Bioinformatic analyses and experimental validation of the role of m6A RNA methylation regulators in progression and prognosis of adrenocortical carcinoma

Figures

• 

  Figure 1. The flow chart of present study. ACC, Adrenocortical carcinoma; DEGs, differentially expressed genes.

• 

  Figure 2. Expression of m6A regulators in ACC. (A) The process and molecular functions of m6A RNA methylation. (B) The expressive heatmap of m6A-related genes. The gene symbols are on the right of the heatmap. L, localized tumor (n = 46). M, metastatic tumor (n = 31). High expression is shown in red and low expression is green. (C–E) Differential expression of METTL3, HNRNPC and RBM15 between localized and metastatic tumor samples. Gene expression is measured by FPKM. FPKM, Fragments Per Kilobase per Million. *P < 0.05, **P < 0.01.

• 

  Figure 3. Screening of m6A prognostic genes. (A) Cox univariate prognostic analyses of m6A-related genes based on TCGA database. (B–N) Prognostic difference of m6A-related genes between high- and low-expression groups based on GEPIA database. Group cutoff is set as median of gene expression.

• 

  Figure 4. Construction of m6A-related risk signature. (A) Identification of m6A-related prognostic DEGs. (B) The survival difference between the high- (n = 39) and low-risk groups (n = 38). (C) Time-dependent ROC curves of m6A risk signature. (D) PCA plot of m6A risk signature. (E) t-SNE plot of m6A risk signature. (F) ROC curves of m6A risk signature and ACC clinical parameters. (G) Risk plots of m6A risk signature. (H) The summary of HNRNPC and RBM15 mutations. (I) The summary of mutation types. (J) The mutation types of HNRNPC. (K) The mutation types of RBM15. (L) The heatmap of m6A risk signature. The relationships between m6A-related risk levels and the clinicopathological characteristics of ACC were determined by Kolmogorov–Smirnov test. Overexpression is presented in red and low expression is green. DEGs, differentially expressed genes; ROC, receiver operating characteristic curve; PCA, principal component analysis; s-SNE, t-distributed stochastic neighbor embedding; *P < 0.05, **P < 0.01.

• 

  Figure 5. The effect of m6A-related risk signature on ACC prognosis in TCGA cohort. (A) The results of cox univariate prognostic analysis. (B) The results of multivariate prognostic analysis. (C) The calibration curves. (D–I) The prognostic differences of ACC patients with the same clinical subgroups. (J) The DCA curve of m6A risk signature. ‘Simple’ curve (red) represents the prognostic model composed of age and clinical stage. ‘Complex’ curve (green) represents the prognostic model composed of age, clinical stage and m6A risk score. (K) The nomogram is used to predict the OSR of ACC patient at 1, 2, 3 year. DCA, decision curve analysis.

• 

  Figure 6. The validation of m6A-related risk signature in GSE33371 cohort. (A) The survival difference between the high- and low-risk groups. (B) Multivariable ROC curves in the validation cohort. (C) The risk plots in the validation cohort. (D–E) identification of independent prognostic factors in the validation cohort. The results of cox univariate regression are shown in green and those of cox multivariate regression are red. (F) The heatmap of m6A risk signature in validation cohort. *P < 0.05.

• 

  Figure 7. The effect of m6A-related risk signature on ACC immune microenvironment. (A) Comparison of the infiltrating levels of 22 immune cells between different risk groups. (B) Comparison of the activity scores of 13 immune-related pathways between different risk groups. (C–D) The relationships between HNRNPC expression and CD274 (PD-L1), CTLA4 expressions. (E–F) The relationships between RBM15 expression and CD274 (PD-L1), CTLA4 expressions. (G–H) The prognostic differences of ACC patients in TCGA cohort between high- and low-CD274 or CTLA4 expression. (I–J) The expressive difference of CD274 or CTLA4 between high and low m6A-risk group. (K–L) The distributions of m6A risk genes in different PAAD immune subtypes. DC, dendritic cell; APC, antigen-presenting cells; CCR, cytokine-cytokine receptor; IFN, interferon.

• 

  Figure 8. The expression and biofunction of HNRNPC in ACC. (A–B) The differential expression of HNRNPC and RBM15 between tumor and normal samples based on GEPIA database. (C–D) The differential expression of HNRNPC and RBM15 in 5 pairs of clinical specimens. (E) Detection of transfection efficiency. (F–G) HNRNPC can promote proliferation of H295R and SW13 cells. (H) HNRNPC can promote migration and invasion of H295R and SW13 cells. *P < 0.05.