Pharmacokinetic Interactions Between the Orexin Receptor Antagonist Almorexant and the CYP3A4 Inhibitors Ketoconazole and Diltiazem
INTRODUCTION
Antagonism of orexin receptors is a new approach for the treat- ment of insomnia, which is currently being tested in humans.1,2 The orexin system was discovered in the late nineties and con- sists of two hypothalamic peptides, orexin-A and orexin-B (also called hypocretins), which exert their effects via two receptors, OX1 and OX2.3,4
This system plays a central role in the regula- tion of arousal and sleep–wake balance.5–7 Following the obser- vation that almorexant could induce sleep in rats and dogs, this compound was the first orexin receptor antagonist to be inves- tigated in man.8,9 In healthy subjects, daytime administration of almorexant resulted in reduction of vigilance, alertness, and visuomotor and motor coordination at single doses of 400 mg and higher.
Furthermore, somnolence was the most frequently reported adverse event. Together these observations provided the basis for the proof-of-concept study in patients with insom- nia. In this double-blind placebo-controlled study, almorexant doses as low as 100 mg significantly improved sleep efficiency.
Effects on secondary endpoints, both objective (polysomnogra- phy measurements) and subjective (self-rating sleep and awak- ening quality questionnaire), were all consistent with the no- tion that almorexant enabled and maintained sleep in insomnia patients.10
The pharmacokinetic profile of almorexant is compatible with the requirements of a sleep-inducing compound. Most no- tably, the time to maximum concentration, tmax, ranged from 0.7 to 2.3 and 8 h after administration almorexant concentrations had decreased to less than 20% of the maximum concentration, Cmax.
The latter is important as it minimizes the risk of reduced next-day performance in patients. Drug disposition was multi- phasic with a distribution phase, which was responsible for the major part of drug disposition from plasma, and an elimination phase characterized by an apparent terminal elimination half- life (t1/2) varying from 13.1 to 19.0 h.9
Following administration of a radioactive dose, 78.0% of the administered radioactivity was recovered in feces and 13.5% in urine. Unchanged almorex- ant was not found in urine and represented 10% of the admin- istered dose in feces.
Four primary metabolites were identified: the isomeric phenols M3 and M8, formed by demethylation; the aromatic isoquinolinium ion M5, formed by dehydrogena- tion; and M6, formed by oxidative dealkylation with loss of the phenylglycine moiety (Fig. 1).
In addition, 44 other metabolites were found that are mainly formed by permutations of the pri- mary metabolites followed by conjugation of the intermediate phenols with glucuronic or sulfonic acid.11 All four major metabolic pathways are primarily dependent on CYP3A4 with only minor contributions of other CYP isoen- zymes (Fig. 1).
Therefore, compounds inhibiting or inducing CYP3A4 activity may affect the pharmacokinetics of almorex- ant. We report here the results of two drug–drug interaction studies performed in healthy subjects investigating the effects of ketoconazole and diltiazem, a strong and moderate inhibitor of CYP3A4, respectively.12
METHODS
Subjects
In order to be eligible for the study, participating male sub- jects in both studies had to be from 18 to 45 years of age, with a body mass index of 18–28 kg/m2, systolic blood pressure of 100–145 mmHg, diastolic blood pressure of 50–90 mmHg, and heart rate of 55–90 beats per min.
They were in good health as further assessed by medical history, physical examination, elec- trocardiogram, and clinical laboratory tests. A known allergy or contraindication to diltiazem or ketoconazole (as appropri- ate), history of alcoholism or drug abuse, smoking, treatment with any medication (including herbal and over-the-counter) within 2 weeks prior to the screening examination, and ex- cessive caffeine consumption were among the exclusion crite- ria.
Restrictions were applied regarding the consumption of methylxanthine- and alcohol-containing beverages, grapefruit, and grapefruit juice.
The sample size in both studies was based on empirical con- siderations. An effect of the CYP3A4 inhibitors on the phar- macokinetics of almorexant was expected. Based on a mixed model with treatment as fixed factor and subject as random factor, the “within subject” standard deviation on a log scale of 0.31 for AUC0–∞ and 0.55 for Cmax of almorexant were esti- mated using data from a previously performed drug–drug in- teraction study.13
With a sample size of 18 evaluable subjects, the lower and upper bounds of the 90% CI for the true ratio test (almorexant with ketoconazole) versus reference (almorexant) were estimated to be within 17% for AUC0–∞ and 27% for Cmax.
Thus, the 90% confidence interval was estimated to be approximately (0.83–1.20) for AUC0–∞ and (0.73–1.37) for Cmax if the estimated geometric mean ratio was 1. Therefore, in each study, 20 subjects were to be enrolled to have at least 18 evaluable subjects.
The almorexant/ketoconazole and almorexant/diltiazem studies were approved by the ethics committees of the A¨ rztekammer Nordrein, Du¨ sseldorf, Germany and Lan- desa¨ rztekammer Baden-Wu¨ rttemberg, Stuttgart, Germany, re- spectively.
All subjects gave written informed consent prior to any study-specific procedures were performed and the stud- ies were conducted in full conformity with the Declaration of Helsinki.
Study Design
The ketoconazole/almorexant and diltiazem/almorexant stud- ies had a single-center, open-label, randomized, two-way crossover design. Treatment A in both studies consisted of a single dose of 100 mg almorexant in the morning. Treatment B in the ketoconazole/almorexant study consisted of 400 mg ke- toconazole once-a-day in the morning for 14 days and a single dose of 100 mg almorexant on day 5.
Each dose of ketoconazole was administered as two tablets of 200 mg Nizoral⃝R . In the diltiazem study, treatment B consisted of 300 mg diltiazem once-a-day for 11 days with concomitant almorexant 100 mg on day 4. Diltiazem was administered as a 300 mg sustained- release capsule (Tildiem⃝R ).
In treatment B, almorexant was administered at the time point when plasma concentrations of the CYP3A4 inhibitor were expected to have reached steady- state conditions and CYP3A4 inhibition to be maximal.14,15 Continued dosing of ketoconazole and diltiazem after intake of the single almorexant dose was carried out in order to main- tain constant CYP3A4 inhibition during the elimination phase.
The results of the ketoconazole/almorexant study were known prior to the conduct of the diltiazem/almorexant study and in- dicated that 8 days of concomitant dosing of almorexant and diltiazem would be sufficient.
On days of concomitant adminis- tration, drugs were administered to subjects in the fasted state, whereas on all other days, the CYP3A4 inhibitor was admin- istered within 30 min after intake of a standard breakfast.16 Food has been shown to modestly affect the pharmacokinetics of almorexant17 and a sleep-inducing agent is generally taken several hours after the evening meal, that is, on an empty stomach.
In both treatment periods, which were separated by a washout of about 2 weeks, subjects were confined to the study center from the evening before until 24 h after almorexant administration. All other assessments were performed on an ambulatory basis for which the subjects returned to the study center.
Data Analysis
Noncompartmental analysis was performed on the individual concentration data using WinNonlin Professional (Version 5.0; Pharsight Corporation, Mountain View, California). The vari- ables tmax and Cmax were directly read from the individual plasma concentration–time curves.
The area under the plasma concentration–time curve (AUC) was calculated using the lin- ear trapezoidal rule. Estimation of the elimination rate con- stant, 8z, was performed by least-squares regression of the log- transformed concentration data in the terminal phase using data from at least three time points. The t1/2 was calculated as ln2/8z.
The variable 8z could not be reliably estimated for all subjects and for all four metabolites and, therefore, AUC from time zero to 24 h after almorexant administration (AUC0–24 h) was calculated as a measure of exposure. For almorexant, the AUC from 0 to infinity (AUC0–∞) was calculated as the sum of AUC0–t plus Clast/8z whereby AUC0–t is the AUC from time zero to the last measurable concentration (Clast).
Pharmacokinetic variables were summarized with geomet- ric mean and 95% confidence limits or for tmax with median and minimum and maximum values. Differences between treat- ments were explored by calculating geometric mean ratios and 90% confidence limits for the variables Cmax and AUC with treatment A as reference.
RESULTS
Subjects
Nineteen men with a mean age of 32 years (range 21–43 years) and a mean body mass index of 23.6 kg/m2 (range 20.0–28.5 kg/m2) were enrolled and received at least one dose of study drug in the ketoconazole/almorexant study.
Of these 19 sub- jects, 17 completed the study as per protocol and were in- cluded in the pharmacokinetic analysis. Because of the occur- rence of two serious adverse events (see under Safety Results), the study was terminated before all 19 subjects had received both treatments.
Two subjects prematurely withdrew from the study. The study population in the diltiazem/almorexant study also consisted of 19 men. They had a mean age of 31.5 years (range 22–44 years), a mean body mass index of 24.0 kg/m2 (range 20.0–27.9 kg/m2) and all but one completed the study according to protocol and were included in the pharmacokinetic analysis.
Pharmacokinetic Results
The pharmacokinetics of almorexant and its metabolites were remarkably similar when comparing the ketocona- zole/almorexant and diltiazem/almorexant studies (Figs. 2–4). Under fasting conditions and in the absence of ketoconazole or diltiazem, almorexant was quickly absorbed with a median tmax of 0.75–1.0 h (Tables 1 and 2).
Following attainment of Cmax, plasma concentrations quickly decreased and were low 8 h after administration. Almorexant plasma concentrations were markedly higher when administered concomitantly with ketoconazole with an increase in Cmax and AUC of 5.8 and 10.5- fold, respectively.
Ketoconazole did not affect tmax but the t1/2 of almorexant was prolonged (Fig. 2 and Table 1). The distribution phase of almorexant did not appear to be affected by ketocona- zole as the mean (SD) plasma concentration 8 h after drug intake, 46.9 (15.2) ng/mL, represented only approximately 15% of Cmax.
In contrast, the tmax of all four metabolites was longer in the presence of ketoconazole. Exposure (both Cmax and AUC) to metabolites M3 and M8 was markedly increased by ketocona- zole whereas formation of metabolites M5 and M6 was reduced (Fig. 3 and Table 1).
When compared with ketoconazole, concomitant administra- tion of diltiazem had qualitatively similar effects on the phar- macokinetics of almorexant and its metabolites but effects were less pronounced.
There was, however, one notable difference between ketoconazole and diltiazem: exposure to M5 slightly increased in the presence of diltiazem whereas it decreased fol- lowing combined administration with ketoconazole (Fig. 5 and Table 2).
DISCUSSION
The fact that CYP3A4 is a major metabolizing enzyme of al- morexant rendered the conduct of drug–drug interaction stud- ies with inhibitors of this enzyme necessary. The effects of both a moderate and strong inhibitor on the pharmacokinetics of almorexant were investigated.
Results from both studies clearly demonstrate an increase in exposure to almorexant when this compound is coadministered with CYP3A4 inhibitors. The effects of diltiazem on almorexant pharmacokinetics were less pronounced than those of ketocona- zole, which confirms the notion that diltiazem and ketoconazole are a moderate and strong inhibitor, respectively, of CYP3A4.12
Cmax was markedly more affected by both inhibitors (5.8-fold in- crease with ketoconazole) than t1/2, which was “only” prolonged by about 50% in the presence of ketoconazole. This indicates that the first-pass effect of almorexant is primarily reduced by ketoconazole and diltiazem and to a lesser extent its elimina- tion.
The consequence is a markedly increased bioavailability of almorexant. Furthermore, the distribution phase is not af- fected by CYP3A4 inhibition as evidenced by low almorexant concentrations when compared to Cmax 8 h after intake.
As- suming appropriate dose adjustment of almorexant is possible, this may indicate that there is no increased risk of reduced next-day performance when almorexant is coadministered with compounds that inhibit CYP3A4.
Consistent with the observed increase in exposure to almorexant, both studies showed an in- creased incidence of almorexant-related adverse events after combined administration when compared with almorexant alone.
In the case where a CYP3A4 substrate is metabolized to form one metabolite, which is not further metabolized, administra- tion of ketoconazole will lead to an increase in exposure to the substrate and a decrease in exposure to the metabolite.
The metabolism of almorexant is, however, far more complex. For- mation of the primary metabolites is dependent on CYP3A4 but all are further metabolized. The subsequent metabolic reactions may be CYP3A4 mediated or not. If mediated by CYP3A4, then most of the primary metabolite, which is still formed because ketoconazole-induced inhibition of CYP3A4 is not 100%, will not be further metabolized, which may result in an increase of the exposure to this primary metabolite.
In contrast, if CYP3A4 is not involved in subsequent metabolic reactions then exposure to the primary metabolite(s) decreases as expected. Plasma concentrations of metabolites M3 and M8 increased after combined administration with ketoconazole, whereas those of M5 and M6 decreased.
The above result in- dicates that the subsequent metabolism of M5 and M6 is not dependent on CYP3A4, whereas the ketoconazole-induced in- crease in exposure to M3 and M8 suggests that CYP3A4 also plays a role in the further disposition of M3 and M8.
Another example of increased exposure to a drug and a primary metabo- lite under conditions of CYP3A4 inhibition when both are sub- strates for this isoenzyme is simvastatin and its metabolite, 6ß-hydroxysimvastatin.19
The metabolite M5 behaved differ- ently with both CYP3A4 inhibitors: AUC and Cmax decreased with ketoconazole but increased slightly with diltiazem. This difference may be explained by the nature of the inhibitors, that is, strong versus moderate inhibition, and/or suggest that the formation of M5 is not purely CYP3A4 dependent and that other CYP isoenzymes may be involved.
Of note, none of the four primary metabolites of almorexant are themselves antag- onists of orexin receptors and do not contribute to the sleep- inducing properties of almorexant.
They may of course be of importance in the safety profile of almorexant but, as far as currently known, almorexant-induced adverse effects in hu- mans are mainly related to its pharmacological action, that is, antagonism of orexin receptors.
In addition, toxicology studies with almorexant had shown a sufficient safety margin for these metabolites in at least one species (Harvey et al., unpublished results).
The absolute oral bioavailability of almorexant is only 11.2%, which is probably because of the extensive first-pass metabolism.18 This feature makes almorexant a likely candi- date for a food–drug interaction with grapefruit juice possibly resulting in increased plasma concentrations and pharmacody- namic effects, both therapeutic and adverse.
Grapefruit juice, or more precisely the furanocoumarins that it contains, de- grades intestinal CYP3A4, which provides the mechanism for the numerous pharmacokinetic interactions with drugs that have been reported.20,21
Nondrug–drug interactions such as with grapefruit juice or St. John’s wort, the latter an inducer of CYP3A4,22 have not received the attention in drug development they may deserve most likely due to an absence of guidance from health authorities.
In conclusion, the exposure to almorexant is increased up to 10.5-fold by concomitant administration of CYP3A4 inhibitors. This indicates that dose adaptation must be considered when almorexant is coadministered with inhibitors of CYP3A4.