Risk score model with two immune infiltration-related long non-coding RNAs to predict prognosis in patients with...

Osteosarcoma (OS) is the most frequent type of cancer, as well as a common malignant tumor in orthopedics. Multiple long non-coding RNAs (lncRNAs) are implicated in immune infiltration in numerous types of cancer. Using single-sample gene set enrichment analysis to assign patients with OS, two immunotypes of OS were identified. Using K-means, spectral and PCA-K-means clustering two immunotypes of OS were demonstrated to be reproducible and represent a biologically meaningful classification of patients with OS based on their tumor immune microenvironment. Cluster 1 was an immune-infiltrating type, while Cluster 2 was an immune 'desert' type (low immune cell infiltration). Between Cluster 1 and 2, 29 common differentially expressed lncRNAs (DELs) were identified and univariate Cox regression (UCR) analysis was performed to identify DELs associated with overall survival. A risk score model was established by performing UCR and the iterative Lasso Cox regression analyses based on two immune infiltration-related lncRNA (IIRLs; LINC01094 and RP11-15K2.2). The risk model was a novel independent prognostic factor for patients with OS. A mutual influence between the expression of two IIRLs and mRNAs (LINC01094 and RP11-15K2.2) was also identified. Gene Ontology demonstrated that these mRNAs were abundant in immune-associated functions and pathways. Cytoscape was used to construct lncRNA-mRNA networks. The discrepancy in tumor-infiltrating immune cell abundance demonstrated that the high-risk group was associated with infiltration of diverse immune cell types. Furthermore, the expression levels of immune checkpoint inhibitors were markedly upregulated in the high- compared with that in the low-risk-cohort. Gene set enrichment analysis revealed that the risk score was associated with nucleotide oligomerization domain-like receptor signaling pathways. In summary, two IIRLs were developed in the present study to predict prognosis in patients with OS. The molecular mechanisms in the high-risk group may influence immune infiltration-linked biological processes.