Alternatives to Laboratory Animals

The two renal cell lines, MDCK (distal tubule) and LLC-PK₁ (proximal tubule), have been used for toxicity testing of three different X-ray contrast media: the ionic monomer Isopaque, the ionic dimer Hexabrix and the non-ionic monomer Omnipaque. The cells were grown to confluency in monolayer cultures in a chemically-defined serum-free medium before the contrast media were added, to give final concentrations corresponding to 0–100mg iodine/ml. Toxicity was assessed by cell viability and by biochemical assays of marker enzymes. The results demonstrate a concentration-dependent toxic effect from the contrast media on cellular appearance, and on the activity of brush border and lysosomal enzymes. The non-ionic X-ray contrast media appeared to be less toxic than the ionic contrast media investigated.

The effects of NiSO₄, K₂CrO₄, PbAc₂, CrCl₃ and CdAc₂ on gap junctional intercellular communication (GJIC) were studied. Two exposure procedures were used — A: an established monolayer of cells was exposed to the compound, and B: the compound was mixed with the cells before seeding. The apparent inhibition of GJIC by CrCl₃ and CdAc₂ could be accounted for by cytotoxicity, i.e. disruption of physical contacts between cells. PbAc₂ had no effect on GJIC. The effects of NiSO₄ and K₂CrO₄ depended on the exposure procedure and cell type studied. Specifically, K₂CrO₄ induced a significant increase in GJIC in BPNi cells by procedure A, and a decrease by procedure B. The cytotoxic effects of NiSO₄ were examined in more detail. The toxic concentrations varied widely between the methods used. The results indicate that the effect of compounds on GJIC may vary with different exposure procedures, and that inhibition of GJIC is not due to general cytotoxic effects.

An established epithelial cell line (A6) from a South African clawed toad (Xenopus laevis) kidney was used as a model for the corneal epithelium of the eye in order to determine ocular irritancy. When grown on Millipore filter inserts, A6 cells form a monolayer epithelium of high electrical resistance and generate a trans-epithelial potential difference. These two easily-measured electrophysiological endpoints showed a dose-related decrease after exposure for 24 hours to seven selected chemicals of different ocular irritancy potential. It was demonstrated that both trans-epithelial resistance and potential ranked closely with in vivo eye irritancy data and correlated well (r = 0.96) with loss of trans-epithelial impermeability of Madin-Darby canine kidney (MDCK) cells, detected by use of a fluorescein leakage assay.

Primary cultured rat skeletal muscle cells were used to determine concentration-dependent effects of the first twenty MEIC chemicals on three endpoints, spontaneous contractility and viability after 1 and 24 hours, and glucose consumption during 24 hours of exposure. The contractions of cultured muscle cells depend on spontaneous electrical activity of the excitable cell membranes. The majority of the test compounds inhibited contractility at concentrations which affected neither viability nor glucose consumption. Most of these compounds are known to interact with excitable membranes in a site-specific or non-site-specific manner, thereby causing therapeutically intended or toxic effects. The results indicate that inhibition of spontaneous contractility of cultured skeletal muscle cells may reflect important non-cytotoxic biological activities of test chemicals which might be more relevant for their acute toxicity than cytotoxic action.

The cytotoxicity of the first twenty chemicals on the MEIC list was evaluated by determining the haemolytic activity of the chemicals. All the tested chemicals, except malathion, were haemolytic. With digoxin and sodium fluoride, however, the maximum degree of haemolysis achieved remained less than 50%. The EC10 for the chemicals varied from 0.38mM (digoxin) to 4,830mM (ethylene glycol). Comparing our data for the first ten MEIC chemicals with data obtained with other in vitro test systems, they correlated well with data reported for human lymphocytes (r = 0.93), rat hepatocytes (r = 0.98) and 3T3 cells (r = 0.95). However, in comparison with these tests, the haemolysis assay was less sensitive. The EC10 for the first ten MEIC chemicals correlated rather well (r = 0.84) with human lethal blood concentrations, but the haemolysis assay underestimated the toxicity of most of the chemicals and particularly underestimated the toxicity of digoxin.

In attempts to overcome the problem for cell culture cytotoxicity tests represented by highly volatile test materials, the wells of 96-well tissue culture plates were sealed individually, either by means of adhesive plastic film spread across the whole plate or by the addition of a small amount of paraffin oil to each well. Use of the adhesive film reduced the growth of control 3T3-L1 murine fibroblast-like cells by nearly 40%, whereas the addition of paraffin oil did not affect cell growth. The latter method was therefore used to compare the cytotoxicities of 27 liquid and 11 water-soluble solid test materials in sealed and unsealed wells, using the FRAME 72-hour kenacid blue method.

Of the 27 liquids tested, acetaldehyde, acetone, acetonitrile, allyl alcohol, ethanol, ethyl acetate, methanol, pyrrole and propanol had significantly greater cytotoxic effects in oil-sealed than in unsealed wells. The cytotoxicities of acetic acid, aniline, butanol, 1,2-dibromomethane, dimethyl formamide, dimethyl sulphoxide, glycerol, hydrochloric acid, 2-methoxyethanol, octanol, pentanol, sulphuric acid, tributyltin chloride, trichlorobenzene, trifluoroacetic acid and Tween 20 were not affected by the layer of paraffin oil. Heptanol and hexanol showed significantly lower cytotoxic effects in oil-sealed than in unsealed wells, possibly because of interactions between these alcohols and the oil, since certain alcohols are more soluble in oil than in culture medium. Only one of the 11 solid materials tested, sodium hydroxide, showed greater cytotoxic effects in oil-sealed wells, possibly because of an effect on the culture medium buffering system.

The use of paraffin oil as a partial well sealant in 96-well tissue culture plates promises to improve the applicability of in vitro cytotoxicity tests to volatile test materials, by improving confidence that their toxicities have not been underestimated.

Cultured hepatocytes and hemisphere neurons from chick embryos and mouse neuroblastoma cells were exposed to carbon tetrachloride (CC1₄; 0, 1, 2, 3 and 4mM) for 1 hour, using a perfusion system developed for studying the effects of volatile substances. In the perfused cultures, three parameters were compared: lipid peroxidation, membrane integrity and cellular respiration. In addition, cytochrome C oxidase activity was determined after incubation of cell homogenates with CC1₄. A concentration-dependent increase in lipid peroxidation and membrane permeability was found in the neuroblastoma cells. The hepatocytes responded to a lesser extent with respect to membrane permeability and their lipid peroxidation did not differ from that of controls. The hepatocytes responded with a 35% decrease in respiration when exposed to 3mM CC1₄, and a 20% decrease in cytochrome C oxidase activity after treatment with 1.5mM CCl₄. In the neuronal cells, much smaller decreases in respiration were found and their cytochrome C oxidase activity remained unaffected. These results are very similar to those obtained after incubation in a closed chamber system. However, the perfused cells were found to be less sensitive to CCl₄ than cells exposed under static conditions.

Progress in the implementation of in vitro methods in testing procedures for non-genetic toxicity is limited by poor or missing validation of the various approaches. Many efforts have focused on determinations of general cytotoxicity in vitro using proliferating, non-differentiated cell lines. The aim of our present studies is to evaluate the relevance to acute systemic toxicity of three in vitro systems using differentiated mammalian cells (bovine spermatozoa, primary cultured rat hepatocytes, primary cultured rat skeletal muscle cells) and also one in vitro system which uses a continuous cell line (BALB/C 3T3 cells). This combination of in vitro systems covers various aspects of cellular toxicity and permits the determination of the intrinsic activity of chemicals with respect to general cytotoxicity, selective cytotoxicity and interference with selected cell-specific functions. In order to contribute to the establishment of a basis for the selection of methods to be included in test batteries in the future, we are contributing to the MEIC programme. In the present communication, results for the first ten chemicals of the MEIC list are presented and discussed.

Cell volume was determined by measuring [¹⁴C]-3-O-methyl glucose uptake in astroglial-enriched primary cultures. Control cell volume was 3.20μl/mg protein. After incubation in 10 ⁵M HgCl₂ for 60 minutes, there was a 71% increase in cell volume. This increase was partially inhibited in the presence of the α₁ receptor agonist, phenylephrine, or by the α₂ receptor agonist clonidine, and was completely reversible by their respective antagonists, prazosine and yohimbine. The β receptor agonist, isoproterenol, which in itself increased cell volume, and 5-hydroxytryptamine (5HT) did not affect the HgCl₂-induced changes in cell volume. 10 ⁵M CH₃HgCl increased cell volume by 26% after 30 minutes of incubation. This increase was not significantly influenced by adrenoceptor agonists or 5HT.

It therefore seems that mercurial-induced changes in cell volume can be regulated by astroglial receptor stimulation.

In vivo exposure to paraquat, a bipyridyl herbicide, is associated with selective necrosis of type I and type II alveolar pneumocytes. We examined the influence of paraquat on cell growth, DNA and protein synthesis in human and rat lung epithelial and fibroblast cell lines, to determine the effects and affinity of the chemical in vitro for cells of various origins. Cells were incubated with [³H]-thymidine or [³H]-proline in the absence or presence of 0.05–10mM paraquat for periods of 6 or 24 hours. Cell cultures grown in the presence of 50–200μM paraquat exhibited significantly lower cell numbers than comparable controls after 3 days. The suppression of DNA synthesis by paraquat was dose-dependent, but did not rely on the length of exposure. Paraquat inhibited protein synthesis and selectively decreased net production of low molecular weight proteins. NADPH levels were the same in both control and treated cultures. [¹⁴C]-Paraquat was accumulated during a 4-hour exposure period via an active non-competitive process. The results suggest that lung fibroblasts are as sensitive to the herbicide as epithelial cells. In addition, paraquat appears to interfere with transcription and/or translation by its ability to generate superoxide radicals, rather than by depletion of reducing equivalents.

Isolated human platelets represent a model system which could be used in in vitro cell toxicology for mechanistic studies on the short-term effects of cytotoxic compounds. We have recorded changes in the concentration of free cytosolic Ca²⁺ ([Ca²⁺]_i) with Fura 2 in isolated human platelets treated with sphingosine, a compound suggested to be a specific inhibitor of protein kinase C (PKC). We huve shown that sphingosine is able to induce a rapid and transient release of Ca²⁺ from the intracellular stores in platelets. Furthermore, a massive efflux of released Ca²⁺ was shown. These effects of sphingosine on the human platelet might leave the cell incapable of performing Ca²⁺-dependent reactions, including PKC-mediated phosphorylation, upon stimulation subsequent to sphingosine-treatment.

BALB/C 3T3 mouse fibroblasts or CHO cells grown in 96-well microtitre plates were used for the neutral red viability assay. The cytotoxicity and ability of microsomes or an S9 mixed function oxygenase to activate cyclophosphamide were compared. A pronounced cytotoxic effect of both activation systems was found using CHO cells, but not using 3T3 cells. 3T3 cells were more sensitive than CHO cells to the active metabolites of cyclophosphamide. Cyclophosphamide was more efficiently converted to toxic metabolites by microsomes than by the S9 mix. A concentration of microsomes corresponding to 0.032mg protein per ml of medium was shown to be optimal for the activation of cyclophosphamide in both cell lines.

The toxicity of allyl alcohol and several glycols (ethylene glycol, 1,2-propanediol, 1,3-propanediol, methoxyethanol, and the glycol ether dioxane) was studied in cultures of 3T3 cells and in co-cultures of 3T3 cells with microcarrier-attached hepatocytes. Metabolism-mediated effects on the cytotoxicity to 3T3 cells were recorded by differences in the growth of the cultures exposed in the presence or absence of hepatocytes. Hepatocyte viability was determined by depletion of intracellular lactate dehydrogenase and effects on the biotransformation ability of hepatocytes were assessed by determination of O-deethylation of 7-ethoxycoumarin (EOD activity). Allyl alcohol was the only substance more toxic to the hepatocytes than to 3T3 cells cultured in the absence of hepatocytes. Toxicity to 3T3 cells of allyl alcohol, ethylene glycol, and 1,3-propanediol, but not of 1,2-propanediol, methoxyethanol and dioxane, was markedly enhanced when the cells were co-cultured with hepatocytes. The results indicate that the toxicity of allyl alcohol, ethylene glycol, and 1,3-propanediol, to 3T3 cells depends on the formation of active metabolites. For ethylene glycol and 1,3-propanediol, growth of 3T3 cells in co-cultures was reduced at concentrations without effects on hepatocyte viability. Co-culture of 3T3 cells with microcarrier-attached rat hepatocytes represents a suitable approach for the in vitro evaluation of metabolism-mediated cytotoxicity.

The toxicities of the therapeutic drugs amitriptyline and diazepam, the metal compounds methyl mercury and triethyllead, and the essential fatty acids linoleic acid and α-linolenic acid, were determined in a system of rat glial cells in culture with microglia activation as the parameter for evaluation of toxicity. The markers of activated microglia — increased expression of complement receptor 3 and transformation of microglia to phagocytes — were evaluated by immunocytochemical staining with the monoclonal antibody OX-42 and after incubation with fluorescent latex beads. Concentrations which activated microglial cells were compared to EC50 values from acute toxicity tests, and to concentrations which in previous studies activated oligodendroglial cells. The results show that the drugs amitriptyline and diazepam, and fatty acids linoleic acid and α-linolenic acid, activated microglia at 10–25 times lower levels when compared to the EC50 values. Microglia activation was induced by α-linolenic acid and amitriptyline at about 10 times lower levels when compared to linoleic acid and diazepam. The microglia activating concentrations of the metal compounds were, 1,000 times lower, when compared to acute toxicity values. The activation of microglia was initiated by the same concentrations of the compounds that in previous studies were found to activate oligodendroglial cells.

Continuous amulatory peritoneal dialysis (CAPD) is a well-established treatment for renal failure. Peritonitis is the most important and serious complication to CAPD. The predisposition of patients to contract peritonitis has been attributed to functional impairment of immunoactive cells in the peritoneal cavity due to cytotoxicity of the dialysis solutions used.

A model for biocompatibility studies of CAPD solutions is presented, in which the migration and phagocytotic capacity of normal isolated human polymorphonuclear granulocytes (PMN) were examined after exposure to the test solutions. The methods proved to be highly reproducible.

The commercially-available lactate-based CAPD solutions tested reduced PMN function, mostly due to their low pH. Cell function was improved by elevation of pH, but not to the control level, when the cells were exposed to RPMI — a standard cell culture medium. A new bicarbonate-based CAPD solution proved to be less cytotoxic than the lactate-based ones, and is at present undergoing clinical evaluation.

The cytotoxicities of the first ten MEIC chemicals were assayed by estimation of IC50 values on cultures of the fungus Geotrichum candidum. Colony radial growth rates were measured between 24 and 48 hours after inoculation on solid media containing different concentrations of the chemicals. The IC50 values obtained in this study were higher than the human lethal blood concentrations obtained for the chemicals, but good correlation (r = 0.958) was found between human lethal blood concentrations and IC50 values obtained from growth rate measurements of G. candidum. Ferrous sulphate, methanol, ethanol and diazepam gave linear relationships between decrease in growth rates and concentration of the chemical. With ethylene glycol, the growth rate decreased linearly with the logarithm of ethylene glycol concentration.

Within the framework of the MEIC (the multicentre evaluation of in vitro cytotoxicity) programme, the long-term cytotoxicity of the first twenty MEIC chemicals was investigated on MRC-5 cells. After treatment of the cells with each chemical for six weeks, the PI50 value was determined. The PI50 is the concentration (mM) of test compound required to induce a 50% reduction in total cell protein content. The results were compared with the acute PI50 values previously measured after treatment for 24 hours with the same compounds in Hep G2 cells. A good correlation between these two sets of data (r² = 0.94) was observed, with the exception of digoxin. Together with other preliminary MEIC validation results, this study indicates that a good alternative cytotoxicity test may be found for the prediction of human long-term toxicity.

The present paper deals with the general anaesthetic effect of organic solvents. The emphasis is on cell membranes as models for predicting the adverse reactions of the human nervous system to organic solvents, and for studying the mechanisms of their anaesthetic action. Human and rat erythrocyte and platelet membranes, rat synaptosomal membranes and cultured neural cell membranes were exposed in vitro to various concentrations (3–30mM) of benzene, toluene, o-xylene and styrene in the incubation mixtures. After exposure to a solvent for 30 minutes, the membrane integral protein integrity was determined by measuring ATPase and acetylcholinesterase activities. All measurements were performed in triplicate from 3 or 4 independent experiments with each exposure of solvent and each membrane isolation. The enzyme inhibition was dose-dependent, and differed after solvent treatment in all membrane preparations tested in increasing order, as follows: benzene, toluene, styrene, o-xylene. The results support the protein theories on the mechanism of the anaesthetic action of various industrial organic solvents. The integral membrane proteins contain hydrophobic pockets which may bind anaesthetics, e.g. solvents which change the properties of integral enzymes. The effects of solvents in peripheral models are quite similar to those in neural membranes in vitro. Therefore, these membrane models can serve as test models when the narcotic potencies of the solvents are being evaluated, or at least for screening purposes and for studies of the mechanisms of solvent neurotoxicity.

This study shows that the Artemia assay, which is usually performed by incubation for a 24-hour period in artificial sea water, can also be performed in phosphate buffered saline (PBS) at pH 7.2, or in a cell culture medium, both of which are used in toxicity assays employing mammalian cells. Thus, by using the equivalent media for incubation, toxic effects in the Artemia assay and toxic mechanisms can more accurately be compared with results obtained in mammalian cell toxicity. Survival of control animals is good for 72 hours, provided that infection can be avoided. A pH greater than 6 is essential for good survival of Artemia Salina, and a pH greater than 10.5 should be avoided. Because of the risk of infection at low saline concentrations, a decreased incubation time of 16 or 12 hours is recommended.

The lethal concentrations (LC10 and LC50) in the 24-hour Artemia assay of the first ten chemicals in the MEIC programme were measured in PBS, and the results compared with those from a previous study of the effects in a 10-minute acute ATP leakage assay, specifically indicating lesions in the cell membrane. The EC10 and EC50 values for the alcohols in the Artemia assay were 35–75% of the corresponding values in the ATP leakage assay. For paracetamol and amitriptyline, the EC10 and EC50 values in the Artemia assay were 2–16% of the corresponding values in the ATP-leakage assay. The greatly increased toxicity of the two compounds in the animals may be related to systemic effects. FeSO₄ was very toxic to Artemia salina at a concentration of lug/ml, but it did not cause leakage of ATP from cultured cells, even at a concentration of 8,000μg/ml, showing that widely different mechanisms of interaction with FeSO₄ are measured by the two assays. A lower toxicity of polygodial, a sesquiterpenoid unsaturated dialdehyde, in cell culture medium, was obvious in the Artemia assay. Thus, some factors in cell culture medium must, by interacting with the toxic molecule, protect the animals against the toxicity of polygodial, as was previously found in cultured cells.

Recent trends in the field of in vitro toxicology have centred around the validation of in vitro methods. The ultimate goal is to obtain pertinent data with the minimum of effort. In our laboratory, we have used toxicological methods based on the evaluation of cell morphology and cell proliferation. A method suitable for this purpose is time-lapse microcinematographic (or video) recording of cellular changes, which we used for many years. For practical in vitro toxicity testing, however, this method is far too complicated. Therefore, we have tried to develop a simple modification for the evaluation of cell morphology and cell proliferation, which would still allow for a basic time-dependent analysis. Comparison of detailed microcinematographic analysis with analysis according to our new proliferation assay is demonstrated with cisplatin as the toxicant. We believe that a time-dependent approach could improve the in vitro assessment of toxicity.

The use of a microporous membrane as an interface between a cell culture and a substance dissolved into a non-aqueous, non-cytotoxic medium made it possible to study the cytotoxicity of non-hydrosoluble substances, without addition of any adjuvant for dispersion or solubilisation. The LC50s of raw materials widely used in cosmetic and pharmaceutical products were established on skin human fibroblasts.

This method was used for the study of cosmetics and toiletries which could be tested directly in their actual galenic forms. A relationship between exposure time and cytotoxicity was established for 50 products and a cytotoxicity index was calculated. The results were compared with those of the Draize irritation test on the rabbit eye. This comparison showed a very good correlation between the calculated cytotoxicity index and the ocular irritation index (p < 0.001).

From the results of this preliminary study, it appears that the method could provide an alternative to animal studies on the ocular tolerance of cosmetics and toiletries.

Normal human epidermal keratinocytes were differentiated on Millipore Millicell CM microporous membranes overlaid with an acellular cross-linked collagen gel. The extracellular calcium concentration was elevated, and stratification of the monolayer to a 20-cell thick multilayer ensued within 5–7 days. Transmission electron microscopy of this stratified, human epidermal model (HEM) revealed typical ultrastructural markers of differentiation, including numerous desmosomes, keratinohyalin granules, keratin filaments, lamellar-type bodies, and squamous-shaped cells in the apical region, but not the basal region, of this HEM. Preparations lifted to the air–liquid interface had electron-dense apical cells suggestive of cornification. An electron-dense lamina appeared on the basal surface in both submerged and air–liquid interface cultures, suggesting that this structure can be produced without the help of underlying dermal cells and is independent of the air–liquid interface conditions. The uniform thickness of the epidermis, the absence of dermal fibroblasts, the lack of contraction of the epidermis, and the thinness of the underlying collagen gel, made this HEM a suitable candidate for use in an automated, fluorescence multiple endpoint assay. Fluorescent probes monitoring lysosomal integrity, epidermal permeability and plasma membrane integrity were monitored using the CytoFluor 2300 fluorescent plate reader. Our data suggest the feasibility of using multilayered human tissues for automated multiple endpoint analysis.

The toxicity of two essential trace elements, the heavy metals zinc and copper, has been studied in Caco2 cells, a human colon carcinoma cell line able to differentiate in culture into mature enterocytes.

The ID50 for zinc and copper in Caco2 cells was determined from dose-effect curves. The ID50 is the same (500μM) for zinc and copper in undifferentiated cells, but in differentiated cells the ID50 for zinc (1mM) is higher than the ID50 for copper (600μM). Total protein synthesis is more severely affected by both metals in undifferentiated Caco2 cells than in differentiated cells. This difference could be explained by the different intracellular accumulation of the two metals at the two stages of differentiation. Copper inhibits total protein synthesis more than zinc, at both stages of differentiation. A protective role of zinc-induced metallothionein is discussed in relation to these results.