Impact of central and peripheral TRPV1 and ROS levels on proinflammatory mediators and nociceptive behavior

Baseline measurements of knee joint temperature, circumference, and nociceptive behaviors were conducted in male Sprague Dawley rats (n = 8). Each animal served as its own control in a repeated measures analysis of variance experimental design (Mann Whitney U post-hoc testing) with the significance level set at p < 0.05. The rats received intrathecal administration of the ROS sensitive dye, dihydroethidium (DHE, 50 μl of 100 μM). The following day, half of the animals received an intraperitoneal (i.p.) injection of the ROS scavenger PBN (100 mg/kg) and half received saline i.p. The left knee joints (condyle fossa) of the hindlimbs of the anesthetized rats were injected 1 h later with 3% carrageenan (plant product and irritant) and 3% kaolin (clay silt) (100 μl in sterile saline) to produce an acute inflammation confined to the one knee joint.

Four hours after the k/c injection, the inflamed knee joint produced an ipsilateral abnormal stance with curled toes and foot eversion at rest. There was a slightly noticeable limp when animals moved slowly. There was no discernable difference from controls in rapid locomotion. Knee joint circumference increased 1 cm on average (4-8 h) on the injected left knee but not on the uninjected knee as previously reported [20]. Skin temperature was elevated over baseline in both knees indicating mirror blood flow increase despite negligible change in joint circumference in the uninjected knee joint (Fig. 1, hatched bars, p < 0.05; 4 h).

In animals pre-treated i.p. with ROS scavenger, PBN, circumference of the injected knee in animals was also increased by a similar amount (0.78 cm ± 0.1 s.e.m., 4-8 h). However, skin temperature of both knees remained at baseline 4 h after induction of inflammation (Fig. 1, black bars). Core body temperature was not increased over baseline in any of the animals (data not shown).

After induction of knee joint inflammation, ROS was observed in the spinal cord of animals pre-injected with the ROS indicator dye DHE (Fig. 3A). No fluorescent DHE was observed in the spinal cord of control animals. The ROS indicator was localized particularly in association with large cells and in the myelin sheaths of incoming afferent nerve fiber bundles (arrows). Localization in nerve bundles implies that ROS derived from oligodendroglia are important in glial/neuronal interactions during emerging inflammation induced pronociceptive states. Systemic pre-treatment with the PBN greatly reduced visualization of the ROS indicator in the spinal cord, concomitant with the decreased scores for nociceptive behaviour and joint temperature in animals with inflamed knee joints (Fig. 3B).

TRPV1 channels have previously been characterized on clonal human type B SW982 synoviocytes using RT-PCR, Western blot and calcium mobilization after activation with TRPV1 agonists, capsaicin and resiniferatoxin [18, 19]. In the present study, the clonal type B synoviocytes were utilized as an in vitro model of knee joint inflammation since these synoviocytes produce inflammatory mediators when activated with another channel activator, glutamate [21]. Unstable hydroxyl radicals (HO•) were induced by TRPV1 agonist resiniferatoxin in cultured clonal SW982 synoviocytes. While the unstable ROS were not measureable, metabolites 2,3- and 2,5-dihydroxybenzoic acid (2,3- and 2,5-DHBA) were measureable in the cell culture media by HPLC with electrochemical detection (Fig. 3C). Cell plates were incubated (37°C × 6 h, n = 4) in saline bath containing 2.5 mM salicylic acid with 100 nM resiniferatoxin or vehicle (n = 4). Samples of extracellular solution were taken and analyzed for ROS metabolites. The TRPV1 activation of the SW982 cells produced a statistically significant 1.68-fold increase in extracellular ROS metabolite 2, 5-DHBA (7.8 ± 0.9 pg/ml, p < 0.05) compared to control cells treated with the vehicle (4.6 ± 0.4 pg/ml).

For comparison, ROS production over time after incubation with H₂O₂ was tested as shown in Fig. 5. An oxidative burst activated by H₂O₂ (1 mM) was prevented when cells were pretreated with the ROS scavenger, PBN (2 mM, 2 h). The average DCF fluorescence (± s.e.m.) was plotted from control (Fig. 5A, lower trace) cells pretreated with PBN as well as from H₂O₂ activated cells, with and without PBN (Fig. 5A middle and upper trace, respectively). Treatment with the ROS scavenger alone had no effect on the synovial cell cultures beyond the gradual spontaneous oxidation of the dye seen over time in vehicle treated cells (Fig. 5A, lower trace).

Activation of synoviocytes with the ROS donor, tBOOH, increased the immunocytochemical staining intensity for TRPV1 almost two-fold (1.9×) within 30 min (Fig. 5B) (control 115.6 ± 10.5 versus activated 219.8 ± 41.49 arbitrary units, p < 0.05). This increase was blocked by the TRPV1 antagonist, SB-366791 (30 μM)(134.5 ± 20.55). The increase in TRPV1 protein was measureable with Western blot at 8 h normalized with β-actin controls (Fig. 5C).

Stimulation of human clonal SW982 synoviocytes with resiniferatoxin, greatly increased immunocytochemical staining for the inflammatory enzyme COX-2 (Fig. 6A) and the chemokine RANTES (not shown) visualized at 24 h. The TRPV1 induced increase in COX-2 staining was reduced by pre-treatment with either TRPV1 antagonist capsazepine (5 μM) or the ROS scavenger PBN (2 mM). Neither inhibitor had an effect on the clonal synoviocytes when applied in otherwise unstimulated cells. Significantly increased staining for TNFα was also evident within 30 min in synoviocytes activated with ROS donor tBOOH (100 μM) (Fig. 6B). The increase in TNFα levels after activation with the ROS donor was less in the presence of the TRPV1 antagonist SB-36791 (30 μM).

We have previously identified, characterized and sequenced TRPV1 and TRPV4 isoforms on clonal and primary human synoviocytes [18]. Live cell imaging of calcium mobilization in the synoviocyte cultures was conducted in the presence of TRPV1 agonists and rapid changes in bath temperature or pH. In the current study, live cell imaging of synoviocyte cultures with a ROS sensitive dye indicated that TRPV1 agonists can elicit ROS elevations. ROS generation in clonal synoviocytes was observed in response to TRPV1 agonists capsaicin and resiniferatoxin, and hydrogen peroxide. These enhanced ROS responses were blocked by a TRPV1 antagonist or ROS scavenger, PBN, indicating that ROS generation in synoviocytes is TRPV1 mediated in the experimental conditions utilized.

Inflammatory products are produced locally in inflamed joints by activated synoviocytes. In the current study, we have detected activation responses mediated by TRPV1 and ROS that include increases in COX-2, RANTES and TNFα. An increase of COX-2 expression in synoviocyte cultures was blocked by either PBN or a TRPV1 antagonist. The ROS induced increase in TNFα was blocked by the ROS scavenger.

Conversely, ROS production in synovial fibroblasts harvested from patients has been shown in a previous study to be increased significantly by IL-6 and inhibited by methotrexate [27]. Modulation of cytokine synthesis and ROS production may contribute to the therapeutic effects of the anti-arthritic agent, methotrexate. These results suggest that the inflammatory events in the synovium of patients with rheumatoid arthritis are augmented locally by ROS generation and the subsequent production of cytokines. ROS released by chondrocytes during inflammation of the synovial membrane have also been associated with cartilage degradation in osteoarthritis (for review see [28]).

We also report here that synoviocytes increase TRPV1 protein expression after administration of ROS donors. We have shown previously that TNFα induces increases in TRPV1 in synoviocytes [19], and that TRPV1 activation increases TNF receptors (TNFR1) in dorsal root ganglia (DRG) through a ROS mediated signalling pathway [29].

It is well known that inflammatory mediators and products (Ex. TNFα, COX-2) can directly sensitize peripheral nerve endings [30‐37], suggesting a potential direct interaction between synoviocytes and peripheral nerves. A metabolite of COX-2, prostaglandin E2, is a components of the "inflammatory soup" used in many studies to activate peripheral nerves and was used combined with low pH in studies by Peter Reeh and colleagues [38]. Bolyard and colleagues have found that continuous exposure of rat sensory neurons to "inflammatory soup" results in continued sensitization [39]. We have found previously that afferent innervation contributes to knee joint inflammation since cutting dorsal roots or spinal cord block of glutamate receptors and dorsal root reflexes reduce inflammation in half [20, 40‐42]. It has been shown previously by others that desensitization of TRPV1 in the knee joint with millimolar doses of resiniferatoxin reduces hyperalgesia, locomotor effects and joint circumference in a similar knee joint inflammation model [43]. These findings suggest that tissue inflammation and injury would activate TRPV1 ion channels and excess ROS would be generated to increase nociceptive responses and other inflammatory response processes.

Other studies have shown that capsaicin acting through TRPV1 generates ROS production in several cell types, including breast epithelial, hepatoma and embryonic kidney cells [1‐3], but TRPV1 may suppress ROS formation in glioblastoma cells [4]. Both TRPV1 and P2X ion channels mediate the sensory transduction of ROS, especially H₂O₂ and hydroxyl radicals (•OH), by capsaicin-sensitive vagal lung afferent fibers [5]. Evidence from several studies has shown that free oxygen radicals stimulate sensory nerve endings through direct actions on capsaicin activated ion channels on primary afferent nerve endings [6, 24, 44‐46]. These studies have shown that administration of capsazepine or a nonspecific TRPV blocker, ruthenium red, prevents afferent nerve activation by free oxygen radicals. Increased primary afferent nerve activity is known to evoke central sensitization as seen following capsaicin injection [47], peripheral injury [48], or inflammation [49]. Importantly, these previous studies have demonstrated that administration of capsazepine has no effect on the baseline activity of either vagal or sympathetic cardiac afferents, but does abolish the response of afferent nerve fibers to capsaicin, H₂O₂, and xanthine + xanthine oxidase.

All of the rats were anesthetized with isoflurane and polyethylene tubing inserted through the foramen magnum intrathecally to the T10 level of the spinal cord for intrathecal injection of the ROS sensitive dye, DHE (dihydroethidium, Sigma; 50 μl in saline:DMSO, 1:1 injected intrathecally). The following day, half the animals were injected i.p. with the ROS scavenger PBN (100 mg/kg, Sigma) and the other half injected i.p. with saline. One hour later, the animals were anesthetized and their left knee joint injected with 3% carrageenan (plant product and irritant) and 3% kaolin (clay silt) (100 μl in sterile saline) to produce an acute inflammation (4-8 h) confined to the knee joint [20]. Inflammation developed in the knee joint within 4 h, and pain related behaviour was assessed.

Baseline measurement of circumference, knee joint temperature and nociceptive behaviours were recorded in male Sprague Dawley rats (250-350 g, n = 8), which served as their own controls in a repeated measures intra-subject experimental design. Experimental measures for comparisons of inflamed vs normal joints in rat included:

For mechanical sensitivity testing, animals were acclimated for 30 min to a cubicle placed on a wire mesh table top. Paw withdrawal responses to two von Frey monofilaments (10 and 47 mN, Semmes-Weinstein Anesthesiometer Kit #18011, Wood Dale, IL), were recorded two days before induction of arthritis and 4 h after the induction by an observer who was blind to the animals' condition. A single trial consists of 10 applications of the filament applied once every second. Trials were repeated no less than 3 min apart to allow the rat to cease any response and return to an inactive posture. Data are presented as a mean response (1-10) and plotted against time after induction of arthritis. An increase in response frequency reflects secondary mechanical allodynia for the 10 mN monofilament and secondary mechanical hyperalgesia for the 45 mN monofilament.

ROS in the spinal cord dorsal horn were visualized using a detector dye dihydroethidium (DHE, Molecular Probes, Eugene, OR). The dye is converted to a stable fluorescent product by ROS and allows visualization of ROS accumulated in spinal cord of rats during the time after knee joint inflammation. The DHE (50 μl, 100 mM in DMSO:saline, 1:1) was injected intrathecally through the foramen magnum of anesthetized rats onto the lumbar level of the spinal cord with a blunt 30 G needle attached to a 250 μl Hamilton syringe. Animals were allowed to recover and knee joint inflammation was induced 24 h later. The tissue stable fluorescent ROS indicator DHE was visualized in aldehyde fixed and cut spinal cord at the conclusion of the behavioral study, i.e. 6 h after induction of arthritis. Animals were anesthetized with an injection of Sleepaway pentobarbital solution and transcardially perfused with 0.9% heparinized saline (500 ml) followed by 4% phosphate buffered paraformaldehyde (1000 ml). The spinal cord was dissected, cryoprotected overnight in 30% buffered sucrose and frozen sectioned (15 μm). Spinal cord sections were cut and mounted on gel coated slides and DHE fluorescence visualized with rhodamine filter settings on a Nikon E1000 photomicroscope equipped with the MetaVue Optical Imaging system.

In the 24-48 h prior to experiments, cells were harvested and plated onto 15 mm quartz glass coverslips. Standard bath solutions consisted of (in mM) 150 NaCl, 5.5 KCl, 1 MgCl₂, 4 CaCl₂, 5 glucose, and 10 HEPES, adjusted to pH 7.4 at RT with NaOH (330 mosmol/l). On the day of the experiment, primary or SW982 synoviocytes were pre-loaded with a ROS-sensitive dye, 2,7-dichlorofluorescein diacetate (DCFDA, 2-4 μM for 30 min at a room temperature, Molecular Probes, Invitrogen, Eugene, OR), a non-fluorescent chemical oxidized intracellularly by ROS to the highly fluorescent derivative DCF. After pre-loading, the cells were washed in a physiological saline solution, allowed to stay from 30 min to 2 h at a room temperature, placed in a continuously superfused thermoregulated chamber (Warner Instruments, Hamden, CT) and then mounted to the stage of inverted Nikon Diphot microscope for recording. Application of chemicals onto individual cell was achieved by gravity flow from a large-bore pipette placed 300-400 μm from the cell. The dye-loaded cells were excited at 488 nm and emitted light collected using 490 nm long pass, or 535 nm (bandwidth 40 nm) filter. Real-time live cell imaging at high power was done with a Polychrome II monochromator (TILL Photonics, Munich, Germany) controlled by X-chart software (HEKA, Heidelberg, Germany) and with an ITC-18 computer interface (Instrutech, Port Washington, NY) for single cell fluorescence excitation. The resulting DCF emissions were detected with a Hamamatsu R928 photomultiplier. Imaging was performed in Dubecco's PBS (pH 7.4, RT) using a Nikon TE2000-S Eclipse quantitative fluorescence live-cell imaging system equipped with a high sensitivity 12 digital monochrome cooled Coolsnap ES CCD camera (Photometrics, Tucson, AZ). Lambda LS illumination and a Lambda 10-B Smart Shutter (Sutter Instruments, Novato, CA) were controlled by Metafluor software (Molecular Devices, Sunnyvale, CA). Metamorph (Molecular Devices, Sunnyvale, CA) software was used for digital image processing, and Excel scientific software (SPSS, Chicago, IL) was used for conversion and analysis of acquired data. Experiments were repeated a minimum of three times each.

Preincubation of the SW982 cells with TRPV1 antagonists capsazepine (5 μM, Calbiochem, La Jolla, CA), SB-366791 (30 μM, Tocris, Ellisville, MO) or ROS scavenger, PBN (2 mM, Sigma, Chemical, St. Louis, MO) was performed 2 h before addition of resiniferatoxin (100 nM) or capsaicin (1 μM, Tocris, Ellisville, MO). After application of resiniferatoxin, the cells were incubated for 2 h at 37°C in an incubator at ambient CO₂. After 2 h, the cell culture supernatant was aspirated and replaced with regular media or with regular media containing capsazepine or PBN at concentrations noted above. The cultures were then incubated overnight (37°C, 5% CO₂) prior to removal of supernatant for Western blot analysis of ROS metabolites and immunocytochemical localization of TRPV1, COX-2, and TNFα.

As an indicator of TRPV1 mediated ROS released into the supernatant, ROS metabolites 2,5-DHBA and 2,3-DHBA were trapped with salicylic acid [50] and the samples analyzed by high performance liquid chromatography with electrochemical detection (HPLC-ECD). Given that the hydroxyl radical is a short-lived, highly reactive chemical, it cannot itself be collected and preserved in a sample tube. The salicylic acid trapping agent (2.5 mM) reacts with the hydroxyl radical to produce DHBAs. The level of hydroxyl radical production in response to the TRPV1 activation was compared to that in media collected from control cultures from three experiments. Comparison of the results based on the quantities of increased 2,3-DHBA were considered reliable because there is some endogenous enzymatic formation of 2,5-DHBA.

After treatments, the cells on glass coverslips were washed with phosphate buffered saline (PBS, with 1% bovine serum albumin (BSA) and 0.9 mM CaCl₂ and 0.5 mM MgCl₂) and fixed with a mixture of methanol and 4% freshly mixed paraformaldehyde (1:1, v/v) for 10 min at room temperature. The coverslips were washed and incubated at room temperature for 30 min with 3% nonspecific species serum. The primary antibody (1:1,000 anti-TRPV1, Alomone, Isreael; 1:1,000 anti-COX-2, BD Transduction Labs San Diego CA; 1:500 anti-TNFα, R&D Systems, Minneapolis, MN) was diluted in PBS with 1.0% normal goat serum (NGS, Sigma Chemical). The primary antibody was allowed to incubate on the glass coverslips for 48 h at 4°C, or overnight at room temperature. The glass coverslips were then washed and incubated for 30-60 min at room temperature with the appropriate species secondary IgG antibody with a fluorescent ALEXA tag (1:1000 dilution, Alexa Fluor 588 green; Molecular Probes, Inc. Eugene OR). After final washes, the stained coverslips were inverted onto a dot of mounting media containing nuclear counterstain DAPI (VectaShield Hardset, Vector Labs, Burlingame, CA). Immunocytochemical controls included the absence of the primary or secondary antibodies, or were done in the presence of matched serum at the same dilution. The specificity of the antibodies was confirmed for TRPV1 by adsorption controls with P-19 N-terminus peptide provided by Santa Cruz, and for TRPV4 with peptide 853-871 supplied by Alomone Labs (1:1 w/w, 30 min). There was no staining in either of the method controls or the peptide blocked controls. ROS donor tBOOH (100 mM) stimulated increase in TRPV1 was greatly reduced by co-incubation with a TRPV1 inhibitor, SB-366791 (30 mM). Stained cells were visualized with a Nikon FXA microscope equipped with a Coolsnap digital low light camera and MetaVue image capture software (Nikon Instruments, Inc. Melville, NY).

Low passage SW982 cell cultures were grown in 75 mm flasks until they reached 70-80% confluency. Proteins were extracted from control cultures without treatment and after treatment with H₂O₂ (1 mM) for 8 h. Western Blot analysis was performed on whole cell protein fractions extracted from synovial SW982 cells using the ProteoExtract Subcellular Proteome Extraction kit (Calbiochem, Cincinnati, OH). Proteins transferred to membranes were probed with anti-TRPV1 antibody (VR1, P-19, 1:1000, Santa Cruz, Burlingame, CA) at 4°C overnight. Amounts of total protein loaded per lane included: 15 μg for cytosolic and membrane fractions, 10-15 μg for whole cell and 5 μg for rat spinal cord lysate as a positive control. Western blot analysis was performed with samples from three experiments.

Statistical analyses for behavioural studies were performed using Prism software. Descriptive statistics for these data are expressed as mean ± S.D. Behavioral data were compared with non-parametric repeated measures analysis of variance using Mann-Whitney U post-hoc tests unless otherwise specified. For measurements of ROS concentrations, the overall percent change from baseline values over time and differences between groups were compared with a multivariate analysis of variance (MANOVA). Sigma Plot and Sigma Stat scientific software (SPSS, Chicago, IL) were used for conversion and analysis of acquired data. Data are reported as means ± SEM. Within subjects, comparisons were made for baseline, time of injection peak, and time of peak increase. All statistical comparisons were evaluated at an alpha level of significance of 0.05.