Utilizing drug-drug interaction prediction tools during drug development: enhanced decision making based on clinical risk.

Several reports in the literature present the utility and value of in vitro drug-metabolizing enzyme inhibition data to predict in vivo drug-drug interactions in humans. A retrospective analysis has been conducted for 26 GlaxoSmithKline (GSK) drugs and drug candidates for which in vitro inhibition parameters have been determined, and clinical drug interaction information, from a total of 46 studies, is available. The dataset, for drugs with a diverse range of physiochemical properties, included both reversible and potentially irreversible cytochrome P450 inhibitors for which in vitro inhibition parameters (IC(50) or K(I)/k(inact) as appropriate) were determined using standardized methodologies. Mechanistic static models that differentiated reversible and metabolism-dependent inhibition, and also considered the contribution of intestinal metabolism for CYP3A4 substrates, were applied to estimate the magnitude of the interactions. Several pharmacokinetic parameters, including total C(max), unbound C(max), as well as estimates of hepatic inlet and liver concentration, were used as surrogates for the inhibitor concentration at the enzyme active site. The results suggest that estimated unbound liver concentration or unbound hepatic inlet concentration, with consideration of intestinal contribution, offered the most accurate predictions of drug-drug interactions (occurrence and magnitude) for the drugs in this dataset. When used with epidemiological information on comedication profiles for a given therapeutic area, these analyses offer a quantitative risk assessment strategy to inform the necessity of excluding specific comedications in early clinical studies and the ultimate requirement for clinical drug-drug interaction studies. This strategy has significantly reduced the number of clinical drug interaction studies performed at GSK.

Casopitant: in vitro data and SimCyp simulation to predict in vivo metabolic interactions involving cytochrome P450 3A4.

Casopitant [1-piperidinecarboxamide,4-(4-acetyl-1-piperazinyl)-N-((1R)-1-(3,5-bis(trifluoromethyl)phenyl)-ethyl)-2-(4-fluoro-2-methylphenyl)-N-methyl-(2R,4S), GW679769] has previously been shown to be a potent and selective antagonist of the human neurokinin-1 receptor, the primary receptor of substance P, both in vitro and in vivo, with good brain penetration properties. On the basis of this mode of action it was evaluated for the prevention of chemotherapy-induced and postoperative nausea and vomiting, and for the chronic treatment of anxiety, depression, insomnia, and overactive bladder. Casopitant is shown to be a substrate, an inhibitor, and an inducer of CYP3A4, and, because of this complex behavior, it was difficult to identify the primary mechanism by which it may give rise to drug-drug interactions (DDIs) of clinical relevance. Moreover, the major circulating metabolite is itself an inhibitor of CYP3A4 in vitro. On the basis of the different clinical indications and the various potential comedications of casopitant, a relevant part of the clinical development plan was focused on the assessment of the importance of clinical DDIs. The present study provides an overview of the DDI potential profile of casopitant, based on in vitro data and clinical evidence of its interaction with CYP3A4 probe substrates midazolam and nifedipine, the strong inhibitor ketoconazole, and the inducer rifampin. Overall, the clinical data confirm the ability of casopitant to interact with CYP3A4 substrates, inhibitors, or inducers. The in vitro data are accurate and robust enough to build a reliable SimCyp population-based model to estimate the potential DDIs of casopitant and to minimize the clinical studies recommended.

Bilirubin degradation by uncoupled cytochrome P450. Comparison with a chemical oxidation system and characterization of the products by high-performance liquid chromatography/electrospray ionization mass spectrometry.

Bilirubin is a protective antioxidant; however, when its conjugation and excretion are impaired, as in neonatal and hereditary jaundice, bilirubin accumulates and may cause severe neurotoxicity. Degradation of bilirubin takes place (a) on interaction with oxidative free radicals and (b) when cytochrome P450 (CYP) enzymes are uncoupled by polyhalogenated substrate analogues. The products of pathways (a) and (b) above have now been characterized by high-performance liquid chromatography/electrospray ionization mass spectrometry (HPLC/ESI-MS) and the mechanisms of fragmentation in part clarified. Oxidation of bilirubin by uncoupled CYP1A5 and by a Fe-EDTA/H2O2 system produced both biliverdin and an identical profile of dipyrrolic fragments, as detected by positive ESI-MS. A similar profile of oxidation products was found from mesobilirubin, all showing the expected increase in mass, thus providing direct evidence for fragmentation at the central methene bridge of the tetrapyrroles. Two degradation products, also detected by negative ESI-MS, were characterized as dipyrroles retaining the central bridge carbon, with one or two oxygen atom(s) bound (probably as the aldehyde and hydroperoxide derivatives). Ions compatible with propentdyopents and bilifuscins were also detected, but here the assignment was less certain. It is concluded that the first step in the oxidation of bilirubin may be hydrogen abstraction at the central methene bridge. This is followed either by loss of another hydrogen to give biliverdin, or by oxygen binding and fragmentation. Fe-EDTA/H2O2 and uncoupled CYP(Fe=O) may both initiate the reaction, the latter in an attempt to reduce the ferryl oxygen to water. These studies shed light on the CYP uncoupling mechanism and are of potential significance for the therapy of severe jaundice.

Up-regulation of advanced glycated products receptors in the brain of diabetic rats is prevented by antioxidant treatment.

Diabetics have at least twice the risk of stroke and may show performance deficit in a wide range of cognitive domains. The mechanisms underlying this gradually developing end-organ damage may involve both vascular changes and direct damage to neuronal cells as a result of overproduction of superoxide by the respiratory chain and consequent oxidative stress. The study aimed to assess the role of oxidative stress on the aldose reductase-polyol pathway, on advanced glycated end-product (AGE)/AGE-receptor interaction, and on downstream signaling in the hippocampus of streptozotocin-treated rats. Data show that, in diabetic rats, levels of prooxidant compounds increase, whereas levels of antioxidant compounds fall. Receptor for AGE and galectin-3 content and polyol flux increase, whereas glyceraldehyde-3-phosphate dehydrogenase activity is impaired. Moreover, nuclear factor kappaB (p65) transcription factor levels and S-100 protein are increased in the hippocampus cytosol, suggesting that oxidative stress triggers the cascade of events that finally leads to neuronal damage. Dehydroepiandrosterone, the most abundant hormonal steroid in the blood, has been reported to possess antioxidant properties. When dehydroepiandrosterone was administered to diabetic rats, the improved oxidative imbalance and the marked reduction of AGE receptors paralleled the reduced activation of nuclear factor kappaB and the reduction of S-100 levels, reinforcing the suggestion that oxidative stress plays a role in diabetes-related neuronal damage.

Ghrelin and des-acyl ghrelin both inhibit isoproterenol-induced lipolysis in rat adipocytes via a non-type 1a growth hormone secretagogue receptor.

Besides possessing a strong growth hormone (GH)-releasing activity, the gastrointestinal octanoylated peptide ghrelin has been reported to antagonize lipolysis in rat adipocytes. It is not yet clear whether this inhibitory activity on lipolysis is also shared by the major circulating isoform, des-acyl ghrelin, that does not activate the ghrelin receptor, namely the type 1a GH secretagogue-receptor (GHS-R1a) and lacks the endocrine effects of the acylated form. Here we show that des-acyl ghrelin, like ghrelin and some synthetic GHS (hexarelin and MK0677) and carboxy-terminally ghrelin fragments such as ghrelin-(1-5) and ghrelin-(1-10), all significantly reduced, over concentrations ranging from 1 to 1000 nM, the stimulation of glycerol release caused in rat epididymal adipocytes by the nonselective beta-adrenoceptor agonist isoproterenol in vitro. The order of potency on stimulated-lipolysis was: des-acyl ghrelin=ghrelin>MK0677=hexarelin>ghrelin-(1-5)=ghrelin-(1-10). This ranking was consistent with the binding experiments performed on membranes of epididymal adipose tissue or isolated adipocytes that did not express mRNA for GHS-R1a. A common high-affinity binding site was recognized in these cells by both acylated and des-acylated ghrelin and also by hexarelin, MK0677, ghrelin-(1-5) and ghrelin-(1-10). In conclusion, these findings provide the first evidence that des-acyl ghrelin, as well as ghrelin, short ghrelin fragments and synthetic GHS, may act directly as antilipolytic factors on the adipose tissue through binding to a specific receptor which is distinct from GHS-R1a.

Chromogenic derivatives of AHTN (6-acetyl-1,1,2,4,4,7-hexamethyltetralin) react with amino acids and protein in vitro. Spectral characteristics of the colour products.

Three chromogenic products of AHTN have been detected after photo-oxidation and shown to react with glycine or albumin, giving rise first to a blue color, followed by pink and, finally, a dark green colour. The absorption spectrum associated with the late green colour showed an increased absorbance in the long wavelength region of the spectrum, similar to the green colour extracted from the liver of AHTN-treated animals. The colour produced by the chromogenic derivatives of AHTN and by o-diacetylbenzene (a model chromogenic compound) were tightly bound to the protein pellet and resistant to acidic pH, unlike the colour from the ninhydrin reaction. The dark green colour produced in the liver by feeding AHTN to rats was also tightly bound to proteins and stable to acidic pH. These results suggest that both the photo-oxidized chromogenic derivatives of AHTN and o-diacetylbenzene, produce coloured derivatives with amino acids and proteins by a mechanism unrelated to ninhydrin. They also suggest that a chromogenic derivative of AHTN, produced in vivo during prolonged treatment with high doses of AHTN, may be responsible for the green colour of the livers and other tissues from treated animals. The data suggest that the AHTN-derived chromogenic material is metabolite-related rather than representative of a toxic process, although further work is necessary to confirm this hypothesis.

Interaction of polyhalogenated compounds of appropriate configuration with mammalian or bacterial CYP enzymes. Increased bilirubin and uroporphyrinogen oxidation in vitro.

Polyhalogenated compounds, such as 2,3,7,8-tetrachlorodibenzo-p-dioxin, are associated with toxic Uroporphyria and cause alleviation of jaundice in the Gunn rat. These effects have been attributed to a microsomal oxidation of uroporphyrinogen and bilirubin for which supportive evidence has been obtained in vitro. CYP1A1 required planar polyhalogenated biphenyls for these oxidative reactions, while CYP1A2 was capable of oxidation in their absence. We have now used rat CYP1A1 and confirmed with the pure enzyme that increased bilirubin oxidation was caused by the addition of 3,4,3',4'-tetrachlorobiphenyl. CYP1A2 was more active than CYP1A1 at oxidizing bilirubin in presence of NADPH alone and reacted to addition of 3,4,3',4'-tetrachlorobiphenyl with a depression rather than a stimulation of bilirubin oxidation. We have also tested a bacterial enzyme, CYP102. Dodecanoic acid and its polyhalogenated analogue (perfluorododecanoic acid) both stimulated NADPH oxidation by CYP102, but only the perfluoro analogue stimulated markedly bilirubin oxidation. The analogue exhibited much greater potency than the normal substrate in stimulating NADPH and bilirubin oxidation and also showed greater affinity for CYP102, as measured by the binding constant, Ks. The molar stoichiometry ratio between NADPH and O(2) consumption was 1 in the case of the substrate, but approximated 2 with the perfluoro analogue. We conclude that halogenated substrate analogues can interact with different CYPs to increase production of oxidative species, probably by an uncoupling mechanism. A role of the ferryl-oxygen intermediate is suggested in the oxidation of biologically important molecules, with possible implications for the therapy of jaundice and for toxic oxidative reactions, such as uroporphyria and cancer.

Bilirubin and uroporphyrinogen oxidation by induced cytochrome P4501A and cytochrome P4502B. Role of polyhalogenated biphenyls of different configuration.

In previous work it was shown that hepatic microsomes from rats treated with 3-methylcholanthrene and similar inducers had increased bilirubin-degrading activity. The activity was further stimulated by addition of 3,4-tetrachlorobiphenyl (TCB), a response specifically dependent on CYP1A1. Here, we compared the effect of adding PCBs of either planar or non-planar configuration on rate of bilirubin degradation, monooxygenase activity and NADPH/O(2) consumption by liver microsomes from animals treated with either phenobarbital or 3-methylcholanthrene/beta-naphthoflavone. We also examined the oxidation of uroporphyrinogen (hexahydro-uroporphyrin) (URO'gen) under these conditions. Polychlorinated biphenyl (PCBs) stimulated the rate of bilirubin and URO'gen oxidation with microsomes expressing high levels of either CYP2B or CYP1A, inhibiting at the same time their monooxygenase activities (PROD and EROD, respectively); however, non-planar di-ortho-substituted PCBs were preferentially active with phenobarbitone-induced microsomes, in contrast to those active with 3-methylcholanthrene/beta-naphthoflavone microsomes, where a planar configuration was required for activity. An antibody raised against CYP2B1 markedly inhibited the PCB-dependent bilirubin degradation and PROD activities of phenobarbital-induced microsomes with similar dose-response curves for the two effects. Increased microsomal utilizations of NADPH and O(2) were also caused by PCBs with both types of induced microsomes and here again PCBs of different configuration were preferentially active. It is concluded that PCBs of the appropriate configuration may interact with either CYP1A1 or CYP2B1, increase production of oxidative species by an uncoupling mechanism, and lead to oxidation of target molecules in the cell, among these uroporphyrinogen and bilirubin.