A novel autophagy-related long non-coding RNAs prognostic risk score for clear cell renal cell carcinoma

Abstract

Background: As the main histological subtype of renal cell carcinoma, clear cell renal cell carcinoma (ccRCC) places a heavy burden on health worldwide. Autophagy-related long non-coding RNAs (ARlncRs) have shown tremendous potential as prognostic signatures in several studies, but the relationship between them and ccRCC still has to be demonstrated.

Methods: The RNA-sequencing and clinical characteristics of 483 ccRCC patients were downloaded download from the Cancer Genome Atlas and International Cancer Genome Consortium. ARlncRs were determined by Pearson correlation analysis. Univariate and multivariate Cox regression analyses were applied to establish a risk score model. A nomogram was constructed considering independent prognostic factors. The Harrell concordance index calibration curve and the receiver operating characteristic analysis were utilized to evaluate the nomogram. Furthermore, functional enrichment analysis was used for differentially expressed genes between the two groups of high- and low-risk scores.

Results: A total of 9 SARlncRs were established as a risk score model. The Kaplan-Meier survival curve, principal component analysis, and subgroup analysis showed that low overall survival of patients was associated with high-risk scores. Age, M stage, and risk score were identified as independent prognostic factors to establish a nomogram, whose concordance index in the training cohort, internal validation, and external ICGC cohort was 0.793, 0.671, and 0.668 respectively. The area under the curve for 5-year OS prediction in the training cohort, internal validation, and external ICGC cohort was 0.840, 0.706, and 0.708, respectively. GO analysis and KEGG analysis of DEGs demonstrated that immune- and inflammatory-related pathways are likely to be critically involved in the progress of ccRCC.

Conclusions: We established and validated a novel ARlncRs prognostic risk model which is valuable as a potential therapeutic target and prognosis indicator for ccRCC. A nomogram including the risk model is a promising clinical tool for outcomes prediction of ccRCC patients and further formulation of individualized strategy.

Keywords: Autophagy; Clear cell renal cell carcinoma (ccRCC); Long non-coding RNA (lncRNA); Prognosis; Risk score.

Fig. 1

A flow diagram depicted in our study

Fig. 2

Construction and validation of the risk score model. A–C Distribution of risk score, scatter plot of survival time, and expression levels of 9 lncRNAs between the high- and low-risk groups in the training cohort, validation cohort, and external validation cohort (ICGC); D–F Kaplan–Meier survival curve between high- and low-risk groups in the training cohort, validation cohort, and external validation cohort (ICGC); G–I PCA based on the nine selected ARlncRs between high- and low-risk ccRCC patients in the training cohort, validation cohort, and external validation cohort (ICGC)

Fig. 3

Prognosis and correlation analysis of the risk score with clinical characteristics in TCGA. A–E Boxplots of the Wilcoxon test of the risk score in clinical characteristics including A age B gender C T stage D M stage E AJCC stages; F–O Kaplan‒Meier survival curve between high-risk and low-risk groups ccRCC patients stratified by F, G age(> 65, <  = 65), H, I gender (female, male), J, K T stage(T1–T2, T3–T4), L, M M stage(M0, M1), N, O tumor stages (stage I–II, stage III–IV)

Fig. 4

Forest plot of the univariate A and multivariate Cox B regression analyses of the correlation between the OS of ccRCC patients and clinicopathological features (including risk score) in the TCGA training cohort. The red squares indicate the HR, and the blue transverse line indicates the 95% CI

Fig. 5

Establishment and assessment of the nomogram. A Nomogram considering age, M stage, and risk score; B–D The calibration curve of the nomogram in the B training cohort; C validation cohort; D external ICGC cohort; E–G ROC curve of independent prognostic indicators in the E training cohort; F validation cohort; G external ICGC cohort

Fig. 6

Establishment of the lncRNA‒mRNA coexpression network. A The ceRNA network of the 9 SARlncRs and their 30 target mRNAs, whose correlation coefficient was no less than 5. B The Sankey diagram of the regulatory relationship between 30 mRNAs, 9 SARlncRs, and risk types (protective or risk); C Heatmap of enriched terms, including GO and KEGG, across 30 input mRNAs, colored by p values. D, E Network of enriched terms for mRNA colored by D cluster ID and E p value. Nodes closer to each other indicated the same cluster ID. Nodes with more remarkable p valuevalues tended to contain more genes

Fig. 7

Functional enrichment analysis of DEGs between high- and low-risk groups. A Biological processes, cell components and molecular functions in the GO analysis; B KEGG signaling pathway analysis with DEGs