Abstract 3873


Disruption of normal DNA methylation, including both gene specific hypermethylation and genome-wide hypomethylation, is found in most malignant tumors. Most epigenetic studies in chronic lymphocytic leukemia (CLL) have been focused in CpG islands and gene promoter regions, and have identified hypomethylated genes, such as BCL2 or TCL1, and hypermethylated genes, such as GRM7. However, the quantification of overall methylation measured as levels of 5-methylCytosine (5mC) has been poorly explored. As compared to their normal counterparts (CD19+ B cells), overall hypomethylation has been observed in CLL neoplastic cells. Importantly, the overall methylation varies among patients but its clinical significance has not been widely investigated. In addition, it is known that microRNA (miRNA) expression is altered in CLL, and that and epigenetic mechanisms, such as methylation, can affect miRNA expression.


To investigate the prognostic impact of overall methylation in patients with CLL and to analyze the correlation of 5mC levels with miRNAs expression.


We analyzed 73 CLL patients (median age, 69 [range, 34–86]; 43% males) diagnosed in our institution between 1992 and 2007. The median follow up was 10.5 years. The level of global methylation in total DNA was estimated after determination of percentage of 5mC using anti-5mC monoclonal antibodies (MethylFlash Methylated DNA Quantification Kit, Epigentek). The expression of 377 mature miRNAs was analyzed using TaqMan Array Human MicroRNA A Card v2.0 (Applied Biosystems). Statistical analysis was performed with SPSS version 15.0.1 and R software version 2.9.0. MaxStat package of R were used to determine the optimal cutoffs and Quantitative trail function in BRB array tools to correlate miRNA expression and methylation levels.


The analysis of methylation levels showed a wide distribution of methylation degree among patients (median: 3.02%, range: 0.58–6.14%). From the clinical standpoint, methylation levels were only correlated with Binet clinical stage, patients with C stage showing a higher degree of methylation (p=0.015).

Using MaxStat, we identified two cutoffs which classified patients as having low, medium or high degree of methylation. Mean progression-free survival (PFS) was 8.4 years (95% CI: 6.4–10.4), 6.2 years (95% CI: 4.7–7.7) and 3.2 years (95% CI: 2.4–4.8) for patients with low, medium, and high methylation levels, respectively (p=0.013). In the multivariate analysis for PFS (including ZAP70, IGHV, Age≤65, cytogenetics and global methylation), high ZAP70 expression (HR: 3; 95%CI: 1.1–7.9; p=0.026) and high global methylation (HR: 5.4 95%CI: 1.7–17.1; p=0.004) were independent unfavorable prognostic factors, while a significant trend was observed for high-risk cytogenetics (17p-, 11q-, +12) (p=0.054). Interestingly, methylation levels retained its prognostic significance in subgroup analysis: clinical stage A (p=0.06) and B/C (p=0.009); mutated (p=0.008) and unmutated IGHV (p=0.028); low (p=0.028) and high ZAP70 (p=0.001); and low-risk (normal karyotype, 13q-)(p=0.008) and high-risk (17p-, 11q-, +12) cytogenetics (p=0.001).

Finally, we identified a 4-miRNA signature associated with global methylation levels: miR-103 (Spearman correlation [SC]: −0.821;p=0.03), miR-132 (SC: 0.786;p=0.05), miR-494 (SC: −0.786; p=0.02), and miR-193a-5p (SC: 0.786; p=0.05). Interestingly, miR-103, miR-132 and miR-494 are located in subtelomeric regions, which are known to be more susceptible to overall methylation changes.


In this study, the degree of global DNA methylation was an independent prognostic factor for PFS in patients with CLL. The analysis of overall methylation could be useful not only for the prognosis of patients with CLL but also in the monitoring of clinical trials in which hypomethylating agents (e.g., decitabine) are being investigated as CLL therapy. The correlation between overall methylation levels and certain miRNAs may be a surrogate marker of epigenetic lesions and deserves further investigation.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.