miR-155 Modifies Inflammation, Endothelial Activation and Blood-Brain Barrier Dysfunction in Cerebral Malaria

Abstract

miR-155 has been shown to participate in host response to infection and neuro-inflammation via negative regulation of blood-brain-barrier (BBB) integrity and T cell function. We hypothesized that miR-155 may contribute to the pathogenesis of cerebral malaria (CM). To test this hypothesis, we used a genetic approach to modulate miR-155 expression in an experimental model of cerebral malaria (ECM). In addition, an engineered endothelialized microvessel system and serum samples from Ugandan children with CM were used to examine an anti-miR-155 as a potential adjunctive therapeutic for severe malaria. Despite higher parasitemia, survival was significantly improved in miR-155^-/- mice vs. wild-type littermate mice in ECM. Improved survival was associated with preservation of BBB integrity and reduced endothelial activation, despite increased levels of pro-inflammatory cytokines. Pre-treatment with antagomir-155 reduced vascular leak induced by human CM sera in an ex vivo endothelial microvessel model. These data provide evidence supporting a mechanistic role for miR-155 in host response to malaria via regulation of endothelial activation, microvascular leak and BBB dysfunction in CM.

Keywords: antagomir; cerebral malaria; microRNAs; microvessels; plasmodium.

Figure 1.

Circulating exosomal miR-155 levels increased over the course of experimental cerebral malaria (ECM) and were associated with ECM susceptibility in P. berghei ANKA (PbA) infection. (A) C57BL/6 mice were more susceptible to PbA infection than BALB/c mice. Survival of C57BL/6 mice was compared with BALB/c mice infected with 1 × 10⁵ parasitized erythrocytes of PbA. Survival was significantly less in PbA-infected C57BL/6 than BALB/c mice (****P < 0.0001; log-rank test n = 9–10/group). (B) Exosomal miR-155 levels increased over the course of infection in ECM-susceptible mice. Median circulating exosomal miR-155 levels, measured by qRT-PCR, were significantly higher in C57BL/6 than BALB/c mice at d 5, immediately before the onset of ECM (****P < 0.0001; two-way ANOVA with Tukey’s multiple comparisons test; n = 4-5/group). Error bars represent standard deviation.

Figure 2.

miR-155 deficiency was associated with improved survival despite increased peripheral parasitemia in P. berghei ANKA (PbA)-induced ECM. (A) Survival was improved in mice deficient in miR-155. Survival of albino C57BL/6 miR-155^-/- versus wild-type miR-155^+/+ littermate controls infected with 1 × 10⁶ parasitized erythrocytes was compared, and miR-155^-/- animals displayed significantly better survival (****P < 0.0001; log-rank test; n = 26/group). (B) Peripheral parasitemia was greater in mice deficient in miR-155. Mean peripheral parasitemia in miR-155^-/- versus wild-type mice infected with 1 × 10⁶ parasitized erythrocytes was compared, and peripheral parasitemia was significantly higher in miR-155^-/- than wild-type mice (****P < 0.0001; linear mixed-effects model; n = 26/group). Experiment was repeated for confirmation of findings. Error bars represent standard deviation.

Figure 3.

miR-155 deficiency was associated with preservation of the blood-brain barrier (BBB) in P. berghei ANKA (PbA)-induced ECM. (A) Brains from mice deficient in miR-155 showed improved BBB integrity (less dye leakage into the brain parenchyma). Evans Blue assay was conducted on albino C57BL/6 miR-155^-/- and wild-type miR-155^+/+ littermate controls on d 7 post-PbA infection (1 × 10⁶ parasitized erythrocytes; n = 10–11/group), and representative photographs of brains from mice following Evans Blue dye assay are shown. (B) Brain microvascular leak was reduced in miR-155–deficient mice. Median levels of Evans Blue in the brain was quantified as an indication of vascular leak, and was significantly less in miR-155^-/- than wild-type mice on d 7 post-PbA infection (1 × 10⁶ parasitized erythrocytes; ****P < 0.0001; Mann-Whitney test; n = 10–11/group). (C) BBB integrity was preserved in miR-155^-/- mice despite increased peripheral parasitemia. Median peripheral parasitemia at the time of Evans Blue assay was measured in miR-155^-/- and wild-type mice (****P < 0.0001; Mann-Whitney test; n = 10–11/group). Experiment was repeated for confirmation of findings. Error bars represent interquartile range.

Figure 4.

miR-155 deficiency is associated with decreased endothelial activation during P. berghei ANKA (PbA)-induced ECM. Median plasma protein levels of endothelial activation markers in albino C57BL/6 miR-155^-/- and wild-type miR-155^+/+ littermate controls on d 6 post-PbA infection (1 × 10⁶ parasitized erythrocytes), measured by ELISA. Ang-1 was significantly higher, while Ang-2, Ang-2:Ang-1 ratio and sE-selectin were significantly lower in miR-155^-/- than wild-type PbA-infected mice (*P < 0.05, **P < 0.01, ****P < 0.0001; Welch’s t test or Mann-Whitney test, where applicable; n = 23–45/group). Experiment was repeated for confirmation of findings. Error bars represent interquartile range. Abbreviations: Ang-1, angiopoietin-1; Ang-2, angiopoietin-2; sE-selectin, soluble E-selectin.

Figure 5.

miR-155 deficiency is associated with increased inflammation during P. berghei ANKA (PbA)-induced ECM. Median plasma protein levels of inflammatory cytokine markers in albino C57BL/6 miR-155^-/- versus wild-type miR-155^+/+ littermate controls on d 6 post-PbA infection (1 × 10⁶ parasitized erythrocytes), measured by Cytometric Bead Array. IFN- γ, IL-6, MCP-1 and TNF:IL-10 ratio were all significantly higher in miR-155^-/- versus wild-type PbA-infected mice (*P < 0.05, **P < 0.01, ****P < 0.0001; Mann-Whitney test; n = 23–25/group). Experiment was repeated for confirmation of findings. Error bars represent interquartile range. Abbreviations: IFN-γ, interferon-gamma; TNF, tumor necrosis factor; IL-6, interleukin-6; MCP-1, monocyte chemotactic-1; IL-10, interleukin-10.

Figure 6.

miR-155 deficiency is associated with decreased expression of granzyme A (a marker of cytotoxic T lymphocytes and natural killer cells) and heme oxygenase-1 (a marker of oxidative stress) in response to P. berghei ANKA (PbA)-induced ECM. Median brain mRNA levels in albino C57BL/6 miR-155^-/- versus wild-type miR-155^+/+ littermate controls on d 6 post-PbA infection (1 × 10⁶ parasitized erythrocytes), measured by qRT-PCR. Gzma and Ho-1 mRNA were significantly decreased in miR-155^-/- versus wild-type PbA-infected mice (*P < 0.05; Mann-Whitney test; n = 7/group). Experiment was repeated for confirmation of findings. Error bars represent interquartile range. Abbreviations: Gzma, granzyme A; Ho-1, heme oxygenase-1.

Figure 7.

Human umbilical vein endothelial cell (HUVEC) microvessels perfused with sera from children with cerebral malaria (CM) illustrate greater dextran leak versus sera control, which can be prevented with preincubation with miR-155 antagomir. (A) Bright field image of human sera perfused through the microvessels: normal pooled sera (left panel), CM sera (middle panel) and microvessels pretreated with antagomir-155 prior to CM sera (right panel) (scale bar: 100 μm). (B). Fluorescence image of microvessels perfused with 40 kDa dextran (as a marker of microvascular leak) after 5 min of perfusion, after different sera treatments. (C) Line cut of intensity profile of dextran-perfused microvessels in B at t = 1 min (blue line) and t = 5 min (red line). The permeability K of three conditions was determined to be 0.016 ± 0.005 μm/s (normal sera), 0.26 ± 0.13 μm/s (CM sera) and 0.065 ± 0.029 μm/s (antagomir-155 + CM sera). (D) Z-stack projection of confocal fluorescence images of microvessels after serum treatment and dextran perfusion, stained with CD31 (cell junction) and Hoechst 33342 (nuclei) (scale bar: 100 μm). Bottom panels: zoomed-in view of dashed square box (scale bar: 50 μm). Confocal images illustrate no differences in PECAM1/CD31 cell junctions despite significant differences in dextran leak. The assay was repeated with a different CM sample and yielded similar results (n = 2 biological replicates).