Endoplasmic Reticulum Stress Proteins Block Oxidant-induced Ca2+ Increases and Cell Death*

Oxidants are important human toxicants. Increased intracellular free Ca2+ may be critical for oxidant toxicity, but this mechanism remains controversial. Furthermore, oxidants damage the endoplasmic reticulum (ER) and release ER Ca2+, but the role of the ER in oxidant toxicity and Ca2+ regulation during toxicity is also unclear.tert-Butylhydroperoxide (TBHP), a prototypical organic oxidant, causes oxidative stress and an increase in intracellular free Ca2+. Therefore, we addressed the mechanism of oxidant-induced cell death and investigated the role of ER stress proteins in Ca2+ regulation and cytoprotection after treating renal epithelial cells with TBHP. Prior ER stress induces expression of the ER stress proteins Grp78, Grp94, and calreticulin and rendered cells resistant to cell death caused by a subsequent TBHP challenge. Expressing antisense RNA targeted to grp78prevents grp78 induction sensitized cells to TBHP and disrupted their ability to develop cellular tolerance. In addition, overexpressing calreticulin, another ER chaperone and Ca2+-binding protein, also protected cells against TBHP. Interestingly, neither prior ER stress nor calreticulin expression prevented lipid peroxidation, but both blocked the rise in intracellular free Ca2+ after TBHP treatment. Loading cells with EGTA, even after peroxidation had already occurred, also prevented TBHP-induced cell death, indicating that buffering intracellular Ca2+ prevents cell killing. Thus, Ca2+ plays an important role in TBHP-induced cell death in these cells, and the ER is an important regulator of cellular Ca2+ homeostasis during oxidative stress. Given the importance of oxidants in human disease, it would appear that the role of ER stress proteins in protection from oxidant damage warrants further consideration.

death has been suggested by some, and the subject has been reviewed (5)(6)(7)(8)(9)(10); yet this proposal remains controversial. On the one hand, it has been suggested that an increase in free Ca 2ϩ is nothing more than a late event associated with loss of membrane integrity and does not contribute appreciably to oxidant-induced cell killing (7)(8)(9). On the other hand, it has been proposed that an early increase in intracellular free Ca 2ϩ exacerbates oxidative stress, damages mitochondria, activates Ca 2ϩ -dependent degradative enzymes, and disrupts the cytoskeleton, all of which play a central role in oxidant-induced cell death (5,10,11). Thus, the contribution of Ca 2ϩ deregulation to oxidant-induced cell death remains unclear.
In addition, in cases where Ca 2ϩ appears to be involved in cell killing, the contribution of intracellular versus extracellular Ca 2ϩ is not clear (5,10). The endoplasmic reticulum (ER) 1 is the major intracellular Ca 2ϩ storage site (12,13). The ER Ca 2ϩ pool plays an important role in the folding and posttranslational processing of secreted and cell surface proteins (14). ER chaperones, including Grp78/BiP (where Grp is named for glucose-regulated protein), Grp94, calnexin, and calreticulin, are Ca 2ϩ binding proteins (15)(16)(17) and regulate ER Ca 2ϩ accumulation and release (18 -21). Inhibiting the ER Ca 2ϩ -ATPase with thapsigargin or adding ionophores releases the ER Ca 2ϩ pool, blocks ER protein processing causing partially folded proteins to accumulate, and activates transcription of ER chaperone genes, e.g. grp78 and grp94 (21)(22)(23)(24). Loss of ER Ca 2ϩ also activates eIF2␣ kinases causing a general inhibition of translation, effects that are attenuated by prior induction of ER stress proteins (25)(26)(27)(28)(29). Overexpression of calreticulin and Grp78, both of which are ER Ca 2ϩ -binding proteins and chaperones, increases the capacity of intracellular Ca 2ϩ stores (20,21) and prevents Ca 2ϩ toxicity (30,31). Oxidants, sulfhydryl active agents, and free radicals also inhibit Ca 2ϩ -ATPases and release Ca 2ϩ from the ER (32)(33)(34)(35)(36), suggesting that ER Ca 2ϩ could play a role in oxidant toxicity. Nonetheless, despite the fact that the ER is the major intracellular Ca 2ϩ store, that ER stress proteins regulate the ER Ca 2ϩ pool, and that increased cellular free Ca 2ϩ may contribute to oxidant-induced cell killing, there is no direct evidence linking regulation of cellular Ca 2ϩ by the ER to cell death after oxidant exposure.
Prior induction of ER chaperones imparts tolerance to the translational block caused by ER Ca 2ϩ depletion and protects against the toxicity of Ca 2ϩ ionophores and other toxic insults (22,37,38). Protection depends in part on expression of grp78 and grp94 because preventing an increase in their expression sensitizes cells to injury (30,31,39,40). Recently, we demonstrated that increasing the expression of ER stress protein genes protects renal epithelial cells against a subsequent challenge with the alkylating and acylating agents, iodoacetamide, or nephrotoxic cysteine conjugates, respectively (31,41). With iodoacetamide, protection required grp78 induction and was linked both to inhibition of lipid peroxidation and maintenance of low intracellular free Ca 2ϩ (31). Furthermore, overexpression of calreticulin, an ER Ca 2ϩ -binding chaperone that increases ER calcium retention (19,42,43), also protected cells from iodoacetamide toxicity (31). These and related studies (21) indicate that control of intracellular Ca 2ϩ by the ER may be important in preventing toxicant-induced cell death. However, the possibility that ER stress proteins blocked oxidative stress directly, thus preventing the rise in Ca 2ϩ , could not be excluded in these studies.
TBHP is a prototypical organic oxidant and has been used extensively to study the role of Ca 2ϩ in oxidant-induced cell death. TBHP treatment causes peroxidation of cellular lipids, oxidation of glutathione, loss of protein thiols, release of ER Ca 2ϩ , a general increase in cytosolic free Ca 2ϩ , a permeability transition in the mitochondrial inner membrane, and lipid peroxidation (36, 44 -47). However, the role of these perturbations in TBHP-induced cell death and in particular the role of Ca 2ϩ remains unclear (5,8,9,44). Our previous studies on the ER stress response provided new insights into cell death induced by alkylating and acylating agents (31). Therefore, we examined the effect of ER stress on TBHP toxicity. The results indicate that prior ER stress or overexpression of calreticulin prevented TBHP-induced cell death and blocked the increase in cellular Ca 2ϩ . Notably, neither manipulation blocked lipid peroxidation pointing to a central role for Ca 2ϩ and regulation of cellular Ca 2ϩ levels by the ER in oxidant-induced cell death.

MATERIALS AND METHODS
The acetoxymethyl esters of EGTA (EGTA-AM) was purchased from Molecular Probes (Eugene, OR). Sigma provided the TBHP. Other common chemical and cell culture reagents were obtained from commercial sources.
LLC-PK1 cells were obtained from American Type Culture Collection (Manassas, VA) at passage 195 and were used from passage 205-215. LLC-PK1 cells expressing either an antisense grp78 construct (pkAS-grp78 cells) or overexpressing calreticulin (pkCRT cells) as well as their counterparts transfected with the same pcDNA3 based plasmid containing no insert (pkNEO cells) were all selected for neomycin resistance and cloned as described (31). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS (complete medium) as described (31). TBHP was added to Earle's balanced salt solution (EBSS), and cultures were treated with TBHP for 40 min in EBSS. Following treatment, cells were washed with phosphatebuffered saline and then returned to complete medium. An ER stress response was produced by treating cells with DTTox (10 mM), A23187 (7 M), thapsigargin (0.3 g/ml), or tunicamycin (1.5 g/ml) as described (31).
Cell death was assessed by measuring the release of lactate dehydrogenase (LDH) into the medium (48). Formation of thiobarbituric acid-reactive substances (TBARS) was used as a measure of lipid peroxidation (49). In general, cells selected for neomycin resistance alone, i.e. pkNEO cells, may differ in their responsiveness to toxicant relative to wild type LLC-PK1 cells, an effect that is likely to be due to the selection process itself as noted before (50), and have slightly lower levels of lipid peroxidation after TBHP treatment (see Table II). Intracellular free Ca 2ϩ measurements were carried out using Fura-2 as described (31). Briefly, cells were treated with TBHP and then loaded with Fura-2 for 1 h, and the intracellular free Ca 2ϩ concentration was determined using a spectrofluorometer.
Significant differences (p Ͻ 0.05) were determined using a one-way analysis of variance followed by a Student-Newman-Keul's test for multiple comparisons. When analysis of variance was performed, letter designations are used in the figures and tables to indicate significant differences. Means designated with a common letter are not different, but different letter designations indicate a significant difference from other means. If more than one letter designation is shown, it indicates that the mean is not different from other means with either letter designation. As an example, means designated C and D are different from each other, but neither is significantly different from a mean designated CD.

RESULTS
Exposing cells to oxidants, including TBHP, increases cellular free Ca 2ϩ and causes lipid peroxidation (44,48). To establish the temporal relationship between these events and cell death in LLC-PK1 cells, we determined the time courses for all three events following TBHP exposure (Fig. 1). Lipid peroxidation increased within the initial TBHP treatment period (40 min) and continued to rise after TBHP removal. An increase in cellular free Ca 2ϩ followed the increase in lipid peroxidation, but both events preceded cell death, which occurred between 2 and 4 h. Adding antioxidants prevented the rise in Ca 2ϩ (data not shown) as reported previously (44), suggesting that lipid peroxidation contributes to the Ca 2ϩ increase. These temporal relationships are consistent with the proposal that oxidative stress preceded the rise in Ca 2ϩ .
We next examined the effect of conditioning cells with prior ER stress on TBHP-induced cell death. Prior treatment with four different ER stress inducers, all of which activate expression of ER stress proteins in these cells (31), prevented cell death caused by a subsequent TBHP pulse treatment (Table I). However, with the exception of A23187, none of these agents blocked lipid peroxidation, even though they blocked the rise in cellular Ca 2ϩ that was observed 2 h after TBHP treatment. The induction of grp78 and the protective effect of prior ER stress in iodoacetamide toxicity are disrupted in pkASgrp78 cells by forced expression of an 0.5-kilobase antisense grp78 construct (31). Therefore, we tested the TBHP sensitivity of three pkAS-grp78 clones relative to pkNEO cells that carry only the neomycin resistance marker (Fig. 2). pkASgrp78 cells were more sensitive to TBHP relative to pkNEO cells. Pretreating pkAS-grp78 cells with ER stress inducers was less effective in preventing TBHP toxicity compared with the pkNEO counter- was determined by spectrofluorometry in cells loaded for 1 h with Fura-2 immediately after TBHP treatment. The time delay due to Fura-2 loading is accounted for in the Ca 2ϩ data shown; thus, the first Ca 2ϩ data were collected 60 min after TBHP removal. Cell death was measured by determining the release of LDH into the medium. The cell death and TBARS data are the averages Ϯ the range of data from two separate experiments. Standard deviations on the TBARS data were smaller than the symbols for each data point. For the Ca 2ϩ determination, data points with standard deviation (Ϫ40, 75, 90,and 120 min) reflect the averages Ϯ the range of data from two separate experiments (n ϭ 2). Two additional points were collected, one in each of the two experiments (60 and 120 min), to further define the shoulders on the curve.
parts. A23187 was again the exception. Thus, the ability of prior ER stress to prevent TBHP toxicity depended on expression of ER chaperones.
Previously we found that pkCRT cells, which overexpress the ER chaperone and calcium binding protein calreticulin, were also resistant to iodoacetamide damage (31). Therefore, we evaluated their response to TBHP toxicity (Table II). pkCRT cells were also less sensitive to TBHP compared with pkNEO cells, and the protection correlated with maintenance of low intracellular free Ca 2ϩ 2 h after TBHP treatment, but at the same time, lipid peroxidation was unaffected. Therefore, increasing the ER chaperone content by two separate mechanisms, prior ER stress or forced expression of calreticulin, attenuated cell death and the increase in cellular free Ca 2ϩ after TBHP treatment without affecting lipid peroxidation.
If Ca 2ϩ is important in TBHP toxicity, then loading cells with EGTA should prevent TBHP toxicity. However, EGTA may chelate iron, thus making it difficult to separate effects on Ca 2ϩ from interference with iron-dependent Fenton generation of free radicals (51,52). To get around this problem, we exploited the time difference between lipid peroxidation (early) and the increase in Ca 2ϩ (later). Cells were treated with TBHP, then washed, and returned to medium containing the cell-permeable form of EGTA, EGTA-AM. Under these conditions, the majority of the peroxidation had already occurred; yet adding EGTA-AM still prevented cell death (Table III). There was only a modest, albeit significant, effect on lipid peroxidation, an effect that is probably associated with the decrease in cell death because in these experiments lipid peroxidation was measured at 6 h, a time when maximal cell death had already occurred. Thus the EGTA experiments further support a role for Ca 2ϩ in TBHPinduced cell death. DISCUSSION These studies provide the first evidence that ER stress proteins protect cells from organic oxidants and provide new insights into the roles of lipid peroxidation and intracellular free Ca 2ϩ and ER Ca 2ϩ handling in cell killing. When taken in context with prior literature (22,31,37), it appears that the role of ER stress proteins in cytoprotection can now be generalized to several distinct classes of toxicants and multiple pathways of cell death. In addition, preventing Ca 2ϩ disturbances during cell injury may be a general mechanism underlying the ability of the ER to protect against toxicants. This hypothesis is in general agreement with the observations that ER stress protects against Ca 2ϩ ionophore toxicity and that the stressed ER is able to accumulate an increased load of Ca 2ϩ (21,30,40,53). Notably, ER stress also prevents apoptosis induced by iodoacetamide and thapsigargin, 2 indicating that protection by ER stress can be extended to multiple pathways of cell death.
By inducing ER stress proteins, we were also able to dissociate lipid peroxidation from TBHP toxicity by preventing the TABLE I ER stress blocks cell death and Ca 2ϩ perturbations but not lipid peroxidation following TBHP treatment Control cells or cells conditioned with DTTox, A23187, thapsigargin, or tunicamycin to cause ER stress were treated with TBHP (1 mM) for 40 min, then washed, and returned to DMEM plus 10% FBS. Intracellular free Ca 2ϩ was determined 2 h after removal of TBHP. TBARS were measured before washing the cells and then again 2 h later as an index of lipid peroxidation. The values shown are nmol/well from cells cultured in 6-well dishes and are the sum of the TBARS that accumulated during the TBHP treatment and that which accumulated during the 2-h recovery period (see Fig. 1). Cell death was determined by measuring LDH release 6 h after removal of TBHP. There was no LDH release during the 40-min TBHP treatment. The data represent the means Ϯ S.D. of data from three independent experiments (n ϭ 3). Significant differences (p Ͻ 0.05) among the means were determined by a one-way analysis of variance with multiple comparisons using the Student-Neuman-Keul's test and are indicated by different letters (see "Materials and Methods").
pkASgrp78 cells are more sensitive to TBHP and do not develop tolerance. The pkASgrp78 cells, which express a 0.5-kilobase antisense grp78 construct, and their pkNEO counterparts were treated with EBSS or TBHP (1 mM) in EBSS for 40 min with or without prior induction of ER stress with A23187, the oxidized form of dithiothreitol (DTTox), tunicamycin (TUNC),or thapsigargin (THAPS) as described under "Materials and Methods." Cells were returned to DMEM with 10% FBS after TBHP cell death was determined 6 h later. The data are the means Ϯ S.D. of data collected from three separate pkNEO clones and three separate pkASgrp78 clones of LLC-PK1 cells (n ϭ 3). Significant differences (p Ͻ 0.05) among means within the pkNEO or pkAS-grp78 treatment groups were determined separately by a one-way analysis of variance with multiple comparisons using the Student-Neuman Keul's test and are indicated by different letters (uppercase for pkNEO cells, and lowercase for pkASgrp78 cells; see "Materials and Methods"). However, the pkASgrp78 cells were significantly more sensitive (p Ͻ 0.05) to TBHP treatment compared with the pkNEO cells when all the means from both the pkNEO and pkASgrp78 treatment groups were analyzed together by a one-way analysis of variance (not shown).

TABLE II
Overexpression of calreticulin blocks cell death and Ca 2ϩ disturbances but not lipid peroxidation Three separate clones of pkCRT cells and pkNEO cells were treated with TBHP (1 mM) for 40 min. Immediately following treatment, cells were washed and returned to DMEM plus 10% FBS. Intracellular free Ca 2ϩ was determined 2 h after washing the cells. TBARS, as a measure of lipid peroxidation, and LDH release, as a measure of cell death, were determined as indicated under "Materials and Methods" and the legend to Table I. In general, TBARS levels are lower after TBHP treatment in cells that have been selected for neomycin resistance (see "Materials and Methods"). The data are the means Ϯ S.D. of the values for each parameter determined for the three individual clones (n ϭ 3). Significant differences were determined using a one-way analysis of variance with the Student-Neuman-Keul's test for multiple comparisons (p Ͻ 0.05) and are indicated by different letters (see "Materials and Methods"). Ca 2ϩ disturbance. Importantly, this was accomplished without pharmacological agents, which have ancillary effects, such as iron chelation or antioxidant properties (52). TBHP causes an early accumulation of oxidized glutathione, due to TBHP metabolism (51), and lipid peroxidation, due to radical formation, both of which precede the increase in cellular free Ca 2ϩ (44,47,48,54). Both TBHP metabolism to radical species and/or accumulation of oxidized glutathione can damage Ca 2ϩ -ATPases in the plasma membrane and the endoplasmic reticulum, disabling the major cellular Ca 2ϩ buffering systems and allowing intracellular Ca 2ϩ to increase (34, 36, 54 -57). Thus, oxidative stress and lipid peroxidation are upstream events that initiate the rise in Ca 2ϩ . Because an ER stress response blocks the Ca 2ϩ increase downstream of the oxidant stress, cell death is prevented without an effect on lipid peroxidation caused by TBHP. Although we do not know the mechanisms coupling increased Ca 2ϩ to cell death, there are several clear possibilities including induction of a permeability transition in the mitochondrial inner membrane and collapse of the membrane potential (45, 46, 58 -60), activation of degradative enzymes such as phospholipases and proteases (10), and further stimulation of oxidant production (61,62), all of which occur after TBHP treatment. This mechanism could differ with much higher concentrations of TBHP or in other cell types but provides a reasonable explanation for the observations reported here and elsewhere. Importantly, ER stress fits neatly into this model because ER chaperones contribute to Ca 2ϩ buffering by the ER (21), the major intracellular Ca 2ϩ storage site. This mechanism resembles the model suggested for iodoacetamide injury with an important exception; after iodoacetamide treatment, lipid peroxidation is a late event relative to the increase in Ca 2ϩ and is blocked by prior ER stress (31,48). Unlike TBHP, iodoacetamide is not an oxidant per se but induces lipid peroxidation after depletion of glutathione, loss of protein thiols, and increased cellular free Ca 2ϩ (31,48,63). Consequently, ER stress blocks lipid peroxidation after iodoacetamide treatment because it prevents the rise in Ca 2ϩ , that is, in conjunction with the loss of glutathione, necessary to cause lipid peroxidation. Thus, in the iodoacetamide and TBHP models different primary events initiate the Ca 2ϩ increase, i.e. thiol depletion in the case of iodoacetamide and free radical production in the case of TBHP, accounting for the difference in the effect of ER stress on lipid peroxidation in the two models.
To our knowledge, this is the first report that calreticulin plays a role in protection against oxidant toxicity. Cells that overexpress calreticulin have increased ER Ca 2ϩ buffering capacity and/or resist Ca 2ϩ toxicity (18 -20, 31). Thus, the protective effect of calreticulin may be due to better ER Ca 2ϩ buffering, decreased Ca 2ϩ release, or an indirect mechanism involving cooperation between Ca 2ϩ uptake by the ER and/or extrusion across the plasma membrane. Elucidating the mechanism whereby calreticulin expression prevents Ca 2ϩ disturbances during injury may provide novel insights into general mechanisms of cellular Ca 2ϩ handling during stress. In this regard, it is interesting to note that calreticulin is also induced by thapsigargin treatment and protects cells from thapsigargin-induced apoptosis. 3 The results with A23187 are also worth noting. A23187, ionomycin, and thapsigargin are all used to perturb cellular Ca 2ϩ and induce ER stress. Clearly, A23187 differs from the other agents used in this study because treatment with A23187 prevented lipid peroxidation and protected pkASgrp78 cells. If Ca 2ϩ ionophore toxicity induces oxidative stress, A23187 treatment may also induce cellular antioxidant defense proteins, in addition to activating ER stress response genes. Because antioxidants block TBHP-induced cell death in LLC-PK1 cells (48), activating antioxidant defense systems would be expected to protect cells.
In conclusion, these studies highlight the importance of Ca 2ϩ and the ER stress response in oxidant-induced cell injury. The results could have broad implications in our understanding of how cells exploit an ER stress response to prevent cell injury. For example, a defect in this type of endogenous protective response could, in conjunction with defects in other antioxidant defense systems or an increase in oxidative stress (1-4), predispose cells and organs to injury. In this regard, it has already been shown that cells from aging rats are less able to upregulate expression of heat shock proteins relative to their younger counterparts (64). Given the importance of oxidants in human disease, it would seem that further investigation of the role of molecular stress responses in regulating oxidative injury is warranted.  (50 M). TBARS were determined as noted in the legend to Table I with the exception that cells were allowed to recover for 6 h before the second TBARS determination. Cell death was measured as described in the legend to Table I. The data represent a summary of three independent experiments (n ϭ 3). Significant differences were determined by a one-way analysis of variance with multiple comparisons using the Student-Neuman-Keul's test for multiple comparisons (p Ͻ 0.05) and are indicated by different letters (see "Materials and Methods").