Safety evaluation of food contact paper and board using chemical tests and in vitro bioassays: role of known and unknown substances.

In vitro toxicological tests have been proposed as an approach to complement the chemical safety assessment of food contact materials, particularly those with a complex or unknown chemical composition such as paper and board. Among the concerns raised regarding the applicability of in vitro tests are the effects of interference of the extractables on the outcome of the cytotoxicity and genotoxicity tests applied and the role of known compounds present in chemically complex materials, such as paper and board, either as constituents or contaminants. To answer these questions, a series of experiments were performed to assess the role of natural substances (wood extracts, resin acids), some additives (diisopropylnaphthalene, phthalates, acrylamide, fluorescent whitening agents) and contaminants (2,4-diaminotoluene, benzo[a]pyrene) in the toxicological profile of paper and board. These substances were individually tested or used to spike actual paper and board extracts. The toxic concentrations of diisopropylnaphthalenes and phthalates were compared with those actually detected in paper and board extracts showing conspicuous toxicity. According to the results of the spiking experiments, the extracts did not affect the toxicity of tested chemicals nor was there any significant metabolic interference in the cases where two compounds were used in tests involving xenobiotic metabolism by the target cells. While the identified substances apparently have a role in the cytotoxicity of some of the project samples, their presence does not explain the total toxicological profile of the extracts. In conclusion, in vitro toxicological testing can have a role in the safety assessment of chemically complex materials in detecting potentially harmful activities not predictable by chemical analysis alone.

Test procedures for obtaining representative extracts suitable for reliable in vitro toxicity assessment of paper and board intended for food contact.

This paper describes the use of a suite of extraction procedures applicable to the assessment of the in vitro toxicity of paper/board samples intended for food-contact applications. The sample is extracted with ethanol, water, or exposed to modified polyphenylene oxide (Tenax) for fatty, non-fatty and dry food applications, respectively. The water extracts are directly suitable for safety assessment using in vitro bioassays. The ethanol extracts of the paper/board and of the exposed Tenax require pre-concentration to give acceptable sensitivity. This is because the in vitro bioassays can tolerate only a small percentage of added organic solvent before the solvent itself inhibits. The extraction procedures have been selected such that they mimic the foreseeable conditions of use with foods and that they are also fully compatible with the battery of in vitro biological assays for the safety assessment of the total migrate. The application of the extraction protocols is illustrated by the results for one of the many paper/board samples provided by the BIOSAFEPAPER project industrial platform members. The assessment indicated that this sample should not be considered as suitable for use with fatty foodstuffs but was suitable for dry and non-fatty foods. Information subsequently received from the manufacturer revealed that this was a non-food-grade product included in the project to test the capabilities of the bioassay procedures. The selection criteria for the test conditions and the suite of methods developed have been prepared in Comité Européen de Normalisation (CEN) format and is currently being progressed by CEN/TC172 as a European Standard.

The BIOSAFEPAPER project for in vitro toxicity assessments: preparation, detailed chemical characterisation and testing of extracts from paper and board samples.

Nineteen food contact papers and boards and one non-food contact board were extracted following test protocols developed within European Union funded project BIOSAFEPAPER. The extraction media were either hot or cold water, 95% ethanol or Tenax, according to the end use of the sample. The extractable dry matter content of the samples varied from 1200 to 11,800 mg/kg (0.8-35.5 mg/dm2). According to GC-MS the main substances extracted into water were pulp-derived natural products such as fatty acids, resin acids, natural wood sterols and alkanols. Substances extracted into ethanol particularly, were diisopropylnaphthalenes, alkanes and phthalic acid esters. The non-food contact board showed the greatest number and highest concentrations of GC-MS detectable compounds. The extracts were subjected to a battery of in vitro toxicity tests measuring both acute and sublethal cytotoxicity and genotoxic effects. None of the water or Tenax extracts was positive in cytotoxicity or genotoxicity assays. The ethanol extract of the non-food contact board gave a positive response in the genotoxicity assays, and all four ethanol extracts gave positive response(s) in the cytotoxicity assays to some extent. These responses could not be pinpointed to any specific compound, although there appeared a correlation between the total amount of extractables and toxicity.

Aflatoxin M1 absorption and cytotoxicity on human intestinal in vitro model.

Aflatoxin M1 (AFM1) is the principal hydroxylated Aflatoxin B1 (AFB1) metabolite and is detected in milk of mammals, after consumption of feed contaminated with AFB1. As it is classified as probable human carcinogen (group 2B of the IARC), most countries have regulated its maximum allowed levels in milk in order to reduce AFM1 risk (50 ng/kg the EU and 500 ng/kg in the USA). It was demonstrated that if AFB1 must be converted into its reactive epoxide to exert its effects, and the protein binding may play an important role in its cytotoxicity. Conversely, the AFM1 epoxidation in human liver microsomes is very limited and studies with human cell line (MCL5), expressing or not expressing cytochrome P450 enzymes, demonstrated a direct toxic potential of AFM1 in absence of metabolic activation. For this reason, while AFM1 is generally considered a detoxification product of AFB1 relatively to carcinogenicity and mutagenicity property, this is not always true for cytotoxicity activity. Aim of this work is to evaluate the intestinal absorption of AFM1 using a human in vitro model, the Caco-2 cell line. Either the parental Caco-2 cell line or its derived clone TC7, with higher metabolic competence, have been used. They were treated with different concentrations of AFM1, that mirror the milk contamination level (0.3-32 nM corresponding to 10-10,000 ng/kg), either in undifferentiated or in differentiated phase of growth. After 48 h of treatment in serum free medium, a dose dependent absorption of AFM1 has been detected in both cell lines, especially in differentiated cells, while, no appreciable effects on cell viability were observed, except for a general cellular suffering, revealed by LDH release, particularly evident in the undifferentiated cells. As well, no metabolites or AFM1 conjugates have been detected. The present results may be crucial for the evaluation of human risk to AFM1 exposure, in particular for children's population, due to their large use of milk and derivatives.

Safety assessment of food-contact paper and board using a battery of short-term toxicity tests: European union BIOSAFEPAPER project.

An European Union (EU)-funded project QLK1-CT-2001-00930 (BIOSAFEPAPER) involves the development, validation and intercalibration of a short-term battery of toxicological tests for the safety assessment of food-contact paper and board. Dissemination of the results to industry, legislators (e.g. DG Consumer Protection, DG Enterprises, DG Research), standardization bodies such as CEN, and consumers will create an agreed risk evaluation procedure. The project involves pre-normative research in order to establish a set of in-vitro cytotoxicity and genotoxicity tests that will be easily adaptable to food-contact fibre-based materials and have endpoints relevant to consumer safety, including sub-lethal cellular events. These tests will be performed on samples representing actual migration conditions from food-contact paper and board with respect to different foodstuffs, and should form an experimental basis for scientifically sound recommendations for a harmonized system of risk evaluation and product testing.

The Caco-2 cell line as a model of the intestinal barrier: influence of cell and culture-related factors on Caco-2 cell functional characteristics.

The human intestinal Caco-2 cell line has been extensively used over the last twenty years as a model of the intestinal barrier. The parental cell line, originally obtained from a human colon adenocarcinoma, undergoes in culture a process of spontaneous differentiation that leads to the formation of a monolayer of cells, expressing several morphological and functional characteristics of the mature enterocyte. Culture-related conditions were shown to influence the expression of these characteristics, in part due to the intrinsic heterogeneity of the parental cell line, leading to selection of sub-populations of cells becoming prominent in the culture. In addition, several clonal cell lines have been isolated from the parental line, exhibiting in general a more homogeneous expression of differentiation traits, while not always expressing all characteristics of the parental line. Culture-related conditions, as well as the different Caco-2 cell lines utilized in different laboratories, often make it extremely difficult to compare results in the literature. This review is aimed at summarizing recent, or previously unreviewed, data from the literature on the effects of culture-related factors and the influence of line sub-types (parental vs. different clonal lines) on the expression of differentiation traits important for the use of Caco-2 cells as a model of the absorptive and defensive properties of the intestinal mucosa. Since the use of Caco-2 cells has grown exponentially in recent years, it is particularly important to highlight these methodological aspects in order to promote the standardization and optimisation of this intestinal model.

Absorption of fumonisin B1 and aminopentol on an in vitro model of intestinal epithelium; the role of P-glycoprotein.

The aim of the present paper is to evaluate the absorption of fumonisin B1 and its principal metabolite, aminopentol on a human intestinal model, Caco-2 cells, cultured on semi-permeable inserts, that reproduces the two different intestinal compartments: luminal (apical) and serosal (basolateral) side. Following separate exposure in apical and in basolateral compartments, aminopentol passage through the cell layer (in particular from basolateral to apical direction) was shown, while it was not observed for the parent compound. The different aminopentol distribution between the two compartments of the culture system, and its variation in presence of verapamil or probenecid (P-gp and MRP inhibitors respectively), strongly suggests the involvement of P-glycoprotein in the influx/efflux mechanisms of aminopentol in the intestinal cells, reducing its oral bioavailability.

Constituents of aromatic plants: carvacrol.

Carvacrol is a component of numerous aromatic plants. Up to now, no toxicological data were available. Carvacrol show a weak activity in the mutagenicity studies. Moreover, in the metabolism study, carvacrol has shown to be excreted with urine after 24 h in large quantities or unchanged or as glucoronide and sulphate conjugates. The available data do not allow the assessment of the NOEL. Further toxicological studies are needed.

Toxicology investigations with cell culture systems: 20 years after.

From almost 20 years the "in vitro" model has gained a wide ground in toxicological investigation, providing advanced tools, reliable protocols, mechanistic information. These advancements have been done thanks to different approaches, addressed at improving chemical testing and validating procedures, at exploring the cellular and molecular basis of toxicity, at studying the modifications that xenobiotics undergo in the cellular environment. In this review the most advanced cellular models, the mechanisms of cell death, the techniques to monitor gene activation, following chemical exposure, is highlighted. Moreover the more recent in vitro models to approach the biotransformation issue will be presented.

Evaluation of Fumonisin B(1) and its metabolites absorption and toxicity on intestinal cells line Caco-2.

The aim of the present paper is to investigate intestinal absorption and toxicity of Fumonisin B(1) (FB(1)) and its partially (PHFB(1) and PHFB(2)) and totally hydrolyzed (HFB(1)) metabolites, using the human intestinal cell line Caco-2, a very well known in vitro model of intestinal epithelium for absorption and metabolism studies. Caco-2 cells were treated for 48 h with several toxin concentrations (in the range of 1-138 microM). At the end of exposure period, no significant variation on cell viability has been observed with all chemicals tested, either in undifferentiated cells or in differentiated ones, suggesting a poor toxicity of these mycotoxins for intestinal cells. In any case, FB(1) appears the most active in this respect. For which concerns the cellular absorption, FB(1), PHFB(1) and PHFB(2) are never detected into Caco-2 cells. On the contrary, a dose-dependent absorption of HFB(1) has been observed in differentiated cells, which express enzymatic and metabolic characteristics of mature enterocytes. Thus HFB(1), losing the tricarballylic acid chain, is more bioavailable than FB(1) on intestinal cell, supporting the hypothesis that in risk evaluation of fumonisins exposure its metabolites are also relevant.

In vitro toxicology methods: impact on regulation from technical and scientific advancements.

The impressive advancement of technologies in biomedical research, and particularly in the area of in vitro experimental models, has opened up new possibilities related to co-cultures, micromass or stem cell cultures. Engineered cells to study specific targets and/or mechanisms are also available. Moreover, a very subtle approach in the study of toxicological effects is represented by the very recent genomics and proteomics techniques. New mechanistically based methods could be established from all these approaches, which, once validated, could enter the regulatory procedure. So far, in toxicology, only a few in vitro tests are accepted for regulatory purposes, such as those related to corrosion, phototoxicity and absorption. Many others are in the pre-validation or validation phase. An area where in vitro tests play a key role is the genetic toxicology. In this context, the most recent testing strategies and test methods will be presented, with particular attention to the recently updated guidelines for food additives by the EU Scientific Committee on Food. An improvement in the implementation of validated methods could arise from a better coordination on the matter at national and international levels, the harmonisation of different legislations, and a strict control of the national rules in order to make them up-to-date with respect to validated methods.

Nitric oxide production in Caco-2 cells exposed to different inducers, inhibitors and natural toxins.

The involvement of the NO pathway in several intestinal inflammatory diseases is under investigation. In vitro models may provide a useful approach to better characterise this pathway at the cellular level. For this purpose, we have used Caco-2 cells, which are able to spontaneously differentiate in long-term culture to small intestine enterocytes. The effect of different NO pathway inducers [gamma-interferon (IFN-gamma) and phorbol myristate acetate (PMA)] has been studied. Our results demonstrate that Caco-2 cells constitutively express NOS at very low levels, while the induction with PMA+IFN-gamma triggers the expression of the inducible isoform with a stronger effect starting from day 14 of differentiation. The use of specific inhibitors of gene expression, at transcriptional and translational level, suggests that new synthesis of iNOS mRNA is required, through direct activation of the gene or new synthesis of transcription-required factors, as indicated by CHX inhibition. The morphological alteration induced by PMA+IFN-gamma is reversed by iNOS inhibitor, suggesting that the NO pathway may be involved in the cytoskeletal alterations. The DSP toxins, OA and DTX-1, induce NO production at levels corresponding to their different toxicity, previously detected in Caco-2 cells.

Identification and coupling to adenylate cyclase of three different [(3)H]CGP 12177 binding sites in Caco-2 cell membranes.

In the present investigation the identification of beta -adrenoceptor (beta -ARs) subtypes in the Caco-2 cell line was performed using radiometric assays. beta -ARs were measured using increasing concentrations of the highly specific beta -AR antagonist (-)[(3)H]CGP 12177 (0.06-4 nM), whereas the beta(1)- and beta(2)-AR subtypes discriminated through selective binding assays using the highly selective unlabelled antagonists CGP 20712A and ICI 118551. Atypical beta -ARs were measured using an incubation system formed by higher concentrations (0.6-20 nM) of (-)[(3)H]CGP 12177. beta - Atypical binding site concentrations (69 +/- 5 fmol mg ml(-1)of membrane protein) were higher than beta(1)-ARs (7 +/- 1) and beta(2)-ARs (24 +/- 2), respectively. The different beta -AR subtype affinities were characterized by binding inhibition experiments and the adrenergic agonists displaced the radioligand from its specific binding sites in the following order of potency: isoproterenol > clenbuterol > dobutamine > SR 58611A; for antagonists the order of potency was: propranolol approximately = ICI118551 approximately = CGP20712A. For atypical beta -ARs the order was: SR 58611A > clenbuterol > dobutamine > isoproterenol for agonists and propranolol > CGP 20712A > ICI 118551 for antagonists. As far as in vitro functional studies are concerned, beta -AR subtypes were shown to be coupled to adenylyl cyclase as their stimulation produced cAMP in an amount significantly higher than basal values. cAMP production after stimulation with dobutamine, clenbuterol, isoproterenol, and SR 58611A was measured using a cAMP radioassay kit. The order of efficacy suggested that the stimulation of beta(2)-ARs was the most effective in inducing the activation of cell signalling mechanisms. The identification of functional beta -ARs in a cancer cell line represents the first step in the study of the possible adrenergic control of cellular activities (e.g. proliferation and/or differentiation), which could suggest the use of this cancer cell line as a model for the study of cell activity or possibly new therapeutic strategies.

Evidence for cytochrome P4501A2-mediated protein covalent binding of thiabendazole and for its passive intestinal transport: use of human and rabbit derived cells.

Thiabendazole (TBZ), an anthelmintic and fungicide benzimidazole, was recently demonstrated to be extensively metabolized by cytochrome P450 (CYP) 1A2 in man and rabbit, yielding 5-hydroxythiabendazole (5OH-TBZ), the major metabolite furtherly conjugated, and two minor unidentified metabolites (M1 and M2). In this study, exposure of rabbit and human cells to 14C-TBZ was also shown to be associated with the appearance of radioactivity irreversibly bound to proteins. The nature of CYP isoforms involved in this covalent binding was investigated by using cultured rabbit hepatocytes treated or not with various CYP inducers (CYP1A1/2 by beta-naphthoflavone, CYP2B4 by phenobarbital, CYP3A6 by rifampicine, CYP4A by clofibrate) and human liver and bronchial CYP-expressing cells. The covalent binding to proteins was particularly increased in beta-naphthoflavone-treated rabbit cells (2- to 4-fold over control) and human cells expressing CYP1A2 (22- to 42-fold over control). Thus, CYP1A2 is a major isoenzyme involved in the formation of TBZ-derived residues bound to protein. Furthermore, according to the good correlation between covalent binding and M1 or 5OH-TBZ production, TBZ would be firstly metabolized to 5OH-TBZ and subsequently converted to a chemically reactive metabolic intermediate binding to proteins. This metabolic activation could take place preferentially in liver and lung, the main biotransformation organs, rather than in intestines where TBZ was shown to be not metabolized. Moreover, TBZ was rapidly transported by passive diffusion through the human intestinal cells by comparison with the protein-bound residues which were not able to cross the intestinal barrier. Consequently, the absence of toxicity measured in intestines could be related to the low degree of TBZ metabolism and the lack of absorption of protein adducts. Nevertheless, caution is necessary in the use of TBZ concurrently with other drugs able to regulate CYP1A2, particularly in respect to liver and lung tissues, recognised as sites of covalent-binding.

Apoptosis evaluation in epithelial cells exposed to different chemicals: relevance of floating cells.

The recent increase in understanding of cell death has promoted new approaches in toxicological studies, mainly those dealing with in vitro systems where the evaluation of cell death has been the most widely adopted end-point in measuring the effects of chemical toxicants. The aim of this study was to investigate the possibility of improving the traditional cytotoxicity test protocols in order to produce more specific information on the type of cell death induced by exposure to toxicants. In particular, we characterized the mode of cell death in an established epithelial cell line, HEp-2 cells, which is frequently used in cytotoxicity testing owing to its easy handling and standardization of culture conditions. Reference chemicals for apoptosis and necrosis were selected as controls, together with other molecules that have been shown, in preliminary studies, to induce various morphological and structural modifications in relation to cell death. The results obtained show that: (a) the floating fraction of treated cells gives the clearest picture of the necrotic/apoptotic distribution; (b) morphological analysis is crucial for characterization of apoptosis; (c) more than one cytotoxic end-point is necessary to clearly identify the type of cell death.

Metabolism of furazolidone: alternative pathways and modes of toxicity in different cell lines.

1. The metabolism and cytotoxicity of the antimicrobial nitrofuran drug furazolidone have been studied in Caco-2, HEp-2 and V79 cell lines. Free radical production, metabolite pattern, formation of bound residues, inhibition of cellular replication and protection by the antioxidant glutathione were compared for the three cell lines. 2. All three cell lines produced the same nitro-anion radical with similar kinetics. Little further metabolic breakdown was observed in V79 cells, whereas Caco-2 and HEp-2 cells showed extensive degradation of furazolidone, but with different end patterns. 3. Under hypoxic conditions, the colony-forming ability was extensively impaired in HEp-2 cells whereas the other two cell lines were less affected, suggesting that irreversible damage to DNA occurred prevalently in HEp-2 cells. In V79 cells the absence of oxygen caused a 25-fold increase in the formation of protein-bound residues. 4. Brief exposure to furazolidone caused a 50% loss of endogenous glutathione in Caco-2 cells, but no loss could be detected in V79 and HEp-2 cells. Consistently, when glutathione was depleted by buthionine-[S,R]-sulphoximine (BSO) and diethylmaleate (DEM) treatment, the viability of V79 and HEp-2 cells was minimally affected by furazolidone, whereas that of Caco-2 cells was substantially reduced. 5. It is concluded that the cytotoxicity of furazolidone in these cell lines can be exerted by a number of different mechanisms, possibly related to different metabolic pathways. The cytotoxicity of nitrofuran drugs, therefore, cannot be ascribed to a single toxic intermediate, but in Caco-2 cells furazolidone is extensively metabolized and detoxified by GSH, in V79 is only partially activated and then bound to proteins, whereas in HEp-2, once activated, may react with DNA.

Toxicity of selected plant volatiles in microbial and mammalian short-term assays.

In this study, several short-term microbial and mammalian in vitro assays were used to evaluate cytotoxicity and genotoxicity of four plant volatiles showing antifungal activity: cinnamaldehyde, carvacrol, thymol and S(+)-carvone. All inhibited viability and proliferation of Hep-2 cells in a dose-dependent manner. IC50 ranged from 0.3 mM (cinnamaldehyde) to 0.7 mM (thymol) in viability tests and from 0.2 mM (carvacrol) to 0.9 mM (carvone) in the proliferation test. The morphological analysis suggested an involvement of apoptosis in the cases of carvone, carvacrol and cinnamaldehyde. At nontoxic doses, carvacrol and thymol increased the number of revertants in the Ames test by 1.5-1.7 times, regardless of metabolic activation. In the SOS-chromotest, none of the four plant volatiles caused DNA damage at non-toxic doses. In the DNA repair test, a marked dose-dependent differential toxicity was observed with carvone and, to a lesser extent, with cinnamaldehyde, while with thymol and carvacrol, this effect was less pronounced. In conclusion, the considered in vitro cytotoxicity assays have shown to be sensitive enough to highlight a variety of toxic effects at the cellular level, which can be rather different between chemically closely related compounds, such as isomers.

A contribution to safety assessment of veterinary drug residues: in vitro/ex vivo studies on the intestinal toxicity and transport of covalently bound residues.

1. The gastrointestinal fate of protein-bound residues of the model compound furazolidone (FZD) was investigated in vitro and ex vivo. Protein-bound residues were generated in rat liver microsomes, isolated by solvent extraction and digested with 0.5% hydrochloric acid and Pronase E. 2. During digestion, 3-amino-2-oxazolidinone (AOZ), the side chain of furazolidone, was partly released from bound residues. 3. The absorption of free AOZ and digested protein-bound residues was tested in isolated perfused rat gut segments (IPGS) and in the intestinal cell line Caco-2. Free AOZ was transfered both in the IPGS model and in Caco-2 monolayer cultures, while no indications for passage of bound residues were obtained. 4. No acute toxicity of AOZ or digested food residues respectively was observed in gut segments and Caco-2 cells at concentrations that were substantially above maximum residue levels to be expected in food of animal origin after administration of therapeutic doses. 5. The results demonstrate that digestive processes can alter the chemical nature of drug residues and yield degradation products that may be bioavailable for the consumer. Thus, the covalent binding of xenobiotics to macromolecular tissue constituents cannot necessarily be regarded as an irreversible endpoint of residue bioavailability and toxicity.