Neural stem cells traffic functional mitochondria via extracellular vesicles

Abstract

Neural stem cell (NSC) transplantation induces recovery in animal models of central nervous system (CNS) diseases. Although the replacement of lost endogenous cells was originally proposed as the primary healing mechanism of NSC grafts, it is now clear that transplanted NSCs operate via multiple mechanisms, including the horizontal exchange of therapeutic cargoes to host cells via extracellular vesicles (EVs). EVs are membrane particles trafficking nucleic acids, proteins, metabolites and metabolic enzymes, lipids, and entire organelles. However, the function and the contribution of these cargoes to the broad therapeutic effects of NSCs are yet to be fully understood. Mitochondrial dysfunction is an established feature of several inflammatory and degenerative CNS disorders, most of which are potentially treatable with exogenous stem cell therapeutics. Herein, we investigated the hypothesis that NSCs release and traffic functional mitochondria via EVs to restore mitochondrial function in target cells. Untargeted proteomics revealed a significant enrichment of mitochondrial proteins spontaneously released by NSCs in EVs. Morphological and functional analyses confirmed the presence of ultrastructurally intact mitochondria within EVs with conserved membrane potential and respiration. We found that the transfer of these mitochondria from EVs to mtDNA-deficient L929 Rho0 cells rescued mitochondrial function and increased Rho0 cell survival. Furthermore, the incorporation of mitochondria from EVs into inflammatory mononuclear phagocytes restored normal mitochondrial dynamics and cellular metabolism and reduced the expression of pro-inflammatory markers in target cells. When transplanted in an animal model of multiple sclerosis, exogenous NSCs actively transferred mitochondria to mononuclear phagocytes and induced a significant amelioration of clinical deficits. Our data provide the first evidence that NSCs deliver functional mitochondria to target cells via EVs, paving the way for the development of novel (a)cellular approaches aimed at restoring mitochondrial dysfunction not only in multiple sclerosis, but also in degenerative neurological diseases.

Fig 1. NSCs shed mitochondrial proteins and mtDNA via EVs.

(a) Overview of multiplex TMT-based proteomic experiment. TMT-based proteomics identified a total of 9,971 proteins, of which 9,951 were quantitated across all conditions. (b) Relative abundance of proteins annotated with indicated GOCC subcellular localisations in EVs compared with NSC whole-cell lysates (NSCs). Annotations were available for 9,049/9,971 cellular proteins identified in the multiplex TMT-based functional proteomic experiment illustrated in a. Boxplots show median, interquartile range, and Tukey whiskers for proteins with the following annotations: extracellular vesicular exosome (GO:0070062, black outline, enriched in EVs); mitochondrion (GO:0005739, red outline, enriched in EVs); and nucleus (GO:0005634, blue outline, depleted in EVs). Data are from N = 3 independent biological replicates. (Data available on ProteomeXchange, identifier PXD024368, and in S3 Data). (c) Relative abundance of proteins from different mitochondrial compartments (outer membrane, matrix, and inner membrane) in EVs compared with NSCs. Volcano plots show statistical significance (y axis) vs. fold change (x axis) for 9,951/9,971 cellular proteins quantitated across all 3 biological replicates (no missing values) in the multiplex TMT-based functional proteomic experiment illustrated in a. Proteins annotated with the following GOCC subcellular localisations are highlighted in red: mitochondrial outer membrane (GO:0005741, enriched in EVs); mitochondrial inner membrane (GO: 0005743, enriched in EVs); and mitochondrial matrix (GO:0005759, enriched in EVs). An FDR threshold of 5% is indicated (proteins with Benjamini–Hochberg FDR-adjusted p-values (q values) < 0.05). (Data available on ProteomeXchange, identifier PXD024368, and in S3 Data). (d) Relative abundance of selected mitochondrial proteins in EVs compared with NSCs. Mitochondrial complex (C) proteins enriched in EVs and encoded in the mitochondrial (upper panel) or nuclear (lower panel) genomes in the multiplex TMT-based functional proteomic experiment include: NADH:ubiquinone oxidoreductase or CI [mtnd2 (ND2 subunit), mtnd5 (ND5 subunit), ndufaf6 (assembly factor 6), ndufa9 (subunit A9), ndufs3 (core subunit S3), ndufb5 (1 beta subcomplex subunit 5)], succinate dehydrogenase or CII [Sdha (Subunit A), sdhd (cytochrome b small subunit), sdhb (iron-sulfur subunit), sdhc (cytochrome b560 subunit)], cytochrome b-c1 or CIII [mt-cyb (cytochrome B), Uqcrfs1 (subunit 5), Uqcrc2 (subunit 2), coq10b (coenzyme Q10B), uqcrq (subunit 8)], cytochrome C oxidase or CIV [mt-co3 (oxidase III), mtco1 (oxidase I), mtco2 (oxidase II), cox15 (subunit 15), cox18 (assembly protein 18), cox6c (subunit 6C), cox5a (subunit 5a)] and ATP synthase or CV [atp5f1 (subunit gamma), atp5s (subunit S), atp5d (subunit delta), atp5h (subunit D)]. Mean abundances (fraction of maximum) and 95% CIs from N = 3 independent biological replicates are shown. *q < 0.05, **q < 0.01, ***q < 0.001 vs. NSCs. (Data available on ProteomeXchange, identifier PXD024368, and in S3 Data). (e) Representative PCR amplification of DNA extracted from NSCs, EVs, and isolated mitochondria (Mito). The mitochondrial encoded gene mt-ND1 (NADH-ubiquinone oxidoreductase chain 1) was found to be present in EVs, Mito, and NSCs (L929 Rho⁰ were used as negative controls). (f) Representative protein expression analysis by WB of NSCs, EVs, and isolated mitochondria (Mito). Mitochondrial complex proteins (CV-ATPase, CII-SDHA, CII-SDHB, CIV-MTCO1, and CIII-UQCRC2), EV positive markers (Tsg-101, Pdcd6ip, and CD9), and negative EV marker (Golga2) are shown, as well as β-actin. CI, II, II, IV, V, complex I, II, II, IV, V; EV, extracellular vesicle; FDR, false discovery rate; GOCC, Gene Ontology Cellular Component; NSC, neural stem cell; PCR, polymerase chain reaction; TMT, Tandem Mass Tag; WB, western blot.

Fig 2. Quality controls of EVs.

(a) Protein expression by WB analysis of EVs isolated using in house protocol, with commercial kits (Qiagen cat. No 76743 and Invitrogen cat. No 4478359), and an alternative protocol with an additional 0.22 μm ultrafiltration step. NSCs and isolated mitochondria (Mito) are used as comparative controls. Mitochondrial complexes proteins (CV-ATPase, CII-SDHA, CII-SDHB, CIV-MTCO1, and CIII-UQCRC2), mitochondrial outer membrane translocase (TOMM20), and exosomal positive (Pdcd6ip and CD9) and negative (Golga2) markers are shown, as well as β-actin. (b) Longer exposure of the lane containing the EVs isolated with the alternative protocol with an additional ultrafiltration step showing mitochondrial proteins. EV, extracellular vesicle; NSC, neural stem cell; WB, western blot.

Fig 3. Proteomic analysis of EVs and exosomes.

(a) Overview of proteins enriched in EVs and/or exosomes. Venn diagram shows overlap of 4,151/9,971 cellular proteins significantly enriched (q < 0.05) in either EVs (green) or exosomes (blue) or both, compared with NSCs. (Data available on ProteomeXchange, identifier PXD024368, and in S3 Data). (b) Relative abundance of proteins in exosomes (fractions 6–9) compared with EVs. Volcano plot shows statistical significance (y axis) vs. fold change (x axis) for 9,951/9,971 cellular proteins quantitated across all N = 3 biological replicates (no missing values) in the multiplex TMT-based functional proteomic experiment illustrated in Fig 1A. A total of 187 proteins were found to be significantly depleted in exosomes vs. EVs (blue), while 25 proteins were significantly enriched (red); q< 0.05 (S1 Data). (Data available on ProteomeXchange, identifier PXD024368, and in S3 Data). (c) Relative abundance of proteins from different mitochondrial compartments (outer membrane, matrix, and inner membrane) in exosomes (fractions 6–9) compared with NSCs. Volcano plots show statistical significance (y axis) vs. fold change (x axis) for 9,951/9,971 cellular proteins quantitated across all 3 biological replicates (no missing values) in the multiplex TMT-based functional proteomic experiment illustrated in Fig 1A. Proteins annotated with the following GOCC subcellular localisations are highlighted in red: mitochondrial outer membrane (GO:0005741, enriched in EVs); mitochondrial matrix (GO:0005759, enriched in EVs); and mitochondrial inner membrane (GO: 0005743, enriched in EVs). An FDR threshold of 5% is indicated (proteins with Benjamini–Hochberg FDR-adjusted p-values (q values) <0.05). (d) Representative protein expression analysis by WB of EV fractions (2–10) obtained via continuous sucrose gradient. Fractions 6–9 (corresponding to the expected exosomal density between 1.13 and 1.21 g/ml) were specifically enriched for mitochondrial complex proteins (CV-ATPase, CII-SDHA, CI-NDUFS88, CII-SDHB, CIV-MTCO1, and CIII-UQCRC2), for the mitochondrial outer membrane translocase TOMM20, and the exosomal marker Tsg-101 (while they were negative for the Golgi marker Golga2). β-actin is also shown. (e) Representative PCR amplification of DNA extracted from EV fractions (2–10) obtained via continuous sucrose gradient. The mitochondrial encoded gene mt-ND1 was found to be present in most of the EV fractions, while the nuclear encoded mitochondrial gene Sdhd was used as negative control. EV, extracellular vesicle; FDR, false discovery rate; mt-ND1, mitochondrial gene NADH dehydrogenase subunit 1; PCR, polymerase chain reaction; Sdhd, succinate dehydrogenase complex subunit D; TMT, Tandem Mass Tag; WB, western blot.

Fig 4. NSC EVs include structurally and functionally intact mitochondria.

(a) TEM data showing size particle analysis and quantification of mitochondria found in the EV preparations (purple dots) compared to non-mitochondrial EVs (grey dots). Data are mean values (± SEM) from N = 2 biological replicates. (Data available in S3 Data). (b) Representative NanoFCM density plot of EVs and EVs^Mito_depl. labelled with the canonical EV marker CD63 and the mitochondrial dye MitoTracker red. (Data available in S3 Data). (c) Representative cryo-TEM image of a free mitochondria labelled with anti-TOMM20 (red arrows) antibody conjugated to 10-nm gold NP. Inset: magnified ROI pseudocolored (or not) to highlight the 2 mitochondrial membranes (white arrowheads). Scale bars: 200 nm. (d) Representative cryo-TEM image of NSC EVs treated with saponin and labelled with anti-CD63 (red arrowheads) and anti-TOMM20 (red arrows) antibodies conjugated to 10 nm and 20 nm gold NP, respectively. Inset: magnified ROI pseudocolored (or not) to highlight the 3 membranes (white arrowheads). Scale bar: 200 nm. (e) Mitochondrial membrane potential of NSCs, EVs, and Mito preparations treated (or not) with the mitochondrial uncoupler CCCP. *p ≤ 0.05. Data are mean values (± SEM) from N ≥ 2 independent experiments. (Data available in S3 Data). (f) Representative mitochondrial respiration of permealised EVs detected by HRR. Representative plot showing O₂ concentration changes over time upon serial additions of selected mitochondrial complexes substrates, inhibitors, and uncouplers, including CI substrates (G+M: ADP), CI inhibitor (rotenone: Rot), CII substrate (succinate: Succ), cytochrome c (Cyt c) to compensate for a possible loss due to outer membrane disruption, CIII inhibitor (antimycin A: AA), CIV electron donor (N,N,N′,N′-tetramethyl-p-phenylenediamine: TMPD) and CIV inhibitor (potassium cyanide: KCN). (Data available in S3 Data). (g) Complex respiratory rate in EVs. Data are mean values (± SEM) from N = 3 independent experiments. (Data available in S3 Data). (h) Representative image of mitochondria respiratory chain native complexes separated by BN-PAGE showing the presence of structurally intact respiratory complexes (CI–V) in NSCs, released EVs, and isolated Mito preparations. (i) Representative image of in situ gel activity of CI-II-IV in NSCs, EVs, and Mito obtained from BN-PAGE gel incubation for 24 hours. ADP, adenosine diphosphate; BN-PAGE, blue native polyacrylamide gel electrophoresis; CCCP, carbonyl cyanide m-chlorophenyl hydrazone; cryo-TEM, cryo-transmission electron microscopy; EV, extracellular vesicle; G+M, glutamate and malate; HRR, high-resolution respirometry; NanoFCM, nano flow cytometry; NP, nanoparticles; NSC, neural stem cell; ROI, region of interest; TEM, transmission electron microscopy.

Fig 5. EV-associated mitochondria revert the auxotrophy of mtDNA-depleted cells.

(a) Experimental setup for in vitro studies with L929 Rho⁰ cells. L929 Rho⁰ cells were deprived of uridine (Uridine⁻) and then treated with either MitoDsRed⁺ EVs (ratio 1:30 or 1:90) or MitoDsRed⁺ Mito (ratio 1:30 or 1:90) after 3 days. (b) Representative confocal images and quantification of Uridine⁻ L929 Rho⁰ cells showing incorporation of MitoDsRed⁺ mitochondria at 24 hours from treatment with either EVs or Mito (ratio 1:30). Orthogonal section (XY) of Z-stacks is shown. Data are mean values (± SEM) from Data are from N = 4 biological replicates per condition (Mann–Whitney). *p < 0.05. Scale bars: 25 μm. (Data available in S3 Data). (c) Representative images and quantification of L929 Rho⁰ cells surviving 5 days after treatment with either EVs or Mito. Data are mean values (± SEM) from N = 4 biological replicates per condition. *p ≤ 0.05 (One-Way ANOVA, followed by Mann–Whitney). Scale bars: 3.25 mm. (Data available in S3 Data). (d) Sanger sequencing chromatograms showing the mitochondrial encoded gene mt-ND3 in L929 Rho⁰ cells at 16 days after treatment with either EVs or Mito. NSCs and L929 Rho⁰ cells are used as positive and negative controls, respectively. EV, extracellular vesicle; mt-ND1, mitochondrial gene NADH dehydrogenase subunit 1.

Fig 6. EV-associated mitochondria integrate in the host mitochondrial network.

(a) Experimental setup for the functional in vitro studies of EVs on Mφ. Mφ^LPS were treated with EVs spontaneously released from MitoDsRed⁺ NSCs (ratio 1:30). Uptake of MitoDSred⁺ particles and functional analyses of Mφ were assessed at 6 hours posttreatment. (b) Flow cytometry–based representative density plots of Mφ and Mφ^LPS at 6 hours after treatment with MitoDsRed⁺ EVs or exosomes. Data are mean % (± SEM). ***p ≤ 0.001. N = 3 independent biological replicates. (Data available in S3 Data). (c) A representative confocal image (orthogonal section (XY) of Z-stacks is shown) of Mφ^LPS treated with EVs at 6 hours, showing uptake of MitoDSred⁺ mitochondria. Nucleus is stained with DAPI (blue). Scale bar: 3 μm. (d) Representative spinning disk micrographs (maximum intensity projection of Z-stacks) and quantification of MitoDsRed⁺ EVs (red) co-localising with the lysosomal marker LAMP1 (green) or the peroxisomal marker PMP70 (blue) in Mφ. Data are mean values (± SEM). *p < 0.05. N ≥ 10 cells per condition (2 independent experiments). Scale bars: 2 μm. (Data available in S3 Data). (e) Representative spinning disk images (orthogonal section (XY) of Z-stacks is shown) and relative quantification showing MitoDSred⁺ EVs (red) attached or included in the mitochondrial network of Mφ (previously stained with MitoTracker Green FM). Inset: magnified 3D surface reconstruction of included mitochondria (Imaris Software). Data are percentage of either attaching or including particles over total MitoDSred⁺ particles in Mφ (± SEM). **p < 0.01. N ≥ 5 cells per condition from N = 3 independent experiments. Scale bars: 5 μm. (Data available in S3 Data). (f) Representative images (orthogonal section (XY) of Z-stacks) of a split FPs showing EVs^{sfCherry2,1–10} fusing with Mφ^sfCherry2,11 (in red) juxtaposed to the host TOMM20⁺ mitochondrial network (in green) at 6 hours from EV treatment. Nuclei are stained with DAPI (blue). Data are mean values (± SEM). **p < 0.01. N = 10 cells per condition. Scale bars: 5 μm. (Data available in S3 Data). (g) Representative CLEM image of Mφ^LPS treated with MitoDsRed⁺-EVs for 6 hours. Top left panel, confocal image (orthogonal section (XY) of 1 Z-stack) showing MitoDsRed⁺ EVs (red) and Mφ^LPS nuclei (blue); bottom left panel, scanning EM image of the Mφ^LPS depicted in the confocal image. Scale bars: 5 μm. Middle panel, magnified ROI of the Mφ^LPS mitochondrial network ultrastructure; right panel, superposition of confocal and scanning EM images showing co-localisation of the MitoDsRed⁺ mitochondria (red) with the host Mφ^LPS mitochondrial network. Nucleus is pseudocolored in blue. Asterisks show MitoDsRed⁺ mitochondria. Scale bars: 1 μm. CLEM, correlative-light electron microscopy; DAPI, 4′,6-diamidino-2-phenylindole; EV, extracellular vesicle; LPS, lipopolysaccharide; ROI, region of interest.

Fig 7. Pro-inflammatory mononuclear phagocytes uptake EV-associated mitochondria via endocytosis.

(a) In vitro experimental setup of EV uptake studies in Mφ^LPS. Mφ^LPS were treated with either Cyto or D/P and then exposed to MitoDsRed⁺ EVs (1:30). Latex beads and liposomes were used as positive controls of phagocytosis and endocytosis, respectively. (b, c) Representative confocal microscopy images (maximum intensity projection of Z-stacks) and quantification of MitoDsRed⁺ EV (red) uptake in Mφ^LPS (stained for F4/80, green) in the presence or absence of endocytosis (D/P) and actin mediated phagocytosis/endocytosis (Cyto) inhibitors. Nuclei are stained with DAPI (blue). Data are mean FI over unstimulated Mφ (± SEM) from N ≥ 8 ROIs per condition. ^#p < 0.05, ^##p < 0.01 vs. unstimulated Mφ. *p < 0.05, **p < 0.01, ***p < 0.001. Scale bars: 10 μm. (Data available in S3 Data). Cyto, Cytochalsin; D/P, Dynasore and Pitstop 2; DAPI, 4′,6-diamidino-2-phenylindole; EV, extracellular vesicle; FI, fold induction; LPS, lipopolysaccharide; ROI, region of interest.

Fig 8. The transfer of EV-associated mitochondria inhibits the metabolic switch of pro-inflammatory mononuclear phagocytes.

(a) Representative spinning disk images and quantification showing Mφ mitochondrial network labelled with TOMM20 (green) polymorphic dynamics after LPS stimulation and/or EV treatment (1:30). Data are expressed as mean % (± SEM). ***p < 0.001. N = 6 biological replicates. Scale bars: 4 μm. (Data available in S3 Data). (b) Top enriched KEGG pathways in genes up-regulated in EV-treated vs. untreated Mφ^LPS at 6 hours. Expression data obtained by microarray analysis. (Data available on ArrayExpress, identifier E-MTAB-8250). (c) XF assay of the OCR during a mitochondrial stress protocol of Mφ^LPS at 6 hours from EV treatment (1:30). Unstimulated Mφ were used as controls. Data are mean values (± SEM). *p < 0.05, **p < 0.01 vs. Mφ^LPS. N = 2 independent experiments. (Data available in S3 Data). (d) XF assay of the basal OCR and ECAR of Mφ^LPS at 6 hours from treatment with EVs or treatment with EVs preexposed to the uncoupling agent FCCP vs. Mφ^LPS. Unstimulated Mφ were used as controls. Data are mean values (± SEM). ^##p < 0.01, ^###p < 0.001 vs. unstimulated Mφ. *p < 0.05, **p < 0.01. N ≥ 4 technical replicates from N ≥ 2 independent experiments. (Data available in S3 Data). (e) Expression levels (qRT-PCR) of pro-inflammatory genes (Il1β, Nos2, and Il6) in Mφ^LPS at 6 hours from treatment with EVs or treatment with EVs preexposed to the uncoupling agent FCCP. Data are mean FI over unstimulated Mφ (± SEM). ^##p < 0.01, ^###p < 0.001 vs. unstimulated Mφ. *p < 0.05, **p < 0.01, ***p < 0.001. N ≥ 3 biological replicates from N ≥ 2 independent experiments. (Data available in S3 Data). ECAR, extracellular acidification rate; EV, extracellular vesicle; FI, fold induction; KEGG, Kyoto Encyclopedia of Genes and Genomes; LPS, lipopolysaccharide; OCR, oxygen consumption rate; qRT-PCR, quantitative real-time polymerase chain reaction; XF, extracellular flux.

Fig 9. Transplanted NSCs transfer mitochondria to mononuclear phagocytes and astrocytes during EAE in vivo.

(a) In vivo experimental setup of EV and NSC treatment in EAE mice. At PD, mice received a single ICV injection of either fGFP⁺/MitoDsRed⁺ NSCs, EVs derived from fGFP⁺/MitoDsRed⁺ NSCs (EVs), or fGFP⁺/MitoDsRed⁺ EVs depleted of mitochondria (EVs^Mito_depl.), or fGFP⁺/MitoDsRed⁺ EVs depleted of CD63+ EVs (EVs^CD63_depl.). Behavioural analysis was carried out daily until the end of the experiment. Neuropathology was performed at 55 dpi. (b) Behavioural outcome showing significant amelioration of the EAE score in mice treated with EVs (N = 5) and NSCs (N = 5), but not in EAE mice treated with EVs^Mito_depl. (N = 4) or EVs^CD63_depl. (N = 5) vs. PBS (N = 8). Data are mean values (± SEM). *p < 0.05. (Data available in S3 Data). (c) Percentage of fGFP^-/MitoDsRed⁺ particles co-localising with GFAP⁺ astrocytes, NeuN⁺ neurons, Olig2⁺ oligodendrocytes, CD3⁺ T cells, or F4/80⁺ mononuclear phagocytes. fGFP⁺/MitoDsRed⁺ NSCs were injected ICV into EAE mice (white bars) and in nonimmunised control mice (hatched bars). Data are mean values (± SEM) from N = 4 biological replicates. *p < 0.05, **p < 0.01, ***p < 0.001. (Data available in S3 Data). (d) Representative pictures of mitochondrial transfer events detected with confocal imaging (maximal projection). Transfer of MitoDsRed⁺ particles (arrowheads) is shown between fGFP⁺/MitoDsRed⁺ NSCs and GFAP⁺ astrocytes (cortex), NeuN⁺ neurons (cortex), Olig2⁺ oligodendrocytes (corpus callosum), CD3⁺ T cells (meninges), or F4/80⁺ mononuclear phagocytes (IV ventricle). Long processes of NSCs can be seen in green, while nuclei are stained with DAPI (blue). Scale bars: 20 μm. DAPI, 4′,6-diamidino-2-phenylindole; dpi, days post immunisation; EAE, experimental autoimmune encephalomyelitis; EV, extracellular vesicle; ICV, intracerebroventricular; NSC, neural stem cell; PD, peak of disease.