In the context of a clinical trial evaluating the safety and efficacy of IPTi-SP as a malaria control tool during infancy, we studied prospectively the cytokine responses to P. falciparum antigens and the concentrations of plasma cytokines and chemokines at multiple time points during a 2 year follow up period. The longitudinal design and the established demographic and morbidity surveillance systems allowed for a rigorous and complete documentation of clinical malaria cases that could provide insights into the acquisition of cytokine/chemokine responses in children living in a malaria endemic area, how they are influenced by age and P. falciparum exposure, and the correlation between these responses and incidence of symptomatic malaria.
One of the main findings of this study was that overall there were no differences in the magnitude of intracellular or plasma cytokines/chemokines between IPTi treatment groups, except for some plasma cytokines and chemokines in which children receiving placebo had higher concentrations than those receiving SP (IL-12 at 9 months, IL-7 at 12 months and G-CSF at 24 months). In these cases, we hypotesize that placebo recipients, who suffered more P. falciparum infections than SP recipients owing to the efficacy of IPTi  had more systemic immune activation and higher levels of some plasma cytokines. However, of these, only IL-12 was directly associated to previous episodes of malaria in the multivariate analysis; to our knowledge, there is no data in the literature on the influence of P. falciparum infection or malaria interventions in the concentrations of IL-7 or G-CSF in peripheral blood. The levels of these cytokines/chemokines at those time points were not clearly associated with subsequent incidence of malaria.
Among the factors that significantly affected the levels of cytokines (both intracellular and plasma), age was the most prominent. Without exception, all cytokines varied significantly in magnitude from the first visit to at least one of the following visits. In all cases, there was a decrease of cytokine values from 5 to 9 months, and some had the highest levels at 24 months. This fluctuation is not surprising as cross-sectional samplings were not scheduled to pick up bursts of production upon infection and most plasma cytokines' half-lives are short , thus may not represent a long lasting response. In addition, cytokine production is stimulated and regulated by many different factors within the cellular responses network, and while some are specific to infection, other are innate or are produced in response to acute or systemic inflammatory conditions.
In contrast to antibody responses that are strongly affected by past or present malaria infection , parasite exposure appeared to have a minor effect on the concentrations of cytokines. Plasma IL-10 increased significantly in children with previous episodes or current infection, and higher levels of IL-12 and IL-13 were also associated with previous episodes. Each of these cytokines belongs to a different family (immuno-regulatory, Th1 and Th2 respectively) therefore it seems that the limited effect of exposure to P. falciparum infection was generalized and not biased towards a specific type. Similar observations have been reported  where malaria patients responded to P. falciparum infected erythrocytes with significant increases in the percentage of IL-2, IFN-γ, and TNF, but also IL-10, positive cells.
We analyzed the correlations and ratios among some representative cytokine families, as they have synergistic and antagonistic effects on each other to regulate the immune system. First, we observed no correlation between intracellular and plasma IL-4, IL-10 or IFN-γ, which is not uncommon as they represent two different measurements. Intracellular cytokines were produced by CD3+ cells after antigen stimulation in vitro (recall response) whereas plasma cytokines reflect what is present ex-vivo in peripheral blood (non malaria specific), and may be produced by many types of cells. In another study , the strength of associations between serum and cellular cytokines varied greatly, suggesting that serum cytokines at best only weakly reflect peripheral blood cell cytokine production and balances. We observed a high ratio between prototype Th1 (IFN-γ) and Th2 (IL-4) cytokines at all ages, and this proportion increased when looking only at the placebo group possibly indicating more parasite exposure. We also observed the relative transition from Th2 to Th1 responses with age .
We also assessed whether cytokines and chemokines were associated with subsequent incidence of malaria, and what factors affected these associations. Overall, high cytokine responses were not associated with a reduced incidence of clinical malaria when analyzed individually. Only when all cytokines where analyzed together in relation to risk of malaria, a significant association towards a decreased risk was found for IL-17 in the second year of life. IL-17 is a Th17 cytokine and its relation to various infectious agents has been described . However to our knowledge, there has not been any study reporting a role for IL- 17 in protection against human malaria.
In fact, the most common finding was that there were no associations, or that higher levels of some cytokines correlated with increased malaria incidence. In particular, IL-10 (intracellular and plasma) and MCP-1 were more consistently associated with incidence of malaria, and this association was not explained by age or previous episodes. IL-10, an immunoregulatory cytokine, is extensively reported in relation to malaria immunopathogenesis [9, 12, 20, 26, 27] and not so much associated with immunity . Higher IL-10 has been associated with less effective parasite clearance . To our knowledge nothing has been reported for chemokine MCP-1 in children in relation to immunity, and it does not correlate with any of the other cytokines associated with increased incidence of malaria. In pregnant women, MCP-1 concentrations were higher in the placentas of primiparous women (more susceptible to malaria) than in those of multiparous women .
In addition, IFN-γ (intracellular and plasma) and IL-13 were also associated with malaria incidence, but these were dependent on age and/or previous episodes. A few other disperse associations were observed during the second year for plasma IL-1β, IL-4, IL-5, IL-7, G-CSF, GM-CSF, when adjusted by previous episodes. These results suggest that the responses measured were not necessarily part of an acquired protective response, but rather could be interpreted as biomarkers of physiopathological processes. In other studies, IL-13, IL-4 and IL-5, Th2-type cytokines, have been associated with reduced parasite killing . Production of IL- 4 but not IFN-γ by activated human T cells has been associated with elevated antibodies to malaria antigens , and consistent with this, our previous studies also found correlations between higher antibody levels and increased malaria risk (Dobaño et al. in press).
Responses associated with reduced malaria incidence would more likely be expected in the antigen-specific cytokines, as measured by ICS. Blood cells from donors in malaria endemic areas stimulated with Plasmodium antigens are known to activate many types of cells, with production of both IFN-γ (Th1 type) and IL-4 (Th2 type) . IFN-γ is thought to be a central mediator of protective immune responses against blood stages of malaria . Surprisingly, in our study, increased frequency of IFN-γ positive CD3+ cells at 5 months was associated with higher incidence of malaria up to 1 year. This is in contrast with other studies since IFN-γ is more often reported to confer protection [30, 32–36] than pathology [9, 12, 74]. In most of these studies IFN-γ was measured in older children or adults and it is likely that the immature immune system of the infant responds differently. In the second year, however, there were no significant associations between intracellular cytokines and incidence of malaria. At a population level, memory-like IFN-γ responses have been measured following malaria infection  and it has been reported that IFN-γ responses are both more prevalent and of greater magnitude at the end of the rainy malaria transmission season , thus showing that IFN-γ might also be a marker of exposure.
The apparent predisposition of children with higher cytokine responses to increased malaria risk might also be influenced by genetic factors. Many studies have shown associations between genetic polymorphisms, immunoregulation, phenotypes and disease risk [76–91], but unfortunately this was not assessed in this study. Other factors not evaluated that could help explain these associations would be malnutrition [92, 93], co-infection with HIV or other pathogens , or prenatal exposure . The duration and/or the nature of antigen exposure in utero appears to govern the outcome with respect to neonatal immune responses, such that placental malaria induce antigen-specific IL-10-producing regulatory T cells that can inhibit Th1-type responses, while antigen-specific IFN-γ production predominate in babies born to mothers successfully treated for malaria during gestation . Prenatal infection could thus contribute to the P. falciparum-specific IFN-γ and IL-10 response pattern in 5-month old children seen in this study.