Changing course of growing axons in the optic chiasm of the mouse

The distribution of retinofugal fibres has been studied by electron microscopy throughout the extent of the developing mouse optic nerve and chiasm at embryonic day (E) 16, in order to determine the course of fibre growth. Growth cones and mature axons, which are randomly distributed in bundles in the extracranial optic nerve, segregate in the juxtachiasmatic optic nerve. Here, growth cones accumulate in subpial regions amongst the endfeet of radial glia, whereas axons lie in the depths of the nerve. Surprisingly, however, growth cones move away from this region toward the ventricular zone in the lateral and midline parts of the chiasm, only to return to subpial regions once more before entering the optic tract, where fibres are again in an age-related order. Superficially, mature axons mingle with growth cones in the chiasm and near the beginning of the optic tract, suggesting that the age-related order begins to be reestablished before growth cones enter the tract. Deep and superficial regions of the pathway were examined in different planes of section. Specialised membrane relationships between retinofugal fibres and radial glial cells were also studied in deep and superficial regions of the lateral part of the chiasm. In addition, the distribution of retinofugal fibre bundles in the adult mouse was looked at by using light microscopy. The changing fibre positions noted in the embryo are maintained in the adult. J. Comp. Neurol. 379:495–514, 1997. © 1997 Wiley-Liss, Inc.     

European Neuroscience Association The changing pattern of fibre bundles that pass through the optic chiasm of mice

The organization of retinofugal fibres in the developing and adult mouse has been studied with transmission electron microscopy, autoradiography and the Bodian silver method. It has previously been shown that all retinal ganglion cell axons are in glial-wrapped bundles in the developing and adult optic nerve, but are not in similar bundles close to the chiasm. In the embryonic mouse this region shows a transition in glial morphology from an interfascicular to a radial type and here retinofugal fibres begin to form a new order related to their age. Growth cones become concentrated at the pial surface of the juxtachiasmatic nerve and older fibres are restricted to deeper regions. This same age-related order is also evident in the optic tract. However, the age-related order is lost within the chiasm, where growth cones, young and old fibres are again mingled in distinct bundles as they cross the mid-line. This study is particularly concerned with the structure of the mid-line bundles. These fibre bundles cross each other at right angles, and are recognizable in fetal and adult mice. In the adult, monocular injections of H³ proline followed by autoradiographic study show that the individual mid-line bundles are monocular and that they fuse again, losing the fascicular structure as they leave the chiasm and enter the tract. In the fetus and in the adult, the bundles generally lack a complete glial wrapping so that growth cones can lie in intimate contact with two crossing bundles, one coming from the left eye, the other from the right. The interesting question about the mechanisms that keep growth cones from entering the wrong bundles when they are in this position remains to be addressed.     

Observations on the early development of the optic nerve and tract of the mouse.

The early development of the retinofugal pathway of mice has been studied by light and electron microscopic methods in order to define the spatial distribution and the structure of the growth cones as they advance from the eye to the brain. We have studied the relationships of the growth cones to each other, to the glia and, in the older individuals, to the nerve fibers that are already terminating in the brain. We have looked at the rate of advance of the growth cones and have paid particular attention to the changing relationships of the growth cones as they approach the optic chiasm. We have also looked to see whether, at early stages, it is possible to recognise any characteristic features distinguishing the fibers destined to be the thickest in the adult, which come from ganglion cells that are generated among the earliest ganglion cells. In transverse sections through the optic stalk about 50-100 microns behind the eye, the first bundles of fibers are seen on embryonic day 12.5 (E12.5) as a mixture of thin (less than 0.5 micron) axons, thicker growth cones, and fine filopodial and foliopodial extensions. During the next two days, as these bundles in the intraorbital nerve increase in size and number, growth cones can be seen in all of the bundles and in all parts of the bundles. They show only a slight preference for one part of the nerve relative to another, and our material provides no evidence for the view that axons are particularly inclined to follow pre-existing bundles. The structure of the pathway changes significantly as it is traced towards the chiasm, and no section or small stretch of sections can be regarded as representative of the nerve as a whole. As the fibers approach the optic chiasm the growth cones come to lie predominantly close to the pial surface, with the deeper regions occupied almost entirely by fine axons. The change occurs in a region where the glial environment also changes, and where a characteristic neural tube-like organization first becomes recognizable. Here the glial cells lie in a periventricular position and send slender radial processes out towards the subpial surface. The newly invading axons in the early optic nerve taper from a broad growth cone back to an extremely slender axon, less than 0.5 micron in diameter. The tapered region is of the order of 100-300 microns in length and advances through the nerve at approximately 60 microns per hour.(ABSTRACT TRUNCATED AT 400 WORDS)     

Early development of the optic chiasm in the gray short-tailed opossum,Monodelphis domestica

We have studied the early development of the uncrssed retinofugal projection in the gray short-tailed opossum. Axons that form the adult uncrossed retinofugal projection arise from the temporal crescent of the retina and reach the optic chiasm on postnatal day 7. The sites at which the uncrossed fibres segregate from the crossed fibres and the pattern of this segregation are very different from those seen in eutherian mammals. In the opossum, the uncrossed fibres segregate from the crossed fibres within the juxtachiasmatic part of the optic nerve before they have encountered either the fibres of the other eye or midline structures of the ventral diencephalon. The uncrossed fibres turn perpendicular to the axis of the nerve and grow dorsoventrally through the crossed projection to gather as a discrete bundle at the ventral edge of the nerve. The abrupt divergence of the uncrossed fibres occurs at a border between two glial cell types: the interfascicular glia that characterise the main part of the optic nerve and the radial glia of the juxtachiasmatic part of the nerve. At the ventral part of the nerve, the bundle of uncrossed fibres turns caudally across the axis of the nerve and enters the ipsilateral optic tract. When retinofugal fibres encounter the border between the interfascicular and radial glia, a very specific axonal reorganisation occurs in marsupials, and this is strikingly different from the axonal reorganisation that occurs at the same site in eutherians, where essentially all retinofugal fibres reorganise, not just the uncrossed component. We believe this to be an important example of an identified cellular element that has quite distinct axon-guidance properties in different species. © 1994 Wiley-Liss, Inc.     

Localization of Nogo and its receptor in the optic pathway of mouse embryos

We have investigated the localization of Nogo, an inhibitory protein acting on regenerating axons in the adult central nervous system, in the embryonic mouse retinofugal pathway during the major period of axon growth into the optic chiasm. In the retina, Nogo protein was localized on the neuroepithelial cells at E12 and at later stages (E13-E17) on radial glial cells. Colocalization studies showed expression of Nogo on vimentin-positive glia in the retina and at the optic nerve head but not on most of the TuJ1- and islet-1-immunoreactive neurons. Only a few immature neurons in the ventricular and peripheral regions of the E13 retina were immunoreactive to Nogo. In the ventral diencephalon, Nogo was expressed on radial glia, most strongly on the dense radial glial midline raphe within the chiasm where uncrossed axons turn and in the initial segment of the optic tract. In vitro studies showed that the Nogo receptor (NgR) was expressed on the neurites and growth cones from both the ventral temporal and dorsal nasal quadrant of the retina. In the optic pathway, NgR staining was obvious in the vitreal regions of the retina and on axons in the optic stalk and the optic tract, but not in the chiasm. These expression patterns suggest an interaction of Nogo with its receptor in the mouse retinofugal pathway, which may be involved in guiding axons into the optic pathway and in governing the routing of axons in the optic chiasm.     

Enzymatic removal of chondroitin sulphates abolishes the age-related axon order in the optic tract of mouse embryos

Retinal axons undergo an age-related reorganization at the junction of the chiasm and the optic tract. We have investigated the effects of removal of chondroitin sulphate on this order change in mouse embryos aged embryonic day 14, when most axons are growing in the optic tract. Enzymatic removal of chondroitin sulphate but not keratan sulphate in brain slice preparations of the retinofugal pathway abolished the accumulation of phalloidin-positive growth cones in the subpial region of the optic tract. The loss of chronotopicity was further demonstrated by anterograde filling of single retinal axons, which showed a dispersion of growth cones from subpial to the whole depth of the tract. The enzyme treatment neither produced detectable changes in growth cone morphology and growth dynamic of retinal neurites nor affected the radial glial processes in the tract, indicating a specific effect of removal of chondroitin sulphate from the pathway to the axon order in the tract. Although chondroitin sulphate was also found at the midline of the chiasm, growth cone distribution across the depth of fibre layer at the midline was not affected by the enzyme treatment. These results suggest a mechanism in which retinal axons undergo changes in response to chondroitin sulphate at the chiasm-tract junction, but not at the midline, that produce a chronotopic fibre rearrangement in the mouse retinofugal pathway.     

Palisading pattern of subpial astroglial processes in the adult rodent brain: relationship between the glial palisading pattern and the axonal and astroglial organization

The architectural organization of the subpial astrocyte processes was examined near the brain surface by single immunostaining methods. The astroglial processes were stained on brain sections made parallel to the pial surface. The astroglial glial fibrillary acid protein (GFAP) antigen was used as a specific marker. We show that these subpial astrocyte processes present a well organized palisading pattern in the adult mouse and rat spinal cord, medulla and pons. This adult astrocyte palisading pattern is compared to the palisading radial glia organization we previously demonstrated in the fetal mouse brain. The observed analogies afford a new and strong argument in favor of a derivation of the subpial astrocytes from radial glia. Double immunostaining methods, using GFAP and neurofilament antigens as glial and neuronal markers respectively, show the close relationship existing between the trajectories of axonal and glial processes. Beside the colinearity already observed between the axon trajectories and the glial palisades we demonstrate a new kind of axon/glia relationship. Axons are closely intermingled, within the palisading glial tufts, with the peripheral processes of the subpial astrocytes progressing to the pial surface. The findings suggest that fetal radial glia organization has a direct and indirect influence on the adult astroglial and perhaps the axonal pattern.     

Early uncrossed component of the developing optic nerve with a short extracerebral course: a light and electron microscopic study of fetal ferrets

During the study of the developing optic nerve described in the preceding paper (Guillery and Walsh, '87), small bundles of nerve fibers were seen passing between the optic nerve and the ipsilateral hypothalamus of 24-to 27-day-old prenatal ferrets. The bundles appear before any other fiber groups of the retinofugal pathway and are identifiable while the main portions of the retinofugal system are growing into the optic tracts. The bundles, made up of 50 or more axons, leave the optic nerve, emerge through the otherwise continuous layer of subpial glia and through the basal lamina of the nerve, run a short, naked, extracerebral course among collagen fibers and presumed fibroblasts, and then re-enter the central nervous system, passing rostrally and dorsally to the superficial parts of the ipsilateral hypothalamus away from the region of the chiasm. These fibers represent the earliest link between the optic nerve and the brain, but their course is not followed by the majority of retinofugal fibers developing later, which pass toward one or the other optic tract.     

Heparan sulfate proteoglycan expression in the optic chiasm of mouse embryos.

Previous studies have demonstrated that heparan sulfate (HS) proteoglycans (PGs) regulate neurite outgrowth through binding to a variety of cell surface molecules, extracellular matrix proteins, and growth factors. The present study investigated the possible involvement of HS-PGs in retinal axon growth by examining its expression in the retinofugal pathway of mouse embryos by using a monoclonal antibody against the HS epitope. Immunoreactive HS was first detected in all regions of the retina at embryonic day (E) 11. The staining was gradually lost in the central regions and restricted to the retinal periphery at later developmental stages (E12--E16). Prominent staining for HS was consistently found in the retinal fiber layer and at the optic disk, indicating a possible supportive role of HS-PGs in axon growth in the retina. At the ventral diencephalon, immunostaining for HS was first detected at E12, before arrival of any retinal axons. The staining matched closely the neurons that are immunopositive for the stage-specific embryonic antigen 1 (SSEA-1). At E13 to E16, when axons are actively exploring their paths across the chiasm, immunoreactivity for HS was particularly intense at the midline. This characteristic expression pattern suggests a role for HS-PGs in defining the path of early axons in the chiasm and in regulating development of axon divergence at the midline. Furthermore, HS immunoreactivity is substantially reduced at regions flanking both sides of the midline, which coincides spatially to the position of actin-rich growth cones from subpial surface to the deep regions of the optic axon layer at the chiasm. Moreover, at the threshold of the optic tract, immunoreactive HS was localized to deep parts of the fiber layer. These findings indicate that changes in age-related fiber order in the optic chiasm and optic tract of mouse embryos are possibly regulated by a spatially restricted expression of HS-PGs.     

Expression of chondroitin sulfate proteoglycans in the chiasm of mouse embryos

Chondroitin sulfate (CS) proteoglycans have been implicated as molecules that are involved in axon guidance in the developing neural pathways. The spatiotemporal expression of CS was investigated in the developing retinofugal pathway in mouse embryos by using the CS-56 antibody. Immunoreactive CS was detected in inner regions of the retina as early as embryonic day 11 (E11). Its expression in subsequent stages of development followed a centrifugal, receding gradient that appeared to correlate with the sequence of axogenesis in the retina. In the chiasm, immunoreactive CS was expressed at E12, before the arrival of retinal axons. When the retinal axons navigated in the chiasm at E13-E14, immunoreactive CS remained at a low level in the optic fiber layer of the chiasm but was observed prominently in the caudal parts of the ventral diencephalon. This pattern followed closely the array of stage-specific-embryonic-antigen-1-positive neurons in the ventral diencephalon, with a V-shaped configuration that bordered the posterior boundary of the retinal axons, and a rostral raphe extension that ran across the decussating axons in the chiasm. Thus, the CS epitope is implicated in patterning the course of early retinal axons and in regulating axon divergence in the chiasm. At the lateral region of the chiasm, where the retinal axons cross the midline and approach the optic tract, a CS-immunopositive region coincided with the region in which active sorting of dorsal retinal axons from ventral retinal axons occurs. Moreover, at the threshold of the optic tract, the immunoreactive CS was restricted only to the deep part of the optic fiber layer, suggesting an inhibitory role of the CS epitope in repelling newly arrived axons to superficial regions of the optic tract during the development of chronotopic order at this part of the retinofugal pathway.     

Growth cone morphology varies with position in the developing mouse visual pathway from retina to first targets

We have labeled the growth cones of retinal ganglion cell axons with HRP in intact mouse embryos. This has allowed us to visualize growth cone morphology during outgrowth along an entire CNS pathway from origin to target; to ask whether growth cone forms, and thus behaviors, differ at various points along the pathway; and to study the relationships of growth cones with the cellular environment. During the major period of axon outgrowth between embryonic day (E) 12 and 15, growth cones in the optic nerve are highly elongated (up to 40 microns) and have lamellopodial expansions, but the majority lack the microspikes or filopodia characteristic of many growth cones. Within the optic chiasm (E13-15), most growth cones shorten and spread, and project several short filopodia. In the optic tract, growth cones become more slender and again lack filopodia, resembling sleeker versions of optic nerve growth cones. Near the first target region (lateral geniculate nucleus), growth cones with filopodia arise from individual axon lengths and turn medially toward the target. Within target regions, the branches of immature axon arbors are tipped by minute swellings rather than by the enlarged growth cones prevalent during outgrowth toward targets. Electron-microscopic analysis of identified labeled growth cones in the optic nerve reveal intimate interactions between growth cones and glia or other growth cones in the form of invaginating contacts. In the optic nerve, growth cones contact immature glial (neuroepithelial) cells somewhere along their length, and also envelop bundles of neurites. In the chiasm, single growth cones simultaneously relate to many different profiles. These results demonstrate that in this single pathway from origin to targets, growth cone morphology varies systematically with position along the visual pathway. During outgrowth, simple growth cones are prominent when axons follow well-defined common pathways, and more elaborate filopodial forms appear when growth cones diverge, as they turn or come to decision regions. Together with observations in vitro and in nonmammalian nervous systems in situ, these data serve as reference points for testing to what extent growth cone form reflects intrinsic factors and interactions with the environment.     

Age-related fiber order in the ferret's optic nerve and optic chiasm

Although the mammalian optic tract shows a grouping of fibers by age, with newer fibers nearer the pial surface, the possible rules for fiber ordering in the mammalian optic nerve have not been well defined. In this study, preferential labeling of the older retinal fibers in the ferret, a close relative of the cat, shows that the age-related fiber order in the ferret's optic tract reflects a systematic sorting of fibers by age that occurs in the optic nerve, and that is maintained through the optic chiasm. The older retinofugal fibers, dispersed throughout the nerve near the retina, come to be limited to the perimeter of the nerve as it passes through the optic foramen, while newer fibers come to lie nearest the center of the nerve. These newest fibers approach the ventral surface of the brain nearer the optic chiasm. In the chiasm, as in the tract, the oldest fibers lie furthest from the pial surface of the brain, while newer fibers lie nearer the surface. The age-related fiber ordering in the ferret's optic nerve, with the newest fibers initially being furthest from the surface at the optic foramen, differs from age-related orderings seen in nonmammalian vertebrates, where the newest fibers are always nearest the surface. The changing patterns of fiber ordering along the ferret's optic nerve may relate to changes in the underlying glial structure of the developing nerve.     

Changes in morphology and behaviour of retinal growth cones before and after crossing the midline of the mouse chiasm – a confocal microscopy study

The growth of retinal axons was investigated in different regions of the optic chiasm in C57 pigmented mouse embryos aged embryonic day 13 (E13) to E15. Individual retinal axons and their growth cones were labelled anterogradely by DiI and imaged using a confocal imaging system. In aldehyde-fixed embryos, retinal growth cones display a simple form in the optic nerve and become more complex in morphology in the chiasm. The complex form is particularly prominent in those axons that turn to the ipsilateral tract in the premidline region of chiasm. Moreover, complex growth cones are also commonly found in axons in the postmidline chiasm, which are markedly different in morphology from those axons in the premidline region, suggesting that the postmidline chiasm contains a novel environment for the pathfinding of retinal axons. In another experiment, the dynamic growth of retinal axons is studied in a brain slice preparation of the living retinofugal pathway. Retinal axons show an intermittent growth across the premidline and postmidline chiasm. Extensive remodelling of growth cone form followed by a shift in growth direction is commonly seen during the pause periods, indicating that signals that guide axon growth across the chiasm are not restricted to the midline, but are laid down throughout the chiasm. Moreover, dramatic changes in axon trajectory are noted first at the premidline chiasm where the uncrossed axons segregate from the crossed axons, and second at the postmidline chiasm where specific sorting of retinal axons according to their position in the dorsal ventral retinal axis and their ages are known to take place. These results show that there are two distinct environments, separated by the midline in the chiasm, where axons show different responses to local guidance cues and develop the distinct fibre orders.     

Prechiasmatic Reordering of Fibre Diameter Classes in the Retinofugal Pathway of Ferrets

This study examines the distribution of fibre diameter classes at various sites along the retinofugal pathway of adult ferrets. Light microscopic observations were made on semi-thin sections, and regional fibre diameter spectra were constructed from diameter measurements taken from electron micrographs of thin sections of the intraorbital optic nerve (2.5 mm from the optic disc), the intracranial optic nerve (1 mm rostral to the fusion of the nerves), and the optic tract (just caudal to the optic chiasm).
Whereas diameter types are relatively evenly distributed behind the eye in the postoptic nerve, they begin to segregate along its prechiasmatic course. Within this prechiasmatic region, coarse and fine calibre fibres are confined increasingly to more ventral locations in the nerve, leaving a dorsal band populated predominantly by intermediate calibre fibres. In conjunction with this redistribution of axon size classes, the fascicular arrangement of axons which is present distally, changes to a non-fascicular organization. The prechiasmatic organization of fibre types approximates that found in the optic tract where the coarse and fine calibre fibres lie further ventrally towards the pial surface.
The prechiasmatic region can be viewed as a region of transition where the order of fibres in the nerve (retinotopic) starts to change to that present in the optic tract (chronotopic), resulting in the first-born beta cell axons becoming segregated dorsally, and rostral to the coarse and fine calibre classes which segregate at further caudal locations. Further, since the sorting of fibres according to diameter appears before the fibres reach the optic chiasm, the segregation of diameter classes is not dependent on the chiasmatic sorting of fibres according to their crossed or uncrossed course.     

Changes in axon arrangement in the retinofugal [correction of retinofungal] pathway of mouse embryos: confocal microscopy study using single- and double-dye label

The changes in quadrant-specific fiber order in the retinofugal pathway of the C57-pigmented mouse aged embryonic day 15 were investigated by using single- (1,1'-dioctadecyl-3,3,3',3'-tetramethyl-indocarbocyanine perchlorate; DiI) and double- (N-4-4-didecylaminostyryl-N-methylpyridinium iodide; 4Di-10ASP in addition to DiI) labeling techniques. At this earliest stage of development, before any fibers arrive at their targets, retinal axons display a distinct quadrant-specific order at the optic stalk close to the eye. This order gradually disappears along the stalk and is virtually lost at the chiasm, as shown in single-label preparations. The double-label preparations, in which the population peaks of fibers from two retinal quadrants are shown simultaneously in an image, show a fiber arrangement at the chiasm that is different from the pattern seen in the single-label preparations. A distinct and consistent preferential distribution of fibers from different retinal quadrants is shown in the chiasm. Before the midline, the central part of the cross section of the chiasm is dominated by dorsal fibers, whereas the rostral and caudal parts of the chiasm are dominated by ventral nasal and ventral temporal fibers, respectively. Moreover, the double-label preparations demonstrate a major reshuffling of fiber position after the fibers cross the midline. Fibers from ventral retina are shifted gradually to a rostral position at the threshold of the optic tract, whereas fibers from dorsal retina are shifted caudally. These changes in fiber position indicate a postmidline location in the chiasm, where fibers are re-sorted in accordance with their origins in the dorsal ventral axis of the retina, and suggest a change in axon response to guidance signals when the fibers cross the midline of the chiasm. These changes in fiber order may also be related to the re-sorting of fibers according to their ages at the postmidline chiasm.     

Axonal regeneration is associated with glial migration: Comparison between the injured optic nerves of fish and rats

The central nervous systems of mammals and fish differ significantly in their ability to regenerate. Central nervous system axons in the fish readily regenerate after injury, while in mammals they begin to elongate but their growth is aborted a the site of injury, an area previously shown to contain no glial cells. In the present study we compared the ability of glial cells to migrate and thus to repopulate the injured area in fish and rats, and used light and electron microscopy in an attempt to correlate such migration with the ability of axons to traverse this area. One week after the optic nerve was crushed, both axonal and glial responses to injury were similar in fish and rat. In both species glial cells were absent in the injured area (indicated by the disappearance of glial fibrillary acidic protein and vimentin immunoreactive cells from the site of injury in rat and fish, respectively), while at the same time axonal growth, indicated by expression of the growth-associated protein GAP-43, was restricted to the proximal part of the nerve. In fish, 2 weeks after the crush, GAP-43 staining (i.e., growing axons) was seen at the site of injury, in association with migrating vimentin-positive glial cells. One week later the site of injury in the fish optic nerve was repopulated by vimentin-positive glial cells, and GAP-43-positive axons had already traversed the site of injury and reached the distal part of the nerve. In contrast, the site of injury in the rat remained devoid of glial fibrillary acidic protein immunoreactive cells, and the expression of GAP-43 by growing axons was still restricted to the proximal part of the nerve. Double-labeling experiments and transmission electron microscopy performed 2 weeks after crush injury of the fish optic nerve revealed that the frontier of axonal growth (i.e., the leading growth cones) appeared to be 200–300 μm ahead of the nearest vimentin-positive glial cells. The leading growth cones were associated with other cells, presumably glial precursors, that seemed to have a high migratory potential. We suggest that the ability of fish glial precursor cells to migrate into the injured area may contribute to the potential of growing axons to traverse this area. The failure of glia and glial precursor cells to migrate into the injured area in the rat may partially account for the failure of rat axons to enter and traverse the injured area.     

Use of brainstem flat-mounts for visualizing DiI-filled axons in the developing rodent visual system

The lipophilic carbocyanine fluorescent label DiI was injected in one eye of aldehyde-fixed embryonic or postnatal hamsters and the brains were examined using flat-mounts of the chiasm region, of the lateral surface of the brainstem, or of the midbrain tectum. Single axons could be discerned within the optic nerves and along the optic tract. Many fibers were tipped by growth cones, ending at various levels of the brainstem. Fine details of retinofugal axon morphology, including varicosities, branch-points and filopodial extensions on growth cones were visible in the flat-mounts. Such preparations allow a high-resolution view of labeled axons which course near the surface of the brain. It is possible, with this method, to simultaneously examine the morphogenesis of multiple collateral arbors on single fibers which project to more than one terminal zone.     

Glial filament protein expression in astroglia in the mouse visual pathway

We have studied the onset of expression of glial filament protein (GFP) in astrocytes along a single axon trajectory, the mouse retinal axon pathway, and the relationship of GFP expression to maturation of astroglial morphology. In fetal optic nerve from embryonic day (E) 12 to E16, primitive glia (neuroepithelial cells) lack GFP, but express vimentin and contain intermediate filaments. GFP is expressed at E17 in two gradients: cells in the optic nerve become GFP-positive first in the borders of the nerve, then in the central nerve by postnatal day (P)0. The second gradient is a distoproximal one, with GFP appearing in the optic nerve at E17, in the optic chiasm by PO, and in the optic tract by P3. The expression of GFP in the optic nerve marks the transformation of radial neuroepithelial cells to multipolar astroglia, accomplished by outgrowth of filament-rich glial processes tipped by a growth cone. Several days after the onset of GFP expression in each portion of the pathway astrocytes exhibit a transient increase in staining, and resemble reactive astrocytes after injury. During this period, filaments are arranged in densely packed bundles, and appear coalesced at points. Thus, primitive glial cells in optic nerve express vimentin. In the retinofugal pathway, GFP is expressed in a distinct spatiotemporal sequence from optic nerve to optic tract. Finally, in contrast to neurons, the extension of astroglial processes is accompanied by the increased expression and assembly of intermediate filaments.     

Position of growth cones within the retinal nerve fibre layer of fetal ferrets.

Optic axons are added to the retinal nerve fibre layer of fish along its vitreal border in a chronotopic manner. Likewise, the optic tract of all vertebrate species acquires axons preferentially along the superficial surface of the pathway. We have examined the developing retina of fetal ferrets (Mustela putorius furo) aged between embryonic day 27 (E27) and E34 to see whether a similar segregation of growth cones is apparent within the mammalian retinal nerve fibre layer. The distributions of growth cone, "wrist" (thick trailing portion of the growth cone), axonal, and glial profiles were determined from electron micrographs, and expressed as a percentage of neural profiles for several retinal locations. The retinal nerve fibre layer of fetal ferrets contains radially elongated bundles of fibres composed of axonal, wrist, and growth cone profiles. Glial processes of varying density divide the adjacent bundles, occasionally subdividing them in the plane of the retina, and give rise to endfeet lining the basal lamina and separating the optic axons from the latter. Growth cones within the developing fibre layer represented about 2.4% of profiles at E28, while at later developmental stages (E34), this value fell to about 0.6%. During this period of axonal outgrowth, growth cones were not preferentially segregated toward the vitreal basal lamina or the glial endfeet within the nerve fibre layer. Rather, they were found scattered throughout the axon bundles of the fibre layer. While there were differences in the proportion of immature profiles found within the vitreal half compared to the scleral half of the fibre layer, such that more growth cones and wrists were found vitreally, there was no clear accumulation of them in association with features of the vitreal margin. The present results show that young and old optic axons course together throughout the depth of the nerve fibre layer. A chronotopic mode of pathway genesis such as seen in the optic fibre layer of fish or in the optic tract of mammals is not present in the nerve fibre layer of ferrets. Differences in growth cone behaviour in the optic fibre layer and tract indicate that the mechanisms governing pathway formation differ along its course.