Effect of Interferon-γ on the Pharmacokinetics of Digoxin, a P-glycoprotein Substrate, Intravenously Injected into the Mouse

Drug transporters play a critical role in the disposition and elimination of xenobiotics. In particular, P-glycoprotein (P-gp) is an important ATP-dependent, efflux membrane transporter with a wide substrate specificity for a number of structurally diverse drugs (Gottesman and Pastan, 1993). P-gp is encoded by the multidrug resistance (MDR) gene family (MDR1 in humans; mdr1a and mdr1b in rodents) and is distributed in normal tissues, particularly in the epithelial cells of the gastrointestinal tract, liver, and kidney, which play important roles in drug disposition. The expression of P-gp in such tissues results in reduced drug absorption and enhanced elimination via the urine and bile.

Conversely, cytochrome P450 3A (CYP3A), a major phase I metabolizing enzyme predominantly expressed in the liver and small intestine, plays an important role in the metabolism of a variety of drugs used in clinical situations. Recent reports have suggested that intestinal CYP3A is involved in reducing the bioavailability of orally administered drugs by first-pass metabolism (Watkins, 1997; Wacher et al., 2001). Moreover, many CYP3A substrates are also substrates of P-gp. Thus, the drug metabolizing enzyme and efflux transporter can act in concert with tissue cells where both proteins are expressed. In addition, the expression of these two functional proteins, P-gp and CY3PA, is modulated by a variety of common compounds (Schuetz et al., 1996). P-gp expression is regulated by various substances, including xenobiotics and hormones (Sukhai and Piquette-Miller, 2000; Sukhai et al., 2000), whereas CYP3A expression is modulated by inflammation and infection (Morgan, 1997) or induced by certain drugs (Guengerich, 1999).

Cytokines, including interferons (IFNs), are also important modulators for both proteins. IFNs, which consist of type I IFN (IFN-α and IFN-β) and type II IFN (IFN-γ), have been used in the treatment of viral and neoplastic diseases (Baron et al., 1991; Heathcote, 1999). It has been demonstrated that hepatic CYP3A expression is significantly reduced by IFN-γ (Carlson and Billings, 1996; Donato et al., 1997), IFN-α (Craig et al., 1993), and IFN inducers (Sakai et al., 1992; Monshouwer et al., 1996). However, until now, the influence of IFNs on P-gp in vivo has received little attention.

In previous studies, we examined the effect of IFN-β and IFN-γ on the expression and drug efflux activity of P-gp in rat primary hepatocytes and found that P-gp-mediated efflux of rhodamine-123, a typical P-gp substrate, was significantly inhibited by IFN-γ treatment, but not by IFN-β treatment, without a reduction in P-gp expression in hepatocytes (Akazawa et al., 2002). Those findings led us to examine whether IFN-γ can modulate P-gp activity and/or its expression in vivo and alter the pharmacokinetics of P-gp substrates, because the in vivo effects of IFN-γ on P-gp activity and expression have not yet been reported. In the present study, we investigated the effect of mouse IFN-γ on the pharmacokinetics of P-gp and its activity and expression in various tissues, including the intestine and liver in the mouse. [³H]Digoxin was used as a model P-gp substrate because of its low toxicity and chemical stability in the mouse (Schinkel et al., 1995, 1997; Mayer et al., 1997; Kawahara et al., 1999).

Materials and Methods

Reagents, Antibodies, and Mice. Digoxin and [³H]digoxin (15.8 Ci/mmol) were purchased from Sigma-Aldrich, Inc. (St. Louis, MO) and PerkinElmer Life and Analytical Sciences, Inc. (Boston, MA), respectively. Recombinant mouse IFN-γ was kindly donated by Shionogi Research Laboratories (Osaka, Japan). Monoclonal antibody against human P-gp (C219) was obtained from DAKO Corp. (Carpinteria, CA), and polyclonal rabbit anti-rat CYP3A1 antibody from Chemicon International, Inc. (Temecula, CA); both cross react with the mouse counterparts. Peroxidase-conjugated rabbit antimouse IgG and goat anti-rabbit IgG antibodies were purchased from Amersham Biosciences Inc. (Piscataway, NJ).

Seven-week-old male ICR mice (28–32 g) were supplied by Shizuoka Agricultural Co-operative Association for Laboratory Animals (Shizuoka, Japan). Animal experiments were carried out in accordance with the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the U.S. National Institutes of Health.

Digoxin Pharmacokinetics Studies. Mice were given a single intraperitoneal (i.p.) injection with 1 × 10⁵ U of IFN-γ in 100 μl of phosphate-buffered saline or saline alone. This dose of IFN-γ was chosen based on the previous report about the curative effect of IFN-γ on both tumor growth and metastasis in mice (An et al., 1996). Twelve hours after IFN-γ or saline administration, mice were not allowed food for another 12 h, and then they were anesthetized with pentobarbital sodium. To collect the bile and urine, the bile duct and urinary bladder were cannulated. [³H]Digoxin was injected into the tail vein at a dose of 0.1 mg/kg, and the bile, urine, and blood were sampled for 4 h at appropriate time intervals. Four hours after digoxin administration, the mice were sacrificed and intestinal contents, plasma, and organs (liver, kidney, intestine, and brain) were collected. Aliquots of the samples were dissolved in 0.7 ml of Solene-350 at 45°C, combined with 0.2 ml of isopropanol, 0.2 ml of H₂O₂, 0.3 ml of 5N HCl, and 5 ml of Clearsol I (scintillation medium), and their radioactivities were measured using a liquid scintillation counter LSA-500 (Beckman Coulter Inc., Tokyo, Japan). The plasma concentration profiles of [³H]digoxin were analyzed by noncompartmental moment analysis using a trapezoidal method with mono-exponential extrapolation to infinite time to calculate the area under the curve (AUC), mean residence time (MRT), total body clearance (CL_tot), and steady-state volume of distribution (V_dss). The total clearance values for urinary (CL_urine), biliary (CL_bile), and intestinal secretion (CL_intestine) were determined by dividing the amounts of [³H]digoxin excreted into urine, bile, and intestine up to 4 h by the AUC value from 0 to 4 h, respectively.

Preparation of Plasma Membrane and Microsome Fractions. After IFN-γ or saline treatment and withdrawal of food as described above (i.e., 24 h after IFN-γ or saline treatment), mice were anesthetized with ether and sacrificed by bleeding, and livers were quickly infused with 1.15% KCl solution with 1 mM of EDTA and 1 mM phenylmethylsulfonyl fluoride via the portal vein and vena cava. The liver, intestine, kidney, and brain were clipped off and individually homogenized in 10 mM potassium phosphate buffer (pH 7.4) containing 250 mM sucrose and 1% (v/v) Sigma protease inhibitor cocktail (Sigma-Aldrich, Inc.) at 4°C. The long intestine was divided roughly into three segments as described (Fig. 5). For the preparation of plasma membranes, homogenates were centrifuged at 9000g for 20 min, and the pellets were resuspended in a small volume of 0.02 M Tris buffer (pH 7.4) with 250 mM sucrose and 1% (v/v) Sigma protease inhibitor cocktail, and stored at -80°C until analysis. For the preparation of microsomes, homogenates were centrifuged at 9000g for 20 min, and the supernatants were again centrifuged at 105,000g for 60 min. The resultant pellets were resuspended in a small volume of 0.02 M Tris buffer (pH 7.4) with 250 mM sucrose and 1% (v/v) Sigma protease inhibitor cocktail, and stored at -80°C until analysis. All procedures were carried out at 4°C. Protein concentrations were determined by the modified Lowry method (Lowry et al., 1951).

Western Blot Analysis of P-gp and CYP3A. Fifty micrograms of protein from plasma membrane preparations and 10 μg of protein from microsomal preparations were dissolved in loading buffer, denatured at 95°C for 3 min, and loaded on 6.5 and 12.5% SDS-polyacrylamide gels, respectively. After electrophoresis, the proteins were transferred onto a polyvinylidene difluoride membrane (Immobilon-P; Millipore Corp., Bedford, MA) by semidry-blotting with Transblot SD (Bio-Rad, Hercules, CA). P-gp protein was detected by primary monoclonal mouse antibody C219 (1:1000) and secondary peroxidase-conjugated rabbit anti-mouse IgG antibody (1:2000). CYP3A protein was detected by polyclonal rabbit anti-rat CYP3A1 antibody (1:1000) and secondary peroxidase-conjugated goat antirabbit IgG antibody (1:1000). Protein bands were visualized by chemiluminescence on the ECL protein detection system (Amersham Biosciences Inc., Piscataway, NJ) followed by exposure to Hyperfilm ECL (Amersham Biosciences Inc.). The densities of bands obtained on the films were re-quantitated using ATTO Image Analysis Software (ATTO Corp., Tokyo, Japan).

Results

Effect of IFN-γ Treatment on Disposition of [³H]Digoxin Intravenously Injected into Mice.Figure 1 shows the time course of the plasma concentration of [³H]digoxin after i.v. injection into control and IFN-γ-pretreated mice. In both cases, the concentration decreased primarily in an exponential manner with time; the decrease, i.e., elimination of plasma digoxin in the IFN-γ-treated mice, was significantly retarded compared with the control mice, and the level at 4 h was higher than that in the control mice.

The tissue distribution of [³H]digoxin 4 h after i.v. injection in the control and IFN-γ-treated mice is shown in Fig. 2. Compared with the control mice, accumulation of [³H]digoxin in the IFN-γ-treated mice increased significantly in the liver and kidney, slightly in the intestine, but not significantly in the brain.

The biliary, urinary, and intestinal excretion was also examined for up to 4 h after i.v. administration of [³H]digoxin. About 20, 50, and 15% of the administered dose of [³H]digoxin was excreted, respectively, into the urine, bile, and intestinal lumen of the control mice within 4 h (data not shown). The cumulative amounts of [³H]digoxin excreted into the urine and bile for up to 4 h after i.v. administration are illustrated in Fig. 3, A and B, and the secreted [³H]digoxin into the intestinal lumen at 4 h after i.v. administration is shown in Fig. 3C. In the IFN-γ-treated mice, both the amounts of [³H]digoxin excreted into the urine and bile were lower than those of control mice (Fig. 3, A and B), whereas the amount excreted from the small intestine was only slightly increased by IFN-γ treatment (Fig. 3C).

Pharmacokinetic Parameters. The pharmacokinetic parameters were determined from the plasma concentrationtime curves (Fig. 1) as summarized in Table 1. Compared with the control mice, the CL_tot value was reduced by half, the V_dss value was unchanged, and the AUC of [³H]digoxin after i.v. injection was increased about 2-fold in the IFN-γ-treated mice. Shown in Fig. 4, the CL values were determined from the biliary, urinary, and small intestinal excretion data in Fig. 3 and the AUC values. The CL_urine and CL_bile values in the IFN-γ-treated mice were 35 and 45% of the values in the control mice, respectively. The CL_intestine was not significantly altered by IFN-γ treatment.

Determination of P-gp and CYP3A Expressions. The expression of P-gp in the liver, kidney, small intestine, and brain were examined by Western blot analysis (Fig. 5). The P-gp expression levels in the liver and intestine were reduced in the IFN-γ-treated mice by approximately 20 and 20 to 30%, respectively, compared with the control mice. The P-gp levels in the kidney and brain were slightly reduced by IFN-γ treatment by approximately 12 and 5%, respectively.

The CYP3A levels in the liver and small intestine were also examined (Fig. 6). As reported previously (Carlson and Billings, 1996; Donato et al., 1997), the CYP3A expression level in the liver was reduced by IFN-γ treatment (by about 22%). The intestinal CYP3A expression in the mice was also reduced by IFN-γ treatment; the reduction was about 29, 28, and 55% in the upper, middle, and lower intestine, respectively.

Discussion

The present study demonstrates that pretreatment of a single therapeutic dose of IFN-γ significantly affects the pharmacokinetics of [³H]digoxin, a typical P-gp substrate administered intravenously into mice. [³H]Digoxin has been used for the in vivo evaluation of P-gp function in various mouse tissues, including mdr1a, mdr1b, and mdr1a/b knockout mice (Schinkel et al., 1995, 1997; Mayer et al., 1997; Kawahara et al., 1999) since the pharmacokinetic properties of this compound are highly P-gp-dependent and no significant metabolism occurs over a short period of time in mice. We observed retardation of the plasma elimination of [³H]digoxin and a concomitant increase of its tissue levels in IFN-γ-treated mice. The amounts of [³H]digoxin excreted into the urine and bile, but not the intestinal lumen, were reduced in IFN-γ-treated mice. Pharmacokinetic analysis revealed that the urinary and biliary excretion clearance values decreased by 65 and 55%, respectively, following IFN-γ treatment, indicating that the drug transport activities mediated by P-gp in the kidney and liver were significantly down-modulated. Meanwhile, there was no significant change in the P-gp-mediated secretion across the intestinal wall.

P-gp is expressed on the apical membrane in epithelial cells of normal tissues, including liver, brain, kidney, adrenal gland, and intestinal tissues (Thiebaut et al., 1987; Gottesman and Pastan, 1993). By contrast, the functional receptors for IFN-γ are expressed on the basolateral surface of epithelial cells, because such cells respond biologically to basolateral stimulation of the cytokine (Adams et al., 1993; Nakanishi et al., 2002). Thus, the investigation of the effects of systemic IFN-γ on epithelial cells makes sense both biologically and clinically.

We thus examined whether the “malfunction of P-gp” is related to the expression level of P-gp on Western blot analysis (Fig. 5). Although the P-gp expression was reduced in the liver of the IFN-γ-treated mice, and to a lesser degree in the kidney, a reduction in P-gp expression comparable with that in the liver was observed in the intestines, where digoxin secretion was not affected significantly by IFN-γ treatment. These results suggest that the apparent impairment of P-gp function observed in IFN-γ-treated mice is not primarily due to the P-gp expression levels in the tissues examined.

Although the mechanism by which IFN-γ affects the P-gp function itself is unclear, several possible mechanisms can be considered. It has been shown that phosphorylation of P-gp by protein kinase C (PKC) is an important step in its drug efflux activity. Sachs et al. (1995) indicated that inhibition of P-gp phosphorylation causes a reduction in P-gp activity with no significant change in the level of expression. IFN-γ may affect P-gp function through modulation of the PKC signal transduction system. In addition, it is well known that nitric oxide (NO) is generated as a result of the expression of the IFN-γ-inducible nitric-oxide synthase (Stark et al., 1998). Donato et al. (1997) showed that NO induced by IFN-γ down-regulates and inactivates CYP3A. NO may cause P-gp malfunction by modulating the process of protein sorting, ATPase activity, PKC activity, etc., but more studies are needed to elucidate these issues.

IFN-γ is capable of exerting its biological activities directly and indirectly. In the liver parenchymal and renal epithelial cells, distribution of IFN-γ to the target cells may be important since i.v. injected IFN-γ is preferentially distributed to the kidney, lung, and liver of the recipients (Cross and Roberts, 1993). It has been established that IFNs (including IFN-γ) increase the release of other cytokines, such as interleukins (ILs) and tumor necrosis factor (TNF-α) in vivo. These cytokines can trigger modulation of the in vivo P-gp function in IFN-γ-treated animals.

It has been shown that, in rat primary hepatocytes, TNF-α induces the expression of P-gp at both the transcription and translation levels, accompanied by enhancement of the transport activity of P-gp (Hirsch-Ernst et al., 1998), whereas IL-6 and IL-1β suppress the expression and activity of P-gp (Sukhai et al., 2000). Lipopolysaccharide and IL-6 cause a significant reduction of hepatic P-gp at the transcription and translation levels in mice (Hartmann et al., 2001). In human colon carcinoma cell lines, IFN-γ, IL-2, and TNF-α cause a reduction in MDR1 gene expression (Walther and Stein, 1994). In human myeloma cells, P-gp expression is unaffected by cytokines such as IFN-α, IFN-γ, and TNF-α, but reduced by IL-2 (Evans and Baker, 1992). IFN-α also induces the expression of P-gp in multidrug-resistant ChR C5 cells, but enhances the cellular uptake and cytotoxicity of doxorubicin, a P-gp substrate, in the presence of verapamil (Kang and Perry, 1994). On the other hand, IFN-α increases Adriamycin-induced cytotoxicity without a significant change in P-gp expression in a multidrug-resistant cell line (Scala et al., 1991). In human macrophages, IFN-γ up-regulates P-gp expression and its activity in a dose- and time-dependent manner (Puddu et al., 1999). These findings indicate that P-gp expression and activity are modulated by some inflammatory cytokines in a manner that differs from one cell type to another, which may account for the differential effects among the mouse organs observed in the present study.

IFN-γ also down-regulates the expression of CYP3A and its enzymatic activity in rat (Craig et al., 1990, 1992, 1993; Tapner et al., 1996) and human (Abdel-Razzak et al., 1993) hepatocytes in vitro. NO production is an important factor in the IFN-γ effect (Donato et al., 1997). In the present study, a similar IFN-γ effect on CYP3A expression was also observed in the mouse liver and intestine. This finding provides valuable information for the oral administration of CYP3A substrates.

In addition, multidrug resistance proteins (MRPs), which are expressed in the liver, kidney, etc., might play a role in digoxin secretion. If so, the decrease of biliary and urinary excretions might be ascribed to down-regulation and/or reduction of the activity of MRPs, although the interaction of MRPs with IFN-γ was not examined in this work. The recent reports, however, have demonstrated that digoxin efflux in intestinal tissues was mediated by P-gp, and the contribution of MRP(s) was minor or negligible (Stephens et al., 2001, 2002).