A detailed kinetic model of glycolysis in Plasmodium falciparum-infected red blood cells for antimalarial drug target identification

Abstract

Upon infection by the malaria parasite Plasmodium falciparum, the glycolytic rate of a red blood cell increases up to 100-fold, possibly contributing to lactic acidosis and hypoglycemia in patients with severe malaria. This dramatic increase in glucose uptake and metabolism was correctly predicted by a newly constructed detailed enzyme kinetic model of glucose metabolism in the trophozoite-infected red blood cell. Subsequently, we expanded the model to simulate an infected red blood cell culture, including the different asexual blood-stage forms of the malaria parasite. The model simulations were in good agreement with experimental data, for which the measured parasitic volume was an important parameter. Upon further analysis of the model, we identified glucose transport as a drug target that would specifically affect infected red blood cells, which was confirmed experimentally with inhibitor titrations. This model can be a first step in constructing a whole-body model for glucose metabolism in malaria patients to evaluate the contribution of the parasite's metabolism to the disease state.

Keywords: glucose transport; glycolysis; malaria; mathematical modelling; red blood cell; systems biology.

Figure 1

The combined parasite and red blood cell model (dutoit1).Orange blocks show enzyme-catalyzed reactions, blue circles are the metabolites (each has an ODE), green circles represent the fixed metabolites for the steady state version of the model. A–F, the metabolic pathways in the red blood cell ((A) glycolysis; (B) PPP; (C) 2,3-BPG shunt; (D) hemoglobin binding; (E) glutathione oxidation and non-glycolytic NADH consumption; (F) magnesium binding) and (G) the glycolytic pathway in the parasite). Note the linking of the parasite metabolism to that of the red blood cell via the glucose transporter (GLCtr - reaction 55) and lactate and pyruvate transporters (LACtr – reaction 69; PYRtr – reaction 68).

Figure 2

Infected red blood cells visualized using fluorescence microscopy. The outlines of the infected red blood cells can be distinguished by the difference in emission intensity of the cytosolic stain (A). Overlaid fluorescence emission (from Hoechst DNA stain) and transmission images of the infected red blood cells show the diameter of the intraerythrocytic parasite compartments (B–I). The pictures shown are representative of ring- (B), early trophozoite- (C), mid-late trophozoite- (D–G), and schizont-stage parasites (H–I). Red lines and accompanying values indicate diameter measurements.

Figure 3

High parasitaemias are obtained using the density gradient separation protocol. Normal culturing protocol showing 15% parasitemia (left), whereas enrichment via density gradient separation produces parasitemia >90% (right). Cells were stained with DNA-based Giemsa stain.

Figure 4

Data and model prediction of glucose runout experiments. Comparison of the model (dutoit2) predicted and experimentally measured values of the change in extracellular (to the iRBC) metabolites over time for the consumption of (A) glucose and production of (B) lactate and (C) glycerol. Blue lines with bands indicate the model’s predicted activity for the average measured trophozoite volume ± SD at 32 to 35 h post-invasion.

Figure 5

Steady-state lactate flux as a function of parasitemia. Steady-state lactate flux as a function of parasitemia determined for trophozoite-stage infected red blood cells in 5 mM glucose. For parasitemia of 30% and above, dilutions were made from the enriched culture (100%), and for lower values of parasitemia, cultures were grown to the % parasitemia shown. Solid blue line with bands indicates the dutoit3 model predicted activity for the average measured trophozoite volume ± SD at 32 to 35 h post-invasion (32 ± 8 fl). The blue dashed line indicates the model prediction when using the 45 fl volume of (29).

Figure 6

Steady-state lactate fluxes as a function of extracellular glucose. Steady-state lactate flux of 100% trophozoite iRBC incubations (black symbols) and uRBCs (red symbols) as a function of extracellular (to the RBCs) glucose, normalized to activity per cell. The model used was dutoit3 with either 0 (red line) or 100% parasitemia for the solid blue line with bands indicating the model prediction for the average measured trophozoite volume ± SD at 32 to 35 h post-invasion (32 ± 8 fl). The blue dashed line indicates the model prediction when using the 45 fl volume of (22).

Figure 7

Comparison of measured and predicted activity of different parasite stages. Comparison of measured and predicted activity of different parasite stages. Model-predicted (blue bars) and experimentally measured fluxes (orange bars) of the iRBC are shown per iRBC (A), and normalized to the respective parasite volumes in the iRBC (B). Experimental results are shown as mean ± SEM. Bars for model predictions indicate the prediction using the mean volume of the parasite stage, with error bars showing the range obtained for the mean ± SD of the volume. The experimental results in (B) were normalized to the average volumes of the relevant parasite stages (5 fl ring, 32 fl trophozoite, 74 fl schizont).

Figure 8

Model simulations for competitive inhibitor titrations of selected glycolytic enzymes. Solid and dashed curves indicate the activities of the parasite and red blood cell enzymes, respectively. Color is used to indicate orthologs. Activities are expressed as % value of uninhibited activity and plotted as a function of the inhibitor concentration (normalized to its K_i value). Disks and circles indicate the [I]/K_i values (x-axis) and the corresponding % activity of the enzyme (y-axis) where inhibition leads to a 50% inhibition of the glycolytic flux in the parasite and red blood cell respectively. Note that for 50% inhibition of flux in the parasite, the required [I]/K_i for a glucose transport inhibitor is ±10 (blue disk) whereas for the red blood cell, this value would be close to 10⁵ (blue circle).

Figure 9

Cytochalasin B inhibitor titrations of glycolytic flux in uninfected and infected red blood cells at different glucose concentrations. Glycolytic flux was measured in terms of the lactate production rate at different cytochalasin B concentrations and is shown as a % of the uninhibited flux. Flux measurements were performed for uninfected (blue) and 100% enriched iRBCs at different glucose concentrations (yellow, green, and orange). The dutoit1 model predictions of the lactate production fluxes are indicated by the solid curves (yellow, green, and orange). The blue line indicates the 0% inhibition seen in the uRBC. For comparison, purple markers show the results of inhibition in the isolated trophozoite, and the purple curve the corresponding vanniekerk1 model simulation (22). Error bars are indicative of SEM where the measurement was performed in two or more biological repeats.

Figure 10

Scheme showing the different volume compartments in the dutoit3 model. The blood plasma/culture volume vBld, the uninfected RBC volume Vrbcu, the infected RBC volume Vrbci, and the parasite volume Vpf. Each volume compartment can consist of multiple cells but is simulated as a total volume.