Insights into physiological roles of unique metabolites released from Plasmodium-infected RBCs and their potential as clinical biomarkers for malaria

Abstract

Plasmodium sp. are obligate intracellular parasites that derive most of their nutrients from their host meaning the metabolic circuitry of both are intricately linked. We employed untargeted, global mass spectrometry to identify metabolites present in the culture supernatants of P. falciparum-infected red blood cells synchronized at ring, trophozoite and schizont developmental stages. This revealed a temporal regulation in release of a distinct set of metabolites compared with supernatants of non-infected red blood cells. Of the distinct metabolites we identified pipecolic acid to be abundantly present in parasite lysate, infected red blood cells and infected culture supernatant. Further, we performed targeted metabolomics to quantify pipecolic acid concentrations in both the supernatants of red blood cells infected with P. falciparum, as well as in the plasma and infected RBCs of P. berghei-infected mice. Measurable and significant hyperpipecolatemia suggest that pipecolic acid has the potential to be a diagnostic marker for malaria.

Figure 1

Global metabolic profiling of supernatants of P. falciparum-infected erythrocytes. (a) Volcano plot depicting fold change (X-axis) in 141 metabolites between normal RBCs and P. falciparum-infected erythrocytes (iRBCs) with their statistical significance indicated by the Y-axis. The scatter points on the left and right of zero in the fold change (FC) axis represent down-regulated and up-regulated metabolites, respectively. Circles towards the top in the Y-axis represent metabolites that have been affected with statistical significance (P- value < 0.05; separated by a grey line) while the circles towards the lower half of the Y-axis represent metabolites with P-value > 0.05. Each circle is representative of a metabolite. Red circles indicate metabolites with a FC > 2 and P-value < 0.05. Green circles indicate metabolites with a FC < 2 and P-value < 0.05. Yellow circles indicate metabolites that are affected with a P-value < 0.05, but not significant with respect to FC threshold. Colored black circles indicate metabolites that are statistically insignificant (P- value > 0.05) and have insignificant FC. Open black circles indicate metabolites that have a fold change < or >2, but are not statistically significant (P-value < 0.05). The data presented correspond to the statistical analysis of n = 5 independent biological replicates and 4 technical replicates. (b) Pathway impact analysis plot showing metabolic pathways affected across time and infection with statistical significance. The radius of the circle represents the number of metabolites identified in each pathway. The color gradient from white to red is representative of the cumulative statistical significance of the pathway and the individual pathway impact on the overall metabolic change observed. (c) Pie chart representing the percentage of metabolites involved in different processes that changed between the normal group and the infected group in Plasmodium culture supernatant. Most of the metabolites that were up/down regulated in the infected group belonged to the lipid (34%) and amino acid metabolic pathways (24%). Redox metabolites, xenobiotics and carbohydrate metabolism also contributed a significant percentage of metabolites that exhibited a change in the two groups. (d) Heat map representation of changes in metabolite levels measured in culture supernatants of P. falciparum-infected RBCs at ring, trophozoite and schizont stages. Green rectangles indicate a down regulation in the metabolite level and red rectangles indicate an increase in metabolite level. Most metabolites show an increase in their levels as P. falciparum intra-erythrocyte development progresses to the schizont stage. (e) The schematic representation of a global metabolic map showing the 141 metabolites identified from P. falciparum-infected RBC culture supernatants. Colours in the map represent the different pathways in which these metabolites are involved. In each of these pathways, the highlighted metabolites have been identified in the current study.

Figure 2

Metabolite profiling of different asexual stages of P. falciparum-infected RBCs. Volcano plot combing statistical significance and fold change observed in metabolites in the culture supernatant of P. falciparum-infected RBC compared to normal RBC at different time points of. Red circles indicate metabolites with a fold change >2 and statistical significance P < 0.05. Green circles indicate metabolites with a fold change <2 and statistical significance P < 0.05. Yellow circles indicate metabolites, which have statistical significance, but insignificant fold change values. Coloured black circles indicate metabolites that are statistically not significant and have insignificant fold change. Open black circles indicate metabolites that have a fold change <or >2, but are not statistically significant. (a) Volcano plot for affected metabolites between culture supernatant of uninfected RBCs and ring-infected RBCs. Only two metabolites show a significant upregulation and one metabolite shows a significant down regulation. (b) Volcano plot depicting affected metabolites between culture supernatant of uninfected RBCs and trophozoite-infected RBCs. A few metabolites show significant upregulation. (c) Volcano plot depicting metabolites affected between culture supernatant of uninfected RBCs and schizont-infected RBCs. At this stage, many metabolites show significant upregulation implying active change in the host metabolome following P. falciparum-infection of RBC. (d) Venn diagram representing a stage-specific metabolic correlation analysis, with each coloured circle representing a specific P. falciparum intra-erythrocyte developmental stage. It represents the metabolites (P < 0.05) that are affected in the supernatants of RBCs infected at the ring, trophozoite and schizont stages and their uniqueness/overlap between different life cycle stages.

Figure 3

P. falciparum infection of RBC leads to efflux of redox metabolites in the culture supernatant. Box plot for change in level of detected redox metabolites at 8 h, 24 h and 40 h in control RBC vs infected RBC culture medium. Y-Axis of the box plots represents scaled intensity and X-Axis indicates the treatment group. (a) Homocysteine in the medium was ~2-fold elevated at 40 h in infected vs control (P = 0.05, n = 5). (b) The level of the amino acid cysteine was moderately upregulated in the infected RBC culture supernatant at 24 h and 40 h (P = 0.0002, n = 5; P = 0.0001, n = 5). (c) 5-oxoproline was mildly affected by P. falciparum-infection of RBC at 40 h (P = 0.0025, n = 5). (d) Cysteine-Glutathione disulfide was also slightly, but significantly, affected at 40 h (P < 0.0005, n = 5). (e) At 40 h oxidized glutathione was observed to be over 2-fold upregulated in the supernatant of P. falciparum-infected RBC compared to normal RBC supernatant (P = 0.0002, n = 5).

Figure 4

Probable metabolite markers of P. falciparum-infection of RBC. Box plot for change in levels of detected metabolites at 8 h, 24 h and 40 h in the spent medium of control RBC vs P. falciparum-infected RBC. Y-Axis of the box plots represents scaled intensity and X-Axis indicates the treatment group. (a) Plot shows absence of pipecolate in the normal group and a significant presence in the infected group, especially at later time points when parasites are in their late trophozoite/schizont stage (~30-fold upregulated at 24 h and ~60-fold upregulated at 40 h; P < 0.0005; n = 5). (b) Plot shows a complete absence of gamma aminobutyrate in the control RBC and in ring-infected RBC. It was elevated at 24 h (10-fold upregulated; P < 0.0005, n = 5) and 40 h (45-fold upregulated; P < 0.0005, n = 5) exclusively in the P. falciparum-infected RBC group. (c) Box plot shows absence of nicotinate ribonucleoside in the supernatants of control RBC and ring-infected RBC. It was significantly elevated in medium obtained from trophozoite-infected RBC (10-fold upregulated; P < 0.0005, n = 5) and schizont-infected RBC (25-fold upregulated; P < 0.0005, n = 5). (d) Box plot shows absence of alpha-ketoglutarate in non-infected RBC supernatants. It was exclusively present in the spent medium of the infected RBC at 24 h (~7-fold upregulated; P < 0.0005; n = 5) and 40 h (~8- fold; P = 0.0001; n = 5).

Figure 5

Probable effects of extracellular metabolites present in culture supernatant of P. falciparum-infected RBCs on cerebral malaria: The lower part of the schematic model summarizes the biochemical pathways and their interconnectivity operating in P. falciparum for the generation of metabolites: pipecolic acid, AKG, GABA, NAD⁺ and homocysteine. On release of these metabolites in the bloodstream, pipecolic acid (red arrow), GABA (green arrow) and homocysteine (cyan arrow) can possibly breach the blood brain barrier and consequently affect the neurons contributing to cerebral malaria pathogenicity. The known effects of these metabolites in the brain have been described with the above-mentioned color coding. Representative image of brain obtained from licensed version of Microsoft Office 365 Powerpoint® Product ID: 00202-51569-36211-AA678. AKG: alpha ketoglutarate; GABA: Gamma amino butyric acid; red circles and blue circles represent red blood cells and intra-erythrocytic stage of Plasmodium falciparum, respectively.

Figure 6

In vitro and in vivo validation of pipecolic acid as a diagnostic marker for malaria. (a) The levels of pipecolic acid increase with time in spent culture medium (Med) of P. falciparum-infected RBC. At 0 h, medium from both non-infected RBC (nMed) and infected RBC (iMed) show similar levels of pipecolic acid (P = ns, n = 3). A small change was observed between nMed and iMed at 8 h (P = 0.0172, n = 4). At 24 h, a 10-fold change was seen between nMed and iMed (P = 0.0101, n = 4), respectively. At 40 h a ~15-fold increase was observed between nMed and iMed (P = 0.0322, n = 3), respectively. (b) Pipecolic acid was measured in nRBC, iRBC and Pf at 40 hrs. A significant 2-fold increase was observed in iRBC compared to nRBC (P- value = 0.0059, n = 4). Importantly, pipecolic acid was present in P. falciparum lysate (n = 12). (c) Pipecolic acid was measured in the plasma of P. berghei infected mice (n = 4 each set). Mice were divided in 5 groups. Pipecolic acid increased with parasite load; Compared with control mice, those with parasitemia <3%, 3–10%, 10–30% and >30%, the observed fold change in pipecolic acid was 2-fold, 2.5-fold, 4-fold and 6-fold, respectively (P = 0.0002, n = 4). (d) Pipecolic acid was measured in nRBC, iRBC from mice with <3% and 3–10% parasitemia and Pb lysate. A significant 6-fold increase was observed in iRBC compared to nRBC (P- value < 0.0001, n = 4). Importantly, pipecolic acid was present in P. berghei lysate. nMed and nRBC represent the supernatant and saponin lysate of normal RBCs, respectively. iMed and iRBC represent the supernatant and saponin lysate of infected RBCs at the mentioned parasitemias, respectively. Pf and Pb correspond to P. falciparum and P. berghei, respectively.