Introduction

Top Abstract Introduction Results Discussion References

The HIV-1 protease inhibitor, saquinavir (SQV),¹ was the first of a new class of agents for the treatment of HIV disease and is a highly selective inhibitor of the HIV protease enzyme (Kitchen et al., 1995; Schapiro et al., 1996). When administered alone or in combination with reverse transcriptase inhibitors, a significant decrease in viral load is observed (Collier et al., 1996). However, the hard gelatin capsule formulation of SQV, Invirase, exhibits low oral bioavailability, measured as 4% in healthy volunteers after a single dose and estimated to be approximately 10% at steady state in HIV-positive patients (Noble and Faulds, 1996; Perry and Noble, 1998). To increase drug exposure, a new soft gel formulation of SQV, Fortovase, has been introduced with a bioavailability approximately three times that of Invirase. SQV is highly protein-bound (97%) (Halifax et al., 1998). Although SQV has been shown to significantly reduce viral loads in patients, treatment failures and the emergence of resistance are problems with this protease inhibitor as well as for all others (Condra et al., 1995; Roberts, 1995; Jacobsen et al., 1996).

Previous studies from this laboratory and others indicate that SQV interacts with the multidrug transport system, P-glycoprotein (P-gp) (Alsenz et al., 1998; Kim et al., 1998; Lee et al., 1998; Washington et al., 1998). P-gp is a plasma membrane efflux pump that transports a variety of amphiphilic and hydrophobic drugs from the cell, including a number of antineoplastic agents. P-gp expression in tumor cells is associated with the development of multidrug resistance. P-gp is also expressed in normal tissues, including the apical membranes of intestinal and renal epithelia and the endothelial cells of the blood-brain barrier (Thiebaut et al., 1987; Cordon-Cardo et al., 1989). Its role in normal tissue is not clearly understood but is believed to have a protective function by preventing the accumulation of exogenous substances inside the cell (Borst and Schinkel, 1996; Schinkel et al., 1996). P-gp is encoded by the MDR1 and MDR3 genes in humans and by the mdr1a, mdr1b, and mdr2 genes in mice. MDR1 in humans and mdr1a encoded P-gp in mice are expressed in similar tissues. A phenotypically normal and viable mouse model with a disruption in the mdr1a gene has been generated [mdr1a(/)] (Schinkel et al., 1994). Studies with these mice have shown increased central nervous system (CNS) permeability of a number of compounds, including digoxin, cyclosporin (Schinkel et al., 1995), loperamide (Schinkel et al., 1995), and vinblastine. These mdr1a(/) mice have also been used to demonstrate the role of P-gp in limiting the oral bioavailability of paclitaxel (Sparreboom et al., 1997).

Our overall goal was to determine whether P-gp plays a role in the accumulation of the protease inhibitor, SQV, into cells or into the CNS. In these studies, the disposition of SQV was compared in mdr1a(/) and mdr1a(+/+) mice. In addition, P-gp-negative and P-gp-positive cells were used to directly determine whether SQV is a substrate for P-gp.

Experimental Procedures

Materials. mdr1a(/) mice (P-gp-deficient) and FVB mice (P-gp-positive) were purchased from Taconic Laboratories (Germantown, NY). Radiolabeled (¹⁴C and ³H) and unlabeled SQV were synthesized at Roche Discovery Welwyn, England and Hoffmann-La Roche, Basle, Switzerland, respectively (Wiltshire et al., 1998). Ritonavir was a gift from Abbott Laboratories (Abbott Park, IL). Cell culture media was purchased from Life Technologies (Gaithersburg, MD). All other chemicals were purchased from Aldrich Chemical Co. (Milwaukee, WI), Fisher Scientific (Fairlawn, NJ), or Sigma Chemical Co. (St. Louis, MO).

Cell Lines. MES-SA (P-gp-negative) cells were derived from a human uterine sarcoma. Dx5 (P-gp-positive) cells were derived from MES-SA cells grown in the presence of doxorubicin, a known P-gp substrate. Cells were purchased from the American Type Culture Collection (Rockville, MD).

Accumulation Studies in MES-SA and Dx5 Cells. MES-SA and Dx5 cells were plated at a density of 1 × 10⁶ cells/well in 6-well plates (Falcon, Franklin Lakes, NJ) and allowed to incubate overnight (37°C, 5% CO₂). Culture media was removed, and cells were washed with PBS and incubated with serum-free McCoy's 5A media. [¹⁴C]SQV (25.6 µCi/mg) or [³H]SQV (742 µCi/mg) was prepared in McCoy's 5A media supplemented with HEPES and added to cells. For the time course studies (see Fig. 1), cells were incubated at 37°C in the presence of [¹⁴C]SQV (100 mM). The reactions were stopped after incubations of either 30 s, 1, 5, 10, 30, or 60 min. Media was removed, and the cells were washed in ice-cold PBS buffer. Cells were solubilized with lauryl sulfate (SDS, 4%) and placed in scintillation vials. Biosafe II scintillation fluid (Mount Prospect, IL) was added, and radioactivity was determined by liquid scintillation counting. Data were normalized with protein determination.

SQV CNS and Cerebrospinal Fluid Distribution Study in Rats. Three male Sprague-Dawley rats were anesthetized with pentobarbital and given a single 5 mg/kg i.v. infusion of [¹⁴C]SQV into the carotid artery over 20 min. Cerebrospinal fluid (CSF) was then collected by cisternal puncture of the atlanto-occipital membrane using a butterfly needle and flexible cannula attached to a syringe and frozen at 20°C. Immediately afterward, blood samples were obtained by cardiac puncture, and the brain was removed. The major blood vessels were dissected away, and the brains rinsed with saline before being stored at 20°C. Plasma was obtained from the blood by centrifugation and stored at 20°C. The brains were homogenized in an equal volume of water (PT10-35 homogenizer, Kinematica GmBH, Littau, Switzerland). Weighed aliquots (~200 µl) were combusted in an auto-oxidizer (Tri-carb B306, Packard Instrument Co., Meriden, CT), and the radioactive carbon dioxide produced was absorbed in Carbosorb (7 ml, Packard) and counted in scintillation cocktail (11 ml, Optisorb, Fisons Corp., Bedford, MA).

SQV Distribution Studies in Mice. SQV distribution studies were carried out in mice that were greater than 4 weeks of age. Three to four mice were studied in each group. [¹⁴C]SQV was dissolved in 5% dextrose (25.6 µCi/mg, i.v. study) or suspended in 10% aqueous succinylated gelatin (0.51 µCi/mg, oral study). Between 0.5 (2 mg/kg) and 2.5 µCi (5 mg/kg) was injected into each mouse via the tail vein, and approximately 6.5 µCi was given by oral gavage. Wild-type and knockout mice were administered the same amount of radioactivity within doses. After appropriate intervals, mice were exsanguinated under methoxyflurane or carbon dioxide narcosis via cardiac puncture. The blood samples were then centrifuged and the serum or plasma was removed. Various tissues and organs were harvested, weighed, and homogenized. The contents of the small intestine, colon, and stomach were emptied before weighing and homogenization. With the exception of the small intestine and liver, whole organs from the i.v. experiment were homogenized in 1 ml of Solvable (Packard). Samples were then incubated overnight at 50°C and then treated with 100 µl of 30% H₂O₂ for 1 h at room temperature. Hionic Fluor scintillation fluid (Packard) was added, and radioactivity was determined by liquid scintillation counting. The radioactivity remaining in the organs obtained from the mice dosed orally was measured in Readysafe liquid scintillator after combustion in a Packard Model 307 sample oxidizer.

Mass Spectrometry of SQV. A previous mass spectrometric method for quantifying SQV in plasma using solid phase extraction and liquid chromatography/mass spectrometry/mass spectrometry (Knebel et al., 1995) was modified for this study. Plasma samples (0.1 ml) were spiked with 0.02 ml of 20 ng/ml D5-labeled Ro-31-8959 internal standard (Wiltshire et al., 1998), extracted, and then dried and reconstituted in the mobile phase. The samples (20 ml) were applied to a Zorbax Rx-C18 (2.1 × 150 mm, 5 m) column at 30°C with gradient elution by 10 mM ammonium acetate/0.1% acetic acid and acetonitrile. The retention time for both SQV and its internal standard was 4.3 min. A Micromass Quattro II triple quadruple mass spectrometer, in positive electrospray mode (ESP+) was used to analyze the samples. The molecular weight/charge ratios of the protonated precursor and product ions (m/z) were 671.3 to 128.0 and 676.2 to 133.0, for SQV and the internal standard, respectively. The quantification range was from 0.5 to 500 ng/ml.

Statistical Methods. The Kruskal-Wallis one-way ANOVA was used to compare the effects of the protease inhibitors on [³H]SQV accumulation. The Student's t test was used to compare the tissue radioactivity of [³H]SQV following i.v. administration in wild-type and knockout mice. A P value of <.05 was considered significant.

	Results

Top Abstract Introduction Results Discussion References

Accumulation Studies. [¹⁴C]SQV displayed enhanced accumulation in MES-SA (P-gp-negative) cells as compared with Dx5 cells (P-gp-positive) (Fig. 1). The effect of SQV, ritonavir, indinavir, and nelfinavir (Fig. 2) on the intracellular accumulation of [³H]SQV was also examined in the MES-SA and Dx5 cell lines. The accumulation of [³H]SQV in the presence of unlabeled SQV (50 µM) and ritonavir (50 µM) in Dx5 cells (P-gp-positive) was comparable with that observed in the MES-SA cells (P-gp-negative) and comparable with that observed in the presence of a known P-gp inhibitor, cyclosporin A. However in the presence of nelfinavir and indinavir (50 µM) in Dx5 cells, the accumulation of [³H]SQV was 55% and 42%, respectively of that observed in the MES-SA cells, indicating that these two compounds are less potent inhibitors then SQV and ritonavir.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   The accumulation of [14C]SQV (100 nM) in MES-SA () and Dx5 () cells. The accumulation of SQV was examined in MES-SA (P-gp-negative) and Dx5 (P-gp-positive) cells. Each data point represents the accumulation of SQV (mean ± S.E.) from three experiments.

View larger version (30K): [in this window] [in a new window]   Fig. 2.   The effect of the anti-HIV protease inhibitors on the accumulation of [3H]SQV (100 nM) in MES-SA and Dx5 cells. Unlabeled SQV (10 and 50 µM), ritonavir (10 and 50 µM), nelfinavir (10 and 50 µM), or indinavir (50 µM), followed by [3H]SQV (100 nM) was added to the Dx5 (P-gp-positive) cells and allowed to incubate for 1 h at 37°C. Bars represent the percentage accumulation from four replicate experiments. Statistical differences in accumulation as compared with the Dx5 group were determined using Kruskal-Wallis one-way ANOVA on ranks, with P < .05 being considered significant.

CNS Penetration in Rats. After a 20-min, 5 mg/kg i.v. infusion of [¹⁴C]SQV, average levels in brain tissue and CSF were 9.4% and 0.2% of the plasma levels, respectively (Table 1), indicating some penetration of the CNS, principally into the more lipophilic material. Thus the relative distribution of SQV between plasma, brain, and CSF is (100:10:0.2). Unchanged SQV accounted for over 60% of the circulating drug-related material; therefore, the figures for penetration into the CNS obtained by measurements of radioactivity are a good guide for those of the parent drug.

Tissue Distribution Studies in Mice. After an i.v. dose of 2 mg/kg (Table 2), there was approximately a 4-fold higher SQV brain concentration in the P-gp-deficient mice versus the wild-type, whereas SQV serum concentrations were comparable (11.49 ± 4.4 ng/ml versus 10.9 ± 2.9 ng/ml). When mice were dosed i.v. with SQV at 5 mg/kg, there was a 3-fold difference in brain concentrations between the P-gp-deficient and normal mice, although this difference did not achieve statistical significance due to the small number of animals dosed. Oral administration of [¹⁴C]SQV at 500 mg/kg to P-gp-deficient [mdr1a(/)] mice resulted in a 5- to 6-fold higher systemic exposure to total plasma radioactivity and to unchanged SQV when compared with normal animals (Table 3). The brain concentrations of drug-related radioactivity in the P-gp-deficient mice were approximately 10-fold those observed in the wild-type mdr1a(+/+) mice.

	Discussion

Top Abstract Introduction Results Discussion References

These studies demonstrate that SQV is not only an inhibitor (Washington et al., 1998) but also a substrate for the multidrug transporter, P-gp. The time course studies (Fig. 1) show that SQV uptake is enhanced in MES-SA (P-gp-deficient) cells as compared with Dx5 (P-gp-positive) cells. Compounds that are substrates for P-gp are actively pumped from cells, and therefore SQV accumulation is reduced in the Dx5 (P-gp-positive) cells. The inhibition studies show that ritonavir also interacts with P-gp and prevents the active efflux of [³H]SQV from cells (Fig. 2). Nelfinavir and indinavir (up to 50 µM) did not significantly enhance the accumulation of [³H]SQV in the Dx5 cells. P-gp transports a number of structurally unrelated compounds out of the cell. These include antineoplastic agents, immunosuppressive agents, and calcium channel blockers. However, it is still not clearly understood by what mechanism P-gp is able to recognize such diverse compounds. A number of studies suggest the presence of more than one drug binding site (Tamai and Safa, 1991; Litman et al., 1997). More recently, Dey et al. (1997) have demonstrated the presence of two nonidentical drug binding sites within P-gp. This may explain why previous studies from this laboratory (Washington et al., 1998) showed that nelfinavir inhibited the efflux of vinblastine and paclitaxel, known P-gp substrates, indicating that nelfinavir interacts with P-gp. Because nelfinavir did not significantly inhibit the efflux of [³H]SQV, this suggests that SQV and the antineoplastic agents bind at different sites. The inability of indinavir to inhibit the efflux of [³H]SQV at similar concentrations agrees with our previous data suggesting that indinavir has, at most, only a weak interaction with P-gp (Washington et al., 1998). However, a recent study by Kim et al. (1998) demonstrated that indinavir is a substrate for P-gp.

The brain concentrations of SQV in rats (Table 1) indicate that the penetration of SQV into the CNS is similar to other protease inhibitors (Lin et al., 1996; Denissen et al., 1997). The much lower concentrations in CSF are consistent with the lipophilic properties of SQV, which can only be found in aqueous solutions when bound to proteins.

Because P-gp has been shown to be expressed in the endothelial cells of the brain, the brush-border membranes of intestinal epithelia, and a number of other tissues, we wanted to determine the role that it plays in the disposition of SQV. The total radioactivity of SQV in tissues was determined in normal and P-gp-deficient mice following i.v. and oral administration. SQV radioactivity appeared to concentrate in the colon, kidney, liver, small intestine, and stomach following i.v. administration, but radioactivity in these organs was not significantly different between the two groups (Table 2). Brain concentrations of SQV, however, were 4- and 10-fold higher following i.v. and oral administration, respectively, in the P-gp-deficient mice. This suggests that P-gp may be responsible for limiting the CNS uptake of SQV. This is noteworthy, because the CNS is exposed to the HIV virus early in the infection process (Price, 1996). Thus, methods to enhance the accumulation of protease inhibitors into the CNS would be valuable in the treatment of AIDS dementia complex or HIV-associated cognitive/motor complex, common neurological complications associated with late HIV infection.

An oral dose of 500 mg/kg was used in these studies to generate concentrations comparable with those obtained after i.v. administration. Following oral administration, the AUC for SQV (Table 3) was significantly lower in the normal mice [mdr1a(+/+)] suggesting that P-gp plays a role in reducing the absorption or enhancing elimination of SQV, possibly at the level of the gastrointestinal tract. P-gp may cause the efflux of SQV back into the intestinal lumen resulting in a higher clearance.

Because SQV is a substrate for P-gp, this may play a role in limiting its oral bioavailability. Previous studies have demonstrated that SQV is also a substrate of the drug-metabolizing enzyme, cytochrome P450 3A in both the gut wall and the liver (Fitzsimmons and Collins, 1997). Cytochrome P450 3A, in combination with P-gp, may account for the large first pass effect observed after the oral administration of SQV.

Additionally, because SQV is a substrate for P-gp, it may be useful to administer P-gp inhibitors along with it to increase bioavailability and hence systemic exposure to the drug, thus resulting in more optimal plasma concentrations. Also, the coadministration of a P-gp inhibitor may lead to higher levels of SQV in the CNS and reduce the development of AIDS dementia complex. The particular effectiveness of the ritonavir/SQV combination in the treatment of AIDS may, in part, result from their mutual inhibition of P-gp, which will enhance the absorption of SQV and the CNS penetration of both compounds.