Axons in the peripheral nervous system (PNS) and central nervous system (CNS) form sprouts after injury. Elongation of regenerating axonal sprouts has been observed as the exception within the adult mammalian CNS but is the rule in the PNS of mammals as well as in the CNS of some fish and amphibians. The relative importance of intrinsic neuronal properties and axonal environment in determining the extent of axonal regrowth is unknown. Neuroglial cells, nerve growth factor and target tissues such as smooth muscle are known to influence neuronal responses to injury. Here we have examined the capacity of transected axons originating in the CNS to regrow into nerve grafts containing Schwann cells.

The success of peripheral and fetal neural tissue in promoting outgrowth of axons from the adult mammalian central nervous system has tended to focus attention on local interactions between extending axons and their environment. The contribution to axon regeneration of biochemical and morphological changes in injured neurones is more difficult to evaluate. We report here that long spinal axons of primary sensory neurones are 100 times more likely to regenerate into peripheral nerve grafts if their peripheral axons are also cut. Regenerative behaviour at the axon tip seems to be strongly influenced by inducible events in the nerve cell.

In this study the actions of NGF in regulating peptide expression were examined in vivo in adult rat primary sensory neurons. The hypothesis that NGF might tonically inhibit expression of some peptides was tested specifically. In situ hybridization and immunohistochemistry were used to detect presence or absence of alpha-CGRP, beta-CGRP, SP, SOM, VIP, CCK, NPY, and GAL as well as their mRNAs. In neurons in normal lumbar DRG alpha-CGRP, beta-CGRP, SP, and SOM are abundantly and heterogeneously expressed whereas few neurons have detectable VIP, CCK, NPY, or GAL. Two weeks following sciatic nerve transection, concentrations of alpha-CGRP, beta-CGRP, SP, and SOM plus their mRNAs have decreased to background in all but a few neurons. In contrast, VIP, CCK, NPY, and GAL are now synthesized in many neurons. Delayed intrathecal infusion of NGF (125 ng/microliter/hr) for 7 d, starting 2 weeks after injury counteracted the decrease in expression of alpha- CGRP, beta-CGRP and SP expression, but not SOM. This lack of influence of NGF on SOM is consistent with the absence of high-affinity NGF receptors and trk mRNA in SOM-positive neurons. Delayed infusion of NGF also reduced the number of neurons expressing VIP, CCK, NPY, and GAL after injury by approximately one-half in each subpopulation. Therefore, we suggest that NGF suppresses expression of these four peptides but only if the neurons also have NGF receptors. The results show that NGF can regulate peptide expression differentially and may also be part of the signal that allows reversion to normal of responses to injury as axons regenerate.

The distributions of mRNAs for the protooncogene trk and the low- affinity NGF receptor (LNGFR) were studied by hybridization with oligonucleotide probes on sections of adult rat primary sensory and sympathetic ganglia. For comparison with high-affinity binding sites, adjacent sections were processed for NGF receptor radioautography. Among neurons in lumbar dorsal root ganglia and trigeminal ganglia, trk mRNA and NGF-binding sites were closely colocalized; this finding together with previous direct evidence in other cell types is taken to indicate that trk protein is an essential component of the high- affinity NGF receptor in adult sensory neurons. In lumbar dorsal root ganglia and trigeminal ganglia, abundant LNGFR mRNA was found in all neurons with strong 125I-NGF labeling and on additional neurons lacking high-affinity NGF-binding sites. The presence of abundant LNGFR in neurons with high-affinity receptors could be the cause and/or consequence of their ability to respond to NGF. Neurons with abundant LNGFR mRNA but few high-affinity NGF-binding sites may have receptors for other members of the neurotrophin family. In nodose ganglia, neurons with high concentrations of LNGFR mRNA greatly outnumbered the small percentage with abundant trk mRNA. Following intrathecal infusion of NGF to otherwise normal dorsal root ganglia, the concentrations of LNGFR mRNA but not those of trk mRNA and NGF-binding sites were increased in NGF-responsive neurons. The usual single normal pattern of frequency histograms of LNGFR labeling indices became bimodal in response to NGF. Concentrations of NGF-binding sites, LNGFR mRNA, and trk mRNA were all decreased by peripheral nerve transection and restored by exogenous NGF, the restoration being complete for LNGFR mRNA and partial for trk mRNA and NGF-binding sites. The data indicate that NGF can regulate both LNGFR and trk mRNAs but do not clarify the possible contribution of the LNGFR protein to high-affinity binding sites.

To investigate regeneration of long spinal axons, the right lateral column of the rat spinal cord was cut at high cervical, low cervical, midthoracic or lumbar level, and one end of an autologous sciatic nerve segment was grafted to the spinal cord at the site of incision. Three to six months after operation, the origin of axons in the grafts was traced retrogradely with horseradish peroxidase injected into the grafts and, in some cases, anterogradely with radioautography of tritiated amino acids injected into the brainstem. Axons from each of the major lateral spinal tracts arising in the brainstem as well as axons ascending from the lower spinal cord succeeded in growing into low cervical grafts. However, long descending axons rarely regenerated after midthoracic or lumbar injury; axons ascending from lumbar segments of the spinal cord usually failed to enter high cervical grafts. Differences in axonal regrowth at the four segmental levels were not simply attributable to dwindling of axonal number in fibre tracts. Axonal regeneration from Clarke's column or the red nucleus was observed only with lesions causing atrophy of many neurons. There was no obvious example of a fibre tract in the lateral spinal columns from which axons failed to regenerate nor from which axons regenerated exceptionally well. Under the conditions of these experiments, the distance from cell body to injury appeared to be an important determinant of axonal regeneration.

RNA from rat dorsal root ganglia was analyzed in search of potentially beneficial cytokines that are induced in dorsal root ganglia by nerve injury. By reverse transcription, the PCR, and Southern blotting, interleukin-6 mRNA was detected during development but not in normal adult dorsal root ganglia, reappeared within 1 d of sciatic nerve transection, was maximally increased after 2 and 4 d, and decreased below the threshold of detection within 1 week. By RNase protection assay, interleukin-6 mRNA was consistently present in RNA from dorsal root ganglia removed from rats 4 d following transection but not in control dorsal root ganglia. Interleukin-6 bioactivity was also present in dorsal root ganglia following nerve injury. By in situ hybridization, interleukin-6 mRNA was localized within large and medium- sized axotomized neurons. In summary, some sensory neurons respond to axotomy with a brisk transient increase in synthesis of interleukin-6. Injury to the sciatic nerve also induced mRNAs for interleukin-1 beta and tumor necrosis factor-alpha in dorsal root ganglia. The inductions of interleukin-1 beta and tumor necrosis factor-alpha mRNAs were later and more sustained than that of interleukin-6 mRNA. The cellular sources of these two cytokines have not been defined.

Although crushed axons in a dorsal spinal root normally regenerate more slowly than peripheral axons, their regeneration can be accelerated by a conditioning lesion to the corresponding peripheral nerve. These and other observations indicate that injury to peripheral sensory axons triggers changes in their nerve cell bodies that contribute to axonal regeneration. To investigate mechanisms of activating nerve cell bodies, an inflammatory reaction was provoked in rat dorsal root ganglia (DRG) through injection of Corynebacterium parvum. This inflammation enhanced regeneration in the associated dorsal root, increasing 4-fold the number of regenerating fibers 17 d after crushing; peripheral nerve regeneration was not accelerated. A milder stimulation of dorsal root regeneration was detected after direct injection of isogenous macrophages into the ganglion. It is concluded that changes favorable to axonal regeneration can be induced by products of inflammatory cells acting in the vicinity of the nerve cell body. Satellite glial cells and other unidentified cells in lumbar DRG were shown by thymidine radioautography to proliferate after sciatic nerve transection or injection of C. parvum into the ganglia. Intrathecal infusion of mitomycin C suppressed axotomy-induced mitosis of satellite glial cells but did not impede axonal regeneration in the dorsal root or the peripheral nerve. Nevertheless, the similarity in reactions of satellite glial cells during 2 processes that activate neurons adds indirect support to the idea that non-neuronal cells in the DRG might influence regenerative responses of primary sensory neurons.

To survey the distribution of neuronal receptors for NGF, sections of the rat brain, spinal cord, and peripheral ganglia were incubated in vitro with radioiodinated NGF and examined by autoradiography. NGF binds selectively with high affinity to most sympathetic neurons and many primary sensory neurons together with their intraspinal or intramedullary axons. In autoradiographs of the brain, labeled neuronal perikarya are seen in the basal forebrain, the caudate-putamen, the medulla oblongata, the ventral cochlear nucleus, and the dorsal nucleus of the lateral lemniscus. The distribution of neurons binding NGF resembles the distribution of cholinergic neurons in the forebrain, but these 2 systems overlap very little in the brain stem. In extracts of the brain or spinal cord enriched for plasma membranes, avid binding sites are regionally manifest with properties similar to those of fetal peripheral neurons. The localization of neurons expressing the high-affinity receptor for NGF defies simple correlation with neurotransmitter function or embryogenesis.