Abstract

Cortical representations of different modalities can be modified by sensory learning. Our previous studies in the barrel cortex showed that expansion of the cortical representation of a row of vibrissae could be induced by pairing stimulation of a row of vibrissae with a tail shock. The plastic change in cortical reactivity to the input used during the training was accompanied by increased density of GABA immunoreactive neurons in the involved row of cortical barrels. Using the same paradigm, the present study examined the pathway of GABA synthesis-expression of GAD67 mRNA and immunoreactivity of GAD67 isoenzyme in the barrel cortex of mice after sensory learning. In situ hybridization revealed that the GAD67 mRNA level was elevated in one row of barrels in the trained group as well as in controls receiving vibrissae stimulation alone. In contrast, elevation of immunoreactivity of the GAD67 protein occurred only in the trained group. The density of GABA-immunoreactive neurons in the hollows of barrels representing the row of vibrissae activated during the training was increased by 50%. These data indicated that sensory stimulation alone affected expression of the 67 kDa glutamate decarboxylase isoenzyme synthesis pathway, whereas the processes involved in cortical plasticity induced by associative learning modified this pathway additionally at the level of translation.

Introduction

γ-Aminobutyric acid (GABA) is the primary fast inhibitory neurotransmitter in the mammalian brain. GABA-mediated transmission processes are considered to be controlled by afferent activity (Jones, 1993). Alterations in GABAergic transmission in the brain cortex change neuronal receptive field properties and affect cortical excitability and plasticity (Sillito, 1984; Jacobs and Donohue, 1991; Hensch et al., 1998). Regulation of GABA synthesis is based upon kinetic features of glutamate decarboxylase (GAD), which has two isoforms (GAD65 and GAD67) (Kaufman et al., 1991), products of two separate, differently regulated genes (Erlander et al., 1991; Bu et al., 1992). In the central nervous system (CNS) the GAD isoforms are represented in various ratios, depending on the structure studied (Erlander and Tobin, 1991; Esclapez et al., 1993).

The present study is designed to examine the alterations in GABA synthesis pathway in the barrel cortex that undergoes a plastic change as a result of sensory learning. We have found previously that pairing stimulation of a row of vibrissae with a tail shock in a classical conditioning paradigm results in an expansion of cortical representation of vibrissae activated during the training (Siucinska and Kossut, 1996). We chose the 67 kDa isoenzyme of GAD as a marker of the functional state of the GABAergic system in the cortical representation of the stimulated vibrissae. GAD67 is involved in synthesis of most of the cellular GABA (Asada et al., 1997). GAD67-synthesized GABA apparently takes part in synaptic transmission, because GAD65–/– knockout mice do not show any major deficits in behavioral tests (Asada et al., 1996) and most receptive fields properties in the visual cortex are normal (Hensch et al., 1998). Moreover, both forms coexist in axon terminals (Feldblum et al., 1993; Sloviter et al., 1996). Because of different suggested mechanisms of activity regulation of both isoforms it was more probable to find changes in GAD67 using methods of cellular resolution such as in situ hybridization (ISH) and immunohistochemistry (Martin, 1987; Erlander and Tobin, 1991).

While numerous studies were devoted to changes in the GABAergic system after the elimination of sensory input, relatively little is known of the effects of increasing sensory input upon alterations of cortical inhibitory system components. Experiments by Hendry (Hendry and Jones, 1986) have demonstrated that opening the eye after unilateral ocular deprivation caused the reversal of reduction of the number of GABA-immunoreactive neurons in monkey visual cortex. The inhibitory system in somatosensory cortex also can be affected by alterations in peripheral stimulation. Experiments by Welker have demonstrated that GAD immunoreactivity (IR) was enhanced in the vibrissal barrel cortex after 4 days of constant whisker stimulation in mice (Welker et al., 1989a). We approached the problem of regulation of GAD67 expression by sensory learning also in the barrel system, which is well suited for studies of cellular correlates of sensory plasticity because cortical representation (barrels) of peripheral receptors (vibrissae) can be easily identified in a simple histological preparation (Woolsey and Van der Loos, 1970). Our recent data indicate that when sensory activation of vibrissae is paired with an aversive stimulus, the response of the cortical GABA system is quite rapid (Siucinska et al., 1999). Three 10-min long sessions of classical conditioning training, which produce an expansion of functional representation of the stimulated row of vibrissae, as seen with 2-deoxyglucose mapping (Siucinska and Kossut, 1996) resulted, 24 h after the training, in a significant increase in the density of GABA-immunoreactive neurons in the cortical barrels representing the stimulated row of vibrissae (Siucinska et al., 1999).

In this study we examined, using ISH, changes in GAD67 mRNA expression in the cortical representation of vibrissae that undergo the plastic changes as a result of sensory conditioning. We also examined with GAD67 immunohistochemistry whether the previously observed increase in GABA IR is due to a change in the level of the GAD67 protein, in order to establish at which level the learning-dependent regulation of GABA synthesis might occur – at the level of GAD67 mRNA or the level of protein synthesis.

Materials and Methods

Animals

Swiss–Webster adult mice of both sexes (25–30 g) were used in all experiments. Food and water were available ad libitum. The animal care, surgery and all experimental procedures on live animals were approved by Animal Care and Use Committees of the Polish Academy of Sciences, performed in accordance with The European Communities Council Directive (86/609/EEC) and conformed to the guidelines of the National Institutes of Health Guide to the Care and Use of Laboratory Animals.

Three groups of mice were used. One group received classical conditioning training, where stimulation of vibrissae was paired with a tail shock (CS + UCS). The second group (the stimulated control) received only stimulation of vibrissae (CS). The third was a naive control group, which received no treatment. The classical conditioning training comprising conditioned (CS) and unconditioned stimulus (UCS) was previously described by Siucinska (Siucinska and Kossut, 1996). Briefly, mice were accustomed to a neck restraint by being placed in a restraining apparatus for 10 min a day for 3 weeks before behavioral conditioning. Each session consisted of 40 strokes (each of 3 s duration) applied to row B of whiskers unilaterally, without touching neighboring rows of whiskers. During the conditioned training a mild electric shock (0.5 s, 0.5 mA) was applied to the mouse tail concurrently with the third stroke (i.e. every 15 s). The CS only group had an identical schedule of vibrissae stimulation but did not receive the tail shock. In both procedures the training session lasted for 1 or 3 successive days, 10 min daily.

Tissue Preparation for ISH

Four or 24 h after a single stimulation session and 1 h, 24 h or 5 days after the third stimulation session the mice were decapitated and brains rapidly removed and frozen in isopentane (–70°C). Frozen hemispheres were cut separately on the cryostat in an oblique coronal plane (55° from the sagittal plane) throughout the barrel field. The oblique plane allowed us to cut a barrel field across all the barrel rows in parallel to the barrel arcs so that barrels belonging to individual rows could be identified. The 10 μm thick sections were mounted (four sections/slide) on silanized slides (APTS, Aldrich, Milwaukee, WI). Air-dried sections were fixed (4% paraformaldehyde in phosphate-buffered saline, PBS, for 5 min; PBS for 5 min) at room temperature and dehydrated in graded ethanol. Every sixth section was kept unfixed and stained for succinyl dehydrogenase (SDH) or Nissl substance. For each hemisphere two slides were used for ISH and one slide for competition control.

Oligoprobe

The antisense oligonucleotide for GAD67 [5′-CGAAGGAGTGGAA- GATGCCATCAGCTCGGTGGTCTT-3′] was custom synthesized (Isogen, The Netherlands). The oligonucleotide was labeled according to Wisden (Wisden and Morris, 1994) with small modifications. Briefly, tailing was performed with a TdT kit (Promega, Madison, WI). A 21 μl volume of labeling mix contained 2 μl of oligo water solution (0.3 pmol/μl), 4 μl of 5× reaction buffer, 12.5 μl of water, 1.5 μl of [γ-35S]dATP (NEN, Boston, MA) and 1 μl of TdT. The reaction was stopped by addition of TNES buffer up to 50 μl, and the probe was purified on Biospin-6 spin columns (BioRad, Hercules, CA). A sample (2 μl) of eluate was moved to scintillation vials to measure incorporated radioactivity. The mix was incubated for 6–20 min to obtain labeling in the range of 1.0–3.0 × 105 dpm/μl. Eluate was then mixed with 2 μl of 1 M dithiothreitol, and stored at –20°C until hybridization.

ISH

In situ hybridization histochemistry was performed according to Wisden (Wisden and Morris, 1994) using synthetic oligodeoxynucleotide probe complementary to the murine complementary DNA sequence. The slide-mounted sections were placed in 4% paraformaldehyde in PBS (5 min, room temperature), 1× PBS (5 min). The sections were then dehydrated with increasing concentrations of ethanol and air-dried. The hybridization mix was made from 100 μl of hybridization buffer (50% formamide, 4× saline–sodium citrate buffer, SSC, 10% dextran sulfate), 1 μl of labeled probe and 1 μl 1 M dithiothreitol. A 100 μl volume of this mix was dropped on each slide and the slides were covered with parafilm coverslips. The slides were incubated in a humid chamber (50% formamide) at 42°C overnight.

After incubation the coverslips were gently washed off and the slides washed in 1× SSC (20 min, 60°C), 1× SSC (3 min, room temperature) and 0.1× SSC (3 min, room temperature). Finally the sections were dehydrated in graded ethanol and air-dried. The competition controls of specificity were performed concurrently by addition of 100× excess unlabelled probe.

Slides were then apposed to mammography X-ray medical film (MH-5, Kodak, Germany), exposed for 4 weeks, developed (D19, Kodak) and fixed (X-ray Fixer, Kodak). In each cassette the slides were exposed along with 14C-microscales (ARC-146A, American Radiolabeled Chemicals Inc., St Louis, MO) in order to calibrate the film.

Analysis of Autoradiograms

Developed and fixed autoradiograms were analyzed with a computer- assisted image analyzer (MCID, Imaging Research, Ontario, Canada). The energy range of  β particles emitted by 35S and 14C is fairly close, which allowed us to calculate the local concentration of the radiolabeled probe from a standard curve created using microscale references as described in detail by Wisden (Wisden and Morris, 1994). In each section, readings of optical density were taken in individual layers (II/III, IV, V, VI) within cortical columns related to each row of barrels (A–E) and respective measurements from four to eight sections were averaged within hemispheres (Fig. 3B). On each section readings of labeling were also taken in each layer of the cingulate cortex, which we chose as a reference structure. Because brains originating from individual groups of subjects were examined in several ISH experiments, using different batches of isotope, significant differences in total radioactivity per section occurred, affecting respective group means and standard deviations. To compensate for this effect, measurements from each film were multiplied by the corresponding correction factor, resulting in equality of mean values of radioactivity across groups for the entire barrel cortex of the control side. The correction factor was calculated by dividing the mean radioactivity in the barrel cortex of the control side by the average radioactivity in the corresponding region from all animals. This procedure was justified by the linear relation between the specific activity of the probe and the radioactivity readings obtained from the calibrated X-ray film. All readings were blind. Then a one-way analysis of variance (ANOVA) and a post hoc Scheffé test were used to determine the difference between row B and each of A, C, D and E rows in every cortical layer in both hemispheres in each group of animals. The results were also expressed without normalization; for each section a ratio of labeling of the examined segment of the barrel cortex to the reference structure was calculated. The results obtained in this way and examined with nonparametric statistics did not differ from those obtained using normalization of the data.

Tissue Preparation for Immunohistochemistry

GAD67 immunohistochemistry was performed only at one time point after training. Approximately 24 h after the end of the conditioning training animals were anaesthetized with an overdose of Nembutal (intraperitoneally) and immediately perfused via the ascending aorta with 0.9% saline, followed by 4% paraformaldehyde in 0.1 M PBS (pH 7.4). The brains were removed immediately, immersed in fresh fixative and stored in a refrigerator and cryoprotected in 30% sucrose. Left and right hemispheres were dissected. Each of them was separately flattened and cut tangentially to the barrel field into sections (30 μm thick) on the cryostat and then collected in PBS. Sections from the barrel field were first observed via transillumination in a microscope, and then processed for GAD67 immunohistochemistry.

Immunohistochemistry

The commercially available polyclonal antibody (K2, Chemicon, Temecula, CA) was used for localization of the 67 kDa isoform of glutamate decarboxylase (GAD67). To block non-specific binding the free-floating sections were kept in PBS buffer (pH 7.4) with 10% normal goat serum (NGS) for 1 h at room temperature. After 48 h of incubation with the primary antibody to GAD67 (1:2500), sections were incubated in biotinylated anti-rabbit IgG at a dilution of 1:100 in PBS (pH 7.4) for ~1 h at room temperature, rinsed in PBS (pH 7.4) and then incubated in avidin–biotin complex reagent (ABC Kit; Vector Laboratories, Burlingame, CA) and DAB as described previously (Kiser et al., 1998). Sections were then viewed and photographed prior to dehydrating in ethanol and embedding in DPX medium. Sections from the two hemispheres of one brain were incubated in one vial. Mice from all groups were processed together.

Methods of Quantitative Analysis

The average number of GAD67 immunoreactive neurons within row B hollows (hollow of barrel B1, B2, B3, B4) in layer IV was determined. To define the morphological shape of each barrel, GAD67- and Nissl-stained sections were overlapped using the blood capillaries as reference marks. Sections were digitized directly from the microscope slides on a Nikon microscope (Nikon, Germany) interfaced with a CCD color video camera and a computerized digitizing morphometry and counting system MCID. Barrels were defined according to the criteria proposed by Woolsey (Woolsey and Van der Loos, 1970). The hollow of each barrel was observed first at low magnification, then at higher magnification (20×). Finally, images of the immunohistochemically processed material were displayed at a magnification of 40× to determine the number of GAD67- immunoreactive neurons belonging to each row B hollow area. We identified at least three consecutive sections from layer IV containing barrels B1, B2, B3, B4. The immunoreactive neurons were counted in the barrel hollows in all sections with clearly visible staining. The areas of the hollows of row B barrels were measured on each section. Results were expressed as number of immunoreactive neurons per area unit. Full stereological analysis was not undertaken. Whereas the employed method is not strictly unbiased, it has been shown to produce results similar to those obtained with the disector method when the analysis concerns the determination of relative changes of neuronal density rather than estimations of absolute number of neurons.

ANOVA followed by post hoc Tukey–Kramer Multiple Comparisons test was used to assess statistical significance.

Results

GAD67 mRNA Level — ISH study

Control Animals

In autoradiograms obtained from oblique coronal sections a heterogeneous expression of GAD67 mRNA was seen throughout the cortical depth (Fig. 1). The maximum signal was confined to cortical layer IV and lower layer III. In each analyzed layer, there were no differences in density of labeling among the barrel columns belonging to different rows (A–E) of vibrissae or between hemispheres. Control hybridizations with an excess of unlabelled oligonucleotides were performed (Fig. 2). The hybridization signal on a section with unlabeled homologous oligonucleotide is at the detection threshold and is homogenous throughout an autoradiogram. Addition of an unrelated, heterologous oligonucleotide did not change the hybridization signal of the GAD67 oligoprobe. These results confirmed the specificity of GAD67 mRNA hybridization signal.

CS + UCS and CS Only Groups of Animals

Three daily sessions of classical conditioning training (CS + UCS) caused a significant increase of hybridization signal for GAD67 mRNA observed in the cortical representation of the ‘trained’ row B of vibrissae as compared with the other hemisphere (Fig. 3A,C,D). Changes in GAD67 mRNA level were limited to layer IV. One hour after the last training session, a 72% increase in labeling of stimulated row B with respect to the representation of non-stimulated rows was observed (Figs 4 and 5). This increase was significant with respect to both non-stimulated rows and the values obtained in row B representation of naive controls (P < 0.01, Scheffé, n = 5). Twenty-four hours later the difference, although smaller (54%), was still significant with respect to non-stimulated rows and naive controls (P < 0.01, Scheffé, n = 5). Either 1 or 24 h after the last session, no changes of labeling intensity were observed in the barrel field ipsilateral to the trained row of vibrissae.

In the CS only group, tactile stimulation without application of unconditioned stimulus affected GAD67 mRNA levels in representation of trained row of vibrissae in a similar way as when it was paired with the tail shock. One hour and 24 h after the last stimulation session an increased labeling in the cortical representation of stimulated row B of vibrissae was observed (78 and 56%, respectively). The increase was significant (P < 0.01, Scheffé, n = 5) with respect to the representation of non-stimulated rows and values obtained in naive controls. Again, in each representation of rows of individual barrels in the control side no changes in GAD67 hybridization signal was observed either 1 or 24 h after the last session.

Time Course of Changes in Intensity of GAD67 Hybridization Signal

To study how quickly the changes in localization of GAD67 mRNA can occur in CS + UCS and CS groups of mice, ISH was performed 4 and 24 h after only one session (Fig. 5). Rapid elevation of GAD67 hybridization signal was observed 4 h after one session in both groups of animals. In the CS + UCS and the CS only groups there were increases of 22 and 32%, respectively, in the cortical representation of trained row B as compared with the non-stimulated rows and to the same cortical area in the control group of animals (P < 0.05, Scheffé, n = 5).

The effect, however, was short-lasting, and 24 h after the last session no difference was observed. Interestingly, in the CS only group an ipsilateral effect was observed – an increase of 22% was noted in ipsilateral row B barrels as compared with the naive controls (P < 0.05, Scheffé, n = 5).

We also examined the existence of delayed effects of vibrissae stimulation. Five days after training lasting 3 days no significant changes in GAD67 mRNA level in the cortical representation of stimulated row B were observed (only the CS + UCS group was examined).

GAD67 Immunohistochemistry Study

Immunoreactivity of the GAD67 protein was examined at one time-point only, i.e. 24 h after three sessions of stimulation in animals receiving CS + UCS and CS alone. In both experimental groups and in naive controls GAD67-immunoreactive cells were found in each layer of the somatosensory cortex. All neurons immunopositive for GAD67 were of non-pyramidal shape (Fig. 6). The nucleus of GAD67-immunoreactive neurons appeared more lightly stained than the surrounding cytoplasm, which was often very strongly stained. In many GAD67-immunoreactive cells, the proximal dendritic arborization and axons were stained. At high magnification GAD67-immunoreactive puncta were clearly visible surrounding the unlabeled cell bodies. The density of GAD67-immunoreactive puncta was not analyzed in the present study. In the barrel cortex GAD67-immunoreactive neurons were not homogeneously distributed throughout the walls and hollows of each barrel. As described previously (Lin et al., 1985), GAD-immunoreactive neurons were more densely concentrated in the barrel walls then in barrel hollows.

The GAD67-immunoreactive neurons were counted only in the hollows of row B barrels, because in the previous study (Siucinska et al., 1999) it was the site where changes in density of GABA-immunoreactive neurons were observed. In barrels belonging to the row of vibrissae stimulated during the training, the average density of GAD67-immunoreactive neurons in the hollows increased by 53% (P < 0.05, Tukey, n = 5) as compared with the hollow of row B on the contralateral side (Fig. 7). The average density of GAD67-immunoreactive neurons for a series of individual hollows in stimulated row B of vibrissae was 49 ± 14 per 105 μm2 and at the control side was 32 ± 8 per 105 μm2. In a group of animals that received only vibrissae stimulation no significant difference in average density of GAD67-immunoreactive neurons in hollows of row B was found between the barrel fields of both hemispheres and those obtained in the naive control group.

Discussion

The two main results of this study are: (i) changes in regulation of the inhibitory neurotransmitter synthesis pathways that occur after sensory stimulation are at the level of transcription and (ii) associative learning specifically modifies the inhibitory neurotransmitter pathway at the post-transcriptional level. Even a short session of tactile stimulation of vibrissae leads to an increase, 4 h later, of the GAD67 mRNA level in the cortical representation of stimulated whiskers. Application of an aversive stimulus at the time of tactile stimulation (classical conditioning) causes a similar effect. However, an increase of the GAD67 protein level, as judged by IR, could be observed only after the classical conditioning paradigm (resulting in associative learning) was employed.

Numerous studies have shown a GAD response to change in activation of the CNS. Increases in GAD67 mRNA expression were observed after pharmacological intervention; for example, in the striatum after chronic application of dopamine antagonists (Laprade and Soghomonian, 1995), in the hippocampus after pilocarpine-induced seizures (Esclapez and Houser, 1999), as well as after the augmented stimulation of a structure; for example, in the motor cortex after short-train tetanic intracortical stimulation (Liang et al., 1996) and in Purkinje cells by increased input (Drengler and Oltmans, 1993). The modification of visual input can influence the expression of GAD67 IR in visual cortex. Short-term monocular deprivation by intraocular application of tetrodotoxin (TTX) in adult monkeys, resulted in a temporary decrease of GAD IR after 4 days (Hendry and Jones, 1986, 1988). However it did not change GAD67 mRNA in deprived ocular dominance columns (Benson et al., 1991). Longer treatment (15 days) caused a decrease of GAD mRNA (Benson et al., 1994). Similar treatment in adult cats did not change the number of cells expressing GAD67 mRNA in deprived ocular dominance columns of the visual cortex (Benson et al., 1989), but the levels of mRNA expression were not measured in the study. Restricted bilateral lesions of retinae also did not influence the level of GAD67 mRNA in corresponding visual cortex in cat (Arckens et al., 1998). In the somatosensory system, changes in sensory input were seen to influence GAD IR and GAD mRNA expression in the cortex. Experiments by Welker have demonstrated a transient decrease of GAD IR in cortical representation of removed vibrissal follicles in adult mice (Welker et al., 1989b). Our previous studies showed that the same procedure resulted in a decrease of GAD67 mRNA level in layer IV of barrel cortex (Gierdalski et al., 1999). GAD IR was reversibly reduced in the barrels corresponding to trimmed vibrissae (Akhtar and Land, 1991). Sciatic nerve injury in rat led to a decrease of the enzymatic activity of GAD and its immunoreactivity in primary somatosensory cortex (Krohn et al., 1992; Warren et al., 1989). On the other hand, intensification of input (constant vibrissal stimulation for 4 days) resulted in temporary increase of GAD IR in respective barrels (Welker et al., 1989a).

The present study showed that the level of GAD67 mRNA was rapidly elevated after a 10-min long physiological stimulation. However, the effect was short-lasting. We chose a 4 h latency period for studying the early effect of a single stimulation session to distinguish the actual response from the resting level of transcript. Activation of the gad67 gene, which is an effector gene, requires a cascade of cellular events, such as activation of immediate early genes and recruitment of other transcription factors. Longer stimulation (three sessions), which is sufficient to provoke plastic rearrangement of cortical sensory maps, led to elevation of mRNA levels that lasted for at least 24 h. The response was also greater, indicating that effects of consecutive stimulations were cumulative. Prolonged decay of mRNA level, after three sessions of stimulation may be attributed to slow breakdown of the transcript. Alternatively, it may be caused by increased or prolonged transcription due to the 3 day long stimulation. The existence of delayed effects was also examined by measuring the GAD67 transcript level 5 days after three training sessions. Previous studies by Siucinska showed that after such a treatment period the learning-induced cortical map rearrangements in the barrel cortex returned to control values (Siucinska and Kossut, 1996).

We also demonstrated increased density of GAD67 immunoreactive neurons in the hollows of barrels representing the row of vibrissae activated during the training. The effect was specific to the group that received CS + UCS. The elements of perceptual learning in animals receiving CS alone can not be excluded, however, we have previously demonstrated that pseudo- conditioning, in which animals received the same number of CS and UCS but in an unpaired manner, did not induce plastic changes in cortical vibrissal representations. The observed increase in density of GAD67 IR neurons is therefore linked to the treatment that induced the plastic change of cortical representation.

In the barrel cortex, both GAD67 and GAD65 IR are found in neuropil and puncta surrounding cell bodies, but only GAD67- immunoreactive somata form the barrel-like pattern (Kiser et al., 1998). We found that the density of GAD67 immunolabeled cell bodies increases in the barrel hollows of the trained row, although it does not obliterate the barrel-like pattern. We attribute this increase in labeled neurons density to enhanced synthesis of GAD67 protein, which allows labeling of a subset of neurons that in naive animals have GAD levels undetectable with the dilution of antibody used.

The physiological importance of increased GABA levels in the cortical representation of the ‘trained’ row of whiskers may lay in buffering the increased cortical activity. It may also have a more instructive meaning, such as the reorganization of interbarrel circuitry via a disinhibitory mechanism. Activity of layer IV neurons is under a strong inhibitory control (Kirkwood and Mear, 1994) and even partial elimination of this restrictive influence from thalamo-cortical input may significantly contribute to induction of plasticity. GABA-containing terminals synapsing with GABAergic neurons that may be the basis for disinhibition were described in electron microscopy studies (Keller and White, 1987). We suggested that local up-regulation of GABA synthesis in a subset of neurons may, by disinhibition, facilitate the propagation of reinforced input to adjacent rows of barrels (Siucinska et al., 1999). It can operate either by direct GABAergic inter-barrel action, that was detected in electrophysiological experiments (Salin and Prince, 1996) or by interactions within the same barrel column (Aroniadou and Keller, 1996). The decrease of inhibition is known to enhance NMDA-mediated activity in juvenile rat somatosensory cortex (Luhmann and Prince, 1990) and may also facilitate the training- induced plastic changes.

It is known that many cortical inter-neuron subtypes operate in a layer-specific fashion (Gonchar and Burkhalter, 1997; Hensch et al., 1998). In the present study, as in our previous investigations of effects of deafferentation upon GAD67 mRNA in the barrel cortex, we observed changes in the level of the transcript only in cortical layer IV. Our GABA IR study localized the up-regulation of density of GABA-immunoreactive neurons to a subset of small diameter cells within the hollows (but not in the sides) of barrels of the expanding cortical representation. As suggested by Hensch, only a specific subset of the elaborate network of inter-neurons may be necessary to sculpt cortical activity (Hensch et al., 1998). Further work is needed to determine the exact type of GABAergic neurons that is affected by learning-dependent plasticity.

We showed a significant increase of the level of GAD67 mRNA in the representation of a stimulated row of vibrissae after only one session of tactile stimulation, as well as after one session of conditioning. An increase of transcript level could be a result of increased transcription rate and/or decreased mRNA degradation.

It is possible that the UCS triggers an intracellular response in barrel cortex GABAergic neurons, which allows an increase in translation rate and promotes utilization of GAD67 mRNA already increased by the CS. An alternative (or additional) mechanism is also possible. Because the gad67 gene has two different alternative promoters, CS and UCS effects might converge on promoter regions of the gene and change their activity, resulting in altered proportions of alternative transcripts available for translation. The way in which the aversive unconditioned stimulus can affect each of these hypothetical mechanisms of regulation of translation is an important issue for understanding the molecular substrate of learning. In vivo electrophysiological recordings have shown that introduction of UCS increases the amplitude of cortical evoked potential (Hall and Mark, 1967; Musial et al., 1998). Cortical activation is controlled by brainstem and basal forebrain structures and indeed stimulation of nucleus basalis can mimic the effect of UCS or operant reinforcement upon cortical responses (Rasmusson and Dykes, 1988; Edeline et al., 1994; Kilgard and Merzenich, 1998). It was recently demonstrated that single units in the barrel cortex can show rapidly induced acetylcholine-dependent plasticity of temporal response properties (Shulz et al., 2000). Acetylcholine may be the means of linking the UCS to intracellular changes affecting translation; the mechanism of its action requires further research.

In summary, our results show that physiological stimulation of the whisker-to-barrel pathway regulates GAD67 mRNA levels and that associative training leads to the increase in GAD67 IR in layer IV of the cortical representation of the stimulated vibrissae. Our results suggest that CS and UCS effects converge at the level of post-transcriptional mechanisms. To the best of our knowledge it is also the first direct report of post-transcriptional regulation of GAD67 protein expression affected by a learning experience.

Notes

We wish to thank Prof. Jolanta Skangiel-Kramska for comments on the manuscript and Dr Linda Porter for language corrections. The research was supported by the State Committee for Scientific Research grant no. 6PO4A01614 to M.K.

Address correspondence to Marcin Gierdalski, USUHS, Department Anatomy, Physiology, Genetics, 4301 Jones Bridge Road, Bethesda, MD 20814, USA. Email: mgierdalski@usuhs.mil.

Figure 1.

In situ hybridization to GAD67 mRNA in sections from naive control animals and animals from groups receiving conditioning (CS + UCS) or stimulation alone (CS) performed 1 and 24 h after the third session. The left column presents adjacent sections from a control animal — the upper section was stained with cresyl violet and shows the pattern of barrels (belonging to rows A, B, C, D and E). In pseudocolored autoradiogram of GAD67 mRNA ISH (lower panel) labeling of layer IV is segmented and accumulations of label corresponding to individual rows of barrels can be identified. In sections from both trained and stimulated animals the cortical representation of the stimulated row B of vibrissae exhibits a stronger hybridization signal than the other rows. In sections from control animal, all barrel rows are marked; in experimental mice the arrow points to the stimulated row B and the barrel field region is marked with an arc. ROD, relative optical density. Scale bar: 500 μm.

Figure 1.

In situ hybridization to GAD67 mRNA in sections from naive control animals and animals from groups receiving conditioning (CS + UCS) or stimulation alone (CS) performed 1 and 24 h after the third session. The left column presents adjacent sections from a control animal — the upper section was stained with cresyl violet and shows the pattern of barrels (belonging to rows A, B, C, D and E). In pseudocolored autoradiogram of GAD67 mRNA ISH (lower panel) labeling of layer IV is segmented and accumulations of label corresponding to individual rows of barrels can be identified. In sections from both trained and stimulated animals the cortical representation of the stimulated row B of vibrissae exhibits a stronger hybridization signal than the other rows. In sections from control animal, all barrel rows are marked; in experimental mice the arrow points to the stimulated row B and the barrel field region is marked with an arc. ROD, relative optical density. Scale bar: 500 μm.

Figure 2.

 Autoradiograms from control set of GAD67 ISH performed with the excess of heterologous (top) or homologous unlabeled oligonucleotide (bottom), along with a proper ISH (middle) demonstrating specificity of hybridization.

Figure 2.

 Autoradiograms from control set of GAD67 ISH performed with the excess of heterologous (top) or homologous unlabeled oligonucleotide (bottom), along with a proper ISH (middle) demonstrating specificity of hybridization.

Figure 3.

 (A) Quantitative results of ISH in individual barrel columns in all cortical layers (II/III, IV, V, VI) of mice from group receiving conditioning training (CS + UCS), 24 h after the third session, in experimental and control hemisphere. Results are expressed as radioactivity (nCi) per mg tissue. Values are normalized as described in the subsection ‘Analysis of Autoradiograms’. A significant difference between representation of stimulated row B of vibrissae and other rows or between experimental and control hemisphere is seen only in layer IV. *P < 0.01. (B) A schematic diagram illustrating how labeling measurements are taken from in situ hybridization autoradiograms. The reading window was circular and it was positioned over the part of autoradiogram corresponding to the histological barrel (insert), and in the supra- and infragranular layers in register with barrels. Additional readings were taken in the cingulate cortex at the same laminar positions. A–E, vibrissal columns. II/III–VI, cortical layers; Ref, cingulate cortex; V, ventral; M, medial. (C) Autoradiogram of GAD67 mRNA ISH from oblique coronal section of a control hemisphere, no significant differences were seen between labeling of five cortical columns in any layer. (D) Autoradiogram from a section of experimental hemisphere; increased hybridization signal was seen in a barrel of row B.

Figure 3.

 (A) Quantitative results of ISH in individual barrel columns in all cortical layers (II/III, IV, V, VI) of mice from group receiving conditioning training (CS + UCS), 24 h after the third session, in experimental and control hemisphere. Results are expressed as radioactivity (nCi) per mg tissue. Values are normalized as described in the subsection ‘Analysis of Autoradiograms’. A significant difference between representation of stimulated row B of vibrissae and other rows or between experimental and control hemisphere is seen only in layer IV. *P < 0.01. (B) A schematic diagram illustrating how labeling measurements are taken from in situ hybridization autoradiograms. The reading window was circular and it was positioned over the part of autoradiogram corresponding to the histological barrel (insert), and in the supra- and infragranular layers in register with barrels. Additional readings were taken in the cingulate cortex at the same laminar positions. A–E, vibrissal columns. II/III–VI, cortical layers; Ref, cingulate cortex; V, ventral; M, medial. (C) Autoradiogram of GAD67 mRNA ISH from oblique coronal section of a control hemisphere, no significant differences were seen between labeling of five cortical columns in any layer. (D) Autoradiogram from a section of experimental hemisphere; increased hybridization signal was seen in a barrel of row B.

Figure 4.

 Sections tangential to the barrel field at the level of the layer IV obtained from the animal of the trained (CS + UCS) group. (A) A section stained for succinyl dehydrogenase (SDH) to reveal the structure of the barrel field. (B) An adjacent section processed for in situ hybridization to GAD67 mRNA. An elevated GAD67 hybridization signal is seen in the cortical representation of the stimulated row B of whiskers (arrows). Scale bar, 500 μm.

Figure 4.

 Sections tangential to the barrel field at the level of the layer IV obtained from the animal of the trained (CS + UCS) group. (A) A section stained for succinyl dehydrogenase (SDH) to reveal the structure of the barrel field. (B) An adjacent section processed for in situ hybridization to GAD67 mRNA. An elevated GAD67 hybridization signal is seen in the cortical representation of the stimulated row B of whiskers (arrows). Scale bar, 500 μm.

Figure 5.

 Quantitative results of ISH in layer IV of mice at different times (1, 4 and 24 h, 5 days) after stimulation alone (CS) or classical conditioning (CS + UCS) of various duration (one and three sessions) in experimental and control hemispheres and in layer IV of unoperated control, in right and left hemispheres. Results are expressed as a relative intensity in row B to the average value in rows A, C, D and E. Asterisks denote groups where results are significantly different from the naive controls. *P < 0.05, **P < 0.01. Values are means±SDs.

Figure 5.

 Quantitative results of ISH in layer IV of mice at different times (1, 4 and 24 h, 5 days) after stimulation alone (CS) or classical conditioning (CS + UCS) of various duration (one and three sessions) in experimental and control hemispheres and in layer IV of unoperated control, in right and left hemispheres. Results are expressed as a relative intensity in row B to the average value in rows A, C, D and E. Asterisks denote groups where results are significantly different from the naive controls. *P < 0.05, **P < 0.01. Values are means±SDs.

Figure 6.

 Upper panel shows higher power magnification (40×) of individual barrels belonging to the row B obtained from tangential sections processed for GAD67 immunohistochemistry. Lower panel shows the drawings of stained cells with outlined barrel hollows. Left – a barrel from the experimental hemisphere from an animal receiving CS + UCS 24 h after three sessions of stimulation. Right – a barrel from a naive control animal. Scale bar, 100 μm.

Figure 6.

 Upper panel shows higher power magnification (40×) of individual barrels belonging to the row B obtained from tangential sections processed for GAD67 immunohistochemistry. Lower panel shows the drawings of stained cells with outlined barrel hollows. Left – a barrel from the experimental hemisphere from an animal receiving CS + UCS 24 h after three sessions of stimulation. Right – a barrel from a naive control animal. Scale bar, 100 μm.

Figure 7.

 Quantitative results of GAD67 IR in row B hollows in both experimental and control cerebral hemispheres. Results are expressed as a number of GAD67 immunoreactive cells per 105 μm2. There is a significant difference between experimental and control hemisphere in group of animals receiving CS + UCS. The asterisk denotes a group where the experimental hemisphere differs significantly from the control one. *P<0.05, n = 5. Data are means±SDs.

Figure 7.

 Quantitative results of GAD67 IR in row B hollows in both experimental and control cerebral hemispheres. Results are expressed as a number of GAD67 immunoreactive cells per 105 μm2. There is a significant difference between experimental and control hemisphere in group of animals receiving CS + UCS. The asterisk denotes a group where the experimental hemisphere differs significantly from the control one. *P<0.05, n = 5. Data are means±SDs.

References

Akhtar ND, Land PW (
1991
) Activity-dependent regulation of glutamic acid decarboxylase in the rat barrel cortex: effects of neonatal versus adult sensory deprivation.
J Comp Neurol
 
307
:
200
–213.
Arckens L, Eysel UT, Vanderhaeghen JJ, Orban GA, Vandesande F (
1998
) Effect of sensory deafferentation on the GABAergic circuitry of the adult cat visual system.
Neuroscience
 
83
:
381
–391.
Aroniadou VA, Keller A (
1996
) Intrinsic inhibitory pathways in the mouse barrel cortex.
NeuroReport
 
7
:
2363
–2368.
Asada H, Kawamura Y, Maruyama K, Kume H, Ding R, Ji FY, Kanbara N, Kuzume H, Sanbo M, Yagi T, Obata K (
1996
) Mice lacking the 65 kDa isoform of glutamic acid decarboxylase (GAD65) maintain normal levels of GAD67 and GABA in their brains but are susceptible to seizures.
Biochem Biophys Res Commun
 
229
:
891
–895.
Asada H, Kawamura Y, Maruyama K, Kume H, Ding RG, Kanbara N, Kuzume H, Sanbo M, Yagi T, Obata K (
1997
) Cleft palate and decreased brain gamma-aminobutyric acid in mice lacking the 67-kDa isoform of glutamic acid decarboxylase.
Proc Natl Acad Sci USA
 
94
:
6496
–6499.
Benson DL, Isackson PJ, Hendry SH, Jones EG (
1989
) Expression of glutamic acid decarboxylase mRNA in normal and monocularly deprived cat visual cortex.
Mol Brain Res
 
5
:
279
–287.
Benson DL, Isackson PJ, Gall CM, Jones EG (
1991
) Differential effects of monocular deprivation on glutamic acid decarboxylase and type II calcium-calmodulin-dependent protein kinase gene expression in the adult monkey visual cortex.
J Neurosci
 
11
:
31
–47.
Benson DL, Huntsman MM, Jones EG (
1994
) Activity-dependent changes in GAD and preprotachykinin mRNAs in visual cortex of adult monkeys.
Cereb Cortex
 
4
:
40
–51.
Bu DF, Erlander MG, Hitz BC, Tillakaratne NJ, Kaufman DL, Wagner-McPherson CB, Evans GA, Tobin AJ (
1992
) Two human glutamate decarboxylases, 65-kDa GAD and 67-kDa GAD, are each encoded by a single gene.
Proc Natl Acad Sci USA
 
89
:
2115
–2119.
Drengler SM, Oltmans GA (
1993
) Rapid increases in cerebellar Purkinje cell glutamic acid decarboxylase (GAD67) mRNA after lesion-induced increases in cell firing.
Brain Res
 
615
:
175
–179.
Edeline J-M, Hars B, Maho C, Hennevin E (
1994
) Transient and prolonged facilitation of tone-evoked responses induced by basal forebrain stimulation in the rat auditory cortex.
Exp Brain Res
 
97
:
373
–386.
Erlander MG, Tobin AJ (
1991
) The structural and functional heterogeneity of glutamic acid decarboxylase: a review.
Neurochem Res
 
16
:
215
–226.
Erlander MG, Tillakaratne NJ, Feldblum S, Patel N, Tobin AJ (
1991
) Two genes encode distinct glutamate decarboxylases.
Neuron
 
7
:
91
–100.
Esclapez M, Houser CR (
1999
) Up-regulation of GAD65 and GAD67 in remaining hippocampal GABA neurons in a model of temporal lobe epilepsy.
J Comp Neurol
 
412
:
488
–505.
Esclapez M, Tillakaratne NJ, Tobin AJ, Houser CR (
1993
) Comparative localization of mRNAs encoding two forms of glutamic acid decarboxylase with nonradioactive in situ hybridization methods.
J Comp Neurol
 
331
:
339
–362.
Feldblum S, Erlander MG, Tobin AJ (
1993
) Different distributions of GAD65 and GAD67 mRNAs suggest that the two glutamate decarboxylases play distinctive functional roles.
J Neurosci Res
 
34
:
689
–706.
Gierdalski M, Jablonska B, Smith A, Skangiel-Kramska J, Kossut M (
1999
) Deafferentation induced changes in GAD67 and GluR2 mRNA expression in mouse somatosensory cortex.
Mol Brain Res
 
71
:
111
–119.
Gonchar Y, Burkhalter A (
1997
) Three distinct families of GABAergic neurons in rat visual cortex.
Cereb Cortex
 
7
:
347
–358.
Hall RD, Mark RG (
1967
) Fear and the modification of acoustically evoked potentials during conditioning.
J Neurophysiol
 
30
:
893
–910.
Hendry SHC, Jones EG (
1986
) Reduction in number of immunostained GABAergic neurones in deprived-eye dominance columns of monkey area 17.
Nature
 
320
:
750
–753.
Hendry SHC, Jones EG (
1988
) Activity-dependent regulation of GABA expression in the visual cortex of adult monkeys.
Neuron
 
1
:
701
–712.
Hensch TK, Fagiolini M, Mataga N, Stryker MP, Baekkeskov S, Kash SF (
1998
) Local GABA circuit control of experience-dependent plasticity in developing visual cortex.
Science
 
282
:
1504
–1508.
Jacobs KM, Donohue JP (
1991
) Reshaping the cortical motor map by unmasking latent intracortical connections.
Science
 
251
:
944
–947.
Jones EG (
1993
) GABAergic neurons and their role in cortical plasticity in primates.
Cereb Cortex
 
3
:
361
–372.
Kaufman DL, Houser CR, Tobin AJ (
1991
) Two forms of the gamma-aminobutyric acid synthetic enzyme glutamate decarboxylase have distinct intraneuronal distributions and cofactor interactions.
J Neurochem
 
56
:
720
–723.
Keller A, White EL (
1987
) Synaptic organization of GABAergic neurons in the mouse SmI cortex.
J Comp Neurol
 
262
:
1
–12.
Kilgard MP, Merzenich MM (
1998
) Cortical map reorganization enabled by nucleus basalis activity.
Science
 
279
:
1714
–1718.
Kirkwood A, Mear MF (
1994
) Hebbian synapses in visual cortex.
J Neurosci
 
14
:
1634
–1645.
Kiser PJ, Cooper NG, Mower GD (
1998
) Expression of two forms of glutamic acid decarboxylase (GAD67 and GAD65) during postnatal development of rat somatosensory barrel cortex.
J Comp Neurol
 
402
:
62
–74.
Krohn K, Rothe T, Biesold D (
1992
) High-affinity uptake of GABA and glutamate decarboxylase activity in rat primary somatosensory cortex after sciatic nerve injury.
Mol Chem Neuropathol
 
16
:
159
–169.
Laprade N, Soghomonian JJ (
1995
) Differential regulation of mRNA levels encoding for the two isoforms of glutamate decarboxylase (GAD65 and GAD67) by dopamine receptors in the rat striatum.
Mol Brain Res
 
34
:
65
–74.
Liang F, Isackson PJ, Jones EG (
1996
) Stimulus-dependent, reciprocal up- and downregulation of glutamic acid decarboxylase and Ca2+- dependent protein kinase II gene expression in rat cerebral cortex.
Exp Brain Res
 
110
:
163
–174.
Lin CS, Lu SM, Schmechel DE (
1985
) Glutamic acid decarboxylase immunoreactivity in layer IV of barrel cortex of rat and mouse.
J Neurosci
 
5
:
1934
–1939.
Luhmann HJ, Prince DA (
1990
) Control of NMDA receptor-mediated activity by GABAergic mechanisms in mature and developing rat neocortex.
Dev Brain Res
 
54
:
287
–290.
Martin DL (
1987
) Regulatory properties of brain glutamate decarboxylase.
Cell Mol Neurobiol
 
7
:
237
–253.
Musial P, Kublik E, Panecki S, Wrobel A (
1998
) Transient changes of electrical activity in the rat barrel cortex during conditioning.
Brain Res
 
786
:
1
–10.
Rasmusson DD, Dykes RW (
1988
) Long-term enhancement of evoked potentials in cat somatosensory cortex produced by co-activation of the basal forebrain cutaneous receptors.
Exp Brain Res
 
70
:
276
–287.
Salin PA, Prince DA (
1996
) Electrophysiological mapping of GABAA receptor-mediated inhibition in adult rat somatosensory cortex.
J Neurophysiol
 
75
:
1589
–1600.
Shulz DE, Sosnik R, Ego V, Haidarifu S, Ahissar E (
2000
) A neuronal analogue of state-dependent learning.
Nature
 
403
:
549
–552.
Sillito AM (1984) Functional considerations of the operation of GABAergic inhibitory process in the visual cortex. In: Cerebral cortex, vol. 2, Functional properties of cortical cells (Peters A, Jones EG, eds), pp. 91–117. New York: Plenum.
Siucinska E, Kossut M (
1996
) Short-lasting classical conditioning induces reversible changes of representational maps of vibrissae in mouse SI cortex – a 2DG study.
Cereb Cortex
 
6
:
506
–513.
Siucinska E, Kossut M, Stewart MG (
1999
) GABA immunoreactivity in mouse barrel field after aversive and appetitive classical conditioning training involving facial vibrissae.
Brain Res.
 
843
:
62
–70.
Sloviter RS, Dichter MA, Rachinsky TL, Dean E, Goodman JH, Sollas AL, Martin DL (
1996
) Basal expression and induction of glutamate decarboxylase and GABA in excitatory granule cells of the rat and monkey hippocampal dentate gyrus.
J Comp Neurol
 
373
:
593
–618.
Warren R, Tremblay N, Dykes RW (
1989
) Quantitative study of glutamic acid decarboxylase-immunoreactive neurons and cytochrome oxidase activity in normal and partially deafferented rat hindlimb somatosensory cortex.
J Comp Neurol
 
288
:
583
–592.
Welker E, Soriano E, Dorfl J, Van der Loos H (
1989
) Plasticity in the barrel cortex of the adult mouse: transient increase of GAD-immunoreactivity following sensory stimulation.
Exp Brain Res
 
78
:
659
–664.
Welker E, Soriano E, Van der Loos H (
1989
) Plasticity in the barrel cortex of the adult mouse: effects of peripheral deprivation on GAD-immunoreactivity.
Exp Brain Res
 
74
:
441
–452.
Wisden W, Morris BJ (1994) in situ Hybridization with Synthetic Oligonucleotide Probes. In: in situ Hybridization Protocols for the Brain (Wisden W, Morris B, eds.) pp 9–34. Academic Press.
Woolsey TA, Van der Loos H (
1970
) The structural organization of layer IV in the somatosensory region (SI) of mouse cerebral cortex. The description of a cortical field composed of discrete cytoarchitectonic units.
Brain Res
 
17
:
205
–242.