Hematopoietic pannexin 1 function is critical for neuropathic pain

Abstract

Neuropathic pain symptoms respond poorly to available therapeutics, with most treated patients reporting unrelieved pain and significant impairment in daily life. Here, we show that Pannexin 1 (Panx1) in hematopoietic cells is required for pain-like responses following nerve injury in mice, and a potential therapeutic target. Panx1 knockout mice (Panx1^-/-) were protected from hypersensitivity in two sciatic nerve injury models. Bone marrow transplantation studies show that expression of functional Panx1 in hematopoietic cells is necessary for mechanical hypersensitivity following nerve injury. Reconstitution of irradiated Panx1 knockout mice with hematopoietic Panx1^-/- cells engineered to re-express Panx1 was sufficient to recover hypersensitivity after nerve injury; this rescue required expression of a Panx1 variant that can be activated by G protein-coupled receptors (GPCRs). Finally, chemically distinct Panx1 inhibitors blocked development of nerve injury-induced hypersensitivity and partially relieved this hypersensitivity after it was established. These studies indicate that Panx1 expressed in immune cells is critical for pain-like effects following nerve injury in mice, perhaps via a GPCR-mediated activation mechanism, and suggest that inhibition of Panx1 may be useful in treating neuropathic pain.

Figure 1. Panx1^−/− mice are resistant to SNI-induced mechanical hypersensitivity.

(A) SNI and CCI models. (Image adapted from55) (B) Absence of mechanical hypersensitivity in Panx1^+/+ mice undergoing sham surgery (n.s., p > 0.46 by two-way RM ANOVA). (C) Wild-type littermates underwent unilateral SNI and developed mechanical hypersensitivity in the operated limb. In contrast, Panx1^−/− mice were protected from mechanical hypersensitivity. (D) Summary data (for C) is the mean (±SEM) of the threshold value for each mouse, averaged over the period from POD7-28 (depicted by the shaded region in the time series). (E,F) Like SNI, mechanical hypersensitivity following CCI is completely absent in Panx1^−/− mice while wild-type littermates experience robust hypersensitivity. In this and all other figures showing the time course of mechanical or thermal sensitivity, two-way RM ANOVA was used for statistical analysis, with Tukey’s test for pairwise comparisons: ^#p < 0.05 versus day 0; *p < 0.05 for contralateral versus ipsilateral paw. For all figures showing summary data, the averaged value from each mouse from POD7-28 was treated as a single data point in calculating group means that were compared by two-way RM ANOVA; *p < 0.05 between indicated groups, with Sidak’s test used for pairwise comparisons.

Figure 2. Bone marrow transplantation reveals a critical role for Panx1 in hematopoietic cells following nerve injury.

(A,B) Study design. (C,D) Mechanical hypersensitivity following SNI was restored in Panx1^−/− mice reconstituted with wild-type marrow and diminished in Panx1^+/+ mice reconstituted with Panx1^−/− marrow. (E,F) SNI-induced mechanical hypersensitivity in mice reconstituted with cells of the same genotype. (G) Flow cytometry showing GFP fluorescence in bone marrow from C57BL/6 mice or UBI-GFP mice. (H) GFP fluorescence after bone marrow transplantation of cells from UBI-GFP or mice of the indicated Panx1 genotypes demonstrating successful irradiation and replacement. Statistical analysis of behavioral data as in Fig. 1.

Figure 3. Removal of Panx1 from macrophages, myeloid cells or T cells does not prevent pain.

(A,C,E) Summary data of effects of SNI on mechanical threshold in Panx1 ^fl/fl mice and CX3CR1-Cre littermates (A), LysM-Cre littermates (C), or CD4-Cre littermates (E). (B,D,F) RT-qPCR for Panx1 (relative to HPRT) from peritoneal macrophages (B, n = 3–7, *p < 0.05, unpaired two-tailed t-test), from peritoneal macrophages (D, n = 5, *p < 0.05, unpaired two-tailed t-test) and from acutely isolated T cells (F, n = 4–5, *p < 0.05, one-way ANOVA followed by Tukey’s multiple comparisons test). See Fig. S4 for time course of changes in mechanical hypersensitivity.

Figure 4. Restoration of Panx1 expression in adult Panx1^−/− hematopoietic cells restores pain following nerve injury; rescue requires tyrosine 198 of Panx1.

(A) Study design. (B) RT-qPCR for Panx1 (relative to HPRT) from bone marrow of uninfected C57BL/6 mice, or mice who received virally-infected hematopoietic cells; *p < 0.05 vs. empty vector. (C,D) Mechanical hypersensitivity develops in mice receiving cells infected with Panx1(WT), but fails to develop in empty-vector treated cells. (E,F) Mechanical hypersensitivity develops fully in mice receiving cells infected with Panx1(TEV), but is reduced in Panx1(YA) treated cells. Statistical analysis of behavioral data as in Fig. 1.

Figure 5. Pannexin channel inhibitors, carbenoxolone (CBX) and trovafloxacin (Trovan), prevent or reverse mechanical hypersensitivity when administered systemically.

(A) Diagram depicting two dosing schemes. (B,D) The effect of early, daily IP CBX (B, 30 mg/kg in saline, n = 10–11 per condition) or Trovan (D, 30 mg/kg in DMSO, n = 10 per condition) and the respective vehicles. At day 7, pain development was significantly blunted by drug treatment, relative to vehicle; after the drug was withdrawn, full hypersensitivity was observed at day 28. (*p < 0.05 by two-way RM-ANOVA followed by Sidak’s multiple comparison test). (C,E) Changes in hypersensitivity after daily IP injections of CBX or saline (C, n = 12–13), and after Trovan or DMSO beginning at POD7 (E, n = 12–13). Mechanical sensitivity was partially relieved by both drugs, comparing pre- and post-treatment values (dotted lines: 95% CI for pre-operation baselines; *p < 0.05 by two-way RM-ANOVA followed by Sidak’s multiple comparison test).