David J. Jollow

Mechanisms of fasting-induced potentiation of acetaminophen hepatotoxicity in the rat☆

The effects of an acute fast on acetaminophen metabolism and hepatotoxicity were investigated in male Long Evans Hooded rats. Histologic studies confirmed that fasting potentiated acetaminophen-induced hepatic necrosis. The previous known fasting-induced decrease in hepatic levels of glutathione and depletion of glycogen levels were also confirmed. Pharmacokinetic studies revealed that, at high dose levels of acetaminophen, fasting decreased the overall rate of elimination as evidence by a longer blood half-life of the drug. The decreased clearance was largely the result of decreases in the apparent rate constants for glucuronidation (ca. 40%) and for sulfation (ca. 30%). Fasting had no significant effects on the apparent rate constants for formation of either acetaminophen mercapturate or the methylthio derivatives. The depression of the nontoxic glucuronidation and sulfation pathways resulted in an increased proportion of the dose converted to the toxic metabolite and, hence, contributed to the potentiation of liver injury in fasted rats. In addition, these studies demonstrated that significant glucuronidation capacity (ca. 60% of that in fed rats) was maintained in fasted rats, indicating that: the glucuronidation capacity was not directly correlated with glycogen levels; and in fasted rats the glucose required for UDP-glucuronic acid formation for acetaminophen glucuronidation was supplied from sources other than glycogen.

Acetaminophen structure-toxicity studies: In vivo covalent binding of a nonhepatotoxic analog, 3-hydroxyacetanilide☆

High doses of 3-hydroxyacetanilide (3HAA), a structural isomer of acetaminophen, do not produce hepatocellular necrosis in normal male hamsters or in those sensitized to acetaminophen-induced liver damage by pretreatment with a combination of 3-methylcholanthrene, borneol, and diethyl maleate. Although 3HAA was not hepatotoxic, the administration of acetyl-labeled [3H or 14C]3HAA (400 mg/kg, ip) produced levels of covalently bound radiolabel that were similar to those observed after an equimolar, hepatotoxic dose of [G-3H]acetaminophen. The covalent nature of 3HAA binding was demonstrated by retention of the binding after repetitive organic solvent extraction following protease digestion. Hepatic and renal covalent binding after 3HAA was approximately linear with both dose and time. In addition, 3HAA produced only a modest depletion of hepatic glutathione, suggesting the lack of a glutathione threshold. 3-Methylcholanthrene pretreatment increased and pretreatment with cobalt chloride and piperonyl butoxide decreased the hepatic covalent binding of 3HAA, indicating the involvement of cytochrome P450 in the formation of the 3HAA reactive metabolite. The administration of multiple doses or a single dose of [ring-3H]3HAA to hamsters pretreated with a combination of 3-methylcholanthrene, borneol, and diethyl maleate produced hepatic levels of 3HAA covalent binding that were in excess of those observed after a single, hepatotoxic acetaminophen dose. These data suggest that the nature and/or the intracellular processing of the reactive metabolites of acetaminophen and 3HAA are different. These data also demonstrate that absolute levels of covalently bound xenobiotic metabolites cannot be utilized as absolute predictors of cytotoxic potential.

DAPSONE-INDUCED HEMOLYTIC ANEMIA*

Dapsone, an old drug introduced and used almost exclusively for the treatment of leprosy, is now utilized in an increasing number of therapeutic situations. However, its hemotoxicity is potentially severe and is often dose limiting. Effective countermeasures, based on resolution of the mechanisms underlying dapsone-induced hemotoxicity, could significantly enhance the therapeutic value of the drug. In studies on rat red cells, we have established that the N-hydroxy metabolites of dapsone, DDS-NOH and MADDS-NOH, are direct-acting hemolytic agents, that they are formed in amounts sufficient to account for the hemotoxicity of the parent drug, and that the action of these toxic metabolites in the red cell induces premature sequestration by the spleen. Incubation of rat red cells with hemolytic concentrations of arylhydroxylamines leads to the generation of hydroxyl, glutathiyl, and hemoglobinthiyl radicals, and the formation of protein-glutathione mixed disulfides. Disulfide-linked adducts are also formed between membrane skeletal proteins and hemoglobin monomers, as well as between the monomeric hemoglobin units forming dimers, trimers, tetramers, and pentamers. Profound morphological changes are seen with change from normal discoidocity to an extreme nonspherocytic enchinocyte shape. Parallel studies with human red cells indicate that the response of human cells is qualitatively similar but that there are notable differences in regard to skeletal membrane effects. A working hypothesis for the mechanism underlying dapsone hemolytic activity is proposed.

Dapsone-induced hemolytic anemia: effect of N-hydroxy dapsone on the sulfhydryl status and membrane proteins of rat erythrocytes.

Dapsone hydroxylamine (DDS-NOH), a known metabolite of dapsone, has recently been shown to be a direct-acting hemotoxin responsible in part for dapsone-induced hemolytic anemia in the rat. The effect of DDS-NOH on the morphology, sulfhydryl status, and membrane skeletal proteins of the rat red cell has been investigated. Exposure of rat red cells to a TC50 of DDS-NOH induced transformation of about 50% of the cells to an extreme echinocyte morphology. Reduced glutathione content of the cells was rapidly lost with concomitant increase in the formation of mixed disulfide between glutathione and the soluble protein of the cell. Oxidized glutathione content of the cells did not increase at any time during exposure to DDS-NOH. Examination of the skeletal membrane proteins by SDS-PAGE indicated that DDS-NOH caused the apparent loss of band 4.2, decrease in peaks 1, 2.1, and 3, and the appearance of new bands at about 16, 27, 40, and 54 kDa. Bands 4.1 and 7 appeared unchanged. Treatment of DDS-NOH altered proteins with dithiothreitol, reversed the protein changes, and indicated that the observed alterations were due to the formation of disulfide-linked adducts between hemoglobin and the various skeletal proteins as well as between hemoglobin monomers. The possible significance of the parallel changes in cell morphology and in membrane skeletal proteins for the premature splenic sequestration of the injured rat red cells is discussed.

Role of metabolites in propanil-induced hemolytic anemia☆

Hemolytic anemia and methemoglobinemia induced by exposure to certain arylamines, such as aniline and dapsone, are known to be mediated by their N-hydroxylamine metabolites. The arylamide propanil (3,4-dichloropropionanilide), a herbicide used extensively in rice fields, is also thought to induce methemoglobinemia through the action of metabolites. However, the hemolytic potential of this compound has not previously been reported. The present studies were undertaken to determine the hemolytic potential of propanil, and, if positive, the role of metabolites in this hemotoxicity. The survival of previously administered 51Cr-labeled erythrocytes in rats was reduced in a dose-dependent manner by ip administration of both propanil and its deacylated metabolite, 3,4-dichloroaniline (ED50 for both ca. 1.8 mmol/kg). When labeled erythrocytes were exposed in vitro to propanil or 3,4-dichloroaniline and then readministered to rats, no decrease in erythrocyte survival was observed, which indicated that these compounds were not direct-acting hemolytic agents. In contrast, erythrocyte survival was markedly reduced by ip administration or in vitro exposure to N-hydroxy-3,4-dichloroaniline. In addition, N-hydroxy-3,4-dichloroaniline was detected in the blood of propanil-treated rats in amounts sufficient to account for the hemolytic activity of the parent compound. These data indicate that N-hydroxy-3,4-dichloroaniline mediates propanil-induced hemolytic anemia, and that occupational exposure to propanil may result in an increased risk of hemolytic episodes.

Mechanism of decreased acetaminophen glucuronidation in the fasted rat

The mechanism by which an acute fast decreases the glucuronidation of hepatotoxic doses of acetaminophen in the rat was examined. Fasting did not depress the level of the enzyme, glucuronyl transferase, or the basal level of the co-substrate, UDP-glucuronic acid (UDPGA). Administration of a hepatotoxic dose of acetaminophen rapidly depleted UDPGA levels in both fed and fasted rats to the same nadir. Fed and fasted rats differed in that the rate of repletion of UDPGA levels was markedly slower in fasted rats. The total hepatic levels of UDP-glucose dehydrogenase and its cofactor, NAD+, were not decreased by fasting. In fasted rats, hepatic levels of the UDPGA precursor, UDP-glucose, were approximately 60% those of fed rats both before and after a hepatotoxic dose of acetaminophen. In fed rats, acetaminophen induced a marked depletion of hepatic glycogen levels and a dramatic increase in blood glucose levels. Acetaminophen induced a similar marked increase in blood glucose levels in fasted rats in spite of the fact that they lacked hepatic glycogen. It is concluded that the fasting-induced decrease in the glucuronidation of hepatotoxic doses of acetaminophen results from decreased production of UDPGA. The decreased synthetic capacity for UDPGA does not appear to be due to the inability of the liver to produce glucose units per se, but rather to the fasting-induced altered activities of the enzymes of carbohydrate metabolism which, in turn, alter the fate of glucose-6-phosphate derived from gluconeogenesis.

Strain differences in susceptibility of normal and diabetic rats to acetaminophen hepatotoxicity

The effects of streptozotocin (STZ)-induced diabetes on acetaminophen metabolism and hepatotoxicity in male Sprague-Dawley (SD) and Long Evans Hooded (LEH) rats were compared. In agreement with earlier studies, normal SD rats were more resistant to acetaminophen-induced hepatic necrosis than normal LEH rats. In contrast to LEH rats, the diabetic state did not protect SD rats from liver injury. Pharmacokinetic studies revealed that normal SD rats eliminated acetaminophen faster than normal LEH rats, and that the diabetic state further enhanced elimination in both strains of rats; however, the effect was much greater in LEH rats. Normal SD rats had a greater capacity to metabolize acetaminophen to nontoxic glucuronide and sulfate conjugates than normal LEH rats. In LEH rats, the diabetic state enhanced acetaminophen glucuronidation and sulfation, whereas in SD rats the diabetic state increased only sulfation; glucuronidation was unaffected. Additional studies revealed that the difference in the glucuronidation capacities between normal LEH and normal SD rats was not due to differences in either the amount of the enzyme, glucuronyl transferase, or basal hepatic levels of the cofactor, UDPGA. Similarly, the diabetes-induced enhancement of glucuronidation in LEH rats was not due to differences in predrug levels of either glucuronyl transferase or UDPGA. Thus, the major difference in susceptibility of the two strains of normal rats to acetaminophen hepatotoxicity appears to be due to the capacity to clear the drug through nontoxic pathways. The greater glucuronidation capacity seen in diabetic LEH rats and in normal and diabetic SD rats as compared to normal LEH rats, appears to be due to a greater ability to produce UDPGA in response to the metabolic demand.

Galactosamine hepatotoxicity: Effect of galactosamine on glutathione resynthesis in rat primary hepatocyte cultures

The effect of galactosamine on the resynthesis of glutathione in rat primary hepatocyte cultures was investigated. Cultured rat hepatocytes were treated with galactosamine (4 mM) 1.5 hr prior to concurrent with, or 1.5 hr after cell attachment; total cellular glutathione was then measured over time. Addition of galactosamine at any of these times suppressed methionine-enhanced glutathione resynthesis in the cultures after a lag period of about 120 min. The lag period was not due to slow uptake of galactosamine by the cultured cells, since cellular UTP levels fell to less than 10% of controls within 60 min, a time frame comparable to that observed in vivo. Neither was the lag period a result of interference with cellular uptake of methionine or with conversion of methionine to cysteine, since the phenomenon was observed regardless of whether methionine or cysteine was used to promote glutathione resynthesis. Addition of uridine, which protects against galactosamine hepatotoxicity in vivo by replenishing hepatic UTP levels, did not prevent the suppression of glutathione resynthesis. The data indicate that (a) galactosamine inhibits the time-dependent resynthesis of glutathione in primary hepatocyte cultures, (b) a lag period exists for this response, and (c) this effect is not directly related to depletion of cellular UTP stores.

Trichloroethylene risk assessment: A review and commentary

Trichloroethylene (TCE) is a widespread environmental contaminant that is carcinogenic when given in high, chronic doses to certain strains of mice and rats. The capacity of TCE to cause cancer in humans is less clear. The current maximum contaminant level (MCL) of 5 ppb (μg/L) is based on an US Environment Protection Agency (USEPA) policy decision rather than the underlying science. In view of major advances in understanding the etiology and mechanisms of chemically induced cancer, USEPA began in the late 1990s to revise its guidelines for cancer risk assessment. TCE was chosen as the pilot chemical. The final guidelines emphasized a “weight-of-evidence” approach with consideration of dose-response relationships, modes of action, and metabolic/toxicokinetic processes. Where adequate data are available to support reversible binding of the carcinogenic moiety to biological receptors as the initiating event (i.e., a threshold exists), a nonlinear approach is to be used. Otherwise, the default assumption of a linear (i.e., nonthreshold) dose-response is utilized. When validated physiologically based pharmacokinetic (PBPK) models are available, they are to be used to predict internal dosimetry as the basis for species and dose extrapolations. The present article reviews pertinent literature and discusses areas where research may resolve some outstanding issues and facilitate the reassessment process. Key research needs are proposed, including role of dichloroacetic acid (DCA) in TCE-induced liver tumorigenesis in humans; extension of current PBPK models to predict target organ deposition of trichloroacetic acid (TCA) and DCA in humans ingesting TCE in drinking water; use of human hepatocytes to ascertain metabolic rate constants for use in PBPK models that incorporate variability in metabolism of TCE by potentially sensitive subpopulations; measurement of the efficiency of first-pass elimination of trace levels of TCE in drinking water; and assessment of exogenous factors’ (e.g., alcohol, drugs) ability to alter metabolic activation and risks at such low-level exposure.

The role of N-hydroxyphenetidine in phenacetin-induced hemolytic anemia

Phenacetin is well known to cause hemolytic anemia and methemoglobinemia in humans. Early mechanistic studies clearly established a causal role for active/reactive drug metabolites in the process but did not unequivocally identify these metabolite(s) or resolve the question of whether these two hemotoxicities are mechanistically linked. As part of ongoing studies on the mechanism underlying arylamine-induced hemotoxicities, we have recently shown that the arylhydroxylamine metabolites of aniline and dapsone mediate the hemolytic activity of aniline and dapsone, respectively. The present study was undertaken to determine if N-hydroxyphenetidine (PNOH), the known arylhydroxylamine metabolite of phenacetin, is responsible for phenacetin-induced hemolytic anemia. As measured by decreased survival of 51Cr-labeled erythrocytes in rats, phenacetin, p-phenetidine, and PNOH were all hemolytic in vivo, with PNOH being significantly the most potent of the three. In vitro exposure of 51Cr-tagged erythrocytes to PNOH, followed by transfusion into isologous rats, resulted in a concentration-dependent reduction in erythrocyte survival, indicating that PNOH is a direct-acting hemolytic agent. Phenacetin and p-phenetidine were inactive. Phenacetin, p-phenetidine, and PNOH all produced dose-dependent methemoglobinemia in rats. In parallel in vitro studies, PNOH elevated methemoglobin levels, p-phenetidine and phenacetin did not. However, attempts to identify PNOH in the blood of phenacetin- and p-phenetidine-treated rats were unsuccessful, despite the use of a highly sensitive analytical method. Hemotoxic concentrations of PNOH were found to be highly unstable in the presence of red cells, though relatively stable in the buffer vehicle alone. Inhibitors of acetylation (p-aminobenzoic acid [PABA]) and deacetylation (bis-[p-nitrophenyl]phosphate [BNPP]), used to alter the cyclic interconversion of phenacetin and p-phenetidine, caused changes in phenacetin hemotoxicity that indicated the hemotoxin was a deacetylated metabolite distal to p-phenetidine. These data are consistent with the hypothesis that PNOH, formed during the metabolic clearance of phenacetin, mediates phenacetin-induced hemolytic anemia and methemoglobinemia through direct toxic actions in the erythrocyte.