A prognostic signature based on seven T-cell-related cell clustering genes in bladder urothelial carcinoma

Abstract

Bladder urothelial carcinoma (BLCA) is one of the most common cancer-related deaths in the world, along with high mortality. Due to the difficult detection of early symptoms, the treatment for this disease is still dissatisfactory. Thus, the current research hotspot is beginning to focus on the immune microenvironment in this disease, aiming to provide guidance for diagnosis and treatment. In this study, the single-cell RNA sequencing data downloaded from the gene expression omnibus database was used to classify the immune cells of BLCA. And the final seven T-cell-related cell clustering genes associated with BLCA prognosis (HSPA2, A2M, JUN, PDGFRB, GBP2, LGALS1, and GAS6) were screened out, and then used for constructing the prognostic model against BLCA based on the Cox and LASSO regression analysis. Satisfactorily, the model could efficiently evaluate the overall survival of BLCA and had the potential to be applied for the clinic treatment. Moreover, we also revealed that the difference in immune infiltration levels and gene mutation might account for the diverse prognosis in BLCA patients. In a word, our findings provided a novel insight for designing efficient immunotherapies for BLCA.

Keywords: bladder urothelial carcinoma; immune microenvironment; prognosis; single-cell RNA sequencing.

Figure 1

Quality control of single-cell RNA sequencing data. (a) Gene expression levels in each cell among four samples. (b) Scatter plot showing the distribution of count number (nCount_RNA) with gene number (nFeature_RNA), percentage of hemoglobin genes (percent.HB), percentage of mitochondrial genes (percent.MT), and percentage of ribosome genes (percent.Ribosome). (c) Scatter plots showing the 2000 high variable genes in red, and the top 10 genes were flagged.

Figure 2

The acquisition of T-cell-related genes. (a) Heatmap showing the expression level of the top five marker genes of each cluster. (b) Umap plots showing the expression level of important marker genes in nine clusters. (c) Umap plots showing the cell-type identification of each cluster.

Figure 3

Enrichment analysis of T-cell-related genes. (a) Bar plot showing the GO analysis of T-cell-related genes. (b) Circos plot showing the log₂ fold change of the main GO terms between BLCA samples and healthy samples. (c) Bar plot showing the KEGG enrichment analysis of T-cell-related genes. (d) Circos plot showing the log₂ fold change of the main KEGG enrichment terms between BLCA samples and healthy samples.

Figure 4

Construction of the prognostic model. (a) Scatter plot showing the univariate COX analysis for screening those genes related to T immune cells with prognostic values. (b) and (c) Line chart showing the LASSO regression analysis of the 12 genes.

Figure 5

Evaluation of the value of the prognostic model. (a) and (b) The Kaplan–Meier survival curve analysis showing the significant difference between the high-risk and the low-risk groups in TCGA and GEO data sets. (c) and (d) Time-dependent ROC analysis showing the significant prognostic value in TCGA and GEO data sets.

Figure 6

Evaluation of the value of the prognostic model. (a) and (b) Risk score distribution of the patients in TCGA and GEO data sets. (c) and (d) Survival status scatter plots showing poorer prognosis of the high-risk group compared with the low-risk group in two data sets. (e) and (f) Multifactorial heat map showing certain clinical features with a significant correlation with the risk score in two data sets. (g) and (h) PCA analysis in two data sets.

Figure 7

Analysis of immune infiltration and immune checkpoint. (a) and (b) The GSEA enrichment analyses of function pathways between two risk groups in two data sets. (c) The expression of the immune checkpoint genes between two risk groups. (d) Boxplot showing the immune score between two risk groups. (e) Boxplot showing the estimated scores between two risk groups. (f) Boxplot showing the tumor purity scores between two risk groups. (g) Boxplot showing the stromal score between two risk groups.

Figure 8

Evaluation of gene mutations between two risk groups: (a) and (b) Heatmaps showing the mutation analysis between high- and low-risk groups in both TCGA and GEO data sets.