NKT Cell–Driven Enhancement of Antitumor Immunity Induced by Clec9a-Targeted Tailorable Nanoemulsion

Invariant natural killer T (iNKT) cells are a subset of preactivated innate immune cells that possess markers of both NK and T cells and play a role in cancer immunity. The activation of iNKT cells requires T-cell receptor recognition of processed glycolipids presented on an MHC class I–like molecule, CD1d. Conventional dendritic cells (cDC) are specialized antigen-presenting cells (APC) that prime the adaptive immune system. However, within this population of cells, CD8α⁺ cDCs are the dominant APC for cross-priming of antigen-specific T cells (1) and also glycolipid presentation on CD1d for the activation of iNKT cells (2). cDCs possess the appropriate machinery for capturing, processing, and presenting glycolipid antigens for the activation of iNKT cells (3). In turn, activation of iNKT cells provides helper signals via CD40–CD40L interactions (4) and chemokine signals (5) that augment DC maturation, production of IL12 and consequently induction of innate immunity, including rapid activation of NK cells, and promotion of adaptive T-cell responses.

The discovery of an NKT-cell glycolipid antigen α-galactosylceramide (αGC) more than 20 years ago has led to a better understanding of the immune adjuvant role of iNKT cells in augmenting simultaneous innate and tumor-specific adaptive immunity (transactivation). However, the use of soluble αGC as a therapeutic has its limitations. Hyporesponsiveness of iNKT cells upon secondary restimulation (6), acute liver toxicity (7), as well as the variability of iNKT cell numbers between individuals hinder clinical utility of αGC. A variety of methods aiming to overcome these limitations and optimize the adjuvanting effects of iNKT cells have been explored over the past decade. One such approach involves the adoptive transfer of αGC-loaded autologous DCs, which can overcome iNKT cell hyporesponsiveness (8–10) and promote lymphocyte infiltration into tumors (8). However, cellular-based vaccines can be costly and labor-intensive to produce and are usually specific to an individual patient (11). Therefore, in vivo targeting of DCs using inert delivery vectors has been explored to drive tumor-specific responses.

Various studies have utilized vectorized αGC and antigen in various passive and active applications (12). Few have explored active and specific targeting strategies delivering αGC and antigen concurrently toward the CD8α⁺ DC subset. Few endocytic receptors have shown specificity for the CD8α⁺ DC subset. However, an endocytic C type lectin receptor known as Clec9a is highly expressed by CD8α⁺ CD103⁺ DCs and to a lesser extent by plasmacytoid DCs (pDC) in mice (13–15). In humans, Clec9a is expressed by the equivalent CD141⁺BDCA3⁺ DC population (16). DC Clec9a regulates T-cell cross-priming, which involves recruitment of early endosomal components and enzymes colocalized with antigens for cross-presentation (17). In the absence of licensing or danger signals, delivery of a recombinant monoclonal antibody (mAb) to Clec9a does not appear to drive cross-priming of antigen-specific CD8⁺ T-cell responses in vivo (15, 18). Anti-CD40 (14, 19) and agonists forTLR3 (19) and TLR9 (20) are necessary for the induction of antigen-specific CTL. Thus, targeting the Clec9a receptor to regulate T-cell responses became attractive for translational approaches in humans. However, only one study has demonstrated the feasibility of codelivering αGC and antigens to CD8α⁺ DCs in nanoparticle systems via the Clec9a receptor. In that study, a Clec9a-decorated anionic poly(lactic-coglycolic acid; PLGA) nanoparticle system generated iNKT cell–driven CD8⁺ T-cell responses against tumor (21) through the simultaneous delivery of αGC and antigen to the same DC (22).

We reported that cationic Clec9a antigen–targeted nanoemulsions (TNE) carrying whole antigen can promote cross-priming of antigen-specific CD8⁺ T cells through CD4⁺ T cell–mediated induction of IFNα and CD40 signaling, resulting in tumor suppression (23). Using this technology, we sought to investigate the feasibility of this “oil-in-water” nanoemulsion system as a safe delivery vector for the simultaneous delivery of lipophilic adjuvants, such as αGC, with tumor antigens for cancer immunotherapy.

C57BL/6 and B6.SJL (CD45.1) mice were purchased from the Animal Research Centre (Perth, WA). iNKT cell–deficient Jα18 knockout mice were bred and maintained onsite at the Translational Research Institute Biological Research Facility. Mice were used at the ages of 6–12 weeks, sex-matched, and housed under specific pathogen-free conditions. All experiments were conducted following the animal ethics guidelines provided by the National Health and Medical Research Council of Australia and approved by the University of Queensland–Health Sciences Animal Ethics Committee (ethics number 301/15).

The following peptides were custom synthesized by GL Biochem to a final purity of >95%: AM1 (Ac-MKQLADSLHQLARQVSRLEHA-NH₂), OVA_257–264 peptide (SIINFEKL), HPV.E7 peptide (RAHYNIVTF, gf001 peptide), and WH peptide (WPRFHSSVFHTHGGGK; ref. 20). Peptide concentration was determined by using HPLC analysis. Miglyol 812 (Cremer Oleochemicals) was a gift from IMCD Australia Limited. Cithrol GMO HP was a gift from Croda Europe Ltd. An alternative surfactant Sorbitan monooleate (Span80) was purchased from Sigma-Aldrich. Reagent-grade 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), zinc chloride (ZnCl₂), n-hexane, and dimethyl sulfoxide (DMSO) were purchased from Sigma-Aldrich. Endotoxin-free albumin from chicken egg white (OVA; 98% purity, <1 EU/mg) was purchased from Hyglos (Bernried/Germany). mPEG-NHS (MW 5,000, protein dispersibility index <1.08, purity >95%) was purchased from Nanocs. RPMI-1640, DMEM, and fetal calf serum (FCS) were purchased from GIBCO. α-Galactosylceramide (αGC) was purchased from Avanti Polar Lipids. CellTrace Violet (CTV), CellTrace CFSE, LIVE/DEAD Fixable Aqua Dead Cell Stain was purchased from Molecular Probes. Collagenase D and DNase I were purchased from Roche. Amicon Ultra-0.5 Centrifugal Filter Units (MWCO 3000) were purchased from Merck Millipore. DAMP4 fused with antibody (mAb-DAMP4) was generated as previously described (13).

Fluorochrome-conjugated mouse mAbs to murine CD3ϵ (145-2C11), CD8α (53-6.7), CD8β (YTS156.7.7), CD19 (1D3 or 6D5), CD4 (RM4-5), CD40 (3/23), CD44 (IM7), CD45.1 (A20), CD45.2 (104), CD69 (H1.2F3), CD80 (16-10A1), CD86 (GL-1), IFNγ (XMG1.2), NK1.1 (PK136), TCRβ (H57-597), CD11c (HL3), I-A/I-E (M5/114.15.2), TNFα (MP6-XT77), and associated isotype control antibodies were purchased from BioLegend, BD Biosciences, or eBioscience. GolgiPlug was purchased from BD Biosciences. Flow-count fluorospheres were purchased from Beckman Coulter. αGC-loaded CD1d tetramer was generously provided by Prof. D. Godfrey (University of Melbourne, Australia). DAMP4 fused with anti-Clec9a and isotype control (mAb-DAMP4) was kindly provided by Dr. I. Caminschi and A/Prof. M. Lahoud at the Burnet Institute (Melbourne, Australia), with DAMP4 previously described as an interfacial anchor on TNE (24).

Single-cell suspensions were prepared from tissues by mechanical dissociation or from blood followed by red blood lysis for 1 or 15 minutes at room temperature, respectively. For analyses of DC subsets in spleens, whole spleens were incubated with 2 mg/mL of collagenase D and 20 μg/mL of DNAse I for 15 minutes at 37°C before mechanical dissociation. Cells were antibody labeled at predetermined optimal concentrations of antibodies for 45 minutes at 4°C in PBS containing 2% FCS and 2 mmol/L EDTA. Flow-count fluorospheres were added to the samples to calculate cell numbers upon acquisition. Intracellular cytokine staining was preceded by the addition of GolgiPlug to the cells for 4 hours to prevent cytokine release from the Golgi/ER complex, unless otherwise stated. Cells were permeabilized and fixed using BD Cytofix/Cytoperm kit, following the manufacturer's instructions. Labeled cells were acquired on Gallios (Beckman Coulter).

Ovalbumin (OVA), SIINFEKL, and RAHYNIVTF (hereafter gf001) peptide solutions (10 mg/mL), and TNEs were prepared as previously described (23, 24). Briefly, endotoxin-free OVA or SIINFEKL/RAHYNIVTF peptide were dissolved in ultrapure water (10 mL) and sonicated 2:1 with Cithrol GMO HP or Span80 solution (1%, w/v) in hexane solution before lyophilization. The antigen-Cithrol GMO HP pellet was dissolved in Miglyol 812 to 5 mg/mL and used as oil phase. Clec9a-mAb or Ig decorated TNEs were prepared as previously described (24). To make αGC-containing TNE, 2 μL of αGC was added to a final concentration of 10 μg/mL and 20 μL of Miglyol 812 was added to an oil volume fraction of 2% (v/v). For TNEs containing OVA/SIINFEKL, 20 μL of protein/peptide in oil was added instead. WH peptide–functionalized targeted emulsions were made in a similar fashion. Briefly, WH peptide was first conjugated to DSPE-PEG-NHS. WH-PEG conjugate was first synthesized by mixing WH peptide with DSPE-PEG-NHS in 25 mmol/L HEPES solution, pH8.0 at 1:1 molar ratio at 4°C for 24 hours. WH-PEG conjugate was then subjected to dialysis into water using a centrifugal dialysis unit of 3000 MWCO. WH-PEG conjugate in water was frozen rapidly in dry ice at an angle for 1 hour before lyophilization overnight. Dry-frozen WH-PEG conjugate was reconstituted in 25 mmol/L HEPES solution (pH 8.0) to a final concentration of 400 nmol/L before use. To make WH-decorated TNE, antigen in Miglyol 812, αGC glycolipid, WH-PEG conjugate, and AM1 peptide were added into an Eppendorf tube to a final concentration of 25 μg/mL in oil volume fraction of 2% v/v, 2.5 μg/mL, 5% v/v, and 400 μmol/L, respectively. The mixture was sonicated for four 1-minute pulses at 60 W to form oil-in-water emulsions. Particle size was measured by a Zetasizer Nano ZS (Malvern Panalytical). Data analysis with DTS software used the nonnegativity constrained least-squares fitting algorithm. Dispersant refractive index and viscosity of the dispersant were assumed to be 1.45 and 1.02 centipoise. Each sample had 10 runs of 10 seconds.

TC-1 cells are a lung epithelial cell line derived from C57BL/6 mice, immortalized by HPV16 E6/E7, and transformed with an activated ras oncogene, a gift originally from T.C. Wu (John Hopkins University, Baltimore, MD; ref. 25). Authentication was determined by reactivity of E7-specific T cells. Cells were sourced from a batch of Mycoplasma-free stock, tested by PCR prior to cryopreservation in 2012. Cells were prepared by passaging twice in vitro for 4 days in complete DMEM containing 5 mmol/L HEPES before inoculation into C57BL/6 mice. Mice were subcutaneously (s.c.) inoculated with 2 × 10⁵ of TC-1 cells in the right flank region (day 0). Mice were then intravenously given 200 μL of antigen (5 μg) and αGC (5–500 ng) in soluble or TNE vector on day 7. Tumors were measured every 2 to 3 days with a digital caliper. Calculation of tumor volume (mm³) was as per formula: [length (mm) × width² (mm²)]/2. Mice were sacrificed when tumor reached a size of 1,000 mm³.

To assess the in vivo antigen-specific CD8⁺ cytotoxic T-cell response, 6 days after immunization with TNE, mice were each injected with 1 × 10⁷ congenic CD45.1⁺ splenocytes unloaded or loaded with 10⁻¹² to 10⁻⁶ mol/L of relevant CD8 peptide for 90 minutes at 37°C in complete RPMI and labeled for 10 minutes at 37°C with CFSE or CTV (0.5 or 5 μmol/L) in 1× PBS at 1:1 ratios. Twelve to 20 hours later, mice were sacrificed and splenocytes from the recipients were analyzed by flow cytometer to assess peptide-specific killing. The percentage peptide-specific lysis was calculated as follows: % specific killing = 100 × [1 – ((Experimental_loaded/Experimental_unloaded)/(Control_loaded/Control_unloaded))].

To assess tumor-specific CD8⁺ T-cell responses, tumor tissue, and tumor-draining lymph nodes were isolated from TC-1 tumor-bearing mice 7 days after vaccination and processed under sterile conditions. Briefly, tumor-draining lymph nodes and tumors were mechanically dissociated with collagenase D (2 mg/mL) and DNAse I (20 μg/mL) to create single-cell suspensions. For ex vivo restimulation, cell suspensions were then seeded into 24-well plates and stimulated with 1 μg/mL of gf001 peptide and 1 in 1,000 dilution (1 μg/mL) of GolgiPlug in complete RPMI media, along with control unstimulated group for 5 hours, unless otherwise stated. To assess long-term effector memory responses in vaccinated mice, mice were bled via the retro-orbital sinus and subjected to red blood cell lysis. Lymphocytes were then seeded into 96-well plates and were either left unstimulated or restimulated with 5 μg/mL of gf001 peptide for 5 hours in the presence of 1 in 1,000 dilution GolgiPlug in complete RPMI media. Cells were then collected from each well into polypropylene FACS tubes and subjected to antibody staining, including permeabilization for intracellular staining of IFNγ and TNFα.

For in vivo depletion of CD8⁺ T cells, 200 μg of anti-CD8β (53-5.8; Bio X Cell) was administered by intraperitoneal injections to vaccinated tumor-bearing mice on days 6 and 10 relative to tumor inoculation. CD8⁺ T cell–depletion efficacy was greater than 95%. For in vivo depletion of CD4⁺ cells, 100 μg of anti-CD4 (GK1.5; Bio X Cell) was administered by intraperitoneal injections to naïve mice one and 4 days prior to immunization. Depletion efficacy of CD4⁺ T cells and NKT cells were greater than 95%. For controls, equivalent doses of 2A3 control immunoglobulin (cIg) were administered.

Results are presented as mean ± SE. Kaplan–Meier plots were used to analyze mouse survival, and a log-rank test was performed to assess the statistical significance of differences between survival curves. For all other data, Student t test, one-way ANOVA were used to assess differences between two and more groups (GraphPad Prism 7 Software).

We generated nanoemulsions that targeted DCs by attaching an antibody against Clec9a, which is expressed on CD8α⁺ DC cell surface, as in ref. 23. Clec9a/OVA TNE can support antigen-specific CTL induction by activating Clec9a⁺ DCs (26, 27) and iNKT cells can be activated by and further amplify DC-mediated signals (28). We therefore investigated whether Clec9a/OVA TNE activates iNKT cells and whether iNKT cells contribute to the self-adjuvanting properties of the TNE. In wild-type (WT) mice, CD1d-Tet⁺ iNKT and NK cells were activated by Clec9a/OVA TNE immunization. The frequency of iNKT and NK cells producing IFNγ was increased, as compared with untreated or soluble OVA-treated mice; however, expression per cell was not changed (Fig. 1A). In contrast, fewer NK cells produced IFNγ upon Clec9a/OVA TNE immunization of iNKT-deficient Ja18^−/− mice (Fig. 1A), demonstrating the partial dependence of NK cell function on iNKT cell activation upon Clec9a/OVA TNE immunization. CD69 is an early activation marker for lymphocytes. Clec9a/OVA TNE treatment saw an increase in frequency in CD69⁺ iNKT cells and overall expression of this molecule (Fig. 1B). Similarly, in WT mice, an increase in the frequency CD69⁺ NK and T cells and expression of CD69 was observed upon Clec9a/OVA TNE treatment (Fig. 1B). These increases were negated in both cell types in the absence of iNKT cells, indicating the activation of NK and T cells is iNKT cell dependent. Activation of iNKT cells induces DC maturation (29). Clec9a/OVA TNE treatment induced upregulation of CD86 and CD80 but not CD40 expression on CD8α⁺ DCs in an iNKT cell–dependent manner (Fig. 1C). These observations coincide with the proportion of activated T cells expressing CD69 (Fig. 1B). To understand the extent to which cross-priming of antigen-specific CD8⁺ T cells was affected in the absence of activated iNKT cells, immunized WT and Jα18^−/− mice were challenged with dye-labeled target cells pulsed with various concentrations of SIINFEKL peptide as a readout for CTL activity in vivo. CTL activity was reduced in the absence of iNKT cells after Clec9a/OVA TNE immunization but was not completely abrogated (Fig. 1D). CD4⁺ T helper cells have also been shown to play a role in the induction of CTL cross-priming upon Clec9a TNE treatment (23). CD4⁺ T cells were depleted in vivo 1 and 4 days prior to TNE immunization (Supplementary Fig. S1a). As a proportion of iNKT cells also expresses CD4, we observed that 2 doses of anti-CD4 were sufficient to deplete CD4⁺ T cells as well as CD4⁺ iNKT cells in the blood (Supplementary Fig. S1b) and in the spleen (Supplementary Fig. S1c). To examine the effect of depletion of CD4⁺ cells on CTL cross-priming, mice were challenged with dye-labeled target cells pulsed with SIINFEKL. After 24 hours, no significant difference in CTL activity was observed between nondepleted and CD4-depleted Clec9a/OVA TNE-treated mice (Supplementary Fig. S1d). These observations demonstrate that iNKT cells enhance NK cell activation and CD8α⁺ DC maturation after Clec9a/OVA TNE immunization. In addition, iNKT cells also enhance cross-priming of antigen-specific CD8⁺ T cells, and this augmentation is independent of CD4⁺ T cells or CD4⁺ iNKT cells in this setting.

Given the adjuvant capacity of iNKT cells in response to Clec9a/OVA immunization, we sought to enhance the adjuvant effects of iNKT cells by incorporating the NKT cell-stimulating glycolipid ligand, αGC into the Clec9a/OVA TNE (Clec9a/OVA/αGC TNE). Physical size and stability of the TNE did not differ between Clec9a/OVA TNE and Clec9a/OVA/αGC TNE (Supplementary Fig. S2a), whereas surface charge was increased with inclusion of αGC (Supplementary Fig. S2b). WT mice were administered i.v. Clec9a/OVA TNE, Clec9a/OVA/αGC TNE, soluble OVA, or OVA plusαGC in PBS. Spleens were analyzed 3 and 24 hours after immunization, for iNKT cell and NK cell activity as well as CD8α⁺ DC maturation. As expected, treatment of mice with soluble OVA + αGC but not OVA alone activated iNKT cells and NK cells, as demonstrated by the production of intracellular IFNγ and TNFα (Fig. 2A), and increased CD69 expression (Fig. 2B). Production of IFNγ and CD69 expression in splenic iNKT and NK cells in mice administered Clec9a/OVA/αGC TNE increased to a similar extent after soluble OVA + αGC (Fig. 2A and B). Clec9a/OVA/αGC TNE administration significantly enhanced the frequency of CD40-, CD86-, and CD80-expressing CD8α⁺ DCs and average expression of these molecules, as compared with Clec9a/OVA TNE (Fig. 2C). CD8α⁺ DC maturation was similar in mice treated with soluble OVA + αGC or Clec9a/OVA/αGC TNE (Fig. 2C). These results demonstrate that incorporation of αGC promotes iNKT/NK activation and DC maturation and that αGC-containing TNE are as effective as nonencapsulated αGC for innate immune cell activation.

We next examined the development of antigen-specific CD8⁺ T-cell responses. Antigen-specific CTL activity, measured by lysis of target cells in vivo, was significantly higher in mice given Clec9a/OVA/αGC TNE than in mice receiving Clec9a/OVA TNE and in mice given soluble OVA + αGC (Fig. 3a). As expected, CTL activity was largely absent in mice receiving soluble OVA alone (Fig. 3a). Thus, activation of iNKT cells through Clec9a/OVA/αGC TNE enhances cross-priming of CD8⁺ T cells and supports previous studies indicating the benefit for simultaneous delivery of αGC and antigen to cross-presenting CD8α⁺ DCs to drive enhanced antigen-specific responses (5, 22). iNKT cell–induced DC maturation requires CD40–CD40L interactions occurring in parallel to CD4⁺ T-cell help (4, 30, 31). This interaction facilitates cross-priming of antigen-specific CD8⁺ T-cell responses (5, 32). To confirm that activation of iNKT cells can promote CTL activity in the absence of CD4⁺ T-cell help, mice were immunized with the minimal MHC I-binding OVA epitope (SIINFEKL), which stimulates a CD8⁺ T-cell response. Clec9a/SIINFEKL/αGC TNE-treated mice could generate a CTL response, whereas CTL responses in mice treated with Clec9a/SIINFEKL TNE were minimal (Fig. 3B). Similarly, mice receiving soluble SIINFEKL + αGC also induced minimal CTL activity in vivo (Fig. 3B). These results are consistent with the concept that codelivery of peptide and αGC to the same Clec9a⁺ DC is beneficial for CTL induction. Likewise, enhanced CTL activity was also observed with a second MHC I-binding peptide, gf001, derived from HPV16E7 protein incorporated into Clec9a/αGC TNE, relative to soluble gf001 + αGC. These observations demonstrate that inclusion of CD4⁺ T helper epitopes is not required in Clec9a/CD8⁺ peptide/αGC TNE to induce CD8⁺ CTLs capable of antigen-specific target cell killing.

In view of the efficient generation of antigen-specific CTL, we tested the therapeutic effect of TNE in the HPV16 E7-expressing TC-1 tumor model. We functionalized TNE with WH peptide that targets Clec9a, as described (20, 23). CTL lytic capacity was similar when comparing Clec9a/gf001/αGC and WH/gf001/αGC TNE (Fig. 4A). One dose of WH/gf001/αGC TNE containing varying concentrations of αGC at 5, 50, and 500 ng was administered intravenously 7 days after subcutaneous TC-1 inoculation. As controls, WH/αGC TNE or soluble αGC with gf001 peptide were administered. Mice vaccinated with WH/gf001/αGC TNE achieved the greatest tumor growth suppression, and this antitumor effect was αGC dose dependent (Fig. 4B). The reduction in tumor growth required TNE encapsulation of both peptide and αGC, as TNE encapsulation of either component alone failed to reduce tumor growth. Furthermore, encapsulation in TNE was more effective than soluble gf001 + αGC. The reduced tumor growth in WH/gf001/αGC-treated mice was associated with improved survival (Fig. 4C).

To determine whether the TC-1 tumor suppression by WH/gf001/αGC TNE was dependent on CD8⁺ T-cell activity, CD8⁺ T cells were depleted from tumor-bearing mice prior to treatment with WH/gf001/αGC^500ng TNE. The therapeutic effect was abrogated in the absence of CD8⁺ T cells (Fig. 5A). Because mature DCs are involved in the recruitment and activation of antitumor CD8⁺ tumor-infiltrating leukocytes (TIL), tumor-draining lymph nodes (TdLN) and tumor tissue were analyzed for the presence of mature DCs and tumor-specific CD8⁺ T cells 7 days after treatment. Indeed, the number of CD80⁺ DCs correlated with the number of effector CD8⁺ T cells in both TdLNs and in the tumors of vaccinated mice (Fig. 5B). Recruitment of DCs and CD8⁺ TILs was most prominent in mice vaccinated with WH/gf001/αGC^500ng TNE (Fig. 5B). The number of NK cells also correlated with the number of mature DCs in the TdLN and in tumor tissue of WH/gf001/αGC TNE-vaccinated mice (Supplementary Fig. S3), suggesting that activation of iNKT cells may help recruit NK cells into tumors. However, NK cell recruitment did not differ between mice receiving WH/gf001/αGC at different doses of αGC. To determine the proportion of functional antigen-specific CTL, lymphocytes from the TdLN and tumor were restimulated with gf001 peptide in vitro and intracellular amounts of IFNγ and TNFα were assessed. The largest proportion of cytokine-producing CD8⁺ T cells was detected in the TdLN and in tumor tissue of mice treated with WH/gf001/αGC^500ng TNE, followed by WH/gf001/αGC^5ng TNE and soluble gf001 + αGC^500ng (Fig. 5C). Thus, the quantity of polyfunctional antigen-specific effector CD8⁺ T cells correlated with tumor suppression induced by the vaccine.

The use of αGC can augment survival of CD8⁺ T memory cell subsets and support recall responses (33, 34). To assess antigen-specific memory CD8⁺ T-cell responses in circulation, surviving mice were bled 30 days after WH/gf001/αGC TNE treatment. Blood leucocytes were restimulated with gf001 peptide for 5 hours and analyzed for CD8⁺ T-cell IFNγ and TNFα production. Without restimulation, a small but distinct population of circulating CD8⁺ T cells produced both IFNγ and TNFα in WH/gf001/αGC^500ng TNE-treated mice, but not in surviving mice receiving WH/gf001/αGC^5ng TNE or soluble gf001+αGC^500ng (Fig. 6). Upon restimulation, a population of tumor-specific CD8⁺ T cells was observed in mice vaccinated with WH/gf001/αGC^500ng TNE (Fig. 6). These observations indicate that Clec9a/CD8 epitope/αGC TNE promote the long-term survival of polyfunctional tumor-specific CD8⁺ T cells and the efficiency of recall response to tumor antigens.

Over the past decade, the understanding of how to harness the immunostimulatory properties of iNKT cell glycolipids for therapy in conditions such as viral infections and tumors has improved. The drawbacks associated with soluble administration of αGC, including iNKT anergy and acute liver toxicity, prompted the development of glycolipid carrier techniques. Some of these nanocarriers such as lipid-based octaarginine-modified (35) or cationic liposomal nanocarriers (36) may improve the delivery and function of hydrophobic αGC (37), and have been used to codeliver αGC with antigen to induce antitumor immunity. Here, we improved on passive targeting systems by functionalizing oil-in-water TNE for the delivery of αGC and tumor-associated antigen directed to Clec9a⁺CD8α⁺ DCs for the optimization of tumor-specific CD8⁺ T-cell responses. Our study (i) provides insight into the role for iNKT cells in the self-adjuvanting responses induced by Clec9a/OVA TNE, (ii) validates Clec9a as an entry receptor into CD8α⁺ DCs for the combinatorial processing of antigens and αGC for cancer immunotherapy, (iii) confirms the adjuvanting effects of iNKT cells for cancer vaccines, and (iv) validates TNE as a biocompatible delivery vector for glycolipids.

We showed that Clec9a-targeted TNE encapsulating protein antigen or pooled CD4 and CD8 epitope peptides could induce antigen-specific CD8⁺ T-cell responses and tumor control in the absence of adjuvant. This self-adjuvanting function was found to be dependent on CD4⁺ T-cell help and IFNα production for DC maturation and cross-priming of CTL (23). Other studies have also demonstrated that TLR9 stimulation, the production of IFNα, and TLR9-mediated production of charged sphingolipids can promote iNKT cell activation (26, 28, 38). Here we show that Clec9a/OVA TNE activate iNKT cells, and NK cells, and that in the absence of iNKT cells, activation of NK cells as well as CD8α⁺ DCs and antigen-specific CD8⁺ T cells was reduced. This observation demonstrates that iNKT cells play a role and suggests they enhance the self-adjuvanting effect of Clec9a/OVA TNE (28). iNKT cells provide maturation signals to DCs via CD40–CD40L interaction (4, 22). The absence of CD86 and CD80 upregulation 24 hours after TNE immunization and the reduced lytic activity of CTLs in iNKT cell–deficient mice suggest that a lack of DC CD40 engagement by iNKT cells prevents the enhanced induction of T-cell costimulation and IL12 production (4). In the absence of iNKT cells, persistent CD40 upregulation may also have been mediated by other innate signals such as proinflammatory cytokines, including TNFα and IFNα (39, 40). Our data thus support the evidence that iNKT cells influence the development of functional CD8α⁺ DCs necessary for CTL induction (41). To then address the contribution of CD4⁺ T-cell help in the promotion of CTL cross-priming after Clec9a/OVA TNE treatment in this setting, including a subset of CD4⁺ iNKT cells, anti-CD4 depletion was performed prior to TNE treatment. Depletion of CD4⁺ cells did not affect CTL activity induced by Clec9a/OVA TNE. Our data therefore indicate that CD4⁻ iNKT cells are required for the production of immunostimulatory IFNα, which was shown previously to be necessary for DC activation and CTL activity after Clec9a/OVA TNE treatment (23, 42).

Clec9a/peptide/αGC TNE and soluble αGC + peptide similarly led to activation of iNKT cells, NK cells, and CD8α⁺ DCs in the spleen. However, antigen-specific CTL activity in vivo was enhanced in mice immunized with Clec9a/OVA/αGC TNE, as compared with soluble αGC and OVA. This observation is supported by previous studies using targeted or nontargeted PLGA nanoparticles encapsulating αGC and OVA antigen (12, 43, 44). In addition, we also show that iNKT cells promote cross-priming of minimal CD8⁺ T-cell epitopes. Our results therefore support the concept of simultaneous delivery of αGC and antigen to the same DCs for optimal CTL induction (22), but may also suggest how alternative T cell-helper signals from iNKT cells synergize with other innate signals such as type I IFNs to enhance cross-priming (5, 23, 45).

Cytotoxic tumor-specific T cells are useful for cancer vaccine immunotherapy. In an oncogenic virus HPV16-E7–driven tumor model TC-1, in which oncogenesis is dependent on the conserved E7 protein (46), we demonstrate the adjuvanting effects of WH/gf001/αGC TNE for iNKT cell stimulation, driving systemic polyfunctional E7-specific CD8⁺ T cells. Ghinnagow and colleagues delivered multiple doses of Clec9a-targeted PLGA nanoparticles to treat a subcutaneous model of B16F10 tumor (21). Here we investigated the efficacy of a single i.v. administration of WH/gf001/αGC TNE with increasing concentrations of αGC. Previously, WH/protein TNE distribution to the spleen, liver, and tumor indicated targeting of WH-TNE to CD8⁺ cross-presenting DCs in these tissues (23). Here we show that CD8⁺ T-cell tumor infiltration, E7-specific CD8⁺ tumor T-cell responses, circulating antigen-experienced memory CD8⁺ T cells, and overall mouse survival correlate with the concentration of administered αGC delivered in WH/peptide TNE. Thus, our method of iNKT cell activation promotes a CD8⁺ T-cell response profile of protective long-term immunity (47). In sum, our observations demonstrate how αGC-dependent iNKT cell activation supports the generation of memory CD8⁺ T cells involved in solid tumor control.

Mature DCs stimulate CD8⁺ T cells associated with effector antitumor responses (48). In a variety of tumors, tumor-infiltrating (Ti-) DCs were often phenotypically immature and functionally suppressive through a variety of mechanisms in the tumor microenvironment (49). By directing αGC to tumor-infiltrating CD8⁺ DCs, TiDCs were CD80⁺. Their frequency and intensity of CD80 expression correlated with that of polyfunctional effector CD8⁺ T cells in an αGC dose–dependent manner. Therefore, WH/peptide/αGC targeting of tumor-infiltrating CD8⁺ DCs promotes DC activation from a normally immature or immune-suppressed state (50). This behavior is associated with cross-talk with activated iNKT cells. Although the effector role of NK cells is not investigated here and long-term suppression of TC-1 tumors appears NK cell independent, recruited NK cells likely activate and recruit DCs and CD8⁺ T cells to the site of tumor (51, 52). Thus, WH/αGC/peptide TNE codeliver antigen and innate immune signals that disrupt tolerizing tumor conditions and drive enhanced tumor-specific systemic responses synonymous with the hallmarks of successful immunotherapy (21). Thus, targeting αGC and peptide to CD8⁺ DCs using Clec9a/αGC/peptide TNE is a flexible and translatable delivery approach with greater efficacy than approaches using soluble glycolipids.

R. Thomas reports receiving other commercial research support from and has received honoraria from speakers bureau of Merck & Co. No potential conflicts of interest were disclosed by the other authors.

Conception and design: P.Y. Lam, B. Zeng, G. Leggatt, R. Thomas, S.R. Mattarollo

Development of methodology: P.Y. Lam, B. Zeng, R. Thomas, S.R. Mattarollo

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P.Y. Lam, T. Kobayashi, M. Soon, B. Zeng

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P.Y. Lam, R. Dolcetti, G. Leggatt, R. Thomas, S.R. Mattarollo

Writing, review, and/or revision of the manuscript: P.Y. Lam, T. Kobayashi, B. Zeng, G. Leggatt, R. Thomas, S.R. Mattarollo

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P.Y. Lam, M. Soon

Study supervision: R. Dolcetti, R. Thomas, S.R. Mattarollo

The authors acknowledge the Translational Research Institute for providing an excellent research environment and core facilities that enabled this research to be conducted. We particularly thank Siok Min Teoh and Daniel Kerage for technical assistance, and the Biological Resources and Flow Cytometry Core Facilities for technical support. Gratitude is extended to Associate Professors Mirelle Lahoud and Irina Caminschi (Monash University, Melbourne) for the provision of the anti-mouse Clec9a IgG2a mAb. This work was supported from grant (1087691) jointly funded by Cancer Australia and Cure Cancer Australia. P.Y. Lam was supported by a University of Queensland International Scholarship. S.R. Mattarollo was supported by an NHMRC Career Development Fellowship (1061429).

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.