Antinociceptive Effects of AS1069562

European Journal of Pharmacology 733 (2014) 54–61

Contents lists available at ScienceDirect

European Journal of Pharmacology

journal homepage: www.elsevier.com/locate/ejphar

Neuropharmacology and analgesia Antinociceptive effects of AS1069562, the (þ)-isomer of indeloxazine, on spinal hypersensitivity induced by intrathecal injection of prostaglandin in mice: Comparison with duloxetine and amitriptyline

Nobuhito Murai n, Mina Tsukamoto, Seiji Tamura, Toshiaki Aoki, Nobuya Matsuoka

Pharmacology Research Laboratories, Drug Discovery Research, Astellas Pharma Inc., 21 Miyukigaoka, Tsukuba-shi, Ibaraki 305-8585, Japan article info abstract

Article history: The (þ)-isomer of indeloxazine AS1069562 exerts multiple pharmacological actions including the Received 22 October 2013 inhibition of serotonin (5-HT) and norepinephrine reuptake and analgesia in experimental animal pain Received in revised form models. Here, we evaluated the antinociceptive effects of AS1069562 and the antidepressants duloxetine 25 February 2014 and amitriptyline in mouse models of prostaglandin-induced spinal hypersensitivity. Prostaglandin E2 Accepted 16 March 2014 (PGE ) and F α (PGF α) were intrathecally administered to induce spinal hypersensitivity, causing tactile Available online 1 April 2014 2 2 2 allodynia in mice. Allodynia induced by PGF2α but not by PGE2 was suppressed by desensitization of Keywords: C-fibers with systemic pretreatment with resiniferatoxin. C-fiber hyperexcitability might therefore play a Pain role in allodynia induced by PGF2α but not PGE2. In the PGE2-induced allodynia model, AS1069562 and Allodynia duloxetine significantly suppressed allodynia, whereas amitriptyline did not. In the PGF α-induced AS1069562 2 allodynia model, AS1069562 and amitriptyline significantly ameliorated allodynia, whereas duloxetine Duloxetine Amitriptyline did not. To demonstrate the broad effects of AS1069562 compared to duloxetine, additional studies were 5-HT receptor conducted to elucidate other target mechanisms of AS1069562 beyond 5-HT and norepinephrine reuptake inhibition. AS1069562 exhibited affinity for both 5-HT1A and 5-HT3 receptors, and the analgesic

effect of AS1069562 on PGF2α-induced allodynia was significantly blocked by the 5-HT1A receptor antagonist (S)–WAY100135 and the 5-HT3 receptor agonist SR57227. Taken together, these results indicate that AS1069562 inhibits both C-fiber- and non-C-fiber-dependent prostaglandin-induced allodynia, while duloxetine inhibits only non-C-fiber-triggered allodynia, and amitriptyline inhibits only C-fiber-triggered allodynia. These broad antinociceptive effects of AS1069562 may be due not only to 5-HT

and norepinephrine reuptake inhibition but also to its effects on 5-HT receptors such as 5-HT1A and 5-HT3 receptors. & 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).

1. Introduction (þ)-isomer of indeloxazine and also exerts inhibitory effects against 5-HT and norepinephrine reuptake (Shimizu-Sasamata Antidepressants such as the serotonin (5-HT) and norepinephr- et al., 1993). We recently found that AS1069562 significantly ine reuptake inhibitor duloxetine and the tricyclic antidepressant improved both mechanical allodynia and thermal hyperalgesia in amitriptyline are currently used to treat chronic pain, such as a rat model of chronic constriction injury (CCI)-induced neuro- neuropathic pain. The analgesic effects of these antidepressants pathic pain. In contrast, duloxetine only induced a significant are based on the enhancement of the serotonergic and noradre- amelioration of mechanical allodynia, while amitriptyline signifi- nergic descending inhibition system (Thor et al., 2007), which cantly ameliorated thermal hyperalgesia. Further, AS1069562 presynaptically inhibit glutamate release at the spinal cord increased the ratio of both 5-HT and norepinephrine to their (Yoshimura and Furue, 2006). metabolites contents in the rat spinal cord, while duloxetine Indeloxazine is a cerebral metabolic enhancer with multiple slightly increased only the ratio of 5-HT to its metabolite contents pharmacological effects, such as 5-HT and norepinephrine reup- (Murai et al., 2014). Although these results imply that AS1069562 take inhibition (Yamamoto, 1990). AS1069562 is the optical may have distinct spinal analgesic mechanisms compared with duloxetine and amitriptyline, no such mechanism has been elucidated. n Corresponding author. Tel.: þ81 29 863 7174; fax: þ81 29 856 2515. Mouse models of intrathecal (i.t.) prostaglandin-induced allodynia E-mail address: [email protected] (N. Murai). are useful in investigating spinal antinociceptive mechanisms of http://dx.doi.org/10.1016/j.ejphar.2014.03.038 0014-2999/& 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). N. Murai et al. / European Journal of Pharmacology 733 (2014) 54–61 55 analgesics, as their evaluation in a pure allodynic state facilitates the and all efforts were made to minimize the number of animals used identification of active sites in the pain pathway. Intrathecal injection as well as their suffering. of prostaglandin E2 (PGE2) or prostaglandin F2α (PGF2α)inducetactile allodynia, a clinically important manifestation of neuropathic pain, 2.3. Binding assay by facilitating glutamate release from presynaptic terminals at the spinal cord dorsal horn via prostaglandin E receptor 1 (EP1) and Binding assays were conducted using rat cerebral cortex (5.0 mg prostaglandin F receptor (FP), respectively (Minami et al., 1992, of protein) for the 5-HT1A receptor, rat whole brain (5.0 mg of 1994a, 1994b; Muratani et al., 2003; Tsukamoto et al., 2010). protein) for the 5-HT1B receptor, human recombinant 5-HT2B recep- In this study, first, we confirmed the dose-dependent induction tor (7.0 mg of protein; Perkin Elmer, Boston, MA, USA), and human of tactile allodynia including their time course via i.t. administra- recombinant 5-HT3 receptor (2.5 mg of protein; Perkin Elmer). For the 3 tion of PGE2 and PGF2α in mice. In addition, to investigate the 5-HT1A receptor binding assay, [Propyl-2,3-ring-1,2,3- H]-8-hydroxy- dependence of C-fiber hyperexcitability in these spinal hypersen- DPAT (0.47 nM; Perkin Elmer) and competitor (5-HT hydrochloride; sitivity models, we examined the effects of resiniferatoxin (RTX), Sigma-Aldrich) were incubated at 37 1Cfor30minin50mMTris- 125 an ultrapotent analog of capsaicin. After these validative experi- HCl buffer (pH 7.4). For the 5-HT1B receptor binding assay, [ I]iodo- ments, we assessed the antinociceptive effects of AS1069562, (7)-cyanopindolol (0.01 nM; Perkin Elmer) and competitor (5-HT duloxetine, and amitriptyline in these models to determine the hydrochloride; Sigma-Aldrich) were incubated at 25 1Cfor60minin different spinal antinociceptive mechanisms of these drugs. 50 mM Tris-HCl buffer (pH 7.4). For the 5-HT2B receptor binding Further, to determine whether or not AS1069562 exerts its activity assay, 2-(þ)-[125I]iodo-lysergic acid diethylamide (0.27 nM; Perkin via different mechanisms from its 5-HT and norepinephrine Elmer) and competitor (5-HT hydrochloride; Sigma-Aldrich) were reuptake inhibition, we investigated the possible involvement of incubated at 37 1Cfor30minin50mMTris-HClbuffer(pH7.4) cross activity for 5-HT receptor subtypes in the antinociceptive containing 4 mM CaCl2.Forthe5-HT3 receptor binding assay, 3 effect using spinal pharmacological tools, a 5-HT1A receptor [N-methyl- H]–GR65630 (0.48 nM; Perkin Elmer) and competitor antagonist and a 5-HT3 receptor agonist. The results of this study (MDL72222, 5-HT3 receptor antagonist; Sigma-Aldrich) were incu- provided novel mechanistic aspects of AS1069562 different from a bated at 25 1Cfor60minin50mMTris-HClbuffer(pH7.4)contain- selective 5-HT and norepinephrine reuptake inhibitor, duloxetine. ing 5 mM MgCl2 and 1 mM EDTA. To determine non-specificbinding, 100 mM 5-HT hydrochloride for 5-HT1A,5-HT1B,or5-HT2B receptor, or 100 mM MDL72222 for 5-HT3 receptor was used. Incubated 2. Materials and methods mixtures were filtered using a cell harvester. The filter paper was rinsed three times with 50 mM Tris-HCl buffer (pH 7.4), and the 2.1. Drugs radioactivity of paper was determined via scintillation counting. Specificbindingwasdefined as a portion of total binding, which

(R)-2-[(1 H-inden-7-yloxy)methyl]morpholine monobenzene- was replaced by 100 mM 5-HT hydrochloride for the 5-HT1A,5-HT1B, sulfonate (AS1069562) and duloxetine were synthesized at Astel- or 5-HT2B receptor, or MDL72222 for 5-HT3 receptor. Dissociation las Pharma Inc. (Ibaraki, Japan). Hydrochloride salts of AS1069562 constant (Kd) and binding site density (Bmax)werecalculatedvia and duloxetine were used as AS1069562 and duloxetine, respec- Scatchard analysis. IC50 values were calculated from the non-linear tively. Amitriptyline, RTX, PGE2, and PGF2α were purchased from regression analysis. Values of the apparent equilibrium dissociation

Sigma-Aldrich (St. Louis, MO, USA). The 5-HT1A receptor antagonist constant of inhibitors (Ki)werecalculatedusingthemethodofCheng – (S) WAY100135 and 5-HT3 receptor agonist SR57227 were pur- and Prusoff (1973). chased from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). For in vivo studies, AS1069562, duloxetine, and amitriptyline were 2.4. Intrathecal prostaglandin-induced allodynia model in mice suspended in distilled water and administered at an oral (p.o.) dose of 10 ml/kg. AS1069562 and duloxetine were administered at The prostaglandin-induced spinal hypersensitivity model was 3, 10, and 30 mg/kg, while amitriptyline was administered at 10, induced as previously described (Tsukamoto et al., 2010). Mice 30, and 100 mg/kg as previously reported (Katoh et al., 1995; were randomly divided into various groups, with eight animals per Shimizu-Sasamata et al., 1993). RTX was suspended in solution of group. A 30-gauge stainless steel needle attached to a microsyr- 10% EtOH, 10% Tween 80, and 80% saline and was administered at a inge was inserted between the L5 and L6 vertebrae of conscious – subcutaneous (s.c.) dose of 0.001 mg/10 ml/kg. (S) WAY100135 mice, and a prostaglandin (PGE2 or PGF2α in saline at a volume of and SR57227 were suspended in saline and administered at an i.t. 5 μl) was slowly administered via i.t. injection into the subarach- μ dose of 2 nmol/5 l/animal. PGE2 and PGF2α were dissolved in noid space (Hylden and Wilcox, 1980). Saline was used as a vehicle ethanol, and an aliquot of the desired solution was evaporated control of PGE2 and PGF2α. Tactile hypersensitivity was assessed 5, under nitrogen gas to remove the ethanol. PGE2 and PGF2α were 10, 15, and 30 min after i.t. injection of prostaglandin by lightly then dissolved in saline and administered at an i.t. volume of 5 μl/ stroking the ﬂank of the mice with a paintbrush. The allodynia animal. PGE2 was administered at 1, 10, and 100 ng/animal, and response was ranked as 0 (no response), 1 (avoidance), or 2 PGF2α at 100, 300, and 1000 ng/animal to investigate dose depen- (vigorous squeaking or strong avoidance), with scores expressed dency. PGE2 was administered at 10 ng/animal and PGF2α at as a percentage of the maximum possible cumulative score for all 300 ng/animal for drug evaluation. Drug concentrations were time points. To conﬁrm the dose-dependent induction of tactile calculated in terms of free base. allodynia, the following doses were administered: PGE2 at 1, 10, and 100 ng/animal and PGF2α at 100, 300, and 1000 ng/animal. 2.2. Animals Doses of prostaglandins (i.t.) that produced a submaximal allodynic

response (PGE2: 10 ng/animal; PGF2α: 300 ng/animal) were selected. Male imprinting control region (ICR) mice (18–22 g; Japan SLC, To investigate the involvement of C-fiber in prostaglandin-induced Hamamatsu, Japan) were used for in vivo experiments. Animals hypersensitivity, we examined desensitization of C-fibers with RTX. were group-housed under a 12-h light-dark cycle (light on: 7:30– RTX (0.001 mg/kg s.c.) was administered for 24 h prior to prosta- 19:30) at room temperature (2371 1C) with free access to food glandin injection, thereby desensitizing C-fibers. A solution of 10% and water. All animal experimental procedures were approved by EtOH, 10% Tween 80, and 80% saline was used as a vehicle control for the Committee for Animal Experiments of Astellas Pharma Inc., RTX. To assess their analgesic effectsonprostaglandin-induced spinal 56 N. Murai et al. / European Journal of Pharmacology 733 (2014) 54–61 hypersensitivity, AS1069562 (3, 10, and 30 mg/kg) was orally admi- 3. Results nistered at 30 min prior, duloxetine (3, 10 and 30 mg/kg) at 240 min prior, and amitriptyline (10, 30 and 100 mg/kg) at 60 min prior to 3.1. Binding affinities for 5-HT receptors prostaglandin injection. Pretreatment with AS1069562, duloxetine, or amitriptyline at these time points enabled almost the maximum To determine whether or not AS1069562 has mechanisms concentration level of each to be obtained (Murai et al., 2014; other than those for 5-HT and norepinephrine reuptake inhibition Yamaguchi et al., 1998). Distilled water was used as a vehicle control contributing to its different ranges of analgesic effects from for AS1069562, duloxetine, or amitriptyline. In the mechanistic study duloxetine, we investigated the binding affinities of AS1069562 of AS1069562, the 5-HT1A receptor antagonist (S)–WAY100135 and duloxetine for 5-HT receptors using rat or human 5-HT1A, (2 nmol) and 5-HT3 receptor agonist SR57227 (2 nmol) were admi- 5-HT1B, 5-HT2B, and 5-HT3 receptors. The Ki values of AS1069562 nistered (i.t.) 30 min prior to the PGF2α injection to determine and duloxetine for 5-HT receptors are listed in Table 1. whether or not AS1069562 activates the 5-HT1A receptor and inhibits In synaptosomal preparations of rat brain, AS1069562 inhibited the 5-HT3 receptor at the spinal cord. Saline was used as a vehicle 5-HT uptake with an IC50 value of 0.83 μM and norepinephrine control of (S)–WAY100135 and SR57227. Eight rats were used in each uptake with an IC50 value of 6.8 μM(Shimizu-Sasamata et al., experimental group. The number of animals was determined in a 1993). Further, duloxetine inhibited 5-HT uptake with an IC50 preliminary study to reproduce the drug efficacy and minimize the value of 0.0026 and norepinephrine uptake with an IC50 value of number of animals necessary for the study. 0.0070 μM(Wong et al., 1993). Compared to the inhibition of 5-HT and norepinephrine uptake, the respective approximate affinities of AS1069562 for each of the 5-HT receptors were as follows: 1.7-

and 0.21-fold lower (5-HT1A receptor), 3.0- and 0.37-fold lower 2.5. Statistical analysis (5-HT1B receptor), 1.8 and 0.22-fold lower (5-HT2B receptor), and 3.1- and 0.38-fold lower (5-HT3 receptor). Compared to the Data are expressed as the mean7the standard error of the inhibition of 5-HT and norepinephrine uptake, the respective mean (S.E.M.). Normal group is i.t. vehicle-treated and s.c. or p.o. approximate affinities of duloxetine for binding 5-HT receptors vehicle-treated group. Control group is i.t. prostaglandin-treated were as follows: 1692- and 629-fold lower (5-HT1A receptor), and s.c. or p.o. vehicle-treated group. Significant differences 1308- and 486-fold lower (5-HT1B receptor), 154- and 57-fold between the two groups (normal group vs. control group lower (5-HT2B receptor), and 5769- and 2143-fold lower (5-HT3 (Figs. 2–5), control group vs. RTX-treated group (Fig. 2), control receptor). These results indicate that the potencies of AS1069562 group vs. AS1069562-treated group (Fig. 5), AS1069562-treated for binding 5-HT receptors and inhibiting 5-HT and norepinephr- group vs. AS1069562 and 5-HT1A antagonist or 5-HT3 agonist- ine uptakes are similar (0.21- to 3.1-fold difference), while those of treated group (Fig. 5)) were assessed using Student's t test, while duloxetine are different (57- to 5769-fold difference). those among more than two groups (i.t. vehicle-treated group vs. prostaglandin-treated groups (Fig. 1B), control group vs. drug- 3.2. Prostaglandin-induced allodynia in mice treated groups (Figs. 3 and 4)) were assessed using one-way analysis of variance followed by Dunnett’s multiple comparison To determine the appropriate time course and dose of pros- post-test. Po0.05 was considered significant. taglandin for drug evaluation, we evaluated the dose dependent

Fig. 1. Tactile allodynia in mice following intrathecal injection of PGE2 (a) and PGF2α (b). (A) Time course of allodynia induction from 5 to 30 min after i.t. injection of PGE2

(1, 10, and 100 ng/animal) (a) or PGF2α (100, 300, and 1000 ng/animal) (b). (B) Dose-dependency of PGE2 (a) or PGF2α (b) in cumulative allodynia scores within 30 min of i.t. stimulation. Intrathecal injection of PGE2 and PGF2α induced allodynia in a dose-dependent manner from 5 to 30 min after i.t. injection. Data (the mean7S.E.M. for 8 mice) are presented as percentage of the maximum cumulative score. Vehicle (black column) represents i.t. injection of saline. nnPo0.01, nnnPo0.001 by Dunnett's test compared with vehicle-treated group. N. Murai et al. / European Journal of Pharmacology 733 (2014) 54–61 57

Fig. 2. Analgesic effects of RTX on PGE2- (A) and PGF2α- (B) induced allodynia in mice. Subcutaneous injection of RTX (0.001 mg/kg) 24 h before i.t. injection of PGE2 (10 ng/ animal) (A) or PGF2α (300 ng/animal) (B). Desensitization of C-ﬁbers with RTX did not suppress PGE2-induced allodynia, but completely suppressed PGF2α-induced allodynia. Data (the mean7S.E.M. for 8 mice) are presented as percentage of the maximum cumulative score. Normal (white column) represents i.t. injection of saline and s.c. administration of solution of 10% EtOH, 10% Tween 80, and 80% saline, while control (black column) represents i.t. injection of prostaglandin and s.c. administration of solution of 10% EtOH, 10% Tween 80, and 80% saline. ###Po0.001 by Student's t test compared with normal group. nnnPo0.001 by Student’s t test compared with control group.

Fig. 3. Analgesic effects of AS1069562 (A), duloxetine (B), and amitriptyline (C) on PGE2-induced allodynia in mice. Oral administration of AS1069562 (3, 10, and 30 mg/kg)

30 min prior (A), duloxetine (3, 10, and 30 mg/kg) 240 min prior (B), and amitriptyline (10, 30, and 100 mg/kg) 60 min prior (C) to i.t. injection of PGE2 (10 ng/animal).

AS1069562 (30 mg/kg) and duloxetine (10 and 30 mg/kg) signiﬁcantly ameliorated PGE2-induced allodynia, while amitriptyline at doses up to 100 mg/kg did not. Data (the mean7S.E.M. for 8 mice) are presented as percentage of the maximum cumulative score. Normal (white column) represents i.t. injection of saline and p.o. administration of distilled water, while control (black column) represents i.t. injection of prostaglandin and p.o. administration of distilled water. ###Po0.001 by Student's t test compared with normal group. nnPo0.01, nnnPo0.001 by Dunnett's test compared with control group.

efﬁcacies in prostaglandin-induced allodynia models over a 30 min of PGE2 and 300 ng of PGF2α, ﬁndings which are consistent with period. Tactile allodynia, which is induced by spinal hypersensitivity, previous reports (Minami et al., 1992, 1994a; Tsukamoto et al., 2010). A was obvious within 5 min of i.t. injection of PGE2 (10 and 100 ng) or time course (5, 10, 15, and 30 min) and i.t. dose for each prostaglandin PGF2α (100, 300 and 1000 ng), and sustained for 30 min (Fig. 1A). Both (PGE2, 10 ng/animal and PGF2α, 300 ng/animal) that produced a PGE2 and PGF2α induced allodynia in a dose-dependent manner submaximal allodynic response were selected for subsequent assess- (Fig. 1B). Near-maximum allodynic effects were observed with 10 ng ment of the effects of AS1069562, duloxetine, and amitriptyline. 58 N. Murai et al. / European Journal of Pharmacology 733 (2014) 54–61

Fig. 4. Analgesic effects of AS1069562 (A), duloxetine (B), and amitriptyline (C) on PGF2α-induced allodynia in mice. Oral administration of AS1069562 (3, 10, and 30 mg/kg)

30 min prior (A), duloxetine (3, 10, and 30 mg/kg) 240 min prior (B), and amitriptyline (10, 30, and 100 mg/kg) 60 min prior (C) to i.t. injection of PGF2α (300 ng/animal).

AS1069562 (30 mg/kg) and amitriptyline (30 and 100 mg/kg) signiﬁcantly ameliorated PGF2α-induced allodynia, while duloxetine at doses up to 30 mg/kg did not. Data (the mean7S.E.M. for 8 mice) are presented as percentage of the maximum cumulative score. Normal (white column) represents i.t. injection of saline and p.o. administration of distilled water, while control (black column) represents i.t. injection of prostaglandin and p.o. administration of distilled water. ###Po0.001 by Student's t test compared with normal group. nPo0.05, nnPo0.01, nnnPo0.001 by Dunnett’s test compared with control group.

Fig. 5. Effects of (S)–WAY100135 (A) and SR57227 (B) on antinociceptive effect of AS1069562 in PGF2α-induced allodynia in mice. Oral administration of AS1069562 (30 mg/ kg) 30 min prior to i.t. injection of PGF2α (300 ng/animal). Intrathecal injection of (S)–WAY100135 (5-HT1A receptor antagonist) (2 nmol/animal) (A) and SR57227 (5-HT3 receptor agonist) (2 nmol/animal) (B) 30 min prior to that of PGF2α. The analgesic effect of AS1069562 in PGF2α-induced allodynia was signiﬁcantly blocked by (S)– WAY100135 and by SR57227. Data (the mean7S.E.M. for 8 mice) are presented as percentage of the maximum cumulative score. Normal (white column) represents i.t. injection of saline and p.o. administration of distilled water, while control (black column) represents i.t. injection of prostaglandin and p.o. administration of distilled water. ###Po0.001 by Student’s t test compared with normal group. nnPo0.01, nnnPo0.001 by Student's t test compared with control group. †Po0.05 by Student's t test compared with AS1069562-treated group.

3.3. Analgesic effects of RTX on prostaglandin-induced allodynia in Table 1 Ki value of AS1069562 and duloxetine for 5-HT receptor subtypes. mice

ﬁ 5-HT receptor subtype 5-HT1A (μM) 5-HT1B (μM) 5-HT2B (μM) 5-HT3 (μM) To investigate the involvement of C- bers in prostaglandin- induced spinal hypersensitivity, we determined the effects of AS1069562 1.4 2.5 1.5 2.6 RTX on PGE2- and PGF2α-induced allodynia. Desensitization of Duloxetine 4.4 3.4 0.40 15 C-ﬁbers with systemic RTX (0.001 mg/kg s.c.) completely sup-

Values are expressed as the mean. pressed allodynia induced by PGF2α (Fig. 2B) but not PGE2 N. Murai et al. / European Journal of Pharmacology 733 (2014) 54–61 59

(Fig. 2A). These results demonstrate that the hyperexcitability of C- two (Minami et al., 1992, 1994a, 1994b; Muratani et al., 2003;

fibers plays a major role in triggering allodynia induced by PGF2α Tsukamoto et al., 2010). In addition, descending 5-HT and norepi- but not PGE2. nephrine systems presynaptically inhibit the glutamate release from the non-C- and C-afferent fibers at the spinal cord dorsal 3.4. Analgesic effects of AS1069562, duloxetine, and amitriptyline on horn (Yoshimura and Furue, 2006). We recently confirmed through prostaglandin-induced allodynia in mice microdialysis assay that AS1069562 and duloxetine elevated extra- cellular 5-HT and norepinephrine levels in rat spinal dorsal horn To investigate the spinal antinociceptive mechanism of AS10 (Murai et al., 2014). We therefore speculate that AS1069562 and 69562 compared with duloxetine and amitriptyline, we assessed duloxetine may exert analgesic effects primarily mediated at the the effects of AS1069562, duloxetine, and amitriptyline on tactile spinal level. To clarify the spinal antinociceptive mechanisms of allodynia in PGE2- and PGF2α-induced allodynia models. Analgesic AS1069562, duloxetine, and amitriptyline, which facilitate descend- effects on PGE2-induced allodynia are shown in Fig. 3, and those ing 5-HT and norepinephrine systems, we evaluated their effects in on PGF2α-induced allodynia in Fig. 4. AS1069562 significantly i.t. PGE2- and PGF2α-induced allodynia models. ameliorated PGE2-induced allodynia at a p.o. dose of 30 mg/kg We investigated the effects of RTX on PGE2- and PGF2α-induced fi (Fig. 3A), and duloxetine also significantly improved PGE2-induced allodynia to clarify the involvement of C- bers. Desensitization of allodynia at p.o. doses of 10 and 30 mg/kg (Fig. 3B). However, C-fibers with systemic RTX significantly suppressed hypersensi- amitriptyline did not significantly improve PGE2-induced allodynia tivity induced by PGF2α but not PGE2, suggesting that hyperexcit- fi even at p.o. doses up to 100 mg/kg (Fig. 3C). AS1069562 signifi- ability of C- bers plays a major role in triggering PGF2α- but not cantly ameliorated PGF2α-induced allodynia at a p.o. dose of PGE2-induced allodynia. PGE2- and PGF2α-induced allodynia 30 mg/kg (Fig. 4A), and amitriptyline also significantly ameliorated models are therefore useful when investigating which type of fi PGF2α-induced allodynia at p.o. doses of 30 and 100 mg/kg ber is involved in the analgesic effect of a drug. Findings from a (Fig. 4C). However, duloxetine did not significantly ameliorated study of capsaicin-treated neonatal mice suggested that capsaicin- β fi PGF2α-induced allodynia even at p.o. doses up to 30 mg/kg insensitive A - bers are involved in PGF2α-induced mechanical (Fig. 4B). These results demonstrate that AS1069562 inhibits both allodynia, while capsaicin-sensitive C-fibers are involved in

PGE2 and PGF2α-induced allodynia, while duloxetine only inhibits PGE2-induced allodynia (Minami et al., 1999). Whether or not non-C-fiber-triggered allodynia and amitriptyline only inhibits the capsaicin-sensitive fibers of neonatal mice destroyed by C-fiber-triggered allodynia. capsaicin treatment are restored at a later age remains To elucidate the mechanism by which AS1069562 induces unclear. However, streptozotocin-induced diabetes in neonatal broader effects on prostaglandin-induced allodynia than dulox- capsaicin-treated mice reportedly caused the reappearance of etine, we co-administered the 5-HT1A receptor antagonist (S)– intraplantar capsaicin-induced biting-licking behavior and WAY100135 and 5-HT3 receptor agonist SR57227 with AS1069562 increased capsaicin receptor-immunoreactive neurons in the dor- in the PGF2α-induced allodynia model. Saline was used as a vehicle sal root ganglia (Rashid et al., 2003). In addition, RTX treatment control for (S)–WAY100135 and SR57227. The analgesic effect of was reported to cause a depolarization block in the short term and

AS1069562 at 30 mg/kg on PGF2α-induced allodynia was signifi- ablation of transient receptor potential vanilloid 1 expressing cantly blocked by 75% following i.t. co-injection of (S)–WAY100135 nerve terminals in the long term (Jeffry et al., 2009). We therefore at 2 nmol (Fig. 5A) and by 66% following i.t. co-injection of administered RTX 24 h prior to prostaglandin injection to desen- SR57227 at 2 nmol (Fig. 5B). These results demonstrate that sitize C-fibers. fi agonism of the 5-HT1A receptor and antagonism of the 5-HT3 AS1069562 signi cantly alleviated both PGE2-andPGF2α-induced fi receptor are involved in the analgesic effects of AS1069562 in a allodynia, while duloxetine signi cantly alleviated only PGE2-induced fi mouse model of PGF2α-induced allodynia. allodynia and amitriptyline signi cantly ameliorated only PGF2α- induced allodynia. Taken together with findings for RTX efficacy

against PGE2-andPGF2α-induced allodynia, these results indicate that 4. Discussion AS1069562 inhibits both C-fiber- and non-C-fiber-triggered allodynia, while duloxetine inhibits only non-C-fiber-triggered allodynia and Prostaglandins are generated by cyclooxygenase (COX)-2 in the amitriptyline inhibits only C-fiber-triggered allodynia. AS1069562 is spinal cord and play an important role in the development and therefore suggested to have broader analgesic effects than duloxetine maintenance of hypersensitivity following nerve injury (Ma et al., and amitriptyline due to amelioration of both C-fiber- and non- 2002; Seybold et al., 2003). Expressions of COX-2 and EP1 receptor C-fiber-triggered allodynia. have been reported to increase in injured human nerves We recently found that AS1069562 significantly improved both (Durrenberger et al., 2006). COX-2 levels were also found to be mechanical allodynia and thermal hyperalgesia in a rat model of elevated in nerve fibers of patients with chronic pain (Durrenberger CCI-induced neuropathic pain (Murai et al., 2014). In that study, et al., 2006). In addition, in rat models of CCI and spinal nerve duloxetine only significantly ameliorated mechanical allodynia ligation-induced neuropathic pain, COX-2 expression was found to be and amitriptyline only thermal hyperalgesia. In addition, desensi- increased in both injured nerves and ipsilateral dorsal spinal cord tization of C-fibers with systemic RTX completely suppressed CCI- (Durrenberger et al., 2004; Zhao et al., 2000), and prostaglandins induced thermal hyperalgesia but only partially suppressed CCI- were produced and released in the spinal cord following peripheral induced mechanical allodynia (Murai et al., 2014). These previous nerve injury (Dirig and Yaksh, 1999). Further, PGE2 signaling via the findings suggest that the hyperexcitability of C-fibers plays a major EP1 receptor has been reported to contribute to human pain role in thermal hyperalgesia and is partially responsible for hypersensitivity (Sarkar et al., 2003). Taken together, these findings mechanical allodynia in CCI rats, and that AS1069562 inhibits suggest that prostaglandins may be involved in the development and hypersensitivity triggered by both C-fibers and non-C-fibers. maintenance of neuropathic pain in humans and rodents. We there- In contrast, duloxetine only inhibits non-C-fiber-triggered hyper- fore consider the prostaglandin-induced allodynia model to be sensitivity and amitriptyline only inhibits C-fiber triggered hyper- predictive of clinical efficacy, particularly for prostaglandin-involved sensitivity, findings are consistent with those of the present pain. Intrathecal administration of PGE2 and PGF2α induces tactile study. These findings suggest that AS1069562 may have an allodynia by facilitating glutamate release from presynaptic terminals antinociceptive mechanism distinct from that of duloxetine and at the spinal cord dorsal horn via different mechanisms between the amitriptyline. 60 N. Murai et al. / European Journal of Pharmacology 733 (2014) 54–61

The mechanism by which AS1069562, unlike duloxetine, inhi- gelatinosa neurons via activation of the α2-adrenergic receptor bits C-fiber-triggered hypersensitivity has not yet been elucidated. found in primary afferent terminals, with inhibition being more To determine whether or not other mechanisms contribute to the effective in non-C-fiber- than C-fiber-mediated transmission relatively broad range of analgesic effects AS1069562 compared to (Kawasaki et al., 2003). These reports suggest that inhibition of duloxetine, we investigated the affinity of AS1069562 and dulox- α2-adrenergic receptors might contribute to the distinct analgesic etine for 5-HT receptors. AS1069562 exhibited affinity for the mechanisms of AS1069562 compared to amitriptyline, as only the

5-HT1A, 5-HT1B, 5-HT2B, and 5-HT3 receptors with similar poten- latter has the potential to inhibit α2-adrenergic receptors, which cies regarding the binding of 5-HT receptors and inhibition of 5-HT contribute to analgesic effects by primarily inhibiting non-C-fiber- and norepinephrine uptake (0.21- to 3.1-fold difference) (Shimizu- triggered hypersensitivity. However, further investigation is Sasamata et al., 1993). In contrast, AS1069562 exhibited no needed prior to the adoption of this hypothesis. appreciable affinity for the 5-HT2A, 5-HT2C, 5-HT4, 5-HT5A, 5-HT6, In conclusion, the primary outcome of this study is that or 5-HT7 receptors (data not shown). Further, the potencies of AS1069562 inhibits both C-fiber- and non-C-fiber-triggered allo- duloxetine for binding 5-HT receptors and inhibiting 5-HT and dynia, while duloxetine inhibits only non-C-fiber-triggered allo- norepinephrine uptake differed substantially (57- to 5769-fold dynia and amitriptyline only C-fiber-triggered allodynia. The difference) (Wong et al., 1993). The activation of the spinal secondary outcome of this study is the suggestion that the

5-HT1A receptor and inhibition of the spinal 5-HT3 receptor are antinociceptive effects of AS1069562 may be attributable to not considered to produce analgesic effects. For example, the 5-HT1A only 5-HT and norepinephrine reuptake inhibition but also its receptor agonist F13640 decreased allodynic responses to tactile effects on spinal 5-HT1A and 5-HT3 receptors. Taken together, these and thermal stimulation in rats with spinal cord or sciatic nerve findings suggest that AS1069562 may have a broader spinal injury (Colpaert et al., 2002). Further, 5-HT3 receptor antagonist analgesic mechanism than that of duloxetine or amitriptyline. reduced the second phase of formalin-induced nociceptive behavior and inhibited thermal and mechanical evoked neuronal responses of wide dynamic range neurons in rats 14 days after Acknowledgments spinal nerve ligation (Okamoto et al., 2004; Suzuki et al., 2004). The authors would like to acknowledge Yoshihiro Keto, Shinji AS1069562 exhibits affinity for 5-HT1A and 5-HT3 receptors and near equal potency for 5-HT reuptake inhibition; and activation of Fukushima, and Kazuhiro Nakato for their expert assistance in the statistical analysis. the spinal 5-HT1A receptor and inhibition of the spinal 5-HT3 receptor have been reported to produce analgesic effects. We therefore investigated the effects of AS1069562 on 5-HT1A and References 5-HT3 receptors to determine whether or not activation of the spinal 5-HT1A receptor and inhibition of the spinal 5-HT3 receptor Ali, Z., Wu, G., Kozlov, A., Barasi, S., 1996. The role of 5HT3 in nociceptive processing contribute to the analgesic mechanisms of AS1069562 differing in the rat spinal cord: results from behavioral and electrophysiological studies. Neurosci. Lett. 208, 203–207. from those of duloxetine. The analgesic effect of AS1069562 on Asano, T., Dohi, S., Ohta, S., Shimonaka, H., Iida, H., 2000. Antinociception by PGF2α-induced allodynia was significantly blocked by i.t. injection epidural and systemic alpha (2)-adrenoceptor agonists and their binding affinity in rat spinal cord and brain. Anesth. Analg. 90, 400–407. of 5-HT1A receptor antagonist (S)–WAY100135 and 5-HT3 receptor Cheng, Y., Prusoff, W.H., 1973. Relationship between the inhibition constant (K1) agonist SR57227, suggesting involvement of agonistic action of the and the concentration of inhibitor which causes 50 per cent inhibition (I50) of spinal 5-HT1A receptor and antagonistic action of the spinal 5-HT3 an enzymatic reaction. Biochem. Pharmacol. 22, 3099–3108. receptor in the antinociceptive effects of AS1069562 in C-fiber- Colpaert, F.C., Tarayre, J.P., Koek, W., Pauwels, P.J., Bardin, L., Xu, X.J., Wiesenfeld- Hallin, Z., Cosi, C., Carilla-Durand, E., Assie´, M.B., Vacher, B, 2002. Large- triggered allodynia. The 5-HT1A and 5-HT3 receptors have been amplitude 5-HT1A receptor activation: a new mechanism of profound central reported to be involved in C-fiber hypersensitivity; for example, analgesia. Neuropharmacology 43, 945–958. 5-HT1A receptor knock-out mice and 5-HT1A receptor antagonist- Dirig, D.M., Yaksh, T.L., 1999. Spinal synthesis and release of prostanoids after fl – treated wild-type mice exhibited increased sensitivity in a hot- peripheral injury and in ammation. Adv. Exp. Med. Biol. 469, 401 408. Durrenberger, P.F., Facer, P., Gray, R.A., Chessell, I.P., Naylor, A., Bountra, C., Banati, R.B., plate test (Kayser et al., 2007), 5-HT3 receptor agonist-treated rats Birch, R., Anand, P., 2004. Cyclooxygenase-2 (Cox-2) in injured human nerve and a exhibited increased excitability of C-fiber axons in their sural rat model of nerve injury. J. Periphr. Nerv. Syst. 9, 15–25. nerve (Lang, et al., 2006), and 5-HT receptor antagonist-treated Durrenberger, P.F., Facer, P., Casula, M.A., Yiangou, Y., Gray, R.A., Chessell, I.P., Day, N.C., 3 Collins, S.D., Bingham, S., Wilson, A.W., Elliot, D., Birch, R., Anand, P., 2006. rats showed reduced neuronal nociceptive reactions to thermal Prostanoid receptor EP1 and Cox-2 in injured human nerves and a rat model of stimuli (Ali et al., 1996), all of which are consistent with our nerve injury: a time-course study. BMC Neurol. 6, 1. Fairbanks, C.A., Stone, L.S., Kitto, K.F., Nguyen, H.O., Posthumus, I.J., Wilcox, G.L., present results. An agonistic action of 5-HT1A receptor and an 2002. Alpha (2C)-adrenergic receptors mediate spinal analgesia and antagonistic action of 5-HT3 receptor might contribute to the adrenergic-opioid synergy. J. Pharmacol. Exp. Ther. 300, 282–290. different analgesic effects of AS1069562 and duloxetine, although Hylden, J.K.L., Wilcox, G.L., 1980. Intrathecal morphine in mice: a new technique. further investigation is required. Eur. J. Pharmacol. 67, 313–316. Jeffry, J.A., Yu, S.Q., Sikand, P., Parihar, A., Evans, M.S., Premkumar, L.S., 2009. The mechanism by which AS1069562, unlike amitriptyline, Selective targeting of TRPV1 expressing sensory nerve terminals in the spinal inhibits non-C-fiber-triggered hypersensitivity has not yet been cord for long lasting analgesia. PLoS One 4, e7021. elucidated. While amitriptyline at a concentration of 1 mM has Katoh, A., Eigyo, M., Ishibashi, C., Naitoh, Y., Takeuchi, M., Ibii, N., Ikeda, M., fi fi α Matsushita, A., 1995. Behavioral and electroencephalographic properties of been reported to have signi cant af nity for the 2-adrenergic duloxetine (LY248686), a reuptake inhibitor of norepinephrine and serotonin, receptor (98% inhibition) (Werling et al., 2007), AS1069562 at in mice and rats. J. Pharmacol. Exp. Ther. 272, 1067–1075. 10 mM had no appreciable affinity for α2-adrenergic receptors Kawasaki, Y., Kumamoto, E., Furue, H., Yoshimura, M., 2003. Alpha 2 adrenoceptor- (o50% inhibition) (data not shown). Given that the α2-adrenergic mediated presynaptic inhibition of primary afferent glutamatergic transmission in rat substantia gelatinosa neurons. Anesthesiology 98, 682–689. receptors are coupled to the Gi/o subunit and are expressed in the Kayser, V., Elfassi, I.E., Aubel, B., Melfort, M., Julius, D., Gingrich, J.A., Hamon, M., spinal cord dorsal horn (Olave and Maxwell, 2004), their activation Bourgoin, S., 2007. Mechanical, thermal and formalin-induced nociception is decreases transmitter release and exerts an inhibitory effect on the differentially altered in 5-HT1A-/-, 5-HT1B-/-, 5-HT2A-/-, 5-HT3A-/- and 5-HTT-/- knock-out male mice. Pain 130, 235–248. sensory transmission in the spinal cord. Further, i.t. activation of Lang, P.M., Moalem-Taylor, G., Tracey, D.J., Bostock, H., Grafe, P., 2006. Activity- α2-adrenergic receptors produces an analgesic action in animal dependent modulation of axonal excitability in unmyelinated peripheral rat models of chronic pain (Asano et al., 2000; Fairbanks et al., 2002). nerve fibers by the 5-HT(3) serotonin receptor. J. Neurophysiol. 96, 2963–2971. fi Ma, W., Du, W., Eisenach, J.C., 2002. Role for both spinal cord COX-1 and COX-2 in In an electrophysiological study, norepinephrine inhibited C- ber- maintenance of mechanical hypersensitivity following peripheral nerve injury. and non-C-fiber-mediated sensory transmission to substantia Brain Res. 937, 94–99. N. Murai et al. / European Journal of Pharmacology 733 (2014) 54–61 61

Minami, T., Uda, R., Horiguchi, S., Ito, S., Hyodo, M., Hayaishi, O., 1992. Allodynia Sarkar, S., Hobson, A.R., Hughes, A., Growcott, J., Woolf, C.J., Thompson, D.G., Aziz,

evoked by intrathecal administration of prostaglandin F2α to conscious mice. Q., 2003. The prostaglandin E2 Receptor-1 (EP-1) mediates acid-induced Pain 50, 223–229. visceral pain hypersensitivity in humans. Gastroenterology 124, 18–25. Minami, T., Uda, R., Horiguchi, S., Ito, S., Hyodo, M., Hayaishi, O., 1994a. Allodynia Seybold, V.S., Jia, Y.P., Abrahams, L.G., 2003. Cyclo-oxygenase-2 contributes to fl – evoked by intrathecal administration of prostaglandin E2 to conscious mice. central sensitization in rats with peripheral in ammation. Pain 105, 47 55. Pain 57, 217–223. Shimizu-Sasamata, M., Yamamoto, M., Harada, M., 1993. Central activating proper- Minami, T., Nishihara, I., Uda, R., Ito, S., Hyodo, M., Hayashi, O., 1994b. Character- ties of indeloxazine HCl and its optical isomers. Pharmacol. Biochem. Behav. 45, – ization of EP-receptor subtypes involved in allodynia and hyperalgesia induced 335 341. by intrathecal administration of prostaglandin E2 to mice. Br. J. Pharmacol. 112, Suzuki, R., Rygh, L.J., Dickenson, A.H., 2004. Bad news from the brain: descending 5- HT pathways that control spinal pain processing. Trends Pharmacol. Sci. 25, 735–740. 613–617. Minami, T., Okuda-Ashitaka, E., Hori, Y., Sakuma, S., Sugimoto, T., Sakimura, K., Thor, K., Kirby, M., Viktrup, L., 2007. Serotonin and noradrenaline involvement in Mishima, M, Ito, S., 1999. Involvement of primary afferent C-fibers in touch- urinary incontinence, depression and pain: scientific basis for overlapping evocked pain (allodynia) induced by prostaglandin E . Eur. J. Neurosci. 11, 2 clinical efficacy from a single drug, duloxetine. Int. J. Clin. Pract. 61, 1349–1355. – 1849 1856. Tsukamoto, M., Kiso, T., Shimoshige, Y., Aoki, T., Matsuoka, N., 2010. Spinal Murai, N., Aoki, T., Tamura, S., Sekizawa, T., Kakimoto, S., Tsukamoto, M., Oe, T., mechanism of standard analgesics: evaluation using mouse models of allody- þ Enomoto, R., Hamakawa, N., Matsuoka, N., 2014. AS1069562, the ( )-isomer of nia. Eur. J. Pharmacol. 634, 40–45. indeloxazine, exerts analgesic effects in a rat model of neuropathic pain with Werling, L.L., Keller, A., Frank, J.G., Nuwayhid, S.J., 2007. A comparison of the unique characteristics in spinal monoamine turnover. J. Pharmacol. Exp. Ther. binding profiles of dextromethorphan, memantine, fluoxetine and amitripty- 348, 372–382. line: treatment of involuntary emotional expression disorder. Exp. Neurol. 207, Muratani, T., Nishizawa, M., Matsumura, S., Mabuchi, T., Abe, K., Shimamoto, K., 248–257. Minami, T., Ito, S., 2003. Functional characterization of prostaglandin F2α Wong, D.T., Bymaster, F.P., Mayle, D.A., Reid, L.R., Krushinski, J.H., Robertson, D.W., receptor in the spinal cord for tactile pain (allodynia). J. Neurochem. 86, 1993. LY248686, a new inhibitor of serotonin and norepinephrine uptake. 374–382. Neuropsychopharmacology 8, 23–33. Okamoto, K., Imbe, H., Tashiro, A., Kumabe, S., Senda, E., 2004. Blockade of Yamaguchi, T., Ohyama, M., Suzuki, M., Ozawa, Y., Hatanaka, K., Hidaka, K.,

peripheral 5HT3 receptor attenuates the formalin-induced nocifensive behavior Yamamoto, M., 1998. Neurochemical and behavioral characterization of poten- in persistent temporomandibular joint inﬂammation of rat. Neurosci. Lett. 367, tial antidepressant properties of indeloxazine hydrochloride. Neuropharmacol- – 259–263. ogy 37, 1169 1176. Olave, M.J., Maxwell, D.J., 2004. Axon terminals possessing alpha 2C-adrenergic Yamamoto, M., 1990. Pharmacological and biochemical properties of indeloxazine – receptors densely innervate neurons in the rat lateral spinal nucleus, which hydrochloride, a new cerebral activator. Brain Dysfunct. 3, 130 147. respond to noxious stimulation. Neuroscience 126, 391–403. Yoshimura, M., Furue, H., 2006. Mechanisms for the anti-nociceptive actions of the Rashid, H., Inoue, M., Bakoshi, S., Ueda, H., 2003. Increased expression of vanilloid descending noradrenergic and serotonergic systems in the spinal cord. J. Pharmacol. Sci. 101, 107–117. receptor 1 on myelinated primary afferent neurons contributes to the anti- Zhao, Z., Chen, S.R., Eisenach, J.C., Busija, D.W., Pan, H.L., 2000. Spinal hyperalgesic effect of capsaicin cream in diabetic neuropathic pain in mice. J. cyclooxygenase-2 is involved in development of allodynia after nerve injury Pharmacol. Exp. Ther. 306, 709–717. in rats. Neuroscience 97, 743–748.