Evidence for a Role of Connexin 43 in Trigeminal Pain Using RNA Interference In Vivo

Abstract

The importance of glial cells in the generation and maintenance of neuropathic pain is becoming widely accepted. We examined the role of glial-specific gap junctions in nociception in the rat trigeminal ganglion in nerve-injured and -uninjured states. The connexin 43 (Cx43) gap-junction subunit was found to be confined to the satellite glial cells (SGCs) that tightly envelop primary sensory neurons in the trigeminal ganglion and we therefore used Cx43 RNA interference (RNAi) to alter gap-junction function in SGCs. Using behavioral evaluation, together with immunocytochemical and Western blot monitoring, we show that Cx43 increased in the trigeminal ganglion in rats with a chronic constriction injury (CCI) of the infraorbital nerve. Reducing Cx43 expression using RNAi in CCI rats reduced painlike behavior, whereas in non-CCI rats, reducing Cx43 expression increased painlike behavior. The degree of painlike behavior in CCI rats and intact, Cx43-silenced rats was similar. Our results support previous suggestions that increases in glial gap junctions after nerve injury increases nociceptive behavior but paradoxically the reduction of gap junctions in normal ganglia also increases nociceptive behavior, possibly a reflection of the multiple functions performed by glia.

INTRODUCTION

Glial cells play a key role in maintaining neuronal integrity and normal function in the CNS. Recently it has become clear that changes in CNS glia (astrocytes and microglia) can be a factor in the pathophysiology of pain and lead to the development and maintenance of neuropathic pain (Cronin et al. 2008; Watkins and Maier 2002). In the peripheral nervous system, neurons located in sensory ganglia are tightly surrounded by a type of glial cell called a satellite glial cell (SGC). Following injury to a peripheral nerve, SGCs undergo changes similar to those found in the CNS and show a marked increase in expression of glial fibrillary acidic protein (GFAP) (Chudler et al. 1997; Stephenson and Byers 1995; Vit et al. 2006). The question that arises from these observations is whether the changes in SGCs following nerve injury contribute to the appearance and/or maintenance of pain similar to the role played by their counterparts in the CNS.

It is known that spontaneous firing and/or hyperexcitability of sensory neurons contribute to neuropathic pain (Kajander et al. 1992; Nordin et al. 1984) and among the functions of SGCs is the regulation of the extracellular perineuronal environment with consequent effects on neuronal activity (Hanani 2005). We recently demonstrated that selective knockdown of an SGC potassium (K⁺) channel in the trigeminal ganglion resulted in the reduction of nociceptive thresholds and an increase in behavioral pain measures (Vit et al. 2008). We proposed that changes in the extracellular ion composition resulting from altered K⁺ buffering capabilities of the SGCs cause a perturbation in neuronal firing. Another major component for regulating the perineuronal ionic environment is the coupling between adjacent SGCs (Huang et al. 2005) via gap junctions that allows rapid transcellular exchange of small molecules. This coupling is known to be dependent on physiological and pathological conditions and that the numbers of gap junctions between SGCs increase markedly in response to nerve injury (Hanani et al. 2002; Huang et al. 2005) and decrease with age (Procacci et al. 2008). These changes led to the hypothesis that gap junctions between SGCs might play a role in the appearance and maintenance of neuropathic pain (Cherkas et al. 2004; Hanani et al. 2002).

The major structural components of gap junctions are connexins and the connexin 43 (Cx43) subunit is expressed preferentially by glial cells in the CNS (Nagy et al. 1997; Ochalski et al. 1997; Yamamoto et al. 1990) and by SGCs in the trigeminal ganglion (Procacci et al. 2008; Vit et al. 2006) and these proteins (including Cx43) have half-lives of 1 to 5 h (Saffitz et al. 2000). Following nerve injury there is an increase in SGC Cx43 expression (Hanani et al. 2002; Pannese et al. 2003), although there is no evidence for a causal connection between increased Cx43 expression and increased pain sensation. It would be predicted that reduction of gap junctions might counter the effects of nerve injury and, indeed, carbenoxolone—a nonselective gap junction inhibitor—has been shown to produce analgesia in different pain models (Dublin and Hanani 2007; Lan et al. 2007; Qin et al. 2006; Spataro et al. 2004). In contrast, however, we recently reported that reducing Cx43 expression in the trigeminal ganglion of normal rats was associated with an increase of painlike behavior. This latter result seems contradictory with the analgesic effect of carbenoxolone and the difference might be that carbenoxolone is not connexin-type specific and the systemic or intrathecal administration of the drug makes it difficult to pinpoint the precise site of action. Further, there is evidence that carbenoxolone has a direct action on neurons (Rouach et al. 2003) in addition to gap junctions.

In the present study, we sought to clarify the role of glial-expressed Cx43 in pain behavior by silencing the expression of glial Cx43 in the trigeminal ganglion. Our results provide evidence that inhibiting Cx43 in the trigeminal ganglion can have different effects depending on whether there is a coincident nerve injury.

METHODS

Animals and surgeries

ANIMALS.

Adult male Sprague–Dawley rats (Charles River Laboratories; http://www.criver.com) weighing between 270 and 330 g were housed on a 12-h light–dark cycle and given food and water without restriction. Procedures for the maintenance and use of the experimental animals conformed to the regulations of University of California San Francisco and Cedars–Sinai Medical Center Committees on Animal Research and were carried out in accordance with the guidelines of the National Institutes of Health regulations on animal use and care (Publication 85–23, Revised 1996).

CHRONIC CONSTRICTION INJURY (CCI) OF THE INFRAORBITAL NERVE (ION).

Rats were anesthetized with a mixture of ketamine (90 mg/kg; http://www.abbott.com) and xylazine (10 mg/kg; http://www.phoenixpharm.com), then placed in the stereotaxic head holder. The skull and nasal bone were exposed through a 2-cm skin incision. The temporal muscle was detached from the rostral upper edge of the orbit. Gentle retraction of the orbital contents exposed the ION. Two 5-0 chromic gut ligatures (2 mm apart) were loosely tied around the exposed nerve. The incision was closed using 6-0 silk sutures or the CCI of the ION was followed by implantation of a guide cannula (see following text).

CANNULA IMPLANT FOR INJECTION IN THE MAXILLARY DIVISION OF THE TRIGEMINAL GANGLION.

The skull was exposed and a bur hole was drilled above the location of the maxillary division of the left trigeminal ganglion at 6.5 mm anterior to interaural zero and 2.3 mm lateral to the midline. A guide cannula pedestal (http://www.plastics1.com) was fixed to the skull over the bur hole using three stainless steel screws (http://www.aaronsmachinescrews.com) and dental acrylic cement. The guide cannula extended into the bur hole 1 mm below the pedestal but did not touch the surface of the cortex. At least 7 days were allowed for recovery from surgery before injection into the trigeminal ganglion.

INJECTION INTO THE TRIGEMINAL GANGLION.

Each animal received a single injection of double-stranded (ds)RNA. Prior to the injection, rats were lightly anesthetized with isoflurane. A 33-gauge beveled stainless steel cannula (http://www.plastics1.com) was inserted through the guide cannula (positioned over the maxillary division of the left trigeminal ganglion as described earlier) to 9.5 mm below the cortical surface. The injection cannula was connected to a 25-μl Hamilton syringe attached to a microinjection pump set to deliver 2 μl over a 1-min period.

SYNTHESIS OF DSRNA AND PREPARATION FOR INJECTION.

Total RNA was extracted from rat brain tissue. A reverse transcription reaction was set up using 1 μg of RNA. Complementary DNA (cDNA) of genes of interest produced by a 30-cycle polymerase chain reaction using gene-specific primers was cloned into a pTOPO vector (Invitrogen). The specific forward and reverse primer sequences for Cx43 corresponded to nucleotides 307–324 and 725–742, respectively (GenBank accession number NM_012567). Rat β-globin sequences were used as nonspecific dsRNA control and were described previously (Bhargava et al. 2004). Sense RNA and antisense RNA were synthesized from cDNA inserts by using a MegaScript RNA kit (Ambion) according to the manufacturer's specification.

Prior to the injection in the trigeminal ganglion, 15 μg of dsRNA was mixed with lipofectamine-2000 (http://www.invitrogen.com) in a final volume of 5 μl at room temperature. After a 30-min incubation, the red fluorescent marker DiI, DiIC₁₈(3), or 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (http://www.probes.invitrogen.com) was added to the mixture at a final concentration of 10 μM.

Behavioral testing

SPONTANEOUS EYE CLOSURE.

The testing was carried out as previously described (Vit et al. 2008). The number of unilateral eye closures and the side (left or right) were recorded for three 2-min periods at 5-min intervals. The testing sessions were videotaped for posttest analysis. The results are expressed as the mean number of eye closures per minute.

MECHANICAL (VON FREY HAIR) TESTING.

Three von Frey hairs—2, 10, and 50 g, corresponding to log units 4.31, 5.07, and 5.88, respectively—were used in the study based on our own preliminary testing and the results of Ren (1999). During testing, mechanical stimulation was done with increasing intensities. Each filament was first applied on the side contralateral to the dsRNA injection or to the CCI of the ION. The stimulation consisted of five to six consecutive applications performed at 5-s intervals in slightly different areas at the center and around the vibrissal pad, as well as in the perioral and perinasal territory. The scoring of the rat behavioral response was based on the method of Vos and colleagues (1994) as follows: 0 = no detection; 1 = detection and exploration of the von Frey hair; 2 = head withdrawal and/or single grabbing movement; 3 = attack and/or escape and/or multiple grabbing movements; and 4 = active asymmetrical grooming directed toward the stimulated facial area. For each hair, the highest score was recorded and the results are presented as the average of the highest score obtained from the three different hairs. The average value is presented because the analysis of scores from each hair separately gave the same results when compared between groups (see Supplemental Fig. S1).1

CONDITIONAL REWARD.

Rats were tested for avoidance of an innocuous stimulus when attempting to obtain a reward that was sugar-sweetened water (20% sucrose) delivered from a drinking tube. Rats that acclimated to sweetened water in their home cage quickly developed a preference and would drink in the test chamber without the need for water deprivation (Davis and Campbell 1973). The testing apparatus (Med Associates) consisted of a chamber (20 × 20 × 20 cm) with a 5 × 5 × 3-cm alcove in one wall with a stainless steel drinking spout located at the back. A conductance-measuring device (Lickometer, Ugo Basile, Comerio, Italy) attached to the drinking spout was used to count the number of licks when the rat drank from the spout. Brushes were positioned on the left side of the drinking alcove (ipsilateral to the injected trigeminal ganglion or to the CCI of the ION), so that the rats had to maintain contact with the brushes while drinking. The number and the position of the brushes were set up to stimulate the vibrissal pad as well as the perinasal and perioral area of the rat face. During testing, the rat was placed in the chamber for a minimum of 5 min and returned to the homecage after 5 min or after the tenth attempt. The observer recorded the number of successful and unsuccessful licking attempts. A licking episode (successful attempt) was defined as an episode where the rat had a minimum of six licks before withdrawing its head from the drinking alcove. A licking episode could be one continuous series of licks or several periods of licking without removing the head from the drinking area. The number of licks was corrected for the weight of animals and the results of the licking behavior are expressed as number of licks per licking episode and per kilogram weight of rat.

Testing schedule

The effect of silencing Cx43 in the trigeminal ganglion was tested in the absence or presence of a CCI of the left ION. The time course of the experiments is presented in Table 1.

TABLE 1.

Time course of experiments

Experiment 1: Effect of Cx43 dsRNA on pain behavior in noninjured rats
Day −21	Days −18, −16	Days −14, −12, −10	Day −8	Day −7
Conditional reward: Introduction of sucrose	Conditional reward: Training (2 trials)	Conditional reward: Training (1 trial) von Frey and eye closure: Habituation to testing environment	Conditional reward and von Frey and eye closure: NAÏV̈E baseline	Cannula implant

Day −3	Day −1	Day 0	Days 1, 3, 5, 7, 10, 13
Conditional reward and von Frey and eye closure: Post-implant training	Conditional reward and von Frey and eye closure: IMPLANT baseline	Injection of dsRNA	Conditional reward and von Frey and eye closure: Effect of dsRNA

Experiment 2: Effect of Cx43 dsRNA on pain behavior in injured rats
Day −14	Days −11, −9	Days −7, −5, −3	Day −1	Day 0
Conditional reward: Introduction of sucrose	Conditional reward: Training (2 trials)	Conditional reward: Training (1 trial) von Frey and eye closure: Habituation to testing environment	Conditional reward and von Frey and eye closure:	CCI of left ION + Cannula implant
			NAÏVE baseline

Days 3, 7	Day 10	Days 11, 13
Conditional reward and von Frey and eye closure: Effect of injury	Conditional reward and von Frey and eye closure: Injection of dsRNA	Conditional reward and von Frey and eye closure: Effect of dsRNA

For experiment 1, behavioral training was started 3 wk prior to the injection (day −21). On day −21, animals were habituated to drink 20% sucrose in water in their homecage. Rats were then tested every other day in the lick-test chamber for five sessions consisting of two trials on days −18 and −16, and a single trial on days −14, −12, and −10. Habituation to both the environment and the procedure for the eye closure test and the orofacial von Frey test was performed during the last three sessions (days −14, −12, and −10). Baseline values of naïve rats for all tests were collected on day −8. Cannulae were implanted on day −7. The lick test was done on day −3 and baseline values of implanted rats were recorded again on day −1 for licking behavior, eye closure, and orofacial von Frey hair testing. On day 0, rats were randomly separated into two groups. One group received an injection of Cx43 dsRNA and the other was injected with globin dsRNA into the maxillary division of the left trigeminal ganglion. Licking behavior, eye closure, and orofacial von Frey hair testing were then performed on days 1, 3, 5, 7, 10, and 13 after injection of the dsRNA.

In experiment 2, after training for the conditional reward test (days −14 to day −3) and habituation to the environment for the orofacial von Frey test and the eye closure test (day −7, −5, and −3), baseline values of naïve rats for all tests were recorded on day −1. On day 0, all rats underwent a CCI of the left ION followed by cannula implantation. The effect of the CCI on pain behavior (three tests) was measured on days 3, 7, and 10. On day 10 (the time point by when all animals had reached maximal behavioral scores), after behavioral testing, the rats were randomly separated into two groups. One group was injected with Cx43 dsRNA and the other with globin dsRNA into the maxillary division of the left trigeminal ganglion. The effect of dsRNA on licking behavior, eye closure, and orofacial von Frey hair sensitivity was then tested 1 and 3 days after the injection (days 11 and 13).

For all experiments, a treatment-blind observer conducted behavioral testing between 10:00 am and 4:00 pm. On each testing day, rats were brought into the behavior room ≥30 min prior to the test session to habituate them to the environment.

Statistical analysis

All results are expressed as means ± SE. All the data were analyzed using SigmaStat software (http://www.systat.com). Results were considered statistically significant at P < 0.05. Differences between the control and Cx43 dsRNA groups and changes over time were analyzed by a mixed repeated-measures (RM) ANOVA followed by Bonferroni post hoc comparisons between groups. In experiment 1, significant changes post-dsRNA injection were established by comparison with baseline values (implanted animals) using Student's t-test for paired data. In experiment 2, the effects of dsRNA after nerve injury were evaluated in each group by comparing data post-dsRNA injection with values from injured animals (10 days post-CCI) using Student's t-test for paired data. For eye closure and von Frey tests, in each group and each experiment, differences between ipsilateral and contralateral sides to the injection and/or injury were tested over time with a two-way RM ANOVA, followed by multiple comparisons between both sides using paired t-test. Moreover, in experiment 1, the magnitude of the effect in dsRNA-injected rats was determined for day 3 (day of maximum effect) by comparison with a 10-day CCI of the ION. The differences between dsRNA/day 3 or CCI of the ION/day 10 and their respective baseline values were calculated and a one-way ANOVA was performed to assess differences between the groups.

Tissue processing

HISTOLOGY.

Rats were deeply anesthetized with 100 mg/kg of pentobarbital (administered intraperitoneally). For light microscopy (LM), rats were perfused transcardially with 10% formalin. The left and right trigeminal ganglia were postfixed in the same fixative for 30 min and then placed in 30% sucrose in phosphate-buffered saline (PBS, pH 7.4) for 48 h. Left and right trigeminal ganglia from each animal were embedded together in tissue freezing medium (TBS; http://www.trianglebiomedical.com) and cut longitudinally at 10 μm on a cryostat.

For electron microscopy (EM), rats were perfused with 2% paraformaldehyde/0.1% glutaraldehyde, then trigeminal ganglia were postfixed for 5 h. Sections (50 μm) were cut on a Vibratome, processed for Cx43 immunocytochemistry using diaminobenzidine (DAB) as a chromogen (see following text), osmicated, dehydrated, and embedded in Epon. Thin sections were stained with uranyl acetate and lead citrate (Ohara et al. 1983).

IMMUNOHISTOCHEMISTRY.

Sections were blocked in 5% normal goat serum (NGS), 0.3% Triton X-100 (http://www.sigmaaldrich.com) in PBS for 1 h, and then incubated in the primary antiserum, Cx43 (Rb, 1:1,000, Zymed and Rb, 1:2,000, Sigma), glutamate aspartate transporter (GLAST; guinea pig, 1:4,000; http://www.chemicon.com) or neurofilament NF160 (mouse, 1:200; Sigma) in 5% NGS, 0.3% Triton in PBS for 24 h. For fluorescence analysis, a fluorescein isothiocyanate (FITC)-tagged secondary antibody (http://www.vectorlabs.com), diluted 1:400 in 0.3% Triton/PBS, was used for 1 h. Sections were then washed, mounted on slides, and coverslipped with Vectashield (http://www.vectorlabs.com). For LM, a biotinylated secondary was used, after which the sections were washed and incubated for 1 h in a 1:400 ABC Elite (#PK 6100; http://www.vectorlabs.com) solution diluted in 0.3% Triton/PBS. To visualize the antibody–antigen complex we used the nickel–DAB protocol. Alternatively, after ABC Elite incubation, amplification was carried out by placing sections in biotinylated tyramide for 5 min, followed by incubation for 1 h in a 1:400 FITC-streptavidin solution (http://www.vectorlabs.com) for immunofluorescence or in ABC Elite solution then nickel–DAB reaction. When a light background appeared the reaction was stopped by washing the sections in PBS.

PROTEIN EXTRACTION AND WESTERN BLOT ANALYSIS.

Proteins from trigeminal ganglia were extracted as described previously (Vit et al. 2006, 2008). Protein amount was quantified using the Bio-Rad protein assay (http://www.bio-rad.com). Protein samples (20 μg) were resolved by 10% SDS-PAGE and transferred onto polyvinylidene fluoride membrane. Equal loading and transfer of proteins were monitored by Ponceau-S staining of the membrane. The following antibodies were used: Cx43 and Cx26 (Zymed; 1:500), GFAP (Chemicon; 1:6,000), and activating transcription factor 3 (ATF-3; Santa Cruz Biotechnology; 1:1,000). Immunoblot analysis was carried out according to standard procedures using enhanced chemiluminescence detection (Amersham). Films were scanned and band density was quantified using Image J software. The number of black pixels was counted in a standardized square positioned over each band.

MACROPHAGE ACTIVATION ASSAY.

To determine whether globin and Cx43 dsRNA could induce an inflammatory response, we evaluated their potential capacity to activate the nuclear factor-kappa B (NF-κB). We tested the effect of the dsRNA on RAW–endothelial leukocyte adhesion molecule macrophages, which harbor a luciferase reporter construct driven by NF-κB activation (Hume et al. 2001). The cells were seeded in 96-well plates and treated with either globin or Cx43 dsRNA in concentrations similar to those injected into the trigeminal ganglion (50 mg/ml). Negative control was untreated cells and positive controls were cells treated with 100, 250, and 500 ng/ml lipopolysaccharide (LPS), a potent activator of NF-κB. After a treatment of 6 h, the cells were lysed and the substrate for luciferase was added to the samples. The degree of NF-κB activation was directly proportional to the luminescence intensity read using a luminometer.

RESULTS

Expression of Cx43 in the trigeminal ganglion from intact and nerve injured rats

In the trigeminal ganglion, the cell bodies of neurons are organized in clusters and each neuron is surrounded by SGCs. SGCs are easily identifiable because they express GLAST, a glia-specific glutamate transporter (Fig. 1A and Vit et al. 2006). Cx43 immunolabeling was present throughout the ganglion and within the neuronal clusters was confined to SGCs. Within SGCs, at the LM and EM levels, Cx43 immunolabeling appeared as diffuse, low-level staining throughout the cytoplasm, most visible in the perinuclear region, with small, intensely labeled foci (presumably representing gap junctions) scattered throughout the cytoplasm (Fig. 1, B–H). Double immunolabeling with GLAST showed that Cx43 immunolabeling was confined to SGCs (Fig. 1, F–H).

FIG. 1.

Connexin 43 (Cx43) is expressed by satellite glial cells (SGCs) in the trigeminal ganglion. A: triple label of the trigeminal ganglion showing the cell bodies of sensory neurons (red, NF160) surrounded by SGCs (green, GLAST) and nuclei DAPI (white). B: Cx43 immunostaining (diaminobenzidine [DAB]) is diffuse in SGC cytoplasm with darker punctuate staining (arrowheads) in attenuated regions of the cytoplasm farther from SGC nucleus (arrow). Neutral red counterstain. C: electron micrograph of a primary sensory neuron (N) closely opposed by 2 SGCs (SGC₁, SGC₂). The cytoplasm shows diffuse Cx43 immunostaining with increased density at gap junctions (arrow). The inset is a high magnification of the gap junction showing the junction of the processes from SGC₁ and SGC₂. A SGC₃ surrounding an adjacent neurons is visible. D and E: Cx43 immunofluorescent images from contralateral and ipsilateral trigeminal ganglia 10 days after chronic constriction injury (CCI) of the infraorbital nerve (ION). The contralateral and ipsilateral ganglia are embedded, cut, and processed side by side, then photographed at the same exposure. Both the diffuse labeling (arrows) and number of bright puncta (arrowheads) are increased 10 days following CCI. F–H: double-label immunofluorescence 10 days after a CCI of the ION showing that Cx43 is confined to SCGs as shown by the overlap with the SGC marker glutamate aspartate transporter (GLAST). The arrow locates the same region of the SGC in the Cx43, GLAST, and merged images.

We next determined whether expression of Cx43 in the trigeminal ganglion was changed 10 days after a CCI of the ION when the pain-related behavior is maximal (see following text). Immunocytochemistry showed an increase in Cx43 immunolabeling, both in the diffuse cytoplasmic labeling (day 3, Ipsilateral vs. Contralateral, P = 0.520, n = 3; day 10, Ipsilateral vs. Contralateral, P ≤ 0.001, n = 3) and in the number of brightly fluorescent foci. The increase was confirmed by Western blot analysis that showed a twofold increase in Cx43 expression following a CCI of the ION (Fig. 2 and Vit et al. 2006). ATF-3 and GFAP were also examined in parallel to confirm the presence of nerve injury. ATF-3 is a transcription factor that is a neuronal marker of nerve injury (Tsujino et al. 2000) and GFAP, a common marker of astrocytes in the CNS, is present at low levels in normal SGCs, although its expression increases following nerve injury (Vit et al. 2006). Ten days after a CCI of the ION, both ATF-3 and GFAP expression were increased in the trigeminal ganglion, confirming the presence of injury (Fig. 2, A and B).

FIG. 2.

Cx43 expression increases in trigeminal ganglion SGCs after CCI of the ION. A and B: Western blot analysis shows that the expression of ATF-3 (a marker of nerve injury) and GFAP (a marker of SGC activation) is significantly increased in the trigeminal ganglion on the side ipsilateral to the CCI 3 days postinjury, whereas Cx43 does not increase until 10 days postinjury. A: P0 and P1 represent the unphosphorylated and phosphorylated forms of Cx43, respectively. C, contralateral; I, ipsilateral. B: the expression level corresponds to the number of black pixels of each band counted using Image J. *P < 0.05, ***P < 0.001 compared with contralateral side.

RNAi-mediated reduction of Cx43 dsRNA

We examined whether reducing Cx43 using RNAi would change pain-related behavior in normal and nerve-injured rats. First, we confirmed that Cx43 immunolabeling was reduced in SGCs 3 days after a single dsRNA injection (Fig. 3A) and that Cx43 immunofluorescence had returned to normal levels 14 days following injection (Fig. 3B). To show that the dsRNA did not spread outside the ganglion we examined the central region of the trigeminal root where the CNS glia cells show a high expression of Cx43 as opposed to the peripheral root. The central root showed intense Cx43 immunolabeling (Fig. 3C), indicating that the dsRNA did not spread along the root.

FIG. 3.

Cx43 doubled-stranded (ds)RNA injection transiently reduces expression of Cx43. A: 3 days following Cx43 dsRNA injection there is a reduction in Cx43 immunolabeling both in the diffuse labeling and in the number of bright immunofluorescent foci. B: 14 days after Cx43 dsRNA injection immunolabeling has returned to normal levels. C: 3 days after Cx43 dsRNA injection the immunolabeling of glial cells in the central part (Cen) of the trigeminal root is unchanged, showing that Cx43 dsRNA does not migrate along the root to enter the brain stem. The inset shows a high magnification of labeled glial element in the central portion of the trigeminal root.

To evaluate the degree of Cx43 silencing following injection of Cx43 dsRNA, the expression of Cx43 was measured by Western blot after injection of globin or Cx43 dsRNA into the trigeminal ganglion. Three days following injection of dsRNA (when behavioral changes are maximal; see following text), Western blot analysis showed that Cx43 expression was significantly decreased in the trigeminal ganglion that received the Cx43 dsRNA injection when compared with the noninjected side (RM ANOVA, F = 56.0, P = 0.005; Fig. 4A). Injection of globin dsRNA had no effect on Cx43 expression in the trigeminal ganglion (RM ANOVA, F = 0.1, P = 0.789; Fig. 4A).

FIG. 4.

The inhibition of Cx43 expression using Cx43 dsRNA is specific. A: Western blot analysis of Cx43 expression after injection of either Cx43 or globin dsRNA shows that Cx43 dsRNA led to a dramatic decrease of Cx43 expression on the side ipsilateral to the injection. Globin dsRNA injection has no effect on Cx43 expression. B: Western blot analysis of Cx26 expression, another subunit of gap junctions expressed in the trigeminal ganglion shows that Cx26 expression is not affected by the injection of Cx43 dsRNA. The expression of the marker of nerve injury activating transcription factor 3 (ATF-3) is unchanged after injection of Cx43 dsRNA.

To confirm the specificity of Cx43 silencing, we assessed the expression of a closely related gap-junction subunit, Cx26. In the CNS, Cx26 is expressed in astrocytes (Altevogt and Paul 2004; Nagy et al. 1997, 2003) and in the peripheral nervous system, Cx26 has been shown to be located in the perineurium of the sciatic nerve (Nagaoka et al. 1999). After injection of Cx43 dsRNA into the trigeminal ganglion, there was no difference in expression of the Cx26 subunit between the injected and uninjected sides (Fig. 4B).

We used the expression of ATF-3 to determine whether the injection procedure and/or the Cx43 dsRNA caused nerve injury. Western blot analysis confirmed that the expression of ATF-3 was identical in both the ipsilateral and contralateral sides to the injection of Cx43 dsRNA (Fig. 4B).

Nociceptive behavior following RNAi of Cx43

We examined the nociceptive behavior of the rats after injection of Cx43 or globin dsRNA into the maxillary division of the trigeminal ganglion. Behavioral testing was performed in the somatotopic area innervated by the sensory neurons from the maxillary division.

To test for evoked mechanical nociceptive responses, we stimulated both sides of the snout with 2, 10, and 50 g von Frey hairs. After injection of Cx43 dsRNA, the rats showed increased sensitivity to mechanical stimulation with all von Frey hairs on the ipsilateral side but not contralateral side to the injection (RM ANOVA, ipsilateral: F = 9.9, P < 0.001; contralateral: F = 1.8, P = 0.126; Fig. 5A). The von Frey score was elevated (i.e., increased sensitivity) from day 1 to day 5 after injection, then returned to baseline (nontreated) by day 7. In animals injected with β-globin dsRNA, the von Frey score remained at preinjection levels for the entire experiment (Fig. 5A).

FIG. 5.

Silencing Cx43 in the trigeminal ganglion of normal rats triggers painlike behavioral changes. A: rats show increased sensitivity to von Frey hair stimulation beginning 1 day after injection of Cx43 dsRNA (day 0). This hypersensitivity reaches a maximum at day 3 and is back to baseline by day 7. B: Cx43 dsRNA reduces the number of licks per episode of drinking in an operant conflict paradigm. C: following injection of Cx43 dsRNA, the number of unilateral eye closures ipsilateral to the injection increases to reach a maximum at day 3. A–C: *P < 0.05, ***P < 0.001 compared with globin dsRNA; #P < 0.05, ##P < 0.01, ###P < 0.001 compared with IMPLANT baseline.

We next evaluated pain behavior using an operant conflict paradigm. To drink sugar-sweetened water, the rats had to maintain contact with a stiff-bristled brush on the side of the snout ipsilateral to the injection of dsRNA. The brush did not affect the drinking behavior of normal rats (data not shown). When compared with globin dsRNA-injected animals, rats injected with Cx43 dsRNA drank less (RM ANOVA, F = 4.4, P = 0.002; Fig. 5B) because they would not maintain contact with the brush to reach the drinking spout. The decrease in sugar-sweetened water consumption in Cx43 dsRNA-injected rats started by day 1, reached a maximum by day 3, and was back to normal by day 7 postinjection.

Spontaneous painlike behavior was assessed by the measurement of eye closures ipsilateral and contralateral to the dsRNA injection. This nonevoked behavior is similar to the unilateral wincing seen in patients with trigeminal neuralgia when they experience shooting pain. The number of unilateral eye closures on the side ipsilateral to the injection was increased in rats with knockdown of Cx43 compared with control rats (RM ANOVA, ipsilateral: F = 10.1, P < 0.001; Fig. 5C). As expected, the number of unilateral contralateral eye closures remained unchanged during the entire experiment for both Cx43 and globin dsRNA-injected animals (RM ANOVA, contralateral: F = 1.7, P = 0.142).

Behavioral effects of silencing Cx43 are comparable to those observed after CCI of the ION

To evaluate the magnitude of the effect of silencing Cx43 on painlike behavior, we compared intact Cx43 dsRNA-injected rats with rats that received a CCI of the ION (Fig. 6). The comparison was done when both treatments reached their maximum (i.e., day 3 for Cx43 dsRNA compared with day 10 for CCI). The von Frey scores for both silencing Cx43 and CCI were similar on the ipsilateral side to the treatment (ANOVA, F = 0.1, P = 0.829; Fig. 6A). In the conditional reward test, the reduction in the consumption of sugar-sweetened water was not different between Cx43 dsRNA-injected and injured rats (ANOVA, F = 0.2, P = 0.645; Fig. 6B). We also compared the number of unilateral eye closures ipsilateral to the treatment between rats 3 days after Cx43 dsRNA injection and rats 10 days after CCI. We found that the increase in eye closures was significantly greater in the CCI rats compared with the Cx43 dsRNA-injected rats (ANOVA, F = 18.0, P < 0.001; Fig. 6C).

FIG. 6.

The effects of inhibiting Cx43 expression in the trigeminal ganglion is similar to those observed after a CCI of the ION. The behavioral effects of Cx43 dsRNA (day 3) were compared with those observed after nerve injury (day 10), when both treatments show their maximal effects. A and B: the magnitude of the effects of Cx43 dsRNA and CCI was similar in the von Frey hair testing (A) and the operant conflict paradigm (B). C: the increase of unilateral eye closures ipsilateral to the after nerve-injured side was double that compared with the injection of Cx43 dsRNA. ***P < 0.001 compared with Cx43 dsRNA.

Inflammatory response following Cx43 dsRNA injection

Because it has been shown that dsRNA can trigger an inflammatory response through the activation of NF-κB, we tested the effect of globin and Cx43 dsRNA with a macrophage activation assay. We treated stable macrophage cells engineered to express luciferase under the control of NF-κB with either globin or Cx43 dsRNA. The level of luminescence produced by the cells was similar to that from untreated cells and was significantly less that the luminescence produced by the cells treated with LPS, an activator of NF-κB response (Fig. 7).

FIG. 8.

Silencing Cx43 in the trigeminal ganglion produces an analgesic effect in a model of nerve injury. A: 10 days after a CCI of the ION, the injection of Cx43 dsRNA in the maxillary division of the trigeminal ganglion has no effect on the increased sensitivity to von Frey hair stimulation observed following CCI of the ION. B: the injection of Cx43 dsRNA 10 days after nerve injury leads to a short-lasting increase of sugar-sweetened water consumption in an operant conflict paradigm. After silencing of Cx43 expression, the number of licks per episode of drinking is similar to baseline level. C: the number of unilateral eye closures ipsilateral to the CCI of the ION is largely decreased after injection of Cx43 dsRNA in the trigeminal ganglion ipsilateral to the injury. A–C: *(black and red) P < 0.05, **(black and red) P < 0.01, ***(black and red) P < 0.001 compared with respective baseline (presurgery); #(red) P < 0.05, ##(red) P < 0.01 compared with CCI day 10; #P < 0.05, ##P < 0.01 compared with globin dsRNA.

Behavioral effects of Cx43 dsRNA following CCI

The reduction in nociceptive threshold following Cx43 dsRNA injection in normal rats contrasts with previous studies showing that carbenoxolone, a nonselective gap-junction inhibitor, produces analgesia. In light of these data, we sought to know what would be the effect of injecting Cx43 dsRNA into the trigeminal ganglion after a CCI of the ION. Ten days following a CCI of the ION, rats were injected with either globin or Cx43 dsRNA. Evoked nociceptive responses were tested using von Frey hairs. After CCI of the ION, increased hypersensitivity to mechanical stimulation occurred ≤10 days postoperation on the side ipsilateral to the injury (Fig. 8A). On the side contralateral to the injury, the von Frey score remained at baseline levels unchanged until day 10 (RM ANOVA, F = 0.4, P = 0.739). Following the behavioral testing on day 10, dsRNA was injected into the maxillary division of the trigeminal ganglion on the side of the injury. Neither globin nor Cx43 dsRNA had any effect on the average von Frey score on the side ipsilateral to the injury and to the injection (RM ANOVA, F = 2.3, P = 0.134; Fig. 8A). The use of the average von Frey score hides a slight analgesic effect of Cx43 dsRNA after CCI of the ION. When von Frey hair scores are analyzed individually, injection of Cx43 dsRNA 10 days after CCI produced a short-lasting decrease of von Frey score only with the 2 g hair (see Supplemental Fig. S2).

We next evaluated the effect of silencing Cx43 on CCI-related pain behavior using the operant conflict paradigm. Following CCI, all rats showed a decreased consumption of sugar-water. One day after injection of Cx43 dsRNA drinking values returned to baseline levels in rats (RM ANOVA, F = 6.8, P = 0.006; Fig. 8B), whereas sugar-water consumption in the globin dsRNA-injected animals remained reduced (RM ANOVA, F = 0.6, P = 0.497).

The effect of dsRNA on spontaneous pain behavior in response to injury was also assessed. After CCI of the ION there was an increase in number of unilateral eye closures ipsilateral to the injury (Fig. 8C). The increased number of ipsilateral unilateral eye closures was similar in both globin and Cx43 groups prior to the injection of dsRNA (RM ANOVA, F = 1.0, P = 0.399). On the contralateral side to the injury, the number of unilateral eye closures remained unchanged until day 10 in both groups (RM ANOVA, F = 0.5, P = 0.697). After injection of globin dsRNA (day 10), the number of ipsilateral unilateral eye closures continued to increase until day 13. One day after injection, Cx43 dsRNA led to a dramatic decrease in the number of unilateral eye closures on the side ipsilateral to the injury and injection (RM ANOVA, F = 5.1, P = 0.017). On the contralateral side to the injury and injection of dsRNA, there was no difference between groups in the number of unilateral eye closures (RM ANOVA, F = 0.2, P = 0.831).

DISCUSSION

The present results show that 1) the expression of Cx43 is increased in the maxillary division of the trigeminal ganglion in a model of orofacial neuropathic pain; 2) reducing Cx43 expression in the trigeminal ganglion leads to spontaneous and evoked painlike behavior similar to that seen after nerve injury; and 3) in a model of orofacial neuropathic pain, inhibiting the expression of Cx43 in the trigeminal ganglion decreases spontaneous and evoked painlike behavior.

The preceding findings are based on experiments using RNAi and some specificity issues associated with RNAi are addressed. We have previously shown that with this long dsRNA approach [which serves as an endogenous substrate for dicer, resulting in several different small interfering (si)RNA molecules and is more potent in silencing at protein and mRNA level than a single siRNA species] knockdown occurs at both mRNA and protein levels (Clifton et al. 2007; la Fleur et al. 2005). We tested for specificity of Cx43 suppression by examining the expression of a closely related protein, the Cx26 subunit (Nagaoka et al. 1999), and found that it was unchanged after Cx43 or control dsRNA injection. Nonspecific effects of the dsRNA that could compromise the viability of cells are unlikely because there was no increase in the cell stress response marker, ATF-3, following injection of either Cx43 or globin dsRNA. It has been suggested that certain dsRNAs do not produce their effect through target-specific inhibition but rather by the induction of an inflammatory response (Cunnington and Naysmith 1975; Wang and Carmichael 2004). Double-stranded RNA can potently activate the NF-κB signaling pathway (Alexopoulou et al. 2001; Kumar et al. 1994; Matsumoto et al. 2002), which triggers an inflammatory response by the production of cytokines (Kumar et al. 1997). There is no evidence that an inflammatory response is a factor with the long Cx43 dsRNA we use. Our results show minimal activation of NF-κB with either globin or Cx43 dsRNA. Further, if an inflammatory response was a component of our behavioral and anatomical data, both Cx43 and globin dsRNA should produce similar effects and this was not the case. Based on these considerations the most parsimonious explanation of the data is that the reported changes are a consequence of reduction of Cx43 expression and not a consequence of unrelated, nonspecific effects.

One way in which connexins are believed to alter nociceptive sensation is by increasing the coupling between adjacent glial cells (Cherkas et al. 2004). We did not directly examine glial coupling, but it is reasonable to infer alterations in glial coupling based on our observed changes in Cx43 expression. Cx43 is one of the protein gap-junction subunits specific to astrocytes (Nagy et al. 2004) and our results show that Cx43 is present in SGCs of the trigeminal ganglion. Electrophysiological and dye-injection studies have shown that coupling between SGCs increases following nerve injury (Cherkas et al. 2004; Huang and Hanani 2005; Huang et al. 2005) and is correlated with lowered SGC membrane potential and pain-related behavior. It has been shown that there is increased coupling between astrocytes in the trigeminal nucleus caudalis after formalin injection into the vibrissal pad in rats (Lan et al. 2007) and Cx43 expression increases in astrocytes in the facial nucleus after peripheral axotomy (Rohlmann et al. 1994). In sensory ganglia, coupling between SGCs has been shown to increase after partial colonic obstruction, complete Freund's adjuvant injection into the paw, and after axotomy (Dublin and Hanani 2007; Hanani et al. 2002; Huang and Hanani 2005).

Given the evidence that Cx43 expression increases in peripheral nerves after injury, our finding that reducing Cx43 expression following nerve injury is analgesic is not surprising. Previous studies have reported similar findings. Injection of the gap-junction blocker carbenoxolone reduces the pain behavior of rats in the facial formalin test (Lan et al. 2007) when injected into the cerebellomedullary cistern. Intrathecal injection of carbenoxolone has also been shown to attenuate inflammatory and neuropathic pain (Qin et al. 2006; Spataro et al. 2004). In these latter studies, it is assumed that carbenoxolone is acting centrally by inhibiting the coupling between astrocytes. Other experiments have shown that intraperitoneal injection of carbenoxolone reduces inflammation-induced pain behavior (Dublin and Hanani 2007). In this latter case the effect is thought to be peripherally mediated because carbenoxolone does not cross the blood–brain barrier (Leshchenko et al. 2006). Our results suggest that the site of peripheral action of carbenoloxone could be on SGCs within the sensory ganglion.

One unexpected finding was that Cx43 dsRNA had opposite effects on nociceptive thresholds in uninjured compared with nerve-injured rats. There are several possible explanations of how these changes might occur. One possibility is that after nerve injury increased glial cell coupling causes increased intercellular passage of small molecules such as cAMP, Ca²⁺, and IP3, all of which can induce nociceptive transmission (Dina et al. 2005). Silencing Cx43 would block the intercellular movement of such small molecules by preventing the formation of additional gap junctions.

In contrast, in nerve-intact rats, reducing the number of gap junctions by silencing Cx43 lowers intracellular K⁺ buffering capacity of SGCs (Cherkas et al. 2004) below a critical level, resulting in an increase in extracellular K⁺. This could result in increased excitability of neurons (Djukic et al. 2007; Kocsis et al. 1983; Walz 2000). We recently showed that reducing a specific K⁺ channel (Kir4.1) in the rat trigeminal ganglion leads to a decrease in nociceptive threshold (Vit et al. 2008) that might be due to an increase in extracellular K⁺. Thus the dominating effect of increasing gap junctions following nerve injury might be to increase the movement of nociceptive-related second-messenger molecules, whereas the K⁺ buffering activities are not changed enough to be significant. If this were the case, reduction of a CCI-induced increase in gap junctions by Cx43 RNAi may interfere with the movement of second messengers enough to restore the nociceptive levels to normal. In contrast, in the normal state, the reduction of K⁺ buffering capacity might be the dominant effect because there is no or little movement of second messengers in this condition. These ideas remain to be investigated.

We suggest that a perturbation of Cx43 expression, whether increased or decreased, is sufficient to cause alterations in neuronal firing and subsequent changes in sensory thresholds. Thus in injury states it is not only increased glial cell coupling that causes the change in behavior; rather any change in glial coupling that results in altered neuronal firing can cause behavioral changes. Nerve injury causes an increase in SGC coupling, although there may be other experimental or naturally occurring conditions that reduce coupling and we would expect such a situation also to be associated with increased nociception.

GRANTS

This work was supported by National Institute of Neurological Disorders and Stroke Grants NS-058479 to P. T. Ohara and NS-051336 to L. Jasmin.

FIG. 7.

Long dsRNA-mediated RNA interference (RNAi) does not trigger an inflammatory response. Using an in vitro macrophage inflammation assay, dsRNA (globin, Cx43, and polyriboinosinic:polyribocytidylic acid [poly I-C]) do not induce the activation of the NF-κB inflammatory pathway. This contrasts with the treatment of the cells with lipopolysaccharide (LPS), a potent activator of NF-κB.