Cell permeability of VSG dye
To ensure that VSG is cell-permeable and that it binds to nucleic acid, a non-synchronized culture of P. falciparum (Fig. 1A) was incubated with VSG dye without fixation and subjected to laser-scanning confocal microscope imaging in which the emitted fluorescent signal of VSG was displayed as green. To deny the possibility of autofluorescence, a sample of unstained P. falciparum-infected erythrocytes was used as a control. There was no green color observed in the control (Fig. 1B, lower panels). At lower magnification, cells having green color were observed, and they accounted for 1.9% of total observed cells (Fig. 1B, upper panels). Higher magnification images revealed green color inside the erythrocytes (Fig. 1B, yellow and blue arrows in middle panels), which suggests cell membrane permeability of VSG. Moreover, the intensity of green color was shown to vary, with intensity roughly grouped into low or high intensity (Fig. 1B, yellow and blue arrows, respectively). Two green dots were also observed in a single erythrocyte similar to those found in the Geimsa-stained thin blood smear, which suggests multiple infection of P. falciparum. These findings indicate that VSG was able to permeate the P. falciparum-infected erythrocytes.
Optimization of VSG stain for flow cytometry
Given that VSG has never been used for flow cytometry, we first had to identify a type of VSG-activating laser and a suitable fluorescence detector. The concentration of VSG was then optimized. In flow cytometry analysis, non-single cells were excluded by gating according to forward scatter (FSC) and side scatter (SSC) characteristics of cells. Briefly, cells were first gated using FSC-A parameter at the X-axis, and using FSC-H parameter at the Y-axis (Fig. 2A, upper panel). Cells having the characteristic of FSC-A equal to FSC-H were gated. Then, SSC-W and SSC-H were set at the X-axis and Y-axis, respectively (Fig. 2A, middle panel), in order to exclude cells having SSC-Whigh, which are not single cells. Cells were then further gated according to FSC-W and FSC-H (Fig. 2A, lower panel). Based on FSC-A and SSC-A, there were two populations: cells having FSC-A lower or higher than 50K (Fig. S1A). Both contained P. falciparum-infected and non-infected erythrocytes (Fig. S1B). Thus, we included both populations for analysis. These initial gating steps aimed to obtain single cells, which increases the accuracy of flow cytometric analysis. Using a FACS Aria II, the 488-nm laser could activate VSG and effectuate emission of a fluorescent signal, whereas the 633-nm and 375-nm laser could not (Fig. 2B). When we used a detector of FITC fluorochrome (500-560 nm), VSG+ cells (green-colored lines) could be separated from the unstained cells (magenta-colored lines). In contrast, when we used detector of PE (543-627 nm) and PE-Texas Red (593-639 nm), VSG+ cells (green-colored lines) overlapped with the unstained cells (magenta-colored lines), which limited our ability to analyze the parasitized cells. Therefore, we decided to use the 488-nm laser for VSG activation, and the FITC detector to read the emitted fluorescent signal.
To determine the optimal concentration of VSG, P. falciparum-infected erythrocytes were incubated with 0.5, 1, 2, 5, 10, and 20 mg/mL of VSG. The optimal VSG concentration was determined based on its ability to fractionate P. falciparum-infected erythrocytes from non-infected cells. As shown in Fig. 2C, 20 and 10 mg/mL of VSG were the concentrations that yielded the highest fluorescence intensity in VSG+ cells. Moreover, different intensity of fluorescence was observed in the 20 and 10 mg/mL VSG-stained samples (Fig. 2C, histogram), which is a finding that is consistent with confocal microscopic data. We excluded the 200, 100, and 50 mg/mL VSG concentrations due to an upward shift in the dots on the flow cytometric profile (data not shown), which suggested an increase in non-specific staining (high background). Microscopic observation of sorted VSG+ cells showed that 10 mg/mL of VSG yielded all stages of intraerythrocytic development of P. falciparum (Fig. 2D). In agreement with Fig. 2D, Giemsa staining of presorted sample showed 10.4% parasitemia that consisted of 9.8% ring form, 0.1% trophozoites, and 0.5% schizonts, which strongly suggests the accuracy of VSG at a concentration of 10 mg/mL. Therefore, 10 mg/mL of VSG was used for other experiments in this study.
Validation of the VSG staining method
To test that each stage of intraerythrocytic development of P. falciparum could be fractionated based on the intensity of VSG, a non-synchronized culture of malaria parasites was prepared. As a standard method, Giemsa staining of thin blood film showed 14% parasitemia that consisted of 13% ring form, 0% trophozoites, and 1.1% schizonts (Fig. 3A). The VSG+ cells were separated according to intensity into low, intermediate, or high (hereafter referred to as VSGlow, VSGintermediate, and VSGhigh, respectively) (Fig. 3B), and their morphologies were examined. Schizonts were observed only in VSGhigh fraction, and ring forms and growing trophozoites were observed only in VSGintermediate and VSGlow fraction (Fig. 3C). Moreover, we could observe different morphology of the P. falciparum parasites in VSGintermediate and VSGlow fraction. The cytoplasm of P. falciparum in the VSGintermediate fraction was thicker than that in the VSGlow fraction, and it contained malarial pigment (Fig. 3D). These findings were in agreement with microscopically examined Giemsa-stained thin blood film that revealed ring form, trophozoites, and schizonts in the culture, which suggested that our protocol was optimal. Thus, fluorescence intensity of VSG depends on the stage of in vitro malaria development.
To test whether VSG-based flow cytometric analysis could distinguish gametocytes from schizonts, we cultured P. falciparum strain K1 in gametocyte-inducing culture medium and performed VSG-based flow cytometric analysis. Cells in VSGlow, VSGintermediate, and VSGhigh fraction were sorted and stained with Giemsa dye. In the VSGhigh fraction, we could observe parasitized erythrocytes having granular distribution of hemozoin that resembled stage-IB gametocytes. Moreover, some were elongated and D-shaped within erythrocytes, which are key characteristics of stage-II gametocytes. Early schizonts having 2 and 6 divided nucleus, and mature schizonts consisting of 14 merozoites were also observed in the VSGhigh fraction, whereas ring forms and trophozoites were observed in the VSGlow and VSGintermediate fractions, respectively (Fig. S2). Thus, VSG-based flow cytometric assay is not able to distinguish gametocytes from schizonts.
Given the ability of VSG to differentiate intraerythrocytic stages, we explored whether change in cell granularity is related to the developmental stages of P. falciparum. VSGlow, VSGintermediate, and VSGhigh cells were gated and analyzed for SSC-A, which is an indicator of cell granularity. As shown in Fig. 3E, the median of SSC-A increased about 2 times when VSGlow and VSGintermediate cells developed into VSGhigh cells. These results suggest that change in cell granularity is related to intraerythrocytic development of P. falciparum, and that this change can be assessed using VSG-based flow cytometry.
Linearity and sensitivity of the VSG-based flow cytometric assay
To evaluate the optimized protocol relative to its ability to enumerate parasitized erythrocytes, we tested whether it could detect malaria-infected erythrocytes in a dose-dependent manner. Various concentrations of malaria-infected erythrocytes were prepared. Two-fold dilutions of infected cells were prepared using non-infected erythrocytes as diluent. That analysis revealed that VSG-based flow cytometry could detect malaria-infected erythrocytes in a dose-dependent manner (Fig. 4A). The relative values correlated well between the two assays (r2=0.75-0.97; p<0.05). The same results were observed from three independent experiments (CV = 11.2%), indicating the reproducibility of linearity measurement.
To assess the sensitivity of VSG-based flow cytometry, parasitized erythrocytes were diluted to 0.001%, which is the limit of detection in routine microscopic diagnosis [20]. As shown in Fig. 4B, two independent cultures were analyzed for each cytometry run. There were 11% and 9% parasitemia enumerated using Giemsa-based microscopy. The parasites were diluted to 0.001% using non-infected erythrocytes as diluent. The diluted samples having 0.001% parasitemia were then subjected to flow cytometry analysis. VSG-based flow cytometry was capable of detecting 0.3% and 1.1% of VSG+ cells, which is 300-1,000 times higher than the detection rate (0.001% parasitemia) by Giemsa-based microscopy. Next, we examined the reproducibility of the developed assay for enumeration of low parasitemia. We prepared three independent settings of malaria culture and diluted them to 0.01% parasitemia, which is a minimum value that correlated well with standard microscopic examination (Fig. 4A). All three independent runs of VSG-based flow cytometry were able to detect 0.9±0.2% of VSG+ cells (CV = 22%, Fig. 4C), implying reproducibility comparable to Giemsa-based microscopy (CV = 21.8%) for detection of low parasitemia.
To examine the variability of VSG-based flow cytometric assay for enumeration and identification of P. falciparum-infected erythrocytes among different sets of parasite culture, we prepared parasite culture on different dates and compared the enumerated values of parasitized cells (mean±SD) obtained from Giemsa-based microscopic analysis with those obtained from VSG-based flow cytometric analysis (Table 1). There were two types of culture: ring-form and trophozoite predominant. In both types of culture, the CV of the VSG-based flow cytometric assay for enumeration of parasitemia was relatively lower than that of the microscopic method, which implies lower variability of the VSG-based flow cytometric assay. When analyzing the variability of assays according to developmental stage, high CV values were obtained from both Giemsa-based microscopy and VSG-based flow cytometry, which is likely due to low parasitemia in each developmental stage. Collectively, VSG-based flow cytometry is a reliable, sensitive, and reproducible assay for enumeration of parasitemia.
Application of VSG-based flow cytometry for synchronicity assessment and drug sensitivity testing
Synchronization of P. falciparum development is a common method used in routine culture, and its aim is to obtain a predominant intraerythrocytic stage of parasites. To explore whether VSG-based flow cytometry is capable of assessing synchronicity of P. falciparum development in a routine culture, synchronized and non-synchronized cultures of P. falciparum were prepared (Fig. 5A), stained with VSG, and subjected to flow cytometry analysis. Given the ability of flow cytometry to detect cell size and granularity using respective FSC and SSC, we hypothesized that synchronized parasites have the same size and granularity, which suggests homogeneity. Thus, we selected a quantile contour plot, which is an effective way to visualize distinct populations regardless of the numbers of cells displayed [22], to assess cell homogeneity. In Fig. 5B, only VSG+ cells were displayed based on their size (as indicated by FSC-A on the X-axis) and granularity (as indicated by SSC-A on the Y-axis). To enhance the visualization of a distinct cell population having various cell size and granularity, histograms of FSC-A and SSC-A are also shown at the top and left side of the contour plots, respectively. Given the ability of contour plot to visualize cells based on the relative frequencies of sub-populations, we were able to locate distinct populations of VSG+ cells using vertical and horizontal lines drawn on the contour plots. There were at least three distinct populations observed in the non-synchronized culture (Fig. 5B, left panel), as follows: (1) cells having small size with various granularity (approximately 0-45K of FSC-A, and 30-170K of SSC-A); (2) cells having a relatively large size with high granularity (approximately 45-185K of FSC-A, and 75-170K of SSC-A); and, (3) cells having a relatively larger size with low granularity (approximately 45-185K of FSC-A, and 20-75K of SSC-A). In contrast, only one minor (indicated as 1) and one major (indicated as 2) population were observed in the synchronized culture. They had a similar size (50-150K of FSC-A), but different levels of granularity (35-240K of SSC-A) (Fig. 5B, right panel). In the left panel of Fig. 5B, the population of VSG+ cells having less than 45K of FSC-A was observed only in the non-synchronized culture (indicated as 1), but not in the synchronized culture (Fig. 5B, lower panel). Based on the intensity of VSG and microscopic images (Fig. S3A), the population number 1, 2, and 3 in the left panel of Fig. 5B are schizonts, trophozoites, and ring forms, respectively. In contrast to observation in the non-synchronized culture, the minor and major populations of the synchronized culture could be separated based on SSC-A, as follows: (1) minor population with SSC-A higher than 160K, and (2) major population with SSC-A lower than 160K (Fig. 5B, right panel). Compared to the non-synchronized culture, we observed a population of VSG+ cells having SSC-A higher than 160K only in the synchronized culture (indicated as 1 in Fig. 5B, lower panel). Based on microscopic images (Fig. S3B), the VSG+ cells with more than 160K SSC-A are infected erythrocytes containing multiple (60%) and single (40%) ring forms, and they had VSG intensity of 11,578; whereas, the VSG+ cells with lower than 160K SSC-A are infected erythrocytes with multiple (35%) and single (65%) ring forms. Disappearance of the population having more than 160K of SSC-A on the contour plot of the non-synchronized culture (Fig. 5B, left panel) and less than 45K of FSC-A on the contour plot of the synchronized culture (Fig. 5B, right panel) resulted from different developmental stage of Plasmodium between the two separate cultures. To confirm heterogeneity in the non-synchronized culture, we statistically analyzed the CV, which is a measure of relative variability, of FSC-A and SSC-A. Despite statistical non-significance (p>0.05), the non-synchronized culture tended to have a higher CV for both FSC-A and SSC-A (Fig. 5C), which confirms the relatively high heterogeneity of VSG+ cells. Thus, VSG-based flow cytometry is an effective alternative method for assessing synchronicity of P. falciparum development in erythrocytes.
To demonstrate the use of VSG for assessment of growth inhibitory effect of antimalarial drug, malaria-infected erythrocytes were incubated with 700 nM DHA for 24 hours following a standard assay [21]. The DHA- and DMSO-treated cells were stained with VSG and analyzed by flow cytometry. In both the presence and absence of the drug, there were VSG+ cells exhibiting VSGintermediate and VSGlow (Fig. 5D), which were likely to resemble trophozoite and ring form, respectively. Our results showed that the number of VSG+ cells decreased following DHA treatment (Fig. 5D, right panel) compared to that of the DMSO-treated control cells (Fig. 5D, left panel and p = 0.02). The majority of DHA-treated VSG+ cells appear as VSGlow, implying that ring form was predominant. In contrast, both VSGintermediate and VSGlow cells were observed in the DMSO-treated control (Fig. 5D, left panel), implying that both ring from and trophozoite were present in the culture. Similar to 700-nM DHA treatment, number of VSG+ cells also decreased after treatment with 350 and 1,400 nM DHA for 24 hours (p = 0.007 and 0.016, respectively); however, there was no difference in number of VSG+ cells among doses (Fig. 5E). Moreover, VSG-based flow cytometry was able to access effect of 700-nM DHA in time-dependent manner (Fig. 5F). According to the VSG-based flow cytometric data, DHA likely inhibited parasite growth. Therefore, the VSG-based flow cytometric assay can be used as an alternative assay for assessment of P. falciparum growth in the presence of antimalarial drug in vitro.