Biomimetic Strategies for Peripheral Nerve Injury Repair: An Exploration of Microarchitecture and Cellularization

Abstract

Injuries to the nervous system present formidable challenges to scientists, clinicians, and patients. While regeneration within the central nervous system is minimal, peripheral nerves can regenerate, albeit with limitations. The regenerative mechanisms of the peripheral nervous system thus provide fertile ground for clinical and scientific advancement, and opportunities to learn fundamental lessons regarding nerve behavior in the context of regeneration, particularly the relationship of axons to their support cells and the extracellular matrix environment. However, few current interventions adequately address peripheral nerve injuries. This article aims to elucidate areas in which progress might be made toward developing better interventions, particularly using synthetic nerve grafts. The article first provides a thorough review of peripheral nerve anatomy, physiology, and the regenerative mechanisms that occur in response to injury. This is followed by a discussion of currently available interventions for peripheral nerve injuries. Promising biomaterial fabrication techniques which aim to recapitulate nerve architecture, along with approaches to enhancing these biomaterial scaffolds with growth factors and cellular components, are then described. The final section elucidates specific considerations when developing nerve grafts, including utilizing induced pluripotent stem cells, Schwann cells, nerve growth factors, and multilayered structures that mimic the architectures of the natural nerve.

Background of Nerve Anatomy, Physiology, Regenerative Mechanisms

Regeneration of tissues within the human nervous system is a particularly intricate process. In comparison with the central nervous system (CNS), the peripheral nervous system (PNS) displays more robust nerve regeneration [1]. Thus, the PNS offers unique opportunities for understanding fundamental principles of nerve injury and growth, along with promising areas for clinical intervention. However, while peripheral nerve cells can regrow, they do so only to a limited extent, and complications such as scarring, misdirection, and inflammation can impair tissue regeneration, abrogate motor and sensory recovery, and lead to significant pain sensation for patients [2]. Approximately 560,000 transected nerve injury procedures are performed each year in the USA [3]. Given the large caseload and the outsized effect of nerve injury on patient quality of life, new solutions to peripheral nerve injury (PNI) and nervous system injury and disease more broadly are sorely needed. Yet no current interventions adequately address peripheral nerve injury repair, especially in cases complicated by long-gap length (> 1 cm), severe PNI according to the Sunderland scale classification, and older injuries that were either surgically treated but unsuccessful at restoring function or not treated at all. These complex cases have limited intervention options and represent around 32% of all peripheral nerve transection procedures [3]. In this article, we will discuss current technology in PNI repair, and the recent developments in stem cell biology, including combinations of induced pluripotent stem cells with novel biomaterial polymer formulations that can augment nerve regeneration [4, 5].

Several previous reviews have focused on various aspects of nerve regeneration, such as design, materials, and fabrication methods [6, 7], drug delivery [8], molecular mechanisms of peripheral nerve regeneration [9], effects of topography [10], bioengineering strategies [10], and surgical options [5, 11]. This article will build on this existing literature and present an integrated overview of biomaterials with an emphasis on peripheral nerve grafts functionalized with biological factors such as growth factors and cellular constituents. Specifically, we aim to expand on the work by Oliveira and co-workers, who provided a detailed analysis of the current state of research to improve the “hollow nerve guidance conduit” [4].

In this review, we first explore the microanatomy of peripheral nerves; next, we describe suitable biomaterials for use in peripheral nerve grafts and conduits; finally, we discuss the incorporation of biochemical factors and live cells into the framework of the biomaterial nerve conduit. In essence, we aim to describe approaches necessary to recapitulate a human nerve in vitro and thus generate a graft with a structure and function generalized to meet the needs of any PNI. We describe repairs in the context of complete nerve transects; however, principles inherent to our discussion may be used to address partial nerve transects, nerve crush injuries, and other nerve pathologies.

Peripheral Nerve Anatomy

In the PNS, axons travel in specific channels called endoneurium (Fig. 1) [12]. Within the endoneurium, axons are surrounded by Schwann cells. Schwann cells are the primary neuron-supporting cells of the PNS and have diverse roles in nerve health and function. During early embryological development, neural crest cells give rise to Schwann Cell Precursors (SCPs) that migrate to early nerves where they develop through an extensive process into non-myelinating Schwann cells and myelinating Schwann cells [13]. Myelinating Schwann cells form tight associations with axons and function primarily to facilitate long-range transmission of nerve impulses (i.e., action potentials) [14]. In contrast, non-myelinating Schwann cells act as support cells in the perineurium by providing essential trophic support to sensory and motor nerves through secreted growth factors and extracellular matrix (ECM) [14].

Fig. 1.

Schematic of nerve anatomy. Nerves are protected by fat and other connective tissues. The outermost layer is the epineurium, which contains nerve fascicles, vasculature, and connective tissue. Nerve fascicles are surrounded by perineurium. Within fascicles, Schwann cells wrap around axons to increase nerve conduction and also play a protective role in axon regeneration following PNI. Adapted from Hendriks et al. [12]

The exterior surface of the endoneurium comprises a fibrous matrix of mostly type I and III collagen [15]. Individual axons, along with their endoneurium components, are collectively gathered into a nerve bundle or “fascicle.” This collection of nerve fibers is surrounded by the next layer of extracellular material known as the perineurium. This layer includes collagenous connective tissue and cellular components. The perineurium thus delineates one fascicle, or nerve bundle, from other nerve bundles within the larger nerve architecture [16]. Each fascicle may project to a different end-target and have a different function. For example, one fascicle within the nerve may contain motor axons that project to neuromuscular junctions, whereas other fascicles within the nerve may contain sensory nerve fibers that project to pain receptors [17]. Interspersed between these fascicles are fat and connective tissue, vasculature, and cells such as fibroblasts (FBs), all of which support ECM structures and the function of the Schwann cells and neurons in varying capacities [4, 17]. This layer can be considered the inner epineurium. Finally, the entire nerve assembly is encased by the outer epineurium [18]. The outer epineurium is also surrounded by fat and blood vessels needed to nourish the external aspects of the nerve. The ability for nutrients and waste products to diffuse throughout the structure is critical to the function of the nerve and a major limiting factor in some nerve repair technologies [12, 18].

Each of these layers—the endoneurium, perineurium, and epineurium—play an integral role in the function of peripheral nerves [13]. The epineurium provides protection and keeps the nerve contents, including growth factors, within the nerve. It is important to note that in injuries where the epineurium is left intact, peripheral nerves have a greater chance to regrow successfully. Conversely, when the epineurium integrity is jeopardized, nerves regenerate more poorly [19]. The perineurium, as described previously, helps delineate and contain specific fascicles with similar functional relevance, while the endoneurium provides protection to individual axons and their respective Schwann cells.

When developing nerve repair technologies, it is also important to recognize the diversity of nerve morphology with respect to various functions [20]. Individual nerve fiber diameters can range from approximately 1 to 20 μm. The diameters of these nerve fibers are correlated with their conduction velocity and, ultimately, nerve function [20]. The endoneurium channels thus scale according to these diameters and these dimensions are an important consideration for biomaterial fabrication. These dimensional characteristics are used in classifying nerve subtypes, which are listed in Table 1 [21].

Table 1.

Peripheral nerve classification and properties

Electrophysiologic classification of peripheral nerves	Classification of afferent (sensory) fibers	Fiber diameter (μm)	Conduction velocity (m/s)
Sensory fibers
Aα	Ia and Ib	13–20	80–120
Aβ	II	6–12	35–75
Aδ	III	1–5	5–30
C	IV	0.2–1.5	0.5–0.2
Motor fibers
Aα	N/A	12–20	72–120
Aγ	N/A	2–8	12–48
B	N/A	1–3	6–18
C	N/A	0.2–2	0.5–2

Adapted from Ref. [21]

Peripheral Nerve Regeneration Mechanism

During PNI, axons of the peripheral nerve will degenerate from the point of injury to their distal termini (e.g., the target muscle, ganglion, or sensory receptor), as shown in Fig. 2 [21]. This means that newly regenerating axons must grow from the injury point all the way to the denervated neuromuscular junction or sensory receptor before they can perform their intended function. Axon regrowth occurs at the rate of 1–3 mm per day; thus, an injury 30 cm proximal to a target muscle could take nearly 1 year to regenerate [22]. During this time of regrowth, adequate perfusion, nutrient delivery, and cellular support must exist for the developing axon [23, 24].

Fig. 2.

Schematic of Wallerian degeneration and peripheral nerve regrowth. At the time of nerve injury (a), biomolecular signals promote degeneration of the distal portion of the axon (b). As regeneration begins, Schwann cells support the growing axon (c), which must regrow from the injury site to its end destination (d). Adapted from Fix [21]

Before regeneration can occur, however, inflammatory byproducts of the injury must be removed. Degraded axons, along with their damaged myelin sheath and any impaired Schwann cells, are phagocytosed by macrophages and nearby resident, non-damaged Schwann cells [25].

During this process, the non-damaged Schwann cells adopt a distinct “repair” phenotype that downregulates myelin and upregulates neurotrophic factors promoting neuron survival and axonal elongation [26]. Some key transcription factors involved in the induction of the repair-state are c-Jun and STAT3 [27]. These pathways can be activated via exogenous application of upstream repair-promoting factors, including nerve growth factor (NGF), brain-derived growth factor (BDNF), glial-derived growth factor (GDNF), and neurotrophin-3 (NT-3), among others. Studies where these factors were applied exogenously to the site of PNI and knock-out models reinforce their crucial role in PNI repair [28].

Additional growth factors implicated in the repair-response are basic fibroblast growth factor (bFGF) and transforming growth factor beta-1 (TGFB-1). bFGF proliferates FBs, stimulates angiogenesis, and contributes to nerve regeneration. Exogenous application of bFGF has also been particularly successful in long nerve gap models of PNI. TGFB-1 is believed to activate the repair-state Schwann cells by suppressing myelin proteins and mediating SC migration from the proximal to the distal stump [11]. Failure of the Schwann cells to sustain this repair phenotype is believed to contribute significantly to poor regeneration outcomes in humans.

Overall Considerations

For a PNI repair intervention to be successful, the engrafted material should be sufficiently similar to human nerve morphology or provide the substructure upon which ingrowing cells can recapitulate the appropriate microanatomical layers and their intervening ECM and vasculature [24]. More specifically, requisite features of such a construct might include analogous structural polymers, growth factors, cellular components, and appropriate dimensions of the nerve microanatomy [24]. Any graft material used should complement nerve microstructure, augment the growth environment, and either remain in place or degrade as endogenous ECM fills in the graft, with minimal immunogenicity [17, 24]. Reducing the immune response to foreign materials is paramount for any transplant, particularly for nerves. Since sensory fibers travel within the nerve sheath, if graft rejection, inflammation, or irritation occurs, pain can be severe due to immediate local activation of sensory neurons [29].

In summary, based on the nerve anatomy, physiology, and regenerative mechanisms, the main considerations in developing a successful graft include (1) A stable outer sheath recapitulating the epineurium, with porosity adequate to allow diffusion of small molecules while preventing unwanted cell infiltration, (2) A luminal composition consistent with the extracellular matrix of human nerves, or an environment which promotes faithful endogenous replication of the original nerve ECM, (3) Cells representative of nerve histology, or the trophic factors necessary to promote migration of cells upon implantation. (4) Growth factors to sustain axon and support cell survival, (5) Vasculature or diffusion properties sufficient to allow egress and ingress of waste and nutrients, and (6) Minimal immunogenicity. These factors and methods for incorporating them into graft development will be the subsequent focus of this article.

Current Available Interventions for PNI Repair

Current approaches to bridging nerve gaps include (1) Direct suturing of severed nerve ends, (2) Grafting a synthetic nerve guidance conduit, and (3) Grafting human nerves from either the patient or an allograft source (Fig. 3) [30]. Short, clean nerve transects can be repaired via direct suturing, either end-to-end, end-side, or side-side, or via synthetic nerve guidance conduits [5]. Longer nerve gaps, however, currently need to be bridged by a grafted material, either the autograft or allograft; synthetic grafts exhibit promising features for long nerve gap repair but are not routinely used in these instances at present.

Fig. 3.

Current interventions for peripheral nerve injury repair. Gap lengths between severed nerve endings are an important factor in determining the appropriate intervention. Direct sutures are recommended for very short gap injuries that span 0.0–0.5 cm in length. Nerve guidance conduits, currently, are recommended for gap injuries in the range of 0.5–1.5 cm. Allografts are indicated in longer PNI gap lengths 1.0–5.0 cm, and autografts are indicated for any nerve injury greater than 1.0 cm [30]. This diagram shows the paucity of intervention options, especially in patients with longer gap injury lengths

Direct suturing has been attempted nearly since antiquity and the use of modern techniques can often successfully repair short nerve transects [31–33]. However, large gaps usually cannot be bridged effectively by direct suturing [17, 32]. Traction on surgically repaired nerves can cause pain and neuroma at the repair site, necessitating further repairs at later times. Given this limitation, end-to-end nerve suturing is simply not practicable for large nerve gaps [17]. Alternative techniques such as end-side and side-side direct suturing can circumvent some of the issues of traction and nerve gap length [5, 34]. In these procedures, nearby nerves or nerve fascicles are sutured together with the injured nerve to provide collateral support and promote regeneration [35]. While often effective, the utility of these surgeries is limited by the availability of adjacent nerves in the anatomical region of the injury, and by the extent of the injury.

For larger nerve gaps, the gold standard treatment is the autograft, whereby surgeons harvest a “less essential” nerve such as a purely sensory nerve, and then graft that nerve into the injured site [5, 34]. For autografts, surgeons often use the sural nerve, a long nerve in the lower leg that provides only cutaneous innervation. Grafting of nerves derived from the patient themselves has the advantages of low immunogenicity, intact nerve microarchitecture, and some residual cells and growth factors. This procedure, however, also poses several problems. First, this creates another nerve injury site with a sensory and/or motor deficit. Additionally, the caliber of the harvested nerve should match or at least approximate the dimension of the injured nerve. When the harvested nerve is too small to match the diameter of the injured nerve site, it can be bundled into a “cable” to increase the diameter of the graft [36]. This, however, adds time and complexity to the surgery and bundling changes the microscopic arrangement of the graft interior. As noted in the previous section, nerve anatomy can vary significantly depending on the body part. The sural nerve, for example, will contain a different arrangement of nerve fibers and fascicles than the median nerve, which controls the muscles of the forearm and hand. Thus, when an autograft is performed, one limiting factor in its success may be the misalignment of nerve fascicles between the injury site and the graft [17]. Moreover, when removed from the donor site, injury signals in the form of cytokines and chemokines present in the autograft may cause apoptosis or cell morphology changes within the graft [29]. Dead cells, axonal debris, and apoptotic molecules within the autograft could dampen axonal regrowth at the repair site [29]. Although considered the “gold standard,” there is no guarantee of full recovery of injured nerve function after autografting [5].

The nerve allograft is another increasingly common option for peripheral nerve injury repair [17]. In this procedure, a human cadaveric nerve is decellularized and then implanted into the injury site. Allografts have the advantage that the extracellular matrix already recapitulates the porosity, density, and mechanical properties of human nerves. It thus provides the dimensions and composition amenable to axon regrowth through the graft. Additionally, since it is decellularized, it has low immunogenicity and does not contain inflammatory chemokines or cytokines. By virtue of the decellularization process, however, the allograft lacks trophic factors and cellular constituents that could aid axonal regeneration.

Finally, synthetic nerve guidance conduits (NGCs), which typically consist of a hollow tube comprised of biocompatible material such as collagen, have been used to repair short nerve transects, typically less than 1.0 cm [17]. Conduits, however, have not proven to be as successful in regenerating nerves over longer distances, likely because there is insufficient guidance structure within the lumen of the conduit, either from ECM, cellular, or trophic factors [37]. More successful grafts such as the autograft and the allograft may exhibit better regeneration due to their faithful composition representative of nerve microanatomy.

Although synthetic NGCs currently provide inferior nerve regeneration for large nerve gaps, they have several potential advantages: The parameters of architecture, trophic factors, and cellularity can be tailored to the exact needs of the patient [37]. Biomaterials can be patterned in vitro, along with cell and growth factor incorporation into the graft prior to implantation. Synthetic NGCs can be seeded with growth factors via blending or chemical linkage, and cells can be grown into the graft as it is being generated [37, 38]. Cells and growth factors can thus be localized to specific areas within the graft such that they serve as guidance cues for ingrowing axons. Localization of cells can be achieved through engineering biomaterials that comprise distinct layers and arrangements [39]. Combined with the ability to match specific dimensions of nerve size and shape, the addition of cells and trophic factors can provide important supplemental guidance for regenerating nerves.

Furthermore, in the case of NGCs, clinicians have the option of using synthetic graft polymers that can be permanent or resorbable [40]. By utilizing resorbable polymers, in vivo ECM remodeling processes can replace synthetic materials over time. Ensuring resorbability and non-immunogenicity can be challenging, but one solution to the immune response is deriving cells from the patient themselves [41]. This could be done using autologous Schwann cells harvested from the patient or by utilizing induced pluripotent stem cells derived from the patient, which can then be induced into the desired cells to be seeded (i.e., Schwann cells) [42–44]. Such induced cells would retain the human leukocyte antigen (HLA) signature of the patient. Similar approaches already exist with Car T-cell therapy [45]. Allogeneic cells from immune-compatible, HLA-matched donors could also be used to seed nerve guidance conduits [46, 47]. Another approach to circumvent immunogenicity could be to develop a conduit substructure replete with the architecture and trophic factors that most efficiently hasten and support endogenous cell regeneration without additional cellular components [23, 38]. Some devices adopting this approach have used xenobiotic materials such as porcine or bovine collagen, as these materials have a history of use in organ repair [48].

Prospects for Improving Neuron Cell Regrowth

In developing a peripheral nerve graft, special attention should be paid to developing the correct balance of axonal support with growth factors, extracellular matrix, and cellular constituents that will maintain axon growth over long time periods [38]. Other organs such as the heart or kidney can restart proper function quickly after a transplant, but nerves require weeks, months, and even years to regrow to their target destination [49].

Materials and fabrication processes geared toward successful graft development should, therefore, generate a structure that is capable of supporting cell growth and nutrient diffusion for the entire duration of recovery, either by those materials remaining unresorbed by the local environment, or being replaced over time by endogenous ECM [50]. The architecture should provide sufficient boundaries within the graft such that axons and Schwann cells can be guided to their respective locations while not occluding the path of ingrowing axons [51]. Previous studies have shown that small variations in the diameter of synthetic endoneurium channels can greatly affect Schwann cell ingress and axonal regeneration. Liu et al. demonstrated that 50-μm-diameter channels supported axon regrowth far better than larger channels over 150 μm in diameter [52]. Thus, it appears that finely tuned parameters of nerve architecture and the complexity and adherence to typical nerve dimensions may be needed to produce a successful synthetic graft. We can see the dual considerations of developing a graft, from the standpoint of nerve guidance: The graft should be of sufficient microstructure to guide incoming axons into fascicular channels, but the graft “fill” or luminal content must not be too dense as to block the growth of nerves. Additionally, the graft must prevent intrusion of unwanted cells and cytokines and promote diffusion of nutrients [53–55].

Biological and biochemical components such as cells and trophic factors that are added to a graft should also continue to function throughout this long duration of recovery [56–59]. Microspheres containing growth factors are one option for long-term release of growth factors [59–61]. Combined with semi-permeable synthetic biomaterials throughout the graft, such time-release factors could strike the appropriate balance between graft self-sustaining nutrients, and diffusion of endogenous biochemicals [62].

Current Technologies in Biomaterials and Fabrication

Much attention must also be given to the choice of biomaterials and fabrication process when generating cellular scaffolds, especially for nerve guidance conduits (NGCs) [6, 63, 64]. First and foremost, the materials must be nontoxic and non-immunogenetic (biocompatible). Optimally, the materials in the scaffold should be resorbable within a year to allow, ultimately, for the complete replacement of the scaffold with endogenous ECM and cellular support [24]. In addition, the chosen material must be adequately assessed for its processibility, mechanical properties (strength, flexibility, suturability, and kink resistance) and ensure the safety of degradative byproducts such that they do not cause inflammation and obstruct nerve regeneration. Finally, the materials chosen to fabricate the NGC, especially for the matrix that fills the lumen, should be tunable to patient specifications [65–68].

Natural materials appear as the ideal choice because they provide a cell-friendly environment. These are either polypeptides such as collagen, gelatin, keratin, and silk, or polysaccharides such as chitin, chitosan, hyaluronic acid, and cellulose [6, 69]. However, NGCs are difficult to fabricate from natural materials. Many natural materials are insoluble and thus are not easily amenable to processing by methods such as solvent casting, dip-coating, porogen leaching, phase separation, freeze drying, combinations of these methods, or most importantly by electrospinning [6]. Commonly used natural materials are also not thermally stable and thus cannot be processed by methods such as extrusion and injection molding, and by 3D printing [46]. Even if these obstacles are overcome by blending (e.g., silk with poly(ethylene glycol), PEG [70], or poly(lactic-co-glycolic acid), PLGA) [71], or by chemical modification [72], the resulting NGCs may not have the desirable mechanical characteristics. For these reasons, synthetic materials are being sought.

The most common synthetic polymers that have been used to fabricate NGCs are polycaprolactone (PCL), poly(L lactic acid), PLLA, PLGA, poly(hydroxy butyrate), and PHB [6]. These polymers are easily processible by a variety of solvent- and thermal-based methods. While they result in NGCs that are mechanically strong, they can be toxic to cells due to the presence of chemical species such as initiators, monomers, and residual solvents [24, 71]. This issue can be addressed by careful analytical evaluation of material and the device. Furthermore, some of these polymers, e.g., PCL, are not degradable over the desired time (< 1 year). Partially degraded crystalline fragments of polymers such as PLLA, as well as the acidic byproducts of degradation of polymers such as PGA and PLA, might cause inflammation. Thus, there is ongoing research to produce a suitable degradable processible polymer that is as benign as natural materials. Blends of synthetic and natural polymers have also been proposed to take advantage of cell-nurturing capacity of natural materials [73]. Conductive polymers are desirable candidates [74], but they are not degradable.

Fabrication methods can be broadly classified as those that use solvents and those that use heat [6]. NGCs can be made from films cast from solution by rolling the flat films into a tube, and thermally sealing the seams, or by dip coating [75, 76]. The walls of the tube can be made porous by salt-leaching and gas foaming [75–77]. A combination of phase separation-salt leaching-freeze drying of the solution in a mold has also been attempted. Of all the solvent-based methods, electrospinning is the most versatile; electrospun tubes are mechanically soft and pliable, and suitable for use as NGCs, while other methods yield mechanically unsatisfactory conduits, some of them being brittle [75, 76]. NGCs can also be made by thermal processes such as compression molding to produce films that can be rolled into a tube, injection molding, extrusion in which the fibers are braided or woven into tubes [78, 79], and by 3D printing [46, 80]. Of these thermal processing methods, the braided extruded fibers are the most versatile because of their flexibility, kink-resistance, and ease of processing [80].

Electrospinning has the added advantage that it can be used to produce oriented surfaces that can facilitate directed growth of axons [77]. In Fig. 4, biomaterial fibers are shown in both unoriented (a) and oriented (b) arrangements, which can be achieved by varying electrical field distribution and spin-speed of the rotating mandrel, among other parameters. Controlling fiber orientation within biomaterial constructs enables researchers to tune porosity and density, thereby selecting the materials that can and cannot traverse the biomaterial barrier [77]. Oriented growth of neurons can also be promoted by creating channels in the hydrogel that fills the lumen [81]. These channels can be produced using a sacrificial material such as sucrose [82], or by directional freezing of water [83, 84]. Furthermore, electrospinning can be used to produce scaffolds with bi-layered structured walls in which the longitudinally oriented fibers on the inside can promote directed neuronal growth and the unoriented fibers on the outside can be used to grow fibroblasts [39, 85].

Fig. 4.

E1001K Bioriented scaffold electro-spun from a tyrosine-derived polymer, E1001(1k), at high speed (2000 rpm) to obtain oriented fibers (a) and then at low speed (200 rpm) to obtain unoriented fiber (b) (unpublished)

Considering endoneurium alignment, researchers should strive to create a substructure which approximates the most general channel and fascicle size and distribution for most nerves [38, 86]. By generating a graft that has a homogeneous distribution of endoneurium channels, in-growing axons may be able to self-sort and recreate their appropriate fascicular arrangements. This will require an understanding of which size and what density of channels best supports axon growth and arrangement [86].

Advances in imaging technology can also be crucial in developing implantable technologies [16]. MRI mapping of nerves at the level of the individual nerve fascicles can provide a template for graft dimensions. Once imaging is performed, 3D printing or similar methods can generate a graft to patient specifications. Presumably, by aligning fascicular bundles and matching the dimensions of the injured nerve, in-growing axons will encounter less mechanical resistance to regeneration while still receiving in situ contact direction [86].

Growth Factors and Cellular Constituents that Allow Schwann Cells, Neurons, and Other Cells to Grow

All cell growth occurs within and is influenced by the local microenvironment. Conditions such as pH, temperature, pressure, oxygenation, osmolarity, and the presence of serum or growth factors all affect the ability of cells to achieve optimal growth, health, and function [87–89]. A fibroblast, the primary ECM-secreting cell of the human body, can behave quite differently in the context of the peripheral nerve compared to the skin [90–93]. These differences are underwritten by the variable expression of genes and production of proteins at these various sites.

A major obstacle to functionalizing synthetic grafts is how cells and growth factors can be localized within the interior. While it is possible to generate channels within the lumen of the graft, getting cells within these spaces is more challenging. One approach is to create channels and then deposit cells via injection or diffusion into the channels [38, 86]. However, the size of cells and their projections may limit cell mobility into narrow channels. If dispersing cells within a pre-constructed structure were possible, then one option would be to seed the allograft with cells and growth factors. Similarly, large proteinaceous molecules such as growth factors may not be able to diffuse ubiquitously throughout the graft [38].

As noted previously, endoneurium channels accommodate axons from 1 to 20 μm in diameter. Schwann cells may have a cell body that is able to fit within this diameter, but Schwann cells have projections that can prevent their movement into channels. A Schwann cell may span 40–50 μm or more in its entirety and thus may prevent neurons from traveling into the pre-formed channels within the graft lumen. Indeed, it is these very projections that myelinate axons, and it is important that they be oriented longitudinally within the graft, as shown in Fig. 5, rather than at random angles [94].

Fig. 5.

Primary Rat Schwann cells. A representative image of rat Schwann cells with highly aligned projections grown in vitro (unpublished)

Alternatively, instead of generating pre-formed channels within the graft lumen, Schwann cells could be seeded onto sacrificial fibers which degrade over a period of days to weeks [95, 96]. Once cells are grown onto such fibers, the fibers can then be immersed within a hydrogel such that the cells are already present within the channels. Several studies have shown that it is possible to grow Schwann cells within hydrogels [97–99], and this approach of using sacrificial fibers may allow scientists to generate a substructure wherein Schwann cells can be incorporated into a dimensionally tuned microstructure [100]. Once Schwann cells are deposited and the sacrificial fibers resorbed, they can survive within the channels and be present at the time of engrafting. Then, axons can travel through the channels with Schwann cells already in place.

Another advantage to the sacrificial fiber method is that nanoparticles containing growth factors could be chemically linked to these fibers prior to channel formation. The presence of growth factors on the sacrificial fibers could sustain Schwann cell growth while developing the graft. A major issue with the incorporation of growth factors is their transient half-life and the need for more protracted growth factor delivery. Nanoparticles offer one solution to this problem and have been shown to improve drug delivery in a variety of circumstances [8, 101].

In pursuing cellularization, the questions being addressed are what combination of these attributes best support nerve regeneration? Are cells seeded prior to implantation necessary and at what density? Prior studies have attempted to seed cells at varying densities with some trials seeding Schwann cells at a density of 10 million cells/cm³, and others seeding up to 80 million cells/cm³ [100]. McGrath et al. utilized a “seed and implant” method whereby Schwann cells are rapidly mixed into the graft lumen and then implanted. This means, however, that cells may not be adequately stabilized within the lumen of the graft prior to implantation. If cells are not growing stably, they can pose a risk of apoptosis and generate inflammatory mediators which may be counteractive to graft function. While this study showed that nerve function recovery was robust in the short-term, long-term recovery could not be sustained.

Proposed Future Directions

Considering the developments in synthetic nerve graft technology to date as discussed in “Current technologies in biomaterials and fabrication” section and described in detail in reviews cited there, two primary requirements arise. First is the need to fabricate a nerve architecture which replicates or approximates in vivo human nerves. Previous attempts have not successfully replicated the complexity of nerve architecture with the complete palette of endoneurial channels, extracellular matrix, growth factors, and vascular components inherent to authentic nerves [4, 17, 36, 38]. Theoretically, a model graft would embody all or most of these architectural features.

Second, the graft should incorporate cellular components such as Schwann cells and fibroblasts needed to secrete growth factors and chemo-attractive agents, produce extracellular matrix, and myelinate in-growing axons [49, 102]. To achieve the appropriate cellularity, two approaches could be utilized: one, seed a graft with these cells prior to implantation, either from autologous sources, stem cell sources, or induced cell sources [103–105]; or two, infuse the graft with growth factors that can hasten the migration of cells into the graft once implanted [106–108].

A pre-cellularized graft presents several advantages, namely the ability to control cell distribution within the graft before implantation [109–111]. Localizing cells within the graft prior to implantation may help prevent undesired cell infiltration by creating a stable microenvironment within the graft [112–114]. However, cells having an HLA signature identical to, or very close to, the patient are critical to prevent graft rejection. Thus, developing a graft without cells, but with the necessary trophic factors to stimulate cell migration once implanted, could provide a viable off-the-shelf option for clinicians.

Generally speaking, NGCs can be classified into three categories; (1) bare biomaterial conduits; (2) constructs with growth factors; and (3) constructs with growth factors and living cells. Nerve regeneration is most effective with growth factors and living cells. However, consideration of fabrication, stability, storage, of the NGC as a clinical device makes NGC without any biologics most expedient to produce and store. Surgeons and hospitals may choose the type of NGC by balancing the cost and benefit of the device. These considerations will also drive the future development of both clinically and cost-effective NGC devices.

Structures with Multiple Architecture and Materials

NGCs are increasingly being considered as viable treatment for the repair of injured or diseased nerves [4, 38, 115, 116]. As mentioned earlier, a large variety of materials and fabrication techniques have been used, either alone or in combination (e.g., copolymers) to produce functioning NGCs that are able to regenerate nerves with varying degrees of success. But most of these are simple tubular structures that fail to mimic the architecture of the biological nerves that can guide the Schwann cell migration and axonal growth [117]. The expectation is that the NGC-regenerated nerves reproduce the functions, if not the architecture, of the nerves that are being repaired. It is difficult to achieve these with a single material and a single architecture. It is apparent that for an NGC to reproduce the function of a biological nerve, it should be a composite structure with multiple architectures, each of them tailored so that the local structure and the material will support the cell type specific to that part of the nerve as it exists in biological structures (Fig. 1). The walls of the conduit need to have different architectures on the inside and the outside to host Schwann cells in the inner walls, and fibroblasts in the outer walls of the structure [82, 118–121]. There also needs to be a scaffolded architecture to promote the vascularization of the nerves support adequate deposition of ECM [122–125]. Bilayer scaffolds with axial orientations on the inner surface, and random orientations on the outside layer are now being produced for different applications including NGCs [39, 85, 126]. Oriented pathways are required in the soft gel-like matrix of the lumen to promote growth and alignment of axons [127]. Multiplicity of such channels could be used to direct motor and sensory nerves to their proper targets. As discussed in the “Current technologies in biomaterials and fabrication” section, there have been attempts to replicate the nerve microarchitecture via channels through a solid lumen or with fibers to mimic substrates onto which Schwann cells may grow [4]. Some of these ideas are illustrated in Fig. 6.

Fig. 6.

Schematic of various bio-engineered, functional conduits. Reproduced with the permission of the Authors from Carvalho et al. [4]

While developing materials that are tailored to match the desired degradation characteristics for nerve regeneration, it is also necessary to understand in detail the effects of porosity in the walls of the conduit on the survival of the cells in the lumen, of the surface topography in the form of grooved walls on cell attachment and axonal growth, and of the modulus of the matrix on cell proliferation. A combination of braiding and electrospinning or braiding and solvent casting are some of combination that can be used. To provide guidance to neurons in the lumen, it is possible to use either channels or guide fibers [4]. Hydrogel matrices continue to be used to fill the lumen [4]. 3D bioprinting, which is now poised to fabricate patient-specific NGCs, requires matching the competing requirements of 3D printing equipment and the material attributes of a functioning NGCs with supporting cells or stem cells [4, 46]. Another area that needs to be addressed is the incorporation of therapeutic drugs into scaffolds for controlled delivery [8, 37, 70]. Potential drugs, such as chondroitinase ABC and antibodies against Nogo, have shown promise in reducing the inhibitory environment created in the central nervous system after injury [8]. Nogo-A and other inhibitory myelin-associated molecules bind to the Nogo-66 receptor, activating a signaling cascade that limits regeneration [8]. Finally, electrical stimulation has shown to be effective in increasing the rate of regeneration [128], but a biocompatible and resorbable material and viable stimulating mechanism still needs to be developed.

Advancements in Stem Cell Technology and Nerve Repair

Induced pluripotent stem cells (iPSCs) present an unparalleled opportunity for regenerative medicine [129]. iPSCs do not suffer from the same ethical implications as embryonic stem cells and can be readily derived from any human patient by transcription factor mediated reprogramming [130]. Pluripotent stem cells have the potential to differentiate into any cell type from the three germ layers and can self-renew in perpetuity given the correct environment, allowing for unlimited cell generation [129]. These properties provide iPSC-derived neural cells with a great advantage over other autologous cells that require invasive surgery to obtain. For instance, iPS cells have now been generated from blood, skin, and even cells within urine [131, 132]. While many patient-specific applications might require autologous iPS cells, there is potential for allogeneic iPS cell repositories that can later be matched to patient HLA subtypes to reduce immunological response [133]. Since induced pluripotent stem cell technology first emerged in 2006, the field has grown to encompass differentiation and reprogramming techniques for a plethora of cell types relevant to peripheral nerve regeneration including neural crest cells, Schwann cells, and peripheral neurons [134].

Neural crest stem cells (NCSC) are an early cell population that gives rise to diverse ectodermal cell types including melanocytes, Schwann cells, peripheral neurons, smooth muscle, bone, and fat [135]. Many protocols have been developed to differentiate NCSC from iPSC [136–138]. NCSC are positive for markers AP2, p75, PAX3, and SOX10 [139]. Several groups have posited that NCSC transplantation could aid nerve regeneration. For example, in a rat sciatic nerve model, Lv et al. found that NCSC promoted regeneration and reconstruction of a transected sciatic nerve by seeding NCSC in a PLLA electrospun conduit [140]. More recently, NCSC-laden nerve scaffolds were shown to attenuate nerve sciatic transection pain, improve motor function recovery, and protect spinal cord from glial activation [141]. An important consideration is whether transplanted cells have the potential to elicit inflammatory host immune responses. To address this issue, several groups have explored the ability of NCSC to stimulate T cells or induce cytokine production. Importantly, iPSC-NCSCs were shown to have negligible immunogenicity offering great potential for clinical application [142, 143].

Schwann cells are a collection of three different cell states encompassing Schwann Cell Precursor cells (SCPs), non-myelinating Schwann cells (Remak cells), and myelinating Schwann cells, as shown in Fig. 7. SCPs derive from NCSC and therefore share similarities in their differentiation protocols. The main difference being that most protocols include Neuregulin-1 (NRG-1) after the initial neural crest induction to guide Schwann cell lineage specification. There is debate about whether SCP are just neural crest cells in disguise as it has been shown that SCP give rise not only to Schwann cells but also parasympathetic and enteric neurons, melanocytes, and endoneurial fibroblasts [144]. To date, several protocols have been published for iPSC-derived SCs using small molecule/growth factor mediated differentiation [136, 145]. More recently, direct reprogramming strategies have been employed using ectopic expression of transcription factors SOX10 and EGR2 [146] or SOX10 alone in combination with small molecule WNT activation to directly convert human fibroblasts into SCPs [147]. Because human neural development is a protracted process, these direct reprogramming strategies have the advantage of shorter differentiation timelines making them potentially more applicable for clinical therapy. In addition to pluripotent stem cell-derived Schwann cells, Schwann-like cells have also been derived from mesenchymal stem cells (MSCs) [148, 149], adipose-derived stem cells (ADSCs) [150, 151], and bone marrow stromal cells (BMSCs) [152, 153].

Fig. 7.

Schematic of growth factors associated with Schwann cell response to axonal myelination and injury. During development and growth, Schwann cells form either myelinating or “bundling” (Remak) phenotypes. In response to injury, Schwann cells facilitate axonal and myelin debris degradation and removal. Specific growth factors participate in these processes, as shown above. Such growth factors may be critical for incorporation into nerve conduits. Reproduced from Salzer et al. [161]; this work is licensed under the Creative Commons Attribution-NonCommerical-ShareAlike 3.0 Unported Licence (http://creativecommons.org/licences/by-nc-sa/3.0/)

While the focus of neural differentiation strategies has mostly focused on CNS cell types, several groups have explored the possibility of creating human somatic cells or iPSC-derived peripheral neurons. Peripheral neurons are subclassed into motor neurons (efferent) and sensory neurons (afferent). Several protocols have been established for the efficient generation of motor neurons [154, 155] and sensory neurons [156–160]. Specifically for Schwann cells, key trophic modulators determine the phenotypic ability to myelinate or encompass axons. Figure 7 suggests a pathway by which immature Schwann cells respond to trophic factors to adopt roles such as myelination and response to injury [161].

Regarding the long-term clinical utility of nerve repair grafts, the following considerations are paramount. First, there must be long-term evaluation of the biomaterial’s biodegradation and resorption properties. In line with this characterization, it is vital to understand from the beginning the host response to the biomaterial, the survivability of engrafted cells, and their ability to remodel into the target tissue. To understand the host response, immunohistochemical characterization of macrophage states in proximity to the device is necessary, in order to determine whether any giant cells are present which indicates immunological activation. This is especially important because it has been found that macrophage activation can recruit and cause proliferation of myofibroblasts and cause conversion of fibroblasts to myofibroblasts [162]. However, myofibroblasts have contractile tendencies which can put strain on the nerve and cause severe pain for patients [163].

Prior studies have identified potential cell sources for the purposes of PNI repair. The use of iPSCs addresses two important issues: (1) induced cells maintain the DNA signature of the patient, reducing the risk of rejection by their immune system and (2) these methods circumvent the need for embryonic stem cells, which carries its own set of ethical and safety concerns.

Conclusions

The field of peripheral nerve regeneration is an area of active research and development. Currently, interventions such as direct nerve suturing, autograft, and allograft offer options for improving nerve regrowth, but they do not adequately restore motor and sensory function in most patients. Nerve guidance conduits (NGCs), typically made of synthetic biomaterials, are also used in some contexts but lack the efficacy of autografts and allografts, particularly for large nerve gap transections (> 1.0 cm).

Despite their present limitations, however, NGCs offer several potential advantages over auto- and allografts, namely the ability to custom-tune the architecture, cellularity, and growth factor composition of the conduit to meet patient specifications. To achieve these desired characteristics, the synthetic hollow tubes that are now commonly available need to be modified, especially for the repair of larger nerve gaps, by using materials with degradation characteristics suited to the timeframe of nerve regeneration, and superior physical properties (strength, flexibility, and permeability). These scaffolds need to be fabricated and functionalized to enhance regeneration, for example by changing the tube’s microarchitecture with internal filaments, filling the lumen with fibrous proteins such as laminin and collagen to support Schwann cells, and incorporating growth factors. These will ultimately enhance nerve regeneration technology and promote NGCs from narrow use to broader clinical application.

Previous studies in animal models have demonstrated that biomaterial constructs combined with cells, growth factors, and luminal fillers can augment axonal regeneration, but this regeneration and subsequent maintenance of nerve function is difficult to maintain [25]. Building on these results is a promising frontier in peripheral nerve repair. In tandem with improved engineering and functionalization of conduits and grafts, biochemical, genetic, and histopathological studies are necessary to gain further understanding of the exact mechanisms of nerve regeneration and to quantify the points and ways in which graft failure occurs. Increased understanding of the molecular biology of neural development and regeneration, along with advances in the areas of drug delivery, gene therapy, and biomaterials will increase the efficacy of these devices. While some may favor a solely cell-based strategy and others a strictly biomaterials solution, it is likely the combination of these two that will yield the most efficacious treatment for patients in the long run.