Multimodal control of dendritic cell functions by nociceptors

Abstract

It is known that interactions between nociceptors and dendritic cells (DCs) can modulate immune responses in barrier tissues. However, our understanding of the underlying communication frameworks remains rudimentary. Here, we show that nociceptors control DCs in three molecularly distinct ways. First, nociceptors release the calcitonin gene-related peptide that imparts a distinct transcriptional profile on steady-state DCs characterized by expression of pro-interleukin-1β and other genes implicated in DC sentinel functions. Second, nociceptor activation induces contact-dependent calcium fluxes and membrane depolarization in DCs and enhances their production of proinflammatory cytokines when stimulated. Finally, nociceptor-derived chemokine CCL2 contributes to the orchestration of DC-dependent local inflammation and the induction of adaptive responses against skin-acquired antigens. Thus, the combined actions of nociceptor-derived chemokines, neuropeptides, and electrical activity fine-tune DC responses in barrier tissues.

Fig. 1.. Nociceptors enhance proinflammatory cytokine production by IMQ-activated DCs in vitro.

BMDCs and nociceptors were either cultured separately or together and treated overnight with proinflammatory stimuli. Indicated cytokines were measured by ELISA in supernatants. (A-C) One representative dose response to IMQ (left panels) and summary of (A) 42 (n=104), (B) 37 (n=74) and (C) 7 (n=16) experiments (right panels) for 1 μg/ml IMQ is shown. (D) BM-macrophages were cultured alone or in the presence of nociceptors and treated with 1 μg/ml of IMQ overnight. Summary of two experiments (n=4) is shown. (E) BMDCs were cocultured with nociceptors, cortical, or hippocampal neurons and treated with 1 μg/ml of IMQ overnight. Summary of three experiments (n=6-10) is shown. (F) BMDCs were cocultured with live or fixed nociceptors and treated with 1 μg/ml of IMQ. Summary of three experiments (n=6) is shown. (G) DCs were cocultured with nociceptors and treated with lidocaine + QX314 and 1μg/ml IMQ. Summary of four experiments (n=8) is shown. (H) Intracellular content of IL-12 p40 in BMDCs left untreated (full histograms) or treated with IMQ (open histograms) in isolation or in a coculture with nociceptors was assessed using flow cytometry. One representative experiment (left) and quantitation of eight independent experiments (right; n=16) is shown. (I) WT or Tlr7-KO DCs were cocultured with WT or Tlr7-KO nociceptors and treated with 1 ug/ml IMQ. Summary of two (n=4) experiments is shown. Across all panels, data represent mean ± s.d. Unpaired t test (A, B, C, and D), one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test (F) or two-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test (E, G, H, and I) were used for statistical analysis: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Fig. 2.. Nociceptors enhance DC cytokine response to various stimuli in a contact-dependent manner.

BMDCs and nociceptors were either cultured separately or together and treated overnight with indicated (A) non-microbial or (B) microbial stimuli. Summary of two to five experiments (n=4-12) is shown. (C) BMDCs were cultured alone, with nociceptors, or in a transferred medium that nociceptors had been stimulated in and treated with IMQ overnight. Summary of three experiments (n=4) is shown. (D) BMDCs were cultured in Transwell plates either alone or with nociceptors in a manner that allowed or prevented direct physical contact between the cell types. Summary of three experiments (n=6) is shown. IL-12 p40 concentration was measured by ELISA in supernatants. Across all panels, data represent mean ± s.d. Unpaired t test (A and B) or one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test (C and D) were used for statistical analysis: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Fig. 3.. Nociceptors attract and physically interact with DC in vitro.

(A) Migration of BMDCs toward WT nociceptor-conditioned, control, and Ccl2^−/− nociceptor-conditioned medium was assessed in a Transwell chemotaxis assay (5-μm pore size). Summary of three experiments (n=6) is shown. Data represent mean ± s.d. Two-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test was used for statistical analysis: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. (B) DCs from MHC-II eGFP mice were cocultured with nociceptors from NaV1.8^Cre/+ ROSA^tdTomato/+ donors and imaged using a lattice light-sheet microscope (see also movies S2 and S3). Arrowheads highlight some of the points of contact between nociceptors and DCs. Scale bars correspond to 10 μM. (C) Transmission electron micrographs of a DC contacting a nociceptor axon. Two representatives of >10 similar images are shown. Scale bars correspond to 1 μM. (D) Dermal sheets from bone marrow chimeric animals expressing tdTomato under the NaV1.8 promoter in somatic cells and eGFP under the Zbtb46 promoter in the hematopoietic lineage were stained with anti-S100 antibody to identify Schwann cells and analyzed using confocal microscopy.

Fig. 4.. Activation of nociceptors induces calcium mobilization and membrane depolarization in DCs.

(A) Cells in nociceptor: DC cocultures were loaded with a calcium indicator dye Fluo-4 and imaged on a spinning disk confocal microscope. Left panel shows the culture before and right panel after the addition of capsaicin. Representative calcium traces of DCs and proximal axons (arrowheads) are shown in bottom panels (see also movies S4 and S5). (B) Calcium traces of representative responder (green), equivocal (yellow), and unresponsive (dark red) DCs (left panel) and quantification of four independent experiments as shown in Fig. 4A, comparing the proportion of responder cells (middle panel) and the magnitude of calcium response in each cell (right panel) are shown. (C) Nociceptors were plated in one compartment (neuronal compartment) of a microfluidic device consisting of two wells that were separated by fluidically-isolated microgrooves. Once neuronal axons had grown across the microgrooves, DCs were added to the second compartment (DC compartment). All cells were loaded with Fluo-4 and the DC compartment was imaged on a spinning disk confocal microscope after capsaicin addition to the neuronal compartment. Some of the responding DCs are highlighted and their calcium traces are shown in the lower panels. One representative and a quantitation of three independent experiments is shown (see also movie S6). (D) Cells in a nociceptor–DC coculture were loaded with Fluo-4 and a membrane potential indicator dye (BeRST) and imaged on a spinning disk confocal microscope. Top panels show the culture before and bottom panels after the addition of capsaicin. Some of the responding DCs are highlighted and their calcium and membrane potential traces are shown in the right panel. One representative of two experiments is shown (see also movie S10). Across all imaging panels, warmer colors represent higher intracellular calcium concentration and, in panel D, lower membrane potential (depolarization).

Fig. 5.. Nociceptors modulate the DC transcriptome.

(A) Principle component analysis of RNA sequencing data of DCs treated with IMQ or capsaicin in the presence or absence of nociceptors. (B) Venn diagram of differentially regulated genes across DC subsets in coculture with nociceptors compared to DC monoculture. (C and D) Heatmap of genes differentially regulated (F.C.≥2, FDR 0.1) between (C) DCs stimulated with IMQ in monoculture and in coculture with nociceptors or (D) DCs left untreated or treated with capsaicin in the presence or absence of nociceptors. Colored bars and sections of the Venn diagrams designate gene-sets that are uniquely differentially regulated in cDC1s (blue), cDC2s (green), or coregulated in the two subsets (red). Darker shades correspond to upregulated genes, lighter shades to downregulated genes.

Fig. 6.. CGRP induces an enhanced sentinel phenotype in DCs.

Intracellular content of pro-IL-1β was determined by flow cytometry in (A) DCs that were directly cocultured with nociceptors, shared medium with nociceptors, or were cultured alone. (B) DCs that were incubated with nociceptor-conditioned medium in the presence or absence of CGRP receptor antagonist (olcegepant). Summary of 3-4 independent experiments (n=6-8) is shown. (C) Principle component analysis of RNA sequencing data of DCs cultured alone, cocultured with nociceptors or treated with CGRP or Substance P. (D) DCs were treated with CGRP, CGRP and dorapamimod (p38 inhibitor), CGRP and H89 (PKA inhibitor), or forskolin (AC activator) overnight. Expression of pro-IL-1β was determined by flow cytometry. A summary of 3-6 experiments per condition is shown (n=6-12). (E) DCs were treated with CGRP or forskolin for indicated periods of time and lysed. Phosphorylation status of p38 was assessed by phospho-Thr180/Tyr182-specific antibody in immunoblot analysis (top), total p38 was used as a loading control (bottom). One representative of three independent experiments is shown. (F) DCs were plated in wells coated with CCL2 and treated with CGRP overnight. Pro-IL-1β upregulation was assessed by flow cytometry. A summary of eight independent experiments (n=16) is shown. Across all panels, data represent mean ± s.d. One-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test (A), two-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test (B and D), or paired t test (F) were used for statistical analysis: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Fig. 7.. Nociceptors control DC functions in murine skin in vivo.

(A to D) Ears of mice in which nociceptors had been left intact, were chemically ablated, or were treated with lidocaine + QX314 were then treated topically with IMQ cream. Cytokine accumulation was analyzed by tissue lysate ELISA. Summary of 3-4 experiments (n=7-9 per group) is shown. (E) Ears of mice in which nociceptors had been left intact or were chemically ablated were treated with IMQ cream and pro-IL-1β expression by dermal DCs was assessed by flow cytometry. One representative experiment (left) and quantification of three independent experiments (n=7 per group; right) are shown. (F) Ears of mice were treated as in (E) and accumulation of mature IL-1β cytokine in the tissues was assessed by by tissue lysate ELISA. Summary of three experiments (n=7 per group) is shown. (G) Ears of mice in which nociceptors had been left intact, or were treated with lidocaine + QX314 were treated topically with IMQ cream and pro-IL-1β expression by dermal DCs was assessed by flow cytometry. Summary of five experiments (n=11 per group) is shown. (H) Ears of mice were treated as in (G) and accumulation of mature IL-1β cytokine in the tissues was assessed by by tissue lysate ELISA. Summary of three experiments (n=9 per group) is shown. (I) Ears of mice in which nociceptors had been left intact or were chemically ablated were treated with capsaicin and analyzed for pro-IL-1β expression by flow cytometry. One representative experiment (left) and quantification of four independent experiments (n=8 per group; right) are shown. (J) Mouse ears were injected intradermally with PBS, oil, or oil supplemented with olcegepant and pro-IL-1β upregulation in DCs was analyzed by flow cytometry 16 hours later. Quantitation of four independent experiments (n=10-11 per group) is shown. Across all panels, data represent mean ± s.d. Two-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test was used for statistical analysis: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

Fig. 8.. Nociceptor-derived CCL2 is important for induction of local inflammatory and adaptive immune responses.

Dermal sheets from (A) NaV1.8^+/+ CCL2-mCherry (control) and (B) NaV1.8^Cre/+ CCL2-mCherry (NaV1.8^ΔCCL2) animals were fixed, stained with anti-β3-tubulin (green) and anti-mCherry (magenta) antibodies and analyzed using confocal microscopy. Examples of neuronal fibers showing CCL2-mCherry signal are highlighted with arrows (top panels). Surface reconstructions of high-resolution images were performed in Imaris to illustrate the vesicle-like staining pattern of CCL2 (bottom panels). Scale bars correspond to 100 μm (top panels) or 2 μm (bottom panels). (C to E) Ears of NaV1.8^ΔCCL2 or littermate control mice were treated with IMQ cream and (C) inflammatory infiltrate, (D) IL12-p40 accumulation, and (E) ratios of DCs between treated and untreated ears for each animal were analyzed. Summary data of six experiments (n=11-12 per group) is shown. (F and G) Mice were sensitized with DNFB or treated with vehicle control (acetone) and challenged 5 days later with DNFB on one ear and acetone on the other. Ear thickness was measured at indicated timepoints post challenge and specific swelling was calculated as the difference between DNFB and vehicle-treated ear for each animal. (E) NaV1.8^ΔCCL2 or littermate control mice (n=7-8 per group). (F) WT mice treated with lidocaine + OX314 prior to, and on two consecutive days after sensitization. All data are presented as mean ± s.d. Two-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test was used for statistical analysis. In panels (F and G) * correspond to statistical comparison between control DNFB-sensitized and control non-sensitized groups, # correspond to statistical comparison between (F) control DNFB-sensitized and NaV1.8^ΔCCL2 DNFB-sensitized groups, or (G) control DNFB-sensitized and lidocaine + QX314 DNFB-sensitized groups. */#P<0.05, **/##P<0.01, ***/###P<0.001, ****/####P<0.001.