Pro- versus Antinociceptive Nongenomic Effects of Neuronal Mineralocorticoid versus Glucocorticoid Receptors during Rat Hind Paw Inflammation

CORTICOSTEROIDS are commonly used drugs for the effective treatment of chronic pain and inflammation.^1,2  They mainly mediate their antiinflammatory and immunomodulatory effects via genomic mechanisms by binding to cytosolic glucocorticoid receptors (GRs), translocating to the nucleus, and influencing the transcription of specific genes, a process that usually takes several hours or days.^3,4  Recently, this traditional model of corticosteroid mechanism was significantly extended due to emerging evidence of rapid nongenomic effects⁵  that are mediated by membrane-bound receptors, e.g., on ion channels within the hippocampus of the brain.^6,7  Consistently, in vitro experiments showed that corticosterone immediately inhibited N-methyl-d-aspartate receptor currents in cultured hippocampal neurons independent of gene expression.⁸  Moreover, Bartholome et al.⁹  showed that normal human peripheral blood mononuclear cells express membrane GRs and that their number is upregulated in patients with rheumatoid arthritis.

Increasing evidence is accumulating that corticosteroids also affect acute postoperative pain.^2,10,11  Perioperative dexamethasone administration acutely reduced postoperative pain and opioid consumption.^2,12  Interestingly, perineural dexamethasone appeared to be more effective than intravenous dexamethasone, suggesting a direct effect on nociceptive neurons.¹²  Indeed, extending previous reports of GR and mineralocorticoid receptor (MR) localization on peripheral neurons,^13,14  we could show that MRs are predominantly expressed in unmyelinated peptidergic nociceptive neurons, whereas GRs are prominent in both peptidergic and nonpeptidergic nociceptive neurons of naive rats. Both receptors apparently exist only marginally in myelinated tyrosine kinase receptor B-immunoreactive (ir) mechanoreceptive and tyrosine kinase receptor C-ir proprioreceptive dorsal root ganglion (DRG) neurons as well as in spinal microglia and astrocytes.^15,16  Within the spinal cord, MR colocalized predominantly with incoming presynaptic calcitonin gene–related peptide (CGRP)—ir nerve terminals in Rexed laminae I and II, while GR localized in pre- and postsynaptic nociceptive neurons within the dorsal horn of the spinal cord.

Here, we investigated the hypothesis whether a condition of persistent nociceptive stimulation, e.g., Freund’s complete adjuvant (FCA)–induced rat hind paw inflammation, leads to changes in GR and MR expression in spinal and DRG neurons as well as glia cells and astrocytes. In addition, the current study examined whether the administration of GR and MR agonists as well as antagonists or their combination would acutely modulate the mechanical hypersensitivity induced by the inflammatory process. To this end, we used dexamethasone as a GR agonist that binds more strongly to the GR than corticosterone¹⁷  and is widely used for its antiinflammatory glucocorticoid effects¹⁸  and mifepristone as a known GR antagonist.¹⁹  In addition, we used aldosterone as a selective MR agonist²⁰  and potassium canrenoate (canrenoate-K), which is a parenteral formulation and prodrug, for the MR antagonist spironolactone.²¹ 

Freund’s complete adjuvant, a water-in-oil emulsion with killed mycobacteria (Calbiochem, USA), was injected (150 µl) intraplantarly during brief isoflurane anesthesia (1.0 to 2.5 vol% isoflurane; Abbott, Germany); aldosterone, canrenoate-K, dexamethasone, and mifepresterone were commercially obtained from Sigma-Aldrich (USA); aldosterone, dexamethasone, and mifeprestone were dissolved in a vehicle composed of 1% dimethyl sulfoxide, 10% ethanol, and 0.9% NaCl as described previously.^15,16  Canrenoate-K was dissolved in 0.9% NaCl. Drugs were administered subcutaneously in a volume of ml/kg body weight. Subcutaneous injections were given under the loose skin of the back just behind and between the shoulder blades.

Experiments were conducted in male Wistar rats (180 to 250 g; breeding facility, Charité University Berlin, Berlin, Germany) following approval by the local animal care committee and in accordance with the European Directive introducing new animal welfare and care guidelines (2010/63/EU; available at http://ec.europa.eu/environment/chemicals/lab_animals/pdf/guidance/animal_welfare_bodies/en.pdf, accessed January 15, 2018). Wistar rats were anaesthetized by brief isoflurane anesthesia and received an intraplantar injection of 0.15 ml FCA (Calbiochem) into the right hind paw. This treatment consistently produces a local inflammation of the inoculated paw characterized by an increase in paw volume, paw temperature, mechanical hypersensitivity, and infiltration of various types of immune cells.²²  All experiments were performed on the fourth day of FCA hind paw inflammation.

The first set of experiments examined the influence of four days of FCA hind paw inflammation on changes in the neuronal versus glial GR and MR expression in the spinal cord and peripheral sensory neurons (see timeline in Supplemental Digital Content 1, http://links.lww.com/ALN/B624). Subsequent experiments assessed the impact of subcutaneous injections of increasing doses of GR agonist dexamethasone (0.1 to 2 mg/kg), GR antagonist mifepristone (1.0 to 4.0 mg/kg), MR agonist aldosterone (0.5 to 2 mg/kg), or MR antagonist canrenoate-K (10 to 40 mg/kg) on mechanical paw pressure thresholds (PPTs). The doses chosen for each drug were based on previous pilot experiments. A final set of experiments investigated the concomitant subcutaneous injection of the calculated ED₅₀ of those drugs that showed significant antinociceptive effects. PPTs were given either in grams or in percentage maximum possible effect MPE (MPE = PPT_{postinjection} – PPT_baseline/70 – PPT_baseline), and values were determined before and 15, 30, 60, and 120 min after drug injections.

The mechanical hyperalgesia after FCA hind paw inflammation was assessed by a pressure apparatus (Ugo-Basile SRL, Italy) with increasing force (measured in grams) applied to the plantar hind paw until a withdrawal reflex was precipitated as described previously.²³  A 250-g cutoff value was assigned in the absence of a response. The investigator was trained to apply the tip perpendicular to the central area of the hind paw with a gradual increase in pressure. The endpoint was characterized by the spontaneous removal of the paw. After paw withdrawal, the intensity of the pressure was recorded manually. The rat was loosely wrapped in a towel and the hind paw placed on the apparatus’s pedestal until the rat totally calmed down (relaxed tail, no squeaking, no tremble response). A plinth was gently lowered onto the hind paw, and then the investigator activated the apparatus. Mechanical paw pressure tests were performed in all groups on the fourth day of FCA hind paw inflammation, but before drug application, to get baseline values in both inflamed and contralateral noninflamed hind paws. Then the mechanical PPTs were reassessed at different time points (0 to 120 min) after subcutaneous drug administrations to determine drug-related behaviors. Each measurement was performed consecutively three times with a time interval of at least 15 s before the next stimulus was applied. The PPTs were defined as the mean of five to seven animals examined before and after subcutaneous drug injection. In all behavioral experiments, drugs were prepared by one of the investigators (M. Shaqura), and the examiner (X.L.) was unaware of the treatment that each animal received by chance; however, no randomization methods were used to assign animals to condition.

Total RNA was extracted from L3–L5 dorsal root ganglia of Wistar rats using an RNeasy Kit (Qiagen, Germany) as previously described.^15,24  The following specific primers were used: for GR, forward primer, 5′-catcttcagaac agcaaaatcga-3′; reverse primer, 5′-aggtgctttggtct gtgggata-3′ (Ensemble Accession No. NM_012576.2); and for MR, forward primer, 5′-CCAAGGTACTTCCAGGAT TTAAAAAC-3′; reverse primer; 5′-AACGATGATAGACACATCCAAGAATACT-3′ (Ensemble Accession No. NM_013131.1). Taqman Real-Time PCR was performed with a SYBR Green kit following the manufacturer’s instructions (Applied Biosystems, USA). Amplification was carried out for 40 cycles, each consisting of 15 s at 95°C; for GR and 18S gene, cycles consisted of 30 s at 60°C. A temperature just less than the specific melting temperature (Tm) was employed for detection of fluorescence-specific products (GR: Tm 83°C, 18s: Tm 88°C; MR: Tm 76°C, 18 s: Tm 83°C). GR and MR messenger RNA (mRNA) were quantified using triplicates of samples and using the delta-delta CT method.¹⁶  The housekeeping gene 18s (forward primer, CGGCTA CCACATCCAAGGAA; reverse primer, GCTGGAATTACCGCGGCT; Accession No. NR_046237) was used as an internal reference gene.

In deep isoflurane anesthesia, rats (n = 15 to 17) were euthanized, and blood samples were taken by intracardiac aspiration into rapid serum tubes (Vacutainer, Becton Dickinson, Germany), then centrifuged at 1,000 rcf and quickly frozen at –80°C. Samples were sent to a commercial chemistry laboratory (SynLab Vet, Berlin, Germany), which determined corticosterone plasma concentrations by liquid tandem mass spectrometry and aldosterone plasma concentrations by chemiluminescence immunoassay.

Dorsal root ganglia and spinal cords from adult rats were solubilized as described by Mousa et al.²⁴  to obtain total cell protein. Western blot analysis was performed as previously described.^15,24  After blotting, the membranes were blocked in 3% bovine serum albumine for 2 h and incubated with rabbit anti-GR (a gift from M. Kawata, Kyoto Prefectural University of Medicine, Kyoto, Japan; 1:4,000 in 3% bovine serum albumine) or rabbit anti-MR overnight at 4°C. This antibody has been proven in different cell lines to be highly specific after cell transfection and knock-down.^25,26  The GR antibody (M. Kawata) has previously been shown in COS-1 cells with or without GR transfection to be highly specific.²⁵  After incubation with the secondary antibody (peroxidase-conjugated goat antirabbit, 1:40,000, Jackson ImmunoResearch, USA) for 2 h at room temperature, reactive protein bands were digitally visualized using enhanced chemiluminescence solutions (SuperSignal West Pico, Thermo Fischer Scientific, Germany) in a ChemiDoc MP Imager (Bio-Rad, Germany). Finally, the blots were reprobed with monoclonal mouse anti–ß-actin antibody (1:20,000; Sigma-Aldrich, Germany) as an internal standard. Experiments were done in three to five animals per group in duplicate.

The Western blot bands specific for MR (107 kDa) or GR (87 kDa) were quantified by Java Image processing and analysis software (ImageJ, open-source image software at http://rsb.info.nih.gov/ij/; accessed June 28, 2010).^15,16  The area and density of pixels within the threshold values representing immunoreactivity were measured, and the integrated density (the product of the area and mean of gray values) was calculated. Integrated immunodensities of controls and treated groups were compared and statistically analyzed.²⁷ 

After transcardial perfusion of rats (four days after FCA inoculation) spinal cord and DRG were removed and further processed as described previously.^15,24  Tissue sections were then incubated overnight with the following primary antibodies (Supplemental Digital Content 2, http://links.lww.com/ALN/B625): polyclonal rabbit anti-GR (private gift from M. Kawata, Kyoto, Japan; this antibody has previously been shown in COS-1 cells with or without GR transfection to be highly specific),²⁵  anti-GR (sc-1004, Santa Cruz Biotechnology, USA),²⁸  anti-MR (M. Kawata),²⁹  or anti-MR (sc-11412, Santa Cruz Biotechnology; this antibody has previously been proven to be highly specific after MR transfection in different cell lines)³⁰  in combination with a polyclonal guinea pig anti-CGRP, monoclonal mouse anti-CD11b (OX-42) or anti-GFAP, anti-GluN1/NR1 (NMDAR), and anti-GluR1 (GluR) or anti-PSD-95; in addition, polyclonal rabbit antibody against GR was examined in combination with the monoclonal mouse anti-MR (private gift from Celso E. Gomez-Sanchez, University of Mississippi, Jackson, Mississippi). Finally, the tissues were washed in phosphate buffer saline, mounted in VECTASHIELD (Vector Laboratories, USA), and imaged on a confocal laser scanning microscope, LSM510 (Carl Zeiss GmbH, Germany), as described previously.¹⁵  To demonstrate specificity of staining, the following controls were included as described in our previous studies:²⁴  omission of either the primary antisera or the secondary antibodies. Reactive microglia or astrocytes in the dorsal horn of the rat spinal cord were identified by hypertrophied cell bodies equipped with poorly ramified, short and thick processes in comparison to inactive ones indicated by weakly stained cells with small cell bodies and ramified thin processes.^31,32 

The quantification of DRG staining has been described previously.^15,24  For quantification of MR immunoreactivity in the dorsal horn of the spinal cord, single optical slice images of red immunofluorescence were obtained using a Zeiss LSM 510 laser scanning microscope (10× or 20× Plan-Neofluar [Carl Zeiss GmbH] air interface or 40× Plan-Neofluar oil interface objective lens) and the image-analysis software package 2.5 SP2 from Zeiss as described in detail elsewhere.²⁷  The settings of the confocal microscope were established using a control section and kept unchanged for all subsequent acquisitions. Images from spinal cord sections were adjusted to a threshold to exclude background fluorescence and gated to include intensity measurements only from positively stained cells.³³  For image analysis, a standardized area was positioned over the Rexed laminae 1 to 5 of all groups to determine the mean product of the area (µm²) and mean intensity of pixels within the threshold value and to calculate the integrated optical intensity (product of area and mean intensity). Quantification of immunofluorescent colocalization of GR or MR with different markers in DRG tissue sections was performed by using the Zeiss Zen 2009 software (Carl Zeiss Micro-Imaging GmbH, Germany). For counting of the total number of neurons, only those immunostained neurons containing a distinct nucleus were counted. In a similar way, the number of GR-ir/total DRG, MR-ir/total DRG, GR/MR-ir/total DRG, CGRP/GR-ir, CGRP/MR-ir, MR/GFAP-ir or M/OX-42-ir, GR/GFAP-ir or GR/OX-42–ir neurons/total DRG neurons were counted in each DRG or spinal cord sections and represented as percentages. Data were obtained from three to four sections per rat and five rats per group using a 40× objective lens.

Dorsal root ganglion membranes were obtained from the lumbar (L3–L5) DRG. Saturation binding experiments were performed using the MR ligand [³H]aldosterone (specific activity 77.4 Ci/mmol, Hartmann Analytic, Germany) or the GR ligand [³H]corticosterone (specific activity 40 Ci/mmol, Hartmann Analytic) as previously described.^15,16  Briefly, 50 to 100 µg membrane proteins were incubated with various concentrations of [³H]aldosterone (1.25 to 40 nM) or [³H]corticosterone together with unlabeled aldosterone (10 µM) or corticosterone (10 µM), respectively, in a total volume of 1 ml of binding buffer (50 mM Tris-HCl, 5 mM EDTA, 5 mM MgCl₂, 100 mM NaCl, 0.2% bovine serum albumin). Nonspecific binding was defined as radioactivity remaining bound in the presence of 10 μM unlabeled corticosterone or aldosterone. At the end of the incubation period (1 h at 22°C for [³H]corticosterone or 2 h at 30°C for [³H]aldosterone), bound and free ligands were separated by rapid filtration over Whatman brand Grade GF/C glass filters (Sigma-Aldrich) under vacuum using a cell harvester (Brandel, USA). Filters were washed three times with 4 ml of cold buffer (50 mM Tris-HCl, pH 7.4). Bound radioactivity was determined by liquid scintillation spectrophotometry after overnight extraction of the filters in 3 ml of scintillation fluid. All experiments were performed in duplicate and carried out at least five times. Maximal specific binding sites and dissociation or affinity konstant values in saturation binding assays were determined by nonlinear regression analysis of concentration-effect curves using GraphPad Prism (GraphPad Software Inc., USA).

All tests were performed using Sigma Stat 2.03 software (SPSS Inc., Germany). Taqman PCR, Western blot, and immunofluorescence data were analyzed as two group comparisons (FCA-treated rats vs. controls) by the two-tailed independent Student’s t test in the case of normally distributed data. The hypothesis tested presumed differences between FCA-treated rats and controls, and a minimum number of five animals per group was chosen based on previous experience. Results of saturation binding experiments were subjected to a nonlinear regression analysis followed by generation of the saturation binding equation according to the formula Y = Bmax · X / (Kd + X) (GraphPad Prism version 6.0). PPTs were determined before and after drug injections, expressed as mean ± SD, and statistically analyzed for alterations over time compared to baseline values by two-way repeated measures ANOVA followed by the Bonferroni test for multiple pairwise comparisons. The hypothesis tested presumed alterations over time immediately after drug injections compared to baseline (0 min), and a minimum number of five animals (commonly five to seven) per group were necessary for statistical analysis based on previous experience. Occasionally, rats were lost during anesthetic events for drug injections (n = 5 rats). For dose-dependent effects, linear regression analysis was performed. For all statistical tests, significance was assumed at P < 0.05.

Western blot analyses of the dorsal horn of the spinal cord identified a single GR-specific protein band at 87 kDa and a single MR-specific protein band at 107 kDa, consistent with previous reports (fig. 1).^15,16  The integrated optical density of the GR and MR protein bands increased significantly by 35 ± 10% and 17 ± 9%, respectively, in FCA-treated rats compared to controls (P < 0.05; fig. 1, A and B). This upregulation is consistent with results from single immunofluorescence confocal microscopy of spinal cord sections showing an increased number of GR-ir neuron-like structures (P < 0.05; fig. 1, C–E) and a significantly elevated integrated optical density of MR immunoreactivity (P < 0.05; fig. 1, F–H).

Immunohistochemical analyses of spinal cord sections from FCA-treated rats demonstrated abundant colocalization of MR with CGRP, a marker for nociceptive neurons, in dorsal horn Rexed laminae I and II (Supplemental Digital Content 1, http://links.lww.com/ALN/B624). GR but not MR predominantly colocalized with several neuronal markers for synaptic proteins: glutamate receptor subtype N-methyl-d-aspartate (NMDA), the glutamate receptor 1 (GluR), and postsynaptic density protein 95 (PSD; Supplemental Digital Content 3, http://links.lww.com/ALN/B626). GR and MR colocalized only scarcely in spinal cord sections with OX-42–ir microglia and GFAP-ir astrocytes of control rats (figs. 2 and 3). After FCA inflammation, there was a minor but not significant rise in the colocalization of GR with OX-42–ir microglia and GFAP-ir astrocytes (3 ± 2% and 3 ± 4%, respectively; P < 0.05; fig. 2, A–H). In addition, immunohistochemical colocalization of MR with OX-42–ir microglia and GFAP-ir astrocytes increased significantly but not considerably after inflammation (only 6 ± 2% and 9 ± 1%, respectively; P < 0.05; fig. 3, A–H). In spite of the higher increases in the number of reactive microglia (FCA: 49 ± 4%; controls: 29 ± 4%) and astrocytes (FCA: 63 ± 5%; controls: 33 ± 5%) in rats with FCA hind paw inflammation (fig. 2, C and G; fig. 3, C and G), there was a relatively small increase in GR and MR colocalization with spinal microglia and astrocytes (fig. 2, D and H; fig. 3, D and H).

Quantitative Taqman RT-PCR of GR and MR mRNA revealed a significant increase in both GR and MR mRNA in dorsal root ganglia of rats with FCA hind paw inflammation compared to naive control rats (P < 0.05; fig. 4, A and B). This was consistent with a significant increase in GR and MR specific protein bands at 87 and 107 kDa, respectively, in dorsal root ganglia of FCA-treated rats compared to controls (P < 0.05; fig. 4, C and D).

Consistently, double immunofluorescence confocal microscopy of DRG sections revealed that the percentage of GR- and MR-ir cells in relation to the total number of DRG cells per section was significantly increased in FCA rats (65 ± 2% and 53 ± 2%, respectively) compared to controls (57 ± 2% and 45 ± 2%, respectively; P < 0.05; fig. 4, E–J). These GR- and MR-ir DRG cells showed strong colocalization with the sensory neuronal marker CGRP (controls: 65 ± 6% and 85 ± 5%, respectively) that further increased after FCA hind paw inflammation (FCA: 83 ± 5% and 97 ± 5%, respectively; Supplemental Digital Content 4, http://links.lww.com/ALN/B627).^15,16  Finally, the percentage of MR-ir DRG neurons that colocalized with GR immunoreactivity significantly increased from 31 ± 2% of total DRG neurons in controls to 42 ± 2% of total DRG neurons in FCA rats (P < 0.05; fig. 4, K–M),

Quantitative Taqman RT-PCR also identified the “prereceptor regulating” enzyme³⁴  11-betahydroxysteroid dehydrogenase type 2 (11-ßHSD2) in DRG neurons of control animals and revealed a significant upregulation of its mRNA in DRG of FCA-treated rats (P < 0.05; fig. 5A). Consistently, protein bands of 11-ßHSD2 were identified at the expected molecular weight of 50 to 57 kDa; however, they were not significantly different between FCA-treated and control rats (fig. 5B). In parallel with MR-ir DRG neurons, which colocalized with CGRP in double immunofluorescence confocal microscopy,¹⁶  11-ßHSD2–ir DRG neurons colocalized with CGRP in 58 ± 5% of control animals and in 62 ± 6% of rats with FCA hind paw inflammation (P < 0.05; fig. 5, C and D). Liquid chromatography and tandem mass spectrometry showed no significant changes in corticosterone and aldosterone plasma concentrations in FCA-treated rats versus controls (P > 0.05, fig. 5E).

Membrane GR and MR receptor–binding experiments in DRG neurons of FCA and control animals demonstrated saturable membrane binding with the radiolabeled GR and MR ligands [³H]corticosterone and [³H]aldosterone, respectively. While the Kd affinity values for ligand receptor binding were not significantly different between the various conditions, the Bmax values per fmol/mg tissue for MR were significantly higher in rats with painful FCA inflammation compared to controls (P < 0.05; fig. 6).

Freund’s complete adjuvant inflammation of the right hind paw resulted in diminished mechanical PPT thresholds in inflamed but not in contralateral noninflamed hind paws (P < 0.05; fig. 7). This enhanced mechanical hypersensitivity was dose-dependently reversed within 15 min after the subcutaneous administration of the GR selective agonist dexamethasone, persisted up to 60 min (P < 0.05; fig. 7A), and could be antagonized by simultaneous administration of the GR selective antagonist mifeprestone (PPT: 70 ± 17.8% MPE for 0.78 mg/kg dexamethasone plus vehicle vs. PPT: 7.2 ± 5.8% MPE for 0.78 mg/kg dexamethasone plus 6 mg/kg mifeprestone, P < 0.05, two-tailed independent Student’s t test). On the contralateral noninflamed side, subcutaneous dexamethasone showed no significant effect on mechanical nociceptive thresholds (fig. 7B), similar to previous reports in naive rats.¹⁶  In contrast to dexamethasone’s effect, subcutaneous administration of the MR selective agonist aldosterone showed a further decrease in the already reduced mechanical PPT thresholds of the inflamed hind paws and a similar decrease in PPT thresholds of the contralateral noninflamed hind paws (P < 0.05; fig. 7, C and D). This effect could be antagonized by simultaneous administration of the MR selective antagonist canrenoate-K (PPT: –19 ± 15.2% MPE for 1.8 mg/kg aldosterone plus vehicle vs. PPT: 3.3 ± 5.9% MPE for 1.8 mg/kg aldosterone plus 13.5 mg/kg canrenoate-K, P < 0.05, two-tailed independent Student’s t test). In contrast, the enhanced mechanical hypersensitivity in the inflamed paw was reversed by subcutaneous administration of the MR selective antagonist canrenoate-K alone (P < 0.05; fig. 8, A and B), but not of the GR selective antagonist mifeprestone alone (fig. 8, C and D). Both substances had no effect on the contralateral noninflamed hind paw, similar to previous reports in naive rats.¹⁵  Dose-dependent increases of the diminished mechanical PPT were demonstrated for the GR agonist dexamethasone and the MR antagonist canrenoate-K (fig. 9). These dose-response curves shifted to the right when combined with the ED₅₀ of the other drug (according to an GR/MR ED₅₀ ratio of 0.78:13.4 = 1:17, respectively, or vice versa) (fig. 9, A–D). Thus, their combination resulted neither in a superadditive (synergistic) nor in an additive but in a subadditive effect in the isobolographic analysis (fig. 9E).

This is the first systematic investigation of alterations in neuronal versus glial GR and MR expression, membrane localization, and subsequent modulation of mechanical hypersensitivity during local inflammation. Our findings demonstrate significant increases in neuronal GR and MR both in the spinal cord and in the dorsal root ganglia ipsilateral to the inflammation (figs. 1 and 4). In parallel, the enhanced mechanical sensitivity of inflamed hind paws was attenuated immediately after GR agonist and MR antagonist administration, suggesting acute nongenomic effects consistent with detected membrane-bound GR and MR.

GR and MR had been identified in peripheral sensory neurons some time ago,^13,14  and their predominant expression in specific subpopulations of nociceceptive neurons has been recently substantiated;^15,16  however, so far there is no information on alterations in their expression under inflammatory painful conditions. Only studies using animal models of neuropathic pain provide some evidence of increased GR expression in DRG and/or spinal cord,^35,36  but information on MR expression is lacking.

Apart from neurons, GR and MR are scarcely expressed in spinal OX-42–ir microglia and GFAP-ir astrocytes of naive animals.^15,16  After FCA-induced inflammation, GR and MR expression in OX-42–ir microglia and GFAP-ir astrocytes is upregulated to only a minor extent (~5 to 15%) compared to their total number in spinal cord sections. This is intriguing in light of the fact that FCA inflammation elicited a nearly twofold increase in the number of reactive spinal OX-42–ir microglia cells and GFAP-ir astrocytes (figs. 2 and 3).³⁷  Thus, enhanced GR and MR expression does not primarily occur in spinal microglia and astrocytes, consistent with a previous report showing a lack of glucocorticoid response in activated microglia, presumably due to low GR expression in a model of optical nerve injury.³⁸  Interestingly, exposure of cultured brain-derived microglial cells to the inflammatory mediator lipopolysaccharid resulted in a downregulation of GR and MR mRNA transcripts.³⁹  Therefore, the expression and relevance of GR and MR in spinal and DRG neurons (~40 to 60% of total neurons) under painful inflammatory conditions appear to be more crucial than in spinal microglia and astrocytes (~5 to 15%). This observation is consistent with findings reported for a mouse model of spinal nerve ligation in which GR expression mainly occurred in neurons (~65%) and much less in microglia (~16%) or astrocytes (~8%).³⁶ 

Most important, our results showed a twofold increase in MR membrane binding sites in DRG neurons of rats with FCA inflammation compared to naive animals,^15,16  while the relatively high number of [³H]corticosterone binding sites remained unchanged (fig. 6). These findings validate previous reports on GR ligand binding to brain membranes⁴⁰  and on GR- and MR-enriched densities on neuronal cell membranes as detected by electron microscopy.^41,42  Since MRs are activated not only by aldosterone but also by corticosterone, and the plasma concentration of corticosterone is at least twofold higher than that of aldosterone,⁴³  MR activation would occur tonically. However, it is established that MR activation is prevented by the corticosterone inactivating enzyme 11-ßHSD2.³⁴  Our experiments identified 11-ßHSD2 mRNA transcripts as well as specific protein bands in nociceptive DRG neurons (fig. 5), similar to kidney tissue.⁴⁴  These 11-ßHSD2–ir neurons predominantly colocalized with the nociceptive neuron marker CGRP (~58 to 62%) in a similar way to MR. After painful inflammation, we found an increase in 11-ßHSD2 mRNA and in the number of 11-ßHSD2–ir neurons, without changes in the plasma concentrations of aldosterone and corticosterone (fig. 5). In the past, neuronal expression of 11-ßHSD2 has been demonstrated mainly in the nucleus of the solitary tract⁴⁵  and in the hypothalamic paraventricular nucleus,⁴⁶  but not in peripheral nociceptive neurons.

Apart from GR’s and MR’s well-known genomic effects,⁴⁷  the question arises: What are the functional consequences of neuronal membrane–bound MR and GR in nociceptive neurons? In recent publications, we have outlined that local application of the GR agonist dexamethasone and MR agonist aldosterone at the peripheral as well as central nociceptive nerve terminals results in rapid modulation of mechanical or chemical painful hypersensitivity.^15,16  While this was tested in naive animals, our goal in this study was to examine the impact of GR and MR activation on inflammatory pain. Subcutaneous application of the GR agonist dexamethasone dose-dependently reversed the FCA-induced mechanical hyperalgesia within 15 min, suggesting a rapid nongenomic effect, whereas the contralateral noninflamed side was not affected. Thus, the GR agonist dexamethasone effectively reversed the mechanical hyperalgesia only under conditions of persistent pain and ongoing firing such as in FCA-induced inflammation (fig. 7). This is in line with previously demonstrated suppression of spontaneous ectopic firing in nerve end neuromas,⁴⁸  with the suppression of adenosine triphosphate–induced DRG cell currents,⁴⁹  and with the inhibition of transmitted electrical impulses in C-nerve fibers.⁵⁰  GR agonists were without effect in nonactivated, silent neurons.⁴⁸ 

In contrast to the GR-mediated effect, administration of the MR agonist aldosterone further enhanced the mechanical hypersensitivity of the inflamed hind paw and decreased the mechanical threshold on the contralateral noninflamed side within 15 min (fig. 8). On the other hand, administration of the MR antagonist canrenoate-K rapidly reversed the FCA-induced mechanical hyperalgesia to same extent as dexamethasone. This is in line with the demonstration of the MR selective antagonist eplerenone’s antihyperalgesic effects in an animal model of neuropathic/neurogenic pain induced by local compression or inflammation of the dorsal root.^51,52  Indeed, the MR antagonist eplerenone inhibited and the MR agonist aldosterone enhanced the number of action potentials evoked by local inflammation and suprathreshold stimulation of DRG neurons.⁵¹  Together, our findings lead to the presumption that there might be an endogenous release of a MR agonist that inadvertently drives inflammation and nociception through the activation of the sensory neuronal MR.⁵³  Finally, we determined the antihyperalgesic ED₅₀ values of the GR agonist dexamethasone and the MR antagonist canrenoate-K and investigated the combination of the two using isobolographic analysis.⁵⁴  ED₅₀ of dexamethasone combined with ED₅₀ of canrenoate-K revealed neither a superadditive (synergistic) nor an additive antihyperalgesic effect, according to the generated isobolographs (fig. 9)⁵⁴ ; the combination of the two rather showed a subadditive effect (fig. 9). This might be explained by the presumption that both GR and MR signaling pathways possibly converge further downstream. Supporting this notion is the fact that 75% of MR-ir DRG neurons colocalize with GR.

In summary, these findings provide the first evidence for the beneficial effects of rapid, nongenomic, neuronal mechanisms of GR agonists or MR antagonists in the control of inflammatory pain in counterbalance to a presumable intrinsic neuronal mechanism of MR-mediated nociception, thus unraveling a yet unconsidered approach of pain relief. In addition, our results seem to extend the “pendulum hypothesis” of Hans Selye⁵⁵  that postulates opposing genomic pro- and antiinflammatory actions of MR and GR⁵  to nongenomic pro- and antinociceptive effects of MR and GR, respectively.

The authors thank Claudia Spies, M.D., Ph.D. (Director and Professor, Department of Anesthesiology and Intensive Care Medicine, Charité University Berlin, Berlin, Germany), for her continuous support, Petra von Kwiatkowski and Susanne Runewitz (technicians, Berlin, Germany) for their technical assistance, and Giulia Schäfer for language editing. We are very grateful for the gift of the GR and MR antibodies from M. Kawata, Ph.D. (Kyoto Prefectural University of Medicine, Kyoto, Japan) and Celso E. Gomez-Sanchez (University of Mississippi, Jackson, Mississippi). The incentive for this work was the previous clinical findings of the beneficial role of corticosteroid receptor agonists/antagonists in patients with inflammation and pain by the Prof. KH René Koczorek Foundation, Neuried, Germany.

This work was supported by the Prof. KH René Koczorek Foundation, Neuried, Germany. Mr. Mohamed was supported by a scholarship from the Faculty of Science at Aswan University, Aswan, Egypt, and the Egyptian Government.

The authors declare no competing interests.