SUBSTANCE-P IN CUTANEOUS PRIMARY SENSORY NEURONS - A COMPARISON OF MODELS OF NERVE INJURY THAT ALLOW VARYING DEGREES OF REGENERATION

We have studied changes in neuropeptide expression in four different models of nerve injury in adult rats. The models involved the cutaneous sural and saphenous nerves, and were associated with different degrees of regrowth and peripheral target reinnervation. These were: simple crush of the nerve, complete cut and self-anastomosis; cut and ligation, and cut and anastomosis of the nerve to an isolated stump of peripheral nerve. Thus, in the first two models a partial or complete reinnervation of peripheral targets was possible, while in the third and fourth it was not. The last model allowed regenerating fibres to come into contact with Schwann cells in the distal stump. We measured substance P-like immunoreactivity in the manipulated nerves (by radioimmunoassay) and the number of manipulated afferents expressing the peptide in dorsal root ganglion cells (by combined immunohistochemistry and retrograde labelling), at time points up to 12 weeks after the nerve manipulations. The retrograde labelling also allowed estimates of cell death. Two weeks after the nerve injuries, when no cell death had occurred, the nerves subjected to a cut lesion (last three models) all showed very low levels of substance P-like immunoreactivity, both in the amounts in peripheral nerve, and in the number of manipulated cell staining positively (P < 0.01). In contrast, the crush model showed no significant change in substance P levels in the nerve (P > 0.05), but a significant increase in the number of immunopositive cells (P < 0.01). Twelve weeks after the nerve manipulations, a variable degree of cell death was seen. Only 9% of afferents in the crush model were lost (P > 0.05 compared with normal) but a 39 and 45% loss was seen in tie and resuture models, respectively, (P < 0.05 for both, compared with normal), and a 63% loss in the stump model (P < 0.01 compared to normal, and P < 0.05 compared to tie and resuture models). An analysis of cell size distributions indicated that cell death affected both large and small cells. At 12 weeks, the levels of substance P in the first two models (associated with peripheral reinnervation) had returned towards, but did not reach, normal (P < 0.01), whilst the stump model showed no significant recovery and the tie model was intermediate. Proportionately more manipulated cells were found to express substance P immunoreactivity in the stump model than expected after allowing for cell death. The maintenance of substance P levels and lack of cell death in the crush model are unlikely to be explained by the regrowing sprouts coming into immediate contact with Schwann cells in the distal nerve. The results invite speculation that cut axons may allow the accumulation and entry of positively harmful factors, and the double lesion associated with the stump model may exaggerate their actions here.