Stress- and Growth-Related Gene Expression AreIndependent of Chemical-Induced Prostaglandin E2

Synthesis in Renal Epithelial Cells

Kelly M. Towndrow, Jozef J. W. M. Mertens,† Jeongmi K. Jeong,‡Thomas J. Weber,§ Terrence J. Monks, and Serrine S. Lau*

Division of Pharmacology and Toxicology, College of Pharmacy, The University of Texas at Austin,Austin, Texas 78712

Received September 9, 1999

Cellular stress can initiate prostaglandin (PG) biosynthesis which, through changes in geneexpression, can modulate cellular functions, including cell growth. PGA2, a metabolite of PGE2,induces the expression of stress response genes, including gadd153 and hsp70, in HeLa cellsand human diploid fibroblasts. PGs, gadd153, and hsp70 expression are also influenced bythe cellular redox status. Polyphenolic glutathione conjugates retain the ability to redox cycle,with the concomitant generation of reactive oxygen species. One such conjugate, 2,3,5-tris-(glutathion-S-yl)hydroquinone (TGHQ), is a potent nephrotoxic and nephrocarcinogenicmetabolite of the nephrocarcinogen, hydroquinone. We therefore investigated the effects ofTGHQ on PGE2 synthesis and gene expression in a renal proximal tubular epithelial cell line(LLC-PK1). TGHQ (200 µM, 2 h) increases PGE2 synthesis (2-3-fold) in LLC-PK1 cells withonly minor (5%) reductions in cell viability. This response is toxicant-specific, since anotherproximal tubular toxicant, S-(1,2-dichlorovinyl)-L-cysteine (DCVC), stimulates PGE2 synthesisonly after massive (68%) reductions in cell viability. Consistent with the ability of TGHQ togenerate an oxidative stress, both deferoxamine mesylate and catalase protect LLC-PK1 cellsfrom TGHQ-mediated cytotoxicity. Only catalase, however, completely blocks TGHQ-mediatedPGE2 synthesis, implying a major role for hydrogen peroxide in this response. TGHQ inducesthe early (60 min) expression of gadd153 and hsp70. However, while inhibition of cyclooxygenasewith aspirin prevents TGHQ-induced PGE2 synthesis, it does not affect TGHQ-mediatedinduction of gadd153 or hsp70 expression. In contrast, a stable PGE2 analogue, 11-deoxy-16,-16-dimethyl-PGE2 (DDM-PGE2), which protects LLC-PK1 cells against TGHQ-mediatedcytotoxicity, modestly elevates the levels of gadd153 and hsp70 expression. In addition, catalaseand, to a lesser extent, deferoxamine mesylate block TGHQ-induced gene expression. Therefore,although TGHQ-induced generation of reactive oxygen species is required for PGE2 synthesisand stress gene expression, acute TGHQ-mediated increases in gadd153 and hsp70 mRNAlevels are independent of PGE2 synthesis.

Introduction

Prostaglandins (PGs)1 modulate a wide variety ofcellular functions, including gene expression, growth, anddifferentiation. In particular, the cyclopentenone pros-taglandins of the A and J series exhibit growth inhibitory

and antitumor activities (1-5). Induction of early stressresponse genes, such as the growth arrest and DNAdamage inducible gene 153 (gadd153), heat shock protein70 (hsp70), c-fos, and Egr-1, is modulated by PGA2, adehydration product of PGE2, in human diploid fibro-blasts (6). Similarly, PGA2 elevates the levels of gadd153and heat shock protein mRNA expression in HeLa cells(3, 7, 8). Prevention or reduction of chemically inducedliver damage by PGs is also well documented (9-11). Forexample, 16,16-dimethyl-PGE2, a stable synthetic PGE2

analogue, protects rats against mild carbon tetrachloride-induced liver and kidney damage (12). Renal proximaltubular anoxia is also associated with increases inphospholipase A2 activity, preceded by arachidonic acidrelease, and the breakdown of membrane phospholipids(13). Misoprostol, a stable PGE1 analogue, is cytoprotec-tive against renal ischemic and mercuric chloride-inducedinjury in male Sprague-Dawley rats (14). However, themechanism(s) of PG-mediated cytoprotection remainsunclear.

Alterations in blood flow, increases in membranestability, changes in toxicant metabolism, and enhance-

* To whom all correspondence should be addressed: Division ofPharmacology and Toxicology, College of Pharmacy, University ofTexas at Austin, Austin, TX 78712. Telephone: (512) 471-5190. Fax:(512) 471-5002. E-mail: slau@mail.utexas.edu.

† Present address: WIL Research Laboratories, Inc., 1407 GeorgeRd., Ashland, OH 44805.

‡ Present address: Laboratory of Cardiovascular Disease Research,National Institute of Health, Rebublic of Korea, #5 Nokbun-dong,Eunpyung-gu, Seoul 122-701, Korea.

§ Present address: Pacific Northwest National Laboratory, 902Batelle Blvd., P7-56, Richland, WA 99352.

1 Abbreviations: AA, arachidonic acid; CAT, catalase; DFX, defer-oxamine mesylate; DDM-PGE2, 11-deoxy-16,16-dimethyl-prostaglandinE2; DCVC, S-(1,2-dichlorovinyl)-L-cysteine; DMEM, Dulbecco’s modifiedEagle’s medium; FBS, fetal bovine serum; GSH, glutathione; gadd,growth arrest and DNA damage inducible gene; hsp, heat shock proteingene; HEPES, N-(2-hydroxyethyl)piperazine-N′-2-ethanesulfonic acid;PG, prostaglandin; PKC, protein kinase C; ROS, reactive oxygenspecies; TX, thromboxane; TGHQ, 2,3,5-tris-(glutathion-S-yl)hydro-quinone.

111Chem. Res. Toxicol. 2000, 13, 111-117

10.1021/tx990160s CCC: $19.00 © 2000 American Chemical SocietyPublished on Web 02/02/2000

ment of a tissue’s regenerative capacity by PGs maycontribute to PG-induced cytoprotection (10, 15). Ad-ditionally, PGs protect isolated hepatic and renal cellsagainst carbon tetrachloride- and hypoxic-induced cyto-toxicity, respectively, suggesting that cytoprotection oc-curs at the cellular level (14, 16). Consistent with theselatter observations, prolonged exposure (8-24 h) to bothPGE2 and 11-deoxy-16,16-dimethyl-PGE2 (DDM-PGE2),a stable PGE2 analogue, protects renal proximal tubuleepithelial cells (LLC-PK1) against 2,3,5-tris(glutathion-S-yl)hydroquinone (TGHQ)-induced cytotoxicity (17). Cy-toprotection appears to be mediated via a receptor whichis pharmacologically distinct from known PGE receptor(EP) subtypes, and is associated with a protein kinase C(PKC)-related signal transduction pathway (17).

In addition to an acute stimulation of PG synthesis,oxidative stress can lead to the induction of early stressresponse genes, such as gadd153 and hsp70 (18, 19).Gadd153 is related to the CCAAT/enhancer-bindingprotein (C/EBP) family of transcription factors, andalthough capable of dimerizing with C/EBP transcriptionfactors, it does not possess intrinsic DNA binding activi-ties (20). Gadd153 is therefore considered a negativeregulator of C/EBP transcription factors and is associatedwith growth arrest (20). Heat shock proteins (HSPs)constitute a large family of stress response genes. HSPsare constitutively expressed, in addition to being in-duced by heat and other stressors (21). The abundanceof HSPs is coupled to their multiple functions, includingprotein chaperoning across membranes, participation inprotein refolding associated with degradation, and pro-tection against oxidative injury (21-23). Of particularinterest is the protective role of HSPs against reactiveoxygen species (ROS) generated during inflammation(24).

TGHQ (see structure) is a redox active metabolite (25)of the nephrocarcinogen, hydroquinone (26, 27).

Administration of TGHQ to male rats causes renalproximal tubular necrosis, and doses of TGHQ as low as7.5 µmol/kg (iv) cause significant increases in blood ureanitrogen levels, enzymuria, and glucosuria (25). Further-more, TGHQ is nephrocarcinogenic (28), and it catalyzes8-oxodeoxyguanosine formation in vitro (29). Incubationof the supF gene with TGHQ, followed by transfectionand replication in human AD293 embryonic kidney cells,causes a significant increase in the mutation frequency(30). Both TGHQ (17) and S-(1,2-dichlorovinyl)-L-cysteine(DCVC; see structure) (31) cause cytotoxicity in LLC-PK1

cells, a proximal tubular epithelial cell line. Here, wehave investigated the ability of TGHQ and DCVC tostimulate cellular PGE2 synthesis, and compared theeffects of this acute in situ production of PGE2, with abrief exposure to exogenous DDM-PGE2 and PGE2, onstress gene expression in LLC-PK1 cells.

Experimental Procedures

Caution: TGHQ is nephrotoxic and nephrocarcinogenic inrats and must be handled with care. This compound shouldtherefore be handled with protective clothing and in a well-ventilated fume hood.

Chemicals and Reagents. TGHQ was synthesized andpurified as previously described (32). 11-Deoxy-16,16-dimethyl-PGE2 (DDM-PGE2), PGE2, and 5,8,11,14-eicosatetraenoic acid[arachidonic acid (AA)] were purchased from Caymen Chemical(Ann Arbor, MI). Plasmid pBluescript containing a 600 bpfragment of cDNA of gadd153 was kindly provided by N. J.Holbrook (National Institute on Aging, Gerontology ResearchCenter, Bethesda, MD). Plasmid pAT153 containing a 2.3 kbfragment of hsp70 was obtained from American Type CultureCollection. S. Fisher (M. D. Anderson Cancer Center, Houston,TX) provided 7S ribosomal RNA cDNA. Procedures for cDNAprobe preparation have been described elsewhere (19). S-(1,2-Dichlorovinyl)-L-cysteine (DCVC) was a generous gift from J.L. Stevens (University of Vermont, Burlington, VT).

Cell Culture and Treatment Conditions. The New Hamp-shire minipig-derived renal proximal tubular epithelial cell line(LLC-PK1) was purchased from the American Type CultureCollection (CL101, passage 185). Cells were maintained inDulbecco’s modified Eagle’s medium with 4.5 g/L D-glucose(DMEM; GIBCO BRL, Grand Island, NY), supplemented with10% fetal bovine serum (FBS; JRH Bioscience, Lenexa, KS) at37 °C in a humidified atmosphere containing 5% CO2. Culturedishes were seeded at starting cell densities of 2 × 105 cells/well for 24-well plates, 9 × 105 cells/well for 6-well plates, and2 × 106 cells/100 mm dishes, and experiments were conductedon cells grown for 7-10 days (approximately 4-7 days post-confluent) with a medium change on day 5. Experiments withTGHQ were conducted in DMEM with 25 mM HEPES (pH 7.4).DCVC treatments were conducted in Earle’s balanced saltsolution [EBSS; 5.4 mM KCl, 1.7 mM MgSO4, 106.4 mM NaCl,26.2 mM NaHCO3, 1 mM NaH2PO4, 5.6 mM D-(+)-glucose, 25mM HEPES, and 1.8 mM CaCl2 (pH 7.4)]. DMEM with 25 mMHEPES supplemented with 10% FBS was used in gene expres-sion studies with DDM-PGE2, according to the PG-mediatedcytoprotection protocol previously described (17). In some cases,cells were pretreated with 20 µM arachidonic acid for 15 minprior to and during both TGHQ and DCVC treatment, toincrease the sensitivity of the medium PGE2 measurements. A0.1 mM aspirin pretreatment (30 min) prior to TGHQ exposure(200 µM, 2 h) was used to inhibit cyclooxygenase activity andhad no effect on cell viability with this dosing regimen. Viabilitywas assessed using the lysosomal neutral red accumulationassay (33).

Quantitation of Prostaglandin E2. Following exposure ofcells to TGHQ, media were immediately removed, frozen on dryice, and stored at -70 °C until further analysis. A PGE2

125Iradioimmunoassay kit (NEN Du Pont, Boston, MA) was usedto quantitate medium PGE2 levels. LLC-PK1 cells produce lowconstitutive levels of PGE2 (34, 35). Commercially availablePGE2

125I radioimmunoassay kits, which allow the directmeasurement of PGE2 as low as 1 pg, offer increased sensitivitycompared to the hepatic radioreceptor assay of Lifschitz et al.,which requires an extraction step with a 75% recovery efficiency,and has a limit of sensitivity to only 62 pg (34-36). Lifschitz(34, 35) suggested that LLC-PK1 cells produce little or noconstitutive PGE2, possibly due to a defect in cyclooxygenaseactivity, and that levels of PGE2 did not increase followingtreatment of LLC-PK1 cells with either calcium ionophore(A23187), butyrate, or arachidonic acid. However, Roszinski etal. (37) later reported the ability to detect elevations in the levelsof PGE2 release utilizing the radioimmunoassay followingcalcium ionophore (A23187) treatment and arachidonic acidsupplementation in LLC-PK1 cells.

Evaluation of Gene Expression by Northern Blot Analy-sis. Cells were treated as described above in a 10 mL culturevolume and were subsequently scraped and homogenized in 3-6mL of RNAzol or UltraSpec RNA reagent (Biotecx Laboratories,

112 Chem. Res. Toxicol., Vol. 13, No. 2, 2000 Towndrow et al.

Inc., Houston, TX) per 100 mm Petri dish. Isolation of total RNA,electrophoresis, and hybridization conditions were as previouslydescribed (19).

Quantitation of Autoradiograms. Data from Northern blotanalysis were quantified by densitometric analysis (ImagingResearch Inc. or NIH Image public domain software) or with aPackard InstantImager Electronic Autoradiography System(Packard Instrument Co., Meriden, CT). All densitometric andimager count values (CPM) were normalized for loading with7S rRNA.

Statistics. Statistical significance (P < 0.05) was determinedwith one-way ANOVA followed by Student-Newman-Keulspost hoc analysis where appropriate.

ResultsEarly Responses to Cellular Injury Induced by

TGHQ. TGHQ is a potent renal proximal tubular toxi-cant in vivo (25) and in vitro (Figure 1A). Lysosomaluptake and accumulation of neutral red dye is a sensitiveand early indicator of chemical insult and generallyoccurs prior to membrane and mitochondrial damage(33). TGHQ (200 µM, 2 h) also induced a 2-3-foldincrease in PGE2 synthesis and release in LLC-PK1 cells(Figure 1B, -AA curve). Supplementing cells with arachi-donic acid (20 µM, AA curve) for 15 min prior to TGHQexposure enhanced the sensitivity of PGE2 measurements(Figure 1B) without affecting cell viability (data notshown). Therefore, cells were supplemented with AA inall subsequent experiments. Increases in PGE2 synthesisoccurred as early as 30 min following TGHQ exposure(data not shown). Interestingly, a 2 h exposure to TGHQ(200 µM), while only causing a 5% reduction in cellviability, produced a 2-3-fold increase in PGE2 levels(Figure 1A). In contrast to TGHQ, another proximaltubular toxicant with a differing toxicological mechanismof action, S-(1,2-dichlorovinyl)-L-cysteine (DCVC), onlyelevated PGE2 synthesis in LLC-PK1 cells after markedreductions in cell viability (Figure 2). A 10% reductionin cell viability by DCVC (500 µM, 3 h) did not alter PGE2

synthesis (Figure 2). Indeed, a 3-4-fold increase in PGE2

levels occurred only after a 9 h exposure of LLC-PK1 cellsto 500 µM DCVC, which was accompanied by a 68%reduction in cell viability (Figure 2).

Role of ROS in TGHQ-Mediated Cytotoxicity andPGE2 Synthesis. Polyphenolic glutathione (GSH) con-jugates retain the ability to redox cycle, and the genera-tion of an oxidative stress is implicated in quinol thioether-mediated cytotoxicity (38). ROS can stimulate arachidonicacid metabolism, leading to the production of PGs (39-41). Cotreatment of LLC-PK1 cells with catalase (10units/mL) completely protected against TGHQ-inducedcytotoxicity (Figure 3A) and blocked PGE2 synthesis fromTGHQ exposure (Figure 4A), suggesting a role for hy-drogen peroxide in both responses. However, althoughchelation of iron with deferoxamine mesylate (DFX, 10mM), which presumably prevents hydroxyl radical for-mation, also diminished TGHQ-induced cytotoxicity (Fig-ure 3B), it did not significantly prevent TGHQ-mediatedPGE2 synthesis (Figure 4B).

Effects of ROS and Prostaglandin E2 on gadd153and hsp70 Gene Expression. Cellular stress, includingDNA damage and oxidative stress, induces the expressionof certain stress response genes, such as gadd153 and

Figure 1. TGHQ-induced cytotoxicity (9, A) and PGE2 syn-thesis (O, A and B) and the effects of arachidonic acid (AA)supplementation on TGHQ-induced PGE2 synthesis (B) in LLC-PK1 cells. Cells [except for the without AA curve (b) in panelB] were supplemented with 20 µM AA (15 min) prior to 0, 200,500, and 1000 µM TGHQ exposure for 120 min in 6-well plates(1 mL volumes). Medium was collected, frozen immediately, andstored at -70 °C until PGE2 analysis by radioimmunoassay.Viability was assessed using the neutral red lysosomal ac-cumulation assay. Mean PGE2 control levels were 7 ( 2 (withoutAA) and 30 ( 5 pg/mL (with AA). PGE2 and cytotoxicity dataare expressed as means ( SE (n ) 3) and are representative ofthree experiments producing similar results. Asterisks andsecants denote values that are significantly different from theirrespective control values at P < 0.05.

Figure 2. Time-dependent effects of DCVC on viability (9) andPGE2 synthesis (O) in LLC-PK1 cells. Cells were supplementedwith 20 µM AA (15 min) prior to 0, 3, 5, 9, and 12 h exposure to500 µM DCVC (1 mL volumes) in 6-well plates. Medium wascollected, frozen immediately, and stored at -70 °C until PGE2analysis by radioimmunoassay. Viability was assessed using theneutral red lysosomal accumulation assay. Mean PGE2 controllevels at 0 h were 127 ( 9 pg/mL. PGE2 and cytotoxicity dataare expressed as means ( SE (n ) 3) and are representative oftwo experiments producing similiar results. Asterisks andsecants denote values that are significantly different from theirrespective control values at P < 0.05.

Figure 3. Effects of catalase (CAT; A) and deferoxaminemesylate (DFX; B) on cell viability after exposure of LLC-PK1cells to TGHQ. LLC-PK1 cells were either pre- (1 h) andcotreated with 10 mM DFX or cotreated with 10 units/mLcatalase and 400 µM TGHQ for 2 h in 24-well plates (0.5 mLvolumes). Viability was assessed using the neutral red lysosomalaccumulation assay. Data are expressed as the means ( SE (n) 4 for CAT; n ) 3 for DFX) and are representative of threeexperiments producing similar results. Asterisks denote valuesthat are significantly different from control values at P < 0.05.Secants denote values that are significantly different fromTGHQ values at P < 0.05.

Toxicant-Induced PGE2 Synthesis and Gene Expression Chem. Res. Toxicol., Vol. 13, No. 2, 2000 113

hsp70 (42-45). TGHQ dramatically induced the expres-sion of both gadd153 and hsp70 mRNA, in a dose- andtime-dependent manner (Figures 5 and 6). Induction ofboth genes occurred as early as 60 min after treatmentwith TGHQ, and was maximal after 120 min (Figure 6).A 120 min exposure to 200 µM TGHQ produced an ∼20-fold increase in gadd153 mRNA and an 8-9-fold increasein hsp70 mRNA levels (Figure 5). Catalase cotreatment(10 units/mL) completely prevented TGHQ-induced el-evations in gadd153 and hsp70 mRNA (Figure 7). WhileDFX partially prevented gadd153 and hsp70 expressionfollowing TGHQ exposure, inhibition of hsp70 geneexpression may be minimized by the stimulatory proper-ties of DFX alone (Figure 7).

To probe the role of PGE2 in the acute expression ofgadd153 and hsp70, we investigated the effects of DDM-PGE2 and PGE2 on gadd153 and hsp70 stress geneexpression in LLC-PK1 cells. Inhibition of cyclooxygenaseby aspirin, while completely blocking TGHQ-mediatedelevations in PGE2 levels in LLC-PK1 cells (Figure 8),

had no effect on TGHQ-induced gadd153 and hsp70 geneexpression (Figure 9). Furthermore, AA supplementationalso had no effect on gene expression (data not shown).

Figure 4. Effects of catalase (CAT; A) and deferoxaminemesylate (DFX; B) on PGE2 synthesis after exposure of LLC-PK1 cells to TGHQ. LLC-PK1 cells were either pre- (1 h) andcotreated with 10 mM DFX or cotreated with 10 units/mLcatalase and 400 µM TGHQ for 2 h. Cells were supplementedwith 20 µM AA (15 min) prior to DFX, CAT, and TGHQexposure. After treatment, medium was collected, frozen im-mediately, and stored at -70 °C until PGE2 analysis byradioimmunoassay. Mean PGE2 control levels were 56 ( 7 (CAT)and 61 ( 4 pg/mL (DFX). Data are expressed as the means (SE (n ) 6) and are representative of at least three experimentsproducing similar results. Asterisks denote values that aresignificantly different from control values at P < 0.05. Thesecant denotes a value that is significantly different from TGHQvalues at P < 0.05.

Figure 5. Concentration-dependent effects of TGHQ on gadd153and hsp70 mRNA expression in LLC-PK1 cells. Cells wereexposed to varying concentrations of TGHQ for 120 min. TotalRNA was isolated and quantified by Northern blot analysis.Data were normalized for loading with 7S rRNA and arerepresentative of at least two experiments producing similarresults.

Figure 6. Time-dependent effects of TGHQ on gadd153 andhsp70 mRNA expression in LLC-PK1 cells. Cells were exposedto 200 µM TGHQ for 0, 30, 60, 120, and 300 min. Total RNAwas isolated and quantified by Northern blot analysis. Datawere normalized for loading with 7S rRNA and are representa-tive of at least two experiments producing similar results.

Figure 7. Effects of catalase and deferoxamine mesylate ongadd153 (A) and hsp70 (B) mRNA expression in LLC-PK1 cellsfollowing TGHQ exposure. LLC-PK1 cells were either pre- (1 h)and cotreated with 10 mM DFX or cotreated with 10 units/mLcatalase and 200 µM TGHQ for 2 h. Total RNA was isolatedand quantified by Northern blot analysis. Data were normalizedfor loading with 7S rRNA and are representative of at least twoexperiments producing similar results.

Figure 8. Effects of cyclooxygenase inhibition by aspirin onTGHQ-mediated PGE2 synthesis in LLC-PK1 cells. After cellswere pretreated with 0.1 mM aspirin (30 min) prior to 200 µMTGHQ (2 h) exposure, medium was collected, frozen im-mediately, and stored at -70 °C until PGE2 analysis byradioimmunoassay. Cells were supplemented with 20 µM AA(15 min) prior to aspirin and TGHQ exposure. Mean PGE2control levels were 97 ( 15 pg/mL. Data are expressed as means( SE of three experiments. Asterisks denote values that aresignificantly different from control values at P < 0.05. Thesecant denotes a value that is significantly different from TGHQvalues at P < 0.05.

114 Chem. Res. Toxicol., Vol. 13, No. 2, 2000 Towndrow et al.

We have previously shown that PGE2 and DDM-PGE2

protect LLC-PK1 cells against TGHQ-induced cellulardamage (17). In contrast to the elevated in situ synthesisof cellular PGE2 following TGHQ exposure (Figure 1)which appeared to be insufficient to stimulate gadd153and hsp70 mRNA (Figure 9), exposure of LLC-PK1 cellsto 20 µM DDM-PGE2 (2 h) induced a 1.6- and 2.5-foldincrease in hsp70 and gadd153 mRNA, respectively(Figure 10). PGE2 (20 µM, 2 h) was much less effectiveat stimulating gadd153 (1.7-fold) and hsp70 (1.2-fold)gene expression (Figure 10). Interestingly, EP receptoragonists, such as sulprostone (EP1/EP3 agonist) and 17-phenyltrinor PGE2 (EP1 agonist), had no effect ongadd153 expression, but slightly stimulated hsp70 geneexpression, to a lesser extent than that of DDM-PGE2

(data not shown).

Discussion

Little is known about the role of endogenous PGE2 inthe acute cellular response to injury. We report here thatLLC-PK1 cells respond to chemical challenge by increas-ing PGE2 synthesis and release. Increases in PGE2 levelsfollowing exposure of LLC-PK1 cells to TGHQ occur priorto the onset of overt cytotoxicity (Figure 1). The biologicalreactivity of quinol thioethers lies in their ability to redoxcycle and to subsequently generate an oxidative stress(38). Deferoxamine mesylate, an iron chelator, and cata-lase, a scavenger of hydrogen peroxide, both preventTGHQ-mediated cytotoxicity in LLC-PK1 cells, support-ing a role for the formation of ROS in this response(Figure 3). ROS can alter arachidonic acid metabolism(39, 40, 46). Cotreatment of LLC-PK1 cells with catalasecompletely blocks TGHQ-mediated induction of PGE2

synthesis (Figure 4A), gadd153 and hsp70 gene expres-sion (Figure 7), and cytotoxicity (Figure 3A). However,whereas DFX protects LLC-PK1 cells against TGHQexposure (Figure 3B), it does not significantly preventincreases in PGE2 synthesis (Figure 4B) and only par-tially prevents TGHQ-induced gadd153 and hsp70 geneexpression (Figure 7). These data suggest that althoughhydrogen peroxide is essential for stimulating PGE2

synthesis and stress gene expression, both hydrogenperoxide and the subsequent formation of hydroxylradicals are required for TGHQ-induced cytotoxicity.Consistent with our findings, treatment of culturedhuman endothelial cells with hydrogen peroxide in-creases PGI2 and thromboxane A2 (TXA2) synthesis (39).Similarly, in rat peritoneal exudate cells, E3330 [(2E)-3-[5-(2,3-dimethoxy-6-methyl-1,4-benzoquinoyl)]-2-nonyl-2-proponoic acid], a quinone derivative, inhibits thesynthesis of TXB2 and LTB4, but induces PGE2 produc-tion (47).

Chemical toxicants, and other stressors, stimulate theexpression of early stress response genes, includinggadd153 and hsp70 (42-44). Although exogenously sup-plied PGs stimulate early stress response gene expres-sion, including gadd153 and hsp70 (3, 6-8), the role ofconstitutive PGE2 synthesis in gene expression is lessclear. Inductions of gadd153, hsp70, and PGE2 are earlyresponses to TGHQ exposure in LLC-PK1 cells, and occurprior to overt cytotoxicity (Figures 1 and 6). Similarly,LLC-PK1 cells undergo growth arrest following exposureto the quinol thioether 2-bromo-bis-(glutathion-S-yl)-hydroquinone, as indicated by formation of DNA single-strand breaks, a decreased level of DNA synthesis, andelevations in gadd153 expression (44). The ability of bothcatalase and deferoxamine to prevent quinol thioether-mediated gadd153 induction is consistent with the modusoperandi of quinol thioethers (44). However, TGHQ-mediated cellular PGE2 synthesis and gadd153 andhsp70 gene expression do not appear to be linked, becauseinhibition of cyclooxygenase with aspirin, while blockingPG synthesis (Figure 8), does not attentuate TGHQ-induced gene expression (Figure 9).

Stimulation of in situ production of PGE2 in LLC-PK1

cells appears to be toxicant-specific. Although TGHQstimulates PGE2 synthesis before the onset of overtcytotoxicity (Figure 1A), increases in the levels of PGE2

synthesis following DCVC exposure only occur after cellviability is compromised and may be the result of overtmembrane damage (Figure 2). The different responsesto TGHQ and DCVC probably relate to differences in the

Figure 9. Effects of cyclooxygenase inhibition by aspirin onTGHQ-mediated gadd153 and hsp70 mRNA expression in LLC-PK1 cells. After cells were pretreated with 0.1 mM aspirin (30min) and 20 µM arachidonic acid (AA, 15 min) prior to 200 µMTGHQ (2 h) exposure, total RNA was isolated and quantifiedby Northern blot analysis. Data were normalized for loadingwith 7S rRNA and are representative of at least two experi-ments producing similar results.

Figure 10. Effects of DDM-PGE2 and PGE2 exposure ongadd153 and hsp70 mRNA expression in LLC-PK1 cells. Cellswere exposed to either 20 µM DDM-PGE2 or 20 µM PGE2 for120 min. Total RNA was isolated and quantified by Northernblot analysis. Data were normalized for loading with 7S rRNAand are representative of at least two experiments producingsimilar results.

Toxicant-Induced PGE2 Synthesis and Gene Expression Chem. Res. Toxicol., Vol. 13, No. 2, 2000 115

mechanisms of action of these two chemicals. The me-tabolism of DCVC by cysteine conjugate â-lyase producesa 1,2-dichlorovinylthiol intermediate, which further rear-ranges to an electrophilic thioketene. Thioketenes arehighly reactive, and interact with cellular macromol-ecules, thereby initiating a series of events, includingmitochondrial dysfunction and lipid peroxidation, thateventually results in cell death (31, 48). Although TGHQpossesses the ability to arylate cellular constituents (49,50), it also retains the ability to redox cycle and generatean oxidative stress (38), which is likely the trigger forPGE2 synthesis (Figure 4A) and stress gene expression(Figure 7). In contrast to TGHQ, DCVC induces earlyelevations in gadd153 (2 h) and hsp70 (0.5 h) mRNAlevels (51, 52), but only induces PGE2 synthesis as a latecellular response (9 h) and only after significant reduc-tions in cell viability (Figure 2).

Studies aimed at investigating the effects of exog-enously added PGs often use concentrations that areusually 3-6 orders of magnitude higher than levelsachieved after stimulation of cellular PGE2 synthesis (1-4, 6-8). In fact, exogenously added concentrations ofPGE2 two orders of magnitude greater than the levels ofPGE2 synthesized in situ following exposure of LLC-PK1

cells to TGHQ (Figure 1) only minimally induced gadd153and hsp70 gene expression (Figure 10). Moreover, thecytoprotective PGE2 analogue, DDM-PGE2, modestlyelevates gadd153 and hsp70 expression when added toLLC-PK1 cultures (Figure 10). Taken together, these datademonstrate differences in the cellular response to exoge-nously supplied PGs and PGs synthesized constitutively.Although it has been shown that exogenously addedPGA2, a metabolite of PGE2, elevates gadd153 and hsp70mRNA levels in HeLa cells (3, 7), the inability of aspirinto prevent TGHQ-induced stimulation of gene expression(Figure 9) indicates that TGHQ-mediated PGE2 synthesisand stress gene expression are unrelated in LLC-PK1

cells. The differences in PG-mediated regulation of stressgene expression between HeLa and LLC-PK1 cells arelikely due to cell type selectivity and/or differences in thegrowth state of the cells.

PGs generally exert their cellular actions via G-protein-coupled receptors (53), and in some cases, multiple PGreceptors with opposite actions can be coexpressed in aparticular cell type (54). Coexpression of both stimulatoryand inhibitory PG receptors has recently been proposedas a possible model for homeostatic control of paracrineand autocrine signaling that enables the cell to bufferagainst locally high agonist concentrations (54). Forexample, PGE2 both stimulates and inhibits adenylcyclase activity in platelets via activation of differentprostanoid receptor subtypes, resulting in a slight netincrease in cAMP levels (55). Therefore, coexpression ofstimulatory and inhibitory prostanoid receptor subtypesand possible differences in prostanoid receptor subtypeaffinities and/or structural stability between DDM-PGE2

and PGE2 may contribute to the differences in the cellularresponse to exogenous and endogenous PGs we observed.Interestingly, certain eicosanoids, such as PGA, PGD,and PGJ2, can activate peroxisome proliferator-activatedreceptors (PPARs), which are nuclear hormone receptors,suggesting that PGs may also exert cellular effectsindependent of cell surface receptor activation (56).Consequently, DDM-PGE2-induced gadd153 and hsp70mRNA expression likely occurs via activation of cellsurface receptors. In contrast, stimulation of cellular

PGE2 synthesis may cause both cell surface and nuclearreceptor-dependent cellular responses.

In summary, LLC-PK1 cells respond to TGHQ-gener-ated ROS by increasing synthesis of PGE2 and byupregulating gadd153 and hsp70 gene expression. Theacute stimulation of PGE2 synthesis in LLC-PK1 cells istoxicant-specific, and despite the ability of exogenous PGsto mildly induce stress gene expression, acute toxicant-induced stimulation of PGE2 synthesis and stress geneexpression are not related.

Acknowledgment. This work was supported by NIHGrants GM56321 (to S.S.L.), ES77084 (NIEHS CenterGrant), and ES07247 (NIEHS Training Grant to K.M.T.).

References

(1) Ohno, K., Fujiwara, M., Fukushima, M., and Narumiya, S. (1986)Metabolic dehydration of prostaglandin E2 and cellular uptakeof the dehydration product: Correlation with prostaglandin E2-induced growth inhibition. Biochem. Biophys. Res. Commun. 139,808-815.

(2) Honn, K. V., and Marnett, L. J. (1985) Requirement of a reactiveR,â-unsaturated carbonyl for inhibition of tumor growth andinduction of differentiation by “A” series prostaglandins. Biochem.Biophys. Res. Commun. 129, 34-40.

(3) Choi, A. M. K., Fargnoli, J., Carlson, S. G., and Holbrook, N. J.(1992) Cell growth inhibition by prostaglandin A2 results inelevated expression of gadd153 mRNA. Exp. Cell Res. 199, 85-89.

(4) Narumiya, S., and Fukushima, M. (1985) ∆12-prostaglandin J2,an ultimate metabolite of prostaglandin D2 exerting cell growthinhibition. Biochem. Biophys. Res. Commun. 127, 739-745.

(5) Bregman, M. D., Funk, C., and Fukushima, M. (1986) Inhibitionof human melanoma growth by prostaglandin A, D, and Janalogues. Cancer Res. 46, 2740-2744.

(6) Choi, A. M. K., Tucker, R. W., Sarlson, S. G., Wiegand, G., andHolbrook, N. J. (1994) Calcium mediates expression of stress-response genes in prostaglandin A2-induced growth arrest. FASEBJ. 8, 1048-1054.

(7) Holbrook, N. J., Carlson, S. G., Choi, A. M. K., and Fargnoli, J.(1992) Induction of HSP70 gene expression by the antiprolifera-tive prostaglandin PGA2: A growth-dependent response mediatedby activation of heat shock transcription factor. Mol. Cell. Biol.12, 1528-1534.

(8) Ohno, K., Fukushima, M., Fujiwara, M., and Narumiya, S. (1988)Induction of 68,000-dalton heat shock proteins by cyclopentenoneprostaglandins. J. Biol. Chem. 263, 19764-19770.

(9) Stachura, J., Tarnawski, A., Ivey, K. J., Mach, T., Bogdal, J.,Szczudrawa, J., and Klimczyk, B. (1981) Prostaglandin protectionof carbon tertachloride-induced liver cell necrosis in the rat.Gastroenterology 81, 211-217.

(10) Ruwart, M. J. (1986) Protection of the liver against variousdamaging agents. In Biological Protection with Prostaglandins(Cohen, M. M., Ed.) pp 229-243, CRC Press, New York.

(11) Funck-Brentano, C., Tinel, M., Degott, C., Letteron, P., Babany,G., and Pessayre, D. (1984) Protective effect of 16,16-dimethylprostaglandin E2 on the hepatotoxicity of bromobenzene in mice.Biochem. Pharmacol. 33, 89-96.

(12) Ruwart, M. J., Rush, B. D., Friedle, N. M., Piper, R. C., and Kolaja,G. J. (1981) Protective effects of 16,16-dimethyl PGE2 on the liverand kidney. Prostaglandins 21 (Suppl.), 97-102.

(13) Portilla, D., Mandel, L. J., Bar-Sagi, D., and Millington, D. S.(1992) Anoxia induces phospholipase A2 activation in rabbit renalproximal tubules. Am. J. Physiol. 262, F354-F360.

(14) Paller, M. S., Manivel, J. C., Patten, M., and Barry, M. (1992)Prostaglandins protect kidneys against ischemic and toxic injuryby a cellular effect. Kidney Int. 42, 1345-1354.

(15) Masaki, N., Ohta, Y., Shirataki, H., Ogata, I., Hayashi, S.,Yamada, S., Hirata, K., Nagoshi, S., Mochida, S., Tomiya, T.,Ohno, A., Ohta, Y., and Fujiwara, K. (1992) Hepatocyte membranestabilization by prostaglandins E1 and E2: Favorable effects onrat liver injury. Gastroenterology 102, 572-576.

(16) Ruwart, M. J., Nichols, N. M., Hedeen, K., Rush, B. D., andStachura, J. (1985) 16,16-dimethyl PGE2 and fatty acids protecthepatocytes against CCl4-induced damage. In Vitro Cell. Dev. Biol.21, 451-452.

116 Chem. Res. Toxicol., Vol. 13, No. 2, 2000 Towndrow et al.

(17) Weber, T. J., Monks, T. J., and Lau, S. S. (1997) PGE2-mediatedcytoprotection in renal epithelial cells: Evidence for a pharma-cologically distinct receptor. Am. J. Physiol. 273, F507-F515.

(18) Aucoin, M. M., Barhoumi, R., Kochevar, D. T., Granger, H. J.,and Burghardt, R. C. (1995) Oxidative injury of coronary venularendothelial cells depletes intracellular glutathione and inducesHSP 70 mRNA. Am. J. Physiol. 268, H1651-H1658.

(19) Jeong, J. K., Stevens, J. L., Lau, S. S., and Monks, T. J. (1996)Quinone thioether-mediated DNA damage, growth arrest, andgadd153 expression in renal proximal tubular epithelial cells. Mol.Pharmacol. 50, 592-598.

(20) Ron, D., and Habener, J. F. (1992) CHOP, a novel developmentallyregulated nuclear protein that dimerizes with transcriptionfactors C/EBP and LAP and functions as a dominant-negativeinhibitor of gene transcription. Genes Dev. 6, 439-453.

(21) Feige, U., and Polla, B. S. (1994) Heat shock proteins: the hsp70family. Experientia 50, 979-986.

(22) Morimoto, R. I., and Santoro, M. G. (1998) Stress-inducibleresponses and heat shock proteins: New pharmacologic targetsfor cytoprotection. Nat. Biotechnol. 16, 833-838.

(23) Polla, B. S., Kantengwa, S., Francois, D., Salvioli, S., Franceschi,C., Marsac, C., and Cossarizza, A. (1996) Mitochondria aresensitive targets for the protective effects of heat shock againstoxidative injury. Proc. Natl. Acad. Sci. U.S.A. 93, 6458-6463.

(24) Jacquier-Sarlin, M. R., Fuller, K., Dinh-Xuan, A. T., Richard, M.J., and Polla, B. S. (1994) Protective effects of hsp70 in inflam-mation. Experientia 50, 1031-1038.

(25) Peters, M. M. C. G., Jones, T. W., Monks, T. J., and Lau, S. S.(1997) Cytotoxicity and cell-proliferation induced by the nephro-carcinogen hydroquinone and its nephrotoxic metabolite 2,3,5-tris(glutathion-S-yl)hydroquinone. Carcinogenesis 18, 2393-2401.

(26) Kari, F. W., Bucher, J., Eustis, S. L., Haseman, J. K., and Huff,J. E. (1992) Toxicity and carcinogenicity of hydroquinone inF344/N rats and B6C3F1 mice. Food Chem. Toxicol. 30, 737-747.

(27) Shibata, M. A., Hirose, A., Tanaka, H., Asakawa, E., Shirai, T.,and Ito, N. (1991) Induction of renal cell tumors in rats and mice,and enhancement of hepatocellular tumor development in miceafter long-term hydroquinone treatment. Jpn. J. Cancer Res. 82,1211-1219.

(28) Lau, S. S., Monks, T. J., Everitt, J. I., and Walker, C. W. (1998)A nephrotoxic metabolite of hydroquinone induces renal celltumorigenesis in Eker rats. Toxicologist 42, 317 (abstract).

(29) Lau, S. S., Peters, M. M. C. G., Kleiner, H. E., Canales, P. L.,and Monks, T. J. (1996) Linking the metabolism of hydroquinoneto its nephrotoxicity and nephrocarcinogenicity. In BiologicalIntermediates V (Snyder, R., et al., Eds.) pp 267-273, PlenumPress, New York.

(30) Jeong, J. K., Wogan, G. N., Lau, S. S., and Monks, T. J. (1999)Quinol-glutathione conjugate induced mutation spectra in thesupF gene-containing pSP189 plasmid, replicated in humanAD293 and bacterial MBL50 cells. Cancer Res. 59, 3641-3645.

(31) Stevens, J., Hayden, P., and Taylor, G. (1986) The role ofglutathione conjugate metabolism and cysteine conjugate â-lyasein the mechanism of S-cysteine conjugate toxicity in LLC-PK1cells. J. Biol. Chem. 261, 3325-3332.

(32) Lau, S. S., Hill, B. A., Highet, R. J., and Monks, T. J. (1988)Sequential oxidation and glutathione addition to 1,4-benzo-quinone: Correlation of toxicity with increased glutathionesubstitution. Mol. Pharmacol. 34, 829-836.

(33) Mertens, J. J. W. M., Gibson, N. W., Lau, S. S., and Monks, T. J.(1995) Reactive oxygen species and DNA damage in 2-bromo-(glutathion-S-yl)-hydroquinone-mediated cytotoxicity. Arch. Bio-chem. Biophys. 320, 51-58.

(34) Lifschitz, M. D. (1982) LLC-PK1 cells derived from pig kidneyshave a defect in cyclooxygenase. J. Biol. Chem. 257, 12611-12615.

(35) Lifschitz, M. D. (1985) Prostaglandins may mediate chlorideconcentration gradient across domes formed by MDCK1 cells. Am.J. Physiol. 250, F525-F531.

(36) Rosenblatt, S. G., Patak, R. V., and Lifschitz, M. D. (1978) Organicacid secretory pathway and urinary excretion of prostaglandin Ein the dog. Am. J. Physiol. 235, F473-F479.

(37) Roszinski, S., and Jelkmann, W. (1987) Effect of PO2 on prostag-landin E2 production in renal cell cultures. Respir. Physiol. 70,131-141.

(38) Monks, T. J., and Lau, S. S. (1992) Toxicology of quinone-thioethers. Crit. Rev. Toxicol. 22, 243-270.

(39) Schimke, I., Griesmacher, A., Weigel, G., Holzhutter, H. G., andMuller, M. M. (1992) Effects of reactive oxygen species oneicosanoid metabolism in human endothelial cells. Prostaglandins43, 281-292.

(40) Wolin, M. S., and Mohazzab-H., K. M. (1997) Mediation of signaltransduction by oxidants. In Oxidative Stress and the MolecularBiology of Antioxidant Defenses (Scandalios, J. G., Ed.) pp 21-48, Cold Spring Harbor Laboratory Press, Plainview, NY.

(41) Rao, G. N., Lassegue, B., Griendling, K. K., Alexander, R. W.,and Berk, B. C. (1993) Hydrogen peroxide-induced c-fos expressionis mediated by arachidonic acid release: role of protein kinaseC. Nucleic Acids Res. 21, 1259-1263.

(42) Fornace, A. J., Jr., Nebert, D. W., Hollander, C. M., Luethy, J.D., Papthanasiou, M., Fargnoli, J., and Holbrook, N. J. (1989)Mammalian genes coordinately regulated by growth arrest signalsand DNA-damaging agents. Mol. Cell. Biol. 9, 4196-4203.

(43) Abe, T., Konishi, T., Katoh, T., Hirano, H., Matsukuma, K.,Kashimura, M., and Higashi, K. (1994) Induction of heat shock70 mRNA by cadium is mediated by glutathione suppressive andnon-suppressive triggers. Biochim. Biophys. Acta 1201, 29-36.

(44) Jeong, J. K., Huang, Q., Lau, S. S., and Monks, T. J. (1997) Theresponse of renal tubular epithelial cells to physiologically andchemically induced growth arrest. J. Biol. Chem. 272, 7511-7518.

(45) Leuthy, J. D., Fargnoli, J., Park, J. S., Fornace, A. J., Jr., andHolbrook, N. J. (1990) Isolation and characterization of thehamster gadd153 gene: Activation of promoter activity by agentsthat damage DNA. J. Biol. Chem. 265, 16521-16526.

(46) Suzuki, Y. J., Forman, H. J., and Sevanian, A. (1997) Oxidantsas stimulators of signal transduction. Free Radical Biol. Med. 22,269-285.

(47) Nagakawa, J., Hishinuma, I., Hirota, K., Miyamoto, K., Ya-manaka, T., Yamatsu, I., and Katayama, K. (1992) Protectiveeffects of E3330, a novel quinone derivative, on galactosamine/tumor necrosis factor-R-induced hepatitis in mice. Eur. J. Phar-macol. 229, 63-67.

(48) Dekant, W., Vamvakas, S., and Anders, M. W. (1994) Formationand fate of nephrotoxic and cytotoxic glutathione S-conjugates:Cysteine conjugate â-lyase pathway. In Advances in Pharmacol-ogy, Volume 27: Conjugation-Dependent Carcinogenicity andToxicology of Foriegn Compounds (Anders, M. W., and Dekant,W., Eds.) pp 115-162, Academic Press, San Diego.

(49) Kleiner, H. E., Riveria, M. I., Pumford, N. R., Monks, T. J., andLau, S. S. (1998) Immunochemical detection of quinol-thioether-derived protein adducts. Chem. Res. Toxicol. 11, 1283-1290.

(50) Kleiner, H. E., Jones, T. W., Monks, T. J., and Lau, S. S. (1998)Immunochemical analysis of quinol-thioether-derived covalentprotein adducts in rodent species sensitive and resistant to quinol-thioether-mediated nephrotoxicity. Chem. Res. Toxicol. 11, 1291-1300.

(51) Chen, Q., Yu, K., Holbrook, N. J., and Stevens, J. L. (1992)Activation of the growth arrest and DNA damage-indicible genegadd153 by nephrotoxic cysteine conjugates and dithiothreitol.J. Biol. Chem. 267, 8207-8212.

(52) Chen, Q., Yu, K., and Stevens, J. L. (1992) Regulation of thecellular stress response by reactive electrophiles. The role ofcovalent binding and cellular thiols in transcriptional activationof the 70-kilodalton heat shock protein gene by nephrotoxiccysteine conjugates. J. Biol. Chem. 267, 24322-24327.

(53) Negishi, M., Sugimoto, Y., and Ichikawa, A. (1995) Molecularmechanisms of diverse actions of prostanoid receptors. Biochim.Biophys. Acta 1259, 109-120.

(54) Ashby, B. (1998) Co-expression of prostaglandin receptors withopposite effects. A model for homeostatic control of autocrine andparacrine signaling. Biochem. Pharmacol. 55, 239-246.

(55) Mao, G. F., Jin, J. G., Bastepe, M., Ortiz-Vega, S., and Ashby, B.(1996) Prostaglandin E2 both stimulates and inhibits adenylcyclase on platelets: Comparison of effects on cloned EP4 andEP3 prostaglandin receptor subtypes. Prostaglandins 52, 175-185.

(56) Yu, K., Bayona, W., Kallen, C. B., Harding, H. P., Ravera, C. P.,McMahon, G., Brown, M., and Lazar, M. A. (1995) Differentialactivation of peroxisome proliferator-activated receptors byeicosanoids. J. Biol. Chem. 270, 23975-23983.

TX990160S

Toxicant-Induced PGE2 Synthesis and Gene Expression Chem. Res. Toxicol., Vol. 13, No. 2, 2000 117