Figures
Abstract
Background
Placental malaria (PM) is a leading cause of maternal and infant mortality. Although the accumulation of parasitized erythrocytes (PEs) and monocytes within the placenta is thought to contribute to the pathophysiology of PM, the molecular mechanisms underlying PM remain unclear. Based on the hypothesis that excessive complement activation may contribute to PM, in particular generation of the potent inflammatory peptide C5a, we investigated the role of C5a in the pathogenesis of PM in vitro and in vivo.
Methodology and Principal Findings
Using primary human monocytes, the interaction between C5a and malaria in vitro was assessed. CSA- and CD36-binding PEs induced activation of C5 in the presence of human serum. Plasmodium falciparum GPI (pfGPI) enhanced C5a receptor expression (CD88) on monocytes, and the co-incubation of monocytes with C5a and pfGPI resulted in the synergistic induction of cytokines (IL-6, TNF, IL-1β, and IL-10), chemokines (IL-8, MCP-1, MIP1α, MIP1β) and the anti-angiogenic factor sFlt-1 in a time and dose-dependent manner. This dysregulated response was abrogated by C5a receptor blockade. To assess the potential role of C5a in PM, C5a plasma levels were measured in malaria-exposed primigravid women in western Kenya. Compared to pregnant women without malaria, C5a levels were significantly elevated in women with PM.
Citation: Conroy A, Serghides L, Finney C, Owino SO, Kumar S, Gowda DC, et al. (2009) C5a Enhances Dysregulated Inflammatory and Angiogenic Responses to Malaria In Vitro: Potential Implications for Placental Malaria. PLoS ONE 4(3): e4953. https://doi.org/10.1371/journal.pone.0004953
Editor: Patricia Bozza, Instituto Oswaldo Cruz and FIOCRUZ, Brazil
Received: July 24, 2008; Accepted: February 14, 2009; Published: March 24, 2009
Copyright: © 2009 Conroy et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by CIHR Team Grant in Malaria (KCK), operating grant MT-13721 (KCK), Genome Canada through the Ontario Genomics Institute (KCK), CIHR Canada Research Chairs (KCK; WCL). Recruitment of women in Kenya was supported by NIH RO1050240 (JMM) and could not have been possible without the invaluable participation of parturient women and dedicated Kenyan staff. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Placental malaria (PM) is a major determinant of maternal and infant health in the developing world. PM, especially in primigravidae, can have profound maternal and fetal health consequences, including anemia, stillbirth, premature delivery, intrauterine growth restriction (IUGR) and delivery of low birth weight (LBW) infants [1], [2]. The accumulation of Plasmodium falciparum parasitized erythrocytes (PEs) within the placenta is believed to be an essential step in the pathogenesis of PM [3]. A subpopulation of PEs that express novel variant surface antigens (VSA-PM), specifically adhere to glycosaminoglycan chondroitin sulfate A (CSA) in the placental intervillous space [3]–[5].
PM is also characterized by the infiltration of the placenta with maternal mononuclear cells (mφ). The sequestered PEs release bioactive molecules including those with associated glycosylphosphatidylinositol anchor molecules (pfGPI) that can stimulate maternal mφ [6]–[10] and fetal syncytiotrophoblast [6], [11], [12] to produce inflammatory cytokines, such as TNF and IFN-γ, and β-chemokines including macrophage-inflammatory protein (MIP)-1α, MIP-1β, monocyte chemoattractant protein-1 (MCP-1), and macrophage migration inhibitory factor (MIF). These cytokines and chemokines further recruit, retain, and activate mφ in the placenta [6]–[8], [12], [13]. The resultant accumulation of activated mφ is believed to contribute to adverse birth outcomes [14]–[16]. Although PEs and mφ collect in the placenta, how they may contribute to fetal and placental injury is unknown.
The maintenance of regulated cytokine responses in the placenta is essential in order to prevent rejection of the semiallogenic fetoplacental unit and ensure appropriate immunological responses to infection [17]. This involves a generalized physiological adaptation that results in a bias towards cytokines promoting humoral immunity at the expense of cell-mediated immunity [18]. PM is associated with heightened TH1-type responses that disrupt the balance of cytokines at the maternal-fetal interface [14], [16], [19], [20]. However, the molecular basis underlying PM-associated cytokine dysregulation, placental immunopathology, and adverse birth outcomes is incompletely understood.
The complement system is an essential component of the innate immune response to a number of infectious agents. The complement cascade can be activated by four distinct pathways, three of which converge at the level of the C3 component, leading to the cleavage of C3 and C5 to their activated forms, C3a and C5a, as well as the formation of the terminal membrane attack complex [21], [22]. Recently, a pathway of C5a generation occurring independently of C3 was described [21], [23]. Several lines of evidence have implicated excessive activation of the complement system, notably generation of the potent pro-inflammatory peptide C5a, in mediating deleterious innate host responses to bacterial and fungal infections and contributing to the development of sepsis (reviewed in [22]) [21]–[26]. Sepsis, similar to severe malaria, is a clinical syndrome characterized by systemic inflammation and endothelial activation in response to infection [27], [28]. Elevated C5a levels in sepsis have been implicated in adverse clinical outcomes and death. Blockade of C5a-C5aR activity in animal models of sepsis prevents end-organ injury and improves survival [29]–[31].
Excessive C5a generation has also recently been identified as a critical mediator of placental and fetal injury in a non-infectious mouse model of spontaneous miscarriage and IUGR, via dysregulation of angiogenic factors required for normal placental development [32]. Activated complement components were shown to directly induce the release of soluble fms-like tyrosine kinase-1 (sFlt-1, sVEGFR-1), a potent anti-angiogenic factor that prevents signaling of vascular endothelial growth factor and placental growth factor [33]. Excessive levels of sFlt-1 inhibit placental differentiation and are thought to play a direct role in pre-eclampsia [34]–[36] and C5a-associated IUGR and pregnancy loss [32]. Of note, Muehlenbachs et al. have reported elevated plasma levels of sFlt-1 in primigravid women with PM [37].
Based on the hypothesis that C5a may play a role in PM by generating angiogenic factors that mediate IUGR and fetal loss, we determined whether malaria parasites can activate C5 and if C5a potentiates the induction of sFlt-1, inflammatory cytokines and chemokines associated with adverse pregnancy outcomes. Here we show that CSA- and CD36-binding PEs activate C5, that C5a and pfGPI cooperate to induce an amplified inflammatory and anti-angiogenic response to P. falciparum malaria in vitro, and that primigravid women with PM have elevated circulating C5a levels in vivo.
Methods
P. falciparum culture
The laboratory isolates CS2 (CSA-binding) and E8B (ICAM-1 and CD36 binding) were cultured in vitro as described [38], [39]. Cultures were routinely treated with mycoplasma removal agent (ICN), and were tested and found negative for mycoplasma by PCR.
Mononuclear cell isolation
PBMCs were isolated from healthy volunteers using a Ficoll gradient as previously described [40]. PBMCs were counted using a haemocytometer and cell viability was determined by Trypan blue exclusion test. Cells were plated at a concentration of 2.0×106 cells/mL (48 well plate) in RPMI 1640 medium supplemented with 10% heat inactivated FBS and gentamycin (R10G). Cells were cultured at 37°C in 5% CO2 for the time indicated.
Isolation and purification of GPIs from P. falciparum
HPLC-purified P. falciparum GPI (pfGPI) was isolated [41], conjugated to gold beads and used at a concentration of 300 ng/ml as previously described [10]. Unconjugated gold beads alone were added as controls to wells containing media, or C5a (endotoxin-free; Biovision, USA). All preparations of purified pfGPIs were tested for endotoxin by Limulus amebocyte lysate assay prior to use.
Complement assays
PEs, uninfected red blood cells (uRBCs), or media control were incubated in the presence of 30% serum from malaria naïve donors. After 30 min, supernatants were collected and analyzed for C5 activation. C5a levels in culture supernatants and human plasma were measured by ELISA (R&D Systems).
Cytokine and chemokine assays
PBMCs were cultured with or without recombinant human C5a in addition to P. falciparum culture supernatants (1∶10 dilution) and R10G alone as media control, or HPLC-purified GPI (300 ng/mL) and unconjugated gold beads as a control. C5a concentrations ranging from 1 nM to 100 nM were selected representing a spectrum of C5a levels ranging from physiologically normal levels (<10 nM) to those associated with sepsis (10–100 nM) [25], [42]. For receptor blockade studies pfGPI, the C5aR blocking antibody (Serotec, S5/1, 5 µg/mL) or an isotype control (eBioscience, eBM2a, 5 µg/mL) was added to culture medium just prior to stimulation with 50 nM C5a. Culture supernatants were collected aseptically at times indicated. IL-6, TNF, IL-1β, and IL-10 were measured by Cytometric Bead Array (CBA) (BD Biosciences, human inflammation kit). IL-8, monocyte chemoattractant protein-1 (MCP-1), and macrophage inflammatory protein-1α and - β (MIP-1α and MIP-1β) and sFlt-1 (sVEGFR-1) were measured by ELISA (R&D Systems).
Flow cytometry
PBMCs (2.0×106) were blocked in 20% serum for 20 min at 4°C. Cells were then stained for CD14-APC (eBioscience, 61D3, 100 uL/mL) and CD88-PE or isotype IgG1-PE (BD Biosciences, D53-1473, 150 uL/mL and eBioscience, P3, 150 uL/mL) for 30 min at 4°C. Cells were fixed in 1% paraformaldehyde/PBS overnight and analyzed on a Becton Dickinson FACS Calibur flow cytometer. Data was analyzed using FlowJo software. A minimum of 50,000 events were collected. Positive background PE staining in isotype samples was subtracted from respective CD88-PE stained samples and then reported as specific CD88 staining.
Statistical analysis
Data are expressed as means±SEM, unless otherwise noted. For continuous variables a Student's t test or one-way ANOVA with Bonferonni post-tests was used to compare means. Non-parametric tests (Mann-Whitney or Kruskall-Wallis) were used to compare groups where parametric test conditions were not met. Synergism was determined by the interaction term of a two way ANOVA (C5a*GPI). Synergism in the time course was determined using the cumulative cytokine levels after 48 hours by performing a two way ANOVA and examining the interaction term. Significance in the C5a receptor blockade experiments was assessed by repeated measures ANOVA to determine whether levels of cytokines differed between cells treated with pfGPI alone, or anti-CD88 plus C5a and pfGPI, or student's t-test by comparing the isotype and anti-CD88 co-treated (C5a and pfGPI) groups after correcting for the background levels observed in media controls. In cases where C5a receptor blockade resulted in values of “0”, a one sample t-test was used to determine whether the isotype C5a and pfGPI co-treated groups differed from a hypothetical value of “0”. Correlation of peripheral parasitemia to placental parasitemia was determined using Spearman correlation. Experiments were performed in duplicate or triplicate and repeated as indicated in the figure legends. Statistical analysis of experimental replicates is found in the supporting information files (Supporting Information S1, S2, S3, S4). Differences with p<0.05 were considered significant. Statistical analysis was performed using GraphPad Prism and SPSS.
Study participants
Plasma samples were collected at time of delivery from the peripheral blood and placental blood of primigravid women, living in an area of holoendemic malaria transmission in Western Kenya, with and without placental malaria infection as described previously [7]. Malaria infection was evaluated by examination of Giemsa-stained peripheral and placental thick and thin blood smears by light microscopy. Percent parasitemia was estimated by counting a minimum of 1000 erythrocytes and calculated using the following formula: (# infected erythrocytes/total # erythrocytes) ×100. Women who tested positive for HIV were excluded. This study was approved by the institutional review boards of the University of Georgia, the Kenya Medical Research Institute and the Centers for Disease Control and Prevention, Atlanta, Georgia, United States. Written informed consent was obtained from all participants.
Results
CSA- and CD36-binding PEs activate C5 and pfGPI induces C5a receptor expression
The pro-inflammatory anaphylatoxin C5a has been implicated in mediating deleterious host responses to bacterial and fungal agents [22], [25]. To confirm that CSA-binding PEs were capable of activating the complement system and thus generating C5a, serum from malaria-naïve donors was added to mature stage CS2 PEs or uRBCS. Compared to uRBCs or controls, C5 activation was only observed in the presence of PEs (Figure 1A) (mean C5a levels: CS2 PEs 24.77 ng/mL [95% confidence interval (CI) 2.229–47.30], E8B PEs 26.19 ng/ml [95% CI −10.44–62.82], uRBCs 7.114 ng/mL [95% CI 5.922–8.306], control 7.242 ng/mL [95% CI 4.116–10.37]; CS2 PEs vs. uRBCs p<0.001 Student's t-test). Non-PM PEs (E8B) also activated C5 (E8B PEs vs. uRBCs p<0.001 Student's t-test).
(A) Serum from malaria-naïve donors was added to mature stage CS2 (CSA binding) PEs or E8B (ICAM-1 and CD36 binding), uninfected red blood cells (uRBCs) or media for 30 min and supernatants were assayed for C5a by ELISA. Increased C5a levels were observed in the supernatants of PEs compared to uRBC or control, (Student's t-test: uRBCs vs. Mature PEs ***p<0.001). Data are presented as means±SEM and are representative of four independent experiments for CS2 parasites and two independent experiments for CD36-binding parasites. (B) Human PBMCs were stimulated with pfGPI or unconjugated gold beads alone as a control for 24 h. Expression of C5a receptor was determined by flow cytometry on monocytes (CD14). The percentage of monocytes expressing C5aR was significantly higher for pfGPI-treated versus media control cells, (Student's t-test: *p = 0.0279). Data are presented as means±SEM and are representative of two independent experiments.
In order to respond to activated C5, cells must express a C5a receptor. Previous studies report that C5a responsiveness increases following cellular stimulation [43]. To determine the effect of pfGPI on monocyte C5aR (CD88) expression, PBMCs were incubated as above with pfGPI or media control for 24 or 48 hours, and C5aR and CD14 expression (used to define monocytes) was determined by flow cytometry. Compared to media only control cells, a significant increase in the number of monocytes expressing C5aR (Figure 1B) was observed (mean: media 1.2% [95% CI −0.2–2.7] vs. pfGPI 3.0 [95% CI 1.5–4.4] p = 0.0279 by Student's t-test). The density of the C5aR, as determined by the mean fluorescent intensity on monocytes already expressing the receptor, was unchanged (data not shown).
C5a potentiates pfGPI-induced inflammatory responses
C5a has been shown to enhance TLR-mediated pro-inflammatory responses to microbial products such as LPS, contributing to sepsis syndromes [22], [25], [44], [45]. The parasite product pfGPI is recognized by TLR2 and is believed to be a mediator of the inflammatory response that characterizes severe malarial syndromes [10], [44]–[46]. To assess whether C5a would amplify pfGPI-induced pro-inflammatory responses, we treated PBMCs with increasing concentrations of recombinant human C5a (0, 1, 5, 10, 50, and 100 nM) with or without pfGPI for 24 hours and determined the production of IL-6 and TNF by ELISA. We observed a dose-dependent potentiation of pfGPI-induced IL-6 and TNF production by C5a (Figure 2A and B; two-way ANOVA (C5a*GPI), p<0.0001).
Human PBMCs were stimulated with varying concentrations of recombinant human C5a (0, 1, 5, 10, 50, 100 nM) for 24 hours, with or without pfGPI. Culture supernatants were collected and assayed for IL-6 (A) and TNF (B) by ELISA. Data were analyzed by two-way ANOVA and demonstrate an interaction effect (C5a*pfGPI) for IL-6 (***p<0.0001) and for TNF (***p<0.0001). Data are presented as means±SEM of three independent experiments.
C5a Receptor blockade inhibits enhanced cytokine responses
Elevated levels of TH1 cytokines have been implicated in PM pathogenesis and adverse pregnancy outcomes [14], [16], [47]. In order to determine the effect of C5a on the kinetics of pfGPI-induced cytokine production, we treated PBMCs with C5a (50 nM) and pfGPI (HPLC-purified, 300 ng/mL) for 48 hours. Supernatants were collected at baseline, 1, 2, 4, 6, 8, 12, 24 and 48 hours and assayed for inflammatory cytokines using a cytometric bead array. To examine the role of C5a-C5aR interactions in the observed inflammatory responses, we also performed parallel experiments including monoclonal antibody blockade of the C5a receptor. pfGPI-stimulated cytokines, including IL-6, TNF, IL-1β and IL-10, were synergistically induced in the presence of C5a (Figure 3A–D: two-way ANOVA on cumulative cytokine levels (C5a*GPI); p<0.0001 for IL-6, p = 0.0124 for TNF, p = 0.0418 for IL-1β and p<0.0001 for IL-10). A significant reduction in cytokine production was observed in experiments using anti-C5aR blocking antibody compared to its isotype control for all cytokines measured (Figure 3E–H: repeated measures ANOVA). Further analysis (repeated measures ANOVA) was performed to determine whether C5aR blockade returned cytokine production to levels for cells treated with pfGPI alone. In the presence of anti-C5aR, production of IL-6, TNF, and IL-1β was returned to levels that were not significantly different from pfGPI alone (p>0.05). However, this was not true for IL-10 (p = 0.0101).
Human PBMCs were cultured with unconjugated gold beads (media control), unconjugated gold beads and C5a (50 nM), pfGPI (300 ng/mL), or a combination of C5a and pfGPI. (A–D) Supernatants were assayed for IL-6 (A), TNF (B), IL-1β (C), and IL-10 (D). A significant interaction effect between C5a and pfGPI was observed for all cytokines tested. p<0.0001 for IL-6, p = 0.0124 for TNF, p = 0.0418 for IL-1β and p<0.0001 for IL-10 (by two-way ANOVA on cumulative cytokine production over 48 h). Data are representative of three independent experiments. (E–H) Human PBMCs were treated with an α-CD88 mAb (5 µg/mL) to block the C5a receptor or an appropriate isotype control and exposed to either unconjugated gold beads (media control) or a combination of C5a (50 nM) and pfGPI (300 ng/mL) for 48 h. Supernatants were assayed for IL-6 (E), TNF (F), IL-1β (G) and IL-10 (H). Cytokine production was significantly reduced upon C5a receptor blockade; p = 0.0005 for IL-6, p = 0.0028 for TNF, p = 0.0479 for IL-1β and p = 0.0191 for IL-10 (repeated measures ANOVA comparing cytokine levels between isotype control mAb and α-CD88 mAb groups co-treated with C5a and pfGPI). Values shown are corrected for background levels observed in the media controls, and are presented as means±SEM. Data are representative of two independent experiments.
C5a potentiates secretion of chemokines and the angiogenesis inhibitor sFlt-1
The infiltration of monocytes into the placenta during PM is associated with adverse birth outcomes [15], [48], [49]. Furthermore, elevated levels of the anti-angiogenic factor sFlt-1 (sVEGFR-1) have been causally implicated in placental dysfunction, growth restriction and fetal loss [32], [37]. In order to determine whether C5a could potentiate the induction of chemokines involved in the recruitment of monocytes and the anti-angiogenic factor sFlt-1, we incubated PBMCs with C5a, pfGPI, or media controls with or without C5aR blocking antibody or an isotype control antibody and measured IL-8 (CXCL8), MCP-1 (CCL2), MIP-1α (CCL3), MIP-1β (CCL4), and sFlt-1 (sVEGFR-1) by ELISA (Figure 4). In the C5a and pfGPI co-treated group, synergistic induction of all factors was observed by two-way ANOVA (C5a*GPI; p = 0.0265 for IL-8, p = 0.0010 for MCP-1, and p<0.0001 for MIP-1 α, MIP-1β and sFlt-1). Blocking C5aR inhibited the augmented secretion of all chemokines with the exception of IL-8 (Figure 4A–E: Student's t-test p = 0.0388for MCP-1, p = 0.0005 for MIP1-α, p = 0.0002 for MIP1-β). Blockade of the C5aR reduced levels of sFlt-1, but the differences did not consistently reach statistical significance (see supporting information file S4).
Human PBMCs were treated with either unconjugated gold beads (media control), unconjugated gold beads and C5a (50 nM), pfGPI (300 ng/mL), or a combination of C5a and pfGPI for 48 h. Where indicated PBMCs were treated with either 5 µg/mL of α-CD88 mAb to block the C5a receptor or an appropriate isotype control and were exposed to C5a and pfGPI. Supernatants were assayed for IL-8 (A), MCP-1 (B), MIP-1α (C), MIP-1β (D) and sFlt-1 (E). A significant interaction effect between C5a and pfGPI was observed; *p = 0.0265 for IL-8, **p = 0.0010 for MCP-1, * p = 0.0177 for sFlt-1, and ***p<0.0001 for MIP-1 α, and MIP-1β (by two-way ANOVA). C5a receptor blockade significantly reduced the levels of MCP-1 (* p = 0.0388), MIP-1α (*** p = 0.0005), and MIP-1β (*** p = 0.0002) (by student's t-test). Data are presented as means±SEM and are representative of three independent experiments.
C5a is increased in PM and is associated with higher placental parasite burdens
In order to extend our observations to the in vivo setting, we examined C5a levels in malaria-exposed women (Table 1 and 2). We analyzed C5a levels in the peripheral and placental plasma of HIV negative primigravid women with or without PM from Kisumu, Kenya. Peripheral C5a levels were significantly increased in women with PM (Figure 5A: mean C5a levels; PM negative: 98.85 ng/mL [95% CI 84.14–113.6] vs. PM positive: 147.1 ng/mL [95% CI 96.81–197.3] p = 0.01 by Student's t-test two-tailed test on log transformed data). Furthermore, C5a was significantly elevated in the placental plasma of women with PM (Figure 5B: mean C5a levels; PM negative: 69.07 ng/mL [95% CI 58.17–79.98] vs. PM positive: 104.0 ng/mL [95% CI 73.10–135.0] p = 0.0264 by Student's t-test two-tailed test on log transformed data). No significant differences were observed between the ages of PM positive vs. PM negative women; however, primigravid women with PM had significantly lower hemoglobin levels, consistent with the destruction of maternal erythrocytes associated with malarial infection (Table 1 and 2).
C5a levels were assayed in peripheral and placental plasma by ELISA. Samples were obtained from HIV negative primigravid women at time of delivery in Kisumu, western Kenya. C5a levels were significantly higher in (A) peripheral plasma of placenta malaria positive (PM+, n = 21) vs. placenta malaria negative (PM−, n = 45) women, *p = 0.01 by two-tailed Student's t-test on log transformed data; and (B) in placenta plasma of placenta malaria positive (PM+, n = 24) vs. placenta malaria negative (PM−, n = 47) women, *p = 0.0264 by Student's t-test two-tailed test on log transformed data.
Discussion
This study provides the first evidence implicating excessive activation of the complement component C5 and the induction of anti-angiogenic factors in the pathogenesis of PM. We demonstrate that PEs and parasite products (pfGPI) activated C5 and induced C5a receptor expression (Figure 1). Further, C5a and pfGPI induced synergistic production of inflammatory cytokines, chemokines and an anti-angiogenic factor associated with adverse pregnancy outcomes (Figures 2, 3&4). A role for C5a-C5aR was confirmed by inhibiting inflammation with C5aR blockade (Figures 3&4). These findings were extended to the in vivo setting, where primigravid women with PM were shown to have elevated C5a levels in both peripheral and placental plasma (Figure 5).
The generation of C5a is normally tightly controlled and, under physiological conditions, C5a enhances the effector function of macrophages and neutrophils, contributing to effective innate responses to microbial pathogens [22]. However, detectable systemic C5a suggests a loss of regulation of complement activation. Elevated C5a levels are associated with a number of deleterious impacts on host innate defence, including defects in phagocyte and endothelial cell function, potentiated chemokine and cytokine secretion to bacterial components such as LPS, and lymphoid apoptosis (reviewed in [22], [25]).
There are several mechanisms by which malaria may induce excessive complement activation, thereby contributing to the pathophysiologic mechanisms underlying PM. We observed that mature stage PEs are capable of generating C5a and a bioactive product of parasite rupture, pfGPI can result in an increase in monocytes expressing the C5aR. These observations suggest the potential for an autocrine loop, resulting in augmented inflammatory and angiogenic responses that may act in concert to mediate or amplify placental and fetal injury.
Our observations of malaria-induced C5a are consistent with studies of severe malaria and experimental malaria challenge models that report activation of the complement system [50]–[52]. In experimental human malaria infection, complement activation was observed prior to microscopic detection of parasitemia [50]. Similarly, a recent study of P. berghei ANKA implicated C5a as an early mediator of experimental cerebral malaria [53]. In the latter study, C5a levels were elevated prior to the onset of symptoms suggesting a role for C5a as an initiating or amplifying factor in malaria pathogenesis. Our observations are further supported by the findings of a recent genome-wide analysis of PM, where selected complement genes as well as C5aR were shown to be upregulated [54].
During sepsis the excess activation of C5 has been associated with enhanced secretion of pro-inflammatory mediators in response to TLR ligands such as LPS [55]. Similarly pfGPI has been shown to induce inflammatory mediators in a TLR2-dependant manner [10], [56]. Additional studies will be needed to determine if the synergistic induction of inflammatory and angiogenic factors observed with pfGPI-C5a are also TLR-dependent.
In our in vitro model, C5aR blockade abrogated the potentiated inflammatory responses that are believed to be central to the pathogenesis of PM. However, a different response was observed for IL-10. While IL-10 production was reduced by C5aR blockade, the levels did not return to those observed with pfGPI alone. We speculate that this may be attributable to the expression of the alternate C5a receptor, C5L2, which has been shown to promote TH2 skewing [57] as well as have inflammatory effects in experimental sepsis [58]. Future studies will be required to investigate the role of pfGPI and other parasite products on the expression of C5L2. When cells were treated with pfGPI and C5a, IL-10 was induced earlier than with pfGPI alone (Figure 3D). This may have important implications in vivo since TGF-β and IL-10 regulate inflammatory responses. Early IL-10 may impair TH1-mediated parasite clearance during malaria infection and increased IL-10 levels have been reported in PM [59].
Monocyte infiltration into the placental intervillous space has been associated with poor birth outcomes including the risk of low birth weight infants [15], [48], [49], [60], [61]. Chemokines function as specific chemoattractants for leukocytes. IL-8 plays a role in neutrophil and monocyte migration, while the β-chemokines are involved in macrophage migration. High levels of macrophage-derived IL-8 during PM have been associated with IUGR [47]. Elevated levels of MIP-1α, MIP-1β and MCP-1 have been observed in PM and it is likely that these chemokines are derived from both maternal (leukocyte) [6], [8] and fetal (syncytiotrophoblast) sources [11], [62]. As with cytokines, C5a and pfGPI synergistically induced chemokines, and this effect was abrogated for the β-chemokines by blocking the C5aR (Figure 4A–D). No significant difference for IL-8 production was found between cells treated with anti-C5aR versus an isotype control due to the variation in the latter group. However, there was no difference between cells treated with pfGPI alone or with anti-C5aR plus C5a and pfGPI, indicating that blocking the C5aR inhibits the effect attributable to C5a.
Peak prevalence of malaria infection during pregnancy occurs between 13–20 weeks gestation and gradually falls as gestation increases [63]. This coincides with the second wave of trophoblast invasion of the maternal spiral arteries (16–18 weeks). It is hypothesized that PM may impair transformation of the maternal vasculature leading to placental dysfunction [64], [65]. One protein that may play a role in impaired angiogenesis is the anti-angiogenic factor sFlt-1. This alternatively spliced soluble variant of VEGFR-1 binds and sequesters placental growth factor and vascular endothelial growth factor. Excess sFlt-1 has been shown to inhibit placental cytotrophoblast differentiation and invasion [33]. Increased sFlt-1 levels have been implicated in the pathogenesis of placentation associated with pre-eclampsia and IUGR [34]–[36] and have been shown to predict the development of pre-eclampsia [35]. sFlt-1 can be produced extra-placentally by monocytes treated with C5a [32]. Therefore, we determined whether C5a and pfGPI would potentiate sFlt-1 production from mononuclear cells. We show that sFlt-1 was synergistically induced by co-stimulation with these molecules and that this effect was reduced but not consistently inhibited by C5aR blockade (Figure 4E). We postulate this may be due to different PBMC donors, as FLT1 genotype has been associated with differing sFlt-1 levels in the peripheral blood of primigravid women, as well as PBMC responsiveness to LPS in vitro [66]. Furthermore, the study by Muehlenbachs et al., provides the first evidence that both fetal and maternal genotype may contribute to immune responses to PM [66]. Taken together, these data suggest a putative mechanism for impaired angiogenesis and placental dysfunction during PM.
In order to extend these observations to human infection, we measured C5a levels in the peripheral blood of a cohort of malaria-exposed primigravid Kenyan women and we measured C5a levels in the placental blood of a similar cohort of malaria-exposed primigravid women at time of delivery (Table 1–2) Significantly increased levels of C5a were found in peripheral blood (Figure 5A) and the placental blood (Figure 5B) of pregnant women with PM compared to those without. These results provide the first evidence that C5a levels are elevated in women with PM. However, these findings will need to be confirmed in larger prospective clinical studies. It will be of interest to also examine how C5a levels change with respect to other disease parameters, including parasitemia and mononuclear cell infiltrates in larger populations.
In summary, we have demonstrated a role for C5a in the enhanced cytokine and chemokine responses that characterize PM and in the induction of sFlt-1, an anti-angiogenic factor associated with abnormal placental development and poor birth outcomes (Figure 6). Our findings that disruption of the C5a-C5aR interaction inhibits inflammatory and anti-angiogenic responses to malaria suggest anti-C5a-C5aR strategies as potential therapeutic approaches for PM [67].
Mature schizonts activate C5 and rupture releasing parasite components containing GPI that induce expression of C5aR and activate macrophages. The concomitant expression of C5aR and generation of C5a represent a potential autocrine mechanism by which C5a can amplify inflammation in P. falciparum malaria resulting in the excessive cytokine and chemokine responses that characterize PM. The dysregulated production of cytokines, chemokines and the anti-angiogenic factor sFlt-1 contribute to placental dysfunction, intrauterine growth restriction and low birth weight infants.
Supporting Information
Supporting Information S1.
Experimental replicates and statistics Figure 1A and B
https://doi.org/10.1371/journal.pone.0004953.s001
(0.03 MB DOC)
Supporting Information S2.
Experimental replicates and statistics Figure 2B (TNF)
https://doi.org/10.1371/journal.pone.0004953.s002
(0.03 MB DOC)
Supporting Information S3.
Experimental replicates and statistics Figure 3
https://doi.org/10.1371/journal.pone.0004953.s003
(0.04 MB DOC)
Supporting Information S4.
Experimental replicates and statistics Figure 4
https://doi.org/10.1371/journal.pone.0004953.s004
(0.05 MB DOC)
Acknowledgments
We thank the Director of the Kenya Medical Research Institute for his permission to publish this work, and all the women who participated in this study.
Author Contributions
Conceived and designed the experiments: ALC LS WCL KCK. Performed the experiments: ALC LS CF. Analyzed the data: ALC LS CF KCK. Contributed reagents/materials/analysis tools: SOO SK CDG WCL JMM KCK. Wrote the paper: ALC KCK. Organized collection of clinical samples and clinical data: SOO JMM.
References
- 1. Gilles HM, Lawson JB, Sibelas M, Voller A, Allan N (1969) Malaria, anaemia and pregnancy. Ann Trop Med Parasitol 63: 245–263.
- 2. Diagne N, Rogier C, Cisse B, Trape JF (1997) Incidence of clinical malaria in pregnant women exposed to intense perennial transmission. Trans R Soc Trop Med Hyg 91: 166–170.
- 3. Fried M, Duffy PE (1996) Adherence of Plasmodium falciparum to chondroitin sulfate A in the human placenta. Science 272: 1502–1504.
- 4. Buffet PA, Gamain B, Scheidig C, Baruch D, Smith JD, et al. (1999) Plasmodium falciparum domain mediating adhesion to chondroitin sulfate A: a receptor for human placental infection. Proc Natl Acad Sci U S A 96: 12743–12748.
- 5. Duffy MF, Byrne TJ, Elliott SR, Wilson DW, Rogerson SJ, et al. (2005) Broad analysis reveals a consistent pattern of var gene transcription in Plasmodium falciparum repeatedly selected for a defined adhesion phenotype. Mol Microbiol 56: 774–788.
- 6. Abrams ET, Brown H, Chensue SW, Turner GD, Tadesse E, et al. (2003) Host response to malaria during pregnancy: placental monocyte recruitment is associated with elevated beta chemokine expression. J Immunol 170: 2759–2764.
- 7. Chaisavaneeyakorn S, Moore JM, Mirel L, Othoro C, Otieno J, et al. (2003) Levels of macrophage inflammatory protein 1 alpha (MIP-1 alpha) and MIP-1 beta in intervillous blood plasma samples from women with placental malaria and human immunodeficiency virus infection. Clin Diagn Lab Immunol 10: 631–636.
- 8. Suguitan AL Jr, Leke RG, Fouda G, Zhou A, Thuita L, et al. (2003) Changes in the levels of chemokines and cytokines in the placentas of women with Plasmodium falciparum malaria. J Infect Dis 188: 1074–1082.
- 9. Nebl T, De Veer MJ, Schofield L (2005) Stimulation of innate immune responses by malarial glycosylphosphatidylinositol via pattern recognition receptors. Parasitology 130: SupplS45–62.
- 10. Krishnegowda G, Hajjar AM, Zhu J, Douglass EJ, Uematsu S, et al. (2005) Induction of proinflammatory responses in macrophages by the glycosylphosphatidylinositols of Plasmodium falciparum: cell signaling receptors, glycosylphosphatidylinositol (GPI) structural requirement, and regulation of GPI activity. J Biol Chem 280: 8606–8616.
- 11. Lucchi NW, Koopman R, Peterson DS, Moore JM (2006) Plasmodium falciparum-infected red blood cells selected for binding to cultured syncytiotrophoblast bind to chondroitin sulfate A and induce tyrosine phosphorylation in the syncytiotrophoblast. Placenta 27: 384–394.
- 12. Lucchi NW, Peterson DS, Moore JM (2008) Immunologic activation of human syncytiotrophoblast by Plasmodium falciparum. Malar J 7: 42.
- 13. Chaisavaneeyakorn S, Lucchi N, Abramowsky C, Othoro C, Chaiyaroj SC, et al. (2005) Immunohistological characterization of macrophage migration inhibitory factor expression in Plasmodium falciparum-infected placentas. Infect Immun 73: 3287–3293.
- 14. Fried M, Muga RO, Misore AO, Duffy PE (1998) Malaria elicits type 1 cytokines in the human placenta: IFN-gamma and TNF-alpha associated with pregnancy outcomes. J Immunol 160: 2523–2530.
- 15. Rogerson SJ, Pollina E, Getachew A, Tadesse E, Lema VM, et al. (2003) Placental monocyte infiltrates in response to Plasmodium falciparum malaria infection and their association with adverse pregnancy outcomes. Am J Trop Med Hyg 68: 115–119.
- 16. Rogerson SJ, Brown HC, Pollina E, Abrams ET, Tadesse E, et al. (2003) Placental tumor necrosis factor alpha but not gamma interferon is associated with placental malaria and low birth weight in Malawian women. Infect Immun 71: 267–270.
- 17. Caucheteux SM, Kanellopoulos-Langevin C, Ojcius DM (2003) At the innate frontiers between mother and fetus: linking abortion with complement activation. Immunity 18: 169–172.
- 18. Bowen JM, Chamley L, Mitchell MD, Keelan JA (2002) Cytokines of the placenta and extra-placental membranes: biosynthesis, secretion and roles in establishment of pregnancy in women. Placenta 23: 239–256.
- 19. Marzi M, Vigano A, Trabattoni D, Villa ML, Salvaggio A, et al. (1996) Characterization of type 1 and type 2 cytokine production profile in physiologic and pathologic human pregnancy. Clin Exp Immunol 106: 127–133.
- 20. Hunt JS (1989) Cytokine networks in the uteroplacental unit: macrophages as pivotal regulatory cells. J Reprod Immunol 16: 1–17.
- 21. Huber-Lang M, Sarma JV, Zetoune FS, Rittirsch D, Neff TA, et al. (2006) Generation of C5a in the absence of C3: a new complement activation pathway. Nat Med 12: 682–687.
- 22. Guo RF, Ward PA (2005) Role of C5a in inflammatory responses. Annu Rev Immunol 23: 821–852.
- 23. Huber-Lang M, Younkin EM, Sarma JV, Riedemann N, McGuire SR, et al. (2002) Generation of C5a by phagocytic cells. Am J Pathol 161: 1849–1859.
- 24. Riedemann NC, Guo RF, Ward PA (2003) The enigma of sepsis. J Clin Invest 112: 460–467.
- 25. Ward PA (2004) The dark side of C5a in sepsis. Nat Rev Immunol 4: 133–142.
- 26. Chenoweth DE, Hugli TE (1978) Demonstration of specific C5a receptor on intact human polymorphonuclear leukocytes. Proc Natl Acad Sci U S A 75: 3943–3947.
- 27. Aird WC (2003) The role of the endothelium in severe sepsis and multiple organ dysfunction syndrome. Blood 101: 3765–3777.
- 28. Pober JS, Sessa WC (2007) Evolving functions of endothelial cells in inflammation. Nature reviews Immunology 7: 803–815.
- 29. Smedegard G, Cui LX, Hugli TE (1989) Endotoxin-induced shock in the rat. A role for C5a. Am J Pathol 135: 489–497.
- 30. Czermak BJ, Sarma V, Pierson CL, Warner RL, Huber-Lang M, et al. (1999) Protective effects of C5a blockade in sepsis. Nat Med 5: 788–792.
- 31. Huber-Lang M, Sarma VJ, Lu KT, McGuire SR, Padgaonkar VA, et al. (2001) Role of C5a in multiorgan failure during sepsis. J Immunol 166: 1193–1199.
- 32. Girardi G, Yarilin D, Thurman JM, Holers VM, Salmon JE (2006) Complement activation induces dysregulation of angiogenic factors and causes fetal rejection and growth restriction. J Exp Med 203: 2165–2175.
- 33. Charnock-Jones DS, Burton GJ (2000) Placental vascular morphogenesis. Baillieres Best Pract Res Clin Obstet Gynaecol 14: 953–968.
- 34. Levine RJ, Maynard SE, Qian C, Lim KH, England LJ, et al. (2004) Circulating angiogenic factors and the risk of preeclampsia. N Engl J Med 350: 672–683.
- 35. Lam C, Lim KH, Karumanchi SA (2005) Circulating angiogenic factors in the pathogenesis and prediction of preeclampsia. Hypertension 46: 1077–1085.
- 36. Zhou Y, McMaster M, Woo K, Janatpour M, Perry J, et al. (2002) Vascular endothelial growth factor ligands and receptors that regulate human cytotrophoblast survival are dysregulated in severe preeclampsia and hemolysis, elevated liver enzymes, and low platelets syndrome. Am J Pathol 160: 1405–1423.
- 37. Muehlenbachs A, Mutabingwa TK, Edmonds S, Fried M, Duffy PE (2006) Hypertension and maternal-fetal conflict during placental malaria. PLoS Med 3: e446.
- 38. Patel SN, Serghides L, Smith TG, Febbraio M, Silverstein RL, et al. (2004) CD36 mediates the phagocytosis of Plasmodium falciparum-infected erythrocytes by rodent macrophages. J Infect Dis 189: 204–213.
- 39. Serghides L, Patel SN, Ayi K, Kain KC (2006) Placental chondroitin sulfate A-binding malarial isolates evade innate phagocytic clearance. J Infect Dis 194: 133–139.
- 40. McGilvray ID, Serghides L, Kapus A, Rotstein OD, Kain KC (2000) Nonopsonic monocyte/macrophage phagocytosis of Plasmodium falciparum-parasitized erythrocytes: a role for CD36 in malarial clearance. Blood 96: 3231–3240.
- 41. Naik RS, Branch OH, Woods AS, Vijaykumar M, Perkins DJ, et al. (2000) Glycosylphosphatidylinositol Anchors of Plasmodium falciparum: Molecular Characterization and Naturally Elicited Antibody Response That May Provide Immunity to Malaria Pathogenesis. J Exp Med 192: 1563–1576.
- 42. Solomkin JS, Jenkins MK, Nelson RD, Chenoweth D, Simmons RL (1981) Neutrophil dysfunction in sepsis. II. Evidence for the role of complement activation products in cellular deactivation. Surgery 90: 319–327.
- 43. Monk PN, Scola AM, Madala P, Fairlie DP (2007) Function, structure and therapeutic potential of complement C5a receptors. Br J Pharmacol 152: 429–448.
- 44. Albrecht EA, Ward PA (2004) Complement-induced impairment of the innate immune system during sepsis. Curr Allergy Asthma Rep 4: 359–364.
- 45. Zhu J, Krishnegowda G, Gowda DC (2005) Induction of proinflammatory responses in macrophages by the glycosylphosphatidylinositols of Plasmodium falciparum: the requirement of extracellular signal-regulated kinase, p38, c-Jun N-terminal kinase and NF-kappaB pathways for the expression of proinflammatory cytokines and nitric oxide. J Biol Chem 280: 8617–8627.
- 46. Niederbichler AD, Hoesel LM, Westfall MV, Gao H, Ipaktchi KR, et al. (2006) An essential role for complement C5a in the pathogenesis of septic cardiac dysfunction. J Exp Med 203: 53–61.
- 47. Moormann AM, Sullivan AD, Rochford RA, Chensue SW, Bock PJ, et al. (1999) Malaria and pregnancy: placental cytokine expression and its relationship to intrauterine growth retardation. J Infect Dis 180: 1987–1993.
- 48. Menendez C, Ordi J, Ismail MR, Ventura PJ, Aponte JJ, et al. (2000) The impact of placental malaria on gestational age and birth weight. J Infect Dis 181: 1740–1745.
- 49. Moore JM, Nahlen BL, Misore A, Lal AA, Udhayakumar V (1999) Immunity to placental malaria. I. Elevated production of interferon-gamma by placental blood mononuclear cells is associated with protection in an area with high transmission of malaria. J Infect Dis 179: 1218–1225.
- 50. Roestenberg M, McCall M, Mollnes TE, van Deuren M, Sprong T, et al. (2007) Complement activation in experimental human malaria infection. Trans R Soc Trop Med Hyg 101: 643–649.
- 51. Wenisch C, Spitzauer S, Florris-Linau K, Rumpold H, Vannaphan S, et al. (1997) Complement activation in severe Plasmodium falciparum malaria. Clin Immunol Immunopathol 85: 166–171.
- 52. Yamada M, Steketee R, Abramowsky C, Kida M, Wirima J, et al. (1989) Plasmodium falciparum associated placental pathology: a light and electron microscopic and immunohistologic study. Am J Trop Med Hyg 41: 161–168.
- 53. Patel SN, Berghout J, Lovegrove FE, Ayi K, Conroy A, et al. (2008) C5 deficiency and C5a or C5aR blockade protects against cerebral malaria. J Exp Med 205: 1133–1143.
- 54. Muehlenbachs A, Fried M, Lachowitzer J, Mutabingwa TK, Duffy PE (2007) Genome-wide expression analysis of placental malaria reveals features of lymphoid neogenesis during chronic infection. J Immunol 179: 557–565.
- 55. Riedemann NC, Guo RF, Hollmann TJ, Gao H, Neff TA, et al. (2004) Regulatory role of C5a in LPS-induced IL-6 production by neutrophils during sepsis. Faseb J 18: 370–372.
- 56. Patel SN, Lu Z, Ayi K, Serghides L, Gowda DC, et al. (2007) Disruption of CD36 Impairs Cytokine Response to Plasmodium falciparum Glycosylphosphatidylinositol and Confers Susceptibility to Severe and Fatal Malaria In Vivo. J Immunol 178: 3954–3961.
- 57. Hawlisch H, Wills-Karp M, Karp CL, Kohl J (2004) The anaphylatoxins bridge innate and adaptive immune responses in allergic asthma. Mol Immunol 41: 123–131.
- 58. Rittirsch D, Flierl MA, Nadeau BA, Day DE, Huber-Lang M, et al. (2008) Functional roles for C5a receptors in sepsis. Nat Med 14: 551–557.
- 59. Kabyemela ER, Muehlenbachs A, Fried M, Kurtis JD, Mutabingwa TK, et al. (2008) Maternal peripheral blood level of IL-10 as a marker for inflammatory placental malaria. Malar J 7: 26.
- 60. Leopardi O, Naughten W, Salvia L, Colecchia M, Matteelli A, et al. (1996) Malaric placentas. A quantitative study and clinico-pathological correlations. Pathol Res Pract 192: 892–898. discussion 899–900.
- 61. Ordi J, Ismail MR, Ventura PJ, Kahigwa E, Hirt R, et al. (1998) Massive chronic intervillositis of the placenta associated with malaria infection. Am J Surg Pathol 22: 1006–1011.
- 62. Lucchi NW, Moore JM (2007) LPS induces secretion of chemokines by human syncytiotrophoblast cells in a MAPK-dependent manner. J Reprod Immunol 73: 20–27.
- 63. Brabin BJ (1983) An analysis of malaria in pregnancy in Africa. Bull World Health Organ 61: 1005–1016.
- 64. Jansson T, Powell TL (2007) Role of the placenta in fetal programming: underlying mechanisms and potential interventional approaches. Clin Sci (Lond) 113: 1–13.
- 65. Dorman EK, Shulman CE, Kingdom J, Bulmer JN, Mwendwa J, et al. (2002) Impaired uteroplacental blood flow in pregnancies complicated by falciparum malaria. Ultrasound Obstet Gynecol 19: 165–170.
- 66. Muehlenbachs A, Fried M, Lachowitzer J, Mutabingwa TK, Duffy PE (2008) Natural selection of FLT1 alleles and their association with malaria resistance in utero. Proc Natl Acad Sci U S A 105: 14488–14491.
- 67. Hillmen P, Young NS, Schubert J, Brodsky RA, Socie G, et al. (2006) The complement inhibitor eculizumab in paroxysmal nocturnal hemoglobinuria. N Engl J Med 355: 1233–1243.