Desbutyl-Lumefantrine Is a Metabolite of Lumefantrine with Potent In Vitro Antimalarial Activity That May Influence Artemether-Lumefantrine Treatment Outcome

Abstract

Desbutyl-lumefantrine (DBL) is a metabolite of lumefantrine. Preliminary data from Plasmodium falciparum field isolates show greater antimalarial potency than, and synergy with, the parent compound and synergy with artemisinin. In the present study, the in vitro activity and interactions of DBL were assessed from tritium-labeled hypoxanthine uptake in cultures of the laboratory-adapted strains 3D7 (chloroquine sensitive) and W2mef (chloroquine resistant). The geometric mean 50% inhibitory concentrations (IC₅₀s) for DBL against 3D7 and W2mef were 9.0 nM (95% confidence interval, 5.7 to 14.4 nM) and 9.5 nM (95% confidence interval, 7.5 to 11.9 nM), respectively, and those for lumefantrine were 65.2 nM (95% confidence interval, 42.3 to 100.8 nM) and 55.5 nM (95% confidence interval, 40.6 to 75.7 nM), respectively. An isobolographic analysis of DBL and lumefantrine combinations showed no interaction in either laboratory-adapted strain but mild synergy between DBL and dihydroartemisinin (sums of the fractional inhibitory concentrations of 0.92 [95% confidence interval, 0.87 to 0.98] and 0.94 [95% confidence interval, 0.90 to 0.99] for 3D7 and W2mef, respectively). Using a validated ultra-high-performance liquid chromatography-tandem mass spectrometry assay and 94 day 7 samples from a previously reported intervention trial, the mean plasma DBL was 31.9 nM (range, 1.3 to 123.1 nM). Mean plasma DBL concentrations were lower in children who failed artemether-lumefantrine treatment than in those with an adequate clinical and parasitological response (ACPR) (P = 0.053 versus P > 0.22 for plasma lumefantrine and the plasma lumefantrine-to-DBL ratio, respectively). DBL is more potent than the parent compound and mildly synergistic with dihydroartemisinin. These properties and the relationship between day 7 plasma concentrations and the ACPR suggest that it could be a useful alternative to lumefantrine as a part of artemisinin combination therapy.

Desbutyl-lumefantrine (DBL) or desbutyl-benflumetol is a 2,3-benzindene compound with antimalarial activity. Although previously considered only a putative metabolite of lumefantrine because of a lack of supportive pharmacokinetic data (20, 24), recent analytical developments have enabled the reliable detection of relatively low concentrations of DBL in samples of plasma from small numbers of patients treated with conventional doses of artemether-lumefantrine combination therapy (11, 15, 18). The ratio of the maximum plasma concentration (C_max) of the parent compound to that of the metabolite in this situation has varied substantially, from 6 (18) to >270 (11).

DBL is more potent in vitro against chloroquine (CQ)-resistant Plasmodium falciparum and Plasmodium vivax field isolates than lumefantrine (13, 17, 23). There is evidence of in vitro synergy between lumefantrine and DBL against P. falciparum but at ratios (999:1 and 995:5) that were presumably selected at a time when plasma concentrations of DBL were assumed to be very much lower than those of the parent compound (23, 24). Interactions between DBL and artemisinin were assessed in a study of schizont maturation in Thai P. vivax field isolates in which antagonism was found at low concentrations and apparent synergy was found at much higher concentrations (13). A subsequent similar field study confirmed the concentration-dependent synergy in strains of P. falciparum (16).

Because of its relative in vitro potency and evidence of low cardiac toxicity (26), DBL has been suggested as an antimalarial drug in its own right (24). There is, however, a need for a further assessment of its interactions with other antimalarial drugs, especially lumefantrine and the artemisinin derivatives, in strains of P. falciparum of differing drug sensitivities. In addition, and given that artemether-lumefantrine is recommended as a first-line treatment for uncomplicated malaria (30), there is also a need to confirm the relative plasma concentrations of lumefantrine and DBL together with their therapeutic implications.

MATERIALS AND METHODS

Antimalarial drugs.

Stock solutions of CQ diphosphate (Sigma Chemicals, St. Louis, MO), piperaquine tetraphosphate (Yick-Vic Chemicals and Pharmaceuticals, Hong Kong), lumefantrine, DBL (Novartis Pharma, Basel, Switzerland), and dihydroartemisinin (Sigma) were prepared in distilled water (CQ), 0.5% (wt/vol) lactic acid (piperaquine), 70% (vol/vol) ethanol (dihydroartemisinin), or a 1:1:1 (vol/vol/vol) mixture of linoleic acid, ethanol, and Tween 80 (lumefantrine and DBL). Stocks and working standards of lumefantrine and DBL were sonicated for 90 s in a water bath in prewarmed medium to facilitate dissolution. Each stock solution was stored in light-proof tubes at −20°C. On the day of the assay, aliquots were thawed and further diluted in RPMI medium to a working standard, and further 2-fold serial dilutions in complete RPMI medium (without hypoxanthine) at double assay concentrations were prepared for CQ (25 to 1,600 nM), piperaquine (6.25 to 400 nM), dihydroartemisinin (0.78 to 51.2 nM), and lumefantrine and DBL (3.12 to 400 nM).

Parasite cultures.

The laboratory-adapted P. falciparum strains 3D7 (CQ sensitive) and W2mef (CQ resistant) were cultured in RPMI 1640 HEPES medium (Sigma-Aldrich, St. Louis, MO) supplemented with 92.6 mg/liter l-glutamine (Sigma-Aldrich), 500 μg/liter gentamicin, 50 mg/liter hypoxanthine (Sigma-Aldrich), and 10% (vol/vol) pooled human plasma. Nonparasitized erythrocytes were added every 2 to 3 days or as required to maintain 0.5 to 5% parasitemia at 5% hematocrit. The cultures were incubated at 37°C in an airtight desiccator cabinet flushed with a gas mixture of 1% O₂ and 5% CO₂ in N₂ balance (BOC Gases, Australia) for 60 to 90 s daily or when opened. The oxygen concentration within the chamber ranged between 3 and 9% over 24 h. This provided a controlled microaerophilic atmosphere for the culture of P. falciparum as described previously (21, 27).

In vitro drug susceptibility.

Drug susceptibility was tested in triplicate. A total of 100 μl of dosed or drug-free medium without hypoxanthine, 90 μl of a red cell suspension (final 0.5% parasitemia and 1.5% hematocrit), and 10 μl of 5 mg/ml tritium-labeled hypoxanthine (Perkin-Elmer, Waltham, MA) were added to each well of a 96-well plate to a final concentration of 0.5 μCi/well. After 48 h of incubation, the plates were subjected to three freeze-thaw cycles for cell lysis and harvested onto a 96-well glass fiber filter mat (Perkin-Elmer, Waltham, MA) using a Harvester 96 instrument (Tomtec Incorporated, Hamden, CT). After air drying, filter mats were sealed in plastic envelopes (Filtermat Bag 1450-432; Perkin-Elmer, Waltham, MA) with 4 ml of beta scintillant (Perkin-Elmer, Waltham, MA) and counted on a 1450 Microbeta Plus liquid scintillation counter (Wallac, Turku, Finland).

Drug interaction studies.

The traditional checkerboard method (1, 2, 4) and a recently described fixed-ratio approach (10) were used to analyze the isobolographic data. Briefly, 5 μM solutions of each drug were used to prepare 17 fixed molar combination ratios in 12-well plates. The molar ratios tested for lumefantrine-DBL and dihydroartemisinin-DBL were 1:0, 1:1, 1:3, 1:10, 1:30, 1:100, 1:300, 1:1,000, 1:3,000, 3,000:1, 1,000:1, 300:1, 100:1, 30:1, 10:1, and 3:1, corresponding to within-well assay concentrations (in nM) of 400:0, 400:400, 133.3:400, 40:400, 13.3:400, 4:400, 1.33:400, 0.4:400, and 0.133:400, and vice versa. Preparations of one drug alone were assayed in triplicate, and other combination ratios were assayed in duplicates. Drug-free control wells were included. All wells contained a final volume of 200 μl including tritium-labeled hypoxanthine, were standardized to 1.5% parasitemia at 1.5% hematocrit, and were incubated for 48 h and processed as described above.

Study site and sample collection.

Plasma concentrations of DBL and lumefantrine were measured in samples taken from children participating in an intervention trial carried out in the Madang Province of Papua New Guinea (PNG) (12). In brief, children aged 0.5 to 5 years and with uncomplicated falciparum or vivax malaria were randomized to one of four treatments, one of which was artemether-lumefantrine (Coartem; Novartis Pharma, Basel, Switzerland) at a dose of 10 mg/kg of body weight lumefantrine twice daily for 3 days (total dose of 60 mg/kg). Scientific and ethical approval was obtained from the Medical Research and Advisory Committee of the Ministry of Health of PNG, and informed consent was obtained from the parents or legal guardians before recruitment and blood sampling. A venous blood sample was taken at 7 days posttreatment for a drug assay, since this is regarded as a useful surrogate marker of the area under the plasma concentration-time curve and the time above the MIC for long-acting antimalarial drugs (19). Whole blood was centrifuged on-site, and separated plasma was stored at −80°C until analysis. For the purposes of the present study, the therapeutic response was considered to be either an adequate parasitological and clinical response (ACPR) or treatment failure (early treatment failure, late parasitological failure, or late clinical failure) during a follow-up of 42 days without PCR correction for reinfection (12, 29).

Drug assays.

Lumefantrine was analyzed by using a validated high-performance liquid chromatography assay (14). Plasma samples of 1.0 ml were spiked with atovaquone as an internal standard and mixed with 7 ml of hexane-diethyl ether (70:30). After centrifugation, the organic layer was separated, evaporated, and reconstituted with 200 μl methanol-acetic acid (98:2). Aliquots of 15 μl were injected onto a Phenomenex C₆-phenyl 4.6- by 150-mm column (Phenomenex, Torrance, CA). A mobile phase of acetonitrile-0.05 M phosphate buffer at pH 2.0 (62:38) with 0.03 M sodium perchlorate was pumped at 1 ml/min. Lumefantrine and atovaquone were measured at 335 nm and quantified by using Chemstation software (version 9; Agilent Technology, Waldbronn, Germany). The linear range for lumefantrine was 20 to 20,000 ng/ml. The between-day variabilities were 4.94%, 4.93%, 7.16%, and 11.23% and intraday variabilities were 2.83%, 4.41%, 4.11%, and 9.55% for 20,000, 2000, 200, and 20 ng/ml, respectively.

DBL was analyzed by using a validated ultra-high-performance liquid chromatography-tandem mass spectrometry (UPLC-LCMS-MS) method using a hexyl analogue as an internal standard (IS). To facilitate protein precipitation, 40 μl of 0.1 M ZnSO₄ was added to a 20-μl sample and briefly vortexed. A 200-μl aliquot of acetonitrile containing the internal standard was added before the sample was centrifuged, and 10 μl of the supernatant was injected on to a 2795/Quattro Premier instrument (Waters Corp., MA) with a Waters XTerra mass spectrometry C₁₈, 2.5- by 50-mm, 3.5-μm column. Gradient elution was performed by using aqueous 2 mM ammonium acetate-0.1% formic acid and methanolic 2 mM ammonium acetate-0.1% formic acid as the mobile phases at 0.4 ml/min. Transitions were monitored by using positive electrospray ionization with multiple-reaction monitoring for DBL and the IS, which were m/z 472.1 and 346.0 and m/z 500.1 and 346.0, respectively. The linear range for DBL was 0.5 to 100 ng/ml, with the lower end taken as the limit of quantitation. The between-day variabilities were 3.36%, 3.47%, 9.98%, and 6.74% and the intraday variabilities were 2.47%, 3.46%, 8.16%, and 3.48% for 50, 10, 1, and 0.5 ng/ml, respectively. When matrix effects were assessed, the between-subject variabilities were 3.37%, 4.47%, and 9.43% at 50, 10, and 1 ng/ml, respectively.

Statistical analysis.

Drug susceptibility and interactions were analyzed by a nonlinear regression of logarithmically transformed concentrations. The concentration that inhibited 50% parasite growth (IC₅₀) was determined for each drug, as was the concentration that inhibited 99% of growth (IC₉₉). The fractional inhibitory concentrations (FICs), representing the concentration of each drug alone or in combination resulting in 50% inhibition, were used to construct isobolograms (2). We used two analytical approaches (6). First, the sum of the fractional inhibitory concentrations (ΣFICs) was calculated using the formula (IC₅₀ of drug A in a mixture resulting in 50% inhibition/IC₅₀ of drug A alone) + (IC₅₀ of drug B in a mixture resulting in 50% inhibition/IC₅₀ of drug B alone) (2). The combination has an indifferent interaction when the ΣFIC is close to 1.0. An ΣFIC of <1 indicates synergy, with data points forming a concave isobole beneath the line of additivity. An ΣFIC of >1 indicates antagonism, as represented by a convex isobole (2, 5, 6, 10). Second, the function y = 1 − x/[x + (1 − x) × exp(−I)] (3, 4) was fitted to the data, where y is the IC₅₀ of drug A combined with drug B, x is the IC₅₀ of drug B combined with drug A, and I is the interaction value. Positive I values indicate synergy, negative values indicate antagonism, and values close to zero indicate an indifferent interaction. Values of I and FIC were considered significantly different from no interaction if both 95% confidence intervals (CIs) of the estimates did not span zero and unity, respectively (6).

RESULTS

The geometric mean IC₅₀s and 95% confidence intervals (CIs) for DBL, lumefantrine, and other drugs are shown in Table 1. W2mef had, as expected, a higher CQ IC₅₀ than did 3D7, but there were no other strain-specific differences. Consistent with the IC₅₀ data, the geometric mean IC₉₉ values for DBL against 3D7 and W2mef were 78 and 56 nM, respectively, and those for lumefantrine were 239 and 226 nM, respectively. Based on both the IC₅₀ and IC₉₉ values, the in vitro activity of DBL was at least 3 times that of lumefantrine regardless of CQ sensitivity.

TABLE 1.

In vitro sensitivity of laboratory-adapted P. falciparum to desbutyl-lumefantrine and other antimalarial drugs^a

Drug	3D7		W2mef
	Geometric mean IC50 (nM)	95% CI	Geometric mean IC50 (nM)	95% CI
Chloroquine	29.6	15.3-57.5	171.4	144.4-203.4
Lumefantrine	65.2	42.3-100.8	55.5	40.6-75.7
Desbutyl-lumefantrine	9.0	5.7-14.4	9.5	7.5-11.9
Piperaquine	16.9	13.4-21.3	17.4	11.8-25.6
Dihydroartemisinin	4.2	2.5-7.0	3.1	1.8-5.3

^aData represent at least six experiments performed in triplicate.

The isobolographic analysis of DBL and lumefantrine combinations showed no interaction in both laboratory-adapted strains (Table 2 and Fig. 1). Drug interactions between DBL and dihydroartemisinin were mildly synergistic, as assessed by both the I and ΣFIC.

TABLE 2.

In vitro efficacy of antimalarial drug combinations against P. falciparum clones 3D7 and W2mef assessed by isobolographic analysis^a

Strain and drug	I (95% CI)	ΣFIC (95% CI)	Interaction
3D7
DBL-lumefantrine	0.41 (−0.24-1.05)	0.99 (0.93-1.05)	Indifferent
DBL-dihydroartemisinin	0.99 (0.71-1.28)	0.92 (0.87-0.98)	Mildly synergistic
W2mef
DBL-lumefantrine	0.79 (0.02-1.56)	1.06 (0.97-1.14)	Indifferent
DBL-dihydroartemisinin	0.92 (0.73-1.10)	0.94 (0.90-0.99)	Mildly synergistic

^aData shown are the interaction factor (I) and the summed fractional inhibitory concentration (ΣFIC) with 95% confidence intervals. The assessment of interactions is based on both I and ΣFIC data (see the text).

FIG. 1.

(Top) Isobolograms showing the effect of desbutyl-lumefantrine in combination with lumefantrine against P. falciparum 3D7 (left) and W2mef (right). (Bottom) Effect of desbutyl-lumefantrine in combination with dihydroartemisinin against P. falciparum 3D7 (left) and W2mef (right). The isoboles are representative of data from three or four experiments in which each of the 17 drug ratios was tested in duplicate.

We quantified both DBL and lumefantrine in 94 available day 7 serum samples from the 127 children with falciparum malaria recruited to the artemether-lumefantrine arm of the intervention trial (12). The mean DBL concentration was 15.5 ng/ml (range, 0.6 to 58.2 ng/ml) or 31.9 nM (range, 1.3 to 123.1 nM); all children had a plasma concentration above the 0.5-ng/ml lower limit of quantitation. For lumefantrine, the mean plasma concentration was 370 ng/ml (range, 26 to 1,720 ng/ml) or 699 nM (range, 49 to 3,251 nM). The lumefantrine-to-DBL ratio ranged from 7.0 to 123.0, with a mean of 27.4.

The relationship between plasma lumefantrine and DBL concentrations and treatment outcome at 42 days is shown in Fig. 2. The mean plasma concentrations of lumefantrine in those children who failed treatment were lower than in those with an ACPR, but this difference was not significant (P = 0.22). In the case of DBL, there was a similar result but at borderline significance (P = 0.053). There was no difference in the case of the lumefantrine-to-DBL ratio (P = 0.97).

FIG. 2.

Box plots summarizing plasma lumefantrine (left) and desbutyl-lumefantrine (right) concentrations in children who had an adequate clinical and parasitological response (ACPR) or who failed treatment with artemether-lumefantrine.

DISCUSSION

The present observations confirm and extend previously published data relating to in vitro and in vivo aspects of the antimalarial activity of DBL. The two laboratory-adapted strains of P. falciparum were more sensitive in vitro to DBL than to lumefantrine, consistent with data from previous studies using field isolates (17, 23, 24). This effect was independent of CQ sensitivity. The lack of an in vitro interaction between DBL and lumefantrine appears to contradict previous reports of synergy in field isolates of P. falciparum using limited numbers of drug combinations (23, 24). However, the synergy between DBL and dihydroartemisinin that we observed using dual methods of isobolographic data analysis parallels data from studies of field isolates of P. vivax (13) and P. falciparum (16). We were able to detect both lumefantrine and DBL at mean concentrations that were well above the IC₉₉ for both laboratory-adapted strains of P. falciparum in day 7 plasma of children treated with artemether-lumefantrine for uncomplicated malaria. The day 7 plasma DBL was a stronger predictor of subsequent therapeutic outcomes than plasma lumefantrine or the plasma lumefantrine-to-DBL ratio. These various observations have potential clinical implications.

When artemether-lumefantrine is administered to patients with malaria, plasma lumefantrine concentrations rise after each of the six doses given over 3 days and then decline, with an elimination half-life of around 4 days (8). There are no equivalent pharmacokinetic data for DBL as yet, but our day 7 plasma concentrations show that both plasma lumefantrine and DBL concentrations remain above the IC₉₉ in most patients for at least three parasite life cycles. Even if there was synergy between lumefantrine and DBL at these concentrations, as found previously by others (23, 24), its relevance at this stage in treatment is unclear, as initial parasite clearance had occurred in all patients allocated this therapy in the trial (12), and the presence of one or another compound at a concentration greater than the IC₉₉ should inhibit low-level (submicroscopic) parasite replication. In addition, the variable day 7 lumefantrine-to-DBL ratios in the children in our study, which are consistent with the marked apparent between-dose variability in lumefantrine bioavailability (9, 28), would lead to inconsistent interactions. Indeed, the data relating day 7 lumefantrine and DBL concentrations to therapeutic outcome suggest that DBL has a stronger role than the parent compound in suppressing recrudescence and/or preventing reinfection. This may reflect the fact that, even at relatively low plasma concentrations, DBL has more potent antimalarial activity than lumefantrine, as exemplified by our in vitro sensitivity data. The fact that the lumefantrine-to-DBL ratio had no relationship with outcome argues against a clinically important synergistic interaction.

Initial parasite clearance is considered to be due primarily to the artemisinin component of artemisinin combination therapy (ACT), and, conversely, its prolongation is taken as evidence of artemisinin resistance (7). It is, however, possible that the longer-half-life partner and interactions between the component drugs enhance the relatively rapid parasiticidal effects of artemisinin derivatives. Synergy between artemether and lumefantrine has been reported (1), and the present data and those described previously by others (13, 16) suggest a similar interaction for DBL and dihydroartemisinin. However, the clinical importance of such effects is uncertain. The synergy that we observed was mild. The relatively short half-lives of artemether and its active metabolite dihydroartemisinin (1 to 2 h) (8, 9) limit the time window for such effects. Interactions can be antagonistic, such as those between artemisinin and DBL at some concentrations (13) and between dihydroartemisinin and 4-aminoquininolines and related drugs (6). The fact that a range of different ACTs show equal efficacy (22) suggests that acute effects on the parasite other than those from the artemisinin drug are minor.

According to the manufacturer, DBL is a degradation product of lumefantrine, and artemether-lumefantrine tablets contain <0.1% of DBL both at the time of manufacture and at the end of the expiry date (H. Gruninger, Novartis Pharma, personal communication). We have confirmed the occurrence of such low relative concentrations of DBL in pure lumefantrine powder by UPLC-LCMS-MS (our unpublished data) and excluded DBL-lumefantrine at 1:10,000 from the drug interaction analysis as a result. Although a small amount of DBL is present at the time of artemether-lumefantrine administration, it is quantitatively far below the level detected in plasma. Thus, this exogenous contribution is insufficient to influence the plasma DBL concentrations in the present study, with the corollary that DBL is a true metabolite.

The present study confirms that DBL has potential as an antimalarial drug in its own right. Its in vitro potency relative to that of the parent compound, its synergy with dihydroartemisinin, and the positive relationship between day 7 plasma concentrations and ACPR suggest that it could be a useful alternative to lumefantrine as a part of ACT. Although the presence of DBL at parasiticidal concentrations during conventional artemether-lumefantrine therapy suggests that it is as safe as lumefantrine, further pharmacokinetic and safety assessments after DBL administration would be required to facilitate the development of optimal dose regimens. In contrast to N-desbutyl-halofantrine, the active metabolite of halofantrine (25), preliminary in vitro cardiotoxicity studies have not raised any significant concerns (26).