Identification of Overexpressed Genes in Malignant Pleural Mesothelioma

Abstract

Malignant pleural mesothelioma (MPM) is a fatal tumor lacking effective therapies. The characterization of overexpressed genes could constitute a strategy for identifying drivers of tumor progression as targets for novel therapies. Thus, we performed an integrated gene-expression analysis on RNAseq data of 85 MPM patients from TCGA dataset and reference samples from the GEO. The gene list was further refined by using published studies, a functional enrichment analysis, and the correlation between expression and patients' overall survival. Three molecular signatures defined by 15 genes were detected. Seven genes were involved in cell adhesion and extracellular matrix organization, with the others in control of the mitotic cell division or apoptosis inhibition. Using Western blot analyses, we found that ADAMTS1, PODXL, CIT, KIF23, MAD2L1, TNNT1, and TRAF2 were overexpressed in a limited number of cell lines. On the other hand, interestingly, CTHRC1, E-selectin, SPARC, UHRF1, PRSS23, BAG2, and MDK were abundantly expressed in over 50% of the six MPM cell lines analyzed. Thus, these proteins are candidates as drivers for sustaining the tumorigenic process. More studies with small-molecule inhibitors or silencing RNAs are fully justified and need to be undertaken to better evaluate the cancer-driving role of the targets herewith identified.

Keywords: MPM; RNAseq; gene signature; malignant pleural mesothelioma; overexpressed genes; therapeutic targets.

Figure 1

Flowchart of the DEG selection process.

Figure 2

(A) Volcano plot. Cut-off criteria: p-value < 0.05 and |log2FC| > 0.38. The differentially expressed genes are in red (high-expressed) and in green (low-expressed), while the insignificantly changed genes are in black. (B) Principal component analysis (PCA) plot of 85 MPM patients (in black) and 3 normal lung samples (in grey). The PCA score plot showed that samples from MPM patients and controls were clustered separately.

Figure 3

Heatmap and one-dimensional hierarchical clustering of the 839 filtered differentially expressed genes (DEGs) across MPM patients and a group of 3 nonmalignant lung samples. The genes are displayed in rows and samples are displayed in columns. High-expressed genes are in red; low-expressed genes are in green.

Figure 4

Bar plots showing significantly enriched GO terms associated with the (A) biological process, (B) molecular function, and (C) pathway categories in MPM. The number of DEGs for each term is indicated in bold alongside.

Figure 5

Kaplan-Meier (KM) survival curve of samples divided into high- (red) and low-risk (blue) groups according to the median (log-rank test p-value < 0.0001). (A) KM of unreduced gene signature; (B) KM of FDR-reduced gene signature; (C) KM of GO-reduced gene signature.

Figure 6

Time-dependent receiver operating characteristic (ROC) curve for predicting 2-year survival.

Figure 7

The densitometry ratios of proteins encoded by genes obtained from in silico analyses normalized versus GAPDH. (A) Proteins overexpressed in at least 5 MPM cell lines; (B) proteins overexpressed in 4 or 3 MPM cell lines; (C) proteins overexpressed in only 2 or 1 MPM cell lines. Data shown in this figure were reproduced independently 3 times. Corresponding blots are reported in Figure S1 and Supplementary File S1.