David C. Thompson

Biological and toxicological consequences of quinone methide formation

Quinone methides are a class of reactive, electrophilic compounds which are capable of alkylating cellular macromolecules. They are formed during xenobiotic biotransformation reactions and are hypothesized to mediate the toxicity of a large number of quinone antitumor drugs as well as several alkylphenols. In addition, oxidation of specific endogenous alkylphenols (e.g. coniferyl alcohol) and alkylcatechols (e.g. N-acetyldopamine, dopa) to quinone methides plays an important role in the synthesis of several complex plant and animal polymers, including lignin, cuticle and melanin. The role of quinone methides in these various processes is reviewed.

Peptide-Based In Vitro Assay for the Detection of Reactive Metabolites

We describe a novel peptide-based in vitro method for the detection of reactive metabolites that is amenable for use with microsomal or purified enzyme systems. Covalently bound adducts are detected by mass spectrometry using a surface-enhanced laser desorption ionizationtime of flight detector. The trapping molecule is an 11 amino acid peptide (ECGHDRKAHYK) that contains cysteine and other nucleophilic amino acid residues, as well as charged residues to enhance binding to a weak cation exchange chip surface used with the detection system. The assay concept was initially tested using rat or human liver microsomes with a series of benzodioxolanes. The assay was refined using human recombinant cytochrome P450 3A4 as the bioactivation system and validated with a series of positive and negative reference compounds. Alternative individual human recombinant P450 enzymes (e.g., 1A1, 2C9, or 2D6) may be used in place of 3A4 as the bioactivation system, or several P450 enzymes can be combined together into a single bioactivation system. We found that a mixture of P450s 3A4, 2C9, and 2D6 was suitable as a rapid general screen for the detection of reactive metabolites that covalently bind to proteins. Combining results from assays of individual P450 enzymes with microsomal systems allows the rapid profiling of metabolic pathways involved in reactive metabolite generation and provides valuable information that can be used to guide structural modifications to minimize the potential for metabolic bioactivation. In addition, non-P450 enzymes may be used as activation systems, such as peroxidases or alcohol dehydrogenase. In summary, this peptide-based assay system is able to detect reactive metabolites generated from a structurally diverse set of drugs and xenobiotics using a variety of microsomal or purified enzyme activation systems.

Studies on the mechanism of hepatotoxicity of 4-methylphenol (p-cresol): effects of deuterium labeling and ring substitution

We recently observed that 4-methylphenol (p-cresol) is toxic to rat liver tissue slices. A possible mechanism involves biotransformation of 4-methylphenol to a reactive quinone methide intermediate which covalently binds to cellular macromolecules and elicits cytotoxicity. In order to obtain further evidence for this proposed mechanism, we studied the effects of deuterium-labeled 4-methylphenol (4-[alpha, alpha, alpha-d3]-methylphenol), and the presence of various ring substituents, on the metabolism and toxicity of 4-methylphenol in precision cut liver slices prepared from male Sprague-Dawley rats. Deuterium-labeled 4-methylphenol was significantly less toxic than the parent compound in rat liver slices (LC50 = 3.36 vs. 1.31 mM, respectively). In addition, the deuterium-labeled compound was metabolized to a reactive intermediate (measured as glutathione conjugate formation) at a slower rate than that of 4-methylphenol in both liver slices and liver microsomal incubations. The presence of electron withdrawing substituents (2-chloro or 2-bromo) markedly enhanced both metabolism and toxicity, with the exception of 2,6-dibromocresol, which was similar to cresol in terms of rate of metabolism and toxicity. Conversely, the presence of electron donating substituents (2-methoxy, 2-methyl or 2,6-dimethyl) diminished metabolism and toxicity. In addition, methylation of the hydroxyl group to form 4-methylanisole, greatly reduced toxicity. These results support the hypothesis that the toxicity of 4-methylphenol is dependent on the formation of a reactive quinone methide intermediate.

Spontaneous hydrolysis of 4-trifluoromethylphenol to a quinone methide and subsequent protein alkylation

4-Trifluoromethylphenol (4-TFMP) was cytotoxic to precision-cut rat liver slices as indicated by loss of intracellular potassium. Intracellular glutathione levels decreased and fluoride ion levels increased in a time and concentration-dependent manner. The cytotoxicity of 4-TFMP did not appear to be due to the release of fluoride, however, since equimolar concentrations of sodium fluoride or potassium fluoride were not toxic. The ortho isomer (2-TFMP), which had a threefold slower rate of fluoride release, was much less toxic to liver slices. In incubations without slices, 4-TFMP spontaneously hydrolyzed in aqueous buffer at physiological pH to form 4-hydroxybenzoic acid via a quinone methide intermediate. The quinone methide was trapped by the addition of glutathione. Analysis of the glutathione adduct indicated that all of the fluorine atoms were lost during the hydrolysis, yielding a cresol derivative with the glutathione moiety attached to a benzylic carbonyl group. The glutathione conjugate was the primary product formed at low alkylphenol/glutathione ratios; however, at higher 4-TFMP concentrations additional unidentified products were observed. 4-TFMP also inhibited the in vitro enzyme activity of purlfied glyceraldehyde-3-phosphate dehydrogenase, a sulfhydryl-dependent enzyme, in a time and concentration-dependentmanner. Loss of thiol residues closely paralleled the loss in enzyme activity. The coaddition of glutathione prevented 4-TFMP-induced loss of enzyme activity. The cytotoxicity of 4-TFMP therefore appears to be due to spontaneous quinone methide formation and subsequent alkylation of cellular macromolecules.

Inhibition of NAD(H)/NADP(H) — requiring enzymes by aurintricarboxylic acid

Aurintricarboxylic acid (ATA), an inhibitor of Ca(2+)-dependent endonuclease activity, is often used to implicate a role for increased intracellular calcium in mechanistic toxicology studies. We report here on the ability of ATA to inhibit the activity of several NAD(H)/NADP(H)-requiring enzymes (purified or cellular homogenates), including lactic dehydrogenase, alcohol dehydrogenase, cytochrome c reductase, ethoxycoumarin o-dealkylase, isocitric dehydrogenase, glutathione reductase and glucose-6-phosphate dehydrogenase. These results were compared with the ability of ATA to inhibit micrococcal nuclease and rat liver Ca(2+)-dependent endonuclease activity in similar incubations. With the exception of alcohol dehydrogenase, ATA was a potent inhibitor of each of the purified enzymes, with IC50s ranging from 0.5 to 82 microM. In cell homogenates, however, ATA was from 10 to 100-fold less potent at inhibiting these enzymes. When exogenous protein was added to purified enzyme incubations, the effect of ATA was similarly diminished. Our results demonstrate that ATA inhibits a wide range of NAD(H)/NADP(H)-requiring enzymes in in vitro incubations using purified enzymes, but that the inhibitory effects are markedly reduced in incubations which more closely resemble a cellular milieu.

Assessment of the Metabolism and Intrinsic Reactivity of a Novel Catechol Metabolite

PH-302 ( 1) demonstrates potent inhibitory activity against the inducible form of nitric oxide synthase (iNOS). The primary metabolite of PH-302 is a catechol ( 2) resulting from oxidative demethylenation of the methylenedioxyphenyl moiety by cytochrome P450 3A4. Concerns regarding subsequent two-electron oxidation of 2 to an electrophilic quinone species and the potential for resulting toxicity prompted additional studies to examine the reactivity and metabolic fate of this metabolite. Contrary to literature reports of catechol reactivity, 2 appeared to be resistant to quinone formation in human liver microsomal incubations, as determined by the lack of detectable glutathione (GSH) adducts and no covalent binding to microsomal proteins. In addition, 2 showed no evidence of depletion of intracellular glutathione or cytotoxicity at concentrations up to 1 mM in primary human and rat hepatocytes. In the presence of tyrosinase, spectral evidence indicated that 2 was oxidized to the ortho-quinone, and upon incubation in the presence of GSH, two conjugates were detected and characterized by LC/MS/MS. Lastly, the metabolic pathways of 2 were investigated in rat and human hepatocytes and found to be similar, proceeding via glucuronidation, sulfation, and methylation of the catechol. Collectively, these studies demonstrate that 2 appears to be resistant to further oxidation to quinone in liver microsomes, as well as spontaneous redox cycling, while the formation of phase II metabolites in hepatocytes suggests that multiple detoxication pathways may provide added protection against toxicity in the liver.

Development and Application of Human Renal Proximal Tubule Epithelial Cells for Assessment of Compound Toxicity

Kidney toxicity is a major problem both in drug development and clinical settings. It is difficult to predict nephrotoxicity in part because of the lack of appropriate in vitro cell models, limited endpoints, and the observation that the activity of membrane transporters which plays important roles in nephrotoxicity by affecting the pharmacokinetic profile of drugs is often not taken into account. We developed a new cell model using pseudo-immortalized human primary renal proximal tubule epithelial cells. This cell line (SA7K) was characterized by the presence of proximal tubule cell markers as well as several functional properties, including transporter activity and response to a few well-characterized nephrotoxicants. We subsequently evaluated a group of potential nephrotoxic compounds in SA7K cells and compared them to a commonly used human immortalized kidney cell line (HK-2). Cells were treated with test compounds and three endpoints were analyzed, including cell viability, apoptosis and mitochondrial membrane potential. The results showed that most of the known nephrotoxic compounds could be detected in one or more of these endpoints. There were sensitivity differences in response to several of the chemicals between HK-2 and SA7K cells, which may relate to differences in expressions of key transporters or other components of nephrotoxicity pathways. Our data suggest that SA7K cells appear as promising for the early detection of renal toxicants.