The human gastric cancer cell lines, MGC803, HGC27, AGS, SGC7901, BGC823, and SNU-216, were purchase from the cell bank of the Chinese Academy of Science and identified by short tandem repeat analysis within 3 years. Cells were maintained at 37°C with 5% CO₂. MGC803, HGC27, SGC7901, and BGC823 were cultured in Dulbecco’s modified Eagle’s medium (DMEM) medium; AGS and SNU-216 were cultured in Roswell Park Memorial Institute (RPMI)-1640 medium; medium (Thermo Fisher Scientific) containing 100 U/ml penicillin–streptomycin (Millipore Sigma); and 10% fetal bovine serum (FBS) (Biological Industries).

CRBN was overexpressed in pLX304-CRBN-V5 vector (39). The short hairpin RNA (shRNA) of CRBN (the sequences: CCGGGCCCACGAATAGTTGTCATTTCTCGAGAAATGA CAACTATTCG) was constructed in the pLKO.1 vector (40). Envelop plasmid pMD2.G (Cat: 12259, Addgene), packaging plasmid psPAX2 (Cat: 12260, Addgene), and CRBN plasmid were cotransfected into 293FT cells. 293FT cells were transfected for 6 h and cultured with fresh medium for 48 h. The viral supernatant was collected and filtered. Lentiviruses were incubated with gastric cancer cells for 24 h. Puromycin or blasticidin (Sigma-Aldrich) was used to screen for stable cell lines.

Gastric cancer cells (1 × 10⁴) were cultured in 96-well plates per well overnight; ARV-825, OTX015, and JQ1 (Cat: HY-16954, HY-15743, HY-13030, MedChemExpress) with different concentrations were added into each well. At 72 h after ARV-825 treatment, CCK8 (Dojindo) was added into 96-well plates per well with previous methods (41). A reader (Thermo Fisher Scientific) read absorbance at 450 nm. Data analysis was conducted by Graph Prism software 8.4.0.

MGC803, HGC27, AGS, and SGC7901 cells (1,000–2,000) were cultured in six-well plates per well, respectively. ARV-825 of different concentrations were added to treat cells after 24 h. Incubated with 5% CO₂ at 37°C for 2 weeks, the gastric cancer cells were fixed with 100% methanol for 15 min and stained with Giemsa for 1 h (Solarbio). The six-well plates were scanned, and the number of clones was counted.

Gastric cancer cells were collected and centrifugated at 180g for 4 min, suspended in cold 70% ethanol overnight, washed with phosphate-buffered saline (PBS), and incubated for 1 h with light-free at room temperature after adding propidium iodide (Cat. P4170, Sigma). The cell cycles were tested by a Beckman Gallios™ Flow Cytometer (Beckman).

Cell apoptosis was assessed as previous protocols (42). Gastric cancer cells were treated with ARV-825 of different concentrations. At 72 h after treatment, the cells were trypsinized, centrifugated at 180g for 4 min and washed with PBS. The cells were suspended in binding buffer and stained using the fluorescein isothiocyanate-Annexin V apoptosis kit (Cat. 556420, BD Biosciences). The cell apoptosis was tested and analyzed by a Beckman Gallios™ flow cytometer.

Western blotting analysis was carried out as previous protocols (42). Gastric cancer cells were seeded into six-well plates per well; ARV-825 of different concentrations were added to treat cells. After treated with ARV-825 for 72 h, the cells were harvested and extracted by in radioimmunoprecipitation assay (RIPA) buffer. The supernatant was added with loading buffer (Cat. B1012-100, Applygen Technologies) and boiled for 10 min at 98–100°C. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was performed on proteins and electrotransferred to polyvinylidene fluoride (PVDF) membranes. The membranes were incubated with 5% skimmed milk powder solution and then incubated with primary antibodies at 4°C overnight: anti-BRD4 (Cat. 13440s), anti-BRD2 (Cat. 5848s), anti-poly-ADP-ribose polymerase (PARP) (Cat. 9542), anticleaved-caspase 3 (Cat. 9664), and c-Myc (Cat. 9402) were purchased from Cell Signaling Technology. Other primary antibodies are listed below: anti-BRD3 (Cat. 11859-1-AP, Proteintech), anti-CRBN (Cat. HPA045910, Sigma-Aldrich), anti-polo like kinase 1 (PLK1) (Cat. ab17056, Abcam), and antiglyceraldehyde-3-phosphate dehydrogenase (GAPDH) (MA3374, Millipore). The membranes were incubated with horseradish peroxidase-conjugated (HRP) secondary antibodies: goat antimouse IgG and goat antirabbit IgG (Jackson ImmunoResearch). The immunoreactive proteins were revealed and analyzed using an ECL detection kit (Pierce) and a LAS 4010 imaging system (GE).

RNA-sequencing (RNA-seq) was implemented using the protocols provided by Novogene (Novogene Co., Ltd.). Total RNA of gastric cancer cells was extracted by the TRIzol^® reagent (Invitrogen). First, RNA was reverse transcribed to cDNA for library construction and sequencing. RNA-seq procedure was performed on HGC27 cells treated with ARV-825 (n = 3) or dimethyl sulfoxide (DMSO) (n = 3). Genes of differential expression (|log₂fold change| > 1 and p < 0.05) were identified using Bioconductor limma analysis according to DAVID Bioinformatics Resources v6.8 (https://david.ncifcrf.gov). Gene set enrichment analysis (GSEA) identified a series of genes that were performed to detect cellular pathways that influenced the cell apoptosis induced by ARV-825 according to the GSEA Application (http://www.broadinstitute.org/gsea/).

The animal experiments complied with the requirements of the Institutional Animal Care and Use Committee of Shanghai Jiao Tong University (No. 201801054). Four-week-old male nude mice (Lingchang BioTech, n = 6 per group) were injected subcutaneously 5 × 10⁶ HGC27 cells in their front flank. Tumor size was measured every 3 days. The calculation formula of tumor volume is (length × width × height)/2. After 2 weeks, when the size of the engrafted tumor came to about 100 mm³, the mice were injected intraperitoneally either ARV-825 at 10 mg/kg or menstruum control (10% Kolliphor^®HS15, BASF) every day. The mice were sacrificed when the tumors in the control group exceeded 1,000 mm³. The tumors were excised and then embedded in paraffin for immunohistochemistry.

Immunohistochemistry was carried out as previous protocols (43). The Ki-67 antibody (Cat. ab15580, Abcam) and HRP/DAB detection kit (Cat. ab64261, Abcam) were used. Tissue sections of immunohistochemical staining were examined using the Olympus BX41 imaging system. The calculation formula of the total scoring (TS) was as follows: TS = the intensity (I) × the percentage of positive cells (P).

SPSS software version 20.0 (IBM) was used to analyze the data. Significant difference of data was analyzed by Student’s t-tests. GraphPad Prism version 8.4.0 was used to draw the figures. In the figures, means ± the standard deviation (SD) is shown; p-values < 0.05, the results were statistically significant, for which *p < 0.05, **p < 0.01, and ***p < 0.001.

Expression of BRD4 was measured in different kinds of tumors, which was displayed using gene expression profiling interactive analysis (GEPIA, http://gepia.cancer-pku.cn/index.html), which illustrated that mRNA expression of BRD4 in tumor tissue was raised remarkably than that in normal tissue only in stomach adenocarcinoma (STAD) and esophageal adenocarcinoma (ESCA, Figure 1A ). High mRNA expression of BRD4 was associated with poor overall survival of patients with gastric cancer ( Figures 1B, C ). The criterion of categorizing high or low expression was median survival time.

BRD4, BRD2, and BRD3 were abundantly expressed in MGC803, HGC27, AGS, SGC7901, BGC823, and SNU-216 cells ( Figure 2A ), implying that the BET proteins were diffusely expressed in gastric cancer cells. ARV-825 consists of CRBN and OTX015. The effect of increasing doses of ARV-825 on gastric cancer cell lines incubated for 72 h was evaluated. CCK8 assays showed that a dose-dependent decrease in gastric cancer cell viability was observed after ARV-825 treatment ( Figure 2B ). The chemical structures of ARV-825, OTX015, and JQ1 are shown in Figure 2C . HGC27 and MGC803 had lower IC50 than other gastric cancer cells. The IC50 values of other BRD4 inhibitors, such as OTX015 and JQ1, were compared with that of ARV-825 in gastric cancer cells ( Figure 2D ). The results showed that ARV-825 had lower IC50 values and showed a better suppression effect on gastric cancer cell viability than OTX015 and JQ1. Decreased quantity and shrinkage of the volume of gastric cancer cell were examined in the group treated with ARV-825 ( Figure 2E ), compared with that in the untreated control group. Clonal formation assay was applied to observe the influence of ARV-825 on the long-term growth of gastric cancer cells ( Figure 3A ); ARV-825 suppressed dose dependently the clonal formation of MGC803, HGC27, AGS, and SGC7901 cells. In conclusion, these findings showed that ARV-825 had antiproliferative effect in gastric cancer cells; the number of clones of ARV-825-treatment groups was remarkably lower compared with that of control groups ( Figure 3B ).

MGC803, HGC27, and BGC823 express substantial amounts of CRBN, while AGS, SGC7901, and SNU-216 cells have relatively low expression of CRBN ( Figure 4A ). The effect of ARV-825 in gastric cancer cells is associated with CRBN expression. Knockdown of CRBN expression in MGC803 and HGC27 cells could decrease partly the growth inhibition influence of ARV-825 ( Figures 4B, D ). By contrast, overexpression of CRBN in gastric cancer cells significantly increased the sensitivity of AGS and SGC7901 cells to ARV-825 ( Figures 4C, E ). These results suggested that CRBN is associated with the growth inhibition activity of ARV-825.

We investigated whether ARV-825 could regulate the cell cycle in gastric cancer cells. MGC803, HGC27, AGS, and SGC7901 cells were treated with ARV-825 for 24 h to perform cell cycle analysis. Compared with the control group, the ARV-825 treatment group showed an increase in the ratio of G1 phase cells and a reduction in the ratio of G2 and S phases cells simultaneously ( Figure 5A ). Apoptosis of gastric cancer cells was also examined after ARV-825 treatment. The Annexin V/PI staining analysis demonstrated that ARV-825 increased cell apoptosis, presenting dose dependence. The ratio of apoptotic cells in the groups with ARV-825 treatment increased dose dependently in contrast to control groups ( Figures 5B, C ). Western blotting analysis revealed that ARV-825 could increase activation of PARP and caspase-3 in the four gastric cancer cell lines ( Figures 6A, B ). These results demonstrated that ARV-825 could induce cell cycle block and apoptosis of gastric cancer cells.

ARV-825 robustly degrades BRD4 protein via the ubiquitin–proteasome system, which consists of a CRBN-recruiting moiety and OTX015. Therefore, we observed the BET protein levels in gastric cancer cells with ARV-825 treatment. Western blotting analysis illustrated that ARV-825 displayed efficient degradation of BRD4 in four gastric cancer cells ( Figures 6A, B ). In addition, ARV-825 could decrease simultaneously BRD2 and BRD3 ( Figures 6A, B ). Gastric cancer cells treated with ARV-825 also resulted in PARP and caspase 3 cleavage. The experimental progress indicated that ARV-825 could downregulate expression level of BET protein in gastric cancer cells.

To determine the potential mechanism of ARV-825, RNA-seq was used to screen and analyze genes (GEO ID: GSE179581). As shown in Figure 7A , under the condition of |log₂fold change | > 1 and an adjusted p < 0.05, compared with those in the control group, 3,584 genes were identified as upregulated and 3,515 genes were downregulated in HGC27 cells of ARV-825 treatment. We continued to analyze the signaling pathways and identify associated genes. ARV-825 markedly downregulated the expression levels of E2F2, CDC45, RBL1, CDC25A, CDK1, PLK1, MYC, SLC19A1, MCM5, MCM4, HK2, SRM, UNG, CDK4, and CDK2 ( Figure 7B ). All the significant GSEA hallmarks have been shown in Supplementary Table S1 ; HALLMARK_G2M_CHECKPOINT and HALLMARK_MYC_TARGETS, which are related with cell-cycle functions, are the top 5 negative hallmarks. GSEA plots showed gene enrichment in HALLMARK_G2M_CHECKPOINT and HALLMARK_MYC_TARGETS signaling pathways after ARV-825 treatment in HGC27 cells. Many genes in the G2M_CHECKPOINT and MYC_TARGETS signaling pathways were downregulated ( Figure 7C ), MYC and PLK1 genes were both downregulated in these two signaling pathways. As expected, Western blotting analysis confirmed that c-Myc and PLK1 protein levels decreased dose dependently with ARV-825 treatment in MGC803, HGC27, AGS, and SGC7901 cells ( Figure 7D ). These results revealed that ARV-825 disturbed BRD4-mediated MYC and PLK1 transcription, resulting in decreasing c-Myc and PLK1 protein levels.

To study the antitumor effect of ARV-825 in vivo, a xenograft model of gastric cancer using HGC27 cells was established. ARV-825 at 10 mg/kg was intraperitoneally injected into mice daily when the volume of subcutaneous tumor achieved about 100 mm³. The tumor burden of the ARV-825 treatment group was significantly reduced ( Figures 8A, C, D ) in contrast to that in the control group.

Meanwhile, the treatment group and control group had no remarkable difference in mouse body weight ( Figure 8B ). Immunohistochemical analysis showed that the ratio of Ki67-positive cells was markedly lower in tumors with ARV-825 treatment than that in the control group ( Figures 8E, F ), illustrating the tumor-inhibiting effect of ARV-825. Furthermore, the levels of BRD4 protein were downregulated significantly by ARV-825 treatment in vivo ( Figures 8G, H ). These results indicated that ARV-825 could remarkably suppress the tumor growth of gastric cancer without obvious side effects.