The work has been reported in line with the ARRIVE guidelines 2.0.

2.1 Animals and ethical considerations

Experiments were performed on 35 adult male Sprague Dawley rats, weighing between 250 and 300 g (Élevage JANVIER®, Centre d’Élevage Roger JANVIER, Le Genest Saint Isle, France), hosted two per cage in smooth-bottomed plastic cages in a laboratory animal house maintained on a 12:12-h light/dark photoperiod and at 22 °C. Drinking water and rat chow (Safe®, Augy, France) were available ad libitum. Animals were housed during 2 weeks before the initiation of the experiment. All animals were weighed before each surgery and once a week until the end of the study.

Experiments were performed according to the French law (Decrees and orders No. 2013-118 of 01 February 2013, JORF No. 0032) on animal care guidelines and after approval by animal Care Committees of Aix-Marseille Université (AMU) and Centre National de la Recherche Scientifique (CNRS). The authorization number granted by the French Ministry of Higher Education, Research, and Innovation (MESRI) is APAFIS#41012. All persons were licensed to conduct live animal experiments and all room have a national authorization to accommodate animals (License No. B13.013.06). Furthermore, experiments were performed in accordance with the recommendations provided in the Guide for Care and Use of Laboratory Animals (U.S. Department of Health and Human Services, National Institutes of Health), with the directives 86/609/EEC and 010/63/EU of the European Parliament and of the Council of 24 November 1986 and of 22 September 2010, respectively, and with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines.

After surgery, the animals were placed under a heat lamp until their thermoregulation was restored. The health status of the animals was monitored daily and animals showing signs of distress such as vocalization, lethargy, hyperactivity, significant weight loss (15–20%), and self-mutilation behavior were euthanized (no animal were excluded). They received subcutaneous injections (3 ml) of glucose-enriched saline solution to replace the fluid lost during the surgical procedure. Buprenorphine (0.03 mg/kg, 0.3 mg/ml, Bruprécare® Multi-dose, Axience Santé Animale SAS, Pantin, France) was administered subcutaneously before and after surgery daily for 3 days. A broad-spectrum antibiotic (Oxytetracycline, 400 mg/l, Sigma Aldrich, Saint-Quentin Fallavier, France) was diluted in the drinking water for 1 week to prevent any infections. Postoperative nursing care also included visual inspection for skin irritation or pressure ulcers, followed by cleaning of the hindquarters with soap and water and rapid towel drying of the fur.

2.2 Protocol design and experimental groups

After a two-week acclimation period, which involved familiarization sessions lasting one hour per day, three days per week, on—a walking corridor -, reference values (PRE-) were measured for Peroneal Functional Index (PFI) test which was used to follow the progress of recovery. Following the acclimation period, the animals were randomly assigned to five groups:

1.

Control group (Control, n = 7): No surgery was performed in this group.

2.

Gold Standard group (GS, n = 7): A 7-mm segment of the left peroneal nerve was excised and immediately autografted in inverted position.

3.

Vein graft (VE) group (n = 7): A 7-mm segment of the peroneal nerve was removed and a 1.5-cm vein conduit was immediately grafted between the two nerve stumps.

4.

Vein graft and freshly isolated extracellular vesicles (VE-fEV) group (n = 7): A 7-mm segment of the peroneal nerve was removed and a 1.5-cm vein conduit was immediately grafted between the two nerve stumps and filled with 5 million fresh EVs.

5.

Vein graft and cryoconserved extracellular vesicles (VE-cEV) group (n = 7): A 7-mm segment of the peroneal nerve was removed and a 1.5-cm vein conduit was immediately grafted between the two nerve stumps and filled with 5 million frozen EVs.

One week (W1) after surgery, animals underwent weekly assessments of sensory and motor recovery in the hindlimb. Evaluations were performed from Week 1 (W1) to Week 12 (W12) and compared to the baseline (PRE-) values. At the end of the twelve-week period the mechanical properties of the Tibialis anterior muscle were quantified using twitch measurements from a motor standpoint. Additionally, sensory aspects of the nerve were studied by analyzing type III-IV metabosensitive afferents in response to (1) electrically induced muscle fatigue (EIF) and (2) intra-arterial injection of potassium chloride (KCl) and lactic acid (LA) solutions. Then, animals were euthanized using an intraperitoneal overdose of pentobarbital sodium (1 ml) at a dosage of 280 mg/kg (i.p., Euthasol® Vet., Dechra Veterinary Products SAS, Suresnes, France) according to ethical recommendations.

Animals were randomly assigned to experimental groups using a randomized table in Excel. Each animal was identified by a number throughout the experiment to prevent disclosure of its experimental group. Group affiliations were revealed after data analysis conducted by blinded researchers. Potential confounders were minimized through standardization of treatment timing, housing conditions, and experimenter identity to ensure that all groups were similarly affected by potential variabilities.

2.3 Surgery procedure

Surgery of the peroneal nerve was performed as previously described [11]. Animals were deeply anesthetized using 3% isoflurane (Isoflurin®, Axience Santé Animale SAS) with a pre-anesthetic injection of buprenorphine (0.03 mg/kg, 0.85 mg/ml) administered 30 min prior. All surgical procedures were performed under aseptic conditions using binoculars.

The animals were positioned in a ventral position, and the left hindlimb was shaved and disinfected using Vétédine® solution 10% (Vetoquinol S.A., Magny-Vernois, France). In the GS, VE, VE-fEV, and VE-cEV groups, the peroneal nerve of the left limb was meticulously dissected from the surrounding tissues and cut to a length of 7 mm. In the GS group, the nerve segment was immediately replaced by the nerve segment in an inverted position and sutured (Ethilon 9–0, Ethicon Inc., Johnson & Johnson, Somerville, New Jersey, USA) at the two free nerve stumps.

For the VE, VE-fEV, and VE-cEV groups, a 1.5 cm branch of the femoral vein was harvested from the contralateral side of the nerve injury. The harvested vein segment was washed in saline solution (NaCl 0.9%) and immediately used. The two nerve stumps were then inserted into the vein, leaving a 7 mm gap between the proximal and distal nerve stumps. To secure the graft, three or four 9–0 monofilament non-absorbable sutures (Ethilon® 9–0, Ethicon Inc.) were used for each stump, along with biological thrombin/fibrinogen glue (Tisseel, Baxter®, Glattpark, Switzerland).

For the VE-fEV, and VE-cEV groups, 5 million fresh or stored OEMSC-derived EVs were suspended in 10 μl of phosphate-buffered saline (PBS). The number of 5 million EVs was determine according to previous in vivo [11] and in vitro [20] studies, and according to literature [21]. The EV were then injected into the vein using a 10 μl Hamilton syringe (Hamilton Company, Bonaduz, Switzerland).

2.4 Isolation and purification of extracellular vesicles from human olfactory ecto-mesenchymal stem cells (OEMSC-derived EV)2.4.1 Collection of olfactory mucosa biopsies

OEMSCs were cultured based on a previously described protocol [3]. Biopsies were unilaterally collected by an ear, nose and throat (ENT) specialist surgeon, with informed consent from the patients participating in the Nose study. The biopsies were taken at the level of the middle turbinate arch using Morscupula forceps and an endoscope. Before the planned procedure and while the patient was under deep anesthesia, a 2 mm2 biopsy was obtained and transferred to a sterile tube filled with culture medium, penicillin, and gentamycin (alpha MEM, Macopharma BC0110020, Tourcoing, France).

2.4.2 Culture of olfactory ecto-mesenchymal stem cells

The olfactory mucosa, located on the nasal septum, was harvested from three healthy donors and placed in a Petri dish filled with DMEM/HAM F12. After triple washing to remove mucus, the biopsies were incubated in a Petri dish containing 1 ml of dispase II solution (2.4 IU/ml) for 1 h at 37 °C. The olfactory epithelium was then separated from the underlying lamina propria using a micro spatula. Once purified, the lamina propria was cut into small pieces using two 25-gauge needles and transferred to a 15-ml tube filled with 1 ml of collagenase NB5 (1U/ml, Nordmark Biochemicals, Uetersen, Germany). Following a 10-min incubation at 37 °C, the tissue was mechanically dissociated, and the enzymatic activity was halted by adding 9 ml of Ca- and Mg-free PBS. After centrifugation at 300 g for 5 min, the cell pellet was resuspended in αMEM (Gibco Thermofisher©, Courtaboeuf, France) supplemented with 10% platelet lysate (Macopharma®) and seeded onto plastic culture dishes. The culture medium was refreshed every 2–3 days. Once reaching 80% confluence, typically between days 7 and 10, the cells were cryopreserved in liquid nitrogen for future use.

2.4.3 Purification of OEMSC-derived EVs

OEMSC used to produce EVs were maintained between passages 2 and 7 to ensure phenotypic stability and consistent EV yield. Cells were cultured in T175 flasks for amplification at the density of 10,000 cells/cm2 in a proliferative medium (αMEM, supplemented with platelet lysate (5%), heparin (2UI/ml), penicillin and streptomycin). On day 4, stem cells were washed with PBS to remove any contaminating vesicles from the platelet lysate. Subsequently, the cells were cultured in containing αMEM supplemented with insulin, transferrin, and selenium (ITS, 1%) for 4 additional days. Then, the cell supernatants were collected and centrifuged at 300 g for 5 min at room temperature and at 2,500 g for 15 min at 20 °C to obtain the vesicular supernatant. Extracellular vesicles were isolated by ultracentrifugation at 100,000 g for 90 min at 4 °C. Pellets of EVs were resuspended in PBS (less than 1 ml). EVs were purified using size exclusion chromatography columns (qEVoriginal Size Exclusion Column®, IZON, Lyon, France). The purification procedure followed the instructions provided by the supplier. Purified OEMSC-derived EVs were either kept at 4 °C for immediate use (referred as "freshly purified extracellular vesicles", fEV) or cryoconserved at − 80 °C for 6 months (referred as "cryoconserved extracellular vesicles”, cEV).

2.4.4 Characterization and quantification of OEMSC-derived EVs

Extracellular vesicles (EVs) derived from OE-MSCs were characterized according to MISEV guidelines using a combination of complementary techniques. Size distribution was assessed by Tunable Resistive Pulse Sensing (TRPS) using an Exoid instrument (Izon Science, Christchurch, New Zealand) equipped with NP400 nanopores. Measurements were performed under controlled conditions (0.3–0.5 V voltage, 0.8 kPa pressure) using PBS filtered through a 0.1 µm membrane as the electrolyte, and calibration was carried out using CPC400 standard beads (mean diameter 350 nm). Surface marker expression was analyzed by high-sensitivity flow cytometry. EVs were incubated with FITC-conjugated Annexin V and antibodies against CD59, CD29, and CD41, and analyzed using a Navios 3-laser cytometer (Beckman Coulter). The protocol was standardized using Megamix-Plus FSC beads to ensure appropriate size-based discrimination. Western blot analysis was performed to confirm the expression of EV markers, including CD63, CD81, CD9, and caveolin-1. Membranes were also probed for cellular contaminants such as β-actin, β-tubulin, GAPDH, and albumin to assess purity. Immunoreactive bands were visualized using enhanced chemiluminescence (ECL) and imaged with a G-Box system (GeneSys, Cambridge, UK).

2.5 Functional assessment of hind limb recovery

To evaluate functional changes, footprints were recorded and analyzed every week from W1 to W12 after the surgery, using a paper track method as previously described [11] and compared to PRE- values. Throughout assessments, experimenters involved in data collection were blinded to the treatment group, ensuring unbiased evaluations.

PFI test. The PFI was calculated using the formula described by [22]: PFI = 174.9 × [(ePL − nPL)/nPL] + 80.3 × [(eTS − nTS)/nTS] − 13.4. This index considers the parameters measured for both the normal (n) and operated (e) feet, including the footprint length (PL), which is the longitudinal distance between the tip of the longest toe and the heel, and the total toe spreading (TS), which is the cross-sectional distance between the first and fifth toes. The recovery rate of the PFI was assessed on a scale ranging from − 100 to − 13.4, where − 13.4 represents normal function and − 100 indicates total failure. Footprints were collected and analyzed from the first (W1) to the twelfth (W12) week post-surgery.

2.6 Electrophysiological recordings

At Week 12, rats were anesthetized using an intraperitoneal injection of Ketamine solution (75 mg/kg) and medetomidine (0.5 mg/kg). The following procedures and measurements were performed as previously described [11]. In brief, the left peroneal nerve was dissected and carefully separated from the surrounding tissues over a length of 3–4 cm. A catheter was inserted into the right femoral artery and advanced to the bifurcation of the descending abdominal aorta. This catheter was used to transport chemicals (potassium chloride, KCl [20 mM in 0.5 ml of saline] and lactic acid, LA [1 mM in 0.1 ml of saline]) to the contralateral muscle. It also allowed free blood flow into the muscles of the lower left limb.

Twitch measurement. Two stimulation electrodes with an inter-electrode distance of 1 mm were placed on the surface of the peroneal nerve. The contractile response of the Tibialis anterior muscle to nerve stimulation was induced using a neurostimulator (S88K stimulator) that delivered rectangular single shocks (duration: 0.1 ms, frequency: 0.5 Hz). The response was measured using an isometric strain gauge (micromanometer 7001) attached to the tendon of the Tibialis anterior muscle. Several parameters, including amplitude (A), maximum contraction rate (MCR), and maximum relaxation rate (MRR), were recorded. MCR and MRR were normalized to the amplitude of the twitch (MCR/A and MRR/A, msec-1). The twitch data was recorded using the Biopac MP150 system (Biopac Systems Inc, Goleta, CA, USA), sampled at 2000 Hz, and analyzed using AcqKnowledge 3.7.3 software (Biopac Systems Inc.).

Afferent activity. A monopolar tungsten electrode was positioned under the Tibialis anterior nerve, which was immersed in paraffin oil. Nerve activity was recorded using a differential amplifier and filtered between 30 Hz and 10 kHz (P2MP®, 5104B, Marseille, France). The afferent discharge was recorded and analyzed using pulse window discriminators and the Biopac AcqKnowledge software. The response of muscle afferents was recorded after a 3-min Electrically-induced muscle Fatigue (EIF) period and intra-arterial bolus injection of KCl (20 mM in 0.5 ml of saline) or LA (1 mM in 0.1 ml of saline) solutions. EIF was elicited using the S88K stimulator, which delivered pulse trains to the muscle surface electrode (pulse duration: 0.1 ms, frequency: 10 Hz, duty cycle: 500/1500 ms). The discharge rate of nerve afferents was averaged for a 30-s period before (regarded as baseline discharge) and after EIF or metabolite injection, and the increase in average afferent discharge rate was expressed as a percentage of the baseline discharge rate. A 20-min recovery period was allowed between different interventions.

2.7 Euthanasia, muscular atrophy measurement and biopsy collection

At the end of the electrophysiological recordings, animals were euthanized using an intraperitoneal overdose of pentobarbital sodium (1 ml) at a dosage of 280 mg/kg (i.p., Euthasol® Vet., Dechra Veterinary Products SAS).

Following euthanasia, the left Tibialis anterior muscle was immediately harvested, weighed and preserved in a PBS solution containing 4% paraformaldehyde (PFA) then transferred in stored in sodium azide solution (0.1%) at 4 °C for further histological investigation. The muscle weight was normalized by dividing it by the body weight, and the muscle/body weight ratio was calculated as an indicator of muscle atrophy.

Additionally, the peroneal nerves (n = 4 per group) were excised, washed in phosphate-buffered saline (PBS) to remove contaminants, and stored in a PBS solution containing 4% paraformaldehyde until histological analysis. Some of the nerves (n = 3) were previously immersed in a 2% glutaraldehyde solution containing PBS for 24 h before being transferred to the 4% paraformaldehyde solution.

2.8 Immunohistochemistry, histology and quantification

Immunohistochemstry: For each group of animals, nerves were cut in three parts (proximal end, medial part, and distal end). Specimens fixed in PFA were processed in HistoGel™ (Epredia™ HG4000012, ThermoFisher Scientific) then embedded in paraffin. After embedding, sections of 5 μm were performed using a vibrating blade microtome (Leica Biosystems, PRID: SCR_016495) and collected on coated slides. Sections were dewaxed in three changes of Histo-Choice® (Sigma Aldrich-Merck) for 10 min each, then rehydrated by transferring in decreasing concentrations of ethanol (100%, 95%, 70%, 50%), twice for 10 min each, and rinsed in two changes of distilled water, for 5 min each. The distal, medial, and proximal sections were immunostained with a mouse monoclonal antibody raised against the light chain of neurofilament protein (NF-L 70 KDa, Sigma Aldrich-Merck, dilution: 1:500). After washing, an appropriate biotinylated-conjugated secondary antibody was applied to the sections. The final staining step was performed using diaminobenzidine kit (DAB substrate kit, ab64238, abcam). Sections were delipidated in xylene and embedded in Eukitt mounting medium (Sigma Aldrich-Merck).

Histology (p-Phenylenediamine staining): For myelin quantification, nerve specimens were washed three times for 5 min each in PBS and immersed into osmium tetroxide (2%) solution for 1 h. After 3 washes in PBS, 5 min each, samples were immersed in increasingly concentrated acetone solutions (50%, 35 min; 70%, 1 h; 95%, 1 h; 100%, 2 h, respectively). Samples were immersed in increasingly concentrated araldite solutions (50%, 3 h; 80%, 5 h, respectively). Specimens were immersed in araldite 100% and placed in heat chamber (80 °C) during 12 h for resin polymerization. After inclusion, semithin sections (0.8 μm) were cut using an ultramicrotome and collected on coated slides. A solution of 1% p-phenylenediamine was prepared in absolute methanol and kept at room temperature in a closed dark glass jar protected from the light and evaporation 5 days before using it. After 5 days, the solution was filtered with a 0.2 μm nylon sterile filter (Millipore, Molsheim, France). Slides were stained by being immersed for 10 min in the staining solution, washed in tap water then in distilled water, and dried for 1 h on a hot plate. Sections were mounted in Eukitt medium.

Microscopy and quantification: The stained sections were examined using an Apotome microscope (Apotome v2, AxioObserver Z1, Zeiss) associated with high-resolution camera (Camera CMOS Orca flash 4.0 v2, Zeiss, Oberkochen, Germany). The slides were digitized and analyzed with ImageJ (NIH) software. Axon numbers were counted for several sections in proximal, medial, and distal parts for each group of animals. Number of myelinated axons was assessed using a deep-learning tool by Axonet 2.0. The G-ratio (i.e., the ratio between the diameter of the axon and the outer diameter of the myelinated fiber) was calculated, using a semi-automated quantification using Myeltracer software [23] by assessing 100 fibers per section, randomly and blindly selected.

2.9 Statistical analysis

Results obtained from behavioral locomotor test (PFI), electrophysiological recordings (EIF, KCl and lactate injection), muscle properties (weight/body weight ratio, contraction properties) and histological data were compared between all experimental groups. Data processing was performed using Rstudio for statistical computing and graphics (Version 2024.04.2 Build 764, 2024). Significant differences were determined using non parametric tests since our data are not normally distributed. For the one-way experimental design, we applied the Kruskal–Wallis one-way analysis of variance, followed by Dunn's test for post-hoc comparisons. For the two-way experimental design (experimental groups × time points), we utilized the Scheirer–Ray–Hare test, an extension of the Kruskal–Wallis test adapted for factorial designs followed by Dunn's test. Data were expressed as median ± SD. Difference was considered significant when p < 0.05.