|Home | About | Journals | Submit | Contact Us | Français|
Bridges for treatment of the injured spinal cord must stabilize the injury site to prevent secondary damage and create a permissive environment that promotes regeneration. The host response to the bridge is central to creating a permissive environment, as the cell types that respond to the injury have the potential to secrete both stimulatory and inhibitory factors. We investigated multiple channel bridges for spinal cord regeneration and correlated the bridge structure to cell infiltration and axonal elongation. Poly(lactide-co-glycolide) bridges were fabricated by a gas foaming/particulate leaching process. Channels within the bridge had diameters of 150 or 250μm, and the main body of the bridge was highly porous with a controllable pore size. Upon implantation in a rat spinal cord hemisection site, cells infiltrated into the bridge pores and channels, with the pore size influencing the rate of infiltration. The pores had significant cell infiltration, including fibroblasts, macrophages, S-100β-positive cells, and endothelial cells. The channels of the bridge were completely infiltrated with cells, which had aligned axially, and consisted primarily of fibroblasts, S-100β-positive cells, and endothelial cells. Reactive astrocytes were observed primarily outside of the bridge, and staining for chondroitin sulfate proteoglycans was decreased in the region surrounding the bridge relative to studies without bridges. Neurofilament staining revealed a preferential growth of the neural fibers within the bridge channels relative to the pores. Multiple channel bridges capable of supporting cellular infiltration, creating a permissive environment, and directing the growth of neural fibers have potential for promoting and directing spinal cord regeneration.
Injury to the spinal cord typically results in permanent functional loss. The initial injury induces a cascade of events, such as inflammation and cyst formation, that are collectively referred to as secondary damage that can inhibit regeneration.1 Although central nervous system (CNS) neurons were first thought to be incapable of regeneration, pioneering studies have demonstrated their capability to regrow axons if given a suitable environment.2,3 Biomaterial scaffolds, termed bridges, have served a central role in many regeneration strategies, either implanted alone or seeded with cells to promote regeneration.4,5 To promote regeneration, the bridge must function as a physical support that apposes both sides of the injury (i.e., rostral and caudal) to minimize further damage to the cord. Additionally, the host response to the bridge should provide a permissive environment. Although bridges alone are likely insufficient to promote complete functional recovery,5,6 the host response to the bridge is a significant component of the local microenvironment that impacts regeneration. The influence of the bridge architecture on this host response has not been well characterized.
The host response at an injury mediates the regeneration process, as cells can secrete either stimulatory or inhibitory factors that determine the permissiveness of the environment. After spinal cord injury, a pronounced cellular inflammatory response is characterized by the accumulation of activated microglia and macrophages.7 Glial cells, such as the native oligodendrocytes, and Schwann cells from the peripheral nervous system are observed and can function to secrete stimulatory factors or remyelinate regenerating axons.8 Another type of resident glial cells, the astrocytes, can become activated by the injury, resulting in the secretion of proteoglycans, which contribute to a scar that stalls axon growth.9,10 Endothelial cell infiltration initiates revascularization of the injury site, which may limit cell necrosis and apoptosis.11
Bridges implanted at a spinal cord injury can influence the host response that determines the permissiveness of the environment. Bridge implantation induces a foreign-body response that involves the same cell types associated with the injury (i.e., macrophages, astrocytes, oligodendrocytes, and endothelial cells), and may also induce migration and accumulation of fibroblasts to the implantation site, though their role for spinal cord regeneration is unclear.12–14 The pore size of the material can influence the rate of cell infiltration, and also cellular responses.15 In the spinal cord, porous bridges support cellular growth into the bridge and provide nutrient exchange with the surrounding tissue. Bridges have been fabricated with porosities ranging from semipermeable to macroporous, which can manipulate the rate of cell infiltration, and the direction from which cells infiltrate.16,17 Initial designs of the bridge had a random structure that did not create a path across the injury site. More recently, bridges with multiple channels have been developed with the objective of maintaining a path for regrowing axons, which may also orient axonal elongation.6,18–22 Channels with diameters ranging from 40 to 600μm have been fabricated for spinal cord regeneration,18–20 though the correlation between channel diameter and regeneration has not previously been determined.
In this report, we investigate the host response to a porous multiple channel bridge, with channels of varying diameter, implanted in a rat spinal cord hemisection model. Multiple channel bridges with controlled pore size and channel diameter were fabricated by a gas foaming/particulate leaching process. The host response was characterized by the distribution and organization of cells within the bridge. Further, immunohistochemical staining was performed to identify the location and relative distribution of astrocytes, Schwann cells, fibroblasts, endothelial cells, and neurites. These studies correlate the bridge structure (e.g., porosity and channel diameter) with the cellular composition and distribution at the implant site, as this composition and distribution has not been well characterized as a function of the material design. Although bridges may not promote regeneration alone, they can function as vehicles for cell transplantation or drug delivery, and the response to the bridge can influence the function of transplanted cells or determine the targets for drug delivery.
Poly(D,L-lactide-co-glycolide) (PLG) with 75:25 molar ratio of lactide to glycolide was obtained from Boehringer Ingelheim Chemical (Resomer 755, i.v.=0.6–0.8, 80–120kDa; Petersburg, VA). Poly(vinyl alcohol) (88% hydrolyzed; average MW 22,000) was purchased from Acros Organics (Morris Plains, NJ). All other reagents were obtained from Fisher Scientific (Fairlawn, NJ) unless otherwise indicated.
Porous PLG bridges were fabricated by a gas foaming/particulate leaching process. PLG microspheres were fabricated by a single emulsion process as previously described.23 The polymer microspheres were mixed with salt particles with size ranges of either (i) less than 38μm or (ii) 63–106μm. The mixture was loaded stepwise into a custom-made mold, where the polymer/salt mixture was loaded into the mold just below a row of pins. Pins (150μm; Carolina Biological Supply, Burlington, NC; or 250μm, Fine Science Tools, Foster City, CA) were inserted across the mold through the pin guides, and then a polymer/salt mixture was added until the pins were covered. The mixture in the mold was manually packed by compression. Upon complete loading of the mixture, the unit was compression molded at 200psi for 15s and transferred to a pressure vessel to equilibrate with CO2 gas (800psi) for 16h. After quenching the pressure, the bridges were removed from the molds, immersed in sterile deionized water for 1h to leach the porogen, dried in a laminar flow hood, and kept in a vacuum chamber until implantation. Bridges were fabricated with a porogen to polymer ratio of 4:1, with the following formulations: (i) 250μm channel configuration, less than 38μm porogen; (ii) 250μm channel configuration, 63–106μm porogen; or (iii) 150μm channel configuration, less than 38μm porogen. The bridges had a length of 4mm, a height of 1.5mm, and a width of 2.6mm. Structural characteristics of scaffolds were imaged with a scanning electron microscope (S-3400N-II; Hitachi, Brisbane, CA) using the variable pressure mode and an ESED II detector (for low vacuum scanning electron microscopy (SEM) imaging). The microscope was operated at an electron voltage of 15kV.
Long Evans rats (female, 8–10 weeks old, 175–200g) were purchased from Charles River Laboratories (Wilmington, MA). For surgery, the animals were deeply anesthetized with intraperitoneal injection of ketamine and xylazine. Animals were placed on a heating pad maintained at 37°C, and ophthalmic ointment (PharmaDerm, Duluth, GA) was used to prevent drying of the eyes during surgery. An incision was made along the dorsal midline, and the paraspinal muscles were exposed and detached from the spinous processes. A laminectomy was performed with rongeurs (FST, Foster City, CA) at the T10 process. The exposed spinal cord at T10 was laterally hemisected on the left side at two sites using a microfeather (Electron Microscopy Sciences, Hatfield, PA) scalpel with a distance of 3mm between the sites (Fig. 1D). The tissue was subsequently excised. After hemostasis was achieved, a bridge was implanted between the rostral and caudal spinal cord stumps. The laminectomy site was covered with sterile Gelfoam (Henry Schein, Melville, NY) to prevent muscle adhesion to the bridge. The muscle layers were closed with 5-0 chromic gut sutures, and the skin was closed with metal wound clips (Stoelting, Wood Dale, IL). A microsurgical microscope (Leica Microsystems, Bannockburn, IL) was used to perform the surgical procedure. The three bridge conditions that were investigated are listed in Table 1. A total of 12 animals were implanted for each condition, with 6 implants retrieved at 2 and 6 weeks. Additionally, two animals were used as negative controls in which all procedures were kept constant with the exception that a bridge was not implanted.
After surgery, animal care consisted of injection with 10mL lactated Ringer's solution subcutaneously, and animals were allowed to recover overnight in cages maintained on a water-jacketed 37°C heating pad. Buprenex (0.01–0.05mg/kg) was administered subcutaneously for pain management every 12h for 48h, and baytril (2.5mg/kg) was administered to prevent bladder infection for 14 days or until recovery of spontaneous micturition. Animals were maintained in sanitized cages with Alpha-Dri bedding (Shephard/Specialty Papers, Watertown, TN); food and water were added ad libitum. All animals survived the surgery until bridge retrieval. Bladders were manually expressed twice daily until voluntary bladder release returned. Animals were treated in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory animals and the Institutional Animal Care and Use Committee (IACUC) protocol at Northwestern University and the University of California, Irvine.
Histological sections were obtained to investigate the bridge structure, integration with the host tissue, cell infiltration, and the presence of neurites. Spinal cord tissue was collected by necropsy at 2 or 6 weeks postimplantation. The spinal column was exposed and the vertebrae removed. The spinal cord between T8 and T12 was removed and snap-frozen by immersion in isopentane kept on dry ice (−55°C) and embedded in optimal cutting temperature compound (Sakura Finetek USA, Torrance, CA). Ten-micrometer-thick longitudinal (also termed sagittal) sections were cut, mounted on gelatin-coated slides, and stored at −20°C until further processing. Cells' infiltration into the bridge and the apposition with the surrounding tissue were viewed with the standard hematoxylin and eosin staining.
Staining of the histological sections was performed to identify the cell types present within the bridge. Nonperfused tissue samples (10μm thickness) were cryosectioned and postfixed with 4% paraformaldehyde (Sigma, St. Louis, MO) for 15min, immunoblocked with 10% normal serum for 30min, and immersed in 0.3% hydrogen peroxidase for 30min, all at room temperature. Astrocytes were identified using polyclonal antibody against glial fibrillary acidic protein (GFAP, 1:500 dilution; Sigma). Macrophages/monocytes were identified using antibodies to anti-ED-1 (dilution 1:150; Chemicon, Billerica, MA). Fibroblasts, Schwann cells, and endothelial cells were identified using the primary antibody anti-prolyl 4-hydroxylase rat prolyl 4-hydroxylase (rPH), dilution 1:200; AcrosAntibodies, Hiddenhausen, Germany), anti-S-100β (dilution 1:500; Sigma), and anti-rat endothelial cell antigen-1 (RECA-1, dilution 1:150; Serotec, Releigh, NC). Neurites and chondroitin sulfate proteoglycans (CSPGs) were also identified using antibodies to neurofilament 200 (dilution 1:350; Sigma), and CS56 (1:350; Sigma, dilution), respectively. To detect the primary antibodies, avidin-biotin immunoperoxidase staining (horseradish peroxidase staining) with Vectastain Elite ABC kit (Vector Laboratories, Burlingame, CA) was used with 3,3′-diaminobenzidine (Vector Laboratories) as the chromogen. The sections were counterstained with Mayer's hematoxylin (Surgipath Medical Ind., Richmond, IL). For immunofluorescence, the secondary antibody was conjugated with Alexa Fluor (1:500 dilution; Invitrogen, Carlsbad, CA). Images were captured by microscopy (Leica Microsystems, Wetzlar, Germany) using Spot image program (Diagnostic Instrument, Sterling Heights, MI) for horseradish peroxidase staining method or MetaVue 5 (Universal Imaging Corporation, Downingtown, PA) program for immunofluorescence. Negative control staining was done with omission of primary antibody during the procedure. Staining intensity was graded by three people, all blinded to the condition, in a semiquantitative fashion. Staining was categorized as negative (−), weakly positive (+), positive (++), and strongly positive (+++). The category+/−indicates that weakly positive staining was observed, yet the staining was not homogeneous throughout the sections.
The staining intensity for neurofilament was quantified in the bridge pores and channels as a measure of neural fiber ingrowth. The bridge area in a tissue section was divided into five regions, with one picture taken at random for each region. Images were captured on a Leica inverted fluorescence microscopy with a cooled CCD camera (Photometrics, Tucson, AZ) using MetaVue 5 acquisition software. Two images of the channels within the bridge were captured, and three images were captured randomly along the bridge/cord interface.
The staining intensity of CSPG was quantified at the bridge/spinal cord interface. All images were then processed using the NIH software, Image J, to record the integrated density for each image. The average integrated density and standard error were calculated. Six tissue sections per animal, and six animals per condition were used for quantification.
The behavioral changes of the rats after spinal cord injury were tested using the Basso, Beattie, Bresnahan (BBB) locomotor rating scale.24 The BBB scale analyzed the early, intermediate, and late phases of locomotor recovery on a 21-point scale at different time points postsurgery up to 8 weeks. Each rat was observed in an open field by two examiners for 4min, as described previously.24 Animal reflexes or movements elicited by touch of an examiner were not scored.
Statistical analyses were performed using the statistical package JMP (SAS, Cary, NC). For multiple comparisons, pairs were compared using ANOVA with post-hoc tests.
Bridges with multiple channels were fabricated and implanted into the rat spinal cord hemisection, and initial studies investigated integration with the surrounding tissue. The bridge had dimensions of 2.6mm in width, 1.5mm in height, and 4mm in length, which fits within the lateral hemisection (Fig. 1). The bridges had channels with diameters of 250 or 150μm, with a total of 7 or 20 channels, respectively (Fig. 1A, B). The bridges were more than 90% porous, as the porogen leached from the scaffold created a porous structure around the channels (Fig. 1C). After 2 and 6 weeks of implantation, all bridges maintained their positions and were well integrated with the host spinal cord tissue, and the bridge channel architecture was preserved (Fig. 2). Hematoxylin and eosin staining of longitudinal sections revealed complete tissue apposition and integration between the bridge and the surrounding spinal cord, with no indication of cavity formation (Fig. 2).
The interconnected open pore structure of the bridge supported cellular ingrowth, with the porogen size affecting the uniformity of cellular infiltration. Bridges fabricated with small porogen (less than 38μm) led to a nonuniform cellular ingrowth throughout the bridge after 2 weeks in vivo (Fig. 3A). However, within bridges fabricated with a larger porogen (63–106μm), cells were found throughout all bridge pores (Fig. 3B). At both 2 and 6 weeks postimplantation, the distribution of cells within the pores and channels was similar between all conditions (data not shown). At 2 weeks, cells in the bridge channels were aligned with the major axis of the channel for both the 250 and 150μm channels (Fig. 3C, D). This alignment within the channel was also observed after 6 weeks (not shown). Additionally, the boundaries of the channels within bridges formed with the small porogen were more easily viewed than the channels within bridges formed with large porogen. The smaller pores resulted in bridges with greater stability relative to large pore bridges, which maintained the channel integrity during removal of the pins used to create the channels and after implantation. For cells within the bridge, we categorized the bridge into two regions: channel and pores. A summary of the cells identified within the bridge pores and channels at both 2 and 6 weeks of implantation can be found in Table 2.
Astrocytes were observed primarily around the implant, with minimal infiltration into the bridge even at 6 weeks. Astrogliosis occurred primarily within the white matter (Fig. 4A), and reactive astrocytes were rarely observed within the bridge channels or pores at 2 weeks. GFAP staining for reactive astrocytes demonstrated a distinctive separation of positively stained cells within the spinal cord from the bridge (Fig. 4B). At 2 weeks, reactive astrocytes were observed in the tissue adjacent to the bridge but not in the bridge itself (Fig. 4C). Astrocytes located closer to the bridge appeared more reactive, based on the morphology and staining intensity. In tissue rostral to the bridge, the astrocytes were not reactive (Fig. 4D). By 6 weeks, the distribution of astrocytes was similar to that observed at 2 weeks; however, a few reactive astrocytes were observed within the bridge that were aligned within the bridge channels (Fig. 4E). The pores of the bridge did not contain GFAP-positive cells (Fig. 4E). In contrast, in tissue regions bordering a hemisection injury without an implanted bridge, GFAP staining indicated a heavy presence of reactive astrocytes (Fig. 4F). Results of this control condition suggest that the bridge may act to reduce the number of reactive astrocytes at the injury site, which may create a less inhibitory environment for axon outgrowth.
Consistent with the distribution of astrocytes, immunofluorescence staining for CSPG, a molecule secreted by reactive astrocytes that can inhibit regeneration, was low within the bridge at both 2 and 6 weeks (Fig. 5A). Further, the intensity of CSPG staining at the distal end of bridge/spinal cord interface (exiting zone) was similar among all the bridge conditions (Fig. 5B).
ED-1-positive (monocytes/macrophages) cells were one of the predominant cell types in the bridge at 2 and 6 weeks. At 2 weeks, ED-1-positive cells were observed equally within the channels and the bridge pores (Fig. 6A). Interestingly, the morphology of these cells varied between the bridge locations. Within the pores, the macrophages appeared large, whereas in the channels, they were small and aligned. After 6 weeks of implantation, the ED-1-positive staining was observed primarily within the pores of the bridge and not within the channels (Fig. 6B). Staining for ED-1 was observed to be closely associated with the polymer (Fig. 6C), and not within the pore interior, with no other significant changes observed (Table 2). The extent of ED-1 staining appeared to decrease between 2 and 6 weeks, which may suggest a decrease in inflammation. All bridge conditions exhibited a similar distribution of macrophages between the pores and channels at 2 and 6 weeks.
S-100β-positive cells, which include both Schwann cells and oligodendrocytes, were observed within the bridges at both time points. At 2 weeks of implantation, S-100β cells were located predominantly within the pores of the bridge and not the channels (Fig. 7A). Cells that were observed within the channel were located primarily near the entrance to the channels at the host tissue/bridge interface. These observations suggest that the cells may have entered the channel, and then migrated into the pores of the bridge. At 6 weeks of implantation, an increase in the number of S-100β cells was observed relative to 2 weeks (Fig. 7B). The cells were again located predominantly within the pores and not within the channels. S-100β-positive cells located in the channel were located only at the entrance to the channels, as the middle of the bridge had no apparent staining. No significant differences in the distribution of S-100β cells at either 2 or 6 weeks was observed for the different bridge conditions.
Fibroblasts (rPH-positive cells) were present within the bridges in quantities similar to macrophages/monocytes. The distribution of fibroblasts did not vary between the bridge conditions investigated. At 2 weeks of implantation, the fibroblasts were present at high levels within the channels (Fig. 8A, B), and to a lesser extent within the pores (Fig. 8A). At 6 weeks of implantation, the extent of rPH staining appeared similar between the channels and pores of the bridge (Fig. 8C). Fibroblasts within the channels at both time points were aligned with the channel (Fig. 8), which contrasts with the cells within the pores adjacent to the channels. At 6 weeks caudal to the implant, few cells stained positive, indicating the specificity of the antibody and the accumulation of fibroblasts within the implant (Fig. 8D).
RECA-1-positive (endothelial) cells had an increasing presence within the bridge between 2 and 6 weeks (Fig. 9). RECA-1 staining was observed predominantly within the channels, particularly for the bridges formed using porogen with a diameter less than 38μm. Importantly, the RECA-1-positive staining within the channels exhibited a linear or branching organization, which suggests the formation of vessel-like structures (Fig. 9B). This organization was observed within both 150 and 250μm channels.
Neurofilament staining was observed within the channels at both 2 and 6 weeks, and indicated a robust and aligned extension of neurites. At 2 weeks, neurofilament staining within the bridge was restricted to the region approximately 500–800μm from the rostral and caudal interface with the host tissue (Fig. 10A, B). No significant neurofilament staining was observed in the middle of the bridge at 2 weeks. However, at 6 weeks, neurofilament staining was observed within the middle of the bridge, a distance ranging from approximately 2–2.5mm from the channel entrance (Fig. 10C). The three bridges investigated had a similar profile of neurofilament staining; however, the bridges formed with the small porogen (less than 38μm) had more consistent staining between the channels. Importantly, for all bridges, neurofilament staining was localized primarily to the channels.
Subsequent studies investigated the functional recovery for animals implanted with multiple channel bridges (Fig. 11). Bridges fabricated with 250μm channels and less than 38μm porogen were implanted into a left lateral hemisection, and the mice were evaluated using the BBB scoring system.24 The animals achieved a score of approximately 10, which corresponds to occasional weight supported stepping with no forelimb to hindlimb coordination. On the injured (i.e., left) side, the BBB score increased during the initial 14 days to a score of 7.1, and had a more gradual increase to the final value obtained at 8 weeks. For the uninjured (i.e., right) side, the BBB scores at days 1, 7, and 14 were significantly greater than the values for the injured side (p<0.05). The BBB score observed at 14 days remained steady thereafter. These values suggest the ability of the bridge to stabilize the implant site and allow for some plastic reorganization of the spinal cord network.
Bridges for spinal cord regeneration must stabilize the injury site, maintain the pathways across the injury, and avoid cyst formation. We have demonstrated that these criteria are attained with a multiple channel bridge, which have full integration of the bridge with the surrounding tissue. The bridge retained its original size throughout the study and cells were present throughout the pores and channels, suggesting that the bridge maintained structural integrity throughout the study. In the absence of a bridge, cyst and cavity formation are frequently observed with spinal cord lesions.1 In this study, the porous PLG bridge fully integrated with the adjacent spinal cord tissue and supported the infiltration of surrounding cells. Bridges formed with the small porogen (less than 38μm) had channels that were more well defined throughout the 6-week study. In contrast to a previous report,19 all channels within the bridge contained cells, and the presence of cells within the channel may be necessary for axonal growth.
Reducing the glial scar and the presence of other inhibitory factors is a desirable function of the bridge, which would aid in creating a more permissive environment. After an injury, the glial scar develops, contributing to a physical and molecular barrier to regeneration. The glial scar consists primarily of reactive astrocytes, secreting inhibitory molecules such as CSPGs.25 Although the glial scar is generally described as an inhibitor to regeneration, it does perform several beneficial functions, such as repairing the blood–brain barrier and restricting the inflammatory response to limit cellular degeneration.9 In our study, minimal reactive astrocytes were present within the bridge, consistent with previous reports.19 However, at the hemisection injury site without a bridge, greater staining for reactive astrocytes was observed within the injury site at 2 weeks. Accordingly, CSPG deposition followed this pattern, and CSPG deposited near the end of the bridge may limit neuronal extension from the bridge back into the host tissue. Strategies to degrade the scar, such as the delivery of chondroitinase ABC, an enzyme that degrades CSPG, may be necessary.26,27
Macrophages, Schwann cells, fibroblasts, and endothelial cells are expected to be located at or near biomaterials implanted into the injured spinal cord; however, their relative distribution within the pores and channels has not been previously reported. Macrophages are commonly present at spinal cord injury sites28 and can function as scavengers of myelin and neuronal debris, and thus may enhance neuronal sprouting in the CNS. Macrophages produce a wide variety of cytokines (e.g., tumor necrosis factor-α and interleukin-1) and growth factors that are both directly and indirectly involved in regeneration.29,30 Implantation of activated macrophages to a completely transected adult rat spinal cord has been reported to partially restore their motor function and electrophysiological activity.31,32 Macrophages, however, can also release inflammatory cytokines and reactive nitrogen and oxygen intermediates that may limit regeneration.7 In total, the role of macrophages and spinal cord regeneration has not been clearly defined.19 Interestingly, macrophages were not observed within the channels at the 6-week time point, yet were observed within the pores. The greater presence of macrophages within the pores of the bridge may reflect their role in the foreign body response.14,15
The presence of S-100β-positive cells within the bridge may function to support and promote regeneration. S-100β-positive cells, which predominately defines Schwann cells and oligodendrocytes, were observed within our bridge pores and channels. Schwann cells may be recruited from the peripheral nervous system through the compromised blood–brain barrier and express and secrete a variety of neurotrophic factors, extracellular matrix molecules, and surface adhesion molecules that can stimulate axonal growth.28 Further, Schwann cells also function to myelinate CNS axons, which is crucial for the functional recovery of regenerated axons.33 Schwann cell transplantation has been widely investigated as a means to promote spinal cord regeneration,34,35 though their direct transplantation is complicated by limited cell survival.36
Fibroblasts and endothelial cells are cell types that would be commonly recruited to injury sites. Fibroblasts were observed within the pores and channels of the bridge, similar to reports by others, and is one of the first cell types to respond after implantation of a foreign material at any implant site.14 Endothelial cells, which are responsible for blood vessel formation and can also synergize with axonal elongation,37 were found sparsely distributed within our bridges, suggesting that revascularization of the injured environment has initiated.38,39 Numerous blood vessel structures were observed within the channels, which can provide nutrients necessary for ingrowth of cells and axons; however, it is unclear whether increasing vascularization could enhance axonal ingrowth.37
Aligned cells were observed within the channels, which could provide physical guidance substrates to restrict and direct axonal growth. Aligned cells and patterned surfaces have proven effective at orienting the extension of neuronal fibers in vitro.40 Alignment of proliferating Schwann cells was obtained by patterning surfaces, which was able to direct neuronal growth.41,42 Recently, in vitro studies indicated that 100μm channel patterns are able to attain linear and guided neuronal extensions more effectively than 250μm channel patterns.21 Further, the 150μm channel design allowed for an increase in number of channels from 7 channels (250μm channel design) to 20 channels. By providing more channels, the sprouting axons may have a higher chance of entering a channel that will guide them through the bridge. Interestingly, we observed significant cell alignment throughout the channel. The use of agarose hydrogels has also led to the alignment of Schwann cells and the orientation of growth for blood vessels and axons.19
A few different types of bridges with multiple linear channels spanning the length have been investigated in vivo. PLG bridges with 500μm channels were implanted into a complete spinal cord transection.18 The identity of infiltrating cells and the alignment within the channels were not investigated. Axons were observed within the channels for bridges seeded with Schwann cells embedded in Matrigel.18 Agarose and alginate-based multiple channel bridges have been fabricated using multiple techniques.16,19,20 The channel density for these hydrogel bridges is larger than with the PLG bridges, as the PLG bridges have thicker walls separating the channels. In this report, we demonstrate that ED-1-positive cells were observed within the pores, but not the channels at the 6-week time point. The hydrogels similarly had low levels of ED-1-positive cells within the channels. Importantly, cell ingrowth into the hydrogels is less consistent than with the PLG bridges.16 Agarose hydrogels had a numerous channels that did not contain significant cell numbers. Cellular infiltration into the channels appeared to be a prerequisite for axonal regrowth. The PLG scaffolds are known to adsorb proteins readily, whereas the hydrogels are generally described as having minimal protein adsorption.20 The increased level of protein adsorption may enhance cell infiltration and integration with the host tissue. Finally, the hydrogel scaffolds have provided some short-term release of neurotrophic factors.19 However, the process for PLG bridge fabrication allows for the encapsulation and sustained release of protein and DNA.23,43 The walls between the channels can serve as a reservoir for the factors, with the potential to provide inductive factors for extended times at the implant site.
Functional recovery after bridge implantation likely results from plasticity within the spinal cord. The objective of the current study was to investigate the relationship between the bridge properties and the host response. The lateral hemisection model is not the ideal model with which to assess functional recovery, as the animals generally recover some mobility after injury. Animals implanted with multiple channel bridges led to a partial functional recovery with a mean BBB score of approximately 10 at 8 weeks, reflecting that the animals had occasional weight-supported plantar steps with no hindlimb to forelimb coordination during locomotion. Thus, the bridge does not inhibit this normal recovery process. The bridge stabilizes the cord, and may limit secondary damage. The cell types recruited to the injury are critical, as they secrete factors that can influence axonal sprouting that can lead to plasticity that supports functional recovery. These bridges can also serve as a platform for the transplantation of cells,44,45 or localized delivery of neurotrophic factors.23,43 Taken together, these bridges implanted into a lateral hemisection provide a versatile system to investigate approaches that enhance axonal extension into and through an injury site, or to investigate the mechanisms limiting spinal cord regeneration.
Multiple channel bridges were implanted into a spinal cord hemisection to investigate and correlate the bridge design with the host response. The bridges maintained their architecture in vivo and supported the infiltration of fibroblasts, macrophages, Schwann cells, and endothelial cells from the surrounding tissue. Cells within the channels were aligned with the longitudinal axis, and axons were observed throughout the channels at 6 weeks of implantation. Relative to hydrogel bridges, the PLG bridges supported cell infiltration throughout the bridge and each channel, likely resulting from improved protein adsorption to PLG relative to the hydrogels. These structures support limited functional recovery, which likely occurs from plasticity within the spinal cord. Additionally, the fabrication process and the porous region between channels can serve as a reservoir for the localized, sustained release of protein and plasmid, which can be employed to target-specific cellular processes. These bridges can be utilized as a versatile tool to present multiple factors (extracellular matrix proteins and neurotrophic factors) and investigate their effect on axon growth into and through the bridge, which may ultimately lead to functional regeneration.
The authors would like to thank the Christopher and Dana Reeve Foundation (CDRF) Spinal Cord Injury Core Facility at the University of California Irvine, and Rebecca Nishi, M.S., Eliza Scott, Amber Nefas, and Hongli Liu for their assistance in training of the hemisection surgical procedure. This research was funded by grants from NIH (RO1 EB005678 and RO1 EB003806), and was also supported by the CDRF.
No competing financial interests exist.