Materials and Methods

Animals.

Eight-week-old, male CF-1 mice were obtained from Charles River Laboratories (Hollister, CA). Animals were housed in plastic, microisolator cages with wood chip bedding and had ad libitum access to water and Purina Rodent Chow (Purina Mills, Richmond, IN). All procedures were approved by the Merck Research Laboratories Institutional Animal Care and Use Committee. Tail tissue was obtained from all animals and the mice were separated into (+/+), (±), or (−/−) genotypes after PstI RFLP (Umbenhauer et al., 1997). This procedure accurately predicted the “sensitive” or “resistant” phenotype with respect to neurotoxicity induced by avermectins in this mouse strain and the phenotype correlated with the deficiency in PGP protein in these animals. Only the homozygous [(+/+) or (−/−) with respect to mdr1a] animals were used in subsequent experiments.

Chemicals and Dose Preparation.

IVM (mixture of 80% B1a and 20% B1b) and [22,23-³H] IVM B1a were synthesized by Merck Research Laboratories. CSA (Neoral microemulsion solution) was obtained from Sandoz Pharmaceutical Co. (East Hanover, NJ). [mebmt β-³H] CSA was obtained from Amersham Life Science, Inc. (Arlington Heights, IL). 8,9-Z-4′Epi-methylamino-avermectin B1a (MK-244) and 4,4-dimethyl-mebmt-CSA (L-674,184) were used as internal standards in IVM and CSA quantification procedures, respectively.

Due to extreme sensitivity of CF-1 PGP (−/−) mice to abamectin/IVM toxicity, a dose below the LD₅₀ (0.3 mg/kg) for abamectin was selected for IVM administration. IVM and CSA were given at doses expected to not saturate transporter systems. Tritiated IVM B1a was diluted with unlabeled compound to yield a specific activity of 400 μCi/mg; thus only the B1a component was radiolabeled. The i.v. dose, 0.2 mg/kg, was evaporated from a solution in ethanol and reconstituted in mouse serum for administration via the tail vein (approximately 3 μCi/mouse, 0.2 ml/mouse). The oral dose, 0.2 mg/kg, was suspended in 0.5% aqueous methylcellulose.

Tritiated CSA was diluted with the Neoral solution in saline to give a specific activity of 100 μCi/mg. The dose for i.v. and oral administration was 1 mg/kg (approximately 3 μCi/mouse, 0.2 ml/mouse). All animals were dosed in the fasted state. Food was returned to the animals 2 h after dosing. All other chemicals and reagents were of the highest grade commercially available.

Verapamil hydrochloride and [³H]IVM was formulated as above and coadministered i.v. The dose was 1 mg/kg verapamil and 0.2 mg/kg IVM. Blood and tissues were collected as described 2 h postdose.

Study Design and Sample Analysis.

Three to four mice from each group were sacrificed at each time point after oral or i.v. administration. Blood was collected by cardiac puncture and the liver, gall bladder, brain, and small intestines were removed. Lumenal contents were obtained by flushing the small intestine (stomach to caecum) with 5 to 10 ml of saline. All samples, except whole blood, were frozen at −70°C until analysis.

Aqueous homogenates were prepared from liver and brain in a Dounce homogenizer, using a 1:3 or 1:4 tissue/water ratio. A 0.5-ml aliquot was air-dried and total radioactivity in these samples was determined by combustion in a Model 307 sample oxidizer (Packard Instrument Co., Downers Grove, IL), followed by scintillation counting. Total radioactivity in whole blood was determined in the same manner. Gall bladders from three animals in each treatment group at each time point were pooled, diluted to 1 ml in water, and homogenized. Total radioactivity in pooled bile and in intestinal wash was determined by scintillation counting. Results are reported as ng equivalent/ml or nanogram equivalent in the collected fraction, mean ± S.D.,n = 3–4. The 2-h liver, brain, and intestinal wash samples from three to four mice were pooled and extracted twice with ethyl acetate. Radioactivity associated with parent compound in the extract was determined by radiochromatographic analysis using a methanol/water gradient mobile phase on a Zorbax SB-C8 column (DuPont, Wilmington, DE; see below).

Liquid Chromatography-Mass Spectrometry (LC-MS) Analysis.

Blood (0.2 ml) was extracted twice with 4 ml ethyl acetate. The extracts were combined, evaporated under nitrogen, and the residue reconstituted in acetonitrile/water (1:1, v/v). All LC-MS and LC-MS/MS experiments were performed with Perkin-Elmer Sciex API 300 mass spectrometers (Perkin-Elmer Cetus Instruments, Eden Prairie, MN), upgraded to API 365 specifications. All HPLC procedures were performed under isocratic conditions using Keystone NH₂50 × 4.6 mm columns (5 mm). The mobile phase for quantitation of IVM was 90% acetonitrile + 10% H₂O containing 10 mM ammonium acetate. For quantitation of CSA the mobile phase contained an additional 0.1% trifluoroacetic acid. The flow rate was 400 μl/min and 80 μl/min went to the mass spectrometer. The electrospray interface was used for ionization. Full scan MS/MS mass spectra were obtained and the following transitions were monitored: 892.4 → 307.2 (IVM), 886.4 → 158.1 (MK-244), 1219.8 → 1202.8 (CSA), and 1233.8 → 1216.8 (L-674,184). Standard curve samples (range 5–500 ng/ml for IVM and 1–600 ng/ml for CSA) and quality control samples were prepared to allow quantitation and validation of the method. The lower limit of quantification was 5 ng/ml for IVM and 1 ng/ml for CSA. Results are reported as ng/ml blood, means ± S.D.,n = 3–4.

Pharmacokinetic Analysis.

Pharmacokinetic parameters were calculated for CSA after i.v. administration. Area under the blood-concentration time curve (AUC) from 0 to 24 h was determined by the linear trapezoidal rule. Clearance was calculated as the Dose_i.v. divided by AUC. The V_dss of CSA was determined as follows: V_dss = Dose_i.v. × AUMC_(0–24)/AUC²_(0–24) and MRT was calculated as AUMC/AUC. Half-life was estimated from the slope of the terminal phase of the log concentration-time points using the 6-, 10-, and 24-h concentrations.

CYP and Enzyme Activity Measurements.

Four samples of liver microsomes were obtained from control, untreated mice, each prepared from three pooled livers. Total CYP content was determined by spectrophotometric analysis following the procedure ofOmura and Sato (1964). The activity of CYP3A was compared in these preparations by monitoring testosterone 6β hydroxylation (Wang et al., 1997) and nifedipine oxidation (Guengerich et al., 1986). Briefly, 0.25 mg/ml of microsomal protein was incubated with 250 μM testosterone or 100 μM nifedipine in phosphate buffer containing an NADPH-regenerating system. The reaction was quenched after 10 min with methanol, centrifuged, and the supernatants were injected directly onto an HPLC column and analyzed following published methods (Wang et al., 1997). The activity of NADPH cytochrome c reductase was measured by recording absorbance at 550 nm for 1 min after addition of microsomes to 0.1 mM NADPH and 0.05 mM cytochrome c in buffer. Cytosolic glutathione-S-transferase activity was determined using 1 mM 1-chloro-2,4-dinitrobenzene as the substrate (Habig et al., 1974). Protein was determined using the BioRad reagent with bovine serum albumin as standard. Statistical analysis was performed using two-tailed Student’s t test. An asterisk (*) denotes statistically different mean values, p < .05.

Concentrations of CYP3A and CYP2B in liver microsomes were determined using enzyme-linked immunosorbent assay (ELISA; Beltz and Burd, 1989) and Western blot analysis. Microsomes (final concentration 3 μg/ml) were incubated in ELISA microplate wells overnight at 4°C. Plates were washed, blocked, and incubated with goat anti-rat CYP2B1 and CYP3A2, or anti-human CYP3A4 serum (Daiichiu Pure Chemicals, Inc., Tokyo, Japan) for 1 h and developed with 3,3′,5,5′-tetramethylbenzidine in aqueous dimethylformamide and H₂O₂ (9:1). The absorbance at 655 nm was recorded in a Bio-Rad (Bio-Rad Laboratories, Inc., Richmond, CA) plate reader.

For Western blot analysis, mouse liver microsomal proteins were separated on a 10% polyacrylamide gel using a Bio-Rad minigel system. Proteins were transferred to nitrocellulose filter paper, washed, and incubated with goat anti-rat CYP3A2 serum in PBS containing 2% dry milk at 4°C overnight. The paper was washed three times with washing buffer and incubated with affinity-purified rabbit anti-goat IgG (H+L) horseradish peroxidase conjugate. The filter was developed with enhanced chemiluminescence detection system reagents (Amersham life Sciences, Inc.). The relative band intensities were quantified by laser densitometry (Model ARCUS II; AGFA Corp., Orangeburg NY).

In Vitro Metabolism of IVM and CSA.

Microsomal metabolism of [³H]IVM and [³H]CSA was performed using mouse liver microsomal fractions with 10 μM substrate and 2 mg/ml protein. After a 30-min incubation, protein was precipitated with methanol, and authentic standard was added for UV identification and analyzed by HPLC (Shimadzu 10A Series HPLC system). The analysis was performed on a Zorbax SB-C8 column with a methanol/water mobile phase. HPLC conditions were: IVM: 0 to 50 min, 86:14 (methanol/water), 50 to 55 min, 100:0; CSA: 0 to 55 min, 80:20, 55 to 60 min, 100:0. The flow rate was 1.0 ml/min. Radioactivity was determined in one-half-minute fraction collections by scintillation counting. Results are reported as percent parent compound remaining after 30 min.

Results

In Vitro Studies.

Tables 1 and2 summarize some physiological parameters associated with the mouse colony and activities of several drug-metabolizing enzymes in subcellular fractions from the livers of PGP (+/+) and (−/−) animals. Both subpopulations are comparable with respect to adult body weight and liver weight as a percentage of body weight (Table 1). Total liver microsomal and cytosolic protein content, as well as total CYP content, are also statistically indistinguishable. The contents of CYP isozymes CYP3A and CYP2B were not different between the (+/+) and (−/−) animals as determined by both ELISA and immunoblot analysis (Table 2, Fig.1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Western immunoblots of PGP (+/+) and (−/−) CF-1 mouse liver microsomes using an antibody recognizing rat CYP3A2.

Microsomal metabolism of testosterone to 6β-hydroxytestosterone and oxidation of nifedipine to its pyridine metabolite were not different between (+/+) and (−/−) animals. In addition, the profiles of other metabolites generated from testosterone biotransformation were qualitatively similar (not shown). The activities of NADPH-cytochrome c reductase and cytosolic glutathioneS-transferase also were not different between the two groups of male mice (Table 2). In female mice, 1.3- to 2-fold higher activities of nifedipine oxidation, NADPH-cytochrome c reductase and glutathione S-transferase were found in (−/−) animals.

The extent of metabolism of IVM and CSA in the presence of liver microsomes was not significantly different between (+/+) and (−/−) mice. In male mouse liver microsomal incubations, 88 to 92% of radioactivity remained as parent IVM after 30 min in both (+/+) and (−/−) groups and 89% and 76% in (+/+) and (−/−) microsomes from female mice, respectively. For CSA, 88 to 90% of radioactivity remained as parent CSA in all groups after 30 min. Radioactivity associated with a more polar peak, eluting about 3 min before parent compound, appeared in all CSA incubation samples. This peak, which was not present in time zero samples, accounted for 11 to 12% of total radioactivity in all groups.

Blood Concentration Profiles.

IVM

None of the animals of either genotype showed signs of neurotoxicity after a single i.v. or oral dose of IVM at 0.2 mg/kg. Blood concentration versus time profiles for total radioactivity and for parent IVM after i.v. or oral administration at 0.2 mg/kg are shown in Fig. 2, A and B, respectively. The profiles of total [³H]IVM after i.v. administration were essentially identical between (+/+) and (−/−) animals; parent concentrations were also similar between groups up to 6 h. Concentrations of total radioactivity were up to 3-fold higher in the blood of (−/−) animals after an oral dose; parent compound concentrations also followed this trend (B). AUC values for parent IVM were not calculated because concentrations were below the lower limit of quantification at 6 or 24 h in some animals.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Concentrations of IVM (parent and total [³H]) in blood of male CF-1 mice after an i.v. (A) or oral (B) dose of 0.2 mg/kg.

CSA.

Blood concentration versus time profiles for [³H]CSA and for parent CSA are shown in Fig.3, A (I.V. dose) and B (oral dose). The AUC, clearance, volume of distribution, and half-life values (Table2) indicate similar pharmacokinetic behavior of CSA in the two groups of mice. Large interanimal variability in concentrations of total radioactivity and parent compound after oral administration of CSA precluded any conclusion relating to differences in blood concentrations between the (+/+) and (−/−) animals. Oral bioavailability values were not calculated because parent CSA concentrations were not quantified past 6 h in the oral dosing protocol.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Concentrations of CSA (parent and total [³H]) in blood of male CF-1 mice after an i.v. (A) or oral (B) dose of 1 mg/kg.

Excretion into Bile and Intestinal Lumen after i.v. Administration.

Total radioactivity recovered in the intestinal lumen of CF-1 mice after i.v. administration of [³H]IVM and [³H]CSA is shown in Fig.4, A and B, respectively. Excretion of total radiolabel was diminished in animals deficient in PGP at every time point compared with PGP (+/+) animals and in addition, excretion of [³H]IVM into intestinal lumen [3.43% and 2.41% of dose, by (+/+) and (−/−) mice, respectively] was 8- to 10-fold greater compared with bile over this period. Radiochromatographic analysis of a pooled sample of intestinal wash showed 13.4% of total extractable radioactivity coeluted with parent IVM in (+/+) mice compared with 5.9% as parent in (−/−) mice. Parent IVM constituted less than 5% of total radioactivity in gall bladder bile from any animal in this study and also in a separate study with bile duct-cannulated anaesthetized mice (not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Intestinal excretion of [³H]IVM (A) and [³H]CSA (B) in male CF-1 mice after an i.v. dose of 0.2 mg/kg (IVM) or 1 mg/kg (CSA).

Results are expressed as ng [³H] equivalents excreted, mean ± S.D., n = 4.

After an i.v. dose of CSA, 11.5% and 3.7% of the administered dose was recovered in intestinal lumenal wash from the (+/+) and (−/−) mice, respectively. Parent CSA constituted 17.5% of total extractable radioactivity in the wash from (+/+) animals, compared with 2.7% in samples from (−/−) mice. Less than 2% of total radioactivity in gall bladder bile coeluted with parent CSA in samples collected from any animal at any time point.

Brain and Liver Concentrations.

IVM

Differences in accumulation of [³H]IVM in the brain, a target tissue for IVM neurotoxicity, was apparent between PGP (+/+) and (−/−) mice at every time point. Concentrations of [3H] were relatively constant over time in (+/+) animals by either route of administration (Fig. 5, solid columns). In contrast, concentrations in brains of PGP (−/−) mice increased with time after an oral dose (Fig. 5A, open columns) and were always higher than concentrations measured in (+/+) mice. At 24 h post oral dose, total radioactivity in brains of (−/−) mice were 141 ng equivalent/g, 70-fold higher compared with wild-type mice (2 ng equivalent/g). Parent IVM constituted 90% and 84% of extractable radioactivity from 2- and 24-h brain homogenates pooled from three to four mice, respectively, from the (−/−) group and recovery of extracted radioactivity was similar among all samples.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

[³H]IVM concentrations in brains of PGP (+/+) and (−/−) mice after i.v. (A) or oral (B) dosing at 0.2 mg/kg.

Results are means ± S.D., n = 4.

When the PGP substrate and inhibitor, verapamil, were coadministered i.v. (1 mg/kg) with [³H]IVM to (+/+) mice, brain concentrations of [³H]IVM increased approximately 3-fold at 2 h postdose compared with (+/+) animals given IVM alone, without affecting the blood concentration of IVM (not shown). Extracts from these brain samples showed that coadministration of verapamil did not alter the relative profile of parent IVM and other radioactivity-containing peaks under our radiochromatographic conditions.

In contrast to the brain concentration-time profile, liver concentrations after i.v. administration were similar in (+/+) and (−/−) mice (Fig. 6A). The fraction of total [³H] remaining as parent IVM at 2 h was also similar (54% in (+/+) mice and 51% in (−/−) mice). Concentrations in livers were generally higher and consistent with higher systemic concentrations in (−/−) mice after oral administration (Fig. 6B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

[³H]IVM concentrations in livers of PGP (+/+) and (−/−) mice after i.v. (A) or oral (B) dosing at 0.2 mg/kg.

Results are means ± S.D., n = 4.

CSA.

The lack of PGP at the blood-brain barrier resulted in higher brain concentrations of [³H]CSA after i.v. and oral administration (Fig. 7, A and B, respectively, open columns), similar to the pattern seen after IVM administration. Concentrations of [³H]CSA equivalents were 15-fold and 10-fold greater in (−/−) mice 24 h after i.v. and oral dosing, respectively. At 2 and 24 h postdose, 86% and 83% of the extractable radioactivity in brains of CF-1 (−/−) mice coeluted with parent CSA by HPLC analysis, respectively. Liver [³H] concentrations were similar between groups after i.v. administration at all time points. At 2 h postdose, parent CSA constituted 52% versus 46% of extractable [³H] in the livers of (+/+) and (−/−) mice. After oral administration, mean liver CSA concentrations were higher in (−/−) mice but the data were not statistically different due to large interanimal variability.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7

[³H]CSA concentrations in brains of PGP (+/+) and (−/−) mice after i.v. (A) or oral (B) dosing at 1 mg/kg.

Results are means ± S.D., n = 4.

Discussion

A survey of several drug-metabolizing enzyme activities, together with results of in vitro metabolism of IVM and CSA, indicated that hepatic metabolic capacity was essentially identical in PGP (+/+) and (−/−) subpopulations of CF-1 mice. Activities and quantities of several CYP isozymes were similar under the normal, non drug-induced state. Because accumulation of both IVM and CSA in brains of PGP-deficient mice could not be explained by the small differences in blood concentrations between the (+/+) and (−/−) mice, we conclude the presence or absence of the efflux protein at the blood-tissue barrier is the key factor resulting in selective tissue accumulation of these PGP substrates. The in vivo consequences of the differences found in in vitro activities in female mice (Table 2) are not known; these differences are under investigation in ongoing in vivo studies.

The blood concentration versus time profiles of IVM and CSA were not different between (+/+) and (−/−) male mice after i.v. administration. Although intestinal efflux of both compounds after i.v. administration was diminished in animals lacking PGP, the extent of drug elimination by this route was not sufficient to markedly affect the pharmacokinetics of IVM or CSA. In contrast with decreased intestinal elimination and enhanced oral absorption of Taxol reported in mdr1a knockout mice (Sparreboom et al., 1997), we cannot unequivocally conclude that oral bioavailabilities of IVM and CSA were enhanced in PGP-deficient CF-1 mice from this study, due to assay sensitivity and limited sampling time points. Higher concentrations of parent (and [³H])IVM and [³H]CSA after oral administration in systemic circulation suggests enhanced net absorption in (−/−) mice. Oral absorption of [³H]IVM is over 60% in mice (Merck & Co., internal report) and absorption of CSA is also reported to be high in preclinical species; however, extensive intestinal and hepatic metabolism contributes to the relatively low and variable bioavailability (Kolars et al., 1991; Wu et al., 1995). The relative contribution of intestinal versus hepatic first pass metabolism for CSA is not known in CF-1 mice and studies are in progress to evaluate intestinal CYP3A activity in these two groups of mice. Other factors to consider in interpretation of results of a compound subject to significant first pass extraction by the intestine include the dose and potential PGP-CYP3A interactions in both liver and intestines. The impact of intestinal PGP on oral bioavailability of drugs may be greatest where poor absorption is the main factor limiting oral bioavailability.

The use of PGP inhibitors (pharmaceutical agents and/or inactive excipients) as a means of enhancing systemic and tissue bioavailability of drugs has been demonstrated in vitro (Chervinsky et al., 1993) and in vivo (Webster et al., 1993; Didier and Loor, 1995; van Asperen et al., 1997) and reversal of multidrug resistance with this strategy was effective in cancer patients treated with the PGP inhibitor PSC 833 (Giaccone et al., 1997). In the present study, coadministration of verapamil with IVM increased the concentrations of IVM in brains of (+/+) mice after an i.v. dose. Because verapamil is also metabolized by CYP3A, experiments are underway to evaluate the pharmacokinetic and metabolic interactions in the presence of multiple PGP/CYP3A substrates. The CF-1 mouse model is thus useful to test the role of this efflux protein in absorption, disposition, metabolism, and elimination of drug candidates. The potential for significant interactions and altered dynamic response, however, argue for further investigations to understand the physiological role of PGP and other efflux transporters.