Materials and Methods

Chemicals.

MDZ (Dormicam, 5 mg/ml solution for intravenous injection) was obtained from Astellas Pharma Inc. (Tokyo, Japan). Tacrolimus (TAC), which was synthesized at our laboratory, was used. Lithium carbonate (Li) was purchased from Kanto Chemical Co., Inc. (Tokyo, Japan). Hydrochlorothiazide (HT), verapamil (VER), propranolol (PRO), and amitriptyline (AMI) were purchased from Wako Pure Chemicals (Osaka, Japan). Dexamethasone (DEX), nifedipine (NIF), quinidine (QID), timolol (TIM), and ibuprofen (IBU) were purchased from Sigma-Aldrich (St. Louis, MO). Liver and intestine microsomes from humans and cynomolgus monkeys were purchased from XenoTech, LLC (Lenexa, KS). All other reagents and solvents were commercial products of analytical grade.

Animals.

Male cynomolgus monkeys (Shin Nippon Biomedical Laboratories, Ltd., Kagoshima, Japan, and Astellas Research Technology, Osaka, Japan) weighing approximately 5 kg were used. The animal experiment was conducted according to the ethical rules of each company.

Selected Drugs and Categorization.

We allocated the 13 drugs into five categories (Type A–E), according to their pharmacokinetic properties in humans, as follows: membrane permeability, P450 isozyme selectivity, and P-gp affinity (Yu, 1999; Kivist et al., 2004; Yang et al., 2006) (Table 1).

Type A.

The drugs categorized as Type A are indicator drugs that undergo no metabolism in humans and are not P-gp substrates. For each of these, almost all of the absorbed drug is excreted into urine as the unchanged form. Li, which has a high F in humans (94.5%) (Arancibia et al., 1986), and HT, which has a moderate F in humans (60.2%) (Patel et al., 1984), were assigned to this category.

Type B.

The drugs categorized as Type B are CYP3A4 substrates, and they have very weak, if any, affinity for P-gp. DEX, which has a high F in humans (81.4%) (Duggan et al., 1975), NIF, and MDZ, which have a moderate F in humans [41.2% (Holtbecker et al., 1996) and 30.0% (Thummel et al., 1996), respectively] were assigned to this category.

Type C.

The drugs categorized as Type C are substrates of both CYP3A4 and P-gp. QID, which has a high F in humans (79.5%) (Greenblatt et al., 1977), as well as TAC and VER, which have a moderate F in humans [23.3% (Moller et al., 1999) and 18.0% (McAllister and Kirsten, 1982), respectively], were assigned to this category.

Type D.

Digoxin (DIG), which is substrate of P-gp but not CYP3A4, was categorized as Type D. DIG has a high F in humans (65.3%) (Hinderling and Hartmann, 1991) and undergoes almost no metabolism in the human body, i.e., it undergoes only P-gp efflux during the absorption process in the intestine.

Type E.

The drugs categorized as Type E are mainly metabolized by the P450 isozyme (except CYP3A4) and have very weak, if any, affinity for P-gp. IBU and TIM, which have a high F in humans [100% (Martin et al., 1990) and 61.0% (Wilson et al., 1982), respectively], as well as AMI and PRO, which have a moderate F in humans [47.7% (Schulz et al., 1983) and 29.0% (Borgstrom et al., 1981), respectively], were assigned to this category. See Table 1 for P450 isozymes that contribute to each drug metabolism.

Pharmacokinetic Study in Cynomolgus Monkeys.

Intravenous and oral administrations were performed with a washout period of at least 7 days between each type of administration. Animals were fasted for approximately 17 h before dosing. Blood samples were collected from the antebrachial vein, kept in an ice-water bath, and then centrifuged at 10,000 rpm for 1 min at 4°C. The plasma samples were kept in a deep freezer (approximately −20°C) until analysis. The experimental conditions for the pharmacokinetic studies, including doses, dosing solution, dosing volume, and sampling time for each drug, are shown in Table 2. Values obtained from the literature were used as the pharmacokinetic parameter values for all selected drugs in humans as well as those for MDZ in cynomolgus monkeys.

Measurement of Model Compounds Plasma Concentration in Cynomolgus Monkeys.

The concentrations of model drugs in cynomolgus monkey plasma were determined by using atomic absorption, enzyme immunoassay analysis, or high-performance liquid chromatography (LC) coupled with tandem mass spectrometry (MS/MS) with sample pretreatment.

Atomic absorption method: Lithium.

The lithium level in the plasma was determined by using atomic absorption in accordance with the method of Pybus and Bowers (1970).

Enzyme immunoassay analysis: Dexamethasone and tacrolimus.

The DEX level in the plasma and the TAC level in the blood were determined by using enzyme immunoassay. After extraction (see below), an aliquot was used as the sample for analysis by enzyme immunoassay (Tamura et al., 1987).

A 50-μl aliquot of plasma was buffered with 1% skim milk/phosphate-buffered saline. After the addition of 1 ml of distilled water, the mixture was extracted with 5 ml of diethyl ether, and the solvent was removed under a stream of nitrogen gas. The residue was then dissolved in 250 μl of skim milk (1%)/phosphate-buffered saline.

LC-MS/MS analysis.

The plasma concentrations of all other drugs were determined by using LC-MS/MS. The LC-system comprised a LC-VP/LC-10A series (Shimadzu, Kyoto, Japan) or HP-1100 series HPLC (Agilent Technologies, Santa Clara, CA). The MS/MS experiments were conducted by using API-2000 or API-3000 LC/MS/MS systems (Applied Biosystems, Foster, CA). The details of the LC-MS/MS conditions, including the machines and columns used for each drug, are shown in Table 3.

Hydrochlorothiazide.

A 200-μl aliquot of plasma was buffered with 500 μl of phosphate buffer (10 mM) adjusted to pH 3.0. After adding 100 μl of acetonitrile and 20 μl of internal standard solution (1 μg/ml diclofenac in 50% acetonitrile), the mixture was extracted with 4 ml of ethyl acetate, and the solvent was removed under a stream of nitrogen gas. Then, the residue was dissolved in 100 μl of mobile phase, and a 40-μl aliquot was injected into the LC-MS/MS (molecular>product: m/z = 296 > 269 [M+H]−).

Nifedipine.

A 50-μl aliquot of plasma, 50 μl of acetonitrile (50%), and 100 μl of internal standard solution (1 μg/ml of in-house compound A in acetonitrile) were mixed well and then centrifuged to remove precipitated protein. The supernatant (100 μl) was then decanted, and 30 μl was injected into the LC-MS/MS (molecular>product: m/z = 347 > 315 [M+H]+).

Quinidine, Verapamil, Propranolol, Amitriptyline, and Timolol.

A 200-μl aliquot of plasma was buffered with 500 μl of saturated sodium bicarbonate solution. After the addition of 50 μl of acetonitrile and 50 μl of internal standard solution (1 μg/ml of in-house compound B in 50% acetonitrile), the mixture was extracted with 3 ml of tert-butyl methyl ether, after which the solvent was removed under a stream of nitrogen gas. The residue was then dissolved in 200 μl of mobile phase, and a 20-μl aliquot was injected into the LC-MS/MS (molecular>product: QID m/z = 325 > 307 [M+H]+, VER m/z = 455 > 165 [M+H]+, TIM m/z = 317 > 261 [M+H]+, AMI m/z = 278 > 117 [M+H]+, PRO m/z = 260 > 116 [M+H]+).

Digoxin.

A 200-μl aliquot of plasma was buffered with 500 μl of phosphate buffer (10 mM) adjusted to pH 3.0. After the addition of 100 μl of acetonitrile and 50 μl of internal standard solution (1 μg/ml digitoxin in 50% acetonitrile), the mixture was extracted with 3 ml of ethyl acetate, and the solvent was removed under a stream of nitrogen gas. The residue was then dissolved in 100 μl of mobile phase, after which a 20-μl aliquot was injected into the LC-MS/MS (molecular>product: m/z = 798 > 391 [M+NH4]+).

Ibuprofen.

A 200-μl aliquot of plasma was buffered with 500 μl of phosphoric acid (5 mM). After the addition of 50 μl of acetonitrile and 50 μl of internal standard solution (1 μg/ml of diclofenac in 50% acetonitrile), the mixture was extracted with 3 ml of tert-butyl methyl ether, and the solvent was removed under a stream of nitrogen gas. The residue was then dissolved in 200 μl of mobile phase, and a 20-μl aliquot was injected into the LC-MS/MS (molecular>product: m/z = 205 > 161, [M+H]−).

Blood-to-Plasma Concentration Ratio.

One milliliter of human and cynomolgus monkey blood was spiked with 10 μl of standard solution (100 μg/ml; 1000 ng/ml final) and preincubated in a shaking water bath at 37°C for 10 min. A 200-μl aliquot was then analyzed to determine the drug concentration in the blood. The remaining samples were centrifuged at 1800g for 10 min at 4°C, after which the drug concentration in 200-μl aliquots of plasma was determined. The Rb was then calculated from the concentrations of drug per milliliter of blood and plasma. All data regarding TAC level in humans and cynomolgus monkeys were determined by blood level base because the Rb value of TAC has been reported to be nonlinear, with values between 10 and 40 depending on the drug concentration in humans (Wallemacq et al., 1993).

Parallel Artificial Membrane Permeability Assay.

The parallel artificial membrane permeability assay (PAMPA) method was carried out by using a PAMPA Evolution instrument from pION INC. (Woburn, MA) (Avdeef et al., 2005). The lipid solution consisted of a 20% (w/v) dodecane solution and lecithin mixture. The donor solutions consisted of test compounds dissolved in 10 mM dimethylsulfoxide diluted in pH 6.5 buffer (final concentration of 50 μM). The acceptor plate was filled with 1% (w/v) SDS in water, and the pH was adjusted to 7.4 with 1N hydrochloric acid. The test plate was incubated for 120 min at 30°C. The concentration of each test compound in the reference, donor, and acceptor plates was measured with a UV plate reader. The permeability coefficient was calculated by using Evolution Library Manager software version 2.2 (pION INC.).

Plasma Protein Binding.

The plasma protein binding (unbound drug fraction in plasma) was determined by using the equilibrium dialysis method or ultracentrifugation method and the following equations: protein binding (%) = (1 − fp) × 100, and fp = concentration in filtrate or supernatant/concentration in serum, where fp is the unbound drug fraction in plasma. The unbound drug fraction in blood (fb) was calculated by dividing fp by Rb.

Equilibrium dialysis method.

A DIANORM dialysis device (Diachema, Zürich, Switzerland), which is impermeable to substances with molecular weights greater than 10,000, was used. Aliquots (3.5-ml) of human and cynomolgus monkey plasma were spiked with 35 μl of standard solution (100 μg/ml; 1000 ng/ml final) and preincubated in a 37°C shaking water bath for 10 min.

One milliliter of mixture and isotonic phosphate buffer solution (pH 7.4) was put into the dialyzing cell and receptor cell, respectively. After 4-h incubation at 37°C, the plasma mixture and buffer sample were stored in 100-μl aliquots at −20°C until analysis.

Ultracentrifugation method.

Ten microliters of standard solution (100 μg/ml) was added to 1000 μl of human or cynomolgus monkey plasma. The calibration samples were prepared by adding 17 μl of acetonitrile (50%) to 1700 μl of human or cynomolgus monkey plasma. These samples were then centrifuged at 436,000g for 140 min at 37°C by using a Beckman Optimal TL ultracentrifuge (Beckman Coulter, Fullerton, CA). After ultracentrifugation, the unbound fp was calculated by dividing the concentration of drugs in the supernatant by that in the plasma.

In Vitro Metabolism in Liver and Intestine Microsomes.

Metabolism study conditions.

The time courses of the unchanged drugs were obtained. Each drug was incubated at 37°C with a reaction mixture (1 ml) containing 500 μl of potassium-phosphate buffer (200 mM; pH 7.4), 100 μl of 1 mM EDTA-NaOH (pH 7.4), 100 μl of liver or intestine microsomes solution (the final concentration of microsomal protein was 0.05 mg/ml for TAC, 0.5 mg/ml for HT and DIG, and 0.2 mg/ml for all other drugs), 190 μl of distilled water, and 10 μl of each compound solution in 50% acetonitrile (final concentration: 0.2 μM).

After a 5-min preincubation, the reaction was initiated by the addition of 100 μl of an NADPH-generating system. The reaction was terminated by adding 100 μl of reaction mixture to 200 μl of acetonitrile including the internal standard at various time periods. After stopping the enzyme reaction, the reaction mixture of TAC and DIG was extracted with 3 ml of tertiary butyl methyl ether, and the solvent was removed under a stream of nitrogen gas. The residue was then dissolved in 150 μl of mobile phase, and a 10-μl aliquot was injected into the LC-MS/MS. The reaction mixture of DEX and NIF was centrifuged at 10,000g for 5 min. The supernatant (100 μl) was then decanted, and 30-μl aliquots were injected into the LC-MS/MS.

The reaction mixtures of all other drugs were centrifuged at 10,000g for 5 min. The supernatants (100 μl) were decanted, and 10-μl aliquots were injected into the LC-MS/MS.

In this experiment, the unchanged concentrations of all drugs were determined by using LC-MS/MS analysis. Mass number of molecular ion and product ion for each compounds were identified as follows (polarity, molecular>product): HT m/z = 296 > 269 [M+H]−; DEX m/z = 393 > 91 [M+H]+; NIF m/z = 347 > 315 [M+H]+; MDZ m/z = 326 > 291 [M+H]+; QID m/z = 325 > 307 [M+H]+; TAC m/z = 821 > 769 [M+NH4]+; VER m/z = 455 > 165 [M+H]+; DIG m/z = 780 > 85 [M+H]−; IBU m/z = 205 > 161 [M+H]−; TIM m/z = 317 > 261 [M+H]+; AMI m/z = 278 > 117 [M+H]+; PRO m/z = 260 > 116 [M+H]+.

The Prominence 2000 series (Shimadzu) was used as the LC-system. The MS/MS analyses were conducted on an API-3200 LC-MS/MS system (Applied Biosystems). For TAC, an Alliance HT Waters 2790 separations module and Micromass Quattro Ultima (Waters Corporation, Milford, MA) were used for the LC-MS/MS analysis.

The Supelco RP-Amide (3 μm, 3.0 × 31 mm; Supelco, Inc., Bellefonte, PA) was used as the analysis column for HT and DIG. The Capcell PAK MG (3 μm, 2.0 × 35 mm; Shiseido Corporation, Kyoto, Japan) HPLC column was used for all other drugs.

The flow rate was 0.3 ml/min. The column temperature was 50°C. The gradient system was used, starting with an ammonium acetate concentration of 20 mM (pH 4.8)/acetonitrile (9:1) for 0.5 min, and increasing the ratio of acetonitrile to 20 mM ammonium acetate (pH 4.8)/acetonitrile (1:9) over 0.5 min, which was then held for 2.5 min. The initial conditions were restored over 0.1 min, after which the column was re-equilibrated for 1 min.

Calculation of CLint_liver.

CLint_liver was calculated by using the following equation based on the time course of the residual ratio of the unchanged drugs as determined using least-squares linear regression (Naritomi et al., 2001): CLint_liver (ml/min/mg protein) = ke/microsomal protein concentration, where ke is the disappearance rate constant.

In the case of liver microsomes study, the units of CLint_liver values were converted to per kilogram of body weight by using the following equation: CLint_liver (ml/min/kg) = CLint_liver (ml/min/mg protein) × SF1 (mg protein/g liver) × SF2 (g liver/kg body weight), where SF1 is the microsomal protein content per gram of liver [48.8 was used for both species (Naritomi et al., 2001), assuming that the SF1 in cynomolgus monkeys is the same as in humans] and SF2 is the liver weight per kilogram of body weight (25.7 and 30.0 were used for humans and cynomolgus monkeys, respectively) (Davies and Morris, 1993).

Calculation of in vitro intrinsic clearance in intestine.

CLint_intestine was calculated by using the following equation based on the time course of the residual ratio of the unchanged drugs as determined using least-squares linear regression (Naritomi et al., 2001): CLint_intestine (μl/min/mg protein) = ke/microsomal protein concentration.

P-gp ATPase Assay.

Each drug was dissolved in dimethylsulfoxide (0.1–100 μM final) and preincubated for 5 min with 2 μg/ml human P-gp membrane (BD Gentest, Woburn, MA) in 50 mM MES buffer (pH 6.8 adjusted with Tris) containing 2 mM EGTA, 2 mM dithiothreitol, 50 mM potassium chloride, and 5 mM sodium azide. Then, the ATPase reaction was started by the addition of 50 mM Mg-ATP solution. After 20-min incubation at 37°C, the reaction was stopped by adding 20 μl of sodium dodecyl sulfate (10%) containing Antifoam A (Sigma-Aldrich). Subsequently, 200 μl of ammonium molybdate/zinc acetate was added for color development, and the mixture was incubated for another 20 min at 37°C. After incubation, the amount of liberated phosphate was measured by using the UV absorption method (630 nm). Baseline activity was determined by reading incubated sodium orthovanadate (100 μM). Finally, ATPase activity was determined as the amount of liberated phosphate per milligram protein per minute. VER was evaluated in all ATPase assays, and the ATPase activity of each drug was normalized by dividing by the VER ATPase activity for each experiment.

Calculation of in Vivo Pharmacokinetic Parameters.

Plasma concentration data were analyzed individually at each point in time, and pharmacokinetic parameters were calculated by using a model-independent method. F, FaFg, and Fh were then calculated from these pharmacokinetic parameters and Rb (see Blood-to-Plasma Concentration Ratio under Materials and Methods) by using the formulas shown below. For Li and HT, we assumed that these drugs underwent almost no in vivo metabolism and that their FaFg values (meaning Fa in this case) were equal to F. The F values for the drugs in cynomolgus monkeys were determined by using the following equation: F(%) = {AUCinf (p.o.)/AUCinf (i.v.)} × (Dose i.v./Dose p.o.) × 100, where AUCinf (i.v.) and AUCinf (p.o.) are the area under the plasma concentration-time curve calculated using the trapezoidal rule with extrapolation from the last measured plasma concentration to infinity after intravenous and oral administrations, respectively.

The Fh of drugs was determined by using the following equation and assuming that the elimination of drugs from the body after intravenous administration consisted of liver metabolism and renal excretion: Fh = 1 − {(CLh/Rb)/Qh}, CLh = CLt × (1 − fe); fe = urinary excretion of unchanged of unchanged drug after intravenous administration, where Qh is the blood flow rate in the liver (the human and cynomolgus monkey Qh values were 20.7 and 43.6 ml/min/kg, respectively) (Davies and Morris, 1993), CLh is hepatic clearance, CLt is total clearance, and fe is the urinary excretion ratio of the unchanged drug after intravenous administration. In cases where the fe value was not available, the CLh was assumed to be equal to the CLt.

The drug FaFg values were determined by using the following equations, assuming that the F was expressed as the product of FaFg and Fh: F(%) = Fa × Fg × Fh × 100, FaFg = {F(%)/100}/Fh. The F, FaFg, and Fh values of each drug in humans were also calculated in a similar manner by using the reported pharmacokinetic parameters.