Induction of MAPK- and ROS-dependent autophagy and apoptosis in gastric carcinoma by combination of romidepsin and bortezomib

Abstract

Proteasome inhibitors and histone deacetylase (HDAC) inhibitors can synergistically induce apoptotic cell death in certain cancer cell types but their combinatorial effect on the induction of autophagy remains unknown. Here, we investigated the combinatorial effects of a proteasome inhibitor, bortezomib, and an HDAC inhibitor, romidepsin, on the induction of apoptotic and autophagic cell death in gastric carcinoma (GC) cells. Isobologram analysis showed that low nanomolar concentrations of bortezomib/romidepsin could synergistically induce killing of GC cells. The synergistic killing was due to the summative effect of caspase-dependent intrinsic apoptosis and caspase-independent autophagy. The autophagic cell death was dependent on the activation of MAPK family members (ERK1/2 and JNK), and generation of reactive oxygen species (ROS), but was independent of Epstein-Barr virus infection. In vivo, bortezomib/romidepsin also significantly induced apoptosis and autophagy in GC xenografts in nude mice. This is the first report demonstrating the potent effect of combination of HDAC and proteasome inhibitors on the induction of MAPK- and ROS-dependent autophagy in addition to caspase-dependent apoptosis in a cancer type.

Keywords: apoptosis; autophagy; bortezomib; gastric carcinoma; romidepsin.

Figure 1. Effects of combination of bortezomib/romidepsin on cell proliferation of GC cells

AGS-BDneo and SNU-719 cells were treated with various combinations of bortezomib/romidepsin for 48 hr. A & D. Data are presented as percentages of cell proliferation as determined by MTT assays. B & E. Synergisms of proliferation inhibition of the two cell lines were analyzed by isobologram analysis. C & F. Percentages of cell proliferation of GC cells upon treatment with combination of 7.5 nM bortezomib and 2.5 nM romidepsin were compared to those treated with either drug alone and combination indexes (CI) were calculated. Percentages of proliferating cells treated with bortezomib/romidepsin were compared with those untreated or treated with either drug alone using One-way ANOVA Dunnett's Multiple Comparison Test. P value < 0.05 was considered statistically significant (***p < 0.001). Error bars represent the standard error of mean (SEM) of data obtained in at least three independent experiments.

Figure 2. Effects of bortezomib/romidepsin on apoptosis of GC cells

A. AGS-BDneo cells were treated with combination of 7.5 nM bortezomib and either 1.25 or 2.5 nM romidepsin or either drug alone for 24 hr and 48 hr, respectively. The treated cells were assayed for apoptosis by annexin V/propidium iodide (AV/PI) staining. B. Percentages of apoptotic cells treated with bortezomib/romidepsin were compared with those untreated or treated with either drug alone using One-way ANOVA Dunnett's Multiple Comparison Test. P value < 0.05 was considered statistically significant (***p < 0.001). C. AGS-BDneo cells were treated with combination of 7.5 nM bortezomib and 2.5 nM romidepsin or either drug alone for 12, 24 and 48 hr followed by detection of expression of PARP, cleaved PARP and cleaved caspase-3, -8 and -9 by western blot analysis. α-tubulin served as loading control. D. AGS-BDneo cells were treated with combination of 7.5 nM bortezomib and 2.5 nM romidepsin or either drug alone for 24 hr. Expression of cleaved PARP (red signals) was detected by immunofluorescent staining. DAPI (blue signals) stained the cell nuclei. Scale bar, 100 μm.

Figure 3. Roles of caspase activation and reactive oxygen species (ROS) generation in the apoptosis of GC cells induced by bortezomib/romidepsin

A. AGS-BDneo and SNU-719 cells were pre-treated with either 50 μM Z-VAD-FMK or 12 mM N-acetyl-cystein (NAC) for 1 hr followed by treatment with combination of 7.5 nM bortezomib and 2.5 nM romidepsin or either drug alone for 24 hr. The treated cells were analyzed for the expression of PARP, cleaved PARP, cleaved caspase-3, -8 and -9 and acetylated histone H4 by western blot analysis. α-tubulin served as loading control. B. Decreases in mitochondrial membrane potential in GC cells upon treatment with bortezomib/romidepsin for 48 hr were analyzed by JC-1 assay. C. Generation of ROS was detected by DCFH-DA staining.

Figure 4. Percentages of caspase-dependent cell death induced by bortezomib/romidepsin in GC cells

AGS-BDneo and SNU-719 cells were pre-treated with either 50 μM Z-VAD-FMK, 12 mM NAC or 50 μM necrostatin-1 for 1 hr followed by treatment with combination of 7.5 nM bortezomib and 2.5 nM romidepsin or either drug alone for 72 hr. Percentages of active caspase-3-positive/death cells was analyzed by Z-DEVD-FMK/PI staining.

Figure 5. Effects of bortezomib/romidepsin on induction of autophagy in GC cells

A. AGS-BDneo and SNU-719 cells pre-treated with either 50 μM Z-VAD-FMK or 12 mM NAC for 1 hr and then treated with bortezomib/romidepsin for 24 hr were analyzed for the expression of PARP, cleaved PARP, LC3-I/II, Beclin-1 and Atg-12. α-tubulin served as loading control. B. AGS-BDneo or C. SNU-719 cells were pre-treated with 10 μM chloroquine (CQ) for 1 hr followed by treatment with bortezomib/romidepsin (BR) for 24 hr. Expression of p62/SQSTM1 aggregates (green signal) was visualized with confocal microscopy. DAPI (blue signals) stained the cell nuclei. Scale bar, 20 μm. D. GC cells were pre-treated with 5 mM 3-MA for 1 hr followed by treatment with bortezomib/romidepsin (combination) for 72 hr. Percentages of death cells were detected by PI staining.

Figure 6. The involvement of MAPK pathways in the induction of apoptosis and autophagy by bortezomib/romidepsin

A. AGS-BDneo and SNU-719 cells were treated with combination of 7.5 nM bortezomib and 2.5 nM romidepsin for 0, 2, 4, 8, 12 & 24 hr. The treated cells were analyzed for the expression of p-ERK1/2, p-JNK, p-p38, Beclin-1, LC3-I/II, PARP and cleaved PARP by western blot analysis. β-actin served as loading control. B. AGS-BDneo and SNU-719 cells were pre-treated with either 50 μM PD98059 (MEK inhibitor), 50 μM SP600125 (JNK inhibitor) and 20 μM SB202190 (p38 MAPK inhibitor) for 1 hr followed by treatment with combination of 7.5 nM bortezomib and 2.5 nM romidepsin or either drug alone for 24 hr. The treated cells were analyzed for the expression of p-ERK1/2, p-c-Jun, LC3-I/II and cleaved caspase-3 by western blot analysis. C. AGS-BDneo and SNU-719 cells were pre-treated with 12 mM NAC for 1 hr followed by treatment with combination of 7.5 nM bortezomib and 2.5 nM romidepsin or either drug alone for 8 hr. The expression of p-ERK1/2 and p-c-Jun was analyzed by western blot analysis. α-tubulin served as loading control.

Figure 7. Effects of bortezomib/romidepsin on tumour growth suppression of GC xenografts in nude mice

SNU-719 cells were subcutaneously injected into the right flanks of nude mice. When the tumours became palpable, the mice were either treated with 60 μg/kg bortezomib (day1-5 per week), 375 μg/kg romidepsin (day 1&4 per week) or their combination for 4 weeks by intraperitoneal injection. A. The size of tumours during the period of experiment was measured twice weekly using a caliper. Data are presented as the mean tumour volumes of mice in both treatment and control groups on the days post-treatment. B. The weight of nude mice was measured every week. C. Picture showing the size of tumours at the end of the experiments D. Average tumour masses of mice of control and treatment groups were shown. The tumour volumes and masses of mice treated with bortezomib/romidepsin were compared with those treated with vehicle control (DMSO) or either drug alone using One-way ANOVA Dunnett's Multiple Comparison Test. P value < 0.05 was considered statistically significant (***p < 0.001 and **p < 0.01). Error bars represent the standard error of mean (SEM) of tumour masses. E. Protein samples were extracted from the tumours and analyzed for the expression of PARP, cleaved PARP, cleaved caspase-3, LC3-I/II, p-c-Jun and p-ERK1/2 by western blotting. α-tubulin served as loading control.