A novel form of immune signaling revealed by transmission of the inflammatory mediator serotonin between dendritic cells and T cells

Abstract

Adaptive immunity is triggered at the immune synapse, where peptide-major histocompatibility complexes and costimulatory molecules expressed by dendritic cells (DCs) are physically presented to T cells. Here we describe transmission of the inflammatory monoamine serotonin (5-hydroxytryptamine [5-HT]) between these cells. DCs take up 5-HT from the microenvironment and from activated T cells (that synthesize 5-HT) and this uptake is inhibited by the antidepressant, fluoxetine. Expression of 5-HT transporters (SERTs) is regulated by DC maturation, exposure to microbial stimuli, and physical interactions with T cells. Significantly, 5-HT sequestered by DCs is stored within LAMP-1+ vesicles and subsequently released via Ca2+-dependent exocytosis, which was confirmed by amperometric recordings. In turn, extracellular 5-HT can reduce T-cell levels of cAMP, a modulator of T-cell activation. Thus, through the uptake of 5-HT at sites of inflammation, and from activated T cells, DCs may shuttle 5-HT to naive T cells and thereby modulate T-cell proliferation and differentiation. These data constitute the first direct measurement of triggered exocytosis by DCs and reveal a new and rapid type of signaling that may be optimized by the intimate synaptic environment between DCs and T cells. Moreover, these results highlight an important role for 5-HT signaling in immune function and the potential consequences of commonly used drugs that target 5-HT uptake and release.

Figure 1.

Expression and developmental regulation of the 5-HT transporter and MAO in mouse DCs. (A) CD11c⁺ CD86^- (immature DCs) or CD11c⁺ CD86⁺ (mature DCs) were purified by cell sorting from BMDC cultures or spleen cells. BMDCs were activated by overnight exposure to LPS (50 ng/mL) prior to sorting CD11c⁺ CD86⁺ (activated DCs). Gene expression for SERT, TPH-1 and TPH-2, MAO-A, and MAO-B was examined by RT-PCR. (B) Density gradient-enriched day 5 BMDCs were immunolabeled and analyzed by confocal microscopy. CD11c (mouse DC marker) and SERT were visualized with antibodies labeled with Alexa Fluor 488 (green; i) and Alexa Fluor 546 (red; ii), respectively. Nuclei were counterstained with TO-PRO-3 (blue). In the red, green, and blue overlay (iii), the yellow signal indicates regions of red/green overlap. The bright-field image (iv) includes a CD11c and SERT double-negative cell with lymphocytic morphology (arrowhead). Scale bar represents 10 μm. Data are representative of 2 independent experiments.

Figure 2.

Serotonin uptake by BMDCs. (A-B) [³H]5-HT uptake by BMDCs and whole mouse brain synaptosomes at 4°C (▪) and 37°C (○) and specific uptake (▵) versus 5-HT concentration. Data are from a single experiment performed in duplicate and are representative of 3 independent experiments. Note that brain protein concentration was about 5 times greater than BMDCs. The smooth curves are fits to a hyperbolic function used to derive K_m and V_max (see “Serotonin uptake by DCs”). The stated K_m values are the means (± SE) from 3 experiments with BMDCs and brain synaptosomes. (C) Mean inhibition of specific [³H]5-HT (100 nM) uptake by fluoxetine (200 nM) in BMDCs and brain. Data are means (± SD) from 2 experiments (^*P < .05, paired t test). (D) Day 5 BMDCs were purified by cell sorting for CD11c-PE-Cy7⁺ cells and then loaded with 5-HT (100 μM). LAMP-1 and 5-HT were visualized by confocal microscopy with antibodies labeled with Alexa Fluor 488 (green; i) and Alexa Fluor 546 (red; ii), respectively. Nuclei were counterstained with TO-PRO-3 (blue). In the red, green, and blue overlay (iii), the yellow signal indicates regions of red/green overlap. Scale bar represents 20 μm. (E) Magnification of bright field (i) and red/green/blue overlay (ii) from a single DC. Scale bar represents 10 μm.

Figure 3.

Ca²⁺-triggered exocytosis of serotonin in BMDCs. (A) [³H]5-HT release (% of control) from BMDCs in response to extracellular ATP (100 μM) and ionomycin (10 μM). Data are means ± SE (n = 3; ^*P < .05; ^**P < .01). (B) BMDCs with apposed 5-μm carbon-fiber electrode; scale bar represents 20 μm. (C) Amperometric spikes evoked by application of extracellular ATP (100 μM) in a cell preloaded with 5-HT. (D) Representative spikes are shown on an expanded time scale. (E) Distribution of spike amplitudes (2 pA bins) from a total of 336 single events taken from 37 cells.

Figure 4.

Serotonin inhibits naive T-cell production of cAMP. (A) Change in levels of cAMP in naive T cells in response to single or combined treatment as indicated with 5-HT (10 μM), the PDE-4 inhibitor, rolipram (30 μM), and the type 1 5-HT receptor inhibitor, S-Way 100135 (1 μM). Each data point is the mean (± SE) of 4 to 5 experiments (^*P < .01 compared with control; †P < .05 compared with other rolipram groups by ANOVA). (B) Dose-response relationship for inhibition of cAMP by 5-HT in T cells pretreated with rolipram. The smooth line is the fit to a Hill equation.

Figure 5.

T cells express the enzyme for 5-HT synthesis and T-cell 5-HT is sequestered by BMDCs. (A) Expression of TPH-1 and TPH-2 was determined by RT-PCR in naive T cells, Con A- (5 μg/mL) activated T cells, and Con A-activated T cells rested for 4 days in IL-2 (10 μg/mL). Comparable quantities of cDNA were ensured by amplification of GAPDH. Data are representative of 2 experiments. (B) BMDC uptake of [³H]5-HT from T-cell supernatants at 0°C, room temperature, and room temperature with fluoxetine (100 nM). Freshly purified, naive T cells were loaded with 1 μM [³H]5-HT (1 hour, 37°C). Cells were washed and resuspended in HBSS (1 hour, 37°C) to passively release [³H]5-HT. BMDCs were resuspended in T-cell supernatant (2 hours, room temperature). Data are means ± SD (n = 2; ^*P < .001,t test).

Figure 6.

SERT expression is coupled to DC functional maturation. (A) BMDCs were incubated for 2 days with LPS (50 ng/mL), cross-linked anti-CD40 mAb (5 μg/mL), or CD152/Fc (40 μg/mL), or left untreated. Solid histograms show the expression of CD86 and MHC II molecules by gated, CD11c⁺ BMDCs relative to isotype controls (dashed histograms). (B) Immature (CD11c⁺ CD86^-) BMDCs were sorted and were cultured for 2 days as described. IL-6 and IL-12 concentration in the culture supernatants was determined by ELISA for treated BMDCs (▪) relative to untreated controls (unT; □). Data are means ± SD from triplicate assays (n = 3; ^*P < .05, t test). (C) The expression of SERT mRNA relative to GAPDH controls was examined by RT-PCR for each treatment group prepared as described in panel B. The optical density of each band relative to untreated BMDCs was determined and normalized for GAPDH signal. Data are representative of 2 independent experiments.

Figure 7.

A conceptual model for 5-HT uptake and release by DCs and the immunologic consequences. Immature DCs (iDC) may sequester 5-HT released by platelets or mast cells at sites of injury and inflammation, which they in turn secrete via a Ca²⁺-dependent exocytotic pathway on encountering naive, cognate T cells. High local concentrations of 5-HT within DC–T-cell synaptic junctions suppress further production of cAMP and facilitate T-cell proliferation. CD40-CD40 ligand signaling up-regulates DC SERT expression and thereby promotes uptake of 5-HT passively released from activated T cells. Mature DCs (mDC) may shuttle 5-HT between activated and naive T cells, thereby amplifying adaptive immune responses.