Terminal arbors of axons that have formed abnormal connections.

Ther terminal arbors of individual retinogeniculate axons that have been induced to grow into an inappropriate geniculate layer have been revealed for light and electron microscopic study by being filled with horseradish peroxidase. After a unilateral ocular enucleation in kittens, single axons from the surviving eye show terminal arbors not only within their own geniculate layers but also in the denervated layers. The new, abnormal arbors arise from the terminal segments of arbors that lie within the nondenervated layer and make patterns of synaptic contacts that appear normal.

Qualitative and quantitative analyses of the patterns of retinal input to neurons in the dorsal lateral geniculate nucleus of the cat.

In the visual system of the cat the projection from the retina to the lateral geniculate nucleus has been studied extensively. However, the patterns of synaptic contacts made by individual axons onto individual cells have not been described. In this study these patterns have been examined for class 1 cells (Guillery: J Comp Neurol 128:21, '66). Retinogeniculate axons and lateral geniculate neurons are labeled with horseradish peroxidase (HRP) via injections into the optic tracts and optic radiations, respectively. Sections are then processed for combined light and electron microscopic analysis. They are examined with the light microscope to identify labeled lateral geniculate neurons that appear to be contacted by labeled retinal axons. These cells and axons are then analyzed by a computerized microscope system, and sites of apparent synaptic contact are recorded. This light microscopic analysis indicates that individual class 1 cells are contacted by many retinogeniculate axons (> 10) and that each of these axons contacts many lateral geniculate neurons (> 20). Some axons make numerous contacts that are concentrated onto a few dendrites, while others make only a few contacts, which are spread over several dendrites. In all cases, the majority of contacts are on the dendritic shafts of relatively thick secondary and tertiary dendrites. Electron microscopic analysis confirms that most of the contacts identified with the light microscope are synaptic. It also reveals that labeled and unlabeled retinal axons can innervate the same dendritic segment. Finally, one cell was studied that had its soma and most of its dendrites in lamina A1 but some of its dendrites extended into lamina A. This cell received input from retinal axons in both layers, thus suggesting that it may have been binocularly excitable.

The morphology of corticofugal axons to the dorsal lateral geniculate nucleus in the cat.

The structural features of corticogeniculate axons were studied in adult cats after labeling them with horseradish peroxidase (HRP). Injections of HRP into the optic radiations near the dorsal lateral geniculate nucleus result in Golgi-like filling of both geniculate relay neurons and corticogeniculate axons. In the present material at least two main types of axons could be defined. The most common type is called the type I axon because it so closely resembles the type I axons described by Guillery ('66, '67) in Golgi preparations. These fine axons have smooth surfaces and consistent fiber diameter. Most terminal swellings are at the ends of short collateral branches and these swellings form asymmetric synaptic contacts onto small and medium-sized dendrites. Type I axons typically innervate more than one lamina as well as interlaminar zones and they clearly arise from the cerebral cortex. The second type of axon is called the beaded axon because of its numerous swellings, en passant. These swellings frequently are larger than those on type I axons and they differ from previously described corticogeniculate axon terminals in their ultrastructural features. That is, their synaptic contacts appear symmetrical and they form axosomatic contacts. Because of these differences, the possibility that beaded axons are of subcortical origin, particularly from the perigeniculate nucleus, is discussed. When type I axons and geniculate relay neurons are filled in the same region of the nucleus it is possible to identify probable sites of synaptic contact by using the light microscope. Such analyses indicate that corticogeniculate axons synapse directly onto relay cells, primarily on peripheral dendritic branches. Further, it appears that single axons contact many geniculate neurons and that single neurons are contacted by many axons.

Abnormal axonal growth in the dorsal lateral geniculate nucleus of the cat.

Retino-geniculate axons in the cat were induced to grow abnormally by cutting one optic nerve in kittens. Surviving optic tract axons that had grown into the denervated regions were then filled in the adults with horseradish peroxidase to reveal the terminal arbors of individual axons. Two types of abnormal axonal growth are described--translaminar growth and monocular segment growth. Translaminar growth is the most common and occurs between laminae in the binocular part to the nucleus. Axons giving rise to translaminar growth do not branch as they pass through the denervated regions of the nucleus. Instead, the abnormal branches originate from portions of the terminal arbor within the normal target lamina. These axons look like normal retino-geniculate axons in terms of their branching patterns, cytological features, and patterns of synaptic contacts except that parts of their terminal arbors have expanded to innervate inappropriate laminae. The distribution of translaminar branches overlaps the distribution of a restricted group of surviving large neurons that have not undergone denervation atrophy. Monocular segment growth invades the lateral pole of the nucleus directly from the optic tract. These branches arise from axons passing through or near the denervated region and appear to represent the formation of new terminal arbors. The synaptic swellings arising from these branches have cytological features like the synaptic swellings arising from translaminar branches and they form similar patterns of synaptic contacts. However, monocular segment branches degenerate more rapidly when damaged and they are not associated with surviving large neurons.

Reconstructions of corticogeniculate axons in the cat.

The terminal arbors of corticofugal axons to the dorsal lateral geniculate nucleus in the cat were filled with horseradish peroxidase and then partially reconstructed through serial sections. The results demonstrate that these arbors are far more complex than was suspected from previous studies of axon segments in individual sections. These axons branch profusely and spread widely within the nucleus. Within laminae A and A1 the terminal arbor of a single axon can be more than 800 micron wide compared with retinogeniculate axons whose terminal arbors range in width from 100 to 410 micron (Sur and Sherman, '82).

Expression of a keratin sulfate proteoglycan during development of the dorsal lateral geniculate nucleus in the ferret.

ABAKAN is a keratin sulfate proteoglycan that was identified in rat brain by monoclonal antibody TED15 (Geisert et al. [1992] Brain Res. 571:165-168). It blocks neuronal attachment and neurite outgrowth in culture, is associated with astrocytes, and marks the boundaries of areas in the developing rat brain (Geisert and Bidanset [1993] Dev. Brain Res., 75:163-173). In the present study TED15 was used to examine the distribution of ABAKAN during laminar development of the dorsal lateral geniculate nucleus in ferrets. This distribution was also compared with that of astrocytes as displayed with antibodies to GFAP. In the adult, TED15 and anti-glial fibrillary acidic protein (GFAP) labeling are similar. Both are fairly uniform in the nucleus although somewhat elevated near the optic tract and in the interlaminar zone between laminae A and A1. During development the pattern is quite different. At postnatal day 1 (P1), before lamination is evident, TED15 and anti-GFAP labeling are light in the nucleus. By P10, when laminae are emerging, both are elevated in the A-A1 interlaminar zone and in the C laminae. At P18, when laminae are distinct, TED15 labels the A-A1 interlaminar zone, and it marks the borders between the ON and OFF leaflets within A and A1 (Stryker and Zahs [1983] J. Neurosci. 3:1943-1951). In comparison, anti-GFAP marks the interlaminar zone but not the ON/OFF leaflets. By 6 weeks the nucleus resembles the adult nucleus. These results show that ABAKAN marks the boundaries of the major functional subdivisions of the lateral geniculate nucleus in the developing ferret and suggest that it plays a role in lamination.

The organization of the pulvinar in the grey squirrel (Sciurus carolinensis). I. Cytoarchitecture and connections.

The posterior neocortex in the grey squirrel, Sciurus carolinensis, includes an extensive region which receives projections from the pulvinar. Previous studies have demonstrated that this cortical region can be subdivided on the basis of differences in cytoarchitecture and electrophysiologically defined representations of the visual field. The main purpose of the present paper was to determine whether these cortical subdivisions could be related to corresponding subdivisions in the pulvinar. The methods used to trace connections included anterograde degeneration, anterograde axonal transport of tritiated amino acids and the retrograde axonal transport of horseradish peroxidase. The results indicate that the pulvinar in this species contains at least three main subdivisions which can be distinguished by their cytoarchitecture and their patterns of connections. A caudal subdivision contains large, evenly-spaced neurons and receives bilateral input from the superficial, retinal-recipient layers of the superior colliculus. This caudal subdivision has reciprocal interconnections with a cytoarchitectonically distinct area in the temporal cortex. A rostro-lateral subdivision contains smaller, more lightly stained neurons which tend to form clusters. This subdivision receives only ipsilateral tectal input and projects to occipital area 18. This subdivision does not receive input from areas 17, 18, and 19, or from the temporal cortex. Finally, a rostro-medial subdivision is cytoarchitectonically similar to the rostro-lateral subdivision but receives little, if any, input from the superior colliculus. This rostro-medial area does, however, receive corticofugal projections from occipital areas 17, 18, and 19, and projects to area 19. These patterns of connections suggest that each of these subdivisions has close associations with the visual system. The question of whether similar subdivision are present in the visual thalamus of other species is discussed.

The organization of the pulvinar in the grey squirrel (Sciurus carolinensis). II. Synaptic organization and comparisons with the dorsal lateral geniculate nucleus.

The purpose of these experiments was to compare the synaptic organization of the subdivisions of the pulvinar defined in the preceding paper (Robson and Hall, '77) with each other and with the organization present in the dorsal lateral geniculate nucleus. The electron microscope was used to analyze normal synaptic arrangements and degenerating axonal terminals resulting from lesions. The dorsal lateral geniculate nucleus in the grey squirrel contains synaptic clusters similar to those described previously for other species. These clusters are characterized by large optic tract terminals which form multiple contacts onto large dendritic processes and other processes containing flat or pleomorphic vesicles. The geniculate lamina adjacent to the optic tract receives projections from the superior colliculus as well are from the retina. The terminals of the superior colliculus axons are small and medium sized and lie outside of the synaptic clusters. The retinal terminals are in the clusters. In the pulvinar, the rostro-medial subdivision contains synaptic clusters which resemble those in the lateral geniculate nucleus. These clusters contain large axon terminals which make multiple contacts onto large dendrites. However, these terminals are not contributed by an ascending sensory pathway but by axons from striate cortex. The rostro-lateral and caudal subdivisions of the pulvinar also contain synaptic clusters, but these clusters consist of a segment of a large dendrite which is ensheathed by medium-sized terminals. Since only a few of these medium sized terminals in any one cluster degenerate after tectal lesions, and none degenerate after cortical lesions, it is suggested that the morphological arrangement of these clusters may permit the convergence of axons from several sources, some of which are unidentified, onto the same dendritic segment.

The effects of monocular deprivation on the size of GAD+ neurons in the cat's dorsal lateral geniculate nucleus.

Previous studies have shown that suturing one eyelid closed in a newborn kitten results in profound changes in the development of the visual system. Among these is a retardation in the growth of neurons in the layers of the dorsal lateral geniculate nucleus receiving retinal input from the closed eye. Moreover, the greatest effect appears to be in the largest neurons. The present study examines the effects of monocular deprivation on the perikaryal size of a select group of small lateral geniculate neurons. GABAergic neurons, that may be interneurons. These cells were selectively labeled by an antiserum to glutamic acid decarboxylase (GAD) and immunocytochemical methods. The results demonstrate that GAD+ neurons are among the smallest in the lateral geniculate nucleus and that they are insensitive to the effects of monocular deprivation. That is, GAD+ neurons in deprived laminae A and A1 are similar in size to those in the corresponding, nondeprived laminae. These findings are consistent with the hypothesis that GAD+ neurons are interneurons and therefore not subject to binocular competition in the visual cortex. This interpretation, however, is complicated by additional studies of the postnatal development of GAD+ neurons which reveal that GAD+ neurons grow to their adult size relatively early, before the onset of the critical period. Thus the insensitivity of the perikarya of GAD+ neurons to monocular deprivation may be attributable to their precocious growth.

Ganglion cell neurogenesis, migration and early differentiation in the chick retina.

Neurogenesis, migration and maturation of ganglion cells in the posterior pole of chick retina have been studied using embryonic incorporation of [3H]thymidine, immunocytochemistry and retrograde labeling. Unlike previous studies, we have examined the neurogenesis of independently identified ganglion cells that have survived the period of naturally occurring cell death (embryonic days 11-16). Embryos were labeled with [3H]thymidine at different embryonic ages (embryonic days 3, 5 and 7). After the chicks hatched, ganglion cells were retrogradely labeled with rhodamine microspheres and the retinas were processed for autoradiography and fluorescent microscopy. The results indicate that 40% of the ganglion cells in the posterior pole undergo a final mitosis by embryonic day 3 and that more than 25% of the ganglion cells are born on or after embryonic day 7. These results also suggest that naturally occurring cell death does not preferentially affect ganglion cells born on specific embryonic days. Using immunocytochemistry with an antibody against neuron-specific beta-tubulin and retrograde labeling with the carbocyanine dye DiI we show that ganglion cells begin to differentiate before the completion of their migration to the presumptive ganglion cell layer. These results suggest the following developmental sequence. (1) Ganglion cells of the posterior pole undergo their final mitosis near the ventricular margin between embryonic days 2 and 8. (2) They maintain contacts with both retinal surfaces and their nuclei move toward the ganglion cell layer. At this time they start to differentiate, expressing a form of neuron-specific tubulin and growing axons that can reach the optic chiasm. (3) Once migration is completed dendritic development commences.

Ultrastructural features of a human neuroblastoma cell line treated with retinoic acid.

This report examines the morphological changes that occur in a line of human neuroblastoma cells (LA-N-5) following treatment with retinoic acid, in vitro. The results demonstrate that retinoic acid induces pronounced differentiation of these cells. Perikarya aggregate into tight clusters and extend long processes that are frequently fasciculated. Growth cones appear at the ends of these processes. Transmission electron microscopy reveals that after 10 days of treatment these long neurites give rise to varicosities which contain clusters of large dense-core vesicles and smaller clear vesicles. After 18 days of treatment the cultures cease to differentiate further. The pattern of neurite outgrowth is very complex by this point and the frequency of growth cones and vesicle-containing varicosities is greatly increased compared with shorter treatments. Most of these varicosities contain a mix of large dense-core vesicles and smaller clear vesicles and in some profiles the clear vesicles are round while in others they are pleomorphic. Despite this increase in the number of vesicle-containing profiles no membrane specializations were seen that resemble mature synapses. The present results demonstrate that retinoic acid can produce morphological changes in these cells in culture, and that these changes closely mimic those of normal differentiating neurons in culture. Considered with previous studies, these findings suggest that this cell line might provide a useful model system for studying neural differentiation.

An autoradiographic analysis of neurogenesis in the chick retina in vitro and in vivo.

Patterns of [3H]thymidine incorporation during neurogenesis of the embryonic chick retina have been compared in vitro and in ovo. Pieces of posterior, undifferentiated retinas were dissected from embryos on day 6 of incubation (E6) and cultured in the presence of [3H]thymidine. Label was added to the medium for 3 h on day 1, 2, 3 or 4 in culture. The retinas were fixed on the fifth day, embedded in epon, sectioned and processed for autoradiography. In parallel experiments, in ovo injections were made on embryonic day 6, 7, 8 or 9 (E6-E9). On E12 the embryos were fixed and a piece of the posterior retina from each eye was dissected and processed for autoradiography as above. Results show that the retinal explants develop well in culture and all of the layers of the neural retina differentiate. However, the cultured retinas are thinner than those grown in ovo. [3H]Thymidine labeling indicates that nearly all retinal neurons undergo their final mitotic divisions between E6 and E9. In addition the patterns of labeling in culture are similar to those in ovo. Most neurons, including the majority of cells in the ganglion cell layer and outer nuclear layer, are labeled on the first three days in culture and in E6-E7 embryos, while labeled cells are restricted to the inner nuclear layer in older specimens. Counts of labeled and unlabeled neurons in the ganglion cell layer suggest that the temporal pattern of neurogenesis in culture lags behind that in the embryo by about one day but that the spatial patterns of cell migration are the same.

Displaced ganglion cells in the rabbit retina.

Displaced ganglion cells were studied in the rabbit retina after filling them with horseradish peroxidase via injections into the optic nerve. These cells are primarily in areas outside of the visual streak. They are larger than other inner nuclear layer cells and are the size of the smallest neurons in the ganglion cell layer. Large labeled cells (greater than 20 microns) are found only in the ganglion cell layer. These findings further illustrate the wide variability in the size and distribution of displaced ganglion cells in the retinas of different mammals.

New method of transplanting purified glial cells into the brain.

This paper describes a new transplantation method for testing the ability of purified populations of glial cells to support axonal growth in the brains of adult animals. Thin tubes, rolled from porous polycarbonate film, are coated with poly-L-lysine and filled with cultured Schwann cells. Schwann cell-filled tubes or control tubes (poly-L-lysine coated only) are then implanted into the brains of adult rats so that one end of the tube is in the thalamus and the other extends extracranially. After survival times of 4-16 weeks horseradish peroxidase (HRP) is applied to the extracranial end of the tube. One or two days later the animal is perfused and the brain is sectioned and processed histochemically. Results show that tubes containing Schwann cells are densely filled with tissue and are well vascularized. Further, neurons in the central nervous system are retrogradely labeled with HRP and most labeled cells are concentrated in regions of the diencephalon near the end of the tube. Control tubes contain very little tissue and show no evidence that they support axonal growth. These results are consistent with the hypothesis that Schwann cells can support axonal growth in the brains of adult rats.

Organization of the dorsal lateral geniculate nucleus in a cat with congenital microphthalmia.

The effects of congenital monocular microphthalmia on the development of the lateral geniculate nucleus were examined in a 10-week-old cat. The left eye and optic nerve in this animal appear normal. The right eye is about 30% smaller in volume than the left and the optic nerve from this eye has a cross-sectional area that is only 15% that of the left. In addition, this nerve contains few, if any, large myelinated axons. Both lateral geniculate nuclei are abnormal and the abnormality differs rostrally and caudally. The caudal portion most closely resembles the normal nucleus. Retinal input from both eyes is segregated into cellular laminae that are separated from each other by cell sparse interlaminar zones. However, the input from the microphthalmic eye seems to be sparse and patchy and it does not support normal cell growth. All neurons, including glutamic acid decarboxylase-positive (GAD+) neurons, in laminae innervated by the small eye are reduced in size in a pattern similar to that seen following the removal of retinal input. In comparison, the rostral portion of the nucleus receives very little input from the microphthalmic eye. Instead the normal eye densely innervates nearly the entire nucleus. In this region, interlaminar zones fail to form but the input from the normal eye is able to support cell growth including the growth of GAD+ neurons.

Up-regulation of a keratan sulfate proteoglycan following cortical injury in neonatal rats.

The up-regulation of the keratan sulfate proteoglycan (ABAKAN) was examined using indirect immunohistochemical methods. Previous studies indicate that the keratan sulfate proteoglycan is associated with astrocytes in the optic nerve and in the developing rat brain. In model culture systems, this proteoglycan is capable of inhibiting the growth of neurites over laminin. To determine whether the proteoglycan is up-regulated specifically during reactive gliosis, stab wounds were made in the cerebral cortex of early postnatal rats, and the up-regulation of the proteoglycan was related to the developmentally regulated gliotic response to injury. Following a stab wound in the cortex of the late postnatal rat, reactive gliosis was consistently observed along with an up-regulation of ABAKAN. When the cortex was injured on postnatal day 2, there was a variable gliotic response and considerable variation in the regulation of proteoglycan expression. Biochemical analysis revealed that ABAKAN is a large proteoglycan with multiple keratan sulfate side-chains, at least one chondroitin sulfate side-chain and at least one additional carbohydrate chain with a terminal 3-sulfoglucuronic acid. Taken together, these data demonstrate that the boundary proteoglycan ABAKAN is also associated with reactive gliosis during early postnatal development.