An automated haematology analyzer XN-30 distinguishes developmental stages of falciparum malaria parasite cultured in vitro

Plasmodium falciparum strains K1 and 3D7 were cultured in RPMI 1640 medium supplemented with 0.5 g/L l-glutamine, 5.96 g/L HEPES, 2 g/L NaHCO₃, 50 mg/L hypoxanthine, 10 mg/L gentamicin, 10% heat-inactivated human serum, and RBCs at 3% haematocrit in an atmosphere of 5% CO₂, 5% O₂, and 90% N₂ at 37 °C as previously described [10]. Ring-form iRBCs were collected using the sorbitol synchronization technique [11]. Sorbitol selectively disrupts iRBCs containing trophozoite or schizont stage and preferentially selects for RBCs with early ring stage parasites. Briefly, the culture contents collected by centrifugation at 840×g for 5 min were suspended in a fivefold volume of 5% d-sorbitol (Nacalai Tesque, Kyoto, Japan) and the cells were washed twice with RPMI 1640 medium to remove d-sorbitol.

Free merozoites were isolated as previously described [12]. Briefly, RBCs infected with late trophozoites and schizonts were enriched using 58.5% (v/v) Percoll (Sigma-Aldrich, St Louis, MO, USA) and late-stage iRBCs were collected using density centrifugation. The Percoll-purified iRBCs were further cultured by resuspension in complete medium without fresh RBCs, for 3 h. The culture was centrifuged at 840×g for 5 min. The supernatant was filtered through a 1.2 μm pore size syringe filter (Minisart; Sartorius Stedim Biotech SA, Aubagne, France) and the filtrate was collected as described [13].

Gametocytogenesis was induced as previously reported [14] with modifications [15]. Briefly, the culture suspension was subjected to stress conditions (high parasitaemia, nutritional stress, RBC lysis product), and incubated with 11.6 mg/mL N-acetyl glucosamine (Sigma-Aldrich). Gametocytaemia was assessed by microscopy of Giemsa-stained smears (Additional file 2: Table S3). The developmental stages of gametocytogenesis were classified as reported [14].

Ring-form iRBCs were collected using sorbitol treatment (see above) [11]; the treatment was repeated two times for tight synchronization. The synchronized culture was separated into three dishes and parasitaemia was measured for each dish every 4 h for 48 h using the XN-30 analyzer and by microscopy of Giemsa-stained smears as described below.

The XN-30 analyzer was equipped with a modified algorithm for cultured falciparum parasites (prototype: software version 01-03, build 16) and used dedicated reagents (CELLPACK DCL, SULFOLYSER, Lysercell M, and Fluorocell M) (Sysmex). In this algorithm, the merozoites are eliminated from scattergrams (SFL-SSC) to avoid counting merozoites as MI-RBC in in vitro culture. The classification of the growth stage has been optimized to suit in vitro culture conditions. Approximately 100 µL of culture suspension was applied to a Capiject Capillary Blood Collection Tube (Terumo, Tokyo, Japan) and loaded onto the XN-30 analyzer as described in the instrument manual. The iRBC counts (total, MI-RBC#; ring-forms, RNG-RBC#; trophozoites, TRPZ-RBC#; schizont, SCHZ-RBC#; and merozoites, MEROZ#) and parasitaemias (total, MI-RBC%; ring-forms, RNG-RBC%; trophozoites, TRPZ-RBC%; and schizont, SCHZ-RBC%) were automatically reported.

For microscopy, a standard thin blood smear was fixed with 100% methanol for 10 min and stained with 10% Giemsa stain working solution, pH 7.2 (Merck KGaA, Darmstadt, Germany), for 13 min. Slides were observed at 1000× magnification using the model BX50 light microscope (Olympus, Tokyo, Japan). The parasitaemia values were obtained by counting the iRBCs in at least 10,000 RBCs.

Twenty microlitres of parasite suspension (K1 strain; 2.5% parasitaemia, 50% haematocrit) and 20 µL Fluorocell M were mixed with 1 mL Lysercell M. As a control, another parasite suspension was incubated with 20 µL Fluorocell M for 30 min at room temperature and then diluted with phosphate buffered saline (PBS). Samples were imaged using an ImageStreamX flow cytometer according to the instrument manual (Amnis-Millipore, Seattle, WA, USA). The data was analyzed using IDEAS Application v6.0 software (Amnis-Millipore). Diameters of iRBCs were measured using Image J software, version 1.8.0 (NIH, Bethesda, MD, USA).

Statistical analyses, including coefficient of determination (R²), mean, standard deviation (SD), and coefficient of variation (CV = SD/mean × 100) were calculated using Excel software (Microsoft, Redmond, WA, USA).

The XN-30 analyzer recognized the developmental stages of the malarial parasites by measuring DNA content and iRBC size as indices. A previous study empirically assumed that permeabilization with Lysercell M shrunk iRBCs depending on the developmental stage of P. falciparum [9]. To confirm this developmental stage-dependent iRBC shrinkage, iRBCs treated with Lysercell M or PBS were imaged using the ImageStreamX flow cytometry system. Only iRBCs treated with Lysercell M shrunk in the ascending order of ring-forms, trophozoites, and schizonts (Fig. 1), indicating that the signal intensity detected by flow cytometry reflected the iRBC size depending on the developmental stage (Fig. 2a(i)).

The XN-30 analyzer recognized ring-forms, trophozoites, and schizonts on the M scattergram, as shown in Fig. 2a(i). Parasitaemias of total and each developmental stages were reported automatically (Fig. 2a(ii) and Additional file 2: Table S1). Purified merozoites were detected at the lower portion of the ring-forms on the M scattergram, and were distinguished from iRBCs on the M(SSC-FSC) scattergram (Fig. 2b and Additional file 1: Figure S2, marked by arrows; compare with Fig. 2a(i)). The XN-30 analyzer definitely detected merozoites (Fig. 2b(ii) and Additional file 2: Table S2). In case of purified gametocytes (see Additional file 2: Table S3), on the other hand, matured gametocytes (stages III/IV/V) were detected above the ring-forms and this portion was inseparable from the trophozoites on the M scattergram. This indivisibility was also found on the M(SSC-FSC) scattergram (Fig. 2c(i), marked by arrows; compare with Fig. 2a(i)). Early stage gametocytes (stages I/II) and undeveloped iRBCs were detected as ring-forms (Fig. 2c(i), marked by arrowheads). Mature and early stage gametocytes were counted as trophozoites and ring-forms, respectively (Fig. 2c(ii) and Additional file 2: Table S4; compare with Additional file 2: Table S3). The observations indicate that the XN-30 analyzer recognizes gametocytes spontaneously generated during the in vitro cultivation as ring-forms and trophozoites. Thus, although it shows inability to distinguish gametocytes from mixed cultures, the XN-30 analyzer is useful for detecting purified gametocytes.

To investigate the reliability of the XN-30 analyzer, parasitaemias of total and each developmental stages were compared between the XN-30 analysis and the gold standard microscopy examination. The correlation of total parasitaemia was high (R² = 0.982; Fig. 3a), confirming prior observations (Additional file 1: Figure S1b and [9]). The coefficients of determination of parasitaemia for ring-forms, trophozoites, and schizonts were 0.979, 0.820, and 0.710, respectively (Fig. 3b–d). These results suggested that the XN-30 analyzer classifies these three developmental stages similar to the microscopic observation, although the coefficients of determination of trophozoites and schizonts were relatively low (Fig. 3c, d). This is likely due to low parasitaemia and/or morphological confusion caused by sequential change of trophozoites to schizonts.

To demonstrate that the XN-30 analyzer reliably measures the proportion of trophozoite and schizont without morphological confusion, the developmental stages from tightly synchronized ring-forms was observed by both the XN-30 analyzer and microscopy. Synchronization of iRBCs to ring-forms revealed two populations on the M scattergram. The left ring-form represented one parasite and the other represented two parasites in a single RBC (Fig. 4a; 0 h). As mentioned, since the XN-30 analyzer can distinguish the amount of nucleic acid in an iRBC, cells are plotted in different populations on M scattergram [9]. Thus, the presence of single and multiple infected parasites is inferred as seen in Fig. 4a. Trophozoites were generated after 20 h and schizonts appeared after 28 h. Ring-forms of infected new RBCs increased after 36 h. The parasitaemia of ring-forms at 32 h was minimal but began to increase after 36 h as parasites infected new RBCs. The parasitaemias of trophozoites and schizonts were highest at 32 and 40 h, respectively (Fig. 4a, b). Comparison of the parasitaemia obtained with the XN-30 analyzer and Giemsa-stained microscopy revealed that the XN-30 analyzer showed smaller variability between the dishes, and slightly higher parasitaemia than microscopy (Fig. 4c and Table 1). These observations suggested that the XN-30 analyzer is more reproducible and sensitive than microscopy.