m6A- and m5C- modified lncRNAs orchestrate the prognosis in cutaneous melanoma and m6A- modified LINC00893 regulates cutaneous melanoma cell metastasis.

BACKGROUND

As the most important modifications on the RNA level, N6-methyladenosine (m6A-) and 5-methylcytosine (m5C-) modification could have a direct influence on the RNAs. Long non-coding RNAs (lncRNAs) could also be modified by methylcytosine modification. Compared with mRNAs, the function of lncRNAs could be more potent to some extent in biological processes like tumorigenesis. Until now, rare reports have been done associated with cutaneous melanoma. Herein, we wonder if the m6A- and m5C- modified lncRNAs could influence the immune landscape and prognosis in melanoma, and we also want to find some lncRNAs which could directly affect the malignant behaviors of melanoma.

METHODS

Systematically, we explored the expression pattern of m6A- and m5C- modified lncRNAs in melanoma from datasets including UCSC Xena and NCBI GEO, and the prognostic lncRNAs were selected. Then, according to the expression pattern of lncRNAs, melanoma samples from these datasets were divided into several subtypes. Prognostic model, nomogram survival model, drug sensitivity, GO, and KEGG pathway analysis were performed. Furthermore, among several selected lncRNAs, we identified one lncRNA named LINC00893 and investigated its expression pattern and its biological function in melanoma cell lines.

RESULTS

We identified 27 m6A- and m5C- related lncRNAs which were significantly associated with survival, and we made a subtype analysis of melanoma samples based on these 27 lncRNAs. Among the two subtypes, we found differences of immune cells infiltration between these two subtypes. Then, LASSO algorithm was used to screen the optimized lncRNAs combination including ZNF252P-AS1, MIAT, FAM13A-AS1, LINC-PINT, LINC00893, AGAP2-AS1, OIP5-AS1, and SEMA6A-AS1. We also found that there was a significant correlation between the different risk groups predicted based on RS model and the actual prognosis. The nomogram survival model based on independent survival prognostic factors was also constructed. Besides, sensitivity to chemotherapeutic agents, GO and KEGG analysis were performed. In different risk groups, a total of 14 drug molecules with different distributions were obtained, which included AZD6482, AZD7762, AZD8055, camptothecin, dasatinib, erlotinib, gefitinib, gemcitabine, GSK269962A, nilotinib, rapamycin, and sorafenib. A total of 55 significantly related biological processes and 17 KEGG signaling pathways were screened. At last, we noticed that LINC00893 had a relatively lower expression in melanoma tissue and cell lines compared with adjacent tissues and epidermal melanocyte, and down-regulation of LINC00893 could promote the malignant behavior of melanoma cells in A875 and MV3. In these two melanoma cell lines, down-regulation of m6A-related molecules like YTHDF3 and METTL3 could promote the expression of LINC00893.

CONCLUSION

We made an analysis of m6A- and m5C- related lncRNAs in melanoma samples and a prediction of these lncRNAs' role in prognosis, tumor microenvironment, immune infiltration, and clinicopathological features. We also found that LINC00893, which is potentially regulated by m6A modification, could serve as a tumor-suppressor in melanoma and play an inhibitory role in melanoma metastasis.