Structural changes of the soleus and the tibialis anterior motoneuron pool during development in the rat.

The morphological development of motoneuron pools of two hindlimb muscles of the rat, soleus (SOL) and tibialis anterior (TA), was studied in rats ranging in age between 8 and 30 postnatal days (P8-P30). Motoneurons were retrogradely labelled by injecting a cholera toxin B subunit solution directly into the muscles. This resulted in extensive labelling of motoneurons as well as their dendritic trees. The distribution of cross sectional areas of neuronal somata was determined for both muscles at various ages. Somal size increased considerably between P8 and P12, whereas growth was moderate between P12 and P20. The size distribution of SOL motoneurons was bimodal from P20, whereas the size distribution of TA motoneurons remained largely unimodal. The morphological development of the dendritic tree was studied qualitatively. The development of dendritic arborization within the SOL and the TA motoneuron pool showed major differences. The arborization pattern of dendrites of TA motoneurons was basically multipolar at all ages. In contrast, dendrites of SOL neurons tended to line up with the rostro-caudal axis and became organized in longitudinal bundles from P16 onwards. The relatively late appearance of dendrite bundles in the soleus motoneuron pool suggests that they might be related to the fine-tuning of neuronal activity rather than patterning of motor activity. The occurrence of dendrite bundles in SOL and not in TA motoneuron pools suggests that they may be related to the different afferent organization of this postural muscle or to its tonic activation pattern.     

Membrane area and dendritic structure in type-identified triceps surae alpha motoneurons

The size and branching structure of the dendritic tree were studied in nine type-identified triceps surae alpha-motoneurons that were labeled intracellularly with horseradish peroxidase and reconstructed from serial sections in the light microscope. The average total membrane area (AN) for motoneurons of type S (slow-twitch) motor units was about 22% smaller than AN for cells of type F units (including both FF and FR motor unit types in this category) (480.1 X 10(3) microns 2 vs. 617.7 X 10(3) microns 2, respectively). Systematic correlations were found between stem dendrite diameter and three measures of dendritic size: dendrite membrane area, combined dendritic length, and number of terminations. All of these correlations were significantly different for the dendrites of F and S motoneurons. Power-function relations between stem diameter and dendritic membrane area were used to estimate AN for a sample of 79 type-identified motoneurons. Mean estimated AN values were significantly different for the F and S motoneuron groups, despite a large overlap in AN values between these groups. The branching structure of dendrites of F and S motoneurons also showed clear differences. Type S motoneuron dendrites showed less-profuse branching and a more-even radial distribution of branch points than found in type F cells. Examination of two forms of the "3/2 power rule" for the relation between the diameters of parent and daughter dendritic branches at branch points showed that the dendrites of type S motoneurons conform less well with the anatomical constraints necessary to represent binary branching trees as equivalent cylinders than do dendrites of type F cells. There was no systematic difference between F and S motoneuron dendrites in the degree of asymmetry of first-order daughter trees. The results overall indicate that the dendrites of F and S motoneuron groups are structurally different, giving rise to a systematic difference in AN between these groups. Such structural differences suggest that the F and S groups of alpha-motoneurons can be viewed as intrinsically distinct cell types and not just large vs. small variants of the same cell species.     

A quantitative analysis of the geometry of cat motoneurons innervating neck and shoulder muscles

The geometry of the somata and dendritic trees of motoneurons innervating neck and shoulder muscles was investigated by using intracellular injections of HRP. In general, these motoneurons did not belong to a homogeneous population of motoneurons. Differences in average primary dendritic diameter, number of primary dendrites, and other measures of dendritic tree size were found between different neck and shoulder motoneuron groups. Several indices of proximal dendritic tree size (number of primary dendrites, sum of dendritic diameters, Rall's dendritic trunk parameter, and the sum of dendritic holes) were weakly correlated with the diameter or surface area of the soma. Some of these correlations depended on the muscle supplied by the motoneuron. The total combined dendritic length ranged from 66,660 to 95,390 microns. There was a weak, but positive, correlation between the diameter of primary dendrites and combined dendritic length. This relationship varied from motoneuron to motoneuron. The diameters of all dendrites of three trapezius motoneurons were examined in detail. The total dendritic surface area examined ranged from 415,000 to 488,100 microns 2 and represented approximately 99% of the total neuronal surface area. Last-order dendrites showed a high degree (39.9%) of taper. Dendritic tapering, by itself, was a major factor in the decrease of the (sum of dendritic diameters)3/2 measured at progressively distal sites from the soma. Although few parent and daughter dendrites obeyed the "three-halves law," the average exponent was 1.57. The diameters of primary dendrites and dendritic surface area were weakly correlated. The correlation between dendritic diameter and combined dendritic length or surface area improved if the weighted average of the diameter of second-order dendrites was used as a measure of dendrite size. Second-order dendrites, whose branches terminated in different regions of the spinal cord, showed different relationships between dendritic diameter and combined dendritic length or surface area. Comparisons between the motoneurons examined in the present study and motoneurons innervating other muscles indicate that, although all spinal motoneurons share several common features (e.g., long dendrites, dendritic tapering), each motoneuron group has a set of unique features (e.g., soma shape, relationship between primary dendrite diameter and dendritic surface area). Thus, the rules governing motoneuron dendritic geometry are not fixed but depend on the species of the motoneuron.     

The dendritic extent of motoneurons in frog brachial spinal cord: a computer reconstruction of HRP-filled cells. With comments on dendritic reconstruction methodologies

A lateral and a medial motoneuron in the brachial spinal cord of the leopard frog, Rana pipiens, were labeled by horseradish peroxidase applied to the ventral root. Their dendritic trees were traced, analyzed, and plotted using a computer-microscope system. Some dendrites of the medial motoneuron crossed the midline of the spinal cord, but no dendrites of the lateral motoneuron crossed the midline. Nevertheless, the total dendritic length of the lateral motoneuron exceeded that of the medial motoneuron. The peak number of dendritic branch segments of the medial motoneuron was located at a greater distance from its soma than that of the lateral motoneuron. Three-dimensional reconstruction and rotation of the dendritic trees revealed that the dendrites of the medial motoneuron had a greater rostrocaudal extent than those of the lateral motoneuron. When compared to reports of Golgi-impregnated motoneurons, our results suggest that the HRP technique labels dendrites more completely. However, use of the HRP technique may introduce greater errors in the subsequent measurement of dendritic segments due to nonuniform tissue shrinkage.     

Morphologic analysis and classification of ganglion cells of the chick retina by intracellular injection of Lucifer Yellow and retrograde labeling with DiI

Retinal ganglion cells (RGCs) of chicks were labeled by using the techniques of intracellular filling with Lucifer Yellow and retrograde axonal labeling with carbocyanine dye (DiI). Labeled RGCs were morphologically analyzed and classified into four major groups: Group I cells (57.1%) with a small somal area (77.5 microm(2) on average) and narrow dendritic field (17,160 microm(2) on average), Group II cells (28%) with a middle-sized somal area (186 microm(2)) and middle-sized dendritic field (48,800 microm(2)), Group III cells (9.9%) with a middle-sized somal area (203 microm(2)) and wide dendritic field (114,000 microm(2)), and Group IV cells (5%) with a large somal area (399 microm(2)) and wide dendritic field (117,000 microm(2)). Of the four groups, Groups I and II were further subdivided into two types, simple and complex, on the basis of dendritic arborization: Groups Is, Ic, and Groups IIs, IIc. However, Group III and IV showed either a simple or complex type, Group IIIs and Group IVc, respectively. The density of branching points of dendrites was approximately 10 times higher in the complex types (18,350, 6,190, and 3,520 points/mm(2) in Group Ic, IIc, and IVc, respectively) than in the simple types (1,890, 640, and 480 points/mm(2) in Group Is, IIs, and IIIs). The branching density of Group I cells was extremely high in the central zone. The chick inner plexiform layer was divided into eight sublayers by dendritic strata of RGCs and 26 stratification patterns were discriminated. The central and peripheral retinal zones were characterized by branching density of dendrites and composition of RGC groups, respectively.     

Three-dimensional architecture of dendritic trees in type-identified alpha-motoneurons

We have studied the spatial distribution of dendrites of type-identified triceps surae alpha-motoneurons, labeled intracellularly with HRP, using a variety of analytical approaches that were designed to quantify the ways in which dendrites occupy three-dimensional space. All of the methods indicated a strong tendency for motoneuron dendrites to project radially. However, regions dorsal and ventral to the somata contained fewer dendritic elements, and less membrane area, than expected for complete radial symmetry. Individual dendrites projecting into these regions tended to be smaller than those projecting rostrocaudally or mediolaterally. Nevertheless, the center of mass of membrane area for five of six fully analyzed cells was within 100 micron of the soma and, in all six cells, was located in the same dorsoventral plane as the cell soma. Maps of the projection of dendritic branches onto concentric shells at various radial distances from the soma showed that some regions have high concentrations of branches, sometimes with considerable overlap between branches arising from different stem dendrites, while other regions have relatively few branches, or none at all. Each motoneuron exhibited a different pattern of projection and there were no systematic differences between fast-twitch (type F, including both types FF and FR units) and slow-twitch (type S) motoneurons evident in the patterns of dendritic concentration. Assessment of the three-dimensional territories of individual dendrites showed that dendrites with larger numbers of terminal branches tended to have larger spatial territories. Despite considerable scatter, the results suggest that the density of branches tends to be approximately the same in large and small dendrites, and in F and S cell groups. The results are discussed in relation to the spatial location of synaptic input to motoneurons.     

Structural and physiological properties of connections between individual reticulospinal axons and lumbar motoneurons of the frog

Although the direct, monosynaptic influence of brainstem projections onto motoneurons is well-known, detailed morphological studies on the synaptic contact systems and a correlation with their functional properties are largely lacking. In this work, 43 pairs, each formed by a reticulospinal fiber contacting a lumbar motoneuron, were identified and studied electrophysiologically. Four of these were successfully labeled intracellularly with horseradish peroxidase (HRP) or neurobiotin and reconstructed using a computer-assisted camera lucida with high resolution. The mean amplitude of excitatory post-synaptic potentials (EPSPs) recorded in these four pairs varied from 100 to 730 microV, spanning most of the range obtained for all pairs (70-1,200 microV; mean +/- SD: 400 +/- 250 microV). Between two and four collaterals of reticulospinal axons established 4-19 close appositions with a labeled motoneuron. Mean distance from the origin of each collateral to any bouton on that collateral was 566-817 microm. A presynaptic action potential must pass 11 branch points on average to reach it. Similarly, the boutons presumably contacting motoneurons were on average 558-624 microm (9-11 branch points) from the origin of the collateral. The distributions of diameters of all boutons and those making putative contacts with stained motoneurons were very similar. The dendritic surface of stained motoneurons was symmetrically distributed along the rostrocaudal axis with more than half the surface being more than 500 microm from the soma. However, the contacts from reticulospinal axons were concentrated ventromedially, 262-356 microm (range of average values for four connections) from the motoneuron soma, in some instances on very proximal dendritic segments. Thus, the location and size of putative contacts in relation to axonal collaterals was not distinguishable from location and size of other boutons, but they occupied specific positions on dendrites of lumbar motoneurons. The number of contacts formed by a reticulospinal axon on a motoneuron in a particular location could be described as the product of the available dendritic surface and the total number of presynaptic boutons in this region. Compartmental models of the reconstructed motoneurons were created, and currents with the time course of an alpha function were injected at the sites of these putative contacts. Despite the restricted volume occupied by contacts from a single fiber, a high variability of their contributions to somatic EPSPs owing to electrotonic attenuation was shown: The coefficient of variation of quantal responses was estimated to be between 60% and 120%, comparable to the variability of the path distance between contacts and soma (50-90%).     

Morphological differences between wild-type and transgenic superoxide dismutase 1 lumbar motoneurons in postnatal mice

Quantitative analysis of the dendritic arborizations of wild-type (WT) and superoxide dismutase 1 (SOD1) postnatal mouse motoneurons was performed following intracellular staining and 3D reconstructions with Neurolucida system. The population of lumbar motoneurons was targeted in the caudal part of the L5 segment, and all labeled motoneurons were located within the same ventrolateral pool. Despite the similar size of the soma and the mean diameter of primary dendrites, the dendritic arborizations of the WT and SOD1 motoneurons showed significant differences in terms of their morphometric parameters. The metric and topological parameters of dendrites show that the total dendritic length and surface area and total number of segments, branching nodes, and tips per motoneuron were significantly higher in SOD1 motoneurons. Our main finding concerns a proliferation of dendritic branches starting at about 100 microm from the soma in the SOD1 motoneurons. However, the longest and mean dendritic paths from soma to terminations were similar, giving a comparable envelope of the dendritic fields. Indeed, the SOD1 motoneurons were larger as a result of abnormal branching. The results suggest that a defect in pruning mechanisms occurs during this developmental period. The abnormal growth of the dendritic arborizations and the reduced excitability of postnatal SOD1 motoneurons could be a neuroprotective response and would represent an early compensatory mechanism against the activity-induced toxicity.     

The number and size of motoneurons in the soleus motor nucleus of the normal and dystrophic (C57BL/6J dy2J/dy2J) mouse

The method of retrograde axonal transport of horseradish peroxidase (HRP) was used to identify the motoneurons that innervate the soleus muscle in normal and dystrophic (C57BL/6J dy^2J/dy^2J) mice. In both normal and dystrophic animals the soleus motor nucleus was located in spinal segments L3 and L4, in a medial position in the lateral division of lamina IX. Measurements of the motoneuron soma area of HRP-labeled cells in normal and dystrophic animals showed that motoneuron size was bimodally distributed in both cases but that mean soma areas of motoneurons of both the large and small cell component from the dystrophic animals were larger. The number of alpha motoneurons was reduced in the dystrophic animals compared with the normal animals, the mean number of alpha motoneurons being 20 in the case of dystrophic animals and 25 in the case of normal animals.     

The number and size of motoneurons in the soleus motor nucleus of the normal and dystrophic (C57BL/6J dy/dy) mouse

The method of retrograde axonal transport of horseradish peroxidase (HRP) was used to identify the motoneurons that innervate the soleus muscle in normal and dystrophic (C57BL/6J dy^2J/dy^2J) mice. In both normal and dystrophic animals the soleus motor nucleus was located in spinal segments L3 and L4, in a medial position in the lateral division of lamina IX. Measurements of the motoneuron soma area of HRP-labeled cells in normal and dystrophic animals showed that motoneuron size was bimodally distributed in both cases but that mean soma areas of motoneurons of both the large and small cell component from the dystrophic animals were larger. The number of alpha motoneurons was reduced in the dystrophic animals compared with the normal animals, the mean number of alpha motoneurons being 20 in the case of dystrophic animals and 25 in the case of normal animals.     

Reduced branching and length of dendrites detected in cervical spinal cord motoneurons of Wobbler mouse, a model for inherited motoneuron disease.

The Wobbler mouse (wr) has been proposed as a model for human inherited motoneuron disease (infantile spinal muscular atrophy). The primary defect is thought to be in the motoneurons. Therefore we undertook a survey of the qualitative and quantitative changes occurring in the cervical spinal motoneurons of Wobbler mice during a late stage of the motoneuron disease compared with age- and sex-matched normal phenotype (NFR/wr) littermates. The Rapid Golgi Method was applied. In control and Wobbler mice, four types of neurons were identified according to their dendritic patterns: multipolar, tripolar, bipolar, and unipolar cells. Unipolar cells were observed more often in the Wobbler specimens than the controls and may represent a final stage in the degeneration of other cell types with greater numbers of primary dendrites. Medium (300-999 microns 2) and large (greater than 1,000 microns 2) impregnated neurons (presumably alpha-motoneurons) showed strong indications of cell degeneration, including statistically significant reductions in the measurements for dendritic length, distribution, and branching, as well as the number of spines. In contrast, the small (less than 300 microns 2) neurons showed only mild signs of degeneration, including slight reductions in dendritic length, but no significant differences appeared in the distribution and branching of dendrites, or in the number of spines. Instead, a small increase could be detected in the number of primary and secondary dendritic branches emanating from the small neurons, as well as in the number of dendritic spines. These findings suggest that sprouting may occur to a slight extent. Although previous studies document that swelling with subsequent vacuolation of motoneurons is the predominant feature characterizing the Wobbler disease, the mean soma area (microns 2) calculated for the impregnated neurons of the Wobbler specimens showed no significant difference from the controls. It is hypothesized that the advanced signs of the Wobbler motoneuron disease are primarily reflected in the degeneration of the dendrites and spines on the medium and large alpha-motoneurons. The small neurons (presumably a mixed population of gamma-motoneurons, interneurons, and Renshaw cells) possess dendrites and spines that seem to be less affected, and instead show signs of sprouting.     

The Extent of the Dendritic Tree and the Number of Synapses in the Frog Motoneuron

Frog motoneurons were intracellularly labelled with cobaltic lysine in the brachial and the lumbar segments of the spinal cord, and the material was processed for light microscopy in serial sections. With the aid of the neuron reconstruction system NEUTRACE, the dendritic tree of neurons was reconstructed and the length and surface area of dendrites measured. The surface of somata was determined with the prolate - oblate average ellipsoid calculation. Corrections were made for shrinkage and for optical distortion. The mean surface area of somata was 6710 microm2; lumbar motoneurons were slightly larger than brachial motoneurons. The mean length of the combined dendritic tree of brachial neurons was 29 408 microm and that of lumbar neurons 46 806 microm. The mean surface area was 127 335 microm2 in brachial neurons, and 168 063 microm2 in lumbar neurons. The soma - dendrite surface area ratio was 3 - 5% in most cases. Dendrites with a diameter of 600 microm from the soma. This suggests that about two-thirds of the synapses impinged upon distant dendrites >600 microm from the soma. The efficacy of synapses at these large distances is investigated on model neurons in the accompanying paper (Wolf et al., Eur. J. Neurosci., 4 1013 - 1021, 1992).     

Dendritic bundling in layer I of granular retrosplenial cortex: intracellular labeling and selectivity of innervation

The extrinsic projections to and from the retrosplenial cortex have been studied in detail, but the intrinsic circuitry within this region has been characterized less completely. To further define the internal connections, small injections of the retrograde, fluorescent tracer Fluorogold were made into the retrosplenial cortex of the rat. These injections label neurons in layers II-V of the contralateral homotopic cortex. In layers III-V, the labeled neurons are present over an area much larger than the injection site, but in layer II neurons are labeled in a very precise homotopic pattern. Following these injections, only the neurons in layer II display heavily labeled apical dendrites, and these labeled dendrites form tight bundles in layer Ic and Ib of the cortex and spread out in layer Ia. An examination of Golgi-stained material demonstrates that most of the neurons in layer II are small pyramidal cells with 2-3 small basal dendrites and a single, large apical dendrite that arborizes extensively in layer Ia. To verify the structure of the layer II neurons, they were intracellularly filled with Lucifer yellow. Examination of these labeled cells confirms the observations from the Golgi-stained material and demonstrates that many apical dendrites of the layer II cells angle acutely, apparently to join a bundle and/or avoid an interbundle space. Tract tracing experiments demonstrate that the anteroventral nucleus of the thalamus appears to project selectively to the region containing the dendritic bundles, whereas intracortical projections appear to terminate in layers Ib and Ic in the 30-200 microns spaces between the bundles. Furthermore, the areas containing the bundles display dense AChE staining, but the interbundle spaces are almost free of AChE staining. These findings demonstrate a form of dendritic bundling that is input and output specific and may play an important role in the regulation of thalamic inputs to the cingulate cortex.     

The size and dendritic structure of HRP-labeled gamma motoneurons in the cat spinal cord.

We report quantitative data obtained from 60 fully reconstructed dendritic trees belonging to eight gamma-motoneurons (gamma-MNs) and six additional gamma-MNs that were not completely reconstructed. The cells were labeled intracellularly with horseradish peroxidase (HRP). These data are compared to measurements from 79 reconstructed dendrites belonging to seven documented alpha-motoneurons (alpha-MNs), supplemented by a larger sample of alpha-MNs labeled intracellularly or by retrograde transport with HRP. As expected from earlier studies, the soma dimensions and total membrane area of gamma-MNs were smaller than those of alpha-MNs. Although gamma-MN dendrites were, on average, slightly but significantly longer than those of alpha-MNs, the former had, on average, smaller diameter stem dendrites, less membrane area, and less profuse branching, and they tended to branch closer to the soma and to terminate farther from the soma. These differences were evident even when subsets of dendrites with similar stem diameters were compared. Some of the anatomical distinctions suggest that gamma-MNs are qualitatively as well as quantitatively different from alpha-MNs, even though the distributions of many of the morphological variables examined showed no abrupt discontinuities between the two motoneuron groups.     

Morphology of dendrite bundles in the cervical spinal cord of the rat: a light microscopic study

Histological staining techniques and Golgi-Cox impregnation revealed three discrete dendrite bundles in the ventral horn of the rat cervical spinal cord. A midline dendrite bundle (MDB) traversed the ventromedial gray matter (C3-6), a central dendrite bundle (CDB) coursed the medial aspect of the ventral horn (C3-5), and a lateral dendrite bundle (LDB) traveled in the ventrolateral gray matter (C2-4). At the light microscopic level, the three dendrite bundles were composed of longitudinally oriented intertwined dendrites that coursed in close apposition among motoneuron perikarya, neuroglia, and capillaries. A gradient of packing density of dendrites in the bundles existed, the MDB displaying the greatest packing density and the LDB forming the most loosely interwoven dendritic plexus. Dendrites contributing to the bundles originated from several different motoneuron pools. Smaller transverse dendrite bundles radiated from the longitudinal dendrite bundles at right angles and appeared to interconnect the MDB, CDB, and LDB. Transverse dendrite bundles also exited the MDB and LDB to course into the anterior and lateral funiculi, respectively. The presence of dendrite bundles among fields of motoneurons suggests that dendrite bundles may provide an anatomical substrate for the synchronization of neuronal activity for coordination of muscle groups involved in particular movements. Dendrite bundles also would provide a means whereby functionally similar motoneurons can receive and integrate similar synaptic inputs, and thus allow these inputs to modulate and coordinate groups of neurons that act as a functional unit. The presence of transverse dendrite bundles interconnecting the longitudinal bundles may permit the fine tuning of motoneuron activity for better coordination of movements involving synergistic and antagonistic muscle groups.     

The organisation of dendritic bundles in the prelimbic cortex (area 32) of the rat

Using an antibody against microtubule associated protein 2 (MAP-2; a specific marker for neuronal dendrites), this paper reports the structural organisation of pyramidal cell apical dendrites in the rat prelimbic (PL) cortex. In the coronal plane, MAP-2-immunoreactive apical dendrites of pyramidal neurons in layers 5, 3 and 2 were found bundled together as they ascended radially through the cortex. These bundles of dendrites dispersed in upper layer 2 to form apical dendritic tufts in layer 1. In tangential cross-section, the immunolabelled bundles were organised into a latticework of discrete clusters of differentially sized profiles. At the boundary between layers 3 and 5, clusters were composed of 26 ± 8 dendritic profiles (group mean value ± S.D., five animals), whereas clusters in layer 2 contained 55 ± 15 profiles. The number of clusters per unit surface area was not significantly different throughout layers 5, 3 and 2 (760 ± 75 per mm²) with the average centre-to-centre intercluster distance in these layers being 44.2 ± 4.9 μm. The data indicate that apical dendritic bundles are a feature of the radial organisation of PL cortex. These structural subunits may subserve specific integrative functions in the PL area of the rat medial prefrontal cortex.     

Dendritic alteration of rat spinal motoneurons after dorsal horn mince: Computer reconstruction of dendritic fields

Mammalian spinal motoneuron dendrites respond with cyclic degeneration and regeneration after ventral root crush. In the following experiments, the cross-sectional dendritic profile of rat lamina IX, medial, motoneurons under the T2 vertebra were analyzed after mincing the dorsal horn (normals, 14, 30, 60, 90 days postoperative (DPO); N = 6 animals/DPO). The spinal cords were impregnated by the tungstate modification of the Golgi technique. Individual lengths along dendritic segments between branching points were measured from coded slides, the data were computerized, and the dendrites were reconstructed by computer. Interanimal statistical comparisons were made by ANOVA a priori and Newman-Keuls test a posteriori. At 14 DPO, there was a statistically significant (P < 0.05) increase in the number of dendritic segments, dendritic and serpentine length, and number of segments emanating from the soma compared with normal intact rats and with all other postoperative days. At 30 and 60 DPO, these parameters returned to normal values; however, there were many long, unbranched dendrites. At 90 DPO, there was a statistically significant (P < 0.05) decrease in motoneuron dendritic serpentine length and segments. These data show that partially deafferented rat spinal motoneurons undergo a biphasic response; an initial growth phase followed by a degenerative phase.     

Dendritic aberrations in the hippocampal granular layer and the amygdalohippocampal area following kindled-seizures

Amygdaloid kindling is a model of human temporal lobe epilepsy, in which excitability in limbic structures is permanently enhanced by repeated stimulations. We report here dendritic aberrations occurring in mice following kindled-seizures. Adult mice received a biphasic square wave pulse [495+/-25.5 (S.E.M.) microA 60 Hz, 200 micros duration, for 2 s] unilaterally in the basolateral amygdaloid complex once a day and mice with electrophysiologically and behaviorally verified seizures were used in the experiments. The hippocampus and amygdaloid complex contralateral to the lesions were observed by immunofluorescence histochemistry with a somatodendritic marker, microtubule-associated protein 2 (MAP2), showing that kindled-seizures caused hypertrophy of proximal dendrites in the granule cells of the dentate gyrus and in neurons of the amygdalohippocampal area. To further characterize the morphological changes of the dendrites, electron micrographic analysis was performed on the contralateral side. (1) In the granular layer of the dentate gyrus and the amygdalohippocampal area, kindled-seizures generated an increase in the number of dendrites containing polymerized microtubules and width of dendritic profiles showing the increase was in the range 0.2-3.0 and 0.2-1.4 microm, respectively. (2) In the granular layer, bundles between dendrites separated by the puncta adhaerentia increased. (3) In the granular layer, the seizure-induced dendritic aberration was more severe in the rostral than the caudal region. These results suggested that growth of dendrites with enriched-stable microtubules is part of the structural plasticity in response to seizure activity in specific areas of the adult brain.     

Distribution of 5-hydroxytryptamine-immunoreactive boutons on α-motoneurons in the lumbar spinal cord of adult cats

Recent studies have shown that at least some of the functional effects of serotonin (5-HT) on motoneuron excitability are direct and are mediated via postsynaptic 5-HT receptors on motoneurons. To determine the spatial distribution of direct inputs from the serotonin system on the proximal and distal dendrites of individual motoneurons, we examined identified motoneurons in vivo with a combination of immunohistochemical localization of 5-HT-immunoreactive boutons and intracellular staining with horseradish peroxidase. Seventeen intracellularly stained motoneurons from 12 adult cats were analyzed with light microscopy. Quantitative analysis of 5-HT boutons apposed to dendrites of five representative motoneurons that were entirely reconstructed in three dimensions (each from the lumbosacral spinal cord of a different animal) revealed a total of 7,848 contacts (1,570 ± 487 contacts/postsynaptic neuron; mean ± SD) over the dendrites of these cells. Analysis of contacts on the soma of two of these cells, and on the somas of an additional 12 intracellularly stained motoneurons, revealed a wide range of somatic contacts (11–211 contacts/cell) on motoneuron cell bodies, with an average of 52 contacts/cell. These results indicate that the vast majority of 5-HT-immunoreactive boutons are apposed to dendritic branches rather than to the somatic surface of motoneurons. The spatial distribution of contacts essentially matched the distribution of surface membrane area of the postsynaptic neuron, resulting in a relatively uniform density of contacts (<1/100 μm²) on proximal and distal dendrites. Consequently, the frequency of contacts was higher on the proximal dendritic compartments where available membrane area is greater. There was no preferential distribution of contacts to particular dendrites. Light/electron microscopic correlations were performed on 21 boutons that contacted dendrites (n = 7) of three motoneurons from different animals. At the electron microscope level, most appositions (18/21; 85.7%) selected by our light microscopic criteria were confirmed as direct contacts when the 5-HT boutons were examined through serial sections. Synaptic junctions, generally small and symmetric, were positively identified in only a subset of these cases (n = 6; 28.6%), in part due to the obscuring effects of the peroxidase histochemical precipitate present in both pre- and postsynaptic profiles. A few 5-HT boutons (3/21; 14.3%) selected as contacts by our light microscopic criteria were in fact separated from the adjacent labeled dendrites; in two of these three cases, the separation was due to intrusion of very thin glial lamellae (<0.3 μm in cross section). These results indicate that the bulbospinal serotonergic system(s) provide a significant, direct synaptic input to spinal motoneurons that innervate hindlimb muscles. The nature of the modulatory actions exerted by such widespread synaptic inputs will affect all regions ofthe somatodendritic membrane and will ultimately depend on the nature of the 5-HT receptors present over different parts of the postsynaptic neuron's dendritic tree. J. Comp. Neurol. 393:69–83, 1998. © 1998 Wiley-Liss, Inc.     

Motoneuron morphology in the dorsolateral nucleus of the rat spinal cord: Normal development and androgenic regulation

The rat lumbar spinal cord contains two sexually dimorphic motor nuclei, the spinal nucleus of the bulbocavernosus (SNB), and the dorsolateral nucleus (DLN). These motor nuclei innervate anatomically distinct perineal muscles that are involved in functionally distinct copulatory reflexes. The motoneurons in the SNB and DLN have different dendritic morphologies. The dendrites of motoneurons in the medially positioned SNB have a radial, overlapping arrangement, whereas the dendrites of the laterally positioned DLN have a bipolar and strictly unilateral organization. During development, SNB motoneuron dendrites grow exuberantly and then retract to their mature lengths. In this experiment we determined whether the adult difference in SNB and DLN motoneuron morphology was reflected in different patterns of dendritic growth during normal development. Furthermore, the development of both these nuclei is under androgenic control. In the absence of androgens, SNB dendrites fail to grow; testosterone replacement supports normal dendritic growth. Thus, we also examined the development of DLN dendrites for similar evidence of androgenic regulation. By using cholera toxin-horseradish peroxidase (BHRP) to label motoneurons retrogradely, we measured the morphology of DLN motoneurons in normal males, and in castrates treated with testosterone or oil/blank implants at postnatal day (P) 7, P28, P49, and P70. Our results demonstrate that in contrast to the biphasic pattern of dendritic development in the SNB, dendritic growth in the DLN was monotonic; the dendritic length of motoneurons increased more than 500% between P7 and P70. However, as in the SNB, development of DLN motoneuron morphology is androgen-dependent. In castrates treated with oil/blank implants, DLN somal and dendritic growth were greatly attenuated compared to those of normal or testosterone-treated males. Thus, while androgens are clearly necessary for the growth of motoneurons in both the SNB and DLN, their different developmental patterns suggest that other factors must be involved in regulating this growth. © 1993 Wiley-Liss, Inc.