The high incidence and mortality rate of GC seriously endangers human health. Surgery is considered to be the only radical cure for GC, but is prone to result in anastomotic leakage, intestinal obstruction, early recurrence and other serious complications, which seriously worsens the patient prognosis and reduces survival rates. m6A modification is accurately regulated by “writers”, “erasers”, and “readers”, which are involved in several mRNA metabolism pathways. Additionally, m6A modifications influence the processing of lncRNA. m6A modifications regulate cellular proliferation and maturation processes, both of which are linked with cancer development. Consequently, the regulation of m6A modification in cancer cells could have a profound impact on the progression of malignant tumor research [13]. An increasing number of studies have revealed the pathological significance of m6A in cancers [14, 15]. LncRNA is a major class of noncoding RNAs. LncRNA plays vital roles in chromatin organization and transcriptional and posttranscriptional regulation [16]. Previous studies have mainly focused on the correlation between specific m6A-related genes or pathways and the diagnosis and treatment of tumors; the lack of systematic analysis of m6A-related lncRNA in gastric cancer ought to be addressed. Therefore, recognition and analysis of m6A-related lncRNA in large cohorts of GC patients are of considerable significance, guiding latent directions and targets for GC research.

In the current study, we extracted m6A-related gene expression data and identified mRNAs and lncRNAs. Additionally, co-expression analysis was carried out to determine the correlation between m6A-related gene expression and lncRNAs. As shown in the co-expression network plot, we found an interesting phenomenon in which several lncRNAs were linked with m6A-related genes in GC. This finding fueled our interest in the expression of m6A-related lncRNAs and their related functions in GC. Prognosis-related lncRNAs were identified and the confidence interval and hazard ratio were calculated. Univariate Cox regression analysis indicated that m6A-related lncRNAs were closely related to the prognosis of GC. Cancers have many essential links with m6A modifications. m6A is considered to influence lncRNA splicing, which might alter cancer progression [17, 18]. There were 17 m6A prognosis-related lncRNAs in our study, the expression of which was different between tumor and normal tissues. Some lncRNAs were highly expressed in tumor tissues, while others were highly expressed in normal tissues (P < 0.05). It has been reported that HBXIP is upregulated in cancers, which plays a role as a tumor promoter in cancer via driving metabolic reprogramming through METTL3-mediated m6A modification [19]. Lately, it has been shown that HBXIP might promote the development of gastric cancer, which is m6A-modified [20]. It was recently identified that lncRNA RP11 expression is upregulated in gastric cancer and functions by promoting migration and invasion via epithelial–mesenchymal transition [16]. m6A modification promotes the upregulation of RP11 expression in GC cells by enhancing epigenetic etiology and pathogenesis of GC [21]. The modification of lncRNA may act an essential role in protein interactome and influence the progress of cancer [22]. The above findings might explain the result of our study in which we found that some m6A lncRNAs are overexpressed in tumors, while others are highly overexpressed in normal tissue. In addition, m6A lncRNAs may act as oncogenes or tumor suppressors.

We further explored the role of m6A lncRNA in GC. Survival analysis according to subtypes of lncRNAs was conducted to evaluate the prognostic value of m6A lncRNAs. Low-risk lncRNAs are beneficial to the prognosis of GC. In addition, lncRNAs are closely related to the survival rate of GC patients. The results of our study are consistent with the conclusion of Wang et al. that m6A-induced lncRNA RP11 expression triggers the malignancy and immunosuppression of gastric cancer cells via upregulation of YAP1 expression [16]. Additionally, Yang et al. [20] showed that the long noncoding RNA HBXIP promotes progression of gastric cancer through METTL3-mediated m6A modification of HIF-1α, also supporting that m6A lncRNA is closely related to the prognosis of GC. The expression of m6A lncRNAs in different clusters was not different, possibly because most of the m6A lncRNAs were expressed at low levels in GC in our study. In addition, studies on m6A modification of lncRNAs are still small in number. Thus, there is an urgent need for further research on lncRNA m6A modification and recognition to validate our results.

Gene ACBD3-AS1(ACBD3 antisense RNA 1) is a member of the lncRNAs located at Chromosome 1: 226,148,003–226,155,071 forward strand, which is ubiquitous expression in stomach, colon and 25 other tissues. Overexpression of ACBD3-AS1 increased feasibility and inhibited apoptosis via stimulating the expression of apoptosis related genes. Research illustrated that overexpression of ACBD3-AS1 launched accumulation of JAK2, indicating potential comic dialog between ACBD3-AS1 and the JAK2 signaling pathway, which revealed that this transcription factor could unleash tumor booster properties in gastric carcinoma [23]. In our study, the expression of ACBD3-AS1 is higher in GC tumor sample, indicating that it might be an oncogene. Additionally, the expression of ACBD3-AS1 was higher in cluster 2, which indicates it may be harmful to prognosis of gastric cancer. Its highly expression in various GC cell types might explain this result. As shown in the analysis of the correlation between the target genes and prognostic m6A lncRNAs in GC, ACBD3-AS1 is closely associated with several m6A lncRNAs. In addition, ACBD3-AS1 was most positively correlated with AC092119.2, AC007038.1, AL139287.1, SNHG12 and C3orf35. With the increased expression of the above m6A lncRNAs in GC cells, the expression of ACBD3-AS1 is increased. This phenomenon further supports our hypothesis that ACBD3-AS1 may be an oncogene. Additionally, the m6A lncRNAs might be therapeutic targets for GC. This is consistent with the result that under the condition of SNHG12 addition, gastric cancer cell proliferation, migration and invasion were notably heightened and cell apoptosis was lessened to accelerate the malignant progression of GC by activating the phosphatidylinositol 3-kinase/AKT pathway [24]. However, the specific molecular mechanism leading to tumorigenesis needs further research to clarify.

Moreover, we explored and calculated the infiltration of different immune cells in the samples to identify the role of immune cell infiltration and the tumor microenvironment in GC. Differential analysis of immune cell infiltration indicated that immune cells such as naïve B cell, Plasma cells, resting CD4 memory T cell were enriched in cluster 2, while Macrophages M2, resting Mast cells, Monocytes, regulates T cells were highly clustered in cluster 1. As stated before, cluster 2 represents high risk GC and poor prognoses. Therefore, the infiltration of naïve B cell, Plasma cells, resting CD4 memory T cell in the tumor microenvironment may be harmful to patient prognosis. Our conclusion is in line with the conclusion of Zhao et al., who found that naïve B cell in cancer tissue correlated with tumor metastases and fully functional regulatory activity against human gastric cancer immunity [25]. Wu et al. suggested that resting CD4 memory T cell is unable to mount sufficient cytotoxic activity against GC cells, indicating that resting CD4 memory T cell is detrimental to GC patients [26]. Differential analysis of the tumor microenvironment in different subtypes was conducted to further explore the purity of tumor cells in the different clusters. All of the scores were higher in cluster 1, indicating lower tumor cell purity and more immune-related cells in the tumor microenvironment of cluster 1. This result is consistent with the conclusion that cluster 1 is low-risk and beneficial for patients. The results of Kemi encourage the use of immune cell score analysis to predict the prognosis of gastric cancer and high score improves the 5-survival rate of GC patients [27]. Research suggested that the immune microenvironments of metastatic tumors was less immunologically active compared to that of primary tumors in gastric cancer patients, which might help establish reliable prognostic signatures due to assessments of stromal and immune components [28]. The results of both studies support our hypothesis that immune cell infiltration in the tumor microenvironment influences the prognosis of GC patients. The higher the immune score is, the lower the purity of the tumor and eventually the better the prognosis.

Next, we carried out GSEA. “ADIPPOCYTOKINE SIGNALING PATHWAY” was the most significantly enriched gene set. Regrettably, there are little researches about this signaling pathway in gastric cancer. As a result, it may be our next research direction to further explore the diagnosis and therapy of GC. Taking into account the above factors, m6A lncRNA may play its function by regulating the “ADIPPOCYTOKINE SIGNALING PATHWAY” to influence the migration and proliferation of GC cells. An m6A lncRNA-related prognostic model was constructed via lasso regression. In both the test group and the training group, the survival rate of the low-risk subtype was higher than that of the high-risk subtype. An m6A lncRNA-related prognostic model can therefore predict the outcome of GC. Additionally, the accuracy of our model for predicting the survival of patients with the disease is considerable. As the risk score increased, the number of deaths increased, and the ratio of high risk increased. Furthermore, our model was independent of other clinical prognostic factors that affect patient outcomes. Also, the model could be applied to different clinical groups. m6A modification of lncRNAs may change the structure of lncRNAs and affect their interaction with proteins, which may mediate gene transcription repression [29, 30]. m6A modification of lncRNAs possibly alters their subcellular dissemination, which regulates lncRNA stability and promotes tumorigenesis and metastasis [30]. In summary, the literature and our research results identified that m6A lncRNA could be a suitable clinical model to predict the outcome of GC. The results of genetic differential analysis indicated that the expression of the ACBD3-AS1 gene was higher in the high-risk group in our model in gastric cancer, which further confirmed that ACBD3-AS1 may be an oncogene in GC. As there are few studies on the ACBD3-AS1 gene and no related study has explored the role of ACBD3-AS1 in GC cells, the results need further study for clarification. Correlation analysis of risk and immune cells was conducted to evaluate the relationship among immune cells and the risk score. Macrophages M2 was positively correlated with the risk score. The conclusion of Takahisa clarified our result that M2 phenotype could contribute to tumor progression and is expected to be a promising target in the treatment of gastric cancer [31]. This finding is consistent with the results of the differential analysis of immune cell infiltration in the different clusters. Resting mast cells might disturb the anti-tumor immune response, and the function might differ depending on its composition via working together with BicC family RNA-binding protein 1 [32].