Celecoxib induces apoptosis through Akt inhibition in 5-fluorouracil-resistant gastric cancer cells

Abstract

Gastric cancer is the fifth leading cause of cancer and a global public health problem. 5-Fluorouracil (5-FU) is the primary drug chosen for the treatment of advanced gastric cancer, but acquired cancer drug resistance limits its effectiveness and clinical use. Proliferation assays showed that a gastric carcinoma cell line, AGS and 5-FU-resistant AGS cells (AGS FR) treated with 3–100 μM 5-FU for 48 h or 72 h showed different sensitivities to 5-FU. Immunoblot assay demonstrated that AGS FR cells expressed more COX-2 and PGE₂-cognated receptor EP2 than AGS cells. AGS FR cells considerably produced PGE₂ than AGS upon stimulation with 5-FU. These results suggest that COX-2 expression is associated with 5-FU resistance. Unlike AGS FR cells, AGS cells showed increased levels of both cleaved caspase-3 and Bax following 5-FU treatment. Treatment of cells with the COX-2 selective inhibitor celecoxib induced cell death of AGS FR cells in a time- and concentration-dependent manner. FACS analysis showed that celecoxib at high doses caused apoptotic cell death, demonstrating a concentration-dependent increase in the cell populations undergoing early apoptosis and late apoptosis. This apoptotic induction was strongly supported by the expression profiles of apoptosis- and survival-associated proteins in response to celecoxib; pro-apoptotic cellular proteins increased while expressions of COX-2 and p-Akt were downregulated in a concentration-dependent manner. An increase in PTEN expression was accompanied with downregulation of p-Akt. Based on the data that downregulation of COX-2 was correlated with the concentrations of celecoxib, COX-2 may play a key role in celecoxib-induced cell death of AGS FR cells. Butaprost, the EP2 agonist, promoted proliferative activity of AGS FR cells in a concentration-dependent manner compared with AGS cells. In cells exposed to butaprost, expressions of COX-2 and p-Akt were increased in a concentration-dependent manner with concomitantly reduced PTEN levels. Taken together, 5-FU-resistance in gastric cancer is correlated with COX-2 expression, and therefore the selective inhibition of COX-2 leads to suppression of cell proliferation of AGS FR cells. Modulation of COX-2 expression and its catalytic activity may be a potential therapeutic strategy to overcome 5-FU-resistant gastric cancer.

Introduction

Gastric cancer develops from the lining of stomach and may spread to other tissues such as liver, lungs, bones and lymph nodes [1]. The most common cause of gastric cancer is Helicobacter pylori, accounting for more than 60% of cases; other risk factors include smoking, dietary factors, and genetic traits [2]. Gastric cancer is the fifth leading cause of cancer worldwide and a global health problem. Chemotherapy, radiotherapy, and surgery have been applied to treat gastric cancer. Several drugs are used in chemotherapy such as 5-fluorouracil (5-FU), doxorubicin, and cisplatin; however, gastric cancer is insensitive to these drugs and may acquire drug resistance. In this study, preliminary experiments were done to evaluate the cell proliferation of gastric cancer cell lines AGS and AGS FR, a 5-FU-resistant cell line. AGS FR cells were more resistant to 5-FU than AGS cells when viability was assessed with 5-FU treatment ranging from 1 to 100 μM. Clinically, either 5-FU monotherapy or combined therapy with other conventional drugs is a standard regimen for gastric cancer, but gastric cancer often relapses after initial response to 5-FU, leading to residual cells with cancer stem cell properties [3]. To delineate the molecular mechanism by which AGS FR cells acquire 5-FU resistance, in this study, both gastric cancer cell lines were examined for the expressions of COX-2 and a PGE₂ cognate receptor EP2 as well as apoptosis-associated proteins like caspase-3 and Bax. Chemoresistance is caused from a variety of factors such as mutations in drug targets, modulation of cell survival and death signaling pathways, and diminished intracellular concentration of drugs [4]. Various studies have explored the underlying molecular mechanism by which gastric cancers become resistant to 5-FU. For instance, thymidine kinase overexpression is suggested to play a major role in 5-FU resistance [5]. RhoGD12 is also proposed to be a plausible candidate to be associated with 5-FU resistance and may be a potential therapeutic target for sensitizing 5-FU-resistant cells [6]. A recent study found that zeste homologue 2 (EZH2) contributes to 5-FU resistance in gastric cancer through suppressing FBXO32 expression [7]. Enhancer of EZH2 is a histone methyltransferase that is aberrantly increased in tumor progression and associated with poor prognosis [8]. Thus, pharmacological inhibition or silencing of EZH2 may hold promise in the treatment of drug-resistant cancers.

COX-2 is overexpressed in solid tumors such as colorectal and liver cancer as well as gastric cancer [9, 10]. COX-2 enzymatically transforms arachidonic acid to produce prostaglandin E₂ (PGE₂), which promotes carcinogen-induced colon tumors in rats [11]. These results have suggested that the COX-2/PGE₂ pathway plays a key role in regulating tumorigenesis. Overexpression of COX-2 contributes to gastric carcinogenesis by enhancing cell proliferation, inhibiting cell apoptosis and elevating metastasis as well as immunosuppression [10]. The COX-2 inhibitor NS-398 induces apoptosis and suppresses the proliferation of gastric cell lines. In vivo administration of nimesulide attenuates chemical-induced gastric tumorigenesis [12]. Notably, the chemopreventive effect of celecoxib, a selective COX-2 inhibitor, was demonstrated in gastric cancer patients and celecoxib is suggested to reverse multidrug resistance in gastric cancer [13, 14].

In this present study, the viability of 5-FU-resistant gastric cancer cells was investigated in the presence or absence of celecoxib. The proliferative activity of 5-FU-resistant cells was reduced by celecoxib treatment in a time- and dose-dependent manner. Moreover, we examined the mechanism by which 5-FU resistant cells are responsive to celecoxib by analyzing expression patterns of death or survival-related cellular proteins. We speculated that production of the COX-2 enzymatic product PGE₂ may play a role in the proliferation of AGS FR cells. We found that the structural analog of PGE₂ butaprost enhanced the proliferation of AGS FR cells more than that of AGS cells. In summary, COX-2 in 5-FU-resistant gastric cancer cell produces PGE₂, which in turn can not only upregulate survival-associated proteins but also downregulate apoptosis-regulated proteins. Therefore, the COX-2/PGE₂ pathway is critical in conferring drug resistance to gastric cancer, and selective inhibition of COX-2 may be promising approach to overcoming 5-FU-resistance in cancer cells.

Materials and methods

Materials

Butaprost, celecoxib and antibodies against COX-2 and EP2 were purchased from Cayman Chemical Co. (Ann Arbor, MI, USA). Antibodies against cleaved caspase-3 and PARP, Bax, p-Akt, and PTEN were obtained from Cell Signaling Technology Inc. (Beverly, MA, USA). A horseradish peroxidase-conjugated secondary antibody was purchased from SantaCruz Biotechnology (Santa Cruz, CA, USA). We purchased 5-FU and β-actin antibody from Sigma-Aldrich (St. Louis, MO, USA).

Establishment of 5-FU resistant AGS cell lines

AGS cells were obtained from American Type Culture Collection (ATCC) and maintained in RPMI 1640 medium supplemented with 10% fetal bovine serum and antibiotics (100 U/ml penicillin G and 100 μg/ml streptomycin) at 37 °C in a humidified incubator containing 5% CO₂ and 95% air.

To establish AGS FR cells, AGS cells were cultured with gradually increasing concentrations (from 5 to 100 μM) of 5-FU. To maintain resistance, AGS FR cells were kept in media containing 100 μM 5-FU. A stock solution of 5-FU was prepared in dimethylsulfoxide (DMSO), aliquoted, and stored at 4 °C.

Cytotoxicity assay

The anti-proliferative effect of celecoxib against AGS FR cells was measured using a solution of tetrazolium compound [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt (MTS) (Promega, Madison, WI, USA). Briefly, cells (2 × 10³) were incubated in triplicate in a 96-well plate in the presence or absence of chemical in a final volume of 0.1 ml at 37 °C for different time intervals. Next, 20 μl of MTS solution was added to each well and samples were incubated for 60 min. The number of viable cells was measured in a 96-well plate at an optical density of 492 nm on a microplate reader (Tecan Trading AG, Switzerland). Cell viability was described as the relative percentage of control.

Annexin V staining

Annexin V staining was performed using the FITC-Annexin V staining kit (BD-Biosciences, San Jose, CA, USA) following the manufacturer’s instructions. Briefly, celecoxib-treated cells were washed with PBS and resuspended in binding buffer containing Annexin V and propidium iodide (PI). Fluorescence intensity was measured using flow cytometry (BD Biosciences, La Jolla, CA, USA).

Western blot analysis

After treatment, cells were lysed in RIPA buffer and the total protein concentrations of cell lysates were quantified using a bicinchoninic acid (BCA) protein assay kit (Pierce Biotechnology, Rockford, IL, USA) and stored at − 20 °C till further use. Protein samples (30 μg) were separated by SDS-PAGE using a 12% (w/v) polyacrylamide resolving gel and transferred to a PVDF membrane. The membrane was blocked using 5% (w/v) skim milk in Tris-buffered saline with Tween-20 buffer (TBS-T) for 1 h at room temperature (RT). The membranes were subsequently incubated with primary antibodies at 4 °C overnight, followed by incubation with the respective secondary antibody for 1 h at RT. The protein-antibody complexes were visualized using SuperSignal Pico Chemiluminescent substrate or dura-luminol substrate (Thermo Fischer Scientific, Waltham, MA, USA) according to the manufacturer’s instruction and visualized with an Imagequant™ LAS 4000 (Fujifilm Life Science, Japan).

PGE₂ assay

AGS and AGS FR cell lines were incubated in 6-well plates at 5 × 10⁴ cells/well and treated with 5-FU at a concentration of 80 μM for 48 h. After the treatment, the cell culture supernatants were collected to monitor the levels of PGE₂. PGE₂ concentration was determined using PGE₂ ELISA Kit (Enzo Life Sciences, Inc., Farmingdale, NY, USA) according to the manufacturer’s protocol. Absorbance was measured at 405 nm using a microplate reader.

Statistical analysis

When necessary, data were expressed as the mean ± SD of at least three independent experiments, and statistical analysis for a single comparison was performed using the Student’s t-test. A p value less than 0.05 was considered as statistically significant.

Results

Cytotoxicity of cells to 5-FU depends on COX-2

Gastric cell lines AGS and AGS FR were exposed to increasing doses of 5-FU for 48 h or 72 h. The cytotoxicity of 5-FU was concentration-dependent in AGS cells, with nearly 40% cell viability at a concentration of 100 μM (Fig. 1A, B). The viability results demonstrated that AGS FR cells were more resistant to 5-FU than AGS cells at low concentrations. AGS FR cells did not show apparent cytotoxic damage at 3 and 10 μM of 5-FU and exhibited less cytotoxicity than AGS cells at the high concentrations of 10 μM or more.

Fig. 1.

Cytotoxic response of AGS and AGS FR cells to 5-FU. AGS and AGS FR cell lines were treated with various doses of 5-FU for 48 (a) or 72 h (b). Cell viability was determined by the MTS assay. Values are expressed as the mean ± SD. *p < 0.05, **p < 0.001, compared with the control

Changes in expression of COX-2, cell death-related signaling molecules and PGE₂ production in AGS and AGS FR cells treated with 5-FU

Both AGS and AGS FR cells were treated with vehicle or 5-FU for 48 h and then protein extracts from cells were prepared for western blot analysis. Regardless of 5-FU treatment, the levels of COX-2, EP2, and p-Akt were higher in AGS FR cells than in AGS cells (Fig. 2a). Treatment with 5-FU enhanced the expressions of cleaved-caspase 3 and Bax in AGS cells, while these proteins were not changed in AGS FR cells compared with the vehicle group. From the differential expression of COX-2 between two cell lines, the levels of PGE₂ were monitored when cells were subjected to 5-FU treatment. The amounts of PGE₂ exists in both cell lines were approximately 100 pg/ml under a non-stimulated condition (data not shown). When stimulated with 5-FU, on the contrary, PGE₂ production from AGS FR cells was considerably increased as compared with that from AGS (Fig. 2b).

Fig. 2.

Changes in levels of COX-2, EP2, PGE₂ and apoptosis-related proteins in AGS and AGS FR cells treated with 5-FU. a Cells were subjected to 5-FU at a concentration of 80 µM or vehicle treatment for 48 h. Cell lysates were examined by western blot analysis using antibodies to COX-2, EP2, p-Akt, cleaved-caspase 3 and Bax. b PGE₂ concentrations presents in cell culture supernatants of AGS or AGS FR treated with 5-FU. AGS and AGS FR cell lines were stimulated with 5-FU (80 μM). The culture media were collected at 48 h after treatment to measure the PGE₂ concentrations using an enzyme immunoassay

Cytotoxic potency of celecoxib in AGS FR cells and analysis of cell death modality

Since AGS FR cells highly express COX-2, we investigated whether their resistance to 5-FU could be affected by inhibition of COX-2 enzyme activity. Celecoxib itself was concentration-dependently cytotoxic in AGS FR for 24 h, 48 h or 72 h treatment (Fig. 3a). AGS FR cells treated with celecoxib (10, 40 or 80 μM) for 48 h were analyzed by flow cytometry using double staining with Annexin V and PI to quantify the population of cells undergoing apoptosis. Treatment of cells with celecoxib markedly increased the percentages of apoptotic cells compared with untreated control cells (Fig. 3b). Quantification of apoptotic cells and statistical analysis of celecoxib-induced apoptosis are presented in Fig. 3c.

Fig. 3.

Cell death profiles of AGS FR cells treated with celecoxib. Cells were treated with celecoxib as a function of dose and time. a Cytotoxicity was determined by relative cell viability of celecoxib-treated group to vehicle-treated group. *p < 0.05, **p < 0.001, compared with control. b The apoptotic index (%) was determined by flow cytometry in cells treated with celecoxib (10, 40 or 80 μM) for 48 h and Annexin V and PI staining. c Graph shows the statistical analysis of apoptotic cells. Data are representative of three independent experiments. *p < 0.05, **p < 0.001, compared with the control

Expression profiles of apoptosis- and survival-related proteins

In AGS FR cells treated with celecoxib at different concentrations for 48 h, cells expressed cleaved-PARP, cleaved-caspase 3 and Bax, though the individual expressions varied depending on the concentration (Fig. 4a). In contrast, the levels of COX-2 and p-Akt were suppressed and PTEN expression was increased in celecoxib-treated AGS FR cells (Fig. 4b).

Fig. 4.

Effects of celecoxib on the changes in expression of apoptosis- or survival-related signaling proteins. Cells were treated with the indicated concentrations of celecoxib for 48 h and then examined by western blot analysis using antibodies against apoptosis-related proteins cleaved-PARP, cleaved-caspase-3, and Bax (a) or survival-related proteins COX-2, p-Akt, and PTEN (b)

Effect of butaprost on cell proliferation and gene expression

Studies performed with EP receptor-specific agonists and antagonists indicate important roles for EP2 in the proliferation of several tumor types [15, 16]. To examine the role of the EP2 receptor in COX-2-mediated proliferation of AGS FR cells, we tested whether butaprost, a selective EP2 prostanoid receptor agonist, could affect the proliferation of AGS and AGS FR cells through activation of EP2 signaling. Butaprost had a more potent proliferative effect in AGS FR cells than AGS cells (Fig. 5a). In line with the proliferation data, expressions of both COX-2 and p-Akt were proportional to the concentration of butaprost while that of PTEN decreased (Fig. 5b).

Fig. 5.

Cytotoxic and signaling response of 5-FU sensitive or -resistant AGS cell lines treated with butaprost. Cells were exposed to butaprost at the indicated doses for 48 h. a Cell viability was determined by the MTS assay. Values are expressed as the mean ± SD. **p < 0.001, compared with control. b To examine changes in signaling molecules, cells were treated with increasing concentrations of butaprost for 48 h and examined by western blot using the indicated antibodies

Discussion

Gastric cancer is mostly diagnosed at the late stage. Furthermore, prognosis of gastric cancer is so poor that the median overall survival is not more than 1 year [3]. The current standard chemotherapy regimen to treat gastric cancer is 5-FU monotherapy or combined therapy with conventional therapeutics. In general, 5-FU chemotherapy is the first treatment chosen for advanced gastric cancer although its efficacy is limited due to drug resistance [3]. 5-FU resistant cells used in this experiment were established by stepwise exposure of cells to 100 μM 5-FU. It is based on the fact that plasma peak concentrations reach approximately between 100 and 1000 μM in patients given with 5-FU at a single dose of 60–500 mg/m² [17, 18]. In the present study, AGS FR cells were refractory to 5-FU compared with parental AGS cells in viability assays (Fig. 1). Like other chemotherapy, 5-FU resistance in patients with gastric cancer causes a serious problem. Accordingly, many studies have been aimed at understanding the resistance mechanisms to make use of them therapeutically [6]. For instance, gastric cancer cells with lower 5-FU sensitivity show higher expression of RhoGD12 than cells with extreme sensitivity to 5-FU. Intriguingly, upregulation of COX-2 is found in gastric cancer compared with normal paired mucosa [19].

Furthermore, the sensitivity of gastric cancer cells to chemotherapy drugs including 5-FU is negatively correlated with COX-2 expression [20]. Like colon cancer cells resistant to 5-FU [21], AGS FR cells showed marked expression of COX-2 and EP2 as compared with AGS cells, while there were little changes in other proapoptotic proteins (Fig. 2).

The COX-2 enzyme is highly expressed in various cancer cell lines and is closely associated with chemotherapy resistance [22]. Therefore, use of a selective COX-2 inhibitor can improve the therapeutic efficacy of anticancer drugs including 5-FU or cyclophosphamide. COX-2 inhibition also augments antiangiogenic cancer therapy and inhibits metastasis [23]. In contrast, a report demonstrated that celecoxib mitigates 5-FU-induced apoptotic cell death in colon cancer cells [24]. The authors suggested that celecoxib arrests cell cycle progression to prevent cell death caused by 5-FU [24]. This contradictory effect of celecoxib in combination with 5-FU may be attributed to differences in cell type and expression profiles of cell cycle regulators.

Beside the inhibition of celecoxib against COX-2 activity, celecoxib treatment also suppresses COX-2 expression in a concentration-dependent manner (Fig. 4). Similarly, the expression of cox-2 mRNA was negatively regulated by celecoxib in MCF-7 and MDA-MB-231 cells [25]. Furthermore, celecoxib inhibited the phorbol ester-induced PGE₂ production and COX-2 expression in mouse skin [26]. Therefore, suppression of both the catalytic activity and expression of COX-2 by celecoxib may effectively lead to a decrease in prostaglandin metabolism.

PGE₂ is a critical cellular mediator that promotes the progression of tumorigenesis by stimulating cell proliferation and producing angiogenic molecules as well as downregulating apoptosis-related proteins [27, 28]. Celecoxib treatment was reported to induce apoptosis by inhibiting 3-phosphoinositide-dependent protein kinase-1 (PDK1) activity in the human colon cancer HT-29 cell line [29]. The authors found that overexpression of active constitutive PDK1 is as protective as the pan-caspase inhibitor from celecoxib-induced cell death. Specifically, PGE₂ inhibits apoptosis in the Caco-2 cell line through Ras-PI3K association, which is dependent on the activation of EP4 receptor and the activation of PKA [27]. PGE₂ also protects gastric mucosal cells from apoptosis via activation of the EP2/EP4 receptor [30], suggesting that the PKA pathway is responsible for PGE₂-mediated inhibition of apoptosis. In contrast, PGE₂ also induced apoptosis of normal lung fibroblasts in a concentration-dependent manner, as determined by Annexin V staining and caspase 3 activity [31]. These conflicting results suggest that PGE₂ plays pleiotropic functions in different cells, experimental conditions, with differential expressions of receptor isoforms and various death or survival stimuli [28, 32–34].

In our results, AGS FR cells considerably produced PGE₂ than AGS upon stimulation with 5-FU, as expected from elevated expression of COX-2 in AGS FR (Fig. 2b). Upregulation of PGE₂ has been found in other different cancers such as colon and esophageal cancers when treated with 5-FU [21, 35]. Increase in PGE₂ level was shown to be related with suppressed apoptosis during chemotherapy or radiotherapy. Furthermore, treatment of AGS FR cells with celecoxib induced early or late apoptosis populations in a concentration-dependent manner (Fig. 3). MTS results demonstrated that celecoxib had concentration-dependent cytotoxic effects for 24, 48 and 72 h. In line with this result, FACS analysis showed a considerable increase in the Annexin V-positive population and Annexin V/PI-double positive population in cells exposed to celecoxib at 80 μM. Selective inhibition of COX-2 through celecoxib seemed to induce cell death but not to inhibit cell growth. Although mechanisms for celecoxib-induced cytotoxicity still remain elusive, some publications have shown that COX-2 inhibition can induce cell death. Celecoxib treatment has been suggested to induce apoptosis in human non-small cell lung cancer cells (NSCLC) via upregulation of the extrinsic death receptor pathway [36]. In a recent article, celecoxib induced apoptosis of NSCLC via endoplasmic reticulum stress [37].

In a previous report, inhibition of COX-2 and 5-LOX suppressed the progression of colorectal cancer by enhancing PTEN and disrupting the PI3K/Akt pathway [38]. In colorectal cancer patients, tumor-specific PTEN/Akt pathway activation is closely associated with poorer cancer specific survival (CSS) when peri-nuclear COX-2 is defective [39]. Moreover, COX-2 inhibitor-upregulated PTEN is proposed to not only antagonize radiation-induced Akt activation by downregulation of Sp1 expression, leading to radiosensitization, but also prevent hepatoma stemness and progression [40, 41].

PGE₂ also promotes cell survival via the PI3K/Akt pathway after binding to its cognate receptor EP4 [33]. TGF-β has an effect on cell migration and invasion of prostate cancer cells via PGE₂ through activation of the PI3K/Akt pathway [42]. TGF-β is reported to enhance COX-2 levels to produce PGE₂ secretion in prostate cancer cells and dental pulp cells [43]. This stimulating effect of TGF-β on COX-2 expression was closely associated with activation of the ALK5/Smad2/3 and MEK/ERK pathways. Another study showed that PTEN removes the phosphate from phosphatidylinositol-3,4,5-triphosphate and functions as a tumor suppressor by negatively regulating the Akt signaling pathway [44]. Akt activation mediates the ubiquitination and degradation of PTEN via the MKRN1 E3 ligase [45]. As shown in Figs. 4b and Fig. 5b, expression of p-Akt was reversely correlated with PTEN expression. Clinically, Akt phosphorylation and PTEN loss is reported to be involved in the development of some rituximab-resistant diffuse large B cell lymphoma (DLBCL) patients and chemoresistance-acquired gastric cancer [46, 47]. In addition, Akt phosphorylation was accompanied by the loss of PTEN in specimens of endometrial carcinomas [48]. Therefore, PTEN inactivation or loss may contribute to carcinogenesis via modulation of the Akt signaling pathway.

The prostaglandin receptor EP2 is involved in cancer cell proliferation and invasion and its receptor regulation can be mediated through binding of agonist or antagonist to control tumorigenesis [16]. The EP2 agonist butaprost more strongly potentiated cell proliferation of AGS FR cells than that of AGS cells (Fig. 5). This suggests that the COX-2/PGE₂ signaling pathway can regulate cell survival or death through the EP2 receptor. Therefore, negative modulation of COX-2 catalytic activity by pharmacological drugs might be a promising therapy to overcome cancer drug resistance.

Compliance with ethical standards

Conflict of interest

The authors declare no conflicts of interest.