Charles A. Tyson

PREDICTION OF IN VIVO DRUG METABOLISM IN THE HUMAN LIVER FROM IN VITRO METABOLISM DATA

As a new approach to predicting in vivo drug metabolism in humans, scaling of in vivo metabolic clearance from in vitro data obtained using human liver microsomes or hepatocytes is described in this review, based on the large number of literature data. Successful predictions were obtained for verapamil, loxtidine (lavoltidine), diazepam, lidocaine, phenacetin and some other compounds where CLint,in vitro is comparable with CLint,in vivo. On the other hand, for some metabolic reactions, differences in CLint,in vitro and CLint,in vivo greater than 5-fold were observed. The following factors are considered to be the cause of the differences: (1) metabolism in tissues other than liver, (2) incorrect assumption of rapid equilibrium of drugs between blood and hepatocytes, (3) presence of active transport through the sinusoidal membrane, and (4) interindividual variability. Furthermore, the possibility of predicting in vivo drug metabolic clearance from results obtained using a recombinant system of human P450 isozyme was described for a model compound, YM796, where the predicted metabolic clearances obtained from the recombinant system, taking account of the content of the P450 isozyme CYP3A4 in the human microsomes, were comparable with the observed clearances using human liver microsomes containing different amounts of CYP3A4. Even in the case where the first-pass metabolism exhibits nonlinearity, it appears to be possible to predict in vivo metabolic clearance from in vitro metabolic data.

cDNA cloning and inducible expression of human multidrug resistance associated protein 3 (MRP3)1

Previously, we cloned rat MRP3 as a candidate for an inducible transporter for the biliary excretion of organic anions [Hirohashi et al. (1998) Mol. Pharmacol. 53, 1068–1075]. In the present study, we cloned human MRP3 (1527 amino acids) from Caco‐2 cells. Human MRP3 is predominantly expressed in liver, small intestine and colon; hepatic expression of MRP3 was observed in humans but not in normal rats. In HepG2 cells, the expression of MRP3 was induced by phenobarbital. These results suggest that MRP3 may act as an inducible transporter in the biliary and intestinal excretion of organic anions.

Prediction of in vivo interaction between triazolam and erythromycin based on in vitro studies using human liver microsomes and recombinant human CYP3A4.

AbstractPurpose. To quantitatively predict the in vivo interaction betweentriazolam and erythromycin, which involves mechanism-basedinhibition of CYP3A4, from in vitro studies using human liver microsomes(HLM) and recombinant human CYP3A4 (REC). Methods. HLM or REC was preincubated with erythromycin in thepresence of NADPH and then triazolam was added. α- and 4-hydroxy(OH) triazolam were quantified after a 3 min incubation and the kineticparameters for enzyme inactivation (kinact and K′app) were obtained.Drug-drug interaction in vivo was predicted based on aphysiologically-based pharmacokinetic (PBPK) model, using triazolam anderythromycin pharmacokinetic parameters obtained from the literature and kineticparameters for the enzyme inactivation obtained in the in vitro studies. Results. Whichever enzyme was used, triazolam metabolism was notinhibited without preincubation, even if the erythromycin concentrationwas increased. The degree of inhibition depended on preincubationtime and erythromycin concentration. The values obtained for kinactand K′app were 0.062 min−1 and 15.9 μM (α-OH, HLM), 0.055 min−1and 17.4 μM (4-OH, HLM), 0.173 min−1 and 19.1 μM (α-OH, REC),and 0.097 min−1 and 18.9 μM (4-OH, REC). Based on the kineticparameters obtained using HLM and REC, the AUCpo of triazolamwas predicted to increase 2.0- and 2.6-fold, respectively, followingoral administration of erythromycin (333 mg t.i.d. for 3 days), whichagreed well with the reported data. Conclusions. In vivo interaction between triazolam and erythromycinwas successfully predicted from in vitro data based on a PBPK modelinvolving a mechanism-based inhibition of CYP3A4.

Variation of Hepatic Methotrexate 7-Hydroxylase Activity in Animals and Humans

This study deals with individual and species variations in the converting activity of methotrexate (MTX) to 7‐hydroxymethotrexate in animals and humans. When MTX 7‐hydroxylase was assayed in six human liver cytosols, a 48‐fold range of intersubject variation of the activity was observed. The variations were correlated to the concentrations of aldehyde oxidase activity in human subjects assayed with benzaldehyde as a substrate. Species differences of liver MTX 7‐hydroxylase activity were also observed. The activity was highest in rabbits, followed by rats, hamsters, and monkeys but was undetectable in dogs. Strain differences of MTX 7‐hydroxylase activity based on aldehyde oxidase activity were also observed in rats and mice. The results suggest that aldehyde oxidase functions as MTX 7‐hydroxylase in livers of animals and humans, and the observed differences of MTX 7‐hydroxylase activity are due to variations in the amount of aldehyde oxidase present.

Response of isolated hepatocytes to organic and inorganic cytotoxins

Thirty-four chemicals-diverse in structure, postulated mechanisms of action, and primary target organs--were tested for cytotoxic response in isolated hepatocyte suspensions from young male Sprague-Dawley rats. Hepatocytes were incubated in the presence and absence of the test chemicals in closed vessels fitted with side arms for serial sampling for up to 5 h at 37 degrees C with gentle shaking under an O2:CO2 (95:5) atmosphere. The parameters evaluated were glutamate-oxaloacetate transaminase and lactate dehydrogenase release from the cells, Trypan blue exclusion, cell count, urea synthesis capability, and steady-state ATP levels. All chemicals cytotoxic in animals following single or short-term repeated exposures caused statistically significant changes in one or more of these parameters in the 0.01-10-mM concentration range. Dimethylnitrosamine and thioacetamide were not as potent in the isolated cell system as expected from their in vivo hepatotoxicity, and the quantitative changes produced with thioacetamide in the hepatocytes were marginal, even at 10 mM. The solvents tested--ethanol, acetone, dimethyl sulfoxide (DMSO), and propylene glycol--were without effect. These results indicate that isolated hepatocyte suspensions are useful for the identification of cytotoxins in general and hepatotoxins in particular, but that their capability for yielding a quantitative index of cytotoxic potential for diverse chemical species remains to be demonstrated.

Primary active transport of organic anions on bile canalicular membrane in humans

Biliary excretion of several anionic compounds was examined by assessing their ATP-dependent uptake in bile canalicular membrane vesicles (CMV) prepared from six human liver samples. 2, 4-Dinitrophenyl-S-glutathione (DNP-SG), leukotriene C4 (LTC4), sulfobromophthalein glutathione (BSP-SG), E3040 glucuronide (E-glu), beta-estradiol 17-(beta-D-glucuronide) (E2-17G), grepafloxacin glucuronide (GPFXG), pravastatin, BQ-123, and methotrexate, which are known to be substrates for the rat canalicular multispecific organic anion transporter, and taurocholic acid (TCA), a substrate for the bile acid transporter, were used as substrates. ATP-dependent and saturable uptake of TCA, DNP-SG, LTC4, E-glu, E2-17G, and GPFXG was observed in all human CMV preparations examined, suggesting that these compounds are excreted in the bile via a primary active transport system in humans. Primary active transport of the other substrates was also seen in some of CMV preparations but was negligible in the others. The ATP-dependent uptake of all the compounds exhibited a large inter-CMV variation, and there was a significant correlation between the uptake of glutathione conjugates (DNP-SG, LTC4, and BSP-SG) and glucuronides (E-glu, E2-17G, and GPFXG). However, there was no significant correlation between TCA and the other organic anions, implying that the transporters for TCA and for organic anions are different also in humans. When the average value for the ATP-dependent uptake by each preparation of human CMVs was compared with that of rat CMVs, the uptake of glutathione conjugates and nonconjugated anions (pravastatin, BQ-123, and methotrexate) in humans was approximately 3- to 76-fold lower than that in rats, whereas the uptake of glucuronides was similar in the two species. Thus there is a species difference in the primary active transport of organic anions across the bile canalicular membrane that is less marked for glucuronides.Biliary excretion of several anionic compounds was examined by assessing their ATP-dependent uptake in bile canalicular membrane vesicles (CMV) prepared from six human liver samples. 2,4-Dinitrophenyl- S-glutathione (DNP-SG), leukotriene C4(LTC4), sulfobromophthalein glutathione (BSP-SG), E3040 glucuronide (E-glu), β-estradiol 17-(β-d-glucuronide) (E2-17G), grepafloxacin glucuronide (GPFXG), pravastatin, BQ-123, and methotrexate, which are known to be substrates for the rat canalicular multispecific organic anion transporter, and taurocholic acid (TCA), a substrate for the bile acid transporter, were used as substrates. ATP-dependent and saturable uptake of TCA, DNP-SG, LTC4, E-glu, E2-17G, and GPFXG was observed in all human CMV preparations examined, suggesting that these compounds are excreted in the bile via a primary active transport system in humans. Primary active transport of the other substrates was also seen in some of CMV preparations but was negligible in the others. The ATP-dependent uptake of all the compounds exhibited a large inter-CMV variation, and there was a significant correlation between the uptake of glutathione conjugates (DNP-SG, LTC4, and BSP-SG) and glucuronides (E-glu, E2-17G, and GPFXG). However, there was no significant correlation between TCA and the other organic anions, implying that the transporters for TCA and for organic anions are different also in humans. When the average value for the ATP-dependent uptake by each preparation of human CMVs was compared with that of rat CMVs, the uptake of glutathione conjugates and nonconjugated anions (pravastatin, BQ-123, and methotrexate) in humans was ∼3- to 76-fold lower than that in rats, whereas the uptake of glucuronides was similar in the two species. Thus there is a species difference in the primary active transport of organic anions across the bile canalicular membrane that is less marked for glucuronides.

Acetaminophen hepatotoxicity: Correspondence of selective protein arylation in human and mouse liver in vitro, in culture, and in vivo

Human and mouse liver were exposed to an APAP-activating system, in vitro. Subsequent immunochemical analysis of electrophoretically separated proteins with an affinity-purified anti-APAP antibody indicated that when a cytosolic fraction from human liver was incubated with APAP, an NADPH-regenerating system, and mouse microsomes selective APAP binding occurred predominantly to proteins of approximately 38, 58, and 130 kDa. To evaluate whether similar proteins are targeted in situ, primary cultures of human hepatocytes were treated with 10 mM APAP for 4 hr prior to immunochemical analysis. APAP binding was again detected in protein bands of approximately 38, 58, and 130 kDa. In addition, selective binding was also noted to other cytosolic protein bands, e.g., approximately 52 and 62 kDa. For mouse liver, the majority of the binding, in vitro or in culture, was to proteins of approximately 44 and 58 kDa with lesser binding to proteins of approximately 33 and 130 kDa among others. By contrast, at the times monitored, little covalent binding was detected in the 44-kDa region in the human liver experiments. Most noteworthy was the finding that when the protein arylation patterns on liver samples from a human APAP fatality were compared to those from a mouse given a hepatotoxic dose of APAP, the binding patterns were similar to those detected after the in vitro and the culture experiments with mouse and human livers. Furthermore, an immunohistochemical analysis revealed that as with the mouse, APAP covalent binding in the human liver exhibited a distinct zonal pattern consistent with centrilobular binding. That APAP arylation of the 58- and 130-kDa proteins was observed in livers from both mice and humans suggests that the mouse provides a valid model for studying the mechanistic importance of covalent binding. Elucidation of the identities and functions of the common targeted proteins may clarify their toxicological significance.

Correlations of in vitro and in vivo hepatotoxicity for five haloalkanes

Five haloalkanes--CBrCl3, CCl4, CHCl3, and 1,1,1-and 1,1,2-trichloroethane (TCE)--were ranked for their relative hepatotoxicity in an in vitro system of isolated hepatocyte suspensions and in vivo by po administration of the test chemical to fasted rats of the same strain and sex as used for the hepatocytes. Cytotoxic parameters used for ranking in the in vitro system were GOT and LDH release, and the results were expressed in terms of EC50 values (the dissolved haloalkane concentration required to release 50% of the cell content of each enzyme after 2 hr of exposure) for rank determination. Cytotoxic parameters measured in vivo were SGOT and SGPT, and the ranking was based on ED50 values (the haloalkane dose that produced an above normal serum transaminase level in 50% of the test animals). With these parameters, the potency rankings in each system were the same except that of 1,1,1-TCE, which was more cytotoxic in the in vitro system than would have been expected from the animal experiments. Purification of the 1,1,1-TCE to remove stabilizers, use of phenobarbital-induced hepatocytes or hepatocytes from starved rats, and administration of the haloalkanes ip instead of po failed to improve the correlation. The discrepancy could be resolved, however, by factoring air: medium partition coefficient data into the EC50 values to take into account differences in the volatility and aqueous and lipid solubility of the chemicals, and hence their retention in vivo. These observations encourage the belief that isolated hepatocyte systems have value for ranking structurally related chemicals as to their cytotoxic potential, even though their mechanisms of action may differ.

Metabolism and cytotoxicity of acetaminophen in hepatocytes isolated from resistant and susceptible species

Acetaminophen (APAP) disposition was studied in vitro using hepatocytes isolated from rats, hamsters, rabbits, and dogs, species that vary markedly in susceptibility to the hepatotoxicity of this drug. Metabolism was assessed by concurrent measurements of glutathione depletion and protein adduct formation (activation pathway) and of total aqueous metabolite production (detoxication pathways). Cytotoxicity was monitored by cell count and by lactate dehydrogenase (LDH) release to culture medium. In agreement with whole animal studies, hepatocytes from hamsters were very susceptible to APAP-induced toxicity whereas rat and rabbit hepatocytes were resistant. In vivo data were unavailable for the dog, but dog hepatocytes were also relatively resistant. Parameters of APAP metabolism generally correlated with the species susceptibility ranking; however, no single parameter was an ideal index of the sensitivity observed. As predicted by the cytotoxicity data, hamster hepatocytes produced more covalent adducts of APAP, were depleted of GSH more rapidly, and detoxified APAP by formation of polar metabolites at a slower rate than rat hepatocytes. On the other hand, rabbit hepatocytes had no detectable covalent adducts, retained higher amounts of GSH, and metabolized more APAP to polar conjugates than the other species. Dog hepatocytes formed low amounts of both covalent adducts and conjugates. These studies indicate that interspecies comparisons using isolated hepatocytes to study xenobiotic metabolism and the resulting cytotoxicity are feasible, but for a clear understanding of observed differences, it is necessary to study the interrelationships between the toxication and detoxication pathways of metabolism.

Oxidative DNA damage induced by potassium bromate in isolated rat renal proximal tubules and renal nuclei

Oxidative damage caused by potassium bromate (KBrO3), a rat renal carcinogen, was investigated using in vitro preparations of rat renal proximal tubules (RPT) and renal nuclear fractions. Release of lactate dehydrogenase and decrease of SH-group content in RPT (1 mg protein/ml) by KBrO3 (0.5-5 mM) in a concentration- and time-dependent manner were observed. Peroxidized arachidonic acid and 8-hydroxydeoxyguanosine (8-OH-dG) levels in RPT were increased after administration of 2 and 5 mM KBrO3. 8-OH-dG formation was observed after incubation of renal nuclei with a lipid-peroxiding system, autooxidized methyl linolenate, or KBrO3. These findings provide support for involvement of lipid peroxidation in producing oxidized DNA damage by KBrO3 directly to RPT, the target site for renal carcinogenesis.