Activation of sphingosine-1-phosphate receptor 1 in the spinal cord produces mechano-hypersensitivity through the activation of inflammasome and IL-1β pathway

Abstract

Sphingosine-1-phosphate (S1P) receptor 1 subtype (S1PR1) activation by its ligand S1P in the dorsal horn of the spinal cord (DH-SC) causes mechano-hypersensitivity. The cellular and molecular pathways remain poorly understood. We now report that activation of S1PR1 with intrathecal injection of the highly selective S1PR1 agonist SEW2871 led to the development of mechano-allodynia by activating the nod-like receptor family, pyrin domain containing 3 (NLRP3) inflammasome (increased expression of NLRP3, cleaved caspase 1 and mature interleukin (IL)-1β) in the DH-SC. The functional S1PR1 antagonist FTY720 blocked NLRP3 activation and IL-1β production. Moreover, inhibiting IL-10 signaling with an intrathecal injection of an anti-IL-10 antibody attenuated the beneficial effects exerted by FTY720. This suggests that disrupting S1PR1 signaling engages beneficial IL-10-dependent pathways. Noteworthy, we found that mice with astrocyte-specific deletions of S1pr1 did not develop mechano-allodynia following intrathecal injection of SEW2871 and exhibited reduced levels of cleaved caspase 1; identifying astrocytes as a key cellular locus for S1PR1 activity. Our findings provide novel mechanistic insights on how S1PR1 activation in the spinal cord contributes to the development of nociception while identifying the cellular substrate for these activities.

Introduction

Dysregulation of sphingolipid metabolism in the dorsal horn of the spinal cord (DH-SC) has been linked to the development of nociceptive behaviors arising from chemotherapy,^{19, 43} traumatic nerve injuries,¹⁰ cancer¹⁷ and opioids.²⁸ The bioactive sphingolipid sphingosine-1-phosphate (S1P) is formed by phosphorylation of sphingosine by the two isoforms of sphingosine kinases (SphKs; SphK1 and SphK2),⁴² which are expressed throughout the central nervous system (CNS) including the spinal cord.^5,50 S1P can act both as an intracellular mediator and an extracellular ligand to its five cognate G protein-coupled receptors, S1PR1-5, in an autocrine/paracrine manner to produce so-called “inside-out signaling”.⁴²

Emerging evidence in a variety of pain models implicate S1PR1, a Gα_i/0-coupled receptor³² found in neurons and glial cells,⁵² as a critical signaling mechanism for the effects of S1P.^{11, 17, 19, 36, 40, 43} Accordingly, pharmacological and genetic inhibition of S1PR1 signaling in the spinal cord attenuates a variety of pain states.^{11, 17, 19, 40, 43} Moreover, a single intrathecal (i.th.) injection of the selective S1PR1 agonist SEW2871³⁷ elicits profound thermal hyperalgesia¹⁴ and mechano-hypersensitivity (hyperalgesia and allodynia)¹⁹ in normal rodents.

However, the cellular substrate for S1PR1 activity and the molecular mechanisms engaged by S1PR1 activation in the spinal cord are not fully understood. Activation of S1PR1 in neurons increases their excitability;^{6, 53} yet, S1PR1 activity also promotes the release of neuroinflammatory/neuroexcitatory substances, such as interleukin (IL)-1β, tumor necrosis factor (TNF) and nitroxidative species, from glial cells in the CNS.^{7, 35, 36} We have recently reported that the attenuation of neuropathic pain with S1PR1 antagonists correlated with a reduction in the expression of markers of glial activity and pro-inflammatory cytokines (e.g., IL-1β, TNF) and a shift toward antiinflammatory IL-10 signaling in the spinal cord.^{17, 19, 43} Moreover, conditional knockout of S1PR1 in astrocytes completely prevents the development of chemotherapy-induced neuropathic pain from administration of the proteosome inhibitor, bortezomib.⁴³ Yet, it is not clear from these studies whether S1PR1 engages neuroinflammatory processes in the spinal cord or simply maintains an environment suitable for the development of neuroinflammation initiated by the underlying neuropathic etiology.

To gain better understanding of S1PR1-dependent mechanisms in the development of neuroinflammation in the spinal cord, we investigated whether direct activation of S1PR1 with an i.th. injection of SEW2871 in naive rodents was sufficient to induce IL-1β signaling in the DH-SC to drive the development of mechano-allodynia. IL-1β is an important mediator in nociception³⁴ as it can stimulate the production of additional inflammatory cytokines²⁷ and enhance neuronal excitatory synaptic transmission.⁵¹ IL-1β activity, however, requires both transcriptional and post translational regulation for its production, release and signaling. The canonical processing pathway of IL-1β is through inflammasome-dependent caspase 1 activity (reviewed in ⁴⁵). The nod-like receptor family, pyrin domain containing subtype 3 (NLRP3) is one of eight known nod-like receptors that can form inflammasomes. Once the NLRP3 inflammasome complex has formed, the cysteine protease caspase-1 is autocatalytically cleaved and activated to process the pro-IL-1β protein to bioactive IL-1β for its released and signaling.⁴⁵ NLRP3 activity has been reported with pain^{9, 16, 46} and recently S1PR1 has been shown to initiate production of active IL-1β through the NLRP3 inflammasome activity in tumor associated macrophages.⁴⁸

We now offer the first evidence that S1PR1 signaling in the DH-SC is sufficient to engage NLRP3 and IL-1β signaling to drive the development of mechano-allodynia. We also identify astrocytes as a critical cellular substrate of S1PR1 activity in the development of S1PR1-induced NLRP3 activation and mechano-alllodynia.

Materials and Methods

Experimental Animals

Rats:

Male Sprague Dawley rats (200-220 g starting weight) were purchased from Harlan Laboratories (Indianapolis IN, USA; Frederick, MD breeding colony).

S1pr1 knockout mice:

Transgenic mouse colonies are descendants of original homozygous S1pr1^fl/fl,GFAP-Cre breeder mice⁷ kindly gifted to us by Dr. Jerold Chun (The Scripps Research Institute, La Jolla CA, USA) and bred as previously described.⁴³ All mice were genotyped before breeding or use in experiments by endpoint PCR of ear-punch DNA as previously described⁴³ and ear-tagged for identification. Extensive morphological and functional characterization of astrocyte-specific S1pr1 knockout mice show that S1PR1 is deleted in CNS cells of astrocyte lineage.⁷ Our own work has further shown that this deletion occurs within the spinal cord, but not in the DRG of naive mice.⁴³ Age-matched male S1pr1 knockout mice and controls were used in all experiments.

Animals were housed 2-4 per cage (for rats) or 5 per cage (for mice) in a controlled environment (12 h light/dark cycle) with food and water available ad libitum. All animals were randomly assigned to their groups and experiments were conducted with the experimenters blinded to treatment conditions. Rats were all healthy and had normal baseline behavior and mechano-allodynia values. All experiments were performed in accordance with the guidelines of the International Association for the Study of Pain and the National Institutes of Health and approvals by the Saint Louis University.

Test Compounds

Fingolimod (FTY720), MCC950 and SEW2871 were purchased from Cayman Chemical (Ann Arbor MI, USA). IL-1Ra was from Amgen (Thousand Oaks CA, USA). Recombinant rat IL-1β was purchased from Bio-Techne (Minneapolis MN, USA). The sphingosine kinase inhibitor, SK-I [2-(p-hydroxyanilino)-4-(p-chlorophenyl) thiazole], was purchased from Calbiochem (San Diego CA, USA). Sheep anti-rat IL-10 IgG antibody was a generous gift from Dr. Linda Watkins at the University of Colorado Boulder. Control sheep serum IgG was obtained from Sigma Aldrich (St. Louis MO, USA). The vehicle used for all intrathecal injections (i.th.) of test agents in rats and mice was 2% DMSO in saline. The vehicle used for all oral administration of test agents was 5%DMSO/0.5% methylcellulose in saline (0.2 ml dosing volume).

Intrathecal drug delivery

Rats and mice were lightly anesthetized with isoflurane. A 30-ga needled, 10 μL Hamilton syringe (Hamilton, Reno, NV) is inserted between the L4/L5 vertebrae puncturing the dura (confirmed by presence of reflexive tail flick)¹⁸ and 5 μl of vehicle or test substance(s) was injected.

Behavioral Testing

Rats and mice were acclimated on an elevated mesh table for at least 30 minutes prior to behavioral measurements. Mechano-allodynia was measured at baseline prior to i.th. injections. After recovering from light anesthetic for i.th. injection of test compounds animals were returned to mesh table for the duration of the study. Behaviors were again measured at 1 hour (h) and 2 h post-injection. To determine mechano-allodynia, the plantar aspect of the hind paw was probed with calibrated von Frey filaments (Stoelting, Wood Dale, IL, USA; mice: 0.07–2.00 g; rats: 0.407–26 g) according to the “up-and-down” method¹² and a paw withdrawal threshold (PWT, g) was calculated.

Preparation of whole tissue lysate.

The dorsal horn of lumbar spinal cords were harvested from saline perfused rats within 30 min following the last behavioral test. The tissues were homogenized in 10 volumes of ice-cold homogenization buffer [50 mM Tris-Cl, pH 8.0, 150 mM NaCl, 0.5% Triton X-100, 0.1% SDS, 1 mM EDTA, 5% glycerol, 1 mM PMSF, 1× protease inhibitor cocktail (Sigma-Aldrich, St. Louis MO, USA; 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, 15 μM pepstatin A, 14 μM E-64, 40 μM bestatin, 20 μM leupeptin, 850 nM aprotinin)]. The homogenates were incubated for 10 min on ice before being pulse sonicated (Sonic Dismembranator 60; Thermo Fisher Scientific, Carlsbad CA, USA). The samples were clarified by centrifugation at 13,000 x g, 15 min, and 4°C. The total protein concentration in the clarified lysates was measured using bicinchoninic acid assay (Thermo Fisher Scientific, Carlsbad CA, USA).

Western blot.

Lysate proteins (40 or 100 μg) were resolved by sodium dodecyl sulphate-polyacrylamide gel electrophoresis and transferred to nitrocellulose. Membranes were blocked for 1 h at room temperature in 5% non-fat dried milk and subsequently probed for target proteins. Knockdown/knockout validated antibodies were used to probe for phospho-p38 (1:1000; Cell Signaling Technologies #9211)² and total p38 (1:1000, Cell Signaling Technologies #9212).³⁸ The antibody used to probe for NLRP3 (1:1000; Novus Biologicals #NBP2-12446, Littleton CO, USA) has been shown to detect increased NLRP3 in LPS-stimulated RAW 264.7 cells with little background signal in spleen tissues from NLRP3 knockout mice.²² However, in our samples, NLRP3 band was detected at 75 kDa as reported in previous publications.^{16, 23} The caspase 1 (p20) (1:500; Santa Cruz Biotechnology #sc-398715, Santa Cruz CA, USA) has been shown to detect increased caspase 1 (p20) in brain, heart and lung tissue of LPS-treated mice.¹³ The IL-1β (1:1000, Cell Signaling #12242, Danvers MA, USA) antibody used detects the same p17 IL-1β protein in tissues as cells with flag-tagged IL-1β.⁴¹ Primary antibodies were diluted in 1X TBS, 5% nonfat dried milk and blots incubated at 4°C overnight. The membranes were washed in 1X TBS-T (15 mM Tris pH 7.6, 150 mM NaCl, 1% Tween-20). The bound antibodies were then visualized following incubation with peroxidase-conjugated bovine anti-mouse IgG secondary antibody (1:1000, Jackson ImmunoResearch #04-18-15, Gaithersburg PA, USA) or peroxidase-conjugated goat anti-rabbit IgG (1:1000-5000 Cell Signaling #7074, Danvers MA, USA) for 1 h at room temperature. Peroxidase-conjugated antibodies were visualized by enhanced chemiluminescence (Bio-Rad, Hercules CA) and documented using Chemidoc XRS+ documentation system and ImageLab^™ software (BioRad, Hercules CA). Blot images were captured using image accumulation mode (100 images; 30-300 s). Then the blots were treated twice for 15 min with 30% hydrogen peroxide to deactivate the HRPO ³⁹ and probed for β-actin (1:5000, Sigma-Aldrich, St. Louis MO, USA) for use as endogenous loading controls.

Relative protein expression was quantified by measuring band densitometry. Images for the blots of a protein of interest and their corresponding β-actin were selected for analyses and presentation based on a predetermined upper grayscale value (35000-40000 units) in order to assure linear densitometric values. Images were analyzed using the lane and band functions of the ImageLab^™ software. For presentation purposes, the grayscale range was set between 0 units and the predetermined upper grayscale value (35000-40000 units) for the protein of interest and corresponding β-actin prior to exporting as an image file. Post-export modifications of the images were limited to cropping to the regions of interest.

S1P ELISA.

S1P levels were measured in spinal cord lysates using commercially available ELISA kits (Echelon Biosciences, Salt Lake City, UT) as previously described.^{19, 28}

Statistical Analysis.

Data are expressed as mean ± SD for n animals. Data were analyzed by two-tailed unpaired Student’s t-test; two-tailed, one-way ANOVA with Bonferroni post hoc comparison or two-tailed, two-way repeated measures ANOVA with Dunnett’s post hoc comparisons. The effect size of treatment or time-dependent treatment in each experiment was assessed using Cohen’s d, eta-squared (η²) or partial eta-squared (η_P²). Significant differences were defined at P<0.05. All statistical analyses were performed using GraphPad Prism (v5.04, GraphPad Software, Inc.).

Results.

Intrathecal SEW2871 induces neuroinflammation and activation of NLRP3 and IL-1β signaling in the DH-SC to drive mechano-allodynia.

We have recently reported that a single i.th injection of SEW2871 causes profound mechano-hypersensitivity.¹⁹ We now extend these findings and show that an i.th. injection of SEW2871 (2 nmol), which caused significant mechano-allodynia (Fig. 1A), was associated with a significant increase in neuroinflammation in the DH-SC. When compared to vehicle-treated rats without mechano-allodynia, the levels of phosphorylated nuclear factor kappa B (NFκB; Fig 1B), phosphorylated p38 (Fig 1C) and mature IL1β (p17; Fig 1D) were significantly increased in the DH-SC of rats 2 h post-SEW2871. Moreover, IL-1β signaling induced by SEW2871 had a functional role in the development of mechano-allodynia. Inhibition of IL-1β signaling with i.th. co-administration of the IL-1 receptor antagonist (IL-1Ra)²¹ prevented SEW2871-induced mechano-allodynia (Fig 1A).

Figure 1. Intrathecal SEW2871 induces neuroinflammation and activation of NLRP3 and IL-1β signaling in the DH-SC to drive mechano-allodynia.

(A) When compared to vehicle (Veh; n=4), SEW2871 (2 nmol; n=6) induced mechano-allodynia in rats, which was blocked by intrathecal IL-1Ra (100 μg; n=5); [F(4,24) = 28.43, p<0.0001, η_p²=0.83]. (B,C) DH-SC harvested from rats 2 h post i.th. SEW2871 (2 nmol) had significantly increased levels of phosphorylated NFκB p65 (B) and p38 (C) when compared to vehicle [B: t(4)=6.0, p=0.004; d=4.92, n=3/group; C: t(10)=2.4, p=0.039; d=1.37, n=6/group]. (D-F) When compared to DH-SC harvested from rats 2 h after i.th. vehicle (D: n=5, E: n=8, F: n=8), the levels of mature IL-1β (p17; D: n=4), NLRP3 (E: n=8) and cleaved caspase 1 (p20; F: n=8) in DH-SC harvested from SEW2871-treated rats were significantly increased. Oral administration of FTY720 (0.1 mg/kg) attenuated SEW2871-induced expression of mature IL-1β (D: n=6), NLRP3 (E: n=8) and cleaved caspase 1 (p20; F: n=8) in the DH-SC. [D: F(2,12)= 10.4, p=0.0024, η²=0.63; E: F(2,21)=6.68, p=0.0057, η²=0.39; F: F(2,21)=4.79, p=0.019, η²=0.31]. SEW2871, FTY720 or vehicle had no effect on β-actin levels (E: [β-actin densitometric units for Veh: 8.21×10⁶±1.53×10⁶ versus SEW2871: 8.32×10⁶±2.7×10⁶ versus SEW2871+FTY720: 9.33×10⁶±2.24×10⁶ [F(2,21)=0.606, p=0.555, η²=0.055]). (G,H) When compared to rats treated with SEW2871 (2 nmol), concurrent i.th. administration of the NLRP3 inhibitor MCC950 (5 μM) attenuated the development of mechano-allodynia (G: [F(4,18)=18.6, p< 0.0001, ηp²=0.81; n=4/group]) and the production of cleaved caspase 1 in the DH-SC (H: [t(6)=5.77, p=0.0012, d=4.08, n=4/group]). Vehicle treatment (n=4; G) had no effect on behavior. SEW2871 or MCC950 had no effects on β-actin levels [H: β-actin densitometric units for SEW2871+Veh: 3.07×10⁶±9.06×10⁵ versus SEW2871+MCC950: 3.71×10⁶±8.29×10⁵; t(6)=1.04, p=0.337, d=0.74]). Mean±SD for (n) and analyzed by two-way repeated measures (RM)-ANOVA with Bonferroni post-hoc comparisons (A,G), Student’s unpaired t-test (B,C,H) or one-way ANOVA with Dunnett’s post-hoc comparisons (D-F). *P<0.05 vs. Oh; #P<0.05 vs. Veh or Veh + Veh and †P<0.05 vs. time-matched SEW2871+Veh.

We investigated whether NLRP3 inflammasome-dependent post-translational processing of IL-1β was activated in the dorsal spinal cord following SEW2871. When compared to vehicle-treated rats, i.th. SEW2871 increased the expression of NLRP3 (Fig 1E) and the levels of cleaved caspase 1 (Fig 1F) in the DH-SC. Our previous work showed that the administration of competitive (NIBR-14) or functional (FTY720 and CYM5442) S1PR1 antagonists attenuated i.th. SEW2871-induced mechano-hypersensitivity in rats.¹⁹ To determine whether the beneficial effects of S1PR1 antagonism on SEW2871-induced mechano-allodynia was associated with the attenuation of NLRP3 post-translational processing of IL-1β, we treated rats with a single oral dose of FTY720 30 minutes prior to i.th. SEW2871 treatment. FTY720 attenuated the levels of SEW2871-induced NLRP3 (Fig 1E), cleaved caspase 1 (Fig 1F) and mature IL-1β (p17; Fig 1D) expression in the DH-SC. The functional significance of such SEW2871-induced NLRP3 activity in the spinal cord was evident in rats given a combined i.th. injection SEW2871 and the small molecule NLRP3 inflammasome inhibitor MCC950.⁸ MCC950 (5 μM)⁴⁶ prevented the development of SEW2871-induced mechano-allodynia (Fig 1G) and reduced levels of activated caspase 1 in the DH-SC (Fig 1H).

IL-1β induces mechano-allodynia through activation of sphingosine kinases and S1PR1 signaling.

IL-1β signaling can promote sphingosine kinase activity by stimulating de novo production or increasing phosphorylation of sphingosine kinases.⁴ To determine whether IL-1β signaling may contribute to sustained S1P production and signaling at S1PR1, we treated rats with an i.th. injection of IL-1β (100 ng) at a dose previously reported to induced mechano-hyperalgesia and mechano-allodynia within one hour.⁴⁴ Intrathecal IL-1β caused an increase in S1P levels in the spinal cord (Fig 2A; measured as described previously^{19, 28}) and induced mechano-allodynia within 1 h (Fig 2B). IL-1β-induced mechano-allodynia was attenuated by i.th. co-administration of the SphK inhibitor, SK-I (1.2 ng)²⁸ or by oral delivery of FTY720 given 30 minutes before IL-1β.

Figure 2. Intrathecal administration of recombinant rat IL-1β engages S1P metabolism and S1PR1 signaling to induce mechano-allodynia.

When compared to its vehicle, i.th. administration of recombinant rat IL-1β (100 ng) caused a significant increase in S1P levels in the spinal cord (A: [t(4)=5.00, p=0.0075, d= 3.16, n=5/group]) and induced mechano-allodynia (B: F(3,16)=31.4, p<.0001, η²=0.85 n=5/group]) in rats by 1 h. The development of mechano-allodynia was attenuated with i.th. administration of sphingosine kinase inhibitor (SK-I; 1.2 ng) or oral administration of FTY720 (1 mg/kg). Mean±SD and analyzed by unpaired Student’s t-test (A) or one-way ANOVA with Dunnett’s post-hoc comparisons (B). *P<0.05 vs. Veh, #P<0.05 vs Baseline and †P<0.05 vs. IL-1β + Veh.

Disrupting S1PR1-mediated NLRP3 and IL-1β-driven neuroinflammation engages anti-inflammatory IL-10 signaling.

IL-10 is one of the most potent anti-inflammatory and neuroprotective cytokines.²⁶ In rats treated with a neutralizing IL-10 antibody (anti-IL-10; 0.2 μg/day), the ability of i.th. IL-1Ra (Fig 3A) or MCC950 (Fig 3B) to attenuate SEW2871-induced mechano-allodynia was lost. These results suggest that the functional consequence of disrupting S1PR1 signaling with S1PR1 antagonists is the engagement of beneficial IL-10-dependent pathways.

Figure 3. Attenuation of mechano-allodynia by disrupting S1PR1/NLRP3/IL-1β signaling is dependent on IL-10 signaling.

When compared to its vehicle (A: n=4; B: n=3), intrathecal administration of SEW2871 (2 nmol; A: n=7; B: n=6) induced mechano-allodynia in rats. The development of mechano-allodynia was attenuated with i.th. administration of intrathecal IL-1Ra (100 μg; A; n=7) or MCC950 (5 μM; B; n=5) in the presence of normal IgG. However, the ability of both IL-1Ra and MCC950 to attenuate mechano-allodynia was lost when rats were treated with i.th. anti-IL-10 (0.2 μg; A: n=4; B: n=5). Mean±SD for (n) rats and analyzed by two-way RM-ANOVA with Bonferroni post-hoc comparisons [A: F(6,36)=12.45, p< 0.0001, η_p²=0.67; B: F(6,30)=15.0, p< 0.0001, η_p²=0.75]. *P<0.05 vs. 0 h and †P<0.05 vs. SEW2871+Veh.

SEW2871-induced mechano-allodynia is dependent on S1PR1 in astrocytes.

Reported RNA-Seq studies of mouse cerebral cortex tissues suggest that the expression of S1PR1 is greater in astrocytes than in neurons and microglia.⁵² We have also recently reported that bortezomib-induced neuropathic pain does not develop in astrocyte-specific knockout mice.⁴³ Thus, astrocytes are a likely target for the action of i.th. SEW2871 in the development of hypersensitivities. However, drugs administered by i.th. routes also can gain access to the dorsal root ganglia (DRG)¹ and activation of S1PR1 on peripheral sensory neurons has been shown increase their sensitivity.⁵³

To determine whether astrocytes are the cellular substrate of i.th. SEW2871 activity, we used astrocyte-specific S1pr1 knockout mice (S1pr1^fl/fl;Gfap-cre).⁷ In control mice (S1pr1^fl/+) with intact S1PR1 expression, i.th. SEW2871 (2 nmol) induced mechano-allodynia within 2 h post injection (Fig 4A) and expression of cleaved caspase 1 p20 was detected (Fig 4B). In contrast, astrocyte-specific S1pr1 knockout mice did not develop mechano-allodynia and had significantly reduced levels of cleaved caspase 1 p20 in the DH-SC following i.th. administration of SEW2871; indicating that astrocyte-specific S1PR1 signaling is necessary for SEW2871-induced NLRP3 activity and mechano-allodynia.

Figure 4. SEW2871-induced mechano-allodynia and caspase 1 activation is dependent on S1PR1 signaling in astrocytes.

(A) Male mice with astrocyte-specific deletion of S1pr1 (S1pr1^fl/fl;Gfap-cre; n=8) of S1pr1 and their controls (S1pr1^fl/+; n=9) were administered intrathecal SEW2871 (2 nmol) and mechano-allodynia was measured 2 h post-SEW2871. Mechano-allodynia developed in control mice, but was attenuated in mice with astrocyte-specific deletion of S1pr1 [F(1,15)=35.9, p< 0.0001, η_p²=0.71]. (B) When measured at 2 h post-SEW2871, the levels of cleaved caspase 1 (p20) in the saline-perfused DH-SC harvested from astrocyte-specific S1pr1 knockout mice (n=8) were significantly lower than in control mice (n=8; [t(14)=2.98, p=0.0099, d=1.49]). Astrocyte-specific knockout of S1pr1 had no effect on β-actin levels [β-actin densitometric units 2.16×10⁷±1.08×10⁶ (S1pr1^fl/+) vs. 2.11×10⁷±1.47×10⁶ (S1pr1^fl/fl;Gfap-cre); t(14)=0.257, p=0.801, d=0.13]. Mean±SD for (n) and analyzed by two-tailed, two-way RM-ANOVA with Bonferroni post-hoc comparisons (A) or unpaired Student’s t-test (B). *P<0.05 vs. 0 h and †P<0.05 vs. S1pr1^fl/+ mice.

Discussion

Our findings provide the first direct evidence that activation of S1PR1 in the DH-SC is sufficient to initiate NLRP3/IL-1β neuroinflammation and drive the development of mechano-allodynia. Moreover, our results are the first to link IL-1β signaling to increased sphingosine kinase activity in the spinal cord, which contributed to the development of S1PR1-dependent mechano-allodynia. Collectively, these findings suggest the establishment of a “feed-forward” cycle in the DH-SC where S1PR1 activity drives NLRP3/IL-1β neuroinflammation that, in turn, enhances sphingosine kinase activity to increase S1P production and sustain S1PR1 signaling (Fig 5). Disrupting this cycle results in shifting the spinal cord towards antiinflammatory state anchored by IL-10 signaling and attenuation of mechano-allodynia.

Figure 5. Summary of proposed mechanistic pathway of S1PR1-induced mechano-hypersensitivity.

Activation of S1PR1 signaling stimulates IL-1β expression and induces NLRP3 inflammasome-dependent post-translational processing of IL-1β. In turn, mature IL-1β promotes further proinflammatory cytokine production and activation of endogenous production of S1P through activation of SphK1. The release of S1P forms a “feed-forward” loop of additional enhancement of proinflammatory cytokine and S1P formation that ultimately leads to the development of mechano-hypersensitivity. Strategies that disrupt this loop and shift the spinal cord toward an anti-inflammatory/neuroprotective state.

Our findings also reveal that a SEW2871-induced NLRP3/IL-1β neuroinflammatory pathway is dependent on S1PR1 in astrocytes. The involvement of astrocytes is critical for the development of hypersensitivities as they regulate excitatory glutamatergic signaling through glutamate reuptake and release and contribute to neuroinflammation.^{20, 47} Under stimuli that induce neuroinflammation, cultured astrocytes are capable of activating NLRP3 and expressing IL-1β.¹⁵ Moreover, direct stimulation of human fetal astrocytes with S1P can increase the activation of NFκB and production of nitroxidative species; two prominent signals in the priming and activation of NLRP3.⁴⁹ Therefore, activating S1PR1 in astrocytes with SEW2871 can be sufficient to trigger NLRP3/IL-1β pathway in astrocytes and drive the development of mechano-allodynia. This is supported by our evidence that caspase 1 activation and SEW2871-induced mechano-allodynia are lost when S1PR1 is deleted from astrocytes. However, in vivo models of pain to date have only identified NLRP3 expression within microglia,¹⁶ indicating a much more complex glia-glia network. Microglia also express S1PR1, although at a lower level than astrocytes,⁵² and there is evidence that exogenous S1P induces IL-1β expression in microglial culture.²⁹ Thus, SEW2871 in our model may prime the NLRP3/IL-1β pathway in microglia, while S1PR1 signaling in astrocytes provides an inflammasome activation signal to trigger microglia IL-1β release. For example, evidence from early studies of the role of S1P in neuropathic pain demonstrated that exogenous S1P or cell-impermeable dihydro-S1P increased Ca²⁺ signaling.¹⁰ Provoked calcium oscillations in astrocytes are propagated between astrocytes to activate microglia at distances from the initial stimuli (reviewed in ²⁵). Propagation of such calcium waves is mediated by purinergic receptors in response to astrocyte release of ATP, which also activate P2X7 purinergic receptors on microglia;²⁵ a well-established trigger of NLRP3 inflammasome oligomerization and its activation.⁴⁵

In addition to activating and driving NLRP3/IL-1β neuroinflammation, S1PR1 signaling in astrocytes may also suppress IL-10 signaling in astrocytes. In models of neuropathic pain, we found that IL-10 levels in the spinal cord significantly increase following inhibition of S1PR1 signaling.^{17, 19, 43} While astrocytes and microglia can both produce IL-10, astrocytes appear to be the predominant cell substrate for IL-10 signaling during neuroinflammation.³⁰ IL-10 signaling in astrocytes has been shown to trigger the release of antiinflammatory mediators, such as tissue growth factor β (TGFβ), that in turn control microglial activity and reduce neuroinflammation.³⁰ S1PR1 signaling in astrocytes sits at the nexus of proinflammatory and antiinflammatory signaling in the CNS. For example, inhibiting S1PR1 signaling in activated astrocytes has been shown to reduce proinflammatory mediator release, while enhancing antiinflammatory mediator release.³⁵ Media obtained from astrocytes where S1PR1 signaling was inhibited was found to block proinflammatory cell activity.³⁵ In contrast, inhibition of S1PR1 signaling had no effect on microglia as they continued to maintain an inflammatory phenotype.³⁵ Under neuroinflammatory conditions, astrocytes also become desensitized to IL-10 and fail to release TGFβ, prolonging microglial activity and neuroinflammation.³¹ IL-10 signaling involves the JAK/STAT pathway to confer its antiinflammatory signaling,²⁶ which can induce SOCS3 expression to attenuate inflammation.³³ S1PR1 signaling has been shown to also activate STAT3, but reduce its inhibitor SOCS3.²⁴ In animal models of neuroinflammation, increased JAK/STAT activation corresponded to increased astrocyte reactivity and neuroinflammation that was attenuated by increasing SOCS3.³ Therefore, in addition to regulating IL-10 expression, S1PR1 may suppress IL-10 signaling in astrocytes by co-opting the STAT3 pathway and guiding astrocytes towards a proinflammatory phenotype.

The significance of a S1PR1-driven NLRP3/IL-1β neuroinflammatory pathway in the DH-SC is its potential for providing context to the antiinflammatory effects of S1PR1 antagonists in the spinal cord during the attenuation of neuropathic pain. Understanding the underlying mechanisms and outcomes of S1PR1 activity in the CNS during the development of pain provides crucial information for developing strategies that target S1PR1 for much needed therapeutic approaches to pain management. Our work here provides the first mechanistic links between S1PR1 activation and neuroinflammatory signaling in the DH-SC that drive the development of mechano-allodynia.

Perspective.

This is the first study to link the activation of NLRP3 and IL-1β signaling in the spinal cord and S1PR1 signaling in astrocytes to the development of S1PR1-evoked mechano-allodynia. These findings provide critical basic science insights to support the development of therapies targeted toward S1PR1.

Highlights.

• S1PR1 mechanisms in the spinal cord producing mechano-hypersensitivity are unknown.

• S1PR1 activation causes mechano-hypersensitivity via NLRP3 neuroinflammation.

• Inhibition of S1PR1 mediated mechano-hypersensitivity is dependent on IL-10.

• Astrocytes are a cellular substrate for S1PR1 activity.