cAMP signaling through protein kinase A and Epac2 induces substance P release in the rat spinal cord

Abstract

Using neurokinin 1 receptor (NK1R) internalization to measure of substance P release in rat spinal cord slices, we found that it was induced by the adenylyl cyclase (AC) activator forskolin, by the protein kinase A (PKA) activators 6-Bnz-cAMP and 8-Br-cAMP, and by the activator of exchange protein activated by cAMP (Epac) 8-pCPT-2-O-Me-cAMP (CPTOMe-cAMP). Conversely, AC and PKA inhibitors decreased substance P release induced by electrical stimulation of the dorsal root. Therefore, the cAMP signaling pathway mediates substance P release in the dorsal horn. The effects of forskolin and 6-Bnz-cAMP were not additive with NMDA-induced substance P release and were decreased by the NMDA receptor blocker MK-801. In cultured dorsal horn neurons, forskolin increased NMDA-induced Ca²⁺ entry and the phosphorylation of the NR1 and NR2B subunits of the NMDA receptor. Therefore, cAMP-induced substance P release is mediated by the activating phosphorylation by PKA of NMDA receptors. Voltage-gated Ca²⁺ channels, but not by TRPV1 or TRPA1, also contributed to cAMP-induced substance P release. Activation of PKA was required for the effects of forskolin and the three cAMP analogs. Epac2 contributed to the effects of forskolin and CPTOMe-cAMP, signaling through a Raf - mitogen-activated protein kinase pathway to activate Ca²⁺ channels. Epac1 inhibitors induced NK1R internalization independently of substance P release. In rats with latent sensitization to pain, the effect of 6-Bnz-cAMP was unchanged, whereas the effect of forskolin was decreased due to the loss of the stimulatory effect of Epac2. Hence, substance P release induced by cAMP decreases during pain hypersensitivity.

Graphical Abstract

Diagram of the proposed pathways by which cAMP controls substance P (SP) release.

Adenylyl cyclase (AC) produces cAMP, which then activates protein kinase A, Epac1 and Epac2. PKA activates Cav2.2 channels (N type), NMDA receptors (NMDAR) and mTOR, all of which contribute to SP release. Epac2 signals through Rap, Raf-1 kinase, MEK and MAPK, leading to the activation of Cav1, Cav2.1 and Cav3.2 channels (L, P, Q and T types), which also contribute to SP release. Epac1 directly inhibits NK1R internalization.

1. Introduction

Recent models of the nociceptive circuitry in the rodent dorsal horn show that neurons in lamina I expressing neurokinin 1 receptors (NK1Rs) serve as a control node for pain transmission: they receive signals from primary afferents and interneurons (Haring et al., 2018; Peirs and Seal, 2016; Sathyamurthy et al., 2018; Yasaka et al., 2010; Yasaka et al., 2014) and send them to the thalamus and the parabrachial nucleus (Todd et al., 2000; Todd et al., 2002). When activated by their main agonist, substance P, NK1Rs increase the excitability of these neurons (Murase and Randic, 1984; Murase et al., 1986) and induce long-term potentiation of their synapses with primary afferents (Ikeda et al., 2003; Liu and Sandkuhler, 1997; Liu and Sandkuhler, 1998). This suggests that NK1Rs are involved in pain hypersensitivity. Indeed, we recently found that NK1Rs are required for the maintenance of latent sensitization to pain (Chen and Marvizon, 2020a).

Substance P is released primarily from the PEP1 population of peptidergic primary afferents (Usoskin et al., 2015), but also from a type of excitatory interneurons that receive monosynaptic input from C-fibers (Dickie et al., 2019; Haring et al., 2018; Sathyamurthy et al., 2018). Since the intensity of nociceptive signals is encoded by the firing frequency of primary afferents (Lever et al., 2001) and substance P release increases with their firing frequency (Adelson et al., 2009; Go and Yaksh, 1987; Marvizon et al., 1997), stronger nociceptive signals elicit more substance P release (Allen et al., 1997). The conventional view is that high-frequency action potentials arriving at the presynaptic terminals cause a depolarization that activates voltage-gated Ca²⁺ (Cav) channels; then, the ensuing entry of Ca²⁺ triggers substance P release. However, other mechanisms for Ca²⁺ entry into primary afferents terminals can induce substance P release, including TRPV1 (Lao et al., 2003; Marvizon et al., 2003) and NMDA receptors (Chen et al., 2014; Chen et al., 2010; Liu et al., 1997; Malcangio et al., 1998; Marvizon et al., 1997). Conversely, substance P release is inhibited by Gi/o-coupled receptors present in primary afferent terminals, including μ-opioid receptors (Kondo et al., 2005; Zhang et al., 2010), δ-opioid receptors (Kondo et al., 2005; Zachariou and Goldstein, 1996a), κ-opioid receptors (Zachariou and Goldstein, 1996b), GABA_B receptors (Malcangio and Bowery, 1993; Marvizon et al., 1999) and α_2A adrenergic receptors (Ossipov et al., 1990; Zhang et al., 2010).

We recently reported (Chen et al., 2018a) that μ-opioid receptors inhibit substance P release not only by inactivating Cav channels but also by inhibiting adenylyl cyclase (AC). We hypothesized that the AC - protein kinase A (PKA) pathway induces substance P release through the activating phosphorylation of NMDA receptors (Lau et al., 2009; Murphy et al., 2014). We tested this hypothesis by determining if activating AC and PKA in spinal cord slices induces NK1R internalization.

Latent sensitization is a model in which certain injuries put the animal in a long-term state of hyperalgesia that is continuously suppressed by opioid and α_2A adrenergic receptors (Marvizon et al., 2015; Taylor and Corder, 2014). In it there is a continuous release of substance P, which increases after blocking μ-opioid receptors (Chen and Marvizón, 2020a). Latent sensitization is mediated by AC and NMDA receptors (Corder et al., 2013). Hence, we also determined whether AC and PKA are involved in this sustained release of substance P during latent sensitization.

2. Material and methods

2.1. Animals

All animal procedures were approved by the Institutional Animal Care and Use Committee of the Veteran Affairs Greater Los Angeles Healthcare System and the Animal Research Committee of UCLA, and conform to NIH guidelines. Efforts were made to minimize the number of animals used and their distress. Rats used were male adult (2–4 months old) Sprague-Dawley (Harlan, Indianapolis, IND).

2.2. Chemicals

ω-Agatoxin, AM-0902, AMG9810, AZ-628, 1,9-dideoxyforskolin, 6-Bnz-cAMP, 8-Br-cAMP, brain-derived neurotrophic factor (BDNF), CE3F4, 8-pCPT-2-O-Me-cAMP sodium salt (CPTOMe-cAMP), ω-conotoxin (CTX) MVIIC, [D-Ala2, NMe-Phe4, Gly-ol5]-enkephalin (DAMGO), ESI-05, ESI-09, diltiazem, forskolin, gabapentin, guanfacine, H89, HJC0350, ifenprodil, KT5720, L-732,138, MK-801, ML-786, NKH-477, (2R/S)-6-PNG, protein kinase inhibitor 14–22 (PKI 14–22), SNX-482, SQ22536, Torin-2, U0126 and U-73122 were from Tocris (Ellisville, MO). Baclofen and capsazepine were from Research Biochemicals International (RBI, now Sigma-Aldrich). Bovine serum albumin, cyanquixaline (6-cyano-7-nitroquinoxaline-2,3-dione) (CNQX), complete Freund’s adjuvant (CFA), ketamine, lidocaine, Phosphatase Inhibitor Cocktail 2 and common reagents were from Sigma-Aldrich. Matrigel was from BD Biosciences (San Jose, CA). Fura2-AM, nerve growth factor, neurobasal media, NuPAGE Tris-Acetate SDS gels, NuPAGE reagents and Pierce™ Protein-Free T20 (TBS) blocking buffer were from Life Technologies, Grand Island, NY. Halt™ Protease Inhibitor Cocktail and Pierce BCA Protein Assay Kit were from Thermo Fisher Scientific. Fetal bovine serum was from Irvine Scientific, Santa Ana, CA. Drugs were prepared as stock solutions of 1–10 mM in DMSO or water and then diluted in aCSF.

2.3. Spinal cord slices

Spinal cord slices were prepared as described (Adelson et al., 2009; Chen et al., 2018a). The spinal cord was extracted from adult rats after euthanasia with 58–78 mg pentobarbital (0.15–0.2 ml of Fatal Plus, Vortech Pharmaceuticals, Dearborn, Michigan). Coronal slices (400 μm, 3–8 per rat) were cut from the lumbar spinal cord (L2-L4) with a vibratome (Integraslice 7550PSDS, Lafayette Instruments, Lafayette, IN). Slices were kept in artificial cerebrospinal fluid (aCSF), which contained (in mM) 124 NaCl, 1.9 KCl, 26 NaHCO3, 1.2 KH2PO4, 1.3 MgSO4, 2.4 CaCl2 and 10 glucose, and was bubbled with 95% O2 / 5% CO2. Vibratome cutting was done in sucrose-aCSF, which was the same medium with 5 mM KCl and 215 mM sucrose instead of NaCl. Slices were left to recover in oxygenated aCSF at 35°C for 1 h and were used within 3 h. Slices were incubated with drugs by placing them on a nylon net glued to a plastic ring inserted halfway down a plastic tube containing 5 ml aCSF at 35 °C, superficially gassed with 95% O₂/5% CO₂ (Chen et al., 2018a; Marvizon et al., 2003).

Electrical stimulation of the dorsal root was done as reported (Adelson et al., 2009; Chen et al., 2018a). Briefly, a spinal cord slice with an attached dorsal root (>8 mm length) was placed in a superfusion chamber where it was superfused with aCSF at 35 °C. The dorsal root was drawn into a side compartment through a hole in a movable partition and placed on top of a bipolar platinum stimulation electrode. The electrode compartment was emptied of aCSF and filled with mineral oil. Electrical stimulation was generated by a Master-8 stimulator and an Iso-Flex stimulus isolating unit (AMP Instruments, Jerusalem, Israel), and consisted of 1000 square pulses of 1–30 V intensity, 0.4 ms duration, delivered at 100 Hz.

2.4. Immunohistochemistry

The NK1R antiserum (Chen et al., 2018a; Chen et al., 2014; Grady et al., 1996) was raised in rabbit against amino acids 385–407 at the C-terminus of the rat NK1R (AB5060, EDM Millipore, Billerica, MA). A rabbit antiserum against the same peptide (94168 from CURE: Digestive Diseases Research Center, University of California Los Angeles) labeled only cells transfected with rat NK1R, its staining was eliminated by preadsorption with the immunizing peptide, and it yielded a single band in Western blots corresponding to a molecular weight of 100 kDa (Grady et al., 1996). Antiserums AB5060 and 94168 produced the same staining pattern in the rat spinal cord.

Spinal cord slices were processed for NK1R immunohistochemistry as described (Adelson et al., 2009; Marvizon et al., 2003). Experiments were ended by immersing the slices in ice-cold fixative (4% paraformaldehyde, 0.18% picric acid in phosphate buffer). The 400 μm slices were cryoprotected, frozen and sectioned at 25 μm using a cryostat. Sections were washed four times and then incubated overnight at room temperature with the NK1R antiserum diluted 1:3000 in phosphate-buffered saline containing 0.3% Triton X-100, 0.001% thimerosal and 10% normal goat serum (Jackson ImmunoResearch Laboratories, West Grove, PA). After three washes, the secondary antibody (1:2000, Alexa Fluor 488 goat anti-rabbit, Invitrogen-Molecular Probes, Eugene, OR) was applied for 2 hours at room temperature. Sections were washed four more times, mounted on glass slides, and coverslipped with Prolong Gold (Invitrogen-Molecular Probes).

2.5. Quantification of NK1R internalization

NK1R internalization was used to measure substance P release in situ. This method has been extensively used and validated (Adelson et al., 2009; Allen et al., 1997; Honore et al., 1999; Mantyh et al., 1995). It was shown to be quantitative and more sensitive than immunoassay (Marvizon et al., 2003). Its main limitation is that it may give a false positive it there is agonistin-dependent NK1R internalization. NK1R neurons in lamina I were visually counted using a Zeiss Axio-Imager A1 (Carl Zeiss, Inc., Thornwood, NY) fluorescence microscope with a 63x (1.40 numerical aperture) oil objective. The criterion for a neuron to have NK1R internalization was the presence in its neuronal soma of ten or more NK1R endosomes. The person doing the counting was blinded to the treatment of the slices. Four sections per slice were used and in each of these sections all lamina I NK1R neurons were counted and classified as with or without internalization. This amounts to counting 50–250 neurons per slice (Marvizon et al., 1999). A detailed explanation of the counting procedure can be found in that reference.

2.6. Confocal microscopy and image processing

A Zeiss LSM 710 confocal microscope (Carl Zeiss, Inc., Thornwood, NY) was used. Objectives were 20x (numerical aperture 0.8) for images of the dorsal horn and 63x oil (numerical aperture 1.4) for individual neurons. Alexa Fluor 488 (emission peak 519 nm) was visualized with the 488 nm line of an argon laser. The emission window of the dichroic mirror was set to 502–570 nm. Pinhole was 31.5 μm for the 20x objective and 50.7 μm for the 63x objective (1.0 Airy unit). Confocal sections were 1024×1024 pixels, separated 0.85 μm for the 20x objective and 0.38 μm for the 63x objective, as determined using the Nyquist equation.

Deconvolution (Cannell et al., 2006; Holmes et al., 2006; Wallace et al., 2001) was used to reduce the blur of some images, consisting of 5 iterations of adaptive point spread function (‘blind’) deconvolution with AutoQuant X 2.0.1 (Media Cybernetics, Inc., Bethesda, MD). Confocal stacks were cropped in three dimensions using Imaris 6.1.5 (x64, Bitplane AG, Zurich, Switzerland). Two-dimension images was generated in Imaris and imported into Adobe Photoshop 5.5 (Adobe Systems Inc., Mountain View, CA), which was used to adjust the contrast and to compose the multi-panel figures.

2.7. Dorsal root ganglion (DRG) neuron culture

Primary cultures of DRG neurons from spinal levels T12-S2 were prepared as previously described (Li et al., 2006). Neurons were plated on coverslips coated with Matrigel in Neurobasal media containing B-27 supplement, 5 ng/ml nerve growth factor, 10% fetal bovine serum and 200 μM ketamine. Ketamine is necessary to preserve NMDA receptor function in cultured DRG neurons (Chen et al., 2014).

2.8. Intracellular Ca²⁺ ([Ca²⁺]_i) measurement

DRG neurons were used 2–3 days after plating. Neurons were loaded for 1 hour at 37 °C on coverslips with 5 μM Fura-2 AM in Neurobasal media supplemented as described above, then washed and incubated for 30 min in Mg²⁺-free Hank’s buffer containing 10 mM Hepes, pH 7.2, 5 mM glucose, 1.2 mM CaCl₂. During the last 15 min, the incubation buffer was supplemented 20 ng/ml BDNF with or without 10 μM forskolin. Coverslips were mounted in an experimental chamber (volume 0.5 ml) and perfused (0.5 ml/min) with the same buffer. The chamber was placed on the stage of an inverted microscope (Axio Observer.A1) with a 40x objective and a digital camera (AxioCam MRm) and operated with associated software (AxioVision, Carl Zeiss, Thornwood, NY). Ratio images (340 nm / 380 nm) were obtained at 1 s intervals, and the ratio values from each cell were stored for offline analysis. Neurons were stimulated by the rapid addition of 250 μM NMDA and 10 μM glycine. KCl (50 mM) was added at the end of the experiment as a means to verify that neurons were able to respond with increases in [Ca²⁺]_i. Responses were quantified as the change in the 340 nm / 380 nm ratio from the initial baseline value to the peak value.

2.9. Western blots

DRG neurons were cultured for 2 days in 6-well plates. Medium was carefully aspirated and 1 ml fresh pre-equilibrated serum-free Neurobasal medium containing either 10 μM forskolin, forskolin plus 10 μM H89, or no addition (control) was slowly added to the wells. Plates were returned to the incubator for 10 min, then removed and placed on ice. Medium was aspirated and 0.75 ml ice-cold PBS added to each well before scraping the neurons off the plates. Samples were centrifuged at 500 g for 5 min at 4°C, after which the supernatant was aspirated and the pellets suspended in ice-cold homogenization buffer containing (in mM): 300 sucrose, 25 Tris-HCl (pH 7.5), 10 Na⁺-glycerophosphate and 1 EDTA, supplemented with Halt™ Protease Inhibitor Cocktail and Phosphatase Inhibitor Cocktail 2. Cells were homogenized by repeated pipetting and vortexing for 1 minute, then set on ice for 10 min before centrifugation at 10,000 g for 10 min at 4°C. Protein concentrations in each sample were measured using the Pierce BCA Protein Assay Kit and each sample combined with sample buffer (NuPAGE) and 10 mM DTT. Samples containing 30 μg protein were electrophoresed on 3–8% NuPAGE Tris-Acetate SDS gels and proteins transferred to PVDF membranes. Blots were blocked with 3% bovine serum albumin in Pierce™ Protein-Free T20 blocking buffer and probed with antibodies (in T20 blocking buffer) to NR1 (1:4000, Abcam, ab109182), p-Ser⁸⁹⁷-NR1 (1:1000, Abcam, ab52184), NR2B (1:2000, Santa Cruz, sc9057) or p-Ser¹¹⁶⁶- NR2B (1:1000, PhosphoSolutions p1516–1166). Blots were developed using peroxidase-conjugated anti-rabbit IgG (1:10,000, 71–035-152, Jackson ImmunoResearch Laboratories, West Grove, PA) and Western Bright Quantum HRP substrate (Advansta, Menlo Park, CA). Band intensities of pNR1 and pNR2B were expressed relative to that of NR1 and NR2B, respectively, and normalized to the controls.

2.10. CFA injections

CFA was injected undiluted in a volume of 50 μl. One hindpaw of the rat was injected subcutaneously with 50 μl undiluted CFA. A 50 μl Hamilton syringe was used, fitted with a 26 G needle that was inserted in the middle of the paw, near the base of the third toe. The needle was held in place for 15 s and then gently withdrawn.

2.11. Mechanical allodynia

Mechanical allodynia was measured using the two-out-of-three method (Chen et al., 2018b; Jarahi et al., 2014; Kingery et al., 2000; Michot et al., 2012; Walwyn et al., 2016). Rats were habituated for 3 days, 30 min daily, to acrylic enclosures on an elevated metal grid (IITC Life Science Inc., CA). A series of von Frey filaments (‘Touch-Test’, North Coast Medical, Inc., San Jose, CA) were applied in ascending order to the plantar surface of the hindpaw for a maximum of 3 s. A withdrawal response was counted only if the hindpaw was completely removed from the grid. Each filament was applied three times, and the minimal value that caused at least two responses was recorded as the paw withdrawal threshold (PWT). The cut-off threshold was the 15 g filament.

2.12. Experimental design

NK1R internalization data were obtained in weekly experiments of 12–24 slices prepared from 2–8 rats (3–8 slices per rat). The number of slices for each data point is indicated in the figures. Replicates in each data point are from slices from more than one rat, and hence the statistical error represents both inter-slice and inter-animal variability. Separate controls were included in each weekly experiment to assess consistency. Control data were later aggregated for the whole experiment, resulting in a larger n for controls than for the results for individual drugs.

2.13. Data analysis

Data were analyzed using Prism 8.03 (GraphPad Software, San Diego, CA) and expressed as mean ± standard error. Statistical significance was set at 0.05. Statistical analyses for NK1R internalization and Western blots consisted of one-way ANOVA followed by Holm-Sidak’s post-hoc tests. Measures of [Ca²⁺]_i in DRG cultures were analyzed with the Mann-Whitney test due to the non-Gaussian distribution of the data.

Concentration-response data were fitted using non-linear regression by the sigmoidal dose-response function Y = bottom + (top-bottom) / (1 + 10^(Log EC₅₀-Log X)), where X is the concentration of drug and EC₅₀ is the concentration of drug that produces half of the effect. The Hill coefficient was assumed to be 1. Control measures (zero concentration of drug) were included in the non-linear regression by assigning them a concentration value three log units lower than the lowest concentration of drug. Parameter constraints were: 0% < top < 100%, 0% < bottom. The statistical error of the IC₅₀ is expressed as 95% confidence intervals (CI).

Time-course data were fitted by non-linear regression to a one-phase decay function: Y = (Y₀ - plateau) * exp(−K * X) + plateau, where K is the rate constant, Y₀ is the value at time 0 and “plateau” is the maximum value. Half-life was calculated as ln(2)/K.

3. Results

3.1. AC activators induced substance P release

Incubating rat spinal cord slices at 35 °C for 60 min with the AC activator forskolin (10 μM) induced NK1R internalization in lamina I neurons, whereas the inactive forskolin analog 1,9-dideoxyforskolin (10 μM) produced no effect (Fig. 1A). Another AC activator, NKH-477 also induced NK1R internalization (Fig. 1A). NKH-477 produced similar amounts of NK1R internalization at 0.1 μM (49 ± 6 %, n=3), 1 μM (46 ± 4 %, n=3) and 10 μM (42 ± 7 %, n=3).

Figure 1. Adenylyl cyclase (AC) activation induces NK1R internalization.

A: Spinal cord slices were incubated for 60 min with aCSF alone (“control”), 10 μM 1,9-dideoxyforskolin (inactive analog), 1 μM NKH-477 (AC activator), 10 μM forskolin, forskolin plus 1 μM L-732,138 (NK1R antagonist), forskolin plus 100 μM SQ22536 (AC inhibitor) or forskolin plus 1 mM lidocaine (Nav channel blocker). ANOVA: p<0.0001, F (6, 54)=71. B: Spinal cord slices were incubated for 60 min with aCSF alone (“control”), or 10 nM 6-Bnz-cAMP, alone or with 1 μM L-732,138 or 1 mM lidocaine. Numbers inside the bars indicate the number of slices in each group (n). ANOVA: p<0.0001, F (3, 45)=49. Holm-Sidak’s post-hoc tests: ** p<0.001, *** p<0.0001 compared to control; # p<0.05, ### p<0.0001 compared to forskolin (A) or 6-Bnz-cAMP (B).

The NK1R internalization induced by forskolin was abolished by L-732,138 (1 μM), a competitive NK1R antagonist that has a 1000-fold selectivity for NK1Rs over NK2 and NK3 receptors (Cascieri et al., 1994; MacLeod et al., 1994). In a previous experiment, L-732,138 produced the same amount of inhibition of NK1R internalization induced by 1 μM substance P in spinal cord slices as the NK1R antagonists L-733,060, L-703,606, WIN-62577 and RP-67580 (all 10 μM). L-732,138 was added to the slices 40 min before inducing substance P release, which was necessary for 1 μM L-703,606 to inhibit NK1R internalization induced by substance P, NMDA and electrical stimulation of the dorsal root (Marvizon et al., 1997), and for 7 μM RP-67,580 to inhibit NK1R internalization induced by capsaicin (Lao et al., 2003). This effect of L-732,138 shows that forskolin-induced NK1R internalization was caused by substance P release and its binding to the NK1R.

Forskolin-induced NK1R internalization was also decreased by the AC inhibitor SQ22536 (100 μM), supporting the idea that is was due to the activation of AC.

Blocking Nav channels with 1 mM lidocaine to eliminate the firing of action potentials slightly decreased the NK1R internalization induced by forskolin (Fig. 1A), indicating that substance P release is induced primarily by AC present in the presynaptic terminals, but with a small contribution of AC outside the synapse.

Fig. 2 shows confocal microscope images of neurons with and without NK1R internalization in slices incubated with 10 μM forskolin (Fig. 2A), 10 nM 6-Bnz-cAMP (Fig. 2B) or 6-Bnz-cAMP plus 100 nM Torin-2 (Fig. 2C), a compound found to largely inhibit the NK1R internalization produced by 6-Bnz-cAMP (see below). Most of the neurons in the slices treated with forskolin or 6-Bnz-cAMP show numerous endosomes containing NK1Rs, the hallmark of receptor internalization. In contrast, in slices in which Torin-2 was added to 6-Bnz-cAMP there are many neurons in which NK1Rs appears only at the cell surface.

Figure 2. Confocal images of NK1R neurons in lamina I.

Spinal cord slices were incubated with 10 μM forskolin (A), 10 nM 6-Bnz-cAMP (B), or 10 nM 6-Bnz-cAMP + 100 nM Torin-2 (C). Images in the main panels are stacks of 6 (A), 4 (B) or 8 (C) optical sections taken with the 20x objective. Images in the insets are stacks of 4–6 optical sections taken with the 63x objective. Insets are all at the same scale (scale bar is 10 μm). Deconvolution was used to reduce blur in the images in panel A. Arrows show the location in the main panels of the neurons shown in the insets. Neurons with NK1R internalization are labeled with “*” and neurons without internalization by “o”.

A concentration-response for forskolin (Fig. 3A) yielded an EC₅₀ of 6.6 μM (95% CI 3.6 – 11.5 μM, R² = 0.898), which is consistent with the reported K_D of forskolin for AC of 9.8 μM (Awad et al., 1983). The efficacy of forskolin was high, with a maximum effect of 80 ± 6.7 % NK1R neurons with internalization.

Figure 3. Concentration-responses of forskolin, 6-Bnz-cAMP, 8-Br-cAMP and CPTOMe-cAMP to induce NK1R internalization.

Spinal cord slices were incubated for 60 min with the indicated concentrations of forskolin (A) or the cAMP analogs (B). A dose-response function was fitted to the data by non-linear regression. Forskolin: EC₅₀=6.6 μM (95% CI 3.6 – 11.5 μM), top=80.0 ± 6.7%, bottom=20.2 ± 1.7%, R²=0.898. 6-Bnz-cAMP: EC₅₀=0.50 pM (95% CI 0.10 – 2.06 pM), top=49.2 ± 1.3%, bottom=21.6 ± 2.4%, R²=0.642. CPTOMe-cAMP: EC₅₀=5.2 μM (95% CI 2.6 – 10.2 μM), top=58.6 ± 3.0%, bottom=20.8 ± 2.1%, R²=0.830. Data for 8-Br-cAMP were fitted significantly better by a biphasic than a single-component dose-response function (Akaike’s Information Criterion, probability >99.99%). Parameters were: EC₅₀(1)=706 pM (95% CI 129 – 2698 pM), EC₅₀(2)=392 μM (95% CI 62 – 652 μM), Fraction (1)=28 ± 13%, top=97.9 ± 36%, bottom=21.7 ± 3.0%, R²=0.788.

These data show that activating AC to produce cAMP induced substance P release in the superficial dorsal horn.

3.2. cAMP analogs induced substance P release

We hypothesized that cAMP induces substance P release by activating protein kinase A (PKA) and exchange protein directly activated by cAMP (Epac) (Gloerich and Bos, 2010; Kawasaki et al., 1998). If so, cell-permeant cAMP analogs should also induce substance P release and NK1R internalization. 6-Bnz-cAMP (10 nM), which selectively activates PKA (Christensen et al., 2003; Rehmann et al., 2003), induced abundant NK1R internalization in spinal cord slices that was abolished by the NK1R antagonist L-732,138 (Fig. 1B), indicating that it was due to substance P release. The effect of 6-Bzn-cAMP was not decreased by 1 mM lidocaine, showing that it did not require the firing of action potentials. This indicates that the release was elicited by PKA in the presynaptic terminals. In fact, lidocaine produced a small but statistically significant increase of the effect of 6-Bnz-cAMP. This could be due to disinhibition or another alteration in the firing of the neurons in the slices.

A concentration-response for 6-Bnz-cAMP (Fig. 3B) showed that it induced NK1R internalization with remarkably high potency (Fig. 3B): its EC₅₀ was 0.50 pM (95% CI 0.1 – 2.0 pM, R² = 0.642). However, its efficacy was lower than that of forskolin. Concentrations of 6-Bnz-cAMP higher than 10 nM resulted in lower amounts of NK1R internalization (43 ± 10 % at 100 nM, 33 ± 6 % at 1 μM, n=3).

Concentration-response curves were also obtained for CPTOMe-cAMP, which selectively activates Epac1 and 2 (Enserink et al., 2002; Singhmar et al., 2016), and 8-Br-cAMP, which activates both PKA and Epac. CPTOMe-cAMP-induced NK1R internalization with an efficacy similar to 6-Bnz-cAMP (Fig. 3B) but with a potency several orders of magnitude lower (EC₅₀ of 5.2 μM, 95% CI 2.6 – 10.2 μM, R² = 0.830). The EC₅₀ of CPTOMe-cAMP was similar to its reported EC₅₀ for Epac, 2.2 μM (Enserink et al., 2002).

8-Br-cAMP induced NK1R internalization with a clearly biphasic concentration-response curve (Fig. 3B), as indicated by a probability of more than 99.99% that the data were fitted by a biphasic instead of a single-phase dose-response function (Akaike’s Information Criterion). The EC₅₀s of these two phases were 706 pM (95% CI 129 – 2698 pM) and 392 μM (95% CI 62 – 652 μM). The overall efficacy of 8-Br-cAMP was very high: 98% NK1R neurons with internalization, with 28% of that value corresponding to the high potency phase and 72% to the low potency phase.

These potent effect of 6-Bnz-cAMP indicates that cAMP induces substance P release by activating PKA. However, the biphasic effect of 8-Br-cAMP and the effect of CPTOMe-cAMP indicate that other signals, possibly Epac also contribute to the induction of substance P release by cAMP. Given the low potency of the low affinity component of 8-Br-cAMP and of CPTOMe-cAMP these possibility was studied in further experiments.

3.3. Time-courses for forskolin and 6-Bnz-cAMP

We studied the time course of the effect of forskolin by incubating the slices with 10 μM forskolin for periods of time between 5 min and 2 h (Fig. 4). A one-phase association function fitted well the data (R² = 0.942), yielding an association constant K of 0.0312 ± 0.0057 min⁻¹ corresponding to a half-time (t_1/2) of 22.2 min (95% CI = 15.6 min - 31.8 min). The effect of forskolin was nearly complete at 60 min, the incubation time used in other experiments.

Figure 4. Time course for forskolin and 6-Bnz-cAMP to induce NK1R internalization.

Slices were incubated for the times indicated with 10 μM forskolin or 10 nM 6-Bnz-cAMP. Curves represent non-linear regression fitting of a one-phase association function to the data. Forskolin: Y0=14.9 ± 1.3%, plateau=64.0 ± 3.0%, K=0.0312 ± 0.0057 min⁻¹, t_1/2=22.2 min (95% CI=15.6 min − 31.8 min), R²=0.942. 6-Bnz-cAMP: Y0=16.2 ± 3.7%, plateau=47.4 ± 2.6%, K=0.89 ± 4.87 min⁻¹, t_1/2=0.77 min (95% CI=? - 8.78 min), R²=0.521.

We also studied the time course of the effect of 6-Bnz-cAMP by incubating the slices with 10 nM 6-Bnz-cAMP for periods of time between 5 and 90 min (Fig. 4). NK1R internalization induced by 6-Bnz-cAMP was already maximal at 5 min. Shorter incubation times were not attempted because NK1R internalization itself requires 2–3 min (Wang and Marvizon, 2002). Indeed, the kinetics of NK1R internalization have a t_1/2 of 71 s, similar to the tentative t_1/2 of 0.77 min (46 s) obtained for 6-Bnz-cAMP. Therefore, substance P release induced by activating PKA with 6-Bnz-cAMP takes place much faster than substance P release induced by activating AC with forskolin.

3.4. AC and PKA contribute to substance P release induced by dorsal root stimulation

In many of our previous studies, substance P release was induced by electrically stimulating the dorsal root entering a spinal cord slice (Adelson et al., 2009; Chen et al., 2014; Kondo et al., 2005; Marvizon et al., 1999; Marvizon et al., 1997; Zhang et al., 2010). The stimulation patterns we used mimic the trains of action potentials in primary afferents induced by noxious stimuli (Adelson et al., 2009; Lever et al., 2001). It is possible that high-frequency action potentials reaching the central terminals of primary afferents activate AC and PKA and that this contributes to the release of substance P. If this is true, then inhibiting AC or PKA should decrease substance P release induced by dorsal root stimulation. Accordingly, we stimulated the dorsal root entering spinal cord slices with 1000 pulses of 20 V, 0.4 ms delivered at 100 Hz, which recruit all types of primary afferents including C-fibers (Adelson et al., 2009). Incubating the slices with the AC inhibitor SQ22536 (100 μM) or the PKA inhibitor KT5720 (10 μM) produced a significant, albeit partial, decrease in the NK1R internalization evoked by the dorsal root stimulation (Fig. 5). KT5720 potently inhibits PKA, but it also inhibits other kinases (Davies et al., 2000). These results indicate that cAMP and PKA participate in the substance P release triggered by incoming action potentials.

Figure 5. Substance P release induced by dorsal root stimulation was decreased by inhibitors of AC and PKA.

Spinal cord slices with one dorsal root were incubated for 60 min in aCSF alone (“control”), 100 μM SQ22536 (AC inhibitor) or 10 μM KT5720 (PKA inhibitor). Then the slices were transferred to a slice chamber, where they were superfused with these compounds for 5 min and the dorsal root electrically stimulated with 1000 pulses of 20 V and 0.4 ms delivered at 100 Hz. NK1R internalization was measured in the dorsal horns ipsilateral and contralateral to the stimulated root. Two-way ANOVA (repeated measures): drugs, p=0.0040, F (2, 23)=7.08; side/stimulation, p<0.0001, F (1, 23)=109; drugs x side, p=0.125, F (2, 23)=2.28. Holm-Sidak’s post-hoc tests: ** p<0.01 compared with “control”.

3.5. NMDA receptors mediate substance P release induced by forskolin and 6-Bnz-cAMP

NMDA receptors induce substance P release (Chen et al., 2018a; Chen et al., 2010; Liu et al., 1997; Malcangio et al., 1998; Marvizon et al., 1999; Marvizon et al., 1997), so we hypothesized that cAMP induces substance P release by activating these NMDA receptors. First, we examined whether substance P release induced by forskolin or 6-Bnz-cAMP is additive with that induced by NMDA (in the presence of its co-agonist D-Ser). NMDA plus D-Ser (both 10 μM) did not increase substance P release induced by forskolin (Fig. 6A) or 6-Bnz-cAMP (Fig. 6B). Second, we studied whether NMDA receptor antagonists inhibit substance P release induced by forskolin and 6-Bnz-cAMP. The NMDA receptor channel blocker MK-801 (1 μM) decreased the substance P release induced by forskolin (Fig. 6A) or 6-Bnz-cAMP (Fig. 6B). However, ifenprodil, a selective antagonist of NR2B subunit-containing NMDA receptors, did not inhibit substance P release evoked by either forskolin (Fig. 6A) or 6-Bnz-cAMP (Fig. 6B). This is surprising because we have previously shown that the NMDA receptor-induced substance P release is potently (IC₅₀ = 1.6 nM), albeit partially (60–70%), inhibited by ifenprodil (Chen et al., 2018a). A concentration-response for ifenprodil on forskolin-induced substance P release (Fig. 6C) yielded an IC₅₀ of 7 μM, indicating that the NR2B subunit is not involved. The AMPA receptor antagonist CNQX (5 μM) did not affect substance P release induced by 6-Bnz-cAMP (Fig. 6B). Therefore, NMDA receptors, but not AMPA receptors, contribute to cAMP-induced substance P release but are not the sole mediators of this effect.

Figure 6. Involvement of NMDA receptors in substance P release induced by forskolin and 6-Bnz-cAMP.

Spinal cord slices were incubated for 60 min with 10 μM forskolin (A) or 10 nM 6-Bnz-cAMP (B), alone (“none”) or with the NMDA receptor agonists NMDA and D-Ser (both 10 μM), the NMDA receptor blocker MK-801 (10 μM), the NR2B-NMDA receptor antagonist ifenprodil (10 nM), or the AMPA receptor antagonist CNQX (5 μM). Numbers inside the bars indicate the number of slices in each group (n). ANOVA (A): p<0.0001, F (3, 46)=16; (B): p<0.0001, F (4, 50)=8.1. Holm-Sidak’s post-hoc tests: *** p<0.001, compared with “none”. C: Spinal cord slices were incubated for 60 min with 10 μM forskolin and the indicated concentrations of ifenprodil. IC₅₀=7.1 μM (95% CI 0.7–16 μM), top=52% (95% CI 51–57%), bottom=0% (95% CI 0–29%), R²=0.407.

3.6. Forskolin increases Ca²⁺ entry in DRG neurons induced by NMDA receptor activation

To test the hypothesis that forskolin-induced substance P release was mediated by an increase in Ca²⁺ entry in primary afferents through NMDA receptors, we studied the effect of forskolin on NMDA-induced increases in [Ca²⁺]_i in cultured DRG neurons. As we previously established (Chen et al., 2018a; Chen et al., 2014), dissociated rat DRG neurons were cultured for 2–3 days in the presence of 200 μM ketamine to preserve NMDA receptor function. The cells were loaded for 1 h with the Ca²⁺ indicator Fura-2 AM (5 μM) and then incubated for 15 min with 20 ng/ml BDNF to increase NMDA receptor function (Chen et al., 2018a; Chen et al., 2014). Increases in [Ca²⁺]_i were evoked by rapid infusion of 250 μM NMDA + 10 μM glycine. Cells were sampled randomly and classified as responders or non-responders based on a cutoff of Log of peak responses of 0.015. The addition of 10 μM forskolin during the 15 min incubation with BDNF significantly increased [Ca²⁺]_i peak responses in responder cells (Fig. 7A), but did not change the number of responder and non-responder cells (Fig. 7B).

Figure 7. Forskolin increases NMDA-induced [Ca²⁺]_i responses in DRG neurons.

A. Cultured DRG neurons were loaded for 1 h with 5 μM Fura-2 AM and then incubated for 15 min before addition of 10 μg/ml BDNF with or without 10 μM forskolin for another 15 min. Responses were evoked by rapid infusion of 250 μM NMDA + 10 μM glycine. Results are graphed on a log scale to better visualize the data and analyzed using the nonparametric Mann-Whitney test (p=0.0009). Lines and error bars are mean ± standard error. B. The number of DRG neurons that responded to NMDA with peak responses larger than 0.015 (cutoff) is indicated by the red bars. Blue bars indicate the number of recorded neurons with no response or responses below the cutoff (p=1.000, Fisher’s exact test).

Therefore, the functionality of NMDA receptors present in DRG neurons is increased by forskolin, which is consistent with the idea that cAMP leads to their activating phosphorylation. This was examined in the next experiment.

3.7. Forskolin increased PKA phosphorylation of NMDA receptors in DRG neurons

Western blots using antibodies against the phosphorylated and non-phosphorylated NR1 and NR2B subunits of the NMDA receptor were performed to determine if forskolin increases their phosphorylation in primary afferents. Cultured DRG neurons were incubated for 10 min in medium alone (control), with 10 μM forskolin, or with forskolin and 10 μM H89, a PKA inhibitor. Gels were probed with antibodies to NR1, p-Ser⁸⁹⁷-NR1 (pNR1), NR2B or p-Ser¹¹⁶⁶-NR2B (pNR2B) (Fig. 8A).

Figure 8. Forskolin increases phosphorylation of the NR1 and NR2B subunits of the NMDA receptor.

Cultured DRG neurons were incubated for 10 min at 37°C with no addition (control, C), 10 μM forskolin (F), or forskolin + 10 μM H89 (F+H). Cells were homogenized and extracts electrophoresed on SDS-PAGE gels. A: Membrane blots of the gels were probed with antibodies to NR1, p-Ser⁸⁹⁷-NR1 (pNR1), NR2B, or p-Ser¹¹⁶⁶- NR2B (pNR2B). B: Band intensities of pNR1 and pNR2B were divided by the intensities of NR1 and NR2B, respectively, and normalized to the controls. Kurskal-Wallis tests: pNR1 p<0.0001, pNR2B p<0.0001. Dunn’s post-hoc tests (uncorrected): *** p<0.0001 compared to control; † p=0.028 compared to forskolin.

Forskolin did not affect the amount of non-phosphorylated NR1 (p=0.607, F(2, 10)=0.52; p>0.73 in all Holm-Sidak’s multiple comparison tests, mixed-effect analysis) or non-phosphorylated NR2B (p=0.145, F(2, 9)=2.4; p>0.16 in all Holm-Sidak’s multiple comparison tests). However, forskolin significantly increased p-Ser⁸⁹⁷-NR1 and p-Ser¹¹⁶⁶-NR2B relative to NR1 and NR2B, respectively (Fig. 8B, Kurskal-Wallis tests: pNR1 p<0.0001, pNR2B p<0.0001). These increases were eliminated by H89.

H89 is considered a selective inhibitor of PKA (Davies et al., 2000), although it also inhibits some other kinases (Limbutara et al., 2019). Taken together, these data indicate that when it is activated by cAMP, PKA phosphorylates the NR1 and NR2B subunits of the NMDA receptor.

3.8. Voltage-gated Ca²⁺ (Cav) channels mediate substance P release induced by forskolin and 6-Bnz-cAMP

Since the NMDA receptor blocker MK-801 did not completely inhibit substance P release induced by forskolin (Fig. 6A) or 6-Bnz-cAMP (Fig. 6B), we studied whether Ca²⁺ entry through Cav channels also contributes to this substance P release.

Forskolin-induced substance P release (Fig. 9A) was decreased by inhibitors of several types of Cav channels: Cav1 (L) channels (1 μM diltiazem), Cav2 (N/P/Q) channels (1 μM CTX MVIIC) (Hannon and Atchison, 2013), Cav2.1 (P/Q) channels (0.3 μM ω-agatoxin TK) (Rusin and Moises, 1995) and Cav3.2 (T) channels (10 μM 2R/S-6-PNG) (Sekiguchi et al., 2018), but not by the inhibitor of Cav2.3 (R) channels SNX482 (0.3 μM) (Qian et al., 2013). Gabapentin (30 μM), an inhibitor of the δ2 subunit that modulates Cav2 channels (Hoppa et al., 2012), produced a slight inhibition. Since the Gαi-coupled receptors μ-opioid receptors, α_2A adrenergic receptors and GABA_b receptors inhibit Cav channels (Heinke et al., 2011; Raingo et al., 2007; Rusin and Moises, 1995; Strock and Diverse-Pierluissi, 2004; Zhang et al., 2010), we studied the effects of their agonists (respectively) DAMGO (1 μM), guanfacine (10 nM) and baclofen (30 μM). The three agonists strongly inhibited forskolin-induced substance P release (Fig. 9A).

Figure 9. Involvement of Cav channels in substance P release induced by forskolin and 6-Bnz-cAMP.

Spinal cord slices were incubated for 60 min with 10 μM forskolin (A) or 10 nM 6-Bnz-cAMP (B), alone (“none”) or with the inhibitors of Cav2 channels diltiazem (1 μM, Cav1 or L type), ω-conotoxin (CTX) MVIIC (1 μM, Cav2 or N, P, Q types), ω-agatoxin TK (0.3 μM, Cav2.1 or P, Q types), SNX482 (0.3 μM, Cav2.3 or R type), (2R/S)-6-PNG (10 μM, Cav3.2 or T type) and gabapentin (30 μM, δ2 subunit), or with the agonists of Gi/o-coupled receptors DAMGO (1 μM, MOR), guanfacine (10 nM, α_2A adrenergic receptor) or baclofen (30 μM, GABA_b receptor). Numbers inside the bars indicate the number of slices in each group (n). ANOVA (A): p<0.0001, F (9, 58)=21.1. ANOVA (B): p<0.0001, F (7, 52)=6.0. Holm-Sidak’s post-hoc tests: ** p<0.01, *** p<0.001, compared with “none”.

The effect of Cav inhibitors on 6-Bnz-cAMP-induced substance P release was markedly different (Fig. 9B). It was significantly decreased by the Cav2 channel blocker CTX MVIIC (1 μM), but not by blockers of Cav1 (1 μM diltiazem), Cav2.1 (0.3 μM ω-agatoxin TK), Cav3.2 (10 μM 2R/S-6-PNG), Cav2.3 (0.3 μM SNX482) or by gabapentin (30 μM). Substance P release induced by 6-Bnz-cAMP was decreased by the μ-opioid receptor agonist DAMGO (1 μM).

It is possible that the differences of the effects of Cav channel blockers between forskolin and 6-Bnz-cAMP are due to the fact that 6-Bnz-cAMP, but not forskolin, was applied at a non-saturating concentration (Fig. 3). However, it is far more likely that these differences are due to the fact that 6-Bnz-cAMP specifically activates PKA, whereas the cAMP produced by forskolin activation of AC has other targets. This possibility was explored in the next experiments.

3.9. Negligible contribution of TRPV1 and TRPA1 to substance P release induced by forskolin and 6-Bnz-cAMP

TRPV1 and TRPA1 are other routes of Ca²⁺ entry into the presynaptic terminals of primary afferents. The TRPV1 inhibitor AMG9810 at 300 nM (IC₅₀=85 nM (Gavva et al., 2005)) did not affect substance P release induced by forskolin (Fig. 10A) or 6-Bnz-cAMP (Fig. 10B). Another TRPV1 inhibitor, capsazepine at 100 μM (IC₅₀=420 nM (Bevan et al., 1992)), did not inhibit the effect of 6-Bnz-cAMP (Fig. 10B). The TRPA1 inhibitor AM-0902 at 300 nM (IC₅₀=71 nM (Schenkel et al., 2016)) did not affect substance P release induced by forskolin (Fig. 10A) but slightly inhibited of the effect of 6-Bnz-cAMP (Fig. 10B).

Figure 10. Contribution of TRPV1 and TRPA1 to substance P release induced by forskolin and 6-Bnz-cAMP.

Spinal cord slices were incubated for 60 min with 10 μM forskolin (A) or 10 nM 6-Bnz-cAMP (B), alone (“none”) or with the TRPV1 antagonists capsazepine (100 μM) or AMG9810 (0.3 μM), or the TRPA1 antagonist AM-0902 (0.3 μM). Numbers inside the bars indicate the number of slices in each group (n). ANOVA (A): p=0.084, F (2, 40)=2.6; (B): p=0.040, F (3, 41)=3.0. Holm-Sidak’s post-hoc tests: * p=0.014 compared with “none”.

These results indicate that TRPV1 and TRPA1 are not the main mediators of cAMP-induced substance P release.

3.10. Mechanisms that mediate substance P release induced by 8-Br-cAMP

NK1R internalization induced by 0.5 mM 8-Br-cAMP, a concentration that activates the high and low potency components of its effect (Fig. 11), was inhibited by the NK1R antagonist L-732,138 (1 μM), indicating that it was due to substance P release. It was decreased by lidocaine (1 mM), showing that it was partly mediated by the firing of action potentials. The PKA inhibitors KT5720 (10 μM) and PKI 14–22 (10 μM) produced full and partial inhibition, respectively, indicating that it is due to the activation of PKA. The NMDA receptor blocker MK-801 (10 μM) and the Cav2 channel blocker CTX MVIIC (1 μM) produced partial inhibitions, whereas the MOR agonist DAMGO (1 μM) produced strong inhibition. The inhibition by DAMGO was stronger than the inhibition produced by CTX MVIIC (p=0.0036 in the Holm-Sidak’s post-hoc test). This could be because 8-Br-cAMP activates several types of Cav channels. Whereas CTX MVIIC primarily inhibits N-type channels, the μ-opioid receptors activated by DAMGO inhibit N-, P- and Q-type Cav channels (Rusin and Moises, 1995).

Figure 11. Substance P release induced by 8-Br-cAMP.

Spinal cord slices were incubated for 60 min with 0.5 mM 8-Br-cAMP, alone (“none”) or with the NK1R antagonist L-732,138 (1 μM), the Nav channel blocker lidocaine (1 mM), the PKA inhibitors KT5720 (10 μM) or PKI 14–22 (10 μM), the NMDA receptor blocker MK-801 (10 μM), the Cav2 channel blocker CTX MVIIC (1 μM) or the MOR agonist DAMGO (1 μM). ANOVA: p<0.0001, F (7, 34)=21. Holm-Sidak’s post-hoc tests: ** p<0.01, *** p<0.001 compared with “none”. Numbers inside the bars indicate the number of slices in each group (n).

3.11. Effects of inhibitors of PKA, Epac1 and Epac2

NK1R internalization induced by forskolin was abolished by the PKA inhibitor KT5720 (10 μM) and reduced by the PKA inhibitor PKI 14–22 (10 μM, Fig. 12A). Likewise, substance P release induced by 6-Bnz-cAMP was abolished by KT5720 and reduced by PKI 14–22 (Fig. 12B). Unexpectedly, the induction of substance P release by CPTOMe-cAMP, a selective activator of Epac, was also decreased by the PKA inhibitors KT5720 and PKI 14–22 (Fig. 12C).

Figure 12. Involvement of PKA, Epac1 and Epac2 in substance P release induced by forskolin, 6-Bnz-cAMP and CPTOMe-cAMP.

Spinal cord slices were incubated for 60 min with 10 μM forskolin (A), 10 nM 6-Bnz-cAMP (B) or 30 μM CPTOMe-cAMP (C), alone (“none”) or with the PKA inhibitors KT5720 (10 μM) or PKI 14–33 (10 μM); the Epac inhibitor ESI-09 (25 μM), the Epac1 inhibitor CE3F4 (30 μM), or the Epac2 inhibitors HJC0350 (3 μM) or ESI-05 (10 μM). Each panel was analyzed by ANOVA: A: p<0.0001, F (6, 58)=22.3; B: p<0.0001, F (6, 49)=6.8; C: p<0.0001, F (6, 29)=11. Holm-Sidak’s post-hoc tests: * p<0.05, ** p<0.01, *** p<0.001 compared with “none”. Numbers inside the bars indicate the number of slices in each group (n).

Another target of cAMP is Epac, of which there are two isoforms, Epac1 and Epac2 (Lezoualc’h et al., 2016). To determine whether they contribute to forskolin-induced substance P release, we incubated spinal cord slices with forskolin and inhibitors of Epac1 and Epac2. ESI-09 inhibits both Epac1 and Epac2 with IC₅₀s of 3 μM and 1 μM, respectively (Almahariq et al., 2013), whereas CE3F4 selectively inhibits Epac1 with an IC₅₀ of 10 μM (Courilleau et al., 2012). Forskolin-induced substance P release was increased by 25 μM ESI-09 and by 30 μM CE3F4 (Fig. 12A). Conversely, it was inhibited by the Epac2 inhibitors HJC0350 at 3 μM (IC₅₀=0.3 μM (Chen et al., 2013)) and ESI-05 at 10 μM (Fig. 12A).

The substance P release induced by 6-Bnz-cAMP was significantly increased by ESI-09, while CE3F4 showed a trend towards increasing it (p=0.100, Fig. 12B). The Epac2 inhibitors HJC0350 and ESI-05 had no significant effects (Fig. 12B).

The substance P release induced by CPTOMe-cAMP was eliminated by the Epac 2 inhibitor ESI-05 (10 μM), but the Epac2 inhibitor HJC0350 (3 μM) only produced a trend (p=0.086) towards inhibition. CE3F4 (30 μM) increased its effect on substance P release, and ESI-09 (25 μM) produced a trend towards an increase (p=0.056).

These results indicate that PKA is the main mediator of cAMP-induced substance P release, but that there is also a sizable contribution of Epac2. Intriguingly, the effects of the selective Epac1 inhibitor CE3F4 and the non-selective Epac inhibitor ESI-09 suggest that Epac1 inhibit substance P release.

3.12. Signals that mediate substance P release downstream from PKA and Epac2

Three signaling pathways have been proposed for Epac: 1) Epac → Rap → Raf-1 → MAPK (Lezoualc’h et al., 2016), 2) Epac → Rap → PLCε → PKCε (Schmidt et al., 2001; Wang et al., 2018) and 3) Epac → Rap → Akt (protein kinase B) → mTOR (Fields et al., 2015; Nijholt et al., 2008).

The PLC inhibitor U-73122 (Salter and Hicks, 1995) at 10 μM did not affect the substance P release evoked by forskolin (Fig. 13A), 6-Bnz-cAMP (Fig. 13B) or CPTOMe-cAMP (Fig. 13C), ruling out a role for pathway 2.

Figure 13. Involvement of Raf, MEK1 and mTOR in substance P release induced by forskolin, 6-Bnz-cAMP and CPTOMe-cAMP.

Spinal cord slices were incubated for 60 min with 10 μM forskolin (A), 10 nM 6-Bnz-cAMP (B) or 30 μM CPTOMe-cAMP (C), alone (“none”) or with the PLCε inhibitor U-73122 (10 μM), the Raf kinase inhibitors AZ-628 (1 μM) or ML-786 (30 nM), the MEK1 inhibitor U0126 (1 μM), or the mTOR inhibitor Torin-2 (100 nM). Numbers inside the bars indicate the number of slices in each group (n). ANOVA (A): p<0.0001, F (5, 47)=9.2; ANOVA (B): p<0.0001, F (5, 46)=9.8. ANOVA; (C): p<0.0001, F (5, 21)=9.5. Holm-Sidak’s post-hoc tests: * p<0.05, ** p<0.01, *** p<0.001 compared with “none”.

To determine if pathway 1 was involved, we inhibited Raf-1 kinase with 1 μM AZ-628 (Wenglowsky et al., 2014) or 30 nM ML-786, and MEK-1 with 1 μM U0126 (Davies et al., 2000). All three inhibitors decreased forskolin-induced substance P release (Fig. 13A) and practically abolished CPTOMe-cAMP-induced substance P release (Fig. 13C). In contrast, 6-Bnz-cAMP-induced substance P release was not affected by the MEK-1 inhibitor U016 or by the Raf-1 inhibitor ML-786, although it was decreased by the Raf-1 inhibitor AZ-628 (Fig. 13B). These results are consistent with the ideas that Epac2 signals through pathway 1 and that the effect of 6-Bnz-cAMP is mediated by PKA and not Epac 2. However, the unexpected inhibition of the effect of 6-Bnz-cAMP produced by AZ-628 suggests that a kinase inhibited by this compound is necessary to support the effect of PKA.

The mTOR inhibitor Torin-2 did not affect CPTOMe-cAMP-induced substance P release (Fig. 13C). It slightly decreased forskolin-induced substance P release (Fig. 13A) and strongly reduced 6-Bnz-cAMP-induced substance P release (Fig. 13B). These results indicate that Epac2 does not signal through pathway 3 and that the effect of PKA requires mTOR.

3.13. Epac1 inhibitors induced NK1R internalization independently of substance P release

Next, we determined if the Epac1 inhibitors CE3F4 and ESI-09 were able to induce substance P release by themselves instead of increasing the effect of forskolin. As shown in Fig. 14A, CE3F4 (30 μM) alone induced a large amount of NK1R internalization. In fact, CE3F4 produced the same effect with and without 10 μM forskolin. ESI-09 (25 μM) alone also induced a large amount of NK1R internalization, similar to that produced by forskolin. When added to forskolin, ESI-09 produced more NK1R than forskolin alone but no significant differences when compared to ESI-09 alone.

Figure 14. The Epac1 inhibitors CE3F4 and ESI-09 induce substance P release.

A: Spinal cord slices were incubated for 60 min with 10 μM forskolin, 30 μM CE3F4, 25 μM ESI-09, forskolin + CE3F4 or forskolin + ESI-09. ANOVA: p<0.0001, F (4, 57)=7.2. Holm-Sidak’s post-hoc tests: ** p<0.01, *** p<0.001 compared with “forskolin”. B: The slices were incubated for 60 min with 30 μM CE3F4 alone (“none”) or with the NK1R antagonist L-732,138 (1 μM), the Cav2 inhibitor ω-conotoxin MVIIC (1 μM), the Cav1 inhibitor diltiazem (1 μM), the Cav2.1 inhibitor ω-agatoxin TK (0.3 μM), the Cav3.2 inhibitor (2R/S)-6-PNG (10 μM), the NMDA receptor blocker MK-801, the TRPA1 antagonist AM-0902 (0.3 μM), the TRPV1 antagonist AMG9810 (0.3 μM) or the PKA inhibitor KT5720 (10 μM). ANOVA: p=0.283, F (9, 53)=1.25. Numbers inside the bars indicate the number of slices in each group (n).

We then investigated if the Epac1 inhibitor CE3F4 induces substance P release by the same mechanisms as forskolin. One key finding was that the NK1R internalization induced by 30 μM CE3F4 was not affected by the NK1R antagonist L-732,138 (1 μM, Fig. 14B), unlike the NK1R internalization induced by forskolin (Fig. 1A), 6-Bnz-cAMP (Fig. 1B) or 8-Br-cAMP (Fig. 13). This indicates that this NK1R internalization does not involve the interaction of an agonist with the NK1R. Moreover, CE3F4-induced NK1R internalization was not affected by any of the compounds expected to inhibit substance P release, such as the Cav channel blockers CTX-MVIIC, diltiazem and ω-agatoxin TK; the NMDA receptor blocker MK-801; the TRPA1 inhibitor AM-0902, or the TRPV1 inhibitor AMG9810 (Fig. 14B). In contrast to NK1R internalization induced by forskolin or 6-Bnz-cAMP, CE3F4-induced NK1R internalization was not affected by the PKA inhibitor KT5720 (Fig. 14B).

All of these data, and especially the lack of effect of the NK1R antagonist, indicate that CE3F4 does not induce substance P release but triggers NK1R internalization by an unidentified mechanism that does not involve Cav channels, NMDA receptors or PKA.

3.14. Changes in cAMP-induced substance P release during latent sensitization

To complete this study, we investigated changes in the ability of AC, PKA and Epac2 to induce substance P release during latent sensitization induced by CFA, a model of persistent hyperalgesia (Campillo et al., 2011; Corder et al., 2013; Marvizon et al., 2015; Walwyn et al., 2016). Latent sensitization is maintained by substance P release and NK1R activation (Chen and Marvizon, 2020a). Rats received 50 μl CFA in one hindpaw and the responses of their ipsilateral and contralateral hindpaws to von Frey filaments were followed until day 24. CFA induced strong mechanical allodynia in the ipsilateral paw that gradually decreased back to baseline, but it did not induce allodynia in the contralateral hindpaw (Fig. 15A). Once the responses were back to baseline, the rats received an injection of naltrexone (NTX, 1 mg/kg s.c.), which induced hyperalgesia in both hindpaws lasting more than 2 h (Fig. 15B). As previously established (Corder et al., 2013; Walwyn et al., 2016), this response to NTX confirmed the presence of latent sensitization.

Figure 15. Forskolin-induced substance P release in rats with latent sensitization (LS).

A. Rats (n=9) were injected in one hind paw with 50 μl CFA and mechanical allodynia was measured with von Frey filaments as the paw withdrawal threshold (PWT). Mixed-effect analysis (repeated-measures by both factors): CFA/time, p<0.0001, F (6, 48)=18; side, p=0.0006, F (1, 8)=29; CFA x side, p<0.0001, F (6, 24)=27. B. The rats received naltrexone (NTX, 1 mg/kg s.c.) and responses to von Frey filaments were measured. Mixed-effect analysis (repeated-measures by both factors): NTX/time, p=0.0003, F (4, 32)=7.2; side, p=0.125, F (1, 8)=2.9; NTX x side, p=0.49, F (4, 20)=0.89. Holm-Sidak’s post-hoc tests: * p<0.05, ** p<0.01, *** p<0.001 compared with time 0. C. Spinal cord slices prepared from rats with LS or from control rats were incubated for 60 min with 10 nM 6-Bnz-cAMP, 10 μM forskolin or 10 μM forskolin plus 3 μM HJC0350. NK1R internalization was measured. Two-way ANOVA: LS, p=0.0006, F (1, 95)=12; drugs, p=0.228, F (2, 95)=1.5; LS x drugs, p=0.0031, F (2, 95)=6.1. Holm-Sidak’s post-hoc tests: *** p<0.0001 comparing control with LS.

During days 25–27, the rats were killed and used to prepare spinal cord slices (6–8 slices per rat). To account for both inter-animal and intra-animal variability, only two slices from the same rat received the same drugs. NK1R internalization in slices from the rats with latent sensitization was compared with that obtained with the same drugs in slices from naïve rats in previous experiments (Fig. 15C). NK1R internalization induced by 6-Bnz-cAMP (10 nM) was the same in control rats and rats with latent sensitization (Fig. 15C), indicating that the ability of PKA to induce substance P release does not change in latent sensitization. However, NK1R internalization induced by forskolin (10 μM) was substantially decreased in rats with latent sensitization (Fig. 15C). Moreover, the inhibition of the effect of forskolin produced by the Epac2 inhibitor HJC0350 (3 μM) in control rats (which was statistically significant in Fig. 12A) disappeared in rats with latent sensitization (Fig. 15C). Substance P release in the presence of forskolin plus HJC0350 (which is due solely to PKA activation) was the same in control rats and rats with latent sensitization. This indicates that the increase in substance P release induced by Epac2 disappears in latent sensitization.

4. Discussion

This study shows that AC activation induces robust substance P release in the dorsal horn mediated by PKA and Epac2.

4.1. Activating AC and PKA induces substance P release

We have previously shown that μ-opioid receptors inhibit substance P release not only by inactivating Cav channels but also by inhibiting AC (Chen et al., 2018a). We now show that AC activation is sufficient to induce substance P release. This effect is mediated by PKA because it was also induced by its activators 8-Br-cAMP and 6-Bnz-cAMP. Moreover, the effects of forskolin, 8-Br-cAMP and 6-Bnz-cAMP were decreased by the PKA inhibitors KT5720 and PKI 14–22. These inhibitors also decreased substance P release induced by dorsal root stimulation, indicating that AC and PKA contribute to the substance P release evoked by noxious stimuli (Adelson et al., 2009).

However, the effect of forskolin was much slower than that of 6-Bnz-cAMP. Moreover, the effect of forskolin, but not the effect of 6-Bnz-cAMP, was partly decreased by lidocaine. This indicates that forskolin activates AC not only in presynaptic terminals but also away from the synapse, probably in the axons of primary afferents or in the substance P-expressing dorsal horn neurons (Dickie et al., 2019; Haring et al., 2018; Sathyamurthy et al., 2018). In contrast, 6-Bnz-cAMP induces substance P release by activating PKA only in presynaptic terminals. This idea is supported by the fact the forskolin-induced substance P release was inhibited by Cav channel blockers that did not inhibit the effect of 6-Bnz-cAMP. The cAMP analog 8-Br-cAMP, which is not selective for PKA, induced substance P release with a biphasic concentration-response curve, supporting the idea that cAMP acts on multiple targets to induce substance P release.

4.2. Source of substance P release

Substance P is released from the PEP1 population of peptidergic primary afferents (Usoskin et al., 2015) but also from ‘DE-11’ excitatory interneurons in lamina II (Dickie et al., 2019; Haring et al., 2018; Sathyamurthy et al., 2018). Since these neurons receive monosynaptic input from C-fibers, it is difficult to establish if substance P is released from them or from the primary afferents. Substance P release elicited by stimulating the dorsal root is not affected by the AMPA receptor antagonist CNQX (Marvizon et al., 1997). Since AMPA receptors mediate neurotransmission between primary afferents and second order neurons, the substance P elicited this way has to be from primary afferents. Unfortunately, this approach cannot be used to determine the source of substance P release induced by applying forskolin or cAMP analogs to the slices, which would reach both primary afferent terminals and dorsal horn neurons. Still, substance P release induced by dorsal root stimulation was decreased by a adenylyl cyclase inhibitor and a PKA inhibitor, showing that cAMP-induced substance P release is present in primary afferent and participates in substance P release evoked by their firing. Whether cAMP-induced substance P release is present in the DE-11 neurons remains to be determined.

4.3. cAMP and PKA induce substance P release by increasing NMDA receptor function

Our initial hypothesis was that AC and PKA induce substance P release by activating NMDA receptors, which are present in primary afferents and mediate substance P release (Lao et al., 2003; Liu et al., 1997; Malcangio et al., 1998; Marvizon et al., 1999; Marvizon et al., 1997). Indeed, the NMDA receptor blocker MK-801 decreased substance P release induced by forskolin, 8-Br-cAMP and 6-Bnz-cAMP. Our other results support the idea that these effects are due to the activating phosphorylation of NMDA receptors by PKA (Lau et al., 2009; Murphy et al., 2014): forskolin increased Ca²⁺ influx into DRG neurons through NMDA receptors and the phosphorylation of their NR1 and NR2B subunits, which was abolished by the PKA inhibitor H89.

Most of the NMDA receptors that induce substance P release have the NR2B subunit, as shown by the potent, albeit partial, inhibition of NMDA-induced substance P release by ifenprodil (Chen et al., 2018a). The NR2B subunit is abundantly expressed in DRG neurons (Marvizon et al., 2002) and is involved in the pro-nociceptive effect of NMDA receptors (Tan et al., 2005). Surprisingly, we found that ifenprodil inhibits forskolin-induced substance P release with the low potency corresponding to NMDA receptors that do not contain the NR2B subunit (Grimwood et al., 2000). Likewise, 10 nM ifenprodil did not inhibit substance P release induced by 6-Bnz-cAMP. A possible explanation is that the NMDA receptors that mediate cAMP-induced substance P release do not have the NR2B subunit but are nevertheless potentiated by PKA phosphorylation of their NR1 subunit. This would mean that there are two populations of NMDA receptors that induce substance P release: one with the NR2B subunit that mediates the effect of NMDA (Chen et al., 2018a; Chen et al., 2010) and another without the NR2B subunit that mediates the effect of forskolin and 6-Bnz-cAMP. In fact, a fraction of the NMDA receptors that induce substance P release are ifenprodil-insensitive (Chen et al., 2018a). Another explanation is that the NMDA receptors that mediate cAMP-induced substance P release contain the NR2B subunit, but PKA phosphorylation greatly decreases their affinity for ifenprodil. This could be similar to the tenfold decrease in the potency of ifenprodil at NMDA receptors induced by their phosphorylation by Fyn kinase (Li et al., 2006).

4.4. cAMP and PKA induce substance P release by activating Cav channels

In addition to NMDA receptors, Cav channels mediate cAMP-induced substance P release. However, there are important differences in the contribution of Cav channels to forskolin-induced and 6-Bnz-cAMP-induced substance P release. Cav1 (L), Cav2.1 (P-Q), Cav2.2 (N) and Cav3.2 (T) channels mediate forskolin-induced substance P release, whereas Cav2.2 are the only Cav channels mediating 6-Bnz-cAMP-induced substance P release. Cav2.3 (R) channels are not involved, whereas the δ2 subunit plays only a minor role in the effect of forskolin. Regulation by PKA of Cav1.2 channels (Dai et al., 2009) and Cav3.2 channels (Kim et al., 2006) has been reported. Inactivation of presynaptic Cav2.1 and Cav2.2 channels mediates the inhibitory effect of μ-opioid receptors in primary afferents (Heinke et al., 2011; Rusin and Moises, 1995). These results support the idea that the effect of 6-Bnz-cAMP is in the presynaptic terminals, while the effect of forskolin is both presynaptic and extrasynaptic.

TRPV1 in primary afferents induce large amounts of substance P release (Go and Yaksh, 1987; Lao et al., 2003; Marvizon et al., 2003; Yaksh et al., 1979) and are potentiated by PKA (Santha et al., 2010). However, we found that TRPV1 does not mediate the substance P release evoked by forskolin or 6-Bnz-cAMP. TRPA1 is also present in primary afferent terminals, although it is scarce in those containing substance P (Usoskin et al., 2015). Results with the TRPA1 inhibitor AM-0902 indicate that TRPA1 marginally contribute to substance P release evoked by 6-Bnz-cAMP, but not forskolin.

4.5. Epac2 increases substance P release

The selective Epac agonist CPTOMe-cAMP (Enserink et al., 2002; Singhmar et al., 2016) induced substance P release with an EC₅₀ of 5.2 μM, similar to its reported EC₅₀ for Epac of 2.2 μM (Enserink et al., 2002). However, the effect of CPTOMe-cAMP was decreased by PKA inhibitors, indicating that it requires PKA activation. Yet, it was inhibited by the Epac2 inhibitor ESI-05 and marginally decreased by the Epac2 inhibitor HJC0305 (Chen et al., 2013). A likely explanation is that substance P release induced by CPTOMe-cAMP is mediated by Epac2 but requires the simultaneous activation of PKA. This idea is supported by the fact that Epac2 inhibitors decreased substance P release induced by forskolin, but not by 6-Bnz-cAMP.

Three signaling pathways have been proposed for Epac: 1) Epac → Rap → Raf-1 → MAPK (Lezoualc’h et al., 2016), 2) Epac → Rap → PLCε → PKCε (Schmidt et al., 2001; Wang et al., 2018) and 3) Epac → Rap → Akt → mTOR (Fields et al., 2015; Nijholt et al., 2008). Since inhibitors of Raf kinase and MAPK decreased the substance P release induced by forskolin and CPTOMe-cAMP, the stimulatory effect of Epac2 is likely mediated by pathway 1 (Fig. 15). Pathway 2 is ruled out by the lack of effect of the PLCε inhibitor U-73122. The mTOR inhibitor Torin-2 did not decrease the effect of CPTOMe-cAMP, arguing against pathway 3 mediating the effect of Epac2. However, Torin-2 strongly inhibited the effect of 6-Bnz-cAMP and produced a smaller decrease in the effect of forskolin, suggesting that mTOR contributes to the effect of PKA on substance P release.

4.6. Epac1 inhibits NK1R internalization

The Epac1 inhibitors CE3F4 and ESI-09 (Singhmar et al., 2016) induced NK1R internalization by themselves. The effect of CE3F4 was not decreased by an NK1R antagonist and was not affected by blockers of Cav channels, NMDA receptors, TRPV1 or TRPA1. Therefore, its effect is likely not due to substance P release but to the direct induction of NK1R internalization. Since two chemically-unrelated Epac1 inhibitors produce the same effect, this effect may be due to a sustained inhibition of NK1R internalization by Epac1 (Fig. 15).

4.7. Signaling pathways that mediate cAMP-induced substance P release

Taken together, our results indicate that the activation of PKA by cAMP induces Ca²⁺ entry in the presynaptic terminals through NMDA receptors and Cav2.2 channels, with a possible minor contribution of TRPA1 (Fig. 15). Activation of mTOR also seems to be required for the effect of PKA, but the signals upstream from mTOR remain to be identified. In addition, cAMP activates another pathway initiated by Epac2 and involving Rap, Raf-1, MEK and MAPK (see Graphic Abstract). This increases substance P release by activating Cav1, Cav2.1 and Cav3.2 channels, likely through their phosphorylation by MAPK. Indeed, there is evidence that the Raf-1 → MEK → MAPK pathway activates Cav channels in DRG neurons (Fitzgerald, 2000). However, this Epac2 pathway seems to require the activation of PKA, because PKA inhibitors decreased substance P release induced by the Epac activator CPTOMe-cAMP.

4.8. Role of cAMP-induced substance P release in hyperalgesia

We recently reported that NK1Rs are necessary for the maintenance of latent sensitization to pain induced by CFA or by spared nerve injury (Chen and Marvizón, 2020a). Thus, the NK1R antagonist RP67580 eliminated the reinstatement of allodynia by NTX even after one week. This is consistent with a study showing that ablating NK1R neurons in the dorsal horn eliminated the prolonged hyperalgesia induced by fentanyl or paw incision (Rivat et al., 2009), two stimuli that induce latent sensitization. Moreover, rats with CFA-induced latent sensitization had higher basal levels of NK1R internalization than control rats, which were increased by NTX (Chen and Marvizon, 2020a). Taken together, these data suggest that increased substance P release maintains pain hypersensitivity during latent sensitization.

There is a discrepancy between animal studies that show that NK1R antagonists have antihyperalgesic effects and clinical trials with some of these antagonists (Hill, 2000; Urban and Fox, 2000). Although this was attributed to species differences, the NK1R neurons in the human spinal cord are similar to the ones in rodents (Ding et al., 1999; Todd et al., 2000). The clinical trials were limited both in the types of pain studied and the number of NK1R antagonists used, and for 20 years there have been no attempts to expand them. Dionne et al. (1998) studied the effect of the antagonist CP-99,994 on short-term pain (4 h after dental extraction) and found that it produced significant analgesia. Three other studies used the NK1R antagonist lanepitant given orally, when it is unclear if it can cross the blood-brain barrier. The effect of lanepitant on migraine was close to statistical significance (p=0.065) (Goldstein et al., 2001a). However, it was ineffective on pain produced by osteoarthritis (Goldstein et al., 2000) and diabetic neuropathy (Goldstein et al., 2001b).

Since latent sensitization is mediated by AC (Celerier et al., 2000; Corder et al., 2013), we hypothesized that AC also mediates the increased substance P release during latent sensitization. This hypothesis predicted that the ability of cAMP to induce substance P release would increase during latent sensitization. However, our results did not confirm this prediction. We found the opposite: substance P release induced by forskolin actually decreased during latent sensitization. This decrease was not due to a decrease in the effect of PKA, since there was no change in 6-Bnz-cAMP-induced substance P release, but to the disappearance of the stimulatory effect of Epac2.

A possible interpretation of these findings is that latent sensitization does require an increase in NMDA receptor-induced substance P release, but that the increase in NMDA receptor function is not driven by the AC-PKA pathway but by a pathway in which BDNF increases NMDA receptor phosphorylation by the Src family kinase Fyn (Abe et al., 2005; Chen et al., 2014; Chen et al., 2010; Guo et al., 2002). Indeed, we found that the Src family kinase inhibitor PP2 abolishes latent sensitization induced by CFA or by spared nerve injury (Chen and Marvizón, 2020b).

4.9. Conclusions

Substance P release in the spinal dorsal horn is induced by cAMP acting on PKA and Epac2, leading to the activation of NMDA receptors and Cav channels. Substance P release induced by cAMP decreases during latent sensitization.

Highlights.

• cAMP induced substance P release through protein kinase A and Epac2.

• Protein kinase A activated NMDA receptors and N type voltage-gated Ca²⁺ channels.

• Epac2 activated L, P, Q and T type voltage-gated Ca²⁺ channels through MAPK.

• In a chronic pain model, signaling through protein kinase A was not changed.

• In a chronic pain model, signaling through Epac2 was eliminated.

Abbreviations:

6-Bnz-cAMP

6-benzo-cyclic adenosine monophosphate

8-Br-cAMP

8-bromo-cyclic adenosine monophosphate

AC

adenylyl cyclase

aCSF

artificial cerebrospinal fluid

Akt

protein kinase B

ANOVA

analysis of variance

BDNF

brain-derived neurotrophic factor

cAMP

cyclic adenosine monophosphate

Cav

voltage-gated calcium channel

CFA

complete Freund’s adjuvant

CI

confidence interval

CNQX

cyanquixaline (6-cyano-7-nitroquinoxaline-2,3-dione)

CPTOMe-cAMP

8-pCPT-2-O-Me-cyclic adenosine monophosphate sodium salt

DAMGO

[D-Ala2, NMe-Phe4, Gly-ol5]-enkephalin

DMSO

dimethyl sulfoxide

DRG

dorsal root ganglia

EC₅₀

concentration of drug that produces half of the effect

Epac

exchange protein activated by cAMP

GABA

gamma-amino-butyric acid

H89

N-[2-[[3-(4-Bromophenyl)-2-propenyl]amino]ethyl]-5-isoquinolinesulfonamide dihydrochloride

IC₅₀

concentration of drug that produces half of the inhibition

NK1R

neurokinin 1 receptor

NMDA

N-methyl-D-aspartate

MAPK

mitogen-activated protein kinase

MEK

mitogen-activated protein kinase kinase

mTOR

mammalian target of rapamycin

PLCε

phospholipase C epsilon

PKA

protein kinase A

PKI 14–22

protein kinase inhibitor 14–22

PKCε

protein kinase C epsilon

Rap

Rap GTP-binding protein

Raf-1

RAF proto-oncogene serine/threonine-protein kinase

SDS

sodium dodecyl sulfate

TBS

Tris-buffered saline

TRPA1

transient receptor potential ankyrin 1

TRPV1

transient receptor potential vallinoid 1