High expression of WTAP leads to poor prognosis of gastric cancer by influencing tumour-associated T lymphocyte infiltration.

BACKGROUND
N6-methyladenosine (m6A) methylation, a well-known modification with new epigenetic functions, has been reported to participate in gastric cancer (GC) tumourigenesis, providing novel insights into the molecular pathogenesis of GC. However, the involvement of Wilms' tumour 1-associated protein (WTAP), a key component of m6A methylation, in GC progression is controversial. Here, we investigated the biological role and underlying mechanism of WTAP in GC.


METHODS
We determined WTAP expression using tissue microarrays and The Cancer Genome Atlas (TCGA) data set, which was used to construct co-expression networks by weighted gene co-expression network analysis (WGCNA). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed by Database for Annotation, Visualization and Integrated Discovery (DAVID). CIBERSORT was used to determine WTAP expression in 22 immune cell types.


RESULTS
Wilms' tumour 1-associated protein was highly expressed in GC, which indicated a poor prognosis, and WTAP expression served as an independent predictor of GC survival. By WGCNA, GO, KEGG and core gene survival analyses, we found that high WTAP expression correlated with RNA methylation and that low expression correlated with a high T cell-related immune response. CIBERSORT was used to correlate low WTAP expression with T lymphocyte infiltration.


CONCLUSION
RNA methylation and lymphocyte infiltration are the main causes of high WTAP expression and poor prognosis, respectively.


| INTRODUC TI ON
Gastric cancer (GC) is one of the most common malignant tumours of the digestive system, and it has a high proportion of malignant tumour-related mortality worldwide, especially in China. According to statistical analysis, GC was ranked 5th in the global incidence of malignant tumours in 2018, but the mortality rate was ranked 3rd. 1 Research suggests that the causes of GC are related to genetic factors, stomach diseases, Helicobacter pylori and lifestyle. 2 Although the treatment of GC has made great progress in recent decades, ranging from interventional therapy or radical resection to targeted therapy or immunotherapy, the treatment results of GC are still not ideal. 3 Therefore, it is necessary to further clarify the molecular mechanism of GC to develop new therapeutic strategies to reduce the mortality of this malignant tumour.
N6-methyladenosine (m6A) methylation modifications are one of the most common methylation modifications in eukaryotes. It accounts for more than 80% of RNA methylation, and its modification site always appears in the conserved sequence RRACH (R = G or A, H = A, C or U). 4 The mammalian Wilms' tumour 1-associating protein (WTAP) is the first nuclear protein associated with the Wilms' tumour 1 inhibitor gene WT1, discovered by Little et al. 5 Wilms' tumour 1-associating protein (WTAP) is a component of the m6A methyltransferase complex that recruits the m6A methyltransferases METTI3 and METTL14 to the corresponding mRNA targets to co-catalyse the formation of m6A. 6 Deregulation of m6A pathway components can affect oncogenic expression, thereby affecting tumourigenesis. 7 Since most studies have focused on the intrinsic carcinogenic pathways of tumours, the potential role of mRNA m6A modification in host antitumour immune responses remains unclear.
Dali Han et al studied the mechanism of the antitumour effect of the mRNA m6A methylation gene YTHDF1 and found that dendritic cells regulate the methylation of mRNA m6A through YTHDF1 and thus play a role in antitumour immunity. 8 WTAP is a component of the m6A methyltransferase complex, and the potential role of WTAP in host antitumour immune responses is unclear, so we need to explore it further. To this end, we studied the expression of WTAP in GC tissue, the effect of WTAP expression on tumour immune cell infiltration and patient prognosis, and then explored the mechanism by weighted gene co-expression network analysis (WGCNA).

| Study subjects
A total of 14 GC patients were recruited from the First Affiliated Hospital of Sun Yat-sen University, including 9 males and 5 females.
The average age was 54.93 ± 7.42 years old, and the age range was 44-68 years old. The study was approved by the Ethics Committee of the First Affiliated Hospital of Sun Yat-sen University. The samples were all obtained with the informed consent of the patients.
Surgical pathological staging criteria were in accordance with the International Disease for Oncology (ICD-O): four cases in stage I and stage II, 10 cases in stage III and stage IV; degree of differentiation: three cases of moderate differentiation, two cases of poor differentiation, seven cases of medium-low differentiation and two cases of high differentiation (see Table 1).  Table 1).

| RNA extraction and RT-qPCR
The mRNA was extracted by TRIzol homogenate from each cancer tissue and its corresponding adjacent tissues, and the concentration was determined. Five micrograms of each sample RNA, 1 µL oligo d(T), 1 µL dNTP and an appropriate amount of DEPC·H2O were added to a total volume of 12 µL and mixed well. The whole system was bathed in a 65°C water bath for 10 minutes, and 4 µL 5x firststrand buffer, 2 µL 1 µmol/L DTT, 1 µL RNase In and 1 µL M-MLV were added. The reverse transcriptase was then inactivated by placing in a 75°C water bath for 5 minutes. The cDNA was amplified using primers. The PCR conditions were as follows: pre-denaturation  Table 2).

| WGCNA co-expression network construction
Gene expression data (mRNA-seq data) were obtained from the TCGA database. A total of 24 991 genes were identified from each sample. Analysis of variance was performed, and the data were sorted from large to small. We calculated the standard deviation values for each gene, sorted them from large to small and then selected the top 5000 genes for WGCNA. The expression data map of these 5000 genes was constructed into a gene co-expression network using the WGCNA package in R software. 9 Using the WGCNA function adjacency, an adjacency matrix is constructed by computing the Pearson correlation between all pairs of genes in the selected sample. In this study, β = 7 (no scale R 2 = .9) was used as a soft threshold parameter to ensure a scale-free network. To further identify the functional blocks in the co-expression network of the 5000 genes, a topological overlap measure (TOM) is computed using the adjacency matrix, which represents the overlap in the shared neighbourhood.

| Identification of clinically significant modules
We

| PPI network construction of key module genes
The hub gene, which is highly interconnected with the nodes in the module, is considered to have important functions. We selected the top 30 hub genes in the module network as candidate genes for further analysis and validation. The STRING data set (https://strin g-db.org/) is an online biological resource that decodes the interaction between proteins to obtain the true functionality of real proteins. 10 The candidate gene was submitted to STRING for protein interaction, and the confidence interval for the cut-off value was set to 0.4. In the Plugin Molecular Complex Detection (MCODE), significance models with strong protein-protein linkages were calculated and selected with default parameters (degree cut ≥ 2, node score cut ≥ 2, K-core ≥ 2, maximum depth = 100). The difference was statistically significant at P < .05.

| Gene Ontology (GO) and pathway enrichment analysis
The

| Survival analysis of hub genes
Kaplan-Meier's plotted network (http://kmplot.com/analy sis/) is a platform containing expression data for 10 tumour genes and clinical survival data for 1065 patients with GC. We used this website to obtain information on core gene expression and patient survival prognosis, which in turn helped us identify core genes that influence the survival of WTAP high-and low-expression groups. 12 To assess the prognostic value of a particular gene, patient samples

| Statistical analyses
The difference in WTAP expression between GC tissues and ad- The P-value was bilateral, and P < .05 was considered statistically significant.

| WTAP was dysregulated in GC
To elucidate the role of WTAP, we first analysed the mRNA expression of WTAP in human GC samples from TCGA data. The results showed that WTAP expression in tumour tissues was significantly increased ( Figure 1A). WTAP expression levels were also significantly up-regulated in GC tissues of our central patients (Figure 1 B).
In addition, we also studied the relationship between WTAP expression and survival prognosis in 318 patients with GC. WTAP expression was significantly associated with survival outcomes in patients ( Figure 2).

| Construction of weighted co-expression network and identification of key modules
To construct a gene co-expression network, the raw data of GC were downloaded from the TCGA database. The background correction and normalization were performed using R, and the same preprocessing was performed on the original data. R-pack annotation matching was performed on the probe and the gene symbol, and the probe matching the plurality of genes was removed. For the plurality of probe-matched genes, the median was taken as the final expression value. Finally, we obtained a total of 24 991 genes.
We calculated the standard deviation value for each gene, sorted the values from large to small and then selected the top 5000 genes for WGCNA. Cluster analysis of 5000 genes was performed using the fashClust function of the WGCNA package ( Figure 3A).
The selection of soft threshold power is an important step in constructing WGCNA. The network topology of 1 ~ 20 threshold weights was analysed, and the scale independence and average connectivity of WGCNA relative equilibrium were determined. As shown in Figure 3B,C, a power value of 9 was selected as the lowest power (0.9) of the scale-free topology ft index, and a hierarchical clustering tree (dendrogram) of 5000 genes was generated.
We set MEDissThres to 0.25 to merge similar modules ( Figure 4A) and generated 65 modules ( Figure 4B). The gene statistics in each module are shown in Table 3. Genes that cannot be included in any module were added to the grey module and rejected in subsequent analyses.

| Correlation between modules and identification of key modules
We analysed the interaction between the 65 modules and plotted the network heat map ( Figure 5A). The results show that each module is independent of each other, each module has high independence, and the gene expression of each module is relatively independent. In addition, we calculated the characteristic genes and clustered them according to their correlation to explore the co-expression similarity of all modules ( Figure 5B). We found that these 65 modules are mainly divided into two clusters. A heat map drawn from the adjacency relationship shows similar results ( Figure 5C). The Salmon module was positively correlated with low WTAP expression, while the dark orange module was positively correlated with high WTAP expression. Figure 6A,B shows the relationship between the number of module members and the GS in the Salmon module and the dark orange module, respectively.

| Identification of hub genes in the Salmon module and dark orange module
We submitted the gene set of the Salmon module and dark orange module to STRING protein interaction analysis, and the cut-off confidence interval was set to 0.4. In the Plugin MCODE, significant models with strong protein-protein linkages were calculated and selected with default parameters (degree cut ≥2, node score cut ≥2, K-core ≥2, maximum depth = 100). The difference was statistically significant at P < .05. The core genes were screened for further analysis by sorting the node degree candidate genes.

| Functional enrichment analysis in the two key modules
To study the roles of the core genes in these two key modules, we  memory-activated T cells can improve patient outcomes ( Figure S7).

| D ISCUSS I ON
m6A RNA modification is a hotspot in the field of regulation in recent years, involving multiple cellular processes such as mRNA maturation, protein translation and molecular structure transformation. 15 There is growing evidence that m6A dysregulation has a profound impact on the pathogenesis of many diseases, including GC. 16 We examined the expression of WTAP in GC samples and found that WTAP expression in tumour tissues was higher than that in adjacent tissues. The data from the TCGA database also confirmed our conclusions. To investigate whether the expression of WTAP has an effect on the prognosis of patients, we found that patients with high expression of WTAP have a poor prognosis and poor patient expression. This shows that the impact of WTAP in GC is distinctive.
WGCNA is a method for constructing gene co-expression networks based on gene expression data. 17 To explore the molecular mechanisms that influence the effect of WTAP expression on prognosis, we used WGCNA to find the core co-expressed genes.  Table 4). The expression of these genes is clearly related to the prognosis of patients. Through enrichment analysis, these genes were found to up-regulate the methylation of mRNA. Studies have found that m6A RNA modifications can affect tumour proliferation and patient prognosis through immunoregulatory effects. 18 Our study also found that T regulatory cells (Treg) and CD4 memory-activated T cells in patients with high WTAP expression were significantly lower than those in patients with low WTAP expression.
There is a clear correlation between the infiltration of these cells and the prognosis of patients. This suggests that tumour immune regulation may be an important cause of poor prognosis in WTAP.
In fact, WTAP overexpression is an important risk factor in many tumours. 19 WTAP is widely expressed in various tissues and plays an important role in cell cycle regulation, RNA alternative splicing, m6A methylation modification, x-chromosome inactivation, eye development, regulation of physiological balance and other physiological processes. 20 The role of WTAP in GC is currently rarely reported and controversial. Zhang C et al found that low m6A levels can lead to the proliferation of GC cells, resulting in GC progression. 21 Wang et al studied the related effects of m6A methylation on the prognosis of GC and found that m6A methylation can lead to the progression of GC and can lead to poor prognosis of GC patients. 22 We studied the effect of WTAP on the prognosis of GC and found that high WTAP expression was associated with poor prognosis of patients. To explore this cause, our study found for the first time that and T-cell cross-priming. 24 Multiple studies have shown that type I interferons (IFNs) play an important role in tumour control by promoting the reactivation (reactivation) of T cells by dendritic cells (DCs). 25 Our study found that high expression of WTAP can reduce T-cell infiltration and inhibit tumour immunity, which may be supported by the above studies. Liang et al found that IFNs may be an important target for PDL1 immunotherapy, and they also showed that methylation genes are a potential target for this research. 26 The study of methylation genes in immune regulation is currently a hot issue, and WTAP plays an indelible role as an important member of the methylation genes. 27 However, WTAP is poorly   It recruits the m6A transferases METTI3 and METTL14 to the corresponding mRNA targets for co-catalysis and m6A formation. 43 It is present in ribosomal RNA (rRNA), transport RNA (tRNA), messenger RNA (mRNA) and non-coding RNA (ncRNA). m6A is the most widely distributed methylation modification in eukaryotic mRNA, and its formation may modulate a series of processes after transcription, such as splicing, transport, degradation and translation of pre-mRNA. The formation of m6A is catalysed by a large methyltransferase complex. 44 Our study also found that high WTAP expression in patients can affect the splicing, transport, degradation and translation of mRNAs by interacting with the DLX2, DLX5, SIX1, HOXB5, HOXC6, HOXC8, RBM48 and KRAS core genes.

| CON CLUS ION
Through this study, it was found that the expression of WTAP was significantly correlated with the survival prognosis of patients. To determine the mechanism, we used WCGNA and enrichment analysis to confirm that high WTAP expression is associated with RNA expression, while low WTAP expression is associated with lymphocyte infiltration. This is also the main cause of high WTAP expression and poor prognosis. Further, we found that lymphocyte infiltration in patients with low WTAP expression has a good correlation with patient prognosis. However, this study has certain limitations and deficiencies. First, due to the lack of certain data in the TCGA database, this study did not provide a good analysis of clinical parameters and prognosis. Second, we only analysed the transcriptome levels from patients and did not perform further in vivo and in vitro experiments. Further research is needed to support our conclusions.

CO N FLI C T O F I NTE R E S T
All authors declare that they have no conflicts of interest.

AUTH O R CO NTR I B UTI O N S
CZ and HL conceived and designed the study. CW, WC, LW, GL and LL performed the data analysis. HL, QS and BL wrote the paper. All authors read and approved the manuscript.

E TH I C S A PPROVA L A N D CO N S E NT TO PA RTI CI PATE
Our research was approved by the ethics committee of the First Affiliated Hospital of Sun Yat-Sen University, and we obtained written informed consent from all these patients.

DATA AVA I L A B I L I T Y S TAT E M E N T
The authors confirm that the data supporting the findings of this study are available within the article and its supplementary.