The effect of ketoconazole on the jejunal permeability and CYP3A metabolism of (R/S)-verapamil in humans

Abstract

Aims

The purpose of this human intestinal perfusion study was to investigate the effect of ketoconazole on the jejunal permeability and first-pass metabolism of (R)- and (S)-verapamil in humans.

Methods

A regional single-pass perfusion of the jejunum was performed using a Loc-I-Gut® perfusion tube in six healthy volunteers. Each perfusion lasted for 200 min and was divided into two periods of 100 min each. The inlet concentration of (R/S)-verapamil was 120 mg l⁻¹ in both periods, and ketoconazole was added at 40 mg l⁻¹ in period 2. (R/S)-verapamil was also administered as a short intravenous infusion of 5 mg, over a period of 10 min. The appearance ratios of the CYP3A formed metabolites (R)- and (S)-norverapamil were also estimated in the outlet jejunal perfusate.

Results

The effective jejunal permeability (Peff) of both (R)- and (S)-verapamil was unaffected by the addition of ketoconazole in period 2 suggesting that ketoconazole had no effect on the P-glycoprotein mediated efflux. However, the appearance ratio of both (R)- and (S)-norverapamil in the outlet jejunal perfusate decreased in the presence of ketoconazole. The rate of absorption into plasma of (R)- and (S)-verapamil increased despite the low dose of ketoconazole added, indicating an inhibition of the gut wall metabolism of (R/S)-verapamil by ketoconazole.

Conclusions

Ketoconazole did not affect the jejunal Peff of (R/S)-verapamil, but it did increase the overall transport into the systemic circulation (bioavailability), probably by inhibition of the gut wall metabolism of verapamil. This might be due to ketoconazole being less potent as an inhibitor of P-glycoprotein than of CYP3A4 in vivo in humans.

Introduction

After oral dosing, the first-pass metabolism of high extraction drugs in the liver demonstrates large interindividual variation. Recently, drugs such as midazolam, cyclosporin, felodipine and verapamil were reported to undergo a significant portion of the first-pass metabolism in the enterocytes [1–4]. These drugs are mainly metabolized by the CYP3 A4 enzymes which are located in the small intestine and the liver [5]. It was recently demonstrated that the CYP3 A4 enzymes expressed in the liver and the intestine are identical [6]. The CYP3 A iso-enzymes contribute to approximately 20% of the total amount of cytochrome P450 in the liver and up to 70% of the amount expressed in the enterocytes of the intestine [5]. For midazolam, the extraction in the gut and liver has been reported to be 44±14 and 43±24%, respectively [7]. This was later confirmed by measuring the midazolam extraction (43±18%) during the anhepatic phase in liver transplant patients [2]. The total amount and activity of CYP3 A (midazolam to hydroxy-midazolam) have been reported for the human small intestine with the highest expression in the jejunum with an amount of 38 nmol and an activity of 148 ml min⁻¹ of [8]. For comparison, the corresponding values estimated for the liver were 5490 nmol and 158 000 ml min⁻¹, respectively [8]. Even if the amount of CYP3 A enzymes that are expressed in the intestine is low compared with the liver, it has been suggested that the relativly high extraction in the intestine is explained by lack of influence of protein binding during the transport across the enterocyte [8].

Considerable attention has been focused on the importance of the expression of P-glycoprotein (Pgp) in the intestinal wall and the impact of Pgp on intestinal absorption and bioavailability [9–11]. The role of Pgp in absorption was shown when the bioavailability of drugs such as paclitaxel, indinavir, nelfinavir and saquinavir was demonstrated to be higher in mdr [−/−] mice compared with wild-type mice [12, 13]. The bioavailability of cyclosporin in humans is also reported to partly be influenced by the efflux transport mediated by intestinal Pgp [14]. In an earlier jejunal perfusion study in humans we reported, that the concentration dependent effective permeability (Peff) of both enantiomers of verapamil could be due to a saturation of an efflux mechanism (presumably Pgp) [15]. Therefore, more mechanistic transport and metabolic studies in humans are needed to clarify the role of this important factor for the variability in drug response.

Ketoconazole is a modulator of Pgp as well as a strong inhibitor of CYP3A enzymes [16, 17]. Verapamil (log D 2.7, octanol/H₂O; pH 7.4; MW 455 Da) has been reported to have a rather high affinity to P-glycoprotein(s) [18]. In the present study we used (R/S)-verapamil with and without coadministration of ketoconazole to investigate the local permeability and metabolism of both enantiomers of verapamil in the jejunum. A single-pass perfusion technique (Loc-I-Gut®) was used for this in healthy volunteers. The disappearance of drugs from the perfused jejunal segment can be used to calculate the effective jejunal permeability (Peff), a parameter that has been shown to predict the extent of intestinal absorption of drugs with high precision [19]. To estimate the effects of Pgp and CYP3A4 inhibition on the overall bioavailability of each enantiomer of verapamil, we collected plasma samples during and after the jejunal perfusion.

The aim of this human study was to examine the in vivo effects of ketoconazole on the intestinal permeability and the presystemic metabolism of each enantiomer of (R/S)-verapamil.

Methods

Subjects

The study was approved by the Ethics and Isotope Committees of the Medical Faculty, Uppsala University, Sweden. Six healthy volunteers participated (five male and one female) aged between 22 and 28 years, and weighing between 63 and 85 kg. The subjects were given written information about the study and all of them gave their informed consent to their participation. Prior to the study all participants underwent a full clinical examination and all had normal clinical and laboratory findings in serum (S) and blood (B) (S-creatinine, S-ASAT, S-ALAT, S-ALP, S-Potassium, S-Sodium, S-Bilirubin, B-Erythrocytes, B-Haemoglobin, B-Trombocytes, B-Leukocytes and HIV). None of the participants received any medication before or during the perfusion experiment other than the drugs under investigation. The study was performed at the clinical research department of the University Hospital in Uppsala, Sweden.

Study medication

(R/S)-verapamil for oral human use was a kind gift from Knoll AG, Darnstadt, Germany. Ketoconazole was purchased from the pharmacy at the University Hospital, Uppsala. Antipyrine was supplied by Astra Läkemedel AB, Södertälje, Sweden, and was used as an absorption marker. The solution for the intravenous administration containing 2.5 mg ml⁻¹ of racemic verapamil (Isoptin®, Knoll AG, Germany) was purchased from the local pharmacy.

The perfusion solution consisted of antipyrine 10 mg l⁻¹, potassium chloride 5.4 mm, sodium chloride 30 mm, mannitol 35 mm, d-glucose 10 mm, PEG 4000 1.0 g l⁻¹, all dissolved in a 70-mm phosphate buffer with pH 6.5. Each perfusion lasted for 200 min and was divided into two periods (P1 and P2) each of 100 min. The inlet concentration of (R/S)-verapamil was 120 mg l⁻¹ in both periods. In the second period (P2) ketoconazole was added to the perfusion solution at a concentration of 40 mg l⁻¹. Each subject was only exposed to (R/S)-verapamil in the first period (P1) in order to avoid carry over effects by the inhibition of ketoconazole. Polyethylene glycol labelled with ¹⁴C ([¹⁴C]-PEG 4000) was purchased from Amersham Laboratories, Buckinghamshire, England and added to the perfusion solution as a nonabsorbable volume marker with an activity of 2.5 μCi l⁻¹.

In vivo perfusion of the human jejunum

After an overnight fast of 10 h a regional single-pass perfusion of the proximal jejunum was performed using a Loc-I-Gut® perfusion tube (Synectics Medical, Stockholm, Sweden). Having applied a local anaesthesia to the oesophagus with lignocaine spray, the tube was introduced through the mouth. During insertion there was a Teflon coated guidewire inside the instrument to facilitate the passage of the tube into the intestine. The position of the tube was checked by fluoroscopy and the perfused segment was located in the proximal part of the jejunum. Along with the Loc-I-Gut® instrument, another tube was positioned in the stomach for drainage of gastric juices during the experiment (Salem sump tube, Sherwood Medical, U.K.). Once the perfusion tube was in place, the two balloons were inflated with approximately 26–30 ml of air creating a 10 cm long segment. A vacuum pump was connected to the proximal drainage channel of the perfusion tube to drain any intestinal fluid above the perfused segment (Ameda suction pump type 23, Ameda AG, Zug, Switzerland). The segment was then rinsed with isotonic saline (37° C) for at least 20 min. When stable perfusion conditions were achieved, the perfusion solution (37° C) was pumped into the jejunal segment at a flow rate of 2.0 ml min⁻¹ using a calibrated syringe pump (model 355, Sage Instr. Orion Research Inc. Cambridge, MA, USA). A more extensive description of this perfusion technique is published elsewhere [20, 21]. The perfusate leaving the segment was quantitatively collected on ice in 10 min fractions and immediately frozen pending analysis. When the perfusion was terminated after 200 min, the jejunal segment was rinsed with 120 ml of saline, and the Loc-I-Gut® instrument was removed. Venous blood samples were collected in one of the forearms immediately before the start of the perfusion and after 20, 40, 60, 80, 100, 120, 140, 160, 180, 200, 240, 300, 360, 480, and 720 min. The blood samples were centrifuged at 3000 rev min⁻¹ for 10 min and the plasma was frozen at −20° C pending analysis. A light meal was served immediately after the Loc-I-Gut® instrument was removed, and again just after the collection of the 480 min blood sample.

Intravenous infusion

After an overnight fast of at least 10 h, each volunteer received a 5 mg dose of (R/S)-verapamil given as a 10 min intravenous infusion in a forearm vein. Venous blood samples were collected in the other arm just before the infusion and after 9, 15, 25, 40, 60, 120, 180, 240, 360, 480 and 720 min. Light meals were served 200 min and 480 min after the start of the infusion. The blood samples were centrifuged at 3000 rev min⁻¹ for 10 min and the plasma was immediately frozen (−20° C) pending analysis.

Stability and adsorption test of (R/S)-verapamil

The stability and adsorption of (R/S)-verapamil have previously been tested by in vitro perfusion using the Loc-I-Gut® tube inside a glass tube at 37° C for 200 min [15]. (R/S)-verapamil was stable in the perfusion solution at 37° C for at least 200 min. Approximately 5% of (R/S)-verapamil was adsorbed to the perfusion instrument during the in vitro perfusion [15]. The adsorption of (R/S)-norverapamil was assumed to be same due to the structural similarities with (R/S)-verapamil.

Analytical methods

Perfusate An enantioselective h.p.l.c.-method for quantifying (R/S)-verapamil and (R/S)-norverapamil was performed on a Chiral AGP-column (4×150 mm, Chrometech, Stockholm, Sweden.), with a Chiral AGP precolumn (3×10 mm) [22]. The pump was a Shimatzu LC-9 A (Kyoto, Japan) and the flow rate applied was 1.0 ml min⁻¹. Temperatures above room temperature (22° C) were achieved by heating the column and the mobile phase in a heat bath (Grant Instrument Ltd, type JB1, Cambridge, England). The detection was made with a Jasco FP-920 fluorescence detector (Tokyo, Japan) with an excitation wavelength of 232 nm and an emission wavelength of 310 nm. The analytical condition used for the analysis of the perfusate was a phosphate buffer with an ionic strength of 0.01 and a pH of 7.6 with 22% (v/v) of acetonitrile at 30° C. The perfusate samples were diluted in the mobile phase and 50.1 μl was injected on the column. The limits of quantification (LOQ) (±s.d.) for both (R)-verapamil and (S)-verapamil were 5.5±0.3 ng ml⁻¹ in the perfusate. The LOQ (±s.d.) for (R)- and (S)-norverapamil in the perfusate were 2.9±0.2 ng ml⁻¹. More details of the enatioselective h.p.l.c. method are available elsewhere [22]. The concentration of antipyrine in the perfusate and perfusion solution was analysed by h.p.l.c. and u.v. detection using a previously validated method [21, 23]. The concentration of the volume marker [¹⁴C-PEG] 4000 was determined by liquid scintillation counting (Mark III, Searle Analytic Inc., Des Plaines, Illinois, USA). The perfusate concentration of d-glucose was analysed using a multichannel analyser (Technicon DAX®, Bayer Diagnostics) at Calab AB, Stockholm, Sweden.

Plasma

Another enantioselective h.p.l.c.-method was used for the analysis of (R/S)-verapamil in plasma, where the metabolites of (R/S)-verapamil reacts with acetic anhydride during the sample preparation [24, 25]. The h.p.l.c. system was the same as for the perfusate samples but the mobile phase consisted of a phosphate buffer with an ionic strength of 0.01 and a pH of 6.6 with 15% (v/v) acetonitrile at 30° C. 1.0 ml of plasma, 0.200 ml of 0.1 m NaOH and 25 μl of internal standard ((+)-methoxyverapamil, 2 μg ml⁻¹ in water) were mixed with 6.0 ml of diethylether for 5 min in a glass tube, and centrifuged for 10 min at 4500 rev min⁻¹. The organic phase was transferred to a new tube and mixed with 100 μl of acetic anhydride for 5 min. The organic phase was the mixed with 150 μl of 0.1 m HCl for 5 min and then centrifuged for 10 min at 4500 rev min⁻¹. The organic phase was removed and the remaining acid was alkalinized with 300 μl of 0.3 m NaOH and extracted once more with 6 ml diethylether (5 min of mixing, 10 min centrifugation at 5000 rev min⁻¹). The organic phase was transfered to a new tube and mixed with 150 μl of 0.1 m HCl, mixed for 5 min and centrifuged for 10 min at 4500 rev min⁻¹. The organic phase was removed and the remaining acid was neutralized with 0.1 m NaOH to a pH between 3 and 5 and 100 μl were injected on the column. The limits of quantification (LOQ)(±s.d.) for the (R)-verapamil and the (S)-verapamil ml in plasma were 0.57±0.07 ng ml⁻¹.

Data analysis

Perfusion All calculations of the effective intestinal permeability (Peff), the net water flux (NWF), the fraction absorbed (fa), the appearance ratio (Ar) and the ratio between the enantiomers (Er) were made from the last five steady state concentrations in the outlet jejunal perfusate at the end of periods one (P1) (50–100 min) and two (P2) (150–200 min). Each sample represents the mean concentration of the aliquots collected for each 10 min interval. The net water flux per cm in the isolated jejunal segment was calculated according to equation 1 for each sample:

1

where PEGin and PEGout are the concentrations of [¹⁴C]-PEG 4000 (dpm/ml) entering and leaving the segment, respectively. Qin is the flow rate of the perfusion solution and L is the length of the perfused jejunal segment (10 cm). The concentration of each compound in the perfusate leaving the intestine was corrected for water flux before fa, Peff and Ar were calculated. The amount that disappeared during the single-passage through the jejunal segment was assumed to have been absorbed (fa):

2

where Cin and Cout are the inlet and outlet concentrations of each compound, respectively. The intestinal effective permeability (Peff) was calculated according to a well-mixed tank model as shown in equation 3 [21]:

3

The surface of the cylinder (2πrL) of the jejunal segment was calculated using the intestinal radius (r = 1.75 cm) and the length (L=10 cm) of the segment. That this estimate of the human jejunal radius is accurate for the perfused segment has been shown by a recently performed fluoroscopic investigation (Knutson et al. personal communication). The appearance ratio (Ar) was calculated for each of the enantiomers of norverapamil (formed by the CYP3A4 route) as shown in equation 4:

4

where Cout (norverapamil) is the concentration of (R)- or S-norverapamil in the perfusate leaving the segment, and Cin (verapamil) is the concentration of the corresponding enantiomer of verapamil entering the segment (on a molar basis). The fraction absorbed (fa) during the perfusion was obtained from equation 2.

Plasma

The total clearance (CL) for each enantiomer of verapamil after intravenous administration was calculated according to equation 5:

5

where AUCi.v. is the area under the plasma concentration-time curve from time zero to infinity after the intravenous infusion. The AUC was calculated using the logarithmic trapezoidal method up to the last observed time point. The area to infinity beyond the last sample was estimated with use of linear regression analysis of the logarithmic concentration vs time for three to five of the last plasma samples. The total blood clearance (CLb) can be estimated by multiplying CL by the ratio of the plasma (Cp) and blood concentrations (Cb) (equation 6):

6

In equation 6 we used previously reported values of the ratio (Cp/Cb) for (R)- and (S)-verapamil of 1.6 and 1.4, respectively [26]. It is assumed that the liver is the main organ for the elimination of (R/S)-verapamil after intravenous administration since the renal elimination is negligible [26, 27]. The hepatic extraction (Eh) of (R/S)-verapamil can then be estimated by dividing the CLb by the hepatic blood flow (Qh)(1500 ml min⁻¹ 70 kg⁻¹ body weight) (equation 7):

7

The bioavailability (F) was calculated using the dose corrected AUC values for the perfusion and infusion doses of each enantiomer of verapamil according to equation 8:

8

where Dose abs is the dose that has disappeared during the jejunal perfusion. This means that F is only influenced by gut wall and hepatic extraction in this particular situation.

Equation 8 can be rewritten to give equation 9, where the bioavailability is dependent on the extent of absorption (Fa), extraction in the gut wall (Eg) and the extraction over the liver (Eh).

9

Fa was determined and incorporated (Doseabs) which means that equation 9 can be rearranged to give the (Eg) extraction of (R/S)-verapamil in the gut wall (equation 10).

10

The extraction in the gut (Eg) was estimated from equation 10 assuming that the extraction in the liver (Eh) was the same after intravenous and oral administration of (R/S)-verapamil. However, it has been reported that the Eg after intravenous administration is substantially lower for midazolam (another CYP3 A substrate) than after oral administration [2].

Exponential functions were fitted to the observed plasma concentration-time profiles for (R)- and (S)-verapamil after the intravenous infusion with the non linear regression computer program MINIM [28]. Discrimination between the pharmacokinetic models was made by standard goodness-of-fit criteria such as residual trends, visual inspection, standard errors of the parameter estimates and the Akaike criterion. These estimates were used to calculate the absorption rate during the perfusion experiment by the method of deconvolution [29]. The half-life (t_1/2) was calculated by a log-linear regression analysis of the three to four data points on the plasma concentration-time curve after the intravenous infusion and jejunal perfusion.

Statistics

The difference between perfusion periods and enantiomers was evaluated using a two tailed paired Student's t-test. Variability is expressed as the standard deviation (s.d.).

Results

Perfusate data

The absorption parameters were calculated from the disappearance of each compound from the jejunal segment during a single-pass perfusion. The results are given in Table 1. There were no statistical differences between period 1 and 2 for the effective jejunal permeability (Peff) or the fraction absorbed (fa) for (R/S)-verapamil, antipyrine or d-glucose (Table 1 and Figure 1a,b). The addition of the Pgp modulator ketoconazole had no effect on the jejunal permeability (Peff) of (R/S)-verapamil (Table 1). The jejunal Peff was high for both (R)- and (S)-verapamil and there was no difference between the two enantiomers of verapamil (Table 1).

Table 1.

The effective human jejunal permeability (Peff) and the fraction absorbed (fa) for (R/S)-verapamil, antipyrine and d-glucose in the two periods of the perfusion (mean±s.d.). In period 2, ketoconazole was added at 40 mg l⁻¹.

Figure 1.

Individual values of the human jejunal effective permeability (Peff) for a) (R)- and b) (S)-verapamil at a luminal concentration of 120 mg l⁻¹ of (R/S)-verapamil (without ketokonazole)(0–100 min) and with the addition of 40 mg l⁻¹ ketoconazole (100–200 min).

The appearance ratios (Ar) were stereospecific. The Ar is a metabolic variable calculated from the outlet perfusate concentrations of the CYP3 A formed metabolite (R/S)-norverapamil. Higher perfusate concentrations were found of (S)-than (R)-norverapamil in both periods (P<0.001) (Table 2 and Figure 2a,b). The Ar-values for both (R)- and (S)-verapamil were lower in period 2 (P<0.001), suggesting that the addition of ketoconazole decreases the jejunal metabolism of both (R)- and (S)-verapamil (Table 2 and Figure 2a,b). Furthermore, the ratio between the Ar for (R)- and (S)-norverapamil was lower in period 2 than in period 1 (P<0.05) (Table 2).

Table 2.

The metabolic variables based on the appearance of the metabolite (R/S)-norverapamil, and the net water flux (NWF) in the perfusate leaving the human jejunal segment during perfusion (mean±s.d.).

Figure 2.

Individual values of the appearance ratio (Ar) in the human jejunum for a) (R)- and b) (S)-norverapamil at a luminal concentration of 120 mg l⁻¹ of (R/S)-verapamil (without ketokonazole)(0–100 min) and with the addition of 40 mg l⁻¹ of ketoconazole (100–200 min).

Plasma data

The individual concentrations of (R)- and (S)-verapamil in plasma after intravenous administration and jejunal perfusion are shown in Figures 3a,b and 4a,b, respectively. The clearance (CL), half-life (t_1/2) and AUC for (R)- and (S)-verapamil from the intravenous infusion, were all significantly different between the two enantiomers of verapamil (Table 3). The AUC was lower, clearance higher and half-life longer for (S)-than (R)-verapamil (Table 3 and Figure 3a,b). The blood clearance (CLb) is based on previously reported values of the in vitro plasma/blood ratios for (R)- and (S)-verapamil of 1.4 and 1.6, respectively [26]. The values of CLb were used to calculate the liver extraction ratios (Eh) of each enantiomers [26]. The Eh was higher for S than (R)-verapamil (Table 3).

Figure 3.

Individual concentrations of a) (R)- and b) (S)-verapamil in plasma after intravenous infusion of 5 mg racemic verapamil over 10 min in six healthy volunteers.

Figure 4.

Individual concentrations of a) (R)- and b) (S)-verapamil in plasma after a jejunal perfusion of racemic verapamil for 200 min in six healthy volunteers.

Table 3.

The pharmacokinetic variables from the plasma concentration-time profiles of (R)- and (S)-verapamil after oral (perfusion) and intravenous administration in humans (mean±s.d.).

The bioavailability (F) was 5.3±4.6 and 18.2±8.2% (P<0.01) for (S)- and (R)-verapamil, respectively (Table 3 and Figure 4a,b). A number of plasma samples of (S)-verapamil taken under and after the jejunal perfusion were under the limit of quantification in some individuals, and the half-life and absorption rates (from the deconvolution of plasma concentrations) were therefore estimated in only three and four subjects, respectively (Table 3 and Figures 4b and 5b). The liver extractions (Eh) were 0.63±0.14 and 0.79±0.16 for (R)- and (S)-verapamil, respectively (Table 3). The estimated extractions in the gut wall (Eg) (from equation 10) were 0.49±0.18 and 0.68±0.19 for (R)- and (S)-verapamil, respectively (Table 3). The total first-pass extraction (calculated as 1-F) estimated from the jejunal perfusion was 0.82±0.08 and 0.94±0.05 for (R)- and (S)-verapamil, respectively (Table 3). The absorption rate of (R)-verapamil into the systemic circulation was 10.4±3.8 and 15±5.1 μg min⁻¹ (P<0.05) in periods 1 and 2, respectively (Table 3). The same variable for (S)-verapamil was 2.65±1.14 and 5.12±2.57 μg min⁻¹ in periods 1 and 2, respectively (Table 3). This clearly demonstrates that even though the dose of the inhibitor was very low (8 mg), ketoconazole affected the systemic availability of verapamil, at least for the (R)-enantiomer.

Figure 5.

Individual values of the absorption rate of a) (R)- and b) (S)-verapamil into plasma after a jejunal perfusion in six healthy subjects with (R/S)-verapamil at a luminal concentration of 120 mg l⁻¹ (without ketokonazole)(0–100 min) and with the addition of 40 mg l⁻¹ of ketoconazole (100–200 min).

Discussion

The main results from the present mechanistic study in humans clearly demonstrate that the potent CYP3A4 inhibitor ketoconazole affects the presystemic metabolism of both (R)- and (S)-verapamil, however, to different degrees. In addition, on the basis of the unchanged Peff of both (R)- and (S)-verapamil in periods 1 and 2 we have found that ketoconazole at 40 mg l⁻¹ in the jejunal lumen does not influence P-glycoprotein mediated efflux of (R/S)-verapamil in the human enterocyte. This suggests that ketoconazole is a significantly more potent inhibitor of CYP3A4 than Pgp in human jejunum.

We have previously reported a concentration dependent decrease in the appearance ratio (Ar) of (R/S)-norverapamil in the outlet jejunal perfusate collected in humans [15]. As the perfusate concentrations of (R/S)-verapamil were very low, 4.0 and 40 mg l⁻¹, we suggested that the decreased Ar was due to an easily saturated metabolism located inside the human enterocyte [15]. The decreased Ar for (R/S)-norverapamil seen in the present study was interpretated as inhibited gut wall metabolism, since the concentration of ketoconazole in the jejunal perfusion solution was very low, 40 mg l⁻¹ and the totally exposed dose was 8 mg during 100 min. It seems rather unlikely that such a low concentration and dose rate of the inhibitor should influence the liver metabolism to any major extent. Consequently, it suggests that the increased absorption rate of (R/S)-verapamil into the systemic circulation (from deconvolution of plasma data) during period 2 is due to an inhibited presystemic CYP3A4 metabolism in the human enterocyte. The overall first-pass metabolism was high for both (R)- and (S)-verapamil, respectively. When the extraction values in the gut wall and the liver were separated, it was clear that they were of the same magnitude. This regional perfusion study confirms earlier reports of a significant contribution of the gut wall to the first-pass extraction of CYP3A substrates.

The absence of an effect of ketoconazole on the jejunal Peff for R and (S)-verapamil, and the fact that the intestinal Peff for (R/S)-verapamil is high, suggests that Pgp has a limited contribution to the presystemic extraction of (R/S)-verapamil in the gut wall.

This is in accordance with the previously reported concentration dependent Peff of verapamil in humans probably mediated by an efflux of Pgp. However, the Peff values for (R)- and (S)-verapamil at all reported concentrations are still high enough to predict a complete absorption of verapamil from the human intestine [15, 30]. The increase in Peff at 400 mg l⁻¹ also indicates that we have not measured the effect of ketoconazole at a concentration of (R/S)-verapamil that completely saturates the P-glycoprotein mediated efflux [15, 30].

The bioavailabilities (F) of (R)- and (S)-verapamil were 18.2 and 5.3%, respectively. This is considerably lower than in a previous report [27]. The estimations of the pharmacokinetic parameters such as clearance (CL) and hepatic extraction (Eh) correlates well with previously reported results [4]. However, the estimation of the gut wall extraction (Eg) was approximately 20% higher than reported by others for both (R)- and (S)-verapamil, which can explain the lower bioavailability [4]. These values from the literature are based on repeated oral dosing with considerably higher doses (240 mg day⁻¹), than our single dose of approximately 28 mg racemic verapamil absorbed during 200 min of jejunal perfusion [4].

Rifampicin has been reported to induce the metabolism of verapamil in both the gut and the liver in humans, with dramatic consequences for the bioavailability of both (R)- and (S)-verapamil [4]. Rifampicin has also been reported to induce the expression of Pgp in the enterocytes of humans [31]. This induction of Pgp was not addressed in the induction study with rifampicin, where the increased gut wall extraction of verapamil instead was explained by an increased metabolic capacity of the enterocytes [4]. It is possible that the increased gut wall extraction might be at least partially explained by an increased efflux by the Pgp in the intestinal wall (or increased apical recycling), and not only by induction of CYP3A4 enzymes in the enterocytes.

In the existing literature, the relative contributions of intestinal and hepatic CYP3A4 to the presystemic metabolism of orally given CYP3A4 substrates differ. For instance, cyclosporin A seems to be mostly dependent on liver CYP3A4, whereas the intestinal CYP3A4 metabolism is crucial for the first-pass extraction of felodipine [14, 32]. These observations may be due to factors other than catalytic activity: (1) other CYP3A enzymes may be differently expressed in the liver and in the intestine (2) the role of Pgp as a diffusion barrier and/or apical recycling mechanism. A third possibility is that as yet unrecognized efflux mechanisms contributes to the variability. The important contribution of the gut wall metabolism to the first-pass extraction of (R/S)-verapamil was also shown in the present regional perfusion and thereby confirms earlier studies. In another perfusion study with fluvastatin (a CYP2C9 substrate) the contribution by the intestine to the first-pass extraction was negligible which confirms that it is mainly CYP3A substrates that are metabolized in the intestinal wall [33].

Ketoconazole has been reported to increase the bioavailability of tacrolimus from about 14–30% [34]. It was suggested that this increase might be caused by a local inhibition of the tacrolimus variables for tacrolimus were obtained after 5 days of repeated administration of ketoconazole. On the basis of the acute effects observed in the present jejunal perfusion study, it seems likely that inhibition of the gut wall metabolism should be the predominant effect. However, we cannot rule out the possibility of a time-lag in order for ketoconazole to inhibit Pgp activity. Assuming that the major part of drug absorption occurs on the tip of the villi, and that enterocytes have a turn-over rate of 5 days, it is possible that ketoconazole mediates its Pgp inhibitory effect by acting on the proliferation of Pgp in immature enterocytes. However, the opposite effect is also plausible since it has been shown that ketoconazole can induce mdr1a mRNA in rats [31].

Conclusions

This study shows that acute exposure of ketoconazole does not affect the jejunal permeability of (R/S)-verapamil. This suggests that ketoconazole is a significantly more potent inhibitor of CYP3A4 than of Pgp in the human intestine. It still remains to investigate whether subjects/patients using verapamil and who are exposed to a pure and more potent Pgp inhibitor, will have a changed or unaffected first-pass extraction. From the present study it seems likely that the first-pass extraction and bioavailability should be unaffected for (R/S)-verapamil. However, a lag-time for the inhibition of Pgp activity by ketoconazole may exist, and therefore we cannot be completely conclusive. In spite of this restriction, the present study suggests that the role of Pgp have been overestimated in the in vivo pharmacokinetics after oral administration of certain drugs (such as verapamil). Further support for our observation is that midazolam undergoes a high extraction in the gut wall without being a Pgp substrate, and that grapefruit juice causes major changes in the gut wall extraction of drugs that are both Pgp and CYP3A substrates without affecting the Pgp [32, 35]. Our study clearly demonstrates the usefulness of this perfusion technique when drug–drug interactions in the presystemic gut wall metabolism is investigated.