Complement System Activation Is a Plasma Biomarker Signature during Malaria in Pregnancy

Abstract

Malaria in pregnancy (MiP) is a public health problem in malaria-endemic areas, contributing to detrimental outcomes for both mother and fetus. Primigravida and second-time mothers are most affected by severe anemia complications and babies with low birth weight compared to multigravida women. Infected erythrocytes (IE) reach the placenta, activating the immune response by placental monocyte infiltration and inflammation. However, specific markers of MiP result in poor outcomes, such as low birth weight, and intrauterine growth restriction for babies and maternal anemia in women infected with Plasmodium falciparum are limited. In this study, we identified the plasma proteome signature of a mouse model infected with Plasmodium berghei ANKA and pregnant women infected with Plasmodium falciparum infection using quantitative mass spectrometry-based proteomics. A total of 279 and 249 proteins were quantified in murine and human plasma samples, of which 28% and 30% were regulated proteins, respectively. Most of the regulated proteins in both organisms are involved in complement system activation during malaria in pregnancy. CBA anaphylatoxin assay confirmed the complement system activation by the increase in C3a and C4a anaphylatoxins in the infected plasma compared to non-infected plasma. Moreover, correlation analysis showed the association between complement system activation and reduced head circumference in newborns from Pf-infected mothers. The data obtained in this study highlight the correlation between the complement system and immune and newborn outcomes resulting from malaria in pregnancy.

Keywords: Plasmodium berghei; Plasmodium falciparum; biomarkers; complement system; malaria; malaria in pregnancy; plasma proteomics.

Figure 1

Comparative label-free proteomics approach in humans and mice with MiP. Translational-comparative approach for label-free proteomics in humans and mice with MiP. (a) Experimental workflow of the murine MiP model. BALB/c females were infected and treated with chloroquine (0.7 mg/animal) and were mated with BALB/c males at forty days post-infection. Pregnancy was detected by weight gain (3–4 g) at thirteen days post-mating. A cesarean was performed on the 19th gestational day, and peripheral blood was collected. Blood was collected in citrate tubes, and plasma was isolated using centrifugation to perform protein digestion (trypsin) and MS analysis. (b) Experimental workflow applied to the cohort of pregnant women that developed MiP. Pregnant women were diagnosed with malaria (gold standard—thin smear and PCR test). After malaria diagnosis, pregnant women were treated with antimalarial drugs, according to Brazilian Ministry of Health (MoH) guidelines. Then, blood from non-infected and infected women was collected, and plasma was isolated to perform protein digestion (trypsin) and then mass spectrometry analysis. No parasitemia was detected in the peripheral plasma at delivery. However, parasite molecules were detected in the placenta (hemozoin).

Figure 2

Clinical parameters of mice and humans with MiP. (a) Clinical parameters of mouse samples—placental vascular space (%) and spleen weight (g) were measured in control (n = 4) and infected (n = 4) groups. Placental vascular space (%) was significantly reduced (p < 0.0001) in infected (n = 12, corresponding to four animals) samples compared to the control group (n = 12, corresponding to four animals). Spleen weight was increased (p = 0.0286) in the infected group (n = 4) compared to the control group (n = 4), an expected effect of Plasmodium berghei ANKA infection. (b) Clinical parameters of the human dataset. Newborn weight (p = 0.0106), head circumference (p < 0.0001), and chest circumference (p = 0.0063) displayed significant differences between the infected (n = 9) and control (n = 9) groups. Mann–Whitney test (p < 0.05) was performed to compare infected and control groups. Statistical analysis was performed using GraphPad Prism version 6. The astherisks indicate statistical significance between the groups. For murine data, **** represents p < 0.0001, * is p = 0.0286. For human data, * represents p = 0.0106, *** for p < 0.0001, and ** p = 0.0063.

Figure 3

Cytokine levels of murine and human samples. (a–d) Cytokine levels of murine samples (n = 4 per group; however, control samples presented N/D values for some cytokines). (e–h) Human peripheral plasma (control = 9, infected = 9). All data were submitted to Robust Regression and Outlier Test (ROUT, Q = 0.1%), and Mann–Whitney was performed to evaluate differences between the two groups. Asterisks indicate statistical significance. Legend: Light pink circles represent the murine in-fected samples and dark blue circles correspond to the human infected samples. * p = 0.0424, ** p = 0.0030.

Figure 4

The number of proteins in murine and human datasets. In mouse and human samples, 279 and 249 proteins were quantified, respectively (represented in dark grey bars), and the regulated proteins in mice (28% of the total proteins) and humans (30% of the total proteins) are depicted in light gray.

Figure 5

Complement system pathways and regulated proteins in the infected group. In the mouse dataset (yellow molecules), classical pathway complement components such as C1q subunits (A, B, and C), C1s, and C2 were up-regulated in the infected group (red triangle). Moreover, alternative pathway complement components were also up-regulated in the infected group (LBP, Factor B, and Factor H). Furthermore, in the MAC complex, Vitronectin (VTN) and C8 subunits (alpha, beta, and gamma) were up-regulated in the murine-infected group. In the human dataset (green molecules), Complement C1r subcomponent-like protein and C4-binding protein from the classical pathway were down-regulated (green square), and Complement C4a was up-regulated in the infected group (red square). From the alternative pathway, Complement Factor D was up-regulated (red square), while the complement inhibitors Complement Factor H-related proteins 1 and 2 were down-regulated in the infected group (green square). Regarding the Lectin pathway, Ficolin-2 was up-regulated in the infected group of the human dataset, and Mannose-binding protein C was identified in both organisms, being up-regulated in the murine-infected group (red triangle) and down-regulated in the human-infected group (green square).

Figure 6

Anaphylatoxin measurement in the peripheral plasma of the infected pregnant women. (a) C3a anaphylatoxin showed a significant increase in the infected group compared to the control group (p < 0.0001). (b) C4a anaphylatoxin also increased in the infected group compared to the control group (p = 0.0005). (c) No statistically significant difference in C5a anaphylatoxin level was detected between the infected and control groups. Astherisks indicate the statistical significance. **** represents p < 0.0001 and *** p = 0.0005.

Figure 7

Matrix of the significant Spearman’s correlation from murine-regulated proteins and clinical parameters.

Figure 8

Matrix of the significant Spearman’s correlation of human regulated proteins and clinical parameters.