Materials and cell culture
The RSC96 cells used in this work were obtained from the Shanghai Cell Bank, Chinese Academy of Sciences, and cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% antibiotics (100 U/mL penicillin and 100 μg/mL streptomycin). FBS, phosphate buffered solution (PBS), DMEM, penicillin-streptomycin, and trypsin-EDTA were obtained from Life Technologies Corporation (29851 Willow Creek Road, Eugene, OR 97402, USA).
Herein, to study the influence of magnetic actuation on the repair phenotypes of Schwann cells, we cultured RSC96 cells (rat Schwann cells, a type of neuroglial cells in peripheral nervous system) on 35 mm imaging ibidi petri dishes (ibidi, 80156, Martinsried, Germany). For in vitro cell experiments, four experimental groups were designed and established: (1) the normal control group (labeled the ‘Normal’ group and consisting of RSC96 cells cultured under normal conditions, without SPIONs and with a null MF), (2) the magnetic actuation group (labeled the ‘SPIONs+MF’ group and consisting of RSC96 cells treated with both SPIONs and a MF), (3) the positive control group (labeled the ‘c-Jun’ group and consisting of RSC96 cells treated with 0.2 µM anisomycin, which activates the JNK pathway and its downstream c-Jun transcriptional regulation mechanism and then induces the repair phenotypes in RSC96 cells), (4) the SPIONs control group (labeled the ‘SPIONs’ group and consisting of RSC96 cells treated with SPIONs and no MF), and (5) the MF control group (labeled the ‘MF’ group and consisting of cells treated with an MF but no SPIONs).
The c-Jun is a major phosphorylation target of JNK (c-Jun N-terminal kinase). Anisomycin is a potent specific agonist of JNK at a concentration of 0.2 µM [28]. In the positive control group, anisomycin was used to activate JNK, thereby activating the expression of the transcription factor c-Jun, which is critical for inducing the repair function of Schwann cells [12]. The c-Jun-activated Schwann cells were used as the positive control for repair Schwann cells. All cultures were maintained in an incubator at 37 °C in a humidified atmosphere with 5% CO2.
Synthesis and characterization of fluorescent-magnetic bifunctional SPIONs
The fluorescent-magnetic bifunctional SPIONs used in our study (Fe3O4·Rhodamine 6G@polydopamine superparticles) were prepared following our previously established protocol [27, 29]. In brief, Fe3O4 NPs (nanoparticles) with average diameter of 5.8 nm were synthesized following the classical thermal decomposition method. Subsequently, Fe3O4 NPs were mixed with Rhodamine 6G to obtain Fe3O4·Rhodamine 6G SPs (superparticles) with an average diameter of 50 nm. Finally, PDA (polydopamine) was coated on the surface of Fe3O4·rhodamine 6G SPs to obtain Fe3O4· Rhodamine 6G@PDA SPs (SPIONs). After the synthesis of SPIONs, their physical, optical and magnetic properties were characterized respectively. The detailed preparation and characterization of SPIONs are presented in the Supporting Information.
In vivo biodistribution and biocompatibility of SPIONs
To evaluate the in vivo toxicities and biodistribution of SPIONs, rats were intravenously injected with SPIONs at a dose of 1 mg/kg body weight through the caudal vein (200 µL). The normal control group was injected with 200 μL normal saline through the caudal vein. The animals were sacrificed by inhalation of carbon dioxide followed anesthetized at various time points (1, 2, 3, 7 and 14 days) after intravenous injection, and the aorta was approached through an abdominal incision and cannulated just distal to the emergence of the renal arteries. Blood samples were collected from each rat for serum biochemical measurements, including ALT (alanine transaminase), AST (aspartate transaminase), TP (total protein), ALB (albumin) and Cr (creatinine). After the blood samples were collected, 500 mL of buffered normal saline was injected, and the right atrium was transected to permit drainage of the blood and injected solution to clean the vascular beds of all major organs of the body. Next, the major organs (heart, liver, spleen, lung, kidney, and brain) were harvested and washed twice with cold PBS. A portion of the organs was collected and fixed in 4% paraformaldehyde (PFA) for histopathological analysis. For conventional histopathological analysis, the major organs of rats were collected, fixed, dehydrated, embedded in paraffin, sectioned, stained with HE (hematoxylin and eosin), and examined using a digital optical microscope (Olympus BX51). The remaining organs were accurately weighed (Wtissue) and used for in vivo biodistribution evaluation. The amount of iron inside the different organs (WFe) was measured by ICP-AES (inductively coupled plasma-atomic emission spectrometry) with a PerkinElmer Optima 3300DV. The iron content per gram of the organ (relative weight, WFe/Wtissue) can be obtained by dividing the mass of iron in the organ (WFe) by the mass of the organ (Wtissue). By subtracting the WFe/Wtissue in the normal control group from the WFe/Wtissue at different time points (1, 2, 3, 7 and 14 days) after injection, the quantity of exogenous iron from intravenous injection (SPIONs) in major organs can be obtained.
In order to further study the neurotoxicity and neuronal affinity of SPIONs, 300 µg/mL SPIONs (20 µL) was injected locally under the epineurium of the sciatic nerve. At various points (1, 3, 7, and 14 days) following local injection of the sciatic nerve, nerve tissue was collected for histopathological analysis to identify possible neurotoxicity of SPIONs. At the same time, fresh-frozen sections were made, and CLSM (confocal laser scanning microscopy) analysis were performed to determine the neuronal affinity of SPIONs. Subsequently, ultrathin sections were made, and TEM (transmission electron microscopy) were performed to determine the localization and distribution of SPIONs in the sciatic nerve microstructure. Finally, the amount of iron in nerve tissues at different time points after local nerve injection was detected by ICP-AES, and the neuronal affinity of SPIONs was further quantitatively analyzed.
Preparation of MF device and quantification of magnetic forces
To obtain optimal results, we designed a simple MF generating device for cell experiments in vitro prior to use in vivo based on our previous studies [27, 29]. A 50 mm × 30 mm × 10 mm perpetual cuboid neodymium magnet (NdFeB N48) was placed on the right side of the ibidi petri dish to expose RSC96 cells to a gradient MF (Figure 1a).
For the in vivo experiment of rats, we designed an MF generating device composed of four circular neodymium magnets (NdFeB N48H) with an inner diameter of 70 mm, an outer diameter of 150 mm and a thickness of 15 mm. Each pair of circular magnets form a group, with a 15-mm gap between the two groups (Figure 2a). There is a large MF gradient inside the circular neodymium magnets (Figure 2b). The gradient magnetic field environment was digitally simulated with Comsol Multiphysics 4.3b software (Comsol Multiphysics GmbH, Goettingen, Germany) (Figure 2c). A digital Gauss meter (Model 425, Lake Shore Cryotronics) was used to measure the magnetic flux density induced by the neodymium magnet setup. The rat was placed inside the circular magnets 30 min a day after surgery, and the sciatic nerves were exposed at the center of the gap between the two groups of circular magnets. SPIONs in the sciatic nerve were induced to interact with the gradient MF to produce nano-scale magnetic stimulation of the sciatic nerve.
The exact explanation of the quantification of magnetic forces has been described in our previous studies [27, 29]. A magnetic particle with a magnetic momentum (m) can generate magnetic force F in a magnetic flux density gradient (∇B):
F = (m · ∇)B Equation (1)
In our experimental setup, we can measure the value of the gradient magnetic field dB/dr, the density ρ, the volume V of SPIONs, and the magnetization M of SPIONs in this field environment. We can assume the net force FSPION of each SPION due to a combination of parameters:
The quantity of exogenous iron in sciatic nerve tissue could be measured, from which the number of SPIONs in the nerve could be calculated. The sciatic nerve will thus be subject to a force Fnerve given by FSPION multiplied by the number of SPIONs in the nerve:
Establishment of rat models of sciatic nerve crush
All works involving animals were in accordance with the National Committee for Science and Technology of Standardized Experimental Animals. Ethical approval for all experiments was granted by the Animal Welfare and Ethical Review Committee of the First Hospital of Jilin University (Approval No. 20210565), and all efforts were made to minimize animal suffering. Eight-week-old Sprague Dawley rats (male, 200–220 g), purchased from Liaoning Changsheng Biotechnology Co., Ltd., were used in all experiments. Rats were deeply anesthetized with an intraperitoneal injection of mixed narcotics (100 mg/kg ketamine plus 10 mg/kg xylazine).
After the rats were fully anesthetized, routine skin preparation and disinfection of the operative field on the lateral aspect of the right thigh were performed. The long sciatic nerve was exposed between the gluteus maximus and quadriceps muscles through a 2 cm long posterolateral longitudinal straight incision on the right thigh. For the nerve crush operation, the sciatic nerve was crushed using a pair of delicate forceps (Fine Science Instruments) two times (30 s each, at 10-second intervals) at the same site 10 mm above the bifurcation into the tibial and common peroneal nerves to create a 2-mm-wide crush lesion. A single 10/0 nylon suture (Mononylon, Ethicon) was passed through the epineurium at the point corresponding to the crushed site to facilitate its identification in the subsequent procedures. For SPION injection, 20 µL SPION solution (300 µg/mL) was slowly injected under the epineurium with a 20 µL Hamilton microsyringe at the distal end of the sciatic nerve crush site. All procedures followed a standard microsurgery technique under a stereomicroscope (Leica Microsystems, Wetzlar, Germany).
Animals were divided into five experimental groups: (1) the normal control group (labeled the ‘Normal’ group and consisting of normal rats without any treatment), (2) the magnetic actuation group (labeled the ‘Crush+SPIONs+MF’ group and consisting of rats that underwent sciatic nerve crush and were then treated with both SPIONs and an MF), (3) the SPIONs control group (labeled the ‘Crush+SPIONs’ group and consisting of animals that underwent sciatic nerve crush and SPION administration but no MF exposure), (4) the MF control group (labeled the ‘Crush+MF’ group and consisting of animals that underwent sciatic nerve crush and MF exposure but did not receive SPIONs), (5) and the surgical control group (labeled the ‘Crush’ group and consisting of animals that underwent only sciatic nerve crush and did not receive SPIONs or MF exposure).
All animals undergoing surgery were given appropriate postoperative analgesia and monitored daily. Animals were housed under a 12/12-hour light/dark cycle with free access to food and water. To investigate whether magnetic stimulation can promote recovery after sciatic nerve crush injury, rats in the magnetic actuation group and MF control group were placed in an MF generating device for 30 min every day after surgery and subjected to a daily gradient MF effect. At 3, 7, 14 and 21 days postoperatively, the histomorphology, motor behavior, electrophysiological function and regeneration-related molecular markers of the sciatic nerve in five experimental groups were measured and analyzed. In this study, to investigate the effect of magnetic actuation on the specific repair phenotypes of denervated Schwann cells after nerve injury, the sciatic nerve tissue segment at the distal end of the crush site was selected and analyzed.
Fresh-frozen sections and CLSM
The distribution and localization of SPIONs in nerve tissue were observed through frozen sections by using CLSM imaging after sub-epineurial injection at different times. Animals were euthanized 1, 3, 7, and 14 days after sub-epineurial injection, and sciatic nerve tissue containing the SPION injection site and its distal 2 cm was harvested as quickly as possible. The sciatic nerve tissue was placed in a special small box (approximately 3 cm in diameter). The tissue was immersed in optimal cutting temperature (OCT) embedding compound, and the container was held steady and fat while beign placed in a small cup containing liquid nitrogen. After the frozen block is made, it can be put into the freezer slicer. On the sciatic nerve, a continuous cross section was made from the injection site to the distal end with a thickness of 10 μm. Nuclei were stained with DAPI (4,6-diamino-2-phenyl indol, Ex/Em: 405/430–470 nm). The localization and persistence of SPIONs (Ex/Em: 488/500−580 nm) in the nerve was analyzed by CLSM. The images were captured with a 60× oil immersion objective at 3.2× magnification. Subsequently, as an additional validation experiment, TEM was used to directly observe the presence and localization of the SPIONs in the nerve ultrastructure.
Optical microscopy and TEM analysis
Semithin and ultrathin sections with optical microscopy and TEM observations were used for the quantitative morphological analysis of sciatic nerve regeneration. Distal sciatic nerves to the crush site at 3, 7, 14 and 21 days after injury were dissected and postfixed with 4% PFA and 3% glutaraldehyde in 0.1 m phosphate buffer. Nerves were osmicated with 1.5% osmium tetroxide for 90 min, dehydrated and embedded in epoxy resin. A series of 5-µm-thick semithin transverse sections 10 mm distal to the lesion were cut using a PowerTome-XL ultramicrotome (RMC, USA) and stained with 1% toluidine blue for 2–3 min for optical microscopy examination (IX51, Olympus Corporation, Tokyo, Japan). All semithin sections were observed and photographed at 200× and 1000× magnification. Ultrathin sections (100 nm) were cut immediately after the series of semithin sections by means of the same ultramicrotome and mounted onto 100 mesh Cu grids coated with formvar. The sections were stained with saturated aqueous solution of uranyl acetate and lead citrate. Pictures were observed using TEM (EP 5018/40/Tecnai Spirit Biotwin 120 kV, FEI Czech Republic s.r.o., Holland) operating at 120 kV and then analyzed using image analysis software (Image-Pro Express, version 6.0.0.319, Media Cybernetics, Silver Springs, MD, USA). The type and number of various nerve fibers as well as the myelin sheath thickness and G-ratio (axon/fiber diameter ratio) of myelinated nerve fibers can be accurately identified and measured. In all the animals, the nerves were crushed at the same site and then sliced at the same location from the crush point.
Immunofluorescent staining of sciatic nerve
At the described timepoints (3rd, 7th, 14th and 21st days) following surgery, nerves were dissected, rinsed and fixed in 4% PFA for 5 h at 4 °C. Then, the sciatic nerves were sectioned using a paraffin section system. Nerve sections were permeabilized with 0.2% Triton X-100 for another 15 min at room temperature and blocked for 1 h in 10% normal goat serum (1:50, DAKO, USA) at room temperature. The sciatic nerve sections were stained for double immunofluorescence using mouse anti-neurofilament heavy chain antibody (1:500, Santa, sc-32729) and rabbit anti-S100β antibody (1:100, Abcam, Ab52642). Nerve sections were coincubated with primary antibodies overnight at 4 °C, followed by incubation with goat anti-mouse immunoglobulin G (IgG) secondary antibody with Alexa Fluor 488 conjugate (1:100; Abcam) and goat anti-rabbit IgG secondary antibody with Alexa Fluor 594 conjugate (1:100; Abcam) for 2 h at room temperature. Nuclei were stained with DAPI (Ex/Em: 405/430–470 nm). All procedures were accompanied by rinsing three times in PBS for 5 min each. Finally, the slides were photographed with CLSM to obtain the nerve microstructure of the sciatic nerve during regeneration.
Sciatic functional index (SFI) analysis
For the SFI analysis, animals were tested in a confined walkway measuring 60 cm long and 10 cm wide, with a dark shelter at the end. White paper was placed on the floor of the rat walking corridor. Prior to any surgical procedure, all rats were trained to walk in the corridor. The rats were held by the back, and their hind feet were pressed down onto a stamp pad soaked with water-soluble black ink. The animals were immediately allowed to walk along the corridor, leaving their footprints on the paper. The tracks left by the walking animals were recorded at days 3, 7, 14 and 21 postoperatively.
The tracks were evaluated for three different parameters: (1) distance from the heel to the third toe, the print length (PL); (2) distance from the first to the fifth toe, the toe spread (TS); and (3) distance from the second to the fourth toe, the intermediary toe spread (ITS). All three measurements were taken from the experimental (E) and normal (N) sides. Using the following formula derived by Bain et al., the SFI is calculated as follows [30]:
SFI=-38.3×(EPL-NPL)/NPL+109.5×(ETS-NTS)/NTS+13.3×(EIT-NIT)/NIT-8.8 Equation (4)
The walking track analysis clearly showed that rat footprint measurements could reflect the muscle functional status of the hind limbs [31]. The PL is dependent on posterior tibial nerve function through gastrocnemius activation, whereas TS and ITS reflect common peroneal nerve innervation of the extensor and intrinsic muscle of the foot. In response to a sciatic nerve lesion, the footprints characteristically demonstrate an increased PL and decreased TS and ITS. The SFI is usually negative after nerve injury, and a higher SFI indicates better function of the sciatic nerve. An SFI score oscillating around 0 is considered normal, whereas an index of -100 indicates total impairment.
Electrophysiological analysis of the sciatic nerve
At the end of the survival period, electroneuromyography (ENMG) evaluation was performed under general anesthesia and was carried out with a Neuromatic 2000 M/C Neuro-Myograph (Dantec Elektronic Medicinsk Og Videnskabeligt Maleudstyr A/S, Skovlunde, Denmark). Set up an electrical stimulator. Tape a pair of acupuncture needles (0.25 mm × 25 mm) with a negligible impedance [<1 Ω]) and 3 mm between them to create electrodes for stimulation. The stimulator and the electrode were connected to a data acquisition unit to take the incoming signals and convert them into digital signals that could be processed with computer software. The sciatic nerves were exposed on both sides under the surgical stereomicroscope as described previously. The ground needle was inserted in the quadriceps femoris muscle of the hindlimb to connect the signal ground plug. Starting with the right hindlimb and placing the recording electrode into the triceps calf, the reference electrode in the Achilles tendon, and the stimulation electrode proximal to the crush lesion site in the sciatic nerves were used. Moisten these electrodes with saline. A stimulation amplitude of 10 mV was chosen, and the CMAPs (compound muscle action potentials) were recorded. The sciatic nerve was stimulated proximal and distal to the crush site twice through stimulation electrodes twice, proximal at the level of the sciatic notch, and distal at the level of the popliteal fossa. The latency of the evoked muscle action potentials was recorded, and the prolongation of latency between two stimuli was calculated. Finally, the distance between the two sets of stimulation electrodes was accurately measured on the sciatic nerve, and the MNCV (motor nerve conduction velocity) was calculated. Both experimental (right) and normal (left) nerves were measured.
Quantitative real-time PCR (qRT-PCR) analysis
Cell or sciatic nerve samples from different groups were washed three times with precooled PBS. After freezing and grinding, 1 mL of TRIzol was added to every 50 mg of tissue, and total RNA of RSC96 cells and sciatic nerve tissues was extracted. The detailed protocol and primer sequence information of qRT-PCR are presented in the Support Information. The relative mRNA expression in different groups was calculated by using the 2-ΔΔCT method. Each reaction was performed three times.
Western blot (WB) analysis
Cell or sciatic nerve samples from different groups were washed with PBS and subsequently resuspended in radioimmunoprecipitation assay (RIPA, Beyotime, Shanghai, China) lysis buffer. Tissue lysates were then collected by centrifugation (12,000 rpm for 15 min at 4 °C). Protein concentration was determined using a bicinchoninic acid (BCA, Thermo Scientific, California, USA) protein assay kit. Proteins were separated by 10% SDS polyacrylamide gel electrophoresis (SDS-PAGE), followed by transfer onto polyvinylidene difluoride (PVDF) membranes (Millipore, Bedford, MA, USA). Then, the PVDF membranes were blocked with 5% nonfat milk in Tris-buffered saline solution for 1 h. Subsequently, the membranes were incubated overnight at 4 °C with the following primary antibodies: rabbit monoclonal anti-Beclin1 (1:1000, Abcam, catalog number ab207612, Cambridge, UK), rabbit polyclonal anti-LC3B (microtubule associated protein1 light chain 3, 1:2000, Abcam, catalog number ab48394, Cambridge, UK), rabbit polyclonal anti-p62 (1:50000, Cell Signaling technology, catalog number 5114S, Danvers, MA, USA), rabbit monoclonal anti-p-MEK (1:5000, Abcam, catalog number ab96379, Cambridge, UK), rabbit polyclonal anti-p-ERK1/2 (1:1000, Wanleibio, catalog number WLP1512, Shenyang, China), rabbit monoclonal anti-MKK7 (1:1000, Abcam, catalog number ab239843, Cambridge, UK), rabbit monoclonal anti-JNK (1:10000, Abcam, catalog number ab124956, Cambridge, UK), rabbit monoclonal anti-mTORC1 (mammalian target of rapamycin complex-1, 1:5000, Abcam, catalog number ab92477, Cambridge, UK), rabbit monoclonal anti-N-cadherin (1:20000, Abcam, catalog number ab76011, Cambridge, UK), rabbit monoclonal anti-NCAM (neural cell adhesion molecules, 1:1000, Abcam, catalog number ab220360, Cambridge, UK), rabbit polyclonal anti-MBP (myelin basic protein, 1:6000, Proteintech, catalog number 10458-1-AP, Rosemont, USA), rabbit polyclonal anti-Periaxin (1:2000, Boster, catalog number A01686, Pleasanton, USA) and mouse monoclonal anti-GAPDH (Proteintech, 1:1000). After three washes in PBST buffer, membranes were incubated at 37 °C for 1 h with the secondary antibodies horseradish peroxidase (HRP)-labeled goat anti-rabbit IgG (H+L) (1:1000, catalog number A0208, Beyotime Biotechnology, Shanghai, China,) and HRP-labeled goat anti-mouse IgG (H+L) antibodies (1:1000, catalog number A0216, Beyotime Biotechnology, Shanghai, China) and then washed three times in PBST. The special bands were visualized using an electrochemiluminescence (ECL) method (Millipore, Bedford, MA, USA). Tanon image analysis software (Tanon Science & Technology, Shanghai, China) was used to conduct grayscale analysis for protein expression. Experiments were carried out in triplicate.
ELISA analysis
Total proteins from different groups were extracted by RIPA lysis buffer, and the protein content was quantified by the BCA method as previously described. Protein concentrations of IL-1α (interleukin-1α), IL-1β (interleukin-1β), LIF (leukaemia inhibitory factor), TNFα (tumor necrosis factor-alpha), BDNF (brain-derived neurotrophic factor), GDNF (glial cell line-derived neurotrophic factor), Olig1 (oligodendrocyte transcription factor 1) and VEGF (vascular endothelial growth factor) were detected using an ELISA kit (Huyu Biological Technology, Shanghai, China). Since the concentration of target protein is directly proportional to the absorbance at 450 nm, the concentration of target protein can be calculated by measuring the absorbance value at 450 nm by ELISA. The detailed protocol of ELISA is presented in the Support Information.
Immunohistochemistry (IHC) analysis
Specimens were fixed with 4% PFA for 5 h and embedded in paraffin. Prior to immunohistochemistry, nerve sections were dewaxed and rehydrated in PBS (pH 7.4). Then, the nerve sections were incubated with 0.6% hydrogen peroxide for 30 min. To block nonspecific immunoreactions, the sections were incubated with 10% normal goat serum (1:50, DAKO, USA). Subsequently, sections were incubated with primary antibodies overnight at 4 °C. According to the parameters to be tested, the following primary antibodies were chosen: rabbit polyclonal anti-BDNF (1:400, Boster Biological Technology Co., Ltd, catalog number PB9075, California, USA), rabbit polyclonal anti-GDNF (1:400, Boster Biological Technology Co., Ltd, catalog number PA1465, California, USA), rabbit polyclonal anti-Olig1 (1:400, Beijing Biosynthesis Biotechnology Co., Ltd. catalog number bs-8548R, Beijing, China), rabbit polyclonal anti-VEGF (1:500, Proteintech Group, Inc, catalog number 19003-1-AP, Rosemont, USA), rabbit polyclonal anti-MBP (1:500, Proteintech Group, Inc, catalog number 10458-1-AP, Rosemont, USA) and rabbit polyclonal anti-periaxin (PRX) antibodies (1:400, Boster Biological Technology Co., Ltd, catalog number A01686, California, USA). They were washed three times with PBS and incubated in biotinylated goat anti-rabbit IgG solution for 1 h at 37 °C. HRP-labeled secondary antibodies were applied for 1 h. Then, all sections were incubated with 3,3'-diaminobenzidine tetrahydrochloride chromogenic substrate solution (DAB, DAKO, USA) for 10 min. The immunohistochemistry results were assessed under an optical microscope.
Statistical analysis
The normality of all data distributions was checked with the Kolmogorov–Smirnov test. Continuous variables presenting a normal distribution, such as quantitative data of protein detected by WB and ELISA, immunohistochemical protein quantification data, qRT-PCR relative gene expression data, amount of iron in the tissues, SFI, CMAPs, and serum biochemical parameters, were represented as the mean ± standard deviation (SD) and compared between groups by one-way analysis of variance (one-way ANOVA). When the presence of significant changes was observed, post hoc multiple pairwise comparisons were carried out using the Student-Neuman-Keuls (SNK) test.
The distributions of the G-ratio of myelinated nerve fibers and the MNCV were found to be nonnormal by the Kolmogorov–Smirnov test; therefore, their values were represented by the median and interquartile range (Q1–Q3). Due to violation of the normality assumption, the G-ratio and MNCV results were analyzed by the nonparametric Friedman test, followed by pairwise comparisons using the Wilcoxon signed-rank test.
All of the analyses were conducted with SPSS (version 18.0, Chicago, IL, USA), and P < 0.05 was considered to be statistically significant.