Scavenging of Extracellular H2O2 by Catalase Inhibits the Proliferation of HER-2/Neu-transformed Rat-1 Fibroblasts through the Induction of a Stress Response*

High levels of reactive oxygen species (ROS) are associated with cytotoxicity. Alternatively, nontoxic levels of ROS like hydrogen peroxide (H2O2) can mediate the transmission of many intracellular signals, including those involved in growth and transformation. To identify pathways downstream of endogenous cellular H2O2 production, the response of Rat-1 fibroblasts exhibiting differential HER-2/Neu receptor tyrosine kinase activity to removal of physiological H2O2 concentrations was investigated. The proliferation of all cells was abolished by addition of the H2O2 scavenger catalase to the culture medium. HER-2/Neu activity was not significantly affected by catalase treatment, suggesting that the target(s) of the H2O2 signal lie downstream of the receptor in our model. ERK1/2 phosphorylation was blocked by catalase in fibroblasts expressing wild type Neu, however such a response did not occur in cells possessing activated mutant Neu. This indicates that the ERK1/2 response contributes little to the growth inhibition observed. By contrast, JNK1 activity increased following the addition of catalase or H2O2, regardless of Neu activity or level of cell transformation. Phosphorylation of p38 MAPK was induced by H2O2 but not by catalase. These observations suggest that scavenging of H2O2 from the cellular environment blocks Rat-1 proliferation primarily through the activation of stress pathways.

High levels of reactive oxygen species (ROS) are associated with cytotoxicity. Alternatively, nontoxic levels of ROS like hydrogen peroxide (H 2 O 2 ) can mediate the transmission of many intracellular signals, including those involved in growth and transformation. To identify pathways downstream of endogenous cellular H 2 O 2 production, the response of Rat-1 fibroblasts exhibiting differential HER-2/Neu receptor tyrosine kinase activity to removal of physiological H 2 O 2 concentrations was investigated. The proliferation of all cells was abolished by addition of the H 2 O 2 scavenger catalase to the culture medium. HER-2/Neu activity was not significantly affected by catalase treatment, suggesting that the target(s) of the H 2 O 2 signal lie downstream of the receptor in our model. ERK1/2 phosphorylation was blocked by catalase in fibroblasts expressing wild type Neu, however such a response did not occur in cells possessing activated mutant Neu. This indicates that the ERK1/2 response contributes little to the growth inhibition observed. By contrast, JNK1 activity increased following the addition of catalase or H 2 O 2 , regardless of Neu activity or level of cell transformation. Phosphorylation of p38 MAPK was induced by H 2 O 2 but not by catalase. These observations suggest that scavenging of H 2 O 2 from the cellular environment blocks Rat-1 proliferation primarily through the activation of stress pathways.
Oxidative stress is known to induce the cellular apoptotic program by the activation of defined ROS 1 -responsive signals: the stress-activated protein kinase (SAPK) and nuclear factor kappa B (NF-B) pathways ((1), for review see Ref. 2). Activation of c-Jun/AP-1 transcription factors by the c-Jun N-terminal kinase (JNK) members of the SAPKs has been shown to induce apoptosis in several cell types (3). This signal is in contrast to the other characterized roles of c-Jun/AP-1 activity in survival and growth. For example, Jun proteins associate with Nrf 1 and 2 transcription factors to up-regulate the antioxidant/electrophile response element-mediated gene expression as a defense against oxidative stress (4). Similarly, NF-B was originally identified as a mediator of tumor necrosis factor ␣and ROS-induced apoptosis but more recently has also been characterized as part of a survival response by stimulating antioxidant response element-regulated and anti-apoptotic protein expression (5,6). Indeed, ROS, like H 2 O 2 , are associated with the regulation of multiple cellular processes via interaction with proteins and lipid species (7). These observations have led to the idea that ROS are required "cofactors" in the regulation of many intracellular signal transduction cascades. Addition of micromolar concentrations of H 2 O 2 to cells in vitro can activate plasma membrane receptor tyrosine kinases, including the epidermal growth factor receptor/ErbB-1 (8), platelet-derived growth factor receptor (9), and insulin receptor (10). The p21 Ras-MAPK (mitogen-activated protein kinase) pathway, a key growth signal linking membrane receptors to the nucleus, is also stimulated by oxidative stimuli such as H 2 O 2 (11,12). Rapid induction of cellular H 2 O 2 generation has been observed to follow the addition of a variety of peptide growth factors to cells, suggesting an autocrine and/or paracrine role for ROS in growth signaling (13)(14)(15). The specific molecular targets that are critical for H 2 O 2 -mediated mitogenesis have yet to be determined; the importance of a given pathway is also dependent on factors such as cell type, environment, and level of differentiation.
It follows that the deregulation of ROS levels may be a key factor in neoplasia. Elevations in O 2 . and H 2 O 2 are seen in some human tumor cell lines (16), and their production from mitochondrial and plasma membrane sources appears to be a requirement of tumor viability and growth (17,18). Mitochondria-localized manganese-containing superoxide dismutase is proposed to act as a both an inhibitor (19) and activator (20) of growth and transformation. A higher oxidative capacity of cancer cells may also accompany an increase in resistance to ROS stress, which could in turn affect drug resistance. For instance, resistance to H 2 O 2 stress by the acquired up-regulation of catalase activity and glutathione levels was shown to be crucial for maintenance of cell viability as well as resistance to cisplatin treatment (21). In theory, the oxidative stress response may be shifted from growth arrest and death to proliferation and transformation for cells exhibiting defects in growth pathway regulation and the apoptotic program (22). If the H 2 O 2 signal is involved in these processes, they should be inhibited by its removal from the cellular environment.
In the present study the requirement of low levels of H 2 O 2 for Rat-1 fibroblast proliferation was investigated. The effects on defined growth-and stress-associated signaling pathways, produced by the scavenging of physiological H 2 O 2 levels from the extracellular environment, were assessed. Rat-1 clones exhibiting differential growth and transformation properties resulting from specific alterations of the HER-2/Neu receptor were utilized as a model for the responses to removal of the H 2 O 2 stimulus (23,24). Multiple signals downstream of Neu are implicated in oncogenesis, and Neu overexpression and/or activating mutation is observed in 25-30% of human breast and ovarian cancers (for review see Ref. 25). The activities of the Neu receptor, extracellular signal-regulated kinase 1 and 2 (ERK1/2) MAPK and JNK1 effector pathways were studied after the addition of the H 2 O 2 scavenger catalase. All Rat-1 clones displayed comparable sensitivities to growth inhibition by this treatment regardless of the level of Neu activity. It was found that scavenging of H 2 O 2 caused stable inhibition of the ERK1/2 signal, transient induction of the JNK1 pathway, and no effect on p38 MAPK. Constitutive Neu activation could rescue the catalase-mediated block of ERK1/2 activity while only slightly increasing resistance to growth inhibition. Therefore, these results indicate that removal of extracellular H 2 O 2 can both down-regulate growth signals and activate stressassociated signals, however, the stress response appears to play a larger role in the anti-proliferative effects observed.

EXPERIMENTAL PROCEDURES
Cell Culture and Treatments-All Rat-1 fibroblast clonal cell lines were kindly provided by Dr. W. J. Muller and have been previously characterized (23,24). Rat-1NeuN (17) clone expresses wild type Neu. The highly transformed Rat-1Neu8142(10) clone expresses Neu having a 12-amino acid (aa 641-652) deletion in the extracellular region of the receptor proximal to the transmembrane domain, resulting in its constitutive dimerization. Rat-1NeuNT (3) is another highly transformed clone expressing Neu with an activating point mutation (V664E) in the transmembrane domain (26). Rat-1NeuNT-NYPD (6) is a transformation-impaired clone expressing the NeuNT receptor with C-terminal autophosphorylation sites (Tyr-1028, Tyr-1144, Tyr-1201, Tyr-1226/7, Tyr-1253) replaced with phenylalanine residues. Cells were maintained at early passage number in 100-mm diameter tissue culture plates (Falcon, Becton Dickinson) at 37°C in a humidified atmosphere of 5% CO 2 , 95% air. Rat-1 lines were grown in Dulbecco's modified Eagle's medium with L-glutamine (Life Technologies, Inc.), and all media were supplemented with 10% heat-inactivated fetal calf serum (Life Technologies, Inc.), except in serum-starved controls. Cells were passaged by trypsinization (1ϫ trypsin-EDTA; Life Technologies, Inc. Cell Proliferation and Active DNA Synthesis Assays-The cell growth rate was assessed by measurement of the total cell number after 4 days of culture. 1000 cells/well were seeded onto a 96-well plate (Falcon, Becton Dickinson) and incubated until adherent. After incubation for 4 days, cells were washed with H 2 O and total cellular DNA was quantitated by the addition of 200 l/well of 10 g/ml Hoechst stain No. 33258 (bisbenzimide H 33258 fluorochrome, Calbiochem-Novabiochem) in TNE buffer (5 mM Tris base, 0.5 mM EDTA, 1 M NaCl, pH 7.4). Fluorescence in wells (excitation , 350 nm; emission , 460 nm) was measured using a CytoFluor series 4000 multiwell plate reader (PerSeptive Biosystems). The cell number was standardized to fluorescence for each cell type by comparison with a standard curve generated from the seeding of known cell numbers.
Quantitation of cellular DNA synthesis was performed on cells by the addition of 10 M 5-bromo-2Ј-deoxyuridine (BrdUrd, Sigma) for 30 min. Cells were then trypsinized and washed 2ϫ with phosphate-buffered saline. 100 l of each cell suspension was added to 2 ml of cold 70% ethanol with vortexing and left on ice for 30 min. 2 ml of 4 N HCl (BDH inc.) was then added, and mixtures were left at room temperature for an additional 30 min. Tubes were centrifuged, and cell pellets were washed in 1 ml of 0.1 M Borax (Sigma, pH 8.5), then stored in 70% ethanol at Ϫ20°C. Pellets were washed 3ϫ in 0.5% Tween 20 (Sigma)/PBS, suspended in 50 l of Tween 20/PBS with 10 l of fluorescein isothiocyanate-conjugated ␣BrdUrd IgG 1 antibody (Becton Dickinson) for 30 min, rewashed 2ϫ, and resuspended in 1 ml of PBS with 5 g/ml propidium iodide (Sigma) for 10 min. The percentage of fluorescent cells in each sample was determined by flow cytometry (Coulter Epics XL, FL1 versus FL3). DNA synthesis as analyzed by [methyl-3 H]thymidine incorporation was performed by plating 50,000 cells in 60-mm plates, adding the indicated cell treatments, then labeling with 1 Ci/ml isotope for 4 h. Cells were trypsinized and added to glass microfiber filters (Whatman), and the filters were washed 3ϫ with water, dried, and immersed in scintillation mixture (Beckman) for assessment of radioactivity using a Beckman LS 1801 scintillation counter.
In Vitro Kinase Assay for the Measurement of JNK1 and p38 MAPK Activity-Cell cultures were washed 2ϫ in ice-cold PBS and lysed in mTNE buffer. 100 g of total cellular protein was added to 60 l of protein G-agarose beads and 0.5 g of rabbit polyclonal ␣JNK1 or ␣p38 antibody (C-17 or C-20; Santa Cruz Biotechnology), and samples were incubated overnight at 4°C for immunoprecipitation. Beads were then washed 2ϫ in cold mTNE, 2ϫ in cold "JNK" kinase buffer (2 mM HEPES, pH 7.2, 15 mM NaCl, 20 mM MgCl 2 , 10 mM ␤-glycerophosphate, 2 mg/ml p-nitrophenyl phosphate, 100 M sodium orthovanadate, 2 mM dithiothreitol) and resuspended in a kinase reaction solution of 30 l of JNK kinase buffer, 2 g of JNK or p38 MAPK substrate: GST fusion protein of c-Jun (1-79 aa) activation domain (Calbiochem) or GST fusion protein of ATF-2 (19 -96 aa) activation domain (Upstate Biotechnology) and 10 Ci of [␥-32 P]ATP (PerkinElmer Life Sciences) at 30°C for 20 min. Kinase reaction was stopped upon addition of SDS gel loading buffer (plus 5 mM EDTA) and heating of samples at 95°C for 10 min. Protein from supernatants was separated by SDS-polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane, and phosphorescence of the c-Jun substrate was quantitated. To compare c-Jun and ATF-2 N-terminal phosphorylation levels with total JNK1 and p38 MAPK levels, the membrane was incubated with ␣JNK1 or ␣p38 antibody, and membrane chemiluminescence was detected.

Low Level Generation of H 2 O 2 by Rat-1 Fibroblasts Is Required for Efficient Growth in
Culture-Addition of catalase preparations to the culture media has been shown by our laboratory to block the proliferation of a variety of immortalized cell types in a dose-dependent and reversible manner. 2 Extracellular catalase treatments have also been utilized to inhibit the effects of exogenous H 2 O 2 in cell culture (27) and forced overexpression of catalase can inhibit DNA synthesis and cell growth (28). To determine whether exogenous catalase activity could produce a similar response in our Rat-1 clones, cells were treated with doses of catalase or heat-inactivated preparation, and the total cell number after 4 days culture was observed (Fig. 1A). The growth of all Rat-1 lines was blocked after catalase addition; removal of this treatment during incubation or heat inactivation of the enzyme reversed this effect. Catalase treatment caused growth arrest without significant induction of the apoptotic program, as observed by cell mor-phology and trypan blue exclusion (not shown). The cell cycle profile of fibroblasts at increasing times of treatment (6 -48 h) was assessed by propidium iodide staining of total DNA and flow cytometry (not shown). A discreet block of the cell cycle at the G 1 /S or G 2 /M checkpoints was not observed, however, an accumulation of cell populations in S and G 2 phases suggests a prolonged S phase and inhibition of entry into mitosis. Coincubation of cells with an H 2 O 2 -generating glucose oxidase preparation in addition to catalase also rescued the proliferation block (Fig. 1B), providing evidence that catalase-mediated effects are indeed a result of changes in H 2 O 2 levels. Treatment with glucose oxidase alone at activities greater than 4 ϫ 10 Ϫ3 units/ml was toxic to cells, suggesting that a defined window of extracellular H 2 O 2 concentration is required for cell survival and growth.
Consistent with the observations of lower cell number and a prolonged S phase and mitosis, active DNA synthesis was slowed ϳ2to 4-fold by 2000 units/ml catalase as observed by BrdUrd ( Fig. 2A) or [methyl-3 H]thymidine (Fig. 2B) incorporation into cellular DNA. Also, the addition of 200 M H 2 O 2 increased DNA synthesis in all lines tested, in accordance with its well-documented growth-promoting effects. Various mammalian cell types have been shown to secrete H 2 O 2 into the surrounding culture medium (15,16). This generation is thought to reflect cell growth potential, because manipulation of cellular H 2 O 2 production results in changes to both the amount of extracellular H 2 O 2 and to levels of proliferation and transformation (19,20 (11,29). We wished to determine whether the removal of H 2 O 2 would cause ERK1/2 MAPK down-regulation, as further evidence for the redox sensitivity of this pathway. Thr-202/Tyr-204-phosphorylated ErK1 and 2 were detected using phospho-MAPK-specific primary antibodies. Levels of phosphorylated ERKs were compared with total ERK1/2 protein levels. Rat-1NeuN cells treated with H 2 O 2 displayed an increase in p44/42 MAPK phosphorylation, whereas the addition of catalase for 24 h (Ϯ10-min exposure to H 2 O 2 ) abolished MAPK activity (Fig. 3A). Heat-inactivated catalase did not produce such an effect. In contrast, this decrease in ERK1/2 phosphorylation was not observed after the identical treatment of Rat-1Neu8142 cells (Fig. 3B). The MEK1 inhibitor PD98059 (30) was added to cells for 1 h as a control for ERK1/2 down-regulation. These observations demonstrate that the alteration of Neu receptor activity can modulate the response of the MAPK pathway to a decrease in H 2 O 2 levels. However, because the growth response of Rat-1Neu8142 cells to catalase treatments differed little from that of Rat-1NeuN cells, this proliferation block appears largely independent of the p44/42 MAPK signal.
The response of Rat-1 clones to a range (0 -500 units/ml) of catalase activities was compared (not shown). Rat-1Neu8142 cells displayed a slightly higher resistance to the effects of catalase than Rat-1NeuN: This may result from differences in MAPK pathway activity or in the amount of endogenous H 2 O 2 production observed between the two clones. To test for the later possibility, SK-OV-3 cells (human ovarian adenocarcinoma) were also examined for resistance to catalase (not shown). These cells generate extracellular H 2 O 2 at a level ϳ6-fold higher than the Rat-1Neu8142 line. This increased H 2 O 2 did not correlate to any increase in resistance; indeed the growth of SK-OV-3 cells was extremely sensitive to the scavenging treatment. Sensitivity of cells to removal of H 2 O 2 in the presence of the MEK1 inhibitor PD98059 was also investigated (not shown). This concurrent treatment blocked the growth of Rat-1NeuN cells to a greater extent than catalase treatment alone, whereas this effect was less marked in the Rat-1Neu8142 line, which exhibited increased p44/42 MAPK phosphorylation. Such increased MAPK activity did not rescue cell growth to a significant degree, however. Thus it appears that, although the H 2 O 2 signal can regulate ERK1/2 activities, there are also other intracellular signals that are responsible for blocking proliferation.
Extracellular Catalase Activates the JNK1 Pathway-Because the effects of H 2 O 2 scavenging could not be fully attributed to the down-regulation of mitogenic MAPK signals, we examined the response of a SAPK signal; the JNK1 pathway. The N-terminal phosphorylation of c-Jun by JNK1 (p46 JNK) is a well-characterized cell stress response (1,31). JNK1 activity was examined by an in vitro kinase assay (Fig. 4). All treatments induced a transient increase in JNK1 kinase activity as quantitated by level of substrate (c-Jun peptide) phosphorylation. In Rat-1NeuN cells (Fig. 4A)  antibody to observe levels of activated p44/42 MAPKs, then stripped and reprobed with anti-ERK antibody. B, identical experiment using Rat-1Neu8142 cells. PPase, incubation with -protein phosphatase (negative control). ERK1/2 phosphorylation was increased in response to catalase treatment, the opposite effect to that observed in the wild type Neu-expressing Rat-1NeuN line. activity occurring 1 h after treatment. Nontoxic levels of H 2 O 2 could not induce JNK1 activity. 500 units/ml catalase (ϳIC 75 ) induced the JNK1 signal ϳ5-fold with peak activity after 30 min. Heat-inactivated catalase (500 units/ml) could also activate JNK1 to a lesser extent (ϳ2-fold over basal levels). Such stimulation did not result in any change in cell viability and growth as discussed. The response of Rat-1Neu8142 cells (Fig.  4B) was similar, except that 2 mM H 2 O 2 increased phosphorylation of the c-Jun substrate in an identical manner as catalase treatment (ϳ5-fold increase, peak at 30 min). These observations demonstrate that opposing directions of H 2 O 2 imbalance, namely, the addition of high concentrations of H 2 O 2 or removal of endogenous H 2 O 2 levels by catalase, can stimulate the JNK1 pathway in a similar fashion. Also, the activation of this SAPK by catalase is unaffected by differential Neu receptor activity, suggesting that it may be a critical component of the growth inhibition observed with all Rat-1 fibroblasts studied. These data indicate that the up-regulation of JNK1 is a common response to removal of extracellular H 2 O 2 in our Rat-1 fibroblast model.
P38 MAPK Phosphorylation Is Induced by H 2 O 2 but Not by Catalase Treatment-In a similar manner to that observed with JNK1 signal stimulation, the stress-induced p38 MAPK was phosphorylated upon addition of exogenous H 2 O 2 in all cells tested (Fig. 5). However, this response to oxidative stress was sustained over time, and no p38 activation was observed after treatment of cells with catalase or heat-inactivated catalase preparations. Levels of induction of p38 phosphorylation in both Rat-1NeuN and Rat-1Neu8142 lines by 2 mM H 2 O 2 was comparable to that observed under hyperosmotic conditions (0.2 M NaCl for 30 min). The phosphorylation of one p38 MAPK substrate, the transcription factor ATF2, was also investigated under the same conditions using an in vitro kinase assay (Fig.  5). No significant changes in phosphorylation of GST-ATF2 peptide by p38 were observed after catalase or H 2 O 2 treatments. It is undetermined whether this finding is a result of a functional impairment of p38 in the cells tested or the inability of the substrate utilized to respond to the p38 signal. Regardless, the absence of any direct p38 activation by catalase at levels observed to affect ERK1/2 and JNK1 suggests that the block in Rat-1 fibroblast proliferation is independent of this signal. and Rat-1Neu8142 (B) cells were treated as shown (30 min Ϫ2 h). Catalase/HI catalase ϭ 500 units/ml. Protein lysates were analyzed by Western blotting. Membranes were labeled with phospho-specific anti-p38 (Thr-180/Tyr-182) antibody to observe levels of activated p38 MAPK, then stripped and reprobed with anti-p38 antibody to compare total protein level in each sample. In both cell lines, p38 phosphorylation was induced by 2 mM H 2 O 2 , but not by catalase treatments. P38 MAPK was also immunoprecipitated from lysates for use in a kinase reaction with p38 substrate (ATF2 (19 -96 aa)) and label ([␥-32 P]ATP). No phosphorylation of this substrate was observed. C-6 extract, untreated C-6 glioma cell protein extract (negative control; NEB); C-6 ϩ Anisomycin, anisomycin-treated C-6 extract (positive control; NEB); ϪATF2, no ATF2 substrate included in reaction; Untr., no treatment.

DISCUSSION
This is the first report describing the effects of the removal of endogenously produced H 2 O 2 upon specific mitogenic and stress-responsive signaling pathways. To remove H 2 O 2 from the cellular environment, a catalase preparation was directly added to the culture media. This strategy was adopted so that the molecular signaling responses observed could be attributed to changes in H 2 O 2 status at the plasma membrane. One potential problem with this approach is enzyme purity, raising concerns that the effects observed could result from contaminants. To address this issue, we compared the effects of catalase to heat-inactivated catalase in all experiments. Incubation with the heat-inactivated solution did not affect cell viability and growth, nor did it affect Neu receptor or ERK activities. JNK1 activity did increase upon the addition of heat-inactivated catalase, but at a 2-to 3-fold lower level than that observed after treatment with active catalase. A comparable A. niger catalase preparation markedly reduced extracellular H 2 O 2 levels, whereas heat inactivation of this treatment blocked the effect. In addition, catalase-induced growth inhibition was rescued upon cotreatment of cells with H 2 O 2 -generating glucose oxidase. The levels of glucose oxidase activity required for this rescue were cytotoxic if added alone, indicating that the two enzymes are inversely affecting H 2 O 2 levels. These results provide evidence that the responses to the catalase preparation used specifically reflect its H 2 O 2 scavenging activity. It is likely that changes in the concentration of extracellular H 2 O 2 will in turn affect intracellular redox status, because H 2 O 2 can freely diffuse across cell membranes. Therefore, the effects of the catalase treatment used here are best described as a response to the alteration of the "pericellular" redox environment, which is both outside and inside of the plasma membrane. Certainly the fact, that Neu receptor activity was not greatly changed by the addition of catalase to cell cultures whereas various MAPK signaling cascades were, indicates that direct manipulation of intracellular signals by this type of treatment occur.
One such affected signal was the ERK1/2 pathway. The inhibition of ERK1/2 phosphorylation observed with catalase addition (and up-regulation observed with H 2 O 2 ) to Rat-1NeuN cells agrees with the previous findings that demonstrate the responsiveness of the MAPK cascade to H 2 O 2 (11,12). Constitutive activation of the Neu receptor in Rat-1 Neu8142 cells induced basal ERK1/2 activity and protected it from catalaseinduced inhibition. Neu interacts with several signaling cascades, including the Ras/Raf/MEK pathway, important in ERK1/2 regulation (25). The rescue of ERK phosphorylation was accompanied by only a slight increase in resistance of fibroblasts to growth inhibition by catalase. This result indicates that, although contributing to the cellular responses observed, ERK1/2 signal down-regulation is not critical for them.
Unlike the ERK response, the stress-activated JNK1 signal was rapidly and transiently induced by treatments that both increased and decreased H 2 O 2 in Rat-1NeuN and Rat-1Neu8142 lines. JNK1 is known to be responsive to oxidative stress (2, 32), but our findings suggest an additional role for this signal in response to an antioxidant. Rac1 and Cdc42, members of the Rho family of small GTPases involved in actin cytoskeleton regulation and cell transformation, have been shown to activate the JNKs through interaction with p21activated kinase (33,34). In addition, Rac1 is a cytoplasmic component of the plasma membrane-localized NAD(P)H oxidase complex that directs the extracellular generation of O 2 . and its dismutase product H 2 O 2 (35). Interestingly, the function of this complex appears to be essential not only for ROS generation but also for cell survival and growth (15,36). It follows that Rac may stimulate the JNK pathway through p21-activated kinase as well as H 2 O 2 production. Because our data demonstrate MAPK inhibition and SAPK activation after catalase treatment, it could be that loss of the H 2 O 2 stimulus blocks the Ras pathway while inducing other JNK-dependent signals like phosphoinositide 3-OH kinase (PI3K) and Rac. Recent evidence suggests that PI3K can play dual roles: the survival-promoting PI3K target Akt kinase (protein kinase B) was shown to phosphorylate Rac1 at serine 71 and abolish its GTPase activity (37). Treatment of our Rat-1 fibroblast lines with a 5 M concentration of the PI3K inhibitor LY294002 (38) inhibited cell proliferation to a comparable extent regardless of HER-2/Neu activity levels (not shown). Thus, changes to the PI3K/Akt signal might contribute to the catalase-induced effects upon JNK1 observed; however, this remains to be determined.
The role of JNK in cell cycle control is not known. c-Jun contributes to a number of contrasting biological processes as part of the AP-1 transcriptional complex (3). The degree of SAPK activity balanced with other AP-1-inducing signals such as the ERKs could decide which process is followed. In our system, ERK inhibition and JNK activation are observed with cell growth arrest. Studies by Wisdom et al. (31) demonstrate that JNK-mediated phosphorylation of c-Jun is specific for stress responses only, whereas c-Jun involvement in cell cycle progression is triggered by other factors. Another report indicates that the JNK signal can prolong the cell cycle at S and G 2 /M phases in a p53-independent manner (39), however, the mechanisms underlying this effect have yet to be elucidated. Indeed, we observed an increase of the proportion of cells in S and G 2 phases after incubation with catalase (results not shown). Because cells remain trapped in the late stages of the cell cycle, apoptosis may then be triggered by a "mitotic crisis." In conclusion, this work demonstrates that H 2 O 2 regulates components of various MAPK cascades to allow for the efficient growth of cultured Rat-1 fibroblasts. ERK1/2 activities are induced upon the addition of H 2 O 2 and inhibited by its removal. Expression of constitutively activated Neu receptor reversed this inhibition but had little effect on the catalaseinduced block of proliferation. However, the stress-activated JNK1 pathway was induced by both the addition and removal of the H 2 O 2 signal, regardless of Neu status. Another stress signal, the p38 MAPK pathway, was activated by toxic levels of H 2 O 2 but not by its removal, indicating that the JNK1 response might be specific. A discreet range of JNK activation may act as a cell sensor of oxidative and reducing stresses, leading to a response of growth arrest and death.