Background
The opioid system, comprised of multiple highly homologous receptor families and their endogenous opioid peptide ligands, is fundamental to the modulation of the sensory and affective aspects of pain [
1]. Three classes of opioid receptors (ORs) have been identified through molecular and pharmacological techniques, namely the mu (μ), delta (δ), and kappa (κ) ORs [reviewed by 2, 3]. Morphine, a classical μOR agonist with remarkable analgesic efficacy, is the current gold standard in the clinical treatment of moderate to severe pain; however, its use in the management of chronic pain may be restricted by the development of analgesic tolerance and the unwanted side effects associated with dose escalation. As such, understanding the mechanisms underlying opioid tolerance has become the primary focus of an extensive research effort with the aim of uncovering novel therapeutic strategies to treat persistent, unremitting pain.
A growing body of evidence identifies the δOR as an instrumental player in the development of morphine-induced analgesic tolerance [reviewed by 4]. Thus, concomitant administration of δOR antagonists with morphine [
5‐
9] or antisense oligodeoxynucleotide treatment directed against the δOR [
10] partially blocked the development of tolerance to morphine antinociceptive effects. In agreement with this data, δOR null mutant mice had a lower propensity to develop antinociceptive tolerance to morphine compared to their wild type littermates [
11,
12]. The mechanism by which δOR modulates μOR analgesic tolerance is not presently known, however, complex interactions between μ and δORs are likely to be relevant in eliciting various opioid-induced physiological responses. For example, direct coupling of μ-δORs in the form of hetero-oligomers has been demonstrated in both expression systems and spinal cord tissue [
13], which was proposed to underlie the antinociceptive synergy between μ and δOR agonists. We, and others, have also demonstrated that chronic activation of the μOR induces a translocation of δORs from intracellular compartments to neuronal plasma membranes and this phenomenon is correlated with an increase in δOR functional competence [
14‐
18]. Taken together, the activation and translocation of δORs may represent an important intermediary step in the development of morphine tolerance; however the mechanism underlying this trafficking remains unknown.
Several studies suggest an intimate and interactive relationship between opioids and glial cells. Once regarded as mere supports cells for CNS neurons, glial cells are now recognized as performing vital and complex functions in response to physiological stressors. Indeed, spinal glial activation has been observed in a number of pathological states including Alzheimer's [
19,
20] and Parkinson's [
21] diseases, HIV-associated dementia [
22‐
24], as well as several persistent pain syndromes [
25‐
30]. Moreover, spinal glial cell activation has been linked to the development of opioid tolerance. Chronic morphine treatment was reported to activate microglial [
31] and astrocytic [
31,
32] cells and to increase pro-inflammatory cytokine levels [
31] in the lumbar spinal cords of tolerant rats. Accordingly, co-administration of a glial modulatory agent with morphine attenuated the spinal immune response and inhibited the loss of morphine analgesic potency [
31,
32], suggesting that spinal glia may contribute to mechanisms responsible for opioid tolerance.
In the current study, we aimed to investigate the functional relationship between δORs and glial cells following prolonged chronic morphine administration. We employed immunohistochemical techniques as well as a behavioural nociceptive paradigm to assess whether prolonged morphine treatment is associated with the activation of spinal glial, and if indeed so, whether this spinal immune response is requisite for the observed enhancement in δOR-mediated antinociception.
Discussion
Opioid agonists are highly efficacious analgesics; however their clinical use is limited by the incidence of adverse effects, particularly the development of analgesic tolerance following repeated use. A growing body of evidence identifies an important role for the δOR in modulating morphine tolerance [
5‐
10] and this phenomenon may involve the trafficking of δORs from internal stores toward the neuronal plasma membrane, thereby enhancing the effects of δOR-selective ligands [
4,
14‐
18,
33]. The mechanism by which this contributes to morphine tolerance is unknown; however recent studies support an active role for spinal glia following chronic morphine treatment [
31,
32]. In the current study, we investigated the relationship between δORs and glial activation and indeed demonstrate a functional role for spinal glia in morphine-induced changes in δOR agonist effects. Moreover, administration of a glial inhibitor effectively blocked these changes in δOR function.
The involvement of spinal glia in the modulation of morphine analgesia has been demonstrated in both preclinical [
25,
27,
31,
32,
34‐
36] and clinical [
37] studies. We hypothesized that the recruitment of glial cells is a gradual response to long-term morphine administration and may be detectable at time points earlier than those at which analgesic tolerance is established. We therefore assessed the spinal immune response using a 48 h morphine dosing schedule; one which has been shown to have substantive effects on δOR trafficking and function [
14‐
16,
38]. This dosing regimen does not produce a state of tolerance [
15]; however it may initiate mechanisms involved early in the cascade of events leading to opioid tolerance. In developing a means of assessing the three dimensional structures of GFAP- and OX42-immunoreactive cells, we observed significant increases in cell volume and surface area of fluorescent GFAP- and OX42-immunoreactive cells in the dorsal spinal cord following prolonged morphine treatment. These results are in accordance with previous studies [
31,
32] illustrating the recruitment of glia in the events precipitating opioid tolerance. Morphine-induced glial hypertrophy was attenuated by co-administration with propentofylline. Interestingly, while propentofylline administration alone had no effect on astrocytes, it produced significant microglial hypertrophy in comparison with saline-treated rats. It is not clear why this occurs, since the combination of morphine and propentofylline did not show such an effect. The neuroprotective role of microglia in the CNS is well known and this cell population is very much attuned to its microenvironment, responding swiftly to even subtle physiological changes [
39]. It is possible that the localized administration of an exogenous compound into the spinal canal, in the absence of any 'pathological' events, was sufficient to produce a microglial response, although such an observation has not been reported previously [
31]. Nevertheless, additional functional studies are necessary to determine whether this propentofylline-induced increase in cell size was indeed accompanied by an inflammatory response. Despite microglial hypertrophy, however, neither baseline tail flick latencies nor deltorphin-mediated analgesia were altered following propentofylline administration alone, suggesting that this increase in microglial cell size was not functionally relevant in our study.
Activation of both glial cells and δORs appears to be important in the mechanisms of morphine tolerance, however it is unknown whether these two events are mutually exclusive or if, in fact, they represent important and related intermediary steps in the development of tolerance. Previous studies demonstrate that δORs are trafficked from internal stores toward the neuronal plasma membrane following morphine treatment, correlating with an increased functional competence of the receptor [
14‐
16]; however it is not known if the spinal immune response observed following morphine is requisite for this δOR trafficking event. Therefore, our second series of experiments aimed to examine the functional role of spinal glia in morphine-induced changes in δOR function. Consistent with earlier reports [
14,
15,
40,
41], we observed a significant augmentation in δOR-mediated effects in rats treated with morphine. This enhancement was effectively blocked by co-administration of morphine with propentofylline, demonstrating an integral role of spinal glial activation in the functional changes in δOR.
Taken together with previous reports that glial inhibition prevents the development of morphine tolerance [
27,
31,
32], this study provides additional evidence for the role of δORs in opioid tolerance and suggests that glial activity may precipitate changes in the δOR, including receptor trafficking. Glial cell activity has been documented to modulate the trafficking of ionotropic channels such as AMPA receptors [
42,
43]; however the current study is the first to our knowledge to suggest such a modulation of a G protein coupled receptor. The precise mechanism by which glial-modulated functional changes in δOR may occur is unclear; however two possible mechanisms include i) increased efficiency with which the receptor couples to intracellular signaling cascades, and/or ii) enhanced cell surface expression of the receptor. Future experiments will be required to investigate these possibilities.
Conclusion
In the present study, we demonstrate a relationship between δOR function and spinal glial activation. Indeed, prolonged administration of morphine induced the activation of astrocytic and microglial cells in the lumbar spinal cord, which correlated with enhanced antinociceptive effectiveness of a δOR agonist. Moreover, attenuation of glial activation with propentofylline, a glial inhibitor, attenuated the enhancement of δOR agonist-mediated effects. These data support an intimate relationship between glial and opioidergic function and provide insight into the mechanisms by which opioid analgesic tolerance develops.
Methods
Animals
Adult male Sprague-Dawley rats (220–300 g; Charles River, Québec, Canada), were housed in groups of two with ad libitum access to food and water, and maintained on a reverse 12/12 h light/dark cycle. All behavioural experiments were performed during the dark phase of the cycle, and animals were handled prior to experimentation in order to reduce stress-related analgesia. All experimental protocols were approved by the Queen's University Animal Care Committee, and complied with the policies and directives of the Canadian Council on Animal Care and the International Association for the Study of Pain.
Drug treatments
Rats were separated into one of four groups receiving i) morphine and intrathecal saline, ii) morphine and intrathecal propentofylline (inhibitor of glial activation), iii) intrathecal propentofylline alone, or iv) intrathecal saline alone (control group). Morphine sulfate (MS) was administered every 12 h by subcutaneous injections of increasing doses (5, 8, 10, 15 mg/kg in saline; Sabex, Kingston General Hospital, Kingston, Ontario, Canada). This treatment protocol was previously shown to induce the trafficking of δORs from intracellular compartments to neuronal plasma membranes in cultured cortical neurons as well as in the spinal cord [
14]. Propentofylline and saline (10 μg/30 μl diluted in saline and 30 μl, respectively; Sigma, St. Louis, MO, USA) were administered by intrathecal injection via lumbar puncture between the L4 and L5 vertebrae under brief isofluorane anesthesia every 24 hours for 5 days based on drug administration protocols required to block morphine tolerance [
31]. Successful drug placement was confirmed by a vigorous tail flick upon injection. All experiments were performed 12 hours following the final morphine injection.
Double-labeling fluorescent immunohistochemistry for confocal microscopy
Rats (n = 3 per group) were deeply anesthesized with sodium pentobarbital (75 mg/kg, i.p.; MTC Pharmaceuticals, Cambridge, Ontario, Canada) and transaortically perfused with 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB; 500 ml, pH 7.4). The spinal cords were removed by spinal ejection and post-fixed in the above fixative for 30 minutes at room temperature and cryoprotected in 30% sucrose in 0.2 M PB for 48 hours at 4°C. Lumbar segments were isolated and cut into 40 μm transverse sections on a freezing sledge microtome and collected in 0.1 M Tris buffered saline (TBS; pH 7.4). Free-floating sections were incubated in a blocking solution containing 10% NGS, 10% BSA followed by incubation with a rabbit polyclonal antisera recognizing glial fibrillary acidic protein (GFAP; 1:2500 working dilution; DakoCytomation, Ontario, Canada) to label activated astrocytes and a mouse monoclonal antisera recognizing OX42 (1:1000 working dilution; Serotec, NC, USA) to label CD3/CDIIB receptors on activated microglia. Spinal cord sections were incubated overnight at 4°C with both primary antibodies, followed by incubation with goat anti-mouse and goat anti-rabbit secondary antibodies (both 1:200 working dilution; Molecular Probes, Invitrogen, Ontario, Canada) conjugated to Alexa 594 and Alexa 488 fluorophores, respectively. To assess non-specific labeling, control sections were processed in the absence of primary antibodies. Sections were mounted on glass slides, air-dried, and cover-slipped using Aquamount (Fisher Scientific, Ontario, Canada).
Immunoreactive cells were visualized in the deep dorsal horn using the Leica TCS SP2 multi photon confocal microscope (100 × magnification; Leica Microsystems Inc, Ontario, Canada) and images acquired and digitalized for quantitative analysis with Leica Confocal Software. Twenty-five to thirty-five serial images were captured in 0.75 μm increments throughout the z plane using identical acquisition parameters and x-y coordinates for each of 12–20 immunoreactive cells per rat for n = 3 rats per experimental group. The serial images were then stacked and reconstructed in three dimensions using Image-Pro Plus v5.0 software (MediaCybernetics, MD, USA). Total cell volume (in pixels3) and cell surface area (in pixels2) were calculated for each cell based on the three dimensional cell reconstruction. Statistical analyses were performed using Excel XP (Microsoft, Ontario, Canada) and Prism 4.01 (Graph Pad, San Diego, CA). The average volume and surface area for cells within each treatment group were calculated and expressed as means ± s.e.m. These values were compared by one-way ANOVA followed by Tukey's post-hoc multiple comparison test. P values less that 0.05 were considered significant.
Behavioural tail flick assay
The effects of a selective δOR agonist, DLT (10 μg/30 μl [i.t.]; Sigma), on thermal nociceptive responses were assessed using the hot water tail flick assay [
15]. The distal 5 cm segment of the rat's tail was immersed in noxious 52°C water, and the latency to a vigorous tail flick was measured. For n = 6 per group, three baseline latencies were measured prior to DLT injection in order to determine the normal nociceptive responses of the animals. A cut-off latency of four times the average baseline response threshold was imposed to avoid tissue damage in the event that the animal became unresponsive following DLT injection. Rats were then injected intrathecally with DLT. Thermal latencies were measured every 10 minutes following drug administration for 50 minutes. The % M.P.E. of DLT was calculated at the 30 minute time point, as this time point corresponded with the maximum analgesic effect of DLT.
% M.P.E. = (post-drug latency - baseline) ÷ (cut-off latency - baseline) × 100
The thermal latencies to respond were analyzed by two-way ANOVA followed by Bonferroni post-hoc and the transformed % M.P.E. data were analyzed by one-way ANOVA followed by Tukey's post-hoc multiple comparison test. All values are expressed as means ± s.e.m. P values less than 0.05 were considered significant.
Acknowledgements
This work was supported by grants from the Canadian Institutes of Health Research (CIHR), the JP Bickell Foundation, Harry Botterell Foundation, and Canadian Foundation for Innovation and Ontario Innovation Trust awarded to CMC. SVH was funded by a fellowship from Queen's University. CMC is a Canada Research Chair in chronic pain. The authors would like to extend their gratitude to Lihua Xue, James Jeong, Matthew Gordon, and Jeffrey Mewburn for technical assistance.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SVH: project conception and design; data analysis and interpretation; writing, editing, revision of manuscript
SAA: major data collection; data analysis; editing, revision of manuscript
AMWT: preliminary data collection; data analysis
CMC: project conception and design; data interpretation; editing, revision of manuscript
*All authors read and approved the final manuscript