Misregulation Of The PRC2 Complex In CML Stem Cells Confers Sensitivity To An EZH2 Inhibitor

Chronic myeloid leukaemia (CML) is a hematological malignancy resulting from the transformation of a primitive hematopoietic progenitor by the fusion oncogene BCR-ABL, a constitutively active tyrosine kinase. In recent years major advances have been made in the treatment of CML with the development of tyrosine kinase inhibitors (TKIs), resulting in high rates of remission in CML chronic phase (CP) patients. However, relapse is driven by quiescent and self-renewing BCR-ABL⁺ CML stem cells (LSCs) that are resistant to TKIs. Consequently, identification of novel proteins or pathways which can be drug-targeted to eliminate the LSCs is a primary goal of current CML research.

Through comparative analysis between CML and non-leukemic samples, we show that components of the repressive Polycomb group (PcG) complex PRC2 are significantly misregulated in CML samples. By performing genome-wide mRNA and epigenetic screens, we demonstrate that this has led to as many as 3-fold more gene repression events in CML cells being associated with gains in the histone modification H3K27me3. This misregulation results in different biological pathways being targeted by PRC2 than those found in non-leukemic samples. We demonstrate that the majority of this misregulation is present in the LSCs.

EZH2 is a key component of the PRC2 complex, responsible for laying down the H3K27me3 mark. To determine the effect of inhibition of the complex on LSC survival we have utilised an inhibitor of EZH2, CPI-625. In the absence and presence of TKI, treatment of CP CML CD34⁺ cells (n=3) with CPI-625 resulted in decreased cell viability (p<0.001 and p<0.05, -/+ TKI respectively) and increased apoptosis (p<0.05 without TKI) in a dose dependent manner. Significantly, there was also a decrease in the number of cells in the undivided, quiescent ‘TKI resistant’ population relative to controls (p<0.01 and p<0.05 -/+ TKI respectively). This was accompanied by an increase in apoptosis (p<0.05 without TKI). Moreover, treatment with CPI-625 resulted in decreasing Colony Forming Cell (CFC) numbers, both in the absence (p<0.05) and presence (p<0.01) of TKI relative to controls. Similar results were seen with treatment of the more primitive CD34⁺38^- cells. Importantly, these effects were not observed in non-leukemic cells. These results demonstrate that CPI-625 is capable of selective targeting of the LSC population.

Our data strongly points to changes in H3K27me3 gene targets in CML as a feature related to misregulation of the PRC2 complex. We have demonstrated that targeting of this complex may have efficacy in the treatment of CML, including eradication of the drug resistant LSCs.