Certain glycolipid antigens for natural killer T (NKT) cells can direct the overall cytokine balance of the immune response. However, the molecular mechanism of Th1- or Th2-biased cytokine secretion by NKT cells is still unknown. Previously, we synthesized isoglobotrihexosylceramide (iGb3) analogues by introducing a hydroxyl group at C4 on the ceramide portion of iGb3 to produce 4-HO-iGb3 or to further deoxylation on the terminal galactose to produce 4‴-dh-iGb3. Both modified iGb3, especially 4‴-dh-iGb3, stimulated more IFN-γ production by hepatic NKT cells, and thus elicited preferential Th1 responses. Here, we found that 4‴-dh-iGb3–loaded bone marrow–derived dendritic cells (DC) could significantly inhibit growth of subcutaneous melanoma and suppress lung metastasis in C57BL/6 mice compared with unmodified iGb3-loaded DCs. In investigating the mechanisms of this improved activity, we found that 4‴-dh-iGb3 stimulation increased STAT1 signaling by NKT cells, whereas the phosphorylation of Th2 type cytokine–associated transcription factor STAT6 signaling was not affected. Analysis of the structures of iGb3 and 4‴-dh-iGb3 revealed that 4‴-dh-iGb3 provides greater stability and affinity between glycolipid and CD1d or NKT TCR complex than iGb3. Thus, 4‴-dh-iGb3 can improve the antitumor effects of a DC-based vaccine possibly by stabilizing the CD1d/glycolipid/TCR complex and stimulating IFN-γ signaling of NKT cells. Furthermore, chemical modification of iGb3 can elicit Th1-biased responses by NKT cells, and 4‴-dh-iGb3 combined with a DC vaccine may serve as a potent new NKT-based therapy against tumors and infectious diseases. Mol Cancer Ther; 10(8); 1375–84. ©2011 AACR.
Natural killer T (NKT) cells are specialized immune cells that express NK markers along with a semiinvariant T-cell antigen receptor (TCR) and display unique characteristics of innate rather than adaptive lymphocytes (1). In mice, the TCR of most NKT cells consists of an invariant Vα chain encoded by the Vα14 and Jα18 gene segments paired with TCRβ chains that belong to a restricted set of Vβ families (2,3). Th1 responses are associated with cell-mediated immunity by Th1 cytokines, which tend to produce the proinflammatory responses responsible for killing pathogens and for perpetuating autoimmune responses, whereas Th2 responses are associated with humoral immunity mediated by Th2 cytokines. Two responses cross-regulate each other: a shift in favor of Th2 may lessen autoimmune damage but also may render viral infection; and shifting from Th2 to Th1 might benefit for clearance of infections and prevention of cancer but also might lead to inflammation and tissue damage (4). NKT cells are implicated in the control of autoimmunity, resistance to tumors, and protection against infectious agents through prompt secretion of large amounts of both T helper 1 (Th1) cytokines (IFN-γ) and T helper 2 (Th2) cytokines [interleukin (IL) 4, IL-5 and IL-13] after activation via TCR engagement (5). Furthermore, promotion of IFN-γ–producing NK cells also occurs after NKT cell activation, accompanied with bystander activation of dendritic cells (DC), conventional T cells and B cells to strengthen Th1 responses.
The first ligand identified for NKT cells was α-galactosylceramide (α-GalCer), a glycolipid derived from a marine sponge. It has been reported that α-GalCer exhibits multiple immunotherapy roles in autoimmune diseases, cancer, atherosclerosis, and infectious diseases (6–8). However, the Th1 and Th2 cytokines that can both be produced by NKT cells after α-GalCer treatment antagonize each other's biological actions. Moreover, α-GalCer is a strong NKT agonist that can cause downregulation of TCR expression and NKT function, leading to anergy or unresponsiveness in mice (9). Such effects largely limit the clinical use of α-GalCer. Isoglobotrihexosylceramide (iGb3) is another glycolipid ligand for NKT cells that has been found to be stimulatory for both mouse and human NKT cells and is assumed to be natural antigenic ligand for NKT cells. The most distinctive difference between α-GalCer and iGb3 is that the latter is a β-linked ceramide and displays 3 sugars. In addition, in terms of the preferential usage of Vβ7>Vβ8>Vβ2 chains, α-GalCer does not have such a bias, whereas iGb3 mimics closely the stimulatory properties of endogenous ligands (10,11). However, iGb3 is a very weak agonist compared with α-GalCer. Recently, there have been reports that structure modifications of α-GalCer induce polarization of cytokine production by NKT cells. For example, OCH, a synthetic analogue of α-GalCer, possessing a truncated sphingosine chain, was shown to be Th2 polarizing (12,13); whereas α-C-GalCer, the C-glycoside of α-GalCer, was found to have a Th1-polarizing effect (14).
In our former study, we synthesized iGb3 analogues by introducing a hydroxyl group at C4 on the ceramide portion of iGb3 to produce 4-HO-iGb3, and further deoxidation at the terminal galactose yielded 4‴-dh-iGb3. We found that both modified iGb3 molecules, particularly 4‴-dh-iGb3, stimulated greater IFN-γ production by hepatic NKT cells and thus elicited preferential Th1 responses compared with the unmodified iGb3 (15). Here, we found that bone marrow–derived DCs (BM-DC) loaded with 4‴-dh-iGb3 exerted a significant greater ability to inhibit growth of subcutaneous melanoma and to suppress lung metastasis in C57BL/6 mice compared with the unmodified iGb3-loaded DC vaccine. We further investigated the underlying mechanisms of this improvement in antitumor effect and the Th1 cytokine polarization driven by 4‴-dh-iGb3 by analysis of the crystal structures of 4‴-dh-iGb3 and iGb3 and expression of key transcription factors by NKT cells.
Materials and Methods
Mice and cell lines
C57BL/6 mice were purchased from the Experimental Animal Center of Beijing University (Beijing, China). All animals used in these experiments were between 6 and 10 weeks of age and were maintained under specific pathogen-free conditions. All animal studies were approved by the Institute Animal Care and Use Committee of Shandong University. The Vα14+ mouse CD1d-specific NKT cell hybridoma, N38-2C12 (2C12), was kindly provided by Dr. Mitchell Kronenberg (La Jolla Institute for Allergy and Immunology, San Diego, CA; ref. 16) in 2008. The hybridoma cells were grown in RPMI 1640 medium (Invitrogen) supplemented with 10% heat-inactivated fetal calf serum (Front Biomedicals), 2 mmol/L l-glutamine (LifeTechnologies), and β-mercaptoethanol (1 × 10−5 mol/L, LifeTechnologies) at 37°C in a humidified atmosphere containing 5% CO2 and were tested and authenticated by flow cytometry to determine the TCRβ expression every time before being used in experiments.
The glycolipids iGb3 and 4‴-dh-iGb3 were synthesized as previously described (the chemical structures are shown in Supplementary Fig. S1; refs. 15,17,18). The glycolipids were dissolved in dimethyl sulfoxide at 1 mg/mL, and the working amount of glycolipids was then dissolved in PBS buffer. For in vivo experiments, the mice were intraperitoneally injected with 100 μg/kg of glycolipids or with vehicle. For in vitro experiments, 100 ng/mL iGb3, 4‴-dh-iGb3, or vehicle was added to the medium. Alexa Fluor 488–labeled tetrameric CD1d molecules loaded with α-GalCer (CD1d tetramer) were kindly provided by the NIH Tetramer Facility (Atlanta, GA). The following monoclonal antibodies (mAb) were from BD Pharmingen: PE-Cy5.5-conjugated anti-CD3 mAb, PE-conjugated anti-IFN-γ mAb, APC-conjugated anti-IL-4 mAb, PE-conjugated anti-Stat1 (pY701) mAb, Alexa Fluor 647–conjugated Stat6 (pY641) mAb, and isotype control Abs. To identify the phenotypes of mature BM-DCs, phycoerythrin (PE)-conjugated anti-MHC class II antibody, fluorescein isothiocyanate-conjugated CD11C and CD80 antibodies (eBioscience) were used.
Preparation of glycolipid-pulsed BM-DCs and coculture with NKT cells
BM-DCs were generated by culturing bone marrow–derived adherent cells with 40 ng/mL of recombinant murine granulocyte macrophage colony-stimulating factor (rmGM-CSF) and 20 ng/mL of mIL-4 (PeproTech Inc.) for 6 days as described (19). To mature the DCs, 50 ng/mL of lipopolysaccharide was added in the medium for 16 hours on day 7. The cells that became nonadherent were recovered and contained more than 60% of mature DCs (CD11chigh, CD80high, and MHC class IIhigh) as identified by flow cytometry. Mature DCs were pulsed with glycolipids (100 ng/mL) for 12 hours, and NKT cell hybridomas were added at 1 × 105 cells per well in the presence of glycolipid- or PBS-pulsed DCs (1 × 105 cells per well) for 6, 12, 24, or 48 hours, separately.
The levels of IFN-γ and IL-4 in cell culture supernatants and serum were detected using standard sandwich ELISA kits (BD Pharmingen) according to the manufacturer's instructions.
For cell surface staining, DC cells were harvested, blocked with anti-FcγR mAb, and stained with the labeled mAbs at 4°C for 45 minutes. For intracellular cytokine staining, hepatic and splenic lymphocytes were cultured in RPMI 1640 containing 10% fetal calf serum and treated with monensin (Sigma) for 4 hours to inhibit extracellular secretion of cytokines. After blocking with anti-FcγR mAbs and staining for surface markers at 4°C for 1 hour, the cells were fixed for 30 minutes with 0.4% paraformaldehyde and permeabilized in buffer containing 0.01% saponin (Sigma) and 0.09% NaN3 (Sigma), followed by incubation for 1 hour in the permeabilization buffer with PE- or adenomatous polyposis coli (APC)-conjugated Ab. For intracellular phospho-STAT analysis, NKT cells were treated with glycolipid-pulsed DCs for 10 to 90 minutes at 37°C, fixed with 1.5% formaldehyde, permeabilized by resuspending in ice-cold MeOH, and then incubated with anti-phospho-STAT1, anti-phospho-STAT6, or isotype Ab for 1 hour at 4°C. All stained cells were analyzed using a flow cytometer (FACScalibur), and the data were processed with WinMDI 2.9 software (Scripps Research Institute, La Jolla, CA).
Real-time PCR analysis
Total RNA of splenic lymphocytes was isolated using TRIzol reagent according to the standard manufacturer's guide (Invitrogen). cDNA was generated with random primers using M-MLV reverse transcriptase (Promega) according to the manufacturer's protocol. The primer sequences were as follows: GATA-3 forward: 5′-AGAACCGGCCCCTTATCAA-3′, reverse: AGTTCGCGCAGGATGTCC; T-bet forward: 5′-CAACAACCCCTTTG-CCAAAG-3′, reverse: TCCCCCAAGCAGCCAAAG; and β-actin forward: 5′-AGAGGGAAATCGTGCGTGAC-3′, reverse: CAATAGTGATGACCTGGCCGT. Triplicate 20 μL PCR reactions were carried out using SYBR Green Supermix (BioRad). The levels of mRNA were normalized to that of β-actin.
Hepatic and splenic lymphocytes were harvested and lysed in ice-cold lysis buffer containing a dissolved protease inhibitor tablet (Sigma) for 30 minutes. The whole-cell extracts were mixed in Laemmli loading buffer, boiled for 5 minutes, and then subjected to SDS-PAGE. After transferring to nitrocellulose membrane, immunoblots were carried out with anti-GATA-3 mAb (clone H48), anti-T-bet mAb (clone H-210), or anti-β-actin mAb (Santa Cruz Biotechnology) followed by incubation with horseradish peroxidase–conjugated secondary antibody and visualization by enhanced chemiluminescence system (Pierce). The results were analyzed by using AlphaEaseFC software (version 4.0.0, Alpha Innotech Corporation) with normalization of each band to their corresponding loading control.
The docking model was applied using the Tripos SYBYL 7.0 package (2009) on a Dell Precision 390 workstation. The glycolipids were constructed with the Sybyl/Sketch module and geometry optimized using Powell's method with the Tripos force field with convergence criterion set at 0.05 kcal/(Å mol) and assigned with the Gasteiger–HÜckel method. The docking studies were conducted using a Sybyl/FlexX module, built with glycolipids docked into the active site of mCD1d and mVα14TCRβ, and the residues in a radius of 7.0 in the cocrystal structure (PDBcode:2DQM) were selected.
Tumor challenge and treatment
B16 melanoma cells (1 × 105) were injected into the right flank of C57BL/6 mice subcutaneously or were administered in the tail vein. The mice were then randomly assigned to 3 groups with 3 mice in each group (PBS-loaded DC treatment control, iGb3-loaded DC treatment group, and 4‴-dh-iGb3–loaded DC treatment group). Mature DCs (5 × 105) that had been exposed in vitro to either 500 or 200 ng of iGb3 and 4‴-dh-iGb3 were vaccinated intravenously on days 3 and 5 respectively, and DCs with a corresponding volume of PBS were injected into the control mice. Four weeks later, the mice were sacrificed, and the tumors or lungs were removed. The tumors were weighed, and the tumor volume (V) was determined by measuring the length (l) and the width (w), then calculated by using the formula: V = l × w2/2.
The superior lobe of the right lung from each tumor-bearing mouse was harvested 4 weeks after the intravenous administration of B16 melanoma cells and fixed by intratracheal instillation of 1 mL buffered formalin (10%, pH 7.2) followed by immersion in 10% neutral-buffered formalin for 24 hours. The lobe was then rinsed in water for 12 hours and subjected to routine histologic processing and embedded in paraffin. Serial 5-μm thick crosssections were obtained from paraffin-embedded blocks and stained with hematoxylin and eosin (H&E) using a standard protocol.
Student's t test was used to compare the differences between 2 different groups, and P < 0.05 was considered statistically significant.
4‴-dh-iGb3–loaded DCs protect against melanoma metastases
Treatment with glycolipids has been reported to induce protection against metastatic tumors in mice, and both NKT and NK cells contribute to the resistance (14, 20, 21). As it seemed that glycolipids, especially those that can induce a Th1 cytokine bias, may be beneficial to antitumor immune responses, we tested whether DCs loaded with the chemically modified 4‴-dh-iGb3 could prime strong NKT-mediated antitumor effects in mice challenged with B16 murine melanoma cells. C57BL/6 mice were inoculated subcutaneously with 1 × 105 B16 cells and then injected intravenously with iGb3, 4‴-dh-iGb3, or PBS-loaded DCs on days 3 and 5. The tumor volumes and weights were examined at day 28. As depicted in Fig. 1A, 4‴-dh-iGb3–loaded DCs exerted a significantly greater inhibitory effect and delayed the growth of B16 melanoma compared with the iGb3-loaded DCs. The tumor volume in PBS-loaded DC group increased to 2,200.4 ± 288.9 mm3, whereas it was only 156.3 ± 150.0 mm3 in the 4‴-dh-iGb3-loaded DC group, and 959.6 ± 249.8 mm3 in the iGb3-loaded DC group. Similarly, the tumor weights were 2.6 ± 0.4, 1.8 ± 0.1, and 0.7 ± 0.2 g in the PBS-loaded DC, iGb3-loaded DC, and 4‴-dh-iGb3-loaded DC group, respectively. These results indicated that both iGb3- and 4‴-dh-iGb3–loaded DC vaccine could inhibit and delay the growth of B16 melanoma, with 4‴-dh-iGb3–loaded DC vaccine being more effective.
We further observed the effects of iGb3- and 4‴-dh-iGb3-loaded DCs in the B16 lung metastasis model. Metastases to the lungs were examined 2 weeks after intravenous administration of DCs loaded with either of the 2 glycolipids. We found that 4‴-dh-iGb3–loaded DCs were more effective than iGb3-loaded DCs in reducing metastases (Fig. 1B). H&E staining confirmed that the metastatic nodules in lungs of mice treated with 4‴-dh-iGb3–loaded DCs were fewer than that in iGb3-loaded DCs treatment mice (Fig. 1B). These results indicated 4‴-dh-iGb3 were more potently inhibitory on the growth and metastasis of B16 melanoma than iGb3 by DC-mediated presentation and activation of NKT cells.
4‴-dh-iGb3 promotes STAT1 activation and T-bet expression by NKT cells
To investigate the mechanisms underlying the improvement of the antitumor effects by 4‴-dh-iGb3, we observed the effect of 4‴-dh-iGb3 on cytokine production by NKT hybridoma in vitro. The freshly prepared DCs were loaded with iGb3, 4‴-dh-iGb3, or the vehicle for 12 hours. The glycolipid-loaded DCs were cocultured with NKT cell hybridomas 2C12 for 6, 12, 24, or 48 hours, and supernatants were collected to determine the secreted levels of IFN-γ and IL-4. As shown in Fig. 2A, the 4‴-dh-iGb3–loaded DCs stimulated more IFN-γ secretion by 2C12 hybridomas than the iGb3-loaded group. There were no differences in IL-4 production. Similar results were obtained with 2C2 cells by assaying intracellular IFN-γ expression analyzed by flow cytometry (Fig. 2B). In an in vivo experiment, the murine serum IFN-γ was enhanced after 4‴-dh-iGb3 stimulation (Fig. 2C), although the percentages of NKT cells were not increased in both liver and spleen (Supplementary Fig. S2). These data strongly suggested that 4‴-dh-iGb3 promoted IFN-γ production and thus polarized the Th1 responses, which might be beneficial for antitumor immunity.
During Th1/Th2 differentiation of conventional Th precursor, signals through the TCR and cytokine receptors can lead to the initiation of a Th1 program via STAT1 or STAT4 activation and the induction of T-bet, which promotes Th1 lineage commitment. Signals that favor the activation of STAT6 induce GATA3 leading to Th2 differentiation (22,23). To explore whether these transcription factors are also involved in the production of Th1/Th2 cytokines in NKT cells and whether the changes of these transcription factors contribute to the preferential induction of IFN-γ by 4‴-dh-iGb3, we examined the phosphorylation of STAT1 and STAT6 in 2C12 hybridomas upon stimulation with glycolipid-loaded DCs at various time points by intracellular staining. The results showed that 4-dh-iGb3–pulsed DCs treatment promoted the phosphorylation of STAT1 in NKT cells from 10 minutes, reaching the maximum at 30 minutes and returning to a lower level at 60 and 90 minutes. The iGb3-pulsed DCs treatment only led to a slight increase at 20 and 30 minutes. The capacity of 4‴-dh-iGb3 to promote STAT1 phosphorytion was stronger than that of iGb3, especially at 30 minutes (Fig. 3). Meanwhile, STAT6 was phosphorylated after 20 minutes, peaked at 30 minutes, and then regressed at 60 minutes in both the 4‴-dh-iGb3 and iGb3 treatment groups, with no significant differences between the 2 groups (Fig. 3).
We next explored whether T-bet, a key downstream target of STAT1 signaling, is involved in the mechanism of preferential induction of IFN-γ by 4‴-dh-iGb3. We used real-time PCR and Western blotting to assess the mRNA and protein levels, respectively, of T-bet as well as GATA-3, a downstream target of STAT6, in splenic lymphocytes from glycolipid and PBS-treated mice. There was an obvious increase in T-bet expression between the resting or stimulated groups at 24 hours, and the expression of T-bet was significantly upregulated upon 4‴-dh-iGb3 treatment, compared with iGb3 treatment, both at the mRNA (Fig. 4A) and protein (Fig. 4B) levels. In contrast, GATA3 levels in splenocytes were comparable between the 2 groups (Fig. 4A and B). Similar results were also obtained from hepatic lymphocytes (data not shown). These data suggested that 4‴-dh-iGb3 presented by DCs upregulated the phosphorylation of STAT1 and expression of T-bet in NKT cells and thus resulted in greater IFN-γ secretion.
Stability and affinity of the CD1d/4‴-dh-iGb3/Vα14TCRβ complexes are augmented
It has been reported that the structural variants of glycolipids, especially the differences in the affinity of TCR binding to the glycolipid/CD1d complex and the stability of glycolipid ligands bound to CD1d molecules, may induce systemic polarization of cytokine production by NKT cells (24). To investigate the mechanisms by which 4‴-dh-iGb3 preferentially induces IFN-γ secretion compared with iGb3, we compared the crystal structures of 4‴-dh-iGb3 and iGb3 to analyze the stability and affinity of the CD1d/glycolipid/Vα14TCRβ complex. 4‴-dh-iGb3 was docked into the active sites of mCD1d and Vα14TCRβ using SYBYL 7.0 software. We encountered the same problems as other researchers as we observed that the iGb3 glycolipid ligand was very different in size and structure with the linear trisaccharide head group (Gal1-3-Gal1-4-Glc), and the ligands in CD1d-iGb3 and CD1d-4‴-dh-iGb3 complexes were less well ordered (10). Therefore, the α-1,3–linked galactose of the iGb3 or 4‴-dh-iGb3 was not visible in the crystal structure. Consistent with the previous reports (25,26), the docking model showed that the acyl chain of 4‴-dh-iGb3 could insert to the pocket A′ by forming hydrophobic interactions with this subsite, and the sphingosine chain could insert into the F′ pocket by forming hydrophobic interactions. Meanwhile, the hydrophilic phenyl group close to the 2 chains could interact with the rip compartment of the CD1d binding groove (Fig. 5A). To further understand the binding mode of 4‴-dh-iGb3 with mCD1d in detail, a two-dimensional picture was created using the LIGPLOT program. As shown in Fig. 5B, similar to iGb3, 4‴-dh-iGb3 could form hydrogen bonds with 3 residues of CD1d: Asp80, Asp153, and Thr156. In addition, the sphingosine chain of 4‴-dh-iGb3 could form hydrophobic contacts with the Trp153, Phe120, Val118, He98, Leu84, and Leu143 residues, and the acyl chain could form hydrophobic contacts with Val160, Ser76, Plm702, and Tyr73. In contrast, iGb3 forms hydrophobic contacts with Arg79, Phe70, Cys12, Leu100, Val98, Leu84, Phe77, and Tyr73 as reported by Zajonc and colleagues (10). We propose that the differences in hydrophobic contacts by the sphingosine chain and acyl chain might increase the binding strength of 4‴-dh-iGb3 with CD1d. Although the hydroxy group introduced at C4 in 4‴-dh-iGb3 does not participate in binding with CD1d in the computed docking model, we speculated that the hydroxyl group may strengthen the interaction with Asp80.
For the interaction of the glycolipid and TCR, we only observed the interaction between the head groups of iGb3/4‴-dh-iGb3 and Vα14TCRβ, as the tail parts of the glycolipid ligands were reported not to take part in the binding (10). The docking results showed that the head groups of 4‴-dh-iGb3 could insert into the “notch” of the TCR pocket as with iGb3 (Fig. 6A). 4‴-dh-iGb3 could also form hydrophobic contacts with Asn23, Tyr110, and Lys167 of the TCR. However, the Lys167 from one side of iGb3 is rotated to the other side of 4‴-dh-iGb3 in the two-dimensional image of the structures displayed by LIGPLOT. This change in spatial configuration led to the increase in index of stability and affinity of glycolipids/TCR from 3.3 to 4.9 points, perhaps due to the removal of the 4‴hydroxyl group in the terminal galactose of iGb3 (Fig. 6B). Together, these results indicated that the structural changes of 4‴-dh-iGb3 may increase the stability of binding to CD1d and the affinity to the NKT TCR compared with that of the unmodified iGb3.
The highly conserved CD1d-restricted NKT cells, identified as a bridge between innate and adaptive immune responses, exert potent immune regulatory functions by releasing a variety immunomodulatory cytokines. Up to now, the response of NKT cells has been studied extensively by multiple groups with α-GalCer that has been proven to be a unique type of adjuvant for vaccine development (7). New analogues of α-GalCer are being synthesized to search for new NKT cell agonists that may have superior properties for the treatment of autoimmune and inflammatory diseases. One of these, α-C-GalCer was found to be more potent in helping mice to defend against mouse malaria and B16 melanoma by inducing a more prolonged IL-12 and IFN-γ response (14). Moreover, α-C-GalCer was reported bind more stably to DCs than α-GalCer, and α-C-GalCer–loaded DCs induced higher levels and longer lasting IFN-γ–producing NKT cell responses and more effective adaptive protective T-cell–mediated immunity (21).
iGb3, the first found natural ligand of NKT cells, is also described as a candidate for mediating the development and function of NKT cells in infections, malignancy, and autoimmunity (27). A recent report showed that iGb3-primed DCs exerted a significant NKT cell–mediated antitumor activity in mice challenged with melanoma cells (28), which supported that iGb3 can display antitumor activities similar to α-GalCer. In our previous work, we found that 2 chemical modification of iGb3 analogues, especially 4‴-dh-iGb3, were much more active than iGb3 for increasing IFN-γ production by hepatic NKT cells (15). In the present study, we found that 4‴-dh-iGb3–loaded BM-DCs exerted a more significant inhibitory effect on the growth of subcutaneous B16 melanoma and better suppression of lung metastasis in C57BL/6 mice when comparing with the unmodified iGb3-loaded DC vaccine. We further confirmed the effect of 4‴-dh-iGb3 on Th1-biased cytokine production by hepatic and splenic NKT cells (data not shown) as well as NKT hybridoma cell lines both in vivo and in vitro. We hypothesize that this Th1 bias status of NKT cells driven by 4‴-dh-iGb3 may be beneficial for antitumor therapy.
The differentiation process of Th1/Th2 cells by conventional T cells is critical for their function in immune responses. Th1 differentiation is promoted by IL-12 signaling, which is mediated by STAT1/4, whereas Th2 differentiation is promoted by IL-4 signals by means of STAT6. In addition, the downstream targets of STAT signaling are important for the differentiation. T-bet has been identified as a key transcription factor for the development of Th1 cells and the induction of IFN-γ production, whereas GATA-3 is crucial for Th2 development (29–31). However, the transcriptional control of the development and differentiation of Th1/Th2 cytokines in NKT cells remains poorly understood. There is evidence that STAT6 signaling pathway acts as an important factor for the expression of many Th2 cytokines by NKT cells (32). GATA-3 is also intrinsically critical in regulating the development, survival, activation, and effector function of NKT cells (33). Moreover, T-bet was also identified as a major regulator necessary for the maturation of NKT cells (34). In this study, we found that, compared with iGb3, 4‴-dh-iGb3 presented by DCs more significantly upregulated the phosphorylation of STAT1 in NKT cells, whereas there were no differences in STAT6 phosphorylation between the 2 glycolipid-loaded DC treatment groups. Our data suggest that STAT1/T-bet signaling may be an important pathway to regulate IFN-γ transcription in NKT cells, leading to preferential Th1-biased responses. These findings are consistent with other studies supporting the idea that T-bet might be responsible for the remodeling of the IFN-γ locus in NKT cells (35).
The spatial configurations of glycolipids apparently affect the differentiation and development of NKT cells. Subtle alterations in a glycolipid antigen, especially differences in the affinity of TCR binding to the glycolipid-CD1d complex and the stability of glycolipid ligands bound to CD1d molecules, can alter the immune responses initiated by NKT cells, possibly by impairing cytokine polarization (24). Here, by docking the 2 glycolipids into the active sites of CD1d and Vα14TCRβ, we proposed that 4‴-dh-iGb3, made by introduction of a hydroxyl group at C4 of iGb3 and removing the 4‴ hydroxyl group of the terminal galactose, possibly stabilizes the CD1d-4‴-dh-iGb3 complexes, augments the interaction with TCR of NKT. Unfortunately, introducing a hydroxy group (O71) at C4 of 4‴-dh-iGb3 does not cause it to bind with any residue of CD1d. We presumed that O71 in the sphingosine chain may stabilize the lipid backbone, together with hydrophobic contacts between chains and additional residues of CD1d that may change the affinity of the mCD1d-glycolipid complexes to TCR. In addition, the change in spatial configuration by removal of the 4‴hydroxyl group in the terminal galactose of iGb3 increases the stability and affinity of glycolipids/TCR, and thus eventually promotes IFN-γ secretion of NKT cells. We have attempted but failed to dock the CD1d-4‴-dh-iGb3/TCR complexes, perhaps due to the high complexity of the glycolipid ligands.
In summary, we provided evidence that 4‴-dh-iGb3 can improve the antitumor effects of a DC-based vaccine, possibly by stabilizing the CD1d/glycolipid/TCR complex and stimulating IFN-γ signaling of NKT cells. Our findings highlight that owing to the privilege of preferential induction of Th1 responses, the 4‴-dh-iGb3–loaded DC vaccine or 4‴-dh-iGb3 as an adjuvant may achieve more effective therapy against tumors and infectious diseases when translated into clinical trials.
Disclosure of Potential Conflict of Interest
No potential conflicts of interest were disclosed.
This work was supported by the Natural Science Foundation of China (90713033) and the National 973 Basic Research Program of China (2007CB815803) and the Important National Science & Technology Specific Projects (2008ZX10002-008).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.