Each year, infection with Plasmodium falciparum
causes 400–600 million cases of malaria and up to 1 million childhood deaths in Africa 
Here, we demonstrate that oAC protects mice from death due to ECM, with associated changes in T cell cytokine production and in whole blood gene expression profiles. Importantly, the changes in immune response and gene expression profile observed following oAC treatment did not impact negatively on host protection against PbA infection per se
. In a randomized controlled open label Phase I study, we found no interference of oAC with the pharmacokinetics of the currently best available anti-malarial drug AS. Collectively, these data pave the way for clinical evaluation of oAC as a new safe, effective and affordable adjunct treatment for severe malaria in man.
Our data indicate that oAC almost completely inhibited the clinical and histopathological signs of ECM in mice. In those mice that did develop ECM, onset was delayed, and in surviving mice oAC also appeared to provide a degree of protection against death due to high parasitemia, as measured by overall survival time. However, further studies are clearly needed to understand the mechanism(s) by which oAC mediates these striking effects. In chronic kidney disease, it has been proposed that oAC serves as a sink to block intestinal absorption of indole, thereby limiting hepatic production of indole sulphate, a regulator of TGFβ production 
. Other data suggest the possibility of bile-mediated transfer of cytokines to the intestinal lumen via the entero-hepatic pathway, with charcoal acting as a presumptive local adsorbant 
Two lines of evidence from the current study suggest that oAC has broad systemic effects on immune and inflammatory processes. First, oAC treatment significantly reduced the number of splenic CD4+
T cells capable of TNF+
production and of CD4+
T cells co-producing TNF and IFNγ, as determined after re-stimulation. Although no effect of oAC was observed on cytokine production measured directly ex vivo, such an observation is in keeping with reports that serum pro-inflammatory cytokines, included those measured above, peak earlier in infection than the time point used in our analysis 
and with the data of others that indicates a greater ability to detect multi-functionality after re-stimulation 
. Similar associations between disease outcome and cytokine levels have been observed in other models in which the course of ECM has been altered. For example, in MyD88−/−
C57BL/6 mice, which are resistant to ECM, production of IFNγ, TNF and IL-17 was reduced 
. Similarly, in mice treated other newly proposed adjunct therapies that show similar levels of efficacy against ECM as demonstrated here for oAC, including panthenine 
and rosiglitazone 
, levels of TNF were also reduced. For rosiglitazone, dampening of inflammation and enhanced parasite clearance have also been observed in humans given this drug as adjunct therapy (with atovaquone/proguanil) for mild malaria 
, confirming the translational potential of studies in murine models of disease.
Second, microarray analysis identified a clear whole blood transcriptional signature that distinguished oAC-treated from untreated PbA-infected mice. Gene expression analysis in ECM has previously been largely restricted to using spleen cells or brain tissue 
making comparisons across either genetically disparate hosts or using different parasite strains. In contrast, in the application of this technology to human malaria has been restricted to analysis of PBMC or whole blood 
, again usually comparing individuals with distinct disease outcome. To our knowledge, no direct comparisons of gene expression profile have been made following interventions that seek to prevent the development of severe disease. Further analysis of our gene expression data, including refinement of the signature associated with oAC treatment, will be published elsewhere. The data shown here nevertheless provides evidence that oAC treatment directly or indirectly affected gene regulation as determined in whole blood from PbA-infected mice. The signature of 68 genes associated with oAC treatment included multiple genes involved in the acute phase response and inflammation, as well as genes involved in heme biosynthesis and erythrocyte function (e.g. Fech
, and Slc25a37
). A case for the development of blood transcriptional biomarkers has been extensively argued elsewhere 
, and the application of a ‘modular’ approach to genomic analysis of the response of SLE patients to treatment has been described recently 
. It will be important in the future to determine whether similar or distinct signatures are associated with other interventions designed to interrupt the progression of ECM, and when oAC is evaluated for protection against CM and severe disease in man, to likewise determine whether such signatures have cross-species predictive value.
Although a molecular mode of action for oAC has yet to be identified, we believe that the experimental data reported here should nevertheless now provide a basis for the evaluation of oAC as a treatment for severe malaria. oAC may provide a first line therapy in the immediate absence of alternate treatment. Severe malaria is an acute illness that may present with neurological symptoms occurring often within 96 h of the onset of fever; much of this time may be spent traveling from remote villages to health clinics and consequently many children arrive in coma. If our observations from the murine model would translate to man, oAC given early in the course of infection could prevent the development of CM and may reduce CM-associated mortality. It is notable that AC has been used for many years in the treatment of poisoning, and can be given orally or via a naso-gastric tube, particularly in powdered form. It is well tolerated, has an excellent, well-documented safety profile, and is relatively inexpensive. Furthermore, AC has a long shelf life and can be stored at ambient temperatures. Should clinical efficacy be proven, these attributes would make AC highly suited for use in remote rural communities where it could be administered at the first point of care, for instance by a village health worker, and would encourage the effective uptake of charcoal therapy in malaria-endemic countries.
Nevertheless, given that specific treatment for malaria is available, and that oAC treatment alone had no anti-malarial activity in mice, the straightforward approach of assessing the efficacy of oral AC as a stand-alone therapy would be unethical in humans. Indeed, further exploration of the potential benefits of oAC as adjuvant treatment have to take into account that oAC is officially recommended to treat intoxication with quinine 
, increasing the elimination of this drug 
. However, it was unknown whether the pharmacokinetics of AS, a drug now regarded to be superior to quinine for the treatment of severe malaria 
and known to have a high endogenous clearance rate 
, would be affected by co-administration of oAC. We found no evidence of reduced plasma levels of AS or DHA due to oAC, when given simultaneously with AS, or when administered 1 h later. In arms 2 (AS and oAC simultaneously) and 3 (oAC 1 h after AS), individuals received their first dose of oAC 12 h prior to starting plasma sampling. The fact that the AS and DHA concentrations found in these groups were not lower than those observed in the control group support the conclusion that even a high concentration of AC in the intestine at the time of administration of AS did not interfere with the pharmacokinetics of AS or DHA. Although the trial did not have an equivalence design, all subjects who started therapy followed the protocol schedule and there was no rescue therapy. This reduces concerns about interpreting the results as if they had come from an equivalence trial. Although the sample size required for an equivalence trial would have been different, the sample size actually achieved (which was reduced by the failure to analyse some samples) is of course reflected in the width of the confidence intervals. These data therefore suggest that clinical trials of combined AS and oAC treatment in patients with malaria could proceed without compromising the anti-malarial activity of AS.
Finally, with the limited resources available at the primary health care level, it is often not possible to reliably assign the correct diagnosis to a severely ill child, especially to differentiate between malaria and a bacterial infection (e.g. pneumonia, typhoid/non-typhoid salmonella or meningitis) 
. Treatment is mainly empirical on a presumptive diagnosis, and clinical assessment is often disturbingly poor 
resulting in an over-diagnosis of malaria at the expense of severe bacterial infections 
. For an adjuvant treatment to be employed successfully, it needs to be safe and, ideally, beneficial irrespective of the precise diagnosis. oAC could be the ideal candidate for this purpose due to its non-specific anti-inflammatory potential, exploited successfully for over a decade to prevent progression of chronic kidney disease 
In conclusion, oAC prevented mortality associated with ECM in a mouse model and when given to humans did not interfere with the pharmacokinetics of AS or DHA. oAC therefore has the potential to be a readily-implemented therapy for the treatment of severe malaria. We suggest there is now an urgent need for controlled clinical trials to evaluate the efficacy of oAC in human malaria. We propose that the next step in the development of oAC should be to evaluate its safety profile as an adjuvant therapy in cases of uncomplicated malaria, a path recently taken for the development of rosiglitazone 
, before progressing to an evaluation of its potential benefit in cases of severe malaria. Incorporation of gene expression profiling into such trials should also yield further insights into the mode of action of AC in man.