Interleukin-6 (IL-6) is a key mediator of inflammation. Inhibitors of IL-6 or of its signal transducing receptor gp130 constitute a novel class of anti-inflammatory drugs, which raise great hopes for improved treatments of painful inflammatory diseases such as rheumatoid arthritis. IL-6 and gp130 may enhance pain not only indirectly through their proinflammatory actions but also through a direct action on nociceptors (i.e., on neurons activated by painful stimuli). We found indeed that the IL-6/gp130 ligand-receptor complex induced heat hypersensitivity both in vitro and in vivo. This process was mediated by activation of PKC-δ via Gab1/2/PI3K and subsequent regulation of TRPV1, a member of the transient receptor potential (TRP) family of ion channels. To assess the relevance of this direct pain promoting effect of IL-6, we generated conditional knock-out mice, which lack gp130 specifically in nociceptors, and tested them in models of inflammatory and tumor-induced pain. These mice showed significantly reduced levels of inflammatory and tumor-induced pain but no changes in immune reactions or tumor growth. Our results uncover the significance of gp130 expressed in peripheral pain sensing neurons in the pathophysiology of major clinical pain disorders and suggest their use as novel pain relieving agents in inflammatory and tumor pain.
Pain management in conditions of chronic inflammation, such as rheumatoid arthritis (RA), remains a major clinical challenge. A large body of preclinical studies has been dedicated toward deciphering the pathophysiological sequelae of events occurring in immune cells during chronic inflammation. However, from the perspective of understanding chronic pain arising in these states, it is also important to understand and address bilateral interactions between immune cells and sensory nerves. Locally released cytokines are a key class of molecules which are particularly relevant in this context. Several lines of evidence show that cytokines play a key role in the pathophysiology of inflammation. A variety of key cytokines, e.g., interleukin-6 (IL-6), IL-11, leukemia inhibitory factor (LIF), oncostatin M (OSM), ciliary neurotrophic factor (CNTF), and cardiotropin-1 (CT-1), use the signal transducer glycoprotein 130 (gp130) for signaling onto mast cells, macrophages, and other types of immune cells (Yoshida et al., 1996; Betz et al., 1998).
Of these, the pleiotropic cytokine IL-6 is particularly worth noting because its levels in healthy individuals are very low, but dramatically elevated in painful inflammatory conditions (Kiefer et al., 2001; Smith et al., 2001). In rats, a direct injection of IL-6 causes hypersensitivity to thermal and mechanical stimuli (Poole et al., 1995; Oka et al., 1995; DeLeo et al., 1996; Brenn et al., 2007). IL-6 complexes with membrane-bound or soluble IL-6 receptor, IL-6R, to activate cells expressing the signal transducer glycoprotein gp130 (Taga et al., 1989; Rose-John and Heinrich, 1994). Interestingly, most cells are devoid of membrane-bound IL-6R and are thus unresponsive to the cytokine IL-6. Such cells, however, can still react to IL-6 complexed with a soluble form of the IL-6R (sIL-6R) to activate gp130, a pathway called “trans-signaling” (Rose-John and Heinrich, 1994). Interestingly, the sIL-6R is generated by apoptotic neutrophils during inflammatory states (Chalaris et al., 2007). In addition to immune and myeloid cells, gp130 is also expressed in rat sensory neurons (Oprée and Kress, 2000; Gardiner et al., 2002; Obreja et al., 2002a, 2005). However, the functional role of gp130 expressed on sensory nerves is not well understood. Furthermore, the signaling events downstream of gp130 in sensory nerves have not been explored, nor is it known how these may potentially impact on key nociceptive molecules, such as ion-channels involved in pain transduction.
We hypothesized that gp130 expressed on sensory nerve may be functionally linked to pathological pain. To specifically delineate the functional role of cytokine signaling via gp130 in sensory neurons, we used a conditional gene targeting approach in mice based on selective deletion of gp130 from peripheral sensory neurons, leaving its expression in other cells intact. Our results suggest that gp130 expressed in sensory nerves not only mediates chronic inflammatory pain, but also contributes significantly to complex interactions between immune cells, tumor cells, and nerves in the context of cancer-evoked pain. Moreover, we identify IL-6 activating gp130, Gab1/Gab2, PI3K, and PKC-δ and regulating TRPV1 as a key mechanism linking cytokine release to sensitization of pain-sensing nerves.
Materials and Methods
Genetically modified mice.
Mice homozygous for the floxed allele of the mouse Il6st gene (gp130fl/fl), which encodes the interleukin-6 signal transducer (Il6st, henceforth gp130) have been described previously (Betz et al., 1998). gp130fl/fl mice were cross-bred with SNS-Cre mice (Agarwal et al., 2004) to obtain homozygous SNS-Cre gp130fl/fl (SNS-gp130−/−) and gp130fl/fl mice (control littermates; the recombinase is homo- or heterozygous). All mice were genotyped using mouse genomic tail DNA with sense primer 5′-TGGCTTGAGCCTCAGCTTGGCTAG-3′ and antisense primer 5′-TGAACAGTCACCATGTACATCTGTACGC-3 to detect the floxed gp130 allele, and sense primer 5′-GAAAGCAGCCATGTCCAATTTACTGACCGTAC-3′ and antisense primer 5′-GCGCGCCTGAAGATATAGAAGA-3′ to detect SNS-Cre transgene expression. SNS-Cre mice, SNS-gp130−/− mice and their corresponding littermates [wild-type (wt) and gp130fl/fl mice, respectively] had the genetic background C57BL/6J. All mice were maintained under SPF conditions. Littermates were used in all experiments to control for background effects, and all animal use procedures were in accordance with ethical guidelines and animal welfare standards according to Austrian law. All behavioral measurements were done in awake, unrestrained, age-matched mice that were >8 weeks old by individuals who were blinded to the genotype of the mice being analyzed.
Genomic DNA from wild-type, gp130fl/fl and SNS-gp130−/− mice dorsal root ganglia (DRGs) was cleaved overnight at 37°C using EcoRI (Roche) in a 10× EcoRI buffer (Roche). DNA was separated on a 0.7% agarose gel containing ethidium bromide, and then blotted onto a nylon membrane (GeneScreen Plus, PerkinElmer Life Sciences) and placed on a UV transilluminator (UV-Stratalinker 2400, Stratagene). The gp130 genomic DNA probe was labeled with α-32P-dCTP (10.0 mCi/ml, PerkinElmer Life Sciences) using the Megaprime DNA Labeling System (GE Healthcare). The membrane was prehybridized for 3 h at 50°C in a hybridization buffer containing 5× Denhardt's solution (bovine serum albumin (Sigma), polyvinylpyrrolidone (Sigma), Ficoll 400 (Sigma), 5× SSC, 1% (w/v) SDS (Sigma), and 100 μg/ml degraded salmon sperm DNA) (Sambrook and Russell, 2001). The probe was denatured, added to the hybridization buffer, and hybridized overnight at 50°C. The membrane was then briefly washed in 2× SSC/0.1% SDS at 50°C followed by 2× SSC/0.1% SDS 2 × 30 min at 50°C and then in 0.1× SSC/0.1% SDS for 10 min at 50°C. Results were examined after exposure to X-ray film (GE Healthcare).
Qualitative and quantitative PCR.
Qualitative PCR was performed, respectively, on RNA, cDNA, isolated from murine dorsal root ganglion (DRG) neurons, treated with TRI Reagent (Sigma), restricted with DNase I (Fermentas), and reverse transcribed with MuLV Reverse Transcriptase (Applied Biosystems). Specific PCR primers were picked according to NCBI referenced sequences. Primer Express Software, Version 3.0 (Applied Biosystems) was used to define primer base constitution, length, and melting temperatures. Oligonucleotides were purchased at MWG (Eurofins MWG Operon). The following sets of primers were used: set 1, sense primer 5′-GATAACCTGCTCTGGGTGGA-3′, antisense primer 5′-TGGAGTTAAAATTGTG CCTTGG-3′; set 2, sense primer 5′-ACCAGATTCCTGTGGACGAC-3′, antisense primer 5′-TGACCACTGGGCAATATGAC-3′. The presence of the desired mutation in the resulting fragments of 400 bp (wild type, gp130fl/fl) and 300 bp (SNS-gp130−/−) was confirmed by sequencing (Microsynth AG). The cycle protocol of the first PCR consists of 1 min at 94°C, 20 three-step cycles of 30 s each at 94°C, 60 s each at 57°C, and 120 s each at 72°C, and a final extension at 72°C of 10 min. The nested PCR was done as aforementioned; only the number of cycles was adjusted to 30. For analysis of gene expression levels, respectively mRNA levels, total RNA was isolated from murine DRG neurons of untreated and treated animals immediately after preparation by using TRI Reagent (Sigma) according to the manufacturer's instructions. RNA was treated with DNase I (Fermentas). Reverse transcription to cDNA was performed using the GeneAmp RNA PCR Kit (Applied Biosystems). Each cDNA sample was analyzed for expression of gp130 by quantitative real-time PCR using the TaqMan 5′ nuclease assays Mm01297292_m1 (gp130, exon boundary 16–17), Mm0129298_m1 (gp130, exon boundary 8–9) and Mm01352363_m1 (Succinate Dehydrogenase Subunit A, SDHA). In case of exon 15 (transmembrane region) a MGB probe (5′-TGTGCTTAGCCTTCC-3′) and the primers 5′-GAAATAGAAGCC-3′ and 5′-CCTGACAACCCTGCTGGGCGT-3′, respectively (all Applied Biosystems), were used. Reactions were performed in a MicroAmp Fast Optical 96-Well Reaction Plate (Applied Biosystems) using the 7500 Fast Real-Time PCR System (Applied Biosystems) for thermal cycling and real-time fluorescence measurements. The PCR cycle protocol consisted of 10 min at 95°C, and 40 two-step cycles of 15 s each at 95°C and of 1 min at 60°. Positive and negative controls were included in all the experiments and each sample was run in triplicate for each PCR. Threshold cycle (CT) values were recorded as a measure of initial template concentration. Relative fold changes in RNA levels were calculated by the ΔΔCT method using SDHA as a reference standard. The range for the target, relative to a calibrator sample was calculated by 2−ΔΔCT.
Culture of primary sensory neurons.
Lumbar DRGs with the cell bodies of primary afferents that project into the hindpaw were harvested from adult mice as previously published (Obreja et al., 2002b; Agarwal et al., 2007). After removal of the connective tissue, ganglia were incubated in Liberase Blendzyme 1 (9 mg/100 ml DMEM, Roche) for 60 min. After washing with PBS (PAA), 1× Trypsin-EDTA (Invitrogen) was added for 30 min, and DRGs were washed TNB medium (Biochrom) containing l-glutamin (Invitrogen), penicillin G sodium, streptomycin sulfate (Invitrogen), and a Protein-Lipid-Komplex (Biochrom). The DRGs were triturated with a fire-polished Pasteur pipette and centrifuged through a 3.5% BSA gradient (Sigma). The sensory neurons were resuspended, plated on coverslips coated with polylysine/laminin (Sigma), and cultivated in supplemented TNB containing mNGF (Alomone Labs, 10 μg/100 ml TNB-medium) at 37°C in 5% CO2 for 24–36 h.
Culture of tumor cells.
Lung carcinoma cells (ETCC clone 1642, European Type Cell Culture Collection) were cultivated in DMEM with l-glutamine and 10% fetal bovine serum (FBS). Before injection, cells were counted, washed twice, and then resuspended in PBS (25 μl). Mice were anesthetized with isofluorane (Baxter) and 7 × 105 lung carcinoma cells were injected subcutaneously in the plantar and dorsal side of the mouse hindpaw.
Immunocytochemistry and live cell labeling.
DRG neurons were dissociated according to the protocol and plated on coverslips. After 24 h in culture, neurons were fixed with 4% PFA for 20 min, permeabilized with 0.01% TX-100 (Sigma) for 2 min and blocked with blocking buffer (BB, 10% goat serum in PBS) for 30 min. Cells were incubated with primary antibodies diluted in BB (1:200) for 1 h at room temperature (RT), washed with PBS, and incubated with secondary antibodies (1:1000) for 30 min at RT. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI, 1:10,000 in PBS) and cells were embedded in Mowiol (Calbiochem). Cell counting was performed at least from triplicates of the experiments. For live cell labeling DRG neurons were incubated on ice with the primary antibody diluted in TNB medium (1:50) for 30 min, washed with TNB medium for 10 min on ice, and incubated with the secondary antibody diluted in TNB medium (1:1000). Subsequently cells were fixed with 4% PFA and proceeded to immunohistochemistry.
Immunohistochemistry on frozen sections.
DRGs were washed with PBS and fixed with 4% PFA for 20 min. After 3 × 10 min washing with PBS, ganglia were frozen in methylbutan. Sections (12 μm) were cut with a Leica CM 1850 Cryomicrotome, permeabilized with 0.1% TX-100, and blocked with blocking buffer. Sections were incubated with primary antibodies overnight at +4°C in a wet chamber. After washing in PBS, sections were incubated with the secondary antibodies, and embedded in Mowiol (Calbiochem). Indirect immune fluorescence was detected with a Zeiss Axiovert 200M microscope and analyzed with the MetaMorph Imaging Software Series 7.1 (Molecular Devices, Visitron Systems). The following antibodies were used: anti-gp130 (1:50, extracellular epitope, Neuromics); anti-Nav1.8 (1:200, was a generous gift from Prof. Dr. H. G. Knaus, Innsbruck, Austria) (see also supplemental Fig. 2, available at www.jneurosci.org as supplemental material) (Liu et al., 2006); anti-CGRP (1:300, ImmunoStar), anti-Isolectin IB4 (1:1000, Invitrogen); Alexa Fluor 488 goat anti-rabbit IgG (1:1000, H+L), Alexa Fluor 594, donkey anti-goat IgG (1:1000, all Invitrogen).
Sensory neurons were plated on poly-l-lysine/laminin-coated dishes, kept in culture for 24 h, stimulated with 5 ng/ml HIL-6 for 10 min in extracellular solution (ECS), and harvested in freshly prepared ice-cold lysis RIPA-buffer (50 mm Tris-HCl, 150 mm NaCl, 50 mm NaF, 5 mm EDTA, 0.5% Deoxycholic Acid, 0.1% SDS, 1% Nonidet P-40, all Sigma). The phosphatase inhibitors sodium-orthovanadate (200 μm) and β-glycerophosphate (40 mm, both Sigma) were added to keep the proteins in phosphorylated state. A protease-inhibitor cocktail (Sigma) was used to protect proteins from proteolysis. SDS-PAGE was performed under standard denaturing conditions using 8.3 × 7.3 cm handcast gels (Mini-PROTEAN, Bio-Rad Laboratories). Equal amounts of protein were loaded to each lane of 10% polyacrylamide gels. Spectra Multicolor Broad Range Protein Ladder (Fermentas) was used as a molecular weight standard. Gels were blotted immediately after electrophoresis onto polyvinylidene fluoride membrane (Hybond-P, GE Healthcare). For immunodetection, membranes were blocked for 1.5 h with 5% (w/v) nonfat dry milk or 5% (w/v) BSA and 0.1% (v/v) Tween 20 in Tris-buffered saline, pH 7.6, at room temperature, and processed according to the instructions of the manufacturers of the antibodies. The following antibodies were used: anti-tubulin (Sigma), anti-phospho-Gab1(Tyr307), anti-phospho-Gab2(Tyr502), anti-phospho-PI3K p85 (Tyr458)/p55 (Tyr199) (all Cell Signaling Technology), anti-Gab1 (H-198), and anti-Gab2 (I-18) (both Santa Cruz Biotechnology). Visualization of blots was performed with enhanced chemiluminescence by using the ECLPlus Western Blotting Detection System (GE Healthcare).
Translocation of PKC-δ.
Cultured neurons were stimulated with 5 ng/ml HIL-6 in ECS for 10 min, fixed in 4% PFA for 20 min, permeabilized in 0.2% TX-100 for 5 min, blocked with 10% NGS, and stained with anti-PKCδ (BD Transduction Laboratories). Images were obtained by a Zeiss Axiovert 200 M laser confocal microscope LCM 510. To quantify the redistribution of PKC-δ after stimulation, line scan profiles of fluorescence intensities were recorded with the MetaMorph Imaging Series 7.1 Software (Visitron Systems). The total length of the profile along the diameter of the cell was set to 100%, and fluorescence intensities of the profile between 0 and 10% and 90 and 100% were determined (periphery) together with intensities between 45 and 55% of the profile (center of cell). The translocation coefficient (K) was calculated as ratio of the values determined for the average of the periphery divided by the value within the center of the cell as previously published (Rathee et al., 2002). K = 1 if the enzyme is equally distributed throughout the cell, K > 1 refers to predominant localization close to the plasma membrane, and K < 1 represents a predominantly cytoplasmic localization.
DRG neurons in culture were used for electrophysiology after 24 h. Whole-cell ionic currents were recorded from isolated neurons in the whole-cell configuration of the patch-clamp technique as previously published (Obreja et al., 2002b, 2005). ECS contained (in mm): 150 NaCl, 5 KCl, 0.1 CaCl2, 1 MgCl2 (all Sigma), 10 glucose, and 10 HEPES (Merck), at pH 7.3 adjusted with NaOH (Merck). Calcium was reduced to 0.1 mm to avoid desensitization of capsaicin-induced currents. Borosilicate glass pipettes (Science Products) pulled with a horizontal puller (Sutter Instruments Company) were filled with internal solution (ICS) (in mm): 148 KCl, 2 MgCl2, 2 Na-ATP, 0.1 CaCl2, 1 EGTA (all Sigma), and 10 HEPES (Merck), at pH 7.3 adjusted with KOH (Merck). After filling, electrode resistance was 4–5 MΩ. Neurons were clamped at −80 mV holding potential. Currents were filtered at 2.9 kHz, sampled at 3 kHz, and recorded with an EPC 9 (HEKA) and the Pulse v8.74 software (HEKA). Experiments were performed at room temperature and only one neuron was tested per Petri dish. A seven barrel system with common outlet was used for fast drug administration and heat stimulation of single neurons (Dittert et al., 1998). Capsaicin-activated inward currents (Icaps) were elicited by applying 0.3 μm capsaicin (Sigma) for 10 s followed by a 2 min washout with control solution. This protocol was repeated 3 times before HIL-6 (1 ng/ml) was applied for 1 min immediately before capsaicin for conditioning stimulation. Recombinant interleukin-6 from mouse (IL-6, Sigma) was solved in PBS containing 10% BSA. IL-6 (5 ng/ml) and BSA (vehicle; 0.1 mg/ml), respectively, were used in the extracellular solution.
Skin-nerve preparation and single fiber recordings.
An in vitro skin nerve preparation was used to investigate the properties of the afferent nerve fibers innervating the skin of the mouse dorsal hindpaw as previously published (Kress et al., 1992; Koltzenburg et al., 1997). Briefly, the preparation was superfused (15 ml/min) with an oxygen-saturated modified synthetic interstitial fluid solution containing (in mm) 108 NaCl, 3.48 KCl, 3.5 MgSO4, 26 NaHCO3, 1.7 NaH2PO4, 2.0 CaCl2, 9.6 sodium gluconate, 5.5 glucose, 7.6 sucrose at temperature of 31 ± 1°C, and pH 7.4 ± 0.05. Action potentials of single sensory neurons were recorded extracellulary from fine filaments dissected from the saphenous nerve, amplified (5000-fold), filtered (low pass 1 KHz, high pass 100 Hz), visualized on oscilloscope, and stored on a PC-type computer with Spike/Spidi software package (Forster and Handwerker, 1990). The fibers were characterized as unmyelinated (C) according to their conduction velocity (c.v. < 1.4 m/s). The receptive field was identified by mechanical probing of the skin with a glass rod; standard heat stimuli linearly rising the intracutaneous temperature from 31 ± 1°C to 47°C were applied. A fiber was considered heat sensitive if 3 or more action potentials were evoked during the stimulus. The heat threshold was defined as the temperature that elicited the third spike of the response. In tumor-injected mice electrophysiological recordings were performed 7–10 d after inoculation when a tumor mass had formed at the injection site.
SNS-gp130−/− and the corresponding gp130fl/fl control mice were transcardially perfused with cacodylate buffer containing 2% freshly depolymerized paraformaldehyde and 2% glutardialdehyde, followed by femoral nerve dissection, osmification, and embedding in Spurr′s medium as described previously (Carenini et al., 2001). Ultrathin sections (80 nm) were contrasted with lead citrate and analyzed using a ProScan Slow Scan CCD (Proscan) camera mounted to a Leo 906 E electron microscope (Zeiss) and corresponding software iTEM (Olympus).
Spinal cord slices of 14 – 20-d-old mice were obtained by a standard procedure. In short, the mice were anesthetized with isoflurane and decapitated, the spinal cord was microdissected in ice-cold ACSF (containing 125 mm NaCl, 26 mm NaHCO3, 1.25 mm NaH2PO4, 2.5 mm KCl, 2 mm CaCl2, 1 mm MgCl2, 10 mm glucose), glued to a gelatin block, and frontal slices were obtained by a vibroslicer. Lumbar slices with intact dorsal rootlets were fixed in a recording chamber with a platinum grid and a bipolar stimulation electrode (250 μm pole distance) was placed on both sides of the rootlet. Synaptic currents of dorsal horn neurons (layer 2/3) in response to stimulation of the primary afferents with short pulses (0.2 ms duration, pulse distance 12 s) were recorded with an EPC10 and patch master software (HEKA) by use of patch pipettes filled with a K-gluconate-based internal solution (Cl− reversal potential −90 mV). The stimulation threshold was identified as the lowest voltage strength with a response probability >50%.
Standard testing procedures were used to quantify signs of pain-like behavior. The area tested was the plantar side of the hindpaw where the tumor cells were inoculated. Heat sensitivity was assessed using the Hargreaves test (Hargreaves et al., 1988). Paw withdrawal latency in response to an increasing heat stimulus was measured automatically (Ugo Basile) before and after cytokine injection or tumor cell injection at time points as indicated in figures. Instead of coinjecting IL-6 together with sIL-6R, the designer peptide hyper-IL-6 (HIL-6) was used to mimic the effects of the IL-6/sIL-6R complex (Fischer et al., 1997). Complete Freund's Adjuvant (CFA, 20 μl, Sigma) was injected into the plantar site of the hindpaw in gp130fl/fl and SNS-gp130−/− mice, and changes in the mechanical and heat sensitivity were measured 6, 24, and 48 h after inoculation. At the end of the observation period, tumors were excised in toto, cleaned from connective tissue, and weighed immediately. For CCI of the right sciatic nerve, mice were anesthetized by intraperitoneal phenobarbital injection, and three ligatures (7-0 prolene) with a distance of 1 mm each were loosely tied around the nerve proximal to the trifurcation until a short flick of the hindpaw was observed.
For detailed statistical analysis the SigmaStat 3 program was used. Data are presented as mean ± SEM and were analyzed using the Wilcoxon signed rank test, the paired t test or one-way repeated-measures ANOVA for comparison between groups and test days, followed by a Student's t test if not otherwise stated. Differences were considered statistically significant at p < 0.05.
Animals deficient for the ubiquitously expressed signal transducer IL-6 receptor β subunit gp130 die after 12.5 d postconception or postnatally, depending on the genetic background (Yoshida et al., 1996; Kawasaki et al., 1997). Therefore, conditional knock-out mice lacking gp130 selectively in nociceptive neurons of the DRG were generated using the Cre recombinase loxP strategy and conditional removal of the transmembrane exon 15 (referred to henceforth as SNS-gp130−/− mice). This was achieved by mating homozygous mice carrying the loxP-flanked (floxed) gp130 allele (130fl/fl) (Betz et al., 1998) with a mouse line expressing Cre recombinase under the transcriptional control of the nociceptor-specific Nav1.8 gene (SNS-Cre). In SNS-Cre mice, gene recombination selectively occurs in >90% of small size (≤28 μm) nociceptive sensory neurons, commenced at birth and does not affect gene expression in other regions (Agarwal et al., 2004). Deletion of gp130 in DRG neurons was detected by Southern blot analysis on genomic DNA of mouse DRGs using a gp130-specific probe and via quantitative PCR analysis (supplemental Fig. 1, available at www.jneurosci.org as supplemental material). To confirm that gp130 protein was absent from the cell membrane, live immune staining was performed with antibodies directed against the extracellular domain of gp130 and the Nav1.8 ion channel in neuron cultures obtained from gp130fl/fl and SNS-gp130−/− mice, which revealed that >90% of the Nav1.8-positive neurons in SNS-gp130−/− mice did not show anti-gp130 immunoreactivity on their cell surface (Fig. 1A,B). Thus these three independent methods suggest an efficient recombination and deletion of functional gp130 in nociceptive neurons. In contrast, expression of gp130 remained intact in other tissues, including immune cells, in SNS-gp130−/− mice (data not shown).
No obvious developmental abnormalities occurred on loss of gp130 from sensory neurons. Since cytokine signaling via gp130 such as IL-6 or LIF are important survival factors and also regulate neuronal excitability (März et al., 1999; Schäfer et al., 1999; Brenn et al., 2007; Kawasaki et al., 2008), we assessed whether the two main classes of nociceptive neurons in the DRG, namely the peptidergic, CGRP-expressing (CGRP+) and the nonpeptidergic, Isolectin B4 binding (IB4+) nociceptors (Caterina and Julius, 1999) were different from wild-type littermates. The morphology and the overall frequency of IB4 and CGRP+ neuron populations were similar between wild-type and SNS-gp130−/− DRGs (Fig. 2A,B,) suggesting that the lack of the gp130 signal transducer did not affect the morphology or general structural properties of sensory neurons. Electron microscopy revealed no pathological alterations in nonmyelinated nerve fibers of SNS-gp130−/− mice (Fig. 2C,D). General morphology of the spinal dorsal horn and electrophysiological parameters of synaptic spinal transmission in the substantia gelatinosa were similar in both mouse strains (supplemental Fig. 3A,B, available at www.jneurosci.org as supplemental material). Overt motor or other behavioral deficits were not observed in the SNS-gp130−/− mice compared with littermates.
Functional consequences of gp130 loss from sensory neurons
No overt difference was observed regarding basal heat pain sensitivity in mice lacking neuronal gp130. Since IL-6 is found at high levels in inflamed tissue, we investigated SNS-gp130−/− mice in a model of inflammation-induced hyperalgesia. Subcutaneous inoculation of 20 μl of CFA (1 mg/ml) into the plantar site of the hindpaw led to development of inflammation and hypersensitivity to noxious heat stimulation, both in gp130fl/fl and SNS-gp130−/− mice within 6 h after injection (Fig. 3A). However, although the degree of swelling, which is an indication of inflammation, was very similar in both mouse strains (Fig. 3B), the heat hyperalgesia was significantly attenuated in SNS-gp130−/− compared with gp130fl/fl mice (p < 0.001, ANOVA). In gp130fl/fl mice, inflammatory hyperalgesia fully developed within 6 h and remained at constant levels for up to 48 h. In contrast, after an initial peak at 6 h, SNS-gp130−/− showed significant recovery from the hyperalgesic state, although inflammation persisted (Fig. 3A).
Significance of neuronal gp130 for cancer pain
In a novel model for cancer pain in mice, we have recently shown that on a subcutaneous injection of LL2 carcinoma cells into hindpaw, the tumor stroma shows signs of an inflammatory reaction and is densely invaded by macrophages, which are known to produce large amounts of cytokines when activated (Constantin et al., 2008). In this model, tumor growth is accompanied by sprouting of peptidergic nerve fibers into the tumor mass, progressive mechanical and heat hyperalgesia, and a significant sensitization of nociceptors in wt C57BL6J mice (Constantin et al., 2008). As the tumors grew, a progressive heat hyperalgesia was observed in gp130fl/fl mice (p < 0.001, ANOVA, n = 23) (Fig. 3C). In SNS-gp130−/− mice, heat hyperalgesia was significantly attenuated, and mild signs of heat hyperalgesia developed only in the very late stages after tumor induction (Fig. 3D). This beneficial effect was not due to a reduction of tumor growth or inflammation because tumor growth and paw swelling were similar in gp130fl/fl and SNS-gp130−/− mice (Fig. 3G).
Neuronal gp130 required for tumor-associated nociceptor sensitization
In wt C57BL6J mice, tumor-induced hyperalgesia is accompanied by a sensitization of heat-sensitive nociceptive fibers innervating the skin in the tumor area. We therefore hypothesized that activation of neuronal gp130 might be critical for the nociceptor sensitization process and performed single fiber recordings from unmyelinated nociceptive primary afferents in a skin-nerve in vitro preparation (Reeh, 1986; Koltzenburg et al., 1997). In vitro, heat-sensitive unmyelinated fibers from gp130fl/fl tumor mice exhibited higher mean discharge rates during heat stimuli compared with fibers from control mice without tumor [4.53 ± 1.00 impulses (imp)/s versus 2.27 ± 0.53 imp/s; p < 0.05; ANOVA] (Fig. 3E). In SNS-gp130−/− mice with tumors, the mean heat response magnitudes were similar to those in nociceptors recorded from naive healthy mice (p > 0.05; ANOVA (Fig. 3F). This suggests that gp130 expressed on nociceptive primary afferents plays an important role in the generation of tumor-induced heat hyperalgesia and directly regulates the sensitivity of pain-sensing neurons. No major contribution of gp130 was observed in a model of neuropathic pain, in which the development of heat hyperalgesia was delayed, but not prevented. This was likely due to lack of a major inflammatory component in this model (supplemental Fig. 4, available at www.jneurosci.org as supplemental material).
Relevance of IL-6 for tumor-associated pain and nociceptor sensitization
The above analyses suggest that gp130 expressed on sensory nerves plays a direct role in pathophysiological processes triggered by cytokines in sensory nerves. To study in depth which molecular events would functionally link cytokine-induced gp130 to nociceptive hypersensitivity, we turned to IL-6, a key cytokine implicated in pathological inflammatory disorders and pain. In the tumor material we also found expression of IL-6 mRNA, which was not detected in LL2 cell cultures before injection into the paw (Fig. 4A), suggesting an upregulation of IL-6 expression in tumor cells or associated immune cells in vivo. Correspondingly, we found significant upregulation of IL-6 protein levels in the tumor tissue compared with healthy control tissue (e.g., skeletal muscle) (Fig. 4B).
In most tissues, IL-6 requires a ligand-binding soluble receptor subunit (sIL-6R), which complexes with the membrane-bound gp130 signaling subunit. A recent study by Chalaris et al. showed that neutrophils, which normally precede the incoming macrophages, are a major source of sIL-6R during inflammatory states (Chalaris et al., 2007). This suggests that both IL-6 and sIL-6R are present in the tumor tissue. A designer fusion peptide, HIL-6, has been shown to bind and activate gp130 and to mimic IL-6/sIL-6R effects in many tissues (Fischer et al., 1997). We used HIL-6 as a tool to investigate whether activation of gp130 affects pain behavior in mice. Injection of HIL-6 (1 μg in 10 μl) into the foot pad of a hindpaw in gp130fl/fl mice led to a significant drop in paw withdrawal latencies (PWLs) to heat stimulation from 8.89 ± 0.43 s before to 4.20 ± 0.43 s 30 min after inoculation (p < 0.001; ANOVA). PWL recovered to preinjection values within 24 h (Fig. 4C). Compared with gp130fl/fl littermates, HIL-6-induced drop in paw withdrawal latencies was significantly attenuated in the nociceptor-specific SNS-gp130−/− mice (Fig. 4D), suggesting that a major part of HIL-6-induced heat hyperalgesia requires gp130 expressed in the membrane of nociceptor terminals or axons. Only a minor part of the sensitization was found to be independent of neuronal gp130, and could result potentially from a release of secondary inflammatory mediators from other gp130-expressing cell types, e.g., those cytokines which signal via gp130-independent receptors, such as the oncostatin M receptor (OSM-R) and the interleukin-27 receptor WSX-1 (Scheller et al., 2005). Based on the magnitude of changes observed in the above experiments, we propose that IL-6 is one of the most important cytokines in this respect. Consistent with the above, pain sensitivity has been reported to be attenuated in mice lacking IL6 or in wild-type mice after neutralization of endogenous IL-6 (Ferreira et al., 1993; Xu et al., 1997; Murphy et al., 1999). Furthermore, we observed that other cytokine activators of gp130, such as leukemia inhibitory factor (LIF), neither induced heat hypersensitivity in vivo nor sensitized nociceptive neurons in vitro (supplemental Fig. 5, available at www.jneurosci.org as supplemental material).
Further evidence for a critical role for IL-6 in the phenotype observed in SNS-gp130−/− mice was found in the following experiments. In the skin-nerve preparation, application of HIL-6 (1 ng/ml) to the receptive field of unmyelinated heat-sensitive fibers induced a significant augmentation of responses to a standard ramp-shaped heat stimulus in 64% (7/11) of heat-sensitive C-fibers in gp130fl/fl mice (Fig. 4E; mean discharge 2.01 ± 0.42 imp/s before vs 4.17 ± 0.72 imp/s after HIL-6; p < 0.01; paired t test). This sensitization occurred within 5 min after HIL-6 application and was accompanied by a significant shift of the activation threshold to a lower temperature (38.0 ± 1.8°C before versus 34.3 ± 0.9°C 5 min after HIL-6; p < 0.05 paired t test). In contrast, none of the fibers recorded from SNS-gp130−/− mice responded to HIL-6 with a potentiated heat response (Fig. 4F; n = 7; n.s.; Wilcoxon rank test). Together, the data suggest, that gp130 expressed in nociceptors is required for IL-6 to induce heat hypersensitivity in vivo and heat sensitization of nociceptors in vitro.
Finally, we performed electrophysiological studies in cultured neurons isolated from wild-type DRGs, which directly showed that HIL-6 sensitizes capsaicin-sensitive neurons to heat stimuli. In particular, we found a significant increase of heat-activated inward currents after conditioning stimulation with HIL-6 in isolated small diameter DRG neurons (Fig. 4G,H), which was accompanied by a shift of activation thresholds toward lower temperatures.
IL-6-induced signaling in sensory neurons
We then sought to clarify the major signaling components underlying the striking functional role we observed for gp130 in pain modulation. The capsaicin- and heat-sensitive ion channel TRPV1, a member of the TRP (transient receptor potential) family of ion channels serves as molecular transducer at the nociceptive nerve terminals (Caterina et al., 1997; Scott and Zuker, 1998; Julius and Basbaum, 2001; Minke and Cook, 2002; Montell, 2005), and there is ample evidence that TRPV1 regulation by inflammatory mediators is essential for the generation of heat hyperalgesia (Davis et al., 2000; Chuang et al., 2001). We hypothesized that TRPV1 could be regulated by gp130 signaling. TRPV1 is activated not only by heat stimuli but also by capsaicin, the pungent ingredient of hot chili peppers, and in small-sized neurons, HIL-6 induced a fast and transient potentiation of capsaicin-activated ionic currents (Icaps) from 1.04 ± 0.34 nA before to 1.82 ± 0.07 nA after HIL-6 (Fig. 5A; supplemental Fig. 6, available at www.jneurosci.org as supplemental material). A similar increase of Icaps was also obtained after conditioning stimulation of gp130fl/fl neurons with recombinant IL-6 (Fig. 5B), suggesting that IL-6 binding receptor α subunits are also expressed in the nociceptive cell membrane. This is in contrast to the rat in which the sensitizing effect of IL-6 requires soluble IL-6 receptor (Obreja et al., 2005). Expectedly, the conditioning effect of IL-6 was fully abrogated in neurons lacking the gp130 protein from SNS-gp130−/− mice (Fig. 5B). Using capsaicin-induced excitatory inward currents as a read-out, we addressed the underlying signaling pathway and proposed a role of Gab1/Gab2, PI3K and PKC-δ (Heinrich et al., 2003). We performed Western blot experiments with phosphospecific antibodies, which revealed that adapter proteins Gab1, Gab2 (Nishida et al., 1999), and PI3K are phosphorylated in DRG neurons after stimulation with HIL-6 (Fig. 5C). In the presence of wortmannin (1 μm), a PI3K inhibitor, HIL-6 no longer induced a sensitization of Icaps (Fig. 5D) suggesting that PI3K was involved in the signaling cascade downstream of gp130 in nociceptive neurons. It is known that protein kinase C (PKC) can be recruited on PI3K activation in SK-N-BE cells (Plo et al., 2004) and translocation of PKC from a cytosolic to a membrane-associated location within the cell is a sensitive indicator of PKC activation. In rat DRG sensory neurons, both PKC-δ and PKC-ε are clearly translocated to the surface membrane on treatment with the PKC activator phorbol-12-myristate-13-acetate (PMA) (Cesare and McNaughton, 1996). Since PKC-δ has been associated with PI3K activation, we investigated whether this isoform was translocated following exposure to HIL-6. In untreated neurons of both gp130fl/fl and SNS-gp130−/− mice, PKC-δ was distributed uniformly throughout the cytoplasm. Stimulation with HIL-6 caused an activation of PKC-δ, which was visible as redistribution from the cytosol to the cell membrane. In contrast, no migration of PKC-δ was observed in sensory neurons of mutant mice (Fig. 5E,F). In line with the translocation of PKC-δ, both the nonselective PKC inhibitor BIM-1 (data not shown) and the PKC-δ inhibitor Rottlerin (Gschwendt et al., 1994) significantly reduced or abolished the HIL-6-induced facilitation of Icaps (Fig. 5G). Together, these data suggest a gp130-mediated activation of PKC-δ via Gab1/2/PI3K, and consecutive regulation of heat and capsaicin activated ion currents.
The phenotypic alterations seen in SNS-gp130−/− (lack of heat hypersensitivity in pathological pain models) and the involvement of PKC prompted us to address the potential modulation of TRPV1 by gp130 signaling in sensory neurons. After HIL-6 injection, wild-type mice reacted with a significant drop in PWL from 15.21 ± 0.91 s before the injection to 5.10 ± 0.77 s after 1 h (p < 0.001; Student's t-Test, n = 6), which recovered to preinjection values within 24 h. In TRPV1−/− mice, the drop in PWL was significantly attenuated (15.44 ± 1.51 s before the injection vs 10.43 ± 0.62 s after 1 h) (Fig. 5H). One hour after HIL-6 injection wild-type littermates showed a significant drop in PWL in comparison with TRPV1−/− mice (5.10 ± 0.77 s in wt vs 10.43 ± 0.62 s in TRPV1−/−; p < 0.001; Mann–Whitney Rank Sum Test, n = 6). These experiments suggest that TRPV1 plays a major role in HIL-6–induced heat hypersensitivity in vivo and revealed that sensitization of nociceptive neurons to heat was due to a regulation of TRPV1. This did not result from differences in TRPV1 expression patterns since mRNA analysis using quantitative PCR revealed that TRPV1 mRNA expression levels were similar in the mutant mouse and the wild type (data not shown).
Together, our data suggest that gp130 expressed in the membrane of nociceptive primary afferents is critically required for the development of pathological pain and hyperalgesia in inflammatory conditions as well as in cancer. We observed that deletion of gp130 specifically in sensory nerves alleviates heat hyperalgesia in vivo in models of pathological pain with inflammatory background independently of an anti-inflammatory action or modulation of tumor growth. Our results show that the gp130-Gab1/Gab2-PI3K-PKC-δ-TRPV1 pathway in peripheral sensory neurons is a novel molecular pathway linking inflammatory cytokines to pain hypersensitivity.
IL-6 and IL-6-related cytokines share a common signaling receptor subunit gp130 which is ubiquitously expressed in the cellular membrane of all cells types. In contrast to IL-6 the other members of the cytokine group, such as LIF, OSM, CNTF, and CT-1, can heterodimerize also with other membrane-bound receptor subunits. The common signaling molecule gp130 is ubiquitously expressed, whereas IL-6R is not (Peters et al., 1998). Therefore, a majority of cell types, including neurons, require the additional presence of soluble IL-6 receptor (Rose-John and Heinrich, 1994; März et al., 1999; Oprée and Kress, 2000).
Application of HIL-6 led to a significant increase in heat responses and a decrease of the activation temperature, which depended on neuronal gp130. The sensitization process occurred at short latency, which suggests the involvement of a fast signaling cascade, rather than the classical cytokine-induced activation of the Janus kinase/signal transducer and activator of transcription pathway, the RAS/mitogen-activated protein kinase pathway and changes in gene expression. The present data favor a gp130 receptor-mediated activation of protein kinases and the consecutive phosphorylation of a heat transducing ion channel, i.e., the heat-transducing vanilloid receptor TRPV1 (Caterina et al., 1997). TRPV1 in its function as a molecular temperature sensor at the cell surface (Julius and Basbaum, 2001) seems to be a likely candidate regulated by IL-6 signaling via PKC-dependent phosphorylation. Inflammatory mediators activating PKC sensitize TRPV1 (Cesare and McNaughton, 1996; Cesare et al., 1999; Premkumar and Ahern, 2000; Sugiura et al., 2002) and PKC inihbitors attenuate the sensitization of heat-activated inward currents by IL-6/gp130 (Obreja et al., 2005). Recent evidence supports the specific involvement of the calcium-independent isoform PKC-δ in IL-6/gp130 signal transduction (Jain et al., 1999; Novotny-Diermayr et al., 2002). Indeed, we observed translocation of PKC-δ after stimulating sensory neurons with IL-6 and IL-6-induced increases in Icaps were significantly reduced in the presence of the rottlerin, which inhibits PKC-δ in neurons (Zhang et al., 2007) but which may have additional effects (Soltoff, 2007). The structure of TRPV1 exhibits several intracellular phosphorylation sites comprising targets for protein kinases like PKA, PKC or PTK (Vellani et al., 2001; Bhave et al., 2002, 2003; Rathee et al., 2002; Jin et al., 2004; Jung et al., 2004; Zhang et al., 2005; Mandadi et al., 2006), and phosphorylation at these sites leaves the channel in a state of increased activity (Numazaki et al., 2002; Mohapatra and Nau, 2003; Mandadi et al., 2006). Additionally, PKC has been shown to regulate exocytosis of TRPV1 channels preformed in cytosolic vesicles into the plasma membrane (Morenilla-Palao et al., 2004) which could cooperate with the sensitization of phosphorylated channel proteins to explain enhanced neuron sensitivity. Our results extend these findings, and we now provide a novel signaling pathway of IL-6/gp130-induced regulation of TRPV1 via activation Gab1/Gab2, PI3K and PKC-δ. PI3K has been reported to mediate the early induction of hyperalgesia and also can sensitize TRPV1, possibly through posttranslational modification (Zhuang et al., 2004). It may also indirectly regulate the activity of the channel by upregulating its translation or trafficking (Stein et al., 2006). Previously it has been shown that PI3K rapidly activates intracellular protein kinases, like PKB/Akt, and some isoforms of PKC like PKC-δ (Balendran et al., 2000; Plo et al., 2004; Zhuang et al., 2004; Zhu and Oxford, 2007). We demonstrate here that this pathway is activated in sensory neurons. Gp130-mediated activation of PKC-δ was indicated by a significant translocation to the cell membrane and thus increased proximity to ion channels with sensory function. Precisely this relocation of PKC isoforms has been shown to regulate TRPV1 via phosphorylation and translocation into the plasma membrane, both of which occur at a fast time scale (Morenilla-Palao et al., 2004; Premkumar et al., 2004; Zhang et al., 2005; Mandadi et al., 2006). Apart from the regulation of TRPV1, IL-6-dependent activation of PKC-δ can interact with STAT3, and this enhances the interaction of STAT3 with the gp130 receptor which is the initial step for STAT3 activation by IL-6 (Novotny-Diermayr et al., 2002). In hematopoietic stem cells, IL-6 regulates expression of membrane-bound and -soluble IL-6R α subunits (Campard et al., 2006). Such effects might also contribute to the IL-6-induced nociceptor sensitization and likely further promote TRPV1 regulation downstream of gp130 activation. IL-6R can exist as a soluble protein (sIL-6R), which binds the ligand IL-6. This soluble complex can bind to gp130 on cells that lack the membrane-bound IL-6R and initiate signaling. The significance of this so called trans-signaling via sIL-6R is generally accepted and very likely exceeds the significance of the membrane-bound IL-6R (for review, see Jones and Rose-John, 2002; Rose-John et al., 2006). Surprisingly, the sIL-R6 was not essentially required in our cellular model, and this is in contrast to previous work in the rat (Obreja et al., 2002a, 2005). It is the general assumption that in many human tissues sIL-6R acts as an agonist in combination with IL-6 resulting in an enhancement of the IL-6 effects, however, sIL-6R may have IL-6-antagonistic effects (for review, see Knüpfer and Preiss, 2008).
In inflammation and cancer, IL-6 biology is coordinated by membrane-bound and -soluble receptors (Rose-John et al., 2006). Not only in a mouse model of peripheral inflammation alone, but also in a mouse cancer pain model, we observed a major contribution of gp130 expressed on sensory nerves in hyperalgesia, independently of its role in inflammation and tumor growth per se. Both models are characterized by inflammatory changes and activation of immune cells secreting proinflammatory cytokines, such as IL-6 (Constantin et al., 2008). In clinical trials, treatment of patients suffering from rheumatoid arthritis and inflammatory pain with a neutralizing anti-IL-6R antibody was reported to ameliorate symptoms of inflammation as well as the associated pain (Smolen et al., 2008). Our results uncover the significance of gp130 expressed in peripheral pain sensing neurons in the pathophysiology of major clinical pain disorders and suggest its promise as a novel therapeutic target.
This work was supported by the Austrian Fond zur Förderung der wissenschaftlichen Forschung (P18444), Tiroler Wissenschaftsfond (UNI-0404/408), and the Wilhelm Sander-Stiftung (AZ1996.058.3) to M.K., and by the Deutsche Forschungsgemeinschaft Germany in SFB415 (project B5) and SFB654 (project C5) and within the Cluster of Excellence “Inflammation at Interfaces” to S.R.J. We thank M. Doblander and I. Lanz for expert technical assistance and G. Gstraunthaler and E. Feifel for general support.
- Correspondence should be addressed to Dr. Michaela Kress, Division of Physiology, Department of Physiology and Medical Physics, Innsbruck Medical University, Fritz-Pregl-Strasse 3, A-6020 Innsbruck, Austria.