Decreased Rate of Plasma Arginine Appearance in Murine Malaria May Explain Hypoargininemia in Children With Cerebral Malaria

Abstract

Background. Plasmodium infection depletes arginine, the substrate for nitric oxide synthesis, and impairs endothelium-dependent vasodilation. Increased conversion of arginine to ornithine by parasites or host arginase is a proposed mechanism of arginine depletion.

Methods. We used high-performance liquid chromatography to measure plasma arginine, ornithine, and citrulline levels in Malawian children with cerebral malaria and in mice infected with Plasmodium berghei ANKA with or without the arginase gene. Heavy isotope–labeled tracers measured by quadrupole time-of-flight liquid chromatography–mass spectrometry were used to quantify the in vivo rate of appearance and interconversion of plasma arginine, ornithine, and citrulline in infected mice.

Results. Children with cerebral malaria and P. berghei–infected mice demonstrated depletion of plasma arginine, ornithine, and citrulline. Knock out of Plasmodium arginase did not alter arginine depletion in infected mice. Metabolic tracer analysis demonstrated that plasma arginase flux was unchanged by P. berghei infection. Instead, infected mice exhibited decreased rates of plasma arginine, ornithine, and citrulline appearance and decreased conversion of plasma citrulline to arginine. Notably, plasma arginine use by nitric oxide synthase was decreased in infected mice.

Conclusions. Simultaneous arginine and ornithine depletion in malaria parasite–infected children cannot be fully explained by plasma arginase activity. Our mouse model studies suggest that plasma arginine depletion is driven primarily by a decreased rate of appearance.

Severe malaria caused by infection with Plasmodium falciparum kills 236 000–635 000 people each year, most of whom are children in sub-Saharan Africa [1]. Decreased endothelial nitric oxide (NO) signaling has been implicated in the pathophysiology of severe malaria and may contribute to impaired vasodilation [2] and microcirculatory abnormalities [3]. L-arginine, the substrate for endogenous NO synthesis, is depleted in patients with severe falciparum malaria [2, 4, 5], and L-arginine supplementation is currently undergoing early stage clinical trials as a strategy to restore endothelial NO signaling [2, 6].

Studies of the human parasite P. falciparum and the murine parasite Plasmodium berghei ANKA in culture have suggested that arginine is consumed by parasite-expressed arginase, which converts arginine to ornithine and urea [7]. Host arginase could also contribute to arginine depletion via intracellular activation by cell-free heme [8] or via direct release into plasma during parasite-mediated hemolysis. Consistent with this hypothesis, increased plasma arginase activity has been associated with severe malaria in studies of Indonesian adults [2], Malian children [9], and Tanzanian children [10]. The latter study also demonstrated elevated peripheral blood mononuclear cell arginase 1 messenger RNA expression [10]. However, unlike sickle cell disease, in which the decreased ratio of plasma levels of arginine to ornithine is consistent with elevated arginase activity [11], patients with severe malaria exhibit profound depletion of both arginine and ornithine [2, 12], leaving the ratio of arginine to ornithine levels unchanged [2]. Thus, arginine depletion may not be explained solely by increased arginase activity.

Citrulline comprises the third major urea cycle intermediate alongside arginine and ornithine (Supplementary Figure 1), but the effect of Plasmodium infection on citrulline metabolism is unknown. In this study, we first report arginine, ornithine, and citrulline levels in Malawian children infected with P. falciparum and then examine the contribution of parasite arginase to arginine depletion in P. berghei ANKA–infected mice. Finally, we use heavy isotope–labeled metabolic tracers to measure in vivo synthesis and interconversion of arginine, ornithine, and citrulline in mice during acute P. berghei ANKA infection.

MATERIALS AND METHODS

Human Participation Ethics Statement

Participants or their parent/guardian gave written informed consent to participate in the protocol “Clinicopathological, Magnetic Resonance and Electrophysiological Correlates of Cerebral Malaria,” approved by the College of Medicine Research Ethics Committee, Blantyre, Malawi (P.11/07/593), and the Michigan State University Institutional Review Board, East Lansing, Michigan (institutional review board number 06-1012). The participation of the National Institutes of Health intramural investigators listed on the protocol was covered by a reliance agreement between Michigan State University and the National Institutes of Health Office of Human Subjects Research Protection.

Children were eligible for the cerebral malaria group if they met the following inclusion criteria: an age of 0.5–12 years; a Blantyre coma score (BCS) of ≤2 on admission; detection of P. falciparum on a blood smear; no evidence of meningitis; no improvement in the BCS after correction of hypoglycemia (if present); and fundoscopy-based detection of retinopathy (defined as whitening, hemorrhage, or vessel changes) by a trained clinician within 6 hours of admission [13]. Exclusion criteria included gross signs of malnutrition, head trauma, preexisting brain injury, clinical manifestations of advanced AIDS or human immunodeficiency virus infection, or achievement of a BCS of >2 within 4 hours of admission. The uncomplicated malaria group was prospectively defined as children with detection of P. falciparum on a blood smear who required hospitalization but did not have any of the following severe manifestations of malaria: impaired consciousness, hyperlactatemia, severe anemia, respiratory distress, hypoglycemia, or hyperparasitemia. Healthy afebrile and aparasitemic Malawian children were enrolled as controls.

Animal Studies Ethics Statement

We performed animal experiments at the National Institute of Allergy and Infectious Diseases Comparative Medicine Branch in accordance with Public Health Service Animal Welfare Assurance A4149-01 guidelines, with ethical approval by the National Institutes of Health Animal Care and Use Committee under identification ASP LMVR18E. Confirmatory experiments performed at the Wellcome Trust Sanger Institute are detailed in Supplementary Materials.

Animals, Housing, Diet, and Parasite Inoculations

Ten-week-old C57BL/6J male mice were obtained from the Jackson Laboratory (Bar Harbor, ME) and 10-week-old Swiss Webster carriers for parasite expansion were obtained from Charles River Laboratories (Frederick, MD). All mice were maintained in a cycle of 12 hours of light and 12 hours of darkness and provided with acidified water and autoclaved rodent feed pellets (Teklad Global 18% Protein Extruded Rodent Diet, 2018SX, Harlan Laboratories). Pellets contained 1% L-arginine. Arginase-knockout (argKO) P. berghei ANKA parasites and the wild-type (WT) parent line were a kind gift from Akhil Vaidya (Drexel Medical School, Philadelphia, PA) [7]. DNA was extracted from whole blood with a DNeasy Kit (Qiagen, Manchester, United Kingdom), and genotypes were assessed by polymerase chain reaction (PCR; Supplementary Methods). For all other experiments, P. berghei ANKA parasites were clone RMgm-29 [14]. Parasites were expanded from frozen stocks by intraperitoneal injection of carrier mice. Experimental mice were infected by intraperitoneal injection of 1 × 10⁶ infected erythrocytes diluted in phosphate-buffered saline.

Plasma Collection and High-Performance Liquid Chromatography (HPLC) Analysis of Amino Acids

Terminal anesthesia (250 mg/kg ketamine plus 25 mg/kg xylazine) was administered intraperitoneally, and 700 µL of blood was collected from the abdominal inferior vena cava, using tubes containing tripotassium ethylenediaminetetraacetic acid for anticoagulation. Blood samples were centrifuged at 3000 g for 7 minutes at 4°C, snap frozen on dry ice, and stored at −80°C until analysis. A modified version of the HPLC method by Wu and Meininger was performed [15]. Details are provided in the Supplementary Materials.

Metabolic Tracer Infusions

Detailed methods are provided in the Supplementary Materials. In brief, mice with jugular vein catheters were infused with 25 µL/hour of 0.9% sterile saline (Hospira, Lake Forest, IL). On day 6 after inoculation, infected mice and controls were infused with L-[U-¹³C₆;U-¹⁵N₄]arginine (prime, 70 µmol × kg⁻¹; continuous infusion, 70 µmol × kg⁻¹ × hour⁻¹), L-[ureido-¹³C]citrulline (prime, 7 µmol × kg⁻¹; continuous infusion, 7 µmol × kg⁻¹ × hour⁻¹), and L-[U-¹³C₅]ornithine (prime, 15 µmol × kg⁻¹; continuous infusion, 15 µmol × kg⁻¹ × hour⁻¹). Priming doses were administered over a period of 5 minutes, and continuous infusions were delivered for a period of 4 hours after which mice were rapidly anesthetized for plasma collection.

Quadrupole Time-of-Flight Liquid Chromatography/Mass Spectrometry (Q-TOF LC/MS) of Metabolic Tracer Enrichment

Underivatized plasma arginine, citrulline and ornithine and heavy isotope-labeled tracers were detected by Q-TOF LC/MS, using an analytical approach adapted from published methods [16, 17]. Detailed methods are provided in the Supplementary Materials.

Calculations of Flux in Stable Isotope Tracer Experiments

Full equations used to calculate the rates of arginine, ornithine, and citrulline appearance and interconversion are provided in Supplementary Methods. In general, the rate of appearance and rate of conversion were calculated using 2 equations.

Equation 1 was

$$Q_X=\frac{I_{X_{m+a}}\times E_{I{X}_{m+a}}}{TT{R}_{X_{m+a}}}-(I_{X_{m+a}}\times E_{I{X}_{m+0}}),$$

where Q_X is the rate of appearance of metabolite X, $I_{X_{m+a}}$ is the rate at which labeled metabolite X tracer with mass m + a is infused, $E_{I{X}_{m+a}}$ is the proportional enrichment of fully labeled metabolite X tracer in the infusion solution, $E_{I{X}_{m+0}}$ is the proportional enrichment of fully unlabeled metabolite X tracer in the infusion solution, and $\text{TT}{\text{R}}_{X_{m+a}}$ is the tracer to tracee ratio of labeled metabolite X tracer with mass m + a to unlabeled metabolite X with mass m + 0. For tracers labeled at multiple positions, $E_{I{X}_{m+0}}\approx 0$; therefore, the term $I_{X_{m+a}}\times E_{I{X}_{m+0}}\approx 0$.

Equation 2 was

$$Q_{X\to Y}=\left(\frac{\text{TT}{\text{R}}_{Y_{m+a}}}{\text{TT}{\text{R}}_{X_{m+a}}}\right)\times Q_Y\times \left(\frac{{\displaystyle Q_X}}{Q_X+I_{X_{m+a}}\times E_{I{X}_{m+0}}}\right),$$

where Q_X→Y is the real rate of substrate-to-product flux, $\text{TT}{\text{R}}_{Y_{m+a}}$ is the tracer-tracee ratio of product Y carrying m + a label derived from substrate X, Q_Y is the total rate of metabolite Y appearance, and all other terms are as described in Equation 1. For tracers labeled at multiple positions, $E_{I{X}_{m+0}}\approx 0$ ; therefore, the term $Q_x/(Q_x+I_{X_{m+a}}\times E_{I{X}_{m+0}})\approx 1$.

The rate of arginine recycling was calculated from the differential rates of carbon skeleton and functional group elimination, as previously described in the context of methionine metabolism [18], and was used to adjust rates of appearance/conversion of other metabolites where required (Supplementary Methods).

Statistical Analyses

All data are presented as median values (interquartile ranges). Statistical analyses were performed with GraphPad Prism 6 software (GraphPad Software, San Diego, CA), and a P value of < .05 was considered significant.

RESULTS

Plasma Arginine, Citrulline, and Ornithine Are Depleted in Plasmodium-Infected Humans and Mice

We used HPLC to assess plasma amino acid concentrations at admission (n = 30) and at follow-up day 28 (n = 17) in Malawian children with cerebral malaria. Children with uncomplicated malaria (n = 29) and healthy children (n = 27) served as comparison groups. Plasma concentrations of arginine, ornithine, and citrulline were significantly decreased in patients with uncomplicated malaria and those with severe malaria, compared with healthy controls (Figure 1). However, owing to similar degrees of arginine and ornithine depletion, the arginine to ornithine ratio was not significantly altered in patients with uncomplicated or cerebral malaria (Table 1). In contrast, the phenylalanine concentration was significantly increased in children with uncomplicated or severe malaria (Figure 1). At a follow-up visit 28 days after admission, children who had recovered from severe malaria demonstrated significant increases in plasma arginine, ornithine, and citrulline levels as compared to admission values (Figure 1).

Figure 1.

Arginine, ornithine, and citrulline were acutely depleted in Malawian children with malaria. Amino acid concentrations were measured in plasma samples obtained at admission and at follow-up visits 28 days later from patients with cerebral malaria. Patients with uncomplicated malaria and healthy Malawian children were also studied. The Mann–Whitney test was used to compare 29 patients with cerebral malaria to 30 with uncomplicated malaria or 27 healthy controls. The Wilcoxon test was used for pair-wise comparison of admission and recovery values in patients with cerebral malaria (15 observations). *P < .05, **P < .01, and ***P < .001. Abbreviation: NS, not significant (P > .05).

Table 1.

Baseline Characteristics and Laboratory Values of Subjects, According to Clinical Status

Characteristic	Healthy		Uncomplicated Malaria		Cerebral Malaria
	Value	Subjects, No.	Value	Subjects, No.	Value	Subjects, No.
Age, mo	77 (64.8–83.8)	27	32.6 (22.4–75.7)	33	43.5 (24.4–50.9)	39
Female sex, no. %	15 (52)	29	11 (33)	33	24 (65)	37
Height, cm	112.5 (107.0–119.5)	27	86.0 (77.0–97.0)	29	93 (81–104)	36
Weight, kg	18.2 (16.4–20.5)	27	12.0 (10.0–18.5)	31	11.9 (9.4–15.2)	39
Temperature, °C	36.2 (35.8–36.7)	27	37.6 (36.7–38.6)	32	37.8 (37.3–38.4)	39
Blantyre coma score						39
0	…	…	…	…	3	…
1	…	…	…	…	16	…
2	…	…	…	…	20	…
Parasitemia frequency, qualitative		29		33		39
Negative findings	29	…	0	…	2	…
Level 1	…	…	1	…	4	…
Level 2	…	…	3	…	6	…
Level 3	…	…	11	…	9	…
Level 4	…	…	17	…	14	…
Level 5	…	…	1	…	4	…
Parasite count, parasites/µL	…		…		36 160 (4380–188 400)	37
Hematocrit, %	37 (34–39)	25	32 (29–36)	32	22 (16–26)	38
Glucose level, mmol/L	4.9 (4.6–5.7)	26	6.0 (5.2–6.5)	30	5.6 (4.8–6.7)	38
Lactate level, mmol/L	2.1 (1.7–2.3)	26	2.2 (1.8–3.0)	29	3.8 (3.0–8.9)	38
Arginine level, µmol/L	64.7 (56.5–73.5)	27	25.7 (21.5–33.8)	30	27.4 (22.5–34.9)	29
Ornithine level, µmol/L	41.9 (36.3–49.7)	27	21.6 (14.5–27.1)	30	19.9 (16.6–25.8)	29
Citrulline level, µmol/L	23.6 (19.1–28.3)	27	10.9 (8.0–16.3)	30	6.9 (6.0–7.9)	26
Phenylalanine level, µmol/L	51.2 (42.2–60.3)	27	115.2 (76.0–157.1)	30	118.2 (87.1–171.5)	29
Arginine:ornithine ratio	1.49 (1.04–1.79)	27	1.27 (0.96–1.79)	30	1.55 (1.02–1.92)	29

Data are median values (interquartile ranges), unless otherwise indicated.

We then assessed plasma amino acids in C57BL/6 mice as P. berghei ANKA infection progressed over 6 days after inoculation (Figure 2A and Supplementary Figure 2). By day 6, plasma levels of arginine, citrulline, and ornithine were significantly decreased in infected mice, compared with uninfected mice (Figure 2A), but the ratio of arginine to ornithine remained unchanged (median, 2.4 [IQR, 1.2–3.5] for infected mice and 1.7 [IQR, 1.1–3.1] for uninfected mice; P > .52). The plasma level of phenylalanine was elevated on day 6 after inoculation (Figure 2A). These changes in plasma amino acid levels were similar to those observed in children with cerebral malaria (Figure 1).

Figure 2.

Depletion of plasma arginine, ornithine, and citrulline during Plasmodium berghei infection. A, Amino acids were measured by high-performance liquid chromatography in plasma from mice culled on days 2, 4, or 6 after inoculation with P. berghei ANKA. Data represent results from 3 independent experiments (uninfected, n = 24; infected, day 2, n = 20; infected, day 4, n = 21; and infected, day 6, n = 20). B, Plasma amino acids were assessed on day 6 after inoculation from mice infected with wild-type (WT) or arginase knockout (argKO) P. berghei ANKA parasites. Data represent results from 10 mice per group. C, Uninfected controls not subjected to dietary restriction were provided with ad libitum access to food. Uninfected mice subjected to dietary restriction received only as much food as was consumed each day by a paired mouse infected with P. berghei ANKA. Data are pooled from 6 experiments, with 22 uninfected controls with ad libitum access to food and 19 pairs of infected mice and uninfected mice subjected to dietary restriction. Box plots depict median values and interquartile ranges. Whiskers extend to the highest and lowest values within 1.5 times the 75th and 25th percentiles, and values outside this range were plotted as individual points (by the Tukey method). *P < .05, **P < .01, and ***P < .001, by the Kruskal–Wallis test, followed by the Dunn multiple comparisons test (A), the Mann–Whitney test (B and C; uninfected with vs uninfected without dietary restriction), or the Wilcoxon matched pairs signed rank test (C; uninfected with dietary restriction vs infected). Abbreviation: NS, not significant (P > .05).

Evaluation of Parasite Arginase Level as a Possible Cause of Arginine Depletion

P. falciparum and P. berghei ANKA both express a functional arginase enzyme that produces ornithine from arginine in vitro [7, 19]. To determine whether parasite-expressed arginase is required for arginine depletion in vivo, we analyzed plasma amino acids from C57BL/6 mice infected with argKO or WT parasites. The presence of argKO was confirmed by PCR detection of parasite DNA isolated from the mice at the end of the experiment (Supplementary Figure 3). The disease severity caused by the argKO parasite was similar to that caused by the WT parasite, as evidenced by behavioral score and survival; however, the argKO parasite density was lower (Supplementary Figure 4). The median plasma arginine levels did not differ between mice infected with argKO parasites and those infected with WT parasites (44.7 µmol/L [IQR, 40.3–49.5 µmol/L] and 48.9 µmol/L [IQR, 41.3–52.4 µmol/L], respective; P = .33; Figure 2B). argKO parasites caused a similar degree of ornithine and citrulline depletion (Figure 2B). These results demonstrate that parasite arginase is not required for arginine depletion in P. berghei ANKA–infected mice.

Evaluation of Reduced Dietary Intake as a Possible Cause of Arginine Depletion

P. berghei ANKA–infected mice consumed less food on days 5 and 6 after inoculation (Supplementary Figure 5). To isolate the effect of decreased dietary intake from other infection-related changes in arginine metabolism, we limited dietary intake in an additional group of uninfected mice to match the intake of a priori–paired P. berghei ANKA–infected mice in 24-hour intervals. We found that dietary restriction induced weight loss (Supplementary Figure 5) and moderate plasma arginine and citrulline depletion in uninfected mice (Figure 2C). However, P. berghei ANKA infection had a greater effect on arginine and citrulline levels than dietary restriction alone (P < .01; Figure 2C).

Quantitative Assessment of Arginine Appearance, Use, and Interconversion

To assess real-time arginine metabolism in vivo, we infused mice with ¹³C- and ¹⁵N-labeled arginine, citrulline, and ornithine tracers and measured plasma enrichment by Q-TOF LC/MS. The total rate of arginine appearance in plasma, representing inputs from dietary absorption, synthesis from citrulline, and proteolysis, was markedly diminished in P. berghei–infected mice (median, 345.0 µmol × kg⁻¹ × hour⁻¹ [IQR, 217.8–436.1 µmol × kg⁻¹ × hour⁻¹]), compared with uninfected controls (median, 503.3 µmol × kg⁻¹ × hour⁻¹ [IQR, 434.8–542.2 µmol × kg⁻¹ × hour⁻¹]) P < .05; Table 2). Dietary restriction alone had a similar impact on arginine appearance (median, 386.5 µmol × kg⁻¹ × hour⁻¹ [IQR, 330.4–440.0 µmol × kg⁻¹ × hour⁻¹]). These results show that P. berghei infection lowers the rate of arginine appearance substantially, and they are consistent with reduced dietary intake as a contributing factor. Our tracer method enabled us to distinguish arginine synthesized from citrulline from arginine derived from other sources, such as proteolysis or enteral absorption. The rate at which arginine was synthesized from citrulline was lower in diet-restricted (P < .05) or P. berghei–infected mice (P < .01), compared with uninfected controls (Table 2 and Figure 3); findings for diet-restricted and P. berghei–infected mice did not differ significantly.

Table 2.

Rates of Urea Cycle Metabolic Flux in Uninfected and Plasmodium berghei ANKA–Infected Mice

Variable	Uninfected, Median (IQR)		Infected, Median (IQR)
	Unlimited Access to Food	Dietary Restriction	Unlimited Access to Food
Arginine appearance
Total
QArgT, µmol × kg−1 × h−1	503.3 (434.8–542.2)	386.5 (330.4–440.0)a	345.0 (217.8–436.1)
From citrulline
QCit→Arg, µmol × kg−1 × h−1	102.5 (76.5–137.6)	59.1 (39.2–71.3)b	40.3 (25.5–59.5)
QArgT, % of total flux	19 (16.2–23.7)	14.4 (13.2–19.1)b	12.6 (10.2–14.1)a
From other sources
Q, µmol × kg−1 × h−1	406.8 (316.3–433.5)	332.9 (273.6–368.3)b	304.1 (190.6–378.2)
QArgT, % of total flux	80.9 (76.3–83.8)	85.6 (80.9–86.9)b	87.4 (85.9–89.8)a
Arginine use
For ornithine synthesis
QArg→Orn, µmol × kg−1 × h−1	106.4 (87.2–117.7)	115.2 (78.4–126.6)	107.4 (59.4–126.3)
QArgT, % of total flux	21.1 (19.9%–22.7%)	26.1 (23.7–32.9)a	29.2 (26.8–33.8)
For nitric oxide synthesis
QArg→Cit, µmol × kg−1 × h−1	1.17 (0.83–2.03)	1.13 (0.86–1.28)	0.66 (0.49–0.82)a
QArgT, % of total flux	0.24 (16–0.37)	0.28 (0.24–0.31)	0.16 (0.15–0.24)
Ornithine appearance
Total
QOrnT, µmol × kg−1 × h−1	163.0 (149.5–187.8)	131.8 (105.3–162.3)	116.3 (66.6–126.6)b
From arginine
QArg→Orn, µmol × kg−1 × h−1	106.4 (87.2–117.7)	115.2 (78.4–126.6)	107.4 (59.4–126.3)
QOrnT, % of total flux	65.2 (54.9–71.9)	80.4 (70.9–83.3)b	94.1 (92.2–101.2)a
From other sources
Q, µmol × kg−1 × h−1	63.1 (45.1–85.2)	30.0 (20.6–45.2)a	5.6 (1.8–8.8)a
QOrnT, % of total flux	34.8 (28.2–45.2)	19.6 (16.7–29.1)b	5.9 (−1.2–7.8)a
Ornithine use
For citrulline synthesis
Q, µmol × kg−1 × h−1	22.9 (19.2–27.9)	17.8 (13.2–21.3)b	15.7 (6.2–17.4)b
QOrnT, % of total flux	13.4 (11.8–17.1)	13.5 (12.5–14.3)	11.9 (10.0–14.1)b
Citrulline appearance
Total
QCitT, µmol × kg−1 × h−1	107.8 (90.6–125.6)	67.5 (60.5–73.0)c	44.2 (25.4–52.7)a
From ornithine
QOrn→Cit, µmol × kg−1 × h−1	22.9 (19.2–27.9)	17.8 (13.2–21.3)b	15.7 (6.2–17.4)b
QCitT, % of total flux	21.7 (20.5–24.6)	27.2 (23.0–32.0)	30.3 (27.4–34.7)
From other sources
Q, µmol × kg−1 × h−1	83.2 (71.5–101.4)	47.0 (42.8–53.5)c	29.3 (19.3–35.7)a
QCitT, % of total flux	78.3 (75.4–79.5)	72.8 (68.0–77.0)b	69.7 (65.3–72.6)
Citrulline use
For arginine synthesis
QCit→Arg, µmol × kg−1 × h−1	102.5 (76.5–137.6)	59.1 (39.2–71.3)b	40.3 (25.5–59.5)
QCitT, % of total flux	97.5 (74.6–109.5)	94.6 (55.8–118.3)	110.4 (91.0–119.7)

Total rates of appearance of plasma arginine (Q_ArgT), ornithine (Q_OrnT), and citrulline (Q_CitT) were calculated from plasma enrichment of continuously infused heavy isotope-labeled metabolic tracers, as measured by quadrupole time-of-flight liquid chromatography–mass spectrometry. Rates of conversion among urea cycle intermediates were calculated on a µmol × kg⁻¹ × hour⁻¹ basis and as a percentage of total flux for each metabolite. Uninfected mice subjected to dietary restriction received only as much food as was consumed each day by a paired mouse infected with P. berghei ANKA. Uninfected controls not subjected to dietary restriction were provided with ad libitum access to food. Data are pooled from 3 experiments for a total of 10 uninfected controls with ad libitum access to food and 10 pairs of infected mice and uninfected mice subjected to dietary restriction. Tracer infusions were performed on day 6 after inoculation. Uninfected controls with or without dietary restriction were compared by the Mann–Whitney test; uninfected controls with dietary restriction were compared to infected mice by the Wilcoxon matched pairs signed rank test.

Abbreviation: IQR, interquartile range.

^a P < .01.

^b P < .05.

^c P < .001.

Figure 3.

Metabolic flux of plasma arginine (Arg), ornithine (Orn), and citrulline (Cit). Rates of appearance and interconversion of plasma Arg, Orn, and Cit were calculated in from plasma enrichment of continuously infused heavy isotope–labeled metabolic tracers, as measured by quadrupole time-of-flight liquid chromatography/mass spectrometry. Uninfected mice subjected to dietary restriction received only as much food as was consumed each day by a paired mouse infected with Plasmodium berghei ANKA. Uninfected controls not subjected to dietary restriction were provided with ad libitum access to food. Data are pooled from 3 experiments, with 10 uninfected controls with ad libitum access to food, and 10 pairs of infected mice and uninfected mice subjected to dietary restriction. Tracer infusions were performed on day 6 after inoculation. Box plots depict median values, interquartile ranges, and ranges. Uninfected controls with or without dietary restriction were compared by the Mann–Whitney test; uninfected controls with dietary restriction were compared to infected mice by the Wilcoxon matched pairs signed rank test. All units in µmol × kg⁻¹ × hour⁻¹. *P < .05, **P < .01, and ***P < .001. Abbreviations: NOS, nitric oxide synthase; NS, not significant (P > .05).

Our metabolic labeling strategy (Supplementary Figure 6) allowed us to distinguish arginine used by NO synthase (NOS) from arginine consumed for ornithine synthesis. Dietary restriction alone had no effect on plasma arginine use by NOS, but P. berghei ANKA infection significantly decreased the rate of plasma arginine use by NOS (median, 1.17 µmol × kg⁻¹ × hour⁻¹ [IQR, 0.83–2.03 µmol × kg⁻¹ × hour⁻¹] for the uninfected group, 1.13 µmol × kg⁻¹ × hour⁻¹ [IQR, 0.86–1.28 µmol × kg⁻¹ × hour⁻¹] for the diet-restricted group, and 0.66 µmol × kg⁻¹ × hour⁻¹ [IQR, 0.49–0.82 µmol × kg⁻¹ × hour⁻¹] for the infected group; P < .01; Table 2 and Figure 3). In contrast, the rate at which plasma arginine was converted into plasma ornithine by arginase was unchanged by dietary restriction or P. berghei infection (median, 106.4 µmol × kg⁻¹ × hour⁻¹ [IQR, 87.2–117.7 µmol × kg⁻¹ × hour⁻¹] for the healthy group, 115.2 µmol × kg⁻¹ × hour⁻¹ [IQR, 78.4–126.6 µmol × kg⁻¹ × hour⁻¹] for the diet-restricted group, and 107.4 µmol × kg⁻¹ × hour⁻¹ [IQR, 59.4–126.3 µmol × kg⁻¹ × hour⁻¹] for the infected group; Table 2 and Figure 3). However, the median rate of synthesis of plasma ornithine from sources other than arginine (eg, from proline and glutamate) was greatest among healthy mice (63.1 µmol × kg⁻¹ × hour⁻¹ [IQR, 45.1–85.2 µmol × kg⁻¹ × hour⁻¹]), compared with that among diet-restricted mice (30.0 µmol × kg⁻¹ × hour⁻¹ [IQR, 20.6–45.2 µmol × kg⁻¹ × hour⁻¹]) and infected mice (5.6 µmol × kg⁻¹ × hour⁻¹ [IQR, 1.8–8.8 µmol × kg⁻¹ × hour⁻¹]; P < .01 vs diet-restricted mice; Table 2 and Figure 3).

The total median rate of appearance of citrulline in plasma decreased substantially across groups, with values of 107.8 µmol × kg⁻¹ × hour⁻¹ (IQR, 90.6–125.6 µmol × kg⁻¹ × hour⁻¹) in healthy mice, 67.5 µmol × kg⁻¹ × hour⁻¹ (IQR, 60.5–73.0 µmol × kg⁻¹ × hour⁻¹) in diet-restricted mice, and 44.2 µmol × kg⁻¹ × hour⁻¹ (IQR, 25.4–52.7 µmol × kg⁻¹ × hour⁻¹) in P. berghei–infected mice (P < .01 vs diet-restricted mice; Table 2 and Figure 3). This was due in part to decreased conversion of ornithine to citrulline but predominantly to decreased enteral synthesis/absorption of citrulline across groups (median, 83.2 µmol × kg⁻¹ × hour⁻¹ [IQR, 71.5–101.4 µmol × kg⁻¹ × hour⁻¹] in healthy mice, 47.0 µmol × kg⁻¹ × hour⁻¹ [IQR, 42.8–53.5 µmol × kg⁻¹ × hour⁻¹] in diet-restricted mice, and 29.3 µmol × kg⁻¹ × hour⁻¹ [IQR, 19.3–35.7 µmol × kg⁻¹ × hour⁻¹] in infected mice; P < .001 vs diet-restricted mice; Table 2 and Figure 3).

DISCUSSION

In this study, we identified simultaneous depletion of arginine, ornithine, and citrulline in a case-control study of Malawian children with cerebral malaria and in a mouse model of severe malaria. In both humans and mice, we observed that Plasmodium infection depleted arginine and ornithine to similar degrees, leaving the arginine to ornithine ratio unchanged. This finding is inconsistent with a primary increase in arginase activity, which would be expected to produce ornithine and decrease arginine to ornithine ratios. We also found that P. berghei ANKA–expressed arginase is not required to induce arginine depletion in infected mice. These findings suggest that changes in host metabolism are greater in magnitude than the effect of parasite arginase, which was previously shown to contribute to arginine depletion in vitro [7]. Real-time metabolic tracer analysis in our mouse model demonstrated that the in vivo rate of conversion of plasma arginine to ornithine was unchanged during infection, demonstrating that arginine catabolism is not the primary mechanism of plasma arginine depletion. Instead, decreased arginine and citrulline appearance seem to be primarily responsible for plasma arginine depletion. Our results do not directly contradict prior observations of increased host- and parasite-derived arginase activity [2, 7, 10]; increased arginase activity may be required to preserve the rate of conversion of plasma arginine to ornithine in the setting of low arginine levels. However, our mouse model results demonstrate that any increase in arginase activity is secondary to the major effect of decreased arginine and citrulline appearance.

Previous clinical studies have reported arginine and ornithine depletion in adults [2] and children [12] with severe malaria, but ours is the first study to identify citrulline depletion. Our mouse model studies suggest that citrulline depletion is driven primarily by decreased appearance from enteral sources, with a minor contribution from decreased citrulline synthesis from plasma ornithine. Recent work has demonstrated that citrulline contributes to NOS substrate availability, owing to rapid recycling to arginine by ASS1 and ASL enzymes in complex with NOS [20–22]. Our metabolic flux data demonstrate that >95% of plasma citrulline is used for plasma arginine synthesis. Taken together, these results suggest that citrulline supplementation could simultaneously improve arginine availability and restore citrulline as a secondary NOS substrate pool [2, 6]. Our data suggest that ornithine depletion is largely driven by decreased arginine appearance from nonarginine sources, such as glutamate or proline. While our tracer method allowed us to quantify conversion of plasma ornithine to citrulline, we were not able to assess other metabolic fates of ornithine. Imbalanced net conversion to proline, glutamate, or polyamines could also contribute to ornithine depletion and should be examined in future studies.

Uninfected mice subjected to dietary restriction exhibited decreased rates of arginine, ornithine, and citrulline appearance. However, plasma arginine, ornithine, and citrulline were further decreased in P. berghei ANKA–infected mice, compared with diet-restricted mice, suggesting that P. berghei ANKA infection produces metabolic disturbances beyond suppression of feeding behavior. Arginine absorbed in the enterocyte may be transported directly to the plasma or converted to citrulline by enteral arginase and ornithine transcarbamylase. Autopsy studies in patients have identified sequestration of infected erythrocytes in mucosal and villous capillaries, which could interfere with normal gut physiology [23, 24]. Further studies are therefore required to investigate the effects of Plasmodium infection on enteral absorption. Dysregulation of transport systems such as the cationic amino acid transporter enzyme family during severe malaria could also result in intracellular sequestration of arginine and ornithine. Our data suggest that impaired protein catabolism is not a major mechanism of decreased arginine appearance, because the phenylalanine level, which is also released during protein degradation, was elevated in Plasmodium-infected patients and mice in our study, as well as in prior studies [12, 25]. Elevation of the plasma phenylalanine level may reflect accelerated protein catabolism or impaired downstream hydroxylation to tyrosine, which cannot be distinguished with static plasma measurements. Future studies may therefore benefit from the inclusion of phenylalanine and tyrosine tracers to assess phenylalanine flux.

Our observation of reduced plasma arginine use by NOS provides the first in vivo metabolic evidence that NO synthesis is impaired during Plasmodium infection. A prior study estimated NO synthesis on the basis of measurements of N¹⁵-labeled urinary nitrate in children with uncomplicated malaria, but without comparison to a healthy control group [26]. Planche et al [26] estimated mean NO production (±SD) to be 0.83 ± 0.50 µmol × kg⁻¹ × hour⁻¹, a value similar to the median value of 0.66 µmol × kg⁻¹ × hour⁻¹ (IQR, 0.49–0.82 µmol × kg⁻¹ × hour⁻¹), we measured in P. berghei–infected mice. Inadequate NO synthesis may contribute to pathology of the vascular endothelium during severe malaria, including impaired vasodilation [2], dysregulation of tight junctions and loss of endothelial barrier integrity [27–30], release of procoagulants [31–33], and increased adhesion molecule expression, which facilitates cytoadherence of parasitized erythrocytes [34–36]. In combination, these factors could result in microcirculatory occlusion and endothelial compromise, which have been directly observed in the retina [37] and rectal mucosa [3] of children with cerebral malaria and via cranial window microscopy in P. berghei–infected mice [38, 39].

In addition to substrate depletion, other enzymatic and metabolic factors may limit NO synthesis in severe malaria. Studies of brain tissue from P. berghei ANKA–infected mice have shown evidence of direct inactivation of the NOS enzyme complex, including decreased eNOS phosphorylation and proportional increases in eNOS and nNOS monomers [40]. NO synthesis could also be impaired by insufficient availability of the essential NOS cofactor tetrahydrobiopterin [41, 42]. In addition, relative accumulation of the endogenous NOS inhibitor asymmetric dimethylarginine has been observed in both children [5, 43] and adults [44] with severe malaria.

Several important limitations to this study should be considered. First, in our clinical study, we did not have sufficient statistical power to rule out a decrease in the arginine to ornithine ratio and did not attempt to reproduce prior clinical assessments of arginase activity or macrophage arginase 1 messenger RNA expression. Second, our mouse model did not account for the possible contributions of release/activation of host erythrocyte arginase [2, 8, 45] because murine erythrocytes do not demonstrate significant arginase activity [46]. Third, in contrast to previous studies [7], we found a significant growth deficit in argKO parasites (Supplementary Figure 3), yet these parasites demonstrated no difference in virulence or capacity to deplete plasma arginine. Fourth, we relied on a previous study in mice that reported plateau enrichment of arginine, citrulline, and ornithine tracers after 3 hours of infusion, using comparable priming and maintenance doses [47]. Fifth, our tracer method does not account for intracellular metabolic steps that occur without free exchange between intracellular and plasma compartments. In particular, protein-derived arginine that is degraded by arginase II [48, 49] before reaching the plasma compartment would be measured as a decreased rate of appearance of plasma arginine. However, the decreased rate of appearance of ornithine in plasma argues against intracellular arginine degradation as a contributor to arginine depletion. Last, differences in metabolism between P. falciparum and P. berghei ANKA parasites and between murine and human hosts may limit clinical inference from our animal model studies, highlighting the need to reproduce similar tracer experiments in P. falciparum–infected patients.

In summary, a decreased rate of arginine appearance, rather than increased catabolism of arginine, is the primary explanation for the depletion of plasma arginine in murine severe malaria. Our findings imply that inhibition of plasma arginase may be an ineffective approach to restoring arginine availability and could further decrease plasma ornithine concentrations. Supplementation with citrulline may represent a more effective means of restoring NOS substrate pools, since nearly 100% of citrulline is converted to arginine in vivo. Our observation that malaria parasite infection reduced the use of plasma arginine by NOS suggests that NO synthesis may be limited by factors other than low substrate availability. Therefore, therapeutic restoration of NO signaling may ultimately require use of NOS-independent pathways.

Supplementary Data

Supplementary materials are available at http://jid.oxfordjournals.org. Consisting of data provided by the author to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the author, so questions or comments should be addressed to the author.