Glycolipid activation of invariant T cell receptor+ NK T cells is sufficient to induce airway hyperreactivity independent of conventional CD4+ T cells

Abstract

Asthma is an inflammatory lung disease, in which conventional CD4+ T cells producing IL-4/IL-13 appear to play an obligatory pathogenic role. Here we show, in a mouse model of asthma, that activation of pulmonary IL-4/IL-13 producing invariant TCR+ CD1d-restricted natural killer T (NKT) cells is sufficient for the development of airway hyperreactivity (AHR), a cardinal feature of asthma, in the absence of conventional CD4+ T cells and adaptive immunity. Respiratory administration of glycolipid antigens that specifically activate NKT cells (alpha-GalactosylCeramide and a Sphingomonas bacterial glycolipid) rapidly induced AHR and inflammation typically associated with protein allergen administration. Naïve MHC class II-deficient mice, which lack conventional CD4+ T but have NKT cells, showed exaggerated baseline AHR and, when challenged with alpha-GalactosylCeramide, demonstrated even greater AHR. These studies demonstrate an expanded role for NKT cells, in which NKT cells not only produce cytokines that influence adaptive immunity but also function as critical effector cells that can induce AHR. These results suggest that NKT cells responding to glycolipid antigens, as well as conventional CD4+ T cells responding to peptide antigens, may be synergistic in the induction of AHR, although in some cases, each may independently induce AHR.

Fig. 1.

Airway responses to pulmonary iNKT cell activation. (a) BALB/c mice show AHR after i.n. challenge with 1.5 μg of α-GalCer. Increasing concentrations of methacholine were used to assess AHR in mice at 2, 6, and 24 h. Data are the mean ± SEM enhanced pause (Penh) values, representative of three experiments (n ≥ 4). (b) iNKT cell-deficient mice do not develop AHR after α-GalCer i.n. challenge, assessed in the same manner as in a for CD1d^−/− and Jα18^−/− BALB/c. Data are the mean ± SEM Penh, representative of three experiments (n ≥ 4). (c) AHR after i.n. α-GalCer challenge includes the lower respiratory track. Invasive measurement of airway resistance was performed in BALB/c and CD1d^−/− mice at 24 h after 1.5 μg i.n. challenge of α-GalCer, shown as airway resistance (R_L) (Right, in cm H₂O per ml per s) and dynamic compliance dynamic compliance (C_dyn) (Left, in ml per cm H₂O). Data represent the mean ± SEM from two experiments (n = 3). (d) Pulmonary but not systemic iNKT cells were essential for AHR. AHR was assessed as in a for BALB/c mice at 24 h after i.v. versus i.n. challenge with 1.5 μg of α-GalCer. Data are the mean ± SEM Penh, representative of three experiments (n ≥ 4). (e) Airway eosinophilia occurs after pulmonary but not systemic iNKT cells activation. BAL fluid from the mice in d was analyzed at 3 h after airway measurements, shown as the number of cells per ml of BAL fluid. TCC, total cell count; MO, monocyte; EOS, eosinophils; LYM, lymphocyte. (f) Serum IL-4 was elevated after pulmonary but not systemic activation of iNKT cells. IL-4 and IFN-γ were measured by ELISA in the sera of mice in d taken at 26 h after challenge of 1.5 μg of α-GalCer. (g) After i.n. α-GalCer challenge, eosinophilic and lymphocytic peribronchiolar inflammation was apparent in wild-type (Center) but not CD1d^−/− mice (Right). Lung tissues from a were sectioned and stained with hematoxylin/eosin; representative sections are shown (×200).

Fig. 2.

AHR depends upon timing of iNKT cell activation. (a) Pretreatment with α-GalCer exhausts iNKT cell capacity to induce AHR. α-GalCer (2 μg) administered i.n. or i.v. 3 days before a second i.n. dose of α-GalCer reduces AHR. AHR was assessed as in Fig. 1a. Data are the mean ± SEM, representative of two experiments (n = 5). (b) Activation of iNKT cells with α-GalCer before i.n. antigen challenge first enhanced (Left) and then inhibited (Right) allergen-induced AHR. α-GalCer (2–4 μg) was administered 6 days after OVA/alum-sensitized mice, 1 day before 3 consecutive days of i.n. challenge with OVA. AHR was assessed on day 8 before the second i.n. challenge and again on day 10. Data are the mean ± SEM Penh, representative of three experiments (n = 3–5). The α-GalCer- and saline-treated groups have significantly different AHR (P < 0.05; Student’s t test).

Fig. 3.

Analysis of serum Ig levels after activation of pulmonary iNKT cells. After i.n. challenge with 1.5 μg of α-GalCer, (a) total serum IgG1 and IgG2a were unchanged at 24 h as measured by ELISA, representative of four experiments (n ≥ 4). (b) Total serum IgM levels increase 2- to 3-fold at 24 h for mice in a. (c) Total serum IgE levels increase after 1.5 μg of α-GalCer challenge, as measured by ELISA at 6, 12, 24, 36, 48, and 60 h. Data represent the mean ± SEM pooled from three experiments (n = 10).

Fig. 4.

Airway response after Sphingomonas antigen PS-30 activation of pulmonary iNKT cells. (a) Molecular structure of α-GalCer and PS-30. α-GalCer, with an unusual α linkage at the 1′ carbon of the sugar to its sphingosphine base, is a more potent activator of iNKT cells than PS-30, which has a biologically more common β linkage at the 1′ carbon of the sugar linked at its sphingosine base. (b) Wild-type but not iNKT cell-deficient mice develop AHR, assessed as in Fig. 1a at 24 h after i.n. challenge with 10 μg of PS-30 or 1.5 μg of α-GalCer in wild-type BALB/c versus CD1d^−/− mice. Data are the mean ± SEM Penh, representative of three experiments (n ≥ 4). (c) Sphingomonas antigen i.n. challenge induces airway eosinophilia. BAL fluid from mice in a was analyzed at 3 h after airway measurements, shown as in Fig. 1c. (d) Total serum IgE levels, measured by ELISA in the sera of mice in b, increase 2- to 3-fold after Sphingomonas antigen i.n. challenge, which also results in (e) eosinophilic and lymphocytic peribronchiolar inflammation and increased eosinophilic cytoplasm of bronchiolar lining cells (Center) not seen in similarly treated CD1d^−/− mice (Right). Images are representative sections (×200).

Fig. 5.

IL-4 and IL-13 but not B cells or IgE are necessary for the full development of AHR. (a) AHR depends upon IL-4 and IL-13. After i.n. challenge with 1.5 μg of α-GalCer, AHR was assessed as in Fig. 1a and was reduced in IL-4^−/− and IL-13^−/− single knockout and absent in IL-4^−/−IL-13^−/− double knockout BALB/c mice. Data are the mean ± SEM Penh, representative of four experiments (n = 3–5). (b) Total serum IgE, collected from mice in a 1–2 h after AHR measurement, was reduced in IL-4^−/− and absent in IL-13^−/− and IL-4^−/− IL-13^−/− versus wild-type mice (measured by ELISA). Data are the mean ± SEM, representative of four experiments (n = 4–5), two of which examined only IL-13^−/− and IL-4^−/−IL-13^−/−mice. (c) AHR does not require B cells. AHR was assessed as in Fig. 1a for B cell-deficient JHD^−/− versus wild-type BALB/c mice at 24 h after 1.5 μg of α-GalCer i.n. challenge. Data are the mean ± SEM Penh, representative of five experiments (n ≥ 4).

Fig. 6.

IL-5 and eosinophils are not required for AHR. (a) BALB/c mice treated with anti-IL-5 mAb blocking antibody (described in Materials and Methods) and challenged with 1.5 μg of α-GalCer i.n. show >85% reduction in BAL eosinophils at 24–27 h versus isotype control-treated mice (P < 0.05; Student’s t test). Results represent three experiments, shown as in Fig. 1c. (b) Mice depleted of lung eosinophils in a show normal AHR at 24 h, assessed as in Fig. 1a. Data are the mean ± SEM Penh, representative of three experiments (n ≥ 4).

Fig. 7.

Conventional MHC class II-restricted CD4 T cells are not required for AHR, and elevated AHR in MHCIIΔ/Δ mice is CD1d-dependent. (a) MHCIIΔ/Δ mice have more pulmonary iNKT cells. Lungs from MHCIIΔ/Δ mice and wild-type C57BL/6 mice were digested and stained with anti-TCRβ mAb, anti-CD4 mAb, and CD1d-tetramer for evaluation by flow cytometry TCRβ ⁺CD1d tetramer⁺ cell numbers were increased in the lungs of MHCIIΔ/Δ versus wild-type mice. Total cell counts (bar graph) show that iNKT cells are the majority of lung CD4⁺ cells in MHCIIΔ/Δ (69%) versus wild-type (14%) mice, representative of three experiments (n = 3–4). (b) MHCIIΔ/Δ mice had exaggerated representative of three experiments (n = 3–4) resting state and postchallenge AHR, assessed in wild-type C57BL/6 versus MHCIIΔ/Δ at 24 h after i.n. challenge with 2 μg of α-GalCer or vehicle controls. Data are the mean ± SEM Penh, representative of three experiments (n = 4–5). (c) Exaggerated AHR seen in MHCIIΔ/Δ mice after i.n. challenge with 2.0 μg of α-GalCer was assessed as in Fig. 1a in C57BL/6 wild-type versus MHCIIΔ/Δ mice at 24 h. Some MHCIIΔ/Δ mice were treated with anti-CD1d blocking or isotype mAb before challenge. Data are the mean ± SEM Penh, representative of three experiments (n = 4). (b) AHR in anti-CD1d-treated resting mice was reduced to wild-type control levels. AHR was assessed in wild-type versus MHCIIΔ/Δ at 24 h after i.n. challenge with vehicle controls or PBS. Some MHCIIΔ/Δ mice were treated with anti-CD1d mAb (indicated). Data are the mean ± SEM Penh, representative of three experiments (n = 3–4).