Enhancing Transcriptional Reprogramming of Mesenchymal Glioblastoma with Grainyhead-like 2 and HDAC Inhibitors Leads to Apoptosis and Cell-Cycle Dysregulation

Abstract

Glioblastoma (GBM) tumor cells exhibit mesenchymal properties which are thought to play significant roles in therapeutic resistance and tumor recurrence. An important question is whether impairment of the mesenchymal state of GBM can sensitize these tumors to therapeutic intervention. HDAC inhibitors (HDACi) are being tested in GBM for their ability promote mesenchymal-to-epithelial transcriptional (MET) reprogramming, and for their cancer-specific ability to dysregulate the cell cycle and induce apoptosis. We set out to enhance the transcriptional reprogramming and apoptotic effects of HDACi in GBM by introducing an epithelial transcription factor, Grainyhead-like 2 (GRHL2), to specifically counter the mesenchymal state. GRHL2 significantly enhanced HDACi-mediated MET reprogramming. Surprisingly, we found that inducing GRHL2 in glioma stem cells (GSCs) altered cell-cycle drivers and promoted aneuploidy. Mass spectrometry analysis of GRHL2 interacting proteins revealed association with several key mitotic factors, suggesting their exogenous expression disrupted the established mitotic program in GBM. Associated with this cell-cycle dysregulation, the combination of GRHL2 and HDACi induced elevated levels of apoptosis. The key implication of our study is that although genetic strategies to repress the mesenchymal properties of glioblastoma may be effective, biological interactions of epithelial factors in mesenchymal cancer cells may dysregulate normal homeostatic cellular mechanisms.

Keywords: EMT; GRHL2; HDAC; aneuploidy; apoptosis; cell cycle; glioblastoma; mesenchymal.

Figure 1

GRHL2 represses mesenchymal proteins in GBM cells. Western blots for GRHL2, Slug, CD44, MMP2, and ZEB1 in LN229 cells with or without doxycycline (200 ng/mL) or 2.5 µM vorinostat (VOR). β-Actin was used as a loading control for protein expression. Graphs depict significant changes in protein expression levels from A. Graphs depict means +/− SEM for n = 3. * p < 0.05 t-test; **p < 0.01 t-test.

Figure 2

GRHL2 enhances apoptosis of HDACi-treated GBM cells. (A) Western blots for CC3, GRHL2, histone 3 lysine 9 acetylation (H3k9ac), and pH2AX in inducible LN229 cells treated with the individual HDACi vorinostat (1 μM) or mocetinostat (1 μM). β-Actin was used as a loading control for protein expression. (B) Quantification of LN229’s CC3 and pH2AX in A. (C) GRHL2 induction in LN229 cells after 72 h of doxycycline induction, treated for 24 h with mocetinostat (1 μM) followed by immunoblot analysis of CC3, pH2AX, and acetylation of histone 3 lysine 9. Data from blots 1 and 2 are the same samples on different membranes. (D) CC3 levels in GRHL2-inducible glioma stem cells from GBM PDX after doxycycline with or without vorinostat (2.5 μM). (E) Apoptosis as measured by annexin V in doxycycline-induced LN229 cells treated with 2.5 μM vorinostat for 56 h. Graphs depict means + SEM for n = 3. * p < 0.05 t-test.

Figure 3

Exogenous GRHL2 expression alters the GBM cell cycle. (A) Inducible LN229 cells were treated with dox for 72 h with or without HDACi for the last 24 h and were immunoblotted for indicators of cell-cycle phases pCDC2, pRB, cyclin B1, and p21. GAPDH was used as a loading control for protein expression. (B) Quantification of data in A. (C) Cell-cycle (propidium iodide) flow cytometry of parental or inducible LN229 treated with dox for 72 h. Graph displays mean values for G1, S, and G2/M phase. (D) Cell-cycle (propidium iodide) flow cytometry of iLN229 treated with dox for 3 or 14 days. Arrow indicates > 4 N cells. (E) Control or dox-treated (21 days) iLN229 cells were stained with CD44 to visualize the cell membrane and DAPI to visualize nuclei. Cells were then counted for multinucleated cells. White arrowheads identify multinucleated cells. Mean values of multinucleated cells are per 300 cells counted. Graphs depict means +/− SEM for n = 3. * p < 0.05 t-test; ** p < 0.01 t-test; *** p < 0.001 t-test.

Figure 4

GRHL2 expression dysregulates cell cycle in glioma stem cells and associates with aneuploidy in patient tumors. (A) Inducible glioma stem cells (GSCs) were treated with dox for 72 h with or without HDACi for the last 24 and were immunoblotted for indicators of cell-cycle phases cyclin B1, and p21. β-Actin was used as a loading control for protein expression. (B) Cell-cycle (propidium iodide) flow cytometry of inducible GSCs treated with dox for 72 h. * p < 0.05, t-test. Arrow indicates > 4 N cells. (C) GRHL2 mRNA expression in patient tumors by cancer type, arranged by mean expression level and color-coded for aneuploidy score. (D) GRHL2 mRNA expression in cancers vs. aneuploidy score. (n = 3077 scores < 6, n = 6579 > 6.) Data in C and D were captured from The Cancer Genome Atlas (TCGA) cBioBortal. **** p < 0.0001 t-test.

Figure 5

GRHL2 expression in glioblastoma cells. GRHL2 suppresses mesenchymal factor expression. HDACi dysregulation of the cell cycle is further promoted by GRHL2, increasing the chance that cells will undergo apoptosis.