Characterization of membrane permeability alterations induced in vero cells by Clostridium perfringens enterotoxin

Alterations in plasma membrane permeability induced by Clostridium perfringens enterotoxin were studied using Vero (African green monkey kidney) cells which were radioactively labeled with four markers of different molecular size. The markers were α-amino[¹⁴C]isobutyric acid (M_r 103), ³H-labeled nucleotide (M_r approx. 300), ⁵¹Cr label (M_r approx. 3000) and [³H]RNA (M_r > 25 000). Over a 2 h period, enterotoxin caused significant release of aminoisobutyric acid, nucleotides and ⁵¹Cr label but not RNA. The effects of enterotoxin on label release were dose- and time-dependent. The rate of release of markers was dependent upon their size. Permeability alterations could be detected within 15 min with a high dose of enterotoxin. Gel chromatography of released material was used to determine that markers of M_r 3000 but not 25 000 leaked from permeabilized cells. It was concluded that enterotoxin is producing functional ‘holes’ of limited size in the membrane. Permeability changes due to enterotoxin treatment differed between confluent and non-confluent (growing) cells. We propose that the primary action of the enterotoxin is to interact with the plasma membrane and produce functional ‘holes’ of defined size. The resultant alterations in membrane permeability cause the loss of essential cellular substances which inhibits processes such as macromolecular synthesis and eventually leads to cell deterioration and death.     

Osmotic stabilizers differentially inhibit permeability alterations induced in Vero cells by Clostridium perfringens enterotoxin

Using a sensitive Vero (African green monkey kidney) cell model system, studies were performed to further investigate whether Clostridium perfringens enterotoxin acts via disruption of the colloid-osmotic equilibrium of sensitive cells. Enterotoxin was shown to cause a rapid loss of intracellular 86Rb+ (Mr approx. 100) with time- and dose-dependent kinetics. The enterotoxin-induced release of intracellular 86Rb+ preceded the loss of two larger labels, 51Cr label (Mr approx. 3500) and 3H-labeled nucleotides (Mr less than 1000). The osmotic stabilizers, sucrose and poly(ethylene glycol), differentially inhibited enterotoxin-induced larger label loss versus 86Rb+ loss. Further, enterotoxin was shown to cause a rapid influx of 24Na+ that was not significantly inhibited by osmotic stabilizers. Additional studies demonstrated that lysosomotropic agents were not protective against characteristic enterotoxin-induced membrane permeability alterations or morphological damage. Taken collectively, these results are consistent with an action for enterotoxin which involves a disruption of the osmotic equilibrium.     

Effect of gamma radiation on resting B lymphocytes. I. Oxygen-dependent damage to the plasma membrane results in increased permeability and cell enlargement

Although the susceptibility of resting B lymphocytes to radiation-induced interphase death is well known, the mechanism by which this occurs is not understood. In this report, we use three measures of plasma membrane integrity (increase in cell volume, uptake of trypan blue, and release of 51Cr) to assess the effect of radiation on the resting B cell plasma membrane. The delivery of 500 to 1000 rad caused the majority of resting B cells to enlarge slightly, whereas 3000 rad caused virtually all of the cells to approximately double in size within 3 to 4 hr. Measurement of the release of 51Cr from resting B cells revealed a similar relationship between the dose of radiation and the loss of radioactive label. Trypan blue exclusion was also found to diminish as a function of radiation dose. An analysis of a variety of lymphoid cells suggested that sensitivity to the membrane damaging effects of gamma radiation was in the order of resting B cells greater than resting T cells greater than a long-term L3T4+ T cell clone greater than a B cell lymphoma. LPS-induced B cell blasts treated with 3000 rad were equivalent to 1000 rad-treated resting B cells. The effects of the gamma radiation could be ameliorated by excluding oxygen (a diradical molecule that can potentially enhance the generation and propagation of highly reactive free radicals) at the time of irradiation, or by adding the free radical scavenging agent cysteamine. These data are compatible with the hypothesis that gamma radiation results in damage to the plasma membrane of resting lymphocytes via the generation of highly reactive free radical species. This damage is reflected in a rapid increase in plasma membrane permeability and swelling of the cells, and may play a major role in causing interphase death.     

Effects of Ca2+ and other cations on the action of Clostridium perfringens enterotoxin

We investigated the role of extracellular Ca²⁺ in the Clostridium perfringens enterotoxin-induced alteration of the permeability of the plasma membrane. Enterotoxin released ⁸⁶Rb and ⁵¹Cr from the Vero cells preloaded with the isotope. In the presence of EGTA, however, it released ⁸⁶Rb but not ⁵¹Cr. The binding of enterotoxin to the cells was not influenced by Ca²⁺ or Mg²⁺. The effects of various cations on the enterotoxin-induced ⁵¹Cr release was also studied. The release depended on extracellular Ca²⁺ but not on Mg²⁺; it was inhibited by each of Zn²⁺, La³⁺ and Co²⁺. Zn²⁺ and Co²⁺ also inhibited ⁵¹Cr release caused by the enterotoxin previously bound to the cell membrane. In contrast, antibody against enterotoxin did not neutralize the toxin once it was bound to the Vero cells. When the cells were treated with enterotoxin, ⁴⁵Ca influx occurred and reached the plateau in a few minutes, as did ⁸⁶Rb release.     

Structural and enzymatic disorganization of biological membranes in rat liver cells in thermal burns

Permeability of hepatocyte cell membrane was studied from the release into blood of hepatospecific enzymes and from 5'-nucleotidase activity in plasma membranes. A study was also made of membrane permeability of mitochondria, lysosomes and microsomes in liver cells of burnt rats from the level of non-sedimented activity and activity of malate dehydrogenase, succinate dehydrogenase, cathepsin D and glucose-6-phosphatase in appropriate organelles. Permeability of cell and lysosomal membranes was demonstrated to be disordered within the first hours after burn. One day after burn generalized disturbance of membrane permeability in the cell was observed, followed by the release into cytosol of organelles template enzymes and a decrease in the activity of membrane-bound enzymes in these organelles. The alterations persisted during 7 days of observation.     

New insights into the cytotoxic mechanisms of Clostridium perfringens enterotoxin

Clostridium perfringens type A isolates producing the 35 kDa enterotoxin (CPE) are an important cause of food poisoning, human non-foodborne gastrointestinal (GI) disease, and some veterinary GI diseases. Studies using CPE knock-out mutants confirmed the importance of enterotoxin expression for the enteric virulence of CPE-positive type A isolates. CPE action involves formation of a series of complexes in mammalian plasma membranes. One such CPE-containing complex (of approximately 155 kDa) is important for the induction of plasma membrane permeability alterations, which are responsible for killing enterotoxin-treated mammalian cells. Those membrane permeability changes damage the epithelium, allowing the enterotoxin to interact with the tight junction (TJ) protein occludin. CPE:occludin interactions result in formation of an approximately 200 kDa CPE complex and internalization of occludin into the cytoplasm. That removal of occludin (and possibly other proteins) damages TJs and disrupts the normal paracellular permeability barrier of the intestinal epithelium, which may contribute to CPE-induced diarrhea. Recent studies demonstrated that low CPE doses kill mammalian cells by inducing a classic apoptotic pathway involving mitochondrial membrane depolarization, cytochrome C release, and caspase 3/7 activation. In contrast, high enterotoxin doses induce oncosis, a proinflammatory event. Thus, inflammation may also contribute to the GI symptoms of patients whose intestines contain high CPE levels. In summary, CPE is a unique, multifunctional toxin with cytotoxic, TJ-damaging, and (probably) proinflammatory activities.     

Classification of microbial, plant and animal cytolysins based on their membrane-damaging effects of human fibroblasts

38 cytolytic agents of mainly microbial origin were investigated with respect to membrane-damaging activity on human diploid fibroblasts. Increased plasma membrane permeability was measured as leakage of three defined cytoplasmic markers of various sizes: alpha-aminoisobutyric acid, uridine nucleotides and ribosomal RNA. The relative leakages of these markers, caused by different concentrations of the various cytolysins, yielded a leakage pattern for each substance. Five distinct types of leakage patterns were obtained. These were transformed into numerical expressions by calculating the ratios between the amounts of cytolysin needed to release 50% of the nucleotide and ribosomal RNA markers and the amounts required to release 50% of the alpha-aminoisobutyric acid marker (ED50 ratios). A classification of the cytolysins into five groups was arrived at on the basis of the different types of leakage patterns with the aid of reference cytolysins with well-known mechanisms of membrane interaction. These groups comprised: (1) detergent-like agents, (2) agents interacting with only certain constituents of the cell membrane, (3) agents interacting with specific receptor molecules in the membrane, (4) agents inducing small functional holes of a definable size, and (5) agents inducing only a very limited increase in plasma membrane permeability. The system may be useful for characterization and differentiation of new cytolytic agents of various sources as it divides membrane-damaging agents into separate groups on the basis of their principal function on intact human cells.     

Polyene macrolide antibiotic cytotoxicity and membrane permeability alterations I. Comparative effects of four classes of polyene macrolides on mammalian cells

The relationship between polyene macrolide-induced early membrane damage and cytotoxicity in B1 (hamster), B82 (mouse), and RAG (mouse) cells has been investigated. Filipin (FIL) induced the greatest immediate damage, as monitored by ⁵¹Cr release, followed by mediocidin (MED), amphotericin B-deoxycholate (Fungizone®) (FZ) and pimaricin (PIM). For long term effect, PIM was the least toxic followed by MED, FZ, and FIL as indicated by 24-hour survival, 72-hour viability, and growth rate of cells. In evaluating polyene macrolide-induced permeability alterations and cytotoxicity two types of interactions with mammalian cells were found: (1) cell toxicity at polyene macrolide levels not eliciting immediate membrane permeability changes; and (2) immediate membrane damage without long range toxicity.     

Limited and selective transfer of plasma membrane glycoproteins to membrane of secondary lysosomes

Radioactive galactose, covalently bound to cell surface glycoconjugates on mouse macrophage cells, P388D1, was used as a membrane marker to study the composition, and the kinetics of exchange, of plasma membrane-derived constituents in the membrane of secondary lysosomes. Secondary lysosomes were separated from endosomes and plasma membrane on self-forming Percoll density gradients. Horseradish peroxidase, taken up by fluid-phase pinocytosis, served as a vesicle contents marker to monitor transfer of endosomal contents into secondary lysosomes. Concurrently, the fraction of plasma membrane-derived label in secondary lysosomes increased by first order kinetics (k = [56 min]-1) from less than 0.1% (background level) to a steady-state level of approximately 2.5% of the total label. As analyzed by NaDodSO4 PAGE, labeled molecules of Mr 160-190 kD were depleted and of Mr 100-120 kD were enriched in lysosome membrane compared with the relative composition of label on the cell surface. No corresponding selectivity was observed for the degradation of label, with all Mr classes being affected to the same relative extent. The results indicate that endocytosis-derived transfer of plasma membrane constituents to secondary lysosomes is a limited and selective process, and that only approximately 1% of internalized membrane is recycled via a membrane pool of secondary lysosomes.     

Apoptosis is induced in cells with cytolytic potential by L-leucyl-L-leucine methyl ester.

Previous studies have demonstrated that the selective toxicity of leucyl-leucine methyl ester (Leu-Leu-OMe) for cytotoxic lymphocytes and myeloid cells is dependent on intracellular conversion to membranolytic metabolites by the acyl transferase activity of the granule enzyme dipeptidyl peptidase I (DPPI) that is enriched in these cells. The mechanism of cell death remained unclear, however, and was the subject of the experiments reported here. When human U937, HL60, or THP-1 myeloid tumor cell lines or murine CTLL-2 cells were treated with Leu-Leu-OMe, early release of both cytosolic 51Cr and soluble [3H]TdR labeled DNA fragments was observed, whereas antibody + C treatment of these cells caused only 51Cr release. Killing of U937 or THP-1 cells by incubation with the lysosomotropic amino acid methyl ester, Phe-OMe also induced only 51Cr release without evidence of DNA fragmentation. Preincubation with Zn2+, a known inhibitor of endonuclease activity prevented Leu-Leu-OMe-induced 51Cr or [3H]TdR release from these cell lines, but had no effect on antibody + C or Phe-OMe-induced 51Cr release. Zn2+ also prevented Leu-Leu-OMe-mediated killing of normal human CD16+ NK cells. Zn2+ had no inhibitory effect on Leu-Leu-OMe uptake or intracellular conversion to (Leu-Leu)n-OMe metabolites by these cell lines. Moreover, Zn2+ did not inhibit 51Cr release from nonnucleated E or nucleated U937 targets induced by extracellular production of DPPI-generated metabolites of Leu-Leu-OMe. Thus, killing of cytotoxic lymphocytes and myeloid cells by Leu-Leu-OMe appears to be dependent on generation of metabolites with membranolytic properties, but cell death induced by this process does not involve simple lysis of the plasma membrane. Rather, intracellular production of DPPI generated (Leu-Leu)n-OMe metabolites appears to trigger, an additional Zn(2+)-sensitive process that is associated with induction of apoptosis in cells with cytolytic potential.     

Responses of mouse lymphocytes to extracellular ATP. II. Extracellular ATP causes cell type-dependent lysis and DNA fragmentation

Extracellular ATP (ATPo) caused dose-dependent lysis of YAC-1 and P-815 mouse tumor cells. This event, assessed by 51Cr release, was accompanied by sustained depolarization of the plasma membrane potential and Ca2+ influx. Plasma membrane depolarization and Ca2+ influx occurred within a few seconds of ATPo addition to both cell types, whereas 51Cr was released without apparent lag in YAC-1 cells and after 2 h in P-815 cells. Furthermore, a rise in [Ca2+]i was required for ATPo-dependent lysis of YAC-1 but not P-815 cells. In P-815 cells, ATPo caused an early and [Ca2+]i-independent DNA fragmentation that occurred at lower nucleotide concentrations than those required to trigger 51Cr release. Instead in YAC-1 cells very low concentrations of ATPo caused early lysis (ED50 for lysis about 200 microM) accompanied by only barely detectable DNA fragmentation. Previous studies disclosed that lymphokine-activated killer cells are fully resistant to the membrane-perturbing effects of ATPo. We show that lymphokine-activated killer cells also do not undergo DNA fragmentation even in the presence of high ATPo concentrations. This study complements previous observations on the lytic effects of ATPo and shows that this nucleotide can also cause DNA fragmentation, one of the earliest target cell alterations observed during CTL-mediated lysis.     

Temperature dependence of membrane permeability in cultured cells exposed to benzene and phenol

The effects of benzene (20mM) and phenol (20mM) on the passive membrane permeability of cultured cells of neuronal (N1E115), glial (138MG), muscle (L6) and liver (BRL123) origin were studied. Permeability changes were assayed as alterations in the release of radioactivity from cells preloaded with [³H]2-deoxy-D-glucose. Temperature plots of the efflux revealed that benzene and phenol caused altered flux rates in N1E115 cells without affecting the activation energy. Also in 138 MG cells the activation energy was unaffected by both compounds, while phenol decreased the flux rates. Benzene had no effect on the efflux in L6 cultures, whereas a low activation energy efflux was induced by phenol. In BRL123 cells both benzene and phenol induced a dual (high and low) activation energy efflux.     

Methods to Assess Beta Cell Death Mediated by Cytotoxic T Lymphocytes

Type 1 diabetes (T1D) is a T cell mediated autoimmune disease. During the pathogenesis, patients become progressively more insulinopenic as insulin production is lost, presumably this results from the destruction of pancreatic beta cells by T cells. Understanding the mechanisms of beta cell death during the development of T1D will provide insights to generate an effective cure for this disease. Cell-mediated lymphocytotoxicity (CML) assays have historically used the radionuclide Chromium 51 (⁵¹Cr) to label target cells. These targets are then exposed to effector cells and the release of ⁵¹Cr from target cells is read as an indication of lymphocyte-mediated cell death. Inhibitors of cell death result in decreased release of ⁵¹Cr. As effector cells, we used an activated autoreactive clonal population of CD8⁺ Cytotoxic T lymphocytes (CTL) isolated from a mouse stock transgenic for both the alpha and beta chains of the AI4 T cell receptor (TCR). Activated AI4 T cells were co-cultured with ⁵¹Cr labeled target NIT cells for 16 hours, release of ⁵¹Cr was recorded to calculate specific lysis Mitochondria participate in many important physiological events, such as energy production, regulation of signaling transduction, and apoptosis. The study of beta cell mitochondrial functional changes during the development of T1D is a novel area of research. Using the mitochondrial membrane potential dye Tetramethyl Rhodamine Methyl Ester (TMRM) and confocal microscopic live cell imaging, we monitored mitochondrial membrane potential over time in the beta cell line NIT-1. For imaging studies, effector AI4 T cells were labeled with the fluorescent nuclear staining dye Picogreen. NIT-1 cells and T cells were co-cultured in chambered coverglass and mounted on the microscope stage equipped with a live cell chamber, controlled at 37°C, with 5% CO₂, and humidified. During these experiments images were taken of each cluster every 3 minutes for 400 minutes.Over a course of 400 minutes, we observed the dissipation of mitochondrial membrane potential in NIT-1 cell clusters where AI4 T cells were attached. In the simultaneous control experiment where NIT-1 cells were co-cultured with MHC mis-matched human lymphocyte Jurkat cells, mitochondrial membrane potential remained intact. This technique can be used to observe real-time changes in mitochondrial membrane potential in cells under attack of cytotoxic lymphocytes, cytokines, or other cytotoxic reagents.     

The JAM-assay: optimized conditions to determine death-receptor-mediated apoptosis

Apoptosis induced by interaction of members of the TNF-/TNF-receptor superfamily has been considered as a major mechanism of cell-mediated cytotoxicity. For functional analysis, the 51Cr release assay has been widely used, which requires loss of membrane integrity in the apoptotic target cell. However, loss of membrane integrity is a late event during apoptosis and therefore only late apoptotic cells will be detected by this method. In contrast, the JAM-assay first described by Polly Matzinger has been demonstrated to be more sensitive than the 51Cr release assay, since this method is dependent on DNA-fragmentation which precedes loss of membrane integrity in most apoptotic cells. The JAM-assay is easier to perform, less expansive, and safer than the current standard (51)Cr release assay. Therefore, this article will focus on optimized conditions of the JAM-assay to detect and quantitate Fas (CD95/Apo-1)-induced apoptosis as an example of death-receptor-mediated cytotoxicity.     

Plasma Membrane and Chromaffin Granule Characteristics in Digitonin-Treated Chromaffin Cells

Digitonin permeabilizes the plasma membranes of bovine chromaffin cells to Ca2+, ATP, and proteins and allows micromolar Ca2+ in the medium to stimulate directly catecholamine secretion. In the present study the effects of digitonin (20 microM) on the plasma membrane and on intracellular chromaffin granules were further characterized. Cells with surface membrane labeled with [3H]galactosyl moieties retained label during incubation with digitonin. The inability of digitonin-treated cells to shrink in hyperosmotic solutions of various compositions indicated that tetrasaccharides and smaller molecules freely entered the cells. ATP stimulated [3H]norepinephrine uptake into digitonin-treated chromaffin cells fivefold. The stimulated [3H]norepinephrine uptake was inhibited by 1 microM reserpine, 30 microM NH4+, or 1 microM carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP). The data indicate that [3H]norepinephrine was taken up into the intracellular storage granules by the ATP-induced H+ electrochemical gradient across the granule membrane. Reduction of the medium osmolality from 310 mOs to 100 mOs was required to release approximately 50% of the catecholamine from chromaffin granules with digitonin-treated chromaffin cells which indicates a similar osmotic stability to that in intact cells. Chromaffin granules in vitro lost catecholamine when the digitonin concentration was 3 microM or greater. Catecholamine released into the medium by micromolar Ca2+ from digitonin-treated chromaffin cells that had subsequently been washed free of digitonin could not be pelleted in the centrifuge and was not accompanied by release of membrane-bound dopamine-beta-hydroxylase. The studies demonstrate that 20 microM of digitonin caused profound changes in the chromaffin cell plasma membrane permeability but had little effect on intracellular chromaffin granule stability and function. It is likely that the intracellular chromaffin granules were not directly exposed to significant concentrations of digitonin. Furthermore, the data indicate that during catecholamine release induced by micromolar Ca2+, the granule membrane was retained by the cells and that catecholamine release did not result from release of intact granules into the extracellular medium.     

Evidence for intracellular sodium pumps in permeabilized Xenopus laevis oocytes.

Ouabain binding was studied in Xenopus laevis oocytes permeabilized by detergents. The behaviour of markers showed that 10 microM-digitonin selectively disrupts the plasma membrane. In the presence of ATP, oocytes permeabilized at 10 microM-digitonin bound no more ouabain molecules than were required to abolish active 86Rb+ uptake in the intact cells. However, the ouabain binding capacity increased approx. 2-fold when inner membranes were disrupted by SDS or excess digitonin, as judged from the accompanying release of the lysosomal marker beta-hexosaminidase. The results suggest that oocytes have a large internal pool of functional sodium pumps.     

Fas/FasL-independent activation-induced cell death of T lymphocytes from HIV-infected individuals occurs without DNA fragmentation

We assessed the effects of activation with phorbol myrystic acetate (PMA) and ionomycin on peripheral blood mononuclear cells (PBMC) from HIV-infected individuals by (51)Cr release, propidium iodide (PI) uptake, electron microscopy, and DNA analysis. Up to 70% (51)Cr release was induced from PBMC of HIV-infected individuals, versus up to 26% (51)Cr release from PBMC of non-HIV-infected volunteers. Flow cytometry identified mostly T cells undergoing activation-induced cell death (AICD). The kinetics of (51)Cr release and the effects of cold target inhibitors were consistent with cell-mediated cytotoxicity. Certain anti-CD3 antibodies or extracellular Ca(2+) chelation prevented AICD, but antagonistic anti-Fas antibodies, caspase inhibitors, and cycloheximide had no effect. The antioxidants thiourea and N-acetylcysteine reduced AICD, indicating a role for oxidative stress. Electron microscopy revealed plasma membrane disruption with nuclear integrity, while DNA analysis showed intact chromosomal DNA. This form of T cell AICD triggered by PMA and ionomycin differs from classical apoptosis in the absence of either caspase involvement or DNA fragmentation.     

Use of peritoneal macrophages labeled with Cr51 and cultivated in vitro as target cells for quantitative assessment of the cytotoxic activity of immune T-lymphocytes]

Peritoneal macrophages previously labeled in a suspension with 51Cr and grown in the form of a discontinuous monolayer on glass, pretreated with poly-l-lysin were used as target cells. This permitted to estimate the cytotoxic activity of immune lymphocytes by 51Cr release for the period of 4 and 20 hours of incubation with the target. 51Cr release from such a target incubated with normal lymphocytes for 20 hours did not exceed 10-20% of the maximum release. Application of a 2% solution of sodium dodecylsulfate ensured a 100 per cent solubilization of labeled macrophages growing on the glass surface, and permitted the cytotoxic effect to be determined by measuring the label remaining in the intact cells.     

The synaptic vesicle membrane: origin, axonal distribution, protein components, exocytosis and recycling

The paper discusses functional and molecular aspects of the synaptic vesicle membrane during its life cycle. The distribution of the synaptic vesicle membrane compartment in an entire cholinergic neuron is monitored using colloidal gold labelling and a monoclonal antibody against the synaptic vesicle membrane protein SV2. This provides new insights concerning vesicle origin and fate in the various compartments of the neuron. A new synaptic vesicle membrane protein (svp25) of Mr 25,000 with properties similar to synaptophysin as well as a synaptic vesicle binding phosphoprotein of the presynaptic membrane (Mr 92,000) likely to be involved in vesicle exocytosis are described. The membrane compartment recycled on induced transmitter release contains synaptic vesicle but not plasma membrane markers and encloses both newly synthesized transmitter and a sample of extracellular medium.     

Stereospecific pharmacokinetics and toxicodynamics of ketorolac after oral administration of the racemate and optically pure enantiomers to the rat

To determine the stereospecific pharmacokinetics and gastrointestinal permeability (GI) changes (surrogate measures of toxicity) in the rat following oral administration of S, R, and racemic ketorolac (KT), optically pure enantiomers (S and R 2.5 mg/kg), and racemic KT (5 mg/kg) were administered orally to male Sprague-Dawley rats and plasma samples were collected for 6 h post-dose for pharmacokinetic assessments. KT-induced changes in GI permeability were assessed using sucrose and 51Cr-EDTA as markers of gastroduodenal and distal intestinal permeability, respectively. After the racemate, R-KT was predominant in plasma (AUC S/R, 0.45). No significant differences in pharmacokinetic indices were evident following administration of the racemate as compared with individual enantiomers. In plasma, there was only negligible S-KT after administration of R-KT. After S-KT, on the other hand, AUC of R-KT was found to be 6.7% of that of S-KT. Both permeability markers showed considerable interanimal variability. Gastroduodenal permeability was significantly increased from baseline by the racemate but not by either of the two enantiomers administered alone. Permeability to 51Cr-EDTA was not significantly increased above baseline for any of the treatments. The plasma concentration of R-KT found after administration of S-KT may be from the < 2% chiral impurity which appears magnified due to its slower clearance as compared with its antipode. There is no evidence of a pharmacokinetic interaction between the enantiomers. Since 2.5 mg/kg S-KT is somewhat less toxic on the gastroduodenum than 5 mg/kg racemate, it may be a safer alternative to the latter, at least in the rat model.