The alleviation of intense inflammatory pain is the most positive effect of opioids and therefore, we analyzed the development of antinociceptive tolerance in the light of the changes that affect the MORs. Upon icv administration, opioids gain access to periventricular areas implicated in the control of ascending pro-nociceptive information. The analgesic test involves the application of a noxious thermal stimulus to promote a flick of the mouse tail, and the administration of analgesic drugs increases the time that elapses between these two events. Whereas, this motor response is still observed in the intercollicular decerebrated animal, stimulus of much higher intensity are needed to evoke this behavior in the spinal animal. Therefore, this response comprises a spinal reflex which is under a facilitatory drive from the brain stem and the MORs in the periaqueductal grey matter (PAG) play an important role in the antinociceptive effects of opioids administered by the icv route 23.

The analysis of MORs in the PAG reveals a series of isoforms that are produced by the alternative splicing of the murine MOR 24, and also by N-glycosylation of these variants 25. Accordingly, these MORs are generally observed at apparent masses of 50–65 kDa, 80–100 kDa and even higher. In our experimental paradigm, icv administration of 10 nmol morphine produced time-dependent antinociception that peaked 30 min after opioid administration and that reached about 80% of the maximum effect measurable in this test. This analgesia had ceased 3 h after the administration of the opioid (Fig. 1). During the time-course of the initial dose of morphine and beyond, the surface levels of MORs remained practically unchanged and the serine 375 displayed moderate phosphorylation. Accordingly, no substantial internalization of MORs was produced and only a small signal could be immunoprecipitated from the supernatant (Fig. 1).

The induction of single-dose tolerance is a time-dependent phenomenon that develops after the first exposure of the animals to the opioids and it can be impaired by inhibitors of protein synthesis 26. Thus, we determined the interval between morphine administrations required to detect this tolerance. When the initial dose of 10 nmol morphine was repeated 3 h after the first dose, the analgesic effect was comparable to that of the first 27. However, at intervals of 6 h or 24 h the analgesia produced by morphine was much weaker (Fig. 2). The antinociceptive potency of the initial dose of morphine was recovered after 4 to 5 days 28. Interestingly, this time frame is that required to recover the analgesic effects from the action of β-funaltrexamine, an irreversible antagonist of MORs 29. These observations suggest that new synthesis of MORs is required to overcome the tolerance that follows an acute dose of morphine.

When a second dose of morphine was icv-injected 6 h after the first, antinociceptive tolerance was accompanied by a reduction in the amount of MORs in the membrane coupled to an increase in the proportion of intracellular receptors. Moreover, whilst the first dose of morphine caused moderate serine375 phosphorylation of MORs, this second dose promoted notable phosphorylation of the receptors that could be detected both at the surface as well as in the internalized MORs (Fig. 2). The Ser-phosphorylated MORs could be detected in the plasma membrane even 24 h after of administration of this second dose of morphine. Therefore, the repeated administration of morphine promoted both phosphorylation and internalization of MORs, and now the decrease of surface MORs correlated with an increase in the internalized receptors. After this second dose of the opioid, further doses morphine spaced at intervals of 24 h produced comparable peak analgesic effects. Thus, after promoting a high tolerance then morphine is able to support the recycling of the MORs, and then tolerance to this abated effect develops at a slow rate. However, the recycling of MORs receptors seems to be only partial, and reductions in the intervals between consecutive doses could augment antinociceptive desensitization.

The maximal effect of an icv dose of 200 pmol DAMGO was similar to that with 10 nmol morphine, producing about 80% of the MPE. However, the analgesic potency of a second dose of DAMGO was maintained when injected 6 h after the first. Only when this interval increased to 24 h was a moderate decrease in the effects of this dose observed (Fig. 3). In contrast to morphine, the initial dose of DAMGO stimulated strong Ser375 phosphorylation and internalization of MORs in PAG neurons. However, an important part of the internalized receptor recycled to the plasma membrane within 3 h of the administration of DAMGO when its analgesic effects had ceased (Fig. 3). While the internalized MORs were no associated with Gαi2 or Gβ1/2 subunits, they did co-precipitate with β-arrestin2 and C-Raf. This observation suggests that the internalized MORs regulate β-arrestin2-dependent pathways in these neurons 4. Thus, our results are in agreement with those describing some internalization of MORs in the adult rodent brain as a consequence of systemic administration of acute doses of etorphine (DAMGO) and the failure of an acute morphine administration to provoke a detectable loss of these receptors 14. Moreover, our observations in PAG neurons are comparable with those with DAMGO in HEK 293 cells expressing MORs. In this model DAMGO produces the robust Ser375 phosphorylation and internalization of MORs 78, and its removal facilitated the recycling of MORs to the plasma membrane and the recovery of the sensitivity to the agonist.

The differences observed in the capacity of opioids to produce desensitization have been attributed to their intrinsic efficacy. This idea assumes that DAMGO activates only a fraction of the MORs required for morphine to produce comparable analgesic effects, thereby promoting much lower tolerance. However, DAMGO could also stimulate antinociceptive tolerance by reducing the number of plasma membrane MORs. To test this possibility we analyzed the capacity of DAMGO and morphine to produce tolerance when a more demanding administration protocol was used. Morphine and related opioids provoke a profound desensitization of their analgesic effects when three consecutive doses are icv-injected into mice at 3 h intervals 27. Unfortunately, this protocol also produced desensitization to the analgesic effects of DAMGO, and the second and third consecutive doses increased the loss of surface receptors, thereby producing desensitization to the analgesic response to this opioid (Fig. 4). Indeed, the successive administrations of morphine also slowly diminished the density of cell surface MORs (not shown). The decrease in surface MORs correlated with an increase in the internalized MORs. While the MORs recycle rapidly, a proportion of the internalized MORs undergoes endocytic sorting to lysosomes and thus, proteolytic degradation 89. The subcellular fractionation of PAG synaptosomes revealed that 30 min after the initial icv-injection of DAMGO, the internalized MORs were detected in the early endosome and the recycling endosome fractions. However, after four consecutive doses administered at 3 h intervals, the internalized receptors accumulated and could be also detected in the late endosome/lysosome fraction for their destruction (Fig. 4).

The MORs in the spinal cord play an important role in development of tolerance induced by systemically administered opioids. Thus, we analyzed these receptors during the development of tolerance to morphine administered subcutaneously by implantation of oily morphine pellets. The morphine released from this suspension reaches levels of 10–13 nmol per mL of serum and of about 10 nmol per g of wet brain between 3 h and 12 h 30. In the hour that followed the implantation of the morphine suspension, the mice exhibited an analgesic response that reached the test cut-off time of 10 s. Subsequently, the mice developed rapid tolerance to this effect which was almost absent 12 h later (Fig. 5). The continuous administration of morphine increased both Ser375-phosphorylation and loss of surface MORs in the dorsal horn of the spinal cord. These changes were rapidly observed after the chronic morphine treatment commenced and persisted over the following 2 days. The reductions were more intense for the MORs that corresponded to protein bands of 75 and 100 kDa. It has been reported that these MOR species are rapidly degraded by the proteosome in HEK 293 cells 31 and our observations suggest a similar situation in spinal neurons. Therefore, the continuous presence of morphine in the receptor environment rapidly shifted the system from the uncoupling of the MORs from the regulated Gα subunits to the phase where GRKs gain access to MORs and promote Ser375 phosphorylation, resulting in the internalization of these receptors. This situation would be comparable to the internalization of MORs observed in the presence of elevated concentrations of morphine in dissociated primary cultures of rat embryonic striatal neurons 18 or mouse dorsal root ganglia neurons 19.

The initial dose of morphine promoted moderate phosphorylation of Ser375 and little or no internalization of the MORs. However, it did alter the association of plasma membrane MORs with Gα subunits (21, present work). Notably, the co-precipitation of these receptors with Gαi2 subunits was greatly diminished and only partially recovered 24 h later. Thus, the morphine-activated Gα subunits seem to have been permanently transferred to other compartment. Here, the RGS9 and RGSZ2 play a relevant role (Fig. 6) 2122. This reduction of MOR-regulated transduction brought about a substantial decrease in the antinociceptive activity of the subsequent doses of the opioid. Notably, after 6 h the subsequent administration of morphine promoted both phosphorylation and internalization of MORs. A good correlation was now observed between the decreases in surface MORs and the corresponding increase in the internalized pool, as well as with changes in their association with Gαi2 subunits (Figs. 6 &7). In these circumstances, the MORs in the membrane activated the Gα subunits that remained after the first dose of morphine but recovered their control after the effects of morphine had ceased. During the time-course of the effects of DAMGO, the Gα subunits underwent a transient transfer to RGSZ2 proteins and later, the recycled MORs in the plasma membrane recovered control over these G proteins and the response to DAMGO was resensitized (Figs. 3 &6). The results indicate that endocytosis and recycling of MORs diminished the permanent transfer (sequestering) of Gα subunits to RGS proteins, in this way reducing the tolerance to the effects of subsequent administrations of the opioids 11121332. This was observed for DAMGO (Fig. 3) and for a second dose of morphine, which now promoted low tolerance to the effects of additional doses (Fig. 2).

In these studies, male albino CD-1 mice weighing 22–25 g were used (Charles River, Barcelona, Spain). PAG synaptosomal membranes were obtained from groups of 6 to 10 mice that were sacrificed by decapitation at various intervals after receiving an icv-injection of DAMGO or morphine. The PAGs were collected and homogenized in 10 volumes of 25 mM Tris-HCl (pH 7.4), 1 mM EGTA and 0.32 M sucrose supplemented with a phosphatase inhibitor mixture (Sigma # P2850), H89 (Sigma, B1427) and a protease inhibitor cocktail (Sigma, P8340), that contained 4-(2-aminoethyl)-benzenesulfonyl fluoride (AEBSF), pepstatin A, transepoxysuccinyl-L-leucylamido(4-guanidino)butane (E-64), bebstatin, leupeptin and aprotinin. The homogenate was centrifuged at 1000 g for 10 min to remove the nuclear fraction, pellet 1 (P1). The supernatant (S1) was centrifuged at 20000 g for 20 min to obtain the crude synaptosomal pellet (P2). The pellet (P2) was resuspended in buffer and centrifuged at 20000 g for an additional 20 min, and the final pellet was diluted in Tris buffer supplemented with a mixture of protease inhibitors (0.2 mM phenilmethylsulphonyl fluoride, 2 μg/mL leupeptin and 0.5 μg/mL aprotinin) before aliquoting and freezing. The supernatant (S2) was centrifuged at 105,000 g for 1 h to obtain the crude microsomal pellet (P3) (Beckman XL-70 ultracentrifuge, rotor Type 70 ti). The S3 supernatant was concentrated in Amicon Ultra-4 centrifugal filter devices (nominal molecular weigh limit NMWL of 10,000 #UFC8 01024, Millipore Iberica S.A., Madrid, Spain), and it was then loaded on a 10–40% continuous sucrose gradient and centrifuged at 225,000 g for 18 h [40 and references therein]. Ten 4 mL fractions were collected, the proteins concentrated, and the MORs immunoprecipitated and analyzed by Western blotting.

To evaluate the presence of MORs in the plasma membrane and in intracellular structures the existing protein interactions were disrupted under denaturing conditions prior to performing immunoprecipitation. Thus, the PAG synaptosomal membranes (P2) and supernatants (S3) were heated in 40 mM Tris-HCl, 1% SDS buffer for 10 min at 100°C in the presence of 2-mercaptoethanol. This mixture was then cooled to room temperature and the SDS concentration reduced 7-fold by adding octylthioglucoside to a final concentration of 20 mM, and the immunoprecipitation of MOR was then performed with biotinylated IgGs (Pierce #21217 & 21339) as described below.

The co-precipitation of signaling proteins with the MORs, RGSZ2 and RGS9 proteins was performed under non-denaturing conditions. PAG preparations were sonicated (2 cycles of 5 s each in 400 μL of buffer containing 50 mM Tris-HCl (pH 7.5), 50 mM NaCl, 1% Nonidet P-40, and 50 μL protease and phosphatase inhibitors and H89). The supernatant was cleared by incubating with 20 μL of streptavidin-agarose (Sigma S1638) pre-equilibrated for 1 h at 4°C, which was recovered by centrifugation at 3000 g for 5 min. The Nonidet P-40 solubilized proteins were incubated overnight at 4°C with about 3 μg of affinity purified biotinylated IgGs directed against the target protein. The immunocomplexes were recovered with streptavidin agarose and the agarose pellets obtained by centrifugation were washed three times, pelleted and resuspended in Nonidet P-40 buffer. The resuspended immunocomplexes were transferred to centrifugal filter devices (5 μm, Amicon Microcon UFC40SV, Millipore) and the Nonidet P-40 buffer was separated from the IgG-agarose complexes by mild centrifugation. This washing step was repeated twice before the agarose-immunocomplexes retained in the filter membrane were heated in Tris-HCl pH 7,5, 1% SDS buffer at 100°C to separate the IgG-agarose from the target proteins. The immunoprecipitated proteins were then recovered by centrifugation and concentrated in centrifugal filter devices (10,000 kDa nominal molecular weight limit, Amicon Microcon YM-10 #42407, Millipore), solubilized in 2 × Laemmli buffer and resolved by SDS-PAGE. The detached anti MOR IgGs retained in the filter membrane were used to normalize the MOR-related signals. This procedure yielded enough protein to load four to six gel lanes. The proteins were transferred to 0.2 μm polyvinylidene difluoride membranes (PVDF; Amersham Biosciences, Spain) and probed with the selected antibodies in DecaProbe chambers (PR 150, Hoefer-AmershamBiosciences, Barcelona, Spain). The procedures used to prepare the PAG enriched synaptosomes, supernatant and the pull-down experiments have been described in detail elsewhere 2122.

Western blots were probed with a series of antibodies raised against signaling proteins. The blots were probed with affinity purified IgGs: antibodies directed against peptide sequences in the murine MOR1 (diluted 1:1000): NT (DSSAGPGNISDCSDP, residues 2–16 of the extracellular N-terminal 25), 2EL (TKYRQGSID, 208–216 of the second external loop 25), anti Gαi2 (1:1000, 25); anti-β-arrestin2 (1:1000, Calbiochem 178600); anti-C-Raf (1:2000, BD Transduction labs., 610151). Rabbit polyclonal IgGs against phospho-μ-opioid receptor (Ser375) (1:1000, Cell Signaling, 3451) were used to analyze the phosphorylation state of the MORs. All the antibodies were diluted in TBS + 0.05% Tween 20 (TTBS) and incubated with the PVDF membranes for 24 h at 6°C. The primary antibodies were detected with the corresponding secondary antibodies conjugated to horseradish peroxidase diluted 1:10000 in TTBS. Antibody binding was visualized with the ECL+plus Western Blotting Detection System (RPN2132, Amersham Biosciences) and the chemiluminescence was recorded with a ChemiImager IS-5500 (Alpha Innotech, San Leandro, California) equipped with a Peltier cooled CCD camera that provided a real time readout of 30 frames per second (-35°C; high signal-to-noise ratio; dynamic range up of 3.4 OD). Densitometry was performed using Quantity One Software (BioRad) and expressed as the mean ± S.E. of the integrated volume (average optical density of the pixels within the object area/mm2). For each treatment, the assays were typically performed twice or three times on samples obtained from independent groups of mice and the results were comparable. For assays in which the immunoprecipitated protein was modified by the opioid treatment, equal loading was verified and where necessary the signal adjusted using that obtained when detecting the heavy chain of the IgGs used to precipitate the target protein. The IgGs present in the immunoprecipitated samples were detached from the target proteins and processed in parallel gels/blots 21.

Male albino CD-1 mice weighing 22–25 g were housed and handled in accordance with the European Community guidelines for the Care and Use of Laboratory Animals (Council Directive 86/609/EEC). Animals were lightly anaesthetized with ether, and the opioids were injected into the lateral ventricle in a volume of 4 μL, as described previously 41. The response of the animals to nociceptive stimuli was determined by the warm water (52°C) tail-flick test. Latencies in seconds were determined both before treatment (basal latency) and also after the administration of the substance under study (test latency). Baseline latencies ranged from 1.5 to 2.2 seconds and a cut-off time of 10 seconds was allotted to minimize the risk of tissue damage. Antinociception was expressed as the percentage of the maximum possible effect (MPE = 100 × [test latency-baseline latency]/[cut-off time-baseline latency]).

Groups of 15–20 mice were subcutaneously (sc) implanted with 10 ml/kg body weight of a suspension containing 50% saline (0.9% NaCl in distilled water), 42.5% mineral oil (Sigma #400-5), 7.5% mannide monooleate (Sigma #M-8546), and 0.1 g/ml morphine base. The oily pellet offers an animal model of opioid tolerance-dependence that reproduces the rapid rise of morphine in serum and brain described for hard pellets 30. Development of tolerance was monitored by measuring the analgesic response promoted by the release of the opioid.