76 Neuroscience Letter~', 106 (1989) 76 82

Elsevier Scientific Publishers Ireland Ltd.

NSL 06443

Co-localization of tyrosine hydroxylase and GABA immunoreactivities in human cortical neurons

S. Trottier 1, M. G e f f a r d 2 a n d B. E v r a r d 1

: INSERM U97, Centre Paul Broca and H6pital Sainte-Anne, Paris (France; and :lnstitut de Biochimie

Cellulaire et Neurochimie and I N S E R M U259, Bordeaux (France)

(Received 27 September 1988: Revised version received 17 July 1989: Accepted 17 July 1989)

Key words." Tyrosine hydroxylase: 7-Aminobutyric acid: Cortical neuron: Human

Samples of human cerebral cortex were stained immunocytochemically for tyrosine hydroxylase (TH)

and 7-aminobutyric acid (GABA), TH-positive neurons were in small number and predominated in the deep infragranular layers V VI contrasting with numerous GABA-posit ive neurons scattered in all layers.

Co-localization of TH- and GABA-like immunoreactivities in a single cell was studied by the double immunolabeling technique with the elution-restaining procedure. Only 50% of the TH-positive neurons

also expressed GABA-like immunoreactivities. The two markers were detectable in the somata and not

ill the processes of the cells. The double-labeled cells were mainly fu~iform and medium-sized and were observed in layer VI. These observations suggest that the TH-positive cells form a mixed neuronal popula-

tion, only a part of which corresponds to the GABAergic class of intrinsic interneurons.

The idea that all catecholaminergic (CA) neurons are located in the mammalian brainstem (for review see ref. 1) was brought into question by the unexpected report of tyrosine hydroxylase-like immunoreactivity (TH-LI) neurons in the basal fore- brain of monkey [8]. Since these observations, TH-LI neurons have been identified outside the brainstem in the human [5] and adult rat cerebral cortex [10]. The coexis- tence of TH- and G A D (glutamic acid deshydrogenase)- or GABA (y aminobutyric acid)-LI in the same cell has been reported in various brain regions in the rat [4, 9, 11]. We were interested to see if the co-localization of TH- and GABA-LI described in rodents could exist in human cerebral cortex.

Subjects for the present study were 3 epileptic patients who underwent surgery for chronic focal seizures; samples were obtained from the second temporal gyrus (area 21) and frontal gyrus (area 9). The cortical samples were fixed immediately after sur- gical removal by immersion in a fixative solution containing 4% paraformaldehyde and 1% or 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH = 7.4) for 6 h at + 4~-'C, After several rinses for 24 h, the cortical blocks were cut into 40-~m-thick vibratome sections. All sections were incubated in 1% sodium borohydride (NaBH4, Wako) in

Correspondence. S. Trottier, INSERM U97, Centre Paul Broca, 2ter rue d'Al6sia, 75014 Paris, France.

0304-3940/89/$ 03.50 :~': 1989 Elsevier Scientific Publishers Ireland Ltd.

77

0.1 M PBS for 20 min at room temperature. They were processed for TH and GABA labeling using two polyclonal rabbit antisera (TH-antiserum, J.Boy, Reims, France and GABA-antiserum [21]): they have been employed in previous studies and meth- ods of antigene purification, rabbit immunization and criteria of antisera specificity have been reported elsewhere [2, 12, 13, 14]. Co-localization of TH- and GABA-LI was demonstrated by the double-labeling technique using the immunoperoxidase staining followed by elution of the first sequence antibodies and the immunofluores- cence staining.

First method: (a) TH-LI was first visualized by incubating the sections in TH anti- serum diluted 1:2000 (in PBS 0.02 M, pH 7.4) for 24 h at +4°C, followed by the streptavidin-biotinylated peroxidase complex (Amersham) procedure; peroxidase ac- tivity was developed with the 4-chloro-1-napthol (4-CN) (40 mg dissolved in 0.1 M tris buffer, pH 7.6 with 0.5 ml dimethylformamide) and hydrogen peroxide (0.075%); TH-LI neurons were photographed and their location in the section was identified with an X-Y drawing plotter. (b) The sections were then subjected to the elution- restaining technique of Tramu et al. [15]: immersion time in acetone and KMnO~ H2SO 4 solution was determined to obtain a complete disappearance of the immuno- labeling. To be sure that all antibodies had been completely eluted and that no new aspecific binding sites in the tissue [3] had occurred which could react with the antibo- dies used in the second sequence the sections were incubated in normal rabbit serum and then in fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit antibody (Byosis) diluted 1:100: no fluorescent cell was observed. (c) GABA-LI was revealed by using GABA-antiserum diluted to 1:2000 (in Tris-buffer 0.05 M, pH 7.4) for 24 h at +4°C and the indirect immunofluorescence technique with a secondary goat anti-rabbit IgG conjugated with FITC diluted to 1:100 for 2 h at + 37°C; the fluores- cent GABA-positive neurons corresponding to the location of TH-LI neurons were again photographed.

Second method." (a) GABA-LI was first revealed by the immunoperoxidase staining using the GABA antiserum diluted to 1:4000, the streptavidin-biotin complex and the 4-CN. (b) The elution-restaining procedure was performed but we tried in this case to obtain a decrease in the immunolabeling intensity and not a complete disap- pearance of the labeling; to check that no interaction could occur between free bind- ing sites of the GABA antiserum used in the first sequence and the FITC-conjugated goat anti-rabbit used in the second sequence the sections were incubated in FITC- conjugated goat anti-rabbit antibody (diluted 1:100); the control with normal rabbit serum was also done: all controls were negative, no immunofluorescent cell was encountered. (c) TH immunolabeling could then be carried out by incubating the sec- tions in TH antiserum (diluted to 1:400) and then in FITC-conjugated goat anti-rab- bit antibody.

Two other immunocytochemical controls were carried out." first, sections were singly stained according to the immunoperoxidase or immunofluorescence procedure to demonstrate immunoreactivity in the tissue for a given antibody; second, each anti- body was replaced by buffer in each step of the immunocytochemical procedure to check the specificity of the antibodies used.

78

Fig. 1. A E: TH-LI neurons in human area 21 in different cortical layers. A: arrangemenl in cluster in layer VI ( x 215). B: multipolar neuron (layer V). C D: fusiform neurons with varicose processes vertically

oriented (layer VI). E: bipolar-like neuron horizontally oriented in layer VI. (B-E: × 450). F,G: GABA-LI neurons in human area 9. F: small round positive somata in layer II. G: immunonegat ive pyramidal cells (P) surrounded by numerous GABA-immunoreact ive puncta (probably terminals) in layer V ( x 280).

79

Clusters of TH-LI nerve cell bodies were concentrated in the deep infragranular layers and mainly in layer VI (Fig. 1A), although a few individual TH-LI cells (less than 10% of the total population) were present in the superficial layers III-IV of the two granular areas. Most of the somata were fusiform (Fig. 1C-E) while the others appeared round, triangular or multipolar (1 B) and of small or medium size (long axis: 13-25 pm, minor axis: 6.5-12.5/~m). The somata gave rise to 2 or 4 processes, sometimes arising from a slender cone-shaped extension of the soma (Fig. 1C), with branching extending for a great distance from the cell body (300-500/tm); usually the processes showed numerous small varicosities. In general, the somata did not show preferential orientation of their long axes except for some fusiform cell bodies which were either parallel to the white matter (Fig. 1E) or vertically situated (Fig. 1C,D).

Numerous GABA-LI neurons were present in all layers exhibiting a peculiar lami- nar distribution with highest densities in layer II but also in layer VI (Fig. 1F). Their somata were multipolar, round or fusiform and small or medium-sized (10-15/lm/ 14-25/lm). Usually several processes emerged from the soma and often displayed varicosities along their course (Fig. 1F). GABA-LI terminals were densely distri- buted throughout the whole neuropile and surrounded GABA-negative immunos- tained cells (Fig. 1G).

With the double-staining procedure, not all TH-LI neurons were found to contain GABA-immunoreactive material and the double-labeled cells constituted only about 50% of the total TH-positive cell population. The double labeling was most often localized in the nerve cell bodies; usually the processes emerging from the cell body were labeled only with the TH antiserum and not with the GABA antiserum, even if the sequence of the immunocytochemical labelings was alternated (Fig. 2A,C,D). Preliminary analysis indicated that (i) the double-labeled cells were most frequently small and fusiform or round and less frequently large multipolar cells (ii) they were observed in the deep layer VI.

These results demonstrate that the procedures employed allow the immunocyto- chemical detection of both TH- and GABA-LI in a single tissue section. The validity of the double-labeling technique is based on the good conservation of the tissue anti- gene immunoreactivities, and on the selectivity of the elution procedure without denaturation of the second antigene to be immunocytochemically detected. We found that low concentrations of glutaraldehyde (1%) allow good preservation of GABA and do not cause evidently drastic loss of TH antigenicity. Selectivity of the elution was investigated by the immunocytochemical controls demonstrating the ab- sence of interactions between the antibodies used in the first and second staining se- quence.

Our results give the first demonstration of colocalization of TH- and GABA-LI in the human cerebral cortex and estimate that only 50% of the TH-LI neurons be- long to GABA-immunolabeled neurons suggesting that the TH- and GABA-LI neu- rons are partially overlapping populations. It must be noted, however, that caution is required when we give a quantitative interpretation of the data: the lack of GABA- LI in some TH-LI neurons may be related to levels of GABA too low to be immuno-

80

r ....

i I I i

i

L

i

I I

Fig. 2. Pairs of micrographs showing colocalization of TH- and GABA-LI in a single neuron in layer VI. A D: cell immunostained (area 21) first with rabbit anti-TH and 4-CN and after antibody complete elution with rabbit ant i -GABA and FITC-conjugated antibody. Note the non-visualization of the process on the GABA fluorescence picture (A,B: x 225; C,D: x 450). E,F: cell stained first with ant i -GABA and 4-CN (E, dark-field illumination) and after partial elution with anti-TH and FITC antibody (F). The pro- cesses were not detectable with the GABA-ant iserum ( x 450). Asterisks indicate blood vessel profiles as

landmarks.

81

cytochemically detected. Moreover, the elution-restaining procedure could be troub-

lesome for the detection of the second antigene: to check that there was no order-

dependent labeling, we reversed the labelings (TH-GABA or G A B A - T H ) and we obtained similar results. The significance of the TH labeling in cerebral cortex neu- rons in primate or human is yet unknown. Although it has been suggested that some periglomerular neurons of the rat olfactory bulb could be both GABAergic and

dopaminergic [9], neurotransmitters related to TH activity or catecholamine (CA)-

synthesizing enzymes (dopadecarboxylase, dopamine-fl-hydroxylase) have never been identified in human or primate TH-LI cortical neurons [5, 8]. Similarly, we have

never been able to visualize immunoreactive cell bodies in the human cerebral cortex

by using an antibody directed against dopamine-fl-hydroxylase although the cortical

noradrenergic terminals were easily labeled. Recently it has been discovered that

most peptide-containing neuronal somata (cholecystokinin, somatostatin, neuropep-

tide Y) of the primate cerebral cortex distributed mainly in layers II and VI belong

to the population of intrinsic neurons that invariably contain GABA as well [7]. Thus, it will be interesting to investigate if (i) the subclass of neurons immunoreactive for both TH and GABA contain also one or several of these peptides and (ii) the

subclass of neurons positive for TH alone could be characterized or not by its neuro-

peptide content. We are aware of a possible bias: these results were obtained from cerebral tissue samples of epileptic patients undergoing surgery but this tissue could

be fixed without the inevitable delay of post-mortem material, thus allowing the

GABA immunocytochemistry procedure. However, it appears that there are striking similarities between our results and those obtained from human [5, 13] and animal

[6, 10} material. We can only suggest at this step of our research that the epileptic

procegs does not seem to prevent the expression of TH and GABA in neurons of the human cerebral cortex.

We thank Dr. B. Berger, P. Gaspar and C. Verney for fruitful discussions, Pr J.-P. Chodkiewicz for surgery and J. Everett for improving the English.

1 Bj6rklund, A. and Lindvall, O., Catecholaminergic brainstem regulatory systems. In F.E Bloom (Ed.), Handbook of Physiology, The Nervous System IV, Oxford University Press, New York, 1986, pp. 155-235.

2 Berod, A., Hartman, B.K., Keller, A., Joh, T.H. and Pujol, J.H., A new double labeling technique using tyrosine hydroxylase and dopamine-fl-hydroxylase immunohistochemistry: evidence for dopami- nergic cells lying in the pons of the beef brain, Brain Res., 240 (1982) 235-243.

3 Ciofi, P., La cholecystokinine dans le syt6me nerveux central: Immunohistochimie, colocalisation avec divers neuropeptides dans l'hypothalamus de quelques mammif6res dont l'homme. Th6se Sciences, Universit6 de Lille, 1987. 122 pp.

4 Everitt, B.J., H6kfelt, T., Wu, J.J. and Goldstein, M., Coexistence of tyrosine hydroxylase-like and gamma-aminobutyric acid-like immunoreactivities in neurons of the arcuate nucleus, Neuroendocrino- logy, (1984) 189-191.

5 Gaspar, P., Berger, B., Febvret, A., Vigny, A., Krieger-Poulet, M. and Borri-Voltanorni, C., Tyrosine hydroxylase-immunoreactive neurons in the human cerebral cortex: a novel catecholaminergic group?, Neurosci. Lett., 80 (1987) 257-262.

6 Houser, C.R., Hendry, S.H., Jones, E.C. and Vaughn, J.E., Morphological diversity of immunocyto- chemically identified neurons in the monkey sensory-motor cortex, J. Neurocytol., 12 (1983) 6 l 7-638.

82

7 Jones, E.G. and Hendry, S.H.C., Peptide-containing neurons of the primate cerebral cortex. In J.B. Martin and J.D. Barchas (Eds.), Neuropeptides in Neurologic and Psychiatric Disease, Raven, New York~1986, pp. 163 178.

8 K6hler, C., Everitt, B.J., Pearson, J. and Goldstein M., Immunohistochemical evidence for a new group of catecholamime-containing neurons in the basal forebrain of the monkey, Neurosci. Lett., 37 (1983) 161 166.

9 Kosaka, T., Hataguchi, Y., Hama, K., Nagatsu, I. and Wu, J.Y., Coexistence of immunoreactivities for glutamate decarboxylase and tyrosine hydroxylase in some neurons in the periglomerular region of the rat main olfactory bulb: possible coexistence of gamma-aminobutyric acid (GABA) and dopa- mine, Brain Res., 343 (1985) 166 17[.

10 Kosaka, T , Hama~ K. and Nagatsu, I., Tyrosine hydroxylase immunoreactive intrinsic neurons in the rat cerebral cortex, Exp. Brain Res., 68 (1987) 393~405.

11 Kosaka, T., Kosaka, K., Hataguchi, Y., Nagatsu, I., Wu, J.Y. and Ottersen, O.P., Catecholaminergic neurons containing GABA-Iike and/or glutamic acid decarboxylase-like immunoreactivities in various brain regions of the rat, Exp. Brain Res., 66 (1987) 191 210.

12 Mouchet, P., Manier, M., Dietl, M., Feurstein, C., Berod, A., Arluison, M., Denoroy, L. and Thibault, J., lmmunohistochemical study ofcatecholaminergic cell bodies in the rat spinal cord~ Brain Res. Bull., 16 (1986) 341 353.

13 Schiffmann, S., Campistron, G., Tugendhaft, P., Brotchi, J., Flament-Durand, J., Geffard, M. and Vanderhaeghen, J.-J., lmmunocytochemical detection of GABAergic nerve cells in the human tempo- ral cortex using a direct gamma-aminobutyric acid antiserum, Brain Res., 442 (1988) 270 278.

14 Seguela~ P, Geffard, M., Buijs, R. and Le Moal, M., Antibodies against gamma-aminobutyric acid: specificity studies and immunocytochemical results~ Proc. Natl. Sci. U.S.A., 81 (1984) 3888 3892.

15 Tramu, G., Pillez, A. and Leonardelli, J., An efficient method of antibody elution for the successive or simultaneous localization of two antigens by immunocytochemistry, J. Histochem. Cytochem., 26 (1978) 322 324.