NSCs were derived and previously established from ICR mouse embryos at E14.5 (Kobayashi et al., 2019). NSCs were frozen several times and then passed about five times before use. NSCs were able to differentiate into neural and glial fates after the inductions (not shown). Active NSCs were cultured in proliferation medium [20 ng/ml epidermal growth factor (EGF); R&D systems, Inc., Minneapolis, MN, United States), 20 ng/ml bFGF (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan), penicillin/streptomycin (Nacalai Tesque, Kyoto, Japan), and N-2 Plus supplement (R&D Systems) in DMEM/F-12 (Gibco, Grand Island, NY, United States)] with 2 µg/ml laminin (Sigma-Aldrich Corp., St. Louis, MO, United States). Quiescence was induced with quiescent medium [50 ng/ml BMP4 (R&D Systems) in proliferation medium minus EGF] after two washings with phosphate-buffered saline (PBS; Nacalai Tesque) and achieved in 3 d. To prepare reNSCs, qNSCs were washed twice with PBS and cultured in proliferating medium for 2 d. For the exosome inhibition, aNSCs were cultured for 1 d and the medium was supplemented either with 10 µM GW4869 (Cayman Chemical Co., Ann Arbor, MI, United States) or dimethyl sulfoxide (DMSO; Nacalai Tesque) as the control. To measure protein synthesis, O-propargyl-puromycin (OP-puro, OPP; MedChemExpress Co., Monmouth Junction, NJ, United States) (Liu et al., 2012) was administrated for 1 h with 50 µM OP-puro in culture medium, followed by fixation with 4% paraformaldehyde (PFA; Nacalai Tesque) in PBS.

Supernatants from 15 cm culture dishes (Greiner, Oberosterreich, Austria) were centrifuged at 2,000 × g at 4°C for 20 min and at 10,000 × g at 4°C for 20 min and passed through a 0.22 µm filter (TPP, Trasadingen, Switzerland). The filtered supernatants were concentrated using a 100 kDa MW centrifugal filter (Amicon Ultra-15; EMD Millipore, Billerica, MA, United States) at 5,000 × g and then ultracentrifuged in a Beckmann optima XE with a SW 41 rotor (Beckman Coulter, Brea, CA, United States) at 100,000 × g at 4°C for 1 h. The pellets were washed with cold PBS and recentrifuged at 100,000 × g at 4°C for 1 h. The exosome pellets were recovered overnight in PBS at 4°C with gentle shaking and subjected to silver staining, western blot, particle analyses, and mass spectrometry. A total exosome isolation kit (Thermo Fisher Scientific, Waltham, MA, United States) was also used to collect exosomes from culture medium after inhibitor treatments by following the manufacturer’s protocol. The exosome pellets were recovered in PBS and subjected to NanoSight (Malvern Instruments, Malvern, United Kingdom) measurements. The exosome concentrations collected by the kit were 5.88E8 ± 6.06E6, 7.51E8 ± 2.51E7, and 7.28E8 ± 3.97E7 particles/ml from 2-d aNSC, 3-d qNSC, and 2-d reNSC cultures, respectively.

Exosome proteins were quantified with a Micro BCA™ protein assay kit (Thermo Fisher Scientific). Bovine serum albumin (BSA) standard solutions and exosome samples were diluted in lysis buffer [50 mM Tris-HCl (pH 8.0), 100 mM NaCl, 5 mM MgCl₂, and 0.5% (w/v) Nonidet P-40] and subjected to the manufacturer’s protocol. The protein concentrations in total cell lysates were quantified by the Lowry method using DC protein assay reagents (Bio Rad Laboratories, Hercules, CA, United States).

The cells were washed with cold PBS, lysed with lysis buffer [50 mM Tris-HCl (pH 8.0), 100 mM NaCl, 5 mM MgCl₂, 0.5% (w/v) Nonidet P-40, Complete™ protease inhibitor cocktail (Roche Diagnostics, Basel, Switzerland), 1 mM phenylmethylsulfonyl fluoride, 250 U/ml Benzonase (Sigma), 10 mM β-glycerophosphate, 1 mM sodium orthovanadate, 1 mM NaF, and 1 mM sodium pyrophosphate] on ice for 30 min, and subjected to SDS-PAGE. The exosome samples were either directly mixed with SDS sample buffer or first precipitated using 10% (v/v) trichloroacetic acid (TCA) and resolved in sample buffer. The following primary antibodies were used for WB: rabbit anti-actin (Sigma-Aldrich), rabbit anti-CD9 (BioVision, Inc., Milpitas, CA, United States), rat anti-MFG-E8 (R&D Systems), mouse anti-CD63 (Novus Biologicals, Littleton, CO, United States), rabbit anti-Sox2 (EMD Millipore), and mouse anti-Cycd1 (Sigma-Aldrich). Horseradish peroxidase (HRP)-conjugated anti-rabbit, anti-mouse, and anti-rat (Thermo Fisher Scientific) were used as the secondary antibodies. Western blots (WBs) were visualized by chemiluminescence using Amersham ECL or ECL prime (Cytiva, Tokyo, Japan), quantified on an LAS3000 image analyzer (Fujifilm, Tokyo, Japan), and normalized against the corresponding intensity of β-actin for cell lysates. Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software, La Jolla, CA, United States). For silver staining, a Pierce silver stain kit was used (Thermo Fisher Scientific).

Cells labeled with OP-puro were fixed with 4% PFA-PBS and stained with Alexa488-azide using Click-iT chemistry (Thermo Fisher Scientific). For immunocytochemistry to detect Ki-67 and MCM2, 4% PFA-fixed cells were immunostained with mouse anti-Ki-67 (BD Pharmingen, Franklin Lakes, NJ, USA) and rabbit anti-MCM2 (Abcam, Cambridge, United Kingdom) as the primary antibodies and then stained with anti-mouse-Alexa488 and anti-rabbit-Alexa594 as the secondary antibodies (Thermo Fisher Scientific) respectively and co-stained with DAPI (Sigma). Cell images were obtained on an AF6000 (Leica Microsystems GmbH, Wetzlar, Germany) and analyzed using ImageJ software (NIH, Bethesda, Maryland, United States). Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software).

Exosome particle size distribution and concentration were evaluated with NanoSight NS300 (Malvern Instruments). The size distribution was determined from five videos recorded by the NanoSight. For the exosome concentration determinations, 20 videos were recorded for more precise results (Parsons et al., 2017).

The exosome samples were subjected to negative staining by being loaded on the collodion surface of mesh grid (Nisshin EM, Tokyo, Japan), coated with 2% (w/v) uranyl acetate, and left to dry on the grid. They were then photographed under a JEM-1400 TEM (JEOL Ltd., Akishima-Shi, Japan).

We investigated the roles of sEVs secreted from NSCs in quiescent and proliferating conditions using cultured NSCs. To purify sEVs from cultured NSCs with high quality, we selected ultracentrifugation as the optimal method to collect abundant and pure sEVs (Figure 1A). In our method, cell culture supernatants were centrifuged to remove cell debris and passed through a 0.22 µm filter to remove larger vesicles since sEVs is usually no larger than 150 nm in size. Nanoparticle tracking analysis (NTA) with NanoSight revealed that the estimated diameter of the purified sEVs was in the range of 100–130 nm (Figure 1B). Transmission electron microscopy (TEM) revealed their donut-shaped morphology (Figure 1C). Results from both NTA and TEM analyses showed that the purified sEVs had typical exosomal characters. Previous reports used the term “exosome” to refer to a mixture of small vesicles because there was heretofore no method of fully separating them (Mathieu et al., 2019). In the present study, we use the term “exosome” to refer the sEVs. We then obtained exosomes from active, quiescent, and reactivating NSCs. BMP was used to induce quiescence in NSCs in vitro (Martynoga et al., 2013). Exosomes from aNSCs (aExo) were purified from aNSC culture supernatants after NSCs were cultured in proliferating media for 2 d. Quiescence was induced by culturing NSCs in BMP-containing quiescent medium for 3 d and exosomes from qNSCs (qExo) were collected from the supernatants. qNSCs were reactivated by changing medium from quiescent to proliferating medium, and exosomes from reNSCs (reExo) were collected after the NSCs were cultured for 2 d. WB showed that exosome markers (CD9, CD63, and MFG-E8) (Thery et al., 2006) were expressed in the purified exosomes (Figure 1D). The transcriptional factor Sox2 was used as a negative control to rule out possible cell debris contamination (Figure 1D). Taken together, the foregoing data suggest that we successfully purified exosome samples suitable for use in the subsequent assays and experiments.

We initially conducted silver staining to visualize proteomic inequalities and profiled the exosome protein contents. Silver staining showed similar protein expression patterns in the exosomes from active, quiescent, and reactivating NSCs (Figure 1E). We then exanimated the exosome profiles by nano-scale liquid chromatography/tandem mass spectrometry (nano LC/MS/MS) (Supplementary Methods). Digested proteins from individual exosome samples were labeled with nine different Tandem mass tag (TMT) 10-plex reagents (aExo vs. qExo vs. reExo; n = 3 each) (Ogata and Ishihama, 2020). Relative protein abundance was quantified based on the “reporter ion intensity.” Among 1,283 proteins we identified, 1,178 proteins were quantified in at least two out of the three replicates in at least one condition and were used for further analysis (Supplementary Table S1). To compare the exosome contents in the databases, we consulted the lists of proteins in Exocarta which includes those involved in MVB biogenesis, those associated with membrane proteins such as CD9 and CD63, exosome markers, and so on (Keerthikumar et al., 2016). There were 372 proteins common to the NSC exosomes and Exocarta (Bardou et al., 2014). Thus, our exosome samples shared many proteins with those previously reported (Figure 2A). The sample comparisons were “aExo vs. qExo” (AQ), “qExo vs. reExo” (QRE), and “aExo vs. reExo” (ARE) (Figures 2B,C). The volcano plots revealed bilateral significantly expressed proteins (log₂ fold change (FC) > 1; p-value < 0.05) in AQ and QRE (Figure 2B; left and center charts). For ARE, however, there were only eight types of highly enriched proteins in aExo and no upregulated proteins in reExo relative to aExo (orange dots in right chart of Figure 2B; Supplementary Table S2).

To obtain functional insights of the differentially expressed protein cargos, we used gene lists of each comparison groups (log₂ FC > 1; p-value < 0.05) in a gene ontology (GO) analysis (Supplementary Table S2) conducted on the Metascape website (https://metascape.org/gp/index.html#/main/step1). We identified enriched GO/KEGG and Reactome terms (Zhou et al., 2019). For both AQ and QRE, vesicle trafficking-related terms were significant in aExo, and translation-related terms were significant in qExo (Figure 2C; Supplementary Table S3). No GO analysis was conducted on ARE as there were too few enriched factors, and the functional clusters could not be calculated. We also applied the “molecular complex detection” (MCODE) algorithm to detect densely related regions in the pool of protein-protein interaction webs (Bader and Hogue, 2003) and explore potential connections among our exosome proteomes. The algorithm speculated high-confidence networks of highly enriched proteins in AQ (Figures 3A,B) and QRE (Figures 3C,D) and applied the enriched terms to annotate each MCODE network (Supplementary Tables S4). MCODE cluster identified endocytosis- and membrane trafficking-related clusters as aExo- and reExo-enriched terms and translation- and ribosome-related clusters as the qExo-enriched terms in both AQ and QRE comparisons. It was consistent with the elevated scores of ribosomal protein subunits in qExo proteome (Supplementary Table S4). The foregoing results disclosed the contrasts between the quiescence and proliferation stages and facilitated the exploration of potential functional units from exosome proteomes.

Our results demonstrated distinct differences in exosome protein contents and enriched pathways between the quiescence and proliferation stages of NSCs. We speculated that the high enrichment of translation-related proteins in qExo may imply the contribution of exosomes to reduce protein translation in qNSCs by discarding the contents from cells. To explore the role of exosomes in this hypothesis, we applied a chemical inhibitor of the exosome (Figure 4A). We used GW4869, a neutral, noncompetitive inhibitor of sphingomyelinase (N-SMase) which reduces ceramide levels and inhibits the ceramide-mediated formation of mature exosomes in MVBs (Verderio et al., 2018). GW4869 is reported to inhibit exosome secretion by HEK293 cells, macrophages, and mesenchymal stem cells (Kosaka et al., 2010; Essandoh et al., 2015; Che et al., 2019). Nanosight measurements revealed that approximately one third of the exosome secretion was blocked in NSCs by GW4869 (Figure 4B). We selected GW4869 for our subsequent analyses and investigated its effects.

We next investigated whether reduced exosome secretion affects the induction of NSC quiescence in proliferating NSCs and the maintenance of quiescence in qNSCs. We tested CycD1 expression levels by WB analysis for the NSC proliferation criteria (Baldin et al., 1993). We treated NSCs with GW4869 and DMSO, as a negative control, dissolved in the quiescent medium. Contrasting CycD1 expression was detected 12 h after inducing quiescence with the exosome inhibitor. The cells treated with GW4869 displayed higher CycD1 expression levels than those in the DMSO control (Figures 4C,E). Hence, interference of exosome biogenesis and secretion impeded the NSCs from entering the G0 phase. We treated qNSCs with GW4869 in the quiescent medium to certify the influence of hindered exosome activity on the maintenance of quiescence. We discovered that after 24 h, CycD1 expression was upregulated in the presence of GW4869 (Figures 4D,F). We investigated two proliferation markers, Ki-67 and MCM2, which are widely used to detect proliferating NSCs, in GW4869-treated NSCs. MCM2 is highly expressed throughout the cell cycle, including the G1 phase and except the G0 phase, whereas Ki-67 expression peaks at M phases and is low at the G1 phase (Miller et al., 2018; Harris et al., 2021). Exosome inhibition increased the percentages of proliferating (Ki-67+ and MCM2+) cells in both quiescent-induced and quiescent-mainteined NSCs (Figures 4G–J). Thus, the blockage of exosome secretion delayed the entry of proliferating NSCs into the G0 phases and destablized the G0 phase.

To elucidate the mechanism through which exosomal inhibition destabilized the G0 phase in NSC quiescence, protein translation was measured using a puromycin analogue, OP-puro (OPP). OPP is effectively incorporated into newly generated proteins and can be used to quantify the protein synthesis rates in cells by detecting OPP-labeled proteins after fixation (Liu et al., 2012). aNSCs and qNSCs were incubated in GW-containing quiescent medium for 24 h, labeled with OPP for 1 h, and analyzed. Exosome inhibition increased the levels of OPP-labeled proteins in quiescent-induced and quiescent-maintained NSCs compared to those in the controls (Figure 4K). This higher translation in GW-treated NSCs was consistent with higher percentages of proliferating cells in those than in controls under quiescent culture conditions (Figures 4G–J). Taken together, these results demonstrated a strong correlation between exosome secretion and cell cycle regulation together with protein synthesis in NSCs quiescence.