Shift of µ-opioid Receptor Signaling in the Dorsal Reticular Nucleus Is Implicated in Morphine-induced Hyperalgesia in Male Rats

All procedures were approved by the Institutional Animal Care and Use Committee of the Faculty of Medicine of the University of Porto (Porto, Portugal) and were performed in accordance with the European Community Council Directive (2010/63/EU) and the ethical guidelines for pain investigation.²⁵  Pathogen-free adult male Wistar rats (Charles River colony, France) were maintained under controlled temperature (22 ± 2°C) and light (12/12h light/dark cycle, lights on between 8:00 h and 20:00 h) conditions with ad libitum access to food and water. We did not use female animals because previous studies found no sex-dependency for µ-opioid receptor implication in opioid-induced hyperalgesia.^4,5  The animals were allowed to acclimate to the housing facility for at least one week before any procedure. All experiments were conducted during the light phase. The subjective bias when allocating the animals to the experimental groups was minimized by arbitrarily housing the animals in pairs upon their arrival, then the animals were randomly picked from the cage for each procedure. After stereotaxic surgeries, the animals were housed individually. No a priori power analysis was performed. The sample sizes were based on common practice of the research group where, by default, six animals per group are used in experiments, giving us approximately 90% power to detect large differences (two standard deviations) between two groups, for continuous outcomes. There were no missing data; all values from animals correctly injected/implanted into the dorsal reticular nucleus and from the animals with misplaced “out sites” injections included in the analysis were available for the analysis. Also, no outliers were detected, and all the values were included in the analysis.

The lentiviral vectors used were produced as previously described.²⁴  Briefly, the cDNA for the µ-opioid receptor was cloned into a lentiviral transfer vector, inserted in antisense orientation relative to the human synapsin promoter which restricts transgene expression to neurons.²⁶  This transfer vector, which is the vector for µ-opioid receptor knockdown, also contained an encephalomyocarditis virus internal ribosome entry site, the enhanced green fluorescent protein, and the woodchuck hepatitis virus posttranscriptional regulatory element. The virus was produced by transfection of human embryonic kidney 293T cells with the transfer vector, a packaging plasmid, a plasmid encoding the rev protein, and a plasmid encoding the vesicular stomatitis virus G glycoprotein. The control vector was constructed similarly, using a transfer vector with the human synapsin promoter driving the expression of the enhanced green fluorescent protein. The titer of the vectors was determined by quantitative real-time polymerase chain reaction, and both vectors were used at 5 × 10⁶ transducing units per microliter.

Opioid-induced hyperalgesia was induced in the rats by the continuous subcutaneous infusion of morphine hydrochloride (generously provided by Dr. Paulo Cruz, Porto Military Hospital, Porto, Portugal) at 45 µg · µL⁻¹ · h⁻¹ for 7 days as described previously.¹⁴  ALZET® osmotic minipumps (model 2001; USA) were used for the delivery of morphine or the saline vehicle solution in control animals at 1 µl/h pump infusion rate for 7 days. The minipumps were implanted in animals weighting 285 to 315 g, and, unless otherwise indicated, the subcutaneous implantation was performed under isoflurane anesthesia.

Rats weighting 285 to 315 g were deeply anesthetized with an intraperitoneal mixture of ketamine hydrochloride (0.06 g/Kg) and medetomidine (0.25 g/Kg) and held in a stereotaxic frame (David Kopf Instruments, USA) for the implantation of a cannula or the injection of lentiviral vectors into the left dorsal reticular nucleus. Immediately after the stereotaxic procedures, the animals were also implanted with osmotic minipumps filled with morphine or saline as described for the induction of opioid-induced hyperalgesia. Upon completion of the latter procedure, the animals received 0.9% NaCl (0.1 ml/kg, subcutaneous) for rehydration followed by atipamezole hydrochloride (0.5 g/Kg, subcutaneous) to revert the anesthesia.

A guide cannula was implanted into the left dorsal reticular nucleus for pharmacologic experiments following the coordinates, determined according to the rat brain atlas²⁷  relative to the interaural line (Anterior-Posterior: −6.0 mm; Medial-Lateral: −1.4 mm; Dorsal-Ventral: −1.5 mm), and procedures described previously.²⁴ 

Stereotaxic injections were performed for the injection of the vector for µ-opioid receptor knockdown in morphine- (n = 6) and saline-infused animals (n = 5) or the injection of the control vector in morphine- (n = 6) and saline-infused animals (n = 7) in two rostrocaudal parts of the left dorsal reticular nucleus following the coordinates of the atlas Paxinos and Watson²⁷  (first injection: Anterior-Posterior: −6.0 mm; Medial-Lateral: −1.4 mm; Dorsal-Ventral: −1.5 mm; second injection: Anterior-Posterior: −6.4 mm; Medial-Lateral: −1.3 mm; Dorsal-Ventral: −1.7 mm), as described previously.²⁴  A total of 0.6 µl was injected per site. The effects of the lentiviral vectors on nociceptive behaviors were assessed by the von Frey and hot-plate tests as described below (see Nociceptive Behavior section) before and 7 days after the stereotaxic injections. The human synapsin promoter was previously shown to be fully active at day 7 after injection at the dorsal reticular nucleus.²⁴  All tests were conducted by an experimenter blinded to the treatments. The primary outcome measures in the studies using lentiviral vectors were mechanical and thermal sensitivities evaluated by the von Frey and hot plate tests, respectively. The results obtained in the von Frey test, presented graphically, were reported as mean withdrawal thresholds (± SD) and as mean percentage of baseline (i.e., before lentiviral injections) ± SD. The results obtained in the hot-plate test were reported as mean withdrawal latency (± SD) and as mean percentage of baseline (i.e., before lentiviral injections) ± SD.

The sustained administration of morphine at the dose regimen used typically induces hypersensitivity to mechanical and thermal stimulation.¹⁴  The von Frey and hotplate tests were used to evaluate mechanical and thermal sensitivity, respectively, in the rats. The animals were habituated to the experimenter and the experimental environment for a period of one week. The von Frey test was performed by placing the animals on an elevated transparent cage with a mesh wire bottom allowing the stimulation of the plantar surface of the left hind paw with calibrated von Frey monofilaments (Stoelting, USA) with logarithmically incremental stiffness ranging from 0.4 g to 60 g. Testing started with the 2-g filament applied perpendicularly to the plantar surface for 3 s. Withdrawal thresholds were determined using the Dixon up-and-down method.²⁸  The hotplate test was performed by placing the animals on a hotplate system (BIO-CHP Cold Hot Plate Test), with a surface temperature of 52°C. The nociceptive threshold was quantified as the latency (in seconds) to licking, retraction of the hind paw, or jump after placement of the rat on the hotplate. A 30-s cut off was used to avoid tissue damage.

All animals were behaviorally evaluated before and 7 days after the implantation of minipumps filled with morphine or saline to confirm the development of mechanical and thermal hypersensitivity. The animals were also monitored for signs of sedation. All animals implanted with morphine minipumps developed mechanical and thermal hypersensitivity. Nonetheless, despite all efforts to maintain the same experimental and environmental conditions throughout the study, basal nociceptive thresholds within saline- and morphine-infused animals were not always consistent. This is likely because the experimental groups were not always performed in the same exact period, owing to the high number of animals, and rodent behavior variation from animal to animal further strengthened this variability.

One experimental group, performed in response to peer review, was performed with the aim of evaluating the analgesic effects of morphine in the early times after the implantation of morphine-minipumps. For that, the animals were behaviorally evaluated before the implantation of saline or morphine minipumps (n = 4 rats each) and at several time points after minipump implantation (5 h and 2, 4, and 7 days).

The rotarod test was used to evaluate the effects of lidocaine at the dorsal reticular nucleus on the motor performance of the rats. The test was performed on naïve animals after training once a day for two consecutive days. Training consisted on placing the rats on a rotating rod (Ugo Basile, Italy) with the rate of rotation set at 10 revolutions per minute, until they fell off or until reaching a cutoff time set at 180 s. The animals that remained on the rod for 180 s were injected with either lidocaine (4% wt/vol) or saline (n = 7 each) at the dorsal reticular nucleus and the test was performed 30 min later. Animals that did not remain on the rod for 180 s were considered to have motor impairments.¹⁴  The test was conducted by an experimenter blinded to the treatment. The results were reported as mean time of permanence on the rod (± SD).

The animals were injected at the dorsal reticular nucleus with either 4% (wt/vol) of lidocaine hydrochloride, 0.1 ng of the µ-opioid receptor agonist DAMGO, 1.5 ng (i.e., an ultra-low-dose) of naloxone hydrochloride or 0.5 µg of N-[2-(4-Bromocinnamylamino)ethyl]-5-isoquinoline (H-89). All drugs were obtained from Sigma-Aldrich (Portugal) and dissolved in saline. Lidocaine or saline was injected at the dorsal reticular nucleus of morphine- (n = 6 per group) or saline-infused animals (lidocaine n = 5; saline n = 6). The animals of this experimental set were tested before and 30 min after injection. The dose and timing of lidocaine action were chosen based on previous studies.¹⁴  In a second set of animals, DAMGO or saline were injected at the dorsal reticular nucleus of morphine- (DAMGO n = 5; saline n = 6) or saline-infused animals (DAMGO n = 7; saline n = 6). The animals were tested before and 15 min after injection. The dose and timing of DAMGO action were chosen based on previous studies performed at the dorsal reticular nucleus and another medullary area.^24,29,30  In a third set of animals, naloxone was injected in saline- (n = 7) or morphine-infused animals (n = 8). The animals were tested before and 30 min after the injection of naloxone. Immediately after testing, morphine-infused animals were further injected with DAMGO, at the dorsal reticular nucleus, and tested 15 min later. The effects of naloxone + DAMGO were compared with the effects of DAMGO injected alone at the dorsal reticular nucleus of morphine-infused animals from the second experimental set (n = 5). In a separate group of morphine-infused animals, that were not behaviorally tested, naloxone (n = 7) or saline (n = 5) were injected 30 min before DAMGO at the dorsal reticular nucleus, and 15 to 20 min later the animals were euthanized for immunodetection of phosphorylated cAMP response element-binding (as described in the Tissue Preparation and Immunohistochemistry section). The ultra-low dose of naloxone and timings of action were chosen based on previous studies.³¹  In a fourth set of animals, performed in response to peer review, H-89 was injected in morphine-infused animals (n = 6). The animals were tested before and 40 min after the injection of H-89. The effects of H-89 were compared with the effects of saline injected alone at the dorsal reticular nucleus of morphine-infused animals from the first experimental set (n = 6). The dose and timing of H-89 were chosen based on previous studies.³² 

The injections were performed 7 days after guide cannula/minipumps implantation using a stainless-steel needle protruding 3 mm beyond the cannula and a volume of 0.5 µl was infused over a period of 1 min. All tests were conducted by an experimenter blinded to the treatments. The primary outcomes in the pharmacologic studies were mechanical and thermal sensitivities evaluated by the von Frey and hot plate tests, respectively. The results obtained in the von Frey test, presented graphically, were reported as mean withdrawal thresholds (± SD). The results obtained in the hot-plate test were reported as mean withdrawal latency (± SD).

After the last behavioral evaluation or drug injection, the animals were deeply anesthetized with an overdose of sodium pentobarbital (70 mg/kg intraperitoneal) and perfused through the ascending aorta with 100 ml of calcium free Tyrode’s solution, followed by 750 ml of a fixative solution containing 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.2. The brainstems were removed, immersed in fixative for 4 h followed by 30% sucrose in 0.1 M phosphate-buffered saline overnight, at 4°C, and sliced at 40 µm in coronal orientation in a freezing microtome.

Two different groups of animals were used: one group of animals which was implanted with morphine- or saline-minipumps (n = 6 each) and a second group of animals which was implanted with morphine- or saline minipumps and further injected with lentiviral vectors into the dorsal reticular nucleus (saline: control vector n = 7; vector for µ-opioid receptor knockdown n = 5; morphine: control vector and vector for µ-opioid receptor knockdown n = 6 each). One in every fourth section encompassing the dorsal reticular nucleus was incubated for 2 h in a blocking solution containing 0.1 M glycine and 10% normal swine serum in 0.1 M phosphate-buffered saline containing 0.3% Triton X-100 follow by an incubation for 48 h, at 4°C, in rabbit polyclonal antibody against µ-opioid receptor (ref: RA10104; Neuromics, USA), diluted at 1:1,000. After washing, the sections were incubated for 1 h in a swine biotinylated anti-rabbit serum diluted at 1:200 (Dako, Denmark). The sections were washed again and incubated for 1 h with the avidin-biotin complex (1:200; Vector Laboratories, USA). After washing in 0.05 M Tris-HCl, pH 7.6, bound peroxidase was revealed using 0.0125% 3,3’-diaminobenzidinetetrahydrochloride (Sigma-Aldrich, USA) and 0.025% H₂O₂ in the same buffer. The antibodies were diluted in 0.1 M phosphate-buffered saline containing 0.3% Triton X-100 and 2% normal swine serum. The avidin-biotin complex was diluted in the same solution without serum. The sections were then dehydrated and mounted in Eukitt. The primary outcome of the immunohistochemical detection of the µ-opioid receptor was the number of µ-opioid receptor-immunoreactive neurons. Five sections encompassing the rostro-caudal extent of the dorsal reticular nucleus were randomly taken from each rat and the numbers of µ-opioid receptor-immunoreactive neurons occurring into the left and right dorsal reticular nucleus using the ×20 objective were counted by an experimenter blinded as to the experimental group. No differences were detected between the left and right side of the dorsal reticular nucleus for either saline- or morphine-infused animals (data not shown); therefore, left and right cell profile counts were summed in each tissue section from this experimental group. In the second group of animals, which was implanted with morphine or saline minipumps and further injected with lentiviral vectors into the left dorsal reticular nucleus, the left and right cell profile counts were also summed in each tissue section. The dorsal reticular nucleus was delimitated in an additional set of immunoreacted sections counterstained with formol-thionin³³  according to the atlas of Paxinos and Watson.²⁷  The specificity of the anti–µ-opioid receptor antibody was previously tested by blocking the antibody with a blocking peptide in immunohistochemistry and western blot analysis.²²  We further tested antibody specificity by performing negative controls with omission of either the primary or the secondary antibodies which blocked all the immunostaining.

Two separate experimental settings were used. In the first experimental setting, phosphorylated cAMP response element-binding expression was evaluated in morphine- and saline- infused animals (n = 6 each). In the second experimental setting, performed in response to peer review, phosphorylated cAMP response element-binding expression was evaluated in morphine-infused animals, pretreated with an ultra-low dose of naloxone (n = 7) or saline (n = 5) at the dorsal reticular nucleus before DAMGO injection at the dorsal reticular nucleus. One in every fourth section encompassing the dorsal reticular nucleus was incubated for 2 h in a blocking solution, as explained above for the immunodetection of the µ-opioid receptor, followed by an incubation for 48 h, at 4°C, in rabbit polyclonal antibody against phosphorylated cAMP response element-binding (ref:06-519; Millipore, USA), diluted at 1:1,000. After washing, the sections were incubated for 1 h in a swine biotinylated anti-rabbit serum diluted at 1:200 (Dako, Denmark). The sections were washed again and incubated for 1 h with the avidin-biotin complex (1:200; Vector Laboratories, USA). After washing in 0.05 M Tris-HCl, pH 7.6, bound peroxidase was revealed using 0.0125% 3,3’-diaminobenzidinetetrahydrochloride (Sigma-Aldrich, USA) and 0.025% H₂O₂ in the same buffer. The antibodies and the avidin-biotin complex were diluted in the solutions described above for the immunodetection of the µ-opioid receptor. The sections were then dehydrated and mounted in Eukitt. The primary outcome of the immunohistochemical detection of phosphorylated cAMP response element-binding was the number of phosphorylated cAMP response element-binding positive nuclei. Five sections encompassing the rostro-caudal extent of the dorsal reticular nucleus were taken from each animal and photomicrographs were taken using a Zeiss light microscope with a high-resolution digital camera. The number of phosphorylated cAMP response element-binding positive nuclei was calculated using an automated cell counter plugin of the ImageJ software. In the first experimental setting, because no differences were detected between the left and right side of the dorsal reticular nucleus for either saline- or morphine-infused animals (data not shown), left and right numbers of phosphorylated cAMP response element-binding positive nuclei were summed in each tissue section. In the second experimental setting, phosphorylated cAMP response element-binding positive nuclei were counted in the left-ipsilateral dorsal reticular nucleus injected with the drugs. To verify whether drug-induced effects were restricted to the dorsal reticular nucleus, in both the first and second experimental setting, the numbers of phosphorylated cAMP response element-binding positive nuclei were additionally counted in the following medullary areas adjacent to the dorsal reticular nucleus, the cuneate nucleus, the nucleus of the solitary tract and the trigeminal subnucleus caudalis. We tested antibody specificity by performing negative controls with omission of either the primary or the secondary antibodies which blocked all the immunostaining.

Seven days after morphine- or saline-minipump implantations (n = 9 per group), rats were deeply anaesthetized with an overdose of sodium pentobarbital (150 mg/kg intraperitoneal) and euthanized by decapitation. The brains were harvested and immediately stored at −80°C. The medulla oblongata was cut into a frozen transverse block (1 mm in depth) from –5.60 to –4.68 mm relative to the Interaural line²⁷  from which the dorsal reticular nucleus (left and right sides) were dissected out using a tissue micropunch (Stoeling, USA). Each sample (n = 3 per group) was prepared by pooling the dorsal reticular nucleus from three animals. Total RNA from the dorsal reticular nucleus was extracted using the PureLink RNA Mini Kit (Thermofisher Scientific, Portugal) by following the manufacturer’s protocol and the RNA integrity verified by agarose gel electrophoresis. The first strand cDNA synthesis was prepared at 42°C during 1h, from 0.8 μg of total RNA using 200 U of reverse transcriptase enzyme (Nzytech, Portugal) and 500 ng of oligo(dT)12-18 (Nzytech, Portugal). To assess for potential contaminants, a control containing all reagents except the reverse transcriptase enzyme was included for each sample. The expression levels of µ-opioid receptor mRNA were then quantified by the standard 2ˆ(–delta delta CT) method using a StepOnePlus Real Time polymerase chain reaction system (Applied Biosystems) and a SYBR green chemistry (SYBR Select master mix, Applied Biosystems). The following intron-spanning primers 5’-GCCATCGGTCTGCCTGTAAT-3’ and 5’-CCAGATTTTCTAGCTGGTGGTTAG-3’ were designed to amplify exon 2 and the junction of exon 3a/4 from the canonical µ-opioid receptor-1 transcript. Normalization was performed by amplification of rat GAPDH using the primers 5´-GCATGGACTGTGGTCCTCAG-3´ and 5´-CCATCACCATCTTCCAGGAG-3´. The thermal cycling conditions included an initial denaturation step at 95°C for 15 s, followed by 45 cycles at 95°C for 15 s, 53°C for 30 s, and 72°C for 1 min. Melting curve analysis of every quantitative polymerase chain reaction was conducted to ensure amplicon specificity. The results were presented as relative differences to µ-opioid receptor mRNA of saline-infused animals at the dorsal reticular nucleus.

After the last behavioral evaluation, the animals used in the pharmacologic experiments were administered 0.5 µl of 0.6% Chicago sky blue dye (Sigma, Portugal) through the guide cannula and euthanized by decapitation, whereas the animals injected with the lentiviral vectors were killed by vascular perfusion for verification of the injection site as previously described.²⁴  In control vector-injected rats, the injection site was identified by direct detection of the enhanced green fluorescent protein labeling (fig. 1, A and B). In rats injected with the vector for µ-opioid receptor knockdown, the injection tract was detected by the formol-thionin staining³³  because the expression of the enhanced green fluorescent protein was faint. In this vector the enhanced green fluorescent gene was inserted into the expression cassette in the second position of the bicistronic lentiviral construct, after an internal ribosome entry site element, and in this type of constructs the enhanced green fluorescent gene expression is lower compared with a vector just containing a promoter and the reporter gene, as observed previously.²⁴  Additionally, the destruction of the antisense RNA, placed in the first position of the bicistronic construct, also likely degrades the enhanced green fluorescent mRNA. We analyzed injection sites encompassing the dorsal reticular nucleus (Supplemental Digital Content, fig. S1, https://links.lww.com/ALN/C402) and for the purpose of control injections, we also analyzed misplaced injections outside the dorsal reticular nucleus, termed here “out sites,” which were located either dorsally in the cuneate nucleus or laterally in the spinal trigeminal nucleus.

A total of three morphine- and three saline-infused animals received lidocaine injections placed outside the dorsal reticular nucleus. A total of four morphine- and four saline-infused animals received DAMGO injection in the out sites. A total of three saline- and four morphine-infused animals received ultra-low dose naloxone injections in out sites. The latter four morphine-infused animals further received an injection of DAMGO 30 min after the microinjection of the ultra-low dose naloxone. In the H-89 experimental group, three animals were injected in out sites with saline and four animals were injected in out sites with H-89. A representative distribution of drug injection sites within the dorsal reticular nucleus and in out sites is depicted in Supplemental Digital Content, figure S2 (https://links.lww.com/ALN/C403).

Control vector injections were placed in out sites in two saline- and two morphine-infused animals. Injections performed with the vector for µ-opioid receptor knockdown were placed in out sites in three saline- and one morphine-infused animals.

The behavioral effects of the drugs located in the dorsal reticular nucleus and in adjacent sites (termed out sites), the effects of lentiviral vectors located in the dorsal reticular nucleus, and the effects of the vector for µ-opioid receptor knockdown in out sites, obtained in the von Frey and hot plate tests, were analyzed by a two-way mixed ANOVA for repeated measurements. Mechanical threshold responses, obtained in the von Frey test, were logarithmic transformed because of their skewed distribution. In case of a significant interaction between group and time, we proceeded with pairwise comparisons using Tukey’s correction for multiple testing. The behavioral results of lentiviral vectors injections located in out sites are shown in the Supplemental Digital Content, figure S3 (https://links.lww.com/ALN/C404); however, no statistical analysis was performed, except for the injections with the vector for µ-opioid receptor knockdown, because of the small number of animals for each lentiviral vector (described in the Histology section). The results from the other lentiviral out site’s injections were inspected from individual value plots. t tests for independent samples were used to compare the mean number of µ-opioid receptor-immunoreactive cells, phosphorylated cAMP response element-binding⁺ nuclei and µ-opioid receptor-mRNA levels between saline- and morphine-infused animals and the mean numbers of phosphorylated cAMP response element-binding⁺ nuclei in Saline+DAMGO- and Naloxone+DAMGO-injected animals. The normality assumption was checked by inspection of the distribution of the variables both with q-q plots and histograms. However, we must acknowledge that the sample size limits the ability to detect departures from normality. The statistical analysis was performed by GraphPad Prism version 7 and SPSS version 24. The significance level was set at 0.05, and all statistical tests were two-tailed.

The subcutaneous administration of morphine at 45 µg · µL⁻¹ · h⁻¹ initially produced antinociception (5 h; fig. 1, A and B), followed by the development of a marked hypersensitivity to mechanical stimuli, from day 4 onwards (fig. 1A), and to thermal stimuli detected at day 7 (fig. 1B), as previously shown.¹⁴  The injection of lidocaine, but not saline, into the dorsal reticular nucleus significantly reversed mechanical (fig. 2, A and C) and thermal hypersensitivity (fig. 2, B and D). Lidocaine injections in out sites produced no effects, except for an increase of mechanical thresholds which did not fully reverse mechanical hypersensitivity (Supplemental Digital Content, fig. S4, A and B, https://links.lww.com/ALN/C405). The administration of lidocaine into the dorsal reticular nucleus did not interfere with the motor function of the animals as shown in the rotarod test (Supplemental Digital Content, fig. S5, https://links.lww.com/ALN/C423).

We next determined the effects of sustained morphine administration in the expression of the µ-opioid receptor by evaluating µ-opioid receptor mRNA levels and the number of µ-opioid receptor-immunoreactive cells at the dorsal reticular nucleus. No significant differences were found in mRNA levels between morphine- and saline-infused animals (fig. 3D). The numbers of µ-opioid receptor-immunoreactive cells were significantly higher in morphine-(103 ± 26 cells) compared with saline-infused animals (63 ± 9 cells; fig. 3, A–C).

We also evaluated the expression of phosphorylated cAMP response element-binding at the dorsal reticular nucleus by analyzing the numbers of positively labeled nuclei. Higher numbers were found in the dorsal reticular nucleus of morphine- (1,329 ± 315 nuclei) compared with saline-infused animals (935 ± 218 nuclei; fig. 4, A–C). To evaluate whether this effect was restricted to the dorsal reticular nucleus, we also analyzed its adjacent medullary areas. No differences were found between morphine- and saline-infused animals (Supplemental Digital Content, fig. S6, A–G, https://links.lww.com/ALN/C406).

We next determined the effects of µ-opioid receptor activation by the agonist DAMGO injected at the dorsal reticular nucleus in saline- and morphine-infused animals. In morphine-infused animals, DAMGO increased mechanical hypersensitivity and did not alter thermal hypersensitivity (fig. 5, A and B). In saline-infused animals, DAMGO reduced mechanical and thermal sensitivity (fig. 5, C and D). DAMGO injection in out sites produced no significant effects (Supplemental Digital Content, fig. S4, C and D, https://links.lww.com/ALN/C405).

Then we determined whether µ-opioid receptor expression was involved in the development of opioid-induced hyperalgesia by evaluating the effects of knocking down the receptor at the dorsal reticular nucleus. In animals injected with the control vector, morphine-infusion increased the numbers of immunoreactive cells compared with saline-infusion. The injection of the vector for µ-opioid receptor knockdown reduced the numbers of immunoreactive cells both in morphine- and saline-infused animals (fig. 6, A–C). Morphine-infused animals injected with the control vector showed decreased mechanical thresholds and heat-evoked withdrawal latencies indicative of the development of developed mechanical and thermal hypersensitivity, respectively (fig. 7, A and B). In saline-infused animals, the control vector produced no effects except for a reduction of mechanical sensitivity (fig. 7, A and B). The knockdown of the µ-opioid receptor attenuated the reduction of mechanical thresholds and heat-evoked withdrawal latencies in morphine-infused animals and produced the opposite in saline-infused animals (fig. 7, C and D). In the morphine group, the magnitude of reduction of mechanical thresholds (−17 ± 8% vs.−40 ± 9%) and heat-evoked withdrawal latencies (−10 ± 5% vs. −32 ± 2%) was lower compared with the control vector (fig. 7, E and F). In the saline group, receptor-knockdown decreased mechanical (31 ± 8% vs. −17 ± 8%) and heat-evoked withdrawal latencies (−24 ± 6% vs. −2 ± 10%) compared with the control vector (fig. 7, E and F). Vector injections in out sites produced no effects (Supplemental Digital Content, fig. S3, https://links.lww.com/ALN/C404).

An ultra-low dose of naloxone administered into the dorsal reticular significantly attenuated mechanical (fig. 8A) and thermal hypersensitivity in morphine-infused animals (fig. 8B). Naloxone produced no effects in saline-infused animals (fig. 8). Saline-vehicle injections were not performed as, in the same experimental conditions, it produced no effects (fig. 2, C and D). Naloxone injections in out sites produced attenuated morphine-induced mechanical and thermal hypersensitivity (Supplemental Digital Content, fig. S4, E and F, https://links.lww.com/ALN/C405). The latter results suggest naloxone diffusion to adjacent areas upon injection.

We next determined the effects of the pretreatment with an ultra-low dose of naloxone on the effects of DAMGO in morphine-infused animals. For that, we compared morphine-infused animals injected with DAMGO alone at the dorsal reticular nucleus with morphine-infused animals that were injected with an ultra-low dose of naloxone at the dorsal reticular nucleus 30 min before DAMGO injection. DAMGO after pretreatment with naloxone, contrary to DAMGO alone, fully reversed mechanical and thermal hypersensitivity (fig. 9, A and B). Injection of naloxone followed by DAMGO in out sites also showed antinociceptive effects for DAMGO (Supplemental Digital Content, fig. S4, G and H, https://links.lww.com/ALN/C405).

We further evaluated the effects of ultra-low dose naloxone in the expression of phosphorylated cAMP response element-binding in morphine-infused animals injected with saline or an ultra-low dose of naloxone into the dorsal reticular nucleus 30 min before the microinjection of DAMGO. Naloxone pretreatment showed significantly lower numbers of positively labeled nuclei (909 ± 77 nuclei) compared with saline (1,099 ± 163 nuclei; fig. 9, C–E) at the dorsal reticular nucleus. Additionally, because naloxone likely diffuses from the injection site, we also evaluated the expression of phosphorylated cAMP response element-binding in medullary areas adjacent to the dorsal reticular nucleus. No differences were found at the adjacent areas (Supplemental Digital Content, fig. S6, H–J, https://links.lww.com/ALN/C406).

To investigate the involvement of the protein kinase A signaling pathway at the dorsal reticular nucleus, we tested the effects of the protein kinase A inhibitor H-89 at the dorsal reticular nucleus of morphine-infused animals. The injection of H-89 into the dorsal reticular nucleus (fig. 10) or in out sites produced no effects (Supplemental Digital Content, fig. S4, I and J, https://links.lww.com/ALN/C405).

This study shows that the dorsal reticular nucleus, a major descending pain facilitatory area of the brain, is involved in the mediation of opioid-induced hyperalgesia. We show that chronic morphine infusion increases the levels of the µ-opioid receptor and phosphorylated cAMP response element-binding, a downstream marker of the excitatory signaling of µ-opioid receptor, at the dorsal reticular nucleus. We further show that µ-opioid receptor activation by DAMGO at the dorsal reticular nucleus enhances opioid-induced hyperalgesia, whereas µ-opioid receptor knockdown produces the opposite. Furthermore, we demonstrate that preventing µ-opioid receptor excitatory signaling attenuates opioid-induced hyperalgesia, restores the analgesic effects of DAMGO, and decreases phosphorylated cAMP response element-binding levels at the dorsal reticular nucleus. Taken together, our results indicate that chronic morphine infusion induces a switch in µ-opioid receptor signaling from inhibitory to excitatory at the dorsal reticular nucleus, which is likely one of the underlying cellular mechanisms of increased descending pain facilitation during opioid-induced hyperalgesia.

The complete reversal of mechanical and thermal hypersensitivity induced by lidocaine at the dorsal reticular nucleus indicates that descending pain facilitation from this area is involved in opioid-induced hyperalgesia. The activation of descending pain facilitatory systems^14,15  is involved in opioid-induced hyperalgesia likely through an enhancement of spinal dorsal excitability.^19,34  The activation of descending pain facilitation from another medullary area, the rostral ventromedial medulla, has been previously shown to be involved in the mediation of opioid-induced hyperalgesia induced by acute³⁵  and sustained opioid administration¹⁴  through local enhanced endogenous cholecystokinin activity and activation of cholecystokinin₂ receptors.¹⁶  We show that one of the mechanisms underlying the involvement of the dorsal reticular nucleus is the activation of the µ-opioid receptor. Whereas µ-opioid receptor expressed in sensory neurons was shown to be critical to the initiation of opioid-induced hyperalgesia,⁵  the sustained activation of the receptor in the dorsal reticular nucleus might contribute to its maintenance by triggering descending facilitatory influences to the spinal dorsal horn. Mechanisms independent of the µ-opioid receptor might also interplay in the dorsal reticular nucleus, because lidocaine fully blocks opioid-induced hyperalgesia whereas µ-opioid receptor knockdown or the ultra-low dose naloxone only produced an attenuation. These mechanisms could involve the NMDA receptor-glutamatergic system^4,20,21  and microglia-to-neuron signaling.³⁶ 

The absence of alterations in µ-opioid receptor mRNA levels is consistent with other studies showing the lack of effect of chronic morphine in the transcriptional regulation of the receptor in the brain.^37,38  This infers that the high number of cells expressing the receptor after morphine infusion could be attributable to increased expression at the protein level³⁹  or a change in the trafficking of the receptor resulting in its accumulation at the plasma membrane. The latter is supported by the fact that morphine is a poor internalizing agonist and the internalization of the receptor is crucial for targeting the receptor to degradation.⁶  Chronic morphine was shown to induce different effects in the density of the receptor with either downregulation, no change, or upregulation.^40–42  Tolerance has been associated, although not exclusively, with downregulation.⁶  At the dorsal reticular nucleus, downregulation of the receptor was associated with tolerance or diminished responsiveness to the inhibitory action of opioids.²⁴  In the present work, the upregulation is likely linked to a switch of signaling of the receptor from inhibitory to excitatory, as suggested by our cellular and behavioral data. At the cellular level, the shift to excitatory signaling is suggested by the increased expression of phosphorylated cAMP response element-binding, and decreased expression after ultra-low dose naloxone treatment, which prevents µ-opioid receptor coupling to a stimulatory guanine nucleotide-binding protein and restores coupling of the receptor to an inhibitory guanine nucleotide-binding protein.⁹  The best-established molecular adaptation to chronic opioid exposure is up-regulation of the cAMP/cAMP-dependent protein kinase A signaling. Additionally, cAMP response element-binding is activated by phosphorylation predominantly by protein kinase A.^7–9  However, the absence of effects of H-89 discards upregulation of this pathway at the dorsal reticular nucleus. Hence, other signaling pathways might be involved; for example, the protein kinase C and extracellular signal-regulated kinase pathways.^3,4,43–45 

At the behavioral level, DAMGO enhanced mechanical allodynia in morphine-infused animals and induced antinociceptive effects in saline-infused animals. Of note, we did not observe an enhancement of thermal hyperalgesia in the hot plate test, after the administration of DAMGO in morphine-infused animals, probably because of methodologic limitations. In the hot plate test, the heat intensity is usually set up to observe responses within 5 to 10 s.⁴⁶  Accordingly, in saline-infused animals, which showed withdrawal latencies on average near to 10 s, increased or decreased withdrawal latencies were observed after DAMGO or µ-opioid receptor knockdown, respectively, but in morphine-infused animals which showed withdrawal latencies on average near to 5 s, shorter latencies, indicative of an enhancement of hyperalgesia, are not likely detected. The antinociceptive effects of DAMGO are likely yielded through the inhibition of dorsal reticular nucleus descending facilitation. Further corroborating this, knockdown of the µ-opioid receptor at the dorsal reticular nucleus of saline-infused animals increased sensitivity to mechanical and thermal stimuli as previously shown.²⁴  It is important to point out that the control vector, although with less magnitude than the vector for µ-opioid receptor knockdown, selectively increased mechanical sensitivity in saline-infused animals. This effect is likely due to the enhanced green fluorescent protein expressed from the vector rather than the lentiviral transduction, because this protein is dose-dependently toxic to some neuronal cells.⁴⁷  The vector for µ-opioid receptor knockdown also carries this reporter gene, but its expression is almost null, therefore the effects of the vector are likely exclusively due to receptor knockdown. The pronociceptive effects of DAMGO in morphine-infused animals are likely mediated through µ-opioid receptor excitatory effects on dorsal reticular nucleus descending facilitation, which is further supported by the attenuation of opioid-induced hyperalgesia by knockdown of the receptor. Finally, the behavioral data are consistent with µ-opioid receptor being expressed on dorsal reticular nucleus-spinally projecting neurons,²²  which are excitatory and engaged in a reciprocal circuitry responsible for pain amplification.^20,21 

The switch of µ-opioid receptor signaling from inhibitory to excitatory is further suggested by the effects of ultra-low dose naloxone on phosphorylated cAMP response element-binding levels selectively at the dorsal reticular nucleus. Ultra-low dose naloxone attenuated morphine-induced hyperalgesia after injection into the dorsal reticular nucleus as well as in adjacent (out) sites. Morphine-infusion increased phosphorylated cAMP response element-binding levels in the dorsal reticular nucleus but not in adjacent areas, and ultra-low dose naloxone pretreatment in morphine-infused animals decreased phosphorylated cAMP response element-binding levels in the dorsal reticular nucleus while it had no effect in adjacent areas. These results indicate the selective involvement of the dorsal reticular nucleus and that the effects of naloxone injected in out sites are likely due to its diffusion to the dorsal reticular nucleus. Additionally, DAMGO in the dorsal reticular nucleus, but not in adjacent areas, produced an enhancement of mechanical hypersensitivity. This is consistent with the alterations of the µ-opioid receptor selectively at the dorsal reticular nucleus. Pretreatment with ultra-low dose naloxone restored the analgesic effects of DAMGO injected in the dorsal reticular nucleus as well as in out sites. Unlike at the dorsal reticular nucleus where morphine potentiates the coupling of the receptor to a stimulatory guanine nucleotide-binding protein, in adjacent areas the coupling of the receptors to inhibitory and stimulatory guanine nucleotide-binding proteins likely occurs dynamically.^8,9  Naloxone injected in out sites likely shifts this balance toward increased inhibitory coupling. Together with naloxone diffusion to the dorsal reticular nucleus, this might explain the analgesic effects of DAMGO in out sites. It is unlikely that the effects of the ultra-low dose of naloxone might be due to the spread of naloxone to the rostral ventromedial medulla, since opioid-sensitive neurons in this area were shown to be resistant to the development of cellular adaptations after local repeated morphine microinjections.⁴⁸  The involvement of the rostral ventromedial medulla in opioid-induced hyperalgesia has been shown to be mediated through cholecystokinin.¹⁶  Furthermore, µ-opioid receptor alterations in the rostral ventromedial medulla and dorsal reticular nucleus, in response to various types of pain, seem to be different.^24,30,49  More generally, in supraspinal pain modulatory areas, µ-opioid receptor seems to also adapt to different types of pain by diverse mechanisms that are nucleus specific.⁵⁰  This study shows that µ-opioid receptor–dependent mechanisms are involved in the activation of descending pain facilitation in opioid-induced hyperalgesia. To better envisage the translational perspectives of this study, future studies should be performed to confirm that the alterations found in µ-opioid receptor function and signaling are the same in female animals. Notwithstanding, uncovering the mechanisms involved in opioid-induced activation of descending facilitatory pathways may offer opportunities for developing new approaches to improve opioid analgesia and prevent paradoxical hyperalgesia.

In summary, our results show that descending pain facilitation from the dorsal reticular nucleus is involved in opioid-induced hyperalgesia and that it entails upregulation of µ-opioid receptor and altered µ-opioid receptor signaling pathways in dorsal reticular nucleus neurons. Because the µ-opioid receptor plays an important inhibitory function at the dorsal reticular nucleus, which accounts for the analgesic effects of systemic opioids,²⁴  the shift to excitatory signaling is likely in the genesis of increased descending pain facilitation.

Support for this study was provided by an International Association for the Study of Pain (Washington, D.C.) Early Career Research Grant; grant NORTE-08-5369-FSE-000026 from European Social Fund (Porto, Portugal); and grant PTCD/SAU-NSC/110954/2009- FEDER/COMPETEFCT from the Portuguese National Funding agency for science (FCT; Lisbon, Portugal).

The authors declare no competing interests.