Effect of sunlight exposure on the release of intentionally ... - Packtox

Food Chemistry 162 (2014) 63–71

Contents lists available at ScienceDirect

Food Chemistry journal homepage: www.elsevier.com/locate/foodchem

Effect of sunlight exposure on the release of intentionally and/or non-intentionally added substances from polyethylene terephthalate (PET) bottles into water: Chemical analysis and in vitro toxicity Cristina Bach a,c,⇑, Xavier Dauchy a, Isabelle Severin b, Jean-François Munoz a, Serge Etienne c, Marie-Christine Chagnon b a

ANSES, Nancy Laboratory for Hydrology, Water Chemistry Department, 40 rue Lionnois, 54000 Nancy, France Derttech ‘‘Packtox’’, Nutox team, AgroSupDijon Nord, 1 Esplanade Erasme, 21000 Dijon, France c Institute Jean Lamour, UMR 7198, Department SI2M, Ecole des Mines de Nancy, University of Lorraine, Parc de Saurupt, CS 14234, 54042 Nancy, France b

a r t i c l e

i n f o

Article history: Received 18 October 2013 Received in revised form 3 March 2014 Accepted 3 April 2014 Available online 13 April 2014 Keywords: PET-bottled waters Migration Sunlight NIAS Genotoxicity Endocrine disruption Aldehydes Antimony Chemical analysis

a b s t r a c t The effect of sunlight exposure on chemical migration into PET-bottled waters was investigated. Bottled waters were exposed to natural sunlight for 2, 6 and 10 days. Migration was dependent on the type of water. Formaldehyde, acetaldehyde and Sb migration increased with sunlight exposure in ultrapure water. In carbonated waters, carbon dioxide promoted migration and only formaldehyde increased slightly due to sunlight. Since no aldehydes were detected in non-carbonated waters, we conclude that sunlight exposure has no effect. Concerning Sb, its migration levels were higher in carbonated waters. No unpredictable NIAS were identified in PET-bottled water extracts. Cyto-genotoxicity (Ames and micronucleus assays) and potential endocrine disruption effects (transcriptional-reporter gene assays) were checked in bottled water extracts using bacteria (Salmonella typhimurium) and human cell lines (HepG2 and MDA-MB453-kb2). PET-bottled water extracts did not induce any toxic effects (cyto-genotoxicity, estrogenic or anti-androgenic activity) in vitro at relevant consumer-exposure levels. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction PET is a polymer with very few additives used for its manufacture; plasticisers and antioxidants are not necessary to produce PET bottles and colorants are added only in small quantities. Acetaldehyde scavengers are used to minimise the formation of acetaldehyde during the melt-process. Also, titanium nitride nanoparticles can be incorporated into PET bottle grade (EFSA, 2012). Even if starting substances and additives are strictly regulated by EU Regulation No. 10/2011, several substances known as NIAS (non-intentionally added substances) can be found in the final plastic material, due to complex formulations of polymers, processes and storage (e.g. impurities, degradation products, breakdown products, etc.) (EU, 2011). These substances can also ⇑ Corresponding author at: ANSES, Nancy Laboratory for Hydrology, Water Chemistry Department, 40 rue Lionnois, 54000 Nancy, France. E-mail address: [email protected] (C. Bach). http://dx.doi.org/10.1016/j.foodchem.2014.04.020 0308-8146/Ó 2014 Elsevier Ltd. All rights reserved.

migrate into foodstuffs. In addition, physical stress applied to a plastic material can modify the structure of its chemical ingredients (with no toxicological concern) and generate NIAS which may have potential estrogenic and/or anti-androgenic activities (Yang, Yaniger, Jordan, Klein, & Bittner, 2011). According to EU Regulation No. 1935/2004 (EU, 2004), ‘‘food contact materials must not transfer their constituents to food in quantities which could endanger human health’’. Furthermore, EU Regulation No. 10/2011 (EU, 2011) states that ‘‘the risk assessment of a substance should cover the substance itself, relevant impurities and foreseeable reaction and degradation products in the intended use’’. A polymer exposed to sunlight may undergo photochemical aging, which is the case with PET, which absorbs sunlight at a wavelength (k) range located at the end of UV light spectra (300 nm 6 k 6 330 nm). Exposing PET bottles to sunlight, which also increases the water’s temperature, raises questions about the formation of by-products and their migration into water, as a possible source of health hazards for the consumers. Few studies

64

C. Bach et al. / Food Chemistry 162 (2014) 63–71

are available on photoproducts released into PET-bottled water exposed to sunlight, and when they are available, case toxicities have not always been assessed in parallel. The presence of aldehydes, phthalates, bisphenol A and 4-nonylphenol in PETbottled waters following sunlight exposure were observed, but with a wide range of concentrations and storage times which makes data comparison difficult. Furthermore, compounds were not systematically presents or their levels were not statistically different in the water samples before and after exposure to sunlight (see review in Bach, Dauchy, Chagnon, and Etienne (2012)). In vitro genotoxicity using plant and eukaryote cell models has been observed in PET-bottled waters exposed to sunlight, but the chemicals responsible for these effects were not identified (Corneanu, Corneanu, Jurescu, & Toptan, 2010; Ubomba-Jaswa, Fernandez-Ibanez, & McGuigan, 2010). PET containers are sometimes exposed to direct sunlight due to poor storage conditions in retail stores and consumers’ homes, which causes degradation of the polymer through thermomechanical and thermo-oxidative processes, generating NIAS which can migrate into the bottled water (Bach et al., 2012). In fact, in a previous study we demonstrated that high temperatures increase migration of formaldehyde, acetaldehyde and Sb into PET-bottled waters. In addition, we identified two NIAS (2,4-ditert-butylphenol and bis(2-hydroxyethyl)terephthalate) in bottled waters. However, bottled water extracts were not found to be cyto/genotoxic, estrogenic or anti-androgenic when using in vitro bioassays (Bach et al., 2013). The objective of the study was to investigate the effect of sunlight on chemical migration into PET-bottled waters and to check the potential toxicities of water extracts using in vitro bioassays in order to avoid any hazard due to unpredictable NIAS (Muncke, 2011). The release of formaldehyde, acetaldehyde and Sb was monitored in bottled waters exposed to sunlight for 2, 6 and 10 days. Other potential migrants linked to plastic packaging (phthalates, nonylphenols, etc.) were also checked. Experiments were performed under realistic conditions of human exposure according to the EU Regulation No. 10/2011 (EU, 2011). Sax (2010) and Yang et al. (2011) mentioned that all plastics may yield endocrine disruptors under regular conditions of use. Next, relevant toxicological endpoints such as cyto/genotoxicity and also endocrine disruption potential were tested in bottled water extracts as a complementary approach to chemical analysis. Bioassays are useful tools to check potential toxicity due to unpredictable NIAS and/or chemical mixtures. Indeed, exhaustive analytical identification and confirmation of all compounds present in the migrates is difficult (Nerín, Alfaro, Aznar, & Domeño, 2013). Ames and micronucleus assays were performed to assess cyto/genotoxicity using prokaryotes and a human cell line (HepG2), respectively. Endocrine disruption potential (estrogenic and anti-androgenic) was assessed by gene reporter assays using human HepG2 and MBA-MB453-kb2 cell lines. Bioassays were chosen in accordance with EFSA and ICCVAM recommendations (EFSA, 2011; ICCVAM, 2003) for their performance. Results of bioassays were then correlated to chemical analysis.

2. Material and methods 2.1. Samples and storage conditions Two French brands of non-carbonated (brand A) and carbonated (brand B) water bottled in PET and in glass purchased from a local store were investigated. Brand A bottles had a light blue colour and were made up of a single PET layer with a pattern in relief on the surface. Brand B bottles had a green colour with a smooth surface and were made up of an immiscible lamellar polyamide (PA) phase within the PET. This PA phase reduces the permeability of O2 and CO2. This type of PET bottle was usually used for carbonated water. Water samples for each brand were from identical batches. For the experiments, three samples were derived from each brand by replacing mineral water by ultrapure water. Bottled waters were exposed to sunlight for 2, 6 and 10 days during July and August 2010 in the Bandol Weathering Station, Southern France. Samples were placed south-facing with an inclination of 45 degrees following the protocol described in the standard method ISO 877 (ISO, 2009). During the experiments, the solar irradiation received by the packaging material for each exposure duration was measured and the temperature of the bottled water was monitored using Thermo-tracersÒ (Oceasoft, Montpellier, France) (Table 1).

2.2. Solid phase extraction (SPE) The presence of 14 compounds found in plastic packaging was evaluated, namely: dimethyl phthalate (DMP), butylated hydroxytoluene (BHT), 2,6-di-tert-butyl-p-benzoquinone, 2,4-di-tert-butylphenol (2,4-dtBP), ethyl-4-ethoxybenzoate, diethyl phthalate (DEP), benzophenone, 4-nonylphenol (NP), 3,5-di-tert-butyl-4-hydroxybenzaldehyde (BHT-CHO), di-iso-butyl phthalate (DiBP), dibutyl phthalate (DBP), 2-ethylhexyl-p-methoxycinnamate, di-2-ethylhexyl adipate (DEHA) and di-2-ethylhexyl phthalate (DEHP) (Table 1S, Supplementary data). One litre of water was spiked with surrogate standards (2,6-di-tert-butyl-d9-4-methylphenol-3,5-d2, benzophenone-d5 and,di-2-ethylhexyl-phthalate-3,4,5,6-d4) at concentrations ranging from 0.5 to 1.6 lg/l depending on the target compounds. The water samples were then loaded on Oasis HLB glass cartridges (6 cc/200 mg, Waters, Milford, USA) previously conditioned with 5 ml of ethyl acetate (EA), methanol (MeOH) and UPLC-grade water (Biosolve, Valkenswaard, the Netherlands). Analytes were eluted with 2 ml of EA directly analysed by GC–MS (Section 2.3). In parallel, bottled samples were extracted for toxicological tests following the same procedure, although deutered standards were not added to the water samples. Chemical analysis and bioassays were then carried out on the EA extracts obtained (concentration factor 500). Preliminary toxicity tests of EA extracts were carried out for the cell lines used in this study (HepG2 and MDA-MB453-kb2 cells) to check the cytotoxicity of the solvent. EA was not cytotoxic at the final concentration of 1% in the culture medium (data not shown).

Table 1 Radiation values and mean temperatures reached in PET-bottled waters. Exposure duration (days)

Irradiation (MJ/m2)

Water temperatures (°C) Brand A bottles

2 6 10 *

Not available.

47.43 119.79 237.90

Brand B bottles

Mean

Min.

Max.

Mean

Min.

25.3 26.3 27.6

16.5 17.0 16.5

42.5 43.5 45.5

*

*

Max. *

27.4

16.5

45.5

*

*

*


2.3. GC–MS analysis A Varian 450 gas chromatograph (GC) coupled to a Varian 240 ion trap mass spectrometer (MS) (Walnut Creek, CA, USA) was used to analyse EA extracts. Large injection volumes (4 ll) in the split mode (1:25) were carried out. The inlet temperature was programmed as follows: 40 °C (hold 1 min) to 300 °C at 100 °C/min and hold at 300 °C for 15 min. Analytes were separated on an Rxi-5MS column (30 m 0.25 mm; 0.25 lm film thickness) connected with a 5 m 0.53 mm deactivated pre-column (Restek, Bellefonte, USA). The oven program was: 40 °C (hold 1 min) to 280 °C at 8 °C/min and 280 °C (hold for 15 min). Helium (carrier gas) was set at 1 ml/min. The transfer line, source and trap temperature were 310 °C, 220 °C and 200 °C, respectively. Data was acquired in full scan mode at a range of 40–600 m/z. The list of ions selected for the quantification is provided in Table 1S (Supplementary data). The LOQs were set on the basis of a signal-to-noise ratio of 10. However, phthalates were observed in blanks. Consequently, the phthalates’ LOQs were calculated to never exceed three times the LOQs of the blank values in order to ensure that the background contamination level remained lower that their limit of detection (LOD). Blanks were prepared with 1 l of UPLC-grade water (Biosolve, Dieuze, France) spiked with the labelled standards at 0.4 lg/l following the extraction procedure described (Section 2.2). The LODs for the analytes were defined as LOQ/3 (ISO/TS13530— Guidance on Analytical Quality Control for Chemical and Physicochemical Water Analysis). For the method employed here, the LOQ ranged from 0.1 lg/l (for most of the target compounds) to 0.3 lg/l (2,4-dtBP and 2-ethylhexyl-p-methoxycinnamate) (Table 1S). The concentration ranges for performing external calibration were from 10 to 1000 lg/l depending on the target compounds. Recovery experiments were carried out with spiked ULPC water and ranged from 44% to 114% (Table 1S). To ensure the validity of quantification during GC–MS analysis, calibration verifications were run for each sample batch. Analytical runs were acceptable if analyte concentrations in the calibration verifications were within ± 20% of the average concentration determined for each compound. For each sample batch, several water samples were fortified (concentrations from 0.5 to 1.6 lg/l depending on target compounds) with labelled standards and analytes to improve the efficiency of extraction and to detect matrix effects, respectively. UPLC blanks were also prepared for each sample batch in order to ensure that the contamination of lab glassware, connections, solvents and the analytical instrument were lower than the LODs. 2.4. Aldehyde analysis in bottled waters Aldehyde (formaldehyde, acetaldehyde, propanal, butanal, crotonaldehyde, pentanal, hexanal, heptanal, octanal, nonanal and decanal) analysis in bottled waters was performed following the protocol previously described by Bach et al. (2013). A derivatisation reaction was carried out with 500 ll of 2,4-dinitrophenylhydrazine (2,4-DNPH) reagent solution (2 mg/ml in acetonitrile (AcCN)) added to water samples (550 ml). The reaction conditions were 4 h at 60 °C without agitation. Carbonated water samples were degassed after derivatisation. The DNPH derivatised aldehydes were loaded through Oasis HLB cartridges (200 mg adsorbent, 6 cc; Waters, Milford, MA, USA) previously conditioned with AcCN (2 5 ml) and citrate buffer solution at 1 M (2 5 ml). The elution was carried out with 6 ml AcCN (2 3 ml). Ultrapure water was used to adjust the extracts to 7 ml prior to analysis. An Agilent 1200 HPLC system with an Agilent 1200 diode array detector (Palo Alto, CA, USA) was used for the aldehyde-DNPH analysis. Chromatographic separation was achieved with a

65

SunFire™ C18 column (250 4.6 mm I.D.; particle size, 5 lm; Waters, Milford, MA, USA) with a binary mixture of AcCN (A) and ultrapure water (B). The gradient program was as follows: isocratic elution at 60% A for 20 min, increase A to 90% over 15 min, and isocratic elution at 90% A. Detection was performed at a wavelength of 360 nm. Matrix-matched calibration was prepared with concentrations from 1 to 10 lg/l. The quantification limits (LOQ) were defined as the tenfold value of results obtained with ultrapure water blanks. The LOQ was 3.5 lg/l for formaldehyde, 2 lg/l for acetaldehyde and octanal, 3 lg/l for nonanal and decanal and 1.5 lg/l for the other aldehydes. 2.5. Analysis of trace metals Bottled water samples were analysed using Series XII inductively coupled plasma mass spectrometry (ICP-MS) (Thermo, Germany) following the ISO 17294-2 standard method (ISO, 2003). The operating conditions were as follows: RF power was 1318 W, the carrier, the auxiliary and the nebulizer argon gas flow were 13.0, 0.88 and 0.69 dm3/min, respectively. Rhodium at a concentration of 1 lg/l was used as the internal standard. The LOQ was 1 lg/l for trace metals, except for Sb (0.2 lg/l), Pb (0.1 lg/l) and V (0.5 lg/l). 2.6. Human cells Routine monitoring showed the cells to be mycoplasma-free (Mycoalert kit from Cambrex, Verviers, France). Stocks of cells were routinely frozen and stored in liquid N2. All experiments were performed using the cell lines on 10 passages after thawing. 2.6.1. HepG2 cell line The HepG2 cell line was obtained from the ECACC (European Collection of Cell Cultures, UK). The cells were grown in monolayer culture in MEM supplemented with 2 mM L-glutamine, 1% nonessential amino acids and 10% FBS in a humidified atmosphere of 5% CO2 at 37 °C. Continuous cultures were maintained by subculturing flasks every 7 days at 2.2 106 cells/75 cm2 flask by trypsination (trypsin (0.05%)–EDTA (0.02%)). 2.6.2. MDA-MB453-kb2 cell line This stable transfected human mammary cancer cell line was obtained from the ATCC (LGC Promochem, Molsheim, France). The cells were grown in monolayer culture in Leibovitz medium (L15) supplemented with 10% FBS in a humidified atmosphere at 37 °C. Continuous cultures were maintained by subculturing flasks every 7 days at 4.0 106 cells/75 cm2 flask by trypsination (trypsin (0.05%)–EDTA (0.02%)) solution from Invitrogen laboratories (Cergy-Pontoise, France). 2.6.3. Cell exposure to extracts Bioassays were performed with concentrated bottled water extracts after 10 days of sunlight exposure (238 MJ/m2 irradiation). Extracts were tested in bioassays under realistic consumer exposure conditions (1 kg of foodstuff/6 dm2 of material surface) in accordance with EU Regulation No. 10/2011 (EU, 2011). Cell sensitivity differs depending on the origins and protocols followed. Since transfected cells are more sensitive to vehicle, for the Ames test and micronucleus assay the final concentration of bottled water extract was 5 times more concentrated (1% of EA) than for the endocrine disruption assays (0.2% of EA). 2.7. Genotoxicity assays 2.7.1. Ames test The Ames test was carried out using the plate incorporation method with or without metabolic activation, with two histidine-

66


dependent auxotrophic mutants of Salmonella typhimurium strains, TA 98 and TA 100, essentially as described by Maron and Ames (1983). The S. typhimurium strains were provided by B. Ames (University of California, Berkeley, USA). The S9 mix was purchased from Trinova Biochem (Giessen, Germany). The protocol used was described by Bach et al. (2013). All the experiments were carried out in triplicate using three extract concentrations. Mutagenic activity was expressed as an induction factor, i.e. as a multiple of the background level. 2.7.2. Micronucleus assay This assay was performed following the protocol by Severin, Jondeau, Dahbi, and Chagnon (2005). HepG2 cells were seeded at 2.5 105 cells/well. After 24 h, cells were treated with 1% of the EA extract and cytochalasin B (4.5 lg/ml) for 44 h. Cells were then washed with PBS and allowed to recover for 1.5 h in MEM with 10% FBS. After washing with PBS, the cells were trypsinised (trypsin (0.05%)–EDTA (0.02%)) solution from Invitrogen laboratories (Cergy-Pontoise, France), fixed in two steps with acetic acid/MeOH (1/3) (v/v), spotted on a glass slide and stained with acridine orange (0.1%) diluted in Sorensen Buffer (1/15, v/v) just before reading. Micronuclei were counted visually in 1000 binucleated cells (BNC) per slide using a fluorescence microscope (Olympus CK40) and two slides per concentration were counted. To identify micronuclei, the criteria established by Kirsch-Volders et al. (2000) was applied: the diameter of micronuclei should be under one-third of that of the main nucleus, they should be clearly distinguishable from the main nucleus and they should have the same staining as the main nucleus. 2.8. In vitro endocrine disruptor potential 2.8.1. Estrogenic activity: Transcriptional activation assay with HepG2 cell line The protocol used was recently described by Bach et al. (2013). Briefly, HepG2 cells were seeded at a density of 1.2 105 cells per

well in 24-well tissue culture plates (Dutscher, France) and maintained in MEM medium without phenol red, supplemented with 10% dextran-coated charcoal fetal calf serum (DCC-FCS), 1% L-glutamin and 1% non-essential amino acids. The microplates were then incubated at 37 °C in a humidified atmosphere of 5% CO2 for 24 h. HepG2 cells were transiently transfected using the Exgen500 procedure (Euromedex) with the following plasmid mix: 100 ng ERE-TK-Luc and 100 ng hERa, 100 ng of pCMV-Gal and pSG5 to a final concentration of 0.5 lg DNA. Then, 2 ll of Exgen500 diluted in NaCl 0.15 M was added to the DNA. After vortex shaking, the microtubes were incubated at room temperature for 10 min. The Exgen500-DNA mixture was then added to OptiMEM without phenol red medium and distributed into the wells (300 ll/well). The microplate was then incubated at 37 °C in a humidified atmosphere of 5% CO2 for 1 h. After incubation, the OptiMEM was removed and replaced by 1 ml of treatment medium (MEM without phenol red, without FCS, 1% glutamin and 1% non-essential amino acids), containing the water extract, or the vehicle EA (1%, negative control), or 17-estradiol (10 8 M, positive control). The plate was then incubated for 24 h. At the end of the treatment, luciferase and -galactosidase activity was determined. 2.8.2. Anti-androgenic activity: Transcriptional activation assay using the human MDA-MB453-kb2 cell line The MDA-MB-453 (AR+) cell line was stably transfected with MMTV-neo-Luc with an (anti)-AR-responsive luminescent reporter gene (Wilson, Bobsein, Lambright, & Gray, 2002). Cells were seeded into a 24-well plate (Dutscher, France) in 1 ml of L15 medium without phenol red, supplemented with 5% of dextran-coated charcoal fetal calf serum (FCS), at a density of 5 104 cells/well. For anti-androgenic activity, 24 h after seeding, the medium was removed and cells were exposed to EA extracts (0.05%, 0.15% and 0.2%) in the presence of the androgenic reference dihydrotestosterone (DHT), (4 10 10 M, prepared in EA). Nilutamide (NIL) (10 6 M, prepared in EA) was used as a positive control for antiandrogenic activity. After 24 h treatment, cells were washed once

Fig. 1. Formaldehyde and acetaldehyde mean concentrations with standard deviations in PET-bottled waters exposed to sunlight for 2, 6 and 10 days. (A and B) Represent aldehyde migration into ultrapure water stored in PET bottles of brands A and B, respectively. (C and D) Correspond to the aldehyde migration in non-carbonated water (brand A) and in carbonated water (brand B), respectively. All analyses were performed in quintuplicate (water from five different bottles).


with 1 ml of phosphate buffered saline. Following 30 min incubation with 200 ll/well lysis buffer at room temperature with shaking, the lysates were briefly vortexed and centrifuged at 3000g at 4 °C for luciferase activity measurement, as described by Stroheker, Picard, Lhuguenot, Canivenc-Lavier, and Chagnon (2004). Ten ll from each well was transferred into an opaque white-walled plate and mixed with 40 ll of luciferase assay reagent. The plate was quickly covered with an adhesive seal and the mixture was immediately analyzed using a luminometer (TopCountNT, Packard). Results were expressed as a percentage of the androgenic positive control (DHT).

3. Results 3.1. Migration of 14 compounds linked to plastic packaging In PET- and glass-bottled waters exposed to the worst-case conditions (10 days of direct sunlight), 2,4-di-tert-butylphenol (2,4-dtBP) was detected but could not be quantified because its content was between the limit of detection (LOD) and the LOQ of the analytical method.

67

3.2. Migration of aldehydes Aldehydes were not detected in glass-bottled water before or after sunlight exposure. Only formaldehyde and acetaldehyde were found in PET-bottled water. The migration results are presented in the following subsections. 3.2.1. Effect of sunlight exposure on formaldehyde and acetaldehyde migration Impact of sunlight exposure on aldehyde migration into bottled water was assessed with ultrapure waters for both brands of PET bottles. In brand A bottles, formaldehyde and acetaldehyde migration increases to 11 lg/l and to 15 lg/l, respectively, after 10 days of exposure (Fig. 1A). However, in brand B bottles, formaldehyde migration was observed only after 10 days while acetaldehyde release was already observed at day 2 (Fig. 1B). At day 10, acetaldehyde concentrations were still higher than formaldehyde (1.4 times higher in brand A bottles and twice as high in brand B bottles). 3.2.2. Effect of water type (non-carbonated or carbonated) on formaldehyde and acetaldehyde migration In non-carbonated water (Fig. 1C), aldehyde migration was not observed, while in carbonated water (Fig. 1D), both aldehydes were already present before exposure (day 0) at 5 lg/l and 45 lg/l, respectively. A weak effect of sunlight on carbonated water with regard to formaldehyde migration was observed only at day 10 with a two-fold increase. In contrast, for acetaldehyde no sunlight effect was observed. This was due to the presence of carbon dioxide, which had already promoted its migration before the exposure experiments. Otherwise a steady concentration of acetaldehyde was observed. 3.3. Migration of trace metals 3.3.1. Effect of sunlight exposure on Sb migration With ultrapure water, sunlight exposure slightly increased Sb migration between 0 and 2 days, and then reached a plateau for both bottle brands (Fig. 2A), leading to a 0.5-fold increase. 3.3.2. Effect of the water type (non-carbonated or carbonated) on Sb migration At day 0, Sb was already present in non-carbonated and carbonated waters at concentration levels of 0.7 lg/l and 1.1 lg/l, respectively (Fig. 2B). In non-carbonated water, a weak effect of sunlight on the migration of Sb was observed (1.4 times concentration increase at day 10). Sb migration was more pronounced in carbonated water (1.8 times concentration increase), probably due to the presence of carbon dioxide. 3.4. Genotoxicity assays 3.4.1. Ames test The results of the Ames test on water extracts are presented in Table 2S (Supporting information). Negative and positive controls were consistent with the laboratory’s historical data. No mutagenic effect due to extracts was observed (induction factors traces 3 months), suggesting that genotoxic compounds undergo degradation in non-genotoxic substances.

With plants models, contradictory results and conclusions using the Allium Cepa test were reported after sunlight exposure. While a 2-fold increase in chromosomal aberrations was showed by Evandri, Tucci, and Bolle (2000) and Corneanu et al. (2010) attributed the chromosomal mutations observed to the mineral salt content of the water as well as the technology used for manufacturing the PET bottles. Our results are not in accordance with these previous studies due to differences in the bioassays, cell models (plants, human cell lines, etc.) and conditions used to perform them. In contrast with ecotoxicology, plants systems are not considered as primary screening tools for extrapolation to mammalian systems (EFSA, 2011). In addition, different sample preparations (cartridges, solvent polarities, etc.) could give different compound extraction efficiencies as demonstrated by Wagner and Oehlmann (2011). Several authors suggest that PET bottles may yield endocrine disruptor chemicals under regular conditions of use, such as long-term storage, high temperatures and exposure to sunlight (Sax, 2010; Wagner & Oehlmann, 2011; Yang et al., 2011). In this study, after 10 days of sunlight exposure, no estrogenic or antiandrogenic activity was detected in PET- and glass-bottled water extracts using HepG2 and MDA-MB453-kb2 cells, respectively. 5. Conclusions The effect of sunlight exposure on chemicals release into PETbottled waters and the potential hazard of water extracts were investigated using in vitro bioassays. The migration of aldehydes and Sb into ultrapure waters increased with sunlight especially after 10 days of exposure without exceeding the current specific migration limits set in Regulation No. 10/2011. However, an offflavour can occur due to the level of formaldehyde in carbonated waters after 10 days in sunlight. In carbonated mineral water, carbon dioxide contributed to migration more than sunlight. Water extracts did not induce any cyto-genotoxic or endocrine-disruption activity in the bioassays under our experimental conditions. Chemical analysis and global approaches using bioassays are complementary tools to identify the potential toxic effects due to unpredictable NIAS and/or chemical mixtures. Acknowledgements This research was financed by the French Agency for Food, Environmental and Occupational Health & Safety (ANSES) and the Institute Jean Lamour of the University of Lorraine. The authors wish to thank the Bandol Weathering Station (SEVN) and the Water Chemistry Department of ANSES’ Nancy Laboratory for Hydrology for their excellent technical assistance. The authors are grateful to C. Dumont, K. Raja, A. Novelli, V. Fessard, C. Tricard, and E. Barthélémy for their collaboration. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.foodchem.2014. 04.020. References Alin, J., & Hakkarainen, M. (2011). microwave heating causes rapid degradation of antioxidants in polypropylene packaging, leading to greatly increased specific migration to food simulants as shown by ESI–MS and GC–MS. Journal of Agricultural and Food Chemistry, 59(10), 5418–5427. Amiridou, D., & Voutsa, D. (2011). Alkylphenols and phthalates in bottled waters. Journal of Hazardous Materials, 185(1), 281–286. Bach, C., Dauchy, X., Chagnon, M. C., & Etienne, S. (2012). Chemical compounds and toxicological assessments of drinking water stored in polyethylene terephthalate (PET) bottles: A source of controversy reviewed. Water Research, 46(3), 571–583.

C. Bach et al. / Food Chemistry 162 (2014) 63–71 Bach, C., Dauchy, D., Severin, I., Munoz, J.-F., Etienne, S., & Chagnon, M.-C. (2013). Effect of temperature on the release of intentionally and non-intentionally added substances from polyethylene terephthalate (PET) bottles into water: Chemical analysis and potential toxicity. Food Chemistry, 139, 672–680. Casajuana, N., & Lacorte, S. (2003). Presence and release of phthalic esters and other endocrine disrupting compounds in drinking water. Chromatographia, 57(9–10), 649–655. Cheng, X., Shi, H., Adams, C. D., & Ma, Y. (2010). Assessment of metal contaminations leaching out from recycling plastic bottles upon treatments. Environmental Science and Pollution Research, 17(7), 1323–1330. Corneanu, M., Corneanu, G., Jurescu, N., & Toptan, C. (2010). Evaluation of the genotoxicity of water bottled in PET recipients. Environmental Engineering and Management Journal, 9(11), 1531–1537. Dabrowska, A., Borcz, A., & Nawrocki, J. (2003). Aldehyde contamination of mineral water stored in PET bottles. Food Additives and Contaminants, 20(12), 1170–1177. De Fusco, R., Monarca, S., Biscardi, D., Pasquini, R., & Fatigoni, C. (1990). Leaching of mutagens into mineral water from polyethyleneterephthalate bottles. Science of the Total Environment, 90, 241–248. Dévier, M.-H., Le Menach, K., Viglino, L., Di Gioia, L., Lachassagne, P., & Budzinski, H. (2013). Ultra-trace analysis of hormones, pharmaceutical substances, alkylphenols and phthalates in two French natural mineral waters. Science of the Total Environment, 443, 621–623. EFSA (2011). Scientific opinion on genotoxicity testing strategies applicable to food and feed safety assessment. EFSA Journal, 9(9), 2379. 2368 pp. EFSA (2012). Scientific Opinion on the safety evaluation of the substance, titanium nitride, nanoparticles, for us in food contact materials. EFSA Journal, 10(3), 2641–2649. EU (2004). Regulation No. 1935/2004 of the European Parliament and of the council of 27 October 2004 on materials intended to come into contact with food and repealing directives 80/590/EEC and 89/109/EEC. EU (2011). Commission regulation (EU) No. 10/2011 of 14 January 2011 on plastic materials and articles intended to come in contact with food. Evandri, M. G., Tucci, P., & Bolle, P. (2000). Toxicological evaluation of commercial mineral water bottled in polyethylene terephthalate: A cytogenetic approach with Allium cepa. Food Additives and Contaminants, 17(12), 1037–1045. Grob, K., Biedermann, M., Scherbaum, G., Roth, M., & Rieger, K. (2006). Food contamination with organic materials in perspective: Packaging materials as the largest and least controlled source? A view focusing on the European situation. Critical Reviews in Food Science and Nutrition, 46(7), 529–536. Honkalampi-Hämäläinen, U., Bradley, E. L., Castle, L., Severin, I., Dahbi, L., Dahlman, O., et al. (2010). Safety evaluation of food contact paper and board using chemical tests and in vitro bioassays: role of known and unknown substances. Food Additives & Contaminants, 27(3), 406–415. ICCVAM (2003). Evaluation of In vitro test methods for detection potential endocrine disruptors: Estrogen receptor and androgen receptor binding and transcriptional activation assays. May, 2003. NIH Pub 03-4503. Available: http:// iccvam.niehs.nih.gov/docs/endo_docs/edfinalrpt0503/edfinrpt.pdf [accessed 7 July 2012]. ISO (2003). ISO 17294-2. Water quality. Application of inductively coupled plasma mass spectrometry (ICP-MS). Part 2: Determination of 62 elements. ISO (2009). International Standard ISO 877:2099. Plastics – Methods of exposure to solar radiation – Part 2: Direct weathering and exposure behing window glass. JORF (2011). Arrêté du 28 décembre 2010 modifiant l’arrêté du 14 mars 2007 relatif aux critères de qualité des eaux conditionnées, aux traitements et mentions d’étiquetage particuliers des eaux minérales naturelle et des eaux de source conditionnées ainsi que de l’eau minérale naturelle distribué en buvette publique. Ministère du travail, de l’emploi et de la santé. Journal Officiel de la République Française, Texte 42. Keresztes, S., Tatár, E., Mihucz, V. G., Virág, I., Majdik, C., & Záray, G. (2009). Leaching of antimony from polyethylene terephthalate (PET) bottles into mineral water. Science of the Total Environment, 407(16), 4731–4735.

71

Kirsch-Volders, M., Sofuni, T., Aardema, M., Albertini, S., Eastmond, D., Fenech, M., et al. (2000). Report from the In Vitro Micronucleus Assay Working Group. Environmental and Molecular Mutagenesis, 35(3), 167–172. Kortenkamp, A. (2007). Ten years of mixing cocktails: a review of combination effects of endocrine-disrupting chemical. Environmental Health Perspectives, 115(1), 98–105. Leivadara, S. V., Nikolaou, A. D., & Lekkas, T. D. (2008). Determination of organic compounds in bottled waters. Food Chemistry, 108(1), 277–286. Maron, D. M., & Ames, B. N. (1983). Revised methods for the Salmonella mutagenicity test. Mutation Research, 113(3–4), 173–215. Monarca, S., De Fusco, R., Biscardi, D., De Feo, V., Pasquini, R., Fatigoni, C., et al. (1994). Studies of migration of potentially genotoxic compounds into water stored in pet bottles. Food and Chemical Toxicology, 32(9), 783–788. Muncke, J. (2011). Endocrine disrupting chemicals and other substances of conerne in food contact materials: and updated review of exposure effect and risk assessment. Journal of Steroid Biochemistry and Molecular Biology, 127(1–2), 118–127. Mutsuga, M., Kawamura, Y., Sugita-Konishi, Y., Hara-Kudo, Y., Takatori, K., & Tanamoto, K. (2006). Migration of formaldehyde and acetaldehyde into mineral water in polyethylene terephthalate (PET) bottles. Food Additives and Contaminants, 23(2), 212–218. Neamtu, M., & Frimmel, F. H. (2006). Photodegradation of endocrine disrupting chemical nonylphenol by simulated solar UV-irradiation. Science of the Total Environment, 369(1–3), 295–306. Nerín, C., Alfaro, P., Aznar, M., & Domeño, C. (2013). The challenge of identifying non-intentionally added substances from food packaging materials: A review. Analytica Chimica Acta, 775, 14–24. OECD (2010). Guideline for the testing of chemicals. No 487: In Vitro Mammalian Cell Micronucleus Test. http://iccvam.niehs.nih.gov/SuppDocs/FedDocs/OECD/ OECD-TG487.pdf. Sax, L. (2010). Polyethylene terephthalate May yield endocrine disruptors. Environmental Health Perspectives, 118(4), 445–448. Schmid, P., Kohler, M., Meierhofer, R., Luzi, S., & Wegelin, M. (2008). Does the reuse of PET bottles during solar water disinfection pose a health risk due to the migration of plasticisers and other chemicals into the water? Water Research, 42(20), 5054–5060. Severin, I., Jondeau, A., Dahbi, L., & Chagnon, M. C. (2005). 2,4-Diaminotoluene (2,4DAT)-induced DNA damage. DNA repair and micronucleus formation in the human hepatoma cell line HepG2. Toxicology, 213(1–2), 138–146. Stroheker, T., Picard, K., Lhuguenot, J.-C., Canivenc-Lavier, M.-C., & Chagnon, M.-C. (2004). Steroid activities comparison of natural and food wrap compounds in human breast cancer cell lines. Food and Chemical Toxicology, 46(6), 887–897. Strube, A., Buettner, A., & Groetzinger, C. (2009). Characterization and identification of a plastic-like off-odor in mineral water. Water Science and Technology: Water Supply, 9, 299–309. Ubomba-Jaswa, E., Fernandez-Ibanez, P., & McGuigan, K. G. (2010). A preliminary Ames fluctuation assay assessment of the genotoxicity of drinking water that has been solar disenfected in polyethylene terephthalate (PET) bottles. Journal of Water and Health, 8(4), 712–719. Wagner, M., & Oehlmann, J. (2011). Endocrine disruptors in bottled mineral water: Estrogenic activity in the E-Screen. Journal of Steroid Biochemistry and Molecular Biology, 127(1–2), 128–135. Wegelin, M., Canonica, S., Alder, C., Marazuela, D., Suter, M. J. F., Bucheli, T. D., et al. (2001). Does sunlight change the material and content of polyethylene terephthalate (pet) bottles? Journal of Water Supply: Research and Technology AQUA, 50(3), 125–133. Wilson, V. S., Bobsein, K., Lambright, C. R., & Gray, L. E. (2002). A novel cell line, MDA-kb2, that stably expresses an androgen- and glucocorticoid-responsive reporter for the detection of hormone receptor agonist and antagonist. Toxicological Science, 66(1), 69–81. Yang, C. Z., Yaniger, S. I., Jordan, V. C., Klein, D. J., & Bittner, G. D. (2011). Most Plastic Products Release Estrogenic Chemicals: A Potential Health Problem That Can Be Solved. Environmental Health Perspectives, 119(7), 982–996.