Expression of basement membrane genes and their prognostic significance in clear cell renal cell carcinoma patients

Abstract

Background: Clear cell renal cell carcinoma (ccRCC) is a malignant tumor with limited treatment options. A recent study confirmed the involvement of basement membrane (BM) genes in the progression of many cancers. Therefore, we studied the role and prognostic significance of BM genes in ccRCC.

Methods: Co-expression analysis of ccRCC-related information deposited in The Cancer Genome Atlas database and a BM geneset from a recent study was conducted. The differentially expressed BM genes were validated using quantitative reverse-transcription polymerase chain reaction (qRT-PCR). Least absolute shrinkage and selection operator regression and univariate Cox regression analyses were performed to identify a BM gene signature with prognostic significance for ccRCC. Multivariate Cox regression, time-dependent receiver operating characteristic, Kaplan-Meier, and nomogram analyses were implemented to appraise the prognostic ability of the signature and the findings were further verified using a Gene Expression Omnibus dataset. Additionally, immune cell infiltration and and pathway enrichment analyses were performed using ImmuCellAI and Gene Set Enrichment Analysis (GSEA), respectively. Finally, the DSIGDB dataset was used to screen small-molecule therapeutic drugs that may be useful in treating ccRCC patients.

Results: We identified 108 BM genes exhibiting different expression levels compared to that in normal kidney tissues, among which 32 genes had prognostic values. The qRT-PCR analyses confirmed that the expression patterns of four of the ten selected genes were the same as the predicted ones. Additionally, we successfully established and validated a ccRCC patient prediction model based on 16 BM genes and observed that the model function is an independent predictor. GSEA revealed that differentially expressed BM genes mainly displayed significant enrichment of tumor and metabolic signaling cascades. The BM gene signature was also associated with immune cell infiltration and checkpoints. Eight small-molecule drugs may have therapeutic effects on ccRCC patients.

Conclusion: This study explored the function of BM genes in ccRCC for the first time. Reliable prognostic biomarkers that affect the survival of ccRCC patients were determined, and a BM gene-based prognostic model was established.

Keywords: basement membrane (BM); clear cell renal cell carcinoma; gene expression analysis; gene expression profile; prognostic biomarkers.

Figure 1

The heatmap displaying the DEGs.

Figure 2

The BM genes with prognostic significance in ccRCC.

Figure 3

The RNA levels of (A) COL4A6, (B) NPNT, (C) SEMA3B, (D) ADAMTS4 in HEK-293 and 786-O cells. "*" represents P < 0.05, "****" represents P < 0.0001.

Figure 4

Establishment of the BM genes-based prognostic signature based on the TCGA dataset. (A). The Kaplan-Meier (K-M) curves of low-risk and high-risk ccRCC patients in the TCGA dataset; (B). The time-dependent ROC curves displaying the 1-year, 3-year, and 5-year OS of ccRCC patients in the TCGA dataset; (C). Survival distributions of the TCGA dataset determined according to the median risk score; (D). Heatmap displaying the divergences between low- and high-risk patients of 16 signature genes in the prognostic model for the TCGA dataset.

Figure 5

Verification of the prognostic signature by utilizing the GEO dataset. (A). The Kaplan-Meier curves of low-risk and high-risk ccRCC patients in the GEO dataset; (B). The time-dependent ROC curves displaying the 1-year, 3-year, and 5-year OS of ccRCC patients in the GEO dataset; (C). Survival distributions of the GEO dataset determined according to the median risk score; (D). Heatmap displaying the divergences between low and high-risk patients of 16 signature genes in the prognostic model for the GEO dataset.

Figure 6

The signature could predict the prognosis of ccRCC patients in the TCGA dataset independently. (A). The univariate Cox regression analysis; (B). The multivariate Cox regression analyses showed the associations of the risk score predicting overall survival with clinicopathological indices.

Figure 7

(A, B). The correlations between clinicopathological features and the gene signature.

Figure 8

The K-M curves showed the differences of OS between low- and high-risk patients with different ages, genders, stages, T stages, N stages, M stages or grades.

Figure 9

Establishment of the nomogram. (A). The nomogram; (B). calibration analaysis for predicting1-, 3- or 5-year OS.

Figure 10

Enrichment analyses of DEGs. (A). GO enrichment analysis; (B). KEGG enrichment analysis.

Figure 11

Gene Set Enrichment Analysis analysis.

Figure 12

Differences in infiltration levels of immune cells between low- and high-risk patients.

Figure 13

The different mRNA levels of immune checkpoint genes between low- and high-risk patients, and the "****" represents P < 0.0001.