Genome-wide DNA methylation and RNA-seq analyses identify genes and pathways associated with doxorubicin resistance in a canine diffuse large B-cell lymphoma cell line

Abstract

Doxorubicin resistance is a major challenge in the successful treatment of canine diffuse large B-cell lymphoma (cDLBCL). In the present study, MethylCap-seq and RNA-seq were performed to characterize the genome-wide DNA methylation and differential gene expression patterns respectively in CLBL-1 8.0, a doxorubicin-resistant cDLBCL cell line, and in CLBL-1 as control, to investigate the underlying mechanisms of doxorubicin resistance in cDLBCL. A total of 20289 hypermethylated differentially methylated regions (DMRs) were detected. Among these, 1339 hypermethylated DMRs were in promoter regions, of which 24 genes showed an inverse correlation between methylation and gene expression. These 24 genes were involved in cell migration, according to gene ontology (GO) analysis. Also, 12855 hypermethylated DMRs were in gene-body regions. Among these, 353 genes showed a positive correlation between methylation and gene expression. Functional analysis of these 353 genes highlighted that TGF-β and lysosome-mediated signal pathways are significantly associated with the drug resistance of CLBL-1. The tumorigenic role of TGF-β signaling pathway in CLBL-1 8.0 was further validated by treating the cells with a TGF-β inhibitor(s) to show the increased chemo-sensitivity and intracellular doxorubicin accumulation, as well as decreased p-glycoprotein expression. In summary, the present study performed an integrative analysis of DNA methylation and gene expression in CLBL-1 8.0 and CLBL-1. The candidate genes and pathways identified in this study hold potential promise for overcoming doxorubicin resistance in cDLBCL.

Fig 1. Genomic distribution of methylation changes in CLBL-1 and CLBL-1 8.0.

(A) Heatmap depicting DNA methylation pattern between CLBL-1 and CLBL-1 8.0 in the promoter and gene-body regions. A greener color indicates the larger value of peak fold enrichment. (B) Bar chart showing the genomic distribution of CpG sites with altered DNA methylation patterns in CLBL-1 8.0 as compared to CLBL-1. “Upstream” is defined as 5000 bp upstream of Transcription start site (TSS), and “Promoter” is defined as 3000 bp upstream of TSS.

Fig 2. Integrative analysis of genes with differential hypermethylation and differential expression in the promoter and gene-body regions in CLBL-1 8.0.

(A) The total number of hypermethylated/downregulated genes in promoter regions. A total of 24 genes were hypermethylated and downregulated in expression. (B) The total number of hypermethylated/upregulated genes in the gene-body areas. A total of 353 genes were hypermethylated and upregulated in expression.

Fig 3. PPI network and analysis of hub genes of hypermethylated/upregulated genes in gene-body regions belonging to the enriched KEGG pathways in CLBL-1 8.0.

(A) PPI network: the thickness of network edges correlates with the confidential score provided by the STRING database. A thicker edge indicates a higher confidence score of the interaction. (B) Clustering analysis of the PPI network. The most important modules and genes obtained using MCODE in Cytoscape. (C) Hub genes identified by cytoHubba tool kits in Cytoscape. The selected nodes are shown with a color scheme from highly essential (red) to crucial (yellow).

Fig 4. DNA hypermethylation in gene-body regions of TGFBR2, SMAD2, SMURF1, UBE2H, UBE4A, and FGF2 increased their mRNA expression levels in CLBL-1 8.0.

(A-F) DNA methylation of these genes analyzed by MSP. Primer pairs were designed to amplify methylated (M) or unmethylated (U) genomes. (a~f) The mRNA expressions of these genes in CLBL-1 and CLBL-1 8.0 cells. GAPDH and OAZ-1 were used as internal controls and showed similar results. Data standardized by GAPDH alone were presented. Error bars represent the SEM of the results from 3 independent experiments. ** P < 0.01.

Fig 5. Inhibiting TGF-β signaling pathway reduced Dox-resistance of CLBL-1 8.0.

(A, B) Cytotoxicity assay in CLBL-1 and CLBL-1 8.0 cells with the treatment of TGF-β receptor inhibitor, SB505124 (5 μM), and doxorubicin (Dox). Cytotoxicity assay was performed in triplicate, and error bars represent the SEM of the means. n.s., p > 0.05; ***, p ≤ 0.001; Dox: doxorubicin; Control: DMSO vehicle. (C) Representative dot plots showing fluorescence channel analysis and (D) quantitative comparison of Dox accumulation in CLBL-1and CLBL-1 8.0 cells after incubation with free Dox or SB505124 (5 μM) for 2 h at 37°C. Control: no doxorubicin treatment. (E) Suppressing TGF-β/Smad signaling axis by SB505124 (5 μM) for 24 hr reduced the quantities of ABCB1/P-gp expression in CLBL-1 8.0 cells.