Characterization of EP receptor subtypes responsible for prostaglandin E₂-induced pain responses by use of EP₁ and EP₃ receptor knockout mice

Abstract

1. Prostaglandin E₂ (PGE₂) is known to be the principal pro-inflammatory prostanoid and play an important role in nociception. To identify PGE receptor (EP) subtypes that mediate pain responses to noxious and innocuous stimuli, we studied them by use of EP₁ and EP₃ knockout (EP₁^−/− and EP₃^−/−) mice.

2. PGE₂ could induce mechanical allodynia in EP₁^+/+, EP₃^+/+ and EP₃^−/− mice, but not in EP₁^−/− mice. N-methyl-D-aspartate (NMDA), the substrate of nitric oxide (NO) synthase L-arginine, or the NO donor sodium nitroprusside administered intrathecal (i.t.) could induce allodynia in EP₃^−/− and EP₁^−/− mice. Activation of EP₁ receptors appears to be upstream, rather than downstream, of NMDA receptor activation and NO production in the PGE₂-induced allodynia.

3. Although PGE₂ produced thermal hyperalgesia over a wide range of dosages from 50 pg to 0.5 μg kg⁻¹ in EP₃^+/+ mice, it showed a monophasic hyperalgesic action at 5 ng kg⁻¹ or higher doses in EP₃^−/− mice. The selective EP₃ agonist, ONO-AE-248, induced hyperalgesia at 500 pg kg⁻¹ in EP₃^+/+ mice, but not in EP₃^−/− mice.

4. Saline-injected EP₁^−/− mice showed hyperalgesia, which was reversed by i.t. PGE₂ in a dose-dependent manner.

5. There was no significant difference in the formalin-induced behaviours between EP₁^−/− or EP₃^−/− mice and the cognate wild-type mice.

6. These results demonstrate that spinal EP₁ receptors are involved in the PGE₂-induced allodynia and that spinal EP₃ receptors are involved in the hyperalgesia induced by low doses of PGE₂. However, the formalin-induced pain cannot be ascribed to a single EP receptor subtype EP₁ or EP₃.

Introduction

Prostanoids are the products of the cyclo-oxygenase pathway of arachidonic acid metabolism and act as local mediators in various tissues under physiological and pathophysiological conditions (Narumiya et al., 1999). Since Vane (1971) first reported that aspirin-like drugs prevented the development of inflammation by blocking the synthesis of prostanoids, it has been widely accepted that prostanoids are involved in pain, fever, oedema, and various aspects of inflammation. Among them, prostaglandin E₂ (PGE₂) is considered to be the principal pro-inflammatory prostanoid and play an important role in nociceptive processing in the spinal cord as well as in the periphery (Yaksh et al., 1999). PGE₂ is released from the spinal cord in vivo upon various noxious stimuli and inflammatory insults. The intrathecal (i.t.) administration of PGE₂ into conscious mice induced hyperalgesia to noxious stimuli (Taiwo & Levine, 1988; Uda et al., 1990; Minami et al., 1994a) and allodynia to tactile innocuous stimuli (Minami et al., 1994a, 1994c). Conversely, i.t. delivery of cyclo-oxygenase inhibitors blocked pain responses induced by subcutaneous formalin, and i.t. substance P and N-methyl-D-aspartate (NMDA) (Malmberg & Yaksh, 1992a, 1992b).

The diversity of PGE₂ actions is believed to be the result of four PGE receptor subtypes EP₁ – EP₄ coupled to different signal transduction pathways (Narumiya et al., 1999). The EP₁ receptor is coupled to intracellular Ca²⁺ mobilization. Whereas EP₂ and EP₄ receptors are coupled to stimulation of adenylate cyclase via Gs, EP₃ receptor is coupled to inhibition of adenylate cyclase via Gi. We previously showed that PGE₂ induced thermal hyperalgesia over a wide range of dosages between 50 pg – 500 ng kg⁻¹ with two apparent peaks of 0.5 ng kg⁻¹ and 500 ng kg⁻¹. While the EP₃ agonist MB28767 showed a monophasic hyperalgesic action at as low as 50 pg kg⁻¹, the EP₂ agonist butaprost induced hyperalgesia at doses higher than 50 ng kg⁻¹, suggesting that the PGE₂-induced hyperalgesia is mediated by the EP₃ receptor at lower doses and by the EP₂ receptor at higher doses (Minami et al., 1994a). We also suggested that PGE₂-induced allodynia is mediated by the EP₁ receptor by use of EP receptor agonists and that the bifunctional EP₁ antagonist/EP₃ agonist ONO-NT-012 is a highly potent, simple competitive antagonist for the PGE₂-induced allodynia (Minami et al., 1994a; 1995b). On the other hand, Malmberg et al. (1994) suggested that i.t. EP₃ agonists produced thermal hyperalgesia and allodynia and that i.t. EP₁ antagonist significantly attenuated the second phase of formalin-induced pain. Previous studies demonstrated that PGE₂ plays versatile roles in pain transmission through different EP receptor subtypes in the spinal cord. However, studies in this area have been limited by the lack of potent and specific EP receptor antagonists so far.

To elucidate the pathophysiological significance of EP receptor subtypes, the gene targeting technique has been employed (Narumiya et al., 1999). In the present study, we examined the roles of EP₁ and EP₃ subtypes in the PGE₂-induced pain responses by use of the respective knockout mice (EP₁^−/− and EP₃^−/−). Furthermore we examined the role of EP₁ and EP₃ subtypes in the formalin test.

Methods

Chemicals

PGE₂ and ONO-AE-248 (Zacharowski et al., 1999) were generous gifts from Ono Central Research Institute (Osaka, Japan). NMDA, L-arginine and sodium nitroprusside (SNP) were purchased from Sigma (St. Louis, MO, U.S.A.). All chemicals were dissolved in sterile saline on the day of the experiments and kept on ice until used. All drugs, including saline, were coded to assure blind testing.

Animals

EP₁^−/− or EP₃^−/− mice were obtained by the gene targeting technique (Ushikubi et al., 1998). The animals were housed under conditions of a 12-h light – dark cycle, a constant temperature of 22±2°C and 60±10% humidity.

Studies on allodynia and hyperalgesia

Studies on allodynia and hyperalgesia were carried out as described previously (Minami et al., 1999). A 27-gauge stainless-steel needle (0.35 mm, o.d.) attached to a microsyringe was inserted between the L₅ and L₆ vertebrae and drugs in vehicle were injected slowly into the subarachnoid space of conscious mice by a slight modification of the method of Hylden & Wilcox (1980).

For allodynia, the mice were divided into various groups (n=5 – 6/group). Drug-treated groups were injected with 5 μl of vehicle containing various doses of test agents. Control mice were given physiological saline (5 μl). After the i.t. injection, each mouse was placed in an individual 13×8.5×13 cm Plexiglas enclosure with wood chips on the floor and observed. Allodynia was assessed once every 5 min over a 5-min period by light stroking of the flank of the mice with a paintbrush. The allodynic response was ranked as follows: 0, no response; 1, mild squeaking with attempts to move away from the stroking probe; 2, vigorous squeaking evoked by the stroking probe, biting at the probe, or strong efforts to escape. The maximum possible score for allodynia of six mice was 2×6=12 in any 5 min period and was taken as 100%.

For hyperalgesia, mice were placed on a hot plate maintained at 52.5°C, and the elapsed time until the mice showed the first avoidance responses (licking the feet, jumping or rapidly stamping the paws) was recorded. The response time of the mice to the hot plate was measured 30 min after i.t. injection, the points of the maximal hyperalgesic effect obtained with PGE₂ (Uda et al., 1990).

Formalin test

Formalin test was carried out as described previously (Nakano et al., 2000), essentially according to the procedure reported by Hunskaar & Hole (1987). The mice were divided into various groups (n=8 – 10/group). Using a minimum of restraint, 20 μl of 2% formalin in 0.9% NaCl was injected subcutaneously into the right dorsal hind paw of the mouse using a microsyringe with a 26-gauge needle. After the formalin injection, each mouse was placed to the observation chamber. The amount of time spent licking and biting the injected paw was measured with a hand-held stop-watch for 5 min from 0 to 30 min. Two distinct periods of high licking activity can be identified, an early phase lasting the first 5 min and a late phase lasting from 15 to 30 min after the injection of formalin.

The animals were used only for one measurement in each experiment.

This study was conducted with the approval of the local ethics committee and in accordance with the guidelines of the Ethics Committee of the International Association for the Study of Pain (Zimmermann, 1983).

Statistics

Data for hyperalgesia were analysed by parametric ANOVA and statistical significance (P<0.05) was further examined by Duncan's test. Data for allodynia and formalin test were analysed by non-parametric ANOVA and statistical significance (P<0.05) was further examined by Williams' test and Dunnett's test, respectively, for multiple comparison.

Results

Effect of i.t. PGE₂ on allodynia in EP₁^−/− and EP₃^−/− mice

To assign a PGE receptor subtype to PGE₂-induced mechanical allodynia, we examined whether PGE₂ could induce allodynia in EP₁^−/− and EP₃^−/− mice. While i.t. administration of saline did not evoke any response to tactile stimuli applied to the flank in wild-type mice, i.t. PGE₂ (500 ng kg⁻¹) resulted in prominent agitation responses, such as vocalization, biting, and escape from the probe. As shown in Figure 1, i.t. PGE₂ (500 ng kg⁻¹) induced allodynia over the 50-min experimental period in EP₁^+/+ and EP₃^+/+ mice. PGE₂ could induce mechanical allodynia in EP₃^−/− mice, but not in EP₁^−/− mice. These results clearly demonstrate that the EP₁ receptor is involved in the PGE₂-induced allodynia.

Figure 1.

Time courses of allodynia induced by PGE₂ in wild-type and receptor-deficient mice. PGE₂ (500 ng kg⁻¹) was injected i.t. in EP₁^+/+ and EP₁^−/− (a), and EP₃^+/+ and EP₃^−/− (b) mice. Assessment of allodynia was made as described under ‘Methods'. The value (mean±s.e.mean) represents the per cent of the maximum possible cumulative score of 5 – 6 mice evaluated every 5 min.

Effect of i.t. NMDA, L-arginine, and SNP on allodynia in EP₁^−/− and EP₃^−/− mice

We previously showed that PGE₂-induced allodynia is mediated by activation of NMDA glutamate receptor and following nitric oxide (NO) production in the spinal cord (Minami et al., 1994b; 1995a; 1999; Eguchi et al., 1999). Intrathecal administration of NMDA (500 ng kg⁻¹), the substrate of NO synthase L-arginine (250 μg kg⁻¹), or the NO donor SNP (500 ng kg⁻¹) induced mechanical allodynia in both EP₁^−/− and EP₃^−/− mice. When the scores of allodynia obtained for the overall 50 min were cumulated and expressed as a per cent of the maximum possible score, the allodynic score of any one agent was beyond 70% (Table 1). These values obtained in EP₁^−/− and EP₃^−/− mice were almost the same as those in wild-type mice.

Table 1.

Allodynia induced by various agents in EP₁^−/− and EP₃^−/− mice

Effect of i.t. PGE₂ on hyperalgesia in EP₁^−/− and EP₃^−/− mice

To characterize a PGE receptor subtype(s) involved in PGE₂-induced hyperalgesia, we examined whether PGE₂ could induce hyperalgesia in EP₁^−/− and EP₃^−/− mice. As shown in Figure 2a,b, there was no difference in the latency periods (17.2±1.2 s and 17.8±0.9 s) (mean±s.e.mean, n=10) between EP₃^+/+ and EP₃^−/− mice 30 min after i.t. saline injection. While i.t. injection of PGE₂ to EP₃^+/+ mice produced a hyperalgesic action over a wide range of dosages from 50 pg to 500 ng kg⁻¹ (Figure 2a), i.t. injection of PGE₂ to EP₃^−/− mice produced a monophasic hyperalgesic action at a narrower range of dosages (5 ng – 5 μg kg⁻¹) (Figure 2b). PGE₂ at a dose of 500 ng kg⁻¹ reduced the latency periods to 10.1±1.1 s and 11.5±0.6 s in EP₃^+/+ and EP₃^−/− mice, respectively. To clarify the difference in dose dependency of PGE₂ for hyperalgesia between EP₃^+/+ and EP₃^−/− mice, we examined the induction of hyperalgesia by the selective EP₃ agonist ONO-AE-248. ONO-AE-248 reduced the latency period to 10.6±0.6 s at 500 pg kg⁻¹, comparable to the potency of PGE₂, and 13.4±1.1 s at 500 ng kg⁻¹ in EP₃^+/+ mice (Figure 2c). On the other hand, ONO-AE-248 had no effect at 500 pg kg⁻¹ and 500 ng kg⁻¹ in EP₃^−/− mice (Figure 2d). Loss of the functional EP₃ receptor in EP₃^−/− mice was verified by the inability of the EP₃ agonist to induce hyperalgesia. Consistent with our pharmacological studies (Minami et al., 1994a), these results demonstrate that the PGE₂-induced hyperalgesia is mediated by the EP₃ subtype at lower doses of PGE₂.

Figure 2.

Hyperalgesia evoked by PGE₂ in EP₃^+/+ (a,c) and EP₃^−/− (b,d) mice. An indicated dose of PGE₂ or ONO-AE-248 (an EP₃ agonist), or saline was injected into the subarachnoid space of conscious mice. Hyperalgesia was assessed 30 min after i.t. injection of agent. Each column represents the mean±s.e.mean (n=8 – 10). ^*P<0.05, ^**P<0.01, compared with the saline-injected control group.

Similar to EP₃^+/+ mice, PGE₂ reduced the latency period in EP₁^+/+ mice over a wide range of dosages (500 pg kg⁻¹ – 5 μg kg⁻¹) with a maximum effect at 50 ng kg⁻¹ (8.5±0.8 s, mean±s.e.mean, n=10), as compared with that (15.1±1.2 s) of the saline-injected control (Figure 3a). Unexpectedly, untreated EP₁^−/− mice and saline-injected EP₁^−/− mice were in a hyperalgesic state (8.7±0.4 s and 6.5±0.7 s, respectively). The latency period was dose-dependently prolonged by i.t. injection of PGE₂ in EP₁^−/− mice (Figure 3b).

Figure 3.

Hyperalgesia evoked by PGE₂ in EP₁^+/+ (a) and EP₁^−/− (b) mice. An indicated dose of PGE₂ or saline was injected into the subarachnoid space of conscious mice. Hyperalgesia was assessed 30 min after i.t. injection of PGE₂. Each column represents the mean±s.e.mean (n=8 – 10). ^*P<0.05, ^**P<0.01, compared with the saline-injected control group.

Formalin test in EP₁^−/− and EP₃^−/− mice

The formalin test is an established model of pain associated with injury and inflammation and is a sensitive method to evaluate various classes of analgesic drugs (Hunskaar & Hole, 1987). Formalin injected subcutaneously into the paw produces two phases of responding, an early phase (0 – 5 min), attributed to a direct effect of the formalin on nociceptors and a late phase (15 – 30 min), related to the subsequent development of inflammation and central sensitization. The formalin injection induced biphasic licking and biting responses of the injected paw, with an early phase determined for the first 5 min (41.7±8.1 s and 44.1±4.5 s) and a late phase determined for 15 – 30 min (171.7±29.8 s and 196.2±28.1 s) in EP₁^+/+ and EP₃^+/+ mice, respectively (Figure 4). The early-phase behaviours (30.2±8.0 s and 46.0±5.1 s) in EP₁^−/− and EP₃^−/− mice were not different from those in wild-type mice in the formalin test. In the late phase, the amount of time spent in licking the injected paw was partly reduced, but not significantly decreased in EP₁^−/− mice (105.5±20.8 s) and in EP₃^−/− mice (127.1±12.5 s). In contrast to the PGE₂-induced allodynia and hyperalgesia, these results demonstrate that licking and biting responses induced by formalin injection are not ascribed to a single EP₁ or EP₃ subtype.

Figure 4.

Formalin test in EP₁^+/+ and EP₁^−/− (a,b), and EP₃^+/+ and EP₃^−/− (c,d) mice. Assessment of the formalin test was made as described under ‘Methods'. Each point represents the mean±s.e.mean (n=9 – 10). ^*P<0.05, compared with the wild-type mice.

Discussion

We previously demonstrated that i.t. PGE₂ induced both allodynia and hyperalgesia. On the basis of rankings of specificities of EP agonists for EP receptor subtypes, we suggested that the PGE₂-induced allodynia is mediated by the EP₁ receptor in the mouse spinal cord and that the PGE₂-induced hyperalgesia is mediated by the EP₃ receptor at lower doses and by the EP₂ receptor at higher doses (Minami et al., 1994a). Consistent with our earlier observations, while PGE₂ induced allodynia in EP₃^−/− mice, it could not induce allodynia in EP₁^−/− mice (Figure 1). This confirmed that the EP₁ receptor is involved in the PGE₂-induced allodynia. PGE₂ stimulates the release of glutamate and NO from the spinal cord in a Ca²⁺-dependent manner (Nishihara et al., 1995; Sakai et al., 1998). The PGE₂-induced allodynia was blocked by NMDA and non-NMDA receptor antagonists and inhibitors for NO synthase and guanylate cyclase (Minami et al., 1994b; 1995a). We recently showed that the PGE₂-induced allodynia was not observed in mice deficient in ε1 subunit of NMDA receptors (Minami et al., 1999). Therefore, we postulated that i.t. PGE₂ induces allodynia through activation of the NMDA receptor and following NO production via the EP₁ receptor. Here we demonstrated that activations of the glutamate-NO system, i.e., NMDA, L-arginine, and SNP, induced allodynia in EP₁^−/− and EP₃^−/− mice (Table 1) to the same extent as in wild-type mice, suggesting that a site(s) of PGE₂ involving the induction of allodynia lies upstream, rather than downstream, of the glutamate-NO system.

Intrathecal injection of PGE₂ produced a monophasic hyperalgesia in EP₃^−/− mice at higher doses than in EP₃^+/+ mice (Figure 2a,b). The selective EP₃ agonist ONO-AE-248 produced a hyperalgesic effect at 500 pg and 500 ng kg⁻¹ in EP₃^+/+ mice, but not in EP₃^−/− mice (Figure 2c,d). ONO-AE-248 at 500 pg kg⁻¹ seemed to be more effective in producing hyperalgesia in EP₃^+/+ mice than at 500 ng kg⁻¹. These results confirm that the PGE₂-induced hyperalgesia is mediated by the EP₃ subtype at lower doses. Whether hyperalgesia induced by high doses of PGE₂ is mediated by EP₂ receptors will be clarified by EP₂^−/− mice. In this context, Kumazawa et al. (1996) previously showed by use of EP agonists that low concentrations of PGE₂ augmented the response of polymodal receptors by bradykinin through EP₃ receptors and that high concentrations of PGE₂ augmented heat responses through EP₂ receptors in the periphery. Unexpectedly, EP₁^−/− mice showed a hyperalgesic action in the hot plate test and the hyperalgesia was alleviated by PGE₂ in a dose-dependent manner (Figure 3). Because i.t. PGE₂ produced hyperalgesia at as low as 500 pg kg⁻¹ (Figures 2a and 3a), the observed hyperalgesic response in EP₁^−/− mice may be due to unopposed activation of EP₃ receptor at the basal level of PGE₂ in the spinal cord. While the exact mechanism of hyperalgesia in EP₁^−/− mice remains to be clarified, the reversal of a hyperalgesic state in EP₁^−/− mice by PGE₂ may be mediated by EP₄. The present study suggests that endogenous PGE₂ may play an inhibitory role in the appearance of hyperalgesia via the EP₁ receptor under physiological conditions and disruption of EP₁ receptors apparently unmasks hyperalgesic EP receptor subtypes.

Formalin injection produces a characteristic biphasic flinching, shaking or licking behaviour of the injected paw. Spinal involvement of PGE₂ in the formalin test has been proposed based on several observations: (i) i.t. injection of cyclo-oxygenase inhibitors produces a dose-dependent antinociceptive effect on the late phase, but limited effect on the early phase of the formalin test in the rats (Malmberg & Yaksh, 1992b); (ii) subcutaneously injected formalin increased PGE₂ release from the spinal cord (Malmberg & Yaksh, 1995); and (iii) i.t. injection of the EP₁ antagonists SC-51089 and SC-51234A produced significant suppression of the second phase, without any effect on the early phase, in the formalin test (Malmberg et al., 1994). Consistent with previous reports, there is no difference in the early response between wild-type mice and knockout mice. Although there was a considerable decrease in flinching behaviour in EP₁^−/− and EP₃^−/− mice 30 min after formalin injection, this difference was not significant when compared with the whole (15 – 30 min) late-phase behaviour (Figure 4). The difference in the extent of contribution of EP₁ in the formalin test may be due to the difference in testing paradigm and animal species. Because the suppressive effect of EP₁ antagonists on formalin-induced pain was weak as compared with those of cyclooxygenase inhibitors (Malmberg & Yaksh, 1992b; Malmberg et al., 1994), the formalin-induced behaviours are likely to be a complex interplay among EP receptor subtypes and other modifying factors at both peripheral and spinal levels.

The use of mice deficient in PG receptors has provided direct assignment to inflammatory responses including pain. In mice lacking the PGI receptor (IP ^−/−), the acetic acid-induced writhing response or carrageenan-induced paw oedema was reduced to the same level as that observed when indomethacin was administered to wild-type mice (Murata et al., 1997). Because the IP ^−/− mice did not show any alteration in thermal nociceptive responses examined by hot plate and tail flick tests, they suggested that PGI₂ is a mediator of inflammatory swelling and pain in the periphery, but not involved in nociceptive transmission at the spinal level. In the present study, we extended our previous studies by use of EP₁^−/− and EP₃^−/− mice and present evidence that PGE₂ induces allodynia through the EP₁ receptor and hyperalgesia through the EP₃ receptor at lower doses. Continued investigations into the specific roles of the receptors that mediate the actions of PGs with other PG receptor knockout mice will further promote our understanding of PG actions in nociception and may potentially lead to the development of more specific and less toxic therapies for pain.

Abbreviations

i.t.

intrathecal

NMDA

N-methyl-D-aspartate

NO

nitric oxide

PGE₂

prostaglandin E₂

SNP

sodium nitroprusside