Antinociceptive Actions of Dexmedetomidine and the Kappa-Opioid Agonist U-50,488H against Noxious Thermal, Mechanical and Inflammatory Stimuli1

Transcription

1 /94/ $3.OO/O Txz Jouai. OF PiwuecowGv m ExpaRnaimi. THnpauTics Copyright 1994 by The American Society for Pharmacology and Experimental Therapeutics.JPET 271: , 1994 Vol. 271, No. 3 Printed in U.S.A. Antinociceptive Actions of Dexmedetomidine and the KappOpioid Agonist U-5,488H against Noxious Thermal, Mechanical and Inflammatory Stimuli1 JUHANA J. IDANPAAN-HEIKKILA, EIJA A. KALSO and TIMO SEPPALA Institute of Biochemist,y, Department of Pharmacology and Toxicology, UnWersity of Helsinki (J.J.!.-H.); Department ofanaesthesia, H&sinkl University Central Hospital (EAK.); and Laborato,y of Pharmacology and Toxicology, National Public Health Institute (T.S., J.J.l.-H.), Helsinki, Rnland Accepted for publication August 26, 1994 ABSTRACT The antinociceptive efficacies of both intrathecally (i.t.) and systemically administered dexmedetomidine (a selective a!- ph2 adrenoceptor agonist) and U-5,488H [trans(±)-3,4- dichloro-n-methyl-n-[2-(1 -pyrrolidlnyl)-cyclohexyl]-benzeneacetamide] (a kappopioid receptor agonist) were studied during peripheral inflammation induced by carrageenan. The antinociceptive tests were the hot plate (HP), the tall flick (IF) and the paw pressure tests (PP). The motor incoordination, If any, produced by both i.t. and s.c. dexmedetomidine were evaluated with the rotarod. The interaction between dexmedetomidine and U-5,488H and between atipamezole (a selective alph2 adrenoceptor antagonist) and U-5,488H were also assessed. The carrageenan injection induced not only peripheral hyperalgesia but also central sensitization, as assessed by decreased PP and TF latencies, respectively. The i.t. dexmedetomidine (.15,.45, 1.35, 4.5) g) resulted in dosedependent increases in the PP thresholds and TF latencies in both the control rats and the rats with unilateral inflammation, without causing changes in motor coordination, whereas on Alph2 adrenoceptor agonists significantly elevate the nociceptive threshold when administered i.t. to both humans and rodents. Their action is thought to be mediated by atp!w-2 adrenoceptors located in the dorsal horn of the spinal chord (Yaksh, 1985; Sullivan et at., 1987, 1992). Spinal application of the alph2 agonist clonidine has produced antinociception in tests with heat as the noxious stimulus (Ossipov et at., 1988; Takano and Yaksh, 1992; Hylden et at., 1991). It has also had analgesic effects in tests that combine peripheral unilateral inflammation with mechanical pressure (Fielding et at., 1978) or thermal stimuli (Hylden et al., 1991). Clonidine is supposed, however, to have mixed atph1 Received for publication February 25, This study was supported by the Pharmacal Research Foundation, Finland (J.J.I.-H.), the Academy of Finland (E.A.K) and the Paulo Foundation, Finland (EAK). S.C. administration of dexmedetomidine (3-i gikg), antino- Ciception was produced in PP only at doses (3 tg/kg) that already interfered with rotarod performance. U-5,488H was ineffective i.t. (5-2 p,g) but, on s.c. administration ( mg/kg), dose-dependent increases were found in the PP thresholds and IF latencies of the rats with unilateral inflammation. Atipamezole, in a dose (3 mg/kg) that has been shown to block the antinociceptive effects of dexmedetomidine, did not modify the antinociceptive effects of U-5,488H. Coadministration of dexmedetomidine (.45 g i.t.) with U-5,488H (2.5 mg/kg s.c.) resulted in additive effects in the PP test on both paws and in the TF test in the inflamed-paw rats. Dexmedetomidine and U-5,488H produced antinociception in the rats with inflammation at doses devoid of any antinociceptive effect in Control rats. This suggests an enhanced antinociceptive effect for both dexmedetomidine and U-5,488H during peripheral inflammation and the consequent central sensitization. Thermal tests seem to be more sensitive to detect this effect. and atph2 effects (Hayes et at., 1986). Dexmedetomidine, a more selective atph2 adrenoceptor agonist, with an atph 2/alphi selectivity ratio of 162 compared with 22 for clonidine (Virtanen et at., 1988; Virtanen, 1989), has also been effective in assays with noxious heat (Kalso et at., 1991; Takano and Yaksh, 1992). Its efficacy in tests that use noxious mechanical and inflammatory stimuli has not, however, been studied to our knowledge. Atipamezole, a selective and specific alph2 adrenoceptor antagonist (Scheinin et at. 1988; Virtanen, 1989), has eliminated the antinociceptive effects ofboth i.t. and systemically administered dexmedetomidine (Kalso et at., 1991; Sullivan et at., 1992). Various studies demonstrate that kappa opioid agonists possess antinociceptive properties after systemic administration in the rat (Leighton et at., 1988; Milan, 1989; Hylden et at., 1991). However, evidence of their efficacy after i.t. ad- Downloaded from jpet.aspetjournals.org at ASPET Journals on September 19, 216 ABBREVIATiONS: HP, hot plate; IF tall flick; PP, paw pressure; i.t., intrathecal; PPT, paw-pressure threshold; MPE%, maximal possible effect in percent; ANOVA, analysis of variance; MW, molecular weight. 136

2 1994 Alph2 Agonists in Inflammation 137 ministration is conflicting. Dose-dependent antinociceptive effects were found when U-5,488H, a selective kappa opioid agonist (VonVoigtlander et at., 1983), was administered i.t. and mechanical PP (Miaskowski et at., 1991) or formalininduced cutaneous pain (Pelissier et at., 199) were used as nociceptive stimuli. On the other hand, no antinociceptive effects were found on i.t. administration of U-5,488H when noxious heat (Piercey and Einspahr, 1989; Leighton et at., 1988), mechanical PP (Leighton et at., 1988) or carrageenaninduced cutaneous hyperalgesia combined with noxious heat (Hylden et at., 1991) were used. The antinociceptive actions of kappa opioids are thought to depend on the intensity and quality of the nociceptive stimulus (Tyers, 198; Milan, 1989, 199), which further complicates interpretation of the previous results. During peripheral unilateral inflammation, the inflamed paw has shown enhanced sensitivity to the antinociceptive action of systemically administered mu and kappa opioids (Stein et at., 1987, 1988; Milan and Colpaert, 1991; Stein, 1993) and to systemic and i.t. clonidine (Hylden et at., 1991). It has been suggested that, during peripheral inflammation, the enhanced analgesic effects of opioids might depend on an interaction with spinal noradrenergic mechanisms (Hylden et at., 1991). This study was designed to investigate 1) the antinociceptive efficacy of both i.t. and systemically administered dexmedetomidine and U-5,488H during peripheral inflammation and 2) the interaction between two selective atph2 adrenoceptor modulating agents (dexmedetomidine and atipamezole) and the kappa opioid agonist U-5,488H during peripheral inflammation. Materials and Methods The study protocol was approved by the Ethics Committee for animal research ofhelsinki University Central Hospital. The International Association for the Study of Pain Ethical Guidelines were adhered to in these studies. Animals Male Sprague-Dawley rats (B&K Universal, Sollentuna, Sweden) that weighed 25 to 3 g were housed five in a cage on a 12-hr light/dark cycle (lights on from 6: km. to 6: P.M.). The ambient temperature was kept at 21#{176}C and the rats had free access to standard laboratory food (Altromin no. 1324, Chr. Petersen AJS, Ringsted, Denmark) and tap water. After the surgery, they were housed individually in the same facilities until tested. Surgery The animals were anesthetized with a combination of s.c. fentanyl citrate.315 mg/mi and fluanisone 1 mg/mi (Hypnorm, Janssen Pharmaceuticals, Beerse, Belgium) at a dose of.8 mi/kg and midazolam 5 mg/mi (Dormicum, Roche, Base!, Switzerland) at a dose of.8 mi/kg. The i.t. catheters were implanted as described previously (Yaksh and Rudy, 1976; Post et at., 1987). In brief, the animals were positioned in a stereotactic apparatus and a fine polyethylene-b catheter(previously stretched to double its length when immersed in water at 75#{176}C; final length, 8. cm) was inserted into the subarachnoid space via the atlanto-occipital membrane. The catheter extended to the rostral end of the lumbar enlargement at the L1-L level. On the third postoperative day, the position of the tip of the catheter was checked for correctness by injecting 1 pj of 5% hyperbaric lidocaine (Lidocain Pond, Medipolar, Oulu, Finland) followed by a flush of 1 d of.9% saline into the catheter. Only animals that developed reversible bilateral hind limb paralysis within 5 ruin were accepted for further studies. The correct location ofthe catheter was finally controlled by dissecting the animalspostmortem. Any animals that showed signs ofneurological deficit after catheter implantation were excluded from the study and sacrificed with an overdose of enflurane (Efrane, Abbott Laboratories, Campoverde, Italy). Nociceptive Tests Before the nociceptive tests were performed, the rats were accimatized to the test situatious for at least 3 days. They were placed and held in the same transparent plexiglass cylinders (Kalso et at., 1991) in which they were restrained during the actual TF and PP testing. The cylinders were provided with two openings on the yentral surface, which allowed both hind paws of the rat to protrude freely. The tail protruded from a third opening at the caudal end of the cylinder. After some minutes, this procedure no longer seemed to cause any discomfort to the subjects because all rats were lying still and breathing normally; most of them had fallen asleep. All experiments were carried out in a quiet room between 8: A.M. and 2: P.M. Thermal nociception. The rats were placed on a HP (Harvard Apparatus, Edenbridge, Kent, UK) calibrated to 52.5 ±.5#{176}C and surrounded by a plexiglass cylinder to prevent the rat from escaping. The response criterion was either a lick of one hind paw or a jump with both hind paws. An absolute cutoiftime of3o sec was used and the average of two consecutive readings was taken. To get them accustomed to the apparatus, the rats were repeatedly placed on a cold plate for 3 consecutive days before testing. Animals with baseline latencies ofmore than 2 sec were excluded from further testing. The TF test (D Amour and Smith, 1941) was done with a Ugo Basile (Comerio, Italy) analgesimeter, an instrument that focuses a noxious beam of light (by a parabolic mirror) on the tail and automatically records the latency to removal. The average of three consecutive readings (6.5, 8 and 9.5 cm from the tip ofthe tail) was recorded and a cutoffti.me of8 sec was used to prevent tissue damage. The average base-line latency on the beam intensity used was 2.53 ±.4 sec (mean ± S.E.M., n = 115). Only rats with base-line latencies less than 4 sec were accepted for further testing. Mechanical nociception. The Randall and Selitto (1957) PP test was used. A mechanical stimulus was applied with a model Randall-Selitto analgesimeter (IITCfLife Science Instruments, Woodland Hifis, CA), which generates a linearly increasing mechanical force applied by a dome-shaped plastic tip into the dorsal surface of the rat s hind paw. The nociceptive threshold, i.e., PPT, was defined as the force in millimeters of mercury at which the rat withdrew its paw. The average of two consecutive trials, with an interval of 1 sec between them, was determined. The cutoff pressure was 3 mm Hg (approximately 1225 g; Sherman et at., 1988). The rats were trained in the test procedure for 3 consecutive days before data collection for at least 2 hr daily. Only rats with base-line PPTs less than 1 mm Hg were accepted for further testing. The average base-line PPT for these rats was ±.81 mm Hg (mean + S.E.M., n = 115). Induction of Inflammation As soon as the base-line thresholds had been measured, the rats were lightly anesthetized with enflurane and a single s.c. injection of lambda carrageenan (2 mg) was given with a 25-gauge needle in the plantar surface of the left hind paw. The injected hind paws had become maximally inflamed 2 hr after the injection, as assessed by the development ofedema, erythema and hyperalgesia to mechanical stimuli. The time course and magnitude of the inflammatory reaction have been described in detail by Hargreaves et at. (1988). The rats that received unilateral injections of carrageenan demonstrated normal eating and grooming behavior and normal levels of locomotor activity (ladarola et at., 1988) but tended to guard their inflamed paw. The number of animals used was kept to a minimum and each animal was used only once in the inflammation study. Downloaded from jpet.aspetjournals.org at ASPET Journals on September 19, 216

3 138 ld#{228}npun-heikkll#{228} et al. Vol. 271 Motor Coordination The rats were placed on a rod, 7 cm in diameter, covered with a rough rubber mat. The rotation speed of the rod was constant (12 rpm). The time the rat spent on the rod before falling off was recorded; the cutoff time was 18 sec. The rats were trained and acclimatized to the test for 3 days before testing by placing them on the rod several times. Drugs Drugs for i.t. administration were mixed; all doses were delivered in a total volume of 1 pi, followed by 1 tl of saline to flush the catheter. The drug solution was separated from the saline solution by an air bubble (.5 d) to prevent mixing. For the s.c. and intraplantar injections, the injection volumes were 1 mi/kg and.1 ml, respectively. The following drugs were used: atipamezole (MW, 348.5) and dexmedetomidine (MW, 236.7), both ofwhich were kindly donated by Dr. Raimo Virtanen (Orion-Farmos Group, Thrku, Finland). U-5,488H (MW, 385) was obtained from Sigma (St. Louis, MO). Doses were calculated as the free base and all solutions were freshly prepared in sterile physiological saline. The carrageenan (C-3889, Sigma) suspension was prepared by slowly adding the powder into sterile saline while vortexing the mixture. Statistics Data are expressed as means ± S.E.M. The data on TF and HP latencies were converted to MPE% with the following equation (Yaksh et at., 1976): MPE% = 1 x (postdrug latency - predrug latency)/(cutoff time - predrug latency). The data on the PP test were expressed as the change (in millimeters ofmercury) in the PPT from the base-line measurements. In rats with inflammation, the PPT measured 2 hr after induction ofinflammation was used as the base line. Because the group size varied within each experiment and a clear-cut normal distribution could not be established, nonparametnc Kruskal and Wallis (1952) one-way ANOVA was used for the statistical analysis of the data and pairwise comparisons between unpaired groups were performed with the Mann-Whitney Utest. The dose-response curves from both TF and PP tests were analyzed by two-way ANOVA with repeated measures to examine the possible influence of the peripheral inflammation on the responses. The statistical procedure was carried out with the pharmacological calculations systems software (Tallarida and Murray, 1987) and a statistical program (SYSTAT, Inc. Evanston, IL). Experiment I The dose-response relationships ofi.t. dexmedetomidine in control rats (rats without inflammation) were assessed in HP, TF and PP tests. The tests were performed in the following order: PP, HP 1, TF and HP 2. After the base-line values had been tested, the rats were given i.t. injections of saline or.15,.45, 1.35 or 4.5 g of dexxnedetomidine and the test responses were re-evaluated 3 mm later. The same rats were used three times for similar experiments but always after a 4-day washout period. The same animal was never given the same dose twice. The effect ofi.t. U-5,488H (5, 15, 5, 1 and 2 gig) in control rats was assessed with the PP and 17 tests. Responses were re-evaluated 3 mm after base-line measurements and administration of the drug. These rats were used only once. Experiment 2 The drug effects in rats with unilateral inflammation were assessed with the PP and TF tests. Base-line latencies were measured first in the PP test and thereafter in the TF. Inflammation was then induced. After 2 hr, the inflammation base lines were measured (PP followed by TF) and the rats received saline or the following drugs: dexmedetomicline (.15,.45, 1.35 and 4.5 g) i.t., U-5,488H (5, 15 and 5,.Lg) i.t. (Miaskowski et at., 1991; Pellisier et at., 199), U-5,488H (2.5, 7.5 and 22.5 mg/kg) s.c. (Stein et at., 1988), atipamezole (15 Lg) i.t. (Sullivan et at., 1992) or (3 mg/kg) s.c. (Scheinin et a!., 1988; Kalso et at., 1991). Responses were re-evaluated 3 mm later. Experiment 3 When the time course of the antinociception produced by saline, dexmedetomidine (3, 1, 3 and 1 pg/kg) or atipamezole given s.c. was evaluated in the rats with inflammation, the responses in the PP and TF tests were re-evaluated 3, 45, 6, 9 and 12 miii (Kalso et at., 1991) after the base-line measurements and drug administration. Experiment 4 The interaction studies were done after the dose-response studies on rats with unilateral inflammation. After induction ofthe inflammation and base-line measurements (PP followed by TF), the rats received saline or U-5,488H (2.5 mg/kg) s.c. They were given a concomitant injection of saline, dexmedetomidine (either 1 or 3,A.g/kg) or atipamezole (3 mg/kg) s.c. to the contralateral side or a concomitant injection ofsaline, dexmedetomidine (.15 or.45.&g)or atipamezole (15 g) i.t. The tests were re-evaluated 3 mm after i.t. administration; additional measurements were made at 45, 6, 9 and b2 mm when both drugs were given s.c. Results Effect of carrageenan-induced inflammation on nociceptive thresholds. In rats with carrageenan-induced inflammation of 2-hr duration, the PPTs of the inflamed paws (32.8 ±.34 mm Hg, n = 115) were significantly lower (P <.1) than the PPTs of the contralateral paws (57.72 ± 1. mm Hg, n = 115). Also the TF latencies were shorter in rats with unilateral inflammation. The mean base-line TF latency was 2.53 ±.4 sec (mean ± S.E.M., n = 115); 2 hr after induction of inflammation, it was 2.1 ±.3 sec (P <.22), which gives an average MPE% of ±.96. Antinociceptive and motor effects of i.t. dexniedetomidine. In control rats, dexmedetomidine produced dosedependent increases in the nociceptive thresholds in all three tests (PP, TF and HP) 3 mm after administration (figs. 1 and 2B). In the HP test, doses of.45, 1.35 and 4.5 g i.t. of dexmedetomidine significantly increased the MPE%. The i.t. administration of dexmedetomithne also resulted in doserelated increases in the nociceptive thresholds in the PP test 15 p125 E E 1 Lu z 75 z 5 C.) I Control ratj IoRight paw Lsft paw Saline #{149} 4.5 Dexmedetomidine (pg) i.t. Fig. 1. Dose-response curve for the antinociceptive effects of l.t. dexmedetomldine in control rats in the PP test at 3 mm after administration. Each point represents the mean ± S.E.M. (n = 8-9) P <.1, *P <.1 (by Mann-Whitney U test) compared with the saline-treated group. No differences between the paws. Downloaded from jpet.aspetjournals.org at ASPET Journals on September 19, 216

4 1994 Alph2 Agonists in Inflammation A Rats with inflammation Control paw 15 A Rats with kiflammation JControl paw p125 E 1 Inflamed paw 125 E E 1 lnflam.d paw a Lu C, 75 z x 5 C.). 25 Control rats Lu C, 75 z x 5 C) Saline Dexmedetomidine (pg) i.t. -25 DRUG l.t. U-5,488H s.c. - SAL ATI DEX DEX SAL DEX DEX ATI Lu U 1 B o Saline. Rats with inflammation * i5 4.5 Dexm.d.tomidine (pg) i.t. Fig. 2. Dose-response curves for the antinociceptive effects of i.t. dexmedetomidine in rats with unilateral inflammation in the PP test (A) and both in the rats with inflammation and in the control rats in the IF test (B). No differences between the paws (n = 8-9 in each group). *P <.5 compared with saline-treated group. For other details, see figure 1. on the rats with unilateral inflammation. It produced a doserelated antinociceptive effect in both paws and there were no statistically significant differences in this effect (two-way ANOVA with repeated measures) between the inflamed and the contralateral paws (F =.435, P = not significant; fig. 2A). Neither were there significant differences between the doseresponse curves of the inflamed paws and the paws of the control rats (F = 1.35, P = not significant; figs. 1 and 2A). In the TF test on the rats with unilateral inflammation, the latencies were already significantly elevated after the dose.15.tg of i.t. dexmedetomidine, whereas in the control rats, a dexmedetomidine dose of.45.tg i.t. was needed for a significant increase (fig. 2B). Atipamezole 15 pg i.t. had no antinociceptive effects in either test (fig. 3). Dexmedetomidine in the dose range.15 to 1.35 pg i.t. had no effect on motor coordination. The i.t. dose of4.5.tg resulted in a clear decline in rotarod performance (fig. 4). Antinociceptive and motor effects of s.c. dexmedetomidine and atipamezole. The antinociceptive effects of s.c. dexmedetomidine in rats with inflammation is shown in figure 5. The s.c. doses of < 3 pg/kg had no effect in the PP test (fig. 5A), whereas in the TF test, a dose of 3 pg/kg s.c. of dexmedetomidine raised the MPE% significantly (P <.5) compared with the saline-treated group (fig. 5B). The doses of : 2-2 DRUG Lt.SAL U-5,488H s.c. - B Rats with inflammation - A ATI DEX DEX aa bb SAL DEX DEX All Fig. 3. Antinociceptive effects of i.t. administered dexmedetomidine (DEX, dose given in micrograms) and atipamezole (ATI, in micrograms) combined with either s.c. saline (SAL) or U-5,488H (2.5 mg/kg) In the PP test (A) and in the TF test (B) in the rats with unilateral inflammation 3 mm after administration. Mean ± S.E.M. (n = 8-9). In all groups treated with U-5,488H, the PPT change in the inflamed paw was significantly higher than the change in the control paw. <.5, <.1 compared with the saline group. 5P <.5 compared with the inflamed paw treated with U-5,488H alone (A), P <.1 compared with the group treated with U-5,488H alone (B). bbp < O1 compared with the inflamed paw treated with DEX.45 p.g alone (A), P <.1 compared with the group treated with DEX.45 g alone (B). C <.5, cccp <.1 compared with the control paw treated with U-5,488H alone (A). 3.tg/kg and 1 pg/kg s.c. of dexmedetomidine increased the PPTs in both inflamed and control paws. No significant differences were seen in this effect between the paws (fig. 5A). Atipamezole (3 mg/kg s.c.) did not have an antinociceptive effect in the PP test (fig. 5A) or in the TF test (fig. 5B). The s.c. doses of 3 p.g/kg ofdexmedetomidine significantly impaired motor performance already at 15 man after administration (Fig. 4). Antmociceptive effects of i.t. and s.c. U-5, The kappa opioid agonist U-5,488H did not demonstrate antinociceptive activity in either the PP or the TF tests when given i.t. to rats with unilateral inflammation (5, 15 and 5 g; n = 6 in each group) 15 or 3 miii after injection or to a separate group of control rats (15, 5, 1 and 2 g; n 6 in each group) 15 or 3 mm after injection (data not shown). The s.c. dose of 2.5 mg/kg of U-5,488H had no effect on PPTs of the control paws or on the PPTs of the control rats but, at higher Downloaded from jpet.aspetjournals.org at ASPET Journals on September 19, 216

5 131 ld#{227}npun-heikkil#{228} at al. VoL I 6,.14 a 12 < #{176} E 175 Lu15#{176} C.) 75 : Saline p E E1 Ui C., 75 z 5#{149} C). 25 a. -25 DRUG s.c.: SAL U-5,488H s.c: B Rats with inflammation U-5,488H : 2-2 DRUG - s.c.: s.c.: Time (mm) Fig. 4. The effects of i.t. and s.c. dexmedetomidine on motor performance of rats measured with the rotarod. Mean ± S.E.M. (n = 4-6). A Rats with inflammation fjcontrol paw lnflamed paw 35* -u1 a d aa All DEX DEX DEX SAL DEX DEX All a a SAL All DEX DEX DEX DEX SAL DEX DEX All Fig. 5. Antinociceptlve effects of s.c. administered dexmedetomidine (DEX, in micrograms per kilogram) or atipamezole (ATI, in mg/kg) cornbined with either subcutaneous saline (SAL) or U-5,488H (2.5 mg/kg) in the PP test (A) and in the TF test (B) in rats with unilateral inflammation at 3 mm after administration. Mean ± S.E.M. (n = 8-9). * <.5, P <.1, <.1 compared with the saline-treated group. 5P <.5, P <.1 compared with the inflamed paw treated with U-5,488H alone. C <.5 compared with the control paw treated with U-5,488H alone (A). 5P <.5, compared with the group treated with U-5,488H alone (B). doses (7.5 and 22.5 mg/kg), dose-related antinociception was observed (fig. 6A). However, in rats with inflammation, dosedependent increases in PPTs of the inflamed paws were Lu U Saline U-5488H (mg/kg) s.c. B erats with inflammation * Control rats U-5488H (mg/kg) s.c. Fig. 6. A) Dose-response curves for the antinociceptive effects of s.c. U-5,488H on the PP test of the rats with unilateral inflammation and the control rats. B) IF test of both the control rats and the rats with unilateral Inflammation. Mean ± S.E.M. (n = 8-9). 4P <.1 compared with the control rats. For further details, see figures 1 and 2. found already with 2.5 mg/kg of U-5,488H (fig. 6A). Significant differences in PPTs were found between the inflamed and control paws (P <.1) after the dose of 2.5 mg/kg and a two-way ANOVA revealed a significant difference between the dose-response curves of the inflamed and contralateral paws (F = 5.52, P <.1) and between the inflamed paws and the paws of the control rats (F = 6.9, P < 1). Similarly, in the TF test, the dose of2.5 mg/kg ofu-5,488h was devoid of effect in the control rats but it resulted in a significant increase in the TF latencies of the rats with unilateral inflammation (fig. 6B). Significant differences were found in the TF latencies between the inflamed rats and the control rats at the dose-levels of2.5 mg/kg (P <.1) and 7.5 mg/kg (P <.1). The difference between the dose-response curves was also significant (F = 4.73, P <.1, by two-way ANOVA; fig. 6B). Effects of s.c. U-5,488H combined with dexmedetomidine and atipamezole. The antinociceptive effect of U-5,488H 2.5 mg/kg s.c. was not modified by the coadministration of atipamezole 15.tg i.t. (fig. 3) or 3 mg/kg s.c. (fig. 5) in either the PP or the TF test in the rats with inflammation. The combination of dexmedetomidine (.45.tg i.t.) and U-5,488H (2.5 mg/kg sc.) was more effective in both the inflamed paw and in the TF test than was dexmedetomidine (.45 ig i.t.) alone (P <.1) or U-5,488H (2.5 mg/kg s.c.) Downloaded from jpet.aspetjournals.org at ASPET Journals on September 19, 216

6 1994 Alph2 Agonists in Inflammation 1311 alone (P <.5 in the PP test, P <.1 in the TF test; fig. 3). In the control paws ofthe inflamed rats, both combinations of the i.t. doses (.15 and.45.tg) of dexmedetomidine with U-5,488H resulted in significantly higher PPTs than when the kappa agonist was given alone (fig. 3A). In the control paws, the dose-dependent antinociceptive effect of s.c. dexmedetomidine was unaffected by coadministration of s.c. U-5,488H, which alone was ineffective (fig. 5A). In the inflamed paw, however, the antinociceptive effect of the combination was subadditive (fig. 5A). The same phenomenon was evident in the TF test of the inflamed rats (fig. 5B). Discussion Carrageenan-induced inflammation causes edema of the paw and enhanced sensitivity of the paw toward mechanical stimuli. It also seems to result in thermal hyperalgesia of the tail, which suggests that the unilateral carrageenan inflammation not only induces peripheral hyperexcitabifity of the injected paw but also results in central sensitization. Our TF test was calibrated to detect antinociception, i.e., increasing latencies after drug treatments. It is therefore not very sensitive in the detection of hyperalgesia, which results in decreased TF latencies compared with the controls. However, the carrageenan inflammation resulted in significantly decreased latencies and negative MPE%s in all inflamed groups. Guilbaud et at. (1992) reviewed several of their previous studies in which carrageenan injection in the hind paw of the rat caused a decrease in the PPT not only in the injected paw but also in the opposite hind limb and even in the two forepaws. We found no clear mechanical hyperalgesia in the contralateral uninflamed paw but the thermal hyperalgesia of the tail suggests that central sensitization might have occurred. Guilbaud et at. (1986) reported increases in the electrophysiological responses of ventrobasal neurons of the thalamus to thermal and mechanical stimuli during peripheral carrageenan inflammation. The mechanisms of this central sensitization under pathological circumstances were reviewed by Woolf (1991). The prolonged changes in central excitability are thought to be produced by C-afferent fibers and carrageenan pretreatment was reported to alter C-fiber responsiveness to both thermal and mechanical stimuli (Handwerker et at., 1987; Woolf, 1992). Both norepinephrine levels and turnover in the dorsal horn ofthe spinal cord were also reported to increase during inflammation (Weil-Fugazza et al., 1986). In our study, during this central sensitization, i.t. dexmedetomidine was already effective in the thermal TF test at doses that produced no effect in the rats without unilateral inflammation. Selective activation of spinal atp/w-2 adrenoceptors during peripheral inflammation thus seems to result in inhibited spinal hyperexcitabiity. In addition, a thermal stimulus seems to be more sensitive than a mechanical stimulus to detect this effect. Our present findings are consistent with previous results on the efficacy of i.t. administered dexmedetomidine in thermal tests (Kalso et at., 1991; Fisher et at., 1991; Saeki and Yaksh, 1992; Takano and Yaksh, 1993). In addition, i.t. administered dexmedetomidine produced antinociception in tests with mechanical pressure as noxious stimulus, both in control rats and in rats with unilateral peripheral inflammation. This action was not due to the motor deficits or sedation produced by dexmedetomidine at higher doses. Also, i.t. clonidine produced dose-dependent antinociception but then the stimulus was thermal, either the TF test (Ossipov et at., 1988), the HP test (Takano and Yaksh, 1992) or noxious heat being applied to the inflamed hind paw of the rat (Hylden et at., 1991). The dose of clonidine needed for a 5% inhibition of spinal chord C-fiber responses was reported to be 5 g (Sullivan et at., 1987) compared with 2.5.tg of dexmedetomidine (Sullivan et at., 1992). In the study oftakano and Yaksh (1992), the ED5 doses of dexmedetomidine and clonidine were 3.2 and 27 g, respectively. After systemic administration, clonidine produced doserelated antinociception in mechanical PP tests in both control and inflamed paws (Fielding et at., 1978; Hayes et at., 1986). This effect of systemic clonidine is enhanced in the inflamed paws and is thought to result from activation of peripheral alphi and alph2 receptors (Hayes et at., 1986). Intraplantar injections of clonidine resulted in antinociception in carrageenan-inflamed hind paws. This effect was suggested to be mediated by activation of peripheral atph2 receptors and the release of an enkephalin-like substance (Nakamura and Ferreira, 1988). However, in this study, the doses of clonidine were extremely high (2.5-1 mg intraplantar) and those of carrageenan (.1 mg) were relatively low. In the present study, s.c. dexmedetomidine, which is far more selective for the atph2 receptors than clonidine, produced antinociception in the PP test only after doses that were clearly sedative and resulted in motor incoordination but no enhanced antinociception in the inflamed paw was found. Taken together, these findings suggest a role for the alphi rather than for the atph2 adrenoceptors in the mediation of the peripheral antinociceptive effect of these compounds in inflamed tissue. This view is supported by Campbell et at. (1992). Intraplantar injections directly into the inflamed tissue might further clarify these peripheral mechanisms. It was interesting that systemic dexmedetomidine significantly increased TF latencies in rats with inflammation at doses (3 i.g/kg) devoid ofeffect in the rotarod test. In previous studies, clearly higher doses (3 pg/kg s.c.) have not resulted in significant antinociception in the TF test on rats without unilateral inflammation (Kalso et at., 1991). This finding suggests an enhanced antinociceptive effect for s.c. administered dexmedetomidine in the thermal test during unilateral inflammation. The mechanism of this action is unclear but it could be due to the high lipophilicity and rapid penetration of dexmedetomidine into the central nervous system after s.c. administration. In previous studies, the kappa opioid agonist U-5,488H produced somewhat discrepant results when administered i.t. to adult rats. In some of the studies, relatively low i.t. doses of5, 5 and 5 ng (Miaskowski et al., 1991) or 2.35 and 6.97 g (Pelissier et at., 199) resulted in antinociception, whereas in others, only very high doses of (Millan et at., 1989), 1 (Suffivan and Dickenson, 1991) or 73.5 p.g (Malmberg and Yaksh, 1993), which possibly activated other than kappa opioid receptors, produced the same effect. Some studies report that even high i.t. doses of U-5,488H are totally devoid of an antinociceptive effect (Leighton et at., 1988; Piercey and Einspahr, 1989; Hylden et at., 1991). The age of rat was also shown to affect the potency of i.t. U-5,488H (Sullivan and Dickenson, 1991). In addition, there are significant differences between various species. Mice, for example, seem to be significantly more sensitive to Downloaded from jpet.aspetjournals.org at ASPET Journals on September 19, 216

7 131 2 ldinp#{227}#{228}n-heikkll#{228} et al. Vol. 271 i.t. administered kappa opioid agonists (Porreca et at., 1987; Takemori et at., 1988; Piercey and Einspahr, 1989). It was suggested that kappa agonists are efficient against nonthermal (e.g., pressure) noxious stimuli but less active against nociception produced by heat (Tyers, 198; Millan, 1989, 199). In addition, it was proposed that the low potency of U-5,488H reflects poor access of the agonist to the dorsal horn after i.t. administration (Sullivan and Dickenson, 1991). Despite the controversy over i.t. administration, U-5,488H showed reliable antinociceptive potency on systemic administration in various studies (Leighton et al., 1988; Milan, 1989; Hylden et at., 1991; Sullivan and Dickenson, 1991). In this case, however, the effects in supraspinal and peripheral sites, in addition to spinal sites, have to be taken into account. In rats with unilateral inflammation, the inflamed paw showed enhanced sensitivity toward the antinociceptive actions of both mu and kappa opioid agonists. This effect is thought to be mediated by activation of peripheral opioid receptors during the inflammation followed by attenuation of the excitability of the nociceptive input terminals and inhibition of the release of excitatory transmitters from central and peripheral endings of primary afferent neurons (Stein et at., 1987, 1988; Milan and Colpaert, 1991; Stein, 1993). In our study, the antinociceptive effect of systemic U-5,488H was clearly enhanced in the inflamed paw and also in the TF test of the rats with unilateral inflammation. These findings reflect the possible effects of the kappa agonist in spinal and supraspinal sites in addition to its peripheral actions. In the spinal cord, kappa receptor agonists depress firing of dorsal horn neurons without diminishing the release of spinal C- fiber transmitters, which suggests a postsynaptic site of action (Go and Yaksh, 1987; Millan, 199; Malmberg and Yaksh, 1993). The precise site of the supraspinal action that results in activation ofdescending antinociceptive systems is not yet known but periaqueductal gray and thalamus were suggested (Milan, 199). On the whole, kappa agonists seem to be effective in the inhibition ofthe facilitated transmission of nociceptive impulses during peripheral inflammation and central sensitization, at least toward both thermal and mechanical stimuli. The enhanced antinociceptive effect of mu opioids during peripheral inflammation was suggested to depend on spinal noradrenergic mechanisms (Hylden et al., 1991). The i.t. administration of noradrenergic agonists, such as norepinephrine (Hylden and Wilcox, 1983) and clonidine (Ossipov et al., 1989), resulted in a potentiation ofmorphine antinociception. The i.t. administration of idazoxan, an alph2 adrenoceptor antagonist, blocked the antinociceptive effect of i.t. morphine in the inflamed paw during peripheral carrageenan-induced hyperalgesia when heat was used as a nociceptive stimulus. This antagonism was evident also when both compounds were administered systemically (Hylden et al., 1991). In the present study, we studied the enhanced antinociceptive effect of a kappa opioid agonist during peripheral inflammation. The doses of 15.tg of atipamezole i.t. (Sullivan et al., 1992) and 3 mg/kg s.c. (Kalso et al., 1991) in other studies previously eliminated the antinociceptive effects of dexmedethmidine. In our study, on i.t. and s.c. administration together with s.c. U-5,488H, atipamezole did not modify the antinociceptive effect of the kappa agonist. The enhanced antinociceptive effects of kappa opioids in the inflamed paw thus do not seem to involve the atph2 adrenoceptors. The i.t. administration of dexmedetomidine with systemic U-5,488H resulted in an additive interaction. However, when both dexmedetomidine and U-5,488H were administered s.c., our results did not indicate any potentiation. On the contrary, the combination of systemic dexmedetomidine with systemic U-5,488H resulted in weaker antinociception than with U-5,488H alone in the TF and in the inflamed paw. However, in the control paw, the antinociception was not modified. One reason for this lack of effect in the inflamed paw may be a pharmacokinetic interaction that results in delayed absorption of U-5,488H into the paw because systemic alph2 agonists are known to cause bradycardia and hypotension, whereas atph2 antagonists stimulate rather than depress the cardiovascular system. In addition, the activation of peripheral alph2 adrenoceptors results in vasoconstriction, which could further delay the effect of U-5,488H. Also the quality of the nociceptive stintulus has to be taken into account. In this study, we used a mechanical, not a thermal stimulus, which might be more sensitive in the detection ofthe antinociceptive effects in the inflamed tissue. In addition, concomitant intraplantar administration of the compounds might be more effective and also further clarify these peripheral interactions. In conclusion, our data show that the selective alph2 adrenoceptor agonist dexmedetomidine can produce antinociception against noxious mechanical stimuli on i.t. administration. This antinociceptive effect is also evident during peripheral unilateral inflammation but is not enhanced in the inflamed paw. However, the efficacy of both i.t. and probably also s.c. dexmedetomidine is enhanced during the central hyperalgesia that results from the peripheral inflammation. Therefore, the effect of dexmedetomidine is likely to be mediated by the activation of central rather than peripheral alph2 adrenoceptors. The antinociceptive effect of systemic U-5,488H is further enhanced during peripheral unilateral inflammation and central sensitization but is not antagonized by an atph2 adrenoceptor antagonist. Moreover, i.t. coadministration of dexmedetomidine with subeutaneous U-5,488H results in additive antinociception toward both thermal and inflammatory stimuli, which further supports the role ofspinal mechanisms for dexmedetomidine and peripheral mechanisms for U-5,488H. Acknowledgments The authors are grateful to Dr. Raimo Virtanen (Orion-Farmos Group Ltd.) for the gift of atipamezole and dexmedetomidine. References CAMPBELL, J. N., Mzvza, R. A., DAVIS, K S. N.: Sympathetically maintained pain a unifying hypothesis. In Hyperalgesia and AllOdynia, ed. by W. D. Willis, Jr., pp , Raven Press, New York, D M&ouit, F. E. Stum, D. L: A method for determining loss of pain sensation. J. PharmacoL Exp. Ther. 72: 74-79, FIELDING, S., Wzuczit, J., Hns, M., Szzwcz.x, M., Novica, W. J., JR. m HARBANS, L.: A comparison of clonidine with morphine for antinociceptive and antiwithdrawal actions. J. Pharmacol. Exp. Ther. 27: , FISHER, B. C., Zomcow, M. N., Yxsti, T. L. Pz rmson, B. M.: Antinociceptive properties ofintrathecal dexmedetomidine in rats. Eur. J. PharmacoL 192: , Go, V. L W. m YAKSH, T. L.: Release ofsubatance P from the cat spinal cord. J. Physiol. (Lond.) 391: , Guuum, G., Bzoir, J. M., KAYSER, V. GAUTRON, M.: Modifications in the responsiveness of rat ventrobasal thalamic neurons at different stages of carrageenan-produced inflammation. Brain Res. 385: 86-98, Gunaun, G., KAYSER, V., Arr, N. n Bzois r, J. M.: Evidence for a central Downloaded from jpet.aspetjournals.org at ASPET Journals on September 19, 216

8 1994 Alph2 Agonists in Inflammation 1313 contribution to secondary hyperalgeaia. In Hyperalgesia and Allodynia, ed. by W. D. Willis, Jr., pp , Raven Press, New York, HANDwERKER, H.., MoN, F., Kocsza, L s Razit, P. W.: Nociceptor functions in intact skin and in neurogemc or non-neurogenic inflammation. Acts Phyaiol. Hung , HAROREAvES, K, Dummt, it, Becwr, F., Fwass, C. n Joan, J.: A new and sensitive method for measuring thermal nociception in cutaneous hyperalgeaia. Pain 32: 77-88, HAYES, A. G., Sicmioiz, M. Tyzas, M. B.: phadrenocepthr-methated antinociception and sedation in the rat and dog. Neuropharmacology 25: , Hmsa, J. L K, Tiiotis, D. A., hautoi, M. J., Numi, R. L m DUBNER, R: Spinal opioid analgesic effects are enhanced in a model ofunilateral inflammation/hyperalgeaia: Possible involvement of noradrenergic mechanisms. Eur. J. Pharmacol. 194: , Hmsa, J. L K rw Wn.cox, G. L.: Pharmacological characterization of substance P-induced nociception in mice: Modulation by opioid and noradrenergic agomats at the spinal level. J. PharmacoL Exp. Ther. 226: , L4DARLA, M. J., Bnny, L S., Daxsci, G. Auin DuBNER, R.: Enhancement of dynorphin gene expression in spinal cord following experimental inflammation: Stimulus specificity, behavioral parameters and opioid receptor bindlag. Pain 35: , KALSO, E. A., P#{246}YHIA,R Aum RosENBERG, P. H.: Spinal antinociception by dexmedetomidine, a highly selective alphadrenergic agonist. Pharmacol. Toxicol. 68:14-143,1991. KRU8KALL, W. H. Wu.:s, W. A.: Use of ranks on one criterion variance analysis. J. Am. Stat. Aasoc. 47: , LEIGHTON, G. E., RODRIGUEZ, R E., Hiu, R. G. m Huouns, J.: x-opioid agomsts produce antinociception after i. v. and i. c. v. but not intrathecal administration in the rat. Br. J. PharmacoL 93:553-56,1988. MALMBERO, A. B. YmH, T. L.: Pharmacology of the spinal action of ketorolac, morphine, ST-91, U-5,488H, and L-PIA on the Formalin test and an isobolographic analysis of the NSAID interaction. Anesthesiology , MIAsxowsKI, C., &rrrzits, K. L, Tiwo, Y.. Aco LzviNE, J. D.: Comparison of the antinociceptive andmotoreffecta ofintrathecal opioid agomsts lathe rat. Brain Res. 553: 15-ice, MILLAN, M. J.: Kappopioid receptor-mediated antinociception in the rat. I. Comparative actions of mu-and kappopioids against noxious thermal, pressure and electrical stimuli. J. Pharmacol. Exp. Ther. 251: , MILLAN, M. J.:,c-Opioid receptors and analgesia. Trends Pharmacol. 5th. 11: 7-76, 199. Mzui, M. J. CouanT, F. C.: Opioid systems in the response to inflammatory pain: Sustained blockade suggests role of ic- but not -opioid recaptore in the modulation of nociception, behaviour and pathology. Neuroscience 42: , MILLAN, M. J., Czwowsia, A., Lipicowsici, A. u Hmz, A.: Kappopioid receptor-mediated antinociception in the rat. II. Supraspinal in addition to spinal sites of action. J. Pharmacol. Exp. Ther. 251: , Nicaiuit, M. m Fmutzm, S. H.: Peripheral analgesic action of clomdine: Mediation by release of endogenous enkephalin-like substances. Eur. J. Pharmacol 146: ,1988. Ossipov, M. H., Suanz, L J. m SPAULDING, T. C.: A comparison of the antinociceptive and behavioral effects ofintrathecally administered opiates, a2-adrenergic agonists andlocsl anesthetics in mice and rats. Anesth. Analg. 67: , Ossipov, M. H., Suanz, L J. co SPAULDING, T. C.: Antinociceptive interactions between alphadrenergic and opiate agonists at the spinallevelin rodents. Aneath. Anaig. 68: , PELISSIER, T., Pznz, C., Soro-Moy4uto, R., Svzniu, H. ru Hmrthcez, A.: Analgesia produced by intrathecal administration of the K opioid agonist, U-5,488H, on Formalin-evoked cutaneous pain in the rat. Eur. J. PharmacaL , 199. PIERCEY, M. F. n ENsPunt, F. J.: Spinal analgesic actions ofkappa receptor agonists, U-5488H and spiradoline ( ). J. Pharmacol. Exp. Ther. 251: , Pomzc, F., Mosazac, H. I., Oies, J. R., Busxs, T. F. in Cow, A.: Supraspinal and spinal potency of selective opioid agonists in the mouse writhing test. J. Pharmacol. Exp. Ther. 24th , PosT, C., GoaDs, T., Mmice, B. G., Aacsza, T. m Fitamti, J.: Antinociceptive effects and spinal cord tissue concentrations after intrathecal injection of guanfacine or clonidine into rats. Anesth. Analg. 66: , RANDALL, L. n Szimo, J. J.: A method for measurement of analgesic activity on inflamed tissue. Arch. hit. Pharmacodyn. 4: , SAlK:, S. uyxss, T. L: Suppression by spinal alph2 agonists ofmotor and autonomic responses evoked by low- and high-intensity thermal stimuli. J. Pharmacol. Exp. Ther. 26: , SCHEININ, H., MACDONALD, E. u Scitxnmi, M.: Behavioural and neurochemical effects of atipamezole, a novel a2-adrenoceptor antagonist. Eur. J. Pharmacol. 151: 35-42, SHERMAN, S. E., Looins, C. W., Miu, B. 4I m CERVENKO, F. W.: Tolerance to intrathecal oxymetazoline-induced analgesia, with paradigm-dependent cross-tolerance to intrathecal morphine. J. Pharmacol. Exp. Ther. 245: , STEIN, C.: Peripheral mechanisms of opioid analgesia. Anesth. Analg. 76: , STErN, C., MILLAN, M. J., Annilyr, R. M., PrrzR, K im Hznz, A.: Alterations in opioid systems and behavior in adjuvant-induced inflammation of the rat hind limb. Soc. Neurosci. Abstr. 13: 641, STEIN, C., Miu, M. J., YssouRmIs, A. m Hmz, A.: Antinociceptive effects of,.- and K-SgoIUSt8 111 inflammation are enhanced by a peripheral opioid receptor-specific mechanism. Eur. J. Pharmacol. 155: , SULLWAN, A. F., DASHWOOD, M. R. m DICKENSON, A. H.: a2-adrenoceptor modulation of nociception in rat spinal cord: Location, effects and interactions with morphine. Eur. J. Pharmacol. 138: , SULLWAN, A. F. n DICKENSON, A. H.: Electrophysiologic studies on the spinal antinociceptive action ofkappa opioid agonists in the adult and 21-day.old rat. J. Pharmacol. Exp. Ther. 256: , SULLIVAN, A. F., Kso, E. A., MCQUAY, H. J. rm DIcicsoN, A. H.: The antinociceptive actions of dexmedetomidine on dorsal horn neuronal respouses in the anaesthetized rat. Eur. J. Pharmacol. 215: , To, Y. AND YAKSH, T.: Characterization of the pharmacology of intrathecally administered alph2 agonists and antagonists in rats. J. Pharmacol. Exp. Ther. 261: , Tuco, Y. AND YxsH, T.: Chronic spinal infusion of dexmedetomidine, ST-91 and clonidine: Spinal alpha2 adrenoceptor subtypes and intrinsic activity. J. Pharmacol. Exp. Ther. 264: , TAKEMORI, A. E., Ho, B. Y., Nzszm, J. S. n PORTOGRESE, P. S.: Norbinaltorphimine, a highly selective kappa opioid antagonist in analgesic and receptor binding assays. J. Pharmacol. Exp. Ther. 246: , TuAiun& R. J. Aco Muiuty, R. B.: Manual ofpharmacologic Calculations with Computer Programs, Springer-Verlag, New York, TYER8, M. B.: A classification of opiate receptors that mediate antinociception in animals. Br. J. Pharmacol. 69: , 198. VIRTANEN, R: Pharmacological profiles of medetomidine and its antagonist, atipamezole. Acts Vet. Scand. 85: 29-37, Vm riz, R., SAvoi, J.-M., So, V. rm Nnw, L.: Characterization of the selectivity and potency ofmedetomidine as an a2-adrenoceptor agonist. Eur. J. Pharmacol , VONVOIGTLANDER, P. F., Lim, R. A. ii Lunmis, J. H.: U-5,488H; a selective and structurally novel non-mu (kappa) opioid agonist. J. Pharmacol. Exp. Ther. 224: 7-12, WanFuzzA J., GODEFROY, F., Mcziu, V. crm BEssoN, J. M.: Increased norepinephrine and uric acid levels in the spinal cord ofarthritic rats. Brain Has. 374: , Woou, C. J.: Generation ofacute pain: Central mechanisms. Br. Med. Bull. 47: , Woou, C. J.: Excitability changes in central neurons following peripheral damage. Role of central sensitization in the pathogenesis of pain. In Hyperalgesia and Allodynia, ed. by W. D. Willis, Jr., pp , Raven Press, New York, YAKSH, T. L: Pharmacology of spinal adrenergic systems which modulate spinal nociceptive processing. Pharmacol. Biochem. Behav. 22: , YAKSH, T. L., NOUEIHED, R. Y. iw Duit, P. A. C.: Studies ofthe pharmacology and pathology of intrathecally administered 4-anilinopiperidine analogues and morphine in the rat and cat. Anesthesiology 64: 54-66, YAKSH, T. L. AND Rury, T. A.: Chronic catheterization of the spinal subarachnoid space. Physiol. Behav. 17: , Send reprint requests to: Juhana J. IdSnpaAn-Heikkil#{228}, M.D., Institute of Biochemistry, Department of Pharmacology and Toxicology, University of Helsinki, P.O. Box 8 (Siltavuorenpenger 1), FIN-14 Helsinki, Finland. Downloaded from jpet.aspetjournals.org at ASPET Journals on September 19, 216