Identification of a methyltransferase-related long noncoding RNA signature as a novel prognosis biomarker for lung adenocarcinoma

Abstract

Background: Lung adenocarcinoma (LUAD) accounts for a high proportion of tumor deaths globally, while methyltransferase-related lncRNAs in LUAD were poorly studied.

Methods: In our study, we focused on two distinct cohorts, TCGA-LUAD and GSE3021, to establish a signature of methyltransferase-related long non-coding RNAs (MeRlncRNAs) in LUAD. We employed univariate Cox and LASSO regression analyses as our main analytical tools. The GSE30219 cohort served as the validation cohort for our findings. Furthermore, to explore the differential pathway enrichments between groups stratified by risk, we utilized Gene Set Enrichment Analysis (GSEA). Additionally, single-sample GSEA (ssGSEA) was conducted to assess the immune infiltration landscape within each sample. Reverse transcription quantitative PCR (RT-qPCR) was also performed to verify the expression of prognostic lncRNAs in both clinically normal and LUAD samples.

Results: In LUAD, we identified a set of 32 MeRlncRNAs. We further narrowed our focus to six prognostic lncRNAs to develop gene signatures. The TCGA-LUAD cohort and GSE30219 were utilized to validate the risk score model derived from these signatures. Our analysis showed that the risk score served as an independent prognostic factor, linked to immune-related pathways. Additionally, the analysis of immune infiltration revealed that the immune landscape in high-risk groups was suppressed, which could contribute to poorer prognoses. We also constructed a regulatory network comprising 6 prognostic lncRNAs, 19 miRNAs, and 21 mRNAs. Confirmatory RT-qPCR results aligned with public database findings, verifying the expression of these prognostic lncRNAs in the samples.

Conclusion: The prognostic gene signature of LUAD associated with MeRlncRNAs that we provided, may offer us a comprehensive picture of the prognosis prediction for LUAD patients.

Keywords: GSEA; immune infiltration; lung adenocarcinoma; methyltransferase-related lncRNAs; prognosis.

Figure 1

Identification of differential genes. (A) The red dots in the plot represent up-regulated genes and blue dots represent down-regulated genes with statistical significance. Gray dots represent no DEGs; (B) The heatmap of top 50 up-regulated and top 50 down-regulated genes in tumor and normal tissue.

Figure 2

Identification of methyltransferase-related lncRNAs. (A) Venn diagram of the intersection between DEGs and methyltransferase-related genes; (B) The mRNA-lncRNA network.

Figure 3

Establishment of methyltransferase-related lncRNAs signature. (A) A forest plot of prognostic methyltransferase-related lncRNAs identified by univariate Cox and Kaplan-Meier survival analysis; (B, C) LASSO regression analysis.

Figure 4

The methyltransferase-related IncRNAs signature was a prognostic biomarker for OS in the TCGA-LUAD cohort. (A) K-M survival of OS according to methyltransferase-related IncRNA signature groups in the training cohorts; (B) AUC of time-dependent ROC curve for the risk score in the training dataset; (C) The OS status and OS risk score plots in the training dataset; (D) The heat map of these 6 methyltransferase-related lncRNAs between the high- and low-risk groups in the training dataset; (E) K-M survival of OS according to methyltransferase-related IncRNA signature groups in the test cohorts; (F) AUC of time-dependent ROC curve for the risk score in the test dataset; (G) The OS status and OS risk score plots in the test dataset; (H) The heat map of these 6 methyltransferase-related lncRNAs between the high- and low-risk groups in the test dataset.

Figure 5

External validation of the risk score in the GSE30219 cohort. (A) The Kaplan-Meier survival analysis; (B) The time-dependent ROC analysis for the risk score in predicting the OS of patients in the GSE30219 cohort; (C, D) The risk score distribution and survival status of patients in the GSE30219 cohort; (E) The heatmap analysis.

Figure 6

The correlation between risk score and clinical characteristics. (A) Age; (B) Sex; (C–F) TNM stage.

Figure 7

Independent value of the prognostic risk model. (A, B) Forrest plots of the univariate Cox regression analysis; (B) Forrest plot of the multivariate Cox regression analysis; (C) The nomogram was established based on the independent prognosis model.

Figure 8

Functional enrichment analyses between the high- and low-risk groups. (A) KEGG_CELL_CYCLE; (B) KEGG_DNA_REPLICATIO; (C) KEGG_P53_SIGNALING_PATHWAY; (D) KEGG_PATHWAYS_IN_CANCER; (E) HALLMARK_G2M_CHECKPOINT; (F) KEGG_CYTOKINE_CYTOKINE_RECEPTOR_INTERACTION; (G) KEGG_JAK_STAT_SIGNALING_PATHWAY.

Figure 9

Immune characteristics of methyltransferase-related lncRNAs-based classifier subgroups. (A) The heatmap of immune infiltrating cells between the high- and low-risk groups; (B) The proportions of 24 infiltrated immune cells and infiltration score in the high-and low-risk groups. ^*P < 0.05; ^***P < 0.001; ^****P < 0.0001.

Figure 10

The Pearson correlation between the risk score and immune cells. (A) CD8 T cells; (B) Cytotoxic cells; (C) DC; (D) Eosinophils; (E) iDC; (F) Macrophages; (G) Mast cells; (H) pDC; (I) T cells; (J) TFH; (K) Tgd; (L) Th2 cells.

Figure 11

Prognostic lncRNA-miRNA-mRNA regulatory network in GC.

Figure 12

Validation of lncRNAs’s expression. (A) Statistical analysis of relative RP11-251M1.1 levels in HCC tissues compared to normal tissue controls; (B) The expression of RP1-78014.1 levels; (C) The expression of LINC00511 levels; (D) The expression of CTD-2510F5.4 levels; (E) The expression of LINC01936 levels; (F) The expression of RP11-750H9.5 levels. ^*P < 0.05; ^**P < 0.01; ^***P < 0.001.