Activation of natural killer T cells by alpha-galactosylceramide rapidly induces the full maturation of dendritic cells in vivo and thereby acts as an adjuvant for combined CD4 and CD8 T cell immunity to a coadministered protein

Abstract

The maturation of dendritic cells (DCs) allows these antigen-presenting cells to initiate immunity. We pursued this concept in situ by studying the adjuvant action of alpha-galactosylceramide (alphaGalCer) in mice. A single i.v. injection of glycolipid induced the full maturation of splenic DCs, beginning within 4 h. Maturation was manifest by marked increases in costimulator and major histocompatibility complex class II expression, interferon (IFN)-gamma production, and stimulation of the mixed leukocyte reaction. These changes were not induced directly by alphaGalCer but required natural killer T (NKT) cells acting independently of the MyD88 adaptor protein. To establish that DC maturation was responsible for the adjuvant role of alphaGalCer, mice were given alphaGalCer together with soluble or cell-associated ovalbumin antigen. Th1 type CD4+ and CD8+ T cell responses developed, and the mice became resistant to challenge with ovalbumin-expressing tumor. DCs from mice given ovalbumin plus adjuvant, but not the non-DCs, stimulated ovalbumin-specific proliferative responses and importantly, induced antigen-specific, IFN-gamma producing, CD4+ and CD8+ T cells upon transfer into naive animals. In the latter instance, immune priming did not require further exposure to ovalbumin, alphaGalCer, NKT, or NK cells. Therefore a single dose of alphaGalCer i.v. rapidly stimulates the full maturation of DCs in situ, and this accounts for the induction of combined Th1 CD4+ and CD8+ T cell immunity to a coadministered protein.

Figure 1.

Maturation of DC surface markers with αGalCer and other stimuli in vivo. (A) 24 h after i.v. administration of LPS (25 μg), CpG ODN (10 μg), agonistic anti-CD40 mAb (50 μg), αGalCer (2 μg), or PBS (black tracing) to C57BL/6 mice, low density spleen cells were stained with PE-CD11c and FITC-CD8α (to identify DCs and their CD8⁺ or CD8⁻ subsets) and with biotinylated isotype control, anti-CD40, CD80, CD86, I-A^b, and DEC205/CD205 followed by streptavidin APC. (B and C) As in (A), but the surface markers of CD8⁻ or CD8⁺ DCs were followed at the indicated time points after i.v. administration of the maturation stimulus; data shown are for CD86 expression. (D) Minimal up-regulation of CD40 and CD86 on lymph node DCs (relative to splenic DCs) 24 h after α-GalCer i.v. Dotted line is the isotype control, gray shadow is antigen expression at time 0, and bold black line is CD40 and CD86 stain at 24 h. (E) Inability of αGalCer to induce the maturation of DCs from bone marrow progenitors, in contrast to LPS, as shown for up-regulation of CD86 on CD11c⁺ DCs from day 6–7 in culture.

Figure 2.

DC maturation by αGalCer in vivo requires NKT cells. (A) Maturation, assessed by increased CD86 expression, did not occur in Jα281^−/− mice (lacking NKT cells) exposed to αGalCer (left), but did occur with LPS (right). (B) Maturation of DCs from MyD88 ^−/− mice in response to αGalCer and LPS, and (C) from mice depleted of NK cells by anti-asialoGM1.

Figure 3.

Functional maturation of splenic DCs from αGalCer treated mice. 8 h after i.v. administration of αGalCer, LPS or PBS, as in Fig. 1, spleen DCs were isolated using anti-CD11c–coated magnetic beads (purity >95 ± 2%). Graded numbers of spleen DCs from C57BL/6 mice were irradiated (30 Gy) and added to 2 × 10⁵ allogeneic BALB/c (A) or syngeneic (B) T cells for 3 d in flat bottomed 96-well plates. In parallel, DCs were fixed with paraformaldehyde for 30 min, to block their maturation during the mixed leukocyte reaction, and also irradiated, followed by addition to allogeneic (C) and syngeneic (D) T cells. Proliferative responses were measured by [³H]-thymidine incorporation (Materials and Methods). The results are representative of four independent experiments. In E, the experiment in B was repeated comparing fixed DCs from wild-type and Jα281^−/− NKT deficient mice given αGalCer or vehicle. The results are averages of three independent experiments. (F and G) Mice were given PBS, αGalCer, or LPS i.v. and 8 h later, CD11c⁺ DC enriched and CD11c⁻ DC-depleted, low density spleen cells were isolated with α-CD11c coated magnetic beads. In parallel, to test cytokine production by DC subsets, the low density spleen cells were first lymphocyte depleted with CD5⁺ and CD19⁺ magnetic beads, and then CD8⁺ and CD11c⁺ cells were selected successively with magnetic beads, providing enriched CD8⁺ and CD8⁻ DC subsets respectively. IFN-γ secretion (pg/ml) by each cell subset was measured by ELISA assays after culturing the cells for 48 h without further stimulation in 96 well plates at 2 × 10⁵/well (F), while IL-12 production was measured in cells cultured 72 h in the presence of 10 μg/ml agonistic FGK45.5 anti-CD40 antibody at 3 × 10⁵/well (G and H). All results required pools of three mice and represent averages of three independent experiments (*P < 0.05).

Figure 4.

DCs capturing OVA protein and matured by αGalCer in vivo efficiently stimulate naive CD4⁺ and CD8⁺ T cells in culture. (A) Schematics for the experiments in Figs. 4, 6, and 7. Mice were given αGalCer or vehicle (PBS) together with OVA protein (5 mg/mouse) i.v. or osmotically shocked syngeneic splenocytes loaded with OVA protein as described (reference 21). 4 h later, CD11c⁺ cells or CD11c⁻ cells were isolated and tested for their capacity to stimulate CD8⁺ and CD4⁺ TCR transgenic T cells from OT-I and OT-II mice, respectively. T cell proliferation was measured after [³H]-thymidine at 36–48 h. In some experiments (Fig. 5), OVA-specific T cell responses in spleen were evaluated 1 wk after immunization, while in others (Fig. 7), CD11c⁺ cells were transferred into naive mice to prime T cells in vivo. (B and C) The responses of CD4⁺ (OT-II) and CD8⁺ (OT-I) T cells to CD11c⁺ (top) and CD11c⁻ (bottom) spleen cells from mice given soluble OVA (left panels) or cell-associated OVA (right panels) in the absence or presence of αGalCer; representative of three independent experiments. (D) As in B, but enriched populations of CD21^high marginal zone B cells and CD21^low follicular B cells were isolated by cell sorting and studied for CD1d and CD86 expression, as well as presentation of the injected soluble OVA to OT-I T cells. CD21^high marginal zone B cells are the upper group of data on the right, while CD21^low follicular B cells are the lower group (note much lower values on y-axis for D vs. B and C).

Figure 5.

αGalCer enhances immunity to OVA-bearing DCs in vivo. (A and B) 1 d after transferring 10⁶ CFSE-labeled OT-I cells i.v. into C57BL/6 mice, four groups of mice were given either no treatment, osmotically shocked and OVA loaded TAP^−/− spleen cells, αGalCer (2 μg/mouse), or both OVA loaded spleen cells and αGalCer. 3 d later, spleen cells were monitored for dilution of fluorescence in the CFSE labeled Vα2⁺ OT-I cells in the spleen. IFN-γ production was monitored in tandem after culturing the spleen cells without or with OVA_257–264 (1 μM) for 6 h in the presence of GolgiPlug. Each panel shows the percentage of cytokine producing, CFSE-labeled, Vα2⁺ cells in the spleen. (C and D) At day 3 after immunization as in A and B, the mice were injected s.c. with EG7 tumor cells, which stably express OVA as a secreted protein. Mean tumor sizes at 2 d intervals were plotted on the left, and the distribution of tumor sizes at day 18 in the individual animals shown on the right. The experiments in the bottom panels were similar to those in the top except that mice were immunized with OVA loaded spleen cells and αGalCer but depleted of individual lymphocyte subsets (from 2 d before tumor inoculation to the end of the experiments) with monoclonal antibodies as indicated.

Figure 6.

Induction of strong immunity in naive mice with cell associated OVA together with αGalCer. (A) Priming of naive mice with the combination of 2 × 10⁷ OVA-loaded, TAP^−/− spleen cells and αGalCer (2 μg/mouse) i.v. 7 d later, spleen cells were cultured without or with OVA_257–264 or OVA_323–339 peptides to stimulate OVA specific CD8⁺ and CD4⁺ T cells respectively and T cell production of IFN-γ was measured by intracellular cytokine staining (the frequency of T cells producing IL-4 was <0.04% for all groups). (B) Same as (A), but means are shown for four individual mice (*P < 0.05, **P < 0.01 for the wild-type mice primed with OVA-loaded, TAP^−/− spleen cells and αGalCer vs. the other groups). CD4⁺ and CD8⁺ T cell responses were elicited by OVA_323–339 and OVA_257–264 peptides respectively. (C) 1 wk after immunization in each group, mice were challenged with either 2 × 10⁶ EG7 or EL-4 tumor cells. The data are mean tumor sizes ± SEM of 6 (EG7) or 3–4 (EL4) individual mice in independent experiments. (TAP^−/−/OVA + αGalCer vs. the other groups at day 18; P < 0.005).

Figure 7.

DCs isolated from mice given OVA together with αGalCer are able to prime naive animals. Mice were given αGalCer or vehicle together with either OVA protein (5 mg/mouse) i.v. (A) or with osmotically shocked TAP^−/− splenocytes without or with OVA loading (B). 4 h after injection, 10⁶ CD11c⁺ DCs from each group were transferred to naive syngeneic mice. In the top panels, 7 d after immunization with DCs but no further antigen or αGalCer, spleen cells were cultured without (open bars) or with (closed bars) OVA peptides with GolgiPlug to allow accumulation of intracellular IFN-γ as described in experimental procedures. Intracellular IFN-γ staining was performed with double labeling for CD4 and CD8. In the bottom panels, the DCs were taken from mice given αGalCer together with either OVA protein (A) or OVA loaded splenocytes (B), but the DCs were injected into either rabbit Ig (control)-treated mice or mice lacking NKT and NK cells (see Results) to test priming of CD4⁺ and CD8⁺, IFN-γ producing effector T cells. The data are mean ± SEM of three individual mice in independent experiments. *P < 0.05.