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.

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.

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.

Toxicology investigations with cell culture systems.

This review concerns some of the cell culture systems that are most frequently used in toxicology investigations. In particular, it sets out to evaluate the effectiveness of these cell culture systems in assessing the toxic potential of chemicals. Metabolic studies and general and specific toxicology investigations are highlighted. Specific toxicology investigations relate to the effects of the tests substances on the highly specialized functions typical of the cell systems chosen. The general toxicology investigations include most of the other studies where differentiated or undifferentiated cells have been used to evaluate the effects of the tested substances on common basic biochemical processes essential for life. Lastly, we have attempted to focus attention on the most promising applications of cell cultures in toxicology studies for the near future and to identify those areas where further research is needed. Because of the several excellent reviews that already exist, we have decided not to consider cell cultures utilized in screening potential mutagens and carcinogens. We have also excluded investigations of drug therapeutic effects and action mechanisms of drugs.

Structural specificities and significance for coeliac disease of wheat gliadin peptides able to agglutinate or to prevent agglutination of K562(S) cells.

Two peptides corresponding to bread wheat A-gliadin fragments 31-43 and 44-55, well known for their ability to damage the coeliac disease intestinal mucosa both in vitro and in vivo, have been confirmed to be very active in inducing in vitro agglutination of K 562 (S) cells. Removal of six amino acid residues from the carboxy-terminal end of the 31-43 peptide, or of five amino acid residues from the amino terminal end of the 44-55 peptide, resulted in a lower, but still very significant, cell agglutination activity. The peptide consisting of ten amino acid residues with a molecular mass of 1157.5 Da, isolated from durum wheat gliadin, was able to prevent agglutination of K 562 (S) cells induced not only by prolamine peptic-tryptic digests from all the cereals toxic in coeliac disease (i.e. bread wheat, rye, barley and oats), but also by the 31-43 and 44-55 peptides. The ability to protect K 562 (S) cells from agglutination was exhibited to the fullest extent also by all the peptides derived from the 1157.5-Da peptide by five progressive deletions of the terminal carboxylic residue, whereas the sixth consecutive deletion yielded a completely inactive peptide. A similar total loss of activity was observed upon addition of a glycine residue to the amino terminal residue of the 1157.5-Da peptide and all the above-mentioned active peptides derived from it. The remarkable sequence homologies existing between peptides able to induce [Gln-Gln-Gln-Pro and -Pro-Ser-Gln-Gln-] or to prevent [H2N-Gln-Gln-Pro-Gln-Asp-COOH] induction of cell agglutination strongly suggest that all these peptides compete for identical or structurally related binding sites on the cell surface.

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.

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.

Effects of gliadin-derived peptides from bread and durum wheats on in vitro cultures of human cell lines. Implications for coeliac disease pathogenesis.

A peptic-tryptic-cotazym digest, obtained from bread (hexaploid) wheat gliadins under experimental conditions mimicking in vivo protein digestion, was found to reduce in vitro viability of human embryo (MRC-5) and tumor cell (Hep-2) lines. Time of onset and extent of cytotoxic effects were largely dependent on initial peptide concentrations in the culture medium. The presence of 2% fetal calf serum was capable of delaying, but not of preventing, the onset of cytotoxic effects only in MRC-5 cultures. A peptic-tryptic-cotazim digest obtained from durum (tetraploid) wheat gliadins and tested under identical conditions did not show any cytotoxic activity on MRC-5 and Hep-2 cell lines. These results indicate that cell systems are useful to investigate pathogenetic mechanisms of coeliac disease (gluten-dependent enteropathy).

Functional alterations induced by the food contaminant furazolidone on the human tumoral intestinal cell line Caco-2.

Caco-2 cells, which are derived from a human colon carcinoma and are able to differentiate in culture, have been used to study the effect of furazolidone (FZ), a chemical belonging to the nitrofuran family which is frequently used for the prevention of animal infections. Its potentially toxic residues could remain in some food products of animal origin and affect human health. Toxicity has been measured by different parameters, either in undifferentiated cells (day 7 of culture), or on differentiated cells (day 21 of culture). Our results indicate that FZ may seriously affect the proliferating portion of the intestinal mucosa, while the differentiated cells appear to be more resistant. However, the slight effect recorded on the aspecific and specific functions of the differentiated cells may suggest that the specialized portion of the intestine can also be compromised by the drug. Caco 2 cells seem a good model for a deeper investigation of the mechanism involved in the toxic action of FZ.

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.

Cytotoxicity of halogenated benzenes and its relationship with logP.

The in vitro toxicity of a series of environmentally relevant halobenzenes was tested using a Chinese hamster lung fibroblast cell line and its relationship with the logarithm of octanol/water partition coefficient (logP) was investigated. Since we wanted to study the direct biological activity of the parent substances, we have used the V-79 cell line that does not express phase I metabolic activities. Moreover, because of the available knowledge on the substances, we decided to perform the colony-forming ability test (CFA) and to analyse the DNA damage by a cytofluorimetric assay. To identify the concentration range at which the toxic effect could be detected, a prescreen with the neutral red assay has been performed. All the substances tested were positive in the CFA, but, according to the concentration values inhibiting this activity by 50%, they can be divided into two groups of differing toxicity. The FACScan analysis shows for the majority of the halobenzenes a clear hypodiploid peak. A good correlation between values describing the concentration that inhibits CFA by 50% and logP was found, indicating that it depends on the hydrophobic properties of the compounds and that logP is a suitable descriptor.

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.

Established cell lines for safety assessment of food contaminants: differing furazolidone toxicity to V 79, HEp-2 and Caco-2 cells.

In vitro models, preferentially derived from human tissues, may be valuable tools to study the biotransformation and toxicity of compounds that may be present as residues in food products. Such residues may represent a risk to human health, and therefore call for increased testing. Three established cell lines were used to study the toxic effect of furazolidone (FZ), a widely used veterinary drug: HEp-2 cells, derived from a human larynx carcinoma, previously used in toxicity screening of several compounds; Caco-2 cells, derived from a human colon adenocarcinoma, able to differentiate partially in culture, and V 79, a fibroblast cell line derived from Chinese hamster lung, widely used to assess direct toxicants. Various toxicity parameters were used, primarily dealing with cell death and cell proliferation. In all cell lines FZ at a concentration of 5 micrograms/ml caused a marked decrease in cell viability and especially in cell proliferation. Inhibition of DNA synthesis has also been observed, even if at higher concentrations. However, only in V 79 cells was the decrease in cell number accompanied by a marked increase in lactate dehydrogenase leakage due to membrane damage. Moreover, the surviving V 79 cells, after removal of FZ, fully recovered from the effect of the drug, as shown by their full capacity to attach to dishes and to form colonies. Surviving cells of the other two cell lines showed much poorer colony-forming ability. Exposure of Caco-2 cells and, to a lesser extent, HEp-2 cells, caused a marked increase in oxygen consumption, that possibly was due to redox cycling of the initially formed radical nitro anion. Biotransformation of the drug by all three cell lines was accompanied by the formation of protein-bound metabolites, HEp-2 being the most active cells. The toxic effects recorded show that cell lines provide a sensitive system in toxicity assessment. Moreover, it may be suggested that a battery of cell lines, including some of human origin, as well as a battery of endpoints, may be of help in addressing further specific mechanistic investigations.

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.

Expression and synthesis of fibronectin and laminin by an intestinal epithelial cell line.

An intestinal epithelial cell line (IEC-17), undergoing a process of progressive morphological differentiation, was analysed for expression and synthesis of the extracellular matrix glycoproteins, fibronectin (FN) and laminin (LM). FN and LM cell surface expression was detected by immunoelectron microscopy, while intracytoplasmic accumulation was shown by immunofluorescence. 35S-methionine metabolic labelling was also performed to demonstrate FN and LM synthesis by IEC-17. We have compared two different maturation stages of the cell culture and have found that either early epithelial monolayer cells or later multistratified organoid structure cells expressed and produced large amounts of both proteins. These results indicate that FN and LM are constantly present during the process of IEC-17 organoid maturation: we can hypothesize that the two proteins act as mediators of cell to cell and cell to substrate adhesion interactions and, probably, have an active regulatory role in the process of intestinal epithelial cell differentiation.

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.

In vitro toxicity and formation of early conjugates in Caco-2 cell line treated with clenbuterol, salbutamol and isoxsuprine.

Caco-2, human intestinal cell line able to differentiate in long-term culture, has been used to assess the cytotoxicity of the beta-agonists clenbuterol, salbutamol and isoxsuprine, also used at high doses to obtain lean meat in food producing animals, and to investigate the eventual in vitro formation of early conjugates of these compounds. For this purpose, the cells have been characterized for the activity of UDP-glucuronyltransferase, which is present and increases in the differentiated cells, and for the beta-receptors' binding characteristics, which are those of beta 1 and beta 2 subtypes. Isoxsuprine was shown to be the most toxic, followed by clenbuterol and salbutamol. Conjugates have been observed after incubation of the cells both with the lowest isoxsuprine and the highest salbutamol concentrations. No conjugates were detected in the case of clenbuterol.

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.