YPEL2 regulates the efficacy of BRD4-EZH2 dual targeting in EZH2Y641mut germinal center-derived lymphoma

Abstract

A significant proportion of diffuse large B cell lymphoma (DLBCL) and follicular lymphoma (FL) cases harbor a gain-of-function, heterozygous somatic mutation of the methyltransferase gene EZH2. While this factor is known to cooperate with the proto-oncogene MYC during malignant B cell development, the effect of interfering with both factors remains underexplored. Here we undertook the simultaneous evaluation of two epigenetic drugs targeting EZH2 methyltransferase activity and BRD4-mediated control of MYC transcription, CPI169 and CPI203, using preclinical models of DLBCL and FL with distinct EZH2 mutational status. We observed a specific and synergistic antiproliferative effect of these compounds in EZH2-mutated cells and mouse xenograft models, that was related to the abrogation of MYC transcriptional program and to tumor cell proliferation blockade at the G1 cell cycle phase. Gene expression profile, exploratory data analysis, and siRNA screening identified the PI3K/AKT-regulated gene and mitosis regulator, YPEL2, as a crucial factor involved in the efficacy of MYC/EZH2 dual targeting both in vitro and in vivo. Altogether, our results provide first pre-clinical evidence that simultaneous targeting of MYC and EZH2 is a safe and efficient approach that can be monitored by specific biomarkers, in aggressive lymphoid tumors of germinal center origin.

Keywords: Biomarkers; Epigenetics; Mouse model; Non-Hodgkin lymphoma; Targeted therapies.

Fig. 1

CPI169 is active in EZH2^Y641mut cell lines of germinal center origin independently of EZH2 expression level and activity. (A) MTT analysis in follicular lymphoma (FL) and diffuse large B cell lymphoma (DLBCL) cell lines treated with the indicated concentrations of CPI169 for 7 days and treatment was renewed at day 4. Data are expressed as mean ± SD of 3 replicates. (B) Western blot analysis of EZH2 and H3K27me3 in EZH2^Y641mut (KARPAS-422, SUDHL-6 and RL) and EZH2 ^wt (SUDHL5) cells. Tubulin was used as a loading control. (C) Western blot analysis of H3K27me in EZH2 ^wt (HT) and EZH2^Y641mut cells (SUDHL-6, RL, and KARPAS-422) treated with increasing doses of CPI169 for 4 days. H3 was used as a loading control. Values refer to the fold-change vs control (DMSO-treated) of the H3K27me3/H3 ratio. (D) Western blot analysis of H3K27me3 in KARPAS-422 treated with increasing doses of CPI169 for 4 and 7 days. H3 was used as a loading control. Values refer to the fold-change vs control (DMSO-treated) of the H3K27/H3 ratio. (E) qPCR analysis of ABAT, MPEG1 and TNFRSF21 mRNA levels in HT and KARPAS-422 cells treated with 1.5 µM CPI169 for 4 days. GAPDH was used as housekeeping gene. Values refer to the fold-change vs control (DMSO-treated) of the above-mentioned genes/GAPDH ratio. Comparisons are made between CPI169-treated vs. control (DMSO-treated) cells. Data are expressed as mean ± SD of 3 replicates. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 2

EZH2 mutation at Y641 regulates cell sensitivity to CPI169. (A) Western blot analysis of H3K27me3 and EZH2 from HT and SUDHL5 cells expressing FLAG-tagged EZH2^Y641mut, WT EZH2, or empty vector (EV). H3 and Tubulin was used as a loading control. Values refer to the fold-change vs control (EV) of the H3K27me3/H3 ratio or the EZH2/Tubulin ratio. (B) Western blot analysis of H3K27me3 from SUDHL5 and HT cells expressing FLAG-tagged EZH2^Y641mut or WT EZH2 and treated with 1.5 µM CPI169 for 4 days. H3 was used as a loading control. Values refer to the fold-change vs control (WT DMSO-treated) of the H3K27me3/H3 ratio. (C) Scatter plot from Gene Expression Profile (GEP) of HTEZH2^Y641mut vs HT-EV cells. (D) MTT analysis in SUDHL5 EZH2^Y641mut and SUDHL5 ^wt cells treated with the indicated concentrations of CPI169 for 7 days and renewed at day 4. Data are expressed as mean ± SD of 3 replicates. (E) GEP scatter plot from EZH2^Y641mut (KARPAS-422, SUDHL-6, RL, and HT EZH2^Y641mut) and EZH2 ^wt (HT and SUDHL5) cells treated with 1.5 µM CPI169 for 4 days. Red and blue dots represent up- and downregulated genes, respectively in EZH2^Y641mut cells. (F) Left panel: GSEA enrichment plots of mutated EZH2 signatures showing a correlation with sensitivity to CPI169. Right panel: heatmap showing genes differentially regulated upon CPI169 treatment in either EZH2 ^wt (HT, SUDHL5) or EZH2^Y641mut (RL, KARPAS-422, SUDHL6, and HT^Y641N) cell lines. (G) qPCR analysis of ABAT, FGGY, GNB4, HSPA1A, KLHL14, ORC1, PECI, RALB, SBF2 and YPEL2 mRNA levels in EZH2 ^wt (HT, SUDHL5) and EZH2^Y641mut (RL, KARPAS-422, SUDHL6) cells treated with 1.5 μM CPI169 for 4 days. β-actin/B2 M were used as housekeeping genes. Values refer to the fold-change vs control (DMSO-treated). Comparisons are made between CPI169-treated vs. control (DMSO-treated) cells. Data are expressed as mean ± SD of 3 replicates. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 3

The combination of CPI169 with the BET inhibitor CPI203 elicits a synergistic cell proliferation blockade in EZH2-mutated FL and DLBCL in vitro and in vivo models. (A) MTT analysis in FL and DLBCL cell lines treated daily with the indicated concentrations of CPI169 for 7 days and 0.1 or 0.5 μM of CPI203 during the last 2 days. Combination indexes (CIs) were determined by the Chou and Talalay's algorithm (Calcusyn software). Combinational effect was considered significantly synergistic when CI < 0.8. Data are expressed as mean ± SD of 3 replicates. (B) Cell Cycle analysis of EZH2^Y641mut cell lines treated with the indicated concentrations of CPI169 for 7 days and 0.1-0.5 μM of CPI203 for 2 days. (C) DLBCL and FL cells were treated with 1.5 µM CPI169 for 96 h and/or 0.5 µM CPI203 for 6 h prior to sample collection for gene expression profiling. Left panel: GSEA enrichment plots of the EZH2^Y641mut signatures showing a correlation with sensitivity to CPI169 and CPI203. Right panel: heatmap showing a global enrichment in upregulated genes in EZH2^Y641mut cells exposed to CPI169/CPI203 combo, when compared to single-agent therapies Top 45 up- and down-modulated genes are represented. (D) Recording of tumor volumes in the different treatment arms during the 3-week treatment, where means ± SEM are plotted. (E) Mean tumor weight at the day of sacrifice in the control (n = 5 animals) and the distinct drug treatment (n = 6 animals) groups. Comparisons are made between treated cells vs. control (DMSO-treated cells or vehicle-treated mice). Data are expressed as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 4

YPEL2 upregulation is required for the synergistic antitumor activity of CPI169 and CPI203 in EZH2^Y641mut B-NHL. (A) siRNA screening of ABAT, KLHL14, PECI, and YPEL2 in KARPAS-422 and RL cells exposed to CPI169 (1.5 μM), CPI203 (0.5 μM) or the drug combination for 72 h, as assessed by MTT assay. Data are expressed as fold change vs control (scramble siRNA, not treated, siCtl, n = 3 biological replicates). (B) Western blot analysis of YPEL2 and (C) qPCR determination of YPEL2, GADD45A and CDKN1A mRNA levels in KARPAS-422 and RL cells treated by 1.5 μM CPI169 for 96 h and/or 0.5 µM CPI203 for 48 h. GAPDH was used as a loading control. Values are referred to control untreated cells. (D) Western blot analysis of YPEL2 in representative tumor samples (N = 3 tumors per group). GAPDH was used as a loading control. Right panel: densitometric quantification of the YPEL2/GAPDH ratio. Values are referred to control (vehicle-treated) tumors. (E) qPCR analysis of YPEL2, GADD45A, and CDKN1A mRNA levels in representative tumor samples (n = 3 tumors per group). GAPDH was used as a housekeeping gene. Values are referred to control (vehicle-treated) tumors. (F) Western blot and (G) qPCR analyses of YPEL2 and cell cycle-related markers in KARPAS-422 and RL cell lines electroporated with either siYPEL2 or siCtl for 8 h. Tubulin and β-actin/B2 M were used as loading control and housekeeping gene, respectively. Values are referred to control (siCtl-transfected) cells. Data are expressed as mean ± SD of 3 replicates. *P < 0.05, **P < 0.01, ***P < 0.001.