Overcoming Resistance to 5-Azacytidine in Acute Myelogenous Leukemia

The DNA methylation inhibitor 5-azacytidine (AZA), which is approved for treatment of myelodysplastic syndrome, is also a potential agent for treatment of leukemia; however, drug resistance is an ongoing problem, and mechanisms underlying developing resistance to AZA are poorly understood. Therefore, clarifying the resistance mechanisms is central to establish effective countermeasures.

To probe the mechanisms of resistance to AZA and to develop an effective method for overcoming them, we first generated two AZA-resistant cell lines, THP-1/AR and HL60/AR, from the human acute myelogenous leukemia cell lines THP-1 and HL60. We then studied variations between the parental and resistant lines.

AZA increased the percentages of sub-G₁ and G₂/M-phase cells in the AZA-sensitive parental cell lines; whereas, it had no similar effect in the resistant lines. Consistent with these results, the AZA-induced increases in the levels of cleaved forms of caspase 3, caspase 7, caspase 9, and PARP seen in sensitive cells were diminished in resistant cells. Furthermore, AZA markedly elevated the level of phospho JNK/SAPK in sensitive cells, but not in resistant cells. These results suggest that AZA induced apoptosis as well as G₂/M arrest due to activation of JNK/SAPK signaling, and that induction of these changes was prevented in resistant cells. We also found that the activity as well as protein levels of DNA methyltransferases (DNMTs), which are the main target molecules of AZA, were suppressed by AZA in sensitive cells. However, in resistant cells, this effect was abrogated; and accordingly, AZA-induced up-regulation of p16 gene expression was also negated. These findings thus suggest that resistance was acquired by a DNMT-dependent mechanism. There was no remarkable difference between resistant cells and sensitive cells in the levels of uridine-cytidine kinase 2 (UCK2), which is a key enzyme for conversion of AZA to active form. However, several point mutations were found restrictedly in exon 4 of the UCK2 gene in both resistant cells. These results raised the possibility that the AZA activation process was perturbed due to reduction of UCK activity; and consequently, AZA failed to suppress DNMT in resistant cells. In addition, by microarray analysis, we identified eleven genes that were expressed at significantly different levels in resistant cells versus sensitive cells. Finally, we showed that the histone deacetylase inhibitor romidepsin induced p16 gene expression and increased the levels of apoptosis-related molecules, while suppressing growth in both sensitive and resistant cell lines. An isobologram analysis demonstrated that simultaneous administration of AZA and romidepsin resulted in an additive inhibitory effect on both AZA-sensitive and AZA-resistant cell growth. These results suggest that romidepsin can overcome AZA resistance; therefore, the combination of AZA and romidepsin not only augments the anti-leukemia effect but also prevents acquisition of resistance to AZA.

Newly established 5-azacytidine-resistant cell lines THP-1/AR and HL60/AR are good models to analyze the mechanisms of drug resistance to 5-azacytidine. Using these cell lines, we revealed that acquisition of resistance is primarily caused by a DNMT-dependent mechanism, which can be surmounted with addition of romidepsin. It is likely that the combination of AZA and romidepsin can prevent patients from acquiring resistance to AZA while augmenting its anti-leukemia therapeutic effect.

No relevant conflicts of interest to declare.