Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Cryptococcus neoformans is a commensal yeast which causes opportunistic infections in immunocompromised hosts and in particular in AIDS¹ patients. It has a marked predilection for the central nervous system. Cryptococcosis may occur as an asymptomatic pulmonary infection. However, in immunocompromised patients, it is usually present as a fatal disseminated disease including meningitis. C. neoformans enters the body by inhalation, but the primary pulmonary infection is often undiagnosed. Unidentified virulence factors allow the yeast to disseminate from the lungs and to reach the brain through capillary embolization and destruction (Huffnagle and McNeil, 1999).

The combination of amphotericin B and flucytosine is recommended for at least 2 weeks as the initial treatment of AIDS-associated cryptococcal meningitis. In the consolidation treatment, fluconazole is recommended for 8 to 10 weeks. Itraconazole may be a suitable alternative for patients unable to take fluconazole (Van der Horst et al., 1997; Saag et al., 2000). Itraconazole (ITC) is an azole broad-spectrum antifungal agent, commercially available as an oral capsule, oral solution, and injectable formulation. Although very lipophilic [partition coefficient in octanol-water (log P) of 5.66], accumulation of ITC in the brain is very limited compared with distribution to other tissues (Heykants et al., 1987). This limited transport to the brain is probably due to an efflux mediated by an efflux protein, P-glycoprotein (P-gp) (Miyama et al., 1998).

P-gp is an efflux plasma membrane protein overexpressed in tumor cells and acting as an efflux pump for anticancer drugs. It confers a multidrug resistance phenotype to these cells. It is also expressed in nonmalignant tissues such as lung, intestine, kidney, liver, epithelia, testis, and brain (Fojo et al., 1987; Thiebaut et al., 1989; Cordon-Cardo et al., 1990). This efflux protein has not only been identified in the endothelial capillaries in the blood brain barrier but also in the parenchyma (Lee et al., 2001). P-gp is encoded by a gene family comprising two mdr genes (MDR1 and MDR3) in humans and three mdr genes (mdr1a, mdr1b, and mdr2) in rodents (Ng et al., 1989), although only the expression of human MDR1 and rodent mdr1a and mdr1b appears to selectively confer multidrug resistance. Other efflux proteins are also expressed in the brain such as multidrug resistance proteins (MRPs) and the more recently described breast cancer resistance protein (BCRP/MXR/ACBG2) (Litman et al., 2001).

Several efflux protein inhibitors, such as valspodar (PSC833), GF120918, LY335979 are currently developed by pharmaceutical companies with the objective of modulating this phenomenon. These modulators were initially developed as specific inhibitors of P-gp. However, studies conducted with these inhibitors showed that other efflux proteins were involved. For the moment, none of the developed modulators are specific to P-gp (Choo et al., 2000; Maliepaard et al., 2001).

Interactions between ITC and P-gp were reported. In brain capillary endothelial cells MBEC4, ITC inhibits the efflux of P-gp substrates such as vinblastine and vincristine. The brain accumulation of ITC in knock-out mdr1a / mice is increased compared with that in mdr1a +/+ mice, suggesting that ITC is a substrate of P-gp (Miyama et al., 1998). In humans, it has been demonstrated that ITC decreases renal clearance of quinidine (Kaukonen et al., 1997) and digoxine (Jalava et al., 1997).

If P-gp is responsible for the low transport of ITC through the blood brain barrier, the combined administration of a P-gp inhibitor with ITC should increase cerebral concentrations of the antifungal agent. Similar results were observed for other drugs such as quinolones (Tamai et al., 2000) and HIV-1 protease inhibitors (Choo et al., 2000; Savolainen et al., 2002). Increased ITC brain concentrations could therefore improve the efficacy of a treatment with ITC for cryptococcal meningitis in humans.

The objective of our work was to demonstrate that a pharmacological inhibition of drug efflux could enhance ITC brain uptake and efficacy. The present study tested this hypothesis, first, by investigating the influence of a cerebral infection with C. neoformans, and the influence of efflux protein inhibition by GF120918, on the plasma and cerebral pharmacokinetics of ITC in mice infected with C. neoformans and, second, by studying the influence of this inhibition on the efficacy of treatment of a cerebral infection with C. neoformans with ITC.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. The parenteral formulation of itraconazole (10 mg/ml; 25 ml) was kindly supplied by Janssen Pharmaceutica (Beerse, Belgium). This formulation was diluted to a final concentration of 4 mg · liters¹ using the following solvent: 1.25 ml of propylene glycol, 188 µl of concentrated HCl, 20 g of hydroxypropyl--cyclodextrin, adjusted to pH 4.5 with concentrated sodium hydroxide, water qs 500 ml.

1

Animals. OF1 (4 to 5 weeks) and BALB/c (6 weeks) mice (Iffa Credo, L'Arbresele, France) were used. They were anesthetized using i.p. pentobarbital (70 mg · kg¹). Intracerebral inoculations were performed according to Blasi et al. (1992). Yeast cells were suspended in pyrogen-free saline and injected (10⁶ C. neoformans in 10 µl) into the brain, 1 mm laterally and posteriorly to the bregma at a depth of 2 mm, using a Hamilton glass micro-syringe connected to a 27-gauge disposable needle (NovoPen, 16 mm 27 gauge; Novo Nordisk Biochem, Franklinton, NC). Surgical mortality was around 1% and occurred within 5 min. After inoculation, mice recovered from the trauma within 30 to 60 min. They received food and water ad libitum. Weight loss and survival were followed after inoculation. All experiments complied with the European Community Guidelines for the use of experimental animals.

Experimental Infection. Two C. neoformans isolates (NIH 52D, H99) were used. NIH 52D was tested on OF1 and H99 on OF1 and BALB/c mice. C. neoformans isolates were subcultured for 48 h on Sabouraud chloramphenicol agar at 32°C. Before inoculation, yeasts were washed three times in saline and suspended in pyrogen-free saline to the desired concentration. Viability of the inoculum was checked by colony forming units counts.

Influence of an Experimental Intracerebral Infection with C. neoformans and of the Efflux Inhibition on the Pharmacokinetics of Itraconazole in the Brain. A preliminary experiment was conducted on OF1 mice by injecting a single i.p. dose of 10 mg · kg¹ of GF120918 to check whether this dose provided sufficient systemic GF120918 concentrations to inhibit P-gp. Twenty mice received a single i.p. dose of GF120918 and two mice were sacrificed at each of the following times: 10, 20, 30, 45 min, 1, 2, 3, 5, 7, and 17 h after GF120918 administration. GF120918 plasma concentrations were determined by liquid chromatography (as described below) and compared with its efficient EC50 concentration for P-gp inhibition, 20 nM (12 ng · ml¹).

Influence of Efflux Inhibition on the Efficacy of Itraconazole in the Treatment of an Experimental Intracerebral Infection of Cryptococcus neoformans in Mice. Forty-nine BALB/c mice were infected intracerebrally by inoculation of 10⁶ C. neoformans (H99) and separated into four groups of treatment (Table 2). Volumes injected were calculated according to mice weight and never exceeded 120 µl. Treatments were injected i.p., GF120918 being administered 20 min prior to ITC to have effective inhibition of P-gp when ITC was injected. A pilot pharmacokinetic study after administration of GF120918 was performed in infected BALB/c mice and showed similar parameters to those observed in noninfected OF1 mice (data not shown). For this reason, we considered that a twice daily administration of the P-gp inhibitor, GF120918 (10 mg · kg¹), should give concentrations always above its EC₅₀ and therefore inhibit the P-gp during the day course. Treatments were repeated twice a day until death. Efficacy of the antifungal treatment was evaluated by daily measurements of mice weights and by following mortality.

Determination of GF120918 in Plasma and ITC in Plasma and Brain Tissue by Liquid Chromatography. For both compounds, the liquid chromatographic equipment consisted of a WISP 717+ automatic sample injector (Waters, Millipore, Saint Quentin, France), a Shimadzu LC10 AV pump (Touzart and Matignon, Les Ulis, France), a Shimadzu SPD10 Av spectrophotometric detector programmed at 263 nm (Touzart and Matignon) and a Class VP automated software system (Touzart and Matignon). The chromatographic separation was performed on a Lichrospher 100 RP-18 (5 µm) (Lichrocart 125-4 HPLC cartridge) with a Lichrospher 100 RP-18 (5 µm) (Lichrocart 4-4) guard column (Merck, Darmstadt, Germany).

1

1

Pharmacokinetic Analysis. Model-independent pharmacokinetic parameters of ITC were determined using WinNonlin Standard edition version 1.5 (Pharsight Corporation, MountainView, CA). Because destructive sampling was used, in each of the treatment group defined in Table 1, the data of the 49 mice were pooled to estimate the mean parameters. Values of maximal concentration (C_max) and time to maximal concentration (T_max) were taken from the raw data. Area under the curve (AUC_last) was calculated from the plasma concentration-time profile by logarithmic trapezoidal method. Individual estimates of the apparent terminal elimination rate constant (_Z) were obtained by regression of terminal portions of the plasma concentration versus time curves. AUC_0- was calculated by the relation AUC_0- = AUC_last + C_last/_Z where C_last was the last measurable concentration and _Z the elimination phase rate constant determined from the raw data. Elimination half-life (t_1/2) was calculated as t_1/2 = 0.693/_Z. The plasma clearance (Cl) corrected with the bioavailability (F) was calculated as Cl/F = Dose/AUC_0-. The plasma volume of distribution (V_Z), based on the terminal phase and corrected with the bioavailability F, was calculated as V_Z/F = Dose/[_Z × AUC_0-].

Efficacy Analysis. Weight versus time curves were compared between the four groups defined in Table 2 using a linear mixed model. As the number of mice at each time decreased because of the deaths, we used a linear mixed model (Proc Mixed, SAS version 8.2; SAS Institute Inc., Cary, NC). The effect of GF120918 on ITC treatment was also evaluated by comparing weight versus time curves between group I (placebo GF120918 + ITC) and group IG (GF120918 + ITC). Survival was estimated in each group using Kaplan-Meier method and was compared between the four groups and specifically between the groups I and IG using a log rank test (Proc Lifetest, SAS version 8.2; SAS Institute Inc.).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Experimental Infection in Mice. After intracerebral injection of NIH 52D strain in OF1 mice, a weight decrease and neurological defects, but no mortality, were observed within 30 days following inoculation preventing assessment of treatment efficacy based on survival criteria. We thus used H99, which provided a more severe weight loss and a mean survival time of 11.7 days (8-14) in OF1 mice and 7.2 days (6-8) in BALB/c mice. While doing the first experiments of the pharmacokinetics study in OF1 mice and assessing the virulence of NIH 52D and H99 in our model, we showed that individual susceptibility to infection with C. neoformans in a model of disseminated infection in OF1 mice is associated with an inflammatory response in the brain of individuals able to eradicate the yeasts. We thus decided to use inbred mice (BALB/c mice) to preclude variations in treatment efficacy related only to individual susceptibility to infection.

HPLC Method Validation. HPLC methods were linear over the following ranges: 0 to 2500 ng/ml, 0 to 4000 µg/ml, and 0 to 10 µg/g for GF120918 in plasma, ITC in plasma, and ITC in brain, respectively. Intraday and interday variabilities were always under 9.5, 4.9, and 8.2% for GF120918 in plasma, ITC in plasma, and ITC in brain, respectively. Inaccuracies were under 15, 12.8, and 11.4% for GF120918 in plasma, ITC in plasma, and ITC in brain, respectively. Limits of quantification were 25 ng/ml, 20 ng/ml, and 25 ng/g for GF120918 in plasma, ITC in plasma, and ITC in brain, respectively. Recoveries were 74, 75, 60% for GF120918 in plasma, ITC in plasma, and ITC in brain, respectively.

Influence of the Experimental Intracerebral Infection with C. neoformans and of the Efflux Inhibition on the Pharmacokinetics of Itraconazole in Plasma and Brain Tissue. Preliminary experiments showed that i.p. injection of a single dose of 10 mg · kg¹ of GF120918 provided plasma concentrations above the EC₅₀ of the drug (12 ng · ml¹) from the first time of the pilot pharmacokinetics (10 min) up to at least 7 h after administration with a maximal concentration of 145 ng · ml¹. Therefore, this dose was adequate to inhibit P-glycoprotein and was later used to study the influence of this inhibition on the cerebral pharmacokinetics of ITC in mice.

1

1

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Mean plasma concentrations (±S.E.M.) of itraconazole versus time in the three following groups: () healthy OF1 mice treated with placebo GF120918 + ITC; () OF1 mice infected with 106 C. neoformans (NIH 52D strain) and treated with placebo GF120918 + ITC; () OF1 mice infected with 106 C. neoformans (NIH 52D strain) and treated with GF120918 + ITC.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Mean cerebral concentrations (±S.E.M.) of itraconazole versus time in the three following groups: () healthy OF1 mice treated with placebo GF120918 + ITC; () OF1 mice infected with 106 C. neoformans (NIH 52D strain) and treated with placebo GF120918 + ITC; () OF1 mice infected with 106 C. neoformans (NIH 52D strain) and treated with GF120918 + ITC.

View larger version (11K): [in this window] [in a new window]   Fig. 3.   Mean brain/plasma concentration ratios (±S.E.M.) of itraconazole versus time in the three following groups: () healthy OF1 mice treated with placebo GF120918 + ITC; () OF1 mice infected with 106 C. neoformans (NIH 52D strain) and treated with placebo GF120918 + ITC; () OF1 mice infected with 106 C. neoformans (NIH 52D strain) and treated with GF120918 + ITC

Influence of Efflux Inhibition on the Efficacy of Itraconazole in the Treatment of an Experimental Intracerebral Infection with C. neoformans in Mice. Weight loss in the four groups is reported in Fig. 4. Comparing the four groups, the mixed model method showed a significant difference in weight versus time curve. The comparison between the groups I (ITC + placebo GF120918) and IG (ITC + GF120918) showed a statistical difference in terms of weight loss, indicating that weights were significantly higher in mice treated with the combination ITC + GF120918, compared with mice treated with ITC alone (p < 0.003). Mice treated with placebos or with GF120918 alone showed superimposed curves with a quick decrease in weight from the third day after inoculation until death. The decrease in weight was slower for mice treated with ITC alone. For mice treated with the combination ITC + GF120918, weight stayed quite stable for 7 days and a decrease similar to the mice treated with ITC alone was observed from the 8th day after inoculation.

View larger version (10K): [in this window] [in a new window]   Fig. 4.   Evolution of mean percentage of weight loss compared with initial weight in the four groups of BALB/c mice infected with 106 C. neoformans (H99 strain) and treated as follows: () placebo GF120918 + placebo ITC; () GF120918 treatment + placebo ITC; () placebo GF120918 + ITC treatment; () GF120918 treatment + ITC treatment (n = 12).

View larger version (11K): [in this window] [in a new window]   Fig. 5.   Survival curve in the four groups of BALB/c mice infected with 106 C. neoformans (H99 strain) and treated as follows: () placebo GF120918 + placebo ITC; () GF120918 treatment + placebo ITC; () placebo GF120918 + ITC treatment; () GF120918 treatment + ITC treatment (n = 12).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The two strains (NIH 52D and H99) showed different degrees of virulence. The more virulent H99 was used to study the influence of efflux inhibition on the efficacy of ITC based on the outcome of the infection, whereas the less virulent NIH 52D was chosen to study the influence of efflux inhibition on the plasma and cerebral pharmacokinetics of ITC. The experimental infection after IC inoculation with C. neoformans has been described and used by others (Blasi et al., 1992; Ding et al., 1997) and was preferred to the i.v. injection to obtain a localized cerebral infection and to focus on the cerebral efficacy of ITC treatment. Blasi et al. demonstrated that, following intracerebral fungal inoculation, mice developed a lethal cerebral infection characterized by a dose-dependent massive colonization and a generalized meningoencephalitis associated with an extensive edematous reaction (Blasi et al., 1992).

As the determination of ITC concentrations in brain required the sacrifice of mice, only one sample of brain and plasma and one determination of cerebral concentration and plasma concentration were available per mice. By using several mice at the different time points of the pharmacokinetics, we were only able to draw a "mean" pharmacokinetic profile and to estimate mean pharmacokinetic parameters, but we were not able to estimate the corresponding variances. This problem might be resolved by using population pharmacokinetics or by using the Bailer's method (Bailer, 1988; Mager and Göller, 1998). The Bailer's method allows for the estimation of the area under the curve of drug concentration versus time when only one sample per animal is available and at least two animals are sampled at each time point. The method also allows for the computation of the variance of this AUC estimate. As population pharmacokinetics require a high number of animals, especially in the case of variable pharmacokinetics, we chose the Bailer's method. For this reason, statistics were only applicable to the AUC parameter (Table 3).

Drug transport through a physiological barrier may be mediated by paracellular or transcellular transfer. In the brain, under physiological conditions, paracellular is restricted to endogenous compounds due to tight junctions between cells. Drug transport is therefore limited to transcellular transfer. Several transporters involved in drug efflux have been described such as the multidrug resistance MDR1/P-glycoprotein, the multidrug resistance-associated proteins MRPs, and the breast cancer resistance protein BCRP/MXR/ABCG2. P-glycoprotein and MRPs have been characterized in brain although no data are available regarding cerebral localization of BCRP (Maliepaard et al., 2001). GF120918 has been described as P-gp and BCRP inhibitor (Maliepaard et al., 2001) but did not show any effect on MRP (Evers et al., 2000). P-glycoprotein is certainly involved in the transport of ITC in the brain as Miyama et al. (1998) observed increased cerebral ITC concentrations in mdr1a (/) knock-out mice. Moreover, the ratio of cerebral concentrations between mdr1a (/) and wild-type mice of about 2.5 obtained by Miyama et al. (1998) is very similar to the ratio we obtained between mice treated with GF120918 or mice treated with the corresponding placebo. It strongly suggests that, in our study, the inhibitory effect of GF120918 on brain efflux of ITC is explained by P-gp inhibition instead of BCRP inhibition. However, a partial involvement of BCRP in the cerebral uptake of ITC cannot be completely ruled out.

Our results showed that ITC brain uptake was increased by the cerebral infection and by P-gp inhibition. Cerebral ITC concentrations were significantly higher in infected mice compared with healthy mice indicating that the blood brain barrier (BBB) was probably affected. Several examples of modifications of the BBB by pathogens involved in meningitis have been reported in literature. In AIDS patients suffering from encephalopathy, tight junctions are altered by transport of macrophages infected with HIV-1 through the BBB (Luabeya et al., 2000). In cerebral malaria, Brown et al. (1999) reported that adhesion proteins (occludine, vinculine, ZO-1) were altered in cellular junctions and were responsible for a BBB rupture. On the contrary, Jong et al. (2001) studied the effect of Candida albicans on the brain and suggested that it could be transported through the BBB by transcytosis without BBB rupture. The transport of C. neoformans through the brain and its effect on the BBB have not yet been described. Only one study by Morris et al. (1998) described that of nine AIDS patients with C. neoformans meningitis, only three showed an alteration of BBB. These data suggested that, in our model, C. neoformans could provoke a rupture of tight junctions allowing ITC to cross BBB by paracellular transport without affecting transcellular transfer.

We observed that P-gp inhibition by GF120918 increased cerebral ITC concentrations without modifying plasma concentrations. Miyama et al. (1998) described similar results in rats by showing that coadministration of the P-gp inhibitors ketoconazole or verapamil increased cerebral concentrations of ITC without affecting plasma levels. In vitro data only reported ITC as a P-gp inhibitor on the accumulation of vinblastine or vincristine in mouse brain capillary endothelial MBEC4 cells (Miyama et al., 1998) and of vinblastine, daunorubicin, and doxorubicin into porcine kidney epithelial LLC-PK1 cells (Takara et al., 1999). But no in vitro information is available for ITC as a P-gp substrate.

Moreover, the expression of P-gp and other efflux proteins may also be affected by pathologies. In liver failure inducted by carbone tetrachloride, P-gp concentrations in the liver showed a 50% increase while they are stable in the kidneys and brain (Huang et al., 2001). Mycobacterium tuberculosis may increase P-gp expression in promonocytic U1 cells infected with HIV-1 (Gollapudi et al., 1994). Similar results were obtained in the same model with Salmonella typhimurium (Andreana et al., 1994). On the contrary, other papers showed that mdr1 expression could be decreased by Toxoplasma gondii after infection of cancer cells in mice (Varga et al., 1999). Our experiments do not allow us to draw direct conclusions on a modulation of the functionality of P-gp by the C. neoformans infection. However, even though we did not check the pharmacokinetics of ITC in noninfected mice pretreated with GF120918 nor compare them to noninfected mice without pretreatment, indirect evidence shows that this alteration may only be partial, as we nevertheless observed an effect of efflux inhibition after GF120918 treatment in infected mice.

Many articles have reported the use of ITC in the treatment of cerebral cryptococcosis (Van Cutsem, 1993) or other brain infections (Al-Abdely et al., 2000; Saulsbury, 2001). Although effective against the pathogens, the efficacy of ITC in the treatment of cerebral infections seems to be limited by its poor brain uptake. Inhibiting the cerebral efflux is a useful strategy to resolve this problem. The multidrug resistance phenomenon was first observed in cancer treatments. P-glycoprotein may be modulated by drugs. Several phase I studies demonstrated the advantages of associating a P-gp inhibitor in the treatment of tumors (Patnaik et al., 2000; Peck et al., 2001). Numerous P-gp inhibitors are currently developed by pharmaceutical companies. P-gp inhibitors are also studied to increase the transport of drugs through the brain in cerebral pathologies. Thus, in mice, the combination of the P-gp inhibitor GF120918 to D-penicillamine^2,5 enkephaline, opioid pentapeptide, improved the antinociceptive effect (Chen and Pollack, 1999). In the present study, we demonstrated that, in mice infected intracerebrally with C. neoformans, in terms of efficacy, groups treated with ITC had a lower weight loss and survived longer than the other groups. The increase of brain ITC concentrations by the P-gp inhibitor GF120918 significantly improved the weight curve. With regard to mortality, median survival was improved by a pretreatment with GF120918, but the number of mice in each treatment group was small (12 mice), and the difference in the survival curve was not significant.

Our results showed that, in mice, ITC brain transport is increased in case of cerebral infection with C. neoformans and is further increased following efflux inhibition by GF120918. Moreover, this efflux inhibition improved the efficacy of ITC in the treatment of cerebral infection with C. neoformans. In all studies reporting a lower efficacy of ITC over fluconazole in the treatment of cryptococcal meningitis, there is little information on plasma concentrations of the antifungal agent. ITC is not recommended in the first line of the fungistatic secondary treatment of meningitis with C. neoformans because of its variable pharmacokinetics, its erratic absorption, and its low transport through the brain. Our data suggest that the inhibition of efflux might improve the efficacy of ITC in the treatment of cerebral infection. Efflux proteins could also be involved in the intestinal transport of ITC, and the inhibition of efflux proteins could also stabilize intestinal transport and decrease the variability of absorption. However, this hypothesis has to be demonstrated.