Role of Prostaglandin Receptor EP¹in the Spinal Dorsal Horn in Carrageenan-induced Inflammatory Pain

The protocol for this study was approved by the Sapporo Medical University Animal Care and Use Committee. The animals used were male Sprague-Dawley rats (weighing 150–250 g; Japan SLC, Hamamatsu, Japan) that were housed individually in a temperature-controlled (21 ± 1°C) room with a 12-h light-dark cycle and given free access to food and water.

The current study consisted of three experiments: behavioral, microdialysis, and [Ca²⁺]ⁱmeasurement experiments.

In this study, we used a rat model of carrageenan-induced inflammatory pain. Unilateral peripheral inflammation was induced by an intraplantar injection of 2 mg carrageenan in a volume of 100 μl into the left hind paw. Experiments were performed with normal rats and rats that had been injected with carrageenan. Carrageenan-injected rats were used in the experiments 3 h (early phase) or 15 h (late phase) after the injection.

During general anesthesia (3% isoflurane in oxygen), a polyethylene intrathecal catheter (PE-10; Becton Dickinson, Sparks, MD) was inserted into the lumbar subarachnoid space at the L4–L5 intervertebrae with the tip of the catheter located near the lumbar enlargement of the spinal cord, using a method described previously. 17Only animals that showed normal behavior and motor function were used in the experiments.

To evaluate the mechanical hyperalgesia, withdrawal threshold to mechanical stimulation was determined using calibrated von Frey filaments (0.0045–75.8580 g in bending force; Stoelting, Wood Dale, IL) that were applied from underneath the cage through openings in the wire mesh floor to the plantar surface of the hind paw on the carrageenan-injected side and to the same area on the contralateral side. Each filament was applied once starting with 0.0045 g and continuing until a withdrawal response occurred. A withdrawal response was considered to be complete lifting of the hind paw off the surface of the cage or flinching. The test was repeated three times at each time point. The minimum force that produced a response to at least one of three applications was considered as the withdrawal threshold.

The effects of the EP¹antagonist ONO-8711 were examined in normal and carrageenan-treated rats (n = 48). Before the carrageenan injection, withdrawal thresholds were measured as control values on each side of hind paw. The carrageenan-treated rats were used in the experiment in the early (3 h) or late (15 h) phase after carrageenan injection. Baseline values of the threshold were determined in the early and late phases. ONO-8711 (1, 10, or 100 μg) or saline in a volume of 10 μl was administered intrathecally. The thresholds were assessed for up to 120 min on both sides of the hind paw.

The magnitudes of inflammatory response to carrageenan were evaluated by measuring the thickness of the dorsal-ventral paw using a vernier micrometer.

A spinal cord dialysis probe was made according to our modification of the method described by Skilling et al.  18The probe was constructed from a 1-cm-long dialysis fiber (ID of 200 μm, OD of 220 μm, and 50-kd molecular weight cutoff; DM-22; Eicom, Kyoto, Japan) that had been coated with an epoxy glue (Devcon, Danvers, MA) along the whole length except for a 2-mm region in the middle. Each end of the fiber was attached to polyethylene catheters (PE-10), and each end of the polyethylene catheter was then attached to a Teflon tube (JT-10, Eicom).

Rats were anesthetized with pentobarbital (50 mg/kg administered intraperitoneally), and an incision was made along the dorsal midline from T2 to L2. The lateral surfaces of vertebra L1 were exposed, and bilateral holes were carefully made through the bone, exposing the spinal cord laterally at the level of the dorsal horn. A dialysis tube was placed through the holes, passing transversely through the dorsal spinal cord. The two distal ends of the probe were tunneled subcutaneously and externalized through the skin in the neck region.

The experiments were performed 24 h after implantation of the dialysis probe. Only animals that showed normal behavior were used (n = 6). The animals were allowed to move freely in a plastic cage during the experiments. The dialysis probe was perfused with artificial cerebrospinal fluid (140 mm NaCl, 4.0 mm KCl, 1.26 mm CaCl², 1.15 mm MgCl², 2.0 mm Na²HPO⁴, 0.5 mm NaH²PO⁴, and pH 7.4) at a constant flow rate of 4 μl/min. Collected samples were frozen at −80°C until used for analysis. The samples were collected as 30-min fractions. After obtaining the three consecutive samples for determination of basal level, carrageenan or saline was injected. Dialysate samples were collected for 15 h after carrageenan or saline injection, and the samples collected at the time points of 3, 8, and 15 h were used.

After each experiment, methylene blue dye was perfused through the dialysis probe to verify the position of the dialysis fiber. The data used for analysis were from rats in which dye remained in the dorsal half area of the dorsal horn.

The concentration of PGE²in dialysate was measured using a commercially available PGE²enzyme immunoassay kit (Cayman Chemicals, Ann Arbor, MI). The limit of quantification was 20 pg/ml.

To measure [Ca²⁺]ⁱin the dorsal horn of the spinal cord, transverse spinal cord slices were prepared according to our modification of the method described previously. 19,20Briefly, during urethane anesthesia (1.5 g/kg administered intraperitoneally), thoracolumbar laminectomy was performed, and the lumbar spinal cord was excised. The spinal cord was mounted on a Vibratome, and two or three 400-μm-thick transverse slices at the level of the L4 root entry were obtained. The slices were mounted in a chamber and were perfused (25 ml/min) continuously with preoxygenated Krebs solution. The slices were then incubated with Krebs solution (37°C) containing 10 μm fura-2/AM (Dojindo, Kumamoto, Japan) for 45 min. Thereafter, slices were perfused with Krebs solution for 5 min and then placed in a recording chamber mounted on an inverted fluorescence microscope (TE300; Nikon, Tokyo, Japan). The recording chamber was perfused (12–15 ml/min) with oxygenated Krebs solution (37°C) for 30 min before starting the study. The composition of Krebs solution was as follows: 117 mm NaCl, 3.6 mm KCl, 2.5 mm CaCl², 1.2 mm MgCl², 1.2 mm NaH²PO⁴, 25 mm NaHCO³, and 11 mm glucose.

Images were viewed through an objective lens (Plan fluor 4×/0.13; Nikon). Light from a Xenon lamp was filtered through either of two different band-pass filters (340 or 380 nm) in the excitation path and conducted to the specimen on the microscope stage through a diachronic mirror. The excitation wavelength was constantly switched between 340 and 380 nm. The fluorescence emitted from the slice was passed through a band-pass filter (510 nm). Video images, passed through a TV lens (C-0.45×; Nikon), were obtained using CCD video camera (C6790–81; Hamamatsu Photonics, Hamamatsu, Japan). The whole area of the spinal cord slice image captured by CCD camera was created by a computerized imaging system (ARGUS-50 HiSCA; Hamamatsu Photonics). Fura-2 fluorescence ratio images (340/380 nm) were calculated in real time (fig. 1A). [Ca²⁺]ⁱin the dorsal horn of the spinal cord slice was calculated according to a previously described method. 20,21Briefly, calibration at high and low [Ca²⁺]ⁱwas made after treating the slices with 0.01% Triton X-100 in Krebs solution followed by exchange with calcium-free medium containing 1 mm EGTA and 2 mm MnCl²in Krebs solution. This produced calcium signals equivalent to the bath concentration (saturated calcium) and to zero calcium. Calculation of [Ca²⁺]ⁱwas performed using the following equation 20,22:

where KD is the dissociation constant of fura-2 (224 nm), R is the ratio of fura-2 fluorescence intensity at 510 nm elicited by excitation at 340 and 380 nm, Rmin is the value in the presence of saturating [Ca²⁺]ⁱ, and F⁰/F¹is the ratio of the 380-nm excitation intensity at zero [Ca²⁺]ⁱto that at saturated [Ca²⁺]ⁱlevels. Thus, [Ca²⁺]ⁱwas imaged and calculated in 400 μm × 200 μm areas of laminae II, III-IV, and V in the spinal dorsal horn (fig. 1B).

Intracellular calcium ion concentration in the spinal dorsal horn in normal rats (n = 6) and in carrageenan-treated rats with inflammation in the early phase (3 h; n = 6) and late phase (15 h; n = 6) were measured, and the concentrations on both sides of the dorsal horn were compared. In some rats (n = 6), 100 μg of ONO-8711 was administered via  an intrathecal catheter that had been implanted. Thirty minutes after the administration of ONO-8711, the spinal cord was removed, and [Ca²⁺]ⁱin the spinal dorsal horn was also measured.

To assess the effects of PGE²on [Ca²⁺]ⁱin the normal and carrageenan-treated rats and the effects of ONO-8711, PGE²alone (n = 18) or a combination of PGE²and ONO-8711 (n = 18) were applied by changing the perfusion solution. After determination of control [Ca²⁺]ⁱon both sides in dorsal horn, 10 μm PGE²was perfused for 7 min, and [Ca²⁺]ⁱwas measured. Using slices from other rats, changes in [Ca²⁺]ⁱwere assessed following perfusion with 10 μm ONO-8711 for 2 min and then perfusion with a combination of PGE²(10 μm) and ONO-8711 (10 μm) for 7 min. After exposure to the drugs, each slice was rinsed with Krebs solution to verify whether [Ca²⁺]ⁱwould return to the initial values in each study.

Prostaglandin E²and ONO-8711 were supplied by Ono Pharmaceutical (Osaka, Japan). ONO-8711 was dissolved in 1 N of NaOH and then diluted to the desired concentration in saline or Krebs solution immediately before use. All other chemicals and solvents were purchased from Sigma Chemical (St. Louis, MO).

The withdrawal threshold was expressed as a median for ordinal data and was analyzed by the Kruskal-Wallis test followed by Dunn test. Changes in PGE²concentrations were presented as means ± SD of the percentages of basal levels. Changes in PGE²concentrations and changes in paw thickness were analyzed by a one- and two-way analysis of variance, respectively, followed by Bonferroni correction. Other data in the current study are shown as means ± SD, and statistical comparisons of values were performed by paired or unpaired two-tailed Student t  tests. P < 0.05 was considered statistically significant.

Intraplantar injection of carrageenan resulted in formations of edema and erythema. Before the carrageenan injection, the paw thickness was 5.0 ± 0.4 mm. Three hours after the injection (early phase), the thickness on the ipsilateral side significantly (P < 0.05) increased (11.2 ± 0.6 mm), and this increase lasted for at least 15 h (late phase, 10.6 ± 0.4 mm). The thickness on the contralateral side did not show any significant changes in either the early or late phase.

Withdrawal thresholds to mechanical stimulation significantly (P < 0.05) decreased in both the early and late phases of inflammation on the ipsilateral side to the injection (figs. 2A and B). In the early phase of car-rageenan-induced inflammation, intrathecal ONO-8711 did not show any effects on hyperalgesia (fig. 2A). However, in the late phase, 10 and 100 μg ONO-8711, but not 1 μg of ONO-8711 or saline, significantly (P < 0.05) increased the withdrawal threshold (fig. 2B). The thresholds on the contralateral hind paw did not show any changes after ONO-8711 administration at all doses administered in both the early and late phases of carrageenan-induced inflammation (figs. 2C and D, respectively). In the normal rats, intrathecal ONO-8711 at the dose of 100 μg did not show any effects on the paw withdrawal threshold to mechanical stimulation (data not shown).

Our preliminary study showed that a dialysis equilibrium was obtained within 120 min after the start of artificial cerebrospinal fluid perfusion at a constant flow rate of 4 μl/min and that basal values were stable for at least 15 h (data not shown). The mean basal value of PGE²concentration in the dorsal horn was 105.78 ± 20.78 pg/ml.

Three hours after carrageenan injection to the hind paw (in the early phase), PGE²concentration in the dialysates increased, but not significantly (fig. 3). A significant (P < 0.05) increase was observed 15 h after the carrageenan (in the late phase) compared with the baseline concentration of PGE²(fig. 3). The concentrations of PGE²in the dialysates did not show significant changes after intraplantar saline injection (data not shown).

The position of the probe after each experiment was verified, and the level of the spinal cord dialysized was L4 in all rats.

Figure 4shows the [Ca²⁺]ⁱin the spinal dorsal horn of normal rats and rats treated with carrageenan. In the normal rats, [Ca²⁺]ⁱon the ipsilateral and contralateral sides of the dorsal horn were comparable (fig. 4A). There was also no difference of [Ca²⁺]ⁱbetween on the ipsilateral and contralateral sides to the carrageenan injection in the rats with early-phase inflammation (fig. 4B). However, in the rats with late-phase inflammation, significantly (P < 0.05) higher [Ca²⁺]ⁱin laminae II and V were observed on the ipsilateral side compared with those on the contralateral side (fig. 4C).

Figure 5shows the effects of pretreatment with intrathecal ONO-8711 on changes in [Ca²⁺]ⁱin the car-rageenan-treated rats with late-phase inflammation. In contrast to the significant (P < 0.05) increases in [Ca²⁺]ⁱon the ipsilateral side of the dorsal horn to the carrageenan in the rats that had not received ONO-8711 (fig. 5A), no significant increases in [Ca²⁺]ⁱon the ipsilateral side were observed in the rats that had been pretreated with intrathecal ONO-8711 (fig. 5B).

The changes in [Ca²⁺]ⁱin the dorsal horn following PGE²perfusion are shown in figure 6. The [Ca²⁺]ⁱin laminae II, III-IV, and V were increased by perfusion with 10 μm PGE²in normal and carrageenan-treated rats. In the normal rats and in the rats with early-phase inflammation, there were no differences between the increase in [Ca²⁺]ⁱon the two sides of the dorsal horn (figs. 6A and B). However, the increases in [Ca²⁺]ⁱwere significantly (P < 0.05) greater in laminae II and V on the ipsilateral side than on the contralateral side in the rats with late-phase inflammation (fig. 6C).

While the magnitudes of PGE²-induced changes in [Ca²⁺]ⁱwere comparable on ipsilateral and contralateral sides in normal rats and in rats with early-phase inflammation, the increase in [Ca²⁺]ⁱin rats with late-phase inflammation were significantly (P < 0.05) greater on the ipsilateral side than on the contralateral side in laminae II and V of the spinal dorsal horn. The increase in [Ca²⁺]ⁱon the ipsilateral side in the rats with late-phase inflammation was significantly (P < 0.05) greater than that on the referred side in lamina II in both the normal rats and rats with early-phase inflammation. Perfusion with ONO-8711 significantly (P < 0.05) suppressed the PGE²-induced elevations in [Ca²⁺]ⁱin normal rats and in rats with early-phase inflammation. In the rats with late-phase inflammation, ONO-8711 significantly (P < 0.05) suppressed the increase in [Ca²⁺]ⁱon both sides of the dorsal horn; the greater elevation in the [Ca²⁺]ⁱin response to PGE²on the ipsilateral side in rats with late-phase inflammation were also suppressed by perfusion with ONO-8711.

We previously reported that peripherally administered ONO-8711 inhibited incision-induced mechanical hyperalgesia, 2suggesting that EP¹receptor-mediated peripheral sensitization of sensory nerve fibers contributes to the generation of mechanical hyperalgesia produced by tissue damage. The findings in the current study suggest that ONO-8711 suppressed inflammation-induced mechanical hyperalgesia at the level of the spinal cord. Taken together, these findings indicate that peripherally and centrally released PGE²following tissue damage and inflammation activates the EP¹receptors at peripheral and spinal sites, respectively.

Recent behavioral studies have demonstrated that pretreatment with intrathecal ibuprofen and cyclooxygenase-2 inhibitors prevented carrageenan-induced thermal hyperalgesia. 23,24However, at 3 h after carrageenan injection, ibuprofen and cyclooxygenase-2 inhibitor were ineffective. 24The results of these studies suggest that spinal prostaglandins are important in the development of thermal hyperalgesia when inflammation begins to develop. In the current study, intrathecal EP¹antagonist had no effect on the mechanical hyperalgesia at 3 h after the carrageenan injection but was effective at 15 h after the carrageenan. Furthermore, a remarkable release of PGE²was observed 15 h after the carrageenan injection. There have been other studies on spinal PGE²release after inflammation. An injection of complete Freund's adjuvant caused an elevation of PGE²content in lumbar spinal cord homogenates with an early peak at 8 h after injection. 25Guhring et al.  26reported an elevation of PGE²release induced by zymosan injection into a mouse hind paw, with a peak at 8 h. In addition, in an inflammation model by intraarticular injection of kaolin and carrageenan, a pronounced enhancement of release of intraspinal PGE²was observed 430–530 min after the injection. 27Thus, it seems that there is a discrepancy in the time courses of previously reported effects of cyclooxygenase inhibitors in behavioral assays and our results, including effects of an EP¹antagonist and PGE²release in the spinal cord. There are several possible explanations for this discrepancy. First, the different time courses of mechanical and thermal hyperalgesia may reflect the different mechanisms that are thought to underlie these two phenomena. Hedo et al.  14reported that both thermal hyperalgesia and mechanical hyperalgesia were induced by carrageenan but that the time courses of the effects were different. The thermal hyperalgesia was maximally developed at 45 min after injection but had returned to baseline levels within 20 h after the injection, whereas significant mechanical hyperalgesia was observed 3 h after the treatment and was still present 6 days after the treatment. Second, the inhibitory effects of intrathecal cyclooxygenase inhibitors on thermal hyperalgesia seen within 3 h after the carrageenan injection might have been mediated by the suppression of production of prostanoids other than PGE², such as PGD², that also cause hyperalgesia when administered spinally.

In the late phase of inflammation, increases in [Ca²⁺]ⁱon the ipsilateral side to the inflammation were observed in spinal slices from which the dorsal root had been removed. This finding suggests that the increase in [Ca²⁺]ⁱpersists for a long time relatively independent of primary afferent inputs once [Ca²⁺]ⁱhas been elevated. Repetitive stimulation of spinal neurons may lead to long-term or sustained increases in [Ca²⁺]ⁱ. Therefore, these increases in [Ca²⁺]ⁱmay have been involved in plastic changes in the spinal neurons of the rats with peripheral inflammation. These changes were also observed in a peripheral neuropathic pain model of rats. 20In the current study, an increase in [Ca²⁺]ⁱwas not observed in the early phase of inflammation, suggesting that central changes induced by carrageenan develop increasingly during the period of inflammation and are different and probably more significant in the late phase than in the early phase, as was also suggested by the results of our behavioral study.

We also showed that the magnitudes of changes in [Ca²⁺]ⁱinduced by PGE²perfusion were greater on the ipsilateral side to inflammation in slices obtained from the rats with late-phase inflammation. The increase in [Ca²⁺]ⁱinduced by PGE²perfusion was inhibited by coperfusion with an EP¹antagonist. Thus, the sensitivity of [Ca²⁺]ⁱin the spinal dorsal horn neurons to PGE²increased in the late phase of carrageenan-induced inflammation, and this increase in the responsiveness is likely mediated by the activation of spinal EP¹receptors.

Autoradiographic studies have indicated that the highest density of spinal PGE²binding sites is in substantia gelatinosa (lamina II). Matsumura et al.  reported that dorsal rhizotomy reduced, but did not eliminate, PGE²binding in lamina II, 28suggesting that receptor sites are located on both presynaptic terminals of nociceptive fibers and postsynaptic dorsal horn neurons. Therefore, PGE²may act presynaptically to facilitate neurotransmitter release 29–31and postsynaptically to directly excite dorsal horn neurons. 32Although we did not examine the PGE²binding sites, especially EP¹receptor sites, in the current study, the superficial and deep layers would contain many PGE²-responsive elements of the spinal cord. The current study showed that [Ca²⁺]ⁱwas increased not only in lamina II but also in lamina V of the dorsal horn by PGE²perfusion. This might be related to the finding that PGE²-induced activation of EP¹receptors in both superficial and deep layers increases [Ca²⁺]ⁱ.

The results of the current study suggest that the late phase of carrageenan-induced inflammation increases the production and release of PGE²and enhances the function of spinal EP¹receptors involved in nociceptive transmission. Thus, it is likely that the activation of EP¹receptors is important for development and maintenance of central sensitization caused by peripheral inflammation. Hedo and colleagues 14reported that, in an isolated spinal cord preparation, nociceptive reflexes were augmented more in the late phase (20 h) than in the early phase (3–6 h) of carrageenan-induced inflammation. A study using in vivo  electrophysiologic assays showed that the effects of a spinally administered N -methyl-d-aspartate (NMDA) receptor agonist and antagonist on spinal nociceptive responses were greater in the late phase than in the early phase of carrageenan-induced inflammation. 15The results of these studies and those of the current study therefore suggest that the hyperalgesia observed during the early phase may be due to the activation of the mechanisms other than spinal EP¹activation, such as glutamate or NK¹receptors, and peripheral sensitization, and that some of the changes in release of PGE², activation of EP¹receptors, and subsequent elevation of [Ca²⁺]ⁱin the spinal dorsal horn in the late phase may contribute to maintenance of hyperalgesia, independent of ongoing peripheral input.

The current study demonstrated that intrathecal treatment with ONO-8711 suppressed not only the increases in [Ca²⁺]ⁱ, but also behavioral hyperalgesia in the rats with late-phase inflammation. These findings suggest that [Ca²⁺]ⁱis maintained at a high level by activation of spinal EP¹receptors during maintenance of central sensitization. The phase-dependent changes in the contribution of EP¹receptor activation to central plasticity may be explained by one or more of the following possibilities: generation of more EP¹receptors, increased affinity for EP¹, increased response to PGE²binding, or an up-regulation of factors triggered by EP¹receptor activation. Regarding generation of EP¹receptors, Donaldson et al.  33found, by using semiquantitative reverse transcription polymerase chain reaction, that EP¹receptor mRNA concentrations in the DRG and spinal cord were unaffected by inflammation following the injection of Freund's complete adjuvant. Further study on the expressions and locations of EP¹receptors in DRG neurons and the spinal dorsal horn is needed to clarify the regulation of EP¹receptors (up-regulated or down-regulated) during peripheral inflammation.

Following peripheral inflammation, excitatory neurotransmitters such as glutamate are released from primary afferents. 34,35The subsequent glutamate-mediated NMDA receptor activation leads to a Ca²⁺influx into the neurons and initiates an enzymatic cascade that finally evokes the release of nitric oxide and PGE². 36–38Nitric oxide and PGE², in turn, have been hypothesized to further increase glutamate release, 30,31,36leading to an ongoing activity in primary afferents and increased sensitivity of dorsal horn neurons. PGE²also directly excites dorsal horn neurons. 32These effects of PGE²contribute, finally, to the development of central sensitization. Thus, it is thought that there exists a positive feedback cascade mediated through PGE²in which increased primary afferent activity with peripheral inflammation sensitizes spinal neurons and increases synaptic glutamate release from primary afferent terminals. In this feedback cascade, glutamate acts on the NMDA receptors on the postsynaptic neurons to evoke an increase in [Ca²⁺]ⁱ. The increase in [Ca²⁺]ⁱactivates phospholipase A², leading to increases in production of arachidonic acid and prostanoids. The released PGE²may interact with EP¹receptors on the presynaptic terminals and postsynaptic neurons, resulting in [Ca²⁺]ⁱelevation, which facilitates the release of glutamate from primary afferent terminals and depolarization of dorsal horn neurons. The fact that an EP¹antagonist suppressed the increase in [Ca²⁺]ⁱin the spinal dorsal horn suggests that the increase in [Ca²⁺]ⁱresulting from the activation of NMDA receptors and EP¹receptors in inflammation can be suppressed by blocking this cascade.

In summary, plastic changes in the spinal dorsal horn mediated by the activation of EP¹receptors were observed in the late phase of carrageenan-induced inflammation. Mechanisms underlying inflammation-induced hyperalgesia may be mediated by neuronal plasticity in the central nervous system resulting from long-term or sustained increases in PGE²concentrations, activation of EP¹receptors, and increases in [Ca²⁺]ⁱin the spinal dorsal horn. It is likely that an EP¹antagonist can block the positive feedback cascade mediated by PGE², resulting in inhibition of carrageenan-induced hyperlagesia. Thus, a selective EP¹receptor antagonist is a potential analgesic for persistent inflammatory pain.

The authors thank Megumu Yoshimura, M.D, Ph.D. (Department of Physiology, Kyushu University School of Medicine, Fukuoka, Japan), for his instruction on how to prepare spinal cord slices, and Mikito Kawamata, M.D., and Tomoyuki Kawamata, M.D. (Department of Anesthesiology, Sapporo Medical University School of Medicine, Sapporo, Japan), for providing helpful suggestions on [Ca²⁺]ⁱimaging and microdialysis studies, respectively.