Toxicology and Applied Pharmacology 244 (2010) 263–272

Contents lists available at ScienceDirect

Toxicology and Applied Pharmacology

j ourna l homepage: www.e lsev ie r.com/ locate /ytaap

Arsenite induces apoptosis in human mesenchymal stem cells by altering Bcl-2family proteins and by activating intrinsic pathway

Santosh Yadav, Yongli Shi, Feng Wang, He Wang ⁎Department of Environmental Health Sciences, School of Public Health and Tropical Medicine, Tulane University New Orleans, LA, USA

⁎ Corresponding author. Department of EnvironmentaAve., TulaneUniversity, BoxTW-2100,NewOrleans, LA701

E-mail address: hwang2@tulane.edu (H. Wang).

0041-008X/$ – see front matter © 2010 Elsevier Inc. Adoi:10.1016/j.taap.2010.01.001

a b s t r a c t

a r t i c l e i n f o

Article history:Received 3 August 2009Revised 4 January 2010Accepted 5 January 2010Available online 18 January 2010

Keywords:Human mesenchymal stem cellsG2-M arrestS-phaseTUNELAnnexin V staining

Purpose: Environmental exposure to arsenic is an important public health issue. The effects of arsenic ondifferent tissues and organs have been intensively studied. However, the effects of arsenic on bone marrowmesenchymal stem cells (MSCs) have not been reported. This study is designed to investigate the cell deathprocess caused by arsenite and its related underlying mechanisms on MSCs. The rationale is that absorbedarsenic in the blood circulation can reach to the bone marrow and may affect the cell survival of MSCs.Methods: MSCs of passage 1 were purchased from Tulane University, grown till 70% confluency level andplated according to the experimental requirements followed by treatment with arsenite at variousconcentrations and time points. Arsenite (iAsIII) induced cytotoxic effects were confirmed by cell viabilityand cell cycle analysis. For the presence of canonic apoptosis markers; DNA damage, exposure ofintramembrane phosphotidylserine, protein and m-RNA expression levels were analyzed.Results: iAsIII induced growth inhibition, G2-M arrest and apoptotic cell death in MSCs, the apoptosis inducedby iAsIII in the cultured MSCs was, via altering Bcl-2 family proteins and by involving intrinsic pathway.Conclusion: iAsIII can induce apoptosis in bone marrow-derived MSCs via Bcl-2 family proteins, regulatingintrinsic apoptotic pathway. Due to the multipotency of MSC, acting as progenitor cells for a variety ofconnective tissues including bone, adipose, cartilage and muscle, these effects of arsenic may be important inassessing the health risk of the arsenic compounds and understanding the mechanisms of arsenic-inducedharmful effects.

l Health Sciences, 1430 Tulane12,USA. Fax:+15049881726.

ll rights reserved.

© 2010 Elsevier Inc. All rights reserved.

Introduction

Mesenchymal stem cells or marrow stromal cells (MSCs) arenon-hematopoietic multipotent stem cells, which have shown hightherapeutic potential because of their ability to self-renew and easeof their isolation and expansion (Prockop, 2009). MSCs possess abroad spectrum for regenerative medicine due to their potential todifferentiate into lineages of mesenchymal tissues such as bone,cartilage and tendon (Lund et al., 2009; Johnstone and Yoo, 1999;Grogan et al., 2009; Schnabel et al., 2009). MSCs reside predom-inantly in the bone marrow but they are also distributedthroughout many other tissues, where MSCs serve as local sourcesof quiescent stem cells. Moreover, these cells with other non-hematopoietic cells such as osteoblast, adipocyte, fibroblast,reticular and endothelial cells possess a functional role inregulating hematopoiesis (Schofield, 1978; Nemeth et al., 2009).Thus, it is obvious that any cellular stress in MSCs can drasticallydiminish the hematopoietic system. It has been previously revealed

that radiation or presences of toxic substances in hematopoieticenvironment can damage genetic material of stromal and hemo-poietic cells (Ludwika et al., 1996) or may interfere in the processof hematopoiesis through the alteration in stromal layer or incytokine homeostasis (Ebert et al., 2006). It is, therefore, extremelyrelevant to explore, carcinogens or toxicant's induced response onMSCs in order to predict whether a functional MSCs is available. Inthis study, we sought to determine the role of inorganic arseniteon MSCs. Inorganic arsenite (iAsIII) is classified as a humancarcinogen by the International Agency for Research on Cancer(IARC, 1980). In some areas of the world, arsenic contaminationlevel in the ground water can reach up to several times higherthan the safe level, as cited by WHO (1996). In portions of thewestern United States, domestic and public wells have shownincreased arsenic levels which correlate to excess fatality risksestimated to be 1 per 9300 to 1 per 6600 (Kumar et al., 2009).

Arsenite and its metabolites can travel to the fetus through theplacental barrier and can compromise the normal developmentalprocess including fetal and neonatal stages (Vahter, 2008; Hill et al.,2009; Lianos and Ronco, 2009; Röllin et al., 2009; Concha et al., 1998).Despite the large number of studies on the developmental toxicologyof arsenite, information about its impact on MSCs is still lacking.

264 S. Yadav et al. / Toxicology and Applied Pharmacology 244 (2010) 263–272

However, studies have shown the potential immunotoxic effects ofiAsIII and its methylated metabolites on human cord blood or murinebone marrow derived progenitor cells (Ferrario et al., 2008, 2009;Pessina et al., 2002; Pessina et al., 2005). To extend these studies, weaimed at establishing MSCs from bone marrow as an adult stem cellmodel for the risk assessment of iAsIII. In this study, we have chosentrivalent form of arsenic over the pentavalent, due to its associationwith the higher cytotoxicity potential (Dopp et al., 2005). This study isfocused on apoptotic stimulation in MSCs by altering Bcl-2 familyproteins and activating intrinsic apoptotic pathway. Arsenite caninhibit cell cycle progression (McCollumet al., 2005) therefore,we alsoexamined the effect of arsenite on cell cycle profile ofMSC. InMSCs, themost of the population exists in G1 phase of the cell cycle in contrast tothe cancer cell lines where ample number of cells exists in S-phase,therefore the analysis of arsenite targeted cell cycle phase in MSCs canbe interesting aspect of the study.

Materials and methods

Culture of MSCs. Mesenchymal stem cells (MSCs) from adult humanbone marrow were obtained from the center for the preparation anddistribution of adult stem cells of Tulane University (New Orleans, LA,USA), as frozen vials of passage one (P-1) (http://www.som.tulane.edu/gene_therapy/distribute.shtml )whichwerewell characterized formultipotent differentiation, and these cells were negative for hemato-poietic markers (CD34, CD36, CD117, and CD45), and positive for CD29(95%), CD44 (N93%), CD49c (99%), CD49 (N70%), CD59(N99%), CD90(N99%), CD105 (N99%), and CD166 (N99%). To cultureMSCs, alpha-MEMsupplemented with FBS (16.5%) (Atlanta Biologicals, Lawrenceville, GAUSA) and 2 mM L-glutamine and penicillin/ streptomycin (1%) (GIBCO,Invitrogen, Carlsbad, CA, USA) was used. Cells were maintained withculture conditions such as 37 °C temperature with 5% CO2. MSCs(1X106/per vial) were plated in a 15-cm diameter dish and allowed toreach up to 70–75% confluence level (3–4 days). After appropriateconfluency level, MSCs were harvested with Trypsin/1 mM EDTA(0.25%) (GIBCO, Invitrogen), and re-plated in various culture platesaccording to the experimental design. In the entire study, MSCs frompassage 2 to 5 were used from the same donor.

Preparation of arsenite treatment. Inorganic arsenite (iAsIII )compound was purchased from Sigma Aldrich (Sigma Aldrich, St.Louis, MO, USA) with the purity of 99%, as stated by Sigma. In eachexperiment, fresh iAsIII treatment was prepared with the variousconcentrations, i.e., 1 to 40 μM in α MEM in the absence of FBS.

Analysis of cell viability by MTT assay. The MTT assay, an index ofcell functionality, is based on the ability of cells to reduceMTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) from a yel-low water-soluble dye to a dark blue insoluble formazan product. TheMTT assay kit was purchased from Sigma (St. Louis, MO, USA). For theassay, MSCs (1 × 104) were plated in 96 multi-well microtiter plates.After overnight incubation, cells were treated with various concen-tration of iAsIII (1 to 40 µM) for 6–48 h. Then the MTT dye was addedto each well for the last 2 h of the incubation; the reaction was thenstopped by the addition of solubilization reagent. The optical densityin each well was then determined at 570 nm using the FLUOstarOptima multi-detection plate reader. Background absorbance ofmedium in the absence of cells was subtracted from all measuredvalues. Untreated (UT) MSCs were considered as 100 % viable cells.

Morphological analysis. MSCs were seeded in 12-well plates incomplete culture medium (α-MEM). After overnight incubation, cellswere treated with various concentrations of iAsIII (1 μM to 40 μM) for24 h. After desired treatment duration (24 h), for morphology theimages were taken using bright field microscope under 20Xmagnification.

BrdU assay for cell proliferation. Inhibition in DNA synthesis by iAsIII

was assessed by bromodeoxyuridine (BrdU) incorporation rate usingenzyme-linked immunosorbent assay kit (Chemicon InternationalInc., Temecula, CA, USA) according to the protocol provided by themanufacturer. In brief, MSCs (1× 104) were plated in 96 well plateand incubated overnight at 37 °C. Next day, cells were treated withiAsIII (1 μM to 20 μM) for 24 h. Subsequently, at 22 h BrdU was addedand further incubated for 2 h. After that, cells were fixed for 30 min, atroom temperature (RT), followed by washing. For detection, anti-BrdU monoclonal (1:200 dilution) was added and incubated for anhour at RT. Cells were washed, and incubated with goat anti-mouseIgG, peroxidase conjugate for 30 min, at RT. After the final wash, thecells were treated with 3,3′,5,5′-tetramethylbenzidine (TMB) perox-idase substrate followed by an incubation of 30 min at RT (in dark).Finally, reading was taken by a spectrophotometer at the wavelengthof 405 nm (FLUOstar plate reader). All samples were assayed induplicate.

Cell cycle analysis. The effect of iAsIII on cell cycle distribution wasassessed by flow cytometry using propidium iodide staining. Briefly,0.3 × 106 MSCs were plated and treated with iAsIII for 24 h. At thecompletion of the treatment, MSCs were harvested using 0.25%Trypsin/EDTA followed by two times washing in cold PBS (0.1%BSA). After that, MSCs were fixed with ice-cold 70% ethanol forovernight at –20 °C. MSCs were then incubated with propidiumiodide containing 80 mg L−1 RNase A and 50 mg L−1 for 30 min.Cell cycle profile was assessed by BD LSRII analyzer (BectonDickinson, USA). Analysis of the cell cycle was carried out usingModFIT software (Becton Dickinson, USA).

TUNEL assay. To examine apoptosis by TUNEL assay, cells (1.5 ×104) were seeded on 8 well chamber slides (Nunc). Apoptotic cellswere detected by in situ cell death detection kit (TMR-red, Roche, IN,USA) according to the manufactures' instructions. In brief, cells werewashed twice with PBS (0.2% BSA) at 4 °C. Then, 0.2 mL fixationsolution (4% formalin) was added to the each well of the chamberslide followed by 30 min incubation on ice, while shaking, and thenwashed with PBS. Cells were permeabilized with 0.1% Triton X–100,(freshly prepared in 0.1% sodium citrate), and fixed in 70% ethanol for30 min at −20 °C. After fixation, cells were washed twice with PBSand incubated in TUNEL reaction mixture (dUTP-FITC) for 60 min at37 °C. Cells were then washedwith PBS (twice) and stained with DAPI(4′,6-diamidino-2-phenyl indole) (Sigma Aldrich, St. Louis, MO, USA)for nuclear localization. Samples were analyzed using fluorescencemicroscope (LICA model), with the range of 570/620 nm and DAPIwith the range of 340/380 nm by measuring TMR-red (dUTP-FITCincorporated fragmented) and DAPI (binding to DNA).

Annexin V-FITC apoptosis detection. MSCs (5×105) were plated in10 cm petri dishes for each data point. After overnight incubation at37 °C, cells were treated with iAsIII (5–20 μM). After 24 h treatmentduration, cells were dissociated using 0.25% trypsin/EDTA for 1 min(keeping all floating cells). Annexin V-FITC staining was performedusing Annexin V-FITC apoptosis detection kit (Sigma Aldrich. St. Louis,MO, USA) in accordance to the manufacturer's instructions. In brief,cells were centrifuged, re-suspended in 1× binding buffer at aconcentration of 1×106/mL in 100 μL of the solution, transferred ina 5 mL FACS tube, combined with 5 μL Annexin V/FITC (conjugatedwith fluorescein isothiocyanate) and 10 μL propidium iodide. Afterincubation for 30 min at RT in dark, 400 μL of 1×Binding Buffer wasadded to each tube and flow cytometry analysis was performed im-mediately. Data acquisition and analysis were performed by BD LSRIIanalyzer (Becton Dickinson, USA) using “CELL Quest” software. Cellsthat were Annexin V (−) and PI (−) were considered viable cells.Cells that were Annexin V (+) and PI (−) were considered early-

Table 1

Nameof gene

Gene forward primer sequence Gene reverse primer sequence

BCl-2a 5′-GGGGAGGATTGTGGCCTTC-3′ 5′-CAGGGCGATGTTGTCCACC-3′Bcl-xl 5′-CTGTGCGTGGAAAGCGTAGA-3′ 5′-AGGGGCTTGGTTCTTACCCA-3′Bax 5′-GGG TGG TTG GGT GAG

ACT C-3′5′-AGA CAC GTA AGG AAA ACGCAT TA-3′

GAPDH 5′-CGAGATCCCTCCAAAATCAA-3′ 5′-GGTTCACACCCATGACGAA-3′

265S. Yadav et al. / Toxicology and Applied Pharmacology 244 (2010) 263–272

stage apoptotic cells. Cells that were Annexin V (+) and PI (+) wereconsidered late-stage apoptotic cells.

Western blot analysis. MSCs (5×105) were plated into the 10 cmdishes for each data point. After overnight incubation at 37 °C, cellswere treated with iAs III (10–20 μM) for the period of 12 and 24 h. Thecells were washed with PBS and collected with a rubber scraper. Aftercentrifugation, the cell pellets were resuspended in RIPA buffer(25 mM Tris, pH 7.5), 50 mM NaCl, 0.5% sodium deoxycholate, 2%Nonidet P-40, 0.2% sodium dodecyl sulfate, 1 mM phenyl-methylsulphonyl fluoride aprotinin, leupeptin, and pepstatin (SigmaSt. Louis, MO, USA). The cells were then cleared by centrifugation(10 min, 14,000×g). For loading, Laemmli sample buffer was usedfollowed by heating at 95 °C for 5 min, the samples were separated on10–12% SDS-PAGE gel and transferred to nitrocellulose membranes(Bio Rad, Hercules, CA, USA). The membranes were blocked with TrisBuffered Saline with 0.1% Tween 20 (TBS-T) containing 5% Non-fatmilk. After blocking nonspecific binding sites, the membranes wereincubated with primary antibodies; anti cyclin D1, anti-p27Waf/Cip1,anti-p53, anti-cyclin B1, anti-cdc1, anti-Bcl-2, anti-Bax, anti-Bad, antiPhospho-Bad, anti-Cytochrome c, anti-PUMA, anti-Cleaved Caspase-3,anti-Caspase-9, and anti-PARP. For measuring total protein levels,blots were stripped and re-proved with anti β-actin, HSP60 or withthe GAPDH (all antibodies were purchased from cell signalingtechnology Inc., MA, USA). For visualization, Infrared IRDye-labeledsecondary antibodies (i.e., with the channel of 680 or 800CW, LI-CORBiosciences, Nebraska, USA) were used, which was prepared in 5%non-fat milk in TBS-T. Blots were visualized and analyzed (fordensitometry) using Odyssey Infrared Imaging System (LI-CORBiosciences, Nebraska USA).

Preparation of cytosolic and mitochondrial fractions. MSC s we r ewashed with ice-cold PBS, incubated on ice for 10 min, and then re-suspended in isotonic homogenization buffer (250 mM sucrose,10 mM KCl, 1.5 mMMgCl2, 1 mM sodium-EDTA, 1 mM sodium-EGTA,1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, and10 mM Tris–HCl, pH 7.4) containing protease inhibitor mixture(aprotinin, leupeptin, pepstatin, Sigma St. Louis Missouri, USA). After80 strokes in a Dounce homogenizer, the unbroken cells were spundown at 30×g for 5 min. Themitochondria fractions were fractionatedat 750×g for 10 min and 14,000×g for 20 min, respectively, from thesupernatant. For Cytosolic fraction after 10 strokes with a loosehomogenizer, supernatant was collected after spun down at 750×g for10 min and followed by a centrifugation at 14,000×g for 20 min. Thesamples were separated on a 12% SDS-PAGE gel as described above.

RT-PCR analysis for gene expression. Total RNA from MSCs (treatedand untreated) was extracted using TRIZOL reagent (InvitrogenCarlsbad, CA, USA) according to the manufacturer's instructions. RT-PCR was performed using one-step RT-PCR kit (iQ-5 Bio-Radlaboratories, Inc., CA, USA) and according to the manufacturer'sinstructions. In brief, total RNA, 10 ng, was added to each RTreaction, containing the kit mix and the appropriate PCR primers(20 pmol) in 50 μL total volume. RT-PCR was carried out for 30 minat 50 °C followed by a 15 min step at 95 °C. Amplifications wereperformed with 30 cycles. Each cycle included denaturation at 94 °Cfor 1 min, annealing at an appropriate temperature for 1 min, andextension at 72 °C for 1 min, followed by a final extension at 72 °Cfor 10 min. For amplification from first -strand cDNA, primers arelisted in Table 1 have been used in this study. The number of cycleswas determined in preliminary experiments to be within theexponential range of RT-PCR amplification. Ct value was calculatedwith the mRNA levels specific vs. the corresponding levels ofGAPDH. All the experiments were repeated at least twice and similarresults were obtained.

Statistical analysis. Results shown are the average of three inde-pendent experiments (n=3). The values are expressed as means±standard deviation. For the determination of statistical significance,results were analyzed by a one-way ANOVA (by Graph Pad Prismversion 4), followed by the Tukey's test procedure for multiplecomparisons. A p value less than 0.05 was judged to be statisticallysignificant.

Results

Effect of iAsIII on growth rate of MSCs

Cell viability by MTT assayTo investigate the viability of MSCs upon iAsIII treatment, we

applied MTT assay. MSCs were treated with various concentrations ofiAsIII ranging from 1 to 40 μM for the following time points 6 to 48 h.Results showed that at lower concentrations, i.e.,≤5 μM,MSCs remaininsensitive towards iAsIII treatment for the period of 6–12 h and nogreater change was seen in MSCs viability. Whereas, at the higherconcentrations, i.e., ≥5 μM for the period of 24 and 48 h, a significantloss in cell viability over the control cells was observed (asdemonstrated in Fig. 1A).

Analysis of DNA synthesis inhibition by BrdU incorporation assayTo assess the effect of iAsIII on DNA synthesis in MSCs, BrdU

incorporation assay was performed. In MSCs, treatment of iAsIII at≥5 μM concentration, the DNA synthesis was found to be lower (Fig.1B). However, at the lower concentration (≤5 μM), it could not alterthe DNA synthesis rate, where BrdU incorporation was almost parallelto the baseline control.

Morphological analysisFigs. 1C and D demonstrate detached MSCs with round and rough

surface and their morphological view by light microscopy underphase contrast field. In the treated group, the most conspicuouschanges were observed in 5 μM of iAsIII treatment, which includedetachment of the cells from the cell culture substratum and loss ofcell–cell interactions; these changes included round and roughmembrane bearing detached cells (Wen et al., 1997) and becamemore remarkable after 24 h with increased concentration of iAsIII

(≥5 μM) treatment. Whereas these changes were absent in untreatedMSCs and very less at low treatment level (≤5 μM) of iAsIII.

Effect of iAsIII on cell cycle profile of MSCs

To understand the mechanism behind iAsIII-induced loss in cellviability and less BrdU incorporation, we examined whether iAsIII

influences distribution of cell population in various phases of cellcycle. Our flow cytometry results revealed that iAsIII treatment, for theperiod of 24 h induced change in cell cycle profile of MSCs ascompared with untreated MSCs. As shown in Fig. 2, iAsIII treatmentarrests MSCs at G2-M Phase of the cell cycle in a dose-dependentmanner. The percentages of cells in G2-M phase (16.5%, Fig 2E) at20 μMof iAsIII treatment were increased from 10 μMof iAsIII treatment(10.61%, Fig. 2D) compared with the untreated MSCs (4.56%, Fig. 2A).At lower concentration, iAsIII had no obvious effect on cell cycle profile

Fig. 1. Effect of iAsIII on growth rate of MSCs. (A) Cell viability was assessed byMTT assay. All cells were incubated with various concentrations of iAsIII (1–40 μM) for the period of 6–48 h. The viability of cells, incubated without iAsIII was defined as 100% and the bars represent themean values (±SD) from three samples per treatment regimen. Asterisks representsignificant difference when compared to the untreated (UT) MSCs (*pb0.05, **pb0.01 or ***pb0.001). (B) Effect of iAsIII on DNA synthesis inhibition, was assessed by BrdU labelingassay and expressed as the absorbance at 405 nm. Measures were taken after treatment of the cells with iAsIII (1–20 μM) for 24 h. The bars represent the mean values (±SD) fromthree samples per treatment regimen. *Significant difference when compared to the untreated MSCs (at pb0.05). (C) Cells which are detached and possess round and roughmorphological appearance. The bars represent the mean values (±SD) from three samples per treatment regimen. (D) Representative images of morphological analysis.

266 S. Yadav et al. / Toxicology and Applied Pharmacology 244 (2010) 263–272

Fig. 2. (A–E) Assessment of DNA content in MSCs after 24 h treatment with iAsIIIand untreated MSCs. DNA content analysis was carried out by propidium iodide staining and flowcytometry after 24 h of treatment. Images correspond to typical histogram distributions for increasing concentrations of iAsIII. A total of 3000 events were counted for eachconcentration. The percentage of cells in the G2/M or in S phase of the cell cycle was estimated using ModFIT software. (F) Demonstrating cell cycle regulatory proteins from the top,Cyclin D1, P27 kip 1, P21 waf1/cip1, Cyclin B1, Cdc2 and loading control (i.e., β-Actin) in iAs III (10 μM) treated MSCs for the period 24 h; UT denotes untreated MSCs. Immunoblotshown is a representative blot from at least three samples per treatment regimen.

267S. Yadav et al. / Toxicology and Applied Pharmacology 244 (2010) 263–272

(as shown in Fig. 2B and C). 25.25% and 27.08% cells were found in S-phase in the cell cycle treated with 10 and 20 μM of iAsIII, respectively,which showed that cells were also arrested in S-phase in treatedMSCsthan in untreated MSCs where the cells in S-phase were only 10.61%.

Treatment of iAsIII decreases expression level of G2-M phase regulatoryCDKs and Cyclins in MSCs

The next attempt was made to characterize the mechanism at themolecular level. In this section, the effects of iAsIII on expression of keyregulators of the G2 to M transition such as Cdc2 and CyclinB1 werestudied. It has been reported that Cyclin B1 and Cdc2 can be altered byarsenite exposure (Chao et al., 2006). To assess the role of iAsIII in G2-M phase regulatory CDKs and Cyclins inMSCs, the expression of CyclinB1 and Cdc2 along with other cell cycle regulatory proteins such asCyclin D1, p27 Kip1 and p21 Waf 1/Cip1 was studied (as demon-strated by western blot analysis in Fig. 2F). Results showed that iAsIII

treatment associatedwith the down-regulation of Cdc2 (fold decreasefrom the control 1.57; 1.86 for 12 and 24 h of treatment durationrespectively) and Cyclin B1 (fold decrease from the control 1.64; 1.71for 12 and 24 h of treatment duration respectively) expression in bothtreatment durations. However, iAsIII had no effect on the expressionlevel of Cyclin D1, p27 Kip1 and p21 Waf 1/Cip1 in MSCs.

Treatment of iAsIII induces apoptosis in MSCs as examined by TUNELand Annexin V-FITC fluorescence-activated cell sorting analysis

The data showed that iAsIII significantly inhibits the growth ofMSCs.Consequently, we sought to assess the effect of this compound on

apoptotic stimulation using TUNEL staining and fluorescence-activatedcell sorting analysis by Annexin V-FITC staining. As shown in Figs. 3Aand B, TUNEL data showed that iAs III treatment induced remarkablenumber of TUNEL-positive cells upon 24 h of the treatment period. Ahigher number of TUNEL-positive cells were found at the concentrationof 10 μM(9.5%)whereas themaximumnumber of TUNEL-positive cellsupon iAsIII treatmentwasnoticed at the concentration of 20 μM(14.5.%)as compared to the untreated MSCs (2.3%). For further confirmation ofapoptotic stimulation by iAsIII, Annexin V-FITC staining was applied todetect apoptotic stimulation. Concordant with TUNEL findings; iAsIII

induced Annexin V-positive cells in a dose-dependent manner for theperiod of 24 h (Figs. 3C and D). Treatment of iAsIII at the concentrationof 10–20 μM showed a significant increase in Annexin V-positive cells(10.3%; 12.9%, respectively) over the untreated MSCs. However, lowerconcentration of iAsIII (5 μM) did not stimulate significant Annexin V-positive cells, but an increase in Annexin V-positive cells has also beennoticed (2.3% Annexin V-positive cells) as compared to the untreatedMSCs.

iAsIII alters expression level of Bcl-2 family proteins and cytochrome crelease from mitochondria

To understand the mechanism behind apoptotic stimulation byiAsIII in MSCs, further evaluation was made to see expression level ofBcl-2 family proteins and cytochrome c. As expected, iAsIII treatmentfor 12 and 24 h up-regulates the expression level of proapoptoticprotein such as Bax, Puma and down-regulates the expression level ofanti-apoptotic such as Bcl-2 family proteins. In contrast, Bad remainedunchanged (data not shown) whereas phospho-Bad fluctuated

Fig. 3. TUNEL analysis in iAsIII -treated MSCs with a concentration of 10 and 20 μM iAsIII for 24 h. (A, i and ii ) DAPI staining in untreated MSCs, no signals were captured in case ofTUNEL staining in untreated MSCs, therefore image is not shown. (iii and iv) DAPI staining in iAsIII treated MSCs. (v and vi) TUNEL-positive nuclei (TMR red-stained) in iAsIII treatedMSCs. (vii and viii) DAPI and TUNELmerged images, from treatedMSCs; arrows (red) indicate TUNEL-positive nuclei. (B) Percentages (%) of TUNEL-positive cells among untreated oriAsIII treatedMSCs. The bars represent themean TUNEL-positive cell percentage (±SD) from three samples per treatment regimen (with the significance level of *pb0.05 & **pb0.01).(C) Percentages (%) of Annexin V-positive cells among all control or iAsIII treated MSCs from duplicate analyses/regimen. A total of 3000 events were counted in each analysis (withthe significance level of **pb0.01). (D) Representative scattergrams from flow cytometry profile represents Annexin V-FITC staining in the x axis and PI in the y axis.

268 S. Yadav et al. / Toxicology and Applied Pharmacology 244 (2010) 263–272

moderately. In addition, as shown in Figs. 4A and B, a slight up-regulation in p53 expression was seen at 24 h of treatment durationand their relative expression (Figs. 4C and D). Furthermore, todetermine the involvement of mitochondrial pathway in this process,redistribution of cytochrome c was examined. Data show that iAsIII

triggers cytochrome c release from mitochondria into the cytoplasmat both time points (12 and 24 h) as shown in Figs. 5A–D.

iAsIII induces activation of Caspase-3, Pro caspase 9 and Poly(ADP-ribose) polymerase (PARP) in MSCs

To elucidate the downstream molecular pathway (s) involved iniAsIII-induced apoptosis, expressions of cleaved caspase-3, pro-caspase-9 and Poly (ADP-ribose) polymerase (PARP) were studiedin MSCs, following 12 and 24 h treatment duration of iAsIII (20 μM).

Fig. 4. Expression levels of Bcl-2 family member proteins in iAsIII-treated MSC for the period of 12 and 24 h. (A, B) From the top, expression level of pBad, Puma, P53, Bax, Bcl-2 andloading controls (i.e., β-actin and GAPDH). Immunoblot shown is a representative blot from at least two generated per treatment regimen. (C, D) Schematic bars demonstrateexpression patterns of (C) Puma, Phospho Bad and P53 (D) Bcl-2 and Bax protein expression (values obtained from integrated density measurement) in untreated and iAsIII-treatedMSCs. The bars in each case represent themean values (±SD); reflective of specific proteins expression as integrated density (measured by an odyssey infrared imaging system fromthree samples per treatment regimen).

269S. Yadav et al. / Toxicology and Applied Pharmacology 244 (2010) 263–272

Cleaved caspase-3 was detected in iAsIII treated MSCs in bothdurations of treatment (12 and 24 h). Results also showed a decreasedexpression level of procaspase-9 but the expression of cleavedcaspase-9 was not detectable (Fig. 5E).

iAsIII induces apoptosis in MSCs by suppressing Bcl-2a and byup-regulating expression of Bax

In addition for inhibition of the cell cycle, the apoptotic effects ofiASIII have been obtained from above data and found to be precededby suppression of Bcl-2a, member of the Bcl-2 family of antiapoptoticproteins. Indeed, Bcl-2a mRNA expression was found suppressed iniAsIII (20 μM) treated MSCs as compared with untreated MSCs (asshown in Table 2). In contrast, Bax m-RNA level was significantly up-regulated at 24 h, showing a good correlation between mRNA andprotein expression levels.

Discussion

Inorganic arsenic is ubiquitous in the environment, where itexists as arsenite (iAsIII) and arsenate (iAsV). At chronic level,arsenite, in the range of 10–50 μg/L, is present in drinking water inNevada, California (United States) (Acien et al., 2009), whereas insome part of the world it is found considerably higher (up to the100 to 1000 μg/L) (Kile et al. 2009; Chen et al. 2007; Navarro et al.,2004). Although water supplies in the United States are low inarsenic contamination, there have been reports in the Southwest

area on arsenic contamination in groundwater, with levels of100 μg/L and in few cases it was more than 1000 μg/L (Smith et al.,2002;Warner et al., 1999). In an overdose study, arsenic concentrationin serum was noticed up to 6.3 μM and in urine it was up to 253 μM(Kim and Abel, 2009). In this study, we tested the biological effects ofarsenic on human mesenchymal stem cells (MSCs) at the concentra-tion between 1 to 40 μM (75–3000 μg/L). The rationale of the study isthat absorbed arsenic is likely to reachbonemarrow to affect theMSCs.The concentrations used in this study overlap with environmentalexposures in chronic or in acute levels.

The responses of mesenchymal stem cells to arsenite exposure atthe selected concentration range were investigated in this study. Atthe concentration of 10 and 20 μM for 24 h exposure, arsenitetriggered apoptosis in mesenchymal stem cells. Previous studies havedemonstrated that arsenite induced apoptosis in different cell types,such as mouse embryonic fibroblasts, embryonic primary ratmidbrain neuroepithelial cells (Newman et al., 2008; Sidhu et al.,2006), as well as on cancerous cell lines such as A375 humanmalignant melanoma cells, SH SY5Y neuroblastoma and in leukemiccell (Ruiz-Ramos et al., 2009; Shavali and Sens , 2007; Nasr et al.,2008). There is a lack of information about interaction betweenarsenic and human MSCs, although, studies have investigated theresponses of arsenite on human progenitor or animal stem cells(Ferrario et al., 2009; Huang et al., 2009; Newman et al., 2008).Progenitor cells from human umbilical cord blood could be tolerant toarsenite exposure at the concentration of 1 μM for 24 h and apoptoticchanges could only be detected after 7 days of exposure at this

Fig. 5. Expression levels of cytochrome c, and Poly (ADP-ribose) polymerase (PARP) caspases (3 and 9) and Poly (ADP-ribose) polymerase (PARP) activity in iAsIII-treated MSC forthe period of 12 and 24 h. (A, B) Expression level of cytochrome c release in mitochondrial fraction and in cytosolic fraction, with their respective loading controls HSP60 and β-actin.(C, D) Schematic bars demonstrate protein expression patterns of (C) mitochondrial cytochrome c and (D) cytosolic cytochrome c expression (values obtained from integrateddensity measurement) in untreated and iAsIII-treated MSCs. (E) Expression level of caspase 3, 9 and PARP activity in MSCs after 12 and 24 h treatment with iAsIII. Immunoblot shownis a representative blot from at least three samples per treatment regimen.

270 S. Yadav et al. / Toxicology and Applied Pharmacology 244 (2010) 263–272

concentration (Ferrario et al., 2009). In our experiments, arseniteexposure at the concentration of 1 μM for 6–48 h could not induceapoptotic changes in MSCs. These studies indicate that stem/progenitor cells are more resistant to arsenic exposure which isconsistent with a previous study with progenitor and differentiatedNG108-15 cells (Yang et al., 2008).

Here, we extended these findings by showing that arsenite is alsoable to trigger apoptosis in MSCs by G2-M arrest and altering Bcl-2family proteins followed by activation of intrinsic apoptotic pathway.Treatment of iAsIII can arrest cell cycle at G2-M phase through a directinteraction with cell cycle regulatory proteins in differentiated cells(Nasr et al., 2008; Li and Yang 2007; Halicka et al., 2002). On the otherhand, it can also accumulate cells at S-phase of the cell cycle

Table 2

Name of gene Genesymbol

Fold-up (denoted as +) and down-regulation(denoted as −) after 24 h treatmentwith iAsIII (200 μM)

Bcl-2-associated Xprotein

Baxl +2.98⁎ (fold change from untreated MSCs)

Bcl-2-related protein a Bcl-2a −4.1⁎⁎ (fold change from untreated MSCs)Bcl-2-associated xlprotein

Bcl-xl −2.25⁎ (fold change from untreated MSCs)

(McCollum et al., 2005). The cyclin B1 with Cdc2 governs cell cycleprogression in G2-M phase. In this study, treatment of iAsIII (≥5 μM)for the duration of 24 h influenced down-regulation of Cdc2 and CyclinB1 expression at the duration of 24 h. Thus, interference in Cyclin B1and Cdc2 expression could directly impair the formation of Cdc2/Cyclin B1 complexes which is essential for the progression or tran-sition of G2-M phase (Porter et al., 2003).

This study demonstrates hallmarks of apoptosis in iAsIII treatedMSCs,including exposure of intramembrane phosphotidylserine and increasedDNA strand breaks. Further molecular mechanism showed alteration inBcl-2 family proteins in the apoptotic event. It is well known thatimbalance of Bcl-2 family member proteins can trigger apoptosis viaintrinsic pathway. Accumulating evidence (Reed and Pellecchia, 2005;Basu and Haldar, 1998) suggests that release of cytochrome c frommitochondria during apoptosis is controlled by the proteins (of Bcl-2family): those that inhibit cell death such as Bcl-2 and including thosethat promote cell death for example Bax (Murphy et al., 2000). Resultsshow the presence of an intrinsic pathway in iAsIII-induced apoptosis inMSCs including up-regulation in Bax and down-regulation in Bcl-2protein expression as compared to untreated controls. Cytochrome creleased from the mitochondria and increased into the cytoplasmicfraction supports this pathway and gets strength with the presence ofcaspase-9 activation (Wang, 2001; Green and Reed 1998). Subsequently,

Fig. 6. The proposed model of iAsIII mediated cell cycle arrest and apoptosis in MSCs. iAsIII induced DNA damage and altered Bcl-2 family proteins. Release of cytochrome c frommitochondria promotes caspase-3 activation causing apoptosis in MSCs. G2/M arrest may be involved in apoptotic induction but further investigation is required.

271S. Yadav et al. / Toxicology and Applied Pharmacology 244 (2010) 263–272

activated caspase-9 triggers downstream caspases, such as caspase-3.These downstream caspases, through the cleavage of several deathsubstrates, such as poly ADP-ribose polymerase (PARP), causedexecution of cell death (Basu and Haldar, 1998). Altogether, our dataclearly delineate arsenite triggered cell death via intrinsic pathway byaltering Bcl-2 familymember proteins. In this study, we found that iAsIII-induced apoptosis in MSCs not only led to accumulation of proapoptoticBax and Puma, but also caused remarkable reduction of anti-apoptoticprotein Bcl-2 as well as in m-RNA expression levels.

The key issue explored in this study is that in future in vitrotoxicology studies, must consider how mesenchymal stem cells mayrespond to toxicants. Traditional cell-based, in vitro toxicity studies arebased on transformed human cell lines or immortalized tumor cells.These cells were obtained from cancerous tissues carrying mutatedgenes or chromosomal instability which is directly involved in cell cyclekinetics, cell death pathways or drug detoxification. The genetic back-ground of these cellsmay complicate the analyses of cytotoxicity relatedcell signaling pathways. MSCs provide a much cleaner system and theuntransformed culture should theoretically provide more accuratemodeling of in vivo condition, ensuring that the results are more com-parable to in vivo effects. In addition, thismodelmay allow prediction ofcytotoxicity at both the developmental and mature stages.

Conclusion

We present that MSC can be useful cell model for the study oftoxicity. In particular, we demonstrate for the first time that iAsIII

induce apoptosis in MSCs by altering Bcl-2 family and activating

intrinsic pathway as outlined in Fig. 6. These findings suggest newperspectives for understanding the underlying mechanism of toxi-cant-induced apoptosis in MSCs, which might contribute to thedevelopment of new predictive screening methods for the riskassessment of developmental toxicity using stem cell technology.

Acknowledgment

This project is supported by Tulane Cancer Centre & EnvironmentalHeath Science, Tulane University, New Orleans, LA, USA.

References

Acien, A.N., Umans , J.G., Howard, B.V., Goessler, W., Francesconi, K.A., Crainiceanu, C.M.,Silbergeld , E.K., Guallar, E., 2009. Urine arsenic concentrations and speciesexcretion patterns in American Indian communities over a 10-year Period: theStrong Heart Study. Env. Health Perspect. 117 (9), 1428–1433.

Basu, A., Haldar, S., 1998. The relationship between Bcl-2, Bax and p53: consequencesfor cell cycle progression and cell death. Mol. Hum. Reprod. 4, 1099–1109.

Chao, J.I., Hsu, S.H., Tsou, T.C., 2006. Depletion of securin increases arsenite-inducedchromosome instability and apoptosis via a p53-independent pathway. Toxicol. Sci.90 (1), 73–86.

Chen, Y., van Geen, A., Graziano, J.H., Pfaff, A., Madajewicz, M., Parvez, F., 2007.Reduction in urinary arsenic levels in response to arsenic mitigation efforts inAraihazar, Bangladesh. Environ. Health Perspect. 115, 917–923.

Concha, G., Vogler, G., Lezcano, D., Nermell, B., Vahter, M., 1998. Exposure toinorganic arsenic metabolites during early human development. Toxicol. Sci. 44,185–190.

Dopp, E., Hartmann, L.M., von Recklinghausen, U., Florea, A.M., Rabieh, S., Zimmermann,U., Shokouhi, B., Yadav, S., Hirner, A.V., Rettenmeier, A.W., 2005. Forced uptake oftrivalent and pentavalent methylated and inorganic arsenic and its cyto-/genotoxicity in fibroblasts and hepatoma cells. Toxicol. Sci. 87 (1), 46–56.

272 S. Yadav et al. / Toxicology and Applied Pharmacology 244 (2010) 263–272

Ebert, R., Ulmer, M., Zeck, S., Meissner-Weigl, J., Schneider, D., Stopper, H., Schupp, N.,Jakob Kassem, M.F., 2006. Selenium supplementation restores the antioxidativecapacity and prevents cell damage in bonemarrow stromal cells in vitro. Stem Cells24 (5), 1226–1235.

Ferrario, D., Croera, C., Brustio, R., Collotta, A., Bowe, G., Vahter, M., Gribaldo, L., 2008.Toxicity of inorganic arsenic and its metabolites on haematopoietic progenitors“in vitro”: comparison between species and sexes. Toxicology 249 (2–3),102–108.

Ferrario, D., Collotta, A., Carfi, M., Bowe, G., Vahter, M., Hartung, T., Gribaldo, L., 2009.Arsenic induces telomerase expression and maintains telomere length in humancord blood cells. Toxicology 16, 132–141.

Green, D.R., Reed, J.C., 1998. Mitochondria and apoptosis. Science 281, 1309–1312.Grogan, S.P., Miyaki, S., Asahara, H., D'Lima, D.D., Lotz, M.K., 2009. Mesenchymal

progenitor cell markers in human articular cartilage: normal distribution andchanges in osteoarthritis. Arthritis Res. Ther. 5, 85.

Halicka, H.D., Smolewski, P., Darzynkiewicz, Z., Dai, W., Traganos, F., 2002. Arsenictrioxide arrests cells early in mitosis leading to apoptosis. Cell Cycle 1, 201–209.

Hill, D.S., Wlodarczyk, B.J., Mitchell, L.E., Finnell, R.H., 2009. Arsenate-induced maternalglucose intolerance and neural tube defects in a mouse model. Toxicol. Appl.Pharmacol. 239, 29–36.

Huang, Z., Li, J., Zhang, S., Zhang, X., 2009. Inorganic arsenic modulates the expression ofselenoproteins in mouse embryonic stem cell. Toxicol. Lett. 187 (2), 69–76.

IARC., 1980. Some metals and metallic compounds (IARC Monographs on theEvaluation of Carcinogenic Risk to Humans, Vol. 23.

Johnstone, B., Yoo, J.U., 1999. Autologous mesenchymal progenitor cells in articularcartilage repair. Clin. Orthop. Relat. Res. 367, S156–162.

Kile, M.L., Hoffman, E., Hsueh, T.M., Afroz, S., Qamruzzaman, Q., Rahman, M., Mahiuddin,G., Ryan, L., Christiani, D.C., 2009. Variability in biomarkers of arsenite exposure andmetabolism in adults over time. Environ. Health Perspect. 17 (3), 455–460.

Kim, L.H., Abel, S.J., 2009. Survival after a massive overdose of arsenic trioxide. Crit. CareResusc. 11 (1), 42–45 2009.

Kumar, A., Adak, P., Gurian, P.L., Lockwood, J.R., 2009. Arsenic exposure in US public anddomestic drinking water supplies: a comparative risk assessment. J Expo SciEnviron Epidemiol. (Epub ahead of print).

Li, J.P., Yang, J.L., 2007. Cyclin B1 proteolysis via p38 MAPK signaling participates in G2checkpoint elicited by arsenite. J. Cell. Physiol. 212 (2), 481–488.

Lianos, M.N., Ronco, A.M., 2009. Fetal growth restriction is related to placental levels ofcadmium, lead and arsenic but not with antioxidant activities. Reprod. Toxicol. 27,88–92.

Ludwika, K., Christoph, S., Ulla, P., Wilhelm, N., 1996. Radiation-induced DNA damage incanine hemopoietic cells and stromal cells as measured by the comet assay.Environ. Mol. Mutagen. 27, 139–145.

Lund, P., Pilgaard, L., Duroux, M., Fink, T., Zachar, V., 2009. Effect of growth media andserum replacements on the proliferation and differentiation of adipose-derivedstem cells. Cytotherapy 11, 189–197.

McCollum, G., Keng, P.C., Christopher States, J., McCabe, M.J., 2005. Arsenite delaysprogression through each cell cycle phase and induces apoptosis following G2/Marrest in U937 myeloid leukemia cells. Am. Soc. Pharmacol. Exp. Ther. 313,877–887.

Murphy, K.M., Ranganathan, V., Farnsworth, M.L., Kavallaris, M., Lock, R.B., 2000. Bcl-2inhibits Bax translocation from cytosol to mitochondria during drug-inducedapoptosis of human tumor cells. Cell Death Differ. 7, 102–111.

Nasr, R., Guillemin, M.C., Ferhi, O., Soilihi, H., Peres, L., Berthier, C., Rousselot, P., Robledo-Sarmiento, M., Lallemand-Breitenbach, V., Gourmel, B., Vitoux, D., Pandolfi, P.P.,Rochette-Egly, C., Zhu, J., de Thé, H., 2008. Eradication of acute promyelocyticleukemia-initiating cells through PML-RARA degradation. Nat. Med. 4 (12), 1333–1342.

Navarro, P.A., Liu, L., Keefe, D., 2004. In vivo effects of arsenite onmeiosis, preimplantationdevelopment, and apoptosis in the mouse. Biol. Reprod. 70, 980–985.

Nemeth, M.J., Mak, K.K., Yang, Y., Bodine, D.M., 2009. Beta-Catenin expression in thebone marrow microenvironment is required for long-term maintenance ofprimitive hematopoietic cells. Stem Cells 27 (5), 1109–1119.

Newman, J.P., Banerjee, B., Fang, W., Poonepalli, A., Balakrishnan, L., Low, G.K.,Bhattacharjee, R.N., Akira, S., Jayapal, M., Melendez, A.J., Baskar, R., Lee, H.W.,Hande, M.P., 2008. Short dysfunctional telomeres impair the repair of arsenite-induced oxidative damage in mouse cells. J. Cell. Physiol. 214, 796–809.

Pessina, A., Malerba, I., Gribaldo, L., 2005. Hematotoxicity testing by cell clonogenicassay in drug development and preclinical trials. Curr. Pharm. Des. 11 (8),1055–1065.

Pessina, A., Albella, M.B., Bayo, J., Bueren, P., Brantom, S., Casati, C., Croera, R.,Parchment, D., Parent-Massin, G., Schoeters, Y., Sibiri, R., Van Den, H., Gribaldo, L.,2002. In vitro tests for haematotoxicity: prediction of drug-induced myelosuppres-sion by the CFU-GM assay. Altern. Lab. Anim. 2, 75–79.

Porter, L.A., Donoghue, D.J., Cyclin B1 and CDK1, 2003. Nuclear localization andupstream regulators. Prog. Cell Cycle Res. 5, 335–347.

Prockop, D.J., 2009. Repair of tissues by adult stem/progenitor cells (MSCs):controversies, myths, and changing paradigms. Mol. Ther. 17 (6), 939–946.

Reed, J.C., Pellecchia, M., 2005. Apoptosis-based therapies for hematologic malignan-cies. Blood 106 (2), 408–418.

Röllin, H.B., Rudge, C.V., Thomassen, Y., Mathee, A., Odland, J.Ø., 2009. Levels of toxicand essential metals in maternal and umbilical cord blood from selected areas ofSouth Africa–results of a pilot study. J. Environ. Monit. 11, 618–627.

Ruiz-Ramos, R., Lopez-Carrillo, L., Rios-Perez, A.D., De Vizcaya-Ruíz, A., Cebrian, M.E.,2009. Sodium arsenite induces ROS generation, DNA oxidative damage, HO-1 andc-Myc proteins, NF-kappaB activation and cell proliferation in human breast cancerMCF-7 cells. Mutat. Res. 31, 109–115.

Schnabel, L.V., Lynch, M.E., van derMeulen, M.C., Yeager, A.E., Kornatowski, M.A., Nixon,A.J., 2009. Mesenchymal stem cells and insulin-like growth factor-I gene-enhancedmesenchymal stem cells improve structural aspects of healing in equine flexordigitorum superficialis tendons. J. Orthop. Res. 27 (10), 1392–1398.

Schofield, R., 1978. The relationship between the spleen colony-forming cell and thehaemopoietic stem cell. Blood Cells 4, 7–25.

Shavali, S., Sens, D.A., 2007. Synergistic neurotoxic effects of arsenic and dopamine inhuman dopaminergic neuroblastoma SH-SY5Y cells. Toxicol. Sci. 102, 254–261.

Sidhu, J.S., Ponce, R.A., Vredevoogd, M.A., Yu, X., Gribble, E., Hong, S.W., Schneider, E.,Faustman, E.M., 2006. Cell cycle inhibition by sodium arsenite in primaryembryonic rat midbrain neuroepithelial cells. Toxicol. Sci. 89 (2), 475–484.

Smith, A.H., Lopipero, P.A., Bates, M.N., Steinmaus, C.M., 2002. Public health. Arsenicepidemiology and drinking water standards. Science 296, 2145–2146.

Vahter, M., 2008. Health effects of early life exposure to arsenic. Basic Clin. Pharmacol.Toxicol. 102, 204–211.

Wang, X., 2001. The expanding role of mitochondria in apoptosis. Gene Dev. 15,2922–2933.

Warner, M.L., Moore, L.E., Smith, M.T., Kalman, D.A., Fanning, E., Smith, A.H., 1999.Increased micronuclei in exfoliated bladder cells of individuals who chronicallyingest arsenic-contaminated water in Nevada. Cancer Epidemiol. Biomarkers Prev.3 (7), 583–590.

Wen, L.P., Fahrni, J.A., Troie, S., Guan, J.-L., Orth, K., Rosen, J.A., 1997. Cleavage of focaladhesion kinase by caspases during apoptosis. J. Biol. Chem. 272, 26056–26061.

WHO guidelines for drinking water quality, 1996. Health Criteria and Other SupportingInformation, 2. WHO, Geneva, pp. 940–949.

Yang, J., Oza, J., Bridges, K., Chen, K.Y., Liu, A.Y, Rosen, J.A., 2008. Neural differentiationand the attenuated heat shock response. Brain Res. 8, 39–50.