• Open Access

Involvement of reactive oxygen species in multidrug resistance of a vincristine-selected lymphoblastoma


To whom correspondence should be addressed. E-mail: cckchao@mail.cgu.edu.tw


Our previous study identified a vincristine-selected multidrug resistance (MDR) cell line, HOB1/VCR, derived from a lymphoblastoma HOB1. The HOB1/VCR cells are resistant to typical MDR drugs and are cross-resistant to P-glycoprotein-independent drugs such as cisplatin (cis-diamminedichloroplatinum [II]). The mechanism of this atypical MDR phenotype is uncertain. The present study provides evidence regarding the contribution of reactive oxygen species (ROS) to the resistance of cells in response to treatments (vincristine, cisplatin and H2O2). Notably, the HOB1/VCR cells were cross-resistant to H2O2. High levels of ROS formed in both sensitive and HOB1/VCR cells by H2O2, and moderate levels of ROS were generated by treatment with cisplatin and vincristine. The ROS level in HOB1/VCR cells was lower than that in sensitive cells following treatments. The ROS level was reduced markedly by a non-toxic concentration of N-acetyl-l-cysteine, a ROS scavenger, in drug-treated cells, and was correlated with reduced cytotoxicity. Furthermore, concentrations of glutathione and glutathione peroxidase, but not superoxide dismutase and catalase, increased in HOB/VCR cells. The dl-buthionine-[S,R]-sulfoximine inhibited formation of glutathione and sensitized both cell types to treatments. Therefore, overexpression of an H2O2-reducing system, glutathione–glutathione peroxidase, has a role in resistance. Experimental results further demonstrate that ROS is likely a primary signal in the acquisition of the MDR phenotype and therefore a potential target when designing drugs for chemoresistance. (Cancer Sci 2007; 98: 1206–1214)

Abbreviations: BSO



cis-diamminedichloroplatinum (II)




ethylenediaminetetraacetic acid


fluorescein isothiocyanate


glutathione peroxidase




2′7′-dichlorodihydrofluorescein diacetate


high-performance liquid chromatography


50% inhibitory concentration


multidrug resistance


multidrug resistance-associated protein




poly(ADP-ribose) polymerase-1


phosphate-buffered saline


reactive oxygen species


superoxide dismutase.

Drug resistance represents a major problem in the use of chemotherapy to treat various cancers.(1) MDR refers to the simultaneous resistance by tumors to numerous structurally and functionally unrelated chemotherapeutic agents.(2) It is commonly encountered in therapeutic courses and is normally associated with overexpression of P-glycoprotein, an ATP-dependent membrane transport protein, which is encoded by the mdr1 gene in humans(3,4) and was first characterized in multidrug-resistant Chinese hamster ovary cells by Ling and Thompson.(5) In addition to P-glycoprotein, other membrane proteins like MRP and, in some cases, ion channels are also involved in the acquired MDR phenotype.(3) These MDR cells are normally resistant to therapeutic compounds derived from plant alkaroids, such as vincristine and taxol, that cause dominant damage to microtubule function but are not resistant to DNA-damaging agents such as cisplatin (reviewed by Ambudkar et al.(6)). However, we previously identified an atypical MDR phenotype in a lymphoblastoma by sequential exposure of the cells to vincristine.(7) The MDR cells acquired resistance not only to typical MDR drugs, but also to P-glycoprotein-independent agents such as cisplatin.(8) The mechanism involved in this atypical MDR phenotype remains unclear.

Apoptosis is the genetic process of cell death that involves a complex mechanism, characterized by various well-defined features, including condensation and fragmentation of chromatin, oligonucleosomal fragmentation, membrane blebbing, loss of mitochondrial transmembrane potential and activation of caspases.(9) Caspase-independent apoptosis has also been reported in some cases.(10) Different agents induce apoptotic death, including death receptor ligands, serum deprivation and anticancer drugs. One of the critical determinants of cellular response to exogenous stimuli is the cellular redox status. Intracellular generation of ROS is tightly regulated by the intrinsic antioxidant defense systems. Oxidants such as H2O2 can induce apoptosis or necrosis depending on the dose added.(11) The exposure of cells to low levels of H2O2 can induce apoptosis.(12) Nevertheless, apoptosis differs significantly from necrosis, which follows a less defined pattern of events.

ROS are involved in many physiological functions, including aging,(13,14) inflammation and cancer.(15) At the cellular level, oxidative damage in DNA is repaired primarily via the base excision repair pathway,(16) and is associated with cell cycle arrest.(17) ROS, such as superoxide anions (inline image), H2O2 and hydroxyl radicals (.OH), are generated in response to anticancer drugs, and are also an essential component of apoptosis.(18) Interestingly, ROS can respond to a dramatic perturbation of the physiology of dying cells, converting late apoptotic steps into steps toward necrotic death.(16,19) These results suggest that the intracellular ROS level can determine the fate of the cells, including cell sensitivity or resistance to anticancer drugs. ROS is best known for its capacity to modify macromolecules such as DNA, lipids and proteins in cells.(20) It is hypothesized that the intracellular level of ROS may account for the atypical MDR phenotype in the lymphoblastoma cell line HOB/VCR, which we identified previously.(8)

Materials and Methods

Cell lines and reagents.  The lymphoblastoma cell line HOB1 and its MDR derivative HOB1/VCR, selected with vincristine, were maintained as described previously.(8) Reagents were purchased from Sigma Chemical Co. (St Louis, MO, USA) unless otherwise indicated.

Measurement of intracellular ROS.  Cells induced to undergo apoptosis were probed with redox-sensitive H2DCFDA (Molecular Probes, Eugene, OR, USA) to detect the intracellular production of ROS.(21,22) Cells (5 × 105/mL) were treated with the indicated concentrations of agents. For some experiments, 20 mM NAC, a non-toxic concentration of ROS scavenger, was included. The treated cells were washed in PBS and then resuspended in 1 mL culture medium with 10 µM/mL H2DCFDA, before being incubated for 40 min at 37°C. After staining, cells were washed in HEPES buffer (5 mM KCl, 140 mM NaCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM Glucose, 5 mM HEPES pH 7.4) at 37°C for 30 min. Dye oxidation (indicated by an increase in fluorescence) was analyzed using a FACScan (Becton Dickinson, San Jose, CA, USA), with excitation and emission wavelengths of 488 and 530 nm, respectively.

Immunodetection of caspases.  Cells were treated with the indicated concentrations of drugs for 24 h, and detected for caspase-3 activation and PARP cleavage by immunoblotting using specific antibodies. Treated cells were first washed with PBS and lysed in RIPA lysis buffer (50 mM Tris-HCl pH 7.4, 1% Nonidet P-40 (NP-40), 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM ethylene glycol bis(βaminoethyl ether)-N,N,N′,N′-tetra-acetic acid (EGTA), 1 mM phenylmethylsulfonyl fluoride (PMSF), protease inhibitor cocktail diluted with RIPA buffer to 1× concentration [Roche, Mannheim, Germany], 1 mM Na3VO4 and 1 mM NaF) on ice for 1 h. Insoluble material was removed by centrifugation at approximately 13 000g for 10 min at 4°C. Protein concentrations were measured using a Bio-Rad protein assay kit (Bio-Rad Laboratory, Hercules, CA, USA). Proteins were separated on 10% acrylamide gels containing sodium dodecylsulfate, transferred onto polyvinylidene fluoride (PVDF) membranes (Pall Co., Pensacola, FL, USA) and incubated with antibodies that were reactive to caspase-3, PARP or β-actin (Santa Cruz Biotechnology, Santa Cruz, CA, USA). The antigen–antibody complexes were visualized with secondary antibodies conjugated with horseradish peroxidase (HRP), using a standard enhanced chemiluminescence reaction (Pierce, Rockford, IL, USA).

Measurement of cell viability.  Cells were treated with cisplatin, vincristine or H2O2 in serum-free medium for 2 h and cultured in normal medium for 3 days. Cell viability was determined by using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide colorimetric assay as described previously.(23,24)

Detection of apoptosis and necrosis.  Cells grown in six-well plates were either left unexposed or exposed to death-inducing agents and incubated at 37°C to assess apoptosis and necrosis. The cells were fixed with methanol and incubated with DAPI (Sigma Chemical Co.) solution for 5 min in darkness, and analyzed using a microscope at 420 nm, as described previously.(25) Apoptotic cells that showed apoptotic morphological features, including chromatin condensation and nuclear fragmentation,(26) were counted in six to eight randomly selected fields. Exactly 900 nuclei of each sample were examined, and the results were expressed as the ratio of the number of apoptotic nuclei to the total number of nuclei counted. To support apoptosis by DAPI analysis, a sub-G1 population of cells was also measured by flow cytometry as described previously.(27) The cells were washed in ice-cold PBS and fixed in 70% (v/v) ethanol. They were resuspended in 0.5 mL PBS and then incubated in 0.5 mL DNA extraction buffer (40 mM Na2HPO4, 20 mM citric acid pH 7.8) at room temperature for 5 min. The cells were washed once with PBS, followed by incubation in PBS containing 20 µg/mL propidium iodide (Sigma Chemical Co.) and 200 µg/mL RNase A (Sigma Chemical Co.) for 30 min at room temperature in the dark. Stained nuclei were then analyzed using a FACScan with 10 000 events/determination. LYSYS II software (Taxarkana, TX, USA) was used to assess cell cycle distribution.

To simultaneously detect apoptotic and necrotic cells, cells were stained with propidium iodide or FITC-conjugated annexin V, and subjected to flow cytometry. Briefly, these cells were washed in PBS and fixed in 70% (v/v) ethanol. Cells were stained with FITC–annexin V (1 µg/mL; Strong Biotech Co., Taipei, Taiwan) and propidium iodide (10 µg/mL; Sigma Chemical Co.) in PBS (pH 7.4), and were subjected to flow cytometry using a FACScan as described previously.(28) It should be noted that the levels of annexin V staining were similar in unfixed cells. Cells were classified as early apoptotic (FL1, green), necrotic (FL2, red) or late apoptotic/necrotic (FL1 and FL2). Experiments were carried out in triplicate. LYSYS II software was used in the two-color flow cytometry to assess cell distribution.

GPx assay.  The GPx activity was measured spectro photometrically at 340 nm by the oxidation of NADPH as described previously.(29,30) Briefly stated, cells were resuspended in 1 mL extraction buffer (100 mM Tris-HCl [pH 7.6]/10 mM EDTA) and disrupted by sonication. The resulting lysate was clarified by centrifugation at 12 000g at 4°C for 15 min, and the supernatant was used for the assay. A typical assay mixture consisted of adequate cell lysate protein in 1 mL of assay buffer (50 mM Tris-HCl [pH 7.6], 5 mM EDTA, 0.12 mM NADPH, 2.5 mM reduced GSH, 0.042% H2O2, and 1 U GSH reductase). The decrease in absorbance at 340 nm was measured. One unit of enzyme was calculated as 1 nmol NADPH depleted per minute.

Catalase activity.  The catalase activity was measured spectrophotometrically at 240 nm in the presence of H2O2 as described previously.(31) Cells were resuspended in 1 mL of extraction buffer (100 mM Tris-HCl [pH 7.6] and 10 mM EDTA) and disrupted by sonication. The lysate was incubated with 50 mM potassium phosphate and 20 mM H2O2 and the change in absorbance at 240 nm was measured spectrophotometrically. Activity was expressed in units/mg protein calculated from the equation:

ΔA240/min × 1000/(43.6 × mg enzyme/mL reaction mixture).

Superoxide dismutase assay.  The level of SOD was assayed using the RANDOX SOD assay kit (Randox Laboratories, Antrim UK) according to manufacturer's instructions.

GSH determination by HPLC.  Cellular levels of GSH were determined by HPLC. BSO, a GSH inhibitor, was included in some experiments. Cells were lysed in 1% metaphosphoric acid, and the resulting lysate was clarified by centrifugation at 12 000g at 4°C for 15 min. An aliquot (10–20 µL) of supernatant was analyzed on an ESA CouloArray HPLC instrument equipped with electrochemical detector. Eight electrochemical cells were arranged in line and set to potentials between 400 and 750 mV. Chromatographic conditions were modified from those described previously.(32,33) A Nova-Pak ODS column (Marion, IN, USA) was used. The mobile phase was composed of 10 mM sodium phosphate (pH 2.7), 0.03 mM octane sulfonic acid and 2% (v/v) acetonitrile.

Statistical analysis.  Data were analyzed using a paired Student's t-test to determine the statistical significance of the data obtained and to compare the means of the two experimental groups. Data are expressed as means ± SD. Differences between two groups were considered significant when P-values were less than 0.05.


Cross-resistance to cisplatin and H2O2 of an MDR lymphoma.  HOB1/VCR, a vincristine-selected MDR cell line, acquired resistance to a variety of anticancer drugs, such as DNA-damaging agents like cisplatin and microtubule-damaging agents like vincristine.(8) These drugs generally control cancers by apoptosis.(34–36) Unexpectedly, HOB1/VCR also displayed cross-resistance to H2O2, a potent ROS in cells (Fig. 1a). The acquired resistance (IC50 of HOB1/VCR divided by IC50 of HOB1) to cisplatin, vincristine and H2O2 was 67-, 214- and 21-fold, respectively. Caspase-dependent apoptosis was also investigated to compare the response of the MDR and its parental HOB1 cells to these treatments. A typical pattern of caspase-3 activation and apoptosis is shown (Fig. 1b). Within a range of cytotoxic concentrations, caspase-3 was activated by cisplatin treatment, as reflected by the appearance of its cleavage products and PARP cleavage in HOB1 cells in a dose-dependent manner, with a concomitant increase in apoptosis (Fig. 1b, top panel). In contrast, the same drug concentration ranges did not cause caspase-3 activation or apoptosis in HOB1/VCR cells. A similar pattern of apoptosis was found in vincristine-treated cells (Fig. 1b, middle panel). HOB1/VCR cells were also resistant to other anticancer drugs, such as adriamycin, which kills cells via apoptosis (data not shown). Notably, the resistant cells intrinsically expressed a reduced level of caspase-3. Dramatic PARP cleavage and slight caspase-3 activation were also detected in HOB1 but not HOB1/VCR cells following treatment with cytotoxic concentrations of H2O2 (Fig. 1b, bottom panel). However, the level of apoptosis was undetected in H2O2-treated HOB1 and HOB1/VCR cells. These results indicate that HOB1/VCR cells are resistant to apoptotic agents like cisplatin and vincristine, and are also resistant to H2O2.

Figure 1.

(a) Cross-resistance to cisplatin (cis-diamminedichloroplatinum [II]) and H2O2 of HOB/VCR cells. The acquired resistance to cisplatin, vincristine and H2O2 was determined by MTT assay. The IC50 of HOB1 divided by the IC50 of HOB/VCR to cisplatin, vincristine and H2O2 was 67-, 214- and 21-fold, respectively. (b) Reduced caspase activation, poly(ADP-ribose) polymerase-1 (PARP) cleavage and apoptosis in multidrug resistant (MDR) lymphoblastoma cells. Top panel, cisplatin-induced caspase activation, PARP cleavage and apoptosis. 50 µg of HOB1 or resistant HOB1/VCR cell extracts, treated with the indicated concentrations of cisplatin, were immunoblotted with specific antibodies. Percentage apoptosis is indicated at the bottom. Middle and bottom panels are vincristine- and H2O2-induced caspase activation, PARP cleavage and apoptosis, repectively. The experimental design was the same as for the top panel, except that vincristine or H2O2 was used instead of cisplatin.

Decreased ROS level in H2O2-treated MDR lymphoma cells.  MDR lymphoma cells are resistant to H2O2, a potent ROS. To explore the role of ROS in drug resistance, the steady-state level of ROS in H2O2-treated sensitive and resistant cell lines was determined using the oxidant-sensitive dye H2DCFDA and flow cytometry. Representative patterns of flow cytometry of time- and dose-dependent responses of cells to H2O2 are shown (Fig. 2a,b). For some experiments, a sublethal concentration of NAC (20 mM), a ROS scavenger, was included. The ROS levels increased with the time course of treatments (Fig. 2a). If the intracellular ROS level was eliminated by the antioxidant system, the ROS levels in pulse-treated cells would decrease with the time. The intracellular ROS levels generated were dependent on the concentrations of H2O2 administered (Fig. 2b). The remaining H2O2, a type of ROS, may have been detected in this condition. The average levels of ROS were calculated (Fig. 2c). These results indicate that the ROS levels accumulated dose dependently in H2O2-treated cells.

Figure 2.

Reduced steady-state level of reactive oxygen species (ROS) in H2O2-treated multidrug resistant (MDR) lymphoblastoma cells. (a) Temporal kinetic patterns of ROS in H2O2-treated cells. The steady-state level of ROS in HOB1 and HOB1/VCR cells treated with 20 mM of H2O2 was examined. (b) Representative patterns of steady-state ROS levels in cells treated with different concentrations of H2O2. Cells were treated for 20 min. N-acetyl-l-cysteine was included in some experiments. (c) Steady-state level of ROS in H2O2-treated MDR lymphoblastoma cells. The average and deviation of ROS levels of triplicate samples were calculated. Best-fit lines are also indicated. **P-value < 0.005.

Inefficiency in apoptotic induction of H2O2-treated MDR lymphoblastoma.  To assess the role of ROS in the MDR phenotype, cell death was investigated in H2O2-treated HOB1 and HOB1/VCR cells. At high concentrations of H2O2, cell death increased with treatment time in both cell lines (Fig. 3a). Cell death reached nearly 90% at 24 h after H2O2 treatment. Dose-dependent cell death was detected in HOB1 cells starting at 12 h after H2O2 treatment. A time-dependent increase in cell death was also observed in H2O2-treated HOB1/VCR cells. Interestingly, H2O2 treatment during the early time caused more severe cell death in HOB1/VCR cells than HOB1 cells. To assess the significance of ROS in cell death, both cell lines were also treated with low concentrations of H2O2 in the presence or absence of NAC (Fig. 3b). H2O2 at a concentration of 0.1 mM caused more cell death in HOB1 cells than in HOB1/VCR cells. Cell death was equally induced in both cell lines by 1 mM H2O2, more severely than 0.1 mM H2O2. In the presence of NAC, H2O2 treatment was less effective at causing cell death in both cell lines. These results reveal that HOB1/VCR cells are resistant to H2O2 and that ROS is likely involved in the MDR phenotype.

Figure 3.

Cross-resistance to H2O2 of multidrug resistant lymphoblastoma cells. (a) Kinetic pattern of H2O2-induced cell death. The percentage of dead cells, including apoptotic and necrotic cells, is included. Data were calculated from triplicate experiments. (b) Attenuation of cell death induced by inhibiting reactive oxygen species. Data were calculated from triplicate experiments. (c) Representative cell distributions following treatment with 0.1 mM H2O2. Cells were stained with propidium iodide (FL2) and fluorescein isothiocyanate–annexinV (FL1), and the fluorescent cells were detected by flow cytometry. Shown are representative dot plots from two-color flow cytometry: lower-left quadrant, viable; upper-left quadrant, necrotic; lower-right quadrant, early apoptotic; upper-right quadrant, late apoptotic. (d) Patterns of necrotic and apoptotic cell death. Only results following treatment with low doses of H2O2 are shown as no apoptosis was detected at high doses.

HOB1/VCR cells are resistant to cisplatin and vincristine (see Fig. 1). H2O2 is a potent necrotic agent. To further assess HOB1/VCR cell sensitivity to H2O2, induced apoptosis and necrosis were compared in these cell lines. Propidium iodide and phosphatidylserine (FITC–annexin V) double staining enabled both necrotic and apoptotic cells, respectively, to be detected by flow cytometry. Fig. 3c presents representative data of 0.1 mM H2O2-treated cells. The distributed cells were divided into four populations. In the control, both HOB1 and resistant cells exhibited low propidium iodide and phosphatidylserine (PS) staining (Fig. 3c, lower left window). Following 8 h of incubation, some cells showed strong propidium iodide staining (Fig. 3c, upper windows); even greater populations of treated cells accumulated in this window following 24 h of incubation. Some cells were in the late apoptotic state (Fig. 3c, upper right windows), especially the HOB1 cells treated with a low concentration of H2O2 (Fig. 3c). Unlike those treated with cisplatin or vincristine (see below), the H2O2-treated HOB1/VCR cells displayed a similar response pattern to that of the HOB1 cells. Only a small fraction of cells were apoptotic compared to necrotic cells (Fig. 3d). Apoptotic cells were detected following treatment with a low concentration of H2O2 for 24 h. Although caspase activation and apoptosis were relatively insensitive to high concentrations of H2O2 (see Fig. 1), low concentrations of H2O2 were effective in eliciting apoptosis in HOB1 cells, although this was dramatically reduced in HOB1/VCR cells. Therefore, minimal apoptosis was activated by H2O2 in both cell lines. In addition to apoptotic resistance, HOB1/VCR cells also showed necrotic resistance to H2O2 depending on drug concentrations.

Decreased induction of ROS and apoptosis in cisplatin-treated MDR lymphoma.  To explore the role of ROS in cisplatin resistance, the intracellular ROS production induced by cisplatin in sensitive and resistant cell lines was compared using the oxidant-sensitive dye H2DCFDA and flow cytometry. Patterns of flow cytometry of time- and dose-dependent responses of cells to cisplatin were determined (data not shown). The average level of ROS production was estimated (Fig. 4a). A dose-dependent increase of ROS production in HOB1 cells was observed. The ROS intensity was around 3–5 relative fluorescence intensity values in HOB1 cells treated with a range of highly toxic concentrations. In contrast, ROS production was completely diminished in the resistant cells. However, the probe may be pumped out of cells in cases of resistant cells that have very active P-glycoprotein systems. The P-glycoprotein inhibiter verapamil was used to test this possibility. The results indicated that P-glycoprotein in the resistant cells was not responsible for their low ROS production (data not shown). In the presence of NAC (20 mM), ROS production was completely blocked (Fig. 4a). Apoptosis was also investigated in cisplatin-treated cells, using the same range of drug concentrations (Fig. 4b). A dose-dependent increase in apoptosis was observed in cisplatin-treated HOB1 cells. Most induced apoptosis was attenuated by NAC in cisplatin-treated HOB1 cells, but only a low level of apoptosis was induced in the resistant cells with cisplatin alone or together with NAC. Measurement of sub-G1 apoptotic cells also indicated reduced levels of ROS-dependent cytotoxicity by cisplatin in the resistant cells (Fig. 4c). The results indicate that ROS production is crucial to cisplatin-induced apoptosis, and that responses to cisplatin are drastically decreased in MDR cells.

Figure 4.

Reduced reactive oxygen species (ROS) production in cisplatin-treated multidrug resistant lymphoblastoma cells. (a) Average ROS production. Cells were treated with different concentrations of cisplatin for 8 h with or without N-acetyl-l-cysteine. The average and deviation of ROS intensity were calculated from triplicate samples. Best-fit lines are indicated. (b) Average apoptosis. Following each treatment, 900 cells were scored for apoptosis according to nuclear morphology. The average and deviation of apoptosis were calculated from triplicate samples. *P-value < 0.05. **P-value < 0.005. (c) Average sub-G1 cells induced by cisplatin. These values represent the total area of each sub-G1 peak modeled by the LYSYS II software. Each value represents the average of three experiments; bars, SD.

Decreased induction of ROS and apoptosis in vincristine-treated MDR lymphoma.  The ROS levels and apoptotic responses in vincristine-treated cells were also investigated (Fig. 5). Vincristine also induced ROS production at a maximum of 5 relative fluorescence intensity values in a dose-dependent manner (Fig. 5a). The induced ROS was nearly saturated when 10 µM of the drug was used, and the amount was partially attenuated by NAC. As for cisplatin treatment, vincristine was ineffective in inducing ROS in the resistant cells. Interestingly, 5 mM vincristine, which induced approximately 3 relative fluorescence intensity values, sufficed to induce 60% of apoptosis (Fig. 5b). The slight inhibition of ROS production by NAC is associated with a moderate reduction in apoptosis in vincristine-treated HOB1 cells. Again, a near basal level of ROS production was associated with low apoptosis in the resistant cells. Measurement of sub-G1 apoptotic cells also indicated decreased levels of ROS-dependent cytotoxicity by vincristine in the resistant cells (Fig. 5c).

Figure 5.

Reduced reactive oxygen species (ROS) production in vincristine-treated multidrug resistant lymphoblastoma cells. (a) Average ROS production. Cells were treated with different concentrations of vincristine for 4 h with or without N-acetyl-l-cysteine. The average and deviation of ROS intensity were calculated from triplicate samples. Best-fit lines for cells treated with drug are plotted. (b) Average apoptosis. The measurement was made as for Fig. 4. *P-value < 0.05. **P-value < 0.005. (c) Average sub-G1 cells induced by vincristine. Data analysis and symbols are the same as for Fig. 4c.

Enhanced expression of GSH and GPx in MDR lymphoma cells.  The fluorescence intensity in the flow cytometry experiments described above were dependent on the cellular GSH levels, owing to oxidation of the ROS probe dichlorohydrofluorescein (DCFH). To ensure the obtained data were accurate, the levels of cellular GSH as well as other antioxidants were determined (Table 1). The GSH levels in the HOB1/VCR cells were approximately two-fold those of the HOB1 cells. Antioxidant enzyme levels were also compared in both cell lines. As shown in Table 1, GPx in the HOB1/VCR cells was approximately 1.4-fold that of the HOB1 cells. However, the levels of SOD and catalase expressed in the HOB1/VCR cells were about the same as in the HOB1 cells. Although the levels of antioxidants and antioxidant enzymes control the rate of oxidation of ROS in cells, only selected redox-related molecules are overexpressed in HOB1/VCR cells.

Table 1. Redox-related molecules
  1. *Significant difference from HOB1 group (P < 0.05). Data are presented as mean ± SD of three measurements. Numbers in parentheses represent relative values of HOB1/VCR to HOB1.

Glutathione (nmol/mg)8.03 ± 1.6014.22 ± 3.06 (1.77)*
Glutathione peroxidase (nmol/min-mg)1.03 ± 0.20 1.41 ± 0.17 (1.37)*
Superoxide dismutase (U/mg)1.41 ± 0.31 0.98 ± 0.05 (0.70)
Catalase (U/mg)1.47 ± 0.70 1.30 ± 0.19 (0.88)

Sensitization of MDR lymphoblastoma cells to drugs by GSH attenuation.  To assess the role of GSH in the MDR phenotype, the levels of GSH and its inhibition in HOB1 and HOB1/VCR cells were investigated. A sublethal concentration of BSO (25 µM), a GSH inhibitor, was equally effective in inhibiting the GSH levels in both cell lines (Fig. 6). More than 85% of the GSH levels were blocked in both cell lines following 24 h incubation with BSO. Under these conditions, cytotoxicity induced by cisplatin, vincristine and H2O2 was investigated. The IC50 of cells treated with these drugs in the presence or absence of BSO was calculated (Table 2). Numbers in parentheses of Table 2 show the modification factors for each treatment in the cell lines. The modification factors for cisplatin-, vincristine- and H2O2-treated HOB1 cells were 1.3, 1.2 and 4, respectively. For HOB1/VCR cells they were 4, 2.5 and 4.4, respectively. These results suggest that GSH is likely to play a role in drug sensitivity. In addition, modification factors of cytotoxicity by BSO are greater in HOB1/VCR cells than in HOB1 cells, suggesting that GSH is likely involved in the acquisition of the MDR phenotype.

Figure 6.

Inhibition of glutathione by dl-buthionine-[S,R]-sulfoximine (BSO) in HOB1 and its multidrug resistant derivatives. Cells were treated with 25 mM BSO. The percentage of glutathione remaining is presented (mean ± SD; n = 3) relative to the glutathione content (nmol/mg protein) of mock-treated cells.

Table 2. Effect of glutathione blockage by dl-buthionine-[S,R]-sulfoximine (BSO) on cell sensitivity to various treatment
Cell lineCisplatin cytotoxicity IC50(µM)Vincristine cytotoxicity IC50(µM)H2O2 cytotoxicity IC50(mM)
  • Data are presented as mean ± SD of three measurements. More than 85% glutathione levels in cells were inhibited under BSO conditions (25 mM, 24 h).

  • Numbers in parentheses represent the modification factor by BSO, calculated as the IC50 of control cells divided by that of BSO treatment. Significantly different from the BSO group:

  • *

    P < 0.1,

  • **

    P < 0.05.

HOB10.15 ± 0.020.12 ± 0.02 (1.3)*0.07 ± 0.010.06 ± 0.01 (1.2)*0.07 ± 0.010.018 ± 0.002 (4)**
HOB1/VCR10.0 ± 2.0 2.5 ± 0.3 (4.0)**15.0 ± 2.0 6.0 ± 0.8 (2.5)** 1.5 ± 0.2 0.33 ± 0.05 (4.5)**


The present study has demonstrated that the concentration of ROS generated is associated with apoptotic sensitivity in anticancer drug-treated lymphoblastoma cells. This MDR lymphoblastoma is resistant to a range of anticancer drugs, including cisplatin (a DNA damaging agent) and vincristine (an actin-microtubule damaging agent).(8) Both cisplatin and vincristine trigger cell death by apoptosis – cisplatin acts primarily through the mitochondria-mediated caspase-9 pathway, whereas vincristine acts through the caspase-9 pathway(9) and by microtubule-associated mechanisms.(37,38) The steady-state concentration of ROS was decreased in the MDR cell line HOB1/VCR following exposure to cisplatin and vincristine. Attenuation of the ROS level by antioxidant NAC resensitized the cell response to these drugs. Experimental findings suggest that overexpression of a ROS-reducing system, GSH–GPx, has a role in cell resistance to these drugs because ROS is readily formed in mitochondria in response to anticancer drugs in cells that likely play a role in the response to cisplatin and vincristine.(39) However, less ROS was generated in cells in response to vincristine than to equitoxic concentrations of cisplatin. The likely contribution of microtubule-associated mechanisms to apoptosis in vincristine-treated cells(37,38) may partially explain the difference in levels of ROS produced. Furthermore, decreased mitochondria-mediated apoptosis by overexpression of antiapoptotic Bcl-2, inactivation of proapoptotic Bax,(40,41) and altered mitochondrial membrane potential was detected in MDR HOB1/VCR cells (data not shown). These experimental findings indicate that the steady-state level of ROS is an important determinant of cell response to chemotherapeutic agents, and identifying ways of attenuating the concentration of ROS in cells is crucial to the development of MDR in vivo.

The selected HOB1/VCR cells also exhibited resistance to H2O2, strongly suggesting that the steady-state level of ROS may have an important role in the MDR phenotype. Experimental results further indicated that low levels of ROS can generate apoptosis, whereas accumulation of high levels of ROS can induce necrosis or cause necrotic-like destruction of apoptosis-committed cells. This experimental finding verifies previous works demonstrating that oxidants such as H2O2 can induce apoptosis or necrosis, depending on the dose administered.(11,12,42) However, caspase activation in H2O2-treated cells was not detected in this study (Fig. 1). The dose of H2O2 administered (Fig. 1) may be too high to induce apoptosis, and therefore may have merely killed cells by necrosis. For treated cells with a low toxic concentration of H2O2 inducing the same ROS level as that with anticancer agents, only a small proportion of cells died by apoptosis. This finding may explain the difficulty in detecting activation of caspase in H2O2-treated cells. Nevertheless, the same levels of ROS induced by H2O2 are generally more potent in killing cells than those induced by cisplatin or vincristine. Additionally, equitoxic concentrations of cisplatin and vincristine generated different levels of ROS in lymphoma cells. Overexpression of P-glycoprotein and enhanced DNA repair are also considered important to this MDR lymphoblastoma.(8) Taken together, these results indicate that ROS may account for MDR.

Experimental results also indicate that overexpression of the H2O2-reducing system GSH–GPx may play a role in MDR. GSH has been characterized as a critical intra- and extracellular protective antioxidant, and is the most rapidly acting and abundant agent against ROS.(18,43) Oxidative stress is induced in cells through inactivation of GPx, a major peroxide-scavenging enzyme.(44) GSH and GPx convert toxic H2O2 to H2O at the expense of GSH. Thus, HOB1/VCR cells were resistant to H2O2 and to apoptotic agents, likely due to overexpression of GSH–GPx. The important role of GSH–GPx overexpression in the resistance is further supported by experimental data indicating that GSH inhibition by BSO markedly increased drug sensitivity in these cells. However, GSH does not exert its effects on cancer cells as an antioxidant alone, it also likely plays a role in drug resistance and contributes to exclusion of cytotoxic drugs out of cells. MRP can be considered a transporter for organic anions, also transporting glutathionyl and cysteinyl conjugates from the glutathione-S-transferase (GST) pathway,(45) and thus has a transport activity that likely depends on intracellular levels of reduced GSH. The MDR phenotype is likely related to MRP. However, the increased MRP levels were undetected in HOB/VCR cells (data not shown). Moreover, NAC is an antioxidant and a cysteine source for GSH synthesis, which plays a role in drug resistance. Despite its non-specificity, NAC has therapeutic value in reversing free radical-mediated diseases.(46) Several studies reported that NAC sensitizes drug resistance.(47,48) NAC also has a role in the elimination of drugs out of cells by spontaneously forming conjugates with drugs.(49) Nevertheless, studies of ROS scavenging by NAC in cells herein demonstrated the role of intracellular levels of ROS in mediating drug sensitivity. Consequently, the acquisition of decreased ROS levels in resistant cells may explain why apoptotically resistant cells are necrotically resistant. These experimental results also suggest that increased ROS levels may be a target for designing drugs against the MDR phenotype.


This study was supported by the National Science Council, ROC (grant NSC91-2320-B182-022) and by Chang Gung University (grants CMRP1025, CMRP32024).