Immunocytochemical evidence for an afferent GABAergic neurotransmission in the guinea pig vestibular system

Brain Research, 589 (1992) 341-348 © 1992 Elsevier Science Publishers B.V. All rights reserved 0006-8993/92/$05.00

341

BRES 25340

Immunocytochemical evidence for an afferent GABAergic neurotransmission in the guinea pig vestibular system Ivfin L 6 p e z a, J a n g - Y e n W u b a n d G r a c i e l a M e z a

a

a Departamento de Neurociencias, Instituto de Fisiologfa Celular, UNAM, Mdxico, D.F. (Mdxico) and t, Department of Physiology and Cell Biology, The University of Kansas, Lawrence, ~KS (USA) (Accepted 9 June 1992)

Key words: Vestibular glutamate decarboxylase; Vestibular GABA transaminase; GAD-like immunoreactivity; GABA transaminase-like immunoreactivity; Vestibular cristae ampullaris; Vestibular ganglion cell; Guinea pig vestibule

To implicate y-aminobutyric acid (GABA) as an afferent neurotransmitter (AN), the localization of GABA synthesizing and degradation en~mes; L-glutamate decarboxylase (GAD) and GABA transaminase (GABA-T) was investigated by light and electron microscopy immunocytochemistry in guinea pig vestibular cristae and ganglion cells (GC). GAD-like immunoreactivity was exclusively confined to the sensory hair cell (HC) cytoplasm, suggesting that GAD synthesizes GABA in the HC. GABA-T-like immunoreactivity was found within HC, nerve calyces, nerve fibers, and GC, suggesting its participation in terminating transmitter action. These results demonstrate the existence of a GABAergic system in the guinea pig vestibule and strongly support GABA as a vestibular AN.

y-Aminobutiric acid (GABA) has long been suspect of being a vestibular hair cell neurotransmitter. GABA synthesis from glutamic acid was first reported in the fish vestibule 6 followed by experiments in the chick labyrinth 2° and more recently in the labyrinth of other vertebrate species such as green frogs ~°'~3, rats L~° and guinea pigs I°, through a well characterized glutamate decarbo~lase (GAD) mediation ~9-2~. GAD vestibular hair cell localization has been suggested because it is demonstrable in the efferent bouton.lacking frog basilar papilla 6 and due to its presence from early development in homogenates of the efferent terminal-devoid vestibular sensory periphery, in chicks 23, rats ~ and guinea pigs9 having already mature hair cells. Moreover, in streptomycin treated guinea pigs showing degeneration of vestibular hair cells and intact nerve terminals, GAD vestibular depletion could be demonstrated 24. Further support to the GABA hypothesis has come from the work of Felix and Ehrenberger 7, reporting that microiontophoretic application of GABA increased single unit spontaneous activity which was diminished by picrotoxin and bicuculline application in

the cat macula sacculi 7 thus suggesting the involvement of GABAA receptors in this response, notion supported by the biochemical demonstration and pharmacological characterization of GABA postsynaptic receptors in chick vestibular membranes 22. Our inference of the hair cell localization of GAD in the chick vestibule was recently confirmed by an immunocytochemically demonstrable GAD localization in the chick vestibular hair cell cytoplasm 2~. Further studies of our laboratory showed a GABA like immunoreactivity (GABA-LIR) in the hair cell cytoplasm, and in some nerve fibers and afferent calyces in guinea pig vestibulary cristae ~Ll2. In controversy, GABA-LIR was reported by others in fibers and bouton-type terminals 3° or in calyceal nerve endings surrounding type 1 hair cells in the rodent labyrinth 5 although with a quite (already discussed) 12 different methodology. In order to contribute to elucidate this controversy and the origin and fate of this GABA pool and since GABA may not be a good GABAergic cell marker due to its possible redistribution or metabolization during the preparation of the tissue, GAD and GABA-trans-

Correspondence: G. Meza, Departamento de Neurociencias, IFIC, UNAM, Apartado Postal 70-600, 04510 M~xico D.F., Mt~xico. Fax: 525-548-0387.

4~

343 aminase (GABA-T), were immunocytochemically and simultaneously investigated in the guinea pig vestibular system. GAD because, as the GABA synthesizing enzyme, it has been clasically considered the marker of choice for GABAergic neurons and GABA-T because, although not usually a GABAergic cell marker, its demonstration may help to explain GABA-LIR distribution in several end organs tt. The methodology followed and the results obtained are hereby described. Some of our findings have previously been presented in abstract form ~4. Six healthy (both sexes) young pigmented guinea pigs (Cavia cobaya) (200-250 g wt.) were used in this study. They were deeply anesthetized with chloral hydrate (1.5 g / k g wt.), exsanguinated transcardiaily with 150 ml of isotonic saline solution, followed by 750 ml of 4% paraformadehyde/0.1% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.3) (SCB) delivered over a 15 rain time period. At the end of the perfusion the auditory bullae were opened, the ampullary cristae and the vestibular ganglion were removed and postfixed in the same fixative at 4°C for 2 h. For GAD immunolocalization, 0.01% of/3-mercapto-ethanol (/3ME) (Sigma, St. Louis, MO, USA) was added to the fixative mixture. After postfixation, the specimens were dehydrated in 100% ethanol and were embedded in paraffin (Paraplast, Polysciences). This procedure allows mounting in the desired orientation and pyramids can then be made and cut accordingly° The tissue was cut longitudinally (3 ttm) and the resulting sections were mounted on gelatin-coated slides. Sections were dewaxed and rehydrated in SCB for 30 min and incubated sequentially in the following solutions: 10% normal goat serum (Vector Labs) in 0.15% Triton X-100 (in SCB) for 30 min at 25°C; GAD or GABA-T antisera in SCB for 48 h at 4°C; biotynilated rabbit anti-goat IgG (for GAD antiserum) (Vector Labs., Inc.) for 60 min at 25°C; Vectastain reagent (Vector Labs., Inc.) for 60 min at 25°C; D A B / H 2 0 2 for 5 min at 25°C. Four rinses of 5 min with SCB were performed between each incubation. After the peroxidase reaction, the sections were coverslipped with Aqua-polymount (Polysciences) and examined under the light microscope. The immunoelectron microscopic localization of

GAD and GABA-T in the guinea pig cristae ampullaris was performed by Somogyi's methodology 2a with some modifications, i.e. the avidine-biotine-peroxidase method (ABC, Vector Labs., Inc. ) was used instead of peroxidase-antiperoxidase complex. Extensively characterized primary antisera against GAD and GABA-T raised in rabbit at dilutions between 1 : 500-1 : 2000 (GAD) and 1 : 2000-1 : 4000 (GABA-T) were used 3~'3z. Control experiments were run in alternate sections as follows: (1) normal rabbit serum or SCB instead of GAD antiserum, and normal goat serum or SCB instead of GABA-T antiserum were performed (negative controls) or (2) the reaction was achieved on transversal section of the cerebellum of the same animal (positive control). From the above methodology it is noticeable that some important modifications were made to the reported protocols 29'a°, i.e., sections of 3/zm were used instead of 15/zm 29'3° for better resolution, a paraffinpost-embedding instead of a pre-embedding (frozen tissue) 29'3° for better morphology, /3-mercaptoethanol was added to fixatives for GAD immunolocalization in order to protect sulfhydride radicals necessary for enzyme activity as described for GAD determination in the chick vestibule 19. SCB, instead of the traditional phosphate buffered saline, lowered the background. Triton X-100 concentration and time of incubation was considerably lowered since using previous reported protocols 29'a° led to deterioration of tissue in such a way that unrecognizable structures were found. The above led to an optimal reproducible trustable preservation of the specimens. The results obtained are the following.

GAD.like immunoreactivity (GAD-LIR). Fig. 1 shows that GAD-LIR is confined to the cytoplasm of both hair cell types (I and ll) (Fig. 1A) with equivalent staining intensity throughout the sensory epithelia in the cristae ampullaris of all specimens studied. No immunoreactivity was found either in supporting cells or in nerve fibers. Sections incubated without antisera (negative control) showed no GADLIR (Fig. 1B). Also, no immunoreactivity was found in vestibular ganglion cells (Fig. 3A). Immunoelectron microscopic localization of GAD-LIR was clearly re-

Fig. 1. GAD-like immunoreactivity in the guinea pig vestibular endorgans. A: immunoreactive hair cells (I and !I) are evenly distributed throughout the sensory epithelia of the cristae ampullaris. No immunoreactivitywas found either in supporting cells (stars) or in nerve fibers (double stars), cell debris (cd), hair tufts (ht). Bar--20 p.m. B: negative control section of the same cristae ampullaris as in A showing no immunoreactivity; at the base some melanin-containingcells are observed (m). GAD antisera was omitted in the immunoreaction. Bar -- 20 ~m. Cell debris (cd) and not hair tufts (ht) show some staining(compare with ht in B). A and B were Nomarski-illuminated.

p ~ ! m o se~ ~s!~ue I-YEr~D "All^!~e~aounmm! ou 8u!~oqs y u? se s!~ellndme ~elSl~a ames jo uol~aas :8 "m~ OZ = ~e8 "(~e~s) SilVa ~u!l~oddn, u! punoj se~ X~!^!],~,aounmm! ou .'(spe~q~o~ae) ~ q ~ ~ , u St~Ole pue (s~oaae) SilVa al~q aql s~ulpunoaans saaXlea u? '(11 pue !) mSeldo]~ '~. ) Sll~a a!eq ~mos jo ms~ido~fo oq~ u? punoj s! Al!^!loe~aounmm! :Y 'sueS~opua ~lnq!~s~A 8!d eau!n8 aq| u! X~!^llaea~ounmm ! ~qll'!"YBVO "Z S!d

~,~

~~

~

i~!~

~~ ~ ~

~

~ ~

~ ~

~

~

~%~~

4

....

~ J ~ t ~-

0

..

VZ

345 stricted to the cytoplasm of hair cells while nerve calyx and supporting cells were unstained (Fig. 4A).

GABA- T-Iike immunoreactivity (GABA- T-LIR). In Fig. 2A, it is shown that GABA-T-LIR was mainly localized within some afferent calyces surrounding type I hair ceils, in the cytoplasm of some hair cells ~tP.d in some, ¢:~" . . . '¢viitl~,llt . ~,:,,L. ~'~.an ,t. . .~.l l lt..,__ t a o ~.. ,.l O i . l l .l U. U t l l U strorfla. No immunoreactivity was found in supporting cells. Moreover, ganglion cells' cytoplasm showed strong GABAT-LIR (Fig. 3B). A characteristic negative control is shown in Fig. 2B. At the electron microscopic level, GABA-T-LIR was present mainly in the nerve calyces, whereas hair cells and supporting cells appeared unstained (Fig. 4B). Sections of the cerebellum showed G! .)-LIR and GABA-T distribution as described in the literature (not shown) 4'27. This is the first report in the literature to study simultaneously the immunocytochemical localization of the enzymes of GABA synthesis and degradation (GAD and GABA-T, respectively) in mammalian ampuilary

cristae and ganglion cells. Both findings provide definitive evidence that GABA fulfills some important criteria for its classification as an afferent neurotransmitter in this system. The exclusive hair cell localization of GAD-LIR (by light and electron microscopy Figs. 1A and 4B) obtained in our study is consistent with that described recently in the guinea pig utricuiar macuiae ~ and the vestibule of the chick 29 and strongly supports our biochemical data L9,~°.13.19-21 and our steady postulation for a GABA afferent neurotransmitter role in the vestibule. Additionally, they help to explain the presence of GABA-LIR in the guinea pig vestibular hair cell cytoplasm ~L~2. GABA-T-LIR distribution resembles GABA-LIR in the guinea pig vestibular crista ampuilaris ~2, suggesting that GABA-T found in hair cells, in nerve calyces and nerve fibers probably represents sites for degradation of GABA. Furthermore, the afferent origin of these fibers was confirmed when GABA-T-LIR (and not GAD-LIR) was found in ganglion cells, which are traditionally their source.

Fig. 3. Absence of GAD-LIR in vestibularganglioncells. A: section from the ganglionof the same animal processed for GAD immunocytochemistry. Bar = 10 /~m. B: GABA-T-like immunoreactivity in the vestibular ganglion. Section from the ganglion of same animal processed for GABA-T showing a view of vestibular ganglion cells (stars). Almost all cells show GABA-T-like immunoreactivity, Bar = 10 /~m. A was Nomarski- and B was bright-field-illuminated.

346

~

.o

A o

Fig. 3. (continued).

In a given GABAergic system, the release of GABA by the nerve terminal and its subsequent binding to its receptor must be followed by a rapid inactivation of the transmitter which is in all cases a Na ÷ and energydependent uptake system for OABA and involves pre and postsynaptic elements sT. Degradation by GABA-T occurs later 4,8 and its sites have a similar distribution to uptake places Is'~. In our case, such an uptake mechanism was demonstrated in isolated vestibular endorgans 2s. GABA degradation, like other GABAergic systems 4, may occur in various, morphologically distinct sites, such as the hair cell itself, the afferent calyx and boutons, hence GABA-T distribution, although our electron microscopic findings of the GABA-T-LIR restricted to the nerve calyx suggest that this site is the principal place for the GABA degradation (Fig. 4b). Preliminary electron microscopy experiments sm demonstrated that non-GABAergic efferent boutons throughout the vestibular sensory periphery, in contrast, showed ,o GABA-LIR (not shown). In conclusion, GAD in the hair cell cytoplasm indicates that GABA is synthesized in this compartment; GABA-T-LIR in the hair cell cytoplasm, nerve calyx,

fibers and ganglion cells suggests that after GABA release and action upon its receptor it can be degraded in perisynaptic elements involved in GABA transmission. These results strongly support the afferent neurotransmitter role of GABA in the guinea pig vestibular system. Its observed excitatory effect in the cat macular afferent fiber ~ cannot be explained at the moment, but increasing number of reports describe a similar action of GABA in various regions of the CNS 2'3J8'2~. An accurate pharmacological characterization of the GABA receptor involved is needed and it is now in progress in our laboratory.

This work was supported, in part, by Direcci6n General del Personal Acad~mico (DGPA), UNAM and Consejo Nacional de Ciencia y Tecnologfa (CONACyT) M~xico, through Grants IN-201791 and Dl13903570, respectively, to G.M. and a doctorate Fellowship (CONACyT) to l.L.. We wish to thank Mrs. Ma. Teresa Cortez for excellent technical help and Mr. Rodolfo Paredes for photographic assistance.

347

Fig. 4. GAD-like immunoreactivity in hair cells and GABA-T-like immunoreaction in nerve calyces by electron microscopy. A: GAD-LIR in the cytoplasm of hair cell ! (HCI) and Ii (HCll) (asterisks). Notice that the nerve calyx (nc) and the supporting cells cytoplasm (sc) are unstained, n, nucleus. Bar -- I p,m. B: GABA-T-LIR is present only in the nerve calyx (nc) (small stars), whereas hair cell cytoplasm (HCI) and supporting cell (sc) are unstained, n, nucleus. Bar -- 1 p.m.

I Acufia, D., Aceves, C., Anguiano, B. and Meza, G. Vestibular site of action of hypothyroidism in the pigmented rat, Brain Res., 536 (1990) 133-138. 2 Alger, B.E. and Nicoll, R.A., Pharmacological evidence for two kinds of GABA receptors on rat hippocampal pyramidal cells studied in vitro, J. Physiol., 328 (1982) 125-141. 3 Brown, D.A. and Schofield, C.N., Depolarization of neurons in the isolated olfactory cortex of the guinea pig by 1,-aminobutyric acid, Br. J. Pharmacol., 65 (1979) 339-345. 4 Chan-Palay, V., Wu, J.-Y. and Palay, L., lmmunocytochemical localization of 1,-aminobutiric acid transaminase at cellular and ultrastructural levels, Proc. Natl. Acad. Sci. USA, 76 (1979) 20672071. 5 Didier, A., Dupont, J. and Cazals, Y., GABA immunoreactivi~.~. of calyceal nerve endings in the guinea pig vestibule, Cell Tissue Res., 260 (1990) 415-419. 6 Flock, A. and Lam, D.M.K., Neurotransmitter synthesis in inner ear and lateral line sense organs, Nature, 249 (1974) 142-144. 7 Felix, D. and Ehrenberger, K., The action of putative neurotransmitter substances in the cat labyrinth, Acta Otolaryngol., 93 (1982) 101-105. 8 Hyde, J.C. and Robinson, N., Localization of sites of GABA catabolism in the rat retina, Nature, 248 (1974) 432-433. 9 Iturbe, A.G. and Meza, G., Asymmetrical development of GABA and acetylcholine synthesis in guinea pig vestibule, Int. J. Neurosci., 4 (Suppl.) (1990) $32. 10 L6pez, I. and Meza, G., Comparative studies on glutamate decarboxylase and choline acetyltransferase activities in the vertebrate vestibule, Camp. Biochem. Physiol., 95B (1990) 375-379. 11 l..6pez, !. Juiz, J.M., Altschuler, R.A. and Meza, G., GABA-like immunoreactivity in the guinea pig vestibule: postembedding light

12 13 14

15 16 17

18 19 20 21

and electron microscopy finding, Sac. Neurosci. Ahstr., 15 (1989) 517. L6pez, 1., Juiz, J.M., Altschuler, R.A. and Meza, G., Distribution of GABA-like immunoreactivity in guinea pig vestibular cristae ampullaris, Brain Res., 530 (1990) 170-175. Ldpez, !. and Meza, G., Neurochemical evidence for afferent GABAergic and efferent cholinergic neurotransmission in the frog vestibule, Neuroscience, 25 (1988) 13-18. L6pez, !., Wu, J.-Y. and Meza, G., Glutamate decarboxylase and GABA-transaminase localization in the guinea pig vestibule. Immunocytochemical support for an afferent GABAergic neurotransmission, Sac. Neurosci. Abstr., 1 (1991)632. L.jungdahl, A., Seiger, A., H6kfelt, T. and OIson, L., ['~H]GABA uptake in growing cerebellar tissue: autoradiography of intraocular transplants, Brain Res., 61 (1973) 379-384. Marshal, J. and Voaden, M., An investigation of the cells incorpo,'-ting [3H]GABA and ['~H]glycine in the isolated retina of the rat, Exp. Eye Res., 18 (1974) 367-370. Martin, D.L. Carrier-mediated transport and removal of GABA from synaptic regions. In E. Roberts, T.N. Chase and D.B. Tower (Eds.), GABA in Nert,ous System Function, Vol. 5, Raven, New York, 1976, pp. 347-386. Mercuri, N.B., Calabresi, P., Stefani, A., Stratta, F. and Bernardi, G., GABA depolarizes neurons in the rat striatum: an in viva study, Synapse, 8 (1991) 38-40. Meza, G., Some characteristics of glutamic acid decarboxylase of chick ampullary cristae, ]. Neurochem., 43 (1984) 634-639. Meza, G., Carabez, A. and Ruiz, M., GABA synthesis in isolated vestibulary tissue of chick inner ear, Brain Res., 241 (1982) 157-161. Meza, G., Characterization of GABAergic and cholinergic neuro-

348 transmission in the chick inner ear, in D.G. Drescher (Ed.), Auditory Biochemistry, Charles C. Thomas, Springfield, IL, 1985 pp. 80-111. 22 Meza, G., Gonzalez-Viveros, M.T. and Ruiz, M., Specific [3H]aminobutyric acid binding to vestibular membranes of the chick inner ear, Brain Res., 337 (1985) 179-183. 23 Meza, G. and Hinojosa, R., Ontogenetic approach to cellular localization of neurotransmitters in the chick vestibule, Hearing Res., 28 (1987) 73-85. 24 Meza, G., Ldpez, l., Paredes, M.A., Pefialoza, Y. and Poblano, A., Cellular target of streptomycin in the internal ear, Acta Otolaryngol., 107 (1989) 406-411. 25 Meza, G., Wu, J.-Y. and L6pez, l., GABA is an afferent vestibular neurotransmitter in the guinea pig. Immunocytochemical evidence in the utricular maculae, Ann. N.Y. Acad. Sci., 656 (1992) 943-946. 26 Misgeld, U., Wagner, A. and Ohno, T., Depolarizing IPSPs and de~31arization by GABA of rat neostriatum cells in vitro, Exp. Brair, ~es., 45 (1982) 108-124. 27 Saito, K., Barber, R., Wu, J.-Y., Matsuda, T., Roberts, E. and Vaughn, J.E., Imlnunocytochemical localization of glutamate decarboxylase in rat cerebellum, Proc. Natl. Acad. Sci. USA, 71 (1974) 269-273.

28 Somogyi, P., and Takagi, H,, A note on the use of picric acidparaformaldehyde-glutaraldehyde fixative for correlated light and electron microscopic immunocytochemistry, Neuroscience 7 (1982) 1779-1783. 29 Usami, S., Hozawa, J., Tazawa, M., lgarashi, M., Thompson, G.C, Wu, J.-Y. and Wenthold, R.J., Immunocytochemical study of the GABA system in chicken vestibular endorgans and the vestibular ganglion, Brain Res., 503 (1989) 214-218. 30 Usami, S., Igarashi, M. and Thompson, G.C., GABA-like immunoreactivity in the squirrel monkey vestibular endorgans, Brain Res., 417 (1987) 367-370. 31 Wu, J.Y., Lin, C.T., Hwang, B., Lin, H.S. and Wei, S., Antibodies against enzymes synthesizing amino acids transmitters. In P. Panula, H. Paivarintia and S. Soinila (Eds.), Neurochemistry: Modern Methods and Applications, Alan R. Liss, New York, 1986, pp. 21-47. 32 Wu, J.Y., Lin, C.T., Lin, H., Xu, Y., Liu, J.W., Hwang, B.H. and Wei, S.C., Immunocytochemical characterization and immunohistochemical localization of glutamate decarboxylase and GABA transaminase in peripheral tissues. In S.L. Erdo and N.G. Bowery (Eds.), GABAergic Mechanism in the Mammalian Periphery, Raven, New York, 1986, pp. 19-34.