Regeneration of adult rat sensory and motor neuron axons through chimeric peroneal nerve grafts containing donor Schwann cells engineered to express different neurotrophic factors

Abstract Large peripheral nerve (PN) defects require bridging substrates to restore tissue continuity and permit the regrowth of sensory and motor axons. We previously showed that cell-free PN segments repopulated ex vivo with Schwann cells (SCs) transduced with lentiviral vectors (LV) to express different growth factors (BDNF, CNTF or NT-3) supported the regeneration of axons across a 1 cm peroneal nerve defect (Godinho et al., 2013). Graft morphology, the number of regrown axons, the ratio of myelinated to unmyelinated axons, and hindlimb locomotor function differed depending on the growth factor engineered into SCs. Here we extend these observations, adding more LVs (expressing GDNF or NGF) and characterising regenerating sensory and motor neurons after injection of the retrograde tracer Fluorogold (FG) into peroneal nerve distal to grafts, 10 weeks after surgery. Counts were also made in rats with intact nerves and in animals receiving autografts, acellular grafts, or grafts containing LV-GFP transduced SCs. Counts and analysis of FG positive (+) DRG neurons were made from lumbar (L5) ganglia. Graft groups contained fewer labeled sensory neurons than non-operated controls, but this decrease was only significant in the LV-GDNF group. These grafts had a complex fascicular morphology that may have resulted in axon trapping. The proportion of FG+ sensory neurons immunopositive for calcitonin-gene related peptide (CGRP) varied between groups, there being a significantly higher percentage in autografts and most neurotrophic factor groups compared to the LV-CNTF, LV-GFP and acellular groups. Furthermore, the proportion of regenerating isolectin B4+ neurons was significantly greater in the LV-NT-3 group compared to other groups, including autografts and non-lesion controls. Immunohistochemical analysis of longitudinal graft sections revealed that all grafts contained a reduced number of choline acetyltransferase (ChAT) positive axons, but this decrease was significant only in the GDNF and NT-3 graft groups. We also assessed the number and phenotype of regrowing lumbar FG+ motor neurons in non-lesioned animals, and in rats with autografts, acellular grafts, or in grafts containing SCs expressing GFP, CNTF, NGF or NT-3. The overall number of FG+ motor neurons per section was similar in all groups; however in tissue immunostained for NeuN (expressed in α- but not γ-motor neurons) the proportion of NeuN negative FG+ neurons ranged from about 40–50% in all groups except the NT-3 group, where the percentage was 82%, significantly more than the SC-GFP group. Immunostaining for the vesicular glutamate transporter VGLUT-1 revealed occasional proprioceptive terminals in ‘contact’ with regenerating FG+ α-motor neurons in PN grafted animals, the acellular group having the lowest counts. In sum, while all graft types supported sensory and motor axon regrowth, there appeared to be axon trapping in SC-GDNF grafts, and data from the SC-NT-3 group revealed greater regeneration of sensory CGRP+ and IB4+ neurons, preferential regeneration of γ-motor neurons and perhaps partial restoration of monosynaptic sensorimotor relays.