Brain endothelial cells exposure to malaria parasites links type I interferon signalling to antigen presentation, immunoproteasome activation, endothelium disruption, and cellular metabolism

Abstract

Introduction: Cerebral malaria (CM) lethality is attributable to induction of brain edema induction but the cellular mechanisms involving brain microvascular endothelium in CM pathogenesis are unexplored.

Results: Activation of the STING-INFb-CXCL10 axis in brain endothelial cells (BECs) is a prominent component of the innate immune response in CM development in mouse models. Using a T cell-reporter system, we show that Type 1 IFN signaling in BECs exposed to Plasmodium berghei-infected erythrocytes (PbA-IE), functionally enhances MHC Class-I antigen presentation through gamma-interferon independent immunoproteasome activation and impacted the proteome functionally related to vesicle trafficking, protein processing/folding and antigen presentation. In vitro assays showed that Type 1 IFN signaling and immunoproteasome activation are also involved in the dysfunction of the endothelial barrier through disturbing gene expression in the Wnt/ß-catenin signaling pathway. We demonstrate that IE exposure induces a substantial increase in BECs glucose uptake while glycolysis blockade abrogates INFb secretion impairing immunoproteasome activation, antigen presentation and Wnt/ß-catenin signaling.

Discussion: Metabolome analysis show that energy demand and production are markedly increased in BECs exposed to IE as revealed by enriched content in glucose and amino acid catabolites. In accordance, glycolysis blockade in vivo delayed the clinical onset of CM in mice. Together the results show that increase in glucose uptake upon IE exposure licenses Type 1 IFN signaling and subsequent immunoproteasome activation contributing to enhanced antigen presentation and impairment of endothelial barrier function. This work raises the hypothesis that Type 1 IFN signaling-immunoproteasome induction in BECs contributes to CM pathology and fatality (1) by increasing antigen presentation to cytotoxic CD8+ T cells and (2) by promoting endothelial barrier dysfunction, that likely favor brain vasogenic edema.

Keywords: antigen presentation; cerebral malaria (CM); endothelial cells; glucose metabolic alterations; malaria; metabolome; proteome; type - 1 interferons.

Figure 1

IFNAR1 signaling is a determinant of immunoproteasome-dependent antigen presentation in brain endothelial cells exposed to PbA-IE. (A) Diagram of in vitro antigen presentation assay depicting primary BEC cultured in presence of interferon-gamma (24h), exposed to synchronized PbA-IE (24h) and incubated with PbA antigen-specific reporter T-cell line that are subsequently stained to count antigen presentation events. (B) Antigen presentation events counts in BEC cultures from wild-type IFNAR1 and IFNB KO mice. (C) Results of two independent experiments counting antigen presentation events in BEC cultures from wild-type and IFNB KO mice in the presence or absence of IFNβ supplementation (80-120 µg/ml). (D) Results of two independent experiments counting antigen presentation events after exposing wild-type BEC cultures to PbA-IE or to reporter cell cognate peptide 10 µM with no supplementation with IFN-γ and in the presence or absence of immunoproteasome inhibitor ONX-0914 (300 nM)(ONX). Kinetics of Psbm8 (E), Psbm9 (F), and Psbm10 (G) gene expression quantified by qPCR (fold change represents 2^ΔΔCT relative to wild-type at 0h) and showing induction in BECS of wild-type, IFNAR1 KO and IFNB KO mice exposed to PbA-IE, in absence of IFN-γ. Statistics: P-value of ANOVA tests of antigen presentation results (black stars) or ANOVA tests of AUC in gene expression induction in IFNAR1 vs wild-type (red stars) and IFNB versus wild-type (brown stars). (*; p<0.05, **; p<0.01, ****; p <0.0001).

Figure 2

IFNAR1 signaling impact the proteome and MHC class I expression in BECs exposed to PbA-IE. Proteome discrimination analysis of BECs cultures from IFNAR1KO and wild-type mice exposed to PbA-IE. Principal component analysis score plot (PC1versus PC2) (A) and orthogonal partial least squares discriminant analysis (OPLS-DA) score plot (B) of all detected protein features in BECs samples (wild-type in green, IFNAR1 KO in red); explained variance is shown in brackets in axes titles. (C) Volcano plot highlighting differentially expressed protein features and showing fold change in IFNAR KO relative to wild-type (FC) log(2)FC > 1.2 (log2(FC) > 0.18, red; log2(FC) < - 0.18, blue) and t-test significance level, P<0.05. Differentially expressed protein features are identified by the respective SwissProt database reference. Proteome analysis is based on three independent samples per genotype, each sample representing BECs obtained from a pool of five mouse brains. (D) Flow cytometry histogram plots of surface staining of MHC class I molecules in BEC of IFNB KO and wild-type mice, cultured in the absence or presence of PbA-IE (24h). Results are representative of 2 independent experiments.

Figure 3

IFNAR signaling a lead to disruption of the brain endothelial barrier through immunoproteosome-dependent Wnt/ß-catenin signaling disturbances. Endothelial barrier integrity was monitored by electric resistance measurements of transwell BEC cultures (TEER) at the indicated time points after exposure or not to PbA-IE (A, B). TEER kinetics in wild-type, IFNAR1 KO BECs (A), or in wild-type BECs in the presence or absence of (GSK) 3 inhibitor CHIR-99021(1 µM) (B). Gene expression of ß-catenin target genes Nkdd1, Axin2, and Apcdd1 was quantified by qPCR in wild-type (C–E) and IFNAR1 KO BECs (F–H) 24 h after exposure to PbA-IE in presence or absence of immunoproteasome inhibitor ONX-0914 (ONX) (300 nM) (I–K). Relative quantification of gene expression is represented as fold change (2^ΔΔCT) using unexposed BECs as controls. Statistics: Significant results of ANOVA tests of AUC TEER in wild-type BECs exposed to PbA-IE versus unexposed (black stars) and in IFNAR1 KO BECs exposed versus unexposed (red stars). Significant results of pairwise comparisons in ANOVA tests in gene expression experiments are shown (*; p<0.05, **; p<0.01).

Figure 4

Glucose metabolism inhibition impairs IFN-ß secretion, immunoproteasome induction and antigen presentation in brain endothelial cells upon contact with PbA-IE. Effects of glycolysis inhibition in BEC cultures in presence 2-DG (10 mM) measured 24 h after exposure to PbA-IE. (A) Quantification of secreted IFN-ß in culture supernatants of BECs exposed or not to PbA-IE in the presence or absence of 2-DG (lower limit of quantitation 0.94 pg/mL). (B–D) Gene expression of immunoproteasome genes Psmb8, Psmb9, and Psmb10 in BECs in the indicated conditions. Cumulative results of independent experiments are represented as fold change (2^ΔΔCT) using unexposed BECs as controls. Antigen presentation assays in BEC cultures unexposed and exposed to PbA-IE in vitro (E) and in BECs from uninfected and infected mice (F), treated or not with 2-DG. Significant results of pairwise comparisons in ANOVA tests are shown (*; p<0.05, **; p<0.01, ***; p<0.001, ****; p <0.0001).

Figure 5

Glucose uptake and metabolism in brain endothelial cells exposed to PbA-IE. Measurement of glucose in supernatants of (A) BECs cultures after 24 h of exposure to PbA-IE in the presence or absence of 2-DG (10 mM) and (B) in cultures from wild-type, IFNAR1 KO and IFNB KO BECs in presence of 2-DG. (C) Multiple pathway targeted analysis showing enrichment ratio (P-value) of metabolic pathways in wild-type BECs exposed to PbA-IE as compared to unexposed BECs (MetaboAnalyst 5.0). Significant results of pairwise comparisons in ANOVA tests are shown (***; p<0.001, ****; p <0.0001).

Figure 6

2-DG treatment increases the survival of infected mice. Clinical scoring (A) and survival curves (B) of wild-type mice infected with PbA-IE and treated (2-DG) or not (PBS) with 2-DG (800 mg/kg) at day 4 post-infection. Log-rank (Mantel-Cox) test (D) (****; p <0.0001).