Suppressed GABAergic signaling in the zona incerta causes neuropathic pain in a thoracic hemisection spinal cord injury rat model

Neuroscience Letters 632 (2016) 55–61

Contents lists available at ScienceDirect

Neuroscience Letters journal homepage: www.elsevier.com/locate/neulet

Research article

Suppressed GABAergic signaling in the zona incerta causes neuropathic pain in a thoracic hemisection spinal cord injury rat model Hyeong Cheol Moon a , Youn Joo Lee b , Chul Bum Cho c , Young Seok Park d,∗ a

Department of Medical Neuroscience and Neurosurgery, College of Medicine, Chungbuk National University, Cheongju, Republic of Korea Department of Radiology, Daejeon St. Mary’s Hospital, The Catholic University of Korea, Republic of Korea c Department of Neurosurgery, Saint Vincent’s Hospital, The Catholic University of Korea, Republic of Korea d Department of Neurosurgery, Chungbuk National University Hospital, Chungbuk National University, College of Medicine, Cheongju, Republic of Korea b

h i g h l i g h t s • To test whether GABAergic signals influence SCI-induced neuropathic pain. • The author recorded and compared in vivo single-unit, neuronal activity between hemisection-SCI and sham-treated rat models. • To test whether muscimol or bicuculline influence neuronal activity in hemisection SCI rat model.

a r t i c l e

i n f o

Article history: Received 11 May 2016 Received in revised form 19 August 2016 Accepted 20 August 2016 Available online 22 August 2016 Keywords: Spinal cord injury Neuropathic pain Zona incerta GABA Neural cell recording

a b s t r a c t Objective: Suppression of the gamma-aminobutyric acid (GABA)ergic activity of the zona incerta (ZI) reportedly plays a role in neuropathic pain after spinal cord injury (SCI). A reduction in GABAergic signaling in the ZI of a thoracic hemisection-SCI rat model has been suggested, but not clearly demonstrated. Accordingly, our objective was to investigate whether GABAergic signals influence SCI-induced neuropathic pain. Methods: In vivo, we recorded and compared single-unit, neuronal activity between hemisection-SCI and sham-operated rat models. Furthermore, we analyzed neuronal activity in both models following treatment with either a GABAA receptor agonist (muscimol) or antagonist (bicuculline). Results: Rats that underwent hemisection SCI exhibited reduced hindpaw withdrawal thresholds, latencies, and decreased ZI neuronal activity compared with sham-operated controls. Importantly, muscimol treatment increased, whereas bicuculline decreased, the firing rates of the ZI neurons. The muscimol treated, hemisection-SCI rats also exhibited increased hindpaw withdrawal thresholds and latencies. Conclusions: These data provide evidence that neuropathic pain after SCI is caused by decreased GABAergic signaling in the ZI. Furthermore, our data demonstrate that infusion of a GABAergic drug into the ZI could restore its inhibitory action and improve neuropathic pain behaviors. © 2016 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Neuropathic pain after spinal cord injury (SCI) is caused by dysfunctional pain signaling. The pain following SCI can be serious for plegic patients and is often resistant to conventional medical treat-

∗ Corresponding author at: Department of Neurosurgery, Chungbuk National University, College of Medicine, 410 SungBong-Ro Heungdeok-gu, Cheongju-si, Chungbuk, Republic of Korea. E-mail addresses: [email protected], [email protected] (Y.S. Park). http://dx.doi.org/10.1016/j.neulet.2016.08.035 0304-3940/© 2016 Elsevier Ireland Ltd. All rights reserved.

ment; however, the pathophysiology of neuropathic pain is poorly understood. The severity and duration of neuropathic pain symptoms are often greater than other types of chronic pain [37], with 5% of adults experiencing debilitation despite the administration of analgesics [31]. Refractory cases of neuropathic pain are clinically meaningful. However, the definition of refractory neuropathic pain remains unclear and such cases are likely underestimated, although some efforts have attempted to measure the epidemiology of refractory neuropathic pain [1,34]. Neuropathic pain is one of the most difficult and troublesome conditions to treat in clinical practice. Chronic neuropathic pain after SCI (SCI pain) often presents spontaneously as widespread

56

H.C. Moon et al. / Neuroscience Letters 632 (2016) 55–61

allodynia and hyperalgesia [5,15,16] that can develop weeks or even months after the initial injury [13,36]. Neuropathic pain can deteriorate the quality of life in patients with an SCI, but the mechanism by which SCI causes dysfunctional pain signaling remains unclear. The specific neuronal populations that are sensitized and associated with SCI-related neuropathic pain are not well known, but certain populations of thalamic nuclei are likely to be involved [22]. Using a rodent model, several researchers have demonstrated that activity in the gamma-aminobutyric acid (GABA)ergic nucleus [17,40,49] of the zona incerta (ZI) is suppressed after SCI [6,25,26,33,38]. As well, the GABAergic nucleus of the ZI suppresses responses in the posterior thalamus of the rodent [38]. Although SCI suppresses activity in the GABAergic nucleus of the ZI, it likewise causes concomitant increased activity in the posterior nucleus of the thalamus (PO), which is one of its primary targets. Further, increased PO activity is correlated with the maintenance and expression of hyperalgesia after SCI [46]. Neuropathic pain after SCI is associated with reduced neuronal activity in the GABAergic ZI, which has connections to the cerebrum, basal ganglia, and spinal cord. Because spinal cord gray matter is terminated in ZI afferents [25,46], we focused on the role of GABAergic neural signaling in the ZI. Although the mechanisms underlying neuropathic pain are not fully understood, it appears that an increase in glutamate and a decrease in GABA content play critical roles in excitotoxicity, cell death and inflammation, neuronal plasticity, and hyperstimulation [3,4,9,14,40]. The ZI plays an important role as the sensory relay neuronal network in neuropathic pain. Accordingly, we hypothesized that the decreased GABA activity in the ZI following SCI influences neuropathic pain; therefore, we monitored neuronal activity and tested pain behavior in a thoracic hemisection-SCI rat model. Moreover, to determine whether the SCI-induced alterations are dependent on GABA signaling, we infused the ZI of hemisection-SCI rats with the GABAA receptor agonist muscimol or the antagonist bicuculline and monitored the effects on ZI neuronal activity and pain behavior. 2. Methods 2.1. Animals All animal use protocols were approved by the Chungbuk National University Institutional Animal Care Committee. Strict aseptic surgical procedures were performed according to the established guidelines. Adult male Sprague-Dawley rats (250–300 g; Deahanbiolink, Eumseong, Korea) were maintained under standard housing conditions (12 h light/dark cycle) and provided laboratory food pellets and water ad libitum. A randomized, double-blind, controlled animal trial design was employed, and all animals were assigned to a treatment group on the same day. To this end, a cage was randomly selected from the pool of cages and primary randomization was performed by an individual other than the surgeon. The animals were removed from the cages following randomization and were assigned a permanent designation by the registrar. 2.2. Surgical preparation for spinal cord injury We used a rodent model of neuropathic pain in which the hemisected lesions of the spinal cord resulted in mechanical hyperalgesia below the lesion site [45]. The rats were transiently anesthetized with a combination of 15 mg/kg tiletamine/zolazepam (Zoletil50® ; Virbac Laboratories, Carros, France) and 9 mg/kg xylazine (Rompun® ; Bayer, Seoul, South Korea) in saline and mounted onto the surgical field. The anesthetized rats underwent a thoracic dorsal T9/T10 laminectomy using a surgical

microscope (Wesco SWF 10×/22, Illinois, USA) for visual guidance. The T9/T10 spinous processes were easily recognized as we have previously performed several laminectomies. The rats were randomly assigned to either the sham-operated group (n = 10) or the hemisection–SCI group (n = 10). The appropriate sample size (n = 10) for conducting a paired t-test between the two groups was determined using the Pass 14 software (NCSS Inc., Kaysville, UT, USA; alpha = 0.05, power = 0.8). The spinal cord hemisection procedures were conducted using microscissors according to the methods described by Christensen et al. [7]. Fleece-coated fibrin glue (TachoComb; Nycomed Austria GmbH, Linz, Austria) was used to seal the dura and to prevent adhesion to the thoracic cord. The wound was closed in anatomical layers, and the skin was secured using 3-0 nylon. The spinal cord hemisection procedure is illustrated in Fig. 1. Sham-operated animals (n = 10) were treated similarly, with the exception of the thoracic spinal cord injury. Post-surgical blinding was conducted, the surgeon completed the surgeries, and the animals were delivered to the registrar. After the surgical procedure, the animals were recovered and monitored in a temperature-controlled space. 2.3. Implantation of guide cannulas After SCI, a longitudinal incision was made along the midline of the skull to expose the bregma and lambda. For recording and the infusion of drugs, a microguide tube (Eicom, Kyoto, Japan) was fixed in place using two bone screws and acrylic resin (Ortho-jet, Lang Dental, USA). The coordinates for implantation were 3.5 mm anterior from the bregma, 2.8 mm lateral to the midline, and 6.8 mm below the surface of the brain (A: −3.5 mm, L: 2.8 mm, D: 6.8 mm), according to the atlas of Paxinos and Watson (Paxinos & Watson, 1986). Fig. 1 shows an illustration of the brain anatomy and the targets for guiding cannula implantation in the SCI rat model. 2.4. Behavioral confirmation of hyperalgesia: hind paw withdrawal threshold (PWT) Mechanical hyperalgesia was assessed by measuring paw withdrawal thresholds (PWT) following the application of an increasing pressure stimulus onto the plantar surface of the paw using an analgesia meter (Ugo Basile, Milan, Italy). The animals were placed in a Plexiglas cage (20 × 20 × 14 cm) with a grid bottom and were allowed to adapt for at least 30 min. Mechanical stimuli were generated by touching the plantar region of the contralateral and ipsilateral hind paws of the rats with a continually increasing pressure (5 g/s). 2.5. In vivo extracellular recordings After the SCI surgery, stainless steel electrodes (diameter, 139 ␮m; exposed tip, 75 ␮m; distance between electrodes, 280 ␮m) were implanted in 10 rats in the ZI contralateral to the spinal lesion. The electrodes were targeted using stereotaxic coordinates (A: −3.5 mm, L: 2.8 mm, D: 6.8 mm) according to Paxinos and Watson [30]. At least 14 days after the spinal lesion surgery, the rats were anesthetized with 15 mg/kg tiletamine/zolazepam (Zoletil50® ; Virbac Laboratories, Carros, France) and 9 mg/kg xylazine (Rompun® ; Bayer, Seoul, South Korea) to prepare for the extracellular recordings. Extracellular recordings were obtained from the ZI in the spinal lesion using a single electrode (A-M Systems, Sequim, USA). We obtained recordings from well-isolated units, digitized (40 kHz) the waveforms using a Digital Lynx SX (NeuraLynx Inc., Montana, USA) data-acquisition system, and sorted units off-line with NeuraLynx SpikeSort 3D using dual thresholds and principle component analyses. Using the SpikeSort 3D program, we identified similar neurons

H.C. Moon et al. / Neuroscience Letters 632 (2016) 55–61

57

Fig. 1. Illustration showing the position of the spinal cord injury and the implantation of the guide tube. A. Hemisection-spinal cord injury model. B. Microinjection guide tube inserted into the ZI and recording using an electrode.

based on waveform shapes and selected similar patterns of spike clusters. We compared and selected similar interspike interval histograms (ISI) using NeuroExplorer® (NeuraLynx Inc., Montana, USA). We obtained proper neuronal activity in six rats, whereas background noise obscured the measurements in the remaining four rats. 2.6. Histological verification of electrode and probe placement After the completion of the experiment, the locations of the electrode and probe sites were verified post-mortem. Transcardiac perfusion was performed with saline followed by 10% formalin in saline. The brains were extracted and fixed in 10% formalin for 1 day before sectioning. After sectioning, the dye spot was located in the 30-␮m serial coronal sections following neutral red staining and identified on the Paxinos and Watson atlas diagrams [30]. Coronal sections (40-␮m) were generated from the thalamic formation using a cryostat microtome (Microm, Germany), stained with crystal violet (Sigma, USA), and examined by light microscopy. 2.7. Statistical analysis We performed paired and unpaired t-tests using a statistical software program (SPSS, version 20.0, Chicago, IL). We used a paired t-test for the behavioral analysis and an unpaired t-test for the firing rates. One-way analysis of variance (ANOVA) was used to compare the neuronal activity between the GABA A agonist and antagonist. In all experiments, P < 0.05 was considered statistically significant. Data are shown as the mean ± SD for the behavioral test and the neural signal processing, and the graphs were prepared using GraphPad Prism (GraphPad Software, San Diego, CA, USA). 3. Results

mechanical thresholds of the hemisection-SCI rats were less compared with the sham-operated rats, decreasing from 34.86 ± 6.69 to 26.73 ± 5.45 g (mean ± SD) in the ipsilateral hindpaw of the hemisection-SCI rats (n = 10; Fig. 2A). In accord with several previous reports, all animals that underwent the SCI surgery exhibited a significant reduction in PWT following mechanical stimuli 14 days after surgery. PWLs were also lower in the hemisection-SCI rats compared to sham-operated rats, decreasing from 6.40 ± 1.64 to 5.23 ± 0.395 s in the ipsilateral paws of lesioned rats on day 14 following surgery (n = 10; Fig. 2B). Importantly, the PWTs and PWLs of the hemisection-SCI rats remained significantly lower than the shamoperated rats throughout the 14-day post-surgical period (Fig. 2).

3.2. Effects of GABA agonist/antagonist treatment on PWTs and PWLs To determine the GABAergic effects on pain control in the ZI, we treated the hemisection-SCI rats with the GABAA receptor agonist muscimol or with the receptor antagonist bicuculline. Following SCI, we implanted a guide cannula into the ZI and microinjected saline (10 ␮L), muscimol (0.5, 1, or 5 mmol/␮L, 10 ␮L), or bicuculline (0.5, 1, and 1.5 ␮mol/␮L, 10 ␮L) using a microdialysis probe. The infusion of muscimol increased the ipsilateral PWTs from 28.87 ± 6.87 to 37.22 ± 8.43 g, whereas infusion of bicuculline decreased PWTs from 30.26 ± 7.29 to 24.71 g. Similar results were obtained for the ipsilateral PWLs compared to vehicle. Moreover, the observed behavioral changes varied depending on the drug concentration. The middle (37.22 ± 8.36 g, PWTs) and high (37.21 ± 8.43 g, PWTs) concentrations of muscimol effectively altered pain behavior, but no changes in ipsilateral hyperalgesia were observed with the middle (27.42 ± 10.22 g, PWTs) or high concentration (25.99 ± 12.86 g, PWTs) of bicuculline despite apparent increases in contralateral hyperalgesia (Fig. 3).

3.1. Behavioral confirmation of neuropathic pain in a hemisection SCI rat model 3.3. Neuronal activity in the ZI following SCI To confirm the induction of neuropathic pain in our hemisection SCI rat model, we performed behavioral assays to assess PWTs and paw withdrawal latencies (PWLs). The hemisection-SCI rats exhibited decreased ipsilateral PWTs following mechanical stimulation of the plantar surface compared with the sham-operated rats, which indicated the development of hyperalgesia after SCI. The

To determine the overall effects of SCI on ZI neuronal output, we monitored the ZI neuronal activity in both the hemisection-SCI and sham-operated rats (n = 6). The average firing rate of ZI neurons in the hemisection-SCI rats (2.29 ± 2.13 Hz) was lower than the corresponding rate in sham-operated rats (7.81 ± 4.43 Hz; Fig. 4A).

58

H.C. Moon et al. / Neuroscience Letters 632 (2016) 55–61

Fig. 2. Alterations in paw withdrawal thresholds (PWTs) and paw withdrawal latencies (PWLs) at days 1, 7, 10, and 14 after spinal cord injury (SCI). The graphs illustrate the PWT and PWL analyses of behavioral responses according to force/weight (A) and time (B), respectively. All data points represent the mean ± SD. A paired t-test was used to assess statistical differences using GraphPad Prism (*p < 0.05). PWTs were assessed by a plantar test. Hindpaw mechanical withdrawal thresholds decreased over time and hyperalgesia developed unilaterally after SCI. Hemisection-SCI rats responded to mechanical stimuli applied to the plantar surface of the ipsilateral hindpaw at significantly shorter latencies than the sham-operated rats.

Fig. 3. The effects of the GABAA receptor agonist muscimol and the GABAA receptor antagonist bicuculline on neuropathic pain-associated mechanical hyperalgesia. Zona incerta (ZI) inactivation, via hemisection-SCI surgery in rats, caused hyperalgesia. However, paw withdrawal thresholds (PWTs) and paw withdrawal latencies (PWLs) in rats were changed following drug infusion through microdialysis cannulas implanted over the right ZI. The infusion of muscimol resulted in increased PWTs and PWLs (A, B). In contrast, the infusion of bicuculline resulted in increased hyperalgesia (C, D).

3.4. Bursting versus tonic firing in ZI neurons following muscimol and bicuculline treatment To analyze its differential roles, the neuronal activity of a single ZI neuron can be divided into burst and tonic firing rates. Thus, to analyze the effects of SCI and GABA modulation on burst and tonic firing rates, we treated hemisection-SCI rats with muscimol or bicuculline and measured burst and tonic firing rates during a 1 h period post-treatment. We defined a burst as a maximum 4 ms interval between spikes, a 100 ms interval between bursts, and a minimum of three spikes. We defined a tonic as a single spike with a regular firing rate. Tonic firing increased to 4.48 ± 1.08 Hz during the 20–40 min period following the injection of muscimol compared to 2.80 ± 0.32 Hz during the same period following the injection of bicuculline. Similarly, burst firing increased

to 0.34 ± 0.12 Hz 20–40 min after the injection of muscimol compared to 0.06 ± 0.01 Hz 20–40 min after the bicuculline injection. The injection of saline did not affect the tonic or burst firing rates over the course of 1 h. The changes in tonic and burst firing rates following GABA drug treatment are depicted in Fig. 4 and summarized in Table 1 mean ± SD). The analysis of the tonic firing rates revealed similarities to the overall firing rate, whereas the burst firing rates were different following GABA-agonist and −antagonist injections. Further, the firing rate was increased to a greater extent by treatment with muscimol compared to bicuculline during the 1 h period after injections.

H.C. Moon et al. / Neuroscience Letters 632 (2016) 55–61

59

Fig. 4. Spontaneous and evoked firing rates in the neurons of the zona incerta (ZI) of hemisection- spinal cord injury (SCI) rats (n = 6) and sham-operated rats (n = 6) (A). The injection of a low concentration GABA drug into the ZI changed the firing rate during a 1 h period (B). The spontaneous firing rate increased following GABAergic drug infusion throughout the monitoring periods. Recordings of tonic and burst action potentials were altered by treatment with a GABAergic drug and a GABA antagonist. The tonic firing rate (C) and bursting firing rate (D) increased after treatment with a GABAergic drug but not with a GABA antagonist. We recorded neuronal activity in the sham and SCI groups. A one-way ANOVA test was used to the compare neuronal activity among the vehicle and GABAA agonist and antagonist. Table 1 The mean firing rates in the zona incerta (ZI) neurons of a spinal cord injury (SCI) model (mean ± SD). Firing Rate(Hz)

Saline-vehicle injection

Bursts Baseline (−10–0 min) Portion I (0–20 min) Portion II (20–40 min) Portion III (40–60 min)

Spikes/s Bursts/s Spikes/s Bursts/s Spikes/s Bursts/s Spikes/s Bursts/s

Spikes/s Spikes/s Spikes/s Spikes/s

2.50 ± 1.67 0.20 ± 0.02 2.28 ± 1.64 0.21 ± 0.01 2.05 ± 1.62 0.08 ± 0.01 1.75 ± 1.59 0.04 ± 0.01

4. Discussion Here, we showed that neuronal activity in the ZI was greatly reduced in a hemisection-SCI rat model but that the reduced activity was recovered by treatment with a GABA-agonist. Moreover, the GABA-agonist treatment also ameliorated the observed reductions in PWTs and PWLs in the hemisection-SCI rats. Together, these data suggest that suppressed GABA neuronal activity in the ZI is the cause of neuropathic pain following SCI. Maladaptive synaptic circuits in the spinal dorsal horn induced by SCI, individually or synergistically, contribute to neuronal hyperexcitability in response to mechanical, chemical, and thermal stimuli [17]. Electrophysiologically, neuronal hyperexcitability or central sensitization is characterized by the enhanced spontaneous or evoked neuronal response properties to external stimuli applied to peripheral receptive fields with lowered activation thresholds, increased peripheral receptive field sizes, or that are increased after discharge activity [8]. Spinal hemisection results in the loss of endogenous GABAergic tone [19,32]. Further, this altered sensory perception to pathologic pain, which is linked to both spontaneous and evoked activities, is regulated by inhibitory input from the GABAergic nucleus ZI [29].

Tonic 2.30 ± 0.20 2.07 ± 0.39 1.97 ± 0.34 1.70 ± 0.29

GABAA receptor Agonist Injection

GABAA receptor Antagonist Injection

Bursts

Bursts

3.37 ± 2.10 0.01 ± 0.01 3.22 ± 3.03 0.20 ± 0.05 4.77 ± 2.44 0.34 ± 0.12 3.21 ± 2.01 0.21 ± 0.04

Tonic 3.35 ± 0.04 3.00 ± 0.46 4.48 ± 1.08 2.91 ± 0.31

3.29 ± 2.10 0.01 ± 0.01 3.26 ± 2.22 0.16 ± 0.03 2.86 ± 1.82 0.06 ± 0.01 2.80 ± 1.94 0.07 ± 0.01

Tonic 3.27 ± 0.07 3.10 ± 0.42 2.80 ± 0.32 2.73 ± 0.26

Progressive neuropathic pain after SCI is thought to be due to hyperexcitability [2], structural alteration in the synaptic circuitry [47], and degeneration of the inhibitory interneurons, including the loss of the GABA-mediated inhibition of neuronal activity that is associated with chronic pain hypersensitivity after SCI [18,21,23,44]. Spinal cord lesions typically cause unremitting pain that can be diffuse and bilateral and may extend to locations caudal to the spinal injury [10,35,48]. Animals with such spinal lesions typically develop mechanical hyperalgesia within 15 days of the spinal lesion surgery [25]. Unlike complete spinal cord transection injuries, the reproducibility of partial spinal cord laceration injuries is somewhat inconsistent [28]. Although these types of laceration injuries are not typically seen in a clinical setting, they can effectively disconnect both the ascending and descending axonal pathways at designated SCI levels [28]. In this study, we investigated the effects of peripheral noxious input after SCI on enhanced mechanical reactivity, a potential behavioral correlate of central sensitization. Thalamic relay cells project into the cerebral cortex and GABAergic interneurons produce local inhibition [22]. Further, SCI induces maladaptive synaptic transmission in the somatosensory system that results in chronic central neuropathic pain [17]. A general consequence of SCI is the development of severe, debilitating neu-

60

H.C. Moon et al. / Neuroscience Letters 632 (2016) 55–61

ropathic pain that is spontaneous and persistent in the absence of an insult, but can also present as hypersensitivity to painful stimuli (hyperalgesia) and hypersensitivity to normally innocuous stimuli (allodynia). Although the cause of SCI pain remains unknown, it has long been hypothesized that chronic pain results from abnormally suppressed inhibition in the thalamus. However, efforts to identify the mechanisms by which GABAergic transmission so powerfully modulates states of arousal and pain must account for the fact that the effects of GABAergic transmission vary by brain region [41,42]. Descending and local interneuron inhibitory pathways, including GABAergic pathways, are essential to the modulation of the balance between excitatory and inhibitory tone in synaptic transmissions. The GABAergic interneurons produce synaptic inhibition, thereby preventing or inhibiting central neuropathic pain, via both the GABAA and GABAB receptors in the spinal dorsal horn [17,20,27]. The spinothalamic tract, which transmits signals that are important for pain localization, and the spinoreticular tract, which is involved in the emotional aspect of pain, are the two main pain pathways. As well, the ZI might serve to integrate diverse sensory input and provide a link to appropriate arousal, attention, and visceral and posture/locomotion responses [24,26]. Among them, the ZI forms part of the “paralemniscal” somatosensory pathway, which is one of the four streams that travel in parallel from the periphery to the cerebral cortex. The ZI projects into a large number of thalamic nuclei [11]. Among them, the ventral division (ZIv) cells are GABAergic and their activity suppresses sensory responses in the medial subdivision of the posterior complex (POm), which serves as the thalamic relay of the paralemniscal pathway [11]. Further, the ZI exerts feed-forward and tonic inhibition on the thalamus [38,39]. Thus, neuropathic pain following SCI is thought to result from decreased neuronal activity in the ZI, which normally inhibits select thalamic nuclei. The loss of ZI activity likewise correlates with a pathological increase in neuronal activity in the posterior thalamus (PO), a somatosensory thalamic nucleus critical for processing nociceptive information [25]. The ZI is a GABAergic nucleus located in the diencephalon and is part of the ventral thalamus. Previous studies have described the connections between ZI neurons, the majority of which are GABAergic, and the spinal cord [26,33]. Recently, Masri et al. described the incertothalamic system, a novel network that is involved in regulating nociceptive processing in the thalamus [25,46]. However, enhanced activity of the thalamic neurons is caused by increased nociceptive input from upstream structures, including sensitized dorsal horn neurons and the somatosensory cortex [46]. Further, the sensitization of thalamic neurons not only reflects a disinhibitory mechanism, but is also related to the upregulation of chemokines, microglia, and unique sodium channels [12,50]. Nonetheless, other “modulatory input,” such as cholinergic or serotonergic afferents, which also affect neuropathic pain in the ZI, cannot be excluded [43]. Our study expands the current understanding of SCI-induced neuropathic pain by demonstrating evidence for the influence of GABAergic signaling on ZI neuronal activity and pain behavior. Furthermore, we demonstrated that the GABAA receptor agonist muscimol increased neuronal activity in the ZI in hemisection-SCI rats, whereas the GABAA antagonist bicuculline did not. These data suggest that treatment with a GABAergic drug could restore the inhibitory action of the ZI, which in turn could decrease thalamic neuronal activity and ameliorate neuropathic pain. However, there were several caveats in this study that should be noted. First, we monitored dose-dependent behavior and neural signaling changes between sham and lesion animals following the administration of muscimol or bicuculline, but we did not monitor changes between the groups following the administration of vehicle alone. Second,

we did not analyze behavioral changes during GABA drug infusion because the insertion of a chronic infusion system into the ZI could alone affect pain behavior. In addition, because of the relatively short half-life of the GABA modulators, it is difficult to maintain a uniform concentration during a pain behavior test. Third, whether or not SCI causes a decrease in GABAA receptor number or the suppression of activity in the GABAergic nucleus in the ZI has not yet been determined. Although we postulated that activity was suppressed, we did not calculate the number of GABAA receptors in the ZI. Likewise, we did not perform GABA staining or calculate the number of GABAergic neurons in our hemisection-SCI or sham-operated rat models; therefore, we cannot report whether neuropathic pain after SCI is due to GABAergic neuronal loss or decreased GABA neuronal activity. Fourth, microscopy-guided thoracic hemisection can be categorized as pure hemisection, suboptimal hemisection, or over-hemisection in anatomical situations. Thus, the possibility exists that suboptimal hemisection occurred in this study. Further, we did not make a morphological correlation in this study. Our study revealed burst firing rates of 0.2–0.4/s in the ZI, but ZI was monitored under anesthesia, which will only reflect spontaneous activity. In rats, the ZI has fewer cells than the thalamus; consequently, is more difficult to record neural cells. Finally, the hypothesis that bursting produces abnormal pain transmission has some inherent limitations. Nonetheless, our observation that neuronal activity from the ZI increased after drug infusion in the hemisection-SCI rat model should shed light on the mechanisms underlying neuropathic pain following SCI. 5. Conclusions In this study, we demonstrated that neuropathic pain after SCI is caused by decreased GABA neuronal activity in the ZI. The spontaneous firing rates of ZI neurons were decreased in the hemisection-SCI rat model compared with sham controls. As well, treatment with the GABAA receptor agonist muscimol increased neuronal activity in the ZI, increased tonic firing rates, and improved pain behaviors, which suggests that GABAergic drugs could be a treatment option for restoring ZI inhibitory action following SCI and for improving symptoms associated with SCIinduced neuropathic pain. Conflict of interest statement The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Acknowledgement This work was supported by grants from the National Research Foundation of Korea (NRF 2015R1C1A2A01053318, NRF 2014K1A3A1A21001372, and NRF 2014R1A1A1A05007768). References [1] Relieving Pain in America A Blueprint for Transforming Prevention, Care, Education, and Research, Washington (DC), 2011. [2] S. Balasubramanyan, P.L. Stemkowski, M.J. Stebbing, P.A. Smith, Sciatic chronic constriction injury produces cell-type-specific changes in the electrophysiological properties of rat substantia gelatinosa neurons, J. Neurophysiol. 96 (2006) 579–590. [3] S.K. Bareiss, M. Gwaltney, K. Hernandez, T. Lee, K.L. Brewer, Excitotoxic spinal cord injury induced dysesthesias are associated with enhanced intrinsic growth of sensory neurons, Neurosci. Lett. 542 (2013) 113–117. [4] R. Baron, A. Binder, G. Wasner, Neuropathic pain: diagnosis, pathophysiological mechanisms, and treatment, Lancet Neurol. 9 (2010) 807–819. [5] D. Bowsher, Central pain: clinical and physiological characteristics, J. Neurol. Neurosurg. Psychiatry 61 (1996) 62–69.

H.C. Moon et al. / Neuroscience Letters 632 (2016) 55–61 [6] M. Cha, Y. Ji, R. Masri, Motor cortex stimulation activates the incertothalamic pathway in an animal model of spinal cord injury, J. Pain 14 (2013) 260–269. [7] M.D. Christensen, A.W. Everhart, J.T. Pickelman, C.E. Hulsebosch, Mechanical and thermal allodynia in chronic central pain following spinal cord injury, Pain 68 (1996) 97–107. [8] M. Costigan, A. Moss, A. Latremoliere, C. Johnston, M. Verma-Gandhu, T.A. Herbert, L. Barrett, G.J. Brenner, D. Vardeh, C.J. Woolf, M. Fitzgerald, T-cell infiltration and signaling in the adult dorsal spinal cord is a major contributor to neuropathic pain-like hypersensitivity, J. Neurosci. 29 (2009) 14415–14422. [9] R. D’Angelo, A. Morreale, V. Donadio, S. Boriani, N. Maraldi, G. Plazzi, R. Liguori, Neuropathic pain following spinal cord injury: what we know about mechanisms, assessment and management, Eur. Rev. Med. Pharmacol. Sci. 17 (2013) 3257–3261. [10] R. Defrin, A. Ohry, N. Blumen, G. Urca, Pain following spinal cord injury, Spinal Cord 40 (2002) 96–97, author rely 98–99. [11] M.E. Diamond, E. Ahissar, When outgoing and incoming signals meet: new insights from the zona incerta, Neuron 56 (2007) 578–579. [12] S.D. Dib-Hajj, T.R. Cummins, J.A. Black, S.G. Waxman, Sodium channels in normal and pathological pain, Annu. Rev. Neurosci. 33 (2010) 325–347. [13] S. Falci, L. Best, R. Bayles, D. Lammertse, C. Starnes, Dorsal root entry zone microcoagulation for spinal cord injury-related central pain: operative intramedullary electrophysiological guidance and clinical outcome, J. Neurosurg. 97 (2002) 193–200. [14] N.B. Finnerup, C. Baastrup, Spinal cord injury pain: mechanisms and management, Curr. Pain Headache Rep. 16 (2012) 207–216. [15] N.B. Finnerup, I.L. Johannesen, A. Fuglsang-Frederiksen, F.W. Bach, T.S. Jensen, Sensory function in spinal cord injury patients with and without central pain, Brain 126 (2003) 57–70. [16] L. Garcia-Larrea, P. Convers, M. Magnin, N. Andre-Obadia, R. Peyron, B. Laurent, F. Mauguiere, Laser-evoked potential abnormalities in central pain patients: the influence of spontaneous and provoked pain, Brain 125 (2002) 2766–2781. [17] Y.S. Gwak, C.E. Hulsebosch, GABA and central neuropathic pain following spinal cord injury, Neuropharmacology 60 (2011) 799–808. [18] Y.S. Gwak, J. Kang, G.C. Unabia, C.E. Hulsebosch, Spatial and temporal activation of spinal glial cells: role of gliopathy in central neuropathic pain following spinal cord injury in rats, Exp. Neurol. 234 (2012) 362–372. [19] Y.S. Gwak, H.Y. Tan, T.S. Nam, K.S. Paik, C.E. Hulsebosch, J.W. Leem, Activation of spinal GABA receptors attenuates chronic central neuropathic pain after spinal cord injury, J. Neurotrauma 23 (2006) 1111–1124. [20] J.H. Hwang, T.L. Yaksh, The effect of spinal GABA receptor agonists on tactile allodynia in a surgically-induced neuropathic pain model in the rat, Pain 70 (1997) 15–22. [21] K. Kahle, J.-F. Schmouth, V. Lavastre, A. Latremoliere, J. Zhang, N. Andrews, T. Omura, J. Laganière, D. Rochefort, P. Hince, G. Castonguay, R. Gaudet, J.C.S. Mapplebeck, S. Sotocinal, J. Duan, C. Ward, A. Khanna, J. Mogil, P. Dion, C. Woolf, P. Inquimbert, G. Rouleau, Inhibition of the kinase WNK1/HSN2 ameliorates neuropathic pain by restoring GABA inhibition, Sci. Signal. 9 (2016) ra32. [22] B. Kumar, J. Kalita, G. Kumar, U.K. Misra, Central poststroke pain: a review of pathophysiology and treatment, Anesth. Analg. 108 (2009) 1645–1657. [23] A. Latremoliere, C.J. Woolf, Central sensitization: a generator of pain hypersensitivity by central neural plasticity, J. Pain 10 (2009) 895–926. [24] R.C. Lin, M.A. Nicolelis, J.K. Chapin, Topographic and laminar organizations of the incertocortical pathway in rats, Neuroscience 81 (1997) 641–651. [25] R. Masri, R.L. Quiton, J.M. Lucas, P.D. Murray, S.M. Thompson, A. Keller, Zona incerta: a role in central pain, J. Neurophysiol. 102 (2009) 181–191. [26] J. Mitrofanis, Some certainty for the zone of uncertainty? Exploring the function of the zona incerta, Neuroscience 130 (2005) 1–15. [27] G. Munro, J.A. Lopez-Garcia, I. Rivera-Arconada, H.K. Erichsen, E.O. Nielsen, J.S. Larsen, P.K. Ahring, N.R. Mirza, Comparison of the novel subtype-selective GABAA receptor-positive allosteric modulator NS11394 [3’-[5-(1-hydroxy-1methyl-ethyl)-benzoimidazol-1-yl]-biphenyl-2-carbonitrile] with diazepam, zolpidem, bretazenil, and gaboxadol in rat models of inflammatory and neuropathic pain, J. Pharmacol. Exp. Ther. 327 (2008) 969–981. [28] S.M. Onifer, A.G. Rabchevsky, S.W. Scheff, Rat models of traumatic spinal cord injury to assess motor recovery, ILAR J. 48 (2007) 385–395. [29] A. Park, K. Hoffman, A. Keller, Roles of GABAA and GABAB receptors in regulating thalamic activity by the zona incerta: a computational study, J. Neurophysiol. 112 (2014) 2580–2596.

61

[30] G. Paxinos, C. Watson, The Rat Brain in Stereotaxic Coodinates Adademic, Sydny, 1986. [31] E.A. Pereira, T.Z. Aziz, Neuropathic pain and deep brain stimulation, Neurotherapeutics 11 (2016) 496–507. [32] E. Rausell, C.G. Cusick, E. Taub, E.G. Jones, Chronic deafferentation in monkeys differentially affects nociceptive and nonnociceptive pathways distinguished by specific calcium-binding proteins and down-regulates gamma-aminobutyric acid type A receptors at thalamic levels, Proc. Natl. Acad. Sci. U. S. A. 89 (1992) 2571–2575. [33] F.Z. Shaw, Y.F. Liao, R.F. Chen, Y.H. Huang, R.C. Lin, The zona incerta modulates spontaneous spike-wave discharges in the rat, J. Neurophysiol. 109 (2013) 2505–2516. [34] B.H. Smith, N. Torrance, J.A. Ferguson, M.I. Bennett, M.G. Serpell, K.M. Dunn, Towards a definition of refractory neuropathic pain for epidemiological research. An international Delphi survey of experts, BMC Neurol. 12 (2012) 29. [35] S. Stormer, H.J. Gerner, W. Gruninger, K. Metzmacher, S. Follinger, C. Wienke, W. Aldinger, N. Walker, M. Zimmermann, V. Paeslack, Chronic pain/dysaesthesiae in spinal cord injury patients: results of a multicentre study, Spinal Cord 35 (1997) 446–455. [36] R.R. Tasker, G.T. DeCarvalho, E.J. Dolan, Intractable pain of spinal cord origin: clinical features and implications for surgery, J. Neurosurg. 77 (1992) 373–378. [37] N. Torrance, B.H. Smith, M.I. Bennett, A.J. Lee, The epidemiology of chronic pain of predominantly neuropathic origin. Results from a general population survey, J. Pain 7 (2006) 281–289. [38] J.C. Trageser, K.A. Burke, R. Masri, Y. Li, L. Sellers, A. Keller, State-dependent gating of sensory inputs by zona incerta, J. Neurophysiol. 96 (2006) 1456–1463. [39] J.C. Trageser, A. Keller, Reducing the uncertainty: gating of peripheral inputs by zona incerta, J. Neurosci. 24 (2004) 8911–8915. [40] C. Ultenius, Z. Song, P. Lin, B.A. Meyerson, B. Linderoth, Spinal GABAergic mechanisms in the effects of spinal cord stimulation in a rodent model of neuropathic pain: is GABA synthesis involved? Neuromodulation 16 (2013) 114–120. [41] G. Vanini, R. Lydic, H.A. Baghdoyan, GABA-to-ACh ratio in basal forebrain and cerebral cortex varies significantly during sleep, Sleep 35 (2016) 1325–1334. [42] G. Vanini, K. Nemanis, H.A. Baghdoyan, R. Lydic, GABAergic transmission in rat pontine reticular formation regulates the induction phase of anesthesia and modulates hyperalgesia caused by sleep deprivation, Eur. J. Neurosci. 40 (2016) 2264–2273. [43] C. Varela, S.M. Sherman, Differences in response to muscarinic activation between first and higher order thalamic relays, J. Neurophysiol. 98 (2007) 3538–3547. [44] L.P. Vera-Portocarrero, J.Y. Xie, J. Kowal, M.H. Ossipov, T. King, F. Porreca, Descending facilitation from the rostral ventromedial medulla maintains visceral pain in rats with experimental pancreatitis, Gastroenterology 130 (2006) 2155–2164. [45] A.A. Webb, G.D. Muir, Compensatory locomotor adjustments of rats with cervical or thoracic spinal cord hemisections, J. Neurotrauma 19 (2002) 239–256. [46] J.L. Whitt, R. Masri, N.S. Pulimood, A. Keller, Pathological activity in mediodorsal thalamus of rats with spinal cord injury pain, J. Neurosci. 33 (2013) 3915–3926. [47] C.J. Woolf, P. Shortland, R.E. Coggeshall, Peripheral nerve injury triggers central sprouting of myelinated afferents, Nature 355 (1992) 75–78. [48] R.P. Yezierski, Pain following spinal cord injury: pathophysiology and central mechanisms, Prog. Brain Res. 129 (2000) 429–449. [49] A.L. Zhang, J.X. Hao, A. Seiger, X.J. Xu, Z. Wiesenfeld-Hallin, G. Grant, H. Aldskogius, Decreased GABA immunoreactivity in spinal cord dorsal horn neurons after transient spinal cord ischemia in the rat, Brain Res. 656 (1994) 187–190. [50] P. Zhao, S.G. Waxman, B.C. Hains, Modulation of thalamic nociceptive processing after spinal cord injury through remote activation of thalamic microglia by cysteine chemokine ligand 21, J. Neurosci. 27 (2007) 8893–8902.