Tracing the dentate gyrus mossy fiber system with horseradish peroxidase histochemistry

The dentate mossy fibers: structural organization, development and plasticity

Hippocampal mossy fibers are the axons of the dentate granule cells and project to hippocampal CA3 pyramidal cells and mossy cells of the dentate hilus (CA4) as well as a number of interneurons in the two areas. Besides their role in hippocampal function, studies of which are still evolving and taking interesting turns, the mossy fibers display a number of unique features with regard to axonal projections, terminal structures and synaptic contacts, development and variations among species and strains, as well as to normal occurring and lesion-induced plasticity and neural transplantation. These features are the topic of this review, which will use the mossy fiber system of the rat as basis and reference in its aim to provide an up-to-date, yet historically based guide to students in the field.

Novel expression of AMPA-receptor subunit GluR1 on mossy cells and CA3 pyramidal neurons in the human epileptogenic hippocampus

Previous immunocytochemical investigations performed in our laboratory on the human hippocampus surgically resected for the treatment of mesial temporal lobe epilepsy (MTLE) have demonstrated an increased expression of the AMPA-receptor subunit GluR1 on neurons in the hilus and area CA3. Light microscopically, many of these neurons exhibited peculiar filamentous extensions and grape-like excrescences that protruded from their somata and proximal dendrites, suggesting that these neurons may be mossy cells and CA3 pyramidal neurons, respectively. The present electron microscopic study was carried out to further characterize these cells. The filamentous extensions were identified as dendrites from which spines often protruded, and the grape-like excrescences represented clusters of closely associated dendrites and spines. A variety of synapses were formed by the GluR1-positive profiles. These arrangements ranged from simple contacts between a single unlabelled axon terminal and a single labelled postsynaptic element, to complex contacts involving multiple unlabelled axon terminals and labelled postsynaptic elements. Many of the axon terminals involved in these arrangements were mossy fibre boutons. Thus, a large proportion of the GluR1-positive neurons were identified as hilar mossy cells and CA3 pyramidal neurons, cells hitherto thought to be absent or greatly reduced in the MTLE hippocampus. Taken together, these data suggest the presence of a highly efficient excitatory circuit involving AMPA receptors, mossy cells and CA3 pyramidal neurons in the sclerotic hippocampus. Such a circuit could be critically involved in the genesis and maintenance of temporal lobe epilepsy.

Ipsilateral associational pathway in the dentate gyrus: an excitatory feedback system that supports N-methyl-D-aspartate-dependent long-term potentiation

Axons from granule cells in the dentate gyrus of the rat hippocampus project to cells in the hilar region, including mossy cells, which project along the longitudinal axis of the hippocampus and synapse in the inner (proximal) one-third of the molecular layer of the dentate gyrus. To study this feedback system, multiple recording electrodes were located along the longitudinal (septo-temporal) axis in the dorsal leaf of the dentate gyrus in urethane-anesthetized rats. Single pulse electrical stimuli delivered to the hilar region evoked negative-going, monosynaptic field potentials that were largest in the inner one-third of the molecular layer (commissural zone). These evoked field potentials (EFPs) were recorded simultaneously at three to five locations. The latency to onset and peak amplitude of the EFP varied linearly with distance from point of stimulation, and EFPs were elicited in both directions along the longitudinal axis. The transmission speed was estimated to be 1.4 m/s. Tetanic stimulation of the hilar region potentiated the EFP slopes (mean = 26%). Potentiation lasted at least 2 hours and was specific to responses from the tetanized stimulating electrode; the responses to other stimulating electrodes in the hilus and the angular bundle of the perforant path changed less than 4%. Combined stimulation of the hilus and the medial perforant path increased the magnitude of recorded field potentials and population spikes, demonstrating that both pathways are excitatory. NMDA antagonist NPC-17742 blocked potentiation of EFP slopes in both the medial perforant path and hilus pathways. The results suggest that the ipsilateral associational system of the dentate gyrus is excitatory and capable of supporting long-lasting NMDA-dependent, synapse-specific plasticity.

Neurons in the medial cortex give rise to Timm-positive boutons in the cerebral cortex of lizards

The origin of Timm-positive presynaptic boutons in the cerebral cortex of the lizard, Podarcis hispanica, was investigated by injections of horseradish peroxidase (HRP)-saponine in Timm-positive areas, i.e. the dorsal and dorsomedial cortices. A broad retrograde labelling of cell somata in the medial cortex was found. Injections of HRP-saponine in the medial cortex resulted in broad anterograde labelling of boutons located in the Timm-positive zones. A double-labelling of the HRP labelled boutons was obtained by using the Neo-Timm or the sulphide-osmium methods. The present results suggest that neurons of the medial cortex send axons that terminate in Timm-positive boutons in the cerebral cortex of lizards.

Brain stem nuclei giving fibers to lobules VI and VII of the cerebellar vermis

The HRP method has been used to identify all the brain stem nuclei, which may project to lobule VI and/or VII of the posterior cerebellar vermis. Three tentative degrees of labeling of the different structures have been assigned: 'massive', 'clear' and 'discrete'. (1) Massive projections have been found to reach lobule VI and VII from the inferior olive and lobule VII only from the nucleus reticularis tegmenti pontis. (2) Clear projections have been found to reach lobule VI only from the pontine nuclei, the nucleus reticularis tegmenti pontis, the nucleus reticularis lateralis and the reticularis paramedianus; lobule VII only from the raphe nuclei, and both VI and VII from the perihypoglossal and vestibular nuclei. (3) Discrete projections have been found to reach lobule VI and VII from the deep cerebellar nuclei; lobule VI only from the nucleus tracti solitarii and nucleus cuneatus externus; lobule VII only from the nucleus lemnisci lateralis pars ventralis, the nuclei parabrachiales and the nucleus subcoeruleus.

A light and electron microscopic analysis of the mossy fibers of the rat dentate gyrus

The axon collaterals of dentate granule cells have been analyzed with the aid of a computerized microscope, following intracellular injections of horseradish peroxidase in hippocampal slice preparations. The axon of each granule cell gives rise to approximately seven primary collaterals; these collaterals usually divide into secondary and tertiary branches, which form an extensive plexus within the hilar region of the dentate gyrus. Individual axon collaterals vary greatly in length, but most have been found to be between 100 and 300 microns long. On average, the summed lengths of the collaterals (exclusive of the parent mossy fiber) are approximately 2,300 microns. Except for an occasional collateral that is given off by a mossy fiber in the proximal part of field CA3 of the hippocampus, the collaterals of the granule cell axons are confined to the hilar region; they are rarely seen in the granule cell layer itself and have never been observed in the molecular layer. In the longitudinal dimension of the dentate gyrus, most of the collaterals are contained within a zone about 400 microns wide. The distribution of the collaterals within the hilar region is correlated with the location of the granule cell body. Those that arise from cells near the tip of the suprapyramidal blade tend to be confined to the region above field CA3; those from cells nearer the crest and from the infrapyramidal blade ramify widely throughout the hilus. Two types of varicosities are present on the collaterals. Numerous small (approximately 2 microns), round varicosities are distributed unevenly along the collaterals; in electron micrographs these varicosities can be seen to make asymmetric synaptic contacts with dendritic shafts. On average, each granule cell collateral plexus has about 160 of these varicosities. The second type of varicosity is irregular in shape and ranges from 2 to 4 microns in diameter; there is usually only one such varicosity per collateral. In all respects except size, these varicosities resemble the expansions found on the parent mossy fibers. Mossy fiber trajectories in the proximal part of field CA3 were studied after extracellular injections of HRP into localized regions of the granule cell layer. Granule cells at different locations around the blade send their mossy fibers to different depths within the pyramidal cell layer in the proximal part of field CA3. However, further distally, mossy fibers from all parts of the granule cell layer contribute to the suprapyramidal bundle that occupies the stratum lucidum.

An HRP study of the connections of the subfornical organ of the rat

The connections of the subfornical organ (SFO) were investigated by using the HRP technique. Injections into the SFO labeled neurons in the medial septum, but not in lateral septum nor in the diagonal band nucleus. Labeled cells were observed in the median preoptic nucleus, below the ependyma of the third ventricle, in the dorsal preoptic region near the anterior commissure, and diffusely throughout the medial preoptic and anterior hypothalamic areas. Fibers were followed from the ventral stalk of the SFO. Precommissural fibers enter the median preoptic nucleus where many of them appear to terminate. Others continue on to the medial septum, the OVLT, the supraoptic nucleus, and suprachiasmatic nucleus. HRP injections into the median preoptic nucleus labeled many neurons in the SFO. Postcommissural fibers reach the hypothalamus by descending along the walls of the ventricle in the subependymal space, by traveling in the columns of the fornix and the medial corticohypothalamic tract, or by passing through the paraventricular nucleus of the thalamus. Some postcommissural fibers turn rostrally and enter the median preoptic nucleus while others join precommissural fibers bound for the supraoptic nucleus. More caudally directed fibers appear to innervate the paraventricular nucleus of the hypothalamus and the medial preoptic and anterior hypothalamic areas. HRP injections into the paraventricular nucleus of the hypothalamus labeled neurons in the SFO. These findings corroborate and extend previous work in describing neural connections between two brain regions that are important for fluid balance.

Fate of the hippocampal mossy fiber projection after destruction of its postsynaptic targets with intraventricular kainic acid

Intraventricular injections of kainic acid were used to create a model of selective cell death in order to study the fate of afferent projections that are deprived of their postsynaptic targets. This treatment rapidly destroyed hippocampal CA3 pyramidal cells, but not those neurons that give rise to their mossy fiber and entorhinal afferents. Light microscopic studies with the Timm's sulfide silver stain indicated that half or more of the mossy fiber boutons in area CA3b were lost within the first 1-3 days after kainic acid administration. This finding was confirmed by electron microscopy. Electron-dense, usually vacuolated mossy fiber boutons accounted for about 10-20% of the total population present at a 4-hour survival time, but were not encountered in control rats nor at survival times longer than 1 day. Other mossy fiber boutons remained electron lucent, but enlarged, became more rounded in shape, and suffered an apparent loss of synaptic vesicles. It is suggested that degeneration of some mossy fiber boutons and resorption of others into the axon may have accounted for the precipitous decline in their number. The dendritic excrescences contacted by these boutons were nearly all undergoing electron-dense degeneration 4 hours after kainic acid administration. In rats that survived 6-8 weeks mossy fiber boutons remained somewhat scarce, individual boutons appeared relatively small, and only one-third the normal percentage were observed to be engaged in more than one synaptic contact within a single cross section. A qualitative electron microscopic study of the entorhinal projection to area CA3 suggested a response to kainic acid treatment similar to that of the mossy fiber projection, except that no entorhinal boutons were seen to become electron dense. These findings suggest that presynaptic fibers in the mature hippocampus adjust the size of their terminal arborizations and number of synaptic contacts to accommodate a reduction in the target cell population. The rapid loss of mossy fiber boutons may be attributable to an unusual fragility of these structures when they are deprived of the mechanical support normally provided by the pyramidal cell. Finally, the ability of kainic acid administration to alter the number and distribution of presynaptic elements must be taken into account whenever this toxin is used to make selective lesions of postsynaptic cells.

Distal infrapyramidal granule cell axons possess typical mossy fiber morphology

Using both the Timm's sulfide silver and anterograde horseradish peroxidase (HRP) histochemical techniques, we have demonstrated that granule cell in the dentate gyrus of the rat have topographical projections to both suprapyramidal and distal infrapyramidal terminal fields in the rostral one-third of the hippocampus. An anterograde HRP technique modified for visualization of axonal morphology indicated that intrapyramidal mossy fiber axons possess classic periodic swellings characteristic of suprapyramidal mossy fibers. Most distal infrapyramidal mossy fiber axons appear to be in position to make synaptic contact with deeper-lying neurons in the pyramidal cell layer.

A new method for application of horseradish peroxidase into a restricted area of the brain

A new method for horseradish peroxidase (HRP) application into the very restricted area of the brain was developed. Crystalline HRP was packed into the tips of fine micropipettes having tip size of about 50 micrometer in outer diameter. The HRP micropipettes were stereotaxically implanted into the globus pallidus of rats. After survival period of 6 to 48 hours, brain was excised and HRP distribution was investigated following conventional procedures. The diffusion area of HRP at the application site was restricted within the limits of 200--250 micrometer in diameter completely inside the globus pallidus. It was known that the slowly solubilized crystalline HRP was taken up by the nerve terminals in the globus pallidus, and transported retrogradely to the cell bodies in the pars compacta of substantia nigra, indicating the existence of nigro-pallidal projections.

Development of the mossy fibers of the dentate gyrus: I. A light and electron microscopic study of the mossy fibers and their expansions

The postnatal development of the axons of the dentate granule cells--the so-called mossy fibers--was studied at the light microscopic level in Timm and Golgi preparations and also by transmission electron microscopy. In the Timm-stained material, there was a distinctive coloration in the hilus and incipient stratum lucidum, indicating the presence of mossy fibers, on the first postnatal day. Over the next two weeks, the stained areas became more extensive, the size and density of the stained particles increased, and the particles became more intensely stained. These signs of progressive development of the mossy fibers appeared to reflect, temporally and topographically, the developmental gradients followed by their parent granule cells. The Golgi material confirmed the presence of mossy fibers in the hilus on the first postnatal day. Fasciculi of mossy fibers were observed in the stratum lucidum of the 3-day-old hippocampus, and although these immature axons were devoid of large synaptic expansions, they did have prominent growth cones at their termini. Small expansions along the lengths of the axons first appeared on day 7 and these grew to approximately an adult size and complexity by about day 14. The postsynaptic component of the mossy fiber synapse, the "thorny excrescence," did not begin to emerge from the proximal portion of the pyramidal cell dendrites until sometime after day 9. At the electron microscopic level we observed, on the first postnatal day, small, immature mossy fiber expansions which made both symmetric and asymmetric contacts directly with dendritic shafts. These profiles, which were only one tenth the size of mature expansions, grew rapidly between postnatal days 1 and 9 and increased their mean area by a factor of five. On or about day 9, as the "thorny excrescences" emerged, the asymmetric synapses came to be associated with these spinous processes. Taken together, the Golgi and electron microscopic analyses support the suggestion that mossy fibers establish synaptic contact with pyramidal cell dendrites early in the postnatal period, several days before there is any indication of spine development. Furthermore, the "thorny excrescences" develop after the more typical, pedicellate spines have appeared on the distal pyramidal cell dendrites. Finally, while it is clear that the mossy fibers in our 21-day-old material are, for the most part, fully matured, a more subtle and protracted development of the system, long into adulthood, is indicated by the increased area and density of stained particles in the Timm preparations from adult animals.

Projections of lamprey spinal neurons determined by the retrograde axonal transport of horseradish peroxidase

The spinal cords of larval sea lampreys (Petromyzon marinus) and adult river lampreys (Ichthyomyzon unicuspis) were injected with horseradish peroxidase through a transection 1 cm caudal to the last gill. Some animals also had a spinal hemisection 1 cm caudal to the injection. After recovery periods of 1 to 52 days, the spinal cords were treated with diaminobenzidene and hydrogen peroxide, and the projections of various cell types determined in wholemount slides. From these observations the following conclusions were drawn. Most dorsal cells (primary sensory cells) are bipolar with a long rostral projection and a short caudal projection of no more than 5-10 mm. Both processes travel in the ipsilateral dorsal column. Their peripheral processes enter the dorsal roots as branches of their central axons. Some dorsal cells send processes out three or more dorsal roots both rostral and caudal to the cell body. Myotomal motoneurons have characteristic locations in the medial gray column and send prominent transversely oriented dendrites into the lateral columns. A few motoneurons are unusually large. In addition to giant interneurons the majority of smaller rostrally projecting interneurons also have decussating axons. A recently described cell type, the oblique bipolar cell, appears to have an exclusively crossed rostral projection. Although most edge cells project rostrally, as many as 20% may have a caudal projection or both rostral and caudal projections. Edge cells project equally to the ipsilateral and contralateral spinal hemicord, but their processes do not extend more than about 18 mm in sea lamprey larvae and 37 mm in adult river lampreys. Lateral cells project exclusively to the ipsilateral caudal hemicord. A few cells which resemble lateral cells in location and in possessing large lateral dendrites, project rostrally. However, these have atypical morphologic features which probably distinguish them from true lateral cells. Thus far, regardless of cell type, all decussating axons seem to pass ventral to the central canal, while decussating medial dendrites pass dorsally.

The decussation of the retinothalamic pathway in the cat, with a note on the major meridians of the cat's eye

We have studied the naso-temporal division of the retinothalamic pathway of the cat by making large unilateral injections of horseradish peroxidase into the lateral geniculate nucleus. In confirmation of previous work, our retinal whole-mounts show a distinct vertical decussation line separating the contralaterally projecting nasal retina from the ipsilaterally projecting temporal retina. The ipsilateral decussation line is quite sharp, while the contralateral decussation is somewhat more diffuse, with numbers of large cells extending a few degrees into the temporal retina. However, in contrast to the results of optic tract section, our material (demonstrating the thalamic component only) does not reveal any significant population of contralaterally projecting small cells across most of the temporal retina. The previous observation of approximately 200 micrometer of naso-temporal overlap in the area centralis is confirmed here, and evidence is presented that this overlap may increase at eccentricities above the horizontal meridian. Taken together with previously published data, this demonstration of the vertical decussation line has allowed us to estimate the relative inclinations of the major meridians of the cat's eye.

A cerebello-pulvino-cortical and a retino-pulvino-cortical pathways in the cat as revealed by the use of the anterograde and retrograde transport of horseradish peroxidase

A cerebello-pulvino-cortical and a retino-pulvino-cortical pathways were revealed in the cat by means of the horseradish peroxidase (HRP) method. The sites of termination of the cerebellofugal and retinofugal fibers in the pulvinar nucleus (Pul) were visualized by the use of the anterograde transport of HRP. The cerebello-pulvinar fibers were found to arise mainly from the parvicellular region of the lateral cerebellar nucleus and to terminate contralaterally in a narrow area at the extreme dorsolateral edge of the Pul at the level of the stereotaxic frontal plane A-7.0. The area of terminal ramification of retino-pulvinar fibers was seen as a thin sheet lying at the extreme lateral edge of the Pul through most of the rostrocaudal extent of the Pul, bilaterally with contralateral predominance. The cerebellorecipient area in the Pul did not seem to overlap with the retinorecipient Pul area; the former appeared to be contiguous ventrolaterally to the latter. The cerebellorecipient and retinorecipient Pul areas were also observed to be connected reciprocally with the cerebral cortical areas; the former was connected with the most posterior part of the area 20, and the latter with the area 19.

Reexamination of the dorsal root projection to the spinal dorsal horn including observations on the differential termination of coarse and fine fibers

Primary afferent fibers in the lumbar, sacral, and caudal spinal segments of several mammals (rat, cat, monkey) were stained by applying horseradish peroxidase to the proximal part of cut dorsal rootlets and reacting the tissue histochemically after several hours of survival. The stained fibers' pattern of termination in the dorsal horn was similar in all three species, with many bouton-like enlargements in the ipsilateral marginal zone, substantia gelatinosa, and nucleus proprius, as well as a few projections at each level to the dorsal commissure and contralaterally to the ventral border of the nucleus proprius. Partial lesions of dorsal rootlets in monkey revealed that the thin fibers comprising the lateral division end principally in the marginal zone and substantial gelatinosa, while the thick fibers of the medial division terminate in the nucleus proprius and deeper regions, contributing little to the substantia gelatinosa and marginal zone. On the basis of the termination patterns observed for whole and partly sectioned rootlets, the superficial dorsal horn can be divided into at least four regions. (1) The marginal zone (lamina I of cat) appears to receive terminations from intermediate (smaller myelinated) fibers; (2) the outer substantia gelatinosa (outer lamina II) receives many terminations from the very finest afferent fibers; (3) the inner substantia gelatinosa (inner lamina II) receives endings from some of the finest fibers and also from intermediate (smaller myelinated) fibers; and (4) the superficial part of the nucleus proprius (lamina III) receives endings from intermediate and large diameter dorsal root fibers.

The cells of origin of the commissural afferents to the area dentata in the mouse

The hippocampal commissural projection to the area dentata of the mouse was studied using the retrograde horseradish peroxidase (HRP) technique. Small volumes of HRP injected into the molecular layer of the fascia dentata or various subareas of regio inferior of the hippocampus (fields CA3a-c) resulted inlabeled perikarya in the contralateral hippocampus and area dentata. The commissural projection to the fascia dentata was observed to originate exclusively from cells within the hilus fasciae dentatae (CA4) of the contralateral area dentata. There was evidence of a considerable spread of commissural innervation along the septotemporal axis preferentially in the septal direction, confirming earlier observations. In contrast to the septotemporal spread, a sharp homotopic spatial organization was found in the mediolateral direction. For example, injections into the lateral portion of field CA3 (CA3a) resulted in HRP-positive cell bodies only in the contralateral field CA3a. When injections were made which apparently labeled all of the commissural fibers, the HRP reaction product was found in neurons both in the entire regio inferior and as far as the innermost point of the hilus fasciae dentatae; the majority of labeled cells were located in hippocampal subfield CA3c. No labeled cells were observed beyond the tip of the mossy fibers in regio superior.

Neuronal plasticity in the limbic system during classical conditioning of the rabbit nictitating membrane response. I. The hippocampus

Hippocampal unit responses were recorded throughout classical conditioning of the rabbit nictitating membrane response to a tone conditioned stimulus (CS) using a corneal air-puff unconditioned stimulus (UCS). Multiple unit analysis revealed that a rapidly developing increase in cell discharges (relative to spontaneous activity) occurs within the first block of paired trials and continues to increment with subsequent training, initially in the UCS period and then in the CS period. The pattern of hippocampal activity within paired trials closely parallels the amplitude-time course of the behavioral response and precedes it temporally. Identical recordsings from animals given unpaired CS-alone and UCS-alone presentations showed no such changes. These control results and additional lines of evidence point to the critical necessity of the learning paradigm for the development of the hippocampal response seen in conditioning animals. A single unit analysis indicates that not all hippocampal neurons exhibit the described conditioned discharge pattern. Hippocampal long-term potentiation is considered as a possible mechanism for mediating this early and rapid neuronal plasticity dependent on specific 'contingent' patterns of stimulation.

Afferents to brain stem nuclei (brain stem raphe, nucleus reticularis pontis caudalis and nucleus gigantocellularis) in the rat as demonstrated by microiontophoretically applied horseradish peroxidase

Using a retrograde tracer technique with microiontophoretically applied horseradish peroxidase (HRP), afferent projections to the brain stem raphe nuclei (BR, raphe magnus, pallidus and obscurus) and to two adjacent reticular nuclei, nucleus reticularis pontis caudalis (nRPC) and nucleus gigantocellularis (nGC) were identified. The most striking difference between the afferent projections to the BR and the adjacent nuclei as determined by this method is that afferents to the BR originate primarily from structures rostral to the pons, especially the mesencephalic central gray and the dorsal and ventral tegmentum. In contrast, the two reticular nuclei studied (nGC and nRPC) received afferent projections within or caudal to the pons-medulla. For example, the nGC receives prominent afferent projections from the gray matter of the spinal cord. In addition, evidence for interconnections between all of the adjacent nuclei (BR, nGC and nRPC) was found. Such afferent projections are compatible with the notion that the brain stem raphe nuclei may serve as connections within the brain stem for a descending system, while the nGC may be a relay in a feedback loop between the spinal cord and the reticular formation.

An explanation of axonal regeneration in peripheral nerves and its failure in the central nervous system

Nerve fibres severed within peripheral nerves are able to regenerate and reinnervate the structures they formerly supplied. Most axons severed within the mammalian central nervous system (CNS) do not regenerate in this way. Regenerative axonal growth begins to occur in the CNS but ceases about two weeks after injury. Five earlier theories purporting to explain this difference are reviewed and found not to account satisfactorily for many experimental observations. A new hypothesis is advanced in which it is held that in order for regeneration to take place, the growing tips of the axons must be surrounded by extracellular fluid containing proteins (of specified identity) derived from the blood plasma. Such proteins are thought to be imbibed by the tips of the fibres and transported retrogradely to the neuronal cell-bodies. With this hypothesis it is possible to explain the success of axonal regeneration in peripheral nerves and its failure in the CNS. It is also possible to account for the exceptional circumstances in which axons do regenerate in the CNS. Various experiments are suggested for testing the validity of the new hypothesis.

Habenular and other midbrain raphe afferents demonstrated by a modified retrograde tracing technique

Afferents to th midbrain dorsal and median raphe nuclei in the rat were studied by means of the horseradish peroxidase (HRP) retrograde transport method. The HRP was given by means of a modified iontophoretic delivery technique. This technique permitted an efficient and localized deposition of a high concentration of HRP into the raphe nuclei. Afferents to the raphe as determined by this method could be categorized into 2 classes; those exclusively to the raphe and those also positive for adjacent reticular formation. The most striking afferent area to the raphe, both in terms of selectivity and density, was the lateral habenula. This result is in accord with previous studies using degeneration methods which indicate an habenular projection to the raphe area. There were afferents exclusively positive for the dorsal raphe nucleus emanating from the nucleus of the solitary tract. Most other raphe afferent areas were also positive for the reticular formation (e.g;, prefrontal cortex, medial forebrain bundle, preoptic nuclei, and reticular formation). The existence of a major afferent system from the lateral habenula to the midbrain raphe is consistent with the concept of a "dorsal pathway" which might be responsible for relaying information from forebrain limbic structures to the "midbrain limbic areas".

Changes in the distribution of the dentate gyrus associational system following unilateral or bilateral entorhinal lesions in the adult rat

The distribution of the dentate gyrus associational system was analyzed in naive adult rats and in those with either unilateral or bilateral lesions of the entorhinal cortex. Horseradish peroxidase histochemistry was used to trace the origin and course of this intrinsic fiber system. The fibers originated in the CA3-4 pyramidal cell field, apparently medial to the origin of the Schaffer collateral system, and followed a trajectory which was essentially identical to that described for this system by Zimmer36. The associational terminal field occupied the inner 26% of the dentate gyrus molecular layer in normal rats and 35-38% of the normal width of that layer following either ipsilateral or bilateral entorhinal lesion. These measurements are quite similar to those previously obtained on the commissural system terminal field in the normal and partially deafferented dentate gyrus. These results are interpreted to reflect axon sprouting by the associational fibers into the adjacent deafferented dendritic field.

The olivocerebellar projection in the cat studied with the method of retrograde axonal transport of horseradish peroxidase

The distribution of labeled cells in the inferior olive of the cat has been mapped following injections of small amounts of horseradish perosidase in the paramedian lobule of the cerebellum. The distribution of labeled cells was plotted in drawings of approximately serial transverse sections. The findings in each case were transferred to a standard diagram of the olive to facilitate comparison of cases. Previous studies of the distribution of retrograde cell loss in the inferior olive following cerebellar lesions (Brodal, '40b) showed that fibers ending in the paramedian lobule come from the caudal part of the ventral lamella of the principla olive. This was confirmed with the peroxidase method, but in addition three other separate and well circumscribed area of the olive showed labeling: one in the dorsal accessory olive, another in the rostral part of the medial accessory olive, a third in the caudal part of the dorsal lamella of the principal olive (fig. 7). There is some degree of topical arrangement within the projection of each of these olivary areas to the paramedian lobule. It is particularly striking that the projection areas of the caudal one-third of the lobule are different from and overlap only little with those of the orstral two-thirds. On account of diffusion of the injected perosidase solution in the folia it could not be decided whether the different olivary areas project to particular longitudinal zones in the paramedian lobule. The main findings can be correlated with the physiological observations of Armstrong et al. ('74). Some of the "paramedian" olivary areas are labeled also following peroxidase injections in other cerebellar parts, among them the nuclei interpositus anterior and posterior. The findings are compatible with the notion that olivocerebellar fibers branch to supply more than one cerebellar region. It is confirmed that the olivocerebellar projection, including that of the nuclei, is almost completely crossed. In the discussion it is emphasized that afferents from several sources converge on all four olivary regions projecting onto the paramedian lobule. The olivocerebellar projection obviously allows for divergence as well as convergence of impulses from the olive to the cerebellum. For further insight into the anatomical organization of the inferior olive, the entire olivocerebellar projection has to be mapped with the peroxidase methods, and further studies of the afferents to the olive are needed. In such studies, as well as in physiological ones, it is essential that findings are described with meticulous reference to the topography of the olivary subdivisions.

Studies on the cerebellar projections from the main and external cuneate nuclei in the cat by means of retrograde axonal transport of horseradish peroxidase

The cerebellar projections from the main and external cuneate nuclei in the cat have been studied by means of retrograde axonal transport of horseradish peroxidase. The main projection from the external cuneate nucleus (ECN) is to the intermediate and, possibly, the small lateral part of lobule V and to the paramedian lobule on the ipsilateral side. The projection from the ECN to the cerebellar regions mentioned is topographically organized. Cells in the caudal part of the ECN send their axons to the caudal parts of lobule V and to the rostral part of the paramedian lobule. Cells in the rostral part of the ECN project to the rostralmost part of lobule V and to the folia in the caudal part of the paramedian lobule. The experimental study also shows that cells in the main cuneate nucleus (MCN) send their axons to the cerebellum. These axons, like those from the ECN, terminate in the intermediate part of lobule V of the anterior lobe and in the paramedian lobule. However, the axons of the cells in the MCN terminate only in the superficial parts of the folia, whereas those from the ECN terminate in the depth of the folia in these two cerebellar areas. The present study also gives evidence that cells in the ventral part of the gracile nucleus send their axons to lobules I and II of the anterior lobe vermis. The observations referred to here are to our knowledge the first anatomical findings demonstrating a projection from the main cuneate and gracile nuclei onto the cerebellar cortex. The observations confirm previous physiological studies.

Electron microscopic observations of horseradish peroxidase transported from the caudoputamen to the substantia nigra in the rat: possible involvement of the agranular reticulum

The intracellular distribution of horeseradish peroxidase (HRP) transported intraaxonally from the caudoputaminal complex to the substantia nigra has been examined with the electron microscope. The reciprocal axonal connections between the caudoputamen and the substantia nigra permitted observation not only of HRP transported retrogradely from axons and axon terminals in the caudoputamen to the cell bodies of origin in the pars compacta of the substantia nigra, but also provided information suggesting that HRP may be transported anterogradely by neurons of the caudoputamen to their terminals, which are especially numerous in the pars reticulata of the substantia nigra. Special attention was focused on observations which might elucidate the manner in which exogenous proteins are compartmentalized and transported intracellularly. It is suggested that the agranular reitculum is involved in the retrograde transport of proteins which are pinocytosed near the axon terminal and ultimately reach lysosomes in the perikaryon. A possible anterograde movement of HRP may also involve the agranular reticulum. The implications such findings have on the use of HRP in neuroanatomical tracing techniques are also discussed.