Comparison of DNA Methylation and Expression Status Prior Azacytidine Treatment and their Relationship to Overall Survival and Clinical Response of MDS Patients

Myelodysplastic syndromes (MDS) are clonal disorders of hematopoietic stem cells characterized by ineffective hematopoiesis. High-risk MDS patients are treated by hypomethylating agents, of which they benefit significantly. However, only half of the patients respond positively to the treatment. Aberrant DNA methylation and mRNA expression in MDS were documented in several studies, but their prognostic impact in response to hypomethylating therapy is still unclear. The aim of the project was to find a relationship between methylation and expression status prior to azacytidine (AZA) treatment and the overall survival and clinical response of MDS patients.

We performed methylation and expression profiling in CD34+ cells from 30 samples from MDS patients before AZA treatment and after 4-8 treatment cycles. HumanMethylation27 BeadChips and HumanHT-12 v4 Expression BeadChips (Illumina) were used to generate profiles. DNA and RNA were isolated from same CD34+ cells separated from bone marrow by magnetic beads. The β-values represent quantitative measurements of DNA methylation levels of specific CpGs, and range from 0 for completely unmethylated to 1 for completely methylated DNA. The nonparametric Mann-Whitney test was used for comparison of β-values and expression levels between responders and nonresponders.

To determine whether DNA methylation and expression might predict a response to AZA treatment, we compared methylation and expression status at baseline with clinical responses in 30 MDS patients. Twelve patients of 30 (40%) achieved complete remission or partial remission, 10 had stable disease (33.3%), and 8 showed progression (26.7%). Median survival after initiation of AZA treatment in progression patient group was 8.7 months, stable group 21.2 months, and group with complete or partial remission 24.5 months. We found significant differences in methylation status in 20 genes (p<0.05) between groups of responders and nonresponders and the largest methylation differences showed CALCA (0.61 vs. 0.16, p<0.05), MAGEE2 (0.71 vs. 0.30, p<0.05), HMP19 (0.62 vs. 0.23, p<0.05), MEOX1 (0.36 vs. 0.84, p<0.05), and KCNQ1DN genes (0.33 vs. 0.84, p<0.05). The aberrant expression status did not correlate with the response to AZA. We also measured methylation changes caused by AZA treatment. In the group of patients with progression, we did not find any change in the methylation profile after treatment. On the contrary, we found significant methylation changes after AZA treatment in the group of patients responding to treatment (e.g. AMT, NOTCH, and WT1genes).

Our finding of different DNA methylation levels at baseline between groups of responders and nonresponders as well as detection of decreased methylation after AZA treatment in the group of patients with clinical response may represent useful prediction markers of treatment success. However, the data require detailed examination along with confirmative cohort of patients.

Supported by grant (NT/13899, NT/14377, NT/14539, NT/13847) and the project for conceptual development of research organization (00023736) from the Ministry of Health of the Czech Republic.

No relevant conflicts of interest to declare.