dysregulation of CD1d-restricted type ii natural killer T cells leads to spontaneous development of colitis in mice

Abstract

Background & aims: CD1d-restricted natural killer (NK) T cells are a subset of immunoregulatory T cells that comprise type I (express the semi-invariant T-cell receptor [TCR] and can be detected using the α-galactosylceramide/CD1d tetramer) and type II (express diverse TCRs and cannot be directly identified). Studies in mouse models of inflammatory bowel disease revealed a complex role for type I NKT cells in the development of colitis. Type II NKT cells have been associated with intestinal inflammation in patients with ulcerative colitis.

Methods: To investigate whether dysregulation of type II NKT cells, caused by increased expression of CD1d, can contribute to colitis, we generated transgenic mice that express high levels of CD1d and a TCR from an autoreactive, type II NKT cell (CD1dTg/24αβTg mice).

Results: CD1dTg/24αβTg mice had reduced numbers of 24αβ T cells compared with 24αβTg mice, indicating that negative selection increases among type II NKT cells engaged by abundant self-antigen. The residual 24αβ T cells in CD1dTg/24αβTg mice had an altered surface phenotype and acquired a cytokine profile distinct from that of equivalent cells in 24αβTg mice. Interestingly, CD1dTg/24αβTg mice spontaneously developed colitis; adoptive transfer experiments confirmed that type II NKT cells that develop in the context of increased CD1d expression are pathogenic.

Conclusions: Aberrant type II NKT cell responses directly contribute to intestinal inflammation in mice, indicating the importance of CD1d expression levels in the development and regulation of type II NKT cells.

Figure 1

CD1d overexpression affects 24αβ T cell development in CD1dTg/24αβTg mice. (A) Thymocytes and spenocytes from indicated mice were stained with mAb against Vα3.2 and Vβ9, and analyzed by flow cytometry. Data are representative of 6 experiments. (B and C) Bar graphs depict the mean ± SEM of the percentages (B) and absolute numbers (C) of Vα3.2⁺Vβ9⁺ cells from 24αβTg (n=6) and CD1dTg/24αβTg (n=6) mice.

Figure 2

Increased CD1d expression leads to enhanced negative selection of 24αβ T cell in CD1dTg/24αβTg mice. (A) Thymocytes from indicated mice were stained with mAb against Vα3.2, Vβ9, CD4 and CD8, and analyzed by flow cytometry. Data shown are representative of 3 experiments. (B) Bar graphs depict the number of DN, DP, CD4^SP and CD8^SP Vα3.2⁺Vβ9⁺ cells in the thymus of 24αβTg (n=8) and CD1dTg/24αβTg (n=8) mice. (C) Histograms depict the expression of Nur77 on thymocytes (solid line) compared to isotype control staining (filled histogram). Data shown are representative of 2 experiments.

Figure 3

Altered 24αβ T cell surface phenotype and function observed in CD1dTg/24αβTg mice. (A) Flow cytometric analysis of Vα3.2 and Vβ9 expression on splenocytes from 24αβTg and CD1dTg/24αβTg mice. (B) Expression of NK1.1, CD4, and CD8 on Vα3.2⁺Vβ9⁺ cells in the spleen and MLN of indicated mice. Data shown are representative of 5 experiments. (C) MLN lymphocytes from indicated mice were stimulated with anti-CD3 and intracellularly stained for the indicated cytokines. Data shown are representative of 3 experiments.

Figure 4

CD1dTg/24αβTg mice develop severe spontaneous inflammation of the large intestine. (A) 24αβTg and CD1dTg/24αβTg mice (n=45 each) were monitored from 3 to 24 weeks of age for symptoms of IBD. (B) Gross observation of representative colon and cecum from 24αβTg and CD1dTg/24αβTg mice (n=10; scale bar=1 cm). (C) H&E stained sections of representative colon samples from indicated mice (200x; scale bar=100 μm). (D) Histological score of colon samples from 24αβTg (n=15) and CD1dTg/24αβTg mice (n=21). (E) Colon mononuclear cells from 24αβTg (n=4) and CD1dTg/24αβTg (n=6) mice were stained with various mAbs to identify 24αβ T cells, non-type II NKT cells (T cells expressing the Tg TCR α or β chain alone), B cells, granulocytes (TCRβ-Gr1^hiCD11b⁺) and dendritic cells (TCRβ⁻CD11b⁺CD11c⁺), respectively. Bar graphs depict the mean ± SEM for the proportion of cells within the indicated gate. (F) Colon mononuclear cells from indicated mice were stimulated and intracellularly stained for IFN-γ and IL-17 by Vα3.2⁺Vβ9⁺ cells. Data shown are representative of 3 experiments.

Figure 5

CD1dTg/24αβTg/RAG^−/− mice spontaneously develop IBD. (A) Splenocytes and colon mononuclear cells from indicated mice were stained with mAb to Vα3.2 and Vβ9, and analyzed using flow cytometry. Data are representative of 3 experiments. (B) 24αβTg/RAG^−/− (n=29) and CD1dTg/24αβTg/RAG^−/− mice (n=27) were monitored from 4 to 28 weeks of age for signs of IBD. (C) Dot plots and histograms depict the expression of indicated markers on Vα3.2⁺Vβ9⁺ cells in the spleen from both genotypes of mice. Data shown are representative of 5 experiments. (D) Hepatic leukocytes from 24αβTg/RAG^−/− (n=7) and CD1dTg/24αβTg/RAG^−/− mice (n=8) were stimulated with PMA/ionomycin and intracellularly stained for IFN-γ production by Vα3.2⁺Vβ9⁺ cells. Data shown are representative of 4 experiments.

Figure 6

Dysregulated type II NKT cell responses directly contribute to intestinal inflammation. (A) Sorted T cells from 24αβTg/RAG^−/− or CD1dTg/24αβTg/RAG^−/− mice were adoptively transferred into either CD1⁺RAG^−/− or CD1dTg/RAG^−/− recipient mice. H&E stained sections are representative of colons from recipient mice (100x; Scale bar=100 μm). (B) Histological scores of colon samples from recipient mice in each experimental group.