Overexpression of the Histone Methyltransferase ΕΖΗ2 in Chronic Lymphocytic Leukemia Confers Protection from Apoptosis and Is Linked to Clinical Aggressiveness

EZH2 is the enzymatic subunit of the polycomb repressive complex 2 (PRC2), which induces gene repression through trimethylation of histone H3 at lysine 27 (H3K27me3). EZH2 over-expression has been reported in a broad range of hematopoietic and solid malignancies and associated with poor prognosis. Recently, we reported for the first time that EZH2 is expressed in CLL and, notably, that EZH2 mRNA and protein levels were up-regulated in clinically aggressive, IGHV-unmutated (U-CLL) stereotyped subset #1 versus indolent, IGHV-mutated (M-CLL) subset #4 and, moreover, that EZH2 expression in subset #1 is regulated by miR-101. Prompted by these preliminary observations, here we sought to investigate in more detail EZH2 expression patterns and functionality in CLL. First, using RQ-PCR, we extended our gene expression analysis to 141 CLL cases, including stereotyped subset #2, the largest stereotyped subset overall (~3% of all CLL) with noted clinical aggressiveness, though mostly concerning M-CLL. We found that bad-prognosis, U-CLL cases express higher EZH2 mRNA levels compared to more indolent, M-CLL cases (fold difference, FD>2, p<0.00001), albeit with variability in EZH2 mRNA levels within each mutational subgroup. Similar results were obtained by Western analysis for EZH2 protein expression, being significantly (p<0.01) higher in U-CLL (n=28) vs M-CLL (n=28), again with intra-subgroup variability. Of note, EZH2 levels where low in the aggressive subset #2 cases, similar to M-CLL, but in sharp contrast to other aggressive, U-CLL cases, including stereotyped subsets #1, #6 and #8. We also extended miR-101 expression analysis to 20 U-CLL and 20 M-CLL cases, in addition to 8 subset #1 and subset #4 cases reported previously, and found that EZH2 mRNA levels were significantly anti-correlated (r=-0.6, p<0.005) with miR-101 levels only in U-CLL, reinforcing the possibility of a regulation of EZH2 expression by miR-101. In order to explore if other polycomb and trithorax complex components, including chromatin modification enzymes and remodeling factors, are dysregulated in CLL, using PCR arrays we analysed the expression of 84 relevant genes in 10 U-CLL and 10 M-CLL cases. Focusing on the main PRC2 components (namely, SUZ12, EED, EZH1, RBBP4/7), we found that their levels were significantly correlated (r=0.49-0.78; p<0.05) with EZH2 levels. Given the variability in EZH2 expression within both M-CLL and U-CLL, for further investigations into the functional impact of EZH2 in CLL, the studied cases were classified into ‘’EZH2 high’’ and ‘’EZH2 low’’ subgroups based on EZH2 mRNA levels (cut-off value determined by ROC curve analysis and Youden Index). In order to explore the effect of EZH2 expression on cell survival, we analyzed the viability of CD19⁺ CLL cells after 9 days in culture and found that CLL clones from 7 ‘’EZH2 high’’ cases displayed significantly (p<0.0001) higher viability compared to 8 ‘’EZH2 low’’ cases. Moreover, ‘’EZH2 high’’ cases (n=19) displayed higher H3K27me3 levels compared to ‘’EZH2 low’’ cases (n=34) (p<0.05). Next, we blocked EZH2 expression using siRNA in CLL cells from 3 ‘’EZH2 high’’ cases and observed downregulation of H3K27me3 levels along with time-dependent increase of cell apoptosis, indicating that EZH2 associates with a survival advantage to CLL cells. In line with this, when we treated CD19⁺ CLL cells from 6 ‘’EZH2 high’’ cases with EZH2 pharmacological inhibitors (GSK-343, EPZ-6438), we found that H3K27me3 levels were decreased in a time- and dose-dependent manner. Moreover, both inhibitors decreased CLL cell viability overtime, suggesting that the histone trimethylation catalytic activity of EZH2 is vital for CLL cell survival. Finally, we searched for potential clinical implications and found that ‘’EZH2 high’’ cases (n=45) showed significantly shorter time-to-first-treatment (p<0.0001) in comparison to ‘’EZH2 low’’ cases (n=62). In conclusion, we demonstrate that EZH2 overexpression is associated with prolonged CLL cell survival and that pharmacological inhibition of EZH2 catalytic activity leads to apoptosis, highlighting a crucial role for EZH2 in CLL cell homeostasis. On these grounds, EZH2 emerges as a novel potential therapeutic target for specific subgroups of CLL.