A Novel Inflammation-Related Gene Signature for Overall Survival Prediction and Comprehensive Analysis in Pediatric Patients with Wilms Tumor

Abstract

Wilms tumor (WT) is a common pediatric renal cancer, with a poor prognosis and high-risk recurrence in some patients. The inflammatory microenvironment is gradually gaining attention in WT. In this study, novel inflammation-related signatures and prognostic model were explored and integrated using bioinformatics analysis. The mRNA profile of pediatric patients with WT and inflammation-related genes (IRGs) were acquired from Therapeutically Available Research to Generate Effective Treatments (TARGET) and Gene Set Enrichment Analysis (GSEA) databases, respectively. Then, a novel prognostic model founded on 7-IRGs signature (BICC1, CSPP1, KRT8, MYCN, NELFA, NXN, and RNF113A) was established by the least absolute shrinkage and selection operator (LASSO) and multivariate Cox regression to stratify pediatric patients with WT into high- and low-risk groups successfully. And a stable performance of the prognostic risk model was verified in predicting overall survival (OS) by receiver-operating characteristic (ROC) curves, Kaplan-Meier (KM) curves, and independent prognostic analysis (p < 0.05). In addition, a novel nomogram integrating risk scores with good robustness was developed and validated by C-index, ROC, and calibration plots. The potential function and pathway were explored via Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and GSEA, with mainly inflammation and immune-related biological processes. The higher-risk scores, the lower immune infiltration, as shown in the single-sample GSEA (ssGSEA) and tumor microenvironment (TME) analysis. The drug sensitivity analysis showed that regulating 7-IRGs signature has a significant correlation with the chemotherapy drugs of WT patients. In summary, this study defined a prognostic risk model and nomogram based on 7-IRGs signature, which may provide novel insights into clinical prognosis and inflammatory study in WT patients. Besides, enhancing immune infiltration based on inflammatory response and regulating 7-IRGs signature are beneficial to ameliorating the efficacy in WT patients.

Figure 1

The workflow chart of this study processes.

Figure 2

Identification of DE-IRGs and IRPGs. (a) Volcanic map for displaying the up- and downregulation of DE-IRGs. (b) Venn plotting IRPGs overlapped between DE-IRGs and OS. (c) Forest plots to show the results of the univariate Cox regression analysis between IRPGs expression and OS. (d) The relevance heatmap revealed the correlation among IRPGs. (e) The heatmap showed the differences of IRPGs between tumor and normal tissues.

Figure 3

Construction of the inflammation-related prognostic signatures using LASSO regression analysis. (a, b) LASSO regression analysis among 23 IRPGs. Each curve corresponds to an IRPG in the LASSO model. Partial likelihood deviance with 10-fold crossvalidation tuning the parameter selection was used to screen the best lambda. Upper X-axes were the number of included IRPGs, while lower X-axes were log lambda (λ) whose greater, the greater the punishment of the linear model. (a) Y-axes mean coefficients in each IRPGs; (b) Y-axes was partial likelihood deviance values that mean the magnitude of the error included in the variable model. (c) Kaplan–Meier curves for the 7-IRGs signature relative to OS outcomes of patients in the high- and low-risk groups. (d) AUC of the ROC curves demonstrated the predictive efficiency of the 7-IRGs signature model. (e, f) The distribution of pediatric patients with WT and survival status was based on the risk scores in the low- and high-risk groups, and the dotted line was the median value of risk scores (g, h) PCA plot and t-SNE analysis with dimension reduction.

Figure 4

Relevance analysis of clinical characteristics. (a) The differences of clinical characteristics in pediatric patients with WT, including stage, events, age, gender, and histology. (b, c) Univariate and multivariate regression analysis outcomes of the relationships between the OS and the clinical parameters.

Figure 5

Construction and validation of the nomogram model. (a) Nomogram integrating the points of clinical characteristics and risk scores to predict the probability of 1-, 3-, and 5-year OS in pediatric patients with WT. (b) Calibration plots for assessing the discrimination ability of the nomogram model. (c) ROC curves for validating the predictive performance of the nomogram model.

Figure 6

Function and pathway analysis based on DE-IRGs. (a) GO analysis is based on biological process (BP), cellular component (CC), and molecular function (MF). (b) KEGG analysis for the pathway.

Figure 7

Function and pathway analysis for WT patients between high- and low-risk groups by GSEA analysis. (a, b) The GSEA analysis is based on GO for high-risk (a) and low-risk (b) groups. (c, d) The GSEA analysis is based on KEGG for high-risk (c) and low-risk (d) groups.

Figure 8

The differences between immune infiltration and tumor microenvironment analysis. (a, b) The results of ssGSEA to compare the differences of immune infiltration between high- and low-risk groups in pediatric patients with WT. ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001; ns: no significance. (c) Stromal score of TME plotted for showing the correlation between the content of stromal cells and risk scores. (d) The immune score of TME was plotted for showing the correlation between the content of immune cells and risk scores.

Figure 9

Drug sensitivity analysis founded on the IRGs signature in pediatric patients with WT.