Exosomes containing miR‐451a is involved in the protective effect of cerebral ischemic preconditioning against cerebral ischemia and reperfusion injury

Abstract Aim To study the role of exosomes in the protective effect of cerebral ischemic preconditioning (cerebral‐IPC) against cerebral I/R injury. Method Mouse models of cerebral‐IPC and MCAO/R were established as described previously, and their behavioral, pathological, and proteomic changes were analyzed. Neuro‐2a subjected to OGD/R were treated with exosomes isolated from the plasma of sham‐operated and cerebral‐IPC mice. The differentially expressed miRNAs between exosomes derived from sham‐operated (S‐exosomes) and preconditioned (IPC‐exosomes) mice were identified through miRNA array, and their targets were identified through database search. The control and OGD/R cells were treated with the IPC‐exosomes, miRNA mimic or target protein inhibitor, and their viability, oxidative, stress and apoptosis rates were measured. The activated pathways were identified by analyzing the levels of relevant proteins. Results Cerebral‐IPC mitigated the cerebral injury following ischemia and reperfusion, and increased the number of plasma exosomes. IPC‐exosomes increased the survival of Neuro‐2a cells after OGD/R. The miR‐451a targeting Rac1 was upregulated in the IPC‐exosomes relative to S‐exosomes. The miR‐451a mimic and the Rac1 inhibitor NSC23766 reversed OGD/R‐mediated activation of Rac1 and its downstream pathways. Conclusion Cerebral‐IPC ameliorated cerebral I/R injury by inducing the release of exosomes containing miR‐451a.


| INTRODUC TI ON
Stroke is a common disorder of the central nervous system, with ischemic stroke accounting for 70-80% of all cases. 1 Although mechanical thrombectomy can significantly increase the chances of successful reperfusion, the long-term prognosis of stroke patients is poor due to ischemia and reperfusion (I/R) injury. [2][3][4] Neuroprotective agents like edaravone and citicoline can alleviate the cerebral I/R injury in animal models but have not shown encouraging results in clinical trials. 5,6 In the recently concluded ESCAPE-NA1 trial, the eicosapeptide nerinetide (NA-1) did not significantly improve the clinical outcomes in acute ischemia stroke patients following thrombectomy and alteplase. 7 Therefore, other strategies have to be explored in order to augment the therapeutic effect of reperfusion.
Ischemic preconditioning (IPC) is induced by transient and repeated (or long-term but moderate) ischemia and reperfusion, which increases the tolerance of the target tissue to long-term ischemia and I/R injury. 8,9 Direct-IPC is induced by repeated I/R of the target organ, and remote-IPC by repeated I/R of distal organs such as the extremities. 10 Although remote-IPC is technically simple, its efficacy has not been conclusively demonstrated in clinical studies. 11 In contrast, direct cerebral-IPC might be induced during transient ischemic attack (TIA) and intracranial atherosclerosis (ICAS). 12,13 In fact, patients with a history of TIA or ICAS tend to have smaller infarct volumes, indicating an endogenous protective effect of cerebral-IPC. Therefore, it is necessary to elucidate the mechanism of cerebral-IPC in order to identify novel neuroprotective targets.
Exosomes are extracellular vesicles measuring 50-150 nm in diameter that transport various bioactive components and thus regulate multiple physiological and even pathological processes. 14 Li et al. found that remote-IPC induces secretion of exosomes containing HIF-1α, which protects against permanent distal middle cerebral artery occlusion. 15 However, the neuroprotective effect of remote-IPC and remote-IPCderived exosomes does not extend to large vascular occlusion and I/R injury. Furthermore, the protective component in remote-IPC-derived exosomes is not equivalent to that in cerebral-IPC-derived exosomes.
In this study, we determined the efficacy of cerebral-IPC and exosomes derived from cerebral-IPC mice (IPC-exosomes) in a mouse model of middle cerebral artery occlusion and reperfusion (MCAO/R), as well as a cellular model of oxygen-glucose deprivation and restoration (OGD/R). To elucidate the underlying mechanisms, the miR-NAs differentially expressed in the IPC-exosomes were screened by miRNA array and confirmed by miR mimics. The target protein of the candidate miRNA was then confirmed, and the downstream pathways were identified.

| Establishment of MCAO/R and cerebral-IPC models
Specific pathogen-free (SPF) C57BL/6 male mice (8-10 weeks old and weighing 25-30 g) were purchased from Beijing Vital River Lab Animal Technology Co. Ltd., and housed in the SPF laboratory of Changhai Hospital Animals at 22°C and 12-h day-night cycle with ad libitum access to water and food. Seventy mice were divided into the sham-operated, cerebral-IPC, middle cerebral artery occlusion and reperfusion (MCAO/R), and cerebral-IPC+MCAO/R groups. Cerebral-IPC was induced by bilateral common carotid artery occlusion as previously described. 16 Two-hour MCAO/R was established 24 h after cerebral-IPC using the suture-occlusion method as previously described. 17 Twenty-four hours after the final operation, the neurological deficit score (NDS) of the mice was evaluated, and their plasma and brains were harvested for various assays (AppendixS 1).

| Infarction volume measurement
The infarcted region in the brain was evaluated by TTC (2,3,5-Triphenyltetrazolium chloride) staining. Briefly, freshly resected brains were sliced into 2 mm-thick sections along the coronal axis and incubated in 2% TTC at 37°C. The vital region appeared red and the infarcted region remained white.

| Measurement of neurological deficit score
The NDS was measured as previously described on a scale of 0-5: 0 -no apparent deficiency, 1 -flexion of contralateral forelimb, 2 -rotated to the opposite side, 3 -falling to the opposite side, 4 -no spontaneous walk, and 5 -dead.

| Histological assessment
Brain tissues were fixed in 4% paraformaldehyde for 24 hours, dehydrated, embedded in paraffin, and sliced into 4μm-thick sections. Hematoxylin and eosin (HE) staining, Nissl staining, and TUNEL staining were performed using specific kits from Servicebio Technology Co. Ltd. (Wuhan, China) as per the manufacturer's instructions.

| ROS (Reactive Oxidative Species) detection
Fresh brain tissues were embedded in optimal cutting temperature compound (OCT) and cut into 8-10 μm-thick slices. The sections were stained using a fluorescent ROS probe using a specific kit from Servicebio Technology Co. Ltd. The ROS+and DAPI+cells were, respectively, detected using 520 nm and 364 nm excitation lasers.

| Western blotting
Proteins were extracted from the cerebral cortex or cultured cells

| Isolation and characterization of plasma exosomes
Blood samples of the sham-operated and cerebral-IPC mice were collected for isolating exosomes. The mice were anesthetized again, and blood was drawn from the apex cordis and transferred into EDTA-K2 tubes. After centrifuging at 2000 × g for 20 minutes at 4°C, the plasma layer was transferred into nuclease-free tubes and centrifuged at 10,000 × g for 20 minutes at room temperature to remove debris. The supernatants were carefully removed and the exosomes were recovered from the pellets using the Total Exosome Isolation Kit (Invitrogen, ThermoFisher Scientific) as per the manufacturer's instructions. The exosomes were characterized as previously described. 18 Briefly, the purified exosomes were re-suspended in 500 μl PBS, and the total protein content was first measured using the BCA Protein Assay Kit (Beyotime Biotechnology). Nanoparticle tracking analysis (NTA) was performed to measure the size and concentration of exosomes as per standard protocols. Western blotting was performed as described above using antibodies against exosomal biomarkers including CD63, CD81, and CD9. Finally, the morphology of the exosomes was examined by electron microscopy.

| Measurement of exosomal miRNAs
Exosomal miRNAs were extracted using the miRcute miRNA Isolation Kit (TianGen Biotech Co. Ltd.). The expression profiles of miRNAs in the different groups were compared by oligonucleotide microarray as described in previous studies and visualized using heatmaps and volcano plots. 19 The results of high-throughout sequencing were confirmed by quantitative real-time PCR (qRT-PCR).
The primer sequences are listed in AppendixS 1.

| Cellular viability assay
Cell Counting Kit-8 (CCK-8; HY-K0301, MedChemExpress) was used to measure the viability of the differentially treated cells according to the manufacturer's instructions. The absorbance of each group relative to the control was calculated.

| Quantitative Real-time PCR
The miRNA and mRNA levels in exosomes or cells were measured by qRT-PCR. Total RNA was extracted using Trizol (Invitrogen, Thermo Fisher Scientific) according to the manufacturer's instructions, and the miRNA extraction protocol has been described above. The RNA samples were quantified at OD 260/280 , and reverse transcribed using the PrimeScriptTM RT reagent Kit (RR037A, TAKARA).
QRT-PCR was performed on LightCycler 96 (Roche Group) using SYBR® Premix Ex Taq™ II (Tli RNaseH Plus) (RR420A, TAKARA), and the fold change in expression levels was calculated by the 2 −△△Ct method. The primer sequences are listed in AppendixS 1.

| MicroRNA mimic and transfection
The N2a cells were transfected with the miR-451a mimic (5'-AAACCGUUACCAUUACUGAGUU-3'). Briefly, miR-451a mimic or negative control was incubated with Lipofectamine 3000 (Thermo Fisher Scientific), and the mixture was then added to the cultured cells. OGD/R was induced at least 24 hours after transfection.

| ROS measurement
ROS levels were measured using the ROS Assay Kit (S0033, Beyotime) according to the manufacturer's instructions. The fluorescence intensity was measured at emission wavelength 535 nm and excitation wavelength 485 nm, and calculated relative to that of the control group.

| GTP-bound GTPase pull-down assay
GTP-bound GTPase pull-down assay was used to determine the activation of Rac1 and Cdc42 as per the kit directions (8815 and 8819, Cell Signaling Technology). Activated Rac1 or Cdc42 were then eluted with sodium dodecyl sulfate and confirmed by Western blotting.

| Statistical analysis
All statistical analysis was performed using GraphPad Prism 7.04.
The normality of the data was verified by Shapiro-Wilk test. Twotailed t test was used to compare two groups, and one-way ANOVA was used for multiple groups. P < 0.05 was considered statistically significant.

| Cerebral-IPC reduced cerebral infarction and ameliorated neurological dysfunction
TTC staining showed that cerebral-IPC did not induce significant infarction in the mouse brain, and in fact markedly decreased the infarcted region following MCAO/R ( Figure 1A Figure S3), which corresponded to a decrease in ROS levels ( Figure 2F,G). Taken together, cerebral-IPC elicited endogenous anti-inflammatory, anti-apoptotic and anti-oxidative pathways in the injured brain, which was accompanied by slight tissue damage.

| Cerebral-IPC increased the exosomal load in plasma
To determine the potential role of exosomes in the protective effect of cerebral-IPC, we extracted exosomes from the plasma of the sham-operated and cerebral-IPC mice. The total amount of exosomal protein was significantly higher in the cerebral-IPC versus the shamoperated group (1.79 ± 0.20 μg/μL vs. 3.05 ± 0.21 μg/μL, P < 0.0001; Figure 3A). Furthermore, NTA showed that the number of particles with diameters ranging from 30 nm to 150 nm was also higher in the cerebral-IPC mice compared to the sham-operated controls (1.18 ± 0.31 × 10 12 particles/ml vs. 3.27 ± 0.38 × 10 12 particles/ml, P < 0.0001; Figure 3B). Finally, cerebral-IPC increased the relative expression levels of the exosomal surface markers CD9, CD63, and CD81 ( Figure 3C). The shape and size of the exosomes are shown in the electron micrographs in Figure 3D. To summarize, cerebral-IPC elevated the number of exosomes in plasma, which likely mediate the protective effects of cerebral-IPC against I/R injury.

| Cerebral-IPC-derived exosomes (IPCexosomes) increased survival of N2a cells after OGD/R
An in vitro model of I/R injury was established by subjecting N2a cells to OGD/R, and the effects of the IPC-exosomes were evaluated. As shown in Figure 3E, the N2a cells effectively internalized the fluorescentlabeled exosomes. Furthermore, IPC-exosomes increased the viability of cells after OGD/R stimulation compared to that of the untreated and S-exosomes-treated OGD/R cells (O+IPCE vs. O and O+SE: 0.69 ± 0.08 vs. 0.37 ± 0.16 and 0.31 ± 0.16, P < 0.0001 and P < 0.0001; Figure 3F).

| MiR-451a targets Rac1 and its downstream pathways
We identified the putative targets of miR-451a by screening the online databases (AppendixS 1). Rac1, Klhl5, AW549877, and Cep55 were identified in all databases, of which Rac1, a member of Ras superfamily, has established pro-oxidative and pro-inflammatory effects that may exacerbate the tissue damage caused by I/R injury.
Furthermore, the downstream molecules of Rac1 include JNK, p38, MMP-2, and MMP-9, which were elevated after MCAO/R and reduced by cerebral-IPC ( Figure 1E,F). Therefore, we hypothesized that the high levels of miR-451a in IPC-exosomes protect against cerebral I/R injury by repressing the Rac1 expression. To confirm this hypothesis, we transfected the N2a cells with the miR-451a mimic and found that Rac1 mRNA level was significantly reduced (1.06 ± 0.15 vs. 0.72 ± 0.17, P = 0.0026; Figure 4G). In addition, cells treated with IPC-exosomes had lower Rac1 mRNA levels (1.06 ± 0.15 vs. 0.71 ± 0.14, P = 0.0014; Figure 4G), which was verified by the Western blotting as well ( Figure 4H,I). Furthermore, the miR-451a mimic promoted cellular survival ( Figure 4J) and decreased ROS levels ( Figure 4K) after OGD/R injury. NSC23766, a specific inhibitor of Rac1, also showed a similar protective effect against OGD/R injury ( Figure 4J,K). These findings clearly indicated that miR-451a exerts its protective effect by downregulating Rac1.

| DISCUSS ION
We found that cerebral-IPC protected mice against cerebral I/R injury, and the protective effects were mediated via the increased plasma load of IPC-exosomes. The latter also exerted a similar protective effect on a cellular model of OGD/R injury. MiR-451a was identified as one of the protective components in IPC-exosomes that targets Rac1 and inhibits its downstream pathways (Graphic Abstract).
Several protective mechanisms of cerebral-IPC have been demonstrated, such as reduction in cerebral metabolic rate, inhibition of excitotoxicity, oxidative stress and inflammation, and activation of neuroprotective or survival pathways. 21,22 In our study, cerebral-IPC mitigated MCAO/R-induced inflammation, apoptosis, and oxidative stress by inhibiting the JNK and P38 pathways.
However, cerebral-IPC by itself also triggered several inflammatory and apoptotic proteins and increased the proportion of active microglia, albeit to a significantly lesser extent compared to MCAO/R.
Since cerebral-IPC is induced by transient and/or repeated ischemic or anoxic insult, it may lead to brain damage on account of the high susceptibility of the brain to ischemia and anoxia. While the efficacy of remote-IPC has been demonstrated as an alternative for cerebral-IPC, it remains to be confirmed clinically. 11 Direct cerebral-IPC is still an indispensable method for studying cerebral endogenous protection. Thus, cerebral-IPC is a double-edged sword, and its underlying mechanisms need further study to augment the benefits and prevent the potential adverse effects.
Cerebral-IPC significantly increased the number of plasma exosomes, which raised the possibility that these IPC-exosomes mediated the neuroprotective effects after I/R injury. Exosomes are extracellular vesicles that transport proteins, miRNAs, lncRNAs, and circRNAs to remotely regulate physiological and pathological processes. 14 Studies show that exosomes derived from endothelial progenitor cells, adipose-derived stem cells, and multipotent mesenchymal stromal cells can manifest neuroprotective or neurotrophic effects by transporting protective miRNAs. [23][24][25] Consistent with this, the plasma exosomes of sham-operated and cerebral-IPC mice in our study showed distinct miRNA profiles. Most miRNAs in the IPC-exosomes were downregulated and only seven were upregulated. Based on previous findings that preconditioned cells or organs release protective factors, we hypothesized that the upregulated rather than the downregulated exosomal miRNAs are neuroprotective. 15,26,27 Indeed, the neuroprotective effect of miR-451a, the most significantly upregulated miRNA in IPC-exosomes, was confirmed in vitro. MiR-486-3p, another upregulated miRNA in the IPCexosomes, is an established biomarker of acute coronary syndrome and should be studied further for its role in cerebral I/R injury. 28 In addition, the downregulated miRNAs such as miR-16-1, miR-15a, and miR-27 may also have functional significance in cerebral I/R injury. For instance, miR-16-1 and miR-15a promote inflammatory and apoptosis, and inhibition of miR-27 enhances neurogenesis. [29][30][31] Down-regulation of these pathological miRNAs may therefore ameliorate the effects of cerebral I/R. Rac1 was confirmed as the target of miR-451a through bioinformatics and molecular analyses. Rac1, or Rac family small GTPase 1, is a GTP-binding protein regulating cell growth and cytoskeletal function. 32 Both protective and detrimental effects of Rac1 have been reported in the neurological system. Several studies have shown that Rac1 protects the brain-blood barrier and exerts a regenerative effect on axons after cerebral ischemia. 33,34 On the other hand, inhibition of Rac1 lowers ROS production, which is a significant pathological factor underlying I/R injury. 35,36 Rac1-GTP, the activated form of Rac1, is an essential component of NADPH oxidase and therefore crucial for ROS generation. 37 Our findings supported a pathological role of Rac1 during OGD/R, which led to a significant increase in ROS levels and exacerbated tissue damage by upregulating MMPs, p-P38, and p-JNK. We hypothesize that cerebral-IPC induces secretion of exosomes containing miR-451a that represses Rac1 expression and its downstream pathways, eventually mitigating the pathological responses. Furthermore, the protective effects of miR-451a mimic and Rac1 inhibitor were similar in the OGD/R model, further underscoring the role of Rac1 in I/R injury. The combination of miR-451a mimic and Rac1 inhibitor inactivated the GTPase Cdc42, whereas either had no effect alone, indicating the selectivity of miR-451a. PAK1 is a downstream molecule of Rac1 and Cdc42, which may involve the negative regulation of NADPH oxidase. 38 Interestingly, the Rac1 inhibitor but not miR-451a mimic significantly reduced the level of p-PAK1 after OGD/R, which indicated that miR-451a predominantly inhibits the total expression of Rac1 and the formation of NADPH oxidase rather than directly reducing the activation of Rac1 and its downstream pathways.
There are some limitations in this study that ought to be addressed. Our study evaluated the protective effect of cerebral-IPC during the acute phase of MCAO/R and its long-term efficacy remains to be determined. Secondly, although we established miR-451a as a protective factor of IPC-exosomes, the presence of other protective molecules cannot be ruled out. We did not perform in vivo experiments to assess the function of IPC-exosomes due to the technical and ethical challenges related to harvesting large amounts F I G U R E 6 MiR-451a suppresses the activation of downstream pathways of Rac1. (A-C) The statistical analysis of the protein expression levels in the indicated groups according to the immunoblot. (* and # P < 0.05 compared to S and O groups, respectively; mimic: miR-451a mimic, NSC: NSC23766) [ of IPC-exosomes. Our future studies will focus on evaluating the long-term effect of cerebral-IPC on the pathological and functional changes of after MCAO/R, and screening for the protective factors in IPC-exosomes. 39 In addition, identifying the cellular origin of IPC-exosomes may allow in vitro production of neuroprotective exosomes. Cultured-cell-derived exosomes and nanoparticles containing bioactive components have significant clinical potential and warrant further investigation.

| CON CLUS ION
Cerebral-IPC protected mice from MCAO/R injury through antioxidant, anti-inflammatory, and anti-apoptotic mechanisms. The IPCexosomes likewise promoted the survival of N2a cells after OGD/R.
MiR-451a was identified as a protective factor in the IPC-exosomes, which repressed the total expression of Rac1 and inactivated its downstream pathological pathways.

D I SCLOS U R E
All authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.