NKT cell-TCR expression activates conventional T cells in vivo, but is largely dispensable for mature NKT cell biology

Abstract

Natural killer T (NKT) cell development depends on recognition of self-glycolipids via their semi-invariant Vα14i-TCR. However, to what extent TCR-mediated signals determine identity and function of mature NKT cells remains incompletely understood. To address this issue, we developed a mouse strain allowing conditional Vα14i-TCR expression from within the endogenous Tcrα locus. We demonstrate that naïve T cells are activated upon replacement of their endogenous TCR repertoire with Vα14i-restricted TCRs, but they do not differentiate into NKT cells. On the other hand, induced TCR ablation on mature NKT cells did not affect their lineage identity, homeostasis, or innate rapid cytokine secretion abilities. We therefore propose that peripheral NKT cells become unresponsive to and thus are independent of their autoreactive TCR.

Figure 1. The Vα14i-TCR knock-in mouse produces large numbers of correctly selected, bona fide NKT cells.

(A) Schematic representation of the knock-in transgene. The Vα14 promoter, loxP (triangle)-flanked STOP cassette, and pre-rearranged Vα14i (Vα14-Jα18, red square) sequences were inserted 3′ of Jα1 and 5′ of the first Cα exon (coding exons are highlighted in blue); 4pA = 4 SV40 polyadenylation sites. AH, arms of homology. Eα, enhancer (black oval). (B) Representative proportions of NKT cells and conventional T cells of total lymphocytes in thymus and spleen. Numbers indicate mean percentages ± SD of at least seven age-matched mice per genotype. (C) Representative proportions of splenic CD4+, CD8+, and DN (CD4− CD8−) NKT cells. Numbers indicate mean percentages ± SD of seven mice per genotype. (D) The Vβ repertoires of splenic NKT cells of the indicated genotypes. Bars indicate means and error bars SD of three independent experiments. (E) Representative flow cytometric analysis of the indicated cell-surface proteins on conventional CD4+ T cells and NKT cells. (F) Intracellular flow cytometric staining of PLZF, GATA-3, ROR-γt, and Th-POK in the depicted NKT cells. Numbers indicate means of the median fluorescence intensities (MFIs), normalized to CD4+ tetramer− T cells of CTR animals, or percentage of ROR-γt+ cells among DN NKT cells; calculated from three animals per genotype. Histograms are representative of three independent experiments with eight mice in total. Throughout the figure, NKT cells were gated as tetramer+ TCRβ+, conventional (conv) T cells as tetramer− TCRβ+; CTR, CD4-Cre or Vα14i^StopF/wt.

Figure 2. Premature Vα14i-TCR expression impairs NKT and conventional T cell development.

(A–E) Absolute cell numbers in thymus and spleen of 7–13 mice of the indicated genotypes: NKT cells (A), CD4+ NKT cells (B), DN NKT cells (C), total cells (D), and DN tetramer− T cells (E). Bars indicate medians. *** p<0.001; ** p<0.01; * p<0.05; ns, not significant; one-way ANOVA. (A) Mean cell numbers are depicted below the scatter blot. (F) The Vβ repertoires of splenic NKT cells of the depicted animals. Data for CD4-Cre Vα14i^StopF/wt are the same as shown in Figure 1D. Bars indicate means and error bars SD of 3–4 mice per genotype of 3–4 independent experiments. Throughout the figure, NKT cells were gated as tetramer+ TCRβ+, conventional (conv) T cells as tetramer− TCRβ+; CTR, CD4-Cre or Vα14i^StopF/wt.

Figure 3. NKT cell overproduction affects their maturation and NK cell homeostasis.

(A) Intracellular IL-4, IL-13, IFN-γ, and TNF expression of splenic CD4+ NKT cells isolated from the depicted animals 90 min after αGalCer injection. Cells were stained directly ex vivo without addition of brefeldin or monensin. Black numbers indicate mean percentages ± SD, and red numbers indicate mean total NKT cell counts expressing the respective cytokine. Data are from three animals per genotype; FSC, forward scatter. (B) Representative proportions of stage 1 (CD44^low NK1.1^low), stage 2 (CD44^high NK1.1^low), and stage 3 (CD44^high NK1.1^high) thymic and splenic NKT cells. Numbers indicate mean percentages ± SD of 10 mice per genotype. (C) Flow cytometric analysis of the depicted markers on thymic and splenic, transgenic, and control NKT cells. Bars indicate means and error bars SD calculated from 4–7 mice. (D) Extracellular and intracellular flow cytometric stainings of CD69 and T-bet in the depicted NKT cell subpopulations. Numbers in representative histogram indicate percentage of CD69^high or T-bet+ cells among the indicated NKT cells calculated from eight animals per genotype (CD69) or three animals per genotype (T-bet). Histograms are representative of at least three independent experiments with each at least seven mice in total. (E) Absolute splenic NK cell numbers (NK1.1+ TCRβ− tetramer–) of age-matched 6–12-wk-old animals (7–16 per genotype). Bars indicate medians. *** p<0.001; ns, not significant; one-way ANOVA. Throughout the figure, NKT cells were gated as tetramer+ TCRβ+, conventional (conv) T cells as tetramer− TCRβ+; CTR, CD4-Cre or Vα14i^StopF/wt.

Figure 4. TCR switch on mature conventional T cells.

(A) Genetic set-up of the TCR switch experiment. In Mx-Cre Cα^F/Vα14i^StopF mice, the endogenous TCRα-chains (Vα(x)Jα(y)) are exclusively expressed from the Cα^F allele. Cre-mediated recombination leads to termination of expression from the Cα^F allele, and simultaneous start of expression of the Vα14i-TCRα-chain from the Vα14i^StopF allele. (B) T-cell-deficient mice were reconstituted with NKT cell-depleted splenocytes of the indicated genotypes. After 2 wk, the TCR switch was induced by poly(I:C) injection. Eight weeks later, percentages of tetramer+ and tetramer− T cells (TCRβ+) were analyzed in spleen and liver. Black numbers indicate percentages of total lymphocytes, red numbers absolute cell number calculated from 9–17 animals. (C) Bars indicate means and SD (error bars) of CD4+, CD8+, or DN (CD4− CD8−) cells among tetramer− and tetramer+ T cells, calculated from at least nine mice per genotype. (D, E) The Vβ repertoires of the depicted splenic CD4+ (D) or CD8+ (E) T cell subsets isolated from T-cell-deficient animals that received NKT cell-depleted Mx-Cre Cα^F/Vα14i^StopF splenocytes. Some of these mice were injected with poly(I:C) 2 wk later to induce the TCR switch. Eight weeks after poly(I:C) injection, the Vβ repertoires were analyzed. Data represent means and SD (error bars) of two independent experiments with a total of three mice (tetramer− without poly(I:C) injection) or eight mice (poly(I:C) injected) per T cell population. Vβs typical for glycolipid selection of NKT cells are highlighted in red.

Figure 5. Signs of sterile inflammation in mice harboring TCR-switched T cells.

T-cell-deficient mice were reconstituted with NKT-cell-depleted splenocytes of the indicated genotypes. Spleen weight (A), absolute splenic cell numbers (B–E, G, H), and serum TNF levels (F) of 3–28 mice per genotype were determined 8 wk after poly(I:C) administration where indicated. Bars indicate medians. Red points show six animals with near absence of B cells and dendritic cells. (B) Total splenocytes; (C) Macrophages/monocytes (Mac1+ Gr1^int SiglecF−); (D) Neutrophils (Mac1+ Gr1^high SiglecF−); (E) Erythroblasts (Ter119+); (G) B cells (B220+ TCRβ−); (H) Dendritic cells (CD11c+). *** p<0.001; ** p<0.01; * p<0.05, one-way ANOVA.

Figure 6. TCR-switched tetramer+ T cells display an activated/exhausted phenotype, but no signs of NKT cell differentiation.

T-cell-deficient mice were reconstituted with NKT-cell-depleted splenocytes of the indicated genotypes. The TCR switch was induced by poly(I:C) administration. Eight weeks later, the animals were analyzed. (A) Expression of intracellular IFN-γ or TNF ex vivo 90 min after αGalCer injection of the indicated mice. Data are representative of two independent experiments with two animals each. (B) Representative histograms of flow cytometric analyses. Surface expression of NKG2D and NK1.1 on T cells (TCRβ+) of the indicated surface phenotypes in comparison to NK cells (NKG2D+ TCRβ− CD5− or NK1.1+ TCRβ− CD5−) are shown. Histograms are representative for at least three independent experiments with at least one mouse each. (C–H) Representative histograms of flow cytometric analyses. T cells (TCRβ+) of the indicated surface phenotypes, and of wild-type splenic CD4+ NKT cells, are shown. Numbers in representative histograms indicate means of the median fluorescence intensities (MFIs), normalized to CD4+ tetramer− T cells of animals that received NKT-cell-depleted Mx-Cre Cα^F/wt splenocytes, 8 wk after poly(I:C) injection. Means were calculated from 6–25 mice. Scatter plots display normalized MFI. Bars indicate medians. Dotted lines indicate medians of the median fluorescence intensities of control CD4+ wild-type NKT cells calculated from 2–6 mice. (C, D) Intracellular PLZF (C) and Egr2 (D) expression. (E–H) Extracellular expression of CD69 (E), PD-1 (F), LAG-3 (G), BTLA (H); *** p<0.001; ** p<0.01; * p<0.05; ns, not significant; one-way ANOVA.

Figure 7. TCR signaling is not required for the steady state homeostasis of mature NKT cells.

(A) Percentages of TCRβ− cells of the depicted T cell subsets 2 wk after poly(I:C) injection into Mx-Cre Cα^F/F mice. Bars show means and SD (error bars) of 3–5 mice. (B) Surface TCRβ expression of splenic CD4+ NKT cells (NK1.1+ CD5+ CD62L^low) 2 wk after poly(I:C) injection. Numbers indicate means ± SD of three independent experiments with altogether five mice per genotype. (C, D) Total cell counts of splenic naïve conventional CD4+ T cells (CD5+ CD44^low NK1.1−; C) or of memory/effector-like CD4+ T cells (CD5+ CD44^high NK1.1−; D) from 26 control Cα^F/F (CTR, TCR+) animals as well as from 24 Mx-Cre Cα^F/F animals, all after poly(I:C) injection (TCR+, TCR−). (E) Splenic CD4+ NKT cell numbers from in total 32 control Cα^F/F animals (CTR, TCR+) as well as TCRβ+ and TCRβ− CD4+ NKT cell numbers from in total 27 Mx-Cre Cα^F/F animals, at the indicated time after poly(I:C) injection. (C–E) Half-lives were calculated with GraphPad Prism software using nonlinear regression, one-phase decay analysis. (F) BrdU was administered for 4 wk via the drinking water, starting 2 wk after poly(I:C) injection. Directly afterwards, animals were sacrificed and BrdU incorporation was measured by flow cytometry. Representative blots of 2 Cα^F/F and 4 Mx-Cre Cα^F/F mice are shown. (G) Bar chart showing proportion of cells that incorporated BrdU of the indicated T cell subtypes. Bars show means calculated from 2 Cα^F/F and means and SD (error bars) 4 Mx-Cre Cα^F/F mice. *** p<0.001; * p<0.05; ns, not significant; one-way ANOVA.

Figure 8. The maintenance of NKT lineage identity does not depend on TCR-signals.

Cα^F/F and Mx-Cre Cα^F/F mice were injected with poly(I:C) and analyzed 6 wk later. (A, B) Intracellular expression of Egr2 (A) and PLZF (B) in T cells from the depicted mice. Plots are representative for at least three independent experiments. (C) Flow cytometric analysis of NK1.1 expression on splenic naïve (CD62L^high CD5+), memory/effector-like (CD62L^low CD5+) CD4+ T cells, and CD4+ NKT cells (NK1.1+ CD5+ CD62L^low), with or without TCR expression. Median fluorescence intensity, normalized to NK1.1 expression of NK cells (NK1.1+ TCRβ− CD5−). Bars indicate medians. *** p<0.001; ** p<0.01; * p<0.05; ns, not significant; one-way ANOVA. (D) Flow cytometric expression analysis of extra- and intracellular markers of splenic T cells. Median fluorescence intensities of at least four mice per analyzed protein were normalized to the expression on/in conventional CD4+ T cells (tetramer− TCRβ+) to account for interexperimental variations. Expression of NK1.1, CD122, FasL, and T-bet were normalized to NK cells (NK1.1+ TCRβ− CD5−) and then set to 1 for naïve T cells. Data are shown as heatmap, calculated by Perseus software. Blue letters, significantly reduced on splenic TCR− CD4+ NKT cells in comparison to TCR+ CD4+ NKT cells from Cα^F/F and Mx-Cre Cα^F/F mice; analyzed by one-way ANOVA.

Figure 9. TCR-signals are not required for the innate activation of NKT cells.

(A) Intracellular Egr2 expression of the depicted splenic T cells, 90 min after PBS or αGalCer injection, or 6 h after LPS injection. Plots are representative for at least three independent experiments. (B) Intracellular IFN-γ expression of the depicted cells stained directly ex vivo without addition of brefeldin or monensin. Splenic cells were isolated 90 min after PBS or αGalCer injection, or 6 h after LPS injection. Plots are representative for at least three independent experiments.