The synaptic organization of terminals traced from individual labeled retino-geniculate axons in the cat

Functional convergence of on-off direction-selective ganglion cells in the visual thalamus

In the mouse visual system, multiple types of retinal ganglion cells (RGCs) each encode distinct features of the visual space. A clear understanding of how this information is parsed in their downstream target, the dorsal lateral geniculate nucleus (dLGN), remains elusive. Here, we characterized retinogeniculate connectivity in Cart-IRES2-Cre-D and BD-CreER2 mice, which labels subsets of on-off direction-selective ganglion cells (ooDSGCs) tuned to the vertical directions and to only ventral motion, respectively. Our immunohistochemical, electrophysiological, and optogenetic experiments reveal that only a small fraction (<15%) of thalamocortical (TC) neurons in the dLGN receives primary retinal drive from these subtypes of ooDSGCs. The majority of the functionally identifiable ooDSGC inputs in the dLGN are weak and converge together with inputs from other RGC types. Yet our modeling indicates that this mixing is not random: BD-CreER⁺ ooDSGC inputs converge less frequently with ooDSGCs tuned to the opposite direction than with non-CART-Cre⁺ RGC types. Taken together, these results indicate that convergence of distinct information lines in dLGN follows specific rules of organization.

Conversations with Ray Guillery on albinism: linking Siamese cat visual pathway connectivity to mouse retinal development

In albinism of all species, perturbed melanin biosynthesis in the eye leads to foveal hypoplasia, retinal ganglion cell misrouting, and, consequently, altered binocular vision. Here, written before he died, Ray Guillery chronicles his discovery of the aberrant circuitry from eye to brain in the Siamese cat. Ray's characterization of visual pathway anomalies in this temperature sensitive mutation of tyrosinase and thus melanin synthesis in domestic cats opened the exploration of albinism and simultaneously, a genetic approach to the organization of neural circuitry. I follow this account with a remembrance of Ray's influence on my work. Beginning with my postdoc research with Ray on the cat visual pathway, through my own work on the mechanisms of retinal axon guidance in the developing mouse, Ray and I had a continuous and rich dialogue about the albino visual pathway. I will present the questions Ray posed and clues we have to date on the still-elusive link between eye pigment and the proper balance of ipsilateral and contralateral retinal ganglion cell projections to the brain.

Synaptic organization of the dorsal lateral geniculate nucleus

A half century after Ray Guillery's classic descriptions of cell types, axon types, and synaptic architecture of the dorsal lateral geniculate nucleus, the functional organization of this nucleus, as well as all other thalamic nuclei, is still of enormous interest. This review will focus on two classic papers written by Ray Guillery: 'A study of Golgi preparations from the dorsal lateral geniculate nucleus of the adult cat', and 'The organization of synaptic interconnections in the laminae of the dorsal lateral geniculate nucleus of the cat', as well as the studies that most directly followed from the insights these landmark manuscripts provided. It is hoped that this review will honor Ray Guillery by encouraging further investigations of the synaptic organization of the dorsal thalamus.

An evolving view of retinogeniculate transmission

The thalamocortical (TC) relay neuron of the dorsoLateral Geniculate Nucleus (dLGN) has borne its imprecise label for many decades in spite of strong evidence that its role in visual processing transcends the implied simplicity of the term "relay". The retinogeniculate synapse is the site of communication between a retinal ganglion cell and a TC neuron of the dLGN. Activation of retinal fibers in the optic tract causes reliable, rapid, and robust postsynaptic potentials that drive postsynaptics spikes in a TC neuron. Cortical and subcortical modulatory systems have been known for decades to regulate retinogeniculate transmission. The dynamic properties that the retinogeniculate synapse itself exhibits during and after developmental refinement further enrich the role of the dLGN in the transmission of the retinal signal. Here we consider the structural and functional substrates for retinogeniculate synaptic transmission and plasticity, and reflect on how the complexity of the retinogeniculate synapse imparts a novel dynamic and influential capacity to subcortical processing of visual information.

Retinal and Tectal "Driver-Like" Inputs Converge in the Shell of the Mouse Dorsal Lateral Geniculate Nucleus

The dorsal lateral geniculate nucleus (dLGN) is a model system for understanding thalamic organization and the classification of inputs as "drivers" or "modulators." Retinogeniculate terminals provide the primary excitatory drive for the relay of information to visual cortex (V1), while nonretinal inputs act in concert to modulate the gain of retinogeniculate signal transmission. How do inputs from the superior colliculus, a visuomotor structure, fit into this schema? Using a variety of anatomical, optogenetic, and in vitro physiological techniques in mice, we show that dLGN inputs from the superior colliculus (tectogeniculate) possess many of the ultrastructural and synaptic properties that define drivers. Tectogeniculate and retinogeniculate terminals converge to innervate one class of dLGN neurons within the dorsolateral shell, the primary terminal domain of direction-selective retinal ganglion cells. These dLGN neurons project to layer I of V1 to form synaptic contacts with dendrites of deeper-layer neurons. We suggest that tectogeniculate inputs act as "backseat drivers," which may alert shell neurons to movement commands generated by the superior colliculus. Significance statement: The conventional view of the dorsal lateral geniculate nucleus (dLGN) is that of a simple relay of visual information between the retina and cortex. Here we show that the dLGN receives strong excitatory input from both the retina and the superior colliculus. Thus, the dLGN is part of a specialized visual channel that provides cortex with convergent information about stimulus motion and eye movement and positioning.

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.

Synaptic development of the mouse dorsal lateral geniculate nucleus

The dorsal lateral geniculate nucleus (dLGN) of the mouse has emerged as a model system in the study of thalamic circuit development. However, there is still a lack of information regarding how and when various types of retinal and nonretinal synapses develop. We examined the synaptic organization of the developing mouse dLGN in the common pigmented C57/BL6 strain, by recording the synaptic responses evoked by electrical stimulation of optic tract axons, and by investigating the ultrastructure of identified synapses. At early postnatal ages (<P12), optic tract evoked responses were primarily excitatory. The full complement of inhibitory responses did not emerge until after eye opening (>P14), when optic tract stimulation routinely evoked an excitatory postsynaptic potential/inhibitory postsynaptic potential (EPSP/IPSP) sequence, with the latter having both a GABA(A) and GABA(B) component. Electrophysiological and ultrastructural observations were consistent. At P7, many synapses were present, but synaptic profiles lacked the ultrastructural features characteristic of the adult dLGN, and little gamma-aminobutyric acid (GABA) could be detected by using immunocytochemical techniques. In contrast, by P14, GABA staining was robust, mature synaptic profiles of retinal and nonretinal origin were easily distinguished, and the size and proportion of synaptic contacts were similar to those of the adult. The emergence of nonretinal synapses coincides with pruning of retinogeniculate connections, and the transition of retinal activity from spontaneous to visually driven. These results indicate that the synaptic architecture of the mouse dLGN is similar to that of other higher mammals, and thus provides further support for its use as a model system for visual system development.

Synaptic organization of thalamocortical axon collaterals in the perigeniculate nucleus and dorsal lateral geniculate nucleus

We examined the synaptic targets of large non-gamma-aminobutyric acid (GABA)-ergic profiles that contain round vesicles and dark mitochondria (RLD profiles) in the perigeniculate nucleus (PGN) and the dorsal lateral geniculate nucleus (dLGN). RLD profiles can provisionally be identified as the collaterals of thalamocortical axons, because their ultrastrucure is distinct from all other previously described dLGN inputs. We also found that RLD profiles are larger than cholinergic terminals and contain the type 2 vesicular glutamate transporter. RLD profiles are distributed throughout the PGN and are concentrated within the interlaminar zones (IZs) of the dLGN, regions distinguished by dense binding of Wisteria floribunda agglutinin (WFA). To determine the synaptic targets of thalamocortical axon collaterals, we examined RLD profiles in the PGN and dLGN in tissue stained for GABA. For the PGN, we found that all RLD profiles make synaptic contacts with GABAergic PGN somata, dendrites, and spines. In the dLGN, RLD profiles primarily synapse with GABAergic dendrites that contain vesicles (F2 profiles) and non-GABAergic dendrites in glomerular arrangements that include triads. Occasional synapses on GABAergic somata and proximal dendrites were also observed in the dLGN. These results suggest that correlated dLGN activity may be enhanced via direct synaptic contacts between thalamocortical cells, whereas noncorrelated activity (such as that occurring during binocular rivalry) could be suppressed via thalamocortical collateral input to PGN cells and dLGN interneurons.

Comparison of the ultrastructure of cortical and retinal terminals in the rat superior colliculus

We compared the ultrastructure and synaptic targets of terminals of cortical or retinal origin in the stratum griseum superficiale and stratum opticum of the rat superior colliculus. Following injections of biotinylated dextran amine into cortical area 17, corticotectal axons were labeled by anterograde transport. Corticotectal axons were of relatively small caliber with infrequent small varicosities. At the ultrastructural level, corticotectal terminals were observed to be small profiles (0.44 +/- 0.27 microm(2)) that contained densely packed round vesicles. In tissue stained for gamma amino butyric acid (GABA) using postembedding immunocytochemical techniques, corticotectal terminals were found to contact small (0.51 +/- 0.69 microm(2)) non-GABAergic dendrites and spines (93%) and a few small GABAergic dendrites (7%). In the same tissue, retinotectal terminals, identified by their distinctive pale mitochondria, were observed to be larger than corticotectal terminals (3.34 +/- 1.79 microm(2)). In comparison to corticotectal terminals, retinotectal terminals contacted larger (1.59 +/- 1.70 microm(2)) non-GABAergic dendrites and spines (73%) and a larger proportion of GABAergic profiles (27%) of relatively large size (2.17 +/- 1.49 microm(2)), most of which were vesicle-filled (71%). Our results suggest that cortical and retinal terminals target different dendritic compartments within the neuropil of the superficial layers of the superior colliculus.

Ultrastructural analysis of projections to the pulvinar nucleus of the cat. I: Middle suprasylvian gyrus (areas 5 and 7)

The mammalian pulvinar nucleus (PUL) establishes heavy interconnections with the parietal lobe, but the precise nature of these connections is only partially understood. To examine the distribution of corticopulvinar cells in the cat, we injected the PUL with retrograde tracers. Corticopulvinar cells were located in layers V and VI of a wide variety of cortical areas, with a major concentration of cells in area 7. To examine the morphology and distribution of corticopulvinar terminals, we injected cortical areas 5 or 7 with anterograde tracers. The majority of corticopulvinar axons were thin fibers (type I) with numerous diffuse small boutons. Thicker (type II) axons with fewer, larger boutons were also present. Boutons of type II axons formed clusters within restricted regions of the PUL. We examined corticopulvinar terminals labeled from area 7 at the ultrastructural level in tissue stained for gamma-aminobutyric acid (GABA). By correlating the size of the presynaptic and postsynaptic profiles, we were able to quantitatively divide the labeled terminals into two categories: small and large (RS and RL, respectively). The RS terminals predominantly innervated small-caliber non-GABAergic (thalamocortical cell) dendrites, whereas the RL terminals established complex synaptic arrangements with dendrites of both GABAergic interneurons and non-GABAergic cells. Interpretation of these results using Sherman and Guillery's recent theories of thalamic organization (Sherman and Guillery [1998] Proc Natl Acad Sci U S A 95:7121-7126) suggests that area 7 may both drive and modulate PUL activity.

A review on electron microscopy and neurotransmitter systems

The purpose of this article is to review the contributions of transmission electron microscopy studies to the understanding of brain circuits and neurotransmitter systems. Our views on the microstructure of connections between neurons have gradually changed, and now we recognize that the classical mental image we had on a chemical synapse is no longer applicable to every neuronal connection. We highlight studies that converge to point out that, while the most prevalent fast transmitters in the brain, glutamate and GABA, are stored in small, clear synaptic vesicles (SSV) and released at synapses, neuropeptides are exclusively stored in large dense core vesicles (LDCV) and released extrasynaptically. Amine transmitters are preferentially, but not exclusively, accumulated in LDCV and may be released at synaptic or extrasynaptic sites. We discuss evidence suggesting that axon terminals from pyramidal cortical neurons and dorsal thalamic neurons lack LDCV and therefore could not use neuropeptides as transmitters. This idea fits with the fast, high temporal resolution information processing that characterizes cortical and thalamic function.

Two distinct types of corticothalamic EPSPs and their contribution to short-term synaptic plasticity

The lateral posterior nucleus (LPN) is innervated by two different morphological types of cortical terminals that originate from cortical layers V and VI. Here we describe two distinct types of excitatory postsynaptic potentials (EPSPs) that were recorded in the LPN after stimulation of corticothalamic fibers. These types of EPSPs differed in amplitude, latency, rise time, and response to increasing levels of stimulus intensity. The most frequently encountered EPSP, type I, displayed a longer latency and slower rise time than the less frequently encountered type II EPSP. Type I EPSPs also showed a graded increase in amplitude with increasing levels of stimulation, whereas type II EPSPs showed an all-or-none response. In response to repetitive stimulation (0.5-20 Hz), type I EPSPs displayed frequency-dependent facilitation, whereas type II EPSPs displayed frequency-dependent depression. Further details of these distinct forms of short-term synaptic plasticity were explored using paired-pulse stimuli. Pharmacology experiments revealed that both N-methyl-d-aspartate (NMDA) and non-NMDA glutamate receptors are involved in corticothalamic synaptic transmission in the LPN and contribute to both synaptic facilitation and depression. Taken together with the results of our previous anatomical studies, these results suggest that type I EPSPs arise from stimulation of layer VI afferents, whereas type II EPSPs arise from stimulation of layer V inputs. Moreover, type I and II EPSPs in the LPN may be functionally similar to corticogeniculate and retinogeniculate EPSPs, respectively.

Ultrastructure and synaptic targets of tectothalamic terminals in the cat lateral posterior nucleus

The recent appreciation of the fact that the pulvinar and lateral posterior (LP) nuclei receive two distinct types of cortical input has sparked renewed interest in this region of the thalamus. A key question is whether the primary or "driving" inputs to the pulvinar/LP complex originate in cortical or subcortical areas. To begin to address this issue, we examined the synaptic targets of tectothalamic terminals within the LP nucleus. Tectothalamic terminals were labeled using the anterograde transport of biotinylated dextran amine (BDA) or Phaselous leucoagglutinin placed in the superior colliculus or using immunocytochemical staining for substance P, a neurotransmitter found to be used by the tectothalamic pathway (Hutsler and Chalupa [ 1991] J. Comp. Neurol. 312:379-390). Our results suggest that most tectothalamic terminals are large and occupy a proximal position on the dendritic arbor of LP relay cells. In the medial LP, tectothalamic terminals labeled by the transport of neuronal tracers or substance P immunocytochemistry can form tubular clusters that surround the proximal dendrites of relay cells. In a rostral and lateral subdivision of the lateral LP nucleus (LPl-2), tectothalamic terminals form more typical glomerular arrangements. When compared with existing physiological data, these results suggest that a unique integration of tectal and cortical inputs may contribute to the response properties of LP neurons.

Relative distribution of synapses in the pulvinar nucleus of the cat: implications regarding the "driver/modulator" theory of thalamic function

To provide a quantitative comparison of the synaptic organization of "first-order" and "higher-order" thalamic nuclei, we followed bias-corrected sampling methods identical to a previous study of the cat dorsal lateral geniculate nucleus (dLGN; Van Horn et al. [2000] J. Comp. Neurol. 416:509-520) to examine the distribution of terminal types within the cat pulvinar nucleus. We observed the following distribution of synaptic contacts: large terminals that contain loosely packed round vesicles (RL profiles), 3.5%; presynaptic profiles that contain densely packed pleomorphic vesicles (F1 profiles), 7.3%; profiles that could be both presynaptic and postsynaptic that contain loosely packed pleomorphic vesicles (F2 profiles), 5.0%; and small terminals that contain densely packed round vesicles (RS profiles), 84.2%. Postembedding immunocytochemistry for gamma-aminobutyric acid (GABA) was used to distinguish the postsynaptic targets as thalamocortical cells or interneurons. The distribution of synaptic contacts on thalamocortical cells was as follows: RL profiles, 2.1%; F1 profiles, 6.9%; F2 profiles, 5.4%; and RS profiles, 85.6%. The distribution of synaptic contacts on interneurons was as follows: RL profiles, 11.8%; F1 profiles, 9.7%; F2 profiles, 2.8%; and RS profiles, 75.6%. These distributions are similar to that found within the dLGN in that the RS inputs (the presumed "modulators") far outnumber the RL inputs (the presumed "drivers"). However, in comparison to the dLGN, the pulvinar nucleus receives significantly fewer numbers of RL, F1, and F2 contacts and significantly higher numbers of RS contacts. Thus, the RS/RL synapse ratio in the pulvinar nucleus is 24:1, in contrast to the 5:1 RS/RL synapse ratio in the dLGN (Van Horn et al., 2000). In first-order nuclei, the lower RS/RL synapse ratio may result in the transfer of visual information that is largely unmodified. In contrast, in higher-order nuclei, the higher RS/RL synapse ratio may allow for a finer modulation of driving inputs.

GABAergic circuitry in the dorsal division of the cat medial geniculate nucleus

Gamma-aminobutyric acid (GABA) is the main inhibitory neurotransmitter of the thalamus. We used postembedding immunocytochemistry to examine the synaptic organization of GABA-positive profiles in the dorsal superficial subdivision of the cat medial geniculate nucleus (MGN). Three groups of GABA-positive profiles participate in synapses: axon terminals, dendrites, and presynaptic dendrites. The presynaptic GABA-positive terminals target mainly GABA-negative dendrites. The GABA-positive postsynaptic profiles receive input primarily from GABA-negative axons. The results indicate that the synaptic organization of GABA-positive profiles in the dorsal superficial subdivision of the MGN nucleus is very similar to that in other thalamic nuclei.

Synaptic and Nonsynaptic Localization of Benzodiazepine/GABAA Receptor/Cl- Channel Complex Using Monoclonal Antibodies in the Dorsal Lateral Geniculate Nucleus of the Cat

The two monoclonal antibodies, bd-17 and bd-24, are specific for beta- and alpha-subunits of the GABAA/benzodiazepine receptor/chloride channel complex respectively. An abundance of both subunits has been revealed in the visual thalamus of the cat by light microscopic immunocytochemistry using these antibodies. The alpha-subunit specific antibody and electron microscopy were used to determine the subcellular distribution of immunoreactivity with respect to specific cell classes in the dorsal lateral geniculate nucleus. Immunoreactivity was always associated with membranes and the degree of immunoreactivity varied greatly between different types of cell as defined by: (i) immunoreactivity for GABA; (ii) soma area; (iii) presence or absence of cytoplasmic laminated bodies (CLB). GABA negative neurons with the smallest soma area showed the strongest immunoreactivity, mainly in the endoplasmic reticulum and also on the somatic plasma membrane. Cytoplasmic laminated bodies could be found in the majority of these neurons. Large GABA negative cells without CLBs were strongly immunoreactive on the plasma membrane of the soma and dendrites, but showed scant if any intracellular immunoreactivity. GABA-positive cells showed weak intracellular immunoreactivity but negligible if any immunoreactivity at the somatic and proximal dendritic plasma membrane. A similar reaction pattern was found in GABA negative cells which contained no CLBs and which constituted a medium sized cell population. It is suggested that the degree of intracellular receptor immunoreactivity is positively correlated with receptor turnover. The dendrites of projection cells, particularly outside the glomeruli, showed strong immunoreactivity on the plasma membrane. The synaptic junctions formed by many boutons (F terminals) establishing symmetrical synapses with dendrites of relay cells were immunopositive, but no immunoreactivity could be detected at the synapses established by the presynaptic dendrites of the local interneurons. Many axo-somatic F1 junctions were also immunoreactive. However, immunoreactivity for the receptor/channel complex was also widely distribution on nonsynaptic plasma membranes of somata and dendrites. Thus GABA may act at both synaptic and non-synaptic sites. Furthermore, the correlation of immunoreactivity for the GABAA receptor complex with previously published properties of physiologically identified cells suggests that the strongly immunoreactive, small, GABA negative cells with CLBs might correspond to the 'lagged' X-type cells, and the large GABA negative receptor outlined cells without CLBs might correspond to some of the Y-type neurons.

Postnatal dendritic development of Y-like geniculocortical relay neurons

We describe the dendritic development of neurons in the dorsal lateral geniculate nucleus (LGNd) projecting to cortical area 18 in the postnatal cat. LGN neurons were identified by retrograde labeling from area 18 with fluorescent latex microspheres and injected in the fixed slice with Lucifer yellow (LY) and horseradish peroxidase (HRP) to visualize their dendritic arborizations. Both topological (measures of the patterns of dendritic branching and their territorial coverage) and metric parameters (measures of the quantitative parameters describing the size, length, extent and diameter of the dendritic arbors) were measured in three-dimensions for 25 LGN neurons in cats between 1 and 18 postnatal weeks. In addition, dendritic growth was compared to the changing dimensions of the LGNd. At all ages, neurons projecting to area 18 have large somata and radiate dendrites. From 1 to 18 weeks neurons increase in size--both soma area and the length of all dendritic segments double during this period. Intermediate and terminal dendritic segments show comparable growth until 5 weeks. However, only terminal segments continue to grow significantly from 5 until 18 weeks. Dendrites become straighter during development, the angle between daughter branches decreases and dendritic segment diameter increases, with terminal segments showing a greater increase relative to intermediate segments. The density of dendritic appendages increases transiently at 5 weeks and a differential redistribution occurs, so that by 18 weeks dendrites further from the soma have a greater density of appendages than those near the soma. Some dendritic relationships remain invariant during development--intermediate segments are always shorter, thicker and straighter than terminal segments. During these changes however, area 18 projecting neurons maintain a constant number of primary dendrites and have, on average, a constant branching pattern. The relative volume of the LGNd occupied by an area 18 projecting neuron increases 2.4-fold between 1 and 18 weeks as the dendrites grow with the result that the coverage of a given point of the LGNd by dendrites of area 18 projecting nearly doubles from 24 to 45 neurons per unit volume. This increased net dendritic overlap provides a substrate for enhanced numerical synaptic divergence of the Y-cell pathway from a point source in the retina to the visual cortex.

GABAergic pretectal terminals contact GABAergic interneurons in the cat dorsal lateral geniculate nucleus

Anterograde tracing techniques combined with postembedding immunocytochemical staining were used to determine the gamma amino butyric acid (GABA) content of pretectogeniculate (PT-LGN) terminals and their postsynaptic targets. The results provide evidence that PT-LGN terminals are GABAergic and that they contact GABAergic interneurons. These results corroborate previous anatomical studies and support the idea that the PT-LGN projection functions to disinhibit thalamocortical cells in the dorsal lateral geniculate nucleus.

Synaptic targets of thalamic reticular nucleus terminals in the visual thalamus of the cat

A major inhibitory input to the dorsal thalamus arises from neurons in the thalamic reticular nucleus (TRN), which use gamma-aminobutyric acid (GABA) as a neurotransmitter. We examined the synaptic targets of TRN terminals in the visual thalamus, including the A lamina of the dorsal lateral geniculate nucleus (LGN), the medial interlaminar nucleus (MIN), the lateral posterior nucleus (LP), and the pulvinar nucleus (PUL). To identify TRN terminals, we injected biocytin into the visual sector of the TRN to label terminals by anterograde transport. We then used postembedding immunocytochemical staining for GABA to distinguish TRN terminals as biocytin-labeled GABA-positive terminals and to distinguish the postsynaptic targets of TRN terminals as GABA-negative thalamocortical cells or GABA-positive interneurons. We found that, in all nuclei, the TRN provides GABAergic input primarily to thalamocortical relay cells (93-100%). Most of this input seems targeted to peripheral dendrites outside of glomeruli. The TRN does not appear to be a significant source of GABAergic input to interneurons in the visual thalamus. We also examined the synaptic targets of the overall population of GABAergic axon terminals (F1 profiles) within these same regions of the visual thalamus and found that the TRN contacts cannot account for all F1 profiles. In addition to F1 contacts on the dendrites of thalamocortical cells, which presumably include TRN terminals, another population of F1 profiles, most likely interneuron axons, provides input to GABAergic interneuron dendrites. Our results suggest that the TRN terminals are ideally situated to modulate thalamocortical transmission by controlling the response mode of thalamocortical cells.

Y retinal terminals contact interneurons in the cat dorsal lateral geniculate nucleus

One of the largest influences on dorsal lateral geniculate nucleus (dLGN) activity comes from interneurons, which use the neurotransmitter gamma-aminobutyric acid (GABA). It is well established that X retinogeniculate terminals contact interneurons and thalamocortical cells in complex synaptic arrangements known as glomeruli. However, there is little anatomical evidence for the involvement of dLGN interneurons in the Y pathway. To determine whether Y retinogeniculate axons contact interneurons, we injected the superior colliculus (SC) with biotinylated dextran amine (BDA) to backfill retinal axons, which also project to the SC. Within the A lamina of the dLGN, this BDA labeling allowed us to distinguish Y retinogeniculate axons from X retinogeniculate axons, which do not project to the SC. In BDA-labeled tissue prepared for electron microscopic analysis, we subsequently used postembedding immunocytochemical staining for GABA to distinguish interneurons from thalamocortical cells. We found that the majority of profiles postsynaptic to Y retinal axons were GABA-negative dendrites of thalamocortical cells (117/200 or 58.5%). The remainder (83/200 or 41.5%) were GABA-positive dendrites, many of which contained vesicles (59/200 or 29.5%). Thus, Y retinogeniculate axons do contact interneurons. However, these contacts differed from X retinogeniculate axons, in that triadic arrangements were rare. This indicates that the X and Y pathways participate in unique circuitries but that interneurons are involved in the modulation of both pathways.

Neuropeptide Y and enkephalin immunoreactivity in retinorecipient nuclei of the hamster pretectum and thalamus

This investigation was stimulated by the historical confusion concerning the identity of certain pretectal nuclei and by large differences reported between species with respect to which nuclei receive retinal innervation. Subcortical visual nuclei were studied using immunohistochemistry to identify retinal projections labeled following intraocular injection of cholera toxin, b fragment. In addition, neuropeptide Y (NPY) or enkephalin (ENK) immunoreactive cells and fibers were also evaluated in the retinorecipient pretectal and thalamic areas. The results confirm the established view that the retina directly innervates the nucleus of the optic tract (NOT), posterior (PPT), and olivary pretectal (OPT) nuclei. However, the retina also innervates the hamster medial (MPT) and anterior (APT; dorsal division) pretectal nuclei, results not previously reported in rodents. A commissural pretectal area (CPT) sparsely innervated by retina is also described. The data show for the first time that the posterior limitans nucleus (PLi) receives a moderately dense, direct retinal input. The PLi does not project to the cortex and appears to be a pretectal, rather than thalamic, nucleus. All retinal projections are bilateral, although predominantly contralateral. The PLi contains a moderately dense plexus of NPY- and ENK-IR fibers and terminals. However, peptidergic fibers also traverse the ATP and connect with the dorsomedial pretectium. The OPT contains ENK- and NPY-IR neurons and fibers, but is specifically identifiable by a moderately dense plexus of ENK-IR terminals. Numerous ENK-IR neurons are found in the NOT and PPT. The latter also has moderate numbers of ENK-IR fibers and terminals, but few NPY-IR neurons or fibers. The MPT contains modest numbers of ENK-IR fibers. The APT has no NPY-IR neurons or terminals, but an occasional ENK-IR neuron is seen and there is sparse ENK-IR innervation. Peptidergic innervation of the visual nuclei does not appear to be derived from the retina. The results show a set of retinally innervated, contiguous nuclei extending from the thalamic ventrolateral geniculate nucleus dorsomedially to the midbrain CPT. These nuclei plus the superior colliculus comprise a dorsal "visual shell" embracing a central core of caudal thalamus and rostral midbrain.

Morphological diversity and glutamate immunoreactivity of retinal terminals in the suprachiasmatic nucleus of the cat

Although the cat visual system has been the subject of intensive investigation, little attention has been given to the morphological features of ganglion cell projections to the suprachiasmatic nucleus. The present study has utilized anterograde transport of horseradish peroxidase and wheat germ agglutinin-conjugated horseradish peroxidase to label ganglion cell terminals in the cat suprachiasmatic nucleus. Visualization of the reaction product was facilitated through the use of gold-substituted silver intensification. Ganglion cell terminals were found to be morphologically diverse, making both asymmetric and symmetric contacts with postsynaptic processes. Synaptic vesicles were either scattered or densely packed, sometimes forming paracrystalline arrays. In contrast to other retinorecipient areas in which ganglion cell terminals have been characterized by the presence of lightly staining mitochondria, many of the retinal terminals in the suprachiasmatic nucleus were seen to contain darkly stained mitochondria. Postembedding antiglutamate immunocytochemistry was used to evaluate the level of endogenous glutamate in these ganglion cell terminals. Although morphologically diverse, all of the retinal terminals in the suprachiasmatic nucleus were glutamate positive, consistent with the postulated role of glutamate as the neurotransmitter of retinal ganglion cells.

Quantitative immunogold evidence for enrichment of glutamate but not aspartate in synaptic terminals of retino-geniculate, geniculo-cortical, and cortico-geniculate axons in the cat

A postembedding immunogold procedure was used on thin sections of the dorsal lateral geniculate nucleus (LGN) and perigeniculate nucleus (PGN) of the cat to estimate qualitatively and quantitatively, at the electron-microscopic (EM) level, the intensity of glutamate or aspartate immunoreactivities on identifiable synaptic terminals and other profiles of the neuropil. On sections incubated with a glutamate antibody, terminals of retinal and cortical axons in the LGN, and of collaterals of geniculo-cortical axons in the PGN, contain significantly higher density of immunogold particles than GABAergic terminals, glial cells, dendrites, and cytoplasm of geniculate cells. By contrast, in sections incubated with an aspartate antibody, terminals of retino-geniculate, cortico-geniculate, and geniculo-cortical axons did not show a selective enrichment of immunoreactivity, but instead the density of immunogold particles was generally low in the different profiles of the neuropil, with the exception of nucleoli. These results suggest that glutamate, but not aspartate, is a neurotransmitter candidate in the retino-geniculo-cortical pathways.

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

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

GABAergic neurons and circuits in the pretectal nuclei and the accessory optic system of mammals

Two classes of GABAergic cell bodies have been described. They probably can be divided into GABAergic local interneurons and GABAergic projection neurons. GABAergic cell bodies receive few terminals which is in contrast to non-GABAergic somata, which receive many synaptic contacts. GABAergic dendrites that originate from GABAergic cell bodies, however, receive numerous terminals, both GABAergic and nonGABAergic. It can therefore be concluded that somatic inhibition is not present on GABAergic neurons, but does occur on nonGABAergic neurons. Furthermore, dendrites traverse large parts of the NOT/DTN forming a complex network that enables sampling and integration from a wide area. The projection to the IO is not GABAergic itself, but cells projecting to the IO receive a substantial GABAergic input, that probably originates in part from the MTN. Further investigation on the distribution of this input over a completely identified neuron would provide the quantitative data that are required to verify the above mentioned hypothesis. A GABAergic projection that originates in the pretectal nuclei is directed towards the superficial layers of the SC in the cat (Appell and Behan, 1990) and rat (Van der Want et al., 1991). A second GABAergic projection derives from the pretectum and reaches the LGN (Cucchiaro et al., 1991). Whether this projection originates from the same GABAergic cell bodies that project to the SC and the LGN or is derived from different populations remains to be determined. The ultrastructural studies of the NOT/DTN complex have shown that GABAergic terminals with different morphological characteristics are present and that the GABA positive F and P terminals are widely distributed over somata and the adjacent neuropil. The P terminals probably originate from dendrites of GABAergic interneurons while the F types originate from GABAergic projection and interneurons (Van der Want and Nunes Cardozo, 1988). One of these sources is located in the MTN differ from the intrinsic GABAergic terminals with respect to their relation to R terminals. GABAergic MTN terminals were never observed to receive R terminal input. This is in contrast with other GABAergic terminals which frequently do receive direct contact from R terminals. Within glomeruli triadic arrangements, formed by a single retinal terminal, a dendritic profile and second axonal profile dendritic profile and second axonal profile synapsing with the dendrite, were frequently encountered in the OPN (Campbell and Lieberman, 1985), but only occasionally in the NOT/DTN (Nunes Cardozo and Van der Want, 1987).(ABSTRACT TRUNCATED AT 400 WORDS)

A quantitative study of synaptic contacts on interneurons and relay cells of the cat lateral geniculate nucleus

The relative proportions of synapses made by retinal and extraretinal terminals on interneurons and relay cells in lamina A of the dorsal lateral geniculate nucleus (LGN) of the cat were estimated quantitatively in a sample of 4003 synapses. Processes of interneurons or relay cells were identified by presence or absence of GABA immunoreactivity, respectively, in thin sections treated with post-embedding anti-GABA immunogold. On the basis of ultrastructural features, synaptic terminals were interpreted as belonging to retinal axons, cortical axons or axon collaterals of relay cells. GABAergic terminals were positively identified by being immunoreactive. GABA(-) terminals with heterogeneous and poorly defined characteristics, which could not be identified in the above classes, were grouped together in an "undetermined" category. Among the total synaptic inputs to interneurons, the following relative percentages of synapses from different terminals were obtained: retinal 25%, cortical 37%, GABAergic 26%, axon collaterals 2%, undetermined 6%. The vast majority of retinal terminals synapse on dendritic appendages of interneurons rather than on their dendritic trunks (about 20:1). By contrast, the majority of cortical terminals synapse on dendrites rather than on dendritic appendages (about 5:1). Virtually all axon-collaterals synapses were established on dendritic appendages. 17% of the dendritic profiles of interneurons contain synaptic vesicles; many of these profiles were seen in postsynaptic relation to cortical axons and in presynaptic relation with relay dendrites. Given the extensive electrotonic lengths of these cells observed by others, and the expected high electric resistance of the slender stalks that are known to connect the dendritic appendages to interneurons, these results suggest that microcircuits involving the interneuronal dendritic appendages with dendrites of relay cells are under predominantly retinal control. The microcircuits established by presynaptic dendritic trunks with relay dendrites, are under predominantly cortical control. The axonal (spiking) output of interneurons would be under control of the few retinal synapses on proximal dendrites of these cells. Among the total synaptic inputs to relay cells, the following relative percentages of different synapses were obtained: retinal 12%, cortical 58%, GABAergic 24%, axon collaterals 0.3%, undetermined 5%. Relay cells receive twice the number of cortical synapses than interneurons, suggesting that direct cortical excitatory influences on relay cells are more preponderant than cortico-interneuron mediated inhibition on these cells. The observed proportions of dendritic profiles of relay cells and interneurons (80% and 20%, respectively) in the geniculate neuropil are similar to the known proportions of somata of both types of cells in the A-laminae.(ABSTRACT TRUNCATED AT 400 WORDS)

Glutamate-like immunoreactivity in retinal terminals in the nucleus of the optic tract in rabbits

The ultrastructural organization of retinal terminals within the nucleus of the optic tract of rabbits was investigated with a combination of anterograde tracing and immunocytochemistry. The anterogradely transported WGA-HRP injected in the vitreous of the eye was visualized with the sensitive gold-substituted silver peroxidase (GSSP) method. Glutamate and GABA immunoreactivity were identified with postembedding colloidal gold particles. Retinal ganglion cell terminals (R-terminals) in the nucleus of the optic tract formed asymmetric synapses and contained spherical vesicles and electron lucent mitochondria. R-terminals were observed in large clusters in the neuropil and in synaptic contact with large initial dendrites and somata. Within the clusters of neuropil the R-terminals formed two types of glomeruluslike arrangements: (1) an R-terminal centrally located and surrounded by small dendritic and axonal profiles and (2) several R-terminals surrounding a single dendrite or a group of dendritic profiles, presumably of interneuronal origin. All R-terminals identified with WGA-HRP as well as those exhibiting similar ultrastructural characteristics showed high levels of glutamate immunoreactivity, but no GABA immunoreactivity. The presence of glutamate and the absence of GABA in R-terminals suggest that glutamate is involved in neurotransmission in the pathway from retina to the nucleus of the optic tract of rabbits.

Active zone organization and vesicle content scale with bouton size at a vertebrate central synapse

A common observation in studies of neuronal structure is that axons differ in the size of their synaptic boutons. The significance of this size variation is unclear, in part because we do not know how the size of synaptic boutons is related to their internal organization. The present study has addressed this issue by using three-dimensional reconstruction of serial thin sections to examine the ultrastructure of synaptic boutons that vary in size. Our observations are based on complete or near-complete reconstructions of 53 synaptic boutons contacting large neurons in the ventromedial gray matter of the upper cervical spinal cord (probable neck motor neurons). We characterized bouton size in terms of volume and total area of membrane apposed to the motor neuron surface (apposition area). Boutons vary in apposition area by a factor of 40, and there is a significant positive correlation between our two measures of bouton size. In addition, bouton size is systematically related to four ultrastructural variables: 1) total active zone area, 2) number of active zones, 3) individual active zone area, and 4) number of synaptic vesicles. The correlations between these variables and both of our measures of bouton size are positive and significant. These data suggest that bouton size may be an index of ultrastructural features that are thought to influence transmitter storage and release.

Fine structure and synaptic architecture of HRP-labelled primary afferent terminations in lamina IIi of the rat dorsal horn

The fine structure and synaptic architecture of the afferent terminations in dorsal horn lamina II are studied using a combined light and electron microscopic procedure after anterograde labelling with horseradish peroxidase. Vibratome parasagittal sections, stained with heavy metal intensified diaminobenzidine after tracer application to the dorsal roots, were flat-embedded in Epon. The five types of labelled terminal arbors occurring in lamina IIi (Cruz et al., '87: J. Comp. Neurol. 261:221-236) were drawn and relocated in 5-microns sections cut serially from the thick sections. Ultrathin sections were then cut from the 5-microns sections so that the terminal fibers and swellings observed in the light microscope could be traced in the electron microscope. The flame-shaped arbors arose from fine myelinated stem fibers. Terminal strands generated large oval central terminals of type II synaptic glomeruli (CII), which established frequent axoaxonal contacts. Similar terminals have been labelled in the cat after tracer injections into hair-follicle fibers (Réthelyi et al., '82: J. Comp. Neurol. 207:381-393). The other four plexuses arose from unmyelinated stem fibers. The swarms of ultrafine boutons consisted of extremely thin terminal fibers generating very small, round, or polygonal glomerular terminals containing tightly packed agranular synaptic vesicles of variable size and one mitochondrion at best. The terminal strands of the bouquet plexus bore long and scalloped central varicosities of type I synaptic glomeruli (CI) with pleomorphic agranular vesicles and a relative abundance of dendroaxonal contacts. These features, together with the location in dorsal lamina IIi, suggest their belonging to the fluoride resistant acid phosphatase (FRAP)-reactive population. The boutons of the undulating fibers and those of the lateral plexus were, like those of the bouquets, scalloped and elongated rostrocaudally (CI), but contained a few large granular vesicles. The occurrence of the swarm, undulating, and lateral plexuses in ventral lamina IIi, which seems to lack FRAP or peptidergic terminals, suggests an origin from other, still unidentified neurochemical populations of fine primary afferents.

Synaptic circuitry identified by intracellular labeling with horseradish peroxidase

During the past two decades new techniques have been developed to directly test the dogma that neuronal structure is correlated with neuronal function. In the earliest experiments, Procion yellow was injected into neurons after they had been characterized physiologically; these neurons were then viewed through the light microscope. Recent advances in the method generally employ horseradish peroxidase as the dye which is injected since it diffuses quite readily throughout the injected neuron and produces a stable reaction product for both light and electron microscopic studies. This review explores the utility of examining synaptic circuitry after physiologically recording from axons or neurons and then injecting horseradish peroxidase into them. As a model system, we studied the cat lateral geniculate nucleus and investigated, at the electron microscopic level, the synaptic contribution to this nucleus from retinogeniculate axons, from interneurons, and from the axon collaterals of neurons that project to visual cortex.

Quantitative immunogold analysis reveals high glutamate levels in synaptic terminals of retino-geniculate, cortico-geniculate, and geniculo-cortical axons in the cat

A postembedding immunogold procedure was used to estimate quantitatively, at the electron-microscopical level, the intensity of glutamate (GLU) immunoreactivity in different identifiable profiles of the lateral geniculate nucleus (LGN) and perigeniculate nucleus (PGN) of the cat. Synaptic terminals of retinal and cortical origins in the LGN, and of axon collaterals of geniculo-cortical relay cells in the PGN, were identified by previously determined ultrastructural features. Processes of interneurons or relay cells were identified by being immunoreactive or non-immunoreactive, respectively, in serial thin section reacted with a GABA antibody. The results showed that synaptic terminals of geniculo-cortical relay cells in the PGN have significantly higher levels of GLU immunoreactivity than their parent somata or dendrites in the LGN; this suggests transmitter storage of this amino acid in these terminals. By contrast, synaptic terminals of interneurons did not show enrichment of GLU relative to their parent somata. This argues against the possibility that the relative enrichment of GLU in relay cells terminals is due to factors other than presynaptic storage. In addition, axon collateral terminals of relay cells in the pGN, as well as retinal and cortical terminals in the LGN, showed significantly higher GLU immunoreactivity than GABAergic terminals. These immunocytochemical results suggest that GLU is a neurotransmitter in the retino-geniculate, cortico-geniculate, and geniculo-cortical pathways in the cat.

Organization of cholinergic synapses in the cat's dorsal lateral geniculate and perigeniculate nuclei

In the preceding article, we showed that cholinergic fibers originating from the brainstem reticular formation provide a dense innervation of the lateral geniculate nucleus. In this report we describe the ultrastructure of these fibers and their relations with other elements in the neuropil of the lateral geniculate nucleus. Cholinergic fibers were labeled with an antibody to choline acetyltransferase (ChAT), the synthesizing enzyme for acetylcholine (ACh). In the A-laminae of the lateral geniculate nucleus, ChAT + profiles are small and contain tightly packed, mostly round vesicles. Some end in encapsulated synaptic zones where they form asymmetrical synaptic contacts with processes of both projection cells and interneurons. Others form synapses upon the shafts of dendrites. Of the four classical types of vesicle-containing profiles identified by Guillery (Z. Zellforsch. Mikrosk. 96:1-38, '69; Vision Res. [Suppl.] 3:211-227, '71), ChAT + profiles most closely resemble RSD profiles (Round vesicles, Small profile, Dark mitochondria). However, as a population, ChAT + profiles can be distinguished from the unlabeled population of RSD profiles because they are larger in size, contain more mitochondria, and make synapses with smaller postsynaptic membrane specializations. Each of these differences is statistically significant and together they indicate that ChAT + profiles are a distinct morphological type of synaptic profile. ChAT + profiles in the perigeniculate nucleus resemble those found in the lateral geniculate nucleus; they also make synapses with obvious postsynaptic thickenings.

Quantitative immunogold analysis reveals high glutamate levels in retinal and cortical synaptic terminals in the lateral geniculate nucleus of the macaque

An immunogold procedure has been used on ultrathin sections of the parvo- and magnocellular layers of the dorsal lateral geniculate of the rhesus monkey to estimate quantitatively at the electron microscopic level the intensity of immunoreactivity to an antibody against glutamate over profiles of retinal, cortical, GABAergic synaptic terminals and glial cells. GABAergic terminals were identified directly by immunogold reactivity to a GABA antibody or by ultrastructural features. The results showed that in both of the main subdivisions of the geniculate the densities of immunogold particles over cortical and retinal terminals were about two- to three-fold higher than those over GABAergic terminals or glial profiles. In addition, cortical and retinal terminals showed higher positive correlations of glutamate immunogold particle densities to synaptic vesicle densities than did GABAergic terminals. These differences suggest higher and lower concentrations of glutamate corresponding to transmitter and metabolic pools of this amino acid in axon terminals of retinal and cortical origins versus GABAergic terminals, respectively, in the dorsal lateral geniculate nucleus of the macaque.

Chemical synapses in the mammalian central nervous system: an introduction

This paper provides a brief review of the historical development of our understanding of synaptic structure in the central nervous system. The basic structure of the synapse is reviewed as well as the development of ultrastructural techniques that have allowed the details of synaptic organization to be elucidated.

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

These experiments examine which morphological features of axon terminals and their synaptic glomeruli are determined by afferent axons, and which by their targets. In normal, adult hamsters, electron microscopy reveals that, with respect to multiple ultrastructural features, the terminals and synaptic glomeruli of retinal afferent axons in the dorsal lateral geniculate nucleus differ from those of ascending auditory and somatosensory afferents in the medial geniculate and ventrobasal nuclei, respectively. These features include: (1) the location of specific sensory axon terminals on the somata and dendrites of their targets neurons, (2) the constitutents of the glomeruli and their synaptic relationships, (3) the number of specific sensory terminal boutons per glomerulus, (4) bouton size, (5) the number of dendritic and somatic appendages contacted by each bouton, and (6) the mitochondrial morphology of the specific sensory afferent boutons. In order to ascertain which of these features are determined by afferent axons and which by their targets, we subjected newborn Syrian hamsters to surgical procedures known to produce permanent, abnormal retinal projections to the main thalamic auditory (medial geniculate) and somatosensory (ventrobasal) nuclei. When the animals were adults, we examined the terminals and synaptic glomeruli of abnormal retino-auditory and retino-somatosensory axons that were anterogradely labeled by intraocular injection of horseradish peroxidase. With respect to all of the preceding features except mitochondrial morphology, the terminals and synaptic glomeruli of retino-medial geniculate and retino-ventrobasal axons more nearly resembled those of normal, auditory and somatosensory afferent axons, respectively, than they did those of normal, retino-lateral geniculate axons. These results demonstrate that the differentiation of all the features that we have examined, except mitochondrial morphology, is determined by factors in target neurons or their environment. This finding suggests that the differentiation of morphological features involved in contacts among neurons (including the type, number and size of interconnected neuronal elements and the loci at which they contact each other) is responsive to interactions among the connected elements, or between neural elements and their environment (e.g., glia, extracellular matrix), whereas the differentiation of structures reflecting intrinsic functions of individual neuronal elements is not responsive to such interactions.

The beta sector of the rabbit's dorsal lateral geniculate nucleus

The beta sector of the rabbit's dorsal lateral geniculate nucleus is a small region of nerve cells scattered among the fibres of the geniculocortical pathway. In its topographical relations it resembles the perigeniculate nucleus of carnivores, which contains neurons driven by geniculate and visual cortical neurons and which sends inhibitory fibres back into the geniculate relay. We have traced retinogeniculate, geniculocortical and corticogeniculate pathways in rabbits by using horseradish peroxidase or radioactively labelled proline and have found that the beta sector resembles the perigeniculate nucleus in receiving no direct retinal afferents, sending no efferents to the visual cortex (V-I), and receiving afferents from the visual cortex. The corticogeniculate afferents are organized so that the visual field map in the beta sector and the main part of the lateral geniculate relays are aligned, as are the maps in the cat's perigeniculate nucleus and the main part of the geniculate relay of carnivores. Electron microscopical studies show similar types of axon terminals in the rabbit and the cat for the main part of the geniculate relay on the one hand and for the beta sector and the perigeniculate nucleus on the other. Earlier observations that the proportion of putative inhibitory terminals (F-type terminals) is lower in the rabbit's than the cat's geniculate region are confirmed. A major difference between the beta sector and the perigeniculate nucleus has been revealed by immunohistochemical staining for GABA. Whereas almost all of the cat's perigeniculate cells appear to be GABAergic, the proportion in the beta sector is much lower, and not significantly different from that found in the main part of the rabbit's geniculate relay. It is concluded that the beta sector shares many of the organizational features of the perigeniculate nucleus. A common developmental origin seems probable, but the functional differences remain to be explored.

Development of corticogeniculate synapses in the cat

The development of corticogeniculate synapses was studied in 16 cats ranging in age from newborn to adult. Tritiated proline was injected into areas 17 and 18 of the visual cortex in order to label corticogeniculate terminals in lamina A of the dorsal lateral geniculate nucleus. The labeled terminals were then characterized ultrastructurally using electron microscopic autoradiography. Labeled synaptic profiles were found in newborn kittens, indicating that corticogeniculate connections are present in the cat at birth. Morphologically, however, many corticogeniculate endings in newborn and 1-week-old kittens are different from those in older animals in that they do not form well-defined terminal boutons, and their synaptic vesicles are often loosely packed. In kittens 2 weeks of age and older, corticogeniculate axons end as RSD terminals exclusively; i.e., they are relatively small in size and contain round, densely packed synaptic vesicles, and occasionally an electron-dense mitochondrion (Guillery: Z. Zellforsch. 99: 1-38, '69). However, not all RSD terminals in the LGN represent input from visual cortex. Injections of 3H-proline into the mesencephalic reticular formation also label RSD terminals selectively in the lateral geniculate nucleus. At all ages corticogeniculate axons make synaptic contacts with dendrites exclusively, and they are always presynaptic. This suggests that the essential pattern of corticogeniculate synapses is formed early and is not altered during subsequent development. Quantitatively, there is no significant change in the size of corticogeniculate terminals or their synaptic vesicles in kittens 2 weeks of age (the youngest measured) and older. In contrast, the synaptic contact lengths of these terminals decreases about 28% between 2 and 12 weeks. During this same period there is approximately a twofold increase in the density of corticogeniculate terminals in the neuropil of lamina A. Since the volume of neuropil in lamina A increases almost fourfold between 2 and 12 weeks, the doubling of corticogeniculate terminal density represents about an eightfold increase in terminal number. After 12 weeks there is little change in the length, density, or number of corticogeniculate synaptic contacts, which suggests that the morphological development of the corticogeniculate pathway is essentially complete by this age.

Target-controlled differentiation of axon terminals and synaptic organization

These experiments investigate the processes regulating the morphological differentiation of synaptic connections. Electron microscopy showed that the terminal boutons and synaptic complexes of retinal afferent axons in the main thalamic visual nucleus, the dorsal lateral geniculate nucleus, differ in their morphology from those of ascending afferent axons in the main thalamic somatosensory (ventrobasal) nucleus. Developing retinal ganglion cell axons in hamsters were made to project permanently to the ventrobasal nucleus, rather than to the lateral geniculate nucleus. With respect to most of the ultrastructural features examined, the terminals and synaptic complexes of mature, anterogradely labeled retino-ventrobasal axons more closely resembled those of normal somatosensory afferents to the ventrobasal nucleus than they did those of normal retinofugal axons within the lateral geniculate nucleus. These results suggest that the ultrastructural differentiation of axon terminals and synaptic complexes is regulated largely by the target environment, although some features appear to be intrinsic to the afferent axons themselves.

A quantitative study of anterograde and retrograde axonal transport of exogenous proteins in olfactory nerve C-fibers

The pike olfactory nerve which consists of a homogeneous population of C-fibers of 0.25 micron diameter or less was used to study quantitatively both anterograde and retrograde axoplasmic transport of wheat germ agglutinin and horseradish peroxidase. It was found that even in these extremely thin axons anterograde and retrograde transport takes place. Activity distribution profiles (transport profiles) for retrograde transport were established and found to be similar to the typical profiles of anterograde transport as they consisted of a small rapidly moving peak and a saddle region followed by the bulk of the material which moved more slowly. Horseradish peroxidase activity profiles were obtained both after injection into the synaptic region and after injection into the perikaryal region. From these transport profiles a maximal velocity of 25 mm/d (19 degrees C) for the leading peak and of about 7 mm/d for the slower component could be determined. There is no significant difference between the velocities for anterograde and retrograde transport. In the case of wheat germ agglutinin, only injection into the synaptic region resulted in typical transport profiles (retrograde transport) with a peak and saddle region. The maximum velocities of retrograde transport were about the same as for horseradish peroxidase [26 mm/d and 7 mm/d (19 degrees C)]. The electron microscopic analysis of horseradish peroxidase revealed that after injection into the olfactory bulb it was taken up into the neurons where it was found mainly in multivesicular bodies (0.5 micron diameter). In longitudinal sections of the nerve similar but slightly more elongated organelles (diameter 0.25 micron, length 0.4 micron) were found in those segments in which the slowly moving bulk of the peroxidase activity was located. The number of these organelles decreased with distance from the site of injection. The horseradish peroxidase transported within the leading peak could not be assigned to specific structures although several electron microscopic-histochemical methods were applied. It was concluded that anterograde and retrograde transport occur simultaneously in these axons, and that, therefore, even the large organelles, each of which almost fills the axon, must be able to pass each other. This would necessitate that the axons are able to transiently enlarge their diameter considerably.

Several morphological types of terminal arborizations of primary afferents in laminae I-II of the rat spinal cord, as shown after HRP labeling and Golgi impregnation

The morphology of the terminal arborizations in laminae I-II of primary afferent fibers was studied in sections stained by the heavy metal (nickel and cobalt) intensification of diaminobenzidine (DAB) after crushing one dorsal root with horseradish peroxidase (HRP) crystals, and with the mixed Golgi method which duplicated the staining provided by the first method. Besides the flame-shaped arbors located in deep lamina IIi as an extension of the arbors of lamina III, which were derived from 1.7-micron thick stem fibers (probably A alpha beta fibers), six types of terminal arbors, all rostrocaudally oriented, arising from fine stem fibers and having preferential locations, were disclosed. The lateral third of laminae I-II contained a longitudinal plexus of parallel 0.8-micron thick stem fibers (C fibers) with longitudinal side branches generating many boutons en passant. Laminae I and IIo, in their middle third, contained dichotomizing longitudinal fibers with elongated boutons, arising from 1-micron thick stem fibers (C or A delta), and, in the medial third, a dense plexus with terminal networks carrying large boutons, which arose from 1.3-micron thick stem fibers (A delta). Fibers ending in terminal bouquets and issuing from 1-micron thick stem fibers (C or A delta) occupied the dorsal part of middle and medial lamina IIi, while the intermediate part contained clusters (swarms) of ultrafine boutons arising from extremely fine fibers. The whole medial lamina IIi also contained fine undulating fibers arising from 0.3 micron-thick stem fibers (C fibers) with large boutons near their ends. The functional meaning of this multiplicity of morphological types and locations is still unclear. It may be clarified when single unit analysis of HRP-injected fine fibers is made possible, or immunocytochemical stainings disclose the neurotransmitters utilized by each fiber type.

Synaptic organization of the nucleus of the optic tract in the rabbit: a combined Golgi-electron microscopic study

The organization of the nucleus of the optic tract was investigated with light and electron microscopy in combination with Golgi impregnation. In Golgi material, neurons ranged in size from 10 to 25 microns with three to seven principal dendrites extending predominantly parallel to the fibres of the optic tract, irrespective of their location within the nucleus. In some areas dendrites extended into the neuropil of the adjacent dorsal terminal nucleus of the accessory optic system and the posterior pretectal nucleus. Occasionally spines and appendages were observed. The fine structure of the nuclei, perikarya and the dendritic arborization did not allow a well-defined distinction between interneurons and projection neurons. The synaptic organization of the nucleus of the optic tract showed great resemblance to the neuropil of the lateral geniculate nucleus and the superior colliculus. Similar types of presynaptic terminals were noticed: (i) R-terminals were either large and scalloped or small and regular in outline with spherical vesicles and electron-lucent mitochondria, and showed asymmetric contact zones; (ii) F-terminals with flattened vesicles, opaque mitochondria and symmetric contact zones; (iii) RLD-terminals with spherical vesicles and electron-dense mitochondria and asymmetric contact zones; (iv) P-terminals with pleomorphic vesicles and electron-lucent or opaque mitochondria and asymmetric synaptic thickenings. These different types of terminal were found isolated in the neuropil or in clusters of synapses. The most striking differences between the nucleus of the optic tract and the lateral geniculate nucleus were the relative scarcity of F-terminals in the clusters, the paucity of triadic arrangements and the relatively small size of the R-terminals. The differences in ultrastructure may be related to retinal W-type ganglion cells, which form the main retinal input to the nucleus of the optic tract and could also be related to the physiologically identified direction-selective units within the nucleus of the optic tract.

Synaptic circuits involving an individual retinogeniculate axon in the cat

In order to describe the circuitry of a single retinal X-cell axon in the lateral geniculate nucleus, we physiologically characterized such an axon in the optic tract and injected it intra-axonally with horseradish peroxidase. Subsequently, we recovered the axon and employed electron microscopic techniques to examine the distribution of synapses from 18% of its labeled terminals by reconstructing the unlabeled postsynaptic neurons through a series of 1,200 consecutive thin sections. We found remarkable selectivity for the axon's output, since only four of the 43 available neurons in a limited portion of the terminal arbor receive synapses from labeled terminals. Moreover, the distribution of these synapses on the four neurons, which we term cells 1 through 4, varies with respect to synapses from other classes of terminals that contact the same cells, including synapses from unlabeled retinal terminals. For cells 1 and 3, the labeled terminals provide 49% and 33%, respectively, of their retinal synapses, and these are located on both dendritic shafts and appendages. Synapses from the injected axon to these cells are thus integrated with those from other retinal axons. For cell 2, the labeled terminals provide 100% of its retinal synapses, but these synapses converge on clusters of dendritic appendages where they are integrated with convergent inhibitory inputs. Finally, for cell 4, the labeled terminals provide less than 2% of its retinal inputs, and these are distally located; we suggest that these synapses are remnants of physiologically inappropriate miswiring that occurs during development. The findings from this study support a concept of selectivity in neuronal circuitry in the mammalian central nervous system and also reveal some of the diverse integrative properties of neurons in the lateral geniculate nucleus.

Effect of light deprivation upon the morphology of axon terminals in the dorsal lateral geniculate nucleus of mouse: an electron microscopical study using serial sections

Two populations of morphologically different large axon terminals have been observed electron microscopically in the dorsal lateral geniculate nucleus of mice raised in complete darkness from birth up to 19 days of age. One population includes larger terminals indistinguishable from the large terminals present in control animals, i.e. they have round synaptic vesicles, rather pale mitochondria, membrane saculae, coated vesicles, and asymmetric contacts with encrusted dendritic spines of portions of dendrites of geniculo-cortical relay neurons. The other population includes large terminals which also have asymmetric contacts with encrusted dendritic spines or portions of dendrites of geniculo-cortical relay neurons, but they show darker mitochondria, absence of both membrane saculae and coated vesicles, and significantly higher synaptic vesicle density and smaller size than the large control ones. We suggest that the latter population of terminals could be inactive due to the absence of visual input.

Morphology of geniculocortical axons in turtles of the genera Pseudemys and Chrysemys

An analysis has been made of the morphology of axons in the geniculocortical pathway of turtles using the anterograde transport of horseradish peroxidase in both in vivo and in vitro preparations. Following injections of HRP into the dorsolateral thalamus, labeled axons could be traced from the dorsal lateral geniculate complex to the telencephalon. They are unbranched and free of varicosities within the diencephalon. They travel in the dorsal peduncle of the lateral forebrain bundle, through the basal telencephalon and dorsally into the pallial thickening. Many axons are situated deep in the pallial thickening and bear numerous varicosities that often appear apposed to the proximal dendrites or somata of neurons retrogradely labeled by thalamic injections of horseradish peroxidase. Individual axons continue from the pallial thickening into the dorsal cortex where they shift dorsally and bear varicosities as they course from lateral to medial in the superficial third of layer 1. These data indicate that the terminal zone of the dorsal lateral geniculate complex within the telencephalon of turtles is more extensive in the mediolateral direction than previously believed. Geniculate axons bear varicosities both within the pallial thickening as well as the dorsal cortex, but have different relationships to potential postsynaptic elements in the two areas. Geniculocortical axons overlie somata and proximal dendrites of neurons in the pallial thickening, but intersect the distal dendrites of neurons in the dorsal cortex.

Ultrastructural organisation of the projection from the superior colliculus to the ventral lateral geniculate nucleus of the rat

Retinorecipient regions of the ventral lateral geniculate nucleus of the thalamus and the superior colliculus of the midbrain are linked by reciprocal axonal projections. In this study we have investigated the ultrastructural characteristics, the distribution, and the postsynaptic targets of the terminals of axons projecting to the ventral lateral geniculate nucleus from the superior colliculus. Horseradish peroxidase was injected into the superior colliculi of adult albino rats, and the Hanker-Yates method was used to visualize anterogradely and retrogradely transported peroxidase in the ventral lateral geniculate nuclei 24 hours following the injection. Labelled terminals were found in the lateral and ventrolateral parts of the external division of the ipsilateral ventral lateral geniculate nucleus. The labelled terminals were confined to areas of simple, nonglomerular neuropil. They were 0.45-1.5 micron in diameter; contained small, dark mitochondria and spherical synaptic vesicles; and established Gray type I (asymmetrical) synaptic contacts with the dendritic shafts, dendritic spines, and occasionally cell bodies of cells with the ultrastructural characteristics of projection cells. A few labelled terminals established synaptic contact with retrogradely labelled cells. Thus, in the rat, the projection from the superior colliculus gives rise to a uniform population of axon terminals in the nonglomerular neuropil of the lateral portion of the ventral lateral geniculate nucleus, which synapse with, and are probably excitatory to, geniculocollicular and other projection cells.

Axon development in mouse cerebellum: embryonic axon forms and expression of synapsin I

A fundamental question in central nervous system development is the timing of synaptogenesis in relation to invasion of targets by afferent axons. A related question is how growth cones transform into synaptic terminals. These two aspects of axon maturation were examined in developing mouse cerebellum, by labeling single axons with horseradish peroxidase, to study their form and cytology, and by immunocytochemical staining of a synaptic vesicle antigen, synapsin I, a phosphoprotein found on synaptic vesicles in all mature CNS synapses. From embryonic day 16 to postnatal day 3, horseradish peroxidase-labeled afferent axons extend well into the cerebellum and have simple forms. At embryonic day 16, axon growing tips are synapsin I-negative. Synapsin I is first expressed at embryonic day 17, and by embryonic day 18, fibers are stained throughout the cerebellum. Synapsin I expression coincides with a general increase in synaptic specializations, although growing tips continue to have the cytology of growth cones. During the period that axons have primitive shapes, synapsin I is distributed throughout the terminal arbor, corresponding to the presence of small vesicles along neurite lengths, even at non-synaptic sites. After postnatal day 3, when synaptic terminals develop into stereotypic shapes and engage in characteristic synaptic relations, synapsin I is restricted to boutons. Thus, the synapse-specific protein synapsin I is expressed in fetal mouse brain, long before nerve endings have the structure and connections of adult brain. In cerebellar axons, the expression of this protein follows axon arrival, coincides with the appearance of elementary synapses, and accompanies the transformation of growing tips into stereotypic synaptic boutons. The time course of expression of synapsin I, a phosphoprotein that may be involved in synaptic efficacy, suggests that transmitter release may influence early axon-target cell interactions.

Ultrastructural analysis of synaptic relationships of intracellularly stained pyramidal cell axons in piriform cortex

Axons of pyramidal cells in piriform cortex stained by intracellular injection of horseradish peroxidase (HRP) have been analyzed by light and electron microscopy. Myelinated primary axons give rise to extensive, very fine caliber (0.2 micron) unmyelinated collaterals with stereotyped radiating branching patterns. Serial section electron microscopic analysis of the stained portions of the collateral systems (initial 1-2 mm) revealed that they give rise to synaptic contacts on dendritic spines and shafts. These synapses typically contain compact clusters of large, predominantly spherical synaptic vesicles subjacent to asymmetrical contacts with heavy postsynaptic densities. On the basis of comparisons with Golgi material and intracellularly stained dendrites, it was concluded that dendritic spines receiving synapses from the proximal portions of pyramidal cell axon collaterals originate primarily from pyramidal cell basal dendrites. Postsynaptic dendritic shafts contacted closely resemble dendrites of probable GABAergic neurons identified in antibody and [3H]-GABA uptake studies. Electron microscopic examination of pyramidal cell axon initial segments revealed a high density of symmetrical synaptic contacts on their surfaces. Synaptic vesicles in the presynaptic boutons were small and flattened. It is concluded that pyramidal cells synaptically interact over short distances with other pyramidal cells via basal dendrites and with deep nonpyramidal cells that probably include GABAergic cells mediating a feedback inhibition. This contrasts with long associational projections of pyramidal cells that terminate predominantly on apical dendrites of other pyramidal cells.

Postnatal changes in arborization patterns of murine retinocollicular axons

The growth and arborization of murine retinocollicular axons have been studied by means of HRP axon filling during postnatal development. Transformations in arborization patterns have been correlated with changes in synaptic density in the superficial collicular neuropil and with the formation of synapses by HRP-filled axons. At all postnatal ages axons of the optic projection are fasciculated and most follow a rostrocaudally aligned path. On the day of birth the axons course through both stratum griseum superficiale (SGS) and stratum opticum (SO); during the following 4 days the axon trunks disappear from SGS and are subsequently found only in SO. From postnatal day (P) 0 to P3, the majority continue far caudally in the colliculus, giving rise to small ascending collaterals at multiple points along their course. Ultimately, usually by P3, one or two collaterals begin to branch profusely and by P5 the majority of axons give rise to a focal terminal ascending arborization. The general configuration of most arborizations at P3 approximates that of the mature axon. However, the richness of terminal branching increases from P3 through the first 2 postnatal weeks. Synaptic density is relatively low in the first postnatal week, and no synapses involving HRP-filled optic axons were identified in this interval. Subsequently, after elaboration of definitive arbors has begun, synaptogenesis in the surrounding neuropil accelerates. Synaptic density in the upper SGS approximates adult values early in the third postnatal week. By this time synaptic junctions involving the terminal arborizations of optic axons are abundant.

Differential effect of visual deprivation on cytochrome oxidase levels in major cell classes of the cat LGN

Cytochrome oxidase histochemistry was used to examine the effects of visual deprivation on the development of neurons in the lateral geniculate nucleus of the kitten. Early postnatal monocular suture results in a decrease in reactivity within the neuropil of visually deprived binocular laminae A, A1, magnocellular C, and medial interlaminar nucleus. Within these regions, monocular suture has a greater effect on the relative numbers of, and the growth of darkly reactive (normally large), presumed Y-cells than on other less reactive geniculate neuronal classes. The decreases in the reactivity of the neuropil may be attributed to the decreases in the number of mitochondria, the number of darkly reactive mitochondria, and/or the number of darkly reactive mitochondria localized within dendrites. Although all classes of dendrites appear to be adversely affected, the decrease in C.O. reactivity was most dramatic within the presumed proximal dendrites of class 1 Y-cells. These dendrites were identified by the type of synaptic contacts they formed with retinal terminals (Rapisardi and Miles, '84, J. Comp. Neurol. 223:515-534; Wilson et al., '84, Proc. R. Soc. Lond. [Biol.] 221:411-436). As with Y-cells, the effects of monocular suture on the large darkly reactive cells were not as dramatic at sites where binocular interactions were either absent or where they had been experimentally eliminated. Based on the present and previously reported findings from several laboratories, it is likely that the selective physiological and morphological effects of monocular suture on Y-cells are accompanied by metabolic deficits involving both dendrites and perikarya. These effects appear to be due more to binocular interactions than to visual deprivation per se.