Formation of the septohippocampal projection in vitro: An electron microscopic immunocytochemical study of cholinergic synapses

Formation of the septohippocampal projection in vitro: An electron microscopic immunocytochemical study of cholinergic synapses

0306-4522/93 $6.00 + 0.00 Pergamon Press Ltd ~7.11993 IBRO Neuroscience Vol. 52, No. 4, pp. 815-827, 1993 Printed in Great Britain FORMATION OF THE ...

14MB Sizes 0 Downloads 7 Views

0306-4522/93 $6.00 + 0.00 Pergamon Press Ltd ~7.11993 IBRO

Neuroscience Vol. 52, No. 4, pp. 815-827, 1993 Printed in Great Britain


of Anatomy,



and M.

of Freiburg,



17, D-7800



Abstract-Cholinergic neurons in the medial septum/diagonal band complex project to the hippocampus and fascia dentata and establish characteristic types of synapses on a variety of target neurons. At present we do not know the principles that underlie the development of this projection and the formation of the cholinergic synapses. Here we have used co-cultured slices of septum and hippocampus of one- to six-day-old rat pups to study the development of the septohippocampal pathway and the formation of cholinergic synapses on hippocampal target neurons in vitro. Slices of septum and hippocampus were incubated together for IO-46 days applying the roller-tube technique. The fluorescent dye dioctadecyltetramethylindocarbocyanine perchlorate and histochemical staining for acetylcholinesterase labeled many fibers connecting both explants. Combined light- and electron-microscopic immunocytochemistry for choline acetyltransferase, the acetylcholine-synthesizing enzyme, revealed multipolar immunopositive neurons with long aspiny dendrites in the septal culture. Numerous varicose immunoreactive, supposedly cholinergic fibers could be followed from the septal to the hippocampal culture where they ramified and formed a three-dimensional network. As in siru, cholinergic terminals formed characteristic symmetric synapses on cell bodies, spines and, most often, on dendritic shafts of the hippocampal target neurons. No immunoreactive fibers and synapses were observed in single cultures of hippocampus. These results demonstrate that the cholinergic septohippocampal projection develops in vitro and that similar types of cholinergic synapses are established on co-cultured hippocampal target neurons as observed in silu.

Cholinergic neurons in the basal forebrain give rise to widespread projections to neocortical and allocortical areas.35 At present the principles that govern the development of these projections as well as synapse formation in the target areas are poorly understood. As far as the cholinergic septohippocampal projection is concerned, it is well known that these fibers arise from large neurons in the medial septal nucleus and in the nucleus of the vertical limb of the diagonal band of Broca, run through the fimbria-fornix, and terminate in all layers of the hippocampal formation.‘2~‘3~28~34 As revealed by immunocytochemistry for choline acetyltransferase (ChAT), the acetylcholine-synthesizing enzyme, cholinergic fibers form characteristic synapses with a variety of target neurons in the hippocampus and fascia dentata.‘.“-13 Thus. ChAT-positive terminals were found to establish synapses with dendritic shafts, spines, and cell

bodies of identified hippocampal pyramidal cells, dentate granule cells and hippocampal and dentate GABAergic non-pyramidal neurons.‘5.33 This is in sharp contrast to the GABAergic component of the septohippocampal projection. As shown by Freund and Antal,’ septohippocampal GABAergic fibers selectively innervate GABAergic neurons in the hippocampus proper and fascia dentata. At present the factors underlying the different target cell selectivities of cholinergic and GABAergic septohippocampal fibers are unknown. In the present study we have tested the capacity of cholinergic neurons in slice cultures of the medial septal nucleus and nucleus of the vertical limb of the diagonal band of Broca to differentiate, project to a co-cultivated hippocampal slice, and form synapses on the hippocampal target neurons. Co-culturing of septal and hippocampal slices has been performed in a number of studies.‘8~20S40In a voltage-clamp analysis, Gtihwiler and Brown’* demonstrated a cholinergic component in hippocampal slice cultures co-cultivated with septum. However, in those studies the formation of the septohippocampal projection in uitro has been demonstrated by acetylcholinesterase (AChE) histochemistry which does not exclusively stain cholinergic neurons and does not reveal cholinergic synapses. Moreover, the extent to which cholinergic neurons intrinsic to the hippocampal

*To whom correspondence should be addressed. Abbreviations: AChE, acetylcholinesterase; bFGF,

basic fibroblast growth factor; BME, basal medium (Eagle); ChAT, choline acetyltransferase; DAB, 3,3’-diaminobenzidine; DiI, I,l’-dioctadecyl-3,3,3’,3’-tetramethylindocarbocyanine perchlorate; DIV, days in vitro; DMSO, dimethylsulfoxide; GBSS, Gey’s balanced salt solution; HBSS, Hanks’ balanced salt solution; NGF, nerve growth factor; PB, phosphate buffer; PHA-L, Phaseolus vulgaris leucoagglutinin. 815



formationI contribute to the observed effects remained to be determined. ChAT is presently the best available marker for cholinergic neurons, fibers and synapses. In the present study, we have used a monoclonal antibody

against ChAT to stain cholinergic neurons in cultured septal slices and cholinergic synapses in co-cultured hippocampal slices.



Preparation of co-cultures of septum and hippocampus

Brains of one- to six-day-old Sprague-Dawley rats (stock of the Anatomical Institute, Freiburg, Germany) were aseptically removed and immersed in cold (4°C) Gey’s balanced salt solution (GBSS) supplemented with glucose to a final concentration of 6.4 m&ml. The septal region and the hippocampi were dissected and blocked. Transverse sections (30&4OO~m thick) were cut by means of a McIlwain tissue sectioner. For cultivation the roller-tube technique, as described in detail by GBhwiler,16,” was applied. Briefly, a hippocampus slice and a septum slice were mounted together on a glass coverslip, embedded in a clot of chicken plasma which was coagulated by the addition of a drop of thrombin. The distance between the two explants was about l-l.5 mm to avoid confluence of the two cultures during cultivation. This way we were still able to identify both co-cultures after extended periods of time following explantation. Unlabeled fibers crossing the gap between the two cultures were then faintly visible under the phasecontrast microscope. For cultivation, the glass coverslips were transferred to plastic tubes containing 0.5 ml of medium composed of 50% basal medium (BME, Eagle), 25% Hanks’ balanced salt solution (HBSS), 25% heat-inactivated horse serum and glucose (6.4 mg/ml medium). The nutrient medium was supplemented with 25 or 50 ng/ml2.5 S nerve growth factor (NGF), purchased from Paesel& Lorei GmbH, Frankfurt. The cultures were fed three times a week, and NGF was given twice weekly. The cultures were incubated at 36°C in dry air for various periods of time. Fiber tracing with DiI

The fluorescent tracer l,l’-dioctadecyl-3,3,3’,3’-tetramethylindocarbocyanine perchlorate (DiI) can be used for anterograde and retrograde labeling of fibers and cell bodies, respectively.26 Co-cultures were incubated for 1l-13 days in vitro (DIV) and were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (PB). Small crystals of DiI were placed under visual control into the septal culture by using a glass micropipette. The co-cultures were then stored in the fixative for four to six weeks at room temperature in the dark to allow for the labeling of fibers extending between the septal culture and the co-cultured hippocampus. Histochemical staining for acetylcholinesterase

Co-cultures from postnatal day two were kept in vitro for two weeks. Thereafter they were fixed on the glass coverslip for 15 min in a solution Containing 4% paraformaldehyde and 0.08% nlutaraldehvde in 0.1 M PB IDH 7.3-7.4). Following rinsing in 0.1 hi PB, they were &ined for kChE. For this, a modified Karnovsky-Roots procedure was applied.36 After rinsing in 0.1 M maleate buffer (2 x 4 min), the co-cultures were immersed for 2 h in an incubation medium containing 0.29 mg promethazine (an inhibitor of unspecific cholinesterases), 46.0 mg maleic acid, 5.88 mg sodium citrate, 3.0mg copper sulfate, and 0.66mg potassium ferricyanide in 4OOml 0.1 M maleate buffer. A droo of glacial acetic acid was added. Immediately before &u&ion, acetylthiocholine iodide was added as a substrate, and the

solution adjusted to pH 8.0.jh The co-cultures were then rinsed two times in 0.1 M Tris buffer, dipped in 0.5% cobalt chloride in 0.1 M Tris buffer for 10 min, and rinsed again in 0.1 M Tris buffer. For visualization we used 3,3’diaminobenzidine (DAB) as a chromogen. The solution consisted of 0.05% DAB olus 0.01% H,O, in 0.1 M Tris buffer. After 3 min the DAb reaction was siopped, and the co-cultures were rinsed three times (3 mins each) in Tris buffer. The co-cultures were finally dehydrated through graded series of ethanol and xylene. and coverslipped with Flo-Texx (Shandon). Choline acetyyltransj~rase immunocytochemistq

For these experiments co-cultures taken from one- to five-day-old animals were used. They were incubated for periods ranging from 1446 days. For selection of co-cultures to be used for ChAT immunocytochemistry we established the following criteria: (i) both septal and hippocampal cultures could be identified; and (ii) the plasma--thrombin clot was found intact which enabled the outgrowing septal fibers to reach the hippocampal target culture. The cocultures were fixed on the coverslip by removing the medium and filling the culture tube with fixative. The fixative contained 4% paraformaldehyde, 0.08% glutaraldehyde and 15% saturated picric acid in 0.1 M phosphate buffer adjusted to pH 7.3-7.4. The fixative solution was changed three times. After rinsing in PB several times, the co-cultures were dissected out of the ensheathing chicken plasma clot and were carefully removed from the underlying glass coverslip. Care was taken to avoid the disruption of the fiber bridge connecting both cultures. Next, the co-cultures were incubated in a graded series of dimethylsulfoxide (DMSO) in 0.1 M phosphate buffer (5, 10,20,40%; 10 min each). This procedure was performed to improve the penetration of the primary antibody. In a previous pilot study we had tried to increase penetration by freezing the tissue in liquid nitrogen which, however, often disrupted the fiber bridge and damaged the cultured tissue. Following the DMSO treatment, the co-cultures were washed in PB several times and then incubated in the primary antibody (monoclonal anti-ChAT, Boehringer; diluted 1: 10, for six days at 4°C). Shorter periods of incubation with the primary antibody failed to give a strong immunolabeling. The incubation solution contained 0.1% NaN, to prevent bacterial contamination. After rinsing in PB, the co-cultures were incubated antirat IgG (Vector Laboratories, in biotinylated Burlingame, CA; diluted 1:250 in PB) at 4°C overnight. Rinsing in PB was followed by incubation in Vectastain ABC reagent for 2 h at room temperature. Thereafter, the septohippocampal co-cultures were again washed in PB several times. For visualization of the immunoreaction. DAB (Sigma) was used as a chromogen. The reaction was intensified following the protocol described by Adams.’ For this, the co-cultures were immersed in a solution of 0.05% DAB, 0.025% CoCl, and 0.02% ammonium nickel sulfate [(NH,),Ni(SO,),] in 0.1 M PB @H 7.4). Each co-culture was incubated separately in an Eppendorf vial filled with 1 mlbf the solution. After 20 min of incubation, 3.3 ~1 3% Hz02 were added to each vial. The DAB reaction was monitored under the microscope and stopped by several rinses in cold (4°C) 0.1 M PB. For control purposes the primary ChAT antibody was omitted. No immunostaining occurred under these conditions. In addition, single slice cultures of either septum or hippocampus and co-cultures of two hippocampal slices were prepared and treated as described for the co-cultures of septum and hippocampus. Single cultures of septum were used to study the survival and differentiation of cholinergic septal neurons in the absence of target-derived trophic factor. *OIn single cultures of hippocampus and in co-cultures of two hippocampal slices we looked for cholinergic (ChAT-positive) fibers originating from hippocampal cholinergic neurons. I4 All cultures were osmicated in 1% 0~0, in PB for 10 min, dehydrated, flat-embedded in Epon

Septohippocampal and photographed. Electron microscopic examination included the hippocampal as well as the septal portions of the co-cultures. Following re-blocking the tissue was cut on a Reichert-Jung Ultracut. Thin-sections were mounted on Formvar-coated single-slot grids, stained with 5% lead citrate and 2% uranyl acetate, and analysed in a Zeiss 109 electron microscope. RESULTS Vital co-cultures of septum and hippocampus were still found after six to seven weeks of cultivation. Both explants have flattened and broadened during the incubation period thereby approaching each other (Fig. la). Numerous cells, mainly astrocytes, have emigrated into the surrounding chicken plasma clot. Thus, in many co-cultures the boundary of the septal explant was not clearly demarcated, but the pyramidal cell layer of the hippocampal explant was detectable using Nomarski optics, so that the two cultures could be distinguished. In other cases a bridge had formed between the two explants (Figs 2,3). The granule cell layer of the fascia dentata only partially developed in many hippocampal explants as already reported in a previous study using Golgi impregnation in order to label the granule cells.*’

Tracing with DiI Co-cultures that were incubated for two weeks, then fixed and labeled with the fluorescent tracer DiI, showed fiber bundles connecting both explants (Fig. 1). The site of DiI injection is visible in the septal culture (Fig. la, b). Anterogradely labeled fibers were seen that crossed the intervening plasma clot and reached the co-cultivated hippocampal slice (Fig. lb, c). Essentially two types of fibers could be differentiated, i.e. fine, beaded fibers and thick fibers with large varicose swellings (Fig. lc). The thin, beaded fibers probably correspond to the fibers stained for ChAT in the bridge between the two explants (Fig. 6a). Thin and thick septohippocampal fibers have similarly been stained by anterograde tracing with Phase&s vulgaris leucoagglutinin (PHA-L) in situ and have been identified as axons of cholinergic and GABAergic septohippocampal projection neurons, respectively.’ Certainly, we cannot exclude that some fibers were retrogradely labeled axons of neurons in the hippocampal culture projecting to the septal explant. However, retrogradely labeled cell bodies in the hippocampal culture were surprisingly rare. Acetylcholinesterase staining of somata and processes Histochemical staining for AChE, the acetylcholine degrading enzyme, was performed in six septohippocampal co-cultures (Fig. 2). In the septal culture, numerous somata of neurons with long processes were stained that formed an irregular meshwork and an extended bridge and finally invaded the hippocampal culture (Fig. 2b) where they ramified extensively in both the hippocampal


in L>itro


and dentate portions of the slice. This fiber staining was not observed in single cultures of hippocampus and in the hippocampal portions of co-cultures which had not developed a bridge. Occasionally AChEstained fibers protruded into the surrounding plasma clot. Choline acetyltran!ferase immunocytochemistr? Immunocytochemistry for ChAT combined with nickel-cobalt intensification specifically stained neuronal elements by a dark blue or black reaction product. Intensification of the immunolabeling with nickel



Fig. 1. DiI tracing of fibers between a septal (S) and a hippocampal (H) slice co-cultured in vitro for 12 days. (a) View of the co-culture with transmitted light using Nomarski optics. Asterisk indicating site of DiI injection. (b) Same culture as in a under the fluorescence microscope. Note strong DiI labeling of the septal culture which gives rise to a fiber projection directed towards the hippocarnpal co-culture. Framed area shown at higher rna~~~tion in c. (c) Higher magnification of framed area in b illustrating single DiI-labeled fibers. Note that two types of fibers are stained, i.e. thick fibers with large varicosities (open arrows) and fine, beaded fibers (arrow).


Fig. 2. Hir ;tochemical staining for AChE s, b eptal i md H, hippocampal parts of sept al cull :ure are directed towards the invading the hippocampal


in ~h-o

of a septohippocampal co-culture incubated for 14 days in vitro. co-culture. Note that AChE-stained fibers originating from the hippo~mpal co-cuiture. The framed area demonstrating fibers co-culture is shown at higher magnification in b.




Fig. 3. Camera lucida drawing of a septohippocampal co-culture stained for choline acetyltransferase (DIV: 46). Representative areas are illustrated by photomicrographs in Figs 4, 6 (framed areas l-4 in this Figure). Note that immunoreactive perikarya are only present in the septal culture (S) and in the adjacent portion of the bridge (B). The majority of immunoreactive fibers in the bridge are orientated towards the hippocampal co-culture. In the hippocampal target region the ChAT-immunopositive fibers outline the curved structure of the Ammon’s horn. The shaded area corresponds to the pyramidal cell layer in CA3 and CA1 which has broadened during the in vitro period. No immunopositive cell bodies are seen in the hippocampal co-culture.


Fig. 4. Photomicrograph in Fig. 3 (framed area



in vitro

of ChAT-immunopositive cell bodies in the septal part of the co-culture shown 1). Immunopositive neurons are arranged in clusters. Multipolar and bipolar neurons are seen that give rise to smooth dendrites.

by micrographs (Figs 4,6). It is obvious from the drawing that immunoreactive cells in the septal culture are arranged in clusters, but some cells appear to have migrated into the bridge. Immunoreactive fibers in the bridge are orientated towards the hippocampal co-culture. In the hippocampal target tissue they outline the characteristic curvature of Ammon’s horn. Staining of perikarya is absent as is a welldiscernible dentate gyrus. Our electron-microscopic analysis of the hippocampal parts of the co-cultures revealed many immunoreactive fibers and presynaptic terminals. We regard it as the main result of the present study that very similar types of cholinergic synapses were observed in hippocampal slices co-cultivated with septum as have been described in a number of studies of cholinergic synapses in situ.5.‘2.‘3.33.45ChATimmunoreactive terminals thus formed symmetric synapses with cell bodies (Fig. 7a), dendritic shafts (Fig. 7b, c), and spines (Fig. 7d) of the hippocampal target cells. We have reason to assume that these ChAT-positive terminals originated from the septal co-culture since immunostaining of fibers and presynaptic terminals was absent in single cultures and in co-cultures that failed to develop a fiber bridge. Incubation of the cultures in DMSO prior to immunolabeling clearly increased antibody penetration when compared with untreated tissue but did

not severely affect the fine-structural the tissue.




The present study has shown that septal cholinergic neurons survive in slice culture, project to a cocultured hippocampal slice, and form similar types of synapses with the hippocampal target neurons as described in studies of cholinergic synapses in situ. In an attempt to interpret these findings we will discuss here: (i) the factors that may play a role in the survival of the cholinergic septal neurons; (ii) the cholinergic nature of at least some of the fibers connecting the septal and the hippocampal culture; and (iii) the origin of the chohnergic synapses in the hippocampal culture from axons of the cholinergic neurons in the septal co-culture and not from hippocampal ChAT-positive neurons.14 Before these points are dealt with in some detail, it should be pointed out that the development of the septohippocampal projection in vitro, as described in the present paper, is rather a regeneration of this fiber tract than a de nom formation. Parallel in vivo studies have shown that first septofugal fibers reach the hippocampus by embryonic day 18 (Linke and Frotscher, unpublished observations). However, the slice cultures for the present experiments were taken from one- to six-dayold rats.



Fig. 5. Input synapses of dendrites of ChAT-immunoreactive neurons in the septal part of a co-culture (DIV: 15) (a) Smooth dendrite (D) covered by presynaptic terminals (arrows) forming synaptic contact. Asterisks label two boutons shown at higher magnification in b. (b) Higher magnification of presynaptic terminals (asterisks) establishing synaptic contact on ChAT-positive dendrite (D) shown in a. (c, d) Examples of what appear to be asymmetric synapses (arrows) formed with ChAT-positive dendrites (D) in the septal part of a co-culture.



in rGtro

Fig. 6. ChAT-positive fibers in the bridge and hippocampal region of the co-culture shown in Fig. 3. (a) Fine, beaded ChAT-positive fibers (arrows) in the bridge (framed area 2 in Fig. 3). cl, plasma clot surrounding the culture. (b) ChAT-positive fibers (arrows) invading the hippocampal culture (framed area 3 in Fig. 3). (c) ChAT-positive fibers (arrows) traversing the broad pyramidal layer (sp) in the b~pp~arn~~i culture (framed area 4 in Fig. 3).




Fig. 7. Choline@ synapses in hippocampal cultures co-cultivated with septum (DIV: 46). (a) ChATimmunoreactive terminal establishing symmetric synapse (arrow) on cell body in the pyramidal layer of the culture. (b, c) ChAT-immunoreactive terminals establishing symmetric contacts (arrows) on dendritic shafts (D). (d) ChAT-positive bouton forming symmetric contact (arrow) on the neck portion of a somatic spine (s).


neurons 01 rim


Could it be that the septal culture itself supplied trophic factor? It has in fact been shown that some neurons in the septal region express the mRNA for A large number of in ciao studies have provided NGF.” Moreover. there is evidence that trophic evidence that septohippocampal neurons are dependent on target-derived trophic factor.‘.6.‘~,‘0.“‘-44 molecules are enriched in CNS lesions.” Weskamp et ~1.~~ found that the postlesional increase in NGF Previous studies thus suggested that the septal cholinwas more dramatic in the septal region than in the ergic neurons die in the absence of trophic factor hippocampus following fimbriaafornix transection. from the target region since transection of the The preparation of the septal slice cultures in the fimbria-fornix resulted in a loss of large cells and in present study similarly results in the transection ol a decrease of ChAT-immunopositive neurons in the septohippocampal fibers which could induce ncuroseptal region.” This conclusion gained support from trophin expression. At present we do not know studies in which the axotomized septohippocampal to what extent glial cells displaying a high proliferaneurons were treated with NGF or basic fibroblast tive activity mainly on the cut surfaces of the slice‘ growth factor (bFGF) in order to substitute for are involved in the expression of neurotrophic molthe loss of trophic factor from target cells. Under ecules. It should bc pointed out in this context that these conditions a large number of septal neurons the slice cultures of the present study were not could be rescued as demonstrated by Nissl stain. subjected to an antimitotic treatment in order to AChE histochemistry and ChAT immunocytochemreduce glial proliferation as routinely done in other istry.‘.X.~‘,‘i.‘4.“.~‘,4’It was concluded that the survival laboratories (Gahwiler, personal communication). of septohippocampal cholinergic neurons is depenAlso, neurotrophins may be contained in the serum dent on trophic molecules supplied by the target used for cultivation, although the serum was heat-inregion. A trophic effect of NGF on septal cholinergic activated. Preliminary studies with the polymerase neurons in slice cultures. accompanied by a remarkchain reaction from this group have at least indicated able increase in ChAT-activity, has been demonthat the mRNA for NGF is expressed in single strated by Gahwiler and co-workers.‘“,” slice cultures of the septal region (Forster CJI rrl.. More recent studies have accumulated evidence unpublished observations). that cholinergic neurons do not die in the absence of target-derived trophic factor but survive, although in Trunvnitter identit~~ of the septo/2ippoc~un~pcrI pro a shrunken state. However, they fail to express the jection in vitro acetylcholine-synthesizing enzyme ChAT. Sofroniew and his associates demonstrated that the almost It is well known that the fimbria, which normally complete excitotoxic destruction of the hippocampus connects the septum and the hippocampus, contains septohippocampal GABAergic fibers,’ commissural did not result in a loss of cholinergic neurons in the septal region. 4’ Moreover, the mRNA for NGF in the fibers from the contralateral hippocampus as well as hippocampal remnant decreased to almost undetecthippocamposeptal axons in addition to the cholinable levels.42 In line with this, Peterson et al.j’ showed ergic fibers. Except the commissural fibers, all other that many septohippocampal neurons that were rettypes of fibers may be present in septohippocampal co-cultures as well. Here we have to ask whether the rogradely labeled prior to fimbria-fornix transection could still be found in the medial septal nucleus and bridge formed between the two cultures really connucleus of the vertical limb of the diagonal band of tains cholinergic fibers originating from the septal Broca after extended periods of time following axoneurons. Histochemistry for AChE. as first applied tomy. More recently. Peterson et al.3’ were able to by Gahwiler and Brown” and Gahwiler et cd.,” does demonstrate that many of these retrogradely labeled not exclude the labeling of fibers other than the cholinergic ones. Our tracing experiments with Dil neurons surviving axotomy displayed a normal finedid not result in the retrograde labeling of many structure and normal input synapses. It appears that hippocampal neurons following tracer application to many septohippocampal cholinergic neurons survive the septal co-culture. Since many fibers can be labeled in the absence of target-derived trophic factor. anterogradely this way, the majority of axons in the Trophic factor from the target region may, however, bridge might derive from septal cells. Moreover. DiI be important for the expression of ChAT in these neurons. tracing revealed two types of fibers that have similarly been labeled by anterograde tracing with PHA-L.” By lmmunostaining for ChAT of many neurons in single slice cultures of the septal region is not using a double-labeling procedure, these authors were able to demonstrate that the thick fibers (their type I sufficiently explained by these studies. We have fibers) were derived from septohippocampal GABAto ask why these cells did not cease to express ergic cells. In line with this, the fibers stained for ChAT, although they were disconnected from their ChAT in our immunocytochemical experiments hippocampal target neurons supplying trophic closely resembled the thin (type II) fibers assumed to factor. It has been pointed out in the Results that be cholinergic by Freund and Antal.’ Experiments ChAT-immunopositive neurons were present in are in progress in which the hippocamposeptal fibers single cultures of septum, regardless of whether are anterogradely labeled by tracer injections into the medium was supplemented with NGF or not. Surriwd

of’ septal



in culture



the hippocampal explant to get an idea about the density of this projection in the co-cultures. Similarly, ongoing experiments employing the anterogradely transported tracer PHA-L in combination with postembedding GABA immunocytochemistry will reveal the extent of a septohippocampal GABAergic projection in this in vitro system. Here we provide direct evidence that ChAT-immunopositive, supposedly cholinergic fibers are actually present in the bridge connecting both slices (Fig. 6a). These fibers invade the hippocampal co-culture and form a threedimensional network reminiscent of that observed in situ. Like in situ, these cholinergic axons are thin, varicose, and are found in all hippocampal layers. At present we do not have an explanation for the large swellings in the course of some ChATpositive fibers (see Fig. 6b, c). Immunoreactive fibers are absent, however, in single cultures of hippocampus. Cholinergic synapses in the hippocampa~ co -culture

The main aim of the present study was to identify the types of synapses formed by the axons of septohippocampal cholinergic neurons. In the hippocampal co-culture we found ChAT-immunoreactive synapses that were very similar to those described in a number of in situ studies on the hip~campal fo~ation.5~iZ~‘3~15.45 As ob,s@ved in perfusion-fixed material, ChAT-positive terminals formed synaptic contacts with the cell bodies, dendritic shafts, and spines of the hippocampal target neurons, probably pyramidal cells and non-pyramidal neurons. Since ultrastructurai characteristics of cultured neurons may change under in vitro conditions as described recently for dentate granule cells,~ the identification of the various types of target cells of cholinergic


septohippocampal terminals in vitro requires additional double-labeling experiments (see Ref. 15). However, the various types of cholinergic synapses in the hippocampal co-culture strongly suggest that there is a similar, complex cholinergic innervation under the present in vitro conditions as described for the hippocampus in situ”-‘3.‘5 and for a number of other brain regions.27.45 It was a crucial question whether the cholinergic synapses in the hippocampal culture were in fact formed by septohippocampal neurons and not by intrinsic cholinergic cells. I4 Single hippocampal cultures stained for ChAT did, however, never reveal any staining of fibers and synapses. Our present in vitro studies are thus in agreement with previous ivr vivo experiments in which we showed that transection of the fimbria-fornix resulted in a virtually complete, permanent loss of ChAT-positive fibers in the hippocampus. lo In these in vivo experiments, the few ChAT-positive hippocampal neurons did not compensate by sprouting for the lack of cholinergic input from the medial septum. It was concluded that the ChAT-positive hippocampal neurons do not contribute to the cholinergic innervation of the hippocampus and fascia dentata.” Interestingly, unequivocally immunopositive neurons were not observed in slice cultures of hippocampus. Our data thus suggest that the ChAT-immunopositive synapses, exclusively found in hippocampal cultures co-cultivated with septum, are in fact formed by the cholinergic neurons in the septal co-culture. Acknowledgemgn~s-ale authors wish to thank U. Geiger, C. Hofmann, and M. Winter for technical assistanee. This study was supported by the Deutsche Forsehungsgemeinschaft (Fr 62012-4 and SFB 325).

REFERENCES 1. Adams J. C, (1981) Heavy metal ~n~nsifi~tion of DAB-based HI@ reaction product. J. H&ofhem. Cytochem. 29,775. 2. Anderson K. J., Dam D., Lee S. and Cotman C. W. (1988) Basic fibrobiast growth factor prevents death of lesioned cholinergic neurons in viva. Nature 332, 360-361. 3. Barde Y.-A. (1989) Trophic factors and neuronal survival. Neuron 2, 1525-1534. neurons and terminals in the rat 4. Bialowas J. and M. Frotscher M. (1987) Choline acetyltransferase-immunoreactive septal complex: a combined light and electron microscopic study. J. camp. Neural. 259, 298-307. 5. Clarke D. J. (1985) Cholinergic innervation of the rat dentate gyrus: an immunocytochemical and electron microscopical study. 3rain Res. 360, 349-354. 6. Cowan W. M., Fawcett J. W., O’Leary D. D. M. and Stanfield B. B. (1985) Regressive events in neurogenesis. Science 225, 1258-1265. 7. de1 Rio J. A., Heimrich B., Soriano E., Schwegler H. and Frotscher M. (1991) Proliferation and differentiation of glial fibrillary acidic protein-immunoreactive glial cells in organotypic slice cultures of rat hippocampus. Neuroscience43, 335-347. 8. Fischer W. and Bjiirklund A. (1991) Loss of AChE- and NGFr-labeling precedes neuronal death of axotomized septal-diagonal band neurons: reversal by intraventricular NGF infusion. Expl Neural. 113.93-108. 9. Freund T. F. and Antal M. (1988) GABA-containing neurons in the septum control inhibitory interneurons in the ~p~p~. Nature 336, 17&173. 10. Frotscher M. (1988) Cholinergic neurons in the rat hippocampus do not compensate for the loss of ~ptohipp~ampal cholinergic fibers. Neurosci. Let?. 87, 18-22. II. Frotscher M. (1989) Central cholinergic synapses: the septohippocampal system as a model. In Central Cholinergic Synaptic Trun.rmisspsion (eds Fro&her M. and Misgeld U.), pp. 3341. Birkhtiuser, Base]. 12. Frotscher M. and Leranth C. (1985) Choline&z innervation of the rat hippocampus as revealed by choline acetyltransferase immunocytochemistry: a combined light and electron microscopic study. J. camp. Neural. 239,237-246. 13. Frotscher M. and Leranth C. (1986) The cholinergic innervation of the rat fascia dentata: identification of target structures on granule cells by combining choline a~yltran~era~ i~un~toche~st~ and Go&i jrnp~~ation. J. camp. Neurol. 243, 58-70.


neurons in aitro


14. Fro&her M., Schlander M. and Leranth C. (1986) Cholinergic neurons in the hippocampus: a combined light and electron microscopic immunocytochemical study in the rat. Cell Tiss. Res. 246, 2933301. 15. Frotscher M., Nitsch R. and Leranth C. (1989) Cholinergic innervation of identified neurons in the hippocampus: electron microscopic double labeling studies. In The Hippocampus-New Visras (eds Chan-Palay V. and Kohler C.), pp. 85-96. Alan R. Liss, New York. 16. Gghwiler B. II. (1981) Organotypic monolayer cultures of nervous tissue. 1. Neurosci. Meth. 4, 3299342. 17. Gahwiler B. H. (1984) Development of the hippocampus in oirro: cell types, synapses and receptors. Neuro~~~ien~e11, 751-760. 18. Gahwiler B. H. and Brown D. A. (1985) Functional innervation of cultured hippocampal neurons by cholinergic afferents from co-cultured septal explants. Nature 313, 577.-579. 19. Gahwiler B. H., Brown D. A., Enz A. and Knijpfel T. (1989) Development of the septohippocampal projection in rirro. In Central Cholinergic Synaptic Transmission (eds Frotscher M. and Misgeld U.), pp. 236250. Birkhluser, Base].

20. Gahwiler B. H., Rietschin L., Kniipfel T. and Enz A. (1990) Continuous presence of nerve growth factor is required for maintenance of cholinergic septal neurons in organotypic slice cultures. Neuroscience 36, 27-31. 21. Gage F. H., Tuszynski M. H., Chen K. S., A~strong D. and Buzsiki G. (1989) Survival, growth and function of damaged cholinergic neurons. In Central Cholinergie Synaptic Tran.~mjssion (eds Frotscher M. and Misgeld U.), pp. 259-274. Birkhluser, Base]. 22. Gnahn H., Hefti F., Heumann R., Schwab M. E. and Thoenen H. (1983) NGF-mediated increase of choline acetyltransferase (ChAT) in the neonatal rat forebrain: evidence for a physiological role of NGF in the brain? Detll Bruin Res. 9, 45-52. 23. Hagg T., Manthorpe M., Vahlsing H. L. and Varon S. (1988) Delayed treatment with nerve growth factor reverses the apparent loss of cholinergic neurons after acute brain damage. Expf Neurof. 101, 303-312. 24. Hefti F. (1986) Nerve growth factor promotes survival of septal cholinergic neurons after fimbriat transections. 1. Neurosci. 6, 21552162. 25. Heimrich B. and Frotscher M. (1991) Differentiation of dentate granule cells in slice cultures of rat hippocampus: Golgi/electron microscopic study. Brain Res. 538, 263-268.


26. Honig M. G. and Hume R. I. (1989) DiI and DiO: versatile fluorescent dyes for neuronal labeling and pathway tracing. Trends Neurosci. 12, 333-341. 27. Houser C. R. (1990) Cholinergic synapses in the central nervous system: studies of the immunocytochemical localization of choline a~tyltransferase. J. Eleczron. Micrusc. ~ech~~q~e IS, 2-19. 28. Houser C. R., Crawford G. D., Barber R. P., Salvaterra P. M. and Vaughn J. E. (1983) Organization and morphological characteristics of cholinergic neurons: an immunocytochemical study with a monoclonal antibody to choline acetyltransferase. Brain Res. 266, 97-119. 29. Ishikawa R., Nishikori K. and Furukawa S. (1991) Appearance of nerve growth factor and acidic fibroblast growth factor with different time courses in the cavity-lesioned cortex of the rat brain. Neurosci. Left. 127, 70-72. 30. Korsching S., Auburger G., Heumann R., Scott J. and Thoenen H. (1985) Levels of nerve growth factor and its mRNA in the central nervous system of the rat correlate with cholinergic innervation. Eur. molec. Biol. Org. J. 4, 1389--1393. 31. Kromer L. F. (1987) Nerve growth factor treatment after brain injury prevents neuronal death. Science 235,214--216. 32. Lauterborn J. C., Isackson P. J. and Gall C. M. (1991) Nerve growth factor mRNA-containing cells are distributed within regions of cholinergic neurons in the rat basal forebrain. J. camp. Neuroi. 306, 439-446. 33. Leranth C. and Frotscher M. (1987) Cholinergic innervation of hippocampal GAD- and somatostatin-immunoreactive commissural neurons. J. camp. Neurol. 261, 3347. 34. Lewis P. R. and Shute C. C. D. (1967) The cholinergic limbic system: projections to hippocampal formation, medial cortex, nuclei of the ascending cholinergic reticular system and the subfornical organ and supraoptic crest. Brain 90, 521~-540. 35. Mesulam M.-M., Mufson E. J., Wainer B. H. and Levey A. I. (1983) Central cholinergic pathways in the rat: an overview based on an alternative nomenclature (Chl-Ch6). Neuroscience 10, 1185-1201. 36. Mesulam M.-M., Geula C. and Moran M. A. (1987) Anatomy of cholinesterase inhibition in Alzheimer’s disease: effect of physostigmine and tetrahydroaminoacridine on plaques and tangles. Ann. Neurol. 22, 683-691. 37. Otto D., Frotscher M. and Unsicker K. (1989) Basic tibroblast growth factor and nerve growth factor administered in gel foam rescue medial septal neurons after fimbria-fornix transection. J. Neurosci. Res. 22, 83-91. 38. Peterson G. M., Lanford G. W. and Powell E. W. (1990) Fate of septohippocampal neurons following fimbria-fornix transection: a time course analysis. Brain Res. Bull. 25, i299137. 39. Peterson G. M., Naumann T. and Fro&her M. (1992) Identified septohippocampal neurons survive axotomy: a fine-structural analysis in the rat. Neurosci. Lerr. 138, 81-85. 40. RimvaIl K., Keller F. and Waser P. G. (1985) Development of cholinergic projections in organotypic cultures of rat septum, hippocampus and cerebellum. Deul Brain Res. 19, 267-278. 41. Sofroniew M. V., Galletly N. P., Isacson 0. and Svendsen C. N. (1990) Survival of adult basal forebrain cholinergic neurons after loss of target neurons. Science 247, 338-342. 42. Sofroniew M. V., Cooper J. D., Svendsen C. N., Crossman P., Ip N. Y., Lindsay R. M., Zafra F., Cast& E.. Thoenen H. and Lindhoim D. (1991) Long-term survival of septal cholinergic neurons after lesions that deplete the hippocampus of cells producing NGF or BDNF mRNA. Soioe.Neurosci. Absfr. 17, 221. 43. Thoenen H. (1991) The changing scene of neurotrophic factors. Trends Neurosci. 14, 165-170. 44. Thoenen H. and Barde Y.-A..(1980) Physiology of nerve growth factor. Physiof. Rev. 60, 12841335. 45. Wainer B. H., Bolam J. P., Freund T. F.. Henderson Z.. Totterdell S. and Smith A. D. (19841 Cholinereic svnauses in the rat brain: a correlated light and electron microscopic immunohistochemical study employing a “mono&ma1 antibody against choline acetyltransferase. Brain Res. 308, 69-76. 46. Weskamp G., Gasser U. E., Dravid A. R. and Otten U. (1986) Fimbria--fornix lesion increases nerve growth factor content in adult rat septum and hippocampus. Neurosei. Letf. 70, 121-126. 47. Williams L. R., Varon S., Peterson G. M., Wictorin K., Fischer W., Bjorklund A. and Gage F. H. (1986) Continuous infusion of nerve growth factor prevents basal forebrain neuronal death after timbria fornix transection. Proc. nom. Acad. Sri. U.S.A. 83, 9231-9235. (Accepted 14 Seprember