Prognostic value of GLCE and infiltrating immune cells in Ewing sarcoma

Fig. 1. Identification of prognostic genes in the training set. (A) Common DEGs between the normal and ES tissues in the GSE17679, GSE68776 and GSE45544 sets. (B) The forest plot for genes with a p value < 0.001 in the univariate Cox regression analysis. (C) The LASSO coefficient profiles for the genes screened by univariate Cox regression analysis in the 10-fold cross-validations. (D) Partial likelihood deviance with changing log(λ) plotted by LASSO regression in 10-fold cross-validations (λ = 1se, the red line in the plot). (E) The error rate varied with the number of trees in RF feature selection: the x-axis represents the number of trees; the y-axis represents the error rate of the RF feature selection. (F) Importance score of genes calculated by RF feature selection: the x-axis represents the importance score; the y-axis represents the gene symbol ranked by the importance score, with GLCE having the highest importance score.

Fig. 2. Screening immune infiltrating cells with prognostic value in the training set. (A) Immune infiltration profile between normal and ES tissues in GSE17679 by the “xCell” package. (B) Forest plot for part of the infiltrating immune cells with a p value < 0.05 in the univariate Cox regression analysis (immune infiltrating cells with too large a hazard ratio value could not be shown in the plot and were marked with a red asterisk instead). (C) Forest plot for the infiltrating immune cells with a p value < 0.05 in the multivariate Cox regression analysis (infiltrating immune cells with too large a hazard ratio value could not be shown in the plot and were marked with a red asterisk instead). (D) The importance score of the infiltrating immune cells varied with the running times of the Boruta feature selection: x-axis refers to the running times of the Boruta feature selection; y-axis refers to the importance score. (E) The importance score of the infiltrating immune cells screened by Cox regression analyses: x-axis refers to the infiltrating immune cells screened by Cox regression analyses; y-axis refers to the importance score.

Fig. 3. Differential infiltration of immune cells in different cohorts. (A) Differential infiltration of immune cells screened by Cox regression analyses between normal and ES tissues in GSE17679. (B) Differential infiltration of immune cells screened by Cox regression analyses between normal and ES tissues in GSE68776. (C) Differential infiltration of the immune cells screened by Cox regression analyses between normal and ES tissues in GSE45544. (Comparison method: Kruskal‒Wallis test; p significance level: no significance (ns), p ≥ 0.05; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001).

Fig. 4. Evaluation of the prognostic value of GLCE, infiltrating mast cells and the risk score in the training set. (A) KM plot for patients in the high and low GLCE expression groups of GSE17679. (B) KM plot for patients in the high and low infiltrating mast cell groups of GSE17679. (C) KM plot for patients in the high and low risk score groups of GSE17679. (D) The characteristics of patients in the high and low risk groups of GSE17679: the left side of the vertical dashed line in the scatter plots refers to the patients in the high score group, the right side refers to the patients in the low score group, and the heatmap refers to the scaled value of GLCE expression and infiltrating mast cells. (E) The nomogram of the Cox proportional hazards model in GSE17679. (F) Calibration analysis of the model in GSE17679: the x-axis refers to the overall survival predicted by the model, the y-axis refers to the observed overall survival rate, and the diagonal (dashed line) is the ideal line. (G) Time-dependent ROC analysis of the model in GSE17679: the x-axis refers to 1-specificity of the model, and the y-axis refers to the sensitivity of the model. (H) The 1-year DCA in GSE17679: x-axis refers…

Fig. 5. Evaluation of the prognostic value of GLCE, infiltrating mast cells and the risk score in GSE63155. (A) KM plot for patients in the high and low GLCE expression groups of GSE63155. (B) KM plot for patients in the high and low infiltrating mast cell groups of GSE63155. (C) KM plot for patients in the high and low risk score groups of GSE63155. (D) The characteristics of patients in the high and low risk groups of GSE63155: the left side of the vertical dashed line in the scatter plots refers to the patients in the high score group, the right side refers to the patients in the low score group, and the heatmap refers to the scaled value of GLCE expression and infiltrating mast cells. (E) The nomogram of the Cox proportional hazards model in GSE63155. (F) Calibration analysis of the model in GSE63155: the x-axis refers to the overall survival predicted by the model, the y-axis refers to the observed overall survival rate, and the diagonal (dashed line) is the ideal line. (G) Time-dependent ROC analysis of the model in GSE63155: the x-axis refers to 1-specificity of the model, and the y-axis refers to the sensitivity of the model. (H) The 1-year DCA in GSE63155: x-axis refers to the…

Fig. 6. Evaluation of the prognostic value of GLCE, infiltrating mast cells and the risk score in GSE63156. (A) KM plot for patients in the high and low GLCE expression groups of GSE63156. (B) KM plot for patients in the high and low infiltrating mast cell groups of GSE63156. (C) KM plot for patients in the high and low risk score groups of GSE63156. (D) The characteristics of patients in the high and low risk groups of GSE63156: the left side of the vertical dashed line in the scatter plots refers to the patients in the high score group, the right side refers to the patients in the low score group, and the heatmap refers to the scaled value of GLCE expression and infiltrating mast cells. (E) The nomogram of the Cox proportional hazards model in GSE63156. (F) Calibration analysis of the model in GSE63156: the x-axis refers to the overall survival predicted by the model, the y-axis refers to the observed overall survival rate, and the diagonal (dashed line) is the ideal line. (G) Time-dependent ROC analysis of the model in GSE63156: the x-axis refers to 1-specificity of the model, and the y-axis refers to the sensitivity of the model. (H) The 1-year DCA in GSE63156: x-axis refers to the…

Fig. 7. Functional analysis of DEGs between normal and ES tissues in the training set. (A) GO enrichment analysis of the common DEGs in the three sets between normal and ES tissues. (B) KEGG enrichment analysis of the common DEGs in the three sets between normal and ES tissues. (C) Top five enriched pathways of interest by GSEA for the DEGs in GSE17679. (D) Chord plot for the top five enriched pathways of interest in GSE17679.

Fig. 8. Functional analysis of DEGs between normal and ES tissues in validation sets. (A) Top five enriched pathways of interest by GSEA for the DEGs in GSE63155. (B) Chord plot for the top five enriched pathways of interest in GSE63155. (C) Top five enriched pathways of interest by GSEA for the DEGs in GSE63156. (D) Chord plot for the top five enriched pathways of interest in GSE63156.

Fig. 9. Functional analysis of the GLCE-correlated DEGs in GSE17679. (A) Lollipop plot for the top GLCE-correlated genes with p value < 0.05 and absolute correlation value ≥ 0.7. (B) GO enrichment analysis of the GLCE-correlated DEGs in GSE17679. (C) Chord plot for the top five enriched pathways of interest for the GLCE-correlated DEGs in GSE17679. (D) Top five enriched pathways of interest by GSEA for the GLCE-correlated DEGs in GSE17679. (E) PPI network analysis of the proteins encoded by the top GLCE-correlated DEGs (p value < 0.05 and absolute correlation value > 0.6) by the STRING website (interaction score≥0.4).

Fig. 10. Histological and immunohistochemical staining of ES and adjacent tissues. Immunohistochemical staining for GLCE in adjacent (A, C) and ES (B, D) tissues at 200 × and 400 × magnification, respectively. Toluidine blue staining of adjacent (E, G) and ES (F, H) tissues at 200 × magnification and 400 × magnification, respectively. The metachromatic granules of mast cells (red arrowhead) were stained red‒purple, and the orthotropic background was stained blue.