lable at ScienceDirect

Environmental Pollution 158 (2010) 3338e3346

Contents lists avai

Environmental Pollution

journal homepage: www.elsevier .com/locate/envpol

Alterations to proteome and tissue recovery responses in fish liver caused bya short-term combination treatment with cadmium and benzo[a]pyrene

P.M. Costa a,*, E. Chicano-Gálvez b, J. López Barea b, T.À. DelValls c, M.H. Costa a

a IMAR-Instituto do Mar, Departamento de Ciências e Engenharia do Ambiente, Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa,2829-516 Monte de Caparica, PortugalbDepartamento de Bioquímica y Biología Molecular, Universidad de Córdoba, Campus de Rabanales, Edificio Severo Ochoa, 14071 Córdoba, SpaincUNESCO/UNITWIN/WiCop Chair-Departamento de Química Física, Facultad de Ciencias del Mar y Ambientales, Universidad de Cádiz, Polígono ríoSan Pedro s/n, 11510 Puerto Real, Cádiz, Spain

The interaction between cadmium and benzo[a]pyrene impairs specifi

c responses to toxicity and tissue repair mechanisms.

a r t i c l e i n f o

Article history:Received 2 February 2010Received in revised form18 July 2010Accepted 21 July 2010

Keywords:MetalPolycyclic aromatic hydrocarbonProteomicsApoptosisHepatic parenchyma

Abbreviations: 1-cysPrx, 1-cys peroxiredoxin; 2-Dresis; ApoA-IV3, apolipoprotein A-IV3; B[a]P, benzo[a]cadmium; CDC48, cell division cycle 48; GPx, glutathH4; MS, mass spectrometry; MTI, metallothioneinhydrocarbon; PCD, programmed cell death; PEPB,binding protein; TIMP2, tissue metalloproteinase inh* Corresponding author.

E-mail address: pmcosta@fct.unl.pt (P.M. Costa).

0269-7491/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.envpol.2010.07.030

a b s t r a c t

The livers of soles (Solea senegalensis) injected with subacute doses of cadmium (Cd), benzo[a]pyrene(B[a]P), or their combination, were screened for alterations to cytosolic protein expression patterns,complemented by cytological and histological analyses. Cadmium and B[a]P, but not combined, inducedhepatocyte apoptosis and Kupfer cell hyperplasia. Proteomics, however, suggested that apoptosis wastriggered through distinct pathways. Cadmium and B[a]P caused upregulation of different anti-oxidativeenzymes (peroxiredoxin and glutathione peroxidase, respectively) although co-exposure impairedinduction. Similarly, apoptosis was inhibited by co-exposure, to which may have contributed a syner-gistic upregulation of tissue metalloproteinase inhibitor, b-actin and a lipid transport protein. Theregulation factors of nine out of eleven identified proteins of different types revealed antagonistic orsynergistic effects between Cd and B[a]P at the prospected doses after 24 h of exposure. The resultsindicate that co-exposure to Cd and B[a]P may enhance toxicity by impairing specific responses and notthrough cumulative damage.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

The mechanisms underlying cellular detoxification and elimi-nation of xenobiotics are complex and are known to depend onmultiple factors such as contaminant class, doses, biologicalspecies, affected tissue and cell types, and co-exposure to othercontaminants. The complexity of these mechanisms in whole-tissue and organs has been an important constraint to many fieldsof xenobiotic research, such as environmental toxicology andhuman occupational health, especially when multiple contami-nants are involved. Most toxicological studies that focused oncontaminant interactions aimed at the effects on common

E, 2-dimensional electropho-pyrene; CatL, cathepsin L; Cd,ione peroxidase; H4, histoneI; PAH, polycylic aromaticphosphatidylethanolamine-

ibitor 2.

All rights reserved.

biomarkers, or potential biomarkers, such as metallothionein (MT),and cytochrome P4501A (CYP1A) induction, activity of antioxidantenzymes, etc. (e.g. Sandvik et al., 1997; Hurk et al., 1998; Orbeaet al., 2002; Marigómez et al., 2005; Costa et al., 2009a; Roesijadiet al., 2009). Such studies revealed the existence of antagonistic(opposite effect) and synergistic (a combination that produces anoutcome that is higher than the sum of each isolated xenobiotic)traits when an organism, organ or cell culture is exposed tomultiple contaminants.

Cadmium (Cd) and benzo[a]pyrene (B[a]P) are commoncontaminants in industrialized areas, especially affecting coastalwater bodies, where organic enriched sediments or suspendedparticles function as a trap for metals such as Cd and hydrophobicorganic xenobiotics like polycyclic aromatic hydrocarbons (PAHs),amongwhich B[a]P is included. Airborne particle-bound Cd and B[a]P(e.g. fumes and ashes) are also of great concern to environmental and,especially, human health (Harris et al., 1985; Viaene et al., 2000). Forsuch reasons, many studies have been carried out to evaluate Cd andB[a]P toxicity in aquatic and terrestrial environments (see Friesenet al., 2008 and Nordberg, 2009 for a review). Both xenobioticshave also been widely employed as model toxicants in in vivo and in

Table 1Summary of the experimental procedure.

Experimental treatment Injected Cd (mg) Injected B[a]P (nmol) N

Control e e 6 (2 � n ¼ 3)Cd 5 e 6 (2 � n ¼ 3)Cd þ B[a]P 5 10 6 (2 � n ¼ 3)B[a]P e 10 6 (2 � n ¼ 3)

N e total number of individuals per experimental treatment.n e number of individuals per replicate.

P.M. Costa et al. / Environmental Pollution 158 (2010) 3338e3346 3339

vitro studies. Nevertheless, interaction studies are scarce and evenscarcer with respect to whole-tissue/organ in vivo effects.

Toxicity of PAHs, including B[a]P, is mostly linked to theproduction of activated PAH forms such as PAH quinones and thehighly genotoxic diol epoxides by the CYP1A monooxygenasecomplex with subsequent release of reactive oxygen species, ROS(Conney, 1982; Flowers-Geary et al., 1996). Consequently, themajority of metal þ PAH interactions have focused on oxidativestress and mutagenicity. Cadmium and other metals have beenfound to reduce the activity and induction of CYP1A in humanHepG2 cells exposed to B[a]P even in presence of non-cytotoxicconcentrations of the metals (Vakharia et al., 2001). Cadmium isknown to be toxic at low levels of exposure but the exact mecha-nisms of toxicity are not yet fully understood. It is thought that Cdmay displace iron and zinc from metallothioneins and otherproteins such as copperezinc superoxide dismutase (CuZn SOD)and from zinc-finger class proteins (Bauer et al., 1980; Asmuß et al.,2000; López-Barea and Gómez-Ariza, 2006). Synergistic andantagonistic effects have been found on lethality induced byintraperitoneally injected Cd and B[a]P in fish, which was specu-lated to be caused by Cd-induced reduction of CYP1A activity andmetallothionein inactivation by B[a]P (Hurk et al., 1998). AlthoughCd is regarded as a genotoxicant on its own, it has been found toenhance DNA damage induced by a B[a]P epoxides metabolite byimpairment of DNA repair in human HeLa cell extracts in a dose-dependent pattern (Mukherjee et al., 2004). Conversely, no inter-action effects were found regarding the occurrence of clastogenicnuclei of mouse bone marrow cells (Lewi�nska et al., 2007).However, the effects of co-exposure to metals and PAHs on cell andtissue structure and biochemistry (including detoxification andregeneration) still remain largely unknown.

In an attempt to study Cd and B[a]P interactions on a speciesrealistically subjected to potential metal and PAH exposures, wehave chosen the Senegalese sole Solea senegalensis Kaup, 1858(Teleostei: Soleidae) as a test species. This species is a commonestuarine flatfish in southeastern Europe that has been employed ina growing number of environmental monitoring and baselinetoxicological studies (e.g. Jiménez-Tenorio et al., 2007; Costa et al.,2008, 2009a, 2009b). The combined effects of metallic and organiccontaminants (especially PAHs) in biomarkers of exposure andeffect in S. senegalensis have already been reported to yield unex-pected results in fish exposed to estuarine sediments, revealing, atsome extent, interaction effects between the two classes ofcontaminants (Costa et al., 2008, 2009a, 2009b). For such reason,the present work intends to explore the mechanisms ofmetal � PAH interaction, employing Cd and B[a]P as modelxenobiotics. The present study aims at the effects and responses ofCd and B[a]P in whole-liver tissue combining proteomics, histologyand cytology as screening tools. It is intended to contribute to theunderstanding of the immediate mechanisms of Cd and B[a]Pco-exposure and their consequences in the hepatic parenchyma ofa species that is exposed in its natural habitat to mixed classes ofcontaminants, including metals and PAHs.

2. Materials and methods

2.1. Experimental procedure

Twenty-four randomly selected laboratory-hatched S. senegalensis juveniles(52.7 � 1.3 mm standard length, 2.2 � 0.4 g total wet mass) from the same cohortwere divided by four experimental treatments: control, Cd, Cd þ B[a]P and B[a]P.Tests were performed in duplicate, with three individuals per replicate. All testedfish were intraperitoneally injected with the same carrier solution containing 5 mLMilli-Q grade ultrapure water þ 10 mL z100% dimethyl sulfoxide (DMSO) þ 85 mLsterilized 50 mM PBS (pH 7.4, with 0.7% sodium chloride). The injected Cd dosagewas obtained by preparing the carrier solution with 5 mL of 1 mg mL�1 cadmiumchloride solution in Milli-Q water diluted from a standard CdCl2 Tritisol solution

(Merck, Darmstadt, Germany). The B[a]P dosage was achieved by preparing thecarrier with 10 mL of 10 nmol mL�1 B[a]P (from Sigma, St Louis, MO, USA) diluted inz100% DMSO. Table 1 summarizes the procedure. Control individuals were injectedwith the carrier solution prepared without any of the test xenobiotics. Real injecteddosages were 24.0 � 4.5 nmol Cd g�1

fish wet mass (2.7 � 0.5 mg Cd g�1) and5.1 � 1.0 nmol B[a]P g�1

fish wet mass (1.3 � 0.3 mg B[a]P g�1). Doses were deter-mined to ensure subacute administration of contaminants, according to pre-existingdata (Hurk et al., 1998). After injection, animals were incubated for 24 h in 15 Lpolyvinyl tanks with 9 L of filtered seawater, with continuous aeration, at constantconditions (temperature ¼ 18 � 1 �C, salinity ¼ 33 � 1, pH 8.0, total ammoniaz0 mg L�1). Animals were collected after the incubation period, euthanized bydecapitation and liver samples collected for subsequent analyses.

2.2. Proteomic analysis

The differential expression of cytosolic proteins was assessed by two-dimen-sional electrophoresis (2-DE), through a protocol adapted from previous research(Romero-Ruiz et al., 2006; Montes-Nieto et al., 2007). For each experimentalcondition z100 mg of liver samples were pooled and homogenized in liquidnitrogen prior to extraction in 20 mM TriseHCl buffer (pH 7.6) with 0.5 M sucroseand 0.15M KCl, complementedwith 20mMdithiothreitol (DTT) as a reducing agent;1 mM phenylmethylsulfonyl fluoride (PMSF), 6 mM leupeptin and 100 mL mL�1 of theProtease Inhibitor Cocktail (Sigma) as proteolytic activity inhibitors. After centri-fuging for 1 min (to remove the lipid supernatant) þ 10 min (4 �C at 14,000�g), thesamples were treated with 500 UmL�1 benzonase endonuclease for 30 min at roomtemperature, followed by centrifuging at 102,000�g for 1 h (at 4 �C) to precipitateremaining non-peptide material. The samples’ total protein was quantified accord-ing to Bradford (1976) in order to load each IPG strip (18 cm, pH 4e7; from GEHealthcare) with 100 mg of protein. From this point onwards, the procedure wasperformed in triplicate per each experimental condition (3 IPG strips þ 3 gels percondition). The IPG strips were incubated for 30 min at room temperature with pH4e7 IPG buffer (GE Healthcare) including 7 M urea, 2% w/v CHAPS (a non-ionicdetergent), 20 mM DTT and z1% w/v bromphenol blue. After passive rehydration(6 h, 20 �C), strips were subjected to isoelectric focusing (IEF) on a Protean IEFapparatus (Bio-Rad). Separation by protein subunit weight (MW) was done bysodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) in20 � 20 cm 12.5% acrylamide/bis-acrylamide gels after equilibration of IPG strips in1.5 M TriseHCl buffer (pH 8.8) with 6 M urea and 2% SDS (as denaturing agents);65 mM DTT, 3% glycerol and z1% w/v bromphenol blue, which was followed bya treatment with 25 mM iodacetamide (IAA) to block sulphydryl groups. Electro-phoresis was run at constant wattage (10 W per gel) in a DodecaCell Plus systemfrom Bio-Rad. The SigmaMarker wide-range protein ladder (SigmaeAldrich) wasused as MW standard. Gels were afterwards stained with Sypro Ruby fluorescentdye (Bio-Rad) and imaged with a Bio-Rad FXImager laser scanner. Gel analyses andprotein regulation factors were achieved using the software ProteomWeaver (Bio-Rad). Spots were selected for protein identification according to the followingcriteria: a significant difference in spot intensity (a¼ 0.05, ManneWhitney U test) ofat least one experimental condition comparatively to control gels; a minimum of150% or 50% up- or downregulation compared to control gels, respectively, consis-tent within the three gels obtained per experimental condition, to ensure goodcontrast between regulation factors (e.g. Montes-Nieto et al., 2007). Excised spotswere treated with DTT and IAA prior to digestionwith porcine trypsin. The productswere sequenced by matrix assisted laser desorption/ionization time-of-flight massspectrometry (MALDI-TOF) using a Voyager De-PRO apparatus (Applied Biosystems)to evaluate digestion quality and eliminate trypsin self-proteolysis noise, followedby de novo sequencing through capillary liquid chromatography electrospray ioni-zation ion trap tandem mass spectrometry (capLC-ESI-ITMS/MS) with the LTQsystem from ThermoFisher Scientific. Peptide search was performed using thesoftware Protein-Protein Blast 2.2.20 (Altschul et al., 1997) to contrast results tochordate, actinopterygian and pleuronectiform peptide sequences from the NCBIAll-Non Redundant (nr) Protein Sequence database and validated by the criteria ofhighest score, lowest e-value, number of matched peptides and representativity inthe prospected taxa sets.

P.M. Costa et al. / Environmental Pollution 158 (2010) 3338e33463340

2.3. Microscopy analyses

Liver samples were fixed in Bouin-Hollande’s solution for 36 h, dehydrated ina progressive series of ethanol, embedded in paraffin and sliced to 1e3 mm thicksections (Martoja and Martoja, 1967). Sections where either stained with haema-toxylin and eosin (H&E) for histological evaluation or with the acridine orange (AO)nucleic acid-binding fluorochrome for determination of apoptosis. Additional liversamples were fixed with a glutaraldehyde and paraformaldehyde mixture in 0.1 Mcacodilate buffer (pH 7.2) with 2.5% w/v NaCl and 6% sucrose and postfixed withosmium tetroxide before dehydration in a progressive series of ethanol andembedding in LR White resin. Sections (200e400 nm thick) where stained withtoluidine blue for structural detailing, sudan black B for lipids (Bronner, 1975) andcoomassie brilliant blue (CBB) R250 for peptides (Fisher, 1968). The proportion ofapoptotic and normal hepatocytes was determined on AO-stained slides (Okadaet al., 2002), observed with UV epifluorescence at � 1000 magnification. At least 1000 cells per individual were counted from eight to twelve sections per slide.Apoptotic cells were scored following morphologic criteria (Häcker, 2000).

All analyses were performed with a DMLB model microscope adapted for epi-fluorescence with an EL6000 light source for mercury short-arc reflector lamps. Theoptical path was equipped with an I3 filter to detect AO fluorescence. All equipmentwas supplied by Leica Microsystems. Image analyses and processing was done withthe software ImageJ (Wayne Rasband National Institutes of Health, USA).

2.4. Statistical analyses

Altered protein regulation factors relatively to controls were analyzed by clusteranalysis based on the 1-Pearson r statistic in order to determine the degree ofcorrelation between the expressions of identified proteins. The differences betweenthe percentages of apoptotic cells in hepatic tissue were assessed through the non-parametric ManneWhitney U test. The statistical significance level was set ata ¼ 0.05 for all analyses. Statistics were computed using the software Statistica(Statsoft, USA).

3. Results

3.1. Proteomic analysis

The analysis of changes in protein regulation patterns by 2-DEyielded twenty-four protein spots that met the assumptionsabovementioned and were thus selected for protein identification.Eleven of these twenty-four spots provided a positive identificationafter contrasting to the NCBI nrProtein database the MS/MS spectraobtained by de novo sequencing of several peptides from eachprotein that had been selected as suitable precursors in theprevious MALDI-TOF analysis. Peptides potentially resulting fromkeratin contamination were eliminated from subsequent analyses.Due to the small number of S. senegalensis sequenced peptidesdeposited in public access databases, only b-actin and trypsin/trypsinogen were directly matched to the species. The best-matched protein was cathepsin L (CatL) with a 90.1 score anda 4.0 � 10�19 e-value (Table 2).

Table 2Protein identification summary after de novo sequencing using ESI-ITMS/MS and peptideover control (� standard deviation) for each identified protein.

Protein ID Abbreviation UniProtAccession

Taxa database a S

1-cys Peroxiredoxin 1-cysPrx B5X838 1,2,3 3Apolipoprotein A-IV3 ApoA-IV3 Q5KSU2 2,3 2Beta-actin b-actin Q1HHC7 1,2,3 3Cathepsin L CatL P79722 1,2,3 9Cell division cycle 48 CDC48 A5JP17 1 3Glutathione peroxidase Gpx Q802G1 2,3 2Histone H4 H4 H4 2,3 3Metallothionein I MT1 MT1 1,2,3 3Phosphatidylethanolamine-binding

proteinPEBP B5DGG2 2,3 3

Tissue metalloproteinase inhibitor 2 TIMP2 B5XCZ1 1,2,3 4Trypsin Trypsin Q5XUG5 1,2,3 2

a nrNCBI database taxa from which peptides were matched: 1-Order Pleuronectiform

Two proteins, 1-cys peroxirredoxin (1-cysPrx) and CatL wereobserved to be significantly upregulated (overexpressed relativelyto control fish) as a consequence of Cd treatment. Cell division cycle48 (CDC48) was observed to be downregulated (underexpressed)as a consequence of the B[a]P treatment. Five proteins weredifferentially expressed in fish treated with Cd þ B[a]P. Theremaining proteins revealed expression patterns significantlydifferent from controls by two of the three different xenobiotictreatments, although in some cases with clear opposite responsessuch as in histone H4, upregulated in Cd and downregulated in B[a]P treatments. Fig. 1 summarizes the different regulation patternsobserved comparatively to controls.

3.2. Liver histology

Fish injected with Cd, B[a]P or both contaminants exhibited veryconsiderable alterations in hepatic parenchyma in comparison tocontrols (Fig. 2). H&E-stained hepatocytes from control fish wereobserved to be polyedric in shape and presented a translucent,virtually unstained, cytoplasm with few inclusions and a sphericalnucleus with a centric nucleolus. The parenchyma contained manysinusoids anastomosing from branches of the hepatic portal vein.Although the basic hepatic architecture did not appear to besignificantly altered in fish treated with Cd, B[a]P or Cd þ B[a]P, thelivers of individuals injected with the isolated xenobiotics pre-sented a higher density of cells caused especially by a profusion ofhepatic-specific macrophages (Kupfer cells). This increment in celldensity caused hepatocytes to compress, which, combined with anincrease in intracellular inclusions (including lipid vacuoles),provided an overall more intense H&E staining of the tissue, makingit difficult to detect cell boundaries and identify cell types (Fig. 2A).Oppositely, the hepatic parenchyma of individuals treated withCd þ B[a]P presented a less significant intrusion of defence cells.Hepatocytes of these fish, however, contained many densely-stained intracellular structures although empty-like lipid vacuoleswhere rare or absent. Necrosis was confined to small, sparse, foci inlivers of fish treated with Cd, B[a]P and Cd þ B[a]P but rarelyobserved in controls. It is not clear if Cd þ B[a]P livers presentedmore necrotic foci or if these were more conspicuous due to lessdense cell packing. Late stage inflammatory responses, such asblood vessel profusion and dilation and retention of erythrocyteswithin the parenchyma, were not observed in any treatment.Apoptotic hepatocytes, especially at a later stage of PCD (pro-grammed cell death), were easily distinguishable in AO-stainedsections (Fig. 2B). Apoptotic cells were observed in the livers of

sequence database search with ProteineProtein Blast plus relative regulation factors

Regulation factors over control

core e-value N� matchedpeptides

Cd Cd þ B[a]P B[a]P

6.7 4.0 � 10�5 2 0.64 � 0.44 0.30 � 0.44 �0.21 � 0.169.1 8.7 � 10�1 6 �0.20 � 0.27 0.54 � 0.23 �0.25 � 0.187.1 3.0 � 10�5 5 �0.51 � 0.10 0.91 � 0.50 0.46 � 0.280.1 4.0 � 10�19 4 0.61 � 0.07 0.40 � 0.06 0.13 � 0.070.3 4.0 � 10�3 2 �0.39 � 0.10 �0.44 � 0.11 �0.52 � 0.048.6 1.2 � 100 2 0.30 � 0.17 0.76 � 0.10 1.42 � 0.325.0 1.5 � 10�2 2 1.93 � 0.98 0.08 � 0.32 �0.50 � 0.058.8 8.0 � 10�3 3 �0.38 � 0.14 �0.51 � 0.04 �0.50 � 0.077.5 2.0 � 10�2 3 �0.48 � 0.03 �0.54 � 0.06 0.35 � 0.14

6.9 4.0 � 10�6 1 0.13 � 0.21 0.52 � 0.48 0.40 � 0.544.8 1.7 � 10�1 5 0.46 � 0.06 0.68 � 0.17 0.12 � 0.23

es; 2-Class Actinopterygii; 3-Phylum Chordata.

Fig. 1. Average protein regulation factors (expressed by an arbitrary unit relatively to control) and respective 2-DE gel photographs for each spot (arrows). [*] significant differencesof spot intensities from controls gels (ManneWhitney U, p < 0.05) and regulation factor up or down 150% and 50%, respectively, relatively to control gels. [y] significant differencesof spot intensities from control gels (ManneWhitney U, p < 0.05) but regulation factors narrower than the imposed thresholds of control �50% protein expression.

P.M. Costa et al. / Environmental Pollution 158 (2010) 3338e3346 3341

individuals subjected to all treatments but with a greater incidencein fish injected the isolated toxicants. Sudan staining of semi-thinsections confirmed the presence of large lipid vacuoles (lipidosis) inhepatocytes of Cd- and B[a]P-treated fish (Fig. 2C). Lipid dropletsin controls were limited to sparse microvesicles and virtuallyabsent from livers treated with Cd þ B[a]P.

3.3. Cytological observations and hepatic cell apoptosis

Kupfer cell intrusions are identifiable by stronger intraplasmaticretention of haematoxylin and AO pigments than hepatocytes dueto a high concentration of phagosomes and lysosomes (Fig. 3A, B).Besides its affinity to nucleic acids, acridine orange is a weak base

known to accumulate in acidic intracellular compartments such asproteolitic lysosomes (Völkl et al., 1993). These macrophages, nor-mally associated to sinusoids, were observed to proliferate(hyperplasia) and intrude into hepatic parenchyma from adjacentblood vessels in Cd- and B[a]P-treated individuals. Kupfer cellhyperplasia was not observed in controls and fish treated withCd þ B[a]P. Apoptotic cells (Fig. 3CeE) were mostly observed, aloneor in clusters, near blood vessels, usually compressed betweennormal hepatocytes and Kupfer cells. Cytological analyses identi-fied intracellular inclusions in macrophages and hepatocytesprobably as lysosomes and phagosomes (deeply stained by thecoomassie blue stain for peptides) containing the remains ofapoptotic, necrotic or authophagic cells (Fig. 3F). Oval (stem) cells

Fig. 2. Overview of hepatic parenchyma of tested individuals. hc) hepatic cord exhibiting the typical rosette-like structure of hepatocytes lining a transversally sectioned sinusoid;hpv) hepatic portal vein branch; lv) lipidic vacuoles; s) sinusoid. Scale bar: 25 mm. (A) Hepatic tissue stained with H&E. Cells of Cd- and B[a]P-injected individuals appear moredensely packed than those of controls and CdþB[a]P due to vacuolation and infiltration of defence cells in parenchyma. Hepatocytes of fish injected with both xenobiotics exhibitmany intraplasmatic inclusions, likely to be phagosomes and lysosomes. Localized foci of necrotic cells were also found no be more frequent in fish subjected to this experimentalcondition. (B) Apoptotic cells (arrows) were found to be distributed without any evident pattern, isolated or in small clusters, typically compressed against blood vessels. Althoughpresent in fish subjected to any of the treatments, apoptotic cells were more frequent in individuals injected with Cd or B[a]P alone. AO stain, viewed with epifluorescent UV light.(C) Sudan black B stain of semi-thin sections. Lipid droplets (black inclusions) could be found in hepatocytes of control and Cd- or B[a]P-injected individuals but were virtuallyabsent in liver cells of fish injected with both substances. However, while in hepatocytes from controls lipid bodies were present as microvesicles, large vacuoles were formed in fishinjected with Cd or B[a]P, indicating severe hepatic lipidosis.

P.M. Costa et al. / Environmental Pollution 158 (2010) 3338e33463342

could often be found intruding the hepatic parenchyma from theportal vein branches, more significantly in Cd and B[a]P treatments.Necrotic and autophagic cells could be observed in livers from alltreatments but were found difficult to pinpoint individually and,occasionally, even to discriminate. For such reason they could notbe objectively quantified. The counting of hepatocytes undergoingapoptosis on AO-stained slides revealed that treatments with Cdand B[a]P significantly increased the percentage of apoptotichepatocytes relatively to control (ManneWhitney U, p < 0.05) butnot in the combined Cd þ B[a]P treatment (Fig. 4). However, nosignificant differences were found between Cd and B[a]P treat-ments (ManneWhitney U, p > 0.05).

4. Discussion

The present findings revealed an impairment in the induction ofhepatocyte apoptosis and tissue regeneration as a consequence ofa simultaneous in vivo short-term treatment with the low dosesof Cd and B[a]P, accompanied by changes in protein expressionpatterns. Apoptosis is a pathway for PCD regarded as a mechanism

to control the elimination of damaged cells avoiding disseminationof potentially harmful degradation products and excessive inflam-matory response (Fedeel and Orrenius, 2005; Häcker, 2000). Inaddition, the observation of Kupfer and oval (stem) cells in fishtreated with the isolated contaminants indicates that the liverrapidly triggered recovery mechanisms. Kupfer cells and adjacenthepatocytes are responsible for the removal of apoptotic bodies andother cellular debris while oval cells infiltrate from periportal areasto replace the regressing parenchyma. The combination of Cd and B[a]P at the given doses impaired the regeneration process begin-ning with the blocking of PCD. Oval cells, on the other hand, areknown to intrude into the hepatic tissue when the normal regen-erative process is insufficient to respond to mechanical or chemicalinsult (Oh et al., 2002). Previous research by Arai and co-workers(2004), for instance, showed that a wide-range of genes arerapidly overexpressed in face of injury and subsequent oval cellproliferation, from energy balance-related genes to growth factorsamong which were included, for instance, apolipoprotein- andhistone-related genes. However, the mechanisms of hepaticregeneration are yet poorly understood.

Fig. 3. Discrimination of hepatic cell types and anomalies. (A) Kupfer cell intrusion in the liver of a B[a]P-injected individual (arrows). H&E stain. Scale bar: 25 mm. (B) Same as previous,stained with AO and viewed under epifluorescent UV light. Kupfer cell hyperplasia (arrows) is clearly discriminated from adjacent tissue due to its strong fluorescence. Side panel:a Kupfer cell on a blood vessel. The cytoplasm of these cells exhibits strong fluorescence, likely due to a high concentration of acidic lysosomes. Batches of AO-stained heterochromatinare visible lining the nuclear envelope. Scale bar: 25 mm. (C) Apoptotic bodies compressed betweenhepatocytes (ab). An early-stage apoptotic nuclei (an) can be observed in an adjacentcell, showing chromatin condensation and compression against the nuclear envelope insteadof the typical concentric nucleolus. H&E stain. Scale bar: 5 mm. (D) An intermediate stage ofhepatocyte apoptosis viewedunder epifluorescentUV light after AO staining. Chromatinfibres hypercondensate and begin to compress against the inner facing of the nuclear envelope,producing a strong fluorescent signal. The plasmatic membrane begins to bleb (arrowheads) as the cell shrinks in volume. Scale bar: 5 mm. (E) Three-dimensional rendering of a clusterof apoptotic cells (AO-stain and UV epifluorescence, obtained from six planes with a depth of approximately 0.5 mm each). The formation of the highly-fluorescent apoptotic bodies isclear on the highlighted cell (arrow). Scale bar: 5 mm. (F) A semi-thin section of the liver of a Cd-injected individual stainedwith toluidine blue. Arrowshighlight twoapoptotic cells at anadvanced stage. The nearly-detaching apoptotic bodies present a heterogeneous colouration due to the varying contents of each apoptotic pouch. A Kupfer cell (kc) is seen infiltratingthe tissue from a nearby sinusoid (s), likely preparing to phagocytize the forming apoptotic bodies. Phagosomes containing indiscriminatematerial could often be found inside normalhepatocytes (ph). Side panel, upper right: twoKupfer cells penetrating the hepatic parenchyma exhibitingmany protein-rich phagosomes (CBB stain). Side panel, lower right: overviewof an apoptotic cell cluster similar to the one shown on the main figure AO stain under UV epifluorescence. Scale bar: 5 mm.

P.M. Costa et al. / Environmental Pollution 158 (2010) 3338e3346 3343

Although Cd and B[a]P treatments resulted in many similaralterations in hepatic tissue, namely apoptosis, lipid vacuolationand Kupfer and oval cell intrusions, the proteomic analysis indi-cates that different biochemical pathways of response occurred.Most metal and PAH interaction studies focused on the effects on

Fig. 4. Average percentage of apoptotic cells per experimental condition. Error barsrepresent 95% confidence intervals. Different letters mean significant differences(ManneWhitney U test, p < 0.05).

known biomarkers, leaving questions regarding the specificbiochemistry of contaminant co-exposure yet to be answered.Marigómez et al. (2005), for instance, on the course of a single-concentration waterborne exposure experiment (500 mg L�1 B[a]Pplus 80 mg L�1 Cd), found that Cd could inhibit lysosomal changestriggered by B[a]P in mussels. Sandvik et al. (1997) observed similarresults in CYP1A induction and a limitation of Cd-induced metal-lothionein upregulation by B[a]P in the liver of flounder intraperi-toneally injectedwith single-doses of themetal and the PAH similarto those employed in the same study (1 mg g�1 and 2.5 mg g�1,respectively). These results are in accordance with the findings onhepatic MT and CYP1A responses in S. senegalensis exposed sedi-ments contaminated by mixtures of organic and metallic contam-inants (Costa et al., 2009a). Interestingly, other authors found thatco-exposure to B[a]P and Cd potentiated MT expression in theintestine of fish fed with contaminated pellets (Roesijadi et al.,2009), which may indicate that these interaction effects maydepend on organ-specific xenobiotic uptake and eliminationprocesses. Bioaccumulation, on the other hand, has been argued tobe unaffected by co-exposure tometals and PAHs (Marigómez et al.,2005). The present findings confirm that the biochemistry ofmetalePAH interactions in a vertebrate liver are complex and mayinvolve a broad-range of protein responses with multiple conse-quences at the histological and cytological levels.

Correlation analysis betweenprotein expression factors revealedpossible links between expression patterns (Fig. 5). Whereas CDC48and MT1 changes in regulation patterns appeared uncorrelated to

Fig. 5. Joining tree plot of all identified proteins based on regulation factors singlelinkage distances using 1-Pearson r correlation statistic as the distance measure(unweighted pair-group average was used as amalgamation rule).

P.M. Costa et al. / Environmental Pollution 158 (2010) 3338e33463344

the remaining proteins’, the other proteins (excluding PEBP) formedtwo distinct clusters. The first cluster included H4, 1-cysPrx, CatLand Trypsin,whose overexpressionmayhave been especially linkedto Cd. The second cluster comprised TIMP2, b-actin, GPx andApoA-IV-3 whose overexpression can be linked to B[a]P or to theinteraction between B[a]P and Cd. CDC48 pertains to the AAAATPase family (ATPases associated with diverse cellular activities),related to energy-dependent unfolding, disassembly and degrada-tion of protein complexes, membrane fusion, microtubule severing,peroxisome biogenesis, signal transduction and the regulation ofgene expression. They form a molecular motor that couples ATPhydrolysis to changes in conformational states of a target substrate,either translocating or remodelling it. Recent studies showed thatcellular depletion of CDC48 affects continuance of cellular cycleby impairing DNA synthesis (Mouysset et al., 2008). Although DNAreplication and induction of cell division by mitogens counterbal-ance apoptosis, the more specific role of CDC48 in the processremains unclear. It is likely that downregulation of CDC48 reflectsa general failure of hepatic cell metabolism. Similarly, someevidence exists thatMTsmight prevent Cd-driven apoptosis (Kondoet al., 1997). Our findings, however, reveal an uncertain relationshipbetween MT isoform 1 and alterations to liver histology eventhough B[a]P, alone or combined with Cd (a strong MT inducer),contributed to a more pronounced MT downregulation.

One of the most important mechanisms of Cd toxicity relates tothis metal’s ability to replace zinc (Zn) in several proteins,rendering them inactive, ranging from MTs (López-Barea andGómez-Ariza, 2006) to the anti-oxidant enzyme superoxide dis-mutase (Bauer et al., 1980) and the nuclear DNA-binding repairenzymes and other zinc-finger class proteins (e.g. Asmuß et al.,2000). It as long been discussed the link between these mecha-nisms and Cd-induced apoptosis (see, e.g., Hamada et al., 1997 andRisso-de Faverney et al., 2001). Holme and co-workers (2007)found that apoptosis induced by B[a]P to be been linked to ROSand DNA-binding B[a]P metabolites but also that the mechanismsmay vary from cell type and contaminant concentration. Both1-cysPrx and GPx are known to protect cells from apoptosis byscavenging oxidative radicals, therefore reducing oxidative stress(Gouazé et al., 2002; Manevich et al., 2002; Pak et al., 2002). At theinjected doses Cd and B[a]P triggered different anti-oxidative stressresponses, generating an antagonistic response during co-expo-sure: whereas Cd promoted 1-cysPrx upregulation, B[a]P upholdedGPx upregulation without affecting 1-cysPrx. The combination ofboth xenobiotics, however, was observed to reduce the induction ofboth antioxidant enzymes, which likely exposed cells to increasedoxidative stress.

Cathepsin L is a protease specialized in the degradation ofcellular and intercellular structural proteins. The activity of thisprotease has been positively related to apoptosis (Fujimoto et al.,2002). Cathepsins, like caspases, are cystein proteases of thepapain family. They are thought to “leak” from lysosomes duringapoptosis, contributing to the regulation of caspase-independentapoptosis via the lysosomes-mitochondrial pathway (seeChwieralski et al., 2006 for a review). Cathepsin L has also beenfound a chief lysosomal protease in cytochrome c-independentapoptosis by activation of caspase-3 in human cell lines (Hishitaet al., 2001). It is noteworthy that Cd has already been foundcapable of triggering caspase-independent apoptosis in normalhuman lung cells, by translocation of apoptosis induction factor(AIF) from the mitochondria to the nucleus (Shih et al., 2003). Eventhough the 2-DE procedure failed to detect caspases or theirprecursors, our findings suggest that CatL may be involved in Cd-induced apoptosis and that Cd and B[a]P had, in this study, anantagonist effect on the regulation of this enzyme. Interestingly,apoMT has been found to be rapidly degraded by cathepsins(including CatL), although metals bound to MT reduce suscepti-bility of hydrolysis, presumably by blocking the cysteine residueswhere cathepsins act (Min et al., 1992). It may be suspected that MTlevels in Cd-treated livers may not reflect MT expression alone, butalso that of cysteine proteinases.

The most striking combination effect observed between Cdand B[a]P in terms of protein expression was related to theobserved upregulation of TIMP2, b-actin and ApoA-IV3. Thedifferences between the regulation patterns of these proteins inCd þ B[a]P-treated fish and individuals treated with the isolatedtoxicants may aid explaining the reduced apoptosis and lack oflipid vacuolation in the livers of fish subjected to Cd þ B[a]P.Although it has been sustained that b-actin, a key element ofcystoskeletal microfilaments, is necessary to PCD, some authorsfound that its transcription is not necessarily linked to apoptosisand that, in some cases, it may actually fall (Naora and Naora,1996). These observations might contribute to explain thereduction in apoptosis in fish treated with Cd þ B[a]P, where b-actin upregulation was observed. On the other hand, TIMPs areknown apoptosis inhibitors by decreasing metalloproteinaseactivity (Barasch et al., 1997; Murphy et al., 2002). The reducedhepatocyte apoptosis observed in individuals treated with Cd þ B[a]P may thus be partially explained by TIMP upregulation.Apolipoproteins such as ApoA-IV3 play a role in lipid transport asstructural components of lipoprotein particles, cofactors forenzymes and ligands for cell-surface receptors. Recently, Apo-IVhas been discovered to reduce oxidative stress-driven apoptosisby modulating the glutathione redox status (Spaulding et al.,2006), which may have contributed to the reduced induction ofapoptosis in fish co-exposed to Cd and B[a]P. Moreover, Apo-IVupregulation has been observed to promote lipid efflux from cells(Fournier et al., 2000), which may have been responsible for thedepletion in intracellular lipid storage observed in the Cd þ B[a]Pcombination treatment. Conversely, Cd- and B[a]P-treated liversdepicted a moderate downregulation of Apo-IV that may havepromoted lipid vacuolation. Trypsin was another protein upre-gulated by the combination treatment. Inhibitors of trypsin-likeserine proteases, for instance, have been found to induce caspase-linked apoptosis (Murn et al., 2004). Conversely, excessive loadsof trypsin and other serine proteases have been found to induceapoptotic-like cell death (Williams and Henkart, 1994). Althoughmuch information is still lacking on the exact role of serineproteases in PCD, some authors have suggested that they mayeither take part in a caspase-independent apoptotic pathway orenhance caspase-induced apoptosis (Stenson-Cox et al., 2003).The present work does not allow further enhancement on the role

P.M. Costa et al. / Environmental Pollution 158 (2010) 3338e3346 3345

of trypsin in programmed cell death, however, upregulation ofthis enzyme in the livers of Cd þ B[a]P-treated animals suggestsa synergistic effect between the two xenobiotics occurred at thetested doses.

The enforced combination treatment didnot reveal any synergisticeffects between the two contaminants for phosphatidylethnaol-amine-binding protein and histone H4. Phosphatidylethanolamine-binding proteins (PEBPs) display several functions, including lipidbinding, neuronal development, serine protease inhibition, and theregulation of several stress-involved cellular signalling processes suchas the MAPK (mitogen-activated protein kinase) and the NF-kB(nuclear factor kappa binding) pathways. PEPBs are known to have ananti-PCD role by protecting themembrane integrity during apoptosis,preventing the phospholipid “flip-flop” through which the phospho-lipidphosphatidylethanolamine isexternalized to theouter faceof theplasmatic membrane (Wang et al., 2004). Cadmium, isolated and,more significantly, combined with B[a]P, caused a dowregulation ofPEBPs that was not observed in the B[a]P treatment. Still, the regula-tion pattern observed for this protein does not clearly relate to theobserved yields of apoptosis attainedwith the applied doses of CdandB[a]P. It is likely that Cd-induced downregulation of this protein mayhave facilitated PCD. Conversely, an antagonistic effect between Cdand B[a]P was observed in H4 expression. Histones are fundamentalproteins during early-stage apoptosis to make chromatin hyper-condensation possible so that, unlike during cell division, hetero-chromatin looses its structure to permit the binding of endonucleases(Robertson et al., 2000). Up- or downregulation of histones is associ-ated to cell proliferation and cell cycle arrest, respectively. Histoneupregulation in the livers of Cd-treated fish could indicate a moreadvancedstageof tissue regeneration than inB[a]P-treatedfish,whichis not entirely confirmed histologically, possibly due to the short-termincubation and to the relatively low doses of the injected xenobiotics.

The present work has shown that in vivo incubation of teleostlivers with low, sublethal, doses of Cd and B[a]Pmay produce a fast-acting response mechanism by promoting apoptosis and infiltra-tion of defence and stem cells to support tissue regeneration.Although the similarity of these responses can be evaluated histoand cytologically, the biochemical machinery responsible for theseresponses may differ. Furthermore, without being responsible forlethality or even a considerable loss of hepatic integrity, thecombination of the two substances very significantly impaired theliver’s defences. Still, it must be noted that much research is yetneeded to disclose the complexity of contaminant interactions inthe vertebrate liver, regarding, e.g., transmembrane transportersinvolved in metabolite excretion like the ATP-binding cassette(ABC) transporters or the cell or the signalling pathways thatcontrol tissue PCD and tissue regeneration by stem cells andmacrophage proliferation. It would also be of importance to exploredifferential exposure routes and multiple doses, as well as differentincubation or exposure periods, which are likely to change themechanisms and magnitude of xenobiotic interactions. In spite ofits limitations caused especially by gaps in peptide databases,proteomics proved to be a relevant tool to identify some of thedifferentially expressed proteins that account for the dissimilarbiochemistry of Cd- and B[a]P-induced apoptosis, anti-oxidativestress and tissue recovery responses without the need ofa complete knowledge on the mechanisms of toxicity. Regardless ofthe usefulness of proteomics in the search for new potentialbiomarkers in toxicological studies (Monsinjon and Knigge, 2007),the present findings showed that “omics” screening techniques areof relevance when breaking way at understanding the fundamentalmechanisms of toxicity. From the present results, it may also beinferred that caution is mandatory when employing commonbiomarkers, such as MT induction, histopathology and detection ofapoptosis in pharmacological and toxicological studies when

different classes of contaminants are involved since interactioneffects may mask the responses to individual contaminants.

Acknowledgements

The present research was approved by the Portuguese Scienceand Technology Foundation (FCT) and POCTI (Programa Oper-acional Ciência, Tecnologia e Inovação, research project ref. POCTI/AMB 57281/104) and financed by FEDER (European Fund forRegional Development). P. M. Costa is supported by a FCT PhD grant(SFRH/BD/28465/2006). The authors thank for their support: PedroPousão (Estação Piloto de Piscicultura, INRB/IPIMAR-Cripsul); JoséAlhama Carmona, Carlos A. Fuentes Almagro and Ricardo Fernán-dez Cisnal (Universidad de Córdoba); Lia Ascensão, Luísa Mota andFrancisco Carrapiço (Faculdade de Ciências da Universidade deLisboa).

References

Altschul, S.F., Madden, T.L., Schäffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J.,1997. Gapped BLAST and PSI-BLAST: a new generation of protein databasesearch programs. Nucleic Acids Research 25, 3389e3402.

Arai, M., Yokosuka, O., Fukai, K., Imazeki, F., Chiba, T., Sumi, H., Kato, M.,Takiguchi, M., Saisho, H., Muramatsu, M., Seki, N., 2004. Gene expressionprofiles in liver regeneration with oval cell induction. Biochemical andBiophysical Research Communications 317, 370e376.

Asmuß, M., Mullenders, L.H.F., Hartwig, A., 2000. Interference by toxic metalcompounds with isolated zinc finger DNA repair proteins. Toxicology Letters112e113, 227e231.

Barasch, J., Yang, J., Qiao, J.Z., Tempst, P., Erdjument-Bromage, H., Leung, W.,Oliver, J.A., 1997. Tissue inhibitor of metalloproteinase-2 stimulates mesen-chymal growth and regulates epithelial branching during morphogenesis of therat metanephros. The Journal of Clinical Investigation 103, 1299e1307.

Bauer, R., Demeter, I., Hasemann, V., Johansen, J.T., 1980. Structural properties of thezinc site in Cu, Zn-superoxide dismutase; perturbed angular correlation ofgamma ray spectroscopy on the Cu, 111Cd-superoxide dismutase derivative.Biochemical and Biophysical Research Communications 94, 1296e1302.

Bradford, M., 1976. A rapid and sensitive method for the quantitation of microgramquantities of protein utilizing the principle of protein-dye binding. AnalyticalBiochemistry 72, 248e254.

Bronner, R., 1975. Simultaneous demonstration of lipids and starch in plant tissues.Biotechnic and Histochemistry 50, 1e4.

Chwieralski, C.E., Welte, T., Bühling, F., 2006. Cathepsin-regulated apoptosis.Apoptosis 11, 143e149.

Conney, A.H., 1982. Induction of microsomal enzyme by foreign chemicals andcarcinogenesis by polycyclic aromatic hydrocarbons. Cancer Research 42,4875e4917.

Costa, P.M., Caeiro, S., Diniz, M.S., Lobo, J., Martins, M., Ferreira, A.M., Caetano, M.,Vale, C., DelValls, T.A., Costa, M.H., 2009a. Biochemical endpoints on juvenileSolea senegalensis exposed to estuarine sediments: the effect of contaminantmixtures on metallothionein and CYP1A induction. Ectoxicology 18, 988e1000.

Costa, P.M., Diniz, M.S., Caeiro, S., Lobo, J., Martins, M., Ferreira, A.M., Caetano, M.,Vale, C., DelValls, T.A., Costa, M.H., 2009b. Histological biomarkers in liver andgills of juvenile Solea senegalensis exposed to contaminated estuarine sedi-ments: a weighted indices approach. Aquatic Toxicology 92, 202e212.

Costa, P.M., Lobo, J., Caeiro, S., Martins, M., Ferreira, A.M., Caetano, M., Vale, C.,DelValls, T.A., Costa, M.H., 2008. Genotoxic damage in Solea senegalensisexposed to sediments from the Sado Estuary (Portugal): effects of metallic andorganic contaminants. Mutation Research 654, 29e37.

Fedeel, B., Orrenius, S., 2005. Apoptosis: a basic biological phenomenon with wide-ranging implications in human disease. Journal of Internal Medicine 258,479e517.

Fisher, D.B., 1968. Protein staining of ribboned epon sections for light microscopy.Histochemistry and Cell Biology 16, 92e96.

Flowers-Geary, L., Bleczinski, W., Harvey, R.G., Penning, T.M., 1996. Cytotoxicity andmutagenicity of polycyclic aromatic hydrocarbon o-quinones produced bydihydrodiol dehydrogenase. Chemico-Biological Interactions 99, 66e72.

Fournier, N., Atger, V., Paul, J.-L., Sturm, M., Duverger, N., Rothblat, G.H., Moatti, N.,2000. Human ApoA-IV overexpression in transgenic mice induces cAMP-stimulated cholesterol efflux from J774 macrophages to whole serum. Arte-riosclerosis, Thrombosis and Vascular Biology 20, 1283e1292.

Friesen, M.C., Demers, P.A., Spinelli, J.J., Le, N.D., 2008. Adequacy of benzo(a)pyreneand benzene soluble materials as indicators of exposure to polycyclic aromatichydrocarbons in a Söderberg aluminium smelter. Journal of Occupational andEnvironmental Hygiene 5, 6e14.

Fujimoto, K., Yamamoto, T., Kitano, T., Abé, S.-I., 2002. Promotion of cathepsin Lactivity in newt spermatogonial apoptosis induced by prolactin. FEBS Letters521, 43e46.

P.M. Costa et al. / Environmental Pollution 158 (2010) 3338e33463346

Gouazé, V., Andrieu-Abadie, N., Cuvillier, O., Malagarie-Cazenave, S., Frisach, M.-F.,Mirault, M.-E., Levade, T., 2002. Glutathione peroxidase-1 protects from CD95-induced apoptosis. The Journal of Biological Chemistry 277, 42867e42874.

Häcker, G., 2000. The morphology of apoptosis. Cell and Tissue Research 301, 5e17.Hamada, T., Tanimoto, A., Sasaguri, Y., 1997. Apoptosis induced by cadmium.

Apoptosis 2, 359e367.Harris, C.C., Vahakangas, K., Newman, M.J., Trivers, G.T., Shamsuddin, A., Sinopoli, N.,

Mann, D.L., Wright, W.E., 1985. Detection of benzo[ a]pyrene diol epoxide-DNAadducts in peripheral blood lymphocytes and antibodies to the adducts inserum from coke oven workers. Proceedings of the National Academy ofSciences of the USA 82, 6672e6676.

Hishita, T., Tada-Oikawa, S., Tohyama, K., Miura, Y., Nishihara, T., Tohyama, Y.,Yoshida, Y., Uchiyama, T., Kawanishi, S., 2001. Caspase-3 activation by lysosomalenzymes in cytochrome c-independent apoptosis in lyelodysplastic syndrome-derived cell line P39. Cancer Research 61, 2878e2884.

Holme, J.A., Gorria, M., Arlt, V.M., Øvrebø, S., Solhaug, A., Tekpli, X., Landvik, N.E.,Huc, L., Fardel, O., Lagadic-Gossmann, D., 2007. Different mechanisms involvedin apoptosis following exposure to benzo[a]pyrene in F258 and Hepa1c1c7cells. Chemico-Biological Interactions 167, 41e55.

Hurk, P.V.D., Faisal, M., Roberts Jr., M.H., 1998. Interaction of cadmium and benzo[a]pyrene in mummichog (Fundulus heteroclitus): effects on acute mortality.Marine Environmental Research 46, 525e528.

Jiménez-Tenorio, N., Morales-Caselles, C., Kalman, J., Salamanca, M.J., González deCanales, M.L., Sarasquete, C., DelValls, T.A., 2007. Determining sediment qualityfor regulatory proposes using fish chronic bioassays. Environment International33, 474e480.

Kondo, Y., Rusnak, J.M., Hoyt, D.G., Settineri, C.E., Pitt, B.R., Lazo, J.S., 1997. Enhancedapoptosis in metallothionein null cells. Molecular Pharmacology 52, 195e201.

Lewi�nska, D., Arkusz, J., Sta�nczyk, M., Palus, J., Dziuba1towska, E., Stepni, M., 2007.Comparison of the effects of arsenic and cadmium on benzo(a)pyrene inducedmicronuclei in mouse bone-marrow. Mutation Research 632, 37e43.

López-Barea, J., Gómez-Ariza, J.L., 2006. Environmental proteomics and metal-lomics. Proteomics 6, S51eS62.

Manevich, Y., Swaeitzer, T., Pak, J.H., Feinstein, S.I., Muzykantok, V., Fisher, A.B.,2002. 1-Cys peroxiredoxin overexpression protects cells against phospholipidperoxidation-mediated membrane damage. Proceedings of the NationalAcademy of Sciences of the USA 99, 11599e11604.

Marigómez, I., Izaguirre, U., Lekube, X., 2005. Lysosomal enlargement in digestivecells of mussels exposed to cadmium, benzo[a]pyrene and their combination.Comparative Biochemistry and Physiology Part C 141, 188e193.

Martoja, R., Martoja, M., 1967. Initiation Aux Techniques De L’Histologie Animal.Masson & Cie, Paris.

Min, K.-S., Nakatsubo, T., Fujita, Y., Onosaka, S., Tanaka, K., 1992. Degradation ofcadmium metallothionein in vitro by lysosomal proteases. Toxicology andApplied Pharmacology 113, 299e305.

Monsinjon, T., Knigge, T., 2007. Proteomic applications in ecotoxicology. Proteomics7, 2997e3009.

Montes-Nieto, R., Fuentes-Almagro, C.A., Bonilla-Valverde, D., Prieto-Álamo, M.J.,Jurado, J., Carrascal, M., Gómez-Ariza, J.L., López-Barea, J., Pueyo, C., 2007. Pro-teomics in free-living Mus spretus to monitor terrestrial ecosystems. Proteomics7, 4376e4387.

Mouysset, J., Deichsel, A., Moser, S., Hoege, C., Hyman, A.A., Gartner, A., Hoppe, T.,2008. Cell cycle progression requires the CDC-48UFD-1/NPL-4 complex for efficientDNA replication. Proceedings of the National Academy of Sciences of the USA105, 12879e12884.

Mukherjee, J.J., Gupta, S.K., Kumar, S., Sikka, H.C., 2004. Effects of cadmium(II) on(�)- anti-benzo[a]pyrene-7,8-diol-9,10-epoxide-induced DNA damage responsein human fibroblasts and DNA repair: a possible mechanism of cadmium’scogenotoxicity. Chemical Research in Toxicology 17, 287e293.

Murn, J., Urleb, U., Mlinaric-Rascan, I., 2004. Internucleosomal DNA cleavage inapoptotic WEHI 231 cells is mediated by a chymotrypsin-like protease. Genes toCells 9, 1103e1111.

Murphy, F.R., Issa, R., Zhou, X., Ratnarajah, S., Nagase, H., Arthur, M.J.P., Benyon, C.,Iredale, J.P., 2002. Inhibition of apoptosis of activated hepatic stellate cells by

tissue inhibitor of metalloproteinase-1 is mediated via effects on matrix met-alloproteinase inhibition. The Journal of Biological Chemistry 277, 11069e11076.

Naora, H., Naora, H., 1996. Differential expression patterns of b-actin mRNA levels incells undergoing apoptosis. Biochemical and Biophysical Research Communi-cations 211, 491e495.

Nordberg, G.F., 2009. Historical perspectives on cadmium toxicology. Toxicologyand Applied Pharmacology 238, 192e200.

Oh, S.-H., Hatch, H.M., Petersen, B.E., 2002. Hepatic oval “stem” cell in liverregeneration. Seminars in Cell and Developmental Biology 13, 405e409.

Okada, M., Akimaru, H., Houl, D.-X., Takahashi, T., Ishii, S., 2002. Myb controls G2/Mprogression by inducing cyclin B expression in the Drosophila eye imaginal disc.The EMBO Journal 21, 675e684.

Orbea, A., Ortiz-Zarragoitia, M., Cajaraville, M.P., 2002. Interactive effects of benzo(a)pyrene and cadmium and effects of di(2-ethylhexyl) phtalate on antioxidantand peroxisonal enzymes and peroxisomal volume density in the digestivegland of mussel Mytillus galloprovincialis Lmk. Biomarkers 7, 33e48.

Pak, J.H., Manevich, Y., Kim, H.S., Feinstein, S.I., Fisher, A.B., 2002. An antisenseoligonucleotide to 1-cys peroxiredoxin causes lipid peroxidation and apoptosisin lung epithelial cells. The Journal of Biological Chemistry 277, 49927e49934.

Risso-de Faverney, C., Devaux, A., Lafaurie, M., Girard, J.P., Bailly, B., Rahmani, R.,2001. Cadmium induces apoptosis and genotoxicity in rainbow trout hepato-cytes though generation of reactive oxygen species. Aquatic Toxicology 53,65e76.

Robertson, J.D., Orrenius, S., Zhivotovsky, B., 2000. Review: nuclear events inapoptosis. Journal of Structural Biology 129, 346e3568.

Roesijadi, G., Rezvankhah, S., Perez-Matus, A., Mitelberg, A., Torruellas, K., vanVeld, P.A., 2009. Dietary cadmium and benzo(a)pyrene increased intestinalmetallothionein expression in the fish Fundulus heteroclitus. Marine Environ-mental Research 67, 25e30.

Romero-Ruiz, A., Carrascal, M., Alhama, J., Gomez-Ariza, J.L., Abien, J., López-Barea, J., 2006. Utility of proteomics to assess pollutant response of clams fromthe Doñana bank of Guadalquivir Estuary (SW Spain). Proteomics 6,S245eS255.

Sandvik, M., Beyer, J., Goksøyr, A., Hylland, K., Egaas, E., Skaare, J.U., 1997. Interactionof benzo(a)pyrene, 2,3,3V,4,4V,5ehexachlorobiphenyl (PCB 156) and cadmiumon biomarker responses in flounder (Platichthys flesus L.). Biomarkers 2,153e160.

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 apoptosisinduced by cadmium in normal human lung cells. Journal of CellularBiochemistry 89, 335e347.

Spaulding, H.L., Saijo, F., Turnage, R.H., Alexander, J.S., Aw, T.Y., Kalogeris, T.J., 2006.Apolipoprotein A-IV attenuates oxidant-induced apoptosis in mitotic compe-tent, undifferentiated cells by modulating intracellular glutathione redoxbalance. American Journal of Physiology e Cell Physiology 290, C95eC103.

Stenson-Cox, C., Fitzgerald, U., Samali, A., 2003. In the cut of apoptosis, serineproteases come of age. Biochemichal Pharmacology 66, 1469e1474.

Vakharia, D.D., Liu, N., Pause, R., Fasco, M., Bessette, E., Zhang, Q.-Y., Kaminsky, L.S.,2001. Polycyclic aromatic hydrocarbon/metal mixtures: effect on PAH inductionof CYP1A1 in human HepG2 cells. Drug Metabolism and Disposition 29,999e1006.

Viaene, M.K., Masschelein, R., Leenders, J., de Groof, M., Swerts, L.J.V.C., Roels, H.A.,2000. Neurobehavioural effects of occupational exposure to cadmium: a crosssectional epidemiological study. Occupational and Environmental Medicine 57,19e27.

Völkl, H., Fridrich, F., Häussinger, D., Lang, F., 1993. Effect of cell volume on acridineorange fluorescence in hepatocytes. Biochemical Journal 265, 11e14.

Wang, X., Li, N., Liu, B., Sun, H., Chen, T., Li, H., Qiu, J., Zhang, L., Wan, T., Cao, X., 2004.A novel human phosphatidylethanolamine-binding protein resists tumournecrosis factor a-induced apoptosis by inhibiting mitogen-activated proteinkinase pathway activation and phosphatidylethanolamine externalization. TheJournal of Biological Chemistry 44, 45855e45864.

Williams, M.S., Henkart, P.A., 1994. Apoptotic cell death induced by intracellularproteolysis. The Journal of Immunology 153, 4247e4255.