Abstract

   Triphenyl phosphate is a high production volume organophosphate flame
   retardant that has been detected in multiple environmental media at
   increasing concentrations. The environmental and health risks of
   triphenyl phosphate have drawn attention because of the multiplex
   toxicity of this chemical compound. However, few studies have paid
   close attention to the impacts of triphenyl phosphate on liver
   metabolism. We investigated hepatic histopathological, metabolomic and
   transcriptomic responses of zebrafish after exposure to 0.050 mg/L and
   0.300 mg/L triphenyl phosphate for 7 days. Metabolomic analysis
   revealed significant changes in the contents of glucose, UDP-glucose,
   lactate, succinate, fumarate, choline, acetylcarnitine, and several
   fatty acids. Transcriptomic analysis revealed that related pathways,
   such as the glycosphingolipid biosynthesis, PPAR signaling pathway and
   fatty acid elongation, were significantly affected. These results
   suggest that triphenyl phosphate exposure markedly disturbs hepatic
   carbohydrate and lipid metabolism in zebrafish. Moreover, DNA
   replication, the cell cycle, and non-homologous end-joining and base
   excision repair were strongly affected, thus indicating that triphenyl
   phosphate hinders the DNA damage repair system in zebrafish liver
   cells. The present study provides a systematic analysis of the
   triphenyl phosphate-induced toxic effects in zebrafish liver and
   demonstrates that low concentrations of triphenyl phosphate affect
   normal metabolism and cell cycle.
     __________________________________________________________________

   Flame retardants are widely used industrial chemicals that are added to
   many products to prevent them from burning and to reduce the risk of
   fires. Since 2004, polybrominated diphenyl ethers (PBDEs), such as
   octaBDE and pentaBDE, which once were used worldwide and which shared a
   large proportion of the worldwide flame retardant market, have been
   gradually phased out of the market because of their environmental
   ubiquity and potential adverse health effects. To satisfy the market
   demand, production of alternative flame retardants, including
   organophosphate flame retardants (OPFRs), has been increased[34]^1.
   Additionally, OPFRs are physically added but not chemically bonded to
   materials, which can lead to easy release of such compounds into the
   environment. Due to their increasing use, OPFRs have been detected in
   multiple environmental media, including indoor and outdoor
   air[35]^2,[36]^3, surface water[37]^4, groundwater[38]^5 and drinking
   water[39]^6, and in aquatic biota and human milk[40]^7. Under this
   circumstance, the environmental and human health risk assessment of
   OPFRs is extraordinarily exigent.

   Triphenyl phosphate (TPhP, also called TPP) is a high production volume
   OPFR used in unsaturated polyester resins, in PVC (polyvinyl chloride),
   and in commercial mixtures, such as FM550 from Chemtura (Firemaster®
   550 flame retardant, a Chemtura proprietary mixture of brominated and
   non-halogen flame retardants) and AC073 from Supresta (commercial flame
   retardant mixture of proprietary aryl phosphates and triphenyl
   phosphate)[41]^8. Due to the increasing use of this compound, TPhP has
   been the most frequently detected OPFR in both environmental media and
   biota[42]^9. According to Zheng et al. and Stapleton et
   al.[43]^10,[44]^11, TPhP is the main OPFR detected in house dust, with
   levels of 9810 and 7360 ng/g in China and the USA, respectively. In
   surface water, the maximum reported level of TPhP was 14000 ng/L in
   influent samples in Norway in 2007[45]^12. Although TPhP is easily
   biodegraded as a non-halogen organophosphate ester and has not been
   considered to be persistent or bioaccumulative[46]^13,[47]^14, the
   constant and abundant use and release of this compound may also cause
   sustained exposure to humans and wildlife. Moreover, TPhP has been
   reported to show multiplex toxicities, including neurotoxicity,
   developmental toxicity and endocrine disrupting ability, in
   organisms[48]^15,[49]^16,[50]^17. Various indications have suggested
   that TPhP poses risks to the environment and to health. However,
   toxicology data are still insufficient to assess all aspects of the
   environmental and health risks of TPhP. The liver is a vital organ that
   plays a major role in metabolism and has numerous functions in
   vertebrates and is sensitive to xenobiotics. The hepatotoxicity of TPhP
   remains unknown, although many studies have investigated TPhP
   metabolism in the liver[51]^18,[52]^19,[53]^20.

   Omics, including transcriptomics, proteomics and metabolomics, have
   been well developed and widely introduced in toxicology studies in the
   past 10 years[54]^21. Compared to traditional toxicological endpoints,
   omics approaches can provide high-throughput data of numerous
   biomolecules and reflect integrated global responses within an
   organism[55]^22. Furthermore, the integration of these omics streams (2
   or 3 of them) can provide a deep mechanistic understanding of how
   compounds perturb normal biological processes and activate defense
   mechanisms[56]^23. Deng et al. have provided a comprehensive view of
   the toxic effects of haloacetamides, which are disinfection
   by-products, by combining metabolomics methods and traditional
   oxidative stress endpoints[57]^24. Omics approaches have also been
   introduced to investigate the toxic effects of OPFRs. Scanlan et al.,
   by using metabolomics and transcriptomics tools, have found that FM550
   impairs nutrient utilization or uptake in Daphnia magna[58]^25. Alam et
   al. have investigated the metabolic influences of tributyl phosphate
   (TBP) and TPhP in a nuclear magnetic resonance (NMR)-based metabolomics
   study of rat urine and have found that TPhP disturbs cellular energy
   metabolism and creatine synthesis in rats[59]^26. These studies have
   revealed that TPhP or TPhP-containing material impairs nutrient
   utilization and cellular energy metabolism in different animals at an
   individual level. Omics is a good methodology to investigate the
   hepatotoxicity of TPhP.

   Zebrafish provide several experimental advantages, such as their low
   cost, ease of raising large numbers of animals, rapid liver
   development, and similar molecular and cellular processes to those of
   humans. Hence, they have been widely employed as a model organism in
   human disease research and in aquatic ecosystem health studies. The
   Zebrafish has also been used as a model organism to study drug-induced
   liver injury[60]^27. To broaden understanding of the hepatotoxicity of
   TPhP, we performed transcriptomics and metabolomics analyses of adult
   zebrafish liver after 7 days of exposure to two different
   concentrations of TPhP. Histopathological liver changes, blood glucose
   levels and blood lipid levels were also investigated to provide a
   systematic understanding of adverse effects induced by TPhP in
   zebrafish.

Methods

   All experimental protocols were approved by the School of the
   Environment, Nanjing University, and the methods were carried out in
   accordance with the approved guidelines.

Fish Care and Exposure

   Five-month-old adult zebrafish (Danio rerio, AB strain; sex ratio, 1:1)
   were purchased from the Institute of Hydrobiology, Chinese Academy of
   Sciences (Wuhan, China). Before the exposure period, the fish were
   cultured for more than 1 week in tap water (aerated and basked for two
   days for dechlorination before use, with oxygen saturation exceeding
   80%) and allowed to acclimate to the following conditions[61]^28:
   12h:12h light/dark regime, 26 ± 1 °C, pH 7.4–8.1 and a daily feeding
   (9:00 am) with dry flake food at a maximum of 2% body weight.

   Before exposure experiments were initiated, the 96h-LC[50] of TPhP to
   adult zebrafish was determined following the guidance of OECD No.
   203[62]^29 (details are shown in [63]Supplementary information, Section
   S2). The 96h-LC[50] of TPhP to adult zebrafish was found to be
   1.026 mg/L; thus, the experimental concentrations of TPhP were set at
   0.050 and 0.300 mg/L. The concentration 0.050 mg/L was close to the
   maximum concentration of TPhP that has been reported in surface water.
   In total, 270 fish were exposed to related concentrations of TPhP
   (99+%, purchased from Sigma–Aldrich, USA) with a solvent control (0.1%
   dimethyl sulfoxide [DMSO], chromatographic grade, purchased from Merck,
   Germany) in 10 L glass tanks. Each tank contained 30 fish (15 males and
   15 females), with 3 parallel tanks per treatment group. The exposure
   experiments were carried out with a semi-static protocol. Half of the
   solution in each tank was renewed daily. Other conditions were the same
   as in the above-mentioned acclimation period. The stability of the TPhP
   concentrations was confirmed using solid phase extraction and
   high-performance liquid chromatography (details are shown in
   [64]Supplementary Information, Section S1). The exposure period lasted
   for 7 days. The fish were fed daily except on the sampling day. When
   the fish were sampled, they were first anesthetized in ice-cold water.

   The exposure experiment was performed twice (540 fish were used), once
   for metabolomic, transcriptomic and histopathological tests and once
   for blood and RT-qPCR (reverse transcription-polymerase chain reaction)
   tests.

Histopathological Observation

   After a one-week exposure period, the livers of zebrafish (10 fish for
   each concentration group) were dissected and fixed in 10% formalin at
   4 °C for 24h. Subsequently, the fixed liver tissues were dehydrated in
   gradient ethanol, hyalinized in xylene, and embedded in paraffin wax at
   56 °C. Then, the paraffin blocks were sectioned at 4-μm thickness. The
   sections were collected on glass slides and stained with hematoxylin
   and eosin (H&E) using an H&E Staining Kit (Nanjing Jiancheng
   Bioengineering Institute, Nanjing, China) following the manufacturer’s
   recommendations (standard H&E staining protocol). Histologic lesions
   were observed using an optical microscope equipped with a digital
   camera.

Metabolomic Analysis

   After a one-week exposure period, the zebrafish were anesthetized on
   ice, and their livers were dissected and used for extraction
   immediately. The extraction was carried out using a
   methanol/chloroform/water system at 4 °C as described by C.Y. Lin et
   al. with minor modifications[65]^30. In detail, approximately 50 mg
   frozen liver tissue (dissected from 12 fish) was homogenized using an
   electric tissue homogenizer in 400 μL methanol (chromatographic grade,
   Merck, Germany) and 85 μL ultrapure water (UPW) in a glass tube. The
   homogenizer was washed with 75 μL UPW; this fluid was also collected in
   the tube. Then, another 200 μL UPW and 400 μL chloroform
   (chromatographic grade, Merck, Germany) were added after the tube was
   vortexed for 60 seconds. Next, the tube was centrifuged at 2000 g at
   4 °C for 5 min, and 500 μL methanol/water (upper) and 350 μL chloroform
   (lower) were separately transferred to new glass tubes followed by
   lyophilization or vacuum drying, respectively. Polar and nonpolar
   residues were redissolved using 400 μL UPW + 100 μL D[2]O (99.9 atom%
   D, containing 0.05 wt% TSP, Sigma–Aldrich, USA) or 500 μL chloroform-d
   (99.8 atom% D, containing 0.03% v/v TMS, Sigma–Aldrich, USA) separately
   before NMR analysis. Six parallel samples (12 fish were sacrificed for
   each sample) were prepared for each treatment/control group.

   ^1H NMR spectra of all samples were acquired using a Bruker AV600 MHz
   spectrometer (Bruker Co., Germany) operating at 600.17 MHz and 298.5 K.
   Spectra of the aqueous samples were acquired using a
   Carr-Purcell-Meiboom-Gill, 1D acquisition with presaturation (CPMGPR1D)
   pulse program with 32 scans (FIDs) collected into 65536 data points
   with Fourier transformation. For the chloroform phase samples, spectra
   were acquired using a zg pulse program with 16 scans (FIDs) collected
   into 32768 data points with Fourier transformation.

   NMR spectra were processed (phase correction, baseline correction and
   ppm shift correction using trisilylpropionic acid [TSP] or
   tetramethylsilane [TMS]) and analyzed (GSD, Global Spectral
   Deconvolution) using MestReNova v6.1.0–6224 software. Metabolite
   resonances were identified using both Chenomx NMR Suite 8.02 and
   published literature[66]^31,[67]^32,[68]^33. NMR spectra and assigned
   metabolites are presented in [69]Figure S1, [70]Figure S2 and [71]Table
   S1. Integrations of water resonance (4.70–5.15 ppm in spectra of
   aqueous samples) and methanol resonance (3.34–3.38 ppm in spectra of
   aqueous samples and 3.30–3.40 ppm in spectra of chloroform phase
   samples) were excluded. Following the removal of these regions, the
   metabolites were then normalized to the total integrated spectral area
   (0.65–9.37 ppm for aqueous samples and 0.60–7.00 ppm for chloroform
   phase samples). Then, the data set was log transformed and Pareto
   scaled (mean-centered and divided by the square root of the standard
   deviation of each variable) before statistical analysis. Significantly
   changed metabolites (SCMs) were identified by t-test and partial least
   squares-discriminant analysis (PLS-DA) using an online tool for
   metabolomics data analysis, MetaboAnalyst 3.0[72]^34.

Transcriptomic Analysis

   The livers of 4 zebrafish were dissected and used for total RNA
   extraction. Total RNA extraction was performed using TRIzol reagent
   (Life Technologies). Two parallel total RNA samples were prepared for
   each treatment group. Then, the transcriptomic analysis was conducted
   by BGI Tech. (Shenzhen, China) using BGI RNA-Seq (Quantification)
   analysis
   ([73]http://bgi-international.com/services/genomics/rna-seq-quantificat
   ion/#tab-id-1). In detail, the total RNA samples were first treated
   with DNase I to degrade any possible DNA contamination. Then, the mRNA
   was enriched by using oligo(dT) magnetic beads (for eukaryotes). The
   mRNA was mixed with the fragmentation buffer and fragmented into short
   fragments (approximately 200 bp). Then, the first strand of cDNA was
   synthesized by using random hexamer primers. Buffer, dNTPs, RNase H and
   DNA polymerase I were added to synthesize the second strand. The
   double-stranded cDNA was purified with magnetic beads. End repair and
   3′-end single nucleotide A (adenine) addition were then performed.
   Finally, sequencing adaptors were ligated to the fragments. The
   fragments were enriched by PCR amplification. During the quality
   control step, an Agilent 2100 Bioanalyzer and ABI StepOnePlus Real-Time
   PCR System were used to qualify and quantify the sample library. The
   library products were used for sequencing via an Illumina HiSeq^TM 2000
   or other sequencer when necessary. Then, the raw data were cleaned by
   removing low-quality reads as well as reads with adaptor sequences and
   reads containing unknown bases more than 10%. Clean reads were mapped
   to reference sequences and/or references. No more than 2 mismatches