Ae

Ga

b

a

ARRA

KCCANGL

1

att(beaiend

0d

Aquatic Toxicology 99 (2010) 73–85

Contents lists available at ScienceDirect

Aquatic Toxicology

journa l homepage: www.e lsev ier .com/ locate /aquatox

poptosis and necroptosis are induced in rainbow trout cell linesxposed to cadmium

erhard Krumschnabela,∗, Hannes L. Ebnerb, Michael W. Hessb, Andreas Villungera

Division of Developmental Immunology, Biocenter, Medical University Innsbruck, Fritz-Preglstr. 3, Innsbruck, AustriaDivision of Histology and Embryology, Medical University Innsbruck, Innsbruck, Austria

r t i c l e i n f o

rticle history:eceived 17 February 2010eceived in revised form 30 March 2010ccepted 2 April 2010

eywords:admiumopperpoptosisecroptosisilliver

a b s t r a c t

Cadmium is an important environmental toxicant that can kill cells. A number of studies have implicatedapoptosis as well as necrosis and, most recently, a form of programmed necrosis termed necroptosis in theprocess of cadmium-mediated toxicity, but the exact mechanism remains ill-defined and may depend onthe affected cell type. This study investigated which mode of cell death may be responsible for cell deathinduction in cadmium-exposed trout cell lines from gill and liver and if this cell death was sensitive toinhibitors of necroptosis or apoptosis, respectively. It was observed that intermediate levels of cadmiumthat killed approximately 50% of the cells over 96–120 h of exposure caused cell death that morphologi-cally resembled apoptosis and was associated with an increase of apoptotic markers such as the numberof cells with diminished DNA content (sub-G1 cells), condensed or fragmented nuclei, and elevation ofcaspase-3 activity. At the same time, however, cells also lost plasma membrane integrity, as indicatedby uptake of propidium iodide, showed a decrease of ATP levels and mitochondrial membrane potential,and displayed cell swelling, signs associated with secondary necrosis, or equally possible, necroptotic celldeath. Importantly, many of these alterations were at least partly inhibited by the necroptosis inhibitornecrostatin-1 and were to a lesser extent also sensitive to the pan-caspase inhibitor zVAD-fmk, indicatingthat multiple modes of cell death are concurrently induced in cadmium-exposed trout cells, including

necroptosis and apoptosis. Cell death appeared to lack concurrent radical formation, consistent withgenetically regulated necroptotic cell death, but was characterized by the rapid induction of DNA damagemarkers, and the early onset of disintegration of the Golgi complex. Comparative experiments evaluat-ing copper-toxicity indicated that in comparison to cadmium much higher concentrations of this metalwere required to induce cell death and that neither necrostatin-1 nor a pan-caspase inhibitor conferred

at add

protection, suggesting thheavy metals.

. Introduction

Cadmium (Cd) is an important toxicant both in terrestrial andquatic environments. Depending on the cell type and concen-ration, it may induce cell death via apoptosis, characterized byypical features such as caspase activation and DNA fragmentationHabeebu et al., 1998; Pulido and Parrish, 2003), or necrosis typifiedy ATP depletion and plasma membrane permeabilization (Lopezt al., 2003; Yang et al., 2007). Furthermore, in some instancesless clearly defined mode of cell death was described involv-

ng both apoptotic and necrotic features (Lee et al., 2006; Sanchot al., 2006; Shih et al., 2003). Recently, a mode of programmedecrosis, termed necroptosis has been identified, in which celleath occurs with largely necrosis-like morphological alterations,

∗ Corresponding author.E-mail address: Gerhard.Krumschnabel@i-med.ac.at (G. Krumschnabel).

166-445X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.aquatox.2010.04.005

itional modes of cell death can be triggered in response to poisoning with

© 2010 Elsevier B.V. All rights reserved.

but following apparently well-defined molecular signaling path-ways that recruit components of the extrinsic as well as intrinsicapoptotic cell death machinery (Christofferson and Yuan, 2010). Acrucial component of this pathway is the serine/threonine kinasereceptor-interacting protein-1 (RIP-1), the pharmacological inhibi-tion of which by the small molecule inhibitor necrostatin-1 (Nec-1)was in fact key to identifying necroptosis (Degterev et al., 2005).A closer characterization of necroptosis was achieved by use ofthe mouse fibrosarcoma cell line L929 in which addition of thephysiological ligand tumor necrosis factor-� (TNF�), or inhibitionof caspases using zVAD-fmk induces cell death that is sensitiveto Nec-1 inhibition (Degterev et al., 2005). In addition to such areceptor-mediated, extrinsic stimulation, Nec-1 sensitive cell death

involving RIP-1 may also be triggered through intrinsic clues, inparticular in response to severe DNA damage (Festjens et al., 2006;Xu et al., 2006; Zong et al., 2004). In line with this, it was recentlyshown that Cd, which may induce DNA damage as well (Bertin andAverbeck, 2006; Joseph, 2009; Viau et al., 2008), induces necropto-

7 atic T

swlwmaRdNpibfd

vdFpntmstdbanfmuitmfimmttwsin(

2

2

wcBd1Hwladfrifa

4 G. Krumschnabel et al. / Aqu

is in Chinese hamster ovary cells (Hsu et al., 2009). Specifically, itas reported that Nec-1 did not prevent the elevation of intracel-

ular calcium levels and reactive oxygen species (ROS) associatedith Cd-exposure, whereas it attenuated Cd-induced reduction ofitochondrial membrane potential and plasma membrane perme-

bilization. This suggests that calcium and ROS act upstream ofIP-1, while mitochondrial alterations and ultimately the break-own of cell integrity are downstream of the kinase. In addition,ec-1 caused significantly enhanced DNA binding activity of thero-survival transcription factor NF�B, which might be taken to

ndicate that Nec-1 may not only protect from Cd-induced toxicityy inhibiting RIP-1 but also ameliorate DNA damaging effects byacilitating transcription of pro-survival genes thereby enhancingamage repair and subsequent cell death signaling.

Similar to the findings reported for mammalian cell systems,arious studies with teleost cells indicated apoptotic and necroticeath upon Cd-exposure (Lyons-Alcantara et al., 1998; Risso-deaverney et al., 2001; Xiang and Shao, 2003). So far, however, theossible occurrence of necroptotic cell death in response to Cd hasot been addressed in cells from teleosts. Thus, despite the facthat apoptosis signaling appears largely conserved among fish and

ammals (Krumschnabel and Podrabsky, 2009), it has not yet beenhown if this also holds for necroptosis. The present study washerefore conducted to elucidate if Cd-exposure of teleost cell lineserived from trout elicits cell death that is sensitive to inhibitiony Nec-1, and to characterize the physiological alterations associ-ted with this. Furthermore, although such an in vitro study doesot reflect a truly environmentally relevant situation, since fish are

requently exposed to this pollutant we also wanted to gain infor-ation regarding the in vivo toxicity of Cd in fish. We therefore

sed both a gill cell line, RTgill-W1, representing the cell type thats usually at first and most directly exposed to environmental pollu-ants, and a liver cell line, RTH-149, representing those cells where

any toxicants including metals may be accumulated and detoxi-ed. The occurrence of specific cell death features in either cell typeay thus provide information pertaining to their use as environ-ental markers of, e.g. the effects of acute versus chronic exposure

o Cd and/or information related to concentration–response rela-ions characteristic for each cell type. For reasons of comparison,e also examined what type of cell death is induced by copper, a

econd relevant pollutant, in these trout cells, since previous stud-es on primary trout hepatocytes indicated mixed apoptotic andecrotic death but did not address the possibility of necroptosisKrumschnabel et al., 2005; Nawaz et al., 2006).

. Materials and methods

.1. Chemicals and antibodies

Leibovitz L-15 medium and fetal bovine serum (FBS)ere from Gibco (Invitrogen, Vienna), l-glutamin, peni-

illin/streptomycin and trypsin/EDTA were from Cambrexioscience (Oberhaching, Germany). Fluorescence indicatorsichlorofluorescein diacetate (DCF-DA), 5,5V,6,6V-tetrachloro-,1V,3,3V-tetraethylbenzimidazolylcarbocyanine iodide (JC-1),oechst 33342, and tetramethylrhodamine methylester (TMRM)ere obtained from Molecular Probes (Leiden, The Nether-

ands). CellTiter-Glo and the Apo-ONE homogeneous caspase-3/7ssay kit were from Promega (Mannheim, Germany), 3-(4,5-imethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)

rom Roche Diagnostics (Vienna, Austria), the poly-(ADP-ibose) polymerase inhibitor 3-aminobenzamide, propidiumodide (PI) and 4′,6-diamidino-2-phenylindole (DAPI) wererom Sigma (Deisenhofen, Germany). Necrostatin-1 (Nec-1)nd the hsp-90 inhibitor 17-(dimethylaminoethylamino)-17-

oxicology 99 (2010) 73–85

demethoxygeldanamycin (17-DMAG) were from Eubio (Vienna,Austria), the pan-caspase inhibitor Z-Val-Ala-dl-Asp(OMe)-fluoromethylketone (zVAD-fmk) from Bachem (Weil am Rhein,Germany), cyclosporine A (CsA) from LC Laboratories (Woburn,MA, USA). Inhibitors of mitogen-activated protein kinases U0126and SB203580 were obtained from Calbiochem (Merck, Darmstadt,Germany), SP600125 from A.G. Scientific Inc. (Munich, Germany).N-(2-Quinolyl)valyl-aspartyl-(2,6-difluorophenosy)methyl ketone(QVD) was from Alexis Biochemicals (Lausen, Switzerland).

Antibodies against �H2AX (#9718), phospho-ATM (#4526),phospho-Chk-2 (#2661) were purchased from New EnglandBiolabs (Frankfurt am Main, Germany), FITC-labeled mouse anti-GM-130 was from BD Transduction Laboratories (Vienna, Austria),and secondary antibodies Alexa Fluor 488 goat anti-rabbit andAlexa Fluor 546 goat anti-mouse from Invitrogen (Vienna, Austria).Cd and Cu were applied from stock solutions of CdCl2 and CuCl2dissolved in distilled water.

2.2. Cell lines and cell culture

Rainbow trout-derived cell lines, the gill cell line RTgill-W1 andthe hepatoma line RTH-149, were obtained from ATCC (Manassas,VA, USA) and were cultured in Leibovitz L-15 medium with 10%FBS, 2 mM l-glutamine and 1% penicillin–streptomycin at 21 ◦C inan air atmosphere. For experimental exposures, which were alsoall conducted at 21 ◦C in air atmosphere, serum concentration wasreduced to 0.5%, which was found adequate to still support cellgrowth but to reduce potential binding of metals.

2.3. Experimental exposures

Cells were initially exposed to concentrations between 10 and500 �M Cd and Cu over a time period ranging from a few hoursto 5 days. Having established metal concentrations and exposuretimes leading to intermediate levels of cell death, a number ofdifferent inhibitors of signaling molecules involved in classicalapoptosis or in necroptosis was applied and the impact of theseinhibitors on viability and cell death-related parameters deter-mined in metal exposed cells was evaluated. Concentrations ofmetals and inhibitors as well as specific time points are as detailedin Section 3 for each experiment.

2.4. Cell viability and caspase activity measurements

Cell viability was determined using the MTT kit from Roche asinstructed by the suppliers. In brief, cells were seeded into 96-wellplates at a density of 15,000/well (RTgill-W1) or 7500/well (RTH-149) and incubated at 21 ◦C overnight prior to experimentation.Then cells were once washed with low-serum medium and subse-quently exposed to the desired treatment as indicated in Section3. If specific inhibitors were used, a 1 h pre-incubation period pre-ceded experimental exposures. At the end of the incubation periodMTT was added and the cells left to metabolize the tetrazolium saltfor 4 h. Following addition of solubilization solution and incubationfor >6 h, absorbance was measured on a multi-well plate reader at595/620 nm.

For the determination of caspase activity, cells were cultured inthe same way and at the end of experimental treatments caspase-3/7 assay mix was added to each well, the reaction incubated for30 min and fluorescence was then measured at 485 nm excitation

and 538 nm emission. Basically the same holds for measurementsof ATP contents, for which after treatments the CellTiter-Glo assaymix was added to wells and after 30 min of incubation lumines-cence values were determined in a multi-well plate reader withluminescence detection.

atic T

D1etfiw5Fg

apb2duttw

2

fit1wisdtw(witaCn6Aetct

2

uettripr

wdqoe5

G. Krumschnabel et al. / Aqu

A marker of apoptotic cells is the occurrence of diminishedNA content, as assessed by the sub-G1 assay (Nicoletti et al.,991). For this purpose cells were cultured in 6 cm cell dishes andxposed to treatments before cells were collected by trypsiniza-ion (including floating cells), pelleted by centrifugation and thenxed in 70% ice-cold ethanol. After one wash with PBS cellsere RNAse digested for 20 min at 37 ◦C and then incubated with

0 �g/ml PI for 10 min before DNA contents were analyzed with aACScalibur flow cytometer (BD Biosciences, Erembodegem, Bel-ium).

For the determination of plasma membrane permeabilizations occurs in necrosis, cells were cultured and treated in 24-welllates, and subsequently co-incubated with 5 �g/ml of the mem-rane impermeable dye PI, which only enters dead cells, and�g/ml Hoechst 33342, a nuclear stain permeating viable andead cells. Following 5 min of incubation, cells were photographednder a fluorescence microscope using excitation/emission fil-ers specific for Hoechst 33342 and PI and the number ofotal and PI-positive cells were analyzed using Image J soft-are.

.5. Immunofluorescence stainings

Cells were seeded onto sterilized glass cover slips and left tormly attach for at least 24 h. The cultures were then exposed tohe desired treatment, briefly washed with PBS, and then fixed for5 min at room temperature with 4% PFA (in PBS). Following threeashes with PBS cells were incubated for 1 h with PBS contain-

ng 0.1% Triton X-100, 1% bovine serum albumin and 10% fetal calferum for permeabilization and blocking. Primary antibodies wereiluted 1:50 (anti-GM-130) or 1:100 (all others) in blocking solu-ion and samples incubated overnight at 4 ◦C. After three furtherashes with PBS, cells were incubated with secondary antibody

1:100 in blocking solution) for 1 h at room temperature, washedith PBS, incubated with DAPI (2 �g/ml, in PBS) for nuclear stain-

ng for 10 min at room temperature and washed one more finalime. Cells were then fixed on microscope slides with Vectashieldntifade mounting medium (Vector Laboratories Burlingame, CA).ell images were collected with a Leica SP5 confocal laser scan-ing microscope (Leica Microsystems, Wetzlar, Germany) using a3× glycerol immersion objective, with acquisition through the LASF acquisition software Version 2.1.0, applying three averages onach channel, a resolution of 1024 × 1024 pixels, and pinholes seto 1 Airy unit. Post-acquisition image processing, i.e. backgroundorrection, adjustment of brightness and contrast and export toif-format, were done with Image J software.

.6. Quantification of mitochondrial membrane potential (MMP)

Alterations of MMP were determined with JC-1 loaded cellssing FACS analysis. In brief, cells were cultured in 6-well plates,xposed to stimuli, trypsinized and then incubated for 30 min inhe dark with 4 �M JC-1 dissolved in Leibovitz medium. Cells werehen spun down in Eppendorf tubes, washed once with PBS ande-suspended in PBS. FACS measurements were done on a FACScal-bur flow cytometer determining red (cells with high mitochondrialotential) and green (cells with low mitochondrial potential) fluo-escence.

As an alternative method for the quantification of MMP cellsere cultured in 96-well plates as above (Section 2.4), treated as

esired and then incubated with 200 nM TMRM for 30 min. Subse-uently the wells were once washed with PBS before re-additionf saline and then fluorescence emitted from TMRM retained innergized mitochondria was determined at 544 nm excitation and90 nm emission.

oxicology 99 (2010) 73–85 75

2.7. Production of reactive oxygen species (ROS)

The formation of reactive oxygen species was examined in cellsseeded and exposed in 96-well plates as described for the MTTviability assay (Section 2.4). After experimental exposure well con-tents were aspired, 50 �l PBS containing 5 �M of the ROS indicatorDCF-DA were added and cells incubated in the dark for 30 minbefore measuring fluorescence at 485 nm excitation and 538 nmemission.

2.8. Electron microscopy

For transmission EM samples were subjected to rapid cryo-immobilisation by means of high-pressure freezing instead ofchemical fixation, which greatly improves the preservation of ultra-structural features. In brief, cells were cultured on carbon-coatedsapphire coverslips, stimulated as required and subjected to high-pressure freezing followed by freeze-substitution and epoxy resinembedding (Hess et al., 2000). Sections were analyzed with aPhilips CM120 EM (F.E.I., Eindhoven, Netherlands) and images wererecorded with a MORADA digital camera (SIS, Münster, Germany).Contrast and brightness of the digital images were optimized byusing grey scale modification and high-pass filtering of the ADOBEphotoshop software (Version 9).

2.9. Time-lapse imaging

For live cell imaging cells were cultured in 12-well plates, chemi-cals added as appropriate, the plates sealed with Parafilm to preventevaporation and mounted on the stage of a Cell-IQ real-time imagecapture system (Chip-Man Technologies Ltd, Tampere, Finland).Images were then captured at 30 min or 1 h intervals and time-stamped and exported as movie files by use of Image J software.

2.10. Statistics

Data shown are means ± SE of n cultures. All experiments wereconducted on at least 3 different sub-cultures individually main-tained before the experiments. In addition, all experiments wererepeated at least twice on cells from different passage numbers.Statistical differences were evaluated by ANOVA followed by theappropriate post hoc tests, with a p value <0.05 considered as sig-nificant.

3. Results

3.1. Concentration–response relationships of metals and cellviability

A first series of experiments investigated the concentration–response relation between the concentrations of Cd or Cu and cellviability. Pilot experiments had shown that rather high concen-trations of Cd (>200 �M) were required to induce significant celldeath within 24 h and this was then almost exclusively necrotic(not shown). Thus, lower concentrations and prolonged exposuretimes were chosen and it was observed that after 120 h of incu-bation a gradual concentration-dependent decline in cell viabilityoccurred in RTgill-W1 cells, with about 50% cell death determinedin the presence of 50 �M Cd (Fig. 1A). RTH-149 hepatoma cells weresomewhat more sensitive and showed nearly 60% cell death after96 h of exposure to 50 �M Cd (Fig. 1B). Cu appeared to be less toxic

to both cell lines and in RTgill-W1 cells we even observed a slightproliferative effect of Cu up to concentrations of 100 �M (Fig. 1A).Above this level, however, Cu reduced cell viability in both cell linesand at 1 mM it killed 85% and 90% of all gill and hepatoma cells,respectively (Fig. 1A and B).

76 G. Krumschnabel et al. / Aquatic Toxicology 99 (2010) 73–85

Fig. 1. Concentration-dependent cell death of trout cells exposed to Cd or Cu. Cell viability, as estimated from MTT conversion, of RTgill-W1 cells exposed to Cd or Cu at thei nel (Co n them depen

acrsiuttea

3

oacwndli

ndicated concentration for 120 h (A) and of RTH-149 cells exposed for 96 h (B). Paf cells with diminished DNA content and in (D) the quantitative relation betweeeans ± SE of 3 (Cd in RTgill-W1 and panel D) and 4–7 cultures used in at least 2 in

Concentration-dependent changes elicited by the metals werelso established for the number of cells with diminished DNAontent, the sub-G1 population, which is generally regarded asepresenting cells that have undergone apoptotic cell death. Ashown for examples in Fig. 1C and summarized in Fig. 1D, Cdnduced a concentration-dependent increase in the sub-G1 pop-lation up to concentrations of 200 �M in RTgill-W1 cells and upo 500 �M in the RTH-149 cells (supplementary Fig. S1). In con-rast, Cu caused no significant increase of the sub-G1 population inither cell type irrespective of the concentration applied (Figs. 1Cnd S1).

.2. Cellular and nuclear morphology

Morphological alterations of gill cells determined at the endf the 120 h incubation period with 50 �M Cd indicated partlypoptotic features, as evidenced by retracting but still attachedell bodies and heavily condensed or fragmented nuclei, and this

as even more pronounced at 200 �M Cd (Fig. 2). However, theumber of apparently apoptotic nuclei could not account for theecrease of viability assessed by the MTT assay and when morpho-

ogical changes were followed over time, using time-lapse imaging,t became clear that many cells underwent necrotic death asso-

) shows examples of FACS analyses of gill cells used to determine the percentageconcentration of Cd and Cu and the number of sub-G1 cells is depicted. Data aredent experiments. N.d.: not determined.

ciated with typical cell swelling (supplementary movie S1). Incontrast, with 200 �M Cu only mild morphological cellular changeswere evident and also nuclear appearance seemed unaltered inRTgill-W1 cells (Fig. 2). A mixed apoptotic/necrotic death pheno-type was also observed in RTH-149 hepatoma cells exposed toCd at 50 �M and 200 �M (Fig. S2 and movie S2). For compari-son, movies of primarily necrotic death occurring during exposureto 500 �M Cd and of the onset of apoptotic death after addi-tion of 0.2 �M of the pan-kinase-inhibitor staurosporine (STS) areshown as supplementary movies S6 and S7 (also compare stills inFig. S2, right panels).

3.3. Mitochondrial membrane potential (MMP)

While changes of cell viability or DNA contents took relativelylong to become apparent, one may expect that pathways signal-ing or mediating cell death should be detectable at earlier timepoints. As such, alterations of MMP may be an early event and

indeed we found that after only 24 h a concentration-dependentincrease in the population of gill cells with diminished MMP wasevident (Fig. 3A). At 48 h of incubation with 50 �M Cd individualcells with condensed perinuclear mitochondria could be observed(Fig. 3B) and after 72 h 23% (50 �M Cd) to 55% (500 �M Cd) of the

G. Krumschnabel et al. / Aquatic Toxicology 99 (2010) 73–85 77

F or 120i .

c(do

3

apdeyqmwa

ig. 2. Cellular and nuclear morphology of RTgill-W1 cells exposed to Cd or Cu fndicated, and of their DAPI-stained nuclei as observed by fluorescence microscopy

ell population had a low MMP, compared to 13% in the controlsFig. 3C and D). In Cu-treated cells no comparable changes were evi-ent at these time points as determined by measuring the retentionf TMRM (not shown).

.4. Induction of DNA damage response markers

Another early event induced by toxic levels of metals is theppearance of DNA damage. In the present context this was ofarticular interest, as the intrinsic leg of programmed necrotic celleath was shown to be triggered upon activation of the DNA repairnzyme poly-(ADP-ribose) polymerase-1 (PARP-1), which by not

et clearly defined mechanisms may activate RIP-1 and in conse-uence necroptosis (Festjens et al., 2006). We thus tested whetherolecules regarded as typical markers of the DNA damage responseould be activated upon exposure of the fish cells to Cd or Cu. Ascontrol treatment known to induce DNA double-strand breaks,

h. Typical examples of the light microscopic appearance of gill cells exposed as

serving to assure DNA damage can be detected with the mam-malian protein-directed antibodies applied, we exposed the fishcells to �-irradiation. As depicted in Fig. 3, both �-irradiation as wellas incubation with Cd or Cu activated the DNA damage responseprotein Ataxia telangiectasia mutated (ATM) to its phosphorylatedform, phospho-ATM, as well as one of its downstream targets, his-tone H2A, to the active, phosphorylated form, �-H2AX (Figs. 4Aand S2). This was not only detected in individual cells, but alsoevident as a right-shift of �-H2AX staining intensity at the popu-lation level determined by FACS analysis (Fig. 4B). These changeswere already evident at 8 h of incubation (Fig. 4A) and persistedthrough at least 24 h of metal exposure (Figs. 4B and S2). Another

molecule downstream of ATM, Checkpoint kinase 2 (Chk-2), wasalso found to be activated by both irradiation and metal expo-sure (Fig. 4C). Together, these observations support that a DNAdamage response is readily elicited upon incubation with Cd orCu.

78 G. Krumschnabel et al. / Aquatic Toxicology 99 (2010) 73–85

Fig. 3. Mitochondrial membrane potential (MMP) of RTgill-W1 cells exposed to Cd. Gill cells were exposed to Cd and changes of their MMP were estimated by use of thepotential-sensitive fluorescence dye JC-1. Panel (A) depicts examples of FACS plots of cells exposed to Cd at the indicated concentrations for 24 h, with the left images showingd rial wd shownu or thei e sum

3

rcacclcap2ilaS1

ot plots of highly energized mitochondria (J-aggregates; FL-2) versus mitochondistribution of red fluorescence. In (B) examples of JC-1 loaded individual cells arepper panels and green monomers in the lower panels. Panel (C) shows examples f

ncubated with Cd for 72 h, and in (D) means ± SE of 3 independent experiments ar

.5. Caspase-3 activity

One of the hallmarks of classical apoptosis, often triggered inesponse to failed DNA damage response, is activation of the exe-utioner caspase-3. Indeed, in RTgill-W1 cells 50 �M Cd elicited anctivation of the protease from day 3 onwards, whereas 200 �M Cdaused only a slight transient elevation at day 1, but no enhancedaspase-3 activity during the subsequent days (Fig. 5A). A completeack of caspase activation was noted with 200 �M Cu, while, foromparison, the typical apoptosis inducing agent STS stimulatedpronounced increase within as little as 12 h. A slightly differenticture was seen in the hepatoma cells, in which both 50 �M and00 �M Cd induced a transient elevation of caspase-3 activity dur-

ng the first 2 days followed by a return to and even below basalevels (Fig. 5B). Furthermore, in these cells Cu caused a pronounced,lbeit transient increase of caspase-3 activity at day 3, whereasTS, again, evoked a rapid elevation of similar magnitude within2 h.

ith low potential (J monomers; FL-1) and the right images showing the frequencyafter exposure to control conditions or to Cd for 48 h, with red J-aggregates in the

decrease of MMP, estimated from the decrease of J-aggregate fluorescence, in cellsmarized.

3.6. The impact of necrostatin-1 (Nec-1) and caspase-inhibitionon cell death

Based on the findings described so far, it appeared that bothapoptosis and necrosis-like death occurred in the fish cells uponmetal exposure. In order to estimate the relative contribution ofeither cell death mode and to elucidate if necroptosis is involvedin this scenario, cells were exposed to the metals in the absenceand presence of the pan-caspase inhibitor zVAD-fmk or the RIP-1inhibitor Nec-1. Further, to reduce unwanted effects such as sec-ondary necrosis, which particularly in vitro may follow apoptosisin the absence of phagocyting cells (Vandenabeele et al., 2006),cell viability was already assessed after 72 h of incubation, a time

when caspase-3 has already been activated or activation was max-imal when present (see Fig. 5). As shown in Fig. 6A for the gillcell line and assessed with the MTT assay, both Nec-1 and zVAD-fmk conferred significant protection against 50 �M Cd, althoughthe effect was clearly less pronounced with the caspase inhibitor.

G. Krumschnabel et al. / Aquatic Toxicology 99 (2010) 73–85 79

F ma cea untero oresce

TwCfNfpa2dacaa

anlamoc

ig. 4. Induction of DNA damage response proteins in RTH-149 cells. Trout hepatond immunostained for �-H2AX and phospho-ATM (A) or phospho-Chk-2 (C) and cof a FACS analysis of cells exposed to Cd are depicted, showing the right-shift of flu

he same was observed when PI-positive cells were examined, inhich the increase to 34% PI-positive cells observed with 50 �Md was reduced to a mean of 12% and 17% with Nec-1 and zVAD-

mk, respectively (Fig. 6B). Partial protection was still seen withec-1 in the presence of 200 �M Cd (MTT assay), whereas zVAD-

mk no longer prevented cell death at this concentration. Betterreservation of cellular and nuclear morphology was even visiblefter 120 h of incubation with 50 �M Cd, but clearly no more with00 �M (Fig. S4). There was no synergistic effect of both inhibitorsetectable against Cd toxicity (Fig. S5) and when Nec-1 was appliedt higher concentrations it became toxic by itself (not shown). Inomparison, in cells treated with Cu no protection at all, but rathern aggravation of cell death was observed when inhibitors werepplied separately (Fig. 6A) or concurrently (not shown).

Basically similar results were obtained with another viabilityssay which is in fact based on cellular ATP contents (Fig. 6C). Aotable difference was that Nec-1 was not protective against the

oss of ATP at 200 �M and that Cu had apparently no impact on ATPt all when given alone. In a further experimental series we deter-ined MMP, as reflected by the retention capability of TRMR, and

btained yet again results with a similar tendency, although mosthanges were largely dampened and thus not significant (Fig. S5B).

lls were exposed to 30 grey �-irradiation, 50 �M Cd, or 200 �M Cu and then fixedstained with DAPI as nuclear marker. Scale bars indicate 20 �m. In panel (B) resultsnce intensity reflecting the increase of �-H2AX levels in the cell population.

In RTH-149 cells, again examined with the MTT assay, wemade qualitatively similar observations as in the gill cells inthat Nec-1 conferred apparent partial protection against Cd,but did not protect or even aggravated Cu-toxicity (Fig. S6A).Unfortunately, Nec-1 appeared to be toxic by itself to thesecells at a concentration when it was effective, reducing via-bility by 20% within 3 days. Its protective effect was thusonly significant at 200 �M Cd, but not at 50 �M. The caspaseinhibitor zVAD-fmk did not significantly protect against any treat-ment, although similar tendencies were seen as with the gillcells.

3.7. The impact of Nec-1 and caspase-inhibition on caspaseactivity and ROS formation

As shown in Fig. 7A, caspase-inhibition with zVAD-fmk workedperfectly well even over 72 h, since there was no increase of caspase

activity seen in its presence irrespective of the metal and its con-centration present. In contrast, Nec-1 had no significant impact oncaspase activity in any treatment, although it caused a slight reduc-tion of the activation seen in the presence of 50 �M Cd. In line withthis, we also observed a protective effect in terms of the number

80 G. Krumschnabel et al. / Aquatic Toxicology 99 (2010) 73–85

Fig. 5. Caspase-3 activity of trout cells exposed to Cd or Cu. RTgill-W1 (A) andRTH-149 (B) cells were exposed to Cd or Cu at the concentrations indicated andcaspase activity was estimated from the increase of fluorescence upon cleavageof zDEVD-R110 after 0–5 days of exposure. For comparison, caspase activity wasalso determined in cells incubated with the pan-kinase-inhibitor staurosporine (STS,0.5 �M) for 12 h. Data are means ± SE of 3 cultures. *p < 0.05 versus control at timez

oc(

w(2iccQsmz

aFtdas

Fig. 6. The impact of inhibition of necroptosis or caspase activity on cell viabilityof RTgill-W1 cells exposed to Cd or Cu. Gill cells were pre-incubated with 10 �Mnecrostatin-1 (Nec-1) or zVAD-fmk (zVAD) for 1 h and then incubated with Cd orCu, in the continued presence of inhibitors, for 72 h. At the end of the period cell

pathways

ero.

f sub-G1 cells, the small increase of which, although not statisti-ally significant, was totally blocked by both Nec-1 or by zVAD-fmkFig. S5C).

Besides inhibiting caspases, zVAD-fmk was shown to interactith components of the mitochondrial permeability transition pore

MPTP) and to thereby promote necrotic cell death (Temkin et al.,006). In order to exclude that this masked the effect of caspase-

nhibition, we also tested QVD, a caspase inhibitor with a differenthemical backbone (Caserta et al., 2003). However, viability of gillells treated with 50 �M Cd in the absence and presence of 20 �MVD amounted to 59 ± 5% and 41 ± 6% (n = 4), respectively. The

ame lack of protection was also indicated in time-lapse moviesade in the presence of the inhibitor (not shown), suggesting that

VAD-fmk is a more potent inhibitor of teleost caspases.The induction of radical production has been described to

ccompany exposure to Cd or Cu (Manzl et al., 2004; Risso-deaverney et al., 2004) and also repeatedly upon induction of necrop-osis (Goossens et al., 1999). Unexpectedly, however, we could notetect an increase of ROS formation with any of the treatments

pplied, neither after 6 h of exposure (Fig. 7B), nor after 24 h (nothown).

viability was estimated by use of the MTT assay (A), by determining PI-uptake (B),or with a viability assay based on cellular ATP contents (C). Data are means ± SE of

3–7 cultures. *p < 0.05 versus control, §p < 0.05 versus metal only.

3.8. Effects of other inhibitors of necroptotic and apoptotic

In addition to RIP, which can not only be directly blocked by useof Nec-1, but also indirectly by interfering with its molecular chap-

G. Krumschnabel et al. / Aquatic Toxicology 99 (2010) 73–85 81

Fig. 7. The impact of inhibition of necroptosis or caspase activity on caspase activityand ROS formation of RTgill-W1 cells exposed to Cd or Cu. Gill cells were pre-incubated with 10 �M Nec-1 or zVAD for 1 h and then further incubated with Cd orCu, in the continued presence of inhibitors, for 72 h (A) or 6 h (B). At the end of theeflp*

e2pasmamfi

qeecp(ncAF

Fig. 8. The impact of inhibitors of hypothetical necroptosis pathway proteins on via-bility of RTgill-W1 (A) or RTH-149 cells (B) exposed to Cd. Cells were pre-incubatedfor 1 h with 1 �M 17-DMAG (hsp-90 inhibitor), 0.5 mM 3-aminobenzamide (PARP

xposure time caspase activity (A) or the formation of ROS (B) were estimated with

uorescence-based assays. Cells treated with 1 mM H2O2 for 1 h were included as aositive control. Data are means ± SE of 3 and 8 (3 for H2O2) cultures, respectively.p < 0.05 versus control.

rone hsp-90, preventing its proteasomal degradation (Lewis et al.,000), other suggested molecular constituents of the necroptoticathway are the DNA repair enzyme PARP-1 and the mitogen-ctivated protein kinase JNK (Deng et al., 2003). Furthermore, iteems that the MPTP or at least a component of it, cyclophilin D,ay be involved in necroptosis (Baines et al., 2005; Nakagawa et

l., 2005). We thus tested the impact of specific inhibitors of theseolecules to assess if any of these findings can be verified for the

sh cells exposed to Cd.In fact, it appeared that inhibition of hsp-90 function had

ualitatively similar effects as Nec-1, but since it exerted consid-rable toxicity against the RTgill-W1 cells by itself, this protectiveffect was not significant (Fig. 8A). In comparison, in the RTH-149ells, for which it was apparently less toxic, it conferred com-lete protection against 50 �M Cd, although not against 200 �M

Fig. 8B). The competitive PARP inhibitor 3-aminobenzamide, aicotinamide analogue, had no effect on cell viability of gillells or hepatoma cells, and the same was true for cyclosporine, an inhibitor of the MPTP component cyclophilin D (Fig. 8).inally, we also tested inhibitors of the MAP kinase JNK and

inhibitor), or 10 �M cyclosporine A (PTP inhibitor). Then cells were exposed to Cd,in the continued presence of inhibitors, for 72 h and cell viability was assessed fromMTT conversion. Data are means ± SE of 11-12 (A) and 4-8 (B) cultures. *p < 0.05versus control.

of MAP kinases p38 and ERK. Although only JNK and not thelatter ones have been implied in necroptosis, they play multi-ple roles in apoptosis (Wada and Penninger, 2004). However,all these treatments, and in particular inhibition of the gener-ally anti-apoptotic ERK were toxic even in the absence of Cd andthus none of these treatments was beneficial for cell survival(Fig. S6B).

3.9. Ultrastructural changes induced by Cd and Cu

In a final series of experiments we examined morphologicalchanges of RTH-149 cells using a rapid cryo-fixation that avoidsfixation artifacts. As depicted in Fig. 9A, controls had homogeneouschromatin in a round-shaped nucleus, elongated mitochondria

with clearly discernible cristae, a modest number of lysosomesand a well-defined Golgi complex with multiple flattened cisternalmembranes. In cells exposed to 200 �M Cu for 48 h, there were nomajor apparent changes visible, although the number of electron-lucent vesicles seemed to have increased in some cells (Fig. 9B). In

82 G. Krumschnabel et al. / Aquatic Toxicology 99 (2010) 73–85

F lls wef croscoG an en

cealGcmtcfa

ig. 9. Ultrastructural changes of RTH-149 cells exposed to Cd or Cu. Hepatoma ceor 48 h, fixed by high-pressure freezing and analyzed by transmission electron miM-130 and counterstained with the nuclear marker DAPI. The lowermost image is

ontrast, while the nuclear morphology was still intact upon 48 hxposure to Cd, the Golgi apparatus appeared profoundly dilatednd seemed to be en route to total disintegration. Given that theatter finding was so prominent, we also tested changes of theolgi structure by immunofluorescence and found the EM findingsonfirmed. Thus, whereas the FITC-labeled antibody for the Golgi

arker GM-130 was nicely localized in the perinuclear region in

he controls (Fig. 9D) and in Cu-treated cells (Fig. S7), it stained aompartment that was largely fragmented and partly more distantrom the nucleus in cells exposed to 50 �M and 200 �M Cd (Figs. 9Dnd S7).

re incubated under control conditions (A) or with 200 �M Cu (B) or 50 �M Cd (C)py. In (D) controls and Cd-treated cells were immunostained for the Golgi markerlargement of the cell marked with a white rectangle in the middle panel.

4. Discussion

4.1. Cd induces multiple modes of cell death in trout cell lines

Rainbow trout cell lines exposed to moderate levels of Cd under-went cell death that at first sight seemed to primarily show signs

of classical apoptosis. That is, upon addition of Cd cells very slowly(after >48 h) lost cell–cell contacts and retracted from the substra-tum and at the end of 96–120 h of incubation many cells displayedheavily condensed and/or fragmented nuclei as well as rounded-up cell bodies. Furthermore, we saw an increase in the sub-G1

atic T

ptoiCctIdwlnbnfiwaeseztbtc(o2qaopCeafltm2cTaaDpottmbstaia

4c

iPae

G. Krumschnabel et al. / Aqu

opulation typically reflecting apoptotic cells and in most condi-ions at least a slight transient elevation of caspase-3 activity wasbserved. At the same time, however, the decrease of cell viabil-ty detected by the MTT assay (specifically 50% viability at 50 �Md) was quantitatively not at all accounted for by the number ofells with an apoptotic nucleus (which was not quantified) or byhe percentage of sub-G1 cells (an increase of 5% at 50 �M Cd).n contrast, the increase of PI-positive cells largely agreed withiminished MTT conversion capacity. In addition, if caspase activityas elevated this was generally moderate, and when morpho-

ogical changes were followed over time it became apparent thatumerous cells displayed cell swelling and not typical apoptoticlebbing. Together, these findings indicate that most cells diedot by apoptosis but rather by some form of necrosis. In support

or this notion, we observed a relatively early decrease of MMPn many cells and a pronounced decline of cellular ATP contents,

hich is usually considered to reflect necrotic cell death (Eguchi etl., 1997). Most importantly, however, the decrease of cell viability,stimated with MTT and PI-uptake, and the loss of ATP were to aignificant extent inhibited by the RIP-1 inhibitor Nec-1 and, gen-rally only to a lesser degree, reduced by the pan-caspase inhibitorVAD-fmk. By the definition used in the present study, Cd-exposedrout cells thus died by a mix of necroptosis and apoptosis, i.e.y Nec-1 sensitive and by zVAD-fmk sensitive cell death, respec-ively. The occurrence of secondary necrosis seems unlikely in thisontext, as this was recently shown to be insensitive to Nec-1Berghe et al., in press). The mixed necroptotic/apoptotic naturef cell death was generally more obvious at 50 �M Cd, whereas at00 �M zVAD-fmk no longer significantly affected the parameter inuestion. In line, caspase activation was absent (gill cells) or moder-te (hepatoma cells) at this concentration and the only indicationf apoptotic death was the occurrence of apoptotic nuclear mor-hology and the increase in sub-G1 cells. Noteworthy, at 200 �Md the latter would in fact agree with the decrease of viability asstimated from the MTT assay, but if sub-G1 cells truly reflectedpoptotic death, then this must have been a caspase-independentorm of apoptosis. Furthermore, in this case PI would then haveabeled cells following secondary necrosis. In support for this, Cd-riggered caspase-independent apoptosis has been described for

any cell models, e.g. rat kidney proximal tubule cells (Lee et al.,006), rat hepatocytes (Pham et al., 2006), a human lymphoblastoidell line (Coutant et al., 2006) and mouse mesangial cells (Liu andempleton, 2008), in each case associate with and possibly medi-ted by the mitochondrial release of apoptosis inducing factor (AIF),n oxidoreductase that is involved in chromatin condensation andNA fragmentation (Cande et al., 2004). But even if this also hap-ened in the trout cells, there was still an additional componentf Nec-1 sensitive cell death observed at 200 �M Cd, indicatinghat while necroptosis persisted also at elevated Cd concentration,he caspase-dependent component of apoptosis vanished. Further-

ore, at both concentrations cell death was seen that could neithere blocked by Nec-1 nor by zVAD-fmk (or a combination of both),uggesting that caspase-independent apoptosis may also have con-ributed at the lower Cd concentration. Taken together, our datare in line with previous reports of multiple modes of cell deathnduced by Cd and extend these observations by adding necroptosiss one pathway besides others.

.2. A DNA damage response is triggered by Cd and may initiateell death

The mechanisms by which Cd may induce cell death are man-fold and include enhanced formation of radicals (Liu et al., 2009;athak and Khandelwal, 2006; Risso-de Faverney et al., 2001), dam-ge of mitochondria with ensuing opening of the MPTP (Belyaevat al., 2006) or with the release of apoptogenic factors independent

oxicology 99 (2010) 73–85 83

of this pore (Lee et al., 2005), disturbance of ion homeostasis, inparticular of Ca2+ (Yang et al., 2007), activation of proteases (Hsu etal., 2009; Lee et al., 2007), and the induction of DNA damage (Bertinand Averbeck, 2006; Viau et al., 2008). Many of these mechanismsare closely intertwined and it is often not generally clear whichprocess is upstream or downstream of others. In the case of thetrout cell lines, a primary role for radicals appears unlikely, as noenhanced ROS formation was detected within the first 24 h of expo-sure. It should be mentioned, however, that ROS formation inducedby Cd may also be a transient phenomenon (Thévenod, 2009) andthus we cannot completely rule out that such a transient eleva-tion of radicals was overlooked. In contrast, we clearly detectedthe activation of proteins involved in the DNA damage response,i.e. of �H2AX, phospo-ATM and phospho-Chk-2, within a time-frame when no other effects of the metal were detectable. In line,it has been reported that 30 �M Cd may induce DNA double-strandbreaks, leading to the appearance of �H2AX foci, within as little as1 h in a human endothelial cell line (Viau et al., 2008). A similarlyrapid induction of DNA damage, as judged from the appearanceof Comet tails, has recently been described for rat kidney proxi-mal tubule cells exposed to 50–100 �M Cd, and this then led toChk-1/2 mediated arrest in G2/M phase (Bork et al., 2010). Suchan arrest was apparently not induced in the trout cells, since time-lapse movies indicated that the cells continued to proliferate forabout 48 h before cell damage became evident.

As outlined in the introduction, extensive DNA damage mayin fact trigger what could be referred to as “intrinsic necroptosis”(Festjens et al., 2006; Xu et al., 2006; Zong et al., 2004). Specifi-cally, it has been suggested that overactivation of PARP could leadto depletion of cytosolic NAD levels and in consequence of cellularATP (Zong et al., 2004), setting in motion a still ill-defined necrop-totic cascade. The decrease of ATP observed in the fish cells wouldagree with such a scenario. However, besides directly damagingDNA, Cd has also been shown to cause genotoxic stress throughinhibition of DNA damage repair enzymes (Bertin and Averbeck,2006; Whiteside et al., 2010), including PARP activity (Hartwig etal., 2002). Thus, enhanced PARP activity may have little effect onATP levels in Cd-exposed fish cells, which would agree with theobservation that inhibition of the enzyme had no effect on cell via-bility. Upon prolonged exposure to Cd, the suppression of PARP andother repair enzymes will nonetheless become a main issue as DNAdamage will accumulate in the cells to an extent that is no longercompatible with proper cell function and ultimately survival.

4.3. Further peculiarities of Cd-induced cell death of trout cells

Irrespective of the role of PARP, it was clear that inhibition of RIP-1, either directly with Nec-1 or indirectly by blocking its molecularchaperone hsp-90, conferred partial protection against Cd-inducedcell death. This is in contrast to the report by Hsu et al. (2009), whereNec-1, but not geldanamycin, an analogue of the hsp-90 inhibitorused here, protected Chinese hamster ovary cells against Cd, a dis-crepancy which remains unexplained at present. In contrast to theunequivocal role of RIP-1, other molecular constituents of pro-grammed necrosis previously identified could not be extended toteleost cells here, as, e.g. the prolonged inhibition of JNK (as well asof ERK and p38) was toxic by itself, and inhibition of the MPTP com-ponent cyclophilin D with CsA was without effect on cell viability.The latter appeared unexpected considering that a decrease of MMPduring necrosis-like cell death is often the result of MPTP opening(Kim et al., 2003). On the other hand, the role of the mitochon-

dria in necroptosis is still largely unclear (Christofferson and Yuan,2010) and may even involve classical apoptotic molecules such asBmf (Hitomi et al., 2008), suggesting possible links between apop-tosis and necroptosis. In this context it seems noteworthy that inthe Cd-treated fish cells, differently from other reports (Han et al.,

8 atic T

2rncflnweah

dopcsa(Acoi

4

oec(wctttiwiCaddlla

4

teteesottfiscpap

4 G. Krumschnabel et al. / Aqu

009), inhibition of either necroptosis or of caspase activity did notesult in the elevation of the respective uninhibited pathway. Thus,either did the presence of Nec-1 enhance apoptotic features (i.e.aspase activity, sub-G1 cells, fragmented nuclei), nor did zVAD-mk increase the number of PI-positive cells or the drop in ATPevels. Curiously, it rather seemed that Nec-1, at least to a small butot significant extent, reduced the above named apoptotic featureshen present. Thus, despite of the lack of an additive protective

ffect of Nec-1 and zVAD-fmk under the conditions tested, thereppeared to be some crosstalk, the nature and importance of whichas yet to be addressed in future studies.

A final peculiarity of Cd-induced cell death in trout cellsetected here is the early appearance of Golgi disintegration,ccurring before the detection of any other sub-cellular mor-hological alterations. Previously, an involvement of the Golgiomplex was reported for receptor-mediated apoptosis or necro-is, where it played a role in the activation of ceramide-generatingcid-sphingomyelinase and the lysosomal protease cathepsin DSchneider-Brachert et al., 2004), and in cytosolic phospholipase2-mediated necrosis (Festjens et al., 2006). Given that the Golgiomplex may act as a sensor and transducer of stress signals in vari-us settings (Hicks and Machamer, 2005), a more detailed look intots potential role in necroptosis is clearly warranted.

.4. The apparent lack of Cu-toxicity

In previous studies on primary trout hepatocytes we hadbserved that the acute toxicity of Cu exceeded that of Cd (Manzlt al., 2003) and that Cu caused opening of the MPTP and inducedell death with a similar mixed phenotype as observed here for CdKrumschnabel et al., 2005; Nawaz et al., 2006). Surprisingly, thisas not at all recapitulated in the trout cell lines used here and

ompared to Cd much higher concentrations of Cu were requiredo induce cell death. Furthermore, cell death was characterized byhe absence of apoptotic markers (except from a transient eleva-ion of caspase activity in RTH-149 cells) and by a relatively minorncrease in PI-positive cells and, most conspicuously, incubation

ith Nec-1 or zVAD-fmk aggravated cell death instead of ameliorat-ng it. Together, this suggests that necroptosis is not induced in theu-exposed trout cell lines and that, differently from primary hep-tocytes (Nawaz et al., 2006), also the inhibition of caspase activityoes not confer any protection. The exact mode of Cu-induced celleath thus remains open here and it appears that much higher Cu

evels will be required to induce faster cell death which is moreikely to involve programmed or secondary necrosis (Festjens etl., 2006).

.5. Cell type specific responses

A previous study has shown higher resistance at low concen-rations and greater sensitivity at high concentrations towards annvironmentally relevant toxicant of RTL-W1 liver cells, anotherrout-derived liver cell line, as compared to RTgill-W1 cells (Shaoat al., 2008). Here, we observed that the gill cell line was gen-rally less sensitive to Cd and Cu than RTH-149 cells, dying at alower rate and/or at higher concentration of the metals. The usef cell lines does quite obviously not perfectly mirror the situa-ion in primary cells from the organs they were derived from andhese differences underscore that extrapolation of these in vitro

ndings to an ecologically relevant situation is most likely not fea-ible. However, as such cell lines guarantee unlimited provision ofells with minimal biological variability, their use for unravelingrinciple mechanisms sure makes them an attractive and usefullternative to the much more limited and labor-intensive use ofrimary cell cultures.

oxicology 99 (2010) 73–85

Acknowledgements

We thank Karin Gutleben for excellent technical assistance. GKand AV were supported by a grant from the Austrian Science Fund(FWF; Y212-B13 START). HLE was supported by grants from theAustrian Science Funds (FWF-P19486-B12) and the Tiroler Wis-senschaftsfonds (P-UNI-0404/100) to M.W. Hess.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.aquatox.2010.04.005.

References

Baines, C.P., Kaiser, R.A., Purcell, N.H., Blair, N.S., Osinska, H., Hambleton, M.A., Brun-skill, E.W., Sayen, M.R., Gottlieb, R.A., Dorn, G.W., et al., 2005. Loss of cyclophilinD reveals a critical role for mitochondrial permeability transition in cell death.Nature 434, 658–662.

Belyaeva, E.A., Dymkowska, D., Wieckowski, M.R., Wojtczak, L., 2006. Reactive oxy-gen species produced by the mitochondrial respiratory chain are involved inCd2+-induced injury of rat ascites hepatoma AS-30D cells. Biochim. Biophys.Acta 1757, 1568–1574.

Berghe, T.V., Vanlangenakker, N., Parthoens, E., Deckers, W., Devos, M., Festjens,N., Guerin, C.J., Brunk, U.T., Declercq, W., Vandenabeele, P., in press. Necropto-sis, necrosis and secondary necrosis converge on similar cellular disintegrationfeatures. Cell Death Differ.

Bertin, G., Averbeck, D., 2006. Cadmium: cellular effects, modifications ofbiomolecules, modulation of DNA repair and genotoxic consequences (a review).Biochimie 88, 1549–1559.

Bork, U., Lee, W.K., Kuchler, A., Dittmar, T., Thevenod, F., 2010. Cadmium-inducedDNA damage triggers G(2)/M arrest via chk1/2 and cdc2 in p53-deficient kidneyproximal tubule cells. Am. J. Physiol. Renal Physiol. 298, F255–F265.

Cande, C., Vahsen, N., Garrido, C., Kroemer, G., 2004. Apoptosis-inducing factor (AIF):caspase-independent after all. Cell Death Differ. 11, 591–595.

Caserta, T.M., Smith, A.N., Gultice, A.D., Reedy, M.A., Brown, T.L., 2003. Q-VD-OPh, abroad spectrum caspase inhibitor with potent antiapoptotic properties. Apop-tosis 8, 345–352.

Christofferson, D.E., Yuan, J., 2010. Necroptosis as an alternative form of programmedcell death. Curr. Opin. Cell Biol. 22, 263–268.

Coutant, A., Lebeau, J., Bidon-Wagner, N., Levalois, C., Lectard, B., Chevillard, S.,2006. Cadmium-induced apoptosis in lymphoblastoid cell line: involvement ofcaspase-dependent and -independent pathways. Biochimie 88, 1815–1822.

Degterev, A., Huang, Z., Boyce, M., Li, Y., Jagtap, P., Mizushima, N., Cuny, G.D., Mitchi-son, T.J., Moskowitz, M.A., Yuan, J., 2005. Chemical inhibitor of nonapoptotic celldeath with therapeutic potential for ischemic brain injury. Nat. Chem. Biol. 1,112–119.

Deng, Y., Ren, X., Yang, L., Lin, Y., Wu, X., 2003. A JNK-dependent pathway is requiredfor TNFalpha-induced apoptosis. Cell 115, 61–70.

Eguchi, Y., Shimizu, S., Tsujimoto, Y., 1997. Intracellular ATP levels determine celldeath fate by apoptosis or necrosis. Cancer Res. 57, 1835–1840.

Festjens, N., Vanden Berghe, T., Vandenabeele, P., 2006. Necrosis, a well-orchestratedform of cell demise: signalling cascades, important mediators and concomitantimmune response. Biochim. Biophys. Acta 1757, 1371–1387.

Goossens, V., De Vos, K., Vercammen, D., Steemans, M., Vancompernolle, K., Fiers, W.,Vandenabeele, P., Grooten, J., 1999. Redox regulation of TNF signaling. Biofactors10, 145–156.

Habeebu, S.S., Liu, J., Klaassen, C.D., 1998. Cadmium-induced apoptosis in mouseliver. Toxicol. Appl. Pharmacol. 149, 203–209.

Han, W., Xie, J., Li, L., Liu, Z., Hu, X., 2009. Necrostatin-1 reverts shikonin-inducednecroptosis to apoptosis. Apoptosis 14, 674–686.

Hartwig, A., Asmuss, M., Blessing, H., Hoffmann, S., Jahnke, G., Khandelwal, S.,Pelzer, A., Burkle, A., 2002. Interference by toxic metal ions with zinc-dependentproteins involved in maintaining genomic stability. Food Chem. Toxicol. 40,1179–1184.

Hess, M.W., Muller, M., Debbage, P.L., Vetterlein, M., Pavelka, M., 2000. Cryoprepa-ration provides new insight into the effects of brefeldin A on the structure of theHepG2 Golgi apparatus. J. Struct. Biol. 130, 63–72.

Hicks, S.W., Machamer, C.E., 2005. Golgi structure in stress sensing and apoptosis.Biochim. Biophys. Acta 1744, 406–414.

Hitomi, J., Christofferson, D.E., Ng, A., Yao, J., Degterev, A., Xavier, R.J., Yuan, J., 2008.Identification of a molecular signaling network that regulates a cellular necroticcell death pathway. Cell 135, 1311–1323.

Hsu, T.S., Yang, P.M., Tsai, J.S., Lin, L.Y., 2009. Attenuation of cadmium-inducednecrotic cell death by necrostatin-1: potential necrostatin-1 acting sites. Toxicol.

Appl. Pharmacol. 235, 153–162.

Joseph, P., 2009. Mechanisms of cadmium carcinogenesis. Toxicol. Appl. Pharmacol.238, 272–279.

Kim, J.S., He, L., Lemasters, J.J., 2003. Mitochondrial permeability transition: a com-mon pathway to necrosis and apoptosis. Biochem. Biophys. Res. Commun. 304,463–470.

atic T

K

K

L

L

L

L

L

L

L

L

M

M

N

N

N

P

P

necrotic cell death through calpain-triggered mitochondrial depolarization and

G. Krumschnabel et al. / Aqu

rumschnabel, G., Manzl, C., Berger, C., Hofer, B., 2005. Oxidative stress, mitochon-drial permeability transition, and cell death in Cu-exposed trout hepatocytes.Toxicol. Appl. Pharmacol. 209, 62–73.

rumschnabel, G., Podrabsky, J.E., 2009. Fish as model systems for the study ofvertebrate apoptosis. Apoptosis 14, 1–21.

ee, W.K., Abouhamed, M., Thevenod, F., 2006. Caspase-dependent and -independent pathways for cadmium-induced apoptosis in cultured kidneyproximal tubule cells. Am. J. Physiol. Renal Physiol. 291, F823–F832.

ee, W.K., Bork, U., Gholamrezaei, F., Thevenod, F., 2005. Cd(2+)-induced cytochromec release in apoptotic proximal tubule cells: role of mitochondrial permeabilitytransition pore and Ca(2+) uniporter. Am. J. Physiol. Renal Physiol. 288, F27–39.

ee, W.K., Torchalski, B., Thevenod, F., 2007. Cadmium-induced ceramide formationtriggers calpain-dependent apoptosis in cultured kidney proximal tubule cells.Am. J. Physiol. Cell Physiol. 293, C839–C847.

ewis, J., Devin, A., Miller, A., Lin, Y., Rodriguez, Y., Neckers, L., Liu, Z.G., 2000.Disruption of hsp90 function results in degradation of the death domainkinase, receptor-interacting protein (RIP), and blockage of tumor necrosis factor-induced nuclear factor-kappaB activation. J. Biol. Chem. 275, 10519–10526.

iu, J., Qu, W., Kadiiska, M.B., 2009. Role of oxidative stress in cadmium toxicity andcarcinogenesis. Toxicol. Appl. Pharmacol. 238, 209–214.

iu, Y., Templeton, D.M., 2008. Initiation of caspase-independent death in mousemesangial cells by Cd2+: involvement of p38 kinase and CaMK-II. J. Cell. Physiol.217, 307–318.

opez, E., Figueroa, S., Oset-Gasque, M.J., Gonzalez, M.P., 2003. Apoptosis and necro-sis: two distinct events induced by cadmium in cortical neurons in culture. Br.J. Pharmacol. 138, 901–911.

yons-Alcantara, M., Mooney, R., Lyng, F., Cottell, D., Mothersill, C., 1998. The effectsof cadmium exposure on the cytology and function of primary cultures fromrainbow trout. Cell Biochem. Funct. 16, 1–13.

anzl, C., Ebner, H., Kock, G., Dallinger, R., Krumschnabel, G., 2003. Copper, but notcadmium, is acutely toxic for trout hepatocytes: short-term effects on energeticsand ion homeostasis. Toxicol. Appl. Pharmacol. 191, 235–244.

anzl, C., Enrich, J., Ebner, H., Dallinger, R., Krumschnabel, G., 2004. Copper-inducedformation of reactive oxygen species causes cell death and disruption of calciumhomeostasis in trout hepatocytes. Toxicology 196, 57–64.

akagawa, T., Shimizu, S., Watanabe, T., Yamaguchi, O., Otsu, K., Yamagata, H., Ino-hara, H., Kubo, T., Tsujimoto, Y., 2005. Cyclophilin D-dependent mitochondrialpermeability transition regulates some necrotic but not apoptotic cell death.Nature 434, 652–658.

awaz, M., Manzl, C., Lacher, V., Krumschnabel, G., 2006. Copper-induced stimu-lation of extracellular signal-regulated kinase in trout hepatocytes: the role ofreactive oxygen species, Ca2+, and cell energetics and the impact of extracellu-lar signal-regulated kinase signaling on apoptosis and necrosis. Toxicol. Sci. 92,464–475.

icoletti, I., Migliorati, G., Pagliacci, M.C., Grignani, F., Riccardi, C., 1991. A rapidand simple method for measuring thymocyte apoptosis by propidium iodide

staining and flow cytometry. J. Immunol. Methods 139, 271–279.

athak, N., Khandelwal, S., 2006. Oxidative stress and apoptotic changes in murinesplenocytes exposed to cadmium. Toxicology 220, 26–36.

ham, T.N., Marion, M., Denizeau, F., Jumarie, C., 2006. Cadmium-induced apoptosisin rat hepatocytes does not necessarily involve caspase-dependent pathways.Toxicol. In Vitro 20, 1331–1342.

oxicology 99 (2010) 73–85 85

Pulido, M.D., Parrish, A.R., 2003. Metal-induced apoptosis: mechanisms. Mutat. Res.533, 227–241.

Risso-de Faverney, C., Devaux, A., Lafaurie, M., Girard, J.P., Bailly, B., Rahmani, R.,2001. Cadmium induces apoptosis and genotoxicity in rainbow trout hepa-tocytes through generation of reactive oxygene species. Aquat. Toxicol. 53,65–76.

Risso-de Faverney, C., Orsini, N., de Sousa, G., Rahmani, R., 2004. Cadmium-inducedapoptosis through the mitochondrial pathway in rainbow trout hepatocytes:involvement of oxidative stress. Aquat. Toxicol. 69, 247–258.

Sancho, P., Fernandez, C., Yuste, V.J., Amran, D., Ramos, A.M., de Blas, E., Susin, S.A.,Aller, P., 2006. Regulation of apoptosis/necrosis execution in cadmium-treatedhuman promonocytic cells under different forms of oxidative stress. Apoptosis11, 673–686.

Shaoa, J., Eckerta, M.L., Leeb, L.E.J., Gallagher, E.P., 2008. Comparative oxygen radicalformation and toxicity of BDE 47 in rainbow trout cell lines. Mar. Environ. Res.66, 7–8.

Schneider-Brachert, W., Tchikov, V., Neumeyer, J., Jakob, M., Winoto-Morbach, S.,Held-Feindt, J., Heinrich, M., Merkel, O., Ehrenschwender, M., Adam, D., etal., 2004. Compartmentalization of TNF receptor 1 signaling: internalized TNFreceptosomes as death signaling vesicles. Immunity 21, 415–428.

Shih, C.M., Wu, J.S., Ko, W.C., Wang, L.F., Wei, Y.H., Liang, H.F., Chen, Y.C., Chen,C.T., 2003. Mitochondria-mediated caspase-independent apoptosis induced bycadmium in normal human lung cells. J. Cell. Biochem. 89, 335–347.

Temkin, V., Huang, Q., Liu, H., Osada, H., Pope, R.M., 2006. Inhibition of ADP/ATPexchange in receptor-interacting protein-mediated necrosis. Mol. Cell. Biol. 26,2215–2225.

Thévenod, F., 2009. Cadmium and cellular signaling cascades: to be or not to be?Toxicol. Appl. Pharmacol. 238, 221–239.

Vandenabeele, P., Vanden Berghe, T., Festjens, N., 2006. Caspase inhibitors promotealternative cell death pathways. Sci. STKE 2006, pe44.

Viau, M., Gastaldo, J., Bencokova, Z., Joubert, A., Foray, N., 2008. Cadmium inhibitsnon-homologous end-joining and over-activates the MRE11-dependent repairpathway. Mutat. Res. 654, 13–21.

Wada, T., Penninger, J.M., 2004. Mitogen-activated protein kinases in apoptosis reg-ulation. Oncogene 23, 2838–2849.

Whiteside, J.R., Box, C.L., McMillan, T.J., Allinson, S.L., 2010. Cadmium and copperinhibit both DNA repair activities of polynucleotide kinase. DNA Repair (Amst.)9, 83–89.

Xiang, L.X., Shao, J.Z., 2003. Role of intracellular Ca2+, reactive oxygen species,mitochondria transmembrane potential, and antioxidant enzymes in heavymetal-induced apoptosis in fish cells. Bull. Environ. Contam. Toxicol. 71,114–122.

Xu, Y., Kim, S.O., Li, Y., Han, J., 2006. Autophagy contributes to caspase-independentmacrophage cell death. J. Biol. Chem. 281, 19179–19187.

Yang, P.M., Chen, H.C., Tsai, J.S., Lin, L.Y., 2007. Cadmium induces Ca2+-dependent

reactive oxygen species-mediated inhibition of nuclear factor-kappaB activity.Chem. Res. Toxicol. 20, 406–415.

Zong, W.X., Ditsworth, D., Bauer, D.E., Wang, Z.Q., Thompson, C.B., 2004. AlkylatingDNA damage stimulates a regulated form of necrotic cell death. Genes Dev. 18,1272–1282.