Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Halofantrine (HF)¹ is a phenanthrene-methanol derivative used in the oral treatment of uncomplicated chloroquine- and multidrug-resistant Plasmodium falciparum malaria in both adults and children (Boudreau et al., 1988). HF is metabolized into N-debutylhalofantrine (HFM; Fig. 1) essentially by cytochrome P-450 (CYP) 3A4 (Halliday et al., 1995). Maximal HF concentrations of plasma are reached at 2 µM (Karbwang and Na Bangchang, 1994). Its bioavailability is very weak and erratic (Karbwang and Na Bangchang, 1994), and hepatic clearance represents a large part of total clearance. Indeed, inhibition of HF metabolism may dramatically increase its plasma concentrations and the cardiotoxicity of this treatment. Significant prolongations in QTc intervals on electrocardiogram (Nosten et al., 1993; Castot et al., 1993) have been reported to be correlated with higher plasma concentrations of HF but not with those of its metabolite HFM (Karbwang and Na Bangchang, 1994). In addition, cardiotoxic effects have been shown to be more pronounced in patients receiving HF after treatment failures with mefloquine (Nosten et al., 1993).

View larger version (7K): [in this window] [in a new window]   Fig. 1.   HF and its major metabolite; HFM.

Metabolic interactions may exist with prophylactic antimalarials administered before HF or with curative antimalarials administered 1 or 2 days after HF in cases where it has been uneffective. Furthermore, many inhibitors and/or inducers of CYP 3A4 (Pichard et al., 1990) are widely used in clinical practice and some of them may also be administered in association with HF for the management of other pathologies or complications of severe falciparum malaria (Panisko and Keystone, 1990).

The aim of this study was thus to investigate the effect of other antimalarial drugs and of a typical CYP 3A4 inhibitor (ketoconazole) on HF metabolism by human liver microsomes.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Drugs and Chemicals. HF and HFM were a generous gift from SmithKline Beecham, Ltd. (Paris, France). Mefloquine was a gift from Dr. Gasser (Hoffmann-La Roche, Ltd., Basel, Switzerland). Proguanil, artemether, artemisine, quinine, and quinidine were supplied by Pr. Lebras (G.H. Bichat, Paris, France). Primaquine, sulfadoxine, and amodiaquine were donated by Dr. Gimenez (G.H. Pitié-Salpêtrière, France). Chloroquine and pyrimethamine were obtained from Specia RP, Ltd. (Paris, France). Doxycycline and all other reagents of the purest grade available were obtained from Sigma (Rueil Malmaison, France) or Bioblock Scientific (Illkirch, France).

Preparation of Human Liver Microsomes. Human Caucasian liver samples were obtained from 11 liver transplant donors. The protocol strictly followed guidelines of French legal and ethical committees. Liver fragments were immediately frozen in liquid nitrogen and stored at 80°C. Epidemiologic data was available for eight donors. They were all males, 15- to 45-years old, with no known drug history. The cause of death was gun shot to the head, cerebro-meningeal hemorrhage, or cranio-cerebral trauma. Microsomal fractions were prepared as described previously (Dragacci et al., 1987) and stored at 80°C until use. The microsomal batches were identified as follows: M10, M19, M22, M23, M25b, M27, M29, M30, M31, M32, and M37. The specific CYP 1A2, CYP 2C8, CYP 2C9, CYP 2D6, and CYP 3A4 contents were determined as described previously for all batches (except M25b, M27, and M37; B. Baune, J.P. Flinois, V. Furlan, F. Gimenez, A.M. Tabenet, L. Becquemont and R. Farinotti, in press) and the mean values of contents were 96 ± 66, 142 ± 77, 170 ± 92, 175 ± 82, and 123 ± 115 arbitrary units (A.U.), respectively. A.U. were calculated on the basis of the CYP content of M23, which was set to 100. Protein content of microsomal preparations was determined by the Bradford method (Bradford, 1976).

Measurement of Kinetic Parameters of HF Metabolism. Incubations (100 µl final volume) in propylene test tubes were carried out at 37°C for 60 min in a shaking water bath. The incubation mixture consisted of 50 µg of microsomal protein, NADPH (1 mM), and phosphate buffer (0.01 M, pH = 7.4). HF (at concentrations ranging from 2.5-200 µM, n = 9) was dissolved in phosphate buffer immediately before incubation. The rate of HF N-debutylation, under initial velocity conditions, was linear over 60 min of incubation and for 0.1 to 1.5 mg/ml of microsomal proteins. The apparent Michaelis-Menten constants, K_m and V_max, were estimated using GRAFIT (Version 3.0, Erithacus Software, Staines, UK), a nonlinear least square regression analysis software by proportional weighting (1/v²).

Effects of Antimalarial Drugs on HF Metabolism. Sulfadoxine, primaquine, artemether, pyrimethamine, artemisine, and amodiaquine in methanolic solution were evaporated to dryness at 30°C. After the addition of phosphate buffer, the solutions were rapidly mixed to dissolve these components. By contrast, mefloquine, quinine, chloroquine, doxycycline (dissolved in water), and proguanil dissolved in methanol/water (7:93 v/v) were directly added to the incubation mixture without the evaporation step. Microsomal preparations and HF (50 µM final concentration) were subsequently added. After a 5-min preincubation period, the reaction was started by the addition of NADPH and stopped after 60 min by the addition of 50 µl 6 N HCl. The antimalarial drugs were studied at concentrations of 10, 100, and 500 µM, with the exception of proguanil (10, 100, and 300 µM) and primaquine (1, 5, and 10 µM). All incubations were performed in duplicate using microsomes from three human livers (M10, M22, M27). The extent of HF N-debutylation was expressed as a percentage of control.

Characterization of Inhibitory Potency of Mefloquine, Quinine, Quinidine, and Ketoconazole. The incubation mixture, in propylene test tubes, consisted of 50 µg of microsomal protein (obtained from three human livers: M10, M19, M22, M25b, M29, M31, or M37), NADPH (1 mM), and phosphate buffer (0.01 M, pH = 7.4). The reaction was carried out in duplicate at 37°C for 60 min in a shaking water bath. HF concentrations ranged from 20 to 90 µM (n = 4). Quinidine and quinine were tested at 0, 25, 75, and 100 µM, mefloquine at 0, 15, 50, and 70 µM, and ketoconazole at 0.025, 0.050, and 0.075 µM. The inhibitor was added at the same time as the substrate. The apparent K_i values were estimated using GRAFIT, a nonlinear least square regression analysis software and using a proportional weighting (1/v²). The data were fitted using conventional relationships for competitive and noncompetitive inhibition (Segel, 1975). The choice of the inhibition model was determined by the size of the random distribution of residuals, the sum of squares of the residuals, and the S.E. of the parameters estimated (K_i, K_m, and V_max).

Predicted In Vivo Inhibition of HF Metabolism by Certain Antimalarials and Ketoconazole. The predicted in vivo inhibition percentage (i) by components (I) of the catalytic activities of the examined CYP were calculated for a noncompetitive inhibition as follows (Segel, 1975): (eq. 1) i = (([I])/([I] + K_i)) × 100. Assuming that this in vitro relationship also applies in vivo, the theoretical percent inhibition of HF clearance can be predicted using maximal plasma ketoconazole, quinine, quinidine, or mefloquine concentrations ([I]) reported in humans after oral administration.

Analytical Method. HF and HFM were quantified according to a previously described HPLC-UV detection method (Gimenez et al., 1992). The extraction procedure was slightly modified: internal standard was quinidine (40 µl at 100 µM in water) for mefloquine and amodiaquine inhibition studies, and mefloquine (25 µl at 40 mg/liter in methanol) for the other inhibition studies. The intra- and interday coefficients of variation did not exceed 4.6% in any of the assays and the detection limit was 0.06 µM for HFM, with a signal-to-noise ratio of 3:1.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of Some Antimalarials on HF Metabolism. Six of the eleven tested antimalarials exhibited an inhibitory effect on HFM formation by human liver microsomes (Table 1). The inhibitory rank order for the other drugs was as follows: primaquine > proguanil > artemether > mefloquine > quinine.

Type of Inhibition of HF Metabolism by Ketoconazole or Some Antimalarials. Ketoconazole, mefloquine, quinine, and quinidine inhibited HFM formation noncompetitively with a mean K_i equal to 0.05 ± 0.02 µM, 70 ± 21 µM, 49 ± 16 µM, and 62 ± 15 µM, respectively (Table 2).

Predicted In Vivo Inhibition of HF Metabolism by Ketoconazole, Mefloquine, Quinine, or Quinidine. Using peak plasma concentration values reported after oral administration to humans of mefloquine (a dose of 250 mg; White, 1985), quinine (25 mg/kg/d; Franke et al., 1987), quinidine (22 mg/kg/d; White, 1985), and ketoconazole (a dose of 200 mg; Borelli et al., 1979; Graybill et al., 1980) in eq. 1, the calculated percentages of inhibition of HFM formation by mefloquine, quinine, quinidine, and ketoconazole were 7%, 49%, 26%, and 99%, respectively.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

As metabolic interactions may exist with prophylactic antimalarial drugs (chloroquine, proguanil, and mefloquine) used before HF or with curative antimalarials (quinine, quinidine, pyrimethamine-sulfadoxine, amodiaquine, artemisine, artemether, and primaquine) administered 1 or 2 days after HF, we have screened these antimalarials for their potency to inhibit HF metabolism. In humans, HF is metabolized into a single metabolite, HFM, mainly by CYP 3A4 (Halliday et al., 1995). Six of the eleven tested antimalarials exhibited an inhibitory effect on HFM formation by human liver microsomes (Table 1).

Sulfadoxine, pyrimethamine, doxycycline, chloroquine, and artemisine with an IC₅₀ value near 500 µM were considered without an inhibitor potency on HF metabolism. The major involvement of CYP 2C8 in chloroquine metabolism (CYP 3A4 has a minor role in this metabolism; J. Ducharme, B. Baune and R. Farinotti, personal data) can explain the lack of an inhibitory effect of chloroquine, whereas the increased formation of HFM by 10 and 100 µM artemisine may be due to CYP 3A4 activation (Kerr et al., 1994).

In vitro, HF metabolism appears to be more inhibited by antimalarial drugs than that of quinine (Zhao and Ishizaki, 1997), mefloquine (Bangchang et al., 1992a), and primaquine (Bangchang et al., 1992b). Amodiaquine, artemether, mefloquine, quinine, quinidine, proguanil, and primaquine may inhibit HFM formation. The inhibitory rank order for these drugs was as follows: primaquine > proguanil > artemether > mefloquine > quinine. The eventual interactions between HF and amodiaquine, proguanil, primaquine, or artemether has not been reported because association of these drugs are uncommon in the clinical setting.

Ketoconazole, mefloquine, quinine, and quinidine inhibited HFM formation noncompetitively with mean K_i values equal to 0.05 ± 0.02 µM, 70 ± 21 µM, 49 ± 16 µM, and 62 ± 15 µM, respectively (Table 2). Quinine, quinidine, and HF are CYP 3A4 substrates (Li et al., 1995; Halliday et al., 1995; Zhang et al., 1997). Theoretically, the inhibition of HFM formation by quinine or quinidine should be competitive. However, HF and quinidine (or quinine) might be bound on different conformers of CYP 3A4, as shown by Koley et al. (1997) for quinidine and nifedipine, and the inhibition may be characterized as noncompetitive. Another hypothesis could be that CYP 3A4 is an allosteric enzyme containing two substrate-binding sites for both HF and quinidine with a single effector site for HF or quinidine, respectively. Indeed, this has been described by Kenworthy et al. (1998) for testosterone and diazepam. These two hypotheses may explain the noncompetitive inhibition fitting of data obtained for the interaction between HF and quinine or quinidine.

The extrapolation formula (eq. 1), which does not take into account hepatic drug uptake, protein binding, rates of elimination, or intrahepatic concentrations of antimalarials (lipophilic drugs), suggests that we could expect the absence of an inhibitory effect of mefloquine in vivo (8% of predicted inhibition). Thus, the increased cardiotoxicity due to HF in patients receiving preliminary mefloquine (Nosten et al., 1993) could be due to a pharmacological rather than to a metabolic interaction. By contrast, quinine, quinidine, and ketoconazole might inhibit HF metabolism in vivo and increase HF-induced cardiotoxicity (maximum predicted inhibition percentages were 49%, 26%, and 99%, respectively). The association of these antimalarials requires a close monitoring of ECG. Furthermore, the coadministration of HF with ketoconazole must be avoided. In the literature, metabolism HF inhibition by ketoconazole in dog has been reported (Khoo et al., 1998). In the same way, arrhythmias have been observed with terfenadine, a drug biotransformed via CYP 3As, when it was administered with ketoconazole and itraconazole (Pohjola-Sintonen et al., 1993; Honig et al., 1993) or erythromycin (Campana et al., 1996). Clinicians should thus be aware of the potential interaction between CYP 3A4 substrates or inhibitors and HF when they are administered concomitantly.

In conclusion, our results show that ketoconazole as well as quinine and quinidine coincubated with HF in human hepatic microsomes leads to the inhibition of HF metabolism and may potentiate HF-induced cardiotoxicity. By contrast, none of the other antimalarials studied inhibited HF metabolism and, by extrapolation, CYP 3A4 activity.