- Email: [email protected]
Liver volatolomics to reveal poultry exposure to γ-hexabromocyclododecane (HBCD)
Chemosphere 189 (2017) 634e642
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Liver volatolomics to reveal poultry exposure to ghexabromocyclododecane (HBCD) re my Ratel a, Christelle Planche a, Fre de ric Mercier a, Patrick Blinet a, Je s Fournier c, Ange lique Travel d, Nathalie Kondjoyan a, Philippe Marchand b, Agne c a, * Catherine Jondreville , Erwan Engel INRA, UR QuaPA, F-63122 Saint-Gen es-Champanelle, France LUNAM Universit e, Oniris, LABERCA, F-44307 Nantes, France INRA, Universit e de Lorraine, UR AFPA, F-54500 Vandoeuvre-les-Nancy, France d INRA, ITAVI, F-37380 Nouzilly, France a
b c
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Poultry liver volatolome is modified by the dietary exposure of animals to g-HBCD. Volatile candidate markers in liver enable to reveal livestock exposure to g-HBCD. Candidate markers are mainly hydrocarbons or oxygenated compounds. Liver volatolomics is a promising approach to strengthen food contamination control.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 June 2017 Received in revised form 8 September 2017 Accepted 15 September 2017 Available online 17 September 2017
Hexabromocyclododecane (HBCD) is a critical emerging brominated flame retardant to which consumers can be exposed at high doses through a single food intake. Based on an animal experiment involving 3 groups of laying hens fed during 70 days with a control diet or g-HBCD-contaminated diets at 0.1 or 10 mg g-HBCD g1 feed, this study aims to use the volatolome of biological samples for revealing markers of livestock exposure to HBCD. Liquid chromatographyetandem mass spectrometry was used to monitor the time-course of HBCD levels in bodily samples. Each liver was analyzed by solid-phase microextractionegas chromatographyemass spectrometry for volatolome profiling. After 70 days, g-HBCD concentrations in egg yolk, fat, liver and serum reached 54 ± 4, 85 ± 6, 31 ± 6, and 32 ± 4 ng g1 lw, respectively, for the low exposure level and 4.6þ/5.7, 7.8þ/6.5, 3.9þ/3.0 and 3.9þ/6.1 mg g1 lw, respectively, for the high exposure level. Isomerization of g-HBCD into a- and b-HBCD was observed in all tissues, at least for the high exposure level. Volatolome data allowed a significant discrimination between control and exposed animals whatever the feed contamination load, demonstrating a liver metabolic response to g-HBCD exposure. The relevance of the twenty nine volatile exposure markers tentatively identified was discussed in light of literature data. © 2017 Elsevier Ltd. All rights reserved.
Handling Editor: Frederic Leusch Keywords: Food safety g-HBCD Animal-derived food Volatolomics Liver SPME-GC-MS
* Corresponding author. E-mail address: [email protected] (E. Engel). https://doi.org/10.1016/j.chemosphere.2017.09.074 0045-6535/© 2017 Elsevier Ltd. All rights reserved.
J. Ratel et al. / Chemosphere 189 (2017) 634e642
1. Introduction Brominated flame retardants (BFRs) are persistent organic pollutants that are increasingly contaminating the environment, wildlife, and people (Birnbaum and Staskal, 2004). Since 2014, the European Commission recommends monitoring foodstuffs for BFRs (European recommendation 2014/118/EU, 2014). As critical emerging contaminant, hexabromocyclododecane (HBCD) is a BFR that recently joined the list of chemicals targeted by the Stockholm Convention (Covaci et al., 2006). This pollutant is now part of Annex A which requires signatories to take measures to eliminate the production and use of HBCD due to the dangers for human health and the environment. Indeed oral exposure to HBCD causes a broad range of adverse effects like endocrine effects, drug-metabolizing enzyme disorders or cancer occurrence by a nonmutagenic mechanism (Covaci et al., 2006; van der Ven et al., 2009). Human exposure to HBCD occurs through multiple routes, but the route of most concern is exposure through diet (Goscinny et al., 2011). In Europe, median chronic dietary exposure estimated attains 0.43 ng kg1 day1 in adults and 1.27 ng kg1 day1 in children (EFSA, 2011). EFSA concluded that current median dietary exposure to HBCD does not raise a health concern given the benchmark dose lower confidence limit for a benchmark response of 10% (BMDL10) of 790,000 ng kg1 day1 (EFSA, 2011). However, HBCD concentrations in animal-derived food products are reported to be subjected to wide variabilities. For example, the literature reports HBCD concentrations as high as 62 (Covaci et al., 2009), 72 (Rawn et al., 2011), 90 (Blake, 2005) or even 2000 (Hiebl and Vetter, 2007) ng g1 lipid in chicken eggs. According to surveillance programs led by the French Directorate General for Food in France in 2008 and 2009, the extreme values reached 3055 ng g1 lipid in poultry meat and 3390 ng g1 in eggs (Travel et al., 2012). Thus, the ingestion of a single contaminated animal-derived food can significantly increase the consumer exposure to HBCD. The gold standard techniques for measuring HBCD in food are LC/MS (Morris et al., 2006) or LC/MS/MS (Tomy et al., 2004), because these approaches allow to determine the different HBCD isomers. However, the demands of an ambitious monitoring program for food safety required the means and capacity to perform measurements that were precise, accurate, robust but also costeffective and not cumbersome (Focant, 2014), and direct HBCD measurement is particularly difficult to implement due to the ubiquitous presence of this contaminant in the environment (Covaci et al., 2006). This prevents the setting up of rapid controls in the production chain to strengthen food safety. Emerging toxicogenomics approaches have been proposed on the basis of the discovery of metabolite markers of contamination (Engel et al., 2015). Research has evidenced that some biological signatures of the human body can be used to reveal the significant perturbations that link exposure and toxicity pathways, such as the use of volatile organic compounds (VOCs) from exhaled air of patients for rapid non-invasive diagnosis of cancers (Hakim et al., 2012). Intensive research has been conducted on the potential use of volatile metabolites emanating from cells and their microenvironment, named “volatolome”, to diagnose pathologies in humans (Broza et al., 2015). Pathophysiologic disorders can impact the composition of the volatolome of living cells or tissues, with a production of new volatile metabolites or a change in the ratio of volatile metabolites normally produced by the body (Hakim et al., 2012). Recently, Berge et al. (2011) evidenced that the exposure of animals to toxic xenobiotics generated an altered pattern of VOC content in their liver, which plays a key role in detoxification function, and they proposed to use this volatolomic signature to reveal a contamination of the food chain. The present study aimed to monitor the time-course of HBCD
635
levels in different tissues of animals dietary exposed to HBCD, and to complete the proof of concept study carried out by Berge et al. (2011) in identifying candidate volatile markers of HBCD chemical contamination. To address this issue, we designed an animal experiment on laying hens dietary exposed or not to g-HBCD, which is the most present HBCD isomer in technical mixtures. 2. Materials and methods 2.1. Feed contamination In the study by Fournier et al. (2012), hens given a diet containing 1 ng g-HBCD g1, laid eggs containing 0.4 ng HBCD g1 lw. Thus, the dose was increased up to 0.1 and 10 mg g1 feed in order to achieve concentrations in eggs consistent with the several tens and thousands recorded in several surveys (Covaci et al., 2009; Rawn et al., 2011; Blake, 2005; Hiebl and Vetter, 2007; Travel et al., 2012). Contaminated diets were prepared from the same maize and soybean meal basal feed including all the nutrient requirements of laying hens. g-HBCD (1,2,5,6,9,10hexabromocyclododecane) obtained from Sigma Aldrich (St. Louis, MO) was dissolved in rapeseed oil then mixed with the basal feed according to Fournier et al. (2012) for spiked feed preparation. Non-contaminated feed was prepared by mixing the same amount of blank rapeseed oil with the basal feed. Feeds were administered as mash. 2.2. Animal test The experiment was conducted under the application of the Directive 2010/63/EU in France. Fifty-six laying hens (Lohmann Brown, 43-week-old; weight 1.77 ± 0.16 kg) were placed in an appropriate facility (agreement n B54-547-15) with individual cages allowing the individual control of feed ingestion and of egg production. Two groups of 24 animals were exposed to the two levels of HBCD. After a one-week adaptation period, the hens were given the contaminated diets during 70 days. The remaining 8 hens were used as a control group given the non-contaminated feed. The daily allowance of feed was adjusted to 6.9% of body weight (bw) and given once in the morning, water was continuously available, room temperature was maintained at 24 C and lighting was set to a 16 h light/8 h dark cycle. HBCD-exposed laying hens were sequentially slaughtered after 1, 12, 21, 36, 54 or 70 days of exposure. Control laying hens were slaughtered the day before the experiment start (n ¼ 4) or on day 70 (n ¼ 4). Hens were slaughtered after a 12-h fasting period by electronarcosis followed by exsanguination. 2.3. Sampling The amount of feed ingested and the number and weight of eggs laid were individually recorded daily. Each feed was weekly sampled and stored at þ4 C. Blood was collected by puncture at the jugular vein in hens slaughtered after 1, 21 and 70 days of exposure, then it was allowed clotting during 24 h at room temperature before serum was prepared by centrifugation (1400 g, 15 min) and stored at 20 C. At slaughter, each hen was weighed, and samples (egg, abdominal fat and liver) were collected and weighed. Egg yolks were freeze-dried, stored at 20 C and finely ground before analysis. Abdominal fat and liver samples were immersed in liquid nitrogen, wrapped in aluminum foil, vacuumpacked and stored at 80 C. Fat and liver were ground for 1 and 3 min, respectively, in liquid nitrogen into a fine homogeneous powder using a home-made stainless steel ball mill. For the analysis of liver volatolome, a 1.2 g aliquot of each powdered liver was
636
J. Ratel et al. / Chemosphere 189 (2017) 634e642
placed in a glass vial (Supelco, Sigma-Aldrich, St. Louis, MO), sealed under a nitrogen flow, and stored at 80 C. 2.4. HBCD quantification HBCD isomers (a, b and g) were quantified by Laberca (ONIRIS, France), the French National Reference Laboratory for most environmental organic contaminants in food. Briefly, after freeze drying, HBCDs were extracted from all matrices by Accelerated Solvent Extraction (ASE300, Dionex, Sunnyvale, CA) with toluene and acetone (70:30 v/v). The extracts were cleaned up using a multilayer silica column (from top to bottom: anhydrous sodium sulfate (5 g), concentrated sulfuric acid acidified silica (25 g), neutral silica (5 g)). HBCDs were assayed by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) according to Fournier et al. (2012). For a-, b- and g-HBCD, limits of quantification (LOQ) were respectively 0.013, 0.004, and 0.038 ng g1 fresh matter (fm) in feed, 0.24, 0.11 and 0.12 ng g1 lipid weight (lw) in liver, 0.045, 0.046 and 0.006 ng g1 lw in abdominal fat, 0.037, 0.013 and 0.037 ng g1 lw in egg yolk and 0.56, 0.15 and 1.1 ng g1 lw in serum. 2.5. Liver volatolome analysis Volatolome of powdered liver samples was analyzed by solidphase microextraction (SPME) coupled to gas chromatography and mass spectrometry (GC-MS) according to Berge et al. (2011). Briefly, the following steps were carried out with an automated sampler (AOC-5000 Shimadzu, Japan): (i) preheating of the sample for 10 min at 60 C in the agitator (500 rpm), (ii) SPME trapping (75 mm carboxen/polydimethylsiloxane, 23-gauge needle, Supelco) of the volatile organic compounds (VOCs) for 30 min at 60 C, and (iii) thermal desorption at 250 C for 2 min in the GC inlet. Further VOC analysis was performed by GC/MS-full scan (GC2010; QP2010þ, Shimadzu). VOCs were injected in splitless mode into a DB-5MS capillary column (60 m 0.32 mm 1 mm; Agilent J&W). Oven temperature was held at 40 C for 5 min, then ramped up to 230 C at a gradient of 3 C.min1, and held at 230 C for 10 min. Temperature of the transfer line between GC and MS was set at 230 C. Temperature was fixed at 180 C in the MS source and 150 C in the MS quadrupole. Electron impact energy was set at 70 eV, and data was collected in the range of 33e250 m/z at a scan range of 10 scans per second. First, SPME-GC-MS chromatograms obtained in full-scan mode were converted into virtual SPME-MS fingerprints to get a quick estimate the informative potential of the volatile metabolome. Second, tentative identification of volatiles was performed from the SPME-GC-MS signals on the basis of both mass spectra, by comparison against the NIST 14 mass spectral library, and retention indices (RI), by comparison against published RI values and those of our internal database. Peak area of the tentatively-identified compounds was integrated with the GC-MS solution software (version 2.53, Shimadzu) using a mass fragment selected as being specific and free of any coelution. 2.6. Data treatment 2.6.1. Body weight, egg production and HBCD quantification Statistical analyses were performed by means of the Statistical Analysis Systems software package (version 9.1, SAS Institute). The hen was considered as the experimental unit. Differences were considered significant for p < 0.05. First, the absence of impact of exposure to g-HBCD on body weight, feed ingested and the number and mass of eggs was checked as described in supplementary material (Table S1). In order to facilitate the statistical comparison
between the two levels of exposure, despite the 100-fold difference between the two feed HBCD concentrations, the concentrations of each isomer in matrices were divided by 100 for hens given the feed containing 10 mg g-HBCD g1 feed. The resulting normalized concentrations were analyzed as repeated measures according to a MIXED procedure as described by Dominguez-Romero et al. (2016). The model included the date of sample collection (time, n ¼ 6), the level of feed contamination (0.1 or 10 mg g1 feed, n ¼ 2), the animal matrix as repeated measure (egg yolk, liver, abdominal fat, serum, n ¼ 4) and their interactions. The animal matrix was introduced in the model using the “repeated” statement, in order to account for the correlation between the tissues collected from the same hen. Compound symmetry was used as covariance structure, thus measurements on the same hen were considered to be similarly correlated. 2.6.2. Liver volatolome Data were processed using Statistica software (version 12, StatSoft) and R software (version 3.3.2., http://www.R-project.org/). The variables used in data treatments were the mass fragments for virtual SPME-MS fingerprints and the VOCs for SPME-GC-MS signals. Datasets were normalized using the systematic ratio normalization (SRN) procedure developed by Lehallier et al. (2012) to improve both sample set discrimination and biomarker discovery. The procedure consists in calculating all the log-ratios between variables for each sample, then selecting the log-ratio(s) that best maximize the discrimination between sample-sets by one-way ANOVA at 1% level of significance corrected by the Bonferroni post-test. Principal component analyses (PCA) were performed on the discriminant ratios selected according to the animal group factor. To assess the test-long time-course of the discriminant variables, ANOVA (stage of exposure; p < 0.05) with 6 modalities (the timepoints from day 1 to day 70) was performed on the SRN data provided by analysis of livers from the animals exposed to 10 mg HBCD g1 feed. 3. Results and discussion 3.1. HBCD in spiked feed The concentrations of a- and b-HBCD were below the LOQ in the control feed, while the concentration of g-HBCD was 0.038 ng g1. In the contaminated feed intended to contain 0.1 mg g-HBCD g1, compliant concentrations of 0.7, 0.5 and 91.6 ng a-, b- and g-HBCD g1, respectively, were measured. No control of the target concentration of 10 mg g-HBCD g1 was performed to avoid any contamination of the analytical device. 3.2. Feed intake, body weight and laying Results are presented in supplementary material (Table S1). Hens exposed more than one day ingested daily 6.11 or 611 mg gHBCD kg1 bw. Body weight, feed intake and laying rate were very similar in control and in exposed hens and were independent of the dietary level of g-HBCD (p > 0.1). Similarly, a 75-day exposure of American kestrels to 510 mg technical HBCD kg1 bw did not alter body characteristics of animals, even if their reproductive performance was disturbed with more numerous but 3% lighter eggs and an increased eggshell porosity (Fernie et al., 2011). 3.3. HBCD in liver, fat, serum and egg yolk HBCD isomers were quantified in bodily samples collected from control hens (Table S2). As expected, concentrations of g-HBCD measured in egg yolk, liver and fat were low, although these
J. Ratel et al. / Chemosphere 189 (2017) 634e642
concentrations tended to increase with time possibly due to a contamination through dust originating from contaminated feed. Fig. 1 reports the experiment-long time-course of the sum of the three HBCD isomers concentrations determined by LC-MS/MS in the tissues of the animals exposed to 0.1 (Fig. 1A) and 10 (Fig. 1B) mg
637
g-HBCD/g feed. The results of the statistical analysis performed on standardized concentrations (Table S3) indicate that the overall increase of HBCD concentration with time in all four tissues (Time, p < 0.001) was independent of the level of feed contamination (Level and Time*Level, p > 0.1). The proportionality of the dose and
Fig. 1. HBCD (sum of a-, b- and g-) quantification by LC-MS/MS in samples from animals exposed to (A) 0.1 and (B) 10 mg HBCD g1 feed. Values are adjusted means (n ¼ 4) ± SE.
638
J. Ratel et al. / Chemosphere 189 (2017) 634e642
the concentration in tissues was already observed in mice orally exposed to doses ranging between 3 and 100 mg g-HBCD kg1 bw (Szabo et al., 2010). Overall, abdominal fat was the most concentrated tissue, and liver and serum the least concentrated, while egg yolk was intermediate (Tissue, p < 0.001). However, time trends differed between tissues (Time*Tissue, p < 0.001): while concentrations were rapidly close to their maximum value in liver and in serum, they increased more progressively in egg yolk and in abdominal fat. At the end of the experiment, the low contaminated diets resulted in HBCD concentrations of 53.6 ± 3.5, 85.3 ± 5.6, 30.5 ± 5.7 and 32.1 ± 3.8 ng g1 lw in egg yolk, abdominal fat, in liver and in serum, respectively. For the high contaminated diet they reached 4.6 ± 0.6, 7.8 ± 0.6, 3.9 ± 0.3 and 3.9 ± 0.6 mg g1 lw, respectively. These values in egg are consistent with those recorded in some surveys (Covaci et al., 2009; Rawn et al., 2011; Blake, 2005; Hiebl and Vetter, 2007; Travel et al., 2012). A similar ranking of tissues was observed in laying hens given feed spiked with a- or gHBCD (Fournier et al., 2012; Dominguez-Romero et al., 2016), and also in American kestrels exposed to a technical mixture of HBCD (Letcher et al., 2015). The well documented izomerization of g-HBCD into a- and bHBCD (Koch et al., 2015) was evidenced in the current study as the three isomers were detected and quantified in all tissues, except bHBCD in liver and in serum from hens exposed to 0.1 mg g-HBCD g1 feed. Time dependent concentrations of the three isomers are presented in Table S4 and in Fig. S1. In accordance with previous results (Fournier et al., 2012), the g-isomer was dominant in all tissues, making around 95% of total HBCD, while a- and b-isomers accounted for around 4.5% and 0.5% of total HBCD, respectively. 3.4. Response of the liver volatolome to g-HBCD exposure 3.4.1. SPME-MS fingerprinting of the volatolome Fig. 2 presents the projection of the two most discriminant ratios of mass fragments in the volatolome fingerprints of hen livers at the maximum HBCD exposure (70 days). The clear distinction of exposed animal groups (0.1 and 10 mg g-HBCD g1 feed) from the control group indicates that dietary contamination by HBCD
triggered changes in liver metabolism and that these changes were perceptible through the volatolome fingerprint. This result is consistent with several previous studies showing that volatolome fingerprints in different animal body compartments were relevant for differentiating animals according to various factors, such as type of animal feeding (Vasta et al., 2007; Sivadier et al., 2008) or occurrence of pollutants in the feed (Berge et al., 2011; Bouhlel et al., 2017). As shown in Fig. 2, separation between the two groups of exposed hens was limited. This may be due to the low resolution of the SPME-MS approach which focuses on the most abundant constituents of the volatolome (Berge et al., 2011). To better explore the information content of liver volatolome, it is necessary to increase the resolution and to reprocess the SPME-GCMS signals of these same samples. 3.4.2. SPME-GC-MS profiling of the volatolome The detailed composition of the liver volatolome was investigated by SPME-GC-MS in both control and 70-day exposed animals, in which 237 volatile organic compounds were determined. Systematic ratio normalization data treatment (Lehallier et al., 2012) revealed 44 ratios discriminating the 3 groups of animals. Fig. 3 reports the first map of PCA performed on the corresponding dataset, and shows that using SPME-GC-MS profiling allows the three groups to be separated, and even two contamination levels to be separated, which was not possible using SPME-MS fingerprints. Table S5 points out the loadings for the first factorial plane. This result shows that the resolution gain provided by the chromatographic separation made it possible to unmask dose-sensitive information in the liver volatolome. Liver volatolome signature was thus significantly different between HBCD-exposed hens and control hens. This change in liver volatolome in response to HBCD exposure is consistent with the results of the proof of concept study carried out by Berge et al. (2011) dealing in particular with exposure to polybrominated diphenyl ethers (PBDEs) which are another well-known class of brominated flame retardants. Berge et al. (2011) proposed that the occurrence of these toxicants, which are known to be quickly metabolized after oral intake, generates a metabolic response in the liver and consequently changes the liver
1,9
Mass fragment ratio 2
1,7
1,5
1,3 0,45
0,55
0,65
0,75
0,85
0,95
Mass fragment ratio 1 Fig. 2. Discrimination of hens exposed or not to HBCD according to virtual SPME-MS fingerprint of the liver volatolome. The first map was plotted on the two most discriminant ratios of mass fragments discriminative of the 3 groups of animals. Control; 0.1 mg g-HBCD g1 feed; 10 mg g-HBCD g1 feed.
J. Ratel et al. / Chemosphere 189 (2017) 634e642
639
3.5. Liver VOC markers of g-HBCD exposure SRN data treatment revealed 43 volatile organic compounds, whose 29 were tentatively identified (Table 1). Among these 29 VOCs, 9 were overexpressed in samples from exposed animals whereas 15 were underexpressed in samples from exposed animals compared to controls. These VOCs which could be considered candidate markers of animal exposure to g-HBCD belong mainly to the following compound families: alkanes and branched-chain alkanes, alcohols, aldehydes, ketones, phenols and alkyl benzenes.
Principal component 2 (9.3%)
3
1
-1
-3 -10
-6
-2
2
6
Principal component 1 (70.1%) Fig. 3. Discrimination of hens exposed or not to HBCD using volatile organic markers of liver volatolome. First map of normed PCA was plotted on the 44 ratios of VOCs discriminative of the 3 groups of animals. Control; 0.1 mg g-HBCD g1 feed; 10 mg g-HBCD g1 feed.
volatolome, and g-HBCD is known to be rapidly biotransformed and eliminated (Koch et al., 2015). To assess the time-course change in the balance between discriminative variables, we highlighted those which are impacted by stage of exposure. When we considered the animal group with the highest HBCD load (10 mg HBCD g1 feed), 15 ratios were found influenced by the stage of exposure. The first map of the PCA (Fig. 4) shows their evolution in the liver of hens exposed to the high-level dose of HBCD during the animal test. Animal samples are fairly well separated, showing that part of the discriminative information for revealing HBCD exposure is thus significantly influenced by stage of exposure. More specifically, these ratios are able to separate certain periods of exposure (1e12, 21, 36e54 and 70 days). The liver volatolome thus appears a very promising focus of efforts to detect hen contamination by g-HBCD.
3.5.1. Alkanes and branched-alkanes Pentane, 2-methylpentane, methylcyclopentane, 2,2,4,6,6pentamethylheptane and 2,2,4,4-tetramethyloctane were identified in animal liver as markers of HBCD exposure. Alkanes and methylated alkanes have been reported at the crossroads of human cell metabolism elicited in response to exposure to toxic xenobiotics (Hakim et al., 2012; Haick et al., 2014; Amann et al., 2014). According to the literature, exposure to toxic xenobiotics induces two main cellular reactions (Hakim et al., 2012). First, oxidative stress can be induced, with synthesis of reactive oxygen species (ROS) that leak from the mitochondria or from peroxidated polyunsaturated fatty acids in the cell membranes. Second, detoxifying enzymes like cytochrome P-450 enzymes (CYP) can be induced, with catalysis of the oxidation of organic substances. These two reactions cause a modulation of the anabolism and catabolism of endogenous VOCs and thus a change in cellular VOC contents. Alkanes are mainly produced through lipid peroxidation by ROS and detoxifying enzymes cause alkane hydroxylation resulting in the production of alcohols (Hakim et al., 2012; Ortiz de Montellano and Voss, 2005). Cellular contents of this compound family in animal liver may thus be very relevant for back-tracing HBCD exposure. Regarding the levels of hydrocarbon markers identified here, 2,2,4,6,6-pentamethylheptane and 2,2,4,4-tetramethyloctane increased in the liver of HBCD-exposed hens while pentane, 2methylpentane and methylcyclopentane decreased in the same samples.
Fig. 4. Time-course change in balance between the volatile organic compounds determined as liver markers of hen exposure to 10 mg g-HBCD g¡1 feed. PCA was performed on the VOC markers of animal exposure to 10 mg HBCD g1 feed after filtering by ANOVA (stage of exposure; p < 0.05) with 6 modalities (day 1 to day 70).
640
J. Ratel et al. / Chemosphere 189 (2017) 634e642
Table 1 Candidate volatile markers to discriminate liver samples from control animals and from animals exposed to g-HBCD (0.1 and 10 mg HBCD/g feed). Compound Alkanes and branched-alkanes pentane 2-methylpentane methylcyclopentane 2,2,4,6,6-pentamethylheptane 2,2,4,4-tetramethyloctane Alcohols 2-methyl-3-buten-2-ol 3-methyl-1-butanol Aldehydes hexanal dimethylbenzaldehyde Ketones 3-cyclohepten-1-one 6-methyl-2-heptanone 1-octen-3-one 2-octanone Phenols phenol Alkyl benzenes ethylbenzene m-xylene o-xylene 1-methylethylbenzene Others 1,1-dichloroethane 2-methylfuran 1,2,5-thiadiazole thiazole unknown 1,2,4,4-tetramethylcyclopentene unknown 2,7-dimethyloxepine a-fenchene unknown unknown dihydro-5,5-dimethyl-2(3H)-furanone unknown unknown unknown unknown 2,2,6-trimethylcyclohexanone unknown unknown unknown unknown unknown 2,6,6-trimethyl-1,3-cyclohexadiene-1-carboxaldehyde b-cyclocitral unknown a b c
IDa
m/z
LRIb
Liver levelc
a,b a,b a,b a,b a,b
72 71 69 57 57
500 561 631 996 1036
þ þ
a,b a,b
71 70
613 737
þ þ
a,b a,b
56 105
800 1203
-
a,b a,b a,b a,b
54 58 70 128
827 955 978 991
-
a,b
94
978
-
a,b a,b a,b a,b
91 91 106 105
868 875 899 930
þ þ þ þ
a,b a,b a a,b
63 82 86 85 71 109 104 122 93 58 98 70 83 97 104 57 97 57 84 95 67 87 107 137 95
579 605 674 738 847 862 919 936 951 966 979 987 1011 1014 1015 1022 1045 1054 1070 1036 1109 1145 1216 1231 1246
-
a,b a,b a,b
a,b
a,b
a,b a,b
þ
-
Tentative identification based on (a) mass spectrum, (b) linear retention index from the literature or internal databank. Linear retention indices on a DB5 capillary column. Level found for tentatively identified compounds higher or lower in liver from animals exposed to HBCD compared to controls.
3.5.2. Alcohols 2-methyl-3-buten-2-ol and 3-methyl-1-butanol were identified as candidate markers of g-HBCD exposure with an increase in liver of exposed animals. Hakim et al. (2012) reported an alteration of the alcohol levels in different body fluids in response to exposure to toxic xenobiotics. They suggested that the metabolism of alcohols could be modified by hydroxylation reactions of hydrocarbons by the CYP450 enzymes or by oxidation reactions to aldehydes by the alcohol dehydrogenase (ADH). They have highlighted the particular interest of the primary and secondary alcohols as candidate markers for VOC-based clinical diagnostics. 3-methyl-1-butanol could be thus a particularly relevant candidate marker of g-HBCD exposure.
3.5.3. Aldehydes Hexanal and dimethylbenzaldehyde were identified as markers of HBCD exposure and decreased in the liver of HBCD-exposed animals. Aldehydes are also involved in the main cellular reactions induced in response to exposure to toxic xenobiotics (Hakim et al., 2012) and could thus be relevant for back-tracing exposure to g-HBCD. Note that hexanal has already been reported as one of the VOCs potentially related to cancer development with a decrease in the headspace of cancer cell lines (Filipiak et al., 2008; Sponring et al., 2009, 2010). According to the literature, aldehydes can be produced in the liver by the metabolization of alcohols induced by ADH or CYP enzymes and by lipid peroxidation through the reduction of hydroxyperoxides by CYP enzymes. They can be also oxidized by aldehyde dehydrogenase (ALDH) to yield carboxylic acids. Aldehyde metabolism could be thus tightly related to
J. Ratel et al. / Chemosphere 189 (2017) 634e642
enzyme induced by exposure to xenobiotics (Ortiz de Montellano and Voss, 2005; Mochalski et al., 2013; Jelski and Szmitkowski, 2008), which makes these compounds particularly relevant candidate markers of g-HBCD exposure. 3.5.4. Ketones 2-octanone, 1-octen-3-one, 6-methyl-2-heptanone, 3cyclohepten-1-one decreased in the liver of HBCD-exposed animals. Like the previous compound families, ketones are involved in hepatocellular reactions in response to xenobiotic exposure (Hakim et al., 2012) and could thus also make relevant candidate markers of exposure to g-HBCD. Ketones are formed in substantial quantities in liver under metabolic conditions similar to those that occur during a toxic exposure, i.e. with a high oxidation rate of fatty acids (Hakim et al., 2012). Protein metabolism has been also shown to be increased in some severe metabolic disorders with amino acid metabolism-induced ketone formation (Hakim et al., 2012). Ketones could thus make particularly relevant candidate markers of g-HBCD exposure. 3.5.5. Phenols Phenol decreased in the liver of HBCD-exposed animals. First, note that CYP-450 enzymes involved in the cellular reactions in response to toxic xenobiotic exposure have also been shown to be involved in the microsomal hydroxylation of phenol in rat liver (Sawahata and Neal, 1983). Second, phenol is derived from the catabolism of tyrosine and tryptophan, and the hepatic degradation of these aromatic amino acids has been demonstrated in human patients with liver function impairment (Dadamio et al., 2012). Liver response to g-HBCD exposure could thus contribute to the change in liver phenol content, which would make phenol a promising candidate marker of g-HBCD exposure. 3.5.6. Alkyl benzenes Four alkyl benzenes (ethylbenzene; m-xylene; o-xylene; methylethylbenzene) were identified as markers of g-HBCD exposure and all increased in the liver of HBCD-exposed animals. The presence of aromatic compounds has already been highlighted in different human clinical samples during serious metabolic disorders. Styrene for example was found in human breath as a marker of lung cancer or liver cirrhosis (Hakim et al., 2012; Dadamio et al., 2012), and methylated benzenes, such as 1,3-dimethylbenzene, ethylbenzene, p-xylene and o-xylene, increased in human melanoma samples compared to controls (Abaffy et al., 2010). These compounds are functionalized usually by CYP-450 enzymes before being conjugated to a more excretable form by other enzyme systems (Hakim et al., 2012). Thus even if these aromatic compounds are considered as deriving from exogenous sources and not from the metabolome (Hakim et al., 2012), the level of these compounds could be impacted by g-HBCD exposure which could modify the activity of detoxifying enzymes such as CYP-450 enzymes. 4. Conclusion These results confirm the informative potential of the liver volatolome for exploring food contamination issues like exposure of hens to g-HBCD, and the relevance of the SPME-GC-MS approach for pointing out candidate markers in the volatolome of animal derived-food products. Recent advances in VOC-sensors (Konvalina and Haick, 2013) could pave the way to the development of routine controls based on an easily and rapidly measurement of the markers in animal tissues or food products. This proof of concept study enabled a screening of VOC markers of animal exposure to gHBCD, but confounding factors cannot be excluded and some complementary studies have to be undertaken. Efforts should turn
641
to assessing robustness and specificity of the candidate markers through the implementation of in vitro assays (Engel et al., 2015), and the analysis of on-market contaminated samples. Another priority prospect of the present study is to assess performance criteria of volatolomic analysis, especially in determining the detection limits for the candidate markers and the range of exposure level as well as time length. As a follow-up to this proof of concept study, statistical robustness will need to be improved for instance by implementing a power analysis that allows for better evaluation of inter-animal variability. Finally, together with the combination of the volatolome contained in different animal fluids, tissues or organs (Amann et al., 2014; Broza et al., 2015), the investigation of the relative balance between VOC markers (Lehallier et al., 2012) might enable to deal with the issue of diagnostic specificity. Acknowledgments The authors are grateful to the French Ministry for Food, Agriculture and Fisheries for financial support from the Compte ciale De veloppement Agricole et Rural (CASDAR d'Affectation Spe project n 7106). They thank N. Besne (PEAT, INRA Val de Loire, Nouzilly, France) for experimental feed manufacturing, P. Hart de Lorraine, meyer and C. Grandclaudon (URAFPA, INRA, Universite Nancy, France) for technical support. Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.chemosphere.2017.09.074. References Abaffy, T., Duncan, R., Riemer, D.D., Tietje, O., Elgart, G., Milikowski, C., DeFazio, R.A., 2010. Differential volatile signatures from skin, naevi and melanoma: a novel approach to detect a pathological process. PLoS One 5, e13813. Amann, A., Mochalski, P., Ruzsanyi, V., Broza, Y.Y., Haick, H., 2014. Assessment of the exhalation kinetics of volatile cancer biomarkers based on their physicochemical properties. J. Breath. Res. 8, 016003. Berge, P., Ratel, J., Fournier, A., Jondreville, C., Feidt, C., Roudaut, B., Le Bizec, B., Engel, E., 2011. Use of volatile compound metabolic signatures in poultry liver to back-race dietary exposure to rapidly metabolized xenobiotics. Environ. Sci. Tech. 45, 6584e6591. Birnbaum, L.S., Staskal, D.F., 2004. Brominated flame retardants: cause for concern? Environ. Health Persp 112, 9e17. Blake, A., 2005. Next generation of POPs: PBDEs and Lindane. IPEN. za, E., Jondreville, C., Bouhlel, J., Ratel, J., Abouelkaram, S., Mercier, F., Travel, A., Bae Dervilly-Pinel, G., Marchand, P., Le Bizec, B., Dubreil, E., Mompelat, S., Verdon, E., rin, T., Rutledge, D.N., Engel, E., 2017. Solid-phase microInthavong, C., Gue extraction set-up for the analysis of liver volatolome to detect livestock exposure to micropollutants. J. Chromatogr. A 1497, 9e18. Broza, Y.Y., Mochalski, P., Ruzsanyi, V., Amann, A., Haick, H., 2015. Hybrid volatolomics and disease detection. Angew. Chem. Int. Ed. 54, 11036e11048. Covaci, A., Gerecke, A.C., Law, R.J., Voorspoels, S., Kohler, M., Heeb, N.V., Leslie, H., Allchin, C.R., de Boer, J., 2006. Hexabromocyclododecanes (HBCDs) in the environment and humans: a review. Environ. Sci. Tech. 40, 3679e3688. Covaci, A., Roosens, L., Dirtu, A.C., Waegeneers, N., Van Overmeire, I., Neels, H., Goeyens, L., 2009. Brominated flame retardants in Belgian home-produced eggs: levels and contamination sources. Sci. Total Environ. 407, 4387e4396. Dadamio, J., Van den Velde, S., Laleman, W., Van Hee, P., Coucke, W., Nevens, F., Quirynen, M., 2012. Breath biomarkers of liver cirrhosis. J. Chromatogr. B 905, 17e22. Dominguez-Romero, E., Cariou, R., Omer, E., Marchand, P., Dervilly-Pinel, G., Le Bizec, B., Travel, A., Jondreville, C., 2016. Tissue distribution and transfer to eggs of ingested a-hexabromocyclododecane (a-HBCDD) in laying hens (Gallus domesticus). J. Agric. Food Chem. 64, 2112e2119. EFSA, 2011. Scientific opinion on hexabromocyclododecanes (HBCDDs) in food. EFSA Panel Contam. Food Chain (CONTAM) EFSA J. 2011 (9), 118. Engel, E., Ratel, J., Bouhlel, J., Planche, C., Meurillon, M., 2015. Novel approaches to improving the chemical safety of the meat chain towards toxicants. Meat Sci. 109, 75e85. European Commission 2014/118/EU, 2014. Commission Recommendation of 3 March 2014 on the Monitoring of Traces of Brominated Flame Retardants in Food (Text with EEA Relevance) (2014/118/EU). Off. J. Euro. Union. Fernie, K.J., Marteinson, S.C., Bird, D.M., Ritchie, I.J., Letcher, R.J., 2011. Reproductive
642
J. Ratel et al. / Chemosphere 189 (2017) 634e642
changes in American kestrels (Falco sparverius) in relation to exposure to technical hexabromocyclododecane flame retardant. Environ. Toxicol. Chem. 30, 2570e2575. Filipiak, W., Sponring, A., Mikoviny, T., Ager, C., Schubert, J., Miekisch, W., Amann, A., Troppmair, J., 2008. Release of volatile organic compounds (VOCs) from the lung cancer cell line CALU-1 in vitro. Cancer Cell Int. 8, 17. Focant, J.F., 2014. Keeping tabs on dioxins. Anal. Sci. 15, 34e40. nisseau, A., Le Bizec, B., Sellier, N., Engel, E., Fournier, A., Feidt, C., Marchand, P., Ve Ratel, J., Travel, A., Jondreville, C., 2012. Kinetic study of g-hexabromocyclododecane orally given to laying hens (Gallus domesticus). Environ. Sci. Pollut. Res. 19, 440e447. Goscinny, S., Vandevijvere, S., Maleki, M., Overmeire, I.V., Windal, I., Hanot, V., Blaude, M.N., Vleminckx, C., Loco, J.V., 2011. Dietary intake of hexabromocyclododecane diastereoisomers (a-, b-, and g-HBCD) in the Belgian adult population. Chemosphere 84, 279e288. Haick, H., Broza, Y.Y., Mochalski, P., Ruzsanyibc, V., Amann, A., 2014. Assessment, origin, and implementation of breath volatile cancer markers. Chem. Soc. Rev. 43, 1423e1449. Hakim, M., Broza, Y.Y., Barash, O., Peled, N., Phillips, M., Amann, A., Haick, H., 2012. Volatile organic compounds of lung cancer and possible biochemical pathways. Chem. Rev. 112, 5949e5966. Hiebl, J., Vetter, W., 2007. Detection of hexabromocyclododecane and its metabolite pentabromocyclododecene in chicken egg and fish from the official food control. J. Agric. Food Chem. 55, 3319e3324. Jelski, W., Szmitkowski, M., 2008. Alcohol dehydrogenase (ADH) and aldehyde dehydrogenase (ALDH) in the cancer diseases. Clin. Chim. Acta 395, 1e5. €tters, T., Rupp, R., Sures, B., 2015. Review of hexKoch, C., Schmidt-Ko abromocyclododecane (HBCD) with a focus on legislation and recent publications concerning toxicokinetics and -dynamics. Environ. Poll. 199, 26e34. Konvalina, G., Haick, H., 2013. Sensors for breath testing: from nanomaterials to comprehensive disease detection. Acc. Chem. Res. 7, 66e76. Lehallier, B., Ratel, J., Hanafi, M., Engel, E., 2012. Systematic ratio normalization of gas chromatography signals for biological sample discrimination and biomarker discovery. Anal. Chim. Acta 733, 16e22. Letcher, R.J., Mattioli, L.C., Marteinson, S.C., Bird, D., Ritchie, I.J., Fernie, K.J., 2015. Uptake, distribution, depletion, and in ovo transfer of isomers of hexabromocyclododecane flame retardant in diet-exposed American kestrels (Falco sparverius). Environ. Toxicol. Chem. 34, 1103e1112. Mochalski, P., Sponring, A., King, J., Unterkofler, K., Troppmair, J., Amann, A., 2013. Release and uptake of volatile organic compounds by human hepatocellular
carcinoma cells (HepG2) in vitro. Cancer Cell Int. 13, 72. Morris, S., Bersuder, P., Allchin, C.R., Zegers, B., Boon, J.P., Leonards, P.E., de Boer, J., 2006. Determination of the brominated flame retardant, hexabromocyclodocane, in sediments and biota by liquid chromatographyelectrospray ionisation mass spectrometry. Trac. Trend Anal. Chem. 25, 343e349. Ortiz de Montellano, P., Voss, J., 2005. Substrate oxidation by cytochrome P450 enzymes. Cytochrome P450 183e245. Rawn, D., Sadler, A., Quade, S., Sun, W.-F., Lau, B.-Y., Kosarac, I., Hayward, S., Ryan, J., 2011. Brominated flame retardants in Canadian chicken egg yolks. Food Addit. Contam. 28, 807e815. Sawahata, T., Neal, R.A., 1983. Biotransformation of phenol to hydroquinone and catechol by rat liver microsomes. Mol. Pharmacol. 23, 453e460. Sivadier, G., Ratel, J., Bouvier, F., Engel, E., 2008. Authentication of meat products: determination of animal feeding by parallel GC-MS analysis of three adipose tissues. J. Agric. Food Chem. 56, 9803e9812. Sponring, A., Filipiak, W., Mikoviny, T., Ager, C., Schubert, J., Miekisch, W., Amann, A., Troppmair, J., 2009. Release of volatile organic compounds from the lung cancer cell line NCI-H2087 in vitro. Anticancer Res. 29, 419e426. Sponring, A., Filipiak, W., Ager, C., Schubert, J., Miekisch, W., Amann, A., Troppmair, J., 2010. Analysis of volatile organic compounds (VOCs) in the headspace of NCI-H1666 lung cancer cells. Cancer Biomark. 7, 153e161. Szabo, D.T., Diliberto, J.J., Hakk, H., Huwe, J.K., Birnbaum, L.S., 2010. Toxicokinetics of the flame retardant hexabromocyclododecane gamma: effect of dose, timing, route, repeated exposure, and metabolism. Toxicol. Sci. 117, 282e293. Tomy, G.T., Budakowski, W., Halldorson, T., Whittle, D.M., Keir, M.J., Marvin, C., MacInnis, G., Alaee, M., 2004. Biomagnification of a-and g-hexabromocyclododecane isomers in a Lake Ontario food web. Environ. Sci. Tech. 38, 2298e2303. bault, Mahe , Grammont, Travel, F., Marchand, Venisseau, Le Bouquin, Allain, The Badreddine, Jurjanz, Feidt, Jondreville, 2012. Transfert de polluants organiques tat des lieux, modalite s et facteurs de persistants vers l’œuf de consommation : e risques. Innov. Agron. 25, 17. van der Ven, L.T., van de Kuil, T., Leonards, P.E., Slob, W., Lilienthal, H., Litens, S., n, R.F., van den Berg, M., 2009. Endocrine effects of Herlin, Håkansson, H., Canto hexabromocyclododecane (HBCD) in a one-generation reproduction study in Wistar rats. Toxicol. Lett. 185, 51e62. Vasta, V., Ratel, J., Engel, E., 2007. Mass spectrometry analysis of volatile compounds in raw meat for the authentication of the feeding background of farm animals. J. Agric. Food Chem. 55, 4630e4639.
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Liver volatolomics to reveal poultry exposure to ghexabromocyclododecane (HBCD) re my Ratel a, Christelle Planche a, Fre de ric Mercier a, Patrick Blinet a, Je s Fournier c, Ange lique Travel d, Nathalie Kondjoyan a, Philippe Marchand b, Agne c a, * Catherine Jondreville , Erwan Engel INRA, UR QuaPA, F-63122 Saint-Gen es-Champanelle, France LUNAM Universit e, Oniris, LABERCA, F-44307 Nantes, France INRA, Universit e de Lorraine, UR AFPA, F-54500 Vandoeuvre-les-Nancy, France d INRA, ITAVI, F-37380 Nouzilly, France a
b c
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Poultry liver volatolome is modified by the dietary exposure of animals to g-HBCD. Volatile candidate markers in liver enable to reveal livestock exposure to g-HBCD. Candidate markers are mainly hydrocarbons or oxygenated compounds. Liver volatolomics is a promising approach to strengthen food contamination control.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 June 2017 Received in revised form 8 September 2017 Accepted 15 September 2017 Available online 17 September 2017
Hexabromocyclododecane (HBCD) is a critical emerging brominated flame retardant to which consumers can be exposed at high doses through a single food intake. Based on an animal experiment involving 3 groups of laying hens fed during 70 days with a control diet or g-HBCD-contaminated diets at 0.1 or 10 mg g-HBCD g1 feed, this study aims to use the volatolome of biological samples for revealing markers of livestock exposure to HBCD. Liquid chromatographyetandem mass spectrometry was used to monitor the time-course of HBCD levels in bodily samples. Each liver was analyzed by solid-phase microextractionegas chromatographyemass spectrometry for volatolome profiling. After 70 days, g-HBCD concentrations in egg yolk, fat, liver and serum reached 54 ± 4, 85 ± 6, 31 ± 6, and 32 ± 4 ng g1 lw, respectively, for the low exposure level and 4.6þ/5.7, 7.8þ/6.5, 3.9þ/3.0 and 3.9þ/6.1 mg g1 lw, respectively, for the high exposure level. Isomerization of g-HBCD into a- and b-HBCD was observed in all tissues, at least for the high exposure level. Volatolome data allowed a significant discrimination between control and exposed animals whatever the feed contamination load, demonstrating a liver metabolic response to g-HBCD exposure. The relevance of the twenty nine volatile exposure markers tentatively identified was discussed in light of literature data. © 2017 Elsevier Ltd. All rights reserved.
Handling Editor: Frederic Leusch Keywords: Food safety g-HBCD Animal-derived food Volatolomics Liver SPME-GC-MS
* Corresponding author. E-mail address: [email protected] (E. Engel). https://doi.org/10.1016/j.chemosphere.2017.09.074 0045-6535/© 2017 Elsevier Ltd. All rights reserved.
J. Ratel et al. / Chemosphere 189 (2017) 634e642
1. Introduction Brominated flame retardants (BFRs) are persistent organic pollutants that are increasingly contaminating the environment, wildlife, and people (Birnbaum and Staskal, 2004). Since 2014, the European Commission recommends monitoring foodstuffs for BFRs (European recommendation 2014/118/EU, 2014). As critical emerging contaminant, hexabromocyclododecane (HBCD) is a BFR that recently joined the list of chemicals targeted by the Stockholm Convention (Covaci et al., 2006). This pollutant is now part of Annex A which requires signatories to take measures to eliminate the production and use of HBCD due to the dangers for human health and the environment. Indeed oral exposure to HBCD causes a broad range of adverse effects like endocrine effects, drug-metabolizing enzyme disorders or cancer occurrence by a nonmutagenic mechanism (Covaci et al., 2006; van der Ven et al., 2009). Human exposure to HBCD occurs through multiple routes, but the route of most concern is exposure through diet (Goscinny et al., 2011). In Europe, median chronic dietary exposure estimated attains 0.43 ng kg1 day1 in adults and 1.27 ng kg1 day1 in children (EFSA, 2011). EFSA concluded that current median dietary exposure to HBCD does not raise a health concern given the benchmark dose lower confidence limit for a benchmark response of 10% (BMDL10) of 790,000 ng kg1 day1 (EFSA, 2011). However, HBCD concentrations in animal-derived food products are reported to be subjected to wide variabilities. For example, the literature reports HBCD concentrations as high as 62 (Covaci et al., 2009), 72 (Rawn et al., 2011), 90 (Blake, 2005) or even 2000 (Hiebl and Vetter, 2007) ng g1 lipid in chicken eggs. According to surveillance programs led by the French Directorate General for Food in France in 2008 and 2009, the extreme values reached 3055 ng g1 lipid in poultry meat and 3390 ng g1 in eggs (Travel et al., 2012). Thus, the ingestion of a single contaminated animal-derived food can significantly increase the consumer exposure to HBCD. The gold standard techniques for measuring HBCD in food are LC/MS (Morris et al., 2006) or LC/MS/MS (Tomy et al., 2004), because these approaches allow to determine the different HBCD isomers. However, the demands of an ambitious monitoring program for food safety required the means and capacity to perform measurements that were precise, accurate, robust but also costeffective and not cumbersome (Focant, 2014), and direct HBCD measurement is particularly difficult to implement due to the ubiquitous presence of this contaminant in the environment (Covaci et al., 2006). This prevents the setting up of rapid controls in the production chain to strengthen food safety. Emerging toxicogenomics approaches have been proposed on the basis of the discovery of metabolite markers of contamination (Engel et al., 2015). Research has evidenced that some biological signatures of the human body can be used to reveal the significant perturbations that link exposure and toxicity pathways, such as the use of volatile organic compounds (VOCs) from exhaled air of patients for rapid non-invasive diagnosis of cancers (Hakim et al., 2012). Intensive research has been conducted on the potential use of volatile metabolites emanating from cells and their microenvironment, named “volatolome”, to diagnose pathologies in humans (Broza et al., 2015). Pathophysiologic disorders can impact the composition of the volatolome of living cells or tissues, with a production of new volatile metabolites or a change in the ratio of volatile metabolites normally produced by the body (Hakim et al., 2012). Recently, Berge et al. (2011) evidenced that the exposure of animals to toxic xenobiotics generated an altered pattern of VOC content in their liver, which plays a key role in detoxification function, and they proposed to use this volatolomic signature to reveal a contamination of the food chain. The present study aimed to monitor the time-course of HBCD
635
levels in different tissues of animals dietary exposed to HBCD, and to complete the proof of concept study carried out by Berge et al. (2011) in identifying candidate volatile markers of HBCD chemical contamination. To address this issue, we designed an animal experiment on laying hens dietary exposed or not to g-HBCD, which is the most present HBCD isomer in technical mixtures. 2. Materials and methods 2.1. Feed contamination In the study by Fournier et al. (2012), hens given a diet containing 1 ng g-HBCD g1, laid eggs containing 0.4 ng HBCD g1 lw. Thus, the dose was increased up to 0.1 and 10 mg g1 feed in order to achieve concentrations in eggs consistent with the several tens and thousands recorded in several surveys (Covaci et al., 2009; Rawn et al., 2011; Blake, 2005; Hiebl and Vetter, 2007; Travel et al., 2012). Contaminated diets were prepared from the same maize and soybean meal basal feed including all the nutrient requirements of laying hens. g-HBCD (1,2,5,6,9,10hexabromocyclododecane) obtained from Sigma Aldrich (St. Louis, MO) was dissolved in rapeseed oil then mixed with the basal feed according to Fournier et al. (2012) for spiked feed preparation. Non-contaminated feed was prepared by mixing the same amount of blank rapeseed oil with the basal feed. Feeds were administered as mash. 2.2. Animal test The experiment was conducted under the application of the Directive 2010/63/EU in France. Fifty-six laying hens (Lohmann Brown, 43-week-old; weight 1.77 ± 0.16 kg) were placed in an appropriate facility (agreement n B54-547-15) with individual cages allowing the individual control of feed ingestion and of egg production. Two groups of 24 animals were exposed to the two levels of HBCD. After a one-week adaptation period, the hens were given the contaminated diets during 70 days. The remaining 8 hens were used as a control group given the non-contaminated feed. The daily allowance of feed was adjusted to 6.9% of body weight (bw) and given once in the morning, water was continuously available, room temperature was maintained at 24 C and lighting was set to a 16 h light/8 h dark cycle. HBCD-exposed laying hens were sequentially slaughtered after 1, 12, 21, 36, 54 or 70 days of exposure. Control laying hens were slaughtered the day before the experiment start (n ¼ 4) or on day 70 (n ¼ 4). Hens were slaughtered after a 12-h fasting period by electronarcosis followed by exsanguination. 2.3. Sampling The amount of feed ingested and the number and weight of eggs laid were individually recorded daily. Each feed was weekly sampled and stored at þ4 C. Blood was collected by puncture at the jugular vein in hens slaughtered after 1, 21 and 70 days of exposure, then it was allowed clotting during 24 h at room temperature before serum was prepared by centrifugation (1400 g, 15 min) and stored at 20 C. At slaughter, each hen was weighed, and samples (egg, abdominal fat and liver) were collected and weighed. Egg yolks were freeze-dried, stored at 20 C and finely ground before analysis. Abdominal fat and liver samples were immersed in liquid nitrogen, wrapped in aluminum foil, vacuumpacked and stored at 80 C. Fat and liver were ground for 1 and 3 min, respectively, in liquid nitrogen into a fine homogeneous powder using a home-made stainless steel ball mill. For the analysis of liver volatolome, a 1.2 g aliquot of each powdered liver was
636
J. Ratel et al. / Chemosphere 189 (2017) 634e642
placed in a glass vial (Supelco, Sigma-Aldrich, St. Louis, MO), sealed under a nitrogen flow, and stored at 80 C. 2.4. HBCD quantification HBCD isomers (a, b and g) were quantified by Laberca (ONIRIS, France), the French National Reference Laboratory for most environmental organic contaminants in food. Briefly, after freeze drying, HBCDs were extracted from all matrices by Accelerated Solvent Extraction (ASE300, Dionex, Sunnyvale, CA) with toluene and acetone (70:30 v/v). The extracts were cleaned up using a multilayer silica column (from top to bottom: anhydrous sodium sulfate (5 g), concentrated sulfuric acid acidified silica (25 g), neutral silica (5 g)). HBCDs were assayed by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) according to Fournier et al. (2012). For a-, b- and g-HBCD, limits of quantification (LOQ) were respectively 0.013, 0.004, and 0.038 ng g1 fresh matter (fm) in feed, 0.24, 0.11 and 0.12 ng g1 lipid weight (lw) in liver, 0.045, 0.046 and 0.006 ng g1 lw in abdominal fat, 0.037, 0.013 and 0.037 ng g1 lw in egg yolk and 0.56, 0.15 and 1.1 ng g1 lw in serum. 2.5. Liver volatolome analysis Volatolome of powdered liver samples was analyzed by solidphase microextraction (SPME) coupled to gas chromatography and mass spectrometry (GC-MS) according to Berge et al. (2011). Briefly, the following steps were carried out with an automated sampler (AOC-5000 Shimadzu, Japan): (i) preheating of the sample for 10 min at 60 C in the agitator (500 rpm), (ii) SPME trapping (75 mm carboxen/polydimethylsiloxane, 23-gauge needle, Supelco) of the volatile organic compounds (VOCs) for 30 min at 60 C, and (iii) thermal desorption at 250 C for 2 min in the GC inlet. Further VOC analysis was performed by GC/MS-full scan (GC2010; QP2010þ, Shimadzu). VOCs were injected in splitless mode into a DB-5MS capillary column (60 m 0.32 mm 1 mm; Agilent J&W). Oven temperature was held at 40 C for 5 min, then ramped up to 230 C at a gradient of 3 C.min1, and held at 230 C for 10 min. Temperature of the transfer line between GC and MS was set at 230 C. Temperature was fixed at 180 C in the MS source and 150 C in the MS quadrupole. Electron impact energy was set at 70 eV, and data was collected in the range of 33e250 m/z at a scan range of 10 scans per second. First, SPME-GC-MS chromatograms obtained in full-scan mode were converted into virtual SPME-MS fingerprints to get a quick estimate the informative potential of the volatile metabolome. Second, tentative identification of volatiles was performed from the SPME-GC-MS signals on the basis of both mass spectra, by comparison against the NIST 14 mass spectral library, and retention indices (RI), by comparison against published RI values and those of our internal database. Peak area of the tentatively-identified compounds was integrated with the GC-MS solution software (version 2.53, Shimadzu) using a mass fragment selected as being specific and free of any coelution. 2.6. Data treatment 2.6.1. Body weight, egg production and HBCD quantification Statistical analyses were performed by means of the Statistical Analysis Systems software package (version 9.1, SAS Institute). The hen was considered as the experimental unit. Differences were considered significant for p < 0.05. First, the absence of impact of exposure to g-HBCD on body weight, feed ingested and the number and mass of eggs was checked as described in supplementary material (Table S1). In order to facilitate the statistical comparison
between the two levels of exposure, despite the 100-fold difference between the two feed HBCD concentrations, the concentrations of each isomer in matrices were divided by 100 for hens given the feed containing 10 mg g-HBCD g1 feed. The resulting normalized concentrations were analyzed as repeated measures according to a MIXED procedure as described by Dominguez-Romero et al. (2016). The model included the date of sample collection (time, n ¼ 6), the level of feed contamination (0.1 or 10 mg g1 feed, n ¼ 2), the animal matrix as repeated measure (egg yolk, liver, abdominal fat, serum, n ¼ 4) and their interactions. The animal matrix was introduced in the model using the “repeated” statement, in order to account for the correlation between the tissues collected from the same hen. Compound symmetry was used as covariance structure, thus measurements on the same hen were considered to be similarly correlated. 2.6.2. Liver volatolome Data were processed using Statistica software (version 12, StatSoft) and R software (version 3.3.2., http://www.R-project.org/). The variables used in data treatments were the mass fragments for virtual SPME-MS fingerprints and the VOCs for SPME-GC-MS signals. Datasets were normalized using the systematic ratio normalization (SRN) procedure developed by Lehallier et al. (2012) to improve both sample set discrimination and biomarker discovery. The procedure consists in calculating all the log-ratios between variables for each sample, then selecting the log-ratio(s) that best maximize the discrimination between sample-sets by one-way ANOVA at 1% level of significance corrected by the Bonferroni post-test. Principal component analyses (PCA) were performed on the discriminant ratios selected according to the animal group factor. To assess the test-long time-course of the discriminant variables, ANOVA (stage of exposure; p < 0.05) with 6 modalities (the timepoints from day 1 to day 70) was performed on the SRN data provided by analysis of livers from the animals exposed to 10 mg HBCD g1 feed. 3. Results and discussion 3.1. HBCD in spiked feed The concentrations of a- and b-HBCD were below the LOQ in the control feed, while the concentration of g-HBCD was 0.038 ng g1. In the contaminated feed intended to contain 0.1 mg g-HBCD g1, compliant concentrations of 0.7, 0.5 and 91.6 ng a-, b- and g-HBCD g1, respectively, were measured. No control of the target concentration of 10 mg g-HBCD g1 was performed to avoid any contamination of the analytical device. 3.2. Feed intake, body weight and laying Results are presented in supplementary material (Table S1). Hens exposed more than one day ingested daily 6.11 or 611 mg gHBCD kg1 bw. Body weight, feed intake and laying rate were very similar in control and in exposed hens and were independent of the dietary level of g-HBCD (p > 0.1). Similarly, a 75-day exposure of American kestrels to 510 mg technical HBCD kg1 bw did not alter body characteristics of animals, even if their reproductive performance was disturbed with more numerous but 3% lighter eggs and an increased eggshell porosity (Fernie et al., 2011). 3.3. HBCD in liver, fat, serum and egg yolk HBCD isomers were quantified in bodily samples collected from control hens (Table S2). As expected, concentrations of g-HBCD measured in egg yolk, liver and fat were low, although these
J. Ratel et al. / Chemosphere 189 (2017) 634e642
concentrations tended to increase with time possibly due to a contamination through dust originating from contaminated feed. Fig. 1 reports the experiment-long time-course of the sum of the three HBCD isomers concentrations determined by LC-MS/MS in the tissues of the animals exposed to 0.1 (Fig. 1A) and 10 (Fig. 1B) mg
637
g-HBCD/g feed. The results of the statistical analysis performed on standardized concentrations (Table S3) indicate that the overall increase of HBCD concentration with time in all four tissues (Time, p < 0.001) was independent of the level of feed contamination (Level and Time*Level, p > 0.1). The proportionality of the dose and
Fig. 1. HBCD (sum of a-, b- and g-) quantification by LC-MS/MS in samples from animals exposed to (A) 0.1 and (B) 10 mg HBCD g1 feed. Values are adjusted means (n ¼ 4) ± SE.
638
J. Ratel et al. / Chemosphere 189 (2017) 634e642
the concentration in tissues was already observed in mice orally exposed to doses ranging between 3 and 100 mg g-HBCD kg1 bw (Szabo et al., 2010). Overall, abdominal fat was the most concentrated tissue, and liver and serum the least concentrated, while egg yolk was intermediate (Tissue, p < 0.001). However, time trends differed between tissues (Time*Tissue, p < 0.001): while concentrations were rapidly close to their maximum value in liver and in serum, they increased more progressively in egg yolk and in abdominal fat. At the end of the experiment, the low contaminated diets resulted in HBCD concentrations of 53.6 ± 3.5, 85.3 ± 5.6, 30.5 ± 5.7 and 32.1 ± 3.8 ng g1 lw in egg yolk, abdominal fat, in liver and in serum, respectively. For the high contaminated diet they reached 4.6 ± 0.6, 7.8 ± 0.6, 3.9 ± 0.3 and 3.9 ± 0.6 mg g1 lw, respectively. These values in egg are consistent with those recorded in some surveys (Covaci et al., 2009; Rawn et al., 2011; Blake, 2005; Hiebl and Vetter, 2007; Travel et al., 2012). A similar ranking of tissues was observed in laying hens given feed spiked with a- or gHBCD (Fournier et al., 2012; Dominguez-Romero et al., 2016), and also in American kestrels exposed to a technical mixture of HBCD (Letcher et al., 2015). The well documented izomerization of g-HBCD into a- and bHBCD (Koch et al., 2015) was evidenced in the current study as the three isomers were detected and quantified in all tissues, except bHBCD in liver and in serum from hens exposed to 0.1 mg g-HBCD g1 feed. Time dependent concentrations of the three isomers are presented in Table S4 and in Fig. S1. In accordance with previous results (Fournier et al., 2012), the g-isomer was dominant in all tissues, making around 95% of total HBCD, while a- and b-isomers accounted for around 4.5% and 0.5% of total HBCD, respectively. 3.4. Response of the liver volatolome to g-HBCD exposure 3.4.1. SPME-MS fingerprinting of the volatolome Fig. 2 presents the projection of the two most discriminant ratios of mass fragments in the volatolome fingerprints of hen livers at the maximum HBCD exposure (70 days). The clear distinction of exposed animal groups (0.1 and 10 mg g-HBCD g1 feed) from the control group indicates that dietary contamination by HBCD
triggered changes in liver metabolism and that these changes were perceptible through the volatolome fingerprint. This result is consistent with several previous studies showing that volatolome fingerprints in different animal body compartments were relevant for differentiating animals according to various factors, such as type of animal feeding (Vasta et al., 2007; Sivadier et al., 2008) or occurrence of pollutants in the feed (Berge et al., 2011; Bouhlel et al., 2017). As shown in Fig. 2, separation between the two groups of exposed hens was limited. This may be due to the low resolution of the SPME-MS approach which focuses on the most abundant constituents of the volatolome (Berge et al., 2011). To better explore the information content of liver volatolome, it is necessary to increase the resolution and to reprocess the SPME-GCMS signals of these same samples. 3.4.2. SPME-GC-MS profiling of the volatolome The detailed composition of the liver volatolome was investigated by SPME-GC-MS in both control and 70-day exposed animals, in which 237 volatile organic compounds were determined. Systematic ratio normalization data treatment (Lehallier et al., 2012) revealed 44 ratios discriminating the 3 groups of animals. Fig. 3 reports the first map of PCA performed on the corresponding dataset, and shows that using SPME-GC-MS profiling allows the three groups to be separated, and even two contamination levels to be separated, which was not possible using SPME-MS fingerprints. Table S5 points out the loadings for the first factorial plane. This result shows that the resolution gain provided by the chromatographic separation made it possible to unmask dose-sensitive information in the liver volatolome. Liver volatolome signature was thus significantly different between HBCD-exposed hens and control hens. This change in liver volatolome in response to HBCD exposure is consistent with the results of the proof of concept study carried out by Berge et al. (2011) dealing in particular with exposure to polybrominated diphenyl ethers (PBDEs) which are another well-known class of brominated flame retardants. Berge et al. (2011) proposed that the occurrence of these toxicants, which are known to be quickly metabolized after oral intake, generates a metabolic response in the liver and consequently changes the liver
1,9
Mass fragment ratio 2
1,7
1,5
1,3 0,45
0,55
0,65
0,75
0,85
0,95
Mass fragment ratio 1 Fig. 2. Discrimination of hens exposed or not to HBCD according to virtual SPME-MS fingerprint of the liver volatolome. The first map was plotted on the two most discriminant ratios of mass fragments discriminative of the 3 groups of animals. Control; 0.1 mg g-HBCD g1 feed; 10 mg g-HBCD g1 feed.
J. Ratel et al. / Chemosphere 189 (2017) 634e642
639
3.5. Liver VOC markers of g-HBCD exposure SRN data treatment revealed 43 volatile organic compounds, whose 29 were tentatively identified (Table 1). Among these 29 VOCs, 9 were overexpressed in samples from exposed animals whereas 15 were underexpressed in samples from exposed animals compared to controls. These VOCs which could be considered candidate markers of animal exposure to g-HBCD belong mainly to the following compound families: alkanes and branched-chain alkanes, alcohols, aldehydes, ketones, phenols and alkyl benzenes.
Principal component 2 (9.3%)
3
1
-1
-3 -10
-6
-2
2
6
Principal component 1 (70.1%) Fig. 3. Discrimination of hens exposed or not to HBCD using volatile organic markers of liver volatolome. First map of normed PCA was plotted on the 44 ratios of VOCs discriminative of the 3 groups of animals. Control; 0.1 mg g-HBCD g1 feed; 10 mg g-HBCD g1 feed.
volatolome, and g-HBCD is known to be rapidly biotransformed and eliminated (Koch et al., 2015). To assess the time-course change in the balance between discriminative variables, we highlighted those which are impacted by stage of exposure. When we considered the animal group with the highest HBCD load (10 mg HBCD g1 feed), 15 ratios were found influenced by the stage of exposure. The first map of the PCA (Fig. 4) shows their evolution in the liver of hens exposed to the high-level dose of HBCD during the animal test. Animal samples are fairly well separated, showing that part of the discriminative information for revealing HBCD exposure is thus significantly influenced by stage of exposure. More specifically, these ratios are able to separate certain periods of exposure (1e12, 21, 36e54 and 70 days). The liver volatolome thus appears a very promising focus of efforts to detect hen contamination by g-HBCD.
3.5.1. Alkanes and branched-alkanes Pentane, 2-methylpentane, methylcyclopentane, 2,2,4,6,6pentamethylheptane and 2,2,4,4-tetramethyloctane were identified in animal liver as markers of HBCD exposure. Alkanes and methylated alkanes have been reported at the crossroads of human cell metabolism elicited in response to exposure to toxic xenobiotics (Hakim et al., 2012; Haick et al., 2014; Amann et al., 2014). According to the literature, exposure to toxic xenobiotics induces two main cellular reactions (Hakim et al., 2012). First, oxidative stress can be induced, with synthesis of reactive oxygen species (ROS) that leak from the mitochondria or from peroxidated polyunsaturated fatty acids in the cell membranes. Second, detoxifying enzymes like cytochrome P-450 enzymes (CYP) can be induced, with catalysis of the oxidation of organic substances. These two reactions cause a modulation of the anabolism and catabolism of endogenous VOCs and thus a change in cellular VOC contents. Alkanes are mainly produced through lipid peroxidation by ROS and detoxifying enzymes cause alkane hydroxylation resulting in the production of alcohols (Hakim et al., 2012; Ortiz de Montellano and Voss, 2005). Cellular contents of this compound family in animal liver may thus be very relevant for back-tracing HBCD exposure. Regarding the levels of hydrocarbon markers identified here, 2,2,4,6,6-pentamethylheptane and 2,2,4,4-tetramethyloctane increased in the liver of HBCD-exposed hens while pentane, 2methylpentane and methylcyclopentane decreased in the same samples.
Fig. 4. Time-course change in balance between the volatile organic compounds determined as liver markers of hen exposure to 10 mg g-HBCD g¡1 feed. PCA was performed on the VOC markers of animal exposure to 10 mg HBCD g1 feed after filtering by ANOVA (stage of exposure; p < 0.05) with 6 modalities (day 1 to day 70).
640
J. Ratel et al. / Chemosphere 189 (2017) 634e642
Table 1 Candidate volatile markers to discriminate liver samples from control animals and from animals exposed to g-HBCD (0.1 and 10 mg HBCD/g feed). Compound Alkanes and branched-alkanes pentane 2-methylpentane methylcyclopentane 2,2,4,6,6-pentamethylheptane 2,2,4,4-tetramethyloctane Alcohols 2-methyl-3-buten-2-ol 3-methyl-1-butanol Aldehydes hexanal dimethylbenzaldehyde Ketones 3-cyclohepten-1-one 6-methyl-2-heptanone 1-octen-3-one 2-octanone Phenols phenol Alkyl benzenes ethylbenzene m-xylene o-xylene 1-methylethylbenzene Others 1,1-dichloroethane 2-methylfuran 1,2,5-thiadiazole thiazole unknown 1,2,4,4-tetramethylcyclopentene unknown 2,7-dimethyloxepine a-fenchene unknown unknown dihydro-5,5-dimethyl-2(3H)-furanone unknown unknown unknown unknown 2,2,6-trimethylcyclohexanone unknown unknown unknown unknown unknown 2,6,6-trimethyl-1,3-cyclohexadiene-1-carboxaldehyde b-cyclocitral unknown a b c
IDa
m/z
LRIb
Liver levelc
a,b a,b a,b a,b a,b
72 71 69 57 57
500 561 631 996 1036
þ þ
a,b a,b
71 70
613 737
þ þ
a,b a,b
56 105
800 1203
-
a,b a,b a,b a,b
54 58 70 128
827 955 978 991
-
a,b
94
978
-
a,b a,b a,b a,b
91 91 106 105
868 875 899 930
þ þ þ þ
a,b a,b a a,b
63 82 86 85 71 109 104 122 93 58 98 70 83 97 104 57 97 57 84 95 67 87 107 137 95
579 605 674 738 847 862 919 936 951 966 979 987 1011 1014 1015 1022 1045 1054 1070 1036 1109 1145 1216 1231 1246
-
a,b a,b a,b
a,b
a,b
a,b a,b
þ
-
Tentative identification based on (a) mass spectrum, (b) linear retention index from the literature or internal databank. Linear retention indices on a DB5 capillary column. Level found for tentatively identified compounds higher or lower in liver from animals exposed to HBCD compared to controls.
3.5.2. Alcohols 2-methyl-3-buten-2-ol and 3-methyl-1-butanol were identified as candidate markers of g-HBCD exposure with an increase in liver of exposed animals. Hakim et al. (2012) reported an alteration of the alcohol levels in different body fluids in response to exposure to toxic xenobiotics. They suggested that the metabolism of alcohols could be modified by hydroxylation reactions of hydrocarbons by the CYP450 enzymes or by oxidation reactions to aldehydes by the alcohol dehydrogenase (ADH). They have highlighted the particular interest of the primary and secondary alcohols as candidate markers for VOC-based clinical diagnostics. 3-methyl-1-butanol could be thus a particularly relevant candidate marker of g-HBCD exposure.
3.5.3. Aldehydes Hexanal and dimethylbenzaldehyde were identified as markers of HBCD exposure and decreased in the liver of HBCD-exposed animals. Aldehydes are also involved in the main cellular reactions induced in response to exposure to toxic xenobiotics (Hakim et al., 2012) and could thus be relevant for back-tracing exposure to g-HBCD. Note that hexanal has already been reported as one of the VOCs potentially related to cancer development with a decrease in the headspace of cancer cell lines (Filipiak et al., 2008; Sponring et al., 2009, 2010). According to the literature, aldehydes can be produced in the liver by the metabolization of alcohols induced by ADH or CYP enzymes and by lipid peroxidation through the reduction of hydroxyperoxides by CYP enzymes. They can be also oxidized by aldehyde dehydrogenase (ALDH) to yield carboxylic acids. Aldehyde metabolism could be thus tightly related to
J. Ratel et al. / Chemosphere 189 (2017) 634e642
enzyme induced by exposure to xenobiotics (Ortiz de Montellano and Voss, 2005; Mochalski et al., 2013; Jelski and Szmitkowski, 2008), which makes these compounds particularly relevant candidate markers of g-HBCD exposure. 3.5.4. Ketones 2-octanone, 1-octen-3-one, 6-methyl-2-heptanone, 3cyclohepten-1-one decreased in the liver of HBCD-exposed animals. Like the previous compound families, ketones are involved in hepatocellular reactions in response to xenobiotic exposure (Hakim et al., 2012) and could thus also make relevant candidate markers of exposure to g-HBCD. Ketones are formed in substantial quantities in liver under metabolic conditions similar to those that occur during a toxic exposure, i.e. with a high oxidation rate of fatty acids (Hakim et al., 2012). Protein metabolism has been also shown to be increased in some severe metabolic disorders with amino acid metabolism-induced ketone formation (Hakim et al., 2012). Ketones could thus make particularly relevant candidate markers of g-HBCD exposure. 3.5.5. Phenols Phenol decreased in the liver of HBCD-exposed animals. First, note that CYP-450 enzymes involved in the cellular reactions in response to toxic xenobiotic exposure have also been shown to be involved in the microsomal hydroxylation of phenol in rat liver (Sawahata and Neal, 1983). Second, phenol is derived from the catabolism of tyrosine and tryptophan, and the hepatic degradation of these aromatic amino acids has been demonstrated in human patients with liver function impairment (Dadamio et al., 2012). Liver response to g-HBCD exposure could thus contribute to the change in liver phenol content, which would make phenol a promising candidate marker of g-HBCD exposure. 3.5.6. Alkyl benzenes Four alkyl benzenes (ethylbenzene; m-xylene; o-xylene; methylethylbenzene) were identified as markers of g-HBCD exposure and all increased in the liver of HBCD-exposed animals. The presence of aromatic compounds has already been highlighted in different human clinical samples during serious metabolic disorders. Styrene for example was found in human breath as a marker of lung cancer or liver cirrhosis (Hakim et al., 2012; Dadamio et al., 2012), and methylated benzenes, such as 1,3-dimethylbenzene, ethylbenzene, p-xylene and o-xylene, increased in human melanoma samples compared to controls (Abaffy et al., 2010). These compounds are functionalized usually by CYP-450 enzymes before being conjugated to a more excretable form by other enzyme systems (Hakim et al., 2012). Thus even if these aromatic compounds are considered as deriving from exogenous sources and not from the metabolome (Hakim et al., 2012), the level of these compounds could be impacted by g-HBCD exposure which could modify the activity of detoxifying enzymes such as CYP-450 enzymes. 4. Conclusion These results confirm the informative potential of the liver volatolome for exploring food contamination issues like exposure of hens to g-HBCD, and the relevance of the SPME-GC-MS approach for pointing out candidate markers in the volatolome of animal derived-food products. Recent advances in VOC-sensors (Konvalina and Haick, 2013) could pave the way to the development of routine controls based on an easily and rapidly measurement of the markers in animal tissues or food products. This proof of concept study enabled a screening of VOC markers of animal exposure to gHBCD, but confounding factors cannot be excluded and some complementary studies have to be undertaken. Efforts should turn
641
to assessing robustness and specificity of the candidate markers through the implementation of in vitro assays (Engel et al., 2015), and the analysis of on-market contaminated samples. Another priority prospect of the present study is to assess performance criteria of volatolomic analysis, especially in determining the detection limits for the candidate markers and the range of exposure level as well as time length. As a follow-up to this proof of concept study, statistical robustness will need to be improved for instance by implementing a power analysis that allows for better evaluation of inter-animal variability. Finally, together with the combination of the volatolome contained in different animal fluids, tissues or organs (Amann et al., 2014; Broza et al., 2015), the investigation of the relative balance between VOC markers (Lehallier et al., 2012) might enable to deal with the issue of diagnostic specificity. Acknowledgments The authors are grateful to the French Ministry for Food, Agriculture and Fisheries for financial support from the Compte ciale De veloppement Agricole et Rural (CASDAR d'Affectation Spe project n 7106). They thank N. Besne (PEAT, INRA Val de Loire, Nouzilly, France) for experimental feed manufacturing, P. Hart de Lorraine, meyer and C. Grandclaudon (URAFPA, INRA, Universite Nancy, France) for technical support. Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.chemosphere.2017.09.074. References Abaffy, T., Duncan, R., Riemer, D.D., Tietje, O., Elgart, G., Milikowski, C., DeFazio, R.A., 2010. Differential volatile signatures from skin, naevi and melanoma: a novel approach to detect a pathological process. PLoS One 5, e13813. Amann, A., Mochalski, P., Ruzsanyi, V., Broza, Y.Y., Haick, H., 2014. Assessment of the exhalation kinetics of volatile cancer biomarkers based on their physicochemical properties. J. Breath. Res. 8, 016003. Berge, P., Ratel, J., Fournier, A., Jondreville, C., Feidt, C., Roudaut, B., Le Bizec, B., Engel, E., 2011. Use of volatile compound metabolic signatures in poultry liver to back-race dietary exposure to rapidly metabolized xenobiotics. Environ. Sci. Tech. 45, 6584e6591. Birnbaum, L.S., Staskal, D.F., 2004. Brominated flame retardants: cause for concern? Environ. Health Persp 112, 9e17. Blake, A., 2005. Next generation of POPs: PBDEs and Lindane. IPEN. za, E., Jondreville, C., Bouhlel, J., Ratel, J., Abouelkaram, S., Mercier, F., Travel, A., Bae Dervilly-Pinel, G., Marchand, P., Le Bizec, B., Dubreil, E., Mompelat, S., Verdon, E., rin, T., Rutledge, D.N., Engel, E., 2017. Solid-phase microInthavong, C., Gue extraction set-up for the analysis of liver volatolome to detect livestock exposure to micropollutants. J. Chromatogr. A 1497, 9e18. Broza, Y.Y., Mochalski, P., Ruzsanyi, V., Amann, A., Haick, H., 2015. Hybrid volatolomics and disease detection. Angew. Chem. Int. Ed. 54, 11036e11048. Covaci, A., Gerecke, A.C., Law, R.J., Voorspoels, S., Kohler, M., Heeb, N.V., Leslie, H., Allchin, C.R., de Boer, J., 2006. Hexabromocyclododecanes (HBCDs) in the environment and humans: a review. Environ. Sci. Tech. 40, 3679e3688. Covaci, A., Roosens, L., Dirtu, A.C., Waegeneers, N., Van Overmeire, I., Neels, H., Goeyens, L., 2009. Brominated flame retardants in Belgian home-produced eggs: levels and contamination sources. Sci. Total Environ. 407, 4387e4396. Dadamio, J., Van den Velde, S., Laleman, W., Van Hee, P., Coucke, W., Nevens, F., Quirynen, M., 2012. Breath biomarkers of liver cirrhosis. J. Chromatogr. B 905, 17e22. Dominguez-Romero, E., Cariou, R., Omer, E., Marchand, P., Dervilly-Pinel, G., Le Bizec, B., Travel, A., Jondreville, C., 2016. Tissue distribution and transfer to eggs of ingested a-hexabromocyclododecane (a-HBCDD) in laying hens (Gallus domesticus). J. Agric. Food Chem. 64, 2112e2119. EFSA, 2011. Scientific opinion on hexabromocyclododecanes (HBCDDs) in food. EFSA Panel Contam. Food Chain (CONTAM) EFSA J. 2011 (9), 118. Engel, E., Ratel, J., Bouhlel, J., Planche, C., Meurillon, M., 2015. Novel approaches to improving the chemical safety of the meat chain towards toxicants. Meat Sci. 109, 75e85. European Commission 2014/118/EU, 2014. Commission Recommendation of 3 March 2014 on the Monitoring of Traces of Brominated Flame Retardants in Food (Text with EEA Relevance) (2014/118/EU). Off. J. Euro. Union. Fernie, K.J., Marteinson, S.C., Bird, D.M., Ritchie, I.J., Letcher, R.J., 2011. Reproductive
642
J. Ratel et al. / Chemosphere 189 (2017) 634e642
changes in American kestrels (Falco sparverius) in relation to exposure to technical hexabromocyclododecane flame retardant. Environ. Toxicol. Chem. 30, 2570e2575. Filipiak, W., Sponring, A., Mikoviny, T., Ager, C., Schubert, J., Miekisch, W., Amann, A., Troppmair, J., 2008. Release of volatile organic compounds (VOCs) from the lung cancer cell line CALU-1 in vitro. Cancer Cell Int. 8, 17. Focant, J.F., 2014. Keeping tabs on dioxins. Anal. Sci. 15, 34e40. nisseau, A., Le Bizec, B., Sellier, N., Engel, E., Fournier, A., Feidt, C., Marchand, P., Ve Ratel, J., Travel, A., Jondreville, C., 2012. Kinetic study of g-hexabromocyclododecane orally given to laying hens (Gallus domesticus). Environ. Sci. Pollut. Res. 19, 440e447. Goscinny, S., Vandevijvere, S., Maleki, M., Overmeire, I.V., Windal, I., Hanot, V., Blaude, M.N., Vleminckx, C., Loco, J.V., 2011. Dietary intake of hexabromocyclododecane diastereoisomers (a-, b-, and g-HBCD) in the Belgian adult population. Chemosphere 84, 279e288. Haick, H., Broza, Y.Y., Mochalski, P., Ruzsanyibc, V., Amann, A., 2014. Assessment, origin, and implementation of breath volatile cancer markers. Chem. Soc. Rev. 43, 1423e1449. Hakim, M., Broza, Y.Y., Barash, O., Peled, N., Phillips, M., Amann, A., Haick, H., 2012. Volatile organic compounds of lung cancer and possible biochemical pathways. Chem. Rev. 112, 5949e5966. Hiebl, J., Vetter, W., 2007. Detection of hexabromocyclododecane and its metabolite pentabromocyclododecene in chicken egg and fish from the official food control. J. Agric. Food Chem. 55, 3319e3324. Jelski, W., Szmitkowski, M., 2008. Alcohol dehydrogenase (ADH) and aldehyde dehydrogenase (ALDH) in the cancer diseases. Clin. Chim. Acta 395, 1e5. €tters, T., Rupp, R., Sures, B., 2015. Review of hexKoch, C., Schmidt-Ko abromocyclododecane (HBCD) with a focus on legislation and recent publications concerning toxicokinetics and -dynamics. Environ. Poll. 199, 26e34. Konvalina, G., Haick, H., 2013. Sensors for breath testing: from nanomaterials to comprehensive disease detection. Acc. Chem. Res. 7, 66e76. Lehallier, B., Ratel, J., Hanafi, M., Engel, E., 2012. Systematic ratio normalization of gas chromatography signals for biological sample discrimination and biomarker discovery. Anal. Chim. Acta 733, 16e22. Letcher, R.J., Mattioli, L.C., Marteinson, S.C., Bird, D., Ritchie, I.J., Fernie, K.J., 2015. Uptake, distribution, depletion, and in ovo transfer of isomers of hexabromocyclododecane flame retardant in diet-exposed American kestrels (Falco sparverius). Environ. Toxicol. Chem. 34, 1103e1112. Mochalski, P., Sponring, A., King, J., Unterkofler, K., Troppmair, J., Amann, A., 2013. Release and uptake of volatile organic compounds by human hepatocellular
carcinoma cells (HepG2) in vitro. Cancer Cell Int. 13, 72. Morris, S., Bersuder, P., Allchin, C.R., Zegers, B., Boon, J.P., Leonards, P.E., de Boer, J., 2006. Determination of the brominated flame retardant, hexabromocyclodocane, in sediments and biota by liquid chromatographyelectrospray ionisation mass spectrometry. Trac. Trend Anal. Chem. 25, 343e349. Ortiz de Montellano, P., Voss, J., 2005. Substrate oxidation by cytochrome P450 enzymes. Cytochrome P450 183e245. Rawn, D., Sadler, A., Quade, S., Sun, W.-F., Lau, B.-Y., Kosarac, I., Hayward, S., Ryan, J., 2011. Brominated flame retardants in Canadian chicken egg yolks. Food Addit. Contam. 28, 807e815. Sawahata, T., Neal, R.A., 1983. Biotransformation of phenol to hydroquinone and catechol by rat liver microsomes. Mol. Pharmacol. 23, 453e460. Sivadier, G., Ratel, J., Bouvier, F., Engel, E., 2008. Authentication of meat products: determination of animal feeding by parallel GC-MS analysis of three adipose tissues. J. Agric. Food Chem. 56, 9803e9812. Sponring, A., Filipiak, W., Mikoviny, T., Ager, C., Schubert, J., Miekisch, W., Amann, A., Troppmair, J., 2009. Release of volatile organic compounds from the lung cancer cell line NCI-H2087 in vitro. Anticancer Res. 29, 419e426. Sponring, A., Filipiak, W., Ager, C., Schubert, J., Miekisch, W., Amann, A., Troppmair, J., 2010. Analysis of volatile organic compounds (VOCs) in the headspace of NCI-H1666 lung cancer cells. Cancer Biomark. 7, 153e161. Szabo, D.T., Diliberto, J.J., Hakk, H., Huwe, J.K., Birnbaum, L.S., 2010. Toxicokinetics of the flame retardant hexabromocyclododecane gamma: effect of dose, timing, route, repeated exposure, and metabolism. Toxicol. Sci. 117, 282e293. Tomy, G.T., Budakowski, W., Halldorson, T., Whittle, D.M., Keir, M.J., Marvin, C., MacInnis, G., Alaee, M., 2004. Biomagnification of a-and g-hexabromocyclododecane isomers in a Lake Ontario food web. Environ. Sci. Tech. 38, 2298e2303. bault, Mahe , Grammont, Travel, F., Marchand, Venisseau, Le Bouquin, Allain, The Badreddine, Jurjanz, Feidt, Jondreville, 2012. Transfert de polluants organiques tat des lieux, modalite s et facteurs de persistants vers l’œuf de consommation : e risques. Innov. Agron. 25, 17. van der Ven, L.T., van de Kuil, T., Leonards, P.E., Slob, W., Lilienthal, H., Litens, S., n, R.F., van den Berg, M., 2009. Endocrine effects of Herlin, Håkansson, H., Canto hexabromocyclododecane (HBCD) in a one-generation reproduction study in Wistar rats. Toxicol. Lett. 185, 51e62. Vasta, V., Ratel, J., Engel, E., 2007. Mass spectrometry analysis of volatile compounds in raw meat for the authentication of the feeding background of farm animals. J. Agric. Food Chem. 55, 4630e4639.