Microglia jointly degrade fibrillar alpha-synuclein cargo by distribution through tunneling nanotubes

Abstract

Microglia are the CNS resident immune cells that react to misfolded proteins through pattern recognition receptor ligation and activation of inflammatory pathways. Here, we studied how microglia handle and cope with α-synuclein (α-syn) fibrils and their clearance. We found that microglia exposed to α-syn establish a cellular network through the formation of F-actin-dependent intercellular connections, which transfer α-syn from overloaded microglia to neighboring naive microglia where the α-syn cargo got rapidly and effectively degraded. Lowering the α-syn burden attenuated the inflammatory profile of microglia and improved their survival. This degradation strategy was compromised in cells carrying the LRRK2 G2019S mutation. We confirmed the intercellular transfer of α-syn assemblies in microglia using organotypic slice cultures, 2-photon microscopy, and neuropathology of patients. Together, these data identify a mechanism by which microglia create an "on-demand" functional network in order to improve pathogenic α-syn clearance.

Keywords: LRRK2; alpha-synuclein; cell-to-cell transfer; clearance; degradation; microglia; synucleinopathies; tunneling nanotubes.

Graphical abstract

Figure S1

Characterization of α-syn fibrils and its transfer between microglia, related to Figures 1, 2, and 3 (A-B) Characterization of α-syn fibrils used throughout this study. Electron micrographs of α-syn fibrils stained by Uranyl acetate, before (A) and after (B) fragmentation. (C) Sedimentation coefficient of fragmented fibrils determined by analytical ultracentrifugation. The sedimentation velocity measurement shows a distribution centered at 110 Svedberg. The sedimentation coefficient is compatible with a molecular species of 16,000 kDa corresponding to ∼1,100 monomers of 14.46 kDa. (D) Proteolytic profile of fibrillar α-syn. Time course of fibrillar α-syn (100 μM) degradation by proteinase K (0.38 μg/mL) analyzed by SDS-PAGE after Coomassie blue staining. (E) All α-syn preparations were confirmed to have an endotoxin concentration below 0.02 endotoxin units/μg (EU/μg). n = 7 independent α-syn preparations. (F) Quantification of the percentage of α-syn monomers containing microglia (left) and the relative individual uptake index per cell (middle) after exposure to ATTO488-labeled α-syn monomers, 2 μM; n = 3 independent experiments per group. Diagram representing the α-syn monomers uptake as measured by FACS (right). (G) Representative immunostaining showing the internalization of ATTO488-labeled α-syn monomers into CD11b-labeled microglia. (H) Representative time-lapse recordings demonstrating that α-syn fibrils are transferred from one microglia to another via thin cellular membrane connections. (I) Representative time-lapse recording demonstrating that α-syn fibrils are transferred from overloaded microglia to naive microglia via cellular connections. (J) Representative particle tracking of aggregates transferred from donors to acceptors as shown in (I) (upper panels). Quantification of the directionality of transferred particles. D = donors, A = acceptors. A total of 37 particle transfer events were analyzed. (K) Quantification of particles that underwent transfer from α-syn-containing microglia toward naive cells for their size, traveling distance, total particle transfer time, and particle transfer velocity. n = 33 individual particles. (L) Quantification of the number of individual cell neighbors and proportion of cells involved in a network before and after α-syn fibrils uptake. Network formation was analyzed using a CellProfiler script, identifying individual cells and measuring the number of adjacent cells. A total of at least 205 cells per condition were analyzed. n = 5-6 individual experiments. Graphs in F are presented as mean ± SEM and were analyzed by one-way ANOVA followed by Tukey’s multiple comparison post hoc test. Graphs in K present individual particles and the mean. Graphs in L were analyzed by t test analysis. ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗p < 0.05 compared to 0 min. Scale bars: 100 nm (A-B), 20 μm (G, I), 10 μm (H).

Figure S2

Cell-to-cell contact favors fibrillar α-syn transfer between microglia and induces cytoskeletal changes, related to Figure 3 (A) Dose-response curve analysis for α-syn transfer capacities from donors to acceptors at increasing concentrations (0.1 – 1 μM). n = 4 with duplicate treatments for all conditions. (B) Quantification of the ROS release of donor (right) and acceptor (left) cells with changing donor: acceptor ratios. (C) Schematic drawing of staining strategy and quantification of the rate of acceptors engulfing dying donors (CellTracer labeled). The schematic was created using BioRender.com and Adobe Photoshop. (D) Quantification of the transfer rate of fibrillar Amyloid-β between microglia. n = 3 with triplicate treatments for all conditions. (E) Quantification of the transfer rate of fibrillar Tau between microglia. n = 3 with duplicate treatments for all conditions. (F) Donors (CellTracer negative) and acceptors (CellTracer positive, blue) were co-cultured for 5 h and immunocytochemical analysis for Connexin 43 (Cx43) were performed. (G) Donors and acceptors were co-cultured for the indicated time and the total length of the F-actin cytoskeleton of acceptors was measured. (H–J) Quantification of the number of trunks (H), branches (I) and the mean trunk to branch end distance (J) of acceptor microglia over time. n = 4 per group. A total of at least 185-400 cells were analyzed. (K) Donors and acceptors were co-cultured for the indicated time and the total length of the F-actin cytoskeleton of donors was measured. (L–N) Quantification of the number of trunks (L), branches (M) and the mean trunk to branch end distance (N) of donor microglia over time. n = 4 per group. A total of at least 185-400 cells were analyzed. Graphs in A-E are presented as mean ± SEM and were analyzed by one-way ANOVA followed by Tukey’s multiple comparison post hoc test. Graphs in G-N are presented as violin plots and were analyzed by one-way ANOVA followed by Tukey’s multiple comparison post hoc test. ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗∗p < 0.01 compared to 0 h. Scale bars: 5 μm.

Figure S3

Effects of the cytoskeleton on α-syn transmission and transcriptomic analysis of microglia, related to Figures 3 and 4 (A) Quantification of the number of neighboring cells per individual cell and the percentage of cells being integrated into a cellular network. Network formation was analyzed using a CellProfiler script, identifying individual cells and measuring the number of adjacent cells. A total of at least 815 cells per condition were analyzed. n = 3 with five replicates per group. (B) Quantification of the mean ROS fluorescence of donors (left) and acceptors (right) that were co-cultured for 5 h and treated with Y-27632, Blebbistatin, and CytD, respectively. n = 3 independent experiments. (C) Quantification of the mean SYTOX fluorescence of donors (left) and acceptors (right) that were co-cultured for 5 h and treated with Y-27632, Blebbistatin, and CytD, respectively. n = 3 independent experiments. (D) STRING protein interaction network for 35 proteins associated to cell-cell adhesion based on the differential gene expression between donors and acceptors. Proteins with highest fold changes in expression levels are highlighted in bold. From the top 10 expressed genes, Sirpb1c is excluded as it was not connected to the network. A maximum of 10 interactors was allowed. Colors represent the membership to clusters based on k-means clustering. (E) Vulcano plot of genes that were differentially regulated when donors were co-cultured with acceptors with direct cell-cell contact. (F) Vulcano plot of genes that were differentially regulated when donors were co-cultured with acceptors without direct cell-cell contact. (G) Bar chart of most enriched pathways for aggregated α-syn induced (red) and suppressed (blue) genes in donors using the transwell insert strategy. Graphs are presented as mean ± SEM and were analyzed by one-way ANOVA followed by Tukey’s multiple comparison post hoc test (A right, B) or by one-way ANOVA followed by Dunn’s multiple comparison post hoc test (A left). ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗∗p < 0.01, ^∗p < 0.05 compared to 0 h.

Figure 1

Uptake of α-syn fibrils results in the induction of an inflammatory profile (A) Quantification of the percentage of phagocytic cells (left) and the individual uptake index per cell (middle) after exposure to fluorescent α-syn fibrils (2 μM); n = 4. Diagram represents the α-syn uptake as measured by FACS (right). (B) Representative immunostaining showing the internalization of α-syn fibrils into CD11b⁺ microglia. (C) Heatmap of 2189 differentially expressed (DE) genes between control and α-syn-treated microglia. (D) Top 15 DE genes belonging to the α-syn signature identified by (Sarkar et al., 2020) in murine microglia plotted as Z-score transformed heatmap of gene expression values. (E) Bar chart of most enriched pathways for aggregated α-syn induced (red) and suppressed (blue) genes. (F) BiNGO enrichment map for DE genes between control and α-syn-treated microglia. Clusters were defined by the Cytoscape tool Wordcloud. FC, fold change; ES, enrichment score. Graphs represent the mean ± SEM and were analyzed by one-way ANOVA followed by Tukey’s multiple comparison post hoc test. ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗∗p < 0.01, ^∗p < 0.05. Scale bar: 20 μm. See also Figure S1; Table S1.

Figure 2

Microglia hesitate to degrade fibrillar α-syn and form a cellular network (A) Heatmap of Z-score transformed gene expression values for DE transcripts between α-syn treated microglia and controls related to the GO term “response to unfolded protein.” (B) Representative immunostaining of F-actin⁺ microglia before and after α-syn fibrils degradation. (C) Quantification of the number of cells (left) and the individual uptake index per cell (right) in microglia after α-syn phagocytosis (15 min) and degradation (24 h). n = 4. (D) Representative chart of fibrillar α-syn phagocytosis and degradation measured by FACS. (E) Immunoblot analysis and quantification of microglial lysates after fibrillar α-syn phagocytosis (15 min) and degradation (24 h). n = 4. (F) Representative immunostaining of microglia demonstrating various cellular F-actin⁺ connections containing α-syn. (G) Representative Electron Microscopy (FIB-SEM) images of membrane-to-membrane contacts of microglia. (H) Representative time-lapse recording demonstrating the transfer of α-syn aggregates from one microglia to another. Graphs represent the mean ± SEM and were analyzed by t test (E) or one-way ANOVA followed by Tukey’s multiple comparison post hoc test (C). ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, Scale bars: 20 μm. See also Figure S2.

Figure 3

α-syn exchange between microglia is mediated by F-actin (A) Quantification of the number of α-syn-positive acceptors over time in co-culture with donors. n = 4. (B) Quantification of the number (left) and size (right) of α-syn aggregates in donors over time. n = 5. (C) Schematic drawing and quantification of α-syn degradation capacity of donors cultured alone (1), in co-culture with acceptors (2), or in co-culture with acceptors in a transwell (3). (D) Schematic depicting the α-syn transfer from donors to acceptors via tunneling nanotube-like structures and gap junctions. The drawing was created using BioRender.com and Adobe Illustrator. (E) Quantification of the number of donor-to-acceptor connections. n = 4 with 185–400 individual cells. (F) Schematic drawing of ROCK signaling and its downstream modulation of the F-actin cytoskeleton via LIM kinase (LIMK) and cofilin dephosphorylation, and the Myosin II actions via phosphorylation of the myosin light chain phosphatase (MLCP). The pharmacological inhibitors Y-27632, Blebbistatin, and Cytochalasin D (CytD) were used to block the downstream effects at different checkpoints. (G) Quantification of the effect of Y-27632 (10 μM) on α-syn transfer (left) and number of α-syn positive acceptors (right) at 5 h of co-culture. n = 4 with duplicate treatments. (H) Quantification of the effect of Blebbistatin (50 μM) on α-syn transfer (left) and number of α-syn positive acceptors (right) at 5 h of co-culture. n = 3 with duplicate treatments. (I) Quantification of the effect of CytD (5 μM) on α-syn transfer (left) and number of α-syn positive acceptors (right) at 5 h of co-culture. n = 3 with duplicate treatments. (J) Representative immunostaining revealing the presence of non-muscle Myosin II and F-actin inside cell-to-cell connections. (K) Schematic drawing of the generation of ROCK1- and ROCK2-knockout (Δ/Δ) microglia from ROCK1^flox/flox and ROCK2^flox/flox cells using a tat-Cre recombinase. (L) Quantification of the effect of ROCK1^Δ/Δ and ROCK2^Δ/Δ on α-syn transfer (left) and number of α-syn positive acceptors (right) at 5 h of co-culture. n = 3 with duplicate treatments. Graphs represent the mean ± SEM and were analyzed by one-way ANOVA followed by Kruskal-Wallis multiple comparison post hoc test for nonparametric data (A–E, L) or by a two-tailed t test (G–I). ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗∗p < 0.01, ^∗p < 0.05. Scale bars: 20 μm. See also Figures S2 and S3.

Figure S4

Transcriptomic analysis of microglia co-cultured in transwell inserts, related to Figure 4 (A) Enrichment Scores for selected GO terms in donors’ transcriptomes over the time of co-culture with acceptors using the transwell insert strategy to prevent direct cell-cell contact. Green line indicates the baseline ES at 0 min. (B) Quantification of the mean ROS fluorescence and percentage of ROS-positive cells of donors (green) and acceptors (blue) that were co-cultured for 5 h using the transwell insert strategy. n = 3 independent experiments. (C) Quantification of the α-syn transfer index from donors to acceptors (left) and the percentage of acceptors containing α-syn (right) after 5 h of co-culture upon treatment with the ROS scavenger N-Acetylcystein (NAC) and hydrogen peroxide (H₂O₂). n = 4 independent experiments with duplicated measurements. (D) Schematic illustrating the co-culture strategy used for experimental results presented in (E) and (F). The schematic was created using BioRender.com and Adobe Illustrator. (E and F) FACS analysis (E) and quantification (F) of the bidirectional transport of mitochondria from donors to acceptors and vice versa. n = 3 independent experiments. (G) Heatmap of Z-score transformed gene expression values for DE transcripts between α-syn treated microglia (“donors”) and control cells (“acceptors”) related to the GO term “intrinsic apoptotic signaling pathway“. All graphs are presented as mean ± SEM and were analyzed by two-way ANOVA (B) or one-way ANOVA followed by Tukey’s multiple comparison post hoc test (C, F). ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗∗p < 0.01.

Figure 4

α-syn activated microglia are rescued from their inflammatory program by naive microglia (A) Heatmap of Pearson Correlation ρ value for the means of gene expression values in acceptors and donors over 5 h of co-culture. (B) 2D PCA and (C) heatmap of 78 DE genes in donors between 0 and 300 min of co-culture. (D) Heatmap for top 15 enriched GO terms sorted according to the Enrichment Scores (ES) in donors at 0 min up to 300 min of co-culture with acceptors. (E) Bar chart of ES for selected GO terms in donors and acceptors at 0 and 300 min. (F) ES for selected GO terms in donors’ transcriptomes over time in co-culture with acceptors. Green line indicates the baseline ES at 0 min. See also Figures S3 and S4.

Figure 5

Mitochondrial propagation reduces microglial ROS (A) Quantification of the ratio of SYTOX penetration and intercalation into acceptors (Ctrl) and donors (α-syn). n = 5. (B) Quantification of the mean SYTOX penetration and intercalation into acceptors and donors over time in co-culture. n = 3 with dublicate measurements. (C) Representative immunostaining and quantification of healthy and condensed mitochondria. n = 5 with duplicate measurements. At least 20 individual cells were analyzed per n. (D) Quantification of the ratio of ROS production in naive acceptors (Ctrl) and α-syn-treated donors (α-syn). n = 5. (E) Quantification of the mean ROS production in acceptors and donors over time in co-culture. n = 3 with dublicate measurements. (F) Quantification of the exchange of mitochondria from acceptors to donors. n = 3. (G) Representative immunostaining demonstrating the presence of mitochondria (MitoTracker) and α-syn inside cell-to-cell connections. (H) Validation of enrichment analysis of the intrinsic apoptotic signaling pathway. (I) Schematic drawing of crossover co-culture experiments using WT and LRRK2 G2019S mutant microglia. The schematic was created using BioRender.com and Adobe Illustrator. (J) Quantification of the ROS production in crossover experiments using WT or LRRK2 mutant microglia. n = 3. (K) Quantification of the individual MitoTracker signal in crossover experiments with WT and LRRK2 G2019S microglia. Acceptors were stained for MitoTracker and their propagation toward donors was assessed. n = 3 with dublicate or triplicate measurements. Graphs represent the mean ± SEM and were analyzed by a two-tailed t test (A, C, D), or by one-way ANOVA followed by Tukey’s multiple comparison post hoc test (B, E, F, J) or by Dunn’s multiple comparison post hoc text (K). ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗∗p < 0.01, ^∗p < 0.05. Scale bar: 20 μm. See also Figures S4, S5, and S6.

Figure S5

Effects of the LRRK2 G2019S mutation on mitochondrial fitness, related to Figure 5 (A) Schematic drawing of the Agilent Seahorse XF Cell Mito Stress Test profile, showing key parameters of mitochondrial function upon inhibition of the Electron Transport Chain complexes. (B) Oxygen Consumption Rate (OCR) of WT microglia (gray) and microglia carrying the LRRK2 G2019S mutation (blue) under basal conditions. n = 3 independent experiments. (C) Oxygen Consumption Rate (OCR) of WT microglia (gray/black) and microglia carrying the LRRK2 G2019S mutation (light blue/dark blue) upon treatment with 2 μM fibrillar α-syn for 24 h. n = 3 independent experiments. (D) Representative immunocytochemical staining and 3D reconstructions of WT microglia and microglia carrying the LRRK2 G2019S mutation demonstrating increased mitochondrial circulation (TOM20, green) following exposure to fibrillar α-syn (orange). (E–G) Quantification of the mitochondrial length (E), elongation score (F), and the mitochondrial area per cell size (G) of WT and LRRK2 G2019S mutant microglia. n = 3 independent experiments. (H) Quantification of the mean ROS release of WT microglia and microglia carrying the G2019S mutation under basal conditions and upon treatment with 2 μM α-syn fibrils for 24 h. n = 6 for WT and n = 4 for LRRK2 G2019S. All graphs are presented as mean ± SEM and were analyzed by two-way ANOVA. ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗∗p < 0.01; ^∗p < 0.05. Scale bar: 10 μm.

Figure S6

α-syn redistribution spread inflammation in microglia carrying the LRRK2 G2019S mutation, related to Figure 5 (A) Heatmap of Z-score transformed gene expression values for DE transcripts between WT and LRRK2 G2019S mutant microglia related to the Hallmark “ROS“. (B) Quantification of the α-syn transfer rate (left) and the number of α-syn positive cells (right) from donors to acceptors using WT (white) or LRRK2 G2019S microglia (blue). n = 4 independent experiments. (C) Quantification of the SYTOX penetration and intercalation into acceptors (blue) and donors (green) over time in co-culture using microglia carrying the LRRK2 G2019S mutation. n = 3 with duplicated treatments for all conditions. (D) Quantification of ROS production in acceptors (blue) and donors (green) over time in co-culture using microglia carrying the LRRK2 G2019S mutation. n = 3 with duplicated treatments for all conditions. (E) Quantification of the exchange of mitochondria from healthy acceptors to affected donors using microglia carrying the LRRK2 G2019S mutation. n = 3 independent experiments with triplicate treatments per condition. (F) Quantification of α-syn transfer from donors to acceptors using the LRRK2 inhibitor GSK 2578215A. n = 3 independent experiments with duplicated or triplicated measurements. Graphs are presented as mean ± SEM and were analyzed by t test (B) or one-way ANOVA followed by Tukey’s multiple comparison post hoc test (D right, E) or one-way ANOVA followed by a Dunn’s multiple comparison post hoc test (C right). ^∗∗∗∗p < 0.0001, ^∗∗∗p < 0.001, ^∗∗p < 0.01, ^∗p < 0.05.

Figure 6

Cell-to-cell transfer of aggregated α-syn in microglia in vivo (A) Schematic illustrating the preparation of organotypic slice cultures (OSC) used for experiments shown in (B) and (C). (B and C) Representative immunostainings and 3D reconstructions of CellTracer labeled microglia containing α-syn injected into the cortex (B) or hippocampus (C) of an OSC connected to tissue-resident microglia with α-syn-positive inclusions. (D) Schematic illustrating in vivo 2-photon imaging used for experiments shown in (E) and (F). (E) Representative recording demonstrating the formation of a microglial network (Cx3cr1^GFP) upon the injection of α-syn. (F) Representative time-lapse recordings demonstrating the transfer of α-syn between microglia via cellular membrane connections. Cellular connections were retracted (asterisk) once α-syn got transferred to the neighboring microglia. Schematics were created using BioRender.com. Scale bars: 20 μm. See also Figure S7.

Figure S7

Formation of a functional microglial network, related to Figure 6 (A) Representative recording demonstrating the formation of a microglial network upon the intracranial injection of α-syn fibrils in Cx3cr1^GFP animals in vivo. (B) Quantification of the number of connected cells per individual cells and the percentage of microglia being integrated into a cellular network upon intracranial injection of α-syn fibrils in Cx3cr1^GFP animals in vivo. n = 2 animals per group with three to four randomly chosen areas that were analyzed for network formation. Interconnected microglia were counted manually. A total of at least 42 microglia were analyzed. (C) Quantification of process movement velocity of microglia recorded by 2-photon imaging with and without intracranial injection of α-syn fibrils in Cx3cr1^GFP animals (left panel). Quantification of process movement velocity of microglia not transmitting or transmitting α-syn aggregated to neighboring cells upon intracranial injection of α-syn fibrils. n = 25-30 individual processes were quantified. (D) Representative recording and 3D reconstruction of the cells in Figure 6F demonstrating the formation of a microglial network (Cx3cr1^GFP, green) upon the injection of α-syn (red). (E) Representative recording and 3D reconstruction of distant microglia (Cx3cr1^GFP, green) containing α-syn (red). (F) Representative time-lapse recording of microglia (Cx3cr1^GFP, green) demonstrating that α-syn (red) is shuffled back into the cell soma when cells could not share the burden of α-syn by attaching to neighbor cells. (G) Quantification of particles that underwent transfer from one cell to another for their traveling distance and particle transfer velocity for cells shown in Figure 6F and Figure S14D. (H) Quantification of particles which transfer to a neighboring cell was unsuccessful for their traveling distance and particle transfer velocity for cells shown in Figure S14E and S14F. All graphs are presented as mean ± SEM and were analyzed by t test. ^∗∗∗∗p < 0.0001, ^∗∗p < 0.01. Scale bars: 20 μm.

Figure 7

Cell-to-cell transfer of aggregated α-syn in human tissue and MDMis (A) Representative imaging of human cortical tissue from DLB patients. Samples were analyzed for Iba1 and α-syn. Arrowheads point toward α-syn aggregates containing microglia-to-microglia connections. (B) Schematic drawing of the isolation and differentiation of PBMCs into MDMis. The schematic was created using BioRender.com and Adobe Photoshop. (C) Charts representing the receptor expression profile of MDMis before and after differentiation as measured by FACS. (D) Quantification and comparison of the percentage and the transfer index of α-syn in human MDMis derived from healthy controls or DLB patients. n = 4. (E) Representative immunostaining of F-actin⁺ MDMis demonstrating the formation of membranous tubular connections between donors and acceptors. Graphs represent the mean ± SEM and were analyzed by t test. ^∗p < 0.05. Scale bars: 20 μm. See also Figure S8.

Figure S8

α-syn aggregates trigger a stronger ROS release in DLB patient-derived monocyte-derived microglia, related to Figure 7 (A) Representative immunohistochemical staining (left panel) and 3D reconstruction (right panel) of human cingulate gyrus samples from MSA patients. Samples were analyzed for Iba1-positive microglia (white) and α-syn (red). DAPI (blue) was used as nuclear counterstain. (B) Representative immunostaining (left panel) and 3D reconstruction (right panel) of human cingulate gyrus tissues from MSA patients. Samples were analyzed for Iba1-positive microglia (white) and α-syn (red). (C) Representative super-resolution imaging of human cortical tissues from DLB patients. Samples were analyzed for Iba1-positive microglia (red) and α-syn (red). (D) Schematic drawing of the use of patient monocytes-derived microglia. (E and F) Quantification of the mean CellROX signal (E) and the percentage of CellROX positive cells (F) using patient-derived monocyte-derived microglia treated for 24 h with 1 μM α-syn fibrils. All graphs are presented as mean ± SEM and were analyzed by a two-way ANOVA in conjunction with Sidak’s multiple comparison test. ^∗p < 0.05. Scale bar: 20 μm.