Postsurgical Latent Pain Sensitization Is Driven by Descending Serotonergic Facilitation and Masked by µ-Opioid Receptor Constitutive Activity in the Rostral Ventromedial Medulla

Abstract

Following tissue injury, latent sensitization (LS) of nociceptive signaling can persist indefinitely, kept in remission by compensatory µ-opioid receptor constitutive activity (MOR_CA) in the dorsal horn of the spinal cord. To demonstrate LS, we conducted plantar incision in mice and then waited 3–4 weeks for hypersensitivity to resolve. At this time (remission), systemic administration of the opioid receptor antagonist/inverse agonist naltrexone reinstated mechanical and heat hypersensitivity. We first tested the hypothesis that LS extends to serotonergic neurons in the rostral ventral medulla (RVM) that convey pronociceptive input to the spinal cord. We report that in male and female mice, hypersensitivity was accompanied by increased Fos expression in serotonergic neurons of the RVM, abolished on chemogenetic inhibition of RVM 5-HT neurons, and blocked by intrathecal injection of the 5-HT₃R antagonist ondansetron; the 5-HT_2AR antagonist MDL-11 939 had no effect. Second, to test for MOR_CA, we microinjected the MOR inverse agonist d-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH2 (CTAP) and/or neutral opioid receptor antagonist 6β-naltrexol. Intra-RVM CTAP produced mechanical hypersensitivity at both hindpaws; 6β-naltrexol had no effect by itself, but blocked CTAP-induced hypersensitivity. This indicates that MOR_CA, rather than an opioid ligand-dependent mechanism, maintains LS in remission. We conclude that incision establishes LS in descending RVM 5-HT neurons that drives pronociceptive 5-HT₃R signaling in the dorsal horn, and this LS is tonically opposed by MOR_CA in the RVM. The 5-HT₃ receptor is a promising therapeutic target for the development of drugs to prevent the transition from acute to chronic postsurgical pain.

SIGNIFICANCE STATEMENT Surgery leads to latent pain sensitization and a compensatory state of endogenous pain control that is maintained long after tissue healing. Here, we show that either chemogenetic inhibition of serotonergic neuron activity in the RVM or pharmacological inhibition of 5-HT₃ receptor signaling at the spinal cord blocks behavioral signs of postsurgical latent sensitization. We conclude that MOR_CA in the RVM opposes descending serotonergic facilitation of LS and that the 5-HT₃ receptor is a promising therapeutic target for the development of drugs to prevent the transition from acute to chronic postsurgical pain.

Introduction

Chronic postsurgical pain affects ∼10% of patients and is often resistant to treatment (Glare et al., 2019). After an incision heals, a state of latent sensitization (LS) continues, whereby spinal nociceptive transmission in the dorsal horn (DH) remains within a state of heightened responsivity (Basu et al., 2021), kept in remission by compensatory signaling through inhibitory GPCRs including the neuropeptide Y Y1 receptor (Fu et al., 2019, 2020), kappa-opioid receptor (Custodio-Patsey et al., 2020; Basu et al., 2021), and µ-opioid receptor (MOR; Corder et al., 2013; Walwyn et al., 2016; Cooper et al., 2022). This endogenous analgesia lasts for long durations, in part because of MOR constitutive activity (MOR_CA). Even when delivered over a year after incision, administration of an opioid receptor antagonist or inverse agonist can unmask LS, precipitating a bilateral reinstatement of mechanical hypersensitivity and ongoing pain (Corder et al., 2013). The long duration of LS and MOR_CA could render studies in animal models particularly relevant to our understanding of the mechanisms that determine the initiation and maintenance of chronic postsurgical pain.

The ascending transmission of spinal nociceptive signals from the periphery to the brain are subject to powerful bulbospinal control. Supraspinal sites contribute to LS and MOR_CA, namely the central nucleus of the amygdala (Cooper et al., 2022). However, the contribution of other brain areas remains unclear. Of particular interest is the rostral ventromedial medulla (RVM). Pain-modulatory signals from higher centers in the brain converge on the RVM before descending to the DH (Porreca et al., 2002; Fields, 2004). Pathways from the RVM can be inhibitory or excitatory, and their net impact determines the modulation of spinal nociceptive signaling (Porreca et al., 2002; Fields, 2004; Chen and Heinricher, 2019). Tissue or nerve injury can shift this balance toward descending facilitation (Vera-Portocarrero et al., 2006; Bee and Dickenson, 2008; King et al., 2009; LaGraize et al., 2010; Wei et al., 2010; Wang et al., 2013). For example, in the setting of inflammation, disruption of pronociceptive signaling by MOR-expressing neurons in the RVM (RVM-MOR neurons) reduces inflammatory hyperalgesia (Kincaid et al., 2006; Cleary and Heinricher, 2013; Carr et al., 2014; Khasabov et al., 2017). RVM-MOR neurons likely mediate the well-known anti-hyperalgesic actions of exogenously administered morphine (Heinricher et al., 2009), but much less clear is their contribution to endogenous opioid receptor signaling such as MOR_CA.

RVM neurons that project to the dorsal horn include 5-HT cells within the raphe magus (RMg; Bowker et al., 1981; Skagerberg and Björklund, 1985). Optogenetic activation of medullary 5-HT neurons induced long-lasting mechanical and thermal hypersensitivity in uninjured mice (Cai et al., 2014), indicating their pronociceptive potential. This potential can be unleashed after nerve injury, with numerous studies suggesting that serotonin release from medullary raphe neurons targets spinal 5-HT_2A and 5-HT₃ receptors to facilitate behavioral signs of peripheral neuropathic pain (Suzuki et al., 2004; Van Steenwinckel et al., 2008; Thibault et al., 2008; Dogrul et al., 2009; Okubo et al., 2013; Kim et al., 2014; Bannister et al., 2015; Patel and Dickenson, 2018). In states of persistent inflammatory pain, however, the contribution of 5-HT₃R-mediated descending facilitation is unclear. Here, we address these questions using chemogenetic and pharmacological approaches to target the activity of medullary 5-HT neurons, MOR_CA in the RVM, and spinal 5-HT₃ receptors in a plantar incision model of latent pain sensitization.

Materials and Methods

Animals

All procedures were approved by the University of Pittsburgh Institutional Animal Care and Use Committee in accordance with American Veterinary Medical Association and International Association for the Study of Pain guidelines. FEV^cre (mice that express Cre recombinase in mid/hindbrain serotonergic neurons; B6.Cg-Tg(Fev-cre)1Esd/J; stock #012712; Scott et al., 2005) and Ai14 (mice that Cre-dependently express tdTomato; B6.Cg-Gt(ROSA)26Sor^{tm14(CAG-tdTomato)Hze}/J; stock #012712; Madisen et al., 2010) mice were obtained from The Jackson Laboratory and bred in our in-house colony. Male and female mice hemizygous for the FEV^cre transgene were used for chemogenetic behavioral experiments. For histology and in situ hybridization studies, FEV^cre mice were crossed with Ai14 mice. Wild-type C57BL/6 (used for all other behavioral pharmacology) and CD1 (used for all other histology) mice were obtained from Charles River Laboratories. Mice ages 6–16 weeks at the beginning of experiments were housed 2–4 per cage and maintained on a 12 h light/dark cycle at 20–22°C and 45 ± 10% relative humidity, with food and water provided ad libitum. Mice were handled and habituated to testing equipment for 30 min/day for 3 consecutive days before experimental manipulations and all procedures were performed during the light cycle (between 7:00 A.M. and 7:00 P.M.).

Viruses

For chemogenetic experiments, we used a control reporter virus that induced Cre-dependent expression of the fluorescent protein mCherry, adeno-associated virus (AAV2)-hSyn-DIO-mCherry (viral prep, lot no. v54505, 1.8 × 1013 vg/ml; catalog #50 459-AAV2, Addgene; RRID:Addgene_50459) or an experimental virus designed to express a neuron-specific, inhibitory G-coupled designer receptors exclusively activated by designer drug (DREADD), AAV2-hSyn-DIO-hM4D(G_i)-mCherry (viral prep, lot no. v68359, 1.5 × 1013 vg/ml; catalog #44362-AAV2, Addgene; RRID:Addgene_44362, a gift from Bryan Roth; Krashes et al., 2011). As described below, either control or experimental virus was targeted to the RVM of FEV^cre mice to generate RVM^FEV-mCherry or RVM^FEV-hM4Di mice, respectively. Viruses were stored in 5 µL aliquots at −80°C and thawed on ice immediately before injection.

Plantar incision model of postsurgical pain

Plantar incision was performed as previously described (Pogatzki and Raja, 2003; Basu et al., 2021). Anesthesia was induced with 5% isoflurane (Abbott Laboratories) and then maintained at 2% isoflurane. Ophthalmic ointment was applied to the eyes, and plantar skin was swabbed with chlorhexidine solution (Chloraprep, BD Healthcare). A 4 mm midline longitudinal incision was made through the glabrous skin of the left hindpaw from the interdigital pads to the heel. The plantaris muscle was separated from underlying tissue, and then a 4 mm midline longitudinal incision was made through the muscle with a no. 11 scalpel blade. The skin incision was closed with two 5-0 PDSII (polydioxanone) sutures (Ethicon), followed by topical application of Neosporin ointment (Johnson & Johnson). Sham-operated mice received isoflurane for the same duration as plantar incision model (PIM)–operated mice, but no incisions were made.

Stereotaxic surgery

Mice received carprofen (2 mg chewable tablet per mouse, per day, 24 h before surgery and for 2 d after, Bio-Serv; and a perioperative injection of buprenorphine (0.1 mg/kg, subcutaneous; Covetrus). Surgical anesthesia was induced with 5% isoflurane and maintained at 2% isoflurane. Mice were placed in a stereotaxic apparatus fitted with blunt mouse ear bars (Stoelting). Ophthalmic ointment (Fisher Scientific) was applied to the eyes, the scalp was shaved, and skin was swabbed with chlorhexidine solution. A midline skin incision exposed the cranium, and with a 0.7 mm dental burr bit, a hole was drilled (World Precision Instruments) above the nucleus RMg of the RVM (coordinates relative to bregma: AP, −5.8–6 mm; ML, 0 mm; DV, −5.6 mm), according to Paxinos and Franklin (2013). Mice were housed in pairs for a recovery period of at least 6–8 d before further experimental manipulations.

Cannulation surgeries were performed 2 weeks after incision and 1–2 weeks before behavioral pharmacology. A 26 Ga, 4.6 mm stainless steel guide cannula (catalog #C315G-SPC, Plastics One) was implanted 1 mm above the RMg. The guide cannula was affixed to the skull with two flat-head jeweler's screws (0–80 × 1/8 inches, Small Parts) and dental cement (RelyX Luting Plus Automix, 3M). Skin was then closed around the base of the cannula using three 5-0 PDSII sutures (Ethicon), followed by insertion of a 4.6 mm stylet (catalog #C315DC-SPC, Plastics One) into the guide cannula to prevent clogging.

Adeno-associated virus (AAV) microinjections were performed 1 week before incision. A 33 gauge needle (Plastics One) attached to a 1 µl microsyringe (Hamilton) with PE-50 tubing (Warner Instruments) was inserted slowly into the RMg (−5.6 mm DV) over 5 min, and then 300 nl of AAV was slowly injected over 5 min. The needle was left in place for a further 5–10 min to prevent backflow of solution up the needle tract and was then slowly retracted over a period of 5 min. Skin was then closed using three 5-0 PDSII sutures and cyanoacrylate glue (Vetbond, 3M).

Drug administration and experimental design

Intracranial drug infusion

Injections were performed using a 33 gauge injection cannula (catalog #C315I-SPC, Plastics One) that extended 1 mm beyond the tip of the guide cannula. The injection cannula was attached to flexible plastic tubing (catalog #C313C, Plastics One), backfilled with mineral oil, and connected to a microliter syringe(Hamilton). d-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH2 (CTAP, 0.3 µg/0.25 µL; Tocris Bioscience), a MOR-selective inverse agonist, was dissolved in sterile saline. 6β-naltrexol hydrate (3 µg/0.25 µL; Sigma-Aldrich, most often described as a neutral opioid receptor antagonist; Raehal et al., 2005; Sirohi et al., 2009; Lam et al., 2011) was dissolved in 10% DMSO in sterile saline. Reports claiming 6β-naltrexol to be an inverse agonist (Sally et al., 2010) were based on recombinant MOR overexpression assays in cell lines and may not recapitulate neutral antagonist activity as occurs in vivo. Drugs were slowly infused using a syringe pump (Harvard Apparatus) at a volume of 0.25 µL over 6 min. The injection cannula was left in place for a further 15 min to prevent backflow. Successful microinjection was confirmed by movement of a small air bubble within the mineral oil along the tubing. CTAP dose was based on our previous data using the intracranial route of administration (Cooper et al., 2022), and 6β-naltrexol dose was based on the molar ratio of CTOP (D-Phe-Cys-Tyr-D-Trp-Orn-Thr-Pen-Thr-NH2) to 6β-naltrexol required to inhibit reinstatement of mechanical hypersensitivity via the intrathecal (i.t.) route (10:1; Corder et al., 2013). Twenty-one days after incision, the first intra-RVM microinjection of drug or vehicle was conducted, and this was followed 7 d later with a crossover injection of vehicle or drug.

Intrathecal injection

A small patch of fur (∼20 × 20 mm) was shaved over the lumber spine. Mice were acclimated to manual restraint at the pelvic girdle within a towel to minimize stress. After insertion of a 30 gauge needle (attached to a 25 µL Hamilton syringe) between the L5 and L6 vertebrae, successful entry was indicated by observation of a tail flick (2/135 injections were excluded). Drug or its vehicle (5 µL) was slowly injected over 20 s into the intrathecal space, and then the needle was held in place for an additional 10 s to minimize backflow (Njoo et al., 2014). The 5-HT₃R antagonist ondansetron (Tocris Bioscience) or 5-HT_2AR antagonist MDL-11 939 (Tocris Bioscience) were dissolved in saline or 0.68% DMSO in saline, respectively. Dosages were selected based on previous literature using the intrathecal route of administration (Pehek et al., 2006; Thibault et al., 2008; Van Steenwinckel et al., 2008; Chang et al., 2013).

Systemic injection for chemogenetics

Clozapine-N-oxide (CNO; Tocris Bioscience) was dissolved in sterile saline at a dose of 3 mg/kg to achieve hM4D activation (Peirs et al., 2015) while minimizing clozapine-mediated adverse effects (Manvich et al., 2018). Mice were randomly allocated to one of four treatment groups [CNO or vehicle, intraperitoneal; naltrexone (NTX) or vehicle, subcutaneous at nape]. Twenty-one days after incision, mice received CNO or its saline vehicle injection followed 5 min later with naltrexone or vehicle, and then tested for mechanical sensitivity. This was followed 7 d later with a crossover injection of vehicle or naltrexone. In the same mice, 35 d later, injections were followed by assessment of heat hypersensitivity and then tested 5 d later with a crossover design.

Behavioral testing

All behavioral measurements were performed by an investigator blinded to experimental treatments by an assistant, who randomly assigned treatment groups.

Von Frey assessment of mechanical allodynia

Hindpaw 50% mechanical withdrawal thresholds were measured with a predefined set of eight von Frey (vF) monofilaments (0.008–6 g, Stoelting) using the up-down method (Chaplan et al., 1994). Mice were acclimated for at least 15 min within an acrylic box, opaque on all sides, atop an elevated wire mesh platform. The vF hair was applied to the proximal region of the glabrous skin at the plantar surface of the hindpaw, just lateral to the incision site. Each trial began with application of an intermediate filament (0.16 g), perpendicular to the skin, causing a slight bending, for 3 s. In case of a positive response (rapid withdrawal or licking of the paw within 3 s of removing the filament but ignoring normal ambulation or rearing), the next smallest filament was tested. In case of a negative response, the next larger filament was tested. Each trial continued until four measurements beyond the first change in response (i.e., no response then response, or vice versa) were taken. A 50% mechanical withdrawal threshold was calculated using the statistical method described by Dixon (1965).

Plantar radiant heat assay (Hargreaves test)

Heat sensitivity was assessed with a radiant heat assay (Ugo Basile). Up to seven mice were tested at a time on an elevated glass platform within acrylic boxes, 7 inches high × 15 inches wide × 35 inches long, transparent on one side to enable the experimenter to observe from the front. Following at least 25 min of acclimation, a radiant heat source was applied through the glass floor to the plantar surface of the hindpaw (Hargreaves et al., 1988). Latency to paw withdrawal was recorded, ignoring normal ambulation. Thermal stimulation was applied for no longer than 20 s to avoid tissue damage. Withdrawal latency was measured three times at 5 min intervals (5 min before, 0 and 5 min after the defined time point) and averaged.

Hotplate test

For chemogenetic studies, hotplate testing was used as an alternative assay of heat hypersensitivity because our pilot studies found that CNO induced a small, DREADD-independent change in thermoregulation when mice were in contact with the glass platform for extended periods of time. Mice were placed on a hotplate (Columbus Instruments) at 52.5°C, and the latency to response (jumping, licking, or rapid withdrawal) at either hindpaw was recorded. At this time, mice were returned to their home cage. Withdrawal latency was measured three times at 10 min intervals (10 min before, 0 and 10 min after the defined time point) and averaged.

Histology

Confirmation of cannulation sites

After completion of intra-RVM behavioral pharmacology experiments, mice were anesthetized with an overdose of pentobarbital (5 ml/kg, i.p.; Fatal Plus, Vortech Pharmaceuticals), perfused with 4% paraformaldehyde (PFA; Sigma-Aldrich), and then received an intra-RVM microinjection of 0.25 µL India ink. After 15 min for dye penetration, brains were removed, postfixed in 4% PFA at 4°C overnight, cryoprotected in 30% sucrose for a further 48 h, and then embedded in optimal cutting temperature (OCT) media (Tissue Tek, Andwin Scientific). Brains were sectioned on a cryostat (Cryostar NX70, Fisher Scientific) at 30 µm, collected on gelatinized slides, counterstained with cresyl violet, and then imaged. The location of staining was cross-referenced with a stereotaxic atlas (Paxinos and Franklin, 2013) to confirm injection site. In all mice, the center of cannula placements was found to be within 0.2 mm from the outer boundary of the RMg with India ink spreading into the RMg, and so all were considered to have been on target.

Fos immunohistochemistry

Twenty-one days after incision or sham surgery, mice were transcardially perfused with 4% PFA. Brains were collected, embedded in OCT, and sectioned on the cryostat at 40 µm coronal cryosections. Free-floating sections were collected in 0.1 m PBS. Six nonadjacent, evenly spaced sections spanning the range between Bregma −5.6 to −6.2 mm (Paxinos and Franklin, 2013) were arbitrarily selected from each mouse. Sections were washed in PBS, then blocked in PBS containing 3% normal goat serum (NGS; MP Biomedicals) and 0.3% Triton X-100 (VWR) for 1 h and then incubated for 18 h at room temperature with either anti-Fos (1:2000; polyclonal rabbit anti-c-Fos, catalog #226003, Synaptic Systems; RRID:AB_2231974) and anti-NeuN (1:1000; Alexa Fluor-488-conjugated mouse anti-NeuN, catalog #MAB377X, Millipore-Sigma; RRID:AB_2149209), or anti-Fos (1:2000; rabbit anti-phospho-c-Fos (Ser32), catalog #5348, Cell Signaling Technology; RRID:AB_10557109), diluted in 1% NGS and 0.3% Triton X-100. Following further washes in PBS, slides were air dried and coverslipped with Vectashield Hard Set Antifade Mounting Medium (Vector Laboratories).

Fluorescence in situ hybridization

FEV^cre::Ai14 mice were administered an overdose of pentobarbital. On cessation of heartbeat, brains were rapidly extracted, embedded in OCT, and frozen on dry ice. Brains were cryosectioned on a cryostat at 8 µm and mounted directly onto slides (Superfrost Plus, Fisher Scientific). Sections were fixed by immersion of slides in ice cold 4% PFA for 15 min and then dehydrated with increasing concentrations of ethanol (50, 70, then 100% for 5 min each). Fluorescence in situ hybridization (FISH) was performed using an RNAscope Multiplex Fluorescent Reagent Kit v2 (catalog #323100, Advanced Cell Diagnostics) following the manufacturer's protocol. Slides were pretreated for 15 min with protease (Advanced Cell Diagnostics), and then incubated and hybridized with Oprm1 mRNA probe (catalog #315841) for 2 h at 40°C in a humidified oven (HybEZ; Advanced Cell Diagnostics). Sections were incubated with three drops each of AMP1, AMP2, AMP3, then AMP4-FL amplification buffers for 30, 15, 30, and 15 min, respectively, at 40°C, with 2 min rinses in wash buffer after each incubation. Slides were then washed in 0.01 m PBS, air dried, and coverslipped with Vectashield Hard Set Antifade Mounting Medium with DAPI (Vector Laboratories).

Imaging

Sections throughout the RMg and raphe pallidus (RPa) were imaged with a Nikon Ti2 inverted epifluorescence microscope equipped with a motorized stage, 10×, 0.45 NA (used for brightfield confirmation of cannulation sites); 20×, 0.75 NA (used for Fos immunohistochemistry); and 40×, 0.95 NA (used for FISH objectives), and a Prime BSI camera (Photometrics). The same exposure time (80–500 ms) was used for all images captured in each channel. For Fos immunohistochemistry performed on FEV^cre::Ai14 tissue, 10–12 z-scans (3 µm separation) of a field of view containing the RVM were acquired. Image capture, stitching, and quantification were performed with NIS-Elements Advanced Research software version 5.02 (Nikon). Quantification of staining was conducted in the RMg and RPa. Anatomical landmarks and rostrocaudal coordinates (from bregma −5.6 to −6.2 mm) were referenced to a mouse brain atlas (Paxinos and Franklin, 2013). Throughout image acquisition and quantification, the investigator, although blind to treatment groups, adjusted brightness and contrast in the same manner for each image.

Quantification

The number of Fos-positive cell profiles were manually quantified in 4–8 mice per experimental group, excluding profiles that were largely outside the plane of view, clearly not representing a soma, or with fluorescence that is readily attributed to artifacts. For each section, a minimum fluorescence intensity was established by examining brainstem nuclei outside of RVM. Profiles with intensity below this threshold likely represented background/nonspecific immunostaining and so were not counted. Fos and tdTomato colocalization was quantified within z-stacks by scrolling back and forth in the z dimension to determine the z position with optimal focus and to determine whether fluorescence in each channel occurred in the same focal plane. A positive cell was defined as a Fos+ nucleus surrounded by a tdTomato+ soma in x, y, and z dimensions. To account for oversampling of Fos+ neuronal profiles in the z-axis, a correction factor was calculated using Abercrombie's formula (ratio of real number to observed number = T/T + h, where T is section thickness and h is mean diameter of objects; Guillery, 2002). Given a section thickness of 40 µm, a mean Fos+ neuronal nuclei diameter of 8.72 µm (determined by measuring the diameter of all Fos+ nuclei in three randomly selected sections from our dataset), a correction factor of 0.82 was applied to all cell counts. Five to six sections per mouse were counted and averaged, with n defined as one mouse. For quantification of FISH, an Oprm1 positive cell was identified by a minimum of four fluorescent puncta within the soma surrounding the nucleus (Snyder et al., 2018).

Statistical analyses

Statistical analyses were performed in Prism 8.1 (GraphPad). Immunohistochemical data were compared with unpaired t tests. Behavioral data were analyzed using two-way repeated measures (RM) ANOVA, examining the interaction of Treatment (incision or sham, and combinations of drugs or vehicle) and Time, unless otherwise specified. If ANOVA revealed a main effect, then Bonferroni's post hoc tests were conducted to compare treatment groups. The threshold for statistical significance was set at p < 0.05. For immunohistochemical studies (see Fig. 2), n represents a single mouse. All behavioral and immunohistochemical results are presented as mean ± SEM.

Figure 2.

Increased Fos expression in medullary raphe 5-HT neurons during NTX-induced reinstatement of hyperalgesia. A, Representative images demonstrate colocalization in the RMg and RPa of NeuN (green) and Fos (red) immunofluorescence 21 d after PIM (30 Fos+, NeuN+ cells) or Sham (9 Fos+, NeuN+ cells) surgery. No NeuN-negative, Fos+ neurons were observed. Scale bars: 100 µm; inset, 10 µm. B, Quantification of Fos+ (red) cells in the RMg and RPa. PIM increased RMg and RPa Fos expression compared with sham-operated mice (unpaired t test; *p < 0.05; n = 8 mice, 5–6 sections per mouse). C, Representative images demonstrating FEV-tdTomato (red, 62 cells) and Fos (green, 36 cells) colocalization (6 FEV-tdtomato+, Fos+ cells) in the RMg and RPa 21 d after incision and 2 h after systemic NTX (3 mg/kg) injection. Scale bars: 100 µm; inset, 10 µm. Images are maximum intensity projections of 11 z-scans. D, Quantification of cells coexpressing FEV-tdTomato and Fos in the RMg and RPa of male and female FEV^cre::Ai14 mice 21 d following incision and 2 h after systemic NTX (3 mg/kg) or saline injection (unpaired t test; ** p < 0.01; n = 4 mice, 5–6 sections per mouse). E, Representation images of FEV-tdTomato (red) and FISH of Oprm1 (green) mRNA in the RMg and RPa. Scale bars: 50 µm; inset, 10 µm. F, Quantification of percentage colocalization of FEV-tdTomato and Oprm1 mRNA in the RVM (n = 15 sections from 2 mice). G, Representative image of Oprm1 mRNA in the spinal dorsal horn. Inset, Cropped, enlarged images of boxed regions. Extended Data Figure 2-1A illustrates that Fos was colocalized with neuronal nuclei marker NeuN.

Results

MOR_CA in the RVM maintains LS in remission

Plantar incision produces mechanical hyperalgesia that peaks within 1–2 d and then gradually resolves over 14–21 d. At this point, LS is in a state of remission that is maintained by ongoing signaling from µ-opioid receptors in the dorsal horn (Corder et al., 2013) and amygdala (Cooper et al., 2022); however, the identity of additional critical brain regions remains a key gap in knowledge. An important supraspinal site in the regulation of chronic inflammatory pain is the RVM (Porreca et al., 2002). The experiments of Figure 1 investigated whether injury recruits MOR signaling in the RVM to maintain LS in remission. Fourteen days after incision, cannulae were inserted into the RVM of male mice (Fig. 1A,B). Incision but not sham surgery evoked a mechanical hypersensitivity that peaked at 2 d and resolved within 21 d (Fig. 1C). Twenty-one days after surgery, mice received an intra-RVM microinjection of the MOR inverse agonist CTAP (0.3 µg/0.25 µL) or vehicle (saline). Intra-RVM CTAP but not saline reinstated mechanical hypersensitivity in incision but not sham mice (Time × Treatment interaction, F_(15,120) = 1.861, p = 0.039, n = 7).

Figure 1.

MOR_CA in the RVM maintains LS in remission. A, Schematic illustration representing the timeline of experimental procedures. B–E, Representative section with postmortem India ink injection (top); location of injection sites in the RVM (bottom) for experiments shown in C (left) and D and E (right). Mechanical thresholds at the ipsilateral hindpaw following PIM or sham surgery (left), and following resolution of PIM-induced hypersensitivity, intra-RVM injection of the MOR inverse agonist CTAP (0.3 µg/0.25 µL) or vehicle (saline; right, n = 7, C). Mechanical thresholds at the ipsilateral (D) and contralateral (E) hindpaw over 21 d following incision (left) and after intra-RVM injection (right) of CTAP (0.3 µg/0.25 µL), the neutral opioid antagonist 6β-naltrexol (3 µg/0.25 µL), both CTAP and 6β-naltrexol, or vehicle (10% DMSO in saline; n = 8). Two-way RM ANOVAs with Bonferroni's post tests; *p < 0.05, **p < 0.01, ***p < 0.001.

To determine whether ligand-dependent or ligand-independent opioid signaling in the RVM maintains LS in remission, we injected either CTAP (0.3 µg/0.25 µL), the neutral opioid antagonist 6β-naltrexol (3 µg/0.25 µL), a combination of both, or vehicle (10% DMSO in saline) into the RVM. Mechanical sensitivity was assessed at both hindpaws. As illustrated in Figure 1, D and E, incision induced a mechanical hypersensitivity in the ipsilateral but not contralateral hindpaw that resolved within 21 d (Time × Side interaction, F_(4,120) = 35.39, p < 0.001, n = 16). When these animals were injected on postsurgical day 21, CTAP but not saline reinstated mechanical hypersensitivity at the ipsilateral hindpaw (Fig. 1D, right) and produced robust hypersensitivity on the contralateral hindpaw as well (Fig. 1E). 6β-naltrexol had no effect when injected alone, indicating that latent sensitization is not suppressed by ligand-dependent MOR signaling in the RVM. By contrast, 6β-naltrexol blocked the hypersensitivity produced by CTAP (ipsilateral, Time × Treatment interaction, F_(15,140) = 2.964, p < 0.001; contralateral, Time × Treatment interaction, F_(4,120) = 2.182, p < 0.001, n = 8), indicating that latent sensitization is suppressed by ligand-independent MOR_CA.

Increased Fos expression in medullary raphe 5-HT neurons during NTX-induced reinstatement of hyperalgesia

Transient application of noxious heat alters the firing of RVM neurons (Heinricher et al., 1989), and persistent chemical nociception evokes neurotransmitter release in the RVM (Taylor and Basbaum, 1995). Furthermore, the complete Freund's adjuvant (CFA) model of inflammatory pain is associated with facilitation of neuronal activity in the RVM (Ren and Dubner, 2002; Heinricher, 2016); however, these experiments were limited to the initial stages of inflammation, typically 1–3 d after induction. To test the hypothesis that incision can produce a longer-lasting neuronal sensitization that is more reflective of the time course of chronic pain, we waited 21 d after incision and then assessed Fos expression as a marker of neuronal activity (Bullitt, 1990) as illustrated in Figure 2. Extended Data Figure 2-1A and Figure 2A illustrate that Fos was colocalized with neuronal nuclei marker NeuN. The number of Fos-positive neurons increased after incision as compared with sham surgery (Fig. 2B; unpaired t test, t₍₁₄₎ = 2.66, p = 0.019, n = 8), indicating a long-lasting increase in RVM neuron activation.

Extended Data Figure 2-1

Colocalization of Fos with NeuN and FEV-tdTomato through the z-axis, Related to Figure 2. A, Colocalization of Fos with NeuN at three z-positions at 6 µm intervals in the RMg. White arrow indicates a neuron positive for both Fos and NeuN immunofluorescence in focus at 0 µm and out of focus ±6 µm in the z dimension. B, Colocalization of Fos with FEV-tdTom at three z positions at 3 µm intervals in the RMg. White arrow indicates an example of a Fos and FEV-tdTom positive neuron in focus at 0 µm, and out of focus ±3 µm in the z dimension. Open arrow indicates a Fos-positive neuron in focus at −3 µm but without FEV-tdTom fluorescence definitively surrounding the nucleus in the same focal plane, thus classified as FEV-tdTom negative. Scale bars: 25 µm. Download Figure 2-1, TIF file ^{(1.6MB, tif)} .

Descending serotonergic facilitation arising from the RVM drives chronic neuropathic pain states (Suzuki et al., 2004; Dogrul et al., 2009; Kim et al., 2014; Bannister et al., 2015; Patel and Dickenson, 2018). To test the hypotheses that serotonergic neurons are activated during reinstatement of hypersensitivity, we examined Fos expression in the medullary raphe (RMg and RPa) of FEV^cre::Ai14 mice (Fig. 2C, Extended Data Fig. 2-1B). Twenty-one days after incision, mice received a subcutaneous injection of NTX (3 mg/kg) or vehicle (saline) and were then allowed a 2 h waiting period to allow Fos expression. NTX increased Fos in serotonergic (FEV-tdTomato+) neurons compared with mice that received saline (Fig. 2D; unpaired t test, (t₍₆₎ = 5.301, p = 0.002, n = 8).

The immunohistochemical evidence for coexpression of MOR and 5-HT in the RVM is contradictory (Gao and Mason, 2000; Sikandar et al., 2012). To readdress this question, we conducted FISH for Oprm1 mRNA in the RMg and RPa of FEV^cre::Ai14 mice. Figure 2, E and F, illustrates that 58.0 ± 3.7% FEV-tdTomato+ neurons expressed Oprm1 mRNA, and 50.7 ± 4.6% Oprm1+ neurons expressed FEV-tdTomato. As a positive control, we also examined Oprm1 mRNA expression in the spinal cord. As illustrated in Figure 2G, Oprm1 mRNA was particularly enriched in the superficial laminae as previously described (Wang et al., 2021). These data support the feasibility of serotonergic neurons as a target for inhibition by MOR_CA.

Chemogenetic inhibition of RVM 5-HT neurons prevents NTX-induced reinstatement of hyperalgesia

Focal lesioning and local anesthesia studies suggest that descending facilitation arising from the RVM contributes to early hypersensitivity on cutaneous inflammation (Urban et al., 1996; Kincaid et al., 2006; Tillu et al., 2008; Carr et al., 2014). However, interpretations of these studies can be confounded by disruption of axons of passage or compensatory changes. Further, these studies did not examine the contribution of RVM 5-HT neurons in a model of long-lasting inflammatory pain. To address these gaps, we chose chemogenetics in our incision LS model as an approach to selectively inhibit RVM 5-HT neurons with temporal control (Fig. 3). As illustrated by the timeline in Figure 3A, we injected a Cre-dependent virus expressing either the inhibitory DREADD hM4Di (AAV2-hSyn-DIO-hM4D(Gi)-mCherry) or a control virus expressing mCherry (AAV2-hSyn-DIO-mCherry) into the RVM of FEV^cre mice. Figure 3B confirmed that hM4D-mCherry expression was largely restricted to RVM 5-HT (Tph2+) neurons, 87.16 ± 3.96% of hM4D-mCherry-expressing neurons colabeled with Tph2 immunofluorescence (n = 4 mice). One week after virus injection, we conducted incision or sham surgery. Twenty-one days later, we first administered CNO (3 mg/kg, i.p.) and then challenged the mice with either NTX (3 mg/kg, s.c.) or vehicle (saline). As illustrated in Figure 3, C and E, incision-induced mechanical and heat hypersensitivity at the ipsilateral hindpaw resolved within 21 d (mechanical, Time × Incision interaction, F_(20,176) = 13.81, p < 0.001; heat, Time × Incision interaction, F_(4,44) = 28.22, p < 0.001, n = 8 sham or 12 PIM RVM^FEV-hM4Di and 5 RVM^FEV-mCherry controls). CNO but not its vehicle abolished NTX-induced reinstatement of mechanical hypersensitivity at the ipsilateral paw of mice with incision but not in (1) sham-operated mice, (2) those that received intra-RVM injection of mCherry control virus, nor (3) mice that did not receive NTX (Fig. 3D, ipsilateral, Time × Treatment interaction, F_(25,190) = 6.074, p < 0.001; Fig. 3E, contralateral, Time × Treatment interaction, F_(25,190) = 7.682, p < 0.001; Time × Treatment interaction, F_(5,37) = 21.46, p < 0.001; both, n = 7–8 RVM^FEV-hM4Di, 5 RVM^FEV-mCherry controls). These data demonstrate that RVM 5-HT neurons maintain LS.

Figure 3.

Chemogenetic inhibition of RVM 5-HT neurons prevents NTX-induced reinstatement of hyperalgesia. A, Schematic illustration representing timeline of experimental procedures. FEV^cre mice received intra-RVM injection of AAV2-hSyn-DIO-hM4Di-mCherry (RVM^FEV-hM4Di) or AAV2-hSyn-DIO-mCherry control (RVM^FEV-mCherry). B, Representative image showing colocalization of AAV2-hSyn-DIO-hM4Di-mCherry expression (red) and Tph2 immunofluorescence (green) in the RVM of FEV^cre mice. Scale bars: 100 µm. C, D, Mechanical thresholds at the ipsilateral (C) and contralateral (D) hindpaws following PIM or sham surgery in FEV^cre mice and effect of CNO or saline administration on NTX-induced reinstatement of mechanical allodynia. PIM-induced hypersensitivity in the ipsilateral hindpaw resolved after 21 d (left, n = 8 sham or 12 PIM RVM^FEV-hM4Di and 5 RVM^FEV-mCherry controls). CNO prevented NTX-induced reinstatement of hypersensitivity in RVM^FEV-hM4Di mice (right, n = 7–8 RVM^FEV-hM4Di, 5 RVM^FEV-mCherry controls). E, Hotplate testing to measure latency of withdrawal in ipsilateral hindpaw 2 and 21 d following PIM or sham surgery (n = 8 sham or 12 PIM RVM^FEV-hM4Di and 5 RVM^FEV-mCherry controls). F, Hotplate testing to measure the effect of CNO or saline administration on NTX-induced reinstatement of heat allodynia of RVM^FEV-hM4Di mice (n = 7–8 RVM^FEV-hM4Di, 5 RVM^FEV-mCherry controls). Two-way RM ANOVAs with Bonferroni's post tests; *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant.

Spinal 5-HT3 but not 5-HT2A receptors contribute to latent sensitization

Both 5-HT_2A and 5-HT₃ receptors contribute to descending serotonergic facilitation of spinal nociceptive signaling and the maintenance of the early stages of injury-induced hyperalgesia (Dogrul et al., 2009; Alba-Delgado et al., 2018; Patel and Dickenson, 2018); here, we determined the contribution of these receptors to longer-lasting hyperalgesia (Fig. 4). As illustrated by the timeline in Figure 4A, we conducted incision or sham surgery and then waited 21–28 d for remission. Incision produced mechanical and heat hypersensitivity at the ipsilateral paw that resolved within 21 d (Fig. 4B, mechanical, Time × Incision interaction, F_(4,84) = 26.31, p < 0.001, n = 9 sham or 14 PIM; Fig. 4D, heat, Time × Incision interaction, F_(2,44) = 46.49, p < 0.001, n = 8 sham or 16 PIM; Fig. 4E, mechanical, Time × Incision interaction, F_(4,112) = 41.29, p < 0.001, n = 15; Fig. 4G, heat, Time × Incision interaction, F_(2,54) = 51.67, p < 0.001, n = 9 sham or 20 PIM). We then intrathecally administered the 5-HT₃R antagonist ondansetron (10 µg/5 µl) or its vehicle (saline), and in a separate study, the 5-HT_2AR antagonist MDL-11 939 (0.5 µg/5 µl) or its vehicle (0.68% DMSO in saline). Five minutes later, we injected NTX (3 mg/kg, s.c.) or vehicle (saline). As illustrated in Figure 4, B–D, NTX led to the reinstatement of mechanical and heat hypersensitivity at the ipsilateral hindpaw, as well as contralateral mechanical hypersensitivity.

Figure 4.

Spinal 5-HT₃Rs but not 5-HT_2ARs antagonists prevent NTX-induced reinstatement of mechanical and heat hypersensitivity. A, Schematic illustration representing timeline of experimental procedure. B, C, Mechanical thresholds at the ipsilateral (B) and contralateral (C) hindpaws following PIM or sham surgery (left; n = 9 sham or 14 PIM). Effect of i.t. ondansetron (5-HT₃R antagonist) or saline on NTX-induced reinstatement of mechanical allodynia at the ipsilateral (B) and contralateral (C) hindpaws (right, n = 6–8). D, Hargreaves testing to measure latency of withdrawal in ipsilateral hindpaw 2 and 21 d following PIM or sham surgery (left, n = 8 sham or 16 PIM). Effect of i.t. ondansetron or saline on NTX-induced reinstatement of heat allodynia in ipsilateral hindpaw (right, n = 6). E, F, Progression of mechanical allodynia at the ipsilateral (E) and contralateral (F) hindpaw following PIM or sham surgery (left, n = 15). Effect of i.t. MDL-11 939 (5-HT_2AR antagonist) or saline on NTX-induced reinstatement of mechanical allodynia at the ipsilateral (E) and contralateral (F) hindpaws (right, n = 8). G, Hargreaves testing to measure latency of withdrawal in ipsilateral hindpaw 2 and 21 d following PIM or sham surgery (left, n = 9 sham or 20 PIM). Effect of i.t. MDL-11 939 or saline on NTX-induced reinstatement of heat allodynia in ipsilateral hindpaw (right, n = 5–7). Two-way RM ANOVAs with Bonferroni's post tests; *p < 0.05, **p < 0.01, ***p < 0.001.

Ondansetron

Two-way RM ANOVA with Bonferroni's post tests revealed that ondansetron blocked NTX-induced reinstatement of mechanical hypersensitivity at the ipsilateral paw (Time × Treatment interaction, F_(20,160) = 2.39, p = 0.001, n = 6–8; Fig. 4B) and the contralateral paw (Time × Treatment interaction, F_(20,160) = 2.45, p = 0.001, n = 6–8; Fig. 4C) as well as heat hypersensitivity (Time × Treatment interaction, F_(16,100) = 5.42, p < 0.001, n = 6; Fig. 4D). Ondansetron did not change sensitivity in sham-operated mice nor in PIM mice that received saline vehicle.

MDL-11 939

In contrast to ondansetron, MDL-11 939 did not change NTX-induced reinstatement of mechanical hypersensitivity at the ipsilateral paw (Time × Treatment interaction, F_(20,175) = 11.46, p < 0.001, n = 8; Bonferroni's post tests comparing PIM plus NTX plus Sal and PIM plus NTX plus MDL, p > 0.9 at all time points; Fig. 4E), the contralateral paw (Time × Treatment interaction, F_(20,175) = 6.88, p < 0.001, n = 8; Bonferroni's post tests comparing PIM plus NTX plus Sal and PIM plus NTX plus MDL, p > 0.9 at all time points; Fig. 4F), nor heat hypersensitivity (Time × Treatment interaction, F_(16,92) = 3.43, p < 0.001, n = 5–7; Bonferroni's post tests comparing PIM plus NTX plus Sal and PIM plus NTX plus MDL, p > 0.2 at all time points; Fig. 4F).

Discussion

Incision produces a long-lasting latent sensitization of RVM 5-HT neurons

Our study is the first to examine the activity of RVM neurons 3 weeks after surgery, during the remission phase of LS. We found that the number of RVM neurons expressing Fos was greater in PIM mice than in sham controls, suggestive of a tonic increase in activity, even in the absence of overt pain-like behavior. Furthermore, we observed greater Fos expression in FEV-tdTomato-positive neurons during NTX-induced reinstatement of hyperalgesia, leading us to conclude that incision produces a long-lasting latent sensitization of RVM 5-HT neurons. These results in our LS model of chronic postoperative pain extend previous studies that had been restricted to noxious stimulus-evoked responses in uninjured animals or in short-term models of persistent pain hypersensitivity (Heinricher, 2016).

The RVM contains three classes of neurons based on their electrophysiological responses to transient noxious stimuli. ON cells are pronociceptive MOR-expressing RVM neurons and display an increase in firing rate before or at the onset of nocifensive behaviors, OFF cells display a transient pause in firing, and neutral cells display no change in firing rate (Fields et al., 1983; Chen and Heinricher, 2019). Because the original hypothesis that MOR and 5-HT provided molecular identification of the ON cell and neutral cell populations, respectively (Fields, 1992; Potrebic et al., 1994; Gao and Mason, 2000), later studies have suggested a more heterogenous distribution (Sikandar et al., 2012). Given that cre expression in FEV^cre mice faithfully recapitulates hindbrain serotonergic neuron populations (Scott et al., 2005), and the molecular identity of ON cells includes expression of MOR (Heinricher et al., 1992), our finding that more than 50% of Oprm1-expressing profiles coexpress FEV-tdTomato supports the idea that 5-HT RVM neurons represent not only neutral cells but also a subpopulation of MOR-expressing ON cells. Further studies are needed to determine whether increased neuronal activity reflects an engagement of LS mechanisms in molecularly defined ON, OFF, and neutral cells.

RVM 5-HT neurons maintain the LS that is masked by endogenous opioid receptor activity

We found that chemogenetic silencing of RVM 5-HT neurons prevented NTX-induced reinstatement of mechanical and heat hypersensitivity in our LS model of chronic postoperative pain. These results are consistent with and extend the work of Carr et al. (2014), who reported that ablation of descending CNS serotonergic neurons with intrathecal 5,7-dihydroxytryptamine partially reduced mechanical hypersensitivity at early time points following ankle injection of CFA; in contrast to this study, we observed complete inhibition of mechanical hypersensitivity at much later time points in a model that more closely mimics the time course of chronic pain. We conclude that RVM 5-HT neurons maintain the LS that is masked by endogenous opioid receptor activity.

Optogenetic activation of RVM 5-HT neurons induces mechanical and thermal hypersensitivity in uninjured mice (Cai et al., 2014). By contrast, our control experiments revealed that chemogenetic inhibition by itself did not increase mechanical or heat hypersensitivity. This indicates that RVM 5-HT neurons do not exert tonic pain inhibition, including during the remission phase of LS.

Spinal 5-HT₃ receptors contribute to latent sensitization of postsurgical pain

We show for the first time that intrathecal injection of the 5-HT₃R antagonist ondansetron blocked NTX-induced reinstatement of both mechanical and heat hypersensitivity when tested 3 weeks after plantar incision. We conclude that spinal 5-HT₃ receptors contribute to latent sensitization of postsurgical pain. This extends what has previously been observed in rodent models of neuropathic pain, where intrathecal ondansetron reduced the mechanical and thermal hypersensitivity and sensitization of dorsal horn neurons following peripheral nerve injury (Suzuki et al., 2004; Dogrul et al., 2009; Kim et al., 2014; Bannister et al., 2015; Patel and Dickenson, 2018). Furthermore, interruption of 5-HT₃R signaling with either global 5-HT₃R knock-out (Zeitz et al., 2002) or shRNA interference of tryptophan hydroxylase-2 (Wei et al., 2010) reduced licking behavior and/or dorsal horn neuronal firing in the intraplantar formalin test. On the other hand, Rahman et al. (2004) reported no effect of ondansetron in the intraplantar carrageenan model of early inflammatory pain, and so it appears that spinal 5-HT₃ receptors maintain neuropathic pain, acute ongoing pain, and long-lasting postsurgical pain, but not short-term inflammatory pain.

Ondansetron blocked NTX-induced reinstatement of hypersensitivity at both hindpaws, ipsilateral and contralateral to unilateral plantar incision. This is consistent with the idea that 5-HT₃R signaling contributes to mirror image pain. Similarly, ondansetron restored diffuse noxious inhibitory controls (DNIC) following nerve injury (Bannister et al., 2015), and intra-RVM injection of lidocaine restored DNIC in the setting of medication-overuse headache (Okada-Ogawa et al., 2009). Further studies measuring forepaw hyperalgesia are needed to test the hypothesis that 5-HT₃R signaling maintains widespread latent sensitization of postsurgical pain.

The RVM 5-HT neuron→spinal 5-HT₃R pathway is just one of many descending pain facilitatory mechanisms (Millan, 2002). Others include descending GABAergic disinhibition (François et al., 2017) and α_1AR-mediated noradrenergic pronociceptive signaling (Taylor and Westlund, 2017; Kohro et al., 2020). Future studies are needed to determine the contribution of these systems to LS.

Spinal 5-HT_2A receptors do not contribute to latent sensitization of postsurgical pain

The 5-HT_2AR antagonist MDL-11 939 did not change NTX-induced reinstatement of mechanical or heat hypersensitivity when tested 3 weeks after plantar incision, consistent with the lack of effect of the 5-HT_2AR antagonist ketanserin on noxious mechanical or heat stimulus-evoked firing of hypothalamic wide dynamic range neurons in normal or neuropathic rats (Patel and Dickenson, 2018). By contrast, others report that intrathecal injection of 5-HT_2A receptor antagonists blocked mechanical hypersensitivity, thermal hypersensitivity, and/or dorsal horn neuronal firing in models of trigeminal nerve injury (Okubo et al., 2013), chemotherapeutic drug administration (Thibault et al., 2008), HIV (Van Steenwinckel et al., 2008), or facial inflammation (Alba-Delgado et al., 2018). Thus, the contribution of spinal 5-HT_2AR signaling may depend on the type (neuropathic vs inflammatory) and duration (hours vs weeks) of the model, as well as modality of hypersensitivity. We conclude that spinal 5-HT_2A receptors do not contribute to long-lasting postsurgical latent pain sensitization.

Incision establishes MOR_CA in the RVM

We report here that microinjection of CTAP into the RVM reinstated hypersensitivity. Our data are consistent with those of De Felice et al. (2011) who reported that a subset of rats displayed no pain-like behavior following spinal nerve ligation, and in these animals, intra-RVM lidocaine induced mechanical hypersensitivity, that is, inhibition of inhibitory RVM signaling unmasked hypersensitivity during latent sensitization. We conclude that injury engages endogenous inhibitory MOR activity within the RVM to maintain LS in a state of remission. This MOR activity could be driven by a ligand-dependent mechanism involving tonic opioid release. Indeed, endogenous opioid peptide signaling in the RVM is integral to the descending inhibitory control of transient nociception. For example, RVM injection of naltrexone blocks the antinociception produced by intra-PAG microinjection of morphine (Kiefel et al., 1993). However, the contribution of endogenous opioidergic mechanisms in the RVM toward the control of injury-induced hyperalgesia is much less clear. For example, MOR signaling in the RVM might not contribute to hyperalgesia in the CFA model of inflammatory pain (Hurley and Hammond, 2001). Here, in the setting of incision, we present two key pieces of data that promote the idea that MOR_CA, rather than opioid release, tonically inhibits postsurgical pain. First, intra-RVM administration of 6β-naltrexol (a neutral opioid receptor antagonist with no intrinsic activity) did not reinstate hypersensitivity. Second, coadministration of 6β-naltrexol prevented CTAP-induced reinstatement of hypersensitivity, arguing that CTAP acts as an inverse agonist with intrinsic activity at MOR. Further ruling out a contribution of endogenous opioids comes from studies in opioid peptide knock-out mice (Walwyn et al., 2016). Germline deletion of pro-enkephalin, pro-endorphin, or pro-dynorphin did not prevent the reinstatement of hypersensitivity that was triggered by systemic blockade of opioid receptors with the CNS-penetrant naloxone. We conclude that injury triggers MOR_CA not only at the dorsal horn of the spinal cord as previously described (Corder et al., 2013; Walwyn et al., 2016) but also at the RVM.

Our use of in vivo brain or intrathecal microinjections precludes the knowledge of opioid or 5-HT₃ receptor antagonist concentrations at their receptors. As a result, and given that concentrations of compounds were several times their IC₅₀ in the injection solution, it is possible that nonspecific receptor activation may have contributed to our observed behavioral effects, and our results should be interpreted with this in mind. However, CTAP, ondansetron, and MDL-11 939 are potent, selective antagonists of MOR (Kramer et al., 1989), 5-HT₃ (Thompson and Lummis, 2006) and 5-HT_2ARs (Pehek et al., 2006), respectively.

Conclusion

As schematized in Figure 5, we conclude that plantar incision establishes acute hypersensitivity that gradually resolves over 3 weeks but is replaced by a latent sensitization that is tonically masked by MOR_CA in the RVM. Latent postsurgical pain can be revealed by administering opioid receptor inverse agonists. Further RVM chemogenetic and intrathecal pharmacology studies then revealed that a bilateral descending serotonergic facilitatory pathway mediates LS and is recruited to induce mechanical and thermal hypersensitivity. This may have translational significance as clinical trials indicate that NTX-induced hypersensitivity might develop in humans (Pereira et al., 2015; Springborg et al., 2020) and could conceivably contribute to episodic hyperalgesia following disruption of endogenous opioid receptor activity such as occurs during stress (Taylor and Corder, 2014) and generalized pain syndromes such as fibromyalgia and irritable bowel syndrome (Reichling and Levine, 2009). 5-HT₃R antagonists have yielded disappointing results in clinical trials for neuropathic pain states (McCleane et al., 2003; Tuveson et al., 2011) possibly because of a lack of CNS availability following intravenous administration (Chiang et al., 2021). However, if further research indicates that LS contributes to the pathogenesis of chronic pain states, then this would encourage future studies to determine whether 5-HT₃R antagonists might be used as pharmacotherapy for chronic inflammatory pain states that rely on LS.

Figure 5.

RVM MOR_CA maintains latent pain sensitization in remission. A, In the absence of injury, the influence of rostral RVM-mediated descending serotonergic input to the DH is minimal (dotted gray line). B, Soon after injury, descending facilitation of spinal nociceptive processing predominates, leading to unilateral hypersensitivity (red arrow). C, Over time, latent sensitization persists (dotted red line) but is masked and kept in remission by RVM MOR_CA. D, Focal or systemic injection of an opioid inverse agonist such as NTX inhibits MOR_CA, unmasking (disinhibiting) descending 5-HT₃ receptor-mediated facilitation, leading to widespread pain reinstatement.