CD1d-mediated lipid presentation by CD11c+ cells regulates intestinal homeostasis

Abstract

Intestinal homeostasis relies on a continuous dialogue between the commensal bacteria and the immune system. Natural killer T (NKT) cells, which recognize CD1d-restricted microbial lipids and self-lipids, contribute to the regulation of mucosal immunity, yet the mechanisms underlying their functions remain poorly understood. Here, we demonstrate that NKT cells respond to intestinal lipids and CD11c⁺ cells (including dendritic cells (DCs) and macrophages) are essential to mediate lipid presentation within the gut ultimately controlling intestinal NKT cell homeostasis and activation. Conversely, CD1d and NKT cells participate in the control of the intestinal bacteria composition and compartmentalization, in the regulation of the IgA repertoire and in the induction of regulatory T cells within the gut. These changes in intestinal homeostasis require CD1d expression on DC/macrophage populations as mice with conditional deletion of CD1d on CD11c⁺ cells exhibit dysbiosis and altered immune homeostasis. These results unveil the importance of CD11c⁺ cells in controlling lipid-dependent immunity in the intestinal compartment and reveal an NKT cell-DC crosstalk as a key mechanism for the regulation of gut homeostasis.

Keywords: CD1d; NKT cell; microbiota.

Figure 1. Intestinal lipids presented by CD11c⁺ cells activate iNKT cells

1. A

   Flow cytometry analysis for GFP expression on iNKT (TCRβ⁺αGalCer:tet⁺) cells and CD4⁺ T cells from Nur77^GFP+ reporter mice (grey histogram) and Nur77^GFP− controls (empty histogram) in the indicated tissues. Numbers indicate percentage of iNKT cells among TCRβ⁺ cells (left panels), of Nur77^GFP+ iNKT cells (middle) and of Nur77^GFP+CD4⁺ T cells (right). Data represent three experiments.

2. B–E

   WT C57BL/6 mice were orally gavaged with αGalCer or PBS, and 16 h later, mLN iNKT cells were analysed. (B) Flow cytometry profiles show intracellular Nur77 expression (grey histogram) or FMO (empty histogram) in iNKT cells from the mLN of αGalCer‐treated or control mice. (C) Frequency of Nur77‐expressing iNKT cells (left), Nur77 MFI (middle) and qPCR analysis of mRNA encoding Nur77 (right) in freshly isolated iNKT cells from the mLN of αGalCer‐treated or control mice. qPCR data are normalized to GAPDH. RE, relative expression. (D) Expression of CD69 in mLN iNKT cells from control (empty histogram) or αGalCer‐treated (grey histogram) mice. Data are from three experiments. (E) Flow cytometry analysis showing GFP expression on iNKT cells from Nur77^GFP+ reporter mice (grey histograms) and Nur77^GFP− controls (empty histogram) from the mLN of αGalCer‐treated or control mice.

3. F, G

   DC populations in the mLN from Cre⁺ and Cre⁻ CD1d^fl/flCD11c^Cre mice were analysed by flow cytometry. Gating strategy (F) and CD1d expression (G) for CD103⁺CD11b⁻ (red), CD103⁺CD11b⁺ (grey) and CD103⁻CD11b⁺ (blue) cells. Data represent four experiments.

4. H

   Cre⁻ and Cre⁺ CD1d^fl/flCD11c^Cre mice were orally gavaged with αGalCer, and 16 h later, Nur77 expression on iNKT cells was analysed in the mLN. Flow cytometry profiles show Nur77 (grey histogram) or FMO (empty histogram) in iNKT cells from the mLN of αGalCer‐treated mice. Right panels show frequency of Nur77‐expressing iNKT cells in αGalCer‐treated or control Cre⁻ and Cre⁺ CD1d^fl/flCD11c^Cre mice. Data are from three experiments.

5. I

   Single‐cell suspension from mLNs from Cre⁻ and Cre⁺ CD1d^fl/flCD11c^Cre mice were prepared and incubated with commensals. Flow cytometry profiles show Nur77 expression in iNKT cells after incubation with commensals (blue histogram) or control (grey histogram). Right panels show Nur77 MFI for iNKT cells. Data are from three experiments.

Data information: Numbers indicate percentage of cells in the indicated gates. Each dot in the bar plots is an individual mouse and bars represent mean ± SEM. *P < 0.05, **P < 0.01, two‐tailed unpaired (C, H) or paired (I) t‐test.

Figure 2. Lipid presentation by CD11c⁺ cells controls iNKT cell subsets

1. A, B

   Analysis of iNKT cells populations in the SI‐LP (A) and mLN (B) from CD1d^fl/flCD11c^Cre mice, showing flow cytometry plots and frequency of total iNKT cells (grey) and NKT1 (RORγt⁻PLZF^loT‐bet⁺; green), NKT2 (RORγt⁻PLZF^hiT‐bet⁻; red), NKT17 (PLZF^intRORγt⁺; blue) and PLZF^loT‐bet⁻ NKT cells (orange; mean ± SEM). Data are from 4–6 experiments.

2. C–E

   C57BL/6 WT mice were orally gavaged with αGalCer, and iNKT cell populations were analysed 3 days later. Flow cytometry plots (C), frequency (D) and total number (E) of iNKT cells and NKT1, NKT2 and NKT17 cells in the mLN are shown. Data are from three experiments.

3. F

   Histograms showing Ki‐67 expression (left) and frequency of Ki‐67⁺ (right) NKT1, NKT2 and NKT17 cells in the mLN from C57BL/6 WT mice 2 days after oral αGalCer administration.

4. G, H

   Frequency (G) and total number (H) of iNKT cells and NKT1, NKT2 and NKT17 cells in the mLN from Cre⁺ and Cre⁻ CD1d^fl/flCD11c^Cre mice 3 days after oral αGalCer administration. Data are from 2–3 experiments.

Data information: Numbers indicate percentage of cells in the indicated gates. Each dot in the bar plots is an individual mouse and bars represent mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, two‐tailed unpaired t‐test.

Figure 3. CD1d and NKT cells regulate the intestinal microbiota

1. A

   Principal coordinates analysis (PCoA) using the Yue–Clayton distances obtained for bacterial samples from the ileum content and ileum wall of CD1d^+/− and CD1d^−/− mice. The axes show the percentage of variation explained by PC1 and PC2. Each dot corresponds to one mouse.

2. B

   Average relative abundance of the most frequent (> 1%) operational taxonomic units (OTUs) of the ileum content and ileum wall from CD1d^+/− and CD1d^−/− mice. Bacterial taxa (at the genus level, or the closest level of classification) are shown, grouped by phylum and labelled with different colours as indicated. UC, unclassified.

3. C, D

   Relative abundance of specified OTUs in the ileum content (C) and of specified taxa in the ileum wall (D) from CD1d^+/− and CD1d^−/− mice.

4. E–G

   C57BL/6 mice were orally administered αGalCer, and faecal bacteria were analysed before (d0) and 10 days (d10) after the treatment. (E) PCoA using the Yue–Clayton distances obtained among faecal samples at d0 and d10. The axes show the percentage of variation explained by PC1 and PC2. Each dot corresponds to one mouse. (F) Average relative abundance of the most frequent (> 1%) OTUs at d0 and d10. Taxa are shown and labelled with different colours as indicated. (G) Relative abundance of the specified phyla, before and 10 days after αGalCer treatment.

Data information: In the boxplots, lines indicate the median, boxes show the 75^th and the 25^th percentiles and whiskers indicate the maximum and minimum values. *P < 0.05, **P < 0.01, two‐tailed Wilcoxon test.

Figure 4. CD1d expression on CD11c⁺ cells shapes the intestinal microbiota

1. Average relative abundance of the most frequent (> 1%) OTUs in the ileum content from Cre⁻ and Cre⁺ mice of the CD1d^fl/flCD11c^Cre strain (n = 4–5). Bacterial taxa (at the genus level, or the closest level of classification) are shown, grouped by phylum and labelled with different colours as indicated.

2. Relative abundance of the specified OTUs in the ileum content from Cre⁻ and Cre⁺ CD1d^fl/flCD11c^Cre mice (n = 4–5). Lines indicate the median, boxes show the 75^th and the 25^th percentiles and whiskers indicate the maximum and minimum values. *P < 0.05, two‐tailed Wilcoxon test.

3. Localization of bacteria (brown) and intestinal cells (blue), detected by RNAscope and haematoxylin staining, respectively, in ileum sections from CD1d^+/− and CD1d^−/− mice. Data represent two experiments. Scale bar, 100 μm.

4. Total cell numbers in the mLN of CD1d^+/− and CD1d^−/− mice. Each dot in the bar plots is an individual mouse and bars represent mean ± SEM. **P < 0.01, two‐tailed unpaired t‐test.

5. Yersinia enterocolitica translocation to the mLNs from CD1d^+/− and CD1d^−/− mice orally infected with Y. enterocolitica (1 × 10⁹ CFU), assessed 24 h after by selective plating. Data are pooled from three experiments (n = 9). *P < 0.05, two‐tailed paired t‐test.

6. Localization of bacteria (brown) and intestinal cells (blue), detected by RNAscope and haematoxylin staining in ileum sections from Cre⁻ and Cre⁺ CD1d^fl/flCD11c^Cre mice. Data represent two experiments. Scale bar, 100 μm.

Figure 5. CD1d expression on CD11c⁺ cells modulates the IgA repertoire

1. IgA levels in serum and faeces from CD1d^+/− and CD1d^−/− mice, determined by ELISA. Data are from more than five experiments.

2. IGHV family usage (left) and frequency of representative IGHV families (right) for IgA in Peyer's patches from CD1d^+/− and CD1d^−/− mice. Data are from four experiments.

3. IGHV family usage (top) and frequency of most frequent IGHV families (bottom) for IgA in Cre⁻ and Cre⁺ CD1d^fl/flCD11c^Cre mice. Data are from three experiments.

4. IGHV family usage (top) and frequency of representative IGHV families (bottom) for IgA in CD1d^+/− and CD1d^−/− mice from a different animal facility. Data are from two experiments.

5. IGHV family usage (top) and frequency of most frequent IGHV families (bottom) for IgA in CD1d^+/− and CD1d^−/− mice treated with antibiotics for 14 days (left) and 14 days after the treatment (right). Data are from two experiments.

Data information: Each dot in the bar plots is an individual mouse and bars represent mean ± SEM. *P < 0.05, **P < 0.01, two‐tailed unpaired t‐test.

Figure 6. iNKTcells/CD1d regulate intestinal immunity by controlling Treg induction

1. A

   Regulatory T cells (Tregs; CD4⁺Foxp3⁺) and induced Tregs (iTregs; Nrp1⁻) in the mLN from CD1d^−/− and CD1d^+/− mice. Flow cytometry plots (top) and frequency and numbers of Treg and iTreg (bottom) are depicted. Data are from three experiments.

2. B, C

   C57BL/6 WT mice were orally gavaged with αGalCer, and the Treg populations were analysed 3 days later in mLN (B) and SI‐LP (C). Flow cytometry plots (top) and frequency and numbers of Treg and iTreg (bottom) are depicted. Data are from three independent experiments.

3. D, E

   Cre⁺ and Cre⁻ CD1d^fl/flCD11c^Cre mice were orally gavaged with αGalCer, and the Treg populations were analysed 3 days later in mLN. Flow cytometry plots (D) and frequency and numbers of iTregs (E) are depicted. Data are from 2–3 experiments.

4. F

   qPCR analyses for the depicted cytokines in freshly isolated iNKT cells from the mLN of C57BL/6 αGalCer‐treated or control mice. Data are normalized to GAPDH. RE, relative expression. Data are from 3–4 experiments.

5. G

   qPCR analyses for Tgfb1 in the mLN of CD1d^+/+, CD1d^−/− or Cre⁺ CD1d^fl/flCD11c^Cre mice after oral αGalCer administration. Data are from three experiments.

6. H, I

   C57BL/6 WT mice were orally gavaged with αGalCer ± αIL‐4 or αTGF‐β blocking antibodies, and the Treg population was analysed 3 days later in mLN. Flow cytometry plots (H) and frequency of iTregs (I) are depicted. Data are from three experiments.

7. J

   Single‐cell suspensions from the mLN of WT mice were stimulated with PMA/ionomycin for 3 h, and secretion of IFN‐γ and IL‐4 was measured by flow cytometry. Plots show production of IFN‐γ and IL‐4 by mLN cells (left) and αGalCer‐loaded CD1d‐tetramer binding for IL‐4⁺ (middle) or IFN‐γ⁺ (right) cells. Data are representative from three experiments.

8. K

   qPCR analyses for IL4 mRNA in freshly isolated CD11c⁺ cells, B cells and T cells sort‐purified from the mLN of WT αGalCer‐treated or control mice. Data are normalized to HPRT‐1. RE, relative expression. Data are from four experiments.

Data information: Numbers indicate percentage of cells in the indicated gates. Each dot in the bar plots is an individual mouse and bars represent mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, two‐tailed unpaired t‐test.