- Email: [email protected]
Immunocytochemical evidence for the presence of enzymes synthesizing GABA and GHB in the same neuron
Neurochem. Int. Vol. 6, No. 3, pp. 333-338, 1984 Printed in Great Britain. All rights reserved
0197-0186/84 $3.00+ 0.00 Copyright ~, 1984 Pergamon Press Ltd
IMMUNOCYTOCHEMICAL EVIDENCE FOR THE PRESENCE OF ENZYMES SYNTHESIZING GABA A N D GHB IN THE SAME N E U R O N D. WEISSMANN-NANOPOULOS,M. F. BELIN*,P. MANDEL and M. MAITREt INSERM U44, 5 rue Blaise Pascal, 67084 Strasbourg Cedex and *INSERM U171, 8 avenue Rockfeller, 69008 Lyon, France (Received 27 July 1983; accepted 10 October 1983)
Abstraet--A specific and sensitive immunocytochemical double staining for visualization of glutamate decarboxylase (GAD) and semialdehyde succinate reductase (SSR2) in the same brain section has been developed. SSR2 is the enzyme responsible for the transformation of succinic semialdehyde into ?-hydroxybutyrate (GHB). GAD was detected using specific rabbit GAD-antibodies and unlabeled antibody enzyme peroxidase antiperoxidase, and SSR2 using specific guinea-pig SSR2 antibodies conjugate to a fluorescein-labeled second antibody. The coexistence of GAD and SSR2 in the same neuron was demonstrated by a peroxidase reaction superimposed on fluorescent compounds. Cell bodies containing both antigens were observed in the cerebellum, dorso-median hypothalamus and raphe nuclei. GHB is present in most GABA containing neurons. Some neurons contain only SSR2; these neurons may synthesize GHB by an active uptake of GABA.
The major route for GABA degradation occurs via transamination to succinate semialdehyde (SSA) in the mitochondria. SSA is dehydrogenated to form succinate (Pitts and Quick, 1965). However, succinate semialdehyde may also be rapidly converted by an N A D P H reduction reaction to 7-hydroxybutyrate (GHB) (Roth and Giarman, 1969). A specific succinate semialdehyde reductase (SSR2) which is responsible for GHB synthesis has been purified from both human and rat brain (Tabakoff and Von Wartburg, 1975; Cash et al., 1979; Rumigny et al., 1980). This enzyme, located in the cytoplasm of neurons, in dendrites and synapses (Weissmann-Nanopoulos et al., 1982) is of great interest since its product, GHB, has been demonstrated in the past few years to exhibit the principal characteristics of a putative neurotransmitter in the rat brain. In fact, the conditions of synthesis, release (Maitre and Mandel, 1982; Maitre et al., 1983), transport (Benavides et al., 1982a) and binding of GHB (Benavides et al., 1982b; Maitre et al., 1983a; Maitre et aL, 1983b) at the synaptic level suggest the possible existence of GHBergic synapses. Measurements of GHB turnover time in rat brain were established to be even more rapid than those reported for whole brain serotonin, dopamine or norepinephrine (Gold and Roth, 1977).
tTo whom correspondence should be addressed.
Thus, it was of interest to investigate a possible coexistence or segregation between G A D ("GABAergic neurons") and SSR2 in "GHBergic neurons". The latter probably requires the presence of GABA to synthesize succinate semialdehyde. To investigate this phenomenon, an immunocytochemical study was carried out using double labelling techniques with glutamate decarboxylase (GAD) and with specific succinate semialdehyde reductase (SSR2). Both antibodies were obtained and characterized in our laboratory (Nanopoulos el al., 1982; Rumigny et al., 1982). To avoid ambiguous results, antibodies against glutamate decarboxylase were obtained from rabbit and SSR 2 antibodies from guinea pig. This facilitates double labelling techniques. The experiments were carried out on three regions of the rat brain: cerebellum, median hypothalamus and raphe dorsalis, which are well characterized and rich in both enzymes (Rumigny et al., 1982). EXPERIMENTAL PROCEDURES
An antiserum to GAD purified from rat brain was produced in rabbits by subcutaneous injections of the antigen emulsified with Freund's adjuvant. The characteristics of the antigen and the antibodies were reported previously (Blinderman et al., 1978; Maitre et al., 1978; Nanopoulos et al., 1982). Pure SSR2 was obtained from rat brain using the method described by Rumigny et al. (1980). Antiserum to specific succinic semialdehyde reductase was raised in the guinea333
334
D. WE1SSVIANN-NANOPOULOSel al.
pig, as previously described for the anti-SSR2 prepared in the rabbit (Weissmann-Nanopoulos et al., 1982). Ten male rats (OFA, 200 g) were anaesthetized by an intraperitoneal injection of sodium pentobarbital (Nembutal, 50 mg/kg body weight) and then the tissues fixed by a 10min intra-cardiac perfusion of 150-200ml freshly prepared fixative composed of a phosphate buffer (100mM. pH 7.4) containing 4')~, paraformaldehyde, and 0.1~i; glutaraldehyde. The dissected brains were postfixed for 12 h in 4~!~; para[brmaldehyde dissolved in the same buffer. After rinsing in an ice-cold phosphate buffer (100 raM, pH 7.4) containing 15,'~, sucrose for at least 12 h, the midbrains were cut on a cryostat at - 18C (sections thickness set at 8 pm). Series of 2 4 consecutive sections were collected, rinsed in phosphate buffer saline (PBS), pH 7.4 and then incubated for 4 h in PBS containing 11}~,normal sheep serum (NSS). GAD immunoreactivity was revealed by the unlabeled antibody enzyme method of Sternberger (peroxidase-antiperoxidase technique, 1 9 7 0 ) . SSR~ immunoreactivity was demonstrated according to the indirect immunofluorescence staining (for details, see Sternberger, 1979). Sections were incubated for 48h at + 4 ' C in PBS containing GAD antiserum diluted 1:2000 and SSR 2 antiserum diluted 1 : 100. Then they were washed in PBS-I°o NSS and incubated for I hr with a swine antirabbit serum (Dakopatts, Denmark) diluted I : 50 with PBS. After washing briefly in PBS, the sections were incubated for l h with a rabbit peroxidase-antiperoxidase complex (Dakopatts, Denmark) diluted 1:50 in PBS. Following a further wash, the sections were treated for 15 min with Tris HCI buffer (50mM, pH 7.5) containing 3,3' diaminobenzidine-4 HC1 (7.5 mg/10 ml) and H202 (0.6 ml of 30°0 HeO2/10ml ). Finally, they were washed in Tris HCI buffer (50 raM, pH 7.5). rhen, the sections were incubated with fluorescein isothiocyanate conjugated goat antiguinea-pig serum immunoglobulin (Cappel) I : 100 in PBS + 0.1" o Triton X-100. Finally, the sections were rinsed again in PBS and mounted in a mixture of glycerin-PBS (3 : I, v/v). They were examined using a fluorescence microscope equipped for incident light illumination. Adjacent sections were either stained for GAD or SSR 2 and then treated identically. In control sections, experiments were performed using preimmune sera.
RESULTS
Methods
Experiments were performed to determine optimal conditions for simultaneous localization of the two antigens. Consecutive sections were incubated either with a n t i - G A D serum, or with anti-SSR2 serum. An optimal double staining on the same section was obtained by decreasing the peroxidase reaction with a shorter revelation time (10min). W h e n sections were incubated with the two antisera and revealed first by immunofluorescence and secondly by peroxidase, the fluorescent staining is altered by the immunocytochemical peroxidase staining procedures,
Double immuno~3'tochemical staining
In sections incubated with b o t h antisera, G A D as well as SSR 2 positive cells were detected simultaneously. A l t h o u g h the fluorescent staining was sometimes weak, n u m e r o u s cell bodies were observed to be reactive to b o t h antibodies. The cell bodies exhibited characteristic G A D positive reaction deposits superimposed o n t o the fluorescent c o m p o u n d making it sometimes difficult to observe the fluorescent dyes. The G A D - p o s i t i v e reaction product was more or less strong and dispersed over the entire cytoplasm (Fig. 1). In the cerebellar cortex, m a n y Purkinje cells were double stained (Fig. lAB). A b o u t 400~ of them exhibit only a SSR2 positive reaction product (Fig.
ICD). In the raphe nuclei, some multipolar neurons contain b o t h antibodies (Fig. 2). The G A D positive reaction product is dispersed in the cytoplasm (Fig. 2AC), and the SSR2 positive reaction p r o d u c t makes p u n c t a deposits a r o u n d the same cells (Fig. 2B). SSR2 can be or is within the cytoplasm (Fig. 2D). A few neurons are only G A D positive (Fig. 2AC) but 80 90,~0 of the cells are double stained. In the reticular formation, some neurons exhibit G A D positive puncta a r o u n d their soma (Fig. 2E), and SSR2 positive reaction products are present within the cytoplasm (Fig. 2F). In the dorso-median h y p o t h a l a m u s , a b o u t 90'~o of the cells are double stained (Fig. 3). The sections incubated with prei m m u n e sera were not stained.
DISCUSSION
This study was carried out using a n t i - G A D and anti-SSR 2 sera from two different species. The antiG A D serum was applied for identification of G A B A ergic neurons in the Purkinje cell of the cerebellum, the substantia nigra and the nucleus raphe dorsalis ( N a n o p o u l o s et al., 1982). The anti-SSR2 serum was applied in order to d e m o n s t r a t e the presence of SSR, in neurons of the hypothalamus, and the nucleus raphe dorsalis ( W e i s s m a n n - N a n o p o u l o s et al., 1982). In this study, second antibodies were specific respectively against rabbit IgG and guinea-pig IgG. Tissue sections were incubated with the two antibodies simultaneously. A n t i - G A D IgG were localized by the peroxidase antiperoxidase technique of Sternberger (1970) which allows a great dilution of the first a n t i b o d y (l:2000). The presence of anti-SSR2 was detected using fluorescent staining. This m e t h o d o l o g y is used in order to avoid interference with the per-
GABA and GHB synthesizing enzymes in the same neuron
335
"qr ~q
ml
P
B
gi
,
P
mi
P Fig. 1. Sections of the cerebellar cortex double stained with antibodies to GAD and SSR2. Bright field illumination (A, C:GAD) and incident-light fluorescence (B, D: SSR2). ml:molecular layer; gl:granular layer; p: Purkinje cell layer. Arrows: cell bodies immunoreactive both to GAD and SSR2. Scale: 20/~m.
oxidase reaction substrates on the fluorochrome. Nevertheless, the development of the peroxidase reaction product may partially mask the fluorescence. For this reason, the peroxidase reaction was only developed during a short period of time, and with a low
concentration of H202. A soma exhibiting a high GAD-positive staining may prevent the access to the second antiserum with its peroxidase deposits. Therefore, a pure GAD-positive cell observed in a region showing, on consecutive sections, SSR2-positive
336
D. WEISSMANN-NANOPOULOSel al.
F
L
L;i
Fig. 2. Sections of the nucleus raphe magnus (A, B, C, D) and pontic reticular formation (E, F) double stained with antibodies to GAD and SSR2. Bright field illumination (A, C, E:GAD) and incident light fluorescence (B, D, F:SSR2). Scale: 10/~m. Arrows show cell immunoreactive both to GAD and SSR 2. Note in A and C, two cell bodies only reactive to GAD.
staining, may be a G A B A element containing inaccessible SSR2 sites. On serial sections, some neurons appear to be only GAD-positive. In this case, we can exclude the presence of SSR2. The double stained cells and the SSR2-positive cells correspond respectively to cells containing G A D and SSR2 alone. No staining is observed on sections treated with preimmune sera. Cells containing both antigens G A D and SSR, are present in the cerebellum, hypothalamus and
raphe nuclei. In the hypothalamus, the majority of these cells are double stained. This result is consistent with a high SSR~ (Rumigny e t al., 1981) and a high G A D activity (Tappaz e t al., 1977) in this area. In the nucleus raphe dorsalis and magnus, most of the cells appeared to be reactive to both antisera. A few neurons are only SSR2-positive. In the cerebellum, several Purkinje cells are double stained; however, some of them are only SSR2 positive.
GABA and GHB synthesizing enzymes in the same neuron
337
Fig. 3. Sections of the dorsomedial nucleus of the hypothalamus (premammilary region) double stained with antibodies to GAD and SSR. Bright field illumination (A, C:GAD) and incident light fluorescence (B, D:SSR2). Scale: 10 #m. Arrows show the cell bodies reactive both to GAD and SSR 2.
The coexistence of the two antigens in most of the neurons in the hypothalamus and raphe area suggests that G A B A is the precursor of G H B in these neurons. As G H B has been qualified as a possible neurotransmitter, these neurons can be either GABAergic and/or " G H B e r g i c " . In the neurons containing only SSR 2, G H B may be synthesized by an active uptake of G A B A from the surrounding cells. In that case, this group of neurons will be purely " G H B e r g i c " . Cells exhibiting SSR2 positive puncta at the periphery of a G A D positive soma could represent a G A B A containing neuron receiving a " G H B e r g i c " input.
In summary, this study demonstrates that G H B is present in most of the G A B A containing neurons. G H B may be a major catabolite of G A B A in neurons while the oxidative route of SSA is predominant in glial cells. Taking into account the potential neurotransmitter characteristics of G H B , this compartmentation might have an important functional role for neuronal activity. REFERENCES
Blindermann J. M., Maitre M., Ossola L. and Mandel P. (1978) Purification and some properties of L-glutamate
338
D. WEISSMANN-NANOPOULOSet al.
decarboxylase from human brain. Eur. J. Biochem. 86, 143 152. Benavides J., Rumigny J. F., Bourguignon J. J., Wermuth C. G., Mandel P. and Maitre M. (1982a) A high affinity, Na + dependent uptake system for 7-hydroxybutyrate in membrane vesicles prepared from rat brain. J. Neurochem. 38, 1570-1575. Benavides J., Rumigny J. F., Bourguignon J. J., Cash C., Wermuth C. G., Mandel P., Vincendon G. and Maitre M. (1982b) High affinity binding site for 7-hydroxybutyrate acid in rat brain. Life Sci. 30, 953-961. Cash C., Maitre M. and Mandel P. (1978) Purification de deux semiald6hydes succinique r6ductases de cerveau humain. C.r. hebd. S~anc. Acad. Sci., Paris 286, 1829-1832. Cash C. D., Maitre M. and Mandel P. (1979) Purification and properties of two NADPH-linked aldehyde reductases which reduce succinic semialdehyde to 4 hydroxybutyrate. J, Neurochem. 33, 1169-1175. Gold B. I. and Roth R. H. (1977) Kinetics of in vivo conversion of ~/-[3H] aminobutyric acid to 7-[3H] hydroxybutyric acid by rat brain. J. Neurochem. 28, 1069-1073. Maitre M., Blindermann J. M., Ossola L. and Mandel P. (1978) Comparison of the structures of L-glutamate decarboxylases from human and rat brain. Biochem. biophys. Res. Commun. vol. 85, 885 890. Maitre M. and Mandel P. (1982) Lib6ration de 7-hydroxybutyrate calcium-dependante apr6s d~polarisation de coupes de cerveau de rat. C.r. hObd. Skanc. Acad. Sci., Paris 295, 741-743. Maitre M., Cash C., Weissmann-Nanopoulos D. and Mandel P. (1983) Depolarization evoked release of ),-hydroxybutyrate from rat brain slices. J. Neurochem. 41, 287 290. Maitre M., Rumigny J. F., Cash C. and Mandel P. (1983a) Subcellular distribution of 7-hydroxybutyrate binding sites in rat brain-principal localization in the synaptosomal fraction. Biochem. biophys. Res. Commun. 110, 262-265. Maitre M., Rumigny J. F. and Mandel P. (1983b) Positive cooperativity in high affinity binding sites for 7-hydroxybutyric acid in rat brain. Neurochem. Res. 8, 113 120.
Nanopoulos D., Belin M. F., Maitre M., Vincendon G, and Pujol J. F. (1982) Immunocytochemical evidence for the existence of GABAergic neurons in the nucleus Raphe Dorsalis. Possible existence of neurons containing serotonin and GABA. Brain Res. 232, 375-389. Pitts F. N, and Quick C. (1965) Brain succinic semialdehyde dehydrogenase. I. Assay and distribution. J. Neurochem. 12, 893-900. Roth R. H. and Gairman N. J. (1969) Conversion in vivo ot 7-aminobutyrate to 7-hydroxybutyrate in the rat. Biochem. Pharmac. 18, 247--280. Rumigny J. F., Maitre M., Cash C. and Mandel P. (1980) Specific and non specific succinic semi aldehyde reductases from rat brain. Isolation and properties. FEBS Lett. 117, 111-116. Rumigny J. F., Maitre M., Cash C. and Mandel P. (19811 Regional and subcellular localization in rat brain of the enzymes that can synthesize 7-hydroxybutyric acid. J. Neurochem. 36, 1433 1438. Rumigny J. F., Cash C., Mandel P. and Maitre M. (1982) Ontogeny and distribution of specific succinic semialdehyde reductase apoenzyme in the rat brain. Neurochem. Res. 7, 555-561. Sternberger L. A. (1970) The unlabelled antibody enzyme method of immunochemistry. J. Histochem. Cytochem. 8, 315 325. Sternberger L. A. (1979) lmmunofluorescence in lmmunocytochemistry, p. 42. John Wiley Medical Publication, U.S.A., Second Edition. Tabakoff B. and Von Wartburg J. P. (1975) Separation of aldehyde reductases and alcohol deshydrogenase from brain by affinity chromatography: Metabolism of succinic semialdehyde and ethanol. Biochem. biophys. Res. Commun. 63, 957-966. Tappaz M. L.. Brownstein M. J. and Kopin I. J. (1977) Glutamate decarboxylase (GAD) and ~,-aminobutyric acid (GABA) in discrete nuclei of hypothalamus and substantia nigra. Brain Res. 125, 109-121. Weissmann-Nanopoulos D., Rumigny J. F., Mandel P., Vincendon P. and Maitre M. (1982) Immunocytochemical localization in rat brain of the enzyme that synthesizes 7-hydroxybutyric acid. Neurochem. Int. 4, 523-529.
0197-0186/84 $3.00+ 0.00 Copyright ~, 1984 Pergamon Press Ltd
IMMUNOCYTOCHEMICAL EVIDENCE FOR THE PRESENCE OF ENZYMES SYNTHESIZING GABA A N D GHB IN THE SAME N E U R O N D. WEISSMANN-NANOPOULOS,M. F. BELIN*,P. MANDEL and M. MAITREt INSERM U44, 5 rue Blaise Pascal, 67084 Strasbourg Cedex and *INSERM U171, 8 avenue Rockfeller, 69008 Lyon, France (Received 27 July 1983; accepted 10 October 1983)
Abstraet--A specific and sensitive immunocytochemical double staining for visualization of glutamate decarboxylase (GAD) and semialdehyde succinate reductase (SSR2) in the same brain section has been developed. SSR2 is the enzyme responsible for the transformation of succinic semialdehyde into ?-hydroxybutyrate (GHB). GAD was detected using specific rabbit GAD-antibodies and unlabeled antibody enzyme peroxidase antiperoxidase, and SSR2 using specific guinea-pig SSR2 antibodies conjugate to a fluorescein-labeled second antibody. The coexistence of GAD and SSR2 in the same neuron was demonstrated by a peroxidase reaction superimposed on fluorescent compounds. Cell bodies containing both antigens were observed in the cerebellum, dorso-median hypothalamus and raphe nuclei. GHB is present in most GABA containing neurons. Some neurons contain only SSR2; these neurons may synthesize GHB by an active uptake of GABA.
The major route for GABA degradation occurs via transamination to succinate semialdehyde (SSA) in the mitochondria. SSA is dehydrogenated to form succinate (Pitts and Quick, 1965). However, succinate semialdehyde may also be rapidly converted by an N A D P H reduction reaction to 7-hydroxybutyrate (GHB) (Roth and Giarman, 1969). A specific succinate semialdehyde reductase (SSR2) which is responsible for GHB synthesis has been purified from both human and rat brain (Tabakoff and Von Wartburg, 1975; Cash et al., 1979; Rumigny et al., 1980). This enzyme, located in the cytoplasm of neurons, in dendrites and synapses (Weissmann-Nanopoulos et al., 1982) is of great interest since its product, GHB, has been demonstrated in the past few years to exhibit the principal characteristics of a putative neurotransmitter in the rat brain. In fact, the conditions of synthesis, release (Maitre and Mandel, 1982; Maitre et al., 1983), transport (Benavides et al., 1982a) and binding of GHB (Benavides et al., 1982b; Maitre et al., 1983a; Maitre et aL, 1983b) at the synaptic level suggest the possible existence of GHBergic synapses. Measurements of GHB turnover time in rat brain were established to be even more rapid than those reported for whole brain serotonin, dopamine or norepinephrine (Gold and Roth, 1977).
tTo whom correspondence should be addressed.
Thus, it was of interest to investigate a possible coexistence or segregation between G A D ("GABAergic neurons") and SSR2 in "GHBergic neurons". The latter probably requires the presence of GABA to synthesize succinate semialdehyde. To investigate this phenomenon, an immunocytochemical study was carried out using double labelling techniques with glutamate decarboxylase (GAD) and with specific succinate semialdehyde reductase (SSR2). Both antibodies were obtained and characterized in our laboratory (Nanopoulos el al., 1982; Rumigny et al., 1982). To avoid ambiguous results, antibodies against glutamate decarboxylase were obtained from rabbit and SSR 2 antibodies from guinea pig. This facilitates double labelling techniques. The experiments were carried out on three regions of the rat brain: cerebellum, median hypothalamus and raphe dorsalis, which are well characterized and rich in both enzymes (Rumigny et al., 1982). EXPERIMENTAL PROCEDURES
An antiserum to GAD purified from rat brain was produced in rabbits by subcutaneous injections of the antigen emulsified with Freund's adjuvant. The characteristics of the antigen and the antibodies were reported previously (Blinderman et al., 1978; Maitre et al., 1978; Nanopoulos et al., 1982). Pure SSR2 was obtained from rat brain using the method described by Rumigny et al. (1980). Antiserum to specific succinic semialdehyde reductase was raised in the guinea333
334
D. WE1SSVIANN-NANOPOULOSel al.
pig, as previously described for the anti-SSR2 prepared in the rabbit (Weissmann-Nanopoulos et al., 1982). Ten male rats (OFA, 200 g) were anaesthetized by an intraperitoneal injection of sodium pentobarbital (Nembutal, 50 mg/kg body weight) and then the tissues fixed by a 10min intra-cardiac perfusion of 150-200ml freshly prepared fixative composed of a phosphate buffer (100mM. pH 7.4) containing 4')~, paraformaldehyde, and 0.1~i; glutaraldehyde. The dissected brains were postfixed for 12 h in 4~!~; para[brmaldehyde dissolved in the same buffer. After rinsing in an ice-cold phosphate buffer (100 raM, pH 7.4) containing 15,'~, sucrose for at least 12 h, the midbrains were cut on a cryostat at - 18C (sections thickness set at 8 pm). Series of 2 4 consecutive sections were collected, rinsed in phosphate buffer saline (PBS), pH 7.4 and then incubated for 4 h in PBS containing 11}~,normal sheep serum (NSS). GAD immunoreactivity was revealed by the unlabeled antibody enzyme method of Sternberger (peroxidase-antiperoxidase technique, 1 9 7 0 ) . SSR~ immunoreactivity was demonstrated according to the indirect immunofluorescence staining (for details, see Sternberger, 1979). Sections were incubated for 48h at + 4 ' C in PBS containing GAD antiserum diluted 1:2000 and SSR 2 antiserum diluted 1 : 100. Then they were washed in PBS-I°o NSS and incubated for I hr with a swine antirabbit serum (Dakopatts, Denmark) diluted I : 50 with PBS. After washing briefly in PBS, the sections were incubated for l h with a rabbit peroxidase-antiperoxidase complex (Dakopatts, Denmark) diluted 1:50 in PBS. Following a further wash, the sections were treated for 15 min with Tris HCI buffer (50mM, pH 7.5) containing 3,3' diaminobenzidine-4 HC1 (7.5 mg/10 ml) and H202 (0.6 ml of 30°0 HeO2/10ml ). Finally, they were washed in Tris HCI buffer (50 raM, pH 7.5). rhen, the sections were incubated with fluorescein isothiocyanate conjugated goat antiguinea-pig serum immunoglobulin (Cappel) I : 100 in PBS + 0.1" o Triton X-100. Finally, the sections were rinsed again in PBS and mounted in a mixture of glycerin-PBS (3 : I, v/v). They were examined using a fluorescence microscope equipped for incident light illumination. Adjacent sections were either stained for GAD or SSR 2 and then treated identically. In control sections, experiments were performed using preimmune sera.
RESULTS
Methods
Experiments were performed to determine optimal conditions for simultaneous localization of the two antigens. Consecutive sections were incubated either with a n t i - G A D serum, or with anti-SSR2 serum. An optimal double staining on the same section was obtained by decreasing the peroxidase reaction with a shorter revelation time (10min). W h e n sections were incubated with the two antisera and revealed first by immunofluorescence and secondly by peroxidase, the fluorescent staining is altered by the immunocytochemical peroxidase staining procedures,
Double immuno~3'tochemical staining
In sections incubated with b o t h antisera, G A D as well as SSR 2 positive cells were detected simultaneously. A l t h o u g h the fluorescent staining was sometimes weak, n u m e r o u s cell bodies were observed to be reactive to b o t h antibodies. The cell bodies exhibited characteristic G A D positive reaction deposits superimposed o n t o the fluorescent c o m p o u n d making it sometimes difficult to observe the fluorescent dyes. The G A D - p o s i t i v e reaction product was more or less strong and dispersed over the entire cytoplasm (Fig. 1). In the cerebellar cortex, m a n y Purkinje cells were double stained (Fig. lAB). A b o u t 400~ of them exhibit only a SSR2 positive reaction product (Fig.
ICD). In the raphe nuclei, some multipolar neurons contain b o t h antibodies (Fig. 2). The G A D positive reaction product is dispersed in the cytoplasm (Fig. 2AC), and the SSR2 positive reaction p r o d u c t makes p u n c t a deposits a r o u n d the same cells (Fig. 2B). SSR2 can be or is within the cytoplasm (Fig. 2D). A few neurons are only G A D positive (Fig. 2AC) but 80 90,~0 of the cells are double stained. In the reticular formation, some neurons exhibit G A D positive puncta a r o u n d their soma (Fig. 2E), and SSR2 positive reaction products are present within the cytoplasm (Fig. 2F). In the dorso-median h y p o t h a l a m u s , a b o u t 90'~o of the cells are double stained (Fig. 3). The sections incubated with prei m m u n e sera were not stained.
DISCUSSION
This study was carried out using a n t i - G A D and anti-SSR 2 sera from two different species. The antiG A D serum was applied for identification of G A B A ergic neurons in the Purkinje cell of the cerebellum, the substantia nigra and the nucleus raphe dorsalis ( N a n o p o u l o s et al., 1982). The anti-SSR2 serum was applied in order to d e m o n s t r a t e the presence of SSR, in neurons of the hypothalamus, and the nucleus raphe dorsalis ( W e i s s m a n n - N a n o p o u l o s et al., 1982). In this study, second antibodies were specific respectively against rabbit IgG and guinea-pig IgG. Tissue sections were incubated with the two antibodies simultaneously. A n t i - G A D IgG were localized by the peroxidase antiperoxidase technique of Sternberger (1970) which allows a great dilution of the first a n t i b o d y (l:2000). The presence of anti-SSR2 was detected using fluorescent staining. This m e t h o d o l o g y is used in order to avoid interference with the per-
GABA and GHB synthesizing enzymes in the same neuron
335
"qr ~q
ml
P
B
gi
,
P
mi
P Fig. 1. Sections of the cerebellar cortex double stained with antibodies to GAD and SSR2. Bright field illumination (A, C:GAD) and incident-light fluorescence (B, D: SSR2). ml:molecular layer; gl:granular layer; p: Purkinje cell layer. Arrows: cell bodies immunoreactive both to GAD and SSR2. Scale: 20/~m.
oxidase reaction substrates on the fluorochrome. Nevertheless, the development of the peroxidase reaction product may partially mask the fluorescence. For this reason, the peroxidase reaction was only developed during a short period of time, and with a low
concentration of H202. A soma exhibiting a high GAD-positive staining may prevent the access to the second antiserum with its peroxidase deposits. Therefore, a pure GAD-positive cell observed in a region showing, on consecutive sections, SSR2-positive
336
D. WEISSMANN-NANOPOULOSel al.
F
L
L;i
Fig. 2. Sections of the nucleus raphe magnus (A, B, C, D) and pontic reticular formation (E, F) double stained with antibodies to GAD and SSR2. Bright field illumination (A, C, E:GAD) and incident light fluorescence (B, D, F:SSR2). Scale: 10/~m. Arrows show cell immunoreactive both to GAD and SSR 2. Note in A and C, two cell bodies only reactive to GAD.
staining, may be a G A B A element containing inaccessible SSR2 sites. On serial sections, some neurons appear to be only GAD-positive. In this case, we can exclude the presence of SSR2. The double stained cells and the SSR2-positive cells correspond respectively to cells containing G A D and SSR2 alone. No staining is observed on sections treated with preimmune sera. Cells containing both antigens G A D and SSR, are present in the cerebellum, hypothalamus and
raphe nuclei. In the hypothalamus, the majority of these cells are double stained. This result is consistent with a high SSR~ (Rumigny e t al., 1981) and a high G A D activity (Tappaz e t al., 1977) in this area. In the nucleus raphe dorsalis and magnus, most of the cells appeared to be reactive to both antisera. A few neurons are only SSR2-positive. In the cerebellum, several Purkinje cells are double stained; however, some of them are only SSR2 positive.
GABA and GHB synthesizing enzymes in the same neuron
337
Fig. 3. Sections of the dorsomedial nucleus of the hypothalamus (premammilary region) double stained with antibodies to GAD and SSR. Bright field illumination (A, C:GAD) and incident light fluorescence (B, D:SSR2). Scale: 10 #m. Arrows show the cell bodies reactive both to GAD and SSR 2.
The coexistence of the two antigens in most of the neurons in the hypothalamus and raphe area suggests that G A B A is the precursor of G H B in these neurons. As G H B has been qualified as a possible neurotransmitter, these neurons can be either GABAergic and/or " G H B e r g i c " . In the neurons containing only SSR 2, G H B may be synthesized by an active uptake of G A B A from the surrounding cells. In that case, this group of neurons will be purely " G H B e r g i c " . Cells exhibiting SSR2 positive puncta at the periphery of a G A D positive soma could represent a G A B A containing neuron receiving a " G H B e r g i c " input.
In summary, this study demonstrates that G H B is present in most of the G A B A containing neurons. G H B may be a major catabolite of G A B A in neurons while the oxidative route of SSA is predominant in glial cells. Taking into account the potential neurotransmitter characteristics of G H B , this compartmentation might have an important functional role for neuronal activity. REFERENCES
Blindermann J. M., Maitre M., Ossola L. and Mandel P. (1978) Purification and some properties of L-glutamate
338
D. WEISSMANN-NANOPOULOSet al.
decarboxylase from human brain. Eur. J. Biochem. 86, 143 152. Benavides J., Rumigny J. F., Bourguignon J. J., Wermuth C. G., Mandel P. and Maitre M. (1982a) A high affinity, Na + dependent uptake system for 7-hydroxybutyrate in membrane vesicles prepared from rat brain. J. Neurochem. 38, 1570-1575. Benavides J., Rumigny J. F., Bourguignon J. J., Cash C., Wermuth C. G., Mandel P., Vincendon G. and Maitre M. (1982b) High affinity binding site for 7-hydroxybutyrate acid in rat brain. Life Sci. 30, 953-961. Cash C., Maitre M. and Mandel P. (1978) Purification de deux semiald6hydes succinique r6ductases de cerveau humain. C.r. hebd. S~anc. Acad. Sci., Paris 286, 1829-1832. Cash C. D., Maitre M. and Mandel P. (1979) Purification and properties of two NADPH-linked aldehyde reductases which reduce succinic semialdehyde to 4 hydroxybutyrate. J, Neurochem. 33, 1169-1175. Gold B. I. and Roth R. H. (1977) Kinetics of in vivo conversion of ~/-[3H] aminobutyric acid to 7-[3H] hydroxybutyric acid by rat brain. J. Neurochem. 28, 1069-1073. Maitre M., Blindermann J. M., Ossola L. and Mandel P. (1978) Comparison of the structures of L-glutamate decarboxylases from human and rat brain. Biochem. biophys. Res. Commun. vol. 85, 885 890. Maitre M. and Mandel P. (1982) Lib6ration de 7-hydroxybutyrate calcium-dependante apr6s d~polarisation de coupes de cerveau de rat. C.r. hObd. Skanc. Acad. Sci., Paris 295, 741-743. Maitre M., Cash C., Weissmann-Nanopoulos D. and Mandel P. (1983) Depolarization evoked release of ),-hydroxybutyrate from rat brain slices. J. Neurochem. 41, 287 290. Maitre M., Rumigny J. F., Cash C. and Mandel P. (1983a) Subcellular distribution of 7-hydroxybutyrate binding sites in rat brain-principal localization in the synaptosomal fraction. Biochem. biophys. Res. Commun. 110, 262-265. Maitre M., Rumigny J. F. and Mandel P. (1983b) Positive cooperativity in high affinity binding sites for 7-hydroxybutyric acid in rat brain. Neurochem. Res. 8, 113 120.
Nanopoulos D., Belin M. F., Maitre M., Vincendon G, and Pujol J. F. (1982) Immunocytochemical evidence for the existence of GABAergic neurons in the nucleus Raphe Dorsalis. Possible existence of neurons containing serotonin and GABA. Brain Res. 232, 375-389. Pitts F. N, and Quick C. (1965) Brain succinic semialdehyde dehydrogenase. I. Assay and distribution. J. Neurochem. 12, 893-900. Roth R. H. and Gairman N. J. (1969) Conversion in vivo ot 7-aminobutyrate to 7-hydroxybutyrate in the rat. Biochem. Pharmac. 18, 247--280. Rumigny J. F., Maitre M., Cash C. and Mandel P. (1980) Specific and non specific succinic semi aldehyde reductases from rat brain. Isolation and properties. FEBS Lett. 117, 111-116. Rumigny J. F., Maitre M., Cash C. and Mandel P. (19811 Regional and subcellular localization in rat brain of the enzymes that can synthesize 7-hydroxybutyric acid. J. Neurochem. 36, 1433 1438. Rumigny J. F., Cash C., Mandel P. and Maitre M. (1982) Ontogeny and distribution of specific succinic semialdehyde reductase apoenzyme in the rat brain. Neurochem. Res. 7, 555-561. Sternberger L. A. (1970) The unlabelled antibody enzyme method of immunochemistry. J. Histochem. Cytochem. 8, 315 325. Sternberger L. A. (1979) lmmunofluorescence in lmmunocytochemistry, p. 42. John Wiley Medical Publication, U.S.A., Second Edition. Tabakoff B. and Von Wartburg J. P. (1975) Separation of aldehyde reductases and alcohol deshydrogenase from brain by affinity chromatography: Metabolism of succinic semialdehyde and ethanol. Biochem. biophys. Res. Commun. 63, 957-966. Tappaz M. L.. Brownstein M. J. and Kopin I. J. (1977) Glutamate decarboxylase (GAD) and ~,-aminobutyric acid (GABA) in discrete nuclei of hypothalamus and substantia nigra. Brain Res. 125, 109-121. Weissmann-Nanopoulos D., Rumigny J. F., Mandel P., Vincendon P. and Maitre M. (1982) Immunocytochemical localization in rat brain of the enzyme that synthesizes 7-hydroxybutyric acid. Neurochem. Int. 4, 523-529.