Alternative approaches to the safety assessment of macronutrient substitutes.

Traditional toxicological procedures have only limited application to the safety assessment of macronutrient substitutes. Experience indicates that spurious effects are often encountered when macronutrients or their replacements are fed to rodents at high dietary levels. These effects may results in nutritional imbalances that lead secondarily to adverse physiological consequences including cancer, renal disease, or reproductive effects. In approaching the safety assessment of macronutrient substitutes, consideration needs to be given to designing and implementing a safety assessment program which acknowledges the physical, chemical, and biological properties of the substance. Factors such as molecular size, physical state, solvent properties, hydrolysis potential, digestibility, absorption potential, and metabolic fate must be well established prior to selection of appropriate test models. Armed with this information, many potential undesirable physiological effects of the substances can be predicted, thus precluding the need for a full spectrum of animal testing. Predicted physiological and metabolic effects, however, should be characterized using in vitro methods and confirmed with in vivo models. Initial short-term toxicity screening tests with rodents should be carried out to identify unanticipated systemic toxicity. Testing in laboratory animals and trials in humans should then proceed with more appropriate models that are specially selected to assess the significance of predicted outcomes, to characterize dose-response relationships, and to identify possible needs to modify the product to mitigate adverse physiological consequences. These might include physical changes to alter particle size, chemical changes to modify digestibility, or nutrient supplementation to overcome impacts on nutrient availability. Thoughtful selection of appropriate and relevant models based on the physical, chemical, and biological properties of test substances will provide more rational approaches to safety assessments and avoid the pitfalls of routine application of traditional tests.

Peroxidase/hydrogen peroxide--or bone marrow homogenate/hydrogen peroxide--mediated activation of phenol and binding to protein.

1. 14C-Phenol was metabolized by rat bone marrow homogenate and H2O2. The homogenate catalyst, however, was inactivated by preincubation with H2O2, presumably due to inactivation of the enzyme(s) involved in phenol metabolism. 2. The majority of the metabolized 14C-phenol was bound to bone marrow proteins. o,o'-Biphenol and p,p'-biphenol were the principal non-protein-bound products. Ascorbate was unable to remove phenol oxidation products bound to protein, although o,o'-biphenol recovery from the reaction mixture was markedly enhanced. Prior alkylation of protein thiols with N-ethylmaleimide decreased the binding of 14C-phenol oxidation products to bone marrow proteins by only 10-20%. 3. 14C-Phenol (200 microM) metabolism by horseradish peroxidase (10 micrograms) and H2O2 (200 microM) also resulted in extensive binding to externally added bovine serum albumin. The absorption spectrum of 14C-phenol oxidation products bound to bovine serum albumin was similar to that of bound oxidation products of o,o'-biphenol but not of p,p'-biphenol. 4. Protease digestion of bovine serum albumin bound 14C-phenol oxidation products, followed by ethyl acetate extraction, extracted 75% of the 14C, indicating that most of the binding is probably non-covalent. Up to 32% of the 14C-phenol oxidation products binding to bovine serum albumin may be covalent, since derivation with dinitrofluorobenzene and extraction under acid, but not alkaline, conditions extracted the 14C. The percentage of metabolites covalently bound to bovine serum albumin was increased to 59% when horseradish peroxidase concentration was decreased to 0.2 micrograms. 5. The thiol groups of bovine serum albumin were unaffected by o,o'-biphenol oxidation products, slightly decreased by phenol oxidation products, but were completely depleted by p,p'-biphenol oxidation products. 6. These results indicate that o,o'-biphenol oxidation products are responsible for much of the 14C-phenol binding to protein.

Modulation of benzoquinone-induced cytotoxicity by diethyldithiocarbamate in isolated hepatocytes.

The copper-chelating thiol drug diethyldithiocarbamate protected isolated hepatocytes from benzoquinone-induced alkylation cytotoxicity by reacting with benzoquinone and forming a conjugate which was identified by fast atom bombardment mass spectrometry as 2-(diethyldithiocarbamate-S-yl) hydroquinone. In contrast to benzoquinone, the conjugate was not cytotoxic to isolated hepatocytes. The thiol reductant dithiothreitol had no effect on benzoquinone-induced alkylation cytotoxicity. However, inactivation of catalase in the hepatocytes with azide and addition of the reducing agent ascorbate markedly enhanced the cytotoxicity of the conjugate but did not affect benzoquinone-induced cytotoxicity. Furthermore, inactivation of glutathione reductase and catalase in hepatocytes greatly enhanced the cytotoxicity of the conjugate and caused oxidation of GSH to GSSG. The conjugate also stimulated cyanide-resistant respiration, which suggests that the conjugate undergoes futile redox cycling resulting in the formation of hydrogen peroxide which causes cytotoxicity in isolated hepatocytes only if the peroxide detoxifying enzymes are inactivated. Diethyldithiocarbamate does, however, protect uncompromised isolated hepatocytes from benzoquinone cytotoxicity by conjugating benzoquinone, thereby preventing the electrophile from alkylating essential macromolecules. Diethyldithiocarbamate therefore changed the initiating cytotoxic mechanism of benzoquinone from alkylation to oxidative stress, which was less toxic.

Molecular mechanisms for bromotrichloromethane cytotoxicity in isolated rat hepatocytes.

1. Bromotrichloromethane added to isolated rat hepatocytes resulted in increased cell death as determined by trypan blue uptake. Toxicity increased in a concentration-dependent fashion between 2.0-5.0 M bromotrichloromethane. 2. Lipid peroxidation (malondialdehyde) increased in a time-dependent fashion but in contrast to toxicity reached a maximum level at 2.0 mM bromotrichloromethane. 3. Hypoxia increased the toxicity of bromotrichloromethane three-fold but only decreased the amount of lipid peroxidation to a small degree. 4. In spite of this poor correlation between toxicity and lipid peroxidation, the antioxidant butylated hydroxyanisole and the iron chelator desferal protected the cells from toxicity under both aerobic and hypoxic conditions and prevented lipid peroxidation. 5. During treatment with bromotrichloromethane, cellular glutathione levels slowly decreased and oxidized glutathione appeared in the media. The addition of cystine to the incubation media prevented the formation of extracellular oxidized glutathione, indicating that cellular glutathione had leaked from the cell during treatment and was oxidized in the incubation media. Although this suggested that glutathione does not play a protective role against bromotrichloromethane toxicity, diethyl maleate-pretreatment of the cells to decrease glutathione levels markedly increased bromotrichloromethane toxicity. 6. The addition of ascorbic acid to the incubation media increased bromotrichloromethane toxicity. This was attributed to the reductive activation of bromotrichloromethane in an iron and oxygen-dependent reaction. 7. It was concluded that peroxidation of essential phospholipids contributes to bromotrichloromethane-induced hepatocyte cytotoxicity.

Myeloperoxidase catalysed cooxidative metabolism of methimazole: oxidation of glutathione and NADH by free radical intermediates.

The myeloperoxidase catalysed oxidation of methimazole in the presence of NADH or GSH resulted in oxygen uptake suggesting that metabolism proceeded via a one electron mechanism. The GSH was oxidised to GSSG and the thiyl radical could be trapped with DMPO while NADH was oxidized to NAD+. Metabolism proceeded without the inactivation of the enzyme myeloperoxidase. Myeloperoxidase catalyzed oxidation of other substrates which proceed via one electron intermediates; 2,6-dimethylphenol, N,N,N',N'-tetramethyl-phenylenediamine and luminol, were all stimulated by methimazole providing further evidence for a methimazole free radical. The presence of iodide stimulated the oxidation of methimazole but inhibited the oxygen uptake in the presence of GSH or NADH suggesting that metabolism in this case proceeded by a two electron mechanism. In contrast, another S-thioureylene drug, thiourea; did not cause oxygen uptake when oxidised in the presence of GSH or NADH indicating that the myeloperoxidase oxidation of thiourea proceeded primarily by a two electron mechanism. The horseradish peroxidase catalysed one electron oxidation of p'p'-biphenol, and 3,3',5,5'-tetramethylbenzidine was reversibly inhibited by methimazole and thiourea by preventing the accumulation of oxidation products via reductive mechanisms whereas the reversible inhibition of guaiacol and luminol oxidation was the result of competitive inhibition. With p,p'-biphenol, and 3,3',5,5'-tetramethylbenzidine unstable adduct formation could be demonstrated.

The metabolism of N-acetyl-3,5-dimethyl-p-benzoquinone imine in isolated hepatocytes involves N-deacetylation.

3,5-Dimethyl-N-acetyl-p-benzoquinone imine (3,5-dimethyl-NAPQI) was cytotoxic to isolated hepatocytes from Sprague Dawley rats at levels between 200 and 300 microM. It rapidly oxidized intracellular glutathione within 10 sec, with the formation of oxidized glutathione. The cytotoxicity of 3,5-dimethyl-NAPQI could be prevented over a 3.5-hr period with the carboxylesterase inhibitor bis(p-nitrophenyl) phosphate, indicating that cytotoxicity involved N-deacetylation. The N-deacetylated product could be trapped with glutathione as 3-(glutathion-S-yl)-4-amino-2,6-dimethylphenol in 3,5-dimethyl-NAPQI-treated hepatocytes but not in hepatocytes pretreated with bis(p-nitrophenyl) phosphate, indicating that N-deacetylation activity had been inhibited. 3,5-Dimethyl-NAPQI was readily N-deacetylated by rat liver microsomes, in contrast to 3,5-dimethylacetaminophen. The latter was also not cytotoxic to hepatocytes at up to 2 mM. The N-deacetylated product 4-amino-2,6-dimethylphenol rapidly underwent autoxidation to form 2,6-dimethylbenzoquinone imine and was highly cytotoxic to hepatocytes at 200-300 microM. The latter reacted with glutathione to give the above conjugate and no glutathione oxidation occurred. Dithioerythritol (2 mM) added at 10, 20, and 30 min after 3,5-dimethyl-NAPQI delayed but did not prevent cytotoxicity. Dithioerythritol also resulted in the partial restoration of GSH, presumably as a result of reduction of protein mixed disulphides. The mechanism of cytotoxicity of 3,5-dimethyl-NAPQI therefore appears to be a result of a combination of oxidative stress and deacetylation resulting in arylation.

Nitrofurantoin-mediated oxidative stress cytotoxicity in isolated rat hepatocytes.

Freshly isolated rat hepatocytes were used to study the mechanism(s) of toxicity of the antimicrobial drug nitrofurantoin. This 5-nitrofuran derivative stimulated hepatocyte oxygen uptake in the presence of the mitochondrial respiration inhibitors KCN or antimycin A. This could indicate the formation of O2- and H2O2, following intracellular nitrofurantoin reduction. Addition of nitrofurantoin to suspensions of isolated rat hepatocytes produced a dose- and time-dependent decrease of cell viability. H2O2 probably plays a significant role in the cytotoxic effects of nitrofurantoin as the catalase inhibitors azide or aminotriazole markedly enhanced cytotoxicity. The loss of cell viability was preceded by glutathione (GSH) depletion and a concomitant and nearly stoichiometric formation of oxidised glutathione (GSSG) that did not occur in hepatocytes lacking glutathione peroxidase activity isolated from rats fed a low-selenium diet. This indicates that H2O2 and the seleno-enzyme glutathione peroxidase are responsible for GSH oxidation. Furthermore, addition of nitrofurantoin to isolated rat hepatocytes produced a reversible inactivation of hepatocyte glutathione reductase activity and explains the maintenance of high GSSG levels. The compromised hepatocytes were also highly susceptible to H2O2. The hepatocyte toxicity of nitrofurantoin may, therefore, be attributed to oxidative stress caused by redox-cycling mediated oxygen activation.

The toxicity of disulphides to isolated hepatocytes and mitochondria.

The disulfide metabolites of thiono-sulfur drugs were found to be about 50 to 100 times more toxic to isolated rat hepatocytes than the corresponding parent drugs. The order of decreasing cytotoxicity for the disulfide metabolites was disulfiram greater than propylthiouracil disulfide greater than formamidine disulfide greater than phenylthiourea disulfide greater than thiobenzamide disulfide greater than cystamine. Depletion of intracellular GSH levels preceded cytotoxicity. GSH could be restored and cytotoxicity averted by adding the thiol reducing dithiothreitol. Depletion of GSH with diethylmaleate potentiated the toxicity of disulfides 3 to 4-fold confirming the protective role of GSH in disulfide toxicity. The toxicity of disulfiram was increased 4-fold in cells pretreated with ATP (0.8 mM) to effect a transient increase in cytosolic Ca2+ suggesting an impairment of Ca2+ homeostasis by the toxicant. Disulfiram (200 microM) rapidly depleted hepatocyte ATP levels within 15 minutes which suggests that ATP production is inhibited. The disulfide effectiveness at causing mitochondrial Ca2+ release was similar to their effectiveness at inducing hepatocyte cytotoxicity. These results suggest that hepatocyte toxicity is the result of oxidative inactivation of membrane protein thiols that regulate intracellular Ca2+ homeostasis.

Glutathione oxidation during peroxidase catalysed drug metabolism.

The peroxidase catalyzed oxidation of certain drugs in the presence of glutathione (GSH) resulted in extensive oxidation to oxidized glutathione (GSSG). Extensive oxygen uptake ensued and thiyl radicals could be trapped. Only catalytic amounts of drugs were required indicating a redox cycling mechanism. Active drugs included phenothiazines, aminopyrine, p-phenetidine, acetaminophen and 4-N,N-(CH3)2-aminophenol. Other drugs, including dopamine and alpha-methyl dopa, did not catalyse oxygen uptake, nor were GSSG or thiyl radicals formed. Instead, GSH was depleted by GSH conjugate formation. Drugs of the former group, e.g. acetaminophen, aminopyrine or N,N-(CH3)2-aniline have also been found by other investigators to form GSSG and hydrogen peroxide when added to hepatocytes or when perfused through an isolated liver. Although cytochrome P-450 normally catalyses a two-electron oxidation of drugs, serious consideration should be given for some one-electron oxidation resulting in radical formation, oxygen activation and GSSG formation.

Oxygen activation during drug metabolism.

The peroxidase-H2O2 catalyzed oxidation of certain drugs in the presence of GSH resulted in extensive oxidation to thiyl radicals and GSSG. NADH or arachidonate in place of GSH was also readily oxidized. Extensive oxygen uptake ensued resulting in the formation of superoxide radicals and H2O2. Only catalytic amounts of drugs and low peroxide levels were required, indicating a radox cycling mechanism. Active drugs included morphine, phenothiazines, aminopyrine, p-phenetidine, acetaminophen and 4-N,N-(CH3)2-aminophenol. Other drugs, including dopamine and methyl-alpha-dopa, did not catalyze oxygen uptake, nor was GSH oxidized to GSSG. Instead, GSH was depleted by GSH conjugate formation. Drugs of the former group, e.g. acetaminophen, aminopyrine or N,N-(CH3)2-aniline, have also been found by other investigators to form GSSG and hydrogen peroxide when added to hepatocytes or when perfused through an isolated liver. Although cytochrome P-450 normally catalyzes a two-electron oxidation of drugs, serious consideration should be given to some one-electron oxidation occurring as well and resulting in radical formation, oxygen activation and GSSG formation.

Glutathione conjugate formation without N-demethylation during the peroxidase catalysed N-oxidation of N,N',N,N'-tetramethylbenzidine.

The mechanism of peroxidative N-dealkylation of alkylamines proceeds via one-electron oxidation to the iminium cation which reacts with water to give the N-hydroxymethyl derivative which decomposes to formaldehyde and the N-demethylated product. This reaction is normally inhibited by glutathione by reduction of the cation radical with subsequent formation of oxidized glutathione (GSSG) with oxygen uptake. It was found that the horseradish peroxidase catalyzed N-demthylation of N,N,N',N'-tetramethylbenzidine (N4-TMB) in the presence of glutathione leads to the formation of water-soluble metabolites identified by high field nuclear magnetic resonance (NMR) and fast atom bombardment (FAB) mass spectrometry as 3,3'-(diglutathion-S-yl) and 2,2'-(diglutathion-S-yl)-N4-TMB. Smaller amounts of (monoglutathion-S-yl)-N4-TMB were also found. Only trace amounts of GSSG were formed and no oxygen uptake was observed. Electron spin resonance (ESR) spectrometry in the presence of 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) did not indicate the presence of a DMPO-glutathionyl adduct. These results indicate that glutathione inhibited the N-demethylation of N4-TMB under the described reaction conditions not by reduction of the cation radical but by conjugate formation. The mechanism of N-demethylation must involve removal of two successive electrons to give the benzoquinone-diimine which undergoes rearrangement to the iminium cation followed by reaction with water.

Peroxidase-catalyzed-3-(glutathion-S-yl)-p,p'-biphenol formation.

Oxidation of p,p'-biphenol with horseradish peroxidase (HRP)-hydrogen peroxide in the presence of bovine serum albumin or with bone marrow cell homogenate-hydrogen peroxide resulted in the formation of reactive products that conjugate with protein. Glutathione prevented the protein binding. Glutathione readily reacted with p,p'-biphenoquinone, the principal oxidation product of p,p'-biphenol in the HRP-hydrogen peroxide system and resulted in the formation of several glutathione conjugates, p,p'-biphenol and small amounts of oxidized glutathione. The major glutathione conjugate was identified as 3-(glutathion-S-yl)-p,p'-biphenol by high field nuclear magnetic resonance and fast atom bombardment mass spectrometry. The same conjugate was formed in the bone marrow homogenate-hydrogen peroxide system. p,p'-Biphenoquinone reduction by glutathione to p,p'-biphenol without glutathione oxidation was explained by the rapid reduction of p,p'-biphenoquinone by 3-(glutathion-S-yl)-p,p'-biphenol.

The metabolism of malondialdehyde.

Interest in malondialdehyde (MDA) metabolism stems from its formation as a product of lipid peroxidation in the diet and in the tissues; its reactivity with functional groups of nucleic acid bases, proteins and phospholipids; its mutagenicity in bacteria, and its reported skin and liver carcinogenicity in animals. Administration of the Na enol salt of MDA in the drinking water of mice over a range of 0.1-10.0 micrograms/g/day for 12 mo produced dose-dependent hyperplastic and neoplastic changes in liver nuclei and increased mortality at the highest level but produced no gross hepatic tumors. Addition of MDA to the medium of rat skin fibroblasts grown in culture caused nuclear abnormalities at concentrations as low as 10(-6) M despite an uptake of only 4%. [1,3-14C]MDA was rapidly oxidized to [14C]acetate in rat liver mitochondria and to 14CO2 in vivo; however, approximately 10% of the radioactivity was recovered in the urine. Chromatographic analysis of rat urine revealed the presence of several compounds which yield MDA on acid hydrolysis. Total MDA excretion increased in response to conditions which stimulate lipid peroxidation in vivo, including vitamin E deficiency, Fe or CCl4 administration, and enrichment of the tissues with PUFA. N-acetyl-e-(2-propenal)lysine was identified as a major urinary metabolite of MDA in rat and human urine. This compound is derived primarily from N-alpha-(2-propenal)lysine released in digestion as a product of reactions between MDA and the epsilon-amino groups of N-terminal lysine residues in food proteins. However, its presence in the urine of animals fasted or fed MDA-free diets indicates that it is also formed in vivo.(ABSTRACT TRUNCATED AT 250 WORDS)

Identification of N alpha-acetyl-epsilon-(2-propenal)lysine as a urinary metabolite of malondialdehyde.

Although orally administered malondialdehyde (MDA), a reactive hepatotoxic and mutagenic product of lipid peroxidation, is extensively metabolized to CO2, a portion is excreted in the urine in acid labile "bound" forms. Since much of the MDA in the diet is apparently bound to protein, the metabolism of protein-bound MDA was investigated. MDA was reacted with serum albumin and fed to rats. A urinary metabolite was detected which was shown to be identical to a metabolite of the lysine-MDA enaminal N epsilon-(2-propenal)lysine. After isolation by ion exchange and high performance liquid chromatography the metabolite was identified using high field nuclear magnetic resonance spectroscopy and fast atom bombardment-mass spectroscopy as N alpha-acetyl-epsilon-(2-propenal)lysine. This compound also was a major urinary metabolite of the Na enol salt of MDA administered by stomach intubation, and was excreted in increased amounts by rats fed a diet containing a highly peroxidizable oil (cod liver oil). It was also detected in the urine of fasted animals after injection with NaMDA, indicating that it is formed as a product of lipid peroxidation in vivo as well as of peroxidation of dietary lipids.

Urinary malondialdehyde as an indicator of lipid peroxidation in the diet and in the tissues.

Although malondialdehyde (MDA) is extensively metabolized to CO2, small amounts are nevertheless excreted in an acid-hydrolyzable form in rat urine. In this study, urinary MDA was evaluated as an indicator of lipid peroxidation in the diet and in the tissues. MDA was released from its bound form(s) in urine by acid treatment and determined as the TBA-MA derivative by HPLC. MDA excretion by the rat was found to be responsive to oral administration of the Na enol salt and to peroxidation of dietary lipids. Urinary MDA also increased in response to the increased lipid peroxidation in vivo produced by vitamin E deficiency and by administration of iron nitrilotriacetate. Chronic feeding of a diet containing cod liver oil led to increases in MDA excretion which were not completely eliminated by fasting or feeding a peroxide-free diet, indicating that there was increased lipid peroxidation in vivo. MDA excretion was not responsive to Se deficiency or CCl4 administration. DPPD, a biologically active antioxidant, but not BHA, a non-biologically active antioxidant, prevented the increase in MDA excretion in vitamin E deficient animals. The results indicate that MDA excretion can serve as an indicator of the extent of lipid peroxidation in the diet and, under conditions which preclude a dietary effect, as an index of lipid peroxidation in vivo.

High tolerance of broilers to vomitoxin from corn infected with Fusarium graminearum.

Corn purposely infected with Fusarium graminearum was found to contain 800 to 900 mg vomitoxin/kg. Contaminated corn was substituted for control corn at 0, 6, 12, 18, and 24% in a corn-soybean meal ration. Broiler cockerels were given each experimental diet from 6 to 11 days of age; then sample groups were necropsied. Remaining birds were subsequently offered commercial starter for 2 days and sample groups again necropsied. Growth and diet consumption were not significantly reduced until contaminated corn exceeded 12% of the ration (116 mg vomitoxin/kg). Alertness, coordination, and feathering appeared normal regardless of treatment. Birds that received contaminated corn exhibited plaques in the mouth and gizzard erosions proportional to the level of substitution. All lesions were generally restricted to the epithelial layer and no liver or kidney involvement could be demonstrated. A short return to uncontaminated feed eliminated most lesions. Fowl appear to be considerably more tolerant of vomitoxin than swine.

High pressure liquid chromatography of zearalenone and zearalenols in rat urine and liver.

A high pressure liquid chromatographic technique with internal standardization has been developed for determining zearalenone and metabolites in rat urine and liver. Following extraction with methylene chloride and solvent partition, samples are cleaned up by applying the extract to a Sephadex LH-20 column and eluting with a mixture of benzene-methanol (85 + 15). Compounds were resolved on 2 Part-isil-10 columns (25 cm x 4.6 mm id) in series with a mobile phase of isooctane-chloroform-methanol (35 + 25 + 3), and detected at 280 nm. The internal standard was 6'alpha-acetoxyzearalane. Limits of detection were about 2.0 ng for zearalenone and 5.0 ng for zearalenols (6'-hydroxyzearalane). Zearalenone and zearalenols were excreted mainly in free form with relatively little glucuronide conjugation. Metabolism of zearalenone to free zearalenol was minor compared with formation of bound forms.