Initially, the crystal structure of CSE protein [Protein Data Bank (PDB) ID: 3COG] was downloaded from the PDB (http://www.rcsb.org/pdb). The binding pocket was identified by the Site Finder module with empirical confirmation. The target library (the SPECS compound library) was prepared with the Wash module in Molecular Operating Environment (MOE) 2014.09 (Chemical Computing Group ULC, Montreal, QC, Canada). High throughput rigid docking, followed by flexible docking with force field refinement, was used to rank the compounds, in order to achieve a balance between accuracy and efficiency.

MOE Dock was used for docking simulations of the ligands and predicting their binding affinity with the homology model. The 2D structures of the ligands were drawn in ChemOffice (2013; PerkinElmer, Inc., Waltham, MA, USA) and converted to 3D in MOE through energy minimization. The flexible docking protocol followed the ‘induced fit’ protocol, in which the side chains of the receptor pocket were allowed to move according to ligand conformations, with a constraint on their positions. The weight used for tethering side chain atoms to their initial positions was 10. Prior to docking, the force field of AMBER12:EHT and the implicit solvation model of Reaction Field (R-field) were selected. The protonation state of the protein and the orientation of the hydrogens were optimized by LigX, at a pH of 7.0 and temperature of 300 K. For rigid docking, the docked poses of the ligands were ranked by London dG scoring. For flexible docking, a force field refinement was conducted on the top 30 poses followed by rescoring of GBVI/WSA dG. MOE Dock, AMBER12:EHT, R-field and GBVI/WSA dG are programs belonging to the MOE 2014.09 software package.

I157172, also known as 2-[(4-(2,5-dimethoxyanilino)-6-{3-nitroanilino}-1,3,5-triazin-2-yl) sulfanyl]-6-ethoxy-1,3-benzothiazole, was purchased from Specs (Zoetermeer, The Netherlands), L-propargylglycine (PAG) was purchased from Sigma-Aldrich; Merck KGaA (Darmstadt, Germany) and doxorubicin (DOX) was obtained from Beijing Solarbio Science & Technology Co., Ltd. (Beijing, China). PAG is an existing CSE inhibitor and DOX is a common clinical antitumor drug; therefore, PAG and DOX were used as positive controls.

A total of 16 breast cancer tumor samples and adjacent non-tumor samples were collected from patients. All of the patients (age, 37–60 years; no smoking history available) had a non-specific type of invasive breast cancer (4 cases of grade II and 12 cases of grade III breast cancer) and were recruited from the Huaihe Hospital (Kaifeng, China). The present study was approved by the ethics committee of the Medical School, Henan University (Kaifeng, China), and patients provided written informed consent. MCF7 and MDA-MB-231 human breast cancer cell lines, and the Hs578Bst mammary epithelial cell line were obtained from the American Type Culture Collection (Manassas, VA, USA). The cells were cultured in Dulbecco's modified Eagle's medium (Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) supplemented with 10% fetal bovine serum (FBS; Zeta Life, Inc., San Francisco, CA, USA) at 37°C in an atmosphere containing 5% CO₂.

H₂S determination was performed using the methylene blue method. Briefly, MCF7 cells (1×10⁶/ml) were seeded into a 6-well tissue culture plate. After 24 h, the cells were treated with 20 µM I157172, alongside 2 mM L-cysteine and 0.5 mM pyridoxal phosphate for 24 h at 37°C. Concurrently, 1% (w/v) zinc acetate (500 µl) was added to filter papers adhered to the lid of the 6-well tissue culture plate for 24 h, in order to absorb H₂S. Then the filter papers were placed into tubes containing 0.2% (w/v) N, N-dimethyl-p-phenylenediamine-dihydrochloride dye (500 µl), 10% (w/v) ammonium ferric sulfate (50 µl) and 3 ml H₂O, and were incubated for 20 min at room temperature. Absorbance at 670 nm was subsequently measured. The production of H₂S was determined using a standard curve of NaHS (0–1 mM; R²=0.9998) and was presented as nmol·min⁻¹ per 1×10⁶ cells. The assay was conducted in triplicate for three independent experiments.

MTS and 5-ethynyl-2′-deoxyuridine (EdU) assays were used to assess cell viability and proliferation, respectively. MCF7 and Hs578Bst cells were plated into 96-well plates at a density of 1×10⁶/ml, and were treated with 0, 2.5, 5, 10, 20, 30 and 40 µM I157172 and 0, 5, 10, 20 and 40 µM PAG or DOX for 24 or 48 h at 37°C. Cell viability was evaluated by determining the number of cells using the MTS assay (Sigma-Aldrich; Merck KGaA), according to the manufacturer's protocol, and images were captured with an inverted microscope (IX53, Olympus Corporation, Tokyo, Japan). The assay was conducted in triplicate for three independent experiments. PAG and DOX served as positive controls.

Cell proliferation was assessed using the EdU assay kit (Guangzhou Ribobio Co., Ltd., Guangzhou, China). The results of a pre-experiment revealed that treatment of cells with I157172 for >24 h cannot accurately reflect cell proliferation. Therefore, in this study, cells were exposed to I157172 for 24 h for the cell proliferation assay. Briefly, MCF7 cells (1×10⁶/ml) were cultured in 96-well plates and were exposed to 0, 10, 20 and 30 µM I157172 for 24 h at 37°C. Subsequently, cells were incubated with 50 µm EdU for 2 h at room temperature, fixed with 4% formaldehyde for 30 min at room temperature, incubated with glycine (2 mg/ml) for 5 min and treated with 0.5% Triton X-100 for 10 min to permeabilize the cells. After being washed with PBS for 5 min, the cells were incubated with 1X Apollo^® 567 (Guangzhou Ribobio Co., Ltd.) for 30 min and treated twice with 0.5% Triton X-100. DNA was stained with Hoechst 33342 for 30 min and visualized with fluorescence microscopy. Five groups of cells in the images were randomly selected.

The scratch wound assay was used to determine cell migration. MCF7 cells were seeded into a 6-well plate and were scraped with a 10-µl pipette tip once they reached ~90% confluence to generate a wound, and were rinsed twice with PBS. Subsequently, the cells were treated with 0, 10, 20 and 30 µM I157172 and were cultured in medium containing 5% FBS for 24 h at 37°C. The distance of wound closure was measured at the beginning of the experiment and after 24 h using an inverted microscope (IX53; Olympus Corporation). The assay was conducted in triplicate for three independent experiments.

A Transwell assay was conducted to analyze invasion. Briefly, 24-well Transwell chambers (pore size, 8 µm; Corning Incorporated, Corning, NY, USA) were coated with Matrigel matrix (BD Biosciences, San Joes, CA, USA) for 30 min at 37°C. MCF7 cells (5×10⁴) suspended in 200 µl DMEM with 1% FBS were seeded into the top chambers, whereas DMEM with 15% FBS was placed into the bottom chambers. MCF7 cells in the top chambers were treated with 0, 10, 20 and 30 µM I157172 for 24 h at 37°C, and the cells on the upper surface of the membrane were removed. The invasive cells attaching to the lower surface of membrane were fixed with 4% polyoxymethylene for 30 min and were stained with 0.1% crystal violet for 15–20 min at room temperature. Images of the stained cells were captured with an inverted microscope (IX53; Olympus Corporation). Subsequently, cells were dissolved in 10% acetic acid and the OD₅₇₀ absorbance was measure at 570 nm using a microplate spectrofluorometer (EnSpire; PerkinElmer, Inc.).

MCF7 cells in 6-well plates (30–40% cell confluence) were transfected with 10 nM scramble siRNA (Sc siRNA), or specific siRNA against human CSE (Invitrogen; Thermo Fisher Scientific, Inc.) at a final concentration of 50 nM, using Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) at 37°C for 48 h. The medium was replaced 6 h post-transfection, and silencing efficiency was determined by western blotting 48 h post-transfection. The sequences for the CSE-specific siRNA were as follows: Sense, 5′-GGUUUAGCAGCCACUGUAAdTdT-3′; antisense, 5′-UUACAGUGGCUGCUAAACCdTdT-3′, which were designed to target the open reading frame region of CSE mRNA. The sequences for the Sc siRNA were as follows: Sense, 5′-GTTCCCTATGCGTGAGAAAdTdT-3′; antisense, 5′-GCTTACAAGGGTCTTCACAdTdT-3′.

Proteins were extracted from human breast cancer and paracancerous tissues, human breast cancer cells (MCF7 and MDA-MB-231; MCF7 cells were transfected with CSE siRNA for 48 h or were treated with 10, 20 and 30 µM I157172 for 24 h) and normal mammary epithelial Hs578Bst cells using radioimmunoprecipitation assay buffer (50 mM Tris-HCl, pH 8.0; 150 mM sodium chloride; 1.0% NP-40; 0.5% sodium deoxycholate; and 0.1% SDS) supplemented with 10 µg/ml phenylmethylsulfonyl fluoride (Sigma-Aldrich; Merck KGaA). The samples were then centrifuged at 12,000 × g for 10 min, and protein concentration was determined using the bicinchoninic acid protein quantitative kit (Solarbio Science & Technology Co., Ltd.). Protein samples (40 µg) were separated by 10% SDS-PAGE and were transferred to polyvinylidene difluoride membrane (EMD Millipore, Billerica, MA, USA) at 70 mA for 2 h at 4°C. The membrane was then blocked in 5% fat-free milk for 2 h at room temperature, and probed with specific primary antibodies against CSE, SIRT1, STAT3, phosphorylated (p)-STAT3, acetyl-STAT3, B-cell lymphoma 2 (BCL-2), matrix metalloproteinase (MMP)-2, MMP-9, protein kinase B (Akt) and p-Akt at 4°C overnight. After incubation with the secondary antibodies for 2 h at room temperature, the proteins were visualized using an EasyBlot Enhanced Chemiluminescence kit (Sangon Biotech Co., Ltd., Shanghai, China) and were detected using a FluorChem Q Multifluor system (ProteinSimple, San Jose, CA, USA) and semi-quantified using ImageJ2× (National Institutes of Health, Bethesda, MD, USAs). GAPDH was used as a loading control. Primary antibodies were as follows: CSE mouse monoclonal antibody (1:1,000, cat. no. ab54573; Abcam, Cambridge, UK) SIRT1 mouse monoclonal antibody (1:1,000, cat. no. #8469), STAT3 mouse monoclonal antibody (1:1,000, cat. no. #9139), p-STAT3 (Tyr705) rabbit polyclonal antibody (1:1,000, cat. no. #9131), acetyl-STAT3 rabbit polyclonal antibody (1:1,000, cat. no. #2523), BCL-2 rabbit monoclonal antibody (1:1,000, cat. no. #2870), MMP-2 rabbit polyclonal antibody (1:1,000, cat. no. #4022), MMP-9 rabbit polyclonal antibody (1:1,000, cat. no. #3852), Akt rabbit monoclonal antibody (1:1,000, cat. no. #4685) and p-Akt rabbit monoclonal antibody (1:1,000, cat. no. #4060) (Cell Signaling Technology, Inc., Danvers, MA, USA). GAPDH mouse monoclonal antibody (1:1,000, cat. no. AG019) was obtained from Beyotime Institute of Biotechnology (Shanghai, China). Secondary antibodies were as follows: Horseradish peroxidase-conjugated goat anti-rabbit (1:10,000, cat. no. SA00001-2) and horseradish peroxidase-conjugated goat anti-mouse (1:10,000, cat. no. SA00001-1) (Proteintech Group, Inc., Chicago, IL, USA).

Each experiment was repeated at least three times. Statistical analyses were performed using SPSS 17.0 software (SPSS, Inc., Chicago, IL, USA). Data are expressed as the means ± standard deviation. Differences between multiple groups were analyzed using one-way analysis of variance followed by Bonferroni post hoc test. Differences between two groups were analyzed by Student's t-test. P<0.05 was considered to indicate a statistically significant difference.

The flowchart of the virtual screening workflow is depicted in Fig. 1A. MOE Dock was used for docking simulations of the ligands and for predicting their binding affinity with the homology model. The final top 100 hits were selected and the top 12 hits are shown as Fig. 1B. I157172 with the lowest binding score (S:-7.9215) possessed the highest binding affinity with the homology model. The 3D and 2D binding mode of a compound (S:-7.9042 in Fig. 1B) are predicted and shown in Fig. 1C.

I157172 was the top hit, as identified by virtual screening, the structure and naming of which are shown in Fig. 2A. To confirm the inhibitory activity of I157172 on CSE, the present study detected the effects of I157172 on H₂S production and CSE protein expression. The results of methylene blue detection indicated that I157172 significantly decreased H₂S release from MCF7 cells (Fig. 2B), thus suggesting that I157172 may inhibit the activity of CSE. Western blot analysis revealed that I157172 markedly inhibited the protein expression levels of CSE in MCF7 cells (Fig. 2C and D). These results confirmed the function of I157172 as a CSE inhibitor.

The results of an MTS assay revealed that I157172 significantly reduced the viability of MCF7 cells in a dose-dependent manner and the half maximal inhibitory concentration (IC₅₀) was 18.51 µM after 48 h (Fig. 3A and B). The inhibitory effects of I157172 on MCF7 cells were significantly stronger than those of the CSE inhibitor PAG and were equivalent to those of the positive control DOX at 40 µM (Fig. 3C). Furthermore, the damaging effects of I157172 on Hs578Bst normal mammary cells (IC₅₀=26.32 µM) were weaker than those of the positive control DOX (IC50=1.62μM) (Fig. 3D). In addition, an EdU assay was performed to further evaluate the inhibitory effects of I157172 on cell proliferation. As shown in Fig. 3E and F, I157172 markedly decreased the number of EdU⁺ MCF7 cells in a dose-dependent manner. In addition, the inhibitory activities of I157172 on cell migration and invasion were observed in MCF7 cells (Fig. 3G-J). Taken together, the novel CSE inhibitor I157172 may effectively inhibit the proliferation, migration and invasion in MCF7 cells.

Our previous study demonstrated the biological functions of the CSE/H₂S system in promoting breast cancer development and progression, and revealed that upregulation of CSE promotes activation of STAT3 (10). Acetylation of STAT3 is essential for its activation, whereas SIRT1 can induce deacetylation of STAT3 (16–22). To further explore the mechanism underlying the effects of the CSE/H₂S system on the promotion of breast cancer development, this study investigated the effects of the CSE/H₂S system on SIRT1 expression and STAT3 acetylation. The association between CSE and SIRT1 expression in breast cancer tissues and paracancerous tissues was initially investigated. As shown in Fig. 4A and B, SIRT1 expression was reduced in breast cancer tissues with high CSE expression. Furthermore, low expression levels of SIRT1 were detected in MCF7 and MDA-MB-231 cells, which had high CSE expression levels (Fig. 4C-E). Furthermore, it was revealed that CSE knockdown induced an increase in SIRT1 levels (Fig. 5A-C), and decreased STAT3 (Fig. 5A and D), acetyl-STAT3 (Fig. A and E), and p-STAT3 (Tyr705) levels (Fig. 5A and F) in MCF7 cells. Subsequently, the target proteins downstream of STAT3 were analyzed and it was revealed that knockdown of CSE resulted in downregulation of p-Akt, BCL-2, MMP-2 and MMP-9 proteins (Fig. 5G-L). These data suggested that CSE/H₂S system might promote the activation of STAT3 via reducing the levels of SIRT1.

To further confirm the role of SIRT1 in the activation of STAT3 and explore the mechanism underlying the effects of the CSE inhibitor I157172 on breast cancer, the effects of I157172 on SIRT1, acetyl-STAT3 and p-STAT3 expression were investigated. Western blotting revealed that I157172 upregulated the expression levels of SIRT1, and downregulated acetyl-STAT3 and p-STAT3 expression in a dose-dependent manner (Fig. 6A-D). Furthermore, the expression levels of STAT3 downstream proteins, p-Akt, BCL-2, MMP-2 and MMP-9, were reduced in a dose-dependent manner in MCF7 cells treated with I157172 (Fig. 6E-K). These data further confirmed the roles of the CSE/H₂S system in deacetylation of STAT3 via SIRT1 in promoting breast cancer progression, and indicated that I157172 promoted SIRT1-mediated deacetylation of STAT3 in breast cancer cells and consequently inhibited the growth of breast cancer cells.