The Presence of Leukocytes in Ex Vivo Assays Significantly Increases the 50-Percent Inhibitory Concentrations of Artesunate and Chloroquine against Plasmodium vivax and Plasmodium falciparum

Abstract

Plasmodium species ex vivo sensitivity assay protocols differ in the requirement for leukocyte removal before culturing. This study shows that the presence of leukocytes significantly increases the 50% inhibitory concentration (IC₅₀) of P. vivax and P. falciparum to artesunate and chloroquine relative to results with the paired leukocyte-free treatment. Although leukocyte removal is not an essential requirement for the conduct of ex vivo assays, its use has important implications for the interpretation of temporal and spatial antimalarial sensitivity data.

Plasmodium falciparum and Plasmodium vivax ex vivo sensitivity assays are an important adjunct to in vivo antimalarial resistance detection. Ex vivo antimalarial sensitivity testing (a specialized form of in vitro sensitivity testing) involves the phenotypic comparison of freshly isolated Plasmodium spp. cultured to a specified endpoint in the presence and absence of antimalarials. Ex vivo assays provide information on the intrinsic sensitivity of malaria parasites, free from confounding variations in patient immune status. Furthermore, in the case of P. vivax, ex vivo assays are also free from the problematic differentiation between relapse, reinfection, or recrudescence which is needed for in vivo testing.

Since the development of ex vivo sensitivity antimalarial assays in the 1960s (20), there have been various modifications to the basic method, including changes in the blood medium mixture (BMM) composition (variations in hematocrit, culture medium used, and percent serum added), BMM volume (i.e., 50 μl versus 200 μl), and even incubation gas environment (candle jar versus commercial gas). Another important modification to ex vivo sensitivity antimalarial assays is the depletion of leukocytes from the patient blood sample prior to addition to the culture medium (26). It was supposed that the depletion of patient leukocytes and platelets removes the possibility that variations in host immune status (i.e., variations in leukocyte count) will confound the result of the sensitivity assay. Additionally, the removal of leukocytes aids in the microscopic examination of thick films which are used to assess the antimalarial effect on parasite development (26). Furthermore, thick and thin films made from filtered isolates produce “noise-free” images better suited for digital analysis. Despite the growing trend toward complete leukocyte depletion prior to culture (1, 8, 11, 13, 14, 17, 21-23, 25, 27), some groups still utilize patient samples with no or only partial depletion of leukocytes (removal of the buffy coat only) (5, 6, 15, 16, 28). Despite this major dichotomy in ex vivo assay methodology, it is not known if the removal of leukocytes influences the antimalarial sensitivity of the parasite or the ex vivo growth of P. vivax or P. falciparum. Consequently, the main objective of this study was to investigate whether the presence of leukocytes affects the IC₅₀ (50% inhibitory concentration) of chloroquine and artesunate against P. vivax and P. falciparum. It was also important to determine if chloroquine and artesunate significantly affected leukocyte phagocytosis (parasite removal), which, if it occurred in a dose-dependent manner, would confound the assay test results. For example, if higher concentrations of antimalarial inhibited phagocytosis, the parasitemia at lower concentrations would be reduced relative to that with the higher-concentration treatments, counter to the presumed effect of the antimalarial.

This study focuses on a modified version of the original WHO schizont maturation microtest (19), as it remains the most reliable and commonly used methodology for the side-by-side antimalarial sensitivity testing of P. vivax and P. falciparum in field locations (11, 13, 14, 17, 21, 24). Consequently, the methodology used for this study is based on that of Russell et al. (22), the only major change being the use of a gas mix (N₂, 90%; O₂, 5%; and CO₂, 5%) rather than a candle jar. Fifteen P. vivax and 10 P. falciparum isolates were collected from malaria patients attending the clinics of the Shoklo Malaria Research Unit (SMRU) Mae Sod region of Tak Province in the northwest of Thailand from October 2008 to January 2010. Isolates were collected only from patients with no prior antimalarial therapy and with a microscopically determined parasitemia of <10,000 parasites/μl. After written consent was obtained, blood samples were collected by venipuncture in 5-ml-volume lithium-heparin tubes and arrived at the culture lab at SMRU within 5 h of collection. An additional 1 ml of blood collected on EDTA anticoagulant was taken from each patient for automated hematology analysis (model pocH-100i; Sysmex Company). Only samples with >90% of the parasites at the early ring stage (approximately 6 to 8 h postinvasion) were chosen for drug sensitivity testing. Due to this strict criterion, only 8/15 and 8/10 P. vivax and P. falciparum isolates, respectively, were included in the study. All of the patient samples used in this study had a normal white blood cell count with minimal variation around the median of 5.6 × 10³ leukocytes/μl (interquartile range [IQR], 4.5 to 6.8 leukocytes/μl) observed. Each of the chosen samples was divided into two, and half of the sample, referred to as the “leukocyte removed” treatment, was depleted of platelets and leukocytes by CF11 filtration (26). It should be noted that the CF11 filtration methodology does not alter the stage composition or viability of the isolates (24). The remaining half of each sample was not filtered, and this was referred to as the “leukocytes present” treatment. Parasites in these two treatments were tested in parallel against chloroquine diphosphate (CQ) (molecular weight [MW], 515.9; Sigma-Aldrich) and artesunate (AS) (base MW, 282.3; Holly Pharmaceuticals Co Ltd.), as previously described (22). Artesunate was used in preference to its metabolite, dihydroartemisin (DHA), as AS is more commonly used in ex vivo studies and is significantly more stable on predosed dried drug plates (9). All of the microscopy work was carried out by a skilled microscopist, and quality assurance was done by an expert microscopist (with 15 years of experience in field microscopy). It was not possible to blind the microscopic reads, since the thick films of the “leukocytes present” treatment are very distinct (26).

The clinical samples examined in this study were collected under the ethical guidelines in the approved protocol OXTREC 027-025 (University of Oxford, Centre for Clinical Vaccinology and Tropical Medicine, United Kingdom).

Dose-response curves and IC₅₀s were calculated by fitting the data to a sigmoidal inhibitory E_max (maximum effect) pharmacodynamic model using the WinNonlin v. 4.1 (Pharsight Corporation, CA) software program using duplicate well data for each drug concentration. The median IC₅₀s presented in Fig. 1 A and 1B and 2 B were compared using a Wilcoxon paired-rank test. The median IC₅₀s presented in Fig. 2C and D were compared using a Kruskal-Wallis test and Dunn's post hoc analysis. Statistical analysis was carried out and graphics produced using the GraphPad Prism v. 5 software program.

FIG. 1.

The effect of leukocyte depletion on the ex vivo sensitivity (solid horizontal lines show median IC₅₀ [nM]) of paired isolates of Plasmodium vivax (n = 8) and Plasmodium falciparum (n = 8) to artesunate (A) or chloroquine (B). Solid symbols indicate the absence of leukocytes.

FIG. 2.

Neutrophil and monocyte phagocytosis of Plasmodium vivax (A) and Plasmodium falciparum significantly reduce the parasitemia in ex vivo cultures over a period of 42 h (B). The presence of artesunate (C) or chloroquine (D) does not significantly affect the ability of leukocytes to phagocytose infected red blood cells or hemozoin relative to results with the drug-free control (phagocytic index). In the image of a Giemsa-stained thin film (A), examples of a neutrophil (N), monocyte (MΦ), phagocytosed infected red blood cell (i), and hemozoin clump (p) are marked on the photomicrograph.

In all cases tested, the presence of leukocytes ex vivo increased the antimalarial IC₅₀ relative to that for the paired leukocyte-free treatment (Fig. 1). We suggest that the increase in IC₅₀s is simply due to leukocytes reducing the total concentration of antimalarial agent free to act on the parasitized red blood cells (i.e., the leukocytes act as a “biological sink” or “bio-sponge” for the antimalarials). Chloroquine is preferentially accumulated into leukocytes, such as neutrophils (which make up the majority of leukocytes present in our isolates) and monocytes (7, 18). However, it is not known if artesunate is preferentially accumulated into or neutralized by leukocytes. Interestingly, an earlier in vitro study has also shown how the presence of certain cell types with a preferential uptake of artemisinin competes with parasitized erythrocytes for the accumulation of antimalarials (10). In this case study, the presence of α-thalassemic erythrocytes, which have a higher accumulation capacity than infected wild-type red cells, resulted in lower artemisinin concentrations available to kill the parasites and a subsequent increase in artemisinin IC₅₀s. Although we suspect that leukocytes are also reducing the availability of artesunate in our study, we cannot discount the alternative possibility, that leukocytes directly increase the actual susceptibility of the parasite to the antimalarials added.

It is notable that although the chloroquine IC₅₀ of P. falciparum was increased in the “leukocyte present” treatment, this change did not reach statistical significance. This is probably due to the fact that chloroquine has little effect on these highly resistant P. falciparum strains. Plasmodium falciparum in this region of Thailand is well known to be highly resistant to chloroquine, where a 300 nM concentration of this antimalarial is not sufficient to inhibit schizont maturation. Therefore, it is less likely that we would be able to detect much of an increase in IC₅₀s over those observed in this study (Fig. 1). In future studies, it would be useful to repeat this experiment with chloroquine-sensitive isolates of P. falciparum, which are extremely rare in the study area.

To examine the effect of antimalarials on leukocyte phagocytosis over 42 h of culture (“leukocytes present” treatments only), a randomly selected sample of 100 leukocytes (consisting of neutrophils, lymphocytes, monocytes, eosinophils, and basophils) in the thin films of each species and drug concentration was examined by light microscopy under 100× oil emersion. The phagocytic index was defined as the percentage of leukocytes containing at least one parasite or obvious granule of hemozoin normalized to that of the drug-free control (Fig. 2). The parasitemia of each drug-free control was determined before and after culture for 42 h.

Prior to the start of our study, only 0 to 1.3% (range) of the leukocytes contained any detectable hemozoin or parasites. However, after 42 h of drug-free culture, a median (range) of 38.2% (13.9 to 71.0%) or 38.9% (3.8 to 93.1%) of leukocytes contained P. vivax or P. falciparum, respectively. This resulted in a significant drop in parasitemias of both species over the 42-h incubation (Fig. 2). It is notable that in addition to free merozoites, entire infected red blood cells (IRBCs) were found in the neutrophils and monocytes postculture (Fig. 2). These findings contradict early reports suggesting that only merozoites are readily phagocytosed by granulocytes and that infected red blood cells are relatively protected from phagocytosis (except if damaged) (2). Although we did not observe IRBCs or merozoites in lymphocytes, it was relatively common to find significant quantities of hemozoin phagocytosed by this cell type. There was no indication that the development of ring stages to healthy schizonts was impeded by the presence of leukocytes; there was only a reduction in overall parasitemia. These results support findings of earlier studies showing no marked inhibition of schizont maturation with peripheral blood mononuclear cells or hyperimmune serum (3, 4). Importantly, the concentrations of chloroquine and artesunate used did not significantly affect the phagocytic index (Fig. 2). Although some studies have found that chloroquine and artesunate significantly reduced the phagocytic function of granulocytes, it should be noted that these effects were noted only at concentrations of drug above 100,000 nM chloroquine (12) and 5,000 nM artesunate (29), both of which are well in excess of concentrations used in our study and in vivo therapeutic ranges for the treatment of malaria.

The results of our study have important implications for ex vivo sensitivity assays. First, the presence of leukocytes can, at least for chloroquine, significantly confound the sensitivity profile of P. vivax and P. falciparum in ex vivo antimalarial assays by reducing the amount of antimalarial free to act on the parasites. If leukocytes are not removed, sample IC₅₀s may be higher than normally expected, especially in patient samples with particularly high leukocyte counts. Furthermore, the presence of leukocytes prior to sensitivity testing will result in significantly lower parasitemias postculture, making it more difficult to observe the antimalarial effect; this is of particular concern with P. vivax isolates, where parasitemias may be close to the microscopic threshold of detection (10 to 50 parasites/μl).

The intention of this article is not to advocate the sole use of protocols using leukocyte depletion for ex vivo studies on P. vivax and P. falciparum. In fact, it would be advisable to retain host leukocytes for studies trying to correlate ex vivo sensitivity with clinical therapeutic response. However, we do wish to draw attention to a major dichotomy in ex vivo sensitivity protocols that if not taken into account may cause confusion when trying to compare spatial and temporal trends in Plasmodium sp. sensitivity profiles. The results of this study are particularly relevant, since there is a trend to move away from the microtest to more automated methods which utilize a differential increase in DNA (stained using dyes such as SYBR green or ethidium bromide) as a measure for parasite growth or inhibition; we predict that the use of leukocyte-depleted isolates (in an effort to remove background signal) will become more frequent.