Search

In order to determine the ongoing role of retinal fibers in the development of dorsal lateral geniculate nucleus (dLGN) neurons during postnatal development, the development of dLGN neurons in the postnatal absence of retinal input was studied in pigmented ferrets using the Golgi-Hortega technique. The development of four dLGN cell classes, defined on the basis of somatic and dendritic morphology, was described previously in normal ferrets (Sutton and Brunso-Bechtold, 1991, J. Comp. Neurol. 309:71-85). The present results indicate that the morphological development of dLGN neurons is strikingly similar in normal and experimental ferrets. The exuberant dendritic appendages that appear after eye opening in normal ferrets are overproduced and eliminated in the postnatal absence of retinal input; however, the final reduction of these transient appendages is delayed. Because exuberant appendages develop in the absence of retinal input, their production cannot depend upon visual experience. Differences in cell body size between normal and experimental ferrets are apparent only after neurons can be classified at the end of the first postnatal month. Cell body size is markedly reduced for class 1 neurons; class 2 cells also are reduced in size but to a far lesser extent. As there is a general trend for class 1 neurons to have the functional properties of Y-cells, it is likely that the dLGN neurons most affected by the absence of retinal input also are Y-cells.

In recent years, the important role of the extracellular matrix in neural development has been increasingly recognized. In order to begin to examine what role might be played by the extracellular matrix in the developing dorsal lateral geniculate nucleus (dLGN), the present study used immunocytochemistry to assess the distribution of a recognized extracellular matrix molecule, fibronectin (FN), during postnatal development of the ferret dLGN. Prior to the segregation of cell layers, no clear pattern of FN distribution can be distinguished within the dLGN. By P16, when layers A and A1 are separated by an interlaminar space, FN is localized in that space. By P24, FN is present not only between layers A and A1, but also within layers A and A1 as bands parallel to the laminar borders. These bands appear to correspond spatially and temporally to the development of sublaminar boundaries in the ferret (Hahm and Sur, Neurosci Abstr 14:460, 1988). By the end of the first postnatal month, immunoreactivity is diminished but still present. FN is no longer present at P44 or in the adult. Adjacent sections were incubated with antiserum to glial fibrillary acid protein (GFAP). When a laminar pattern of GFAP can be distinguished, it coincides with the interlaminar distribution of FN. The findings described here are consistent with a role for FN in the development of a laminar organization in the dLGN.

The development of neurons in the dorsal lateral geniculate nucleus (dLGN) of pigmented ferrets was studied by using the Golgi-Hortega technique. In adult ferrets, four dLGN cell classes were defined on the basis of somatic and dendritic morphology. Classes 1 and 2 were divided into stellate and oriented subtypes. Class 1 and 4 cells are characterized by filiform appendages, class 2 cells by club-like appendages, and class 3 cells by stalked appendages. At birth, dLGN neurons have simple dendritic arbors. During the first postnatal week, dendritic length and proximal branching density increase markedly. By postnatal day 21 (P21), dendritic morphology begins to take on mature characteristics and by the time of eye opening (P30-P35), most neurons can be classified. Also by that time, dLGN cells are covered with abundant filiform appendages. Developmental changes in appendage density were quantified for class 1 stellate cells. These data reveal that appendage density reaches a peak at P56, decreases sharply until P90, and then gradually declines to mature levels by P180. Elaboration and elimination of transient appendages occurs centrifugally; at maturity appendage density remains greater distally.

In normal development, cell layers in the dorsal lateral geniculate nucleus (dLGN) segregate from a relatively homogeneous cell group. If all retinal input is removed prior to this segregation, the layers fail to form. In the present study, we used ultrastructural and morphometric analyses to study dLGN development in the tree shrew following neonatal removal of retinal input. The goal of the present study was to determine whether there are differences between normal animals and enucleates in the development of dLGN cells and their interrelationships with each other and/or with the surrounding glia, which might explain the failure of cellular lamination in enucleated animals. The results indicate that although the development in enucleated animals may take place somewhat more slowly, by P90 cell size and density are not significantly different from normal. These results, coupled with the observation that the dLGN in enucleates is smaller than in normals, suggest that the removal of retinal input results in dLGN cell loss. At both the light and electron microscopic level, cells in the developing normal dLGN are arranged in bands of immediately adjacent cells. In enucleates, dLGN cells are less frequently in immediate contact and are arranged in small groups or clumps which may be separated by degenerating cells. The present data suggest that the presence of retinal input may be necessary to allow dLGN cells to maintain the intercellular relationships necessary for laminar segregation to take place.

In the present study, we examined the colliculogeniculate projection in normal adult tree shrews and in adults that were bilaterally enucleated at birth. We injected lectin-conjugated HRP into superficial superior colliculus and then mapped the pattern of anterogradely transported enzyme in the ipsilateral dLGN. In normal adult tree shrews, the results confirm that the colliculogeniculate projection is laminated and terminates predominantly in small-celled layers 3 and 6 and in the interlaminar space between layers 4 and 5 (Fitzpatrick et al., 1980); we report an additional sparse projection to layer 4. In bilaterally enucleated animals, the colliculogeniculate projection is unlaminated and tends to terminate in the lateral two-thirds of the dLGN even though synaptic sites are vacated throughout the nucleus. We suggest that this preference may be due to a specificity of the colliculogeniculate fibers for the lateral two-thirds of the dLGN, which, in normal adult tree shrews, contains cells with similar physiological characteristics. We further suggest that the normal lamination of the colliculogeniculate projection in the lateral two-thirds of the dLGN may be due to competition with retinogeniculate fibers so that colliculogeniculate fibers terminate predominantly in layers containing small, W-like cells.

In order to gain insight into the mechanisms involved in the formation of groupings of functionally similar cells in the developing nervous system, we have studied the formation of cell layers in the developing dorsal lateral geniculate nucleus (dLGN). To examine the possibility that a higher affinity or adhesion between cells in individual layers may play a role in laminar segregation, we studied cellular interrelationships in the dLGN of tree shrews before (P0), during (P4 and P7), and just after (P15) laminar segregation has taken place. We compared our observations at these stages of development with similar observations in the adult. In none of the cases do we see evidence of gap junctions either between adjacent neurons or between neurons and processes in the surrounding neuropil. However, we frequently observe the presence of puncta adherentes between adjacent neurons at all stages of development. These profiles are also present between neurons and cellular processes in the neuropil. We also see subsurface cisternae in all of our cases, although these are more pronounced before and during interlaminar space formation. As with the puncta adherentes, these are found both between adjacent neurons as well as between neurons and other elements in the neuropil. We also see some evidence of what appear to be cytoplasmic bridges between adjacent neurons; these are quite rare but appear to be present only before and during laminar segregation. Finally, we frequently see cytoplasmic processes interdigitated between otherwise immediately adjacent cells. These processes also are often found oriented along other portions of the neuronal plasmalemma. Whether these processes are portions of neuronal growth cones or glial processes is impossible to determine at this time. Because of the potential role glial processes may play in the formation and maintenance of laminar cell groupings during layer formation, we have also made a preliminary survey of whether glial cells can be distinguished ultrastructurally at the stages we have studied.

We demonstrated in a previous study that, following neonatal bilateral enucleation in the tree shrew, interlaminar spaces (ILSs) in the dorsal lateral geniculate nucleus fail to form. In the present study we sought to determine which aspects of ILS formation are dependent upon retinal input. Accordingly, we studied the degree of ILS formation in tree shrews which were bilaterally enucleated either during ILS formation on postnatal day 3 (P3) or just after all ILS were apparent but before they had reached a mature width (P15). Our results indicate that retinal input is necessary for the initial formation of each ILS, but that it is not required for the maturation or maintenance of ILSs which have already begun to form.

In the present study, we quantified the distribution of growth cones and synapses in 2 developing layers, as well as in the intervening interlaminar space of the dorsal lateral geniculate nucleus (dLGN) in tree shrews. Our goal was to gain insight into mechanisms involved in the segregation of dLGN cells into layers during development. We sacrificed tree shrews before (P0), during (P4 and P7), and after (P15) laminar segregation as well as at maturity (P90). The dLGN from each animal was sectioned horizontally, and all tissue for analysis was blocked from the middle third of the nucleus along the dorsoventral axis. Each micrograph was coded and blindly scored for the number of growth cones and synapses in layers 4 and 5 and the intervening interlaminar space. We also measured each growth cone and classified synapse type. Statistical analyses of these data reveal that neither growth cones nor synapses are significantly more common in the interlaminar space early in the period of laminar segregation (P4). By nearly a week after the interlaminar space can first be distinguished (P7), there are more growth cones in the interlaminar space than in the layers, but this difference is no longer present at (P15). These results suggest that, although neuropil development at the laminar borders may not play a role in the onset of laminar segregation, it may contribute to the widening of the interlaminar spaces once this process has begun. In addition, growth cones continuously decrease in number and become less bulbous and more linear in shape with development. Synapses, on the other hand, continuously increase in number with age and pass through a transient period characterized by heavy spinous terminations.

The role of retinogeniculate afferents in the development of patterns of connections between visual cortex and the lateral geniculate nucleus (LGN) was addressed by studying the effect of bilateral enucleation at birth on those patterns of connections in tree shrew. In normal adult tree shrews there are six LGN cell layers separated by cell-sparse interlaminar spaces. The reciprocal connections between the LGN and visual cortex are restricted to a column running across all six LGN layers; the geniculocortical projection arises from the cell layers while the corticogeniculate projection terminates primarily in the interlaminar spaces. At birth, when the experimental animals were bilaterally enucleated, the retinogeniculate fibers have begun to segregate by eye but neither the cytological characteristics of individual layers nor the interlaminar spaces have yet formed, and the corticogeniculate fibers have not entered the nucleus. Bilateral enucleation does not prevent the development of the cytological characteristics of individual layers but the interlaminar spaces do not develop. The results of [3H]proline/HRP injections into visual cortex in animals bilaterally enucleated at birth indicate that in the absence of retinogeniculate fibers, and thus interlaminar spaces, the corticogeniculate fibers do not concentrate at the laminar borders but instead spread across all six LGN cell layers. Despite the failure of this projection to concentrate at laminar borders, the corticogeniculate fibers do terminate within a restricted projection column.

We examined the ultrastructural development of the lateral geniculate nucleus (LGN) in postnatal tree shrews to distinguish which features, if any, show a correlation with the development of cell layers. Our data indicate that synaptogenesis has begun at birth (PO) which is prior to the development of cell layers. At postnatal day 8 (P8), when laminar borders can be distinguished, the majority of synaptic profiles are still immature. Although some mature synaptic profiles can be identified at P8 and especially at P15, complex synaptic arrangements characteristic of adults are absent at that time. Growth cones are present at all 3 ages but are less prevalent at P8 and P15 than at birth and appear to be present in slightly higher concentrations in the interlaminar spaces. These results suggest that LGN cell layer formation does not correlate with the beginning of synaptogenesis but that it may correlate with an increase in growth cones in the interlaminar spaces.

Three characteristics distinguish the six layers of the adult tree shrew dorsal lateral geniculate nucleus (LGN). First, interlaminar spaces divide the nucleus into cell layers. Second, input from the two eyes projects to the nucleus such that two layers (1 and 5) receive input from the ipsilateral retinal and four layers (2, 3, 4, and 6) receive input from the contralateral retina. Finally, distinct cytological features characterize individual layers. In this report, we describe the postnatal development of LGN layers in the tree shrew in terms of the development of these three characteristics. At birth, the nucleus appears homogenous in Nissl-stained sections. Thus, no interlaminar spaces are present and all cells look similar in shape, size, and staining intensity. However, autoradiographic data show that, at birth, retinal afferents are segregated in an adult-like pattern. Interlaminar spaces begin to be evident between layers innervated by opposite eyes on postnatal day 2. Several days later, interlaminar spaces between layers innervated by the same eye (i.e., the borders of layer 3) appear, while the others continue to widen. Although some cytological maturation begins before interlaminar space formation, it is not until interlaminar spaces are apparent that features such as differential staining intensity and cell size can be used to distinguish individual layers. The results suggest that the three characteristics that define LGN layers in the tree shrew may be temporally separate events in the developing nucleus. Thus, retinal afferents are segregated prior to interlaminar space formation which, in turn, is initiated prior to final maturation of the cytological features that characterize the cell layers. This may indicate a degree of developmental independence among these maturational events.