151: Genome-Wide CpG Methylation Profiling of Leukemia Cells Treated With Curcuminoids

Statement of Purpose, Innovation or Hypothesis: Dimethoxycurcumin (DMC) is a metabolically stable analog of curcumin (Cur) that demonstrated antitumor activity in leukemia cells. Current DNA methylation analysis methods do not differentiate between DNA methylation and hydroxymethylation.  In this study, we are using the oxidative bisulfite sequencing method, that can differentiate between DNA methylation and hydroxymethylation, to test the effect of DMC and its parent compound Cur on the genome-wide DNA methylation of U937 leukemia cells.

Description of Methods & Materials: Genomic DNA was extracted from control and treated U937 leukemia cells. Treated samples included Cur (5 and 10 uM) and DMC (1 and 2 uM).  The samples were digested with MspI enzyme. Next Generation Sequencing (NGS) was performed using a modified method of Reduced Representation Bisulfite Sequencing (RRBS), where the samples were oxidized before bisulfite treatment. The oxidation step converted cytosine hydroxymethylation into 5-fluorocytosine (5fC), which was converted into thymine after bisulfite treatment. On the other hand, 5mC was not affected by the oxidation step and represented “true” methylation. Bioinformatics analysis: the raw sequence data was quality checked was done using FastQC v 0.11.9. Quality filtering was done using cutadapt v 2.10, and trim_galore python functions, to remove adapter content and simultaneous trimming of low quality ends and filtering for reads below Q20. Diversity trimming of reads is done using NuMetRRBS package using recommended settings for paired end reads. hg38 version of the human genome was used for aligning the reads. Before the alignment the genome, was converted to bisulfite compatible versions using bismark v 0.22.3. Reads were aligned to the genome sequences using bowtie2 v 2.4.4 via the bismark scripts. Methylation calling and extraction of methylation levels was done using bismark methylation extractor. After combining all samples for common regions, 615024 CpG sites were obtained for analysis. Excluding sites on random chromosomes and filtering out sites with less than 2 standard deviation among samples, and sites with known C2T mutations there were, 288097 CpG sites, which were further used for analyzing differential methylation sites.

Data & Results: Both Cur and DMC induced DNA methylation changes in different genomic regions. Cur induced DNA hypomethylation in 118 CpG  islands and increasing its concentration does not significantly increase DNA hypomethylation. On the other hand, DMC induced DNA hypomethylation in 57 CpG islands and increasing its concentration significantly increased DNA hypomethylation and affected 166 CpG islands. The CpG islands showing DNA hypomethylation were mainly located in gene promoter regions as illustrated in Figure 1.

Interpretation, Conclusion or Significance: Cur and DMC induced DNA hypomethylation in specific genomic regions but not in the whole genome. It is not clear yet why DMC and Cur are effective in reversing DNA methylation at specific sites and not able to reverse it at other sites. Dose increase has a significant effect on DMC methylation reversal, while Cur dose increase showed no significant effect. The genes associated with the CpG islands are currently under investigation to identify genes that contribute to leukemia progression.

Disclosures: The authors have no conflicts to disclose.

Citations/References:

Additional Information/Authors: None.