Click chemistry extracellular vesicle/peptide/chemokine nanocarriers for treating central nervous system injuries

Abstract

Central nervous system (CNS) injuries, including stroke, traumatic brain injury, and spinal cord injury, are essential causes of death and long-term disability and are difficult to cure, mainly due to the limited neuron regeneration and the glial scar formation. Herein, we apply extracellular vesicles (EVs) secreted by M2 microglia to improve the differentiation of neural stem cells (NSCs) at the injured site, and simultaneously modify them with the injured vascular targeting peptide (DA7R) and the stem cell recruiting factor (SDF-1) on their surface via copper-free click chemistry to recruit NSCs, inducing their neuronal differentiation, and serving as the nanocarriers at the injured site (Dual-EV). Results prove that the Dual-EV could target human umbilical vascular endothelial cells (HUVECs), recruit NSCs, and promote the neuronal differentiation of NSCs in vitro. Furthermore, 10 miRNAs are found to be upregulated in Dual-M2-EVs compared to Dual-M0-EVs via bioinformatic analysis, and further NSC differentiation experiment by flow cytometry reveals that among these miRNAs, miR30b-3p, miR-222-3p, miR-129-5p, and miR-155-5p may exert effect of inducing NSC to differentiate into neurons. In vivo experiments show that Dual-EV nanocarriers achieve improved accumulation in the ischemic area of stroke model mice, potentiate NSCs recruitment, and increase neurogenesis. This work provides new insights for the treatment of neuronal regeneration after CNS injuries as well as endogenous stem cells, and the click chemistry EV/peptide/chemokine and related nanocarriers for improving human health.

Graphical abstract

Click chemistry extracellular vesicle/peptide/chemokine nanomissiles repair central nervous systems (CNS) injuries by targeting blood vessels, recruiting neural stem cells (NSCs) and inducing their differentiation into neurons.

1. Introduction

Endogenous stem cells are pluripotent cells located in specific tissues or circulating in the blood. They can self-renew and differentiate into different cell types, and thus are the basic unit of tissue repair after injury¹. After an injury, local inflammatory cells gather to release a large number of cytokines, and subsequently recruit and activate stem cells, which then initiate regeneration^2,3. These endogenous stem cells can effectively repair diseases such as bone fractures, acute blood loss, and skin damage4, 5, 6. However, their effectiveness is greatly limited in central nervous system (CNS) diseases such as stroke, brain injury, and spinal cord injury. CNS diseases are essential causes of death and long-term disability and are difficult to cure. One obstacle is that the density of local neural stem cells (NSCs) is extremely low, leading to a lack of sufficient cell sources for nerve regeneration⁷. The second obstacle is the lack of neural differentiation-inducing factors in the lesion, thus leading to the lack of directionality in stem cell differentiation, and causing local lesion fibrosis, resulting in irreversible loss of lesion function and the formation of glial scar⁸. Therefor, some studies considered introducing exogenous NSCs for local repair9, 10, 11, 12, 13. However, exogenous NSCs also face the problem of undirected differentiation and may even have complications such as malignant proliferation and immune rejection^12,14. Therefore, in view of the shortcomings of stem cell therapy, the design of constructing a new tissue repair factory construction strategy to achieve the rapid and effective recruitment of NSCs and the directed differentiation of neurons is of great value to the advancement of clinical treatment of CNS injuries.

Extracellular vesicles (EVs) are extracellular membrane vesicles secreted by cells, which can regulate cell functions and biological behaviors through active molecules such as RNA and proteins¹⁵. Because of its low immunogenicity and the ability to perform similar physiological functions of parent cells, EVs as a new type of cell-free replacement therapy have received extensive research attention. Taking the EVs secreted by mesenchymal stem cells (MSC) as an example, they can promote the recovery of nerve function through methods such as white matter repair and inhibition of neuroinflammation^16,17. However, the direct mechanism of MSC-EVs is to protect relatively healthy nerve cells and accelerate the original nerve remodeling process, which cannot fully stimulate the repair potential of NSCs¹⁸. Moreover, the NSCs-derived EVs mainly promote neurological recovery through mechanisms such as immunomodulation and edema reduction but lack the neuronal differentiation-related effect on endogenous NSCs^19,20. Therefore, it is of great need to develop EVs that can specifically stimulate NSCs to fully activate the potential of NSCs as a repair factory and continue to promote neuron regeneration.

M2 phenotype microglia is an important type of immune cell that could inhibit neuroinflammation and accelerate neurogenesis^21,22. The medium of M2 microglia has been reported to promote the differentiation of neural stem cells into neurons²¹, suggesting that the EVs secreted by them have the potential to promote NSC differentiation and can become the basis of the neuron regeneration factory. On the other hand, endogenous NSCs are generated from the subventricular zone (SVZ) and migrate to the injured site⁷. Due to the long migration distance, their migration to the injured area is limited and the absolute number in the lesion is insufficient, which limits the effectiveness of the neuron regeneration factory⁷. Therefore, constructing EVs that can recruit NSCs is expected to exert better function of inducing NSCs differentiation by increasing the number of local stem cells. In addition, due to the widespread presence of microglia in nerve tissues^23,24, direct and systemic administration of EVs is prone to be widely taken up, resulting in only a very small amount of EVs that can reach the lesion area, which greatly restricts the scale of the local neuron regeneration factory. Therefore, how to regulate the in vivo behavior of EVs and increase their targeting to the nerve injury area, thereby improving the precise recruitment of local NSCs and the performance of inducing neuron regeneration is another issue to be addressed. The modification methods for EVs include genetic engineering of parent cells and direct modification of EVs²⁵. However, the former method is complicated and costly, and genetic modification may bring potential ethical risks. Direct modification of EVs, such as the hydrophobic membrane insertion method, lacks specificity and selectivity²⁶. Copper-free click chemistry has the advantages of non-toxicity, high-efficiency, and high-specificity²⁷ which can be an ideal method for modifying EVs. Therefore, it is of great need to construct a targeted EV that can accurately recruit and induce NSC differentiation simultaneously through a copper-free click chemistry method and achieve the purpose of nerve repair by locally establishing a high-efficiency neuron regeneration factory.

Using copper-free click chemistry, the high-affinity ligand of damaged blood vessels DA7R (^DR^DP^DP^DL^DW^DT^DA)²⁸, and a class of stromal cell-derived factor-1α (SDF-1) that can recruit NSCs are simultaneously modified on the EVs secreted by M2 microglia to achieve rapid and effective recruitment and differentiation transformation of NSCs and construct an engineered neuron regeneration factory (DA7R-SDF-1-EV, Dual-EV nanocarriers, Fig. 1) for ischemic stroke treatment. Firstly, the azide group was introduced to DA7R peptide and SDF-1 factor via the amidation reaction. Secondly, the dibenzocyclooctyne (DBCO) group was modified on EVs by the reaction between the amino group of EV membrane proteins and DBCO-terminated PEGylated N-hydroxysuccinimidyl ester (DBCO-PEG₄-NHS). Finally, the functionalized EVs modified with DA7R and SDF-1 were developed by copper-free azide-alkyne cycloaddition reaction and the physicochemical properties were verified. In vitro cell experiments were used to verify Dual-EV's abilities to recruit stem cells, target vascular endothelial cells, and potentially induce NSCs to differentiate into neurons. Its ability to target the stroke site in vivo was verified by constructing the transient middle cerebral artery occlusion mice (tMCAO). Finally, the neurobiological evaluation and immunofluorescence staining were applied to assess the neurorestorative efficacy and mechanism of Dual-EV on nerve repair in mice. In summary, we describe a new strategy that can effectively recruit NSCs and induce them to differentiate into neurons for treating CNS injury.

Figure 1.

Schematic illustration of DA7R-SDF-1-EV nanomissile (Dual-EV) treatment on ischemic stroke. BV2 cells were treated with IL-4 to induce M2 polarization, then the M2 microglia-derived EVs (M2-EVs) were isolated for further modification. First, the DBCO group was modified on the surface of EVs and azide groups were introduced into DA7R peptide and the SDF-1 chemokine. Then the Dual-EVs nanomissile was engineered via copper-free click chemistry. After intravenous administration, the Dual-EV could target injured vascular endothelial cells in stroke lesions and accumulate in the lesion site via the VEGFR2-mediated effect. Then NSCs were recruited from the lateral ventricle to the ischemic region via the function of SDF-1. Finally, the Dual-EV induced the NSCs to differentiate into neurons to achieve neurogenesis after stroke.

2. Materials and methods

2.1. Materials

Interleukin 4 (IL-4, cat. CK74, Novoprotein, Shanghai, China), IBA-1 (cat. 019-19741, WAKO, Osaka, Japan), arginase-1 (ARG, cat. SC-271430 Santa Cruz, CA), DBCO-PEG₄-NHS (No. 764019, Sigma–Aldrich, USA), CD206 (cat. ab64693, Abcam, USA), recombinant murine SDF-1α (Cat. 250-20A, Peprotech, USA), CD63 (cat. SC-15363, Santa Cruz, CA), tumor susceptibility gene 101 (TSG-101, cat. ab83, Abcam, USA), glial fibrillary acid protein (GFAP, cat. MAB5804, Merck, German), Tuj 1 (cat. MAB1637, Merck, German), CD31 (cat. AF3628, R&D, USA), MAP2 (cat. MAB3418, Merck, German), DCX (cat. ab207175, Abcam, USA), Nestin (cat. MAB353, Merck, German). TryplE was obtained from Life Technologies (USA). Azide-DA7R and azide-DA7R-FITC peptides were purchased from Apeptide (Shanghai, China), Azidoacetic acid NHS ester (N₃-NHS, Lot. X-CL-1210), and sulfo-Cyanine 5.5 NHS ester (Cy5.5-NHS, Lot. Y-R-3060) were obtained from Shaanxi New Research Biosciences Co., Ltd. (Shaanxi, China).

2.2. BV2 microglia culture, polarization, and identification

BV2 microglial cells were purchased from the National Laboratory Cell Resource Sharing Platform (Wuhan, China). BV2 cells were polarized by 20 ng/mL IL-4 for 48 h, and then the M2 phenotype of BV2 cells was characterized by immunofluorescence staining of IBA-1, CD206, and ARG antibodies. The Western blot assay was also conducted to compare the expression of CD206 and ARG between IL-4-treated and untreated BV2 cells.

2.3. EVs isolation

The EVs were isolated according to our previous studies²⁹. First, the EVs-deplete FBS were prepared via ultracentrifuging at 100,000×g (SW 32 Ti rotor, Beckman Coulter Life Sciences, GER) for 16 h at 4 °C to avoid the contamination of FBS EVs. Next, the EVs-free medium prepared by the EVs-depleted FBS was used for culturing M2-BV2 cells for 48 h. Then, the cell culture supernatant was harvested and followed by differential ultracentrifugation under 4 °C at 300×g for 10 min to remove cells, 2000×g for 10 min to remove cell debris, 10,000×g for 30 min to remove large cellular vesicles, and 100,000×g for 70 min to precipitate EVs. Finally, the obtained EVs were purified through suspending in PBS and ultracentrifuging again at 100,000×g for 70 min under 4 °C, and resuspended with PBS for further modification and characterization.

2.4. Construction of functionalized EVs

The dual-functionalized DA7R-SDF-1-EVs were achieved by copper-free click chemistry. First, a hetero-bifunctional crosslinker was used to decorate reactive DBCO groups on the surface of EVs. Briefly, 0.15 mg/mL EVs were reacted with 10 μmol/L DBCO-PEG₄-NHS in PBS at room temperature (R.T.). 4 h later, the mixture was filled in an ultrafiltration tube (10 kDa, Thermo Scientific) to centrifuge at 12,000×g for 10 min, and washed with PBS for 4 times to remove excessive DBCO-PEG₄-NHS. Next, the azide-modified SDF-1 was synthesized through the reaction between SDF-1 and NHS-N₃. Briefly, NHS-N₃ (12 μL, 64 ng/μL) was added into 10 μg SDF-1 in PBS and reacted at R.T. for 4 h on a magnetic stirrer. And the unconjugated N₃-NHS was discarded with an ultrafiltration tube 4 times (3 kDa, Millipore). Finally, the azide-labeled SDF-1 was first added into EVs-DBCO at 4 °C for 4 h on a magnetic stirrer, then azide-DA7R was added to react together overnight. The obtained functionalized EVs were purified using centrifugation with ultrafiltration tubes 4 times at 14,000 rpm for 10 min for removal of undecorated ligands. Then DA7R-SDF-1-EV was suspended in PBS and stored at −80 °C for further use. To estimate the successful conjugation of SDF-1 and DA7R on EVs, N₃-Cy5.5-NHS was applied to label SDF-1, DA7R-N₃ was marked with FITC, and EVs were stained with Dil-red. Thus, this triple-fluorescence-tagged Dual-EV was photographed by confocal laser scanning microscopy (CLSM, Zeiss, German).

2.5. Characterization of functionalized EVs

Different EVs, including unmodified EVs, DA7R-EV, SDF-1-EV, Dual-EV were detected by a BCA Protein Assay Kit to assure the protein concentration, and EVs markers, TSG101 and CD63, were analyzed by the Western blot assay. The morphology of EVs was observed by a transmission electron microscope (TEM, Tecnai G2 SpiritBiotwin, FEI, USA), the EVs were dropped onto a carbon-coated copper grid for 1 min, stained with 1% uranyl acetate dihydrate for another 1 min, and then applied for observation. The size and particle number were analyzed by Nanoparticle tracking analysis (NTA, Brookhaven, New York, USA). EVs were diluted with PBS, and the Brownian motion was traced, drawing size distribution data with the Strokes–Einstein formula, and graphs were acquired with Origin 8.0.

2.6. Quantification of peptides on the decorated EVs

Azide-DA7R was labeled with one FITC by introducing a lysine at the C-terminal, which was purchased from Apeptide Co., Ltd. (Shanghai, China). This N₃-DA7R-FITC was modified on the surface of EVs by the same copper-free click chemistry. A standard curve of free N₃-DA7R-FITC was calculated by a microplate reader (BioTek, USA). And the concentration of DA7R in EVs was measured. Furthermore, the number of DA7R on the EVs was determined by connecting the particle concentration of EVs. Besides, SDF-1 was quantified by the Elisa kit (ab100741, Abcam, USA).

2.7. NSC culture, differentiation, and migration assay

NSCs were isolated from the cortex of embryonic Day 16–18 fetal mice from pregnant ICR mice (P16-18 fetal ICR) as previously described³⁰. The cells were digested with trypsin-ethylene diamine tetraacetic acid (trypsin-EDTA), and then suspended in normal DMEM-F12 containing 1% B27, 20 ng/mL endothelial growth factor (EGF), and 20 ng/mL basic fibroblast growth factor (bFGF) at 37 °C in an incubator with 5% CO₂.

NSCs differentiation was induced following described protocols. First, NSCs at P1 generation were digested with Accutase and seeded into a 24-well plate at a density of 2 × 10⁵/mL in proliferation medium and cultured for two days. Then the medium was switched into the differentiation medium (neurobasal medium, 2% B27 and 1% 100× GlutaMAX) with PBS, 10 ng/mL M0-EV, M2-EV, or Dual-EV. The medium was exchanged every two days and on Day 7, the NSCs were applied for immunofluorescence staining (DAPI for nuclei, GFAP for astrocytes, and Tuj 1 for neuron). The percentage of NSC differentiation was calculated using ImageJ software.

To estimate the ability of Dual-EV to induce NSCs migration, 3 × 10⁴ NSCs per well were cultured in the upper chamber with an 8 μm pore size membrane. Then blank medium, EV, SDF-1-EV, or Dual-EV was added in the lower chamber. Forty-eight hours later, the medium was removed and cells were washed twice with PBS. Next, use a cotton swab to completely wipe the nonmigratory NSCs on the upper side of the membrane, followed by staining NSCs on the lower side of the membrane with DAPI for 10 min. Lastly, after another three times washing with PBS, cells were visualized under inverted fluorescence microscopy (Leica, Germany) and three random fields were captured to quantify the migratory NSCs. The number of NSCs crossing the membrane was calculated by counting the cells using ImageJ software (n = 3 in each group). The migrating percentage was displayed as the ratio of the control group.

2.8. Bioinformatic analysis of the dual-M2-EVs

First, miRNA-sequencing of Dual-M2-EVs and Dual-M0-EVs was tested using bioinformatic analysis. Then heatmap of differentiated expressed miRNAs was computed using a function of heatmap.2 in ggplot 2 by R platform. Next, the target genes of 10 upregulated miRNA were predicted in miRbase databases (https://www.mirbase.org/) and obtained the miRNA-mRNA mapping via Cytoscape 3.9.0.

2.9. Flow cytometry

4 × 10⁵/mL NSCs were seed on the 12-well plates and transfected with different 100 nmol/L miRNAs, including miR-NC, miR-151-3p, miR-30b-3p, miR-222-3p, miR-221-3p, miR-129-5p, or miR-155-5p by lipofectamine™ 3000 (Invitrogen) on Days 1 and 4 for 6 h in aforementioned neuronal differentiation medium. Then on Day 7, the cells were first washed with PBS, and trypsinized with 0.25% EDTA-free trypsin (Gibco, cat. 15050-065) for 5 min, and terminated by the addition of DMEM medium containing 10% FBS. Next, the cell suspension was centrifuged at 350×g for 5 min, discard the supernatant and fix cells in 0.5 mL/tube Fixation Buffer (BioLegend Cat. No. 420801) in the dark for 20 min at R.T. Then resuspend fixed cells in Intracellular Staining Perm Wash Buffer (BioLegend, Cat. No. 421002) and centrifuge at 350×g for 8 min and repeated this step twice. Next, resuspend permeabilized cells in Intracellular Staining Perm Wash Buffer and add the Alexa Fluor 488 anti-Tuj 1 antibody (BioLegend, Cat. No. 801203) for 20 min in the dark at R.T. Finally, washed twice with Intracellular Staining Perm Wash Buffer and centrifuge at 350×g for 5 min, and suspend in 350 μL PBS for flow cytometry analysis.

2.10. Cellular internalization of Dual-EV

Human umbilical vascular endothelial cells (HUVECs) were used to assess the in vitro targetability of Dual-EV as in previous studies^31,32. EVs were first stained with Dil-red, and 0.25 μmol/L EVs, DA7R-EV, or Dual-EV was incubated with HUVEC at 37 °C for 4 h. Flow cytometry assay was carried out to determine the percentage of Dil-red-positive cells. First, HUVECs were washed with PBS and trypsinized to collect the HUVEC, and then cells were suspended in PBS and applied for flow cytometry.

For qualitative analysis, HUVECs were cultured in a 24-well glass-bottom culture plate (NEST Biotechnology Co., Ltd.) at 37 °C for 12 h, cells were incubated with different EVs medium for 4 h. HUVECs were then washed twice, and fixed with 4% paraformaldehyde for 15 min and stained with DAPI. Fluorescent images were captured by CLSM.

2.11. tMCAO mice model

All procedures involving animal usage were performed strictly and carefully according to the guidelines approved by the Institutional Animal Care and Use Committee (IACUC) of Shanghai Jiao Tong University with permission number (Bioethics 2,012,022).

Adult male ICR mice (30–35 g) were used to construct tMCAO based on our previous reports^33,34. Briefly, the common carotid artery, internal carotid artery, and external carotid artery were carefully isolated. A silicon-coated 6-0 nylon suture was inserted from the external carotid artery to the internal carotid artery and reached the origin of the middle cerebral artery. Blood flow was monitored using a Laser Doppler flowmetry. The success of occlusion was determined when cerebral blood flow was decreased to 80% of baseline. Reperfusion was carried out via withdrawing the suture 90 min after MCAO, and the blood flow restored to 70% of baseline was considered as an indicator of successful reperfusion. In this study, 83 ICR mice were used to establish the tMCAO model in our study, 15 mice died after 90-min tMCAO, so the mortality was 18.07%.

2.12. In vivo biodistribution of Dual-EV

To investigate the site-specific delivery of Dual-EV in tMCAO model mice, 24 h after reperfusion, different Cy5.5-tagged EVs, including EVs, DA7R-EV, and Dual-EV at a concentration of 200 μg were separately intravenously administrated into the model mice. 2 or 6 h later, the mice were sacrificed and perfused subsequently with 0.9% saline and 4% paraformaldehyde. The organs were then harvested for NIRF imaging using an IVIS Spectrum image instrument (PerkinElmer, USA). And the ROI analysis was conducted using the Living Image software for semi-quantitative analysis.

Additionally, the immunofluorescence assay was performed after intravenously injecting different EVs (EV, DA7R-EV, Dual-EV) for 2 or 6 h. The brains were harvested and stained with CD31, GFAP, MAP2, or IBA-1 to assess the co-localization of Dual-EV with endothelial cells, astrocytes, neurons, and microglia.

2.13. Neurobehavioral tests

The neurobehavioral experiments were conducted to assess the effect of Dual-EV on neurological function recovery as previously described^35,36. One day post-tMCAO, mice were divided into 4 groups (n = 8), including PBS, M2-EV, DA7R-EV, and Dual-EV groups at random. EVs (20 μg/mouse) were intravenously injected on Days 1, 3, 5, 7 post tMCAO. The modified Neurological Severity Score (mNSS), rotarod test, and hanging-wire test were conducted by an investigator blinded to the experiment to evaluate the stroke severity, motor coordination, endurance, and muscular strength of mice. For mNSS evaluation, the mice were graded on a scale ranging from 0 to 14, the lower the score, the better is the performance. For the rotarod test³⁷, mice were placed on a rotating rod, and the duration of mice on the rod was recorded. The mean duration was assessed by repeating 3 times. For the hanging-wire test, the mice's forelimbs were placed on a wire to record the times of reaching either end of the wire and the times of falling within 180 s. The basic score was 10, if the mice reached the end, one point was plus, if the mice fell, one point was minus.

2.14. Infarct volume assessment

Mice were euthanized on Day 14 post tMCAO operation. Brains were harvested, coronally sliced into 30 μm, and stained with 0.05% Cresyl violet acetate. The area of the ipsilateral and contralateral hemispheres was measured using ImageJ software. Infarct volume was calculated by subtracting the area of the ipsilateral hemisphere from that of the contralateral hemisphere, and multiplied by the section interval thickness according to the Eq. (1):

V = Σ h/3[ΔS_n+(ΔS_n × ΔS_n+1)^1/2+ΔS_n+1] (1)

where V is the volume, h is the distance between the two adjacent sections, ΔS_n and ΔS_n+1 are the areas between the two adjacent sections.

2.15. Immunofluorescence assay

According to the previous study³⁸, brain sections were permeabilized using 1% Triton-X for 10 min and blocked with 1% BSA for another 60 min at R.T., and incubated with the primary antibodies, DCX (1:200) or Nestin (1:200), at 4 °C overnight. Then the secondary antibodies were used AlexFluor 488 donkey anti-rabbit (1:500) for DCX, and AlexFluor 488 donkey anti-mouse (1:500) for Nestin. The statistical analysis was conducted by calculating the intensity of fluorescence density of each group section with three random visions in SVZ or striatum.

2.16. Statistical analysis

All results were shown as mean ± standard deviation (SD). Statistical comparison among multiple groups was analyzed by one-way ANOVA with Tukey's multiple comparisons test. The statistical comparisons were performed using GraphPad Prism 8 (GraphPad Software, California, USA) and a P value less than 0.05 was considered as statistical significance.

3. Results and discussion

3.1. Preparation and characterization of Dual-EV

To collect the M2-phenotype microglia-derived EVs, BV2 microglial cells were induced with 20 ng/mL IL-4 to polarize them into the M2 phenotype. The M2 phenotype markers ARG and CD206 were examined using immunofluorescence and Western blot. As shown in Fig. 2A and B, increased expression of ARG and CD206 was detected in the IL-4 treated BV2 microglia. M2-EVs were then isolated from the M2-BV2 conditioned medium for further modification.

Figure 2.

Isolation, modification, and characterization of Dual-EV. (A) BV2 microglia were induced with 20 ng/mL IL-4 for 48 h or not, and then double immunostaining with IBA1 (microglial marker)/CD206 (M2 marker), or IBA1/ARG (M2 marker). Scale bar = 50 μm. (B) Western blotting assay to detect the expression of CD206 and ARG in BV2 cells induced with/without IL-4. (C) Schematic design of dual-modification of DA7R and SDF-1 on the surface of M2-EVs by copper-free click chemistry method. (D) Fluorescence images of Dual-EV. EVs were stained with Dil-red, DA7R was labeled with FITC and SDF-1 was marked with Cy5.5. Scale bar = 0.5 μm. Size distribution was detected by NTA and TEM images of unmodified EVs (E) and Dual-EV (F). Scale bar = 100 nm. (G) Western blot assay of EVs markers, TSG101 and CD63, from EVs and Dual-EV. The cell lysis was set as a negative control.

To modify EVs with DA7R peptide and SDF-1 protein, a copper-free click chemistry reaction was proposed (Fig. 2C). First, DBCO-PEG₄-NHS was covalently conjugated with amino groups of EV proteins to introduce DBCO groups on the surface of EVs. Then, the DBCO-EVs were incorporated with SDF-1-N₃ and DA7R-N₃ to form stable triazole linkages. To verify the successful modification of SDF-1 and DA7R on EVs, three different fluorescent dyes were used to mark: Cy5.5 for SDF-1, FITC for DA7R, and Dil-red for EVs. As shown in Fig. 2D, these three dyes were almost completely merged, indicating SDF-1 and DA7R peptides were co-localized into EVs. Besides, we quantified the conjugation of SDF-1 and DA7R onto the EVs using the SDF-1 Elisa kit and the fluorescence quantification of DA7R-FITC. It turned out that an average 1 mg/mL modified EVs were conjugated with about 45.4 μg/mL DA7R and 9.46 μg/mL SDF-1.

Furthermore, we examined whether the EVs modification would affect their morphology and diameter using TEM and NTA. The results showed that both modified and unmodified EVs were in the cup shape with an average diameter of 120 nm, revealing that dual-modification did not significantly change the shape and diameter of EVs (Fig. 2E and F). Furthermore, results of Western blot assay displayed obvious expression of EV markers, TSG101, and CD63, in unmodified EVs and modified EVs, but not detected in M2-BV2 cell lysis (Fig. 2G).

3.2. Dual-EV promoted NSC neuronal differentiation in vitro

First, the effect of modified and unmodified EVs on NSC neuronal differentiation was examined, where EVs derived from microglia cells that were not polarized by IL-4 (M0-EV) were used as experimental controls. Tuj 1 could indicate the axon and cell body of immature neurons39, 40, 41 and did not express in NSCs^42,43, therefore, was widely used as the neuron-specific marker. As shown in Fig. 3, 9.4 ± 0.15% of NSC treated with Dual-M2-EV differentiated into neurons (Tuj1-positive cells), which held a similar portion (9.6 ± 1.3%) of unmodified M2-EV and was significantly higher than that of M0-EV treated group (3.1 ± 1.6%) and control group (3.8 ± 0.29%). Besides, after being treated with Dual-M2-EV, only a small portion (29.0 ± 4.3%) of NSC differentiated toward astrocytes (GFAP-positive cells), while above 80% of NSCs treated with blank medium or M0-EV differentiated into astrocytes (Fig. 3C). These results indicate that M2-EV can promote NSC neuronal differentiation, and dual-modification of DA7R and SDF-1 did not affect the differentiation effect of EVs on NSCs.

Figure 3.

Effect of M2-EV on NSC differentiation. (A) Representative fluorescent images of NSCs treated with blank medium, M0-EV, M2-EV or Dual-M2-EV for 7 days. M2-EVs promote NSC to differentiate into Tuj 1⁺ neurons (B) and suppress GFAP⁺ astrocyte formation (C). Data are presented as mean ± SD, n = 3; ∗∗∗P < 0.001. Scale bar = 100 μm.

To investigate the potential mechanisms of this effect on NSC differentiation, the miRNA-sequencing analysis was carried out and revealed that 10 miRNAs including miR-23a-5p, miR-151-3p, miR-30b-3p, miR-222-3p, miR-501-3p, miR-362-5p, miR-221-3p, miR-129-5p, miR-155-5p and miR-744-5p in Dual-M2-EV were upregulated compared to Dual-M0-EV (Fig. 4A). Among these upregulated miRNAs, miR-222 was proved to promote neurite outgrowth via inhibiting the neuronal growth inhibitor, chromosome 10 (PTEN) and activating the expression and phosphorylation of cAMP response element-binding protein (CREB)⁴⁴. miR-30 was identified as one of the most important miRNAs to directly convert of somatic cells into neurons⁴⁵. Also, miR-30 could protect neurons against ischemic injury via suppressing autophagy by negatively regulating Beclin-1, Atg-5 and LC3B^46,47. miR-362 could inhibit neurons from apoptosis via targeting Rho-related coiled-coil containing protein kinase 2 (ROCK2) in ischemic injury⁴⁸. Overexpressing miR-221 in stem cells from human deciduous teeth induced their differentiation into neurons through activating the Wnt/β-catenin pathway by binding to CHD8⁴⁹, and in PC12 cells also enhanced neuronal differentiation⁵⁰. In contrast, suppression of miR-221/222 cluster was required for neuronal differentiation of glioma cells⁵¹. Besides, miR-221 was shown to promote neural stem cell proliferation via inhibiting PTEN and activating the AKT pathway⁵². Also, miR-221-3p could inhibit neuron apoptosis via targeting CDKN1B⁵³ or activating transcription factor 3 (ATF3)⁵⁴. miR-129-5p was shown to modulate neuronal migration via targeting Fragile X Mental Retardation gene 1 (Fmr1)⁵⁵. miR-155-5p was identified as up-regulated during neural stem cell differentiation⁵⁶ and could promote neuron axon growth via activating cAMP/PKA pathway⁵⁷. In all, by crossing the potential target genes involved in the nervous system of the above 10 upregulated miRNAs, 2198 target genes were obtained (Fig. 4B).

Figure 4.

Mechamisms of Dual-M2-EV on NSC neuronal differentiation. (A) The heatmap of differentially expressed miRNA in Dual-M2-EV compared with Dual-M0-EV. (B) The network of miRNA-mRNA. The gray circle represented the miRNA, and the green circle meant the target mRNA. (C) Quantitative analysis and representative images by flow cytometry of NSC-differentiated Tuj 1⁺ cells induced by selected miRNAs. Data are presented as mean ± SD, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, ns meant non-significance.

Since miRNAs exert their functions through negatively regulating their target genes, 6 miRNAs among the 10 upregulated miRNAs that may be involved in negative regulation of NSC differentiation and neuron development were selected for further investigation, including miR-151-3p, miR-30b-3p, miR-222-3p, miR-221-3p, miR-129-5p, and miR-155-5p. These miRNAs were respectively transfected into NSCs in vitro to evaluate their effect on neuronal induction by flow cytometry analysis. As shown in Fig. 4C, the Tuj 1⁺ cells were significantly higher in the miR30b-3p, miR-222-3p, miR-129-5p, and miR-155-5p groups than the control miR-NC group, suggesting that these miRNAs in Dual-M2-EVs may exert the effect of inducing NSC to differentiate into neurons. Also, which miRNA is supposed to be the key factor needs further investigation.

3.3. Targetability and NSC recruitment effect of Dual-EV

To enhance the function of M2-EVs on nerve repair, the surface modification of EVs was modified with a targeting ligand to increase their accumulation in the ischemic region and promote NSC recruitment. Peptide ligands are widely applied in the construction of targeted nanocarriers to achieve CNS disease treatment58, 59, 60, which may also be applied in modifying the EVs to improve their targetability. DA7R peptide, as a high binding ligand of VEGFR2, is a retro-inverso peptide of d-amino acids which endowed it with enhanced proteolytic stability during circulation compared with natural l-peptide^61,62. Previous studies have proven that DA7R peptide held a high binding affinity with VEGFR2, highly expressed in HUVECs^62,63. Also, HUVECs are widely used as the cell lines to investigate the effect of EVs or miRNAs on endothelial cell function in stroke^64,65, thus the in vitro targeting ability of Dual-EV could be assessed using this cell line. First, it is proved that the FAM-labeled DA7R peptide could be effectively internalized by HUVECs (Supporting Information Fig. S1), suggesting that modification of DA7R on the EVs is supposed to enhance its uptake by HUVECs. To evaluate the effect of DA7R peptide on blood vessel targeting, different Dil-red-labeled EVs, DA7R-EV, and Dual-EV were separately incubated with HUVECs for 4 h. Then the level of EVs internalization was detected using the fluorescence microscopy and flow cytometry. Quantification of the confocal fluorescent images showed that EV uptake efficiency was increased from 78.3% to 89.5% when modified with DA7R peptide (Fig. 5A and B), suggesting the DA7R modified EVs exhibited higher target ability toward HUVECs. Surprisingly, the additional decoration of SDF-1 on EVs further enhanced the internalization level of DA7R-SDF-1-EV by HUVECs. This may be attributed to the C–X–C motif chemokine receptor 4/7 (CXCR4/CXCR7) expressed in HUVECs that both are SDF-1 receptors66, 67, 68.

Figure 5.

Internalization of different EVs by HUVECs and its recruitment effect on NSCs. (A) Representative confocal fluorescent images of HUVECs treated with 0.25 μmol/L EVs, DA7R-EV, or Dual-EV for 4 h. Scale bar = 10 μm. (B) Quantitative analysis using flow cytometry to quantify EVs internalized by HUVECs in the overlaid type. (C) Schematic illustration of NSC Transwell migration assay. (D) NSCs were seeded on the Transwell and were treated with blank medium, unmodified EVs, SDF-1-EV and Dual-EV for 48 h. The DAPI-stained NSCs represented these migrated on the lower side of Transwell membrane. Scale bar = 200 μm. (E) percentage of NSCs migration calculated by ImageJ software. Date are represented as mean ± SD for three random division. ∗∗∗P < 0.001. Scale bar = 200 μm.

Since SDF-1 was reported as a class of chemokine that could mediate cell migration, effectively inducing NSCs to migrate to the lesion site during nerve injury⁶⁹. The Transwell assay was conducted to assess whether SDF-1-modified EVs could still promote the migration of NSCs. As shown in Fig. 5C, NSCs were cultured in the upper chamber and treated with a blank medium, M2-EV, SDF-1-EV, or Dual-EV at an SDF-1 concentration of 100 ng/mL in the lower chamber. After 48 h of incubation, the NSCs on the lower side of the chamber membrane were stained with DAPI, and the cell number was counted and compared. In Fig. 5D and E, compared to the control group, a similar ratio of NSC migration was observed in the M2-EV group, while a remarkably higher ratio of NSC migration in the SDF-1-EV group (2.9-fold, P < 0.001) or Dual-EV (3.4-fold, P < 0.001). The above results indicate that the modification of SDF-1 endowed the EVs with the ability to enhance NSC migration. Moreover, the decoration of DA7R did not affect the migratory potential of SDF-1.

Besides, the Transwell assay was further carried out to prove whether the increased targeting ability of Dual-EVs could contribute to a better effect on NSC differentiation. First, the HUVECs were seeded on the upper chamber of Transwell and proliferated for 3 days. Then M2-EV and Dual-M2-EV were respectively added to the upper chamber and incubate to penetrate the HUVECs, 6 h later, the mediums in the lower chamber were harvested and applied to the seeded NSCs in another 24-well plate to conduct the differentiation assay by the immunofluorescent staining method. Finally, the mediums were changed every two days with freshly harvested medium containing M2-EV or Dual-M2-EV. As shown in Supporting Information Fig. S2, the Dual-M2-EV group significantly promoted the NSCs to differentiate into neurons (Tuj 1-positive cells) instead of astrocytes (GFAP-positive cells) when compared with both M2-EV and control groups, suggesting that modification with DA7R and SDF-1 helped EVs to target and penetrate the endothelial cells and adjust the function of NSCs in vitro.

3.4. Biodistribution of Dual-EV in tMCAO model mice

To assess the targetability of Dual-EV to the ischemic stroke brain in vivo, free EVs, DA7R-EV, and Dual-EV were first tagged with near-infrared dye Cy5.5, and then respectively systemically injected into the tMCAO mice via tail vein 1-day post tMCAO. Brains were dissected 2 h later for IVIS imaging. The typical lesion region was marked as an ipsilateral area and the matching non-ischemic region was termed the contralateral region (Fig. 6A). Obvious fluorescent signals were observed in Dual-EV and DA7R-EV compared to free EVs, and the calculated mean fluorescent intensity was also higher than free EVs (Fig. 6B and D). Furthermore, Dual-EV and DA7R-EV predominantly accumulated in the ipsilateral site (Fig. 6E). These may be due to the highly upregulated VEGFR2 in brain endothelial cells after stroke and the high binding affinity of DA7R with VEGFR2^70,71. These results indicate that DA7R modification remarkably promoted the targetability of EVs to the ischemic region. We examined the ex vivo fluorescent signals in the heart, liver, lung, spleen, and kidney. As shown in Supporting Information Fig. S3, compared to unmodified EVs, the distribution of DA7R-modified EVs increased in the liver but decreased in the kidney, which may also be attributed to the VEGFR2 expression in the liver endothelial cells⁷². However, there was no difference in EVs biodistribution in the heart, spleen, and lung. These results suggest that the clearance way of EVs may be changed after modification of DA7R. Besides, 6 h biodistribution results showed that a remarkable amount of engineered EVs including the DA7R-EV and Dual-EV still accumulated in the brain compared to the unmodified EVs at 6 h (Supporting Information Fig. S4).

Figure 6.

Ischemic brain-targeting ability of Dual-EV in vivo. (A) a photo of a brain harvested from tMCAO mice illustrated the lesion region (left hemisphere: ipsilateral region, right hemisphere: contralateral region). (B) Representative ex vivo IVIS images of brains harvested from model mice which respectively received free EVs, DA7R-EV and Dual-EV labeled with Cy5.5 2 h after administration. (D) Semi-quantitative analysis of fluorescent intensity in the ischemic brain. (E) The ratio of fluorescent signals in the ipsilateral to the contralateral area. Representative images (C) and statistical analysis (F) of different Dil-red-labeled EVs (red) co-localized with CD31-positive vascular endothelial cells (green) in the ischemic region 2 h after administration. Data are presented as mean ± SD, n = 3, ∗P < 0.05, ∗∗∗P < 0.001. Scale bar = 25 μm.

3.5. Localization of Dual-EV in the ischemic lesion site

To investigate the co-localization efficiency of Dual-EV to blood vessels in the brain after stroke, Cy5.5-marked free EVs, DA7R-EV, or Dual-EV were intravenously injected into the stroke mice 24 h after tMCAO. Then the brains were harvested for immunofluorescence assay 2 h after injection. As revealed in Fig. 6C and F, evident Dual-EV accumulated in the ipsilateral site and about 58% were co-localized with CD31-marked blood vessels, significantly higher than unmodified EVs (~14.9%) and similar to DA7R-EV (~48.8%). Interestingly, after 6 h of blood circulation, a distinct decrease of Dual-EV co-localized with blood vessels was observed, but appeared in the brain cells nearby instead (Supporting Information Fig. S5). These results verified that Dual-EV could effectively target endothelial cells, penetrate blood vessels, and accumulate in the ischemic hemisphere. According to previous reports^73,74, The fate of EVs that enter cells is that some internalized EVs can escape lysosomal degradation by being re-secreted either via the early endocytic recycling pathway or by fusion of early multivesicular endosomes with the plasma membrane. Therefore, in this study, EVs could be taken up by endothelial cells and then released into brains to affect NSCs. It is also reported by Tian et al.⁷⁵ that engineered EVs could target the blood vessels and delivery the curcumin to brain tissues in ischemic stroke to inhibit the inflammatory. The results of our study are consistent with these previous studies.

Furthermore, the internalization of Dual-EV nanocarriersby other types of brain cells was examined, including microglia (IBA1⁺ cells), neurons (MAP2⁺ cells), and astrocytes (GFAP⁺ cells). As shown in Supporting Information Fig. S6, Dual-EV nanocarrierscould also be uptaken by all these three types of cells, which may be due to the different EVs uptake mechanisms including receptor-mediated endocytosis, lipids rafts, macropinocytosis, or micropinocytosis⁷⁶.

3.6. Dual-EV nanocarrier treatment promoted in vivo neurogenesis post-stroke mice

To evaluate the therapeutic efficacy of the tMCAO mice treated with Dual-EV, we performed the behavioral tests on Days 1, 3, 5, 7, and 14 after stroke and examined the brain infarct volumes (Fig. 7A). As shown in Eq. (1) and Fig. 7B and C, the Dual-EV group profoundly reduced brain atrophy volume to 5.26 ± 1.16 mm³ in comparison with DA7R-EV (11.2 ± 2.05 mm³), M2-EV (19.6 ± 4.4 mm³), and PBS (24.9 ± 4.10 mm³) groups. This indicates that EVs functionalized with DA7R and SDF-1 peptides greatly potentiated their therapeutic effect on ischemic stroke. For the 90-min tMCAO model, brain atrophy volume after 14 days is relatively apparent. Although there is a glial scar, the ischemic brain tissue cannot be completely replaced by this proliferated glial scars⁷⁷. As shown in Fig. 7D and E, the neurological score of the Dual-EV group was gradually decreased and finally significantly lower compared to PBS, EV groups on Day 14 post-stroke. Besides, both Dual-EV and DA7R-EV groups showed better performance in rotarod and hanging-wire tests, as revealed by a prolonged latency to fall and higher hanging-wire scores compared with PBS, M2-EV, and DA7R-EV groups (Fig. 7F‒G, Supporting Information Fig. S7). These results suggest that Dual-EV treatment enhanced neurological function recovery in tMCAO mice.

Figure 7.

Treatment of Dual-EV decreased brain atrophy volumes and improved neurobiological recovery induced by an ischemic stroke. (A) Schematic illustration of ischemic stroke treatment. (B) Brain sections stained with cresyl violet treated with Sham, PBS, EV, DA7R-EV, or Dual-EV on Days 1, 3, 5 and 7 at a dosage of 20 μg EV/day after tMCAO. The dashed line indicated the infarct area, was first drafted along the outer edge of the contralateral hemisphere, then flipped vertically to the ischemic hemisphere. (C) Infarct volumes of brains in each group (n = 4–5). (D) Neuroscores evaluation on Days 1, 3, 7 and 14 post-reperfusion. (E) mNSS of model mice in Day 14, (F) Rotarod test (G) Hanging-wire test. n = 8, data are presented as mean ± SD, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001.

To further verify the effect of dual-modified EVs on NSC recruitment and neurogenesis, the immunofluorescence assay was performed by staining Nestin and DCX, which indicates neuronal precursor cells (NPCs) and migrating immature neurons. As shown in Fig. 8A and C, there was a significant increase of Nestin-positive cells in the SVZ region treated with Dual-EV compared with other groups, suggesting that Dual-EV treatment stimulated the activation of NPCs. Furthermore, the number of DCX-positive cells was remarkably increased in the infarct areas, including SVZ and striatum regions in comparison with other groups (Fig. 8B, D and E). Moreover, the migrating distance of migrating immature neurons from SVZ to striatum in the Dual-group was greatly lengthened than in other groups (Fig. 8F), revealing that Dual-EVs could effectively recruit NSCs and induce their neuronal differentiation. These results demonstrated that EVs with modification of DA7R and SDF-1 enhance its accumulation in the ischemic area, recruitment ability of NSCs, and finally promote neurogenesis.

Figure 8.

Treatment of Dual-EV enhance neurogenesis in tMCAO mice. Representative fluorescent images of (A) NPCs (Nestin-positive cells, green); (B) Migrating immature neurons (DCX-positive cells, green) in groups treated with PBS, EVs, DA7R-EV or Dual-EV on Day 14 post-stroke in the SVZ (top) or striatum (bottom) region of ischemic mice brains. DAPI (blue): cell nuclei. Scale bar = 50 μm. (C) Relative integrated density (IntDen) of Nestin-positive cells in SVZ region. Data are presented as the ratio of average IntDen of PBS group. Quantitative analysis of DCX⁺ cells in the SVZ (D) and striatum (E) regions of the ipsilateral ventricle. (F) DCX cells migration distance. Data are presented as mean ± SD, n = 3; ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001.

In this study, engineered M2-EV modified with SDF-1 and DA7R were developed via copper-free click chemistry for neurogenesis in stroke. Despite the inspiring results, some limitations still exist for future clinical translation. First, the M2-EV was proved to hold the ability to induce NSCs to differentiate into neurons. Though bioinformatic analysis has been conducted to predict the possible mechanisms of neuronal differentiation, further experiments still need to be performed to figure it out. Secondly, the recycling rate of the engineered EVs was about 60%, how to increase the yield and avoid the loss of engineered EVs during the modification process is another issue that requires to be addressed. Third, the EVs tend to be quickly cleared during blood circulation, which would restrict their therapeutic capabilities. This may be solved by utilizing PEGylation⁷⁸ or cell membranes or polymer hybrids like DNA tethers⁷⁹ to engineer EVs to modulate the pharmacokinetic property. Nevertheless, our work paves the way toward the rational design of functionalized EVs for CNS injury treatment and more broadly related nanocarriers and Nature-derived or biologically modified nanomaterials for improving human health80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100.

4. Conclusions

Overall, M2 microglia-derived EVs were for the first time proven to hold the potential of inducing NSC to differentiate into neurons. Moreover, we have engineered DA7R peptide and SDF-1 cytokine-modified M2-EVs via copper-free click chemistry reaction to endow M2-EVs with the ability of vascular endothelial cells-homing and NSCs recruitment, more effectively realizing the nerve function repair and neuron regeneration after ischemic stroke. Our work paves the way toward endogenous neuron regeneration therapy and rational design of functionalized EVs for CNS injury treatment and more broadly, extracellular vesicles/peptide/chemokine and related nanocarriers and biologically modified nanomaterials for improving human health.

Author contributions

Huitong Ruan, Yongfang Li, Xingcai Zhang, Yaohui Tang, and Wenguo Cui proposed the ideas and designed the study. Huitong Ruan, Yongfang Li, Cheng Wang, Yixu Jiang, Dandan Zheng, and Jing Ye performed the experiments, analyzed the data, and designed figures. Yulong Han, Yiwei Li, and Ming Guo participated in the discussion. Huitong Ruan wrote the manuscript. Yongfang Li, Yaohui Tang, Xingcai Zhang, Wenguo Cui, Yulong Han, Yiwei Li, Gang Chen, Guo-yuan Yang, Lianfu Deng and Ming Guo revised the manuscript. All authors have read and approved the final manuscript.

Conflicts of interest

The authors have no conflicts of interest to declare.

Appendix A. Supplementary data

The following is the Supplementary data to this article: