Adult male Sprague–Dawley rats weighing 250–300 g were purchased from the Animal Center of the Chinese Academy of Sciences, Shanghai, China. The animal experimental protocols, including care, breeding, and operative procedures, were approved by the Animal Care and Use Committee of Soochow University and complied with the Guide for the Care and Use of Laboratory Animals approved by the National Institutes of Health.

Primary rat MSCs were isolated from male rats weighing 80–100 g. Briefly, the bone marrow of the femurs and tibias was flushed out with PBS followed by centrifugation. The pellet was suspended in Dulbecco’s modified Eagle medium (DMEM; Life Technologies, United States) with 10% heat-inactivated fetal bovine serum (FBS; Life Technologies) and 1% penicillin–streptomycin, and was then incubated under a humidified atmosphere with 5% CO₂ at 37°C. The medium was replaced every 3 days, and the MSCs were passaged when the cultures reached 90% confluence.

MSCs were transfected with miR-133b mimic and negative control using Lipofectamine 3000 (Invitrogen, United States) in serum-free medium according to the manufacturer’s instructions. At 72 h after transfection, exosomes were obtained from MSC supernatants using the ExoQuick-TC Kit (System Biosciences, United States). Subsequently, exosome pellets were resuspended in PBS at a total protein concentration of 10 μg/μl. Moreover, the exosomes were characterized by western blotting of exosome surface markers, including CD81, CD63, and CD9. The sequence of the miR-133b mimic was 5’-UUUGGUCCCCUUCAACCAGCUA-3′.

Male rats were anesthetized by chloral hydrate (400 mg/kg body weight). Following dissection of the paraspinal muscles, a laminectomy from T9–T11 was performed. Subsequently, SCI was inflicted with an aneurysm clip of 35 g closing force for 60 s at the T10 level as previously described (Figley et al., 2014; Soubeyrand et al., 2014). Finally, the incision was closed in layers with silk sutures. The sham-operated rats only received laminectomy. After surgery, all animals received penicillin and an analgesic for 3 days, and the bladders were manually voided thrice daily. At 24 h following trauma, the animals received treatments by tail intravenous injection of miR-133b exosomes (100 μg exosomes in 0.5 mL of PBS), miR-con exosomes (100 μg exosomes in 0.5 mL PBS), or PBS (0.5 ml) as previously described (Zhang et al., 2015).

The rats were randomly divided into four groups, and testing was performed by blinded observers: (1) sham group (the rats were subjected to sham operation), (2) control group (the rats received SCI and were treated with PBS), (3) miR-con group (the rats were subjected to SCI and treated with miR-con exosomes), and (4) miR-133b group (the rats were subjected to SCI and treated with miR-133b exosomes).

Total protein was extracted with RIPA lysis buffer (Beyotime Institute of Biotechnology, China). We used the BCA protein assay kit to assess protein concentrations (Beyotime Institute of Biotechnology). For in vivo studies, a 10-mm long segment of the spinal cord containing the injury site was harvested at 4 days following SCI (four animals per group). For western blot analysis, 30 μg of protein was separated by 12% SDS–PAGE and transferred to PVDF membranes (Merck Millipore, Germany). Following blocking with 5% non-fat milk, the membranes were incubated with primary antibodies overnight at 4°C: mouse anti-CD63 (Abcam, Cambridge, United Kingdom), rabbit anti-CD81 (Abcam), rabbit anti-CD9 (Abcam), mouse anti- neurofilament (NF) (Abcam), rabbit anti-growth-associated protein 43 (GAP43) (Abcam), rabbit anti-p-signal transducer and activator of transcription 3 (STAT3) (Cell Signaling Technology, United States), mouse anti-STAT3 (Cell Signaling Technology), rabbit anti-p-cAMP response element-binding protein (CREB) (Abcam), rabbit anti-CREB (Abcam), rabbit anti-RhoA (Cell Signaling Technology), or rabbit anti-β-actin (Cell Signaling Technology). Horseradish peroxidase (HRP)-conjugated goat anti-rabbit or mouse was used as secondary antibody (Cell Signaling Technology). Finally, bands were visualized by enhanced chemiluminescence (ECL) Plus (Thermo Scientific, United States), and the relative band densities were determined by Image Lab (Version 2.0.1).

Total RNAs from MSCs, exosomes, or spinal cords (four animals per group) were extracted by TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. For in vivo studies, a 10-mm long spinal cord segment containing the injury epicenter was harvested. Total RNA was reverse-transcribed to cDNA using the PrimeScript RT reagent kit (TaKaRa, Japan). Quantitative real-time PCR was performed using the miScript SYBR Green PCR kit (QIAGEN, Germany). The relative expression of miRNA normalized to U6 was calculated using the 2^−ΔΔCT method.

The recovery of hindlimb locomotor function was evaluated by using the Basso–Beattie–Bresnahan (BBB) locomotor rating scale of 0 (no motor activity) to 21 (normal locomotion) (Basso et al., 1995). Two independent investigators blinded to the treatment observed the movement and scored the locomotor function preinjury and at days 1, 3, 5, 9, and 14 post-injury as previously described (Hu et al., 2013; Datto et al., 2015).

Animals were anesthetized terminally with an overdose of inhalation isoflurane at day 4 after injury. The spinal cords were embedded in an optimal cutting temperature compound. The T9–T11 spinal cord segments near the epicenter of the lesion were collected for histological evaluation. HE staining was performed in accordance with the manufacturer’s instructions to quantify the area of lesion cavity using Imagepro-Plus software. The T9–T11 spinal cord segments in transverse sections were dissected in six rats per group, and every eighth section derived from each animal was used to determine the area of cystic cavity in each group as previously described (Chen et al., 2014; Wang et al., 2014).

For immunohistochemistry, briefly, longitudinal sections containing the lesion site were deparaffinized with xylene and hydrated in graded alcohol, followed by boiling in citrate buffer (pH 6.0) twice for 5 min. Subsequently, to inactivate the endogenous peroxidase, the sections were cooled off and incubated in 3% H₂O₂ for 15 min at room temperature. The slides were then blocked with 10% FBS for 10 min and incubated with primary antibodies overnight at 4°C, including rabbit anti-GAP43, rabbit anti-NF, rabbit anti-MBP (Abcam), and mouse anti-NeuN (Abcam). Following washing with TBS, the sections were incubated with fluorescence-labeled secondary antibodies (Abcam). Finally, the sections were stained with DAPI and visualized under a confocal laser-scanning microscope (Olympus LSM-GB200, Tokyo, Japan).

Data were analyzed with SPSS 19.0 (SPSS Inc., Chicago, IL, United States). All data represent at least three independent experiments and are expressed as means ± standard deviations (SD). One-way analysis of variance (ANOVA) with Tukey’s post hoc test was used to compare the levels of different groups. BBB scores were analyzed by repeated measures ANOVA followed by Bonferroni post hoc corrections. Statistical significance was set at p < 0.05.

The expression levels of miR-133b in injured spinal cords were measured by qRT-PCR at 12 h, 24 h, 2, 3, 4, 5, and 7 days after acute SCI. The results revealed that the expression of miR-133b was significantly downregulated at 24 h or later following SCI (Figure 1A).

MSCs were cultured as described above and characterized as being positive for CD73, CD90, and CD105, but negative for CD34 and CD45 (data not shown). MSCs were then transfected with miR-133b mimic and negative control (the transfection efficiency was approximately 90%). Exosomes were isolated from MSC supernatants at 72 h after transfection. To characterize the exosomes, we carried out western blot analysis, which showed that exosomes expressed common exosomal marker proteins, including CD9, CD63, and CD81 (Figure 1B), as previously described (Roccaro et al., 2013). As shown in Figure 1C, qRT-PCR revealed that the expression levels of miR-133b were approximately 2.5-fold higher in exosomes derived from miR-133b-transfected MSCs than in those from MSCs transfected with miR-con. These results demonstrated that MSCs efficiently packaged miR-133b into secreted exosomes.

We first explored using qRT-PCR whether the tail vein injection of exosomes would change the miR-133b expression in spinal cords. As shown in Figure 2A, compared with the control groups, the expression of miR-133b was significantly increased in the miR-133b exosome-injected group at day 4 following SCI.

To investigate whether miR-133b exosomes had a beneficial effect on the recovery of hindlimb locomotor function after acute SCI, the BBB locomotor grading scale was used at different time points after SCI. Immediately after SCI, the BBB score was approximately 0–1, indicating that the SCI model was successful. Spontaneous functional recovery was observed in all groups after SCI, as described previously (Pinzon et al., 2008). After 5 days, improved recovery with significant differences between miR-133b and miR-con exosome-injected rats was observed (day 5, p < 0.01; day 9, p < 0.001; and day 14, p < 0.01), indicating that the tail vein injection of miR-133b exosomes improved the recovery of hindlimb locomotor function after SCI (Figure 2B).

We next evaluated the effect of miR-133b exosomes on the volume of the lesion and the preservation of NeuN+ neurons after SCI by immunohistochemistry. HE staining showed that miR-133b exosomes significantly decreased the area of lesion cavity when compared to miR-con or the control group (Figures 3A,C). At day 4 after SCI, neurons in the injured spinal cord were stained with antibodies directed against NeuN, a specific marker of mature neurons. As shown in Figures 3B,D, the results showed significantly increased mature neuron numbers in injured rats receiving miR-133b exosomes compared to rats receiving miR-con exosomes (p < 0.01).

To investigate the potential effect of miR-133b exosomes on the outgrowth of axons, an immunohistochemistry study of GAP43 and NF was carried out (Van der Zee et al., 1989). The results showed that the expression of GAP43 was increased in the miR-133b exosomes group compared to the miR-con or control group (Figure 4). Moreover, we observed that treatment with miR-133b exosomes enhanced NF expression at day 4 after SCI (Figure 5). Finally, western blot results confirmed significant increases in the protein levels of GAP43 and NF in the miR-133b exosomes group compared with those in the miR-con group (Figures 6A,B).

Our previous study indicated that RhoA is a direct target of miR-133b (Lu et al., 2015). Moreover, RhoA has been shown to be involved in the death of neuronal cells in the spinal cord (Wu et al., 2016). To assess the effect of miR-133b exosomes on the expression of RhoA after SCI, we detected the levels of RhoA expression by using western blot. The results demonstrated that the expression of RhoA was significantly decreased in the injured spinal cords of rats receiving miR-133b exosome injections compared to those in miR-con exosome-treated animals (p < 0.05) (Figure 6C). These results are consistent with those of other in vitro and in vivo studies (Niu et al., 2016; Theis et al., 2016). Moreover, the phosphorylation of ERK1/2, another cell survival-related pathway protein, was significantly increased in the injured rats treated with miR-133b exosomes after SCI (Figure 6D).

The transcription factors CREB and STAT3 have been shown to be involved in the outgrowth of neurites (Gao et al., 2004; Qiu et al., 2005). In this study, we observed that injecting miR-133b exosomes increased the protein phosphorylation levels of STAT3 and CREB in injured spinal cords after SCI (p < 0.05 for p-STAT and p < 0.05 for p-CREB; Figures 6E,F).