Introduction

It is now well established that plastic changes occur in the somatosensory cortex in response to surgical manipulations inducing drastic changes in sensory inflow—such as amputation or nerve section—or to exposure to novel experiences. These procedures cause a shrinkage of the cortical representation of the affected body part associated with a reorganization of topographic maps and with an expansion of receptive fields (Buonomano and Merzenich 1998). Hindlimb unloading (HU) is a ground-based model of weightlessness. The HU conditions are generally obtained in the rat by hindlimb suspension (Wronski and Morey-Holton 1987), in which contact of the plantar sole with the ground is prevented. This model can therefore be considered as a “functional deafferentation”, since the tactile information from the paw and the proprioceptive input from the limb are dramatically reduced, at least during the first week of HU (De-Doncker et al. 2005; Kawano et al. 2004). A 14-day period of HU produces a remapping of cortical maps (Langlet et al. 1999) and a higher excitation of cortical cells to peripheral stimulation (Dupont et al. 2003; Langlet et al. 2001).

Mechanisms of the cortical functional reorganization after a change in sensory experience have become the focus of interest. In the visual cortex, results from a large number of studies have indicated that cortical plasticity might be linked to a change in the balance of excitation and inhibition, with a loss of inhibition and/or a higher excitation level. In particular, several studies have shown the role of γ-aminobutyric acid (GABA) inhibition and glutamate excitation in the maintenance of cortical maps (Buonomano and Merzenich 1998; Jones 2000). In the somatosensory cortex, many studies have also emphasized the role of GABA and glutamate in the cortical reorganization following amputation or nerve section. For instance, in the rat (Lane et al. 1997) and raccoon (Tremere et al. 2001), it has been shown that GABAergic activity is involved in controlling receptive field size after reorganization. Moreover, a significant increase in GABAA receptors was observed 2 weeks after a single-digit amputation in the raccoon (He et al. 2004) or several years after an upper limb amputation in humans (Capaday et al. 2000). Furthermore, glutamate seems to be essential for the development of cortical plasticity since the reorganization that follows peripheral nerve injury in adult monkeys is dependent on N-methyl-d-aspartate (NMDA) receptors (Garraghty and Muja 1996; Kano et al. 1991). In addition, the level of α-amino-3-hydroxy-5-methyl-4-isoxazole propionate (AMPA) receptors is increased 1–9 days after single-digit amputation in the raccoon (He et al. 2004).

All these experiments concerned a total deprivation of input to the somatosensory cortex. By contrast, our model of functional deafferentation (HU) is very different from those described above (visual deprivation, nerve section, amputation), since it affects a very restricted part of the cortex, and it does not totally silence this cortical zone. The role of amino acid neurotransmitters following this type of sensorimotor restriction is less documented and concerns exclusively GABA. In particular, data obtained after 14 days of HU suggest a down-regulation of GABAergic function: immunoreactivity of local circuit GABAergic neurons in the cortical hindlimb representation was reduced (D’Amelio et al. 1996) and an electrophysiological work has shown that cells supposed to be GABAergic were less frequently encountered in HU rats than in control ones (Dupont et al. 2003).

The aim of the present study was to determine whether functional deafferentation (HU) can change the concentration in excitatory and inhibitory amino acid neurotransmitters in the deprived cortex. Moreover, these results were compared to those obtained following deafferentation by surgical means (selective dorsal rhizotomy).

Materials and methods

Animals

Fifty-six adult male Wistar rats weighing 280–380 g were divided into four groups: control (CON, n=14), submitted to 7 days of HU (HU-7, n=14), submitted to 14 days of HU (HU-14, n=14), deafferented by dorsal rhizotomy (DRH, n=14). The animals were housed individually under temperature- and light-controlled conditions (23°C, 12-h light/12-h dark cycle). All procedures described below were approved by both the Agricultural and Forest Ministry and the National Education Ministry (veterinary service of health and animal protection, authorization 59-00999).

Hindlimb unloading

Hindpaw unloading was obtained using the model adapted from Wronski and Morey-Holton (1987). The tail of each rat was cleaned, dried and wrapped in antiallergenic adhesive plaster. This cast was secured to an overhead swivel that permitted 360° rotation and allowed the rats to walk freely on their forelimbs and have free access to food and water. The rats were unloaded by the tail at a ~30° head-down angle in order to avoid contact of the hindlimbs with the ground. They were unloaded during 14 days.

Rhizotomy

Surgical procedure has already been described in a previous paper (Picquet and Falempin 2003). Briefly, the animals were deeply anaesthetized with sodium pentobarbital (60 mg kg−1, ip). Supplementary doses were given when necessary. Under aseptic conditions, the midline dorsal musculature was retracted and a laminectomy was performed between L2 and S1. The dorsal roots L3–L5 were identified and sectioned bilaterally at their entry into the spinal cord. Dorsal musculature and skin were then sutured, and an antiseptic (Betadine) was applied to the cutaneous incision area. Antibiotics (Trisulmix, 100 μl/kg) and analgesic (Metacam, 100 μl/kg) were orally administered for 6 days following surgery. Rats were killed 17 days following surgery.

Hindpaw somatosensory cortex removal and tissue extraction

The animals were killed by decapitation. The head was immediately surrounded with ice and placed in a stereotaxic frame. A craniotomy was performed to expose the somatosensory cortex. The dura mater was incised and resected. A plastic cylinder (1 mm inner diameter) mounted on a syringe was used to remove a column of hindpaw somatosensory cortex by aspiration at the stereotaxic coordinates Anterior −1 and Lateral 3 mm with Bregma as the bone reference point. These coordinates correspond to the functional center of the hindpaw sensory cortical map in CON and HU rats (Canu et al. 2003). In addition, a sample of visual cortex was taken at the stereotaxic coordinates Anterior −7 and Lateral 3 mm with Bregma as the bone reference point to serve as internal control. The cortex was weighed and immediately frozen and stored at −80°C for later use. The total duration of the cortex removal did not exceed 4 min.

High performance liquid chromatography

The tissue extracts were homogenized by adding 5 μl of water per milligram of tissue and by sonification. The homogenate was then centrifuged (15,000g × 15 min, 4°C) and 10 μl of the supernatant was diluted in 990 μl of water and filtered with a 1 ml syringe through a 0.2 μm Millex-GV filter (Nihon Millipore Ltd., Yonezawa, Japan).

The amino acid concentrations were assessed by high performance liquid chromatography with electrochemical detection. The system was configured for AAA-Direct (Dionex SA, Voisins Le Bretonneux, France). The anion-exchange column set was the AminoPac PA 10 (2× 250 mm; P/N 55406; Dionex) with its PA10 guard (2× 50 mm; P/N 55407; Dionex). A gradient mixer GM4 (2 mm; P/N 49135; Dionex) was added to homogenize the eluent mix. A disposable gold working electrode was used.

All chromatography eluents, standards and samples were prepared using 18 MΩ-cm-deionized water, free of electrochemically active impurities. Eluents (water in channel A; 250 mM NaOH in channel B; 1 M sodium acetate in channel C) were prepared according to the supplier’s recommendations. All amino acids were separated using a flow rate of 0.25 ml/min and a column temperature of 30°C. The standards and the samples were placed in an autosampler (Dionex SA) for a 24-h maximal period. The sampling sequence was randomly determined to avoid a possible assay effect and to make sure that each sample would remain less than 24 h in the autosampler. The chromatograms were analysed and integrated with Chromeleon software (version 6.40; Dionex Corporation).

Statistical analysis

The protein content was determined using the method of Folin-Lowry to express the results as μmol amino acid per g protein ± SEM. Values obtained in the two unloaded groups (HU-7 and HU-14) were compared by a Mann–Whitney test. Since the values were very similar (P>0.37 to P>0.91), they have been pooled. A one way analysis of variance (ANOVA) followed by a Dunnett’s post hoc test served to compare the CON, HU and DRH groups. A P value of less than 0.05 was considered significant.

Results

Eighteen amino acids were identified (Fig. 1). We will focus on the main neurotransmitters, i.e. GABA, glycine, glutamate, glutamine, aspartate and taurine. For all other amino acids, no change was observed after treatment.

Fig. 1
figure 1

Typical chromatograms obtained in CON, HU and DRH rats

Full size image

The visual cortex was taken as an internal control. None of the treatments (neither unloading nor rhizotomy) modified the amino acid content studied in the visual cortex (Table 1). In consequence, the changes in the amino acid levels described below were specific to the somatosensory cortex and did not result from a global effect of HU or rhizotomy on the whole brain.

Table 1 Determination of amino acid concentrations (in μmol/g protein) in the sensorimotor (S) and visual (V) cortex of the rat
Full size table

In the somatosensory cortex, the GABA level was decreased (ANOVA: P=0.003) in both unloaded (−40%; P<0.001) and rhizotomized rats (−41%; P<0.001). Mean values are presented in Table 1. Values were, respectively, 0.286±0.033 and 0.247±0.029 for the HU-7 and HU-14 groups (Mann–Whitney: P=0.371). The level of GABA precursor glutamine was slightly decreased, although not significantly (unloaded −17%; rhizotomized −14%). The level of taurine was also altered by the treatment (ANOVA: P=0.031); unloading and rhizotomy induced a decrease of, respectively, −20% (P<0.05) and −25% (P<0.05) in the taurine concentration. The value was 0.327±0.017 for the HU-7 group and 0.365±0.038 for the HU-14 group (Mann–Whitney: P=0.647). By contrast, no significant difference could be detected in the glycine concentration after unloading (−24%; P>0.05) nor after rhizotomy (−13%; P>0.05).

Concerning excitatory amino acid, deafferentation tended to decrease the level of aspartate (ANOVA: P=0.068; unloaded −34%; rhizotomized −34%), whereas no difference was noticed for glutamate (unloaded −5%; rhizotomized −13%).

Discussion

The present findings indicate that a functional (HU) or surgical (rhizotomy) deafferentation of the hindlimb representation in the somatosensory cortex results in a decrease in inhibitory amino acids without any changes in excitatory amino acids.

Amino acids concentration

Deafferentation of the somatosensory cortex resulted in a decrease in the GABA content. At the cortical level, there are numerous observations indicating that GABA containing neurons are the major controlling elements determining receptive field size and location (Dykes 1996; Jones 1993). The administration of a blocker of GABA receptors results in a dramatic increase in the receptive fields of low threshold (Alloway et al. 1989; Hicks and Dykes 1983; Tremere et al. 2001). Many studies have described an increase in the size of receptive fields after surgical (Buonomano and Merzenich 1998 for review) or functional deafferentation by immobilization (Coq and Xerri 1999) or HU (Langlet et al. 1999). Based on these observations, these authors suggested that the level of GABA in the deafferented cortex might be reduced. Previous studies have provided indirect arguments to support this statement. Firstly, a study performed in the rat after unloading has shown that the cells with thin spikes, supposed to be GABAergic, were less frequently encountered in rats submitted to HU than in control ones (Dupont et al. 2003). Secondly, an up-regulation of GABAA receptors has been reported 2–7 weeks following single-digit amputation in the raccoon (He et al. 2004), or several years after an upper limb amputation in humans (Capaday et al. 2000). An increase in the strength of inhibitory synapses (He et al. 2004) or a homeostatic process to compensate for a decrease in GABA levels (Capaday et al. 2000) have been put forward to explain the increase in GABA receptors. Our results are in favour of the second hypothesis. It is indeed unlikely that the decrease in GABA level corresponds to an increased activity of inhibitory synapses. The measure of neurotransmitter level thus brings additional information to the experiments where receptor levels were evaluated. Thirdly, a decrease in the number of immunoreactive cells for GABA has also been frequently described in the visual cortex of monkey following visual deprivation (Hendry and Jones 1986), and in the somatosensory cortex following neonatal denervation in the mouse (Kossut et al. 1991), or after a 14-day period of HU (D’Amelio et al. 1996). However, it was not established whether this depletion was due to an alteration in its synthetic activity or to an increased release. In the present work, we measured directly the concentration of GABA in the deafferented cortex. Since GABA levels were decreased, the hypothesis of an increase of GABA seems unlikely. On the contrary, a decrease in the synthesis process and/or an increase in the enzymatic catabolism can be considered. To further support this statement, we have shown that the level of a GABA precursor, glutamine, was unchanged in both unloaded and rhizotomized rats. Thus, the decrease in the GABA level is more likely due to a down-regulation of its synthesizing enzyme, glutamic acid decarboxylase (GAD). As a matter of fact, a deafferentation of the cortex by partial vibrissectomy leads to a decrease in the expression of GAD 67 mRNA (Gierdalski et al. 1999).

Taurine is one of the most abundant free amino acids in the brain (Della Corte et al. 2002). Although a specific receptor has not been clearly characterized, it gathers many of the criteria of a neurotransmitter. Taurine exerts its inhibitory action through GABAA or glycine receptors (Della Corte et al. 2002). Therefore, the decrease in taurine concentration further reinforces the idea of a release of inhibition in the deafferented cortex.

The participation of NMDA receptors in nerve injury-induced plasticity is still a matter of debate (Kano et al. 1991; Garraghty and Muja 1996; Myers et al. 2000), and we have no data concerning experience-dependent plasticity. In the present work, the levels in aspartate, glutamate and glycine were unchanged following HU or rhizotomy. Aspartate is an excitatory neurotransmitter, acting on NMDA receptors with a weak affinity. However, the precise role of aspartate in excitatory transmission is currently unknown (Leonard 2003). Glutamate and glycine subserve two important roles as neurotransmitters in the brain. Glutamate is not only the main excitatory neurotransmitter, acting through NMDA receptors, but is also a precursor of GABA. Glycine has a fast inhibitory action via specific receptors, not only in the spinal cord and in the brain stem, but also in other regions such as the sensorimotor cortex (Levi et al. 1982); in addition, it participates in excitatory neurotransmission as a co-agonist of glutamate on NMDA receptors (Fletcher et al. 1990). The lack of changes in the level of these neurotransmitters suggest that reorganization does not necessitate activation of excitatory mechanisms. Our results are in agreement with those of Levy et al. (2002) who reported, in addition to a decrease in GABA concentration, a lack of change in glutamate level in the cortex following ischemic nerve block. To conclude, the present results support the idea that cortical changes are more likely due to a release from inhibition than due to an increased excitation.

Comparison of unloaded and rhizotomized rats

All changes described above were similar with either the deafferentation mode used here or the unloading duration. Unloading results in a temporary decrease in sensory input within the first week of unloading, with a gradual recovery afterwards (De-Doncker et al. 2005; Kawano et al. 2004) whereas rhizotomy corresponds to a section of dorsal roots, and in consequence should suppress totally sensory inflow in the long term. Therefore, we could have expected that the latter model would have a deeper impact on the cortex. The fact that different procedures have similar effects at the cortical level suggests that a plasticity phenomenon might occur at different levels of the sensory pathway, i.e. spinal cord, brain stem, or thalamus.