Toxicology xxx (2006) xxx–xxx

Cytotoxicity and cell signalling induced by continuous mildhyperthermia in freshly isolated mouse hepatocytes

Maria Joao Santos-Marques a,!, Felix Carvalho a,!, Carla Sousa a, Fernando Remiao a,Rui Vitorino b, Francisco Amado b, Rita Ferreira c, Jose Alberto Duarte c,

Maria de Lourdes Bastos a

a REQUIMTE, Toxicology Department, Faculty of Pharmacy, University of Porto, Rua Anıbal Cunha, 164, 4099-030 Porto, Portugalb Chemistry Department, University of Aveiro, 3810-123 Aveiro, Portugal

c Biochemistry Department, Faculty of Sport Sciences and Physical Education, University of Porto,Rua Dr. Placido Costa, 91, 4200-450 Porto, Portugal

Received 3 January 2006; received in revised form 20 March 2006; accepted 13 April 2006

Abstract

An increasing body of data has been demonstrating that mammalian cells have elaborate networks of molecular signalling incounteracting heat shock and in developing adaptation to oxidative stress to avoid cell death. However, the precise mechanismslinking heat shock, oxidative stress and cell survival/cell death mechanisms are not yet clearly understood. The purpose of thisstudy was thus to study the time course of hyperthermia-induced oxidative stress and cellular signalling through the activation ofheat shock factor 1 (HSF1) and heat shock protein 70 (HSP70), using freshly isolated mouse hepatocytes. The results accomplishedin this work demonstrated that mild continuous hyperthermia (41") leads to oxidative stress and loss of cellular viability in atime-dependent manner, with significant effects already observed at the first hour of incubation. These toxic effects developedconcomitantly with activation of HSF1 and emerged before the formation of HSP70 levels. Thus, although cell signalling wastriggered through the transcriptional activation of HSP70 via HSF1, this putative protective process did not modify the trend ofhepatotoxic effects mediated by this type of hyperthermic challenging.© 2006 Elsevier Ireland Ltd. All rights reserved.

Keywords: Hyperthermia; Mouse hepatocytes; Oxidative stress; Heat shock; Cell signalling

1. Introduction

Exposure of mammalian cells to high temperatures,either in short term (heat shock) or long term hyperther-mia, usually triggers several signalling pathways, somethat facilitate cell survival and others that initiate cell

! Corresponding authors. Tel.: +351 222078922;fax: +351 222003977.

E-mail addresses: mjsantosmarques@hotmail.com(M.J. Santos-Marques), felixdc@ff.up.pt (F. Carvalho).

death programs. The outcome of a cell’s exposure to heatshock may be either a development of a state of tolerance,if survival pathways are strong enough, or cell death, ifdeath pathways prevail (Gabai and Sherman, 2002).

The mechanisms responsible for causing cell deathafter thermal injury remain a matter of debate. Evenso, there is a consensus that heat shock may result inaltered generation of reactive oxygen species as well asalterations in intracellular antioxidant capacity. Mam-malian cells and tissues exposed to hyperthermia havebeen shown to suffer several modifications of the redox

0300-483X/$ – see front matter © 2006 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.tox.2006.04.028

TOX-49422; No. of Pages 9

2 M.J. Santos-Marques et al. / Toxicology xxx (2006) xxx–xxx

status and subsequent damage: (i) increased conversionof xanthine dehydrogenase to the superoxide-generatingenzyme xanthine oxidase; (ii) release of iron from ferri-tine; (iii) increased formation of reactive oxygen (ROS)and nitrogen species (RNS); (iv) alterations in glu-tathione homeostasis and lipid, protein and DNA oxi-dation; (v) increased sensitivity to heat-induced cellkilling when GSH is depleted; (vi) increased resistanceto heat-induced cell killing in stable H2O2-resistantcell types which overexpress cellular antioxidants (glu-tathione, catalase, superoxide dismutase, glutathioneperoxidase, and glutathione transferase) (Carvalho et al.,2001; Flanagan et al., 1998; Powers et al., 1992; Skibbaet al., 1988, 1991). On the other hand, the thermal energyrequired to induce hyperthermic cell death is closelycorrelated to that required for cellular protein denatu-ration. This has led to the hypothesis that the cytotoxiceffect of hyperthermia is mainly based on denatura-tion of cytoplasmatic and membrane proteins (Gerner,1985; Sapareto, 1987). The most relevant modificationsinduced by hyperthermia were described to occur at theplasma membranes (alteration of lipid properties, mem-brane transport, membrane receptors), proteins (proteindenaturation, unfolding, aggregation and insolubiliza-tion), nucleic acids (chromosomal aberrations, mitoticdysfunctions, inhibition of repair of DNA damages) andcytoskeleton (morphological changes such as cell round-ing and blebbing) (Lepock, 2003).

Whereas synthesis of most cellular proteins isimpaired under hyperthermic conditions, this is not thecase with the so-called heat shock proteins (HSPs).Indeed, most living cells are sensitive to sudden eleva-tions in temperature and respond to this environmentalinsult by activating the heat shock response that leadsto the synthesis of HSPs. HSPs are highly conservedproteins that play a role in the normal cell functionsas diverse as protein folding (in their role as molec-ular chaperones), signal transduction, cell growth anddifferentiation, and the regulation of the actin cytoskele-ton, therefore contributing for cellular homeostasis andpromoting survival (Morimoto et al., 1997; Nollen andMorimoto, 2002; Pirkkala et al., 2001). Among theseproteins, 70-kDa heat shock proteins (HSP70s) assist awide range of folding processes, including the foldingand assembly of newly synthesized proteins, refolding ofmisfolded and aggregated proteins, membrane translo-cation of organellar and secretory proteins, and controlof the activity of regulatory proteins (Mayer and Bukau,2005). Therefore, it seems that regulation of HSP tran-scription by heat-activated signalling pathways may bean important factor affecting balance between death andsurvival programs.

In mammalian cells, triggering of the heat shockresponse (HSR) requires activation and nuclear translo-cation of transregulatory proteins, the heat shock factors(HSFs) (Ali et al., 1998; Farkas et al., 1997; Morimoto,1998). HSFs, when activated and translocated into thenucleus, acquire DNA binding activity at the heat shockelement (HSE), thereby mediating transcription of theheat shock genes responsible for the synthesis of HSPs(Pirkkala et al., 2001). Heat shock factor 1 (HSF1) is thetranscription factor that plays a central role in the HSRby regulating expression of HSPs. In fact, genetic anal-ysis of hsf1#/# cells established that HSF1 is the onlyheat-inducible HSF in mice (Ahn et al., 2001). Undernormal conditions, HSF1 forms heterocomplexes withregulatory proteins such as HSP70 and HSP90 in thecytosol, which interfere with HSF1 transactivation. It isgenerally accepted that accumulation of non-native pro-teins caused by heat stress is a proximal signal for HSF1activation. As a result of competition with non-nativeproteins for these chaperones (e.g. HSP70 and HSP90),which prefer these to HSF1, HSF1 is released and acti-vated through a multistep process that involves con-version from the inactive monomer to the homotrimer,translocation into the nucleus, binding to heat shock ele-ment (HSEs), a consensus sequence located upstream ofheat shock genes, and target gene activation (Park et al.,2005). Additionally, the activity of HSF1 is regulatedby reversible phosphorylation in many distinct sites.In this regulation many kinases are involved, includ-ing heat-activated Akt, ERK, and JNK kinases (Gabaiand Sherman, 2002). For instance, the activity of HSF1is increased by heat-induced activation of Akt (Chu etal., 1996), but heat-induced activation of ERK and JNKresults in a decreased HSF1 activity through phosphory-lation at multiple sites (Dai et al., 2000).

Thus, an increasing body of data has demonstratedthat mammalian cells have elaborate networks of molec-ular signalling in counteracting heat shock and in devel-oping adaptation to oxidative stress to avoid cell death.However, the precise mechanisms linking heat shock,oxidative stress and cell survival/cell death mechanismsare not yet clearly understood. Liver cells are probablywell suited for such studies. In fact, mammalian liverpossesses an extraordinary capacity for compensatorysignalling mechanisms in response to conditions thatinduce loss of cell viability by physical, infectious, ortoxic injury (Jones and Czaja, 1998). However, the num-ber of reports in the literature correlating hyperthermia-induced toxicity with cell signalling in freshly isolatedor cultured hepatocytes is still very scarce and incom-plete, involving the use of hepatoma cell line HepG2(Bauer et al., 2003), human liver cell line (de Vera et

M.J. Santos-Marques et al. / Toxicology xxx (2006) xxx–xxx 3

al., 1996a) primary cultures of rat hepatocytes (Diez-Fernandez et al., 2002; Kurz et al., 1998; de Vera et al.,1996b; Wu et al., 1993), freshly isolated rat hepatocytes(Gutsmann-Conrad et al., 1999; Heydari et al., 1993,2000; Landry et al., 1982), and rainbow trout hepato-cytes (Airaksinen et al., 1998). Curiously, from thesereports, the concomitant study of HSF and HSP wasonly performed in those of Heydary’s group (Gutsmann-Conrad et al., 1999; Heydari et al., 1993, 2000), whichevidences the scarcity of studies in this field. Moreover,at the best of our knowledge, the deleterious effects ofmild continuous hyperthermia reflected in the develop-ment of oxidative stress, together with the activation ofcellular protective mechanisms, like HSF and HSP is stillto be studied in hepatocyte cell models.

The purpose of this study was thus to study the timecourse of hyperthermia-induced oxidative stress and cel-lular signalling through the activation of heat shockfactor 1 (HSF1) and heat shock protein 70 (HSP70),using freshly isolated mouse hepatocytes. Freshly iso-lated mouse hepatocytes were not tested before forthis type of study, but may have some advantages overother hepatocyte systems, since they are more economic,and do not require large animal house or laboratoryspace for handling. Furthermore, the mouse is the sec-ond most used laboratory animal species in toxicol-ogy studies, just after the rat species, which requiresa good knowledge about its responsiveness to stressfulconditions.

2. Materials and methods

2.1. Chemicals

All reagents used were of analytical grade. Collage-nase (grade I), bovine serum albumin (fraction V), pyru-vic acid, reduced glutathione (GSH), oxidized glutathione(GSSG), glutathione reductase (E.C. 1642), reduced nicoti-namide adenine dinucleotide (!-NADH), reduced nicoti-namide adenine dinucleotide phosphate (!-NADPH), 5,5-dithio-bis(2-nitrobenzoic) acid (DTNB), 2-vinylpyridine, 4-(2-hydroxyethyl)-1-piperazineethanesulphonic acid (HEPES),ethylene glycol bis(2-aminoethyl ether)-N,N,N$N$-tetraaceticacid (EGTA), trypan blue and 2-thiobarbituric acid, wereobtained from Sigma Chemical Company (St. Louis, MO,USA). Nuclear extract from HeLa cells (gel-shift assay grade)was obtained from Promega Corporation (Madison, USA). Thenuclear extract kit was obtained from Active Motif (Rixensart,Belgium). Oligodeoxynucleotides (specific probes and com-petitors) were obtained from Amersham Pharmacia Biotech(Uppsala, Sweden). The antibody anti-HSF 1 was obtainedfrom Stressgen Bioreagents (Victoria, Canada). Perchloric acidand trichloroacetic acid and all other chemicals were obtainedfrom Merck (Darmstadt, Germany).

2.2. Animals

Adult male Charles River CD1 mice (25–30 g b.w. and 6weeks of age) were obtained from Charles River Laboratories(Barcelona, Spain), at least 1 week before experiments. Theseanimals were housed in polyethylene cages with free accessto standard chow and water, maintained on a standard 12:12 hlight: dark cycle. The temperature of the holding room was nor-mally maintained at 20 ± 1 "C and humidity between 40 and60%. Surgical procedures were performed under diethyl etheranaesthesia and occurred always between 9.00 and 10.30 a.m.Housing and experimental treatment of the animals were inaccordance with National Institutes of Health guidelines (ILAR1996). The experiments complied with the current laws of Por-tugal.

2.3. Isolation and Incubation of hepatocytes

Hepatocyte isolation was performed by collagenase perfu-sion as previously described (Carvalho et al., 2001). At thebeginning of the experiments, cell viability was between 85and 96%.

Cell incubations were performed in a shaking water bath(90 oscillations/min) using 106 cells/mL in Krebs-Henseleitbuffer supplemented with 25 mM HEPES (pH 7.4) and aer-ated every 30 min with an air stream of humidified carbogen.The hepatocytes suspensions were always pre-incubated for30 min at 37 "C before the beginning of the experiments. Cellswere then incubated for 4 h at two different incubation condi-tions, normothermia (37 "C) and hyperthermia (41 "C). Samplealiquots were taken at time 0 and thereafter each 60 min dur-ing 4 h of incubation for measurement of cell viability, lipidperoxidation, reduced and oxidized glutathione, HSF-1, andHSP70.

2.4. Biochemical analysis

2.4.1. Cellular viabilityCell viability was determined after isolation by the Trypan

Blue exclusion test. During the course of the experiments cellviability was determined by the measurement of LDH (lactatedehydrogenase) leakage, which was randomly confirmed bythe Trypan Blue exclusion test.

2.4.2. Lipid peroxidationThe extent of lipid peroxidation, which can be character-

ized by the formation of malondialdehyde, was measured bythe thiobarbituric acid reactive substances (TBARS) assay at535 nm as previously described (Carvalho et al., 1997).

2.4.3. Reduced and oxidized glutathioneGSH and GSSG measurement was performed by the

DTNB-glutathione reductase recycling assay as describedbefore (Carvalho et al., 2004; Vandeputte et al., 1994). Briefly,cell suspension aliquots (300 "L) were precipitated with per-chloric acid (5% in final volume), and centrifuged at 6000 rpm,

4 M.J. Santos-Marques et al. / Toxicology xxx (2006) xxx–xxx

4 "C, for 10 min. The supernatant was then kept frozenat #80 "C until assay. Before quantification of glutathione,samples were diluted 20% in 5% perchloric acid. Samples(200 "L), standards (0.625–5.00 nmol/mL as GSH equiva-lents) and blank were then neutralized with 200 "L of KHCO3

0.76 M and centrifuged at 13,000 rpm, 4 "C, for 1 min. Stockbuffer (Na2HPO4·2H2O 71.5 mM, NaH2PO4·2H2O 71.5 mM,EDTA 0.62 mM, H2O, pH 7.5) was prepared periodicallyand stored at 4 "C. NADPH (0.24 mM), DTNB (0.70 mM)and glutathione reductase (10 U/mL) were prepared dailyin stock buffer. Standards, samples and freshly preparedreagent were kept on ice until being transferred into a96-well microplate. For measurement of total glutathione,100 "L/well of blank, standards or samples were added in trip-licate to the 96-well microplates, followed by 65 "L/well offreshly prepared NADPH/DTNB reagent. Microplates werethen incubated in a thermomax plate reader (PowerWaveX;Bio-Tek, Winooski, VT, USA) in the dark, for 15 min and at30 "C. The enzymatic reaction was started by quickly adding40 "L/well of glutathione reductase (10 U/mL in phosphatebuffer). The formation of TNB (5-thio-2-nitrobenzoic acid)was followed kinetically at 415 nm for 3 min, at 30 "C. GSSGwas measured after GSH derivatization by 2-vinilpyridine.2-Vinilpyridine (10 "L) was added to 200 "L of samples, stan-dards (0.50–8.00 nmol/mL as GSH equivalent) and blank andmixed continuously in a vortex for 1 h at 4 "C. GSSG was thenquantified as described above for total glutathione. The GSHlevels were calculated by subtracting GSSG content from totalglutathione levels, in GSH equivalents.

2.4.4. Fluorescent gel-shift assays for analysis of HSF1Transcriptional activation of HSF1 was determined by fluo-

rescent electrophoretic mobility shift assay (fEMSA). Whole-cell protein extracts from control and heat-treated hepatocyteswere used in the transcription assay. The whole-cell extract wasobtained as described in the extraction kit by Active Motif®

(Rixensart, Belgium). Protein quantification was determinedby the Lowry assay and each cell extract aliquot was diluted to2 mg protein/mL. For gel-shift assays, 1 pmol of Cy5-labeledspecific probe was incubated with 10 "g of extracted proteinin assay buffer (200 mM HEPES, 500 mM KCl, 10 mM EDTAand 10 mM DTT) and 50% glycerol for 2 h at 4 "C. The fol-lowing Cy5-labeled specific double-stranded probe was usedfor HSF-1 gel-shift assay: 5$-GAT CCT CGA ATG TTC GCGAAA AG-3$. Specificity was confirmed by the addition of a50-fold excess of either unlabeled specific competitor (spe-cific probe without the Cy5 label) or unlabeled nonspecificcompetitor and also by performing the supershift assay in thepresence of the specific antibody for HSF1. Nine microlitresof each mixture were separated on a 5% nondenaturing poly-acrylamide gel at 10 "C, 800 V, 50 mA, and 30 W for 2 h inTBE buffer 1% (90 mM Tris base, 2 mM EDTA, pH 8.3)using an ALFexpress II DNA Analyser (Amersham Pharma-cia Biotech, Sweden). The temperature was regulated by anexternal ALFexpress II Cooler system (Amersham PharmaciaBiotech, Sweden). The supershift assays were conducted using

specific antibody for HSF1. In the supershift assays the cellularextracts were pre-incubated with the specific anti-HSF1 anti-body on ice for 15 min before adding the specific probe to themixtures. Signals were analyzed by ALFwin Fragment Anal-yser 1.03 (Amersham Pharmacia Biotech, Sweden) in fEMSA(Rosenthal, 1995; Ruscher et al., 2000).

2.4.5. Western immunoblot analysis of HSP70To determine the levels of HSP70 in freshly isolated mouse

hepatocytes, whole-cell extracts were isolated from hepato-cytes, as described previously for fEMSA. Equal amounts ofthe extracted proteins, diluted in SDS-PAGE reducing buffer(4% (w/v) SDS, 0.125 M Tris pH 6.8, 20% (v/v) glycerol, 0.1%(w/v) bromophenol blue, 10% (v/v) mercaptoethanol), weresubjected to a SDS-polyacrylamide gel (12.5%) electrophore-sis (SDS-PAGE). The proteins in the gel were then transferredonto nitrocellulose membranes (Hybond ECL; AmershamPharmacia Biotech, Buckinghamshine, UK) using transferbuffer pH & 8.6 consisting of 25 mM Tris, 169.5 mM glycine,and 20% methanol. Membranes were then blocked overnightwith 5% skim milk in TBS-T 10% (100 mM Tris, 1.5 mMNaCl, 0.5% Tween 20). Primary and secondary antibodies werediluted in the blocking solution to the appropriate concentra-tions, 1:2500 and 1:500, respectively. The blots were incubatedwith the first antibody—anti-HSP70 mouse monoclonal anti-body (H5147, monoclonal BRT1-22, Sigma, St. Louis, USA),for 2 h at room temperature with shaking. After three washeswith TBS-T, they were incubated with anti-mouse IgG peroxi-dase labeled secondary antibody (1:500, Amersham PharmaciaBiotech, Buckinghamshine, UK), for 2 h at room tempera-ture and then washed with TBS-T for 3%. The bands werevisualized by treating the immunoblots with ECL chemilu-minescence detection reagent (Amersham Pharmacia Biotech,Buckinghamshine, UK) according to the supplier’s instruc-tions, followed by exposure to X-ray film (Kodak Biomax LightFilm, Sigma, St. Louis, USA). The films were scanned (Densit-ometer FX-710, Biorad) and analyzed using the QuantityOnesoftware (BioRad). Optical density results were expressed asthe percentual variation of control values.

2.5. Statistical analysis

Results are given as mean ± S.D. from five different exper-iments (each experiment was performed at both temperatureswith cells obtained from a different mouse). Statistical analysis,between results obtained at normothermic and hyperthermicconditions, were performed in each time point using two-wayANOVA followed by Bonferroni post-test. The level of signif-icance was set at p < 0.05.

3. Results

The increase in incubation temperature from 37 to41 "C induced oxidative stress and loss of cellular via-bility in freshly isolated mouse hepatocytes in a time-

M.J. Santos-Marques et al. / Toxicology xxx (2006) xxx–xxx 5

Fig. 1. Effect of normothermia (37 "C) and hyperthermia (41 "C) onlactate dehydrogenase (LDH) leakage from freshly isolated mousehepatocytes. Results are given as mean ± S.D. from five differentexperiments (cells obtained from each isolation procedure were testedat both temperatures). *p < 0.05 and ***p < 0.001, between normother-mic and hyperthermic conditions.

dependent manner. Fig. 1 illustrates the viability of hep-atocytes incubated in normothermic (37 "C) and hyper-thermic (41 "C) conditions. Cell viability was main-tained fairly stable under normothermia, at all testedtimes. In contrast, hyperthermia induced a significantdecrease in cellular viability (down to &30% after 4 h ofexposure, p < 0.001). An extensive and time-dependentlipid peroxidation was observed when hepatocytes weresubjected to hyperthermic conditions as measured byTBARS assay (p < 0.001 at all tested times of incubation)(Fig. 2). Significant variations in reduced (GSH) andoxidized (GSSG) glutathione intracellular levels werealso observed. GSH depletion was observed in bothconditions but clearly more extensive for cells exposed

Fig. 2. Effect of normothermia (37 "C) and hyperthermia (41 "C) onthiobarbituric acid reactive substances (TBARS) content in freshly iso-lated mouse hepatocytes. Results are given as mean ± S.D. from fivedifferent experiments (cells obtained from each isolation procedurewere tested at both temperatures). ***p < 0.001, between normother-mic and hyperthermic conditions.

to hyperthermia (p < 0.01) (Fig. 3A). This effect wasaccompanied by an increase in GSSG intracellular lev-els; cells incubated at 41 "C presented a considerableand significant increase in GSSG formation (p < 0.001)(Fig. 3B), which did also occur but in a much smallerextent in control cells. Comparison of both figures indi-cates a strong decrease in the GSH/GSSG ratio afterexposure to hyperthermia (data not shown), which cor-roborates the oxidative stress status.

Hyperthermia resulted on activation of HSF1 inmouse hepatocytes, as can be observed in Fig. 4. How-ever, it was not observed a linear time dependence,with a small decrease at the third hour and increasingagain at the fourth hour of incubation. For normother-

Fig. 3. Effect of normothermia (37 "C) and hyperthermia (41 "C) on reduced glutathione (GSH) (A) and oxidized glutathione (GSSG) (B) content infreshly isolated mouse hepatocytes. Results are given as mean ± S.D. from five different experiments (cells obtained from each isolation procedurewere tested at both temperatures). **p < 0.01 and ***p < 0.001, between normothermic and hyperthermic conditions.

6 M.J. Santos-Marques et al. / Toxicology xxx (2006) xxx–xxx

Fig. 4. Effect of normothermia (37 "C) and hyperthermia (41 "C)on the activation of heat shock transcription factor 1 (HSF1) infreshly isolated mouse hepatocytes. Results are given as mean ± S.D.from five different experiments (cells obtained from each iso-lation procedure were tested at both temperatures). *p < 0.05,**p < 0.01 and ***p < 0.001, between normothermic and hyperthermicconditions.

mic incubation conditions no significant variations intime were observed for HSF1 along time. In accordancewith the results obtained by fEMSA, the HSP70 proteinlevel exhibited a significant raise in hepatocytes incu-bated at hyperthermic incubation conditions versus con-trol cells (normothermia conditions) (Fig. 5), althoughonly perceptible at the second hour of incubation.Noteworthy, HSP70 was also significantly expressed in

control conditions from the third hour of incubation(p < 0.001).

4. Discussion

The results accomplished in this work demonstratedthat mild continuous hyperthermia (41 "C) leads tooxidative stress and loss of cellular viability in freshlyisolated mouse hepatocytes in a time-dependent man-ner, with significant effects already observed 1 h aftertreatment. These toxic effects developed concomitantlywith activation of HSF1 and emerged before the forma-tion of HSP70. Concerning cell signalling, the presentresults are in accordance with previous findings usingfreshly isolated hepatocytes from another species, therat (Gutsmann-Conrad et al., 1999; Heydari et al., 1993,2000), showing that heat shock is followed by a rapidoligomerization of HSF1 monomers to trimers that bindHSE and activate HSP70 transcription. Moreover, itextends these previous findings by correlating signallingdata with that of hyperthermia-induced oxidative stressand cell death.

In addition to freshly isolated hepatocytes, hyper-thermia-induced cell signalling through the activation ofHSP70 transcription has also been studied using hepato-cyte cultures, namely hepatoma cell line HepG2 (Baueret al., 2003), human liver cell line (de Vera et al., 1996a),primary cultures of rat hepatocytes (Diez-Fernandez et

Fig. 5. Effect of normothermia (37 "C) and hyperthermia (41 "C) on the expression of heat shock proteins (HSP70s) in freshly isolated mousehepatocytes. (A) Representative result, as obtained by Western Blot measurements. Equal amounts of protein from different groups were loaded onthe same gel to avoid any potential intergel variation. Alpha-tubulin western blot is included as a loading protein control. (B) Density measurementsof Western Blot obtained spots. Results expressed in (B) are given as mean ± S.D. from five different experiments (cells obtained from each isolationprocedure were tested at both temperatures). ***p < 0.001, between normothermic and hyperthermic conditions.

M.J. Santos-Marques et al. / Toxicology xxx (2006) xxx–xxx 7

al., 2002; Kurz et al., 1998; de Vera et al., 1996b; Wu etal., 1993), freshly isolated rat hepatocytes (Gutsmann-Conrad et al., 1999; Heydari et al., 1993, 2000; Landry etal., 1982), and rainbow trout hepatocytes (Airaksinen etal., 1998). Freshly isolated hepatocytes and primary cul-tures of hepatocytes are considered important in vitrocellular systems for studying various aspects of hep-atic metabolism and xenobiotic-induced hepatotoxicity.However, some biochemical processes are altered dur-ing the process of culturing. A major limitation to the useof hepatocyte cultures as an alternative to toxicity stud-ies employing experimental animals is the knowledgethat these cellular systems rapidly lose their cytochromeP450 (CYP) contents (Paine, 1990). Also, the expressionof the genes coding for the antioxidant enzymes catalaseand glutathione peroxidase decline markedly when hep-atocytes are cultured (Van Remmen et al., 1996). Thischange in the balance of antioxidant enzymes would beexpected to make the cultured hepatocytes more vulner-able to pro-oxidant toxicants, like ROS because it wouldalter the balance between the rate of H2O2 formationvia superoxide dismutase and its removal by catalaseand glutathione peroxidase, leading to oxidative stress,as previously observed (Van Remmen et al., 1996). Onthe other hand, it was also demonstrated that the expres-sion of HSP70 genes increases in cultured hepatocytes(Van Remmen et al., 1996), which corresponds to anadaptation of the cellular defense systems. Of note, itwas previously shown that the expression of HSP70 isnot immediately induced when hepatocytes are isolatedfrom rats by in situ collagenase perfusion (Heydari et al.,1993; Takahashi et al., 1994; Wu et al., 1993), which wasalso confirmed in the present study, since the formationof HSP70 was only evident at the second hour of incu-bation. We have previously reported that freshly isolatedhepatocytes from mice are a good model for the studyof hyperthermia-induced oxidative stress and loss of cellviability (Carvalho et al., 1997). In the present study, weclearly show that this model is also suitable for studyinghyperthermia-induced cell signalling through the activa-tion of HSE and formation of HSP70.

In the present study, hyperthermia resulted on acti-vation of HSF1 in mouse hepatocytes, in a time inde-pendent manner between the second and third hour ofincubation. This apparent lack of linearity in the responsemay be due to the observed cell death or to a possible ther-motolerance by HSPs-induced inactivation of HSF-1. Infact, it was previously shown that the heat shock tran-scriptional response attenuates upon prolonged exposureof cells at intermediate heat shock temperature (up to42 "C) or upon return to physiological temperature; thisattenuation is accompanied by conversion of the active

trimeric form of HSF to the non-DNA binding monomerand by a return to the normal subcellular distribution(Heydari et al., 2000; Morimoto, 1993).

The heat shock response is a finely regulated andhighly conserved mechanism to protect cells againstdifferent types of injury, including temperature andoxidative stress (Jolly and Morimoto, 2000). The cel-lular exposure to stress induces the rapid and transitoryexpression of HSPs. HSPs are present in both prokaryoticand eukaryotic cells, and their high level of conservationsuggests that they play an important role in fundamen-tal cell processes (Kregel, 2002; Nollen and Morimoto,2002; Schlesinger, 1990). However, in the present study,oxidative stress, expressed by the reduced/oxidised glu-tathione ratio and lipid peroxidation, and cellular viabil-ity were also measured showing a concomitant increaseof these biomarkers of cellular damage along with theheat shock response. As mentioned before, it has beenrecognized that the sustained generation of ROS and con-sequent oxidative stress is implicated in heat-inducedcell death. Our findings also demonstrate a decreaseddetoxification capacity via the action of glutathioneredox cycle with time, thus increasing the vulnerabil-ity of hepatocytes to further oxidative stress. The factthat the increase of HSP70 did not alter the hepatocytedamage trend is indicative that the heat shock protec-tive effect is perhaps better achieved in short hyperther-mic challenges as previously demonstrated in culturedrat hepatocytes (43 "C for 20 min) (Diez-Fernandez etal., 2002) and whole-body hyperthermia (42.5 "C for10 min) (Mokuno et al., 2004), rather than in continu-ous mild hyperthermia.

In conclusion, the results accomplished in this workshow that continuous mild hyperthermia leads to a sus-tained increase of oxidative stress, leading to cell deathin freshly isolated mouse hepatocytes. Furthermore,although cell signalling was triggered through the tran-scriptional activation of HSP70 via HSF1, this putativeprotective process did not modify the trend of hepa-totoxic effects mediated by this type of hyperthermicchallenging.

Acknowledgment

This work received financial support from FCT andPOCTI-FEDER European Community funding (projectPOCTI/ACT/43562/2001).

References

Ahn, S.-G., Liu, P.C.C., Klyachlo, K., Morimoto, R.I., Thiele, D.J.,2001. The loop domain of heat shock transcription factor 1 dictates

8 M.J. Santos-Marques et al. / Toxicology xxx (2006) xxx–xxx

DNA-binding specificity and responses to heat stress. Genes Dev.15, 2134–2145.

Airaksinen, S., Rabergh, C.M., Sistonen, L., Nikinmaa, M., 1998.Effects of heat shock and hypoxia on protein synthesis in rainbowtrout (Oncorhynchus mykiss) cells. J. Exp. Biol. 201, 2543–2551.

Ali, A., Bharadwaj, S., O’Carroll, R., Ovsenek, N., 1998. HSP90 inter-acts with and regulates the activity of heat shock factor 1 in Xenopusoocytes. Mol. Cell. Biol. 18, 4949–4960.

Bauer, I., Rensing, H., Florax, A., Ulrich, C., Pistorius, G., Redl,H., Bauer, M., 2003. Expression pattern and regulation of hemeoxygenase-1/heat shock protein 32 in human liver cells. Shock 20,116–122.

Carvalho, F., Remiao, F., Soares, M.E., Catarino, R., Queiroz, G.,Bastos, M.L., 1997. d-Amphetamine-induced hepatotoxicity: pos-sible contribution of catecholamines and hyperthermia to theeffect studied in isolated rat hepatocytes. Arch. Toxicol. 71, 429–436.

Carvalho, M., Carvalho, F., Bastos, M.L., 2001. Is hyperther-mia the triggering factor for hepatotoxicity induced by 3,4-methylenedioxymethamphetamine (ecstasy)? An in vitro studyusing freshly isolated mouse hepatocytes. Arch. Toxicol. 74,789–793.

Carvalho, M., Remiao, F., Milhazes, N., Borges, F., Fernandes, E.,Carvalho, F., Bastos, M.L., 2004. The toxicity of N-methyl-alpha-methyldopamine to freshly isolated rat hepatocytes is preventedby ascorbic acid and N-acetylcysteine. Toxicology 200, 193–203.

Chu, B., Soncin, F., Price, B.D., Stevenson, M.A., Calderwood,S.K., 1996. Sequential phosphorylation by mitogen-activated pro-tein kinase and glycogen synthase kinase 3 represses transcrip-tional activation by heat shock factor-1. J. Biol. Chem. 271,30847–30857.

Dai, R., Frejtag, W., He, B., Zhang, Y., Mivechi, N.F., 2000. c-JunNH2-terminal kinase targeting and phosphorylation of heat shockfactor-1 suppress its transcriptional activity. J. Biol. Chem. 275,18210–18218.

de Vera, M.E., Wong, J.M., Zhou, J.Y., Tzeng, E., Wong, H.R., Billiar,T.R., Geller, D.A., 1996a. Cytokine-induced nitric oxide synthasegene transcription is blocked by the heat shock response in humanliver cells. Surgery 120, 144–149.

de Vera, M.E., Kim, Y.M., Wong, H.R., Wang, Q., Billiar, T.R., Geller,D.A., 1996b. Heat shock response inhibits cytokine-inducible nitricoxide synthase expression in rat hepatocytes. Hepatology 24,1238–1245.

Diez-Fernandez, C., Andres, D., Cascales, M., 2002. Attenuatingeffects of heat shock against TGF-beta1-induced apoptosis in cul-tured rat hepatocytes. Free Radic. Biol. Med. 33, 835–846.

Farkas, T., Kutskova, Y.A., Zimarino, V., 1997. Intramolecular repres-sion of mouse heat shock factor 1. Mol. Cell. Biol. 18, 906–918.

Flanagan, S.W., Moseley, P.L., Buettner, G.R., 1998. Increased fluxof free radicals in cells subjected to hyperthermia: detection byelectron paramagnetic resonance spin trapping. FEBS Lett. 431,285–286.

Gabai, V.L., Sherman, M.Y., 2002. Interplay between molecular chap-erones and signalling pathways in survival of heat shock. J. Appl.Physiol. 92, 1743–1748.

Gerner, E.W., 1985. Definiton of a thermal dose. In: Overgaard, J. (Ed.),Hyperthermic Oncology. Taylor & Francis, London, pp. 253–262.

Gutsmann-Conrad, A., Pahlavani, M.A., Heydari, A.R., Richardson,A., 1999. Expression of heat shock protein 70 decreases with agein hepatocytes and splenocytes from female rats. Mech. AgeingDev. 107, 255–270.

Heydari, A.R., Wu, B., Takahashi, R., Strong, R., Richardson, A., 1993.Expression of heat shock protein 70 is altered by age and diet atthe level of transcription. Mol. Cell. Biol. 13, 2909–2918.

Heydari, A.R., You, S., Takahashi, R., Gutsmann-Conrad, A., Sarge,K.D., Richardson, A., 2000. Age-related alterations in the activa-tion of heat shock transcription factor 1 in rat hepatocytes. Exp.Cell. Res. 256, 83–93.

Jolly, C., Morimoto, R.I., 2000. Role of the heat shock response andmolecular chaperones in oncogenesis and cell death. J. Natl. CancerInst. 92, 1564–1572.

Jones, B.E., Czaja, M.J., 1998. Intracellular signaling in response totoxic liver injury. Am. J. Physiol. 275, G874–G878.

Kregel, K.C., 2002. Heat shock proteins: modifying factors in phys-iological stress responses and acquires thermotolerance. J. Appl.Physiol. 92, 2177–2186.

Kurz, A.K., Schliess, F., Haussinger, D., 1998. Osmotic regulation ofthe heat shock response in primary rat hepatocytes. Hepatology 28,774–781.

Landry, J., Chretien, P., Bernier, D., Nicole, L.M., Marceau, N.,Tanguay, R.M., 1982. Thermotolerance and heat shock proteinsinduced by hyperthermia in rat liver cells. Int. J. Radiat. Oncol.Biol. Phys. 8, 59–62.

Lepock, J.R., 2003. Cellular effects of hyperthermia: relevance tothe minimum dose for thermal damage. Int. J. Hyperthermia 19,252–266.

Mayer, M.P., Bukau, B., 2005. Hsp70 chaperones: cellular func-tions and molecular mechanism. Cell. Mol. Life Sci. 62, 670–684.

Mokuno, Y., Berthiaume, F., Tompkins, R.G., Balis, U.J., Yarmush,M.L., 2004. Technique for expanding the donor liver pool: heatshock preconditioning in a rat fatty liver model. Liver Transpl. 10,264–272.

Morimoto, R.I., 1993. Cells in stress: transcriptional activation of heatshock genes. Science 259, 1409–1410.

Morimoto, R.I., Kline, M.P., Bimston, D.N., Cotto, J.J., 1997. The heat-shock response: regulation and function of heat-shock proteins andmolecular chaperones. Essays Biochem. 32, 17–29.

Morimoto, R.I., 1998. Regulation of the heat shock transcriptionalresponse: cross talk between a family of heat shock factors,molecular chaperones, and negative regulators. Genes Dev. 12,3788–3796.

Nollen, E.A.A., Morimoto, R.I., 2002. Chaperoning signalling path-ways: molecular chaperones as stress-sensing ‘heat shock’ pro-teins. J. Cell Sci. 115, 2809–2816.

Paine, A.J., 1990. The maintenance of cytochrome P450 in rat hepa-tocyte culture: some applications of liver cell cultures to the studyof drug metabolism, toxicity and the induction of the P450 system.Chem-Biol. Interact. 74, 1–31.

Park, H.G., Han, S.I., Oh, S.Y., Kang, H.S., 2005. Cellular responsesto mild heat stress. Cell. Mol. Life Sci. 62, 10–23.

Pirkkala, L., Nykanen, P., Sistonen, L., 2001. Roles of the heat shocktranscription factors in regulation of the heat shock response andbeyond. FASEB J. 15, 1118–1131.

Powers, R.H., Stadnicka, A., Kalbfleisch, J.H., Skibba, J.L., 1992.Involvement of xanthine oxidase in oxidative stress and ironrelease during hyperthermic rat liver perfusion. Cancer Res. 52,1699–1703.

Rosenthal, N., 1995. Molecular medicine. Recognizing DNA. N. Engl.J. Med. 333, 925–927.

Ruscher, K., Reuter, M., Kupper, D., Trendelenburg, G., Dirnagl,U., Meisel, A., 2000. A fluorescence based non-radioactive elec-trophoretic mobility shift assay. J. Biotechnol. 78, 163–170.

M.J. Santos-Marques et al. / Toxicology xxx (2006) xxx–xxx 9

Sapareto, S.A., 1987. Thermal isoeffect dose: addressing the problemof thermotolerance. Int. J. Hyperthermia 3, 297–305.

Schlesinger, M.J., 1990. Heat shock proteins (minireview). J. Biol.Chem. 265, 12111–12114.

Skibba, J.L., Powers, R.H., Stadnicka, A., Kalbfleisch, J.H., 1988.The effect of hyperthermia on conversion of rat hepatic xan-thine dehydrogenase to xanthine oxidase. Biochem. Pharmacol.37, 4592–4595.

Skibba, J.L., Powers, R.H., Stadnicka, A., Cullinane, D.W., Almagro,U.A., Kalbfleisch, J.H., 1991. Oxidative stress as a precursor to theirreversible hepatocellular injury caused by hyperthermia. Int. J.Hyperthermia 7, 749–761.

Takahashi, R., Heydari, A.R., Gutsmann, A., Sabia, M., Richardson,A., 1994. The heat shock transcription factor in liver exists in a

form that has DNA binding activity but no transcriptional activity.Biochem. Biophys. Res. Commun. 201, 552–558.

Vandeputte, C., Guizon, I., Genestie-Denis, I., Vannier, B., Loren-zon, G., 1994. A microtiter plate assay for total glutathione andglutathione disulfide contents in cultured/isolated cells: perfor-mance study of a new miniaturized protocol. Cell Biol. Toxicol.10, 415–421.

Van Remmen, H., Williams, M.D., Heydari, A.R., Takahashi, R.,Chung, H.Y., Yu, B.P., Richardson, A., 1996. Expression of genescoding for antioxidant enzymes and heat shock proteins is altered inprimary cultures of rat hepatocytes. J. Cell. Physiol. 166, 453–460.

Wu, B., Gu, M.J., Heydari, A.R., Richardson, A., 1993. The effect ofage on the synthesis of two heat shock proteins in the hsp70 family.J. Gerontol. 48, B50–B56.