HSPB8 is a Potential Prognostic Biomarker that Correlates With Immune Cell Infiltration in Bladder Cancer

Abstract

Background: Heat shock protein B8 (HSPB8) is expressed in various cancers. However, the functional and clinicopathological significance of HSPB8 expression in bladder cancer (BC) remains unclear. The present study sought to elucidate the clinicopathological features and prognostic value of HSPB8 in BC. Methods: A BC RNA-seq data set was obtained from The Cancer Genome Atlas Urothelial Bladder Carcinoma (TCGA-BLCA) database, and the external validation dataset GSE130598 was downloaded from the GEO database. Samples in the TCGA-BLCA were categorized into two groups based on HSPB8 expression. Differentially expressed genes (DEGs) between the two groups were defined as HSPB8 co-expressed genes. Gene set enrichment analysis (GSEA), protein-protein interaction networks, and mRNA-microRNA (miRNA) interaction networks were generated to predict the function and interactions of genes that are co-expressed with HSPB8. Finally, we examined immune cell infiltration and constructed a survival prediction model for BC patients. Results: The expression level of HSBP8 has a significant difference between cancer samples and normal samples, and its diagnosis effect was validated by the ROC curve. 446 differential expressed genes between HSBP8 high-expression and HSBP8 low expression groups were identified. Gene enrichment analysis and GSEA analysis show that these differential gene functions are closely related to the occurrence and development of BC and the metabolic pathways of BC. The cancer-related pathways included Cytokine-cytokine receptor Interaction, Focal adhesion, and Proteoglycans in cancer. PPI and protein-coding gene-miRNA network visualized the landscape for these tightly bounded gene interactions. Immune cell infiltration shows that B cells, CD4+T cells, and CD8+T cells have strongly different infiltration levels between the HSBP8 high exp group and low exp group. The survival prediction model shows that HSBP8 has strong prognosis power in the BLCA cohort. Conclusion: Identifying DEGs may enhance understanding of BC development's causes and molecular mechanisms. HSPB8 may play an essential role in BC progression and prognosis and serve as a potential biomarker for BC treatment.

Keywords: HSPB8; biomarker; bladder cancer; microarray; prognosis.

FIGURE 1

Differentially expressed target genes and diagnostic validation in bladder cancer. (A) mRNA expression profile from the GSE130598 dataset before normalization; (B) mRNA expression profile of the GSE130598 dataset after normalization; (C) mRNA expression profile from the TCGA-BLCA dataset before normalization; (D) mRNA expression profile of the TCGA-BLCA dataset after normalization; (E) Box plot of the differences in HSPB8 expression between the tumor and normal groups in theTCGA-BLCA dataset. Each point represents a sample. Blue represents normal tissue,and red represents tumor tissue; (F) Box plot of the differences in HSPB8 expression between the tumor and normal groups in the GSE130598 dataset. Each point represents a sample.Blue indicates normal tissue and red indicates tumor tissue. (G) Receiver operating characteristic(ROC) curve of HSPB8 in the TCGA-BLCA dataset; (H) HSPB8 distribution bodymap in tumor and normal samples, red represents tumor tissue (left), green represents normal tissue (right).

FIGURE 2

Differentially expressed gene distribution in the TCGA-BLCA dataset. (A) Heatmap of the HSPB8 expression between the high and low expression groups from the TCGA-BLCA dataset; (B) Volcano plot of gene expression in the high and low HSPB8 expression groups from the TCGA-BLCA dataset; (C) Heatmap of gene expression between the high and low HSPB8 expression groups in the GSE130598 dataset; (D) Volcano plot of gene expression between the high and low HSPB8 expression groups in the GSE130598 dataset.

FIGURE 3

Gene ontology and KEGG enrichment analysis. (A) Venn diagram showing the co-expressed DEGs in the TCGA-BLCA and GSE130598 datasets; (B) Summary of functional similarities of the co-expressed genes; (C) GO enrichment analysis bar graph. The length of the bar represents the number of enriched genes, and the color represents the significance level (increasing from blue to red); (D) GO enrichment analysis bubble chart. The bubble size represents the number of enriched genes, and the color represents the significance level (increasing from blue to red); (E) KEGG enrichment analysis bar graph. The length of the bar represents the number of enriched genes, and the color represents the significance level (p.adjust < 0.05, increasing from blue to red, p-value adjusted for multiple comparisons); (F) KEGG enrichment analysis network diagram. Each point represents an enrichment term, and the color represents the significance level (p < 0.05, increasing from green to blue, p-values adjusted for multiple comparisons).

FIGURE 4

KEGG enrichment analysis showing (A) cytokine-cytokine receptor interaction; (B) vascular smooth muscle contraction; (C) focal adhesion; (D) protein digestion and absorption; and (E) proteoglycans in cancer pathway diagram.

FIGURE 5

Molecular interaction networks. (A) Protein-protein interaction network analysis was performed on DEGs between the high and low HSPB8 expression groups using the STRING database. Cytoscape was used for visualization. (B–E) The MCODE plug-in was used to analyze hub genes, which include the four groups with the largest number of clusters; (F) The most closely linked hub genes were used as target genes, and the miRNAs that interact with the target genes were predicted by the TarBase, miRecords, and miRTarBase databases to construct a molecular interaction network.

FIGURE 6

Gene set enrichment analysis (GSEA). (A) Bubble chart showing the GO terms enriched between the high and low HSPB8 expression groups in the TCGA-BLCA dataset; (B) Enrichment plots of the GO terms between the high and low HSPB8 expression groups in the TCGA-BLCA dataset; (C) Bubble chart of the enriched KEGG terms between the high and low HSPB8 expression groups in the TCGA-BLCA data set; (D) Enrichment plot of the KEGG terms between the high and low HSPB8 expression groups in the TCGA-BLCA dataset; (E) Bubble chart of the enriched GSEA entries between the high and low HSPB8 expression groups in the TCGA-BLCA dataset; (F) Chordogram of GSEA term enrichment between the high and low HSPB8 expression groups in the TCGA-BLCA dataset.

FIGURE 7

The relationship between HSPB8 expression and immune cell composition in the TCGA-BLCA data set. (A) Immune cell composition in the high and low HSPB8 expression groups. The proportion of composition by 22 immune cell types in tumor samples is shown in the stacked histogram; (B) The relationship between immune cell composition and HSPB8 expression; (C) The correlation of immune cell composition in the 22 samples. Red represents positive correlations, and blue represents negative correlations; (D) Immune cell composition in the high and low HSPB8 expression groups, as analyzed using the CIBERSORT algorithm.

FIGURE 8

The relationship between HSPB8 expression and immune cell composition in the GSE130598 data set. (A) Immune cell composition in the high and low HSPB8 expression groups. The proportion of composition by 22 immune cell types in tumor samples is shown in the stacked histogram; (B) The relationship between immune cell composition and HSPB8 expression; (C) The correlation of immune cell composition in the 22 samples. Red represents positive correlations, and blue represents negative correlations; (D) Immune cell composition in the high and low HSPB8 expression groups, as analyzed using the CIBERSORT algorithm.

FIGURE 9

Construction of a prognostic risk model of immune-related genes co-expressed with HSPB8. (A) HSPB8 expression affects the overall survival of BC patients; (B) The intersection of HSPB8 co-expressed genes with immune-related genes; (C) Combination of TCGA-BLCA clinical data, HSPB8 expression, and co-expressed immune-related genes were used to construct the risk prediction model; (D) The risk curve of HSPB8 co-expressed, immune-related gene risk model (top), survival status (middle), and risk heatmap (bottom); (E) Forest plot showing univariate Cox regression analysis of the predictive power of risk scores combined with clinicopathological characteristics of patients for BC prognosis; (F) Forest diagram showing multivariate Cox analysis of the predictive ability of risk scores combined with clinicopathological characteristics of patients for BC prognosis.

FIGURE 10

Decreased expression of HSPB8 predicts a poor prognosis in patients with BLCA. (A) Overall survival between the high and low-risk groups from the TCGA-BLCA database; (B) The BLCA prognostic value of each clinical variable and its accuracy ROC curve; (C) Construction of a clinical prediction nomogram; (D) Correlation analysis of clinical subgroup variables based on the risk score and the differences between the clinical subgroups shown as a heatmap.