Sphingosine-1-phosphate receptor 1 activation in astrocytes contributes to neuropathic pain

Abstract

Neuropathic pain afflicts millions of individuals and represents a major health problem for which there is limited effective and safe therapy. Emerging literature links altered sphingolipid metabolism to nociceptive processing. However, the neuropharmacology of sphingolipid signaling in the central nervous system in the context of chronic pain remains largely unexplored and controversial. We now provide evidence that sphingosine-1-phosphate (S1P) generated in the dorsal horn of the spinal cord in response to nerve injury drives neuropathic pain by selectively activating the S1P receptor subtype 1 (S1PR1) in astrocytes. Accordingly, genetic and pharmacological inhibition of S1PR1 with multiple antagonists in distinct chemical classes, but not agonists, attenuated and even reversed neuropathic pain in rodents of both sexes and in two models of traumatic nerve injury. These S1PR1 antagonists retained their ability to inhibit neuropathic pain during sustained drug administration, and their effects were independent of endogenous opioid circuits. Moreover, mice with astrocyte-specific knockout of S1pr1 did not develop neuropathic pain following nerve injury, thereby identifying astrocytes as the primary cellular substrate of S1PR1 activity. On a molecular level, the beneficial reductions in neuropathic pain resulting from S1PR1 inhibition were driven by interleukin 10 (IL-10), a potent neuroprotective and anti-inflammatory cytokine. Collectively, our results provide fundamental neurobiological insights that identify the cellular and molecular mechanisms engaged by the S1PR1 axis in neuropathic pain and establish S1PR1 as a target for therapeutic intervention with S1PR1 antagonists as a class of nonnarcotic analgesics.

Keywords: S1P receptor subtype 1; astrocytes; interleukin 10; sphingosine-1-phosphate; traumatic nerve injury-induced neuropathic pain.

Fig. 1.

S1PR1 antagonism, but not agonism, attenuated and reversed neuropathic pain states. (A) In mice with CCI, administration of TASP0277308 (3 mg/kg, ▲; n = 5) or FTY720 (1 mg/kg, ▼; n = 5), but not SEW2871 (20 mg/kg, ■; n = 5) or their vehicle (●; n = 5), attenuated the development of mechano-allodynia. (B–G) In mice with established neuropathic pain, mechano-allodynia was reversed by an acute oral administration (p.o.) on day 10 post CCI of selective S1PR1 antagonists (B) TASP0277308 (0.3, ●; 0.6, ■; 1, ▲ and 3 mg/kg, ▼; n = 6/group) or (C) NIBR-15 (0.3, ●; 0.6, ■; 1, ▲ and 3 mg/kg, ▼; n = 6/group) or S1PR1 functional antagonists (D) FTY720 (0.3, ●; 0.6, ■; 1, ▲ and 3 mg/kg, ▼; n = 6/group), (E) siponimod (0.5, ●; 1.5, ■; 5, ▲ and 15 mg/kg, ▼; n = 5/group), (F) ponesimod (0.3, ●; 0.6, ■; 1, ▲ and 3 mg/kg, ▼; n = 6/group) or (G) KRP-203 (0.4, ●; 1.3, ■ and 4 mg/kg, ▲; n = 4/group). (H) In mice, an acute oral administration of NIBR-15 (3 mg/kg) on day 10 post CCI reversed mechano-allodynia in the ipsilateral paw. This reversal was not prevented (P = 0.79) by naloxone (8 mg/kg; i.p.; n = 5) given 30 min before NIBR-15 (n = 5). (I) CCI-induced mechano-allodynia (○; n = 5) was not attenuated by SEW2871 (20 mg/kg i.p., ●; n = 5). Data are expressed as mean ± SEM for (n) male mice and analyzed by (A–G and I) two-tailed, two-way ANOVA with Bonferroni comparisons or (H) two-tailed, Welch’s corrected Student’s t test. *P < 0.05 vs. d 0, ^#P < 0.05 vs. Vehicle and ^†P < 0.05 vs. d 10.

Fig. 2.

S1PR1 in the spinal cord mediates the development and maintenance of neuropathic pain. (A) The levels of S1P in the spinal cord are increased by D7 in the DH-SC of rats with CCI (n = 5/group). (B) On D14 and D15 post CCI (peak pain), rats were treated with an intrathecal (i.th.) dose of nontargeting control siRNA (siNT; 2 μg; n = 4) or S1PR1-targeting siRNA (siS1PR1; 2 μg; n = 5). When measured on D16, S1PR1 expression was increased in the DH-SC ipsilateral to CCI compared with the contralateral side of rats treated with siNT, but not those treated with siS1PR1. (C) In male rats, intrathecal (i.th.) administration of S1PR1-targeting siRNA (siS1PR1; 2 μg) on days 7 and 8 post CCI attenuated neuropathic pain (n = 5) compared with rats receiving nontargeting control siRNA (siNT; n = 5). (D) Compared with vehicle (n = 8), daily i.th. administration of FTY720 (3 nmol, n = 6) or TASP0277308 (10 nmol, n = 6), but not SEW2871 (2 nmol/d, n = 5), prevented the development of CCI-induced mechano-allodynia in male rats. (E) Acute i.th. administration of NIBR-15 (3 nmol, n = 6) or FTY720 (3 nmol, n = 6), but not vehicle (n = 6), on day 7 reversed CCI-induced mechano-allodynia in rats. (F) Acute intrathecal administration of SEW2871 (2 nmol; n = 5) or its vehicle (n = 5) on D7 post CCI did not reverse mechano-allodynia in male mice. Data are expressed as mean ± SEM for (n) male rats and analyzed by two-tailed, Welch’s corrected Student’s t test (A and B) or two-tailed, two-way ANOVA with Bonferroni comparisons (C–F). *P < 0.05 vs. d 0; ^#P < 0.05 vs. siNT and ^†P < 0.05 vs. d 7.

Fig. 3.

The beneficial effects of S1PR1 antagonist are dependent on IL-10 signaling. (A) At peak mechano-allodynia (day 7), IL-1β levels increased in the spinal cord of rats and was attenuated 1 h after oral administration (p.o.) of NIBR-15 (3 mg/kg; n = 5/group). (B) IL-10 expression the spinal cord of rats did not significantly change from baseline levels at peak mechano-allodynia (day 7), but increased 1 h after administration of NIBR-15, TASP0277308 (TASP), FTY720, or ponesimod (3 mg/kg; p.o.; n = 6/group). (C) s.c. infusion of NIBR-15 (0.3 mg/kg/d, 7 d; n = 4) or FTY720 (0.3 mg/kg/d, 7 d; n = 5) in rats reversed mechano-allodynia. Intrathecal (i.th.) delivery of an anti-IL-10 antibody (αIL-10; 0.2 µg in 5 µl) on days 8, 13, and 15 attenuated the anti-allodynic effects of NIBR-15 (n = 5) and FTY720 (n = 5); nonspecific IgG (n = 5) had no effect on mechano-allodynia. (D and E) CCI produced mechano-allodynia in ipsilateral hind paws of IL-10KO (n = 5) mice and wild-type cohorts (WT; n = 5) by day 7. (D) Acute administration of TASP0277308 (TASP; 3 mg/kg, i.p; n = 5) or FTY720 (3 mg/kg, i.p; n = 5) on day 7 rapidly reversed CCI-induced hypersensitivity in WT mice but was ineffective in IL-10KO mice. (E) In contrast, morphine (1 mg/kg; s.c.; n = 5) analgesic effects were similar in both mouse genotypes. (F and G) S1PR1 antagonist attenuates S1PR1-induced mechano-allodynia in an IL-10-dependent manner: (F) Mice were given a combined i.th. injection of SEW2871 (2 nmol) with vehicle (Veh) or TASP0277308 (TASP; 30 nmol) and αIL-10 (0.2 µg) or IgG (n = 5/group). (G) Mice were given an i.th. injection of SEW2871 (2 nmol) and at peak mechano-allodynia (1 h) were given a second i.th. injection of vehicle (Veh) and IgG or TASP0277308 (TASP; 30 nmol) with αIL-10 or IgG (n = 5/group). Data are expressed as mean ± SEM for (n) male rats or mice and analyzed by one-way ANOVA with Dunnett’s comparisons (A and B) or two-tailed, two-way ANOVA with Bonferroni comparisons (C–G). *P < 0.05 vs. d 0, ^†P < 0.05 vs. d 7, ^#P < 0.05 vs. daily baseline (BL), ^§P < 0.05 vs. 1 h.

Fig. 4.

Astrocytes are a primary cellular target of S1PR1 activity in the spinal cord. When measured on day 7, male (A; n = 13) and female (B; n = 13) astrocyte-specific S1pr1 knockout mice (S1pr1^fl/fl;Gfap-cre) did not develop the CCI-induced mechano-allodynia observed in ipsilateral paws of their sex-matched controls (S1pr1^fl/fl; A, n = 9; B, n = 14). (C) Female control mice (n = 4), but not astrocyte-specific S1pr1 knockout (n = 3), mice developed mechano-allodynia by day 7. Intrathecal administration of anti-IL-10 antibody (αIL-10) on day 7 precipitated mechano-allodynia within 1 h in astrocyte-specific S1pr1 knockout mice. The administration of αIL-10 had no effect on mechano-allodynia in control mice. (D and E) Working hypothesis: Traumatic nerve injury induces increased S1P levels in the spinal cord from various sources (D). Activation of S1PR1 signaling on astrocytes activates NLRP3 inflammasome leading to release of IL-1β, which can then dampen IL-10 expression. Attenuation of IL-1β production following S1PR1 antagonism may remove a brake on IL-10 production and engage IL-10-dependent signaling leading to an attenuation of the neuroinflammatory response and reduced neuronal excitability (E). Data are expressed as mean ± SEM for (n) mice and analyzed by two-tailed, two-way ANOVA with Bonferroni comparisons. *P < 0.05 vs. d 0 and ^†P < 0.05 vs. SEW2871+Veh+IgG.