ASAP3 disrupts ASAP1-ARHGAP12 to inhibit RhoA and yes-associated protein/transcriptional coactivator with PDZ-binding motif, suppressing gastric cancer progression

Abstract

BACKGROUND

Gastric cancer (GC) has high global mortality, with limited efficacy of advanced therapies due to unclear pathogenesis. While other ASAP family members like ASAP1 are recognized oncoproteins, the role of ASAP3 in GC is poorly understood.

AIM

To investigate the functions and mechanisms of ASAP3 in GC.

METHODS

The biological effects of ASAP3 were assessed in vitro using AGS and HGC-27 GC cell lines with ASAP3 overexpression or knockdown, evaluating proliferation, apoptosis, migration, and invasion. A subcutaneous xenograft model was used for in vivo validation. Underlying mechanisms were explored via transcriptomics, proteomics, and molecular biology techniques including co-immunoprecipitation and RhoA activation pull down.

RESULTS

ASAP3 expression in HGC-27 cells was significantly higher than that in AGS cells. Functional experiments demonstrated that ASAP3 overexpression suppressed GC cell proliferation, migration, invasion, induced S arrest/apoptosis, and inhibited tumor growth, while its knockdown promoted these malignant phenotypes. Then, transcriptomic and proteomic analyses respectively identified 382/714 differentially expressed messenger RNAs and 98/66 differential expressed proteins between short hairpin (sh)-ASAP3-HGC-27/over-expression (oe)-ASAP3-AGS and sh-negative control (NC)-HGC-27/oe-NC-AGS cells, which significantly enriched in Hippo pathway, and GTPase regulation. Real-time quantitative polymerase chain reaction showed ASAP3 silencing in HGC-27 cells upregulated CCN1, AMOTL2 while downregulated CCN2, while ASAP3 overexpression in AGS cells reversed these trends. Western blot further showed ASAP3 silencing reduced the phosphorylation of MST1/MST2, LATS1/LATS2, yes-associated protein (YAP), and transcriptional coactivator with PDZ-binding motif (TAZ), whereas ASAP3 overexpression in AGS cells enhanced their phosphorylation. Finally, ASAP3 overexpression could inhibit RhoA activity and thus suppress YAP/TAZ activation by interfering with the ASAP1-ARHGAP12 interaction.

CONCLUSION

ASAP3 may inhibit GC oncogenesis/progression by disrupting ASAP1-ARHGAP12 to suppress RhoA/YAP/TAZ, serving as a potential GC therapeutic target.

Key Words: Gastric cancer; ASAP3; RhoA activity; ASAP1-ARHGAP12; Yes-associated protein/transcriptional coactivator with PDZ-binding motif

Core Tip: Our study identified ASAP3 as a novel tumor suppressor in gastric cancer (GC) that may suppress RhoA activity by disrupting the ASAP1-ARHGAP12 interaction, thereby activating the Hippo pathway and suppressing yes-associated protein (YAP)/transcriptional coactivator with PDZ-binding motif (TAZ)-mediated oncogenesis. These findings clarify the context-dependent function of ASAP3, uncover a new regulatory mechanism of the RhoA/YAP/TAZ axis, and provide a potential prognostic biomarker and therapeutic target for GC. Future studies will focus on validating ASAP3’s prognostic value in clinical cohorts and developing small-molecule modulators of the ASAP3 interfered ASAP1-ARHGAP12 interaction for GC therapy.

INTRODUCTION

Gastric cancer (GC) ranks among the malignant tumors with persistently high incidence and mortality rates worldwide, imposing an especially severe disease burden in East Asia[1]. According to the latest epidemiological data, although measures such as Helicobacter pylori eradication and dietary structure improvement have led to a gradual decline in GC incidence in some regions, the 5-year survival rate of patients with advanced GC remains below 30%[2,3]. Tumor invasion, metastasis, and therapeutic resistance continue to be the core bottlenecks contributing to clinical treatment failure[4]. Currently, clinical management of GC encompasses surgical resection, chemotherapy (e.g., the fluorouracil-leucovorin-oxaliplatin-docetaxel regimen), human epidermal growth factor receptor 2-targeted therapy, and immune checkpoint inhibitors[5,6]. However, due to the complex pathogenesis and high heterogeneity of GC, existing treatment strategies exhibit limited efficacy in a subset of patients, particularly those with advanced disease[7]. Therefore, in-depth elucidation of the molecular mechanisms underlying GC initiation and progression, as well as the identification of novel diagnostic biomarkers and therapeutic targets, represent critical challenges in the current field of GC research.

The occurrence and development of tumors are accompanied by extensive reprogramming of gene expression[8]. Members of the adenosine diphosphate (ADP)-ribosylation factor GTPase-activating protein (ArfGAP) protein family play important roles in regulating membrane transport, cytoskeleton recombination, and cell signal transduction[9,10]. ArfGAP with SH3 domain, Ankyrin repeat, and PH domain (ASAP) family proteins, as key regulators of ArfGAP activity, participate in biological processes such as intracellular membrane trafficking, signal transduction, and cytoskeletal homeostasis by modulating the guanosine diphosphate/guanosine triphosphate (GDP/GTP) binding status of Arf proteins[9,11]. Among them, ASAP1 has been confirmed to active IQ motif containing GTPase activating protein 1/cell division cycle 42 homolog (IQGAP1/CDC42) pathway to promote the progression of GC and chemotherapy resistance[12]. Additionally, the interaction between ASAP1 and RhoA can regulate the yes-associated protein/transcriptional coactivator with PDZ-binding motif (YAP/TAZ) pathway and promotes tumor progression[13,14]. Furthermore, circ ASAP2 is reported to be highly expression in GC, and knockdown of circ ASAP2 can inhibit development of GC cells via the miR-770-5p/CDK6 axis, thereby suppressing GC tumor growth[15]. For ASAP3, it is also known as ACAP4, and is the GTPase activating protein of ADP ribosylation factor 6. There is little or no expression of ASAP3 in normal epithelial cells, but it is reported that ASAP3 is significantly increased in a variety of human cancers (including lung cancer, colon cancer and breast cancer)[16-18], which may lead to poor clinical outcomes in cancers. These effects may be attributed to the role of ASAP3 in regulating cell migration and subsequently regulating cancer cell invasion. Su et al[19] showed that ASAP3 was overexpressed in glioblastomas, and glioma patients with high expression of ASAP3 messenger RNA (mRNA) had poorer overall survival and progression free survival. However, the expression profile, biological function, and molecular mechanism of ASAP3 in GC remain unclear.

In the signaling pathways of tumor progression, the Hippo signaling pathway is a conserved pathway discovered in recent years that plays a central role in organ size regulation and tumorigenesis[20]. Its downstream effector molecules, YAP/TAZ, upon receiving upstream signals, can translocate into the nucleus and activate a series of target genes (such as CCN1, CCN2, etc.) that promote cell proliferation and inhibit apoptosis[21]. Aberrant inactivation of this pathway is closely associated with the progression of various tumors. Notably, RhoA, as an important member of the small GTPase family, regulates cytoskeletal reorganization and the expression of adhesion molecules, directly participating in the control of tumor cell motility[22,23]. RhoA is highly expressed in GC tissues, and its expression level is significantly correlated with tumor stage, lymph node metastasis, and poor patient prognosis[24]. ARHGAP12, a member of the RhoGAP protein family, can catalyze the hydrolysis of GTP bound to RhoA, thereby inhibiting its activity, forming a “RhoA-ARHGAP12-YAP/TAZ” regulatory axis[25]. Abnormal activation of this pathway has been confirmed as a key driver in the progression of poorly cohesive GC (a subtype with an extremely poor prognosis)[26], suggesting that targeting the RhoA/YAP/TAZ axis may represent a potential direction for GC therapy. Notably, ASAP3 exhibits a key structural difference from ASAP1/ASAP12: It lacks the SH3 domain at its C-terminus[27]. Based on this unique structural feature, we proposed the following scientific hypothesis: ASAP3 may drive the malignant progression of GC by regulating the GTPase activity of RhoA, which in turn affects the downstream Hippo/YAP signaling pathway.

To validate this hypothesis, we systematically evaluated the effects of ASAP3 on the biological behaviors (including cell proliferation, apoptosis, cell cycle progression, migration, and invasion) of GC cells and tumor growth through in vitro and in vivo functional experiments using the construction of stable cell lines with ASAP3 knockdown or overexpression, as well as subcutaneous xenograft experiments in NSG mice. Furthermore, we combined transcriptomic and proteomic analyses with molecular experiments to explore the potential molecular mechanism of ASAP3 in GC progression. Our findings are expected to provide novel theoretical insights into the pathogenesis of GC and identify potential molecular targets for the precise diagnosis and targeted therapy of GC.

MATERIALS AND METHODS

Cell culture and construction of stable transfection cell lines

Human GC cell lines AGS and HGC-27 were acquired from Obio Technology. Among them, AGS cells were cultured in Ham’s F-12K (Kaighn’s modification) medium (Gibco, Grand Island, NY, United States) supplemented with 10% fetal calf serum (FBS) (Gibco) and 1% penicillin/streptomycin (Gibco); as well as HGC-27 cells were maintained in RPMI-1640 medium (Gibco) containing 10% FBS (Gibco) and 1% penicillin/streptomycin (Gibco). Both the cells were cultured in an incubator with 5% carbon dioxide at 37 °C. Upon the cells were reaching 80%-90% confluence, the AGS and HGC-27 cells were passaged.

The pcSLenti-CMV-MCS-3xFLAG-PGK-Puro-WPRE3 vector [over-expression oe-negative control (NC)] and pCLenti-U6-shRNA(NC)-CMV-Puro-WPRE vector [short hairpin (sh)-NC] were employed to construct ASAP3 overexpression (oe-ASAP3) in AGC cells and ASAP3 knockdown (sh-ASAP3) in HGC-27 cells, respectively. The methods of lentivirus packaging sh-ASAP3 (pCLenti-U6-ASAP3 shRNA2-CMV-Puro-WPRE) and oe-ASAP3 (pcSLenti-CMV-ASAP3-3xFLAG-PGK-Puro-WPRE3) vectors were described as previously[28]. Briefly, 100 μg recombinant sh-ASAP3 and oe-ASAP3 vectors were transfected into 293T cells using Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, MA, United States) based on the manufacturer’s protocols. Following 48 hours of transfection, the viral supernatant was gathered and centrifuged at 1000 rpm for a 5-minute duration. Thereafter, the supernatant was filtered into a new tube using a 0.45-μm polyvinylidene difluoride (PVDF) membrane, and a second centrifugation step was performed at 50000 g for 2 hours at 4 °C. The pellet obtained was resuspended with 200 μL phosphate-buffered saline (PBS), and stored at -80 °C for long-term preservation. Then, AGS or HGC-27 cells were seeded in a 6-well plate at a density of 5 × 10⁵ cells/well and cultured overnight. The following day, the medium was replaced with serum-free medium, and the cells were infected with 50 μL of the packaged lentiviral supernatant (oe-NC, oe-ASAP3, sh-NC and sh-ASAP3). Forty-eight hours post-infection, the medium was discarded, and medium supplemented with 2 μg/mL puromycin was added to select resistant cells. The oe-ASAP3-AGS and sh-ASAP3-HGC-27 stably transfected cell lines were isolated, expanded in culture, and used for subsequent experiments. Transfection efficiency was evaluated by detecting ASAP3 expression via real-time quantitative polymerase chain reaction (RT-qPCR), and the sequences of ASAP3 are provided in Table 1.

Table 1 Sequences of all primers.

Primer	Sequence (5’-3’)
CCN1	F: ATAGTATCAAGGACCCCA
	R: CTCCATTCCAAAAACAGG
CCN2	F: TCTGTCACTTCGGCTCCCC
	R: CTCGTCACACACCCACTCC
AMOTL2	F: AGGTGCCCCAGGTGGTAGC
	R: GGAATGTCGGATGAGAGTG
ASAP3	F: GAGTGTGAGGCGTGGGTG
	R: CTTTGTGAGGTCGTGCGG
GAPDH	F: GGAGCGAGATCCCTCCAAAAT
	R: GGCTGTTGTCATACTTCTCATGG

F: Forward; R: Reverse.

Cell viability, apoptosis and cycle assays

The viability of cells with different treatments was determined using cell counting kit-8 (CCK-8) (Beyotime Biotechnology, Shanghai, China) in line with the manufacturer’s instructions. Briefly, oe-NC-AGS, oe-ASAP3-AGS, sh-NC-HGC-27 and sh-ASAP3-HGC-27 cells were expanded in culture. Once the cells reached the logarithmic growth phase, their concentrations were adjusted to 0.8 × 10⁴ cells/mL, and they were seeded into 96-well plates at a volume of 100 μL/well, followed by incubation at 37 °C in a 5% carbon dioxide incubator for 24 hours. Then, the supernatant in each well was aspirated and replaced with 200 μL of fresh medium. After cultured for 24 hours, 48 hours and 72 hours, 10 μL of CCK-8 reagent was added to each well, and the plates were incubated in the dark for 1 hour. Finally, the absorbance at 450 nm was quantified using a microplate reader.

Flow cytometry was employed to evaluate the cell apoptosis and cycle. For cell apoptosis, the Annexin V-fluorescein isothiocyanate (FITC) apoptosis assay kit (Beyotime Biotechnology) was used. In brief, cells were resuspended in 195 μL of Annexin V-FITC binding buffer, followed by the addition of 5 μL of Annexin V-FITC and 10 μL of propidium iodide (PI). After 20 minutes of dark incubation, flow cytometry was performed to acquire cell data, and total apoptosis rate was computed. For cell cycle, cells were resuspended in 200 μL of PBS, and 4 mL of pre-cooled 70% ethyl alcohol was added. After fixation at 4 °C overnight, the cell suspension was centrifuged at 1000 rpm for 3 minutes and washed with PBS. Following centrifugation, 500 μL PBS containing RNase (100 μg/mL) was added, and the mixture was incubated at 37 °C for 30 minutes. PI was then added to reach a final concentration of 50 μg/mL, with dark staining for 30 minutes. After washing and centrifuged at 1000 rpm for 3 minutes, cells were resuspended with 500 μL PBS, and cell cycle analysis was conducted using a flow cytometer within 24 hours.

Cell migration and invasion assays

Cell migration and invasion capabilities were assessed using Transwell chambers (8 μm pore size; Guangzhou Jet Bio-Filtration Co., Ltd., Guangzhou, Guangdong Province, China). For the cell invasion assay, the Transwell chambers were pre-coated with Matrigel prior to use. The different cell populations were harvested and then seeded into the upper compartments of the Transwell chambers, while the lower chambers contained medium supplemented with 10% FBS. Following a 48-hour incubation period, the cells were fixed with 4% paraformaldehyde for 20 minutes. After washing with PBS, crystal violet was applied to the cells and allowed to incubate for 20 minutes. Excess dye was removed, and images were captured under a microscope to analyze the relative number of cells.

Cell scratch and clone formation assays

For cell scratch, horizontal lines (1 cm in length, with no fewer than five lines per well) were pre-drawn on the back of 6-well plates using a marker pen. The plates were then seeded with cells at a density of 5 × 10⁵ cells per well. After cultured overnight, the cells in each well were scratched with a pipette tip, with the scratches made perpendicular to the pre-drawn horizontal lines. The cell culture medium was discarded and replaced with serum-free medium, and the cells were cultured for another 48 hours. Images of the cells were captured at the 0-hour and 48-hour time points.

Subsequently, the colony-forming ability of various cells was assessed. The oe-NC-AGS, oe-ASAP3-AGS, sh-NC-HGC-27 and sh-ASAP3-HGC-27 cells were seeded into 6-well plates and incubated in a 5% carbon dioxide incubator for 24 hours. The medium was changed every 2 days, and once distinct cell colonies had formed, the supernatant was discarded. After washing twice with PBS, the cells were fixed with 4% paraformaldehyde at room temperature for 15 minutes. Following another two washes with PBS, the cells were stained with 0.5% crystal violet for 10 minutes, and then observed and imaged under a microscope.

Transcriptome sequencing and analysis

Total RNA was extracted from oe-NC-AGS/oe-ASAP3-AGS, and sh-NC-HGC-27/sh-ASAP3-HGC-27 cells using Trizol reagent (Thermo, United States). The integrity of the isolated RNA was evaluated with an Agilent 2100 bioanalyzer (Agilent Technologies, United States), and the concentrations of the isolated RNA were accurately determined using a NanoDrop spectrophotometer (Thermo Scientific, United States). For library construction, mRNA was isolated from total RNA using oligo-dT magnetic beads, and the captured mRNA was subjected to fragmentation. First-strand and second-strand complementary DNAs (cDNAs) were then synthesized using reverse transcriptase. The reverse transcription products underwent end repair, followed by the addition of an A base to the 3’ end. Subsequently, these fragments were ligated with sequencing adapters. After purifying the ligation products to remove incompletely ligated products and empty adapter self-ligated products, PCR amplification was performed using primers complementary to the adapter sequences. Finally, the sequencing library was obtained through magnetic bead purification. After library construction, the library concentration was measured using Qubit, and the library fragment length was detected using an Agilent Fragment Analyzer to ensure library quality. Once the library passed quality inspection, paired-end 150bp sequencing was performed on the Illumina Novaseq 6000 platform.

Following sequencing, raw reads underwent quality control using the fastp software, with clean reads obtained by filtering out low-quality data and joint contamination. Subsequently, HISAT2 software was utilized to map and annotate mRNAs, and then the gene expression [fragments per kilobase of transcript per million mapped reads (FPKM) value] was determined using featureCounts software. Thereafter, the DESeq2 algorithm was used to identify differentially expressed mRNAs (DEMs) with the thresholds of |log₂fold change (FC)| > 1 and adjusted P value < 0.05. Additionally, the identified DEMs were submitted for Gene Ontology (GO) term enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. The criterion for classifying GO terms and KEGG pathways as significantly enriched was established as P < 0.05.

Protein extraction and proteomics analysis

The different cells were added with cracking buffer containing 1% Triton X-100, 1% protease inhibitor, and 1% phosphatase inhibitor, and cracked by ultrasound (80 W) on ice for 2 minutes. After centrifuged at 12000 rpm for 10 minutes at 4 °C, the supernatant was transferred to a new tube, which was the total protein solution. Then, the concentrations of the extracted total protein were measured using a bicinchoninic acid (BCA) assay kit (Beyotime Biotechnology). Afterwards, 10 μg protein samples were separated by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (180 V, 45 minutes) and examined by Coomassie Blue Staining using the EStain LG protein staining instrument. Based on the determined protein concentrations, the protein concentration in each sample was adjusted to the same with the cracking buffer, and the final concentration of 5 mmol/L dithiothreitol was added. After incubated at 55 °C for 30 minutes, the samples were cooled on ice until they reached room temperature, and then iodoacetamide was added to make the final concentration of 10 mmol/L. After incubated in the dark for 15 minutes, six times the volume of acetone was added to precipitate the protein, which was left at -20 °C overnight. After centrifuged at 8000 g for 10 minutes, the precipitates were resuspended with 100 μL of 50 mmol/L ammonium bicarbonate, and then trypsin (1 mg/mL) was added at a ratio of 1:50 (protease: Protein, m/m). After enzymolysis at 37 °C overnight, the peptides were desalted by the C18 Cartridge. C18 Cartridge was used to desalinate the peptide segments. After freeze-drying, 40 μL of 0.1% formic acid solution was added for reconstitution, and the peptide segments were quantified (optical density = 280 nm) for further liquid chromatography-tandem mass spectrometry (LC-MS/MS).

For LC-MS/MS, the preprocessed peptides were dissolved in mobile phase A and then separated using a NanoElute ultra-high performance liquid chromatography (UHPLC) system. Mobile phase A consisted of an aqueous solution containing 0.1% formic acid, while mobile phase B was acetonitrile solution with 0.1% formic acid. The chromatographic column was equilibrated with 95% of mobile phase A. The sample was separated on a C18 reversed-phase analytical column (Thermo Scientific EASY column, 25 cm, ID 75 μm, 1.6 μm) at a flow rate of 300 nL/minute. The liquid phase gradient was programmed as follows: 0-75 minutes, 2%-22% B; 75-80 minutes, 22%-37% B; 80-85 minutes, 37%-80% B; 85-90 minutes, 80% B. The peptides separated via UHPLC system were injected into a capillary ion source for ionization, followed by analysis using a timsTOF Pro mass spectrometer. The detection method was positive ion, with the ion source voltage set at 1.5 kV. The secondary mass spectrometry (MS/MS) scan range was configured to 100-1700 m/z, ion drift time was 0.6-1.6 Vs/cm², and collision energy range was 20-59 eV. Data acquisition was performed using the parallel cumulative serial fragmentation (PASEF) mode. After primary mass spectrometry data collection, the PASEF mode was used to acquire 10 secondary spectra for parent ions with charge numbers ranging from 0 to 5. The dynamic exclusion time for the tandem mass spectrometry scan was set to 24 seconds to prevent the repeated scanning of the parent ion.

The original LC-MS/MS files were imported into MaxQuant (version 1.6.17.0) for database search with the uniprot-Homo sapiens-9606-2023.2.1.fasta. For the acquisition of high-quality analytical outcomes, we implemented rigorous filters: 1% 1% false discovery rate at spectrum, peptide, and protein levels; and a requirement for at least one unique peptide per identified protein. Then, the intensity value of each protein in different samples was obtained, and differential expressed proteins (DEPs) were screened based on the criteria of P < 0.05 and FC ≥ 2 or FC ≤ 0.5. Thereafter, the screened DEPs were subjected to GO terms and KEGG pathways analyses. The filtered DEPs were matched against the STRING database (https://string-db.org/), and differential protein interaction pairs were retrieved with a confidence score exceeding 0.7 (high confidence). Subsequently, the protein-protein interaction network was constructed and visualized using the visNetwork tool within the R package.

Animal experiments

Prior to animal experiments, the oe-NC-AGS, oe-ASAP3-AGS, sh-NC-HGC-27, and sh-ASAP3-HGC-27 cells in the logarithmic growth phase were collected. The culture medium was removed, and the cell concentration of each group was adjusted to a final density of 5 × 10⁷ cells/mL using PBS. A total of 24 specific pathogen free NSG female mice aged 5-6 weeks were purchased from Shanghai Model Organisms Center, Inc. (Shanghai, China). All the mice were housed under controlled environmental conditions: Temperature at 24 ± 2 °C, humidity at 50% ± 10%, and a 10 hours: 14 hours light/dark cycle. Throughout the experiment, the mice had free access to food and water. After 7 days of acclimatization, the mice were randomly divided into four groups (n = 6 for each group): Oe-NC, oe-ASAP3, sh-NC, and sh-ASAP3. The mice in the oe-NC, oe-ASAP3, sh-NC, and sh-ASAP3 groups were respectively subcutaneously injected with 0.2 mL oe-NC-AGS, oe-ASAP3-AGS, sh-NC-HGC-27, and sh-ASAP3-HGC-27 cells (1 × 10⁷ cells) on the right axilla of the mice (a single injection). After feeding for 4 weeks, tumor volume was calculated weekly based on measurements of tumor length and short diameter. At 4 weeks, mice were euthanized. For tissue collection, mice (with an average body weight of 18-23 g at this stage) were first deeply anesthetized via intraperitoneal injection of sodium pentobarbital (80 mg/kg body weight). Once unresponsive to toe-pinch, euthanasia was confirmed by cervical dislocation. After which tumors were stripped, weighed, and imaged. All animal studies were performed in line with the Animal Research: Reporting in vivo Experiments guidelines 2.0, and were approved by the Institutional Animal Care and Use Committee at Obio Technology (Shanghai) Corp., Ltd. (approval No. OBIO-AUF-182).

The harvested tumor tissues were fixed in 4% paraformaldehyde, and dehydrated using gradient concentrations of ethanol. Following paraffin embedding, the tissues were sectioned into 5 μm-thick slices, and stained with hematoxylin and eosin. An optical microscope (Carl Zeiss Microscopy, Jena, Germany) was used to observe and capture the stained tissue images. In addition, the tumor sections were baked at 62 °C for 1 hour, followed by dewaxing, rehydration, antigen retrieval, and blocking of endogenous peroxides at 37 °C for 30 minutes. Next, the sections were incubated with anti-ASAP3 antibody (1:1000, Santa Cruz Biotechnology, United States) at 4 °C overnight. After washing, the sections were incubated with Cy3-labeled secondary antibody (1:1000, Jackson ImmunoResearch, PA, United States) at 37 °C in the dark for 30 minutes. Following PBS with Tween-20 washing, the sections were stained with 4’,6-diamidino-2-phenylindole dye liquor in the dark for 5 minutes, and then sealed with antifade mounting medium for 10 minutes. Finally, an inverted fluorescent microscope was employed to observe and acquire the cell images, and Image J software (National Institutes of Health) was used to analyze the fluorescence intensity.

RT-qPCR

Total RNA was extracted from tissue samples or cell suspension using the Trizol reagent (Thermo, United States), and was reverse-transcribed into cDNA with the PrimeScript™ II 1^st Strand cDNA Synthesis Kit (Takara, Beijing, China), in accordance with the kit’s protocols. RT-qPCR was performed with 1:5-diluted cDNA on the Applied Biosystems QuantStudio 6 Flex Real-Time PCR System (Thermo Fisher Scientific), using SYBR^® Premix Ex Taq^TM (Takara Bio) for the reaction. All primers, whose sequences were synthesized and provided by Sangon Biotech (Shanghai, China), are shown in Table 1. Relative mRNA expression levels of the related genes were determined using the 2^-ΔΔCt method, with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as the reference gene.

Western blot

Total protein was isolated from cell suspension using the radio immunoprecipitation assay (RIPA) lysis buffer, and was quantified using a BCA assay kit (Beyotime). Subsequently, the protein samples (20 μg) were separated by 8% or 12% SDS-PAGE, and transferred to PVDF membranes. After blocked with 5% skim milk for 2 hours, the membranes were incubated with primary antibodies at 4 °C overnight, followed by horseradish peroxidase-goat anti-rabbit or mouse secondary antibodies (1:5000, cat. no. ZB2301 or ZB2305, ZSGB-BIO, Beijing, China) for 1 hour. After developed by an enhanced chemiluminescence (ECL) assay kit (Thermo Fisher Scientific), the protein bands were visualized utilizing the integrated chemiluminescence imaging instrument (ChemiScope 5300 Pro, United States), while Image J software (National Institutes of Health) was employed to quantify the protein bands, with GAPDH serving as the internal reference for normalization. The primary antibodies included ASAP1 (1:1000, cat. no. sc-81896, Santa Cruz, United States), MST1/MST2 (1:1000, cat. no. PA5-36100, Invitrogen, United States), phosphorylated (p)-MST1 (Thr183)/MST2 (Thr180) (1:2000, cat. no. 80093-1-RR, Proteintech), LATS1/LATS2 (1:1000, cat. no. PA5-115498, Invitrogen), p-LATS1 (Thr1079)/LATS2 (Thr1041) (1:500, cat. no. bs-7913R, Bioss, Wuhan, Hubei Province, China), YAP/TAZ (1:1000, cat. no. 93622T, Cell Signaling Technology, United States), p-YAP (Ser127)/TAZ (Ser89) (1:1000, cat. no. 13008T, Cell Signaling Technology), ARHGAP12 (1:2000, cat. no. MG939263, Abmart, Shanghai, China), and GAPDH (1:50000, cat. no. 60004-1-Ig, Proteintech).

Co-immunoprecipitation/western blot

Co-immunoprecipitation (co-IP) combined with western blot was performed to detect the interaction between ASAP1 and ARHGAP12 in sh-ASAP3 and oe-ASAP3 cells. Briefly, cells at 80%-90% confluence were rinsed twice with pre-cooled PBS, and after centrifuged, the sediments were resuspended with IP lysis buffer. Then, 4 μL phenylmethylsulfonyl fluoride (250 mmol, 1:250) and 10 μL protease inhibitor cocktail (100 mmol, 1:100) were added, and after votex three times, cell lysates were centrifuged at 12000 rpm for 10 minutes at 4 °C, and the supernatants (total protein extracts) were collected for co-IP. For preclearing to reduce non-specific binding, protein extracts (50 μL) were incubated with 50 μL protein A/G agarose beads (pre-equilibrated with PBS) for 1 hour at 4 °C with gentle rotation. After centrifugation, the supernatants were transferred to new tubes and incubated overnight at 4 °C with 2 μg of IP-grade anti-ASAP1 antibody; parallel control reactions were set up with isotype-matched immunoglobulin G to exclude non-specific antibody binding. The next day, 50 μL protein A/G agarose beads were added to each tube, followed by incubation for 4 hours at 4 °C to capture antibody-protein complexes. The beads were then washed five times with pre-chilled RIPA buffer and once with PBS to remove unbound proteins. Immunoprecipitated complexes were eluted by boiling the beads in 2 × SDS loading buffer for 10 minutes, and the eluates were subjected to SDS-PAGE electrophoresis. Proteins were transferred to PVDF membranes, blocked with 5% non-fat milk, and probed with anti-ARHGAP12 antibody to detect interacting proteins, with anti-ASAP1 antibody used to confirm successful immunoprecipitation. Total protein extracts without IP (input control) were included to verify the expression of ASAP1 and ARHGAP12 in all cell groups. Chemiluminescent signals were detected using an ECL system, and band intensities were analyzed to compare interaction strength across sh-ASAP3 and oe-ASAP3 cells.

RhoA activation pull down/western blot

To determine RhoA activity in sh-ASAP3-HGC-27 and oe-ASAP3-AGS cells, RhoA activation pull down/western blot was performed using a commercial RhoA Activity Assay Kit (cat. no. 80601, Wuhan NewEast Biosciences Co. Ltd., Wuhan, Hubei Province, China) following the manufacturer’s protocols. Briefly, cells were cultured in 6-well plates until reaching 80%-90% confluence. After removing the culture medium, cells were rinsed twice with pre-chilled PBS and lysed on ice for 15 minutes using the kit-supplied lysis buffer supplemented with protease inhibitor cocktail (to prevent protein degradation) and phosphatase inhibitors (to preserve phosphorylation status). Cell lysates were centrifuged at 12000 g for 10 minutes at 4 °C, and the supernatants (total protein extracts) were collected for pull down/western blot. A portion of the extracts (50 μL) was mixed with 2 × SDS loading buffer, boiled for 10 minutes, and stored as the “total RhoA” sample for subsequent western blot analysis. The remaining extracts were quantified using a BCA assay kit to ensure equal protein input (300 μg/sample) across groups. For pull down of active RhoA (GTP-bound RhoA), pre-equilibrated Rhotekin Rho-GTP binding domain (RBD)-agarose beads (50 μL) were added to the normalized protein extracts. The mixture was incubated at 4 °C with gentle rotation for 1 hour to allow specific binding between Rhotekin RBD (a RhoA effector domain) and active RhoA. After incubation, the beads were centrifuged at 3000 g for 30 seconds at 4 °C, and the supernatant was discarded. The beads were then washed three times with pre-chilled wash buffer to remove unbound proteins, and the bound complexes were eluted by resuspending the beads in 2 × SDS loading buffer and boiling for 10 minutes. The eluted samples (containing active RhoA) and the pre-stored “total RhoA” samples were subjected to western blot with the anti-RhoA antibody (1:1000, Cell Signaling Technology). The intensity of active RhoA and total RhoA bands was quantified using ImageJ software, and the ratio of active RhoA to total RhoA was calculated to compare RhoA activity in sh-ASAP3-HGC-27 and oe-ASAP3-AGS cells.

Bioinformatics analysis of ASAP3/ASAP1/ARHGAP12

The correlation of ASAP3/ASAP1/ARHGAP12 expression levels with clinical data and survival outcomes in GC was performed. First, GC RNA-sequencing gene expression data (FPKM/transcripts per million normalized) and corresponding clinical information [including overall survival time, survival status, age, gender, tumor node metastasis (TNM) stage, histological grade, etc.] were downloaded from The Cancer Genome Atlas (TCGA) database via the Genomic Data Commons Data Portal or R/Bioconductor packages (e.g., TCGA biolinks), followed by extraction of the ASAP3/ASAP1/ARHGAP12 expression values and key clinical variables. If necessary, gene expression values were subjected to log₂ transformation to approximate normal distribution, categorical clinical variables were appropriately coded, and survival time was calculated in months from diagnosis to last follow-up or death. Then, Spearman’s rank correlation (for non-normally distributed data) or Pearson correlation (for normally distributed data) was used to assess associations between continuous gene expression levels and continuous clinical variables (e.g., age), while non-parametric tests (Mann-Whitney U test for two groups, Kruskal-Wallis test for multiple groups) were applied to compare gene expression differences across categorical clinical variable groups, with false discovery rate or Bonferroni correction used to adjust P values for multiple comparisons. For survival analysis, patients were stratified into high- and low-expression groups for each gene based on median expression, Kaplan-Meier survival curves were generated, differences in survival between groups were assessed by log-rank test. All analyses were implemented using R (version 4.x or later) with packages including survival, survminer, ggplot2, stats, and corrplot. Ethical compliance was ensured as TCGA data are publicly available and de-identified, requiring no additional institutional review board approval and adhering to TCGA publication guidelines.

Statistical analyses

Each experiment was conducted in triplicate, and all results were expressed as mean ± SD. Statistical analyses were performed using SPSS 22.0 software (SPSS Inc., Chicago, IL, United States), while GraphPad Prism 8.0 (GraphPad, CA, United States) was used for figure generation. Student’s t-test was applied for comparisons between two groups, whereas one-way analysis of variance with subsequent Tukey’s post hoc test was used to compare three or more groups. A P value < 0.05 was considered to indicate a statistically significant difference.

RESULTS

ASAP3 expression in GC, and cell transfection efficiency

To investigate the basal expression profile of ASAP3 in GC, the relative expression level of ASAP3 was first detected in two widely used GC cell lines (AGS and HGC-27). As shown in Figure 1A, the expression of ASAP3 in HGC-27 cells was significantly higher than that in AGS cells (P < 0.05), implying a potential association between ASAP3 expression and the biological characteristics of GC cells.

Figure 1 ASAP3 expression in gastric cancer, and cell transfection efficiency. A: Relative expression of ASAP3 in gastric cancer (GC) cell lines (AGS and HGC-27 cells); B: Cell transfection efficiency assessed by determining the messenger RNA (mRNA) expression of ASAP3 in AGS transfected with over-expression-ASAP3 plasmids; C: Cell transfection efficiency assessed by determining the mRNA expression of ASAP3 in HGC-27 transfected with short hairpin-ASAP3 plasmids; D: Proliferation of GC cells (HGC-27 and AGS) after transfection using cell counting kit-8 after cultured for 24 hours, 48 hours, and 72 hours. ^aP < 0.05 vs AGS cell. ^bP < 0.05 vs HGC-27 cell. ^cP < 0.05 vs short hairpin-negative control-HGC-27 cell. ^dP < 0.05 vs over-expression-negative control-AGS cell. mRNA: Messenger RNA; oe: Over-expression; sh: Short hairpin; NC: Negative control.

To explore the functional roles of ASAP3 in GC in vitro, ASAP3-overexpressing and ASAP3-silenced cell models were established. It was found that no significant differences in the ASAP3 expression between the AGS and oe-NC-AGS groups, as well as between the HGC-27 and sh-NC-HGC-27 groups (P > 0.05, Figure 1B and C). After transfected with oe-ASAP3 plasmids, the ASAP3 expression was about 220 times higher than that in the control AGS cells, which showed ASAP3 expression in the oe-ASAP3-AGS group was significantly elevated compared with that in the parental AGS group (P < 0.05, Figure 1B). Furthermore, the mRNA expression of ASAP3 in the sh-ASAP3-HGC-27 group was remarkably reduced relative to the parental HGC-27 group and the sh-NC-HGC-27 group (P < 0.05, Figure 1C). These results indicated the ASAP3-overexpressing AGS and ASAP3-silenced HGC-27 cells were successfully constructed, and could be used for subsequent experiments.

Effects of ASAP3 on the growth of GC cells

The effects of ASAP3 on the growth of GC cells were then evaluated. Firstly, CCK-8 was performed to assess the proliferation activity of GC cells after transfection. As shown in Figure 1D, in HGC-27 cells, the proliferation activity of ASAP3-silenced HGC-27 cells were significantly higher than that of the sh-NC cells at 24 hours, 48 hours, and 72 hours (P < 0.05). In contrast, in AGS cells, the proliferation activity of the ASAP3-overexpressiong AGS cells was significantly lower than that of the oe-NC AGS cells at the same time points (P < 0.05). These results preliminarily indicated that ASAP3 may exert an inhibitory effect on the proliferation of GC cells.

For cell cycle, compared with the sh-NC HGC-27 cells, silencing ASAP3 in HGC-27 cells significantly increased the proportions of cells in the S phase (P < 0.05); whereas in AGS cells, overexpressing ASAP3 evidently reduced the proportions of cells in the S phase (P < 0.05, Figure 2A). Flow cytometry was also used to detect cell apoptosis. The results (Figure 2B) showed that the apoptosis rate of HGC-27 cells after silencing ASAP3 was significantly lower than that in the sh-NC-HGC-27 group (P < 0.05). Conversely, the apoptosis rate of AGS cells after overexpressing ASAP3 was significantly higher than that in the oe-NC-AGS group (P < 0.05, Figure 2B). Collectively, these findings demonstrated that ASAP3 could arrest the cell cycle and promote apoptosis of GC cells.

Figure 2 Effects of ASAP3 on the growth of gastric cancer cells. A: Cell cycle of HGC-27 and AGS cells after transfection with short hairpin (sh)-ASAP3 or over-expression (oe)-ASAP3 plasmids using a flow cytometry; B: Apoptosis of HGC-27 and AGS cells after transfection with sh-ASAP3 or oe-ASAP3 plasmids using a flow cytometry; C: Migration of HGC-27 and AGS cells after transfection with sh-ASAP3 or oe-ASAP3 plasmids by cell scratch assay; D: Migration of HGC-27 and AGS cells after transfection with sh-ASAP3 or oe-ASAP3 plasmids by Transwell assay; E: Invasion of HGC-27 and AGS cells after transfection with sh-ASAP3 or oe-ASAP3 plasmids by Transwell assay; F: The colony-forming ability of HGC-27 and AGS cells after transfection with sh-ASAP3 or oe-ASAP3 plasmids using clone formation assay. ^cP < 0.05 vs short hairpin-negative control-HGC-27 cell. ^dP < 0.05 vs over-expression-negative control-AGS cell. oe: Over-expression; sh: Short hairpin; NC: Negative control.

The scratch and Transwell assays were both conducted to assess cell migration. As shown in Figure 2C and D, after 24 hours of culture, the migration of HGC-27 cells after ASAP3 silencing was significantly enhanced compared with sh-NC HGC-27 cells (P < 0.05); whereas the migration of AGS cells after ASAP3 overexpression was markedly suppressed in comparison with the oe-NC AGS cells (P < 0.05). For cell invasion, the number of the invaded cells was significantly increased in the sh-ASAP3-HGC-27 group compared to the sh-NC-HGC-27 group (P < 0.05), but was remarkedly declined in the oe-ASAP3-AGS group relative to the oe-NC-AGS group (P < 0.05, Figure 2E). Finally, the clone formation assay was performed to evaluate the clonogenic potential of GC cells. As shown in Figure 2F, the number of colonies formed by HGC-27 cells in the sh-ASAP3-HGC-27 group was significantly greater than that in the sh-NC-HGC-27 group (P < 0.05). On the other hand, the number of colonies formed by AGS cells in the oe-ASAP3-AGS group was significantly fewer than that in the oe-NC-AGS group (P < 0.05). These findings suggested that ASAP3 could inhibit the migration, invasion and in vitro colony-forming ability of GC cells.

Effects of ASAP3 on GC tumor growth in vivo

To validate the in vivo effect of ASAP3 on GC tumor growth, a nude mouse xenograft model of GC cells was established. Monitoring of mouse body weight (Figure 3A) revealed no significant differences between the sh-NC-HGC-27 and sh-ASAP3-HGC-27 groups, and between the oe-NC-AGS and oe-ASAP3-AGS groups during the experimental period, ruling out potential adverse effects of the experimental treatments on the overall health status of the mice. Mice injected with sh-ASAP3 HGC-27 cells exhibited increased tumor size and volume over time compared to the sh-NC HGC-27 cells, whereas mice injected with oe-ASAP3 AGS cells showed slightly suppressed tumor growth (Figure 3B and C). Then, hematoxylin-eosin staining results displayed that in the tumor tissues of different groups, there were cells with nuclear atypia, large nuclei, basophilic and tightly arranged distribution. The tumor tissues of the sh-NC-HGC-27 group contained a small number of mitotic phases, a small amount of punctate cell necrosis, and the nuclei were condensed, deeply stained, fragmented, or dissolved and disappear (Figure 3D). In contrast, tumor tissues in the sh-ASAP3-HGC-27 group had a small amount of mitotic phase, small scale necrosis, with nuclei shrinking, staining, fragmentation, or dissolution disappearing, and many irregular cavities visible (Figure 3D). For AGS cells-induced tumors, the oe-ASAP3-AGS group exhibited a small amount of nuclear fission phase, large scale necrosis, with nuclei shrinking, staining, fragmentation, or dissolution disappearing, and numerous irregular cavities visible compared with the oe-NC-AGS group (Figure 3D). These implied that ASAP3 could inhibit the proliferative activity of tumor tissues and promotes their necrosis.

Figure 3 Effects of ASAP3 on gastric cancer tumor growth in vivo. A: Body weight changes of mice after injected with HGC-27 or AGS cells after transfection with short hairpin (sh)-ASAP3 or over-expression (oe)-ASAP3 plasmids for different days; B: Tumor size of mice after injected with HGC-27 or AGS cells after transfection with sh-ASAP3 or oe-ASAP3 plasmids; C: Tumor volume changes of mice after injected with HGC-27 or AGS cells after transfection with sh-ASAP3 or oe-ASAP3 plasmids for different days; D: Hematoxylin and eosin staining of tumor tissues with different treatments. Purple arrow: Irregular cavity; Red arrow: Necrosis; Black arrow: Nuclear fission phase; E: ASAP3 positive cells in tumor tissues of mice after injected with HGC-27 cells after transfection with sh-ASAP3 plasmids using immunofluorescence; F: ASAP3 positive cells in tumor tissues of mice after injected with oe-ASAP3 plasmids using immunofluorescence. ^cP < 0.05 vs short hairpin-negative control-HGC-27 cell. ^dP < 0.05 vs over-expression-negative control-AGS cell. oe: Over-expression; sh: Short hairpin; NC: Negative control.

Finally, immunofluorescence staining was used to detect ASAP3-positive cells in tumor tissues of different groups. It was obvious that the number of ASAP3-positive cells in the sh-ASAP3-HGC-27 group was significantly lower than that in the sh-NC-HGC-27 group (P < 0.05, Figure 3E), while the number of ASAP3-positive cells in the oe-ASAP3-AGS group was significantly higher than that in the oe-NC-AGS group (P < 0.05, Figure 3F). This further validated the regulatory effect of ASAP3 expression in the in vivo model.

Underlying mechanisms of ASAP3 in GC cells using transcriptome sequencing

To uncover the underlying mechanisms, we performed transcriptome sequencing in HGC-27 and AGS cells with modulated ASAP3 expression. Principal component analysis (PCA) showed distinct differences in gene expression patterns between the sh-NC-HGC-27 and sh-ASAP3-HGC-27 groups, with good sample clustering, indicating that ASAP3 silencing significantly alters the transcriptomic characteristics of HGC-27 cells (Figure 4A). After differential analysis, 382 DEMs were identified between the sh-NC-HGC-27 and sh-ASAP3-HGC-27 cells, including 264 upregulated DEMs (IGFBP3, L1CAM, FOS) and 118 downregulated DEMs (DCHS1, RRN3, LRP2) (Figure 4B). Heatmap analysis further visualized the expression differences of these DEMs between the two groups, showing a clear grouping trend (Figure 4C). Then, the identified DEMs were submitted for GO and KEGG analyses. GO analysis revealed that the DEMs were mainly enriched in biological process (BP) such as “multicellular organismal process”, “biological regulation”, and “cell differentiation”; molecular function (MF) such as “protein binding”, “signaling receptor binding”, and “carbonyl reductase (nicotinamide adenine dinucleotide phosphate hydrogen) activity”; and cellular component (CC) such as “cell periphery”, “cell junction”, and “extracellular matrix” (Figure 4D). KEGG pathway enrichment analysis showed that the identified DEMs were closely associated with “relaxin signaling pathway”, “cyclic adenosine monophosphate signaling pathway”, “forkhead box O signaling pathway”, “complement and coagulation cascades”, “tumor necrosis factor (TNF) signaling pathway”, “estrogen signaling pathway”, and “Hippo signaling pathway” (Figure 4E).

Figure 4 Underlying mechanisms of ASAP3 in HGC-27 cells using transcriptome sequencing. A: The principal component analysis of the short hairpin (sh)-negative control (NC)-HGC-27 and sh-ASAP3-HGC-27 cells; B: The volcano plot of the differential expressed messenger RNAs (DEMs) between the sh-NC-HGC-27 and sh-ASAP3-HGC-27 cells; C: The heatmap of the identified DEMs; D: The Gene Ontology terms enriched by the identified DEMs; E: Kyoto Encyclopedia of Genes and Genomes pathway analysis of the identified DEMs. PCA: Principal component analysis; PC: Principal component; GO: Gene Ontology; NADPH: Nicotinamide adenine dinucleotide phosphate hydrogen; KEGG: Kyoto Encyclopedia of Genes and Genomes; BP: Biological process; MF: Molecular function; CC: Cellular component; cAMP: Cyclic adenosine monophosphate; AGE-RAGE: Advanced glycation end products-receptor for advanced glycation end products; TNF: Tumor necrosis factor; FoxO: Forkhead box O; sh: Short hairpin; NC: Negative control.

Transcriptome analysis of AGS cells in the oe-NC-AGS and oe-ASAP3-AGS groups showed that PCA results exhibited significant differences in gene expression patterns between the two groups, with good sample separation (Figure 5A). Based on |log₂FC| > 1 and adjusted P value < 0.05, 714 DEMs were identified, including 477 upregulated (LRP8, GATM, REG4) and 237 downregulated (LCN2, IFFO2, S100P) ones (Figure 5B). Heatmap analysis clearly displayed the expression differences of DEMs between the oe-NC-AGS and oe-ASAP3-AGS groups, with obvious grouping clustering (Figure 5C). GO functional enrichment analysis indicated that the DEMs were mainly enriched in BP such as “multicellular organismal process”, “cellular response to chemical stimulus”, and “regulation of biological quality”; and MF such as “chemokine activity”, “signaling receptor activator activity”, and “CXCR chemokine receptor binding”; as well as CC like “extracellular region”, “plasma membrane”, and “extracellular exosome” (Figure 5D). Additionally, KEGG showed the identified DEMs were involved in “cell cycle”, “mineral absorption”, “interleukin-17 signaling pathway”, “mitogen-activated protein kinase signaling pathway”, “hypoxia inducible factor-1 (HIF-1) signaling pathway”, “Rap1 signaling pathway”, “TNF signaling pathway”, “cytokine-cytokine receptor interaction”, and “Hippo signaling pathway” (Figure 5E).

Figure 5 Underlying mechanisms of ASAP3 in AGS cells using transcriptome sequencing. A: The principal component analysis of the over-expression (oe)-negative control (NC)-AGS and oe-ASAP3-AGS cells; B: The volcano plot of the differential expressed messenger RNAs (DEMs) between the oe-NC-AGS and oe-ASAP3-AGS cells; C: The heatmap of the identified DEMs; D: The Gene Ontology terms enriched by the identified DEMs; E: Kyoto Encyclopedia of Genes and Genomes pathway analysis of the identified DEMs. PCA: Principal component analysis; PC: Principal component; oe: Over-expression; GO: Gene Ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; BP: Biological process; MF: Molecular function; CC: Cellular component; TNF: Tumor necrosis factor; IL: Interleukin; MAPK: Mitogen-activated protein kinase; NC: Negative control.

Underlying mechanisms of ASAP3 in GC cells using proteomics

Proteomic analysis was conducted on HGC-27 cells in the sh-NC-HGC-27 and sh-ASAP3-HGC-27 groups. PCA results showed significant differences in protein expression patterns between the two groups, with good sample reproducibility (Figure 6A). Volcano plot analysis identified 98 DEPs, including 57 downregulated (CMPK1, UBA6, SFRS11) and 41 upregulated (S100A16, SNTB1, HMGA2) DEPs (Figure 6B). The heatmap clustering of the identified DEPs showed the identified DEPs could significantly distinguish the oe-NC-HGC-27 samples from the oe-ASAP3-HGC-27 samples (Figure 6C). Subcellular localization analysis showed that the DEPs were mainly distributed in the nucleus (49 proteins), cytoplasm (31 proteins), plasma membrane (19 proteins), mitochondria (18 proteins), and extracellular space (14 proteins) (Figure 6D), suggesting that ASAP3 may exert its functions by regulating protein functions in different subcellular compartments. Functional analysis of GO terms revealed that the identified DEPs were mainly enriched in BP such as “signal release”, “glycosylphosphatidylinositol (GPI) anchor metabolic process”, and “neurotransmitter secretion”; CC like “gap junction”, “endoplasmic reticulum lumen”, and “GPI-anchor transamidase complex”; as well as MF such as “CoA-ligase activity”, “calcium ion binding”, and “carbohydrate binding” (Figure 6E). KEGG pathway enrichment analysis showed that the DEPs were significantly enriched in pathways related to “HIF-1 signaling pathway”, “mineral absorption”, “fatty acid biosynthesis”, “glycolysis/gluconeogenesis”, “arginine and proline metabolism”, and “steroid biosynthesis” (Figure 6F).

Figure 6 Underlying mechanisms of ASAP3 in HGC-27 cells using proteomics. A: The principal component analysis of the short hairpin (sh)-negative control (NC)-HGC-27 and sh-ASAP3-HGC-27 cells; B: The volcano plot of the differential expressed proteins (DEPs) between the sh-NC-HGC-27 and sh-ASAP3-HGC-27 cells; C: The heatmap of the identified DEPs; D: Subcellular localization analysis of the DEPs; E: The Gene Ontology terms enriched by the identified DEPs, including biological process, cellular component, and molecular function; F: Kyoto Encyclopedia of Genes and Genomes pathway analysis of the identified DEPs. PC: Principal component; sh: Short hairpin; GO: Gene Ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; BP: Biological process; MF: Molecular function; CC: Cellular component; NC: Negative control.

Further, proteomic analysis of AGS cells in the oe-NC-AGS and oe-ASAP3-AGS groups found PCA results exhibited significant differences in protein expression patterns between the two groups, with good sample separation (Figure 7A). Then, a total of 66 DEPs including 28 upregulated DEPs (PAGE4, TYMS, TNFAIP2) and 38 downregulated DEPs (NEFF4, PHLDB2, DKFZp686P11128) were screened (Figure 7B). The heatmap displayed the expression differences of DEPs between the oe-NC-AGS and oe-ASAP3-AGS groups, with obvious grouping clustering (Figure 7C). Afterwards, subcellular localization analysis showed that the DEPs were mainly distributed in the nucleus (39 proteins), cytoplasm (22 proteins), extracellular space (18 proteins), plasma membrane (8 proteins), and mitochondria (7 proteins) (Figure 7D). There were partial differences in the subcellular localization of DEPs compared with those in HGC-27 cells, suggesting that the mechanism of ASAP3 may be cell-type-specific in different GC cell lines. In addition, the identified DEPs were closely related to GO terms of BP like “regulation of osteoclast differentiation”, “body fluid secretion”, and “regulation of myeloid leukocyte differentiation”; CC such as “sperm midpiece”, “C/EBP homologous protein–CCAAT/enhancer-binding protein complex”, and “activating transcription factor 4-cyclic adenosine monophosphate response element binding protein 1 transcription factor complex”; MF such as “transposase activity”, “jistone acetyltransferase binding”, and “thymidylate synthase activity” (Figure 7E); as well as KEGG pathways of “estrogen signaling pathway”, “glucagon signaling pathway”, “TNF signaling pathway”, “parathyroid hormone synthesis, secretion and action”, and “cyclic guanosine monophosphate-protein kinase G signaling pathway” (Figure 7F).

Figure 7 Underlying mechanisms of ASAP3 in AGS cells using proteomics. A: The principal component analysis of the over-expression (oe)-negative control (NC)-AGS and oe-ASAP3-AGS cells; B: The volcano plot of the differential expressed proteins (DEPs) between the oe-NC-AGS and oe-ASAP3-AGS cells; C: The heatmap of the identified DEPs; D: Subcellular localization analysis of the DEPs; E: The Gene Ontology terms enriched by the identified DEPs, including biological process, cellular component, and molecular function; F: Kyoto Encyclopedia of Genes and Genomes pathway analysis of the identified DEPs. PC: Principal component; GO: Gene Ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; BP: Biological process; MF: Molecular function; CC: Cellular component; oe: Over-expression; NC: Negative control.

Validation of the related mechanisms using RT-qPCR, western blot, co-IP and RhoA activation pull down

Based on the omics data pointing towards GTPase regulation and the Hippo pathway, we performed mechanistic validation experiments. RT-qPCR analysis showed CCN1, and AMOTL2 expression was significantly higher, while CCN2 expression was markedly lower in the ASAP3-silencing HGC-27 cells that those in the sh-NC HGC-27 cells (P < 0.05, Figure 8A). However, the expression trend of CCN1, CCN2, and AMOTL2 in different AGS cells were opposite to that in HGC-27 cells (Figure 8A). Western blot analysis showed the p-MST1/MST2/MST1/MST2 level was evidently decreased in HGC-27 cells after ASAP3 silencing (P < 0.05), whereas there was no significant difference in the p-MST1/MST2/MST1/MST2 Level between the oe-ASAP3-AGS and oe-NC-AGS groups (P > 0.05, Figure 8B). Furthermore, in HGC-27 cells, silencing ASAP3 significantly decreased the protein expression levels of p-LATS1/LATS2/LATS1/LATS2, p-YAP/YAP, p-TAZ/TAZ, and ARHGAP12 (P < 0.05), while increasing ASAP1 protein expression (P < 0.05); however, in AGS cells, overexpressing ASAP3 remarkedly increased the levels of p-LATS1/LATS2/LATS1/LATS2, p-YAP/YAP, p-TAZ/TAZ, and ARHGAP12 (P < 0.05), while downregulating the ASAP1 protein (P < 0.05, Figure 8B). After that, co-IP results showed that overexpression of ASAP3 reduced the interaction of ASAP1-ARHGAP12, while knockdown of ASAP3 increased the interaction of ASAP1-ARHGAP12 (Figure 8C). Finally, a RhoA activation pull down assay demonstrated that ASAP3 overexpression inhibited RhoA activity, while ASAP3 silencing enhanced RhoA activity (Figure 8D), suggesting ASAP3 overexpression could inhibit RhoA activity and thus suppress YAP/TAZ activation by interfering with the ASAP1-ARHGAP12 interaction.

Figure 8 Validation of the related mechanisms using real-time quantitative polymerase chain reaction, western blot, co-immunoprecipitation and pull down. A: The messenger RNA expression of CCN1, CCN2 and AMOTL2 in gastric cancer (GC) cells with different treatments measured using real-time quantitative polymerase chain reaction; B: The protein expression of phosphorylated (p)-MST1/MST2, p-LATS1 + LAST2, p-yes-associated protein, p-transcriptional coactivator with PDZ-binding motif, ASAP1 and ARHGAP12 in GC cells with different treatments detected by western blot; C: The relationship between ASAP3 and ASAP1-ARHGAP12 interaction determined by co-immunoprecipitation; D: The correlation between ASAP3 and RhoA activity examined by RhoA activation pull down. ^cP < 0.05 vs short hairpin-negative control-HGC-27 cell. ^dP < 0.05 vs over-expression-negative control-AGS cell. mRNA: Messenger RNA; oe: Over-expression; sh: Short hairpin; NC: Negative control; YAP: Yes-associated protein; TAZ: Transcriptional coactivator with PDZ-binding motif; p-YAP: Phosphorylated yes-associated protein; p-TAZ: Phosphorylated transcriptional coactivator with PDZ-binding motif; GAPDH: Glyceraldehyde-3-phosphate dehydrogenase; IP: Immunoprecipitation; WB: Western blot.

Correlation between ASAP3/ASAP1/ARHGAP12 and clinicopathologic features using bioinformatics

Eventually, we analyzed the correlations between ASAP3/ASAP1/ARHGAP12 and clinicopathologic features (including age, gender, TNM stage, tumor size and overall survival) using bioinformatics. As shown in Figure 9A, no significant relationships were found between ASAP3 and age (P = 0.54) as well as between ASAP1 and age (P = 0.53). However, the ARHGAP12 expression had significantly positive correlation with age (R = 0.16, P = 0.00096). Moreover, there were no significant differences in ASAP3 (P = 0.42), ASAP1 (P = 0.75) and ARHGAP12 (P = 0.87) expression between male and female patients (Figure 9B). Kruskal-Wallis test showed statistically significant differences in ASAP1 (P = 0.000) and ARHGAP12 (P = 0.028) expression among the patients at stages I, II, III and IV, but no obvious change in ASAP3 (P = 0.815) expression among the four stages (Figure 9C). For tumor size, no significant correlations were observed between ASAP3 (P = 0.053)/ASAP1 (P = 0.11)/ARHGAP12 (P = 0.26) and tumor size (Figure 9D). In addition, we found that high expression of ASAP3 and ASAP1 had lower overall survival than their low expression; but the opposite observation was shown in ARHGAP12 expression (Figure 9E).

Figure 9 Correlation between ASAP3/ASAP1/ARHGAP12 and clinicopathologic features using bioinformatics. A: Correlation between ASAP3/ASAP1/ARHGAP12 and age; B: Correlation between ASAP3/ASAP1/ARHGAP12 and gender; C: Correlation between ASAP3/ASAP1/ARHGAP12 and tumor node metastasis stage; D: Correlation between ASAP3/ASAP1/ARHGAP12 and tumor size (mm); E: Correlation between ASAP3/ASAP1/ARHGAP12 and patient overall survival. TNM: Tumor node metastasis; FPKM: Fragments per kilobase of transcript per million mapped reads.

DISCUSSION

GC remains a leading cause of cancer-related mortality globally, with limited therapeutic options for advanced disease primarily due to incomplete understanding of its molecular pathogenesis[29,30]. In this study, we identified a novel tumor-suppressive role and mechanism for ASAP3, a member of the ArfGAP family, in GC oncogenesis and progression. We demonstrated that ASAP3 expression was variable in GC cell lines and that its enforced expression significantly suppressed, while its knockdown promoted, the malignant phenotype of GC cells both in vitro and in vivo. Mechanistically, ASAP3 overexpression may exert its anti-tumor effects by disrupting the ASAP1-ARHGAP12 interaction, inhibiting RhoA activity, and regulating the downstream Hippo/YAP/TAZ signaling pathway, thus providing a novel regulatory axis for GC progression and a potential therapeutic target.

Prior studies have reported elevated ASAP3 expression in multiple solid tumors, including lung adenocarcinoma, colorectal cancer, and glioblastoma, where it correlates with aggressive phenotypes and poor prognosis[17-19]. For instance, Guo et al[31] showed ASAP3 was upregulated in the colorectal cancer, and its expression was positively correlated with lymph node metastasis. Meanwhile, a meta-analysis of Willis et al[32] found high ASAP3 mRNA levels in ovarian serous carcinomas were associated with shorter overall and progression-free survival. However, our study revealed a contrasting tumor-suppressive role of ASAP3 in GC: ASAP3 expression was lower in AGS cells (a well-differentiated GC cell line) compared to HGC-27 cells (a poorly differentiated, highly invasive line), and modulation of ASAP3 expression exerted opposite effects on GC cell biology. Specifically, ASAP3 overexpression in AGS cells inhibited proliferation, migration, and invasion, induced S cell cycle arrest and apoptosis, and suppressed xenograft tumor growth; conversely, ASAP3 silencing in HGC-27 cells enhanced these malignant phenotypes. This discrepancy in ASAP3’s role across cancer types highlights the context-dependent nature of ArfGAP family proteins, which may arise from tissue-specific signaling networks, tumor heterogeneity, or differences in downstream effector molecules[33,34]. For GC, our in vitro and in vivo data collectively establish ASAP3 as a negative regulator of malignant progression, challenging the prevailing view of ASAP3 as a universal oncogene and expanding our understanding of its tissue-specific functions.

The Hippo/YAP/TAZ pathway is a conserved regulator of organ size and tumorigenesis, and its aberrant activation is a hallmark of GC, particularly in aggressive subtypes like poorly cohesive GC[35,36]. YAP/TAZ translocation to the nucleus drives the expression of proliferation-promoting genes (e.g., CCN1, CCN2) and suppresses apoptosis, directly contributing to cancer cell invasion and metastasis[37,38]. RhoA, a small GTPase, is a critical upstream activator of YAP/TAZ. Its activity is tightly controlled by RhoGAP proteins (e.g., ARHGAP12), which catalyze the hydrolysis of GTP-bound RhoA to its inactive GDP-bound form[39,40]. In GC, high RhoA expression correlates with advanced tumor stage, lymph node metastasis, and poor prognosis[41,42], making the RhoA/YAP/TAZ axis a promising therapeutic target.

Our mechanistic studies reveal that ASAP3 intersects with this axis at multiple levels. First, transcriptomic and proteomic analyses identified enrichment of Hippo signaling pathway components (e.g., MST1/MST2, LATS1/LATS2) and YAP/TAZ target genes (CCN1, CCN2, AMOTL2) among the identified DEMs and DEPs following ASAP3 modulation. CCN1 has been reported to play a key role in the regulation of proliferation, differentiation, apoptosis, angiogenesis and fibrosis[43]. CCN2 has been proven to be essential for the fibrotic function (activation, proliferation and migration) of fibroblasts through the transforming growth factor beta receptor/SMAD pathway[44]. CCN1 and CCN2 were the targets of YAP/TAZ, and a previous study demonstrated that wogonin could enhance the radio-sensitivity of hepatocellular carcinoma through regulating the Hippo-YAP/TAZ signaling pathway (CCN1 and CCN2 expression)[45]. AMOTL2, a kind of motin family proteins, plays a critical role in angiogenesis, tumorigenesis, and neurogenesis by regulating multiple cellular signaling pathways, including Hippo signaling[46]. Our RT-qPCR analysis confirmed that ASAP3 silencing in HGC-27 cells upregulated CCN1 and AMOTL2 while downregulated CCN2, while ASAP3 overexpression in AGS cells reversed these trends, consistent with YAP/TAZ activation or suppression, respectively. Furthermore, western blot further showed that ASAP3 silencing reduced the phosphorylation of MST1/MST2, LATS1/LATS2, YAP, and TAZ, whereas ASAP3 overexpression in AGS cells enhanced their phosphorylation. MST1 is mainly expressed in immune cells, while MST1/MST2 regulates lymphocyte development, transport, survival, and T cell and effector T cell differentiation through non-standard Hippo pathways or alternative pathways, thereby balancing immune activation and tolerance[47]. Zhang et al[48] demonstrated that MISP could inhibit ferroptosis through MST1/MST2 kinase and promote YAP activation in non-small cell lung cancer. LATS1/LATS2 are the Hippo pathway kinases, and the loss of them can facilitate immune escape of endometrial cancer tumors by downregulating major histocompatibility complex class I[49]. YAP and TAZ are the key determinants of malignant tumors. It has been found that transforming growth factor beta-activated kinase 1 can suppress the proteasomal degradation of YAP/TAZ proteins to promote their K63-ubiquitination and K48-ubiquitination, thereby controlling the progression of pancreatic cancer[50]. Taken together, we speculated that ASAP3 may restrain GC development via activating the Hippo pathway to suppress YAP/TAZ transcriptional activity.

Critically, we found that ASAP3 overexpression exerts this effect by inhibiting RhoA activity. The RhoA activation pull down assay demonstrated that ASAP3 overexpression in AGS cells suppressed the level of active RhoA-GTP, while ASAP3 silencing in HGC-27 cells enhanced it, with no changes in total RhoA protein levels. This finding aligns with the role of RhoA as an upstream activator of YAP/TAZ[51]. Notably, ARHGAP12 (a RhoGAP) was identified as a direct interacting partner of ASAP3 via co-IP. ARHGAP12 has been shown to inhibit RhoA activity and suppress tumor progression in other cancers[52,53], and our data suggested that ASAP3 may stabilize or enhance ARHGAP12-mediated RhoA inhibition. Collectively, these findings delineate a novel regulatory cascade in GC, wherein ASAP3 may interact with ARHGAP12 to promote RhoA inactivation, which in turn activate the Hippo signaling pathway and ultimately suppress YAP/TAZ-mediated transcriptional activity, providing a mechanistic basis for ASAP3’s tumor-suppressive role in GC progression.

ASAP1, a homolog of ASAP3, has been previously reported to drive GC progression by activating the IQGAP1/CDC42 pathway to enhance cell proliferation and chemotherapy resistance[12], as well as interacting with RhoA to promote YAP/TAZ activation[54]. Our study extended this understanding by uncovering a unique regulatory mechanism wherein ASAP3 interferes with the ASAP1-ARHGAP12 interaction. Notably, we did not observe a direct interaction between ASAP3 and ARHGAP12, suggesting that ASAP3 may target ASAP1 itself to disrupt the complex. Supporting this, a prior work has demonstrated that both ASAP1 and ASAP3 contain Bin/Amphiphysin/Rvs domains that mediate binding to Rab11 family interacting protein 3 (class II) (FIP3) (a Rab11 effector protein)[55]. FIP3 serves as a scaffold protein involved in endosomal trafficking and cytoskeletal reorganization[56], and the shared ability of ASAP1 and ASAP3 to bind FIP3 implies they may compete for this common interacting partner. Such competition could alter the subcellular localization or conformational stability of ASAP1, thereby impairing its ability to form a complex with ARHGAP12. Our western blot analysis also revealed a reciprocal relationship between ASAP3 and ASAP1: ASAP3 silencing in HGC-27 cells upregulated ASAP1 expression, while ASAP3 overexpression in AGS cells downregulated it. Co-IP experiments further confirmed that ASAP3 directly interfered with the ASAP1-ARHGAP12 interaction, with this interaction being strengthened by ASAP3 silencing and weakened by ASAP3 overexpression. Therefore, we propose a “competitive sequestration” model wherein ASAP3 may disrupt the ASAP1-ARHGAP12 complex by outcompeting ASAP1 for ARHGAP12 binding, influencing RhoA activity and Hippo pathway suppression, thereby inhibiting GC progression. To our knowledge, this is the first report of a competitive interaction between ASAP family members regulating RhoGAP function, adding a new layer of complexity to the modulation of RhoA/YAP/TAZ signaling in GC.

Our bioinformatics analysis of public clinical datasets offers preliminary insights into the clinical relevance of the ASAP3-ASAP1-ARHGAP12 axis. ARHGAP12 expression correlated positively with patient age, consistent with its RhoA-inhibitory and potential tumor-suppressive role[52,57]. Notably, ASAP1 and ARHGAP12 expression varied significantly across TNM stages, suggesting involvement in GC progression[12], while ASAP3 expression remained stable across stages, likely reflecting its context-dependent expression in different GC subtypes, as observed in our in vitro models. Furthermore, survival analysis showed high ASAP3 and ASAP1 expression associated with poorer overall survival, whereas high ARHGAP12 correlated with better prognosis. This aligns with our functional data (ASAP1 as an oncoprotein[58], ARHGAP12 as a tumor suppressor[25]), while the apparent contradiction between ASAP3’s tumor-suppressive in vitro role and poor survival association may stem from clinical tumor heterogeneity, modulation by coexisting signaling alterations, or limitations of public datasets (e.g., lack of subtype/treatment stratification). These bioinformatics findings support the axis’s clinical potential but require validation with clinical tissues (via immunohistochemistry/RT-qPCR) and large-scale cohort analysis to confirm protein expression patterns and refine prognostic utility, laying the groundwork for their translational application as GC biomarkers or therapeutic targets.

Our findings have important translational implications. First, assessing ASAP3 expression levels in GC patient tissues could potentially serve as a prognostic biomarker, identifying tumors with dysregulated RhoA-YAP/TAZ signaling. Future studies with large clinical cohorts should validate whether ASAP3 expression predicts GC patient survival or response to therapy. Second, targeting the ASAP3-ARHGAP12-RhoA axis could represent a novel therapeutic strategy. For example, small molecules that mimic ASAP3’s ability to disrupt the ASAP1-ARHGAP12 interaction or enhance ARHGAP12 activity may inhibit RhoA/YAP/TAZ signaling in GC. Given the role of YAP/TAZ in GC therapeutic resistance[59,60], such agents might also sensitize GC cells to chemotherapy or immunotherapy.

Despite the robust data presented, our study has certain limitations. First, the investigation was primarily conducted in two GC cell lines (AGS and HGC-27), with ASAP3 manipulated in a unidirectional manner (knockdown in HGC-27, overexpression in AGS) without reciprocal validation, and no additional GC cell lines or normal gastric epithelial line (e.g., GES-1) for baseline comparison were included, which limits the generalizability of our conclusions on ASAP3’s universal role in GC. Second, mechanistic causality of the ASAP3 to ASAP1/ARHGAP12 to RhoA to YAP/TAZ axis lacks rescue experiment validation (e.g., ARHGAP12 knockdown, constitutively active RhoA, YAP/TAZ overexpression, transcriptional enhanced associate domain reporter assays, YAP/TAZ nuclear localization immunofluorescence, which are essential to confirm causal links between pathway steps. Third, the RhoA activation pull-down assay, a core experiment for verifying the downstream mechanistic axis of ASAP3, should be performed with three independent biological replicates to further ensure statistical robustness of the RhoA activity detection results. Fourth, our in vivo experiments used subcutaneous xenografts, which do not recapitulate the complex tumor microenvironment or metastatic spread of GC. Orthotopic GC models or patient-derived xenografts would provide more clinically relevant insights into ASAP3’s role in metastasis. Fifth, the study lacks clinical validation of ASAP3 expression, its correlation with RhoA/YAP/TAZ pathway molecules, and its prognostic value in GC patients based on large cohorts. Additionally, while we demonstrated a competitive interaction between ASAP3 and ASAP1 for ARHGAP12, the structural basis of this interaction remains unclear. Structural biology studies (e.g., X-ray crystallography) could identify critical residues mediating ASAP3 interfered ASAP1-ARHGAP12 binding, facilitating the design of targeted small molecules.

CONCLUSION

Our study identified ASAP3 as a novel tumor suppressor in GC that may suppress RhoA activity by disrupting the ASAP1-ARHGAP12 interaction, thereby activating the Hippo pathway and suppressing YAP/TAZ-mediated oncogenesis. These findings clarify the context-dependent function of ASAP3, uncover a new regulatory mechanism of the RhoA/YAP/TAZ axis, and provide a potential prognostic biomarker and therapeutic target for GC. Future studies will focus on validating ASAP3’s prognostic value in clinical cohorts and developing small-molecule modulators of the ASAP3 interfered ASAP1-ARHGAP12 interaction for GC therapy.