Severe malarial anemia: innate immunity and pathogenesis.

Greater than 80% of malaria-related mortality occurs in sub-Saharan Africa due to infections with Plasmodium falciparum. The majority of P. falciparum-related mortality occurs in immune-naïve infants and young children, accounting for 18% of all deaths before five years of age. Clinical manifestations of severe falciparum malaria vary according to transmission intensity and typically present as one or more life-threatening complications, including: hyperparasitemia; hypoglycemia; cerebral malaria; severe malarial anemia (SMA); and respiratory distress. In holoendemic transmission areas, SMA is the primary clinical manifestation of severe childhood malaria, with cerebral malaria occurring only in rare cases. Mortality rates from SMA can exceed 30% in pediatric populations residing in holoendemic transmission areas. Since the vast majority of the morbidity and mortality occurs in immune-naïve African children less than five years of age, with SMA as the primary manifestation of severe disease, this review will focus primarily on the innate immune mechanisms that govern malaria pathogenesis in this group of individuals. The pathophysiological processes that contribute to SMA involve direct and indirect destruction of parasitized and non-parasitized red blood cells (RBCs), inefficient and/or suppression of erythropoiesis, and dyserythropoiesis. While all of these causal etiologies may contribute to reduced hemoglobin (Hb) concentrations in malaria-infected individuals, data from our laboratory and others suggest that SMA in immune-naïve children is characterized by a reduced erythropoietic response. One important cause of impaired erythroid responses in children with SMA is dysregulation in the innate immune response. Phagocytosis of malarial pigment hemozoin (Hz) by monocytes, macrophages, and neutrophils is a central factor for promoting dysregulation in innate inflammatory mediators. As such, the role of P. falciparum-derived Hz (PfHz) in mediating suppression of erythropoiesis through its ability to cause dysregulation in pro- and anti-inflammatory cytokines, growth factors, chemokines, and effector molecules is discussed in detail. An improved understanding of the etiological basis of suppression of erythropoietic responses in children with SMA may offer the much needed therapeutic alternatives for control of this global disease burden.

1. Human Malaria

Human malaria is caused by unicellular obligate intracellular protozoan parasites of the genus Plasmodium. Although malaria was once prevalent throughout most of the world, malaria is currently endemic in the tropical zones with extensions into the sub-tropical regions of Asia, Africa, South and Central America. However, about half of the world's population (3.3 billion people) is at risk of malaria in more than 100 countries 1. Four primary species of malaria parasites infect humans: P. falciparum, P. ovale, P. malariae and P. vivax. In addition, studies in Southeast Asia have shown that P. knowlesi, a malaria parasite that typically involves monkeys as the natural reservoir, can also infect humans, and in some cases, result in fatal disease (reviewed in 2). Malaria due to P. vivax, P. ovale and P. malariae is less severe than that experienced by P. falciparum infections and collectively, these three species account for slightly less than 10% of the worldwide malaria cases 3. The most virulent of the human malaria parasites is P. falciparum which is responsible for the bulk of the malaria-related morbidity and mortality. P. falciparum accounts for 91% of malaria cases worldwide of which the majority (i.e., 86%) occurs in the African region 3. Consistent with the high rate of morbidity within Africa, 90% of the P. falciparum-attributable malaria deaths also occur in the African region, primarily sub-Saharan Africa 3. Approximately 247 million malaria infections are estimated annually, resulting in greater than one million deaths, primarily in African children under the age of five 1.

2. Clinical Spectrum of Plasmodium falciparum Infections

The clinical spectrum of P. falciparum malaria in African children encompasses a wide range of pathophysiological derangements and includes multiple organ involvement and systemic disorders. Falciparum malaria ranges from asymptomatic infections to the classic symptoms of malaria (e.g., fever, chills, sweating, headache and muscle aches), which in a sub-population of the cases, results in severe life-threatening complications such as hyperparasitemia, hypoglycemia, hyperlactatemia, kidney failure, metabolic acidosis, cerebral malaria, severe malarial anemia [SMA, hemoglobin, (Hb)<5.0 g/dL], and respiratory distress (RD) 4, 5. Although the pathophysiology of malaria is multifactorial and only partially understood, development of a pathogenic versus protective outcome, once an infection occurs, is mediated by host and parasite interactions of which the following appear critically important: endemicity patterns, acquisition of naturally acquired malarial immunity, parasite virulence, multiplication rate, antigenic variation and polymorphic variability in both the host (human) and parasite 6. The age of the individual when they first acquire a malarial infection also plays a significant role in the clinical outcomes of the disease process. For example, children typically display enhanced susceptibility to severe anemia and hypoglycemia, while non-resident malaria-naïve adults are more likely to present with jaundice and progress to renal failure and respiratory distress due to pulmonary edema 7.

3. Infant and Childhood Anemia in Developing Nations

Although this review will focus primarily on SMA and the innate immune responses that condition the development and outcomes of the disease, it is important to understand the definitions of anemia and the importance of geographic context. A general definition of anemia is a reduction in Hb levels in relation to the age, gender, and physiological status of the individual within a defined geographical context 8. In western countries, anemia is defined by a Hb concentration <12.0 g/dL, while in developing countries the standard definition of anemia for children <5 years of age is Hb<11.0 g/dL 8. Throughout much of the developing world, particularly in regions of sub-Saharan Africa with high rates of malaria and human immunodeficiency virus (HIV), the majority of infants and young children suffer from anemia 9. Anemia in the developing world is largely a product of inadequate feeding practices, frequent infections, and micronutrient deficiencies which culminate in high rates of mortality in infants and young children 9.

4. Disease Burden of Severe Malarial Anemia

The World Health Organization (WHO) defines SMA as Hb concentrations <5.0 g/dL (or a hematocrit <15.0%) in the presence of any density parasitemia 5. SMA is a major public health problem in developing countries where it contributes 3-46% of inpatient pediatric fatalities in referral care facilities 10. Despite efforts aimed at ameliorating the anemia burden, SMA remains an important childhood health burden in sub-Saharan Africa 11. Previous studies in endemic areas of Africa demonstrated that the annual rate of presentation to hospital with SMA was 7.6/1000 in children aged 0-4 years with a case fatality of 9.7% 12. Additional studies illustrate that the risk for SMA peaks at 1 year of age in high (holoendemic) transmission regions and at approximately 2 years of age in areas with moderate and low transmission intensities, such that (in general) the overall risk of SMA decreases with increasing age 13. Multicentre studies indicate that SMA affects 7.5-34% of the African children that acquire malaria with an overall prevalence of 21.2% and case fatality rate of 8.4% 14.

5. Etiological Basis of Severe Malarial Anemia

The etiology of SMA can include a number of distinct, as well as overlapping features, including lysis of infected and uninfected RBCs 15-18, splenic sequestration of RBCs 19, dyserythropoiesis and bone marrow suppression 20, 21, co-infections with bacteremia, HIV-1, and hookworm 22-26, and chronic transmission of malaria in holoendemic regions. It is important to note that some or all of these factors can culminate in the chronically low Hb values observed in infants and young children residing in holoendemic regions. As such, the degree of parasitemia is typically a poor indicator of malaria disease severity in these locales, especially considering that peripheral parasitemia is a “snapshot” in time of the complex and continuously evolving disease process. However, it is important to stress that high levels of parasitemia, particularly in non-immune individuals, can certainly result in massive lysis and clearance of RBCs, resulting in profound anemia 21, 27.

6. Impaired Erythropoietic Responses in Severe Malarial Anemia

Although parasite-driven hemolysis will contribute to a reduction in Hb levels in childhood malaria, one of the primary mechanisms responsible for low Hb levels in children with SMA is impaired and/or ineffective erythropoiesis. Loss of appropriate production of erythrocytes translates into a failure in the ability to replenish the reduced pool of erythrocytes due to parasite- and/or anti-malarial-driven hemolysis. Earlier studies demonstrated that there is parenchymal damage of bone marrow, ineffective erythropoiesis, and a reduced rate of erythropoietic proliferation in patients with acute falciparum malaria 28. Subsequent studies in Gambian children demonstrated that SMA was defined by erythroid hyperplasia with dyserythropoiesis 6. Examination of bone marrow from children with SMA revealed hypercellularity, mild to normal erythroid hyperplasia, and abnormal features of late erythroid progenitors, but absence of damage to burst forming unit-erythrocyte (BFU-E) or colony forming unit-erythrocyte (CFU-E) early stages 6. In addition, we have shown that children with SMA have an inefficient reticulocyte production index (RPI) characterized by levels <2.0 29. The RPI is a measure of the extent to which the reticulocyte count has risen (or not) in response to the level of anemia 30. Taken together, these investigations demonstrate that suppression of erythropoiesis is a primary cause of severe anemia in children with P. falciparum infections.

7. Role of Inflammatory Mediators in Impaired Erythropoiesis

Although the precise mechanisms responsible for reduced reticulocyte responses in children with SMA have been somewhat elusive, it is clear that one important cause of reduced erythropoiesis in children with SMA is due to an imbalance in inflammatory mediators. In an attempt to control the parasitemia, the host releases an array of pro- and anti-inflammatory cytokines, chemokines, growth factors, and effector molecules as part of the innate immune response. Depending on the magnitude and timing of inflammatory mediator release, the immune response to malaria can result in either successful control of the parasitemia or alternatively, an inappropriate balance in the inflammatory milieu that can induce damage to the host, including suppression of the erythropoietic response. As such, although malaria is typically viewed as an 'acute' infection, in regions with high levels of falciparum endemicity, SMA is often a more 'chronic' condition in which the immunological response to infection drives an inflammatory milieu that can promote suppression of eythropoiesis. This premise is supported by observations showing that persistent childhood P. falciparum infections are associated with bone marrow suppression 31.

8. Role of Parasitic Products in Stimulating the Innate Immune Response

There are a number of key parasitic products that drive the innate immune response in the malaria-infected human host, including malarial pigment (hemozoin, Hz), glycosylphosphatidylinositols (GPIs), and parasitic antigens. While it is clear that GPIs, which form the connection between the parasite's cellular membrane and external antigens 32, as well as an array of parasitic antigens play an important role in activating the immune response, we will focus our discussion on the role of P. falciparum-derived Hz (PfHz) in promoting the host immune response and the potential implications that this process has on suppression of the erythropoietic response. Excellent reviews on the means by which parasite GPIs 33 and malarial antigens stimulate the innate immune response can be found elsewhere 34.

9. Hemozoin Formation

The importance of hemozoin in malarial infections was recognized over five decades ago following the discovery that accumulation of phagocytosed malarial pigment in bone marrow produces a brown/black appearance in patients with repeated malaria infections 35. Formation of PfHz occurs during the intraerythrocytic asexual replication cycle in which P. falciparum metabolizes host Hb as a source of amino acids 36, 37, leaving the iron-rich heme portion designated ferriprotoporphyrin IX (FP-IX). P. falciparum then promotes aggregation of the toxic FP-IX molecules, using heme polymerase 38, 39, into an insoluble product known as PfHz 40-42. Once the schizont has gone through replication within the host RBC, the erythrocyte ruptures 43, releasing PfHz along with merozoites, with the newly-formed merozoites infecting other RBCs, and PfHz being phagocytosed by monocytes/macrophages and neutrophils.

10. Hemozoin Accumulation as a Causal Factor in Suppression of Erythropoiesis

Accumulation of phagocytes containing malaria pigment in the microvasculature of deep tissues such as the bone marrows of patients with acute malaria was recognized in earlier studies 44. Intraleukocytic malaria pigment is a useful direct diagnostic marker with the amount of phagocytosed pigment being a good indirect measure of the sequestered parasite burden, recent schizogony, and disease severity 45. Studies by others and our group have also found an association between the presence of PfHz-containing monocytes and suppression of erythropoiesis in children with P. falciparum-induced anemia 46, 47. In addition, our studies in western Kenya, using multivariate regression analyses, controlling for age, gender, and parasitemia, revealed that elevated levels of pigment-containing monocytes (PCM) were a significant predictor of SMA 47. A recent multi-site study that included 26,296 hospitalized children with P. falciparum malaria demonstrated that the percentage of PfHz-containing monocytes was negatively correlated with hematocrit 48. Additional evidence that PfHz-containing cells are an important source for promoting erythroid suppression is supported by histological observations of bone marrow in children with SMA, and those dying from malaria, showing that developing erythroid progenitors located within the vicinity of pigmented macrophages have abnormal cellular development 46. Consistent with these observations, we recently demonstrated that elevated levels of PfHz in monocytes are associated with inefficient erythropoiesis in Kenyan children with malaria 49.

11. Severe Malarial Anemia is Characterized by Dysregulation in Innate Inflammatory Mediators

One of the primary factors for stimulating the innate immune response to P. falciparum is phagocytosis of PfHz by circulating monocytes and neutrophils, and resident macrophages. One of the primary means by which PfHz generates an innate immune response is through the toll-like receptors (TLRs). A recent review discussing the role of hemozoin in stimulating the TLRs can be found elsewhere 50. The studies discussed below support a model in which phagocytosis of PfHz, and the host immune response patterns that ensue following this event, are a critical step in the promotion of the reduced erythroid responses observed in children with SMA (Figure 1). Although it is clear that a wealth of important cellular biology is yet to be elucidated about the mechanisms by which phagocytosis of PfHz shapes the innate immune response, the ability of PfHz to alter inflammatory mediator profiles is central to its action. However, it is important to note that clarification of strict protective versus pathological roles for inflammatory mediators remains poorly defined and extremely difficult to quantify in human malaria in which manipulation of the biological systems is typically not feasible.

Figure 1

. Based on concomitant measurement of innate inflammatory mediators (using multiplex technologies) in children with varying severities of malarial anemia, we developed a model to describe how dysregulation in innate inflammatory mediators promotes suppression of erythropoiesis in children with SMA. Central to the model is the fact that phagocytosis of hemozoin (PfHz) by monocytes is one of the primary causes of altered production of innate inflammatory mediators. Elevated inflammatory mediators are shown in green text, while those that are decreased in children with SMA are shown in red text. Solid lines indicate positive signaling (up-regulation), whereas dashed lines indicate suppression (down-regulation). Children with SMA have decreased levels of IL-12 in response to ingestion of parasitized red blood cells (pRBC) and/or hemozoin by monocytes. Suppression of IL-12 in children with SMA is due to PfHz-induced IL-10 over-production. Children with SMA have increased circulating levels of TNF-α, IFN-γ, IL-6, MIP-1α, and MIP-1β. Although TNF-α can induce PGE2 and nitric oxide (NO), these effector molecules are suppressed in children with SMA. Suppression of PGE2 allows over-production of TNF-α, which is associated with enhanced severity of anemia. In addition, MIF is suppressed in children with falciparum malaria, an event associated with phagocytosis of PfHz by monocytes, and enhanced severity of anemia. Circulating levels of IFN-α, IL-1β, RANTES, and SCGF are also decreased in children with SMA. Reduced production of these innate inflammatory mediators, along with increased TNF-α, IL-6, MIP-1α and MIP-β, likely contribute to the development of SMA by suppressing the erythropoietic response. Lastly, although the reduced NO and reactive oxygen species (ROS) generation reported in children with falciparum malaria may promote ineffective parasite killing and, thereby, prolong parasitemia, children with malarial anemia have elevated levels of NO and ROS that can directly inhibit erythropoiesis.

Pro-inflammatory Mediators

A successful type 1 response to malaria requires a well-timed and proportional release of interleukin (IL)-12, interferon (IFN)-γ, and tumor necrosis factor (TNF)-α to minimize parasitemia and preserve erythropoiesis 51, 52. The pro-inflammatory phase should be followed by an equally timely abrogation of this response by type 2 cytokines such as IL-10, transforming growth factor (TGF)-β, and IL-4, to avoid inflammatory host damage 53.

TNF-α is the molecule typically associated with pathology in malaria and was first hypothesized to be involved in the host immune response to malaria in 1978 54. Several studies have shown an association between elevated TNF-α levels and morbidity and mortality in individuals with malaria 55, 56. However, TNF-α is critical for parasite killing and preventing parasite replication 55-59. In addition to its direct effects 60, 61, TNF-α also mediates its effects by inducing production of macrophage migration inhibitory factor (MIF) 62, 63 and nitric oxide synthase type 2 (NOS2, inducible nitric oxide synthase, iNOS) 64, which through generation of nitric oxide (NO) also has direct parasite killing effects 65. TNF-α can also exacerbate inflammation by inducing cyclooxygenase (COX)-2 and subsequently generated effector molecules, such as prostaglandins 66. Many of the signs and symptoms associated with malaria, such as fever, headache, nausea, vomiting, diarrhea, anorexia, myalgias, and thrombocytopenia can be linked to TNF-α 67.

IFN-γ is produced by natural killer and αβ-T cells, as well as the regulatory γδ-T cells, during the early phases of the immune response to a malaria infection 68-70. This prototypical type 1 cytokine is a key molecule for protecting against infection in childhood malaria 71 and in non-immune volunteers experimentally infected with malaria 72. The ability to generate IFN-γ from mononuclear cells exposed to asexual malaria parasites has recently been attributed to the significant resistance to P. falciparum malaria in the Fulani people of West Africa compared to other tribes in the Mali, Burkina Faso, and Sudan regions 73. Consistent with a protective role, IFN-γ responses to CD8+ T cell epitopes from pre-erythrocytic antigens are associated with higher Hb levels, and reduced prevalence of SMA in Kenyan children 74.

Although both TNF-α and IFN-γ appear to play a protective role in children and adults during the early stages of a P. falciparum infection through their ability to stimulate monocyte/macrophage activation and aid in controlling parasitemia 75, over-production of these innate inflammatory mediators is also associated with anemia 76, 77. In addition, persistent macrophage activation is significantly greater in children with complicated malaria 78. Excessive release of IFN-γ and TNF-α, along with NO, also promote enhanced malarial anemia pathogenesis by contributing to bone marrow suppression, dyserythropoiesis, and erythrophagocytosis 79.

IL-1 is a potent endogenous pyrogen that promotes an acute inflammatory response and provides a first line of defense against invading pathogens 80. IL-1β and IL-1α synergize with TNF-α to enhance the production of NO and IFN-γ in murine models of malaria 81. However, high levels of sustained IL-1β production in inflammatory diseases can induce a number of hematological abnormalities, including anemia 82, 83. In murine models, administration of recombinant IL-1 can inhibit development of the pre-erythrocytic stages of malaria 84, protect against the development of cerebral malaria, and aid in controlling parasitemia 85. Although several studies have reported elevated peripheral blood levels of circulating IL-1β levels in humans with severe malaria 86-88, an additional investigation failed to find significant changes in IL-1β levels in children with SMA 76. Studies in Gambian children with falciparum malaria illustrated that IL-1α and TNF-α levels, measured upon admission to hospital, were positively correlated with venous blood lactate concentrations, which were approximately two-fold higher in fatal cases compared to survivors 89. A study by our group investigating the role of IL-1β in the immunopathogenesis of SMA revealed that children with SMA had significantly lower levels of IL-1β than parasitized children without SMA 90. In addition, haplotypes of IL-1β promoter polymorphisms that were associated with significantly greater risk of developing SMA were also associated with reduced IL-1β production, whereas those haplotypes associated with protection against SMA produced higher levels of IL-1β 90. Thus, although sustained production of IL-1β can promote anemia 82, 83, it appears that 'high producing' haplotypes of IL-1β provide protection against SMA.

Elevated levels of IL-6 in the peripheral blood of patients with severe P. falciparum malaria was recognized two decades ago 57. This finding has since been corroborated by a number of studies showing elevated IL-6 levels in children with severe malaria 75, 76, 88. An investigation in Gabonese children further demonstrated that peripheral blood mononuclear cells (PBMC) are a primary source of increased IL-6 production during acute malaria 91. However, studies in murine models illustrate that IL-6 mediates protective immunity against the pre-erythrocytic stages of malaria by inducing IL-1β and TNF-α, and during the erythrocytic stage of disease by controlling parasitemia through boosting of specific immunoglobulin (Ig) G antibodies 84. Experimental infections with P. falciparum in humans support these findings in that early IL-6 production is associated with protective effects 92. Consistent with the protective effects of IL-6, lower circulating levels of IL-6 are associated with hyperparasitemia in Malian children with falciparum malaria 76. Taken together, these studies support a protective role for IL-6 during the early stages of disease by controlling parasitemia. However, lack of control over the parasitemia and the resulting progression towards severe disease may explain the association between elevated levels of IL-6 and enhanced pathophysiology.

MIF was the first soluble mediator described in malaria 93, but has only recently been more fully explored in the context of malaria pathophysiology by our group and a number of others 47, 94-98. MIF is a ubiquitous cytokine produced in response to pro-inflammatory stimuli by T cells 99, 100, monocytes/macrophages 101, and the anterior pituitary gland 102. However, unlike most cytokines, MIF is constitutively expressed at high levels and stored in preformed vesicles, and as such can therefore be rapidly released without de novo gene expression 102, 103. MIF has potent pro-inflammatory properties that are important for both innate and adaptive immune responses to bacterial and parasitic infections 100, 101, 104-106. Murine models of malaria demonstrated that elevated MIF levels were associated with enhanced disease severity 107 with the MIF gene knock-out mice having less anemia and higher survival rates following infection with P. chabaudi compared to wild types 108.

Previous investigations in humans showed that MIF was elevated in intervillous blood during placental malaria 109, 110, thoracic blood vessels of Malawian children with cerebral malaria 111, and peripheral blood from Zambian children with acute malaria 108. However, results from our laboratory were the first to show that elevated MIF protein (in circulation) and MIF transcripts (in PBMC) were associated with less severe forms of falciparum malaria 94. We subsequently confirmed these results in a larger study of children (aged <3 years, n=357) with P. falciparum-induced malarial anemia in western Kenya 47. In the Kenyan cohort, circulating MIF concentrations declined with increasing severity of anemia and significantly correlated with peripheral blood leukocyte MIF transcripts. Interestingly, MIF concentrations in peripheral blood were not significantly correlated with reticulocyte responses in these children. However, multivariate regression modeling, controlling for age, gender, and parasitemia, demonstrated that elevated levels of PCM were significantly associated with both SMA and decreased MIF production. As a complement to the in vivo studies in children with malarial anemia, additional experiments were conducted in PBMC from malaria-naïve individuals which showed that phagocytosis of PfHz caused dysregulation in MIF production in a apoptosis-independent manner 47. Taken together, investigations in Gabonese and Kenyan children with malarial anemia demonstrate that elevated levels of monocytic PfHz are associated with suppression of peripheral blood MIF production and enhanced severity of anemia.

Another pro-inflammatory mediator that appears important in conditioning the pathogenesis of SMA is IL-23. Although largely unexplored in the context of SMA, IL-23 is important in mediating the development of anemia in autoimmune diseases 112 and chronic inflammation 113. IL-23 is composed of two subunits, p19 and p40 114. IL-23 shares a number of common properties with IL-12, including the p40 subunit 115, the ability to bind to the IL-12Rβ1 receptor 116, release from activated myeloid antigen presenting cells, promotion of a type 1 immune response 114-118, and suppression by both IL-10 119, 120 and IL-12p40 homodimers 115, 121. In addition to the common properties IL-23 shares with IL-12, there are also distinct immunological roles in that IL-23 acts on activated memory CD4+ T cells, while IL-12 promotes Th1 differentiation of naïve CD4+ T cells 114, 122. Based on the common and distinct roles of IL-23 and IL-12, along with the well established importance of IL-12 in the pathogenesis of malarial anemia (discussed below in detail), we explored the relationships among these cytokines in Kenyan children with varying severities of malarial anemia 123. Children with malarial anemia had increased peripheral blood levels of IL-23 and suppressed IL-12 relative to healthy controls. Additional experiments in cultured PBMC revealed that hemozoin caused a sustained induction of IL-23p19 transcripts over 72 h, while IL-12p40 and IL-10 transcripts peaked at 24 h, and rapidly declined thereafter. This line of investigation suggests that elevated IL-23 levels may be important in the pathogenesis of SMA, and that both IL-10 and IL-12 may regulate IL-23 production during an infection with P. falciparum.

Perhaps one of the most important innate inflammatory mediators in the pathogenesis of SMA is IL-12, a heterodimeric protein composed of 35 and 40 kDa subunits, and a prototypical cytokine of the type 1 immune response 124, 125. IL-12 is secreted from dendritic cells, monocytes, and B-cells in response to bacterial cell wall components, intracellular pathogens, and CD40 ligation 124-126. IL-12 stimulates production of IFN-γ and TNF-α from T-cells and natural killer (NK) cells 124, 125, thereby further augmenting type 1 responses. A number of cytokines and chemokines can promote IL-12 [e.g., granulocyte macrophage-colony stimulating factor (GM-CSF) and IFN-γ], while others decrease IL-12 production [e.g., IL-4, IL-10, IL-11, IL-13, monocyte chemotactic protein (MCP)-1/CCL2, and TGF-β] 125, 126. As such, the overall ability of the innate immune response to generate IL-12 is an important event that mediates the development of malarial anemia.

Previous studies in the murine models showed that administration of recombinant IL-12 and chloroquine ameliorated blood-stage malaria and severe anemia, and induced immunity against re-infection 127. Additional murine malaria studies demonstrated that deficient IL-12 production is associated with severe anemia and dyserythropoiesis 128. The protective responses associated with IL-12 against blood-stage malaria appear to be due to increased IL-12 production from splenic macrophages and NK cells 129, 130 and the ability of IL-12 to stimulate antibody production 131. Central to the role of IL-12 in malaria is its ability to act as a hematopoietic growth factor 132, 133. In concert with IL-3, IL-12 along with IL-6, granulocyte-macrophage colony-stimulating factor (GM-CSF), and IL-11 can bolster colony formation in dormant hematopoietic progenitors during times of cytopenic crisis 132, 133. Consistent with this role, our previous studies in children with falciparum malaria showed that suppression of circulating IL-12 is associated with enhanced malarial anemia 77, 134. Suppression of IL-12 in these children was also associated with increasing concentrations of PfHz-containing leukocytes 134. Additional studies from our laboratories demonstrated that children with SMA have reduced IL-12 levels through a mechanism that involves, at least in part, phagocytosis of PfHz which promotes up-regulation of monocyte-derived IL-10 that, in turn, suppresses IL-12p40 subunits 135.

Anti-inflammatory Mediators

High levels of anti-inflammatory cytokines, such as IL-10, appear to provide protective effects against SMA by preventing the over-production of pro-inflammatory mediators. Anti-inflammatory cytokines are typically produced during the later stages of the innate immune response to P. falciparum in which they down-regulate the potentially pathogenic pro-inflammatory responses that are important for controlling parasitemia 136. Previous investigations show that a low IL-10 to TNF-α ratio is associated with enhanced severity of malarial anemia 77, 137, suggesting that the timing and magnitude of pro-inflammatory cytokine production, relative to the anti-inflammatory cytokine response, is an important determinant of the clinical outcomes of malaria. Although plasma IL-10 levels are elevated in Malian children with SMA relative to healthy controls 76, studies in Ghana showed that plasma IL-10 levels were significantly lower in children with SMA compared to children with cerebral malaria, uncomplicated malaria, or moderate malarial anemia 138. Additional studies have shown significant associations between circulating IL-10 levels and pigment-containing phagocytes in the peripheral blood 134, suggesting that malarial pigment plays an important role in governing the systemic pathology of malaria through up-regulation of IL-10 production.

TGF-β1 is an anti-inflammatory cytokine (and growth factor), which down-regulates the production of TNF-α and IL-10 and protects against severe murine malaria 139. TGF-β appears to be important in malaria pathogenesis 86, 110, 140, and is attributed to both positive 141 and negative 141-143 effects on erythropoiesis. Previous studies from our group showed that TGF-β1 levels in peripheral blood are significantly reduced in Gabonese children with severe malaria 77. However, studies in Burkina Faso revealed opposite effects in which severe childhood malaria was associated with increased plasma concentrations of TGF-β1 144. The reason for differing results may be due to the differences in malaria endemicity since the rural site in Lambaréné, Gabon, has a high level of P. falciparum transmission, whereas the urban region of Ouagadougou, Burkina Faso, is a mesoendemic area for P. falciparum. Alternatively, it is unclear whether or not 'platelet poor' samples were used to measure TGF-β1 in the Burkina Faso study, an important consideration since platelets contain high levels of TGF-β1. A more recent study further supports the importance of the TGF-β family in malaria pathogenesis by showing that serum levels of the soluble form of the TGF-β co-receptor, endoglin (sEng or CD105/ TGF-βRIII), was significantly elevated in children with severe falciparum malaria 145. In addition to TGF-β1, a recent investigation in Angolan children illustrates that polymorphic variability in TGF-β2 conditions susceptibility to the risk of progressing to cerebral malaria 146. It remains to be definitively determined if TGF-β2 is also an important mediator of SMA pathogenesis.

Chemokines

Chemotactic cytokines, or chemokines, are primarily known for their chemotactic properties, but also play important roles in immune activation, hematopoiesis, angiogenesis, and antimicrobial activities 147. Burgmann et al. were the first to investigate chemokines in the context of malaria in 1995 by measuring the C-C chemokine, macrophage inflammatory protein (MIP)-1α/CCL3, and the C-X-C chemokine, IL-8/CXCL8, in the serum of acutely infected adult patients with P. falciparum infections in which they found a positive correlation between parasitemia and IL-8/CXCL8 148. A subsequent investigation revealed that Malian children with severe malaria had ten-fold higher concentrations of IL-8/CXCL8 compared to either healthy controls or individuals with uncomplicated malaria 76. IL-8 is an important neutrophil activating chemokine that was also previously shown to be elevated in Thai patients with severe, non-fatal malaria 149. Additional studies in Gabonese children and adults illustrated that higher plasma IL-8 levels were associated with acute malaria and a slow rate of cure after malaria chemotherapy 75.

It appears that phagocytosis of PfHz is an important signal for promoting chemokine production and/or suppression. For example, PfHz treatment of a bone marrow-derived murine cell line increased transcript levels of MIP-1α/CCL3, MIP-1β/CCL4, MIP-2/CXCL2, and MCP-1/CCL2 150. Concomitant studies from our group demonstrated that Gabonese children with severe malaria had elevated levels of MIP-1α/CCL3 and MIP-1β/CCL4 protein (measured in circulation) and transcripts (determined in ex vivo PBMC) 151. Additional experiments in cultured PBMC from healthy, malaria-naïve donors revealed that PfHz promoted increased MIP-1α/CCL3 and MIP-1β/CCL4 production 151.

Regulated on activation, normal T-cell expressed and secreted (RANTES, CCL5) also appears to play a critically important role in the pathogenesis of SMA. RANTES is secreted by a number of cell types including monocytes, macrophages, fibroblasts, NK and T cells, and CD34+ hematopoietic progenitors 152-155. RANTES protein is sequestered in the α-granules of platelets and is released by thrombin stimulated platelets 156, indicating that RANTES is involved in both innate and adaptive immune response. In addition, RANTES stimulates hematopoiesis, angiogenesis, cell proliferation, and development 157. An earlier study by our group, which was the first to examine RANTES in the context of malaria, demonstrated that RANTES was suppressed in Gabonese children with severe malaria, at least in part, through PfHz-induced down-regulation of RANTES transcripts in PBMC 151. The inherent ability to produce RANTES/CCL5 also appears important in conditioning susceptibility to severe malaria. For example, our investigation in Gabon revealed that healthy children with prior mild malaria produced significantly higher RANTES transcripts and protein than children with a history of severe malaria 151. These investigations were then confirmed in Kenya where we showed that RANTES was significantly suppressed in children with SMA 29. Suppression of RANTES in these children was also significantly associated with inefficient erythropoiesis and malaria-induced thrombocytopenia 29. Subsequent studies from our laboratories determined that naturally acquired PfHz by monocytes promotes suppression of RANTES in children with malarial anemia through an IL-10-dependent mechanism 49. Taken together, these findings suggest that thrombocytopenia may be an important source of reduced RANTES which appears to contribute to suppression of erythropoiesis in children with SMA.

Growth Factors

Since a number of growth factors influence the erythropoietic cascade, and SMA is clearly a disease process in which phagocytosis and lysis of RBCs necessitates the production of new RBCs to recover from anemia, growth factors will clearly emerge as critical determinants of clinical outcomes. However, the literature is largely lacking with respect to the importance of growth factors in conditioning the development of SMA. A time-course study in patients with P. falciparum malaria in Thailand showed that serum levels of granulocyte-colony stimulating factor (G-CSF) were significantly elevated in individuals with complicated disease on day 0 that then declined to within the normal range by day 7; G-CSF at day 0 was correlated with procalcitonin, parasite density, and erythropoietin 158. Although not specifically examined in children with SMA, the potential negative consequences associated with over-production could certainly emerge since G-CSF has a negative impact on erythropoiesis 159-161.

GM-CSF is important for promoting erythropoiesis 162 and synergizes with TNF-α to increase the killing capabilities of neutrophils on blood-stage malaria parasites 163. In addition, enhanced pathology in a murine model of malaria (characterized by parasitemia and anemia) was associated with elevated levels of erythropoietin (EPO), which were strongly correlated with the level of anemia, and negatively correlated with GM-CSF concentrations 164. The influence that GM-CSF has on anemia outcomes in children with SMA remains to be determined.

As part of our investigations aimed at identifying genes/gene pathways that could play a role in the pathogenesis of SMA, we utilized gene expression profiling from pooled fractions of human PBMCs stimulated with PfHz. These experiments revealed that human stem cell growth factor [SCGF, C-type lectin domain family member 11A, (CLEC11A)] was up-regulated following treatment with PfHz 165. SCGF is a hematopoietic growth factor, expressed primarily by myeloid cells and fibroblasts that possess burst-promoting activity for human bone marrow erythroid progenitors 166. Human SCGF-α is a 323-amino acid protein, while SCGF-β is a 245-amino acid protein that results from cleavage of the conserved carbohydrate domain 167. After determining the in vitro kinetics of SCGF expression in response to PfHz, we then examined circulating SCGF levels in Kenyan children with malarial anemia. SCGF levels in circulation and in cultured peripheral blood were significantly suppressed in children with SMA, with circulating SCGF levels being positively correlated with Hb concentration and the RPI 165. SCGF was significantly lower in children with a suppressed erythropoietic response and in children with high levels of naturally acquired monocytic PfHz 165. Our additional investigation showed that a novel SCGF promoter variant (-539C/T, rs7246355) was significantly associated with susceptibility to SMA and reduced erythropoietic responses with the 'high producing' TT genotype protecting against development of SMA and suppression of erythropoiesis in parasitized children 168. Taken together, these results illustrate that SCGF is an important mediator of SMA pathogenesis that may offer the potential for immunotherapy in future clinical trials.

Effector Molecules

As described above, the clinical outcomes of malaria are largely conditioned by the relative expression of inflammatory mediators. The relative timing and magnitude of pro- and anti-inflammatory cytokines, chemokines, and growth factors released into the inflammatory milieu have direct actions on the cellular response as well as the 'down-stream' effector molecules that ultimately get produced. As such, effector molecules play a critical role in the pathogenesis of SMA. One important effector molecule in malaria is the toxic free radical, NO. NO and equimolar amounts of L-citrulline are generated by catalysis of L-arginine by the NO synthases (NOS) 169. In the context of an acute inflammatory disease, such as malaria, much of the NO produced comes from the cytokine inducible isoform, nitric oxide synthase type 2 [NOS2 or inducible NO synthase (iNOS)] present in monocytes, macrophages, and neutrophils 170. In general, pro-inflammatory cytokines (e.g. IL-12, IFN-γ, and TNF-α) increase NOS2-generated NO production, whereas anti-inflammatory cytokines (e.g. IL-10 and TGF-β) down-regulate NOS2 expression (for review see 171). Although the role of NO in the pathogenesis of malaria has been debated for nearly a decade, it is apparent that NO is both protective and pathogenic. For example, NO is protective in that it has potent parasiticidal properties against P. falciparum 65 and can thereby limit parasitemia 172. A protective effect is also illustrated by our previous investigation showing that healthy, malaria-exposed Gabonese children with a history of mild malaria have significantly higher levels of ex vivo PBMC NO production and NOS enzymatic activity than their age-matched cohort with a history of severe malaria 173. However, our follow-up investigation in the same population revealed that ex vivo and in vitro NOS activity in PBMCs was significantly higher in children with malarial anemia in which there was an inverse association between NOS enzyme activity and hemoglobin levels 174. Additional experiments confirmed that PfHz was an important source of NOS2 transcripts and NO production 174. Thus, although NO serves an important role in controlling parasitemia, it is likely that sustained, high levels of NO production also promotes anemia. This premise is supported by the fact that generation of NO during a malarial infection can promote severe anemia through bone marrow suppression, dyserythropoiesis, and erythrophagocytosis (for review see 79).

In addition to NO, reactive oxygen species (ROS) also appear to be both protective and pathogenic in human malaria. High levels of oxygen radical production are associated with accelerated clearance of parasitemia in Gabonese children with falciparum malaria 175. In addition, ROS are important for controlling peripheral parasitemia in children with severe malaria 176. A pathogenic role for ROS is illustrated by a study in Kenyan children showing that ROS cause damage to the erythrocytic membrane (demonstrated by measurement of α-tocopherol and polyunsaturated fatty acid levels in the erythrocyte membrane) 177. A recent investigation in Indian children (<15 years of age) infected with P. falciparum (in which the primary disease manifestation of severe malaria was SMA) revealed that severe malaria cases had significantly elevated markers of oxidant stress, including malondialdehyde, protein carbonyl, nitrite, ascorbic acid, and copper levels 178. Although a wealth of data exists on the topic of free radicals in malaria, such a discussion is beyond the scope of this review. However, studies outlined here highlight both the protective and pathogenic roles of reactive nitrogen and oxygen intermediates.

Prostaglandin (PG)E2 is synthesized from arachidonic acid (AA) through the catalytic activity of cyclooxygenase (COX) enzymes also known as prostaglandin-H2 (PGH2) synthase, which exists in two isozymes: COX-1 (PGH synthase-1) and COX-2 (PGH synthase-2). Constitutively expressed COX-1 catalyses immediate biosynthesis of PGE2 and other prostanoids involved in physiological homeostasis, whereas inducible COX-2 catalyses delayed formation of PGE2 and prostanoids involved in regulating the inflammatory response and immunity to invading pathogens 179. Formation of PGH2, the committed step in prostanoid biosynthesis, promotes generation of primary prostanoids [i.e., PGE2, PGH2, thromboxane A2, PGD2, PGF2α, and prostacyclin (PGI2) through the action of respective terminal prostanoid synthases 179. Our previous study illustrates that intervillous blood mononuclear cell (IVBMC) PGE2 production is reduced in parasitemic women of all gravidae due, at least in part, to acquisition of intraleukocytic PfHz 180. Additional studies from our group have shown that plasma bicyclo-PGE2 (a stable end metabolite of PGE2) and ex vivo PBMC COX-2 gene expression are significantly reduced in Gabonese children with severe malaria 181. Studies in Tanzanian children also showed that suppression of systemic bicyclo-PGE2 production (measured in urine) was suppressed in children with cerebral malaria 182. In addition, in vitro experiments in our laboratories revealed that reduced PGE2 biosynthesis in children with falciparum malaria was largely due to inhibition of de novo COX-2 transcripts following phagocytosis of PfHz by monocytes 183, 184. Further investigation of the role of prostaglandins in childhood malaria showed that suppression of PGE2 by PfHz and commonly used antipyretics to treat the malarial fever promoted over-production of TNF-α, an event associated with enhanced malaria pathogenesis 183, 184.

12. In vitro Models for Investigating Suppression of Erythropoiesis in Malarial Anemia

In vitro studies have shown that PfHz directly (or in synergy with TNF-α) inhibits erythroid cell development 46. In addition to inducing erythropoiesis-inhibiting cytokines, PfHz has direct effects on the erythropoietic cascade through its ability to cause apoptosis of erythroid precursor cells via oxidative stress 185. We recently developed an in vitro model of erythropoiesis using CD34+ stem cells isolated from human peripheral blood to investigate the effects of inflammatory mediators on erythroid development 186. This model showed that PfHz only slightly suppressed erythroid cell proliferation and maturation, marked by decreased expression of glycophorin A (GPA) 186. However, the addition of PfHz-stimulated PBMC-conditioned media (PfHz-CM), recombinant TNF-α, and NO donors significantly inhibited erythroid cell proliferation 186. The decreased proliferation witnessed in cells treated with PfHz-CM and NO was accompanied by increased apoptosis of erythropoietin-stimulated CD34+ cells 186. Interestingly, the addition of NO donors significantly inhibited erythroid cell maturation, whereas TNF-α failed to impact on maturation. These results demonstrate that PfHz suppresses erythropoiesis by acting directly on erythroid cells, and to a greater extent, through indirect effects in which phagocytosis of PfHz generates inflammatory mediators which have adverse effects on erythroid development.

13. Conclusion

Studies outlined here support a model in which the pathogenesis of SMA is largely driven by dysregulation in pro- and anti-inflammatory cytokines, chemokines, growth factors, and effector molecules. Altered patterns of these innate inflammatory mediators is due, at least in part, to the phagocytosis of PfHz by monocytes, resident macrophages (including those in bone marrow), and neutrophils. The mechanisms that lead to the profoundly low Hb concentrations witnessed in children with SMA are due to hemolysis and phagocytosis of parasitized and non-parasitized RBCs, and to a large extent, by suppression of erythropoiesis that is driven by PfHz-generated dysregulation in innate inflammatory mediators. While it is clear that overcoming the global burden of SMA will continue to present a serious challenge, it is our hope that gaining an improved understanding of pathogenic events that cause suppression of erythropoiesis, particularly dysregulation in innate immunity, may offer future treatment strategies to combat the unacceptable rates of morbidity and mortality associated with SMA.