Rubral astrocytic reactions to proximal and distal axotomy of rubrospinal neurons in the rat

Different macaque brain network remodeling after spinal cord injury and NT3 treatment

Graph theory-based analysis describes the brain as a complex network. Only a few studies have examined modular composition and functional connectivity (FC) between modules in patients with spinal cord injury (SCI). Little is known about the longitudinal changes in hubs and topological properties at the modular level after SCI and treatment. We analyzed differences in FC and nodal metrics reflecting modular interaction to investigate brain reorganization after SCI-induced compensation and neurotrophin-3 (NT3)-chitosan-induced regeneration. Mean inter-modular FC and participation coefficient of areas related to motor coordination were significantly higher in the treatment animals than in the SCI-only ones at the late stage. The magnocellular part of the red nucleus may reflect the best difference in brain reorganization after SCI and therapy. Treatment can enhance information flows between regions and promote the integration of motor functions to return to normal. These findings may reveal the information processing of disrupted network modules.

Strapping the spinal cord: an innovative experimental model of CNS injury in rats

Experimental models of spinal cord (SC) lesion are essential for understanding a few of the primary and secondary mechanisms of injury and functional recovery of the central nervous system (CNS). We have developed an experimental model of SC injury in adult rats (n=32), that involves the use of a device (SC-STRAPPER) that straps the SC and promotes gradual and controlled SC injury similar to clinical compressive SC injuries. SC strapping is a less-invasive procedure in comparison to other SC injury models, and it performs compression with smaller infection risk and undetectable paravertebral or vertebral lesions. The survival of the rats was 100%, minimizing the suffering of the animals. We have analyzed the histopathological changes that occur during experimental SC compression, as well as the immunohistochemical labeling for glial fibrillary acidic protein (GFAP). Animals survived for 21 days being thereafter anesthetized and perfused with aldehydes. SC lesions were associated with motor deficits and local increase in GFAP immunolabeling proportionate to the severity of the compression. This experimental model represents a potential contribution for neuroscientific research, providing a low-cost and rather simple system of controllable and reproducible SC experimental damage.

A subpopulation of reactive astrocytes at affected neuronal perikarya after hypophysectomy in adult rats

Intermediate filaments (IFs) of nestin and vimentin are expressed in immature astrocytes. In this study, we examined the re-expression of these early glial traits in rat reactive astrocytes in affected neuronal perikarya in supraoptic (SON) and paraventricular (PVN) nuclei induced by hypophysectomy. Double-labeling immunofluorescence confocal laser microscopy demonstrated that by 7 days post-lesion, both nestin and vimentin were present intensely in hypertrophied GFAP-IR reactive astrocytes in the area of hypophysectomized magnocellular neurons in SON and PVN, while nestin and vimentin are absent in the normal or sham-operated animals. As the gliotic reaction progressed, the morphology of nestin or vimentin-positive reactive astrocytes in SON but not PVN changed from stellate form at 7 days to thin and elongated shape, morphologically compatible with radial glia during development, at 14 days post-lesion. By 28 days post-lesion, while vimentin-IR persisted in reactive astrocytes in SON and PVN, nestin-IR could hardly be detected. The spatiotemporal pattern of nestin-IR and/or vimentin-IR in reactive astrocytes suggests astrocytes attempt to revert to a more primitive glia form indicated by changes in morphology and phenotype following hypophysectomy, which may contribute to neuronal trophism and plasticity in the lesioned HNS favoring neuronal maintenance and fiber outgrowth.

Proximity of lesioning determines response of facial motoneurons to peripheral axotomy

We recently found that rubrospinal (RS) neurons, which typify central neurons projecting within the central nervous system (CNS), exhibited different neuronal and glial reactions to axotomy at proximal as opposed to distal sites. To determine whether distance also determines the reaction to axonal injury of central neurons projecting to the periphery, we studied the temporal expression of four free-radical-related enzymes as well as the severity of cell loss, perineuronal astrocytic and microglial reactions, and degeneration of the proximal central axons of facial motoneurons after axotomies performed at various sites on the brainstem surface and in the stylomastoid foramen, respectively. Distal lesions resulted in upregulation of these neurons' expression of nitric oxide synthase (NOS) and persistent downregulation of their expression of the NOS-activating enzyme calcineurin. It also led to transient upregulation of their expression of manganese-dependent superoxide dismutase (Mn-SOD), and resulted in a mild neuronal loss. Proximal axotomy led to an upregulation of NOS but a transient downregulation in the expression of calcineurin and Mn-SOD at 4 weeks after injury. This was accompanied by severe cell loss and swelling of mitochondria at 2-4 weeks postinjury. However, neither proximal nor distal axonal lesioning led to nuclear fragmentation or TUNEL staining of neurons. Proximal as opposed to distal axotomy produced an earlier transformation of glial morphology, including the hypertrophy of astrocytic processes and metamorphosis of ramified microglia to amoeboid cells. We unexpectedly found that unlike RS neurons, whose central axons degenerated slowly and in an anterograde manner only after the severe cell loss induced by proximal axotomy, the central axons of facial motoneurons degenerated rapidly and in a retrograde manner independently of the severity of loss of these neurons after axotomy. However, degeneration began sooner after proximal than after distal axotomy. Since the central axons of both rubrospinal neurons and facial motoneurons lie within the CNS, the differences in whether and how they degenerated after axotomy suggests that central neurons that project within and outside the CNS are inherently different. The significance of these and also the free radical environment regulation differences between these two types of neurons following close and distant axotomies remains to be explored.

Spinal Axonal Injury Induces Brief Downregulation of Ionotropic Glutamate Receptors and No Stripping of Synapses in Cord-Projection Central Neurons

Spinal cord injury often damages the axons of cord-projecting central neurons. To determine whether their excitatory inputs are altered following axonal injury, we used rat rubrospinal neurons as a model and examined their excitatory input following upper cervical axotomy. Anterograde tracing showed that the primary afferents from the cerebellum terminated in a pattern similar to that of control animals. Ultrastructurally, neurons in the injured nucleus were contacted by excitatory synapses of normal appearance, with no sign of glial stripping. Since cerebellar fibers are glutamatergic, we examined the expression of ionotropic receptor subunits GluR1-4 and NR1 for AMPA and NMDA receptors, respectively, in control and injured neurons using immunolabeling methods. In control neurons, GluR2 appeared to be low as compared to GluR1, GluR3, and GluR4, while NR1 labeling was intense. Following unilateral tractotomy, the levels of expression of each subunit in axotomized neurons appeared to be normal, with the exception that they were lower than those of control neurons of the nonlesioned side at 2-6 days postinjury. These findings suggest that axotomized neurons are only temporarily protected from excitotoxicity. This is in sharp contrast to the responses of central neurons that innervate peripheral targets, in which both synaptic stripping and reduction of their ionotropic glutamate receptor subunits persist following axotomy. The absence of an injury-induced trimming of afferents and stripping of synapses and the lack of a persistent downregulation of postsynaptic receptors might enable injured cord-projection neurons to continue to control their supraspinal targets during most of their postinjury survival. Although this may support neurons by providing trophic influences, it nevertheless may subject them to excitotoxicity and ultimately lead to their degenerative fate.

Glial fibrillary acidic protein expression in the hamster red nucleus: effects of axotomy and testosterone treatment

Testosterone propionate (TP) administration coincident with facial nerve axotomy in the hamster attenuates glial fibrillary acidic protein (GFAP) expression in the facial nucleus that is normally increased following axotomy alone. This ability of TP to modulate astrocyte activity has been linked to the ability of steroid hormones to enhance the regenerative response of injured motor neurons. In an ongoing study designed to examine the potential influences of steroid hormones on centrally projecting motoneurons, the astrocyte reaction in the red nucleus was examined. In the present study, in situ hybridization was used to assess changes in GFAP mRNA in the hamster red nucleus following spinal cord injury (SCI) and TP treatment. Castrated male hamsters were subjected to right rubrospinal tract (RST) transection at spinal cord level T1, with half the animals implanted subcutaneously with Silastic capsules containing 100% crystalline TP and the remainder sham implanted. The uninjured red nucleus served as an internal control. Postoperative survival times were 1, 2, 7, and 14 days. Qualitative-quantitative analyses of emulsion autoradiograms were accomplished. Axotomy alone resulted in a significant but transient increase in GFAP mRNA levels at 2 days postoperative in the injured red nucleus compared with the contralateral uninjured red nucleus. However, in TP-treated animals, GFAP mRNA levels were no different than control levels at 2 dpo but were significantly increased at 7 dpo relative to contralateral control. Additionally, the increase in GFAP mRNA levels following TP treatment was significantly smaller than following axotomy alone. These data suggest that testosterone both delays and reduces the astrocytic reaction in the red nucleus following rubrospinal tract axotomy, and confirms a difference between peripheral and central glial responses to axotomy and steroid administration.

Close axonal injury of rubrospinal neurons induced transient perineuronal astrocytic and microglial reaction that coincided with their massive degeneration

To learn more about the pathophysiology of axonal injury and the significance of axon collaterals on the survival of axotomized cord-projection central neurons, we studied the survival rate, surrounding astrocytic and microglial reactions, and bouton coverage on rat rubrospinal cell bodies following their axonal lesion at the brain stem and upper cervical level. The brain stem lesion disconnected most rubrospinal neurons from all their targets, while the upper cervical lesion spared their supraspinal collaterals. Much higher cell loss accompanied by robust astrocytic and microglial reaction was found following brain stem than upper cervical lesion starting 4 days postaxotomy. The reaction of astrocytes had subsided while microglial reaction remained relatively robust by 10 weeks postaxotomy when the cell loss had slowed down. Ultrastructural observation revealed that reactive astrocytes covered 40%, an increase from the 20% of control, of brain stem-axotomized rubrospinal cell body surface at 4 days and 2 weeks and returned to normal levels by 10 weeks postlesion. An increase of apposition by axons and dendrites and a moderate decrease of round and flattened vesicle-containing bouton contacts at 4 days and 2 weeks and returning to normal levels at 10 weeks postaxotomy accompanied this. It appears that although axotomy induced robust astrocytic reaction around cord-projection central neurons, this, unlike their periphery-projection counterparts, failed to effectively strip their somatic synapses. In effect, this might in part determine neuronal fate following axonal injury.

Central nervous system pathology in 25 dogs with chronic degenerative radiculomyelopathy

The neuropathology of 20 German shepherd dogs and five German shepherd dog crosses with chronic degenerative radiculomyelopathy were analysed by conventional techniques, immunocytochemistry and electron microscopy. There were previously unrecognised changes in brain nuclei. In the spinal cord, both motor and sensory tracts were involved, principally in their more distal regions. Wallerian degeneration affected the corticorubrospinal pathways in the lateral columns and the ventral funiculi, predominantly in the caudal thoracic and lumbar segments, although more cranial involvement was also observed. The dorsal columns were affected in the caudal lumbar region and the cervical fasciculus gracilis. The regional distribution was variable between cases. Within the brain, abnormalities, including chromatolysis, gliosis and neuronal loss were observed in the red nucleus, lateral vestibular nucleus and, occasionally, in the dentate nucleus. The changes in brain nuclei were compared with those found in dogs at various times after a focal spinal injury. The neuronal changes in the brain may be related to the primary site of damage, and possible aetiological mechanisms are discussed.

Fate of the supraspinal collaterals of cord-projection neurons following upper spinal axonal injury

In investigating the fate of the cord-projecting CNS neurons following spinal axonal injury, we have demonstrated that surviving rat rubrospinal neurons have altered electrical membrane properties so that their input/output relationship was increased. Further, we found that the synaptic inhibition they received from nearby reticular formation was also reduced following injury. Whether or not these property changes were functional was dependent on the output connections of injured neurons. In the current communication, we examined the supraspinal efferents of the injured neurons recognizing that normal neurons innervate not only spinal but also supraspinal targets. To this end we conducted anterograde tracing on the injured red nucleus 8 weeks following spinal lesion. Results showed that injured rubrospinal neurons still innervated the same supraspinal targets, targeted by normal neurons. We subsequently evaluated the relative intensity of the sustained supraspinal connectivity by examining, in detail, the cerebellar projection of rubrospinal neurons of similarly injured animals using retrograde tracing technique. Here our data revealed that the number, distribution and labeling intensity of rubrospinal neurons projecting to the cerebellum were unchanged following cord injury. In conclusion, although spinal cord injury deprive cord-projecting CNS neurons of their spinal targets, injured neurons survived with altered electrical membrane properties and intact supraspinal projections. The sustained supraspinal connections might allow injured cord-projecting CNS neurons to exert a different weight of influence on higher centers following spinal cord injury.

Androgenic regulation of the central glia response following nerve damage

Current research on the effects of gonadal steroids on the brain and spinal cord indicates that these agents have profound trophic effects on many aspects of neuronal functioning, including cell survival, growth and metabolism, elaboration of processes, synaptogenesis, and neurotransmission (Jones et al., 1985; Luine, 1985; Nordeen et al., 1985; Matsumoto et al., 1988a,b; Gould et al., 1990). Since many of the aspects of normal neuronal functioning altered by gonadal steroids are affected by injury to the nervous system, we initiated a series of experiments designed to exploit the trophic capabilities of steroids as therapeutic agents in neuronal injury and repair (Kujawa et al., 1989, 1991; Kujawa and Jones, 1990). Three steroid-sensitive model systems were used for these studies: the hamster facial motoneuron, the rat sciatic motoneuron, and the hamster rubrospinal motoneuron. The results of our initial series of experiments suggest that androgens, and possibly estrogens, act either directly or indirectly on the injured motoneuron and enhance elements of the neuronal reparative response that are critical to successful recovery of function. Recently, we discovered that gonadal steroids may also modulate the central glia response to nerve damage. In this review, a summary of our data identifying a therapeutic role for androgens in enhancing the reparative response of motoneurons to injury is presented. This is followed by a discussion of the effects of androgens on the glial response to injury.

Clusterin upregulation following rubrospinal tract lesion in the adult rat

We have examined the expression of the multifunctional protein clusterin in the axotomized red nucleus, at the lesion site in the lateral funiculus of C3, as well as along the Wallerian degeneration in the lateral funiculus of T1. There was a marked increase in clusterin-immunoreactivity (IR) and clusterin mRNA in red nucleus nerve cell bodies. An early, transient occurrence of large, heavily clusterin-IR globules were found in axons in the spinal cord at the lesion site in C3 as well as a marked upregulation of mRNA for clusterin, presumably associated with reactive astrocytes and oligodendrocytes from 1 to 4 weeks postoperatively. Clusterin-IR and its mRNA were markedly increased in the zone of Wallerian degeneration at T1, where some strongly expressing cells were identified as oligodendrocytes. Taken together with previous changes in clusterin expression following peripheral nerve and dorsal root injury, we suggest that this protein is involved in regenerative as well as degenerative neural responses.

Central neuron-glial and glial-glial interactions following axon injury

Axon injury rapidly activates microglial and astroglial cells close to the axotomized neurons. Following motor axon injury, astrocytes upregulate within hour(s) the gap junction protein connexin-43, and within one day glial fibrillary acidic protein (GFAP). Concomitantly, microglial cells proliferate and migrate towards the axotomized neuron perikarya. Analogous responses occur in central termination territories of peripherally injured sensory ganglion cells. The activated microglia express a number of inflammatory and immune mediators. When neuron degeneration occurs, microglia act as phagocytes. This is uncommon after peripheral nerve injury in the adult mammal, however, and the functional implications of the glial cell responses in this situation are unclear. When central axons are injured, the glial cell responses around the affected neuron perikarya appears to be minimal or absent, unless neuron degeneration occurs. Microglia proliferate, and astrocytes upregulate GFAP along central axons undergoing anterograde, Wallerian, degeneration. Although microglia develop into phagocytes, they eliminate the disintegrating myelin very slowly, presumably because they fail to release molecules which facilitate phagocytosis. During later stages of Wallerian degeneration, oligodendrocytes express clusterin, a glycoprotein implicated in several conditions of cell degeneration. A hypothetical scheme for glial cell activation following axon injury is discussed, implying the injured neurons initially interact with adjacent astrocytes. Subsequently, neighbouring resting microglia are activated. These glial reactions are amplified by paracrine and autocrine mechanisms, in which cytokines appear to be important mediators. The specific functional properties of the activated glial cells will determine their influence on neuronal survival, axon regeneration, and synaptic plasticity. The control of the induction and progression of these responses are therefore likely to be critical for the outcome of, for example, neurotrauma, brain ischemia and chronic neurodegenerative diseases.

Membrane properties and inhibitory connections of normal and upper cervically axotomized rubrospinal neurons in the rat

Membrane properties and inhibitory synaptic connections of normal and axotomized rat rubrospinal neurons were examined using a coronal slice preparation. Rubrospinal neurons were axotomized at the C2 vertebral level in vivo. Retrograde labelling in vivo and intracellular biocytin injection following recording were combined to identify recorded axotomized rubrospinal neurons. Their input resistances decreased three and four days and became higher than normal four and 10 weeks following lesioning which coincided with a sequential increase and decrease of their soma area. On the other hand, although their membrane time-constant was reduced three and four days following lesioning, it returned to normal value four and 10 weeks following axotomy. Other than these, their membrane current-voltage relationship including an inward rectification in the hyperpolarizing direction was not altered. Normal rubrospinal neurons generated very fast spikes which were not affected by axotomy. Both normal and axotomized cells generated trains of repetitive spikes with a fast spike frequency adaptation at the beginning upon suprathreshold current injection. However, the slope of the steady-state spike frequency and applied current relationship was increased four and 10 weeks following axotomy which also showed an increased steady-state spike frequency in response to high-amplitude current injection. Synaptically, the amplitude and duration of the monosynaptic inhibitory potential evoked from nearby reticular formation were reduced following axotomy. In addition, fewer rubrospinal neurons were found to receive this inhibition 10 weeks following axotomy. Thus, our results show that spinal axotomy induces a time-dependent modification of the membrane properties and spike generating behaviour of rubrospinal neurons which probably represents an initial decrease and a later increase of their excitability. This is accompanied by a persistent decrease of synaptic inhibition which is expected to affect structures that remained innervated by the undamaged axon collaterals of these spinally axotomized neurons.