0306-4522/91 $3.00 + 0.00 Pergamon Press plc © 1991 IBRO
Neuroscience Vol. 42, No. 2, pp. 497-507, 1991
Printed in Great Britain
I N S I T U HYBRIDIZATION HISTOCHEMISTRY REVEALS
A DIVERSITY OF GABAA RECEPTOR SUBUNIT mRNAs IN NEURONS OF THE RAT SPINAL CORD A N D DORSAL ROOT GANGLIA E. PERSOnN, P. MALHERBEand J. G. RICHARDS* Pharma Research CNS, F. Hoffmann-La Roche Ltd, CH-4002 Basel, Switzerland Abstract--The distribution and relative abundance of gene transcripts for diverse GABAA receptor subunits (~q-3.~,fll-3, 72) in neurons of the rat cervical spinal cord and dorsal root ganglia were determined by in situ hybridization histochemistry using 35S-labeled 60mer oligonucleotide probes. The receptor proteins (mapped by benzodiazepine receptor radioautography and immunohistochemistry with [3H]flumazenil and a monoclonal antibody for the f12+ f13 subunits, respectively) were most abundant in the dorsal horn (layers II and III) and in layer X around the central canal. Although diverse receptor subunit mRNAs were detected (to varying degrees) in neurons throughout layers II-X of the spinal cord, motoneurons in layer IX were particularly strongly labeled. The )'2 mRNA was the most ubiquitous and abundant of the subunit variants investigated. The labeling of motoneurons in layer IX was particularly strong for ~2, moderate for f13 and 72 and extremely weak for ~tt and ~ta. In layers VII, VIII and X the f13 and 72 transcripts were moderately expressed whereas the ~q and f12 transcript levels differed markedly among the cells of these layers. Although the mRNAs of all subunit variants could be detected in layers IV-VI, only ~t3,~ts, f13 and 72 hybridization signals were observed in layers II and III. In the dorsal root ganglia, whereas ~t2 transcripts were abundant in virtually all large sensory neurons and to a much lower degree in the small diameter cells, 72 transcripts were confined to a subpopulation of large and small neurons. Furthermore, f12 and ~q transcripts were even more restricted in their distribution. The findings provided a basis for the mediation of synaptic inhibition in the spinal cord by diverse GABA^ receptors and further strong evidence for the long-established view that presynaptic inhibition of inter- and motoneurons, via axoaxonic synapses between GABAergic interneurons and primary afferent terminals, is mediated by GABAA receptors. The physiological roles and pharmacological implications of this receptor diversity have yet to be determined.
Inhibitory neurotransmission in the mammalian spinal cord is mediated primarily by the amino acids glycine and G A B A (Fig. 1). Synaptic receptors for the latter, G A B A ^ and G A B A a receptors (reviewed by Bowery, 4 Haefely, ~s Olsen and Tobin36), regulate Cl--conductances and K ÷- as well as Ca2+-conduc tances, respectively. Recent evidence from cloning studies 22,24,3°'3L44,45,ss'57,6°'7° has revealed a structural heterogeneity of GABAA receptors in the mammalian brain which could have diverse functional roles as well as pharmacological implications. Whether or not this receptor heterogeneity extends to other parts of the CNS, such as the spinal cord, is the subject of the present investigation. The anatomical distribution of G A B A and its r e c e p t o r s L2,2s'32'34,35,38'46-49'64,66,73,74 as
GABAergic neurotransmission 5'~°'2°'39 in the spinal cord have been extensively studied. Since G A B A is supposed to mediate the inhibitory control of primary afferents via axoaxonic synapses, 1'3'~2'~7'25'4°we investigated not only the spinal cord but also the dorsal root ganglia where the cell bodies of these sensory fibers are located. Here we describe the anatomical distribution and cellular localization of m R N A s specifically encoding receptor subunit variants using the technique of in situ hybridization histochemistry.
well as the electrophysiology and pharmacology o f
Male rats [specific pathogen-free (SPF) albino, FiiUinsdoff, Switzerland] weighing 120-130 g were decapitated, the spinal cords and dorsal root ganglia (C 3 and C 5 segment levels) were removed, immediately frozen on dry-ice and stored at -80°C until used; to facilitate sectioning, the ganglia were embedded in kidney tissue. Cryostat sections (12/zm) were mounted on slides, previously coated with 2% 3-aminopropyltriethoxysilane in acetone, then fixed in 4% paraformaldehyde (in phosphate-buffered saline, PBS, pH 7.4) for 40 min followed by three 5 min washing steps in PBS. Sections were then stored at -20°C until used. Gloves were worn when handling the microscope slides to protect against RNAase contamination of the tissue sections.
EXPERIMENTAL PROCEDURES
Tissue preparation
*To whom correspondence should be addressed. Abbreviations: C 3 and C5, cervical segments 3 and 5; DRG,
dorsal root ganglion; DTT, 1,4-dithio-t~L-threitol; EDTA, ethylenediaminetetra-acetic acid; GAD, glutamate decarboxylase; mRNA, messenger ribonucleic acid; PBS, phosphate-buffered saline; Poly A, polyadenylic acid; [35S]dATP, deoxyadenosine 5(thio)triphosphate 35S; SSC, standard saline citrate; tRNA, transfer ribonucleic acid. 497
498
E. PERSOHNet al.
Oligonucleotide labeling We used oligodeoxyribonucleotide probes, prepared on a DNA synthesizer by Med prob A.S., Oslo, with the following sequences:
ctt bases 1144-119729 ct2 bases 1137-1186 (Malherbe, in preparation) ct3 bases 1535-158530 ct5 bases 1486-153731 fit bases 1200-125029 f12 bases 1203-1261 TM f13 bases 1283-134570 72 bases 1039-10893~ 6 bases 1174-12336°. To minimize cross-hybridization, the least conserved sequences of the M3-M 4 cytoplasmic loop were chosen for all probes. The oligomers were labeled at the 3' end using terminal deoxynucleotidyl transferase (BRL) and [35S]dATP (New England Nuclear). The reaction mixture (30#1 total) contained 5:5ktl [35S]dATP, 10#1 tailing buffer (BRL: 500mM potassium cacodylate pH7.2, 10mM CoC12, 1.0 mM DTT), 1 #1 oligomer (100 ng/#l), 1 #1 bovine serum albumine (2mg/ml RNase/DNase free), 6.5 #1 terminal deoxynucleotidyl transferase (BRL, 15 U/lt 1) and 26 #1 H20. The cocktail was transferred to 37°C for 5min. The reaction was stopped by adding 5 #1 EDTA (0.5 M) and transferring the mixture to 75°C for 10 min. The labeled probe was separated from unincorporated nucleotides with a Biogel P30 spun-column (4 min at 1600 x g twice, Sorvall SW 24). In situ hybridization histoehemistry A modified version of the procedures of Young 72 and Lewis etlal. 26 was used. Sections of spinal cord and ganglia were brought to room temperature for 1 h before carrying out hybridization. For hybridization, sections were incubated with 50/11 of a solution with the following constitutents: 4 × standard saline citrate (SSC; 1 × SSC is 0.15 M NaCI, 0.015 M trisodium citrate, pH 7.0), 20% dextran sulfate, 0.25 mg/ml tRNA (Boehringer Mannheim), 0.25mg/ml Poly A (Sigma), 0.25mg/ml Herring sperm DNA (denat.), 50% deionized formamide (Bethesda Research Laboratories), 0.1 M DTT (Fluka), 0.5 × Denhard's solution and 35S-labeled probe (3 x 105 c.p.m.). Sections were covered with Fujifilm® coverslips and incubated in a moist chamber at 43°C for 21~24 h. After removal of the coverslips, the sections were washed twice in a solution containing 1 x SSC and 10 mM DTT for 15 min at 55°C, then in 0.5 x SSC with 10 mM DTT twice for 15 min at 55°C and afterwards in 0.5 x SSC with 10mM DTT once for 15 min at room temperature. After a dip in double distilled water, sections were dehydrated in ethanol, exposed (for up to I0 days) to sheet film (Hyperfilm, fl-Max ®, Amersham) and then dipped in a nuclear track emulsion [Ilford (Warnington, PA) K5, diluted I: 1 with distilled water] to reveal the regional and cellular localization of the probes, respectively. The film and emulsion were developed in Kodak PL12 and Kodak D19 respectively, then transferred to Kodak rapid fix. Nissl-stained sections were examined microscopically with bright-field and darkfield optics using a Zeiss Axiophot. In vitro binding and immunohistoehemistry The distribution of benzodiazepine receptors in sections of rat spinal cord and dorsal root ganglia was revealed by [3H]flumazenil binding in vitro 48 and by immunohistochemistry with bd-17, a monoclonal antibody specific for f12 and f13 subunits of the GABA A receptor complex) 4,49
RESULTS
Spinal cord
The distribution of GABAA receptors in sections of the rat spinal cord (C3 and C~ segment levels33), revealed by [3H]flumazenil binding and m A b bd-17 immunohistochemistry in vitro (Fig. 2a, b), was similar to that reported previously. 47~9,73 The receptor density in gray matter was highest in layers II and III of the dorsal horn and in layer X proximal to the central canal. Moderate to low densities were present in layers IV-VI and V I I - I X , respectively. In dorsal root ganglion ( D R G ) neither [3H]flumazenil binding nor G A B A A receptor immunoreactivity could be detected. In adjacent spinal cord sections labeled with 35Santisense oligonucleotide probes (Figs 2c-f, 3 a - f and 4), the distribution of the diverse receptor m R N A s (~1 3.5, fit 3, ),2) differed markedly. A discrete cellular localization was observed throughout layers I I - X but most markedly in the ventral horn. Control sections (labeled with 35S-antisense probes in the presence of a 50-fold excess of the respective non-radioactive probes, or with 3SS-sense probes) were homogeneously labeled to only background levels (results not shown). No cross-hybridization was detectable when using, for example 35S-antisense ~ probe in the presence of a 50-fold excess of non-radioactive ~2 probe. The ~2 m R N A was the most strongly expressed of the subunit variants investigated (Figs 2d and 4). The hybridization signal was particularly intense in layer IX of the ventral horn where large multipolar neurons (motoneurons) were labeled (Fig. 3e, f). Other layers did not express the ~2 transcript (Figs 2d and 4). A more moderate level of hybridization was observed with the ~3 probe in layers I I - V I I and X (Figs 2e and 4). Although the cellular labeling in layer II could be resolved from background levels, it was clearly much weaker than that found in other (more ventral) regions. Layers VII and VIII were the most prominently labeled regions expressing ~ m R N A , although the levels differed markedly among cells (Figs 2c, 3a, c and 4). Small to large neurons in layers VII and VIII contained a moderate to intense hybridization signal. Whereas lower transcript levels (in some multipolar neurons) were observed in layers I V - V I and X, little or no labeling was detected in layer IX. The ):2 probe labeled the majority of large multipolar neurons (motoneurons) in layer IX to a strong degree (Figs 2f, 3b, d and 4). Medium-sized (but not small) neurons in layers VII and VIII were moderately labeled. Medium-sized multipolar cells in layers IV-VI and cells in layers II and III contained a low but detectable signal. Transcripts for ~5 were similarly distributed but were less abundant (results not shown). A moderate hybridization signal was obtained with the f12 probe (results not shown). However,
GABAA receptor mRNAs in spinal cord and ganglia
499
_~- III \
LSN
Pyr VII \
IX
\
IX J
- -
VIII
\
Fig. 1. Neuronal connections in the cervical spinal cord (level C5): a schematic representation (modified from Molander et al?3) of its afferents, efferents and interneurons with particular emphasis on GABAerglc neurons and their local projections (in red). GABAerglc interneurons mediate pre- and postsynaptic inhibitory control of spinal inter- and motoneurons via primary sensory afferents, whose cell bodies are located in the dorsal root ganglia. I-X, Rexed layers; IM, intermediomedial nucleus; Liss, Lissauer's tract; LSN, lateral spinal nucleus; Pyr, pyramidal tract. Neurotransmitter (putative)
Symbol
Neuronal identity
- -