Spotlight on differentially expressed genes in urinary bladder cancer

Abstract

Introduction: We previously identified common differentially expressed (DE) genes in bladder cancer (BC). In the present study we analyzed in depth, the expression of several groups of these DE genes.

Materials and methods: Samples from 30 human BCs and their adjacent normal tissues were analyzed by whole genome cDNA microarrays, qRT-PCR and Western blotting. Our attention was focused on cell-cycle control and DNA damage repair genes, genes related to apoptosis, signal transduction, angiogenesis, as well as cellular proliferation, invasion and metastasis. Four publicly available GEO Datasets were further analyzed, and the expression data of the genes of interest (GOIs) were compared to those of the present study. The relationship among the GOI was also investigated. GO and KEGG molecular pathway analysis was performed to identify possible enrichment of genes with specific biological themes.

Results: Unsupervised cluster analysis of DNA microarray data revealed a clear distinction in BC vs. control samples and low vs. high grade tumors. Genes with at least 2-fold differential expression in BC vs. controls, as well as in non-muscle invasive vs. muscle invasive tumors and in low vs. high grade tumors, were identified and ranked. Specific attention was paid to the changes in osteopontin (OPN, SPP1) expression, due to its multiple biological functions. Similarly, genes exhibiting equal or low expression in BC vs. the controls were scored. Significant pair-wise correlations in gene expression were scored. GO analysis revealed the multi-facet character of the GOIs, since they participate in a variety of mechanisms, including cell proliferation, cell death, metabolism, cell shape, and cytoskeletal re-organization. KEGG analysis revealed that the most significant pathway was that of Bladder Cancer (p = 1.5×10(-31)).

Conclusions: The present work adds to the current knowledge on molecular signature identification of BC. Such works should progress in order to gain more insight into disease molecular mechanisms.

Figure 1. Representative immunohistochemical analysis for tumors of T1/grade 2, T1/grade 3, T2/grade 3.

Tumors were stained with anti-cerbB2, anti-Ki67, anti-p53 and anti-Cyclin D1. H&E, Representative hematoxylin-eosin slides.

Figure 2. Two-way average-linkage hierarchical clustering with Euclidian distance for the genes of interest (GOIs) in 10 TCC samples vs. 5 controls.

Each row in the diagram represents a gene and each column a tumor sample. The color saturation represents differences in gene expression across the tumor samples; red indicates expression higher than the median (black), and green indicates expression lower than the median. The color intensity indicates degree of gene regulation (A). Fold expression of the GOIs with respect to tumor histology, as detected by microarray analysis (B).

Figure 3. The Scatter Plot graphs the expression level (2^∧-ΔCt) of each gene in BC samples versus the control samples.

The black line indicates fold changes of 1. The pink lines indicate the threshold in gene expression fold-change (2-fold difference). qPCR revealed up- and down-regulated genes in urinary bladder cancer. Total RNA from the normal adjacent tissue and urinary bladder cancer were characterized in technical triplicates, and the relative expression levels for each gene in the two tissue types were plotted against each other in the Scatter Plot. VEGFA, OPN, p14^INK4A, p6^CDKN2A, NRAS and TGFβ1 were up-regulated, while FGF2 and EGF were down-regulated by at least two-fold (outside the purple lines). The genes KRAS, MMP2, AKT1, p53, EGFR, MMP9 and HRAS exhibited equal expression between bladder cancer and normal tissue (A). The Volcano Plot graphs the log₂ of the fold change in expression of each gene between the BC samples versus its p-value from the t-test. The black line indicates fold changes of 1. The pink lines indicate the threshold in gene expression fold-change (2-fold difference). The blue line indicates the threshold for the p-value of the t-test (0.01) (B).

Figure 4. Scatterplot depicting the mRNA levels of the genes that were over-expressed in various urinary bladder cancers of different stage/grade, versus normal tissue.

Groups pairs were statistically compared using the Mann-Whitney U test. Bars depict the median values (A). Scatterplot depicting the mRNA levels of the genes that were equally expressed among urinary bladder cancers of various stage/grade, and normal tissue. Groups pairs were statistically compared using the Mann-Whitney U test. Bars depict the median values (B). Scatterplot depicting the mRNA levels of the genes that were under-expressed in various urinary bladder cancers of different stage/grade, versus normal tissue. Group pairs were statistically compared using the Mann-Whitney U test. Bars depict the median values (C).

Figure 5. Representative bands depicting the expression of SPP1/OPN (58 kDa), TGFβ1 (25 kDa), VEGFA (15 kDa), CDKN2A (19 kDa), HRAS (21 kDa) and p53 (53 kDa) proteins in 4 BC and 4 control samples, using Western blotting.

β-actin (43 kDa) was used as loading control.

Figure 6. Kaplan-Meier curves depicting overall survival (%) of the urinary bladder cancer patients, regarding the mRNA expression of the GOIs.

The cases whose tumors exhibited low levels of VEGFA (median value, 0.141) and EGFR (>median value, 0.0051) mRNA levels [for VEGFA, p = 0.04; for EGFR, p = 0.02; Log-rank (Mantel Cox) test]. Differences in survival were assessed using Log-rank (Mantel Cox) test. Statistical significance was set at the 95% level (p<0.05).

Figure 7. GO analysis of the selected genes revealed functions related to cell death, cell proliferation, metabolism and signal transduction.

Figure 8. Further resolution of GO functions showed that the majority of genes (∼30%) are related to metabolism.

However, despite the small size of the dataset that we used, it appeared that those selected genes participate significantly in a variety of cellular functions.