A semi-automated method for measuring the potential for protein covalent binding in drug discovery.

INTRODUCTION: Covalent protein binding of metabolically reactive intermediates of drugs has been implicated in drug toxicity including the occurrence of idiosyncratic drug toxicity. Investigators therefore would prefer to avoid developing compounds that produce significant amounts of reactive metabolites. By incubating the radiolabeled drug of interest with liver microsomes it is possible to evaluate the propensity of a drug candidate to covalently bind to proteins. METHODS: Here we present a semi-automated method in which a Brandel cell harvester is used to collect and wash proteins that have been incubated with radiolabeled drug. This method utilizes glass fiber filter paper to capture precipitated protein, rather than the more traditional exhaustive extraction/centrifugation approach. Using model compounds (including [14C]diclofenac, [3H]imipramine, [14C]naphthalene, and [14C]L-746530) we compare the covalent binding results obtained using this method to results generated using the traditional method and we performed cross-laboratory testing of assay reproducibility. RESULTS: It was found that results from new method correlated highly with the traditional method (R2=0.89). The cross-laboratory testing of the method showed an average interlaboratory coefficient of variation of only 18.4%. DISCUSSION: This method provides comparable results to the more traditional centrifugation-based method with considerable time and labor savings.

Assessment of P450 induction in the marmoset monkey using targeted anti-peptide antibodies.

The identity and expression of hepatic P450 enzymes in marmosets was investigated using a panel of anti-peptide antibodies originally targeted against human P450 enzymes. In immunoblotting, of 12 antibodies examined, 10 bound specifically to bands in marmoset liver microsomal fraction corresponding to P450 enzymes. It is proposed that these represent marmoset CYP1A1, CYP1A2, CYP2A, CYP2B, CYP2C forms (CYP2C-1 and CYP2C-2), CYP2D19, CYP3A21 and another CYP3A form (CYP3A-m). The antibodies, together with an anti-marmoset CYP2E1 antibody, were used to investigate the expression of 10 P450 enzymes in marmosets treated with P450-inducing chemicals. Treatment with phenobarbitone caused CYP2B, CYP2C-2 and CYP3A21 levels to increase, rifampicin caused increases in CYP2B and CYP2C-1 and a decrease in CYP3A21 levels, whereas dioxin caused CYP1A1 and CYP1A2 levels to increase and CYP2E1 levels to decrease. Clofibric acid did not induce any P450. P450 enzyme activities were assessed using 8 different substrates and increases were found after treatment with phenobarbitone, rifampicin, and dioxin. However, due to species differences in substrate selectivity, it proved difficult to ascribe these changes to individual P450 enzymes. Thus, the use of anti-peptide antibodies provides a more informative way of assessing the levels of specific P450 enzymes than enzyme activity measurements.

Selective and potent inhibition of human CYP2C19 activity by a conformationally targeted antipeptide antibody.

A conformationally targeted anti-peptide antibody was produced by immunizing a rabbit with a cyclized peptide corresponding to a loop region of human CYP2C19 (residues 250-261). In an enzyme-linked immunosorbent assay, the antibody bound strongly to recombinant CYP2C19 and poorly to recombinant CYP2C8, CYP2C9, and CYP2C18. In immunoblotting studies, the antibody bound strongly to recombinant CYP2C19 and weakly to recombinant CYP2C8. No binding to recombinant CYP1A2, CYP2C9, CYP2C18, CYP2D6, CYP2E1, and CYP3A4 was detected. In immunoinhibition experiments, the anti-peptide antibody targeted against CYP2C19 potently inhibited (S)-mephenytoin 4'-hydroxylase activity of human hepatic microsomal fraction (>90%). It had no appreciable effect on ethoxyresorufin O-deethylase (CYP1A2), tolbutamide methyl-hydroxylase (CYP2C9), dextromethorphan O-demethylase (CYP2D6), 4-nitrophenol hydroxylase (CYP2E1), or testosterone 6beta-hydroxylase (CYP3A4) activity of human hepatic microsomal fraction. However, large amounts of purified IgG fractions were able to inhibit up to 35% of paclitaxel 6alpha-hydroxylase (CYP2C8) activity. In conclusion, we have demonstrated that an anti-peptide antibody targeted against residues 250 to 261 of human CYP2C19 selectively and potently inhibited CYP2C19 activity of human hepatic microsomal fraction.

Structure-function analysis of human CYP3A4 using a specific proinhibitory antipeptide antibody.

An anti-peptide antibody targeted against residues 253 to 269 of human CYP3A4 was produced that specifically and potently inhibited its activity in human hepatic microsomal fraction (>90%). The function of this region in P450 catalysis was investigated. Antibody binding to CYP3A4 was unable to affect the magnitude of the Type I spectrum on addition of testosterone. It also had no effect on the K(m) of the enzyme for testosterone, but it did cause a marked decrease in V(max) (>90%) of testosterone 6 beta-hydroxylation. There was no change in the ability of the antibody-bound CYP3A4 to form the steady-state level of the enzymatically or chemically reduced P450-CO complex or even the steady-state level of the dioxy-ferrous complex during testosterone metabolism, but the oxidation of NADPH by CYP3A4 in the presence of antibody was 60% that of CYP3A4 in the absence of antibody. The binding of the antibody also resulted in potent inhibition of cumene hydroperoxide-supported testosterone 6 beta-hydroxylase activity of human liver microsomal fraction (>90%). Our conclusion is that the loop region targeted in CYP3A4 is not involved in substrate binding, in reductase binding, in the transfer of the first or second electron from the reductase to CYP3A4, or in the binding of molecular oxygen. We speculate that antibody binding to CYP3A4 inhibits enzyme activity by destabilizing the ternary hydroperoxo complex, by interfering with the second proton transfer, and/or by interfering with the conformational changes that are suggested to be induced by substrate binding.

Affinity and potency of proinhibitory antipeptide antibodies against CYP2D6 is enhanced using cyclic peptides as immunogens.

A series of antipeptide antibodies directed against CYP2D6 were produced by immunizing rabbits with peptides that were sterically unrestrained (linear) or conformationally restricted by cyclization. A variety of sites within the region comprising residues 254 to 290 of CYP2D6 were targeted. In immunoblotting studies, each of the antibodies against the linear and cyclic peptides recognized only a single immunoreactive band of 54 kDa in human liver microsomal fraction and bound to recombinant CYP2D6, but not recombinant CYP1A1, CYP1A2, CYP2A6, CYP2B6, CYP2C9, CYP2C19, CYP2E1, or CYP3A4. However, the relative intensity of immunoreactive bands was considerably stronger for those antibodies raised against cyclic peptides. Similarly, in an enzyme-linked immunosorbent assay, antibodies raised against cyclic peptides bound 10 to 100 times more strongly to recombinant CYP2D6 than antibodies raised against the corresponding linear peptides. None of the antibodies raised against linear peptides had any effect on debrisoquine 4-hydroxylase activity of human hepatic microsomal fraction; however, anticyclic peptide antibodies targeted against residues 254 to 273, 261 to 272, and 257 to 268 of CYP2D6 inhibited enzyme activity by a maximum of 60, 75, and 91%, respectively. In contrast, despite binding strongly to CYP2D6, an anticyclic peptide antibody directed against residues 278 to 290 did not inhibit enzyme activity. The epitope of the proinhibitory anticyclic peptide antibody directed against residues 257 to 268 of CYP2D6 included Thr-261 and Trp-262, and indicates a role for these residues in enzyme inhibition. In conclusion, immunization with peptides conformationally restricted by cyclization to mimic loop regions of CYP2D6 resulted in strongly binding antibodies that when targeted appropriately were able to inhibit CYP2D6-catalyzed activity.

Polymorphic debrisoquine 4-hydroxylase activity in the rat is due to differences in CYP2D2 expression.

The female Dark Agouti rat is widely used as an animal model for the CYP2D6 poor metabolizer phenotype, males of other strains such as Sprague Dawley or Wistar serving as models for the extensive metabolizer phenotype. To determine the relative level of expression of CYP2D enzymes in the liver of female and male Dark Agouti, Sprague Dawley and Wistar rats, anti-peptide antibodies were raised in rabbits against short synthetic peptides representing the C-termini of the rat P450 enzymes CYP2D1, CYP2D2, CYP2D3, CYP2D4 and CYP2D5. In immunoblotting studies, it was found that the hepatic expression of CYP2D1 was greater in Dark Agouti rats than Sprague Dawley or Wistar rats. In contrast, hepatic CYP2D2 was 30-40-fold less abundant in female Dark Agouti than female Sprague Dawley or Wistar rats and six- to eightfold less abundant in male Dark Agouti than male Sprague Dawley or Wistar rats. No hepatic CYP2D3 could be detected in either sex of any of the three strains. Hepatic CYP2D4 expression was generally greater in male than female rats, and higher in Dark Agouti compared with Sprague Dawley or Wistar strains. CYP2D5 was expressed in the livers of female and male Dark Agouti rats but not in female Sprague Dawley or Wistar rats. This form was variably expressed in livers of male Sprague Dawley and Wistar rats. Hepatic debrisoquine 4-hydroxylase activity was markedly reduced in female and male Dark Agouti rats as compared to Sprague Dawley or Wistar rats and correlated (r = 0.88; P < 0.001) with the hepatic CYP2D2 content. Recombinant CYP2D2 was 18-fold more active at catalysing the 4-hydroxylation of debrisoquine than CYP2D1. Furthermore, quinine markedly inhibited CYP2D2-mediated debrisoquine and metoprolol oxidation, while quinidine, its diastereoisomer, inhibited the reactions to a lesser extent. In conclusion, these results show that impaired debrisoquine 4-hydroxylase activity in the female Dark Agouti rat is due to low levels of CYP2D2.

Expression and localisation of CYP2D enzymes in rat basal ganglia.

P450 enzymes in the CYP2D subfamily have been suggested to contribute to the susceptibility of individuals in developing Parkinson's disease. We have used specific anti-peptide antisera and peroxidase immunohistochemistry to investigate the expression of CYP2D enzymes in the rat brain and some possible factors that may affect their regulation. In male Wistar rats, CYP2D1 was not detected in the basal ganglia or in any other brain region. CYP2D2 was weakly expressed within neurones of the subthalamic nucleus, substantia nigra and interpeduncular nucleus as well as in the hippocampus, dentate gyrus, red nucleus and pontine nucleus. CYP2D3 and CYP2D4 were absent from the basal ganglia, although moderate amounts of CYP2D3 were detected within fibres of the oculomotor root, and very low levels of CYP2D4 were present in white matter tracts. In contrast, CYP2D5 was extensively expressed in the basal ganglia, including neurones in the subthalamic nucleus, substantia nigra and interpeduncular nucleus, as well as other areas of the brain, including the ventral tegmental area, piriform cortex, hippocampus, dentate gyrus, medial habenular nucleus, thalamic nucleus and pontine nucleus. Lesioning of the nigro-striatal tract to cause almost a complete loss of tyrosine hydroxylase containing neurones in the substantia nigra, also reduced the number of neurones expressing CYP2D5 by 50%, indicating that CYP2D5 is expressed in dopaminergic neurones. Castration of pre-pubertal or adult Wistar rats had no effect on the number of CYP2D5-positive neurones in the substantia nigra. Although Dark Agouti rats lack hepatic CYP2D2, expression in the midbrain was similar to that of Wistar rats; furthermore, there was no difference in expression or distribution between male and female rats. In contrast to naive rats, extensive expression of CYP2D4 was found throughout the basal ganglia and in other brain nuclei in Wistar rats treated with not only clozapine, but also saline, suggesting that CYP2D4 may be induced as a result of mild stress. The function of CYP2D enzymes in the brain remains unknown, but their selective localisation suggests a physiological role in neuronal activity and in adaptation to abnormal situations.