The mRNA level of WTAP in HCC was analyzed using the Tumor Immune Estimation Resource (TIMER) database (https://cistrome.shinyapps.io/timer/) and Genomic Data Commons (https://portal.gdc.cancer.gov/). Data of 474 RNA sequences and relevant clinical information of patients with HCC were retrieved from the HCC project of TCGA. Unavailable or unknown clinical data were considered a missing value.

Tumor samples were further stratified into a high or low group in accordance with the median WATP mRNA expression level. All RNA-sequencing (RNA-Seq) data were converted to the transcripts per million reads format. After log2 transformation, two sets of t-test were implemented, and statistical significance was fixed at P < 0.05. Data analysis was assisted by the R program (version 3.6.3), and the R package “ggplot2” was used to draw box plots.

Tissue microarrays were purchased from Chaoxing Biotechnology (Shanghai, China). The tissue microarrays contained 180 samples including liver tumor and healthy adjacent tissues. All samples had detailed clinical information including sex, prognosis, age, and grade. Immunohistochemical staining was carried out in strict compliance to instructions stipulated by the manufacturer. Nuclear and cytoplasmic staining intensities were graded based on the following scale: 0, no staining; 1, weak; 2, medium; 3, strong. The level of staining was graded from 0% to 100%. Scores of 160 or lower indicated a low WTAP level and scores of 161 or higher indicated a high WTAP level.

Immunohistochemistry images of WTAP expression in HCC normal tissues and tumor tissues were gathered on the website The Human Protein Atlas (http://www.proteinatlas.org/) and analyzed.

According to the median expression of WTAP, all participants were stratified according to low or high WTAP expression. To assess the impact of WTAP expression on the clinical outcomes of patients with HCC, we constructed a prognostic classifier using Kaplan-Meier survival curves and compared the survival disparities using R packages “survminer” and “sunival”. To verify the results, we further constructed Kaplan-Meier survival curves using the data extracted from the TIMER and Gene Expression Profiling Interactive Analysis databases (http://gepia.cancer-pku.cn/).

Univariate logistic regression analysis to delineate whether the WTAP affected the overall survival of patients in the two cohorts has been performed. A multivariate Cox regression analysis was utilized to determine whether WTAP is an independent prognostic factor for the survival of patients with HCC. Results with P < 0.05 were considered as statistically significant. Further, a receiver operating characteristic curve generated using the pROC package was also utilized to assess whether WTAP has a prognostic value in patients with HCC.

To screen for proteins interacting with WTAP, genes that were co-expressed with WTAP were verified by the cBioPortal database (https://www.cbioportal.org/), and the top 50 of genes are shown in a heatmap. The top 10 significant genes that correlated with WTAP were subjected to additional analysis using the TIMER database to determine if they were related to WTAP expression.

PPI analysis was carried out by GeneMANIA (http://genemania.org/), and the results were confirmed by The Search Tool for the Retrieval of Interacting Genes/Proteins website (https://string-db.org/). By importing the WTAP data into this online tool, we obtained the PPI network information and collected the top 10 interacting proteins. The confidence scores of these proteins were >0.9. Scores of >0.7 were considered significant.

After identifying the genes co-expressed with WTAP using the cBioPortal database, we imported them into the online analysis tool Hiplot (https://hiplot.com.cn) and performed enrichment analyses in terms of cellular components, molecular functions, biological processes, and Kyoto Encyclopedia of Genes and Genomes (KEGG).

Correlation analysis of WTAP expression level and immune-infiltrating cells was using the TIMER database. Using the “GSVA” package of R software, Single-Sample Gene Set Enrichment Analysis (ssGSEA) was performed. The co-expression analysis of WTAP and immune-related genes also have been performed, including genes encoding immune checkpoints, major histocompatibility complexes, chemokines, chemokine receptors, and immune activation and immunosuppression molecules.

The human HCC cell lines used in this study were stored in Zhuhai Precision Medical Center. In a 37°C incubator containing 5% CO₂, MHCC97-H and SMCC-7721 were incubated in Dulbecco’s modified Eagle medium (DMEM; Gibco, MA, USA) or RPMI 1640 (Gibco, MA, USA) that contained 10% fetal bovine serum. The HCC cells were transfected with siRNA (RiboBio, Guangzhou, China) using Lipofectamine 3000 Reagent (Invitrogen, CA, USA) to produce a working concentration of 100 nM, following to the protocols of the manufacturer. The sequence of siRNAs used is as follows: siWTAP#1:5’-GGAACAGACTAAAGACAAA-3’; and siWTAP#2:5’-CTAAGAGAGTCTGAAGAAA-3’.

The RNA-Quick Purification Kit (ES Science, Shanghai, China) was used to isolate the RNA from cells. The Revert Aid First Strand cDNA Synthesis Kit (Thermo, MA, USA) was used to obtain cDNA from RNA. SYBR Green Master Mix II was used for RT-qPCR analysis, with GAPDH as the internal control. The 2^-ΔΔCT approach was used to analyze the data. The sequence of the PCR primer used was as follows: WTAP-forward: 5’-GCAACAACAGCAGGAGTCTGCA-3’, WTAP-reverse: 5’-CTGCTGGACTTGCTTGAGGTAC-3’.

In 96-well plates, 800 cells per well were incubated with the indicated reagents. Thereafter, 20 µL CellTiter 96 AQueous One Solution (Promega, WI, USA) was added to each well. The mixture was incubated for 3 h. Optical density was measured at 490 nm.

500 treated cells were coated into 6-well plates with three repetitions. After incubation for 14 days, these plates were washed with PBS twice, fixed by 4% polymethanol for 10 min and stained with 0.1% crystal violet solution within 10 min. A cluster of 50 cells was defined as a colony.

To examine cell migration, 1 × 10⁵ treated cells mixed in 100 µL of serum-free DMEM or 1640 medium were placed in the upper Transwell chamber. Thereafter, 500 µL of the complete culture medium was placed in the lower chambers. After incubating for 24 h, cells that migrated were fixed for 30 min with 4% paraformaldehyde and subjected to 20 min of 0.1% crystal violet staining, whereas non-migrated cells were remove with a cotton swab. The wells were then washed with water and air-dried before quantifying the total migrated cells.

Data were analyzed using SPSS 23.0 and R version 3.6.1. Data are presented as mean ± SD. Intergroup differences were established using independent and paired sample t-tests. Logistic and Cox regression models were used to analyze the relationship between WTAP expression and clinical characteristics. Statistical significance was set at P < 0.05.

RNA-Seq data of various human cancers from the TIMER database were analyzed for their WTAP transcription level ( Figure 1A ). HCC samples presented considerably increased levels of WTAP mRNA compared with healthy liver samples. WTAP was also highly expressed in the tissues of the following cancer types: colon adenocarcinoma, cholangiocarcinoma, glioblastoma multiforme, head and neck squamous cell carcinoma, lung squamous cell carcinoma, esophageal carcinoma, kidney renal clear cell carcinoma, and stomach adenocarcinoma. In contrast, WTAP expression was downregulated in breast invasive carcinoma, kidney renal papillary cell carcinoma, kidney chromophobe, bladder urothelial carcinoma, pancreatic adenocarcinoma, thyroid carcinoma, lung adenocarcinoma, and uterine corpus endometrial carcinoma samples compared with that in non-tumor tissues. We conclude that WTAP is aberrantly expressed across several tumors. We then investigated WTAP expression in the TCGA datasets. The results show that WTAP was highly expressed in both unpaired and paired tumor samples with HCC ( Figure 1B, C ).

At the protein level, WTAP was expressed at moderate to high levels in HCC tissues and low level in normal liver tissues as evidenced by immunohistochemistry staining based on The Human Protein Atlas database ( Figure 1D ). To confirm the above result, we used immunohistochemical staining to detect WTAP level in 90 pairs of HCC and the adjacent normal tissue. HCC tumor tissue presented higher WTAP levels, in both cytoplasm and nucleus, compared with healthy liver samples ( Figure 1E-G ).

The clinicopathological characteristics of the 90 patients stratified by low or high WTAP expression are summarized in Table 1 . We found a significant association between low cytoplasmic WTAP expression and tumor grade (P = 0.04). Conversely, nuclear WTAP level did not correlate with tumor grade ( Table 2 ). Moreover, WTAP expression did not appear to be related to other clinical parameters such as sex, age, tumor size, T stage, and TNM stage.

Next, the prognostic value of WTAP expression in HCC was investigated. As indicated in Figure 2A , a higher WTAP expression was significantly associated with shorter overall survival based on the TCGA data (hazard ratio = 1.70, P = 0.003). These results were verified using TIMER2.0 and GPEIA, and similar results were obtained ( Figures 2B, C ). We then analyzed the prognostic data of tissue microarrays. The WTAP level in the cytoplasm and nucleus negatively corelated with survival time ( Figures 2D, E ). Then, the univariate and multivariate analysis results demonstrated that the WTAP expression level is an independent predictor of unfavorable prognosis in patients with HCC ( Tables 3 and 4 ). Furthermore, the result of receiver operating characteristic curves (ROC)obtained an area under the curve value of 0.731 (confidence interval: 0.670–0.792), which representing good performance of ROC curves ( Figures 2F ).

We identified the genes that were co-expressed with WTAP through data mining. The top 10 significant genes correlated with WTAP were TBP, PPIL4, ZUP1, VTA1, FBXO5, HSF2, KATNA1, HDAC2, FBXO30, and PHF10 ( Figure 3A ). The relationships between WTAP and these genes was confirmed by the TIMER database ( Figure 3B ).

To elucidate the molecular mechanism underlying HCC malignancy and its progression, we constructed the protein interaction networks of the WTAP protein ( Figure 3C ). The top 10 proteins and their corresponding gene names, annotations, and scores are shown in Figure 3D . These genes included CBLL1, VRIMA, METTL14, METTL3, NCBP1, NCBP2, RBM15, RBM15B, ZC3H13, and ZNF645.

The molecular functions of WTAP in HCC was further analyzed using Gene Ontology and KEGG to predict WTAP-related signaling pathways. The most enriched terms were “cellular developmental process”, “molecular function regulator”, and “nucleoplasm” in the biological processes (BP), molecular functions (MF), and cellular components (CC), respectively ( Figures 4A–C ). According to the results of KEGG analysis, the most relevant signaling pathways included “cytokine–cytokine receptor interaction”, “viral protein interaction with cytokine and cytokine receptor”, and “osteoclast differentiation” ( Figure 4D ). Moreover, WTAP closely correlated with the incidence of cell cycle and immune diseases such as inflammatory bowel disease and rheumatoid arthritis.

Additional analyses were performed using the TIMER database to elucidate the core action between WTAP expression and HCC immune cell infiltration. As shown in Figure 5A , WTAP expression strongly correlated to WTAP level of different infiltrating immune cell types, including neutrophils (P < 0.001), dendritic cells (P < 0.001), macrophages (P < 0.001), CD4+ T cells (< 0.001), B cells (P < 0.001), and CD8+ T cells (P < 0.001). Furthermore, the ssGSEA showed that WTAP expression positively correlated with tumor-infiltrating immune cells ( Figures 5B, C ), including T helper cells, Th2 cells, follicular helper T cells, macrophages, Th1 cells, eosinophils, central memory T cell, effective memory T cell, NK CD56 bright cells, B cells, and T cells. In contrast, WTAP expression negatively correlated with Th17 cells.

We hypothesized that the expression of WTAP may be associated with that of immune checkpoint genes. Heatmap and correlations between WTAP and immune checkpoints, including PD-1, PD-L1, PD-L2, CTLA4, LAG3, TIGIT, and HAVCR4, are shown in Figure 6A . there were significant and positive correlation between all these immune checkpoint genes and WTAP expression in HCC samples from the TCGA database. Then, the gene co-expression analysis showed that the expression of WTAP correlated with that of genes encoding immunosuppressive genes, immune activation genes, major histocompatibility complexes genes, chemokines receptor genes, and chemokines genes ( Figures 6B–F ). The resulting heatmap showed that many immune-related genes were co-expressed with WTAP, and the majority were positively correlated with WTAP, except for KIR2DL1, IL6R, CXCL2, CXCL17, CCL15, CCL25, and CCL27.

Transfection efficiency was examined ( Figure 7A ). WTAP knockdown considerably suppressed SMMC-7721 and MHCC97-H cell proliferation ( Figures 7B, C ) as evidenced by MTS and colony formation assay results. WTAP knockdown also considerably reduced HCC cell invasion and migration ( Figure 7D ).