0306-4522/92 $5.00 + 0.00
Neuroscience Vol. 47, No. 2, pp. 395-407, 1992 Printed in Great Britain
Perpmon Press plc 0 1992IBRO
EVIDENCE OF GABA-IMMUNOPOSITIVE NEURONS IN THE DORSAL PART OF THE LATERAL GENICULATE NUCLEUS OF REPTILES: MORPHOLOGICAL CORRELATES WITH INTERNEURONS J.-P. Rro,*t J. RnPaaANr,t$ R. WARD,$$ D. MICELI~ and M. MEDINA$ TINSERM U 106, Laboratoire de Neuromorphologie, Paris, France SLaboratoire d’Anatomie Comparee, MNHN, Paris, France $UQTR, Laboratoire de Neuropsychologie, Trois-Rividres, Canada Ahstrati-The distribution and staining pattern of y-aminobutyric acid immunoreactivity have been examined by both light and electron microscopy in the dorsal part of the lateral geniculate nucleus of three reptilian species: the turtle Chinemys reevesi, the lizard Ophisaurus apodus and the snake Vipera aspis. After perfusion of the animals with 1% paraformaldehyde and 1% glutaraldehyde and polyethyleneglycol embedding of the brains, the analysis of sections processed immunocytochemically with an anti-GABA antiserum has revealed a moderate-to-dense labeling of the neurons of the dorsal part of the lateral geniculate complex in these species. Labeled cell bodies are small-sized, either rounded or fusiform and the GABA-positive dendrites emerging from them are not preferentially oriented in any particular direction. Quantitative studies in Vipera indicate that GABA-positive neurons make up about 14% of the population of neurons of the dorsal part of the lateral geniculate nucleus. Electron microscopy of specimens treated by either pre- or post-embedding techniques has confirmed that these cells corresponded to neurons. No glial cells were ever observed to be immunopositive. These GABA-positive neurons, characterized by the presence of pleiomorphic synaptic vesicles localized either in their perikaryon or more often in presynaptic dendrites, established symmetrical synaptic contacts. In this case., the latter were involved both pre- and postsynaptically in serial and, more rarely, in triadic arrangements, a synaptic organization specific to intemeurons. The involvement of such GABA-positive neurons in local circuits is discussed.
The dorsal part of the lateral geniculate complex (GLd) in reptiles is the retino-recipient thalamic structure which projects to the telencephalon. In lizards,** as in snakes,55 the GLd projects to the lateral part of the dorsal ventricular ridge, while in turtles it projects to the dorsal pallidum’~” and the reptilian GLd has been compared to the GLd of mammals and to the anterior dorsolateral nucleus (DLA4’s5*)of birds. Numerous studies carried out in the mammalian GM, using a variety of histological techniques, and particularly the Golgi method, have revealed the presence of interneurons in this structure, the proportion of which varies according to the
*To whom correspondence should be addressed at: INSERM U 106, Laboratoire de Neuromorphologie, HBpital de la Salp&iere, 47 Bd. de l’H6pita1, 75651 Paris Cedex 13, France. Abbreviations: BSA, bovine serum albumin; DAB, diaminobcnxidine tetrahydrochloride; DLA, anterior dorsolateml nucleus; DMA, anterior dorsomedial nucleus; DP, nucleus of the dorsal peduncle of the lateral forebrain bundle; GAD, glutamate decarboxylase; GLd, dorsal part of the late&l geniculate complex; GLdl, lateral molecular GM: GLdm. medial cellular GLd; GLv, ventral part of ‘the lateral genictdate complex; HRP, horseradish peroxidase; PAP, peroxidascantioeroxidase: PBS. ohosnhate-buffered saline; PEG, polyethyleneglycol; ke, nucleus reuniens; Rt, nucleus rotundus; TBS, Tris-butfered saline; TOM, marginal optic tract; VLT, ventrolateral thalamic nucleus. 395
species (for review, see Ref. 35), and recent immunohistochemical studies have shown these neurons to be GABA-immunopositive. Moreover, the inhibitory function of such neurons with interneurons has been electrophysiologically correlated in the cat’s lateral geniculate nucleus.49 In spite of the numerous investigations which have been devoted to the primary visual system of reptiles (for review, see Ref. 43), the presence of interneurons in the primary visual centers is still a matter of controversy. Though their presence has been undeniably reported both in the superticial layers of the optic tectum6.‘* and in the nucleus of the basal optic root,“*45 authors do not agree about their presence in the GLd. While Ramon’ described a few neurons with short axons in the chameleon with the Golgi technique, more recent experimental investigations carried out in turtles, using either horseradish peroxidase (HRP) or tritiated GABA, led Rainey and Ulinski,‘* and Ulinsk?’ to conclude that interneurons do not exist in the GM of turtles. Results obtained with immunohistochemical methods are also discordant: in the thalamus of Cuiman, and particularly in the GLd, Pritz and Stritzel~ have not observed any GABA-positive neurons and axon terminals and concluded that interneurons are absent in these regions. On the other hand, a recent GABA-immunohistochemical study,2 carried out in the chameleon, has revealed the presence of GABA-immunopositive
396
J.-P. Rto ef ut’
neurons in the GLd, which these authors have interpreted as interneurons. The present study was therefore undertaken in an attempt to resolve these ambiguities. Three reptilian species were investigated, the turtle Chinemys reevesi, the lizard Ophisaurus apodus and the snake Vipera aspis by using immunocytocbemical methods both at light (Chinemys, Ophisaurus and Vipera) and electron (Vipera) microscopic levels. Some of the results have already
been presented
in abstract
EXPERIMENTAL
form.46
PROCEDURES
Four specimens of Chinemys, ~p~i~aur~ and Vipera (kindly supplied from a MNHN expedition) were used for light microscopic investigations, and five additional specimens of Vipera were used for electron microscopic studies. The lack of availability of specimens of the other species prevented us from carrying out more extensive ultrastructural investigations. The primary antiserum was a polyclonal anti-GABA (Immunotech, France) conjugated to glutaraldehyde with bovine serum albumin (BSA). Light microscopy. Four animals of each species were fixed either by intracardial (Chinemys, CJ&isaurus) or intracarotidal (Vi&a) perfusion with 1% paraformaldehyde and 1% alutaraldehvde in 0.12 M nhosnhate buffer. DH 7.3. Heads &em storei overnight in the- same fixative at 4°C and the brains were dissected out and embedded in polyethylene~ycol (PEG). Frontal sections (15 pm) were cut and processed immun~ytochem~caIly by using the peroxidase-antiperoxidase (PAP) method of Stemberger er al.s’ Free-floating sections were collected in phosphatebuffered saline (PBS) containing 0.2% gelatin, 0.25% Triton X-100 and 0.1 M lysin, and subsequently incubated overnight at 4°C in the anti-GABA antiserum diluted 1: 6000 in PBS containing 0.2% gelatin and 0.25% Triton X-100, buffer-rinsed, incubated for I h in a goat anti-rabbit immunoglobulin (1: loo), buffer-rinsed and incubated for I h in a rabbit PAP complex (I :200). The peroxidase was visualized with 0.03% diaminobenzidine tetrahydrochloride (DAB) and 0.005% hydrogen peroxide in 0.1 M Tris-HCI buffer, pH 7.6. Sections were mounted onto gelatin-coated slides and coverslipped before exa~nat~on with an Axiophot Zeiss microscope. Addition~ly, Titan sections (I pm>, obtained from plastic-embedded material in the three Vipera specimens, were mounted on glass slides and processed by the immunoperoxidase method. These sections were first treated in saturated sodium ethanolate and hydrogen peroxide (10% in PFJS) prior to incubation with the anti-GABA antiserum diluted I:5000 in the same PBS buffer. The peroxidase activity was visualized as for light microscopy. Sections were stained with Toluidine Blue, coverslipped and camera lucida drawings of labeled and unlabeled somata with an immersion objective of x 100 in I-urn-thick plastic sections were made in order to calculate their diameter. All neurons, both GABA-positive (n = 56) and GABA-negative (n = 343), visible in two series (separated by 50pm) of four serial scmithin sections (1 am>, processed ~mun~yt~herni~l~y and stained with Totuidine Blue, were measured; ghal cells, clearly identifiable by morphological specific criteria proposed at the light microscopic levek2’ were not measured. Similar measurements of GM neurons were made in consecutive l-pm sections that were not processed for immunocytochemistry. The second series of measurements provides an estimate of the number of the GM neurons which was not statistically different from the overall number of GABA-positive and -negative neurons estimated in the first series. The specificity of the immunohistochemical procedures was verified on every 10th section, either by omitting the
primary antiserum, replacing it with a non-imnlunc normal serum or by immunoadsorption of free GABA (200 mM) on polyacrylamide beads with glutaraldehyde as a linking reagent. Under none of these conditions was any labeling ever observed. Electron microscopy. Two specimens of Vipera were prepared for pre-embedding and three for post-embedding immunocytochemistry. Pre-embedding. The perfusion and fixation of snakes were the same as those used for light microscopy. Following several PBS rinses of the brains after perfusion, the thalamic region was Vibratome-sliced (SOpm), immunoprocessed according to the same PAP method, except that the antiGABA antiserum dilution was 1:2000. Sections were subsequently postfixed in 2% buffered osmium tetroxide, stained en bloc with uranyl acetate, dehydrated in a graded series of ethanol and flat epoxy-em~dded. Thin sections (70-80 nm) were cut and mounted on copper grids before examination with a Philips 400 electron microscope. Structures and profiles considered as GABA-immunopositive were those localized further than 30pm from the dense nonspecific labeling occurring in the cut surface. The same controls as those employed at the light microscopic level gave identical negative results. Post-embedding. Following perfusion with 1% paraformaldehyde and 1% glutaraldehyde in 0.12 M phosphate buffer, pH 7.3, small frontally cut pieces of the thalamus were prepared for conventional electron microscopy. Thin sections, mounted on nickel grids, were previously etched with 1% periodic acid (7 min) and I % sodium metaperiodate (7 min), pretreated in 10% BSA in Tris-buffered saline (TBS; pH 7.6) for 30 min and incubated overnight at 4°C in an anti-GABA antiserum diluted 1: 2000 in TBS at the same pH. After several rinses in TBS., sections were reacted with a goat anti-rabbit IgG colloidal gold (15 nm; Janssen, Belgium) diluted 1:75 and were double-stained with uranyl acetate and lead citrate prior to examination. The intensity of background labeling was estimated from GABAimmunonegative profiles such as retinal terminals,27 blood vessels and empty resin. 32 In some cases, semi-serial thin sections were performed in order to visualize the synaptic contact established by presynaptic dendrites. As for the other procedures, the same control experiments resulted in the absence of labeling. RESULTS
Light microscopy
Chinemys. The GM of turtles differs from that of other reptilian species in that it can be clearly subdivided into two parts: (i) a lateral molecular (GLdl) and (ii) a medial cellular (GLdm) region which correspond, respectively, to the lateral magnocellular and medial parvocellular sublaminae described in the cell plate region by Rainey and Ulinski.3* Our examination has revealed that only neurons located in the GM1 were GABA-immunopositive and evenly dispersed, whereas the medial region (GLdm) was devoid of immunoreacti~ty (Fig. 1A). The modest thickness of our sections cl5 pm) prevented us from consistently following the dendritic arborizations of GABA-positive neurons over long distances. The presence of numerous fine punctate dots was observed in the neuropil. Among the other nuclei of the thalamus, we observed immunopositive cells in the ventral part of the lateral geniculate complex (GLv) and the nucleus of the dorsal peduncle of the lateral forebrain bundle
GABA-immunopositive
(DP), while no immunoreactive cells were observed in the anterior dorsolateral nucleus (DLA), anterior dorsomedial nucleus (DMA), nucleus rotundus (Rt) and nucleus reuniens (Re). Ophisawus. In this lizard, the GM is a wellcircumscribed structure located in the laterodorsal part of the thalamus which comprises three cytoarchitectonic subdivisions. I9 Immunoreactive neurons were small-sized, rounded and dendrites did not display any orientation (Fig. 1B). Among the other thalamic structures, we observed a dense immunoreactivity in the GLv, DP and ventrolateral thalamic nucleus (VLT), while the DLA, DMA and Rt were immunonegative. Vipera. The GM of this species constitutes an oval-shaped mass of the dorsal thalamus, clearly separated from the GLv (Fig. lC), located in the concavity of the marginal optic tract (TOM) and in which cytoarchitectonic subdivisions are difficult to define.” The GABA-positive neurons are homogeneously distributed throughout the GLd (Fig. 1C); they are either small or rounded, though some fusiform neurons and less intensely stained cell bodies were observed. The dendritic arborizations were uneasy to define precisely, although in some sections primary dendritic stems did*&t display any topographical orientation. The analysis of immunoprocessed semithin sections revealed that the distribution of labeled neurons appeared to be homogeneous throughout the GLd (x2 = 2.98, with P > 0.90; Fig. ID, E), and it thus appears that about 14% of Vipera GLd neurons are GABA-immunopositive. Labeled cell bodies are smaller (8.3 If: 1.8 pm) than unlabeled cell bodies (12.5 + 3.3 pm; t, = 16.4, P
Neuroscience Vol. 47, No. 2, pp. 395-407, 1992 Printed in Great Britain
Perpmon Press plc 0 1992IBRO
EVIDENCE OF GABA-IMMUNOPOSITIVE NEURONS IN THE DORSAL PART OF THE LATERAL GENICULATE NUCLEUS OF REPTILES: MORPHOLOGICAL CORRELATES WITH INTERNEURONS J.-P. Rro,*t J. RnPaaANr,t$ R. WARD,$$ D. MICELI~ and M. MEDINA$ TINSERM U 106, Laboratoire de Neuromorphologie, Paris, France SLaboratoire d’Anatomie Comparee, MNHN, Paris, France $UQTR, Laboratoire de Neuropsychologie, Trois-Rividres, Canada Ahstrati-The distribution and staining pattern of y-aminobutyric acid immunoreactivity have been examined by both light and electron microscopy in the dorsal part of the lateral geniculate nucleus of three reptilian species: the turtle Chinemys reevesi, the lizard Ophisaurus apodus and the snake Vipera aspis. After perfusion of the animals with 1% paraformaldehyde and 1% glutaraldehyde and polyethyleneglycol embedding of the brains, the analysis of sections processed immunocytochemically with an anti-GABA antiserum has revealed a moderate-to-dense labeling of the neurons of the dorsal part of the lateral geniculate complex in these species. Labeled cell bodies are small-sized, either rounded or fusiform and the GABA-positive dendrites emerging from them are not preferentially oriented in any particular direction. Quantitative studies in Vipera indicate that GABA-positive neurons make up about 14% of the population of neurons of the dorsal part of the lateral geniculate nucleus. Electron microscopy of specimens treated by either pre- or post-embedding techniques has confirmed that these cells corresponded to neurons. No glial cells were ever observed to be immunopositive. These GABA-positive neurons, characterized by the presence of pleiomorphic synaptic vesicles localized either in their perikaryon or more often in presynaptic dendrites, established symmetrical synaptic contacts. In this case., the latter were involved both pre- and postsynaptically in serial and, more rarely, in triadic arrangements, a synaptic organization specific to intemeurons. The involvement of such GABA-positive neurons in local circuits is discussed.
The dorsal part of the lateral geniculate complex (GLd) in reptiles is the retino-recipient thalamic structure which projects to the telencephalon. In lizards,** as in snakes,55 the GLd projects to the lateral part of the dorsal ventricular ridge, while in turtles it projects to the dorsal pallidum’~” and the reptilian GLd has been compared to the GLd of mammals and to the anterior dorsolateral nucleus (DLA4’s5*)of birds. Numerous studies carried out in the mammalian GM, using a variety of histological techniques, and particularly the Golgi method, have revealed the presence of interneurons in this structure, the proportion of which varies according to the
*To whom correspondence should be addressed at: INSERM U 106, Laboratoire de Neuromorphologie, HBpital de la Salp&iere, 47 Bd. de l’H6pita1, 75651 Paris Cedex 13, France. Abbreviations: BSA, bovine serum albumin; DAB, diaminobcnxidine tetrahydrochloride; DLA, anterior dorsolateml nucleus; DMA, anterior dorsomedial nucleus; DP, nucleus of the dorsal peduncle of the lateral forebrain bundle; GAD, glutamate decarboxylase; GLd, dorsal part of the late&l geniculate complex; GLdl, lateral molecular GM: GLdm. medial cellular GLd; GLv, ventral part of ‘the lateral genictdate complex; HRP, horseradish peroxidase; PAP, peroxidascantioeroxidase: PBS. ohosnhate-buffered saline; PEG, polyethyleneglycol; ke, nucleus reuniens; Rt, nucleus rotundus; TBS, Tris-butfered saline; TOM, marginal optic tract; VLT, ventrolateral thalamic nucleus. 395
species (for review, see Ref. 35), and recent immunohistochemical studies have shown these neurons to be GABA-immunopositive. Moreover, the inhibitory function of such neurons with interneurons has been electrophysiologically correlated in the cat’s lateral geniculate nucleus.49 In spite of the numerous investigations which have been devoted to the primary visual system of reptiles (for review, see Ref. 43), the presence of interneurons in the primary visual centers is still a matter of controversy. Though their presence has been undeniably reported both in the superticial layers of the optic tectum6.‘* and in the nucleus of the basal optic root,“*45 authors do not agree about their presence in the GLd. While Ramon’ described a few neurons with short axons in the chameleon with the Golgi technique, more recent experimental investigations carried out in turtles, using either horseradish peroxidase (HRP) or tritiated GABA, led Rainey and Ulinski,‘* and Ulinsk?’ to conclude that interneurons do not exist in the GM of turtles. Results obtained with immunohistochemical methods are also discordant: in the thalamus of Cuiman, and particularly in the GLd, Pritz and Stritzel~ have not observed any GABA-positive neurons and axon terminals and concluded that interneurons are absent in these regions. On the other hand, a recent GABA-immunohistochemical study,2 carried out in the chameleon, has revealed the presence of GABA-immunopositive
396
J.-P. Rto ef ut’
neurons in the GLd, which these authors have interpreted as interneurons. The present study was therefore undertaken in an attempt to resolve these ambiguities. Three reptilian species were investigated, the turtle Chinemys reevesi, the lizard Ophisaurus apodus and the snake Vipera aspis by using immunocytocbemical methods both at light (Chinemys, Ophisaurus and Vipera) and electron (Vipera) microscopic levels. Some of the results have already
been presented
in abstract
EXPERIMENTAL
form.46
PROCEDURES
Four specimens of Chinemys, ~p~i~aur~ and Vipera (kindly supplied from a MNHN expedition) were used for light microscopic investigations, and five additional specimens of Vipera were used for electron microscopic studies. The lack of availability of specimens of the other species prevented us from carrying out more extensive ultrastructural investigations. The primary antiserum was a polyclonal anti-GABA (Immunotech, France) conjugated to glutaraldehyde with bovine serum albumin (BSA). Light microscopy. Four animals of each species were fixed either by intracardial (Chinemys, CJ&isaurus) or intracarotidal (Vi&a) perfusion with 1% paraformaldehyde and 1% alutaraldehvde in 0.12 M nhosnhate buffer. DH 7.3. Heads &em storei overnight in the- same fixative at 4°C and the brains were dissected out and embedded in polyethylene~ycol (PEG). Frontal sections (15 pm) were cut and processed immun~ytochem~caIly by using the peroxidase-antiperoxidase (PAP) method of Stemberger er al.s’ Free-floating sections were collected in phosphatebuffered saline (PBS) containing 0.2% gelatin, 0.25% Triton X-100 and 0.1 M lysin, and subsequently incubated overnight at 4°C in the anti-GABA antiserum diluted 1: 6000 in PBS containing 0.2% gelatin and 0.25% Triton X-100, buffer-rinsed, incubated for I h in a goat anti-rabbit immunoglobulin (1: loo), buffer-rinsed and incubated for I h in a rabbit PAP complex (I :200). The peroxidase was visualized with 0.03% diaminobenzidine tetrahydrochloride (DAB) and 0.005% hydrogen peroxide in 0.1 M Tris-HCI buffer, pH 7.6. Sections were mounted onto gelatin-coated slides and coverslipped before exa~nat~on with an Axiophot Zeiss microscope. Addition~ly, Titan sections (I pm>, obtained from plastic-embedded material in the three Vipera specimens, were mounted on glass slides and processed by the immunoperoxidase method. These sections were first treated in saturated sodium ethanolate and hydrogen peroxide (10% in PFJS) prior to incubation with the anti-GABA antiserum diluted I:5000 in the same PBS buffer. The peroxidase activity was visualized as for light microscopy. Sections were stained with Toluidine Blue, coverslipped and camera lucida drawings of labeled and unlabeled somata with an immersion objective of x 100 in I-urn-thick plastic sections were made in order to calculate their diameter. All neurons, both GABA-positive (n = 56) and GABA-negative (n = 343), visible in two series (separated by 50pm) of four serial scmithin sections (1 am>, processed ~mun~yt~herni~l~y and stained with Totuidine Blue, were measured; ghal cells, clearly identifiable by morphological specific criteria proposed at the light microscopic levek2’ were not measured. Similar measurements of GM neurons were made in consecutive l-pm sections that were not processed for immunocytochemistry. The second series of measurements provides an estimate of the number of the GM neurons which was not statistically different from the overall number of GABA-positive and -negative neurons estimated in the first series. The specificity of the immunohistochemical procedures was verified on every 10th section, either by omitting the
primary antiserum, replacing it with a non-imnlunc normal serum or by immunoadsorption of free GABA (200 mM) on polyacrylamide beads with glutaraldehyde as a linking reagent. Under none of these conditions was any labeling ever observed. Electron microscopy. Two specimens of Vipera were prepared for pre-embedding and three for post-embedding immunocytochemistry. Pre-embedding. The perfusion and fixation of snakes were the same as those used for light microscopy. Following several PBS rinses of the brains after perfusion, the thalamic region was Vibratome-sliced (SOpm), immunoprocessed according to the same PAP method, except that the antiGABA antiserum dilution was 1:2000. Sections were subsequently postfixed in 2% buffered osmium tetroxide, stained en bloc with uranyl acetate, dehydrated in a graded series of ethanol and flat epoxy-em~dded. Thin sections (70-80 nm) were cut and mounted on copper grids before examination with a Philips 400 electron microscope. Structures and profiles considered as GABA-immunopositive were those localized further than 30pm from the dense nonspecific labeling occurring in the cut surface. The same controls as those employed at the light microscopic level gave identical negative results. Post-embedding. Following perfusion with 1% paraformaldehyde and 1% glutaraldehyde in 0.12 M phosphate buffer, pH 7.3, small frontally cut pieces of the thalamus were prepared for conventional electron microscopy. Thin sections, mounted on nickel grids, were previously etched with 1% periodic acid (7 min) and I % sodium metaperiodate (7 min), pretreated in 10% BSA in Tris-buffered saline (TBS; pH 7.6) for 30 min and incubated overnight at 4°C in an anti-GABA antiserum diluted 1: 2000 in TBS at the same pH. After several rinses in TBS., sections were reacted with a goat anti-rabbit IgG colloidal gold (15 nm; Janssen, Belgium) diluted 1:75 and were double-stained with uranyl acetate and lead citrate prior to examination. The intensity of background labeling was estimated from GABAimmunonegative profiles such as retinal terminals,27 blood vessels and empty resin. 32 In some cases, semi-serial thin sections were performed in order to visualize the synaptic contact established by presynaptic dendrites. As for the other procedures, the same control experiments resulted in the absence of labeling. RESULTS
Light microscopy
Chinemys. The GM of turtles differs from that of other reptilian species in that it can be clearly subdivided into two parts: (i) a lateral molecular (GLdl) and (ii) a medial cellular (GLdm) region which correspond, respectively, to the lateral magnocellular and medial parvocellular sublaminae described in the cell plate region by Rainey and Ulinski.3* Our examination has revealed that only neurons located in the GM1 were GABA-immunopositive and evenly dispersed, whereas the medial region (GLdm) was devoid of immunoreacti~ty (Fig. 1A). The modest thickness of our sections cl5 pm) prevented us from consistently following the dendritic arborizations of GABA-positive neurons over long distances. The presence of numerous fine punctate dots was observed in the neuropil. Among the other nuclei of the thalamus, we observed immunopositive cells in the ventral part of the lateral geniculate complex (GLv) and the nucleus of the dorsal peduncle of the lateral forebrain bundle
GABA-immunopositive
(DP), while no immunoreactive cells were observed in the anterior dorsolateral nucleus (DLA), anterior dorsomedial nucleus (DMA), nucleus rotundus (Rt) and nucleus reuniens (Re). Ophisawus. In this lizard, the GM is a wellcircumscribed structure located in the laterodorsal part of the thalamus which comprises three cytoarchitectonic subdivisions. I9 Immunoreactive neurons were small-sized, rounded and dendrites did not display any orientation (Fig. 1B). Among the other thalamic structures, we observed a dense immunoreactivity in the GLv, DP and ventrolateral thalamic nucleus (VLT), while the DLA, DMA and Rt were immunonegative. Vipera. The GM of this species constitutes an oval-shaped mass of the dorsal thalamus, clearly separated from the GLv (Fig. lC), located in the concavity of the marginal optic tract (TOM) and in which cytoarchitectonic subdivisions are difficult to define.” The GABA-positive neurons are homogeneously distributed throughout the GLd (Fig. 1C); they are either small or rounded, though some fusiform neurons and less intensely stained cell bodies were observed. The dendritic arborizations were uneasy to define precisely, although in some sections primary dendritic stems did*&t display any topographical orientation. The analysis of immunoprocessed semithin sections revealed that the distribution of labeled neurons appeared to be homogeneous throughout the GLd (x2 = 2.98, with P > 0.90; Fig. ID, E), and it thus appears that about 14% of Vipera GLd neurons are GABA-immunopositive. Labeled cell bodies are smaller (8.3 If: 1.8 pm) than unlabeled cell bodies (12.5 + 3.3 pm; t, = 16.4, P
