Endometrial cancer ranks as the fifth most common cause of cancer death in the United States, increasingly, threatening female patients’ lives [29]. The discovery of objective and susceptible indicators is crucial to optimize clinical diagnosis and has instructive significance in judging the prognosis of EC accurately.
LncRNAs have been confirmed to play a key role in numerous pathological courses and remain stable in blood circulation, hence, lncRNAs are deemed biomarkers, offering guidance for therapy and determining eventual results in various carcinomas [30]. Although the abnormal expression of several lncRNAs in endometrial and ovarian cancer has been revealed in prior studies, realizing a comprehensive probe on this topic in gynecological cancers still needs long way off [31]. Pan et al. pointed out a two-way feedback ring composed of LINC01016 and miRNA that mediated cell biology phenotype transformation in EC [32]. Simultaneously, various lncRNAs have been indicated to participate in adjusting the immune system [33]. The immune response exerts a great impact on the pathophysiology and progression of solid tumors, including endometrial carcinoma. Likewise, immunotherapy, an innovative therapy modality with great promise, has recently been reported as a research hotspot [34]. Moreover, the immunoregulatory mechanism determines the occurrence and intensity of the immune response, and many transcription factors act as the main regulators to regulate the immune response process. Zaiss discussed the core role undertaken by Forkhead box transcription factors (FTFs) in the regulation of immune responses and homeostasis[19]. In addition, years of research have adequately established the immense modulatory potential of transcription factor activity as a trigger of cancer, such as the transcription factors SNAIL1 and ZEB1, mediating the epithelial-to-mesenchymal transition-related signaling pathway [35]. Kingwell, K et al. concluded that elF4F promotes the expression capacity of the transcription factor of STAT1, thus potentiating the immune escape of melanoma [36].
In this study, we constructed a network in EC by making use of the TCGA database to explore the interactions between DE TFs and immune-OS-related genes. There were 6267 differentially expressed genes between tumor tissues and normal tissues obtained from the TCGA. Afterwards, the network of 29 DE TFs and 17 immune-OS-related genes was established on the basis of the results from the coexpression analysis. Furthermore, by conducting enrichment analyses, we explored 102 DE TFs and 53 DE genes associated with OS and the immune system, which were both included in the 6267 DE genes. The results of GO functional analysis and KEGG enrichment analysis revealed that the DE genes are significantly clustered in the transcription factor complex, DNA-binding transcription activator activity, intracellular receptor signaling pathway, MAPK signaling pathway, PI3K-AKT signaling pathway, and PD-L1 and PD-1 checkpoint pathway, which also accounted for a major portion of the enrichment characteristics. Among them, the PTEN/PI3K/AKT/mTOR pathway is the main signaling pathway participating in the metastasis of EC [37].
Moreover, we detected the association between 204 immune-related lncRNAs and EC patient survival prognosis by performing a series of analysis processes, such as univariate, LASSO and multivariate Cox analyses. The 4 lncRNAs, FP671120.4, LINC02381, LNCTAM34A and AC074212.1, showed notable prognostic value for EC patients in the training dataset. Next, the 4 immune-relevant lncRNAs were integrated by adopting risk scoring methods, and the results suggested that the signature could forecast patient survival independently. Furthermore, we verified the accuracy of the signature’s prognostic value in the testing dataset and the entire dataset, and the results demonstrated that the development of the 4-lncRNA signature model was successful in both high-level robustness and improved repeatability in two key aspects. Our research showed that the immune score is closely linked with adverse outcomes in patients diagnosed with endometrial cancer. In addition, in accordance with univariate and multivariate analyses results, the 4-lncRNA signature was further proven to serve as an independent predictor of endometrial cancer survival prognosis. Compared with the age and grade curve, the four lncRNA signature ROC curve displayed a greater AUC. Therefore, we infer that it is reasonable to consider immune-related lncRNA that is independent of other traditional clinical features as beneficial in terms of the accuracy of measuring the prognosis of EC patients. Previous studies have constructed other prognostic models of endometrial cancer [25–28]. By comparing these models, we confirmed the prognostic ability of our immune-related lncRNA signature was higher than that of previous gene signature. It indicated that our four-lncRNA signature was better indicator for making prognosis assessment of endometrial cancer. In addition, we quantitatively analyzed these four lncRNAs by conventional qRT-PCR, and the results were consistent with the above-mentioned bioinformatics results. Therefore, the biological signature we constructed can be identified as a valuable and reliable predictor.
In subsequent steps, to provide better application to clinical diagnosis, we estimated the expression differences of individual lncRNAs among the 4-lncRNA signature between carcinoma tissues and normal tissues and performed KM survival analysis for each of the four lncRNAs. The results showed that patients with high LNCTAM34A expression had a longer survival time than those with low expression, while there was no significant difference in LNCTAM34A expression between cancer and normal tissues, perhaps because the data sample we used was limited. In addition, LNCTAM34A has been found to mediate the high expression of miR34a to protect cells from outside stress stimuli [38]. Moreover, the expression results of two genes in PCR were not identical to each of those analyzed in database. It may be due to the small sample size of normal endometrial tissue in the database or heterogeneity of the tumor.
Nevertheless, there remain certain limitations to our study. To solve the problem of insufficient samples, our signature should be adequately validated in other databases and in studies with larger amounts of endometrial cancer data. The major research analysis approach we employed is bioinformatics technology, which has emerged as an effective and reliable tool, while the functional mechanism and interaction network of lncRNAs are complicated. Further functional studies on the lncRNAs we explored, the acquisition of additional experimental data (in vitro and in vivo) and long-term follow-up observations are essential to estimate the accuracy of our signature and confirm our findings.