Changes in metabolite profiles in the cerebrospinal fluid and in human neuronal cells upon tick-borne encephalitis virus infection

Abstract

Background: Tick-borne encephalitis virus (TBEV) is a significant threat to human health. The virus causes potentially fatal disease of the central nervous system (CNS), for which no treatments are available. TBEV infected individuals display a wide spectrum of neuronal disease, the determinants of which are undefined. Changes to host metabolism and virus-induced immunity have been postulated to contribute to the neuronal damage observed in infected individuals. In this study, we evaluated the cytokine, chemokine, and metabolic alterations in the cerebrospinal fluid (CSF) of symptomatic patients infected with TBEV presenting with meningitis or encephalitis. Our aim was to investigate the host immune and metabolic responses associated with specific TBEV infectious outcomes.

Methods: CSF samples of patients with meningitis (n = 27) or encephalitis (n = 25) were obtained upon consent from individuals hospitalised with confirmed TBEV infection in Brno. CSF from uninfected control patients was also collected for comparison (n = 12). A multiplex bead-based system was used to measure the levels of pro-inflammatory cytokines and chemokines. Untargeted metabolomics followed by bioinformatics and integrative omics were used to profile the levels of metabolites in the CSF. Human motor neurons (hMNs) were differentiated from induced pluripotent stem cells (iPSCs) and infected with the highly pathogenic TBEV-Hypr strain to profile the role(s) of identified metabolites during the virus lifecycle. Virus infection was quantified via plaque assay.

Results: Significant differences in proinflammatory cytokines (IFN-α2, TSLP, IL-1α, IL-1β, GM-CSF, IL-12p40, IL-15, and IL-18) and chemokines (IL-8, CCL20, and CXCL11) were detected between neurological-TBEV and control patients. A total of 32 CSF metabolites differed in TBE patients with meningitis and encephalitis. CSF S-Adenosylmethionine (SAM), Fructose 1,6-bisphosphate (FBP1) and Phosphoenolpyruvic acid (PEP) levels were 2.4-fold (range ≥ 2.3-≥3.2) higher in encephalitis patients compared to the meningitis group. CSF urocanic acid levels were significantly lower in patients with encephalitis compared to those with meningitis (p = 0.012209). Follow-up analyses showed fluctuations in the levels of O-phosphoethanolamine, succinic acid, and L-proline in the encephalitis group, and pyruvic acid in the meningitis group. TBEV-infection of hMNs increased the production of SAM, FBP1 and PEP in a time-dependent manner. Depletion of the metabolites with characterised pharmacological inhibitors led to a concentration-dependent attenuation of virus growth, validating the identified changes as key mediators of TBEV infection.

Conclusions: Our findings reveal that the neurological disease outcome of TBEV infection is associated with specific and dynamic metabolic signatures in the cerebrospinal fluid. We describe a new in vitro model for in-depth studies of TBEV-induced neuropathogenesis, in which the depletion of identified metabolites limits virus infection. Collectively, this reveals new biomarkers that can differentiate and predict TBEV-associated neurological disease. Additionally, we have identified novel therapeutic targets with the potential to significantly improve patient outcomes and deepen our understanding of TBEV pathogenesis.

Keywords: Cerebrospinal fluid; Chemokines; Human motor neurons; Metabolomics; Neuroinflammation; Pro-inflammatory cytokines; Tick-borne encephalitis virus.

Fig. 1

Demographics and reported symptoms of encephalitis (blue) and meningitis (purple) groups. Data were collected upon hospital admission with consent (n = 52 biologically independent patients). (A) No differences in biological age were observed across encephalitis and meningitis groups (ns, p = 0.675). (B) A pie chart shows the percentage of gender representation in each group, with females in green and males in orange. (C) Gender association with encephalitis (blue) and meningitis (purple). Statistical significance was determined using a two-proportion z-test (z = 1.98, *p = 0.048); ns, not significant. (D) Frequency and (E) peak hospitalisation days post-tick bite for the encephalitis and meningitis groups. (F) Non-neurological symptoms for patients in the encephalitis (blue) and meningitis (purple) groups. Proportion and peak severity of non-neurological symptoms experienced by infected patients since acquiring TBEV. Proportion and peak severity of neurological symptoms over the hospital administration period for patients in (G) encephalitis and (H) meningitis groups. Statistical significance was determined using a two-way ANOVA followed by Bonferroni multiple comparison test (*p < 0.05, **p < 0.01 ***p < 0.001, ****p < 0.0001); ns, not significant

Fig. 2

Human proinflammatory chemokine profiling in the cerebrospinal fluid of hospitalised TBEV patients and controls. Panels consisting of 13 human chemokines (IL-8, IP-10, CCL11, CCL17, CCL2, CCL5, CCL3, CXCL9, CXCL5, CCL20, CXCL1, CXCL11 and CCL4) were assayed using a multiplex bead-based system. Significance was determined using a one-way ANOVA followed by Bonferroni multiple comparison test (*p < 0.05, **p < 0.01 ***p < 0.001, ****p < 0.0001); ns, not significant. Data are the mean ± standard deviation of 3 technical repeats and 2 independent experiments

Fig. 3

Cytokine profiling in the cerebrospinal fluid of hospitalised TBEV patients and controls. Panels consisting of 13 human cytokines (TSLP, IL-1α, IL-1β, GM-CSF, IFN-α2, IL-23, IL-12p40, IL-12p70, IL-15, IL-18, IL-11, IL-27, and IL-33) were tested using a multiplex bead-based assay. Significance was determined using a one-way ANOVA followed by Bonferroni multiple comparison test (*p < 0.05, **p < 0.01 ***p < 0.001, ****p < 0.0001); ns, not significant. Data are the mean ± standard deviation of 3 technical repeats and 2 independent experiments

Fig. 4

Metabolomic profiles in cerebrospinal fluid samples from control, encephalitis and meningitis TBEV patients. (A) 2D PLS-DA plots with 95% confidence ellipses showing the effects of encephalitis and meningitis on the CSF metabolome. The metabolites contributing to the separation of samples are labelled with arrows indicating their direction and magnitude. Cross-validated 2D PLS-DA score plot comparing TBEV patients with (B) encephalitis (with follow-up) and control, or TBEV patients with (C) meningitis, follow-up, and control groups on the CSF metabolome. Volcano plots comparing a two-fold change in concentration of CSF metabolites in (D) encephalitis vs. meningitis, or (E) control vs. encephalitis or (F) meningitis patients. Significantly altered metabolites are highlighted in red (increased) or blue (decreased). Data were compared using a two-sided Mann-Whitney U test followed by Benjamini-Hochberg multiple comparison test with FDR < 0.05 and fold-change > 1. (G) Heatmap showing CSF metabolites of significantly higher (red) or lower (blue) abundance in TBEV-patients with encephalitis vs. meningitis. Metabolite alterations are represented by colour intensity. Borders are colour-coded according to statistical significance (p < 0.05)

Fig. 5

Changes in the cerebrospinal fluid metabolomes in encephalitis and meningitis TBEV. (A) Area under the curve (AUC) in the top panel and box plot representation in the lower panel, showing cut-off values of 4 CSF biomarkers with significant up-regulation (S-Adenosylmethionine: AUC = 0.818, p = 0.0016348; 1-Methylnicotinamide: AUC = 0.78, p = 0.0006835; Phosphoenolpyruvic acid: AUC = 0.759, p = 0.0031707; Fructose 1,6-bisphosphate: AUC = 0.755, p = 0.012209) identified in encephalitis vs. meningitis. Statistical significance was determined using a one-way ANOVA followed by Bonferroni multiple comparison test (*p < 0.05, **p < 0.01 ***p < 0.001, ****p < 0.0001); ns, not significant. (B) KEGG metabolic pathway enrichment analysis and (C) scatter plot of pathway impact versus pathway significance, where each node represents a significant metabolic pathway contributing to the differences between encephalitis and meningitis TBEV-patient groups. The white to red node colour scale indicates the p-value, and node size reflects the pathway impact score. Red circles indicate the most highly affected metabolic pathways in TBEV patients, with the highest number of metabolite hits (9) from the pathway analysis (Impact score = 0.65877). (D) Network diagrams of metabolites in the CSF from encephalitis and meningitis patients. (E) Magnification of a new cluster of the PPP visualized from (D) in Cytoscape

Fig. 6

Identified metabolites impact TBEV infection in human motor neurons (hMNs). (A) Schematic showing neuronal differentiation achieved from human iPSCs with a combination of CHIR99021, an activator of Wnt signaling and purmorphamine a SHH activator. (B) Validation of pluripotency by immunostaining of Nanog and Oct. Cells were counterstained with DAPI. (C) qRT-PCR analysis of Nango, Sox2, Oct, cMyc and klf4 expression in human iPSCs. (D) Validation of hMNs generated from (A-C). Expression of HB9 and Tuj-1 quantified by immunofluorescence. (E) Fluorescent intensities are shown for Tuj-1 and HB9 from four technical repeats in two independent experiments. (F) Confirmation of TBEV infection at an MOI of 0.1 in hMNs. Cells were stained with anti-E proteins (red) and Tuj-1 (green) at 24- and 96-h post-infection (hpi). (G) Quantification of the % of infected cells and (H) cell viability studies. (I) Bar graphs of soluble SAM, (J) FBP1 and (K) PEP levels in infected neurones at 24- and 96-hpi. Infection of hMNs with TBEV at 1 MOI, followed by ELISA. (L) Schematic of the assay design. (M) hMNs were treated with increasing concentrations of Sinefungin, MB05032 [0.0, 0.1, and 1.0 µM], NaF [0.0, 0.1, and 1.0 mM]) for 24 h and infected with TBEV at an MOI of 0.1 PFU, plaque-forming units. Viral titres at 72 hpi. were determined via plaque assay. (N) For viability studies, hMNs were treated with inhibitors and infected with TBEV at an MOI of 1 PFU. Statistical significance was determined using a one-way ANOVA followed by Bonferroni multiple comparison test (*p < 0.05, **p < 0.01 ***p < 0.001, ****p < 0.0001); ns, not significant. Data are the mean ± standard deviation of 3 technical repeats in 3 independent experiments. A and L created using BioRender.com