Abstract
Cdc25 dual-specificity phosphatases coordinate cell cycle progression and cellular signaling. Consequently, Cdc25 inhibitors represent potential anticancer agents. We evaluated >10,000 compounds for inhibition of human Cdc25 phosphatases and identified many potent and selective inhibitors, which all contained a quinone. Bioreductive enzymes frequently detoxify or activate quinones. Therefore, we evaluated the effect of NAD(P)H:quinone oxidoreductase-1 (NQO1) and reductase-rich microsomes on the activity of three quinone-containing Cdc25 inhibitors: 2-(2-hydroxyethylsulfanyl)-3-methyl-1,4-naphthoquinone (Cpd 5, compound 5; NSC 672121), 2,3-bis-(2-hydroxyethylsulfanyl)-1,4-naphthoquinone (NSC 95397), and 6-chloro-7-(2-morpholin-4-yl-ethylamino)quinoline-5,8-dione (NSC 663284). Each inhibitor was reduced by human NQO1 (Km of 0.3–0.5 μM) but none by microsomes. Compounds were evaluated with six cancer cell lines containing different amounts of NQO1: HT-29 (1056 nmol/mg/min), HCT116 (660 nmol/mg/min), sublines HCT116-R30A (28 nmol/mg/min) and HCT-116R30A/NQ5 (934 nmol/mg/min), MDA-MB-231/Q2 (null NQO1), and subline MDA-MB-231/Q6 (124 nmol/mg/min) but containing similar amounts of microsomal cytochrome P450 reductase and cytochrome b5 reductase. Growth inhibition and G2/M arrest by Cpd 5 was proportional to NQO1 levels, requiring 4- to 5-fold more Cpd 5 to inhibit HCT-116 or HCT-116R30A/NQ5 compared with HCT-116R30A. In contrast, in all tested cell lines irrespective of NQO1 level, growth inhibition and G2/M arrest by NSC 95375 and NSC 663284 were similar (average IC50 of 1.3 ± 0.3 and 2.6 ± 0.4 μM, respectively). NSC 95375 and NSC 663284 also caused similar Cdk1 hyperphosphorylation, indicating similar Cdc25 inhibition. However, lower Cpd 5 concentrations were needed to produce Cdk1 hyperphosphorylation in sublines with minimal NQO1. Thus, NQO1 detoxified Cpd 5, probably by reducing it to a less active hydroquinone, whereas NSC 95397- and NSC 663284-generated cytotoxicity was unaffected by NQO1.
Mammalian cells depend on protein kinases and phosphatases for communication and growth. Among the protein phosphatases, the dual-specificity Cdc25 phosphatases regulate cell cycle progression (Nilsson and Hoffmann, 2000) and may participate in mitogen activate protein kinase-mediated cell signaling (Vogt et al., 2001). The three human Cdc25 homologues have been assigned different roles in cell cycle progression (Sadhu et al., 1990; Galaktionov and Beach, 1991; Nagata et al., 1991). Cdc25C completes the G2/M transition by dephosphorylating and activating the mitotic kinase Cdk1/cyclin B, which is required for entry into mitosis (Millar et al., 1991). Cdc25A is important for the entry into S phase (Jinno et al., 1994) and may participate in mitosis (Donzelli et al., 2002). Cdc25B is essential for preinitiating G2/M transition and S phase progression (Lammer et al., 1998).
Tumor cells frequently overexpress Cdc25 phosphatases (Wu et al., 1998; Cangi et al., 2000), making them attractive candidate targets for anticancer agents. Earlier studies (Tamura et al., 2000; Wang et al., 2001) showed that 2-(2-hydroxyethylsulfanyl)-3-methyl-1,4-naphthoquinone (Cpd 5, NSC 672121), a vitamin K analog with arylating capability, inhibited Cdc25 and arrested cell cycle progression at G2/M and G1 phases. More recently, we evaluated >10,000 compounds in the National Cancer Institute chemical repository for their ability to inhibit recombinant human Cdc25B phosphatase in vitro. Several potent phosphatase inhibitors were identified with in vitro inhibitory concentrations <1 μM (Lazo et al., 2001, 2002). Many of these active compounds were shown to contain a quinone substructure. Besides Cpd 5, two other potent quinone-containing Cdc25 inhibitors, 2,3-bis-(2-hydroxyethylsulfanyl)-1,4-naphthoquinone (NSC 95397) and 6-chloro-7-(2-morpholin-4-yl-ethylamino)quinoline-5,8-dione (NSC 663284) (Fig. 1) were also evaluated extensively (Lazo et al., 2001, 2002).
Several clinically used antitumor drugs, such as mitomycin C and doxorubicin, are also quinones. They are known to be activated or inactivated by a 2e- reduction catalyzed by NAD(P)H:quinone oxidoreductase-1 (NQO1) or by a 1e- reduction catalyzed by NAD(P)H:cytochrome P450 reductase (P450 reductase) (Bachur et al., 1979; Pan et al., 1984; Siegel et al., 1990; Walton et al., 1991). In addition, cytochrome b5 reductase (Powis and Appel, 1980; Hodnick and Sartorelli, 1993) and carbonyl reductase (Wermuth et al., 1986) are capable of reducing quinones. NQO1 is frequently elevated in human tumors compared with the surrounding normal tissue (Schlager and Powis, 1990; Cresteil and Jaiswal, 1991; Malkinson et al., 1992; Jarrett et al., 1998; Mikami et al., 1998). Furthermore, the level of NQO1 in human tumors can be 5- to 25-fold greater than other reductive enzymes, including P450 reductase and cytochrome b5 reductase (Smitskamp-Wilms et al., 1995). Therefore, sensitivity of quinones to NQO1 reduction could markedly alter, limiting or enhancing, the utility of any new quinone anticancer agents. Because several newly found Cdc25 inhibitors have a quinone substructure, we examined the role NQO1 might have in determining their cellular action in tumor cells. Using cell lines containing different levels of NQO1 activity, we evaluated three potent quinone-containing Cdc25 inhibitors, NSC 672121 (Cpd 5), NSC 95397, and NSC 663284 (Fig. 1), for their inhibition of cell growth, arrest of cell cycle progression, and inhibition of Cdk1 dephosphorylation.
Materials and Methods
Reagents. Cpd 5 (NSC 672121) was synthesized as described previously (Tamura et al., 2000) (Fig. 1). NSC 95397 and NSC 663284 (Fig. 1) were provided by Dan Zaharevitz and Jill Johnson (Development Therapeutics Program of the National Cancer Institute, Bethesda, MD). Mouse monoclonal anti-Cdk1 antibody (SC 54) was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Rabbit polyclonal anti-phospho-Cdk1 (#9111) was purchased from New England BioLabs (Beverly, MA). Sheep anti-mouse IgG and donkey anti-rabbit IgG, both conjugated to horseradish peroxidase, recombinant human microsomal cytochrome b5, and other biochemicals were purchased from Sigma-Aldrich (St. Louis, MO). Renaissance kits were purchase from PerkinElmer Life Sciences (Boston, MA). Hybond-ECL nitrocellulose membranes were purchase from Amersham Biosciences Inc. (Patacasway, NJ).
Cell Lines. Six human tumor cell lines were selected to study the effect of Cdc25 inhibitors on cellular growth. These cell lines provide a broad spectrum of different NQO1 genotypes and levels of activity (Table 1). Thus, the three known human NQO1 genotypes are represented: 1) wild type (NQO1*1), 2) 609C>T single nucleotide polymorphism (SNP) (NQO1*2), and 3) 465C>T SNP (NQO1*3) (Ross et al., 2000; Pan et al., 2002). HCT-116R30A, a mitomycin C-resistant subline with minimum NQO1 activity, was derived from human colon cancer cell line HCT-116 (Pan et al., 1992). HCT-116R30A/NQ5 cells were generated from HCT-116R30A cells by a transfection of human NQO1*1 cDNA and selected for stable expression (Pan et al., 2002). HCT-116 and all of the sublines contain one or two alleles of NQO1*3. MDA-MB-231/NQ2 and MDA-MB-231/NQ6, sublines of human breast cancer cell line MDA-MB-231 with homozygous NQO1*2, contain stably transfected vector or a human NQO1*1 cDNA, respectively (gifts from Dr. David A. Boothman, Department of Radiation Oncology, Case Western Reserve University, Cleveland, OH). The human colon cancer cell line HT-29 contains two NQO1*1 alleles with high amounts of NQO1 activity. All colorectal cancer cell lines were maintained and subcultured in modified McCoy's 5A medium supplemented with 10% heat-inactivated fetal bovine serum, 2 mM glutamine, and 100 units/ml penicillin and 100 μg/ml streptomycin. The MDA-MB-231 sublines were maintained in RPMI 1640 medium containing the same supplements as McCoy's 5A. All cell lines were incubated at 37°C at 5% CO2 and 95% humidity.
Cytotoxicity of Cdc25 Inhibitors. The cytotoxicity of Cdc25 inhibitors was assessed by a modified colony formation assay described previously (Pan et al., 1992). Briefly, cells were plated in six-well plates at 500 to 1000 cells per well in 3 ml of medium. After overnight incubation, designated concentrations of compounds were added to the wells, and an equal volume of vehicle (ethanol) was added to the control well. After further incubation for 5 days to allow colony formation, medium was removed. Wells were washed with phosphate-buffered saline (PBS) and stained with 0.25% crystal violet in 10% formalin and 90% methanol. Excess stain was removed and the wells were washed with water and dried. Colonies (>50 cells) were counted. The drug concentration producing 50% inhibition of cell growth for each cell line was calculated with a Sigmoid-Emax model implemented in the ADAPT II computer program (D'Argenio and Schumitzky, 1979).
Isolation and Assays of NQO1. Recombinant human NQO1 was prepared and purified as described previously (Pan et al., 1995) and used to determine the kinetics of reduction of Cdc25 inhibitors. Purified NQO1 had a specific activity of 598 μmol/mg/min in reducing menadione-cytochrome c at pH 7.5 at room temperature. Due to interfering absorbance peaks of Cdc25 inhibitors and to allow for comparison with menadione, we adapted a method described to measure reduction of menadione by NQO1 (Ernster, 1967). Cdc25 inhibitors were used as intermediate electron acceptors, and cytochrome c was the terminal electron acceptor. The reaction mixture of 1 ml contained 50 mM Tris-HCl buffer at pH 7.5, 200 μM NADH, 80 μM cytochrome c, 5 μM FAD, 0.07% bovine serum albumin, various concentrations of Cdc25 inhibitors, and 100 ng NQO1. The rate of reduction of cytochrome c was measured at room temperature by monitoring the increase of absorbance at 550 nm for 2 min. Absorbance was measured with a Beckman Coulter DU-640 spectrophotometer, and each assay was performed in triplicate. Each assay was also repeated in the presence of 10 μM dicoumarol, a NQO1 inhibitor. Enzymatic activity attributed to NQO1 was that inhibited by dicoumarol. We also assayed the reduction of menadione, the commonly used substrate of NQO1 containing the core naphthoquinone structure. The best curve fit for Lineweaver-Burk plots, and values of Km and Vmax were determined using SigmaPlot 8.0 (SPSS Science, Chicago, IL). NQO1 activity in the cytosol was determined by the same assay, whereas 20 μl of cytosol was used in place of purified NQO1. Cytosol of each cell line was prepared as described previously (Pan et al., 1992).
Isolation of Microsomes and Assay for Cytochrome P450 Reductase (P450 reductase) and Cytochrome b Reductase (cyto-b Reductase). Microsomes, rich in P450 reductase and cyto-b reductase, were isolated from each cell line according to a method described previously (Pan et al., 1992). Freshly prepared microsomes were used to assay the activity of P450 reductase and cytochrome b reductase. All reactions were carried out at room temperature. Each enzymatic activity was determined with microsomes from three independent preparations, and each determination was carried out in triplicate.
P450 reductase activity was determined by cytochrome c reduction using NADPH as the electron donor. The reaction mixture of 1 ml contained 100 mM phosphate buffer at pH 7.5, 400 μM NADPH, 80 μM cytochrome c, 20 μM KCN (to inhibit cytochrome c oxidase), and 100 μg of microsome. Reduction of cytochrome c was monitored at 550 nm for 2 min. To analyze the reduction of Cdc25 inhibitors by microsomes, the reaction mixture was identical except that NADPH or NADH concentration was decreased to 150 μM, and cytochrome c was replaced with various concentrations of Cdc25 inhibitors. The rate of reducing Cdc25 inhibitors was monitored by the oxidation of NADPH or NADH as reflected by the decrease of absorbance at 340 nm for 10 min.
The activity of cyto-b reductase was determined in a 1-ml reaction mixture containing 10 mM phosphate buffer at pH 6.6, 1 mM NADH, 500 μM ferricyanide, and 100 μg of microsomes as described previously (Hodnick and Sartorelli, 1993). Reduction of ferricyanide was monitored at 420 nm. For confirmation, cyto-b reductase activity was also determined in a 100-μl reaction mixture containing 10 mM phosphate buffer at pH 6.6, 100 μM NADH, 2 μM cytochrome b5, and 10 μg of microsome. Reduction of cytochrome b5 was monitored at 427 nm for 10 min as described previously (Yubisui and Takeshita, 1982; Holmans et al., 1994). Purified rat P450 reductase was analyzed for its ability to reduce ferricyanide in the same manner with NADPH or NADH as an electron donor.
Flow Cytometric Analysis. The analysis of cell cycle by flow cytometry was conducted with asynchronous cells. All cell lines were plated at 4 × 106 cells in a 75-cm2 flask containing 10 ml of medium. After cells were allowed to adhere to the monolayer by overnight incubation at 37°C, either the compounds or vehicle was added to cells, which were incubated for an additional 12 h. Treated cells were harvested by trypsin digestion, washed twice with cold PBS, suspended in cold 70% ethanol, and incubated at –20°C for 30 min. Cells were washed again with PBS containing 1% bovine serum albumin before being stained at 5 × 105 cells/ml with a PBS solution containing 50 μg/ml propidium iodide and 250 μg/ml RNase A. After incubating overnight at 4°C, cells were analyzed with a Beckman Coulter Epics XL flow cytometer (Beckman Coulter, Fullerton, CA). Each Cdc25 inhibitor was analyzed at least three independent times. Cells treated with the vehicle, 0.5% dimethyl sulfoxide, served as a negative control.
Preparations of Cell Lysates for the Analysis of Cdk1. HCT-116, HCT116R30A, MDA-MB-231/Q6, and MDA-MB-231/Q2 cells were plated at 4 × 106 cells in a 75-cm2 flask (10 ml of medium) and incubated overnight at 37°C to allow cells to adhere. Compounds at 10 and 20 μM or vehicle (ethanol) were added to cells, which were incubated for an additional 6 h, medium was removed, and cells were washed twice with cold PBS. Drug-treated cells were lysed with 1.5 ml of lysis buffer (50 mM Tris-HCl at pH 8.0, 150 mM NaCl, 5 mM EDTA, 1% NT-40, 0.1% sodium deoxycholate, and protease inhibitors, including 100 μg of phenylmethylsulfonyl fluoride, 100 μg of benzamide hydrochloride, and 2 μg/ml each of aprotinin, leupeptin, antipain, and pepstatin). Protein concentrations were determined by the Bradford method (Bradford, 1976).
Western Blotting. Proteins (40 μg) were separated by SDS-polyacrylamide gel and transfer from the gel to Hybond-ECL nitrocellulose membranes as described previously (Pan et al., 1995). Total Cdk1 was detected by anti-Cdk1 at 1:100 dilution, and phosphorylated Cdk1 was detected by anti-phospho-Cdk1 at 1:1000 dilution. Anti-mouse IgG and anti-rabbit IgG conjugated to horseradish peroxidase at 1:2000 dilution were used as secondary antibodies, respectively. All incubations were performed at room temperature. Membranes were exposed to primary antibodies for 2 h, blocked with a PBS solution containing 5% nonfat milk and 1% Tween 20 for 1 h, and washed with PBS until free of milk. Then membranes were exposed to secondary antibody for 1 h and rinsed two times with PBS. Protein bands in membranes were detected by chemiluminescence produced by the oxidation of luminol, using a Renaissance kit, and captured by exposing blots to Kodak X-OMAT Blue film. Loading equivalence was monitored by the detection of β-actin bands on the same membrane.
Results
Activities of Reductases in Cell Fractions of Different Cell Lines. Cytosolic NQO1 activity in six cell lines showed a broad spectrum of level of NQO1 (Table 1). The difference between cell lines with high and low activity was as much as 60-fold. However, the activities of microsomal P450 reductase and cyto-b reductase activity among the cell lines did not vary significantly (Table 1). To address the possibility that other NADH-using reductases might contribute to the reduction of ferricyanide, which was used for measuring the activity of cyto-b reductase, we analyzed purified P450 reductase for its ability to reduce ferricyanide. We found that purified P450 reductase with activity equivalent to microsomes of HCT-116 reduced only 0.05 μM ferricyanide using NADPH or NADH as electron donor. Therefore, P450 reductase in microsomes seemed to contribute only minimally to the reduction of ferricyanide. The reduction of cytochrome b5 by microsomes also showed that cyto-b reductase activity among the tested cell lines did not vary significantly. Thus, the ferricyanide-reductase activity shown by microsomes (Table 1), which represents total one-electron reduction by microsomes, seems to be mainly a reflection of cyto-b reductase activity.
Reduction of Cdc25 Inhibitors by Recombinant Human NQO1 and Microsomes. All Cdc25 inhibitors shared similar Km values for reduction by purified recombinant human NQO1 (Table 2). Menadione, which is a substrate commonly used to measure the activity of NQO1 had a significantly higher Km. Cpd 5 exhibited a 2- and 6-fold higher maximum rate of reduction than did NSC 95397 and NSC 663284, respectively. The Vmax values of Cpd 5 and menadione for human NQO1 were similar. Oxidation of NADPH as well as NADH could not be detected after incubating microsomes for 10 min with Cdc25 inhibitors. Therefore, reduction of the three Cdc25 inhibitors by microsomes was minimal under conditions used in this study.
Relationship between NQO1 Activity and Cytotoxicity of Cdc25 Inhibitors. The three Cdc25 inhibitors caused a concentration-dependent cytotoxicity in all six cell lines tested with a colony formation assay (Fig. 2). Comparing the isogenic cell groups, we found HCT-116 and MDA-MB-231/NQ6, which had higher NQO1 activity, were more resistant to Cpd 5 than were HCT-116R30A and MDA-MB-231/NQ2 cells (Fig. 2A). HCT-116 cells were approximately 5-fold more resistant compared with HCT-116R30A cells and MDA-MB-231/NQ6 cells were 2-fold more resistant compared with MDA-MB-231/Q2 cells (Table 3). Restoration of NQO1 activity in HCT-116R30A/NQ5 cells regenerated resistance to Cpd 5 as reflected by an IC50 value similar to that seen in the parental HCT-116 cells. Among the six cell lines, HT-29 had the highest NQO1 activity and was the most resistant toward Cpd 5. In contrast, all cell lines irrespective of NQO1 content were equally sensitive to NSC 95397 and also equally sensitive to NSC 663284 (Fig. 2, C and D). The IC50 values of NSC 95397 and NSC 663284 for all six cell lines did not show significant differences, exhibiting average values of 1 to 1.8 μM and 2 to 3 μM, respectively.
Effect of Dicoumarol on IC50 of Cpd 5. Treatment with 10 μM dicoumarol, a potent inhibitor of NQO1, enhanced the sensitivity of HT-29 and HCT-116 cells to Cpd 5 (Fig. 2B), decreasing their IC50 values from 6.7 ± 0.3 and 2.3 ± 0.3 to 0.5 ± 0.1 and less than 0.5, respectively. Treatment with dicoumarol did not alter the IC50 values of Cpd 5 for HCT-116R30A and MDA-MB-231/Q2 cells (data not shown), consistent with the hypothesis that NQO1 inactivates Cpd 5 in cells. In addition, dicoumarol treatment did not alter the IC50 values of NSC 95397 and NSC 663284 (data not shown), consistent with the hypothesis that these two compounds were not affected by NQO1.
Cell Cycle Progression at G2/M. Because Cdc25 has an essential role at the G2/M interface, mitotic arrest is a valuable beacon of Cdc25 inhibition within cells. HCT-116 and HCT-116R30A cells were used to analyze the effect of NQO1 activity on the ability of Cdc25 inhibitors to cause cell cycle arrest. All three Cdc25 inhibitors blocked cell cycle progression of the two cell lines at the G2/M interface (Fig. 3). As anticipated by the cytotoxicity data, HCT-116R30A cells were more sensitive to Cpd 5 than HCT-116 cells. Cpd 5 at 2.5 μM clearly blocked HCT-116R30A cells at G2/M, whereas only slightly affecting HCT-116 cells (Fig. 3, A and B). Cpd 5 at 5.0 μM caused 92% of HCT-116R30A cells to arrest in G2/M phase, but only 32% of HCT-116 cells. Too few HCT-116R30A cells survived exposure to 7.5 μM Cpd 5 to permit an accurate analysis, whereas 34% HCT-116 cells accumulated in G2/M phase without prominent cell loss. HCT-116 and HCT-116R30A cells showed similar sensitivity to G2/M block by NSC 95397 and NSC 663284 (Fig. 3, C–F). More than 40% of the total cells were arrested at G2/M with either 7.5 μM NSC 95397 or NSC 663284.
Inhibition of Cdk1 Dephosphorylation. Dephosphorylation of the mitotic inhibitor Cdk1 by Cdc25 B and C is obligatory for entry into mitosis (Sadhu et al., 1990; Nilsson and Hoffmann, 2000). Thus, inhibition of Cdc25 phosphatases should prevent the dephosphorylation of Cdk1. Consequently, we compared the effect of the three quinoid Cdc25 inhibitors on the phosphorylation status of Cdk1 in the isogenic HCT-116 pair. Exposure of HCT-116R30A cells to Cpd 5 resulted in a pronounced increase of phosphorylated Cdk1 but only a modest change in Cdk1 phosphorylation status in HCT-116 cells (Fig. 4A). In contrast, Cdk1 phosphorylation was markedly increased in both HCT-116 and HCT-116R30A cells after either NSC 95397 or NSC 663284 exposure (Fig. 4, B and C). A similar inhibition of dephosphorylation of Cdk1 was seen in MDA-MB-231/Q2 and MDA-MB-231/Q6 cells after treatment by the three inhibitors (data not shown).
Discussion
Using relatively small, publicly available compound libraries, we have previously identified the para-naphthoquinone and quinolinedione scaffolds as promising lead structures for the design of inhibitors of dual-specificity Cdc25 phosphatases (Lazo et al., 2001, 2002). Both electrostatic and steric issues are important for potent and selective inhibition of Cdc25. However, the quinone substructure can be reduced by various enzymes and such metabolism might enhance or reduce the activity of a compound.
In this report, we have evaluated the ability of microsomes and an important reductive enzyme, NQO1, to affect the cytotoxicity and growth arrest of three potent, quinone-containing Cdc25 inhibitors. Microsomes are rich in both P450 reductase and cyto-b reductase that catalyze one-electron reduction of quinones to semiquinones. The failure to detect consumption of NADPH or NADH when microsomes were incubated with Cdc25 inhibitors strongly suggests that reduction of Cpd 5, NSC 95397, and NSC 663284 by microsomes isolated from human tumor cell lines was minimal. The amount of these microsomal reductases in the tested cell lines did not vary significantly, although the cells exhibited different sensitivity to the cytotoxic effects of the compounds. Thus, the cytotoxicity of the three inhibitors cannot be easily explained by P450 reductase or cyto-b reductase. In contrast, recombinant human NQO1 was capable of reducing all three Cdc25 inhibitors, although with some differences in the rate of their reduction. The fact that the Km values of these inhibitors are in the range of 0.3 to 0.5 μM indicates that the affinity of these compounds toward the recombinant NQO1 is similar, but much higher than that of the commonly used substrate and vitamin K analog menadione. Earlier publications (Tamura et al., 2000; Lazo et al., 2001, 2002) showed that in vitro Ki values for Cdc25A, Cdc25B, and Cdc25C were 15, 1.7, and 1.3 μM for Cpd 5, 32, 96, and 40 nM for NSC 95397, and 29, 95, and 89 nM for NSC 663284, respectively. The Km values for NQO1 are about 100- and 50-fold higher than the Ki values for Cdc25 of NSC 95397 and NSC 663284, respectively. This difference may influence the effect of these compounds on cells.
Indeed, our studies of Cdc25 inhibitors in selected human tumor cells showed that the effect of NQO1 on the activity of Cpd 5 differed from its effects on the activity of NSC 95397 and NSC 663284. The presence of NQO1 in each cell line had a minimal effect on the inhibition of cell growth by NSC 95397 and NSC 663284. On the other hand, cell lines with more NQO1 activity, such as HT-29 and HCT-116, were more resistant to Cpd 5 than cell lines with minimal NQO1 activity. The involvement of NQO1 in the cytotoxicity of Cpd 5 was confirmed by two additional experimental approaches. First, restoration of NQO1 activity increased the resistance for the HCT-116R30A/NQ5 cells to Cpd 5. Second, inhibition of NQO1 by dicoumarol rendered HT-29 and HCT-116 cells more susceptible to Cpd 5. Two other analyses were conducted to study the effect of cellular NQO1 on the function of Cdc25 inhibitors. Arrest of cell cycle progression at G2/M and inhibition of Cdk1 dephosphorylation, indicative of inhibition of Cdc25B and Cdc25C (Lammer et al., 1998; Nilsson and Hoffmann, 2000; Tamura et al., 2000), were observed for cells treated with Cpd 5, NSC 95397, and NSC 663284. Again, the activities of NSC 95397 and NSC 663284 were unaffected by the presence of NQO1 in cells, whereas the activity of Cpd 5 was decreased by the presence of NQO1. Reduction of Cpd 5 by cellular NQO1 decreases its activity against Cdc25 phosphatase and its cytotoxicity, indicating a detoxification process of Cpd 5.
The loss of Cpd 5 function may be due to its reduction to the hydroquinone. Cpd 5 is an analog of vitamin K with arylating capability. Among several vitamin K analogs, only those with arylating capability inhibited the action of Cdc25 (Wang et al., 2001), and as a consequence of the reduction by NQO1, Cpd 5 hydroquinone has lost its electrophilic properties. The chemical structure of NSC 95397 differs from that of Cpd 5 by one additional thioethanol residue, whereas NSC 663284 differs from Cpd 5 by having a quinolinedione rather than a naphthoquinone substructure. Both NSC 95397 and NSC 663284 were reduced by NQO1 at a slower rate than Cpd 5. However, the Vmax and Km values of both compounds observed are not different enough from Cpd 5 to account for the difference in NQO1 independence for the two compounds in tumor cells. Apparently, the reduction of these two compounds in cells, if any, does not affect their activity in inhibiting Cdc25 phosphatases. One possible factor attributed to the difference between Cpd 5 and the other two compounds may be the Ki values mentioned above. The much lower Ki values of NSC 95397 and NSC 663284, compared with Cpd 5, for Cdc25 phosphotases probably make these two compounds more effective inhibitors in cells, which is independent of NQO1. Another factor may be redox cycling. After reduction by NQO1 or other reductases, these three compounds may be capable of undergoing redox cycling at a different rate in the presence of oxygen to cause cytotoxicity. Cell lines with less NQO1 are more susceptible to Cpd 5, which may be an indication that reduced Cpd 5 is not capable of being recycled and becomes less toxic. All tested cell lines being equally sensitive to NSC 95397 and NSC 663284 could mean that the inhibitory function of these two compounds is not affected by the rate of redox cycling. In addition to NQO1 and microsomes, mitochondrial reductases also could have a role in the toxicity of these Cdc25 inhibitors.
Currently, our study has demonstrated that reduction of NSC 95397 and NSC 663284 by NQO1 had little effect on their inhibitory functions. NSC 95397 and NSC 663284 are not efficiently reduced by microsomes. However, the full extent of cellular metabolism of these compounds requires additional studies. It is unclear whether these compounds can be substrates for other quinone reductases or dehydrogenase, such as carbonyl reductase and xanthine dehydrogenase. Hypoxia may also play a role in the metabolism of these compounds. In addition, the effect of phase II metabolism, such as glucuronidation of reduced products of quinones, could occur in vivo. Studies are continuing to address these issues.
Acknowledgments
We thank Dr. Merrill J. Egorin for critically reviewing the manuscript, and Ezekiel Woods for excellent secretarial assistance in the preparation of this manuscript.
Footnotes
-
This study was supported in part by a U.S. Public Health Service National Institutes of Health Grant CA 61862 (to S.P.).
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
DOI: 10.1124/jpet.103.059477.
-
ABBREVIATIONS: NQO1, NAD(P)H:quinone oxidoreductase-1; NQO1*1, wild type NQO1; SNP, single nucleotide polymorphism; NQO1*2, 609C>T SNP; NQO1*3, 465C>T SNP; P450, cytochrome P450; PBS, phosphate-buffered saline; cyto-b reductase, cytochrome b5 reductase.
- Received September 2, 2003.
- Accepted December 15, 2003.
- The American Society for Pharmacology and Experimental Therapeutics