Schwann cells (M1700-57, ScienCell) isolated from postnatal day 8 C57BL/6 mouse sciatic nerves were cultured in Schwann cell medium (1701, ScienCell). To overexpress miR-214 in Schwann cells (SCs), a lentiviral vector-carrying the human pre-microRNA expression construct Lenti-miR-214 (hsa-miR-214 ACAGCAGGCACAGACAGGCAGT, PMIRH 214PA-1, System Biosciences) was packaged using Lenti-X™ Packaging Single Shots (631276, Takara Bio USA, Inc.) according to the manufacture’s protocols. A lentiviral vector-carrying scramble construct (PMIRH000PA-1, System Biosciences) was used as a control. To maintain the purity of transfected Schwann cells, cells were cultured in the presence of 1.1 ug/ml puromycin. When Schwann cells reached approximately 60%∼80% confluence, culture medium was changed to a medium containing 5% exosome-depleted fetal bovine serum-contained medium (SBI System Bioscience), and Schwann cells were maintained for an additional 48 h. The conditioned medium was then collected for isolation of EVs.

Extracellular vesicles (EVs) were isolated using a differential ultracentrifugation approach and characterized in compliance with the guideline of Minimal information for studies of extracellular vesicles 2023 (Welsh et al., 2024). Briefly, the collected supernatant was filtered through a 0.22 μm filter before being centrifuged at 10,000 g for 30 min. Subsequently, ultracentrifugation was conducted at 100,000 g (Optima XE-100 Ultracentrifuge, SW 32 Ti Rotor) for 2 h, and the resulting pellet was resuspended with 100 μl of sterilized saline. The concentration and size of EVs were determined using the NanoSight analysis system (Malvern Panalytical). Western blot analysis was performed to measure proteins of Alix, heat shock protein (HSP 70), TSG101, CD9, CD63, and calnexin in the isolated EVs (Wang et al., 2020; Zhang et al., 2025).

Transmission electron microscopy (TEM) was performed to examine the ultrastructure of the isolated EVs (Zhang et al., 2025). Briefly, freshly isolated EVs were placed onto copper grids coated with Formvar carbon film (Catalog #FCF400-Cu, Electron Microscopy Sciences, Hatfield, PA, USA) and allowed to adhere for 45 s. Unbound particles were removed by rinsing the grids twice with distilled water. The samples were then briefly stained with 1% osmium tetroxide (OsO4, Catalog #19100, Electron Microscopy Sciences) for 30 s. Prepared grids were subsequently examined under a JEOL 1400 Flash transmission electron microscope (JEOL, Tokyo, Japan).

All experimental procedures were approved by the Institutional Animal Care and Use Committee of Henry Ford Hospital (IACUC #1502) and were conducted according to NIH Guidelines for the Care and Use of Laboratory Animals. All experiments and data analyses were conducted by investigators who were unaware of the treatment assignments.

To induce the T2D model, male C57BL/6 mice at the age of 10 weeks were fed a 60-kcal% high-fat diet (HFD, D12492; Research Diets). After 8 weeks on HFD, mice were treated with two low doses of streptozotocin (STZ, 75 mg/kg followed with 50 mg/kg 3 days later) (Yorek et al., 2015, 2017). Mice with blood glucose >250 mg/dl were considered diabetic. These mice remained on the HFD for the duration of the study. WT group was fed the standard diet for the duration of the study. To evaluate the effect of naïve-EVs, 214-EVs, and Scra-EVs on neurological recovery of DPN mice, 16 weeks post HFD/STZ treatment, the mice were randomly divided into one of the following treatment groups: (1) WT + saline (WT); (2) T2D + saline (T2D); (3) T2D + naïve-SC-EVs (EVs); (4) T2D + miR-214 enriched SC-EVs (214-EVs); (5) T2D + scramble-SC-EVs (Scra-EVs), where n = 10 mice/group was selected to ensure the study had an 80% power to detect an effect size of 1.48, assuming two-sided test, alpha = 0.05 and an equal-space means (Cohen, 1988). EVs (2 × 10¹⁰particles in 0.2 ml saline/mouse) or an equal volume of saline were intravenously administered via a tail vein once every week for 8 consecutive weeks, respectively. All mice were sacrificed 8 weeks after the initial treatment. The doses of EVs were selected based on our published studies (Wang et al., 2020, 2024).

Motor nerve conduction velocity (MCV) and sensory nerve conduction velocity (SCV) in sciatic nerve were measured every 4 weeks, as previously described (Wang et al., 2012, 2024). Briefly, mice were anesthetized with 1.5% isoflurane, and electrodes were positioned at the sciatic notch and knee. Single square wave current pulses were applied using a pulse stimulator, while simultaneous electromyography recordings were obtained through two sterilized electrodes inserted into the dorsum of the foot. Data were collected using a 4-channel amplifier (Natus UltraPro S100 EMG/NCS/EP Neurodiagnostic System). To maintain a stable body temperature of 37°C ± 1.0°C during the measurements, a water-based heater was used. MCV and SCV were analyzed using the Natus Elite software supplied provided by the manufacturer (Wang et al., 2012, 2024).

Thermal sensitivity was evaluated biweekly using the Hargreaves method with a thermal stimulation meter (plantar test, Model 336 TG, IITC Life Science), following established protocols (Wang et al., 2012, 2020, 2024). Mice were allowed to acclimate on a transparent glass surface for at least 20 min before testing. A thermal stimulator was positioned under the plantar surface of the hind paw, and withdrawal latency in response to radiant heat at 15% intensity was measured. Each mouse underwent three trials, with a 15-min interval between measurements. The average withdrawal latency was then calculated for each animal.

For immunofluorescent staining, sciatic nerve tissues were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned at a thickness of 6 μm. A total of three cross-sections with each at 60 μm interval per animal were employed for immunochemistry study with a primary antibody against CD68 (1:30 dilution, BIO-RAD, Catalog #MCA341) (Wang et al., 2024).

To assess intraepidermal nerve fiber density (IENFs), hind paw plantar skin tissues were fixed in Zamboni Fixative, following our published protocol (Wang et al., 2020; Zhang et al., 2021a). 20 μm-thick cryosections of footpad tissue were immunostained with an anti-protein gene product 9.5 (PGP9.5) antibody (1:1,000; MILLIPORE) and imaged using a FluoView FV 1200 laser scanning confocal microscope (Olympus) with a 40x objective. Images were analyzed using the MicroComputer Imaging Device (MCID) system (Imaging Research Inc.). Nerve fibers crossing the dermal-epidermal junction were counted, and intraepidermal nerve fiber density was expressed as the number of fibers per millimeter of epidermal length (Fan et al., 2020; Wang et al., 2020; Zhang et al., 2021a).

For morphometric analysis of the sciatic nerve, tissue samples were processed, as described previously (Wang et al., 2011, 2020, 2024). Transverse sciatic nerve sections (2 μm thick) were stained with toluidine blue and randomly imaged using a 100 × oil immersion lens (Olympus Optical Co., Ltd.). Myelin sheath thickness, myelinated fiber and axon diameter were quantified using the MCID system, following established protocols (Wang et al., 2011, 2020, 2024).

Dorsal root ganglia (DRG) neurons were harvested from 26-week-old WT and T2D mice and cultured under normal glucose (RG, 5 mM) and high glucose (HG, 30 mM) conditions, according to published protocols (Wang et al., 2020, 2024). To evaluate the impact of EVs on neurite outgrowth, DRG neurons (2,000 cells/cm²) were plated on glass coverslips and treated with one of the following conditions: (1) WT + saline (WT); (2) T2D + saline (T2D); (3) T2D + naïve-SC-EVs (EVs); (4) T2D + miR-214 enriched SC-EVs (214-EVs); (5) T2D + scramble-SC-EVs (Scra-EVs). EVs were applied at a concentration of 6 × 10⁹ particles/ml. After 72 h in culture, neurons were immunostained with an anti-neurofilament heavy subunit (NF-H) antibody (1:500, BioLegend). Images were captured at 10 × magnification using a digital camera. Neurite outgrowth was quantified by measuring the total neurite length of 20 neurons per group using the MCID system, and the average neurite length was calculated (Wang et al., 2020).

Schwann cells were grown to 90% confluence in six-well plates under normal (5 mM) and high glucose (30 mM) medium. A scratch wound was created using a 10 μL pipette tip, and cells were incubated for 18 h. The wound closure was then assessed by capturing images and measuring the gap distance using the MCID system (Wang et al., 2020).

Total RNA was extracted from EVs and sciatic nerve tissues using the miRNeasy Mini Kit (Qiagen) and subsequently reverse transcribed. Quantitative real-time RT-PCR (qRT-PCR) was carried out using a TaqMan miRNA assay kit and TaqMan PCR reagents, following the protocol outlined in previous studies (Wang et al., 2020, 2024). The following hydrolysis miRNA primers were used: has-miR-214-3p (Assay ID: 002306, mature sequence: ACAGCAGGCACAGACAGGCAGU). U6 snRNA (Assay ID: 001973, mature sequence: GTGCTCGCTTCGGCAGCACATATACTAAAATTGGAACGATA CAGAGAAGATTAGCATGGCCCCTGCGCAAGGATGACACGC AAATTCGTGAAGCGTTCCATATTTT). The relative expression levels of miRNAs were determined using the 2-ΔΔCt method (Livak and Schmittgen, 2001), with U6 snRNA (Applied Biosystems) serving as the endogenous control.

Western blots were conducted following established protocols (Wang et al., 2015, 2020). In brief, samples were lysed and the protein concentration in the supernatant was quantified using a bicinchoninic acid (BCA) assay kit (Pierce Biotechnology). Equal amounts of proteins were separated by SDS-PAGE and transferred onto Polyvinylidene difluoride (PVDF) membranes. After blocking, the membranes were incubated overnight with primary antibodies at 4°C and followed with secondary antibody (1:1000). A complete list of antibodies employed in this study is provided in Supplementary Table 1. The signals were detected using an enhanced chemiluminescence detection kit (Pierce Biotechnology) and visualized with the FluorChem E System (ProteinSimple).

Generalized estimating equations (GEE) were used to evaluate the effects of group and time on neurological test outcomes. The interaction between group and time was assessed first. If the interaction was not significant, the main effects of group and time were analyzed separately. For multiple group comparisons, one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test was performed. For pairwise group comparisons, Student’s t-tests were used. Data are presented as mean ± standard error of the mean (SE). A p-value of < 0.05 was considered statistically significant.

To investigate the effect of 214-EVs on DPN, we generated and characterized 214-EVs. EVs were isolated from exosome free conditional media of cultured Schwann cells transfected with a miR-214-containing lentiviral vector or a control scramble vector by means of differential ultracentrifugation (Wang et al., 2020, 2024). Following the EV guidelines (MISEV2023) (Welsh et al., 2024), the isolated EVs were characterized using Nanosight nanoparticle tracking analysis (NTA), TEM, and Western blotting. We found that there were no significant differences in size distributions and morphology between 214-EVs and EVs isolated from the media of Schwann cells transfected with the control scramble vector (Scra-EVs). These EVs contained EV marker proteins (Alix, HSP70, TSG101, CD9 and CD63), but not Calnexin, a negative EV control protein (Figure 1). Quantitative RT-PCR analysis revealed that miR-214 levels were significantly elevated in 214-EVs compared to Scra-EVs (Figure 1), indicating that we successfully generated SC-EVs enriched with miR-214.

Next, we examined the effect of 214-EVs on DPN of the HFD/STZ-induced T2D mice. Compared to wild-type mice (WT), 16 weeks post-HFD/STZ administration, the HFD/STZ-induced T2D mice displayed significant DPN symptoms, characterized by marked reductions in motor and sensory nerve conduction velocities (MCV and SCV), and thermal sensitivity as measured by plantar test (Figure 2). We thus randomly treated (I.V.) HFD/STZ mice with 214-EVs and Scra-EVs weekly for 8 consecutive weeks starting 16 weeks post HFD/STZ when mice exhibited DPN. Compared to the saline treatment, HFD/STZ mice treated with naïve-SC-EVs and Scra-EVs exhibited significant attenuation of diabetes-induced reductions in MCV and SCV, along with improvements in thermal sensitivity starting 4 weeks after the treatment, and the improvement persisted for at least 8 weeks. These data are consistent with our previous findings of the therapeutic efficacy of SC-EVs in DPN animal models. However, 214-EVs treatment led to further significant improvement of neurological function at 8-weeks post-treatment compared to Scra-EVs (Figure 2). Additionally, treatment with 214-EVs and Scra-EVs, did not significantly affect blood glucose levels and animal body weight (Table 1). These data indicate that the 214-EVs’ therapeutic benefit is superior to naïve-SC-EVs.

To examine whether 214-EVs treatment affects nerve fibers, morphometric changes of nerve fibers were analyzed. Immunohistochemical analysis revealed that compared to WT mice, HFD/STZ mice at age of 34 weeks exhibited a significant reduction in intraepidermal nerve fiber density (IENFD), as quantified by PGP9.5-positive sensory nerve fibers in footpad tissues. Naïve-SC-EVs and Scra-EVs treatment significantly augmented IENFD compared to saline treatment. However, treatment with 214-EVs led to a further significant increase in IENFD compared to the Scra-EVs treatment, indicating enhanced distal nerve fiber regeneration (Figure 3).

Morphometric analysis of sciatic nerves using toluidine blue staining revealed that HFD/STZ mice exhibited substantial reduction of nerve fiber diameter and myelin sheath thickness, and elevated g-ratio (axon diameter/fiber diameter). Treatment with naïve-SC-EVs and Scra-EVs significantly improved these parameters compared to saline treated HFD/STZ mice. Importantly, 214-EVs treatment resulted in further significant improvements in nerve fiber diameter, myelin thickness, and g-ration, suggesting superior promotion of axonal regeneration and remyelination in diabetic sciatic nerve tissues (Figure 4).

To investigate the cellular effects of 214-EVs on Schwann cells and DRG neurons, we conducted in vitro experiments. DRG neurons isolated from 26-week-old T2D mice and Schwann cells were incubated with high glucose (30 mM) with or without SC-EVs (6 × 10⁹ particles/ml). We found that naïve-SC-EVs and Scra-EVs significantly decreased the inhibitory effects of high glucose on Schwann cell migration and DRG neurite outgrowth. However, treatment with 214-EVs resulted in substantially enhanced Schwann cell migration and DRG neurite outgrowth compared to Scra-EVs treated cells (Figure 5). These in vitro results support for our in vivo finding, indicating that 214-EVs amplify the effect of naïve-SC-EVs on Schwann cell migration and DRG neurite outgrowth.

Quantitative RT-PCR analysis of sciatic nerve tissues revealed that miR-214 expression was substantially reduced in HFD/STZ mice compared to WT controls. Compared with saline treatment, naïve-SC-EVs and Scra-EVs significantly increased miR-214 levels. However, 214-EVs treatment led to a further significant increase in miR-214 expression in HFD/STZ mice (Figure 6).

PTEN and c-Jun have been identified as target genes of miR-214, known to inhibit axonal growth and myelination, respectively (Chen et al., 2019; Zeng et al., 2023). Western blot analysis of sciatic nerve tissues showed markedly elevated PTEN and c-Jun protein levels in HFD/STZ mice compared to WT controls. Treatment with naïve-SC-EVs and Scra-EVs significantly reduced the expression of both inhibitory proteins, which is consistent with our previous studies (Wang et al., 2020, 2024). However, treatment with 214-EVs resulted in a further reduction in PTEN and c-Jun levels (Figure 6). These data demonstrate that diabetes suppresses miR-214 expression and elevates its target proteins, PTEN and cJUN, in sciatic nerve tissue. The enhanced therapeutic effect of 214-EVs appears to be mediated through increased delivery of miR-214, which more effectively downregulates PTEN and c-Jun, thereby promoting axonal regeneration and remyelination.

Inflammation plays a crucial role in the pathogenesis of DPN, with macrophages acting as key neuroinflammatory regulators affecting peripheral nerve tissue integrity (Farmer et al., 2012; Panou et al., 2025). Immunohistochemical analysis revealed a significant increase in CD68 + macrophage accumulation in the sciatic nerves of HFD/STZ mice. Treatment with naïve-SC-EVs and Scra-EVs significantly reduced CD68 + cell infiltration, however, 214-EVs showed a superior reduction in CD68 + cells compared to Scra-EVs treatment (Figure 7).

Western blot analysis of sciatic nerve tissue showed that HFD/STZ mice exhibited elevated expression of pro-inflammatory mediators, TLR4 and NFκB. Naïve-SC-EVs and Scra-EVs treatment reduced these inflammatory markers. However, 214-EVs completely abolished the diabetes-induced elevation of TLR4 and NFκB expression (Figure 7). Bioinformatic analysis identified TLR4 as a predicted target of miR-214, suggesting that 214-EVs suppress neuroinflammation by decreasing CD68 + macrophages and inactivating the TLR4/NF-kB signaling pathway.