- Email: [email protected]
Thalamic nuclei that project to reptilian telencephalon lack GABA and GAD immunoreactive neurons and puncta
154
Brain Red,earth, 457 ( I ~ ) 154 159 l-.Iscvicr
BRE 23024
Thalamic nuclei that project to reptilian telencephalon lack GABA and GAD immunoreactive neurons and puncta Michael B. Pritz and Mark E. Stritzel Division gf Neurological Surgery, California College of Medicine, University of CaliJornia lrvine Medical Center, Orange, CA 92668 (U.S.A.) (Accepted 26 April 1988) Key words: 7-Aminobutyric acid; Glutamic acid decarboxylase: lmmunocytochemistry; Inhibition: Local circuit neuron: Relay cell; Reptile: Thalamus
Brains of reptiles, Caiman crocodilus, were processed by standard immunocytochemical methodology using a polyclonal antibody to y-aminobutyric acid (GABA) raised in rabbit and a polyclonal antibody to glutamic acid decarboxylase (GAD) raised in sheep. No GABA(+) or GAD(+) cells or puncta were observed over any thalamic nucleus known to project to the telencephalon in Caiman. These findings suggest that all thalamic nuclei that project to the telencephalon in Caiman lack intrinsic cells and presumably direct inhibitory input mediated by GABA.
Previously, we investigated the percentage of relay and intrinsic neurons in 3 thalamic nuclei in reptiles, Caiman crocodilus, by means of large injections of horseradish peroxidase ( H R P ) into either the general cortex r or the dorsal ventricular ridge ( D V R ) l~'. Retrogradely labeled neurons were felt to represent relay cells while unlabeled cells were considered to be intrinsic or local circuit neurons. W h e n examined in this way, we found that nucleus dorsolateralis anterior, which projects bilaterally to cortex I; and which probably does not receive direct sensory information, contains less than 1% intrinsic cells located ipsilaterally Iv while nucleus reuniens pars centralis (audition) and nucleus rotundus (vision), which project to the D V R L~l-'. each contain less than l % local circuit neurons ~('. While this approach has provided a reasonable index of the percentage of relay and intrinsic neurons in cat dorsal lateral geniculate nucleus 5 when compared with other i n d e p e n d e n t techniques 4"~'~'~2~, certain limitations are well appreciated l(''lv. Therefore, we sought other ways to independently verify our H R P quantitative observations. W e chose immu-
nocytochemistry with antibodies against y-aminobutyric acid ( G A B A ) and glutamic acid decarboxylase ( G A D ) , the synthesizing enzyme in G A B A production, for 2 reasons. First, cells immunoreactive for G A B A , G A B A ( + ) , and G A D , G A D ( + ) , a r e thought to identify the intrinsic neurons of the dorsal thalamus 7. Second, in experiments in which massive injections of H R P were made into Rhesus m o n k e y striate cortex and semi-thin sections through the dorsal lateral geniculate nucleus were examined, G A B A ( + ) cells lacked H R P grains and the vast majority of G A B A ( - ) cells were densely labeled with H R P reaction product I~. Experiments were p e r f o r m e d on 19 juvenile Cai,tan crocodilus. Animals weighed between 23.5 and 1160 g. S n o u t - v e n t length ranged from 10 to 38 cm. Animals received a lethal overdose of sodium pentobarbital intraperitoneally. Before perfusion, each animal received between 300 and 2000 U sodium heparin intraperitoneally and between 300 and 3500 U sodium heparin transcardiaily based on each animal's weight. A variety of different perfusates were e m p l o y e d in
Correspondence: Michael B. Pritz, Division of Neurological Surgery, University of California Irvine Medical Center. 101 City Drive South, Orange, CA 92668, U.S.A. ()006-8993/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)
155 the 14 GABA experiments. The optimum combination was: 0.1 M phosphate buffer, pH 7.2, at ambient temperature followed by cold solutions of graded increases in sucrose solutions of 0%, 10% and 30% in phosphate-buffered saline and fixative (0.4% glutaraldehyde and 4% paraformaldehyde). In the 5 GAD experiments, each animal was perfused with 0.1 M phosphate buffer in 5% sucrose, pH 7.2, at ambient temperature followed by cold solutions of graded increase of sucrose solutions of 5%, 10% and 30% in 0.1 M phosphate buffer and fixative (0.2% glutaraldehyde, 4% paraformaldehyde, 0.1 M lysine, and 0.01% sodium periodate). Brains were then blocked in a standard plane 11, removed from the skull, and left overnight in a solution of 30% sucrose, 0.1 M phosphate buffer at pH 7.2, and fixative. The following day, frozen sections were cut on a sliding microtome at 15 or 20/~m for the GABA experiments or at 30 ~tm for the GAD studies. Sections were collected serially into cooled trays containing 0.1 M phosphate buffer at pH 7.2. For the GABA experiments, free floating sections were processed by the peroxidaseanti-peroxidase (PAP) technique (4 cases) and by the avidin-biotin complex (ABC) method (10 cases). All GAD experiments were done by the ABC method. Free-floating sections processed for G A B A immunocytochemistry used a polyclonal rabbit anti-GABA antibody (ImmunoNuclear, Stillwater, MN) and followed a protocol similar to that described by others 2 except that the blocking solution contained 5% or 10% normal goat serum and accordingly a biotinylated goat anti-rabbit immunoglobulin (Vector Laboratories, Burlingame, CA) was used. Free floating sections processed for GAD immunocytochemistry used a polyclonal sheep anti-GAD antibody (gift from D. Schmechel) and followed a protocol similar to that described by others 2 except that 10% normal rabbit serum was used in the blocking solution. Concentrations of the rabbit anti-GABA antibody varied from 1/1000 to 1/30,000 with the optimal range of dilutions obtained at 1/1000, 1/2000, and 1/4000. The ABC method provided better results than the PAP technique. Concentrations of sheep anti-GAD antibody varied from 1/500 to 1/16,000 with the best resuits obtained at concentrations of 1/500, 1/1000, and 1/2000. Controls for the GABA experiments were: pre-incubation of primary antibody (1/2000) with 'free' G A B A (100 ~tg/ml); substitution of normal rab-
bit serum or rabbit IgG for the primary antibody at the same or higher concentrations than that of the primary; and omission of one of the following reagents - - primary antibody, biotinylated secondary IgG, or ABC. In addition, preabsorption of the primary antibody with bovine serum albumin (BSA) (200pg/ml) was also used in the GABA experiments. Controls for the GAD experiments included: omission of one of the following reagents - - primary antibody, biotinylated secondary IgG, or ABC and substitution of pre-immune sheep serum for the primary antibody at concentrations of 1/250 or 1/500. Colchicine pretreatment was used in 4 animals - - 1 processed for GABA and 3 processed for GAD. Each animal was anesthetized by cold narcosis and placed in a stereotaxic headholder modified for use with Caiman. A craniectomy was made over the posterior telencephalon and rostral optic tectum. After the thalamus was exposed, 10 pl of a 1% colchicine solution was directly instilled into the third ventricle. The incision was closed. Animals were housed in aquaria in which the water temperature varied from 19 to 29 °C and survived approximately 24 h before sacrifice. Animals were then perfused and tissue was processed as described above. Neurons considered immunoreactive for GABA or GAD fullfilled 2 criteria. First, both the cell body and primary dendrites required visualization. Second, control sections of substituted normal or pre-immune serum at concentrations equal to or greater than that of the primary antibody did not label cells considered to be G A B A ( + ) or G A D ( + ) . In the GABA experiments, we found that background activity was high. This was attributable to non-specific reaction between rabbit serum and Caiman brain tissue. To circumvent these problems, several adjustments were made. These included: changes in perfusate, changes in fixatives, use of 1% normal goat serum with all rinses since rabbit anti-serum was followed by biotinylated goat anti-rabbit immunoglobulin, and pre-absorption with BSA since the anti-GABA antibody was produced by injecting rabbits with GABA conjugated to BSA by glutaraldehyde. Colchicine produced no appreciable change in GABA staining. Nevertheless, no G A B A ( + ) cells or puncta were seen over either nucleus rotundus, nucleus reuniens pars centralis, or nucleus dorsolateralis anterior (Fig. 1A-C) or over any other thalam-
156
g
A
A i v
D
O
G"
% ~o
).
Z
#
°/
E
L
5
Fig. 1. GABA and GAD immunoreactivity in the thalamus. Transverse sections of nucleus rotundus (A), nucleus reuniens pars centralis (B), and nucleus dorsolateralis anterior (C) in which the concentration of rabbit anti-GABA antibody was 1/1000 are shown. Transverse sections of nucleus reuniens pars centralis (D), nucleus rotundus (E), and nucleus dorsolateralis anterior (F) in which the concentration of sheep anti-GAD antibody was 1/1000 in animals pre-treated with colchicine are illustrated. Reaction products in A, C and E (arrowheads) are within blood vessels and represent endogenous peroxidase activity. Bar in A-C and D-F = 100/~m.
ic nucleus known to project to the telencephalon in
Caiman 3,13-15. In the G A D experiments, colchicine had a profound enhancing effect not only on the impregnation of soma and dendrites but also on the visualization of puncta. These differences were readily seen in the cerebellum (Fig. 2 A , B ) and in the olfactory bulb (Fig. 2 C - E ) . This pattern and morphology of immunoreactive cells and puncta were never seen in control sections even at concentrations of pre-immune serum 2 - 4 times that of the primary antibody•
When sections through nucleus rotundus, nucleus reuniens pars centralis, or nucleus dorsolateralis anterior (Fig. 1 D - F ) were examined, neither G A D ( + ) cells nor G A D ( + ) puncta were seen. Furthermore, neither G A D ( + ) neurons nor G A D ( + ) puncta were observed over other thalamic nuclei (nucleus dorsomedialis anterior, nucleus diagonalis, nucleus reuniens pars diffusa, nucleus posterocentralis, or the medialis complex) known to project to the telencephalon3A 3-15. Before discussing our results, several possible
157 sources of error need to be addressed. First, the lack
G A B A findings, the staining of tertiary dendrites,
of G A B A and G A D immunoreactivity in the thalamus might be attributed to technical errors. While such an argument might be raised to explain the
axons, soma, and puncta seen in the G A D experiments (Fig. 2) processed in the same tissue wells as the n o n - i m m u n o r e a c t i v e thalamic section suggests
• Fig. 2. GAD immnnoreactivity in the cerebellum and olfactory bulb. Sagittal cerebellar sections of colchicine pre-treated control (A: pre-immune sheep serum at concentration of 1/1000) and primary antibody stained sections (B: sheep anti-GAD antibody concentration of 1/1000) are shown. In B, observe the intense puncta present throughout the molecular layer, labeling of Purkinje cell soma and dendrites, and the staining of presumed Golgi Type II neurons (arrows) and of axons (arrowhead). Sagittal olfactory bulb sections of control (C: pre-immune sheep serum at concentration of 1/2000) and primary antibody stained sections (D-E: sheep anti-GAD antibody concentration of 1/2000) are illustrated. Asterisk in D and E marks a blood vessel for orientation of higher magnification photo shown in E. Note the numerous GAD(+) puncta, GAD(+) cells, and dendritic staining in D and E as compared with control section, C. In E, thin arrow points to branching of a primary dendrite into a secondary dendrite while the thick arrow marks branching of a secondary into a tertiary dendrite. Arrowheads denote a labeled tertiary GAD(+) dendrite. Abbreviations: G, granule cell layer; GL, glomerular layer; M, molecular layer; ONL, olfactory nerve layer; P, Purkinje cell layer. Bar = 100/~min A and B; 75 ~m in C and D; and 25/~m in E.
158 that purely methodologic problems are not the explanation for our results. Second, tissue processed for G A B A immunocytochemistry exhibited non-specific background staining secondary to cross-reactivity between rabbit serum and Caiman brain tissue. Under these circumstances, one might have expected false positive staining in the thalamus. Clearly, this was not the case as the dorsal thalamus (Fig. 1 A - C ) was conspicuously free from any sort of G A B A ( + ) immunoreactivity. Third, cross-species variation in response to the commercially available rabbit antiG A B A antibody might have led to an incomplete identification of all G A B A populations. In our reptilian species, Caiman crocodilus, this might have resulted in a lack of G A B A ( + ) cells and puncta. While an anti-GABA antibody raised in an animal other than rabbit might have answered this possible concern, none was available to us. Therefore, we performed additional experiments using a polyclonal antibody to G A D raised in a different animal, sheep, since GAD is localized in neurons that appear to use G A B A as a neurotransmitter 6. Obvious G A D ( + ) cells and puncta were found in the cerebellum (Fig. 2B) and olfactory bulb (Fig. 2D,E) in a fashion similar to that described for mammals TM. Admittedly, antibodies that would specifically recognize Caiman GAD might have revealed G A D ( + ) cells and puncta that we did not observe. However, such an antibody has yet to be produced. Nevertheless, this polyclonal antibody to G A D has been used by numerous investigators and is considered to provide reliable and reproducible data. Moreover, these immunocytochemical results independently confirm and extend the findings of our previous quantitative HRP studies 16,17. Bearing in mind the above qualifications, these data suggest the following. First, neither G A B A ( + ) nor G A D ( + ) neurons or puncta are present in nucleus rotundus, nucleus reuniens pars centralis, or nucleus dorsolateralis anterior. Second, no other tha-
1 Barbaresi, P., Spreafico, R., Frassoni, C. and Rustioni, A., GABAergic neurons are present in the dorsal column nuclei but not in the ventroposterior complex of rats, Brain Research, 382 (1986) 305-326. 2 Blanton, M.G., Shen, J.M. and Kriegstein, A.R., Evidence for the inhibitory neurotransmitter y-aminobutyric acid in aspiny and sparsely spiny nonpyramidal neurons of
lamic nucleus that projects to the telencephalon in Caiman contains either G A B A ( + ) or G A D ( + ) cells or puncta. These immunocytochemical experiments combined with previous HRP studies 16'17 indicate that all telencephalic-projecting thalamic nuclei in Caiman lack the classical intrinsic or local circuit neurons that have been described in the thalamus of mammals 7. These results raise a number of interesting and biologically significant questions. First, if no classical intrinsic neurons are present in the dorsal thalamus of Caiman, does this imply that inhibition within each of these telencephalic-projecting thalamic nuclei is absent as well? Or are intrinsic neurons present more peripherally as has been described in rat ventrobasal complex which lacks G A D ( + ) cells while approximately 25% G A D ( + ) cells are present in the dorsal column nucleusl? Second, if inhibition is present in the thalamus, what is its mechanism? Third, if inhibition is present in the thalamus of Caiman, is it mediated by a transmitter(s) other than G A B A ? Last, are these features of a 'pure' relay cell population in the dorsal thalamus of Caiman, a feature unique to Caiman or common to other reptiles or vertebrates as well? Answers to these questions are the focus of future research. We are deeply indebted to Drs. D. Schmechel for a generous supply of anti-GAD antibody and pre-immune serum; E.G. Jones for introducing us to immunocytochemistry and providing laboratory facilities and reagents for initial experiments; S. Hendry for his advice, guidance, and patience; and C. Ribak for his suggestions and comments. We also thank J. Blanks for manuscript preparation, R. Robertson for loan of microscope lens, and L. Sutherland and the Department of Pathology for the use of photographic facilities. Supported by General Surgery Grant 102, CCM Neurobiology Grant, and funds from the Blakely Compensation Plan.
the turtle dorsal cortex, J. Comp. Neurol., 259 (1987) 277-297. 3 Brauth, S.E. and Kitt, C.A., The paleostriatal system of Caiman crocodilus, J. Comp. Neurol., 189 (1980) 437-465. 4 Fitzpatrick, D., Penny, G.R. and Schmechel, D.E., Glutamic acid decarboxylase-immunoreactiveneurons and terminals in the lateral geniculate nucleus of the cat, J. Neuro-
159
sci., 4 (1984) 1809-1829. 5 Geisert, E.E., Jr., Cortical projections of the lateral geniculate nucleus in the cat, J. Comp. Neurol., 190 (1980) 793-812. 6 Gottlieb, D.I., Chang, Y.-C. and Schwob, J.E., Monoclonal antibodies to glutamic acid decarboxylase, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 8808-8812. 7 Jones, E.G., The Thalamus, Plenum, New York, 1985, 935 PP. 8 LeVay, S. and Ferster, D., Proportion of interneurons in the cat's lateral geniculate nucleus, Brain Research, 164 (1979) 304-308. 9 Madar~isz, M., Somogyi, G., Somogyi, J. and H~imori, J., Numerical estimation of ~,-aminobutyric acid (GABA)containing neurons in three thalamic nuclei of the cat: direct GABA immunocytochemistry, Neurosci. Lett., 61 (1985) 73-78. 10 Montero, V.M., The interneuronal nature of GABAergic neurons in the lateral geniculate nucleus of the rhesus monkey: a combined HRP and GABA-immunocytochemical study, Exp. Brain Res., 64 (1986) 615-622. 11 Pritz, M.B., Ascending connections of a thalamic auditory area in a crocodile, Caiman crocodilus, J. Comp. Neurol., 153 (1974) 199-214. 12 Pritz, M.B., Anatomical identification of a telencephalic visual area in crocodiles: ascending connections of nucleus. rotundus in Caiman crocodilus, J. Cornp. Neurol., 164 (1975) 323-338.
13 Pritz, M.B. and Northcutt, R.G., Anatomical evidence for an ascending somatosensory pathway to the telencephalon in crocodiles, Caiman crocodilus, Exp. Brain Res., 40 (1980) 342-345. 14 Pritz, M.B. and Stritzel, M.E., Thalamic projections to the noncortical telencephalon in a reptile, Soc. Neurosci. Abstr., 11 (1985) 1309. 15 Pritz, M.B. and Stritzel, M.E., Reptilian somatosensory midbrain, Soc. Neurosci. Abstr., 12 (1986) 106. 16 Pritz, M.B. and Stritzel, M.E., Percentage of relay and intrinsic neurons in two sensory thalamic nuclei projecting to the non-cortical telencephalon in reptiles Caiman crocodilus, Brain Research, 376 (1986) 169-174. 17 Pritz, M.B. and Stritzel, M.E., Percentage of intrinsic and relay cells in a thalamic nucleus projecting to general cortex in reptiles, Caiman crocodilus, Brain Research, 409 (1987) 146-150. 18 Shepherd, G.M., The Synaptic Organization of the Brain, 2nd edn., Oxford University Press, New York, 1979, 152-183 pp. 19 Sterling, P. and Davis, T.L., Neurons in cat lateral geniculate nucleus that concentrate exogenous [3H]~,-aminobutyric acid (GABA), J. Comp. Neurol., 192 (1980) 737-749. 20 Weber, A.J. and Kalil, R.E., The percentage of interneurons in the dorsal lateral geniculate nucleus of the cat and observations on several variables that affect the sensitivity of horseradish peroxidase as a retrograde marker, J. Comp. Neurol., 220 (1983) 336-346.
Brain Red,earth, 457 ( I ~ ) 154 159 l-.Iscvicr
BRE 23024
Thalamic nuclei that project to reptilian telencephalon lack GABA and GAD immunoreactive neurons and puncta Michael B. Pritz and Mark E. Stritzel Division gf Neurological Surgery, California College of Medicine, University of CaliJornia lrvine Medical Center, Orange, CA 92668 (U.S.A.) (Accepted 26 April 1988) Key words: 7-Aminobutyric acid; Glutamic acid decarboxylase: lmmunocytochemistry; Inhibition: Local circuit neuron: Relay cell; Reptile: Thalamus
Brains of reptiles, Caiman crocodilus, were processed by standard immunocytochemical methodology using a polyclonal antibody to y-aminobutyric acid (GABA) raised in rabbit and a polyclonal antibody to glutamic acid decarboxylase (GAD) raised in sheep. No GABA(+) or GAD(+) cells or puncta were observed over any thalamic nucleus known to project to the telencephalon in Caiman. These findings suggest that all thalamic nuclei that project to the telencephalon in Caiman lack intrinsic cells and presumably direct inhibitory input mediated by GABA.
Previously, we investigated the percentage of relay and intrinsic neurons in 3 thalamic nuclei in reptiles, Caiman crocodilus, by means of large injections of horseradish peroxidase ( H R P ) into either the general cortex r or the dorsal ventricular ridge ( D V R ) l~'. Retrogradely labeled neurons were felt to represent relay cells while unlabeled cells were considered to be intrinsic or local circuit neurons. W h e n examined in this way, we found that nucleus dorsolateralis anterior, which projects bilaterally to cortex I; and which probably does not receive direct sensory information, contains less than 1% intrinsic cells located ipsilaterally Iv while nucleus reuniens pars centralis (audition) and nucleus rotundus (vision), which project to the D V R L~l-'. each contain less than l % local circuit neurons ~('. While this approach has provided a reasonable index of the percentage of relay and intrinsic neurons in cat dorsal lateral geniculate nucleus 5 when compared with other i n d e p e n d e n t techniques 4"~'~'~2~, certain limitations are well appreciated l(''lv. Therefore, we sought other ways to independently verify our H R P quantitative observations. W e chose immu-
nocytochemistry with antibodies against y-aminobutyric acid ( G A B A ) and glutamic acid decarboxylase ( G A D ) , the synthesizing enzyme in G A B A production, for 2 reasons. First, cells immunoreactive for G A B A , G A B A ( + ) , and G A D , G A D ( + ) , a r e thought to identify the intrinsic neurons of the dorsal thalamus 7. Second, in experiments in which massive injections of H R P were made into Rhesus m o n k e y striate cortex and semi-thin sections through the dorsal lateral geniculate nucleus were examined, G A B A ( + ) cells lacked H R P grains and the vast majority of G A B A ( - ) cells were densely labeled with H R P reaction product I~. Experiments were p e r f o r m e d on 19 juvenile Cai,tan crocodilus. Animals weighed between 23.5 and 1160 g. S n o u t - v e n t length ranged from 10 to 38 cm. Animals received a lethal overdose of sodium pentobarbital intraperitoneally. Before perfusion, each animal received between 300 and 2000 U sodium heparin intraperitoneally and between 300 and 3500 U sodium heparin transcardiaily based on each animal's weight. A variety of different perfusates were e m p l o y e d in
Correspondence: Michael B. Pritz, Division of Neurological Surgery, University of California Irvine Medical Center. 101 City Drive South, Orange, CA 92668, U.S.A. ()006-8993/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)
155 the 14 GABA experiments. The optimum combination was: 0.1 M phosphate buffer, pH 7.2, at ambient temperature followed by cold solutions of graded increases in sucrose solutions of 0%, 10% and 30% in phosphate-buffered saline and fixative (0.4% glutaraldehyde and 4% paraformaldehyde). In the 5 GAD experiments, each animal was perfused with 0.1 M phosphate buffer in 5% sucrose, pH 7.2, at ambient temperature followed by cold solutions of graded increase of sucrose solutions of 5%, 10% and 30% in 0.1 M phosphate buffer and fixative (0.2% glutaraldehyde, 4% paraformaldehyde, 0.1 M lysine, and 0.01% sodium periodate). Brains were then blocked in a standard plane 11, removed from the skull, and left overnight in a solution of 30% sucrose, 0.1 M phosphate buffer at pH 7.2, and fixative. The following day, frozen sections were cut on a sliding microtome at 15 or 20/~m for the GABA experiments or at 30 ~tm for the GAD studies. Sections were collected serially into cooled trays containing 0.1 M phosphate buffer at pH 7.2. For the GABA experiments, free floating sections were processed by the peroxidaseanti-peroxidase (PAP) technique (4 cases) and by the avidin-biotin complex (ABC) method (10 cases). All GAD experiments were done by the ABC method. Free-floating sections processed for G A B A immunocytochemistry used a polyclonal rabbit anti-GABA antibody (ImmunoNuclear, Stillwater, MN) and followed a protocol similar to that described by others 2 except that the blocking solution contained 5% or 10% normal goat serum and accordingly a biotinylated goat anti-rabbit immunoglobulin (Vector Laboratories, Burlingame, CA) was used. Free floating sections processed for GAD immunocytochemistry used a polyclonal sheep anti-GAD antibody (gift from D. Schmechel) and followed a protocol similar to that described by others 2 except that 10% normal rabbit serum was used in the blocking solution. Concentrations of the rabbit anti-GABA antibody varied from 1/1000 to 1/30,000 with the optimal range of dilutions obtained at 1/1000, 1/2000, and 1/4000. The ABC method provided better results than the PAP technique. Concentrations of sheep anti-GAD antibody varied from 1/500 to 1/16,000 with the best resuits obtained at concentrations of 1/500, 1/1000, and 1/2000. Controls for the GABA experiments were: pre-incubation of primary antibody (1/2000) with 'free' G A B A (100 ~tg/ml); substitution of normal rab-
bit serum or rabbit IgG for the primary antibody at the same or higher concentrations than that of the primary; and omission of one of the following reagents - - primary antibody, biotinylated secondary IgG, or ABC. In addition, preabsorption of the primary antibody with bovine serum albumin (BSA) (200pg/ml) was also used in the GABA experiments. Controls for the GAD experiments included: omission of one of the following reagents - - primary antibody, biotinylated secondary IgG, or ABC and substitution of pre-immune sheep serum for the primary antibody at concentrations of 1/250 or 1/500. Colchicine pretreatment was used in 4 animals - - 1 processed for GABA and 3 processed for GAD. Each animal was anesthetized by cold narcosis and placed in a stereotaxic headholder modified for use with Caiman. A craniectomy was made over the posterior telencephalon and rostral optic tectum. After the thalamus was exposed, 10 pl of a 1% colchicine solution was directly instilled into the third ventricle. The incision was closed. Animals were housed in aquaria in which the water temperature varied from 19 to 29 °C and survived approximately 24 h before sacrifice. Animals were then perfused and tissue was processed as described above. Neurons considered immunoreactive for GABA or GAD fullfilled 2 criteria. First, both the cell body and primary dendrites required visualization. Second, control sections of substituted normal or pre-immune serum at concentrations equal to or greater than that of the primary antibody did not label cells considered to be G A B A ( + ) or G A D ( + ) . In the GABA experiments, we found that background activity was high. This was attributable to non-specific reaction between rabbit serum and Caiman brain tissue. To circumvent these problems, several adjustments were made. These included: changes in perfusate, changes in fixatives, use of 1% normal goat serum with all rinses since rabbit anti-serum was followed by biotinylated goat anti-rabbit immunoglobulin, and pre-absorption with BSA since the anti-GABA antibody was produced by injecting rabbits with GABA conjugated to BSA by glutaraldehyde. Colchicine produced no appreciable change in GABA staining. Nevertheless, no G A B A ( + ) cells or puncta were seen over either nucleus rotundus, nucleus reuniens pars centralis, or nucleus dorsolateralis anterior (Fig. 1A-C) or over any other thalam-
156
g
A
A i v
D
O
G"
% ~o
).
Z
#
°/
E
L
5
Fig. 1. GABA and GAD immunoreactivity in the thalamus. Transverse sections of nucleus rotundus (A), nucleus reuniens pars centralis (B), and nucleus dorsolateralis anterior (C) in which the concentration of rabbit anti-GABA antibody was 1/1000 are shown. Transverse sections of nucleus reuniens pars centralis (D), nucleus rotundus (E), and nucleus dorsolateralis anterior (F) in which the concentration of sheep anti-GAD antibody was 1/1000 in animals pre-treated with colchicine are illustrated. Reaction products in A, C and E (arrowheads) are within blood vessels and represent endogenous peroxidase activity. Bar in A-C and D-F = 100/~m.
ic nucleus known to project to the telencephalon in
Caiman 3,13-15. In the G A D experiments, colchicine had a profound enhancing effect not only on the impregnation of soma and dendrites but also on the visualization of puncta. These differences were readily seen in the cerebellum (Fig. 2 A , B ) and in the olfactory bulb (Fig. 2 C - E ) . This pattern and morphology of immunoreactive cells and puncta were never seen in control sections even at concentrations of pre-immune serum 2 - 4 times that of the primary antibody•
When sections through nucleus rotundus, nucleus reuniens pars centralis, or nucleus dorsolateralis anterior (Fig. 1 D - F ) were examined, neither G A D ( + ) cells nor G A D ( + ) puncta were seen. Furthermore, neither G A D ( + ) neurons nor G A D ( + ) puncta were observed over other thalamic nuclei (nucleus dorsomedialis anterior, nucleus diagonalis, nucleus reuniens pars diffusa, nucleus posterocentralis, or the medialis complex) known to project to the telencephalon3A 3-15. Before discussing our results, several possible
157 sources of error need to be addressed. First, the lack
G A B A findings, the staining of tertiary dendrites,
of G A B A and G A D immunoreactivity in the thalamus might be attributed to technical errors. While such an argument might be raised to explain the
axons, soma, and puncta seen in the G A D experiments (Fig. 2) processed in the same tissue wells as the n o n - i m m u n o r e a c t i v e thalamic section suggests
• Fig. 2. GAD immnnoreactivity in the cerebellum and olfactory bulb. Sagittal cerebellar sections of colchicine pre-treated control (A: pre-immune sheep serum at concentration of 1/1000) and primary antibody stained sections (B: sheep anti-GAD antibody concentration of 1/1000) are shown. In B, observe the intense puncta present throughout the molecular layer, labeling of Purkinje cell soma and dendrites, and the staining of presumed Golgi Type II neurons (arrows) and of axons (arrowhead). Sagittal olfactory bulb sections of control (C: pre-immune sheep serum at concentration of 1/2000) and primary antibody stained sections (D-E: sheep anti-GAD antibody concentration of 1/2000) are illustrated. Asterisk in D and E marks a blood vessel for orientation of higher magnification photo shown in E. Note the numerous GAD(+) puncta, GAD(+) cells, and dendritic staining in D and E as compared with control section, C. In E, thin arrow points to branching of a primary dendrite into a secondary dendrite while the thick arrow marks branching of a secondary into a tertiary dendrite. Arrowheads denote a labeled tertiary GAD(+) dendrite. Abbreviations: G, granule cell layer; GL, glomerular layer; M, molecular layer; ONL, olfactory nerve layer; P, Purkinje cell layer. Bar = 100/~min A and B; 75 ~m in C and D; and 25/~m in E.
158 that purely methodologic problems are not the explanation for our results. Second, tissue processed for G A B A immunocytochemistry exhibited non-specific background staining secondary to cross-reactivity between rabbit serum and Caiman brain tissue. Under these circumstances, one might have expected false positive staining in the thalamus. Clearly, this was not the case as the dorsal thalamus (Fig. 1 A - C ) was conspicuously free from any sort of G A B A ( + ) immunoreactivity. Third, cross-species variation in response to the commercially available rabbit antiG A B A antibody might have led to an incomplete identification of all G A B A populations. In our reptilian species, Caiman crocodilus, this might have resulted in a lack of G A B A ( + ) cells and puncta. While an anti-GABA antibody raised in an animal other than rabbit might have answered this possible concern, none was available to us. Therefore, we performed additional experiments using a polyclonal antibody to G A D raised in a different animal, sheep, since GAD is localized in neurons that appear to use G A B A as a neurotransmitter 6. Obvious G A D ( + ) cells and puncta were found in the cerebellum (Fig. 2B) and olfactory bulb (Fig. 2D,E) in a fashion similar to that described for mammals TM. Admittedly, antibodies that would specifically recognize Caiman GAD might have revealed G A D ( + ) cells and puncta that we did not observe. However, such an antibody has yet to be produced. Nevertheless, this polyclonal antibody to G A D has been used by numerous investigators and is considered to provide reliable and reproducible data. Moreover, these immunocytochemical results independently confirm and extend the findings of our previous quantitative HRP studies 16,17. Bearing in mind the above qualifications, these data suggest the following. First, neither G A B A ( + ) nor G A D ( + ) neurons or puncta are present in nucleus rotundus, nucleus reuniens pars centralis, or nucleus dorsolateralis anterior. Second, no other tha-
1 Barbaresi, P., Spreafico, R., Frassoni, C. and Rustioni, A., GABAergic neurons are present in the dorsal column nuclei but not in the ventroposterior complex of rats, Brain Research, 382 (1986) 305-326. 2 Blanton, M.G., Shen, J.M. and Kriegstein, A.R., Evidence for the inhibitory neurotransmitter y-aminobutyric acid in aspiny and sparsely spiny nonpyramidal neurons of
lamic nucleus that projects to the telencephalon in Caiman contains either G A B A ( + ) or G A D ( + ) cells or puncta. These immunocytochemical experiments combined with previous HRP studies 16'17 indicate that all telencephalic-projecting thalamic nuclei in Caiman lack the classical intrinsic or local circuit neurons that have been described in the thalamus of mammals 7. These results raise a number of interesting and biologically significant questions. First, if no classical intrinsic neurons are present in the dorsal thalamus of Caiman, does this imply that inhibition within each of these telencephalic-projecting thalamic nuclei is absent as well? Or are intrinsic neurons present more peripherally as has been described in rat ventrobasal complex which lacks G A D ( + ) cells while approximately 25% G A D ( + ) cells are present in the dorsal column nucleusl? Second, if inhibition is present in the thalamus, what is its mechanism? Third, if inhibition is present in the thalamus of Caiman, is it mediated by a transmitter(s) other than G A B A ? Last, are these features of a 'pure' relay cell population in the dorsal thalamus of Caiman, a feature unique to Caiman or common to other reptiles or vertebrates as well? Answers to these questions are the focus of future research. We are deeply indebted to Drs. D. Schmechel for a generous supply of anti-GAD antibody and pre-immune serum; E.G. Jones for introducing us to immunocytochemistry and providing laboratory facilities and reagents for initial experiments; S. Hendry for his advice, guidance, and patience; and C. Ribak for his suggestions and comments. We also thank J. Blanks for manuscript preparation, R. Robertson for loan of microscope lens, and L. Sutherland and the Department of Pathology for the use of photographic facilities. Supported by General Surgery Grant 102, CCM Neurobiology Grant, and funds from the Blakely Compensation Plan.
the turtle dorsal cortex, J. Comp. Neurol., 259 (1987) 277-297. 3 Brauth, S.E. and Kitt, C.A., The paleostriatal system of Caiman crocodilus, J. Comp. Neurol., 189 (1980) 437-465. 4 Fitzpatrick, D., Penny, G.R. and Schmechel, D.E., Glutamic acid decarboxylase-immunoreactiveneurons and terminals in the lateral geniculate nucleus of the cat, J. Neuro-
159
sci., 4 (1984) 1809-1829. 5 Geisert, E.E., Jr., Cortical projections of the lateral geniculate nucleus in the cat, J. Comp. Neurol., 190 (1980) 793-812. 6 Gottlieb, D.I., Chang, Y.-C. and Schwob, J.E., Monoclonal antibodies to glutamic acid decarboxylase, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 8808-8812. 7 Jones, E.G., The Thalamus, Plenum, New York, 1985, 935 PP. 8 LeVay, S. and Ferster, D., Proportion of interneurons in the cat's lateral geniculate nucleus, Brain Research, 164 (1979) 304-308. 9 Madar~isz, M., Somogyi, G., Somogyi, J. and H~imori, J., Numerical estimation of ~,-aminobutyric acid (GABA)containing neurons in three thalamic nuclei of the cat: direct GABA immunocytochemistry, Neurosci. Lett., 61 (1985) 73-78. 10 Montero, V.M., The interneuronal nature of GABAergic neurons in the lateral geniculate nucleus of the rhesus monkey: a combined HRP and GABA-immunocytochemical study, Exp. Brain Res., 64 (1986) 615-622. 11 Pritz, M.B., Ascending connections of a thalamic auditory area in a crocodile, Caiman crocodilus, J. Comp. Neurol., 153 (1974) 199-214. 12 Pritz, M.B., Anatomical identification of a telencephalic visual area in crocodiles: ascending connections of nucleus. rotundus in Caiman crocodilus, J. Cornp. Neurol., 164 (1975) 323-338.
13 Pritz, M.B. and Northcutt, R.G., Anatomical evidence for an ascending somatosensory pathway to the telencephalon in crocodiles, Caiman crocodilus, Exp. Brain Res., 40 (1980) 342-345. 14 Pritz, M.B. and Stritzel, M.E., Thalamic projections to the noncortical telencephalon in a reptile, Soc. Neurosci. Abstr., 11 (1985) 1309. 15 Pritz, M.B. and Stritzel, M.E., Reptilian somatosensory midbrain, Soc. Neurosci. Abstr., 12 (1986) 106. 16 Pritz, M.B. and Stritzel, M.E., Percentage of relay and intrinsic neurons in two sensory thalamic nuclei projecting to the non-cortical telencephalon in reptiles Caiman crocodilus, Brain Research, 376 (1986) 169-174. 17 Pritz, M.B. and Stritzel, M.E., Percentage of intrinsic and relay cells in a thalamic nucleus projecting to general cortex in reptiles, Caiman crocodilus, Brain Research, 409 (1987) 146-150. 18 Shepherd, G.M., The Synaptic Organization of the Brain, 2nd edn., Oxford University Press, New York, 1979, 152-183 pp. 19 Sterling, P. and Davis, T.L., Neurons in cat lateral geniculate nucleus that concentrate exogenous [3H]~,-aminobutyric acid (GABA), J. Comp. Neurol., 192 (1980) 737-749. 20 Weber, A.J. and Kalil, R.E., The percentage of interneurons in the dorsal lateral geniculate nucleus of the cat and observations on several variables that affect the sensitivity of horseradish peroxidase as a retrograde marker, J. Comp. Neurol., 220 (1983) 336-346.