Gut Microbiota Species Can Provoke both Inflammatory and Tolerogenic Immune Responses in Human Dendritic Cells Mediated by Retinoic Acid Receptor Alpha Ligation

Abstract

Dendritic cells are considered as the main coordinators of both mucosal and systemic immune responses, thus playing a determining role in shaping the outcome of effector cell responses. However, it is still uncovered how primary human monocyte-derived DC (moDC) populations drive the polarization of helper T (Th) cells in the presence of commensal bacteria harboring unique immunomodulatory properties. Furthermore, the individual members of the gut microbiota have the potential to modulate the outcome of immune responses and shape the immunogenicity of differentiating moDCs via the activation of retinoic acid receptor alpha (RARα). Here, we report that moDCs are able to mediate robust Th1 and Th17 responses upon stimulation by Escherichia coli Schaedler or Morganella morganii, while the probiotic Bacillus subtilis strain limits this effect. Moreover, physiological concentrations of all-trans retinoic acid (ATRA) are able to re-program the differentiation of moDCs resulting in altered gene expression profiles of the master transcription factors RARα and interferon regulatory factor 4, and concomitantly regulate the cell surface expression levels of CD1 proteins and also the mucosa-associated CD103 integrin to different directions. It was also demonstrated that the ATRA-conditioned moDCs exhibited enhanced pro-inflammatory cytokine secretion while reduced their co-stimulatory and antigen-presenting capacity thus reducing Th1 and presenting undetectable Th17 type responses against the tested microbiota strains. Importantly, these regulatory circuits could be prevented by the selective inhibition of RARα functionality. These results altogether demonstrate that selected commensal bacterial strains are able to drive strong effector immune responses by moDCs, while in the presence of ATRA, they support the development of both tolerogenic and inflammatory moDC in a RARα-dependent manner.

Keywords: CD1a; CD1d; T cell; all-trans retinoic acid; gut microbiota; interferon regulatory factor 4; monocyte-derived dentritic cell; retinoic acid receptor alpha.

Figure 1

The effects of all-trans retinoic acid (ATRA) on human monocyte-derived dentritic cell (moDC) differentiation. Two-day moDCs were differentiated in the absence or presence of 1 nM ATRA. The relative gene expression levels of retinoic acid receptor alpha (RARα), retinoic X receptor alpha (RXRα), peroxisome proliferator-activated receptor gamma (PPARγ), ALDH1A2, and CD1d (A), interferon regulatory factor (IRF)4 (F), and IRF8 (G) were measured by quantitative real-time PCR, and the cell surface expression level of CD1a, CD1d (B), DC-SIGN (C), CD14 (D), and CD11b (E) was measured by flow cytometry. Mean values of relative mRNA levels and the ratio of moDCs positive for the measured cell surface proteins were calculated from five independent experiments +SD. Student’s unpaired two-tailed t-test was used in the statistical analysis with significance defined as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Figure 2

Monitoring monocyte-derived dentritic cell (moDC)-mediated T-lymphocyte polarization induced by commensal stimuli. The T cell polarizing capacity of moDCs was monitored in moDC stimulated with Escherichia coli Schaedler, Morganella Morganii, and Bacillus subtilis or lipopolysaccharide (LPS) followed by co-culturing the cells with freshly isolated autologous T cells for 4 days. The number of cytokine producing T-lymphocytes induced by LPS or by moDCs exposed to commensal bacteria was measured by interferon gamma (IFNγ) (A) and interleukin (IL)-17 (B) enzyme-linked ImmunoSpot assays. T corresponds to T cells cultured without dendritic cells as negative control. The mean value of spot numbers was calculated from five independent experiments +SD. ANOVA followed by Bonferroni’s multiple comparison tests was used in the statistical analysis with significance defined as *P < 0.05, **P < 0.01, and ***P < 0.001.

Figure 3

All-trans retinoic acid (ATRA) shifts the cell surface expression pattern of CD1, gut-related receptors, and the phagocytic capacity of monocyte-derived dentritic cells (moDCs) in an ATRA and commensal strain-dependent manner. Human moDCs were differentiated in the presence of granulocyte-macrophage colony-stimulating factor and interleukin-4 with or without 1 nM ATRA for 2 days. The surface expression level of CD1a and CD1d was measured on resting cells and moDCs activated with live commensal bacteria for 24 h (A) by flow cytometry. Histogram overlays show results derived from 1 representative donor of 10. The cell surface expression level of the mucosa-related CX₃CR1 (B,C) and CD103 (D,E) was measured by flow cytometry followed by a 24-h activation period with live commensal bacteria or lipopolysaccharide (LPS) served as a positive control. Mean values showing the ratio of moDCs positive for the measured surface protein were calculated from five independent experiments +SD. To monitor the phagocytic capacity of moDCs, on day 2, moDCs were co-cultured with heat-inactivated and fluorescein-isothiocyanate (FITC)-labeled bacteria at 37°C or at 4°C for 3 h at a moDC:bacteria ratio of 1:20. (F,G) Dot plots show one of four independent experiments. The ratio of moDC positive for heat-inactivated and FITC-labeled bacteria was measured by flow cytometry. The number of moDCs carrying FITC-labeled bacteria was calculated from four independent experiments +SD. ANOVA followed by Bonferroni’s multiple comparison tests was used in the statistical analysis with significance defined as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Figure 4

Characteristics of the inflammatory and migratory potential of monocyte-derived dentritic cell (moDC) populations induced by commensal bacteria. Two-day moDCs were co-incubated with live commensal strains or with 250 ng/ml lipopolysaccharide (LPS) used as control for 24 h. Expression of the moDC-associated activation marker CD83 (A) and CCR7 (B) was measured by flow cytometry. Mean values were calculated from five to seven independent experiments +SD. The concentration of TNF-α, interleukin (IL)-1β, IL-6 pro-inflammatory cytokines, and the chemokine CXCL8 (C) was measured by ELISA followed by a 24-h activation of moDC in five independent experiments. Mean values +SD are shown. ANOVA followed by Bonferroni’s multiple comparison tests was used in the statistical analysis with significance defined as *P < 0.05, ***P < 0.001, and ****P < 0.0001.

Figure 5

The T-lymphocyte activating and polarizing capacity of monocyte-derived dentritic cells (moDCs) activated by selected commensal bacteria. Two-day moDCs were co-incubated with live commensal strains or with 250 ng/ml lipopolysaccharide (LPS) used as control for 24 h. The expression levels of HLA-DQ and HLA-DR (A), the co-stimulatory proteins CD80 and CD86 (B), CD40 (D), and the inhibitory molecule PD-L1 (E) was measured by flow cytometry. Mean values of median fluorescence intensities (MFIs) were calculated from five to seven independent experiments +SD. The concentration of interleukin (IL)-12, IL-23, and IL-10 cytokines was measured by ELISA followed by a 24-h activation of moDC and was tested in seven independent experiments (C). Mean values +SD are shown. ANOVA followed by Bonferroni’s multiple comparison tests was used in the statistical analysis with significance defined as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Figure 6

The selective inhibition of retinoic acid receptor alpha (RARα) prevents the all-trans retinoic acid (ATRA)-induced signature of microbiota-generated immune responses mediated by monocyte-derived dentritic cells (moDCs). To analyze how ATRA acts on the moDC-mediated immune response against microbiota species, the cells were treated with the RARα antagonist BMS614 prior to treating the cell culture medium with ATRA. The cell surface expression level of CD1a, CD1d, and CD103 was measured by flow cytometry in 2-day moDCs (A). The concentration of TNF-α, interleukin (IL)-6, and IL-1β (B) and IL-23 (C) was measured by ELISA followed by a 24-h activation of moDC performed in seven independent experiments. Mean values +SD are shown. The cell surface expression level of HLA-DQ was measured by flow cytometry followed by a 24-h incubation period with live commensal bacteria. (D) Mean values of cells positive for the measured cell surface molecules were calculated from the results of seven independent donors +SD. Analysis of interferon regulatory factor 4 (IRF4) expression in moDCs. (E) Two-day moDCs were activated by live commensal bacteria for 24 h, and the relative expression levels of IRF4 protein was measured by Western blotting. Bar graphs show IRF4/β-actin ratios measured after 24 h of stimulation. Mean values of protein densities were calculated from five independent experiments +SD. The T cell polarizing capacity of moDCs was monitored in moDCs activated with the selected commensal strains or with lipopolysaccharide (LPS) followed by co-culturing the cells with autologous T cells. Freshly isolated peripheral blood lymphocytes were co-cultured with autologous moDCs for 4 days. The number of cytokine producing T-lymphocytes, induced by LPS or moDCs exposed to Escherichia coli Schaedler and Bacillus Subtilis, was measured by interferon gamma (IFNγ) (F) and IL-17 (G) enzyme-linked ImmunoSpot assays. The mean value of spot numbers was calculated from five independent experiments +SD. Statistical analysis was performed by the Student’s unpaired two-tailed t-test with significance defined as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Figure 7

The role of retinoic acid receptor alpha (RARα) in guiding monocyte-derived dentritic cell (moDC) development and microbiota-induced immune responses. All-trans retinoic acid (ATRA) modifies the differentiation of moDCs that could be prevented by the selective inhibition of RARα. (A) In the presence of granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin (IL)-4, monocytes differentiate to CD1d⁻CD1a⁺ DCs (43). Peroxisome proliferator-activated receptor gamma (PPARγ) and RARα regulate the gene expression of CD1d and ALDH1A2 both directly and indirectly in human moDCs (20). Interferon regulatory factor 4 (IRF4) mediates the differentiation and antigen-presenting capacity of human moDCs, which is downregulated by the ligation of RARα resulting in decreased mRNA and protein levels of IRF4 together with CD1a. Selected microbiota species provoke different types of immune responses mediated by moDCs. (B) Escherichia coli Schaedler induces full maturation in moDCs leading to strong inflammatory and microbiota-induced effector helper T (Th) responses, while Bacillus subtilis induces inflammation in the absence of IL-12 and IL-23 and provokes decreased effector Th1/Th17 immune responses. ATRA downmodulates the immunogenicity of moDCs resulting in diminished Th1 and undetectable Th17 responses, which effect can be restored in moDCs by the prior inhibition of RARα in moDCs. Solid lines represent known mechanisms; dotted lines indicate unknown molecular interactions.