Adipose tissue stem cells in peripheral nerve regeneration—In vitro and in vivo

After peripheral nerve injury, Schwann cells (SCs) are crucially involved in several steps of the subsequent regenerative processes, such as the Wallerian degeneration. They promote lysis and phagocytosis of myelin, secrete numbers of neurotrophic factors and cytokines, and recruit macrophages for a biological debridement. However, nerve injuries with a defect size of >1 cm do not show proper tissue regeneration and require a surgical nerve gap reconstruction. To find a sufficient alternative to the current gold standard—the autologous nerve transplant—several cell‐based therapies have been developed and were experimentally investigated. One approach aims on the use of adipose tissue stem cells (ASCs). These are multipotent mesenchymal stromal cells that can differentiate into multiple phenotypes along the mesodermal lineage, such as osteoblasts, chondrocytes, and myocytes. Furthermore, ASCs also possess neurotrophic features, that is, they secrete neurotrophic factors like the nerve growth factor, brain‐derived neurotrophic factor, neurotrophin‐3, ciliary neurotrophic factor, glial cell‐derived neurotrophic factor, and artemin. They can also differentiate into the so‐called Schwann cell‐like cells (SCLCs). These cells share features with naturally occurring SCs, as they also promote nerve regeneration in the periphery. This review gives a comprehensive overview of the use of ASCs in peripheral nerve regeneration and peripheral nerve tissue engineering both in vitro and in vivo. While the sustainability of differentiation of ASCs to SCLCs in vivo is still questionable, ASCs used with different nerve conduits, such as hydrogels or silk fibers, have been shown to promote nerve regeneration.


| INTRODUC TI ON
Peripheral nerve injuries often affect young patients, leading to high costs in the health-care system. They have various causes.
Besides accident-related short-or long-distance transection injuries (e.g., knife cut, saw wound) and crush injuries, also birth traumata, ischemia, or malignant processes can lead to nerve lesions (Bahm et al., 2009;Farley et al., 2019;Gilbert & Whitaker, 1991;Myers et al., 1993;Reinhardt et al., 2010). Depending on which anatomical structure of the peripheral nerve is lesioned, surgical therapy for peripheral nerve regeneration has to be performed to regain nerve function.
In the case of nerve lesions, up to a gap of approx. 1 cm, the proximal and distal nerve ends can usually be surgically reconnected without tension to restore sensibility and motor functions in the affected target area. This is called coaptation (Figure 1). It has to be noted that most experimental investigations on nerve regeneration were performed on defect sizes that do not exceed 10 mm.
As explained, most nerve defects up to 10 mm can be coapted in the clinical scenario, meaning no need for nerve reconstruction. Gap sizes larger than 10 mm have to be assumed as critical, as peripheral nerve reconstruction is needed. Tension after coaptation leads to poor results, as Schwann-cell activation and axon regeneration are impaired, meaning decreases in fiber count and nerve density (Mackinnon, 1989;Millesi, 1984;Sunderland et al., 2004). The nerve regeneration rate is about 1-3 mm per day, depending on the location along the axon (proximal lesion sites tend to regenerate faster than at distal locations; Menorca et al., 2013). When distances are too vast, functional recovery is usually incomplete as prolonged denervation may induce damage to the target areas (Pfister et al., 2011). If no surgical therapy is applied after neurotmesis, neuroma develops, and spontaneous healing is doubtful (Figure 2).
The clinical gold standard for bridging a nerve gap that cannot be coapted without tension is the autologous nerve transplantation (ANT). It was first performed in 1870 by Phillipeaux and Vulpian. In a canine model, they bridged a 2 cm resection of the N. hypoglossus with an autologous nerve transplant from the N. lingualis. The first results were poor, as in only two of seven animals, nerve regeneration could be observed (Dellon & Dellon, 1993).
The advantages of ANT are the availability of several donor nerves in a patient and the nonimmunogenicity. Even for defect sizes of 10-20 cm, the ANT has shown excellent results in restoring motor and sensory function, for example, after brachial plexus reconstruction (Dahlin, 2008;Socolovsky et al., 2011). A commonly used nerve for ANT is the N. suralis. For the reconstruction of nerves with a large caliber, even the Nn. surales from both legs can be used.
Due to the limited number of donor nerves and differences in caliber between the donor and the injured nerve, the therapy is particularly problematic when treating extensive nerve defects. After sacrificing a nerve for transplantation, patients suffer from donor site morbidities such as loss of sensibility, paresthesia, pain, allodynia, cold sensitivity, and functional impairment (Hallgren et al., 2013;Ijpma et al., 2006;Tada et al., 2020).
Thus, the development of a potent alternative to ANT or nerve transfer is of high clinical relevance. The ideal tissue-engineered nerve Significance Peripheral nerve injuries lead to neurological deficits distal to the lesion side. They are caused by systemic diseases or localized damage. Affected patients experience sensory deficits, restriction of motor functions, or both. While short transection injuries are able for spontaneous recovery, long-distance injuries require surgical treatment.
The standard procedure is nerve grafting, which remains suboptimal, for example, because of donor site morbidity.
Tissue-engineered nerve grafts seeded with adipose tissue stem cells hold great potential as substitutes for autologous nerve grafts to bridge peripheral nerve defects since these cells stimulate axon regrowth and myelination.
F I G U R E 1 21-year-old patient with broken glass cut at the base of the left small finger. (a) Intra-operative finding: neurotmesis of the radial digital nerve. (b) Intra-operative aspect after microsurgical primary end-to-end peri-neural suture (coaptation) of the nerve [Color figure can be viewed at wileyonlinelibrary.com] should fulfill several criteria. First, the material must be biocompatible and not provoke a foreign body reaction. Second, structural similarity to nerve tissue is found useful to enhance directed nerve growth.
Permeability and conductivity should be enabled. The tissue-engineered nerve should also resemble the physiological nerve tissue mechanically with respect to features such as flexibility.
F I G U R E 2 (a) Schematic drawing of a peripheral nerve. Each myelinated axon is surrounded by a connective tissue layer, the so-called endoneurium. The enclosed fibers are grouped and organized into motor or sensory fascicles. The perineurium surrounds these nerve fascicles. Several fascicles may be bundled together within the inner epineurium, whereas the outer epineurium is the outermost layer surrounding a peripheral nerve. The inner epineurium also contains blood vessels and a small amount of adipose tissue. (b) Schematic representation of the classification of peripheral nerve injury. The severity of peripheral nerve injury as classified by Seddon (Neurapraxia, Axonotmesis, Neurotmesis) and by Sunderland (I -V). Peripheral nerve injuries differ depending on the impaired structures, which determines if surgical therapy is needed. Sunderland I corresponds to neurapraxia, and Sunderland V corresponds to neurotmesis. Axonotmesis is further subdivided into three subgroups. In Sunderland II, only the axon is injured, while the perineurium, fascicles, and endoneurium remain intact. In Sunderland III, besides the axon also the endoneurium is destroyed. As the inner guiding structure of the nerve is impaired. In Sunderland IV, only the perineurium is preserved, while the inner structures of the peripheral nerve are destroyed [Color figure can be viewed at wileyonlinelibrary.com] During the last decades, multiple nerve conduits have been tested both in vitro and in vivo. Beginning in the 19th century, several autologous materials such as arteries (performed first by Bünger in 1891), veins (performed first by Wrede in 1909), and bone (probably first performed by Glück in 1881/Vanlair 1889) were used as an outer guidance structure for peripheral nerve regeneration (Chiu & Strauch, 1990;Ijpma et al., 2008). Primarily veins are still used for bridging small nerve gaps. For longer nerve gaps, veins have not produced satisfying results as they tend to collapse. Muscle fibers have also been used as a nerve-guiding structure, as well as veins filled with mucle fibers (Marcoccio et al., 2010). Even though the donor site morbidity with those tissues is lower than for ANT, results were of inferior quality, especially for more extended defects.
Additionally, cell therapy has been used to enhance nerve growth.

| SCHWANN CELL THER APY FOR NERVE REG ENER ATI ON
Schwann cells (SCs) are crucial for peripheral nerve regeneration.
They are specific glial cells and aligned along the axons; they are able to adjust their physiology to generate appropriate feedback responses to support and control neuronal function when releasing neurotrophic factors (Figure 3), such as the nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3 (Nt3), ciliary neurotrophic factor (CNTF), glial cell-derived neurotrophic factor (GCNF), artemin, and vascular endothelial growth factor (VEGF)-especially during Wallerian degeneration (Bunge, 1993;Nocera & Jacob, 2020;Samara et al., 2013;Waller, 1851). NGF acts on the growth of sensory and sympathetic neurons in the peripheral nervous system (PNS) as it leads to an increased innervation density, neuron cell body size, axonal terminal sprouting, and dendritic outgrowth (Rocco et al., 2018). BDNF is essential for synaptic plasticity in the central nervous system, and it acts neuroprotective during hypoxia (Binder & Scharfman, 2004;Chen et al., 2013). Nt3 provides axon outgrowth (Maisonpierre et al., 1990). CNTF and artemin are secreted after injuries or stress to the PNS and prevent the neurons from degeneration (Baloh et al., 1998;Sendtner et al., 1990Sendtner et al., , 1992). An antioxidant effect was described for GCNF, as it counteracts oxidative stress (Mishchenko et al., 2018). SCs can be distinguished into myelinating SCs (mSCs) or nonmyelinating SCs. The former ones ensheath nerves in a layer of myelin, and the latter wrap several axons, forming a Remak bundle . By a process involving rotation, the mSC membrane is elongated and wrapped up to hundreds of times around the underlying axon (Aguayo et al., 1976;Ritchie & Rang, 1983). Several cytoskeletal and adhesion proteins, such as F I G U R E 3 A composite diagram summarizing features of the myelination process (upper row) and the myelin sheath in the PNS (magnified section below). Schwann cells (SCs) develop from precursor cells (Nc. -nucleus), which are involved in neuronal development and maintenance by the release of soluble factors (NGF -nerve growth factor, VEGF -vascular endothelial growth factor, BDNF -brain-derived neurotrophic factor, GDNF -glial cell-derived neurotrophic factor, Nt3 -neurotrophin-3, CNTF -Ciliary Neurotrophic Factor). The diagram depicts the arrangement of a myelinating SC (mSC) surrounding an axon. The mSC ensheathes the axon in a layer of myelin by rotation. Short radius: The mSC membrane is separated from the axonal membrane by the periaxonal space. Adhesion proteins (p75 -p75 neurotrophin receptor, MAG -Myelin-Associated Glycoprotein) mediate connectivity and further interactions between the axon and the mSC. The myelin sheath is formed by the apposition of the external and internal surfaces of a myelin bilayer that constitute the intraperiodic line (IPL) and the major dense line (MDL). The myelin bilayer has an asymmetric composition of cytoskeletal and adhesive protein, with P0 (P0 -myelin protein zero) in the IPL and PMP22 and MBP in the MDL (PMP22 -Peripheral Myelin Protein 22, MBP -Myelin Basic Protein) [Color figure can be viewed at wileyonlinelibrary.com] myelin basic protein (MBP), peripheral myelin protein 22 (PMP22), and myelin-associated glycoprotein (MAG), participate in the myelination process and are found extracellular in the intraperiodic line (IPL) and intracellular in the major dense line (MDL) of the myelin sheath ( Figure 3; Boggs, 2006;Quarles, 2007;Snipes et al., 1992).
The myelin sheath wrapped by one mSC encapsulates an axon and forms an internodal region, while the remaining gap in the sheath forms when neighboring mSCs extend and wrap the myelin. This is called the node of Ranvier, a tiny stretch of axon membrane (~1 mm wide) more freely accessible for the extracellular solution, with a high density of sodium channels (Baker, 2002). Myelin sheaths and the formation of internodes increase the velocity of nerve conduction by limiting the sites of ionic transfer along the axon, which results in a process termed saltatory conduction (Muzio & Cascella, 2020).
Upon injury, Wallerian degeneration occurs at the distal nerve stump during the first few days (Chen et al., 2015;Perry et al., 1987).
Initially, SC responses can be detected within a few hours after injury (Jessen & Mirsky, 2016). They demyelinate, proliferate, and transdifferentiate into the so-called repair SCs, depending on the ubiquitin-proteasome system (Nocera & Jacob, 2020). The ubiquitin-proteasome system alters the gene expression profile of the cells, meaning that genes like L1, p75NTR, and GFAP (glial fibrillary acidic protein gene) are upregulated while myelin-associated genes (myelin transcription factor Egr2, MAP, MAG, and periaxin) are downregulated (Trapp et al., 1988). Thus, SCs promote lysis and phagocytosis of myelin that contains inhibitors of neurite outgrowth, such as MAG and oligodendrocyte-myelin glycoprotein (Beuche & Friede, 1984).
MAG inhibits the neurotrophin receptor p75 in SCs, which leads to cell death. It also inhibits SC migration (Chaudhry et al., 2017).
The expression of cytokines, such as tumor necrosis factor (TNF)-α, leukemia inhibitory factor, interleukin (IL)-1α, -1β, or monocyte chemotactic protein (MCP)-1, is upregulated during Wallerian degeneration (Rotshenker, 2011). These cytokines recruit macrophages and other immune cells to the side of injury and stimulate the SCs to proliferate.
The macrophages support the SCs with removing myelin debris, and the neurotrophic factors secreted by the SCs contribute to the survival of injured neurons and axon elongation (Lindholm et al., 1987).
Peak proliferation occurs around 4 days post-injury. Proliferating SCs are confined to their basal lamina tubes, where they align to form the so-called bands of Büngner, which provide a supportive substrate and growth factors for regenerating axons (Salzer & Bunge, 1980). The bands induce nerve regeneration from the injury site to the target areas. The success of functional nerve regeneration is dependent on several factors: the sprouting fibers have to direct toward the correct target organ, and the distal stump must have neuronal contact. Otherwise, the SCs denervate chronically and do not participate in nerve regeneration anymore. The bands of Büngner disappear (Weinberg & Spencer, 1978). Then, the denervated target organ is exhausted of trophic factors, muscle fiber atrophy, and satellite cells undergo apoptosis (Jejurikar et al., 2002;Lee et al., 1995).
SCs have been applied for nerve regeneration in vivo, as well as for peripheral nerve tissue engineering to enhance nerve growth (Hyung et al., 2019). They promote peripheral nerve regeneration when used with an outer guiding structure; for example, bridging a 6 cm critical gap size of the N. peroneus in rabbits with a vein filled with autologous SCs shows no significant difference between an intact nerve and the graft with respect to the number of myelinated fibers 4 months after reconstruction. Furthermore, no SCs or axons are detected at the distal nerve stump in the control group (vein without SCs; Strauch et al., 2001). Also, a 4 cm gap of the N. tibialis in rabbits could be bridged using a vein filled with SCs, though leading to inferior results concerning motor nerve conduction velocity and fiber count, when compared to the autogenous nerve graft (Zhang et al., 2002).
Also, the reconstruction of a sizeable sciatic nerve injury (7.5 cm) in a 25-year-old human patient with ANT combined with autologous SCs has led to functional recovery of all motor and sensory functions (Levi et al., 2016). As in such clinical settings, control groups are not possible; it is not unambiguously comprehensible whether the recovery would have been similar without the addition of SCs.
Decellularized nerve grafts of different sizes have been used in combination with SCs. For a 1 cm nerve gap in the rat sciatic nerve, similar results are found for decellularized nerve grafts with SCs and the ANT regarding myelinated fiber number and myelin sheath thickness (Sun et al., 2009). In another study, a 14 mm defect of the rat sciatic nerve was bridged with a decellularized nerve graft combined with allogeneic rat SCs. In this study, SCs harvested from motor nerves were compared to cells harvested from sensory nerves. Via PCR, it could be found that VEGF, pleiotrophin, protein kinase C iota, and neurofilament were upregulated in motor nerve SCs, while BDNF, GDNF, MBP, and neural cell adhesion molecule were upregulated in SCs of sensory nerve origin. Only NGF was similarly upregulated in both SC types. Forty-two days after injury, histomorphometric analysis shows similar numbers of regenerating nerve fibers in decellularized nerve grafts with SCs when compared to autografts. Eighty-four days post-injury, muscle force generation was also similar in those two groups. SC source did not affect nerve fiber counts or muscle force generation (Jesuraj et al., 2014). In a comparable study, a 2 cm and a 3 cm nerve gap of the rat sciatic nerve has been bridged using a decellularized nerve graft seeded with autologous SCs. For the 2 cm defect, no significant difference regarding regenerated axons and fiber density was found compared to the autograft 10 weeks after injury (Hoben et al., 2014). However, 12 weeks after injury, the motor functions did not recover in the 3 cm defect, although nerve sprouting was observed histologically (Aszmann et al., 2008). A 6 cm ulnar nerve defect in Macaca fascicularis has been bridged with cold-preserved allografts seeded with autologous SCs (Hess et al., 2007). Six months after injury, increased fiber counts, nerve density, and percentage of neural tissue (including axons, SCs, and myelin) are found. In conclusion, the mentioned studies indicate that resident SCs support nerve repair not only during physiological regeneration but also after secondary insertion. It is also shown that with an increasing defect size, the regeneration results are of lower quality.

TA B L E 1 (Continued)
Also, various biodegradable nerve conduits have been combined with SCs. When a 7 mm gap in the rat sciatic nerve model is bridged with a polylactide-co-glycolide foam conduit seeded with SCs, the conduit does not degenerate until 6 weeks post-surgery.
The total fiber number is significantly smaller than in the autologous nerve transplantation group, but the axon diameter has been measured 3.7 µm, which is significantly higher than found for autografts (2.3-µm; Hadlock et al., 2000). Also, a poly-l-lactic acid-based multiwall carbon nanotube has been seeded both with rat SCs, and curcumin encapsulated chitosan nanoparticles (Jahromi et al., 2020). Bridging a 10 mm sciatic nerve defect in rats with this tube has confirmed a specific SC-and curcumin-induced enhancement of sciatic nerve regeneration. The sciatic function index is comparable between the scaffold and the autograft. Also, the weight of the musculus gastrocnemius is similar to the lesion site for these groups.
Chitosan has also been used together with SCs to promote peripheral nerve regeneration. A critical size defect of 15 mm in rat sciatic nerve of healthy and diabetic rats can be bridged using a second-generation chitosan film-enhanced chitosan nerve guide with or without fibroblast growth factor (FGF)-2-overexpressing SCs.
With a very long observation time of 120 days, it has been shown that the addition of FGF-2-overexpressing cells leads to inferior results concerning sensory recovery, muscle weight ratios for the gastrocnemius muscle and the tibialis anterior muscle when compared to the implantation of cell-free second-generation chitosan film-enhanced chitosan nerve guide. However, it is not clear whether the SCs have survived inside the nerve guide . This is in contrast to the observation that bridging a 15 mm critical size nerve defect in a rat sciatic nerve with a chitosan nerve guide filled with FGF-2-overexpressing SCs containing hydrogel increases functional recovery 17 weeks after injury .
The neurotrophic effects of SCs become evident when compared to sham groups, especially in histological measures of nerve density, percentage of neural tissue, and total fiber counts. This indicates that the neuro-enhancing properties of SCs, such as the upregulation of axonotrophic cellular adhesion molecules, production of NGF, BDNF, and GDNF, also lead to axonal outgrowth and provide trophic support to regenerating growth cones even after SC transplantation. However, when aiming at clinical application, the cultivation of SCs in vitro is demanding, and the proliferation is slow (Kingham et al., 2007). Likewise, the harvesting of autologous SCs in the necessary amount results in donor site morbidity, so that it is of particular interest to identify cells with similar characteristics but lower donor site morbidity and better availability. Finally, harvesting autologous SCs in an acute trauma situation with a nerve defect would lead to a delay for the nerve reconstruction, which is also disadvantageous.

| AD IP OS E TISSUE S TEM CELL S AND PERIPHER AL NERVE REG ENER ATI ON
The injection of fat grafts around a coapted sciatic nerve in rats leads to reduced perineural adherence and minor scar formation (Cherubino et al., 2017). Furthermore, ASCs have been solely investigated for peripheral nerve regeneration (for an overview, see Table 1).
After crush lesion of the rat sciatic nerve, the transplantation of canine ASCs to the injured area accelerates functional motor recovery, as shown by sciatic function index gait analysis and electromyography during 3-week post-injury (Rodríguez Sánchez et al., 2019). In this study, the ASC-treated group achieved values close to healthy animals; a significant improvement was observed between weeks 2 and 4 after injury. In electromyography, the latency was higher than in the healthy control group, meaning no improvement during the 4-week post-surgery. An ultrasound-controlled epineural injection of allogenic rat ASCs to the rat sciatic nerve after crush injury accelerates nerve regeneration. Myelin thickness, fiber diameter, and fiber density ratios were highest in the ASC-treated group, while the relative gastrocnemius weight ratio was inferior compared with functional analysis, indicating that function and morphology did not correlate (Tremp et al., 2018).
As already described for SCs, autologous veins filled with ASCs were also used for peripheral nerve regeneration. Bridging an 8 mm gap in the sciatic nerve with this conduit leads to discouraging results. Compared to the sham group, the sciatic function index is significantly lower 30 and 60 days after the nerve reconstruction. Also, a loss of large myelinated fibers, together with an increase in smaller myelinated fibers is observed, which indicates that axonal sprouting is enhanced by the presence of ASCs. However, the axonal sprouting occures both intra-and extra-fascicular, and the myelin sheath is thinner than found for the sham group (Fernandes et al., 2018).
These findings can be interpreted that treatment with undifferen-  (Lasso et al., 2015). In contrast, neuronal differentiation of ASCs is confirmed after bridging a 12 mm gap in the rat sciatic nerve with a genipin-cross-linked gelatin conduit annexed with tricalcium phosphate ceramic particles loaded with ASCs using immunofluorescence staining of the previously labeled cells (Shen et al., 2012).

| ACELLUL AR NERVE SC AFFOLDS AND AD IP OS E TISSUE S TEM CELL S
Acellular nerve scaffolds seeded with nerve leachate-treated ASCs were seeded on acellular nerve scaffolds and used for repair of a 1 cm injury of the rat sciatic nerve . The highest number of regenerated nerve fibers with thick myelin sheaths can be detected after treatment with nerve leachate-treated ASCs on acellular nerve scaffold and autograft groups when compared to untreated ASCs. Regarding the neurophysiology, there is no difference between the generation of action potentials or the conduction velocity between the ANT and the acellular nerve scaffold with pretreated ASCs. The treatment with nerve leachate offers the additional advantage that nerves of different species can be used without provoking any donor site morbidity . Moreover, ASCs labeled with luciferase by lentiviral particles could be detected at the side of implantation for at least 29 days. They do not migrate to other tissues, indicating that those cells enhance nerve regeneration . Other studies have also addressed ASC survival after implantation in vivo and in peripheral nerve regeneration, which showed different time spans of cell survival. This also depended on the conduit used with the cells. ASCs seeded on a Type 1 collagen scaffold as well as a fibrin conduit survive for 21 days after implantation into the dorsal muscle in rats and for 20 days after subcutaneous implantation, respectively (Venugopal et al., 2017;Wolbank et al., 2007). By contrast, when applying ASCs within a poly-3-hydroxybutyrate conduit to reconstruct a 1 cm sciatic nerve gap in rats, already 14 days after implantation, a significant loss in the number of transplanted cells was shown via cell labeling and PCR analysis (Erba et al., 2010). Subsuming those results, the fate of ASCs after implantation in vivo remains unclear. While several nerve conduits profit from the addition of ASCs regarding functionality after nerve injury, this indicates positive effects of the cells on peripheral nerve regeneration. Nevertheless, cell tracking and measures of ASC vitality remain heterogeneously.
Gene expression analyses have indicated that the neurotrophic markers NGF and GCNF are downregulated when ASCs have been seeded on a decellularized (motor) nerve graft, suggesting that upon adhesion to the substrate, ASCs become less involved in sensory nerve growth. In contrast, all myelination and angiogenesis markers are increased compared to control cells . Avascular nerve scaffolds seeded with ASCs have also been used to bridge the gap between the contralateral C7 nerve root and C5-6 nerve in a rat brachial plexus injury model. The avascular nerve scaffolds seeded with ASCs accelerated nerve regeneration when compared to cellfree avascular nerve scaffolds. Admittedly, in the given study, the ANT led to the best results (Yang et al., 2019).

| AD IP OS E TISSUE S TEM CELL D IFFERENTIATI ON INTO SCHWANN CELL-LIK E CELL S
ASCs can be differentiated into Schwann cell-like cells (SCLCs) by stimulation with different soluble factors (Faroni et al., 2016). A frequently used differentiation protocol is described by Kingham et al. (2007), which uses 5 ng/ml platelet-derived growth factor, 10 ng/ml basic fibroblast growth factor, 14 μM forskolin, and GGF-2 252 ng/ml (Kingham et al., 2007). Also, other differentiation media with glial growth factors are described (Wong et al., 2020).
Alternative protocols have detected that SCLC differentiation of ASCs is possible by treatment with SC-conditioned medium, olfactory ensheathing cell-conditioned medium, the addition of insulin and progesterone, or nerve leachate (Fu et al., 2016;Kang et al., 2019;Liu et al., 2020;Xie et al., 2017). Intermittent induction and cell-imprinting methods are described for ASC differentiation into SCLCs (Moosazadeh Moghaddam et al., 2019;Sun et al., 2018). It has also been demonstrated that lentivirus transfection of GCNF induces SCLC differentiation of ASCs (Zheng & Liu, 2019). Interestingly, the factors applied for SCLC differentiation determined the features of the resulting cells. The addition of forskolin induces the expression of S100ß and P75, while medium with β-mercaptoethanol and without forskolin leads to the expression of S100ß, MAG, and P0, indicating a promyelinating phenotype (Schuh et al., 2014). Furthermore, the secretory activity of SCLCs is influenced by the cells' microenvironment, for example, two self-assembling hydrogels (PeptiGel-Alpha 1 and PeptiGel-Alpha 2) have been seeded with human SCLCs . The hydrogels differed concerning stiffness and charge.
Compared to ASCs cultured on collagenase I conduits, the cells express higher levels of BDNF, neuregulin, and NGF, while there is no increase in GDNF and Nt3 . Glial cell markers, such as S100ß, glial fibrillary acidic protein, MBP, and p75 neurotrophin, are expressed by the SCLCs in vitro and can be used for identification via immunostaining or western blotting. During in vitro differentiation, the morphology of the ASCs changes from a monolayer of large cells with a flat morphology to a small number of bipolar or tripolar spindle-shaped cells (Fu et al., 2016;Kingham et al., 2007;Mathot et al., 2020). The change in shape is enhanced by muscarinic cholinergic receptor M2 activation. Muscarinic receptors are present both in SCs and in SCLCs (Piovesana et al., 2019). Contradictory findings were reported by Faroni et al. (2016), that after an 18 days of differentiation period, the shape of the SCLCs change back to fibroblastlike. This indicates a redifferentiation to ASCs, which is accompanied by a decreased expression level of BDNF (Faroni et al., 2016). These findings might suggest a fragile differentiation of ASCs to SCLCs and put the durability of the differentiation into question, as well as the possible effects of the gained neurotrophic features as well.
Similar to SCs, SCLCs promote neurite outgrowth and elongation (Kingham et al., 2007). In a coculture with the NG108-15 motor neuron-like cell line, the number of NG108-15 cells expressing axons increases, as it is the case for the mean length of the longest neurite.
This effect is not induced by ASCs. Also, the myelin-forming ability of SCLCs has been examined. In vitro, myelin protein P0 is increased in SCLCs upon elevated levels of intracellular cAMP (Xu et al., 2008).
SCLCs can form myelin with PC12 cell neurites, which has not been shown for ASCs (Xu et al., 2008). In vivo, a fibrin conduit seeded with SCLCs has been applied to bridge a 10 mm gap in the rat sciatic nerve (Di Summa et al., 2018). Twelve weeks after injury, the myelinated portions of the mid and distal conduit segments are thinner than after ANT but significantly stronger than after insertion of the cellfree conduit. Also, the amount of collagen is similar to the autograft, indicating that only little scar formation due to collagen IV occurs (Di Summa et al., 2018). In another study, a 15 mm sciatic nerve defect in rats was bridged with polyglycolic acid (PGA-c) tubes seeded with SCLCs. Eight weeks post-surgery, the axons exposed to SCLCsinjected PGA-c tubes reach the distal nerve stump, which indicates directed axonal sprouting. By contrast, the wet weight ratio of the gastrocnemius muscle (8 and 12 weeks after reconstruction) has been measured significantly lower in the SCLC-treated group than after ANT. This means that no complete innervation of the muscle occurred. Additionally, the SCLCs could be detected via cell tracing 8 weeks after injury (Yamamoto et al., 2020). Nerve conduction examination of the reconstructed sciatic nerve also showed a decrease in conduction velocity and a decreased compound muscle action potential when compared to the ANT. SCLCs derived from human ASCs have been seeded on graphene oxide substrates (Verre et al., 2018).
The combination of SCLCs and avascular nerve scaffolds inhibits the degradation of motor endplates, increases the expression of neurotrophic factors, and downregulates the expression of Janusactivated kinase 2, which is supposed to induce improved peripheral nerve regeneration (Fu et al., 2019). By contrast, Sowa et al. filled gelatin hydrogel tubes with ASCs from P0-Cre and glial fibrillary acidic protein-Cre/Floxed-double reporter mice and transplanted the scaffolds into wild-type mice. Afterwards, they could not find the green fluorescent protein-positive cells during histology, which indicated that the cells did not differentiate into SCs in the regenerating nerve tissue (Sowa et al., 2016). As also undifferentiated ASCs are able to secrete neurotrophic factors (Tomita et al., 2013), it remains questionable how permanent differentiation of ASCs to SCLCs is, how long the cells remain differentiated in vivo and how important the differentiation of ASCs to SCLCs is, although beneficial effects have been shown after differentiation as described above. sheets. These sheets stimulate SC migration, neuron proliferation, and neurite outgrowth in vitro. They also improve the functional recovery, nerve reinnervation, axon regeneration, and remyelination in the rat sciatic nerve injury model in vivo (Hsu et al., 2019).

| NERVE CONDU ITS AND AD IP OS E TISSUE S TEM CELL S
A critical size gap of 15 mm in the rat sciatic nerve is bridged using a poloxamer hydrogel inside a polycaprolactone conduit with human ASCs leading to promising results regarding the increase in expression of neurofilament protein and S100ß when compared to the cell-free conduits. Also, the conduit loaded with ASCs shows significantly higher numbers of myelinated nerve fibers than the control group. Similar results are shown for polycaprolactone conduit with a poloxamer hydrogel loaded with SCs. Furthermore, the best structural outgrowth from the proximal stump is found in the combined hydrogel and ASC group (Allbright et al., 2018). A scaffold made from a biodegradable cryo-polymerized gelatin methacryloyl gel loaded with ASCs and used for bridging a 10 mm sciatic nerve gap in rats leads to similar results in functional recovery and axonal regeneration (regarding sciatic function index, electrophysiological results, and nerve and muscle fiber diameters) when compared to ANT (Hu et al., 2016). Even though these results indicate that a scaffold in combination with ASC promotes nerve regeneration similar to the autologous nerve transplantation, this is not true for all nerve conduits for bridging a 10 mm nerve gap. For example, a 10 mm sciatic nerve gap in rats has been bridged with NeuraGen ® (Integra LifeSciences Corporation) with and without ASCs. NeuraGen ® is an FDA-approved nerve conduit, consisting of collagen type I and thus being absorbable. Increased S100 immunoreactivity is detectable in the ASC-seeded conduit group 6 months post-injury, but not in the control group without ASCs. Axon arrangement inside the conduits was found more organized, whereas fewer single axons with ineffective regeneration and foreign body responses have been documented for the control group without ASCs (Klein et al., 2016).
SCs and ASCs can be seeded on silk fibers, which is another promising scaffold in peripheral nerve regeneration (Resch et al., 2018). Rat SCs cocultured with ASCs are seeded on a silk fibroin/collagen scaffold to construct a tissue-engineered nerve scaffold for bridging a 10 mm gap in the rat sciatic nerve. Twelve weeks post-surgery, the compound muscle action potential shows no significant difference between the autograft and the scaffold animals. Nevertheless, muscle atrophy, tactile allodynia, and thermal hyperalgesia are not improved using this scaffold (Xu et al., 2016). Also, Fernandes et al. did not detect improved nerve regeneration in animals treated with Matrigel ® (Corning Inc.) conduits containing ASCs when compared to cell-free Matrigel ® .
Morphometrically, the nerve fibers sprouting the Matrigel ® conduits show a minor diameter, and fewer motor neurons can be found (Fernandes et al., 2018).

| CON CLUS ION
The procedure of ANT as the clinical gold standard for peripheral nerve repair is associated with certain difficulties, including donor nerve sacrifice and nerve mismatch. During the last years, the approaches for cell-assisted nerve regeneration have gained F I G U R E 4 Schematic design of adipose tissue stem cells harvest for cell-based therapy on peripheral nerve regeneration with a nerve conduit seeded with cells. After peripheral nerve injury, harvested adipose tissue is taken for isolation of adipose stem cells. Adipose tissue stem cells are multipotent stem cells that can be differentiated upon specific stimulation into Schwann cell like-cells. Adipose tissue stem cells and/or Schwann cell like-cells can be seeded on a scaffold serving as a conduit for peripheral nerve regeneration (see text for details) [Color figure can be viewed at wileyonlinelibrary.com] significant progress. The increasing understanding of the regenerative character of adult stem cells, that is, adipose tissue stem cells, allowed for experimental investigations on the use of these cells for therapeutic application after peripheral nerve injuries.
ASCs can differentiate into multiple phenotypes, including the SCLCs. Furthermore, ASCs, as well as SCLCs are able to support and/or promote peripheral nerve regeneration when combined with various nerve scaffolds in vivo. While therein, the differentiation of ASCs to SCLCs is assumed to mimic the biological features of natural SCs, the durability and stability of the SCLC phenotype have not been approved in vivo or in vitro. Importantly, the regenerative effects of ASC-seeded conduits can exceed the effects of SCLC or SC-seeded conduits. Though not every aspect of the mechanism of ASCs supporting peripheral nerve regeneration is understood, these cells remain promising candidates for supporting peripheral nerve regeneration.

ACK N OWLED G M ENT
Open access funding enabled and organized by Projekt DEAL.

CO N FLI C T O F I NTE R E S T
No potential conflict of interest, financial or otherwise are declared by the authors.

AUTH O R CO NTR I B UTI O N S
All the authors had full access to all the data in the study and take re-

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1002/jnr.24738.