Epithelial colonization by gut dendritic cells promotes their functional diversification

Abstract

Dendritic cells (DCs) patrol tissues and transport antigens to lymph nodes to initiate adaptive immune responses. Within tissues, DCs constitute a complex cell population composed of distinct subsets that can exhibit different activation states and functions. How tissue-specific cues orchestrate DC diversification remains elusive. Here, we show that the small intestine included two pools of cDC2s originating from common pre-DC precursors: (1) lamina propria (LP) CD103⁺CD11b⁺ cDC2s that were mature-like proinflammatory cells and (2) intraepithelial cDC2s that exhibited an immature-like phenotype as well as tolerogenic properties. These phenotypes resulted from the action of food-derived retinoic acid (ATRA), which enhanced actomyosin contractility and promoted LP cDC2 transmigration into the epithelium. There, cDC2s were imprinted by environmental cues, including ATRA itself and the mucus component Muc2. Hence, by reaching distinct subtissular niches, DCs can exist as immature and mature cells within the same tissue, revealing an additional mechanism of DC functional diversification.

Keywords: T cell activation and tolerance; antigen presentation; epithelium; immature and mature dendritic cells; niche; small intestine; transmigration.

Graphical abstract

Figure 1

LP and intraepithelial CD103b⁺CD11b⁺ cDC2s have distinct transcriptomic profiles (A) Gating strategy used to characterize DCs in the epithelium and lamina propria of the whole small intestine from live CD45⁺ cells: cDC1s (CD103⁺CD11b⁻), cDC2s (CD103⁺CD11b⁺ and CD103⁻CD11b⁺), and double negative CD103⁻CD11b⁻ DCs. Example of C57BL/6J mouse. (B) Image of fixed slices from the duodenum of Itgax: Cre/R26^mTmG mice. Sections represent CD11c (yellow), CD11b (cyan), CD103 (magenta), and laminin (blue). Bottom right panel represents merged images. Scale bar, 20 μm. Representative of 3 independent experiments. (C and D) Purified CD11c⁺CD103⁺CD11b⁺ cDCs were analyzed by single-cell RNA-seq using a drop-seq approach. Colors represent samples identification (C) and unbiased clustering from graph-based clustering with resolution of 1 (D). Each dot represents an individual cell from a single experiment. tSNE analysis of individual cells for total cells (n = 1,263). (E) Barplot represents the percentage of cells in each cluster found in each sample. (F) Violin plots showing the expression of Itgax (A), Itgam (B), and Itgae (C) in different clusters obtained by unbiased analysis.

Figure 2

While LP CD103⁺CD11b⁺ cDC2s express proinflammatory markers, intraepithelial cDC2s display an anti-inflammatory phenotype (A) Heatmap of scaled expression (log normalized UMI counts) of the top 20 most differentially expressed genes for each cluster. Antimicrobial genes are highlighted in orange, migration genes in red, inflammatory genes in blue, and cyclin genes in green. (B) Violin plots showing the expression of inflammatory genes: Il1b and Rel proto-oncogene (NF-κB subunit). (C) Violin plots showing the expression of Ccr7, CD83, and CD209a among clusters. (D) Left panel: plot of percentage of GFP positive cells in CD103⁺CD11b⁺ cDC2s as reporter of CCR7 expression in Ccr7^gfp mice lamina propria and epithelium. Data are pooled from 4 independent experiments with n = 3–4 mice per experiment. Mean ± SEM, data were compared using mixed-effects analysis and Sidak’s multiple comparisons test, ∗∗∗∗p < 0.0001. Right panel: geometric mean fluorescence intensity of CD83 comparing the expression in LP versus epithelial CD103⁺CD11b⁺ DCs. Mean ± SEM, data are pooled from 3 independent experiments with n = 2–3 mice per experiment. (E) Geometric mean fluorescence intensity of TNFα and IL-1β measured by flow cytometry intracellular cytokine staining. Data are pooled from 3 independent experiments with n = 2–3 mice per experiment. For IL-1β, data were normalized by experiment. Data were compared using t test. (F) Geometric mean fluorescence intensity of the costimulatory molecules CD86 and CD80 comparing the expression in CD103⁺CD11b⁺ DCs from epithelium and lamina propria. Data are pooled from 3 independent experiments, with n = 2–3 mice per experiment and were compared using Mann-Whitney test or t test. (G) Left panel: geometric mean fluorescence intensity of CD209a comparing the expression in LP versus epithelial CD103⁺CD11b⁺ DCs. Data are pooled from 3 independent experiments normalized by experiment with n = 2–3 mice per experiment and compared using t test. Right panel: Defa24 gene expression measured by quantitative real-time PCR. Fold change of epithelial over LP CD103⁺CD11b⁺ DCs expression normalized against the housekeeping Hprt. Data correspond to 3 independent experiments, with RNA obtained after pooling 4 mice per experiment. In (E)-(G), mean ± SEM, ∗∗p < 0.01, ∗∗∗ p < 0.001, ∗∗∗∗ p < 0.0001.

Figure 3

LP and intraepithelial CD103⁺CD11b⁺ cDC2s originate from common pre-DC precursors (A) Pseudotime reconstruction of the trajectory analysis performed with monocle 2 based on all the differentially expressed genes of the unbiased clustering, by samples identification (top panel) and by cluster (bottom panel). (B) RNA velocity analysis of single-cell RNA-seq data showing that transitional cluster 2 might represent an intermediate cDC2 state between lamina propria cluster 3 and epithelial clusters 0 and 1. (C) Plot of numbers of CD103⁺CD11b⁺ in live CD45⁺ CD11c⁺ MHCII⁺ CD64⁻ cells from the epithelium of the whole small intestine of Flt3l^−/− and Flt3l^+/+ mice. Mean ± SEM, data are pooled from two independent experiments and compared using t test, ∗p < 0.05. Each symbol represents one mouse. (D) Extent of DNGR 1 fate mapping measured as percentage of YFP⁺ cells in CD103⁺CD11b⁺ cells (live CD45⁺ CD11c⁺ MHCII⁺ CD64⁻) from the lamina propria or epithelium of the whole small intestine of Clec9a^+/CreRosa^+/EYFP mice. In total, n = 6 mice corresponding to one independent experiment. Mean ± SEM, data were compared using t test. (E) Flow cytometry dot plots of CD45.2 pre-DC transfer experiments showing differentiation dynamics in the small intestine of syngeneic CD45.1 mice. CD45.2 mice were injected i.v. with pre-DCs sorted from BM of CD45.1 donor mice, and the differentiation into CD103⁺CD11b⁺ DCs was followed after 4, 7, and 10 days after transfer. Representative of 3 independent experiments. (F) Quantification of pre-DC transfer experiments showing the percentage of donor-derived CD45.2⁺ CD103⁺CD11b⁺ in live CD45⁺ CD11c⁺ MHCII⁺ CD64⁻ cells. Mean ± SEM, data are pooled from three independent experiments with n = 2–4 mice per time point per experiment. Data were compared using two-way ANOVA, ∗p < 0.05. (G) Pre-DCs arriving from BM in the small intestinal lamina propria could differentiate first in transitional cluster 2 cDC2s, which constitute an intermediate state between lamina propria cDC2s (cluster 3) and intraepithelial cDC2s (clusters 0 and 1). Please also see Figure S1.

Figure 4

The PILRα paired receptor and the actomyosin cytoskeleton are needed for epithelium colonization by CD103⁺CD11b⁺ cDC2s (A) Violin plot representing the expression of transendothelial migration-related genes among clusters of single-cell RNA-seq analysis: Pilra, Pilrb, and Amica1. (B) Plots of number of CD103⁺CD11b⁺ in live CD45⁺ CD11c⁺ MHCII⁺ CD64⁻ cells from the epithelium and lamina propria of the whole small intestine in Pilra^−/− and Pilra^+/+ female mice. Data are pooled from three independent experiments, each symbol represents one mouse. (C) Pseudotime reconstruction showing the expression of the migratory related gene Abi3. (D) Plots of number of CD103⁺CD11b⁺ in live CD45⁺ CD11c⁺ MHCII⁺ CD64⁻ cells from the epithelium and lamina propria of the whole small intestine in Arpc4^−/− (Arpc4^flox/flox × Itgax^Cre+) and Arpc4^+/+ (Arpc4^flox/flox × Itgax^Cre−) mice. Data are pooled from three independent experiments, each symbol represents one mouse. (E) Pseudotime reconstruction showing the expression of the migratory related gene Myh9 (MyoIIA). (F) Plots of number of CD103⁺CD11b⁺ in live CD45⁺ CD11c⁺ MHCII⁺ CD64⁻ cells from the epithelium and lamina propria of the whole small intestine in MyoIIA deficient (Myh9^flox/flox × Itgax^Cre+) and WT (Myh9^flox/flox × Itgax^Cre−) mice. Data are pooled from three independent experiments, each symbol represents one mouse. In (B), (D), and (F), mean ± SEM, data were compared using Mann-Whitney test or t test, ∗p < 0,05, ∗∗∗p < 0,001. Please also see Figures S2 and S3.

Figure 5

CD103⁺CD11b⁺ cDC2s transmigration into the epithelium depends on ATRA (A) Adult SPF C57BL/6 mice were gavaged with PBS, antibiotic cocktail (Atb) or fluconazole (Fluc) for 10 days, and the numbers of CD103⁺CD11b⁺ cells in epithelium and lamina propria were determined by flow cytometry. Mean ± SEM, data are pooled from 3 independent experiments and compared using one-way ANOVA and Tukey’s multiple comparisons test. Each symbol represents one mouse. (B) Plots of number of CD103⁺CD11b⁺ in live CD45⁺ CD11c⁺ MHCII⁺ CD64⁻ cells from the epithelium (left panel) or lamina propria (right panel) of small intestinal duodenum, jejunum, and ileum. Data are pooled from two independent experiments. Mean ± SEM, data were compared using Kruskal-Wallis test or paired one-way ANOVA and Tukey’s multiple comparisons test, ∗p < 0.05. Each symbol represents one mouse. (C) Violin plot showing the expression of Spi1 gene among clusters of single-cell RNA-seq analysis of CD103⁺CD11b⁺ DCs. Cluster 0 versus 3: p value 3,11E-11; cluster 1 versus 3: p value 2,84E-18. (D) Retinoic acid receptor alpha expression measured by intracellular staining of CD103⁺CD11b⁺ DCs comparing epithelium and lamina propria. Mean ± SEM, data are pooled from 3 independent experiments with n = 2–3 per experiment and compared using t test, ∗p < 0.05. (E) Adult SPF C57BL/6J mice were gavaged with olive oil or bisdiamine for 2 days. Plots represent the number of CD103⁺CD11b⁺ cells in the epithelium (left) and lamina propria (middle) from the entire small intestine determined by flow cytometry. Mean ± SEM, data are pooled from 2 independent experiments and representative of 3 independent experiments. Each symbol represents one mouse. Data were compared using t test, ∗∗p < 0.01, ∗∗∗∗p < 0.0001. Right panel: percentage of intraepithelial DCs among total CD103⁺CD11b⁺ dendritic cells in control or bisdiamine-treated mice. (F) Flow cytometry analysis of CD103⁺CD11b⁺ DC numbers from the small intestinal epithelium analyzed in duodenum, jejunum, and ileum of SPF C57BL/6J mice fed with vitamin A-deficient, excess vitamin A, or control diet for 3 months. Mean ± SEM, data are pooled from 2 independent experiments and compared using two-way ANOVA, ∗p < 0.05. Each symbol represents one mouse. (G) Quantification of cells inside microchannels of 4 × 4 μm crossing constrictions of 1.5–2 μm. Lamina propria CD103⁺CD11b⁺ DCs were sorted from GFP-tagged Myh9 mice and let them migrate inside microchannels with constrictions overnight. Control CD103⁺CD11b⁺ DCs or CD103⁺CD11b⁺ DCs treated with 1 nM ATRA were included in the experiment (left panel). Speed of lamina propria CD103⁺CD11b⁺ DCs while migrating into the constrictions measured as μm/min (right panel). (H) Quantification of myosine IIA mean intensity of lamina propria CD103⁺CD11b⁺ DCs while migrating through microchannels. Values were normalized against control average (left panel). Front/back ratio of myosin IIA distribution inside CD103⁺CD11b⁺ LP DCs while migrating through microchannels (right panel). In (G) and (H), mean ± SEM, data are pooled from 2 independent experiments. Total number of cells analyzed: 157 cells (control), 316 (ATRA). Data were compared using Mann-Whitney test, ∗∗p < 0.01, ∗∗∗∗p < 0.0001. (I) Upper panels: zoom-in of CD103⁺CD11b⁺ LP DCs migrating intro microchannels and passing through constrictions of between 1.5 and 2μm for control and ATRA conditions. Scale bar, 40 μm. Lower panels: montage of CD103⁺CD11b⁺ LP DCs migrating into microchannels. White stars show CD103⁺CD11b⁺ DCs zoomed-in on the upper panels. Scale bar, 100 μm. (J) Heatmap of myosin IIA distribution of LP CD103⁺CD11b⁺ DCs passing through the constrictions. Please also see Figure S4.

Figure 6

Epithelial colonization imprints CD103⁺CD11b⁺ cDC2s with an immature-like phenotype (A) Transwell experiments were performed with purified lamina propria CD103⁺CD11b⁺ DCs, and geometric mean fluorescence intensity of CCR7 and CD86 was analyzed by flow cytometry after overnight transmigration. DCs transmigrated from the upper to the lower compartment passing through pores of 3 μm. Mean ± SEM, data are pooled from 3 independent experiments and compared using t test. (B) Freshly obtained SI epithelial cells were obtained from the small intestine of C57BL/6J mice, as well as sorted lamina propria CD103⁺CD11b⁺ DCs. Gut epithelial cell supernatant (SN) was obtained after 6 h of incubation at 37°C with 5% CO₂. Sorted LP CD103⁺CD11b⁺ DCs were either incubated alone, with gut epithelial cells or their supernatant overnight. Differentially expressed markers were analyzed by flow cytometry. (C) Geometric mean fluorescence intensity of CCR7 and CD86 comparing LP CD103⁺CD11b⁺ DCs incubated with small intestine epithelial cells, supernatant of small intestine epithelial cells, or control LP CD103⁺CD11b⁺ DCs alone. Mean ± SEM, data are pooled from 3 independent experiments and compared using one-way ANOVA or Kruskal-Wallis test, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. (D) Small intestinal duodenal organoids were derived from a membrane-fluorescent mice reporter, and after 4 days of culture, LP CD103⁺CD11b⁺ DCs purified from Itgax-EYFP reporter mice were added to the culture. Images of the interaction between DCs and organoids. Scale bar, 80 μm. Representative of two independent experiments. (E) Geometric mean fluorescence intensity of CCR7 and CD86 comparing LP CD103⁺CD11b⁺ DCs incubated with organoids derived from duodenum or control LP cDC2s. Mean ± SEM, data are pooled from 3 independent experiments and compared using t test, ∗∗∗∗p < 0.0001. (F) Sorted LP CD103⁺CD11b⁺ DCs were incubated alone or with freshly obtained gut epithelial cells overnight, with either ATRA 1 nM, bisdiamine 45 μM, or anti-TGF-β 10 μg/ml. Differentially expressed markers were analyzed by flow cytometry, and geometric mean fluorescence intensity of CCR7 and CD86 is represented. Mean ± SEM, data are pooled from 3 independent experiments and compared using one-way ANOVA and Tukey multiple comparisons test, ∗∗∗p < 0.001. (G) Sorted LP CD103⁺CD11b⁺ DCs were incubated overnight with the mucus protein Muc2 (50 μg/ml), and CCR7, CD86, and CD209a expression was analyzed by flow cytometry. Mean ± SEM, data are pooled from 3 independent experiments (CCR7 and CD86) or two independent experiments (CD209a). Data were compared using t test (∗∗∗p < 0.001) or Mann-Whitney test (∗p < 0.05). Please also see Figure S5.

Figure 7

Intraepithelial CD103⁺CD11b⁺ cDC2s are endowed with tolerogenic properties (A) Left panel: percentage of CD69 positive CD4⁺ T cells in live cells after 18 h of incubation with OVA preincubated CD103⁺CD11b⁺ dendritic cells. Right panel: percentage of CFSE⁻ CD4⁺ T cells in live cells after 3 days of incubation with OVA or OVA peptide II preincubated CD103⁺CD11b⁺ dendritic cells, analyzed by flow cytometry. OT-II antigen presentation assay was performed with sorted CD103⁺CD11b⁺ DCs from both small intestine lamina propria and epithelium from C57BL/6J mice. Mean ± SEM, data are pooled from five independent experiments and compared using multiple t test or Mann-Whitney test, ∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. (B) IL-2 secretion by OT-II T cells measured by Luminex. Representative of 2 independent experiments. Mean ± SEM, data were compared using Sidak’s multiple comparisons test, ∗∗p < 0.01, ∗∗∗p < 0.001. (C) CD4⁺ OT-II T cells were incubated overnight with CD103⁺CD11b⁺ DCs (previously incubated with OVA or OVAp). T cells were incubated with anti-CD3 and anti-CD28 antibodies, and the percentage of CD69 positive cells (left panel) or CFSE negative cells (right panel) was measured after overnight or 2 days, respectively. Data are pooled from four independent experiments. Fresh OT-II T cells preactivated or not with anti-CD3 and anti-CD28 antibodies were used as controls (two independent experiments). (D) Plot of percentage of OVA-Texas Red positive cells after incubation of sorted CD103⁺CD11b⁺ DCs from lamina propria and epithelium with fluorescent OVA, measured by flow cytometry. Mean ± SEM, data are pooled from three independent experiments and compared using two-way ANOVA and Sidak’s multiple comparisons test, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. (E) Images of cytospin and MGG staining of CD103⁺CD11b⁺ DCs from lamina propria and epithelium after sorting. Representative of 2 independent experiments. (F) Geometric mean fluorescence intensity of the costimulatory molecule ICOS-L comparing the expression in CD103⁺CD11b⁺ DCs from epithelium and lamina propria. Mean ± SEM, data are pooled from two independent experiments, normalized by experiment, and compared using t test, ∗∗∗p < 0.001. Please also see Figure S6.