Differential GABAergic disinhibition during the development of painful peripheral neuropathy

Neuroscience 184 (2011) 183–194

DIFFERENTIAL GABAERGIC DISINHIBITION DURING THE DEVELOPMENT OF PAINFUL PERIPHERAL NEUROPATHY S. P. JANSSEN,* M. TRUIN, M. VAN KLEEF AND E. A. JOOSTEN

An impaired spinal inhibitory neuronal activity is thought to be involved in neuropathic pain (NPP). Several lines of evidence suggest that loss of GABAergic inhibition plays a pivotal role. In this context, pharmacologically-induced inhibition of spinal GABAA and GABAB receptor-mediated inhibitory transmission is known to result in NPP-like behaviors (Yaksh, 1989; Malan et al., 2002; Lee et al., 2010). Furthermore, substantially reduced primary afferent-evoked inhibitory post-synaptic currents (IPSCs), which are primarily GABAA receptor-mediated, have been observed predominantly in the substantia gelatinosa (Rexed lamina II) of the spinal dorsal horn after peripheral nerve injury-induced NPP (Moore et al., 2002; Miletic et al., 2003; Scholz et al., 2005; Yowtak et al., 2011). From literature it is clear that the role and mechanisms of spinal GABAergic disinhibition in NPP may not only be different due to the animal model used and severity of mechanical/thermal hypersensitivity induced but might also be related to the duration after peripheral nerve injury studied. For instance, reduced levels of the GABA synthesizing enzyme glutamate decarboxylase (GAD)65, but not GAD67, together with decreased GABA-immunoreactive (GABA-IR) neuronal profiles were reported between 3 and 6 days post-injury (Ibuki et al., 1997; Eaton et al., 1998; Moore et al., 2002). On the other hand, evidence of increased dorsal horn GABA levels is also documented at this time point (Satoh and Omote, 1996). Interestingly, a down-regulation of GAD65-67 (Moore et al., 2002) and extracellular GABA levels (Stiller et al., 1996), decreased GABAB receptor binding properties (Castro-Lopes et al., 1995) and GABA-IR cell profiles (Castro-Lopes et al., 1993; Ibuki et al., 1997; Eaton et al., 1998) as well as apoptosis of GABAergic interneurons per se (Scholz et al., 2005) have been described 2 weeks post-injury. However, data demonstrating unaltered numbers of GABAergic interneurons and nerve boutons at GABAA synapses, intracellular dorsal horn GAD65-67 and GABA levels, GABA transporters 1 and 3, and GABAA and GABAB receptors have also been reported (Somers and Clemente, 2002; Polgar et al., 2003; Engle et al., 2006; Miletic and Miletic, 2008; Polgar and Todd, 2008). Furthermore, there is evidence of enhanced spinal GABA levels, GABAA receptors and GABAA receptor binding properties 2 weeks postinjury (Castro-Lopes et al., 1995; Satoh and Omote, 1996; Moore et al., 2002). Recently, it has been documented that spinal GABAA receptor-mediated IPSCs are critically dependent on the activity of the post-synaptic neuronal K⫹ Cl- cotransporter 2 (KCC2). Reduced KCC2 expression has shown to disrupt dorsal horn intracellular Cl- homeostasis causing nor-

Pain Management and Research Center, Department of Anesthesiology, Maastricht University Medical Center, P. Debyelaan 25, PO Box 5800, 6202 AZ Maastricht, The Netherlands

Abstract—An impaired spinal GABAergic inhibitory function is known to be pivotal in neuropathic pain (NPP). At present, data concerning time-dependent alterations within the GABAergic system itself and post-synaptic GABAA receptormediated inhibitory transmission are highly controversial, likely related to the experimental NPP model used. Furthermore, it is unknown whether the severity of NPP is determined by the degree of these GABAergic disturbances. In the present study we therefore examined in one experimental animal model whether anatomical changes within the spinal GABAergic system and its GABAA receptor-mediated inhibitory function are gradually aggravated during the development of partial sciatic nerve injury (PSNL)-induced NPP and are related to the severity of PSNL-induced hypersensitivity. Three and 16 days after a unilateral PSNL (early and late NPP, respectively), GABA-immunoreactivity (GABA-IR) and the number of GABA-IR neuronal profiles were determined in Rexed laminae 1–3 of lumbar spinal cord cryosections. Additionally, the efficiency of dorsal horn GABAA receptor-induced inhibition was examined by cation chloride cotransporter 2 (KCC2) immunoblotting. NPP-induced hypersensitivity was only observed at the ipsilateral side, both at early and late time points. During early NPP, a decrease in ipsilateral dorsal horn GABA-IR was observed without alterations in the number of GABA-IR neuronal profiles or KCC2 protein levels. In contrast, bilateral increases in spinal GABA-IR accompanied by an unchanged number of GABA-IR interneurons were observed during late NPP. This was furthermore attended with decreased ipsilateral KCC2 levels. Moreover, the degree of hypersensitivity was not related to disturbances within the spinal GABAergic system at all time points examined. In conclusion, our anatomical data suggest that a dysfunctional GABA production is likely to be involved in early NPP whereas late NPP is characterized by a combined dysfunctional GABA release and decreased KCC2 levels, the latter suggesting an impaired GABAA receptor-mediated inhibition. © 2011 IBRO. Published by Elsevier Ltd. All rights reserved. Key words: rat, neuropathic pain, spinal cord, inhibition, GABA, KCC2. *Corresponding author. Tel: ⫹31-433881057; fax: ⫹31-433671096. E-mail address: [email protected] (S. P. Janssen). Abbreviations: CCI, chronic constriction injury; GABA-IR, ␥-aminobutyric acid-immunoreactivity; GAD, glutamate decarboxylase; IPSC, inhibitory post-synaptic current; KCC2, cation chloride cotransporter 2; NPP, neuropathic pain; PSNL, partial sciatic nerve ligation; SCS, spinal cord stimulation; SNI, spared nerve injury; TBS, Tris-buffered saline; TBS-T, Tris-buffered saline with 0.3% Triton X-100.

0306-4522/11 $ - see front matter © 2011 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2011.03.060

183

184

S. P. Janssen et al. / Neuroscience 184 (2011) 183–194

mally inhibitory GABA-mediated currents to become less inhibitory or even excitatory within a time frame of hours to weeks after peripheral nerve injury-induced NPP (Coull et al., 2003; Miletic and Miletic, 2008). However, unaltered spinal KCC2 protein levels at intermediate time points post-nerve injury (Miletic and Miletic, 2008) furthermore strengthen the earlier discussed notion of duration and animal model-specific alterations in the role of spinal GABAergic inhibition. In view of the ongoing discussion of the exact role of the GABAergic system in NPP, we decided to study the role of NPP-induced dysfunctional GABAergic inhibition using one experimental model in which various time points after injury as well as the severity of nerve injury-induced hypersensitivity were included. The particular model selected consisted of partial sciatic nerve ligation (PSNL) due to the fact that this is an often used model in studies of NPP and moreover because this model is exclusively used in studies on the analgesic properties of spinal cord stimulation (SCS). In this context, previous studies showed that the pain-relieving effects of experimental SCS were accomplished by increased extracellular spinal GABA levels occurring 16 days after induction of PSNL (Stiller et al., 1996; Cui et al., 1997). These biochemical data strongly suggest that GABA release from GABAergic interneurons is deregulated in chronic NPP. Besides this, it is known that the severity of PSNL-induced hypersensitivity 16 days post-injury predicts the pain-relieving effect of SCS (Smits et al., 2006) and one could consequently formulate the hypothesis that the severity of NPP is determined by the degree of GABAergic disinhibition. Furthermore, with respect to the timing after nerve injury, it is important to note that recent studies have shown a differential effect in SCS-mediated analgesia after PSNL (Smits et al., 2006; Truin et al., 2011). The latter may directly be related to a time-dependent differential role of GABA-mediated inhibition after PSNL. In the present study we therefore examined whether quantitative changes in dorsal horn intracellular GABA levels, the number of GABA-IR interneurons and KCC2 protein levels are gradually aggravated during the developmental course of NPP and whether these anatomically-determined GABAergic-related alterations are related to the degree of hypersensitivity.

EXPERIMENTAL PROCEDURES Animals Fifty adult male Sprague–Dawley rats (10-weeks-old; Charles River, Maastricht, The Netherlands) weighing 250 –300 g at the beginning of the study were used. All animal procedures were approved by the Animal Care Committee, University of Maastricht, The Netherlands and adhered to the guidelines of the National Institute of Health Guide for the Care and Use of Laboratory Animals. All efforts were made to minimize the number of animals used and their suffering. Animals were housed individually after surgery in a climate-controlled room with a 12-h light-dark cycle. Food and water were allowed ad libitum.

Neuropathic surgery All operations were carried out under sterile conditions. PSNL was performed as described earlier by Seltzer et al. (Seltzer et al., 1990). Hence, animals were anesthetized by inhalation of a mixture of isoflurane (5% induction, 3% maintenance) and air enriched with 100% oxygen at a constant flow rate of 250 ml/min. Body temperature was maintained at 37.5 °C using an automatic heating pad. After disinfection of the trimmed skin with iodine, a small incision was made in the left hind paw to expose the sciatic nerve which was freed from surrounding connective tissue under optical magnification. One third to one half of the diameter of the left sciatic nerve was then tightly ligated with 8/0 non-absorbable silk suture just a few millimeters proximal to the little fat pad, assuring that the unilateral partial ligation occurred at the monofascicular side of the sciatic nerve. Afterward, the wound was sutured externally with 4/0 non-absorbable silk suture and disinfected with iodine (Smits et al., 2006, 2009; Truin et al., 2007, 2009). In sham-ligated animals, the left sciatic nerve was only freed from surrounding connective tissue without performing a partial ligation.

Behavioral assessment Mechanical hypersensitivity of the hind paw was determined by testing withdrawal responses to tactile stimuli with von Frey monofilaments of logarithmically increasing stiffness (ranging from 0.16 to 100 g, North Coast Medical, Inc., Morgan Hill, CA, USA). The filaments were applied to the midplantar surface of both hind paws until slight bending of the filaments was obtained. Application of the filaments always occurred five times for each consecutive filament, starting with the lowest filament. A response was considered positive when three out of five stimuli induced a clear withdrawal (Cui et al., 1997; Smits et al., 2006). All testing procedures occurred by an observer (SJ) who was blinded for the implemented surgery. Animals were placed in elevated individual Plexiglas cages exposed on a wire mesh floor and were allowed to adapt 15 min before testing. Rats which responded positively to filaments ⬍60 g before surgery (pre-injury, day 0) were withdrawn from the study (n⫽3) while 100 g was used as a cut off value during the whole experiment. Non-allodynic animals with withdrawal thresholds ⱖ60 g at day 3 and day 16 post-operative were also excluded (n⫽4). Ipsi- and contralateral withdrawal thresholds of sham-ligated animals and PSNL-induced NPP animals, respectively, were compared both before surgery and 3 or 16 days post-injury. Ipsilateral withdrawal thresholds of PSNL animals were also compared to ipsilateral values of sham-ligated animals 3 and 16 days postinjury. Similar comparisons were performed for the contralateral withdrawal thresholds.

Experimental design and tissue processing After behavioral examination at 3 and 16 days post-PSNL, defined as early and late NPP respectively, animals were sacrificed for immunohistochemistry and immunoblotting procedures. For immunohistochemistry experiments, rats (3 days postinjury: n⫽12; 16 days post-injury: n⫽11) were perfused transcardially with a mixture of 4% paraformaldehyde and 15% picric acid in 0.2 M phosphate buffer (pH 7.6) after anesthesia with pentobarbital (180 mg/kg body weight) to perform anti-GABA immunohistochemical stainings and stereology-based cell counting of GABA-IR cell bodies in laminae 1–3 of the spinal dorsal horn. Lumbar spinal cord L4-L5 regions were removed by laminectomy, post-fixated overnight at 4 °C and consecutively cryoprotected in 10% and 25% sucrose at 4 °C. Tissues were subsequently embedded in Tissue-Tec glue®, frozen in liquid carbon dioxide and stored at ⫺80 °C. Ten-␮m thick transverse cryosections were mounted on gelatin-coated glass slides and stored at ⫺30 °C until staining procedures were performed.

S. P. Janssen et al. / Neuroscience 184 (2011) 183–194 For immunoblotting experiments, rats (n⫽10 for 3 and 16 days post-injury, respectively) were anesthetized with pentobarbital (180 mg/kg body weight) after which 1.5 cm of the lumbar spinal cord enlargement including L1-L6 regions was removed by hydroextrusion using ice cold saline in order to examine the levels of spinal dorsal horn KCC2 protein. Therefore, the spinal cord was freed from its spinal meninges after which it was divided into ipsilateral and contralateral halves and subsequently into dorsal and ventral quadrants. All tissue processing occurred on ice. Tissue quadrants were frozen in liquid nitrogen and stored at ⫺80 °C until use.

GABA immunohistochemistry All immunohistochemical procedure steps were performed at room temperature. Glass slides were air-dried for 2 h and consecutively washed for 10 min in Tris-buffered saline (TBS, 0.1 M, pH 7.6) including 0.3% Triton X-100 (TBS-T), TBS and TBS-T. Blocking was performed by a 1-h incubation period with 2% normal donkey serum (Sigma-Aldrich, Zwijndrecht, The Netherlands, D9663) diluted in TBS-T. A polyclonal rabbit anti-GABA antibody (1:500 diluted in TBS-T; Sigma-Aldrich, Zwijndrecht, The Netherlands, A2052) was used for overnight incubation. Excess of primary antibody was washed away with TBS during three times for 10 min. A secondary Alexa Fluor® 488 donkey anti-rabbit IgG antibody (1:100 diluted in TBS-T; Invitrogen, Breda, The Netherlands, A21206) was subsequently incubated for 2 h. After three times 10 min of washing in TBS, the sections were incubated for 30 min with a bis-benzimide H33342 trihydrochloride fluorochrom (Hoechst, 1:500 diluted in TBS; Sigma-Aldrich, Zwijndrecht, The Netherlands, B2261) to visualize cell nuclei. Finally, the glass slides were washed three times 10 min with TBS and coverslipped with TBS/glycerol (20%/80%). Both for the early and late phase of NPP, four sham-ligated rats served as controls. For examination of early NPP, eight allodynic animals were used, while examination of late NPP was based on seven allodynic animals.

Immunofluorescence microscopy Photomicrographs of ipsi- and contralateral lumbar L4-L5 spinal cord immunostained sections were taken using a Provis AX70 fluorescent microscope (Olympus, Hamburg, Germany) connected to a black and white digital video camera (U-CMAD-2, Olympus), equipped with CellP© software (Soft Imaging Systems, Münster, Germany). Per animal at least five ipsi- and contralateral spinal cord sections per lumbar L4 and L5 level, respectively, were photographed. All sections were photographed with a 4⫻ objective and a 12.5⫻ condenser in such a way that the entire dorsal and ventral horns together with the central canal were visualized in the picture. Microscope settings were maintained constant during photographing, with the exposure time of the camera set to values that prevent saturation.

Determination of dorsal horn Rexed laminae 1, 2, and 3 Schematic drawings of Rexed laminae 1, 2, and 3 of the ipsi- and contralateral L4 and L5 lumbar dorsal horns were carried out on the 4⫻ photomicrographs according to Paxinos and Watson (Paxinos and Watson, 1998; Fig. 1A). In detail, the circumference of the dorsal horn photomicrographs including the central canal was drawn on transparent slides. The central and lateral secants of lamina 1 with the dorsal horn circumference were marked according to Paxinos and Watson (Paxinos and Watson, 1998) and their distance to the central canal was measured. According to the table in Fig. 1B, an average correction factor was calculated for those two reference points and used for further drawings of laminae 1, 2 and 3. More precisely, a connection line was drawn

185

between the central and lateral secant of lamina 1, which was then used as the foundation for a perpendicular running through the middle of this connection line. A third marker for the medial part of lamina 1 could consequently be placed on the perpendicular after conversion of the medial part table measurement with the correction factor. The lateral, medial and central secants of laminae 2 and 3 with the dorsal horn circumference or perpendicular were subsequently marked after conversion of the measurement table with the correction factor. At the end, each lamina consisted of three marked reference points which were then connected following the curvature of the dorsal horn. The complete transparent drawing could finally be drawn on the 4⫻ photomicrographs to delineate laminae 1, 2 and 3 (Fig. 1C).

Anti-GABA immunoreactivity analysis Analysis of grey scale ipsi- and contralateral spinal cord pictures was performed with the AnalySIS software programm CellP© (Soft Imaging Systems, Münster, Germany). As most GABAergic inhibitory interneurons are known to reside in the spinal dorsal horn superficial layers (Magoul et al., 1987), laminae 1, 2 and 3 were defined as regions of interest (ROI), in which the average grey value of each individual lamina was calculated. Similar analysis was performed for laminae 1, 2 and 3 together. Combined L4 and L5 contralateral grey values of sham-ligated animals were expressed as 100% since the contralateral grey value might refer to values obtained in normal naive rats. Ipsi-and/or contralateral combined L4 and L5 grey values of shamligated animals and PSNL animals were consequently expressed as a percentage of the contralateral sham-ligated side.

Stereology-based GABA-IR cell counting An adapted stereology-based protocol was used to count the number of GABA-IR neuronal profiles as cells were only counted at one depth on the Z-axis. The same sections as those used to determine anti-GABA-IR based on dorsal horn grey values were utilized. The adapted stereology-based GABA-IR cell counting was performed with a stereology workstation consisting of a modified fluorescence microscope (Olympus BX50) and stereology software (StereoInvestigator, MBF Bioscience). Rexed laminae 1, 2 and 3 were delineated according to Paxinos and Watson (Paxinos and Watson, 1998) as described above under 4⫻ magnification on live microscopic video images displayed on a monitor. GABA-IR cells were counted in the different Rexed laminae at 40⫻ magnification using counting frames of 2500 ␮m2. GABA-IR cells were identified positive if the total cell cytosol was stained above background levels or if an intense fluorescent cell surface staining was observed (Dougherty et al., 2009). Hoechst staining was used to avoid positive counting of staining artifacts. The total number of GABA-IR cells was calculated individually for laminae 1, 2 and 3, but also for all laminae together. The number of ipsi- and contralateral GABA-IR cells of sham-ligated animals was compared to the ipsi- and contralateral side of allodynic animals, respectively.

KCC2 immunoblotting Tissue extracts of the spinal dorsal horn quadrants were prepared as described earlier by Coull et al. (Coull et al., 2003) with small modifications. Homogenates were first centrifuged at 3000 g, 4 °C for 20 min. The remaining supernatants were subsequently centrifuged at 10,000 g, 4 °C for 30 min. The resulting pellets, indicative of membrane-bound proteins, were used for further analyses. Total protein content was determined by Biorad DC protein assay (Bio-Rad Laboratories, Veenendaal, The Netherlands; # 500-0112). Equal amounts of protein (10 ␮g of total protein per well) were diluted in sample buffer, preheated at 37 °C for 30 min and separated by 10% SDS-PAGE electrophoresis. Transfer of

186

S. P. Janssen et al. / Neuroscience 184 (2011) 183–194

Fig. 1. Methodological method for the topological determination of Rexed laminae 1, 2 and 3. (A) Schematic drawings of L4 and L5 lumbar spinal cords adapted from Paxinos and Watson with modifications. (B) Table indicating the distances of the secants of laminae 1, 2 and 3 with the central, medial and lateral part of the dorsal horn circumference of L4 and L5 lumbar spinal cords. (C) After determination of the correction factor according to the magnification of the gray scale photomicrographs used, laminae 1, 2 and 3 were drawn on the photomicrographs. Scale bar⫽500 ␮m.

proteins occurred overnight at 30 V, 4 °C onto pure nitrocellulose membranes (Bio-Rad Laboratories, Veenendaal, The Netherlands; # 162-0115). All immunoblotting steps were performed at room temperature unless stated otherwise. After 10 min of washing in TBS with 0.1% Tween20 (TBS-tween), PonceauS staining was used to visualize transferred proteins. The blots were then cut into two pieces according to the different primary antibodies used: rabbit anti-KCC2 (1:2000, Millipore, Amsterdam, The Netherlands; # 07-432) and rabbit anti-␤-actin (1:1000, Cell Signaling, Leiden, The Netherlands; # 4967). The membranes for KCC2 detection were blocked for 30 min in 5% non-fat dry milk dissolved in TBS-tween, while those for ␤-actin were blocked for 1 h in 5% bovine serum albumin in TBS-tween. Primary antibody incubations occurred overnight at 4 °C in 5% non-fat dry milk dissolved in TBS-tween. After several washes in TBS-tween, the KCC2 membranes were incubated for 30 min with anti-rabbit IgG peroxidase conjugate (1:2000 in 5% non-fat dry milk, Sigma, Zwijndrecht, The Netherlands; A6154) while the membranes for ␤-actin were incubated for 1 h with anti-rabbit IgG HRP-linked (1:2000 in 5% non-fat dry milk, Cell Signaling, Leiden, The Netherlands; # 7074). Finally, the membranes were washed in TBS-tween and

chemiluminescent bands were detected using enhanced chemiluminescence technology (GE Healthcare, Eindhoven, The Netherlands; RPN2109). Images were captured onto high performance chemiluminescence films using different exposure times to avoid saturation of the images (GE Healthcare, Eindhoven, The Netherlands; 28906836). Densitometric quantification of immunoreactive bands was performed using ImageJ 1.43r software. All KCC2 protein concentrations were normalized to ␤-actin protein. Both for the early and late phase of NPP, dorsal horn quadrants of five sham-ligated animals and five PSNL animals were examined. Ipsi- and contralateral values of sham-ligated animals were compared to ipsi- and contralateral values of allodynic animals, respectively.

Statistical analyses Statistical analyses were carried out using Graphpad Prism 4.0 software. Data were tested for normality and student’s paired t-test, unpaired t-test or Mann–Whitney test were used wherever appropriate. Correlations were performed using Spearman or Pearson correlation test and linear regression analysis wherever appropriate. All data are represented as mean⫾standard error of the mean (SEM) or

S. P. Janssen et al. / Neuroscience 184 (2011) 183–194

187

Fig. 2. Developmental course of NPP-induced mechanical hypersensitivity. Changes in hind paw withdrawal thresholds to von Frey monofilaments after unilateral partial peripheral nerve ligation or sham-ligation at both early and late time points, respectively. (A) Three days post-injury, decreased withdrawal thresholds were observed at the ipsilateral hind paw of PSNL rats compared to the contralateral hind paw or the ipsilateral hind paw of sham-injured rats. Data are mean⫾SEM of nine sham-ligated animals and 13 PSNL animals. (B) The decreased withdrawal thresholds observed at the ipsilateral hind paw of PSNL rats 3 d post-injury remained decreased up to 16 d when compared to the contralateral hind paw or the ipsilateral hind paw of sham-injured rats. Data are mean⫾SEM of nine sham-ligated animals and 12 PSNL animals. *** P⬍0.001.

mean⫾95% confidence interval (CI). Asterisks in the figures indicate levels of significance (* P⬍0.05; ** P⬍0.01; *** P⬍0.001).

RESULTS Developmental course of PSNL-induced hypersensitivity Early signs of pain behaviour. Ipsi- and contralateral pre-injury withdrawal thresholds were similar in sham-operated and PSNL rats, respectively. Three days after induction of PSNL, withdrawal thresholds were significantly decreased at the ipsilateral hind paw compared to its contralateral side (10.0⫾1.7 g vs. 87.7⫾5.3 g; P⬍0.001) or compared to the ipsilateral side of sham-ligated animals (68.9⫾5.9 g; P⬍0.001). In contrast, contralateral withdrawal thresholds of PSNL rats did not change compared to the contralateral level of sham-injured rats. Moreover, ipsi- and contralateral hind paw withdrawal thresholds of sham-injured rats remained unaffected 3 days post-injury (Fig. 2A). All animals recovered well after surgery. No signs of autotomy were noted 3 days post-injury, although slight alterations in paw gesture were observed at the ipislateral hind paw. Long-lasting pain behaviour. Ipsi- and contralateral pre-injury withdrawal thresholds did not differ between sham-ligated and PSNL rats, respectively. Sixteen days after PSNL, withdrawal thresholds were significantly decreased at the ipsilateral hind paw compared to its contralateral side (10.8⫾2.5 g vs. 72.9⫾7.8 g; P⬍0.001) or compared to the ipsilateral side of sham-ligated animals (64.4⫾4.4 g; P⬍0.001). In contrast, contralateral withdrawal thresholds of PSNL rats were not changed compared to the contralateral value of sham-injured rats. Moreover, ipsi- and contralateral hind paw withdrawal thresholds of sham-injured rats remained similar 16 days post-injury (Fig. 2B). Analogously as for the early phase of NPP, signs of autotomy remained absent whereas altered paw gesture was still observed at the ipsilateral hind paw.

Early NPP PSNL reduces ipsilateral spinal dorsal horn intracellular GABA levels. GABA-IR was primarily observed in laminae 1–3 of the lumbar L4-L5 spinal cord dorsal horn, including a clear distinction between GABA-positive cell bodies and GABAergic neuropil (Fig. 3A). Three days after nerve injury, ipsi- and contralateral GABA-IR was similar at the L4-L5 spinal cord dorsal horn of sham-ligated rats in laminae 1, 2, 3 and 1–3 (Fig. 3B–E). Interestingly, PSNL rats showed a significantly decreased ipsilateral GABA-IR in laminae 1–3 compared to the ipsilateral side of sham-injured rats (90.5⫾6.2% vs. 103.1⫾10.9%; ⫺12%; P⬍0.05; Fig. 3E). This decrease was mainly due to reductions in GABA-IR in lamina 3 (P⬍0.05; Fig. 3D), as no significant decreases were observed in laminae 1 and 2 (P⬍0.07; Fig. 3B, C). Moreover, linear regression analysis did not reveal a significant correlation between the degree of NPP-induced hypersensitivity and GABA-IR in laminae 1, 2, 3 separately (data not shown) and laminae 1–3 at the ipsilateral side after nerve ligation (P⫽0.48, r2⫽0.09; Fig. 3F). Contralateral GABA-IR did not differ between sham-injured and PSNL animals in all three laminae examined (Fig. 3B–E). PSNL does not induce changes in the number of spinal dorsal horn GABA-IR cells. GABA-IR cell bodies were primarily observed in laminae 1–3 of the spinal dorsal horn (Fig. 3A). Three days after induction of PSNL, the number of GABA-IR cell bodies in laminae 1–3 did not differ between sham-injured animals and PSNL animals at the ipsi-(47.5⫾1.1 vs. 43.8⫾3.3) and contralateral (50.1⫾3.2 vs. 43.8⫾3.0) side, respectively (Fig. 4). Similar findings were observed when laminae 1, 2 and 3 were compared individually (data not shown). In line with this, the number of GABA-IR neuronal profiles in all laminae was demonstrated to be independent of the degree of mechanical hypersensitivity (data not shown). PSNL does not alter spinal dorsal horn KCC2 protein levels. Three days after induction of NPP, normalized KCC2 protein levels did not change when comparing the

188

S. P. Janssen et al. / Neuroscience 184 (2011) 183–194

Fig. 3. GABA-IR in the L4-L5 spinal dorsal horn during early NPP. (A) Representative photomicrographs of the ipsilateral dorsal horn, subdivided in laminae 1-2-3, of a sham-injured animal and a PSNL animal. Scale bar⫽100 ␮m. (B–E) GABA-IR in laminae 1, 2, 3 and 1–3, respectively. Significant decreases in the gray value mean ROI (expressed as % sham contralateral) were only observed at the ipsilateral dorsal horn in laminae 3 and 1–3 of PSNL animals compared to the ipsilateral dorsal horn of sham-injured animals. Data are mean⫾SEM of four sham-ligated animals and eight PSNL animals. (F) Correlation analysis between the severity of NPP (ipsi- and contralateral) and GABA-IR in laminae 1–3. Note that the severity of NPP is independent of the changes in GABA-IR. Data are mean⫾CI of eight PSNL animals. * P⬍0.05.

S. P. Janssen et al. / Neuroscience 184 (2011) 183–194

Fig. 4. Number of GABA-IR cells in the L4-L5 spinal dorsal horn during early NPP. Number of GABA-IR cells in laminae 1–3. No differences were observed between the number of GABA-IR cells in sham-injured animals compared to PSNL animals. Data are mean⫾SEM of four sham-ligated animals and eight PSNL animals.

ipsilateral dorsal horn of the lumbar enlargement of PSNL animals and sham-injured animals (1.2⫾0.3 vs. 0.9⫾0.2; Fig. 5B). Similar unchanged KCC2 protein contents were found at the contralateral side of the spinal dorsal horn after induction of NPP (1.0⫾0.2 vs. 1.0⫾0.1; Fig. 5A). Additionally, pain-related severity-dependent alterations in dorsal horn KCC2 expression were absent (data not shown).

189

PSNL does not induce changes in the number of spinal dorsal horn GABA-IR cells. Sixteen days after induction of PSNL, the number of GABA-IR cell bodies in laminae 1–3 did not differ between sham-injured animals and PSNL animals at the ipsi- (47.4⫾2.1 vs. 48.6⫾2.0) and contralateral (49.7⫾2.5 vs. 49.2⫾1.5) side, respectively (Fig. 7). Similar findings were observed when laminae 1, 2 and 3 were compared individually (data not shown). Concomitantly, no correlations were observed between the degree of NPP and the number of ipsilateral GABA-IR neuronal profiles in all dorsal horn laminae (data not shown). PSNL strongly decreases ipsilateral spinal dorsal horn KCC2 protein levels. Sixteen days after induction of NPP, normalized KCC2 protein levels significantly decreased by more than 50% when comparing the ipsilateral dorsal horn of the lumbar enlargement of PSNL animals and sham-injured animals (0.6⫾0.1 vs. 1.3⫾0.1; P⬍0.01; Fig. 8A/C). Noteworthy, the alterations in KCC2 levels were independent of the degree of mechanical hypersensitivity (data not shown). In contrast, unchanged KCC2 protein contents were observed at the contralateral side of the spinal dorsal horn after induction of NPP (0.9⫾0.1 vs. 1.1⫾0.1; Fig. 8B).

Late NPP PSNL increases spinal dorsal horn intracellular GABA levels bilaterally. Sixteen days post-injury, ipsi- and contralateral GABA-IR was similar at the L4-L5 spinal cord dorsal horn of sham-injured rats in laminae 1, 2, 3 and 1–3 (Fig. 6B–E). In contrast to the minor decreased ipsilateral GABA-IR during early NPP, PSNL rats showed a larger increased GABA-IR in laminae 1–3 of the ipsilateral dorsal horn compared to sham-injured rats (119.7⫾7.9% vs. 100.5⫾5.9%; ⫹19%; P⬍0.05; Fig. 6A/E). Interestingly, a similar increase in GABA-IR was observed at the contralateral side when PSNL rats were compared to sham-ligated rats (117.9⫾12.2% vs. 100.0⫾4.2%; ⫹18%; P⬍0.05; Fig. 6E). Both ipsi- and contralateral increases were mainly due to augmentations in the GABA-IR in laminae 1 and 2 (P⬍0.05; Fig. 6B, C), as no significant increases were observed in lamina 3 (Fig. 6D). Linear regression analysis revealed no differences in ipsilateral GABA-IR in relation to the severity of NPP-induced hypersensitivity in either lamina separately or in laminae 1–3 combined (Fig. 6F).

DISCUSSION In the present study we demonstrated differential alterations in the spinal dorsal horn GABAergic system and KCC2-related GABAA receptor-mediated inhibition during the early and late phase of PSNL-induced NPP. In detail, we noted a decrease in spinal GABA-IR at the ipsilateral side without alterations in the number of GABA-IR neuronal profiles per se during the early phase of NPP. At the same time, KCC2 protein concentrations remained unaltered. In contrast, we observed increases in spinal GABA-IR at both the ipsi- and contralateral side during the late phase of NPP. Additionally, NPP-induced changes in the total number of GABA-IR interneurons were absent whereas robust decreases in ipsilateral KCC2 protein levels were detected. With respect to the degree of PSNLinduced hypersensitivity, we observed no correlations with GABA-IR, the number of GABA-IR neuronal profiles as well as with KCC2 protein levels during the specified developmental course of NPP. The present anatomical/bio-

Fig. 5. Dorsal horn KCC2 protein concentrations of the lumbar enlargement during early NPP. (A) Normalized KCC2 protein concentrations in the contralateral dorsal horn. (B) Normalized KCC2 protein concentrations in the ipsilateral dorsal horn. Unchanged KCC2 protein levels were observed both in the ipsi- and contralateral dorsal horn of PSNL animals compared to sham-injured animals. Data are mean⫾SEM of five animals in each group.

190

S. P. Janssen et al. / Neuroscience 184 (2011) 183–194

Fig. 6. GABA-IR in the L4-L5 spinal dorsal horn during late NPP. (A) Representative photomicrographs of the ipsilateral dorsal horn of a sham-injured animal and a PSNL animal. Scale bar⫽100 ␮m. (B–E) GABA-IR in laminae 1, 2, 3 and 1–3, respectively. Significant increases in the gray value mean ROI (expressed as % sham contralateral) were observed at both the ipsilateral and contralateral dorsal horn in laminae 1, 2 and 1–3 of PSNL animals compared to the ipsilateral and contralateral dorsal horn of sham-injured animals, respectively. Data are mean⫾SEM of four sham-ligated animals and seven PSNL animals. (F) Correlation analysis between the severity of NPP (ipsi- and contralateral) and GABA-IR in laminae 1–3. Note that the severity of NPP is independent of the changes in GABA-IR. Data are mean⫾CI of seven PSNL animals. * P⬍0.05.

S. P. Janssen et al. / Neuroscience 184 (2011) 183–194

Fig. 7. Number of GABA-IR cells in the L4-L5 spinal dorsal horn during late NPP. Number of GABA-IR cells in laminae 1–3. No differences were observed between the number of GABA-IR cells in shaminjured animals compared to PSNL animals. Data are mean⫾SEM of four sham-ligated animals and seven PSNL animals.

chemical findings strongly suggest that NPP develops by a severity-independent decrease in spinal intracellular GABA levels switching to an increase in intracellular GABA levels and impaired GABAA receptor-mediated inhibition over time. The pivotal role of GABAergic inhibitory interneurons (Moore et al., 2002) located in spinal laminae 1–3 (Magoul et al., 1987) interfering with nociceptive activation of painsignaling neurons has already been pointed out decennia ago (Melzack and Wall, 1965). In line with this, pharmacological data have clearly confirmed that loss of endogenous GABAergic synaptic activity by blocking presynaptic

191

metabotropic GABAB receptors or postsynaptic ionotropic GABAA leads to tactile and thermal hypersensitivity (Yaksh, 1989; Malan et al., 2002) by enhancing excitatory neurotransmitter release from primary afferents or impairing hyperpolarization of spinal pain transmission neurons, respectively (Malcangio and Bowery, 1996). With respect to the GABAA receptor-mediated function, it is known that IPSCs are critically dependent on the activity of KCC2 which normally triggers a high intracellular Cl- balance. Recently, it has been shown that sciatic cuff-induced nerve injury downregulates ipsilateral KCC2 expression levels 2–3 weeks post-injury (Coull et al., 2003, 2005). The resulting shift in transmembrane Cl- gradient consequently rendered GABAergic input less inhibitory or even excitatory (Coull et al., 2003, 2005). Thus besides the available extracellular GABA levels and the concentrations of both GABAA and GABAB receptors, spinal GABAergic inhibition seems to be affected by the functional status of GABAA receptor-mediated transmission. In view of the former, NPP-induced alterations in GABAA and GABAB receptor protein expression are largely uncertain (Engle et al., 2006; Polgar and Todd, 2008) or might be a cellular response due to altered extracellular GABA levels (Moore et al., 2002). Nevertheless, the increased analgesic effect of GABAB receptor activation over GABAA receptor activation has been repeatedly demonstrated in pharmacological experiments (Cui et al., 1996; Hwang and Yaksh, 1997;

Fig. 8. Dorsal horn KCC2 protein concentrations of the lumbar enlargement during late NPP. (A) Immunoblot representing KCC2 protein concentrations in the ipsilateral dorsal horn of PSNL and sham-injured animals. (B) Normalized KCC2 protein concentrations in the contralateral dorsal horn. (C) Normalized KCC2 protein concentrations in the ipsilateral dorsal horn. Significantly decreased KCC2 protein levels were observed in the ipsilateral dorsal horn of PSNL animals compared to sham-injured animals. No differences were observed in the contralateral dorsal horn of PSNL animals compared to sham-injured animals. Data are mean⫾SEM of five animals in each group. ** P⬍0.01.

192

S. P. Janssen et al. / Neuroscience 184 (2011) 183–194

Malan et al., 2002) and might be due to the fact that GABAA receptor-mediated inhibition can be modulated by NPP-induced KCC2 down-regulation. In PSNL-induced NPP, measurement of tactile/mechanical hypersensitivity is known to result in a consistent and reproducible pain development (Wang and Wang, 2003) which already occurs within hours after ligation and may last for over several months (Seltzer et al., 1990; Yakhnitsa et al., 1999; Smits et al., 2006; Song et al., 2009) while bilateral patterns or mirror image hypersensitivity might appear occasionally (Seltzer et al., 1990). Our behavioral data are in line with the above-described findings as 3 and 16 days post-injury time points respectively reflect similar and long-lasting increases in tactile hypersensitivity at the ipsilateral side only. In the present study, the decreased ipsilateral intracellular GABA levels 3 days post-PSNL might point to a dysfunctional GABA production as it has already been shown that ipsilateral levels of the GABA synthesizing enzyme GAD65 are decreased early after chronic constriction injury (CCI) or spared nerve injury (SNI). In addition, no alterations in the ipsilateral levels of the other GAD enzyme, GAD67, were reported (Moore et al., 2002). Although contradictory at first sight, it might be permitted to preferentially link decreases in GAD65, and not GAD67, with decreased intracellular functional GABA levels as it has been argued that GAD67 is primarily cytosolic, localized to neuronal cell bodies and dendrites, and involved in non-vesicular GABA release whereas GAD65 is preferentially targeted to membranes and nerve endings where it may synthesize GABA for vesicular release (Soghomonian and Martin, 1998). In contrast to the present findings, non-reconciling increased total spinal GABA levels have also been demonstrated 3 days post-CCI (Satoh and Omote, 1996). However, the unaltered number of GABA-IR interneurons observed in the present study further strengthens the hypothesis of possible dysfunctional GABA production during early NPP. On the other hand, our findings are in contrast with minor bilateral decreases in GABA- or GAD67-IR cell profiles 3 days after induction of CCI and transient spinal cord ischemia (Zhang et al., 1994; Ibuki et al., 1997; Eaton et al., 1998). Interestingly, in these studies the number of GABA-IR interneurons reached its minimum 2 weeks after induction of CCI, gradually increasing to pre-injury levels within 2 months (Ibuki et al., 1997; Eaton et al., 1998) whereas in the spinal cord ischemia model the number of GABA-IR interneurons already reached its minimum 3 days post-injury, and returned to basal levels 2 weeks later (Zhang et al., 1994). Therefore, the above-described findings based on semi-quantitative analysis might depend on the experimental NPP model used but nevertheless suggest that decreased numbers of GABA-IR cell profiles are not likely at later stages of NPP. In the present study, spinal KCC2 levels were unaltered both at the ipsi- and contralateral side 3 days post-injury, thereby suggesting that only a decreased ipsilateral GABA production, as proposed earlier, without any disturbances in GABAA receptor-mediated inhibitory signaling itself accounts for the observed pain behavior at early time points

after PSNL. At first sight, this seems to be in contrast with earlier published data showing ipsilateral decreases in KCC2 protein levels already 4 h after CCI whereas 7 days later KCC2 levels were similar to pre-operative levels (Miletic and Miletic, 2008). Clearly the timetable of KCC2 alterations after CCI differs from that after PSNL and might reside in the fact that pain symptoms after CCI are not as long-lasting as those after PSNL (Bennett and Xie, 1988; Seltzer et al., 1990). In contrast to the early phase of NPP, we observed bilateral increases in intracellular spinal dorsal horn GABA levels 16 days after PSNL. With respect to the PSNL model in particular, one study showed decreased extracellular GABA levels at the ipsilateral spinal dorsal horn 2–3 weeks post-injury (Stiller et al., 1996). The ipsilateral increase in intracellular GABA as we have observed is likely to be in line with the above-mentioned decrease in ipsilateral extracellular GABA (Stiller et al., 1996). From this it is concluded that the late phase of PSNL-induced NPP is characterized by a dysfunctional GABA release at the ipsilateral side. In support of this, down-regulation of GAD65-67 enzymes has been demonstrated at the ipsilateral side 2 weeks after SNI or CCI (Moore et al., 2002), and might thus furthermore strengthen the proposed concept of dysfunctional GABA release instead of enhanced GABA production at the ipsilateral side during late NPP. The exact mechanism behind disturbed GABA release was however not examined here, but decreases in vesicular GABA transporter proteins are likely to be excluded (Polgar and Todd, 2008). The rationale behind the contralaterally increased intracellular GABA levels observed in the present study is difficult to interpret but is in line with long-lasting bilateral increases of total dorsal horn GABA after CCI (Satoh and Omote, 1996). However, as mirror image pain was not observed at the contralateral side in the present study, a dysfunctional contralateral GABA release might be excluded. The statement of an impaired ipsilateral GABA release during the late phase of NPP is moreover strengthened by the unaltered number of GABA-IR interneurons as observed in the present study. An unchanged number of GABA-IR interneurons was also observed 2 weeks after CCI or transient spinal cord ischemia (Zhang et al., 1994; Polgar et al., 2003) together with unaltered numbers of GABA inhibitory nerve boutons at GABAA-synapses 4 weeks after SNI (Polgar and Todd, 2008). On the other hand, conflicting evidence of bilaterally depleted GABA- or GAD67-IR cell profiles at the lumbar dorsal horn 2 weeks after neurectomy and CCI (CastroLopes et al., 1993; Ibuki et al., 1997; Eaton et al., 1998), and apoptosis of ipsilateral GAD67 mRNA-positive neurons 4 weeks after SNI (Scholz et al., 2005) have also been reported. However, both unaltered and decreased GABAergic cell numbers in combination with increased intracellular GABA levels as observed in the present study strongly indicate a dysfunctional ipsilateral GABA release. Furthermore, in the present study we observed very significant decreases in KCC2 protein levels at the ipsilateral side, whereas alterations were absent at the contralateral side. This is in line with previously discussed data (Coull et

S. P. Janssen et al. / Neuroscience 184 (2011) 183–194

al., 2003, 2005) and strongly suggests a decreased GABAA receptor-mediated inhibition, which might eventually result in GABAA receptor-mediated excitation, on top of the earlier mentioned dysfunctional GABA release at the ipsilateral side in the late phase of NPP whereas GABA released at the contralateral side likely exerts a normal GABAA receptor-mediated inhibitory functioning. The latter is in line with the absence of mirror image pain. Nevertheless, to ascribe decreases in ipsilateral KCC2 protein levels as highly important contributors to late PSNL-induced hypersensitivity, further research is necessary examining both the functionality of spinal GABAA receptors and the ratio of GABAA/GABAB receptor activation. So far, it is unknown whether the severity of NPP is determined by the degree of spinal GABAergic dysfunction. We hypothesized that both during the early as well as during the late phase of NPP hypersensitive animals would suffer from a more significant GABAergic dysfunction compared to non-hypersensitive animals. Literature data concerning (experimental) SCS-mediated pain relief might support the hypothesis of a severity-related (Li et al., 2006; Smits et al., 2006; van Eijs et al., 2010) GABAergic dysfunction as re-activation of the dysfunctional GABAergic system is suggested to occur during SCS-mediated analgesia in animals suffering from PSNL-induced NPP (Cui et al., 1996; Stiller et al., 1996; Cui et al., 1997). However in the present study, no correlation was detected between the GABA-IR, the number of GABA-IR cell profiles or KCC2 protein levels and the severity of NPP in the early phase, suggesting that the degree of pain behavior is not determined by possible differences in GABA production. Nevertheless, from the present data it cannot be excluded that GABA release itself might be affected in a severitydependent manner early during the development of NPP. Also during the late phase of NPP, we did not observe any differences in intracellular GABA levels, the number of GABA-IR cell profiles and KCC2 protein concentrations between animals suffering from different degrees of tactile hypersensitivity. These findings therefore exclude the involvement of a severity-dependent GABA release and GABAA receptor-mediated inhibition in the late phase of NPP. However, it should be stressed that the development of NPP is based on the balance between inhibitory and excitatory neurotransmitters (Latremoliere and Woolf, 2009), where the latter has been shown to play a pivotal role in determining the degree of hypersensitivity (Ultenius et al., 2006). In conclusion, our anatomical data strongly suggest that the impaired inhibitory role of GABA in the modulation of PSNL-induced NPP might be related to a dysfunctional GABA production mediated by GABAergic interneurons during early NPP whereas both a dysfunctional GABA release and an impaired GABAA receptor-mediated inhibition contribute to late NPP. As the GABAergic system is differentially deregulated during the developmental course of NPP, clinical therapies targeting GABA-mediated disinhibition should take these time-depending changes into account.

193

Acknowledgments—The authors report no conflict of interest. We sincerely thank Medtronic Europe Sarl for their financial research support without interfering with the study design, data collection and interpretation, report writing and submission of the article. Carolien Ceulemans and Sarina Gerard deserve gratitude for their technical support.

REFERENCES Bennett GJ, Xie YK (1988) A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man. Pain 33:87–107. Castro-Lopes JM, Tavares I, Coimbra A (1993) GABA decreases in the spinal cord dorsal horn after peripheral neurectomy. Brain Res 620:287–291. Castro-Lopes JM, Malcangio M, Pan BH, Bowery NG (1995) Complex changes of GABAA and GABAB receptor binding in the spinal cord dorsal horn following peripheral inflammation or neurectomy. Brain Res 679:289 –297. Coull JA, Boudreau D, Bachand K, Prescott SA, Nault F, Sik A, De Koninck P, De Koninck Y (2003) Trans-synaptic shift in anion gradient in spinal lamina I neurons as a mechanism of neuropathic pain. Nature 424:938 –942. Coull JA, Beggs S, Boudreau D, Boivin D, Tsuda M, Inoue K, Gravel C, Salter MW, De Koninck Y (2005) BDNF from microglia causes the shift in neuronal anion gradient underlying neuropathic pain. Nature 438:1017–1021. Cui JG, Linderoth B, Meyerson BA (1996) Effects of spinal cord stimulation on touch-evoked allodynia involve GABAergic mechanisms. An experimental study in the mononeuropathic rat. Pain 66:287–295. Cui JG, O’Connor WT, Ungerstedt U, Linderoth B, Meyerson BA (1997) Spinal cord stimulation attenuates augmented dorsal horn release of excitatory amino acids in mononeuropathy via a GABAergic mechanism. Pain 73:87–95. Dougherty KJ, Sawchuk MA, Hochman S (2009) Phenotypic diversity and expression of GABAergic inhibitory interneurons during postnatal development in lumbar spinal cord of glutamic acid decarboxylase 67-green fluorescent protein mice. Neuroscience 163: 909 –919. Eaton MJ, Plunkett JA, Karmally S, Martinez MA, Montanez K (1998) Changes in GAD- and GABA-immunoreactivity in the spinal dorsal horn after peripheral nerve injury and promotion of recovery by lumbar transplant of immortalized serotonergic precursors. J Chem Neuroanat 16:57–72. Engle MP, Gassman M, Sykes KT, Bettler B, Hammond DL (2006) Spinal nerve ligation does not alter the expression or function of GABA(B) receptors in spinal cord and dorsal root ganglia of the rat. Neuroscience 138:1277–1287. Hwang JH, Yaksh TL (1997) The effect of spinal GABA receptor agonists on tactile allodynia in a surgically-induced neuropathic pain model in the rat. Pain 70:15–22. Ibuki T, Hama AT, Wang XT, Pappas GD, Sagen J (1997) Loss of GABA-immunoreactivity in the spinal dorsal horn of rats with peripheral nerve injury and promotion of recovery by adrenal medullary grafts. Neuroscience 76:845– 858. Latremoliere A, Woolf CJ (2009) Central sensitization: a generator of pain hypersensitivity by central neural plasticity. J Pain 10:895– 926. Lee J, Back SK, Lim EJ, Cho GC, Kim MA, Kim HJ, Lee MH, Na HS (2010) Are spinal GABAergic elements related to the manifestation of neuropathic pain in rat? Korean J Physiol Pharmacol 14:59 – 69. Li D, Yang H, Meyerson BA, Linderoth B (2006) Response to spinal cord stimulation in variants of the spared nerve injury pain model. Neurosci Lett 400:115–120. Magoul R, Onteniente B, Geffard M, Calas A (1987) Anatomical distribution and ultrastructural organization of the GABAergic system

194

S. P. Janssen et al. / Neuroscience 184 (2011) 183–194

in the rat spinal cord. An immunocytochemical study using anti-GABA antibodies. Neuroscience 20:1001–1009. Malan TP, Mata HP, Porreca F (2002) Spinal GABA(A) and GABA(B) receptor pharmacology in a rat model of neuropathic pain. Anesthesiology 96:1161–1167. Malcangio M, Bowery NG (1996) GABA and its receptors in the spinal cord. Trends Pharmacol Sci 17:457– 462. Melzack R, Wall PD (1965) Pain mechanisms: a new theory. Science 150:971–979. Miletic G, Miletic V (2008) Loose ligation of the sciatic nerve is associated with TrkB receptor-dependent decreases in KCC2 protein levels in the ipsilateral spinal dorsal horn. Pain 137:532–539. Miletic G, Draganic P, Pankratz MT, Miletic V (2003) Muscimol prevents long-lasting potentiation of dorsal horn field potentials in rats with chronic constriction injury exhibiting decreased levels of the GABA transporter GAT-1. Pain 105:347–353. Moore KA, Kohno T, Karchewski LA, Scholz J, Baba H, Woolf CJ (2002) Partial peripheral nerve injury promotes a selective loss of GABAergic inhibition in the superficial dorsal horn of the spinal cord. J Neurosci 22:6724 – 6731. Paxinos G, Watson C (1998) The rat brain in stereotaxic coordinates, 4th ed. San Diego: Academic Press. Polgar E, Todd AJ (2008) Tactile allodynia can occur in the spared nerve injury model in the rat without selective loss of GABA or GABA(A) receptors from synapses in laminae I–II of the ipsilateral spinal dorsal horn. Neuroscience 156:193–202. Polgar E, Hughes DI, Riddell JS, Maxwell DJ, Puskar Z, Todd AJ (2003) Selective loss of spinal GABAergic or glycinergic neurons is not necessary for development of thermal hyperalgesia in the chronic constriction injury model of neuropathic pain. Pain 104: 229 –239. Satoh O, Omote K (1996) Roles of monoaminergic, glycinergic and GABAergic inhibitory systems in the spinal cord in rats with peripheral mononeuropathy. Brain Res 728:27–36. Scholz J, Broom DC, Youn DH, Mills CD, Kohno T, Suter MR, Moore KA, Decosterd I, Coggeshall RE, Woolf CJ (2005) Blocking caspase activity prevents transsynaptic neuronal apoptosis and the loss of inhibition in lamina II of the dorsal horn after peripheral nerve injury. J Neurosci 25:7317–7323. Seltzer Z, Dubner R, Shir Y (1990) A novel behavioral model of neuropathic pain disorders produced in rats by partial sciatic nerve injury. Pain 43:205–218. Smits H, Kleef MV, Honig W, Gerver J, Gobrecht P, Joosten EA (2009) Spinal cord stimulation induces c-Fos expression in the dorsal horn in rats with neuropathic pain after partial sciatic nerve injury. Neurosci Lett 450:70 –73. Smits H, Ultenius C, Deumens R, Koopmans GC, Honig WM, van Kleef M, Linderoth B, Joosten EA (2006) Effect of spinal cord stimulation in an animal model of neuropathic pain relates to degree of tactile “allodynia”. Neuroscience 143:541–546.

Soghomonian JJ, Martin DL (1998) Two isoforms of glutamate decarboxylase: why? Trends Pharmacol Sci 19:500 –505. Somers DL, Clemente FR (2002) Dorsal horn synaptosomal content of aspartate, glutamate, glycine and GABA are differentially altered following chronic constriction injury to the rat sciatic nerve. Neurosci Lett 323:171–174. Song Z, Ultenius C, Meyerson BA, Linderoth B (2009) Pain relief by spinal cord stimulation involves serotonergic mechanisms: an experimental study in a rat model of mononeuropathy. Pain 147: 241–248. Stiller CO, Cui JG, O’Connor WT, Brodin E, Meyerson BA, Linderoth B (1996) Release of gamma-aminobutyric acid in the dorsal horn and suppression of tactile allodynia by spinal cord stimulation in mononeuropathic rats. Neurosurgery 39:367–374; discussion 374 –365. Truin M, Smits H, Joosten EA, Van Kleef M (2007) Spinal cord stimulation in a mouse chronic neuropathic pain model. Neuromodulation 10:358 –362. Truin M, van Kleef M, Verboeket Y, Deumens R, Honig W, Joosten EA (2009) The effect of spinal cord stimulation in mice with chronic neuropathic pain after partial ligation of the sciatic nerve. Pain 145:312–318. Truin M, van Kleef M, Linderoth B, Smits H, Janssen SP, Joosten EA (2011) Increased efficacy of early spinal cord stimulation in an animal model of neuropathic pain. Eur J Pain 15:111–117. Ultenius C, Linderoth B, Meyerson BA, Wallin J (2006) Spinal NMDA receptor phosphorylation correlates with the presence of neuropathic signs following peripheral nerve injury in the rat. Neurosci Lett 399:85–90. van Eijs F, Smits H, Geurts JW, Kessels AG, Kemler MA, van Kleef M, Joosten EA, Faber CG (2010) Brush-evoked allodynia predicts outcome of spinal cord stimulation in complex regional pain syndrome type 1. Eur J Pain 14:164 –169. Wang LX, Wang ZJ (2003) Animal and cellular models of chronic pain. Adv Drug Deliv Rev 55:949 –965. Yakhnitsa V, Linderoth B, Meyerson BA (1999) Spinal cord stimulation attenuates dorsal horn neuronal hyperexcitability in a rat model of mononeuropathy. Pain 79:223–233. Yaksh TL (1989) Behavioral and autonomic correlates of the tactile evoked allodynia produced by spinal glycine inhibition: effects of modulatory receptor systems and excitatory amino acid antagonists. Pain 37:111–123. Yowtak J, Lee KY, Kim HY, Wang J, Kim HK, Chung K, Chung JM (2011) Reactive oxygen species contribute to neuropathic pain by reducing spinal GABA release. Pain 152:844 – 852. Zhang AL, Hao JX, Seiger A, Xu XJ, Wiesenfeld-Hallin Z, Grant G, Aldskogius H (1994) Decreased GABA immunoreactivity in spinal cord dorsal horn neurons after transient spinal cord ischemia in the rat. Brain Res 656:187–190.

(Accepted 26 March 2011) (Available online 7 April 2011)