Function of γ-Aminobutyric Acid Receptor/Channel ρ1 Subunits In Spinal Cord*

γ-Aminobutyric acid (GABA) receptor/channel ρ1 subunits are important components in inhibitory pathways in the central nervous system. However, the precise locations and roles of these receptors in the central nervous system are unknown. We studied the expression localization of GABA receptor/channel ρ1 subunit in mouse spinal cord and dorsal root ganglia (DRG). The immunohistochemistry results indicated that GABA receptor/channel ρ1 subunits were expressed in mouse spinal cord superficial dorsal horn (lamina I and lamina II) and in DRG. To understand the functions of the GABA receptor/channel ρ1 subunit in these crucial sites of sensory transmission in vivo, we generated GABA receptor/channel ρ1 subunit mutant mice (rho1-/-). GABA receptor/channel ρ1 subunit expression in the rho1-/- mice was eliminated completely, whereas the gross neuroanatomical structures of the rho1-/- mice spinal cord and DRG were unchanged. Electrophysiological recording showed that GABA-mediated spinal cord response was altered in the rho1-/- mice. A decreased threshold for mechanical pain in the rho1-/- mice compared with control mice was observed with the von Frey filament test. These findings indicate that the GABA receptor/channel ρ1 subunit plays an important role in modulating spinal cord pain transmission functions in vivo.

␥-Aminobutyric acid (GABA) receptor/channel 1 subunits are important components in inhibitory pathways in the central nervous system. However, the precise locations and roles of these receptors in the central nervous system are unknown. We studied the expression localization of GABA receptor/channel 1 subunit in mouse spinal cord and dorsal root ganglia (DRG). The immunohistochemistry results indicated that GABA receptor/channel 1 subunits were expressed in mouse spinal cord superficial dorsal horn (lamina I and lamina II) and in DRG. To understand the functions of the GABA receptor/channel 1 subunit in these crucial sites of sensory transmission in vivo, we generated GABA receptor/ channel 1 subunit mutant mice (rho1 ؊/؊ ). GABA receptor/channel 1 subunit expression in the rho1 ؊/؊ mice was eliminated completely, whereas the gross neuroanatomical structures of the rho1 ؊/؊ mice spinal cord and DRG were unchanged. Electrophysiological recording showed that GABA-mediated spinal cord response was altered in the rho1 ؊/؊ mice. A decreased threshold for mechanical pain in the rho1 ؊/؊ mice compared with control mice was observed with the von Frey filament test. These findings indicate that the GABA receptor/ channel 1

subunit plays an important role in modulating spinal cord pain transmission functions in vivo.
␥-Aminobutyric acid (GABA) 1 is the major inhibitory neurotransmitter in the vertebrate central nervous system (CNS). GABA receptors have been classified into three distinct subtypes based on their pharmacological properties. GABA A receptor/channels are inhibited by bicuculline (1), GABA B receptors are sensitive to baclofen (an agonist) (2), whereas GABA C receptor/channels are insensitive to both bicuculline and baclofen (3,4). Both GABA A and GABA C receptors form ligand-gated chloride channels, whereas the GABA B receptor belongs to the G protein-coupled receptor family and is coupled to the K ϩ and Ca 2ϩ channels. GABA receptor/channel heterogeneity in the CNS is achieved by at least 19 different subunits, including ␣ 1-6 , ␤ 1-3 , ␥ 1-3 , ␦, , , ⑀, and 1-3 (5)(6)(7)(8)(9)(10). It is generally believed that the GABA C receptor is formed homooligomerically by the 1 subunit or heterooligomerically by both 1 and 2 subunits (9 -14). The 1 subunit is thought to form the GABA C receptor/channel because its pharmacological profile, physiological properties, and histological distribution match extensively with those of GABA C receptors (10,(15)(16)(17)(18). Recent evidence indicates that the 1 subunit is able to assemble with the GABA A receptor ␥ 2 subunit in the brain and spinal cord to form a novel hybrid GABA receptor/channel (19 -21). The hybrid receptor/channel very likely possesses pharmacological and electrophysiological properties that are distinct from those of typical GABA A and GABA C receptor/channels (19). Therefore, the GABA C receptor/channel formed by the 1 subunit is suggested to be a unique subclass of the GABA-gated ionotropic receptor/channel family. The 1 subunit is also considered to be an important component of the GABA receptor/channel heterogeneity with special functional properties.
GABA receptor/channel 1 subunit is highly enriched in the retina, where strong immunoreactivities of 1 and 2 subunits have been found in the inner plexiform layer, and weaker immunoreactivities have been found in the outer plexiform layer and cell bodies of bipolar cells (16,17,22). The 1 subunit is also expressed elsewhere, including the hippocampus (23), cerebellum (24), anterior pituitary (25), superior colliculus (24), and spinal cord (26). However, the precise localizations of the 1 subunit within these regions have not been defined.
The functions of the GABA receptor/channel 1 subunit are mainly studied in the retina, where it is thought that the GABA C receptor/channel formed by the 1 subunit is involved in mediating lateral inhibition (30,34) and local gain control circuitry (34,35). In other regions of the CNS, the functions of the GABA receptor/channel 1 subunit are largely unknown. Nevertheless, GABA has been long known for its important roles within the spinal cord, controlling sensory input via modulation of primary afferent transmitter release (36). Indeed, the activation of GABA A and GABA B receptors is antinociceptive in a variety of rodent models (36,37). The effects are thought to be mediated through the presynaptic inhibition of the primary afferent fibers by GABA A and GABA B receptors (36,38). However, the functional significance of the GABA receptor/channel 1 subunit in the spinal sensory pathway remains to be resolved.
To address this issue, we have conducted RT-PCR and immunohistochemical studies to define the expression localization of the GABA receptor/channel 1 subunit within mouse spinal cord and dorsal root ganglia (DRG). To assess the functional roles of 1 subunit in the sensory pathway, we have generated mutant mice carrying a deletion of the first exon of the 1 gene. RT-PCR analyses indicate that the 1 subunit gene transcription was completely abolished in tissues from the rho1 Ϫ/Ϫ mice. Immunohistochemical results demonstrate that subunits expression was eliminated completely in the rho1 Ϫ/Ϫ mice spinal dorsal horn and DRG. Electrophysiological studies indicate that GABA-inhibited spinal cord response was altered in rho1 Ϫ/Ϫ mice. Moreover, mechanical pain sensitivity in the rho1 Ϫ/Ϫ mice was significantly changed compared with control mice despite the apparently normal neuroanatomical structures of the mutant mice. These findings suggest that the GABA receptor/channel 1 subunit is essential in modulating spinal cord nociceptive functions in vivo.

MATERIALS AND METHODS
Immunohistochemistry and Nissl Stain-Adult mice (rho1 Ϫ/Ϫ and rho1 ϩ/ϩ ) were overdosed with 5% sodium pentobarbital and then transcardially perfused with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 15 min. Spinal cord lumbar segments (L4 -L6) and DRG were removed and placed overnight in 10, 20, and 30% sucrose PBS solution series for cryoprotection. The spinal segments and DRG were embedded in OCT® compound (Fisher), and 10 -20-m sections were cut on a cryostat (Zeiss). The sections were thaw-mounted onto slides, air-dried, and stored at Ϫ80°C. For immunohistochemistry experiments, the slides were warmed to room temperature and washed four times for 5 min in PBS. The sections were encircled with hydrophobic resin (PAP Pen) and incubated at room temperature for 1 h in a solution of 1:10 normal goat serum in PBS with 0.5% Triton X-100 and 1% bovine serum albumin to block nonspecific antibody binding. Then rabbit anti-antiserum in the blocking solution at dilution of 1:100 was applied to the sections at 4°C for overnight. After washing with PBS five times for 5 min, the sections were incubated with Alexa 488-goat anti-rabbit IgG (Molecular Probes) 1:250 in PBS at room temperature for 2 h. After washing with PBS five time for 5 min, the sections were air-dried and coverslipped with VectaShield (Vector). When fluorescent Nissl counter stain was desired, the sections were washed an additional three times for 10 min in PBS and for 10 min in PBS with 0.1% Triton X-100. NeuroTrace TM red fluorescence Nissl stain (Molecular Probes) 1:100 solution in PBS with 0.1% Triton X-100 was applied to the sections at room temperature for 20 min. Then the sections were washed for 10 min in PBS with 0.1% Triton X-100, followed by three washes for 10 min and one 2-h wash with PBS. Finally, the sections were air-dried and coverslipped with VectaShield (Vector). The slides were viewed under Leica microscope, and the images were captured by Leica TCS SP2 laser scanning confocal microscope with proper filter sets (Leica, Germany).
RT-PCR Analysis-Spinal cord lumbar segments, retina, and liver tissues from wild-type or rho1 Ϫ/Ϫ mice were homogenized in Ultraspec RNA reagent (Biotecx Laboratories, Houston, TX). RNA samples were then extracted and purified. After treatment with DNase I, 1 g of each RNA sample was reverse transcribed into cDNA using 1 M of oligo(dT) primer and 2 units of Omniscript reverse transcriptase (Qiagen). 0.25 or 0.5 g of cDNA was used in the subsequent PCRs with a pair of 1 gene-specific primers (forward, 5Ј-GAATCTATGTTGGCTGTCCAGA-3Ј; reverse, 5Ј-TGGTGTGGAATTCTTGAATGAG-3Ј). PCRs were run for 40 cycles with 94°C for 1 min, 58°C for 1 min, and 72°C for 1 min. 40 ng of the same cDNA samples was used in control PCRs with primers specific to the mouse GAPDH gene. All PCR products were visualized in 0.9% agarose gels after electrophoresis and ethidium bromide staining.
Mouse GABA C Receptor 1 Subunit Gene, Targeting Construct, and Embryonic Stem Cell Homologous Recombinants-Two oligonucleotide primers were used in PCRs using mouse 129 genomic DNA as template.
The sequences of these primers are 5Ј-GTTGGCTGTCCAGAATAT-GAAAT-3Ј and 5Ј-CTTTCCTAGATGGCTCATGAAC-3Ј, respectively (39). The resulting 120-bp PCR product containing part of the 1 gene first exon was used to clone mouse 1 gene and its flanking sequences from a 129 genomic library. To make the gene targeting construct, four pieces of DNA were ligated together. With plasmid pBluescript (Stratagene) as the backbone, a 3.8-kb BamHI-EcoRI fragment from the 5Ј end of the 1 gene, a 1.8-kb DNA containing a neo gene driven by the phosphoglycerate kinase promoter, and a 5.5-kb XbaI fragment from 3Ј of the 1 first exon were assembled in order. To obtain homologous recombinants, mouse J1 embryonic stem (ES) cells were transfected with 50 g of linearized targeting construct by electroporation using a Bio-Rad Gene Pulser at 800 V and 3 microfarads. G418 selection was applied 24 h after the transfection at 200 g/ml. G418-resistant colonies were picked 6 days after transfection (40). Genomic DNA from the ES cell colonies was analyzed by NcoI digestion and hybridization with a 3Ј 1 gene-specific probe. rho1 Ϫ/Ϫ Mice-Three identified homologous recombinant ES cell clones were amplified and subsequently injected into blastocysts isolated from C57BL/6 female mice. The injected blastocysts were implanted back into the uteri of B6 ϫ DBA2 F1 foster mothers (40). The resulting male chimeric mice were bred repeatedly with C57BL/6 females, and germ-line transmission was identified initially by screening for agouti offspring. Mice heterozygous for the 1 gene mutation were confirmed by genomic Southern analyses of DNA isolated from their tail biopsies. Finally, mice homozygous for the 1 gene mutation were produced by crossing heterozygous mutants. The genotypes of those animals were identified by Southern analyses with a 3Ј probe and were further confirmed by an exon 1 probe. Mutants and their wild-type and heterozygous control littermates 8 weeks of age were used in all subsequent analyses.
Recording of Spinal Cord Response-Spinal cord preparations were obtained from 8-week-old rho1 ϩ/ϩ and rho1 Ϫ/Ϫ mice. The animals were anesthetized by intraperitoneal administration of 10% urethane (0.3-0.5 ml). An incision was made on the back of the mouse, and the spinal cord lumbar segment (L2-L6) was dissected out. The isolated spinal cord was transferred in ice-cold Kreb's solution composed of 115 mM NaCl, 2.5 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 1 mM KH 2 PO 4 , 25 mM NaHCO 3 , 11 mM glucose, pH 7.4. The spinal cord preparation was placed in a perfusion voltage clamp chamber with the dorsal side up. The chamber was grounded through an agar bridge connected with an Ag/AgCl wire. Stimulation pulse (0.1 ms, 0.01 Hz, 5-10 V) was delivered by a stimulating electrode to the L3-L5 dorsal root. A glass recording electrode containing Kreb's solution was placed in the dorsal region of the spinal cord, and light suction was applied into the electrode. Spinal cord depolarization potential was recorded by a direct current-coupled amplifier (Axon labs, Foster City, CA). The data were filtered through a 4-pole filter at 1 kHz and stored in a digital recorder (32).
Mechanical Pain Sensitivity-A set of force-calibrated von Frey filaments (Stoelting Co., Wood Dale, IL) was used to assess mechanical pain sensitivity (41,42). The mice were placed on a wire mesh platform inside individual round plastic containers (7 cm in diameter) and allowed to acclimate for 30 min before the test. A series of von Frey filaments (0.09 -6 g) were applied through the wire mesh onto the hairless plantar surface of a hind paw. A response was indicated by the brisk withdrawal or flinching of the paw. In the absence of a response, the filament of the next greater force was used. In the presence of a response, the filament of the next lower force was used. Each monofilament was applied five times at 30-s intervals to each hind paw, and the response threshold was defined as the lowest force that caused at least three withdrawals of the five consecutive applications.
Data Analysis-The results are shown as the means Ϯ S.E. Student's t test and analysis of variance were performed to determine significant differences. Statistic significance was considered as p Ͻ 0.05.

Expression of GABA Receptor/Channel 1 Subunit in Spinal
Cord and DRG-Total RNA samples from mouse spinal cord, retina, and liver tissues were analyzed with RT-PCR using 1 -specific primers. The results clearly demonstrated the presence of 1 subunit expression in the spinal cord. The specificity of the PCR products was affirmed by the positive control obtained from the retina RNA sample and by the negative control obtained from the liver RNA sample. The integrity of RNA samples was assured by the intactness of control GAPDH gene products (Fig. 1A). Our result was consistent with previous findings of the existence of the 1 subunit in the spinal cord of mice (23,26). The precise localizations of the GABA receptor/channel 1 subunit were studied using immunohistochemistry. GABA receptor/channel subunit immunoreactivity revealed distinct punctate immunofluorescence in lamina I and II (LI and LII) of the superficial dorsal horn, as well as in ventral LIX motoneurons (Fig. 1B). Immunoreactivity of 1 subunit was most abundant in the superficial dorsal horn of the spinal cord, labeling LI or LII neurons, and the surrounding network of neuronal processes and synaptic terminals (Fig. 1, B and C). On LI and LII neurons, 1 subunit immunoreactivity displayed continuous membrane labeling (Fig. 1C).
Because abundant 1 subunit immunoreactivity was exhibited in LI and LII where the primary afferent fibers are known to terminate, cell bodies of primary sensory neurons in the DRG were also examined for 1 subunit immunoreactivity. Noticeably, a considerable amount of 1 subunit immunoreactivity was detected in DRG neurons (Fig. 1B).
Generation of rho1 Ϫ/Ϫ Mice-A restriction map of the first exon of the 1 gene and its flanking sequences is shown in Fig.  2A. Sequence analysis indicated that the first exon of the 1 gene includes the entire published 5Ј-untranslated region and 105 bp of coding region. The cloned first exon is 100% identical to the published mouse 1 sequence (39).
The gene targeting construct was designed to delete a 2.5-kb genomic DNA sequence including the first exon of the 1 gene and a portion of its promoter region ( Fig. 2A) that will disrupt the transcription of the gene. To assemble the targeting construct, a BamHI-EcoRI fragment from the 5Ј end of the 1 gene and a XbaI fragment from the 3Ј of the 1 gene first exon were ligated to the 5Ј and the 3Ј ends of a selectable marker ( Fig.  2A). When homologous recombination takes place, one allele of the 1 gene exon 1 is replaced by the neo gene, and this recombination creates a shift of the NcoI fragment from its native 7.5 kb to 6.5 kb when detected by a 3Ј end probe (Fig. 2, A and B).
After transfection with the gene targeting construct and G418 selection, we picked and screened the ES cell colonies for homologous recombinants by genomic Southern analysis using the 3Ј probe. Twelve ES cell clones harboring the desired homologous recombination were identified. ES cells amplified from three of the homologous recombinants were used to generate male chimeric mice, which were bred with C57BL/6 females to obtain heterozygous mutant mice (rho1 ϩ/Ϫ ). rho1 Ϫ/Ϫ mice were obtained by sibling mating of heterozygotes. The genotypes of the mice were identified by genomic Southern blotting with the 3Ј probe. rho1 Ϫ/Ϫ mice were identified with a single 6.5-kb band, a shift from the single 7.5-kb band from the rho1 ϩ/ϩ mice, whereas the rho1 ϩ/Ϫ mice showed double bands with both sizes (Fig. 2B).
The desired gene mutation was further confirmed by genomic Southern blotting with the exon 1 probe, showing the successful deletion of the 1 gene first exon from the genome of rho1 Ϫ/Ϫ mice (Fig. 2C). As expected, the deletion of the first exon and part of the promoter region from the mouse genome disrupted the transcription of the 1 gene completely, which was clearly demonstrated by RT-PCR assays. Total RNA from spinal cord lumbar segments and retina was subjected to analysis, and the integrity of the RNA samples was assured by the intactness of control GAPDH gene products (Fig. 3). Our result confirmed that the 1 subunit was absent from spinal cord in rho1 Ϫ/Ϫ mice (Fig. 3). The 1 subunit gene transcripts were also missing from the retina in rho1 Ϫ/Ϫ mice (Fig. 3). Thus, deletion of the 1 gene first exon and its promoter region disrupted the gene transcription, and consequently the expression of the 1 subunit was abolished in rho1 Ϫ/Ϫ mice. rho1 Ϫ/Ϫ mice appeared healthy and did not show obvious growth abnormalities compared with their wild-type or heterozygous littermates. Both male and female rho1 Ϫ/Ϫ mice were fertile, and the rho1 Ϫ/Ϫ females nursed their offspring until they were weaned and separated from the parents at 3 weeks after birth.
Alteration of GABA Effects on Spinal Cord Response in rho1 Ϫ/Ϫ Mice-To investigate the consequences of the 1 gene mutation, we used current clamp to compare the effect of GABA on spinal cord responses. Polysynaptic reflexes were elicited by delivering a stimulatory pulse to L3-L5 dorsal roots through a stimulating electrode in the absence and presence of GABA (500 M). Spinal cord responses were characterized by a depolarization potential (upward deflections) measured though a recording electrode located in the dorsal root region (Fig. 4). These depolarization potentials may be caused by direct responses of spinal cord neurons or via a local network of interneurons. Upon application of GABA in the spinal cord of rho1 ϩ/ϩ mice, the spinal cord response was significantly inhibited in wild-type mice (p Ͻ 0.05) but was less inhibited in rho1 Ϫ/Ϫ mice (Fig. 4A). This inhibitory effect of GABA on the depolarization potential was decreased by applications of 10 M TPMPA in the current clamp chamber (p Ͻ 0.05, comparing GABA with and without TPMPA; Fig. 4B). However, application of TPMPA alone did not affect the depolarization potential (p Ͼ 0.1, comparing control to TPMPA alone). In contrast, in rho1 Ϫ/Ϫ mice the effect of GABA on depolarization potentials was much less effective than in rho1 ϩ/ϩ mice (p Ͻ 0.05), and TPMPA had no effect on the GABA-mediated inhibition of depolarization potentials in the spinal cord (p Ͼ 0.1; Fig. 4B). Depolarization potential recorded from wild-type and rho1 Ϫ/Ϫ mice were traced and amplified to overlap each other in a proportionally enlarged scale to compare altered responses among these mice (Fig. 4C). Statistical analysis demonstrated that there was a significant effect of TPMPA on the GABA response in rho1 ϩ/ϩ mice (n ϭ 7, p Ͻ 0.05) but not in rho1 Ϫ/Ϫ mice (Fig. 4D). These results indicate that the effect of GABA on the spinal cord response is partially mediated by the 1 subunit and that the contribution of the 1 subunit in GABAmediated modulation of the spinal cord response was abolished in rho1 Ϫ/Ϫ mice.
Altered Mechanical Pain Sensitivity in rho1 Ϫ/Ϫ Mice-Because 1 is expressed in dorsal root ganglia and in the spinal superficial dorsal horn, which are crucial sites for pain sensory transmission, we examined rho1 Ϫ/Ϫ mice for their sensitivity to mechanical nociceptive stimuli. Interestingly, determination of mechanical pain threshold with the von Frey filament test indicated that the threshold for mechanical pain production in rho1 Ϫ/Ϫ mice had decreased to 0.6 g of force from the threshold of 2.5 g of force in rho1 ϩ/ϩ mice (Fig. 5). Therefore, significantly smaller forces are needed to generate pain responses from rho1 Ϫ/Ϫ mice than from rho1 ϩ/ϩ mice, suggesting an increase in mechanical pain sensitivity in rho1 Ϫ/Ϫ mice compared with control mice.
Absence of 1 Subunit in Superficial Dorsal Horn and DRG in rho1 Ϫ/Ϫ Mice-Our discovery of altered spinal cord response and decreased mechanical pain threshold in rho1Ϫ/Ϫ mice led us to believe that the 1 subunit is functionally involved in the modulation of the pain sensory pathway. To provide anatomical evidence for the altered sensory phenotype in rho1Ϫ/Ϫ mice, we examined the 1 subunit expression in DRG and spinal cord superficial dorsal horn in rho1Ϫ/Ϫ and rho1 ϩ/ϩ mice using immunohistochemistry. It was evident that the subunit immunoreactivity was completely absent in the spinal cord dorsal horn and DRG compared with the plentiful immunoreactivity in rho1 ϩ/ϩ mice (Fig. 6). The fluorescent Nissl counter stain assured that the images were taken at similar focal planes. It also demonstrated the grossly unchanged spinal cord and DRG neuroanatomical structures between rho1Ϫ/Ϫ and rho1 ϩ/ϩ mice (Fig. 6).

DISCUSSION
The biological functions of the relatively recently discovered GABA receptor/channel 1 subunit (33,43,44) are poorly understood. Because its expression is highest in the retina, most effort has been focused on examining the subunit homooligomer GABA C receptor activity and its precise localization in the visual system. For example, during the preparation of this manuscript, McCall et al. (45) published a study addressing the role of GABA C activity in retina. However, potential GABA receptor/channel 1 subunit functions outside of the visual system are virtually unexplored. In this investigation, we have discovered expression of the 1 subunit within the mouse spinal cord and DRG. Detailed immunohistochemical studies localized the 1 subunit in the spinal superficial dorsal horn. The expression of the 1 subunit in these sites implied the functional involvement of the 1 subunit in the nociceptive transmission pathway. To address this issue in vivo, we generated rho1 Ϫ/Ϫ mice by gene targeting. The genomic cloning and sequence analysis demonstrated that the first exon of the 1 gene included the entire 5Ј-untranslated region, the start codon, the leader sequences, and the first 10 amino acids of the tip of N terminus of the gene. Deletion of the first exon and 1 kb of the promoter region of the 1 gene most likely disrupted its transcription. Therefore, no truncated form of the 1 gene was expressed, avoiding the possibility that the truncated 1 subunit assembles with 2 subunits or other GABA A receptor subunits to form heterooligomeric receptors that restore the channel activity.
GABA Receptor/Channel 1 Subunit Expressed in Spinal Cord and DRG-The inhibitory response in spinal cord is mainly subserved by GABA receptors and glycine receptors. The cell type diversity and the functional diversity in spinal cord demand high heterogeneity in biophysical properties of these postsynaptic receptors. The diverse properties of the ionotropic GABA A receptors in spinal cord are achieved through their various subunit combinations (46), and GABA A receptor subunit expression patterns appear to be cell typespecific (47). In recent years, subunits, a new class of ionotropic GABA receptor subunit, have received attention because of their unique biophysical properties and their ability to form homooligomeric GABA receptor/channels. Because the new homooligomeric GABA receptor/channels do not respond to most of the GABA A and GABA B receptor antagonists, agonists, and modulators, they were given the name GABA C receptors. How-FIG. 2. Generation of GABA receptor/channel 1 subunit mutant mice. A, the genomic DNA locus surrounding the first exon (black box) of the 1 gene, the targeting vector, the mutant 1 gene locus, and the 3Ј hybridization probe. The restriction enzyme sites used in the gene targeting construct are shown. The exon 1 probe contains 120 bp from the first exon. The NcoI fragment size shift from 6.5 to 7.5 kb in the mutant 1 gene is indicated schematically. B, identification of rho1 Ϫ/Ϫ mice. Tail DNA samples from one litter of heterozygous intercross were digested with NcoI and hybridized with the 3Ј probe. C, genomic Southern analysis with the exon 1 probe. The rho1 ϩ/ϩ , rho1 ϩ/Ϫ , and rho1 Ϫ/Ϫ mouse genomic DNA samples were digested with either EcoRI or BamHI and were hybridized with the exon 1 probe. ever, recent evidence indicates that the 1 subunit is able to assemble with the GABA A receptor 2 subunit in the brain and spinal cord to form a novel hybrid GABA receptor/channel (19 -21). The pharmacological and electrophysiological properties of the hybrid receptor/channel are very likely distinct from those of the GABA A and GABA C receptor/channels (19), further increasing the functional diversity of the ionotropic GABA receptor/channel in CNS (48). These discoveries suggest that 1 subunits may be a unique component of GABA receptor/channel heterogeneity. Hence, we decide to refer to it as GABA receptor/channel 1 subunit throughout this manuscript.
In this study, we have confirmed previous findings that the GABA receptor/channel 1 subunit is present in mouse spinal cord (Fig. 1A) (26). Furthermore, we have precisely located 1 subunit in the LI and LII of the spinal superficial dorsal horn and in DRG (Fig. 1B). In LI and LII, 1 subunit immunoreactivity was richly present on LI and LII neurons as well as the surrounding neuronal processes and fibers (Fig. 1, B and C). It is well known that central nociceptive neurons are located in dorsal LI (also called the marginal layer) and LII (the substantia gelatinosa). The majority of these neurons receive direct synaptic input from nociceptive A␦ and C fibers (49,50). Many of the neurons in lamina I respond exclusively to noxious stimulation and project to higher brain centers (51), whereas many of the neurons in lamina II are interneurons, responding only to nociceptive inputs (49,50). Nociceptive afferent fibers also terminate predominantly in LI and LII. Therefore, it is likely that the 1 subunit immunoreactivity LI and LII is present on central nociceptive neurons in lamina I, on nocicepion-modulating interneurons in lamina II, or on primary nociceptive afferent fibers. Moreover, abundant 1 subunit immunoreactivity was detected in DRG (Fig. 1B), where the cell bodies of all of the primary nociceptive neurons are located. All of this evidence suggested that GABA receptor/channel 1 subunit was actively involved in the process of pain sensory transmis-sion. To study this issue in vivo, we have generated GABA receptor/channel 1 subunit mutant mice rho1 Ϫ/Ϫ .
Alterations of GABA-mediated Spinal Cord Responses-GABA is well known for its important roles in mammalian spinal cord function. It is thought that the effects of GABA are mediated through GABA A and GABA B receptors (36). Recently, a newly developed GABA receptor/channel 1 subunit-specific antagonist TPMPA was employed in electrophysiological studies, indicating that the 1 subunit plays a role in modulating GABA-mediated spinal cord response (26,31,32). In the present study, application of TPMPA partially abolished the GABAinduced inhibition of spinal cord depolarization potential in rho1 ϩ/ϩ mice. TPMPA-sensitive depolarization potential contributes to about 20% of total depolarization potential in the spinal cord (Fig. 4C). However, TPMPA had no effect on the spinal cord response in rho1 Ϫ/Ϫ mice, indicating that the effect of TPMPA on the spinal cord response is dependent on the GABA receptor/channel 1 subunit. We have also observed a significant difference in GABA-mediated inhibition of the spinal cord response in rho1 ϩ/ϩ and rho1 Ϫ/Ϫ mice. In spinal cord preparations obtained from rho1 ϩ/ϩ mice, application of GABA inhibited nearly 60% of stimulation-evoked spinal cord response, whereas it only inhibited 30% of the response in rho1 Ϫ/Ϫ mice (Fig. 4C). The spinal cord responses recorded from our wild-type and mutant mice are consistent with other electrophysiological studies of GABA 1 subunit (32). The inhibitory effect of TPMPA on GABA-sensitive spinal cord response is not a hugely or dramatically change as we have seen in the knock-out rho1 Ϫ/Ϫ mice. The GABA 1 subunit is of only one of several GABA receptor subunits, and it is just such more subtle effects that need to be understood by modern neurobiology. The smaller sensitivity to GABA inhibition in rho1 Ϫ/Ϫ mice suggests that any compensatory increase in GABA A expression resulting from a loss of 1 subunit is unlikely. Recently, studies in our lab and from other labs indicated that the 1 subunit could interact with 2 subunits of GABA A receptor/ channel in brains and spinal cord (19). In addition, the 1 subunit can also form functional heteromeric receptor/channel with 2 subunits of the GABA A receptor/channel when heterologously expressed in Xenopus oocytes (19 -21). These findings may explain why the inhibitory effect of GABA was less in rho1 Ϫ/Ϫ mice than in rho1 ϩ/ϩ mice. Altered Mechanical Pain Threshold-Our findings indicate that 1 is expressed in dorsal root ganglia and spinal cord superficial dorsal horn, two essential sites for nociceptive input. Our results of altered mechanical pain sensitivity in rho1 Ϫ/Ϫ mice (Fig. 5) provide for the first time strong evidence that indeed 1 subunit modulates nociception. Mechanical pain threshold was decreased (therefore sensitivity increased) in rho1 Ϫ/Ϫ mice (Fig. 5). This is consistent with the well documented antinociceptive role that GABA A and GABA B receptors play (36,37). For example, mechanical pain is increased by GABA A and GABA B antagonists (52) and inhibited by GABA A and GABA B agonists (53).
GABA receptor/channel 1 subunits possesses some very unique physiological characteristics compared with their GABA A receptor counterparts, including their much higher sensitivity to GABA, their longer mean open time, and their lack of significant desensitization even at very high agonists concentrations (4,(27)(28)(29)(30). All of these properties make 1 subunit ideal for maintaining the normal inhibitory tone and pain threshold in spinal cord. Therefore, we report that when GABA receptor/channel 1 subunits are eliminated in rho1 Ϫ/Ϫ mice (Fig. 6), the inhibitory tone in the sensory pathway is attenuated (Fig. 4). This can cause long term increased sensitivity and overexcitation of the primary nociceptive neurons. The atten- uated inhibition in the superficial dorsal horn in conjunction with persistent excessive firing of the primary nociceptive neurons can lead to progressively increased responses from the superficial dorsal horn neurons and eventually cause long term hyperexcitability of superficial dorsal horn neurons. All of these alterations in the biochemical properties and excitability of primary afferent neurons and dorsal horn neurons have been linked to the decrease of the threshold for the production of pain and mechanical allodynia (54,55), which is observed in rho1 Ϫ/Ϫ mice (Fig. 5).
The Nissl stain of the spinal cord and DRG have indicated grossly normal neuroanatomical structures of the spinal cord Depolarization potential traces from wild-type and rho1 Ϫ/Ϫ mice were proportionally enlarged and overlapped to demonstrate the altered response in these mice (C). Statistical analysis of GABA and TPMPA effects on spinal cord depolarization potentials in rho1 ϩ/ϩ and rho1 Ϫ/Ϫ mice is presented in D. The data were collected from 7 rho1 ϩ/ϩ and 5 rho1 Ϫ/Ϫ mice, respectively. Asterisks indicate significant differences (p Ͻ 0.05). and DRG in rho1 Ϫ/Ϫ mice, despite the complete depletion of immunoreactivity in these sites (Fig. 6). It seems that the altered pain threshold in rho1 Ϫ/Ϫ mice can be attributed specifically to the absence of GABA receptor/channel 1 subunit functions. Interestingly, the antibody we used in the immunohistochemistry study recognizes both 1 and 2 subunit (56). The complete absence of subunit immunoreactivity indicates that the elimination of 1 subunit by gene targeting results in the absence of 2 subunit expression. This finding is consistent with the similar discovery of McCall et al. (45) in retina, although a completely different 1 gene targeting strategy was employed in their case. It is possible that the 2 subunit has to assemble with 1 subunit to form a functional channel, and 1 subunit possesses the unique assembly motif. As a result, when the 1 subunit is absent, the 2 subunit cannot form functional channels and eventually is degraded. It is also possible that genomic structural changes introduced by the gene targeting interfere with the 2 subunit expression, because both 1 and 2 genes are linked on chromosome 4, 40 kbp apart (15,57). Similar changes in neighboring gene expression have been observed for GABA A receptor/channel ␣ 1 and ␤ 2 subunit after the knock-out of ␣ 6 subunit (58,59). The exact nature of the absence of 2 subunit expression in dorsal spinal cord and DRG remains to be determined.
In summary, we have precisely located the GABA receptor/ channel 1 subunit expression in mouse spinal superficial dorsal horn and DRG. We have also generated rho1 Ϫ/Ϫ mice to study the function of the GABA receptor/channel 1 subunit in pain transmission in vivo. Altered pain threshold for mechanical stimulation in rho1 Ϫ/Ϫ mice indicates the functional roles of 1 subunit in modulating nociceptive input and transmission. These findings represent the first in vivo evidence of modulatory roles of GABA receptor/channel 1 subunit in pain sensory pathways in the spinal cord. Future studies with mice with a congenic background will allow us to define more precisely these findings. The rho1 Ϫ/Ϫ mice will also provide a FIG. 5. Altered pain threshold for mechanical stimulation in rho1 ؊/؊ mice. Mechanical allodynia was assessed with von Frey filaments. The rho1 Ϫ/Ϫ mice responded to forces significantly lower than those for the rho1 ϩ/ϩ mice, indicating an increase in mechanical sensitivity. The data were collected from at least eight individual mice for each genotype. The asterisk indicates statistical significance (Student's t test, p Ͻ 0.05) compared with wild-type mice. valuable tool to study other possible functions of GABA receptor in the CNS and GABA receptor subunit physiology.