THE JOURNAL OF COMPARATIVE NEUROLOGY 292:163-177 (1990)

Disynaptic Olfactory Input to the Hippocampus Mediated by Stellate Cells in the Entorhinal Cortex WALTER K. SCHWERDTFEGER, EBERHARD H. BUHL, AND PETER GERMROTH Max-Planck-Institut fur Hirnforschung, Deutschordenstraae 46,6000 FrankfudM.71, Federal Republic of Germany

ABSTRACT Electrophysiological and anatomical studies indicate functional relationships between the olfactory bulb and the hippocampus, mediated by the lateral olfactory tract and perforant path. Fibres from the lateral olfactory bulb terminate in the molecular layer of the lateral entorhinal cortex, which contains stellate and pyramidal cells that project to the hippocampus. Therefore this study was performed to analyze whether a trineuronal, disynaptic chain connects the olfactory bulb and the hippocampus. In adult rats, Fast Blue was unilaterally injected into the septa1 hippocampus to label cells of origin of the entorhinohippocampal pathway. Lesions of the ipsilateral olfactory bulb induced anterograde terminal degeneration in the entorhinal cortex of the same animals. Fast Blue labelled, and thus hippocampally projecting entorhinal neurones in fixed vibratome slices of the operated brains were injected with Lucifer Yellow. Most of these neurones were stellate layer I1 and pyramidal layer 111 cells; in addition there were some sparsely spinous multipolar cells in layers I1 and 111 and sparsely spinous horizontal cells at the layer 1/11 border. Injected cells were photoconverted and processed for electron microscopy. Olfactory bulb lesions resulted in electron-dense degeneration of abundant terminal boutons in the outer zone of entorhinal layer 1. The relative frequency of degenerating boutons decreased towards deeper zones of the layer. In the outer zone, degenerated terminals predominantly contacted dendritic spines. These contacts could be seen on injected stellate cells but not on pyramidal cells. This study shows that the area dentata of the rat is reached by disynaptic afferent input from the olfactory bulb and thus is likely to process olfactory information. Oligosynaptic pathways might provide the hippocampus also with visual and auditory inputs; such fast transmitted polysensory information could be essential for the proposed participation of the hippocampus in attention-related mechanisms. Key words: anterogradedegeneration, retrograde transport, Fast Blue, intracellular injection, Lucifer Yellow, electron microscopy

Of all mammalian brain structures, the function of the hippocampus has probably aroused the most controversy. It has been related to memory and learning (Milner, '59), anxiety (Gray, '77), and spatial (O'Keefe and Nadel, '78) and temporal (Solomon, '79) orientation. It has also been proposed that the hippocampus may participate in attention (Douglas, '75; Foreman and Stevens, '87)-in the sense of focusing central nervous activity onto relevant external and internal stimuli, perhaps after comparing stored with new information (Vinogradova, '75). There is general agreement that the hippocampus processes polysensory information in

0 1990 WILEY-LISS, LNC.

the fulfillment of its functions (e.g., Berger et al., '80; Winocur, '80; Deadwyler et al., '87; Eichenbaum and Cohen, '88). Since a major function of attention-related mechanisms is to provoke fast reactions to relevant stimuli, it is reasonable to assume that if the hippocampus is involved in these mechanisms, sensory input could reach it by both diffuse, polysynaptic channels, and fast, oligosynaptic pathways. No direct pathways from primary sensory structures have been Accepted August 18, 1989.

W.K. SCHWERDTFEGER ET AL.

164

Figures 1 and 2

DISYNAPTIC OLFACTORY PATHWAY TO THE HIPPOCAMPUS

165

Fig. 3. Somata of retrogradely labelled neurones in the entorhinal cortex after Fast Blue injection into the hippocampus. Note that layer I1 neurones are intensely labelled. White bars indicate the pial surface and the borders between layers I, I1 and 111. x 100.

reported so far, whereas there are indications of two possible disynaptic routes: 1)the retina projects to the anterior dorsal (AD) and lateral dorsal thalamic (LD) nuclei (Itaya et al., '86; Reperant et al., '87), which in turn send axons to the hippocampus (Wyss et al., '79; Schwerdtfeger, '84); and 2) mitral cells in the olfactory bulb project via the lateral olfactory tract to the lateral entorhinal cortex (Haberly and Macrides and Schneider '82) where the fibres terPrice, '77; minate on apical dendrites in layer I, which belong to stellate and pyramidal cells of layers I1 and 111 (Wouterlood and Nederlof '83) and on nonpyramidal, GABAergic cells in layer I (Wouterlood et al., '85); earlier studies have shown

Fig. 1. Rat brain in dorsal view. Left arrow: site of Fast Blue injection aimed at septa1 hippocampus; right arrow: location of olfactory bulb incision. x 10.

Fig. 2. Frontal vibratome section a t Fast Blue injection site (brightfield illumination). The injection predominantly affected the area dentata (AD), and to a lesser extent Ammon's horn (AH). x30. Inset shows fluorescence micrograph of the same section. x 15.

that axons of layer I1 stellate cells and layer IT1 pyramidal cells terminate in the molecular layers of the area dentata and field CA1 of Arnrnon's horn, respectively (e.g., Steward and Scoville, '76; Witter and Groenewegen, '84). These axons constitute the perforant path, an excitatory projection (Doller and Weight, '82; Robinson and Racine, '82) that uses glutamate as a putative neurotransmitter (Nadler and Smith, '81; Dolphin et al., '82). There is no firm anatomical evidence for oligosynaptic relations between primary sensory structures and the hippocampus. Layers I1 and 111 neurones in the entorhinal cortex (Witter and Groenewegen, '86a,b) and AD and LD neurones (Sripanidkulchai and Wyss, '86) project to a variety of cortical and subcortical targets. Therefore the recipient cells of sensory information may transmit to nonhippocampal targets, and the hippocampally projecting neurones may receive other than sensory input. Since the pathways both from the olfactory bulb to the entorhinal cortex and from the entorhinal cortex to the hippocampus are fairly prominent, they were studied here to determine whether they actually form the substrate of a trineuronal chain. For this purpose, in adult rats lesions of the olfactory bulb were followed by hippocampal injection of the fluorescent tracer Fast Blue. Retrogradely labelled somata in the entorhinal cortex were intracellularly

166

W.K. SCHWERDTFEGER ET AL.

i

C

Fig. 4. Horizontal and multipolar entorhinal cells that project to the hippocampus. The horizontal cells (aand b) are located at the outer margin of layer I1 whereas multipolar cells are located in layers 11 (cell c ) and I11 (cell d). Bar = 100 pm.

injected with the fluorescent dye Lucifer Yellow, which was then transformed into an osmiophilic polymer by a simple photoconversion procedure, allowing ultrastructural examination of the injected neurones. Thus, anterogradely degenerated olfactory terminals could be found in synaptic contact with identified neurones projecting to the hippocampus.

MA'IERIAIS AND METHODS In eight adult rats, anaesthetized with an intraperitoneal injection of 4 % aqueous chloral hydrate (1 m1/100 g body weight), the left olfactory bulb was transversely cut a t an intermediate rostrocaudal level. Then, 0.25 pl of the fluorescent tracer Fast Blue (FB; Bentivoglio et al., '80; 5%#dis-

167

DISYNAPTIC OLFACTORY PATHWAY TO THE HIPPOCAMPUS

5a

Fig. 5. Pyramidal (a,b)and spiny stellate cells (c,d) of the entorhinal cortex that project to the hippocampus. Note the bifurcating apical dendrites of the pyramidal cells. Bar 100 pm. ~

I68

Fig. 6. Electron micrograph showing various degenerating boutons in the entorhinal cortex after lesions of the olfactory bulb. A: Outer zone of molecular layer. Arrows point at degenerating houtons in synaptic contact with dendritic elements. Open arrow (bottom right) points a t degenerating fiber enclosed by myelin sheath. White asterisks indicate

W.K. SCHWERDTFEGER ET AL.

more degenerating houtons. Bar = 2 Mm. B Detail of A. Arrows point at synaptic contacts between degenerating boutons and dendritic spines. C: Degenerating bouton in synaptic contact (arrows) with a dendritic shaft (open star). Bars in B and C = 1&m.

DISYNAPTIC OLFACTORY PATHWAY TO THE HIPPOCAMPUS

169

Fig. 7. Degenerating olfactory boutons in layer I1 of the entorhinal cortex. A: Degenerating boutons (arrows) close to proximal dendritic segment of an injected stellate cell. Bar = 1fim. B: Lower power micrograph showing the position of A (framed area). Bar = 5 fim.

solved in distilled water) were stereotaxically injected (caordinates after Paxinos and Watson, '82) into the presumed hilar region of the left hippocampus (Figs. 1, 2). After a 3 day survival period the animals were transcardially per-

fused with 300 ml physiological saline followed by 500 ml fixative containing 4 % paraformaldehyde and 0.5% glutaraldehyde in 0.1 M phosphate buffer (PB; pH 7.4). The brains were immediately removed and serially cut a t 100 Frn

W.K. SCHWERDTFEGER ET AL.

170

Figure 8

171

DISYNAPTIC OLFACTORY PATHWAY TO THE HIPPOCAMPUS in the frontal plane. For subsequent use slices were stored in PB at 4OC. To determine the extent of the injection site, corresponding sections were mounted on slides and coverslipped with glycerol. The preparations were then examined under epifluorescence and photographically documented on high-speed black and white films. Micropipettes were pulled from omega dot glass capillaries (1.0 mm 0.d. x 0.86 mm id.). These electrodes were filled with a 3 4 % aqueous solution of Lucifer Yellow (LY; Stewart, '78, %1), and a platinum-iridium or silver wire was inserted. The resistance usually ranged between 90 and 300 MR in PB. Fixed slices were transferred in PB, floated on a slide, and immobilized in a Petri dish by means of a fenestrated millipore filter. For dye injection this slice chamber was transferred to the stage of a Zeiss ACM fixed stage microscope equipped with epifluorescence illumination and long-distance objectives, which had working distances ranging between 8 and 10 mm (Zeiss filter combination for LY and Fast Blue: BP 400-440, FT 460, L P 470; Zeiss objectives UD 20/0.57, UD 40/0.65). The electrodes were attached to a motor-driven high-precision micromanipulator, which could be moved in three axes. Then the LY pipettes were advanced towards retrogradely labelled and therefore fluorescing neurones, which were impaled under visual control whereby one can directly observe the entrance of LY into the injected FB-labelled cell. A successful penetration was monitored by applying a short negative current pulse, which led to rapid intracellular diffusion of LY. Then iontophoretic dye injection was continued for 5-15 minutes (1-3 nA negative constant current) until all fine dendritic branches appeared brightly fluorescent. For a more detailed description of the injection procedures see Tauchi and Masland ('84) and Schwerdtfeger and Buhl ('86). After one cell had been extensively stained the slice was immersed for 5 minutes in PB containing 1mg diaminobenzidine (DAB)/ml. Following preincubation, the filled neurone was irradiated in the same solution with the LY excitation wave length for 20-25 minutes until all visible fluorescence had faded (Maranto, '82). Photo-oxidation of LY resulted in the intracellular formation of a brown reaction product, which was homogeneously distributed throughout the cell's dendritic arbor. After three rinses in PB the slice was osmicated in 0.5% osmium tetroxide in PB for 10 minutes. Following three further washes in PB the sections were dehydrated and flat-embedded in Durcupan (Fluka, Switzerland) between two acetate sheets. Prior to ultrathin sectioning, photo-converted neurones were serially photographed and drawn with a camera lucida. Then filled cells were trimmed, mounted on blocks, and cut serially with an ultramicrotome. Sections were contrasted with lead citrate on Formvar-coated single-slot grids. Finally, the preparations were analysed with a Zeiss EM10 electron microscope for the presence of electron-dense DAB precipitate and anterogradely degenerating boutons. For further method-

Fig. 8. Low-power EM photo-reconstruction of injected spiny stellate cell with assemblage of one apical dendrite that can be traced (arrows) without interruption up to a distance of about 30 p m from the pia (complete cell drawn in B). A Lower; B upper part of reconstructed neurone. In B, arrows indicate dendritic course. Arrowheads mark one isolated distal segment of the dendrite. Dashed line indicates upper horder of Fig. A. Boxes (a,b,c) enclose areas containing synaptic contacts of degenerating olfactory boutons with spines of the injected cel1 (enlargements of the boxes are shown in Fig. 6A, B, and C, respectively). Dotted line indicates pial surface. Note darkly stained mitochondria within the dendritic profile (see also Fig. 7B). Bar = 5 pm.

ological details see Buhl and Schlote ('87), Buhl and Lubke ('88),and Buhl et al. ('89).

RESULTS Entorhinal neurones projecting to the hippocampus After injection of FB into the hippocampus, retrogradely labelled somata in the entorhinal cortex were detected nearly exclusively in ipsilateral layers I1 and I11 (Fig. 3). A few labelled cell bodies were found in contralateral layers I1 and 111 and in the deep ipsilateral layers. A total of 54 cells in layers I1 and I11 in the ipsilateral side were intracellularly injected with LY. Injection led to penetration of the dye even into fine dendritic ramifications and a t least into the initial part of the axon, thus producing a Golgi-like appearance of the neurone. We were therefore able to classify the retrogradely labelled cells into four groups: stellate cells, pyramid-like cells, multipolar cells, and horizontal cells (Figs. 4,5). On the basis of' their soma shape, the arrangement of their dendritic tree, and their densely spined dendritic surfaces, the former two cell groups can be compared with the stellate and pyramidal cells that have been described in Golgi material (e.g., Lorente de No, '33) and that are known to project, respectively, to the area dentata and Ammon's horn of the hippocampus (e.g., Steward and Scoville, '76; Schwartz and Coleman, '81). The multipolar and horizontal neurones were sparsely spinous, multipolar cells occurring throughout layers I1 and 111. Horizontalbipolar cells were seen only a t the outer rim of layer 11.Since the olfactory bulb projection terminates in the outer zone of layer I on the apical dendrites of stellate and pyramidal cells, three stellate cells and three pyramidal cells were selected for electron microscopic analysis. These neurones extended apical dendrites close to the pia in the termination zone of the bulbar projection. Occasionally the complete extent of labelling of LY-injected cells in osmicated and flat-embedded preparations i s not seen with the light microscope; however, the reaction product remains detectable in the electron microscope (see Figs. 8,9). Distal dendritic segments can be identified by their darkly stained mitochondria and granular internal structure, which displays a slightly higher electron density than that of unlabelled dendrites (Figs. 8B, 10B). We considered as stellate cells those spiny neurones whose somata were generally located in layer I1 (only a few of them were found in layer 111) and whose dendritic tree was formed by roughly equally sized primary dendrites. In contrast to the pyramidal cells, these neurones did not display one single main apical dendrite. Figure 5 shows that one of the selected stellate cells extended several thick dendrites without any preferential orientation, while the other had only ascending principal dendrites. Dendrites of both cells were intensely ramified and covered with spines. Basally directed dendrites entered layer I11 without extending into deeper layers. Somata of retrogradely labelled pyramidal cells lay in layer 111;one of the pyramidals selected for electron microscopy occupied a superficial position and the other a central position within the layer. The apical dendrite of the pyramidal cells bifurcated into secondary and tertiary spiny dendrites that gave rise to shorter side branches (Fig. 5). Basal dendrites did not extend beyond layer IV.

W.K. SCHWERDTFEGER ET AL.

172

Figure 9

DISYNAPTIC OLFACTORY PATHWAY TO THE HIPPOCAMPUS

Termination of olfactory bulb projection in lateral entorhinal cortex Lesions performed in the olfactory bulb led to dense degeneration of abundant terminal boutons in the molecular layer of the lateral entorhinal cortex as defined by Blackstead ('72). The termination zone ceased at about 50 pm from the rhinal fissure. The greater part of the degenerating boutons were located in the outer 30 pm of the molecular layer (Fig. 6A), but, although with progressively decreasing density, some of them were seen also in deeper zones up to the border with layer I1 (Fig. 7). Degenerating boutons formed asymmetric synaptic contacts on their postsynaptic elements, which in most cases was a dendritic spine (Fig. 6B). Less frequently, terminals were seen in synaptic contact with dendritic trunks (Fig. 6C). Densely degenerated boutons generally contained remnants of dark, apparently swollen mitochondria and in some cases clear mitochondria with a degenerated internal structure (Fig. 6B,C). The morphology of synaptic vesicles could not be determined in such densely degenerated boutons. However, other terminal boutons that displayed a somewhat lighter axoplasm, and thus probably represented earlier stages of degeneration, possessed clear vesicles of roughly spherical shape (see Fig. 9F).

Ultrastructure and synaptic contacts of selected neurones After photoconversion, LY-injected cells could easily be identified in the electron microscope by the dark D A B reaction product in their somata (Figs. 7B, 8A, 10A) and dendrites (Fig. IOB). Nuclei of injected stellate cells displayed a round-to-oval shape (Fig. 8A),while those of pyramidal cells were more elongated (Fig. 10A). In both cell types the karyolemma was largely smooth with only shallow indentations and enveloped an euchromatic karyoplasm with one electron-dense nucleolus (Figs. 8A, IOA). The long axis of the cells measured approximately 20 pm and that of the nuclei about 12 pm. However, nuclei of stellate cells appeared slightly smaller, so that their cytoplasmic perikaryal rim was more prominent than that of the pyramidal cells. The perikaryon of both types contained numerous mitochondria and cross-sectioned profiles of granular endoplasmic reticulum. Various amounts of electron-dense precipitate were apposed t o the postsynaptic densities and thus impeded identification of the morphology of synaptic contacts. However, the large sample provided by the serial sections revealed that the dendritic spines of the injected cells received asymmetric contacts (Figs. 9, lOB), while nearly all of the synapses formed on the somata were symmetric and contained pleomorphic vesicles. A few of the boutons contacting stellate cells displayed dense-core vesicles. In the outer half of the molecular layer, dendrites of the inspected three stellate cells but not of the three selected pyramidal cells were contacted by densely degenerated terminal boutons

Fig. 9. A,B,C Enlargements of the areas in the upper, middle, and lower boxes in Figure 1OB. Black arrows point at labelled spines in synaptic contact with degenerating boutons. Open arrow in C points at synapse between normal bouton and labelled spine. D,E,F Examples for spines of other identified stellate cells in synaptic contact with degenerating boutons. Boutons in F contain a somewhat lighter axoplasm and thus probably represent earlier stages of degeneration. Bars in A-F 0.5 Fm.

173

forming asymmetric contacts and displaying the typical morphology described above (Fig. 9). All of these contacts were established on spines.

DISCUSSION Degeneration of terminals arising from the olfactory bulb The degenerating terminals found after olfactory bulb lesions probably belong to mitral cell axons, since no other cell type projects to the entorhinal cortex (Haberly and Price, '77; Macrides and Schneider, '82). They contained spherical vesicles and formed asymmetric synaptic contacts. This result supports earlier findings (Westrum, '66; Wouterlood and Nederlof, '83) and supports the observation that the olfactory bulb projection is excitatory (Ldmo, '71; Boeijinga and Van Groen, '84). The fibres from the olfactory bulb terminate in the outer molecular layer of the entorhinal cortex (e.g. Wouterlood and Nederlof, '83; Room et al., '84). However, earlier studies limited the extent of termination to the outer zone of the layer, whereas our olfactory bulb lesions showed that a considerable part of the endings are located in deeper zones. We cannot exclude that the deep degenerating boutons belong to axons of neurones undergoing rapid transneuronal degeneration after olfactory bulb lesion, as reported for neurones in the prepiriform cortex (Price, '73; Heimer and Kalil, '78). However, such an effect is more likely to occur in neurones that depend preferentially on the olfactory bulb input. Such neurones probably exist in the prepiriform cortex, which is the most important termination zone of the olfactory bulb fibers. The entorhinal cortex receives afferents from a variety of extrinsic sources (e.g. Beckstead, '78; Insausti et al., '87a,b) and, additionally, the olfactory input terminates nearly exclusively on a short, distal segment of entorhinal neurones, which makes rapid cell death of the latter caused by olfactory bulb lesions rather improbable. Most of the degenerated terminals make asymmetric synaptic contacts on dendritic spines and thus probably with stellate and pyramidal cells. The dendritic trunks in layer I contacted by some of the boutons may belong to GABAergic neurones that receive afferent input from the olfactory bulb (Wouterlood et al., '85).

Entorhind neurones projecting to the hippocampus Results from our Fast Blue injections corroborate earlier findings concerning the topography and laminar distribution of entorhinal cortical neurones that comprise the perforant path to the hippocampus (Germroth et al., '89a). Our injections were placed in the anterior half of the hippocampus and resulted in retrograde labelling of somata in the lateral part of the lateral entorhinal cortex. This result confirms the report of Witter and Groenewegen ('84), who described the entorhinohippocampal projection as following a Iateromedial to rostrocaudal gradient. Similarly, confirmation that a large proportion of the injected Fast Blue entered the area dentata was reflected in the finding that most of the retrogradely labelled cells were stellate cells. These cells are largely confined to layer I1 of the entorhinal cortex and project to the area dentata, whereas field CA1 of Ammon's horn is reached by axons that arise from pyramidlike cells in layer 111(Steward and Scoville, '76; Witter and Groenewegen, '84). In addition, our experiments revealed that two groups of sparsely spinous nonpyramidal cells,

174

W.K. SCHWEBDTFEGER ET AL.

Fig. 10. A Injected pyramidal cell (open star) next to an unlabelled neuron (black star) in layer I11 of the entorhinal cortex. Note t h a t Lucifer Yellow, i.e., its reaction product, remains reliably in the injected cell and its appendages. Bar = 5 pm. B Apical dendrite of injected pyramidal cell. Note the numerous spines (arrowheads). One spine

shows an asymmetric synaptic contact with a terminal houton in this section (arrow). Open arrow points a t axodendritic shaft synapse. White asterisks indicate mitochondria, which stain intensely black during photoconversion. Bar = 1 pm.

namely the multipolar and horizontal-bipolar cells, also send axons to the hippocampus (Germroth et al., '89a). Wouterlood and Nederlof ('83) observed that pyramidal and stellate cells as well as sparsely spinous multipolar cells and GABAergic layer I cells receive olfactory bulb afferents. The morphology of the sparsely spinous cells closely resembles that of GARAergic neurones (e.g., Ribak et al., '81),which are generally considered to be interneurones with a locally distributing axon. GABA-immunoreactive somata occur in all entorhinal layers including layer I (Kohler et al., '85; authors' unpublished observations). Some of them are located at the border with layer I1 like the horizontally

oriented bipolar cells that project to the hippocampus. Apparently, the GARAergic cells in hippocampus (e.g., Alger and Nicoll, '82) inhibit principal neurones as they do in neocortex (e.g., Fonnum and Storm-Mathisen, '78). Van Groen et al. ('87) found electrophysiological evidence for recurrent inhibition of entorhinal principal cells after stimulation of the olfactory bulb. Stimulation of the lateral olfactory tract (LOT) may also result in feed-forward inhibition of entorhinal cells (Finch et al., '88). Hence, part of the LOT fibres end on GABAergic neurones as demonstrated by Wouterlood et al. ('85). Recently we were able to demonstrate by postembedding immunocytochemistry that a t least part of the hori-

175

DISYNAPTIC OLFACTORY PATHWAY TO THE HIPPOCAMPUS

zontal-bipolar and multipolar neurones that project to the hippocampus are GABA-immunoreactive (Germroth et al., '89b). Thus the perforant path apparently mediates not only excitatory, but to a lesser extent, also direct inhibitory influences. There is increasing evidence for other areas of the mammalian brain also that GABAergic neurones possess not only locally arborizing axons, but also a long, projecting axon. Thus, septa1 GABAergic neurones project to the hippocampus (Kohler et al., '84), and GABAergic neurones in the hilus of the hippocampus send an axon to the contralateral hemisphere (Seress and Ribak, '83; Ribak et al., '86).

Bulboentorhinohippocampd projection This study presents anatomical evidence for a disynaptic pathway that connects the olfactory bulb with the hippocampus, thus providing a neuronal substrate for hippocampal processing of olfactory input. This type of processing was already suggested a t the end of the last century, when the hippocampus was considered part of the so-called rhinencephalon (Zuckerkandl, 1887). More recent, comparative anatomical studies indicate that there is no correlation between the phylogenetic development of the olfactory cortices and that of the hippocampus (Stephan, '75; Stephan et al., '88). Likewise, the numerous studies on hippocampal afferents that have been performed in several species proved that no monosynaptic connections exist between the olfactory bulb and the hippocampus. However, the current view that polysensory information is processed in the hippocampus (e.g., Winocur, '80; Deadwyler et al., '87) includes the utilization of olfactory information. Electrophysiological studies support the opinion that such information is relayed in the entorhinal cortex (Wilson and Steward, '78; Habets et al.. '80). Our results are a direct demonstration of bulbohippocampal relations mediated by stellate cells in layer I1 and thus fit the observation that electrical stimulation in the lateral olfactory tract (Wilson and Steward, '78) and in the prepiriform cortex (Habets et al., '80) evokes potentials generated in the distal molecular layer of area dentata where the axons of entorhinal layer I1 cells end (e.g., Steward and Scoville, '76). Entorhinal stellate and pyramidal cells, like other cortical projection neurones, mainly receive asymmetric synaptic contacts on their dendrites, and predominantly symmetric contacts on their somata. The latter are likely to belong to GABAergic, inhibitory neurones (Kohler et al., '85). We did not detect degenerating boutons in contact with the two pyramidal cells selected for electron microscopy, although a complete series of sections was obtained from these cells whose dendrites extended into the zone where degenerating terminals were abundant. This negative result may arise from one of the following reasons: 1) labelling of the dendritic tree, although virtually complete at the light microscopic level, failed to involve just those branches that receive the olfactory bulb d e r e n t s ; 2) subpopulations of pyramidal cells exist whereby some are contacted by the olfactory fibers, while others are not; 3 ) pyramidal cells are not reached by the projection from the olfactory bulb. Based on our data none of these possibilities can be excluded. The second possibility may receive some support by Golgi studies reporting that several morphologically distinct types of pyramidal layer 111 cells exist (e.g., Lorente de No, '33). Since the stellate cells project to the area dentata and hippocampal field CA3, and the pyramidals to field CAI (Witter and Groenewegen, '84), the third possibility may be favored by the finding that stimulation of

the lateral olfactory tract evokes potentials in the area dentata, but not in field CAI (Wilson and Steward, '78). However, future studies on a larger sample of cells are indispensable to confirm or contradict the negative result.

Methodological aspects By combining retrograde tracing, intracellular staining, intracellular injection, and anterograde degeneration a novel approach has been taken to reveal a disynaptic pathway in the same piece of tissue. Essentially the protocol is technically easy, works reliably, and is highly selective due to visually guided penetration of labelled neilrones (Buhl and Lubke, '88; Buhl et al., '89). Due to the additional use of glutaraldehyde in the fixative, the quality of tissue preservation has been considerably improved (cf. Buhl and Schlote, '87; Buhl and Lubke, '88; Buhl et al., '89), thus facilitating ultrastructural analysis. Previous studies have combined tract tracing with either horseradish peroxidase or fluorescent dyes with Golgi impregnations (Freund and Somogyi, '83; Somogyi and Smith, '79; Catsicas et al. '86). However, this procedure critically depends on the randomness of the Golgi technique. In addition, the latter method, when applied for EM purposes (Fairen et al., '77), is technically difficult and requires considerably more steps than the simple LY-EM protocol. Recently, retrograde tracing and intracellular staining have been combined in an in vitro slice preparation (Katz '87). However, due to the rapid onset of degenerative ultrastructural changes (Frotscher e t al., '81; Crunelli et al., '87), the evaluation of anterogradely degenerating boutons in this material would render equivocal results. Furthermore, the occurrence of dye coupling could be potentially confusing with regard to the identity of labelled cells and processes.

CONCLUSIONS Olfactory input is probably not the only sensory information transmitted by the perforant path to the hippocampus. It has been proposed that visual stimuli from primary and secondary cortical areas may reach the entorhinal cortex via the presubiculum (Vogt and Miller, '83). Auditory input may be relayed from the medial geniculate nucleus to the perirhinal cortex (Room and Groenewegen, '86), which, in turn, projects to the entorhinal cortex (Witter et al., '86). In the monkey, the latter has been reported to receive fibres originating in prefrontal and temporal association fields (Insausti et al., '87a,b). All of these putative sensory pathways, however, constitute rather indirect, polysynaptic relations. The polysynaptic channels are likely to transmit highly preprocessed and premodulated stimuli and thus may yield stored sensory information; the latter, according to one of the models of hippocampal function, may be compared with current sensory input (Vinogradova, '75; Schwerdtfeger, '84; Deadwyler et al., '87), which, however, ought to reach the hippocampus by faster oligosynaptic routes. Further work is required to elucidate whether more direct pathways connect not only the olfactory but also other sensory receptors with the hippocampus. Interestingly, a direct projection has been found from the retina to the anterior dorsal thalamus (Itaya e t al.. '86; Reperant et al., '87) from where, in turn, axons run to the hippocampus (Wyss et al., 1979; Schwerdtfeger, '84; Reperant et al., '87). The methods used in this study are not limited to investigating the existence of proposed disynaptic

W.K. SCHWERDTFEGER ET AL.

176

relations; they may also be combined with immunocytochemistry, Golgi methods, and anterograde tracing and may thus help to identify comprehensively the afferent and efferent connections of individual neurones.

ACKNOWLEDGMENTS The authors are grateful to R. Kraup, G.-S. Nam, A. Wennekers, and W. Hofer for valuable technical assistance, and Dr. J.F. Dann for reading the manuscript. We thank Prof. W. Singer for providing the injection facilities. This study was supported by the Deutsche Forschungsgemeinschaft.

L I T E R A m CITJ3D Alger, B.E., and R.A. Nicoll(1982) Pharmacological evidence for two kinds of GABA receptor on rat hippocampal pyramidal cells studied in vitro. J. P h ysiol. 328~125-141. Beckstead, R.M. (1978) Afferent connections of the entorhinal area in the rat as demonstrated by retrograde cell-labeling with horseradish peroxidase. Brain Res. 152:249-264. Bentivoglio, M., H.G.J.M. Kuypers, C.E. Catsman-Berrevoets, H. Loewe, and 0. Dann (1980) Two new fluorescent neuronal tracers which are transported over long distances. Neurosci. Lett. 18:2&30. Rerger, T.W., G.A. Clark, and R.F. Thompson (1980) Learning-dependent neuronal responses recorded from limbic system brain structures during classical conditioning. Physiol. Psychol. 8t155-167. Rlackstead, T.W. (1972) Commissural connections of the hippocampal region in the rat with special reference to their mode of termination. J. Comp. Neurol. 105:417-539. Boeijinga, P.H., and Th. Van Groen (1984) Inputs from the olfactory bulb and olfactory cortex to the entorhinal cortex in the cat. 11. Physiological studies. Exp. Brain Res. 57~4048. Buhl, E.H., and W. Schlote (1987) Intracellular Lucifer Yellow staining and electron microscopy of neurones in slices of fixed epitumourous human cortical tissue. Acta Neuropathol. 75:14@146. Buhl, E.H., and J. Lubke (1988) Intracellular Lucifer Yellow injection in fixed brain slices combined with retrograde tracing, light and electron microscopy. Neuroscience 28t3-16. Buhl, E.H., W.K. Schwerdtfeger, P. Germroth, and W. Singer (1989) Combining retrograde tracing, intracellular injection, anterograde degeneration and electron microscopy to reveal synaptic links. J. Neurosci. Methods (in press) Catsicas, S., P.J. Berbel, and G.M. Innocenti (1986) A combination of Golgi impregnation and fluorescent retrograde labeling. J. Neurosci. Methods 18325-332. Crunelli, V., N. Leresche, J.W. Hynd, N.M. Patel, and J.G. Parnavelas (1987) An in vitro slice preparation of the cat lateral geniculate nucleus. J. Neurosci. Methods 20:211-219. Deadwyler, S.A., Th.C. Foster, and R.E. Hampson (1987) Processing of sensory information in the hippocampus. CRC Crit. Rev. Clin. Neurobiol. 2335-355. Doller, H.J., and F.F. Weight (1982) Perforant pathway activation of hippocampal CA1 stratum pyramidale neurons: Electrophysiological evidence for a direct pathway. Brain Res. 237:l-13. Dolphin, A.C., M.L. Errington, and T.V.P. Bliss (1982) Long-term potentiation of the perforant path in vivo is associated with increased glutamate release. Nature 297:496498. Douglas, R.J. (1975) The development of hippocampal function: Implications for theory and for therapy. In R.L. Isaacson and K.H. Pribram (eds): The Hippocampus. Vol. 11, Neurophysiology and Behavior. New York: Plenum Press, pp. 327-362. Eichenhaum, H., and N.J. Cohen (1988) Representation in the hippocampus: What do hippocampal neurons code? TINS 11~244-248. Fairen, A., A. Peters, and J. Saldanha (1977) A new procedure for examining Golgi impregnated neurons by light and electron microscopy. J. Neurocytol. 6:311-~337. Finch, D.M., A.M. Tan, and M. Isokawa-Akesson (1988) Feedforward inhibition of the rat entorhinal cortex and subicular complex. J. Neurosci. 8:2213-2226. Fonnum, F., and J. Storm-Mathisen (1978) Localization of GABA-ergic neurons in CNS. In L.L. Iversen, S.D. Iversen, and S.H. Snyder (eds): Handbook of Psychopharmacology. New York: Plenum Press, pp. 357401.

Foreman, N., and H. Stevens (1987) Relationships between the superior colliculus add hippocampus: Neural and behavioral considerations. Behav. Brain Sci. 10:101-152. Freund, T.F., and P. Somogyi (1983) The section-Golgi impregnation procedure. 1. Description of the method and its combination with histochemistry after intracellular iontophoresis or retrograde transport of horseradish peroxidase. Neuroscience 9:463- 474. Frotscher, M., U.Misgeld, and C. Nitsch (1981) Ultrastructure of mossy fiber endings in vitro hippocampal slices. Exp. Brain Res. 41:247-255. Germroth, P., W.K. Schwerdtfeger, and E.H. Buhl (1989a) Morphology of identified entorhinal neurons projecting to the hippocampus. A light microscopical study combining retrograde tracing and intracellular injection. Neuroscience (in press). Germroth, P., W.K. Schwerdtfeger, and E.H.Buhl (1989b) GABAergic neurons in the entorhinal cortex project to the hippocampus. Brain Res. (in press). Gray, J.A. (1977) Drug effects on fear and frustration: possible limbic site of action of minor tranquillizer. In L.L. Iversen, S.H. Jversen, and S.H. Snyder (eds): Handbook of Psychopharmacology. New Y o r k Plenum Press, pp. 433-515. Haberly, L.B., and J.L. Price (1977) The axonal projection patterns of the mitral and tufted cells of the olfactory bulb in the rat. Brain Res. 129:152157. Habets, A.M.M.C., F.H. Lopes Da Silva, and W.J. Mollevanger (1980) An olfactory input to the hippocampus of the c a t Field potential analysis. Brain Res. 18247-64. Heimer, L., and R. Kalil(1978) Rapid transneuronal degeneration and death of cortical neurons following removal of the olfactory bulb in adult rats. J. Comp. Neurol. 178:559-610. Insausti, R., D.G. Amaral, and W.M. Cowan (1987a) The entorhinal cortex of the monkey: 11. Cortical afferents. J. Comp. Neurol. 264:356-395. Insausti, R., D.G. Amaral, and W.M. Cowan (1987h) The entorhinal cortex of the monkey: 111. Subcortical afferents. J. Comp. Neurol. 264t396-408. Itaya, S.K., G.W. Van Hoesen, and L.A. Benevento (1986) Direct retinal pathways to the limbic thalamus of the monkey. Exp. Brain Res. 61:607613. Katz, L.C. (1987) Local circuitry of identified projection neurons in cat visual cortex brain slices. J. Neurosci. 7t1223-1249. Kohler, C., V. Chan-Palay, and J.-Y. Wu (1984) Septa1 neurons containing glutamic acid decarboxylase immunoreactivity project to the hippocampal region in the rat brain. Anat. Embryol. 169.41-44. KShler, C., J.-Y. Wu, and V. Chan-Palay (1985) Neurons and terminals in the retrohippocampal region in the rat’s brain identified hy anti-y-aminobutyric acid and anti-glutamic acid decarboxylase immunocytochemistry. Anat. Embryol. 173:35-44. LBmo, T. (1971) Patterns of activation in a monosynaptic cortical pathway: The perforant path input to the dentate area of the hippocampal formation. Exp. Brain Res. 12:1&45. Lorerite de Nb, R. (1933) Studies on the structure of the cerebral cortex. J. Psychol. Neurol. 45:381438. Macrides, F., and S.P. Schneider (1982) Laminar organization of mitral and tufted cells in the main olfactory bulb of the adult hamster. J. Comp. Neurol. 208~419430. Maranto, A.R. (1982) Neuronal mapping: A photooxidation reaction makes Lucifer Yellow useful for electron microscopy. Science 21 7t953-955. Milner, B. (1959) The memory defect in bilateral hippocampal lesions. Psychiatr. Res. Rep. 11:43-52. Nadler, J.V., and E.M. Smith (1981) Perforant path lesion depletes glutamate content of fascia dentata synaptosomes. Neurosci. Lett. 25:275280. O’Keefe, J., and L. Nadel (1978) The Hippocampus as a Cognitive Map. Oxford Clarendon. Paxinos, G., and C. Watson (1982) The Rat Brain in Stereotaxic Coordinates. London: Academic Press. Price, J.L. (1973) An autoradiographic study of complementary laminar patterns nf termination of afferent fibers to the olfactory cortex. J. Comp. Neurol. 150t87-108. Repbrant, J., C. Weidner, A. Pakhomova, A.-M. Desroches, N. Vesseklin, and M. Lemire (1987) A retino-thalamo-hippocampal pathway in the rat. C.R. Acad. Sci. Paris. 305:601-604. Rihak, C.E., J.E. Vaughn, and R.P. Barber (1981) Immunocytochemical localization of GABAergic neurones a t the electron microscopical level. Histochem. J. 13~555-582. Rihak, C.E., L. Seress, G.M. Peterson, K.B. Seroogy, J.H. Fallon, and L.C. Schmued (1986) A GABAergic inhibitory component within the hippocampal commissural pathway. J. Neurosci. 6~3492-3498.

DISYNAPTIC OLFACTORY PATHWAY TO THE HIPPOCAMPUS Robinson, G.B., and R.J. Racine (1982) Heterosynaptic interactions between septa1 and entorhinal inputs to the dentate gyrus: Long-term potentiation effects. Brain Res. 249t162-166. Room, P., and H.J. Groenewegen (1986) Connections of the parahippocampal cortex in the cat. 11. Subcortical afferents. J. Comp. Neurol. 251t451-473. Room, P., H.J. Groenewegen, and A.H.M. Lohman (1984) Inputs from the olfactory bulb and olfactory cortex to the entorhinal cortex in the cat. I. Anatomical observations. Exp. Brain Res. .56:48-96. Schwartz, S.P., and P.D. Coleman (1981) Neurons of origin of the perforant path. Exp. Neurol. 74:30%312. Schwerdtfeger, W.K. (1984) Structure and fiber connections of the hippocampus. Adv. Anat. Embryol. Cell. Biol. 83. Berlin: Springer-Verlag, pp. 1-74. Schwerdtfeger, W.K., and E. Buhl (1986) Various types of non-pyramidal hippocampal neurons project to the septum and contralateral hippocampus. Brain Res. 386r146-154. Seress, L., and C.E. Ribak (1983) GABAergic cells in the dentate gyms appear to he local circuit and projection neurons. Exp. Brain Res. 50~17% 182. Solomon, P.R. (1979) Temporal versus spatial information processing theories of hippocampal function. Psychol. Bull. 86:1271-1279. Somogyi, P., and A.D. Smith (1979) Projection of neostriatal spiny neurons to the substantia nigra. Application of a combined Golgi-staining and horseradish peroxidase transport procedure at both light-and electron-microscopic levels. Brain. Res. J78:3-15. Sripanidkulchai, K., and J.M. Wyss (1986) Thalamic projections to retrosplenial cortex in the rat. J. Comp. Neurol. 2541143-165. Stephan, H. (1975) Allocortex. In W. Bargmann (ed): Handbuch der mikroskopischen Anatomie des Menschen, Vol IV/9. Berlin: Springer-Verlag. Stephan, H.,G. Baron, and H.D.Frahm (1988) Comparative size of brains and brain components. Cornp. Primate Biol. 4:l-38. Steward, O., and S.A. Scoville (1976) Cells of origin of entorhinal cortical afferents to the hippocampus and fascia dentata of the rat. J. Cornp. Neurol. 269:347-370. Stewart, W.W. (1978) Functional connections between cells as revealed by dye-coupling with a highly fluorescent naphtalimide tracer. Cell 24:741759. Stewart, W.W. (1981) Lucifer dyes-highly fluorescent dyes for biological tracing. Nature 292r17-21. Tauchi, M., and R.H. Masland (1984) The shape and arrangement of the cholinergic neurons in the rabbit retina. Proc. R. Soc. Lond. [Biol.] 223:lOl-119.

177

Van Groen, Th., F.H. Lopes da Silva, and W.J. Wadman (1987) Synaptic organization of olfactory inputs and local circuits in the entorhinal cortex: A current source density analysis in the cat.. Exp. Brain Res. 67:615-622. Vinogradova, O.S. (1975) Functional organization of the limbic system in the process of registration of information: Facts and hypotheses. In R.L. Isaacson and K.H. Pribram (eds): The Hippocampus, Vol 11. Neurophysiology and Behavior, New York: Plenum, pp. 3-70. Vogt, B.A., and M.W. Miller (1983) Cortical connections between rat cingulate cortex and visual, motor, and postsubicular cortices. J. Comp. Neural. 216r192-210. Westrum, L.E. (1966) Electron microscopy of degeneration in the prepiriform cortex. J. Anat. 100t683485. Wilson, R.C., and 0. Steward (1978) Polysynaptic activation of the dentate gyms of the hippocampal formation: An olfactory input via the lateral entorhinal cortex. Exp. Brain Res. 33:523-534. Winocur, G. (1980) The hippocampus and cue utilization. Physiol. Psychol. 8:280-288. Witter, M.P., and H.J. Groenewegen (1984) Laminar origin and septotemporal distribution of entorhinal and perirhinal projections to the hippocampus in the cat. J. Cornp. Neurol. 224:371-385. Witter, M.P., and H.J. Groenewegen (1986a) Connections of the parahippocampal cortex in the cat. In. Cortical and thalamic efferents. J. Comp. Neurol. 252:l-31. Witter, M.P., and H.J. Groenewegen (1986h) Connections of the parahippocampal cortex in the cat. IV. Subcortical efferents. J. Comp. Neurol. 2513-77. Witter, M.P., P. Room, H.J. Groenewegen, and A.H.M. Lohman (1986) Connections of the parahippocampal cortex in the cat. V. Intrinsic connections; comments on input/output connections with the hippocampus. J. Comp. Neurol. 252%--94. Wouterlood, F.G., and J. Nederlof (1983) Terminations of olfactory afferents on layer I1 and I11 neurons in the entorhinal area: degeneration-Golgielectron microscopic study in the rat. Neurosci. Lett. 36:105-110. Wouterlood, F.G., E. Mugnaini, and J. Nederlof (1985) Projection of olfactory bulb efferents to layer I GABAergic neurons in the entorhinal area. Combination of anterograde degeneration and immunoelectron microscopy in rat. Brain Res. 343~283-296. Wyss, J.M., L. Swanson, and W.M. Cowan (1979) A study of subcortical d e r ents to the hippocampal formation in the rat. Neuroscience 41463-476 Zuckerkandl, E. (1887) Uber das Riechzentrum. Eine vergleichend-anatomische Studie. Stuttgart: Enke.