EZH2 upregulates the PI3K/AKT pathway through IGF1R and MYC in clinically aggressive chronic lymphocytic leukaemia

ABSTRACT

EZH2 is overexpressed in poor-prognostic chronic lymphocytic leukaemia (CLL) cases, acting as an oncogene; however, thus far, the EZH2 target genes in CLL have not been disclosed. In this study, using ChIP-sequencing, we identified EZH2 and H3K27me3 target genes in two prognostic subgroups of CLL with distinct prognosis and outcome, i.e., cases with unmutated (U-CLL, n = 6) or mutated IGHV genes (M-CLL, n = 6). While the majority of oncogenic pathways were equally enriched for EZH2 target genes in both prognostic subgroups, PI3K pathway genes were differentially bound by EZH2 in U-CLL versus M-CLL. The occupancy of EZH2 for selected PI3K pathway target genes was validated in additional CLL samples (n = 16) and CLL cell lines using siRNA-mediated EZH2 downregulation and ChIP assays. Intriguingly, we found that EZH2 directly binds to the IGF1R promoter along with MYC and upregulates IGF1R expression in U-CLL, leading to downstream PI3K activation. By investigating an independent CLL cohort (n = 96), a positive correlation was observed between EZH2 and IGF1R expression with higher levels in U-CLL compared to M-CLL. Accordingly, siRNA-mediated downregulation of either EZH2, MYC or IGF1R and treatment with EZH2 and MYC pharmacological inhibitors in the HG3 CLL cell line induced a significant reduction in PI3K pathway activation. In conclusion, we characterize for the first time EZH2 target genes in CLL revealing a hitherto unknown implication of EZH2 in modulating the PI3K pathway in a non-canonical, PRC2-independent way, with potential therapeutic implications considering that PI3K inhibitors are effective therapeutic agents for CLL.

KEYWORDS:

• Chronic lymphocytic leukemia
• EZH2
• ChIP sequencing
• PI3K pathway
• IGFR1

Acknowledgments

The sequencing was performed by the SNP&SEQ Technology Platform in Uppsala. The facility is part of the National Genomics Infrastructure (NGI) Sweden and Science for Life Laboratory. This work is supported by grants from LUA/ALF (FoU Västra Götalandsregionen) and Swedish Cancer Society (CAN 2016/424 to M.K and CAN 2016/414 to R.R.). This work was also supported in part by the grants from the Knut and Alice Wallenberg Foundation (KAW 2014.0057 and KAW 2016.0373) and Swedish Research Council (Dnr 2017-02834 to C.K and Dnr 2016-02064 to R.R.); “ Co-financed by Greece and the European Union (European Social Fund- ESF) through the Operational Programme «Human Resources Development, Education and Lifelong Learning» in the context of the project “Reinforcement of Postdoctoral Researchers” (MIS-5001552), implemented by the State Scholarships Foundation (ΙΚΥ); TRANSCAN CGH-CLL; and a research grant from Gilead Sciences Hellas (Asklepios award) to N.P.

Authorship contributions

MK, SK and MH performed research, analyzed the data and wrote the paper. SK did all the bioinformatics analysis in this study. NS and LM provided clinical samples and associated data. NP and CK analyzed data and wrote the paper. KS, RR and MK contributed to the design and analysis of the study and wrote the paper.

Disclosure statement

KS received research support from Jannsen Pharmaceutica. Remaining authors have no conflicts of interests or disclosures.

Supplementary material

Supplemental data for this article can be accessed here.

Additional information

Funding

This work was supported by the Cancerfonden [CAN 2016/424].