Biochimica et Biophysica Acta 1830 (2013) 4778–4789

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

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

Molecular insights into 4-nitrophenol-induced hepatotoxicity inzebrafish: Transcriptomic, histological and targeted geneexpression analyses

Siew Hong Lam a,b,⁎,1, Choong Yong Ung a,d,1, Mya Myintzu Hlaing a, Jing Hu a, Zhi-Hua Li a,Sinnakaruppan Mathavan c, Zhiyuan Gong a,b,⁎⁎a Department of Biological Sciences, National University of Singapore, 14 Science Drive 4, Singapore 117543, Singaporeb NUS Environmental Research Institute (NERI), TL #02-02, Engineering Drive 1, Singapore 117411, Singaporec Genome Institute of Singapore, Agency for Science Technology and Research, Genome, 60 Biopolis Street, Singapore 138672, Singapored Bioinformatics Programme, Institute of Biological Sciences, University of Malaya, 50603 Kuala Lumpur, Malaysia

⁎ Correspondence to: S.H. Lam, Department of BiologicalDrive 4, National University of Singapore, Singapore65167379; fax: +65 67792486.⁎⁎ Correspondence to: Z. Gong, Department of BiologicalDrive 4, National University of Singapore, Singapore65162860; fax: +65 67792486.

E-mail addresses: dbslsh@nus.edu.sg (S.H. Lam), dbs1 Equal contribution from both authors.

0304-4165/$ – see front matter © 2013 Elsevier B.V. Alhttp://dx.doi.org/10.1016/j.bbagen.2013.06.008

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 21 November 2012Received in revised form 3 May 2013Accepted 9 June 2013Available online 17 June 2013

Keywords:4-NitrophenolZebrafishLiverTranscriptomicMicroarrayToxicity

Background: 4-Nitrophenol (4-NP) is a prioritized environmental pollutant and its toxicity has been investigatedusing zebrafish, advocated as an alternative toxicological model. However, molecular information of 4-NP in-duced hepatotoxicity is still limited. This study aimed to obtainmolecular insights into 4-NP-induced hepatotox-icity using zebrafish as a model.Methods: Adult male zebrafish were exposed to 4-NP for 8, 24, 48 and 96 h. Livers were sampled for microarrayexperiment, qRT-PCR and various histological analyses.Results: Transcriptomic analysis revealed that genes associated with oxidative phosphorylation and electrontransport chain were significantly up-regulated throughout early and late stages of 4-NP exposure due to oxi-dative phosphorylation uncoupling by 4-NP. This in turn induced oxidative stress damage and up-regulatedpathways associated with tumor suppressors Rb and p53, cell cycle, DNA damage, proteasome degradationand apoptosis. Pathways associated with cell adhesion and morphology were deregulated. Carbohydrate andlipid metabolisms were down-regulated while methionine and aromatic amino acid metabolisms as well as

NFKB pathway associatedwith chronic liver conditions were up-regulated. Up-regulation of NFKB, NFAT and in-terleukin pathways suggested hepatitis. Histological analyses with specific staining methods and qRT-PCR anal-ysis of selected genes corroborated with the transcriptomic analysis suggesting 4-NP induced liver injury.Conclusion: Our findings allowed us to propose a plausible model and provide a broader understanding of themolecular events leading to 4-NP induced acute hepatotoxicity for future studies involving other nitrophenolderivatives.General significance: This is the first transcriptomic report on 4-NP induced hepatotoxicity.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

4-Nitrophenol (para-nitrophenol; 4-NP) is widely used in the man-ufacture of pesticides, fungicides, paints, dyes, leather preservative anddrugs; and it has been found commonly inmany industrialwastewatersand diesel exhaust [1,2]. Due to the high-volume production, applica-tions and potential toxicity, 4-NP has been classified as ‘organic priority

Sciences, S3-Level 5, 14 Science117543, Singapore. Tel.: +65

Sciences, S3-Level 5, 14 Science117543, Singapore. Tel.: +65

gzy@nus.edu.sg (Z. Gong).

l rights reserved.

pollutants’ [1,2]. Moreover, since 4-NP is an intermediate in the synthe-sis of acetaminophen which is used as a popular analgesic and antipy-retic in many pharmaceutical formulations, the presence of 4-NP as animpurity in pharmaceuticals and its impact on consumer's health isalso of concern [3]. 4-NP is known to resist biodegradation and there-fore can accumulate andmaymagnify from lower to higher tropic levelsin both aquatic and terrestrial organisms [4–7]. It has been detected inhuman urine likely due to exposure to organophosphorus pesticides[7,8]. Recently, perhaps of great concern, it has been shown that 4-NPhas endocrine-disrupting activities perturbing steroidal hormone sig-naling pathways in rodent models [9–11]. Although 4-NP has relativelylow toxicity, there are still concerns of the increased accumulation of4-NP in water, soil and air, through the degradation of widely used or-ganophosphorus pesticides and diesel exhaust emissions that couldcause adverse effects on wildlife and human health [6,7]. Consequently,there is a concern of 4-NP entering the food chain, since besides surface

4779S.H. Lam et al. / Biochimica et Biophysica Acta 1830 (2013) 4778–4789

water and sediments, the compound had been detected in vegetables[12] and fish [13].

It has been shown that the biological responses of zebrafish tochemicals, such as small molecules, drugs and environmental toxicants,are similar to that of mammals [14–16]. Therefore, in view of the envi-ronmental health concern of 4-NP as well as the advantages and in-creasing advocacy of using zebrafish as an alternative model fortoxicology studies [17–19], the zebrafish has been used to investigate4-NP metabolism and toxicity. The 96-hour acute toxicity test for4-NP had been performed in adult zebrafish and the LC50 was reportedbetween 10 and 14 mg/L [20,21]. In mammals, 4-NP is metabolized pri-marily in the liver into conjugated glucuronide and sulfate ester formsbefore being excreted from the body [1]. 4-NP can be metabolized byhepatic microsomal enzyme systems via glucuronidation and sulfation[22]. In zebrafish, glucuronidation and sulfation of 4-NP into respectiveglucuronide and sulfate conjugates had been demonstrated [23]. Thisfurther suggests some similarity of 4-NP metabolism in zebrafish andmammals. The zebrafish had been used for ultrastructural study of hep-atotoxicity effects after a prolonged 2–3 month exposure to 4-NP [20].It was observed that 25% of the fish treated with 1 and 5 mg/L of 4-NPshowed degenerative transformation of the liver tissue, and deforma-tion of the nuclear membrane with partial lysis of the mitochondriacould be observed at 5 mg/L [20]. However, mechanistic informationon the 4-NP induced hepatoxicity in the zebrafish, as well as in mam-mals, is still limited at the molecular level. Since the liver is the mainorgan that performs detoxification processes as well as regulation ofmetabolic pathways, it is important to further understand the in vivotoxic effects induced by 4-NP in the liver.

The primary aim of this study was to obtain molecular insights into4-NP-induced hepatotoxicity using zebrafish as a model. DNA microar-ray was used to investigate the hepatic transcriptomic changes inducedby 4-NP. The transcriptomic approach is applied in toxicology to deter-mine the relative changes in gene expression levels induced by exposureto toxicant(s). In zebrafish, transcriptomic technology has been appliedto capture differential gene expression profiles induced by specific toxi-cants [16,24]. Moreover, toxicant-induced global gene expression pro-files of zebrafish have been demonstrated to be useful for identifyingbiomarkers of effects and deregulated signaling pathways, as well asfor inferring possible affected biological functions, perturbed physiolog-ical systems and increased health-risks for mechanistic and predictivetoxicology [16,24–28]. For example, it has been shown that genes thatare predictive of prenatal arsenic exposure in humans were also identi-fied as zebrafish orthologs in arsenic-treated zebrafish embryos usinggene expression profiling [29]. To our knowledge, there has been nopublication on the global gene expression profiling study to investigatethe transcriptomic changes induced by 4-NP. Therefore, this is the firsttranscriptomic report on 4-NP induced hepatotoxicity. Furthermore,we corroborated our transcriptomic analysis with various histologicalanalyses to further understand the toxicity mechanism of 4-NP. Wealso conducted gene-targeted validation using real-time PCR performedon an independent batch of fish samples treated with different concen-trations of 4-NP, and these validated genes were used to further supportthe transcriptomic analysis of 4-NP induced hepatotoxicity. Taken to-gether, our findings allow us to propose a plausible integrated modeldepicting 4-NP induced hepatotoxicity and provide a broader under-standing of themolecular events leading to acute hepatotoxicity for fur-ther studies.

2. Experimental procedures

2.1. 4-NP exposure and fish sampling

Four main experiments were performed independently in thisstudy, i.e. preliminary acute toxicity test, microarray experiment, histo-logical analyses, and real-time PCR targeted gene expression validation.To determine suitable sub-lethal and low lethal concentrations for the

experiment, male adult zebrafish were treated with 4-NP (Sigma-Aldrich, USA) at nominal concentrations of 1, 5 and 7 mg/L for 96 h at27 ± 2 °C in a static condition. Control group was maintained inwater as 4-NP is soluble in water at the concentration used and no or-ganic vehicle was used. Fresh chemical and water were renewed daily.For themicroarray experiment, a new batch of male adult fish were ex-posed to 4-NP at a low-lethal concentration of 7 mg/L for 96 hwith a re-spective control group maintained in a similar condition as describedabove. Triplicate pooled liver samples, where each replicate was pooledfrom four individual livers, were obtained for each sampling time-pointat 8, 24, 48 and 96 h treatment for themicroarray experiments. For his-tology and real-time PCR targeted gene expression validation experi-ments, another new batch of male adult fish were exposed to vehicle0.5 and 5 mg/L of 4-NP for 96 h maintained with a control groupunder a similar condition as above. Livers from five fish were sampledindividually for targeted gene expression validation experiment at96 h. All liver samples were snap-frozen in liquid nitrogen and storedat−80 °C for total RNA extraction formicroarray and targeted gene ex-pression validation experiments. For histological analysis, livers fromfour fish of each concentration of the 4-NP treated group and the controlgroup were fixed in Bouin's solution for hematoxylin and eosin (H & E)staining, another four livers from each group were fixed in formalin so-lution (10%, Neutral Buffered, Sigma-Aldrich) for immunohistochemicaland periodic acid-Schiff (PAS) staining, and another four livers fromeach group were freshly frozen for Oil-Red O staining.

2.2. Total RNA extraction

Total RNA was extracted from the fish liver samples using Trizol re-agent (Invitrogen, USA) according to the manufacturer's instructions.The integrity of RNA samples was verified by gel electrophoresis andby UV spectrophotometer (Nanodrop 2000; Thermo Scientific, USA).The concentrations were also determined by UV spectrophotometer.

2.3. Microarray hybridization

The arrays contained 16.4 K oligonucleotide probes. The probeswere resuspended in 3× SSC at 20 μM concentration and spotted ontoin-house poly-L-lysine-coated microscope slides using a custom-builtDNA microarrayer in GIS. A two-color microarray hybridization formatbetween test versus common reference, and control versus common ref-erence was used. Common reference RNA for microarray hybridizationwas obtained by pooling total RNA extracted from untreated wholeadult male and female fish. Both experimental (test and control) sam-ples and common reference RNAswere reverse-transcribed and labeledwith fluorescent dyes Cy-5 and Cy-3 (Amersham, USA), respectively.After hybridization at 42 °C for 16 h in hybridization chambers (GeneMachines, USA), themicroarray slides were washed in a series of wash-ing solutions (2× SSC with 0.1% SDS; 1× SSC with 0.1% SDS; 0.2× SSCand 0.05× SSC; 30 s each), dried using low-speed centrifugation andscanned for fluorescence detection using the GenePix 4000Bmicroarrayscanner (Axon Instruments, USA). Further detailed protocol for the mi-croarray hybridization and data acquisition had been previously de-scribed [30,31]. The microarray data had been submitted to the GeneExpression Omnibus (GEO) with accession number GSE30058.

2.4. Microarray data processing and transcriptome analysis

The raw microarray data was normalized using Lowess method inthe R package (http://www.braju.com/R/) followed by Student's t-test.To increase statistical power, triplicate arrays of 8 and 24 hwere pooledand analyzed together as ‘early responsive stage’ whereas arrays of 48and 96 h were pooled and analyzed together as ‘late responsive stage’.Although the resolution time/kinetics is reduced into two phases, thiswill help to capture genes that are showing a similar expression trend,hence more consistent and robust, by considering two sampling time

4780 S.H. Lam et al. / Biochimica et Biophysica Acta 1830 (2013) 4778–4789

points, i.e. the early (between 8 and 24 h) and the late (between 48 and96 h). Gene set enrichment analyses (GSEA) as described in detail bySubramanian et al. [32] was used to determine biological pathwaysperturbed by 4-NP. Genes were ranked based on their p-values usingStudent t-test. The ‘GSEA Preranked’ option of GSEAwas used. The rank-ing metric used was log10 (1/P) where P is the p-value of a given gene.Up-regulated genes will carry positive values of log10 (1/P) whereasdown-regulated genes will carry negative values of log10 (1/P). Thegenes were then ranked in descending order based on values of log10(1/P) taking into account of the positive and negative assignments.The ranked list of genes will then be compared to pre-defined genesets or signatures for canonical pathways.

An enrichment score (ES) that reflects the degree to which apre-defined gene set is over-represented at the top or bottom rankof the ranked whole transcriptome profile was calculated by walkingdown the ranked profile, increasing a running-sum statistic when agene in a pre-defined gene set was encountered and decreasing itwhen the gene was absent. The ES is either the maximum or the min-imum deviation from zero encountered in the random walk thatcorresponds to a weighted Kolmogorov–Smirnov-like statistic. Thestatistically significant value (nominal p-value, NP) of the ES was es-timated by using an empirical phenotype-based permutation testprocedure. The phenotype labels were permuted and the ES of thegene set for the permuted data was recomputed to generate a nulldistribution for the ES. The empirical, nominal p-value of the ob-served ES was then calculated relative to this null distribution. The es-timated significance level was adjusted with multiple hypothesistesting. The ES for each gene set was first normalized to the size ofthe set yielding a normalized enrichment score (NES) with the fol-lowing relation:

NES ¼ actual ES=mean ESs against all permutations of the datasetð Þ:

The number of permutation used was 1000. Pathways with nominalp-value (NP) b 0.05 are considered statistically significant. Positive andnegative values of normalized enrichment scores (NES) indicate genesof pathways that were up- and down-regulated, respectively.

2.5. Histological sample preparation and imaging analysis

Formalin-fixed as well as Bouin's-fixed liver samples were washedseveral times with 70% ethanol, followed by dehydration in a series ofascending concentrations of ethanol (70%–100%) before clearing inHistoclear and embedding in paraffin. The paraffin-embedded tissueswere sectioned longitudinally at 5 μm thickness. The sections werestained with Mayer's H & E (Sigma-Aldrich, USA) for qualitative andquantitative histological analysis. Imaging of liver sections was carriedout using a compound microscope, Axioskop 2, Zeiss® equipped withan imaging system and were analyzed with a computer-assistedimage analyzer program (Axiovision, Zeiss). Liver sections from treatedfishwere comparedwith control fish, both quantitatively (i.e. nucleatedcell density) and qualitatively (i.e.descriptive observation) from the he-matoxylin and eosin-stained section. The number of nucleated cells (no.of nucleated cells/11,000 μm2) was counted in both the 4-NP-treatedgroup and the control group. Three images from each portion (anterior,middle and posterior portions) of liver sections (at 1000× magnifica-tion) for each liver from four control and four 4-NP-treated fish weredetermined for each image. Student's t-test was performed to analyzethe data for statistical significance (P b 0.05; n = 4 replicates).

2.6. Immunohistochemical staining for E-cadherin and Cytokeratin

For immunohistochemistry of paraffin sections, antigen retrievalwas performed by heating in 10 mM sodium citrate buffer for 10 min.After cooling down at room temperature for 30 min, peroxidaseblocking was done with 3% hydrogen peroxide for 10 min, followed

by blocking buffer (5% goat serum-PBS) for 1 h to block non-specificbinding. The sectionswere incubatedwith primary antibodies includingmouse monoclonal anti-human E-cadherin (1:300 dilution; BD Biosci-ences, USA) and mouse monoclonal anti-human Cytokeratin-18(1:500 dilution; Millipore, USA) at 4 °C overnight. Secondary antibody,rabbit anti-mouse antibody conjugated with horseradish peroxidase(Zymax grade; DAKO, Denmark), was used for color-reactivity detec-tion by 3,3′-diaminobenzidine tetrahydrochloride (DAB) substrate-chromogen (DAKO, Denmark). The color reaction was stopped assoon as positive-reactivity was revealed by the formation of abrown-colored precipitate at the target antigen in the section. Sectionswere counterstained with Mayer's hematoxylin for 1 min. Images ofliver sections were captured at 1000× magnification using an Axiovertmicroscope (Axivision; Zeiss, Germany).

2.7. ApopTag staining

DNA fragmentations associated with structural changes in cellularmorphology in apoptosis induced by 4-NP were detected usingApopTag® Plus Fluorescein In Situ Apoptosis Detection Kit (Chemicon,USA) according to the manufacturer's protocol. The 3′-OH ends ofdouble-stranded or single-stranded DNA were labeled with thedigoxigenin (DIG)-nucleotide and then allowed to bind an anti-digoxigenin antibody (anti-DIG) that is conjugated to fluorescein or al-kaline phosphatase. In this experiment, we used anti-DIG conjugatedwith fluorescein and alkaline phosphatase. For using anti-DIG con-jugated with alkaline phosphatase, the DNA fragmentations localizedin apoptotic bodies were detected by enzymatically using 5-bromo-4-chloro-3-indolyl phosphate (BCIP)/nitroblue tetrazolium (NBT) sub-strate. DNA fragmentations localized in apoptotic bodies induced by4-NP were viewed as the intense insoluble purple color (400× and1000× magnifications) and captured using an Axiovert microscope(Axivision; Zeiss, Germany).

2.8. Periodic acid-Schiff (PAS) staining

PAS staining is mainly used for staining structures containing a highproportion of carbohydrate macromolecules (glycogen, glycoprotein,proteoglycans) typically found in connective tissue, mucus and basallamina. Staining for glycogen was performed using Alcian BluePAS stain kit (BioGenex, USA) without diastase according to themanufacturer's protocol. Images of PAS stained sections (1000×magni-fication) were captured using an Axiovert microscope (Axivision; Zeiss,Germany).

2.9. Oil-Red O staining

For staining of lipid globules accumulated in hepatocytes, cryostatsectioning was performed from fresh frozen liver samples followed byOil-Red O staining according to the manufacturer's protocol (Sigma-Aldrich, USA). Images of Oil-Red O stained sections (1000× magnifica-tion) were captured using an Axiovert microscope (Axivision; Zeiss,Germany).

2.10. cDNA synthesis and quantitative real-time PCR validation

Equal amounts of total RNA samples from the 4-NP-treated experi-ments were reverse transcribed to cDNA using SuperScript™ II RT Kit(Invitrogen, USA) according to the manufacturer's instructions. ThecDNA samples were used for quantitative real-time PCR analysis usingthe Lightcycler 480 system (Roche Applied Science) with Lightcycler480 SYBR Green I Master kit (Roche Applied Science) according to themanufacturer's instructions. Primers used for PCR are listed inAdditionalFile 1. Statistical comparison of the relative mean expression level foreach gene between a treated group and a control group was performed

Fig. 1. Gene set enrichment analysis (GSEA) profiles of the enriched functional categories in the liver of fish treated with 4-nitrophenol (4-NP). Normalized enrichment scores (NES)generated from GSEA for early (8 and 24 h) and late (48 and 96 h) stages of 4-NP exposure were used to represent the ‘meta-activities’ of pathways where significant (P b 0.05) up-and down-regulation of the pathways were indicated by respective dark red and dark blue cells. The NES values for these pathways are given in Additional File 2.

4781S.H. Lam et al. / Biochimica et Biophysica Acta 1830 (2013) 4778–4789

using Student's t-test (n = 5 replicates). Log2 fold change between eachtreatment group and control group is shown.

3. Results and discussion

3.1. General acute toxicity test for 4-NP in zebrafish

Acute toxicity testwas performed to determine the appropriate con-centrations of 4-NP for DNA microarray experiments. Adult malezebrafish were exposed to 4-NP at 1 mg/L, 5 mg/L and 7 mg/L for

96 h. Based on the reported LC50 for 96 h in zebrafishwhich is between10 and 14 mg/L of 4-NP [20,21], we have chosen slightly lower concen-trations with the aim of determining sub-lethal and low lethal concen-trations. No mortality was observed at 1 mg/L and 5 mg/L during theexperimental period, but 25% mortality was observed at 7 mg/L within48 h and thereafter nomortality was observed suggesting that an adap-tive response to toxicity has occurred. For a subsequent microarray ex-periment, we therefore exposed fish to 7 mg/L 4-NP and sampled theliver tissue at 8, 24, 48 and 96 h. The exposure to this low lethal concen-tration and sampling timewould likely capture a spectrum of responses

D

10 μμm

C

A B

EN

um

ber

of

nu

clea

ted

cel

ls

Concentration of 4-nitrophenol

****

real -time PCR data

microarray data

* = p value ≤ 0.1** = p value ≤ 0.05

00.5

11.5

22.5

3

Lo

g2F

old

-Ch

ang

eV

ersu

s C

on

tro

l

Concentration of 4-NP

tp53

**

****

ccng1

****

bax

**

***

0

1

2

3

4

5mdm2

**

**

**

0

1

2

3

4

casp8

**

**

**

0.5 mg/L 5 mg/L 7 mg/L012345

0.5 mg/L 5 mg/L 7 mg/L0

0.51

1.52

2.53

0.5 mg/L 5 mg/L 7 mg/L

0.5 mg/L 5 mg/L 7 mg/L 0.5 mg/L 5 mg/L 7 mg/L

0

20

40

60

80

100

120

Control 0.5 mg/L 5 mg/L

Fig. 2. Molecular changes associated with decreased nucleated cell number in the liver of zebrafish treated with 4-nitrophenol (4-NP) for 96 h. (A) Gene expression profiles forderegulated leading edge genes involved in DNA damage signaling, apoptosis, and proteasome degradation as determined by GSEA. Color scale represents log2 fold-change above the cor-respondingmatch time-point control group where red, green and black cells represent up-regulation, down-regulation and no change in gene expression, respectively. (B) Quantitativehistological analysis of nucleated-cell count from H & E stained sections. Nucleated-cell count in the liver significantly (**p value b 0.05; n = 4 replicates) decreased in aconcentration-dependent manner in fish treated with 0.5 and 5 mg/L of 4-NP when compared to the control group. (C–D) Apoptag staining (1000×) indicated the presence ofdark-staining apoptotic bodies (D) suggesting DNA damage induced by apoptosis in the liver of fish treated with 5 mg/L of 4-NP when compared to control (C). (E) Quantitativereal-time PCR for tp53, ccng1, bax, mdm2, and casp8 indicated up-regulation of gene expression which corroborated with microarray data (*p-value ≤ 0.1; **p-value ≤ 0.05; n = 5replicates).

4782 S.H. Lam et al. / Biochimica et Biophysica Acta 1830 (2013) 4778–4789

to 4-NP induced hepatotoxicity ranging from early to late/advancedtoxic effects that resulted in mortality, which would providemechanis-tic insights into 4-NP induced hepatotoxicity. Lower sub-lethal concen-trations (0.5 mg/L and 5 mg/L) of 4-NP exposure to capture dose-dependent effects and infer low dose effects of 4-NP were selectivelycarried out for qRT-PCR of specific genes and histological analyses.

3.2. Transcriptomic analysis revealed the existence of major functionalcategories deregulated by 4-NP

Gene set enrichment analysis (GSEA) was carried out on thetranscriptomic data (see Experimental procedures) and revealed that

4-NP affected pathways and processes that belonged to certain func-tional classes in fish livers. Normalized enrichment scores (NES)obtained from GSEA were used to represent ‘meta-activities’ of path-ways. Down-regulated pathways are enriched with down-regulatedRNA transcripts associated with the pathways and therefore carryingnegative NES values, and vice versa for the up-regulated pathwayswhich represented positive NES values (see Experimental procedures;Additional File 2). Pathways with nominal p-value b 0.05 were consid-ered significantly deregulated by 4-NP. As shown in Fig. 1, the pathwayswere grouped into five major related categories i.e. (i) ‘bioenergetic andoxidative phosphorylation’, (ii) ‘cell cycle, DNA damage, proteasomedegradation, and apoptosis’, (iii) ‘cell adhesion, cytoskeletal

4783S.H. Lam et al. / Biochimica et Biophysica Acta 1830 (2013) 4778–4789

organization and cell morphology’, (iv) ‘metabolisms’, and (v) ‘otherpathways’. Some of these pathways/processes were significantlyderegulated only in an early (8 and 24 h) responsive stage or a late(48 and 96 h) responsive stage while others were affected throughoutearly and late responsive stages. We further validated some of thesefunctional categories using histological analyses and qRT-PCR ontargeted genes, in order to corroborate with the transcriptomic analysis.

B

10 μμm

E F

A

C

G H

10 μm

10 μm

Fig. 3. Histological andmolecular changes in zebrafish liver treatedwith 4-nitrophenol (4-NP) fcell adhesion and cytoskeleton as determined by GSEA. Color scale represents log2 fold-changerepresent up-regulation, down-regulation and no change in gene expression, respectively. (B–sections revealed a less compact and less homogeneous distribution of hepatic parenchyma cell4-NP (C), when compared with the liver of control fish (B). (E and F) Immunohistochemical sthepatocytes in fish treated with 5 mg/L 4-NP treatedmale fish (F) compared to the control groand less intense staining (brown precipitate) in hepatocytes of fish treated with 5 mg/L 4-NP

3.3. 4-NP deregulates oxidative phosphorylation and induces cell death

GSEA of transcriptomic data revealed significant deregulation inbioenergetic pathway by 4-NP (Fig. 1). The expression of genes asso-ciated with oxidative phosphorylation and electron transport chainwere significantly up-regulated throughout early and late stages of4-NP exposure. This can be attributed to oxidative phosphorylation

D

or 96 h. (A)Microarray expression profiles for leading edge genes that are associatedwithabove the correspondingmatch time-point control group where red, green and black cellsD) Histological examination of representative hematoxylin and eosin (H & E) stained livers in fish treatedwith 5 mg/L of 4-NP (D), but less apparent in fish treatedwith 0.5 mg/L ofaining for E-cadherin showed weak signals (brown precipitate) on the cellular borders ofup (E). (G and H) Immunohistochemical staining for cytokeratin showed less wide spreadtreated male fish (G) compared to the control group (H).

CB

10 μμm

D E

A

F

real-time PCR data

microarray data

* = p value ≤ 0.1** = p value ≤ 0.05

0.5 mg/L 5 mg/L 7 mg/L

Lo

g2F

old

-Ch

ang

eV

ersu

s C

on

tro

l

Concentration of 4 - NP

pklr

* ** -3-2.5

-2-1.5

-1-0.5

00.5

0.5 mg/L 5 mg/L 7 mg/L

acads

**

**

-2

-1.5

-1

-0.5

0

0.5

0.5 mg/L 5 mg/L 7 mg/L

elovl6

**** -2

-1.5

-1

-0.5

0

0.5 mg/L 5 mg/L 7 mg/L

nqo1

**

**

0.5 mg/L 5 mg/L 7 mg/L

hsdl2

**

**

0.5 mg/L 5 mg/L 7 mg/L

fads2

****

0.5 mg/L 5 mg/L 7 mg/L

acox3

** ** 0.5 mg/L 5 mg/L 7 mg/L

cpt2

** **

-1.5-1

-0.50

0.51

1.52

2.5

-2

-1.5

-1

-0.5

0

0.5

1

-5-4-3-2-10123

-1

-0.8

-0.6

-0.4

-0.2

0

-0.8-0.6-0.4-0.2

0

-1

4784 S.H. Lam et al. / Biochimica et Biophysica Acta 1830 (2013) 4778–4789

4785S.H. Lam et al. / Biochimica et Biophysica Acta 1830 (2013) 4778–4789

uncoupling by 4-NP. Phenolic compounds, including 2,4-dinitrophe-nol a derivative of 4-NP, have been shown to uncouple oxidativephosphorylation [33,34]. The defining feature of oxidative phosphor-ylation uncouplers is that they inhibit ATP synthesis via decouplingthe activity of F0F1 ATP synthase from the electron transport chain,leading to rapid depletion of intracellular ATPs [33]. The adaptive re-sponse to cope with the energetic demands was to increase oxidativephosphorylation and electron transport chain activities, and to in-crease generation of ATPs via glycolysis or breakdown of other sub-strates such as lipids [35]. In addition, uncouplers of mitochondrialoxidative phosphorylation had been reported to increase mito-chondrial respiration, increase ROS production and deplete cellularglutathione rapidly, resulting in oxidative stress and inducingcell cycle arrest and apoptosis [34–36]. This further explained theincreased apoptosis and reduced nucleated cell count observed inthis study.

The transcriptomic analysis using GSEA also revealed that 4-NP in-duced hepatotoxicity up-regulated tumor suppressor retinoblastoma(RB) pathway, p53 pathway, cell cycle and G2 pathways, DNA damagesignaling and apoptosis-related pathways including Death and BADpathways (Fig. 1). It has been demonstrated that several deregulatedDNA damage pathways were associated with an elevation in reactiveoxygen species (ROS) suggesting oxidative stress-induced DNA andprotein damage as a plausible component to hepatotoxicity in rodent[37]. The uncoupling of oxidative phosphorylation by 4-NP increasedROS that could cause DNA and protein damage, as well as induce cellcycle arrest and apoptosis [34–36]. Consequently, oxidative stress in-duced damage and apoptosis will also cause elevated amounts ofmisfolded proteins and partially folded protein remnants that are likelyto increase proteasome degradation activities (Fig. 1). As observed, DNAdamage signaling was significantly enriched at late responsive stageonly, hence suggesting that 4-NP is unlikely to have cause DNA damagedirectly but rather indirectly via oxidative damage and inducedmitochondrial-dependent apoptosis. This observation is also consistentwith a recent study indicating that 4-NP is not genotoxic [3].

Our transcriptomic data showed that the expression of genes in-volved in apoptosis was deregulated by 4-NP treatment [Fig. 2A]. Quan-titative histological analysis on nucleated hepatic parenchyma cellcount based on H & E stained sections revealed that the number of he-patic parenchyma cells decreased significantly at 0.5 and 5 mg/L of4-NP (Fig. 2B). To verify if apoptosis has occurred, we performedApoptag staining on liver sections from fish treated with 5 mg/L of4-NP and compared them with the control group. We observed appar-ent staining for DNAbreakage and fragmentations induced by apoptosisin the liver of fish treatedwith 5 mg/L of 4-NP (Fig. 2D)when comparedto control (Fig. 2C) confirming that apoptosis had indeed occurred.Quantitative real-time PCR for tumor suppressor protein p53 (tp53),cyclin G1 (ccng1), Bcl2-associated X protein (bax), transformed 3T3 celldouble minute 2 homolog (mdm2), and caspase 8 (casp8), indicatesup-regulation of these genes (Fig. 2E) which is consistent with themicroarray data (Fig. 2A). The p53 protein is known to activate apo-ptotic factor Bax in order to induce mitochondrial-dependent intrin-sic apoptotic pathway [38,39]. Moreover, it was observed that mostof these genes were significantly up-regulated at 0.5 and 5 mg/L of4-NP which corroborated with the significant decrease in nucleatedhepatic cell count. Therefore, it is likely that the hepatocyte cellnumber decreased via apoptosis as a result of 4-NP induced acutehepatotoxicity.

Fig. 4. Deregulation of metabolic pathways by 4-nitrophenol (4-NP). (A) Gene expression prodetermined by GSEA. Color scale represents log2 fold-change above the corresponding matdown-regulation and no change in gene expression, respectively. (B and C) Histological anacogen content (indicated by light pink staining) in the liver section from fish treated with 5ysis of zebrafish liver stained with Oil-Red O staining revealed less lipid content (stained bri(D). (F) Quantitative real-time PCR for genes involved in carbohydrate and lipid metabolism

3.4. 4-NP affects cellular adhesion, cytoskeletal organization and cellmorphology

Results from GSEA suggested that gene sets correlated with cellularadhesion and communication were down-regulated by 4-NP treatmentat late responsive stage. These adhesion-related gene sets were ‘celladhesion’, ‘cell communication’, ‘ECM receptor interaction’, ‘tight junc-tion’, ‘focal adhesion’, and ‘regulation of actin cytoskeleton’ (Fig. 1). Ex-pression profile for the deregulated genes for these processes is shownin Fig. 3A, suggesting an overall decrease in cellular integrity and cell–cell interactions as the experiment progressed. Histological examina-tions of liver sections stained with H & E revealed a less compact andless homogeneous distribution of hepatic parenchyma cells thatappeared more dissociated and irregular in shape in fish treated with5 mg/L of 4-NP (Fig. 3D), but less apparent in fish treated with0.5 mg/L of 4-NP (Fig. 3C), when compared with the liver of controlfish (Fig. 3B). Eosin staining of cytoplasmic proteins appeared to bemore granulated in hepatocytes of treated fish compared to thesmooth-pink cytoplasmic staining in hepatocytes of the control group.Immunohistochemical staining for E-cadherin, a transmembrane glyco-protein thatmediates epithelial cell-to-cell adhesion, showedweak sig-nals on the cellular borders of hepatocytes in fish treated with 5 mg/Lof 4-NP (Fig. 3F) when compared with the control group (Fig. 3E). Like-wise, immunohistochemical staining for cytokeratin revealed a lesswidespread and less intense staining in the liver of fish treated with5 mg/L of 4-NP (Fig. 3H) when compared with the control group(Fig. 3G). The histological findings thus corroborated with GSEA resultsindicating that intercellular adhesion and tissue integrity as well ascell morphology in the liver were affected by 4-NP induced acutehepatotoxicity.

3.5. 4-NP acts as a metabolic disruptor in the liver

The transcriptomic analysis using GSEA showed that carbohydrateresponsive element binding protein (ChREBP) pathway, which is in-volved in both carbohydrate and lipid metabolisms, was up-regulatedearly followed by down-regulation of feeder pathway associated withcarbohydrate metabolism and mitochondrial fatty acid beta-oxidationat the late responsive stage (Fig. 1), suggesting rapid depletion of cellularcarbohydrate and lipid. Microarray data showed that genes involved incarbohydrate and lipid metabolisms were deregulated (Fig. 4A). To de-termine if glycogen, a storage form of carbohydrate, and lipid contentin hepatocyte of fish treated with 5 mg/L of 4-NP were indeed lowerthan the control group, periodic acid-Schiff (PAS) staining for glycogenand Oil-Red O staining for lipid were performed. PAS staining revealedthat the liver of fish treated with 5 mg/L 4-NP had very weak toundetectable staining for cellular glycogen content (Fig. 4C) when com-pared to control fish (Fig. 4B) suggesting that the treated fish had lowerglycogen content in hepatocytes. This result also corroborated with pre-vious ultrastructural findings that sublethal exposure to 5 mg/L of 4-NPcaused a depletion of hepatic glycogen in zebrafish [20]. Furthermore,Oil-Red O staining for lipid showed that there were smaller and muchless lipid globules in the liver of fish treated with 5 mg/L of 4-NP(Fig. 4E) when comparedwith the control group (Fig. 4D). The depletionof hepatic glycogen and lipid content was likely caused by theuncoupling effects of oxidative phosphorylation described earlier. Un-couplers of oxidative phosphorylation such as 2,4-dinitrophenol havebeen shown to increase metabolism [40], and increase beta-oxidation

files for deregulated leading edge genes involved in lipid and xenobiotic metabolism asch time-point control group where red, green and black cells represent up-regulation,lysis of zebrafish liver stained with periodic acid Schiff staining (PAS) showed less gly-mg/L 4-NP for 96 h (C) when compared to control fish (B). (D and E) Histological anal-ght red) from fish treated with 5 mg/L 4-NP for 96 h (E) when compared to control fishs (*p-value ≤ 0.1; **p-value ≤ 0.05; n = 5 replicates) compared to microarray data.

4786 S.H. Lam et al. / Biochimica et Biophysica Acta 1830 (2013) 4778–4789

of fatty acids that leads to a reduction in adipose tissue and therefore hasbeen used as a slimming agent, however, with fatal consequences due toits toxicity [41]. Further targeted validation using quantitative real-timePCR confirmed that the genes associated with carbohydrate and lipidmetabolism are down-regulated at concentrations of 4-NP at 5 mg/L,but not at 0.5 mg/L concentration, when compared to the controlgroup (Fig. 4F). Deregulation of key glycolytic regulator pklr suggests re-duced glycogen content in the liver. Down-regulation of genes involvedin lipidmetabolism such as fatty acid desaturase 2 (fads2), acyl-CoenzymeA oxidase 3 (acox3), carnitine O-palmitoyltransferase II (cpt2) further cor-roborated with the transcriptomic analysis indicating that pathways as-sociated with biosynthesis of steroid, cholesterol, and PPAR signalingwere down-regulated in the late response phase. Moreover,down-regulation of nqo1 and hsdl2 that is involved in steroid biosynthe-sis pathway suggests that the production of steroid hormonesmay be af-fected and corroborated with previous findings of endocrine disruptionand perturbation of steroidal hormone signaling pathways in rodentmodels by 4-NP [9–11]. Besides carbohydrate and lipid metabolisms,our GSEA analysis also revealed up-regulation of pathway associatedwith amino acid metabolism particularly methionine, phenylalanine, ty-rosine, tryptophan, glycine, serine and threonine (Fig. 1). Elevated levelsof methionine and aromatic amino acids (phenylalanine, tyrosine, tryp-tophan) are hallmarks of liver disease [42] and deregulated amino acidmetabolism has been an indicator of liver pathological conditions suchas cirrhosis [43].

3.6. Other genes and pathways deregulated by 4-NP suggest liver injury

Finally, we have validated other deregulated genes such as selectinP (selp; leukocyte-endothelial cell adhesion molecule 3), thioredoxinreductase 1 (txrnd1), ribosomal protein L19 (rpl19), chloride intracellularchannel 4 (clic4), protein phosphatase 2, catalytic subunit, beta isoform(ppp2cb), activating transcription factor 3 (atf3), mitogen-activated pro-tein kinase 3 (mapk3), glutamic-oxaloacetic transaminase 1 (got1), andalcohol dehydrogenase (adh8a) which are associated with several otherpathways (Fig. 5A and B). Among other pathways, GSEA also indicatedthat protein synthesis and immune-related pathways are deregulatedby 4-NP (Fig. 1). The early down-regulation of pathways associatedwith t-RNA synthetases and aminoacyl t-RNA biosynthesis at the 24 htreatment suggests a direct acute response to 4-NP. A rapid acute de-crease in levels and activities of aminoacyl t-RNA synthetases andt-RNA had been documented in rat liver administered with aflatoxinB1 [44,45] which is also known to uncouple oxidative phosphorylation[46]. Moreover, it has been reported that a high speed supernatantobtained from ischemic livers contained reduce t-RNAs suggestingthat oxidative stress injury may play a role in regulating t-RNA levels[47]. However, the direct mechanistic link between the uncoupling ofoxidative phosphorylation and oxidative stress induced liver injurywith the deregulation of t-RNA synthetases and aminoacyl t-RNAbiosynthesis is not known and therefore warrants further study. Never-theless, a recent study has identified ‘aminoacyl t-RNA biosynthesis’pathway as among the top significant hepatotoxic-specific pathwaysderegulated in drug-induced liver injury [48]. The up-regulation of ribo-somal proteins at late responsive stage may be associated with the in-creased synthesis of ribosomal proteins and ribosomal RNA after liverinjury by partial hepatectomy [49] or chemical-induced inflammation[50] which was attributed to liver regeneration. Recently, the increasedsynthesis of ribosomes has been demonstrated to play a role in modu-lating MDM2–p53 signaling [51], presumably resulting in lesser avail-able ribosomal proteins binding to inhibit MDM2 from attenuatingp53 signaling [52]. Our data indicate that both mdm2 and p53 wereup-regulated in the liver of fish treated with 5 and 7 mg/L 4-NP(Fig. 2) suggesting that attenuation of p53 signaling has been activatedto prevent massive cell death and is possibly important for liver regen-eration at the late responsive stage. The deregulation of severalimmune-related pathways such as nuclear factor activated T-cells

(NFAT), nuclear factor kappa B-cells (NFKB), fibrinolysis, interleukin(IL) 6 and 7, further suggested acute liver inflammation. The analysiscorroborated with Braunbeck et al. [20] who reported infiltration oflymphocytes and macrophages in the livers of both male and femalezebrafish treated with 1 and 5 mg/L of 4-NP. Moreover, it has alsobeen documented that chronic administration of mononitrophenolscan cause hepatitis in mammals [53]. Furthermore, NFKB pathway,known to regulate cell survival/death, is up-regulated by ROS and cellu-lar stress [54] and has been found to be activated in virtually everychronic liver disease [55] suggesting similarity between mammals andzebrafish with regards to the role of NFKB pathway and liver damage.

3.7. An integrated model of 4-NP-induced toxicity mechanism in thezebrafish liver

An integrated model with plausible mechanistic insights into4-NP-induced acute hepatotoxicity in adult male zebrafish inferredfrom transcriptome analysis, histological analyses, and targeted genevalidation together with other related published findings is summa-rized in Fig. 6. Upon the entrance of 4-NP into the liver cells, it causesuncoupling of mitochondrial oxidative phosphorylation preventingATP synthesis hence resulting in the depletion of intracellular ATP. Asa homeostatic feedback to copewith the energy demand, both carbohy-drate and lipid content are used rapidly to generate ATP, and when thesubstrates are depleted, glycolytic feeder pathway and mitochondrialfatty acid beta-oxidation are rapidly down-regulated. This sets the cel-lular energetic environment into a catabolic, growth suppressing statewith the increased expression of tumor suppressor genes Rb and p53to arrest cell-cycle. Expression of genes involved in oxidative phosphor-ylation and electron transport chain are up-regulated to increase ATPproduction although the ATP needs would not be met due to theuncoupling effect of 4-NP and this becomes a vicious cycle. This will fur-ther increasemitochondrial respiration, increase production of reactiveoxygen species (ROS), deplete antioxidant glutathione rapidly and,when overwhelmed, will result in oxidative damage. In turn, oxidativestress and damage will activate p53-promoting Bax and Caspase-8 ac-tivities, culminating with apoptosis and increase in proteasome degra-dation. Consequently, cellular stress and cell death further affectedcellular adhesion, cell morphology and tissue integrity (Fig. 1). Theacute phase response also triggered inflammation in the liver with re-lease of cytokines and infiltration of leukocytes. Concomitant deregula-tion of glucose, lipid/steroidal and amino acid metabolisms cancollectively lead to hepatic metabolic disorders. Overall, our findings re-vealed an integrated action of 4-NP-induced molecular events and themechanistic routes leading to disruption of liver function and injury.

3.8. Limitation and strength of present study

This study consisted of three 4-NP exposure experiments, wherethe first exposure experiment was for the general acute toxicity test,the second for the microarray analysis, and the third for histologicaland real-time PCR validation analyses. These experiments were singleand independent as each exposure experiment was performed onlyonce, but with different batches of fish divided into treated and con-trol groups, each of which had several biological replicates. Sincethese experiments involved nominal exposures of 4-nitrophenol, theactual amount of 4-NP in the water was not determined. Thus, the ab-sence of exposure replicates within each experiment and the lack ofchemistry verification of 4-NP raises the concern that variation in ex-posure conditions between experiments were not accounted for.However, the exposure experiments were straight forward and theconditions were well controlled to minimize effects of variation bycomparison with respective control groups in each experiment. Thecomparable survival rates and outcome of each independent exposureexperiments with biological replicates and corresponding controlgroups, as well as multiple platform analyses by microarray, real-

0.5 mg/L 5 mg/L 7 mg/L

Log

2 F

old

-Ch

ang

eV

ersu

s C

on

tro

l

Concentration of 4-NP

selp

**

**

0.5 mg/L 5 mg/L 7 mg/L

txnrd1

**

**

0.5 mg/L 5 mg/L 7 mg/L

rpl19

**

**

0.5 mg/L 5 mg/L 7 mg/L

clic4

**

**

0.5 mg/L 5 mg/L 7 mg/L

ppp2cb**

***

0

1

2

3

4

5

6

0.5 mg/L 5 mg/L 7 mg/L

atf3

*

*

**

real-time PCR data

microarray data

* = p value ≤ 0.1** = p value ≤ 0.05

0.5 mg/l 5 mg/l 7 mg/L

mapk3

**

*

**

0

0.2

0.4

0.6

0.8

1

0.5 mg/l 5 mg/l 7 mg/L

got1

**

-3

-2.5

-2

-1.5

-1

-0.5

0

0.5 mg/L 5 mg/L 7 mg/L

adh8a

***

A

B

00.5

11.5

22.5

33.5

00.5

11.5

22.5

33.5

-0.5

0

0.5

1

1.5

2

2.5

0

0.5

1

1.5

2

0

0.2

0.4

0.6

0.8

1

1.2

1.4

00.20.40.60.8

11.21.41.61.8

Fig. 5. Validation of several genes that are deregulated by 4-nitrophenol (4-NP) in zebrafish liver. A. Gene expression profiles for deregulated genes associated with several otherpathways. Color scale represents log2 fold-change above the corresponding match time-point control group where red, green and black cells represent up-regulation,down-regulation and no change in gene expression, respectively. Quantitative real-time PCR for the deregulated genes (*p-value ≤ 0.1; **p-value ≤ 0.05; n = 5 replicates) com-pared to microarray data.

4787S.H. Lam et al. / Biochimica et Biophysica Acta 1830 (2013) 4778–4789

time PCR and histological approaches, corroborated and substantiatedthat the findings are valid and reproducible. Furthermore, other pub-lished studies (e.g. ultra-structural study in zebrafish [20] and severalreports that had shown 4-NP induced oxidative phosphorylation

uncoupling [33,34], metabolic disruption [40,41], liver inflammationand injury [53]) also corroborated and confirmed the validity of ourfindings. Therefore, the present study which involved three single in-dependent exposure experiments that were assayed or validated with

Fig. 6. An integrated model with plausible mechanistic insights into 4-nitrophenol (4-NP) induced acute hepatotoxicity in adult male zebrafish (refer to manuscript text for detaildescription).

4788 S.H. Lam et al. / Biochimica et Biophysica Acta 1830 (2013) 4778–4789

different methods should provide robust evidence for 4-NP-inducedhepatotoxicity in zebrafish.

4. Conclusion

This is the first transcriptomic study on 4-NP toxicity in the liver andit has provided insights and broader understanding of the molecularevents leading to acute hepatotoxicity induced by 4-NP. Our study pro-vided us a plausible model to depict these molecular events of 4-NP in-duced hepatotoxicity and could serve as a reference for investigatinghepatotoxicity induced by other forms of nitrophenol derivatives. Theincreasing use of nitrophenol derivatives in slimming drug formulation[41] or uncoupling oxidative phosphorylation for obesity treatment [56]could lead to similar 4-NP induced hepatotoxicity. In view of the meta-bolic disruption caused by 4-NP, a future study at the metabolomicslevel would provide a more integrated system perspective of thetranscriptome and metabolome. The up-regulation of pathways associ-ated with NFKB, methionine and aromatic amino acid metabolism in4-NP induced hepatoxicity in zebrafish are hallmark similarities sharedwith chronic liver diseases reported in human [42,55]. This further un-derscores the potential use for zebrafish as an alternative model to in-vestigate liver injury or diseases.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbagen.2013.06.008.

Acknowledgement

This work was supported by the Singapore National Research Foun-dation under its Environmental &Water Technologies Strategic ResearchProgramme and administered by the Environment & Water IndustryProgramme Office (EWI) of the PUB, grant number R-154-000-328-272.

References

[1] United States Environmental Protection Agency (US EPA), Ambient Water QualityCriteria for Nitrophenols, Office of Water Regulations and Standards, Criteria andStandards Division, Washington, DC, 1980. , (EPA 440/5-80-063).

[2] Agency for Toxic Substances Disease Registry (ATSDR), Toxicological Profile forNitrophenols: 2-Nitrophenol and 4-Nitrophenol, U.S. Department of Health andHuman Services, Public Health Service, Atlanta, GA, 1992. (104 pp.).

[3] G. Eichenbaum, M. Johnson, D. Kirkland, P. O'Neill, S. Stellar, J. Bielawne, R. DeWire,D. Areia, S. Bryant, S. Weiner, D. Desai-Krieger, P. Guzzie-Peck, D.C. Evans, A. Tonelli,Assessment of the genotoxic and carcinogenic risks of p-nitrophenol when it ispresent as an impurity in a drug product, Regul. Toxicol. Pharmacol. 55 (2009)33–42.

[4] R.I. Buccafusco, S.J. Ells, G.A. Le Blanc, Acute toxicity of priority pollutants to blue-gill (Lepomis macrochirus), Bull. Environ. Contam. Toxicol. 26 (1981) 446–452.

[5] D.J. Call, L.T. Brooke, P.Y. Lu, Uptake, elimination, and metabolism of three phenolsby fathead minnows, Arch. Environ. Contam. Toxicol. 9 (1980) 699–714.

[6] D.K. Richards, W.K. Shiehs, Biological fate of organic priority pollutants in theaquatic environment, Water Res. 20 (1986) 1077–1090.

[7] R.C. Loehr, R. Krishnamoorthy, Terrestrial bioaccumulation potential of phenoliccompounds, Hazard. Waste Hazard. Mater. 5 (1988) 109–119.

[8] D.O. Hryhorczuk, M. Moomey, A. Burton, K. Runkle, E. Chen, T. Saxer, J. Slightom, J.Dimos, K. McCann, D. Barr, Urinary p-nitrophenol as a biomarker of householdexposure to methyl parathion, Environ. Health Perspect. 110 (2002) 1041–1046.

[9] C. Furuta, A.K. Suzuki, S. Taneda, K. Kamata, H. Hayashi, Y. Mori, C. Li, G.Watanabe, K. Taya, Estrogenic activities of nitrophenols in diesel exhaust parti-cles, Biol. Reprod. 70 (2004) 1527–1533.

[10] C. Li, S. Taneda, A.K. Suzuki, C. Furuta, G. Watanabe, K. Taya, Estrogenic andanti-androgenic activities of 4-nitrophenol in diesel exhaust particles, Toxicol.Appl. Pharmacol. 217 (2006) 1–6.

[11] X. Li, C. Li, A.K. Suzuki, S. Taneda, G. Watanabe, K. Taya, 4-Nitrophenol isolatedfrom diesel exhaust particles disrupts regulation of reproductive hormones inimmature male rats, Endocrine 36 (2009) 98–102.

[12] T.E. Archer, Dissipation of parathion and related compounds from field-sprayedlettuce, J. Agric. Food Chem. 23 (1975) 858–860.

[13] J. Camanzo, C.P. Rice, D.J. Jude, R. Rossmann, Organic priority pollutants in near-shore fish from 14 Lake Michigan tributaries and embayments, 1983, J. GreatLakes Res. 13 (1987) 296–309.

[14] A.J. Hill, H. Teraoka, W. Heideman, R.E. Peterson, Zebrafish as a model vertebratefor investigating chemical toxicity, Toxicol. Sci. 86 (2005) 6–19.

[15] S.H. Lam, Y.L. Wu, V.B. Vega, L.D. Miller, J. Spitsbergen, Y. Tong, H. Zhan, K.R.Govindarajan, S. Lee, S. Mathavan, K.R. Murthy, D.R. Buhler, E.T. Liu, Z. Gong,Conservation of gene expression signatures between zebrafish and human livertumors and tumor progression, Nat. Biotechnol. 24 (2006) 73–75.

[16] S.H. Lam, S. Mathavan, Y. Tong, H. Li, R.K. Karuturi, Y. Wu, V.B. Vega, E.T. Liu, Z.Gong, Zebrafish whole-adult-organism chemogenomics for large-scale predictiveand discovery chemical biology, PLos Genet. 4 (2008) e1000121.

[17] A. Flemming, Zebrafish as an alternative model organism for disease modellingand drug discovery: implications for the 3Rs, NC3Rs 10 (2007) 1–7.

[18] S. Scholz, S. Fischer, U. Gündel, E. Küster, T. Luckenbach, D. Voelker, The zebrafishembryo model in environmental risk assessment—applications beyond acutetoxicity testing, Environ. Sci. Pollut. Res. Int. 15 (2008) 394–404.

[19] N.S. Sipes, S. Padilla, T.B. Knudsen, Zebrafish: as an integrativemodel for twenty-firstcentury toxicity testing, Birth Defects Res C Embryo Today 93 (2011) 256–267.

4789S.H. Lam et al. / Biochimica et Biophysica Acta 1830 (2013) 4778–4789

[20] T. Braunbeck, V. Storch, H. Bresch, Sex-specific reaction of liver ultrastructure inzebra fish (Brachydanio rerio) after prolonged sublethal exposure to 4-nitrophenol,Aquat. Toxicol. 14 (1989) 185–202.

[21] H. Bresch, Investigation of the long-term action of xenobiotics on fish with specialregard to reproduction, Ecotoxicol. Environ. Saf. 6 (1982) 102–112.

[22] T. Mizuma, M. Hayashi, S. Awazu, p-Nitrophenol sulfation in rat liver cytosol:multiple forms and substrate inhibition of aryl sulfotransferase, J. Pharm. Dyn.(6) (1983) 851–858.

[23] T. Kasokat, R. Nagel, K. Urich, The metabolism of phenol and substituted phenolsin zebrafish, Xenobiotica 17 (1987) 1215–1221.

[24] L. Yang, J.R. Kemadjou, C. Zinsmeister, M. Bauer, J. Legradi, F. Müller, M. Pankratz,J. Jäkel, U. Strähle, Transcriptional profiling reveals barcode-like toxicogenomicresponses in the zebrafish embryo, Genome Biol. 8 (2007) R227.

[25] S.H. Lam, C.L. Winata, Y. Tong, S. Korzh, W.S. Lim, V. Korzh, J. Spitsbergen, S.Mathavan, L.D. Miller, E.T. Liu, Z. Gong, Transcriptome kinetics of arsenic-inducedadaptive response in zebrafish liver, Physiol. Genomics 27 (2006) 351–361.

[26] H. Sukardi, X. Zhang, E.Y. Lui, C.Y. Ung, S. Mathavan, Z. Gong, S.H. Lam, Liver X recep-tor agonist T0901317 induced liver perturbation in zebrafish: histological, gene setenrichment and expression analyses, Biochim. Biophys. Acta 1820 (2012) 33–43.

[27] S.H. Lam, S.G. Lee, C.Y. Lin, J.S. Thomsen, F.Y. Fu, K.R. Murthy, H. Li, K.R.Govindarajan, L.C. Nick, G. Bourque, Z. Gong, T. Lufkin, E.T. Liu, S. Mathavan,Molecular conservation of estrogen-response associated with cell cycle regulation,hormonal carcinogenesis and cancer in zebrafish and human cancer cell lines, BMCMed. Genet. 16 (2011) 41.

[28] F.A. Tilton, S.C. Tilton, T.K. Bammler, R.P. Beyer, P.L. Stapleton, N.L. Scholz, E.P.Gallagher, Transcriptional impact of organophosphate and metal mixtures onolfaction: copper dominates the chlorpyrifos-induced response in adult zebrafish,Aquat. Toxicol. 102 (2011) 205–215.

[29] C.J. Mattingly, T.H. Hampton, K.M. Brothers, N.E. Griffin, A. Planchart, Perturbationof defense pathways by low-dose arsenic exposure in zebrafish embryos, Environ.Health Perspect. 117 (2009) 981–987.

[30] S.H. Lam, S. Mathavan, Z. Gong, Zebrafish spotted-microarray for genome-wideexpression profiling experiments. Part I: array printing and hybridization,Methods Mol. Biol. 546 (2009) 175–195.

[31] S.H. Lam, M.K.R. Krishna, Z. Gong, Zebrafish spotted-microarray for genome-wideexpression profiling experiments: data acquisition and analysis, Methods Mol.Biol. 546 (2009) 197–226.

[32] A. Subramanian, P. Tamayo, V.K. Mootha, S. Mukherjee, B.L. Ebert, M.A. Gillette, A.Paulovich, S.L. Pomeroy, T.R. Golub, E.S. Lander, J.P. Mesirov, Gene set enrichmentanalysis: a knowledge-based approach for interpreting genome-wide expressionprofiles, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 15545–15550.

[33] H. Terada, Uncouplers of oxidative phosphorylation, Environ. Health Perspect.87 (1990) 213–218.

[34] Y.H. Han, S.Z. Kim, S.H. Kim,W.H. Park, 2,4-Dinitrophenol induces apoptosis in As4.1juxtaglomerular cells through rapid depletion of GSH, Cell Biol. Int. 32 (2008)1536–1545.

[35] C. Vidau, R.A. González-Polo, M. Niso-Santano, R. Gómez-Sánchez, J.M. Bravo-SanPedro, E. Pizarro-Estrella, R. Blasco, J.L. Brunet, L.P. Belzunces, J.M. Fuentes,Fipronil is a powerful uncoupler of oxidative phosphorylation that triggers apo-ptosis in human neuronal cell line SHSY5Y, Neurotoxicology 32 (2011) 935–943.

[36] Y.H. Han, S.W. Kim, S.H. Kim, S.Z. Kim, W.H. Park, 2,4-Dinitrophenol induces G1phase arrest and apoptosis in human pulmonary adenocarcinoma Calu-6 cells,Toxicol. in Vitro 22 (2008) 659–670.

[37] J. Nishimura, Y. Dewa, M. Muguruma, Y. Kuroiwa, H. Yasuno, T. Shima, M. Jin, M.Takahashi, T. Umemura, K. Mitsumori, Effect of fenofibrate on oxidative DNAdamage and on gene expression related to cell proliferation and apoptosis inrats, Toxicol. Sci. 97 (2007) 44–54.

[38] T.G. Cotter, Apoptosis and cancer: the genesis of a research field, Nat. Rev. Cancer9 (2009) 501–507.

[39] D.R. Green, G. Kroemer, Cytoplasmic functions of the tumour suppressor p53,Nature 458 (2009) 1127–1130.

[40] S.E. Dominguez, J.L. Menkel, A. Fairbrother, B.A. Williams, R.W. Tanner, The effectof 2,4-dinitrophenol on the metabolic rate of bobwhite quail, Toxicol. Appl.Pharmacol. 123 (1993) 226–233.

[41] J. Grundlingh, P.I. Dargan, M. El-Zanfaly, D.M. Wood, 2,4-Dinitrophenol (DNP): aweight loss agent with significant acute toxicity and risk of death, J. Med. 7 (2011)205–212.

[42] M. Charlton, Branched-chain amino acid enriched supplements as therapy forliver disease, J. Nutr. 136 (2006) 295S–298S.

[43] L. Fontana, E. Moreira, M.I. Torres, M.I. Fernández, A. Ríos, F. Sánchez de Medina, A.Gil, Serum amino acid changes in rats with thioacetamide-induced liver cirrhosis,Toxicology 106 (1996) 197–206.

[44] G. Wagner, A.M. Unterreiner, Inhibition of rat liver aminoacyl-tRNA synthetasesin vitro after acute and chronic aflatoxin B1 administration in vivo, Chem. Biol.Interact. 37 (1981) 233–244.

[45] G. Wagner, A.M. Unterreiner, Synthesis of transfer RNA in rat liver after acuteand chronic aflatoxin B1 administration, Chem. Biol. Interact. 41 (1982)353–360.

[46] M. Ramachandra Pai, N. Jayanthi Bai, T.A. Venkitasubramanian, Effect of aflatoxins onoxidative phosphorylation by rat liver mitochondria, Chem. Biol. Interact. 10 (1975)123–131.

[47] F. Cajone, L. Schiaffonati, A. Bernelli-Zazzera, Protein synthesis in liver injury.Soluble factors of protein synthesis in the cytosol from ischemic rat liver, Lab.Invest. 34 (1976) 387–393.

[48] C.W. Chang, F.A. Beland, W.M. Hines, J.C. Fuscoe, T. Han, J.J. Chen, Identificationand categorization of liver toxicity markers induced by a related pair of drugs,Int. J. Mol. Sci. 12 (2011) 4609–4624.

[49] S. Chaudhuri, I. Lieberman, Control of ribosome synthesis in normal andregenerating liver, J. Biol. Chem. 243 (1968) 29–33.

[50] T.C. Vary, S.R. Kimball, Regulation of hepatic protein synthesis in chronic inflam-mation and sepsis, Am. J. Physiol. 262 (1992) C445–C452.

[51] G. Donati, S. Bertoni, E. Brighenti, M. Vici, D. Treré, S. Volarevic, L. Montanaro, M.Derenzini, The balance between rRNA and ribosomal protein synthesis up- anddownregulates the tumour suppressor p53 in mammalian cells, Oncogene 30(2011) 3274–3288.

[52] C. Deisenroth, Y. Zhang, The ribosomal protein–Mdm2–p53 pathway and energymetabolism: bridging the gap between feast and famine, Genes Cancer 2 (2011)392–403.

[53] National Research Council, Drinking Water & Health, Volume 4National AcademyPress, Washington, DC, 1981. 241.

[54] A.R. Brasier, The NF-kappaB regulatory network, Cardiovasc. Toxicol. 6 (2006)111–130.

[55] T. Luedde, R.F. Schwabe, NF-κB in the liver—linking injury,fibrosis and hepatocellularcarcinoma, Nat. Rev. Gastroenterol. Hepatol. 8 (2011) 108–118.

[56] J.A. Harper, K. Dickinson, M.D. Brand, Mitochondrial uncoupling as a target fordrug development for the treatment of obesity, Obes. Rev. 2 (2001) 255–265.