Three-dimensional culture sensitizes epithelial ovarian cancer cells to EZH2 methyltransferase inhibition

Abstract

Inhibitors of EZH2 methyltransferase activity have been demonstrated to selectively suppress the growth of diffused large B cell lymphoma (DLBCL) cells with gain-of-function mutations in EZH2, while exhibiting very limited effects on the growth of DLBCL cells with wild-type EZH2. Given that EZH2 is often overexpressed but not mutated in solid tumors, it is important to investigate the determinants of sensitivity of solid tumor cells to EZH2 inhibitors. In the current study, we show that three-dimensional (3D) culture of epithelial ovarian cancer (EOC) cells that overexpress EZH2 sensitizes these cells to EZH2 methyltransferase inhibition. Treatment of EOC cells with GSK343, a specific inhibitor of EZH2 methyltransferase, decreases the level of H3K27Me3, the product of EZH2's enzymatic activity. However, GSK343 exhibited limited effects on the growth of EOC cells in conventional two-dimensional (2D) culture. In contrast, GSK343 significantly suppressed the growth of EOC cells cultured in 3D matrigel extracellular matrix (ECM), which more closely mimics the tumor microenvironment in vivo. Notably, GSK343 induces apoptosis of EOC cells in 3D but not 2D culture. In addition, GSK343 significantly inhibited the invasion of EOC cells. In summary, we show that the 3D ECM sensitizes EOC cells to EZH2 methyltransferase inhibition, which suppresses cell growth, induces apoptosis and inhibits invasion. Our findings imply that in EZH2 wild-type solid tumors, the ECM tumor microenvironment plays an important role in determining sensitivity to EZH2 inhibition and suggest that targeting the ECM represents a novel strategy for enhancing EZH2 inhibitor efficacy.

Keywords: 3D culture; EZH2; EZH2 inhibitor GSK343; apoptosis; epithelial ovarian cancer.

Figure 1. The EZH2 inhibitor GSK343 exhibits limited effects on the growth of human EOC cells under conventional 2D monolayer culture. (A) Expression of EZH2 and β-actin in two individual batches of normal human ovarian surface epithelial (HOSE) cells and indicated human EOC cell lines was determined by immunoblotting. (B) Same as (A) but for H3K27Me3 and histone H3 expression determined by immunoblotting. (C) Expression of H3K27Me3, H3K9Me3 and histone H3 was determined in the indicated EOC cell lines by immunoblotting after 3 d of treatment with GSK343 (1 μM) or vehicle control (0.1% DMSO). (D) Same as (C) but for EZH2 and β-actin expression determined by immunoblotting. (E–G) Cell growth curves for OVCAR10, UPN289 and SKOV3 cell lines treated with GSK343 (1 μM) or vehicle control (0.1% DMSO) over 12 d. Media was changed and cells counted every 3 d, time points represent the mean of three independent experiments with SD, *p < 0.05.

Figure 2. The EZH2 inhibitor GSK343 significantly suppresses the growth of human EOC cells in 3D cultures. (A) Indicated EOC cells were cultured in 3D Matrigel and treated with GSK343 (1 μM) or vehicle control (0.1% DMSO) for 12 d. 3D acini were examined by phase contrast (top) or stained with the nuclei fluorescence dye DAPI (bottom). Note differences in acini size and shape. Arrows indicate examples of invasive structures observed in EOC 3D cultures. (B) Quantification of acini size formed by the indicated EOC cells treated with GSK343 (1 μM) or vehicle control (0.1% DMSO) after 12 d of growth in Matrigel. Mean of three independent experiments with SD, *p < 0.05. (C) Same as (A) but stained for Ki-67 expression Bar = 40 μm. (D) Quantification of (C). Ki-67 positive staining cells in the indicated EOC cells treated with GSK343 (1 μM) or vehicle control (0.1% DMSO) after 12 d of growth in Matrigel. Numbers represent mean number of cells counted in 5 acini for each condition. Mean of three independent experiments with SD, *p < 0.05.

Figure 3. The EZH2 inhibitor GSK343 suppresses invasion of human EOC cells. (A) Equal number of SKOV3 cells treated with 1 μM GSK343 or vehicle control (0.1% DMSO) were assayed for migration through uncoated control membrane or invasion through Matrigel-coated membrane. The cells migrated through control membrane or invaded through Matrigel-coated membrane were stained with 1% crystal violet in PBS. (B) Quantification of (A). Number of SKOV3 cells migrated through control membrane or invaded through Matrigel-coated membrane. Numbers represent mean of three independent experiments. *p < 0.05.

Figure 4. The EZH2 inhibitor GSK343 induces apoptosis and decreases the G₂/M phase population of EOC cells cultured in 3D conditions. (A) Representative cell-cycle distribution as determined by FACS analysis for SKOV3 cells treated with 1 μM GSK343 (red) or 0.1% DMSO vehicle control (blue) after 4 d of growth on tissue culture plastic (left) or in Matrigel (right). (B) Quantitation of (A). The percentage of sub-G₁, G₁, S and G₂/M of SKOV3 cells treated with 1 μM GSK343 (black bars) or 0.1% DMSO vehicle control (white bars) after 4 d of growth in 2D monolayer culture. Mean of three independent experiments with SD (C) Same as (B) but for cells cultured in 3D in Matrigel. Mean of three independent experiments with *p < 0.05. (D) Same as (B and C) but stained for Annexin V, a marker of apoptosis. Percentage of Annexin V positive cells was indicated. (E) Quantification of (D). Mean of three independent experiments with SD, *p < 0.05. (F and G) Same as (B and C) but examined for relative HRK expression in the indicated EOC cells treated with 1 μM GSK343 (black bars) or 0.1% DMSO vehicle control (white bars) after 4 d of growth in 3D in Matrigel as determined by q-RT PCR. Numbers represent expression relative to β-2-microglobulin (B2M) expression. Mean of three independent experiments with SD, *p < 0.05.