Simultaneous phenotyping of CYP2E1 and CYP3A using oral chlorzoxazone and midazolam microdoses

Abstract

Aims

Chlorzoxazone is the paradigm marker substrate for CYP2E1 phenotyping in vivo. Because at the commonly used milligram doses (250–750 mg) chlorzoxazone acts as an inhibitor of the CYP3A4/5 marker substrate midazolam, previous attempts failed to combine both drugs in a common phenotyping cocktail. Microdosing chlorzoxazone could circumvent this problem.

Method

We enrolled 12 healthy volunteers in a trial investigating the dose–exposure relationship of single ascending chlorzoxazone oral doses over a 10,000‐fold range (0.05–500 mg) and assessed the effect of 0.1 and 500 mg of chlorzoxazone on oral midazolam pharmacokinetics (0.003 mg).

Results

Chlorzoxazone area under the concentration–time curve was dose‐linear in the dose range between 0.05 and 5 mg. A nonlinear increase occurred with doses ≥50 mg, probably due to saturated presystemic metabolic elimination. While midazolam area under the concentration–time curve increased 2‐fold when coadministered with 500 mg of chlorzoxazone, there was no pharmacokinetic interaction between chlorzoxazone and midazolam microdoses.

Conclusion

The chlorzoxazone microdose did not interact with the CYP3A marker substrate midazolam, enabling the simultaneous administration in a phenotyping cocktail. This microdose assay is now ready to be further validated and tested as a phenotyping procedure assessing the impact of induction and inhibition of CYP2E1 on chlorzoxazone microdose pharmacokinetics.

What is already known about this subject

• Chlorzoxazone is the paradigm marker substrate for CYP2E1 phenotyping in vivo.

• Currently, doses of 250–750 mg are used for this purpose with the disadvantage of saturation of metabolism and a drug interaction with the prototypical CYP3A marker substrate midazolam.

What this study adds

• Oral chlorzoxazone exhibits linear pharmacokinetics in the low dose range from 0.05–5 mg and nonlinear pharmacokinetics between 50 and 500 mg.

• An oral chlorzoxazone microdose of 0.1 mg does not lead to a pharmacokinetic interaction with a midazolam microdose.

• A chlorzoxazone microdose can be integrated into a phenotyping cocktail without interfering with the other important probe drugs and may prove to be superior for phenotyping than the milligram doses that cause saturation of CYP‐dependent metabolism and ambiguities of phenotyping results.

1. INTRODUCTION

The cytochrome P450 isozyme (CYP) 2E1 is a major drug‐metabolising enzyme that accounts for 9–37% of total CYP expression in the liver.1, 2 It is also expressed in several extra‐hepatic tissues such as kidney, lung and lymphocytes but not in the intestinal epithelium.1, 2, 3 CYP2E1 is involved in the metabolism of endogenous substrates such as acetone and fatty acids4 and is a key metabolic elimination pathway for inhalation narcotics (e.g. enflurane, halothane),5, 6 theophylline 7 and the centrally acting spasmolytic chlorzoxazone.8 CYP2E1 contributes to the detoxification of organic compounds such as aniline, benzene and ethanol.9, 10, 11 While ethanol induces CYP2E1 expression,12 disulfiram and clomethiazole, which are used in the acute and chronic detoxification of alcohol‐dependent patients, are inhibitors of CYP2E1.13, 14 Clomethiazole is also a CYP2E1 substrate, and its exposure and hence toxicity is increased in patients who have severe alcohol‐induced liver cirrhosis (10‐fold area under the concentration–time curve [AUC] increase)15 and in healthy subjects when coadministered with ethanol (2‐fold AUC increase).16

The reaction of chlorzoxazone to 6‐hydroxy‐chlorzoxazone is almost exclusively catalysed by CYP2E1 with only minor, but concentration‐dependent, involvement of CYP1A2 and CYP1A1 in its metabolism.8, 17, 18 Chlorzoxazone is therefore used as a specific CYP2E1 probe19, 20 and recommended as a marker substrate by regulators for the in vitro and in vivo investigation of CYP2E1‐dependent drug metabolism.21 Because of the short half‐life (t_½) of approximately 1 hour, sampling periods of 8–10 hours are usually long enough to capture the whole AUC, but limited sampling methods depending on only 1 plasma sample drawn at the expected time of peak concentration (C_max) have also been proposed.19, 22

Currently, phenotyping with chlorzoxazone has 2 unaddressed drawbacks. (i) The use of different doses for CYP2E1 phenotyping yields inconsistent activity measures due to the nonlinear pharmacokinetics; in the dose range from 250 to 750 mg (as a tablet formulation), the dose‐normalised chlorzoxazone AUC was ~ 30% higher after a 750‐mg oral dose than after 250 mg.23 In another small series of male volunteers each taking 250, 500 and 750 mg of chlorzoxazone, nonlinear dose‐dependent pharmacokinetics of the parent compound and metabolite 2 hours after dosing were reported,24 which was presumably caused by saturation of metabolism.24

(ii) To accelerate and facilitate CYP phenotyping in interaction trials, drug cocktails enabling the simultaneous activity measurement of multiple drug‐metabolising enzymes and drug transporters have been devised.25 Until now, the incorporation of chlorzoxazone into phenotyping cocktails that contained the CYP3A marker substrate midazolam has failed due to a pharmacokinetic interaction between chlorzoxazone and midazolam.26 CYP inhibition has been shown to be concentration‐dependent and, therefore, also dose‐dependent,27, 28, 29 hence choosing a pharmacologically inactive dose may circumvent this drug interaction problem. Recently, the use of microdosed CYP marker substrates for in vivo phenotyping has been proposed30, 31 and cocktails incorporating microdosed substrates for the assessment of drug transporter activity have been developed.32

To investigate the potential of CYP2E1 phenotyping with microdosed chlorzoxazone, we conducted a clinical trial to investigate the linearity of chlorzoxazone pharmacokinetics from 0.05 to 500 mg and assessed the impact of regular and microdosed chlorzoxazone on the pharmacokinetics of oral midazolam.

2. METHODS

2.1. Study design

We enrolled 12 healthy volunteers into this randomised, open‐label dose escalation and drug interaction trial. First treatment started within 21 days of the screening visit. Participants were randomised to 1 of 2 dose arms (6 in each arm) after enrolment. To reduce the duration of participation and the blood volume drawn per participant, we randomised the volunteers into 2 arms: participants in trial arm A received 0.05, 0.1, 5 and 250 mg of oral chlorzoxazone. participants in trial arm B received 0.1, 0.5, 50 and 500 mg of oral chlorzoxazone (Figure 1). The washout time between each dosing was at least 48 hours. In the subsequent drug interaction part, all 12 participants received 0.003 mg of oral midazolam alone, then 0.003 mg of oral midazolam with 0.1 mg of oral chlorzoxazone followed by 0.003 mg of oral midazolam with 500 mg of oral chlorzoxazone (Figure 1).

Figure 1.

Trial design of the 1:1 randomised, open‐label single ascending dose trial of oral chlorzoxazone (CZZ) followed by a drug interaction trial assessing chlorzoxazone's perpetrator characteristics with the oral CYP3A substrate midazolam (MDL)

2.2. Regulatory and ethical requirements

Before enrolment of the first participant, the clinical trial protocol was approved by the national competent authority (Bundesinstitut für Arzneimittel und Medizinprodukte, Bonn, Germany) and the responsible Ethics Committee of the Medical Faculty of Heidelberg University. The trial was registered in the European clinical trial database (EudraCT 2014–003348‐11) prior to enrolment of participants. The trial was conducted at the Clinical Trial Centre of the Department of Clinical Pharmacology and Pharmacoepidemiology, which is certified according to DIN EN ISO 9001, and followed all principles of the Declaration of Helsinki (Version 2013, Fortaleza, Brazil), the guidelines of the International Conference on Harmonisation, especially ICH E6, the principles of good clinical practice, the corresponding European regulation (Directive 2001/20/EC), and all pertinent national legal requirements and regulations (Arzneimittelgesetz, Bundesdatenschutzgesetz and GCP‐Verordnung).

2.3. Study population

Healthy volunteers of both sexes, aged between 18 and 60 years, able to understand and willing to comply with the trial interventions and restrictions were considered for this trial. Participants were fully informed about trial procedures before they had to voluntarily sign the informed consent. They were physically and mentally healthy as determined by medical assessment consisting of a medical history, physical examination, an echocardiogram and a laboratory evaluation that did not reveal clinically relevant abnormalities. Males and females of child‐bearing potential were only included if they used reliable contraception with a Pearl Index <1% (i.e. 2 independent effective contraceptive methods) during the study and 2 weeks after the last administration of study medication.

Any regular drug treatment within the last 14 days before screening and during the trial was not allowed (except oral contraception) as was any use or abuse of drugs or substances such as prescription drugs, nonprescription drugs, and herbal or illicit substances. Also, regular smoking of >5 cigarettes daily and alcohol consumption in the 14 days before trial participation and during the trial was forbidden. Pregnant and lactating women were barred from trial participation.

2.4. Study drugs

Chlorzoxazone microdoses and low mg doses (dose levels 0.05, 0.1, 0.5, 5 and 50 mg) were administered as an oral solution with a volume of 200 mL. The different dose levels were produced from an aqueous stock solution that was diluted with tap water. The aqueous stock solution had a concentration of 0.025 mg/mL and was produced by using 250 mg Paraflex tablets (Biophausia AB, Stockholm, Sweden) by the central pharmacy of Heidelberg University Hospital. The 250‐ and 500‐mg dose levels were administered as crushed and stirred chlorzoxazone 250‐mg tablets (Paraflex) in 200 mL tap water. This was necessary because of the limited solubility of chlorzoxazone, which would necessitate that participants have to ingest excessive fluid volumes. Midazolam was administered as an oral solution as previously described.33 When both, midazolam and chlorzoxazone were administered, participants ingested the chlorzoxazone oral solution/tablet first and then, within 2 minutes, the oral midazolam solution. All study drugs were administered under fasting conditions, food intake was allowed 4 hours postdose.

2.5. Blood sampling and sample handling

During the dose escalation part, 4.9 mL of blood was drawn into a heparin tube from an indwelling venous catheter in an antecubital vein before chlorzoxazone dosing and 0.5, 1, 2, 3, 4, 5, 6, 7 and 8 hours postdose. During the drug interaction part, 7.5 mL of blood was drawn into a heparin tube before midazolam dosing and 0.5, 1, 2, 2.5, 3, 4 , 5, 6, 7, 8, 10 and 24 hours postdose. The tubes were centrifuged at 2500 g and 4°C for 10 minutes within 5 minutes of sampling. The supernatant (plasma) was transferred to 1 mL tubes using a pipette and then stored at −20°C until further analysis.

2.6. Quantification of midazolam

Midazolam was quantified as previously described.34 The lower limit of quantification was 93 fg/mL for midazolam and 255 fg/mL for the 1′‐hydroxy metabolite. For both analytes the batch‐to‐batch (within‐batch) accuracy ranged from 97.0 to 108% (95.5 to 110%) with precision between 4.4 and 12.2% (2.4 to 13.5% coefficient of variation [CV]).

2.7. Quantification of chlorzoxazone

Chlorzoxazone and its 6‐hydroxy metabolite was quantified using the previously published ultrasensitive ultraperformance liquid chromatography–tandem mass spectrometry method.35 The lower limit of quantification was 2.5 pg/mL for chlorzoxazone and 5 pg/mL for the 6‐hydroxy metabolite. For both analytes the batch‐to‐batch (within‐batch) accuracy ranged from 98.3 to 106% (94.9 to 111%) with precision between 3.6 and 8.5% CV (1.5–11.3% CV).

2.8. Pharmacokinetics

Pharmacokinetic parameters of chlorzoxazone and midazolam were calculated using Kinetica 5.0 (Thermo Fisher Scientific, Waltham, MA, USA). The AUC from time 0 to infinity (AUC_0–inf) was calculated using the log‐linear model implemented in Kinetica 5.0. C_max was the maximum concentration measured and T_max was the time to reach C_max. The t_½ was calculated as ln 2/elimination constant. The apparent volume of distribution at steady‐state (V_ss/F) was calculated as (dose/mean residual time)/AUC_0–inf where F is the bioavailability. Apparent oral clearance (CL/F) was calculated as CL/F = dose/AUC_0–inf.

2.9. Statistics

The statistical analysis was conducted using Prism 6.07 (Graphpad, La Jolla, CA, USA). Pharmacokinetic parameters of chlorzoxazone and midazolam are presented as geometric mean and 95% confidence interval, except for T_max, which is given as median and range. Concentration–time profiles are presented as mean and standard deviation; only time points for which concentrations from all participants were available are shown. Geometric mean ratios and corresponding 90% confidence intervals for chlorzoxazone and midazolam pharmacokinetic parameters were calculated after dose‐normalisation for C_max and AUCs. We ran a 1‐way ANOVA after log‐transformation followed by Dunnett's multiple comparisons test to assess differences between chlorzoxazone doses. The 0.1‐mg dose was the control group for the chlorzoxazone doses. The 0.003‐mg group without coadministration of chlorzoxazone served as the control group for the comparison of midazolam administration. When testing for significant differences between the pharmacokinetic parameters of the 0.1‐mg chlorzoxazone dose of trial arm A and Bas well as for female and male participants we used an unpaired t‐test of log transformed values using Prism 6.07. A P‐value <.05 was considered significant.

2.10. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY,49 and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18.50

3. RESULTS

3.1. Population

The study population consisted of 12 healthy volunteers (6 females) of Caucasian origin, aged 22–54 (median 26) years. Their mean weight was 70.1 (± 13.5) kg, mean height was 173 (± 9.5) cm, and mean body mass index was 23.2 (± 3.0) kg/m². One participant, with body mass index of 30.6 kg/m² had the highest chlorzoxazone clearance.

3.2. Pharmacokinetics of ascending doses of chlorzoxazone

The resulting plasma concentrations could be measured in each individual participant up to 5 hours for the 0.05 mg dose and for 8 hours when doses ≥0.1 mg were used. The resulting 6‐hydroxy metabolite concentrations could be measured in each individual participant up to 1 hours for the 0.05 mg dose, up to 2 hours for the 0.1 mg dose, up to 4 hours for the 0.5 mg group and for 8 hours when doses ≥5 mg were used. Pharmacokinetics of parent and metabolite were nonlinear in the investigated dose range (Figures 2 and 3; Tables 1 and 2). After dose normalisation, the dose–exposure relationship did not change in the dose range from 0.05 to 5 mg but started to become nonlinear with the 50 mg dose level (Tables 1 and 2) with an approximately 2‐fold increased geometric mean ratio (GMR) of AUC and C_max, while t_½ remained unchanged. Concomitant administration of a midazolam microdose (0.003 mg) did not change chlorzoxazone pharmacokinetics (Table S1). Participants in trial arm A had a slightly higher t_½ than participants in trial arm B (Table S2). The participants' sex did not influence chlorzoxazone pharmacokinetics (Table S3).

Figure 2.

Plasma concentration–time profiles (mean ± standard deviation) of (A) chlorzoxazone and (B) 6‐hydroxy‐chlorzoxazone after oral administration of ascending chlorzoxazone doses in 2 independent dose arms of 6 healthy volunteers (▲,▼); all 12 participants received the 0.1‐mg dose (O). LLOQ, lower limit of quantification

Figure 3.

Plasma concentration–time profiles of chlorzoxazone and 6‐hydroxy‐chlorzoxazone dose‐normalised to 1 mg in healthy volunteers after oral doses ranging from 0.05 to 500 mg. All dose groups included 6 participants except the 0.1 mg dose group, which included 12 participants

Table 1.

Descriptive statistics of the pharmacokinetic parameters of single ascending doses of oral chlorzoxazone in healthy volunteers

Parameter (unit)	Oral chlorzoxazone dose group
Oral dose (mg)	0.05	0.1	0.5	5	50	250	500
Dosage form	Oral solution	Oral solution	Oral solution	Oral solution	Oral solution	Tablets	Tablets
n =	6	12	6	6	6	6	6
C max (ng/mL)	0.834 [0.546–1.27]	1.97 [1.63–2.37]	7.52 [3.44–16.5]	97.3 [57.1–166]	1740 [1020‐2950]	6680 [3250‐13 700]	9970 [4030‐24 700]
C max (ng/mL) (dose‐normalised to 1 mg of CZZ)	16.7 [10.9–25.4]	19.7 [16.3–23.7]	15.0 [6.88–33.0]	19.5 [11.4–33.2]	34.8 [20.4–59.1]	26.7 [13.0–54.8]	19.9 [8.06–49.4]
T max (h)	0.50 [0.50–0.50]	0.50 [0.50–0.50]	0.50 [0.50–1.00]	0.50 [0.50–1.00]	0.50 [0.50–0.50]	1.00 [0.50–2.00]	2.00 [1.00–6.00]
AUC obs (ng/mL * h)	1.09 [0.638–1.86]	2.72 [2.10–3.52]	10.3 [5.01–21.3]	135 [66.1–276]	2150 [1450‐3200]	15 000 [8200‐27 300]	25 100 [11 200‐55 900]
AUC obs (ng/mL * h) (dose‐normalised to 1 mg of CZZ)	21.8 [12.8–37.2]	27.2 [21.0–35.2]	20.6 [10.0–42.6]	27.0 [13.2–55.2]	43.0 [29.0–64.0]	60.0 [32.8–109]	50.2 [22.4–112]
AUC 0–inf (ng/mL * h)	1.09 [0.641–1.87]	2.73 [2.10–3.54]	10.4 [5.01–21.4]	136 [65.9–281]	2160 [1450‐3210]	15 200 [8310‐27 800]	25 500 [11 600‐56 100]
CL/F (mL/min)	762 [446–1300]	612 [472–793]	805 [390–1660]	612 [296–1270]	386 [260–573]	274 [150–502]	326 [148–718]
t ½ (h)	0.845 [0.698–1.02]	0.89 [0.81–0.99]	0.840 [0.708–0.997]	0.948 [0.718–1.25]	0.805 [0.687–0.943]	0.910 [0.755–1.10]	0.845 [0.614–1.16]
V ss /F (L)	66.4 [44.1–99.9]	58.0 [46.6–71.9]	74.1 [34.7–158]	61.6 [37.0–103]	33.2 [21.9–50.6]	39.8 [21.2–75.0]	54.1 [19.4–151]

AUC_obs, observed area under the concentration–time curve; AUC_0–inf, area under the concentration–time curve extrapolated to infinity; C_max, observed peak plasma concentration; CL/F, apparent oral clearance; CZZ, chlorzoxazone; F, oral bioavailability; t_½, half‐life; T_max, time of peak concentration; and Vss/F, apparent oral volume of distribution. Data are reported as geometric mean [95% confidence interval] except for T_max, which is given as median [range].

Table 2.

Pharmacokinetics of ascending chlorzoxazone doses compared with a chlorzoxazone dose of 0.1 mg

	Statistical comparison of chlorzoxazone dose groups
	0.05 mg vs 0.1 mg	5 mg vs 0.1 mg	250 mg vs 0.1 mg	0.5 mg vs 0.1 mg	50 mg vs 0.1 mg	500 mg vs 0.1 mg
n =	6	6	6	6	6	6
Trial arm	A	A	A	B	B	B
GMR of C max (dose‐normalised to 1 mg of CZZ)	0.88 [0.74–1.05]	1.03 [0.84–1.26]	1.41 [0.96–2.06]	0.74 [0.43–1.28]	1.71 [1.11–2.62]	0.98 [0.51–1.88]
P‐value	.4037	.9877	.2738	.5871	.1201	.9997
GMR of AUC obs (dose‐normalised to 1 mg of CZZ)	0.79 [0.59–1.06]	0.98 [0.70–1.38]	2.17 [1.54–3.05]	0.78 [0.50–1.23]	1.64 [1.11–2.42]	1.91 [1.16–3.13]
P‐value	.3451	.9984	.0141	.5979	.1163	.1055
GMR of t ½	0.85 [0.76–0.91]	0.95 [0.79–1.14]	0.91 [0.78–1.07]	1.05 [0.93–1.18]	1.01 [0.83–1.21]	1.06 [0.78–1.43]
P‐value	.0615	.8943	.5369	.7262	.9996	.9590
GMR of V ss /F	1.10 [0.89–1.37]	1.02 [0.78–1.33]	0.54 [0.39–0.73]	1.33 [0.78–2.25]	0.59 [0.40–0.88]	0.96 [0.47–1.98]
P‐value	.7205	.9961	.0257	.6102	.1016	.9994

AUC_obs, observed area under the concentration–time curve; AUC_0–inf, area under the concentration–time curve extrapolated to infinity; C_max, observed peak plasma concentration; CL/F, apparent oral clearance; CZZ, chlorzoxazone; F, oral bioavailability; GMR, geometric mean ratio [90% confidence interval] of pharmacokinetic parameters; t_½, half‐life; T_max, time of peak concentration; and Vss/F, apparent oral volume of distribution. To test for statistically significant differences, the log‐transformed pharmacokinetic values were tested using ANOVA followed by Dunnett's test with the 0.1 mg group serving as a control group. Statistically significant values (P < .05) are highlighted in boldface type.

3.3. Pharmacokinetics of midazolam microdoses

Midazolam exposure increased 2.5‐fold when coadministered with 500 mg of oral chlorzoxazone but was unchanged when co‐administered with an oral chlorzoxazone microdose (Tables 3 and 4). Concurrently, midazolam C_max increased 1.6‐fold when co‐administered with 500 mg of oral chlorzoxazone, whereas it was unchanged when a chlorzoxazone microdose was given (Tables 3 and 4).

Table 3.

Pharmacokinetics of single midazolam microdoses (0.003 mg) with and without coadministration of oral chlorzoxazone in 12 healthy volunteers

	0.003 mg midazolam	0.003 mg midazolam +0.1 mg chlorzoxazone	0.003 mg midazolam +500 mg chlorzoxazone
C max (pg/mL)	13.1 [10.8–15.9]	14.2 [11.3–17.9]	21.2 [17.8–25.2]
T max (h)	0.75 [0.5–1.0]	0.75 [0.5–1.00]	1.00 [0.5–3.00]
AUC obs (pg/mL * h)	36.0 [28.3–45.8]	34.8 [26.9–45.1]	89.8 [75.8–106]
AUC 0–inf (pg/mL * h)	37.7 [29.4–48.4]	35.5 [27.3–46.0]	103 [77.7–138]
CL/F (mL/min)	1330 [1030–1700]	1410 [1090–1830]	483 [363–644]
t ½ (h)	3.47 [2.57–4.68]	3.39 [2.39–4.81]	5.61 [3.51–8.94]
V ss /F (L)	356 [286–442]	325 [275–383]	207 [167–258]

AUC_obs, observed area under the concentration–time curve; AUC_0–inf, area under the concentration–time curve extrapolated to infinity; C_max, observed peak plasma concentration; CL/F, apparent oral clearance; F, oral bioavailability; t_½, half‐life; T_max, time of peak concentration; and Vss/F, apparent oral volume of distribution. Geometric mean [95% confidence interval], except for T_max, which is given as median [range].

Table 4.

Comparison of pharmacokinetic parameters of 0.003 mg midazolam in the absence and presence of chlorzoxazone (0.1 or 500 mg)

	0.003 mg midazolam vs 0.003 mg midazolam (+ 0.1 mg chlorzoxazone)	0.003 mg midazolam vs 0.003 mg midazolam (+ 500 mg chlorzoxazone)
N =	12	12
GMR of midazolam C max	1.08 [0.88–1.34]	1.61 [1.42–1.83]
P‐value	.7307	<.0001
GMR of midazolam AUC obs	0.97 [0.77–1.21]	2.50 [2.09–2.98]
P‐value	.8525	<.0001

AUC_obs, observed area under the concentration–time curve; C_max, observed peak plasma concentration; and geometric mean ratios (GMR) [90% confidence interval] of pharmacokinetic parameters. To test for statistically significant differences, the log‐transformed values for each parameter of each group were tested using ANOVA followed by Dunnett's test; the 0.003 mg midazolam group without chlorzoxazone served as a control. Statistically significant values (P < .05) are highlighted in boldface type.

3.4. Safety

No serious adverse event and no adverse events of grade 3 or higher occurred. No participant discontinued trial participation because of an adverse event. Overall, the rate of adverse events was low and all adverse events were self‐limiting and resolved. Events considered related to study drug administration were the following: Two participants reported myalgia (1 × grade 1, 1 × grade 2) after taking the highest chlorzoxazone dose. One participant reported dizziness and fatigue (both grade 1) after taking the highest chlorzoxazone dose.

4. DISCUSSION

This is the first trial demonstrating the feasibility of a chlorzoxazone microdosing approach to measure CYP2E1 activity with doses as low as 0.05 mg that result in picomolar plasma concentrations. This is also the first report of pharmacokinetic parameters of chlorzoxazone over a 10 000‐fold dose range. Exposure after a 250‐ and a 500‐mg dose in this trial was comparable to previously published results both with respect to AUC and C_max for a 250‐ or 500‐mg oral dose of chlorzoxazone in healthy volunteers.23, 36, 37, 38, 39 Also, the terminal elimination t_½ is in line with values previously reported (0.9–1.2 hours) in healthy volunteers without interacting comedication.23, 36, 37, 38, 39

When we increased the chlorzoxazone dose from 5 to 50 mg, the dose–exposure relationship became nonlinear. Nonlinear pharmacokinetics of 250‐, 500‐ and 750‐mg oral doses of chlorzoxazone were also previously reported.23, 24 The 5‐ and 50‐mg dose levels were administered as an oral solution and had comparable concentration–time profiles. Dose‐normalised C_max and dose‐normalised AUC were both 2‐fold higher for the 50‐mg dose relative to the 5‐mg dose, while t_½ was similar. Because chlorzoxazone clearance is stable over short (weeks) and longer periods of time (years),24 time‐dependent variability cannot explain the observed differences in chlorzoxazone CL/F in our trial. The Michaelis–Menten constant (K_M) of CYP2E1‐catalyzed hydroxylation of chlorzoxazone is up to 10‐fold lower than the systemic C_max following 250‐ and 500‐mg doses.19, 20 Human intestinal and portal vein concentrations of chlorzoxazone have not been reported, but are expected to be considerably higher during the absorption phase than the systemic C_max. Considering dose‐independent protein binding of 96%,40 saturation of presystemic and possibly also systemic CYP2E1‐dependent chlorzoxazone metabolism is conceivable at doses of 50 mg and higher. Our findings are indicative of a 2‐fold increase of bioavailability of higher chlorzoxazone doses, further supporting the notion of saturable first‐pass. The concentration–time profiles of the 6‐hydroxy metabolite are also indicative of saturation of presystemic metabolism when comparing the 0.05–5 mg doses with the 50–500 mg doses.

In vitro, some8, 18 but not all studies20 suggested that in addition to CYP2E1 also CYP1A2 may be involved in the metabolism of chlorzoxazone and K_M values of these enzymes indicate that the CYP2E1‐mediated contribution to overall chlorzoxazone metabolism steeply increases when chlorzoxazone concentrations exceed 10 μM (1700 ng/mL), i.e. at peak concentrations observed after administration of oral doses ≥50 mg (Table 1). While the contribution of other CYP isozymes in vivo is subject to debate,19 the suggested variable contribution of different enzymes might compromise the reproducibility of the phenotyping procedure using high chlorzoxazone doses and suggests that microdoses might lead to more reproducible results, which will have to be shown in a dedicated study.

Chlorzoxazone oral solution results in less variable plasma concentration–time profiles, especially during the absorption phase as the wider spread of T_max of the oral tablets demonstrates. Also, with doses of 5 mg or smaller, pharmacokinetics are linear and no saturation of (pre)systemic metabolism occurs. It was previously discussed whether a high dose of body weight‐adjusted chlorzoxazone (10 mg/kg) to saturate presystemic metabolism or a lower dose (250 mg) mostly avoiding saturation of presystemic metabolism is more appropriate for CYP2E1 phenotyping.24 Under circumstances of nonlinear pharmacokinetics with dose‐dependent metabolite formation rates, direct comparison of AUC or clearance values obtained with different doses is difficult as will be the interpretation of molar metabolite‐to‐parent ratios from single‐point measurements (e.g. 2 hours after dosing).19 Metabolite formation rate is constant and intrinsic clearance is independent from drug concentration, when the unbound drug concentration is well below K_M. When drug concentration approaches or exceeds K_M, intrinsic clearance becomes concentration dependent. Hence, when CYP‐dependent chlorzoxazone metabolism is partly saturated, net CYP2E1 activity will be misinterpreted particularly if the metabolite‐to‐parent ratio is used as an activity marker. This will only modestly impact interpretation of trial results that reported large effects on CYP2E1 activity such as the coadministration of disulfiram14 or ethanol,12 but the bias could be considerable in trials reporting smaller effects such as the effect of obesity or diabetes mellitus on CYP2E1 activity.19, 20

Because more concentrated oral solutions were unavailable, the 250‐ and 500‐mg doses had to be given as tablets to avoid administration of excessive fluid volumes. Therefore, the dosage form differed from the lower dose levels (oral solution), complicating the interpretation. While C_max is quickly achieved when using an oral solution (T_max ~ 0.5 hours), the absorption of the solid formulation occurs more slowly (T_max ~ 1–2 hours) and C_max is thus lower. The relative bioavailability of an oral chlorzoxazone solution compared to a tablet formulation is unknown for doses of 250 and 500 mg, but a relative bioavailability of 0.7 (tablet vs oral solution) has been suggested for a 750‐mg oral dose comparing data of different cohorts originating from the mid‐1980s to 1990s.40, 41 At that time, this observation remained unexplained, but could well reflect saturation of presystemic metabolism. The more rapidly absorbed oral solution will cause higher portal vein concentrations and may, thus, cause more profound saturation of presystemic metabolism than the more slowly releasing tablets.

Compared to the microdose, the dose‐normalised C_max ratio of the 250‐ and 500‐mg doses is lower due to the slower dissolution and absorption of tablets, and, hence, these values cannot readily be compared. The slower and longer lasting absorption after the tablet formulation contribute to the differences in the shape of the concentration–time profile after the 250‐ and 500‐mg doses, whereas t_½ was similar across all dose levels (Table 2).

We also demonstrated the absence of a drug interaction between midazolam and chlorzoxazone when both marker drugs are administered simultaneously. In contrast, coadministration of 500 mg chlorzoxazone doubled midazolam AUC, which is well in line with the previously reported 1.82‐fold increase in midazolam AUC when milligram doses were included in a 6‐drug cocktail containing chlorzoxazone.26 When administering the cocktail without chlorzoxazone, midazolam pharmacokinetics was unchanged.26 For this reason, phenotyping cocktail protocols either omit a CYP2E1 probe drug42, 43, 44 or the CYP3A probe drug,45 or separate the administration of chlorzoxazone from midazolam,46 or administer midazolam intravenously.47, 48 Because the oral midazolam solution was ingested within 2 min after the chlorzoxazone tablet(s), it is conceivable that midazolam could partially escape from the inhibiting effect of chlorzoxazone, and hence the magnitude of drug interaction would increase if chlorzoxazone is taken well ahead of CYP3A substrates. Still, the simultaneous intake reflects the administration schedule 1 would choose when designing a phenotyping cocktail. Our microdose approach enables the simultaneous administration of oral midazolam and oral chlorzoxazone thus providing a substantial improvement of future multi‐CYP phenotyping cocktails.

4.1. Limitations

The most important limitation is the absence of pharmacokinetic data of a chlorzoxazone microdose with and without a CYP2E1 inhibitor, e.g. disulfiram.14 This will also help answering the question of whether a putative contribution of CYP1A2 to chlorzoxazone metabolism will influence the accuracy of the phenotyping procedure e.g. when used in drug interaction trials. In this context, microdoses could be particularly advantageous because in the low concentration range, chlorzoxazone concentrations will not cause fluctuations of the relative contribution of CYP2E1 to chlorzoxazone metabolism. Therefore, the next necessary step in the validation process of the CYP2E1 phenotyping procedure with chlorzoxazone microdoses is the assessment of its sensitivity for estimating the impact of inhibitors and inducers. Because of nonsaturated CYP, we expect that the microdosed probe will sensitively reveal more subtle changes in enzyme activity with possibly improved reproducibility.

Other limitations of this trial are the small number of participants per dose group during dose escalation and the use of 2 different dosage forms for technical reasons. Because of the large pharmacokinetic variability of therapeutic chlorzoxazone doses, the number of 6 participants per group in the dose escalation part was too low to detect an approximately 2‐fold difference in AUC at an α = 5% level. Hence, GMR and 90% confidence interval of dose‐normalised AUC and C_max were outside bioequivalence borders when comparing the 50‐, 250‐ and 500‐mg dose groups with the 0.1 mg dose group. Due to low solubility and thus large volumes needed to administer high therapeutic doses, we had to switch to the approved tablet formulation when establishing pharmacokinetics of the 250‐ and 500‐mg doses. Therefore, the absorption profiles are not readily comparable and greater emphasis should be given to the results obtained at the 50‐mg dose level.

Finally, metabolite pharmacokinetics from our trial is not readily comparable to other published trials, because of a methodological difference. We measured 6‐hydroxy‐chlorzoxazone in plasma without prior deglucuronidation, whereas previous methods sometimes,26, 45, 48 albeit not always,39 assessed plasma metabolic ratios in samples subjected to glucuronidase treatment before analysis. We chose this methodology to be able to analyse the CYP2E1‐dependent formation rate of 6‐hydroxy metabolite circulating in plasma and not the combined phase I and II reaction and potential impact of renal dysfunction leading to accumulation of chlorzoxazone conjugates.41

In conclusion, we have demonstrated that chlorzoxazone pharmacokinetics is dose‐proportional from 0.05 to 5 mg and found nonlinearity when chlorzoxazone doses were further increased from 50 mg up to 500 mg. A chlorzoxazone microdose did not interact with the CYP3A marker substrate midazolam, enabling the simultaneous administration in a phenotyping cocktail. The lack of saturation of CYP2E1‐dependent presystemic metabolism and the lack of interaction with CYP3A of a chlorzoxazone microdose are possibly both superior characteristics compared to the 250–750 mg doses currently used in phenotyping procedures. The microdose has now to be further tested for its utility as a phenotyping procedure by assessing the impact of induction and inhibition of CYP2E1 on chlorzoxazone microdose pharmacokinetics.

COMPETING INTERESTS

N.H., J.B., Y.S., G.M. and W.E.H. have no competing interests to declare.

CONTRIBUTORS

N.H., G.M. and W.E.H. designed the trial; N.H. and W.E.H. wrote the clinical trial protocol; N.H., A.B., G.M. and W.E.H. conducted the trial; Y.S. and J.B. analysed the samples, N.H., Y.S., A.B., J.B., G.M. and W.E.H. analysed the data; and N.H., Y.S., A.B., J.B., G.M. and W.E.H. wrote the manuscript.

Supporting information

Table S1:

Geometric mean ratio and 90% confidence interval of chlorzoxazone pharmacokinetic parameters when given with and without midazolam.

Table S2: Comparison of pharmacokinetic parameters of 0.1 mg of oral chlorzoxazone of group A (n = 6) and group B (n = 6).

Table S3: Comparison of pharmacokinetic parameters of 0.1 mg of oral chlorzoxazone of female (n = 6) vs male (n = 6) participants.

Hohmann N, Blank A, Burhenne J, Suzuki Y, Mikus G, Haefeli WE. Simultaneous phenotyping of CYP2E1 and CYP3A using oral chlorzoxazone and midazolam microdoses. Br J Clin Pharmacol. 2019;85:2310–2320. 10.1111/bcp.14040