Introduction

5-Fluorouracil (5FU) has been used for decades as a treatment for patients with advanced and metastatic cancers of the stomach and colon. Because response to 5FU was not always satisfactory, trials to improve response rates led to the use of combination therapy with two or three different drugs. One combination, 5FU and MMC, has been used widely for the therapy of advanced gastric cancer [1, 2]. However, there are limited data regarding the precise protocol that might offer the best therapeutic effect. One study used MMC on day 1 with subsequent addition of 5FU [3], while another pretreated with 5FU [4]. However, neither combination was compared to each other nor to other drug schedules. The mode of action of both drugs differs, suggesting that better understanding of the combination might lead to a more correct use of the drugs.

The basis of the therapeutic effect of 5FU depends upon metabolic conversion into three active metabolites. One of the metabolites, fluorodeoxyuridine monophosphate, blocks DNA synthesis by inactivating thymidylate synthase, leading to a reduced concentration of thymidylate and DNA degradation [5, 6]. The other two metabolites, fluorouridine triphosphate and fluorodeoxyuridine triphosphate, result in an inhibition of RNA maturation and an induction of single-strand breaks in DNA, respectively [7, 8]. MMC has two major modes of action. First, it inhibits DNA topoisomerase-II. Second, this 1,4-quinone forms a semiquinone radical that can transfer an electron to oxygen with generation of a superoxide radical. This reaction is catalyzed by flavoenzymes such as NADPH-cytochrome P-450 reductase [9, 10]. Both the semiquinone or superoxide radicals can generate hydroxyl radicals that can covalently bind to DNA and cause strand breaks, which inhibits DNA replication and repair [11, 12]. In addition to the crucial toxic effect of DNA intercalation, under aerobic conditions reactive oxygen species (ROS) generated in the redox cycle of MMC are also thought to participate in MMC-induced cellular toxicity [9]. This effect has been demonstrated in cultured cells exposed to the anthracycline antibiotics, adriamycin [13] and idarubicin [14]. In experiments using tumor cells, it has been suggested that ROS-dependent apoptotic signaling including mitochondrial dysfunction and caspase activation is involved in toxic events induced by MMC [15, 16].

Addition of 5FU to gastric cancer cells treated with a platinum-based drug, cisplatin, was effective in controlling cell growth [17]. Moreover, apoptosis induced by 5FU were associated with p53 gene status in gastric cancer cell lines [18]. Although the potency of a variety of combinations has so far been examined by numerous investigators both clinically and in cell cultures, the molecular mechanism(s) of the combined effects of these drugs on toxicity against gastric cancer cells still remains unclear. In this paper, we have therefore focused on the combined use of 5FU and MMC against gastric cancer MKN45 cells, and investigated the addition sequence of the two drugs, and the mechanism of the toxic effect with special attention to alterations in apoptotic-related factors, changes in antioxidant enzymes, and ROS production as factors of potential importance.

Materials and methods

Materials

5FU and MMC were kindly supplied by Kyowa Hakko Chemical. 2-(4-Iodophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium mono sodium salt (WST-1) was purchased from Wako Pure Chemical Industries; 6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate di(acetoxymethylester) (DCFH) was from molecular probes; Cu, Zn-superoxide dismutase (SOD) was from UBE industrials; and acetyl Asp-Glu-Val-Asp p-nitroanilide and acetyl Leu-Glu-His-Asp p-nitroanilide were from Sigma-Aldrich. All other chemicals were of the highest grade that could be obtained commercially.

Cell culture and treatment

Human gastric cancer cell line MKN45 and MKN28 (Health Science Research Resources Bank) were cultured in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and antibiotics in a 5% CO2 incubator. The cells were seeded at densities of 2 × 105 cells/dish and 1 × 104 cells/well into a 60-mm dish and a 96-well microplate, respectively. After reaching 70–80% confluence of the cells, the medium was replaced with DMEM containing 2% FBS 24 h before treatment of drugs. The cells were treated concomitantly with 5FU and MMC (FM treatment), or with MMC after 24-h pretreatment with 5FU (F24M treatment).

Cell viability was measured by the cytotoxicity assay, based on the activity of mitochondrial dehydrogenase reducing 3-(4,5-dimethyl-2-thizalyl)-2,5-diphenyl-2H tetrazolium bromide (MTT) to a purple insoluble formazan compound, but using WST-1 as the chromogen [19].

Western blot analysis

For the analysis of cytochrome c release from the mitochondria, the cells were washed twice with Dulbecco’s phosphate-buffered saline (DPBS) supplemented with 40 μg/ml saponin, and homogenized by passing the suspension through a 26 gauge needle (20 strokes). The cytosolic fraction of the homogenate was obtained by centrifugation at 12,000×g for 15 min. In the analyses of other proteins, the cells washed with DPBS were suspended before sonication in DPBS containing 0.5% Triton X-100 and 0.3 mM phenylmethanesulfonyl fluoride. The cell extracts were isolated by centrifuging at 12,000×g for 15 min, and protein concentration in the extracts was determined with a bicinchoninic acid Protein Assay System (Pierce). The proteins in the cytosolic fractions and extracts of the cells were electrophoretically separated on a 7.5 or 12% SDS-polyacrylamide gel under reducing conditions, and then transferred to a PVDF membrane (Millipore) by electroblotting. After blocking with 5% skim milk, the membrane was allowed to react with primary monoclonal antibodies against human Bcl-2 (BD Transduction Laboratories), Bcl-XL (BD Transduction Laboratories), Bax (Santa Cruz Biotechnology), cytochrome c (Trevigen), CuZnSOD (Binding Site) and MnSOD (Binding Site). The immunoreactive proteins were detected by being reacted with peroxidase-conjugated secondary antibody (Bio-rad), in which the peroxidase activity was visualized by means of an enhanced chemiluminescence substrate system (Amersham Biosciences).

Assay of enzyme activity

Cells were washed twice with DPBS, suspended in 10 mM Tris–HCl, pH 7.4, containing 5 mM 2-mercaptoethanol and 20% glycerol, and then homogenized by passing the suspension through a 26 gauge needle. The homogenate was centrifuged at 12,000×g for 15 min, and MMC reductase activity in the supernatant was determined by monitoring oxidation rate of NADPH at 340 nm. The assay mixture consisted of 0.1 M potassium phosphate, pH 7.4, 0.1 mM NADPH, 9 μM MMC and the cell extract, in a total volume of 2.0 ml. The rate was corrected for the nonenzymatic NADPH oxidation by MMC in the mixture without the extract. One unit (U) of enzyme activity was defined as the amount of enzyme that catalyzes the formation or oxidation of 1 μmol NADPH per minute at 25°C.

For assay of caspase activity, the cell extracts were prepared as described above, except that 50 mM HEPES–NaOH, pH 7.4, containing 5 mM 3-[(3-cholamidopropyl) dimethylammonio]-propanesulfonic acid and 5 mM dithiothreitol was used as the homogenized solution. The activities of caspase-3 and caspase-9 were spectrophotometrically measured at 415 nm using acetyl Asp-Glu-Val-Asp p-nitroanilide and acetyl Leu-Glu-His-Asp p-nitroanilide, respectively, as the substrates [20].

Superoxide dismutase activity in the cell extract was determined using a SOD Activity Detection Kit (Wako Pure Chemical) according to the manufacturer’s instruction.

Measurement of ROS

Level of intracellular hydrogen peroxide was estimated by DCFH oxidation as previously reported [21]. Following washes with DPBS, the cells were incubated for 20 min in fresh medium containing 10 μM DCFH, and again sufficiently washed with DPBS. DCF fluorescence-positive cells were counted using a FACScan equipped with a fluorescein isothiocyanate filter and Cell Quest Pro software (Becton-Dickinson).

Superoxide anions generated by reduction of MMC in the cell extract were determined by the method of McCord and Fridovich [22]. The assay mixture consisted of 0.1 M potassium phosphate buffer, pH 7.4, 0.5 mM NADPH, 0.1 mM EDTA, 50 μM ferricytochrome c, 9 μM MMC and the cell extract, in a total volume of 2.0 ml. The formation rate of ferrocytochrome c at 37°C was spectrophotometrically monitored at 550 nm.

Measurement of 8-hydroxy-deoxyguanine

8-Hydroxy-deoxyguanine (8-OH-dG) concentration in the cell extracts was determined using a 8-OH-dG enzyme-linked immunosorbent assay kit according to the manufacturer’s instructions (Japan Institute for the Control and Aging).

Measurement of intracellular reduced glutathione

Glutathione (GSH) concentration was measured by HPLC as its o-phthalaldehyde adduct [23]. Cells were washed twice with DPBS and lysed by sonication for 1 min. Following centrifugation at 12,000×g for 15 min, GSH in the supernatant was derivatized by incubating for 30 min with o-phthalaldehyde, and an aliquote of the derivatized preparation was subjected to HPLC using a C18 column, which was eluted isocratically with a mobile phase consisting of 0.15 M sodium acetate/methanol (91.5:8.5) at a flow rate of 0.5 ml/min. The o-phthalaldehyde-GSH adduct was monitored using a fluorescence detector operating at excitation and emission wavelengths of 250 and 410 nm, respectively.

Statistical analysis

Data are expressed as the means ± SD of at least three independent experiments. Statistical evaluation of the data was performed by using the unpaired Student’s t test and ANOVA followed by Fisher’s test. A p value <0.05 was considered statistically significant.

Results

Pretreatment with 5FU potentiates MMC-induced toxicity against gastric cancer MKN45 cells

Treatment with 9 μM MMC decreased viability of MKN45 cells, reaching 34% by 72 h, although the viability was less affected by treating with 23 μM 5FU (Fig. 1a). Concomitant treatment with 5FU (FM treatment) slightly strengthened the MMC-induced cytotoxicity. In contrast, pretreatment of the cells for 24 h with the same concentration of 5FU (F 24 M treatment) drastically increased the susceptibility to MMC. As a dose–response experiment of MMC in the F24M treatment is shown in Fig. 1b, the marked reduction in the viability was observed when the cells were treated for 48 h with higher concentrations (≥9 μM) of MMC. The potentiating effect of the 5FU pretreatment for 24 h on the MMC-induced cytotoxicity was almost the same as that of the longer time (≥48 h) pretreatment, and was stronger than that of the MMC pretreatment on the 5FU-induced toxicity (data not shown). Therefore, the 24-h pretreatment with 5FU was adopted in all subsequent experiments.

Fig. 1
figure 1

Viability of MKN45 cells in response to drug treatment. a Time-course experiment. Cells were treated for the indicated times with 5FU (23 μM) alone (open circle), MMC (9 μM) alone (filled circle) or FM (open square). In the F24M treatment, the cells were pretreated for 24 h with 5FU prior to treating with MMC (filled square). The values are normalized to those in cells treated with dimethylsulfoxide alone. b Dose–response. The cells were treated with MMC alone (filled circle), FM (open square) and F24M (filled square), and the treatments were terminated 48 h after adding the indicated concentrations of MMC to the medium

Full size image

Activation of p53-dependent apoptotic mechanisms in the F24M-treated cells

To examine the involvement of p53 in the cytotoxicity induced by MMC and 5FU, we compared difference in susceptibility to the two drugs and their combined treatments (FM and F24M) between poorly differentiated MKN45 and highly differentiated MKN28 cells that bear wild-type p53 and its mutated form, respectively (Fig. 2a). No significant difference in the viability between the two cells was observed in the treatments of single drug (5FU or MMC) and FM, whereas in the F24M treatment the viability of MKN45 cells was significantly low compared to that of MKN28 cells. Levels of Bcl-2 and Bcl-XL were low and that of Bax was greatly increased in the F24M-treated cells, compared to the cells treated with the single drug and FM (Fig. 2b). In addition, the release of cytochrome c into the cytosol of the F24M treated cells was observed at 12 h after the incubation with MMC (Fig. 2c). Marginal activations in caspase-3 and caspase-9 were furthermore induced only by the F24M treatment (Fig. 2d). It should be noted that release of cytochrome c and activation of caspases observed during treatment of MMC alone or FM were much less than those during the F24M treatment.

Fig. 2
figure 2

Apoptotic signaling in MKN45 cells induced by F24M treatment. a Comparison of toxicities of 5FU and MMC between MKN45 (open bar) and MKN28 cells (filled bar). The cells were treated with 5FU, MMC alone, FM, or F24M, and the treatments were terminated 48 h after initiation of MMC (9 μM). The values are normalized to that in cells treated with dimethylsulfoxide. *Difference from the F24M-treated MKN45 cells, p < 0.05. b Alterations in Bcl-2, Bcl-XL and Bax. The cells were treated as described in a, except that the incubation with MMC was terminated at 24 h. Cell extracts (20 μg of protein) were analyzed by Western blots probed with antibodies against Bcl-2, Bcl-XL and Bax. c Release of cytochrome c into cytosol after F24M treatment of MKN45 cells. The cells were pretreated for 24 h with 5FU and then treated for 0, 6, 12, 24 and 48 h with MMC (9 μM). The cytosolic fraction (40 μg of protein) was analyzed by Western blotting. d Activation of caspases. Cells were treated as noted in a, and caspase-3 (open bar) and caspase-9 (filled bar) activities were measured. Each activity is expressed as the fold increase in the ratio of the activity in the treated group compared with that in the untreated group

Full size image

Pretreatment with 5FU increases ROS production and GSH depletion by MMC

The DCFH oxidation assay was used to estimate amounts of intracellular H2O2-derived oxidants. The DCF-derived fluorescence was slightly enhanced in the cells treated for 24 h either with MMC alone or FM (Fig. 3a, panels a and b). In contrast, 5FU pretreatment greatly augmented the ratio of DCF-positive/total cells following addition of MMC (panel c). The ratio of DCF-positive/total cells was normalized by addition of polyethyleneglycol-conjugated catalase (PEG-cat) (panel d), confirming that the DCF oxidation reflected H2O2 production under the experimental conditions. The ratio was not so strong as that observed in cells treated with glucose oxidase, which causes continuous H2O2 production during the oxidation of glucose (panel e). In addition, F24M treatment of MKN45 elicited an increase in 8-OH-dG level (Fig. 3b) and a decrease in reduced GSH (Fig. 3c) in the cells compared to treatment of MMC alone. These results demonstrated that ROS-dependent apoptotic signaling is highly activated by 5FU pretreatment and participates in the MMC-induced toxicity against MKN45 cells. Pretreatment with ROS inhibitors, cell-membrane permeable GSH ethyl ester (GSH ester) and PEG-cat, almost completely suppressed endogenous cytochrome c release in the F24M-treated cells (Fig. 4a), and improved cell viability resulted from the 5FU pretreatment (Fig. 4b).

Fig. 3
figure 3

Intracellular ROS generation enhanced by F24M treatment. a DCFH oxidation. MKN45 cells were treated with MMC (a), FM (b), or F24M (c), and the treatments were terminated 24 h after initiation of MMC (9 μM). PEG-cat (200 U/ml) was added to the culture media of (c) 2 h before treating with MMC (d). The cells were also treated for 4 h with glucose oxidase (20 mU/ml) in the presence of glucose (20 mM) (e). b 8-OH-dG formation. The cells were treated as noted in a, and in the 5FU group the cells were treated for 72 h with 5FU (23 μM). The values are expressed as the fold increase in the ratio of 8-OH-dG concentration in the treated cells compared with that in the untreated cells. c Intracellular GSH level. The cells were treated as noted in a, and GSH level was measured by HPLC analysis. The values are normalized to that in untreated cells and expressed as the percentage of reduced GSH level in the treated cells to that in the untreated cells

Full size image
Fig. 4
figure 4

Effect of ROS inhibitors on cytochrome c release and cytotoxicity induced by F24M treatment. a Cytochrome c release. Cells were pretreated for 24 h with 5FU and then treated for 24 h with MMC (9 μM). GSH ester (2 mM) or PEG-cat (200 U/ml) was added to the culture medium 2 h before initiation of MMC treatment. Cytosolic cytochrome c was assayed as in Fig. 2. b Cell viability. The cells were pretreated for 24 h with 5FU and then treated for 48 h with MMC (9 μM). The indicated concentrations of GSH ester or PEG-cat were added to the culture medium 2 h before initiation of MMC treatment. *Difference from the F24M group (no inhibitor), p < 0.05

Full size image

Mechanism of superoxide anion generation by F24M treatment

The NADPH-linked reductase activity of MMC was detected in the extracts of MKN45 cells, and was not significantly changed by MMC and F24M treatments (Fig. 5a). When the reductive rates of cytochrome c coupled with the in vitro reduction of MMC by the cell extracts were compared, the F24M-treated cells showed higher reduction rate than the cells treated with MMC alone (Fig. 5b, curve a) and without drugs (data not shown). The F24M treatment resulted in a noticeable decrease in SOD activity, whereas such decrease was not detected in the cells treated with 5FU or MMC alone (Fig. 6a). In the F24M-treated cells, the protein levels of CuZnSOD and MnSOD were decreased by 36 and 67%, respectively, of the basal level at the time point of 48 h after the addition of MMC, as estimated on densitometric analysis of the Western blotting results (Fig. 6b, c). The loss of SOD activity was partially reversed by ROS inhibitors (Fig. 6d).

Fig. 5
figure 5

MMC reduction in MKN45 cells. a MMC reductase activity. The cells were pretreated for 24 h without (MMC) or with 5FU (F 24 M), and followed by treating for 24 h with MMC (9 μM). MMC reductase activity in cell extracts (100 μg) was spectrophotometrically determined by monitoring oxidation rate of NADPH. b Cytochrome c reduction coupled with MMC reduction. The extracts (100 μg) prepared in a were incubated with cytochrome c and 9 μM MMC in the absence (curve a) or presence (curve b) of 200 U CuZnSOD. The reduction rate of cytochrome c was spectrophotometrically monitored at 550 nm. No significant reduction of cytochrome c was observed without MMC (curve c)

Full size image
Fig. 6
figure 6

F24M treatment attenuates SOD activity in MKN45 cells. a Time-course changes in SOD activity. The cells were treated with MMC (filled circle) or F24M (open square), and the treatments were terminated 24 h after initiation of MMC (9 μM). The cells were also treated for 72 h with 5FU (23 μM) alone (open circle). The values are normalized to that in cells before treatment and expressed as the percentage of the activity in the treated group to that in the untreated group. b Alterations in CuZnSOD and MnSOD. The proteins (20 μg) in the F24M-treated cell cytosol prepared in a were analyzed by Western blotting. c Densitometric data of the bands of CuZnSOD (open bar) and MnSOD (filled bar) shown in b. The values are expressed as the percentage of the band density in the treated group to that in the untreated group. d Effect of ROS inhibitors on SOD activity. The cells were pretreated for 24 h with 5FU and then treated for 24 h with MMC (9 μM). GSH ester (2 mM) or PEG-cat (200 U/ml) was added to the culture medium 2 h before initiation of MMC treatment. The values are expressed as the percentage of the activity in the treated group to that in the untreated group. *Difference from the F24M group (no inhibitor), p < 0.05

Full size image

Discussion

In this study, we address the optimal conditions to maximize toxicity against gastric cancer in combination of 5FU with MMC, and have found that the pretreatment for 24 h with 5FU results in a great enhancement of the susceptibility of MKN45 cells to a range of MMC commonly used in clinical treatment (Fig. 1). Previous literature showed that MMC is known to upregulate thymidine phosphorylase, a tumoral enzyme that converts 5FU into its active form, fluorouridinemonophosphate [24]. It is therefore speculated that the MMC-induced metabolic activation of 5FU also plays an important role in the combined chemotherapy of 5FU plus MMC. The results using the MKN45 and MKN28 cells (Fig. 2a) suggest that the additive effect of 5FU and MMC is mainly mediated via p53-dependent apoptotic signaling. In the experiment, the demonstration that mutation of p53 is involved in chemoresistance of gastric cancer toward the combined treatment with 5FU and MMC is consistent with the report that combined treatment with 5FU and cisplatin induces apoptosis in MKN45 cell but not in MKN28 cells [17]. The up-regulation of Bax protein, which is transcriptionally regulated by p53 [25], increases mitochondrial permeability via an alteration of balance between Bax and Bcl-2 (or Bcl-XL), and releases mitochondria-derived proteins, such as cytochrome c, leading to caspase activation [26]. As expected, the F24M treatment can alter these proteins and caspases involved in the p53-dependent signaling pathway (Fig. 2b–d). A variety of apoptotic stimuli, such as ceramide [21], hydrogen peroxide [27] and other chemotherapeutic agents [28] produce similar changes in cultured cell lines. During 5FU treatment, p53 activates mitochondrial ferredoxin reductase [29], which sensitizes cells to oxidative stress-induced apoptosis [30]. Therefore, the activation of ferredoxin reductase might be responsible for the synergic effect of 5FU on the MMC-induced cell death.

Malignant cells are sensitive to oxidative stress injury by exogenous ROS inducers, although cancer cells also have counter-regulatory mechanisms to quench intrinsic oxidative stress [31, 32]. In this report, pretreatment with 5FU augmented the MMC-induced ROS production and the enhanced ROS triggered further oxidative changes in gastric cancer cells (Fig. 3a, b). Since p53 is believed to be up-regulated and activated by 8-OH-dG formation, an indicator of oxidative damage for DNA [33], ROS generated during the F24M treatment might contribute to the increased to the p53-dependent apoptosis. GSH plays an essential role in many detoxification processes, and accumulation of oxidants causes its depletion through excessive GSSG formation or efflux from cells, linking to apoptotic initiation [34]. In the F24M-treated cells the intracellular level of GSH decreased proportionately with an increase in intracellular hydrogen peroxide (Fig. 3c). In addition, pretreatment with exogenous GSH ester and PEG-cat partially reversed the F24M treatment-induced cytochrome c release into the cytosol, the reduced cell viability, and the impaired SOD activities (Figs. 4, 6d). These results demonstrate that ROS-dependent pathways participate in the F24M treatment-induced cytotoxicity. However, the lack of complete reversal by anti-oxidant compounds could suggest that ROS-independent pathways also contribute to the toxic effects of the combined drugs.

Reactive oxygen radicals are chemically reactive and potentially toxic to cells when high levels accumulate, and need to be promptly eliminated from cells by a highly regulated antioxidant system that includes enzymes such as SOD, catalase, and GSH peroxidase. Our current study showed that the F24M treatment does not elicit any significant change in the MMC reductive activity in the cells (Fig. 5a). This result is not consistent with the hypothesis that accelerated turnover of free electrons in the MMC-redox cycle was driving the accumulation of ROS. Another explanation for the accumulation of ROS is that increased formation of intracellular ROS can reduce the levels of antioxidant proteins, such as SOD and GSH peroxidase [35]. In addition, inhibition of RNA synthesis by 5FU is theoretically thought to quench de novo syntheses of cellular proteins. It is therefore suggested that the marked reduction of SOD by the F24M treatment is mediated by both oxidant-induced degradation and inhibition of the de novo synthesis of the enzyme. The 5FU-induced inhibition of the de novo synthesis of SOD is supported by the data in Fig. 6a, in which the expression of SOD was transiently enhanced, possibly due to a cytoprotective effect, in the cells treated for 12 h with MMC, whereas such enhancement was not detected in the F24M treatment group. Thus, the data in the current study favor a decrease in anti-oxidant function (levels of GSH and CuZnSOD) as the major cause of increased ROS accumulation, rather than an increase in ROS production per se.

Collectively, our study shows that in combined chemotherapy of 5FU and MMC, 24-h pretreatment with 5FU maximizes the MMC-induced toxicity against gastric cancer, and the mechanism includes ROS accumulation and the resultant reduction in anti-oxidant enzymes, especially CuZnSOD. By comparison to previous data [3, 36], it is needed to evaluate efficacy of our observations in the clinical setting.