells were treated with different concentrations of midazolam for 24 h. Cell viability was measured by CCK-8 assay. As shown in Figure 2B, even if treatment with high-dose midazolam (40 μM), there was no significant difference compared with the control group.

To further verify the inductive effect of midazolam on Nrf2, the protein levels of Nrf2 in 293T-Nrf2-luc cells with or without midazolam treatment were determined. As shown in Figure 3A, Western blotting analysis showed that the protein expression of Nrf2 was elevated with the increasing concentrations of midazolam. It is well-known that Nrf2 is a transcriptional factor that needs to migrate to the nucleus to exert its function (Dinkova-Kostova et al., 2002). Under basal conditions Nrf2 is largely bound in the cytoplasm to Keap1, which is anchored to the actin cytoskeleton. When inducers disrupt the Keap1–Nrf2 complex, Nrf2 migrates to the nucleus where, in heterodimeric combinations with other transcription factors, it binds to antioxidant response element (ARE) regions of phase two genes and accelerates their transcription (Dinkova-Kostova et al., 2002). After the proteins in the nucleus and cytoplasm were extracted respectively, western blot analysis showed that with the increase of midazolam concentration, the Nrf2 protein level in the nucleus increased (Figure 3B) and the Nrf2 protein level in the cytoplasm decreased (Figure 3C). Midazolam promoted the translocation of Nrf2 to the nucleus.

Next, the protective effects of midazolam on hepatocytes were investigated by using CCl₄-induced HepG2 cells were used as model cells. After pretreatment of HepG2 cells with different concentrations of midazolam for 24 h, CCl₄ (IC₅₀ = 1.5%, Supplementary Figure S1) was added to induce cell damage. As shown in Figure 4, CCl₄ significantly decreased HepG2 cells activity by more than 50% compared with the control group (p < 0.01). By contrast, midazolam could significantly improve HepG2 cells activity compared with the CCl₄ group (p < 0.01) even at low concentration (2.50 μM), as depicted in Figure 4. These findings clearly demonstrated that midazolam had good protective effects on CCl₄-induced hepatocytes.

Next, we evaluated the protective effects of midazolam (4 and 8 mg/kg/day, lasting for 3 days, i. p. injection) on CCl₄ induced acute liver injury in mice. As shown in Figure 5, compared with the normal mice group, the activity levels of ALT and AST in serum of the CCl₄ group were significantly increased (p < 0.01). In sharp contrast, the activity levels of ALT and AST in serum collected from midazolam pretreatment group were significantly decreased (p < 0.01), when compared with that from the CCl₄ group (Figures 5A,B). Furthermore, H&E staining also suggested that the hepatocytes of the CCl₄ group were swollen, while the liver cells were necrosis and a large number of inflammatory cells infiltrated. By contrast, following Midazolam pretreatment, the hepatocyte necrosis and the inflammatory cell infiltration were significantly reduced (Figure 5C). The histopathological scores showed a significant lower in both 4 and 8 mg/kg midazolam pretreatment compared to the CCl₄ group (p < 0.01, Figure 2S). These findings clearly demonstrated that midazolam could attenuate CCl₄ induced acute liver injury in mice.

As shown in Figure 6, Western blot analysis showed that the expression levels of Nrf2 and HO-1 in the CCl₄ group were down-regulated evidently when compared to that in the control group. However, following administration of midazolam, the protein levels of both Nrf2 and HO-1 were significantly enhanced. These findings clearly suggested that Nrf2-related signaling pathway in mice liver could be enhanced following administration of midazolam.

It is well-known that SOD and MDA are typical biomarkers of antioxidation and oxidative stress, respectively. As shown in Figure 7, compared with the control group, MDA levels in liver were increased (p < 0.05), while SOD activity levels were decreased at 24 h after CCl₄ treatment (p < 0.05). Compared with CCl₄ group, midazolam pretreatment group reduced MDA levels (p < 0.05) and increased SOD activity levels (p < 0.05) (Figures 7A,B). Meanwhile, it was also found that CCl₄ treatment significantly increased TNF-α and IL-6 expression levels compared with the control group (p < 0.01). As expected, midazolam dose-dependently reduced IL-6 expression levels (p < 0.01) when compared with that in the CCl₄ group, while the serum levels of TNF-α were also significantly reduced following administration with high dose midazolam (8 mg/kg) (p < 0.05) (Figures 7C,D).

To further verify the hepatoprotective effect of midazolam, we investigated the ability of midazolam to affect the survival rate of mice treated with lethal dose of CCl₄. For this purpose, a lethal dose of CCl₄ for mice (2.6 ml/kg, 50%, i. p.) was used, while midazolam (8 mg/kg) was administrated for three consecutive days (once per day). As shown in Figure 8, midazolam significantly improved the survival rate of mice when compared with the CCl₄ group (p < 0.05).

Finally, to decipher the binding modes and the molecular mechanism of midazolam in enhancing of Nrf2 signaling pathway, molecular docking simulations of midazolam into Keap1 were performed. Under homeostatic conditions, Nrf2 is kept inactive being bound to its endogenous inhibitor Keap1 (Bellezza et al., 2018). As shown in Figure 9, midazolam could be well-docked into the kelch-like site (KS) of Keap1, forming strong hydrophobic interactions, electrostatic and hydrogen bonding interactions with several key residuals in the inner wall of KS. Remarkably, the unsaturated and aromatic functional groups of midazolam interacted with Ala556 and Tyr525 via strong hydrophobic interactions including Alkyl, Pi-Alkyl and Pi-Pi stacked interactions (Figure 9). It was particular that three phenyl containing lipophilic residues (Phe577, Tyr572 and Tyr525) stretched the side chains orienting the sheer center direction of KS, resulting a regional hydrophobic surface of KS inner wall (Figure 9). In addition, the group of chlorobenzene and fluorobenzene were vital for the binding of midazolam. The group of chlorobenzene interacted with Arg415 via Pi-Cation interactions, while the fluorobenzene could interact with Gln530 via hydrogen bonding. These observations clearly suggest that midazolam could stably and tightly bind to Keap1, which in turn, blocking the coalition between Keap1 and Nrf2.