Enkephalin-producing regulatory T cells in the skin restrain local inflammation through control of nociception

Abstract

The skin integrates diverse signals discerned by sensory neurons and immune cells to elicit adaptive responses to a range of stresses. Considering interactions between nervous and immune systems, we examined whether regulatory T (T_reg) cells, which suppress systemic and local inflammation, can modulate activation of peripheral neurons. Acute T_reg cell "loss of function" increased neuronal activation to noxious stimuli independently of their immunosuppressive function. This activity was mediated by a T_reg cell subset capable of production of enkephalins encoded by the gene Penk, whose expression is facilitated by combined TCR and glucocorticoid receptor signaling. Punctual selective depletion of Penk-expressing T_reg cells or specific ablation of Penk in T_reg cells increased neuronal activation in response to noxious stimuli and associated inflammation. Our study indicates that a population of tissular T_reg cells exhibits neuromodulatory activity to restrain local inflammation in the skin.

Figure 1.. T_reg cell depletion increases nociception.

A-D, Ear skin from Foxp^creERT2R26^LSLTdTomato mice was analyzed by confocal microscopy. A, Representative image; 100μm sections were stained for Tubβ3 (white), CD31 (blue) and TdTomato (yellow). Scale bar, 10 μm. B, T_reg cell distance to nearest neuron and endothelial cell. Each dot represents a cell. C, Distribution of T_reg cell distances to nearest neuronal and endothelial cells. Data is shown as percentage of T_reg cells in 5 μm bins. D, Spatial relationship between T_reg cell distance to neurons and endothelial cells. Each point represents a single cell, X-axis indicates distance to nearest neuron and Y-axis indicates distance to the nearest endothelial cell. A-D Data compiled from ear skin of 3 mice. E-I, Foxp3^DTR mice were treated with DT or PBS control for 18 hours. E-F, Vehicle or capsaicin was applied to ear skin and DRG were isolated 30 min post-treatment and analyzed by confocal microscopy. E, Representative confocal images of DRG. Scale bars, 10 μm. F, Percentage of cFos⁺ nuclei of neurons (Tubβ3⁺ cells) in DRG. Each dot represents average cFos⁺ nuclei from cervical DRGs per mouse. G, Latency to respond to thermal stimulus. H, Duration of aversive response to cold stimulation. G-H Each dot represents response time per mouse. I, Number of responses to varying forces von Frey filaments out 10 trials in male (n=9 per group) and female mice (n=11 per group). J-K, Foxp3^DTR mice were treated daily with topical IMQ or petroleum jelly (control) on both ears and treated with PBS or DT on 2^nd day of treatment. Mice were analyzed after 3 days of IMQ. J, Percentage of cFos⁺ nuclei of Tubβ3⁺ cells in DRG. K, IL-23 (p40) per g of ear tissue quantified by ELISA. Bars show means, error bars show SEM. **, p<0.01; ***, p<0.001. P-values were calculated using unpaired t-test (G, H), one-way ANOVA (F, J, K) or Two-way ANOVA (I). Tukey’s test and Sidak post-test were performed for one-way and two-way ANOVA, respectively.

Figure 2.. An effector population of T_reg cells expressing Penk is enriched in the skin.

A, Frequency of TdTomato⁺ T_reg cellsfrom Penk^CreR26^LSL-TdTomatoFoxp3^gfp mice. B, RT-qPCR analysis of Penk expression in T_reg cells from spleen, LN and skin. C Frequency of tdTomato⁺ T_reg cells in specified tissues after 3 day IMQ treatment. D-G Ear skin T_reg cells from Penk^CreR26^LSL-YFPFoxp3^Thy1.1 were analyzed by plate-based SMART-Seq2 scRNA-seq analysis. D, 2D force-directed graph layout of T_reg cells colored according to cluster (nearest-neighbor clustering). E, YFP expression per cluster. F, 2D force-directed graph layout of T_reg cells from ear skin. Shading (pink/low to purple/high) indicates cell pseudotime. G, Selected gene expression per cluster. H, T_reg cells from secondary lymphoid organs of Foxp3^DTR mice were stimulated by plate coated anti-CD3/CD28 (activated) or anti-CD28 coated (Control) for 48 hours in the presence of IL-2. Penk expression determined by RT-qPCR relative to Control (2^−ΔΔCT). I, Penk promoter region, defined as 500 bp upstream and 200 bp downstream of transcription start. Penk transcripts (blue), GR motifs (NR3C1) (yellow), motifs associated with TCR signaling (FOS, RELB, JUN and Nfatc2) (magenta) and Stat5 (Stat5a and Stata5b) motifs (grey). J-K, T_reg cells from secondary lymphoid organs of Foxp3^DTR mice were stimulated by plate-coated anti-CD3/CD28 with dexamethasone (0–1 μM) in the presence of IL-2 for 24–48 hours. J, Relative Penk expression determined by RT-qPCR after 24 hours. Data shows relative expression to dexamethasone untreated control (2^−ΔΔCT). k, Met-Enkephalin measured in the supernatant by competitive ELISA after 48 hours. L, Penk transcripts from skin T_reg cells isolated from Nr3c^1fl/flFoxp3^Cre and littermate control mice. Data from (39). M, Frequency of TdTomato⁺ T_reg cells in cervical DRG from Penk^CreR26^LSL-TdTomatoFoxp3^gfp mice after 3 day IMQ treatment. A and M CD45 IV-labeling was performed to exclude circulating cells. A, C, H, J and K, L, M Dots represent individual mice, bars show means, error bars show SEM. * p<0.05, ** p<0.01; *** p<0.001. P-values were calculated using unpaired t-test (K, L, M), one-way ANOVA (B, J) or two-way ANOVA (C). Tukey’s test and Sidak post-test were performed for one-way and two-way ANOVA, respectively.

Figure 3.. Acute ablation of Penk-expressing T_reg cells results in increased nociception and inflammation.

A-F, Tcrb^−/− mice were irradiated and reconstituted with a mix of bone marrow from Tcrb^−/− and R26^iDTRPenk^Cre at 4:1 ratio (referred to as Penk^posDTR). 6 weeks post reconstitution Penk^posDTR mice were treated with IMQ for 3 consecutive days. DT or PBS control was administered on day 1 and 2. Tissues were collected on day 3. A, Schematic for bone marrow chimeric mice and depletion of Penk-expressing T_reg cells. B-C, Number of T_reg cells found in the cervical LN (B) and skin (C) following DT. D, Latency to respond to thermal stimulus 18 hours post DT or PBS control treatment. E, Percentage of cFos+ nuclei of Tubβ3 + cells in DRG in IMQ treated Penk^posDTR mice. F, IL-23 quantified by ELISA in ear skin from IMQ treated Penk^posDTR mice. G-L, Foxp3^fl-DTRPenk^Cre mice (referred to as Penk^negDTR) were treated with IMQ for 3 consecutive days (start: day 0). DT or PBS control was administered on day 1 and 2. Tissues were collected on day 3. G, Schematic of depletion of non-expressing Penk T_reg cells in Penk^negDTR mice.H-I, Number of T_reg cells found in the cervical LN (H) and skin (I) following DT treatment in Penk^negDTR mice. J, Latency to respond to thermal stimulus in male Penk^negDTR mice 18 hours post DT or PBS control treatment. K, Percentage of cFos+ nuclei of Tubβ3 + cells in DRG in IMQ treated Penk^negDTR mice. L, IL-23 quantified by ELISA in ear skin from IMQ treated from Penk^negDTR mice. M, Comparison of IL-23 fold change DT treated and PBS treated (DT/Ctrl) Foxp3^DTR, Penk^posDTR and Penk^negDTR mice following daily IMQ treatment. Each dot represents data from a mouse, bars show mean, error bars show SEM. *, p<0.05; **, p<0.01; ***, p<0.001. P-values were calculated using unpaired t-test.

Figure 4.. T_reg cell Penk expression is required to prevent excessive nociception and inflammation.

A-L, Penk^fl/wtFoxp3^creERT2 (Ctrl) and Penk^fl/flFoxp3^creERT2 (PenkΔ) mice were treated with tamoxifen 7 day prior to all treatments and tests. A, Latency to respond to thermal stimulus in male Ctrl and PenkΔ mice. B, Number of responses to varying forces von Frey filaments out 10 trials in male Ctrl (n=7) and PenkΔ mice (n=10). C, Duration of aversive response to cold stimulation in male Ctrl and PenkΔ mice. D-E, Ctrl and PenkΔ mice were treated daily with IMQ or control (petroleum jelly) for 3 days. D Percentage of cFos⁺ nuclei of DRG neurons (Tubβ3⁺ cells). E, Concentration of IL-23 in ear tissue quantified by ELISA. F-J, Ctrl and PenkΔ mice were treated with IMQ for 7 days. F, Percentages of dermal γδ T cells producing IL-17A 3 hours post restimulation with PMA/ionomycin. G, Number of neutrophils in skin tissue. H, Ear thickness at indicated time points, data shows percent of thickness prior to IMQ treatment (Ctrl n=9, PenkΔ n=7). I, Representative histological sections of IMQ treated ears at day 7 stained by H&E. Scale bars 200 μm. J-L, Ctrl and PenkΔ mice were treated daily with IMQ for 3 days. Cells from cervical DRG were analyzed by sn-RNA-seq on day 4, pooled from 5 mice per group. J, UMAP of subclustered neuronal populations by subtype. Colors indicate clustering and classification based on expression of canonical markers. K, Expression of canonical markers across clusters. L, Heatmap of scaled expression for differentially expressed genes for MrgprD-expressing neuronal cluster. Bold gene names indicate genes associated with pain, itch and inflammation. Each dot represents data from a mouse, bars show mean, error bars show SEM. *, p<0.05; ***, p<0.001. P-values were calculated using unpaired t-test (A, C, F, G), one-way ANOVA (D,E), and two-way ANOVA (B,H). Tukey’s test and Sidak post-test were performed for one-way and two-way ANOVA, respectively.