The MGC-803 and BGC-823 cells were obtained from Cell Bank of Chinse Academy of Sciences at Shanghai, China. Cells were maintained in Dulbecco's Modified Eagle Medium (DMEM) supplied with 10% fetal bovine serum (FBS) in the presence of 100 IU/ml penicillin and 100 μg/ml streptomycin in a 37°C humidified incubator with 5% CO₂.

The non-radioactive, colorimetric assay system using (3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2-H-tetrazolium bromide) MTT was performed to evaluate the antiproliferative activity of monobenzone on MGC-803 and BGC-823 cells. First, cells were plated into 96 wells plates at a density of 10³ cells per well, then cells were incubated in medium containing diverse concentrations of compounds as indicated. After that, 20 μL MTT was dispensed to each well. After 4 h culture, the supernatant was discarded and 200 μL dimethyl sulfoxide (DMSO) was used to dissolve the insoluble formazan. Finally, the result was read out at 490 nm on the multiplate reader (H1, BioTek, United States).

The KDM1A assay was performed using a KDM1A Inhibitor Screening Assay Kit from Cayman Chemical (Cayman, United States). Briefly, test compounds were dissolved in 100% DMSO and 10 μL of the diluted drug sample (final concentration at 10 μM), 20 μL horseradish peroxidase (HRP), 140 μL assay buffer, 20 μL KDM1A recombinant and 10 μL fluorometric substrate (10-acetyl-3,7-dihydroxyphenoxazine, named as Amplex Red) were added to a black 96-well plate. Then, 20 μL peptide (dimethyl K4 peptide corresponding to the first 21 amino acids of the N-terminal tail of histone H3) was added to each well except the background well so that to initiate the reaction. After 30 min incubation at 37°C, amount of product (resorufin) was analyzed on an multiplate reader (H1, BioTek, United States) with an excitation wavelength of 530 nm and an emission wavelength of 595 nm.

After treatment, cells were collected and and resuspended in RIPA buffer [50 mM Tris HCl pH 8.0, 150 mM NaCl, 1% Nonidet P 40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 1:100 protease inhibitor (Sigma-Aldrich, United States), 1 mM NaV, and 1 mM NaF in H₂O] on ice for 20 min. Then, the lysate was centrifuged at 12,000 rpm for 15 min at 4°C so that to remove debris. Finally, the supernatant was collected and subjected to BCA assay kit (Beyotime, China) for protein quantification.

Sample was mixed with 1 × SDS loading dye (Bio-Rad, United States) and boiled at 95°C for 10 min. Then this mixture was subjected to 10% SDS-PAGE gel. After running, gel was transferred for 2.0 h at 1000 V, and membrane was blocked with 5% milk in PSBT (0.1% Tween-20 in 1 × PBS) for 1 h at room temperature. Then, membrane was incubated with appropriate primary antibody in PBST at 4°C overnight. Following three times washing with PBST, the membrane was incubated with corresponding secondary antibody for 2 h at room temperature. Then, wash the membrane 3 times for 10 min with PBST again, and get the image with X-ray film with the aid of SuperSignal West Sensitivity Substrate (Thermo Scientific, United States).

After cells were treated as indicated, cells were collected and subjected to RNA extraction (CWBio, China). Then, cDNA was synthesized by cDNA synthesis kit (CWBio, China) with the extracted RNA. qPCR was performed using the SYBR mix (CWBio, China) with under the Applied Biosystems Q6 (USA). The 2^−ΔΔCt method was applied to calculate the relative amount mRNA as compared to controls.

Scratch assay was used to examine the wound restoring capacity. Cells were seeded into 24 wells plate (10⁵ cells each well) and grown until reached to 80% confluence. After that, scratches were made using sterile 10 μL pipette tip and cells were re-washed with PBS. Then, media was replaced with serum-free DMEM as well as indicated compound and cells were imaged with microscope (Ts100, Nikon, Japan). After incubating with indicated time, pictures were taken in the same position before and after incubation to document the wound-healing process. Scratch experiments were repeated three times at least and representative pictures are included in this study.

10⁵ cells were suspended in cell culture medium without FBS, and then seeded into the upper chamber. Meanwhile, 700 µL medium supplied with 20% FBS was used as a chemoattractant and added to the lower chamber. Finally, cells on the upper chamber was removed with swab, and cells on the lower side of the membrane were fixed in 95% ethanol and stained with 2 μg/ml DAPI (Beyotime, China), followed with the analysis by IN Cell Analyzer 6500 HS high content analysis (GE, United States).

Structure of KDM1A/LSD1 (PDB ID: 6KGK) was obtained from the Protein Data Bank. Molecular docking was performed by Molecular Operating Environment (MOE) modeling software, 2019.0102. Ligand binding pocket residues were selected by graphical tools in MOE, and chemical structures of compounds were written with MOE either. In the docking calculation, potential energy maps of the receptor were calculated using default parameters. Compounds were imported into MOE. Conformational sampling was based on the Monte Carlo procedure, and finally the lowest-energy and the most favourable orientation of the ligand was selected.

Data were analyzed by Student’s t tests and log rank test for group differences, and by two-way ANOVA for condition and group differences together using GraphPad Prism software. *p < 0.05, **p < 0.01.

To discover KDM1A inhibitor with novel skeleton, an FDA approved drug library containing 1971 compounds from APExBIO was subjected to KDM1A inhibitor screening assay. In the first-round screening, all compounds were tested at 10 μM. Then, the 12 compounds with inhibitory rate more than 50% were collected and subjected to KDM1A inhibitor screening assay with H₂O₂ instead of KDM1A recombinant, so that false positive compound that can react with H₂O₂ or HRP can be excluded. After that, KDM1A inhibitory activity of all these obtained compounds were tested, and monobenzone (Figure 1A) was characterized as one of the most potent KDM1A inhibitors with IC₅₀ = 0.4507 μM (Figure 1B). To evaluate the selectivity, monobenzone was subjected to monoamine oxidase A/B (MAO-A/B) assay as KDM1A is a homology of MAO-A/B, and results suggested that monobenzone failed to inhibit 5% of MAO-A/B activity even at 100 μM. To explore the binding mechanism, monobenzone was subjected to the dilution assay and dialysis assay to test the KDM1A binding reversibility, and results in Figures 1C,D indicated that when monobenzone was incubated with KDM1A recombinant at a high concentration at 30 μM, depletion of monobenzone by either 80 folds dilution or dialysis can refresh the activity of KDM1A recombinant, suggesting that monobenzone may bind and inhibit KDM1A in a reversible manner as CC-90011, which is a reversible KDM1A inhibitor in clinic trials (Dai et al., 2020). Further docking analysis using Molecular Operating Environment (MOE, 2019.0102) revealed that monobenzone may penetrate into the FAD binding cavity of KDM1A, and overlap with the position of FAD in the catalytic cavity (Figures 1E,F). Moreover, ligand interaction analysis for the docking result using Discovery Studio Visualizer 4.5 indicated that monobenzone may form two hydrogen bonds with Arg316 and Thr624, respectively. Additional cell thermal shift assay also confirmed that monobenzone can bind KDM1A in MGC-803 cells (Figure 1G). All these results gave a support that monobenzone is a potent and reversible KDM1A inhibitor.

Once monobenzone has been characterized as a potent and reversible KDM1A inhibitor, whether monobenzone is cellular active remains unknown. To answer this question, as KDM1A was reported to contribute to gastric cancer progression (Ma et al., 2015; Dai et al., 2020; Ma et al., 2020), monobenzone was applied to gastric cancer cell lines MGC-803 and BGC-823, two LSD1 overexpressed cell lines as reported (Zheng et al., 2013; Fang et al., 2017), for 5 days treatment, respectively. Results indicated that monobenzone can inhibit MGC-803 and BGC-823 cells proliferation with IC₅₀ as 7.82 ± 0.55 μM and 6.99 ± 0.51 μM, respectively. Hence, inhibitory activity of monobenzone against cellular KDM1A was further evaluated. As H3K4me1/2 and H3K9me1/2 are KDM1A substrates, and KDM1A could demethylate H3K4me1/2 and H3K9me1/2 in a FAD dependent manner (Amente et al., 2013), to evaluate whether the monobenzone can effectively inhibit the enzyme activity of KDM1A, we detected the change of KDM1A substrates after monobenzone treatment. As shown in Figure 2, monobenzone treatment for 5 days could significantly induce the accumulation of H3K4me1/2 and H3K9me1/2 in a dose dependently manner, which confirmed that the cellular activity of KDM1A was inhibited by monobenzone treatment.

As in vitro experiments confirmed that monobenzone inhibited KDM1A enzyme activity effectively, and KDM1A plays as a contributor in the migration of gastric cancer (Pan et al., 2019; Zhang et al., 2019), we assessed the effects of monobenzone on the migration ability of gastric cancer cells. To assess the inhibitory activity of monobenzone on gastric cancer migration, wound healing assay was performed, and results in Figures 3A,B suggested that monobenzone can inhibit the wound healing in a dose dependent manner in either MGC-803 (Figure 3A) or BGC-823 (Figure 3B) cell lines. To further validate the inhibitor role of monobenzone on gastric cancer migration, transwell experiment was applied in combination with high contenting screening, and result in Figure 3C showed that monobenzone can weaken the migration ability of MGC-803 and BGC-823 cells in a concentration dependent manner either. Taken together, these experiments proved that monobenzone treatment could effectively repressed migration of gastric cancer cells.

As we have confirmed that monobenzone treatment could effectively repress migration of gastric cancer cells, the mechanism behind is still unknown. As reported, KDM1A is a driver of epithelial–mesenchymal transition (EMT) and it can contribute to cell migration and invasion (Ambrosio et al., 2017; Boulding et al., 2018; Das et al., 2019), mRNA expressions of EMT biomarkers, including CDH1, ZO-1, N-Cadherin and Vimentin, were investigated when gastric cancer cells were exposed to different concentrations of monobenzone for 36 h. As shown in Figures 4A,B, monobenzone can significantly induce the mRNA accumulation of CDH1 and ZO-1, two epithelial cell markers; while it can also suppress the mRNA amount of N-Cadherin and Vimentin, two mesenchymal cell markers. All these data suggested that monobenzone can reverse EMT of gastric cancer cells.