GABA-immunoreactive basal forebrain afferents innervate GABA-immunoreactive non-pyramidal cells in the cerebral cortex of the lizard podarcis hispanica

Neuroscience Vol. 51, No. 2, pp. 4 2 5 4 3 7 , 1992

0306-4522/92 $5.00 + 0.00 Pergamon Press Ltd © 1992 IBRO

Printed in Great Britain

GABA-IMMUNOREACTIVE BASAL FOREBRAIN AFFERENTS INNERVATE GABA-IMMUNOREACTIVE N O N - P Y R A M I D A L CELLS IN THE CEREBRAL CORTEX OF THE LIZARD P O D A R C I S HISPANICA F. J. MARTiNEz-GuIJARRO*~'~ and T. F. FREUND* *Institute of Experimental Medicine, Hungarian Academy of Sciences, Budapest, POB 67, H-1450 Hungary tUnidad de Biologia Celular, Facultad de Ciencias Biol6gicas, Universidad de Valencia, Burjasot, Spain Abstract--The basal forebrain projection to the cerebral cortex was studied in the lizard Podareis hispanica by anterograde transport of Phaseolus vulgaris leucoagglutinin. After injections of the lectin into the septal-basal forebrain area, Phaseolus vulgaris leucoagglutinin-labelled fibres were mainly detected in the outer plexiform layer of the medial cortex and in the inner plexiform layer of the dorsal and dorsomedial cortices. Ultrathin sections from these areas were obtained and processed for postembedding immunogold staining for GABA. Most of the Phaseolus vulgaris leucoagglutinin-labelled boutons in the dorsal and dorsomedial cortical areas were GABA immunoreactive and all the double-labelled boutons established symmetric synaptic contacts on cell bodies and dendrites that were also found to be GABA immunoreactive in all cases. In contrast, Phaseolus vulgaris leucoagglutinin-labelled varicosities in the outer plexiform layer of the medial cortex made asymmetric synaptic contacts on GABA-immunonegative profiles and they were themselves negative for GABA. In double-labelled sections, GABA-, calbindin D28kand neuropeptide Y-immunoreactive neurons were found to be innervated by multiple Phaseolus vulgaris leucoagglutinin-labelled varicosities in the dorsal and dorsomedial cortical areas, whereas in the medial cortex Phaseolus vulgaris leucoagglutinin-labelled fibres were not observed in contact with any subpopUlation of GABAergic cells. The results demonstrate that in lizards the septaPbasal forebrain projection to the cortex has a GABAergic component, which selectively terminates on GABAergic non-pyramidal cells including the neuropeptide Y- and the calbindin D28k-containing subpopulations. This synaptic organization is remarkably similar to that in mammals, and suggests that the mechanisms of control of the cortical activity by the basal forebrain have been highly preserved during phylogeny.

The cerebral cortex of lizards is formed by four areas, namely the medial (MC), dorsomedial (DMC), dorsal (DC) and lateral (LC) cortices. All of these areas show a three-layered pattern with an intermediate cell layer located between the outer and inner plexiform layers. This lamination pattern is similar to that of the m a m m a l i a n archicortex. 34 The cerebral cortex of lizards has been suggested to correspond to the mammalian hippocampal formation on the basis of their similar cytoarchitecture, chemoarchitecture, connectivity and ontogeny. 24 Retrograde tracing experiments have demonstrated that neurons located in the septum, z22 and in the basal forebrain (vertical limb of the diagonal band of

:~To whom correspondence should be addressed at: Unidad de Biologia Celular, Facultad de Ciencias Biol6gicas, Universidad de Valencia, c/o Dr Moliner, 50, 46100 Burjasot Spain. Abbreviations: CaBP, calbindin-D2sk; CHAT, choline acetyltransferase; DAB, 3,Y-diaminobenzidine; DC, dorsal cortex; DMC, dorsomedial cortex; LC, lateral cortex; MC, medial cortex; NPY, neuropeptide Y; PHAL, Phaseolus vulgaris leucoagglutinin; TBS, Tris-buffered saline. NSC 51/2--H

425

Broca and bed nucleus of the medial forebrain bundle) 2 project to the cerebral cortex in lizards. Recently, these nuclei have been reported to contain choline acetyltransferase (CHAT)- 16 as well as G A B A - i m m u n o r e a c t i v e neurons. 12,38 In mammals, the septohippocampal pathway is known to have a cholinergic and a GABAergic component. 18'19'31'45 The former terminates both on principal cells and on non-pyramidal cells, whereas the latter was demonstrated to innervate predominantly G A B A e r g i c non-pyramidal cells. 8'13'14a Moreover, the G A B A e r g i c component of the septohippocampal pathway w a s found to terminate on GABAergic interneurons involved in perisomatic inhibition as well as on those mediating distal dendritic inhibition. 8,13 The similarities and presumed homology of the lizard cortex with the m a m m a l i a n hippocampus and the occurrence of both cholinergic and GABAergic neurons in the septal-basal forebrain nuclei projecting to the cortex, support the hypothesis that in lizards, as in mammals, the septocortical pathway may also have a dual, cholinergic and GABAergic, nature.

426

F.J. MARTiNEz-GU1JARROand T. F. FREUND

The aim o f this work was to investigate w h e t h e r the projection from the septal-basal forebrain region to the cerebral cortex has a G A B A e r g i c c o m p o n e n t in lizards, and, if so, w h e t h e r this projection terminates selectively o n cortical G A B A e r g i c n o n - p y r a m i d a l cells. In lizards, different subsets of G A B A e r g i c n e u r o n s have been recently described. 5'27'29 Similarly to m a m mals, these different subsets include n e u r o n s involved in perisomatic inhibition as well as n e u r o n s m e d i a t i n g distal dendritic i n h i b i t i o n Y U s i n g the calcium binding protein c a l b i n d i n D28k (CaBP) a n d n e u r o p e p t i d e Y ( N P Y ) as m a r k e r s we also investigated the types of G A B A e r g i c n e u r o n s receiving the s e p t a l - b a s a l forebrain input. EXPERIMENTAL PROCEDURES

Four adult male lizards of the species Podarcis hispanica, captured in the surroundings of our laboratory, were used in this study. Under ether anaesthesia, Phaseolus vulgaris leucoagglutinin (PHAL) (2.5%, Vector) was iontophoretically injected H in the septal-basal forebrain area. Three days later, animals were perfused through the heart under ether anaesthesia, first with saline (0.9% NaC1) for 1 min and then with 30 ml of fixative containing 1% glutaraldehyde, 2.5% paraformaldehyde and 0.2% picric acid in 0.1 M phosphate buffer, pH 7.4, for 30 min. Brains were removed and transversely cut at 60/~m on a Vibratome. After extensive washes in 0.1 M phosphate buffer, pH 7.4, the sections were immersed in 0.05 M phosphate buffer, pH 7.4, containing 25% sucrose and 10% glycerol, and freeze-thawed in liquid nitrogen to enhance penetration of the antisera. The sections were treated with 1% NaBH 4 for 30 min to reduce free aldehyde groups and double bonds. 2°

Immunocytochemistry Double immunostaining46 was used to visualize PHALlabelled fibres and the GABA-, the CaBP- or the NPY-immunoreactive cells in the same sections. First, the sections were incubated in a mixture of the primary antibodies for two days; biotinylated goat anti-PHAL (1:200, Vector) was mixed either with mouse anti-GABA41 (1:50) or rabbit anti-CaBW a (1:1000) or rabbit anti-NPY 4 (I:1000). The second layer (overnight) was a mixture of avidinbiotinylated horseradish peroxidase complex (l:100, Vector) and rabbit anti-mouse IgG (1:50, Dakopatts; for the PHAL/GABA combination) or goat anti-rabbit IgG (1 : 50, ICN; for the PHAL/CaBP and PHAL/NPY combinations). PHAL-labelled axons were visualized with a 3,3'-diaminobenzidine (DAB) reaction intensified by ammoniumnickel sulphate (black reaction product). In some sections, which were selected for single PHAL immunostaining for further electron microscopy study, peroxidase reaction was developed using DAB alone. The third layer was mouse

peroxidase-antiperoxidase complex (1:100, Dakopatts; for GABA) or rabbit peroxidase-antiperoxidase complex (1:100, Dakopatts; for CaBP and NPY). The CaBP- and NPY-immunoreactive cells were visualized using DAB alone, which gives a brown end-product. For the dilution of the antisera a n d for the washes 50 mM Tris-buffered saline (TBS; pH 7.4) containing 1% normal rabbit (for GABA immunoreaction) or 1% normal goat (for CaBP and NPY immunoreaction) serum was used. In the case of sections processed for light microscopy the solutions contained 0.5% Triton X-100 to enhance penetration of the antibodies. The sections for light microscopy were thoroughly washed, mounted on gelatine-coated slides and covered with XAM neutral medium. Sections selected for electron microscopy were osmium postfixed, dehydrated and fiatembedded in Durcupan (ACM, Fluka). Sections immunostained only for PHAL were re-embedded for ultrathin sectioning. The sections were collected on nickel Formvarcoated single-slot grids and were processed for immunogold staining for GABA.

Postembedding immunogold staining for GABA The immunogold staining procedure followed that described by Somogyi and Hodgson 4° with small modifications using a well-characterized antiserum against GABA. ~4bThe steps were carried out on droplets of Millipore-filtered solutions in humid Petri dishes, as follows: 1% periodic acid (H5IO6) for 10 min; wash in several changes of double-distilled water, 2% sodium metaperiodate (NalO 4 BDH) for 15 min; wash in several changes of double-distilled water, three times 2 min in TBS (pH 7.4); 30 min in 1% ovalbumin disolved in TBS; three times 10 min in TBS containing 1% normal goat serum; 2 h in a rabbit anti-GABA antiserum (code no. 914b)diluted 1:3000 in TBS containing 1% normal goat serum; two times 10 min TBS; 10 min in 50 mM Tris buffer (pH 7.4) containing 1% bovine serum albumin and 0.5% Tween 20; goat anti-rabbit IgG-coated colloidal gold (15 nm, BioClin) for 2 h (diluted 1:20 in the same buffer); two times 5 min wash in double-distilled water; saturated uranyl acetate for 30 min; wash in four changes of doubledistilled water; staining with lead citrate; wash in distilled water. RESULTS

Location o f injection sites and distribution o f Phaseolus vulgaris leucoagglutinin-labelled axons in the cerebral cortex The injection sites consist of a central h o m o geneous core s u r r o u n d e d by cells labelled in a Golgilike m a n n e r . The core o f the injection was located in the vertical limb of the diagonal b a n d of Broca, the bed nucleus o f the medial forebrain bundle and the precomissural septum in each animal (Fig. 1A). Golgi-like labelled cells were observed t h r o u g h o u t the whole extension of the septum.

Fig. I. Dark-field microphotographs of transverse sections of the telencephalon of the lizard Podarcis hispanica after PHAL injection into the septal-basal forebrain area. (A) Precommissural level showing the injection site. The bulk of the injection (asterisk) affected the vertical limb (white dots) of the diagonal band nucleus (DB) and the bed nucleus of the medial forebrain bundle (bMFB). The diffusion area involved the septum (SPT) and the horizontal limb (black dots) of the diagonal band nucleus. PHAL-labelled fibres are seen in the outer plexifom layer of the MC, DC and DMC areas (small arrows) and in the inner plexiform layer of the DC and DMC (large arrow). The cell layer (el) contains only a few fibres. The LC receives a sparse innervation. (B, C) Commissural and post-commissural telencephalic levels showing the distribution of PHAL-immunoreactive fibres. Labelled fibres are especially concentrated in the outer plexiform layer of the MC (small arrows) and in the inner plexiform layer of the DC (large arrow). Scattered fibres are also seen in the DMC. D x 65.

Fig. 1 427

428

F.J. MARTiNEz-GU1JARROand T. F. FREUND

At precommissural levels, PHAL-labelled fibres were observed in all cortical areas (Fig. 1A). They appeared in the outer plexiform layer of the MC, DC and DMC and in the inner plexiform layer of the DC and DMC. Some fibres were also observed in the superpositio lateralis between the DC and LC. At commissural and post-commissural levels (Fig, tB, C), fibres were mainly detected in the outer plexiform layer of the MC and in the inner plexiform layer of the DC and DMC. They were especially concentrated in the inner plexiform layer of the DC, where they displayed abundant varicosities. PHAL-labelled varicosities were not found in close association with somata of the cell layer where mainly principal cells are located. By contrast, they surrounded cell bodies of neurons located in the inner or in the outer plexiform layers (Fig. 5). The number of boutons forming a basket arround a single target cell ranged between four and 12. GABA-immunoreactive Phaseolus vulgaris leucoagglutinin-labelled boutons synapse on GABAimmunoreactive postsynaptic structures At the electron microscopic level, PHAL-labelled structures were found to include both myelinated and non-myelinated axonal segments as well as synaptic boutons. They were identified by their electron-dense appearance, which is due to the DAB deposit in the axoplasm, on the synaptic vesicles and on the outer membrane of mitochondria. To ascertain whether the septal-basal forebrain afferents to the cerebral cortex included GABAergic fibres, ultrathin sections from the inner plexiform layer of the DC and DMC (the zone with the highest concentration of PHAL-labelled varicose fibres) were obtained and processed for postembedding immunogold staining for GABA. After G A B A immunostaining, most of the PHALlabelled profiles showed immunogold labelling. Gold particles were almost restricted to mitochondrial profiles where the DAB deposit was absent. In addition, cell bodies, dendrites and synaptic boutons displaying G A B A immunoreactivity but without PHAL labelling were frequently found. GABAimmunonegative profiles were also observed, and corresponded to cell bodies, spiny dendrites, synaptic boutons making asymmetric synaptic contacts and glial processes. Forty-five PHAL-labelled GABA-immunoreactive synaptic boutons were analysed in serial ultrathin sections, and the postsynaptic profiles were found to be G A B A immunoreactive in all cases. The postsynaptic structures included cell bodies (Fig. 2) as well as thin (Fig. 3A-C) and thick (Fig. 3D, E) dendritic shafts. Synaptic contacts established by the PHAL-labelled GABA-immunoreactive boutons were always of the symmetric type. Ultrathin sections taken from the outer plexiform layer of the MC were also processed for postembed-

ding immunogold staining for GABA. All the PHALlabelled boutons analysed in this layer were immunonegative for GABA, and established asymmetric synaptic contacts on GABA-immunonegative profiles (Fig. 4). They showed ultrastructural features (e.g. larger synaptic vesicles and rare mitochondria) different from those of the GABA-immunoreactive boutons found in the inner plexiform layer of the DC and DMC. Distribution o f GABA-, neuropeptide Y- and calbindin D28k-immunoreactive neurons innervated by septal afferents In order to map the distribution of GABAergic neurons receiving septal-basal forebrain input, and to examine which subsets of GABA-immunoreactive neurons are the targets of this pathway, alternate Vibratome sections (60/~m thick) were doublestained for P H A L and either GABA, NPY or CaBP (Fig. 5). The distribution of GABA-, NPY- and CaBPimmunoreactive cells observed in these experiments was in agreement with earlier studies. 5'27'36'37'42 GABA-immunoreactive cells were found throughout the cerebral cortex. The highest density of GABA-immunoreactive cell bodies in the MC, DC and DMC was found in the inner plexiform layer, followed by the outer plexiform layer. In the cell layer, GABA-immunoreactive cells were sparse and usually appeared near to the upper or lower rims of the layer. In the LC, GABA-immunoreactive cells were mainly located in the outer plexiform and cell layers, but they were rare in the inner plexiform layer. Punctate structures equivalent to axon terminals were also stained for GABA in large numbers concentrated around principal cell bodies, but they were also numerous in the plexiform layers. NPY-immunoreactive cells were detected in all cortical areas, increasing their frequency from rostral to caudal levels. In the MC, DC and DMC they were most frequently found in the inner plexiform layer, although in the DC and DMC they also occurred in the outer plexiform and cell layers. In the LC they were mainly located in the outer plexiform and cell layers. NPY-immunoreactive puncta resembling axon terminals were detected in the superficial portion of the outer plexiform layer in the MC, DC and DMC, and in the outer plexiform and cell layers of the LC. CaBP-immunoreactive cells appeared in the DC, DMC and LC, but were absent in the MC. In the DC and DMC they were most frequently found in the inner plexiform layer and less frequently in the outer plexiform and cell layers. CaBP-immunoreactive cells were never observed in the most superficial part of the outer plexiform layer in the DMC and DC. The number of CaBP-immunoreactive cells in these two cortical areas increased at caudal levels. In the LC, CaBP-immunoreactive cells were observed predominantly in the outer plexiform and cell layers.

Basal forebrain input to lizard cortical GABAergic neurons

Fig. 2. Electron micrographs of two (A, B) adjacent G A B A immunogold-stained ultrathin sections, showing a PHAL-immunoreactive bouton making a symmetric synaptic contact (large arrow) on a cell body (Cb6ABA). Both the cell body and the PHAL-labelled bouton show an accumulation of gold particles, and are therefore considered G A B A immunoreactive. The asterisk labels a GABA-negative bouton. Scale bars = 0.25 # m .

429

430

F.J. MARTiNEz-GUIJARROand T. F. FREUND

In double-labelled sections a differential innervation of GABAergic neurons was observed in the different cortical areas. In the MC, PHAL-labelled fibres were abundant in the outer plexiform layer, but they were not observed in contact with the GABAergic cell bodies located in this layer. In contrast, in the DMC and especially in the DC, GABA-, NPYand CaBP-immunoreactive cells were frequently found to be innervated by multiple PHAL-labelled varicosities. Most of these target cells were located in the inner plexiform layer, where the density of PHAL-labelled varicose fibres was higher, although they were also observed in the outer plexiform layer. The innervated cell body was usually surrounded by several boutons, but proximal and distal dendritic segments were also frequently followed by PHALlabelled varicose fibres. DISCUSSION The major findings of this study are that (i) the septal-basal forebrain projection to the cerebral cortex of lizards has a GABAergic component, and (ii) this GABAergic component selectively terminates on GABAergic non-pyramidal cells which include the N P Y - and the CaBP-containing subpopulations.

The origin of Phaseolus vulgaris leucoagglutininlabelled fibres in the cortex PHAL is is known to be transported anterogradely by intact axons if delivered iontophoretically, ll but in some systems it may be taken up and transported by fibres of passage. 3 However, in mammals this transport has been described to be of shorter range; the labelled axons were small in number and stained more weakly than those labelled following uptake of the lectin by the cell body or dendrites. In the case of lizards the distance between the injection site and the projection fields is very short (1-2mm). Thus, the possibility of anterograde labelling as a result of uptake by fibres of passage has to be considered. There are several possible sources of subcortical fibres reaching the cortex, and most of these are known to project through the injection site; e.g. different nuclei located in the thalamus, hypothalamus and raphe, the locus coeruleus and the reticular formation. 44Moreover, the terminal pattern observed in the outer plexiform layer of the MC in the present study matches that reported for terminal degeneration after electrolytic lesion in the septum. These lesions also interrupted the fibres of passage coming from the dorsolateral anterior nuclei of the thalamus,

which pass through the medial forebrain bundle and terminate in the cortex. 17'22 Similar results were obtained after anterograde transport of horeseradish peroxidase injected into the thalamus. 25 Further support for the presumed thalamic origin of the PHAL-labelled boutons in the outer plexiform layer of the MC comes from the finding that they were always negative for GABA, and established asymmetrical synaptic contacts with GABA-negative profiles, mostly with dendritic spines (Fig. 4). None of the previously mentioned subcortical areas have been reported to contain GABA- or glutamate decarboxylase-immunoreactive neurons. It has been reported in crocodiles that thalamic nuclei projecting to the cortex do not contain GABAergic neurons at all. 33 The raphe nuclei are considered to be the origin of a serotonergic projection, whereas the locus coeruleus and some hypothalamic nuclei contain noradrenergic and dopaminergic projection neurons. 39 It has been reported that neurons in the reticular formation give rise to axons which terminate in the cortex forming asymmetric synaptic contacts, and contain round clear vesicles. 43 Thus, it is unlikely that any of these nuclei contribute to the GABAergic subcortical input to the cortex. Another technical aspect requiring discussion is the immunogold staining of the PHAL-labelled boutons. Immunostaining was largely restricted to the mitochondria and only in a few cases were gold particles observed over the synaptic vesicles or in the axoplasm. The same observation has been reported in mammals for PHAL-labelled fibres; i.e. the stronger the DAB reaction resulting from the pre-embedding staining, the weaker the immunogold labelling for GABA. 7 This may be due to the masking of epitopes by the reaction product from the immunostaining for PHAL. Therefore, GABA can only be detected in those sites of fibres and boutons which are devoid of DAB deposits, i.e. in the mitochondria. Thus, PHALlabelled boutons showing a selective accumulation of gold particles in the mitochondria were considered GABA immunopositive. Other GABA-immunopositive boutons which were not PHAL labelled, also showed a higher density of gold particles over the mitochondria than in the axoplasm. During perfusion, GABA is fixed by glutaraldehyde to basic amino acid side chains, which are most numerous in the nuclei (histone proteins), but also in any membranes. This may explain why GABA immunoreactivity is always higher in the nuclei and over mitochondria. On the other hand, boutons making asymmetric synaptic contacts, which are

Fig. 3. Electron micrographs of adjacent GABA immunogold-stained ultrathin sections, showing two PHAL-immunoreactive boutons, one making a symmetric synaptic contact (large arrow) on a thin dendrite (A-C), and the other (D, E) on a thick dendrite (d~ABA).The PHAL-labelled boutons as well as the dendrite are GABA immunoreactive. Asterisks mark GABA-negative boutons, some of them making asymmetric synaptic contacts on dendritic spines (arrowheads). GABA-positive, PHAL-negative boutons are labelled by stars. Scale bars = 0.25 pm.

Fig. 3 431

432

F . J . MARTirCEz-GUIJARROand T. F. FREUNO

Fig. 4. Electron micrographs of two (A, B) PHAL-immunoreactive GABA-negative boutons from the outer plexiform layer of the medial cortex. They make asymmetric synaptic contacts (open arrows) on GABA-immunonegative postsynaptic elements. GABA-positive, PHAL-negative profiles are labelled by stars. Scale bars = 0.25/am.

most unlikely to be GABAergic, never showed a n a c c u m u l a t i o n o f gold particles in any parts o f the terminal (Figs 2, 3).

The septal-basal forebrain projection to the cortex has a GABAergic component R e t r o g r a d e tracing studies d e m o n s t r a t e d the existence o f n e u r o n s located in the septum a n d basal

forebrain t h a t projected to the cortex. 2'22 Both cholinergic a n d G A B A e r g i c n e u r o n s are k n o w n to be present in these areas, ~2'~6'38suggesting that, similarly to the m a m m a l s , the septocortical pathway in the lizard m a y have b o t h a cholinergic a n d a G A B A e r g i c component. O u r results d e m o n s t r a t e the existence of a G A B A ergic pathway, which arises from n e u r o n s in the

Fig. 5. Light micrographs of cortical Vibratome sections double-stained for PHAL and NPY (a~l), PHAL and CaBP (e, f) or PHAL and GABA (g, h). PHAL-immunoreactive varicose fibres are stained in black (Ni-DAB reaction) while cell bodies, dendrites and fibres immunoreactive for NPY, CaBP and GABA are brown (DAB reaction). (a) Survey of the DMC at a caudal telencephalic level showing an NPY-immunoreactive cell body surrounded by PHAL-immunoreactive varicosities (open arrow). (b) High power light micrograph of the cell body shown in a, contacted by many PHAL-labelled varicosities (arrowheads). (c) A PHAL-immunoreactive fiber is climbing on an NPY-immunoreactive dendrite and forms varicosities (arrowheads) at the points of contact, in the outer plexiform layer of the DC. (d) Photomicrograph from the inner plexiform layer of the DC, a region similar to that indicated by a large arrow in Fig. 1C. A group of four NPY-immunoreactive cells are shown contacted by several PHAL-labelled varicosities (arrowheads). (e) Distal dendritic shaft of a CaBP-immunoreactive cell in the outer plexiform layer of the DMC is contacted by many PHAL-labelled varicosities (arrowheads). (f) Photomicrograph of the inner plexiform layer of the DC (a zone similar to that indicated by a large arrow in Fig. 1C) showing PHAL-labelled boutons (arrowheads) in contact with a CaBP-immunoreactive cell body. (g) Photomicrograph of the inner plexiform layer of the DC (a zone similar to that indicated by a large arrow in Fig. 1C) showing PHAL-labelled boutons (arrowheads) in contact with GABA-immunoreactive cell bodies. (h) GABA-immun0reactive cell body surrounded by PHAL-labelled varicosities (arrowheads). Scale bars = 15/am.

Basal forebrain input to lizard cortical GABAergic neurons

Fig. 5

433

434

F.J. MARTiNEZ-GUI~IARROand T. F. FREUND

septum-basal forebrain complex and projects predominantly to the DC and DMC. Further experiments of retrograde tracing combined with G A B A or glutamate decarboxylase immunostaining are required to identify the precise location of the parent cells of this pathway. Available data derived from studies using different experimental approaches point to the diagonal band of Broca and the septum as the most probable sources of the GABAergic projection. They fulfil the two prerequisits: they project to the cortex and have GABAergic cells. In contrast, the bed nucleus of the medial forebrain bundle has not been reported to contain GABAergic cells. However, in contrast to the results obtained in a related lizard species by other authors, 12 we were unable to find GABA-immunoreactive cells in the diagonal band of Broca in the present material. Nevertheless, it has been described in mammals that cell bodies in the medial septum-diagonal band of the Broca complex, giving rise to the GABAergic septohippocampal afferents, show very weak, if any, immunoreactivity for G A B A or glutamate decarboxylase,32 similarly to the somata of other GABAergic neurons with distant projections (e.g. commissural GABAergic neurons in the hippocampus, Purkinje cells, striatonigral cells, etc.). Thus, the possibility that the somata of GABAergic projection neurons contain a level of GABA just below the detection threshold of immunocytochemistry has also to be considered in lizards. In a recent study in the striate cortex of the cat, ChAT-immunoreactive boutons, most of them presumably originating in the basal forebrain, were found to have a moderate proportion (27.5%) of their postsynaptic elements immunoreactive for GABA. Furthermore, approximately 8% of these ChAT-immunoreactive boutons were GABAimmunoreactive as well. lb Thus the possibility that some of the PHAL-labelled GABA-immunoreactive boutons found in lizards are also cholinergic has to be considered. However, this possibility is highly unlikely, since, in lizards, the area with the highest density of PHAL-labelled GABA-immunoreactive boutons (the inner plexiform layer of the DC) lacks ChAT-immunoreactive fibres, 16whereas the area with the highest density of ChAT-immunoreactive fibres (the outer plexiform layer of the MC), although it contained PHAL-labelled boutons contained GABAnegative fibres. In addition, a major difference is that in the cat only 25% of the double-labelled CHAT- and GABA-immunoreactive boutons had GABA-immunoreactive postsynaptic elements, whereas in the lizard all the postsynaptic elements for PHALlabelled GABA-immunoreactive boutons were positive for GABA. Nevertheless, double-labelling studies for ChAT and G A B A should be performed in lizards to estimate the degree of co-localization, if any, in cortical boutons as well as in neurons of the basal forebrain.

The apparent lack of PHAL labelling of the cholinergic basal forebrain-cortical pathway is difficult to explain. Nevertheless, the same rather selective labelling of the GABAergic component of the same pathway has also been observed in the rat 9 and the cat; 1° and it has been attributed to the relatively larger axonal diameter, and/or to specific uptake mechanisms that may help the GABAergic pathway to acquire significantly more labelling by the anterograde tracer. Subcortical control o f inhibitory cortical circuits In the cerebral cortex of lizards the cell layer is formed by principal spiny projection neurons, while the outer and inner plexiform layers contain aspiny or sparsely spinous GABAergic non-principal neurons. 23'28'36'37'42 GABA-immunoreactive boutons are largely concentrated in the cell layer where they are distributed in a basket-like fashion around the principal cells and in the superficial third of the outer plexiform layer.36 This distribution of GABA-positive boutons matches the terminal fields of different subpopulations of GABAergic non-pyramidal cells which contain different calcium binding proteins or neuropeptides. 5'27'29 It has been recently reported in the DC and DMC, the areas receiving the GABAergic subcortical input, that a subpopulation of GABAergic neurons contains three calcium binding proteins, parvalbumin, calbindin D28k and calretinin. 27 Boutons of these neurons have a basket-like distribution around cell bodies of principal cells; they have been found synapsing on cell bodies 29 and axon initial segments (Martinez-Guijarro et al., unpublished observations), thus, they are involved in perisomatic inhibition. Another subpopulation of GABAergic neurons found in the cortex of lizards includes somatostatin- /NPY-immunoreactive cells.5 These cells have been considered to be involved in distal dendritic inhibition, since somatostatin/NPYimmunoreactive puncta are abundant in the superficial zone of the outer plexiform layer where the distal dendritic tufts of principal cells are located. 27Calcium binding protein-containing cells and somatostatin/ NPY-immunoreactive cells represent non-overlapping subpopulations of GABAergic cells (MartinezGuijarro et al., unpublished observations). In this PHAL-tracing study we used antisera against CaBP and NPY for double-labelling as markers for these two distinct subsets of GABAergic cells, which are likely to correspond to functionally different subpopulations of inhibitory cells. Moreover, considering that both of these markers may be absent from some cortical GABAergic cells (e.g. cholecystokinin-, vasoactive intestinal polypeptide-,fl-endorphin-containing cells) (L6pez-Garcia et al., unpublished observations), PHAL tracing was also combined with immunostaining for GABA. In the DC and DMC, calcium binding protein-containing cells and somatostatin/NPY-immunoreactive

Basal forebrain input to lizard cortical GABAergic neurons neurons extend their dendrites into the projection fields of two presumed excitatory corticocortical pathways. One is the zinc-rich, glutamate-immunoreactive projection arising from the granule cells of the M C which has been compared to the hippocampal mossy fibres. 23'3° Thus, both subsets of GABAergic neurons may participate in feed-forward inhibition of principal pyramidal cells, mediated by the zinc-rich axonal fibres. The other is the commissural-association pathway originating from pyramidal cells of the D M C and D C which terminates in the superficial third of the outer plexiform layer. ~5'26'28In this case, calcium bindin protein-containing neurons and somatostatin/NPY-immunoreactive cells may be involved in both feed-forward and feed-back inhibition of pyramidal cells mediated by commissural fibres and recurrent collaterals of local pyramidal neurons, respectively. In this study we showed that cortical GABAergic non-principal neurons, which may be involved in feed-forward and/or feed-back inhibitory circuits, and mediate perisomatic or distal dendritic inhibition, are under a powerful G A B A e r g i c septal control. Thus, since G A B A has been proved to be inhibitory in the cerebral cortex of reptiles, 21 it can be expected that activation of this GABAergic input to the cortical G A B A e r g i c non-pyramidal cells results in disinhibition of the principal cells in the D C and DMC. The anatomical substrate of a possible disinhibitory action of the septal-basal forebrain input to

435

the cerebral cortex found in lizards is comparable to that described in the mammalian hippocampus s'~3'~4a and neocortex. 9'1°This finding suggests that the mechanisms of control of the cortical activity by the basal forebrain have been h i g h l y preserved during phylogeny. It is interesting to note that in mammals the developmentally younger cortical areas (neocortex) have the smallest, and the archicortex the largest innervation from the GABAergic component of the basal forebrain-cortical pathway. 35 Mesocortical areas (cingulate, retrosplenial cortex) have an intermediate innervation. In mammals, it has been suggested that the septohippocampal pathway may be involved in learning and m e m o r y processes by influencing the induction of long-term potentiation, s'13 The cerebral cortex of lizards also seems to be implicated in spatial memory. 6 Whether the septal-basal forebrain input to the cortex has any influence on spatial memory in lizards remains to be established. thank Drs K. G. Baimbridge (University of British Columbia, Vancouver, Canada), T. J. G6rcs (Semmelweis University Medical School, Budapest, Hungary) and P. Somogyi (MRC Anatomical Neuropharmacology Unit, Oxford, U. K.) for gifts of antisera against calbindin D28k, NPY and GABA. We are grateful to Ms E. Bor6k and Ms I. Weisz for excellent technical assistance. The work was supported by grants from the Human Frontier Science Program (TFF), OTKA (no. 2920) Hungary, University of Kuopio (Finland) and the Spanish DGICYT (PB90-0422). F.J.M.-G. was the recipient of a travel fellowship from the Spanish DGICYT. Acknowledgements--We

REFERENCES

la. Baimbridge K. G. and Miller J. J. (1982) Immunohistochemical localization of calcium-binding protein in the cerebellum, hippocampal formation and olfactory bulb of the rat. Brain Res. 2,45, 223-229. lb. Beaulieu C. and Somogyi P. (1991) Enrichment of cholinergic synaptic terminals on GABAergic neurons and coexistence of immunoreactive GABA and choline acetyltransferase in the same synaptic terminals in the striate cortex of the cat. J. comp. Neurol. 304, 666-680. 2. Bruce L. L. and Butler A. B. (1984) Telencephalic connections in lizards. I. Projections to cortical areas. J. comp. Neurol. 229, 585-604. 3. Cliffer K: D. and Giesler G. J. Jr (1988) PHA-L can be transported anterogradely through fibers of passage. Brain Res. 458, 185-191. 4. Csiff~iry A., G6rcs T. J. and Palkovits M. (1990) Neuropeptide Y innervation of ACTH-immunoreactive neurons in the arcuate nucleus of the rat: a correlated light and electron microscopic double immunolabeling study. Brain Res. 506, 215-222. 5. D~tvila J. C., de la Calle A., Gti~rrez A., Megias M., Andreu M. J. and Guirado S. (1991) Distribution of neuropeptide Y (NPY) in the cerebral cortex of the lizards Psammodromus algirus and Podarcis hispanica: co-localization of NPY, somatostatin, and GABA. J. comp. Neurol. 308, 397-408. 6. Font E., Garcia-Verdugo J. M., Martinez-Guijarro F. J., Alc~ntara S. and L6pez-Garcia C. (1989) Neurobehavioral effects of 3-acetylpyridine in the lizard Podarcis hispanica. Eur. J. Neurosci. 26, 148. 7. Freund T. F. (1989) GABAergic septohippocampal neurons contain parvalbumin. Brain Res. 478, 375-381. 8. Freund T. F. and Antal M. (1988) GABA-containing neurons in the septum control inhibitory interneurons in the hippocampus. Nature 336, 170-173. 9. Freund T. F. and Guly~is A. I. (1991) GABAergic interneurons containing calbindin D28k or somatostatin are major targets of GABAergic basal forebrain afferents in the rat neocortex. J. comp. Neurol. 314, 187-199. 10. Freund T. F. and Meskenaite V. (1992) GABAergic basal forebrain neurons innervate inhibitory interneurons in the neocortex. Proc. natn Acad. Sci. U.S.A. 89, 738-742. 11. Gerfen C. R. and Sawchenko P. E. (1984) An anterograde neuroanatomical tracing method that shows the detailed morphology of neurons, their axons and terminals: immunohistochemical localization of an axonally transported plant lectin, Phaseolus vulgaris leucoagglutinin (PHA-L). Brain Res. 290, 219-238. 12. Guirado S., D~ivila A., Guti~rrez A. and de la Calle A. (1989) GABA structures in the telencephalon of the lizard Psammodromus algirus. An immunohistochemical study. Eur. J. Neurosci. 2S, 21. 13. Gulyfis A. I., G6rcs T. J. and Freund T. F. (1990) Innervation of different peptide-containing neurons in the hippocampus by GABAergic septal afferents. Neuroscience 37, 31-44.

436

F. J. MARTiNEZ-GVrIARROand T. F. FREUND

14a. Gulyfis A. I., Seress L., T6th K., Acsfidy L., Antal M. and Freund T. F. (1991) Septal GABAergic neurons innervate inhibitory interneurons in the hippocampus of the macaque monkey. Neuroscience 41, 381 390. 14b. Hodgson A. J., Penke B., Erdei A., Chubb I. V. and Somogyi P. (1985) Antisera to y-aminobutyric acid. I. Production and characterization using a new model system. J. Histochem. Cytochem. 33, 229-239. 15. Hoogland P. V. and Vermeulen-Van der Zee E. (1988) Intrinsic and extrinsic connections of the cerebral cortex of lizards. In The Forebrain of Reptiles. Current Concepts of Structure and Function (eds Schwerdtfeger W. K. and Smeets W. J. A.), pp. 20-29. Karger, Basel. 16. Hoogland P. V. and Vermeulen-Van der Zee E. (1990) Distribution of choline acetyltransferase immunoreactivity in the telencephalon of the lizard Gekko gecko. Brain Behav. Evol. 36, 378-390. 17. Hoogland P. V., ten Donkelaar H. J. and Cruce J. A. F. (1978) Efferent connections of the septal area in a lizard ( Tupinambis nigropunctatus). Neurosci. Lett. 7, 61q55. 18. Kitt C. A., Mitchell S. J. DeLong M. R., Wainer B. H. and Price L. D. (1987) Fiber pathways of basal forebrain cholinergic neurons in monkeys. Brain Res. 406, 192-206. 9. K6hler C., Chan-Palay V. and Wu J.-Y. (1984) Septal neurons containing glutamic acid decarboxylase immunoreactivity project to the hippocampal region in the rat brain. Anat. Embryol. 169, 41-44. 20. Kosaka T., Nagatsu I., Wu J.-Y. and Hama K. (1986) Use of high concentrations of glutaraldehyde for immunocytochemistry of transmitter-synthesizing enzymes in the central nervous system. Neuroscience 18, 975-990. 21. Kriegstein A. R., Shen J. M., Lee G. M. and Blanton M. G. (1988) Development of inhibitory function during corticogenesis in the turtle. In The Forebrain o f Reptiles. Current Concepts of Structure and Function (eds Schwerdtfeger W. K. and Smeets W. J. A.), pp. 131-141. Karger, Basel. 22. Lohman A. H. M. and Van Woerden Verkley I. (1978) Ascending connections to the forebrain in the tegu lizard. J. comp. Neurol. 182, 555-594. 23. L6pez-Garcia C. and Martinez-Guijarro F. J. (1988) Neurons in the medial cortex give rise to Timm-positive boutons in the cerebral cortex of lizards. Brain Res. 463, 205-217. 24. L6pez-Garcia C., Molowny A., Garcia-Verdugo J. M. and Ferrer I. (1988) Delayed postnatal neurogenesis in the cerebral cortex of lizards. Devl Brain Res. 43, 167-174. 25. Martinez-Garcia F. and Lorente M. J. (1990) Thalamo-cortical projections in the lizard Podarcis hispanica. In The Forebrain in Non-mammals. New Aspects o f Structure and Development (eds Schwerdtfeger W. K. and Germroth P.), pp. 93-102. Springer, Berlin. 26. Martinez-Garcia F., Amiguet M., Schwerdtfeger W. K., Olucha F. E. and Lorente M. J. (1990) Interhemispheric connections through the pallial commissures in the brain ofPodarcis hispanica and Gallotia stehlinii (Reptilia, Lacertidae). J. Morph. 205, 17-31. 27. Martinez-Guijarro F. J. and Freund T. F. (1992) Distribution of GABAergic interneurons immunoreactive for calretinin, calbindin D2sk and parvalbumin in the cerebral cortex of the lizard Podarcis hispanica. J. comp. Neurol. 322, 449-460. 28. Martinez-Guijarro F. J., Desfilis E. and L6pez-Garcia C. (1990) Organization of the dorsomedial cortex in the lizard Podarcis hispanica. In The Forebrain in Non-mammals. New Aspects of Structure and Development (eds Schwerdtfeger W. K. and Germroth P.). pp. 77-92. Springer, Berlin. 29. Martinez-Guijarro F. J., Soriano E., Del Rio J. A. and L6pez-Garcia C. (1991) Parvalbumin-immunoreactive neurons in the cerebral cortex of the lizard Podarcis hispanica. Brain Res. 547, 339-343. 30. Martinez-Guijarro F. J., Soriano E., Del Rio J. A. and C. L6pez-Garcia (1991) Zinc-positive boutons in the cerebral cortex of lizards show glutamate immunoreactivity. J. Neurocytol. 20, 834-843. 31. Mesulam M. M., Mufson E. J., Levey A. I. and Wainer B. H. (1983) Cholinergic innervation of cortex by the basal forebrain: cytochemistry and cortical connection of the septal area, diagonal band nuclei, nucleus basalis (substantia innominata), and hypothalamus in the rhesus monkey. J. comp. Neurol. 214, 170-197. 32. Onteniente B., Tago H., Kimura H. and Maeda T. (1986) Distribution of ~-aminobutyric acid-immunoreactive neurons in the septal region of the rat brain. J. comp. Neurol. 248, 422-430. 33. Pritz M. B. and Stritzel M. E. (1988) Thalamic nuclei that project to reptilian telencephalon lack GABA and GAD immunoreactive neurons and puncta. Brain Res. 457, 154-159. 34. Ram6n y Cajal S. (1911) Histologie du Systeme Nerveux de l'Homme et des Vertebres. Maloine, Paris. 35. Rye D. B., Wainer B. H., Mesulam M. M., Mufson E. J. and Saper C. B. (1984) Cortical projections arising from the basal forebrain: a study of cholinergic and noncholinergic components employing combined retrograde tracing and immunohistochemical localization of choline acetyltransferase. Neuroscience 13, 627q543. 36. Schwerdtfeger W. K. and L6pez-Garcia C. (1986) GABA-ergic neurons in the cerebral cortex of the brain of a lizard (Podarcis hispanica). Neurosci. Lett. 68, 117-121. 37. Schwerdtfeger W. K. and Lorente M. J. (1988) Laminar distribution and morphology of gamma-aminobutyric acid (GABA)-immunoreactive neurons in the medial and dorsomedial cortex of the lizard Podarcis hispanica J. comp. Neurol. 278, 473-485. 38. Schwerdtfeger W. K., L6pez-Garcia C., Martinez-Guijarro F. J. and Tineo Roberto P. L. (1986) GABAergic neurons in the septum of the lizard, Podarcis hispanica. Brain Res. 348, 184-188. 39. Smeets W. J. A. J. (1988) The monoaminergic systems of reptiles investigated with specific antibodies against serotonin, dopamine, and noradrenaline. In The Forebrain of Reptiles. Current Concepts o f Structure and Function (eds Schwerdtfeger W. K. and Smeets W. J. A.), pp. 97-109. Karger, Basel. 40. Somogyi P. and Hodgson A. J. (1985) Antisera to 7-aminobutyric acid. III. Demonstration of GABA in Golgi-impregnated neurons and in conventional electron microscopic sections of cat striate cortex. J. Histochern. Cytochem. 33, 249-257. 41. Szabat E., Happ61a O., Linnala A., Soinila S. and Virtanen I. (1992) A new monoclonal antibody against gamma-aminobutyric acid shows immunoreactivity in sensory neurons. Neuroscience 47, 409-420. 42. Teruel V., Villeta E., Martinez-Garcia F. and Lorente M. J. (1990) The GABAergic system of the dorsal cortex of lizards: a combined G A B A - H R P immunohistochemistry study. Neurosci. Lett. 109, 13-17. 43. Ulinski P. S. (1981) Thick caliber projections from brainsteam to cerebral cortex in the snakes Thamnophis sirtalis and Natrix sipedon. Neuroscience 6, 1725-1743. 44. Ulinski P. S. (1990) The cerebral cortex of reptiles. In Cerebral Cortex (eds Jones E. G. and Peters A.), Vol. 8A, Comparative structure and evolution of cerebral cortex, Part I, pp. 139-215. Plenum Press, New York.

Basal forebrain input to lizard cortical GABAergic neurons

437

45. Wainer B. H., Levey A. I., Rye D. B., Mesulam M. M. and Mufson E. J. (1985) Cholinergic and non-cholinergic septohippocampal pathways. Neurosci. Lett. 54, 45 52. 46. Wouterlood F. G., Bol J. G. J. M. and Steinbush H. W. M. (1987) Double-label immunocytochemistry: combination of anterograde neuroanatomical tracing with Phaseolus vulgaris leucoagglutinin and enzyme immunocytochemistry of target neurons. J. Histochem. Cytochem. 35, 817-823. (Accepted 16 June 1992)