Synaptic organization of anomalous retinal projections to the somatosensory and auditory thalamus: target-controlled morphogenesis of axon terminals and synaptic glomeruli

Nuclei-specific differences in nerve terminal distribution, morphology, and development in mouse visual thalamus

Background

Mouse visual thalamus has emerged as a powerful model for understanding the mechanisms underlying neural circuit formation and function. Three distinct nuclei within mouse thalamus receive retinal input, the dorsal lateral geniculate nucleus (dLGN), the ventral lateral geniculate nucleus (vLGN), and the intergeniculate nucleus (IGL). However, in each of these nuclei, retinal inputs are vastly outnumbered by nonretinal inputs that arise from cortical and subcortical sources. Although retinal and nonretinal terminals associated within dLGN circuitry have been well characterized, we know little about nerve terminal organization, distribution and development in other nuclei of mouse visual thalamus.

Results

Immunolabeling specific subsets of synapses with antibodies against vesicle-associated neurotransmitter transporters or neurotransmitter synthesizing enzymes revealed significant differences in the composition, distribution and morphology of nonretinal terminals in dLGN, vLGN and IGL. For example, inhibitory terminals are more densely packed in vLGN, and cortical terminals are more densely distributed in dLGN. Overall, synaptic terminal density appears least dense in IGL. Similar nuclei-specific differences were observed for retinal terminals using immunolabeling, genetic labeling, axonal tracing and serial block face scanning electron microscopy: retinal terminals are smaller, less morphologically complex, and more densely distributed in vLGN than in dLGN. Since glutamatergic terminal size often correlates with synaptic function, we used in vitro whole cell recordings and optic tract stimulation in acutely prepared thalamic slices to reveal that excitatory postsynaptic currents (EPSCs) are considerably smaller in vLGN and show distinct responses following paired stimuli. Finally, anterograde labeling of retinal terminals throughout early postnatal development revealed that anatomical differences in retinal nerve terminal structure are not observable as synapses initially formed, but rather developed as retinogeniculate circuits mature.

Conclusions

Taken together, these results reveal nuclei-specific differences in nerve terminal composition, distribution, and morphology in mouse visual thalamus. These results raise intriguing questions about the different functions of these nuclei in processing light-derived information, as well as differences in the mechanisms that underlie their unique, nuclei-specific development.

Cortical GABAergic interneurons in cross-modal plasticity following early blindness

Early loss of a given sensory input in mammals causes anatomical and functional modifications in the brain via a process called cross-modal plasticity. In the past four decades, several animal models have illuminated our understanding of the biological substrates involved in cross-modal plasticity. Progressively, studies are now starting to emphasise on cell-specific mechanisms that may be responsible for this intermodal sensory plasticity. Inhibitory interneurons expressing γ-aminobutyric acid (GABA) play an important role in maintaining the appropriate dynamic range of cortical excitation, in critical periods of developmental plasticity, in receptive field refinement, and in treatment of sensory information reaching the cerebral cortex. The diverse interneuron population is very sensitive to sensory experience during development. GABAergic neurons are therefore well suited to act as a gate for mediating cross-modal plasticity. This paper attempts to highlight the links between early sensory deprivation, cortical GABAergic interneuron alterations, and cross-modal plasticity, discuss its implications, and further provide insights for future research in the field.

Alterations of the visual pathways in congenital blindness

We used whole brain MRI voxel-based morphometry (VBM) to study the anatomical organization of the visual system in congenitally blind (CB) adults. Eleven CB without a history of visual perception were compared with 21 age- and sex-matched normal-sighted controls (NS). CB showed significant atrophy of the geniculo-striate system, encompassing the optic nerves, the optic chiasm, the optic radiations and the primary visual cortex (BA17). The volume decrease in BA17 reached 25% in both hemispheres. The pulvinar and its projections to the associative visual areas were also dramatically altered, BA18/19 and the middle temporal cortex (MT) showing volume reductions of up to 20%. Additional significant white matter alterations were observed in the inferior longitudinal tract and in the posterior part of the corpus callosum, which links the visual areas of both hemispheres. Our data indicate that the afferent projections to the visual cortex in CB are largely atrophied. Despite the massive volume reductions in the occipital lobes, there is compelling evidence from the literature (reviewed in Noppeney 2007; Ptito and Kupers 2005) that blind subjects activate their visual cortex when performing tasks that involve somatosensory or auditory inputs, suggesting a reorganization of the neural pathways that transmit sensory information to the visual cortex.

Dual axonal terminations from the retrosplenial and visual association cortices in the laterodorsal thalamic nucleus of the rat

Light and electron microscopic tracing studies were conducted to assess the synaptic organization in the laterodorsal thalamic nucleus (LD) of the rat and the laminar origins of corticothalamic terminals from the retrosplenial and visual association cortices to LD. A survey of the general ultrastructure of LD revealed at least three types of presynaptic terminals identified on the basis of size, synaptic vesicle morphology, and synaptic membrane specializations: (1) small axon terminals with round synaptic vesicles (SR), which accounted for the majority of terminal profiles and made asymmetric synaptic contacts predominantly with small dendritic shafts and spines; (2) large axon terminals with round synaptic vesicles (LR), which formed asymmetric synaptic contacts mainly with large dendritic shafts; and (3) small to medium-size axon terminals with pleomorphic synaptic vesicles (SMP), which symmetrically synapsed with a wide range of postsynaptic structures from cell bodies to small dendrites. Synaptic glomeruli were identified, whereas no presynaptic dendrites were found. To characterize and identify corticothalamic terminals arising from the retrosplenial and visual association cortices that project to LD, wheat germ agglutinin conjugated to horseradish peroxidase (WGA-HRP) was injected into these cortices. Axons anterogradely labeled with WGA-HRP ended in both SR and LR terminals. On the other hand, dextran-tetramethylrhodamine injected into LD as a retrograde fluorescent tracer labeled large pyramidal cells of layer V as well as small round or multiform cells of layer VI in the retrosplenial and visual association cortices. These findings provide the possibility that corticothalamic terminations from cortical neurons in layer V end as LR terminals, while those from neurons in layer VI end as SR boutons.

Cerebellar grafting in the oculomotor system as a model to study target influence on adult neurons

In the last decades, there have been many efforts directed to gain a better understanding on adult neuron-target cell relationships. Embryonic grafts have been used for the study of neural circuit rewiring. Thus, using several donor neuronal tissues, such as cerebellum or striatum, developing grafted cells have been shown to have the capability of substituting neural cell populations and establishing reciprocal connections with the host. In addition, different lesion paradigms have also led to a better understanding of target dependence in neuronal cells. Thus, for example, axotomy induces profound morphofunctional changes in adult neurons, including the loss of synaptic inputs and discharge alterations. These alterations are probably due to trophic factor loss in response to target disconnection. In this review, we summarize the different strategies performed to disconnect neurons from their targets, and the effects of target substitution, performed by tissue grafting, upon neural properties. Using the oculomotor system-and more precisely the abducens internuclear neurons-as a model, we describe herein the effects of disconnecting a population of central neurons from its natural target (i.e., the medial rectus motoneurons at the mesencephalic oculomotor nucleus). We also analyze target-derived influences in the structure and physiology of these neurons by using cerebellar embryonic grafts as a new target for the axotomized abducens internuclear neurons.

Fine structure of long-term changes in the cochlear nucleus after acoustic overstimulation: Chronic degeneration and new growth of synaptic endings

The companion study showed that acoustic overstimulation of adult chinchillas, with a noise level sufficient to damage the cochlea, led to cytological changes and degeneration of synaptic endings in the cochlear nucleus within 1-16 weeks. In the present study, the same stimulus was used to study the long-term effects on the fine structure of synaptic endings in the cochlear nucleus. For periods of 6 and 8 months after a single exposure to a damaging noise level, there ensued a chronic, continuing process of neurodegeneration involving excitatory and inhibitory synaptic endings. Electron microscopic observations demonstrated freshly occurring degeneration even as late as 8 months. Degeneration was widespread in the neuropil and included the synapses on the globular bushy cell, which forms part of the main ascending auditory pathway. Neurodegeneration was accompanied by newly formed synaptic endings, which repopulated some of the sites vacated previously by axosomatic endings on globular bushy cells. Many of these synaptic endings must arise from central interneurons. The findings suggest that overstimulation can induce a self-sustaining condition of progressive neurodegeneration accompanied by a new growth of synaptic endings. Noise-induced hearing loss thus may progress as a neurodegenerative disease with the capacity for synaptic reorganization within the cochlear nucleus.

Intrinsic determinants of synaptic phenotype: an experimental study of abducens internuclear neurons connecting with anomalous targets

The present experiments investigate the role of postsynaptic neurons in the morphological differentiation of presynaptic terminals that are formed de novo in the adult CNS. Abducens internuclear neurons in the adult cat were chosen as the experimental model. These neurons project onto the contralateral medial rectus motoneurons of the oculomotor nucleus. Abducens internuclear axon terminals were identified by their anterograde labeling with biocytin and analyzed at the electron microscopic level. To promote the formation of new synapses, two different experimental approaches were used. First, after the selective ablation of medial rectus motoneurons with ricin, abducens internuclear neurons reinnervated the neighboring oculomotor internuclear neurons. Second, after axotomy followed by embryonic cerebellar grafting, abducens internuclear axons invaded the implanted tissue and established synaptic connections in both the molecular and granule cell layer. Boutons contacting the oculomotor internuclear neurons developed ultrastructural characteristics that resembled the control synapses on medial rectus motoneurons. In the grafted cerebellar tissue, abducens internuclear axons and terminals did not resemble climbing or mossy fibers but showed similarities with control boutons. However, labeled boutons analyzed in the granule cell layer established a higher number of synaptic contacts than controls. This could reflect a trend towards the mossy fiber phenotype, although labeled boutons significantly differed in every measured parameter with the mossy fiber rosettes found in the graft. We conclude that at least for the abducens internuclear neurons, the ultrastructural differentiation of axon terminals reinnervating novel targets in the adult brain seems to be mainly under intrinsic control, with little influence by postsynaptic cells.

When the auditory cortex turns visual

We studied visually guided behavior and the visual response properties of single auditory cortex (A1) neurons in neonatally operated hamsters with surgically induced, permanent, ectopic retinal projections to auditory thalamic nuclei and to visual thalamic nuclei which normally receive little direct retinal input. The surgically induced retino-thalamo-cortical pathways can mediate visual guided behaviors whose normal substrate, the pathway from the retina to the primary visual cortex via the primary thalamic visual nucleus, is missing. The visually evoked response properties of A1 neurons resemble in many respects those of neurons in V1 of normal hamsters: many A1 neurons have well-defined visual receptive fields and preferences for orientation or direction of movement. In addition, some visually responsive cells in A1 are bimodal--they also respond to auditory stimuli. The visually responsive neurons in A1 probably account for the capacity of the auditory cortex to mediate visual behavior in 'rewired hamsters'.

Developmental changes in BDNF protein levels in the hamster retina and superior colliculus

Quantitative studies of ontogenetic changes in the levels of brain-derived neurotrophic factor (BDNF) mRNA and its effector, BDNF protein, are not available for the retinal projection system. We used an electrochemiluminescence immunoassay to measure developmental changes in the tissue concentration of BDNF within the hamster retina and superior colliculus (SC). In the SC, we first detected BDNF (about 9 pg/mg tissue) on embryonic day 14 (E14). BDNF protein concentration in the SC rises about fourfold between (E14) and postnatal day 4 (P4), remains at a plateau through P15, then declines by about one-third to attain its adult level by P18. By contrast, BDNF protein concentration in the retina remains low (about 1 pg/mg tissue) through P12, then increases 4.5-fold to attain its adult level on P18. The developmental changes in retinal and collicular BDNF protein concentrations are temporally correlated with multiple events in the structural and functional maturation of the hamster retinal projection system. Our data suggest roles for BDNF in the cellular mechanisms underlying some of these events and are crucial to the design of experiments to examine those roles.

Surgically created neural pathways mediate visual pattern discrimination

Combined lesions of retinal targets and ascending auditory pathways can induce, in developing animals, permanent retinal projections to auditory thalamic nuclei and to visual thalamic nuclei that normally receive little direct retinal input. Neurons in the auditory cortex of such animals have visual response properties that resemble those of neurons in the primary visual cortex of normal animals. Therefore, we investigated the behavioral function of the surgically induced retino-thalamo-cortical pathways. We showed that both surgically induced pathways can mediate visually guided behaviors whose normal substrate, the pathway from the retina to the primary visual cortex via the primary thalamic visual nucleus, is missing.

Anatomic, intrinsic, and synaptic properties of dorsal and ventral division neurons in rat medial geniculate body

Anatomic, intrinsic, and synaptic properties of dorsal and ventral division neurons in rat medial geniculate body. Presently little is known about what basic synaptic and cellular mechanisms are employed by thalamocortical neurons in the two main divisions of the auditory thalamus to elicit their distinct responses to sound. Using intracellular recording and labeling methods, we characterized anatomic features, membrane properties, and synaptic inputs of thalamocortical neurons in the dorsal (MGD) and ventral (MGV) divisions in brain slices of rat medial geniculate body. Quantitative analysis of dendritic morphology demonstrated that tufted neurons in both divisions had shorter dendrites, smaller dendritic tree areas, more profuse branching, and a greater dendritic polarization compared with stellate neurons, which were only found in MGD. Tufted neuron dendritic polarization was not as strong or consistent as earlier Golgi studies suggested. MGV and MGD cells had similar intrinsic properties except for an increased prevalence of a depolarizing sag potential in MGV neurons. The sag was the only intrinsic property correlated with cell morphology, seen only in tufted neurons in either division. Many MGV and MGD neurons received excitatory and inhibitory inferior colliculus (IC) inputs (designated IN/EX or EX/IN depending on excitation/inhibition sequence). However, a significant number only received excitatory inputs (EX/O) and a few only inhibitory (IN/O). Both MGV and MGD cells displayed similar proportions of response combinations, but suprathreshold EX/O responses only were observed in tufted neurons. Excitatory and inhibitory postsynaptic potentials (EPSPs and IPSPs) had multiple distinguishable amplitude levels implying convergence. Excitatory inputs activated alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) and N-methyl-D-aspartate (NMDA) receptors the relative contributions of which were variable. For IN/EX cells with suprathreshold inputs, first-spike timing was independent of membrane potential unlike that of EX/O cells. Stimulation of corticothalamic (CT) and thalamic reticular nucleus (TRN) axons evoked a GABAA IPSP, EPSP, GABAB IPSP sequence in most neurons with both morphologies in both divisions. TRN IPSPs and CT EPSPs were graded in amplitude, again suggesting convergence. CT inputs activated AMPA and NMDA receptors. The NMDA component of both IC and CT inputs had an unusual voltage dependence with a detectable DL-2-amino-5-phosphonovaleric acid-sensitive component even below -70 mV. First-spike latencies of CT evoked action potentials were sensitive to membrane potential regardless of whether the TRN IPSP was present. Overall, our in vitro data indicate that reported regional differences in the in vivo responses of MGV and MGD cells to auditory stimuli are not well correlated with major differences in intrinsic membrane features or synaptic responses between cell types.

Brainstem inputs to the ferret medial geniculate nucleus and the effect of early deafferentation on novel retinal projections to the auditory thalamus

Following specific neonatal brain lesions in rodents and ferrets, retinal axons have been induced to innervate the medial geniculate nucleus (MGN). Previous studies have suggested that reduction of normal retinal targets along with deafferentation of the MGN are two concurrent factors required for the induction of novel retino-MGN projections. We have examined, in ferrets, the relative influence of these two factors on the extent of the novel retinal projection. We first characterized the inputs to the normal MGN, and the most effective combination of neonatal lesions to deafferent this nucleus, by injecting retrograde tracers into the MGN of normal and neonatally operated adult ferrets, respectively. In a second group of experiments, newborn ferrets received different combinations of lesions of normal retinal targets and MGN afferents. The resulting extent of retino-MGN projections was estimated for each case at adulthood, by using intraocular injections of anterograde tracers. We found that the extent of retino-MGN projections correlates well with the extent of MGN deafferentation, but not with extent of removal of normal retinal targets. Indeed, the presence of at least some normal retinal targets seems necessary for the formation of retino-MGN connections. The diameters of retino-MGN axons suggest that more than one type of retinal ganglion cells innervate the MGN under a lesion paradigm that spares the visual cortex and lateral geniculate nucleus. We also found that, after extensive deafferentation of MGN, other axonal systems in addition to retinal axons project ectopically to the MGN. These data are consistent with the idea that ectopic retino-MGN projections develop by sprouting of axon collaterals in response to signals arising from the deafferented nucleus, and that these axons compete with other sets of axons for terminal space in the MGN.

An ultrastructural study of the neural circuit between the prefrontal cortex and the mediodorsal nucleus of the thalamus

Synaptic connectivity between the prefrontal cortex (PFC) and the mediodorsal thalamic nucleus (MD) of the rat has been investigated with the electron microscope after labeling both the pre- and postsynaptic elements. Prefrontal corticothalamic fibers end exclusively as small axon terminals with round synaptic vesicles (SR boutons), which make asymmetrical synaptic contacts with distal dendritic segments of MD neurons. Thalamocortical terminals from MD in PFC are also of the SR type and form asymmetrical synaptic contacts predominantly with dendritic spines arising from the apical or basal dendrites of pyramidal cells whose somata reside in layers III, V and VI. At least some pyramidal cells in layer III that receive MD afferents are callosal cells, whereas deep layer pyramidal cells projecting to MD receive directly some of the thalamocortical terminations from MD, suggesting that the recurrent loop to MD is monosynaptically mediated. Thus, taken together with recent evidence that both the PFC-MD and MD-PFC pathways are glutamatergic and excitatory, the cortical excitation exerted by afferent fibers from MD is transferred, not only back to MD itself through deep pyramidal cells, but also the contralateral prefrontal cortex via pyramidal cells in layer III of the ipsilateral prefrontal cortex. Concerning modulatory and inhibitory inputs, fibers to MD from the ventral pallidum and substantia nigra pars reticulata have been shown to be inhibitory and GABAergic. In addition, fibers from the ventral tegmental area preferentially make symmetrical membrane thickenings (i.e. inhibitory synapses) on deep pyramidal cells in PFC that receive synaptic endings from MD. From these morphological grounds, therefore, cells in the ventral pallidum, the substantia nigra pars reticulata and the ventral tegmental area may mediate, to some extent, an inhibitory effect on the reverberatory excitation between PFC and MD.

Target- as well as source-derived factors direct the morphogenesis of anomalous retino-thalamic projections

Neonatal tectal lesions in hamsters result in the elimination of a major central target of retinal axons, massively denervate the lateral posterior nucleus of the thalamus (LP), and lead to a marked increase of the retino-LP projection. In such animals, retino-LP axons show all of the normally-occurring terminal types. In addition, large clusters of varicosities, whose tubular configuration resembles the major type of tecto-LP terminals observed in normal animals, are also noted if the tectal lesion is made on the day after birth (P1). If, however, the neonatal lesion occurs on P5 rather than on P1, terminals resembling normal tecto-LP endings are rarely observed; rather, the distribution and morphology of retino-LP terminals bear a greater resemblance to those seen in normal hamsters, but the size and complexity of the terminals, particularly those that form string-like arrangements, is significantly increased. Our findings suggest that the altered morphology of some abnormally induced retino-LP terminals may be orchestrated by target-associated signals. However, there are age-related limitations on the degree to which afferent systems can vary their terminal morphology; these restrictions may derive from the target, or may be a function of intrinsic changes within the cells of origin of the afferent fibers.

Dendritic development of retinal ganglion cells after prenatal intracranial infusion of tetrodotoxin

The dendritic form of a cell may be established by many factors both intrinsic and environmental. Blockade of action potentials along the course of axons and in their postsynaptic targets dramatically alters the development of axonal morphology. The extent to which blockade of target cell activity retrogradely alters the dendritic morphology of the presynaptic cells is unknown. To determine whether the establishment of dendritic form by developing retinal ganglion cells depends on activity within their targets, the sodium channel blocker, tetrodotoxin (TTX), was administered via minipumps to the diencephalon of cat fetuses from embryonic day 43 (E43) to E57. At E57 retinae were removed and living retinal ganglion cells injected in vitro with Lucifer yellow to reveal their dendritic morphology. In the TTX-treated animals both alpha and beta types of retinal ganglion cells were present, as were putative gamma cells. Overall, the dendrites of retinal ganglion cells in TTX-treated animals appeared qualitatively and quantitatively similar to those of untreated animals. The only significant change in the TTX-treated cases was a small increase in the number of dendritic spines on the non-beta cells. These results indicate that the acquisition of basic dendritic form of developing ganglion cells is not influenced by the action potential activity within their targets, and that it is also independent of the terminal branching patterns of their axons.

Experimentally induced retinal projections to the ferret auditory thalamus: development of clustered eye-specific patterns in a novel target

We have examined the relative role of afferents and targets in pattern formation using a novel preparation, in which retinal projections in ferrets are induced to innervate the medial geniculate nucleus (MGN). We find that retinal projections to the MGN are arranged in scattered clusters. Clusters arising from the ipsilateral eye are frequently adjacent to, but spatially segregated from, clusters arising from the contralateral eye. Both clustering and eye-specific segregation in the MGN arise as a refinement of initially diffuse and overlapped projections. The shape, size, and orientation of retinal terminal clusters in the MGN closely match those of relay cell dendrites arrayed within fibrodendritic laminae in the MGN. We conclude that specific aspects of a projection system are regulated by afferents and others by targets. Clustering of retinal projections within the MGN and eye-specific segregation involve progressive remodeling of retinal axon arbors, over a time period that closely parallels pattern formation by retinal afferents within their normal target, the lateral geniculate nucleus (LGN). Thus, afferent-driven mechanisms are implicated in these events. However, the termination zones are aligned within the normal cellular organization of the MGN, which does not differentiate into eye-specific cell layers similar to the LGN. Thus, target-driven mechanisms are implicated in lamina formation and cellular differentiation.

Patterning and specification of the cerebral cortex

Regionalization of the cerebral cortex occurs during development by the formation of anatomically and functionally discrete areas of the brain. Descriptive evidence based on expression of molecules and structural features suggests that an early parcelation of the cerebral wall may occur during fetal development. Experimental strategies using tissue transplants and cell culture models have explored the nature of the timing of areal specification. New signaling systems displaying the sensitivity of precursor cells to environmental cues that define the fate of neurons destined for specific areas of the cortex have been discovered. Studies in the field now suggest mechanisms of regulating cell phenotype in the cortex that are common to all parts of the neuraxis.

Functional properties of regenerated optic axons terminating in the primary olfactory cortex

When the optic nerve of Rana pipiens is cut and deflected into the telencephalon, the regenerating fibers terminate selectively in the superficial neuropil of the primary olfactory cortex. These redirected fibers and their terminals on the dendrites of the cortical cells appear normal by LM and EM criteria. Electrical recording, done 2-16 months after surgery, shows visually evoked activity in the superficial neuropil (Layer I) of the olfactory cortex, and visually excited responses in the deep cortical cell layer (Layer II). In the normal frog, the electrical activity seen in the neuropil of the olfactory cortex consists of small transients about 2-3 x the noise level of the electrode contact. These occur spontaneously and are also excited by puffs of air to the nose. There is no such excitation by visual stimuli. Larger initially negative spikes cell above noise level are recorded in the cell layer next to the ependymal surface, and these are also spontaneous, or excited by puffs of air to the nose, but not by visual stimuli. In the operated frog, the small transients in the neuropil appear and are excited by the puffs of air and by visual stimuli. Similarly the responses in the cell layer are excited by both sorts of stimuli. But new types of electrical signals appear in the neuropil; they are driven only by visual stimuli presented to the affected eye. These are very large transients of the kind found in the tectal neuropil and have the two characteristic shapes which were classified as B and C types in the tectum. Such large transients are never seen in the neuropil of the olfactory cortex in normal frogs. The receptive fields of the small visually driven transients in the neuropil are not easy to make out because the signal levels are so close to the noise level that different units cannot be reliably distinguished from each other. But the receptive fields of the much larger B and C type unit responses are as easy to classify and plot as they are in tectum, even though on the average they are only about 2/3 as large as in tectum. The single-unit receptive fields belong to one or another of the several types of retinal ganglion cell classes distinguished in optic-nerve recordings. Four of the major classes normally project to the tectum and a fifth projects to the lateral geniculate complex. But all five are present in the ectopic projection to the olfactory cortex.(ABSTRACT TRUNCATED AT 400 WORDS)

Retinal ganglion cells in normal hamsters and hamsters with novel retinal projections. I. Number, distribution, and size

We examined the number, spatial distribution, and size of ganglion cells in the retinae of normal Syrian hamsters and hamsters with retinal projections to the auditory and somatosensory nuclei of the thalamus, induced by neonatal surgery. As revealed by retrograde filling with horseradish peroxidase, there are about 64,600 contralaterally projecting retinal ganglion cells (RGCs) and 1,700 ipsilaterally projecting RGCs in the retinae of normal adult hamsters. Contralaterally projecting RGCs are distributed throughout the retina and have two local density peaks located within a central streak of high RGC density that is oriented approximately along the nasal-temporal axis. RGC density falls above and below the central streak, with a steeper gradient towards the upper retina. Ipsilaterally projecting RGCs are diffusely distributed within a crescent at the inferotemporal retinal periphery and are most dense at the internal border of the crescent. The soma diameter of contralaterally projecting RGCs ranges from 6 to 25 microns; the diameter distribution is unimodal, with a peak in the 10-13 microns range and is skewed toward smaller values, with an elongated tail towards higher values. Contralaterally projecting RGCs tend to be smaller in regions of higher density. Ipsilaterally projecting RGCs tend to be larger than contralaterally projecting RGCs both globally and within the temporal crescent, and their size distributions tend to be less regular and less well related to local density. The retinae of neonatally operated hamsters with novel retinal projections to the auditory. and somatosensory systems contain about one-fourth the normal number of contralaterally projecting RGCs, whose relative density distribution is approximately normal despite the drastic reduction of absolute RGC density. The range and distribution of RGC soma diameters are similar in normal and neonatally operated hamsters, and, in operated as in normal hamsters, contralaterally projecting RGC somata tend to be smaller in regions of higher density. Our results in normal hamsters suggest a role for intraretinal mechanisms in the determination of RGC size. Our findings in neonatally operated hamsters suggest that, despite the reduced number of RGCs in these animals, the same types of RGCs are found in the retinae of normal and neonatally operated hamsters.

Morphology of retinal axon arbors induced to arborize in a novel target, the medial geniculate nucleus. II. Comparison with axons from the inferior colliculus

Specific neonatal lesions in ferrets can induce retinal axons to project into the medial geniculate nucleus (MGN). In the accompanying paper (Pallas et al., this issue), we described the morphology of these retinal ganglion cell axons. Those results and others (Roe et al. [1993] J. Comp. Neurol. 334:263) suggest that these axons belong to the W class of retinal axons. In this paper, the retino-MGN axons are compared with the normal inputs to the MGN from the brachium of the inferior colliculus (BIC). We first sought to determine further the extent to which a novel target might influence retinal axon arbor morphology. The second issue concerns retinal topography. Ferrets with retinal projections to the MGN have a two-dimensional retinotopic map in the MGN and the primary auditory cortex rather than the one-dimensional tonotopic map normally present (Roe et al. [1990] Science 250:818). To investigate whether there might be an anatomical substrate for a two-dimensional retinotopic map in the MGN, we compared the space-filling characteristics of the retino-MGN axons with the IC-MGN axons. Our results show that the branched retino-MGN axons resemble normal retinal W axons much more closely than they resemble the normal inputs to MGN. In addition, most of the axon arbors from the BIC are elongated along the rostrocaudal (isofrequency) axis, whereas the branched retino-MGN axons are more spatially restricted, suggesting an anatomical substrate for a retinotopic map in the MGN of the rewired ferrets.

Development of the retinofugal projections in the embryonic and larval zebrafish (Brachydanio rerio)

Studies of the projection from the vertebrate retina have contributed significantly to current concepts of neural development. The zebrafish has recently become a favored system for the study of development in general and neural development in particular. Although the development of both the optic nerve and the retinotectal projection of the zebrafish has been described, the retinofugal projection in its entirety has not. This paper describes it and also addresses the issue of projectional exuberance: i.e., transient projections to targets that are not innervated in the adult. The retinofugal projection of embryonic and larval zebrafish (32 hours to 7 days post-fertilization) was labeled by intraocular injection of DiI (1,1'-dioctadecyl-3,3,3',3',tetramethylindocarbocyanine perchlorate) and then studied in wholemounts and sections. The first optic axons crossed the chiasm at 32 hours post-fertilization and projected in a straight line to reach the tectum at about 44 hours. At 48 hours, a few optic axons deviated along either the tract of the posterior commissure or the tract of the postoptic commissure. By 72 hours (about the time of hatching) optic axons arborized in ten distinct regions, termed arborization fields. At 6-7 days post-fertilization, the same ten arborization fields (nine contralateral, one bilateral) were evident. Most of the arborization fields were located in the superficial neuropil and were not associated with morphologically identifiable clusters of somata. On the basis of various landmarks, the ten arborization fields are identified as precursors of retinorecipient nuclei previously described in other adult cypriniform fishes. The development was characterized by the nearly complete absence of any transient projections. Thus, the idea that axonal outgrowth is initially exuberant and trimmed back later is not supported by these results.

Maintenance of pharmacologically-immature glutamate receptors by aberrant synapses in the chick cochlear nucleus

Surgical destruction of the otocyst in chick embryos prevents formation of the *** ear, abolishes normal cochlear input to the cochlear nucleus (nucleus magnocellularis, NM) and results in axons from the contralateral NM forming (in addition to their normal bilateral endings in nucleus laminaris, NL) a novel and functional aberrant projection to the deafferented NM. We studied the pharmacology of synaptic transmission at aberrant synapses in an in vitro preparation of the brainstem in chick embryos and hatchlings. Transmission at the aberrant synapses (as with cochlear nerve synapses in NM and NM synapses in NL) is blocked by the quinoxalinedione antagonists CNQX and NBQX, confirming the presence of excitatory amino acid receptors of the non-NMDA subtype. At cochlear nerve synapses in NM, the antagonist potency of NBQX normally decreases rapidly after embryonic day (E)18 (IC50 = 0.69 +/- 0.06 microM, mean +/- S.E.M.), reaching an asymptotic value by E21 (IC50 = 2.7 +/- 0.4 microM) that is maintained at least through posthatching day (P)14 (IC50 = 3.6 +/- 0.3 microM). In the case of the aberrant endings, the potency of NBQX remained (from E21 [IC50 = 0.6 +/- 0.1 microM] through at least P14[IC50 = 0.5 +/- 0.1 microM]) at levels that are statistically indistinguishable from the E18 value for normal cochlear nerve synapses.(ABSTRACT TRUNCATED AT 250 WORDS)

Experimentally induced visual projections to the auditory thalamus in ferrets: evidence for a W cell pathway

We have previously reported that following specific neonatal brain lesions in ferrets, a retinal projection is induced into the auditory thalamus (Sur et al., Science 242:1437, '88). In these "rewired" ferrets, a novel visual pathway is established through auditory thalamus [the medial geniculate nucleus (MGN)] and primary auditory cortex (A1); cells in both MGN and A1 are visually responsive and exhibit properties similar to those of visual cells in the normal visual pathway. In this paper, we use three approaches--physiological, anatomical, and developmental--to examine which of the retinal ganglion cells project to the MGN in these rewired ferrets. We find that: 1) physiological response properties of postsynaptic visual cells in the MGN are W-like; 2) retinal ganglion cells back-filled from the MGN are small and similar to soma sizes of subsets of the normal retinal W cell population; and 3) subpopulations of these small cells can be preferentially rerouted to the MGN in response to different surgical manipulations at birth, consistent with differential W cell projection patterns in normal animals. These data suggest that retinal W cells come to project to the MGN in rewired animals. These findings not only provide a basis on which to interpret functional properties of this novel visual pathway, but also provide important information about the developmental capabilities of specific retinal ganglion cell classes and the regulation of their projections by target structures in the brain during development.

Changes in synaptic density after developmental compression or expansion of retinal input to the superior colliculus

The retinal projection to the superior colliculus can be made abnormally dense by inducing a "compressed" retinal projection into a subnormal tectal volume, or abnormally sparse by monocular enucleation early in development. Any or all of the features of cell number, axonal arbor, dendritic arbor, and synaptic density could potentially be adjusted to compensate for such variations in the convergence of one cell population on another. We have examined the consequences of neonatal partial tectal ablation or monocular enucleation for synaptic length, density, and relative numbers of synapse classes in the superficial gray layer of the hamster superior colliculus. Monocular enucleation resulted in a reduction of synaptic density in the superficial gray layer of the colliculus ipsilateral to the remaining eye. This decrease in density was entirely accounted for by a reduction of the number of synapses with round vesicles, large asymmetric terminal specializations, and pale mitochondria characteristic of retinocollicular terminals (RLP synapses). There was no compensatory increase in any other synaptic class. RLP synapses were larger in monocular enucleates. Partial tectal ablation had no effect on synaptic density, nor on the relative proportions of different synaptic types. Synapses of the RLP class were slightly smaller than normal. These results suggest that synaptic density is normally at a maximum that cannot be altered by increases in potential input. However, density may be reduced by decreasing the number of inputs. Terminal classes do not appear to compete with each other within the collicular volume, suggesting that postsynaptic cells controls both the classes and numbers of their potential inputs.

Altered distribution of synaptic densities at aberrant synapses in the chick cochlear nucleus

The role of axon-target cell interactions in shaping the prevalence and distribution of synaptic densities was studied in an experimentally-induced aberrant functional projection from the chick cochlear nucleus (nuc. magnocellularis, NM) to the contralateral NM. Contact with an abnormal target appears to induce in the aberrant axons a pattern of presynaptic densities resembling that in normal cochlear nerve endings in NM. NM axon terminals induced similar numbers of postsynaptic densities (PSDs) per unit length of membrane apposition in both their normal and abnormal targets but the longer membrane apposition in the highly invaginated aberrant terminal in NM results in a significantly greater amount of postsynaptic density per ending. These auditory neurons thus appear able to adjust a variety of features to permit assembly and maintenance of a novel functional synapse.

Transient pattern of exuberant projections of olfactory axons during development in the rat

The purpose of our study was twofold: (1) to trace the development of the olfactory axons from early embryonic stages until the mature pattern of connectivity and (2) to determine whether a transient penetration of them exists beyond the olfactory glomeruli. Two techniques were employed: DiI applied in the olfactory epithelium after aldehyde fixation, and olfactory marker protein (OMP) immunostaining. At E13 and E14 olfactory axons were observed spreading over the telencephalic vesicle and entering deeply into the prospective olfactory bulb, extending near the ventricular zone. Growth cones were seen at the end of these axons. At E15, the bundles of olfactory axons form a network, in which axons, growth cones and cells were seen. Some of these axons entered the olfactory bulb. Using OMP immunostaining olfactory axons were observed along the external plexiform layer, the mitral cell layer and in the granular layer from E19 to P6. At P9 some OMP immunoreactive axons were observed in the external plexiform layer. No OMP immunostained axons could be observed outside the glomeruli at P10. Our conclusions are that a transient immature pattern of early invasion over the telencephalic vesicle and of the olfactory bulb by olfactory axons occurs in the olfactory system. By the second postnatal week the glomerular layer reaches its mature configuration, and no olfactory fibers are seen outside the glomerular layer.

Neuronal and synaptic composition of the mediodorsal thalamic nucleus in the rat: a light and electron microscopic Golgi study

The distribution and dendritic domain of neurons in each segment of the mediodorsal thalamic nucleus (MD) have been studied in the rat with the Golgi technique. In addition, a combined Golgi method-electron microscopic (Golgi-EM) study was undertaken to determine the distribution of morphologically distinct synapse types along the dendrites of individual identified neurons in MD. All the subdivisions or "segments" of MD (medial, central, lateral) contained both stellate and fusiform cells. The dendritic domain of both types of cells was predominantly restricted to the same segment of MD that contained the cell body of the neuron. Typical stellate neurons were found near the center of each segment, with radiating dendrites that extended to but not across the boundaries of the segment. Fusiform cells were usually located close to the segmental or nuclear boundaries and tended to have dendrites oriented parallel to those borders; again, the dendrites tended not to extend across borders between segments or at the outer edge of MD. In the medial segment of MD many fusiform cells had especially bipolar dendritic configurations, generally with a dorsoventral orientation. Because no small neurons were identified that might correspond to thalamic interneurons, all the impregnated cells in MD are presumed to be thalamocortical projection neurons. These results indicate that cells and their major dendrites are confined to a single segment of MD, with little dendritic overlap across segmental or nuclear borders. The segments of MD may therefore be considered to be relatively independent subnuclei. The distribution of the four types of synapses previously identified in MD (Kuroda and Price, J. Comp. Neurol., 303:513-533, 1991) was determined along several identified dendrites studied with the Golgi-EM method. Primary dendrites were contacted mostly by large axon terminals, including both large, round vesicle (LR) terminals and large, pleomorphic vesicle (LP) terminals, as well as a few small to medium sized terminals with pleomorphic vesicles (SMP). No small terminals with round vesicles (SR terminals) were observed to make synapses with primary dendrites. Secondary and tertiary dendrites received synapses from all types of axon terminals. Higher order dendrites were contracted predominantly by SR boutons, but they also carried some LR and SMP terminals. In addition, SMP boutons were often found to form symmetric contacts with cell somata.

Distribution of the piriform cortical terminals to cells in the central segment of the mediodorsal thalamic nucleus of the rat

A Golgi electron microscopic study was undertaken to investigate the distribution of terminals from the piriform cortex that synapse on identified dendrites of neurons in the central segment of the mediodorsal thalamic nucleus of the rat. The piriform cortical terminals were identified as degenerating terminals following lesions in the cortex. They consisted of two types, i.e., large (LR type) and small (SR type) presynaptic terminals, both of which had round synaptic vesicles and formed asymmetric synaptic contacts. SR boutons terminated preferentially onto distal dendrites and never synapsed on primary dendrites. LR terminals synapsed preferentially on proximal dendrites, but were also found on more distal dendritic segments.

Regeneration of adult rat CNS axons into peripheral nerve autografts: ultrastructural studies of the early stages of axonal sprouting and regenerative axonal growth

If one end of a segment of peripheral nerve is inserted into the brain or spinal cord, neuronal perikarya in the vicinity of the graft tip can be labelled with retrogradely transported tracers applied to the distal end of the graft several weeks later, showing that CNS axons can regenerate into and along such grafts. We have used transmission EM to examine some of the cellular responses that underlie this regenerative phenomenon, particularly its early stages. Segments of autologous peroneal or tibial nerve were inserted vertically into the thalamus of anaesthetized adult albino rats. The distal end of the graft was left beneath the scalp. Between five days and two months later the animals were killed and the brains prepared for ultrastructural study. Semi-thin and thin sections through the graft and surrounding brain were examined at two levels 6-7 mm apart in all animals: close to the tip of the graft in the thalamus (proximal graft) and at the top of the cerebral cortex (distal graft). In another series of animals with similar grafts, horseradish peroxidase was applied to the distal end of the graft 24-48 h before death. Examination by LM of appropriately processed serial coronal sections of the brains from these animals confirmed that up to several hundred neurons were retrogradely labelled in the thalamus, particularly in the thalamic reticular nucleus. Between five and 14 days after grafting, large numbers of tiny (0.05-0.20 microns diameter) nonmyelinated axonal profiles, considered to be axonal sprouts, were observed by EM within the narrow zone of abnormal thalamic parenchyma bordering the graft. The sprouts were much more numerous (commonly in large fascicles), smoother surfaced, and more rounded than nonmyelinated axons further from the graft or in corresponding areas on the contralateral side of animals with implants or in normal animals. At longer post-graft survival times, the number of such axons in the parenchyma around the graft declined. At five days, some axonal sprouts had entered the junctional zone between the brain and the graft. By eight days there were many sprouts in the junctional zone and some had penetrated the proximal graft to lie between its basal lamina-enclosed columns of Schwann cells, macrophages and myelin debris. Within the brain, sprouts were in contact predominantly with other sprouts but also with all types of glial cell.(ABSTRACT TRUNCATED AT 400 WORDS)

Isthmotectal axons make ectopic synapses in monocular regions of the tectum in developing Xenopus laevis frogs

During the development of binocular maps in the tectum of Xenopus laevis, axons that relay input from the ipsilateral eye via the nucleus isthmi undergo a prolonged period of shifting connections. This shifting accompanies the dramatic change in eye position that takes place as the laterally placed eyes of the tadpole move dorsofrontally. There is a concomitant expansion of the proportion of tectum that receives contralateral retinotectal input corresponding to the binocular portion of the visual field. Electrophysiological recording demonstrates that ipsilateral units are present in those rostral tectal zones, and anatomical methods show that the isthmotectal axons arborize densely in the rostral region but also extend sparser branches into the caudal zone, which is occupied by contralateral inputs with receptive fields in the monocular zone of the visual field. A mechanism that aligns the ipsilateral and contralateral maps is activity-dependent stabilization of isthmotectal axons that exhibit firing patterns correlated with those of nearby retinotectal axons. In order for activity patterns to function in stabilizing correct connections and promoting the withdrawal of incorrect connections, synaptic communication of some sort is hypothesized to be essential. We have investigated whether isthmotectal axons make morphologically identifiable synapses during development and where such synapses are located. We find evidence for morphologically identifiable synapses in all regions of the tectum, along with many growth cones and structures that are probably immature synapses. As in the adult, the synapses contain round, clear vesicles, have asymmetric specializations, and terminate on structures that appear to be dendrites. In both adult and tadpole, the rarity of serial synapses involving isthmotectal terminals suggests that the interactions between retinotectal and isthmotectal inputs are mediated by postsynaptic dendrites.

Postnatal development of type I retinal ganglion cells in hamsters: a lucifer yellow study

The postnatal development of a population of superior colliculus projecting retinal ganglion cells with large somata in hamsters aged from postnatal day (P) 4 to adult was studied by the intracellular injection of Lucifer Yellow. This population of cells was interpreted as Type I cells based on their large soma sizes and dendritic morphology resembling that of mature Type I cells. In addition to the growth of the soma and the dendritic field, transient morphological features such as intraretinal axon collaterals and exuberant dendritic spines, but not somatic spines, were frequently observed on this population of cells in hamsters during development. None of them exhibited any intraretinal axon collaterals after P7. The number of transient spine-like processes on dendrites increased from P4 onwards to reach a peak at P16, decreased abruptly within a few days after the peak, and stabilised to reach the adult level by P30. These developing cells attained the maximum number of dendritic branches by P16 and there seems to be little, if any, reduction in the number of branch points after this time point. In addition, the length of individual branches of dendrites was not increased excessively during development and then shortened during maturation. Thus, the dendritic remodeling of these cells after P16 seems to be mainly the increase of the length of dendrites and the removal of exuberant dendritic spines.

Neonatal whisker removal in rats stabilizes a transient projection from the auditory thalamus to the primary somatosensory cortex

A normally transient cross-modal thalamocortical projection from the magnocellular subdivision of the medial geniculate nucleus (MGm) to the primary somatosensory (SI) cortex of rats was found to remain unchanged throughout adulthood following unilateral removal of whiskers in newborn animals. The normal MGm projection to the auditory cortex is not lost in these neonatally whisker-deprived adults rats but some of the MGm neurons send collaterals to both primary auditory and SI cortices. Parallel electrophysiological experiments demonstrated the multimodal character of some MGm neurons, since they responded to both auditory and cutaneous stimulation. These results suggest that the areal distribution in the cortex of thalamocortical projections arising from a multimodal thalamic nucleus, such as the MGm, may be determined during early postnatal development by the normal flow of sensory information from the periphery to the thalamus and that an early postnatal somatosensory deprivation may prevent the normal withdrawal of a cross-modal projection from the MGm to the SI.

Ultrastructure and synaptic organization of axon terminals from brainstem structures to the mediodorsal thalamic nucleus of the rat

The ultrastructural characteristics and synaptic organization of afferent terminals from the brainstem to the mediodorsal thalamic nucleus (MD) of the rat have been studied with the electron microscope, by means of anterograde transport of wheat germ agglutinin-horseradish peroxidase (WGA-HRP). Labeled fibers were seen predominantly in the lateral portion of MD after the injections of WGA-HRP into the substantia nigra pars reticulata (SNr), the superior colliculus (SC), and the dorsal tegmental region (DT). The boutons arising from the SC were relatively small (less than 1.5 microns in diameter), formed asymmetric synaptic contacts with small dendrites and dendritic spines, and contained round synaptic vesicles. The axon terminals from the DT were mostly large boutons (2-4.5 microns) with asymmetric synaptic specializations and round vesicles. These boutons and their postsynaptic targets formed synaptic glomeruli that were entirely or partially ensheathed by glial lamellae. The ultrastructural features are almost identical to those of boutons in the medial and central segments of MD that were previously shown to originate from the basal amygdaloid nucleus and the piriform cortex. The boutons from the SNr had a wide range in size, but the majority were medium-sized to large (1.5-4 microns). The nigral boutons established symmetric synaptic contacts with dendritic shafts and occasionally with somata, and contained pleomorphic vesicles. However, like the DT terminals, they participated in glomerular formations. The nigral terminals closely resemble previously described terminals in the medial part of MD from the ventral pallidum, except that the nigral terminals formed en passant and axosomatic synapses as well as axodendritic synapses. A combined immunohistochemistry and WGA-HRP tracing study revealed that the nigral inputs were immunoreactive for glutamic acid decarboxylase and the axon terminals from the DT were immunoreactive for choline acetyltransferase. In a separate study, the colliculothalamic fibers have been shown to take up and transport the transmitter specific tracer [3H]-D-aspartate, and are therefore putatively glutamatergic and/or aspartatergic. Taken together with this, the present results suggest that the collicular afferents are excitatory and glutamatergic and/or aspartatergic, that the inputs from the DT are also excitatory and cholinergic, while the nigral inputs are inhibitory and GABAergic.

Regenerated synapses persist in the superior colliculus after the regrowth of retinal ganglion cell axons

Synapse formation by retinal ganglion cell axons was sought in the superior colliculus of four adult rats 16-18 months after the optic nerve was transected and replaced by a peripheral nerve graft that guided regenerating RGC axons from the eye to the superior colliculus. The terminals of retinal ganglion cell axons were labelled by intravitreal injections of tritiated amino acids and studied by light and electron microscopic autoradiography. We found that (i) retinal ganglion cell axons had extended from the tips of the peripheral nerve grafts into the superior colliculus for approximately 350 microns; (ii) within the superior colliculus, some regenerated retinal ganglion cell axons became ensheathed by CNS myelin; (iii) retinal ganglion cell terminals formed asymmetric synapses with dendrites of neurons in the superficial layers of the superior colliculus, mainly the stratum griseum superficialis. Regenerated (n = 418) and normal retinal ganglion cell terminals (n = 1775) in the superior colliculus were compared in terms of their size (area, perimeter, and maximum diameter), contacts per terminal, contacts per 10 microns terminal perimeter, and post-synaptic structure contacted (dendritic spine, shaft, or soma). No statistically significant differences in the ultrastructural characteristics of the pre-synaptic profiles were apparent between the two groups. The post-synaptic structures contacted by axon terminals were similar in regenerated and control animals, although there were quantitative differences in the distributions of these contacts among dendritic spines and shafts. These results suggest that the regeneration of retinal ganglion cell axons in adult rats can lead to the formation of ultrastructurally normal synapses in the appropriate layers of the superior colliculus. The re-formed connections appear to persist for the life-span of these animals.

Transient retinal axon collaterals to visual and somatosensory thalamus in neonatal hamsters

We have studied the postnatal development of individual axons in the optic tract and thalamus of the Syrian hamster, concentrating attention on retinal ganglion cell axons that make a transient projection to the main somatosensory nucleus, the ventrobasal complex. We bulk-filled axons with horseradish peroxidase in hemithalami maintained en bloc, in vitro. After processing and reaction with diaminobenzidine, we reconstructed individual axons from serial sections. In hamsters and other rodents, the optic tract is composed of superficial and internal components, either or both being possible sources of the retino-ventrobasal projection. Both project to the midbrain, but in normal adults only the superficial optic tract maintains collaterals in the thalamus. We found that the axons of the internal component bear numerous transient thalamic collaterals on postnatal days 0, 1, and 2, and some of these extend into the ventrobasal complex. Axons in the superficial optic tract also bear collaterals on days 0 to 2, but these are confined to the superficial half of the dorsal lateral geniculate nucleus. Thus the transient retino-ventrobasal projection comprises solely transient collaterals originating from axon trunks in the internal optic tract. On days 1 and 2, some collaterals from the superficial optic tract appear to have begun to arborize in the lateral geniculate nucleus. In contrast, collaterals from internal optic tract axons to the ventrobasal complex branch little if at all as they traverse the lateral geniculate nucleus, and at no time prior to their elimination do they develop an appreciable terminal arbor. These long collaterals often terminate in growth cones that include lamellopodia. Our HRP-impregnation method also revealed some transient non-retinofugal axons that pass medially from the ventral lateral geniculate nucleus to the ventrobasal complex but then return without terminating or branching. By day 4, they are absent, as are collaterals from the internal optic tract to the ventrobasal complex.

Synaptic connections of the auditory nerve in cats: relationship between endbulbs of held and spherical bushy cells

This report focuses on a class of large synaptic endings, the endbulbs of Held. These endings are located in the anteroventral cochlear nucleus and arise from the axons of type I spiral ganglion neurons. Axons were stained with horseradish peroxidase (HRP) using intracellular injections of single fibers or extracellular injections into the auditory nerve. Individual endbulbs or pairs of endbulbs that converged onto the same spherical bushy cell were examined with the aid of a light microscope and subjected to morphometric analyses. Endbulbs of fibers having low spontaneous discharge rates (SR, less than or equal to 18 spikes/sec) have a more complex shape than those of high SR fibers (greater than 18 s/s), a feature represented by systematic differences in endbulb silhouette perimeter without differences in endbulb silhouette area. Consequently, the ratio, silhouette area divided by silhouette perimeter, yields a "form factor" separating endbulbs of high SR from those of low SR. High SR fibers had ratios greater than 0.52 (mean = 0.63 +/- 0.09), whereas low SR fibers had ratios less than 0.52 (mean = 0.45 +/- 0.06). Pairs of endbulbs with unknown physiological properties had similar form factor values, despite the wide range of values observed in the endbulb population. These data imply that endbulbs converging upon the cell body of a spherical bushy cell arise from fibers of the same SR group. Electron microscopic examination was conducted on the endbulb of one physiologically characterized and intracellularly stained auditory nerve fiber (CF = 1.4 kHz; SR = 55 s/s) and its unstained endbulb mate with the aid of serial ultrathin sections. In addition to the well-known axosomatic synapses, these endbulbs formed axodendritic synapses: 11.7% for the HRP-labeled endbulb and 13.3% for the unlabeled endbulb. The axodendritic synapses appear to occur on dendrites of nearby spherical bushy cells and may represent a mechanism whereby single endbulbs can disperse activity to multiple neurons in the cochlear nucleus. We propose that axosomatic synapses preserve fiber SR groupings, whereas axodendritic synapses may not.

Synaptic organization of projections from basal forebrain structures to the mediodorsal thalamic nucleus of the rat

The synaptic organization of the mediodorsal thalamic nucleus (MD) in the rat was studied with the electron microscope, and correlated with the termination of afferent fibers labeled with wheat germ agglutinin conjugated to horseradish peroxidase (WGA-HRP). Presynaptic axon terminals were classified into four categories in MD on the basis of the size, synaptic vesicle morphology, and synaptic membrane specializations: 1) small axon terminals with round synaptic vesicles (SR), which made asymmetrical synaptic contacts predominantly with small dendritic shafts; 2) large axon terminals with round vesicles (LR), which established asymmetrical synaptic junctions mainly with large dendritic shafts; 3) small to medium axon terminals with pleomorphic vesicles (SMP), which formed symmetrical synaptic contacts with somata and small-diameter dendrites; 4) large axon terminals with pleomorphic vesicles (LP), which made symmetrical synaptic contacts with large dendritic shafts. Synaptic glomeruli were also identified in MD that contained either LR or LP terminals as the central presynaptic components. No presynaptic dendrites were identified. In order to identify terminals arising from different sources, injections of WGA-HRP were made into cortical and subcortical structures known to project to MD, including the prefrontal cortex, piriform cortex, amygdala, ventral pallidum and thalamic reticular nucleus. Axons from the amygdala formed LR terminals, while those from the prefrontal and insular cortex ended exclusively in SR terminals. Fibers labeled from the piriform cortex formed both LR and SR endings. Based on their morphology, all of these are presumed to be excitatory. In contrast, the axons from the ventral pallidum ended as LP terminals, and those from the thalamic reticular nucleus formed SMP terminals. Both are presumed to be inhibitory. At least some terminals from these sources have also been identified as GABAergic, based on double labeling with anterogradely transported WGA-HRP and glutamic acid decarboxylase (GAD) immunocytochemistry.

Plasticity of GABA- and glutamate-containing terminals in the mouse thalamic ventrobasal complex deprived of vibrissal afferents: an immunogold-electron microscopic study

GABA and glutamate immunogold staining demonstrated that nerve cells of the thalamic ventrobasal complex (VB) of mice were positive exclusively for glutamate. None of the neuronal perikarya reacted the GABA antibody. By using alternate thin sections of the normal VB, it was also shown that large "specific" somatosensory and small corticothalamic terminals, both of which contained spherical synaptic vesicles, exhibited only glutamate-like immunoreactivity. A third axonal type, containing flat-ovoid synaptic vesicles, stained only for GABA. Seventy-five days after coagulation of the vibrissal follicles in newborn mice, a characteristic multiplication of GABA positive axon terminals was observed. In addition, it was demonstrated that, similarly to modified cortical endings (Hámori et al., J. Comp. Neurol. 254:166-183, '86), many GABA positive terminals appeared as specific afferent endings, replacing the missing "specific" vibrissal afferents. This finding shows a remarkable plasticity of inhibitory GABA axons during developmental synaptogenesis and provides further evidence that the size, location, and the type of attachment of presynaptic terminals are dependent on their postsynaptic target.

Regeneration of adult dorsal root axons into transplants of fetal spinal cord and brain: a comparison of growth and synapse formation in appropriate and inappropriate targets

Cut dorsal root axons regenerate into transplants of embryonic spinal cord and form synapses that resemble those found in the dorsal horn of normal spinal cord. One aim of the present study was to determine whether these axons also regenerate into and establish synapses within transplants of embryonic brain. A second aim was to compare the patterns of growth in embryonic brain and spinal cord transplants. Embryonic spinal cord or brain was transplanted into the lumbar enlargement of adult Sprague-Dawley rats, the L4 or L5 dorsal root was cut, and the cut root was juxtaposed to the transplant. The transplants included whole pieces or dissociated cell suspensions of embryonic day 14 (E14) spinal cord, or whole pieces of E14 neocortex, E18 occipital cortex, E15 cerebellum, or E18 hippocampus. One month later the regenerated dorsal root axons were labeled by immunocytochemical methods to demonstrate calcitonin gene-related peptide (CGRP). CGRP-immunoreactive axons regenerated into all the transplants examined and formed synapses in the neocortex and cerebellum transplants in which they were sought. Synapses were far rarer in neocortex and cerebellum than we had observed previously in transplanted spinal cord, and the patterns of growth differed in transplants of spinal cord and brain. In solid transplants of spinal cord, regenerated axons remained relatively close to the interface with the dorsal root, branched, and formed bundles. Areas of dense ingrowth were separated by regions with few labeled axons. In transplants of brain regions, the regenerated axons were few, unbranched, and appeared as individual fibers rather than in bundles, but they were distributed widely in neocortex transplants. The results of quantitative studies confirmed these observations. The area fraction occupied by regenerated axons in solid spinal cord transplants was significantly larger than in occipital cortex or cerebellum transplants. Distribution histograms of the area occupied in transplants demonstrated that regenerated axons were distributed sparsely but homogeneously in transplants of brain, whereas spinal cord transplants were heterogeneous for regenerated axons and contained areas in which growth was dense or sparse. In contrast, several measurements of axon distribution, including area, longest axis, and length of lateral extension, indicated that CGRP-labeled axons spread more widely in occipital cortex transplants than in solid transplants of spinal cord or cerebellum. The results indicate that embryonic CNS tissues that are not normal targets support or enhance the growth of severed dorsal roots and suggest that the conditions that constitute a permissive environment for regenerating axons are relatively nonspecific.(ABSTRACT TRUNCATED AT 400 WORDS)

Abnormal retinal projections alter GAP-43 patterns in the diencephalon

In Syrian hamsters, mature retinal terminals contain only low levels of the growth-associated protein, GAP-43, whereas the lateral posterior nucleus (LP) of the thalamus contains high levels of this protein. Damage to the superior colliculus in neonatal hamsters induces retinal terminals to form dense patches of innervation in the LP, an area which otherwise receives little if any direct retinal input. The present study used GAP-43 antibodies to examine the interaction between abnormally routed optic fibers and the cells in the anomalous thalamic target zone. Immunohistochemistry revealed very little GAP-43 in the abnormal retinal projection to the LP, indicating that the normal developmental decline in GAP-43 levels occurs even in an inappropriate extracellular environment. Moreover, retinal fibers were found to exclude the protein from its normal territory, forming negatively-stained islands in those regions of the LP containing the retinal terminals. In order to identify the normal source of GAP-43-positive terminals in the LP, we surgically removed two major extrinsic afferents to this region, or we chemically eliminated local interneurons. Whereas removing projections from the SC or posterior cortex did not alter GAP-43 immunoreactivity in the LP, destruction of local interneurons with ibotenic acid resulted in markedly diminished levels of this protein. These results show that retinal terminals induced to form in an abnormal target area undergo their normal diminution of GAP-43, and that these retinal projections displace other GAP-43-rich terminals in the LP that appear to arise from local interneurons.

Induction of somatic sensory inputs to the lateral geniculate nucleus in congenitally blind mice and in phenotypically normal mice

The incidence of aberrant innervation of the lateral geniculate nucleus by ascending somatic sensory axons was examined following injections of wheat germ agglutinin conjugated with horseradish peroxidase into the dorsal column nuclei of adult mice which were: (1) normal; (2) normal, but bilaterally enucleated on the day of birth; (3) normal, but received a large unilateral lesion of the rostral cortex on the day of birth; (4) normal, bilaterally enucleated, as well as unilaterally lesioned in the rostral cortex on the day of birth; (5) homozygous for an ocular retardation mutation (orj/orj); or (6) homozygous for the orj mutation and received a large unilateral lesion of the rostral cerebral cortex on the day of birth. In the phenotypically normal animals which were untreated, no somatic sensory inputs enter into the dorsal lateral geniculate nucleus. A few labeled axons enter into and arborize within the dorsal lateral geniculate nucleus in normal animals which received bilateral enucleations or unilateral rostral cortical lesions on the day of birth. However, in congenitally blind animals and in phenotypically normal animals which received bilateral enucleations as well as unilateral rostral cortical lesions on the day of birth, a significant number of labeled axons enter into and arborize within the dorsal lateral geniculate nucleus. Among all these experimental groups, the densest innervation of the lateral geniculate nucleus occurred in congenitally blind animals which received rostral cortical lesions on the day of birth. In these, robust arborizations of labeled somatic sensory axons occupy a substantial extent of the lateral geniculate nucleus. These results not only demonstrate that ascending somatic sensory axons can be rerouted to the lateral geniculate nucleus, but also indicate that the ability of a thalamic afferent pathway to undergo extensive reorganization and to innervate inappropriate thalamic targets following early perturbations is not unique to the retinal projection (in which this has previously been demonstrated), and may be a more general characteristic of the major thalamic afferent systems.

Grafting of fetal CA3 neurons to excitotoxic, axon-sparing lesions of the hippocampal CA3 area in adult rats

Hippocampal CA3 neurons from fetal rats were grafted to excitotoxic lesions in the CA3 subfield of the adult rat hippocampus and the formation of graft-host brain nerve connections examined. The excitotoxic lesions were induced by localized, stereotaxic injection of ibotenic acid (IA), a glutamic acid agonist, into CA3 of the dorsal hippocampus. The result was a so-called axon-sparing lesion with localized degeneration of nerve cells, but preservation of the extrinsic afferent fibers, now deprived of their targets. One week after the lesion a suspension of embryonic (E18-20) CA3 cells was grafted to the lesion site. Six weeks or more later the recipient brains were processed and analyzed by ordinary cell stains, histochemistry for acetylcholinesterase (AChE) and heavy metals (Timm staining), immunohistochemistry for the neuropeptides cholecystokinin and somatostatin and glial fibrillary acidic protein (GFAP) for astroglia, electron microscopy, and axonal tracing with retrogradely axonal transported fluorescent dyes or lesion-induced, anterograde degeneration combined with silver staining or electron microscopy. More than 90% of the grafts survived. They contained the normal types of CA3 neurons, which are mainly pyramidal cells, in addition to some normal, peptidergic, cholecystokinin- and somatostatin-reactive neurons. The grafts were innervated by AChE-positive, host cholinergic fibers, Timm-positive mossy fiber terminals from the host fascia dentata, and host commissural fibers traced by axonal degeneration. Efferent transplant projections were traced to the ipsilateral host CA1 (Schaffer collaterals) and the contralateral host hippocampus by retrograde axonal transport of fluorochromes injected into these host brain areas. All grafts analyzed by electron microscopy contained axonal varicosities resembling axonal growth cones even after long survival times. The results demonstrate that fetal rat hippocampal neurons, grafted to excitotoxic, axon-sparing lesions in the adult brain, can become both structurally and connectively well incorporated in the mature host central nervous system.

Auditory pathway and auditory activation of primary visual targets in the blind mole rat (Spalax ehrenbergi): I. 2-deoxyglucose study of subcortical centers

The blind mole rat Spalax ehrenbergi is a subterranean rodent that shows striking behavioral, structural, and physiological adaptations to fossorial life including highly degenerated eyes and optic nerves and a behavioral audiogram that indicates high specialization for low-frequency hearing. A 2-deoxyglucose functional mapping of acoustically activated structures, in conjunction with Nissl/Klüver-Barrera-stained material, revealed a typical mammalian auditory pathway with some indications for specialized low-frequency hearing such as a poorly differentiated lateral nucleus and a well-developed medial nucleus in the superior olive complex. The most striking finding was a marked 2-deoxyglucose labeling of the dorsal lateral geniculate body and of cortical regions that correspond to visual areas in sighted rodents. The results render the blind mole rat a good model system for studying natural neural plasticity and intermodal compensation. In this report, we confine ourselves to the subcortical levels. The cortical level will be dealt comprehensively in a following paper.

Visual responses of neurons in somatosensory cortex of hamsters with experimentally induced retinal projections to somatosensory thalamus

These experiments investigate the capacity of thalamic and cortical structures in a sensory system to process information of a modality normally associated with another system. Retinal ganglion cells in newborn Syrian hamsters were made to project permanently to the main thalamic somatosensory (ventrobasal) nucleus. When the animals were adults, single unit recordings were made in the somatosensory cortices, the principal targets of the ventrobasal nucleus. The somatosensory neurons responded to visual stimulation of distinct receptive fields, and their response properties resembled, in several characteristic features, those of normal visual cortical neurons. In the visual cortex of normal animals and the somatosensory cortex of operated animals, the same functional categories of neurons occurred in similar proportions, and the neurons' selectivity for the orientation or direction of movement of visual stimuli was comparable. These results suggest that thalamic nuclei or cortical areas at corresponding levels in the visual and somatosensory pathways perform similar transformations on their inputs.