A Novel Defined RAS-Related Gene Signature for Predicting the Prognosis and Characterization of Biological Function in Osteosarcoma

Abstract

Background: Osteosarcoma (OS) is the most common primary bone malignancy in children and adolescents with a high incidence and poor prognosis. Activation of the RAS pathway promotes progression and metastasis of osteosarcoma. RAS has been studied in many different tumors; however, the prognostic value of RAS-associated genes in OS remains unclear. On this basis, we investigated the RAS-related gene signature and explored the intrinsic biological features of OS.

Methods: We obtained RNA transcriptome sequencing data and clinical information of osteosarcoma patients from the TARGET database. RAS pathway-related genes were obtained from the KEGG pathway database. Molecular subgroups and risk models were developed using consensus clustering and least absolute shrinkage and selection operator (LASSO) regression, respectively. ESTIMATE algorithm and ssGSEA analysis were used to assess the tumor microenvironment and immune penetrance between the two groups. A comprehensive review of gene ontology (GO) and KEGG analyses revealed inherent biological functional differences between the two groups.

Results: The consistent clustering showed stratification of osteosarcoma patients into two subtypes based on RAS-associated genes and provided a robust prediction of prognosis. A risk model further confirmed that RAS-related genes are the best prognostic indicators for OS patients. GO analysis showed that GDP/GTP binding, focal adhesion, cytoskeletal motor activity, and cell-matrix junctions were associated with the RAS-related model group. Furthermore, RAS signaling in osteosarcoma based on KEGG analysis was significantly associated with cancer progression, with immune function and tumor microenvironment particularly affected.

Conclusion: We constructed a prognostic model founded on RAS-related gene and demonstrated its predictive ability. Then, furtherly exploration of the molecular mechanisms and immune characteristics proved the role of RAS-related gene in the dysregulation in OS.

Figure 1

Flow diagram of the data analysis process.

Figure 2

Identification of RAS-related prognostic genes. (a) Venn diagram showing the 14 intersection RAS-related prognostic genes. (b) The optimal number of clusters (K = 2) determined from cumulative distribution function (CDF) curves, and the classification effect is the best. (c) Heatmap of the expression of 14 RAS-related prognostic genes in the two clusters. (d) PCA confirming that osteosarcoma can be clearly separated into two clusters based on the expression of 14 RAS-related prognostic genes. (e) Kaplan–Meier curves for survival prediction of patients in the two clusters.

Figure 3

Differentially expressed gene (DEG) analysis and functional analysis. (a) Volcano plot showing the DEGs between the two subgroups. (b) Circle plot visualizing the biological processes enriched by gene ontology (GO) analysis. (c) The KEGG pathways associated with immune, stromal, and oncogenic signatures highly enriched in C1 versus C2 identified by GSEA.

Figure 4

Construction of risk signature. (a) Univariate Cox regression analysis of OS for each pyroptosis-related gene, and 14 genes with P < 0.05. (b) LASSO regression of the 5 OS-related genes. (c) Cross-validation for tuning the parameter selection in the LASSO regression. (d) Kaplan–Meier curves for the OS of patients in the high and low-risk groups. (e) ROC curves demonstrating the predictive efficiency of the risk score.^∗P value less than 0.05 is statistically significant.

Figure 5

Univariate and multivariate Cox regression analyses for the risk score. (a) Univariate analysis. (b) Multivariate analysis. (c) Heatmap (green, low expression; red, high expression) for the connections between clinicopathologic features and the risk groups (^∗P < 0.05).

Figure 6

Immune analyses in the two subgroups. (a) Stromal score, (b) immune score, (c) ESTIMATE score, and (d) tumor purity calculated by the ESTIMATE algorithm. (e) Comparison of the enriching level of 29 immune-related cells evaluated by the ssGSEA algorithm, between low (yellow box) and high-risk (blue box) group. ^∗P < 0.05; ^∗∗P < 0.01; and ^∗∗∗P < 0.001.

Figure 7

Functional analysis based on the DEGs between the two risk groups. (a) Gene ontology (GO) highly enriched in the high-risk group. (b) Gene ontology (GO) highly enriched in the low-risk group.