There is controversy about whether axotomized neurons undergo death or only severe atrophy after spinal cord injury (SCI) in mammals. Lampreys recover from complete spinal transection, but only about half of the severed spinal-projecting axons regenerate through the site of injury. The fates of the unregenerated neurons remain unknown, and until now death of axotomized spinal-projecting neurons has not been described in the lamprey brain. We now report that in animals allowed to survive for 12 or more weeks after spinal cord transection, several identified reticulospinal (RS) neurons were missing in Nissl-stained or neurofilament-immunostained brain whole mounts. At earlier times, these neurons were swollen and pale in Nissl-stained preparations. Retrograde fluorescent labeling from the site of transection combined with TUNEL histochemistry suggested that neuronal death, including that of the identified RS neurons, began in animals 4 weeks posttransection, reaching a peak at 12-16 weeks. This was not seen in untransected animals. The TUNEL positivity suggests that some cells were dying by apoptosis. Of special interest, among the identified neurons, this delayed cell death was restricted to neurons that at earlier posttransection times have a low probability of regeneration. These data show that SCI induces delayed cell death in lamprey spinal-projecting neurons and suggest that the reason why some neurons are "bad regenerators" is that they are already undergoing apoptotic cell death. Thus protection from apoptosis may be necessary in order to enhance axonal regeneration after SCI.

Paralysis following spinal cord injury (SCI) is due to interruption of axons and their failure to regenerate. It has been suggested that the small GTPase RhoA may be an intracellular signaling convergence point for several types of growth-inhibiting extracellular molecules. Even if this is true in vitro, it is not clear from studies in mammalian SCI, whether the effects of RhoA manipulations on axon growth in vivo are due to a RhoA-mediated inhibition of true regeneration or only of collateral sprouting from spared axons, since work on SCI generally is performed with partial injury models. RhoA also has been implicated in local neuronal apoptosis after SCI, but whether this reflects an effect on axotomy-induced cell death or an effect on other pathological mechanisms is not known. In order to resolve these ambiguities, we studied the effects of RhoA knockdown in the sea lamprey central nervous system (CNS), where after complete spinal cord transection (TX), robust but incomplete regeneration of large axons belonging to individually identified reticulospinal (RS) neurons occurs, and where some RS neurons show unambiguous delayed retrograde apoptosis after axotomy. RhoA protein was detected in neurons and axons of the lamprey brain and spinal cord, and its expression was increased post-TX. Knockdown of RhoA in vivo by retrogradely-delivered morpholino antisense oligonucleotides (MOs) to the RS neurons significantly reduced retrograde apoptosis signaling in identified RS neurons post-SCI, as indicated by Fluorochrome Labeled Inhibitor of Caspases (FLICA) in brain wholemounts. In individual RS neurons, the reduction of caspase activation by RhoA knockdown began at 2 weeks post-TX and was still seen at 8 weeks. RhoA knockdown slowed axon retraction and possibly increased early axon regeneration in the proximal stump. The number of axons regenerating beyond the lesion more than 5 mm at 10 weeks post-TX also was increased. Thus RhoA knockdown both enhanced true axon regeneration and inhibited retrograde apoptosis signaling after SCI.

Apoptosis is a major feature in neural development and important in traumatic diseases. The presence of active caspases is a widely accepted marker of apoptosis. We report here the development of a method to study neuronal apoptotic death in whole-mounted brain preparations using fluorochrome-labeled inhibitors of caspases (FLICA). As a model we used axotomy-induced retrograde neuronal death in the CNS of larval sea lampreys. Once inside the cell, the FLICA reagents bind covalently to active caspases causing apoptotic cells to fluoresce, whereas nonapoptotic cells remain unstained. The fluorescent probe, the poly caspase inhibitor FAM-VAD-FMK, was applied to whole-mounted brain preparations of larval sea lampreys 2 weeks after a complete spinal cord (SC) transection. Specific labeling occurred only in identifiable spinal-projecting neurons of the brainstem previously shown to undergo apoptotic neuronal death at later times after SC transection. These neurons also exhibited intense labeling 2 weeks after a complete SC transection when a specific caspase-8 inhibitor (FAM-LETD-FMK) served as the probe. In this study we show that FLICA reagents can be used to detect specific activated caspases in identified neurons of the whole-mounted lamprey brain. Our results suggest that axotomy may cause neuronal apoptosis by activation of the extrinsic apoptotic pathway.

The sea lamprey recovers from spinal cord transection by a process that involves directionally specific regeneration of axons. The mechanisms underlying this specificity are not known, but they may involve molecular cues similar to those that guide the growth of spinal cord axons during development, such as netrins and semaphorins. To test the role of guidance cues in regeneration, we cloned netrin and its receptor UNC-5 from lamprey central nervous system (CNS) and studied their expression after spinal cord transection. In situ hybridization showed that (1) mRNA for netrin is expressed in the spinal cord, primarily in neurons of the lateral gray matter and in dorsal cells; (2) mRNA for UNC-5 is expressed in lamprey reticulospinal neurons; (3) following spinal cord transection, UNC-5 message was dramatically downregulated at two weeks, during the period of axon dieback; (4) UNC-5 message was upregulated at three weeks, when many axons are beginning to regenerate; and (5) axotomy-induced expression of UNC-5 occurred primarily in neurons whose axons regenerate poorly. Because the UNC-5 receptor is thought to mediate the chemorepellent effects of netrins, netrin signaling may play a role in limiting or channeling the regeneration of certain neurons. These data strengthen the rationale for studying the role of developmental guidance molecules in CNS regeneration.

To explore the role of axon guidance molecules during regeneration in the lamprey spinal cord, we examined the expression of mRNAs for semaphorin 3 (Sema3), semaphorin 4 (Sema4), and netrin during regeneration by in situ hybridization. Control lampreys contained netrin-expressing neurons along the length of the spinal cord. After spinal transection, netrin expression was downregulated in neurons close (500 mum to 10 mm) to the transection at 2 and 4 weeks. A high level of Sema4 expression was found in the neurons of the gray matter and occasionally in the dorsal and the edge cells. Fourteen days after spinal cord transection Sema4 mRNA expression was absent from dorsal and edge cells but was still present in neurons of the gray matter. At 30 days the expression had declined to some extent in neurons and was absent in dorsal and edge cells. In control animals, Sema3 was expressed in neurons of the gray matter and in dorsal and edge cells. Two weeks after transection, Sema3 expression was upregulated near the lesion, but absent in dorsal cells. By 4 weeks a few neurons expressed Sema3 at 20 mm caudal to the transection but no expression was detected 1 mm from the transection. Isolectin I-B(4) labeling for microglia/macrophages showed that the number of Sema3-expressing microglia/macrophages increased dramatically at the injury site over time. The downregulation of netrin and upregulation of Sema3 near the transection suggests a possible role of netrin and semaphorins in restricting axonal regeneration in the injured spinal cord.

The sea lamprey recovers normal-appearing locomotion after spinal cord transection and its spinal axons regenerate selectively in their correct paths. However, among identified reticulospinal neurons some are consistently bad regenerators and only about 50% of severed reticulospinal axons regenerate through the site of injury. We previously suggested (Shifman and Selzer, 2000) that selective chemorepulsion might explain why some neurons are bad regenerators and others not. To explore the role of additional chemorepulsive axonal guidance molecules during regeneration, we examined the expression of the repulsive guidance molecule (RGM) and its receptor neogenin by in situ hybridization and quantitative PCR. RGM mRNA was expressed in the spinal cord, primarily in neurons of the lateral gray matter and in dorsal cells. Following spinal cord transection, RGM message was downregulated in neurons close (within 10 mm) to the transection at 2 and 4 weeks, although it was upregulated in reactive microglia at 2 weeks post-transection. Neogenin mRNA expression was unchanged in the brainstem after spinal cord transection, and among the identified reticulospinal neurons, was detected only in “bad regenerators, Neurons that are known to regenerate well never expressed neogenin. The downregulation of RGM expression in neurons near the transection may increase the probability that regenerating axons will regenerate through the site of injury and entered caudal spinal cord.

No abstract

Spinal cord injury (SCI) is a devastating condition that leads to permanent disability because injured axons do not regenerate across the trauma zone to reconnect to their targets. A prerequisite for axonal regeneration will be the prevention of retrograde degeneration that could lead to neuronal death. However, the specific molecular mechanisms of axotomy-induced degeneration of spinal-projecting neurons have not been elucidated yet. In lampreys, SCI induces the apoptotic death of identifiable descending neurons that are “bad regenerators/poor survivors” after SCI. Here, we investigated the apoptotic process activated in identifiable descending neurons of lampreys after SCI. For this, we studied caspase activation by using fluorochrome-labeled inhibitors of caspases, the degeneration of spinal-projecting neurons using Fluro-Jade C staining, and the involvement of the intrinsic apoptotic pathway by means of cytochrome c and Vα double immunofluorescence. Our results provide evidence that, after SCI, bad-regenerating spinal cord-projecting neurons slowly degenerate and that the extrinsic pathway of apoptosis is involved in this process. Experiments using the microtubule stabilizer Taxol showed that caspase-8 signaling is retrogradely transported by microtubules from the site of axotomy to the neuronal soma. Preventing the activation of this process could be an important therapeutic approach after SCI in mammals.