Light activated approach for large gap peripheral nerve repair
Fairbairn, Neil G.
Introduction: Conventional suture repair of peripheral nerves following injury is associated with several limitations such as technical difficulty, intra- and extra-neural scar formation, axonal escape and the leakage of neurotrophic factors. These limitations are particularly relevant following nerve grafting when regenerating axons must traverse two coaptation sites. Outcomes following suture repair are notoriously poor, providing large impetus for the development of alternative methods. Photochemical tissue bonding (PTB) uses visible light to create sutureless, non-thermal bonds between two closely apposed tissue surfaces stained with a photoactive dye. When used with a human amnion nerve wrap for end-to end nerve repair, this technique results in superior functional and histological outcomes in comparison to conventional epineurial suture. When initially applied to large gap injury and nerve grafting, outcomes were unsuccessful due to proteolytic degradation of amnion and photochemical bonds during extended periods of recovery. Chemical crosslinking of nerve wraps prior to PTB may improve wrap durability and efficacy of technique. This thesis provides a comprehensive three-phase assessment of the efficacy of this novel approach when applied to the repair of large gap injuries with nerve grafts. Phase 1 assesses the ex vivo biomechanical properties of nerve wraps and light activated bonds in addition to the in vivo performance of photochemically sealed crosslinked nerve wraps against several other clinically relevant fixation methods in a rodent sciatic nerve isograft model. Following major multi-limb injury and amputation, demand for autogenous nerve graft may exceed that which can be supplied by the patient. Acellular nerve allograft (ANA) is an alternative option in these circumstances although outcomes are typically inferior to autograft. Phase 2 assesses the performance of the optimum repair strategy from phase 1 against conventional epineurial suture when applied to ANA. Most studies investigating the efficacy of novel repair techniques tend to perform repairs immediately following injury, a situation that rarely occurs clinically. Delays of weeks or months are not uncommon and have been shown to have a detrimental effect on regeneration and outcome. Phase 3 assesses the efficacy of PTB when applied to delayed nerve grafting. Additional work investigating a novel imaging technique for visualizing nerve revascularisation following injury and repair has been included. Optical frequency domain imaging (OFDI) uses low power infrared light to provide real time in vivo imaging of tissue microvasculature and flow characteristics. Originally applied to the study of tumour biology, this technique may prove useful for outcome assessment in preclinical research and eventually for the assessment of nerve viability in the clinical setting. Experiments investigating the early development of a brain body interface system (BBI) for upper limb reanimation following spinal cord injury (SCI) have also been included. The ultimate aim of this project is to restore autonomous motor control in a non-human primate (NHP) using cortically driven stimulation of peripheral nerves via implantable nerve cuffs. The experiments reported in this thesis detail the development of a selective, reversible paralysis model of elbow flexion in a NHP and demonstrate selective fascicular stimulation using acute and chronically implanted nerve cuffs in rodent and murine models. Methods: Phase 1: Three candidate nerve wraps (human amnion (HAM), crosslinked human amnion (xHAM), crosslinked swine intestinal sub-mucosa (xSIS)) and 3 fixation methods (suture, fibrin glue, PTB) were investigated. Crosslinking was performed using (1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC)/N-hydroxysuccinimide (NHS). Biomechanical tests were performed using a tensiometer. Ex vivo wrap durability was assessed using a type- 2 collagenase degradation assay. Under isoflurane anaesthesia, 110 inbred male Lewis rats had 15mm left sciatic nerve defects created and repaired with reversed isografts. 9 groups (n=10) had isografts secured by one of the aforementioned wrap/fixation combinations. PTB repairs had nerve wraps and nerve ends stained with photoactive dye (Rose Bengal) and, once nerve ends were apposed and wrapped circumferentially, the interface was illuminated with a 532nm laser. Fibrin repairs had nerve ends apposed, wrapped circumferentially and secured with Tisseel fibrin glue. Suture repairs had nerve ends apposed, wrapped circumferentially and then secured with two 10-0 nylon sutures at each coaptation site (one either side of each repair). Positive and negative control groups (n=10) were repaired with graft+suture (10-0 nylon) and no repair respectively. Phase 2: 20 sciatic nerves were harvested from Sprague Dawley rats and sent to AxoGen Inc. for decellularisation. An additional 20 male inbred Lewis rats were randomized into 2 groups (n=10). All rats had 15mm left sciatic nerve defects created and repaired with processed ANA. 1 group had nerves secured using conventional epineurial suture. The remaining group had ANA secured using photochemically sealed amnion wraps. Phase 3: 40 inbred male Lewis rats were randomized into 4 groups (n=10). All 40 rats had 15mm left sciatic nerve gaps created and reconstructed with reversed isografts harvested from donor Sprague Dawley rats. In groups 1 and 2, nerve gaps were repaired immediately with either conventional epineurial suture or photochemically sealed amnion wraps, respectively. In groups 3 and 4, repair took place 30- days following injury using either conventional epineurial suture or photochemically sealed amnion wraps, respectively. All outcomes were assessed using walking track analysis and calculation of sciatic function index (SFI). Walking track analysis and SFI was performed pre-operatively, after the 30-day delay (phase 3) and at 30-day intervals following surgery. Following sacrifice after 5-months, left (experimental) and right (control) gastrocnemius muscles were excised and weighed for calculation of muscle mass retention. Nerves were excised for histomorphometric analysis including axon count, fiber diameter, axon diameter, myelin thickness and G-ratio. For all in vivo experiments, statistical analysis was performed using ANOVA, repeated measures ANOVA and the post hoc Bonferroni test. Optical Frequency Domain Imaging (OFDI) pilot study: eight rodents were randomized into 4 groups (n=2): (1) crush injury, (2) transection and end-to-end repair, (3) transection and repair of 10mm nerve gap using contralateral autograft, (4) transection and repair of 10mm nerve gap using ANA. Under ketamine/xylazine anaesthesia, all rodents had sciatic nerves exposed through hind limb dorsolateral incisions. Imaging was performed immediately pre-injury, immediately post-injury and on post-operative days 1, 3, 5 and 7. Rodents were secured firmly to polystyrene platforms in order to reduce movement artifact during imaging Brain-Body Interface (BBI) experiments: In the upper limb of a Rhesus macaque nonhuman primate, the median nerve branch to brachialis and radial nerve branch to brachioradialis were transected, leaving elbow flexion entirely reliant on the musculocutaneous nerve. The musculocutaneous nerve was transposed into a subcutaneous position. Ultrasound guided nerve block resulted in a highly selective, reversible paralysis of elbow flexion. Under ketamine/xylazine anaesthesia, Sprague Dawley rats (n=5) and C57 Black 6 mice (n=5) had sciatic nerves exposed through dorsolateral, muscle splitting incisions. 8-channel stimulating cuff electrodes were wrapped around sciatic nerves and connected to a Tucker14 Davies stimulation/recording system. Electromyography (EMG) needle electrodes were inserted into the tibialis anterior (TA) and gastrocnemius (G) muscles to record muscle activity. Single pulses and pulse trains were delivered to each animal whilst pulse parameters were systematically varied. In one survival rat, a stimulating nerve cuff and EMG recording electrodes were implanted chronically to assess equipment biocompatibility and recording stability. Results: Phase 1: Optimal tensile strength, photochemical bond strength and resistance to collagenase degradation were achieved using 4mM EDC/1mM NHS. Following sacrifice, crosslinked nerve wraps were still present. Un-crosslinked material was completely degraded. xHAM+PTB repairs recovered greatest mean SFI although this was not statistically significant compared with standard repair (-67.9+/-1.6 vs -71.7+/-1.6 (mean+/-SEM)). xHAM+PTB repairs also recovered greatest muscle mass retention and this was significant in comparison to standard repair (67.3% +/- 1.4 vs 60.0% +/- 1.6; p=0.02). No significant difference in axon diameter existed between treatment groups. Fiber and axon diameter and myelin thickness were all significantly greater in the xHAM+PTB group in comparison to standard repair (6.87μm+/-0.04 vs. 5.47μm+/-0.03, p<0.0001; 4.51μm+/-0.04 vs. 3.50μm+/- 0.03, p<0.0001; 2.35μm+/-0.01 vs. 1.96μm+/-0.01, p<0.0001). Phase 2: Following sacrifice, all ANAs were in continuity and, on gross observation, showed evidence of regeneration. ANA repaired photochemically had less extraneural scar tissue formation in comparison to standard epineurial suture. SFI did not differ significantly between ANA+suture and ANA+PTB groups after 5-months follow up (-80.3+/-1.3 vs. -78.3+/-1.6 respectively; p=0.3). ANA+suture and ANA+PTB recovered significantly less SFI than corresponding isograft+suture and isograft+PTB. Gastrocnemius muscle mass retention did not differ significantly between ANA+suture and ANA+PTB groups (53.3%+/-1.8 vs. 55.2%+/-1.6 respectively; p=0.5). ANA+suture and ANA+PTB recovered significantly less muscle mass than isograft+suture and isograft+PTB, respectively. Muscle mass retention was statistically equivalent between ANA+PTB and isograft+suture (55.2+/-1.8% vs. 60+/-1.6%; p=0.22). With the exception of G-ratio, no significant differences existed between ANA+suture and ANA+PTB for any remaining histomorphometric parameter. ANA+PTB was statistically equivalent to isograft+suture for all histomorphometric parameters. Phase 3: After 5-months, in both immediate and delayed repairs, PTB resulted in greater mean SFI values but these results were not statistically significant (-72.3+/-1.5 vs. -68.5+/-1.5; p=0.4 and -80.1+/-1.4 vs. -77.3+/-1.5; p=1). Both suture and PTB fixation methods resulted in significantly greater recovery of SFI when performed immediately rather than after a 30-day delay after injury (- 72.3+/-1.5 vs. -80.1+/-1.4; p=0.003 and -68.5+/-1.5 vs. -77.3+/-1.5; p=0.002). Significantly greater muscle mass retention occurred following PTB fixation in both immediate and delayed repairs (64.9%+/-1.8 vs. 59.0%+/-1.1; p=0.03 and 60.2%+/-1.4 vs. 54.1%+/-1.7; p=0.04). No significant difference existed between immediate suture and delayed suture, or immediate PTB and delayed PTB groups. Muscle mass retention was not significantly different between immediate suture and delayed PTB (59.0+/-1.1% vs. 60.2+/-1.4%; p=1). No significant differences in axon count or density existed between groups. Fiber diameter, axon diameter, myelin thickness and G-ratio were not significantly different between immediate suture and delayed PTB. With the exception of G-ratio, all other histomorphometric parameter comparisons were significantly different between treatment groups, with immediate PTB achieving greatest recovery and delayed suture being poorest. OFDI: OFDI successfully provided real time images of nerve microvasculature. Vessels of the mesoneurial and epineurial longitudinal plexus were readily identifiable in uninjured nerve. Marked tortuosity of these vessels was apparent. Different light scattering properties of muscle and nerve permitted easy differentiation between tissues. Injury sites were easily visible as areas of relative hypovascularity. Following crush injury, the longitudinal intrinsic plexus remained intact. Following end-to-end neurorrhaphy, although initially hypovascular, by post-operative day 7, a florid angiogenic response had occurred at the repair site. Following repair of 10mm autografts, the re-establishment of longitudinal vessels was apparent by day 4. Following repair of 10mm ANA, grafts remained relatively avascular. A predominance of inosculated longitudinal vessels existed with a relatively minor centripetal contribution from the surrounding muscle bed. BBI: Highly selective, reversible paralysis of non-human primate (NHP) elbow flexion was achieved. Application of nerve cuff electrodes around small animal sciatic nerves was technically simple and permitted muscle stimulation in all animals. Needle EMG electrodes successfully recorded muscle activity in all animals. In 9 animals, selective activation of tibial and common peroneal fascicles was possible, allowing the plotting of muscle recruitment curves as a function of stimulation amplitude. Following 6 months of chronic implantation, the nerve cuff and EMG electrodes were well tolerated and stimulation and recording was still possible. Conclusions PTB: Chemical crosslinking of biological nerve wraps improves tensile strength and in vivo resistance to biodegradation, whilst preserving the formation of light-activated bonds. Light activated sealing of crosslinked amnion around nerve isograft coaptation sites results in significant improvements in muscle mass retention and nerve histomorphometry in comparison to conventional suture repair. SIS wraps were unsuitable for bonding around small diameter nerves although when applied to larger caliber nerves, may still be useful. Outcomes following ANA were inferior to those achieved following isograft reconstruction in phase 1. No significant difference was detected between photochemically sealed ANA and sutured ANA although the increase in mean values for muscle mass retention and histomorphometry were comparable to epineurial suture, the current standard of care. Immediate repairs performed significantly better than delayed repairs and light-activated repairs performed significantly better than sutured repairs. Delayed PTB repairs were statistically comparable to immediate suture. Light-activated sealing leads to superior outcomes in comparison to suture when applied to nerve isografts and may improve outcomes following the use of ANA and when repairs occur following a surgical delay. OFDI provides a real time, in vivo imaging modality for the assessment of nerve revascularisation following injury and repair. This has potential applications for pre-clinical outcome assessment in peripheral nerve research and also as a means of assessing peripheral nerve viability in the clinical setting. BBI: We have shown that selective fascicular stimulation of sub-1mm peripheral nerves can be achieved using implantable cuff electrodes. In combination with a highly selective and reversible paralysis of elbow extension in a non-human primate model, these cuffs will be used to deliver cortically recorded signals directly into upper limb nerves. Through operant conditioning and feedback, this BBI should enable the NHP to regain movement during paralysis.