Susceptibility of human Plasmodium knowlesi infections to anti-malarials

Background Evidence suggests that Plasmodium knowlesi malaria in Sarawak, Malaysian Borneo remains zoonotic, meaning anti-malarial drug resistance is unlikely to have developed in the absence of drug selection pressure. Therefore, adequate response to available anti-malarial treatments is assumed. Methods Here the ex vivo sensitivity of human P. knowlesi isolates in Malaysian Borneo were studied, using a WHO schizont maturation assay modified to accommodate the quotidian life cycle of this parasite. The in vitro sensitivities of P. knowlesi H strain adapted from a primate infection to in vitro culture (by measuring the production of Plasmodium lactate dehydrogenase) were also examined together with some assays using Plasmodium falciparum and Plasmodium vivax. Results Plasmodium knowlesi is uniformly highly sensitive to artemisinins, variably and moderately sensitive to chloroquine, and less sensitive to mefloquine. Conclusions Taken together with reports of clinical failures when P. knowlesi is treated with mefloquine, the data suggest that caution is required if using mefloquine in prevention or treatment of P. knowlesi infections, until further studies are undertaken.


Background
From its natural simian hosts in Southeast Asia, Plasmodium knowlesi has emerged as a significant human pathogen, particularly in Malaysian Borneo [1][2][3]. Human P. knowlesi infections cause febrile illnesses that can rapidly progress to severe and sometimes fatal outcomes [4]. Ominously, the incidence of P. knowlesi malaria is increasing in geographic areas where Plasmodium falciparum and Plasmodium vivax are coming under control, thereby threatening the aim of eliminating malaria [5]. Determining the efficacy of conventional anti-malarials against P. knowlesi is a priority, particularly as there are no reports of anti-malarials assessed against human isolates of P. knowlesi ex vivo.
Here, the drug sensitivity profiles of P. knowlesi isolates obtained from patients being recruited into a study of the pathophysiology of knowlesi malaria in an endemic area of Sarawak, Malaysian Borneo were investigated. Currently, the WHO recommends artemisinin-based combination therapy (ACT) as first-line treatment for malaria in most endemic areas, so artemisinin and its clinically useful derivatives artesunate, dihydroartemisinin (DHA) and artemether were tested. Mefloquine, used as a partner drug in certain artemisinin-based combinations and in prophylaxis against malaria, and chloroquine that is recommended for treatment of Plasmodium malariae (the species which P. knowlesi is often confused with when diagnosed by microscopy) were also included. In addition, results from P. falciparum and P. vivax studied contemporaneously are presented, together with those obtained with the H strain laboratory isolate of P. knowlesi (cultured in vitro in rhesus erythrocytes) to confirm the methodologies used. Insights into the drug susceptibility patterns of this important emerging parasite, may prove useful in guiding the best choice of anti-malarial treatment regimens for P. knowlesi infection.

Patient recruitment
Plasmodium isolates were obtained from patients presenting to hospitals in Sarikei and Sibu. Informed written consent was obtained from all patients entered into this study, which was approved by the Malaysian Ministry of Health's Medical Research and Ethics Committee, and the Sarawak State Planning Unit. Infecting species was confirmed by Plasmodium species-specific nested-PCR assays [6] and only patients with single species infections were retained in the study.

Blood collection and ex vivo parasite development in growth assays
Pre-treatment venous blood from each patient was collected into EDTA. Parasitaemia and the asexual stage of development were determined by Giemsa-stained thin film microscopy. Whole blood (~2.5 ml) was washed twice without centrifugation to avoid haemolysis, before resuspending in RPMI 1640 complete medium supplemented with 20 mM D-glucose, 40 mM HEPES, 25 mg/l gentamicin sulphate, and 15% v/v human AB plasma with 0.2 mM hypoxanthine.
Growth inhibition by anti-malarials was assessed by quantifying schizont maturation using an adapted WHO Mark III assay [7]. Species of Plasmodium tested and the time lag in maturation seen previously ex vivo were allowed for in these assays [8]. While more complex counting procedures have been used to study P. vivax parasites that, like P. knowlesi, have mature parasites present in circulating blood [9], only the timing of assays was altered. This is because the P. knowlesi isolates contained predominantly immature parasites (Table 1) and results from a parallel study on tightly synchronised immature laboratory P. knowlesi H strain parasites with artesunate using the pLDH (see below) and Mark III assays were comparable with each other and with the data derived from isolates (see Results).
Aliquots (100 μl) of no drug control and serial dilutions of the anti-malarial compounds in culture medium (final concentrations of 0.25 to 25 nM for artemisinin and derivatives and 1.25 nM to 1 μM for chloroquine and mefloquine) were dispensed into 96-well plates, and parasites were added (100 μl at 2% haematocrit). Incubation was in 5% O 2 , 5% CO 2 , 90% N 2 at 37°C until the majority of parasites reached the schizont stage (with at least 3 nuclei after 12-17 h for P. knowlesi). Monitoring of maturation was undertaken every 2 h by examination of fixed, Giemsa stained thin blood films, taken from a parallel culture to those of the drug exposed cultures. When at least half of the parasites in the monitoring culture had reached the schizont stage of development, the drug exposed cultures and their controls were harvested, as thick films on glass slides, and examined. The duration of in vitro development for each isolate (as determined above), which also equates to the time of drug exposure, is given in Table 1. Thick films were fixed by air drying for at least 24 h, stained with Giemsa, and mounted to protect slides during transportation.
As thick films are easily damaged, five replicates were prepared for each drug concentration and 3 replicates then counted for each experimental condition. Counting was in a blinded fashion to avoid bias. Thick films were counted according to the WHO Mark III protocol. At least 200 asexual parasites were counted. Fields of view were consecutive, starting at the left edge of each blood film and moving stepwise in a uniform direction (to ensure no overlap). Asexual parasites were grouped into either trophozoites or schizonts (defined as asexual parasites displaying 1-2 nuclei and 3 or more nuclei, respectively). Growth assay measuring Plasmodium lactate dehydrogenase (pLDH) Rhesus monkey red blood cells for in vitro P. knowlesi culture were obtained under protocols approved by the independent institutional ethical committee (DEC) according to Dutch and European laws. The efficacy of the anti-malarial compounds was also assessed in vitro against the laboratory maintained P. knowlesi H strain [10,11], by measuring the production of pLDH [12]. pLDH catalyzes the conversion of 3-acetylpyridine adenine dinucleotide (APAD) and lactate to APADH and pyruvate. The enzyme diaphorase subsequently converts nitro blue tetrazolium (NBT) to nitro blue formazan (NBF), using APADH as a reducing agent. NBF can be measured at a wavelength of 655 nm. Serial dilutions of the anti-malarial compounds (100 μl in culture medium) and including a no drug control were dispensed into 96-well plates, to which tightly synchronized (by alanine lysis) early ring-stage infected erythrocytes (100 μl at 2% haematocrit and 2% parasitaemia in culture medium) were added. The plates were then placed in 3% O 2 , 7% CO 2 , 90% N 2 at 37°C for 22 h to mature (just prior to parasite release). Growth was halted and drug removed, by washing twice in ice-cold PBS. Erythrocytes were lysed by freezing at −20°C and thawing. Aliquots of 0.5 mg/ml of NBT, 1 U/ml diaphorase and 50 μg/ml APAD in LDH buffer (100 mM Tris-HCl, pH 8.0, 50 mM Na L-lactate, 2.5% v/v Triton X-100) were added to the thawed cell pellets and incubated for 30 min in the dark at room temperature with shaking. The optical density (OD) at 655 nm of each well was measured in a BioRad 680 microplate reader. OD 655 values were used to calculate growth by comparing the OD 655 values at each drug concentration with that of the no drug control. To test the efficacy of artemisinin, artemether, artesunate and DHA, preparations of final concentrations of 0.1 nM to 25 nM were used, and for chloroquine and mefloquine, final concentrations of 1.25 nM to 1 μM were used.

IC 50 values and sequence alignments
Dose-response data were modeled using a four-parameter fit and a variable slope, using Prism (Version 4.0a). Goodness of fit was assessed by R 2 , and either noted in the text or highlighted if < 0.8 in tables and figures. For the schizont maturation assay, the top parameter was constrained to 100% for assays using artemisinins but not for mefloquine or chloroquine. For the pLDH assay, the top and bottom parameters were constrained to 100 and 0, respectively, for all drugs and, for the 3 Hhypoxanthine incorporation assay, the top parameter was constrained to 100. Data are summarized as a mean and 95% CI for replicates of single parasite assays and a mean ± SEM for results from multiple independent parasite assays.
The P. knowlesi and P. vivax orthologues of P. falciparum CRT, MDR1, and ATP6 (proteins associated with modulating sensitivities to chloroquine, mefloquine and artemisinins [13]), were aligned with P. falciparum 3D7 strain, as a drug sensitive control. H strain and Sal-1 strain sequences were used for P. knowlesi and P. vivax, respectively. Wild-type and polymorphism sequence data were taken from PlasmoDB [14]. Alignments were performed in MacVector (version 11.0.2).

Patient recruitment
Patient isolate data from those obtained in Sarikei and Sibu hospitals, between March and September 2010, are shown in Table 1. A total of nine patients were recruited into this study, six with P. knowlesi, two with P. vivax and one with P. falciparum monoinfections. Infections were initially diagnosed by microscopy and confirmed later by nested PCR [2,6].

Parasite development
In vitro development data for each isolate used during growth assays are shown in Table 1. P. knowlesi field isolates were seeded into growth assays with a starting parasitaemia ranging from 0.4 to 7%. The percentage of schizonts present at the start of the assays ranged from 6 to 33%. Good in vitro development was demonstrated for all P. knowlesi field isolates, with 53 to 94% of asexual parasites reaching the schizont stage, over a development period of 12 to 17 h. Development was similar for P. falciparum and P. vivax field isolates (Table 1).

Growth assays: schizont maturation
Dose-response curves were used to derive IC 50 values, which are presented in Figure 1, together with mean values in Table 2. All the isolates from humans were sensitive to artemisinin and its derivatives, with IC 50 values in the low nM range (from 0.38 to 6.8 nM). The IC 50 values for chloroquine on P. knowlesi from humans were relatively high, with values ranging from 11 to 38 nM but were similar to those derived for the single P. falciparum isolate (28 (20 to 38) nM; R 2 = 0.93) and the two P. vivax isolates (33 (10 to 110) and 18 (11 to 27) nM; R 2 = 0.45 and 0.95, respectively) that were studied.
Plasmodium knowlesi isolates were least sensitive to mefloquine, with a mean IC 50 value of 26 nM ( Table 2). The IC 50 value for mefloquine on the single P. falciparum field isolate (6.3 (2.5 to 16) nM; R 2 = 0.97) was four-fold less than that of the average P. knowlesi IC 50 value and 2.5-fold less than the lowest P. knowlesi IC 50 value (16 (12 to 22) nM; R 2 = 0.96).

Growth assays: pLDH
To validate the results of the drug assays on P. knowlesi from patients, anti-malarials were assessed against the well characterized P. knowlesi H strain grown in vitro in rhesus erythrocytes. To determine whether data generated by the pLDH method (as used to study the H strain) would be comparable with the adapted schizont maturation method (as used in the field study) a parallel study was performed, using artesunate. The IC 50 value for artesunate against the H strain was 2.0 (0.93 to 4.2) nM (R 2 = 0.39), using the pLDH assay (from a single experiment repeated in quintuplicate), and 1.2 (0.88 to 1.6) nM (R 2 = 0.95), using the schizont maturation assay (from a single experiment repeated in triplicate), demonstrating that the two assay methods are comparable, at least in the case of artemisinins. Further IC 50 values derived from pLDH assays are presented in Table 2. All IC 50 values for field isolates and the P. knowlesi H strain for artemisinins are highly comparable.
For mefloquine, the IC 50 value against the laboratory H strain (25 (7.4 to 81) nM; R 2 = 0.40) was nearly identical to the mean IC 50 value of human isolates (26 nM). The IC 50 values suggest that P. knowlesi may be intrinsically insensitive to mefloquine. To ensure that the mefloquine stock used in field studies had not degraded during the study, an aliquot was tested after shipping it back to St. George's, University of London after the study. Using hypoxanthine incorporation as a measure of growth [15] in a single assay (performed in quadruplicate) with P. falciparum (3D7) parasites cultured in vitro in human erythrocytes, the IC 50 value was 9.7 (5.5 to 17) nM (R 2 = 0.89), confirming that potency of mefloquine was maintained.
Interestingly, the H strain was over 7-fold more sensitive to chloroquine than the average value for field isolates, having an IC 50 value of 3.2 (2.2 to 4.7) nM (R 2 = 0.86). This value was also well below the lowest calculated IC 50 value (11 (5.3 to 24) nM; R 2 = 0.94) for a field isolate.

Sequence alignments
Point mutations associated with change in the sensitivity of P. falciparum to chloroquine, mefloquine and the artemisinins [13] were analysed, and similarities and differences in sequences encoded by P. knowlesi and P. vivax homologues are reported in Figure 2. Known loci at which mutations can reduce drug sensitivity were highly conserved between P. knowlesi H, P. vivax Sal-1 MDR1 homologues and MDR1 in drug sensitive P. falciparum 3D7. PfATP6 orthologues demonstrated polymorphism in amino acids in about two thirds of the   residues previously examined for effects on drug sensitivity, including L263 [16][17][18]. The CRT orthologues were fairly well conserved with the exceptions of S219A, A220S, and R371M in the chloroquine sensitive strain. These differences do not confer chloroquine resistance to the P. knowlesi H strain. Agreement in sequences in these polymorphic regions was higher between P. knowlesi H and P. vivax Sal-1 than between either strain and P. falciparum 3D7 consistent with a closer phylogenetic relationship between the former species [19].

Discussion
In vitro culture of P. knowlesi has only recently been achieved in human erythrocytes [20,21]. However, drug sensitivity assays of natural human infections can only be assessed in short term cultures. These allowed the successful application of drug sensitivity micro-assays such as the WHO Mark III micro-assay test after adaptation to the quotidian life cycle of the parasite and frequent (every 2 to 2.5 h) monitoring of development to allow assessment of when most parasites were mature schizonts in control samples.
The excellent in vitro efficacy of the artemisinins against both human P. knowlesi (mean IC 50 values < 2.2 nM) and the laboratory (H) strain is consistent with recent observations made on patients treated with artesunate, where no mortality was observed after treatment [22]. These results also agree with data from animal models, which demonstrate the successful clearance of P. knowlesi parasites from infected rhesus monkeys, by artemisinin in combination therapy with naphthoquine [23]. In addition, William et al. [24] in a retrospective analysis of clinical cases of knowlesi malaria in Sabah, reported the successful treatment of patients by artemether-lumefantrine combination therapy. This study also noted that where knowlesi malaria developed signs of severity, intravenous artesunate was effective [24].
Intriguingly, chloroquine IC 50 values from this field study, including those derived against P. knowlesi (with values ranging from 11 to 38 nM), were higher than that of the laboratory P. knowlesi H strain (3.2 nM), although the reason for this is unclear. In general, the values are comparable with chloroquine-sensitive laboratory strains of P. falciparum, having IC 50 values ranging from 8 to 15 nM [25][26][27]. These values all fall below the 100 nM threshold used to define chloroquine resistance [27,28] and there is no evidence for clinical chloroquine resistance reported in P. knowlesi. Chloroquine is effective both as a monotherapy and when used in combination with primaquine [1,24,29]. Consistent with these findings, Tyagi et al. [30] have recently reported that CRT (and DHFR) sequences from P. knowlesi clinical isolates collected in the Andaman and Nicobar Islands, India, were all found to be wild-type (with close homology to the CRT sequence of chloroquine sensitive P. falciparum parasitessee Figure 2 and below). Nevertheless, continued monitoring of chloroquine sensitivity in P. knowlesi might be prudent bearing in mind the history of chloroquine resistance development in other malarial species.
Evidence suggests that transmission of P. knowlesi to humans in Sarawak remains zoonotic and, thus, ostensibly free from mefloquine drug selection pressure. However, results with mefloquine consistently showed a low sensitivity in P. knowlesi field isolates, when compared with that of P. falciparum. The mean IC 50 value for mefloquine calculated for the 6 P. knowlesi isolates is 26 nM, which is just above the value used to define mefloquine resistance in P. falciparum (>24 nM) in some reports [31,32] but well below that reported by others (>119 nM) [33]. Importantly, the reduced P. knowlesi response in human isolates was also observed in the P. knowlesi H experimental line. Given the efficacy of the mefloquine used was confirmed after the end of the study, these results indicate an innate tolerance of P. knowlesi to mefloquine. These findings also suggest the strong possibility of treatment failure if mefloquine is used as mono or combination therapy for P. knowlesi and is supported by reports of mefloquine treatment failure in rhesus monkeys infected with P. knowlesi [34], as well as recent cases of mefloquine treatment failure in humans with knowlesi malaria [35]. On this basis, mefloquine should be used with caution for the treatment of knowlesi malaria, or indeed for prophylaxis against malaria in areas where acquiring knowlesi is a risk until larger studies have been undertaken.
There is reasonable conservation of MDR1 sequences between P. vivax Sal-1 and P. knowlesi H and the P. falciparum 3D7 reference strain. Previous studies on P. falciparum found that increased mdr1 copy number conferred a mefloquine resistant phenotype [36] and risk of treatment failure, although P. knowlesi mdr1 copy number in the isolates reported in the current study have not been determined. Alignments of the P. vivax Sal-1 and P. knowlesi H CRT and ATP6 orthologues with the P. falciparum 3D7 sequence revealed several polymorphic differences between the P. vivax and P. knowlesi alignments and that of the P. falciparum. These substitutions do not alter sensitivity of P. knowlesi to artemisinins and give insights into the possible contributions of these residues to artemisinin sensitivity in P. falciparum.
Here, it has been established that it is possible to culture in vitro P. knowlesi in human erythrocytes in the short term, when taken ex vivo. In addition, the successful adaption of the schizont development assay to determine anti-malarial drug sensitivities of P. knowlesi field isolates has been shown. Using this adapted method, this study has demonstrated that chloroquine and artemisinin