Protection of renal epithelial cells against oxidative injury by endoplasmic reticulum stress preconditioning is mediated by ERK1/2 activation.

We investigated the role of the endoplasmic reticulum (ER) stress response in intracellular Ca2+ regulation, MAPK activation, and cytoprotection in LLC-PK1 renal epithelial cells in an attempt to identify the mechanisms of protection afforded by ER stress. Cells preconditioned with trans-4,5-dihydroxy-1,2-dithiane, tunicamycin, thapsigargin, or A23187 expressed ER stress proteins and were resistant to subsequent H2O2-induced cell injury. In addition, ER stress preconditioning prevented the increase in intracellular Ca2+ concentration that normally follows H2O2 exposure. Stable transfection of cells with antisense RNA targeted against GRP78 (pkASgrp78 cells) prevented GRP78 induction, disabled the ER stress response, sensitized cells to H2O2-induced injury, and prevented the development of tolerance to H2O2 that normally occurs with preconditioning. ERK and JNK were transiently (30-60 min) phosphorylated in response to H2O2. ER stress-preconditioned cells had more ERK and less JNK phosphorylation than control cells in response to H2O2 exposure. Preincubation with a specific inhibitor of JNK activation or adenoviral infection with a construct that encodes constitutively active MEK1, the upstream activator of ERKs, also protected cells against H2O2 toxicity. In contrast, the pkASgrp78 cells had less ERK and more JNK phosphorylation upon H2O2 exposure. Expression of constitutively active ERK also conferred protection on native as well as pkAS-grp78 cells. These results indicate that GRP78 plays an important role in the ER stress response and cytoprotection. ER stress preconditioning attenuates H2O2-induced cell injury in LLC-PK1 cells by preventing an increase in intracellular Ca2+ concentration, potentiating ERK activation, and decreasing JNK activation. Thus, the ER stress response modulates the balance between ERK and JNK signaling pathways to prevent cell death after oxidative injury. Furthermore, ERK activation is an important downstream effector mechanism for cellular protection by ER stress.

When the kidney is rendered ischemic or is obstructed, it is protected against subsequent ischemia for at least 15 days from the time of the initial insult (1,2). The mechanisms of the protection afforded by prior ischemic stress are not known; however, many of the downstream signals that mediate ischemia/reperfusion injury have been well characterized. For example, reactive oxygen species (ROS) 1 are major mediators of ischemia/reperfusion injury (3)(4)(5). Although cells contain antioxidant defenses that minimize susceptibility to ROS, ROS generation and oxidative stress often exceed the cell's antioxidant capacity (6). Oxidative stress causes a rapid increase in intracellular free Ca 2ϩ concentration ([Ca 2ϩ ] i ) in a number of cell types, including cells derived from the renal tubule (7,8). Increases in [Ca 2ϩ ] i can result in enhanced Ca 2ϩ influx into mitochondria, disrupting mitochondrial metabolism and leading to cell death (9,10). Changes in [Ca 2ϩ ] i also modulate gene transcription and proteases and nucleases that control cell apoptosis (9,10). Recent investigations using renal epithelial cells indicate that the endoplasmic reticulum (ER) stress response can modulate both oxidative stress and [Ca 2ϩ ] i after treatment with organic hydroperoxides and alkylating agents (11,12). Likewise, increased attention has been paid recently to the possibility that ER stress influences the pathophysiology of acute ischemia in the brain, heart, or kidney (13)(14)(15). Given the association between ROS, oxidative stress, and ischemic injury in the kidney, we investigated the association between the ER stress response and H 2 O 2 in renal epithelial cells.
ROS-induced cell injury has been attributed, in part, to the change in activation of intracellular signaling molecules, including mitogen-activated protein kinases (MAPKs). MAPKs, which include extracellular signal-regulated protein kinase (ERK), c-Jun N-terminal kinase (JNK)/stress-activated protein kinase (SAPK), and p38 subfamilies, are important regulatory proteins that transduce various extracellular signals into intracellular events (16,17). The ERK, JNK/SAPK, and p38 subfamilies are all activated in response to oxidative injury. In 1996, Guyton et al. (18) implicated ERK activation as a survival factor following oxidant injury. Subsequent studies from a number of laboratories confirmed these findings in other cell types and with other injurious agents (19,20). The influence of JNK activation on cell survival following oxidative stress remains complex and highly controversial (6). Some studies have shown that JNK activation is correlated with cell death or apoptosis induced by agents that act, at least in part, via generation of ROS (21). Thus, ERK and JNK can have counteracting influences on cell survival during stress, including oxidant injury (18,20); and the balance between JNK and ERK activation may determine cell fate after renal ischemia/reperfusion injury (1,22). We investigated the association between ER stress and JNK or ERK phosphorylation.
The signals that activate the ER stress response are well characterized, as is the link between ER Ca 2ϩ release and cell injury. The ER is involved in both Ca 2ϩ signaling and posttranslational protein folding and maturation. Release of Ca 2ϩ from the ER may contribute to the ischemia/reperfusion injury of the brain, heart, and kidney (15,23,24). The response of the ER to unfolded proteins, known as the unfolded protein response (UPR), is currently the best understood model of ER stress signaling. The UPR has been shown to modulate expression of ER chaperones, allowing the cell to tolerate the accumulation of unfolded proteins (15,(25)(26)(27). In mammalian cells, the UPR is activated by agents that prevent protein glycosylation (tunicamycin) and disulfide bond formation (DTTox) and by agents that deplete ER Ca 2ϩ stores such as thapsigargin and the Ca 2ϩ ionophore A23187. The UPR also regulates many genes that affect diverse aspects of cell physiology. In the yeast Saccharomyces cerevisiae, 381 of 6000 genes were found to participate in the UPR, including 208 genes for which some functional information is available (28). Thus, the current view of the ER stress pathway has broadened from a pathway that simply regulates ER molecular chaperones to one that impacts many aspects of cell physiology (29). It is not entirely clear, however, how ER stress protects cells. In particular, it is not known whether the cellular MAPK and Akt/protein kinase B (PKB) pathways are effectors of the ER stress response (30,31).
Although the ER stress response has been implicated in the pathophysiology of ischemic injury (13)(14)(15)32), the role of individual ER stress proteins has not been addressed. Glucoseregulated proteins are the prototypical ER chaperones induced by ER stress (33). Induction of glucose-regulated proteins has been associated with protection against an increase in [Ca 2ϩ ] i (11,12,15) and facilitation of protein folding (25,27,34). Overexpression, antisense, and ribozyme approaches in tissue culture systems have led to the conclusion that GRP78, GRP94, and Adapt78 protect cells against cell death (11,(35)(36)(37)(38). Overexpression of the ER molecular chaperones also correlates with increased survival of renal epithelial cells subjected to ATP depletion (39). Therefore, glucose-regulated proteins might also be involved in ER protection from ischemic injury.
In this study, we investigated the effect of preconditioning the renal epithelial cell line LLC-PK 1 with different inducers of ER stress to determine whether ER stress can protect against injury caused by H 2 O 2 , the prototypical ROS thought to contribute to ischemia/reperfusion injury. Our results demonstrate that ER stress and GRP78 protect epithelial cells against oxidative stress by preventing the increase in [ preconditioning is linked to an increase in ERK1/2 and a decrease in JNK activation. These results place ERK as a distal mediator of the ER stress response in protection against oxidative injury and may be important for understanding how renal preconditioning affects mechanisms of ischemia/reperfusion injury in vivo.
Cell Cultures and Experimental Treatments-LLC-PK 1 cells (a porcine renal epithelial cell line with proximal tubule epithelial characteristics) were obtained from American Type Culture Collection (Manassas, VA). LLC-PK 1 cells were maintained in DMEM supplemented with 10% fetal calf serum. ER stress preconditioning was produced as described previously (11,12). Briefly, confluent cells were treated with DTTox (10 mM) for 3 h and returned to complete medium for 12-16 h. Alternatively, cells pretreated for 12-16 h in complete medium containing tunicamycin (1.5 g/ml), thapsigargin (0.3 g/ml), or A23187 (7 M) were washed with phosphate-buffered saline and returned to complete medium. To prevent the bias of pre-selection, cell injury was measured both immediately and 24 h after returning to complete medium. There was no significant difference in lactate dehydrogenase (LDH) release in cells with and without ER stress preconditioning at either time point (data not shown). Preconditioned and control cells were treated with 1 mM H 2 O 2 for 15 min in Earle's balanced salt solution (EBSS) and allowed to recover in complete medium. Alternatively, preconditioned cells were serum-deprived for 3-4 h and then treated with 250 M H 2 O 2 by directly adding H 2 O 2 to the medium without changing the medium to prevent an effect of medium change on signaling pathways during oxidative stress. The later protocol was used when activation of members of the MAPK family was determined to avoid serum stimulation upon refeeding. Specific inhibitors were added 1 h before H 2 O 2 treatment. Cell injury was determined by LDH release as a percent of total LDH as described previously (11).
In some experiments, the effect of altering expression of specific ER stress proteins was examined. LLC-PK 1 cells expressing an antisense RNA targeted to GRP78 (pkASgrp78 cells) or overexpressing human calreticulin (pkCRT cells) as well as controls transfected with the same pcDNA3 plasmid (used to construct both cell lines) containing no insert (pkNEO cells) were established as described previously (11,12). When the effects of ER stress on H 2 O 2 -induced cell injury and MAPK activation were tested using pkASgrp78 or pkCRT cells, three independent clones of both lines as well as three pkNEO lines were compared to avoid bias that might have occurred randomly through selection of individual clones.

Measurement of Intracellular Free Calcium Concentration-[Ca 2ϩ
] i was determined with the Ca 2ϩ -sensitive fluorescent dye Fura-2/AM according to Chen et al. (40) with modifications. Cells grown on coverslips coated with bovine collagen type I were rinsed with phosphatebuffered saline and loaded with 3 M Fura-2/AM in EBSS. Pluronic F-127 (20%) at a 1:1000 (v/v) dilution was added to Fura-2/AM to facilitate cell loading. In addition, 2 mM probenecid was added to prevent intracellular compartment transport or extrusion of Fura-2-free acid. After incubation with Fura-2/AM for 1 h at 37°C, cells were washed two to three times with EBSS in the presence of probenecid. The coverslips were positioned in a quartz cuvette, containing 2.5 ml of EBSS with probenecid, for fluorescence analysis using a Shimadzu RF-5000 spectrofluorophotometer. [Ca 2ϩ ] i was calculated as equal to K d (224 nM) ϫ (R Ϫ R min )/(R max Ϫ R) according to Grynkiewicz et al. (41). Fluorescence emission was monitored at 505 nm. R is the ratio of the fluorescence at 340 nm excitation to that at 380 nm excitation.
Construction of Recombinant Adenoviral Vectors-MEK1-DD, a constitutively active mutant of MAPK/ERK kinase-1 (MEK1), the upstream activator of ERK1/2, was created by PCR using primers to substitute serines 218 and 222 with aspartic acid residues as previously described (42). This mutant has been shown to activate ERK1/2 when expressed in COS-7 cells as well as in NIH3T3 cells (42,43). A recombinant adenoviral vector carrying the MEK1-DD cDNA (AdMEK1-DD) was constructed as previously described (44). Protein expression was confirmed by immunoblotting and by assay of ERK activity in mouse mesangial cells. 2 The recombinant adenovirus carrying the Escherichia coli LacZ gene (AdLacZ) encoding ␤-galactosidase was kindly provided by Dr. Roger Hajjar (Cardiovascular Research Center, Massachusetts General Hospital, Boston, MA).
Statistical Analyses-Student's t test was used to determine whether there was a significant difference between two groups (p Ͻ 0.05). When multiple means were compared, significance (p Ͻ 0.05) was determined by analysis of variance, followed by Fisher's protected least significant difference test. For analysis of variance, letter designations are used to indicate significant differences. Means with a common letter designation are not different, and those with a different letter designation are significantly different from all other means with different letter designations. StatView software (SAS Institute, Cary, NC) was used as a statistical tool in this study. LDH release, was not observed until 2 h and increased steadily through 4 h (Fig. 1). Thus, the increase in [Ca 2ϩ ] i preceded significant cell injury in LLC-PK 1 cells treated with H 2 O 2 .

ER Stress Preconditioning Prevents H 2 O 2 -induced Cell
Injury and Ca 2ϩ Accumulation-We next examined the effect of prior ER stress on H 2 O 2 -induced cell injury. LLC-PK 1 cells preconditioned with different ER stress inducers were protected against subsequent H 2 O 2 -induced cell injury at 4 h ( Fig.  2A). Prior ER stress induced by DTTox, tunicamycin, thapsigargin, or A23187 prevented the rise in [Ca 2ϩ ] i normally occurring 2 h after H 2 O 2 treatment (Fig. 2B). To confirm the biological significance of the protection afforded by ER stress preconditioning, we assessed cell injury at later time points under two different conditions of H 2 O 2 exposure. Under conditions of transient (Fig. 3, A and B) or continuous (C and D) H 2 O 2 exposure, the protective effect afforded by previous ER stress persisted up to 24 h. Thus, ER stress preconditioning modified the magnitude of oxidant-induced cell injury and not just the kinetics of the process.
Blocking Expression of GRP78 Disrupts the Effect of ER Stress Preconditioning-Because prior ER stress prevented both the increase in [Ca 2ϩ ] i and cell death after H 2 O 2 , we determined whether disruption of the ER stress response alters the sensitivity to oxidant injury and prevents adaptation. We tested the H 2 O 2 sensitivity of pkASgrp78 cells, which express antisense grp78 and are unable to increase grp78 gene expression after ER stress, relative to pkNEO cells, which carry only the neomycin resistance marker. Cell death was much greater in pkASgrp78 cells compared with pkNEO cells (Fig. 4A). ER stress did not result in protection of pkASgrp78 cells subsequently exposed to H 2 O 2 , in contrast to pkNEO cells (Fig. 4B). Thus, the ability of ER stress preconditioning to prevent H 2 O 2induced cell injury depended on an intact ER stress response and expression of GRP78.
Overexpression of Calreticulin Blocks Cell Injury and Ca 2ϩ Disturbance-If one mechanism underlying ER stress preconditioning is to prevent increased [Ca 2ϩ ] i , then increasing expression of Ca 2ϩ -binding proteins in the ER might produce a similar effect. Previously, we demonstrated that overexpression of the ER calcium-binding protein calreticulin in LLC-PK 1 (pkCRT) cells prevents deregulation of Ca 2ϩ in response to alkylating agents and organic hydroperoxides (11,12). pkCRT cells were also more resistant to H 2 O 2 -induced cell injury relative to pkNEO cells (Fig. 5A) (Fig. 6A). The phosphorylation states of ERK, JNK, p38, and Akt/PKB were evaluated between 0 and 60 min of H 2 O 2 exposure by employing phosphospecific antibodies (Fig. 6B). There was strong basal phosphorylation of Akt/PKB with no detectable change in phosphoryl-ation of Akt/PKB upon 250 M H 2 O 2 exposure. Anti-phospho-ERK1/2 antibody recognized two bands corresponding to ERK1 (44 kDa) and ERK2 (42 kDa). Increased phosphorylation of ERK was noted at 10 min, and maximum phosphorylation occurred at 30 min of H 2 O 2 exposure. Although ERK phosphorylation decreased at a later time point (60 min), it remained above the levels seen in the absence of treatment. Activation of JNK, reflected by phosphorylation of the 46-and 54-kDa bands, occurred within 30 min of H 2 O 2 exposure and reached a maximum at 60 min prior to a significant increase in LDH release. There was weak basal phosphorylation of p38 before treatment. After treatment, the phosphorylation of p38 decreased and then increased to a level above the base line after 30 and 60 min of H 2 O 2 exposure. Exposure to H 2 O 2 did not alter the levels of total Akt/PKB, ERK, JNK, and p38 proteins, which also confirmed equal protein loading on the gels (Fig. 6B).
ER Stress Modulates Cellular MAPK Activation-We investigated whether prior ER stress modulates the MAPK signaling pathways and their response to H 2 O 2 . When cells were pretreated with tunicamycin and then exposed to H 2 O 2 after serum starvation, the ER stress-preconditioned cells were more resistant to H 2 O 2 (Fig. 7A) and had much stronger and persistent phosphorylation of ERK1/2 (Fig. 7C). The phosphorylation of MEK1/2, an upstream kinase of ERK, was increased in tunicamycin-pretreated cells relative to control cells (Fig. 7C), indicating that ER stress preconditioning results in activation of the ERK pathway by regulating kinases upstream of ERK. Prior induction of ER stress in LLC-PK 1 cells also resulted in a diminished level of JNK phosphorylation in response to H 2 O 2 exposure (Fig. 7C). Total ERK and JNK proteins were equivalent in ER-stressed and control cells before and after H 2 O 2 treatment, confirming equal protein loading (Fig. 7C). Expression of GRP94, another prototypical ER stress protein, is shown in the lower panel of Fig. 7C to confirm that the ER stress response was activated in response to tunicamycin. These effects of ER stress on MAPKs are not specific to tunicamycin because pretreatment of LLC-PK 1 cells with another ER stress inducer, thapsigargin, protected cells against oxidative injury (Fig. 7B) and, like tunicamycin, up-regulated the MEK/ERK and down-regulated the JNK pathways in response to H 2 O 2 (Fig. 7D).
Influences of Akt/PKB, ERK, and JNK Pathways on Cell Survival in H 2 O 2 -treated Cells-Preincubation of LLC-PK 1 cells with a specific phosphatidylinositol 3-kinase inhibitor (LY294002, 10 M) completely abolished the phosphorylation of Akt/PKB (Fig. 8A). However, this agent did not alter the cell injury seen after H 2 O 2 exposure (Fig. 8B). Thus, Akt/PKB phosphorylation did not influence cell survival in LLC-PK 1 cells treated with H 2 O 2 .
To evaluate whether ERK phosphorylation contributes to cell survival in response to oxidative stress, cells were infected with adenovirus encoding MEK1-DD, a constitutively active upstream kinase of ERK. LLC-PK 1 cells infected with constitutively active MEK1-DD had increased levels of ERK1/2 phosphorylation (Fig. 8C) and were resistant to H 2 O 2 toxicity (Fig.   8D). Therefore, ERK phosphorylation may promote cell survival in LLC-PK 1 cells exposed to H 2 O 2 . It was not possible to use U0126, an inhibitor of MEK1, because U0126 is a free radical scavenger and also inhibited the activation of JNK that occurred at 30 -60 min upon H 2 O 2 exposure (data not shown).
We next examined whether JNK activation is linked to cell injury of H 2 O 2 -treated LLC-PK 1 cells. Preincubation of LLC-PK 1 cells with a specific JNK inhibitor (SP600125, 40 M) 1 h prior to addition of H 2 O 2 completely abolished the phosphorylation of JNK (Fig. 8E) and protected cells from oxidative injury (Fig. 8F). Thus, ERK1/2 activation protected against oxidative injury, whereas JNK activation potentiated cell death.
Cells Overexpressing Calreticulin Have More ERK Phosphorylation in Response to H 2 O 2 Treatment-As indicated previously by the dose response to H 2 O 2 (Fig. 5A) and by the time course of LDH release after exposure to 500 M H 2 O 2 (Fig. 9A), pkCRT cells were more resistant to H 2 O 2 -induced cell injury than were pkNEO cells. To clarify whether the protective effect afforded by calreticulin overexpression is through the modulation of MAPK pathways, we examined the activation pattern of ERK and JNK in pkCRT cells in response to H 2 O 2 treatment. In comparison with pkNEO cells, there was more phosphorylation of ERK in pkCRT cells after H 2 O 2 exposure. The phosphorylation of JNK was not significantly different between pkNEO and pkCRT cells (Fig. 9B). Thus, overexpression of calreticulin resulted in enhanced ERK activation in response to H 2 O 2 treatment in LLC-PK 1 cells. Significant differences (*, p Ͻ 0.05) between two groups at various time points were determined by Student's t test. In parallel cells were lysed at different times before significant cell injury (C and D). MEK1/2, ERK, and JNK activation in cell lysates was determined with phospho-specific antibodies that recognize phosphorylated (p) MEK1/2, ERK, and JNK. Total (T) MEK1/2, ERK, and JNK were also measured by specific antibodies. Expression of GRP94 confirmed the induction of ER stress response by tunicamycin and thapsigargin. The blots shown are representative of three separate experiments in which similar results were observed.

GRP78 May Play a Role in Modulating Cell
Signaling Pathways in ER Stress-preconditioned Cells-As indicated previously by the dose response to H 2 O 2 (Fig. 4A) and by the time course of LDH release after exposure to 250 M H 2 O 2 (Fig.  10A), pkASgrp78 cells were more susceptible to H 2 O 2 -induced cell injury. The phosphorylation of ERK in pkASgrp78 cells after 30 min of H 2 O 2 exposure was less than that in pkNEO cells (Fig. 10B). After pretreatment with tunicamycin, the pk-NEO cells also had a significant increase in phosphorylation of ERK in response to H 2 O 2 , similar to wild-type LLC-PK 1 cells (Fig. 7C). In contrast, there was only a mild increase in ERK phosphorylation in response to H 2 O 2 in pkASgrp78 cells (Fig.  10B). In contrast to the effect of GRP78 in decreasing ERK activation, the phosphorylation of JNK in response to H 2 O 2 was much greater in pkASgrp78 cells in comparison with pk-NEO cells. After pretreatment with tunicamycin, the activation of JNK in both cells decreased, but remained greater in pkAS-grp78 cells than in pkNEO cells (Fig. 10B).
To evaluate whether ERK activation can rescue the pkAS-grp78 cells exposed to H 2 O 2 , we infected them with adenovirus carrying constitutively active MEK1-DD to enhance phosphorylation of ERK in these cells (Fig. 10C). After adenoviral infection, the pkASgrp78 cells with constitutively active ERK were more resistant to oxidative injury compared with cells infected with adenovirus carrying the LacZ gene only (Fig.  10D). Thus, the enhanced sensitivity of pkASgrp78 cells to oxidative stress (Figs. 4A and 10A) was associated with less phosphorylation of ERK in response to H 2 O 2 (Fig. 10B), an effect that could be overcome by expression of constitutively active MEK1 (Fig. 10D). Therefore, ERK pathway activation is a downstream effector of the protective response mediated by ER stress and GRP78 expression. Cells were lysed at different times, and Akt/PKB activation in cell lysates was determined using phospho-specific antibody recognizing phosphorylation (p) at Ser 473 of Akt/PKB. Total (T) Akt/PKB was also measured with anti-total Akt/PKB antibody to confirm equal protein loading. B, at different times, cell injury was measured by LDH release. C, 80% confluent LLC-PK 1 cells in DMEM with 2% FCS were infected for 48 h with a recombinant adenoviral vector carrying MEK1-DD cDNA (AdMEK-DD) that constitutively activates ERK. Other cells were infected with the adenovirus carrying the E. coli LacZ gene (AdLacZ) encoding ␤-galactosidase as a negative control. The multiplicity of infection was ϳ100. After infection, cells were allowed to recover in DMEM with 10% FCS for 24 h and then serum-deprived for 3 h before addition of 250 M H 2 O 2 . Cells were lysed at different times, and ERK activation in cell lysates was determined with phospho-specific antibody. Total ERK was determined to confirm equal protein loading. D, at times indicated, cell injury was determined by measurement of LDH release. E, LLC-PK 1 cells were pretreated with 40 M SP600125 for 1 h before addition of 250 M H 2 O 2 . Cells were lysed at times indicated, and JNK activation in cell lysates was determined by phospho-specific antibody recognizing dually phosphorylated activated JNK. Total JNK was also measured using anti-total JNK antibody to confirm equal protein loading. F, at different times, cell injury was determined by measurement of LDH release. The LDH data are the means Ϯ S.E. of duplicate measurements from three different experiments. Significant differences (*, p Ͻ 0.05) between two groups at various time points were determined by Student's t test. The blots shown are representative of three separate experiments in which similar results were observed. At time points indicated, cell injury was determined by LDH release. The LDH data are means Ϯ S.E. from results with three separate pkNEO clones and three separate pkCRT clones of LLC-PK 1 cells. Significant differences (*, p Ͻ 0.05) between groups at various time points were determined by Student's t test. B, at time points indicated after H 2 O 2 treatment, cells were lysed, and phosphorylated (P) and total (T) ERKs and JNKs were determined with specific antibodies. The blots shown are representative of the results from three separate clones of pkCRT or pkNEO cells. Confirmation of overexpression of calreticulin (CRT) is also shown.

DISCUSSION
Increased amounts of RNA transcripts for GRP78 and other ER stress proteins are characteristic of rat models of brain, kidney, and heart ischemia/reperfusion (13)(14)(15)48). Overexpression of the ER molecular chaperones correlates with increased survival of cells subjected to conditions modeling ischemia/reperfusion (39) as well as injury by other agents (27,49,50). Despite efforts of a number of laboratories, however, the mechanism of cytoprotection due to ER stress remains unclear. The results from this investigation demonstrate that the protection that ER stress affords against oxidative injury is mediated by enhanced ERK pathway activation.
The ability of prior ER stress to prevent a rise in [Ca 2ϩ ] i in response to H 2 O 2 is consistent with the facts that the ER serves as a major storage site of intracellular calcium (51) and that several of the ER molecular chaperones bind calcium (38,52,53). Our data are consistent with previous studies showing that induction of ER stress moderates rises in [Ca 2ϩ ] i that occur in H 2 O 2 -induced cell injury as well as oxidative stress caused by glutathione depletion and organic oxidants, hence reducing the threat of oxidative stress to the cell (11,12,54,55). Although other ER proteins may contribute to this cytoprotection, our evidence suggests that GRP78 is a critical aspect of the integrated ER stress response. Blocking induction of GRP78 sensitizes the epithelial cell to oxidative stress. There are previous reports that the protective effects of prior ER stress against iodoacetamide and tert-butyl hydroperoxide toxicity are also disrupted in pkASgrp78 cells (11,12). The levels of expression of other stress proteins in pkASgrp78 cells, including GRP94 and another KDEL protein, increase when cells express the antisense GRP78 RNA (56). Despite the increases in these other ER stress proteins, they cannot overcome the effect of disabling the ER stress response by GRP78 depletion.
Calreticulin is one of the major ER Ca 2ϩ -binding proteins in non-muscle cells and has been shown to modulate ER Ca 2ϩ release (57). Cells that overexpress calreticulin have increased ER Ca 2ϩ -buffering capacity and/or resist Ca 2ϩ toxicity (11,12,58). Calreticulin is increased markedly when NIH3T3 cells are preconditioned with thapsigargin (59); however, calreticulin expression is not altered appreciably in pkASgrp78 cells (56). Overexpression of calreticulin protects LLC-PK 1 cells against H 2 O 2 -induced cell injury and an increase in [Ca 2ϩ ] i . Thus, the fact that increased ER Ca 2ϩ buffering protects against oxidative injury supports the suggestions that the protective effect of ER stress is due, in part, to better Ca 2ϩ buffering, decreased Ca 2ϩ release, or an indirect mechanism involving cooperation between Ca 2ϩ uptake by ER and net efflux across the plasma membrane. Each of these mechanisms could prevent a change in [Ca 2ϩ ] i .
Renal epithelial LLC-PK 1 cells have been used extensively to investigate cytotoxicity and stress gene activation (11,12,56,60). However, the influence of cell signaling pathways on stress pathways and cell survival during oxidative stress is not well characterized. In comparison with native LLC-PK 1 cells, cells Cells were lysed after a 0-or 30-min exposure to H 2 O 2 , and phosphorylated (p) and total (T) ERKs and JNKs were determined with specific antibodies. C and D, constitutively active MEK1 rescued pkASgrp78 cells from H 2 O 2 -induced cell injury. pkASgrp78 cells at 80% confluence were infected with a recombinant adenoviral vector carrying the MEK1-DD cDNA (AdMEK1-DD) that constitutively expresses MEK1, the upstream activator of ERK, in DMEM with 2% FCS for 48 h. Infection with a recombinant adenovirus carrying the E. coli LacZ gene (AdLacZ) encoding ␤-galactosidase served as a negative control. The multiplicity of infection was ϳ100. Cells were allowed to recover in DMEM with 10% FCS for 24 h and then serum-deprived for 3 h before addition of 250 M H 2 O 2 . Cells were lysed at time points indicated, and phosphorylated and total ERK activation in cell lysates was determined using specific antibodies (C). At 0, 30, and 60 min after addition of H 2 O 2 , cell injury was determined by measuring LDH release (D). The LDH data are means Ϯ S.E. of results with three separate pkNEO clones and three separate pkASgrp78 clones of LLC-PK 1 cells. Significant differences (*, p Ͻ 0.05) between groups at various time points were determined by Student's t test. The blots shown are representative of three separate clones of pkASgrp78 or pkNEO cells in which similar results were observed. preconditioned with ER stress have greater ERK and less JNK phosphorylation following oxidative stress. The activation of the ERK pathway results from up-regulation of MEK1, the upstream kinase. GRP78 modulates these MAPK signaling pathways because prevention of GRP78 induction in pkAS-grp78 cells alters the effects of H 2 O 2 on ERK and JNK pathways. There is less ERK and more JNK activation upon exposure to H 2 O 2 in pkASgrp78 cells, and these cells are more sensitive to oxidative injury. Constitutively active MEK1 rescues the pkASgrp78 cells from H 2 O 2 -induced cell injury. Thus, ERK1/2 phosphorylation is a downstream effector of the protection against oxidative injury afforded by GRP78. Our findings demonstrate for the first time that the ER stress response and specifically GRP78 expression act via ERK and JNK signaling pathways to modulate cell injury in response to oxidative stress.
Whether or not changes in [Ca 2ϩ ] i following H 2 O 2 treatment are also involved in modulating JNK activation and cell death in a coordinate manner is not clear. The fact that [Ca 2ϩ ] i can modulate MAPK signaling and cell injury may provide a link between these two events and ER stress. The ability of ER stress to suppress JNK activation in response to subsequent oxidative stress might be related to the prevention of an increase in [Ca 2ϩ ] i by GRP78 and other ER stress proteins. There is precedent for an interaction among proteins that regulate oxidative stress, [Ca 2ϩ ] i , and JNK activation. Under normal conditions, the redox regulatory protein thioredoxin has been shown to bind and inhibit the activity of apoptosis signalregulated kinase-1 (ASK1), a MAPK kinase kinase involved in both JNK and p38 kinase activation (6). Furthermore, in Ha-CaT keratinocytes, increased [Ca 2ϩ ] i leads to oxidation of thioredoxin, dissociation of the thioredoxin-ASK1 complex, and activation of JNK (61). By preventing a rise in [Ca 2ϩ ] i , the ER stress response, including up-regulation of GRP78 expression, could prevent this dissociation of the thioredoxin-ASK1 complex and diminish activation of the JNK pathway. Although such a mechanism is speculative at this time, it represents one plausible mechanism linking oxidative stress and deregulation of [Ca 2ϩ ] i to JNK activation and cell death. An increase in [Ca 2ϩ ] i is, however, not the only contributor to the increased phosphorylation of JNK. This is demonstrated by our experiments with calreticulin-overexpressing cells. The cells were protected against H 2 O 2 -induced cell injury associated with prevention of an increase in [Ca 2ϩ ] i . Despite the absence of a change in [Ca 2ϩ ] i , JNK phosphorylation was increased in the cells to an extent equivalent to the increase seen in control cells.
The mechanism by which ER stress preconditioning and increases in ER GRP78 or calreticulin can result in ERK pathway activation is also not defined by these studies. Given the recent observation that Ras restricted to the ER compartment can activate Raf and cellular ERK pathways (62), it is possible that GRP78 or calreticulin leads to increased Ras/Raf activation in response to H 2 O 2 . This may occur because of increased local Ca 2ϩ concentration in the ER of GRP78-expressing cells since it is well known that Ca 2ϩ regulates the Ras/Raf/MEK1/ ERK signaling pathways (63). Regardless of the mechanisms, the fact that ERK1/2 activation is modulated by ER stress, and infection with an adenovirus expressing constitutively active MEK1 protects pkASgrp78 cells, indicates that ERK1/2 activation is a downstream effector of ER stress-induced survival during oxidative stress.
In conclusion, over the past 10 years, much attention has been directed toward understanding the mechanisms and physiological significance of the ER stress response or UPR (31,64). However, our knowledge of how ER stress impacts cell signaling pathways and mediates its protective effect remains incomplete. In this study, we have demonstrated that expression of ER stress proteins (and in particular, GRP78 and calreticulin) attenuates H 2 O 2 -induced cell injury in LLC-PK 1 cells by preventing increases in [Ca 2ϩ ] i . Induction of ER stress preconditioning and GRP78 expression potentiates ERK and diminishes JNK signaling pathway activation during oxidative stress. Activation of the ERK pathway rescues cells not able to induce GRP78 so that they resist oxidant injury. These results indicate that the MAPK signaling pathways are critical downstream effectors of the protection afforded by ER stress proteins against oxidative stress in renal epithelial cells.