The use of pig hepatocytes for biotransformation and toxicity studies

The three main objectives of this study were, (1) to investigate the possibility to isolate viable hepatocytes from liver samples of pigs, (2) to study their use for biotransformation and toxicity studies, and (3) to demonstrate the value of this model, in particular in the field of residue toxicology.The main result from this study was the fact that hepatocytes can be successfully isolated from samples of pig livers obtained at the slaughterhouse, and that the cells retained a high viability as demonstrated by parameters like the exclusion of trypan blue. Furthermore, the cells were able to attach themselves to untreated culture dishes, and to form monolayer cultures, which could be kept alive for at least seven days.Even more important was the fact that the cells retained the capability of biotransforming xenobiotics into metabolites that were also observed in vivo.Initial studies with suspension cultures (Chapter 3) demonstrated that these cells were capable of biotransforming a number of different substrates by very diverse pathways, like the deethylation of 7-ethoxy-coumarin and subsequent glucuronidation of the product 7-hydroxy-coumarin, the acetylation of sulfadimidine and deacetylation of N 4-acetyl-sulfadimidine, and the transformation of furazolidone, partly resulting in the formation of the open chain cyano-metabolite. In vivo the N 4-acetyl metabolite of sulfadimidine has been shown to be the only significant metabolite. In the case of furazolidone, the only in vivo metabolite identified sofar, is the cyano-metabolite, although similar to in vitro it appears to be formed in relatively small amounts.In comparison with suspension cultures, monolayer cultures have a number of additional advantages, such as the extra viability criterium of the ability of cells to form a monolayer, the longer survival of the cells, and the physical separation between the cells and the culture medium. However, by the results from studies with hepatocytes from other animal-species, it was shown that the prolonged use of monolayer cultures is often limited by a decrease in the activity of certain biotransformation enzymes (Chapter 1). Therefore, special attention was paid to the possible existence and impact of this phenomenon in monolayer cultures of pig hepatocytes, by incubating cells with the substrates described above. There was a decrease in absolute levels of cytochrome P-450, and in the cytochrome P-450 related deethylation of 7-ethoxy-coumarin. In the case of the N 4-acetylation of sulfadimidine, an increase in the activity was observed, which was paralleled and possibly caused by a decrease in the reverse activity, i.e. the deacetylation of the N 4-acetyl metabolite. Furazolidone was converted relatively slowly during the first 24 h of the monolayer cultures, as compared to the next 72 h, where little variation was observed. Therefore, the effect of ageing apparently varies from one compound to another, which may be a strong argument for the use of freshly isolated cells. However, a recovery period may allow the cells to recover from the isolation procedure. It was e.g. shown that intracellular levels of reduced glutathione (GSH) were sometimes very low in freshly isolated cells, but increased rapidly during the first 24-48 h of the cultures (Chapters 3, 5 and 7). Since GSH has an important role in protecting the cells against the adverse effects of many xenobiotics (see also Chapter 7), freshly isolated cells may be oversensitive to the toxic properties of certain drugs. Further experiments were therefore primarily performed with monolayer cultures, after a 24 h recovery period.In the case of the anabolic steroid B-nortestosterone, a substantial change in the metabolite pattern, due to the ageing of the cells, was only observed after 72 h (Chapter 4). This compound was rapidly oxidized by the cells, initially resulting in the formation of norandrostenedione, which was than hydroxylated and glucuronidated at its C-15 position. No differences were observed between cells from female and castrated male pigs, which is consistent with the "feminizing" effect of castration. In the absence of essential information on the in vivo biotransformation of this compound, three pigs were injected with the hormone. The fact that the two main metabolites could subsequently be identified in the urine of these pigs, clearly demonstrates the value of the model for the identification of important metabolites.The high viability of the cells made the model very useful for toxicity studies. In Chapter 5, a number of different cellular functions were chosen as end-points to study and compare the effects of five different nitrofuran drugs. These parameters, i.e. the leakage from the cells of the cytosolic enzyme lactate dehydrogenase (LDH), due to membrane damage, the incorporation of 14C-leucine into newly synthesized proteins, the intracellular levels of reduced and oxidized glutathione, and the metabolism of pyruvate, could all be measured easily and reproducibly. Using these parameters, clear similarities but also differences were observed between the drugs. At the highest dose levels there was a small but significant increased leakage of LDH, especially in the case of nitrofurantoin. With all drugs, this was accompanied by a decrease in intracellular LDH-levels, that could only partly be accounted for by the loss into the medium. This might be explained by the decreased incorporation of 14C-labelled leucine into proteins after exposure to all nitrofurans. This inhibiting effect was especially clear on the proteins excreted into the medium. Furthermore, incubation with all nitrofurans resulted in an increase in intracellular levels of oxidized glutathione (GSSG), which is in agreement with the fact that reduction of the nitro-group of these drugs may result in redox-cycling and the subsequent formation of H 2 O 2 (Fig. 7.1). Somewhat unexpected was the fact that, with the exception of nitrofurantoin, there was no decrease in intracellular GSH levels. The most sensitive, and possibly most specific parameter, was the inhibition of the pyruvate metabolism by all nitrofurans, except nitrovin, as concluded from the accumulation of pyruvate and lactate in the medium. In a more detailed study with furazolidone, it was shown that the cells recovered only very slowly from the latter effect, which is consistent with previous results from in vivo studies. These and some of the other effects, might underlie some of the many adverse effects reported in both human and animal patients treated with these type of drugs, like eg. the increased plasma levels of pyruvate and lactate. However, in general such in vivo information was too limited to completely validate the model at this point, and further in vivo studies should be carried out for this purpose. Nevertheless, the study demonstrates the value of the model for studying the toxicity and underlying mechanisms of xenobiotics.In Chapter 6, another specific toxic effect of furazolidone observed in vivo but not in vitro, i.e. the inhibition of the enzyme monoamine oxidase (MAO), was studied in more detail. In this study, intact cells were used instead of homogenates for measuring the MAO-inhibition, thereby offering the advantage of the presence of an intact biotransformation potential. Using p-tyramine as a substrate for the MAO-enzymes, the effect of the known inhibitors clorgyline and iproniazid could easily be measured. Incubation of hepatocytes with furazolidone resulted in a thusfar unknown reversible MAO-inhibition. However, the irreversible inhibition reported in in vivo studies, was only observed after incubation of cells with 3-amino-2-oxazolidinone (AOZ), the side chain of furazolidone, and the proposed metabolite β-hydroxyethylhydrazine (Fig. 6.1). The time-related increase in the MAO-inhibition by 3-amino-2-oxazolidinone, indicated that the compound must be transformed into a metabolite capable of inactivating MAO-enzymes irreversibly, possibly by covalent binding. The results from this study are in support of the hypothesis of Stem et al., with regard to the role of the side-chain in the MAO-inhibition observed in vivo.The great perspectives of the model in the field of residue toxicology are further demonstrated in Chapters 7 and 8, where the different features of the model were used to obtain more information on the identity and properties of the protein-bound metabolites of furazolidone. As a follow-up of the microsomal studies performed by Vroomen et al., the role of glutathione in the biotransformation and toxicity of this drug was further investigated (Chapter 7). Using 14C-labelled furazolidone, the drug was shown to be extensively transformed by monolayer cultures, resulting in the formation of the cyano metabolite, a large number of more hydrophilic compounds, and proteinbound metabolites. These results are in agreement with previous in vivo studies, and confirm the value of the model.Similar to the studies described in Chapter 5, furazolidone had apparently no effect on intracellular GSH-levels. However, both by inhibition of the GSH-synthesis and by labelling of the intracellular GSH with 35S, it could be demonstrated that exposure of cells to furazolidone actually resulted in an increased loss of intracellular GSH, but that this effect was compensated for by an increased GSH synthesis. The increased GSH loss was accompanied by increased medium concentrations of GSSG and a second compound, most likely formed after excretion of GSH by the cells. However, no evidence could be obtained for the formation and excretion of the glutathione- conjugate of furazolidone that was previously detected in microsomal incubations. Similarly, the presence of unstable protein-bound thiol conjugates could not be demonstrated. Therefore, it seems unlikely that these type of metabolites are responsible for the previously observed high bioavailability of protein-bound metabolites in rats, that had been fed with muscle tissue of a piglet treated with furazolidone, as well as the subsequent formation of protein-bound metabolites in tissues of those rats.An alternative explanation for this observation was given in Chapter 8, where the time- and dose- related formation of protein-bound metabolites of furazolidone in pig hepatocytes was studied in more detail. Furthermore, by using 14C-furazolidone, either labelled in the nitrofuran or in the oxazolidinone part of the molecule, strong indications were obtained for the presence of both parts of the parent drug in proteinbound metabolites. Therefore, a new method for the detection of these type of compounds was developed, based on the release and detection of the 3-amino-2-oxazolidinone side-chain of furazolidone. Using this relatively simple method it was demonstrated that at least 75% of the metabolites bound to protein of hepatocytes still contained this side-chain, showing for the first time that at least part of these metabolites resemble the original structure of the parent compound. With the same method 3-amino-2-oxazolidinone could be released from 15-25% of the protein-bound metabolites present in liver samples from piglets treated with 14C-furazolidone, even after a storage period of seven years at -40 °C.The difference between in vivo and in vitro studies, with respect to the fraction of bound metabolites from which AOZ can be released, might be ascribed to the role of the stomach in the in vivo biotransformation of furazolidone. The acid hydrolysis of the azomethine bond of furazolidone may result in the release of 3-amino-2-oxazolidinone (Fig. 9.1), which is subsequently absorbed and metabolized into a compound, possibly via β-hydroxyethylhydrazine (Fig. 6.1), capable of binding covalently to proteins (type I). As pointed out in Chapter 6, there are strong indications that this mechanism underlies the irreversible type of MAO-inhibition. The bound metabolites observed in vivo would therefore partly be related to the formation of a reactive intermediate from the sidechain (type I), and partly to intermediates containing an intact AOZ side-chain, because they are formed by the reduction of the nitro-group (type II), as shown in Fig. 7.1. In the case of incubation of furazolidone with microsomes and hepatocytes at pH 7, it seems likely that the formation of bound metabolites can be ascribed to the latter type of reactive intermediates (type II). Since AOZ can only be released from this type of metabolites, this difference could explain the discrepancy with in vivo . At this point it cannot be excluded that a similar mechanism, ie. the acid hydrolysis of the azomethine bond and subsequent biotransformation of AOZ, is responsible for the high bioavailability of bound residues and the formation of new bound metabolites as observed in rats after consumption of meat from pigs treated with furazolidone.This is the first time that, based on experimental data, strong indications are obtained for the fact that bound residues of a compound may still exert a toxic action, once orally absorbed by the consumer. Therefore, the presence of such protein-bound residues should be reduced as much as possible, and adequate analytical methods, such as presented in this study, should be developed in order to control this. It should be mentioned, that the possible toxic potential of bound-metabolites of other compounds, depends very much on the structure and properties of those compounds, and should therefore be considered from one case to another. In general these results subscribe the importance of including metabolites in the risk-assessment of compounds used in veterinary practice.The results of the present study clearly demonstrate the large potential of in vitro models with hepatocytes from large domestic animals. Regarding the high costs and special requirements that go along with in vivo studies with larger animals, the use of such in vitro models may eventually not only result in a reduction in the number of animals used for experiments, but even more, offer an improved chance to obtain the more detailed information on the species-related biotransformation and toxicity of xenobiotics, as requested more and more by regulatory authorities. In general, the use of such models may therefore contribute to the development of new drugs with a minimal health risk for the consumer and less adverse effects for the animal.Furthermore, this study shows that knowledge on the identity and nature of metabolites, may result in the development of new analytical methods, that ran be used to control the sound use of drugs and growth-promoting agents in veterinary practice.