Hydrogen sulfide stimulates activation of hepatic stellate cells through increased cellular bio-energetics

Hepatic fibrosis is caused by chronic inflammation and characterized as the excessive accumulation of extracellular matrix (ECM) by activated hepatic stellate cells (HSCs). Gasotransmitters like NO and CO are known to modulate inflammation and fibrosis, however, little is known about the role of the gasotransmitter hydrogen sulfide (H2S) in liver fibrogenesis and stellate cell activation. Endogenous H2S is produced by the enzymes cystathionine β-synthase (CBS), cystathionine γ-lyase (CTH) and 3-mercaptopyruvate sulfur transferase (MPST) [1]. The aim of this study was to elucidate the role of endogenously produced and/or exogenously administered H2S on rat hepatic stellate cell activation and fibrogenesis. Primary rat HSCs were culture-activated for 7 days and treated with different H2S releasing donors (slow releasing donor GYY4137, fast releasing donor NaHS) or inhibitors of the H2S producing enzymes CTH and CBS (DL-PAG, AOAA). The main message of our study is that mRNA and protein expression level of H2S synthesizing enzymes are low in HSCs compared to hepatocytes and Kupffer cells. However, H2S promotes hepatic stellate cell activation. This conclusion is based on the fact that production of H2S and mRNA and protein expression of its producing enzyme CTH are increased during hepatic stellate cell activation. Furthermore, exogenous H2S increased HSC proliferation while inhibitors of endogenous H2S production reduce proliferation and fibrotic makers of HSCs. The effect of H2S on stellate cell activation correlated with increased cellular bioenergetics. Our results indicate that the H2S generation in hepatic stellate cells is a target for anti-fibrotic intervention and that systemic interventions with H2S should take into account cell-specific effects of H2S.


Introduction
Chronic inflammation occurs in many liver diseases, e.g. non-alcoholic steatohepatitis (NASH), viral infection or chronic alcohol consumption. Liver fibrosis can be viewed as an uncontrolled wound healing response. Hepatic stellate cells (HSCs) play an important role in the onset and perpetuation of liver fibrosis. Under normal conditions, HSCs are quiescent and are the principal vitamin A storing cells in the liver [1]. In conditions of chronic inflammatory liver injury, quiescent hepatic stellate cells (qHSCs) transform into proliferative myofibroblast-like cells called activated HSCs (aHSCs). During activation, HSCs lose their vitamin A content and start to produce large amounts of extracellular matrix (ECM) [2]. When the inflammatory response is not suppressed, the excessive accumulation of ECM can lead to hepatic fibrosis, cirrhosis and eventually hepatocellular carcinoma. At present, there is no effective treatment for hepatic fibrosis, leaving liver transplantation as the only viable treatment option. Therefore, it is important to understand the mechanisms that lead to hepatic stellate cell activation and hepatic fibrosis [3,4]. Gasotransmitters like nitric oxide (NO) and carbon monoxide (CO) have been shown to play an important role in chronic liver inflammation and liver fibrosis [5,6]. Recently, interest has been focused on another gasotransmitter, hydrogen sulfide (H 2 S) [7][8][9].
In the last two decades, H 2 S has been identified as a gasotransmitter that is generated in many mammalian cells and is involved in various physiological and pathophysiological processes as a signaling molecule similar to NO and CO [10]. H 2 S has also been implicated to modulate inflammation and fibrosis, although its role in liver fibrosis and hepatic stellate cell activation is still not completely elucidated. enzymes cystathionine β-synthase (CBS), cystathionine γ-lyase (CTH) and 3-mercaptopyruvate sulfur transferase (MPST) [11,12]. It has been shown to regulate hepatic fibrosis via its anti-oxidative and anti-inflammatory properties and by inducing cell-cycle arrest, apoptosis, vasodilation and reduction of portal hypertension [8,9,[13][14][15][16]. However, most of these experiments were performed in vivo conditions and did not focus directly on the process of fibrogenesis and HSCs activation. Furthermore, conflicting results have been reported depending on the concentration or type of H 2 S donor used. Based on the H 2 S release rate, H 2 S releasing donors can be categorized as fast (NaHS; Na 2 S) or slow (GYY4137; ADT-OH) releasing donors, often yielding contrasting results [17][18][19]. For instance, some studies reported pro-inflammatory and anti-apoptotic properties of H 2 S and in some studies H 2 S was shown to increase mitochondrial bioenergetics and promote cell proliferation [20][21][22][23]. Therefore, there are still major gaps in our understanding of the actual effects of H 2 S on HSCs and liver fibrosis.
The aim of the current study was to elucidate the effects of H 2 S on HSCs by investigating how endogenously produced and/or exogenously administered H 2 S affects primary rat HSCs and its proliferation. Furthermore, we tried to elucidate the dynamics of endogenous production of H 2 S and H 2 S synthesizing enzymes during HSCs activation.

Hepatic stellate cell isolation and culture
Specified pathogen-free male Wistar rats were purchased from Charles River (Wilmington, MA, USA) and housed in a 12hr light-dark cycle under standard animal housing conditions with free access to chow and water. HSCs were isolated from rats weighing 350-450 g, anesthetized by isoflurane and a mixture of Ketamine and Medetomidine. The liver was perfused via the portal vein with a buffer containing Pronase-E (Merck, Amsterdam, the Netherlands) and Collagenase-P (Roche, Almere, the Netherlands). The HSC population was isolated by density centrifugation using 13% Nycodenz (Axis-Shield POC, Oslo, Norway) solution. Isolated HSCs were cultured in Iscove's Modified Dulbecco's Medium supplemented with Glutamax (Thermo Fisher Scientific, Waltham, MA, USA), 20% heat inactivated fetal calf serum (Thermo Fisher Scientific), 1% MEM Non Essential Amino Acids (Thermo Fisher Scientific), 1% Sodium Pyruvate (Thermo Fisher Scientific, Waltham, MA, USA) and antibiotics: 50 μg/mL gentamycin (Thermo Fisher Scientific), 100 U/mL Penicillin (Lonza, Vervier, Belgium), 10 μg/mL streptomycin (Lonza) and 250 ng/mL Fungizone (Lonza) in an incubator containing 5% CO 2 at a 37°C [24]. Quiescent HSCs (day 1) spontaneously activate when cultured on tissue culture plastic and reached complete activation (increased proliferation, loss of retinoids and increased synthesis of extracellular matrix components) after 7 days of culture. Day 3 cultured HSCs are considered intermediately activated.

Experimental design
Culture-activated HSCs (aHSCs) were treated with H 2 S donors or inhibitors for 72 h. All treatments with H 2 S donors and inhibitors were performed in fresh medium containing 20% FCS and other supplements. H 2 S releasing donors GYY4137 (kind gift from Prof. Matt Whiteman, University of Exeter, United Kingdom) and NaHS (Sigma-Aldrich, Zwijndrecht, the Netherlands) were diluted in distilled water and prepared freshly. NaHS was added every 8 h to the cells because of its rapid evaporation. The CBS inhibitor O-(carboxymethyl) hydroxylamine, AOAA (Sigma-Aldrich) was prepared as a 200 mmol/L stock solution and diluted in distilled water at neutral pH. CTH inhibitor DLpropargylglycine, DL-PAG (Sigma-Aldrich) was freshly prepared.

Measurement of H 2 S concentration
The accumulation of H 2 S in the culture medium was measured as described previously [25,26]. After 72 h incubation, medium samples were collected in 250 μL of 1% (wt/vol) zinc acetate and distilled water was added up to 500 μL. Next, 133 μl of 20 mmol/L N-dimethyl-p-  [26][27][28][29][30][31][32][33] phenylenediamine sulfate in 7.2 mmol/L hydrogen chloride and 133 μl 30 mmol/L ferric chloride in 1.2 mmol/L hydrogen chloride were added. After incubation for 10 min at room temperature, protein was removed by adding 250 μL trichloroacetic acid and centrifugation at 14000g for 5 min. Spectrophotometry was performed at 670 nm light absorbance (BioTek Epoch2 microplate reader) in 96 well-plates. All samples were assayed in duplicated. Concentrations were calculated against a calibration curve of NaHS (5-400 μmol/L) in culture medium.

Quantitative real-time polymerase chain reaction
Hepatic stellate cell RNA was isolated using Tri-reagent (Sigma-Aldrich) according to the manufacturer's protocol. RNA concentrations were measured by Nano-Drop 2000c (Thermo Fisher Scientific, Waltham, MA, USA) and 1.5 μg of RNA was used for reverse transcription (Sigma-Aldrich). cDNA was diluted in RNAse-free water and used for real-time polymerase chain reaction on the QuantStudio™ 3 system (Thermo Fisher Scientific). All samples were analyzed in duplicate using 18S and 36b4 as housekeeping genes. The mRNA levels of Cth, Cbs, Mpst (Invitrogen) were quantified using SYBR Green (Applied Biosystems), other genes were quantified by TaqMan probes and primers. Relative gene expression was calculated via the 2 -?Ct method. The primers and probes are shown in Table 1.

Cell toxicity determination by Sytox Green
Cell necrosis was measured by Sytox Green nucleic acid staining (Invitrogen, the Netherlands) at a dilution of 1:40.000 in culture medium or HBSS for 15 min at a 37°C. Necrotic cells have ruptured plasma membranes, allowing entrance of non-permeable Sytox green into the cells. Sytox green then binds to nucleic acids. Fluorescent nuclei were visualized at an excitation wavelength of 450-490 nm by a Leica microscope. Hydrogen peroxide 1 mmol/L was used as a positive control.

Cell proliferation measurement
Proliferation of aHSCs was measured by Real-Time xCELLigence system (RTCA DP; ACEA Biosciences, Inc., CA, USA) and by colorimetric BrdU cell proliferation ELISA kit (Roche Diagnostic Almere, the Netherlands). Cells were seeded in a 16-well E-plate and treated as indicated. Cell index was determined by measuring the change of impedance on the xCELLigence system.
For BrdU incorporation assay, aHSCs were seeded in a 96-well plate and treated as indicated. BrdU incorporation was determined according to manufacturer's instructions and quantified by light emission chemiluminiscence using the Synergy-4 machine (BioTek).

Bile duct ligation
Male Wistar rats were anaesthetized with halothane/O 2 /N 2 O and subjected to bile duct ligation (BDL) as described by Kountouras J et al. [27]. At the indicated times after bile duct ligation (BDL), the rats (n = 4 per group) were sacrificed, livers were perfused with saline and removed. Control rats received a sham operation (SHAM). Specimens of these livers were snap-frozen in liquid nitrogen for isolation of mRNA and protein.

Statistical analysis
Results are presented as mean ± standard deviation (mean ± SD). Every experiment was repeated at least 3 times. Statistical significance was analyzed by Mann-Whitney test between the two groups and Kruskal-Wallis followed by post-hoc Dunn's test for multiple comparison test. Statistical analysis was performed with GraphPad Prism 7 (GraphPad Software, San Diego, CA, USA).

Hydrogen sulfide production is increased upon activation of hepatic stellate cells
In order to determine the dynamics of H 2 S production and H 2 S producing enzymes during HSC activation, mRNA expression of H 2 S synthesizing enzymes was measured in quiescent (q) and activated (a) HSCs and compared to the expression of these enzymes in hepatocytes and Kupffer cells. As shown in Fig. 1, the H 2 S producing enzymes Cth, Cbs and Mpst were expressed at low levels in qHSCs compared to hepatocytes and Kupffer cells. Upon activation, Cth gene expression increased in HSCs (Fig. 1A) while Cbs and Mpst mRNA levels were not changed ( Fig. 1B and C). In line with this, the accumulation of H 2 S in culture medium was increased during HSC activation. In Fig. 1D, values are normalized for cell number since the morphology and proliferation rate of quiescent and activated stellate cells are very different (Fig. 1D). Western blotting results showed similar trend as observed for the mRNA expression data. Protein expression of CTH was increased during HSCs activation (Fig. 1E).
Cth, Cbs and Mpst mRNA expression was determined in HSCs at day 1, 3, and 7 and compared to primary rat hepatocytes and Kupffer cells (A-C). The cytosolic enzymes Cth and Cbs were abundantly expressed in hepatocytes, while their expression was relatively low in HSCs. Upon HSCs activation, Cth expression was induced 7-fold, and Mpst slightly upregulated, whereas expression of Cbs was downregulated. Expression levels are relative to 18S expression. D. Production of H 2 S in activated and quiescent HSCs. The production of H 2 S was increased upon activation of HSCs. Results were normalized with respect to the number of cells. E. Protein expressions of CTH, CBS, MPST of HSCs at different time point and hepatocytes and Kupffer cells. Equal protein loading was confirmed by Ponceau S staining and Western blot for GAPDH.

Effect of H 2 S on activation markers in hepatic stellate cells
In order to avoid confounding effects of cell toxicity, we optimized the concentration of H 2 S donors and inhibitors by Sytox green staining. At concentrations twice as high as used in the experiments, none of the donors or inhibitors were toxic to HSCs (Fig. 2).
Toxicity of the compounds was checked by Sytox Green staining. Hydrogen peroxide (1 mmol/L; 6 h exposure) was used as a positive control. The compounds DL-PAG (Cth inhibitor), AOAA (Cbs inhibitor), GYY4137 (slow releasing donor) and NaHS (fast releasing donor) were not toxic for HSCs. Duration of the treatment was 24hrs.
We next evaluated the effect of H 2 S on activation markers in aHSCs. Inhibitors of H 2 S producing enzymes (DL-PAG, AOAA) decreased the expression of the fibrogenic markers Col1α1 and Acta2 (Fig. 3A). The H 2 S donors GYY4137 and NaHS did not affect the expression of Col1α1. However, GYY4137 slightly, but significantly, reduced Acta2 mRNA expression (Fig. 3B). Interestingly, both of the two enzyme inhibitors also downregulated the expression of Cth mRNA. The changes in mRNA expression were reflected in similar changes in protein expression of COL1α1 but not ACTA2 (Fig. 3C). Accumulation of H 2 S in culture medium was reduced by inhibitors, whereas GYY4137 increased H 2 S accumulation. Because of the fast release, no accumulation of H 2 S was measured in NaHS-treated group. In Fig. 3D, we did not normalize values to the number of cells (in contrast to Fig. 1), because experiments were performed with only activated stellate cells over a limited time span, in which it can be assumed that cell numbers will not differ significantly (Fig. 3D). The H 2 S synthesizing enzyme inhibitors DL-PAG and AOAA downregulated Col1α1, Acta2 and Cth mRNA expression while the H 2 S donors GYY4137 and NaHS did not affect Cth and Col1α1 mRNA expression (A, B). In contrast, GYY4137, but not NaHS reduced Acta2 mRNA expression slightly. 18S was used as a housekeeping gene. The inhibitors also reduced COL1α1 protein level but not ACTA2 protein level (C). GAPDH was used as loading control for protein analysis. The accumulation over 72 h of H 2 S in culture medium was measured in the experimental groups (D). DL-PAG and AOAA significantly reduced the accumulation of H 2 S. Because of its fast release, no accumulation of H 2 S was measured in the NaHS-treated group. Accumulation of H 2 S was detected with the slow releasing donor GYY4137.

H 2 S promotes hepatic stellate cell proliferation
The effect of H 2 S on rat HSC proliferation was assessed using realtime cell analyzing xCelligence and BrdU incorporation ELISA assays. H 2 S donors promote, whereas H 2 S synthesizing enzyme inhibitors inhibit aHSCs proliferation, indicating a stimulatory effect of H 2 S on HSC  proliferation (Fig. 4). Culture-activated HSCs were treated with H 2 S donors and enzyme inhibitors over period of 72 h. Cell proliferation was monitored by realtime xCELLigence system (A) and confirmed with BrdU incorporation ELISA assay (B). Inhibition of endogenous production of H 2 S suppressed cell proliferation, whereas H 2 S donors increased aHSCs proliferation. Data are presented ± SD.

H 2 S increases cell metabolic activity
H 2 S at low concentrations can increase cellular bioenergetics as an electron donor in mitochondrial oxidative phosphorylation [20,28]. Since enhanced bioenergetics is associated with HSC activation, we investigated the effect of H 2 S on the bioenergetics of aHSCs. Two parameters of cellular metabolic activity, oxygen consumption rate (OCR) for mitochondrial oxidative phosphorylation and extracellular acidification rate (ECAR) for glycolysis, were determined using the Seahorse Extracellular Flux analyzer (Fig. 5). The H 2 S donors GYY4137 and NaHS increased both the OCR and ECAR and ATP production, whereas the enzyme inhibitors DL-PAG and AOAA decreased metabolic activity of HSCs and ATP production.
Effect of H 2 S donors and enzyme inhibitors on bioenergetics of aHSCs. Treatments with donors and inhibitors was for 48hrs. OCR and ECAR are represented as mean ± SEM of a representative experiment (A, B). Results were normalized with respect to the total amount of protein. Fold change of normalized maximal and basal level of OCR and ECAR between conditions were analyzed in 3 different experiments. For each experiment, every condition was repeated at least two times (C, D). Production of ATP was calculated using Seahorse XF Cell Mito Stress Test Report Generator software. Fold change of ATP production in experimental groups was calculated in 3 independent experiments (E).

Cth is specifically induced in hepatic stellate cells during fibrogenesis
We next evaluated the expression of H 2 S synthesizing enzymes in the bile duct ligation model, an experimental model of chronic inflammation leading to fibrosis [29]. mRNA levels of all H 2 S synthesizing enzymes, Cth, Cbs and Mpst, decreased progressively in the bile duct ligation model (Fig. 6A-C). As expected, expression of the profibrogenic cytokine TGFβ1 increased progressively in the bile duct ligation model (Fig. 6D). We next evaluated the effect of TGFβ1 on the mRNA expression of H 2 S synthesizing enzymes in different liver cell populations. TGFβ1 decreased mRNA expression of all H 2 S synthesizing enzymes in hepatocytes. In contrast, TGFβ1 increased mRNA expression of Cth in HSCs and did not change the mRNA expression of Cbs and Mpst in HSCs (Fig. 6E-G).
Comparison of H 2 S synthesizing enzymes mRNA levels during fibrosis in vivo and in vitro. Cth, Cbs, Mpst were downregulated in total liver in the BDL model of liver fibrosis (A,B,C). Tgfβ1 expression is increased in fibrosis (D). 36b4 was used as a housekeeping gene. TGFβ1 reduced the expression of H 2 S synthesizing enzymes in hepatocytes (F,G), but it specifically induced Cth mRNA expression in HSCs in vitro (E).

Discussion
The main message of our study is that H 2 S promotes hepatic stellate cell activation. This conclusion is based on the fact that production of H 2 S and expression its producing enzyme cystathionine γ-lyase (Cth) expression are increased during hepatic stellate cell activation and on the fact that exogenous H 2 S increased HSC proliferation while inhibitors of endogenous H 2 S production reduce proliferation of HSCs. Although the inhibitors we used are not completely specific for one of the H 2 S producing enzymes, e.g. the CBS inhibitor AOAA is also a potent inhibitor of CTH [30] it is important to note that reducing H 2 S production leads to reduced stellate cell activation. In addition, since  [26][27][28][29][30][31][32][33] CTH is the sole enzyme to upregulated during HSCs activation, it is likely that the effect of AOAA is mediated via inhibition of CTH. The effect of H 2 S on stellate cell activation correlated with increased cellular bioenergetics. Previous in vivo studies reported that H 2 S has anti-fibrotic properties due to its antioxidant and/or anti-inflammatory actions and its ability to reduce portal hypertension in the liver. In models of (experimental) fibrosis and cirrhosis, reduced expression of H 2 S producing enzymes are observed and an anti-fibrotic effect as well as reduction of portal hypertension of systemically administered H 2 S donors has been reported [7,[13][14][15]. In line with this, in vitro studies, using the fast-releasing H 2 S donor NaHS have demonstrated that H 2 S inhibits stellate cell proliferation, possibly via decreasing the phosphorylation of p38 MAP-Kinase and increasing the phosphorylation of Akt [9,15]. In another study, the natural H 2 S donor diallyl trisulfide suppressed activation of HSCs through cell cycle arrest at the G2/M checkpoint associated with downregulation of cyclin B1 and cyclindependent kinase 1 in primary rat HSCs [8]. However, the results described above were obtained using potentially toxic, fast-releasing H 2 S donors, which is not representative of the continuous production of low levels of H 2 S by cells. Furthermore, the use of systemically administered donors or inhibitors does not allow to distinguish effects of H 2 S on different cell types present within one organ. Therefore, we applied 2 different H 2 S releasing donors, GYY4137 and NaHS at concentrations 5 times as lower as in some in vitro studies. Furthermore, most studies used exogenous H 2 S donors to study the role of H 2 S in stellate cell activation and fibrogenesis and the importance of endogenous production of H 2 S in HSC activation has not been properly addressed. Therefore, we also used 2 inhibitors of H 2 S synthesizing enzymes (DL-PAG and AOAA) and we determined H 2 S production by HSCs during the process of activation [8,9].
Our observations of increased expression of H 2 S synthesizing enzyme CTH and increased H 2 S production during HSC activation indicates a role for H 2 S in HSC activation and fibrogenesis. Indeed, inhibition of endogenous H 2 S production in HSCs reduced proliferation and expression of activation markers. These results are in line with the observation that platelet-derived growth factor BB (PDGF-BB) induced proliferation of rat mesangial cells via induction of CTH [31] and the observation that homocysteine, a precursor in H 2 S synthesis, enhances activation of rat HSCs via activation of the PI3K/Akt pathway [32]. In contrast, an anti-fibrotic role has been proposed for cystathionine-βsynthase (CBS), another PLP-dependent enzyme which is involved in H 2 S synthesis in the liver [33,34].
Since our results demonstrated a pro-fibrogenic effect of H 2 S on HSCs, whereas most in vivo studies reported an anti-fibrotic role for H 2 S, we investigated in more detail the H 2 S generating capacity in different liver cell types. First, we determined that expression of H 2 Ssynthesizing enzymes in hepatocytes and Kupffer cells is much higher than in HSCs. Next, we determined the expression of H 2 S-synthesizing enzymes in the bile duct ligation model of liver fibrosis. We observed a down-regulation of total hepatic expression of both Cth and Cbs in our bile duct ligation model. As expected, the pro-fibrogenic cytokine Tgfβ1 was increased in the bile duct ligation model. Finally, we studied the effect of TGFβ1 on the expression of H 2 S-synthesizing enzymes in different liver cell types. Of note, we observed that TGFβ1 decreases Cth and Cbs mRNA expression in hepatocytes, but increased Cth mRNA expression in stellate cells. These findings could explain the contradictory results between in vivo and in vitro studies with regard to the role of H 2 S in fibrogenesis: since hepatocytes are the major source of H 2 S in total liver, the increased expression of Tgfβ1 will lead to an overall reduction in the hepatic expression of Cth and Cbs and H 2 S production, whereas at the same time it will increase expression of Cth and H 2 S production in hepatic stellate cells. The cell-specific and local increase in H 2 S generation also explains the effect of H 2 S donors and inhibitors of H 2 S-synthesizing enzymes on HSC proliferation and activation. Recently, Szabo et al. reported that a low exogenous dose of H 2 S or endogenously produced H 2 S increases mitochondrial oxidative phosphorylation [35,36]. In accordance, Katalin et al. described that low concentrations of H 2 S stimulates mitochondrial bio-energetics via S-sulfhydration of ATP-synthase in HepG2 and HEK293 cell lines [20]. Activation of stellate cells is also accompanied by increased bioenergetics [37]. We have extended these findings by demonstrating that H 2 S increases cellular bioenergetics in hepatic stellate cells.  [26][27][28][29][30][31][32][33] In summary, we demonstrate that stellate cell activation is accompanied by increased generation of H 2 S via induction of the H 2 S-synthesizing enzyme CTH, leading to increased cellular bioenergetics and proliferation of HSCs. In addition, the response of H 2 S-synthesizing enzymes to the fibrogenic cytokine Tgfβ1 is liver cell-type specific. Our results indicate that the H 2 S generation in hepatic stellate cells is a target for anti-fibrotic intervention and that systemic interventions with H 2 S should take into account cell-specific responses to H 2 S.