It has been elucidated that several lncRNAs, miRNAs, and mRNAs are differentially expressed in AML, and the ceRNA network might be an instrumental knot in the regulation of tumorigenesis [23]. Therefore, it is urgent to identify key genes and understand the pathogenesis of AML. This study used a series of methods to identify and analyze the data downloaded from TCGA and GTEx databases and verify the obtained results.
Here, the differentially expressed genes were obtained by analyzing the downloaded data. GO and KEGG enrichment analysis revealed that differentially expressed RNAs and notable pathways of enrichment were all associated with the progression of AML. A ceRNA network constructed by 164 regulatory axes was screened. A total of six lncRNAs (ZBTB20-AS4, ZMYM4-AS1, RP11-402L5.1, CTB-193M12.1, KIZ-AS1, and CTD-2260A17.1) were identified from the constructed ceRNA network through survival analysis, which were significantly associated with overall survival rate (P < 0.05) (Fig. 7). Patients with high expression of CTB-193M12.1 indicated poor prognosis (Fig. 7a). The correlation analysis showed that the interaction between CTB-193M12.1 and NFAT5 was significantly strong (r = 0.82, P < 0.01) (Fig. 8c). A potential CTB-193M12.1/hsa-mir-206/NFAT5 regulatory axis with the highest correlation was constructed ultimately, which was basically in line with the ceRNA theory (Fig. 9). The results of qRT-PCR validated the differentially expressed of hsa-mir-206 and NFAT5 (Fig. 10). The analysis aimed to provide new ideas for a comprehensive and deeper understanding of the molecular mechanism of AML.
LncRNA, a functional RNA molecule that plays a role in various biological functions of the human body, contributes significantly to the aforementioned regulatory network. Since studies focused on lncRNAs are still limited, only three have official human genome nomenclature committee symbols (ZBTB20-AS4, ZMYM4-AS1, and KIZ-AS1) and none have been reported. However, cluster analysis showed that CTB-193M12.1 and NFAT5 expression were increased in AML patients while hsa-mir-206 was decreased (Fig. 3). Our bioinformatics analysis results also revealed a strong correlation between CTB-193M12.1 and AML (Fig. 7a). Besides, several studies indicate that abnormal expression of lncRNAs is often associated with the development of AML [3, 24]. Therefore, it is reasonable to suggest CTB-193M12.1 as a regulatory gene of AML.
As a binding site for multiple RNAs, miRNA are also indispensable in tumorigenesis. Interestingly, our results revealed that of 163 lncRNA-miRNA-mRNA regulatory axes, 145 were regulated by hsa-mir-206, which indicated that hsa-mir-206 may be the strongest regulator in the pathogenesis of AML. Recently, studies on hsa-mir-206 have become a hot topic, but studies on AML are still lacking. Notably, we found that a recent study by Chen et al. [25] on occupational asthma (OA) proposed a regulatory axis similar to that in this study. They pointed out that miR-206-3p, as a key factor in calcineurin/NFAT signaling in macrophages and bronchoalveolar lavage cells, affects the transcription of iNOS and thus regulates the development of OA. This proven hsa-mir-206/NFAT regulatory axis correlates with our findings. Further, Chen et al. also used the THP-1 cell line. Therefore, it is reasonable to suggest that hsa-mir-206 can act as a key site in regulating the development of AML by mediating the expression of NFAT5. Moreover, it has been reported that low expression of hsa-mir-206 is closely associated with the poor prognosis of pediatric AML patients [26]. This is consistent with our finding that the downregulation of hsa-mir-206 predicts poor prognosis in AML. It is noteworthy that our study was conducted in adults, which not only verified the above view but also provided powerful evidence that hsa-mir-206 may have an impact on the pathogenesis of AML. Also, hsa-mir-206 plays an important role in other cancers. For example, downregulation of miR-206 leads to poor prognosis in endometrial cancer [27] and promotes the occurrence and development of laryngeal cancer [28]. High expression of miR-206 inhibits osteosarcoma cell proliferation [29], as well as the proliferation and migration of prostate cancer cells [30]. Overall, these results suggested that hsa-mir-206 can regulate multiple cancer processes and might play an important role in AML through the hsa-mir-206/NFAT5 regulatory axis.
Similarly, NFAT5, which is a target gene of hsa-mir-206 in the ceRNA network, has been shown to regulate multiple tumor processes. For example, NFAT5 can act as a ceRNA regulated by circFOXO3, plays a role in glioblastoma through sponging miR-138-5p and miR-432-5p [31], and may also regulate the progression of colon cancer through the MALAT1/miR1295p/NFAT5 axis [32]. Here, results from our investigation and the GEPIA website both showed that NFAT5 has a strong relationship with CTB-193M12.1 (Fig. 8c) and that the low expression of NFAT5 can lead to a higher survival rate. Moreover, one study also found that the downregulation of NFAT5, which is regulated by an upstream gene, led to a more optimistic prognosis in lung cancer [33]. This demonstrates that high expression of NFAT5 can lead to poor prognosis in cancer and may also participate in the progression of multiple cancer types as a ceRNA. Therefore, it is reasonable to suggest that NFAT5 may regulate the progression of AML through the CTB-193M12.1/hsa-mir-206/NFAT5 ceRNA regulatory axis proposed in this study.
Recently, studies on the association of ceRNA with AML have emerged, but most of them only focused on the association between lncRNA dysregulation and AML [34, 35]. Among a study aimed at pediatric AML, although the biomarkers based on the ceRNA network have been analyzed [36], whether the biomarkers suggested in the study applies to adult AML still need further validation. Another research proposed by Wang et al. [37] constructed a coexpression ceRNA network and identified several cancer-related genes, but the interactional relationships between differentially expressed genes still not been analyzed. Our current research aimed to establish a more integrated regulatory network and select the most relevant regulatory axis to an in-depth study, in order to identify several prognostic biomarkers and understand the new "language" of communication between RNA transcripts from multiple dimensions in AML.
However, there are still some limitations in our study. The interaction between RNAs has not been experimentally confirmed. It is necessary to create experiments to verify the CTB-193M12.1/hsa-mir-206/NFAT5 regulatory axis in future studies.