Expression profile analysis identifies key genes as prognostic markers for metastasis of osteosarcoma

Abstract

Background: OS is the most common malignant tumor of bone which was featured with osteoid or immature bone produced by the malignant cells, and biomarkers are urgently needed to identify patients with this aggressive disease.

Methods: We downloaded gene expression profiles from GEO and TARGET datasets for OS, respectively, and performed WGCNA to identify the key module. Whereafter, functional annotation and GSEA demonstrated the relationships between target genes and OS.

Results: In this study, we discovered four key genes-ALOX5AP, HLA-DMB, HLA-DRA and SPINT2 as new prognostic markers and confirmed their relationship with OS metastasis in the validation set.

Conclusions: In conclusion, ALOX5AP, HLA-DMB, HLA-DRA and SPINT2 were identified by bioinformatics analysis as possible prognostic markers for OS metastasis.

Keywords: Biomarker; Gene expression; Metastasis; Molecular classifier; Osteosarcoma; Prognosis.

Fig. 1

Flow chart of study design

Fig. 2

Construction and identification of modules associated with the clinical traits. a Clustering dendrogram of OS samples and the clinical traits. For age and grade, white means a low value, red a high value, and grey a missing entry; for gender and metastasis, white means female or non-metastasis, red means male or metastasis. b Hierarchical clustering based on the dynamic tree, each branch above represented a gene, and each color below represented a gene co-expression module. Grey module color is a reserved one for genes that are not part of any module. c Number of genes in different gene co-expression modules. Note that genes in the grey module were identified as not co-expressed. d Heatmap of the correlation between module eigengenes and clinical traits. Each row corresponds to a module eigengene, column to a trait. Each cell contains the corresponding correlation and p-value. The table is color-coded by correlation according to the color legend. The blue module was significantly correlated with metastasis

Fig. 3

Functional enrichment analysis of blue module. a GO analysis of all genes in blue module. b KEGG pathway analysis of all genes in blue module. c GSEA in Reactome gene sets. d, e enrichment plots for cell cycle (d) and cell cycle mitotic gene set (e)

Fig. 4

Identification of key genes based on training set. a A scatterplot of Gene Significance (GS) for weight vs. Module Membership (MM) in the blue module. There is a highly significant correlation between GS and MM in this module. b Volcano plot of significance of gene expression difference between metastasis and non-Metastasis patients. A gene is considered significantly differentially expressed if its |log(FC)| > 1 and p-value < 0.05. c Overall survival analysis of 4 key genes. Expression levels of ALOX5AP, HLA-DMB, HLA-DRA and SPINT2 are significantly related to the overall survival of patients with OS (P < 0.05). d Boxplot of significance of gene expression levels of 4 key genes. ALOX5AP, HLA-DMB, HLA-DRA and SPINT2 are significantly downregulated in metastasis OS compared with non-metastasis OS. The **** represents P < 0.0001. e ROC curves analysis of 4 key genes diagnosis. ROC curves and AUC statistics are used to evaluate the capacity to discriminate OS with or without metastasis with excellent specificity and sensitivity. f The network illustrates the relationship of 4 key genes and the 36 most frequently altered neighbor genes. The 4 key genes are presented in red and orange depending on the gene importance defined as the degree of connectivity. The other genes are represented in blue and green

Fig. 5

Evaluation and validation of the 4-gene signature risk model of OS. a The ROC curves are shown for risk score model in training set. b, c Kaplan–Meier analysis for the overall survival of OS patients in training set (b) and validation set (c)