Introduction

In addition to altering neuronal networks, peripheral nerve injury leads to the activation of spinal glial cells, resulting in increased production of inflammation- and pain-related molecules, including pro-inflammatory cytokines (tumor necrosis factor- (TNF), interleukin (IL)-1, IL-6), nitric oxide (NO), prostaglandins, and ATP (for reviews, see refs. 1, 2). Activated glial cells thus appear to be an important spinal site for the production of algogenic mediators, and several mitogen-activated protein kinase pathways have been suggested to participate in the functional changes of glial cells.^{3, 4, 5} The nuclear factor B (NF-B) system may contribute to these changes as it plays a key role in the regulated production of pro-inflammatory cytokines and in the cytokine-induced expression of major proteins of inflammatory and immune responses (for review, see ref. 6). In addition to the involvement of NF-B in inflammation,^7,8 several recent reports suggest the NF-B pathway has a more direct pain-related role. For example, in models of peripheral inflammation–induced pain or pharmacologically induced central pain, intrathecal administration of NF-B inhibitors resulted in transitory attenuation of pain.^{9, 10, 11} Repeated systemic administration of an inhibitor of the NF-B–associated kinase complex (IB kinase), which prevents NF-B activation, also reduced hyperalgesia in rats.¹²

NF-B is a pleiotropic factor also involved in transcriptional control of a wide variety of genes. Convergent studies demonstrated that it participates in central nervous system homeostasis, playing an important antiapoptotic and antiexcitotoxic role in neurons (for review, see refs. 13 and 14). Therefore, to evaluate the role of NF-B in functional changes of glial cells, in the increased production of pain-related molecules, and in chronic neuropathic pain, cell category non-discriminating inhibition of NF-B appears inadequate and we hypothesized that a viral vector–targeted blockade of NF-B restricted to spinal glial cells would be a valuable approach.

In most cells, NF-B is kept inactive in the cytoplasm by an inhibitory protein of the IB family. Numerous stimuli may trigger the activation of IB kinase, which phosphorylates IB, leading to its ubiquitination and proteasomal degradation. Subsequently, NF-B translocates into the nucleus and induces transcription of target genes. IB thus offers one possible means to block the NF-B pathway. In particular, the "super-repressor" (sr) IB, a degradation-resistant form of IB, has been shown to block NF-B activity.¹⁵ We generated pseudotyped lentiviral (LV) vectors to drive expression of srIB, and we assessed the possibility of a local and glial cell–oriented blockade of the NF-B pathway as well as its impact in a well-characterized rat model of neuropathic pain.

Results

SrI B production from LV-srIB

Infection of 293T cells with LV-srIB resulted in an accumulation of IB that was strongly attenuated when cells were incubated in the presence of reverse-transcriptase inhibitor AZT (3'-azido-3'-deoxythymidine) (Figure 1). Production of LV-srIB–derived IB-like immunoreactive material (IB-IR) was further demonstrated in both 293T cells and rat glial cell primary cultures (data not shown). Western blot analysis of total proteins extracted from glial cells infected with increasing concentrations of LV-srIB revealed viral titer–dependent production of IB protein (Figure 1).

Figure 1.

LV-srIB drives the production of srIB in glial cells in vitro. Control (non-infected) or LV-srIB-infected (300 of p24/ml) glial cell primary cultures were incubated (or not) in the presence of 10 M reverse-transcriptase inhibitor AZT. The presence of mRNA encoding sr-IB was assessed using semi-quantitative RT-PCR on 0.5 g total RNA extracted from cell cultures 48 hours after infection. srIB specific amplification was compared with amplification from control GPDH mRNA. Production of srIB protein was further verified in glial cells infected with LV-srIB (300 or 30 ng of p24/ml) using western blot analysis, which showed a viral titer–dependent accumulation of srIB.

Full figure and legend (9K)

LV-srIB prevents lipopolysaccharide-evoked NF-B intranuclear translocation

In lipopolysaccharide (LPS)-treated mixed glial cell cultures, NF-B–like immunoreactive material (NF-B-IR) massively translocated from the cytoplasm to the nucleus (Figure 2a,b, in red). In contrast, in glial cells infected with LV-srIB (Figure 2c, IB-IR in green), LPS failed to induce NF-B translocation, as demonstrated by the cytoplasmic localization of NF-B-IR (Figure 2d, in red) and its co-localization with IB-IR (Figure 2e, in yellow). This cytoplasm-restricted localization of NF-B-IR under LPS activation was observed in all cells infected with LV-srIB. Infection of glial cells with equivalent concentrations of control LV–enhanced green fluorescent protein (LV-EGFP) vector (Figure 2f, green fluorescence) did not affect the LPS-induced nuclear translocation of NF-B (Figure 2g, in red), supporting the specific inhibitory effect of LV-srIB. The co-localization of transgene-derived EGFP and NF-B-IR accumulating in nuclei of LPS-stimulated cells was demonstrated in double-labeling experiments (Figure 2h, in yellow). Western blot analysis of nuclear fraction proteins extracted from glial cell culture (Figure 2i) confirmed that pre-treatment of cells with LV-srIB completely inhibited the LPS-induced nuclear translocation of NF-B. The potency of LV-srIB to inhibit NF-B–induced gene expression was demonstrated in the luciferase reporter gene assay. LPS stimulation of glial cells co- transfected with plasmid bearing 2 X NFB-FLuc reporter gene and plasmid containing Renilla luciferase gene driven by herpes simplex virus thymidine kinase promoter resulted in a selective induction of NFB-FLuc expression. Compared with controls (non-infected or control LV-EGFP–infected cells), infection of cells with LV-srIB 48 hours before LPS addition specifically inhibited the evoked 2 X NFB-FLuc reporter gene activity (Figure 2j).

Figure 2.

LV-srIB–mediated overexpression of srIB in primary cultures of glial cells blocks intranuclear translocation of NF-B and luciferase reporter gene activity evoked by LPS. (a) In control (untreated) glial cell cultures, NF-B–like immunoreactive material (NF-B-IR, p65 subunit, in red) is distributed throughout the cytoplasm. (b) After treatment with LPS (1 g/ml, 1 hour) NF-B-IR is detected mainly in cell nuclei. (c) Forty-eight hours after infection with LV-srIB (300 ng of p24/ml), the majority of the cells (bold arrows) showed transgene-derived IB-IR (in green). Dashed arrows point to cells without detectable IB-IR (presumably non-infected). (d) Double-labeling experiments revealed that LPS treatment evoked NF-B-IR (in red) translocation and accumulation only in the nucleus of uninfected cells, devoid of IB-IR (dashed arrow). In contrast, in cells overproducing IB-IR, nuclear translocation of NF-B-IR was completely abolished and NF-B-IR was sequestered in the cell cytoplasm (bold arrow). After stimulation with LPS, LV-EGFP pre-treated glial cell cultures (300 ng of p24/ml, 48 hours; EGFP green fluorescence) (e) showed strict nuclear localization of NF-B-IR (in red) (f). Panels (g) and (h) represent combined IB-IR/NF-B-IR immunofluorescence and EGFP fluorescence/NF-B immunofluorescence, respectively. Scale bar = 20 m. (i) In a comparable set of experiments, we analyzed nuclear protein fraction by western blotting. LPS-induced accumulation of NF-B-IR in nuclear fraction was prevented in LV-srIB–infected cells. (j) In cells transfected with 2 X NFB-FLuc reporter gene plasmid and TK-RLuc plasmid, and infected (or not) with control LV-EGFP, LPS treatment (1 hour) induced Luc activity. On the other hand, infection of transfected cells with LV-srIB completely prevented the LPS-induced Luc activity. Results represent ratios of luminescence from the NFB-FLuc to TK-RLuc luminescence (means SEM from three independent experiments).

Full figure and legend (59K)

Semi-quantitative reverse-transcriptase polymerase chain reaction (RT-PCR) experiments performed on RNA extracted from glial cells (Figure 3) showed that a 3-hour incubation of cell cultures in the presence of LPS resulted in a large increase of IL-6, IL-1, TNF, and inducible nitric oxide synthase (iNOS) messenger RNA (mRNA) concentrations. Infection of cells with LV-srIB 48 hours before LPS stimulation prevented or significantly reduced the LPS-induced enhancement of cytokines and iNOS mRNA levels compared with controls. Comparable results were obtained in LV-srIB–infected glial cells stimulated for 1 hour with TNF (data not shown). On the other hand, infection of cells with the same amount of LV-EGFP had no inhibitory effect on LPS-evoked accumulation of any of the analyzed mRNA (Figure 3).

Figure 3.

Infection of glial cell cultures with LV-srIB reduces LPS-induced expression of pro-inflammatory cytokines and iNOS. Semi- quantitative RT-PCR was performed with specific primers on 0.5 g of total RNA extracted from control, LV-EGFP, and LV-srIB–infected cells incubated with LPS (1 g/ml, 3 hours). Data (n = 3 for each group) are expressed in arbitrary units representing 260 nm optical density of cytokines or iNOS specific RT-PCR products/260 nm optical density of GPDH RT-PCR products. Data represent means SEM of three independent experiments.

Full figure and legend (18K)

LV vectors allow dorsal horn–restricted transgene expression in glial cells

Saline or control LV-EGFP (60 ng of p24) micro-injection using a glass capillary needle induced limited lesion (250 m rostro-caudally) of the spinal cord. As shown in Figure 4a, injection of 2 l LV-EGFP into the right dorsal horn of the spinal cord resulted 1 week later in EGFP fluorescence strictly restricted to the dorsal horn parenchyma. EGFP production remained constant throughout the 4 weeks of the experimental procedure and was still present 6 months after injection (data not shown). In the lumbar enlargement, transgene-derived protein was visualized approximately 2–3 mm rostral and 2–3 mm caudal to the site of injection (Figure 4b). One week after LV-EGFP injection, most of the EGFP staining co-labeled with the astrocyte-specific marker glial fibrillary acidic protein (GFAP) (Figure 4c–e). Numerous EGFP-positive cells also co-stained for the microglial marker Ox42 (Figure 4f–h). However, only 8.1% of cells were double-labeled with EGFP and the neuronal marker NeuN (Figure 4i–k). This pattern of labeling remained similar throughout the 4 weeks of observation (data not shown). Moreover, in neurons, EGFP fluorescence was always weaker. These in vivo observations are comparable to those from in vitro experiments using primary neuron/glia (50%/50%) co-cultures prepared from rat spinal cord. Indeed, infection of cell co-cultures with LV-EGFP also resulted in EGFP fluorescence localized mainly in GFAP-positive astrocytes and rarely in NeuN-positive neurons (data not shown).

Figure 4.

In vivo injection of lentiviral-derived vectors into the rat spinal cord allowed dorsal horn–restricted expression of transgene preferentially in glial cells. (a) Micro-injection of LV-EGFP resulted in EGFP fluorescence restricted to the ipsilateral dorsal horn of the spinal cord. (b) Single injections of viral suspension resulted in a rostro-caudal distribution of the vector through approximately 5 mm. Immunofluorescence experiments performed with antibodies raised against specific markers of astrocytes (GFAP), microglia (Ox42) or neurons (NeuN) showed that most EGFP fluorescent cells (c,d,e) contained also GFAP-LM (f) or Ox42-LM (g). Only scarce cells with weak EGFP fluorescence (dashed arrow, e) were colabeled with NeuN antibodies (h). Panels i, j and k represent combined fluorescence for EGFP and GFAP, Ox-42 or NeuN immunofluorescence, respectively. Scale bar = 50 m or 20 m in small windows. Estimated from spinal cords of four LV-EGFP–injected rats, neuronal profiles containing EGFP represented approximately 8% of total EGFP expressing cells (graph in k).

Full figure and legend (103K)

Sensitive RT-PCR experiments failed to detect EGFP transcripts in right lumbar (L3–L5) dorsal root ganglion or brainstem, suggesting that neurons projecting to the dorsal spinal cord were not infected by LV vectors (data not shown).

Similarly, injection of LV-srIB (60 ng of p24) into the right dorsal horn of the spinal cord resulted in accumulation of IB-IR restricted to the ipsilateral side of the injection (Figure 5a), and the pattern of IB-IR distribution was comparable to the pattern of EGFP staining in control LV-EGFP–infected animals (astrocytes and microglial cells; data not shown). Western blot analysis performed 1 week after viral vector injection revealed overproduction of IB protein in the spinal cord dorsal horn of rats injected with LV-srIB (Figure 5b).

Figure 5.

LV-srIB delivery into the rat spinal cord resulted in IB-IR accumulation in the dorsal horn of the spinal cord. (a) IB-IR–positive cellular profiles were observed only in LV-srIB–infected right (ipsilateral) dorsal horn 2 weeks after vector injection. Scale bar = 25 m. (b) Western blot analysis of protein extracts from the right dorsal part of the lumbar spinal cord (L3–L5) of naïve, control LV vector–injected rats or animals (two distinct rats) injected with LV-srIB was performed 1 week after injection. The blot was successively incubated with IB and -tubulin antibodies.

Full figure and legend (21K)

srIB inhibits NF-B nuclear translocation and attenuates IL-6 and iNOS expression

Western blot analysis of nuclear fraction proteins extracted from ipsilateral lumbar dorsal spinal cord showed that, compared with sham animals, chronic constriction injury (CCI) induced an early (after 6 hours) NF-B nuclear translocation that was still present 48 hours after CCI (Figure 6a). As shown in Figure 6b, CCI-evoked enhanced NF-B nuclear translocation (48 hours after the nerve lesion) was inhibited in animals infected with LV-srIB (60 ng of p24) 1 week before nerve lesion.

Figure 6.

Chronic constriction injury (CCI)–induced nuclear accumulation of NF-B in the dorsal lumbar spinal cord was abolished in rats injected with LV-srIB. (a) Western blot analysis of nuclear fraction proteins extracted from the right dorsal part of the lumbar spinal cord (L3–L5) of naïve animals or animals 6 and 48 hours after CCI. (b) In animals injected with LV-srIB (CCI-LV-srIB, two distinct rats) and subjected 1 week later to CCI, western blot analysis revealed complete inhibition of NF-B accumulation in nuclear extracts (48 hours after the nerve lesion) from the right dorsal horn of the spinal cord. The blot was successively incubated with NF-B and -tubulin antibodies.

Full figure and legend (8K)

Real-time RT-PCR experiments (Figure 7) showed that 7 days after the sciatic nerve constriction, relative concentrations of IL-6 and IL-1 mRNA in the dorsal horn of the spinal cord (ipsilateral to the lesion) were approximately 7- (P < 0.005, n = 3) and 2.5-fold (P < 0.05, n = 3) higher, respectively, than in naïve rats (n = 3). On the other hand, TNF and iNOS mRNA levels remained comparable to those measured in naïve rats (data not shown). Interestingly, 4 weeks after CCI, iNOS mRNA levels were markedly increased in the dorsal horn of the spinal cord, almost eightfold (P < 0.02, n = 3) those measured in naïve rats (n = 4). At this time point, IL-6 and IL-1 mRNA levels were no longer significantly different from those determined in the dorsal spinal cord of naïve rats (data not shown). Whereas IL-1 mRNA levels remained unaffected by viral vector injection (P = 0.59, n = 4), the CCI-induced increased levels of IL-6 mRNA at 7 days (P < 0.005, n = 4) and the enhanced concentration of iNOS mRNA at 28 days after sciatic nerve injury (P < 0.02, n = 3) were almost completely suppressed by LV-srIB treatment (Figure 7a,b).

Figure 7.

Injection of LV-srIB into the dorsal spinal cord prevented chronic constriction injury (CCI)–evoked enhanced expression of IL-6 and iNOS but not of IL-1 in the dorsal lumbar spinal cord. Real-time RT-PCR was performed on total RNA extracted from the right dorsal horn of the lumbar (L3–L5) spinal cord of naïve, CCI, or LV-srIB–treated CCI rats (n = 3–4 for each group). (a) IL-6 and IL-1 mRNA relative quantities 7 days after CCI. LV-srIB treatment prevented the CCI-induced accumulation of IL-6 mRNA. (b) The late increase of iNOS mRNA relative concentrations observed 28 days after CCI was also abolished in LV-srIB–treated CCI rats. Each sample, PCR-amplified in triplicate, was normalized with GPDH as reporter gene. Data represent mean SEM.

Full figure and legend (9K)

srIB attenuates CCI-evoked thermal hyperalgesia and mechanical and cold allodynia

Hindpaw skin temperatures, measured with a JADE Middle Wave Infra-Red (3–5 m) camera, did not significantly differ between naïve (27.5 1.5 °C), CCI (contralateral paw 27.2 1.6 °C; ipsilateral paw 27.0 1.4 °C), and spinal cord injected CCI (contralateral paw 27.5 0.7 °C; ipsilateral paw 27.3 0.7 °C) rats. Injection of LV-EGFP or LV-srIB into the dorsal horn of the spinal cord of control (non-CCI) animals did not affect their sensitivity to mechanical or thermal stimulations. The right hindpaw (lesioned side) of CCI animals showed reduced paw withdrawal latencies (PWLs) to heat stimulation compared with baseline latency (before CCI; Figure 8a). Indeed, a significant (P < 0.001) and constant thermal hyperalgesia was established at 7 days (PWL = -4.02 0.42 seconds, n = 10), 14 days (PWL = -4.45 0.55 seconds, n = 10), and 21 days (PWL = -4.51 0.32 seconds, n = 10) after the sciatic nerve constriction. Starting from the first time point of behavioral evaluation (i.e., 7 days after sciatic nerve constriction), rats infected with LV-srIB showed significantly reduced CCI-evoked thermal hyperalgesia (PWL = -2.48 0.30 seconds, P < 0.02, n = 18) compared with paired control-CCI rats. This reduction of CCI-induced thermal hyperalgesia in rats injected with LV-srIB persisted during the whole observation period, i.e., for 3 weeks (14 days PWL = -1.98 0.42 seconds, P < 0.005, n = 15; 21 days PWL = -2.22 0.39 seconds, P < 0.01, n = 15). Two-factor analysis of variance revealed a significant effect of treatments (control, control-CCI, and LV-srIB-CCI; F = 68.7, P < 0.0001) but neither a significant effect of time (7 days, 14 days, 21 days; F = 0.016, P = 0.98) nor a treatment–time interaction (F = 0.44, P = 0.78).

Figure 8.

Rats injected with LV-srIB into the dorsal horn of the spinal cord exhibited attenuated chronic constriction injury (CCI)–induced thermal hyperalgesia and mechanical and cold allodynia. Behavioral studies were performed in control (representing both naïve and LV-EGFP–injected animals), control-CCI (representing both naïve and LV-EGFP–injected rats with CCI surgery), and LV-srIB–infected CCI rats. (a) Nociceptive responses elicited by radiant heating were measured as paw withdrawal latencies (PWLs) in seconds, and for each animal PWL was calculated as the difference between PWL at days 7, 14, and 21 and the baseline PWL (day 0 before CCI). Seven days after LV vector spinal injection (day 0 before CCI), PWLs were comparable between different groups. Starting at day 7 and then throughout the experimental procedure (21 days), PWLs were significantly different between control and control-CCI groups (P < 0.0001). Note that single LV-srIB injection resulted in constant significant antihyperalgesic effect throughout the 21 days of experiment. (b) Sensitivity to mechanical stimulation was assessed using von Frey filaments. The mean of first reaction (increasing testing) and lowest test result (decreasing testing) was taken as the mechanical paw withdrawal threshold (PWT). These data were log transformed, and the percentage decrease of the withdrawal threshold was then calculated in relation to the withdrawal threshold before surgery (baseline): PWT = (CCI paw - baseline paw)/baseline paw 100. (c) Reactivity of animals to cold stimulus was measured after application of acetone (100 l) onto the right footpad. The time that animals spent with paw withdrawn (PWt) was measured during the next 120-second period. Data represent mean SEM.

Full figure and legend (14K)

Mechanical hypersensitivity to stimulation with von Frey filaments was fully developed 14 days after the lesion (7 days PWT = -3.1 2.2%, n = 8; 14 days PWT = -53.8 5.4%, n = 8; 21 days PWT = -42.8 2.6%, n = 8) compared with the baseline thresholds before CCI (Figure 8b). In comparison with paired control-CCI rats, LV-srIB–treated animals exhibited reduced mechanical allodynia 14 days (PWT = -33.6 3.3%, P < 0.01, n = 8) and 21 days (PWT = -24.7 6.8%, P < 0.02, n = 7) after CCI.

Finally, as shown in Figure 8c, 7, 14, and 21 days after CCI, right ipsilateral paws were hypersensitive to cold stimuli (duration of paw withdrawal (PWt) at 7 days = 86.8 8.1 seconds, n = 7; 14 days PWt = 74.8 9.1 seconds, n = 7; 21 days PWt = 35.7 4.6 seconds, n = 7). Infection of CCI rats with LV-srIB reduced significantly the nerve injury–evoked cold allodynia at 7 days (PWt = 59.8 9 seconds, P < 0.05, n = 9) and 14 days (PWt = 48.5 7.1 seconds, P < 0.05, n = 9) compared with paired control-CCI rats. This effect was no longer significant 21 days after CCI (PWt = 27.3 7.2 seconds, P > 0.05, n = 9).

Discussion

The present study shows that local, targeted LV-mediated production of srIB, inhibiting NF-B activity strongly preferentially in glial cells of the dorsal horn of the spinal cord, resulted in prolonged antihyperalgesic and antiallodynic effects in a rat model of CCI-induced neuropathic pain. Furthermore, the sciatic nerve injury–induced enhanced intranuclear translocation of NF-B in the spinal cord was inhibited after local administration of LV-srIB, which also prevented the CCI-evoked early accumulation of pro-inflammatory cytokine IL-6 mRNA and the delayed increase of iNOS mRNA.

In contrast to the known anti-inflammatory potency of NF-B inhibition (for review, see ref. 6), several recent reports have suggested that this transcription factor might also be related to exaggerated pain. For example, the NF-B inhibitor pyrrolidinedithiocarbamate, a free radical scavenger with moderate specificity, showed an antiallodynic effect in models of pain evoked by intrathecal dynorphin or perisciatic zymosan administration.^9,11 Intrathecal delivery of NF-B "decoy" oligodeoxynucleotides also revealed NF-B to be implicated in peripheral inflammation–associated hyperalgesia.¹⁰ Repeated intraperitoneal administration of a specific inhibitor of IB kinase also showed antiallodynic potency in a CCI model of neuropathic pain.¹² However, in these studies, the contribution of peripheral compared with central inhibition of NF-B and the impact of the anti-inflammatory effect of the NF-B blockade remain unclear. Moreover, the approaches used in these latter studies lead to global NF-B blockade without considering the clearly deleterious consequences of NF-B inhibition in neurons,^14,16 making it difficult to evaluate the specific role of NF-B in spinal glial cells and its implication in neuropathic pain, as assessed in our study. Using a spinal micro-injection procedure and LV-derived vector, we were able strongly and with an important selectivity to inhibit NF-B activity in glial cells of the dorsal spinal cord.

Consistent with previous reports showing that primo-injection of LV vectors into the brain evoked a negligible inflammatory reaction,¹⁷ we found that direct injection of LV vector into the rat spinal cord induced only minor injury of the spinal parenchyma without glial activation and scar formation far from the injection site. The harmlessness of intraspinal delivery was further supported by the lack of behavioral alterations in rats injected with saline or control LV-EGFP vector. Although human immunodeficiency virus (HIV)-1–derived vectors pseudotyped with vesicular stomatitis virus envelope and conveying cytomegalovirus (CMV) promoter were originally reported to transduce mostly neuronal cells in brain structures,^17,18 several recent reports have also demonstrated an important proportion of infected astrocytes.^19,20 Combining fluorescence and immunohistofluorescence studies, we show here that injection of LV-(CMV)-EGFP into the dorsal horn of the rat spinal cord led to strong and preferential EGFP expression in astrocytes and microglial cells, with only approximately 8% of neurons expressing EGFP. Our observations fit well with a recent report showing that administration of similar vesicular stomatitis virus–pseudotyped HIV-1–derived vector into the spinal cord mainly resulted in transduction of glial cells,²¹ suggesting that the vector's tropism may be different in distinct regions of the central nervous system.

Although CCI of the sciatic nerve has been reported to be associated with NF-B activation in dorsal root ganglion neurons and in Schwann cells,²² we also observed accumulation of NF-B-IR in dorsal spinal cord nuclear extracts of CCI rats, which suggests a rapid and prolonged spinal activation of NF-B as well as an early implication of NF-B in the spinal response to peripheral nerve injury. At the spinal level, activation of NF-B has also been reported in peripheral inflammation–induced pain and after intrathecal administration of HIV envelope glycoprotein gp120, which evokes behavioral changes characteristic of ongoing pain.^{10, 11, 12} Interestingly, after spinal cord injury associated with NF-B activation, transgenic mice expressing the IB inhibitory protein under the control of the GFAP promoter show a complete loss of NF-B signaling and of upregulation of pro-inflammatory cytokines at the spinal level, suggesting that, in such conditions, NF-B activation occurred essentially in spinal cord astrocytes.²³

Peripheral nerve injury is also associated with rapid and enhanced production of pro-inflammatory cytokines in the spinal cord, where they clearly contribute to nociceptive processing (for review, see ref. 24). In agreement with a previous study,²⁵ IL-1 and IL-6 mRNA levels in the dorsal spinal cord were strongly enhanced 7 days after CCI. However, we observed that CCI was associated with a strong but delayed increase of iNOS mRNA concentration 4 weeks after nerve lesion. We thus assessed both the early and prolonged possible effects of LV-srIB administration on the spinal expression of these proteins. Interestingly, LV-srIB injection almost completely prevented the CCI-induced increase of IL-6 mRNA concentrations in the dorsal spinal cord but not that of IL-1 mRNA. Even though in vitro observations do not always translate to in vivo conditions, we found that the intrinsic efficacy of LV-srIB could not explain this marked differential effect in vivo, as LV-srIB prevented with similar efficacy the LPS (or TNF)-evoked expression of both cytokines in primary glial cell cultures. Our data rather suggest that CCI-induced IL-1 overproduction may be mediated/compensated for by a distinct signaling pathway or that IL-1 expression was induced in cell types different from those infected with LV-srIB. Participation of an NF-B–independent mechanism in the regulation of IL-1 expression was also shown in human synoviocytes, where adenoviral vector–mediated production of IB inhibited IL-6 but not IL-1 overproduction.²⁶ The p38 mitogen-activated protein kinase, known to participate in the control of IL-1 production, may represent such a parallel pathway independent of NF-B signaling,²⁷ especially as p38 mitogen-activated protein kinase is activated in spinal glial cells after peripheral nerve injury.⁴ Given that the CCI-induced overproduction of IL-1 precedes that of IL-6,²⁵ that IL-1 is an important inductor of IL-6, and that IL-1 requires NF-B binding sites to induce IL-6 production in several cell types, including rat astrocytes,^{28, 29, 30} we can speculate that the CCI-evoked IL-6 overexpression in spinal glia is NF-B dependent and that IL-1 represents an important link.

In addition to cytokine expression upregulation, we observed that CCI was also associated with a strong but delayed increase of iNOS mRNA concentrations in dorsal spinal cord 4 weeks after nerve lesion. At the spinal level, overproduction of iNOS expression also has been demonstrated in activated astrocytes after peripheral inflammation.³¹ In fact, the presence of reactive astrocytes producing NO in response to the induced iNOS expression, in particular by IL1- and TNF, appears to be a hallmark of traumatic, neurotoxic, or inflammatory brain injury (for review, see ref. 32). The CCI-associated late increase of iNOS mRNA concentrations was also markedly reduced in LV-srIB–injected rats, suggesting the involvement of an NF-B–dependent pathway, and supporting the long-term efficacy of LV-srIB. These latter data are consistent with previous reports showing that in vitro the induction of transcription of the iNOS gene in human primary astrocytes by various cytokines is NF-B dependent and requires IL1-,³³ and that in vivo intrathecal administration of IL1- induces iNOS expression in rat spinal cord.³⁴

The LV-srIB–mediated strong inhibition of NF-B together with the almost complete prevention of CCI-induced increase of IL-6 and iNOS expression at the spinal level strongly supports our behavioral results. Indeed, both IL-6 and iNOS are known to participate in pain processing. In addition to directly sensitizing nociceptors and spinal cord neurons,^35,36 IL-6 stimulates glial cells through a positive feedback loop, promoting its own production and participating in glial activation.^37,38 The nitric oxide synthase product NO has also been shown to play an important role in spinal nociceptive processing (for review, see ref. 39). Spinal glia–derived NO, in particular, has been proposed to participate in long-term presynaptic facilitation of primary afferent fibers.⁴⁰ In addition to the enhanced expression of spinal iNOS after peripheral inflammation,³¹ the participation of spinal NO in inflammatory pain is further supported by data showing that iNOS contributes to the late phase of thermal hyperalgesia.⁴¹ Our demonstration of a peripheral nerve injury–associated increase of iNOS mRNA concentrations in the dorsal spinal cord suggests that the putatively glial cell–derived NO might also participate in CCI-induced prolonged pain.

An important finding of our study is that local inhibition of NF-B activity in glial cells of the dorsal spinal cord was sufficient to attenuate the thermal hyperalgesia and mechanical allodynia that develop after CCI of the sciatic nerve. The antihyperalgesic and antiallodynic effects of LV-srIB, mediated, at least in part, through the prevention of IL-6 and iNOS overexpression, remained constant for 21 days after lesion. A moderate but significant antiallodynic effect to cold stimulus was also observed 7 and 14 days after CCI but not 21 days after the nerve injury. LV-mediated transgene production was sustained throughout the 4 weeks of experimental procedure and was still detected 6 months after spinal administration. However, further experiments are necessary to evaluate the long-term impact of NF-B blockade in spinal glial cells. Although LV-srIB treatments strongly inhibited the CCI-associated increase of NF-B activity and the enhanced expression of IL-6 and iNOS, single administrations of LV-srIB attenuated but did not completely abolish nociceptive behavior in CCI rats. In fact, in addition to the persistent overexpression of IL-1, known as a potent hyperalgesic agent,^42,43 NF-B–independent signaling pathways may also have contributed to the spinal plastic changes that participate in the development of neuropathic pain.^4,44

The activation of the NF-B signaling pathway in neurons has been shown to have important consequences for neuronal survival and plasticity.^14,16 The effect of its activation in glial cells is, however, less clear. It was thus important to assess its role in the production of glial pro-algesic molecules and its impact on neuropathic pain. Our study shows that selective inhibition of NF-B activity in glial cells of the dorsal spinal cord resulted in prolonged antihyperalgesic and antiallodynic effects, possibly through the prevention of CCI-associated expression of IL-6 and iNOS. Thus, our results demonstrate the active role of this glial pathway in exaggerated pain states that develop after a peripheral nerve lesion.

Materials and Methods

Plasmids. The expression plasmid pTrip-CMV-WPRE⁴⁵ (gift from Dr. Hamid Mammeri, UMR CNRS 7091, Paris, France) was used to produce LV vectors. The plasmid containing the IB super-repressor (srIB, i.e., IB in which Ser-32 and Ser-36 were substituted by alanine) was a gift from Dr. Anning Lin (University of Chicago, IL). The coding sequence of srIB or EGFP was inserted (BamHI/XhoI) under the transcriptional control of CMV promoter into pTrip-CMV-WPRE. The encapsidation plasmid, p8.91, and the vesicular stomatitis virus-G–encoding plasmid, pMD-G, have been described previously.⁴⁶

Lentiviral vector production. Pseudotyped HIV vector encoding EGFP (LV-EGFP) was produced as described previously.¹⁹ Because of the capacity of srIB to interfere with HIV-derived vector production,⁴⁷ the protocol for LV-srIB production was slightly modified. Co-transfection of 293T cells with p8.91 and pTrip-srIB was performed in the proportion 2:1. The S1 supernatants with LV vectors were harvested 40 hours later and stored at 4 °C. Cultures were further incubated in fresh medium, and 24 hours later (64 hours after transfection), S2 supernatants were harvested and pooled with S1 before final vector concentration. LV-srIB titers were thus increased at least 10-fold. Viral suspension was titrated and normalized for the capsid p24 antigen, assayed by enzyme-linked immunosorbent assay to approximately 35 ng of p24/l (Beckman Coulter, Roissy, France). The efficacy of LV-EGFP and LV-srIB in driving the expression of transgene-derived respective mRNA was confirmed in infected 293T cell cultures treated or not with 10 M reverse-transcriptase inhibitor AZT (Sigma-Aldrich, Saint Quentin Fallavier, France).

Glial cell primary cultures and neuron/glia co-cultures. Primary mixed glial cultures were prepared, as described in Supplementary Materials and Methods, from the cerebral cortex or the spinal cord of 4-day-old rat pups (Sprague-Dawley, Centre d'Elevage R. Janvier, Le Genest-St. Isle, France) following the procedure of Goslin et al.⁴⁸ with slight modifications. Spinal cord neuron/glia co-cultures were prepared, as described in Supplementary Materials and Methods, using essentially the same protocol as for glial cells, from embryonic (E17–18) spinal cords.

Luciferase reporter gene assay. The luciferase reporter gene assay was performed on primary glial cells using the Dual-Glo Luciferase assay system (Promega) as described in Supplementary Materials and Methods.

Animal treatments and behavioral analysis. Rats (200 g) were deeply anesthetized with chloral hydrate (400 mg/kg, i.p.). The thoracic T13 vertebra⁴⁹ was accessed through skin (20 mm) and dorsal muscle (10 mm) incisions under semi-sterile conditions. Surgery was then continued under a Zeiss operation microscope (10–25). A hole of 1 mm diameter was carefully drilled through the vertebra without disrupting protective layers of the spinal cord. A hole of approximately 150 m was then opened through dura mater and arachnoid mater with a micro-fine point scalpel. Saline or LV vectors were delivered using an automatic micro-injection device (KDS 310; KD Scientific, Holliston, MA) via a heat-pulled glass capillary needle (external diameter 60 m) introduced into the spinal cord parenchyma at an angle of 50°, parallel with the sagittal plane. Two microliters of vector (approximately 60 ng of p24) was injected at a rate of 0.5 l/min, 0.5 mm aside from the midline of the cord and at a depth of 0.8 mm. The needle was then left for 2 minutes in the parenchyma before gentle withdrawal. Muscles were sutured using resorbable 4/0 ethicon stitches, and regular 5/0 ethicon perma-hand sutures were used to close the skin (Johnson and Johnson, New Brunswick, NJ). If necessary, CCI⁵⁰ surgery was performed 1 week later. Thermal hyperalgesia, mechanical and cold allodynia were evaluated as described in Supplementary Materials and Methods. At the end of the behavioral experiments, correct placement of the injection capillary needle was verified by histological examination. All experiments were performed in conformity with the institutional guidelines, which are in compliance with national and international law and policies for use of animals in neuroscience research (European Communities Council directive No. 87848, October 1987, Ministère de l'Agriculture et de la Forêt, Service Vétérinaire de la Santé et de la Protection Animale; Permission No. 75-1179 to M.P.).

Antibodies. Primary antibodies used in this study included goat anti-NF-B p65 and rabbit anti-IB (1:100 and 1:250; Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti-glial fibrillary acidic protein (anti-GFAP; 1:150; DAKO, Glostrup, Denmark), monoclonal anti-Ox42 (1:100; Serotec, Oxford, UK), and monoclonal anti-NeuN (1:1000; Chemicon International, Temecula, CA). Secondary antibodies were CY3-conjugated donkey anti-goat immunoglobulin G (IgG) and Alexa 488-conjugated goat anti-rabbit IgG (1:800 for cells or 1:400 for tissues; Interchim, Montlucon, France). Slides with cells or tissue sections were observed and images were generated using a Leica DMRD microscope.

Quantification of infected neurons. The proportion of neurons expressing EGFP was determined in spinal cords from four LV-EGFP-injected rats at different (1- to 4-week) intervals after viral injection. Twenty-micrometer serial sections were prepared from the lumbar enlargement of the spinal cord, and every fourth section was stained and examined. A total of 40 sections were analyzed with Photoshop software (Adobe Systems ver. 7.0, San Jose, CA) and the proportion of infected neurons was determined by counting the total number of EGFP-stained cells (total infected cells) and those co-labeled with the neuronal marker NeuN (infected neurons).

Conventional RT-PCR and real-time RT-PCR analyses. Rats were killed by decapitation and the lumbar enlargement of the spinal cord (L3–L5) was divided into left and right parts and then into their dorsal and ventral zones in cold conditions (0–4 °C). Tissues were frozen immediately in liquid nitrogen and stored at -80 °C until they were used. Total RNA was extracted using the Nucleospin RNA II Purification Kit (Macherey-Nagel, Hoerdt, France) and its concentration was evaluated by optical density at 260 nm. Conventional RT-PCRs were performed on 0.5 g of each RNA sample in the presence of 40 pmol of specific primers using the Access RT-PCR system (Promega, Charbonnières, France) as described in Supplementary Materials and Methods. For the real-time PCR application, first-stranded cDNA synthesis (0.5 g total RNA per 20 l reaction) was carried out with the SuperScrip III reverse-transcriptase and random primers (ribosomal phosphoprotein at 0.25 g per reaction), as recommended by the manufacturer (Invitrogen, Cergy Pontoise, France). Real-time PCR amplification of each sample in triplicate was performed on the ABI Prism 7300 apparatus according to the manufacturer's protocol (Applied Biosystems, Courtaboeuf, France) using ABgene ABsolute QPCR ROX Mix (ABgene, Courtaboeuf, France) and the Assays-on-Demand Gene Expression probes (Applied Biosystems) for target genes: IL-6 (Rn00561420-m1), IL-1 (Rn00580432-m1), TNF (Rn00562055-m1), iNOS (Rn00561646-m1), and glyceraldehyde-3-phosphate dehydrogenase (GPDH) (Rn99999916-s1). To perform semi-quantitative studies, GPDH was used as reporter gene.

Western blot analysis. Total and nuclear proteins were extracted from glial cells or from the dorsal part of the rat lumbar spinal cord as described in ref. 12. Equal concentrations of proteins, as determined by Bio-Rad protein assay (Bio-Rad, Paris, France), were mixed with standard Laemmli buffer, sonicated, heated at 95 °C for 5 minutes, then separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (12% acrylamide) and electro-transferred (Trans-Blot SD, Bio-Rad) onto a nitrocellulose membrane (Bio-Rad). Membranes were saturated in blocking solution (5% non-fat dry milk, 0.1% Tween 20 in PBS 1) for 1 hour at room temperature and then incubated (overnight, 4 °C) with primary antibodies directed against IB, NF-B p65 (1:250; Santa Cruz Biotechnology, Tebu, Le Perray en Yvelines, France), or -tubulin (1:10,000; Amersham Biosciences, Paris, France) in the blocking solution. After rinsing in blocking solution, blots were incubated (40 minutes at room temperature) with horseradish peroxidase–linked secondary antibodies (anti-goat IgG (1:10,000; Interchim), anti-rabbit IgG (1:5,000), or anti-mouse IgG (1:25,000; Sigma-Aldrich). Blots were finally washed in PBS containing 0.1% Tween 20, and then in PBS. Membranes were processed with the ECL Plus kit and exposed to MP-ECL film (Amersham Biosciences).

Statistical analyses. Data are presented as means SEM. One-way analysis of variance used for RT-PCR data revealed statistically significant differences between different groups (control, LPS, LV-srIB + LPS; F = 34.7, P < 0.0001 for IL-6; F = 23, P < 0.0005 for IL-1; F = 7.2, P < 0.02 for TNF; F = 5.1, P < 0.02 for iNOS). Fisher PLSD was used as post hoc test. Behavioral data were subjected to Student's t-test or to the two-factor analysis of variance followed by unpaired Fisher's PLSD tests. Statistical evaluation was carried out using the StatView 5.2 software (Abacus Concepts, Berkeley, CA). When P > 0.05, the corresponding difference was considered to be non-significant.

References

REFERENCES

1. Watkins, LR, Milligan, ED and Maier, SF (2003). Glial proinflammatory cytokines mediate exaggerated pain states: implications for clinical pain. Adv Exp Med Biol 521: 1–21. | PubMed | ChemPort |
2. Tsuda, M, Inoue, K and Salter, MW (2005). Neuropathic pain and spinal microglia: a big problem from molecules in "small" glia. Trends Neurosci 28: 101–107. | Article | PubMed | ISI | ChemPort |
3. Ma, W and Quirion, R (2002). Partial sciatic nerve ligation induces increase in the phosphorylation of extracellular signal-regulated kinase (ERK) and c-Jun N-terminal kinase (JNK) in astrocytes in the lumbar spinal dorsal horn and the gracile nucleus. Pain 99: 175–184. | Article | PubMed | ChemPort |
4. Jin, SX, Zhuang, ZY, Woolf, CJ and Ji, RR (2003). p38 mitogen-activated protein kinase is activated after a spinal nerve ligation in spinal cord microglia and dorsal root ganglion neurons and contributes to the generation of neuropathic pain. J Neurosci 23: 4017–4022. | PubMed | ISI | ChemPort |
5. Zhuang, ZY, Wen, YR, Zhang, DR, Borselo, T, Bonny, C, Strichartz, GR et al. (2006). A peptide c-Jun N-terminal kinase (JNK) inhibitor blocks mechanical allodynia after spinal nerve ligation: respective roles of JNK activation in primary sensory neurons and spinal astrocytes for neuropathic pain development and maintenance. J Neurosci 26: 3551–3560. | Article | PubMed | ChemPort |
6. Makarov, SS (2000). NF-kappaB as a therapeutic target in chronic inflammation: recent advances. Mol Med Today 6: 441–448. | Article | PubMed | ISI | ChemPort |
7. Auphan, N, DiDonato, JA, Rosette, C, Helmberg, A and Karin, M (1995). Immunosuppression by glucocorticoids: inhibition of NF-kappaB activity through induction of I kappaB synthesis. Science 270: 286–290. | Article | PubMed | ISI | ChemPort |
8. Yin, MJ, Yamamoto, Y and Gaynor, RB (1998). The anti-inflammatory agents aspirin and salicylate inhibit the activity of I(kappa)B kinase-beta. Nature 396: 77–80. | Article | PubMed | ISI | ChemPort |
9. Laughlin, TM, Bethea, JR, Yezierski, RP and Wilcox, GL (2000). Cytokine involvement in dynorphin-induced allodynia. Pain 84: 159–167. | Article | PubMed | ISI | ChemPort |
10. Lee, KM, Kang, BS, Lee, HL, Son, SJ, Hwang, SH, Kim, DS et al. (2004). Spinal NF-B activation induces COX-2 upregulation and contributes to inflammatory pain hypersensitivity. Eur J Neurosci 19: 3375–3381. | Article | PubMed |
11. Ledeboer, A, Gamanos, M, Lai, W, Martin, D, Maier, SF, Watkins, LR et al. (2005). Involvement of spinal cord nuclear factor kappaB activation in rat models of proinflammatory cytokine-mediated pain facilitation. Eur J Neurosci 22: 1977–1986. | Article | PubMed |
12. Tegeder, I, Niederberger, E, Schmidt, R, Kunz, SGühring, H, Ritzeler, O et al. (2004). Specific inhibition of IkappaB kinase reduces hyperalgesia in inflammatory and neuropathic pain models in rats. J Neurosci 24: 1637–1645. | Article | PubMed | ISI | ChemPort |
13. Karin, M and Lin, A (2002). NF-kappaB at the crossroads of life and death. Nat Immunol 3: 221–227. | Article | PubMed | ISI | ChemPort |
14. Bhakar, AL, Tannis, LL, Zeindler, C, Russo, MP, Jobin, C, Park, DS et al. (2002). Constitutive nuclear factor-kappaB activity is required for central neuron survival. J Neurosci 22: 8466–8475. | PubMed | ISI | ChemPort |
15. Purcell, NH, Tang, G, Yu, C, Mercurio, F, DiDonato, JA and Lin, A (2001). Activation of NF-kappaB is required for hypertrophic growth of primary rat neonatal ventricular cardiomyocytes. Proc Natl Acad Sci USA 98: 6668–6673. | Article | PubMed | ChemPort |
16. Fridmacher, V, Kaltschmidt, B, Goudeau, B, Ndiaye, D, Rossi, FM, Pfeiffer, J et al. (2003). Forebrain-specific neuronal inhibition of nuclear factor-kappaB activity leads to loss of neuroprotection. J Neurosci 23: 9403–9408. | PubMed | ISI | ChemPort |
17. Naldini, L, Blomer, U, Gage, FH, Trono, D and Verma, IM (1996). Efficient transfer, integration, and sustained long-term expression of the transgene in adult rat brains injected with a lentiviral vector. Proc Natl Acad Sci USA 93: 11382–11388. | Article | PubMed | ChemPort |
18. Blomer, U, Naldini, L, Kafri, T, Trono, D, Verma, IM and Gage, FH (1997). Highly efficient and sustained gene transfer in adult neurons with a lentivirus vector. J Virol 71: 6641–6649. | PubMed | ISI | ChemPort |
19. Zennou, V, Serquera, C, Sarkis, C, Colin, P, Perret, E, Mallet, J et al. (2001). The HIV-1 DNA flap stimulates HIV vector-mediated cell transduction in the brain. Nat Biotechnol 19: 446–450. | Article | PubMed | ChemPort |
20. Baekelandt, V, Clayes, A, Eggermont, K, Lauwers, E, De Strooper, B, Nuttin, B et al. (2002). Characterization of lentiviral vector-mediated gene transfer in adult mouse brain. Hum Gene Ther 13: 841–853. | Article | PubMed | ChemPort |
21. Pezet, S, Krzyzanowska, A, Wong, LF, Grist, J, Mazarkis, ND, Georgievska, B et al. (2006). Reversal of neurochemical changes and pain-related behavior in a model of neuropathic pain using modified lentiviral vectors expressing GDNF. Mol Ther 13: 1101–1109. | Article | PubMed |
22. Ma, W and Bisby, MA (1998). Increased activation of nuclear factor kappa B in rat lumbar dorsal root ganglion neurons following partial sciatic nerve injuries. Brain Res 797: 243–254. | Article | PubMed | ChemPort |
23. Brambilla, R, Bracchi-Ricard, V, Hu, WH, Frydel, B, Bramwell, A, Karmally, S et al. (2005). Inhibition of astroglial nuclear factor B reduces inflammation and improves functional recovery after spinal cord injury. J Exp Med 202: 145–156. | Article | PubMed | ISI | ChemPort |
24. Marchand, F, Perretti, M and McMahon, SB (2005). Role of the immune system in chronic pain. Nat Rev Neurosci 6: 521–532. | Article | PubMed | ChemPort |
25. Lee, HL, Lee, KM, Son, SJ, Hwang, SH and Cho, HJ (2004). Temporal expression of cytokines and their receptors mRNAs in a neuropathic pain model. Neuroreport 15: 2807–2811. | PubMed | ChemPort |
26. Amos, N, Lauder, S, Evans, A, Feldmann, M and Bondeson, J (2006). Adenoviral gene transfer into osteoarthritis synovial cells using the endogenous inhibitor IB reveals that most, but not all, inflammatory and destructive mediators are NFB dependent. Rheumatology 45: 1201–1209. | Article | PubMed | ChemPort |
27. Wu, CY, Hsieh, HL, Jou, MJ and Yang, CM (2004). Involvement of p42/p44 MAPK, p38 MAPK, JNK and nuclear factor-kappaB in interleukin-1beta-induced matrix metalloproteinase-9 expression in rat brain astrocytes. J Neurochem 90: 1477–1488. | PubMed |
28. Sparacio, SM, Zhang, Y, Vilcek, J and Benveniste, EN (1992). Cytokine regulation of interleukin-6 gene expression in astrocytes involves activation of an NF-kappa B-like nuclear protein. J Neuroimmunol 39: 231–242. | Article | PubMed | ISI | ChemPort |
29. Tsukihara, S, Harada, T, Deura, I, Mitsunari, M, Yoshida, S, Iwabe, T et al. (2004). Interleukin-1beta-induced expression of IL-6 and production of human chorionic gonadotropin in human trophoblast cells via nuclear factor-kappaB activation. Am J Reprod Immunol 52: 218–223. | Article | PubMed |
30. Zhang, YH, Lin, JX and Vilcek, J (1990). Interleukin-6 induction by tumor necrosis factor- and interleukin-1 in human fibroblasts involves activation of nuclear factor binding to a kappa-B like sequence. Mol Cell Biol 10: 3818–3823. | PubMed | ChemPort |
31. Maihöfner, C, Euchenhofer, C, Tegeder, I, Beck, KF, Pfeilschifter, J and Geisslinger, G (2000). Regulation and immunohistochemical localization of nitric oxide synthase and soluble guanylyl cyclase in mouse spinal cord following nociceptive stimulation. Neurosci Lett 290: 71–75. | PubMed |
32. Saha, RN and Pahan, K (2006). Regulation of inducible nitric oxide synthase gene in glial cells. Antioxid Redox Signal 8: 929–947. | Article | PubMed | ChemPort |
33. Jana, M, Anderson, JA, Saha, RN, Liu, X and Pahan, K (2005). Regulation of inducible nitric oxide synthase in proinflammatory cytokine-stimulated human primary astrocytes. Free Radic Biol Med 38: 655–664. | Article | PubMed | ChemPort |
34. Sung, CS, Wen, ZH, Chang, WK, Ho, ST, Tsai, SK, Chang, YC et al. (2004). Intrathecal interleukin-1beta administration induces thermal hyperalgesia by activating inducible nitric oxide synthase expression in the rat spinal cord. Brain Res 1015: 145–153. | Article | PubMed | ChemPort |
35. Obreja, O, Biasio, W, Andratsch, M, Lips, KS, Rathee, PK, Ludwig, A et al. (2005). Fast modulation of heat-activated ionic current by proinflammatory interleukin 6 in rat sensory neurons. Brain 128: 1634–41. | Article | PubMed | ChemPort |
36. Wang, S, Lim, G, Zeng, Q, Sung, B, Yang, L and Mao, J (2005). Central glucocorticoid receptors modulate the expression and function of spinal NMDA receptors after peripheral nerve injury. J Neurosci 25: 488–495. | Article | PubMed | ChemPort |
37. Campbell, IL, Abraham, CR, Masliah, E, Kemper, P, Inglis, JD, Oldstone, MB et al. (1993). Neurologic disease induced in transgenic mice by cerebral overexpression of interleukin 6. Proc Natl Acad Sci USA 90: 10061–10065. | Article | PubMed | ChemPort |
38. Okada, S, Nakamura, M, Mikami, Y, Shimazaki, T, Mihara, M, Ohsugi, Y et al. (2004). Blockade of interleukin-6 receptor suppresses reactive astrogliosis and ameliorates functional recovery in experimental spinal cord injury. J Neurosci Res 76: 265–276. | Article | PubMed | ChemPort |
39. Meller, ST and Gebhart, GF (1993). Nitric oxide (NO) and nociceptive processing in the spinal cord. Pain 52: 127–136. | Article | PubMed | ChemPort |
40. Ikeda, H and Murase, K (2004). Glial nitric oxide-mediated long-term presynaptic facilitation revealed by optical imaging in rat spinal dorsal horn. J Neurosci 24: 888–9896.
41. Osborne, MG and Coderre, TJ (1999). Effects of intrathecal administration of nitric oxide synthase inhibitors on carrageenan-induced thermal hyperalgesia. Br J Pharmacol 126: 1840–1846. | Article | PubMed | ISI | ChemPort |
42. Sommer, C, Petrausch, S, Lindenlaub, T and Toyka, KV (1999). Neutralizing antibodies to interleukin 1-receptor reduce pain associated behavior in mice with experimental neuropathy. Neurosci Lett 270: 25–28. | Article | PubMed | ChemPort |
43. Samad, TA, Moore, KA, Sapirstein, A, Billet, S, Alichorne, A, Poole, S et al. (2001). Interleukin-1beta-mediated induction of Cox-2 in the CNS contributes to inflammatory pain hypersensitivity. Nature 410: 471–475. | Article | PubMed | ISI | ChemPort |
44. Zhuang, ZY, Gerner, P, Woolf, CJ and Ji, RR (2005). ERK is sequentially activated in neurons, microglia, and astrocytes by spinal nerve ligation and contributes to mechanical allodynia in this neuropathic pain model. Pain 114: 149–159. | Article | PubMed |
45. Bemelmans, AP, Bonnel, S, Houhou, L, Dufour, N, Nandrot, E, Helmlinger, D et al. (2005). Retinal cell type expression specificity of HIV-1-derived gene transfer vectors upon subretinal injection in the adult rat: influence of pseudotyping and promoter. J Gene Med 7: 1367–1374. | Article | PubMed | ISI | ChemPort |
46. Zufferey, R, Nagy, D, Mandel, RJ, Naldini, L and Trono, D (1997). Multiply attenuated lentiviral vector achieves efficient gene delivery in vivo. Nat Biotechnol 15: 871–875. | Article | PubMed | ISI | ChemPort |
47. Kwon, H, Pelletier, N, DeLuca, C, Genin, P, Cisternas, S, Lin, R et al. (1998). Inducible expression of IkappaBalpha repressor mutants interferes with NF-kappaB activity and HIV-1 replication in Jurkat T cells. J Biol Chem 273: 7431–7440. | Article | PubMed | ChemPort |
48. Goslin, K, Asmussen, H and Banker, G (1998). Rat hippocampal neurons in low-density culture. In: Banker, G and Goslin, K (eds). Culturing Nerve Cells. MIT Press: Cambridge, MA. pp. 339–370.
49. Waibl, H (1973). Contribution to the topography of the spinal cord of the albino rat (Rattus norvegicus). Adv Anat Embryol Cell Biol 47: 1–42. | PubMed |
50. Bennett, GJ and Xie, YK (1988). A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man. Pain 33: 87–107. | Article | PubMed | ISI | ChemPort |

Acknowledgements

We are grateful to François Cesselin, André Bogdan, Joao Braz and Chamsy Sarkis for critical reading of the manuscript and helpful discussions. We thank Justine Masson for her assistance with primary neuron-glia co-cultures. We thank Anning Lin and Hedi Mammeri for providing us with plasmids. This work was supported by grants from INSERM, Université Pierre et Marie Curie–Paris 6, Fondation pour la Recherche Médicale, Institut UPSA de la Douleur, and Institut pour la Recherche sur la Moelle épinière et l'Encéphale.