Hydrogen Sulfide Protects Against Ammonia-Induced Neurotoxicity Through Activation of Nrf2/ARE Signaling in Astrocytic Model of Hepatic Encephalopathy

Objective: Hepatic encephalopathy (HE) characterized by neuropsychiatric abnormalities is a major complication of cirrhosis with high mortality. However, the pathogenesis of HE has not been fully elucidated. This study aimed to determine endogenous hydrogen sulfide (H2S) in the blood of HE patients and investigate the role of H2S in an astrocytic model of HE. Methods: Patients with and without HE were recruited to determine plasma H2S levels and blood microbial 16S rRNA gene. Rat astrocytes were employed as a model of HE by treatment of NH4Cl. Exogenous H2S was preadded. Cell viability was measured by Cell Counting Kit-8 (CCK-8) assay, and cell death was evaluated by lactate dehydrogenase (LDH) release. Apoptosis was determined by Hoechst 33342/Propidium Iodide (PI) Double Staining and Western blot analysis of apoptosis-related protein expression. Intracellular reactive oxygen species (ROS) levels were assessed by flow cytometer. Expressions of Nrf2 and its downstream regulated genes were examined by immunofluorescence staining and Western blot, respectively. Nrf2 gene knockdown was performed by antisense shRNA of Nrf2 gene. Results: There was a significant decrease in H2S levels in cirrhotic patients with HE compared with without HE. Blood microbiota analyses revealed that certain strains associated with H2S production were negatively correlated with HE. In vitro, H2S markedly attenuated NH4Cl-induced cytotoxicity, oxidative stress, and apoptosis. This effect was mediated by Nrf2/ARE signaling, and knockdown of Nrf2 expression abolished the antagonistic effect of H2S on NH4Cl-induced neurotoxicity in astrocytes. Conclusion: Levels of H2S and bacteria associated with H2S production are decreased in HE, and H2S functions as the neuroprotector against NH4Cl-induced HE by activating Nrf2/ARE signaling of astrocytes.


INTRODUCTION
Hepatic encephalopathy (HE) is a common complication of cirrhosis, leading to low quality of life and high mortality (Vilstrup et al., 2014). Feature of HE is a neuropsychiatric syndrome covering a broader range of disturbances including alterations in intellectual function, conscience, and motor function and coordination (Bustamante et al., 1999). Excessive ammonia generated by enteric bacteria crossing the blood-brain barrier and causing astrocyte swelling is a key driver of HE (Williams, 2007;Felipo and Butterworth, 2002). However, its pathogenesis is incompletely understood, and effective clinical treatments are still in development.
Oxidative stress (OxS) plays a major role in brain injury in a patient with cirrhosis (Heidari et al., 2018;Swaminathan et al., 2018). During ammonia toxicity, nicotinamide adenine dinucleotide phosphate (NADPH) oxidase isoforms (Poznyak et al., 2020) and the mitochondrial permeability transition pore (Bai et al., 2001;Rama Rao et al., 2003) are the main sources for the reactive oxygen species (ROS). Astrocyte swelling induced by intracellular glutamine accumulation triggers OxS through an Nmethyl-D-aspartic acid (NMDA) receptor-and Ca 2+ -dependent mechanism, and in turn, the activation of NMDA receptor and OxS enhances astrocyte swelling. This self-amplifying cycle between astrocyte swelling and OxS results in neuronal dysfunction, ranging from trivial lack of awareness to coma (Häussinger et al., 1994, Häussinger, 2006Zielińska et al., 2003;Reinehr et al., 2007). Nuclear factor erythroid 2-related factor 2 (Nrf2), termed the master regulator of antioxidant responses, is a transcription factor found to be frequently dysregulated in OxS (Arefin et al., 2020). During OxS, the static binding between Nrf2 and Kelch-like ECH-associated protein 1 (Keap1) in the cytoplasm is dissociated, which allows the translocation of Nrf2 into the nucleus, and Nrf2 interacts with the antioxidant response element (ARE) to initiate the transcription of target genes to alleviate OxS (Shen et al., 2015). Interestingly, Nrf2 has been suggested as a crucial target of H 2 S and a vital mediator for H 2 S to attenuate OxS (Keum, 2011;Liu et al., 2012;Li et al., 2013;Yang et al., 2015). Therefore, Nrf2 might play an important role in the protection of H 2 S on HE.
In the current study, we determine plasma H 2 S levels and the microbiota associated with H 2 S production in the blood of HE patients. Meanwhile, we try to investigate the neuroprotective effect of H 2 S against ammonia toxicity to astrocyte and the role of the Nrf2/ARE signaling pathway in this putative cytoprotective function.

Ethics Statement
The clinical study protocol was approved by the Human Research Ethics Committee, the First Affiliated Hospital, Wenzhou Medical University, China. All participants or guardians signed the consent form in accordance with the Declaration of Helsinki. The animal experiments were approved by the Animal Ethics Committee and carried out in accordance with the established Guiding Principles for Animal Research.

Study Subjects
All cirrhotic patients involved in this study were diagnosed through biopsy and/or radiological evidence.
We included cirrhotic patients with exclusion of individuals in coma (including HE grade 4); individuals on current or past specific treatment for HE; individuals with malignancy, heart failure, hematological or autoimmune diseases, and HIV infection; and individuals with organ transplants. The healthy controls exhibited no disease symptoms. Individuals who had consumed alcohol within 3 months, individuals with systemic antibiotics within 6 weeks, and individuals with yogurt/probiotic consumption within 2 weeks were also excluded. All participants were recruited from the First Affiliated Hospital of Wenzhou Medical University, and cirrhotic patients were divided into two groups according to with/without HE (Ferenci et al., 2002). The diagnosis of HE depended on clinical symptoms of brain dysfunction based on a careful and comprehensive neuropsychiatric evaluation addressing consciousness, orientation, cognitive function, and sensory and motor function, together with the knowledge of the patient's history.

Clinical Data Collection
All clinical data were collected through face-to-face interviews with hepatologists. All measurements and questionnaires were voluntary. Weight and height of all subjects were measured by an attending physician, and then the body mass index (BMI) was calculated. Blood ammonia, biochemistry, and coagulation index were determined at the hospital biochemistry laboratory of the First Affiliated Hospital of Wenzhou Medical University. The Model for End-Stage Liver Disease (MELD) score was calculated according to a previous study (Malinchoc et al., 2000).
NaHS was used as an H 2 S donor. When NaHS is dissolved at pH 7.35-7.45, HS − is released and binds to H + to form H 2 S. This provides a solution of H 2 S at a concentration at about one-third of the original concentration of NaHS (Reiffenstein et al., 1992).

Plasma Levels of H 2 S Assay
Plasma was prepared after blood collection, and H 2 S level was measured within 2 days after collection by employing a human H 2 S ELISA kit (Wuhan MSK Biotechnology Company Limited, China) with a microplate reader at a wavelength of 450 nm.

Detection of Microbial 16S rRNA Gene in the Blood
Blood samples were collected and flash frozen. As reported previously (Païssé et al., 2016), microbial DNA was extracted. Universal primers linked with indices and sequencing adaptors were used to amplify the V4 regions of the 16S rDNA. Agencourt AMPure XP magnetic beads were used to purify the PCR amplification products and then labeled to complete the establishment of the library. The amplicon sequencing libraries were sequenced on an Illumina platform and clustered into operational taxonomic units (OTUs) for further microbial analysis.

Identification of Bacteria Associated With H 2 S Production
Sulfur occurs in various oxidation states ranging from +6 in sulfate to −2 in sulfide. The pathway map of the Kyoto Encyclopedia of Genes and Genomes (KEGG) 1 reveals that the oxidation states of sulfur from +6 to −2 (the production of H 2 S) are through the energy consuming assimilatory and the energy producing dissimilatory pathways. Gene information including the lineage was known by gene search on the National Center for Biotechnology Information (NCBI) 2 . Bacteria carrying the gene involved in the pathway above were identified as association with H 2 S production, and the identification was extended from species to higher taxonomic levels. For example, Odoribacter splanchnicus is known to produce H 2 S, and this designation was extended to all members of the Odoribacter genus (Nguyen et al., 2020).

Primary Astrocyte Culture and Treatment
Postnatal 1-to 2-day Sprague-Dawley (SD) rats were provided by the Laboratory Animal Center of the First Affiliated Hospital of Wenzhou Medical University. Primary astrocytes were prepared from the cerebral cortices of SD rats as described previously (Yin et al., 2011;Yu et al., 2015;Yuntao et al., 2016). Briefly, cerebral cortices were freed of meninges by microscope, minced, dissociated with 0.125% trypsin (Gibco, Grand Island, USA) for 15 min, passed through sterile nylon sieves, and placed in Dulbecco's modified Eagle's medium (DMEM; Gibco, Grand Island, USA) with 10% fetal bovine serum (FBS; Gibco, Grand Island, USA) and 1% penicillin/streptomycin (Gibco, Grand Island, USA). After centrifugation at 500 g for 5 min, the cell pellets were resuspended and seeded on dishes. The culture was maintained at 37 • C in a humidified 5% CO 2 /95% air incubator. The culture medium was changed every 4 days. Upon reaching confluence (10-12 days), cells were harvested for further research. Immunostaining revealed that >95% of cells were glial fibrillary acidic protein (GFAP)-positive astrocytes (Supplementary Figure 1).
In most experiments, confluent cells were detached from dishes and seeded into new dishes with complete culture medium, which were divided into four groups randomly: control, H 2 S, NH 4 Cl, and H 2 S + NH 4 Cl groups. The NH 4 Cl group was used as a cellular model of HE , where astrocytes were incubated in media with NH 4 Cl. Cells in the H 2 S + NH 4 Cl group and H 2 S group were pretreated with NaHS for 1 h, washed twice with phosphate-buffered saline (PBS), and then respectively incubated in complete culture medium with and without NH 4 Cl. This excludes the effect of H 2 S as direct reductant or oxidant scavenging action. Incubation time of cells in NH 4 Cl is depending on the approach before further analysis.

Cell Viability Assay
Cell viability was measured with the CCK-8 assay. Astrocytes were seeded into a 96-well culture plate at a density of 1 × 10 4 cells/well in 100 µl of culture medium overnight and then were treated as above for 24 h. CCK-8 solution (10 µl) was added to each well of the plate and incubated for 2 h in an incubator. Absorbance was measured at 450 nm using a microplate reader.

Lactate Dehydrogenase Release Assay
LDH release assay (Decker and Lohmann-Matthes, 1988) was also performed to measure cytotoxicity. Briefly, after cells were treated as above for 24 h, cell medium was transferred into a 96-well plate, and LDH release kit was used to detect LDH release activity of damaged cells. Absorbance was measured at 450 nm.

Assessment of Intracellular Reactive Oxygen Species Generation
Intracellular ROS levels were examined using the DCFH-DA staining method based on the conversion of non-fluorescent DCFH-DA to the highly fluorescent DCF upon intracellular oxidation by ROS. This cell-permeable fluorogenic probe is useful for the detection of H 2 O 2 , O − 2 , and OH − and for the determination of the degree of overall OxS. At least 2 × 10 5 cells in each group were treated as indicated for 4 h, and 10 µmol/l DCFH-DA in serum-free medium was added and incubated for 25 min at 37 • C in the dark. DCF fluorescence was measured using a flow cytometer with excitation at 484 nm and emission at 530 nm.

Detection of Apoptosis by Hoechst 33342/PI Double Staining
Hoechst 33342/PI Double Staining was used to detect cell apoptosis . Astrocytes were seeded at 2 × 10 4 cells/well in 24-well plates. After the indicated treatments for 8 h, cells were stained with Hoechst 33342 and PI dye according to the manufacturer's protocol. The percentage of apoptotic cells was observed using fluorescence microscopy.

Immunofluorescence
Astrocytes were cultured in 24-well plates on glass slides at a density of 2 × 10 4 cells/well with complete culture medium and cultured to 50-60% confluency. After treatment for 24 h, the cells were fixed in 4% paraformaldehyde for 15 min and then permeabilized with 0.3% Triton X-100 for 20 min. After blocking with goat serum for 2 h at 37 • C, cells were incubated with primary antibodies against Nrf2 (1:500) at 4 • C overnight. This was followed by incubation with goat anti-rabbit IgG H&L (1:1,000) for 1 h and counterstaining with DAPI (Thermo Fisher Scientific, Waltham, MA, USA). Cells were observed with a fluorescence microscope.

Knockdown of Nrf2 Expression With Small Hairpin RNA
Adenovirus with antisense of Nrf2 small hairpin RNA (shRNA) was obtained from Genechem Company Limited, Shanghai, China). The sequences of the two shRNAs for Nrf2 target genes were as follows: shRNA1, 5 -aaGCAGCATACAGCAGGACAT-3 and shRNA2, 5 -gaGCAAGAAGCCAGATACAAA-3 . Astrocytes were seeded in 6-well plates at a 50-60% confluence, and cells were transfected with adenovirus at a multiplicity of infection (MOI) of 100 for 5 h later, then supplemented with fresh medium, and continuously cultured for an additional 48 h. Transfection efficiency was confirmed by Western blot analysis of Nrf2 protein expression.

Statistical Analysis
Data were expressed as median (interquartile range) or mean ± SEM. The comparisons of H 2 S levels, blood ammonia, ages, BMI, and MELD score were made using Kruskal-Wallis test. The correlation between plasma H 2 S level and HE grade was using Spearman. The comparison of the relative abundance of the HE and nonhepatic encephalopathy (NHE) patients was performed using the Wilcoxon rank-sum test. Multiple comparisons in astrocyte cultures were carried out using one-way ANOVA, Dunnett's test, or the Student-Newman-Keuls (SNK) test as a post hoc test, as appropriate. All statistical analyses were performed using SPSS (22.0). Values of P < 0.05 were considered statistically significant.

Characteristics of the Studied Groups
A total of 89 subjects (healthy = 26, HE = 30, and NHE = 33) were studied. There was no difference in age or BMI among the three groups, nor was MELD score in HE and NHE patients ( Table 1). As expected, the blood ammonia level in the HE group was significantly higher than that in the NHE group and healthy group (P < 0.001). Measurement of plasma H 2 S revealed a lower level in HE patients than in the NHE and healthy ones [39.8 (10.0) of HE vs. 43.2 (15.9) pg/ml of NHE or 54.1 (21.4) pg/ml of the healthy, P < 0.05; Figure 1A] and a negative correlation between plasma H 2 S level and HE grade (r = −0.662; Figure 1B).

Bacteria Associated With H 2 S Production in HE
Blood microbial 16S rRNA gene detection was performed in 22 HE and 29 NHE patients. Three taxa with significant different abundances between the HE and NHE groups were found to be association with H 2 S production. Genus Staphylococcus and phylum Chloroflexi were increased in the blood of HE patients; conversely, an enrichment of species Pseudomonas alcaligenes was detected in the NHE group. Among the genes implicated in the production of H 2 S, genus Staphylococcus carries genes Sat and CysC, and phylum Chloroflexi is only with gene Sat, whereas Continuous data are presented as median (interquartile range) and categorical data as n (%). MELD, Model for End-Stage Liver Disease. species P. alcaligenes negatively associated with HE carries the most genes related to H 2 S production, such as CysH, CysD, CysN, CysC, and AprA (Figure 2).

H 2 S Attenuates NH 4 Cl-Induced Cytotoxicity in Primary Rat Astrocytes
The CCK-8 assay indicated that treatment with NH 4 Cl (2-10 mM) for 24 h significantly inhibited the proliferation of astrocytes in a concentration-dependent manner (P < 0.05; Figure 3A). Therefore, we used 5 mM NH 4 Cl to construct the HE models in follow-up studies. To assess the efficacy of H 2 S in attenuating the NH 4 Cl-induced cytotoxicity, astrocytes were treated with NH 4 Cl in the absence or presence of NaHS-an H 2 S donor. As shown in Figure 3B, the NH 4 Cl-induced decrease in cell viability was significantly attenuated by NaHS (200-800 µM; P < 0.05), and at 400 µM, NaHS exhibited the strongest effect. In addition, LDH in the cytoplasm is released if the cell is damaged. As shown in Figure 3C, pretreatment with NaHS (400 µM) significantly reduced NH 4 Cl-induced cell LDH release (P < 0.05).

H 2 S Prevents NH 4 Cl-Induced Oxidative Stress in Primary Astrocytes
We examined the effect of H 2 S on NH 4 Cl-induced OxS in astrocytes by measuring ROS levels. As shown in Figure 4, NH 4 Cl significantly increased intracellular ROS generation compared with the control group (P < 0.05). Notably, pretreatment with NaHS significantly inhibited the NH 4 Cl-induced increase in ROS levels (P < 0.05).

H 2 S Suppresses NH 4 Cl-Induced Apoptosis of Astrocytes
Double staining with Hoechst 33342 and PI was performed to investigate whether NaHS could mitigate astrocytic apoptosis induced by NH 4 Cl. Our result showed that NaHS significantly inhibited the apoptotic rate of astrocytes (P < 0.05; Figure 5A). To further observe the antiapoptotic effect of H 2 S, the expression of apoptosis-related protein was detected by Western blot . Activation of caspase proteases has been considered as an important mechanism in apoptosis with caspase-3 accounted as an essential executioner. Western blotting demonstrated that the expression of active caspase-3 fragment (17 kDa cleaved caspase-3) was up-regulated by NH 4 Cl, and H 2 S significantly suppressed the NH 4 Clinduced increase in cleaved caspase-3 level (P < 0.05; Figure 5B). Since pro-and anti-apoptotic members of the Bcl-2 family arbitrate the death or survival decision, the expressions of Bcl-2 (26 kDa) and Bax (21 kDa) were also examined. After treatment with NH 4 Cl, the expression of Bcl-2 decreased. In contract, the expression level of Bax increased. These observations were reversed by H 2 S (P < 0.05; Figure 5C). These results suggested that H 2 S suppressed FIGURE 2 | Analyses of correlation network among microbiota, H 2 S production, and HE. If the correlation is negative, the connecting line is black with dashes, whereas if positive, it is black. The yellow node represents hepatic encephalopathy, whereas the red one is the microbial community Pseudomonas. The big blue box shows the H 2 S production including assimilatory sulfate reduction pathway and dissimilatory sulfate reduction and oxidation pathway. The little boxes in green or orange represent genes, and the ones in orange are highly correlated with Pseudomonas. It clearly indicates a positive correlation between Pseudomonas and H 2 S production as well as a negative correlation between HE and H 2 S production.

Activation of Nrf2/ARE Signaling Pathway Mediates the Protective Effects of H 2 S Against NH 4 Cl-Induced Neurotoxicity
During OxS, Nrf2 enters the nucleus and binds to the ARE to initiate the antioxidant process. Western blotting analysis revealed a dramatical increase in the nuclear fraction of Nrf2 (110 kDa; P < 0.05) in the H 2 S + NH 4 Cl group, with a concomitant decrease in the cytoplasm (P < 0.05; Figure 6A). These results suggested that H 2 S could promote the translocation of Nrf2 from the cytosol to the nucleus. Similarly, immunofluorescence staining displayed that nuclear Nrf2 staining (green) was more abundant in the H 2 S + NH 4 Cl group than in any other groups where Nrf2 appeared to be mainly localized to the cytoplasm ( Figure 6B). Moreover, the most essential downstream target genes of Nrf2 are HO-1 and GCLC. The expression of HO-1 (33 kDa) and GCLC (73 kDa) was significantly elevated in NaHS-pretreated cells after being exposed to NH 4 Cl (P < 0.05; Figure 6C). This provides further evidence that H 2 S activates the Nrf2/ARE signaling pathway against NH 4 Cl-induced neurotoxicity.

Nrf2 Knockdown Blocks the Protective Effect of H 2 S in Primary Astrocytes
To further confirm whether Nrf2/ARE signaling was involved in the protective effect of H 2 S against ammonia toxicity to astrocytes, RNA inhibition was employed by shRNA1 and shRNA2 targeting Nrf2. As shown in Figure 7A, decreased expression of Nrf2 in astrocytes was observed at 27.54 ± 8.63 and 19.97 ± 6.15% relative to Nrf2 level in control after shRNA1 and shRNA2 treatments, respectively. Expressions of Nrf2 downstream genes HO-1 and GCLC were also down-regulated (P < 0.05; Figure 7A). Moreover, proliferation and apoptosis of astrocytes were examined. The proliferation of cells was inhibited, and they were more prone to apoptosis (albeit not statistically significant; Supplementary  Figure 2). Furthermore, inhibition of Nrf2 significantly abolished the protective effect of H 2 S on the NH 4 Clinduced decrease in cell proliferation ( Figure 7B) and so to the NH 4 Cl-induced elevation of LDH release ( Figure 7C). In addition, knockdown of Nrf2 expression blocked H 2 S protective effect on NH 4 Cl-induced apoptosis observed in both Hoechst 33342/PI staining and Western blotting (Figures 7D-F).

DISCUSSION
The results of the current study reveal the decrease of endogenous H 2 S and the effective protective role of H 2 S in HE astrocyte model. Our results suggest that: (i) there is a decrease in H 2 S production and bacteria associated with H 2 S production in the blood of HE patients; (ii) H 2 S attenuates NH 4 Cl-induced cytotoxicity, OxS, and apoptosis in primary rat astrocytes; and (iii) the Nrf2/ARE signaling pathway mediates the cytoprotection of H 2 S against NH 4 Cl-induced neurotoxicity.  Understanding of neurological disorder and H 2 S is emerging with more evidences suggesting that abnormal H 2 S generation can lead to neuronal dysfunction (Hu et al., 2010;Vandini et al., 2019). It is yet to be established whether HE characterized by neuropsychiatric abnormalities is associated with H 2 S synthesis. In the current study, the endogenous H 2 S levels were measured. Plasma H 2 S levels in HE patients were lower than those in cirrhotic patients without HE and healthy ones, and a negative correlation was found between the H 2 S level and HE grade. This suggests that the decline of H 2 S can promote neuronal dysfunction of HE disease. In addition, our results showed that H 2 S levels in the NHE group were lower than those in the healthy group, but statistically insignificant, indicating that the decrease of H 2 S is more closely associated with dysfunction of the brain than with the liver. Undoubtedly, a decrease in enzymes in the brain and liver is responsible for FIGURE 5 | H 2 S suppresses NH 4 Cl-induced apoptosis of astrocytes. Panel (A) shows the immunostaining of astrocyte with Hoechst 33342 and Propidium Iodide (PI). Chromatin condensation in apoptotic cell is indicated by the red arrow, and necrotic cell stained in red is indicated by the green arrow, respectively. Cell apoptosis ratio (%) = The quantity of positive cells/The quantity of total cells. Panel (B) indicates the expression of cleaved caspase-3 in astrocytes treated with control, H 2 S, NH 4 Cl, and combination of H 2 S and NH 4 Cl, respectively. Panel (C) displays the expression of Bcl-2 and Bax in astrocytes treated with control, H 2 S, NH 4 Cl, and combination of H 2 S and NH 4 Cl, respectively. Data are presented as mean ± SEM of three independent experiments. *P < 0.05. the reduction of H 2 S synthesis in HE. It is noteworthy that recent studies by Shen and colleagues found that the absence of microflora is connected with a significantly reduced CSE activity in many tissues coincident with an increase in tissue cysteine levels (Shen et al., 2013). These observations suggest an interesting hypothesis that bacteria could possibly influence enzyme activity or expression. A growing evidence indicated that a significant amount of H 2 S is produced by bacteria in the host. Systemic bioavailability and metabolism of H 2 S is also profoundly linked to bacteria (Rowan et al., 2009;Medani et al., 2011). Hence, HE in a patient with cirrhosis could be related to bacteria. In line of these findings, microbiota in the blood is mainly derived from gut microbiota and oral microbiota. Detection of blood microbial 16S rRNA gene, to our surprise, revealed that certain strains and functions of bacteria, especially P. alcaligenes, carry most genes related to the H 2 S generation but have negative correlation with HE. This suggested that certain strains associated with H 2 S production are reduced in HE, such as P. alcaligenes, and can up-regulate the H 2 S concentration by mediating the production of H 2 S. Thus, one of the main reasons for the decline of H 2 S in HE patients is the reduction of microbiota that can mediate the production of H 2 S. H 2 S exhibits therapeutic efficacy in many neurological disorders (Hu et al., 2010;Gheibi et al., 2014;Zhao et al., 2018). Our results demonstrated that H 2 S synthesis and the microbiota associated with H 2 S production are decreased, however the role of H 2 S in HE remains unclear. Study on the role of H 2 S in astrocytic model of HE demonstrated that H 2 S mitigates NH 4 Cl-induced cytotoxicity, OxS, and cell apoptosis in primary astrocytes, which indicates its neuroprotective effect against HE.
Ammonia intoxication impairs mitochondrial function and activates NADPH oxidase, leading to the formation of ROS (Bai et al., 2001;Rama Rao et al., 2003;Poznyak et al., 2020). In the current study, the level of intracellular ROS increased significantly with ammonia treatment, suggesting that ammonia induces OxS in astrocytes. Previous studies have established that H 2 S works as an endogenous scavenger for ROS under OxS. Detection of FIGURE 6 | Activation of Nrf2 signaling by H 2 S during NH 4 Cl-induced toxicity in astrocytes. Panel (A) shows the expression of Nrf2 in the cytoplasm (left) and nucleus (right) with histograms on the top and typical Western blot pictures at the bottom. Panel (B) displays the immunostaining of Nrf2 and DAPI as well as merge pictures in astrocytes treated with control, H 2 S, NH 4 Cl, and combination of H 2 S and NH 4 Cl, respectively. Panel (C) indicates the expression of Nrf2 downstream genes HO-1 and GCLC treated with control, H 2 S, NH 4 Cl, and combination of H 2 S and NH 4 Cl, respectively, with typical Western blot picture on the left and histograms on the right. Data are presented as mean ± SEM of three independent experiments. *P < 0.05, and # P > 0.05.
intracellular ROS levels showed that H 2 S ameliorates the disrupted redox state induced by NH 4 Cl, characterized by decreased ROS.
Apoptosis is a physiological process of cell death and plays a critical role in many biological systems. It has been considered as an important molecular basis for ammonia-FIGURE 7 | Effects of Nrf2 and H 2 S on astrocytes with and without NH 4 Cl treatment. Panel (A) shows that the antisense shRNA of Nrf2 attenuated the expression of Nrf2, HO-1, and GCLC in astrocytes with Western blot on the left and histogram on the right. Panel (B) shows the astrocyte viability, and panel (C) displays the lactate dehydrogenase (LDH) release in astrocytes treated with different combinations of control shRNA (nc-shRNA), antisense shRNA1 of Nrf2 (Nrf2-shRNA1), antisense shRNA2 of Nrf2 (Nrf2-shRNA2), H 2 S, and NH 4 Cl, respectively. Panel (D) indicates the cleaved caspase-3, and panel (E) shows the Bcl-2/Bax in astrocytes treated with different combinations of control shRNA (nc-shRNA), antisense shRNA1 of Nrf2 (Nrf2-shRNA1), antisense shRNA2 of Nrf2 (Nrf2-shRNA2), H 2 S, and NH 4 Cl, respectively, with Western blot at the bottom and histogram on the top. Panel (F) displays the immunostaining of Hoechst 33342 and PI in astrocytes under different cultures of control shRNA (nc-shRNA), antisense shRNA1 of Nrf2 (Nrf2-shRNA1), H 2 S, and NH 4 Cl, respectively, with staining picture on the top and histogram at the bottom. Data are presented as mean ± SEM of three independent experiments. *P < 0.05, and # P > 0.05.
induced cell death Zhang et al., 2018). Apoptosis induced by ammonia is regulated through a variety of signaling pathways, such as interruption of intracellular calcium ion (Ca 2+ ) homeostasis and activation of the p53 pathway . Our results displayed that apoptosis of astrocytes was increased sharply with ammonia treatment, and that this was significantly attenuated by H 2 S. Recent studies on apoptosis popularly focused on OxS, and it has been reported that ROS can trigger apoptotic pathways (Chen et al., 2006;Zhang et al., 2018). In the present study, it displayed a consistent trend between apoptosis and ROS production, suggesting that apoptosis of astrocytes may be closely correlated with the overproduction of ROS. Based on these findings, we can conclude that H 2 S exerts neuroprotective effects by relieving astrocytic toxicity, OxS, and apoptosis against ammonia-induced HE.
H 2 S itself is a reductant that can neutralize free radicals. Recent studies have confirmed that the Nrf2/ARE signaling pathway plays a pivotal role in the antioxidative effect of H 2 S (Calvert et al., 2009;Yang et al., 2013;Kimura, 2014). During OxS, H 2 S can induce S-sulfhydration of Keap1, which contributes to Nrf2 dissociation from Keap1 and migration of Nrf2 into the nuclei to up-regulate the transcription of antioxidant genes (Calvert et al., 2009;. Herein, results from both immunofluorescence and Western blotting analysis attested for the first time that pretreatment with NaHS contributes to nuclear translocation of Nrf2 and improves the expression of its downstream genes, such as GCLC and HO-1, against NH 4 Cl-induced neurotoxicity. When NH 4 Cl induces excessive ROS, H 2 S is able to prevent this oxidative damage by promoting nuclear translocation of Nrf2, which up-regulates the expression of a series of antioxidant enzymes, such as HO-1, to enhance detoxification and attenuate OxS. Our results showed that H 2 S does not increase nuclear Nrf2 expression by itself. This suggested that OxS is a prerequisite for activation of Nrf2/ARE signaling by H 2 S. In addition, knockdown of astrocytic Nrf2 quelled the protective effect of H 2 S in NH 4 Clinduced cytotoxicity and cell apoptosis, which indicated that the Nrf2/ARE signaling pathway mediates the neuroprotection of H 2 S against HE. Although the current study demonstrates a role of H 2 S in NH 4 Cl-induced HE astrocyte model and provides a new mechanistic explanation for the potential therapeutic value of H 2 S in the treatment of HE, there are still some limitations. The first limitation is the group division in patient study. Although the HE and NHE groups were divided based on clinical symptoms of brain dysfunction, patients with minimal HE have few recognizable clinical symptoms of brain dysfunction, which could lead to the possibility of patients with HE in the NHE group. The possibility could be minimized by excluding those currently or previously diagnosed with or suspected of HE in the NHE group. The second limitation is that patients in coma (including HE grade 4) were excluded since most coma patients experience hemodynamic instability or multi-organ dysfunction that could not reflect their ''real'' states, especially H 2 S levels. The third limitation is that the study could be better to have H 2 S level in cerebrospinal fluid (CSF); however, it is impossible to have CSF in patients with HE. Since there are evidences suggesting consistent H 2 S level between CSF and blood (Eto et al., 2002), the findings in the current investigation are still reliable. In addition, in the study of cell viability, NaHS showed the stronger effect at 400 µM than at 800 µM, indicating that high concentrations of NaHS might be cytotoxic and H 2 S exerts its neuroprotective effect within a proper range of concentrations.
In conclusion, the current investigation demonstrates that H 2 S acts as neuroprotection against NH 4 Cl-induced HE model by activating Nrf2/ARE signaling of astrocytes, and this contributes to the prevention of HE with the possible activation of the antioxidative defense system in astrocytes.

DATA AVAILABILITY STATEMENT
The sequence data have been deposited in the National Center for Biotechnology Information (NCBI) BioProject database with project number PRJNA640495 (https://www.ncbi.nlm.nih.gov/ Traces/study/?acc=PRJNA640495).

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Human Research Ethics Committee, the First Affiliated Hospital, Wenzhou Medical University, China. The patients/participants provided their written informed consent to participate in this study. The animal study was reviewed and approved by the Animal Ethics Committee, Wenzhou Medical University, China.

AUTHOR CONTRIBUTIONS
LX and YC designed the protocol. XJ and DC performed the experiments and edited the manuscript. FW and LZ directed and participated in the sample detection. YH, ZL, RW, and XW analyzed the data. XJ wrote the manuscript, which was also edited by DC. All authors contributed to the article and approved the submitted version.

ACKNOWLEDGMENTS
We heartily appreciate all those who contributed to this research.
SUPPLEMENTARY FIGURE 1 | GFAP immunostaining of primary astrocytes (shown in green).

SUPPLEMENTARY FIGURE 2 | Effects of Nrf2 on astrocytes. Panels (A,B)
show the antisense shRNA of Nrf2 regulated astrocyte viability and LDH release. Panel (C) displays the immunostaining of Hoechst 33342 and PI in astrocytes of control, control shRNA (nc-shRNA), and antisense shRNA1 of Nrf2 (Nrf2-shRNA1) groups, respectively, with staining picture on the top and histogram at the bottom. Data are presented as mean ± SEM of three independent experiments. Panel (D) shows that the antisense shRNA of Nrf2 regulated the expression of cleaved caspase-3 and Bcl-2/Bax, respectively, with Western blot at the bottom and histograms on the top, respectively. # P > 0.05.