Self-reactivities to the non-erythroid alpha spectrin correlate with cerebral malaria in Gabonese children

Abstract

Background: Hypergammaglobulinemia and polyclonal B-cell activation commonly occur in Plasmodium sp. infections. Some of the antibodies produced recognize self-components and are correlated with disease severity in P. falciparum malaria. However, it is not known whether some self-reactive antibodies produced during P. falciparum infection contribute to the events leading to cerebral malaria (CM). We show here a correlation between self-antibody responses to a human brain protein and high levels of circulating TNF alpha (TNFalpha), with the manifestation of CM in Gabonese children.

Methodology: To study the role of self-reactive antibodies associated to the development of P. falciparum cerebral malaria, we used a combination of quantitative immunoblotting and multivariate analysis to analyse correlation between the reactivity of circulating IgG with a human brain protein extract and TNFalpha concentrations in cohorts of uninfected controls (UI) and P. falciparum-infected Gabonese children developing uncomplicated malaria (UM), severe non-cerebral malaria (SNCM), or CM.

Results/conclusion: The repertoire of brain antigens recognized by plasma IgGs was more diverse in infected than in UI individuals. Anti-brain reactivity was significantly higher in the CM group than in the UM and SNCM groups. IgG self-reactivity to brain antigens was also correlated with plasma IgG levels and age. We found that 90% of CM patients displayed reactivity to a high-molecular mass band containing the spectrin non-erythroid alpha chain. Reactivity with this band was correlated with high TNFalpha concentrations in CM patients. These results strongly suggest that an antibody response to brain antigens induced by P. falciparum infection may be associated with pathogenic mechanisms in patients developing CM.

Figure 1. IgG immunoreactivity profiles from malaria patients.

A. Example of reactivities of IgG from patients sera in the EUIC (3), UI (2), UM (2), SNCM (2) and CM (3) groups, respectively at day 0, day 7 and/or day 30 for each patient. B. Frequency of patients in each group recognizing ranges of 0–4, 5–8, 9–12 and more than 12 bands, respectively. 1: UI, 2: UM, 3: SNCM, 4: CM.

Figure 2. PCA factor loads.

Relative contributions of reactivity bands to the first two PCA factors calculated on unadjusted profiles of IgG immunoreactivity to brain proteins separated by 10% SDS-PAGE. A. PCA factor 1. B. PCA factor 2. C. localisation of the band-0 on Western blot profile obtain after the computer analysis of membrane N°5. Bands are ordered from high to low molecular weight (between about 230 kDa and 20 kDa).

Figure 3. PCA factor scores from unadjusted IgG immunoreactivity profiles.

A. Groupwise distribution of factor 1 scores. B. Groupwise distribution of factor 2 scores. C. Frequency of patients in each group with above-average factor 1 scores. D. Frequency of patients in each group with above-average factor 2 scores.

Figure 4. Properties of adjusted IgG anti-brain reactivity in terms of PCA factor 1 scores.

A. Relationship to total plasma IgG concentration, and dissection into three subgroups : α (IgG<25 mg/ml; F1<0.5), β (IgG<25 mg/ml; F1>0.5), δ (IgG>25 mg/ml; F1<0.5). B. Frequencies of patients above one year of age in each group, over the three subgroups. C. Groupwise distribution of PCA factor-1 scores among patients older than one year, with horizontal bars indicating median values. The significance of the difference between patients with CM and with non cerebral clinical malaria (UM+SNCM) is indicated.

Figure 5. Mass fingerprinting.

Proteins from human brain extract were separated on a 10% acrylamide gel. After Coomassie staining, three bands corresponding to reactive bands detected by Western blot were cut and their protein contents analyzed by peptide mass fingerprinting as described in the material and methods section. Three independent experiments were carried out and conducted to identical protein identifications. Results in figure 5 correspond to those obtain in one typical experiment.

Figure 6. Intensity of IgG reacting to band-0 in children older than one year in each group.

Horizontal bars indicate medians; the significant difference between patients with cerebral malaria (CM) and with UM+SNCM is indicated.

Figure 7. Total reactivity to brain proteins and intensity of IgG reacting to band-0 as a function of plasma TNFα levels in children older than 1 year.

A. Positive correlation between PCA factor 1 and TNFα concentrations in the CM group, indicated by dashed regression line. B. Positive correlation between unadjusted reactivity to band 0 and TNFα levels. C. High band-0 reactivity was observed most frequently in CM patients with TNFα levels above 100 pg/mL (indicated as high-TNFα in a separate group on the right). Horizontal bars indicate medians. Only children over the age of 1 year are shown.