Plasmodium falciparum antigenic variation: relationships between widespread endothelial activation, parasite PfEMP1 expression and severe malaria

Abstract

Background: Plasmodium falciparum erythrocyte membrane protein 1(PfEMP1) is a family of variant surface antigens (VSA) that mediate the adhesion of parasite infected erythrocytes to capillary endothelial cells within host tissues. Opinion is divided over the role of PfEMP1 in the widespread endothelial activation associated with severe malaria. In a previous study we found evidence for differential associations between defined VSA subsets and specific syndromes of severe malaria: group A-like PfEMP1 expression and the "rosetting" phenotype were associated with impaired consciousness and respiratory distress, respectively. This study explores the involvement of widespread endothelial activation in these associations.

Methods: We used plasma angiopoietin-2 as a marker of widespread endothelial activation. Using logistic regression analysis, we explored the relationships between plasma angiopoietin-2 levels, parasite VSA expression and the two syndromes of severe malaria, impaired consciousness and respiratory distress.

Results: Plasma angiopoietin-2 was associated with both syndromes. The rosetting phenotype did not show an independent association with respiratory distress when adjusted for angiopoietin-2, consistent with a single pathogenic mechanism involving widespread endothelial activation. In contrast, group A-like PfEMP1 expression and angiopoietin-2 maintained independent associations with impaired consciousness when adjusted for each other.

Conclusion: The results are consistent with multiple pathogenic mechanisms leading to severe malaria and heterogeneity in the pathophysiology of impaired consciousness. The observed association between group A-like PfEMP1 and impaired consciousness does not appear to involve widespread endothelial activation.

Figure 1

Plasma ang-2 level and clinical malaria (N = 213). A) Plasma ang-2 level and severe malaria. Shown in blue is the median and interquartile range, P-value determined by Mann–Whitney U test. B) Relationship between ang-2 plasma level and severe malaria syndromes: Plot of regression coefficient and 95% confidence interval obtained from an age-adjusted multi-variable regression model that predicted ang-2 using impaired consciousness and respiratory distress as explanatory variables (N = 213).

Figure 2

The relationship between ang-2 and group A-like expression and the rosetting phenotype of the infecting parasite. Scatter plots showing the relationship between A) ang-2 and group A-like expression, B) ang-2 and rosetting frequency for all the samples. C & D and E & F are repeats of A and B analysis within severe and non-severe cases respectively.

Figure 3

Group A-like var expression and ang-2 have independent association with impaired consciousness. A plot of odds ratio and 95% CI obtained from age-adjusted logistic regression models predicting impaired consciousness using; A) Group A-like expression before and after adjusting either for ang-2 or IE surface antibodies B) plasma ang-2, before and after adjusting for either group A-like expression or IE surface antibodies C) IE surface antibodies, before and after adjusting for either group A-like expression or ang-2. Asterisk indicate significance. Scatter plots showing the relationship between D) group A-like expression and peripheral parasite density, E) group A-like expression and PfHRP2.

Figure 4

Relationship between rosette, ang-2 and respiratory distress. A plot of odds ratio and 95% CI obtained from age-adjusted logistic regression predicting respiratory distress using; A) rosetting frequency (unadjusted) and ang-2-adjusted, B) ang-2(unadjusted) and rosette frequency-adjusted. Asterisk indicate significance. Scatter plots showing the relationship between C) rosetting frequency and peripheral parasitemia, D) rosetting frequency and PfHRP2, E) ang-2 and peripheral parasitemia, F) ang-2 and PfHRP2.