Bioinformatics Analysis to Screen the Key Prognostic Genes in Tumor Microenvironment of Bladder Cancer

Abstract

Bladder cancer (BLCA) is the fifth most common cancer and has the features of low survival rate and high morbidity and mortality. The Cancer Genome Atlas (TCGA) is a pool of global gene expression profile and contains huge amounts of cancer genomics data, which makes it possible to inquire the relationship between gene expression and prognosis of a series of malignant tumors including BLCA. Immune and stromal cells are two major components of tumor microenvironment (TME) which play an important role in judging the prognosis of tumor and influencing the progression of malignant, inflammatory, and metabolic disorders. In our study, we conducted a quantitative analysis of immune and stromal elements based on the ESTIMATE algorithm and thus divided BLCA cases into high and low groups. Then the differentially expressed genes closely related to tumor prognosis between groups were identified and had been shown to correlate with immune response and stromal alterations, which was further confirmed by functional enrichment analysis and protein-protein interaction networks. We validated those genes through BLCA dates downloaded from ArrayExpress and thus got the marker genes to predict prognosis of BLCA. Additionally, immune cell infiltration analysis explored the correlation between the verified genes and immune cells. In conclusion, we identified a series of TME-related genes that assess the prognosis and explored the interaction between TME and tumor prognosis to guide clinical individualized treatment.

Figure 1

Immune scores and stromal scores are associated with smoking, race, the malignancy, and the survival probability. (A, B) Distribution of immune scores (a) and stromal scores (b) of different duration of cigarette smoking (n = 405, P < 0.05). (C, D) Distribution of immune scores (c) and stromal scores (d) for Asian, Black, and White race (n = 405, P < 0.05). (E, F) Distribution of immune scores (e) (n = 405, P=0.1471) and stromal scores (f) (n = 405, P < 0.05) of different degrees of malignancy cases. (G, H) Kaplan–Meier curves for survival probability of bladder cancer patients with low versus high immune scores (g) (n = 405, P=0.07) and stromal scores (h) (n = 405, P < 0.05).

Figure 2

Gene expression profile is of great relevance to immune scores and stromal scores in BLCA. (a, b) Heatmaps show that differentially expressed genes profiles between high and low immune scores/stromal scores groups. Red represents higher expression genes, green represents lower expression genes, black represents same expression genes (fold change >1.5 and P < 0.05). (c, d) Venn diagrams show the number of coupregulated (c) or codownregulated (d) DEGs in immune and stromal score groups. (e, f, and g) The major relevant terms. P < 0.05.

Figure 3

Association between DEGs expressions and overall survival in TCGA. Kaplan–Meier curves for OS (d) of bladder cancer patients with low versus high immune/stromal scores were made to select the DEGs. P < 0.05.

Figure 4

The top four remarkable PPI networks of CD27, TBX21, SLC39A5, and HMHB1 modules.

Figure 5

GO function analysis and KEGG pathway analysis for DEGs associated with overall survival. (a) Cellular components (CC). (b) Biological processes (BP). (c) Molecular functions (MF). (d) KEGG pathways.

Figure 6

Validation of correlation of DEGs associated with prognosis extracted from TCGA database with overall survival in ArrayExpress database. Kaplan–Meier survival curves of the extracted genes from TCGA were generated to verify the prognostic significance in ArrayExpress database. P < 0.05.

Figure 7

Correlation analysis between the expressions of mainly identified DEGs (TNFAIP6, CTSE, COMP, and DSG1) and infiltration levels of B cell, CD8+ T cell, CD4+ T cell, macrophage, neutrophil, and dendritic cell.