A common surfactant used in food packaging found to be toxic for reproduction in mammals.

A common surfactant used in food packaging found to be toxic for reproduction in mammals Cristina Nerin a,∗,1 , Elena Canellas a,1 , Paula Vera a,1 , Estefanía Garcia-Calvo b,1 , José Luis Luque-Garcia b,1 , Carmen Cámara b,1 , Raquel Ausejo c,1 , Joaquín Miguel c,1 , Noelia Mendoza c,1 a I3A, University of Zaragoza, Campus Rio Ebro, Maria de Luna 3, 50018 Zaragoza, Spain b Dept. Analytical Chemistry, Complutense University of Madrid, Faculty of Chemical Sciences, 28040, Madrid, Spain c Biotechnology Research and Development Department, Magapor SL, Parque Científico Tecnológico Valdeferrín, Ejea de los Caballeros, Zaragoza, Spain ARTICLE INFO Keywords: Food packaging Reprotoxicity Surfynol Proteomic Mammals LC-MS ABSTRACT Migration from a multilayer plastic material intended for food contact showed that 2,4,7,9-tetramethyl-5-decyne-4,7-diol mixture (surfynol), used as a surfactant in the adhesive employed to build the multilayer, was transferred to water and other food simulants in contact with the plastic. When these multilayer plastics were used for containing seminal doses for artificial insemination, it was found that fertility was seriously damaged in terms of motility, acrosome integrity, mitochondrial activity and penetration capacity in the cells, thus affecting male fertility. Quantitative proteomic analysis of exposed germinal cells demonstrated the inhibition of key proteins involved in the fertilization capacity by affecting the cytoskeleton, sperm motility, the energy machinery and sperm defense mechanisms against oxidation, therefore confirming the surfactant-induced male infertility. These results open up new and interesting perspectives for the study of reprotoxicity caused by different chemicals common in our daily lives. Significance: This paper demonstrates the toxicity for reproduction of a common surfactant used in food packaging and the scientific reasons why the sperm loses reproductive capacity in presence of this chemical. So, the surfactant affects the male fertility. The surfactant is present in many adhesives used either for building multilayer materials or to glue paper and plastic in food packaging. This is the first time that reprotoxicity is demonstrated for this compound. According to the theoretical approach Threshold of Toxicological Concern (TTC) the compound is highly toxic but experimental data did not exist so far. The study described in this paper and the results obtained open a door to further research in which male infertility caused by chemicals could be demonstrated. 1. Introduction In 2010, a migration study of several food packaging materials reported a concentration of 621.0 μg/kg food of 2,4,7,9-tetramethyl-5decyne-4,7-diol (TMDD) in solid simulant Tenax (Canellas et al., 2010). No experimental data about the toxicity of TMDD were available, but the theoretical prediction according to the theoretical approach of the threshold of toxicological concern (TTC) (Barlow, 2005) classified this compound as Cramer class III, the highest toxicity level, for which the value of 90 μg/kg food should not be surpassed. According to the document written by Susan Barlow (2005), TTC is a principle that refers to the establishment of a generic human exposure threshold value for chemicals below which there would be no appreciable risk to human health. The concept proposes that such a value can be identified for many chemicals, including those of unknown toxicity when considering their chemical structures, such as the presence of aromatic rings, double or triple bonds, heterocycles and heteroatoms. The use of the TTC principle would eliminate the necessity of extensive toxicity testing and safety evaluations when human intakes of a chemical are below a certain level of concern. Obviously, those compounds taken in over the proposed limits and theoretically classified as toxic should be evaluated. Toxicity prediction for humans can be also valuable to select potential damage of chemicals in other mammals. Then, the migration of chemicals from materials used for artificial insemination should be taken with care, as all chemicals transferred from the packaging will be directly introduced into the animal. Previous studies carried out in vivo T 2.2. Samples An aqueous dispersion of adhesive containing 7% surfynol, a mixture of 2,4,7,9-tetramethyl-5-decyne-4,7-diol (TMDD) monomer and its ethoxylated polymers as a surfactant was used to glue a 35, 60 or 90 μm low-density polyethylene (LDPE) layer to a 12 μm polyethylene terephthalate (PET) layer. Surfynol was provided by Samtack (Barcelona). Different multilayers were obtained with 3 and 4 g/m 2 of grammage for adhesive in the multilayer. Ultra-high-performance liquid chromatography with quadrupole and time of flight mass spectrometry (UPLCMS-Q-TOF) was used for analysis of surfynol, the adhesive and food simulants after the exposure to plastic multilayers for 10 days. A standard of the TMDD monomer (Sigma-Aldrich (Madrid, Spain) was used as a calibrant for quantitative purposes. 2.3. Migration tests Bags of 10 cm × 5 cm were manufactured with the multilayer laminate described above and filled with 30 mL of ethanol 10% and acetic acid 3%. Three replicates of both simulants were prepared. These samples were kept in the oven at 60 °C for 10 days according to Directive 10/2011/EU. Finally, the extracts were analyzed by UPLCMS-Q-TOF using the method explained above. 2.4. Analysis by UPLC-MS-Q-TOF Chromatography separation was carried out in an AcquityTM system. The column used was an Acquity UPLC BEH C18 column of 17 μm particle size (2.1 mm × 100 mm) from Waters (Milford, MA, USA). Methanol and water were used as the mobile phase, both with 0.1% formic acid. The column flow and column temperature were 0.3 mL/min and 40 °C, respectively. The gradient was 5–95% methanol and 0.1% formic acid (0–25 min), and the volume of sample injected was 5 μL. An API source (atmospheric pressure ionization) with an electrospray interface (ESI) coupled to a Xevo G2 mass spectrometer, which consisted of a hexapole, a quadrupole, a collision cell and a time of flight analyzer (QTOF) from Waters (Milford, MA, USA), was used. Electrospray was operated in positive (ESI+) mode with two different cone voltages, 30 and 70 V, and negative (ESI-) mode to detect as many compounds as possible. The corona voltage was 2.5 kV for (ESI+) and 0.5 kV for (ESI-). Other MS parameters were as follows: the mass range was from 10 to 1000 Da, the source temperature was 150 °C, the desolvation gas temperature was 450 °C, and the desolvation gas flow was 650 Lh-1. MSE mode was selected for the acquisition. The collision ramp energy was selected from 15 to 40 V. MassLynx v.4.1 software (Waters, Milford MA, USA) was used to analyze the samples. To determine the migrants coming from the laminate in each simulant, a comparison between the chromatograms of the laminate and their respective blanks was carried out. The selected peaks coming only from the laminate were identified using Elemental Composition, Mass Fragment TM and MSE software as well as the chemical databases ChemSpider [www.chemspider.com] and Scifinder [scifinder.cas.org]. Using these tools, the potential candidates were selected. Confirmation was done with the pure standards analyzed under the same conditions. 2.5. Sperm collection Semen was manually collected by the double gloved hand technique, using a gauze filter to remove the bulbourethral gland gel secretion. All ejaculates were collected in different Spanish boar studs, diluted 1:10 in commercial boar semen extender Duragen ® with antibiotics and then immediately sent to Magapor SL quality control laboratories. Twenty ejaculates were collected from 10 different animals. The ejaculates were used individually. Samples were collected into a prewarmed insulating collection flask, within which was a 450 mL with other migrants (Nerin et al., 2014) demonstrated the importance of chemicals in contact with sperm cells. Justification of those packaging materials with TMDD or its mixture called surfynol, present in the market in 2010, was based on the small amount of adhesive used to build such materials. However, the market evolved, and packaging materials currently can contain more of this surfactant, as polyurethane adhesives are being substituted for aqueous-based in many plastic applications. Thus, the whole surface of the plastic layer will be coated with adhesive to glue the additional layers in the laminated structure, and the presence and concentration of this surfactant can be very high. Surfynol is a common surfactant used in coatings, inks and adhesives employed in many food-packaging applications. Industrial surfynol is a mixture of ethoxylated compounds, the major compounds being those with 1 and 2 ethoxy units. The surfactant is not applied on the surface in direct contact with the food but behind the plastic layer in multilayer (laminated) structures. However, migration occurs through the different layers, either paper or plastic, to both solids and liquids in contact with the packaging, as was demonstrated in several publications (Aznar et al., 2011; Canellas et al., 2015; Felix et al., 2012; Isella et al., 2013; Nerin et al., 2014; Vera et al., 2011, 2013). Many different formulas of aqueous dispersion adhesives, which are probably the new generation of environmentally friendly adhesives, use this surfactant because of its efficiency, availability and price. An in-depth study was carried out on some multilayers containing this surfactant in the adhesive used to build flexible material. It was found that seminal doses packaged in plastic bags containing this surfactant in the structure caused the inactivation and lack of fertility of spermatozoa. For this reason, migration of this surfactant to aqueous and ethanol solutions and its toxic effect on spermatozoa from mammals were investigated. The cellular model NTERA2, consisting of germinal cells of testicular embryonal carcinoma, was used for the toxicity test. In addition, to obtain a deeper insight into the biomolecular mechanisms underlying the potential toxicity of surfynol, a quantitative proteomic approach named SILAC was carried out. Stable isotopic labeling by amino acids in cell culture (SILAC) is one of the most widely used alternatives for relative protein quantitation due to its high accuracy and because it offers the possibility for the identification and quantitation of proteins within the same experiment. SILAC involves the addition of 12 Cand 13 C-labeled lysine and arginine to the growth media of separately cultured cells, giving rise to cells containing “light” or “heavy” proteins, respectively, which are further identified and quantified by mass spectrometry (Luque-Garcia et al., 2011). The reproduction capacity and integrity of spermatozoa were also evaluated by studying acrosome integrity, mitochondrial activity and penetration capacity. All experimental data confirmed the negative effect on spermatozoa of this compound. On the one hand, the present study demonstrates that direct exposure of chemicals to living cells, such as spermatozoa, is a feasible and cheap method for reprotoxicity studies concerning male fecundity. On the other hand, this research notes the risk that untested packaging materials can represent for assisted reproduction, either in animals or humans. The results obtained are shown and discussed. 2. Materials and methods 2.1. Ethical statement No human participants were involved in the research. The use of animals is limited to collecting semen from male Spanish boars and harvesting of ovaries from pigs at the official slaughterhouse of the Council. The authors confirm that all methods were carried out in accordance with relevant guidelines and regulations. The sampling procedures complied with Ethics Committee for Animal Research of the University of Zaragoza. milliliter). Samples were incubated at 37 °C in darkness for 15 min. The argon laser and filters of 525 and 675 nm were used to avoid overlap. 2.10. Mitochondrial activity A stain technique with MitoTracker (10 μM in DMSO, Invitrogen) was used to evaluate mitochondrial membrane potential (ΔΨm). Two microliters of dye was added to 300 μL(4×10 7 ) of sperm samples and incubated at 37 °C in darkness for 30 min. MitoT emissions were collected with a 675-nm filter to avoid spectral overlap. 2.11. Sperm penetration assay Sow ovaries were collected at the slaughterhouse and transported to the laboratory in sodium chloride 0.9% at RT. Oocytes were collected by the slicing and puncture techniques. Oocytes with no cumulus cells and intact zona pellucida were selected, distributed randomly, and placed in the wells of a four-well Petri dish with 400 μL of fertilization medium. These were kept at 39 °C and 5% CO2 in a humidified atmosphere until use. Spermatozoa were diluted in fertilization medium and added to the oocytes, with a final concentration of 5 × 10 5 cells per milliliter in each well. The wells were covered with mineral oil and kept in a humidified atmosphere with 5% CO 2 at 39 °C for 24 h. Special plastic bottles for seminal doses were used for incubation. After incubation, the oocytes were fixed in 1.5% glutaraldehyde for 15 min and stained with Hoechst 33342 (1 μg/mL; Sigma) for another 15 min at 37 °C. Groups of five to six oocytes were placed on a slide under a cover slide and examined with a fluorescence microscope at 400 × . The number of oocytes penetrated among total mature oocytes was counted and recorded. 2.12. Cell viability assay NTERA2 (ATCC), consisting of germinal cells of testicular embryonal carcinoma, were selected as an in vitro model since germ cells are the precursors of sperm. NTERA2 cells were seeded in 100 mm Petri dishes and cultured in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 100 units per mL of penicillin/streptomycin in 5% CO 2 at 37 °C to test cell viability. One day after seeding, cells were exposed to different concentrations of surfynol for 72 h. Surfynol concentrations ranging from 1 to 50 mg/L were selected based on preliminary cytotoxicity assays. After the exposure time, 20 μL of MTT reagent (5 mg/L) was added to each well and incubated for 5 h. Then, the MTT solution was removed, and 100 μLof dimethyl sulfoxide (DMSO) was used to redissolve the formazan crystals. The cell viability was quantified by measuring the absorbance at 595 nm in a microplate absorbance reader (Sunrise, Tecan). 2.13. SILAC experiment NTERA2 cells were maintained in DMEM medium supplemented with 10% dialyzed FBS, 100 units per mL of penicillin/streptomycin and either naturally occurring isotope (“light”) or stable isotope-labeled (“heavy”) 13 C 6 arginine and 13 C 6 lysine amino acids. After 6 cell population doublings, the full incorporation of the labeled amino acids was verified by MS analysis of a protein digest (data not shown). After differential labeling, control and cells exposed to 20 mg/L surfynol for 72 h were mixed in a 1:1 ratio. Cells were lysed with a buffer containing Tris 50 mM, NaCl 150 mM, EDTA 0.2 mM, and protease inhibitors (Roche). Protein extracts were separated by SDS-PAGE on 10% SDSpolyacrylamide gels and visualized by Coomassie blue staining, and the gel lanes were cut horizontally into 20 sections. In-gel protein digestion was carried out with 12.5 ng/μL trypsin solution in 25 mM ammonium bicarbonate and incubated overnight at 37 °C. Peptides were extracted using acetonitrile and 5% formic acid. The extracts were dried by vacuum centrifugation and reconstituted in 12 μL of 2% acetonitrile in plastic vessel with a filter and 100 mL of prewarmed (37 °C) extender to minimize thermal shock and sperm agglutination (Rodríguez-Martínez et al., 2005). Only the richest fraction was collected because it contains 80–90% of spermatozoa in the ejaculate. The extended semen was placed in plastic bags, which were sealed and transported to the laboratory at 17 ± 2 °C in a refrigerated cabinet to maintain a stable temperature and protect them from temperature drops. 2.6. Sperm sample treatments Semen volumes of 90 mL were incubated with different amounts of surfynol for 10 days at 16 °C. The semen doses were selected from the same sample. Another aliquot was maintained in the same conditions without adding surfynol as a control sample. To evaluate sperm quality, several parameters were analyzed, including viability, mitochondrial potential, acrosome integrity and early apoptosis by flow cytometry and motility by a CASA system, at 1, 4, 7 and 10 days. Finally, penetration in vitro assays were performed with the treated and control samples. The fertilization medium consisted of TCM 199 (tissue culture media 199) supplemented with sodium bicarbonate, glucose, sodium pyruvate, bovine serum albumin, penicillin and kanamycin. 2.7. Sperm motility analysis Sperm motility analysis was performed using a commercial computer-assisted sperm analysis system (ISAS 1.0.4; Proiser, Spain). This system was based on the analysis of 25 consecutive digitalized photographic images obtained from a single field at magnification ×100 in a negative phase contrast field with a heated stage (Nikon Eclipse 50i; 10 × 0.30 PLAN objective lens). All materials used were prewarmed to 36–37 °C. Aliquots of 1 mL were warmed up for 5 min at 37 °C. Samples (6 μL) were placed between prewarmed slides and cover slides and maintained at 37 °C during analysis by a hot plate. From a single field, 10 consecutive digitalized images were analyzed. The following motion characteristics were compared between control and treated groups: the percentage of total motile (TM) and progressive motile (PM) spermatozoa; curvilinear velocity (VCL), i.e., the mean path velocity of the sperm head along its actual trajectory (μm/ sec); linear velocity (VSL), i.e., the mean path velocity of the sperm head along a straight line from its first position to its last position (μm/ sec); mean velocity (VAP), i.e., the mean velocity of the sperm head along its average trajectory (μm/sec); linearity coefficient (LIN), i.e., (VSL/VCL) × 100 (%); straightness coefficient (STR), i.e., (VSL/ VAP) × 100 (%); wobble coefficient (WOB), i.e., (VAP/VCL) × 100 (%); mean amplitude of lateral head displacement (ALH), i.e., the mean value of the extreme side-to-side movement of the sperm head in each beat cycle (μm); and beat cross frequency (BCF), i.e., the frequency with which the actual sperm trajectory crosses the average path trajectory (Hz). 2.8. Flow cytometry analysis All the measurements were performed on a BD Accuri™ C6 (Becton Dickinson, Madrid, Spain) with BD software. At least 40,000 events were counted in every experiment. The sperm population was gated for further analysis on the basis of its specific forward (FS) and side scatter (SS) properties; other non-sperm events were excluded. 2.9. Viability/acrosome integrity A double stain technique with propidium iodide, a nuclear dye that penetrates inside the damaged plasmatic membrane in nonviable spermatozoa, and FITC-PNA, a staining that penetrates when spermatozoa undergo the acrosome reaction, was used. Five microliters of each stain ([FITC; 1 μg/mL; Sigma] and propidium iodide [PI; 0.75 mM; Sigma]) was added to 300 μL of sperm samples (4 × 10 7 cells per 117 0.1% formic acid. Peptide mixtures were analyzed using nanoflow LC-MS/MS. Peptides were loaded onto a 0.3 × 10 mm C18 precolumn (SGE) and separated on a reverse-phase column (75 μm × 15 cm fused silica capillary C18 HPLC PepMap column, 3 μm, 100 A, Thermo) with a linear gradient of 120 min of 5–95% acetonitrile in 0.1% aqueous solution of formic acid. The peptides were ionized with a stainless steel nanobore emitter (Thermo Scientific), scanned and fragmented with an LTQ XL linear ion trap mass spectrometer (Thermo Scientific) operated in datadependent ZoomScan and MS/MS switching mode using the three most intense precursors detected in a survey scan from 400 to 1600 u (three μscans). Normalized collision energy was set to 35%, and dynamic exclusion was applied for 3 min to avoid repetitive fragmentation ions. The generated.raw files were converted to.mgf files. A database containing the NCBI nr Human sequences with 35,586 entries (01/01/12) was searched using MASCOT Software (version 2.3 MatrixScience) for protein identification. The search parameters were as follows: oxidation of methionine and 13 C 6 -Arg and 13 C 6 -Lys as variable modifications, trypsin as the specific enzyme and one missed cleavage allowed. Minimum precursor and fragment-ion mass accuracies of 1.2 and 0.3 Da were used. A requirement of at least one bold (unique) red peptide (i.e., the highest scoring peptide matches the protein with the highest total score) was required for protein identification, and at least two bold red (unique) peptides were required for quantification. Cut-offvalues for MASCOT scores of peptides and proteins were set to 40 (P < 0.05) and 47 (P < 0.01), respectively. Relative quantitation ratios of identified proteins were calculated using QuiXoT (version 1.3.26). The Zoom Scan mass window was set to 12 Da, enabling monitoring of the entire 12 C/ 13 C isotopic envelope of most doubly and triply charged peptides. SILAC ratios were defined by the area of the heavy peptides ( 13 C) divided by the area of the light peptides ( 12 C). As observed in previous studies, a proportion of 13 C 6 -Arg was converted to 13 C 5 -Pro, leading to a reduction in the intensity of the isotope-labeled peptide peak; this was corrected for all peptides containing one or more proline residues. Molecular and cellular functions of the proteins found to be deregulated by SILAC were assigned based on the biological knowledge available in Gene Ontology (GO) annotations. 2.14. Statistical analyses The results are shown as the mean ± SEM (standard error of the mean) of the number of samples indicated in each case. All binomial and nominal variables, expressed as percentage, were analyzed using ANOVA, with DMS as a post hoc test after a Kolmogorov-Smirnov normality test (Stat View 5.0 Software, SAS Institute INC). 3. Results and discussion 3.1. Identification of migrants from packaging Multilayer materials are quite common in food packaging, as barrier properties cannot be achieved with only one plastic layer, and a combination of different materials is currently used. Most of the multilayers are built by lamination, that is, by sticking layers of different materials together with adhesive, covering the whole surface of the materials. It has been demonstrated that all chemicals with an atomic mass unit (amu) lower than 1000 can migrate through plastic layers (Aznar et al., 2011; Canellas et al., 2010, 2015; Felix et al., 2012; Isella et al., 2013; Nerin et al., 2014; Vera et al., 2011, 2013) and reach the packaged product. Thus, to ensure safety in the use of these materials, migration tests were carried out using ethanol 10% and acetic acid 3% as simulants. These food simulants have been established to mimic the behavior of acidic or hydrophilic foodstuffs and were selected in this study as the worst-case scenario, as surfynol is a polar surfactant. The analysis of simulants by UPLC-MS-Q-TOF showed evidence of migration of TMDD monomers and their ethoxylated compounds (Fig. 1). Thus, there was no doubt that the main components of surfynol were present in the food simulants after the exposure. This surfactant is not a pure substance but a mixture of several ethoxylated compounds. From 8 to 8.5 min, the following accurate masses were detected: 249.1839 Da, 293.2098 Da, 337.2354 Da and 381.2612 Da. The elemental Fig. 1. UPLC-MS-Q-TOF of Surfynol (A), blank (B) and migration to acetic acid 3% (C) after exposure to plastic multilayer containing Surfynol. 118 composition was elucidated from these accurate masses and the isotopic fit. It was concluded that they corresponded to the sodium adducts of the monomer 2,4,7,9-tetramethyl-5-decyne-4,7-diol and the polymers 2,4,7,9-tetramethyl-5-decyne-4,7-diol (TMDD) ethoxylate n = 1 m = 0, TMDD ethoxylate n = 1 m = 1 and TMDD ethoxylate n = 2 m = 1. Two isomers of each polymer were detected. These compounds diffused through the LDPE layer and were dissolved in the simulant, resulting in the migration phenomena. Different thicknesses of LDPE were tested in order to check if a higher thickness would delay the migration. Table 1 shows the quantitative results obtained. Only 4 compounds were found in migration solutions, with three of them coming from surfynol. The TMDD monomer and its ethoxylated compounds are not degraded in acidic medium and remain stable for at least 15 days. As seen, the migration values depended on both the compound and the LDPE thickness. As expected, higher migration values were obtained for the thinner PE (35 μm), as surfynol is behind the LDPE layer and has to diffuse through a shorter distance. However, migration values were also found even using 90 μm of LDPE, which demonstrates that LDPE cannot be used as a barrier. Higher migration values were obtained for 10% ethanol than for 3% acetic acid in water, as TMDD and its ethoxylated compounds are organic compounds and more soluble in ethanol. Higher migration was obtained for ethoxylated compounds than for the original monomer, with 17.07 mg/ kg for TMDD ethoxylated n = 1 m = 1 ∗ (sum of isomers) being the highest migration value in the food simulant. 1-Hexanol-2-ethyl is included on the positive list of European Regulation 10/2011/EU, with a specific migration limit of 30 mg/kg. The migration values found in this study are well below this limit. However, none of the ethoxylated compounds are on the positive list of EU Regulation of food contact materials. Thus, the migration value should be lower than 10 μg/kg, after checking that the compound is not carcinogenic, mutagenic or reprotoxic (CMR). Toxicity data are required for the risk assessment of the material. Toxicological values were searched for all migrants. The TMDD monomer was tested on aquatic organisms and animals. Using data from a 28-day dietary study and applying a very conservative safety factor of 1,000, an acceptable daily intake (ADI) of 15 mg per day can be estimated. The TMDD monomer was reviewed by the U.S. Environmental Protection Agency (USEPA), which determined in 2008 that the TMDD monomer was a low priority for further review in the chemical assessment and management program (Labunska et al., 2012; US-EPA, 2008). The USEPA recognized the potential for the general population to be exposed to environmental releases due to the commercial use of the TMDD monomer and its low biodegradation rate, as it is very persistent in water. Based on the low human health hazard, there was a low concern for potential risks to the general population from environmental releases, which means a lowest observed adverse effect level (LOAEL) of 200 mg/kg body weight/day. The NOAEL can be calculated as 3 times lower than the LOAEL. The self-derived specific migration limit (SML) was then calculated by applying an assessment factor of 100 to the NOAEL value and taking into account a body weight of 60 kg (PlasticsEurope, 2014). Therefore, the self-derived SML would be 40 mg/kg food for adults and 0.6 for toddlers of 10 kg body weight. Consequently, the migration values of the TMDD monomer were below the limit for adults but above this selfderived SML for toddlers. Concerning the ethoxylated TMDD migrants, neither SML nor NOAEL data were found. TTC can then be applied for risk assessment. Toxtree (Ideaconsult, 2011) software was used for Cramer class assignation. These compounds were classified as Cramer class III with an estimated dairy intake (EDI) of 0.09 mg/person/day. Considering a dairy intake of 1 kg of food, the self-derived SML of 0.09 mg/kg food is obtained for all compounds of Cramer class III, if CMR substances are discarded. As seen in Table 1, the migration values obtained for the surfactant ethoxylated polymers were much higher than this value. Thus, the migration could represent a risk for human health, and additional tests are required. None of these compounds were degraded in acidic medium. As these surfactants are very common in food packaging materials, an in-depth study was carried out on their reprotoxicity. The results obtained are described in the following paragraphs. 3.2. Quantitative proteomics suggests potential reproductive toxicity induced by TMDD in a cellular model An MTT assay was carried out to test the cytotoxicity of surfynol in the NTERA2 model. A decrease in cell viability from 85% to 5% was observed with increasing concentrations of surfynol from 1 mg/L to 50 mg/L respectively. The highest concentration tested (50 mg/L) caused the cellular death of almost the entire cell culture. To evaluate the effect on the cells of this TMDD mixture, but without drastically compromising the cell viability, a concentration of 20 mg/L was used to carry out the SILAC experiment. To identify the proteins altered after TMDD mixture exposure, two large-scale SILAC experiments were performed (Fig. 2). Mass spectrometry analysis identified a total of 2207 proteins with at least two peptides, of which 1147 passed the criteria established for quantitation. The mean relative standard deviation (RSD) of the ratios obtained from replicates was lower than 20%, indicating good agreement between experiments. Most of the quantified proteins presented a SILAC ratio close to 1, as expected for a 1:1 mixture. Using 1.5 as the threshold ratio, 108 proteins were found to be altered upon surfynol exposure, 34 of which were found to be upregulated and 74 downregulated (Table S1). Several structural proteins were found to be downregulated. These structural proteins are major components of the cytoskeleton that provide cellular structural support and cell motility (Cheng, 2008). Among these proteins, a significant inhibition of tubulin (R SILAC =−1.54) was found. Since this protein constitutes 70% of the axoneme, which is the active organelle of the sperm flagellum, this inhibition might negatively affect spermatozoa development (Cosson, Compound/migration mg/Kg Mass Q-TOF LOD Ethanol 10% PE(35 μm)-adh (4 g/m 2 )- PET(12 μm) Ethanol 10% PE(60 μm)-adh (4 g/m 2 )- PET(12 μm) Ethanol 10% PE(90 μm)-adh (3 g/m 2 )- PET(12 μm) AC 3% PE(35 μm)-adh (4 g/m 2 )- PET(12 μm) AC 3% PE(60 μm)-adh (4 g/m 2 )- PET(12 μm) AC 3% PE(90 μm)-adh (3 g/m 2 )- PET(12 μm) 1-hexanol-2-ethyl 0.001 0.005 0.003 < LOD 0.011 0.007 0.003 2,4,7,9-Tetramethyl-5-decyne-4,7diol (sum of isomers) 249.1827 0.012 1.38 0.58 < LOD 0.72 0.33 0.14 2,4,7,9-Tetramethyl-5-decyne-4,7diol ethoxylate n=1m=1 ∗ (sum of isomers) 337.2353 0.012 17.07 11.08 2.90 13.40 6.42 5.67 2,4,7,9-Tetramethyl-5-decyne-4,7diol ethoxylate n = 2 m = 1*(sum of isomers) 381.2614 0.012 15.89 8.81 1.68 12.70 4.77 4.40 119 2012). Other structural proteins found to be downregulated were actin (R SILAC =−1.95) and several actin-related proteins, including actin crosslinkers such as filamin C (R SILAC =−2.37), transgelin (R SILAC =−2.60) and transgelin 2 (R SILAC =−1.88); an actin/myosin binding protein named caldesmon (R SILAC =−2.31); zyxin (R SILAC =−2.13), which mediates actin adhesion; and fascin (R SILAC =−1.62), which is involved in the organization of actin filaments. Actin is a key protein present in the head, neck and tail of mammalian spermatozoa. Thus, given the crucial role of the actin cytoskeleton during spermatogenesis and, therefore, in the gradual differentiation process and function of sperm for fertilization (Virtanen et al., 1984; Xiao and Yang, 2007), inhibition of all these proteins induced by surfynol can be related to an actin cytoskeletal dysfunction, affecting the fertilization capacity of exposed cells. In addition, upregulation of DAZ-associated protein 1 (R SILAC = 2.47) and hepatomaderived growth factor (RSILAC = 1.54) have been previously linked to abnormal development of germ cells, which also correlates well with the results found in the experiment (Fu et al., 2015; Kuroda et al., 1999). The TMDD mixture induced downregulation of additional proteins related to the sperm cytoskeleton such as tropomyosin (R SILAC =−2.28), which is an actin-dependent motor protein associated with cargo movement; catenin (R SILAC =−1.73), which associates with cadherin and produces a complex linked to actin filaments; talin (R SILAC =−1.53), which is a protein involved in cytoskeletal organization; and vimentin (R SILAC =−1.79), located in the head of the spermatozoa and with a crucial role in the fusion of the sperm with the surface membrane of the ovum during fertilization. This result also supports the potential lack of fecundity associated with exposure to the TMDD mixture. Another set of proteins inhibited upon exposure to the TMDD mixture that might affect the fertilization capacity of spermatozoa includes annexin A1 (R SILAC =−4.61), which forms heterodimers with S100A11 (R SILAC =−2.74), annexin A2 (R SILAC =−2.43) which forms heterodimers with S100-A10 (R SILAC =−4.03), and annexin A6 (R SILAC =−1.52) (De Jonge and Barratt, 2006; Gibbons et al., 1978; Tourmente and Roldan, 2015). These proteins are related to the binding and release of Ca 2+ ions, which regulate mammalian fertilization, as sperm requires extracellular Ca 2+ for two processes that precede fertilization: motility hyperactivation and the acrosome reaction. In mammals, sperm motility depends on the propulsive force generated by moving the flagellum, which is reached through ATP hydrolysis (Tourmente and Roldan, 2015). Between 10 and 15% of the global protein mass of axonemes are dynein-ATPases. Dynein 1 (R SILAC =−1.82), which was found to be inhibited after surfynol exposure, acts as a micromotor in charge of transforming the chemical energy (ATP) into mechanical work. More than 70% of the total ATPase activity is directly coupled to cell motion (Cosson, 2012). Previous studies have demonstrated that inhibitors of dynein 1, such as vanadate, induce inhibition of cilia and sperm flagella motility (Gibbons et al., 1978). Therefore, a similar negative effect on sperm motility Fig. 2. Experimental set-up of the performed SILAC experiment. 120 regulate defense mechanisms under different kinds of stress (Kregel, 2002), or the apoptotic chromatin condensation inducer in the nucleus isoform 1 (2.10), which induces damaged cells to undergo apoptotic cell death (Sahara et al., 1999). Overall, the results obtained from the SILAC experiment have shown a significant downregulation of several key proteins involved in the fertilization capacity by affecting the cytoskeleton, the sperm motility, the energy machinery and the sperm defense mechanisms against oxidation. These results demonstrate the potential toxicity of surfynol to reproduction, as it affects male fertility. 3.3. Reprotoxicity study Artificial insemination uses semen doses stored in multilayer plastic bags, where migration of toxic compounds from the packaging to the semen occurs (Nerin et al., 2014). Thus, to demonstrate the influence of packaging on sperm and perform reprotoxicity studies, one clever and easily available approach is the direct exposure of the compounds under study to living cells of spermatozoa. The quality and main characteristics of sperm can be measured, and then, conclusions about the toxic effects that the chemicals cause on spermatozoa can be determined. To better represent the effects on mammals, semen from boars was used for the tests. Semen was spiked with different concentrations of surfynol, covering the range found in previous migration studies, and motility, acrosome integrity, mitochondrial activity and penetration tests were performed. 3.4. In vitro seminal analysis To investigate the mechanisms of surfynol, several parameters were analyzed (total and progressive motility, viability, acrosome status and mitochondrial activity). The samples were prepared with Duragen ® (Magapor SL), a specific extender of extra-long duration that conserves the samples in optimal conditions of viability. Spermatozoa in direct contact with the potential toxic compound at different concentrations for 48 h, 3 days, 8 days and 11 days were tested. Sperm quality parameters were assessed and are shown in Fig. 3 A) Total Motility (n = 10), Fig. 3 B) Progressive Motility (n = 10), Fig. 3 C) Viability (FITC-PNA-/IP-; n = 10), Fig. 3 D) Acrosome Reaction (FITC-PNA+; n = 10), and Fig. 3 E) Mitochondrial Activity (n = 10). Significant differences in the results were observed in the groups relative to the control, *P < 0.05, **P < 0.01 and ***P < 0.001, and are stated below. The addition of surfynol caused a significant decrease (P < 0.001) of both total and progressive motility in all treated samples (Fig. 3 Ay B) compared to the control. The highest effect was found at 48 h (P < 0.001) in progressive and total motility, which is interesting from the sperm point of view due to the significant effect on motility (P < 0.001) in a short period of time. This effect increases with the incubation time until reaching the maximum inhibition after 8 days. The inhibition is as interesting as the speed with which it occurs because seminal doses are used at most after 72 h after extraction. Surfynol decreases viability and produces a significant (P < 0.001) increase in the percentage of spermatozoa with reacted acrosomes in relation to the control after 11 days of exposure to surfynol. The negative effect of surfynol on the viability is not significant. Mitochondrial activity was also determined, given the significance of mitochondria in sperm motility. The results obtained showed that the proportion of viable spermatozoa with high mitochondrial activity (MitoT+) in the samples (90% ± 4%) decreased significantly (P < 0.001) after incubation with surfynol at 8 days (control samples, 58% ± 4%). Our findings showed that the values of all analyzed characteristics (except viability) were significantly lower in the samples exposed to surfynol compared to control samples. This might be due to the toxic effect of surfynol on spermatozoa. caused by TMDD mixture-induced inhibition of dynein 1 can be expected. Other proteins involved in sperm motility, such as Ras (−1.69), the dual specificity mitogen-activated protein kinase 1 (−1.57) and a small GTP-binding protein named Rho (−1.76), whose inactivation blocks sperm motility (Cheng, 2008; De Jonge and Barratt, 2006), were also found inhibited in NTERA2 cells upon surfynol exposure. Supporting these results that indicate a toxic effect of the TMDD mixture on the cellular energy machinery and sperm motility, considering that glycolysis is the principal metabolic pathway for ATP production in mature sperm (Cosson, 2012; De Jonge and Barratt, 2006), several proteins involved in the glycolysis pathway were also found to be downregulated upon surfynol exposure: hexokinase (R SILAC = −1.60), pyruvate kinase isozymes M1 (R SILAC = −1.64) and M2 (R SILAC =−1.75) and lactate dehydrogenase A chains 1 (R SILAC = −1.94) and 3 (R SILAC =−1.94), which transform pyruvate into lactate. Inhibition of these proteins might affect ATP production and induce lactate deficiency. Pyruvate and lactate have been shown to maintain a high rate of protein synthesis in spermatozoa and are involved in the effects of follicle-stimulating hormone (FSH) during spermatogenesis (Jutte et al., 1983). Furthermore, germ cells use lactate as a substrate for ATP production in mitochondrial oxidative phosphorylation; thus, the lack of lactate affects male germ cell viability (Erkkilä et al., 2002). Additional proteins participating in the respiratory chain were also found to be downregulated: NADH dehydrogenase (R SILAC =−1.66), cytochrome c (R SILAC =−2.04) and cytochrome c oxidase subunits 5A (R SILAC = −1.57), II (R SILAC = −1.82) and 5B (R SILAC =−1.93). Maintenance of ATP levels during motility can also be provided by creatine kinase (CK), which is a phosphotransferase protein that catalyzes the transference of a phosphate group to generate ATP. This protein acts as a temporal and spatial ATP buffer by mitigating mismatches of ATP supply and demand(Cosson, 2012). CK was found to be upregulated (1.54), which might indicate an impairment in ATP production during glycolysis and mitochondrial oxidative phosphorylation. Antioxidant proteins are necessary for germ cell survival, playing a key role in protecting the cells from oxidative damage and, in the case of sperm, from loss of motility. Many constitutive proteins of this defense mechanism were also found to be inhibited after exposure to the TMDD mixture: glutathione peroxidases 7 (R SILAC =−1.51) and 1 (R SILAC = −1.90), which are needed for sperm motility, and other wellknown antioxidant proteins, glutathione S-transferase 3 (R SILAC = −1.64) and omega-1 (R SILAC =−1.79). In this context, ascorbic acid, a powerful lipophilic antioxidant, also plays an absolutely vital role for the maintenance of mammalian spermatogenesis. Low levels of ascorbic acid have been related to infertile seminal plasma and low sperm motility (Fanaei et al., 2014; Yazama et al., 2006). There are two molecular pathways to obtain ascorbic acid in the organism, both beginning with the transformation of glucose in glucose 6-phosphate by the enzyme hexokinase. This enzyme was found to be downregulated in our experiment, which might be directly related to low levels of ascorbic acid in the cell and, therefore, low sperm motility induced by the exposure to the TMDD mixture. This fact is supported by two additional enzymes also found to be inhibited: phosphomannomutase (R SILAC = −2.72), which transforms mannose-6-P into mannose-1-P, and UDP-glucose 6-dehydrogenase (R SILAC = −1.62), which converts UDP-D-glucose to UDP-D-glucuronate, with both steps being necessary in the ascorbic acid biosynthesis pathways. Immunosuppressors, such proteasomes, polyamines (spermine and spermidine) and transforming growth factors are also involved in spermatozoa protection from a loss of motility induced by reactive oxygen species (Pegg, 2014; Yazama et al., 2006). Spermine synthetase, a protein that regulates the formation of spermine, was also found to be inhibited after TMDD mixture exposure (R SILAC = −1.88). Additionally, other proteins found to be upregulated indicated a general cellular response against TMDD mixture toxicity. This was the case with heat shock 70 kDa (1.78) and 75 kDa (1.53) proteins, which 121 It is worth noting that the addition of surfynol to sperm samples decreases the quality values, which suggests that these spermatozoa may have worse survival rates. Our findings are in accordance with the protein studies described above and with the theoretical classification as class III (high toxicity) from Cramer. It has already been hypothesized that this harmful effect is related to the apoptotic role of surfynol, as deduced from the significantly high effect on the penetration rate (number of oocytes penetrated over the total mature oocyte number) in just 48 h. It has been proved that the harmful effect of surfynol is related to mitochondrial status and motility, as we found that the proportion of motile sperm decreased in the presence of surfynol. This is an interesting result because mitochondrial activity is directly related to motility, which affects fertilization capacity, as we have seen. Previous studies carried out on the exposure of spermatozoa to other migrants coming from plastic bags caused strong reprotoxicity in sows. However, the spermatozoa were not apparently affected, and the toxic effect was demonstrated in the in vivo analysis. In fact, the chemicals tested in the previous study caused reproductive failures, but the spermatozoa were as fertile as the controls in the in vitro tests. In this case, the surfynol surfactant directly affects the fecundity of spermatozoa, and this effect is concentration dependent. 3.5. In vitro penetration assay The aim of this test was to highlight the toxicity of surfynol based on an oocyte penetration assay. For this reason, the tests were carried out in different groups of sperm: sperm stored in bags spiked with different concentrations of toxic compounds and sperm stored in control bags. The use of homospecific oocytes and spermatozoa in an artificial system allows the analysis of all the phases of the fertilization process in great detail. It has been suggested that this approach would be a more adequate method for predicting the fertilization capacity of spermatozoa. Howard et al. (1991) and Mattioli et al. (1990) demonstrated for the first time that immature oocytes can be penetrated even though the spermatozoid that has penetrated is unable to decondense its nuclear material as a result of the immature cytoplasm of the oocyte. This possibility allows the introduction of the immature oocyte penetration test as a way to test the fecundity of porcine spermatozoa (Gadea et al., 1998; Martinez et al., 1993; Matás et al., 1996). The in vitro penetration assay in combination with the analysis of other important aspects of sperm function, such as movement and mitochondrial activity, is determinant for fecundity potential. The results showed significant differences (P < 0.001) in terms of penetration rate (90% ± 3.2 vs. 40% ± 4.2) and in penetrated sperm per oocyte between the control and surfynol groups (Fig. 4). In total, 50 oocytes per sample were scored. Although at 48 h, the effect on the viability and acrosome reaction was not significant, we could find a very significant decrease (P < 0.001) in the sperm penetration rate, as other authors have described (Mohamed et al., 2010; Rahman et al., 2015) for other compounds. The results obtained showed that all surfynol concentrations produced a great effect and significant damage in all parameters studied (except viability) not being concentration dependent in the concentration range studied, as the differences appear between the samples with surfynol (at any concentration studied) and the control. The harmful effect of surfynol targets the action of the plasma membrane, the acrosome, and the mitochondria and therefore affects the motility of the spermatozoa. Surfynol begins to act at the biochemical level in the spermatozoa at 24 h, but its maximum effect is seen after 48 h. However, the effect on fertilization ability (assessed by the penetration test) is observed at 24 h. This effect, added to the decrease in motility due to low mitochondrial activity, produces a decrease in the penetration rate. Linear regression analysis demonstrated a positive correlation between the percentage of spermatozoa with mitochondrial activity and total motility. These results support the hypothesis of the relationship between mitochondria and motility. With this correlation Fig. 3. Effect of different amounts of TMDD on functionality markers of boar sperm at 48 h, 3 days, 8 days and 11 days. A) Total Motility (n = 10). B) Progressive Motility (n = 10). C) Viability (FITC-PNA-/IP-; n = 10). D) Acrosome reacted (FITC-PNA+; n = 10). E) Mitochondrial Activity (n = 10). Significant difference related to control: *P < 0.05, **P < 0.01, ***P < 0.001. 122 we see (Fig. 5) that the surfynol action that affects the production of mitochondrial energy can lead to the alteration of motility. A relationship between both parameters supports the hypothesis of the intracellular mitochondrial action of surfynol. Therefore, the mechanism of action of surfynol must be intracellular, as it does not significantly affect sperm viability (evaluated as stability of the plasma membrane). Surfynol probably exerts its effect on a molecular target belonging to a mitochondrial metabolic pathway or directly affecting mitochondria, which would explain the significant decrease in mitochondrial activity and the consequent effect on sperm motility. These results are in agreement with the proteomic data obtained in the cellular model, as several of the proteins that are downregulated upon surfynol exposure are involved in glycolysis and the mitochondrial respiratory chain. Mitochondria, in addition to playing a key role in sperm motility, are the fundamental organelles in the intrinsic apoptotic pathway. Thus, the toxic effect of surfynol on the mitochondria could trigger the activation of this apoptotic pathway, which would explain the significant harmful effect on the sperm penetration rate. The fecundating capacity of the spermatozoa is altered by the surfynol effect on the mitochondria. The decrease in mitochondrial activity produces a chain reaction of several molecules involved in the apoptotic pathway that would eventually cause damage to DNA. This would explain why the damage in the acrosomes is not significant, as other authors have observed (Mohamed et al., 2011), up to 11 days after treatment. Thus, they are not the cause of the decline in fertilization capacity, but the cause is probably damage in the DNA derived from the mitochondrial alteration produced by the harmful effect of surfynol. The present research did not involve DNA studies and so this theoretical approach is not confirmed. To determine the correct mechanism of action of surfynol on spermatozoa, it would be necessary to perform more studies on the mitochondrial effect and the actuation of the intrinsic apoptotic pathway. To provide new insights into the mechanisms through which surfynol is able to produce a toxic effect on spermatozoa, the relationship between different biochemical parameters of the spermatozoa has been examined in this study. The results obtained could help in understanding the mechanisms of action of toxic compounds seen by other authors (Park et al., 2011; Rahman et al., 2016, 2017). The results of this research open up new and interesting perspectives for the study of reprotoxicity of different chemicals, which are common in our daily lives and probably cause many of the miscarriages and lack of male fertility occurring currently in humans in developed countries. Author contributions C. Nerín, P. Vera and E. Canellas performed the migration study, identification and quantitation of migrants from multilayers and LCMS-QTOF analysis; E. Garcia-Calvo, J.L. Luque-Garcia, C. Cámara did the proteomic analysis and R. Ausejo, J. Miguel, N. Mendoza did the reproduction study. Competing financial interests The author(s) declare no competing financial interests. Acknowledgments Authors wish to thank the financial support from the Ministry of Economy and Competitiveness (Projects IPT-2012-0261-420000 and CTQ2014-55711-R), Gobierno de Aragón and European Social Funds given to GUIA group T-10 (University of Zaragoza). Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx. doi.org/10.1016/j.fct.2018.01.044. Transparency document Transparency document related to this article can be found online at http://dx.doi.org/10.1016/j.fct.2018.01.044. References Aznar, M., Vera, P., Canellas, E., Nerin, C., Mercea, P., Stormer, A., 2011. Composition of the adhesives used in food packaging multilayer materials and migration studies from packaging to food. J. Mater. Chem. 21, 4358–4370. Barlow, S., 2005. Threshold of Toxicological Concern: a Tool for Assessing Substances of Unknown Toxicity Present at Low Levels in the Diet. ILSI Press, Washington DC and Brussels. Canellas, E., Aznar, M., Nerin, C., Mercea, P., 2010. Partition and diffusion of volatile compounds from acrylic adhesives used for food packaging multilayers manufacturing. J. Mater. Chem. 20, 5100–5109. Canellas, E., Vera, P., Nerín, C., 2015. Risk assessment derived from migrants identified in several adhesives commonly used in food contact materials. Food Chem. Toxicol. 75, 79–87. Cosson, J., 2012. ATP, the Sperm Movement Energizer. Nova Publisher Inc, USA. Cheng, C.Y., 2008. Molecular Mechanisms in Spermatogenesis. Springer Science +Business Media. Fig. 4. Sperm Penetration Rate at 48 h (n = 5). Significant difference related to control: *P < 0.05, **P < 0.01, ***P < 0.001. Fig. 5. Linear regression analysis. Positive correlations between the percentage of spermatozoa with total motility and mitochondrial activity in all treated samples. 123