Abstract
Artemisone is one of the most promising artemisinin derivatives in clinical trials. Previous studies with radiolabeled artemisinin and dihydroartemisinin have measured uptake in Plasmodium falciparum-infected erythrocytes. Uptake is much greater in infected than in uninfected erythrocytes, but the relative contributions of transport, binding, and metabolism to this process still await definition. In this study, we characterized mechanisms by which [14C]artemisone is taken up into uninfected and P. falciparum-infected human erythrocytes in vitro. Radiolabeled artemisone rapidly enters uninfected erythrocytes without much exceeding extracellular concentrations. Unlabeled artemisone does not compete in this process. Radiolabeled artemisone is concentrated greatly by a time- and temperature-dependent mechanism in infected erythrocytes. This uptake is abrogated by unlabeled artemisone. In addition, the uptake of artemisone into three subcellular fractions, and its distribution into these fractions, is examined as a function of parasite maturation. These data are relevant to an understanding of the mechanisms of action of this important class of drugs.
Artemisinins are in the frontline of treatments for malaria (14), a disease that is responsible for nearly a million deaths annually. These sesquiterpenes contain a peroxide bridge that is essential for their antimalarial activity, although their exact mechanism of action is still a controversial subject (14). Resistance to artemisinins by different mechanisms in laboratory-derived plasmodial parasites has been developed by several groups (e.g., see references 3 and 28). More worryingly, and despite a policy of using artemisinins in combination with other antimalarials, there has been a recent emergence of parasites that are also becoming less susceptible to artemisinins by unknown mechanisms in patients in Western Cambodia (6, 19). This disturbing development has increased the urgency for researchers to understand better how these drugs kill parasites.
One aspect worthy of consideration in determining how artemisinins work is their uptake (i.e., transport and any additional accumulative processes) into parasitized erythrocytes. The bulk uptake of artemisinin (25) and dihydroartemisinin (DHA) (1, 9) in uninfected erythrocytes is most likely to occur through passive diffusion, rapidly reaching concentrations similar to those in the external medium. In Plasmodium falciparum-infected erythrocytes, the uptake of artemisinins is much greater than that in uninfected erythrocytes and occurs via a concentrative mechanism that is saturable, temperature sensitive, and largely irreversible (1, 9, 25).
Currently planned for use in clinical trials where decreased susceptibility to artesunate has been reported (for more information, see the Medicines for Malaria Venture website at http://www.mmv.org/), artemisone is a promising, “second-generation,” semisynthetic artemisinin derivative with a longer half-life, lower curative dose (18), and superior bioavailability compared with all other derivatives. It displayed no neurotoxicity in preclinical testing (21). Here, we examine the uptake of radiolabeled (15-14C) artemisone (Fig. 1) in P. falciparum-infected and uninfected human erythrocytes. This includes elucidating the time course of uptake into uninfected and infected erythrocytes (the latter at both the immature and mature trophozoite stages) and an analysis of the subcellular fractions in which artemisone accumulates. Insights gained from this work help our understanding of how these drugs may exert their antimalarial actions.
FIG. 1.
Structure of [15-14C]-artemisone. The asterisk indicates the radiolabel.
MATERIALS AND METHODS
Materials.
Triton X-100 (Triton), dibutyl phthalate (DBP), trichloroacetic acid (TCA), hydrogen peroxide, and EDTA were obtained from Sigma (Poole, United Kingdom). [15-14C]artemisone, containing 42.1% 14C and with a specific activity of 2.42 MBq/mg (973 MBq/mmol, or 26.3 mCi/mmol), was prepared at Bayer Pharma (Wuppertal, Germany) according to a method developed at the Hong Kong University of Science and Technology for preparing deuterated artemisone (12). Unlabeled artemisone was prepared according to a method reported previously (12). Ecoscint was obtained from National Diagnostics (Hessle, United Kingdom), and Picofluor 40 and Solvable were obtained from Perkin-Elmer (Beaconsfield, United Kingdom).
In vitro culture of P. falciparum.
P. falciparum (line A4 [20])-infected erythrocytes were cultured and harvested as described previously (22). By using 65% (vol/vol) Percoll, the harvesting of mature parasites (35 to 40 h postinvasion) resulted in parasitemias in excess of 70%. For experiments comparing immature and mature parasites, the same culture, following sorbitol synchronization, was used on two consecutive days (taking advantage of the 48-h intraerythrocytic life cycle of P. falciparum). As immature parasites are difficult to harvest, the culture was used without a concentrative process, giving parasitemias of approximately 10%. Cell counts were made by using an improved Neubauer counting chamber. Parasitemia was estimated from methanol-fixed, Giemsa-stained smears.
Uptake assays.
Uptake time courses were initiated by the addition of a 50-μl aliquot of radiolabeled artemisone in RPMI 1640 culture medium to 950 μl of a 0.5 to 2% hematocrit erythrocyte suspension (5 to 10% for experiments comparing uptake in immature infected erythrocytes to that in mature infected erythrocytes), also in RPMI 1640 culture medium (both preincubated to the appropriate temperature), in 1.5-ml microcentrifuge tubes. The amount of radiolabeled drug was varied to give final concentrations of 20, 4, or 0.4 μM, without altering the specific activity (i.e., only radiolabeled artemisone was used and not unlabeled drug). At each time point, a 100-μl sample was layered over 300 μl DBP (chilled to 4°C) in a microcentrifuge tube and spun down (10,000 × g for 10 s), separating the cells from the remaining radiolabeled drug in solution (and terminating the uptake). In experiments involving the competition of radiolabeled and unlabeled artemisone, the unlabeled drug was added to the experimental tubes to a concentration of 100 or 500 μM 10 min prior to experimentation. At the end of each experiment, a further 100-μl sample was taken to determine the total amount of radiolabel. Alternatively, a “dummy” experiment was set up (i.e., in the absence of cells), and a sample was taken.
Preparation of samples.
The aqueous supernatant solution in each sample was removed by aspiration, and the radioactivity remaining on the tube's walls was removed by rinsing the tubes four times with water. The DBP was aspirated, and the cell pellets were then lysed with 0.1% (vol/vol) Triton (0.1 ml) and deproteinized by the addition of 5% (wt/vol) TCA (0.1 ml), followed by centrifugation (10,000 × g for 10 min). The supernatants (containing the “Triton-soluble, non-TCA-precipitable” fraction) were poured off into scintillation vials containing 4 ml of Ecoscint scintillation fluid. The remaining pellets contained the “Triton-insoluble” and the “Triton-soluble, TCA-precipitable” fractions. For experiments comparing uptakes in immature and mature infected erythrocytes, the samples were centrifuged after the addition of Triton so as to pellet the Triton-insoluble fraction. In this case, the supernatants were moved into fresh tubes before the addition of TCA and subsequent centrifugation to separate the Triton-soluble, TCA-precipitable (pellet) and the Triton-soluble, non-TCA-precipitable (supernatant) fractions.
Treatment of pellets.
Pellets were washed in Triton or TCA, as appropriate, to remove the remaining supernatant. Washed samples were heated to 60°C for 1 h with 1 ml of Solvable (to dissolve the pellets). The samples were then transferred into 20-ml scintillation vials, and 100 μl EDTA (0.1 mM) (to prevent foaming) and 100 μl hydrogen peroxide (70%, vol/vol, to bleach the samples) were added. The samples were left for 15 min before being heated to 60°C for 1 h while bleaching occurred. Subsequently, 15 ml of Picofluor 40 scintillation fluid was added to each tube. Radioactivity was measured by using a β-scintillation counter (Tri-Carb; Perkin-Elmer, Beaconsfield, United Kingdom).
Parasite growth assays.
The efficacy of unlabeled and radiolabeled artemisone was determined by an in vitro growth assay using [3H]hypoxanthine incorporation as a measure of parasite growth (5). Serial dilutions of artemisone (90 μl in culture medium) were dispensed into 96-well plates, to which tightly synchronized immature (∼15 h postinvasion) P. falciparum-infected erythrocytes (100 μl at 2% hematocrit and 2% parasitemia in culture medium) were added. The plates were then placed into a gas-tight box under 3% O2, 7% CO2, and 90% N2 at 37°C for 24 h. At this point, [3H]hypoxanthine (10 μl at 0.3 μCi/well in culture medium) was added so that labeling could occur over a subsequent 24-h period. Plates were then harvested, and the incorporated radioactivity was measured by using a 1450 MicroBeta Plus scintillation system (Wallac, Turku, Finland).
RESULTS
Artemisone uptake in uninfected erythrocytes.
Initially, the uptake of [14C]artemisone in uninfected erythrocytes was studied. At the end of each uptake experiment, cell samples taken at each time point were divided into supernatant (containing lipids, low-molecular-weight peptides, solutes predominantly) and pellet (containing protein predominantly in anucleate human erythrocytes) fractions after lysis in Triton and deproteinization with TCA (see Materials and Methods). At an initial starting concentration of 20 μM [14C]artemisone, uninfected erythrocytes took up artemisone into the supernatant fraction to a concentration of 10 to 15 μM (assuming that 1013 erythrocytes equates to a volume of approximately 1 liter), a value just below the (initial) extracellular concentration, within the first (15 s) time point (Fig. 2). The measured uptake of radiolabeled artemisone was not affected significantly by time (up to 180 min), the presence of unlabeled artemisone (500 μM), or a reduction of the experimental temperature from 37°C to 4°C (P > 0.05 by analysis of variance [ANOVA] with Tukey's posttest). In the pellet fraction, [14C]artemisone uptake increased significantly compared with that of the supernatant fraction (P = 8 × 10−14 by an unpaired, two-tailed, Student's t test), by approximately 4-fold at the initial 15-s time point. As with the supernatant fraction, [14C]artemisone uptake was not affected significantly by time, the presence of unlabeled artemisone, or temperature reduction (P > 0.05 by ANOVA with Tukey's posttest). [14C]artemisone uptake into the pellet fraction equated to approximately 2.6 × 106 molecules per cell under the conditions used here.
FIG. 2.
Effect of time (up to 180 min), temperature (37°C versus 4°C), and presence of unlabeled drug (500 μM) on the uptake of radiolabeled artemisone in uninfected erythrocytes. Cell samples were divided into supernatant (open bars) and pellet (closed bars) fractions after lysis in Triton and deproteinization with TCA. The initial concentration of artemisone was 20 μM. Data are derived from 3 independent experiments and are shown as means ± SEM.
At an initial starting concentration of 4 μM [14C]artemisone, essentially identical data were obtained for the uninfected erythrocyte fractions as described above but with 5-fold-lower values (data not shown). At an initial starting concentration of 0.4 μM, the uptake of [14C]artemisone was not measurable in either uninfected erythrocyte fraction due to the low level of radiolabel present.
Artemisone uptake in infected erythrocytes.
By using the same separation protocol as that described above (note that in these experiments, the pellet fraction, in addition to protein, will contain DNA, RNA, and insoluble hemozoin predominantly), the uptake of [14C]artemisone in mature (∼39 h postinvasion) P. falciparum-infected erythrocytes was studied. At initial concentrations of 20 and 4 μM radiolabeled drug, uptake was greater in both fractions of infected erythrocytes compared with those of the supernatant and pellet fractions of uninfected erythrocytes. In contrast to experiments with uninfected erythrocytes, the uptake of artemisone into both fractions of mature infected erythrocytes was measurable at an initial starting concentration of 0.4 μM. At each concentration, approximately 90% of the total artemisone taken up was found in the pellet fraction. In all cases, the presence of unlabeled artemisone (up to 500 μM) and/or an experimental temperature reduction from 37°C to 4°C resulted in substantially less uptake of radiolabeled artemisone (as shown in Fig. 3 for an initial starting concentration of 0.4 μM). These actions left minor components of (nonspecific) artemisone uptake that were similar to those measured for uninfected erythrocyte fractions at a given initial artemisone concentration. Using Avogadro's number (∼6 × 1023) and assuming 1013 erythrocytes approximates to 1 liter, cellular concentrations of [14C]artemisone can be calculated by dividing the uptake value (in millions of molecules per cell) by 0.06, giving values in μM. By back calculation, 0.4, 4, and 20 μM initial starting concentrations of radiolabeled artemisone equate to 0.024, 0.24, and 1.2 million molecules per cell, respectively, under the conditions used here. These values would be expected if artemisone uptake into infected erythrocytes reached equilibrium in the absence of a concentrative process. However, at all initial starting concentrations and for all fractions of infected erythrocytes, the values were considerably higher than those expected for a nonconcentrative process, demonstrating that artemisone is accumulated greatly in infected erythrocytes. This finding is supported further by the large reduction of external radiolabel during uptake experiments using infected erythrocytes (see below).
FIG. 3.
Effect of time (up to 90 min), temperature (37°C versus 4°C), and presence of unlabeled drug (100 μM) on the uptake of radiolabeled artemisone in mature infected erythrocytes. Cell samples were divided into supernatant and pellet fractions after lysis in Triton and deproteinization with TCA, with the data presented for the pellet fraction only. Data are corrected for 100% parasitemia. The initial concentration of artemisone was 0.4 μM. Circles, 37°C; triangles, 4°C; squares, 37°C and in the presence of unlabeled artemisone (100 μM); inverted triangles, 4°C and in the presence of unlabeled artemisone (100 μM). Data are derived from 3 independent experiments and are shown as means ± SEM.
Figure 4 shows the specific uptake of artemisone into the pellet fraction of mature infected erythrocytes over time at the three drug concentrations used. The specific uptake was calculated by subtracting the nonspecific uptake (defined as uptake at 4°C in the presence of unlabeled artemisone) from the total uptake measured at 37°C. Using a starting concentration of 20 μM [14C]artemisone, uptake reached a steady state only toward the final (180-min) time point. At starting concentrations of 4 and 0.4 μM [14C]artemisone, uptake reached a steady state at around 60 and 10 min, respectively. By using an exponential (rise to maximum) equation to fit the data [y = a (1 − e−bx)], the uptake maximum (a) and rate constant (b) were derived for each concentration in each fraction (Table 1). Maximum values of uptake into the pellet fraction increased 10-fold when the initial [14C]artemisone concentration was increased from 0.4 to 4 μM. However, the maximum uptake increased by less than 4-fold when the initial [14C]artemisone concentration was increased from 4 to 20 μM, suggesting that a saturation of specific uptake might have been reached. The rate constants were inversely related to the starting concentration. An essentially identical trend was observed for the specific uptake of artemisone into the supernatant faction of mature infected erythrocytes, although the maximum uptake values were approximately 10-fold lower (Table 1). Note that the rate constants were not statistically different for uptake into the supernatant fraction compared with uptake into the pellet fraction of mature infected erythrocytes at 0.4, 4, and 20 μM [14C]artemisone starting concentrations (P = 0.74, 0.15, and 0.20, respectively, by an unpaired, two-tailed, Student's t test).
FIG. 4.
Effect of starting concentration on the specific uptake of radiolabeled artemisone in mature infected erythrocytes over time. Cell samples were divided into supernatant and pellet fractions after lysis in Triton and deproteinization with TCA, with the data presented for the pellet fraction only. The specific uptake was calculated by subtracting the nonspecific uptake (defined as accumulation at 4°C in the presence of unlabeled artemisone) from the total uptake measured at 37°C. Data are corrected for 100% parasitemia. The initial radiolabeled artemisone concentrations used were 20 μM (triangles), 4 μM (squares), and 0.4 μM (inset) (circles). Data are derived from 3 or 4 independent experiments and are shown as means ± SEM.
TABLE 1.
Maximum specific artemisone uptake and rate of uptake into infected erythrocytes (parasitemia corrected) at a range of initial external concentrations of radiolabeled artemisone
| Concn (μM) | Mean uptake max (million molecules per cell) for parasite fraction ± SEM |
Mean rate constant (h−1) for parasite fraction ± SEM |
||
|---|---|---|---|---|
| Pellet | Supernatant | Pellet | Supernatant | |
| 0.4 | 2.7 ± 0.7 | 0.20 ± 0.06 | 17 ± 2 | 16 ± 3 |
| 4 | 30 ± 5 | 2.7 ± 0.3 | 2.3 ± 0.2 | 3.6 ± 0.7 |
| 20 | 110 ± 20 | 10 ± 2 | 1.1 ± 0.1 | 1.5 ± 0.3 |
Maximal uptake of radiolabel.
The total uptake of radiolabel in mature infected erythrocytes was not maximal. At the steady-state maxima, only 50 to 60% of the available radiolabel was taken up during any given experiment at initial starting concentrations of 0.4 and 4 μM [14C]artemisone (note that a lower percentage of radiolabel was taken up in experiments using 20 μM radiolabeled artemisone). To investigate the ability of the radiolabel remaining in solution to be taken up into infected erythrocytes, an additional assay was developed (Fig. 5). Using an initial concentration of 0.4 μM radiolabeled artemisone, two uptake experiments with mature infected erythrocytes were set up in parallel. In the first experiment, samples were taken over time to measure radiolabel uptake (Fig. 5, solid lines and circles), in which approximately 60% of radiolabel was taken up. In the second experiment, no samples were taken, and after 20 min, when maximum uptake was complete, the infected cells were removed from suspension by centrifugation, and the remaining solution was used in a subsequent uptake experiment with fresh mature infected erythrocytes. In this experiment, the radiolabel (approximately 40% of the original amount) failed to be taken up into fresh mature infected erythrocytes (Fig. 5, long dashed lines and squares). This contrasts with a control experiment in which fresh mature infected erythrocytes were exposed to an equivalent concentration (0.16 μM) of fresh radiolabeled artemisone (Fig. 5, short dashed lines and triangles).
FIG. 5.
Maximal uptake of radiolabel in mature infected erythrocytes. Data for three uptake experiments with mature infected erythrocytes are presented. In the first experiment (solid lines and circles), the total average amount of radiolabel in each sample (horizontal solid line) and the amount associated with mature infected erythrocytes only over time are presented (shown as counts per minute [CPM]), which were determined by using a starting concentration of 0.4 μM radiolabeled artemisone (note that ∼60% of radiolabel is taken up by the cells in this example). In the second experiment (long dashed lines and squares), data are presented as described above but were determined by using medium containing 0.4 μM radiolabeled artemisone that had first been exposed to mature infected erythrocytes for 20 min before use (note that only ∼40% of the original radiolabel is measureable after this procedure and that this radiolabel is unable to be taken up into fresh mature infected erythrocytes). In the third experiment (short dashed lines and triangles), data are presented as described above for the first experiment but were determined by using a starting concentration of 0.16 μM radiolabeled artemisone, which served as a control.
To ensure that the quality of stock radiolabeled drug was not at fault, a sample was returned to the manufacturer 1 year after supply (and after the majority of the experiments presented here were performed) to be used to investigate its purity. Their analysis, using a high-pressure liquid chromatograph equipped with a Ramona 92 radiochemical detector, showed the stock to be 83% pure at that time.
The ability of the remaining radiolabel to inhibit parasite growth in vitro was also tested, in an effort to determine if the radiolabel was associated with artemisone or an inactive complex or metabolized version. In initial assays, unlabeled artemisone inhibited parasite growth with a 50% inhibitory concentration (IC50) of 1.0 ± 0.2 nM, whereas the IC50 determined for radiolabeled artemisone was 3.9 ± 0.5 nM (corrected for purity; mean ± standard error of the mean [SEM]; n = 3). By using a protocol similar to that described above, an uptake experiment with mature infected erythrocytes was set up with an initial radiolabeled artemisone concentration of 4 μM. After 1 h (a point at which maximum uptake would have occurred) (Fig. 4), the cell suspension was centrifuged to remove the cells. The supernatant, which would have at least 40% radiolabel still present (based on previous experiments) (see above and Fig. 5), was then used as a stock solution to prepare an in vitro growth assay to test its effect. In parallel, a control stock solution of fresh 1.6 μM radiolabeled artemisone (i.e., containing 40% of the radiolabel found in a 4 μM solution) was used. Note that the uptake of the radiolabeled artemisone did not affect significantly the measurement of incorporated [3H]hypoxanthine, as determined by control experiments performed in the presence of radiolabeled artemisone but in the absence of [3H]hypoxanthine (data not shown). For each of three experiments, there was at least a 10-fold increase in the IC50 for the effect of radiolabeled artemisone that had been in the presence of mature infected erythrocytes for 1 h previous on parasite growth compared with the controls (data not shown). Given that the control experiments contained the same or slightly less radiolabel, these data suggest that the remaining radiolabel in solution after a preincubation with infected erythrocytes is not associated with active artemisone.
Parasite stage.
Artemisone uptake in erythrocytes infected with parasites of different ages was investigated. Using an initial artemisone concentration of 0.4 μM, uptake was measured in immature (∼15 h postinvasion) and mature (∼39 h postinvasion) infected erythrocytes over time. At the end of each experiment, samples taken were divided into three fractions: Triton insoluble; Triton soluble, TCA precipitable; and supernatant (i.e., Triton soluble and non-TCA precipitable). Figure 6 shows the specific uptake of artemisone into the former two fractions in immature and mature infected erythrocytes over time, while the kinetic constants are presented in Table 2.
FIG. 6.
Uptake of radiolabeled artemisone in immature (A) and mature (B) infected erythrocytes. Cell samples were divided into three fractions, with data presented for the Triton-insoluble fraction (triangles); the Triton-soluble, TCA-precipitable fraction (squares); and the two combined fractions (circles) only. Data are corrected for 100% parasitemia. The initial concentration of artemisone was 0.4 μM. Data are derived from 3 independent experiments and are shown as means ± SEM.
TABLE 2.
Maximum specific artemisone uptake and rate of uptake (0.4 μM initial radiolabeled artemisone concentration) into infected erythrocytes (parasitemia corrected) of different ages
| Parasite stage and fractiona | Mean uptake max (million molecules per cell) ± SEM | Mean rate constant (h−1) ± SEM |
|---|---|---|
| Immature | ||
| TSTP | 1.2 ± 0.2 | 2.3 ± 0.7 |
| TI | NA | NA |
| Mature | ||
| TSTP | 1.6 ± 0.3 | 17 ± 4 |
| TI | 2.5 ± 0.4 | 11 ± 3 |
TSTP, Triton soluble, TCA precipitable; TI, Triton insoluble; NA, not analyzed.
In mature infected erythrocytes, more than half of the artemisone taken up was present in the Triton-insoluble fraction, with uptake into the two presented fractions occurring at the same rate (P = 0.32 by an unpaired, two-tailed, Student t test). In immature infected erythrocytes, the majority of the artemisone taken up was confined to the Triton-soluble, TCA-precipitable fraction. Total specific artemisone uptake in the combined Triton-insoluble and Triton-soluble, TCA-precipitable fractions was approximately 3-fold higher in mature infected erythrocytes than in immature infected erythrocytes, suggesting a saturation of uptake in immature infected erythrocytes. In addition, while the maximum uptakes into both immature and mature infected erythrocytes in the Triton-soluble, TCA-precipitable fraction were similar (P = 0.31 by an unpaired, two-tailed, Student's t test), the rate of uptake into immature infected erythrocytes was significantly lower (P = 0.03 by an unpaired, two-tailed, Student's t test).
DISCUSSION
Despite their importance for treating malaria and other diseases, research into the mechanisms of action of artemisinins has been a relatively neglected subject compared with other antimalarials. Here, we report how radiolabeled artemisone is distributed in uninfected and infected erythrocytes. These types of experiments have been useful for determining the uptake (and localization) of artemisinins, which is an important factor in their mechanisms of action and which can also be used to identify factors that reduce drug efficacy. For example, an artemisinin-resistant but unstable Plasmodium yoelii strain, which accumulated 43% less [3H]DHA, was 4-fold less susceptible to artemisinin (26). An increased uptake of [14C]artemisinin in uninfected α-thalassemic erythrocytes was identified as the reason behind the reduced effect of artemisinins on P. falciparum parasites cultured in these cells (reducing the degree of artemisinin available to concentrate within the parasites [13]). The increased uptake of artemisinins in α-thalassemic erythrocytes is associated with the competitive decomposition of the artemisinins by heme-Fe(II) and hemoglobin (Hb)-Fe(II) (24).
The temperature- and time-independent uptake of [14C]artemisone in uninfected erythrocytes presented here is unaffected by competition with unlabeled drug and is directly proportional to the starting concentration. The uptake is consistent with a nonconcentrative, bulk diffusion component (observed for the supernatant fraction after Triton lysis and TCA precipitation) that proceeds with an additional nonspecific concentrative component of ∼3-fold above the external drug concentration (observed for the TCA-precipitated fraction after Triton lysis and, thus, most likely associated with precipitated proteins). Note that while data for the former component are consistent with diffusion, it is not possible to rule out that the drug is in a sequestered form. These observations are consistent with those described previously by Gu et al. (9), who reported that [3H]DHA was concentrated in uninfected cells by slightly less than 2-fold above the external concentration, in keeping with a small concentrative uptake component, and with those described previously by Vyas et al. (25), who concluded that artemisinin enters uninfected cells via passive membrane diffusion.
Additional mechanisms of artemisone distribution are observed after infection of erythrocytes by P. falciparum. Infected erythrocytes concentrate [14C]artemisone by a saturable, competitive, and time- and temperature-dependent mechanism. This process is largely independent of the energy status of the parasites (data not shown), consistent with previous work (13) and suggestive of a nonactive uptake mechanism. The fact that the majority of radiolabeled artemisone was recovered from pellet fractions that were repeatedly washed suggests the artemisone is covalently bound to precipitable (i.e., protein) or insoluble material, although other forms of interaction cannot be ruled out without additional experimentation. Covalent binding would be best described by irreversible binding kinetics. In this case, binding will carry on until all available binding sites are occupied or the available drug is bound fully. At 0.4 and 4 μM initial artemisone concentrations in the presence of mature infected erythrocytes, the evidence suggests that no “active” drug is left in the medium, even though 40 to 50% of the radiolabel is still present (associated with inactive versions of artemisone) (see “Maximal uptake of radiolabel” in Results and below). In line with this, there is a 10-fold increase in the maximal artemisone uptake as the initial concentration is increased from 0.4 to 4 μM (Table 1). This suggests that the latter is true (i.e., all active drug is bound). However, at a 20 μM initial artemisone concentration in the presence of mature infected erythrocytes as well as at a 0.4 μM initial artemisone concentration in the presence of immature infected erythrocytes, there is a deviation away from the expected maximal uptake (i.e., at 20 μM, the expected maximal uptake would be 5 times that measured at 4 μM, but the observed uptake is only 3-fold) (Table 1). This suggests that “active” artemisone is still present and that the former is true (i.e., all artemisone binding sites are occupied) and, therefore, that saturation has been reached.
The distributive properties of artemisone are dependent on the stage of parasite maturation. Immature (ring-stage) parasites are highly susceptible to the action of artemisinins in vitro as well as in vivo (2, 23). These early stages do not display visible (insoluble) pigment. Consistent with this, most of the artemisone taken up is in the pellet fraction of ring-stage-infected erythrocytes after Triton lysis, the removal of any insoluble components, and subsequent TCA precipitation. This fraction contains most of the cell's protein, DNA, and RNA, but as previous work has shown a negligible binding of artemisinin to DNA and RNA (29), artemisone is most likely to be associated with protein. If this uptake is assumed to be related directly to parasite killing, then at an IC50 of 4 nM, this amount of artemisone equates to 12,000 molecules per cell (or 3,000 molecules per cell for an IC50 of 1 nM, assuming that 400 nM artemisone distributes as 1.2 million molecules per immature infected cell from our measurements, which is at saturation). These figures would have to be doubled to approximate IC90 values. These are relatively low numbers of drug molecules per parasite, suggesting that artemisone's activity is potent at the target(s) and may be confined to a limited number of targets at this stage of parasite development.
Erythrocytes carrying mature (pigmented) stages of parasite development also show a time- and temperature-dependent uptake of artemisone. However, this is highly concentrative compared with uptake in uninfected cells, and the majority of artemisone (∼55%) is associated with the Triton-insoluble pellet (predominantly hemozoin but also including associated heme and protein and detergent-insoluble membrane regions). This finding is highly consistent with data from a previous study which showed that 60% of [14C]artemisinin accumulation occurs in the Triton-insoluble fraction of P. falciparum-infected erythrocytes (17). A further 35% of uptake occurs in the pellet derived after subsequent precipitation with TCA, and 10% is detected in the remaining supernatant fraction. Artemisone uptake demonstrates similar kinetic behaviors in all three fractions, suggesting a common mechanism. As noted above, the saturation of total uptake was likely observed with an initial extracellular artemisone concentration of 20 μM, which equates to 110 million molecules per cell. This is nearly 100-fold greater than that observed for infected cells parasitized with immature parasites and 10,000-fold greater than that required to kill parasites. The latter finding suggests that mature infected erythrocytes can act as a sump for artemisinins (removing bioactive drug). This effect might be further compounded by the observation that mature infected erythrocytes take up artemisone much faster than do the more sensitive ring-stage-infected erythrocytes.
The results presented here are in general agreement with results previously reported for [3H]DHA (1, 9) and [14C]artemisinin (25) uptake. Gu et al. (9) showed previously that both artemether and unlabeled DHA inhibit the uptake of [3H]DHA, supporting the idea that the artemisinins are taken up through a common pathway, such that data from work on one derivative is relevant to work on others. The similar uptake processes reported for [14C]artemisinin, [3H]DHA, and now [14C]artemisone add further support to this hypothesis. However, two of these studies hypothesized that artemisinin uptake was consistent with carrier-mediated transport (1, 25). Results for artemisone uptake do little to support the presence of carrier-mediated artemisone transport. First, given that uptake into uninfected erythrocytes is consistent with diffusion, it would be surprising if the same process did not occur in infected erythrocytes. Second, given the energy-independent, concentrative process of uptake, the data are more suggestive of a mechanism that traps artemisone within the parasite by metabolism/decomposition and/or by specific, tight, or irreversible binding to cellular components (the latter being the most likely endpoint, given the location of the majority of radiolabel in insoluble or precipitated fractions). By whatever mechanism it occurs, the concentration of artemisinins in infected cells could contribute to their selective toxicity for parasites.
Overall, Hb-Fe(II), heme-Fe(II), and heme-Fe(III) are unlikely to be associated with artemisone's mode of action, since artemisinins are highly active against hemozoin-poor ring-stage parasites (2, 23) and also act against parasites that do not degrade hemoglobin, such as Toxoplasma gondii and Babesia spp. (7, 8). Interestingly, while chloroquine (an accepted inhibitor of hemozoin formation) is only a fewfold less potent than DHA against sensitive P. falciparum strains in vitro, it is nearly 200 times less potent than DHA against Babesia gibsoni (16). Furthermore, artemisone itself is not readily decomposed in model studies involving heme-Fe(II) under aqueous conditions (10). Other artemisinins (artemisinin, DHA, and artesunate) that are susceptible to decomposition by Hb- or heme-Fe(II) display enhanced activities against malaria parasites cultured under CO, a reagent that blocks any reaction of Hb- and heme-Fe(II) with the artemisinins by virtue of the formation of stable Fe(II)-CO complexes (4). Therefore, the role of heme-Fe(II) within the malaria parasite is to decompose the susceptible artemisinins (10).
Artemisinins have also been suggested to act by inhibiting P. falciparum ATP6 (PfATP6), a protein that is expressed at all asexual stages of the life cycle (14, 15). Other hypothesized targets for artemisinins include mitochondria (27), and very recently, they have been shown to act as efficient oxidizers of reduced flavin cofactors associated with disulfide reductases crucial for redox homeostasis in the parasite cytosol. The hypothesis is validated by the observation that artemisinin accelerates the turnover of NADPH during the functioning of glutathione reductase (11). As the artemisinins are converted into benign reduction products after reduction, these are likely to remain in the cytosol and not be sequestered by heme or proteins. The cofactor model is also consistent with the observation here that young ring-stage parasites do not take up radiolabel into insoluble fractions. In mature parasites, competing degradation pathways that are not associated with this mechanism may enhance the extent of uptake. The production of benign reduction products is consistent with the 40 to 50% of radiolabel left in the supernatant of experiments performed under conditions of maximal uptake with no saturation of specific binding sites (i.e., the uptake of low concentrations of artemisone into mature infected cells) that is not able to be taken up into fresh infected cells or kill parasites (and which cannot be explained by autoradiolysis).
Our methodologies lay the foundation for future studies to characterize the various molecules with which artemisinins associate, such as those in the Triton-soluble, TCA-precipitated fraction, and to investigate their interactions with these drugs (e.g., it will be interesting to determine the level and locations of uptake in artemisinin resistance models).
Acknowledgments
This work was supported by the Wellcome Trust (grant no. 077441), by European Commission projects ANTIMAL (grant no. 018834) and MALSIG (grant no. 223044), and by the Open Laboratory of Chemical Biology of the Institute of Molecular Technology for Drug Discovery and Synthesis with financial support from the Government of the HKSAR University Grants Committee Areas of Excellence Fund, project no. AoE P/10-01/01-02-I and AOE/P-10/01-2-II, and University Grants Council grant no. HKUST 6493/06 M and 600507. H.M.S. is a Wellcome Trust career development fellow.
Footnotes
Published ahead of print on 6 December 2010.
REFERENCES
- 1.Akompong, T., J. VanWye, N. Ghori, and K. Haldar. 1999. Artemisinin and its derivatives are transported by a vacuolar-network of Plasmodium falciparum and their anti-malarial activities are additive with toxic sphingolipid analogues that block the network. Mol. Biochem. Parasitol. 101:71-79. [DOI] [PubMed] [Google Scholar]
- 2.Angus, B. J., K. Chotivanich, R. Udomsangpetch, and N. J. White. 1997. In vivo removal of malaria parasites from red blood cells without their destruction in acute falciparum malaria. Blood 90:2037-2040. [PubMed] [Google Scholar]
- 3.Chavchich, M., et al. 2010. Role of pfmdr1 amplification and expression in induction of resistance to artemisinin derivatives in Plasmodium falciparum. Antimicrob. Agents Chemother. 54:2455-2464. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Coghi, P., et al. 2009. Interaction of artemisinins with oxyhemoglobin Hb-FeII, Hb-FeII, carboxyHb-FeII, heme-FeII, and carboxyheme FeII: significance for mode of action and implications for therapy of cerebral malaria. ChemMedChem 4:2045-2053. [DOI] [PubMed] [Google Scholar]
- 5.Desjardins, R. E., C. J. Canfield, J. D. Haynes, and J. D. Chulay. 1979. Quantitative assessment of antimalarial activity in vitro by a semiautomated microdilution technique. Antimicrob. Agents Chemother. 16:710-718. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Dondorp, A. M., et al. 2009. Artemisinin resistance in Plasmodium falciparum malaria. N. Engl. J. Med. 361:455-467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Dunay, I. R., W. C. Chan, R. K. Haynes, and L. D. Sibley. 2009. Artemisone and artemiside control acute and reactivated toxoplasmosis in a murine model. Antimicrob. Agents Chemother. 53:4450-4456. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Goo, Y. K., et al. 2010. Artesunate, a potential drug for treatment of Babesia infection. Parasitol. Int. 59:481-486. [DOI] [PubMed] [Google Scholar]
- 9.Gu, H. M., D. C. Warhurst, and W. Peters. 1984. Uptake of [3H]dihydroartemisinine by erythrocytes infected with Plasmodium falciparum in vitro. Trans. R. Soc. Trop. Med. Hyg. 78:265-270. [DOI] [PubMed] [Google Scholar]
- 10.Haynes, R. K., et al. 2007. The Fe2+-mediated decomposition, PfATP6 binding, and antimalarial activities of artemisone and other artemisinins: the unlikelihood of C-centered radicals as bioactive intermediates. ChemMedChem 2:1480-1497. [DOI] [PubMed] [Google Scholar]
- 11.Haynes, R. K., et al. 2010. Facile oxidation of leucomethylene blue and dihydroflavins by artemisinins: relationship with flavoenzyme function and antimalarial mechanism of action. ChemMedChem 5:1282-1299. [DOI] [PubMed] [Google Scholar]
- 12.Haynes, R. K., et al. 2006. Artemisone—a highly active antimalarial drug of the artemisinin class. Angew. Chem. Int. Ed. Engl. 45:2082-2088. [DOI] [PubMed] [Google Scholar]
- 13.Kamchonwongpaisan, S., G. Chandra-ngam, M. A. Avery, and Y. Yuthavong. 1994. Resistance to artemisinin of malaria parasites (Plasmodium falciparum) infecting α-thalassemic erythrocytes in vitro. Competition in drug accumulation with uninfected erythrocytes. J. Clin. Invest. 93:467-473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Krishna, S., L. Bustamante, R. K. Haynes, and H. M. Staines. 2008. Artemisinins: their growing importance in medicine. Trends Pharmacol. Sci. 29:520-527. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Krishna, S., S. Pulcini, F. A. Fatih, and H. M. Staines. 2010. Artemisinins and the biological basis for the PfATP6/SERCA hypothesis. Trends Parasitol. 26:517-523. [DOI] [PubMed] [Google Scholar]
- 16.Matsuu, A., M. Yamasaki, X. Xuan, H. Ikadai, and Y. Hikasa. 2008. In vitro evaluation of the growth inhibitory activities of 15 drugs against Babesia gibsoni (Aomori strain). Vet. Parasitol. 157:1-8. [DOI] [PubMed] [Google Scholar]
- 17.Meshnick, S. R., A. Thomas, A. Ranz, C. M. Xu, and H. Z. Pan. 1991. Artemisinin (qinghaosu): the role of intracellular hemin in its mechanism of antimalarial action. Mol. Biochem. Parasitol. 49:181-189. [DOI] [PubMed] [Google Scholar]
- 18.Nagelschmitz, J., et al. 2008. First assessment in humans of the safety, tolerability, pharmacokinetics, and ex vivo pharmacodynamic antimalarial activity of the new artemisinin derivative artemisone. Antimicrob. Agents Chemother. 52:3085-3091. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Noedl, H., et al. 2008. Evidence of artemisinin-resistant malaria in western Cambodia. N. Engl. J. Med. 359:2619-2620. [DOI] [PubMed] [Google Scholar]
- 20.Roberts, D. J., et al. 1992. Rapid switching to multiple antigenic and adhesive phenotypes in malaria. Nature 357:689-692. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Schmuck, G., M. Temerowski, R. K. Haynes, and B. Fugmann. 2003. Identification of non-neurotoxic artemisinin derivatives in vivo and in vitro. Recent Res. Dev. Antimicrob. Agents Chemother. 3:35-47. [Google Scholar]
- 22.Staines, H. M., et al. 1999. Passive Ca2+ transport and Ca2+-dependent K+ transport in Plasmodium falciparum-infected red cells. J. Membr. Biol. 172:13-24. [DOI] [PubMed] [Google Scholar]
- 23.ter Kuile, F., N. J. White, P. Holloway, G. Pasvol, and S. Krishna. 1993. Plasmodium falciparum: in vitro studies of the pharmacodynamic properties of drugs used for the treatment of severe malaria. Exp. Parasitol. 76:85-95. [DOI] [PubMed] [Google Scholar]
- 24.Vattanaviboon, P., N. Siritanaratkul, J. Ketpirune, P. Wilairat, and Y. Yuthavong. 2002. Membrane heme as a host factor in reducing effectiveness of dihydroartemisinin. Biochem. Pharmacol. 64:91-98. [DOI] [PubMed] [Google Scholar]
- 25.Vyas, N., B. A. Avery, M. A. Avery, and C. M. Wyandt. 2002. Carrier-mediated partitioning of artemisinin into Plasmodium falciparum-infected erythrocytes. Antimicrob. Agents Chemother. 46:105-109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Walker, D. J., et al. 2000. Mechanisms of artemisinin resistance in the rodent malaria pathogen Plasmodium yoelii. Antimicrob. Agents Chemother. 44:344-347. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Wang, J., et al. 2010. Artemisinin directly targets malarial mitochondria through its specific mitochondrial activation. PLoS One 5:e9582. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Witkowski, B., et al. 2010. Increased tolerance to artemisinin in Plasmodium falciparum is mediated by a quiescence mechanism. Antimicrob. Agents Chemother. 54:1872-1877. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Yang, Y. Z., B. Little, and S. R. Meshnick. 1994. Alkylation of proteins by artemisinin. Effects of heme, pH, and drug structure. Biochem. Pharmacol. 48:569-573. [DOI] [PubMed] [Google Scholar]