Transcriptomic data of bevacizumab-adapted colorectal adenocarcinoma cells HCT-116

Abstract

A bioinformatic approach was applied to evaluate the effect of treatment with Bevacizumab on the gene expression profile of colorectal adenocarcinoma cells. The transcriptomic profile of Bevacizumab-adapted HCT-116 (Bev/A) colorectal adenocarcinoma cells was determined and compared with that of the corresponding control cell line by Agilent microarray analysis. Raw data were preprocessed, normalized, filtered, and subjected to a differential expression analysis using standard R/Bioconductor packages (i.e., limma, RankProd). As consequence of Bevacizumab adaptation, 166 differentially expressed genes (DEGs) emerged, most of them (123) resulted downregulated and 43 overexpressed. The list of statistically significant dysregulated genes was used as an input for functional overrepresentation analysis using ToppFun web tool. Such analysis pointed at cell adhesion, cell migration, extracellular matrix organization and angiogenesis as the main dysregulated biological process involved in Bevacizumab-adaptation of HCT116 cells. In addition, gene set enrichment analysis was performed using GSEA, searching for enriched terms within the Hallmarks (H), Canonical Pathways (CP), and Gene Ontology (GO) gene sets. GO terms that showed significant enrichment included: transportome, vascularization, cell adhesion and cytoskeleton, extra cellular matrix (ECM), differentiation and epithelial-mesenchymal transition (EMT), inflammation and immune response. Raw and normalized microarray data were deposited in the Gene Expression Omnibus (GEO) public repository with accession number GSE221948.

Keywords: Bevacizumab; Bioinformatic analysis; Gene expression; HCT116.

Fig. 1

Preliminary plots for sample quality control. Hierarchical clustering (Ward's minimum variance method with Euclidean distance measure) and Principal Component Analysis (PCA) were used as inter-sample distance metrics to detect possible outliers and batch effects (A,B) Dendrogram and PCA representing the full expression dataset in its initial form. Both representations clearly indicated the presence of an outlier sample (Ctrl_5) as well as a marked batch effect separating samples 1 and 2 from the others. (C,D) Dendrogram and PCA representing the expression dataset after outlier removal and batch effect correction. Overall, the procedure was quite effective (as shown by the much shorter height range of C compared to A) and enabled a good separation of the two experimental groups in the principal component space. Blue dots are the Ctrl group representing HCT-116 cells; Red dots are the Res group representing HCT-116-Bev/A (adapted to Bevacizumab) cells.

Fig. 2

Volcano plot showing the results of the differential expression analysis (DEA). DEA was carried out using the RankProd Bioconductor R package to perform a rank product-based statistical analysis. For each unfiltered probe in the array, the –log₁₀(p-value) was plotted against its log₂FC in Res as compared to Ctrl group. Statistically significant DEGs (i.e., genes with an adjusted p-value < 0.05 and |log₂FC| > 0.5) are shown in red. Genes with a positive log₂FC were upregulated in HCT-116-Bev/A compared to HCT-116 control cells.

Fig. 3

Bar chart showing the statistically significant functional terms as returned by the OverRepresentation Analysis (ORA) of our DEG lists (ToppFun by ToppGene Suite, https://toppgene.cchmc.org/, accessed on 25 December 2022). Statistically significant (q-value < 0.05) functional terms related to the Gene Ontology (GO) of Cellular Components (blue bars) and Biological Processes (red bars) have been ranked according to their q-value, represented in the graph as bar length in terms of –log₁₀(FDR). Overall, cell adhesion, cell migration, extracellular matrix organization, and angiogenesis were the main dysregulated biological processes. Details about the precise composition of the enriched gene sets can be found in Table SM2.

Fig. 4

Results of Gene Set Enrichment Analysis (GSEA software v4.2.2 and MSigDB database v7.4) are shown here as a selection of enrichment plots for some of the most representative GO terms belonging to four different functional categories: cell adhesion and cytoskeleton, transportome, inflammation, and immune response. Normalized Enriched Score (NES) for down-regulated terms (the majority) are in blue, while up-regulated terms are in red. In any case, NES was computed using the standard (weighted) scoring method. For a more complete list of all statistically significant (q-value < 0.25) enriched terms returned by GSEA see Table SM3.