Hepatocyte resistance to oxidative stress is dependent on protein kinase C-mediated down-regulation of c-Jun/AP-1.

The prevention of injury from reactive oxygen species is critical for cellular resistance to many death stimuli. Resistance to death from the superoxide generator menadione in the hepatocyte cell line RALA255-10G is dependent on down-regulation of the c-Jun N-terminal kinase (JNK)/AP-1 signaling pathway by extracellular signal-regulated kinase 1/2 (ERK1/2). Because protein kinase C (PKC) regulates both oxidant stress and JNK signaling, the ability of PKC to modulate hepatocyte death from menadione through effects on AP-1 was examined. PKC inhibition with Ro-31-8425 or bisindolylmaleimide I sensitized this cell line to death from menadione. Menadione treatment led to activation of PKCmicro, or protein kinase D (PKD), but not PKCalpha/beta, PKCzeta/lambda, or PKCdelta/. Menadione induced phosphorylation of PKD at Ser-744/748, but not Ser-916, and translocation of PKD to the nucleus. PKC inhibition blocked menadione-induced phosphorylation of PKD, and expression of a constitutively active PKD prevented death from Ro-31-8425/menadione. PKC inhibition led to a sustained overactivation of JNK and c-Jun in response to menadione as determined by in vitro kinase assay and immunoblotting for the phosphorylated forms of both proteins. Cell death from PKC inhibition and menadione treatment resulted from c-Jun activation, since death was blocked by adenoviral expression of the c-Jun dominant negative TAM67. PKC and ERK1/2 independently down-regulated JNK/c-Jun, since inhibition of either kinase failed to affect activation of the other kinase, and simultaneous inhibition of both pathways caused additive JNK/c-Jun activation and cell death. Resistance to death from superoxide therefore requires both PKC/PKD and ERK1/2 activation in order to down-regulate proapoptotic JNK/c-Jun signaling.


INTRODUCTION
The ability of the cell to resist injury from excessive levels of reactive oxygen species (ROS) 1 is a critical survival mechanism in response to a variety of environmental stresses. Until recently oxidative stress was thought to trigger cell death through the adverse effects of biochemical reactions between oxidants and cellular macromolecules. However, it is now known that oxidant-induced death pathways are far more complex with death also resulting from the effects of oxidants on signal transduction pathways (1,2). Central among these signal transducers of oxidant-induced death are the mitogen-activated protein kinases (MAPKs). In the hepatocyte cell line RALA255-10G, resistance to toxicity from the ROS superoxide depends on activation of the MAPK extracellular signal-regulated kinase 1/2 (ERK1/2). Treatment of these cells with the superoxide generator menadione induces ERK1/2 activation (3). Inhibition of ERK1/2 signaling causes sustained activation of the c-Jun N-terminal kinase (JNK)/c-Jun/AP-1 pathway, resulting in cell death from normally nontoxic concentrations of menadione (3).
Overactivation of JNK/AP-1 signaling is known to mediate cell death from a number of stimuli in both hepatocytes and non-hepatic cells (4,5). Restricting the duration of this pro-apoptotic AP-1 activation following superoxide-generated cellular stress is required for hepatocyte resistance to oxidative stress.
The critical nature of cellular resistance to oxidant stress suggests the likelihood that redundant or complementary signaling pathways exist in order to protect hepatocytes against oxidant injury. However, upstream inhibitors of AP-1 activation other than ERK1/2 have not been identified after oxidative stress in hepatocytes. In addition to their effects on MAPK signaling, oxidants have been demonstrated to phosphorylate and thereby activate protein kinase C (PKC) isoforms. Multiple PKC isoforms are phosphorylated in response to oxidative stress induced by hydrogen peroxide (6,7), including PKCµ or protein kinase D (PKD) (8,9). Although originally described as a PKC family member, PKD has distinct features that make it part of a separate kinase family that also includes PKD2 and PKD3 (10). Both serine and tyrosine phosphorylation of PKD have been reported to result from hydrogen peroxide treatment (8,9).
Hydrogen peroxide-induced phosphorylation of Ser744/748 within the PKD activation loop occurs by a PKC-dependent mechanism (11,12). In addition to phosphorylation, PKD activation involves translocation from the cytoplasm to other cellular compartments including the nucleus and mitochondria (13,14). PKD activation has been reported to up-regulate NF-κB signaling, and the protective effects of PKD activation against death from hydrogen peroxide were associated with PKD-dependent NF-κB activation (8). Interestingly, PKD has also been reported to regulate JNK/c-Jun signaling (15)(16)(17)(18), suggesting the possibility that PKD activation induced by oxidative stress may also regulate the AP-1 pathway.
The objective of the present study was to examine whether PKC is an upstream regulator of the AP-1 death pathway in a hepatocyte cell line exposed to the superoxide generator menadione. The studies demonstrate that menadione causes a PKC-dependent activation of PKD. Inhibition of PKC/PKD activation leads to increased toxicity from menadione associated with sustained activation of the JNK/AP-1 pathway. Death resulting from PKC/PKD inhibition is blocked by the c-Jun dominant negative TAM67, suggesting that PKD-dependent resistance to menadione toxicity is the result of down-regulation of AP-1 signaling. These data therefore demonstrate for the first time a critical physiologic role for PKC/PKD in the regulation of AP-1 signaling.

MTT assay
Cell death was determined by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (22). At 24 h after treatment, the cell culture medium was aspirated and an equal volume of a 1 mg/ml MTT solution, pH 7.4 in DMEM was added to the cells. After incubation at 37°C for 1 h, the MTT solution was removed, and 1.5 ml of N-propanol was added to solubilize the formazan product. The absorbance of this compound was measured at 560 nm in a spectrophotometer. The percentage of cell death was calculated by dividing the optical density of a treatment group by the optical density for untreated, control cells, multiplying by 100, and subtracting that number from 100.

Fluorescence microscopy
The numbers of apoptotic and necrotic cells were quantified by fluorescence microscopy after costaining with acridine orange and ethidium bromide (23), as previously described (24).
Cells with a shrunken cytoplasm, and a condensed or fragmented nucleus as determined by acridine orange staining were considered apoptotic. Necrotic cells were detected by positive staining with ethidium bromide. A minimum of 400 cells per dish were examined, and the numbers of apoptotic and necrotic cells expressed as a percentage of the total number of cells counted.
For immunoprecipitations, cells were lysed in a buffer containing 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, pH 7.6, 1% Triton X-100, 1 ug/ml leupeptin, 2 ug/ml pepstatin and 1 mM phenylmethylsulfonyl fluoride. Protein determination was performed as above and 350 ug of protein were immunoprecipitated by a 1 h incubation with 2 ug of antihemagglutinin (anti-HA) antibody purified from the 12CA5 hybridoma. Samples were then incubated with protein A/G agarose (Sigma) for 30 min. The immune complexes were washed five times with 20 mM Tris, pH 7.5, 500 mM sodium chloride, and resolved on Western blots as described subsequently.

Transient transfections for PKD overexpression
RALA hepatocytes were transiently transfected with an expression vector containing the Escherichia coli β-galactosidase gene, CMV-βGal, or with PKD.SS744/748EE. PKD.SS744/748EE expresses an HA-tagged mutant PKD in which the activation loop residues have been replaced with negatively charged Glu resulting in a constitutively active PKD (27).
Cells were plated at a lower density and cultured at 37°C for a shorter period of time than in the other experiments to allow for a less confluent culture necessary for optimal transfection efficiency.
Transfections were performed with FuGENE 6 (Roche Applied Science, phosphorylation-independent c-Jun antibody (Santa Cruz Biotechnology). Proteins were visualized using a secondary antibody and chemiluminescent substrate as described above.

Adenoviruses
The adenoviruses Ad5LacZ containing the β-galactosidase gene (29), and Ad5TAM that expresses TAM-67, a dominant negative c-Jun (30), were employed. The adenoviruses were grown in 293 cells, purified by banding twice on CsCl gradients, and titered by plaque assay as previously described (31). RALA hepatocytes were infected at an MOI of 20 as previously described (24).

Statistical analysis
All numerical results are expressed as mean ± SE and represent data from three independent experiments with duplicate dishes in each treatment group. Statistical significance was determined by the Student's t-test. Calculations were made with Sigma Plot 2000 (SPSS Science, Chicago, IL).

PKC inhibitors sensitize RALA hepatocytes to death from menadione
Menadione is a quinone compound that undergoes redox cycling resulting in the formation of superoxide (32,33). To insure that death was secondary to PKC inhibition, the effect of a second PKC inhibitor, bisindolylmaleimide I (Bis I) (35), on menadione toxicity was investigated. Bis I alone was non-toxic (data not shown), but sensitization of RALA hepatocytes to death from menadione occurred with Bis I pretreatment. Death from Bis I/menadione cotreatment was 28% for 20 uM menadione, and 47% for 25 uM menadione at 24 hours (Fig. 1B), similar to findings for Ro-31-8425/menadione. In contrast, chelerythrine, another purported chemical PKC inhibitor (36), failed to sensitize RALA hepatocytes to death from menadione (data not shown).

Resistance to menadione toxicity requires early PKC signaling
To delineate the temporal involvement of protective PKC signaling in the menadione death pathway, RALA hepatocytes were examined for menadione-induced cell death after different times of Ro-31-8425 treatment. Conversion of the Ro-31-8425 1 h pretreatment to a 1 h post-treatment still sensitized the cells to significant toxicity from menadione, but reduced death from Ro-31-8425/menadione treatment by 36% (Fig. 3). When Ro-31-8425 treatment was delayed to 2 h after menadione administration, the amount of cell death was not significantly different from that of 25 uM menadione alone (Fig. 3). These data indicate that PKC-dependent signaling mediates an immediate protective response against menadione-induced oxidative stress.

Menadione causes selective PKD Ser744/748 phosphorylation and nuclear translocation
To identify the PKC isoform mediating RALA hepatocyte resistance to menadione toxicity, levels of active, phosphorylated PKC were examined after menadione treatment.
Menadione induced an increase in phospho-PKCµ  or phospho-PKD within 1 h after menadione treatment (Fig. 4A). Menadione-induced phosphorylation was specific for Ser744/748 as no change was detected in the levels of phosphorylation at the Ser916 residue. Levels of total PKD were also unaffected by menadione treatment. Menadione had no effect on the levels of phosphorylated PKCα/β, PKCζ/λ, or PKCδ/θ (Fig. 4A). Selective PKD Ser744/748 phosphorylation was induced by both nontoxic and toxic concentrations of menadione (Fig. 4B).
Once activated, PKD has been reported to translocate from the cytoplasm to cellular organelles including the nucleus and mitochondria (13,14). The levels of active, phosphorylated PKD and PKC were examined in nuclear and cytosolic fractions from RALA hepatocytes after menadione treatment. PKD phosphorylated at Ser744/748 was undetectable in the nuclei of untreated cells, but increased markedly within 1 h after menadione treatment (Fig. 4C).
Significant amounts of Ser916 phosphorylated and total PKD were present in the nucleus of untreated cells, but these levels were unchanged by menadione treatment (Fig. 4C). The exclusive presence of the nuclear protein Nopp140 in the nuclear fractions, and of PDI in the cytosolic fractions, demonstrated both the relative purity of the isolates and the equivalence of loading among samples (Fig. 4C). Phosphorylated PKCα/β, PKCζ/λ, and PKCδ/θ were all present in the nuclear fraction of untreated cells, and their levels were unaffected by menadione treatment (data not shown). Menadione treatment failed to lead to mitochondrial translocation of PKD or any PKC isoform (data not shown). Menadione treatment was therefore associated with the translocation of Ser744/748 phosphorylated PKD to the nucleus.

Ro-31-8425-induced sensitization to death from menadione is prevented by PKD expression
To mechanistically link inhibition of PKD activation with cell death from menadione, the effect of PKD expression on death from Ro-31-8425/menadione was determined. RALA hepatocytes were transiently transfected with a β-galactosidase expressing control vector or with the vector PKD.SS744/748EE that expresses a constitutively active PKD (27). Similar to primary hepatocytes, RALA hepatocytes are difficult to transfect, and the transfection efficiency was only 30% as assessed by β-galactosidase staining of control vector transfected cells.
Transfection with PKD.SS744/748EE resulted in PKD expression as demonstrated by immunoprecipitations with an anti-HA antibody followed by Western blotting for PKD (Fig.   6A). Expression of the constitutively active PKD decreased cell death from Ro-31-8425/menadione by 40% (Fig. 6B). While the inhibition of death was incomplete, this percentage of inhibition was commensurate with the transfection efficiency. The relative amount of cell death from Ro-31-8425/menadione was higher than in the previous experiments as less confluent cultures were employed in order to maximize transfection efficiency, and cell death from menadione is proportional to cell density. These data directly link PKC-dependent PKD activation to hepatocellular resistance to menadione-induced cell death.

Ro-31-8425 pretreatment results in overactivation of ERK1/2 and JNK MAPKs in response to menadione
The early temporal involvement of PKC/PKD signaling in the regulation of RALA hepatocyte death from menadione suggested that their activation may affect other cell signals that ultimately mediate resistance to toxic oxidative stress from menadione. Previous studies identified ERK1/2 activation as critical for RALA hepatocyte resistance to menadione toxicity (3), and PKD overexpression in non-hepatic cell types has been shown to induce ERK1/2 activation (16). In light of these findings, the possibility that PKC/PKD inhibition sensitized RALA hepatocytes to death from menadione by blocking ERK1/2 activation was examined by Western blotting for phospho-ERK1/2. Surprisingly, PKC inhibition by Ro-31-8425 led to activation rather than inhibition of ERK1/2 at a low dose of menadione that by itself failed to significantly affect ERK1/2 phosphorylation (Fig. 7A). ERK1/2 activation that occurred with toxic concentrations of menadione was further increased by Ro-31-8425 cotreatment (Fig. 7B).
Sensitization to menadione toxicity by PKC inhibition therefore could not be explained by a block in ERK1/2 MAPK signaling.
The mechanism of ERK1/2-mediated resistance to menadione toxicity is through the down-regulation of pro-apoptotic JNK/c-Jun/AP-1 signaling (3). Despite high levels of ERK1/2 activation, Ro-31-8425/menadione cotreatment led to sustained JNK1/2 and c-Jun activation as reflected by increased levels of these phospho-proteins on immunoblots (Fig. 7C). Levels of total JNK1/2 and c-Jun were unaffected by Ro-31-8425 or menadione treatment. JNK activation as measured by an in vitro kinase assay was markedly increased in both Ro-31-8425/menadioneand Bis I/menadione-treated cells as compared to cells treated with menadione alone (Fig. 7D and 7E). JNK activity was unaffected by administration of Ro-31-8425 at times later than 1 h after menadione treatment (Fig. 7F), corresponding to the inability of delayed administration of this inhibitor to sensitize cells to death from menadione (Fig. 3). Thus, PKC/PKD inhibition converted the RALA hepatocyte response to menadione to one of sustained JNK/c-Jun overactivation despite increased activation of ERK1/2.

Death from Ro-31-8425/menadione is mediated by c-Jun/AP-1 overactivation
Increased phosphorylation of c-Jun leads to its transcriptional activation as a subunit of the transcription factor AP-1. To assess levels of AP-1 activity, RALA hepatocytes were transiently transfected with an AP-1 driven luciferase reporter gene Coll73-Luc. Treatment with 25 uM menadione had no effect on AP-1 dependent luciferase activity (Fig. 8A). Ro-31-8425 treatment alone led to a modest increase in activity, while cotreatment with menadione led to a 2.5-fold increase in AP-1-dependent gene expression (Fig. 8A).
To determine whether increased AP-1 activity resulting from PKC/PKD inhibition mediated cell death from menadione, the effect of blocking c-Jun function was examined by adenoviral expression of the c-Jun dominant negative TAM67. TAM67 expression has been previously demonstrated to effectively inhibit AP-1 transcriptional activity in RALA hepatocytes (38). Cells were infected with the adenovirus Ad5LacZ as a control for the nonspecific effects of viral infection. Ad5LacZ-infected cells were sensitized to toxicity from menadione by Ro-31-8425 similar to uninfected cells (Fig. 8B). Infection with the TAM67-expressing adenovirus Ad5TAM completely blocked death from PKC inhibition and menadione treatment at the 25 uM concentration, and inhibited death at 30 uM menadione by 50% (Fig. 8B). PKC/PKD inhibition therefore sensitized RALA hepatocytes to death from menadione through overactivation of the c-Jun/AP-1 pathway.

ERK1/2 and PKC/PKD are independent signals for resistance to menadione toxicity
The present data together with previous studies (3), indicate that ERK1/2 MAPK and PKC/PKD signaling are both critical for RALA hepatocyte resistance to menadione toxicity.
These two signals may act sequentially or in parallel. ERK1/2 signaling was not downstream of PKC activation as inhibition of PKC increased rather than decreased ERK1/2 activation in response to menadione ( Fig. 7A and 7B). Inhibition of ERK1/2 signaling by U0126 similarly failed to affect menadione-induced PKD activation (Fig. 9A). These data suggested that the two signaling pathways acted independently to protect RALA hepatocytes from menadione toxicity.
To examine this possibility, the effect of co-inhibition of ERK1/2 and PKC on cell death from menadione was determined. At two concentrations of menadione, co-administration of Ro-31-8425 and U0126 led to a significantly increased amount of cell death over that from either inhibitor alone (Fig. 9B). In addition, cotreatment with both inhibitors led to a greater increase in phospho-JNK1/2 and phospho-c-Jun levels in response to menadione than did either inhibitor by itself (Fig. 9C). These data indicate that ERK1/2 and PKC/PKD are independent signals that down-regulate JNK/c-Jun after menadione treatment.

DISCUSSION
The present study demonstrates that resistance to superoxide toxicity in a hepatocyte cell line is mediated through a PKC-dependent serine phosphorylation and activation of PKD that temporally restricts pro-apoptotic AP-1 signaling. The critical findings that support this conclusion are: (1) the ability of the PKC inhibitors Ro-31-8425 and Bis I but not chelerythrine to sensitize RALA hepatocytes to death from normally nontoxic concentrations of menadione; (2) the specific phosphorylation of PKD at Ser744/748 by menadione, and prevention of this Menadione specifically induced PKD phosphorylation at Ser744/748 but not at Ser916.
PKD activation by phosphorylation of the Ser744/748 site has been previously demonstrated to occur by a PKC-dependent mechanism in non-hepatic cells (11,12). In contrast, phosphorylation at Ser916 occurs by PKC-independent autophosphorylation (40). It is also known that the PKC inhibitors employed in this study do not inhibit PKD directly (41 phosphorylation of PKD in other cell types (12,16). However, we were unable to examine this possibility experimentally because of the unavailability of rat-reactive antibodies for these activated PKC isoforms.
Hydrogen peroxide has been previously demonstrated to induce phosphorylation of both serine and tyrosine sites on PKD (8,11,27). Both serine and tyrosine phosphorylation activate PKD (11,42), and together they lead to synergistic activation (27). In contrast to the PKC dependence of Ser744/748 phosphorylation, tyrosine phosphorylation of PKD occurs through Src-Abl signaling (8,42). We were unable to examine for changes in tyrosine phosphorylation in our cells because of the lack of cross-reactivity between the phospho-tyrosine specific PKD antibody (42), and rat cells. However, while we cannot exclude a role for tyrosine phosphorylation of PKD in RALA hepatocyte resistance to superoxide toxicity, the PKCdependent nature of our findings strongly suggests that PKD Ser744/748 phosphorylation mediates PKD activation in our model.
In addition to its regulation by phosphorylation, PKD activity is a function of its translocation to different cellular compartments in response to stimuli. PKD predominantly resides in the cytoplasm although smaller amounts have been reported in Golgi and mitochondria in some cell types (10,13,14). In response to an activating stimulus, PKD moves briefly to the plasma membrane, returns to the cytoplasm, and then translocates to the nucleus (43). While