Expression of NAD(P)H:quinone oxidoreductase 1 in HeLa cells: role of hydrogen peroxide and growth phase.

The aim of this work was to study the role of H(2)O(2) in the regulation of NAD(P)H:quinone oxidoreductase 1 (NQO1, DT-diaphorase, EC ) with relation to cell density of HeLa cells cultures and the function played by NQO1 in these cells. Levels of NQO1 activity were much higher (40-fold) in confluent HeLa cells than in sparse cells, the former cells being much more resistant to H(2)O(2). Addition of sublethal concentrations of H(2)O(2) (up to 24 microm) produced a significant increase of NQO1 (up to 16-fold at 12 microm) in sparse cells but had no effect in confluent cells. When cells reached confluency in the presence of pyruvate, a H(2)O(2) scavenger, NQO1 activity was decreased compared with cultures grown to confluency without pyruvate. Inhibition of quinone reductases by dicumarol substantially decreased viability of confluent cells in serum-free medium. This is the first demonstration that regulation of NQO1 expression by H(2)O(2) is dependent on the cell density in HeLa cells and that endogenous generation of H(2)O(2) participates in the increase of NQO1 activity as cell density is higher. This enzyme is required to promote survival of confluent cells.

NAD(P)H:(quinone acceptor) oxidoreductase (DT-diaphorase, EC 1.6.99.2) (1, 2) is a cytosolic flavoenzyme widely distributed and ubiquitously present in all the tissues of nearly all animal species (3,4). Among the various cytosolic NAD(P)H: (quinone acceptor) oxidoreductases described so far, the isoform 1 (NQO1) 1 is the best studied enzyme (4). Several properties of NQO1 make it an unique flavoenzyme (5). These include its nonspecific reactivity toward pyridine nucleotides substrates (NADH or NADPH) and broad electron acceptor specificity, its extreme sensitivity to inhibition by the anticoagulant dicumarol (6), and its obligatory two-electron reaction mechanism, which results in a direct hydride transfer to a variety of quinone substrates to give their corresponding hydroquinones (7,8).
NQO1 is generally regarded as a protective enzyme that has been shown to prevent the formation of highly reactive quinone metabolites, detoxify benzo(a)pyrene quinone, and reduce chromium (VI) toxicity (4,9). Recent reports have indicated that NQO1 activity maintains the reduced states of ubiquinones (10 -13) and ␣-tocopherolquinone (14), thereby promoting their antioxidant function in membranes. On the other hand, metabolism of a number of quinones by NQO1 (and other related reductases as well) produce unstable hydroquinones that can be readily autoxidized to generate reactive oxygen species (ROS) and can rearrange into bioalkylating compounds causing further damage to the cells (5,15,16). Thus, the actual antioxidant or prooxidant role of NQO1 in the cell depends upon the chemical nature and reactivity of the particular hydroquinone generated in the reactions it catalyzes (5,17). Very recently, this enzyme has received a renewed interest because of the demonstration of novel roles for NQO1, which include the regulation of the intracellular redox state by controlling the NAD(P)H:NAD(P) ϩ ratio (18) and the regulation of tumor necrosis factor-␣ and p53-mediated apoptotic cell death, most likely through to its ability to control the intracellular redox environment (17,19,20).
Published papers have documented changes in NQO1 expression with relation to the growth phase in cultured cells. In this way, an increase in NQO1 activity has been reported to occur at high densities in normal BALB/c 3T3 cells, and the proposal was made that this increase could be also associated to density-dependent inhibition of growth (21). NQO1 expression and activity are also significantly elevated in confluent cell cultures and spheroids of human colon carcinoma HT-29 cells (22). Although the demonstration that expression of NQO1 is elevated in confluent cells and inside multicellular spheroids has important implications in the area of bioreductive drug metabolism (22), environmental stimuli responsible for causing elevated NQO1 expression at high density have not been investigated. Furthermore, after the demonstration of NQO1 participation in the maintenance of intracellular redox balance controlling cell growth and apoptosis, the characterization of those factors involved in density-regulated expression of NQO1 may give new insights in the knowledge of the factors that regulate cell growth.
Expression of the NQO1 gene is positively or negatively regulated by a number of transcription factors (such as c-Jun, Jun-B, Jun-D, c-Fos, Fra1, Nrf1, and Nrf2) that bind to several cis-elements of the NQO1 gene promoter, including an antiox-idant response element that contains AP-1 and AP-1-like elements, a basal element, and AP-2 element (4,23). NQO1 expression is coordinately induced with other genes by 3-methylcholantrene, dioxin, trans-stilbene, phenobarbital, azo dyes, aromatic diamines, aminophenols, and phenolic antioxidants (4,24). A common feature of all phenolic antioxidants that transcriptionally activate gene expression via the antioxidant response element sequence is their ability to undergo redox cycling to form superoxide radicals and H 2 O 2 , which have been recognized as secondary messengers (25)(26)(27). Because these ROS are endogenously generated by many metabolic reactions and released constitutively by tumor cells (28 -30) and can activate NQO1 expression through the antioxidant response element (4,(31)(32)(33), this study was set to test the putative role of H 2 O 2 in the regulation of NQO1 expression with relation to density of HeLa cell cultures.
Our results have shown that NQO1 activity is considerably increased when HeLa cells reach high density, and this increase correlates with enhanced resistance of cells against H 2 O 2 . We report evidence that endogenous generation of H 2 O 2 contributes to the rise of NQO1 activity, and this enzyme is required to maintain cell viability at high cellular densities.

EXPERIMENTAL PROCEDURES
Cell Cultures-Cultures of HeLa cells were maintained at the laboratory in Dulbecco's MEM (Sigma) supplemented with 10% fetal calf serum (Flow Laboratories), 100 units/ml penicillin, 100 mg/ml streptomycin, and 2.5 mg/ml amphotericin B (Sigma) at 37°C in a humidified atmosphere of 5% CO 2 and 95% air. Prior to experiments using H 2 O 2 , the cells were changed to iron-free and pyruvate-free MEM (Sigma) (34). The culture medium was changed every 2 days until the cells reached the densities required for each experiment. The cells were grown to a density of about 8,000 viable cells/cm 2 (low density, sparse) or 100,000 cells/cm 2 (high density, confluent). In some experiments, 10 mM sodium pyruvate was added to the culture medium to study the effect of H 2 O 2 scavenging (35)(36)(37). Stock solutions of the NQO1 inhibitor dicumarol (2 mM) were prepared in 6 mM NaOH and added to cells to a final concentration of 20 M. The same amount of vehicle was added to controls. The viability of cells was estimated by the trypan blue exclusion assay after detaching cells from culture dishes using a nonenzymatic solution (Sigma).
Cell Proliferation Assay-Proliferation was measured from the ability of cells to incorporate thymidine. Briefly, the cells were incubated with 0.25 Ci/ml [methyl-3 H]thymidine for 8 -24 h. After incubation, the culture plates were put on ice, and the cells were washed with cold 0.9% NaCl. Cold trichloroacetic acid was added to a final concentration of 5%, the supernatants were discarded, and the cells were then lysed with 0.1 N NaOH. Lysates were used for measuring incorporation of radioactivity using a liquid scintillation counter (Beckman, Palo Alto, CA). Incorporation was referred to cell number to obtain specific values.
Preparation of Cytosolic Fractions-All procedures were carried out at 4°C. The cells were separated from culture dishes as described above, concentrated by centrifugation at 1,000 ϫ g for 5 min, and washed with cold 130 mM Tris-HCl, pH 7.6, containing 1 mM EDTA, 0.1 mM dithiothreitol, and 1 mM PMSF. The cells were centrifuged again and resuspended in 1 ml of hypotonic lysis buffer (10 mM Tris-HCl, pH 7.6, containing 1 mM EDTA, 0.1 mM dithiothreitol, 1 mM PMSF, and 20 g/l each of chymostatin, leupeptin, antipain, and pepstatin A). Homogenization of cells was carried out for 5 min with the aid of a glass-glass potter and then for 30 s with a mechanical cell homogenizer. After disruption of the cells, the concentration of the lysis buffer was raised to 100 mM Tris by adding enough volume of 250 mM Tris buffer, pH 7.6, containing 1 mM EDTA, 0.1 mM dithiothreitol, 1 mM PMSF, and chymostatin, leupeptin, antipain, and pepstatin A. Unbroken cells and debris were separated by centrifugation at 800 ϫ g for 5 min, and the supernatant was saved. Cytosolic fractions were separated from membranous material by ultracentrifugation at 100,000 ϫ g for 30 min.
NQO1 Activity Assay-NQO1 (DT-diaphorase) activity was measured in cytosolic fractions from the NADH and menadione-dependent dicumarol-inhibitable reduction of cytochrome c (6). Assays were carried at 37°C with constant gentle stirring. The assay mixture (1 ml) contained 70 g of cytosolic protein in 50 mM Tris-HCl (pH 7.5), 0.08% Triton X-100, 0.5 mM NADH, 10 M menadione, and 77 M cytochrome c. Assays were carried out either in the absence or in the presence of 10 M dicumarol, and absorbance was recorded at 550 nm in a Beckman DU-640 UV-visible spectrophotometer. NQO1 activity was calculated from the difference in reaction rates obtained with and without dicumarol. An extinction coefficient of 18.5 mM Ϫ1 cm Ϫ1 was used in calculations of specific activities (6). Protein determinations were carried out by the dye binding method described by Stoscheck (38).

RESULTS
Cell Cultures-HeLa cells were seeded at a density of 1,500 viable cells/cm 2 on 50-cm 2 culture dishes in MEM supplemented with 10% fetal calf serum. After 3 days growing, viable cells reached a density of about 8,000/cm 2 (sparse). The cells from separate plates were allowed to grow for an additional period of 3 days until they reached a density of about 100,000 viable cells/cm 2 (confluent) (Fig. 1A). Sparse cells were in exponential phase of growth ( Fig. 1), but the growth rate of confluent cells was significantly lower, as estimated from a substantial decrease in thymidine incorporation (Fig. 1B). All cultures used in our experiments exhibited similar viabilities of about 87%. Growing HeLa cells to densities above 200,000 cells/cm 2 resulted in a considerably decrease in viability (Fig.  1A). Thus, reaching these cellular concentrations was avoided.
Role of Cell Density on NQO1 Activity-Cytosolic fractions were obtained from sparse or confluent cells and used for assaying NQO1 activity. Very little NQO1 activity was detected in cytosols obtained from low density HeLa cells, but this activity showed a dramatic increase of up to 40-fold in confluent cultures (Fig. 2). To study the kinetics of NQO1 increase with relation to cell density, cytosols were obtained from HeLa cells cultured to different cell densities ranging between 8,000 and 100,000 viable cells/cm 2 . As shown in Fig. 3, a 10-15-fold increase in NQO1 activity was obtained in only 1 day of culture, with the density of cells increasing from about 8,000 to 15,000 cells/cm 2 . Thereafter, the stimulation of NQO1 continued until a maximal activation was achieved in confluent cultures (about 100,000 viable cells/cm 2 ). It is noteworthy that a significant increase of NQO1 (10 -15-fold) was observed well before cells reached confluence, and thus, different factors in addition to cell-to-cell contacts could account for the observed increment in NQO1 activity. A plausible interpretation is that some diffusible agent(s) could mediate the rise of NQO1 as the cell density increases.
Cell Density and Resistance against H 2 O 2 Cytotoxicity-Because HeLa cells generate and release H 2 O 2 constitutively during growth (28,30), we tested the possibility that H 2 O 2 could be involved in the cell density-related rise of NQO1 activity. As a first step in addressing the putative role of H 2 O 2 in density-regulated expression of NQO1, we checked the resistance of HeLa cells to addition of exogenous H 2 O 2 , both in sparse and confluent cultures.
The cells were treated with increasing concentrations of H 2 O 2 for 8 h, and then, the viability of cells was estimated from the trypan blue exclusion assay. As shown in Fig. 4 Fig. 4). Addition of H 2 O 2 to HeLa cells at low density produced a significant increase of NQO1 activity that was proportional to the concentration of H 2 O 2 used up to 12.5 M. Increasing the concentra-tion of H 2 O 2 up to 24 M did not result in a further increase of NQO1 activity, but a stabilization was observed (Fig. 5). When similar experiments were carried out with cells cultured to high density, the addition of H 2 O 2 to the culture medium did not result in a significant change of NQO1, and the activity remained elevated at levels identical to those found for confluent cells not treated with H 2 O 2 (Fig. 5).
Because H 2 O 2 stimulated NQO1 activity in sparse but not in confluent cultures, we tested the possibility that endogenous generation of H 2 O 2 was involved in the observed rise of NQO1 activity as the cell density increased. To test the putative role of endogenous H 2 O 2 , cells were allowed to grow to confluency in the presence of 10 mM pyruvate. Culturing HeLa cells in the presence of 10 mM pyruvate did not affect significantly their growth rate with respect to cultures grown in its absence (Fig.  3B). However, a much lower degree of NQO1 activation was observed in HeLa cells cultured in the presence of pyruvate (Fig. 3A). Whereas scavenging of endogenous H 2 O 2 by pyruvate resulted in a significant prevention of NQO1 increase, the activity was still stimulated about 13-fold with relation to low density cultures, which indicates that endogenous H 2 O 2 is not the sole factor involved in the rise of NQO1 but that additional factors related with cell-to-cell interactions and/or growth conditions at high cellular densities should be considered (21,22).
Protective Role of NQO1 at High Cellular Densities-HeLa cells showed density limitation of growth in such a way that cells exhibited a markedly reduced rate of proliferation at high density, measured as thymidine incorporation. Furthermore, substantial cell death was observed at cellular densities above 200,000 cells/cm 2 , even in the presence of fresh medium (Fig. 1).
To test whether increased NQO1 activity induced cell death or, instead, played a protective role, cultures of confluent HeLa cells were treated for 24 h with dicumarol at 20 M, and then the viability and proliferation rate of the cells was scored. This concentration of dicumarol, which fully inhibited NQO1 activity in in vitro enzymatic assays, was chosen to avoid inhibition of other related quinone reductases that are much more resistant to this compound (39). When the treatment was carried out in complete culture medium (containing 10% serum), no effect of dicumarol either on the viability or proliferation rate of cells was observed. However, treatment with the inhibitor in serumfree MEM resulted in a considerable decrease of the viability of cells and a reduction in their proliferation rate. Intermediate results were obtained when cells were treated with the inhibitor in the presence of 1% serum. The protective effect of serum reversing the effect of dicumarol treatment was likely due to serum albumin because treatment of cells with dicumarol in the presence of bovine serum albumin mimicked the protective effect of serum (Table I and Fig. 6). Thus, we can conclude that NQO1 plays a protective role, and this activity is necessary to maintain the viability of confluent HeLa cells in the absence of serum. DISCUSSION Since originally discovered by the group of Lars Ernster (1,2), much work has been focused on the characterization of NQO1 (DT-diaphorase) functions in quinone metabolism. Considerable progress has also been made in the knowledge of the regulation of basal and stimulated expression of this enzyme (4,5), its role in detoxification and bioactivation of xenobiotic compounds (4,5,15,16), and its role in the maintenance of antioxidant hydroquinones (10 -14). Recent studies with NQO1 knock-out mice have enabled us to demonstrate for the first time in vivo functions of NQO1 as an endogenous factor in protection against benzo(a)pyrene carcinogenicity (9) and in the control of intracellular redox state (18). The balance in the redox state is required to maintain the appropriate cell environment permissive for signaling (17,19). Accordingly, inhibition of NQO1 by dicumarol and treatment with high hydroquinone concentrations cause the blockade of stress-activated protein kinase/c-Jun NH 2 -terminal kinase and NFB pathways (19). Furthermore, NQO1 has been shown to play a regulatory role in several models of apoptotic cell death, most likely because of its ability to regulate the NAD(P) ϩ /NAD(P)H ratio (17,19,20). Taken together, these data support the idea that NQO1 plays important roles in the control of cell growth and death.
The aim of this work was to study the factors that regulate the expression of NQO1 in relation to cell density using HeLa cell cultures, an adenocarcinoma cell line growing as monolayers. NQO1 was dramatically increased by about 40-fold in confluent compared with sparse cultures, which is in contrast with results of Schlager et al. (21), who reported that an increase in NQO1 activity at high density occurs in normal BALB/c 3T3 cells but not in transformed cells derived from this line. However, our results agree with those of Phillips et al. (22), who showed that plateau cultures of the adenocarcinoma cell line HT-29 exhibited a modest increase in NQO1 activity of 2-3-fold. Thus, it seems clear that the increase in NQO1 observed at confluency is not restricted to normal cells but that it also occurs at least in these adenocarcinoma cell lines.
As shown for other proteins (40), the demonstration that expression of NQO1 is elevated in confluent adenocarcinoma cells is relevant to the area of bioreductive drug metabolism, because plateau phase cultures of tumor cells have been proven as valuable models that mimic many characteristics of the tumor microenvironment, such as reduced pH, poor nutrient status, low cell proliferation rates, and high catabolite concentrations (22,41). Accordingly, higher levels of NQO1 gene expression have been observed in many tumors when compared with normal tissues of the same origin (42)(43)(44). However, the demonstration of novel regulatory roles for NQO1 has opened new perspectives in the function played by this enzyme in confluent cells.
Because H 2 O 2 is constitutively generated and released by tumor cells (28,30) and this compound can directly activate NQO1 expression (4, 31-33), we investigated the putative role of H 2 O 2 in the regulation of NQO1 expression with relation to cell density. We first evaluated the sensitivity of HeLa cells to various concentrations of this oxidant. Low concentrations of H 2 O 2 had a slightly stimulatory effect, whereas higher concentrations produced a significant increase in the number of dead cells in sparse but not in confluent cultures. It has been firmly established that low levels of ROS constitute important growth regulatory signals in various cell lines (including HeLa cells), and exogenous administration of H 2 O 2 can elicit growth responses in these cells, but progression to a more prooxidant state results in increased cell death (28,30,45). Also, low levels of H 2 O 2 are able to act as a sort of "life signal" to maintain cell proliferation and to protect against apoptosis of U937 cells (46). Consistent with a role for the maintenance of certain levels of ROS in promoting cell growth, a decrease in steady-state levels of intracellular ROS has been related to density-dependent inhibition of cell growth in fibroblasts (47).
The increase in NQO1 activity at high density correlated with enhanced resistance to H 2 O 2 , thus suggesting that NQO1 may play a role in the protection of cells against H 2 O 2 cytotoxicity. This is apparently in contrast with results of Siemankowski et al. (17), who showed that transfectants of human breast adenocarcinoma-derived MCF-7 cells expressing different levels of NQO1 did not differ in their sensitivity to H 2 O 2 .
However, a putative increase of endogenous NQO1 activity in these cells at the different densities used in cytotoxicity experiments was not considered. Also, conditions used to evaluate cytotoxicity of H 2 O 2 in this and our study are inherently different. According to our results, the concentrations of H 2 O 2 used by these authors (300 -900 M) are probably too high to reveal differences between different transfectants. For instance, a correlation between catalase levels in bacteria and their sensitivity to H 2 O 2 could be only observed when the H 2 O 2 concentration was not excessive in relation to the amount of catalase (48). Also, because treatments were carried out in Dulbecco's MEM, which contains both pyruvate and iron, results may be difficult to interpret because of the simultaneous occurrence of H 2 O 2 scavenging by pyruvate and iron-catalyzed generation of hydroxyl radicals (34). In agreement with our results, Okuda et al. (49) have documented that neurons exhibiting NADPH-diaphorase activity are spared from toxicity because of 3-hydroxykynurenine, a neurotoxin causing H 2 O 2mediated cell death. Also, ubiquinol prevents adriamycininduced generation of H 2 O 2 in isolated hepatocytes (10,11), and this protection is abolished by the NQO1 inhibitor dicumarol (10). Finally, strains of Schizosaccharomyces pombe defective in ubiquinone biosynthesis display enhanced sensitivity to H 2 O 2 (50,51) and enrichment of hepatocytes with ubiquinol protects cells against H 2 O 2 (52). In sum, these data support the possibility that NQO1 may contribute to protect against H 2 O 2 via the generation and maintenance of hydroquinones.
Clearly, it is very likely that NQO1 is not the only enzyme responsible for the increased resistance of confluent HeLa cells against H 2 O 2 . Other antioxidant enzymes such as manganese superoxide dismutase, catalase, and glutathione peroxidase are elevated during confluency in several nonmalignant cell lines. Although elevation of these enzymes has not been observed in confluent cultures of mouse teratocarcinoma HR-9 cells (53), preliminary results obtained in our laboratory have shown that catalase activity is also significantly increased in confluent HeLa cells. 2 Increases of additional proteins, such as the small stress protein HSP27, which accumulates in confluent human colorectal cancer HT-29 and Caco2 cell lines and blocks cell death by decreasing ROS levels (40), could also occur in confluent HeLa cells.
Sublethal concentrations of exogenous H 2 O 2 produced a significant increase of NQO1 activity in sparse but not in confluent cells. This is the first demonstration that the regulation of NQO1 expression by H 2 O 2 is dependent on cell density in HeLa cells and strongly suggests that endogenous generation of H 2 O 2 TABLE I Effect of dicumarol on the viability of HeLa cells at high density HeLa cells cultured in MEM were supplemented with 10% serum until they reached confluence. The medium was replaced by serum-free MEM, and 24 h later, the cells were treated as listed for an additional period of 24 h. The cells were then detached from culture plates, and the viability was estimated by the Tripan blue exclusion test. The effect of dicumarol was estimated from the differences of viability between cells treated with inhibitor, and the cells were cultured under the same conditions but in its absence. The assays were carried out in duplicate. The data are the means Ϯ S.D. could be involved in density-mediated increase in NQO1 activity. Pyruvate is a well characterized H 2 O 2 scavenger that has been used to decrease intracellular levels of H 2 O 2 (35)(36)(37). Accordingly, the presence of pyruvate induced a substantial decrease in NQO1 stimulation as cells reached confluence, indicating that endogenous H 2 O 2 is a factor that mediates part of NQO1 increases, yet additional factors related with cell-tocell interactions and/or growth conditions should be also considered. This is in accordance with Phillips et al. (22), who proposed that a combination of factors, rather than a single triggering stimulus, may lead to elevated NQO1 mRNA levels in plateau phase cultures of adenocarcinoma HT-29 cells.
What is the function played by NQO1 in confluent cells? The increase of NQO1 activity in confluent fibroblasts has been related to density-dependent growth inhibition (21). Because a considerable decline in the ability of HeLa cells to proliferate was observed at high density, NQO1 might also play a role in growth control of HeLa cells. On the other hand, according to several recent reports, increased NQO1 activity could either potentiate or inhibit cell death. Although the inhibition of NQO1 by dicumarol strongly potentiates tumor necrosis factor-␣-induced apoptosis in HeLa cells (19), tumor necrosis factor-␣ sensitivity is also increased by overexpression of NQO1 in adenocarcinoma-derived MCF-7 cells (17). Furthermore, inhibition of NQO1 by dicumarol has been shown to decrease p53 stability in several cell lines and to increase cell viability of M1-t-p53 myeloid leukemic cells that overexpress a temperature-sensitive p53 transgene (20).
In HeLa cells, inhibition of NQO1 by dicumarol resulted in a significant decrease in viability and proliferation rate in serum-free medium, thus supporting the idea that NQO1 plays a protective role in confluent HeLa cells. However, no effect was obtained in the presence of 10% serum. Serum contains antioxidants that could contribute to protection against cell death induced by dicumarol, but a major protective effect could be attributable to albumin. Although albumin is a scavenger of ROS such as peroxide (54), it is reasonable to argue that a substantial part of protection is the result of the strong complexation of dicumarol by serum albumin. It has been reported that anticoagulants such as dicumarol and warfarin bind to albumin at high affinity sites, and this prevents uptake of these anticoagulants by cells (55)(56)(57). Complexation of dicumarol by albumin has not been taken into account in many studies set to prove the biological role of NQO1 in cultured cells. Furthermore, in many studies dicumarol concentration exceeds that required to inhibit the enzyme, without reference to whether serum was present during treatment (20). Reinvestigation of either the specific effect of dicumarol on NQO1 at these high concentrations (58) or the effective concentration that can enter the cells under experimental conditions should be therefore considered.
In summary, we have shown that regulation of NQO1 expression by H 2 O 2 is dependent on the cell density in HeLa cells. The endogenous H 2 O 2 participates in elevating the NQO1 activity as a function of increasing cell density. Our data showed that this enzyme is required to promote survival of confluent cells.