Patients with neuropathic pain show reduced endogenous analgesia induced by a conditioned noxious stimulus. Here, the authors tested whether peripheral nerve injury impairs descending noradrenergic inhibition from the locus coeruleus (LC) after L5–L6 spinal nerve ligation (SNL) in rats.
A subdermal injection of capsaicin was used to examine noxious stimulation–induced analgesia (NSIA), evoked LC glutamate and spinal noradrenaline release, and evoked LC neuronal activity in normal and SNL rats. The authors also examined the role of presynaptic metabotropic glutamate receptors or the astroglial glutamate transporter-1 (GLT-1).
SNL increased basal extracellular glutamate concentration in the LC (170.1%; 95% CI, 44.7 to 295.5; n = 15) and basal spinal cord noradrenaline release (252.1%; 95% CI, 113.6 to 391.3; n = 15), which was associated with an increased tonic LC neuronal activity and a down-regulation of GLT-1 in the LC. SNL reduced NSIA (−77.6%; 95% CI, −116.4 to −38.8; n = 14) and capsaicin evoked release of glutamate in the LC (−36.2%; 95% CI, −49.3 to −23.2; n = 8) and noradrenaline in the spinal cord (−38.8%; 95% CI, −45.1 to −32.5; n = 8). Capsaicin-evoked LC neuronal activation was masked in SNL rats. Removing autoinhibition of glutamatergic terminals by metabotropic glutamate receptor blockade or increasing GLT-1 expression by histone deacetylase inhibition restored NSIA in SNL rats. SNL-induced impairment of NSIA was mimicked in normal rats by knockdown of GLT-1 in the LC.
These results suggest that increased extracellular glutamate in the LC consequent to down-regulation of GLT-1 contributes to LC dysfunction and impaired pain-evoked endogenous analgesia after nerve injury.
The authors investigated the circuitry for impaired noxious stimulation–induced analgesia in rats with neuropathic pain. The authors demonstrate glutamate dysregulation in the locus coeruleus, and the resulting impairment of descending inhibitory control contributes to the loss of noxious stimulation–induced analgesia in neuropathic pain.
Patients with chronic pain have an impaired noxious stimulation–induced analgesia (NSIA), the ability to reduce painful stimuli through activation of the endogenous analgesia system
The authors investigated the circuitry for impaired noxious stimulation–induced analgesia (NSIA) in rats with neuropathic pain
The authors demonstrate glutamate dysregulation in the locus coeruleus, and the resulting impairment of descending inhibitory control contributes to the loss of NSIA in neuropathic pain
PERIPHERAL nerve injury can result in chronic neuropathic pain, which is associated with a plethora of changes in the peripheral and central nervous systems related to pain signaling. The gate control theory of pain posits a modulatory role of descending systems on spinal cord sensory processing, which is important to endogenous analgesia.1 Patients with established neuropathic pain have a reduced ability to physiologically recruit descending inhibition,2 and patients with reduced descending inhibition are at greater risk for chronic pain after surgery,3 suggesting that disruption of endogenous analgesia plays a permissive role in chronic pain. However, the underlying mechanisms for disruption of endogenous analgesia remain unclear.
Although endogenous analgesia is mediated by several peripheral and central mechanisms,4–6 descending noradrenergic inhibition is essential to endogenous analgesia and recovery from hypersensitivity after nerve injury in rats.7 As an important endogenous analgesic neurotransmitter, noradrenaline is released in the spinal cord from bulbospinal axons that originate in the locus coeruleus (LC) and adjacent nuclei in the brain stem and suppresses activation of spinal nociceptive neurons by actions on α2-adrenoceptors.8 Among various neurochemical inputs in the LC, glutamate is the primary excitatory neurotransmitter on noradrenergic neurons, acting through α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors.9 Glutamate also inhibits its own release via group 2 and group 3 metabotropic glutamate receptors (mGluRs) located on glutamatergic terminals.10 Extracellular glutamate is classically removed by two types of astroglial glutamate transporters, primarily glutamate transporter-1 (GLT-1) and also glutamate–aspartate transporter.11,12 Knockdown or blockade of GLT-1 increases basal extracellular glutamate concentrations in the LC of normal rats. Although peripheral nerve injury also increases basal extracellular glutamate concentrations in the LC,13 whether this reflects changes in GLT-1 has not been examined. We previously demonstrated that GLT-1 expression decreases in the spinal cord after L5–L6 spinal nerve ligation (SNL), associated with increased extracellular glutamate concentrations, and that inhibition of histone deacetylase (HDAC) restores both GLT-1 and glutamate to normal.14 Based on these results, the current study focused on the LC and tested the hypothesis that peripheral nerve injury increases extracellular glutamate by down-regulating GLT-1 to inhibit evoked glutamate release in the LC via activation of mGluRs and thereby diminishes evoked LC neuronal activity from noxious input important to pain-induced endogenous analgesia.
To measure endogenous analgesia in rats, we used a validated method, noxious stimulation–induced analgesia (NSIA), which is reduced in animals after neuropathic injury and in humans with neuropathic pain.2,4 The current study examined whether SNL impairs NSIA and reduces noxious stimuli–evoked release of glutamate in the LC and subsequent LC neuronal activation and spinal noradrenaline release. We reasoned based on the adaptive gain theory of LC function15 that increasing tonic LC activity after SNL would reduce the phasic response to the spinal cord in response to this noxious stimulation and necessary for NSIA. We tested mechanisms by which these impairments occur by the use of presynaptic mGluRs blockade, HDAC inhibition, and quantification of AMPA receptor subunits, glutamate receptors 1 and 2 (GluR1 and GluR2), in spinally projecting noradrenergic neurons in the LC.
Materials and Methods
Male and female Sprague-Dawley rats (6 weeks old at arrival) from Harlan Industries (USA), housed under a 12-h light–dark cycle with free access to food and water, were used. All experiments were approved by Animal Care and Use Committee at Wake Forest University School of Medicine (Winston-Salem, North Carolina).
Surgeries and LC Cannula Verification
Unilateral L5 and L6 SNL was performed as described previously.16 Briefly, animals were anesthetized with 2.0% isoflurane in oxygen, the right L6 transverse process was removed, and the right L5 and L6 spinal nerves were tightly ligated using 5-0 silk suture. Three to four weeks after SNL surgery, some animals were anesthetized with 2% isoflurane and placed securely in a stereotaxic frame (KOPF, USA). A sterile stainless steel guide cannula (CXG-8 for microdialysis, Eicom Co., Japan, or C315G for drug or small interfering RNA (siRNA) injection, Plastic One, USA) was implanted into the left or right LC as described previously.17 The coordinates for placement of the tip of the guide cannula were 9.8 mm posterior and 1.4 mm lateral to the bregma, and 6.5 mm ventral from the surface of the dura mater, according to the rat brain atlas.18 Animals were allowed to recover from the surgery at least 1 week before the experiment. After the experiment, 175 animals with LC guide cannulae except 18 animals for Western blotting received an intra-LC injection of methylene blue (0.5 μl) and were euthanized by an intravenous or intraperitoneal injection of pentobarbital (150 mg/kg). The brain was removed, postfixed with 4% buffered formaldehyde, and sectioned, and the placement of the cannula was verified visually. In the current study, six rats were excluded from the behavioral and microdialysis experiments (three rats in each experiment) due to cannula malplacement.
Drugs and siRNA Treatments
Animals were randomly assigned to drug or siRNA treatment. For intra-LC drug administration, (2S)-2-Amino-2-[(1S,2S)-2-carboxycycloprop-1-yl]-3-(xanth-9-yl) propanoic acid (LY341495; Tocris Bioscience, USA), 6-cyano-7-nitroquinoxaline-2,3-dione disodium salt (CNQX; Tocris Bioscience), and vorinostat (Tocris Bioscience) were dissolved in 0.033 M NaOH, double distilled water, and dimethyl sulfoxide, respectively, diluted with saline to the final concentrations, and injected through the LC guide cannula at a rate of 0.1 μl/min. For intrathecal drug administration, idazoxan hydrochloride (Tocris Bioscience) was dissolved in saline and injected in the L5–L6 intervertebral space using a 30-gauge needle. Knockdown of GLT-1 in the LC was performed as we previously described.13 A mixture of siRNAs for rat GLT-1 (SMARTpool #M-091209-02, Thermo Fisher Scientific Inc., USA) or a nontargeting siRNA pool (#D-001206-14, Thermo Fisher Scientific Inc.) was dissolved in double distilled water, diluted with the transfection reagent (i-Fect; Neuromics, USA) to achieve a final concentration of 8.3 pmol/0.5 μl, and injected through the LC guide cannula for five consecutive days.
Behavior Test: NSIA
Nociceptive mechanical thresholds in the hindpaw were measured with a Randall-Selitto analgesiometer (Ugo Basile, Italy) as previously described.19 A cutoff pressure of 250 g was used to avoid potential tissue injury. All animals were trained for 3 days with this apparatus before baseline values were measured. NSIA was performed as recently reported.7 Under brief isoflurane anesthesia (2%), capsaicin (150 μg/50 μl/rat; Sigma-Aldrich, USA) or vehicle (50 μl/rat; 50% ethanol in saline) was injected into left forepaw, and the withdrawal thresholds in the left hindpaw (contralateral to SNL injury) were measured. In the current study, no animal showed distress or self-mutilating behavior in the forepaw after capsaicin injection. The authors (M.K. and T.S.) who conducted behavior tests were blinded to treatments and groups.
Western Blotting for GLT-1 in the LC
Western blotting for GLT-1 in the LC was performed as previously reported.13 Under deep isoflurane anesthesia (5%), animals were euthanized by decapitation and the brain stem was quickly removed and placed in ice-cold saline. Brainstem slices (2 mm thickness) containing the LC were obtained using a precision brain slicer (RBM-4000C, ASI Instruments, Inc., USA), and the region of the left LC was carefully dissected under the surgical microscope. The LC and adjacent tissue were homogenized, lysed, and centrifuged for 10 min at 4°C at 1,000g. Protein content in each supernatant was measured using a standard method.20 Samples (25 μg protein) were placed on 10 to 20% gradient gels (Criterion Tris-HCl Gel; Bio-Rad, USA), run at 100 V for 1 h, and transferred to polyvinylidene difluoride membrane (Bio-Rad). The membrane was blocked with 1% bovine albumin serum in Tris-buffer saline containing 0.1% Tween 20 and incubated overnight at 4°C with a guinea pig anti-GLT-1 (1:5,000, AB1783; Millipore, USA) or a rabbit α-tubulin (1:5,000, 2125S; Cell Signaling, USA). After washing with Tris-buffer saline containing 0.1% Tween 20, the membrane was incubated for 1 h at room temperature with a corresponding horseradish peroxidase–conjugated secondary antibody (1:5,000; anti-guinea pig or 1:5,000; anti-rabbit, Santa Cruz Biotechnology, Inc., USA), treated for 1 min with West Pico hemiluminescence substrate (Thermo Fisher Scientific Inc.), and exposed to x-ray film (Kodak BioMax film, Sigma-Aldrich). The density of each specific band was measured using a computer-assisted imaging analysis system (MCID Elite7.0 image analysis system, Imaging Research Inc., Canada).
Microdialysis for LC Glutamate and Spinal Noradrenaline
Simultaneous microdialysis in the spinal cord and LC was performed according to our previous reports with minor modifications.13,14 On the day of the experiment, anesthesia was induced with 2% isoflurane and then maintained with 1.5% isoflurane during the study. A heating blanket was used to maintain rectal temperature at 36.5° ± 0.5°C, and the right jugular vein was cannulated for saline infusion (3.2 ml kg−1 h−1). For microdialysis in the left lumbar spinal dorsal horn, the L3–L6 level of spinal cord was exposed by the T13–L1 laminectomy. Microdialysis probes (CX-I-8-01, outer diameter = 0.22 mm, inner diameter = 0.20 mm, length = 1 mm; Eicom Co.) were inserted into the left spinal dorsal horn and left LC 1 h before the experiment and perfused with Ringer’s solution (1.0 μl/min). Then, after two 30-min baseline samples, capsaicin, at the same dose used for NSIA, or vehicle was injected into the left forepaw. Glutamate and noradrenaline contents in the microdialysates were measured by separate high-pressure liquid chromatography systems with electrochemical detection (HTEC-500, Eicom Co.), as we previously reported.14,17
Capsaicin-evoked Neuronal Activation in the LC.
Under brief isoflurane anesthesia (2%), capsaicin, at the same dose used for NSIA, or vehicle was injected into the left forepaw. Thirty minutes later, animals were euthanized with an intraperitoneal injection of pentobarbital (150 mg/kg) and perfused with 0.01 M phosphate-buffered saline containing 1% sodium nitrite, followed by 4% formalin in 0.1 M phosphate-buffered saline. The brain stems were dissected out, postfixed in the same fixative, cryoprotected with 30% sucrose in 0.1 M phosphate-buffered saline, and sectioned at a 16 μm thickness. Immunostaining for phosphorylated cyclic adenosine monophosphate response element binding protein (pCREB) and dopamine-β-hydroxylase (DβH) in the LC was performed as previously described.17 After being pretreated with 1% normal donkey serum (Jackson Immuno Research Laboratories, USA), the sections were incubated with a rabbit monoclonal anti-pCREB antibody (1:500, 06-519; Millipore) and a mouse monoclonal anti-DβH antibody (1:500, MAB 308; Millipore) overnight. Subsequently, the sections were incubated with Cy2 conjugated anti-mouse IgG (1:200, Jackson Immuno Research Laboratories) and Cy3 conjugated anti-rabbit IgG (1:600, Jackson Immuno Research Laboratories).
Retrograde Tracing and Immunostaining for GluR1/2 in LC Neurons.
T13–L1 laminectomy and intraspinal injection were performed as previously described.21 Red-fluorescent beads (F8793; Life Technologies, USA) were filled in a glass micropipette (10 to 20 μm tip diameter) and four pairs of bilateral injections (50 nl/site, each 500 μm apart in the rostrocaudal axis) were performed at 500 μm from the midline and 500 μm under the surface of the dorsal spinal cord using a micro injector (Nanoinject, Drummond Scientific Co., USA). Seven days after injection, the brainstem sections were prepared as described in the previous paragraph. The sections were pretreated with 3% normal donkey serum, incubated with a combination of rabbit monoclonal anti-GluR1 antibody (1:500, 04-855; Millipore) with mouse monoclonal anti-DβH antibody (1:500), or mouse monoclonal anti-GluR2 antibody (1:1,000, MAB397; Millipore) with rabbit monoclonal anti-tyrosine hydroxylase (TH) antibody (1:500, P40101-0; Pel-Freez Biologicals, USA) overnight. Subsequently, the sections were incubated with donkey anti-rabbit or anti-mouse biotin (1:500, Jackson Immuno Research Laboratories), followed by Cy2 conjugated streptavidin (1:200, Jackson Immuno Research Laboratories) and AlexaFluor 647 conjugated anti-mouse or anti-rabbit IgG (1:500, Jackson ImmunoResearch Laboratories).
For quantification, four to five brainstem sections containing the LC were randomly selected from each rat and digital images were captured using a Nikon Eclipse Ni fluorescent microscope system (Nikon, Japan). All images for each experiment were taken at the same time with the same camera settings, and the authors performing the image analysis (M.K. and C.E.M.-U.) were blinded to treatment or group. Positive immunostaining for pCREB, GluR1, or GluR2 was defined using a constant threshold for each protein applied across all sections. DβH- or TH-immunoreactive (IR) cells with or without pCREB, GluR1, or GluR2 immunostaining were counted in the entire LC from four to five sections per rat. Results are reported as % of total DβH- or TH-IR cells per rat. DβH or TH-IR cells containing fluorescent beads were defined as spinally projecting LC neurons.
A limitation of the current study is that power analysis was not conducted before the study to determine appropriate sample sizes. Data are presented as mean ± SD. Behavioral data and microdialysis data were analyzed by two-way repeated-measures ANOVA where time and group (surgery or treatment) were considered as independent variables, and Tukey post hoc test was used for multiple comparisons using SigmaPlot software (Systat Software Inc., USA). Other data were analyzed by one-way or two-way ANOVA followed by Tukey post hoc test with two-tailed hypothesis. P value less than 0.05 adjusted by SigmaPlot software was considered significant.
Role of LC-spinal Noradrenergic Pathway in NSIA.
In intra-LC (left side) and intrathecal saline-treated normal rats, forepaw injection of capsaicin into the left forepaw significantly increased withdrawal thresholds in the left hindpaw 30 min after injection compared with vehicle (fig. 1), demonstrating NSIA. This NSIA effect was significantly reduced by intra-LC injection of the AMPA antagonist CNQX, indicating importance of glutamatergic signaling in the LC. NSIA was also significantly reduced by intrathecal injection of the α2-adrenoceptor antagonist idazoxan, indicating the importance of spinal noradrenaline release. Neither intra-LC CNQX nor intrathecal idazoxan affected withdrawal thresholds in forepaw vehicle-injected rats. Consistent with our recent observation,7 these results confirm that activation of LC-spinal noradrenergic pathway after capsaicin injection is essential to NSIA in normal rats.
Impact of SNL on NSIA in Male and Female Rats.
In male and female rats, there was a significant main effect of group on withdrawal thresholds in the right hindpaw ipsilateral to SNL (fig. 2A; F5, 48 = 35.47; P < 0.001). Post hoc testing revealed that normal female rats showed slightly but significantly lower withdrawal thresholds in the right hindpaw than those of normal male rats, that SNL resulted in mechanical hypersensitivity in the ipsilateral hindpaw 2 to 6 weeks after surgery in male rats and 6 weeks after surgery in female rats, and that withdrawal thresholds in the ipsilateral hindpaw 6 weeks after SNL showed no sex difference (P = 0.92).
In male rats treated with forepaw injection of capsaicin (fig. 2B), there was a significant main effect of time (F3, 102 = 39.97; P < 0.001) on the withdrawal thresholds in the contralateral hindpaw and a significant group × time interaction (F9, 102 = 3.59; P < 0.001), but no main effect of group (F3, 102 = 2.13; P = 0.12). Post hoc testing revealed that significant NSIA was observed only at 30 min after injection and that the group × time interaction reflected a progressive reduction in NSIA as a function of time since SNL injury. Similarly, in female rats, SNL resulted in a reduction of NSIA compared with normal rats (group: F1, 14 = 123.40, time: F1, 14 = 116.05, group × time: F1, 14 = 15.73; P < 0.01 for all F values). There was no sex difference in NSIA in either normal (P = 0.60) or SNL conditions (P = 0.49). Based on this lack of sex difference in SNL-induced hypersensitivity and impairment of NSIA, we used male rats 6 weeks after SNL for the following studies.
Presynaptic Inhibition on LC Glutamate Release after SNL.
An intra-LC injection of the presynaptic mGluR antagonist LY341495 had no effect on withdrawal threshold ipsilateral to injection in normal rats (P = 0.68), but increased it bilaterally and dose dependently in SNL rats, an effect reversed by intrathecal idazoxan (fig. 3A). Also in SNL animals, intra-LC injection (left) of a low dose (0.4 μg) of LY341495, which did not alter withdrawal threshold alone (P = 0.55), restored NSIA (fig. 3B). These results demonstrate SNL-induced autoinhibition of glutamatergic terminals in the LC, consistent with tonically elevated extracellular glutamate concentrations in the LC after SNL.
Down-regulation of GLT-1 in the LC after SNL.
Expression of GLT-1 protein in the LC from saline-treated SNL rats was significantly lower than that of normal rats (fig. 4A; P < 0.01). Intra-LC vorinostat treatment significantly increased GLT-1 expression (P = 0.03) and NSIA (P = 0.04) (fig. 4B) compared with saline in SNL rats, while not affecting baseline withdrawal thresholds in the hindpaw contralateral to SNL (vorinostat: 127 ± 10 g vs. saline: 129 ± 20 g; P = 0.77; n = 9 in each group), consistent with our previous observations in the spinal cord that the reduced GLT-1 expression after SNL was reversed by HDAC inhibitors.14,22 In normal rats, repeated LC injections of GLT-1 siRNA, at a dose that reduces GLT-1 expression and increases basal extracellular glutamate level in the LC,13 mimicked the effect of SNL on NSIA, while not affecting baseline withdrawal threshold (GLT-1 siRNA: 139 ± 12 g vs. control siRNA: 135 ± 16 g; P = 0.52; n = 9 in each group). These results suggest that down-regulation of GLT-1 in the LC plays a critical role for impaired NSIA after SNL.
Increased Tonic and Decreased Evoked LC Activity after SNL
LC Glutamate and Spinal Noradrenaline Release.
Basal glutamate concentrations in microdialysates were significantly greater from the left LC of SNL rats (2.7 ± 2.3 ng/30 μl; n = 15) than those of normal rats (1.0 ± 0.8 ng/30 μl; P = 0.01; n = 14), consistent with current behavioral results that demonstrate autoinhibition of glutamatergic terminals in the LC after SNL via mGluRs. Similarly, basal noradrenaline concentrations in microdialysates from the left lumbar spinal dorsal horn of SNL rats (4.0 ± 2.5 pg/30 μl) were significantly greater than those of normal rats (1.5 ± 0.7 pg/30 μl; P < 0.005). These results suggest that SNL increases tonic activity of spinally projecting LC neurons.
For capsaicin-evoked glutamate release in the LC, there were significant main effects of group (F3, 50 = 11.64; P < 0.001), time (F2, 50 = 11.08; P < 0.001), and the group × time interaction (F6, 50 = 3.01; P = 0.014) (fig. 5A). For capsaicin-evoked noradrenaline release in the spinal cord, there was a significant main effect of group (F3, 50 = 19.69; P < 0.001), but no main effect of time (F2, 50 = 0.55; P = 0.58) or group × time interaction (F6, 50 = 1.56; P = 0.18) (fig. 5B). Capsaicin-evoked release of glutamate in the LC and noradrenaline in the spinal cord were reduced in SNL rats compared with normal rats at peak time points (LC glutamate at 30 min: −36.2%; 95% CI: −49.3 to −23.2 and spinal noradrenaline at 60 min: −38.8%; 95% CI: −45.1 to −32.5).
LC Neuronal Activity.
Photomicrographs in figure 6A depict pCREB-IR in noradrenergic LC neurons, identified by DβH-IR, in normal and SNL rats treated with forepaw injection of vehicle or capsaicin at the same dose used for NSIA. For pCREB-IR in the LC, there was a significant main effect of group (F3, 40 = 31.07; P < 0.001), but no main effect of side (F1, 40 = 0.00003; P = 0.996) or the group × side interaction (F3, 40 = 0.11; P = 0.96) (fig. 6B). Post hoc testing revealed that vehicle-treated SNL rats showed an increase in pCREB-IR in LC neurons compared with vehicle-treated normal rats (P < 0.001 for both sides) and that capsaicin increased pCREB-IR in LC neurons compared with vehicle in normal rats (P < 0.001 for both sides), whereas no such increase occurred in SNL rats (P = 0.81), consistent with current NSIA and the microdialysis results. These results suggest that SNL bilaterally increases tonic but reduces evoked LC neuronal activity, effectively taking the LC “off-line.”
Expression of AMPA Receptor Subunits in the LC
Photomicrographs in figure 7, A and B depict GluR1-IR and GluR2-IR in spinally projecting LC neurons, identified by the spinally injected retrograde tracer-positive cells with DβH-IR or TH-IR, in normal and SNL rats. Since GluR1- and GluR2-IR showed no difference between the sides of LC from SNL rats in our preliminary experiment (data not shown), the current study summed cell counts from left and right LC to obtain a single value for each rat. For GluR1-IR in the LC, there was a significant main effect of surgery (F1, 18 = 12.94; P < 0.01), but no main effect of tracer (F1, 18 = 0.53; P = 0.48) or the surgery × tracer interaction (F1, 18 = 0.11; P = 0.74) (fig. 7C). Post hoc testing revealed that SNL significantly increased GluR1-IR in LC neurons compared to normal (fig. 7D; P < 0.01). For GluR2-IR in the LC, there was no significant main effect of surgery (F1, 20 = 1.85; P = 0.19), tracer (F1, 20 = 0.88; P = 0.36), or surgery × tracer interaction (F1, 20 = 0.06; P = 0.81) (fig. 7D). These results suggest that SNL increases rather than decreases expression of GluR1 in overall LC neurons and do not support the idea that down-regulation of AMPA receptors in LC neurons contributes to impaired NSIA after nerve injury.
Peripheral nerve injury leading to hypersensitivity is associated with a plethora of changes in pain processing at many levels of the nervous system, from the primary sensory afferent to the spinal cord to the brain stem and cortex. In the past 15 yr, pain research has primarily focused on facilitatory changes in pain processing after nerve injury, including spinal neural and glial activation and descending facilitation, yet the clinical relevance of these phenomena in animals is unknown. By contrast, descending inhibition important to endogenous analgesia has been shown to be reduced in patients with neuropathic pain,2 yet its mechanisms have been less studied. Interestingly, although LC activity in neuropathic pain may facilitate supraspinal pain pathways,4–6 many drugs approved to treat neuropathic pain, including gabapentinoids, opioids, noradrenaline re-uptake inhibitors, and clonidine, activate, augment, or mimic the LC-spinal descending noradrenergic pathway to produce analgesia,17,23–25 suggesting that this pathway is a key pharmacological target to treat neuropathic pain. As presented in the scheme (fig. 8), the current results demonstrate that down-regulation of astroglial glutamate transporters in the LC after nerve injury dysregulates descending noradrenergic activation, which is important to endogenous and exogenous analgesia.
To measure endogenous analgesia in rats, we chose forepaw capsaicin injection remote from nerve injury as the stimulus because this models conditioned pain modulation in humans.2,3,6 Our results with this test in SNL rats mirror the reduction in conditioned pain modulation in patients with chronic neuropathic pain.2 We recognize we could have applied the conditioning noxious stimulus to the nerve injured hindpaw,26,27 but it would be difficult to match the nociceptive input from this site after SNL to that in the normal animal in which peripheral and central sensitization are not present. We also recognize that all behavioral measures in the current study reflect threshold withdrawal to an external stimulus. Although this is especially relevant to NSIA, future studies should examine more complex behaviors that may better reflect pain itself. The current and previous studies showed that unilateral peripheral nerve injury affects both LCs, demonstrated by a bilateral increase in neuronal activation markers in the LC after injury.17,28,29 These results provide a rationale to study LC dysregulation in the side contralateral to injury.
Previous studies in rats demonstrated that noxious stimuli evoke glutamate release in the LC to activate noradrenergic neurons.9,30 The current study confirmed these observations by demonstrating that forepaw injection of capsaicin in normal rats increased LC glutamate and spinal noradrenaline release and LC neuronal activity, consistent with NSIA which was reversed by the blockade of AMPA receptors in the LC or α2-adrenoceptors in the spinal cord. By contrast, SNL diminished these effects of capsaicin, while it increased basal extracellular LC glutamate and spinal noradrenaline levels and tonic LC neuronal activity. Although several peripheral and central mechanisms are involved in endogenous analgesia,4–6 these results suggest that the descending noradrenergic pathway is essential to pain-evoked endogenous analgesia in normal conditions and that SNL bilaterally increases tonic activity of the LC-spinal noradrenergic pathway and impairs evoked LC activity and NSIA, effectively taking the LC “off-line.” This is consistent with previous work in animals showing that cortical responses to nonnoxious sensory stimuli are modulated by tonic LC activity, with reduced response in the presence of high tonic activity,31 consistent with the adaptive gain concept.15 Similar effects of LC tonic activity as inferred by pupillometry in humans on the role of the LC in decision-making32,33 support this theory.
In contrast to the stable hypersensitivity in the ipsilateral hindpaw, NSIA was reduced in the contralateral hindpaw in a delayed manner, present at 4 and 6 weeks, but not at 2 weeks after SNL surgery. A previous study demonstrated that sciatic nerve injury in rats results in increased bursting activity of LC neurons in a delayed manner, 4 weeks after injury.34 Similarly, SNL increased pCREB expression in the LC and basal extracellular noradrenaline concentrations in the spinal cord 6 weeks after nerve injury. By contrast, early after peripheral nerve injury (1 to 2 weeks), there is no change in spontaneous discharge rate of LC neurons.26,27 These results suggest the delayed appearance of functional plasticity in tonic and evoked LC neuronal activity after nerve injury. They also suggest that phasic responses of the LC to noxious events which occur during activity daily living may be more important to recovery from injury than increased tonic activity, even though both processes result in norepinephrine release in the spinal cord. Future studies should examine this relationship more directly via LC neuronal recording to confirm the inferred activity from pCREB expression.
Similar to our previous observations in the spinal cord,14,22 the current results showed that SNL down-regulated GLT-1 expression and increased basal extracellular glutamate concentrations in the LC. Although glutamate is the primary excitatory neurotransmitter on noradrenergic neurons in the LC,9 increased synaptic concentrations of glutamate can activate presynaptic mGluRs to autoinhibit glutamate release from its terminals.10 As such, intra-LC mGluR blockade had no effect on withdrawal threshold in normal rats, but restored impaired NSIA and reduced hypersensitivity in SNL rats, and this is dependent on downstream release of noradrenaline in the spinal cord, acting on α2-adrenoceptors. Restoring down-regulated GLT-1 expression in the LC by HDAC inhibition restored NSIA in SNL rats. Knockdown of GLT-1 by a selective siRNA, which increases basal extracellular glutamate concentrations in the LC,13 mimicked the effect of SNL on NSIA in normal rats. These results strongly suggest a causal relationship between increased basal extracellular glutamate concentrations in the LC and impairment of NSIA after SNL. Taken together, the current observations support the hypothesis that peripheral nerve injury increases extracellular glutamate by down-regulating GLT-1 to inhibit evoked glutamate release in the LC via activation of mGluRs, thereby diminishing evoked LC neuronal activity important to NSIA.
Based on previous data in the rat, GluR1, GluR2, GluR3, and GluR4 AMPA receptor subunits are all expressed in the LC.35 The current study also examined an alternative hypothesis that impaired NSIA after nerve injury is due to reduced expression of AMPA receptor subunits in noradrenergic LC neurons, rendering them less responsive to glutamate. However, SNL increased GluR1 expression and did not affect GluR2 expression in both spinally projecting and other LC neurons. Although we did not examine GluR3-4 subtypes and expression of phosphorylated GluR1, which is a more relevant marker for AMPA receptor activity,36,37 the current results suggest that down-regulation of AMPA receptors in the LC after SNL is unlikely to occur.
In summary, down-regulation of astroglial glutamate transporters in the LC is critical to impairment of pain-evoked endogenous analgesia after peripheral nerve injury in rats. Since the descending noradrenergic pathway is not only essential to endogenous analgesia but also an important target for many drugs approved to treat neuropathic pain,17,23–25 restoring impaired responsiveness of descending noradrenergic pathway may have considerable clinical significance.
This work was supported by grants DA27690 (to Dr. Hayashida) and GM99863 (to Dr. Peters) from the National Institutes of Health (Bethesda, Maryland).
The authors declare no competing interests.