Innate and cytokine-driven signals, rather than microbial antigens, dominate in natural killer T cell activation during microbial infection

Abstract

Invariant natural killer T cells (iNKT cells) are critical for host defense against a variety of microbial pathogens. However, the central question of how iNKT cells are activated by microbes has not been fully explained. The example of adaptive MHC-restricted T cells, studies using synthetic pharmacological α-galactosylceramides, and the recent discovery of microbial iNKT cell ligands have all suggested that recognition of foreign lipid antigens is the main driver for iNKT cell activation during infection. However, when we compared the role of microbial antigens versus innate cytokine-driven mechanisms, we found that iNKT cell interferon-γ production after in vitro stimulation or infection with diverse bacteria overwhelmingly depended on toll-like receptor-driven IL-12. Importantly, activation of iNKT cells in vivo during infection with Sphingomonas yanoikuyae or Streptococcus pneumoniae, pathogens which are known to express iNKT cell antigens and which require iNKT cells for effective protection, also predominantly depended on IL-12. Constitutive expression of high levels of IL-12 receptor by iNKT cells enabled instant IL-12-induced STAT4 activation, demonstrating that among T cells, iNKT cells are uniquely equipped for immediate, cytokine-driven activation. These findings reveal that innate and cytokine-driven signals, rather than cognate microbial antigen, dominate in iNKT cell activation during microbial infections.

Figure 1.

Antigen- versus cytokine-driven pathways of iNKT cell activation. (A–C) iNKT cell lines were cultured with WT (filled squares) and CD1d-deficient (A), MyD88-deficient (B), or IL-12p35–deficient (C; all open squares) BM-derived DCs and stimulated with various concentrations of LPS, CpG, or GSL-1 for 16–24 h. Cytokine concentrations in culture supernatants were measured by ELISA. Data are presented as means of duplicate values ± SD and are representative of at least three independent experiments. Bar graphs show percent inhibition of IFN-γ secretion in culture supernatants comparing CD1d-deficient (A), MyD88-deficient (B), or IL-12p35-deficient (C) DCs (filled bars) to WT DCs after stimulation with 2 ng/ml LPS, 2 µg/ml CpG, or 10 µg/ml GSL-1, and data are summarized from four independent experiments (mean ± SD). Open bars in C represent the percentage of IFN-γ secretion in the presence of 10 µg/ml of blocking anti–IL-12 antibodies. Data are summarized from two independent experiments. No significant inhibition was observed with control antibodies (not depicted). (D and E) Experiments showing IL-4 production in response to GSL-1, LPS, or CpG using CD1d- (D) and MyD88- and IL-12–deficient (E) BM DCs were performed as in A–C. IL-4 concentrations in culture supernatants were measured by ELISA. n.a. indicates not applicable because IL-4 levels with WT DCs were very low. Data are presented as means of duplicate values ± SD and are representative of at least three independent experiments.

Figure 2.

Detection of microbial lipid antigens expressed by bacteria. Lipids were extracted from bacteria and analyzed by ESI mass spectrometry. (A–C) [M – H]⁻ adducts of the GSL-1 (GlcAGSL) antigen in S. capsulata, N. aromaticivorans, and S. yanoikuyae. (D) [M + CH3COO]⁻ adducts of the GalDAG (BbGL-II) antigen in B. burgdorferi. (E) [M + Na]⁺ adducts of the GlcDAG and GalGlcDAG antigens in S. pneumoniae. For annotation of sphingosine and acyl chain composition see Table S1. For corresponding LC and MS data of lipid antigens expressed by bacteria see SI Fig. 1.

Figure 3.

Cytokine responses of iNKT cells to diverse bacteria. iNKT cell lines were cultured with WT (filled squares) or CD1d-deficient (open squares) DCs in the presence of heat-inactivated bacteria. (A and B) IFN-γ (A) and IL-4 (B) concentrations were measured in culture supernatants by ELISA after 16–24 h. Data represent means of duplicate values ± SD and are representative of at least three independent experiments.

Figure 4.

Mechanism of iNKT cell activation in response to diverse bacteria (A) iNKT cell lines were cultured with WT or MyD88-deficient DCs in the presence of heat-inactivated bacteria. Data for IFN-γ secretion in culture supernatants are shown as percent inhibition comparing MyD88-deficient to WT DCs after stimulation with 10 bacteria per DC for E. coli (Ec), P. aeruginosa (Pa), and S. typhimurium (St) and with 100 bacteria per DC for S. pneumoniae (Sp), L. monocytogenes (Lm), S. aureus (Sa), B. burgdorferi (Bb), S. capsulata (Sc), N. aromoticivorans (Na), S.yanoikuyae (Sy), and M. tuberculosis (Mt). Data are summarized from four independent experiments (mean ± SD). (B) Concentration of IL-12p40 (ng/ml, filled bars) and IL-12p70 (pg/ml, open bars) measured by ELISA in culture supernatants of DCs stimulated with microbial products or heat-killed bacteria as in Fig. 1 and Fig. 4 A, respectively (mean ± SD). No IL-18 was detected in culture supernatants and IFN-β concentrations were not different between unstimulated and stimulated conditions (not depicted). (C) iNKT cell lines were cultured with WT or IL-12p35–deficient DCs in the presence of stimuli as described in A. Data are shown as percentage of IFN-γ secretion in culture supernatants comparing IL-12–deficient to WT DCs after stimulation (filled bars; mean ± SD). Open bars represent the percentage of IFN-γ secretion in the presence of 10 µg/ml blocking mAb against IL-12. No significant inhibition was observed with isotype-matched control antibodies (not depicted). Data are summarized from two independent experiments. (D) Expression of the early activation marker CD25 by iNKT cells after 24 h of co-culture with WT or IL-12–deficient DCs in the presence of stimuli as in A. GSL-1 was used at 2 µg/ml, LPS at 10 ng/ml, and CpG at 2 µg/ml. (E) IL-4 production by iNKT cells in response to heat-inactivated bacteria using WT (filled squares), MyD88-deficient (open squares), or IL-12–deficient (open circles) DCs. Experiments were performed as in Fig. 1 (D and E). IL-4 concentrations in culture supernatants were measured by ELISA. Data are presented as means of duplicate values ± SD and are representative of at least three independent experiments.

Figure 5.

iNKT cell response to live in vitro infection. DCs were incubated with 2 µg/ml CpG, 10 µg/ml GSL-1, P. aeruginosas, S. typhimurium (both 10 live bacteria per DC), L. monocytogenes, B. burgdorferi, S. capsulata, or N. aromaticivorans (all 100 live bacteria per DC) for 3 h, followed by washing and plating with iNKT cell lines. (A and B) IFN-γ (A) or IL-4 (B) concentrations were measured in culture supernatants by ELISA after 16–24 h. Data are displayed as means of triplicate measurements ± SD and are representative of at least two independent experiments.

Figure 6.

In vivo activation of iNKT cells during S. yanoikuyae infection. (A) CD69 surface staining on CD1d tetramer–positive iNKT cells isolated from the livers of uninfected WT mice (dotted line) or 18 h after i.v. infection with S. yanoikuyae (bold line). (B) IFN-γ secretion by CD1d tetramer–positive liver iNKT cells from uninfected animals (left, bold line) or 18 h after i.v. infection with S. yanoikuyae (right, bold line) in comparison with staining with isotype control antibodies (dotted line). (C and D) Comparison of IFN-γ secretion (C) and CD69 expression (D) by CD1d tetramer–positive liver iNKT cells in WT and IL-12–deficient mice 18 h after i.v. infection with S. yanoikuyae (mean ± SD). (E) WT or IL-12-deficient mice were injected i.v. with 20, 5, or 1.25 µg GSL-1, and IFN-γ secretion by CD1d tetramer–positive liver iNKT cells was determined after 80 min. Means ± SD are shown. Data are pooled from three independent experiments and represent four to six mice per condition. (F) IL-4 secretion by liver iNKT cells 45 min after injection of 2 µg α-GalCer. Similar results were observed with GSL-1 (not depicted). (G) IL-4 secretion by liver iNKT cells from uninfected animals or 18 h after i.v. infection with S. yanoikuyae in comparison with staining with isotype control antibodies. Data are representative of two independent experiments using three to four mice per condition.

Figure 7.

In vivo activation of iNKT cells during S. pneumoniae infection. WT or Jα18-deficient mice were infected intranasally with S. pneumoniae. (A) Survival was recorded daily for 2 wk for WT (filled squares; n = 9) and Jα18^−/− (open squares; n = 9) mice. Results are representative of two independent experiments. (B) The number of CFU was determined in lung tissues of WT (open bars; n = 6) or Jα18^−/− (filled bars; n = 6) on days 3 and day 6 after infection. Data are pooled from two independent experiments (mean ± SD). (C and D) Lymphocytes were isolated from lungs of mice infected intratracheally with S. pneumoniae or from uninfected mice and analyzed by flow cytometry. Staining for surface expression of CD69 (C) or secreted IFN-γ or IL-4 (D) on CD1d tetramer–positive lymphocytes are shown. Data represent means ± SD (n = 6) and are pooled from two independent experiments. (E and F) WT or IL-12p35–deficient mice were infected intratracheally with S. pneumoniae and secretion of IFN-γ (E) or expression of CD69 by iNKT cells (F) was determined as described in C and D on day 2 after infection. Data represent means ± SD for three to four mice per group and one experiment of two similar experiments is shown.

Figure 8.

Expression and function of IL-12 receptor on iNKT cells. (A) IL-12Rβ1 (left) and IL-12Rβ2 (right) mRNA expression in purified splenocyte subsets or iNKT cell subsets purified from spleen or liver was assessed by microarray. Data are representative of two to four independent experiments and are shown as mean ± SD. (B) Flow cytometry staining of CD19⁻ splenocytes with CD1d tetramers and anti-CD3 antibodies (left). The tetramer⁻/CD3⁻ gate contains NK cells, the tetramer⁻/CD3⁺ gate contains MHC-restricted CD4 and CD8 αβT cells and γδT cells, and the tetramer⁺/CD3⁺ gate contains iNKT cells. Surface expression of IL-12Rβ1 was determined by flow cytometry on gated lymphocyte subpopulations. (C) Naive and memory CD4 T cells, NK cells, and iNKT cells were isolated from spleens of WT mice and purified by cell sorting. Cells were stimulated in the presence of 1 ng/ml recombinant IL-12 for 1 h and subsequently stained with antibodies against the phosphorylated form of STAT4 (solid lines) or isotype control antibodies (dotted lines). (D) iNKT cell lines were incubated with WT (filled squares) or CD1d^−/− (open squares) DCs in the presence of various concentrations of recombinant IL-12. Data are presented as means of duplicate values ± SD and are representative of at least three independent experiments. (E) iNKT cell lines were incubated with WT DCs and various concentrations of the microbial lipid antigen GSL-1 in the absence or presence of various concentrations of recombinant IL-12. IFN-γ concentrations were measured in culture supernatants by ELISA after 16–24 h. Data are presented as means of triplicate values ± SD and are representative of two independent experiments.

Figure 9.

Innate and cytokine-driven iNKT cell activation during microbial infection. iNKT cell activation during microbial infection is dominantly driven by innate TLR-mediated signals and IL-12, which is released by DCs after stimulation with microbial products. In addition, TCR-mediated stimulation contributes to iNKT cell activation. However, the TCR-mediated signal alone, provided either by recognition of CD1d-presented self- or microbial antigens, is not sufficient to result in iNKT cell IFN-γ production in the absence of IL-12 stimulation. The constitutive expression of high levels of IL-12 receptor endows iNKT cells with the ability to respond rapidly to cytokine-mediated stimulation and ensures immediate iNKT cell activation in response to virtually any infectious agent that induces the production of IL-12, irrespective of the expression of microbial lipid antigens, thus allowing iNKT cells to overcome their restricted TCR specificity.