Differential Effects of Encapsulation of Rat ADSCs in Fibrin Matrix and Combination Delivery of BDNF and Gold Nanoparticles on Peripheral Nerve Regeneration

Background: Fibrin as an extracellular matrix feature like biocompatibility, creates a favorable environment for proliferation and migration of cells and acts as a reservoir for storage and release of growth factors in tissue engineering. Methods: In this study, the inner surface electrospun poly (lactic-co-glycolic acid) (PLGA) nanobrous conduit was biofunctionalized with laminin containing brain derived neurotrophic factor (BDNF) and gold nanoparticles in chitosan nanoparticle. The rats were randomly divided into ve groups, including autograft group as the positive control, PLGA conduit coated by laminin and lled with DMEM/F12, PLGA conduit coated by laminin and lled with adipose-derived stem cells (rADSCs) , PLGA conduit coated by laminin containing gold-chitosan nanoparticles (AuNPs-CNPs), BDNF-chitosan nanoparticles (BDNF-CNPs) and lled with rADSCs or lled with rADSCs suspended in brin matrix, and they were implanted to bridge a 10 mm rat sciatic nerve gap. Eventually, axonal regeneration and functional recovery were assessed after 12 weeks. Results: After 3months post-surgery period, the results showed that in the PLGA conduit lled with rADSCs without brin matrix group, positive effects were obtained as compared to other implanted groups by increasing the sciatic functional index signicantly (p < 0.05). In addition, the diameter nerve bers had a signicant difference mean in the PLGA conduit coated by laminin and conduit lled with rADSCs in brin matrix groups relative to the autograft group (p< 0.001). However, G-ratio and amplitude (AMP) results showed that brin matrix might have benecial effects on nerve regeneration but, immunohistochemistry and real-time PCR outcomes indicated that the implanted conduit which lled with rADSCs with or without BDNF-CNPs and AuNPs-CNPs had signicantly higher expression of S100,


Page 3/19
Fibrin matrix has been used as a ller within the lumen of conduits with biodegradability, biocompatibility, non-reactiveness, lower toxicity and stability properties which may impact the tissue repair in a positive pathway [8]. A study reported that application of autologous brin glue is valuable e cacy in repairing of sciatic nerves injury in rabbits [9]. It has been shown that brin containing Schwann cells (SCs) has been used as a luminal ller and can increase axonal repair [10]. Additionally, neurotrophins along with SCs suspended in brin matrix have been used for rat sciatic nerve regeneration [11]. Moreover, brin matrix as a natural polymer may be appropriate for this purpose via retaining the cells in the damage zone, providing interaction of cells with extracellular matrix and physical connection to the nerve end [12].
Adipose-derived stem cells (ADSCs) are the promising candidate to be applied as cellular components in a nerve conduit [13]. Therefore, we used rat adipose-derived stem cells (rADSCs) to promote nerve repair, owing to the features of secreting growth factors, myelin formation, and improving nerve repair [14,15].
It has been shown that factors like brain derived neurotropic factor (BDNF) contribute to formation of neurites, increase axonal outgrowth, prevent apoptosis, and form chemotactic factors guiding regenerating axon after nerve damage [16,17].
In addition, one previous study showed that gold nanoparticles (AuNPs) could improve axonal growth and myelination, as well as prevent neuronal death [18].
Recently, it has been proven that some factors, such as AuNPs can be added on immature neuronal cell line and stimulate cell adhesion, proliferation, differentiation, stimulate axonal elongation, and sprouting axons [19].
To optimize the transfer of rADSCs in nerve conduit, a matrix is need to ll the nerve conduit, leading to homogenous distribution of cells over the conduit [20]. This matrix must have some properties such as preserving the cells, not disturbing the axonal growing pathway into the lumen of the conduit, being consistent with cell proliferation, and remaining alive [21,22].
In our previous study, we found that coating of BDNF and AuNPs encapsulated in chitosan on nano bers lead to their continuous release for 7 days and this could enhance proliferation and differentiation of adipose derived stem cells into Schwann cells in vitro. [23,24]. Therefore, we decided to fabricate PLGA nano brous scaffold with random outer surface and aligned with inner surface orientation by electrospinning. The electrospun nano ber lms were made into nerve conduits after being coated with laminin containing BDNF and AuNPs encapsulated in chitosan nanoparticles of the inner surfaces of PLGA nerve conduits. Then, rADSCs were loaded in brin gel and injected into PLGA conduit implanted into transected sciatic nerve rat model. Then, we assessed effects of encapsulation cells in brin matric and combination of gold-chitosan nanoparticles (AuNPs-CNPs) and BDNF-chitosan nanoparticles (BDNF-CNPs) factors on sciatic nerve regeneration at 12-weeks post-surgery. Experiments for all tests were repeated three times.
Characterization of PLGA nerve conduit and injected rADSCs The structure and morphology of fabricated scaffolds were observed using a scanning electron microscope (SEM). The inner surface of PLGA that directly contacted the growing nerve was electrospun in an aligned form, while the outer surface had randomly oriented nano bers (Fig. 1A). Furthermore, the SEM image showed that the overall thickness of the PLGA scaffold was approximately ~74 µm (using an Image J analysis software). While, the mean thickness of laminin-coated sheets with BDNF/AuNPs encapsulated chitosan nanoparticles was measured ~78 µm. Oure results indicated that the fabricated conduits had an internal diameter of 2 mm with a length of ~1.4 cm (Fig. 1Ba). After 3 passage primary cultured rADSCs were exhibited irregular, at, and spindle-shape broblast-like morphology under light microscopy (Fig. 1Bb).
In the PLGBC group, the presence of three factors such as AUNPs, BDNF and rADSCs with the conduit exhibited more proper appearance of nerve formation (Fig.1C), whereas the injection of cells with brin (PLGBCF) created an incomplete repair morphology in the nerve (Fig. 1D).

Sciatic function index outcomes
Behavior analysis such as sciatic functional index and pinprick test was performed to assess the sensory and motor recovery of the sciatic nerve in rats, respectively. Motor recovery of the sciatic nerve was evaluated by the sciatic functional index after the 1st, 2 nd , 4 th , 8 th and 12 th weeks after surgery ( Fig. 2A). All experimental groups improved over time. The autograft group showed a signi cantly higher SFI value (-15.44 ± 0.96) than that of the other implanted nerve conduit groups at the 4 th , 8 th and 12 th weeks after surgery (p < 0.001). However, three months after the transplantation, repair with PLGBCF had a lower signi cant SFI value (-74.44 ± 2.69) than that of all experimental groups (p < 0.001). At the 12th week after surgery, the PLGBC group showed a signi cantly increase mean SFI value (-31.3 ± 2.34) than that of the other implanted conduit groups (p < 0.05).

Pinprick test outcomes
To assess of sensory recovery, the pinprick test was performed. The results of pinprick test show that there was no signi cant difference mean pinprick score between autograft group (3 ± 0.0) and PL (2.5 ± 0.2), PLC (2.6 ± 0.2) and PLGBC (2.8 ± 0.1) groups at 12 weeks' post-surgery. However, there was a signi cant difference mean of pinprick score in the autograft group (3 ± 0.0) compared with PLGBCF (1.8 ± 0.4) group (p <0.01). All autograft animals received score 3 at 12 weeks, regardless of type of nerve repair (Fig. 2B).

Outcomes of electrophysiological assessment
Twelve weeks following surgery, the CMAP latency and CMAP amplitude were measured on the implanted side in experimental grafted groups by electrophysiological assessment (Fig. 2C-D). Our results revealed that the mean CMAP latency in the PLGBCF group was signi cantly higher than that of the other implanted and autograft groups (p <0.05), but no statistically signi cant differences were found between the autograft group and the other treated groups except PLGBCF (Fig. 2C).
The CMAP amplitude ratio showed no signi cant difference between autograft and conduit implanted groups. In fact, these CMAP amplitude values were higher in the autograft group (6.60 ± 1.8), then the PL (5.04 ±1.73), the PLC (5.78 ± 1.5) the PLGBC (5.64 ± 0.5) and the PLGBCF (3.21 ± 0.6) groups. However, these differences in the mean of CMAP amplitude values were not signi cant between the experimental groups ( Fig. 2D).

Histological analyses of gastrocnemius muscle
To assess atrophy in the gastrocnemius muscle, owing to sciatic nerve transection, Masson's trichrome staining was performed for muscle in ve groups (Fig. 3A). Denervation of target muscles led to decrease muscle ber diameters. Compared to muscle morphology in the autograft group, the mean muscle bers diameter suffering from denervation was degenerated in the PLGBCF group (11.06 ± 0.9) and showed a signi cant smaller muscle cell diameter (p < 0.001). However, the mean muscle bers diameter of PL (13.02 ± 0.3), PLC (14.41 ± 0.6) and PLGBCF groups had a signi cant decrease compared to that in the autograft group (19 ± 0.4) (p < 0.001), but the mean of muscle ber diameter in PLGBC group (16.5 ± 0.7) was not signi cant different with the autograft group (19 ± 0.4) (p 0.05) (Fig. 3B).
Twelve weeks after implantation, the mean wet weight of gastrocnemius muscle in the autograft group (19 ± 0.4) was signi cantly decreased than in the conduit-implanted groups (p < 0.001). In addition, the mean muscle mass in the PLGBCF group (11.06 ± 0.9) was signi cantly smaller than that in the autograft and other experimental groups (p < 0.001). The mean weight of the gastrocnemius muscle in the PLGBC group (16.5 ± 0.7) was signi cant higher than that in PL and PLGBCF groups (p < 0.001). Furthermore, the atrophy in PLGBC was signi cantly less than that in the PLC group (p < 0.01) (Fig. 3C).
Nerve histomorphometry Figure 4A depicts the results of toluidine blue staining for the cross-section of regenerated nerves. Numerous bers of regenerated nerves can be found in the autograft group. We observe that the mean nerve bers diameter had no signi cant difference between PLC (6.08 ± 0.3) and PLGBC (6.2 ± 0.5) groups and the autograft group (7.2 ± 0.5) (p 0.05), but this difference mean was signi cant between PL and PLGBCF ( 5.1 ± 0.2) groups and the autograft (7.2 ± 0.5) group (p < 0.001) (Fig. 4B). Furthermore, all conduit implanted groups showed that the mean diameter of myelinated axons was signi cantly smaller than that in autograft (3.5 ± 0.2) groups; this difference was more evident in PL (1.9 ± 0.1) and PLGBCF (2.5 ± 0.1) groups (p < 0.001) than PLC (2.7 ± 0.2) and PLGBC (2.9 ± 0.3) groups (p < 0.05) with autograft groups (Fig. 4C). Additionally, quantitative analysis shows that the mean thickness of myelin sheet in PL (2.7 ± 0.2) and PLGBC (3.3 ± 0.2) groups was not signi cantly different from that of the autograft (3.6 ± 0.2) group (p > 0.05), but there was a signi cant different mean in this analysis between PL (p < 0.05) and PLGBCF (2.5 ± 0.2) (p < 0.01) groups and autograft group (Fig. 4D). Finally, there was no signi cant different mean in the G-ratio between all conduit-implanted groups and autograft groups (Fig. 4E).

Immunohistochemical analysis of regenerated nerves
At 12th week, the presence of rADSCs throughout the gap area is clearly apparent in longitudinal sections, suggesting that they contributed to the process of tissue regeneration in vivo. The micrographs depict the immunohistochemical cross sections of the middle region of the implanted laminin coated PLGA conduit loaded with rADSCs in different experimental groups. Then, to evaluate the immunopositivity expression intensity for S100, MBP and NF200 cells in autograft and all implanted conduits, three immunohistochemical markers, S100, MBP and NF200, were applied to all the samples (Figs. 5& 6). To visualize axon bers, neuro laments were labeled with NF-200 exhibiting a red color. Schwann cells labeled with S100 and cells labeled with MBP appeared as green. Finally, cells nuclei were labeled with DAPI exhibiting a blue color. Microscopic elds were randomly selected from each slice for measuring the immunohistochemical intensity value of positive cells.
The comparison of the mean percentage of intensity value S100, MBP and NF200 intensity in experimental groups relative to autograft group have shown that in Table 2.
We found that the mean intensity of S100 positive cells was signi cantly higher in the PLGBC group (106.25 ± 0.63) than in PL (89.18 ± 0.43), PLGBCF (48.07 ± 0.58) groups and the autograft group (p < 0.001) 12 weeks after nerve after grafting. While, the intensity of S100 positive cells in the PLGBCF group is less than that in the other treated groups (48.07 ± 0.58).
The statistics of MBP-positive intensity indicated a signi cant difference between PL (78.6 ± 0.88) and PLGBCF (45.05 ± 0.52) groups and other experimental groups (p < 0.001), whereas this intensity in the PLC group (96.47 ± 0.48) was signi cantly different from that in the autograft group (p < 0.01). Furthermore, in PLGBC groups, it was not signi cantly different from that of the autograft group (p > 0.05) ( Table 2). While the mean percentage of intensity of NF200 positive cells in PLGBC was higher than that in the other experimental groups (125.17 ± 1.16), and this difference was signi cant (p < 0.001). Meanwhile, a signi cant difference mean of the NF200 intensity was obtained in PL, PLC and PLGBCF groups compared with the autograft groups (p < 0.05) ( Table 2).

Discussion
In the present study, the PLGA scaffold with aligned inner surface and random outer surface orientated bers were fabricated by the electrospinning method. The inner surfaces of PLGA nerve conduit coated with laminin containing BDNF and AuNPs encapsulated in chitosan nanoparticles were rolled up to form a tubular nerve conduit. Then, the brin gel was loaded with rADSCs and injected into the lumen of the PLGA conduit implanted into the 10 mm transected sciatic nerve rat model. The nerve repair status with or without of rADSCs, brin matrix and BDNF/AuNPs was investigated by different methods in a 12-week post-surgery period.
In the previous study, we demonstrated that the releasing rate of BDNF was 83.28 ± 2.22 (%) of chitosan/TPP particles during the 7 days [24]. Also, a sustained release of BDNF and AuNPs from PLGA scaffolds functionalized with coated laminin containing chitosan nanoparticles encapsulating either AuNPs or BDNF was detected over a period of 7 days. The release rate of BDNF and AuNPs from nano bers was 74 ± 2.42% and 47.24 ± 1.78%, respectively. In addition, our results showed that biofunctionalized scaffold enhanced proliferation, differentiation and myelinization of h-ADSCs into SClike cells [23].
Previously researches have demonstrated that, BDNF may be effect on ADSCs differentiation into Schwann-like cells phenotype during 7 days and growth factors released by differentiated SCs help to neurite outgrowth [25]. Therefore, our hypothesis was that it might be effective in peripheral nerve regeneration.
Strategies of using the PLGA conduit containing with ADSCs were useful to regenerate transected nerve.
It is shown that, the ADSCs results increase growth factor secretion, axonal myelination and nally nerve repair [14,15]. Furthermore, exogenous factors such as AuNPs and BDNF were used for more recovery in nerve repairing [16,18].
Fibrin matrix is one of these hydrogels, which has been extensively used for nerve reparation, and it has been shown that the nerve conduits fabricated by brin affect axonal growth and myelination [26,27].
According to one study, autologous brin glue in combination with several growth factors might effectively increase peripheral nerve regeneration in 15 mm rabbit peroneal nerve defect [9].
In addition, autologous brin glues have some advantages, including no risk of infection or hypersensitivity. Since brin glues are usually prepared from whole blood therefore, their use leads to increased risk of transmission of the pathogen from the donor to the recipient [28]. However, disadvantages of available commercial brin glues include viral transmission risk, expensiveness, reduced bioactivity and growth factors, which are not present in commercial products [29,30].
Ra jah et al. have demonstrated that the use of collagenous conduits lled with brin glue in 10 mm rat sciatic nerve gap promotes axonal repair and functional recovery 12 weeks after surgery [31]. Therefore, according to the previous studies, we purposed that laminin coated PLGA nerve conduit containing BDNF and gold nanoparticle was loaded with rADSCs in brin matrix, and can improve peripheral nerve regeneration.
We observed that the PLGA conduit was degraded completely 12 weeks after implantation.
The evaluation of electrophysiological in all implanted groups, as indicator of conduction function of peripheral nerves, was performed for detectable CAMP latency and amplitude analysis. The latency of CMAP represents the thickness of the myelin sheath, whereas the CMAP amplitude shows the number of nerve bers [32]. The comparison in the CMAP latency between all groups showed that the speed of transport in action potential in PLGBCF group was signi cantly lower than that of the other groups (p < 0.05), but the comparison in the CMAP amplitude between four conduit implanted groups provided further evidence that their functional recovery was more close to that in the autograft group. The lack of amplitude differences mean indicates that the number of regenerated nerve bers in all groups is approximately similar.
Beside in, we found that motor and sensory nerve recovery by the SFI and pinprick tests, consistent with electrophysiology outcomes.
These documents indicate that the presence of rADSCs with BDNF and AuNPs can lead to nerve repair, but the loaded of brin along with these factors may have non synergistic effects. Our results are inconsistent with the data obtained by Masgutov et al. indicating that adipose-derived mesenchymal stem cells loaded in brin glue promote peripheral nerve regeneration in sciatic nerve injury [12]. In our study, a combination of factors was used that each of them alone increased nerve regeneration, but we have shown that they may have an inhibitory impact on each other's functions and reduce the effect of each other.
Histological and wet weight analyses in gastrocnemius muscle evaluations indicated that PLGBC, PLC, PL and PLGBCF achieved more reconstruction, respectively. However, the results of these methods between conduit-implanted groups were not comparable to those in the autograft group (p < 0.001). Nevertheless, the mean diameter muscle ber in the PLGBC group was not signi cant in the autograft group (p > 0.05). These results are in line with previous researches. McGrath et al. showed that gastrocnemius muscle weight was decreased in the conduit lled with brin matrix, but treatment with cyclosporine A or cyclosporine A with human mesenchymal stem cells induced recovery of the muscle weight [33].
In the present study, the nerve histological data showed that the mean thickness of myelin sheath and diameter of nerve bers were not signi cantly different mean in PLGBC and PLC groups, compared to in the autograft group (p 0.05). However, the mean diameter of myelinated axons in PLGBC (2.97 ± 0.34) and PLC (2.73 ± 0.22) was higher among the conduit implanted groups but comparisons have different signi cant with autograft group (p < 0.05).
In this study, application of brin matrix in conduit hasn't promoted effect in nerve regeneration and despite poor results that indicates the brin group is limited nerve regenerated, but, the non-signi cant difference between all groups in the parameters of G-ratio and amplitude (AMP) was observed. It maybe indicates brin can repair sciatic nerve in terms of number of axons and the thickness of the myelin in received conduit.
The mean number of myelinated axons in PL and brin groups was signi cantly lower than that in the autograft group, respectively (p < 0.001). However, there were no differences between PLGBC and PLC groups. As a result, in the groups containing AuNPs and BDNF, there were a higher number of axons, suggesting that the encapsulated nanoparticles and rADSCs can differentiate between SCs and lead to repair of the damaged nerve.
However, increases in toxicity levels for the use of intraluminal llers such as brin must be kept in mind. A study that supports this case was supported by Wood et al. This study used a brin matrix which degraded after a four-week period and despite its approximately short residence time showed bene cial effects on nerve repair [34]. This further complements the idea that intraluminal llers like brin are bene cial, but only at the early stages of regrowth, after which it can be hypothesized that they become an inhibitory material.
The presence of rADSCs in the PLGA conduit in combination with BDNF/AuNPs without the presence of brin may lead to their differentiation into SCs and, as a result, high expression of S100 in immunohistochemical and real-time PCR methods.
The results of the expression of MBP and NF200 were con rmed by the rate of S100 expression.
Therefore, myelinated axons (MBP positive cells) and regenerating axons (NF200 positive cells) in conduit-implanted groups containing cell and growth factor were higher than those in the other conduit implanted groups, and it seems that in the PLGBCF group, brin interfered with other factors, thereby preventing to improve nerve regeneration.

Conclusion
Overall, we found that laminin-coated PLGA conduit could improve axonal regeneration in peripheral nerve using some exogenous factors such as BDNF, AuNPs, and ADSCs in brin matrix. Although there was some evidence that brin matrix promoted peripheral nerve regeneration, but distinctive effect of brin was not observed in this study. It should be noted that brin matrix may interfere with other exogenous factors. It is recommended that future works determine the molecular mechanism of brin matrix in peripheral nerve regeneration and its clinical outcomes following injury.

Electrospinning and characterization of PLGA nano bers
All of materials which used in this study were prepared from Sigma-Aldrich, St. Louis, MO, USA except it was mentioned in the text. Electrospun scaffolds with mild modi cation were fabricated as described in our previous study [35]. PLGA (80:20, Mw: 50,000-75,000) was dissolved in chloroform: DMF (Merck, Darmstadt, Germany) mixture with a volume ratio of 80: 20 to obtain a concentration of 20% (w/v). High voltage electric eld of 21 kV was used to draw the polymer solution fed at a rate of 250 µL/h into nano bers over a distance of 15 cm from the needle tip to an aluminum-wrapped rotating drum. The drum rotation speed was gently decreased from 2500 to 300 rpm. Therefore, a highly aligned nano brous inner surface was formed, and as the drum rotation speed was decreased, ber orientation gradually diminished. Electrospinning process continued for 9 h, and randomly oriented bers of outer surface endowed the scaffold with isotropic mechanical properties, while the gradual alteration in ber orientation could avoid sheet delamination. Scanning electron microscopy (SEM) (SEM, Seron Technology AIS 2500, India) was used to assessed the morphology of PLGA. Therefore, electrospun PLGA samples were coated with a thin layer of gold prior to taking SEM images.
PLGA sheets coating with laminin containing BDNF and AuNPs encapsulated chitosan nanoparticles AuNP or BDNF encapsulated chitosan nanoparticles were produced by the ionotropic gelation method . After emulsi cation of 50 ppm AuNPs (US Research Nanomaterials, Inc, Houston, TX, USA) [18] or 5 µg/ ml BDNF (R&D systems, Minneapolis, MN, USA) with 0.1 % chitosan solution, the solution was mixed on a magnetic stirrer for 15 min. Then, 0.03% TPP of an aqueous solution as a cross linker was added dropwise into the previous solution.The 20 µg/ ml of laminin was used for supply suspention of chitosan nanospheres containing BDNF or AuNPs, and nally was coated on the PLGA scaffold at 4°C temperature for 24 h [24] .
Design and fabrication of nerveconduit PLGA sheet was cut to the size (14×20 mm) and exposed to UV light for sterilization. The membrane was coated with chitosan nanospheres encapsulated BDNF or AuNPs mixed by 20 µg/ ml laminin. Next, the electrospun mat rolling 2.5 rounds around a mandrel to create a tubular structure. The edge of conduit was sealed with cyanoacrylate glue. All these procedures were performed under sterile conditions.

Rat-ADSCs isolation, expansion
All of the experimental procedures involving animals were conducted in accordance with the guidelines given by National Under sterile conditions, the isolated tissue was placed on ice to the stem cell laboratory. Brie y, the adipose tissue surrounding the inguinal region was cut into tiny segments and treated with 0.075% collagenase type I and shaken at 37°C for 35 to 40 min. The resultant to neutralizing enzyme activity was added to the DMEM (Gibco Grand Island, NY, USA) containing 10% fetal bovine serum (Gibco Grand Island, NY, USA) into each tube. After centrifugation (1200 rpm for 5 min), the upper fat tissue layer and supernatant were discarded. The cell pellet was transferred into culture asks containing DMEM + 10% FBS and transferred into an incubator at 5% CO2 and 37°C and saturated humidity. After 3-4 days, the medium was refreshed, and the cells were sub-cultured until passage 4. In this study, h-ADSCs obtained from passage 3-4 were used for the experiments.
Operation procedure and experimental groups In this study, 40 mature (12-week-old) Wistar male rats (weighing ~200-250 g) purchased from Pasteur Institute, Tehran, Iran and were housed under standard conditions (at 18-24°C, 12 h light/dark) with free access to laboratory pellet chow and water. Acclimatization of animals began 10 days before the experiment date. Each rat was kept in one cage during experimental period.
All of the experimental procedures involving animals were conducted in accordance with the guidelines given by International Council for Laboratory Animal Science (ICLAS) for the Care and Use of Laboratory (PL), PLGA conduit coated by laminin and lled with 2×10 6 rADSCs according to previous study [36] (PLC), PLGA conduit coated by laminin containing BDNF-CNPs, AuNPs-CNPs and lled with 2×10 6 rADSCs (PLGBC) and PLGA conduit coated by laminin containing BDNF-CNPs, AuNPs-CNPs and lled with 2×10 6 rADSCs suspended in brin matrix (PLGBCF) transplanted in 10 mm sciatic gap.
Each rat was anesthetized with intraperitoneal injection of Xylazine 10mg/kg and Ketamine 100mg/kg. The thigh areas on left sides were shaved and disinfected with povidone-iodine.
Then, the sciatic nerve was exposed by skin and muscle-splitting incision. Then, the muscle and skin were closed with 7-0 nylon sutures. Finally, the rats had free access to water and food, and housed in controlled animal rooms under arti cial 12-hour light/dark cycles.

Sciatic function index
Functional recovery of the sciatic nerve was evaluated by the sciatic functional index (SFI) obtained from rat footprints on the white paper covered on the box oor. Three obvious footprints were selected from each rat, and nally three different parameters were measured: (i) heel to the third toe (print length, PL), E is the experimental sides, and N is the normal sides. Generally, a value of close to 0 indicates the normal function and a value of close to -100 indicates the higher impairment [37].

Pinprick test
To determine of sensory recovery, the pinprick test was performed. The experimental limb of rats was pinched with standardized forceps from the toe to the heel. In the pinprick test limb withdraw response to painful stimulus was graded from 0 to 3. Pinprick test was evaluated from heel to toes and measured values include: 0 = no response, 1 = heel, 2 = dorsum of foot (mid-foot), and 3 = toes [38] .

Electrophysiology
For each animal, electrophysiological testing was performed in a 12-week post-surgery period. After anesthesia , the electrophysiologic method was performed by placing stimulation electrode in the proximal end of the regenerated nerve and recording electrodes in the mid of gastrocnemius muscle. The compound muscle action potential (CMAP) has been used to estimate the numbers of repaired motor nerve bers. The CMAP parameters were analyzed for each rat; for instance, the CMAP latency parameter was assessed from the stimulation site to the start of the response and measured in milliseconds.
In addition, the amplitude (AMP) of CMAPs was calculated by the potential difference between maximum negative and positive peaks of the CMAP signal in millivolts [39]. Electrophysiological analysis was calculated using the eProbe software. Finally, the CMAP parameters were compared among different groups.

Muscle mass
After sacri cing of animals (with 100 mg/kg sodium pentobarbital i.p.), gastrocnemius muscles were removed from normal and injured sides and weighed while still wet by an electronic balance (A&D Weighing EK-3000I Portable Balance, 3000g Capa city). The recovery rate and muscle atrophy were calculated by the wet weight of gastrocnemius muscle on the experimental side/the wet weight of gastrocnemius muscle on the normal contralateral side × 100% [40].

Histological study
Operated nerve and gastrocnemius muscle of both limbs were dissected from the surrounding tissues 12 weeks after conduit implantation. The middle segment of muscles and nerve were xed in 10% neutral buffered formalin for 48 hours, and after tissue processing, para n blocks were cut into 5 μm thick sections. The transvers sections of gastrocnemius muscle were stained with Masson's trichrome and examined under a light microscope. The mean diameter of the muscle bers was calculated using the Digimizer software by an image analysis system. Nerve samples were stained with toluidine blue to label the myelin sheath. Brie y, the xed nerves were dehydrated in a graded ethanol series and embedded in para n. The specimens were cut into 5 μm thick cross sections with a microtome, stained with 0.125% toluidine blue solution, observed and evaluated by under a light microscope [41].
The G-ratio (the axon-to-ber diameter ratio) is an important parameter to estimate degree of myelination [42].To calculate the mean number of myelinated axons, nine images were randomly captured at 400× magni cation for each group, and 5 elds of all images were counted to obtain the mean numbers of myelinated nerve bers. The average diameter of the sciatic regenerated nerve bers, axon, myelin and the mean number of myelinated axons were calculated using the Digimizer software by an image analysis system.

Immunohistochemistry staining
In this study, S100 β as a speci c marker for Schwann cells, myelin basic protein (MBP) as a myelin sheath marker, and neuro laments-200 (NF200) as a growing axon marker were used. Nerve specimens were xed with 4% paraformaldehyde for 6 h and embedded in para n and then cut into 3-μm thick sections. For double immunostaining analysis of NF200 and MBP or NF200 and S100 in the sections, antigen unmasking was performed by 10 mM sodium citrate buffer, pH= 6.0, at 90 °C for 10 min.
Finally, all sections were incubated with DAPI for cell nuclei staining, and immunohistochemistry results were examined under a uorescent microscope (Olympus BX51, Japan). Then, the total area of the images was measured for the intensity of NF200, S100 and MBP staining using the Image-j software [43].

Real-time reverse transcription polymerase chain reaction
The level of S100, MBP, GFAP and Nestin expression in the rat sciatic nerve tissue in were assessed using real time RT-PCR. Total RNA was isolated using the Total RNA Prep Kit (BIOFACT). RNA was reverse transcribed using the BioFact™ 5X RT Pre-Mix cDNA Synthesis Kit (BIOFACT) according to the manufacturer's protocol. The real-time PCR was performed using BioFact™ 2X Real-Time PCR Master Mix Kit (BIOFACT) and the StepOne Plus™ quantitative Real Time PCR Detection System (Applied Biosystems). The level of β-actin was used as the control housekeeping gene. Table 1 lists the sequence of the used primers (metabion, Germany) [23].

Statistical analysis
Statistical differences between the different groups were tested by the one-way ANOVA followed by posthoc, tukey's test using SPSS software version 22.0 (SPSS Inc., Chicago, IL, USA). Data were presented as mean ± standard error and P-values less than 0.05 were considered statistically signi cant. Table 1. The list of primer sequences of S100; Schwann cell marker, Nestin; Marker of neuronal progenitor cells, Gfap; Glial brillary acidic protein, Mbp; Myelin basic protein, β-actin; As the control housekeeping gene for Real time RT-PCR analysis.