Angiogenesis and evading immune destruction are the main related transcriptomic characteristics to the invasive process of oral tongue cancer

Abstract

Metastasis of head and neck tumors is responsible for a high mortality rate. Understanding its biochemistry may allow insights into tumorigenesis. To that end we carried out RNA-Seq analyses of 5 SCC9 derived oral cancer cell lines displaying increased invasive potential. Differentially expressed genes (DEGs) were annotated based on p-values and false discovery rate (q-values). All 292 KEGG pathways related to the human genome were compared in order to pinpoint the absolute and relative contributions to the invasive process considering the 8 hallmarks of cancer plus 2 new defined categories, as well as we made with our transcriptomic data. In terms of absolute contribution, the highest correlations were associated to the categories of evading immune destruction and energy metabolism and for relative contributions, angiogenesis and evading immune destruction. DEGs were distributed into each one of all possible modes of regulation, regarding up, down and continuum expression, along the 3 stages of metastatic progression. For p-values twenty-six genes were consistently present along the tumoral progression and 4 for q-values. Among the DEGs, we found 2 novel potentially informative metastatic markers: PIGG and SLC8B1. Furthermore, interactome analysis showed that MYH14, ANGPTL4, PPARD and ENPP1 are amenable to pharmacological interventions.

Figure 1

Regulation of the differentially expressed genes. (A) Exclusive, down regulated, continuum and up regulated for p- and q- values. (B) Clusters of gene expression after comparison of the regulated genes from parental and its derived cell lines. In white, DEGs without regulation of expression (exclusive to parental or to derived cell lines and continuum), in blue, down regulated DEGs and in red, up regulated DEGs; light purple, intersections between down regulated and up regulates DEGs.

Figure 2

Contributions of each hallmark of cancer to invasion and metastasis, based on the KEGG pathways from protein coding genes. The intersections show the number of KEGG pathways (upper left panels) or genes (upper right panels). (A) All 292 KEGG pathways of the human genome (upper left panel) or genes related to 35238 human cDNAs annotated in Ensemble (upper right panel) and the distribution of the differentially expressed genes in all-3 comparisons of transformed cell lines (bottom panels, ZsG vs. LN1 left, LN1 vs. LN2 middle and LN2 vs. LN3 right), for p-values; (B) all 292 KEGG pathways of the human genome (upper left panel) or related to 3717 differentially expressed genes after FDR correction found in this study (upper right panel) and the distribution of the differentially expressed genes in all 3 comparisons of transformed cell lines (bottom panels, ZsG vs. LN1 left, LN1 vs. LN2 middle and LN2 vs. LN3 right) for q-values. In upper left panels were highlighted (in red) the changes of the number of KEGG pathways found in the transformed cell lines comparisons.

Figure 3

Individual contributions and percentages, (in parenthesis), of each hallmark to the invasive process. The total percentages are the mean of the 11-percentages of the clusters of gene expression (CoGE) for each hallmark, for (A) p-values, and (B) q-values.

Figure 4

Interactome showing the contributions of the 26-common DEGs displaying altered expression of the metastatic model of OTSSC to the consensual biomarkers associated to epithelial-mesenchymal transition. Highlighted circles represent the common DEGs displaying altered-expression detected in our transcriptome analysis. Highlighted circles in blue are down regulated genes, purple, continuum genes and red, up regulated genes.

Figure 5

Expression profiles of the selected common DEGs displaying altered-expression for each cell line, of 3 independent experiments. (A) RT-PCR data and (B) FPKM data. Turkey’s multiple comparisons test, ****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05.

Figure 6

Clinical data of head and neck cancer from TCGA (n = 248), relative to proposed molecular targets. Low expression of (A) MYH14 and (B) ANGPTL4 (red line) and high expression (black line); (C) low and high expression (red lines) of PPARD; and high (red line) and low expression (black line) of ENPP1, in alive patients. Dotted gray lines represent gene expression of dead patients.