A total of 24 paired GC tissues and corresponding adjacent normal tissues were collected from the First Affiliated Hospital of Chongqing Medical University (Chongqing, China). Pathological examination was performed to confirm the resected samples were GC tissues. Human ethnics approval was obtained from the Ethics committee of the First Affiliated Hospital of Chongqing Medical University. The informed consents were obtained from all patients.

RNA-seq data was acquired from the Cancer Genome Atlas (TCGA) database. This data included 27 patient-matched tumor and normal samples, 32 unpaired normal samples, and 375 unpaired tumor samples.

The GC cell lines MGC803 and AGS were purchased from the Type Culture Collection of the Chinese Academy of Sciences (Beijing, China). All cell lines were validated by STR profiling and tested negative for mycoplasma. Cells were cultured in a humidified incubator at 37 °C with 5% CO${}_2$ using RPMI 1640 medium (#01-100-1ACS, BI, Kibbutz Beit Haemek, Israel) containing 10% certified Fetal Bovine Serum (VivaCell, Shanghai, China).

Total RNA from tissues and cells was isolated using Trizol reagent (#9108, Takara, Osaka, Japan) according to manufacturer’s protocol. The concentration and quality of RNA was tested by using a nanodrop (#K5600C, KAIAO, Beijing, China). The RNA was used in cDNA synthesis reactions using PrimeScript RT reagent kit (#RR037A, Takara, Osaka, Japan). Relative RNA levels of YTHDF1 was determined using Ultra SYBR Mixture (#RR42WR, Takara, Osaka, Japan) and GAPDH mRNA was used as internal reference. Primers used in qRT-PCR are listed in Supplementary Table 1.

Small interfering RNA (siRNA) were used to knockdown gene expression, and siRNA targeting YTHDF1 and TCF7 was purchased from Sangon Biotech (Sangon Biotech, Shanghai, China).

Small hairpin RNA (shRNA) targeting YTHDF1 were ligated into the psi-LVRU6GP plasmid and a full-length YTHDF1 cDNA was cloned into the phosphoglycerate dehydrogenase-Cytomegalovirus (pCDH-CMV) plasmid. All plasmids were transfected into cells using Lipofectamine 2000 (#11668019, Invitrogen, Carlsbad, CA, USA). YTHDF1 overexpression and knockdown plasmids, together with psPAX2, pMD2.G plasmids, were transfected into 293T cells to generate lentiviral particles, and these lentivirus were subsequently used to transduce GC cell lines. Puromycin (4 µg/µL, Beyotime, Shanghai, China) was used to select stably transduced GC cells. Information relevant to siRNA, shRNA, and primers used is shown in Supplementary Table 2.

Flow cytometry was conducted at Chongqing Medical University, College of Life Sciences and was used to evaluate cell cycle dynamics and apoptosis of GC cells. For CCK-8 assays, cells were seeded in 96-well plates at a density of 2 $\times$ 10${}^3$ cells per well. Adherent cells were treated with 10 uL CCK-8 reagent (#CK04, DOJNDO, Tokyo, Japan) at 0, 24, 48 and 72 hours after seeding. Following a 2-hour incubation with CCK-8 reagent at 37 °C, the absorbance at 450 nm was measured using a spectrophotometer (Synergy2, BioTek, VT, USA). For colony formation assays, GC cells were seeded at 8 $\times$ 10${}^2$ cells/well into 6-well plates and further cultured. After 10 days, cells were fixed using 10% paraformaldehyde (#P0099, Beyotime, Shanghai, China) and stained using 0.1% crystal violet (#C0121, Beyotime, Shanghai, China) prior to counting. For the EdU staining, a total of 3 $\times$ 10${}^4$ GC cells per well were planted into a 96-well plate. After 24 hours, cell proliferation was determined following the manufacturer’s protocol (RiboBio, Guangzhou, China). Where noted, cells were imaged using fluorescence microscopy (#Leica DMi8, Leica, Wetzlar, Germany).

RIP was conducted according to the manufacturer’s instructions included in the Magna RIP Kit (Merck Millipore, Billerica, MA, USA). RIP experiments were conducted with anti-YTHDF1 antibody (Proteintech, Wuhan, China). Co-precipitated TCF7 was analyzed using qRT-PCR.

Tissues from GC patients and xenografts were fixed using 10% paraformaldehyde and subsequently embedded in paraffin. Immunohistochemistry staining was conducted as previously described [19]. Primary antibodies used were anti-YTHDF1 (1:200, Proteintech, Wuhan, China) and anti-Ki-67 (1:500, Cell Signaling Technology, Boston, USA). IHC images were captured with an inverted Leica microscope (#Leica DMI3000B, Leica, Wetzlar, Germany).

Proteins from cell and tissue samples were extracted using RIPA lysis buffer (1:100) containing PMSF (Beyotime, Shanghai, China). Proteins were subsequently added to SDS-PAGE gels prior to conducting electrophoresis. Proteins in SDS-PAGE gels were electrotransferred to PVDF membranes (#03010040001, Millipore, USA). Then, the PVDF membrane was blocked using 5% skim milk (Beyotime, Shanghai, China) for 1 hours. The primary antibodies used in this study are anti-GAPDH (1:6000, Proteintech, Wuhan, China), anti-YTHDF1 (1:3000, Proteintech, Wuhan, China), anti-MYC (1:8000, Proteintech, Wuhan, China), anti-Cyclin D1 (1:5000, Proteintech, Wuhan, China), anti-MMP7 (1:2500, Proteintech, Wuhan, China), and anti-TCF7 (1:5000, Proteintech, Wuhan, China). These antibodies were incubated with PVDF membranes at 4 °C for 12 hours. The next day, the membrane was incubated with corresponding secondary antibody (1:6000, Proteintech, Wuhan, China) for 2 hours. Proteins were visualized using an enhanced chemiluminescence solution ECL (#K-12049-D50, Advansta, CA, USA).

Six female nude mice were bought from the Animal Experimental Laboratory of Chongqing Medical University. MGC803 cells stably overexpressing YTHDF1 and control cells were injected (7 $\times$ 10${}^6$ cells) into the flank of each mouse. Tumor volume was assessed every 7 days using a vernier caliper. After four weeks, mice were anesthetized with pentobarbital sodium and then euthanized by cervical dislocation. Then, the tumors were harvested for immunohistochemistry and western blot analysis. Animal experiments were approved by the Ethics Committee of the First Affiliated Hospital of Chongqing Medical University.

All statistical analyses were conducted using GraphPad Prism 5 (GraphPad Software, San Diego, CA, USA), SPSS 19.0 (SPSS, Chicago, IL, USA) and R (version 4.1.0, Auckland, New Zealand) software. Data are expressed as mean $\pm$ SD. *p 0.05, **p 0.01 and ***p 0.001 were considered to have statistical significance.

To explore a potential role for YTHDF1 in GC, we first accessed gene expression and survival data from the TCGA database. We compared the expression level of YTHDF1 in 375 tumor tissues and 32 normal tissues and result indicated that the mRNA level of YTHDF1 is significantly higher in tumor tissues (Fig. 1A). Next, we compared expression of YTHDF1 between 27 paired GC tissues and adjacent normal tissues and similar results showing higher expression in GC tumors was observed in these tissue pairs (Fig. 1B). We then compared overall survival differences between the low YTHDF1 expression group and the high YTHDF1 expression groups. Here, we observed that GC patients with higher YTHDF1 expression have poorer overall survival (Fig. 1C).

Fig. 1.

YTHDF1 is differentially expressed in GC. (A,B) RNA level of YTHDF1 in unpaired and paired GC tissues and adjacent normal tissues in the TCGA database. (C) Survival differences of GC patients with high or low expression of YTHDF1 were analyzed using Kaplan-Meier analyses. (D) RNA expression level of YTHDF1 in 24 patient-matched GC and normal tissues. (E,F) The protein level of YTHDF1 in 24 paired GC tissues and corresponding normal tissues were detected using immunohistochemical and western blot assays (scale bar (200$\times$) = 200 µm, scale bar (400$\times$) = 100 µm). ***, p 0.001. YTHDF1, YTH domain-containing family protein 1; GC, gastric cancer; TCGA, the cancer genome atlas; GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

We also collected 24 paired GC cancer/normal tissues from our hospital as an independent validation of the differential expression of YTHDF1 observed in the TCGA dataset. Results demonstrated that YTHDF1 mRNA was elevated in most GC tissues compared with adjacent normal tissues (Fig. 1D). In addition, immunohistochemistry and western blotting were used to directly evaluate the level of YTHDF1 protein in these tissues. Results obtained showed that YTHDF1 was upregulated in 20 out of 24 GC tissues in our panel when compared with adjacent normal tissues (Fig. 1E,F).

To further explore the role of YTHDF1 in GC, we constructed GC cell lines stably overexpressing YTHDF1. qRT-PCR and western blot assays were utilized to confirm overexpression (Fig. 2A).

Fig. 2.

Overexpression of YTHDF1 promotes proliferation and inhibits apoptosis of GC Cells. (A) qRT-PCR and western blot assays was conducted to evaluate the over expression efficiency of YTHDF1 in GC cells. (B,C) CCK-8 results proved that YTHDF1 promote the proliferation ability of GC cells. (D,E) Colony formation assays were used to evaluate the effect of YTHDF1 on proliferation ability of GC cells. (F,G) Flow cytometry was utilized to assess the cell cycle of GC cells after over expression of YTHDF1. (H,I) EdU staining was performed to evaluate the DNA synthesis abilities of GC cells after over expression of YTHDF1 (scale bar = 120 µm). (J,K) Flow cytometry was conducted to determine whether YTHDF1 effects the apoptotic population in cultured GC cells. *, p 0.05; **, p 0.01. qRT-PCR, quantitative real time polymerase chain reaction; CCK-8, cell counting kit-8; EdU, 5-Ethynyl-2${}^{\prime }$-deoxyuridine.

CCK-8 and colony forming assays indicated that proliferative ability of GC cells was increased following YTHDF1 overexpression (Fig. 2B–D). Flow cytometry indicated that YTHDF1 overexpression induced a higher proportion of S phase cells than that seen in normal control (NC) cells (Fig. 2E–G). In addition, EdU staining was utilized to measure cell proliferation in GC cells overexpressing YTHDF1. Consistent with other results, YTHDF1 overexpression significantly enhanced the rates of DNA synthesis in GC cells (Fig. 2H,I).

We also examined whether YTHDF1 could affect apoptosis in GC cells. The results suggested that YTHDF1 overexpression partially inhibited GC cell apoptosis (Fig. 2J,K). Taken together, these results suggest that YTHDF1 may act in promoting GC tumor growth.

To further examine the potential tumor promoting role of YTHDF1, we designed siRNAs targeting YTHDF1. Knockdown efficiency was measured using qRT-PCR and western blotting (Fig. 3A). CCK-8 and colony formation assays were used to evaluate the proliferative ability of GC cells. We observed that YTHDF1 knockdown significantly inhibited GC cell growth (Fig. 3B–D). Flow cytometry results demonstrated that MGC803 and AGS cells transfected with siRNA specific for YTHDF1 were arrested in the G0/G1 phase of the cell cycle (Fig. 3E–G). Result of EdU assay also showed that proliferation ability of the YTHDF1 knockdown group was significantly inhibited when compared to EdU incorporation than the NC cell group (Fig. 3H,I). Analysis of apoptosis by flow cytometry showed that the number of apoptotic GC cells was increased after knockdown of YTHDF1 (Fig. 3J,K). Taken together, these results suggest that YTHDF1 exert a crucial role in establishing/maintaining GC cell malignancy.

Fig. 3.

Knock down of YTHDF1 inhibits proliferation and promotes apoptosis of GC Cells. (A) Knockdown efficiency of YTHDF1 in GC cells was measured using qRT-PCR and western blot analysis. (B,C) CCK-8 assays indicate that knockdown of YTHDF1 inhibits the proliferation of GC cells. (D,E) Colony formation assays were used to evaluate the effect of YTHDF1 knockdown on GC cell colony formation ability. (F,G) Flow cytometry was utilized to assess the effect of YTHDF1 knockdown on the cell cycle of GC cells. (H,I) EdU staining was performed to evaluate the proliferation status of GC cells after knockdown of YTHDF1 (scale bar = 120 µm). (J,K) Flow cytometry was conducted to determine the effects of YTHDF1 knockdown on GC cell apoptosis. *, p 0.05; **, p 0.01. siRNA, small interfering RNA; NC, normal control.

To further explore the underlying mechanism mediated by YTHDF1, we conducted transcriptome sequencing on GC cells with YTHDF1 knockdown (Supplementary Table 3). Using the filter condition of ($|$log2FC$|$ $>$ 1 and padj 0.05), we identified a total of 649 differentially expressed genes, including 394 upregulated genes and 225 downregulated genes. The differentially expressed genes were visualized using a heatmap and volcano plot (Fig. 4A,B).

Fig. 4.

Identification of potential target pathways regulated by YTHDF1. (A,B) Transcriptome sequencing was conducted after YTHDF1 knockdown. (C) mRNA expression differences between NC and YTHDF knockdown groups were visualized using a heatmap and a volcano map. GO and KEGG enrichment analyses were conducted on the differently expressed genes between the NC group and the YTHDF knock down group. (D) meRIP-seq, RNA-seq and RIP-seq data collected on MGC803 cells in a previously published study were acquired and interogated to obtain the potential target genes of YTHDF1. (E) GO and KEGG enrichment results of 991 YTHDF1 potential target genes from the above analyses. (F) Cytoscape software was utilized to visualize core genes involved in top four signal pathways and two signal pathways correlated with GC. mRNA, messenger RNA; NC, normal control; GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; meRIP-seq, methylated RNA immunoprecipitation sequencing.

Next, we performed Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis on these differentially expressed genes. We found that several tumor-associated signaling pathways such as MAPK signaling, Hippo signaling, WNT signaling, and TGF-$\beta$ signaling pathways were significantly enriched (Fig. 4C). This finding suggests that the tumor promoting function of YTHDF1 is possibly associated with one or more of these pathways.

Considering that YTHDF1 was reported to enhance the translational efficiency of its target genes, we speculated that YTHDF1 might promote target gene translation and thereby affect these tumor related pathways. To identify YTHDF1 direct target genes, we acquired the meRIP-seq, RNA-seq and RIP-seq datasets obtained from the analysis of MGC803 cells from a previously published study [20]. Our analysis of these data sets uncovered 991 genes that are potential direct targets of YTHDF1 (Fig. 4D). Furthermore, we found that these genes were also enriched in Hippo signaling, WNT signaling, and in cells responding to human papillomavirus (HPV) infection (Fig. 4E). To acquire the core genes mediated by YTHDF1, we queried the genes involved in the top six signaling pathways associated with GC. These efforts identified eight (8) genes as core components within these pathways (Fig. 4F).

Among these core genes, we noted that the transcription factor TCF7 was associated with all six signaling pathways and, thus, TCF7 was selected for subsequent analyses. We found that TCF7 has potential m${}^6$A methylation site (Fig. 5A,B). In addition, m${}^6$A peak enrichment was noted in TCF7 mRNA (Fig. 5C). This led us to speculate that YTHDF1 can bind to mRNA of TCF7 through its m${}^6$A modification, thereby affecting TCF7 expression. To further explore the regulatory mechanism of YTHDF1-mediated YCF7 control we conducted a RNA immunoprecipitation (RIP) assay. RIP assay revealed that YTHDF1 co-precipitates with TCF7 mRNA (Fig. 5D).

Fig. 5.

YTHDF1 regulate the translation efficiency of TCF7, thereby affecting WNT pathway. (A) An online database (https://www.cuilab.cn/) was used to visualize the predicted m${}^6$A modification sites within the TCF7 mRNA and the motif of m${}^6$A modification. (B) The combined scores between YTHDF1 and potential m${}^6$A modification site of TCF7 mRNA was predicted using SRAMP (https://www.cuilab.cn/sramp). (C) m${}^6$A peak enrichment of TCF7 mRNA. (D) RIP assay was conducted to prove that YTHDF1 binds to TCF7 mRNA. (E) The level of TCF7 mRNA was determined after YTHDF1 knockdown using siRNA. (F) A heatmap was used to display the changes in TCF7 target gene mRNA levels after TCF7 knockdown. (G) Effect of YTHDF1 knockdown on TCF7 protein level in GC cells. (H) The expression of TCF7 and its target genes after YTHDF1 knockdown was determined using western blot analysis. **, p 0.01. TCF7, transcription factor 7; MYC, MYC proto-oncogene bHLH transcription factor; MMP7, matrix metalloproteinase 7.

Further, we also observed that the mRNA level of TCF7 was not altered after YTHDF1 knockdown (Fig. 5E), and the TCF7 protein level was elevated after overexpression of recombinant YTHDF1 as judged by western blotting (Fig. 5F). Of note, we observed TCF7 protein level downregulation following YTHDF1 knockdown (Fig. 5G).

We also examined the effect of TCF7 on the WNT pathway. After knockdown of TCF7 expression, the mRNA expression levels of WNT downstream target genes, including c-myc, Cyclin D1, Survivin, MMP7 and CD44, were decreased (Fig. 5H). Consistent with these findings, YTHDF1 knockdown in MGC803 and AGS cells resulted in decreased expression of TCF7 protein as were the levels of Cyclin D1, MMP7, and MYC (Fig. 5G). Taken together, these results suggest that YTHDF1 activates the WNT pathway by upregulating TCF7, and that this molecular event likely promotes GC development and/or progression.

To directly verify that YTHDF1 serves in a tumor promoting role in vivo, MGC803 cells with stable knockdown of YTHDF1 were injected into 4-weeks female nude mice. These MGC803 knockdown cells were engineered by transducing cells with lentivirus encoding a YTHDF1-specific shRNA or normal control (NC) shRNA sequences. Tumor volume was measured every 7 days and after 28 days, we found that the tumor size in the YTHDF1 knockdown tumors was significantly smaller than those in the NC group (Fig. 6A) and that this was observed in four measurements (Fig. 6B). Tumor weight in the YTHDF1 knockdown group was also markedly lower compared with the NC group (Fig. 6C). In addition, compared with the NC group, expression of YTHDF1 and the proliferation marker Ki-67 were significantly lower in the YTHDF1 knockdown group as indicated by IHC staining (Fig. 6D). The protein level of YTHDF1, as well as its target gene TCF7, were also found to be lower in the YTHDF1 knockdown group than that of NC group as measured by western blotting (Fig. 6E). Taken together, these results strongly suggest that YTHDF1 contributes to GC malignancy both in vitro and in vivo.

Fig. 6.

YTHDF1 promote GC growth both in vivo. (A) Knockdown of YTHDF1 effectively inhibits GC tumor growth in vivo. (B,C) Tumor volume and tumor weight was decreased in the YTHDF1 knockdown group when compared to the normal control group. (D) Expression of YTHDF1 and Ki-67 in NC and YTHDF1 knockdown tumors was determined by IHC staining (scale bar = 200 µm). (E) Western blotting was performed to evaluate the protein level of YTHDF1 and TCF7 after YTHDF1 knockdown. *, p 0.05. IHC, immunohistochemistry; shRNA, small hairpin RNA.