Prostaglandin E2 (PGE2) is both an inflammatory mediator released at the site of tissue inflammation and a neuromodulator that alters neuronal excitability and synaptic processing. The effects of PGE2 are mediated by four G-protein-coupled EP receptors (EP1-EP4). Here we show that the EP4 receptor subtype is expressed by a subset of primary sensory dorsal root ganglion (DRG) neurons, and that its levels, but not that of the other EP1–3 subtypes, increase in the DRG after complete Freund' adjuvant-induced peripheral inflammation. Administration of both an EP4 antagonist [AH23848, (4Z)-7-[(rel-1S,2S,5R)-5-((1,1′-biphenyl-4-yl)methoxy)-2-(4-morpholinyl)-3-oxocyclopentyl]-4-heptenoic acid] and EP4 knockdown with intrathecally delivered short hairpin RNA attenuates inflammation-induced thermal and mechanical behavioral hypersensitivity, without changing basal pain sensitivity. AH23848 also reduces the PGE2-mediated sensitization of capsaicin-evoked currents in DRG neurons in vitro. These data suggest that EP4 is a potential target for the pharmacological treatment of inflammatory pain.
Nonsteroidal anti-inflammatory drugs (NSAIDs) are the most commonly used class of analgesics. Traditional NSAIDs inhibit prostaglandin synthesis through a nonselective blockade both of the constitutively expressed cyclooxygenase COX-1 and the inducible isomer COX-2. However, their clinical use is hampered by side effects, most notably gastric erosion, ulceration, and hemorrhage. Although selective COX-2 inhibitors were designed to prevent the adverse effects mediated by inhibition of COX-1, prolonged use of COX-2-selective inhibitors may, like NSAIDs, confer a risk for cardiovascular events, including myocardial infarction and stroke. The cause of the cardiovascular adverse effects is uncertain but may include an imbalance in prostacyclin and thromboxane levels in the endothelium and blockade of prostanoid actions on renal function. Identification of therapeutic targets downstream of COX may provide an opportunity for the development of analgesics that interfere with prostanoid proinflammatory and pronociceptive actions but with less cardiovascular risk.
Prostaglandin E2 (PGE2) is the principal proinflammatory prostanoid and contributes, in particular, to one of the key features of inflammation, pain hypersensitivity. At the site of inflammation, PGE2 sensitizes peripheral nociceptors through activation of EP receptors present on the peripheral terminals of these high-threshold sensory neurons, reducing threshold and increasing responsiveness, the phenomenon of peripheral sensitization (Omote et al., 2002a). PGE2 is also produced in the spinal cord after tissue injury (Samad et al., 2001), where it contributes to central sensitization (Minami et al., 2001), an increase in excitability of spinal dorsal horn neurons that produces pain hypersensitivity.
Four PGE2 G-protein-coupled receptor subtypes (EP1, EP2, EP3, and EP4) that are products of different genes have been identified, and these mediate the diverse effects of the prostanoid based on their differential tissue distribution and coupling to intracellular signal transduction pathways. EP2 and EP4 are coupled to Gs, some splice variants of EP3 are coupled to Gs, and some are coupled to Gi, whereas EP1 is coupled to Gq/G11. Studies performed either in mutant mice lacking the individual PG receptors (Murata et al., 1997; Minami et al., 2001; Stock et al., 2001) or with EP receptor-specific ligands (Minami et al., 1994) have not yet provided a coherent picture of which EP receptor(s) are responsible for inflammatory pain. This is partly due to the fact that PGE2 facilitates nociception at multiple levels in the neuraxis and that multiple receptors are involved. EP1, EP2, EP3, and EP4 (Oida et al., 1995) receptors are expressed by dorsal root ganglion neurons, whereas EP2 receptors are expressed by spinal cord neurons (Kawamura et al., 1997). We have now investigated the expression and regulation of EP4 in DRG neurons after peripheral inflammation and have explored its involvement in inflammatory pain by either administering an EP4 antagonist or by knockdown of EP4 in vivo with intrathecally delivered shRNA.
Materials and Methods
Animals. All experiments were carried out with the approval of the Massachusetts General Hospital subcommittee on research animal care. Adult male (∼200 g) Sprague-Dawley rats were used. Peripheral inflammation was induced by intradermal injection of 0.1 ml of complete Freund's adjuvant (5.0 mg/ml heat-killed Mycobacterium butyricum in mineral oil, CFA) into the plantar surface of the left hind paw. The adjuvant was injected into rats anesthetized with 3% isoflurane. Under 2.5% isoflurane anesthesia in an intrathecal catheter (32G; Recathco, Allison Park, PA) was implanted into the intrathecal space of the vertebral column with its tip at the lumbar (L3–L5) level through a cisternal incision. The catheter was connected to an osmotic minipump (model 2001 delivering 1 μl/h; Alzet, Cupertino, CA). EP4 shRNA and MS shRNA were continuously infused at 200 μg/day for 7 days. The EP4 antagonist AH23848 [(4Z)-7-[(rel-1S,2S,5R)-5-((1,1′-biphenyl-4-yl)methoxy)-2-(4-morpholinyl)-3-oxocyclopentyl]-4-heptenoic acid hemicalcium salt] (Sigma-Aldrich, St. Louis, MO) (Davis and Sharif, 2000) was injected intraperitoneally, and inflammatory nociceptive hypersensitivity was measured 30 min postinjection.
shRNA. shRNA was synthesized using the T7 Ribomax express RNA system (Promega, Madison, WI). Short interfering RNA (siRNA) sequences were designed with the aid of shRNA Target Designer (Promega). The top strand for the hairpin shRNA1 was 5′-GGATCCTAAT ACGACTCACT ATAGGAAGAC TGTGCTCAGT AATTCAAGAG ATTACTGAGC ACAGTCTTCC-3′, and the bottom strand was 5′-AAGGAAGACT GTGCTCAGTA ATCTCTTGAA TTACTGAGCA CAGTCTTCCT ATAGTGAGTC GTATTAGGAT CC-3′. The top strand for the hairpin shRNA2 was 5′-GGATCCTAAT ACGACTCACT ATAGACTGGA CCACCAACGT AATTCAAGAG ATTACGTTGG TGGTCCAGTC-3′, and the bottom strand was 5′-AAGACTGGAC CACCAACGTA ATCTCTTGAA TTACGTTGGT GGTCCAGTCT ATAGTGAGTC GTATTAGGAT CC-3′. The top strand for the mismatch control shRNA was 5′-GGATCCTAAT ACGACTCACT ATAGACTGCT CCAGCAACCT AATTCAAGAG ATTAGGTTGC TGGAGCAGTC-3′, and the bottom strand was 5′-AAGACTGCTC CAGCAACCTA ATCTCTTGAA TTAGGTTGCT GGTCCAGTCT ATAGTGAGTC GTATTAGGAT CC-3′. The concentration of shRNA was determined by spectrophotometry (OD 260).
Immunohistochemistry. Sections were fixed with 4% paraformaldehyde in 0.1 M phosphate buffer for 15 min at 4°C. For colocalization of EP4 and NF200 or peripherin, sections were immunostained for EP4 using tyramide signal amplification (PerkinElmer Life and Analytical Sciences, Boston, MA). Following primary antibody incubation (1:2000 EP4; PerkinElmer Life and Analytical Sciences) in TNB buffer (0.1 M Tris-buffered saline, pH 7.4, containing 1% blocking reagent) overnight at RT, sections were incubated with biotin-conjugated anti-rabbit antiserum (1:100 in TNB buffer, 2 h at RT), streptavidin-horseradish peroxidase (1:100 in TNB buffer, 30 min at RT), and fluorophore tyramide (1:50 in amplification diluent, 10 min at RT). After washing, sections were incubated with mouse anti-NF200 (1:1000; Sigma) or mouse antiperipherin (1:1000; Sigma) antibody in 0.1 M PBS containing 1% bovine serum albumin and 0.5% Triton X-100 for 16 h at RT, followed by rhodamine-conjugated anti-rabbit incubation (1:200 in 0.1 M PBS for 2 h at RT).
Image Analysis. Four EP4 antibody-stained sections (minimal separation of 90 μm) were selected from the L5 DRG from four animals in each group to analyze the proportion of EP4-labeled neuronal profiles and determine size-frequency distribution. Signals were analyzed under 400× magnification. Signal intensity and area were calculated using the NIH Image software system. Only neurons with visible nuclei were analyzed. Neurons were determined as EP4-positive if their signal intensity was 3-fold higher than background. All counting was performed blinded.
In Situ Hybridization. A digoxigenin (DIG)-labeled RNA probe was transcribed from a human EP4 fragment cloned into pCRII (Invitrogen, Carlsbad, CA) using 10× DIG-labeling mixture and SP6 RNA polymerase (Roche, Indianapolis, IN). Signal specificity was confirmed by the absence of signal when sections were hybridized with the sense probe. Tissue preparation, sectioning, and fixation proceeded as described for the immunohistochemistry. After washing in 0.1 M PBS, sections were acetylated in 0.25% acetic anhydride in 0.1 M triethanolamine and 0.9% NaCl for 15 min at room temperature. Sections were then incubated with RNA probe in hybridization buffer (Sigma) at 60°C for 12 h. After three washes in 50% formamide in 2× SSC at 60°C for 15 min, sections were incubated with anti-DIG antibody (1:1000; Roche) diluted in DIG buffer 1 (100 mM Tris-HCl, pH 7.5, and 150 mM NaCl), 1% blocking reagent, and 0.3% Triton X-100 for 12 h at 4°C. Signals were visualized by incubating with 4.5 ml/ml 5-bromo-4-chloro-3-indolyl phosphate and 3.5 ml/ml nitro blue tetrazolium in DIG buffer 3 (100 mM Tris-HCl, pH 7.5, 100 mM NaCl, and 50 mM MgCl2.
Western Blots. DRGs or HEK293T cells transfected with EP4 were homogenized-sonicated in a lysis buffer (200 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40, and 10% glycerol) containing a mixture of proteinase inhibitors (Roche). Protein samples were separated on a SDS-polyacrylamide gel electrophoresis gradient gel (4–15%; Bio-Rad, Hercules, CA) and transferred to a polyvinylidene difluoride filter (Immobilon-P; Millipore Corporation, Billerica, MA). The blots were blocked with 6% skim milk (Difco, Detroit, MI) in TBS-T (Tris-buffered saline, 0.15% Tween 20) for 1 h and incubated with EP4 antibody (1:6000) overnight at 4°C. The blots were incubated with horseradish peroxidase-conjugated anti-rabbit secondary antibody (1:2000; GE Healthcare, Little Chalfont, Buckinghamshire, UK) for 1 h at room temperature, developed in ECL solution (PerkinElmer Life and Analytical Sciences) for 1 min, and exposed on X-ray film (GE Healthcare) for 1 to 10 min. For densitometric analysis, blots were scanned using the Model GS-710 imaging densitometer, and results were expressed as the ratio of EP4 immunoreactivity to β-actin immunoreactivity.
Behavioral Assessment of Thermal and Mechanical Stimulation. All behavioral tests were performed blinded to the treatment. The thermal nociceptive threshold was measured using a radiant heat technique (Hargreaves et al., 1988). Latency between the application of a focused light beam and a hind paw-withdrawal response was measured to the nearest 0.1 s, with a cut-off time of 20 s. The punctate mechanical withdrawal threshold was determined using a series of calibrated von Frey filaments (Stoelting, Inc., Wood Dale, IL), ranging from 0.23 to 59.0 g. Animals were placed in a plastic cage with a metal mesh floor, allowing them to move freely. The animals were habituated to the testing environment and allowed to acclimatize to this environment for at least 20 min before the experiment, and the filaments were presented to the midplantar surface. Withdrawal threshold measurements were collected using an up-down method.
Electrophysiology. Primary adult rat DRG neuron cultures were prepared as described and incubated in Neurobasal medium containing nerve growth factor (100 ng/ml; Sigma). Whole-cell patch-clamp recordings were performed 2 to 4 days after dissociation of the DRG. Conventional whole-cell patch clamp was used. In brief, only small DRG neurons (<30 μm) were recorded. A typical recording pipette resistance was approximately 2 MΩ; once the whole-cell configuration was achieved, the neurons were held at –70 mV, and voltage error was not compensated. Signals were recorded by Axon 200A (Molecular Devices, Sunnyvale, CA) in gap-free mode filtered at 5 KHz and sampled at 2 KHz with pClamp 8 (Molecular Devices). External solution (pH 7.4 with NaOH) was as follows: 145 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM glucose, and 10 mM HEPES. Internal solution (pH 7.4 with KOH) was as follows: 15 mM KCl, 130 mM potassium aspartate, 1 mM MgCl2, 0.5 mM CaCl2, 10 mM EGTA, 10 mM HEPES, and 2 mM glucose. Capsaicin was delivered for 30 s through a custom-made multibarrel fast exchange drug applicator positioned ∼300 μm from the recording neurons to achieve a final concentration of ∼0.2 μM. To avoid inadvertent desensitization due to previous application of capsaicin, only one cell was studied from each culture dish. Between two capsaicin applications, cells were treated with either vehicle or PGE2 (10 μM, 3 min). In the EP4 antagonist AH23848 study, AH23848 (10 μM; Tocris Cookson Inc., Bristol, UK) was pretreated for 30 min before recording. Capsaicin receptor desensitization was expressed as the ratio of the 2nd peak response to the 1st peak.
Statistics. All behavior and Western data are presented as mean ± S.D. ANOVA followed by Tukey's post hoc test was used for statistical analysis, with P < 0.05 considered significant. The capsaicin response data are presented as mean ± S.E.M., and P values were analyzed by one-way ANOVA.
EP4 Receptors Are Expressed by a Subset of Dorsal Root Ganglion Neurons. The expression of EP4 receptors was analyzed in the lumbar (L5) DRG and spinal cord of naive rats. EP4 was detected by immunohistochemistry in many cellular profiles in the DRG (Fig. 1A). Double label indicated expression of EP4 in both small unmyelinated (peripherin-positive) and myelinated (NF200-positive) primary sensory neurons (Fig. 1B) but not in spinal cord neurons (data not shown). EP4 mRNA expression in DRG neurons as detected by in situ hybridization had a neuronal expression pattern very similar to that of immunolabeled EP4 (Fig. 1C). A Western blot using the same antibody as that for the immunohistochemistry revealed a single band of the expected size (Fig. 1D). This antibody also recognized recombinant EP4 expressed in HEK cells (data not shown). EP4 mRNA levels analyzed using real-time quantitative PCR revealed much higher expression in the DRG than the spinal cord (>12-fold), and both immunohistochemical and Western blot analyses found only low levels of EP4 protein in the dorsal horn (data not shown). The DRG/dorsal horn mRNA ratios for EP1, 2, and 3 were 1.2, 2.3, and 1.3, respectively.
Increased Expression of EP4 Receptors in the DRG after Peripheral Inflammation. After induction of a localized area of peripheral inflammation in the hind paw by an intraplantar injection of CFA, EP4 protein levels increased significantly in the L5 DRG ipsilateral to the inflamed hind paw but not in the contralateral noninflamed hind paw (Fig. 2A). The increase in EP4 immunoreactivity was present from days 2 through 7 (the longest time measured). In the L5 DRG ipsilateral to a CFA-treated hind paw (2 days), 57 ± 9.8% peripherin-positive neuronal profiles were found to be EP4-positive compared with 39 ± 6% in the contralateral DRG (P < 0.05, ipsilateral versus contralateral) (Fig. 2B). However, the proportion of NF200-positive cells that are EP4-positive did not differ between DRGs ipsilateral and contralateral to the inflamed hind paw (data not shown). Real-time reverse transcriptase-PCR analysis of naive and inflamed L5 DRGs confirmed an increase in EP4 mRNA expression in the DRG but showed that mRNA for the other EP receptors, EP1, EP2, and EP3, did not change in the DRG (Fig. 2C). In the spinal cord, no change in EP1, EP2, EP3, or EP4 mRNA levels occurred after the CFA-induced inflammation (Fig. 2C).
The EP4 Antagonist AH23848 Attenuates CFA-Induced Pain Hypersensitivity. The inflammation produced by intraplantar CFA results in pronounced thermal and mechanical hypersensitivity with a reduction in the thermal (radiant heat) withdrawal latency (Fig. 3A) and in the von Frey mechanical threshold for eliciting a withdrawal response (Fig. 3B). The ipsilateral thermal withdrawal latency was reduced from 17.4 ± 1.8 to 10.9 ± 0.6 s, the mechanical threshold was reduced from 20.8 ± 2.5 to 8.1 ± 1.3 g (n = 7 or more). Intraperitoneal injection of the EP4 antagonist AH23848 (Davis and Sharif, 2000) resulted in a dose-dependent reduction in the inflammatory nociceptive hypersensitivity measured 30 min postinjection (Fig. 3). AH23848 produced no change in thermal latency or mechanical withdrawal thresholds on the contralateral noninflamed paw (data not shown).
AH23848 Attenuates PGE2 Sensitization of Capsaicin Currents in DRG Neurons. To test whether PGE2 has direct actions via EP4 in DRG neurons, we used the in vitro PGE2 sensitization of capsaicin currents as a model system. Capsaicin is a ligand for the TRPV1 receptor. Repeated exposure of TRPV1 to capsaicin results in a desensitization of the ion channel that is reduced by exposure of the cells to PGE2 (Bhave et al., 2002). More than 90% TRPV1-positive cells observed in DRG cultures were found to express EP4 (Fig. 4A). This indicates that EP4 is expressed by heat-sensitive nociceptors.
To study the desensitization of capsaicin-activated current in DRG neurons, we applied a series of brief (30-s) pulses of 0.2 μM capsaicin at 3-min intervals and measured current using whole-cell patch-clamp recordings. Capsaicin-responsive DRG neurons showed a pronounced desensitization on repeated exposure. The mean current amplitude on the second application was 50 ± 5% (n = 11) of that found with first application (Fig. 4, B and E). The extent of the desensitization was significantly decreased when the cells were pretreated for 3 min with 10 μM PGE2. In this situation, the current amplitude on the second application of capsaicin was 74 ± 10% (n = 9) of the first application (Fig. 4, C and E). The EP4 antagonist AH23848 (1 μM) blocked PGE2 reduction of the desensitization (49 ± 6%, n = 11) without producing an effect by itself (eight cells) (Fig. 4, D and E).
Intrathecal shRNA Infusion Knocks Down EP4 and Reduces CFA Pain Hypersensitivity. In preliminary experiments, we injected DIG-labeled EP4 shRNA intrathecally and found extensive label in the lumbar DRG but only limited penetration into the spinal cord (data not shown). To assess potential neurotoxicity, increasing doses of the EP4 shRNA were continuously delivered intrathecally to naive rats (n = 3 per dose) via an indwelling cannula with its tip positioned at the levels of L4/L5 and attached to an osmotic minipump. Doses up to 8.3 μg/h did not elicit any signs of hind limb paralysis, sensory loss, vocalization, or anatomical damage to the spinal cord and were used in all subsequent experiments. This dose of two independent EP4 shRNAs resulted in a down-regulation of EP4 mRNA (Fig. 5A) and protein levels (Fig. 5B) in the L5 DRG (69% reduction). The MS shRNA and vehicle controls had no effect (Fig. 5). The shRNA targeted against EP4 had no effect on the levels of the mu-opiate receptor, another G-protein-coupled receptor (Fig. 5B). EP4 shRNAs delivered intrathecally, but not MS shRNA or vehicle, produced within 12 h a significant reduction (P < 0.05) in the CFA-induced thermal and mechanical hyperresponsiveness (Fig. 5, C and D). To address the issue of the specificity of EP4 shRNA, we quantified by real-time reverse transcriptase-PCR levels of EP1, EP2, EP3, and EP4 mRNA following EP4 shRNA treatment. Only EP4 mRNA levels showed a significant decrease (P < 0.05) (Fig. 5E). Similar behavioral effects were produced by the EP4 antagonist and shRNA. The shRNA was administered intrathecally to target lumbar spinal cord and DRG cells that produce the EP4 receptor to reduce the synthesis of the protein, whereas the antagonist was administered systemically to block activity of receptors distributed in both the peripheral and central terminals of sensory neurons.
PGE2 is a major mediator of peripheral inflammation. This prostanoid, when produced at the site of tissue injury and inflammation after induction of COX-2 and prostaglandin synthases, acts on capillaries to produce vasodilatation and edema and on macrophages to stimulate release of cytokines (McCoy et al., 2002). Therefore, inhibition of PGE2 synthesis produces an anti-inflammatory effect. However, PGE2 also has a direct action on the peripheral terminals of nociceptor sensory neurons, reducing their threshold to peripheral stimuli by promoting the phosphorylation of TRPV1 and other TRP channels, and increasing terminal membrane excitability by phosphorylating voltage-gate sodium channels (Lopshire and Nicol, 1998; Smith et al., 2000). Treatment with a monoclonal anti-PGE2 antibody reverses carrageenan-induced hyperalgesia to the same extent as NSAIDs, indicating that PGE2 plays a key role in inflammatory pain (Zhang et al., 1997). Peripheral sensitization contributes to pain hypersensitivity within the confines of inflamed tissue, the zone of primary hyperalgesia, and inhibiting PGE2 synthesis or action produces an analgesic action by reducing peripheral sensitization.
The central action of PGE2 includes a presynaptic action, increasing transmitter release from the central terminals of nociceptors (Vasko, 1995), and postsynaptic actions, including a blockade of glycinergic inhibition (Harvey et al., 2004). The low constitutive levels of COX-1 and COX-2 in the spinal cord generate a short latency, activity-dependent release of PGE2, and peripheral inflammation leads after several hours to a large induction of COX-2 in the spinal cord that contributes to inflammation-induced central sensitization and pain (Samad et al., 2001).
PGE2 exerts its cellular effects through four different G-protein-coupled EP receptors. Activation of the EP1 receptor leads to calcium influx and stimulation of EP2, and some isoforms of EP3 and EP4 receptors all increase intracellular cAMP through activation of adenylate cyclase, whereas activation of other EP3 receptor splice variants inhibit adenylyl cyclase (Negishi et al., 1995; Funk, 2001). We find that EP4 receptors are, as described previously (Oida et al., 1995), expressed by a subset of dorsal root ganglion but not by spinal cord neurons. The actions of PGE2 on peripheral sensitization could include an EP4 component, but any central action of PGE2 mediated by EP4 could act only presynaptically.
Dissecting out the specific contributions of the different EP receptors to inflammatory pain and indeed other functions has proven relatively difficult. Specific antagonists for each receptor are not available, and in consequence, knockout mice have been extensively used to study the function of each receptor. PGI2 receptor and EP3, but not EP1, EP2, or EP4 receptors contribute to abdominal writhing in response to an intraperitoneal proinflammatory irritant (Ueno et al., 2001), but another study found a contribution of EP1 in a similar model (Stock et al., 2001). Intrathecal PGE2 induces allodynia in wild-type and EP3-deficient mice but not in EP1-deficient mice (Minami et al., 2001). EP2-deficient mice also show a lack of intrathecal PGE2-evoked hyperalgesia (Reinold et al., 2005). Inoculation of mice with herpes simplex virus type-1 elicits both acute herpetic pain and a delayed “postherpetic” pain, and deficiency of EP3, but not EP1, PGI2 receptor, or TXA2 receptor, prostanoid receptors diminishes both the acute herpetic pain-like behavior and the delayed postherpetic pain (Takasaki et al., 2005). Local PGE2-induced thermal hyperalgesia is reduced in EP1-deficient mice (Moriyama et al., 2005), whereas EP4-deficient mice display a marked reduction in the development of joint inflammation in a model of rheumatoid arthritis (McCoy et al., 2002). It is difficult to get a clear view of the specific role of the EP receptors from these studies. Interpreting phenotypic differences in null mutations is problematic due to variations in genetic background, developmental effects, genetic compensation, and lack of regional or cell specificity.
Intrathecal injection of the EP1-specific antagonists SC-51089 and SC-51234A results in suppression of the second phase of formalin test (Malmberg et al., 1994). Another EP1-specific antagonist ONO-8711 decreases incision-induced mechanical hyperalgesia (Omote et al., 2002b), chronic constriction injury-induced hyperalgesia (Kawahara et al., 2001), and inflammatory pain (Nakayama et al., 2002). AH23848 is reported to be selective for EP4 receptors (Coleman et al., 1994). Binding studies using HEK293 cells expressing cloned EP2 and EP4 receptors produced Ki values of >100,000 and 8010 nM, respectively, for AH23848 (Boie et al., 1997), and AH23848B produced a >90% inhibition of PGE2-stimulated cAMP formation in human lens epithelial cells (Mukhopadhyay et al., 1999). Paradoxically, ONO-AE1–329, a selective EP4 receptor agonist (Ki = 9.7 nM for EP4, with selectivity at least 100-fold over EP1, EP2, and EP3), inhibits hyperalgesia and swelling in CFA-induced arthritis (Omote et al., 2002a).
One mechanism with a major role in inflammatory heat pain sensitivity is sensitization of TRPV1 receptors. Both EP2 and EP4, by virtue of their coupling with Gs to activate adenylate cyclase, increase intracellular adenosine 3′,5′-cyclic monophosphate levels, which in turn activate protein kinase A (PKA). Activation of PKA by PGE2 in DRG neurons enhances both capsaicin and heat-mediated responses (Lopshire and Nicol, 1998; Smith et al., 2000), an effect mimicked by membrane-permeant cAMP analogs (Southall and Vasko, 2001). A null mutation of the type I PKA regulatory subunit eliminates PGE2-induced heat pain hypersensitivity (Malmberg et al., 1997). Our study demonstrates that the EP4-specific antagonist AH23848 attenuates PGE2 sensitization of capsaicin currents in DRG culture and CFA-induced hypersensitivity in vivo. In HEK293 cells and DRG neurons, PGE2 increases TRPV1 activity via EP1 receptors in a PKC-dependent manner and through PKA via EP4 (Moriyama et al., 2005). These results suggest that both EP4 and EP1 receptors may contribute to the acute PGE2-induced peripheral sensitization of sensory neurons. EP4, because of its increased levels after peripheral inflammation, may play a more substantial role in peripheral chronic inflammatory pain. In DRG cultures pretreated with antisense oligonucleotides directed against EP3C and EP4 receptor subtype mRNA, the PGE2-augmented release of substance P and calcitonin gene-related peptide is abolished (Southall and Vasko, 2001), suggesting that these receptors mediate neurogenic inflammation.
RNAi induces sequence-specific gene silencing by production of double-stranded RNA (dsRNA), a phenomenon first discovered in Caenorhabditis elegans (Fire et al., 1998). Long dsRNA is processed by the nuclease “dicer” into 21 to 23 duplexes called siRNA. siRNAs associate with a RNA-induced silencing complex nuclease complex to degrade mRNA in a sequence-specific manner. A similar process occurs in mammals (Billy et al., 2001). However, in cultured mammalian cell lines, uninterrupted RNA duplexes longer that 30 bp trigger nonspecific deleterious cellular responses as the result of the activation of the dsRNA-dependent protein kinase (protein kinase R) and RNase L (Stark et al., 1998). This can be avoided by direct administration of 21 nucleotide siRNAs to elicit only a sequence-specific RNAi-mediated inhibition of gene expression (Elbashir et al., 2001). shRNA produced in vitro or expressed in vivo silences genes as effectively as short dsRNAs (Yu et al., 2002). The shRNA system is more effective than the tandem system, and shRNAs are more potent inducers of RNAi than siRNA (Siolas et al., 2005). Here we show that intrathecal delivery of shRNA suppresses EP4 expression in DRG neurons. Why shRNA penetrates the DRG more successfully than the spinal cord is not clear, although it may reflect in part the relative position of the tip of the intrathecal cannula targeted more caudally than the lumbar enlargement. Nevertheless, considering its effectiveness in vivo and relative lack of toxicity even at high concentrations, shRNA can be used to validate potential novel targets in sensory neurons where small molecule antagonists or inhibitors are not available.
In conclusion, our data indicate that the EP4 receptor is expressed by primary sensory neurons and that EP4 levels increase in the DRG after peripheral inflammation. Both administration of an EP4 antagonist and EP4 knockdown with shRNA attenuate inflammatory pain hypersensitivity. Therefore, EP4 receptor-specific antagonists or EP4 shRNA may be useful drugs for treating inflammatory pain, including conditions like rheumatoid and osteoarthritis. EP4 antagonists as analgesics may have an advantage over other approaches that target prostaglandins, such as NSAIDs, COX-2 inhibitors, and prostaglandin synthase inhibitors. This will be determined in part by the specific role of EP4 in other PGE2-mediated functions, such as renal homeostasis, endothelial and platelet function, control of blood pressure, and gastrointestinal mucosal function, which remain to be systematically investigated. The overall safety profile of selective EP4 antagonists should be nevertheless different from COX-2 inhibitors and NSAIDs that nonselectively inhibit all prostaglandin syntheses. In particular, EP4 antagonists would not be expected to block PGE2/PGI2 biosynthesis or affect the PGI2/TXA2 balance that may be an underlying cause of renal or cardiovascular side effects associated with COX-2 inhibitors. In the late phase of inflammation, COX-2 may actually facilitate the production of anti-inflammatory PGs, such as PGA1, that play roles in the resolution of inflammation by suppressing nuclear factor-kB activation and thereby COX-2 gene expression (Mandal et al., 2005). Unlike COX-2 inhibitors, therapies targeting the EP4 receptor may selectively reduce pain without prolonging inflammation.
This work was supported by Pfizer and the National Institutes of Health.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
ABBREVIATIONS: NSAID, nonsteroidal anti-inflammatory drug; PG, prostaglandin; PGE2, prostaglandin E2; EP receptor, prostaglandin E2 G-protein-coupled receptor; DRG, dorsal root, ganglion; COX, cyclooxygenase; CVS, cardiovascular system; CFA, complete Freund's adjuvant; siRNA, short interfering RNA; shRNA, short hairpin RNA; MS, mismatched; PKA, protein kinase A; TXA2, thromboxane A2; ANOVA, analysis of variance; AH23848, (4Z)-7-[(rel-1S,2S,5R)-5-((1,1′-biphenyl-4-yl)methoxy)-2-(4-morpholinyl)-3-oxocyclopentyl]-4-heptenoic acid; RT, room temperature; PBS, phosphate-buffered saline; PCR, polymerase chain reaction; DIG, digoxigenin; HEK, human embryonic kidney; RNAi, RNA interference; dsRNA, double-stranded RNA; TRPV, transient receptor potential vanilloid; SC-51089, dibenz(b,f)(1,4)oxazepine-10(11H)-carboxylic acid, 8-chloro-, 2-(1-oxo-3-(4-pyridinyl)propyl)hydrazide, monohydrochloride; SC-51234A, 8-chloro-dibenz(b,f)oxazepine-10(11H)-carbonyl-(2, 2-difluoro-3-hydroxy-3-(4-pyridyl)propanoylhydrazide; ONO-8711, 6-((2S,3S)-3-(4-chloro-2-methylphenylsulfonylaminomethyl)-bicyclo(2.2.2)-octan-2-yl)-5Z-hexanoic acid; ONO-AE1-329, 16-(3-methoxymethyl)phenyl-ω-tetranor-3,7-dithia prostaglandin E1.
- Received March 31, 2006.
- Accepted September 7, 2006.
- The American Society for Pharmacology and Experimental Therapeutics