Noradrenergic innervation of the rat spinal cord caudal to a complete spinal cord transection: Effects of olfactory ensheathing glia

Transplantation of olfactory bulb-derived olfactory ensheathing glia (OEG) combined with step training improves hindlimb locomotion in adult rats with a complete spinal cord transection. Spinal cord injury studies use the presence of noradrenergic (NA) axons caudal to the injury site as evidence of axonal regeneration and we previously found more NA axons just caudal to the transection in OEG- than media-injected spinal rats. We therefore hypothesized that OEG transplantation promotes descending coeruleospinal regeneration that contributes to the recovery of hindlimb locomotion. Now we report that NA axons are present throughout the caudal stump of both media- and OEG-injected spinal rats and they enter the spinal cord from the periphery via dorsal and ventral roots and along large penetrating blood vessels. These results indicate that the presence of NA fibers in the caudal spinal cord is not a reliable indicator of coeruleospinal regeneration. We then asked if NA axons appose cholinergic neurons associated with motor functions, i.e., central canal cluster and partition cells (active during fictive locomotion) and somatic motor neurons (SMNs). We found more NA varicosities adjacent to central canal cluster cells, partition cells, and SMNs in the lumbar enlargement of OEG- than media-injected rats. As non-synaptic release of NA is common in the spinal cord, more associations between NA varicosities and motor-associated cholinergic neurons in the lumbar spinal cord may contribute to the improved treadmill stepping observed in OEG-injected spinal rats. This effect could be mediated through direct association with SMNs and/or indirectly via cholinergic interneurons.     

Kappa opioid receptor density is consistent along the rostrocaudal axis of the female rat spinal cord

Kappa opioid receptors (KORs) were immunocytochemically localized at four different levels of the spinal cord of normally-cycling female rats in estrus or diestrus. KOR labeling was primarily observed in fine processes and a few neuronal cell bodies in the superficial dorsal horn and the dorsolateral funiculus. Quantitative light microscopic densitometry of the superficial dorsal horn revealed that there were no significant differences in KOR densities among spinal segments C1--C2, T2, T13--L1, and L6--S1 in either the estrus or diestrus phases. These results suggest that the potential for KOR-mediated antinociceptive responses is consistent along the rostrocaudal axis of the female rat spinal cord.     

Immunohistochemical study of a dopaminergic system in the spinal cord of the ray, Raja radiata

Using a specific antibody we have demonstrated the presence of an intrinsic dopamine (DA)-containing system in the spinal cord of a cartilaginous fish. The DA neurons are distributed throughout the spinal cord in the subependymal layer that surrounds the central canal. These cells extend a thick process into the canal and thinner neurites throughout the gray matter; they resemble the liquor-contacting neurons reported for a range of vertebrates.     

Norepinephrine throughout the spinal cord of the cat: I. Normal quantitative laminar and segmental distribution

Norepinephrine (NE) concentrations were measured by radioenzymatic assay in microdissected individual laminae of each segment of the cat spinal cord. Norepinephrine was detected in all areas of the spinal gray matter and showed more than a 7-fold difference in concentration between the laminae with the highest and lowest NE. The cervical, thoracic, and lumbosacral spinal regions showed significant interlaminar differences in NE. Intersegmental changes in NE were seen within single laminae of the thoracic and lumbosacral spinal cord, but not in the cervical spinal cord. A significant rostral to caudal, increasing regional gradient of NE was observed from the cervical to lumbosacral spinal cord in laminae I-III, V, VI, VII, and IX. In the intermediolateral cell column (IML), epinephrine concentrations were 2 to 5% of NE. Neither neurotransmitter showed a significant intersegmental variation in the IML. These data should prove useful in further defining the precise role of NE in specific regions of the spinal cord that mediate sensory, motor, autonomic, or propriospinal functions.     

Transient decrease of acetylcholinesterase in ventral horn neurons caudal to a low thoracic spinal cord hemisection in the adult rat

Light microscopic enzyme: histochemistry was employed to study the alterations of acetylcholinesterase (AChE) within lumbosacral ventral horn neurons at survival times of 1, 4, 7, 14, 28, 60, and 90 days after low thoracic spinal cord hemisection in adult rats. The intensity of histochemical staining was quantified using densitometric techniques. Virtually all ventral horn neurons of sham-operated and unoperated animals, which served as controls, displayed intense AChE staining. Hemisection of the spinal cord induced a transient, ipsilateral decrease of AChE staining in most neuronal cell bodies and in the neuropil of lamina IX at all segmental levels caudal to the lesion. Quantitative analysis of representative segments revealed a reduction of AChE in the ventral horn during a postoperative (p.o.) period of 1 to 28 days followed by a phase of recovery over the next two months. AChE activity still remained slightly reduced, even at 90 days p.o. The transient decrease in ACNE is a well-known metabolic response of axotomized motoneurons. However, the observed changes of AChE reactivity in intact motoneurons ipsilateral and caudal to the hemisection are presumably induced by the interruption of supraspinal descending pathways. These metabolic changes may functionally affect the whole motor unit and be involved in the disturbances of motor function following spinal cord injury.     

Temporal plasticity of dorsal horn somatosensory neurons after acute and chronic spinal cord hemisection in rat

Unilateral T13 hemisection of the rat spinal cord produces a model of chronic spinal cord injury (SCI) that is characterized by bilateral hyperexcitability of lumbar dorsal horn neurons, and behavioral signs of central pain. While we have demonstrated that responsiveness of multireceptive (MR) dorsal horn neurons is dramatically increased at 28 days after injury, the effects of acute hemisection are unknown and predicted to be different than observed chronically. In the present study, the consequences of T13 hemisection are examined acutely at 45 min in MR neurons both ipsilateral and contralateral to the site of injury, and compared to the same class of cells at 28 days after injury (n=20 cells total per group: 2–3 cells/side of the cord from n=5 animals). Acutely, ipsilateral to the hemisection, both spontaneous and evoked activity of MR neurons were significantly increased, whereas contralaterally, only evoked activity was significantly increased. In animals 28 days after hemisection, spontaneous activity of MR neurons was comparable to intact levels ipsilaterally, and cells exhibited hyperexcitability to evoked stimuli bilaterally. Expansion of cutaneous receptive fields was observed only in hindpaws ipsilateral to the lesion, acutely. These results demonstrate dynamic plasticity in properties of dorsal horn somatosensory neurons after SCI.     

Differential effects ofp-chlorophenylalanine on indoleamines in brainstem nuclei and spinal cord of rats. I. Biochemical and behavioral analysis

The involvement of endogenous serotonergic pathways in the mediation of antinociception has been indicated by electrophysiological, pharmacological and behavioral experiments. However, manipulation of the indole pathway, either by lesioning of raphe nuclei or drug intervention, often produces disparate results. In particular, serotonin (5-HT) synthesis inhibition withp-chlorophenylalanine (PCPA) has been reported to produce either hyperalgesia or analgesia, depending upon the type of pain measurement examined. In the present study, we sought to evaluate the effects of PCPA on (1) behavioral responses to noxious stimulation, and (2) levels of serotonin, tryptophan and 5-hydroxyindoleacetic acid (5-HIAA) in raphe nuclei (pallidus, obscurus, magnus and dorsalis) and spinal cord regions by HPLC with electrochemical detection. Treatment of rats with 400 or 600 mg/kg of PCPA for 3 consecutive days resulted in significant elevations in pain thresholds assessed by tail withdrawal from radiant heat as well as vocalization to electric shock of the tail. The effect of PCPA on vocalization threshold was particularly striking, for the majority of animals showed a nociceptive-specific attenuation of this response. Although the PCPA induced changes in indole content of the various raphe nuclei were not unequivocally dose-dependent, differential reductions of serotonin and 5-HIAA were clearly detected in the various raphe regions. Nuclei raphe pallidus and obscurus were depleted of 5-HT and 5-HIAA to the greatest extent, whereas levels detected in nuclei raphe magnus and dorsalis were reduced by 30–40% from control values. Metabolism of 5-HT and 5-HIAA appeared unaffected by PCPA in all regions examined except the dorsal portion of the spinal cord. These findings collectively suggest that the effects of PCPA are not uniform throughout the central nervous system and raise the possibility that discrepancies in the behavior literature may be attributed to drug-induced changes in some, but not all serotonergic pathways.     

When is the maximal effect of pre-administered systemic morphine on carrageenin evoked spinal c-Fos expression in the rat?

This study evaluated, in awake rats, the time course of the expression of c-Fos in spinal cord neurons, in the L4–L5 segments, at various time points after intraplantar carrageenin (0.5 h, 1 h, 1.5 h, 2 h and 2.5 h). In addition, the effects of pre-administered morphine (3 mg/kg, i.v.) on the c-Fos expression, at the various time points, were studied. Very few Fos-like immunoreactive (Fos-LI) neurons were observed 0.5 h after carrageenin. However, spinal c-Fos expression increased initially (at 1 h), in the superficial laminae (I–II) of the spinal dorsal hom, and incrementally increased both in the superficial and deep (V–VI) laminae at later time points after carrageenin. Systemic morphine did not significantly decrease the number of superficial Fos-LI neurons observed 1 h after carrageenin, whereas it significantly reduced the number of superficial Fos-LI neurons induced at 1.5 h and 2 h after carrageenin (58 ± 3% and 57 ± 10% reduction, P < 0.001, respectively). In addition, morphine reduced the number of deep Fos-LI neurons at 1.5 h and 2 h after carrageenin (86 ± 4%, P < 0.01 and 82 ± 8%, P < 0.001 reduction as compared to control carrageenin expression, respectively). In contrast, morphine was less efficacious in decreasing the number of Fos-LI neurons observed in the superficial and deep laminae at 2.5 h after carrageenin (34 ± 6% and 59 ± 6% reduction, P < 0.001, respectively). Thus, the peak effect of pre-administered morphine on carrageenin evoked c-Fos expression was observed 1.5 h and 2 h after intraplantar carrageenin, with a weaker effect observed at 2.5 h after carrageenin. The pharmacokinetic complications between the time course of the antinociceptive effects of morphine and c-Fos expression is discussed. These results clearly demonstrate that studies of c-Fos expression with pharmacological investigations should take into consideration this finding since one delay after the stimulation does not give a full indication of the full potential of the drug tested.     

Inhibition of spinal nociceptive transmission from the midbrain, pons and medulla in the rat: activation of descending inhibition by morphine, glutamate and electrical stimulation

It is generally believed that morphine activates a descending system(s) of inhibition, an effect contributing significantly to the analgesia produced. There has arisen, however, considerable controversy on this point. To address whether morphine inhibits spinal nociceptive transmission when given into the brainstem, the effects of focal electrical stimulation and monosodium S-glutamate (Glu) given in the periaqueductal gray (PAG), the locus coeruleus/subcoeruleus (LC/SC) and/or the nucleus raphe magnus (NRM) on spinal unit responses to noxious heating (50 °C) of the skin were examined and compared with effects produced by morphine (Mor). Focal electrical stimulation in 46 sites in the midbrain, dorsolateral pons and ventromedial medulla reliably inhibited unit responses to noxious heating of the skin (mean 34% of control). Microinjections of Glu (50 nmol, 0.5 μl) were made into 17 sites in the midbrain, 10 sites in the LC/SC and 11 sites in the NRM, inhibiting unit responses to a mean 57% at 22 of the 38 sites of microinjection. Mor (10–20 μg, 0.5 μl) was microinjected into 15 sites in the midbrain, 13 sites in the LC/SC and 11 sites in the NRM, inhibiting unit responses to heat to 63% of control at 24 sites of microinjection. The effects of morphine were shown to be receptor specific by antagonism with naloxone administered either intravenously or into the brainstem at the same site of microinjection as morphine. In 31 sites in the midbrain, dorsolateral pons and ventromedial medulla, microinjections of both Mor and Glu into the same sites attenuated unit responses to heating of the skin to a mean 77% and 71% of control, respectively. The results support the hypothesis that Mor acts supraspinally to modulate spinal nociceptive transmission by activating an endogenous descending inhibitory system(s). Focal electrical stimulation, glutamate and morphine modulated spinal nociceptive transmission by activation of descending inhibitory systems whose cell bodies of origin are in the PAG, the LC/SC or the NRM.     

Serotonin and norepinephrine in the spinal cord of man

The content of serotonin (5-HT), 5-hydroxyindoleacetic acid (5-HIAA) and norepinephrine (NE) was analysed in 71 human spinal cords obtained post-mortem. The highest content of 5-HT, 5-HIAA and NE was found in the lumbar enlargement of the spinal cord. 5-HT and 5-HIAA content increased from fetal to adult spinal cord whereas the content of NE decreased. Characteristic segmental distribution of measured monoamines was present in adult spinal cord only. In two patients spinal cord lesion led to the reduction in spinal cord content of 5-HT, 5-HIAA and NE and loss of characteristic segmental distribution of these substances. These results are in general agreement with observations on spinal cord of different animal species.     

Motor neuron destruction in guinea pigs immunized with bovine spinal cord ventral horn homogenate: experimental autoimmune gray matter disease

Guinea pigs immunized with bovine spinal cord ventral horn homogenate develop muscle weakness with electromyographic and morphologic evidence of denervation. Pathological examination demonstrates a loss of motoneurons and scattered inflammatory foci primarily localized to the spinal cord. Immunohistochemical techniques document the presence of immunoglobulin G at the motor end plate and around the external membrane and within the cytoplasm of motoneurons. This syndrome of experimental autoimmune gray matter disease (EAGMD) differs from experimental autoimmune motor neuron disease induced by inoculation with purified motoneurons and also differs from experimental autoimmune encephalomyelitis. The existence of two different forms of immune-mediated motoneuron destruction suggests that a number of cytoplasmic and membrane antigens may give rise to an immunologically based attack on the motor system.     

Substance P releases and augments the morphine-evoked release of adenosine from spinal cord

The effects of substance P on the morphine-evoked release of adenosine were examined. Substance P alone produced a multiphasic effect on release of adenosine, with release occurring at low nanomolar concentrations and at a micromolar concentration, but not at intermediate concentrations. An inactive dose of substance P augmented the morphine-evoked release of adenosine at a nanomolar concentration of morphine. Release of adenosine by substance P alone (1 nM) or substance P/morphine (100 nM/10 nM) was Ca²⁺-dependent and originated from capsaicin-sensitive nerve terminals.     

Increases in protein kinase C gamma immunoreactivity in the spinal cord of rats associated with tolerance to the analgesic effects of morphine

Our previous studies have indicated a critical role of protein kinase C (PKC) in intracellular mechanisms of tolerance to morphine analgesia. In the present experiments, we examined (1) the cellular distribution of a PKC isoform (PKCγ) in the spinal cord dorsal horn of rats associated with morphine tolerance by utilizing an immunocytochemical method and (2) the effects of theN-methyl-d-aspartate receptor antagonist MK-801 on tolerance-associated PKCγ changes. In association with the development of tolerance to morphine analgesia induced by once daily intrathecal administration of 10 μg morphine for eight days, PKCγ immunoreactivity was clearly increased in the spinal cord dorsal horn of these same rats. Within the spinal cord dorsal horn of morphine tolerant rats, there were significantly more PKCγ immunostained neurons in laminae I–II than in laminae III–IV and V–VI. Such PKCγ immunostaining was observed primarily in neuronal somata indicating a postsynaptic site of PKCγ increases. Moreover, both the development of morphine tolerance and the increase in PKCγ immunoreactivity were prevented by co-administration of morphine with 10 nmol MK-801 between Day 2 and Day 7 of the eight day treatment schedule. In contrast, PKCγ immunoreactivity was not increased in rats receiving a single i.t. administration of 10 μg morphine on Day 8, nor did repeated treatment with 10 nmol MK-801 alone change baseline levels of PKCγ immunoreactivity. These results provide further evidence for the involvement of PKC in NMDA receptor-mediated mechanisms of morphine tolerance.     

5-Hydroxytryptamine in the spinal cord of the domestic fowl

5-Hydroxytryptamine (5HT) immunoreactive fibers and varicosities are present in the gray and white matters of the adult domestic fowl spinal cord. These immunoreactive structures are densest in laminae I and II, the area around the central canal, and in the ventral horn. 5HT fibers and varicosities surround certain laminae I and II cells and large ventral horn cells. The apparent intimate relationship between dorsal horn cells and numerous 5HT structures may render them good models to study the possible role of 5HT in the modulation of nociception in the dorsal horn.     

Morphological and biochemical differences expressed in separate dissociated cell cultures of dorsal and ventral halves of the mouse spinal cord

The neuronal properties of separate dissociated cell cultures of dorsal and ventral halves of the embryonic mouse spinal cord (E 13.5) were investigated. Ventral-half cultures grew on a variety of substrates and in a variety of media; dorsal-half cultures required a non-neuronal feeder layer and supplemented medium for survival. The two types of cultures differed in their morphological and biochemical properties. Ventral-half neurons remained well separated on the culture plate, whereas dorsal-half neurons tended to aggregate. Lucifer yellow fills showed that ventral-half neurons were substantially larger and had more processes than dorsal-half neurons. Because of the large size and good separation of the neurons, ventral-half cultures provide an especially attractive system for electrophysiologic and morphologic studies. Ventral-half cultures were highly enriched for choline acetyltransferase (ChAT) activity and had more neurons that stained for intracellular acetylcholinesterase (AChE); dorsal-half cultures were enriched for glutamic acid decar☐ylase (GAD) activity, and high-affinity γ-aminobutyric acid (GABA) uptake. The clear differences between the two cultures indicate that many morphological and biochemical properties are already specified on embryonic day 13.5.     

Pain and rehabilitation after spinal cord injury: the case of sensory spasticity?

Sixty percent of patients with posttraumatic para- or tetraplegia suffer from severe, continuous burning and/or lancinating pain. Multiple sclerosis produces pain in more than 30%. This pain can be as important as the absent mobility or sexual function as a cause of lowered quality of life. Two unique types of longstanding neuropathic pain can be recognized in persons with spinal cord injury: (1) segmentally distributed pain at the lesion; and (2) pain in the body below the lesion, often with late onset. The first type could be produced by nerve root entrapment or by direct segmental deafferentation. The second type probably contains several forms of central pain, evoked either by the original spinal lesion, by an expanding syrinx in the spinal cord or by secondary changes at higher levels of the somatosensory systems. Patients with central pain almost always have stimulus-independent pain. Its intensity may vary independently, be related to the presence of visceral activity/inflammation or be constant. In addition, stimulus-dependent pain is sometimes present, usually because skin areas or viscera below the lesion are allodynic. Partial spinal lesions, especially centrally in the cervical spinal cord, may be more prone to produce pain than are complete lesions. There is limited analgesic effectiveness in controlled studies of serotonin reuptake inhibitors, of sodium channel blockers (lidocaine, tetracaine), of the GABA receptor agonist baclofen (one study) and of the NMDA-receptor antagonist ketamine (one study). There are anecdotal reports on oral carbamazepine, on gabapentin, on intrathecal opiates and also on the ₂-agonist clonidine, being effective in central neuropathic pain. Neurostimulation is effective only if it evokes paraesthesia in the painful area; hence TENS may give relief of segmental pain. Neurodestructive procedures and central neurostimulation have been largely unsuccessful. As in other longstanding pain, improved coping through cognitive–behavioural rehabilitation may be helpful for the clinical outcome.     

Inhibition of spinal nociceptive responses after intramuscular injection of capsaicin involves activation of noradrenergic and opioid systems

Extracellular recordings of wide dynamic range neurones in the dorsal horn driven by electrical stimulation of the sciatic nerve were performed in intact urethane-anaesthetized Sprague–Dawley rats. The electrically evoked neuronal responses were defined as A- and C-fibres responses according to latencies, and the effect of a deep nociceptive conditioning stimulus induced by 200 μg capsaicin (8-methyl-N-vanillyl-6-noneamide) injected into the contralateral gastrocnemius–soleus muscle was studied for at least 30 min. Independent of the size and location of the receptive field of the neurone under study, a clear inhibition of the neuronal responses was observed. The electrically evoked C-fibre responses were inhibited to 53% of baseline 15–30 min after injection of capsaicin. This inhibition was only slightly attenuated by 125 nmol of the α-adrenoceptor antagonist phentolamine or 250 nmol of the opioid receptor antagonist naloxone applied directly onto the spinal cord when the two compounds were administered separately 5 min before capsaicin. In contrast, when a mixture of the two compounds was given 5 min before capsaicin, the effect of capsaicin was completely abolished. These results indicate that activation of the capsaicin-sensitive afferents in the gastrocnemius–soleus muscle inhibits the electrically evoked C-fibre responses in the dorsal horn by activating noradrenergic and opioidergic inhibitory systems. Moreover, our data indicate that the activation of these two systems following injection of capsaicin has a sub-additive inhibitory effect on the wide dynamic range neurones in the spinal cord. We conclude that only one of these systems is sufficient for the inhibition to occur.     

The caudal end of the rat spinal cord: transformation to and ultrastructure of the filum terminale

Contrary to the current belief, the spinal cord of the rat does not terminate with the conus terminalis (CT), but its basic components (central canal, gray matter, white matter) continue in the filum terminale (FT). Proceeding caudally in the conus terminalis, first the motoneuron cell column discontinues in the ventral horn. More caudally the dorsal horns separate from the intermediate zone, and discontinue. The ensuing filum terminale consists of the slit-like central canal lined by ciliated ependymal cells, the periventricular gray matter and the peripheral white matter. Uniform small size neurons and glial cells populate the gray matter. Ultrastructural analysis revealed various types of axodendritic and axosomatic synapses as well as fine unmyelinated axons. The white matter consists mainly of myelinated nerve fibers. The neuronal components of the filum terminale, if they occur also in the human spinal cord, should be involved in the diagnosis and treatment of various diseases, e.g. tethered spinal cord syndrome, vascular malformations and disraphysm.     

Intense peripheral electrical stimulation evokes brief and persistent inhibition of the nociceptive tail withdrawal reflex in the rat

In a study of modulation of nociception by sensory inputs, electrical stimulation was applied to specific sites in the hindlimb and effects on the nociceptive tail withdrawal reflex were monitored in the lightly anaesthetized rat. Stimulation was applied to previously defined sites in the hindlimb, meridian points femur-futu (ST-32), fengshi (GB-31) and zusanli (ST-36). It consisted of a 4 Hz train of 2 ms square pulses given for 20 min at 20 × the threshold intensity required for muscle twitch. Tail withdrawal was provoked by application of a noxious heat stimulus applied to the tip of the tail. Results were expressed as a percentage of the maximal possible inhibition which is achieved when the post-treatment latency is 2 × the pre-treatment latency otherwise known as the cut off. During stimulation, the latency of the withdrawal increased to ≈ 70% of the maximal possible inhibition. Following stimulation, the inhibition persisted for > 1 h. Stimulation at 2 or 6 Hz elicited similar effects but stimulation at 8 Hz evoked inhibition during the stimulation only. Stimulation applied to sites away from defined meridian points inhibited tail withdrawal during the stimulation; no post-stimulation effect was produced. In acutely transected animals (≤ 48 h), stimulation of meridian points elicited a small, brief increase in latency but during stimulation only. At 7 and 14 days after spinal transection, this response during stimulation was greater in magnitude and a brief post-stimulation increase was also observed. The return of the of this latter effect was coincident with the return of bladder function. These data suggest that high intensity, low frequency electrical stimulation of hindlimb meridian points in the lightly anaesthetized rat produces both brief and persistent inhibitory effects on the nociceptive tail withdrawal reflex. These effects appear to be elicited by different mechanisms. The persistent effect may represent a plastic change in central inhibitory mechanisms. Data from spinal animals indicate a major participation of supraspinal structures but that spinal mechanisms are also capable of sustaining both types of effect.     

Structural changes of anterior horn neurons and their synaptic input caudal to a low thoracic spinal cord hemisection in the adult rat: a light and electron microscopic study

Structural changes in lumbosacral ventral horn neurons and their synaptic input were studied at 3, 10, 21, 42, and 90 days following low thoracic cord hemisection in adult rats by light microscopic examination of synaptophysin immunoreactivity (SYN-IR) and by electron microscopy. There was an ipsilateral transient decrease in SYN-IR at the somal and proximal dendritic surfaces of anterior horn neurons which extended caudally from the site of injury over a postoperative (p.o.) period of 42 days. Concomitantly, at 21 days p.o., perineuronal SYN-IR started to recover in upper lumbar segments. By 90 days p.o., a normal staining pattern of SYN was noted in upper and mid lumbar segments, but the perineuronal SYN-IR was still slightly below normal levels in low lumbar and sacral segments. Electron microscopy revealed ultrastructural changes coincident with the alterations in SYN-IR. At 3 days p.o., phagocytosis of degenerating axon terminals by activated microglial cells was observed at the somal and proximal dendritic surfaces of ventral horn neurons. These changes were most prominent up to two segments caudal to the lesion. At 10 days p.o., advanced stages of bouton phagocytosis were still detectable in all lumbosacral motor nuclei. Additionally, abnormal axon terminals, with a few dispersed synaptic vesicles and accumulations of large mitochondria, appeared at the scalloped somal surfaces of anterior horn neurons. At 21 days p.o., several large lumbosacral motoneurons had developed chromatolysis-like ultrastructural alterations and motoneuronal cell bodies had become partially covered by astrocytic lamellae. At 42 days p.o., there was a transient appearance of polyribosomes in some M-type boutons. In addition, at 42 and 90 days p.o., a few degenerating motoneurons were detected in all lumbosacral segments, but most displayed normal neuronal cell bodies contacted by numerous intact synapses as well as by astrocytic processes. In contrast to these striking alterations of synaptic input at somal and proximal dendritic surfaces of motoneurons, relatively few degenerating boutons were detected in the neuropil of motor nuclei at all the p.o. times studied. We suggest that the preferential disturbance of the predominantly inhibitory axosomatic synapses on ventral horn neurons may be involved in the mechanisms which influence the well-established increase in motoneuronal excitability after spinal cord injury.