Chimeric CNS-Targeting-Peptide Engineered Exosomes for Experimental Allergic Encephalomyelitis Therapy

Background: Multiple sclerosis (MS), an in�ammatory disease of the central nervous system (CNS), leads to demyelination, neuronal injury, and loss of white matter, yet still can't be cured. Exosomes are double-layered membrane vesicles of 30–200 nm in size, which can easily penetrate the blood–brain barrier (BBB). Exosomes derived from umbilical cord mesenchymal stem cells exosomes (UMSC-exos) has been shown to treat experimental autoimmune encephalomyelitis (EAE) through the action of anti-inammatory and immunomodulatory, but its clinical translation has been hampered by their ine�cacious accumulation in CNS. Therefore, we developed a TAxI-peptide-chimeric UMSC-exos termed TAxI-exos for CNS-specic accumulation and curative effect in EAE. Methods: We used an EAE model in vivo, and actived T cells and BV-2 cells models in vitro. After two immunizations to establish the EAE model, UMSC-exos, TAxI-exos or DiR labeled exosomes were administered to EAE mice EAE mice for one dose (150μg) before the peak at day 15. On day 30, the mice were sacri�ced to collect spinal cords, spleens, and blood for analysis of demyelination, in�ammation, microglia, the proportions of T-cell subsets, and the expression of in�ammatory cytokines. In vitro, for immune mechanism analyses, PBMCs and splenocytes isolated from healthy C57BL/6 mice, were activated and incubated with 0.15mg/mL UMSC-exos or TAxI-exos. Activated BV-2 cells were used to explore the targeting-ability and polarization-regulating ability of UMSC-exos and TAxI-exos. Results: As expected, TAxI-exos had signi�cant curative effects in EAE mice compared with UMSC-exos via an enhanced targeting-ability. The treatment alleviated in�ammation, facilitated microglial cell polarization from M1 to M2, reduced the proportions of T-cell subsets, increased the expression levels of


Background
MS is an autoimmune disease of the central nervous system (CNS), which leads to demyelination, neuronal injury, loss of white matter, and eventually leaves patients permanently disabled and blind, who are always young women [1].More than 2 million people suffer from MS worldwide, with approximately 30 thousand patients in China [2].Currently, 19 disease-modifying therapies (DMTs) have been approved for treatment of MS [3].However, they are only effective for some patients in relapsing remitting multiple sclerosis (RRMS) [4], and only one drug was recently approved for secondary progressive multiple sclerosis (SPMS).Additionally, long-term treatment leads to tolerance, and increasing the types and dosages of medicines [5].Therefore, it is critical to develop a novel and effective therapeutic drug.
The experimental autoimmune encephalomyelitis (EAE) is one of the most ideal animal models in MS research.The clinicopathological features of the model are similar to MS patients [6].At the onset of MS and EAE, in ammatory cells in ltrate in brain and spinal cord tissues, and release cytokines along with damaging of the BBB [7].T helper (Th) 1 cells [8] and Th17 cells [9] play the main roles in the initiation of EAE.When IL-6 and IFN-γ are locally present, naive CD4+ T cells are selectively induced to differentiate into Th1 and Th17 cells [10] to activate macrophages/microglia, which contributes to cellular immunity by secreting IFN-γ, IL-2, TNF-α, and IL17A [11].An M1/M2 microglial cell polarization imbalance is also a critical cellular mechanism of MS pathogenesis.M1 microglia produce pro-in ammatory cytokines (TNFα and IL-6) to accelerate the EAE course, while M2 microglia have an anti-in ammatory effect [12].
Mesenchymal stem cells (MSCs) are a type of highly proliferative multipotent cells that can be isolated from various tissues, have a great potential in the elds of biotherapy, immunoregulation and tissue regeneration [13].At early stages of EAE, an underlying mechanism is the ability of MSCs to induce a CD4 + CD25 + Foxp3 + regulatory T cells (Tregs) population and suppress the differentiation processes of Th1 and Th17 cells [14].MSCs exosomes, a type of lipid bilayer membrane vesicles, with a diameter of 50-200 nm, harbor many bioactive factors such as proteins, lipids, and microRNAs secreted from their parent cells [15].Because of their excellent BBB-penetrating ability, humoral stability, and low immunogenicity compared with parent stem cells, umbilical cord MSCs exosomes (UMSC-exos) have become a treatment option for EAE.However, many studies have con rmed that intravenously injected exosomes are mainly distributed in the liver or spleen, and have very low tropism to other organs, even the CNS in EAE [16].Thus, how to enhance the CNS-targeting ability of UMSC-exos is required before using UMSC-exos for MS therapy.
The SACQSQSQMRCGGG peptide (targeted axonal import, TAxI) is a novel targeting ligand with high a nity for axons, which delivers functional proteins to spinal cord motor neurons after peripheral administration [17].Accordingly, we developed CNS-targeting TAxI-exos by modifying UMSC-exos with the TAxI peptides for MS therapy.In this study, we assessed the therapeutic effects of TAxI-exos and UMSCexos in EAE mice by evaluating differences in the targeting ability, clinical symptoms, histology, microglial polarization, proportions of T cell subsets, and expression levels of in ammatory factors.Furthermore, we investigated the underlying mechanisms of these effects.

Methods
EAE induction and exosomes treatment 8 to10-week-old female C57BL/6 mice (18-22 g) were purchased from Wu's Experimental Animals Inc. (Shanghai, China).Animals were cared in accordance with the guidelines of the Animal Ethics Committee of Xiamen Medical College ( Approval ID: 20190302007).
UMSC-exos or TAxI-exos (150 μg or 2×10 10 particles) was intravenously injected (i.v.) once before the peak of the disease.Each mouse was assessed daily by a clinical score that ranged from 0 to 5 [18]: No clinical disorder in motor functions compared with non-immunized mice (0); partial tail weakness or slight loss of muscle tone (0.5); limp tail (1.0); limp tail and hind legs paralysis with gait abnormality (2.0); limp tail and complete paralysis of hind legs (3.0); limp tail and complete paralysis of hind legs and partial front leg paralysis (3.5); front and hind limb paralysis (4.0); moribund state (5.0).Mice were monitored for at least 30 days after immunization.

Cells
Human umbilical cords were provided by Zhongshan Hospital a liated to Xiamen University as the source of human umbilical cord tissue-derived MSCs (UC-MSCs) and UMSC-exos used in this study.UC-MSCs were isolated from discarded umbilical cords and prepared by following an established protocol [19].
For in vitro experiments, microglial cell line BV-2 was cultured in DMEM-high glucose (Meilunbio) supplemented with 10% FBS.The cells were incubated at 37 °C in a humidi ed atmosphere with 5% CO 2 .

Isolation of exosomes
As described above, conditioned medium from cultures of UC-MSCs was collected and centrifuged at 300 × g for 10 min and then 2000 × g for 20 min at 4 °C to remove cell debris pellet and larger vesicles.Supernatants were transferred to centrifuge bottles (Centrifuge bottles polypropylene with caps 29 × 104 mm, Beckman Coulter, CA, USA) and centrifuged in a high speed refrigerated centrifuge (Avanti JXN-26 centrifuge, Beckman Coulter) for 20 min at 12,000 × g at 4 °C (Type JA-25.50 rotor, Beckman Coulter), to remove microvesicles.The supernatants were ltered through a 0.22-μm lter (Millipore, MA, USA) and then carefully collected in 40 PA tubes (26 × 90 mm, TKY, Japan) and ultracentrifuged for 1.5 h in a Type R30AT rotor at 110,000 × g at 4 °C (High-speed refrigerated centrifuge CR30NX, Himac).The exosomes (UMSC-exos) were reconstituted in PBS and stored at −80 °C.For experiments, exosomes prepared from three different UC-MSCs batches were pooled to reduce any potential effect of batch-to-batch variations on molecular markers.
Nanoparticle tracking analysis (NTA) NTA was performed using a nanoparticle tracking analyzer (ZetaView PMX120, Particle Metrix, Bavaria Germany).UMSC-exos and TAxI-exos suspended in PBS were diluted 1000-fold prior to analysis.The precise diameter of exosomes were characterized by NTA that tracks the particle movement to obtain size information.A 60-s video was recorded and subsequently analyzed using NTA software.

Biodistribution of exosomes in vivo
Wild-type (WT) and EAE mice were used to access the biodistribution of DiR labeled UMSC-exos or TAxIexos.Mice (n = 3) were intravenously administered with 150 μg (or 2 × 10 10 particles) DiR-labeled UMSCexos or TAxI-exos.Fluorescence images of the mouse body were recorded and analyzed by the IVIS® Lumina XRMS III and analysis software (Perkin-Elmer, MD, USA) at 24 and 48 h after injection.

Histopathological examination
For immunohistochemical staining, at the time of sacri ce, 4% paraformaldehyde-xed spinal cords were removed and cut into 4 µm thick sections.At least six sections were examined for each mouse.The sections were incubated with a primary antibody against Iba-1 (1:200 dilution; Huabio, Hangzhou, China) at 4 °C overnight.Then, the samples were incubated with a secondary antibodies (HRP conjugated goat anti-rabbit IgG, 1:500 dilution; Servicebio, Wuhan, China) for 1 h at RT.The sections were observed under an Axio Imager (Zeiss, Aalen, Germany).Quanti cation of Iba-1 staining was performed in three sections each from three mice of one group using ImageJ-Fiji IHC Toolbox plugins (The University of Nottingham, UK).
Luxol fast blue (LFB, Leagene®, China) staining determines myelin differentiation.Spinal cord sections went through a series of staining steps in accordance with the manufacturer's instructions.Quanti cation of myelinated axon areas was performed using three sections each from three mice of one group.All spinal cord sections were observed under the Axio Imager (Zeiss) and analyzed using ImageJ-Fiji.
For hematoxylin and eosin staining, spinal cord sections were stained and went through a series of ethanol washes, and analyzed under the Axio Imager (Zeiss).In ammatory cell in ltration was scored as follows [12]: no in ltrating cells = 0; a few scattered in ltrating cells = 1; organization of in ammatory cell in ltration around blood vessels = 2; extensive perivascular cu ng with widespread cell in ltration = 3.

Protein quanti cation
Exosomes were diluted in PBS and were analyzed using a microBCA protein assay kit (Sangon, Shanghai, China).Tissues and cells were lysed in cold RIPA buffer (Meilunbio) with a protease inhibitor cocktail (Meilunbio).After centrifugation at 4 °C for 10 min, supernatant was collected and subjected to quanti cation using a BCA protein assay kit (Solarbio, Beijing, China) and western blot.The plate was analyzed using a SpectraMax i3x and software (Molecular Devices, SV, USA).
Finally the washed membranes were incubated with Fg super sensitive ECL luminescent reagent (A nity) and analyzed using a Chemidoc MP (Bio-Rad Laboratories, CA, USA).

Flow cytometry (FCM)
UMSCs were harvested from culture plates and washed with 1% BSA in PBS.MSCs were then incubated with PE-labeled antibodies against CD73, CD90, or CD105, an APC-labeled anti-CD44, and FITC-labeled anti-HLA-DR or an isotype as the control (BioGems, PeproTech) for 40 min at 4 °C.The samples were analyzed using a CytoFLEX LX (Beckman Coulter).Data were processed using FlowJo Software v.10 (Tri Star, Ashland, OR, USA).
For ow cytometry, the cells isolated from spleens or spinal cords of mice were washed with PBS.Cells were suspended and ltered through a 200-mesh sieve, prior to centrifuging at 500 × g for 5 min to collect.Peripheral blood mononuclear cells (PBMCs) were isolated from blood using a Mouse Peripheral Lymphocyte Separation Kit (Meilunbio) for medium density gradient centrifugation and stained for markers of T cell subsets.The following antibodies were used: anti-IL-17-PE, anti-IFN-γ-PE, anti-CD4-FITC, anti-CD25-PE, anti-Foxp3-APC or an isotype as control (eBioscience, CA, USA).

T-cell proliferation assay
PBMCs were isolated from blood samples of healthy female mice using density gradient centrifugation.

Treg induction assay
Splenocytes from healthy female mice were ltered through a 200-mesh sieve, and then centrifugated at 500 × g for 5 min.Suspensions were prepared and stimulated with anti-mouse CD3 SAFIRE Puri ed (3 μg/mL) and murine recombination IL-2 (30 ng/mL).The splenocytes further cultured with UMSC-exos or TAXI-exos (30 μg/well) or unconditioned medium as the control.Tregs are marked by expression of FOXP3 and high level of CD25 [23].After 3 days, cells were stained for analysis of CD4 + CD25 + FOXP3 + T cells by ow cytometry using the CytoFLEX LX and FlowJo V10.
For the cytometric bead array (CBA) assay, 50 μL of PBMC culture supernatant or 10 μL of a mouse serum sample were incubated with 50 μL assay beads and 50 µL of reagent using a CBA Mouse Th1/Th2/Th17 Cytokine Kit (Cat No. 560485, BD, USA).Results were analyzed with FCAP Array Software (BD) and then reported as the mean uorescence intensity (MFI).

Quantitative reverse transcription-PCR (qRT-PCR)
Total RNA was isolated from spinal cord (lumber enlargement) tissue using Column Animal RNAOUT reagent (Tiandz, Beijing, China) in accordance with the manufacturer's instructions.Reverse transcription of total RNA to cDNA and quantitative real-time PCR were performed using a One Step qRT-PCR SYBR Green Kit (Vazyme, Nanjing, China).The primers (Invitrogen, Suzhou, China) are listed in Table 1.The results were analyzed by the 2−ΔΔCt method and represented as fold changes, normalized to GAPDH.
Table .1 Sequences of primers used for qPCR.

Statistical analysis
Each experiment was repeated at least three times.All data were represented as the mean ± SEM using GraphPad Prism 5 software (GraphPad Inc., USA).Comparisons of two groups were performed by the ttest.One-way ANOVA with Bonferroni's or Newman-Keuls post-hoc tests were applied to compare multiple groups.p-values of less than 0.05 were considered statistically signi cant.

Identi cation of UMSCs
Before subcultured UMSCs were used in experiments, they were con rmed by MSCs standards [24].When primary cells had migrated from umbilical cord tissues and adhered to the bottom of the culture ask, UMSCs gradually began to display a broblast-like and spindle-shaped morphology, which gathered together in a circinate appearance, as observed under an inverted phase contrast microscope (Fig. 1a).Surface markers on P3-P5 UMSCs were detected by FCM, which were positive for CD44, CD73, CD81, CD90, and CD105, but negative for HLA-DR (Fig. 1b).The purity of MSCs was up to 95% and no obvious morphological changes were observed.

Isolation and identi cation of UMSC-exos
A large number of UMSCs were cultured in medium without FBS for 48 h.Next, we puri ed exosomes from culture supernatants using differential centrifugation [25].Exosome pellets were reconstituted in PBS and stored at −80 °C before use.Isolated UMSC-exos were characterized by WB, TEM, and NTA.CD9, CD63, CD81 and TSG101 were enriched in UMSC-exos, which was con rmed by WB (Fig. 1c).An absence of Calnexin (endoplasmic reticulum marker) suggested that there was no contamination from endoplasmic reticulum-derived vesicles (Fig. 1c).TEM showed that UMSC-exos had a spherical shape and bilayer membrane structure with a diameter of approximately 130 nm (Fig. 1d).Furthermore, the precise sizes of exosomes were characterized by NTA.A screenshot of the NTA video is presented in Figure 1e.The average size of exosomes was determined to be approximately 116.6 nm in diameter (Fig. 1f).Taken together, these data demonstrated successful isolation of UMSC-exos that expressed several common exosomal markers.
TAxI peptides were identi ed quantitatively and qualitatively by high performance liquid chromatography and mass spectroscopy.The purity was 95.05% and the molecular weight was 1880.2 (Fig. 2a, b).The structures of DSPE-PEG2000-TAxI were identi ed by infrared spectroscopy and the nuclear magnetic resonance (Fig. 2c, d).
To obtain applicative TAxI-exos, UMSC-exos were modi ed with DSPE-PEG2000-TAxI by a post-insertion method, which were named TAxI-exos (Fig. 2e).The morphology of UMSC-exos and TAxI-exos were assessed by TEM.They were both round spherical vesicles and TAxI-exos had a larger particle diameter than UMSC-exos (Fig. 2f).The average diameters and concentrations of UMSC-exos and TAxI-exos were analyzed by NTA, which were 138.4 and 150.6 nm (Fig. 2g).These results indicated that UMSC-exos maintained integrity after the TAxI modi cation and had a larger particle diameter because of the TAxI peptides insertion.
Targeting capabilities of UMSC-exos and TAxI-exos in vivo and in vitro After preparation of TAxI-exos, we con rmed the targeting ability of TAxI-exos in vivo.First, female C57BL/6 mice were immunized twice on days 0 and 2, and approximately 15~20 days later, EAE mice displayed the peak of the disease, which showed complete paralysis of the tail and hind limbs with a attened posture.Next, each mouse was assigned a clinical score every day, which ranged from 0 to 5 in the monitoring stage (Fig. 3a).UMSC-exos and TAxI-exos were labeled with DiR.Then, 150 µg (or 2 × 10 10 particles) DiR-labeled UMSC-exos or TAxI-exos were injected (i.v.) into EAE mice at day 15 (n = 3).The intensity of uorescence signals was recorded at 24 and 48 h by a live imaging system (Fig. 3b).The results show that the uorescence intensity in the brain region of TAxI-exo-treated mice was signi cantly higher than that after UMSC-exos treatment and DiR-labeled UMSC-exos could not be detected after 24 or 48 h (Fig. 3c).In the TAxI-exos group, no signi cant difference was observed in the uorescence intensity between 24 and 48 h, which was the same in the UMSC-exos group (Fig. 3c).Additionally, in EAE animals, the uorescence intensity in the TAxI-exo-treated group had a slightly stronger trend, whereas in WT animals, this trend was weakened between 24 and 48 h (Fig. 3c).These results suggested that TAxI peptides enhanced the blood stability and target-site residential capability of UMSC-exos.
Next, we assessed the internalization ability of TAxI-exos in target cells of the CNS in vitro.To this end, we chose BV-2 cells for CFDA SE staining and incubated with 0.4mg/mL (or 5×10 10 particles mL −1 ) DiRlabeled UMSC-exos or TAxI-exos and then the uorescence signal was observed under a uorescence microscope (Fig. 3d).Stronger signals occurred in BV-2 cells incubated with TAxI-exos than in those with UMSC-exos (Fig. 3e), which suggested that TAxI-exos had a signi cant BV-2 cell-targeting ability in vitro.
In this study, we applied in vitro and in vivo models to speci cally demonstrate the CNS-targeting properties of TAxI-exos that accumulated in CNS sites more e ciently than UMSC-exos.
UMSC-exos and TAxI-exos improve clinical functional signs in EAE We next examined the therapeutic effects of UMSC-exos and TAxI-exos in a well-established EAE model.
To determine whether UMSC-exos and TAxI-exos could alleviated clinical symptoms associated with EAE, clinical scores and body weight measurements were recorded after EAE induction and exosomes treatments.Higher clinical scores and lower weights indicated deterioration of motor dysfunction and increased disease severity.

UMSC-exos and TAxI-exos alleviate demyelination in EAE mice
Clinical symptoms in EAE mice are associated with demyelination in the spinal cord [7].Therefore, we analyzed the severity of demyelination in spinal cord from each experimental group of mice.
The levels of MBP (a structural protein of myelin) in spinal cords from TAxI-exo-and UMSC-exo-treated EAE mice were analyzed by WB (Fig. 5c).Similarly, the results showed that both TAxI-exos and UMSCexos had signi cantly increased the expression of MBP compared with the EAE group (Fig. 5d).
These results suggested that TAxI-exos and UMSC-exos alleviated demyelination associated with the progression of EAE.
These results suggested that UMSC-exos and TAxI-exos treatments not only signi cantly inhibited the conversion of microglia to the M1 phenotype and promoted polarization to the M2 phenotype in vitro, but also decreased macrophage/microglia in ltration in mouse spinal cords.
This study applied in vitro and in vivo models to speci cally demonstrate the anti-in ammatory properties of UMSC-exos and TAxI-exos.Both suppressed activation and proliferation of T cells and inhibited the development of Th1 and Th17 cells in EAE mouse spinal cords.Surprisingly, TAxI-exos performed better than UMSC-exos.

UMSC-exos and TAxI-exos induce Tregs to exert immune tolerance in vitro and in vivo
Regulatory T cells (Tregs) are thought to play a critical role in the maintenance of peripheral immune tolerance.Therefore, we assumed that UMSC-exos and TAxI-exos would induce immune tolerance by facilitating the proliferation of Tregs.
UMSC-exos and TAxI-exos reduce expression levels of in ammatory cytokines in vitro and in vivo To verify the mechanism through which exosomes acted on T cells, cytokines were measured in the T-cell culture medium, serum, and spinal cords of each group.
In culture medium, the levels of IL-2 and IL-17 were signi cantly reduced by TAxI-exos and UMSC-exos treatments (Fig. 9a) and a decreasing trend of IFN-γ was only found in TAxI-exos group (Fig. 9a).We were unable to detect other cytokines (data not shown),because their concentrations were below the limit of detection of our assays.

Discussion
Although DMTs development has been continuous for MS in recent years, a limitation of the drugs is that they are only applicable to RRMS [28].Therefore, we focused on nding a more effective and safer drug.
As a classical animal model of MS, EAE was used in this study, which well represents the immune mechanism in MS patients.Because axonal loss is limited to the spinal cord and the minimal demyelination is seen in EAE of C57BL/6 mice [29], which is considered to be a hallmark of MS [30], spinal cord tissues were selected for our analysis.MSCs have been widely reported to possess better clinical curative effects on numerous in ammatory diseases through the paracrine functions of bioactive factors and differentiation into certain functional cells [31,32].Previous studies con rmed that UMSCexos exert an immunoregulatory effect on multiple autoimmune imbalances [33,34], attenuate tissue injury, and promote tissue repair and regeneration [35].Meanwhile, UMSC-exos have great BBBpermeability and are suitable for MS therapy [36].Because of these properties, we chose UMSC-exos for our study.However, when intravenous exosomes are used as a treatment, rapid accumulation occurs in peripheral organs, such as liver and spleen, in EAE [16].Thus, the targeting characteristic of MSC-exos requires improvement before they can be used for CNS disease therapies.Direct modi cation of the exosome surface by a CNS-targeting ligand has been shown to enhance CNS tissue retention of exosomes [37].On the basis of these ndings [17,38], we successfully established TAxI-exos by a postinsertion method using DSPE-PEG2000 [39], that targeted axons of the CNS in EAE models.Furthermore, we determined that systemic injection of TAxI-exos compared with that of UMSC-exos resulted in more e cient CNS accumulation and BV-2 cells internalization, more sustained clinical recovery with enhanced improvement of motor skills, and more reduction of neuroin ammation and demyelination in EAE mice.
We also investigated the underlying mechanisms.In terms of EAE pathogenesis, we have brie y summarized this with a diagram (Fig. 10).T cells play a major role in inducing and regulating MS pathophysiology, and the development of therapeutics for MS has often focused on targeting factors that mediate T cell functions [40].
Th1 and Th17 cells are the main pathological T cells in the innate immune response of MS and EAE [41].They produce IFN-γ and IL-17, and in ltrate the brain and spinal cord prior to the development of clinical symptoms in EAE, which activates macrophages/microglia [42].Thus, the reductions of CD4 + T cells (Th1), CD8 + T cells and Th17 cells may contribute to ameliorating EAE.Our study supported these ideas and demonstrated that TAxI-exos and UMSC-exos inhibited activation and proliferation of T cells and decreased the proportions of CD4 + T and CD8 + T cells in vitro.Additionally, our results suggested that TAxI-exos and UMSC-exos decreased the proportions of Th1 and Th17 cells in the EAE mouse spinal cord and relieved in ammatory cell in ltration of the CNS.Surprisingly, we found that TAxI-exos performed signi cantly better than UMSC-exos, probably because TAxI peptides provided UMSC-exos with blood stability and targeting ability.
Regulatory T cells (Tregs) are also involved in the pathogenesis of CNS autoimmune in ammation.It is believed that Tregs regulate immune responses in the periphery predominantly by suppressing effector CD4 + T cell subsets that mediate autoimmune responses [43].Tregs also promote oligodendrocyte differentiation and remyelination [44], which revealed a new regenerative prospect.Studies of MS have shown that a decline expression in CD25 + cells leads to Treg dysfunction during disease progression [45].
Consequently, restoring the Tregs functions prevents development of MS and EAE.In our study, we con rmed that that UMSC-exos and TAxI-exos induced CD4 + CD25 + FOXP3 + Tregs in vitro in the absence of the critical factors.The same result was obtained in vivo.These results suggest that TGFβ and IFN-γ or both from exosomes were provided to the Tregs surface or indirectly other other signals provided on lymphocytes and accessory cells (e.g., B cells, dendritic cells, and antigen-presenting cells), which is in agreement with recent studies showing that Tregs cannot be induced with puri ed and activated CD4 + T cells [46,47].Generally, these data support our hypothesis that UMSC-exos and TAxI-exos ameliorate EAE by suppressing pathological T cell subset activation and induce immune tolerance by promoting Tregs to a certain extent.
Microglia as early immune response cells in gray matter of the spinal cord and brain [48] become activated abnormally, which leads to the demyelination and neurodegeneration in MS and EAE [49].The next step in microglia activation is an M1/M2 microglia polarization imbalance that is also a major cellular mechanism of MS pathogenesis [50].M1 microglia cause tissue damage through releasing proin ammatory cytokines (TNF-α, IL-6, and iNOS) to accelerate the EAE course [51].M2 phenotype microglia produce anti-in ammatory cytokines (IL-10 and Arg1) to promote tissue regeneration [12].
When microglial M2 polarization is blocked and halted at the M1 phenotype, the demyelination in EAE may be exacerbated [52].Conversely, if polarization of microglia from M1 to M2 is induced, immune homeostasis can be restored and neurological functions improved.Our study focused on the role of TAxIexos and UMSC-exos in M1/M2 polarization of microglia.The results showed that TAxI-exos enhanced internalization ability of BV-2 cells, reduced the number of CD86 + CD206 + BV-2 cells and CD86 + BV-2 (M1) cells, and elevated the number of CD206 + (M2) cells in vitro.Most likely, the reason for the appearance of CD86 + CD206 + BV-2 cells under LPS stimulation is that BV-2 cells were directed to a transition state between M1 and M2 phenotypes.Additionally, decreasing expression of iNOS and increasing expression of Arg1 were found in spinal cords in vivo.These results also correspond to the expression level of Iba-1 in the spinal cord in which TAxI-exos had a better suppressive effect than UMSC-exos.Taken together, these data indicate that TAxI-exos provide a better delay in the onset of the EAE, which is bene cial to dampen in ammation and promote relief.
Some cytokines are involved in EAE mouse pathology, such as proin ammatory cytokines.For example, IL-1β is secreted from monocytes, IFN-γ and TNFα from Th1 cells, and IL-17A from Th17 cells.
Additionally, anti-in ammatory cytokines are involved, such as IL-4, IL-6, and IL-10 from Th2 cells, and TGF-β that is important for Tregs proliferation.IL-6 and IL-1β are also important activators of Th17 cells that differentiate from native CD4+ T cells [9].A major function of IL-17A in autoimmunity of CNS is to recruit IL-1β-secreting myeloid cells that prime pathological γδT17 and Th17 cells [53].IL-6 is also a Th2 cell-secreted cytokine that is positively correlated with the numbers of Th2 cells that inhibit Th1 cell proliferation and participate in humoral immunity by secreting anti-in ammatory cytokines such as IL-10 [54].IDO1 is an immunosuppressive enzyme screened from exosomes with good immunosuppressive effects on T cell proliferation [55,56].Our study analyzed the levels of these cytokines in vitro and in vivo.
The data demonstrated that treatment with TAxI-exos or UMSC-exos increased the mRNA levels of IL-6, IL10, IDO-1, and TNF-α in the spinal cord; reduced the concentrations of IL-1β, IL-6, and TNF-α, and increased the concentration of IL-4, IL-10, and TGF-β in the serum of EAE mice; and reduced the concentrations of IL-2, IL-17A, and IFN-γ in culture medium.These data helped to reveal the antiin ammatory mechanisms of exosomes.We found that the levels of IL-4 and IL-10 were commonly increased in periphery and in spinal cord after TAxI-exo-treatment.It indicated that the anti-in ammatory effect of TAxI-exos was partially mediated by boosting the production of cytokines from Th2 or Th2 cells themselves.Additionally, the declining levels of IL-2, IFN-γ, and IL-17A in culture medium indicated that Th1 and Th17 cells were effectively inhibited by TAxI-exos or UMSC-exos, which contributed to immunosuppression of EAE.By the way, the expression level of TNF was down-regulated in serum, but up-regulated in spinal cord.One possible reason is that TNF was ablated in monocytes/macrophages but not in microglia, which delayed the onset of EAE and was associated with reduced acute spinal cord in ltration of monocytes [57].Ultimately, the therapeutic effects of TAxI-exos or UMSC-exos were achieved partially by regulating the levels of in ammatory cytokines in our study.
Because of their small size and enrichment steps, compared with the MSCs, exosome-based therapeutics have many advantages in terms of safety and convenience, and can be considered as a surrogate treatment to cell therapy.Our results suggest that TAxI-exos perform better than native exosomes (UMSC-exos) in many aspects because of their central nervous system-targeting ability.A study has shown that TAxI peptides have no adverse effects on mice [17].The successful establishment and application of TAxI-exos in the mouse EAE model may contribute to treating other immune diseases such as MS.They can also be used as a targeted drug delivery vehicle for CNS diseases in the future.
Although our results suggest that TAxI-exos have a good therapeutic effect on EAE in mice, this study has some limitations.Despite the axon-targeting ability of TAxI, the speci c target cells in axons are not yet clear.We only chose microglia as the candidate in our study.More precise target cells of the TAxI peptide need to be explored further.The pathogenesis of MS involves many immune cells, and we just focused on CD4 + T cells, CD8 + T cells, Th17 cells, and Tregs both in vivo and in vitro.Therefore, it is necessary to explore the mechanisms with of immune cells in EAE.Additionally, the therapeutic effect of TAxI-exos in other suitable animal models, such as spinal cord injury, should be assessed in future studies.

Conclusion
Our study is the rst to develop a CNS-targeting cell-free therapy for MS and EAE using TAxI peptides and UMSC-exos, named TAxI-exos.TAxI-exos enhanced the therapeutic effect of UMSC-exos by improving the CNS-targeting ability.Our work also helped to understand the mechanism of immune regulation, immunosuppression, and induction of immune tolerance of UMSC-exos in EAE.

Figures
Figures

Figure 8 UMSC
Figure 8 TAxI-exos have great promise for other autoimmune and neurodegenerative diseases in the future.
DeclarationsEthic approval and consent to participate