Cross section of polymer conduits (original magnification ×30). Top, Single-lumen conduit (1-mm scale). Bottom, Five-lumen conduit.
Electron micrographs of polymer conduit (original magnification ×790). Top, Junction of 3 lumina. Bottom, View inside 1 lumen. Note smooth, nonporous surface.
Light micrographs of Schwann cells cultured onto polylactic–co-glycolic acid polymer films. Top, In a mitogenic medium, the cells are nearly confluent (original magnification ×50). Bottom, In an ordinary medium, the cells are adherent, extending processes (original magnification ×100).
Toluidine blue–stained cross sections of regenerated nerve cable (original magnification ×660). Top, Single-lumen conduit. Bottom, Autograft.
Graph illustrating the percentage of the cross section represented by neural tissue (axons and myelin sheaths) for the 3 single-lumen conduits (bars 1-3) and the 2 autografts (bars 4 and 5).
Hadlock T, Elisseeff J, Langer R, Vacanti J, Cheney M. A Tissue-Engineered Conduit for Peripheral Nerve Repair. Arch Otolaryngol Head Neck Surg. 1998;124(10):1081-1086. doi:10.1001/archotol.124.10.1081
Peripheral nerve repair using autograft material has several shortcomings, including donor site morbidity, inadequate return of function, and aberrant regeneration. Recently, peripheral nerve research has focused on the generation of synthetic nerve guidance conduits that might overcome these phenomena to improve regeneration. In our laboratory, we use the unique chemical and physical properties of synthetic polymers in conjunction with the biological properties of Schwann cells to create a superior prosthesis for the repair of multiply branched peripheral nerves, such as the facial nerve.
To create a polymeric facial nerve analog approximating the fascicular architecture of the extratemporal facial nerve, to introduce a population of Schwann cells into the analog, and to implant the prosthesis into an animal model for assessment of regeneration.
Tubes of poly-L-lactic acid (molecular weight, 100,000) or polylactic–co-glycolic acid copolymer were formed using a dip-molding technique. They were created containing 1, 2, 4, or 5 sublumina, or "fascicular analogs." Populations of Schwann cells were isolated, expanded in culture, and plated onto these polymer films, where they demonstrated excellent adherence to the polymer surfaces. Regeneration was demonstrated through several constructs.
A tubular nerve guidance conduit possessing the macroarchitecture of a polyfascicular peripheral nerve was created. The establishment of resident Schwann cells onto poly-L-lactic acid and polylactic–co-glycolic acid surfaces was demonstrated, and the feasibility of in vivo regeneration through the conduit was shown. It is hypothesized that these tissue-engineered devices, composed of widely used biocompatible, biodegradable polymer materials and adherent Schwann cells, will be useful in promoting both more robust and more precisely directed peripheral nerve regeneration.
PERIPHERAL NERVE injuries continue to be among the most challenging problems faced by surgeons. Within otolaryngology, it has been particularly difficult to develop appropriate strategies for the management of facial paralysis, recurrent laryngeal nerve injury, and the morbidity associated with insensate oropharyngeal and esophageal flap reconstructions after ablative surgery for malignancy. The development of a synthetic nerve guidance conduit would be of enormous utility in the management of these and other nerve-related problems in otolaryngology.
The standard approach to repair a peripheral nerve when a gap is present is to bridge the severed ends with a segment of autologous donor nerve. When the caliber of the donor autograft matches that of the recipient nerve, in both total diameter and average axon density, superior results are obtained.1,2 While this approach incurs some donor site morbidity and usually involves a secondary surgical site, its main drawback is insufficient functional outcome. In the case of the facial nerve, autografted repairs rarely, if ever, result in a recovery that is better than House-Brackmann grade III.3
Attention has long been directed at alternatives to autografting. Researchers in the past have attempted to bridge severed nerve endings with a wide variety of autologous biological tubular structures, including artery, vein, inside-out vein conduits, and decalcified bone channels.4- 8 Some have used cadaveric nerve allografts or xenografts to bridge long nerve defects.9 Others have used skeletal muscle grafts for this purpose.10- 12 While investigators have achieved promising results using many of these techniques, to our knowledge none has matched or surpassed those achieved by autograft repair.
Entubulation repair, by which the proximal and distal ends of the severed nerve are inserted into either end of a hollow conduit, has been attempted extensively.13- 19 It is hypothesized that isolation of the local regenerative milieu may encourage more focused regeneration toward the distal stump. Mechanisms for this increased regeneration include decreased access of inflammatory cells to the stumps and decreased diffusion of distally elaborated neurotrophins into surrounding tissues, which results in increased local concentrations of these substances at the proximal stump, where they exert their effects. This entubulation model has helped to delineate the molecular and cellular events that occur in the environment of the injured and regenerating peripheral nerve.
Studies of regenerating peripheral nerves have uncovered a variety of influences that affect the behavior of individual nerve growth cones at the proximal stump. For example, it was discovered that the presence of electrical field stimulation promotes robust neurite extension20 and that certain neurotropic and neurotrophic substances are secreted by the cut nerve ends after injury.21,22 A very large number of seemingly unrelated molecules and peptides also possess neurotrophic properties or promote axonal elongation after injury, including extracellular matrix components, testosterone, vasoactive intestinal peptide, and the immunosuppressive agent tacrolimus, among others.23- 28
Recently, focus has been directed toward the development of nerve guidance conduits made from synthetic biomaterials. Investigators now hope to use the controllable properties of synthetic materials to create regenerative environments that provide known neurotrophic influences, including electroconductivity and the ability to release nerve growth–promoting substances over time. It is also feasible to manipulate conduit architecture, surface properties, porosity, and biodegradability to optimize the regenerative conditions.
Entubulation repair with polymeric conduits possessing electroconductive properties has demonstrated increased regeneration over neutral controls. Polymeric channels engineered to slowly release growth factors from their walls have also proven beneficial. The addition of Schwann cells to the lumen of a hollow guidance channel has provided a benefit,29,30 based on the fact that Schwann cells express both surface proteins and elaborate soluble factors conducive to the extension of the regenerating axon tips. The beneficial presence of Schwann cells was found to be concentration dependent, with tubes loaded with higher numbers of cells leading to increased axonal bridging through conduits. Certain lumen-occupying substances, including extracellular matrix components, soluble neurotrophins, or combinations of these, have also proven useful to the bridging of neuronal gaps. However, despite findings that convincingly demonstrate the benefit of these influences that are able to be introduced through biomaterials, no hollow conduit model been able to bridge a neural gap more effectively than control autografts.
In this study, we used a polymer dip-molding technique to create a biocompatible, biodegradable polymer conduit whose ability both to release growth-promoting substances over time and to harbor Schwann cells would exceed that of simple hollow conduits. This was accomplished by creating several thin-walled compartments within the conduit. This increase in surface area will ultimately allow more extensive delivery of neurotrophins throughout the conduit and may allow the selective delivery of different neurotrophins into different regions of the conduit. The larger surface area provided by the introduction of subluminal architecture also significantly increases the surface area available for Schwann cell adherence.
We also studied the adherence of both neonatal and adult Schwann cells to poly-L-lactic acid (PLLA) and polylactic–co-glycolic acid (PLGA) surfaces to discern whether the surface environment provided by conduits of these materials would be favorable to inhabitance by Schwann cells. While neonatal cell populations expand more rapidly and give higher yields, it is important for tissue-engineering applications to demonstrate the feasibility of cell harvest from adult tissues, as implant devices would have to be developed from biopsy specimens. We also explored in vivo regeneration through the conduits over a long gap compared with autografts.
Poly-L-lactic acid with a molecular weight (MW) of 100,000 and a copolymer of PLGA with a lactide-glycolide ratio of 85:15 were purchased commercially (Polysciences, Worthington, Pa). Stock solutions of 2.5% to 5% (wt/vol) of the polymers were generated by dissolution in either chloroform (CHCl3) or methylene chloride (CH2Cl2). Tubes were prepared by a modification of a standard dip-molding technique. Briefly, for the single lumen conduits, a thin-walled polytef sheath (outer diameter, 1.8 mm) was passed over a glass capillary tube. Inside the fume hood, the tube was secured to a dipping apparatus and systematically dipped into the polymer suspension every 15 seconds, for a total of 20 dips. The tubes were placed into and removed from the solution at a constant rate and were suspended for a total of 10 seconds between each subsequent dip for solvent evaporation. After the dipping procedure, the tubes remained in the fume hood overnight for evaporation of solvent at atmospheric pressure, followed by an additional 24-hour period under low vacuum for the removal of residual solvent, and then stored in a dessicator for later characterization.
A similar approach was taken for the creation of tubes with multiple sublumina. Smaller polytef sheaths (OD, 600 µm) were passed over stainless steel wires rather than glass capillary tubes. Each sublumen was made separately, as described above, and then bundled together for creation of the final multiluminal channel. For example, to create a 5-lumen conduit, 5 tubes were prepared separately. The group of tubes was then bound together with a suture at either end and dipped as a bundle for a total of 20 additional dips to achieve a single prosthesis with 5 separate compartments. The bundled conduits were redried and stored as described above.
Polymer films for the plating of Schwann cells were prepared from the same polymer stock solutions as the conduits. They were formed by a standard solvent casting technique. Several milliliters of the polymer suspension was placed into a 100-mm glass Petri dish in the fume hood, and the solvent was allowed to slowly evaporate. The resultant films were gas sterilized with ethylene oxide and used for the plating of Schwann cells.
The tubes were examined grossly for mechanical strength and pliability. Scanning electron microscopy using an environmental scanning electron microscope (Jeol USA Inc, Peabody, Mass) was used to examine luminal surfaces and porosity. Gel permeation chromatography using a peristaltic pump (Hewlett-Packard Co, Palo Alto, Calif) and commercially purchased columns (Polymers Inc, Burlington, Vt) was carried out to determine the MW of the processed polymer.
Schwann cells from both neonatal and adult Lewis rats were isolated and expanded in culture using previously described methods.31 Briefly, for neonatal cell harvests, 2 litters of day 2 Lewis pups were killed, and their sciatic nerves were removed. After enzymatic digestion with collagenase and dispase, the crude cell suspension was plated in Dulbecco modified Eagle medium, enriched with 10% heat-inactivated fetal bovine serum and antibiotics. The cells were treated for the ensuing 2 to 3 days with cytosine arabinoside for the elimination of fibroblasts. On day 5, a thy1.1 antibody purification was used to remove residual fibroblasts, and the purified Schwann cell population was expanded on polylysine-coated dishes over a period of 6 weeks using mitogenic media enriched with a cyclic AMP activator (Forskolin, Sigma Chemical Co, St Louis, Mo) and glial growth factor (Biomedical Technologies, Wilmington, Mass). Cultures were assessed for purity using an immunocytochemical stain against S100 protein, a glial-specific marker.
For adult Schwann cell isolation a previously described explant technique was used.32 Lewis rats (weighing 200-250 g) were killed, and the sciatic were nerves removed. The epineurial tissue was microscopically peeled away, and the remaining tissue was cut into 1-mm segments. These segments were placed on tissue culture plastic in Dulbecco modified Eagle medium and explanted on a weekly basis. After 6 weeks, the explants were enzymatically digested, purified, and expanded as described above. Yields from these explants were much lower, as expected, and Schwann cells for the majority of the ensuing studies were performed using the neonatal preparations.
For the adherence studies, polymer films prepared and sterilized as described above were placed onto tissue culture plates. A Schwann cell suspension was then plated onto the films, and the cultures were examined using light microscopy on a daily basis for observations of adherence, survival, and proliferation behavior.
The polymer conduits, with or without inhabitant syngeneic Schwann cell populations, were implanted into 20-mm gaps in the left sciatic nerve of 10 adult Lewis rats. After induction of anesthesia with inhalational methoxyfluorane, the animal was shaved and prepped. An incision was made over the left hind limb, and the gluteal muscles were divided to expose the sciatic nerve. Under the operating microscope, a 20-mm segment of nerve was removed, and the proximal and distal stumps were secured to either end of the polymer conduit using several 9-0 nylon microsutures. The muscle and skin were closed in layers, and the animals recovered in a controlled environment. They were given food and water ad libitum and examined for signs of autotomy. Institutional guidelines regarding animal experimentation were followed. After 3 months, the animals were evaluated for functional sensory recovery using standard pinch testing and perfused with 3% glutaraldehyde, and the implants were harvested. Observations were made regarding tubular continuity both along the length and at the anastomoses. The specimens were plastic embedded, cut into 1-µm-thick sections, and stained with toluidine blue for assessment of both the axonal contents of the conduit and the degree of surrounding inflammation. For comparative purposes, 2 control animals underwent 20-mm autografts rather than tubular implants. Surgical and harvest techniques were identical to those described above, except that when the 20-mm segment of sciatic nerve was removed, it was rotated 180° and sewn in as a reversed autograft.
Both PLLA and PLGA conduits were created containing 1 to 5 sublumina (Figure 1). All conduits displayed favorable gross pliability characteristics. They could be deformed approximately 30° without kinking or fracturing. We found that the materials were soft enough to pass 9-0 nylon suture through readily, but were thick enough that the sutures held without tearing.
Electron micrographs demonstrated smooth-walled, nonporous tubular structures, as expected from solvent casting preparation techniques (Figure 2). Wall thickness ranged from 50 to 100 µm. Internal lumen diameters correlated precisely with the outer diameter of the polytef sheaths used in the conduit preparation, suggesting that no shrinkage or deformation of the tubular conduits took place during drying. Molecular weight determinations using gel permeation chromatography demonstrated MWs close to those of the starting materials.
This similarity suggests that no significant material breakdown occurs during the polymer processing technique.
Both neonatal and adult Lewis rat Schwann cell cultures appeared to contain more than 90% Schwann cells after the purification steps. The cells adhered equally well to films of either pure PLLA or the PLGA copolymer. When plated in mitogenic media, the cells divided to reach confluence over the same time course as similar cells plated onto polylysine-coated tissue culture plastic. When plated in ordinary media, the cells extended processes and were morphologically indistinguishable from cells plated onto control dishes of polylysine-coated tissue culture plastic (Figure 3).
Implantation studies revealed that 1 of the tubular conduits had become dislodged at the proximal anastomosis, and 2 of the single-lumen channels had fractured at the junction of the proximal one third and the distal two thirds. All other conduits remained intact, as did the control autografts.
Histologic evaluation revealed minimal inflammatory reaction in all specimens. A thin fibrotic band was seen to surround the outer surface of the tubes. Cross sections of the conduits, at a point 5 mm from the proximal anastomosis, revealed the presence of a regenerating axon cable in all the intact single-lumen channels, as well as the autografts (Figure 4). Figure 5 shows the percentage of neural tissue in these regenerating cables. The multiluminal conduits demonstrated sparse islands of regenerating axons, although there were too few to quantify. Sensory testing at 12 weeks revealed withdrawal to nociceptive stimulus at the fifth digit of all experimental animals, but withdrawal to nociceptive stimulus of the third digit occurred in only 1 animal in the autograft group.
Synthetic polymer conduits are a promising approach to artificial nerve guide development. They possess many properties that can be finely manipulated in order to effect axonal elongation. It has been well demonstrated that a number of growth factors and other small molecules promote neurite extension, and polymer technology has been shown to be capable of delivering a number of these substances effectively.
The current model for delivery of neurotrophins involves an entubulation of the nerve into a single-lumen conduit. Growth-promoting substances are then administered in 1 of 3 ways. Either they are loaded into the walls of the hollow conduit, so that they are slowly released in a circumferential fashion, or they are loaded into an aqueous medium and delivered into the lumen of the conduit at the time of surgery. Alternatively, they have been delivered systemically or regionally by injection.
We have designed a fully biodegradable polymer conduit whose multiple-channel architecture is conducive to a more effective local slow release of neurotrophins. There is enormous benefit to the use of copolymers of lactic and glycolic acid, because the rate of biodegradation can be tightly controlled by altering the ratio of the 2 monomers. Degradation, in turn, controls the rate of release of growth-promoting substances, as they are released primarily in association with degradation of the polymer walls.
A second advantage to this polyluminal conduit is that it provides increased area for Schwann cell adherence. While several studies have introduced Schwann cells into the conduit prior to implantation, the surface area for Schwann cell adherence through these hollow conduits has been a limiting factor. We have demonstrated in this study that Schwann cells adhere, survive, and can be directed to divide on polymer surfaces of PLLA and PLGA. In our 5-lumen conduits, the surface area for Schwann cell adherence rises from 1.1 to 1.9 cm2, a factor of nearly 2. By decreasing the internal diameter of each of the sublumina and increasing their number, it is feasible to increase surface area for cellular adherence by as much as an order of magnitude.
The in vivo implantation study demonstrated that the polymeric substances were well tolerated in the rat hind limb. The tubes experienced some fracturing and anastomotic disruption, suggesting that the 100-µm-thick walls may be unacceptably thin for these long peripheral nerve gaps. The presence of a neural regenerate in many of the conduits is a starting point from which we may make improvement with the addition of neurotrophic substances. It is possible that the lack of contiguous neural cables through the multiluminal conduits stems from the lack of porosity or the higher residual solvent presence in these conduits compared with the single-lumen conduits. Both these issues are currently being addressed in our laboratory. We have begun to use highly porous tubular structures and have eliminated chlorinated hydrocarbons from our preparation technique. We have also developed an extensive battery of behavioral tests that will offer more information regarding the functional recovery we seek to achieve. We recognize the limitations of the rat sciatic nerve model, in which neural regeneration occurs more robustly than in higher species, and expect to move to more analogous mammalian species once the architecture and properties of the prosthesis have been optimized.
We hypothesize that these polyluminal structures will have more thorough neurotrophin release profiles and higher inhabitant Schwann cell numbers, and will ultimately lead to the bridging of longer gaps than currently possible using existent artificial nerve conduits.
Accepted for publication June 23, 1998.
The authors wish to thank Daniel A. Hunter, RT, for preparation of the histologic specimens.
Reprints: Tessa Hadlock, MD, Department of Otolaryngology, Massachusetts Eye and Ear Infirmary, 243 Charles St, Boston, MA 02114.
This technique appears very promising but certainly is no better than cable nerve autographs. With further study in primates, it may well fit into our surgical armamentarium in the future.—Harold C. Pillsbury III, MDChapel Hill, NC