Decitabine (5-aza-2′-deoxycytidine, DAC) is a cytidine analogue that inhibits DNA methylation by trapping DNA methyltransferases. DAC is cytotoxic at high doses, hypomethylating at low doses, and has clinical activity in myeloid malignancies that appears to be optimal at low doses. In a phase II study, we have been evaluating the activity of single agent DAC (at 15 mg/m2 IV over one hour daily, 5 days a week for two weeks) in Imatinib refractory or intolerant chronic myelogenous leukemia (CML). Thirty five patients were enrolled (12 in chronic phase [CP], 17 in accelerated phase [AP] and 6 in blastic phase [BP]). Complete hematologic responses were seen in 12 patients (34%), partial hematologic responses in 7 patients (20%) and hematologic improvement in 4 patients (11%), for an overall hematologic response rate of 65% (83% in CP, 59% in AP and 50% in BP). Major cytogenetic responses were observed in 7 patients (20%), and minor cytogenetic responses were seen in 9 patients (26%) for an overall cytogenetic response rate of 46%. Median response duration was 3 months, ranging from 2 to 13+ months, and responses are ongoing in 4 patients. The only common grade 3 or 4 toxicities observed were related to myelosuppression. Neutropenic fever developed in 18 patients (51%), in 27/115 (23%) courses of therapy. There were two deaths on study, both related to thrombocytopenia and hemorrhage. We used a bisulfite/pyrosequencing assay to measure DNA methylation across LINE repetitive elements in peripheral blood mononuclear cells as a surrogate marker for global methylation in CML. LINE methylation had a mean +/− standard error of the mean of 71.2+/−1.2% prior to therapy (range 56.5–85.8). Methylation decreased to 60.1+/−1.44% after 1 week, 49.7+/−2.9 after 2 weeks, 50.1+/−2.7% after 3 weeks and returned to 67.2+/−2.1% at recovery of counts (median, 46 days). The dynamics of hypomethylation following DAC (decrease of about 10% at week 1, 10–20% at weeks 2 and 3 and recovery by week 6) were similar after cycle 2 and cycles >2, with no evidence of cumulative hypomethylation or development of resistance to the hypomethylating effect of DAC. LINE methylation at the end of week 1 did not correlate with subsequent responses. However, the degree of LINE hypomethylation at the end of therapy (day 12) was paradoxically higher in patients who did not subsequently respond to therapy. Thus, the absolute decrease in methylation was 17.7% vs. 28.6% in responders vs. non-responders (p=0.009) and the relative decrease in methylation was 25.0% vs. 39.5% responders vs. non-responders (p=0.016). These results are consistent with a cell death mechanism of response, whereby the most hypomethylated cells die rapidly in patients sensitive to therapy, while cells resistant to DAC can withstand higher degrees of hypomethylation without cell death. Consistent with this concept, in patients who responded and subsequently lost their response while on therapy, the degree of hypomethylation achieved in the cycle immediately prior to relapse was similar or higher than that achieved in prior cycles, indicating that DAC resistance in this population is not pharmacologic, but likely biologic resistance to hypomethylation-induced cell death. We conclude that DAC induces hypomethylation in-vivo and has significant clinical activity in Imatinib refractory CML. The inverse correlation between hypomethylation at day 12 and response is consistent with a cell death mechanism of action of DAC in CML.

Author notes

Corresponding author

Sign in via your Institution