Jos University Teaching Hospital (JUTH) and Florida State University Human Subjects Committees approved the protocol for a physician-based malaria team from Jos, Nigeria, to perform a malaria clinical outreach/study of children under six years of age in the Barkin Ladi Village Clinic 45. The Barkin Ladi Village Clinic is located in a region mesoendemic to P. falciparum infection with average spleen rate of 28-30% during high transmission season and hyperparasitaemia occurring in 0.6% of the patients 46 and represents the situation found in many parts of sub-Saharan Africa, where advanced hospital facilities are not available to the patient, physician, or researcher.

Patients presenting with slide positive falciparum malaria were evaluated by a physician for general signs of neurological well-being, hepatomegaly, splenomegaly, fever, and wasting. All patients in the study were characterized as having uncomplicated falciparum malaria through established criteria 264647. Patients were given a single therapeutic dose of sulphadoxine-pyrimethamine (SP, Fansidar^®), previously assessed by members of the current research team as the drug of choice for the Plateau region of Nigeria because of minimal trophozoite recrudescence 46. Patients placed on the study received a seven-day post-treatment follow-up visit by the physicians. All patients were given an insecticide treated mosquito net to reduce the rate of re-infection whether they participated in the study or not.

Blood from slide-positive children under six years of age that was sero-negative for HIV (Oraquick Rapid Antibody HIV Test #3001-0951, OraSure Technologies Inc, Bethlehem, PA), and that had a packed cell volume greater than 25% 47 was collected pre- and seven-days post SP-treatment. Blood samples were collected in sterile blood collecting vacutainers with 15% EDTA, 3.8% Na⁺ citrate, or without additive for serum collection (Becton Dickinson, Franklin Lakes, N.J.). The EDTA-containing samples were immediately placed on ice. The Na⁺ citrate and serum samples were placed in a foam-insulated cooler with an ice pack tethered inside to maintain a cool environment without contact of the tubes with the coolant. Blood samples were taken at the end of the day so that all samples collected were centrifuged within an hour and a half of collection. Na⁺ citrate and serum samples were spun at ambient temperature (27°C) for 15 minutes at 1000 × g, whereas EDTA samples were spun at 4°C for 15 minutes at 1000 × g in a refrigerator centrifuge on site (powered with a generator). The centrifuge had adequate bucket space for simultaneous spinning of all serum and Na⁺ citrate samples at ambient temperature first, followed by centrifugation of EDTA samples at 4°C. Retrieved serum and plasma samples were transferred aseptically to Nalgene cryovials (Fisher Scientific) and stored at -70°C in a freezer at JUTH. Samples were transported by carrier in a Saf T Case (STP 350, SAF-T-PAK, Inc) with all enclosed samples, containers, and gel packs (Arctic Pack, Packaging Products Corp., New Bedford, Mass) frozen at -70°C before air transport to FSU for analysis.

Films of freshly collected finger-prick blood samples were stained in Giemsa solution diluted in pH 7.3 phosphate buffer. A rapid staining protocol using double strength Giemsa (4%) for 10 minutes was used initially during screening to select those who had malaria parasitaemia. For parasite density estimation, a second slide was stained with 2% Giemsa solution for 30 minutes. In both screening and subsequent parasite density estimation, the slides were fixed with absolute methanol and allowed to air dry before staining 48. Two trained microscopists evaluated the slides for the presence of falciparum malaria. Peripheral blood smears from patients present at the clinic, but slide negative for falciparum malaria, were used as staining controls. A leukocyte differential analysis of slides shipped to the US was performed by two microscopists, one of whom was a physician with expertise in abnormal leukocyte fragility, detection of NETs dispersed among normal leukocytes, and the identification of neutrophil toxic granulation. The pre-treatment differential between metamyelocytes, segmented neutrophils, bands, hypersegmented neutrophils, eosinophils, basophils, lymphocytes, monocytes, NETs, and smudge forms was calculated per 100 leukocytes for each child studied pre treatment (n = 21) and post treatment (n = 19 – two slides were unavailable for differential count in U.S.). The slides were scored for Rouleaux and coined-agglutinin aggregation and for the presence or absence of NET-associated extracellular fibers sequestering parasites.

To determine the presence of chromatin in NET formation, slides were destained by the protocol of Kobayashi et al, 1989 49. Briefly, the slides were immersed in 50% ethanol at 37°C for one hour and 100% methanol at 37°C for one hour. The slides were then stained with DAPI (0.01 mg/ml in Tris-EDTA buffer solution with 10 mM 2-mercaptoethylamine, pH 7.4) to visualize the DNA.

Slides were viewed with a Nikon Microphot-FX (Nikon, Inc, Melville, N.Y.) microscope. Images were recorded with a Zeiss AxioCam MR camera (Carl Zeiss Vision GmbH, Germany). Minor contrast and gamma adjustments of the images were made with Photoshop software.

Parasite counts standardized per 200 leukocytes were obtained from thick/thin blood films. The number of parasites per microliter of blood was calculated by assuming an average white blood cell count of 10,000/μl in children.

Plasma and serum samples from children ≤5 years of age (n = 21) were analysed in duplicate both pre-and post-treatment to assess individual cytokine changes. ELISA kits were used to measure IL-10, IL-6, TNF, IL-2, IFN-γ (Pierce Biotechnology, Inc., Rockford, IL), TGF-β, and CRP (US Biological, Swampscott, MA) according to manufacturers' protocols. Positive and negative controls were included. Standard curves were determined from serial dilutions of recombinant human cytokines included in each test kit.

Serum samples from children ≤5 years of age (n = 21) were analysed evaluated by the ELISA using an assay kit for all antinuclear antibodies (ANA) as well as a kit specific for IgG ANA to dsDNA, according to the manufacturer's instructions (Wampole Laboratories, Princeton, N.J.). Both pre-and post-treatment samples were analysed to assess changes in individual patient general and anti-dsDNA ANA levels. Additionally, blood obtained during the season of low parasite transmission 46 from age-matched children (n = 18) that were slide negative for detectable trophozoites or gametocytes was evaluated for general ANA. Scoring of ANA levels was based on ANA kit index values: negative ≤ 0.90; equivocal 0.91 – 1.09; positive ≥ 1.10.

Sigma Plot was used to analyse data and assemble graphs. Independent cytokine and ANA values for pre- (n = 21) and post-treatment (n = 21) samples, and anomalies associated with NETs and cytokine induced increased fragility of lymphocytes and neutrophils, as noted on peripheral slide evaluation by leukocyte differential analysis were tested for equality of variances by Levenne's test and equality of means by Student's t test (2-tailed). Results were expressed at a 95% confidence level as the mean ± standard deviation. Cramer's V test was used for measures of nominal association, as might be found in aggregation phenomena and/or the presence of extracellular fibers in fibrinoid complexes. The results of the Cramer's V tests were evaluated according to the following criteria: <0.10 = no relationship, 0.10 – <0.20 = weak association, 0.20 – <0.25 = moderate association, 0.25 – <0.30 = moderately strong association, 0.30 – <0.35 = strong association, 0.35 – <0.40 = very strong association, 0.40 – <0.45 extremely strong relationship or the two variables are measuring the same concept, 0.45 – <0.99 = two variables probably measuring the same concept, 1.00 = perfect relationship; independent variables will predict the dependent variable.

Blood samples from 21 children five years of age and younger diagnosed with clinically uncomplicated slide-positive P. falciparum malaria were obtained before and seven days after treatment with SP during the season of high transmission. Peripheral slide analysis of Giemsa-stained thick and thin blood films revealed the children had parasitaemias ranging from 2,000/mL to 90,000/mL (mean = 30,738) before SP treatment and negligible trophozoite levels seven days after SP treatment in all but one patient (who had <250 trophozoites/mL). Gametocytes were present in 10% (2/21) of the pre-treatment and 32% (6/19) of the post treatment patients, consistent with the inactivity of SP as a gametocidal drug (data not shown). Seven (34%) of the 21 children presented with hepatomegaly (2 cm in 5 children; 4 cm in 2 children). Two of the children with 2 cm hepatomegaly also presented with 3 cm splenomegaly. No statistical correlation was found between cytokine levels and the presence of gametocytes, hepatomegaly, and/or splenomegaly.

Percentages (number per 100 leukocytes counted) of normal and abnormal lymphocytes and neutrophils were determined by a leukocyte differential count before and seven days after initiation of SP treatment with special attention to the presence of NETs (Figure 1). The post-treatment differential analysis was based on slides from 19 of 21 children, because two post treatment slides were not available for analysis in the U.S. Haematologic values were compared to a normal differential reference range 50. Remarkably, monocyte involvement was low in the children both pretreatment (mean of 1.21% ± 2, range of 1-5/100), in 29% (6/21) of the children, and post treatment (2.0% ± 1.8, 1-6/100), in 68% (13/19) of the children. Monocyte levels did not change significantly after treatment (P < 0.105). Levels of eosinophils and basophils were normal in the children both before and after treatment.

Lymphocytes were observed in all (21/21) the children pretreatment (31.7% ± 13, range of 6-61/100), and most (17/19) children post treatment (36% ± 16, range of 0-63/100), with no significant change overall between pre- and post-treatment (P < 0.359). Lymphocyte fragility was evident as leukemoid bare nuclei in smudge cells 51 to varying degrees in 95% (20/21) of the children pre-treatment (10.6% ± 8.2, range of 1-21/100). The mean increase to 18.9% ± 10.7 (range of 4-28/100) in 100% (19/19) of the children post treatment was significant (P < 0.012). Stained control smears from children presenting with fever but lacking evidence of P. falciparum on the slides exhibited no evidence of fragile leukocytes (Figure 2A).

Neutrophils dominated the pre-treatment leukocyte population with segmented neutrophils (29.3% ± 14.4, range of 9-68/100) in 100% (21/21) of the children pre-treatment and (18.4% ± 6.6, range of 2-28/100) in 100% (19/19) of the children post treatment; the post-treatment decrease in the mean number of segmented neutrophils was significant (P < 0.005). Hypersegmented neutrophils were observed in only one child (1/21). Immature neutrophils were found as metamyelocytes (1.3% ± 1.9, range of 1-7/100) in 43% (9/21) of the children pre-treatment and in 11% (2/19) of the children post-treatment (0.2% ± 0.7, range of 0-3/100) and as bands (7.2% ± 5.3, range of 1-21/100) in 95% (20/21) of the children pre-treatment and in 84% (16/19) of the children post-treatment (2.7% ± 2.5, range of 1-8/100). The post-treatment drop in bands was significant (P < 0.002), but the drop in metamyelocytes was only borderline significant (P < 0.053). Neutrophils from 14 (67%) of the children exhibited toxic granulation (Figure 2B).

All the children exhibited evidence of aggregates of circulating net-like or stranded material with adhered parasitized erythrocytes and trophozoites (Figure 2C–F). The net-like or stranded material stained with DAPI (Figure 2G–I), confirming it contains DNA. The status of the neutrophils and presence of DNA in the circulating aggregates supports the conclusion that these aggregates are NETS. This evidence for NET formation was found in 86% (18/21) of the children pretreatment (18.5% ± 9.9, range of 6-38/100) and in 100% (19/19) of the children post treatment (14.5% ± 6.4, range of 4-26/100). The slight decrease in the mean number of NETs post treatment was not significant (P < 0.155).

To determine whether the presence of DNA in the circulating NETs stimulated an immune response in the infected children, levels of ANA in the samples from the 21 falciparum-infected children were measured before and after SP treatment. ANA values determined for the children infected with malaria were compared to ANA levels in a control group of age-matched children obtained during the season of low transmission that were slide negative for detectable levels of trophozoites or gametocytes. Significantly elevated ANA levels were found in 86% (18/21) of the falciparum-infected children pre-treatment and in 100% (21/21) of the post-treatment serum samples (Figure 3A), but in only 33% (6/18) children of a control group of children whose blood was collected in a season of low transmission (Figure 3). Levels of ANA reactive with dsDNA that are predictive of autoimmunity (>60 IU/ml) were found in 81% (17/21) of the falciparum infected children pre-treatment and post-treatment (Figure 3B).

Pre- and post-SP treatment plasma samples from all 21 children also were evaluated by ELISA for levels of various cytokines: IL-10, a cytokine directed toward increased Th2 activity; IL-6, a cytokine associated with an acute immune response; TNF a multifunctional cytokine; IL-2, a cytokine associated with T-cell clonal expansion; IFN-γ, a NK- or Th1 cell-mediated adaptive immune response cytokine; TGF-β, a cytokine with immunoregulatory activity; and CRP, an acute phase complement cascade-activating protein. Mean pre-SP treatment levels of all cytokines tested except IL-2 were high compared to healthy US adult standard ranges (Figure 4). Levels of IL-2 found in all pre- and post-treatment samples were below the range of detection in the ELISA assay.

Decreased parasitaemia due to SP treatment correlated with significant decreases in the mean (and median) levels of IL-10 (~30-fold decrease in mean, p = 0.000), IL-6 (~30-fold decrease in mean, p < 0.001), TNF-a (~5-fold decrease in mean, p < 0.006), and IFN-g (~20-fold decrease in mean, p < 0.004) (Figure 3). The dramatic drops in these cytokine levels strongly suggest the pre-SP treatment levels were elevated due to effects associated with P. falciparum trophozoite parasite infection, or cytokine inhibition by SP. SP treatment had no significant effect on the post-treatment levels of either TGF-b (p < 0.607) or CRP (p < 0.023) after seven days (Figure 3).

IL-6 levels strongly correlated with the erythrocyte Rouleaux aggregation phenomenon (Cramer's V = .36, p = .104, data not shown) and very strongly correlated with the presence of extracellular fibers (Cramer's V = 0.6, p = .006), which appeared to coat circulating erythrocytes and sequester free-circulating parasites (Figure 2G–I). Extracellular fibers were observed in 13 (62%) of these children, many of whom demonstrated appliqué forms (Figure 2H) of advanced stage P. falciparum trophozoites 52.

Further analysis of the pre-SP treatment samples revealed that grouping the samples according to child age revealed significant differences (p = 0.078) in mean levels of only one of the cytokines tested IFN-γ (group data not shown separately). The ten children under 24 months old had a mean IFN-γ level of 190 pg/ml (SD ± 338 pg/ml), with undetectable levels in only two of those children (25% of children under 24 months old), whereas the eleven children 24 to 60 months old had a mean IFN-γ level of 0.47 pg/ml (SD ± 1.1 pg/ml), with undetectable levels in eight of those children (73% of the children over 24 months old). Analysis of individual pre-SP treatment samples revealed that one child, a one-year old female, had the highest levels of IL-6 and TNF (highest of the two outlier points in both Figure 4 box plots), the next to highest level of IL-10 (the second highest outlier point in the Figure 4, IL-10 box plot), and the next to highest IFN-γ (the second highest outlier point in the Figure 4, IFN-γ box plot).

Relating multiple variables in the pre-SP-treatment samples revealed a triad correlation only for levels of IL-6, IL-10, and TNF (Figure 5). The effect of IL-6, while controlling for IL-10, on TNF is statistically significant (virtually without random sampling error; p = 0.009, r = 0.852, r² = 0.726). Relating TGF-β levels to CRP levels revealed a moderately strong inverse correlation (p = 0.005, r = 0.6, r² = 0.353) (Figure 6).

Elevated TNF-a levels which did not exceed 324 pg/ml (20/21 children) correlated with NET formation (r² = 0.512) quantified by the leukocyte differential percentage of NETs and smudge forms determined by peripheral slide analysis (Figure 7). Analysis of individual pre-SP treatment samples revealed that the one year-old female described earlier generated the single outlier point.