Potential therapeutic effect of targeting glycogen synthase kinase 3β in esophageal squamous cell carcinoma

Abstract

Esophageal squamous cell carcinoma (ESCC) is a common gastrointestinal cancer and is often refractory to current therapies. Development of efficient therapeutic strategies against ESCC presents a major challenge. Glycogen synthase kinase (GSK)3β has emerged as a multipotent therapeutic target in various diseases including cancer. Here we investigated the biology and pathological role of GSK3β in ESCC and explored the therapeutic effects of its inhibition. The expression of GSK3β and tyrosine (Y)216 phosphorylation-dependent activity was higher in human ESCC cell lines and primary tumors than untransformed esophageal squamous TYNEK-3 cells from an ESCC patient and tumor-adjacent normal esophageal mucosa. GSK3β-specific inhibitors and small interfering (si)RNA-mediated knockdown of GSK3β attenuated tumor cell survival and proliferation, while inducing apoptosis in ESCC cells and their xenograft tumors in mice. GSK3β inhibition spared TYNEK-3 cells and the vital organs of mice. The therapeutic effect of GSK3β inhibition in tumor cells was associated with G0/G1- and G2/M-phase cell cycle arrest, decreased expression of cyclin D1 and cyclin-dependent kinase (CDK)4 and increased expression of cyclin B1. These results suggest the tumor-promoting role of GSK3β is via cyclin D1/CDK4-mediated cell cycle progression. Consequently, our study provides a biological rationale for GSK3β as a potential therapeutic target in ESCC.

Figure 1

Comparative analysis for the expression and phosphorylation of GSK3β in human ESCC cells (TE-1, TE-5, TE-8, TE-9, TE-10, TE-15, KES), normal esophageal squamous epithelial cells (TYNEK-3), and normal squamous mucosa and primary tumors from ESCC patients. (A) Expression of GSK3β and of its phosphorylated forms (pGSK3β^S9, inactive form; pGSK3β^Y216, active form) were examined by Western blotting. β-actin expression was monitored as a loading control in each sample. (B) Representative findings for the expression of GSK3β and its Y216 phosphorylated fraction (pGSK3β^Y216) in the primary tumor and corresponding normal squamous mucosa of ESCC patients. The scale bar indicates 100 μm in length. Immunohistochemical images were captured using Keyence BZ-X700 Analyzer (Version 1.3). The two right hand graphs generated using GraphPad Prism 5.0 (GraphPad Software, Inc. CA) show statistical comparison of the immunohistochemistry (IHC) scores for GSK3β and pGSK3β^Y216 between the primary tumor (T) and normal mucosa (N) of ESCC patients. A horizontal bar in each group shows the mean value of IHC scores. (C) Expression of GSK3β mRNA in normal esophageal tissues (N) and primary ESCC tumor tissues (T) based on the TCGA database. The data was generated using the analysis tool UALCAN (https://ualcan.path.uab.edu/). n, number of patients; **P < 0.01. Full-length blots for (A) are shown in Supplementary Information, Fig. S9.

Figure 2

Effects of GSK3β inhibition on cell survival, proliferation and apoptosis in ESCC (TE-5, TE-8, TE-10) and normal esophageal squamous TYNEK-3 cells. (A) The respective ESCC cells and TYNEK-3 cells were treated with DMSO or the indicated concentration of AR-A014418 or SB-216763 for the designated times. The relative number of viable cells at each time point was examined by WST-8 assay. Mean values with standard deviations of triplicate experiments were compared between cells treated with DMSO and the indicated GSK3β inhibitor at different concentrations. *P < 0.05, **P < 0.01. (B) The incidence of EdU-positive proliferating cells was compared between ESCC cells treated with DMSO (open column), 25 μmol/L AR-A014418 (gray column) and SB-216763 (closed column) for 24 h (left panel), and between the cells transfected with 20 nmol/L control (open column) and GSK3β-specific siRNA (dotted column) for 72 h (right panel). The mean percentages of EdU-positive proliferating cells in 5 fluorescence microscopy fields (shown in Supplementary Information, Fig. S5) were calculated with SDs and statistically compared. (C) The mean relative number ± SDs of apoptotic cells measured by DNA fragmentation assay in triplicate was compared between ESCC cells with the same treatment as shown in B. (B,C) * P < 0.05; ** P < 0.01. (D) Western blotting analysis for the amount of PARP and c-PARP in ESCC cells treated with DMSO or the indicated concentrations of AR-A014418 for 24 h. The amount of each protein sample was monitored by the expression of β-actin. Full-length blots for (D) are shown in Supplementary Information, Fig. S9.

Figure 3

Effects of GSK3β inhibition on the cell cycle profiles and expression of cell cycle-regulating molecules in ESCC cells. (A) Representative flow cytometry findings of cell cycle profile of TE-8 cells treated with DMSO, 25 μmol/L AR-A014418 or SB-216763 for 48 h. The data were generated using a FACS Canto II (BD Biosciences). (B) Comparison of DNA histograms for each cell cycle fraction of TE-8 cells treated with DMSO (open column), 25 μmol/L AR-A014418 (gray column) or SB-216763 (closed column), or cells transfected for 72 h with non-specific (open column) or GSK3β-specific siRNA (dotted column), respectively. Cellular DNA content was analyzed using FACSDiva software (Version 8.0, BD Bioscience). Data are the mean percentages of cell populations in the respective cell cycle phases with SDs in five separate tests. *P < 0.05; **P < 0.01. (C) Western blotting analysis for expression of GSK3β, cyclin D1, CDK4 and cyclin B1 and GSK3β Y216 phosphorylation (pGSK3β^Y216) in ESCC cells treated with the indicated concentrations of AR-A014418 for 48 h. (D) Western blotting analysis for expression of GSK3α/β, cyclin D1, CDK4 and cyclin B1 in ESCC cells transfected with non-specific (N) or GSK3β-specific (S) siRNA, respectively, for 72 h. (C, D) The amount of each protein sample was monitored by the expression of β-actin. Full-length blots for (C, D) are shown in Supplementary Information, Fig. S9.

Figure 4

Effect of GSK3β inhibitors on proliferation of ESCC cell xenograft tumors in mice. (A) Time course of the effects of DMSO (open circle), AR-A014418 (2 mg/kg body weight, open square; 5 mg/kg body weight, closed square), and LY2090314 (1 mg/kg body weight, open triangle; 2.5 mg/kg body weight, closed triangle) on tumor size of TE-8 cell xenografts in mice. (B) Gross appearance of xenograft tumors removed at autopsy from mice after 5 weeks of treatment with different doses of AR-A014418 (AR) or LY2090314 (LY). The left lower insets show the tumors removed from two mice 4 weeks after treatment with DMSO following animal experimentation ethics guidelines as described in the “Results”. **P < 0.01.

Figure 5

Representative histological, immunohistochemical and histochemical findings of xenograft tumors from mice treated with DMSO, 5 mg/kg body weight AR-A014418 (AR) or 2.5 mg/kg body weight LY2090314 (LY). Serial paraffin sections of the respective tumors were stained with hematoxylin and eosin (HE), immunostained for GSK3β, pGSK3β^Y216, GS, pGS^S641, cyclin D1 and Ki-67, and histochemically stained by the TUNEL method. A scale bar in each panel indicates 100 μm. Histological, immunohistochemical and histochemical images were captured using Keyence BZ-X700 Analyzer (Version 1.3). The graphs on the right side show statistical comparison of the mean percentages with SDs of tumor cells positive for the corresponding molecules, and TUNEL results for xenograft tumors from mice treated with DMSO, AR or LY. *P < 0.05; **P < 0.01.