Peripheral nerve injury can induce a state of chronic pain characterized by allodynia, pain sensation evoked by nonnoxious stimuli, applied to the affected limb. In some patients with chronic pain in the affected limb, altered pain sensation appears in contralateral noninjured regions, where the areas are identical to the ipsilateral chronic pain region. The pain is called mirror image pain. The mechanisms of mirror image pain have been examined by using animal models of chronic pain.20,26 However, despite recent recognition of a role of cortical plasticity in chronic pain symptoms, there have been few studies to date investigating cortical mechanisms of mirror image pain.
The primary somatosensory cortex (S1) plays a role in coding of location and intensity of pain,3 and we and others have reported increased neuronal activity in the S1 contralateral to the nerve injury24 and inflammation.16 In humans with chronic pain, functional magnetic resonance imaging and positron emission tomography reveal increases in activity-dependent blood flow in the S1 contralateral to the injured body side.27,35,36 Nerve injury or inflammation in rodent models induced increases in activity of both excitatory and inhibitory neurons,17 as well as increases in astrocyte activity and spine turnover.24,25 Indeed, the transiently increased astrocyte activity triggered the formation of new spines, which potentially form part of an abhorrent neural circuit that mediates aspects of chronic allodynia.25 Whether changes in the structure of synapses, or in cell activity, also occur in the ipsilateral S1 that may correspond to mirror pain are unclear.
In addition to astrocytes, inhibitory neuronal function in the S1 has critical roles in chronic pain. Inhibitory neurons regulate cortical excitatory neuronal activity and prevent exaggerated excitatory activity.44 Our previous report suggests that although inhibitory neuronal activity increases under the inflammatory chronic pain condition, excitatory neuronal activity increases due to the reduction of efficacy of γ-aminobutyric acid (GABA) inhibition.17 Moreover, a recent report demonstrates that the activity of somatostatin (SOM) neurons, which is one of the inhibitory neurons, decreased in the S1 of mice model of chronic pain and chronic selective activation of SOM neurons just after nerve injury can prevent induction of chronic pain behavior.10 Thus, inhibitory neuronal function in the S1 is supposed to contribute to induction and maintenance of mirror image pain. Therefore in this study, we examined cellular changes in the ipsilateral cortex after peripheral nerve injury, using 2-photon imaging, and correlated these changes with mechanical sensitivity of the paw of the noninjured limb. As we and others reported previously,10,24 nerve injury increased activity in the contralateral S1 cortex that was associated with chronic allodynia. In this study, we found that activities of astrocytes and local interneurons were also increased in the ipsilateral cortex, and this arose from transcallosal inputs. When S1 inhibition was pharmacologically blocked, we observed an increase in spine formation and decreased mechanical thresholds in the uninjured paw—mirror image pain—was revealed. Our data demonstrate that enhanced astrocytic activity in the S1 cortex ipsilateral to nerve injury primes circuits for remodeling and induces potential mirror image pain that is suppressed by local inhibition.
All animal experiments were approved by the Animal Research Committee of the National Institutes of Natural Sciences. Every effort was made to minimize the number and suffering of animals.
For clarity, we define the following terminology used. The nerve injury (ligation) was applied to the right-side sciatic nerve, defined as the injured or partial sciatic ligation (PSL) hind paw. The left hind limb is defined as the uninjured or intact hind paw. The left S1 cortex is contralateral to the injury and referred to as the contralateral (or contra-) S1. The right S1 cortex is ipsilateral to the injury and referred to as the ipsilateral (or ipsi-) S1. Diagrams in some of the figures indicate the cortical hemisphere used for imaging and drug application.
Adult male C57BL/6 mice (postnatal 8-12 weeks old), C57BL/6 mice expressing enhanced green fluorescent protein (EGFP) under the control of the thy-1 promoter in a small subset of cortical neurons (M-line mice),18 inositol trisphosphate receptor type 2 (IP3R2) knockout mice raised on a C57BL/6 background,19 and C57BL/6 mice expressing Venus fluorescent protein under the control of the vesicular GABA transporter (VGAT) promoter in inhibitory neurons of the central nervous system (VGAT-Venus mice)39 were used. All mice were housed in cages (5 mice per cage) with ad libitum access to food and water. The room was maintained on a 12-hour light/12-hour dark cycle.
2.2. Partial sciatic nerve ligation and behavioral testing
Under isoflurane anesthesia (1.5%-2.0%) or ketamine (0.13 mg/g, intraperitoneally [i.p.]) and xylazine (0.01 mg/g, i.p.), the right sciatic nerve of mice was exposed, and one-third to one-half the diameter of the nerve was ligated with 9-0 suture (PSL-injured mice). As for the control (sham-operated mice), the sciatic nerve similarly was exposed but did not receive any ligation. For assessing tactile allodynia, the mice were habituated for 30 minutes in a transparent box with a mesh floor, and 6 calibrated von Frey hairs (0.16, 0.4, 0.6, 1.0, 1.4, and 2.0 g bending force) were pressed perpendicularly against the plantar surface of the hind paw until slight buckling. The experiment started by testing the response to the 1.0 g filament. If mice show a paw withdrawal response, the filament with the next lower force was subsequently tested. If mice showed no withdrawal response to 1.0 g, the filament with the next higher force threshold was used. Behavior tests were finished when 4 stimuli were applied after the first positive response and were conducted during the daytime. To conduct behavior experiments with long-term drug application to the S1 through Elvax (noninflammatory drug delivery system, EV40W; DuPont-Mitsui Polychemicals) (see section 2.9), drug-soaked Elvax was applied to the brain surface on the ipsilateral S1 at the same time, 14 days, or 28 days after PSL surgery. These behavioral tests were performed during the daytime.
2.3. Repeated in vivo imaging of dendritic spines
For repeated in vivo dendritic spine imaging, M-line mice were deeply anesthetized with ketamine (0.13 mg/g, i.p.) and xylazine (0.01 mg/g, i.p.), and the head surface was cleaned of hair and skin to enable a custom-made thin metal imaging frame to be secured to the skull using acrylic glue and dental acrylic cement (Quick Resin; Shofu, Kyoto, Japan). The frame had a central hole giving access to the skull and allowed the mice to be secured under the microscope during subsequent imaging. A craniotomy was performed above the ipsilateral S1 and this small region (diameter <2-3 mm) of the brain was covered with a glass coverslip (diameter = 4.0 mm, thickness 0.12 mm, Matsunami Glass, Osaka, Japan). Dental acrylic cement and adhesive glue secured the coverslip to the surrounding skull and metal frame. In vivo imaging of dendrites was performed using a 2-photon microscope (Nikon A1R MP+, Tokyo, Japan) with a water-immersion objective lens (×25, numerical aperture [NA] 1.10; Nikon, Tokyo, Japan). To repeatedly obtain images of the same dendritic spine, the imaging area was identified by the blood vessel patterns at the brain surface obtained during the first imaging session, which could be readily found and aligned in subsequent imaging sessions. Because only a small number of neurons express EGFP in the cortex of M-line mice, the same dendrites can be identified with certainty. A Ti:sapphire laser (Mai Tai HP; Spectra-Physics, Santa Clara, CA) was tuned to 950 nm, the wavelength for 2-photon excitation of EGFP. Dendrites (512 × 512 pixels, 0.08-0.16 μm/pixel, 1.0-0.5 μm z-step) were imaged at depth within 100 μm from the brain surface (layer 1).
2.4. Analysis of dendritic spines
Metamorph (Molecular Devices, San Jose, CA) and Image J (http://rsbweb.nih.gov/ij/) were used to analyze individual spines in the same dendrites from 3-dimensional image stacks. Only spines that clearly protruded from the dendrite (>0.4 μm laterally) were included in the analysis. Spine length was measured from the dendrites to the tip. As described previously,24,25 all dendritic spines that were such identified were included in the analysis, regardless of morphological category. The spine gain and loss rates were defined as the number of the newly appeared spines (Ngain) and the number of spines that disappeared (Nlost) from before to after drug application (SR95531, SR95531, and fluoroacetate, dimethyl sulfoxide [DMSO]) in PSL or sham mice. These rates were calculated relative to the total number of spines before drug application (Ntotal) by the following equations24:
Total spine turnover rate was calculated using the following equation24:
2.5. Dye loading and in vivo calcium imaging of astrocytes and neurons
The dye solution for in vivo imaging contained the calcium indicator, Oregon Green 488 BAPTA-1 AM (OGB-1AM, 50 μg) dissolved in a small volume of Pluronic F-127 (10 μL, P3000PM; Invitrogen, Carlsbad, CA) and mixed with a solution of artificial cerebrospinal fluid (ACSF) (34 μL), and 1 mM SR101 (6 μL), an astrocyte marker.16,17,28 The final concentrations of OGB-1 AM and SR101 were 0.8 mM and 0.12 mM, respectively. The composition of ACSF is as followed: 125 mM NaCl, 5 mM KCl, 10 mM glucose, 10 mM HEPES, 2 mM CaCl2, and 2 mM MgSO4 (pH 7.4).
In vivo calcium imaging of S1 astrocytes and neurons was performed through a cranial window under continual visual guidance with a 2-photon microscope.25 Mice 3 to 6 days after PSL or sham surgery were anesthetized with urethane (1.7 g/kg body weight i.p.) and atropine (0.4 mg/kg body weight i.p.), and a craniotomy was performed above the ipsilateral S1 as described above with slight modification. Just before sealing the brain with a coverslip, a glass pipette (tip diameter 5-10 μm) was filled with the dye solution and inserted into the S1 to a depth of 50 to 100 μm from the brain surface. The dye solution was slowly pressure ejected (about 30 kPa, 60-80 seconds) using a microinjector (IM 300; Narishige, Tokyo, Japan). Calcium imaging was performed in anesthetized mice beginning straight after dye injection by using a 2-photon microscope (Nikon A1R MP+ or Olympus FV1000MPE laser scanning microscope, Tokyo, Japan) with an excitation wavelength of 800 nm for OGB-1 AM and SR101, and 950 nm for Venus. Laser power (mW) measured after the objective lens (×25, NA 1.10, Nikon or ×60, NA 0.9; Olympus) was <10 mW. Fluorescence was imaged for 10 minutes to quantify spontaneous astrocytic activity (0.5 frame/seconds, 1.2-1.9 μm/pixels, 512 × 512 pixel) and layer 1 inhibitory neuronal activity (4 frame/seconds, 1.2-1.9 μm/pixels, 512 × 256 pixel).
To identify inhibitory GABAergic neurons, we used VGAT-Venus mice. Before dye injection, we initially scanned a series of Z-stack images to identify the extent and location of Venus-positive neurons. After calcium imaging, we aligned these Z-stack images with the OGB-1 AM responses to identify activity in layer 1 inhibitory neurons. Astrocyte activity was quantified in cells with colocalised SR101 and OGB-1 AM fluorescence.
To visualize evoked activity in layer 2/3 neurons, the SR101 and OGB-1 AM dye solution was similarly pressure injected, but at a depth of about 250 μm from the brain surface while visualizing fluorescence under the 2-photon microscope. Calcium transients were recorded in response to electrical stimulation (20 V, 1 second, SEN-7203; Nihon Kohden, Tokyo, Japan) to the intact hind paw. Responses were imaged for a total of 15 seconds; 5 seconds before stimulation and for 10 seconds after stimulus. A hind paw was stimulated 10 times at each imaged z-plane (4 frame/seconds, 1.2-1.9 μm/pixels, 512 × 256 pixel).
2.6. Virus injection and in vivo calcium imaging of cortical neurons with GCaMP6f
Spontaneous activity in layer 2/3 neurons was imaged using GCaMP6f fluorescence. Mice were anesthetized with i.p. injections of ketamine (0.13 mg/g) and xylazine (0.01 mg/g), and a craniotomy was performed above the hind paw area of the ipsilateral or contralateral S1 (0.5 mm posterior, 1.5 mm lateral to the bregma). After craniotomy, the adeno-associated virus encoding the synapsin promoter-driven calcium indicator protein GCaMP6f (AAV9-Syn-GCaMP6f-WPRE-SV40; Penn Vector Core, Philadelphia, PA) was pressure injected at 250 to 300 μm from the cortical surface. After virus injection, the cranial window was sealed with a glass coverslip (diameter 1.8-4.0 mm, thickness 0.12 mm, Matsunami Glass, Osaka, Japan), and the coverslip was fixed by dental cement and adhesive glue (Aron Alpha, Konishi Co., Osaka, Japan). Calcium imaging was conducted 3 to 4 weeks after surgery using the 2-photon microscope (Nikon A1R MP+ laser) with a ×25, NA 1.10, Nikon objective lens. For imaging, mice were anesthetized with isoflurane (0.9%-1.0%) and layer 2/3 areas expressing GCaMP6f were imaged for 3 minutes to quantify spontaneous neuronal activity (4 frame/seconds, 1.2-1.9 μm/pixels, 512 × 256 pixel). Partial sciatic nerve ligation injury or sham surgery was conducted after the first calcium imaging session, and the same region was imaged 3 days after PSL or sham. The same area was identified from blood vessel patterns and from the location of GCaMP6f-positive cells as landmarks.
2.7. Analysis of calcium transients in neurons and astrocytes
We defined the outlines of cells as regions of interest based on cell morphology from averages Venus (inhibitory neurons), SR101 (astrocytes), and OGB-1 AM or GCaMP6f (layer 2/3 neurons) fluorescent signals using Olympus Fluoview software or Nikon NIS-Elements software. For the analysis of astrocytic calcium (OGB-1 AM) signals, the baseline fluorescence was defined as the lower two-thirds of the total fluorescence intensity histogram obtained over the 10-minute imaging period (F0).38 The amplitude of fluorescent transients was calculated as (ΔF = F − F0), where F is the instantaneous fluorescent signal. An astrocytic calcium transient was defined when ΔF/F0 went above 2.5 SDs of the baseline for at least 6 seconds. For layer 1 inhibitory neuron calcium transients, the F0 baseline was similarly defined in Venus-positive neurons, and a transient was detected when ΔF/F0 was more than 2 SD above the baseline value. The frequency of astrocytic or inhibitory neuron calcium transients was calculated from the number of spontaneous transients detected during the 10-minute imaging period. The amplitude of both astrocytic and inhibitory neuronal calcium transients was calculated as the peak ΔF/F0. To analyze the calcium transients of layer 2/3 neurons that are loaded with OGB-1 AM and evoked by electrical stimulation of the hind paw, the baseline intensity F0 was obtained by averaging the intensity values during the prestimulus period (5 seconds). The amplitude of OGB-1 AM signals was calculated as ΔF/F0 (ΔF = F − F0). Cells were considered responsive if the fluorescence intensity change (ΔF/F0) was >2 SD of the baseline value. The probability of response to stimulation of the hind paw was calculated as the proportion of stimuli that elicited a response, with a total of 10 stimuli applied. To analyze spontaneous calcium transients in layer 2/3 neurons expressing GCaMP6f, the baseline intensity F0 was obtained by averaging the GCaMP6 fluorescence intensity values across the total 3-minute imaging session. A spontaneous calcium transient was detected when the fluorescent intensity (ΔF/F0, ΔF = F − F0) was above 2 SD of the baseline fluorescence. The frequency of calcium transients was calculated as the number of calcium transients during the 3-minute observation and expressed as per second. The amplitude of these calcium transients was calculated as ΔF/F0.
Mice were deeply anesthetized with ketamine (0.13 mg/g, i.p.) and xylazine (0.01 mg/g, i.p.) and secured in a stereotactic frame. The skull was exposed, and a small piece of the skull along the sagittal suture was excised using a drill (0.5 mm rostral from the bregma, 2.0 mm caudal from the bregma, 1.0 mm both side from sagittal suture). The exposed brain surface was kept moist by application of ACSF. Direct insertion of a blade into this sagittal suture cavity causes bleeding, presumably from the superior sagittal sinus. To avoid this, the blade was inserted into the left hemisphere (0.5 mm lateral from the sagittal suture) into the secondary motor cortex, and a shallow (0.5 mm from the brain surface) cut was made across the cortex. After this initial cut, the razor blade was angled towards the sagittal suture, and a deeper cut (2.5 mm from the brain surface) was made to dissect the corpus callosum. After this callosotomy, the skin was sutured and the mouse allowed to recover for 7 days before behavior experiments and in vivo 2-photon calcium imaging. To visualize the cutting path and the transected axons, the lipophilic fluorescent dye (DiI; Sigma-Aldrich, St. Louis, MO) was injected into layer 1 to 2/3 under ketamine and xylazine anesthesia. Three to 7 days after this fluorescent dye injection, deeply anesthetized mice were perfused with 4% paraformaldehyde (PFA) and postfixed overnight in 4% PFA. Coronal cortical slices (50 μm) were cut with a vibratome (VT1000S; Leica, Wetzlar, Germany), and sections were mounted with VECTASHIELD mounting medium (H-100; Funakoshi, Tokyo, Japan). Slices were imaged under a fluorescence microscope (BX63, Olympus).
2.9. Acute drug application
Under urethane (1.7 g/kg body weight i.p.) and atropine (0.4 mg/kg body weight i.p.) anesthesia, 2 cranial windows were made over the bilateral somatosensory cortex hind paw area. For imaging activity in the ipsilateral S1, a bolus of the OGB-1 AM and SR101 dye solution was injected as described above. Drugs were pressure injected as required (IM 300; Narashige, Tokyo, Japan). Specifically, the sodium channel blocker tetrodotoxin (TTX, 5 μM, 1 μL) or vehicle (ACSF, 1 μL) was injected into the contralateral S1. As for application of ACSF containing SR95531 (10 μM) and fluoroacetate (astrocytic Krebs cycle inhibitor; 2.6 mM),2 or SR95531 alone, a cranial window was made over the ipsilateral S1 and drugs were applied on the surface of the ipsilateral S1 to block local GABAergic and astrocyte activity, respectively. After drug and dye application, the cranial window was sealed with suitably sized cover glass by dental cement and adhesive glue.
2.10. Preparation of Elvax drug solutions and long-term drug application
Elvax is a noninflammatory polymer used for slow release of drugs in tissues, including for sustained brain delivery.22 Elvax beads (100 mg) were dissolved in 1 mL of dichloromethane and mixed with 10 μL of DMSO, or 10 μL solution containing either 100 μM SR95531 or 100 μM SR95531 and 26 mM fluoroacetate. The mixture was stirred with a vortex mixer for 1 hour after which the solution was plated on a glass dish, frozen quickly, kept at –80°C for 1 hour, and then placed at –20°C overnight to allow the dichloromethane to evaporate. Small pieces of Elvax–drug matrix were cut (2 mm, approximately 0.5 mm2) just before use. Final concentrations of the drugs in Elvax were approximately 10 μM (SR95531) or 2.6 mM (fluoroacetate).
2.11. Evaluation of the effect of long-term drug application on spines and calcium responses in cortical cells
For evaluation of the effects of drugs on spine turnover rate or neuronal activity, a cranial window was made over the ipsilateral S1, as described above, with the exception that the glass coverslip was secured with KWIK-SIL (World Precision Instruments, Sarasota, FL), which could be more easily removed during subsequent steps. Predrug control images were obtained, and the glass coverslip sealed with KWIK-SIL was removed to allow for application of Elvax. A small piece of drug-soaked Elvax was gently placed on top of the dura, as we had performed previously.22,25 The margin of the cranial window, which was not covered with Elvax, was sealed with KWIK-SIL, and then, the incised scalp was sutured. Three to 7 days after Elvax application, the scalp was incised again, and the Elvax and KWIK-SIL gently removed. The cranial window was then sealed with a coverslip. Dendritic spine imaging was performed 3 days after the start of Elvax application, neuronal and astrocyte calcium imaging was performed from 5 to 7 days after Elvax.
Immunohistochemistry was performed to validate the Elvax-mediated sustained drug delivery system. In these experiments, Elvax was loaded with the fluorescent dye, Fluorescein (FITC; Thermo Fisher Scientific, Waltham, MA, 5 μΜ) using the same approach as described above. The FITC-loaded Elvax was applied to the ipsilateral S1 for 3 days. Mice were anesthetized with ketamine/xylazine and perfused with 4% PFA. Brains were excised and incubated with 4% PFA for 1 day. Then, brains were cut with a vibratome (VT1000S; Leica) at 50-μm thickness. Brain slices were counterstained with DAPI diluted in 0.1 M phosphate buffer (Hp7.2) (2 μg/mL) for 5 minutes. After washing in 0.1 M phosphate buffer, sections were mounted in PermaFluor (Thermo Fisher Scientific). Slices were subsequently imaged under a fluorescence microscope (BX51; Olympus).
2.13. Statistical analyses
Data are presented as mean ± SEM. Statistical tests were performed using a paired t test, unpaired t test (2 variables), 1-way analysis of variance (ANOVA), followed by a Dunnett or Bonferroni multiple comparisons test or 2-way ANOVA, followed by the Bonferroni test (≥3 variables) as indicated in the text. Behavioral data were analyzed using 2-way repeated-measure ANOVA, followed by the Dunnett or Bonferroni multiple comparisons test. In all cases, a P value of less than 0.05 was considered as statistically significant.
3.1. Inhibitory neuronal activity in the ipsilateral S1 is increased after partial sciatic nerve ligation
Peripheral nerve injury induced by PSL resulted in prolonged mechanical allodynia in response to stimulation of the injured hind paw, as we24,25 and others23 have previously reported. A dramatic decrease in the mechanical threshold for paw withdrawal was seen from 3 days after injury, and this persisted for more than 27 days (Fig. 1A). By contrast, the response to stimulation of the intact or uninjured hind paw was not different after PSL (Fig. 1A). Hence, mirror image pain is not typically manifest in mice with this PSL nerve injury, as also reported for the spared sciatic nerve injury model in mice.10 By contrast, mirror image pain is present in the rat spinal root ligation model4,6 and the rat carrageenan or complete Freund adjuvant–induced paw inflammation models.7,8,20
One reason for the absence of mirror image pain in this mouse model is that it is suppressed by cortical inhibition. Hence, we conducted in vivo calcium imaging in the ipsilateral S1 (layer 1) using OGB-1 AM fluorescence, whereas GABAergic neurons were identified by using a transgenic mouse in which Venus fluorescence was driven by a GABA neuron-specific promoter (VGAT). The astrocyte fluorescent marker, SR101, was also used in these experiments, and inhibitory neurons were defined as SR101-negative (Fig. 1B). Three to 6 days after PSL, the frequency of spontaneous calcium transients in these inhibitory neurons was increased as compared to sham-operated mice (sham, 0.028 ± 0.004/seconds, 79 cells in 5 mice; PSL, 0.054 ± 0.008/seconds, 89 cells in 5 mice; P = 0.008) (Figs. 1C and D). The amplitude of these calcium transients did not change (sham, 0.180 ± 0.007, 79 cells in 5 mice; PSL, 0.180 ± 0.002, 89 cells in 5 mice; P = 0.97) (Fig. 1E). Hence, PSL causes an increase in activity of inhibitory neurons in the ipsilateral S1, as we have also reported for the contralateral S1.16
The activity of excitatory neurons in the contralateral S1 is known to increase after PSL, and we proposed that increased in the activity of transcallosal fibers projecting to the ipsilateral S1 may excite local inhibitory neurons.30 To test this, we silenced the contralateral S1 with local application of a sodium channel blocker TTX, 5 μM (Fig. 1F). Application of TTX (contralaterally) reduced the frequency of spontaneous calcium transients in inhibitory neurons in the ipsilateral S1 (0.022 ± 0.006/seconds, 66 cells in 6 mice), and this was a significantly lower frequency as compared to that seen after local ACSF application to the contralateral S1 (0.061 ± 0.011/seconds) (Figs. 1G and H). The amplitude of these calcium transients did not change with TTX application (PSL + ACSF; 0.16 ± 0.005, PSL + TTX; 0.15 ± 0.006) (Figs. 1G and I). Therefore, transcallosal inputs appear to induce an increase in the activity of ipsilateral S1 inhibitory neurons after PSL.
3.2. Excitatory neuronal activity in the ipsilateral S1 is decreased after partial sciatic nerve ligation
To further examine changes in ipsilateral S1 cortical activity after PSL, we imaged calcium transients in layer 2/3 excitatory neurons. In the contralateral S1, the activity of both layer 2/3 and layer 5 excitatory neurons is known to increase after peripheral nerve injury.10,16 To image spontaneous activity of layer 2/3 neurons, we injected adeno-associated virus encoding the synapsin promoter-driven calcium-sensitive protein GCaMP6f (AAV-Syn-GCaMP6f), into layer 2/3 of the contralateral or ipsilateral S15 (Figs. 2A and B). We imaged, in separate experiments, both the activity of contralateral and ipsilateral layer 2/3 neurons 3 to 4 weeks after AAV injections, both before, and 3 days after PSL (Figs. 2A and B, see also experimental protocol). The frequency of spontaneous neuronal calcium transients in layer 2/3 of the contralateral S1 was increased after PSL as expected (before PSL; 0.044 ± 0.002/seconds [270 cells in 5 mice], after PSL; 0.068 ± 0.003/seconds [268 cells in 5 mice]; P < 0.0001) (Figs. 2C and D). The mean amplitude of these calcium transients did not change (before PSL; 0.28 ± 0.007 [278 cells in 5 mice], after PSL; 0.29 ± 0.009 [248 cells in 5 mice]; P = 0.189) (Figs. 2C and E). By contrast, the frequency of spontaneous calcium transients in layer 2/3 neurons in the ipsilateral S1 actually decreased after PSL. The frequency of calcium transients was 0.063 ± 0.003/seconds before PSL (469 cells in 12 mice), and 0.050 ± 0.002/seconds (409 cells in 12 mice) after PSL (P = 0.0002) (Figs. 2C and F). The amplitude of spontaneous calcium transients in layer 2/3 neurons in the ipsilateral S1 also decreased after PSL, from 0.32 ± 0.01 before PSL (n = 469 cells from 12 mice), to 0.25 ± 0.01 (n = 409 cells from 12 mice) after PSL (P < 0.0001) (Fig. 2G). No differences in either the frequency or amplitude of calcium transients were seen in sham-operated mice, in either the contralateral or ipsilateral S1 (Figs. 2D–G). This result suggests that the enhanced activity of layer 1 inhibitory neurons in the ipsilateral S1 after PSL resulted in a reduced basal level of excitatory neuron activity in the S1, thereby suppressing the expression of any mirror image pain behavior.
3.3. Astrocyte activity in the ipsilateral S1 is increased after partial sciatic nerve ligation
Astrocyte activity in layer 1 of the contralateral S1 cortex increases after nerve injury,25 and this contributes to the induction of chronic allodynia in response to mechanical stimulation of the injured hind paw. To examine whether nerve injury induces any changes in astrocytes in the ipsilateral S1 (layer 1), we imaged calcium transients in cells that were loaded with OGB-1 AM and were positive for fluorescence of the astrocyte marker SR101 (Fig. 3A). The frequency of spontaneous astrocytic calcium responses in the ipsilateral S1 hind paw area was significantly increased 3 days after PSL (Figs. 3B and C), as compared to that in the ipsilateral S1 hind paw area of sham-operated mice, or as compared to other ipsilateral S1 areas (the barrel cortex) after PSL. The frequency of astrocyte calcium transients was: sham 0.44 ± 0.03/10 minutes (465 cells in 6 mice); PSL 3 days, 1.76 ± 0.07/10 minutes (365 cells in 5 mice, P < 0.0001); and PSL 3 days (the barrel cortex), 0.59 ± 0.04/10 minutes (437 cells in 5 mice, P > 0.05) (Fig. 3C). The increase in frequency of astrocytic calcium responses was still significantly increased at 14 days after PSL (1.18 ± 0.07/10 minutes, P < 0.0001, 313 cells in 5 mice) and returned to control values at 28 days after PSL (0.58 ± 0.07/10 minutes, 160 cells in 4 mice, P > 0.05) (Fig. 3C). Hence, as in the contralateral S1,25 nerve injury also induces a transient increase in astrocyte activity.
We next examined the origin of this facilitated astrocytic activity. Calcium responses in astrocytes can be triggered by a variety of signaling pathways that includes activation of metabotropic glutamate receptors by glutamate released from neurons.40 As the activity of local excitatory neurons was not increased in the ipsilateral S1 after PSL as shown above, we examined whether excitatory inputs from the hyperactive contralateral hemisphere may contribute to this increased astrocyte activity in the ipsilateral S1 (layer 1). To test this, we monitored astrocyte calcium responses in the ipsilateral S1 while acutely blocking neuronal activity in the contralateral S1 by local injection of TTX (5 μM, 1 μL) (Fig. 4A). As a control, local application of ACSF to the contralateral S1 had little apparent effect on the elevated astrocyte activity in the ipsilateral S1 after PSL (Fig. 4B). However, local application of TTX into the contralateral S1 markedly reduced the frequency of astrocytic calcium signals in the ipsilateral S1, to a level similar to that seen in sham-operated mice as shown above. The frequency of ipsilateral astrocytic calcium signals during contralateral ACSF injection was 1.57 ± 0.10/10 minutes, 200 cells in 5 mice, and significantly lower when TTX was applied to the contralateral S1 (0.33 ± 0.03/10 minutes, 506 cells in 8 mice; P < 0.0001) (Figs. 4A–C). To further explore the role of contralateral inputs in triggering astrocyte activation, we cut the corpus callosum fiber tracts (callosotomy) 7 days before PSL. To validate that this callosotomy effectively prevented cross-hemispheric axonal communication, we injected the retrograde axonal tracer, DiI, into the contralateral S1. As shown in Figure 4D, callosotomy effectively prevented axonal transport of DiI from the contralateral S1 across to the ipsilateral S1. A sham callosotomy (control) left cross-hemispheric axonal transport of DiI intact (Fig. 4D). Furthermore, callosotomy by itself did not alter the paw withdrawal thresholds in response to mechanical stimulation of either hind paw (Figure 4E) nor did callosotomy alter the typical responses to stimulation of either hind paw after PSL (Figure 4F). However, as with TTX application, callosotomy virtually abolished any PSL-induced increase in the frequency of astrocytic calcium signals in layer 1 of the ipsilateral S1, whereas increased astrocyte activity was still seen in a sham callosotomy control group. The frequency of ipsilateral S1 astrocytic calcium transients after PSL was 1.32 ± 0.10/10 minutes (227 cells in 5 mice) in the sham group, and 0.47 ± 0.04/10 minutes (331 cells in 5 mice) in the callosotomy group (P < 0.0001) (Figs. 4G–I). Together, these consistent TTX and callosotomy results suggest that enhanced contralateral S1 activity after PSL drives the increase in ipsilateral S1 astrocyte activity.
3.4. Sustained block of local inhibition in the ipsilateral S1 induces neuronal hyperactivity and synaptic remodeling
Our results showed that PSL induces a transient increase in ipsilateral S1 astrocyte activity. We have recently reported that a similar transient increase in astrocyte activity in the contralateral S1 triggers synaptic remodeling (spine formation) and allodynia.25 We hypothesized that astrocyte-induced synaptic remodeling in the ipsilateral S1 after PSL was suppressed by the concurrent increase in inhibitory neuron activity observed above. To examine this hypothesis, we sought to chronically block GABAergic activity in the ipsilateral S1 after PSL. To do this, we applied the GABAA receptor antagonist SR95531 to the surface of the ipsilateral S1 using the noninflammatory, Elvax drug delivery system. The Elvax system results in the sustained release of drugs for about 6 days when implanted in the brain in vivo.22 Although we have previously used the Elvax system for sustained drug release,25 we first further examined the penetration of released molecules through cortical layers. We loaded Elvax with the fluorescent dye, FITC, and applied the Elvax-FITC complex to the dura surface (Fig. 5A). Three days later, we excised the brain and examined tissue distribution with fluorescence microscopy. Dye fluorescence could be found at a depth of 0 to 300 μm from the surface (layer 1 and layer 2/3) (Fig. 5B). Hence, the Elvax delivery system can likely effectively apply SR95531 across layers 1 to 3 to chronically block inhibitory GABAergic transmission in the ipsilateral S1 upper cortical layers. To further validate this, we imaged layer 2/3 neuronal activity in the ipsilateral S1 evoked by stimulation of the uninjured paw, after both sham surgery and PSL (Fig. 5C). Electrical stimulation of the intact paw evokes a modest response in layer 2/316 and even after PSL of the other hind limb, stimulation of the intact paw evokes a small response in ipsilateral layer 2/3 neurons in about 20% of the trials (Figs. 5D and E, DMSO). Application of SR95531 through Elvax markedly increased the mean probability of evoking a calcium transient in sham-operated mice and further increased it in PSL-injured mice (Figs. 5D and E). Response probabilities in the 3 experimental groups were: PSL + DMSO, 0.18 ± 0.01%, (408 cells in 5 mice); sham + SR95531, 0.44 ± 0.02%, (253 cells in 5 mice); PSL + SR95531, 0.63 ± 0.01%, (517 cells in 6 mice); and P < 0.0001 (PSL + DMSO vs Sham + SR95531), P < 0.0001 (PSL+ DMSO vs PSL+ SR95531). The amplitude of calcium transients in the ipsilateral S1 evoked by hind paw stimulation was similarly increased by application of SR95531 to sham mice, and further increased in PSL-injured mice, with mean values being (PSL + DMSO; 0.029 ± 0.002, [408 cells in 5 mice], sham + SR95531; 0.077 ± 0.003, [253 cells in 5 mice], PSL + SR95531; 0.105 ± 0.002, [518 cells in 6 mice]; and P < 0.01 [PSL + DMSO vs sham + SR95531], P < 0.0001 [PSL + DMSO vs PSL + SR95531]) (Fig. 5F). Thus, chronic inhibition of GABAA receptors in the ipsilateral S1 increased the excitability of neurons in layer 2/3 neurons, validating the effectiveness of our approach and indicating the extent to which excitatory neuron activity is modulated by inhibitory inputs.
In the contralateral S1 to nerve injury, the PSL-induced increase in neuronal activity is associated with dendritic spine remodeling.25 Therefore, we next examined whether this validated sustained block of GABA receptors and a concurrent increase in excitability in the ipsilateral S1 would be associated with any synapse remodeling. We assayed this by examining dendritic spine plasticity, defined as changes in spine formation and elimination rates. To visualize dendritic spines in the ipsilateral S1, we used M-line mice in which some layer 5 neurons express EGFP. Layer 5 neurons extend their apical dendrites into layer 1, and these apical dendrites were imaged just before, and 2 to 3 days after concurrent PSL surgery and Elvax application (Elvax loaded with DMSO or SR95531; Fig. 6A). Given the sparsity of EGFP labeling of layer 5 neurons in M-line mice, it was readily possible to detect changes over time in the same dendrite (Fig. 6B), as we and others have previously performed.21,24,25 In PSL and sham mice with DMSO or SR95531, respectively, applied to the ipsilateral S1 the rate at which new spines appeared between the first and second imaging session was similar, being 13.1 ± 1.73%, (15 dendrites from 4 mice) and 7.9 ± 1.91%, (10 dendrites from 4 mice), respectively (Fig. 6D). However, when SR95531 was applied locally to the ipsilateral S1 of PSL-injured mice, there was a modestly larger increase in this spine gain rate (Fig. 6D), being 18.3 ± 2.15%, (15 dendrites from 5 mice); P < 0.05 (sham + SR95531 vs PSL + SR95531). The rate at which spines were lost between the first and second imaging session was also larger in the PSL + SR95531 group, being 12.5 ± 1.88%, which was significantly larger than that in the PSL + DMSO group (6.3 ± 1.04%, P < 0.05). Spine loss rate in the sham + SR95531 group was 7.5 ± 1.31%. These spine gain and loss rates were averaged to give an overall turnover rate, which was significantly greater in the PSL-injured mice treated with SR95531, as compared to both the PSL + DMSO and the sham + SR95531 groups (Fig. 6F). Thus, when inhibition in the ipsilateral S1 is blocked, an increase in excitability and an increase in spine plasticity in response to PSL are unmasked. The total density of spines did not change much across all 3 groups (Fig. 6G), reinforcing that it was spine turnover, and not just gain or loss that was changed. Interestingly, although we showed above that sham + SR95531 was able to significantly elevate ipsilateral S1 excitability and increase the evoked layer 2/3 neuron responses, this was not sufficient by itself to increase spine plasticity.
3.5. Sustained block of local inhibition in the ipsilateral S1 unmasks mirror pain behavior after partial sciatic nerve ligation
We next examined the behavioral consequences of this sustained block of GABAergic receptors in the ipsilateral S1 by Elvax-mediated sustained application of SR95531. Paw withdrawal thresholds were measured in response to mechanical stimulation of the intact, noninjured hind paw using von Frey filaments (Fig. 6H). The mechanical thresholds were decreased at 3 days after PSL in mice treated with SR95531, and remained low for up to 28 days after PSL (Fig. 6I), long after delivery of SR95531 through Elvax would be expected to have ceased.22 Such allodynia or mirror pain was not seen in the sham-operated mice with SR95531, nor in the PSL-injured mice treated with DMSO (Fig. 6I). Thus, blocking inhibition in the ipsilateral S1 to PSL induced long-lasting mirror image pain-like behavior in the intact hind paw. This result suggests that the enhanced inhibitory neuron activity in the ipsilateral S1 subsequent to nerve injury is suppressing the induction of mirror image pain.
3.6. Reducing astrocytic calcium signaling inhibits synaptic remodeling in the ipsilateral S1
Given that astrocytes can secrete spinogenic factors and contribute to structural plasticity at spines,11 and the key role they play in synaptic remodeling in the contralateral S1 after PSL,25 we hypothesized that the enhanced astrocyte activity we observed in the ipsilateral S1 might similarly be involved in the synaptic remodeling described above. To test this hypothesis, we used Elvax to apply fluoroacetate, which inhibits astrocyte metabolism and thereby blocks calcium transients in astrocytes,2 to the ipsilateral S1. To enable assessment of fluoroacetate on synapse remodeling, we coapplied SR95531 to induce increased spine turnover in the ipsilateral S1 after PSL (Fig. 7A). As we demonstrated above, there was an increased frequency of astrocytic calcium signals in the ipsilateral S1 in PSL-injured mice, and this increase was similar in PSL mice treated with SR95531, being 1.51 ± 0.09/10 minutes (170 cells in 3 mice). On the other hand, the coapplication of fluoroacetate significantly decreased the frequency of ipsilateral S1 astrocyte transients to 0.26 ± 0.03/10 minutes (288 cells in 5 mice; P < 0.0001) (Figs. 7B and C). As an independent test of astrocytic calcium release being causative for the astrocyte calcium transients, we examined PSL-induced changes in the inositol trisphosphate receptor type 2 (IP3R2) knockout mouse, in which astrocytic calcium signaling is absent.38 Astrocytic calcium signals in the ipsilateral S1 in PSL-injured mice treated with SR95531 was completely absent in IP3R2 knockout mice (IP3R2 KO) (Fig. 7B; 89 cells in 2 mice). Furthermore, spine plasticity in the ipsilateral S1 after PSL was also significantly decreased by the coapplication of fluoroacetate as compared to that with SR95531 alone (Figs. 7D–H). The rate of new spine appearance in the ipsilateral S1 after PSL was 15.4 ± 1.0% in SR95531 alone (15 dendrites) and 8.4 ± 0.9% in PSL mice treated with both SR95531 and fluoroacetate (21 dendrites, P = 0.0009) (Fig. 7F). The spine elimination rates were also reduced by fluoroacetate (PSL + SR95531; 18.3 ± 2.2%, PSL + SR95531 + fluoroacetate: 9.2 ± 1.4%, P = 0.024) (Fig. 7G). The combined turnover rate was consequently also significantly decreased by fluoroacetate (PSL + SR95531; 12.5 ± 2.2%, PSL + SR95531 + fluoroacetate: 7.6 ± 1.1%, P < 0.0001) (Fig. 7H). Thus, enhanced astrocyte activity seems necessary for PSL and SR95531-induced spine plasticity in the ipsilateral S1.
3.7. Reducing astrocytic calcium signaling alleviates mirror pain behavior after partial sciatic nerve ligation
We next measured the behavioral consequences of inhibiting astrocyte calcium signaling. We applied the 3 experimental models we have shown above to reduce astrocyte activation in the ipsilateral S1 subsequent to PSL: callosotomy, local application of fluoroacetate, and the IP3R2 KO mouse. Mirror pain behavior was induced by Elvax-mediated application of SR95531 to the ipsilateral S1, and mechanical thresholds in response to stimulation of the intact paw were examined after PSL (Fig. 8A). With SR95531 application alone in control PSL mice, the paw withdrawal thresholds were markedly decreased, and this allodynia was maintained for at least 28 days (ie, mirror pain). Consistently, when astrocytic activity was reduced or prevented by these 3 different approaches, there was a significant reduction in the extent of mirror pain behavior (Fig. 8B). Astrocytic hyperactivity gradually disappears after PSL injury as shown above. Thus, to further assess the involvement of astrocytic activities in mirror image pain, we examined the temporal relationship between PSL and SR95531 induction of mirror pain. As also shown in Figure 8B, application of SR95531 to the ipsilateral S1 coincident with PSL induced strong and sustained mirror pain. When SR95531 was applied to the ipsilateral S1 at 14 days after PSL, the extent of the mirror pain behavior was reduced. On the other hand, application of SR95531 at 28 days after PSL did not change withdrawal thresholds. This indicates that delayed block of GABAergic signaling after PSL does not induce mirror pain (Fig. 8C). In summary, we conclude that both astrocyte activity and block of inhibition in the ipsilateral S1 is needed for mirror image pain behavior, and the block of inhibition is only effective when applied coincident with the PSL.
Integration of sensory responses requires coordination of neural activity across both hemispheres, and inhibitory neurons in the S1 receive transcallosal excitatory inputs from the contralateral S1 to regulate excitatory cortical pyramidal neurons.30 Therefore, alteration of the contralateral S1 neuronal activity affects inhibitory neuronal activity in the ipsilateral S1. Indeed, denervation of the sciatic nerve induces an increased functional magnetic resonance imaging signal in both the contralateral and ipsilateral S1,32 and enhancement of inhibitory neurons contributes to this facilitated brain function after denervation.33 In this study, we demonstrated that inhibitory neuronal activities increased in the ipsilateral S1 to injured hind paw after PSL, and the enhanced activities were abolished in the presence of TTX in the contralateral S1 to injury. Because PSL enhances excitatory neuronal activities in the S1 contralateral to the injured hind limb,24 callosal inputs from the contralateral S1 can be increased and facilitate inhibitory neuronal activities in the ipsilateral S1. On the other hand, excitatory neuronal activities decreased in layer 2/3 of the ipsilateral S1 after PSL, and SR95531-induced disinhibition of the ipsilateral S1 to injury caused enhancement of excitatory neuronal activities in PSL-injured mice compared with those in wild-type mice. Because layer 1 inhibitory neurons make synapses onto layer 2/3 excitatory neurons and can modulate excitatory neuronal activities,42 those results suggest that increased activities of inhibitory neurons in the ipsilateral S1 may prevent induction of exaggerated excitatory neuronal activities. Furthermore, SR95531-induced disinhibition of the ipsilateral S1 also causes synaptic remodeling and chronic pain–like behavior. Enhanced excitatory neuronal activities induce synaptic remodeling in the S1 contralateral to the injured hind limb after PSL, which in turn causes mechanical allodynia.24 A recent report demonstrates that activities of SOM neurons, which are one of the inhibitory neurons, reduce in the S1 of nerve-injured mice and enhance dendritic activities. Moreover, long-term activation of SOM neurons prevents induction of nerve injury-induced chronic pain.10 Therefore, facilitated inhibitory neuronal activities reduce excitatory neuronal activities and inhibit synaptic remodeling, resulting in preventing an induction of mirror image pain in the normal hind limb after PSL and disruption of the inhibitory neuronal function may lead to mirror image pain.
What are the mechanisms of SR95531-induced synaptic remodeling in the ipsilateral S1? Astrocytes, which are one of the glial cells, are well known to modulate synaptic transmission34 and also are involved in synaptic remodeling, such as synaptogenesis and synapse elimination.1,9,14,15 In addition, activated astrocytes in the S1 release a synaptogenic molecule and cause synaptic remodeling in layer 5 pyramidal neurons and subsequently induce chronic pain.25 Here, we found that astrocytic activities increased in the ipsilateral S1 after PSL, and callosotomy and TTX application in the contralateral S1 abolished the astrocytic hyperactivities, suggesting that callosal inputs cause the astrocytic hyperactivities. Astrocytes express various types of receptors of neurotransmitters.31 Among them, glutamate is a major molecule that can modulate astrocytic activities through mGluR5 receptor,12,13 and astrocytes can sense sensory information by glutamate signaling in vivo.40 Furthermore, in chronic pain, expression of mGluR5 increases in the S1 and responsiveness to glutamate is enhanced.25 Thus, the enhanced astrocytic activities may be mediated by glutamate signaling from the contralateral S1. Further studies are necessary to clarify whether glutamate signaling contributes to the hyperactivities of astrocytes.
Astrocytic activation can induce synapse remodeling.25 Calcium responses in astrocytes can regulate synaptic plasticity through the release of cytokines, D-serine, glutamate, and thrombospondin. These molecules combine with neurotransmitters released depending on neuronal activity to modulate neural circuits.29,31 Furthermore, thrombospondins released from astrocytes can directly generate spines15 and thus increase spine turnover.25 Indeed, thrombospondin 1 released from astrocytes in the S1 cause synaptogenesis and results in chronic pain–like behavior.25 Although astrocytes were clearly and strongly activated by PSL in the current study, morphological spine plasticity was only induced when the local inhibitory influence was blocked. In addition, SR95531 treatment at 28 days after PSL, when astrocytic activities return to normal level, did not induce chronic pain in the intact hind paw. Hence, increased astrocytic activities increase the plasticity in the neuronal circuit of the S1 and make ready to induce synaptic remodeling. And structural plasticity may require an appropriate enhanced local excitatory neuronal activity combined with astrocytic activation.
Recent reports indicate that plastic changes in cortical areas have critical roles in chronic pain. In the insular cortex, which integrates sensory and emotional aspects of pain, activity-dependent synaptic plasticity occurs after nerve injury37 and inhibition of those plastic changes works as analgesic. As for the S1, attenuation of neuronal activities or blockade of synaptic remodeling induced by astrocytes alleviates chronic pain.16,25 In this study, we show that intracortical rewiring of neuronal network induced by astrocytes in the ipsilateral S1 to injury caused allodynia in the intact hind paw. This indicates that just only cortical remodeling by itself cause chronic pain-like behavior. Indeed, epilepsy originated from the somatosensory cortex causes pain sensation in human patients.43 In addition, activation of the S116 and rewiring of the S1 induced by thrombospondin cause allodynia-like behavior.25 Therefore, our results support the idea that rewiring of cortical network contributes to induction and maintenance of chronic pain.
Some individuals with chronic pain can suffer from mirror image and/or referred pain, where the pain is sensed or evoked at a site that is contralateral to the actual injury.41 Accumulation of evidence suggests that peripheral and spinal mechanisms are involved in mirror image pain. Peripheral inflammation induces bilateral allodynia by activation of spinal glial cells.4,26 In addition, satellite glial cells in the contralateral side to the injured hind limb are activated by TNF-alpha released from the ipsilateral side and facilitate ipsilateral peripheral neuronal activity, which in turn causes mirror image pain.6 The present results demonstrate that in addition to peripheral and spinal mechanisms, cortical reorganization induced by disinhibition and glial activation may have critical roles in mirror image pain. Thus, it will be interesting to examine whether patients suffering from mirror image pain have a deficit in the ability of the local inhibition to apply a break to the astrocyte-induced allodynia, as we have observed here. Regardless, the present data emphasize the importance of considering the balance, and interplay between, excitatory neuron activity, local inhibitory neuron activity and astrocyte activation in determining the cellular and molecular mechanisms that may mediate the establishment of chronic pain symptoms.
Conflict of interest statement
The authors have no conflict of interest to declare.
This study was supported by Core Research for Evolutional Science and Technology (CREST) grant from the Japan Agency for Medical Research and Development (AMED) and from the Japan Science and Technology Agency (JST) (to J. Nabekura); Grant-in-Aids for Scientific Research (A) (22240042, to J. Nabekura), Grant-in-Aid for Young Scientists (15K21604 to T. Ishikawa) (15K21603 to K. Eto), Grant-in-Aids for Scientific Research (C) (17K09051, to K. Eto), Grant-in-Aid for Scientific Research on Innovative Areas (16H01399, to K. Eto), Grant-in-Aid for Scientific Research on Innovative Areas-Resource and technical support platforms for promoting research Advanced Bioimaging Support (JP16H06280). This work was also supported by Spectrography and Bioimaging Facility, NIBB Core Research Facilities.
. Allen NJ, Bennett ML, Foo LC, Wang GX, Chakraborty C, Smith SJ, Barres BA. Astrocyte
glypicans 4 and 6 promote formation of excitatory synapses via GluA1 AMPA receptors. Nature 2012;486:410–14.
. Andersson M, Blomstrand F, Hanse E. Astrocytes play a critical role in transient heterosynaptic depression in the rat hippocampal CA1 region. J Physiol 2007;585:843–52.
. Bushnell MC, Duncan GH, Hofbauer RK, Ha B, Chen JI, Carrier B. Pain perception: is there a role for primary somatosensory cortex
? Proc Natl Acad Sci U S A 1999;96:7705–9.
. Cao J, Li Z, Zhang Z, Ren X, Zhao Q, Shao J, Li M, Wang J, Huang P, Zang W. Intrathecal injection of fluorocitric acid inhibits the activation of glial cells causing reduced mirror pain in rats. BMC Anesthesiol 2014;14:119.
. Chen TW, Wardill TJ, Sun Y, Pulver SR, Renninger SL, Baohan A, Schreiter ER, Kerr RA, Orger MB, Jayaraman V, Looger LL, Svoboda K, Kim DS. Ultrasensitive fluorescent proteins for imaging neuronal activity. Nature 2013;499:295–300.
. Cheng CF, Cheng JK, Chen CY, Lien CC, Chu D, Wang SY, Tsaur ML. Mirror-image pain is mediated by nerve growth factor produced from tumor necrosis factor alpha-activated satellite glia after peripheral nerve injury. PAIN 2014;155:906–20.
. Choi HS, Roh DH, Yoon SY, Moon JY, Choi SR, Kwon SG, Kang SY, Han HJ, Kim HW, Beitz AJ, Oh SB, Lee JH. Microglial interleukin-1 b in the ipsilateral dorsal horn inhibits the development of mirror-image contralateral mechanical allodynia through astrocyte
activation in a rat model of inflammatory pain. PAIN 2015;156:1046–59.
. Choi HS, Roh DH, Yoon SY, Choi SR, Kwon SG, Kang SY, Moon JY, Han HJ, Beitz AJ, Lee JH. Differential involvement of ipsilateral and contralateral spinal cord astrocyte
D-serine in carrageenan-induced mirror-image pain: role of σ 1 receptors and astrocyte
gap junctions. Br J Pharmacol 2018;175:558–72.
. Chung WS, Clarke LE, Wang GX, Stafford BK, Sher A, Chakraborty C, Joung J, Foo LC, Thompson A, Chen C, Smith SJ, Barres BA. Astrocytes mediate synapse elimination through MEGF10 and MERTK pathways. Nature 2013;504:394–400.
. Cichon J, Blanck TJJ, Gan WB, Yang G. Activation of cortical somatostatin interneurons prevents the development of neuropathic pain. Nat Neurosci 2017;20:1122–32.
. Clarke LE, Barres BA. Emerging roles of astrocytes in neural circuit development. Nat Rev Neurosci 2013;14:311–21.
. Ding S, Fellin T, Zhu Y, Lee SY, Auberson YP, Meaney DF, Coulter DA, Carmignoto G, Haydon PG. Enhanced astrocytic calcium signals contribute to neuronal excitotoxicity after status epilepticus. J Neurosci 2007;27:10674–84.
. Ding S, Wang T, Cui W, Haydon PG. Photothrombosis ischemia stimulates a sustained astrocytic calcium signaling in vivo. Glia 2009;57:767–76.
. Diniz LP, Almeida JC, Tortelli V, Vargas Lopes C, Setti-Perdigão P, Stipursky J, Kahn SA, Romão LF, de Miranda J, Alves-Leon SV, de Souza JM, Castro NG, Panizzutti R, Gomes FC. Astrocyte
-induced synaptogenesis is mediated by transforming growth factor β signaling through modulation of D-serine levels in cerebral cortex neurons. J Biol Chem 2012;287:41432–45.
. Eroglu C, Allen NJ, Susman MW, O'Rourke NA, Park CY, Ozkan E, Chakraborty C, Mulinyawe SB, Annis DS, Huberman AD, Green EM, Lawler J, Dolmetsch R, Garcia KC, Smith SJ, Luo ZD, Rosenthal A, Mosher DF, Barres BA. Gabapentin receptor alpha2delta-1 is a neuronal thrombospondin receptor responsible for excitatory CNS synaptogenesis. Cell 2009;139:380–92.
. Eto K, Wake H, Watanabe M, Ishibashi H, Noda M, Yanagawa Y, Nabekura J. Inter-regional contribution of enhanced activity of the primary somatosensory cortex
to the anterior cingulate cortex accelerates chronic pain
behavior. J Neurosci 2011;31:7631–6.
. Eto K, Ishibashi H, Yoshimura T, Watanabe M, Miyamoto A, Ikenaka K, Moorhouse AJ, Nabekura J. Enhanced GABAergic activity in the mouse primary somatosensory cortex
is insufficient to alleviate chronic pain
behavior with reduced expression of neuronal potassium-chloride cotransporter. J Neurosci 2012;32:16552–9.
. Feng G, Mellor RH, Bernstein M, Keller-Peck C, Nguyen QT, Wallace M, Nerbonne JM, Lichtman JW, Sanes JR. Imaging neuronal subsets in transgenic mice expressing multiple spectral variants of GFP. Neuron 2000;28:41–51.
. Futatsugi A, Nakamura T, Yamada MK, Ebisui E, Nakamura K, Uchida K, Kitaguchi T, Takahashi-Iwanaga H, Noda T, Aruga J, Mikoshiba K. IP3 receptor types 2 and 3 mediate exocrine secretion underlying energy metabolism. Science 2005;309:2232–4.
. Gao YJ, Xu ZZ, Liu YC, Wen YR, Decosterd I, Ji RR. The c-Jun N-terminal kinase 1(JNK1) in spinal astrocytes is required for the maintenance of bilateral mechanical allodynia under a persistent inflammatory pain condition. PAIN 2010;148:309–19.
. Holtmaat AJ, Trachtenberg JT, Wilbrecht L, Shepherd GM, Zhang X, Knott GW, Svoboda K. Transient and persistent dendritic spines in the neocortex in vivo. Neuron 2005;45:279–91.
. Kakizawa S, Miyazaki T, Yanagihara D, Iino M, Watanabe M, Kano M. Maintenance of presynaptic function by AMPA receptor-mediated excitatory postsynaptic activity in adult brain. Proc Natl Acad Sci U S A 2005;102:19180–5.
. Kawasaki Y, Xu ZZ, Wang X, Park JY, Zhuang ZY, Tan PH, Gao YJ, Roy K, Corfas G, Lo EH, Ji RR. Distinct roles of matrix metalloproteases in the early- and late-phase development of neuropathic pain. Nat Med 2008;14:331–6.
. Kim SK, Nabekura J. Rapid synaptic remodeling in the adult somatosensory cortex following peripheral nerve injury and its association with neuropathic pain. J Neurosci 2011;31:5477–82.
. Kim SK, Hayashi H, Ishikawa T, Shibata K, Shigetomi E, Shinozaki Y, Inada H, Roh SE, Kim SJ, Lee G, Bae H, Moorhouse AJ, Mikoshiba K, Fukazawa Y, Koizumi S, Nabekura J. Cortical astrocytes rewire somatosensory cortical circuits for peripheral neuropathic pain. J Clin Invest 2016;126:1983–97.
. Milligan ED, Twining C, Chacur M, Biedenkapp J, O'Connor K, Poole S, Tracey K, Martin D, Maier SF, Watkins LR. Spinal glia and proinflammatory cytokines mediate mirror-image neuropathic pain in rats. J Neurosci 2003;23:1026–40.
. Moisset X, Bouhassira D. Brain imaging of neuropathic pain. Neuroimage 2007;37:S80–8.
. Nimmerjahn A, Kirchhoff F, Kerr JN, Helmchen F. Sulforhodamine 101 as a specific marker of astroglia in the neocortex in vivo. Nat Methods 2004;1:31–7.
. Ota Y, Zanetti AT, Hallock RM. The role of astrocytes in the regulation of synaptic plasticity and memory formation. Neural Plast 2013;2013:185463.
. Palmer LM, Schulz JM, Murphy SC, Ledergerber D, Murayama M, Larkum ME. The cellular basis of GABA(B)-mediated interhemispheric inhibition. Science 2012;335:989–93.
. Pannasch U, Rouach N. Emerging role for astroglial networks in information processing: from synapse to behavior. Trends Neurosci 2013;36:405–17.
. Pelled G, Chuang KH, Dodd SJ, Koretsky AP. Functional MRI detection of bilateral cortical reorganization in the rodent brain following peripheral nerve deafferentation. Neuroimage 2007;37:262–73.
. Pelled G, Bergstrom DA, Tierney PL, Conroy RS, Chuang KH, Yu D, Leopold DA, Walters JR, Koretsky AP. Ipsilateral cortical fMRI responses after peripheral nerve damage in rats reflect increased interneuron activity. Proc Natl Acad Sci U S A 2009;106:14114–19.
. Perea G, Araque A. Properties of synaptically evoked astrocyte
calcium signal reveal synaptic information processing by astrocytes. J Neurosci 2005;25:2192–203.
. Petrovic P, Ingvar M, Stone-Elander S, Petersson KM, Hansson P. A PET activation study of dynamic mechanical allodynia in patients with mononeuropathy. PAIN 1999;83:459–70.
. Peyron R, Schneider F, Faillenot I, Convers P, Barral FG, Garcia-Larrea L, Laurent B. An fMRI study of cortical representation of mechanical allodynia in patients with neuropathic pain. Neurology 2004;63:1838–46.
. Qiu S, Chen T, Koga K, Guo YY, Xu H, Song Q, Wang JJ, Descalzi G, Kaang BK, Luo JH, Zhuo M, Zhao MG. An increase in synaptic NMDA receptors in the insular cortex contributes to neuropathic pain. Sci Signal 2013;6:ra34.
. Takata N, Mishima T, Hisatsune C, Nagai T, Ebisui E, Mikoshiba K, Hirase H. Astrocyte
calcium signaling transforms cholinergic modulation to cortical plasticity in vivo. J Neurosci 2011;31:18155–65.
. Wang Y, Kakizaki T, Sakagami H, Saito K, Ebihara S, Kato M, Hirabayashi M, Saito Y, Furuya N, Yanagawa Y. Fluorescent labeling of both GABAergic and glycinergic neurons in vesicular GABA transporter (VGAT)-venus transgenic mouse. Neuroscience 2009;164:1031–43.
. Wang X, Lou N, Xu Q, Tian GF, Peng WG, Han X, Kang J, Takano T, Nedergaard M. Astrocytic calcium signaling evoked by sensory stimulation in vivo. Nat Neurosci 2006;9:816–23.
. Watkins LR, Maier SF. Beyond neurons: evidence that immune and glial cells contribute to pathological pain states. Physiol Rev 2002;82:981–1011.
. Wozny C, Williams SR. Specificity of synaptic connectivity between layer 1 inhibitory interneurons and layer 2/3 pyramidal neurons in the rat neocortex. Cereb Cortex 2011;21:1818–26.
. Yazawa S, Ikeda A, Sawamoto N, Terada K, Fukuyama H, Tanaka M, Shibasaki H. Painful focal sensory seizure arising from the primary somatosensory cortex
. Intern Med 2003;42:875–9.
. Zhu Y, Stornetta RL, Zhu JJ. Chandelier cells control excessive cortical excitation: characteristics of whisker-evoked synaptic responses of layer 2/3 nonpyramidal and pyramidal neurons. J Neurosci 2004;24:5101–8.