Parkinson's disease-derived α-synuclein assemblies combined with chronic-type inflammatory cues promote a neurotoxic microglial phenotype

Abstract

Parkinson's disease (PD) is a common age-related neurodegenerative disorder characterized by the aggregation of α-Synuclein (αSYN) building up intraneuronal inclusions termed Lewy pathology. Mounting evidence suggests that neuron-released αSYN aggregates could be central to microglial activation, which in turn mounts and orchestrates neuroinflammatory processes potentially harmful to neurons. Therefore, understanding the mechanisms that drive microglial cell activation, polarization and function in PD might have important therapeutic implications. Here, using primary microglia, we investigated the inflammatory potential of pure αSYN fibrils derived from PD patients. We further explored and characterized microglial cell responses to a chronic-type inflammatory stimulation combining PD patient-derived αSYN fibrils (F^PD), Tumor necrosis factor-α (TNFα) and prostaglandin E₂ (PGE₂) (TPF^PD). We showed that F^PD hold stronger inflammatory potency than pure αSYN fibrils generated de novo. When combined with TNFα and PGE₂, F^PD polarizes microglia toward a particular functional phenotype departing from F^PD-treated cells and featuring lower inflammatory cytokine and higher glutamate release. Whereas metabolomic studies showed that TPF^PD-exposed microglia were closely related to classically activated M1 proinflammatory cells, notably with similar tricarboxylic acid cycle disruption, transcriptomic analysis revealed that TPF^PD-activated microglia assume a unique molecular signature highlighting upregulation of genes involved in glutathione and iron metabolisms. In particular, TPF^PD-specific upregulation of Slc7a11 (which encodes the cystine-glutamate antiporter xCT) was consistent with the increased glutamate response and cytotoxic activity of these cells toward midbrain dopaminergic neurons in vitro. Together, these data further extend the structure-pathological relationship of αSYN fibrillar polymorphs to their innate immune properties and demonstrate that PD-derived αSYN fibrils, TNFα and PGE₂ act in concert to drive microglial cell activation toward a specific and highly neurotoxic chronic-type inflammatory phenotype characterized by robust glutamate release and iron retention.

Keywords: Chronic inflammation; Cystine/glutamate transporter; Iron metabolism; Metabolomic; Microglia; Neurotoxicity; Parkinson’s disease; Protein misfolding cyclic amplification; Transcriptomic; α-Synuclein.

Fig. 1

aSYN fibrils derived from PD patients hold more potent inflammatory properties than de novo-generated counterparts. A Representative images of CD11b immunostaining (grey) and Hoechst nuclear staining (blue) showing the morphological changes in microglial cells following inflammatory stimulation by the prototypal inflammogen LPS (10 ng/mL), de novo assembled human αSYN fibrils (3 µM; F^S) and fibrils derived from PD patients (3 µM; F^PD) compared to nonstimulated cells (NCS). Note that inflammatory stimulation results in cellular flattening and shortening of cell processes under all three inflammatory conditions. Scale bar: 50 µm. B Dose–response analysis of TNFα, IL-6, IL-10 and glutamate release by microglial cells exposed or not (NSC) to increasing concentrations (from 0.01 to 3 µM) of αSYN fibrils derived from PD patients (F^PD). The data are presented as the means ± SEM (n = 6). *p < 0.05; **p < 0.01 vs. NSC (Tukey’s test). C Quantification of TNFα, IL-6, IL-10 and glutamate release by microglial cells exposed or not (NSC) to similar concentrations of F^S, F^PD or F^DLB (3 µM). The data are presented as the means ± SEM (n = 6). *p < 0.05; **p < 0.01; ***p < 0.001 vs. F^S or F^PD (Tukey’s test)

Fig. 2

Regulation of human macrophage-associated TPP-specific inflammatory markers in microglial cells. A Representative images showing CD25 immunostaining (green) with Hoechst nuclear stain (blue) in nonstimulated microglial cells (NSC) and cells exposed to IL4 (10 ng/mL), IFNγ (20 ng/mL) or a combination of TNFα (17 ng/mL), PGE₂ (1 µg/mL) and Pam3CSK4 (1 µg/mL) (TPP stimulation). Scale bar: 60 µm. B Quantification of CXCL5 and IL1α release by microglial cells exposed or not (NSC) to the indicated treatments. Data are means ± SEM (n = 4). *p < 0.05 vs. TPP (Tukey’s test). C Quantification of ROS production measured via the NBT reaction in microglial cells exposed or not (NSC) to the indicated treatments. The data are expressed as fold increase relative to nonstimulated cells (NSC) and are represented as the means ± SEM (n = 4). *p < 0.05 vs. TPP (Tukey’s test)

Fig. 3

TPF^PD-related chronic-type inflammatory stimulation of microglia is associated with lower cytokine but increased glutamate release. A Representative images of CD11b immunostaining (grey) and Hoechst nuclear staining (blue) of microglial cells following inflammatory stimulation by PD patient-derived αSYN fibrils (1.5 µM; F^PD) or TPF^PD as compared to nonstimulated cells (NSC). Note that F^PD and TPF^PD induce similar morphological changes of microglial cells. Scale bar: 50 µm. B Quantification of TNFα, IL-6, IL-10 and IL-1β release by microglial cells exposed or not (NSC) to TNFα (17 ng/mL), PGE₂ (1 µg/mL), TNFα + PGE₂ (TP), F^PD (1.5 µM), TNFα + PGE₂ + F^PD (TPF^PD) or LPS (10 ng/mL). The data are represented as the means ± SEM (n = 3–6). *p < 0.05; **p < 0.01; ***p < 0.001 (Tukey’s test). C Quantification of ROS production measured by the NBT reaction in microglial cells exposed or not (NSC) to F^PD (1.5 µM), TPF^PD or LPS (10 ng/mL). The data are expressed as a % of the NSC control. The bars are the means ± SEM (n = 6). *p < 0.05 vs. NSC (Student’s t-test). D Quantification of glutamate release by microglial cells exposed or not (NSC) to F^PD (1.5 µM), TPF^PD or LPS (10 ng/mL). The data are expressed as a % of the NSC control. The bars are the means ± SEM (n = 3–6). *p < 0.05 vs. TPF^PD (Tukey’s test)

Fig. 4

Specific transcriptional reprogramming of microglial cells upon TPF^PD chronic-type inflammatory activation. A Principal component analysis (PCA) of gene expression data based on the first two PCs shows, on the one hand, a clear distinction between nonstimulated (NSC) and activated cells (LPS and TPF^PD) and, on the other hand, strong differences between M1-type (LPS) and chronic-type (TPF^PD) inflammatory activation. B Hierarchical clustering and heatmap of differentially expressed genes (DEGs; n = 413) in TPF^PD-activated microglia versus nonstimulated (NSC) and LPS-treated cells. The scaled expression value (Z-score transformed) is shown in a blue-red color scheme with red indicating higher expression, and blue lower expression. Biological replicates are indicated in brackets. C PCA analysis of pathway-related data generated by the Pathifier method [55] demonstrates clear clustering and separation of stimulated and nonstimulated (NSC) cells and of M1-type (LPS) and chronic-type (TPF^PD) inflammatory activation. D Bubble chart showing the enrichment of the KEGG Pathway in TPF^PD-treated microglial cells (adjusted p < 0.05). The bubble size indicates the number of genes annotated in the indicated KEGG pathway. The colors represent pathway enrichment (% of overlapping genes) in TPF^PD-exposed cells. E Volcano plot depicting individual DEGs (|log₂ Fold Change| > 2) in TPF^PD-treated microglia versus M1-type (LPS) activated cells. The red and blue dots show upregulated and downregulated genes in TPF^PD-treated cells, respectively. Individual genes of interest are indicated. F Fold change in individual gene expression level (qPCR) of LPS-activated and TPF^PD-exposed microglial cells relative to nonstimulated cells (NSC). The data are represented as the means ± SEM (n = 5 biological replicates from independent experiment). *p < 0.05 vs TPF^PD (Student’s t test)

Fig. 5

Microglial metabolic reprogramming evoked by TPF^PD is closely related to M1-type polarized cells. A Hierarchical clustering analysis and heatmap visualization of 248 differentially regulated metabolites (28 annotated and 220 unknown/nonannotated metabolites) in stimulated cells (LPS and TPF^PD) compared to those in the NSC control group. The number of biological replicates is indicated in brackets. Clear changes in the metabolic profile between stimulated and nonstimulated cells are observed. B Hierarchical clustering analysis and heatmap visualization of the 28 annotated metabolites extracted from A. Despite these subtle differences, the TPF^PD- and LPS-associated metabolic signatures are closely related. C Box plots showing the relative amount of α-ketoglutarate, succinate, fumarate, glutamate and glutathione disulfide (GSSG) in control (NSC) and in LPS- and TPF^PD-treated microglial cells. Similar decreases in the α-ketoglutarate/succinate and fumarate/succinate ratios were found in LPS- and TPF^PD-stimulated cells. Likewise, similar increases in glutamate and GSSG levels were observed under both stimulatory conditions. *p < 0.05 (Kruskal–Wallis and Wilcoxon test)

Fig. 6

Compared with mouse microglia, TPF^PD-exposed human microglial-like cells exhibit both conserved and distinct gene regulation. A Representative images showing CCR2 (red) and CX3CR1 (green) double immunostaining with Hoechst nuclear stain (blue) in nondifferentiated (ND) and GM-CSF/IL34-differentiated human monocytes (iMGs) after 16 days of culture. The right panel shows the quantification of the immunosignal CX3CR1/CRR2 ratio in ND and GM-CSF/IL34-differentiated cells. The data are presented as the means ± SEM (n = 4–5). **p < 0.01 vs. ND (Student’s t-test). Scale bar: 40 µm. B Fold change in individual gene expression level (qPCR) of LPS-, F^PD- and TPF^PD-activated iMGs relative to nonstimulated (NSC) cells. The data are presented as the means ± SEM (biological replicates n = 3). *p < 0.05 vs. NSC (Student’s t test)

Fig. 7

TPF^PD-stimulated microglial cells trigger more excitotoxic cell death than F^PD-exposed cells on dopaminergic neurons. A Schematic representation of the experimental setup used to analyze the neurotoxic potential of chronic-type inflammatory microglia. Microglial cell cultures were maintained in DMEM^S-based astrocyte-conditioned medium (ACM) and exposed or not (NSC) for 6 h to either F^PD or TPF^PD. Following inflammatory stimulation, the culture medium was fully replaced with fresh DMEM^S-based ACM (washout step), and the microglial cells were left for 24 h before the glutamate assay (B). Midbrain neuronal cultures were prepared from E13.5 mouse embryos and maintained for 2 days in Nb^S. On day in vitro 2 (DIV2), the culture medium was fully replaced with Nb^S-based ACM and the neurons were left to mature until DIV6. At DIV6, 30% (v/v) of the Nb^S-based ACM media was replaced with microglia-conditioned media (MCM) and dopaminergic neuron survival was assayed 24 h later. The data are presented as the means ± SEM (biological replicates n = 13; 3 independent experiments). ***p < 0.001 vs NSC and F^PD (One-way ANOVA followed by Tukey’s test). C Low- and high-magnification (dashed brown squares) images of tyrosine hydroxylase (TH) immunostained DIV7 midbrain cultures exposed for 24 h to MCM from F^PD and TPF^PD-stimulated or unstimulated cells (NSC). Brown arrows point to dystrophic cell bodies and processes with varicosities. Scale Bar = 100 µm. D Quantification of TH+ neurons in DIV7 midbrain cultures exposed for 24 h to MCM from F^PD and TPF^PD-stimulated or unstimulated cells (NSC). In sister cultures, neurons were exposed to MCM from F^PD and TPF^PD-stimulated microglial cells treated with the xCT inhibitor Sulfasalazine (250 µM) during the 24 h washout period. In additional sister cultures, neurons exposed to MCM from F^PD and TPF^PD-stimulated microglial cells were cotreated with the N-methyl-D-aspartate (NMDA) receptor pore blocker MK-801 (2 µM). The data are expressed as a % of neurons grown in Nb^S only and are presented as the means ± SEM (biological replicates n = 5–8; 2 independent experiments). *** p < 0.001 vs. NSC or F^PD (One-way ANOVA followed by pairwise multiple comparisons using the Holm-Sidak method). E TH+ dopaminergic neuron survival is inversely correlated with the concentration of glutamate within the transferred MCM (linear regression analysis)