Convergence of stress granules and protein aggregates in hippocampal cornu ammonis 1 at later reperfusion following global brain ischemia.

The delayed and selective vulnerability of post-ischemic hippocampal cornu ammonis (CA) 1 pyramidal neurons correlates with a lack of recovery of normal protein synthesis. Recent evidence implicates sequestration of translational machinery into protein aggregates and stress granules as factors underlying persistent translation arrest in CA1 neurons. However, the relationship between protein aggregates and stress granules during brain reperfusion is unknown. Here we investigated the colocalization of protein aggregates and stress granules using immunofluorescence microscopy and pair-wise double labeling for ubiquitin/T cell internal antigen (TIA-1), ubiquitin/small ribosomal subunit protein 6 (S6), and TIA-1/S6. We evaluated the rat dorsal hippocampus at 1, 2 or 3 days of reperfusion following a 10 min global brain ischemic insult. At 1 day of reperfusion, ubiquitin-containing aggregates (ubi-protein clusters) occurred in neurons but did not colocalize with stress granules. At 2 days' reperfusion, only in CA1, cytoplasmic protein aggregates colocalized with stress granules, and ubiquitin-containing inclusions accumulated in the nuclei of CA1 pyramidal neurons. Functionally, a convergence of stress granules and protein aggregates would be expected to sustain translation arrest and inhibit clearance of ubiquitinated proteins, both factors expected to contribute to CA1 pyramidal neuron vulnerability.

Immunohistochemical mapping of total and phosphorylated eukaryotic initiation factor 4G in rat hippocampus following global brain ischemia and reperfusion.

Partial proteolysis and phosphorylation of the translation initiation factor eukaryotic initiation factor 4G (eIF4G) occur in reperfused brain, but the contribution of eIF4G alterations to brain injury has not been established. A component of the complex delivering mRNA to the small ribosomal subunit, eIF4G is also found in stress granules. Stress granules sequester inactive 48S preinitiation complexes during stress-induced translation arrest. We performed double-labeling immunofluorescence histochemistry for total or ser 1108 phosphorylated eIF4G and the stress granule component T-cell internal antigen following normothermic, 10 min cardiac arrest-induced global brain ischemia and up to 4 h reperfusion in the rat. In cornu ammonis (Ammon's horn; CA) 1 at 90 min and 4 h reperfusion, eIF4G staining transformed from a homogeneous to an aggregated distribution. The number of eIF4G-containing stress granules differed between CA1 and CA3 during reperfusion. In hippocampal pyramidal neurons, phosphorylated eIF4G appeared exclusively in stress granules. Supragranular interneurons of the dentate gyrus showed a large increase in cytoplasmic eIF4G(P) following reperfusion. Immunoblot analysis with antisera against different portions of eIF4G showed a large increase in phosphorylated C-terminal eIF4G fragments, suggesting these accumulate in the cytoplasm of dentate gyrus interneurons. Thus, altered eIF4G subcellular compartmentalization may contribute to prolonged translation arrest in CA1 pyramidal neurons. Accumulation of phosphorylated eIF4G fragments may contribute to the vulnerability of dentate interneurons. Ischemia and reperfusion invoke different translational control responses in distinct hippocampal neuron populations, which may contribute to the differential ischemic vulnerabilities of these cells.

Adjuvant therapies using normobaric oxygen with hypothermia or ethanol for reducing hyperglycolysis in thromboembolic cerebral ischemia.

BACKGROUND AND PURPOSE: Normobaric oxygen (NBO), ethanol (EtOH), and therapeutic hypothermia (TH) delivered alone or in combination have neuroprotective properties after acute stroke. We used an autologous thromboembolic rat stroke model to assess the additive effects of these treatments for reducing the deleterious effects of hyperglycolysis post-stroke in which reperfusion is induced with recombinant tissue plasminogen activator (rt-PA). METHODS: Sprague-Dawley rats were subjected to middle cerebral artery (MCA) occlusion with an autologous embolus. One hour after occlusion, rt-PA was administered alone or with NBO (60%), EtOH (1.0 g/kg), TH (33 °C), either singly or in combination. Infarct volume and neurological deficit were assessed at 24h after rt-PA-induced reperfusion with or without other treatments. The extent of hyperglycolysis, as determined by cerebral glucose and lactate levels was evaluated at 3 and 24h after rt-PA administration. At the same time points, expressions of glucose transporter 1 (Glut1), glucose transporter 3 (Glut3), phosphofructokinase1 (PFK-1), and lactate dehydrogenase were (LDH) measured by Western blotting. RESULTS: Following rt-PA in rats with thromboembolic stroke, NBO combined with TH or EtOH most effectively decreased infarct volume and neurological deficit. As compared to rt-PA alone, EtOH or TH but not NBO monotherapies significantly reduced post-stroke hyperglycolysis. The increased utilization of glucose and production of lactate post-stroke was prevented most effectively when NBO was combined with either EtOH or TH after reperfusion with rt-PA, as shown by the significantly decreased Glut1, Glut3, PFK-1, and LDH levels. CONCLUSIONS: In a rat thromboembolic stroke model, both EtOH and TH used individually offer neuroprotection after the administration of rt-PA. While NBO monotherapy does not appear to be effective, it significantly potentiates the efficacy of EtOH and TH. The similar neuroprotection and underlying mechanisms pertaining to the attenuation of hyperglycolysis provided by EtOH or TH in combination with NBO suggest a possibility of substituting EtOH for TH. Thus a combination of NBO and EtOH, which are widely available and easily used, could become a novel and effective neuroprotective strategy in the clinical setting.

Prolonged translation arrest in reperfused hippocampal cornu Ammonis 1 is mediated by stress granules.

Global brain ischemia and reperfusion cause phosphorylation of the alpha subunit of eukaryotic initiation factor 2alpha, a reversible event associated with neuronal translation inhibition. However, the selective vulnerability of cornu Ammonis (CA) 1 pyramidal neurons correlates with irreversible translation inhibition. Phosphorylation of eukaryotic initiation factor 2alpha also leads to the formation of stress granules, cytoplasmic foci containing, in part, components of the 48S pre-initiation complex and the RNA binding protein T cell internal antigen-1 (TIA-1). Stress granules are sites of translationally inactive protein synthesis machinery. Here we evaluated stress granules in rat hippocampal formation neurons after 10 min global brain ischemia and 10 min, 90 min or 4 h of reperfusion by double-labeling immunofluorescence for two stress granule components: small ribosomal subunit protein 6 and TIA-1. Stress granules in CA3, hilus and dentate gyrus, but not CA1, increased at 10 min reperfusion and returned to control levels by 90 min reperfusion. Dynamic changes in the nuclear distribution of TIA-1 occurred in resistant neurons. At 4 h reperfusion, small ribosomal subunit protein 6 was solely localized within stress granules only in CA1 pyramidal neurons. Both TIA-1 and small ribosomal subunit protein 6 levels decreased approximately 50% in hippocampus homogenates. Electron microscopy showed stress granules to be composed of electron dense bodies 100-200 nm in diameter, that were not membrane bound, but were associated with endoplasmic reticulum. Alterations in stress granule behavior in CA1 pyramidal neurons provide a definitive mechanism for the continued inhibition of protein synthesis in reperfused CA1 pyramidal neurons following dephosphorylation of eukaryotic initiation factor 2alpha.

Insulin induces dephosphorylation of eukaryotic initiation factor 2alpha and restores protein synthesis in vulnerable hippocampal neurons after transient brain ischemia.

Brain reperfusion causes prompt, severe, and prolonged protein synthesis suppression and increased phosphorylation of eukaryotic initiation factor 2alpha [eIF2alpha(P)] in hippocampal CA1 and hilar neurons. The authors hypothesized that eIF2alpha(P) dephosphorylation would lead to recovery of protein synthesis. Here the effects of insulin, which activates phosphatases, were examined by immunostaining for eIF2alpha(P) and autoradiography of in vivo 35S amino acid incorporation. Rats resuscitated from a 10-minute cardiac arrest were given 0, 2, 10 or 20 U/kg of intravenous insulin, underwent reperfusion for 90 minutes, and were perfusion fixed. Thirty minutes before perfusion fixation, control and resuscitated animals received 500 microCi/kg of 35S methionine/cysteine. Alternate 30-microm brain sections were autoradiographed or immunostained for eIF2alpha(P). Controls had abundant protein synthesis and no eIF2alpha(P) in hippocampal neurons. Untreated reperfused neurons in the CA1, hilus, and dentate gyrus had intense staining for eIF2alpha(P) and reduced protein synthesis; there was little improvement with treatment with 2 or 10 U/kg of insulin. However, with 20 U/kg of insulin, these neurons recovered protein synthesis and were free of eIF2alpha(P). These results show that the suppression of protein synthesis in the reperfused brain is reversible; they support a causal association between eIF2alpha(P) and inhibition of protein synthesis, and suggest a mechanism for the neuroprotective effects of insulin.

Effect of brain ischemia and reperfusion on the localization of phosphorylated eukaryotic initiation factor 2 alpha.

Postischemic brain reperfusion is associated with a substantial and long-lasting reduction of protein synthesis in selectively vulnerable neurons. Because the overall translation initiation rate is typically regulated by altering the phosphorylation of serine 51 on the alpha-subunit of eukaryotic initiation factor 2 (eIF-2 alpha), we used an antibody specific to phosphorylated eIF-2 alpha [eIF-2(alpha P)] to study the regional and cellular distribution of eIF-2(alpha P) in normal, ischemic, and reperfused rat brains. Western blots of brain postmitochondrial supernatants revealed that approximately 1% of all eIF-2 alpha is phosphorylated in controls, eIF-2(alpha P) is not reduced by up to 30 minutes of ischemia, and eIF-2(alpha P) is increased approximately 20-fold after 10 and 90 minutes of reperfusion. Immunohistochemistry shows localization of eIF-2(alpha P) to astrocytes in normal brains, a massive increase in eIF-2(alpha P) in the cytoplasm of neurons within the first 10 minutes of reperfusion, accumulation of eIF-2(alpha P) in the nuclei of selectively vulnerable neurons after 1 hour of reperfusion, and morphology suggesting pyknosis or apoptosis in neuronal nuclei that continue to display eIF-2(alpha P) after 4 hours of reperfusion. These observations, together with the fact that eIF-2(alpha P) inhibits translation initiation, make a compelling case that eIF-2(alpha P) is responsible for reperfusion-induced inhibition of protein synthesis in vulnerable neurons.

Quantitative analysis of age-related dendritic changes in medium spiny I (MSI) striatal neurons of C57BL/6N mice.

Our study used quantitative morphometric analysis and Golgi staining methods to evaluate postnatal changes in the dendritic architecture of MSI neurons of the striatum between 1 and 30 months of age. Morphological changes and chronological age were also correlated with functional testing in order to identify subpopulations of aged mice with dendritic alterations that may be more characteristic of a motor deficit rather than the normal aging process. We found that the overall size of the dendritic arbor of MSI neurons in the rostral striatum remained stable with age, while caudal MSI neurons exhibited a significant elongation of terminal dendritic segments between 25 and 30 months of age. In addition, our correlation analysis of motor performance and chronological age found that neither striatal-motor deficits nor their associated anatomical correlates were inevitable consequences of senescence but were characteristic for a select subpopulation of aged mice with striatal-motor deficits. We found that mice that tested poorly on the balance rod had a significant increase in the number of MSI neurons with small dendritic arbors in various stages of atrophic degeneration. Conversely, 30-month-old mice that had no functional impairment showed no significant change in the number of neurons with atrophic dendrites. These data reinforce the premise that the correlation of structure and function plays an important role in the analysis of an aging population since data may vary based on the number of functionally impaired or unimpaired mice that make up an experimental group.

The phrenic nucleus of th albino rat: a correlative HRP and Golgi study.

The phrenic nucleus of the adult albino rat was studied by utilizing the O-dianisidine method for the retrograde transport of horseradish peroxidase in conjunction with the zinc chromate modification of the Golgi technique. Application of HRP to the transected phrenic nerve in the neck labeled a column of phrenic motor neurons from C3 to C5 in the ipsilateral spinal cord. However, when HRP was applied to the phrenic nerve intrathoracically, labeled neurons were found from C3 to C6. The long axis of the column of phrenic neurons was oriented tangentially from rostral to caudal poles. There was a gradual shift of the column from posterior to anterior and from lateral to medial positions in the ventral horn. The peroxidase material was also used to localize impregnated phrenic motor neurons in the Golgi sections and to provide quantitative data on phrenic motor neurons. In Golgi-impregnated material two types of phrenic neurons were distinguished on the basis of dendritic morphology and orientation. These neurons were designated (1) large neurons with smooth, radially oriented dendrites, and (2) smaller neurons with varicose, tangentially oriented dendrites. Both types of neurons had a small number of spines and bulbous appendages issuing from the dendritic trunks and branches. The dendritic fields of adjacent phrenic neurons overlapped extensively with one another and with dendrites of more distally placed ventral horn motor neurons. In peroxidase-labeled sagittal sections the dendrites of phrenic neurons were primarily oriented in the rostrocaudal plane. The mean total number of peroxidase-labeled neurons in the phrenic nucleus was 415.75 +/- 18.36 cells. In sagittal sections the mean long axis diameter of phrenic cell bodies was 34.5 micrometers. In frontal sections the mean long axis diameter of phrenic cell bodies was 22.5 micrometers. Thus, from direct measurement, the phrenic neurons were 34% longer in the sagittal plane than in the frontal plane. In the present study each phrenic nucleus contributed fibers only to the ipsilateral phrenic nerve, and no evidence for peripheral crossing of fibers was found.

The radial fibers in the globus pallidus.

In our Golgi collection of adult monkey brains the striatal efferents, i.e., the radial fibers in the globus pallidus and the "comb" bundle fibers in the internal capsule and in the cerebral peduncle, are well impregnated in the horizontally sectioned brain and in a sagittal sectioned brain. Since collaterals emerging from radial fibers are seen only in the horizontal series and not in the saggittal series, the interpretation is that they proceed anteriorly and posteriorly only, following the curvature of the pallidal segments, and do not run superiorly or inferiorly as they emerge. Although radial fibers emitting collaterals in the lateral segment and in the medial segment of the globus pallidus have been observed, it has not been possible to observe the same radial fiber emitting collaterals in both pallidal segments and the prospects of ever doing so are not good. The radial fibers converging in the globus pallidus pursue many radii and there is little coincidence between the plane of section and the planes in which they travel. At most only severed radial fiber segments 100-150 microns in length can be found in the horizontal sections needed to observe the collaterals. Moreover, sagittal sections trodorsally, as they pass through the internal medullary lamina to enter the medial segment of the globus pallidus. The radial fibers in the medial segment of the globus pallidus are continuous with the "comb" bundle fibers and appear to be thinner than the radial fibers in the lateral segment of the globus pallidus. It is not proved; nonetheless, the view expressed here is that the radial fibers are thinner in the medial segment of the globus pallidus because they may be the same fibers that gave off collaterals in the lateral segment of the globus pallidus. This is discussed in the light of the electrophysiological disclosure of Yoshida et al. ('71, '72) that caudatopallidal fibers are collaterals off caudatonigral fibers. The afferent plexuses of fine, "bouton en passage" fibers, which completely ensheath the long radiating dendrites in the globus pallidus (Fox et al., '66) are well impregnated in the horizontal series. Obviously, they are formed by a number of ultimate branches converging from the collateral brances of a number of different radial fibers. The divergence, too, in this system must be considerable; however, its true extent can only be surmised from the several radial fibers and radial fiber collaterals seen in the incompletely impregnanted Golgi section. Continued.

The neurons in the primate subthalamic nucleus: a Golgi and electron microscopic study.

In Golgi preparations of the adult monkey (Macaca mulatta) local interneurons and two varieties of principal neurons, radiating and elongated fusiform, are found in the subthalamic nucleus. The cell bodies of the radiating neurons have a few delicate, somatic spines some of which are occasionally bilobed and trilobed. Five to eight dendritic trunks give rise to branching, tapering dendrites, which may extend for over 400 microns. These dendrites are much thinner than the dendrites in the globus pallidus and the substantia nigra. Some neurons have many and some neurons have few dendritic spines. When numerous the dendritic spines are concentrated on the dendritic trunks and proximal dendrites. The relatively few elongated fusiform neurons are found not only in the capsule but also in the center of the nucleus. Most dendrites emerge from the opposite poles of their smooth surfaced cell bodies. They have a few dendritic spines. Some of these dendrites extend for more than 750 microsn. In 1-micron thick plastic sections lipofuscin granules are present in some but not all principal neuron cell bodies of the monkey (Macaca mulatta); but these granules are present in all principal neuron cell bodies of the pig-tail monkey (Macaca nemestrina) and of the squirrel monkey (Saimiri sciureus). The local interneurons have small cell bodies and a few relatively long undulating dendrites. The dendrites have bulbous dendritic appendages of varying complexity and beaded axon-like processes. The dendritic appendages and axon-like processes are more numerous distally and on the distal ends of the dendrites they form complex entanglements. Axons coming from the cell body have not been observed. The cell bodies of the local interneurons are identified in cresyl violet stained sections of the monkey (Macaca mulatta), in 1-micron thick plastic sections and electron micrographs of the squirrel monkey (Saimiri sciureus). They have relatively large nuclei surrounded by a thin rim of cytoplasm rich in polyribosomes.

A Golgi study of the monkey paraventricular nucleus: neuronal types, afferent and efferent fibers.

The neuronal organization of the paraventricular nucleus (PVN) was examined in Golgi impregnations of adult monkey. Results showed that at least six types of neurons could be identified in the nucleus on the basis of morphological features of the somata, dendrites, and axons. Four types of neurons with sparse to densely spined cell bodies and dendrites exhibited long axons and included large neurons (types I and II), medium-sized to large neurons (type III), and small to medium-sized cells (type IV). Axons of type I, III, and IV neurons had different diameters and were followed out of the PVN. Axon collaterals that arborized within the PVN were seen on the axons of types III and IV cells. Two types of interneurons with small somata were also found. One (type V) exhibited varicose dendrites and a profusely arborizing local axon. The other cell (type VI) had recurved dendrites with long appendages and no impregnated axon. Afferent fibers were also identified. Type 1 was a fine-caliber axon that coursed long distances in the PVN and exhibited numerous short branches. Additional observations suggested that type 1 afferents originated from the stria terminalis. The other afferent axon (type 2) was thicker and gave rise to terminal arborizations containing clusters of small swellings. The efferent fibers of the PVN were also examined in impregnations of the paraventriculosupraopticohypophysial tract. Fibers formed an extensive plexus as they coursed ventrally and passed through the lateral hypothalamus. Axons coursing more laterally in the tract were much larger than those more medially located. Our findings show a diverse organization of neuronal types within the monkey PVN with evidence for intrinsic connections through axon collaterals of efferent neurons and the locally arborizing axons of interneurons. Correlations are proposed between morphological subtypes of neurons seen in this Golgi study and the known functional output pathways of the PVN.

Golgi study of the mouse striatum: age-related dendritic changes in different neuronal populations.

The Van der Loos modification of the Golgi-Cox method and morphometric analyses were used to study the neuronal types in the striatum of adult (3, 6, and 10 months) and aged (20, 25, and 30 months) C57BL/6N mice. In adult mice six types of striatal neurons were distinguished primarily on the basis of the morphology of their cell body and dendrites. Each of these types was compared with morphologically similar neurons from previous Golgi classifications in other species and discussed within the framework of recent immunocytochemical work. With similar methods the age-related changes occurring on the dendrites of three of the six striatal types were also analyzed. In the medium-sized neuron with spine-laden dendrites, various dendritic tree shapes and sizes were distinguished in all age groups studied. Qualitative observations as well as measurements of total dendritic length per cell suggested that the dendrites in this type may both grow and regress throughout the life span, although signs of dendritic atrophy and regression were observed only in the aged groups. In the other two types of neuron, one a medium aspiny cell with thin varicose dendrites and the other a large spiny neuron with many dendrites, measurements of total dendritic lengths revealed sustained growth of the tree well into advanced age, followed by moderate regression in the oldest groups. The present findings also indicate that the dendrites of each type of striatal neuron follow unique temporal patterns of growth and regression during the life span of the mouse.

Ultrastructural changes in olfactory receptor neurons following olfactory nerve section.

Unilateral olfactory nerve section was performed in the salamander, Ambystoma tigrinum. An ultrastructural study was performed to investigate the changes occurring during degeneration and replacement of the mature olfactory receptor neurons. Experimental and contralateral control tissues were examined following postoperative survival periods ranging from 12 hours to 90 days. Normal bipolar receptor neurons have a fusiform cell body containing a thin rim of cytoplasm and an ovoid nucleus with a characteristic "checkerboard" chromatin pattern. A single apical dendrite projects to the surface of the epithelium, where numerous cilia extend from its apex into the overlying mucus. A single, unmyelinated, unbranching axon originates at the basal pole of the cell. After nerve section, retrograde degeneration of the mature neurons occurs. Early degenerative changes include pronounced condensation of the nuclear chromatin, increased number of nuclear membrane infoldings, and dilation of the space between the membranes of the nuclear envelope. At a later stage, the cytoplasm of the cell increases in volume and its organelle systems break down, resulting in accumulation of various forms of cell inclusions. Subsequently, proliferation of cells in the basal region of the epithelium occurs. Between 3 week and 2 months following nerve section, these cells differentiate into mature neurons. By 3 months, neurons within the epithelium have resumed their normal ultrastructure. Correlation of the time course of the ultrastructural changes with previously reported neurophysiological studies indicates that neuronal activity of the epithelium is dependent upon the presence of fully differentiated olfactory receptor neurons.

Neuronal and glial changes in the rat phrenic nucleus occurring within hours after spinal cord injury.

The present study describes specific morphological changes in the normal ultrastructure of the rat phrenic nucleus which occur within 4 hours after an ipsilateral spinal cord hemisection rostral to the nucleus. Phrenic neurons were identified at EM levels by retrograde HRP labeling. Ultrastructural features of the phrenic nucleus in uninjured animals and at 4 hours and 1, 2, and 4 days after injury were qualitatively analyzed and then quantitated with a computerized morphometric system. Our results indicated that by 4 hours posthemisection, there was a significant increase in the number of double synapses. Furthermore, the number of double synapses remained significantly higher than normal at all the other posthemisection periods. A significant increase in the length of dendrodendritic membrane appositions was also noted as early as 4 hours posthemisection. The mean normal appositional length of 1.42 +/- 0.09 microns increased to 1.89 +/- 0.12 microns at 4 hours and further increased to 2.20 +/- 0.20 microns by 1 day posthemisection. The increase in the length of membrane appositions was most likely due to an active retraction of astroglial processes from their normal position in between the dendrites. Although there was an increase in the mean length of the dendrodendritic appositions, the mean percentage of the appositions (expressed as the total number of appositions divided by the total number of dendrites in each sample) was not increased significantly over normal values during the early posthemisection periods. By 2 and 4 days posthemisection, however, the percentage of dendrodendritic appositions increased to significantly higher values than normal. Normally, 4.68 +/- 0.69% of the dendrites in the phrenic nucleus were found to be in apposition, and this number increased significantly to 7.27 +/- 1.06% by 2 days and 7.46 +/- 0.79% by 4 days posthemisection. At these later posthemisection periods, the mean length of the appositions decreased to levels which were no longer significantly higher than normal. A distribution analysis of the length of each dendrodendritic apposition in both the normal and spinal hemisected rats showed that there were more dendrodendritic appositions in the phrenic nucleus at the later posthemisection periods. It also showed that their mean length was decreased because many of the new appositions were relatively short. The above neuronal and glial alterations of the phrenic nucleus have never before been described as a response to injury of the mammalian spinal cord. Furthermore, the possibility that the above changes could represent the morphological substrate for the unmasking of functionally ineffective synapses in ou

Computer measurements of axis cylinder diameters of radial fibers and "comb" bundle fibers.

In electron micrographs of the squirrel monkey (Saimiri sciureus) brain in the striatal efferents were observed at two different levels in their course: (1) in cross-sectioned radial fiber bundles just before they enter the globus pallidus; (2) in cross-sectioned "comb" bundle fibers just before they enter the substantia nigra. In the radial bundles nearly all of the fibers are myelinated; in the "comb" bundle most are unmyelinated. The polarity of all the "comb" bundle fibers is descending. None of them degenerate following a large lesion in the substantia nigra but they do degenerate following a large lesion in the striatum. Also following this latter lesion the endings with large synaptic vesicles, which make up most of the endings in the globus pallidus and the substantia nigra, degenerate. For computer measurements, electron micrographs of the radial bundle were enlarged photographically to a final magnification of 20,000; those of the "comb" bundle to times 50,000. Measurements of 1309 radial fibers revealed a mean axis-cylinder diameter of 0.68 microns, and measurements of 749 unmyelinated "comb" bundle fibers gave a mean axis-cylinder diameter of 0.21 microns. Myelinated fibers were not included in the "comb" bundle measurements because it contains myelinated fibers of pallidal origin in addition to myelinated fibers of striatal origin. The results here indicate that the striatal efferents undergo a decided decrease in axis-cylinder diameter during their transit through the globus pallidus. It is suggested that the large non-spine bearing neurons in the striatum are the source of the striatal efferents and that they send their axons into the substantia nigra and enroute spend a great quantity of their axoplasm by extending extensive collaterals in both segments of the globus pallidus. This could account for the decreased caliber of the striatal efferents in the "comb" bundle and other findings in the striatum, globus pallidus and substantia nigra.

Organization of the zona incerta in the macaque: a Nissl and Golgi study.

The zona incerta has been implicated in the control of the initiation of saccadic eye movements in the primate. Complex interactions within the zona incerta must take place to integrate its varied inputs and to produce a coherent efferent signal in order for this function to occur. However, whether the anatomical substrates exist within the zona incerta to allow this integration to take place has not been established. The zona incerta in monkeys (Macaca mulatta) was examined in frontally, horizontally, and sagittally sectioned preparations stained for Nissl, myelinated fibers, or cytochrome oxidase, or impregnated by the Golgi technique. This nucleus can be separated into dorsal and ventral laminae on the basis of staining and morphological differences between these two subdivisions. Neurons are more densely packed, more darkly stained, and larger in the ventral lamina. In addition, the neuropil of the ventral lamina is much more intensely stained after cytochrome oxidase histochemistry. Two neuronal types, principal cells and interneurons, were identified on the basis of neuronal cell body, dendritic, and axonal features in Golgi-impregnated preparations. Principal cells have fusiform or polygonal somata (long axis from 18 to 40 microns) and dendrites that extend for up to 750 microns within the lamina in which the cell bodies are located. Putative local interneurons have small (12-16 microns), round or oval cell bodies with wavy dendrites (up to 400 microns). Numerous multilobed appendages and axon-like processes originate from these dendrites and make apparent contacts with other interneurons or with dendrites of principal cells. Dendrites of most neurons in both laminae are oriented preferentially along the principal axis, dorsolateral-to-ventromedial, of the nucleus. Therefore, within the limits of light microscopy, the zona incerta appears to possess the morphological heterogeneity to form complex intrinsic interactions. These interactions are hypothesized to form the integrative substrate for the large array of incertal inputs that are utilized to produce an efferent signal involved in the initiation of saccadic eye movements.

Morphological study of the rostral interstitial nucleus of the medial longitudinal fasciculus in the monkey, Macaca mulatta, by Nissl, Golgi, and computer reconstruction and rotation methods.

We have studied the morphology of silver-impregnated neurons (rapid Golgi technique) in the rostral interstitial nucleus of the medial longitudinal fasciculus (riMLF), a center involved in the control of vertical and torsional saccadic eye movements. This morphological study of riMLF neurons in the rhesus monkey was undertaken to further our understanding of the functional circuitry of the oculomotor system. Our study employed Nissl, Golgi, and computer-assisted methods. The cytoarchitectonic boundaries of the riMLF and its relationships to neighboring structures were determined in both Nissl and Golgi preparations. Five (I-V) distinct morphological types of riMLF neurons were distinguished in the Golgi impregnations on the basis of soma size, dendritic size, numbers of primary dendrites, number of dendritic branch points, as well as form, number, and distribution of dendritic appendages. Type I neurons impregnated most frequently and had the most extensive and highly branched dendritic tree. Type II neurons displayed thick dendrites with complex dendritic appendages, but the dendritic tree was much more compact than that of type I cells. Type III and type V cells had fusiform somas and relatively unbranched dendritic trees but differed greatly in size as well as dendritic morphology. The type IV cell was the smallest neuron and had many characteristics of the local interneurons found in other thalamic, subthalamic, hypothalamic and midbrain centers. The type V was the largest neuron, least frequently impregnated, and found only at rostral riMLF levels. Digitized reconstructions of each type of neuron were rotated by the computer, which revealed that the dendritic trees of types I, III, and V occupy a disk-like compartment in the riMLF neuropil. In contrast, the tree of types II and IV occupy a roughly spherical compartment. We suggest that three of the cell types are well suited for specific purposes: type II cells for receiving topographically organized inputs that contain spatial information, type I cells for short-lead burst neuron output to the motor neurons or other premotor centers, and type IV cells for inhibitory inputs to type I cells.

Intermediate and deep layers of the macaque superior colliculus: a Golgi study.

We studied the intermediate and deep layers of the macaque superior colliculus by means of the Golgi technique in an attempt to better understand the structural features of this important oculomotor center. For this study, we examined the optic (stratum opticum, SO), intermediate gray (stratum griseum intermedium, SGI), intermediate white (stratum album intermedium, SAI), and deep gray (stratum griseum profundum, SGP) layers. These are the four layers in which neurons having saccade-related activity are localized. We identified eight neuronal types on the basis of differences in somatic and dendritic morphologies: large multipolar neurons (Type I); large pyramidal neurons (Type II); large fusiform neurons (Type III); medium fusiform neurons with spiny, radially oriented dendrites (Type IV); medium round neurons with fan-shaped dendritic trees (Type V); medium stellate neurons with varicose dendrites (Type VI); medium multipolar neurons with robust, spiny dendrites (Type VII); and local interneurons (Type VIII). Most neuronal types possessed features that are homologous to presynaptic dendritic features in other brain centers. With the exception of the medium stellate neurons (Type VI), which are aspinous, and the local interneurons (Type VIII), which are sparsely spinous, all other types had a moderate number of spines on their dendrites. Dendrites that terminated in the optic layer had specializations not observed elsewhere, suggesting that these tips may sample a tectal afferent that is not present in the more ventral layers. These eight types comprise all the neuronal morphologies observed in a large number of Golgi-impregnated macaque brains (n = 50). We suggest that they represent the full range of neuron types in the saccade-related layers of the macaque tectum.

Sources of endothelin-1 in hippocampus and cortex following traumatic brain injury.

Endothelin 1 (ET-1) exerts normally a powerful vasoconstrictor role in the control of the brain microcirculation. In altered states, such as following traumatic brain injury (TBI), it may contribute to the development of ischemia and/or secondary cell injury. Because little is known of ET-1's cellular compartmentalization and its association to vulnerable neurons after TBI, we assessed its expression (both mRNA and protein) in cerebral cortex and hippocampus using correlative in situ hybridization and immunocytochemical techniques.Sprague-Dawley male rats were killed at 4, 24 or 48 h after TBI (450 g from 2 m, Marmarou's model). Semiquantitative analysis of our in situ hybridization results indicated a 2.5- and a 2.0-fold increase in ET-1 mRNA content in the hippocampus and cortex respectively which persisted up to 48 h post TBI. At 4 and 24 h after TBI enzyme-linked immunosorbent assay showed a tendency for increased ET-1 synthesis. In animals subjected to TBI, qualitative immunocytochemical analysis revealed a shift in ET-1 expression from astrocytes (in control animals) to endothelial cells, macrophages and neurons. Astrocytes and macrophages were identified unequivocally by using double immunofluorescence revealing ET-1 and glial fibrillary acidic protein or ED-1, respectively, the markers being specific for these cellular types. While this redistribution was most prominent at 4 and 24 h post TBI, at 48 h the endothelial cells remained strongly ET-1 immunopositive. The results suggest that cellular types which in the intact animal synthesize little or no ET-1 provide novel sources of the peptide after TBI. These sources may contribute to the sustained cerebrovascular hypoperfusion observed post TBI.

Motor balance and coordination training enhances functional outcome in rat with transient middle cerebral artery occlusion.

The goal of this study was to determine if relatively complex motor training on Rota-rod involving balance and coordination plays an essential role in improving motor function in ischemic rats, as compared with simple locomotor exercise on treadmill. Adult male Sprague-Dawley rats with (n=40) or without (n=40) ischemia were trained under each of three conditions: (1) motor balance and coordination training on Rota-rod; (2) simple exercise on treadmill; and (3) non-trained controls. Motor function was evaluated by a series of tests (foot fault placing, parallel bar crossing, rope and ladder climbing) before and at 14 or 28 days after training procedures in both ischemic and normal animals. Infarct volume in ischemic animals was determined with Nissl staining. Compared with both treadmill exercised and non-trained animals, Rota-rod-trained animals with or without ischemia significantly (P<0.01) improved motor performance of all tasks except for foot fault placing after 14 days of training, with normal rats having better performance. Animals trained for up to 28 days on the treadmill did not show significantly improved function. With regard to foot fault placing task, performance on foot placing was improved in ischemic rats across the three measurements at 0, 14 and 28 days regardless of training condition, while the normal group reached their best performance at the beginning of measurement. No significant differences in infarct volume were found in rats trained either with Rota-rod (47+/-4%; mean+/-S.E.), treadmill (45+/-5%) or non-exercised control (45+/-3%). In addition, no obvious difference could be detected in the location of the damage which included the dorso-lateral portion of the neostriatum and the frontoparietal cortex, the main regions supplied by the middle cerebral artery. The data suggest that complex motor training rather than simple exercise effectively improves functional outcome.