Enhanced ZBTB10 expression induced by betulinic acid inhibits gastric cancer progression by inactivating the ARRDC3/ITGB4/PI3K/AKT pathway

GC tissues and paired adjacent normal tissues were collected from patients diagnosed with GC who underwent surgical resection at the First Affiliated Hospital of Sun Yat-sen University. All tissue specimens were confirmed by pathologic examination and were separated and frozen at -80 °C or formalin-fixed. Written informed consents were obtained from all the patients, and the study was approved by the Ethics Committee of the First Affiliated Hospital of Sun Yat-sen University.

Human GC cell lines (AGS, MKN1, MKN28, HGC-27 and MGC803) and normal human gastric epithelial cells-1 (GES-1) were purchased from the Chinese Academy of Sciences, Shanghai Branch Cell Bank. All human cell lines have been authenticated using STR profiling within the last three years, and all experiments were performed with mycoplasma-free cells. AGS cells were cultured in DMEM/F12 (Gibco, Waltham, MA, USA), and MKN1, MKN28, HGC-27, MGC803, and GES-1 cells were cultured in RPMI-1640 (Gibco).

BA (B8936, Sigma Aldrich, St. Louis, MO, USA) was dissolved in DMSO (Sigma Aldrich) as a 20 mM stock solution and stored at -20 °C. LY294002 (HY-10108, MCE, Shanghai, China) was dissolved in DMSO as a 5 mM stock solution and stored at -80 °C. MG-132 (HY-13259, MCE) was dissolved in DMSO as a 10 mM stock solution and stored at -80 °C.

For the construction of stable overexpression and knockdown GC cell lines, we purchased lentivirus from iGene Biotechnology (Guangzhou, Guangdong, China) and infected GC cells with it. Subsequently, complete media containing 3 µg/mL puromycin (Sigma-Aldrich) were used to screen stably overexpressing and knocked down GC cell lines. Overexpression plasmids were constructed using pEZ-Lv201 (iGene Biotechnology), and knockdown plasmids were constructed using LVRU6GP (iGene Biotechnology). All plasmids and siRNAs were constructed from iGene Biotechnology (Supplementary Table 2). Plasmids and siRNAs transfection were performed using Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions.

For the proliferation assays, GC cells (1 × 10³/well) were seeded in 96-well plates with six replicates. Proliferation rates were determined using the Cell Counting Kit-8 (CCK-8, Boster Biological Technology, Wuhan, Hubei, China), and measure was performed on a microplate reader (Bio-TEK, Winooski, VT, USA). For the cell viability assay, 5 × 10³ cells per well were inoculated into 96-well plates. After attaching for about 24 h, the culture medium was replaced with complete media containing different concentrations of BA and LY294002 for 24 h. Finally, the absorbance was measured at 450 nm.

5 × 10³ cells per well were inoculated into 96-well plates. The EdU incorporation assay was conducted using an EdU Staining Kit (RiboBio, Guangzhou, Guangdong, China) according to the manufacturer’s instructions. Finally, the cell nucleus was visualized using DAPI, and fluorescence signals were acquired using a fluorescence microscope (Leica, Wetzlar, Germany).

Cells were seeded separately in 6-well plates at a density of 500/well. The medium was changed twice per week. After 14 days, cells were fixed in 4% paraformaldehyde for 30 min, stained with 0.5% crystal violet for 15 min, rinsed three times with PBS to remove excess dye, photographed, and counted.

5 × 10⁴ cells resuspended in 400 µL serum-free medium were plated in the upper chamber, while the lower chamber was filled with the complete culture medium. Finally, the cells were fixed with 4% paraformaldehyde and stained with 0.5% crystal violet. Images were acquired using a microscope (Olympus, Tokyo, Japan).

Cell apoptosis was evaluated by flow cytometric analysis using annexin V-FITC/PI staining (BD Biosciences, San Jose, CA, USA). Apoptotic cells were analyzed using FACSCalibur flow cytometry (BD Biosciences), and FlowJo software (Tree Star Corp, Ashland, WI, USA) was used to analyze the results.

Total RNA was extracted from tissues and cells using TRIzol reagent (Takara, Beijing, China) according to the manufacturer’s protocol. cDNA was generated using the Master Mix cDNA Synthesis Kit (Accurate Biotechnology, Changsha, Hunan, China). qPCR was performed using SYBR Green I (Accurate Biotechnology). The expression levels of the target genes were measured using the 2^–ΔΔCt method and normalized to GAPDH as a reference. Each experiment was performed in triplicate and qPCR primers are listed in Supplementary Table 3.

Western blot was performed as described by Liu [21]. Proteins were extracted from tissues and cells using RIPA lysis buffer (Beyotime, Guangzhou, Guangdong, China) following the manufacturer’s protocol and quantified with the bicinchoninic acid protein assay kit (Beyotime). GAPDH was used as a loading control. The antibodies are listed in Supplementary Table 4.

Concentrations of p-PI3K in AGS and MKN1 cell lysates with ZBTB10 overexpression and knockdown were measured by the ELISA Kits (ml060625, Mlbio, Shanghai, China) according to the manufacturer’s instructions.

The ChIP assay was performed using a ChIP assay kit (Cell Signaling Technology, Boston, MA, USA) according to the manufacturer’s instructions. Briefly, GC cells were crosslinked with 1% formaldehyde and quenched with glycine (Solarbio, Beijing, China) at room temperature, after which they were collected, washed and resuspended in lysis buffer (Cell Signaling Technology). The digested chromatin was incubated with anti-IgG (DIA-AN, Wuhan, Hubei, China) and anti-ZBTB10 (Abcam) for immunoprecipitation. DNA was purified and analyzed by qRT-PCR.

Plasmid preparation was completed by Genechem (Shanghai, China). AGS and MKN1 cells were transfected with corresponding plasmids using Lipofectamine 3000 (Invitrogen). After 48 h, 20 µM BA was added to the medium and treated for 24 h. Luciferase activity was detected by the Multifunctional enzyme marker (Thermo).

For the ubiquitination assay, the GC cells were pretreated with the indicated plasmids for 48 h. After treated with 20 µM MG-132 for 6 h, the cells were lysed by IP lysis buffer. Next, immunoprecipitation against the corresponding protein was performed. Finally, the immunoprecipitants were subjected to WB.

AGS cells with stable ARRDC3 overexpression and negative control were seeded onto the chamber slides. After attaching for about 24 h, the cells on chamber slides were washed with PBS and fixed in 4% paraformaldehyde for 15 min at room temperature, followed by exposure to 0.3% Triton X-100 for 20 min. Subsequently, the cells were blocked in 2% goat serum for 1 h and incubated with specific primary antibodies overnight at 4 °C. The next day, cells were incubated fluorescent secondary antibodies corresponding to the species in dark for 2 h and DAPI was used for nuclear staining. Finally, images were acquired using a confocal microscope (FV3000, Olympus).

The tissues were fixed with 10% paraformaldehyde and embedded in paraffin. After deparaffinization, rehydration, and antigen retrieval, the samples were incubated with primary antibodies using the corresponding antibodies. Images were taken with an Axio Scope.A1 398 vertical microscope (Leica). The final immunoreactive score (IRS) was obtained by multiplication of the intensity score and the quantity score. Briefly, two investigators independently scored the percentage of positive-stained cells (%PC) from 0 to 4 (0: no positive cells; 1: <10%; 2: 10–50%; 3: 51–80%; 4: >80%) and the staining intensity from 0 to 3 (0: no staining; 1: weak staining; 2: moderate staining; 3: strong staining). The IRS was then calculated as the product of these two scores (range, 0–12). We defined an IRS score of less than 6 as low expression and the rest as high expression.

Female BALB/c nude mice (5 weeks old) were obtained from Specific Pathogen-Free Biotechnology Co., Ltd. (Beijing, China) and housed under specific pathogen-free conditions. All animal experiments were performed with the approval of the Animal Care and Use Committee of the Sun Yat-sen University.

For the xenograft model, MGC803 cells (5 × 10⁶ in 100 µL PBS) with ZBTB10 overexpression and negative control were subcutaneously injected into each flank of nude mice (six mice/group). Tumor size was measured every 3 days (volume = length × width² × 1/2). The mice were sacrificed 21 days after tumor cell implantation. The tumors were weighed, imaged, fixed in 4% paraformaldehyde. To investigate the function of BA, xenograft models were established using MGC803 cells with knockdown ZBTB10 and negative control according to the methods in previous studies [17]. Seven days later, the tumors grew to approximately 100 mm³, and tumor-bearing mice were randomly assigned to four groups.

For the lung metastasis model, MGC803 cells (1 × 10⁶ in 100 µL PBS) overexpressing ZBTB10 and the negative control were injected into the tail vein of nude mice. The mice were sacrificed 8 weeks after GC cell injection and pulmonary metastatic nodules were counted. Finally, the lungs were resected, photographed, and fixed in 4% paraformaldehyde for further analyses. Similarly, we explored the effect of BA on lung metastasis according to a previous study [22].

For the popliteal lymph node metastasis model, MGC803 cells (1 × 10⁶ in 20 µL PBS) with ZBTB10 overexpression and negative control were injected into the left footpad of mice (twice in 3 days). The mice were sacrificed after 8 weeks, and the lymph node transfer ratio was determined. The footpad and popliteal lymph nodes of the left leg were resected, photographed, and fixed in 4% paraformaldehyde for further analyses.

To identify genes closely associated with the development and progression of GC, we conducted RNA-sequencing on five pairs of GC tissues and adjacent normal tissues. Our analysis revealed 442 differentially expressed genes (p < 0.01, log₂FC > 2.0), including 280 upregulated and 162 downregulated genes in GC tissues (Fig. 1A, and Supplementary Table 6). While ZBTB protein family has been implicated in various tumors, their research in GC is limited. Notably, our transcriptome data identified ZBTB10 as one of the most significantly downregulated genes in GC (Fig. 1A). Therefore, ZBTB10 was chosen for further exploration. To validate these findings, we examined ZBTB10 mRNA expression in 59 GC cases using qRT-PCR. The results showed a marked downregulation of ZBTB10 mRNA levels in GC tissues compared to adjacent normal tissues (n = 59, p = 0.00073; Fig. 1B). Moreover, consistent with the mRNA levels, ZBTB10 protein levels were also significantly downregulated in GC tissues (Fig. 1C, D, and Supplementary Fig. 1A). Additionally, we also quantified ZBTB10 expression levels in GES-1 and common GC cell lines. The results showed that ZBTB10 expression levels were lower in GC cells compared to that in GES-1 (Fig. 1E, F). Overall, all of the above results indicate that ZBTB10 is lowly expressed in GC.

Next, to elucidate the clinical prognostic relevance of ZBTB10 in GC, we performed IHC staining on a tissue microarray containing samples from 94 GC patients. We categorized ZBTB10 expression into high and low expression groups based on IRS (Fig. 1G). Analysis of clinicopathologic parameters revealed that lower ZBTB10 expression was associated with worse T stage and more advanced tumor stage (Supplementary Table 1). Furthermore, Kaplan–Meier survival analysis showed that lower ZBTB10 expression in GC tissues contributed to poorer prognosis (n = 94, p < 0.001; Fig. 1H). More importantly, multifactorial Cox regression analysis confirmed ZBTB10 as an independent prognostic predictor (Fig. 1I), suggesting its potential as a valuable diagnostic biomarker for GC. Taken together, these findings indicate that ZBTB10 is significantly downregulated in GC and that lower ZBTB10 expression is correlated with worse prognosis.

We next attempted to clarify the role of ZBTB10 in GC progression. Firstly, we constructed AGS, MKN1, HGC-27, and MGC803 cells with stable ZBTB10 overexpression and knockdown, and the efficiencies of stable ZBTB10 overexpression and knockdown were verified by qRT-PCR and WB (Fig. 2A, B, and Supplementary Fig. 1D, 2A). Then, gain or loss-of-function experiments were conducted. The CCK-8, EdU, and colony formation assays demonstrated that ZBTB10 overexpression significantly inhibited GC cell growth within 4 days and reduced the percentage of EdU cells and the number of colonies, and vice versa (Fig. 2C–H, and Supplementary Fig. 1B, 1E–G, 2B, 2C). Next, to examine the effect of ZBTB10 on the metastatic capabilities of GC cells, we performed transwell migration and matrigel invasion assays. The results indicated that ZBTB10 knockdown significantly increased the migration and invasion capabilities of GC cells, whereas ZBTB10 overexpression inhibited these processes (Fig. 2I, J, and Supplementary Fig. 1C, 1H, 2D). Meanwhile, apoptosis assays also revealed that the proportion of apoptotic cells was significantly reduced upon ZBTB10 knockdown, whereas the proportion of apoptotic cells was increased upon ZBTB10 overexpression (Fig. 2K–L, and Supplementary Fig. 1I, 2E). The above results suggest that ZBTB10 could inhibit GC cells proliferation, metastasis, and anti-apoptotic capabilities in vitro.

To confirm our in vitro findings and explore the biological role of ZBTB10 in vivo, we constructed a BALB/c nude mouse model using MGC803 cells with stable ZBTB10 overexpression and negative control cells. In the subcutaneous xenograft model, compared with tumors from negative control cells, tumors derived from ZBTB10-overexpressing cells exhibited significantly reduced growth, both in terms of volume and weight (Fig. 3A–C, and Supplementary Fig. 3A). IHC analysis further revealed increased ZBTB10 expression and a higher proportion of apoptotic cells in the ZBTB10-overexpression group, alongside reduced Ki-67 expression (Fig. 3D). Furthermore, we also explored whether ZBTB10 could affect the metastasis of GC in vivo by constructing a lung metastasis model and a footpad-popliteal lymph node (LN) metastasis model. As expected, ZBTB10 overexpression significantly inhibited lung metastasis, with fewer metastatic nodules observed in the ZBTB10-overexpression group compared to the negative control (Fig. 3E–G, and Supplementary Fig. 3B, C). In the footpad-popliteal LN metastasis model, LNs from mice injected with ZBTB10-overexpressing cells were smaller than those from the negative control cells (Fig. 3H–J, and Supplementary Fig. 3D). Moreover, HE staining revealed that LNs of the ZBTB10 overexpression group had a lower percentage of metastasis compared with the negative control group (Fig. 3K, L). Collectively, these observations demonstrate that ZBTB10 also plays a crucial role in inhibiting GC growth and metastasis in vivo.

To pursue the underlying molecular mechanism of ZBTB10 inhibiting GC progression, we conducted whole genome RNA-sequencing and subsequent functional enrichment analysis between AGS cells with the negative control and ZBTB10 overexpression. Our sequencing data revealed that ZBTB10 overexpression resulted in altered expression of 323 genes (Supplementary Fig. 4A, and Supplementary Table 7). Further pathway analysis of these differentially expressed genes showed that ZBTB10 overexpression inhibited the activation of multiple cancer-promoting pathways, such as the PI3K/AKT pathway, the MAPK pathway, and the Ras pathway (Supplementary Fig. 4B). Among them, the PI3K/AKT pathway was the most affected pathway by ZBTB10 overexpression (Supplementary Fig. 4B). To further investigate this relationship, we performed a Phospho-proteomic profiling to detect changes in the expression levels of common phosphorylated proteins upon ZBTB10 overexpression. The results showed that phosphorylation of AKT was significantly inhibited after ZBTB10 overexpression (Fig. 4A). Furthermore, ELISA and WB also demonstrated that ZBTB10 overexpression reduced the protein levels of p-PI3K and p-AKT in AGS and MKN1 cells, whereas ZBTB10 knockdown had the opposite effect (Fig. 4B, C). Consequently, we propose the hypothesis that ZBTB10 inhibits GC progression through inactivating the PI3K/AKT pathway.

Subsequently, we constructed in vitro functional rescue experiments. We used the PI3K inhibitor LY294002 (LY2; 10 µM) [23] to block the PI3K/AKT pathway in ZBTB10-knockdowning cells and detected the proteins levels by WB (Fig. 4D, F). Inactivation of the PI3K/AKT pathway reversed the promoting effect on GC cells induced by ZBTB10 knockdown (Fig. 4E, G, and Supplementary Fig. 4D–F). Similarly, targeting AKT with siRNAs in ZBTB10-knockdown cells yielded results consistent with PI3K inhibition (Fig. 4H-K, and Supplementary Fig. 4G–I). Through these data, we summarize that ZBTB10 inhibits the proliferation, metastasis, and anti-apoptotic capabilities of GC cells through inactivating the PI3K/AKT pathway.

We then explored the detailed mechanism of ZBTB10 inactivating the PI3K/AKT pathway. Through further comprehensive analysis of sequencing data, we found that ZBTB10 overexpression promoted ARRDC3 expression (Supplementary Fig. 5A). Furthermore, Chip-sequencing data from GEO (GSE105183) and the Genecards database (https://​www.​genecards.​org/​) both revealed that there were the direct interaction sites between ZBTB10 and the ARRDC3 promoter region. Notably, previous studies have shown that ARRDC3 inhibited liver fibrosis and epithelial-to-mesenchymal transition via the ITGB4/PI3K/AKT pathway [24]. Thus, we hypothesized that ZBTB10 might promote ARRDC3 expression through transcriptional activation, thereby inhibiting the PI3K/AKT pathway and GC progression. To confirm our hypothesis, we quantified ARRDC3 mRNA and protein levels in cells with ZBTB10 overexpression and knockdown using qRT-PCR and WB. The results indicated that ZBTB10 overexpression led to elevated ARRDC3 levels, while ZBTB10 knockdown reduced them (Fig. 5A, B). Moreover, IHC staining of the previous subcutaneous tumor samples also showed that ZBTB10 overexpression upregulated ARRDC3 expression (Fig. 5C). We then assessed ARRDC3 expression in GC tissues and cell lines; the results revealed that ARRDC3 was downregulated in GC tissues and cell lines, and there was a significantly positive correlation between ZBTB10 and ARRDC3 expression in GC tissues (Fig. 5D, E, and Supplementary Fig. 5B). Meanwhile, lower ARRDC3 expression was associated with shorter overall survival in patients with GC (Fig. 5F, and Supplementary Fig. 5C). Gain- and loss-of-function studies further demonstrated that ARRDC3 overexpression inhibited GC cell proliferation and metastasis while promoting apoptosis, whereas ARRDC3 knockdown had the opposite effects (Supplementary Fig. 5D–I). To verify that ZBTB10 promoted ARRDC3 expression through a transcription mechanism, we conducted ChIP and dual-luciferase reporter assays. ZBTB10 overexpression enhanced the binding ability of ZBTB10 to the promoter region of ARRDC3 (Fig. 5G), whereas ZBTB10 knockdown reduced it (Fig. 5H). Correspondingly, ZBTB10 overexpression increased reporter activity (Fig. 5I), whereas ZBTB10 knockdown decreased it (Fig. 5J), supporting the hypothesis of ZBTB10-mediated transcriptional activation of ARRDC3. Finally, we performed rescue experiments by knocking down ARRDC3 in LCN2-overexpressing AGS cells and overexpressing ARRDC3 in ZBTB10-knockdown MKN1 cells. Results showed that ARRDC3 knockdown rescued the proliferation, migration, and invasion abilities of ZBTB10-overexpressing AGS cells (Fig. 5K, and Supplementary Fig. 5J, K), and ARRDC3 overexpression eliminated the promoting effect of ZBTB10 knockdown in GC progression (Fig. 5L, and Supplementary Fig. 5L, M). These results suggest that ZBTB10 functions as a TF to upregulate ARRDC3, which is a potential tumor suppressor gene and a key downstream target in ZBTB10-regulated GC progression.

As previously demonstrated, ARRDC3 inhibited tumor progression by negatively regulating ITGB4 [25, 26], and the integrin family was known to be one of the regulators of the PI3K/AKT pathway [27]. To explore whether ARRDC3 demonstrates the above mechanism in GC, we examined the expression of ITGB4, p-PI3K, and p-AKT following ARRDC3 overexpression and knockdown. WB analyses showed that ARRDC3 overexpression markedly reduced ITGB4, p-PI3K, and p-AKT levels, whereas ARRDC3 knockdown had the opposite effect (Fig. 5M). Moreover, IF analysis showed that ITGB4 surface levels were significantly down-regulated in cells with ARRDC3 overexpression (Fig. 5N). It has been reported that ARRDC3 promoted protein degradation by mediating ubiquitination levels [25, 28]. Therefore, we next sought to elucidate whether ARRDC3 inhibits ITGB4 expression by promoting its ubiquitination level. We found that treatment with the proteasome inhibitor MG-132 (20 µM) prevented the ARRDC3-mediated decrease in ITGB4 levels (Fig. 5O). Furthermore, Co-IP assays also indicated that MG-132 could reverse the ubiquitination of ITGB4 induced by ARRDC3 (Fig. 5P). More importantly, we found that ARRDC3 overexpression could counteract the elevation in ITGB4, p-PI3K, and p-AKT levels induced by ZBTB10 knockdown, indicating that ARRDC3 was indispensable for the ZBTB10-mediated inhibition of the ITGB4/PI3K/AKT pathway (Fig. 5Q). Collectively, these results demonstrate that ZBTB10 transcriptionally activates ARRDC3 and then enhanced ARRDC3 directly binds to ITGB4 leading to its ubiquitination and degradation, which in turn downregulates PI3K and AKT phosphorylation.

Emerging evidences have suggested that ZBTB10 might be a target of multiple anti-cancer drugs [29, 30], but this has not been reported in GC. To explore whether ZBTB10 is the target for a drug with anti-GC effects, we made predictions using the connectivity map (CMap) database [31] combined with the sequencing data of control cells and ZBTB10-overexpressing cells (Fig. 6A). The results showed that the differentially expressed genes after ZBTB10 overexpression were basically consistent with those affected by BA treatment (Supplementary Table 5). Additionally, ZBTB10 has been reported as a target of BA in breast and colon cancers. Therefore, we hypothesized that ZBTB10 might also be a target of BA in GC. Firstly, we performed cell viability assays and found that BA inhibited GC cell proliferation in a concentration- and time-dependent manner, with IC-50 values of 40.46 µM and 40.17 µM for AGS and MKN1 cells, respectively, after 24 h of treatment (Supplementary Fig. 6A, B). Subsequently, treatment of AGS and MKN1 cells with 20 µM BA [18] for 24 h confirmed that BA increased ZBTB10 mRNA and protein levels, while decreasing p-PI3K and p-AKT levels in a concentration-dependent manner (Fig. 6B, C). The above results indicate that ZBTB10 may be a potential target of BA in GC, and that BA may inhibit GC progression through the ZBTB10/PI3K/AKT pathway.

To further validate this, we conducted rescue experiments with DMSO and 20 µM BA in control and ZBTB10 knockdown cells. BA treatment decreased the upregulation of p-PI3K and p-AKT induced by ZBTB10 knockdown (Fig. 6D, I), and ZBTB10- and ARRDC3-knockdown GC cells exhibited increased resistance to BA (Fig. 6E, J, and Supplementary Fig. 6C, D). Moreover, ZBTB10 knockdown partially rescued the inhibition of colony formation, metastasis, and anti-apoptotic effects induced by BA treatment in GC (Fig. 6F–H, K–M, and Supplementary Fig. 6E, F). Collectively, these findings confirm that ZBTB10 is a target of BA in GC, and that BA inhibits GC progression via the ZBTB10/ARRDC3/ITGB4/PI3K/AKT pathway.

Finally, we assessed whether BA could inhibit GC progression in vivo and whether this effect depends on ZBTB10. We constructed a rescue mouse model with MGC803-shNC and MGC803-shZBTB10 cells, treated with or without BA. In the subcutaneous xenograft model, BA treatment delayed tumor growth in terms of both volume and weight, and ZBTB10 knockdown partially reversed this effect (Fig. 7A–C, and Supplementary Fig. 7A). Correspondingly, Ki-67 expression decreased, and ZBTB10 expression and apoptosis rates increased with BA treatment (Fig. 7G). Most importantly, BA did not cause significant weight loss and induce any morphological variations nor functional changes in the liver, heart, spleen, lung, and kidney throughout the experiment (Fig. 7D–F). Similarly, in the lung metastasis model, BA significantly inhibited lung metastasis, and ZBTB10 knockdown diminished this effect (Fig. 7H, I, and Supplementary Fig. 7B). These results confirm that BA inhibits GC tumorigenicity and metastasis by promoting ZBTB10 expression in vivo.