Su Xiang Duan

Protease Inhibitors as Inhibitors of Human Cytochromes P450: High Risk Associated with Ritonavir

Four protease inhibitor antiviral agents (ritonavir, indinavir, nelfinavir, saquinavir) were evaluated as in vitro inhibitors of the activity of six human cytochromes using an in vitro model based on human liver microsomes. Ritonavir was a highly potent inhibitor of P450‐3A activity (triazolam hydroxylation), having inhibitory potency slightly less than ketoconazole. Indinavir was also a potent 3A inhibitor, while nelfinavir and saquinavir were less potent. Ritonavir had high inhibition potency against cytochrome P450‐2C9 (tolbutamide hydroxylation), −2C19 (S‐mephenytoin hydroxylation), and −2D6 (dextromethorphan O‐demethylation and desipramine hydroxylation), while the other protease inhibitors had one or more orders of magnitude lower inhibitory activity against these reactions. None of the protease inhibitors had important inhibitory potency against P450‐1A2 (phenacetin O‐deethylation) or −2E1 (chlorzoxazone hydroxylation). Thus, among available protease inhibitors, ritonavir carries the highest risk of incurring drug interactions due to inhibition of cytochrome P450 activity.

Midazolam Hydroxylation by Human Liver Microsomes In Vitro: Inhibition by Fluoxetine, Norfluoxetine, and by Azole Antifungal Agents

Biotransformation of the imidazobenzodiazepine midazolam to its α‐hydroxy and 4‐hydroxy metabolites was studied in vitro using human liver microsomal preparations. Formation of α‐hydroxy‐midazolam was a high‐affinity (Km = 3.3 μmol/L) Michaelis‐Menten process coupled with substrate inhibition at high concentrations of midazolam. Formation of 4‐hydroxy‐midazolam had much lower apparent affinity (57 μmol/L), with minimal evidence of substrate inhibition. Based on comparison of Vmax/Km ratios for the two pathways, α‐hydroxy‐midazolam formation was estimated to account for 95% of net intrinsic clearance. Three azole antifungal agents were inhibitors of midazolam metabolism in vitro, with inhibition being largely consistent with a competitive mechanism. Mean competitive inhibition constants (Ki) versus α‐hydroxy‐midazolam formation were 0.0037 μmol/L for ketoconazole, 0.27 μmol/L for itraconazole, and 1.27 μmol/L for fluconazole. An in vitro‐in vivo scaling model predicted inhibition of oral midazolam clearance due to coadministration of ketoconazole or itraconazole; the predicted inhibition was consistent with observed interactions in clinical pharmacokinetic studies. The selective serotonin reuptake inhibitor (SSRI) antidepressant fluoxetine and its principal metabolite, norfluoxetine, also were inhibitors of both pathways of midazolam biotransformation, with norfluoxetine being a much more potent inhibitor than was fluoxetine itself. This finding is consistent with results of other in vitro studies and of clinical studies, indicating that fluoxetine, largely via its metabolite norfluoxetine, may impair clearance of P450‐3A substrates.

Inhibition of alprazolam and desipramine hydroxylation in vitro by paroxetine and fluvoxamine : comparison with other selective serotonin reuptake inhibitor antidepressants

In vitro preparations of human liver microsomes were used to study the inhibiting effects of two selective serotonin reuptake inhibitor (SSRI) antidepressants, paroxetine and fluvoxamine, on metabolism via hydroxylation of alprazolam and of desipramine. These reactions are mediated by Cytochromes P450-3A4 and P450-2D6, respectively. Paroxetine was a highly potent inhibitor of desipramine hydroxylation; the inhibition constant (Ki) value of 2.0 microM indicated greater inhibiting potency than fluoxetine or norfluoxetine. The in vitro data predicted in vivo impairment of desipramine clearance by coadministration of paroxetine which was in the same range as observed in a clinical study. Fluvoxamine, by contrast, was a much weaker inhibitor of desipramine hydroxylation, having a Ki value (16.6 microM) similar to those of sertraline and desmethylsertraline. For hydroxylation of alprazolam, paroxetine was a relatively weak inhibitor, approximately comparable to fluoxetine, whereas fluvoxamine showed inhibiting capacity similar to that of norfluoxetine. The in vitro data predicted the degree of impairment of alprazolam clearance observed in vitro model can therefore provide clinically relevant data on prediction of potential drug interactions with SSRIs.

In Vitro Prediction of the Terfenadine‐Ketoconazole Pharmacokinetic Interaction

Biotransformation of the peripherally acting H‐1 histamine antagonist, terfenadine, to its desalkyl and hydroxy metabolites was studied in vitro using microsomal preparations from six separate human livers. These metabolic reactions are mediated by the specific cytochrome P450‐3A4. Addition of ketoconazole to the reaction mixtures reduced the rate of formation of both metabolites in a manner consistent with competitive inhibition. Ketoconazole inhibition constants (Ki) averaged 0.024 μM for the desalkyl terfenadine pathway, and 0.237 μM for the hydroxy terfenadine pathway. A mathematical model, based on the in vitro Ki values and the usual clinical range of plasma ketoconazole concentrations (1–5 μg/mL; 1.88 − 0.94 μM), predicted that plasma terfenadine levels during coadministration of ketoconazole would increase by a factor ranging from 13‐fold to 59‐fold relative to the same dose of terfenadine given without ketoconazole. Actual plasma terfenadine levels during terfenadine‐ketoconazole coadministration in a clinical pharmacokinetic study were close to those predicted by the model. These plasma levels were associated with prolongation of the corrected QT interval, thereby explaining the potentially life‐threatening ventricular arrhythmias reportedly associated with terfenadine‐ketoconazole cotherapy. Thus, data from studies of drug metabolism in vitro can be used to predict and thereby possibly avoid clinically important drug interactions.

Inhibition of Human Cytochrome P450 Isoforms by Nonnucleoside Reverse Transcriptase Inhibitors

The capacity of three clinically available nonnucleoside reverse transcriptase inhibitors (NNRTIs) to inhibit the activity of human cytochromes P450 (CYPs) was studied in vitro using human liver microsomes. Delavirdine, nevirapine, and efavirenz produced negligible inhibition of phenacetin O‐deethylation (CYP1A2) or dextromethorphan O‐demethylation (CYP2D6). Nevirapine did not inhibit hydroxylation of tolbutamide (CYP2C9) or S‐mephenytoin (CYP2C19), but these CYP isoforms were importantly inhibited by delavirdine and efavirenz. This indicates the likelihood of significantly impaired clearance of CYP2C substrate drugs (such as phenytoin, tolbutamide, and warfarin) upon initial exposure to these two NNRTIs. Delavirdine and efavirenz (but not nevirapine) also were strong inhibitors of CYP3A, consistent with clinical hazards of initial cotreatment with either of these drugs and substrates of CYP3A. The in vitro microsomal model provides relevant predictive data on probable drug interactions with NNRTIs when the mechanism is inhibition of CYP‐mediated drug biotransformation. However, the model does not incorporate interactions attributable to enzyme induction.

Phenacetin O-deethylation by human liver microsomes in vitro: inhibition by chemical probes, SSRI antidepressants, nefazodone and venlafaxine

Abstract Biotransformation of phenacetin via O-deethylation to acetaminophen, an index reaction reflecting activity of Cytochrome P450-1A2, was studied in microsomal preparations from a series of human livers. Acetaminophen formation was consistent with a double Michaelis-Menten system, with low-Km (mean Km1 = 68 μM) and high-Km (mean Km2 = 7691 μM) components. The low-Km enzyme accounted for an average of 96% of estimated intrinsic clearance, and was predicted to contribute more than 50% of net reaction velocity at phenacetin concentrations less than 2000 μM. Among index inhibitor probes, α-naphthoflavone was a highly potent inhibitor of the low-Km enzyme (Ki1 = 0.013 μM); furafylline also was a moderately active inhibitor (Ki1 = 4.4 μM), but its inhibiting potency was increased by preincubation with microsomes. Ketoconazole was a relatively weak inhibitor (Ki1 = 32 μM); quinidine and cimetidine showed minimal inhibiting activity. Among six selective serotonin reuptake inhibitor (SSRI) antidepressants, fluvoxamine was a potent inhibitor of 1A2 (mean Ki1 = 0.24 μM). The other SSRIs were more than tenfold less potent. Mean Ki1 values were: fluoxetine, 4.4 μM; norfluoxetine, 15.9 μM; sertraline, 8.8 μM; desmethylsertraline, 9.5μM; paroxetine, 5.5 μM. The antidepressant nefazodone and four of its metabolites (meta-chloro-phenylpiperazine, two hydroxylated derivatives, and a triazoledione) were very weak inhibitors of P450-1A2. Venlafaxine and its O- and N-desmethyl metabolites showed minimal inhibitory activity.

Inhibition of terfenadine metabolism in vitro by azole antifungal agents and by selective serotonin reuptake inhibitor antidepressants: relation to pharmacokinetic interactions in vivo.

Biotransformation of the H-1 antagonist terfenadine to its desalkyl and hydroxy metabolites was studied in vitro using microsomal preparations of human liver. These metabolic reactions are presumed to be mediated by Cytochrome P450-3A isoforms. The azole antifungal agent ketoconazole was a highly potent inhibitor of both reactions, having mean inhibition constants (Ki) of 0.037 and 0.34 microM for desalkyl- and hydroxy-terfenadine formation, respectively. Itraconazole also was a potent inhibitor, with Ki values of 0.28 and 2.05 microM, respectively. Fluconazole, on the other hand, was a weak inhibitor. Six selective serotonin reuptake inhibitor antidepressants tested in this system were at least 20 times less potent inhibitors of terfenadine metabolism than was ketoconazole. An in vitro-in vivo scaling model used in vitro Ki values, typical clinically relevant plasma concentrations of inhibitors, and presumed liver:plasma partition ratios to predict the degree of terfenadine clearance impairment during coadministration of terfenadine with these inhibitors in humans. The model predicted a large and potentially hazardous impairment of terfenadine clearance by ketoconazole and, to a slightly lesser extent, by itraconazole. However, fluconazole and the six selective serotonin reuptake inhibitors (SSRIs) at usual clinical doses were not predicted to impair terfenadine clearance to a degree that would be of clinical importance. Caution is nonetheless warranted with the coadministration of SSRIs and terfenadine when high doses of SSRIs (particularly fluoxetine) are administered. Also, some individuals may be unusually susceptible to metabolic inhibition for a variety of reasons.

Inhibition of triazolam clearance by macrolide antimicrobial agents: In vitro correlates and dynamic consequences

Macrolide antimicrobial agents may impair hepatic clearance of drugs metabolized by cytochrome P4503A isoforms. Potential interactions of triazolam, a substrate metabolized almost entirely by cytochrome P4503A in humans, with 3 commonly prescribed macrolides were identified using an in vitro metabolic model. The actual interactions, and their pharmacodynamic consequences, were verified in a controlled clinical study.

Citalopram and desmethylcitalopram in vitro: human cytochromes mediating transformation, and cytochrome inhibitory effects

BACKGROUND Biotransformation of citalopram (CT), a newly available selective serotonin reuptake inhibitor antidepressant, to its principal metabolite, desmethycitalopram (DCT), and the capacity of CT and DCT to inhibit human cytochromes P450, were studied in vitro. METHODS Formation of DCT from CT was evaluated using human liver microsomes and microsomes from cDNA-transfected human lymphoblastoid cells. Cytochrome inhibition by CT and DCT in liver microsomes was studied using isoform-specific index reactions. RESULTS Formation of DCT from CT in liver microsomes had a mean apparent K(m) of 174 mumol/L. Coincubation with 1 mumol/L ketoconazole reduced reaction velocity to 46 to 58% of control values, while omeprazole, 10 mumol/L, reduced velocity to 80% of control. Quinidine produced minimal inhibition. DCT was formed from CT by heterologously expressed human P450-2D6, -2C19, -3A4. After accounting for the relative abundance of individual cytochromes, 3A4 and 2C19 were estimated to make major contributions to net reaction velocity, with a possible contribution of 2D6 at therapeutic CT concentrations. CT and DCT themselves produced negligible inhibition of 2C9, 2E1, and 3A, and only weak inhibition of 1A2, 2C19, and 2D6. CONCLUSIONS Formation of DCT from CT is mediated mainly by P450-3A4 and 2C19, with an additional contribution of 2D6. CT at therapeutic doses in humans may produce a small degree of inhibition of P450-1A2, -2C19, and -2D6, but negligible inhibition of P450-2C9, -2E1, and -3A.

Potent mechanism-based inhibition of human CYP3A in vitro by amprenavir and ritonavir: comparison with ketoconazole

AbstractObjective: Biotransformation of triazolam to its α-hydroxy and 4-hydroxy metabolites by human liver microsomes in vitro was used as an index of human cytochrome P450 3A (CYP3A) activity. Results: The reaction was strongly inhibited by co-incubation with the viral protease inhibitors ritonavir (IC50=0.14 μM) and amprenavir (IC50=2.5–2.9 μM), and by the azole derivative ketoconazole (IC50 = 0.07 μM). Pre-incubation of microsomes with ritonavir or amprenavir increased inhibitory potency (IC50 reduced to 0.07 μM and 1.4 μM, respectively). This was not the case with ketoconazole. Conclusions: Thus, ritonavir and amprenavir are highly potent mechanism-based inhibitors of human CYP3A isoforms.