Tumor tissues and adjacent tissues were collected from 80 patients who underwent radical gastrectomy for GC at Tianjin Medical University General Hospital (Tianjin, China). The tissues were stored in liquid nitrogen for RNA extraction and were fixed in 4% formalin at room temperature for 24 h and embedded in paraffin for immunohistochemistry (IHC). Median patient age was 67 years (range, 39–85 years). The histopathological diagnosis and grading were performed by two experienced pathologists, and clinicopathological characteristics were recorded. Survival data, such as overall survival (OS) and recurrence-free survival (RFS), were also collected to perform survival analysis via the Kaplan-Meier method and values were compared using the log-rank test (18). Written informed consent was obtained from all patients or their relatives for specimen collection. Ethics approval for this project was granted by the Investigation and Ethics Committee of Tianjin Medical University General Hospital (approval no. IRB2020-KY-640).

The human GC cell lines NCI-N87 (cat. no. CL-0169) and MKN45 (cat. no. CL-0292), and the noncancerous gastric epithelial cell line GES-1 (cat no. AY09234) were obtained from the Laboratory of General Surgery, Tianjin Medical University General Hospital which purchased them from Procell Life Science & Technology Co., Ltd. (NCI-N87 and MKN45) and Ai-yan Biotechnology Co. Ltd. (GES-1). The human GC cell lines HGC27 (cat. no. CL-0107) and AGS (cat. no. CL-0022) were provided by Procell Life Science & Technology Co., Ltd. Short tandem repeat analysis was performed on these human cell lines for authentication. The GES-1, NCI-N87, MKN45 and HGC27 cells were cultured in RPMI-1640 (cat no. 11875093), whereas AGS cells were cultured in F12 (cat no. 11765054) supplemented with 10% fetal bovine serum (cat no. A5670701) and 1% penicillin/streptomycin (cat no. 15140122) (all from Gibco; Thermo Fisher Scientific, Inc.) at 37°C in a humidified atmosphere with 5% CO₂.

Total RNA was extracted using RNA Isolater Total RNA Extraction Reagent (cat no. R401; Vazyme Biotech Co., Ltd.), and was converted to cDNA with the HiScript^® III 1st Strand cDNA Synthesis Kit (cat no. R312; Vazyme Biotech Co., Ltd.) or the miRcute Plus miRNA First-Strand cDNA Kit (cat no. KR211; Tiangen Biotech Co., Ltd.), according to manufacturers' protocols. qPCR was performed using ChamQ Universal SYBR RT-qPCR Master Mix (cat no. Q711; Vazyme Biotech Co., Ltd.) or the miRcute Plus miRNA RT-qPCR Kit (cat no. FP411; Tiangen Biotech Co., Ltd.) according to the manufacturer's protocol. For the detection of miRNA, the qPCR cycling conditions were as follows: Initial denaturation step at 95°C for 15 min, followed by 45 cycles at 94°C for 20 sec (denaturation) and 60°C for 34 sec (annealing and extension), concluding with a melting curve analysis. For mRNA detection, the qPCR cycling conditions included an initial denaturation step at 95°C for 30 sec, followed by 40 cycles at 95°C for 20 sec (denaturation) and 60°C for 30 sec (annealing and extension), also concluding with a melting curve analysis. The reverse primers for miRNA were provided in the qPCR kit. The copy number of miR-424-5p in patient tissues was calculated via absolute quantification methods (19), and relative quantification was performed to evaluate the expression of miRNAs and mRNAs in cell lines, which was analyzed using the comparative 2^−ΔΔCq method (20). The amplification curve and standard curve of miR-424-5p are shown in Fig. S1. The primer sequences are shown in Table SI, and GAPDH and U6 were used as housekeeping genes for mRNA and miRNA expression, respectively. All procedures were performed in triplicate.

The miR-424-5p mimics and inhibitor, and the negative controls (ncs) were constructed by Shanghai GenePharma Co., Ltd. HGC27 and AGS cells were transfected in 6-well plates with the miR-424-5p inhibitor, mimics or ncs (50 pmol/ml) using Lipofectamine^® 2000 (cat no. 11668019; Invitrogen; Thermo Fisher Scientific, Inc.) once they reached 30–50% confluence, following the manufacturer's instructions, at 37°C in CO₂ for 24 h. After 24 h, the cells were harvested for subsequent experiments. The lentivirus used to knock down SMURF1 [short hairpin (sh)SMURF1] and its negative control (sh-nc; cat. no. 20221126) was constructed using the pCDH-CMV-MCS-EF1-copGFP-T2A-Puro vector (Shanghai GenePharma Co., Ltd.) as a backbone, whereas the SMURF1 overexpression lentivirus (LV-SMURF1) and its negative control (LV-nc) were constructed using the pLKO.1-copGFP-2A-PURO vector (Shanghai GenePharma Co., Ltd.). All lentiviral constructs were produced by Shanghai GenePharma Co., Ltd. using the 3rd generation system. The AGS and HGC27 cells were then infected with lentiviral particles (MOI=10) with the addition of 5 µg/ml polybrene (cat no. H8761; Beijing Solarbio Science & Technology Co., Ltd) for 8 h at 37°C in a humidified atmosphere containing 5% CO₂. The culture solution was fully replaced at 24 h. The stable cell lines were selected with puromycin (cat. no. HY-B1743; MedChemExpress) after 3 days. RT-qPCR and western blotting were used to determine the transfection and infection efficiencies. All of the aforementioned sequences are presented in Table SII.

HGC27 and AGS cells transfected with miR-424-5p inhibitors, mimics or ncs were seeded into 6-well plates at a density of 1,000 cells/well, and were cultured at 37°C in a humidified atmosphere containing 5% CO₂, with the culture media replaced every 3 days. Colony formation was assessed after culturing the cells for 14 days. After washing the cells three times with phosphate-buffered saline, the colonies were fixed with 4% paraformaldehyde for 20 min and stained with 0.1% crystal violet (cat no. G1063; Beijing Solarbio Science & Technology Co., Ltd.) staining solution for 5 min at room temperature. Formations with >50 cells were identified as colonies under a dissecting microscope.

Cells were seeded into 6-well plates at a density of 2×10⁵ cells/well and were allowed to attach until they reached 80–90% confluence at 37°C in a humidified atmosphere with 5% CO₂. Subsequently, a straight scratch was made using a 20-µl pipette tip in the cell monolayer, and cell migration was monitored and documented every 24 h under an inverted fluorescence microscope. The migration of the cells was calculated by measuring the distance covered by the cells at each time point via ImageJ 1.54g (National Institutes of Health). The cells were serum starved for 12 h prior to the migration assay to negate the effect of cell proliferation before migration, and then cultured with fresh medium containing 1% fetal calf serum.

Cells (5×10⁴/well) were seeded into the upper chambers of a 24-well Corning^® 6.5 mm Transwell^® with 5.0 µm Pore Polycarbonate Membrane Insert (cat. no. 3421; Corning, Inc.) to detect cell migration and invasion. To assess cell invasion, the insert was pre-coated with Matrigel (cat no. 354248; BD Biosciences) at 37°C for 1 h. In addition, medium containing 10% fetal bovine serum was added to the lower chambers. After incubation for 24 h at 37°C in a humidified atmosphere containing 5% CO₂, the migratory and invasive cells were stained with 0.1% crystal violet at room temperature for 30 min, and images were captured under an inverted fluorescence microscope. All assays were repeated at least three times.

Total protein was isolated from cells using RIPA buffer (cat no. R0010; Beijing Solarbio Science & Technology Co., Ltd.) supplemented with protease inhibitor (cat no. A32955; Thermo Fisher Scientific, Inc.). After the protein concentration was measured using the BCA assay, the protein samples were boiled and ~15 µg protein was separated by SDS-PAGE on 10% gels and transferred to PVDF membranes. After blocking non-specific binding sites with 5% non-fat dry milk (cat. no. D8340; Beijing Solarbio Science & Technology Co., Ltd.) at room temperature for 1 h. The membranes were probed with antibodies against SMURF1 (1:1,000; cat. no. 2174; Cell Signaling Technology, Inc.), caspase 3 (1:1,000; cat. no. 9662; Cell Signaling Technology, Inc.), ING2 (1:1,000; cat. no. 11560-1-AP; Proteintech Group, Inc.), p53 (1:5,000; cat. no. 10442-1-AP; Proteintech Group, Inc.) and β-actin (1:20,000; cat. no. 66009-1-Ig; Proteintech Group, Inc.) overnight at 4°C. After washing, the membranes were incubated with the HRP goat anti-rabbit IgG (1:5,000; cat no. SA00001-2; Proteintech Group Inc.) or HRP-conjugated goat anti-mouse IgG (1:5,000; cat no. SA00001-1; Proteintech Group Inc.) at room temperature for 1 h. An ECL substrate (cat. no. E411-04; Vazyme Biotech Co., Ltd.) was used to visualize the protein bands and the data were analyzed using ImageJ 1.54g.

GC and adjacent noncancerous tissues previously preserved in 4% formalin were embedded, then the sections were prepared to a thickness of 5 µm. Subsequently, sections were permeabilized with 0.1% Triton X-100 (cat. no. T8200; Beijing Solarbio Science & Technology Co., Ltd.) at room temperature for 30 min, allowing antibodies to enter the cells, and endogenous peroxidase was blocked with 3% hydrogen peroxide at room temperature for 10 min. BSA (5%; cat. no. SW3015; Beijing Solarbio Science & Technology Co., Ltd.) was used to block non-specific binding sites at room temperature for 1 h. SMURF1 antibody (1:200; cat no. ab57573; Abcam) was used for IHC staining, and sections were incubated with this antibody overnight at 4°C. The horseradish peroxidase-conjugated goat anti-mouse IgG (1:5,000; cat. no. ZB2305; Beijing Zhongshan Jinqiao Biotechnology Co., Ltd.) was used to incubate the sections at room temperature for 1 h. Chromogenic detection was performed using DAB (cat no. DA1010; Beijing Solarbio Science & Technology Co., Ltd.). Three fields were randomly selected to determine the percentage of positive cells and the staining intensity by light microscope. The expression grade of SMURF1 was calculated by multiplying the positivity score by the intensity score (21). Cells with <10% staining were rated as 1, cells with 10–49% staining were rated as 2, cells with 50–74% staining were rated as 3 and cells with 75–100% staining were rated as 4. The staining color was scored as light-yellow particles, 1; brown-yellow particles, 2; and brown particles, 3. The final score was defined as staining number score multiplied by staining color score. Patients were subsequently divided into high expression level (grade >3) and low expression level (grade ≤3) groups.

The miR-424-5p binding site in the 3′-UTR of SMURF1 was predicted using starBase v2.0 (http://starbase.sysu.edu.cn/). psiCHECK2-wild type (WT)-SMURF1 and psiCHECK2-mutant-SMURF1 were synthesized by Hanbio Biotechnology Co., Ltd. and were cotransfected with miR-424-5p mimics or nc into 293T cells (cat. no. CL-0005; Procell Life Science & Technology Co., Ltd) Lipofectamine 2000. After culture for 48 h, the dual-luciferase system (cat no. E2920; Promega Corporation) was used to measure the firefly luciferase (F-Luc) and Renilla luciferase (R-Luc) values, and the relative fluorescence values of firefly luciferase were calculated as F-Luc/R-Luc.

The cells were inoculated into 96-well plates (4,000 cells/well) (Corning, Inc.) for 24 h. Then, the cells were treated with different concentrations (0, 0.005, 0.05, 0.5, 5, 50 and 500 µg/ml) of CDDP (cat. no. P4394; MilliporeSigma) for 48 h at 37°C in a humidified atmosphere containing 5% CO₂. The CCK-8 assay (cat. no. CA1210; Beijing Solarbio Science & Technology Co., Ltd.) was used to measure cell viability and the half maximal inhibitory concentration (IC₅₀) was determined using GraphPad Prism 8 software (Dotmatics).

The cells were divided into an experimental group and a control group, and were seeded into 6-well plates (1×10⁵ cells/well; Corning Life Sciences) for 24 h. The experimental group was treated with CDDP (10 µg/ml in AGS cells and 2 µg/ml in HGC27 cells) for 48 h at 37°C and the control group was treated with DMSO for the same duration. Annexin V-FITC/7-AAD (cat no. 40311ES60; Shanghai Yeasen Biotechnology Co., Ltd.) was used for apoptosis analysis; the samples were stained with Annexin V-FITC at room temperature for 5 min in the dark, after which, 7-AAD was added and the cells were immediately analyzed by flow cytometry (BD FACSCanto II; BD Biosciences). The data were analyzed using FlowJo (v10.8.1; FlowJo LLC) and GraphPad Prism 8 software. The apoptosis rate was represented by the sum of the early apoptotic rate and the late apoptotic rate.

SPSS version 26 (IBM Corp.) and GraphPad Prism 8 were used to analyze the data. The numerical data are presented as the mean ± standard deviation of at least three experimental repeats. The normality of the data was tested using the Kolmogorov-Smirnov test, and the comparisons among groups were analyzed using either unpaired or paired Student's t-test or one-way ANOVA followed by Bonferroni's post hoc test. Clinicopathological data were analyzed using the χ² test or Fisher's exact test. Correlations between variables were tested using linear regression and Spearman rank tests. P<0.05 was considered to indicate a statistically significant difference.

The expression of miR-424-5p in noncancerous gastric epithelial cells (GES-1) and GC cell lines (HGC27, MKN45, NCI-N87 and AGS) was assessed using RT-qPCR. As shown in Fig. 1A, the expression of miR-424-5p in GC cells was lower than that in GES-1 cells. In addition, 80 pairs of GC tissues and adjacent noncancerous tissues were collected and the absolute expression of miR-424-5p was assessed by miRNA RT-qPCR. Compared with those in paired adjacent tissues, the expression levels of miR-424-5p in human GC tissues were decreased (Fig. 1B). The samples were subsequently divided into high- and low-expression groups on the basis of their miR-424-5p expression levels, according to the median value. As shown in Table I, owing to one missing data point, 40 patients were included in the high-miR-424-5p group and 39 patients were included in the low-miR-424-5p group. Low miR-424-5p expression was associated with increased pathological N stage. The survival analysis of all samples revealed a weak association with poor prognostic outcome, but the results were not significant (Fig. 1C). Subgroup analysis revealed that lower expression of miR-424-5p was associated with a poor OS and RFS in patients treated with adjuvant chemotherapy (Fig. 1D).

To investigate the effect of miR-424-5p on chemotherapy resistance in GC, the present study used AGS and HGC27 cells to verify the biological function of miR-424-5p. The results of miR-424-5p expression and the IC₅₀ values of CDDP in AGS and HCG27 cells are shown in Fig. 2A. The IC₅₀ value in AGS cells was greater than that in HGC27 cells, which indicated that AGS cells may be more resistant to CDDP than HGC27 cells. According to the miR-424-5p expression status in GC cells shown in Fig. 1A, miR-424-5p exhibited the highest expression in HGC27 cells, whereas its expression was the lowest in AGS cells. Therefore, the HGC27 cells were selected for miR-424-5p silencing, while AGS cells were chosen for miR-424-5p overexpression in subsequent experiments. The miR-424-5p mimics and inhibitor were constructed and transfected into AGS and HGC27 cells, respectively. RT-qPCR was performed to verify the transfection efficiency in the cell lines (Fig. 2B). As expected, miR-424-5p was significantly upregulated in AGS cells and significantly downregulated in HGC27 cells post-transfection compared with that in the nc groups. First, the colony formation, wound healing, Transwell and invasion assays revealed that there was no significant difference in cell proliferation, invasion and migration after altering the expression of miR-424-5p (Fig. S2). A CCK-8 assay, flow cytometry and western blotting were used to evaluate the effects of miR-424-5p on CDDP resistance in GC cells. The CCK-8 assay results revealed that the sensitivity of AGS cells transfected with the miR-424-5p mimics to CDDP was increased compared with that of the cells transfected with miR-nc, whereas HGC27 cells transfected with the miR-424-5p inhibitor presented the opposite results (Fig. 2C and D). Following CDDP treatment, flow cytometry revealed that the apoptosis rate of the miR-424-5p mimics-transfected cells was greater than that of the miR-nc-transfected cells (Fig. 2E). By contrast, a lower percentage of apoptotic cells was detected in the miR-424-5p inhibitor group (Fig. 2F). In addition, to evaluate apoptosis, the expression levels of caspase 3 and cleaved caspase 3 were assessed. In contrast to the nc groups, the cleaved caspase 3/caspase 3 rate was elevated in the miR-424-5p mimics group and was reduced in the miR-424-5p inhibitor group, which indicated that the miR-424-5p could promote apoptosis following CDDP treatment (Fig. 2G and H).

In our previous study, SMURF1 was found to be downstream of miR-424-5p (11). The miR-424-5p binding site in the 3′-UTR of SMURF1 was predicted using starBase v2.0. A dual-luciferase reporter assay revealed that the luciferase activity in the SMURF1 3′UTR WT group was significantly lower than that in the nc group (P<0.05; Fig. 3A). Thus, these results confirmed that miR-424-5p regulated luciferase expression through this binding site. To clarify the mechanism by which miR-424-5p regulates SMURF1 in GC, RT-qPCR and western blotting were used to assess the expression levels of SMURF1 in different treatment groups. The qPCR results revealed that the mRNA expression levels of SMURF1 in AGS cells transfected with miR-424-5p mimics were significantly lower than those in the miR-nc group (Fig. 3B). By contrast, the mRNA expression levels of SMURF1 in HGC27 cells transfected with the miR-424-5p inhibitor were significantly increased. Consistently, the western blotting results were the same as the RT-qPCR results. These results suggested that miR-424-5p may suppress the expression of SMURF1.

The present study next assessed the expression of SMURF1 in GC cell lines and GES-1 cells by RT-qPCR and western blotting. As shown in Fig. 4A and B, the mRNA expression levels of SMURF1 in GC cells, such as NCI-N87, AGS and MKN45, were greater than those in GES-1 cells, while SMURF1 protein levels were higher in NCI-N87, AGS, MKN45 and HGC27 cells. IHC analysis revealed that SMURF1 was expressed at higher levels in GC tissues than in adjacent tissues (Fig. 4C). Furthermore, miR-424-5p was revealed to be negatively correlated with SMURF1 in GC tissues, although the correlation was weak (Fig. 4D). In addition, the association between SMURF1 expression levels and clinicopathological characteristics was analyzed. High expression of SMURF1 was associated with increased Tumor-Node-Metastasis stage (Table I).

The present study confirmed that miR-424-5p regulated CDDP resistance in GC cells. To further confirm whether miR-424-5p regulated drug resistance in GC cells by partially inhibiting SMURF1 expression, AGS cells overexpressing SMURF1 and HGC27 cells with SMURF1 knockdown were constructed, and the effects of its overexpression and knockdown were verified by RT-qPCR and western blotting (Fig. S3). In addition, the expression of miR-424-5p was detected by RT-qPCR following SMURF1 knockdown or overexpression and it was found that the knockdown or overexpression of SMURF1 did not influence miR-424-5p expression (Fig. S4). Then, four groups of AGS cells, namely, the miR-nc + LV-nc, miR-424-5p mimics + LV-nc, miR-nc + LV-SMURF1 and miR-424-5p mimics + LV-SMURF1 groups, were constructed via cotransfection. Similarly, miR-nc + sh-nc, miR-424-5p inhibitor + sh-nc, miR-nc + sh-SMURF1 and miR-424-5p inhibitor + sh-SMURF1 HGC27 cells were constructed. In terms of drug response, the CCK-8 assay results revealed that CDDP resistance was increased after SMURF1 was upregulated in AGS cells and decreased after SMURF1 was downregulated in HGC27 cells (Fig. 5A and B). The CDDP resistance induced by LV-SMURF1 could be reversed by overexpressing miR-424-5p in GC cells. By contrast, the CDDP sensitivity induced by sh-SMURF1 was not be altered by the miR-424-5p inhibitor. In addition, flow cytometry demonstrated that the overexpression of SMURF1 in GC cells reduced the number of apoptotic GC cells and this effect was reversed by the transfection of miR-424-5p mimics (Fig. 5C). Moreover, the miR-424-5p inhibitor was found to reduce the apoptosis rate. By contrast, sh-SMURF1 increased the apoptosis rate, and this could be slightly attenuated by the miR-424-5p inhibitor (Fig. 5D). The western blotting results revealed a consistent trend in caspase 3 changes; the effects of SMURF1 expression on apoptosis in response to CDDP treatment were validated by the protein expression levels of cleaved caspase 3. The results indicated that miR-424-5p mimics enhanced the cleaved caspase 3/caspase 3 ratio. By contrast, SMURF1 overexpression significantly reduced this ratio, an effect that could be reversed by miR-424-5p mimics. Similarly, the miR-424-5p inhibitor decreased the cleaved caspase 3/caspase 3 ratio, whereas sh-SMURF1 increased the ratio, with this increase being modestly attenuated by the miR-424-5p inhibitor (Fig. 6A and B).

p53 serves an important role in chemosensitivity and ING2 activates p53 in neoplasms. Moreover, SMURF1 interacts with and targets ING2 for polyubiquitination and proteasomal degradation (14). Western blotting data revealed that the protein expression levels of ING2 and p53 were partially inhibited or promoted by the miR-424-5p mimics or inhibitor, respectively (Fig. 6C and D). It was found that the expression of ING2 and p53 in AGS cells transfected with LV-SMURF1 decreased and this change nearly disappeared following transfection with the miR-424-5p mimics. Similarly, the levels of ING2 and p53 were increased in HGC27 cells transfected with sh-SMURF1 and were partially restored by transfection with the miR-424-5p inhibitor. These results indicated that the overexpression of miR-424-5p may inhibit the expression levels of SMURF1 and increase the expression levels of p53.