Dendritic cells (DCs) and invariant natural killer T (iNKT) cells play important roles in linking innate immunity and adaptive immunity. Mature DCs activated by Toll-like receptor (TLR) agonists directly activate iNKT cells and the iNKT ligand α-galactosylceramide (α-Galcer) can induce DC maturation, resulting in enhanced protective immune responses. In the present study, we aimed to boost anti-tumour immunity in a murine colon cancer model by synergizing DCs and iNKT cells using α-Galcer-loaded tumour cells (tumour–Gal) and the TLR9 agonist cytosine-phosphorothioate-guanine (CpG1826). The vaccine strategy was sufficient to inhibit growth of established tumours and prolonged survival of tumour-bearing mice. Importantly, the immunization induced an adaptive memory immune response as the survivors from primary tumour inoculations were resistant to a tumour re-challenge. Furthermore, injection of tumour–Gal with CpG1826 resulted in iNKT cell activation and DC maturation as defined by interferon (IFN)-γ secretion by iNKT, natural killer (NK) cells and interleukin (IL)-12 by DCs. Immunohistochemistry analysis revealed that cluster of differentiation (CD)4+ T-cells and CD8+ T-cells played important roles in anti-tumour immunity. Additionally, the vaccine redirected Th2 (T-helper cell type 2) responses toward Th1 (T-helper cell type 1) responses with increases in IL-2, IFN-γ expression and decreases in IL-4 and IL-5 expression after immunization with tumour–Gal with CpG1826. Taken together, our results demonstrated a novel vaccination by synergizing tumour–Gal and CpG1826 against murine colon cancer, which can be further developed as tumour-specific immunotherapy against human cancer.
Invariant natural killer T (iNKT) cells are a subset of innate-like lymphocytes that recognize CD1d lipid antigens, such as α-galactosylceramide (α-Galcer), a non-mammalian glycosphingolipid. iNKT cells produce large amounts of interferon (IFN)-γ and interleukin (IL)-4 which lead to downstream activation of dendritic cells (DCs), natural killer (NK) cells, B-cells and conventional T-cells [1–3]. Through this unique property, iNKT cells play important roles in modulating different immune responses such as inflammation, tumours and autoimmune diseases.
Soluble α-Galcer was initially used to stimulate iNKT cells in vivo which led to iNKT cell anergy as a result of apoptosis and T-cell receptor (TCR) down-regulation [4,5]. Thereafter, DCs pulsed with peptides, also loaded with α-Galcer, was used as a vaccine . This approach could prevent natural killer T (NKT) cells’ anergy and induce abundant IFN-γ production. Recent studies showed that α-Galcer-loaded tumour cells (tumour–Gal) led to generations of innate immunity and enhanced tumour-specific adaptive immunity . In several studies, α-Galcer-loaded tumour cells effectively inhibited tumour growth and prolonged survival [7,8] and the vaccine partially delayed tumour growth in a therapeutic setting . Furthermore, this strategy induced more enhanced protective immunity than equivalent α-Galcer-loaded DCs , suggesting that tumour cells incorporated with appropriate adjuvants may be efficient sources of tumour antigens.
However, in the absence of pathogen-derived signals, α-Galcer-activated NKT cells promote tolerogenic DC maturation leading to immune tolerance in vivo . The mechanisms for the tolerogenic action are mainly through DC–CD1d signalling and ERK1/2 (extracellular signal-related kinase 1/2) pathway . When synergizing with Toll-like receptor 4 (TLR4) stimulation, iNKT cells triggered pro-inflammatory DC maturation and IL-12 secretion . Multiple pieces of evidence indicate that α-Galcer and TLR agonists (such as TLR4, TLR8 and, TLR9 [12–14]) acting co-operatively could induce DC maturation and T-cell priming.
Activation of TLR9 in conventional DCs has been used for vaccine applications. Through activating TLR9, unmethylated CpG oligodeoxynucleotides (ODNs) can induce DC maturation and drive them to perform subsequent immunostimulatory functions, especially in enhancing co-stimulatory molecules expression and increasing IL-12 and IFN-α production [15,16]. However, it remains unclear whether α-Galcer-loaded tumour cells combined with TLR9 agonist could enhance tumour-specific immune response. Therefore, in the present study, we investigated α-Galcer-loaded MC38 (major histocompatibility complex) combining with CpG1826 [tumour–α-Galcer (T/GC)+TLR9] to synergize iNKT cells and DCs as a therapeutic vaccine against murine-established colon cancers. These data demonstrate that several immune cells could be synergistically activated by this combination to generate potent anti-tumour immunity.
MATERIALS AND METHODS
Mice and cell lines
Specific pathogen-free C57BL/6 female mice at 6–8 weeks of age from the Shanghai Laboratory Animal Centre at the Chinese Academy of Sciences were used in the present study. The mice were maintained under specific pathogen-free conditions and treated in compliance with institutional guidelines. MC38 murine colon carcinoma cell line (H-2b) was generously provided by Dr Steven A. Rosenberg (National Cancer Institute, Bethesda, MD, U.S.A.). The retroviral vector Lentivirus-elongation factor 1-alpha (LV–EF1a)–EGFP/Puro and a Plat-E packaging cell line were purchased from GenePharma Co. After the introduction of full-length cDNA of murine CD1d to LV–EF1a–EGFP/Puro, it was retrovirally transduced into MC38 tumour cells by lipofection and the cells were subsequently sorted based on the expression of GFP by FACSVantage. All cell lines were grown in Dulbecco's modified Eagle's medium (DMEM) containing high glucose (HyClone) and 10% FBS (Gibco).
The iNKT cell ligand α-Galcer was purchased from Avanti Polar Lipids and was solubilized in DMSO according to the manufacturer's instructions. The TLR9 ligand agonist CpG DNA 1826 (5′-TCCATGACGTTC CTGACGTT-3′) was synthesized by Sangon Biotech Co. cytometric bead array (CBA) Mouse Th1/Th2 Cytokine Kit (BD Pharmingen) was used for detecting serum cytokines of IL-2, IL-4, IL-5, tumour necrosis factor (TNF)-α and IFN-γ. The following mouse monoclonal antibodies (mAbs) were purchased from eBioscience: anti-α-Galcer–CD1d complex, anti-CD1d, anti-CD11c, anti-CD86 (B7-2), anti-MHC (major histocompatibility complex) class I (H-2Db), anti-CD154 (CD40 ligand), anti-CD69, anti-NK1.1, anti-CD3, anti-IFN-γ and anti-IL-12/IL-23 p40, Armenian hamster IgG, rat IgG1κ. For flow cytometry of iNKT cells, we used recombinant soluble dimeric mouse CD1d:Ig, anti-mouse IgG1 and anti-mouse TCRβ (BD Pharmingen).
For loading of α-Galcer, CD1dMC38 tumour cells were cultured for 48 h in the presence of 1 μg/ml α-Galcer. These α-Galcer-loaded tumour cells were washed three times before injection. To evaluate the binding efficiency of α-Galcer, we co-cultured CD1dMC38 and α-Galcer in several concentrations (0, 0.2, 0.5 and 1 μg/ml) for different periods of time (24, 36, 48 and 60 h). To isolate mononuclear cells (MNCs), the livers were homogenized into a single suspension and resuspended in a 40%/70% Percoll solution (GE Healthcare) for centrifugation for 20 min at 900 g to float the MNCs [9,17].
Real-time quantitative RT-PCR (reverse transcription polymerase chain reaction)
Total RNA was extracted using an Ultrapure RNA kit (CWBiotech). For cDNA preparation, 1.0 μg of total RNA was reverse-transcribed by PrimeScript® RT Master (TaKaRa) following the manufacturer's instructions. The forward 5′-AAGAAGACTATCCCATTG-3′ and reverse 5′-CAGAATCTCACGACATAT-3′ primers for murine CD1d and the forward 5′-AAAGGGTCATCATCTCTG-3′ and reverse 5′-GCTGTTGTCATACTTCTC-3′ primers for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were purchased from Invitrogen (Life Technologies). CD1d and GAPDH transcripts were quantified by real-time quantitative PCR using SYBR® Green Real-Time PCR Master Mixes (Applied Biosystems, Life Technology) and 7500 Real-Time PCR system (Applied Biosystems, Life Technologies) according to the manufacturer's instructions. For each sample, the mRNA abundance was normalized to the amount of GAPDH.
The serum concentrations of cytokines were measured by a flow CBA using Mouse Th1/Th2 Cytokine CBA kit for mouse Th1/Th2 cytokines (IL-2, IL-4, IL-5, IFN-γ and TNF-α; BD Pharmingen) at 2, 6, 10 and 20 h after administration of vehicle, CpG1826, tumour–Gal, α-Galcer and T/GC+CpG1826. Samples were processed according to manufacturer's protocols and analysed using BD CBA software (version 1.4).
IL-12 secreted by splenocytes was assessed by ELISA. Briefly, spleens was harvested from the animals 24 h after different treatments. Cells isolated from different groups of mice were cultured in complete Roswell Park Memorial Institute (RPMI) 1640 medium containing 10% heat-inactivated FBS in 96-well plates at 5×106 cells/well for 72 h and the supernatants were measured for IL-12 and by ELISA.
Intracellular cytokine staining
The liver MNCs were pre-incubated with 2.4G2 culture medium to block FcγR, washed, incubated with anti-CD1d dimer–Gal followed by anti-mouse IgG1–PE (R-phycoerythrin) and TCRβ–FITC for iNKT cells or NK1.1–PE and CD3-APC (allophycocyanin)–eFluor mAb for NK cells. After the cell surface was labelled with mAbs, cells were permeabilized in Cytofix-Cytoperm Plus (BD Biosciences) and stained with anti-IFN-γ–PE. In some experiments, splenocytes from the animals 2 h after different immunizations were cultured with brefeldin A (BD Biosciences) for 6 h. IL-12 produced by DCs was then determined by cytoplasmic staining with anti-IL-12–PE mAb (eBioscience), followed by cell-surface staining with anti-CD11c–APC mAb (eBioscience).
Flow cytometric staining and analysis
To analyse the expression of various surface markers, the liver MNCs concentrated to approximately 106 cells per 50 ml, stained in FACS staining buffer containing anti-mouse CD16/CD32 mAb (2.4G2) in order to block non-specific binding to Fcγ receptors and incubated for 10 min at 4°C. Then, the staining for iNKT was done using dimer–α-Galcer–PE, anti-TCRβ–FITC, anti-CD69–PerCP–cy5.5 and anti-CD154 (CD40L)-APC (eBioscience).
Freshly isolated splenic DCs were stained using anti-CD11c-APC, anti-MHC class II–PE, anti-CD40–PE, anti-CD80–FITC and anti-CD86–FITC or with respective isotype controls (eBioscience) in a staining buffer (PBS containing 2% FBS) on ice for 30 min in the dark. After staining, the cells were fixed using 2% paraformaldehyde for 1 h, washed twice and suspended in PBS with 2% FBS and 1 mM EDTA. For analysis, a CyAn™ ADP Analyzer (Beckman) and FlowJo (version 7.6) software were used.
Generation of DCs and co-culture of DC and liver MNCs with or without α-Galcer-loaded MC38 tumour cells
Bone-marrow-derived DCs were generated in the presence of GM–CSF (granulocyte-macrophage colony-stimulating factor) as previously described . DCs (1×106 cells/ml) were stimulated with α-Galcer (10 ng/ml) or with CpG1826 (2 μg/ml) for 16 h, extensively washed with PBS and cultured for 48 h with liver MNCs in the presence or absence of α-Galcer-loaded MC38 tumour cells (T/GC) (1×105 DCs+5×105 MNCs+2×105 T/GC per well) in round-bottom 96-well plates in RPMI 1640 medium supplemented with 5% FBS. Co-culture supernatants were collected and IFN-γ, IL-4 and IL-12 concentrations were measured by ELISA. For intracellular FACS staining of iNKT cells and DCs, stimulated DCs were co-cultured with liver MNCs in the presence or absence of T/GC for 16 h and afterwards brefeldin A was added for another 4 h. iNKT cells were labelled with anti-TCR-β mAb and anti-CD1d dimer–Gal, fixed and permeabilized for intracellular IFN-γ production. DCs were labelled with anti-CD11c mAb and IL-12 production was then determined by cytoplasmic staining with anti-IL-12 mAb.
Spleen cell-mediated cytotoxicity
GFP-transfected MC38 cells were used as target cells. The cells were washed twice and plated in a round-bottom 96-well plate with each well containing 1×104 cells. Splenocytes as effector cells were derived from mice, 7 days after the immunizations (three mice in each group). Effector cells were then added to the wells at the indicated effector to target cell ratios (E–T) of 5:1, 10:1, 20:1 and 40:1. After incubation for 6 h at 37°C with 5% CO2, the total cell population was harvested from the plate and collected into polystyrene tubes. Immediately before flow cytometric analysis, 10 μg/ml propidium iodide (PI; Sigma–Aldrich) was added to each of the tubes and all cells were then assessed by FACS analysis. Analysis was on gated GFP-positive cells only. The percentage of cytotoxic activity was calculated using the following formula: percentage specific death=[percentage of total GFP+PI+ (dead) cells–percentage of spontaneous GFP+PI+ (dead) cells]/[100%–percentage of spontaneous GFP+ PI+ (dead) cells] × 100%. The percentage of spontaneous dead cells is measured in the tube containing target cells only .
Immune assays and in vivo tumour studies
Mice were divided into five groups with eight mice per group: (a) PBS; (b) 1 μg of α-Galcer; (c) α-Galcer loaded with CD1dMC38 (1×105; T/GC); (d) CpG1826: 50 μg; and (e) α-Galcer-loaded CD1dMC38 combined with CpG1826 (T/GC+TLR9).
To establish tumour models for assessing anti-tumour efficacy in vivo by tumour–Gal and CPG1826 immunization, mice were inoculated subcutaneously with 1×106 MC38 cells (in 100 μl of PBS) in the right flank. The immunizations above were given twice at day 3 and day 10 after tumour implantation. Perpendicular diameters of individual tumour were monitored every 3 or 4 days and tumour volume was determined as (short diameter)2×long diameter×0.52. Animals were killed when tumour size exceeded 5000 mm3. For tumour growth curves, we used eight mice per group and performed the experiments. For immunohistochemical analysis, animals were humanely killed after 30 days of different immunizations and tumours were harvested preserved in 10% neutral phosphate-buffered formalin.
We followed a previously described analysis  with the following modifications: we used mAbs against CD4 (Clone RM4-5; Santa Cruz Biotechnology) at dilution 1:50 and mAbs against CD8 (Clone 5H10-1, Santa Cruz Biotechnology) at dilution 1:50 for immunohistochemical reactions. The number of CD4/CD8-positive T-cells was determined by counting in ten high power fields of view (×400) in each tumour tissue.
Results are expressed as means ± S.D. The Kaplan–Meier method was used to determine the statistical significance of differences in survival time. We performed the log-rank test using SPSS 16.0. To compare the differences between two groups, Student's t test was used. One-way ANOVA was conducted to compare multiple groups.
Transduction of the murine CD1d gene into MC38 tumour cells and loading α-Galcer into CD1dMC38 cell line
The murine CD1 locus is duplicated and encodes two CD1 molecules designated CD1d1 and CD1d2, which share approximately 90–95% sequence homology at the nucleotide level. CD1d1 mRNA can be detected in colon adenocarcinoma MC38 cells, but not CD1d2 mRNA . After transfecting MC38 cells with CD1d1 and GFP doubly expressing retroviral vectors, the stable CD1d1 high-tumour cell line was then selected by sorting on a FACSVantage instrument. GFP expression in CD1dMC38 cells was observed under a fluorescence microscope, the results showed that almost all the cells were GFP-positive (Figure 1A). We then quantified the expression of CD1d for the transfectant by RT-PCR (Figure 1B) and flow cytometric analysis (Figure 1C). The CD1d mRNA expression on CD1dMC38 cells was much higher than in the MC38 cell line (9698.33±149.54 compared with 1; Figure 1B); whereas the percentage of CD1d molecule-positive cells was 96.87±1.35%.
Analysis of CD1d expression on tumour cell lines
Next, to investigate the time course and concentration kinetics of α-Galcer loaded on to tumour cells, we set different concentrations of α-Galcer and then co-cultured with CD1dMC38 cells for a designated period of time (24, 36, 48 and 60 h). The proportion of tumour cells that stably combined with α-Galcer was tested by PE-conjugated anti-mouse α-Galcer: CD1d using flow cytometry. As shown in Figure 1(D), the binding efficiency of the tumour cells for α-Galcer was time- and dose-dependent. According to this result, we prepared CD1dMC38 tumour cells (T/GC) co-culturing with 1 μg/ml α-Galcer for 48 h in the following experiments.
Anti-tumour effects of α-Galcer-loaded tumour cells combined with TLR9 agonist CpG1826 (T/GC+TLR9)
MC38 tumour cells were used to assess the effectiveness of α-Galcer-loaded tumours combined with CpG1826 as a therapeutic vaccine against established colon cancer in mice. After 48 h of pulsing with α-Galcer, CD1dMC38 cells were irradiated to arrest proliferation of the cells before inoculation. The immunizations (PBS, α-Galcer, T/GC, TLR9, T/GC+TLR9) were given two times at day 3 and day 10 after tumour implantation.
Intravenous vaccine treatment with T/GC+TLR9 was sufficient in delaying growth of MC38 colon cancer tumours for at least 15 days, whereas the tumours were detected at 5 days in the PBS treatment group. Notably, T/GC+TLR9 or single T/GC could protect mice free from tumours (nine and seven mice in the two groups free from tumour respectively; Figure 2A). Next, to detect immune memory response, the mice without tumour were re-challenged with 1×106 MC38 tumour cells. Again, six out of nine mice in group T/GC+TLR9 and two out of seven in group T/GC showed resistance (Figure 2B). Treatment with T/GC+TLR9 and single T/GC can both significantly prolonged survival compared with the other groups (P<0.01); and mice in T/GC+TLR9 had longer survival compared with T/GC (P<0.05; Figure 2C).
Therapeutic efficacy of different immunizations in subcutaneous MC38 colon cancer models
Next, we used immunohistochemical analysis to detect CD4+ and CD8+ tumour-infiltrating lymphocytes (TILs) in MC38 colon cancer. The tumour sections were subjected to immunohistochemical staining using brown substrate for determination of TILs. As shown in Figures 2(D) and 2(E), more CD4+ and CD8+ cells could be detected in tumours of mice treated with T/GC and T/GC+TLR9 compared with other groups (P<0.01). The results indicated CD4+ T-cells and CD8+ T-cells may play important roles in anti-tumour response for T/GC+TLR9.
Th1-biased immune response was induced in mice treated with T/GC+TLR9
Immune response polarization towards a Th1 or Th2 phenotype is crucial for the defence against pathogens, autoimmune disease and tumour. Additionally, Th1-dominant immunity is critically important for development of tumour-specific immunotherapy . Therefore, in the present study, we detected Th1/Th2 cytokine production in serum after mice were treated with different immunizations at indicated time points.
The serum concentrations of IL-2 (Figure 3A), TNF-α (Figure 3B) and IL-4 (Figure 3C) reached a peak 2 h after immunization, whereas IL-4 (Figure 3D) and IFN-γ (Figure 3E) peaked at 6 and 10 h respectively. The results indicated that iNKT cells were activated to secrete large amounts of cytokines at an early time [4,9]. The concentrations of IL-2 (Figure 3A), TNF-α (Figure 3B) and IFN-γ (Figure 3E) in mice treated with T/GC+TLR9 were comparable with that of α-Galcer which is a positive control here (P>0.05). However, the secretion of IL-4 (Figure 3C) and IL-5 (Figure 3D) in the T/GC+TLR9 group was much lower than that in the α-Galcer group (P<0.01, for IL-4 at 2 and 6 h; P<0.01, for IL-5 at all times indicated). The concentration of Th1 cytokines (IL-2, TNF-α and, IFN-γ) was improved in the T/GC+TLR9 group compared with T/GC. This result showed that T/GC+TLR9 were able to promote an immune response polarizing towards a Th1 bias.
Serum Th1 and Th2 cytokine productions in response to T/GC+TLR9
T/GC+TLR9 activated both NK and NKT cells in vivo
To investigate whether iNKT cells and NK cells respond to tumour–Gal+TLR9 in vivo, liver MNCs, a rich source of iNKT cells (Figure 4) from mice with different immunizations, were used. After different treatments, iNKT surface-activation markers and the synthesis of IFN-γ by iNKT and NK cells were tested using flow cytometry. As CD40L and CD69 are early activation markers of iNKT cells [7,23], we evaluated the extent of their up-regulation at 2 h after injection of T/GC+TLR9. As shown in Figures 4(A) and 4(B), the percentage of CD69-positive iNKT cells was 33.56±5.7 and 39.12±3.23 for CD40L (P<0.01 compared with any other groups).
In vivo NK and iNKT cell responses to tumour cells loaded with α-Galcer combined with TLR9 agonist CpG1826
At 2 h, we also used FACS assays to measure IFN-γ production by both iNKT cells and NK cells at the single-cell level and we noted comparable induction of IFN-γ to that seen after injection of α-Galcer alone (Figures 4C and 4D). The result was consistent with the results in Figure 3(E), the secretion of IFN-γ was low at 2 h and came mainly from early activators of iNKT, NK cells. There was no significant difference between T/GC+TLR9 and GC. These data indicated that tumour–Gal was able to stimulate innate iNKT and NK immunity in vivo.
T/GC+TLR9 enhanced co-stimulatory surface marker expression and IL-12 production by splenic DCs
To test whether T/GC+TLR9-induced maturation of DCs in vivo, we detected co-stimulatory surface marker expression and IL-12 production by splenic DCs. After 24 h intravenous (i.v.) administration of T/GC+TLR9, DCs up-regulated several markers consistent with maturation, which included molecules involved in T-cell co-stimulation (CD40, CD86), as well as antigen capture and presentation (MHC class II). As shown in Figure 5(A), the MHC-II and CD86 expression of T/GC+TLR9 paralleled that seen with i.v. injection of α-Galcer, known stimuli for DCs maturation in vivo . For CD40 expression, it was much higher in the T/GC+TLR9 group than in other groups (P<0.01; Figure 5A).
Tumor cells loaded with α-Galcer combined with TLR9 agonist CpG1826 induced maturation of splenic DCs
Maturation stimuli can also prime DCs to produce large amounts of immune-enhancing cytokines such as IL-12. Soluble IL-12 can directly stimulate iNKT cells through the IL-12 receptor even in the absence of α-Galcer . In the present study, we examined IL-12 production by splenic DCs in vivo and in vitro. For the in vivo experiment, IL-12 production was tested by intracellular cytokine staining 24 h after i.v. injection of T/GC+TLR9. As shown in Figure 5(B), the percentage of IL-12-positive CD11c DCs was 28.7±3.8, which was significant compared with other groups (P<0.01). For the in vitro experiment the following was carried out 24 h after i.v. treatment: spleen cells from mice were harvested, cultured in 96-well plates, the supernatants were collected and measured for IL-12 by ELISA 72 h later (Figure 5C). The concentration of IL-12 after i.v. T/GC+TLR9 was 371±46.8 pg/ml (P<0.01; compared with other groups).
To examine whether i.v. injection of T/GC+TLR9 would generate antigen-specific T-cell cytotoxicity, splenocytes were isolated from different groups of mice 7 days after immunization. GFP-expressing MC38 cells were used as target cells when co-cultured with splenocytes for 20 h (Figure 5D). CTLs from mice treated with T/GC+TLR9 showed 25–45% lysis of tumour cells at different E–T, which were much higher than other groups (P<0.01) when at 5:1, 20:1 and 40:1.
Tumour–Gal+TLR9 activated both iNKT cells and DCs in vitro
To further verify whether T/GC+TLR9-induced activation of iNKT cells and maturation of DCs in vitro, bone-marrow-derived DCs were exposed to α-Galcer (10 ng/ml) or with TLR9 agonists (2 μg/ml) for 16 h, after extensive washing, we co-cultured the DCs with liver MNCs in the presence or absence of tumour–Gal for 48 h and measured the supernatants for IFN-γ (Figure 6A), IL-4 (Figure 6B) and IL-12 (Figure 6C) production. The result showed that liver MNCs were activated by both tumour–Gal and DCs stimulated with CpG–ODN (TLR9–DC) to produce both IFN-γ and IL-4 (Figures 6A and 6B). The use of tumour–Gal and TLR9–DC resulted in an obvious increase in IFN-γ (Figure 6A) and IL-12 (Figure 6C) but not the IL-4 (Figure 6B), statistically significant compared with other groups (P<0.01). Notably, although α-Galcer-pulsed DCs (α-GC–DC), used in the present study as a positive control, promoted IL-4 synthesis by liver MNCs cells, DCs stimulated with TLR9 (TLR9–DC) combined with T/GC agonists failed to do so (Figure 6B); this is consistent with previous experiments in vivo (Figure 3C).
In vitro co-cultures of α-Galcer or CpG ODN-activated DCs with primary NKT cells in the presence or absence of α-Galcer-loaded MC38 tumour cells (T/GC)
To further confirm that iNKT cells produce IFN-γ and DCs produce IL-12 in this co-culture setting, intracellular staining was performed. After DC-liver MNCs with or without T/GC culturing for 16 h, cells were labelled with the α-Galcer–CD1d dimer, a probe that exclusively stains iNKT cells, plus an anti-TCRβ mAb. As shown in Figure 6(D), TCRβ+ dimer+ cells (iNKT cells) produced much more IFN-γ intracellularly in response to CpG OD N-stimulated DCs combined with T/GC (TLR9–DC+T/GC). DCs were labelled with anti-CD11c mAb and IL-12 production of DCs by intracellular cytokine staining is shown in Figure 6(E), TLR9–DC combined with T/GC produced more IL-12 than other groups (P<0.01). These data indicate that tumour–Gal combined with TLR9 agonist stimulate innate iNKT cells and DCs in vitro, which is consistent with results in vivo (Figure 5).
The injection of a single α-Galcer is sufficient to activate iNKT cells; however, they become unresponsive over a period of time. Additionally, α-Galcer-loaded DCs were considered since α-Galcer–DCs were able to activate an innate immune response and trigger an adaptive immune response . However, when α-Galcer–DCs were combined with peptides for co-delivery, the antigen-specific T-cell levels were low , which suggested that insufficient antigen was provided. To overcome this problem, Shimizu et al. [7,9] applied tumour cells loaded with α-Galcer as APCs, which successfully induced an antigen-specific T-cell immune response. This strategy is based on the fact that tumour–Gal is killed by activated iNKT cells followed by a release of various tumour antigens. In fact, in a prophylactic vaccine setting, the vaccine induced protection against tumour challenge and B16 lung metastases [7,9,26].
In some cases, exogenous microbial glycolipid antigens are not needed to stimulate iNKT cells through their TCRs. For instance, DCs stimulated by TLR4 can activate iNKT cells . In addition, previous studies showed that DCs activation through TLR7–TLR9 led to iNKT cells stimulation [12,14,28]. Additionally, a previous study  showed that iNKT cells activated by α-Galcer increased expression of several TLR genes. Stimulation with TLR9 agonists and α-Galcer enhanced activation of iNKT cells, which induced maturation of DCs through CD40L interaction with CD40 on the DC. Thus, the previous studies suggested that a combination of α-Galcer with TLR9 signalling pathway could enhance functions of DCs and iNKT cells [12,14].
In the present study, we demonstrated that α-Galcer-loaded tumour cells combined with CpG1826 (T/GC+TLR9) vaccine approach was effective against mouse models of colon cancer. As shown in Figures 2(A) and 2(C), T/GC+TLR9 substantially inhibited the development and outgrowth of colon cancers and significantly prolonged survival of mice. It should be noted that there was no tumour growth in five out of eight mice for T/GC+TLR9, compared with four out of eight mice for T/GC. Next, the mice without tumours were re-challenged at day 40; four mice treated with T/GC+TLR9 did not grow a tumour, whereas only one mouse did not grow a tumour when treated with T/GC (Figure 2B). The result suggested that long-lived adaptive immunity by T/GC+TLR9 was generated sufficiently to protect mice from the same tumour again.
We also noted that single i.v. injection of α-Galcer could inhibit the outgrowth of colon cancers compared with PBS, which was not consistent with the previous reports [4,30]. According to earlier reports, tumour cells with down-regulated CD1d expression can escape from iNKT cell-mediated tumour cell lysis since MC38 cells expressed low amounts of endogenous CD1d mRNA (Figure 1B) and low surface expression (Figure 1C). Additionally, there have been reports that CD1d1 expression could be up-regulated by cytokines, especially IFN-γ . Importantly, a previous study showed that iNKT cells played a direct cytotoxic role in CD1d-positive tumour lysis . iNKT cells were activated when treated with i.v. α-Galcer, which identified CD1d-positive MC38 tumours, thus directly killing them leading to inhibition of tumour growth.
Stimulation of DCs by TLR agonists results in increasing expression of co-stimulatory molecules and secretion of cytokines (such as IL-12 or TNF-α), which leads to enhanced activation of iNKT cells . Activated iNKT cells up-regulate expression of CD69 and CD40L (CD154) and produce IL-4 and IFN-γ . And CD40L–CD40 interaction is important in modulating DCs and iNKT cells in vitro and in vivo [9,33]. In addition, IFN-γ and IL-12 are critical for anti-tumour response after being given an α-Galcer-loaded tumour since IFN-γ and IL-12 expand peripherally from iNKT cells and NK cells . In combination, α-Galcer and CpG1826 act co-operatively in the induction of DCs maturation and iNKT cell priming.
To investigate whether the vaccine strategy (T/GC+TLR9) was able to activate innate immune cells in vivo, we analysed the expression of surface molecules on iNKT cells (CD69, CD40L; Figures 4A and 4B) and DCs (CD40, CD86, MHC-II; Figure 5A), as well as intracellular IFN-γ on iNKT and NK cells (Figures 4C and 4D) and IL-12 on DCs in vivo and in vitro (Figure 5). CD69 and CD40L surface expression rose on iNKT cells at 2 h after immunization, which was consistent with the fact that NKT cells were the main effector cells in the early stage of immune response. To further demonstrate that the vaccine could activate iNKT cells and DCs in vitro, we therefore undertook experiments to co-culture DCs and liver MNCs with α-Galcer-loaded MC38 tumour cells. The results indicate that T/GC combined with TLR9 was able to activate iNKT cells to produce IFN-γ (Figures 6A and 6D) and promote DCs to play a more intense immune response of IL-12 (Figures 6C and 6E). The effect of immune activation by T/GC+TLR9 in vitro was similar with results in vivo (Figure 5). These data suggest that NKT cells activated by T/GC+TLR9 can secrete amounts of IFN-γ, followed by IFN-γ production by NK cells (Figure 7, upper part) and lead to activation of DCs, including IL-12 production. Thereafter, DCs activated by TLR9 agonist in the presence of IFN-γ express high levels of surface molecules and released high amounts of IL-12 (Figure 7, lower part). In turn, IL-12 and CD40 can directly activate NKT cells via corresponding receptors on NKT cells even without α-Galcer (Figure 7)
Proposed pathway of T/GC+TLR9 anti-tumour response
iNKT cells rapidly produce large amounts of Th1 cytokines (IFN-γ, TNF-α, IL-2, etc.) and Th2 cytokines (IL-4, IL-10, IL-13, etc.) soon after activation [32,33]. There are various analogues of α-Galcer which are able to polarize iNKT cells towards a Th1 and/or a Th2-response. For example, OCH was reported to be Th2-polarizing [34,35]. In contrast, α-C-galactosylceramide (α-C-Galcer), had a Th1-polarizing response [36,37]. In general, Th1 cytokines, such as IL-2 and IFN- γ, play important roles in the treatment of tumours and infectious diseases; whereas Th2 cytokines, such as IL-4 and IL-10, are necessary for prevention of auto immune diseases. Indeed, CpG ODNs are widely used in the induction of type I immune response [38,39]. Interaction of CpG ODNs with TLR9 expressed on DCs lead to their maturation and enhances immunostimulatory functions . Furthermore a previous study reported that CpG1826-induced iNKT cell activation and promoted Th1 response . Therefore, it is important to determine whether the combination of T/GC and CpG1826 could have an enhanced Th1 effect in the anti-tumour immune response.
Next, we analysed the Th1 and Th2 cytokine production after immunization with tumour–Gal and CpG1826. Our results were consistent with previous studies, as shown in Figures 3 and 4 [42,43]. When α-Galcer was combined with CpG1826, the concentration of Th1 cytokines in serum, such as IL-2, IFN-γ and TNF-α, were higher than CpG1826 alone, whereas IL-4 and IL-5 were relatively lower than α-Galcer alone. Also, the percentages of IFN-γ-positive cells of iNKT and NK cells were higher in T/GC+TLR9 (Figures 4C and 4D). These findings indicate that Th2 responses could be selectively redirected toward Th1 responses by combining α-Galcer with CpG1826.
TILs composed of several different lymphocyte subsets (including CD4+ and CD8+ T-cells, iNKT cells and NK cells) play significant roles in the host immune response to cancer [44,45]. In the present study, we detected CD4+ and CD8+ T-cells in tumours by immunohistochemical analysis. As seen in Figure 2, CD4+ and CD8+ T-cells increased significantly after i.v. immunization with T/GC+TLR9 and T/GC and they play important roles for anti-tumour immune response (Figure 7). Considerable evidence has accumulated suggesting that tumour- infiltrating CD8+ T-cells were associated with improved clinical outcomes in various types of cancers [46,47]. On the other hand, CD4+ T-cells, which secrete various types of cytokines, are composed of different subsets including CD4+ Th1 cells, CD4+ Th2 cells, Th17 cells and CD4+ CD25+ Treg (the regulatory T cells) cells, which elicit different stimulatory or inhibitory/regulatory effects on anti-tumour immunity [48–50]. For various vaccine strategies, anti-tumour response may depend on different cell types.
A previous study  showed that α-Galcer-loaded mouse B-cell lymphoma cells inhibited the development and outgrowth of tumours; the authors revealed that both innate (iNKT cell, NK cells) and adaptive (CD8+ T-cells) cells were necessary for anti-tumour immunity in this approach. Chung et al.  reported that α-Galcer-loaded A20 lymphoma cells elicited potent anti-tumour immunity and established memory-type immune response to A20 tumours; however, only CD4+ T-cells were required in this vaccine strategy. Shimizu et al.  reported that live tumour cells loaded with α-Galcer induced tumour resistance that mainly depended on innate immunity, including NK and iNKT cells. It is possible that the type of tumour cells affects the type of effector mechanism. Future studies are needed to clarify the exact mechanism of anti-tumour immunity elicited by our vaccine, especially for different CD4+ cell types.
In the present study, α-Galcer-loaded tumour cells combined with CpG1826 was used to treat colon cancer-bearing mice. The results showed that the vaccine suppressed tumour growth and, more importantly, induced long-term memory anti-tumour activity in which iNKT cells, NK cells, CD8+ T-cells and CD4+ T-cells played important roles (Figure 7). We found that T-cell priming can be broken down into two parallel phases: an initial phase of NKT/NK cell activation by T/GC and functional maturation of DCs by CpG1826, followed by synergistic activation through CD40–CD40L and IL12–IL-12L interaction and a second phase in which naive T-cells are cross-primed by endogenous mature DCs through CD28–B7 interaction (Figure 7). The vaccine is mainly a synergy of the iNKT cells with DCs to enhance the anti-tumour effect; however, the specific mechanisms require further studies.
In conclusion, a combination of T/GC and TLR signalling was able to activate several aspects of innate and adaptive immunity including NKT cells, NK cells and DCs, as well as CD4+ T- and CD8+ T-cells. Such extensive immune activation may be critical for exploiting immune system to harness cancer progression and recurrence. The present study is expected to introduce a potent vaccine strategy for tumour therapy and the strategy could be applied to patients with tumours in combination with conventional immunotherapy.
Tiangeng Dong, Tuo Yi, Xinqiang Hong, Mengxuan Yang, Shengli Lin and Wenxiang Li performed the experiments. Xingyuan Xu and Jianwei Hu analyzed the data. Tiangeng Dong and Weixin Niu wrote the paper.
We are grateful to Dr Steven A. Rosenberg for his generous provision of the cell lines.
We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work. There is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the position presented in the manuscript.
This work was supported by the Science and Technology Commission of Shanghai Municipality [grant number 09411960700]; and the Shanghai Municipal Education Commission [grant number 15ZZ003].
cytometric bead array
effector to target cell ratios
invariant natural killer T
natural killer T
tumour necrosis factor
α-Galcer- loaded tumour cells
These authors contributed equally to this study.