Zhiyong Wang, Weiya Ma, Jean-Guy Chabot, Remi Quirion * PAIN Ò 151 (2010) abstract

Transcription

1 PAIN Ò 151 (2010) Calcitonin gene-related peptide as a regulator of neuronal CaMKII CREB, microglial p38 NFjB and astroglial ERK Stat1/3 cascades mediating the development of tolerance to morphine-induced analgesia Zhiyong Wang, Weiya Ma, Jean-Guy Chabot, Remi Quirion * Douglas Mental Health University Institute, McGill University, Montreal, Quebec, Canada H4H 1R3 article info abstract Article history: Received 18 January 2010 Received in revised form 2 June 2010 Accepted 10 July 2010 Keywords: Calcitonin gene-related peptide Morphine tolerance Transcription factor N-methyl-D-aspartate receptor Microglia Astrocyte Tolerance to morphine-induced analgesia is an intractable phenomenon, often hindering its prolonged applications in the clinics. The enhanced pronociceptive actions of spinal pain-related molecules such as calcitonin gene-related peptide (CGRP) may underlie this phenomenon and could be a promising target for intervention. We demonstrate here how CGRP regulates the development of morphine analgesic tolerance at the spinal level. A 7-day treatment with morphine led to tolerance to its analgesic effects and enhanced expression of CGRP and its receptor subunits calcitonin receptor-like receptor (CRLR) and receptor activity modifying protein 1 (RAMP1). Activation of several cell-type-specific kinase transcription factor cascades is required to mediate this tolerance, including calcium/calmodulin-dependent protein kinase II (CaMKII) and camp response element-binding protein (CREB) in neurons, p38 and nuclear factor kappa B (NFjB) in microglia and extracellular signal-regulated protein kinase (ERK) and signal transducer and activator of transcription 1 and 3 (Stat1/3) in astrocytes, because inhibitors of CaMKII, p38 and ERK pathways correspondingly reduced the increases in phosphorylated CREB, acetylated-nfjb and phosphorylated Stat1/3 levels and attenuated the development of tolerance. Interestingly, these cascades were linked to the regulation of glutamatergic N-methyl-D-aspartate (NMDA) receptor expression. Chronic morphine-induced behavioural responses and biochemical events were all subjugated to modulation by disrupting CGRP receptor signaling. Together, these data suggest that CGRP contributes to the development of tolerance to morphine-induced analgesia by regulating the activation of the neuronal CaMKII CREB, microglial p38 NFjB and astroglial ERK Stat1/3 cascades. Targeting CGRP-associated signaling molecules may prolong or restore morphine s analgesic properties upon a chronic exposure. Crown Copyright Ó 2010 Published by Elsevier B.V. on behalf of International Association for the Study of Pain. All rights reserved. 1. Introduction Although opiates including morphine are the most widely used drugs in the management of chronic pain induced by injury and inflammation, the development of tolerance to its analgesic properties hinders its prolonged safe use [29]. The molecular and cellular mechanisms responsible for this phenomenon remain elusive and may involve mu-opioid receptor desensitization [3], mu-delta opioid receptor interaction [70] and the activation of protein kinase A and C [39] and mitogen-activated protein kinases (MAPKs) [5]. Decrease in morphine analgesia seen during its prolonged exposure may also be related to the relative increase in the * Corresponding author. Address: Douglas Mental Health University Institute, 6875 Boulevard LaSalle, McGill University, Montreal, Quebec, Canada H4H 1R3. Tel.: x2934; fax: addresses: remi.quirion@douglas.mcgill.ca, quirem@douglas.mcgill.ca (R. Quirion). sensitivity of spinal nociceptive neurons, which may emanate from enhanced signaling efficacy of pain-related substances such as calcitonin gene-related peptide (CGRP) and glutamate. Indeed, CGRP-immunoreactive materials and glutamate receptors are upregulated at the spinal level following chronic morphine while blockade of CGRP or glutamate receptor signaling attenuates the development of tolerance [34,41,59]. Here we chose to investigate the intracellular signal transduction cascades underlying the role of CGRP in the development of tolerance to morphine-induced analgesia. The neuropeptide CGRP is widely distributed in the peripheral and central nervous system [25]. As a G-protein coupled receptor, functional CGRP receptors are composed of the 7-transmembrane protein CRLR and two associated components RCP and RAMP1. RAMP1 is required for CGRP binding to CRLR and determines receptor ligand profile [4]. RCP facilitates camp formation by interacting with CRLR [17]. Upon stimulation of CGRP receptors, the adenylate cyclase (AC) camp PKA cascade is usually activated as /$36.00 Crown Copyright Ó 2010 Published by Elsevier B.V. on behalf of International Association for the Study of Pain. All rights reserved. doi: /j.pain

2 Z. Wang et al. / PAIN Ò 151 (2010) well as other pathways under certain conditions [4,50]. CGRP is an important pain mediator in spinal nociceptive processing [10,69] and an interaction between CGRP and opiates has been suggested [48,65]. Protein kinases including ERK and p38 MAPKs as well as CaM kinases are important signal transducers involved in morphine s analgesia [5,6,11,18]. Additionally, the stimulation of glia and the subsequent up-regulation of proinflammatory cytokines such as interleukin-1b (IL-1b) and IL-6 have been implicated in the development of tolerance [27,53,56]. We have previously shown a strong association between MAPK activation and the up-regulation of glia-derived cytokines in the development of tolerance, CGRP being an upstream mediator [62]. However, little information is available about the intracellular events occurring between the activation of these kinases and downstream target protein expression at the spinal level. Analyses of transcription factors such as camp response element-binding protein (CREB), nuclear factor kappa B (NFjB) and signal transducer and activator of transcription (Stat) have revealed that these molecules are involved in various opiates-induced effects [7,34,61] and thus could be promising candidates between those aforementioned kinase activation and target protein up-regulation (such as cytokines). The present study was undertaken to investigate the possible cascade association between these kinases and transcription factors occurring in the development of morphine analgesic tolerance and whether those cascades are under modulation by CGRP receptor signaling at the spinal level. In addition, the glutamatergic NMDA receptor is a key mediator in morphine tolerance [38,59], the possible modulation of NMDA receptor expression by CGRP receptor signaling was also explored. 2. Materials and methods 2.1. Experimental animals Adult male Sprague Dawley rats weighing g were obtained from Charles River Breeding Laboratories (St. Constant, QC, Canada). Rats were individually housed in cages on a 12 h alternating light dark cycle (lights on at 7:00 AM) with food pellets and water available ad libitum. All experimental protocols were approved by the Animal Care and Use Committee of McGill University and Canadian Council on Animal Care and were in accordance with the guidelines of the International Association for the Study of Pain concerning the use of laboratory animals Intrathecal catheter implantation and drug administration Animals were intrathecally implanted with a polyethylene (PE)- 10 catheter through the gap between vertebrae L5 and L6 and extended to the subarachnoid space under anaesthesia with a cocktail of ketamine (10 mg/kg; Bioniche, Belleville, ON, Canada), acepromazine (1 mg/kg; Ayerst, Guelph, ON, Canada) and xylazine (5 mg/kg, Novopharm, Toronto, ON, Canada) according to a previously described method [62]. Animals were allowed to recover for 7 9 days and habituated to the test environment at least 3 times before behavioural testing. Rats displaying signs of motor dysfunction (hind limb paralysis) were excluded from the experiments. Drugs were delivered through the exterior end of the catheter in a total volume of 10 ll followed by a flush of 15 ll saline. The doses used here were selected on the basis of previous and our pilot studies [11,49,62,72]. The following drugs were purchased from Sigma Aldrich (St. Louis, MO, USA): morphine sulfate (1.5 lg/ll in saline), 4-(4-fluorophenyl)-2-(4-methylsulfinylphenyl)-5-(4-pyridyl)1H-imidazole (SB203580; 1 lg/ll in 20% DMSO) and 2 0 -amino-3 0 -methoxyflavone (PD98059; 1 lg/ll in 20% DMSO). 2-[N-(2-hydroxyethyl)]-N-(4-methoxybenzenesulfonyl)]amino-N-(4-chlorocinnamyl)-N-methylbenzylamine), Phosphate (KN93, 1.5 mm in 20% DMSO) and 2-[N-(4-methoxybenzenesulfonyl)]amino-N-(4-chlorocinnamyl)-N-methylbenzylamine, Phosphate (KN92, 1.5 mm in 20% DMSO) were from Calbiochem (San Diego, CA, USA). [r-(r*,s*)]-n-[2-5-amino-1-4-(4-pyridinyl)- 1-piperazinyl]pentyl]amino]-1-[(3,5-dihydro-2-oxo-3(2H)-quinazolinyl)-1-piperidinecarboxamide (BIBN4096BS, lg/ll in 50% DMSO) was kindly provided by Dr. Henri Doods (Boehringer Ingelheim Pharma KG, Biberach, Germany) Induction of morphine tolerance and behavioural testing A daily intrathecal delivery of 15 lg morphine for 7 days was used to induce tolerance to the analgesic effects of morphine, as described previously [27, 41, 62]. Either tail-flick or paw-withdrawal test was used to evaluate the analgesic potential of morphine or morphine in combination with other agents. For testing tail-flick response, animals received radiant thermal stimulus at the tail position as described previously [41,62]. The intensity of thermal stimulus was adjusted to produce a baseline response latency of 2 3 s and a cut-off time of 10 s was set to avoid any tissue damage. For paw-withdrawal latency test, a radiant heat beam was focused on the plantar skin of the hindpaw as previously described [23], with thermal stimulation intensity adjusted to produce a baseline latency of s. A cut-off time of 25 s was used to prevent tissue damage. The response latency was defined as the time from the initiation of the radiant heat to tail or paw withdrawal and measured automatically by a digital meter. Behavioural testing was performed before and 30 min after drug treatment every other day starting from day 1. Changes in response latency in both tests were standardized by expressing values as percentage of maximal possible effect (%M.P.E.) using the following formula: [(post-drug latency-baseline latency)/(cut-off time-baseline latency)] 100. On the eighth day, ascending doses of morphine were given every 30 min until a maximal antinociceptive level was achieved in tailflick test. Cumulative dose response curve for the antinociceptive action of acute intrathecal morphine in animals receiving chronic intrathecal infusions was constructed accordingly Immunohistochemical studies Animals were perfused through the ascending aorta with ml saline followed by 300 ml 4% paraformaldehyde in 0.1 M phosphate buffer (ph 7.4). L4 L5 spinal cord segments were removed, postfixed in the same fixative for 4 h at 4 C and cryoprotected in 30% sucrose overnight. Transverse 35-lm thick sections were cut on a cryostat and processed for immunohistochemistry as previously described [64]. The specificity of antibodies used was checked by western blotting and/or omission of the primary antibodies Single immunostaining for CGRP, p-creb, NFjB and p-stat3 For CGRP immunostaining, free-floating sections were treated with 0.3% H 2 O 2 in 0.01 M PBS for 15 min to abolish endogenous peroxidase activity. Sections were then rinsed in PBS, blocked with 10% goat serum in 0.3% Triton X-100 for 1 h at room temperature (RT) and incubated overnight at 4 C with rabbit anti-cgrp antibody (1:2000, Peninsula Labs, San Carlos, CA, USA). Thereafter, sections were processed with biotinylated goat anti-rabbit IgG (1:200) for 1 h followed by incubation with avidin biotin peroxidase complex (ABC) for 30 min using Vecastain ABC kit (Vector, Burlingame, CA, USA). The immunoprecipitates were developed in 0.05% diaminobenzidine with 0.01% hydrogen peroxide. The chromogen was intensified by the addition of 0.02% nickel ammonium sulphate. Immunofluorescent labelling was performed using a primary

3 196 Z. Wang et al. / PAIN Ò 151 (2010) rabbit anti-p-creb (1:100; Catalogue No. 9191, Cell Signaling Technology, Danvers, MA, USA), rabbit anti-nfjb p65 (1:200; Santa Cruz Biotechnology, Santa Cruz, CA, USA) or rabbit anti-p-stat3 antibody (1:100; Catalogue No. 9134, Cell Signaling Technology). Staining was visualized using Alexa 568-conjugated secondary antibody Double labelling procedures For double immunofluorescence, sections were incubated with a mixture of two primary antibodies followed by a mixture of Alexa 568- and Alexa 488-conjugated secondary antibodies. Specifically, to identify cell types expressing p-creb (1:100, Cell Signaling Technology), NFjB (1:200, Santa Cruz Biotechnology), p-stat3 (1:100, Cell Signaling Technology) and p-camkii (1:100, Catalogue No. 3361, Cell Signaling Technology), each of these four antibodies was mixed with mouse anti-glial fibrillary acidic protein (anti- GFAP, astrocyte marker, 1:1000, Millipore, Billerica, MA, USA), mouse anti-ox-42 (microglial marker, 1:100, AbD Serotech, Raleigh, NC, USA) or mouse anti-neun (neuronal marker, 1:1000, Millipore). To determine if p-erk (rabbit, 1:100, Catalogue No. 9101, Cell Signaling Technology), p-p38 (rabbit, 1:100, Catalogue No. 9211, Cell Signaling Technology) and p-jnk (rabbit, 1:100, Catalogue No. 9251, Cell Signaling Technology) are expressed in neurons, antibodies for these proteins were mixed with neuronal marker MAP-2 (mouse, 1:1000, Millipore) or NeuN. To substantiate whether p-camkii and p-creb, p-p38 and NFjB as well as p-erk (Catalogue No. 9106, Cell Signaling Technology) and p-stat3 are co-localized in the same cells, sections were incubated with a mixture of corresponding antibodies. Images were captured using a confocal microscopy system (PCM2000; Nikon, Tokyo, Japan) Western blotting Animals were rapidly perfused through the ascending aorta with ml saline under anaesthesia, and the dorsal horn of the L4 L6 spinal cord segments and corresponding L4 L6 dorsal root ganglions (DRGs) were removed, frozen on dry ice and stored at 80 C. Tissues were homogenized in modified RIPA buffer (50 mm Tris HCl, 150 mm NaCl, 1 mm EDTA, 1% Igepal CA-630, 0.1% SDS, 50 mm NaF, and 1 mm NaVO 3 ) combined with a mixture of proteinase inhibitors (all from Sigma Aldrich) including PMSF (2 mm), leupeptin (10 lg/ml) and aprotinin (10 lg/ml). The protein concentration of tissue lysates was determined using a BCA Protein Assay Kit (Pierce Biotechnology, Rockford, IL, USA). Equal amounts of protein samples (20 lg/lane) were loaded and separated on SDS PAGE gels (4 20%; Invitrogen, Carlsbad, CA, USA), and electroblotted onto Hybond-C nitrocellulose membranes (Amersham Biosciences, Piscataway, NJ, USA). Membranes were incubated overnight at 4 C with various indicated primary antibodies following incubation with 5% skim milk to block non-specific binding sites. These membranes were further incubated with HRP-conjugated secondary antibody (1:1000, Santa Cruz Biotechnology) for 1 h at room temperature (RT), developed in ECL solution for 1 5 min, and exposed onto X-films for 1 5 min. Antibodies used here including goat anti-crlr (1:1000), rabbit anti-ramp1 (1:800), goat anti-nr1 (1:800), goat anti-nr2a (1:800), goat anti-nr2b (1:800), rabbit anti-nfjb p65 (1:1000) and rabbit anti-ijb (1:1000) were from Santa Cruz Biotechnology. Antibodies including rabbit anti-p-creb (Ser 133, Catalogue No. 9191), rabbit anti-p-camkii (Thr 286, Catalogue No. 3361), rabbit anti-p-nfjb p65 (Ser 536, Catalogue No. 3033), rabbit anti-p-p38 (Catalogue No. 9211), rabbit anti-p-ijb (Ser 32/36, Catalogue No. 9246), rabbit anti-acetyl-nfjb p65 (Lys 310, Catalogue No. 3045), rabbit anti-p-erk (Catalogue No. 9101), rabbit anti-p-stat3 (ser 727, Catalogue No. 9134), rabbit anti-p-stat3 (Tyr 705, Catalogue No. 9145), rabbit anti-p-stat1 (ser 727, Catalogue No. 9177), rabbit anti-p-stat1 (Tyr 701, Catalogue No. 9171), rabbit anti-p-stat2 (Tyr 690, Catalogue No. 4441), rabbit anti-p-stat5 (Tyr 694, Catalogue No. 9351) and rabbit anti-p-stat6 (Tyr 641, Catalogue No. 9361) were from Cell Signaling Technology and used at a 1:1000 dilution. The rabbit for loading controls, membranes were incubated with a stripping buffer and reprobed using antibodies to b-actin (1:3000) from Santa Cruz Biotechnology, CaMKII (1:1000) from Millipore and CREB (Catalogue No. 9197), ERK (Catalogue No. 4695), p38 (Catalogue No. 9212) and Stat3 (Catalogue No. 9132) from Cell Signaling Technology at a 1:1000 dilution. Specific bands were captured using a Sierra Scientific MS-4030 video camera (Sierra Scientific, Sunnyvale, CA, USA) and quantified by densitometric analysis using MCID version 6 Elite imaging software (Imaging Research, St. Catharines, ON, Canada) and normalized with their corresponding internal loading control bands Statistics Results are expressed as means ± SEM. Statistical significance was analyzed for behavioural data using a two-way repeated ANO- VA followed by a post hoc student Newman Keuls test or Tukey s test to determine differences between groups. Results from the Western blot were compared with those of a two-way ANOVA, where each of the four gels included all of the related treatments and thus gels and the treatments were considered as independent variables and the relative density as dependent variable. ED 50 values were determined using a non-linear regression analysis followed by one-way ANOVA with a student Newman Keuls post hoc test for multiple comparisons between groups. p < 0.05 was considered significant. For simplicity and clarity, all statistics and group size information are presented in the figure legends. 3. Results 3.1. CGRP signaling in the development of tolerance to morphineinduced analgesia We and others have previously shown that chronic morphine treatment increases CGRP contents in the spinal cord dorsal horn using immunohistochemistry (Supplementary Fig. 1) and ELISA [20,41,49,62]; the two methods do not differentiate between the various precursors and mature forms of CGRP. To further identify the role of CGRP in the development of morphine tolerance, we examined first changes in CGRP and CGRP receptors by western blotting following repeated exposure to morphine. A 7-day intrathecal injection of morphine (15 lg/day) induced a significant increase in the immediate precursor (13 kda) of mature CGRP, an effect inhibited by the co-administration of the non-peptide CGRP receptor antagonist, BIBN4096BS (0.025 and 0.1 lg/day; Fig. 1A and C). In contrast, the fully mature form of CGRP (5 kda) was undetectable. The two subunits of the CGRP receptor complex, CRLR and RAMP1, were also up-regulated by the 7-day morphine treatment (Fig. 1A and C). An acute, single dose of morphine (15 lg) did not alter the expression of CGRP and its two receptor components in the spinal dorsal horn (Fig. 1A and C). It has been reported that the DRG is the main source of CGRP in the lumbar spinal dorsal horn [9]. In that context, we also observed an increase in CGRP and RAMP1, but not CRLR in the DRG following the 7-day treatment with morphine, an effect inhibited by BIBN4096BS (Fig. 1A and B). We investigated next the effects of interrupting CGRP receptor signaling on the development of tolerance to morphine-induced analgesia using two nociceptive tests, namely the tail-flick and paw-withdrawal tests. As shown in Fig. 2, morphine-induced analgesia was gradually decreased in both models over the 7-day treatment, indicative of the development of

4 Z. Wang et al. / PAIN Ò 151 (2010) Fig. 1. Chronic morphine-induced up-regulation of CGRP, CRLR and RAMP1 is inhibited by the co-administration of a CGRP receptor antagonist, BIBN4096BS. (A and B) Western blot analyses of CGRP and its receptor components CRLR and RAMP1 in DRG (Left) as well as in lumbar spinal dorsal horn (Right and C) 30 min after daily intrathecal injection of saline (NS), DMSO (20%), morphine (M; 15 lg) or morphine with BIBN4096BS (M/B; and 0.1 lg) for a 1- or 7-day treatment. b-actin (Actin) is used as a loading control. Data are mean ± SEM of 4 rats. *p < 0.05 vs saline group at day 7. **p < 0.01 vs saline group at day 7. # p < 0.05 vs morphine group at day 7. ## p < 0.01 vs morphine group at day 7. tolerance. A treatment with BIBN4096BS significantly blocked the development of tolerance. Furthermore, the acute morphine ED 50 values were largely increased in animals receiving a 7-day treatment of morphine (15 lg) compared to saline group (Table 1), suggesting a loss of morphine potency following chronic morphine treatment. However, this increase was significantly prevented by a co-treatment with BIBN4096BS (0.1 lg) (Table 1), implying that this co-treatment can prevent the loss of morphine potency associated with the development of tolerance. These data suggest an association of CGRP signaling with the development of tolerance to morphine-induced analgesia Evidence that the neuronal CaMKII CREB cascade mediates CGRPassociated development of tolerance to morphine-induced analgesia We examined next the neuronal intracellular signaling cascade(s) possibly involved in CGRP-associated development of morphine s tolerance. Both CaM kinase and CREB have been implicated in chronic morphine-induced effects [34,63,67]. As shown in Fig. 3, a 7-day treatment with morphine (15 lg/day) increased the phosphorylation/activation of CaMKII and CREB. BIBN4096BS (0.025 and 0.1 lg; Fig. 3A and B) blocked these effects (also see Supplementary Fig. 10). An acute, single dose of intrathecal morphine (15 lg) only inhibited p-camkii levels without affecting p-creb levels (Fig. 3A and B). We then asked whether CaMKII is an upstream kinase candidate mediating CREB activation following a 7-day morphine treatment. A co-localization study showed that activated CREB (p-creb) and CaMKII (p-camkii) were both localized in neurons (Fig. 4Ad f and Supplementary Fig. 2) and were co-localized together (Fig. 4B). A treatment with KN93 (1.5 and 15 nmol), a CaMKII inhibitor, dose-dependently inhibited chronic morphine-induced increase in p-creb levels (Fig. 4C). The co-treatment of morphine with KN93 (15 nmol), but not its inactive form KN92, attenuated the development of tolerance (Fig. 4D) as well as prevented the increase of acute morphine ED 50 values induced by chronic morphine treatment (Table 1). These data suggest that the activation of the CaMKII CREB cascade occurring in neurons mediates the CGRP-associated development of tolerance. CaMKIV has also been proposed to be involved in the development of morphine s tolerance [31]. However, our chronic morphine treatment did not increase p-camkiv level (data not shown) even if it was localized in p-creb-expressing cells (Fig. 4B). MAPKs including ERK, p38 and JNK may also be implicated in this phenomenon [5,11,62] and are other upstream candidates involved in the activation of CREB. However, we did not observe any colocalization of these three activated kinases with neuronal markers following chronic morphine treatment (Supplementary Fig. 3) and inhibitors of these kinases did not alter chronic morphine-induced increase in p-creb levels (Supplementary Fig. 4) Evidence that the microglial p38 NFjB cascade mediates CGRPassociated development of tolerance to morphine-induced analgesia Activated p38 MAPK in microglia has been suggested to be involved in the development of tolerance to morphine-induced analgesia [11,62]. However, limited information is currently available on downstream target(s) of p38 activation in CGRP-associated morphine tolerance. NFjB is a promising candidate in view of its association with p38 in various in vitro systems [26,68]. We demonstrate here that the increase in NFjB levels induced by repeated treatment with morphine (15 lg/day for 7 days) was inhibited by a co-treatment with a p38 inhibitor, SB (10 lg; Fig. 5B). In accordance with our previous results [62], chronic morphine treatment enhanced p38 activation/phosphorylation (p-p38; Fig. 5A) and the inhibition of p38 activity also attenuated the development of tolerance in the paw-withdrawal test (Supplementary Fig. 5) as well as prevented chronic morphine-induced increase in the acute morphine ED 50 values (Table 1). Chronic morphine-induced increase in p-p38 and NFjB was blocked by interrupting CGRP receptor signaling using BIBN4096BS (Fig. 5A and B; also see Supplementary Fig. 10). A co-treatment with SB produced no effects on chronic morphine-induced increase in the phosphorylation of p38 (Fig. 5A), consistent with the fact that SB is an inhibitor of p38 activity, but not p38 activation. An acute treatment with a single dose of morphine on day 1 (15 lg) did not affect p-p38 or NFjB (Fig. 5A and B). A confocal study showed that

5 198 Z. Wang et al. / PAIN Ò 151 (2010) Fig. 2. Inhibition of the development of morphine analgesic tolerance by the coadministration of BIBN4096BS. Either tail-flick (A) or paw-withdrawal latency is tested 30 min after injection every other day from day 1 following a daily intrathecal delivery of saline (NS), DMSO (20%), BIBN4096BS only (B; 0.1 lg), morphine (M; 15 lg) or morphine with BIBN4096BS (M/B; and 0.1 lg) for 7 days. Morphine and BIBN4096BS are mixed in equal volume before injection. Data are mean ± SEM of 6 8 animals. *p < 0.01 vs morphine group at identical time point. Fig. 3. Chronic morphine-induced increases in p-camkii and p-creb are reversed by co-administration of BIBN4096BS. (A and B) Western blot analyses reveal that a 7-day treatment with morphine (M; 15 lg/day) increases p-camkii (A) and p-creb (B) levels in the lumbar spinal dorsal horn, effects inhibited by a co-treatment with BIBN4096BS (M/B; and 0.1 lg). NS, saline. CaMKII or CREB levels are used as loading controls. Data represent mean ± SEM of 4 rats. *p < 0.05 vs saline group at identical time point. **p < 0.01 vs saline group at day 7. # p < 0.05 vs morphine group at day 7. ## p < 0.01 vs morphine group at day 7. Table 1 Effect of various indicated inhibitors on the development of tolerance to morphineinduced analgesia. Group ED 50 (lg) 95% CI (lg) Saline 3.6 * BIBN4096BS (0.1 lg) 3.9 * KN93 (15 nmol) 4.2 * SB (10 lg) 4.3 * PD98059 (10 lg) 2.5 * Morphine Morphine/BIBN4096BS (0.1 lg) 4.5 * Morphine/KN93 (15 nmol) 6.9 * Morphine/SB (10 lg) 5.5 * Morphine/PD98059 (10 lg) 4.7 * Morphine/KN92 (15 nmol) Data represent ED50 values with the 95% confidence interval (CI) for the antinociceptive action of acute intrathecal morphine in animals on day 8 following a 7-day treatment with morphine (15 lg) alone or in combination with indicated inhibitors (KN93, a CaMKII inhibitor; SB203580, a p38 inhibitor; PD98059, an ERK activation inhibitor; KN92, an inactive form of KN93). * p < 0.01 compared to morphine group (n = 6 per group). OX-42-positive microglial cells expressed NFjB (Fig. 6A) and that both NFjB and p-p38 were co-localized in the same cells (Fig. 6B). We also examined the phosphorylation of cytosol inhibitor jb (p-ijb) and acetylation of NFjB (acetylated-nfjb), an important modification in the nucleus which enhances its transactivation potency [42], following chronic morphine treatment. As shown in Fig. 7, a 7-day treatment with morphine (15 lg) induced a pronounced increase in phosphorylation levels of IjB and acetylation levels of NFjB. The p38 inhibitor SB (10 lg) or CGRP receptor antagonist BIBN4096BS (0.025 or 0.1 lg) significantly blocked these increases. Interestingly, we detected no change in IjB and a reduction of p-nfjb at serine 536 or serine 276 (data not shown), another post-translational modification of NFjB p65 subunit occurring in the cytoplasm to enhance its nuclear translocation [47], following repeated treatment with morphine (Supplementary Fig. 6). It has been reported that nuclear p65 can be rapidly dephosphorylated at ser536, suggesting the necessity for NFjB to return to the cytoplasm for reactivation [47]. This could explain a decrease observed in p65 phosphorylation in the present study. In addition, NFjB activation requires IjB dissociation from its complex and subsequently IjB undergoes degradation. No alteration of IjB levels in our study could be due to its rapid synthesis, thereby compensating its possible loss following degradation. However, further experiments are needed to clarify this.

6 Z. Wang et al. / PAIN Ò 151 (2010) Fig. 4. Immunohistochemistry shows p-creb-ir cells (red) in the spinal dorsal horn (Aa). Double immunofluorescence shows that p-creb (red) is co-localized with NeuN (green; arrows, Ad f), but not with GFAP (green; Ab) or OX-42 (green; Ac). p-creb-ir cells (B) also express p-camkii or p-camkiv. Scale bars, 50 lm (Aa); 20 lm (Ab f); 10 lm (B). (C) Reversal of chronic morphine-induced increases in p-creb levels by co-treatment with KN93, a CaMKII inhibitor. Western blot analyses of p-creb levels in the lumbar spinal dorsal horn 30 min after daily intrathecal injection of saline (NS), DMSO (20%), KN93 alone (15 nmol), morphine alone (15 lg) or morphine with KN93 (M/ KN93; 1.5 and 15 nmol) for 7 days. The CREB level is used as a loading control. Data are mean ± SEM of 4 rats. *p < 0.01 vs saline group. # p < 0.01 vs morphine group. (D) Effects of KN93 on the development of morphine analgesic tolerance. Paw-withdrawal latency is tested 30 min after injection every other day from day 1 following a daily intrathecal delivery of KN93 alone (15 nmol), morphine alone (M; 15 lg), morphine with KN93 (1.5 and 15 nmol) or morphine with KN92 (an inactive form of KN93; 15 nmol). KN93 and KN92 are administered 30 min before morphine injection. Data are mean ± SEM of 6 8 animals. *p < 0.01 vs morphine group at identical time point Evidence that the astroglial ERK Stat3 cascade is involved in CGRP-associated development of tolerance to morphine-induced analgesia The activation of ERK in astroglia has been linked to CGRP-associated development of tolerance to morphine-induced analgesia [62]. Downstream targets from ERK activation remain to be established. Since we have excluded the role of CREB as a substrate of ERK in the development of morphine tolerance (Supplementary Figs. 3 and 4), we postulate the signal transducer and activator of transcription (Stat) as a possible target [8,12,66]. Stat3 has been suggested to be involved in the pathogenesis of neuropathic pain, which has shown common mechanisms with morphine tolerance [15,39]. We thus examined a possible link between ERK and Stat3. As shown in Fig. 8, an acute, single dose of morphine (15 lg) produced no effects on p-erk levels, but reduced p-stat3(ser) levels when compared with the saline group. However, a chronic treatment with morphine (15 lg/day for 7 days) increased both p-erk and p-stat3(ser) levels (also see Supplementary Fig. 10). PD98059 (10 lg), a MEK (ERK upstream kinase) inhibitor, blocked these effects as well as the development of morphine tolerance (Supplementary Fig. 5) and chronic morphine-induced increase in the acute morphine ED 50 values (Table 1). A co-localization study of p-stat3 with OX-42, a microglial marker, NeuN, a neuronal marker, and GFAP, an astroglial marker, showed that p-stat3 was predominantly enriched in astrocytes (Fig. 9Ad f). p-stat3 and p-erk were expressed in the same cells (Fig. 9Ba c). Moreover, chronic morphine-induced increase in p-erk and p-stat3 levels was also inhibited by blockade of CGRP receptors using BIBN4096BS (0.025 and 0.1 lg; Fig. 8A and B), suggesting a role for CGRP receptor signaling in modulating the ERK Stat3 cascade during the development of tolerance to morphine-induced analgesia. In addition to the serine(ser) phosphorylation, we also investigated the tyrosine(tyr) phosphorylation of Stat3. Interestingly, the tyrosine phosphorylation of Stat3 was reduced following a repeated treatment with morphine (Supplementary Fig. 7). This inhibitory effect was reversed by the co-administration of PD98059, suggesting the modulation of tyrosine phosphorylation by ERK activation. Similarly, BIBN4096BS blocked this effect of morphine on tyrosine phosphorylation of Stat3 (Supplementary Fig. 7). Additionally, the serine phosphorylation of another Stat, Stat1, was increased and the tyrosine phosphorylation was decreased following chronic morphine treatment (Supplementary Figs. 8 10). Interestingly, the serine phosphorylation of Stat1, but not its tyrosine phosphorylation, was modulated by PD98059, suggesting an ERK-dependent effect on serine phosphorylation of

7 200 Z. Wang et al. / PAIN Ò 151 (2010) Fig. 6. Distribution and cellular localization of NFjB. Immunohistochemistry shows NFjB p65-ir cells (NFjB; red) in the spinal dorsal horn (Aa). Double immunofluorescence reveals that NFjB (red) is co-localized with OX-42(green; Ad f), but not with GFAP (green; Ab) or NeuN (green; Ac). NFjB-ir cells (red) also express p-p38 (green; Ba c). Scale bar, 50 lm (Aa); 20 lm (Ab f and Ba c). Fig. 5. Chronic morphine-induced increases in p-p38 and NFjB are reversed by the co-administration of BIBN4096BS. (A and B) Western blot analyses reveal that a 7- day treatment with morphine (M; 15 lg/day) increases p-p38 (A) and NFjB p65 (NFjB; B) levels in the lumbar spinal dorsal horn, effects being inhibited by a cotreatment with BIBN4096BS (M/B; and 0.1 lg). The inhibition of p38 activity by SB (SB; 10 lg), a specific p38 inhibitor, decreased chronic morphineincreased NFjB p65, but not p-p38 levels. NS, saline. p38 or b-actin levels are used as loading controls. Data represent mean ± SEM of 4 rats. *p < 0.01 vs saline group at day 7. # p < 0.05 vs morphine group at day 7. ## p < 0.01 vs morphine group at day 7. Stat1. The cellular co-localization of p-stat1 showed its preferential localization in astrocytes (data not shown). As shown in Supplementary Fig. 9, we did not observe any significant changes in the tyrosine phosphorylation levels of other Stats including Stat2, 5 and 6 following repeated treatment with morphine Modulation of NMDA receptor expression by CGRP receptor signaling in the development of tolerance to morphine-induced analgesia It has been well established that NMDA receptors play a crucial role in mediating the development of tolerance to morphine-induced analgesia [38,59]. Accordingly, we investigated if our chronic morphine treatment altered NMDA receptor expression and whether these changes are affected by the modulation of CGRP receptor signaling. Three NMDA receptor subunits NR1, NR2A and NR2B failed to be altered by an acute treatment with morphine (15 lg; Fig. 10). In contrast, chronic morphine treatment (15 lg/ day for 7 days) up-regulated all the three receptor subunits. These effects were reversed by the co-administration of the CGRP receptor antagonist, BIBN4096BS (0.1 lg; Fig. 10). To identify further the possible mechanisms underlying changes in NMDA receptor Fig. 7. Chronic morphine-induced increases in p-ijb and acetyl-nfjb are reversed by the co-administration of BIBN4096BS or SB Western blot analyses reveal that a 7-day treatment with morphine (M; 15 lg/day) increases p-ijb and acetyl- NFjB p65 (Acetyl-NFjB) levels in the lumbar spinal dorsal horn, these effects being inhibited by co-treatment with BIBN4096BS (M/B; and 0.1 lg) or SB (SB; 10 lg), a specific p38 inhibitor. b-actin is used as a loading control. Data represent mean ± SEM of 4 rats. *p < 0.01 vs saline group at day 7. # p < 0.05 vs morphine group at day 7. ## p < 0.01 vs morphine group at day 7. expression following chronic morphine treatment, we examined the possible role of CaMKII, p38 and ERK. SB (10 lg), a p38 inhibitor, suppressed chronic morphine-induced up-regulation of NR1, NR2A and NR2B (Fig. 10). The ERK activation inhibitor, PD98059 (10 lg), and the CaMKII inhibitor, KN93 (15 nmol), did not affect morphine-induced NR1 up-regulation, but suppressed increase in NR2A and NR2B expression (Fig. 10).

8 Z. Wang et al. / PAIN Ò 151 (2010) Fig. 9. Distribution and cellular localization of p-stat3. Immunohistochemistry shows p-stat3-ser-ir cells (p-stat3; red) in the spinal dorsal horn (Aa). Double immunofluorescence reveals that p-stat3 (red) is co-localized with GFAP (green; Ad f), but not with OX-42 (green; Ab) or NeuN (green; Ac). p-stat3 (red) is also colocalized with p-erk (green; Ba c). Scale bar, 50 lm (Aa); 20 lm (Ab f and Ba c). Fig. 8. Reversal of chronic morphine-induced increases in p-erk and p-stat3-ser by co-administration of BIBN4096BS or PD (A and B) Western blot analysis reveals that a 7-day treatment with morphine (M; 15 lg/day) increases p-erk (A) and p-stat3-ser (B) levels in the lumbar spinal dorsal horn, these effects being inhibited by co-treatment with BIBN4096BS (M/B; and 0.1 lg) or PD98059 (PD; 10 lg), an ERK activation inhibitor. ERK1, ERK2 or Stat3 levels are used as loading controls. Data represent mean ± SEM of 4 rats. *p < 0.05 vs saline group at day 1. **p < 0.01 vs saline group at day 7. # p < 0.05 vs morphine group at day 7. ## p < 0.01 vs morphine group at day Discussion Activation of CGRP receptors has been coupled to the activation of AC/cAMP/PKA, CaMKII and MAPK as well as the subsequent induction and activation of transcription factors that regulate gene expression [4,21,50,51,55]. It has been suggested that CGRP is involved in the development of tolerance to morphine-induced analgesia [41,49,62]. However, the extent to which those related intracellular molecular cascade(s) associated with the activation of CGRP receptors contributes to CGRP involvement in the development of tolerance has yet to be determined. In that context, the present study provides the first evidence that differential activation of intracellular signaling cascades occurring in neurons and glia mediates the development of tolerance to morphine-induced analgesia, CGRP being an important upstream regulator. Chronic exposure to morphine stimulated the expression of CGRP in the central afferent terminals of DRG in the lumbar spinal dorsal horn. CGRP receptor components, CRLR and RAMP1, were also increased, leading to a loss of efficacy of morphine antinociceptive properties. Treatment with a non-peptide CGRP receptor antagonist reduced these effects. Upon stimulation of functional CGRP receptors Fig. 10. Differential modulation of NMDA receptor expression following chronic morphine treatment by BIBN4096BS, PD98059, SB and KN93. Western blot analyses reveal that a 7-day treatment with morphine (M; 15 lg/day) induced increases in NMDA receptor NR1, NR2A and NR2B subunits in the lumbar spinal dorsal horn. NR1 up-regulation is decreased by a co-treatment with the CGRP receptor antagonist, BIBN4096BS (M/B; 0.1 lg) or SB (SB; 10 lg), a specific p38 inhibitor while increases in NR2A and NR2B are reduced by co-delivery of BIBN4096BS (M/B; and 0.1 lg), SB (10 lg), PD98059 (10 lg) or KN93 (KN; 15 nmol). b-actin (Actin) is used as a loading control. Data represent mean ± SEM of 4 rats. *p < 0.05 vs saline group at day 7. **p < 0.01 vs saline group at day 7. # p < 0.01 vs morphine group at day 7. expressed on neurons and glia, released CGRP regulated the activation of neuronal CaMKII CREB, microglial p38 NFjB and astroglial

9 202 Z. Wang et al. / PAIN Ò 151 (2010) ERK Stat1/3 cascades. Inhibitors of CaMKII, p38 and ERK upstream kinase correspondingly reduced p-creb, NFjB and p-stat1/3 levels and attenuated the development of tolerance. Taken together, these data established the relative contribution of neurons, microglia and astroglia in CGRP-associated development of morphine tolerance Activation of neuronal CaMKII CREB cascade following chronic morphine The nuclear transcriptional factor CREB mediates the transcriptional regulation of various genes encoding peptides and proteins, which requires CREB phosphorylation at Ser 133. Both MAPK and CaMKII have been shown to regulate CREB phosphorylation at Ser 133 [19,57] and are associated with tolerance to morphine-induced analgesia [11,34,62,63]. However, little is known about the association between these kinases and their downstream transcription factors in the development of tolerance. We propose here for the first time that the CaMKII CREB cascade activated in neurons mediates the development of tolerance. Chronic morphine led to the activation of neuron-localized CaMKII and CREB. The inhibition of CaMKII blocked the development of tolerance and the increase in p-creb. In addition, we did not observe any increase in p-camkiv induced by chronic morphine or co-localization of p-mapks with neurons, thereby excluding these kinases being the upstream candidates of neuronal CREB activation. Additionally, treatment with CGRP activated both CaMKII and CREB in vitro [55]. Our in vivo data showed that the CaMKII CREB cascade implicated in the development of tolerance was modulated by blockade of CGRP receptors, demonstrating a role for the CaM- KII CREB cascade in CGRP-associated development of tolerance Activation of microglial p38 NFjB cascade following chronic morphine The transcription factor NFjB p50/p65 is the most common heterodimer, of which p65 subunit possesses transcriptional activity. In an inactive state, NFjB resides in the cytoplasm as a latent form, associated to its inhibitory protein, IjB. NFjB activation requires the occurrence of a sequential series of events, including IjB phosphorylation and the subsequent dissociation from the complex for degradation and the translocation of NFjB to the nucleus to regulate gene expression [42,44]. Aside from IjB, the NFjB p65 subunit is also subjected to several post-translational modifications such as phosphorylation and acetylation to enhance transcription [47,60]. We reported here that chronic morphine enhanced IjB phosphorylation and NFjB p65 acetylation. These effects were regulated by inhibition of p38 or CGRP receptors. p-p38 and NFjB were up-regulated following chronic morphine, and were also modulated by CGRP receptor signaling. Inhibition of p38 attenuated the development of morphine analgesic tolerance. Collectively, these results demonstrate that the development of tolerance involves the p38 NFjB cascade in microglia and this cascade was regulated by disrupting CGRP receptor signaling. Our observations are in line with previous studies that p38 activates NFjB activity in cultured microglia [26,68]. Additionally, glia-derived IL-6 was proposed to be involved in the development of tolerance [52,53]. Subsequent analysis has associated the up-regulation of IL-6 to the activation of p38 in microglia [62]. Moreover, the promoter region of IL-6 gene contains NFjB binding sites and is controlled by NFjB to initiate gene transcription [14,71]. Taken together, the p38 NFjB-IL- 6 cascade is likely involved in the development of tolerance under modulation by CGRP receptor signaling. To add complexity to the role of NFjB in morphine-induced analgesia, a few neurons in the spinal dorsal horn are NFjB p65- positive (data not shown). The cellular distribution of NFjB p65 was similar to a previous study [2]. However, the extent to which neuronal NFjB contributes to the development of tolerance awaits further clarification Activation of astroglial ERK Stat1/3 cascade following chronic morphine It is traditionally proposed that the tyrosine phosphorylation of Stats is a prerequisite for their DNA binding to activate gene transcription and the serine phosphorylation of some Stats such as Stat1 and Stat3 (Stat1/3) complementarily regulate their transcriptional activity by positively or negatively influencing the tyrosine phosphorylation [13]. Within this context, increase in serine phosphorylation and decrease of tyrosine phosphorylation of both Stat1/3 occurred simultaneously following chronic morphine. Interestingly, Stat1/3 contains a consensus sequence at Serine 727, which can be targeted by ERK [8,66]. Increase in serine phosphorylation of Stat1/3 following repeated morphine was inhibited by ERK inactivation. Moreover, the inhibition of ERK activation attenuated chronic morphine-induced decrease of tyrosine phosphorylation (at least for Stat3), indicating that the increased serine phosphorylation is ERK-dependent, which may contribute to decrease in tyrosine phosphorylation of Stat3. These observations are consistent with previous findings that ERK-dependent serine phosphorylation of Stats negatively modulates their tyrosine phosphorylation [8]. In contrast, emerging evidence shows that serine 727 phosphorylation of Stats per se, independent of their tyrosine phosphorylation is responsible for their transcriptional activity and even mediates gene expression [35,43]. In the present study, chronic morphine activated Stat1/3 by enhancing their serine 727 phosphorylation, which is ERK dependent and observed mostly in astrocytes. The morphological localization of p-stat3 is consistent with a previous study [45]. In addition, astrocyte-localized ERK activation was enhanced following chronic morphine and the inhibition of ERK activation blocked the development of tolerance, suggesting an involvement of the astroglial ERK Stat1/3 cascade in this phenomenon. All associated changes in the ERK Stat1/3 cascade were influenced by blockade of CGRP receptors, suggestive of an association with CGRP receptors. Several astrocyte-related molecules such as the glial glutamate transporters (GLAST and GLT-1) and IL-1b are implicated in morphine analgesic tolerance. Chronic morphine down-regulates astroglia-localized GLAST and GLT-1, but up-regulates IL-1b levels while activation of GLAST and GLT-1 or inhibition of IL-1 receptors attenuates the development of tolerance [27,40,52,58,62]. In parallel to the down-regulation of GLAST and GLT-1 and up-regulation of IL-1b, attenuation of tyrosine phosphorylation and enhancement of serine phosphorylation of astrocyte-localized Stat1/3 occur following chronic morphine. Furthermore, binding sites for Stat1/3 [16] have been identified in the promoter regions of GLAST, GLT-1 and IL-1b [22,33]. Activated Stat1/3 has been linked to the up-regulation of IL-1b [33]. Combined with the findings that both astrocyte-localized ERK Stat1/3 cascade (present data) and astrocyte-localized ERK-IL-1b cascade [62] are activated following chronic morphine, it is conceivable that ERK Stat1/3(serine) cascade regulating IL-1b expression may occur and be involved in the development of tolerance. In contrast, the tyrosine phosphorylation of Stat1/3 may be coupled to astrocyte GLAST and GLT-1 expression, but future experiments will be required to establish this hypothesis Modulation of NMDA receptor expression by CGRP signaling following chronic morphine Several reports have shown that chronic morphine increases NMDA receptor subunits NR1 levels and NR2B phosphorylation

10 Z. Wang et al. / PAIN Ò 151 (2010) Fig. 11. Schematic model of the intracellular signaling cascades underlying CGRP involvement in the development of tolerance to morphine-induced analgesia. Chronic intrathecal treatment with morphine induces increased synthesis of CGRP in the central terminals of primary afferents and of CGRP receptors in the dorsal horn of the spinal cord. CGRP then acts on CGRP receptors located on neurons, microglia and astrocytes [21,37,51,54], promoting the preferential activation of CaMKII CREB, p38 NFjB and ERK Stat1/3 cascades, respectively. This in turn leads to increased levels of NMDA receptors, IL-6 and IL-1b by binding of these transcription factors (CREB, NFjB and Stat1/3) to their corresponding binding sites in the promoter regions of genes encoding NMDA receptors, IL-6 and IL-1b, respectively. The combination of these various factors subserves the development of tolerance to morphine-induced analgesia by facilitating excitatory synaptic transmission [28]. [34,36]. However, underlying mechanisms regulating NMDA receptor expression following chronic morphine are still unclear. We provide here novel data showing that increase in NMDA receptor subunits NR2A and NR2B induced by chronic morphine treatment is modulated by neuronal CaMKII. Moreover, the neuronal CaMKII CREB cascade was shown here to be involved in the development of tolerance and CREB binding sites have been found in the promoter region of NR2B gene and NR1 gene which can be regulated by CREB activation [1,30,32]. Altogether, these data suggest the existence of a CaMKII CREB cascade targeting NMDA receptor expression in the development of morphine analgesic tolerance. In addition, the expression of NMDA receptor subunits following chronic morphine is modulated by microglia-localized p38 and astroglia-localized ERK activity, suggesting a glial contribution to NMDA receptor expression. Since the CaMKII CREB, p38 NFjB and ERK Stat1/3 cascades were modulated by the blockade of CGRP receptors, it is not surprising that NMDA receptor expression may also be regulated by disrupting CGRP receptor signaling. Our data support this hypothesis. In summary, CGRP acts as a regulator of the development of tolerance to morphine-induced analgesia, as shown in Fig. 11. CGRP receptor antagonists could thus prove useful to be adjuncts in morphine-based management of various types of chronic pain, especially considering the clinical effectiveness and safety of CGRP receptor antagonists such as BIBN4096BS and telcagepant in the treatment of migraine [24,46]. Acknowledgements This study was supported by a grant from the Canadian Institutes of Health Research to R.Q. The authors also thank Mira Thakur for editing assistance. The authors declare that they have no conflicts of interest related to this work. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi: /j.pain References [1] Bai G, Kusiak JW. Cloning and analysis of the 5 0 flanking sequence of the rat N- methyl-d-aspartate receptor 1 (NMDAR1) gene. Biochim Biophys Acta 1993;1152: [2] Bethea JR, Castro M, Keane RW, Lee TT, Dietrich WD, Yezierski RP. Traumatic spinal cord injury induces nuclear factor-kappab activation. J Neurosci 1998;18: [3] Bohn LM, Gainetdinov RR, Lin FT, Lefkowitz RJ, Caron MG. Mu-opioid receptor desensitization by beta-arrestin-2 determines morphine tolerance but not dependence. Nature 2000;408: [4] Brain SD, Grant AD. Vascular actions of calcitonin gene-related peptide and adrenomedullin. Physiol Rev 2004;84: [5] Chen Y, Geis C, Sommer C. Activation of TRPV1 contributes to morphine tolerance. involvement of the mitogen-activated protein kinase signaling pathway. J Neurosci 2008;28: