The XCL1/XCR1 axis is upregulated in type 1 diabetes and aggravates its pathogenesis

Abstract

Type 1 diabetes (T1D) is precipitated by the autoimmune destruction of the insulin-producing β cells in the pancreatic islets of Langerhans. Chemokines have been identified as major conductors of islet infiltration by autoaggressive leukocytes, including antigen-presenting cells and islet autoantigen-specific T cells. We have previously generated a road map of gene expression in the islet microenvironment during T1D in a mouse model and found that most of the chemokine axes are chronically upregulated during T1D. The XCL1/XCR1 chemokine axis is of particular interest, since XCR1 is exclusively expressed on conventional DCs type 1 (cDC1) that excel by their high capacity for T cell activation. Here, we demonstrate that cDC1-expressing XCR1 are present in and around the islets of patients with T1D and of individuals with islet autoantibody positivity. Furthermore, we show that XCL1 plays an important role in the attraction of highly potent DCs expressing XCR1 to the islets in an inducible mouse model for T1D. XCL1-deficient mice display a diminished infiltration of XCR1+ cDC1 and, subsequently, a reduced magnitude and activity of islet autoantigen-specific T cells, resulting in a profound decrease in T1D incidence. Interference with the XCL1/XCR1 chemokine axis might constitute a novel therapy for T1D.

Keywords: Autoimmunity; Chemokines; Dendritic cells; Immunology; T cells.

Figure 1. XCL1 and XCR1 are upregulated in the islets of RIP-GP mice and NOD mice.

(A) Volcano blots of the gene expression profile of chemokine ligands and their receptors of laser-dissected islets from RIP-GP mice at day 1, 3, 7, 10, 14, and 28 after the LCMV infection in comparison with uninfected mice. XCL1 is highlighted in red and XCR1 in blue. (B) Quantification of the data obtained by gene array (curves) and reverse transcription PCR (RT-PCR) (bars) of laser-dissected islets for the expression of XCL1 and XCR1. Data are shown as mean ± SD. (C) Duplex RNAScope in situ hybridization XCL1 (red) and XCR1 (blue) of pancreas tissue sections from RIP-GP mice at days 0, 7, 10, 14, 28 after infection (n = 3 mice per time point). Representative images are displayed per time. Original magnification, 63× (oil). Scale bar: 20 μm. (D) Duplex RNAScope in situ hybridization for XCL1 (red) and XCR1 (blue) of pancreas tissue sections from NOD mice obtained at different age and disease state (n = 2–3). Mice were considered diabetic with nonfasting BG levels of > 300 mg/dL. Original magnification, 40×. Scale bars: 25 μm. (E) Quantification of XCL1 (left) and XCR1 (right) expression in the islets of RIP-GP mice at day 0, 7, 10, 14, and 28 after infection (n = 3 per time point). Data are shown as mean ± SEM, and significant P values (2-way ANOVA) are indicated. (F) Quantification of XCL1 (left) and XCR1 (right) expression in the islets of NOD mice at different time points (n = 2–3 per time point). Data are shown as mean ± SEM.

Figure 2. XCR1⁺ cDC1 are present in the islets of patients with T1D and individuals with islet autoantibodies.

Human pancreas slides were obtained through the HPAP program. Groups of individuals were divided as follows: nondiabetic (ND), nondiabetic with known T1D familiarity (T1D familiarity), autoantibody positive (Aab⁺), and patients with diabetes (T1D). (A) Representative pictures of duplex RNAScope in situ hybridization for XCL1 (red) and XCR1 (blue). In the left column, the original picture with a magnified example of a positive cell (square). On the right, XCR1-expressing cells are highlighted with a light blue dot, and XCL1-expressing cells are highlighted with a red dot. Islets are indicated with black lines. Original magnification, 63×. Scale bars: 20 μm. (B) Number of XCR1-expressing cells per islet microenvironment. Number of islets per section analyzed was 6–63. Each islet is represented by 1 dot. Data are shown as the mean ± SEM number of XCR1-expressing cell in one individual. (C) Mean ± SEM of XCR1-expressing cells per islet microenvironment in the groups of ND, T1D familiarity, Aab⁺, and T1D. Number of islets per section analyzed was 6–63. The mean for every individual is represented by dots.

Figure 3. Fewer CD103⁺ cells are present in islets of XCL1-deficient mice.

(A) Duplex RNAScope in situ hybridization for XCL1 (red) and XCR1 (blue) of pancreas tissue sections obtained from RIP-GP × XCL1^–/– mice at different times after infection, demonstrating that XCL1-deficient mice do not express XCL1. Original magnification, 63× (oil). Scale bars: 20 μm. (B) IHC staining for CD103 of quick-frozen pancreas sections at different times after LCMV infection, comparing RIP-GP with RIP-GP × XCL1^–/– mice. Representative pictures are shown. Original magnification, 40×. Scale bars: 25 μm. (C) Immunofluorescence double-staining for CD11c (green) and CD103 (red) of quick-frozen pancreas sections obtained at different times after LCMV infection, comparing RIP-GP with RIP-GP × XCL1^–/– mice. Nuclei are stained with DAPI (blue). White arrows indicate the double-positive cells (cDC1). Representative pictures are shown. Original magnification, 40×. Scale bars: 20 μm. (D) Quantification of the CD103 staining in RIP-GP (light gray) and RIP-GP × XCL1^–/– (dark gray) mice shown in B, expressed as a percentage of positive CD103 cell area per islet area. Numbers of mice per group are indicated. Data are shown as mean ± SEM. (E) Quantification of the CD103/CD11c double-staining in RIP-GP (light gray) and RIP-GP × XCL1^–/– (dark gray) mice shown in C, expressed as a percentage of CD103/CD11c–double-positive cell area per CD11c⁺ cell area. Numbers of mice per group are indicated. Data are shown as mean ± SEM, and significant P values (Mann-Whitney t test) are indicated.

Figure 4. Numbers of DC and T cells are reduced in the islets but accumulate in the lymph nodes of XCL1-deficient mice.

Quantification of different cell subtypes per organ (spleen, pancreatic draining lymph nodes, and islets) obtained via flow cytometric analysis at day 7 and day 28 after infection, comparing RIP-GP mice with RIP-GP × XCL1^–/– mice. (A) Representative dot plots of the most substantial changes in the islet-infiltrating cells, comparing RIP-GP mice with RIP-GP × XCL1^–/– mice. Selected populations are CD103⁺XCR1⁺CD11c⁺ cells (left panel) and islet autoantigen–specific CD8 T cells (right panel) at day 7 and at day 28. (B) Quantification of the total number of CD11c⁺MHCII⁺ and CD103⁺XCR1⁺CD11c⁺ cells per organ as indicated. Values are displayed as mean ± SEM, and significant P values (Mann-Whitney t test) are indicated (n = 6–9). (C) Quantification of the total number of CD8 T cells and islet autoantigen–specific CD8 T cells per organ as indicated. Islet autoantigen–specific CD8 T cells have been identified by intracellular cytokine staining for IFN-γ after stimulation with the immunodominant LCMV-GP epitope GP33. Values are displayed as mean ± SEM, and significant P values (Mann-Whitney t test) are indicated (n = 6–9). (D) Organ-specific redistribution of cDC1 (left panel) and islet autoantigen–specific CD8 T cells (right panel) in the presence or absence of XCL1. Note that without XCL1, cDC1 and islet autoantigen–specific CD8 T cells are partially remained in the PDLN.

Figure 5. Switch to a Treg milieu in the islets of XCL1-deficient mice.

(A) Frequencies of CD8 T cells expressing perforin (Perf), granzyme B (GrB), PD-1, or KLRG1 of total CD8 T cells or LCMV-GP33–specific CD8 T cells. Data were obtained via flow cytometric analysis of islet-infiltrating cells of RIP-GP mice and RIP-GP × XCL1^–/– mice at day 7 and day 28 after infection. Results are shown as mean ± SEM, and P values (Mann-Whitney t test) are indicated when significant (n = 7–9). (B) In vivo cytotoxicity assay, comparing RIP-GP uninfected mice (d0) to RIP-GP and RIP-GP × XCL1^–/– mice at day 28 after infection. Differently labeled GP33-loaded and unloaded target splenocytes were injected i.v. at a 1:1 ratio. At 10 minutes and 1, 4, 24, and 48 hours after injection, blood was taken, and the ratio of GP33-loaded and unloaded target cells was determined by flow cytometry. The obtained data were normalized against uninfected mice (baseline). (C) Calculated half-life of GP33-loaded target cell turnover (left) and different visualization of the GP33⁺/GP33^– ratio at 48 hours after the i.v. injection for the infected RIP-GP and RIP-GP × XCL1^–/– at day 28 after infection. Values are shown as mean ± SEM. Number of mice used are displayed in brackets. (D) Frequencies of FoxP3⁺ cells among CD8⁺ cells (left) and CD4⁺ cells (right) obtained via flow cytometric analysis of islet-infiltrating cells of RIP-GP and RIP-GP × XCL1^–/– mice at day 7 and day 28 after infection. Results are displayed as mean ± SEM. (E) Ratio of total FoxP3⁺ cells and total autoaggressive (IFN-γ⁺) CD8 T cells. Results are displayed as mean ± SEM. Number of mice and significant P values (Mann-Whitney t test) are indicated.

Figure 6. XCL1-deficient mice show less T1D incidence.

(A) T1D incidence study comparing RIP-GP and RIP-GP × XCL1^–/– mice. Left panel: Percentage of diabetic mice at each time point after infection. Mice with nonfasting blood glucose (BG) levels of > 300 mg/dL were considered diabetic. Note that some mice reverted from a diabetic to a nondiabetic state over time. Right panel: Mean BG levels over time. Significant differences (2-way ANOVA) and the number of mice are indicated. (B) IHC staining of insulin in quick-frozen pancreas sections of RIP-GP and RIP-GP × XCL1^–/– mice. Representative images are shown for days 7, 14, and 28 as well as week 12 after infection. Original magnification, 40×. Scale bars: 25 μm. Note that, for RIP-GP mice, it was not possible to acquire an image at week 12 (not done; n.d.), since all the mice had to be sacrificed earlier due to severe T1D. (C) Mean insulitis scores determined from insulin staining shown in B. Islets were scored as lined out in Methods. Insulitis in RIP-GP and RIP-GP × XCL1^–/– mice was compared at days 7, 14, and 28 and at week 12 after infection. Number of mice are displayed in brackets.

Figure 7. Three-dimensional analysis of total β cell content shows more functional islets in XCL1-deficient than regular RIP-GP mice.

Total β cell content was determined by staining of the entire pancreas with anti-insulin antibody, followed by clearing of the tissue and scanning of the transparent pancreas by light sheet fluorescence microscopy. (A) Comparison between uninfected RIP-GP and LCMV-infected RIP-GP and RIP-GP × XCL1^–/– mice at week 8–12 after infection. Representative images are shown. Scale bars: 200 μm. The insulin-producing β cell content (red) of the representative pancreas tissues shown are indicated as a percentage of the total pancreas volume (green autofluorescence) (Supplemental Videos 1–3). (B) Quantification of the volume of the insulin producing cells per total pancreas volume, expressed as a percentage. Data obtained from diabetic mice are displayed in red. Values are shown as mean ± SEM. P values (2-way ANOVA) and numbers of mice are indicated. (C) Islet volume analysis: islets were separated in 4 groups according to their volume. Note that 8–12 weeks after infection, XCL1-deficient mice have larger and more insulin-producing islets left than regular RIP-GP mice. Strikingly, in comparison with uninfected RIP-GP and infected XCL1-deficient RIP-GP mice, infected regular RIP-GP mice showed a massive reduction of large (>500 μm³) islets. Values are shown as mean ± SEM. P values (2-way ANOVA) and numbers of mice are indicated.