Peptide‐Based Cancer Vaccine Delivery via the STINGΔTM‐cGAMP Complex

Abstract

With the advent of bioinformatic tools in efficiently predicting neo‐antigens, peptide vaccines have gained tremendous attention in cancer immunotherapy. However, the delivery of peptide vaccines remains a major challenge, primarily due to ineffective transport to lymph nodes and low immunogenicity. Here, a strategy for peptide vaccine delivery is reported by first fusing the peptide to the cytosolic domain of the stimulator of interferon genes protein (STINGΔTM), then complexing the peptide‐STINGΔTM protein with STING agonist 2'3' cyclic guanosine monophosphate–adenosine monophosphate (cGAMP). The process results in the formation of self‐assembled cGAMP‐peptide‐STINGΔTM tetramers, which enables efficient lymphatic trafficking of the peptide. Moreover, the cGAMP‐STINGΔTM complex acts not only as a protein carrier for the peptide, but also as a potent adjuvant capable of triggering STING signaling independent of endogenous STING protein—an especially important attribute considering that certain cancer cells epigenetically silence their endogenous STING expression. With model antigen SIINFEKL, it is demonstrated that the platform elicits effective STING signaling in vitro, draining lymph node targeting in vivo, effective T cell priming in vivo as well as antitumoral immune response in a mouse colon carcinoma model, providing a versatile solution to the challenges faced in peptide vaccine delivery.

Cytosolic stimulator of interferon genes protein‐peptide + 2'3' cyclic guanosine monophosphate–adenosine monophosphate complexes are synthesized via recombinant protein technology for vaccination against cancer. The complexes achieve high levels of vehicle‐free delivery in vitro, antigen presentation ex vivo, and T‐cell priming in vivo. Additionally, they show increased accumulation of adjuvant and epitope peptide in the draining lymph nodes and inhibited tumor growth.

1. Introduction

In the past decade, the field of oncology has been revolutionized by cancer immunotherapy, which utilizes the antigen‐directed cytotoxicity of T cells to eliminate tumor cells. Neoantigens, cancer‐specific peptides that are absent from the human genome, have thus emerged as promising antigenic targets for T cells to generate anti‐tumor responses.^{[ 1 ]} Various techniques have enabled the efficient identification and optimization of neoantigens,^{[ 2 ]} which can then be easily synthesized and characterized before they are administered to the patient as personalized cancer vaccines.^{[ 3 ]} However, despite advances in neoantigen research, the potency of peptide antigens are in general lacking due to insufficient trafficking to the draining lymph nodes and poor immunogenicity from ineffective activation of antigen presenting cells (APCs).^{[ 4 ]} These challenges have motivated studies on the development of immune‐stimulatory adjuvants and carriers, including the fusion of peptide epitopes to transport proteins^{[ 5 ]} or antibodies^{[ 6 ]} for enhanced targeting and accumulation in draining lymph nodes, as well as the co‐delivery of peptides with adjuvants in synthetic cargos such as polymersomes,^{[ 7 ]} liposomes,^{[ 8 ]} gold nanoparticles,^{[ 9 ]} and hydrogels.^{[ 10 ]}

Among these adjuvants, those that interact with the stimulator of interferon genes (STING) pathway have gained increasing attention as a therapeutic target in cancer immunotherapy, due in large part to STING's potent activation of APCs and antitumoral T cell responses.^{[ 11 ]} Consequently, STING agonists such as 2’3’ cyclic guanosine monophosphate (GMP)–adenosine monophosphate (AMP) (cGAMP) have been included in various cancer vaccines to enhance the efficacy of checkpoint blockade.^{[ 12 ]} Our group has recently reported the use of STING as a biologic and carrier, where the cytosolic, C‐terminal domain (CTD) of STING protein (STINGΔTM, which lacks the transmembrane component) is repurposed as a fully functional platform for cGAMP delivery.^{[ 13 ]} We found that STINGΔTM self‐assembles with cGAMP at physiological conditions to form a well‐defined tetrameric complex that enters the cytosol and is capable of triggering STING signaling even in STING‐deficient cell lines, and also observed an effective transport of the cGAMP‐STINGΔTM complex to the draining lymph nodes after tail base injection in mice.^{[ 13 ]}

In light of the aforementioned challenges in peptide vaccine delivery, we leveraged this biologically informed cGAMP‐STINGΔTM signaling complex to deliver a model antigen epitope from chicken ovalbumin (OVA) amino acids 252–272: GLEQLESIINFEKLTEWTSS (denoted as SIINFEKL) by fusing the peptide to the N‐terminus of the STINGΔTM and then complexing the fusion protein with cGAMP. This approach of recombinant protein technology likewise introduces the possibility of including additional epitopes or functional peptides in a modular vaccine platform. The co‐localized delivery of antigen epitope, adjuvant cGAMP, and cGAMP's functional carrier STINGΔTM resulted in a similar self‐assembled signaling complex of SIINFEKL‐STINGΔTM with cGAMP that could activate type‐I interferon in vitro, activate APCs ex vivo and in vivo, and traffic to the inguinal lymph nodes. It was also observed to elicit robust antigen‐specific T‐cell responses and a robust anti‐tumoral response in a prophylactic MC38 colon cancer mouse model, demonstrating the potential of this platform in addressing the poor immunogenicity and ineffective lymph node trafficking of peptide vaccines.

2. Results

2.1. Overview of the Delivery Strategy

Previously, we reported the self‐assembly of pre‐dimerized STINGΔTM to a well‐defined cGAMP‐STINGΔTM tetramer,^{[ 13 ]} which resulted in lymphatic accumulation several times higher than that of an equal amount of STINGΔTM protein injected (possibly due to the increase of size from STINGΔTM dimer to cGAMP‐STINGΔTM tetramer). Inspired by the platform's performance in lymph node trafficking and its function as a highly potent adjuvant, we sought to adapt it for peptide vaccine delivery by fusing the model antigen epitope peptide to the N‐terminus of the STINGΔTM protein. Its C‐terminus‐fused counterpart was also synthesized as part of a pilot vaccination study in BL/6 mice (Figure S1, Supporting Information), though fusion at the N‐terminus resulted in a significantly higher level of tumor necrosis factor (TNF)‐α ⁺/ CD8⁺ cells relative to an OVA + cGAMP control. As a result, all subsequent experiments were conducted with peptides fused at the N‐terminus of STINGΔTM.

In this study, we used the SIINFEKL peptide–the class I (Kb)‐restricted peptide epitope of chicken ovalbumin (OVA) presented by class I MHC molecules–as our model antigen. Figure  1A illustrates the predicted structure of SIINFEKL‐STINGΔTM by protein homology modeling. To verify that the fusion of SIINFEKL to cytosolic STING will not alter its tetramerization–and in turn, its desirable size for lymph node accumulation–we analyzed SIINFEKL‐STINGΔTM protein with fast protein liquid chromatography (FPLC) and stepwise titration with increasing molar ratios of cGAMP (Figure 1B).

Figure 1.

cGAMP binding induces self‐assembly of peptide‐STINGΔTM tetramer. A) Structural prediction of dimerized STINGΔTM protein with a peptide antigen fused at the N‐terminus (STINGΔTM dimer structure predicted with protein homology‐modeling by SWISS modeling).^{[ 18 ]} FPLC analyses of B) mouse SIINFEKL_STINGΔTM and C) mouse SIINFEKL_STINGΔTM_R237A/Y239A mutant (which is not capable of binding cGAMP) in PBS, titrated with various molar equivalences of cGAMP.

This process resulted in the formation of a monodisperse, tetrameric cGAMP‐SIINFEKL‐STINGΔTM complex (≈120 kD) from the SIINFEKL‐STINGΔTM dimers (≈60 kD), with excess cGAMP (674D) eluting off after all SIINFEKL‐STINGΔTM had been saturated. In contrast, the R237A/Y239A mutant of SIINFEKL‐STINGΔTM, which is unable to bind cGAMP, results in no peak shift upon cGAMP titration (Figure 1C). These results corroborate our conclusion that the tetramerization of SIINFEKL‐STINGΔTM is induced through cGAMP's specific interaction with STINGΔTM.

2.2. cGAMP‐SIINFEKL‐STINGΔTM Activates STING Signaling In Vitro

Subsequently, we sought to ascertain the functionality of the complex as a peptide vaccine adjuvant that triggers STING signaling. Both cell lines with and without endogenous STING were used to evaluate interferon activity resulting from this treatment, in order to account for potential epigenetic silencing of STING in certain cancer cells^{[ 14 ]} and the possibility of deficient downstream signaling from the HAQ mutation, which exists in 19% of the human population.^{[ 15 ]} Figure  2A illustrates the expression of cyclic GMP–AMP synthase (cGAS), STING, TANK‐binding kinase 1 (TBK1), and interferon regulatory factor 3 (IRF3) in the STING‐deficient, interferon (IFN)‐luciferase reporter cell line HEK293T, as well as the mouse macrophage cell line RAW264.7 and mouse dendritic cell line DC2.4, both of which have fully functional endogenous STING.

Figure 2.

Peptide‐fused STINGΔTM‐cGAMP complex effectively activates STING signaling in vitro. A) Immunoblotting of endogenous expression of cGAS, STING, TBK1, and IRF3 in DC2.4, RAW264.7, and HEK293T cells. B) HEK293T cells (n = 3) treated with SIINFEKL peptide‐fused STINGΔTM‐cGAMP complexes along with mutant controls, with the help of TransITx2. Interferon‐luciferase activity was measured 24 h post treatment, both ***p < 0.0001. DC2.4 cells (n = 3) treated with SIINFEKL peptide‐fused STINGΔTM‐cGAMP complexes along with mutant controls C) with (p values left to right: **p = 0.0089, *p = 0.03, **p = 0.0087) and D) without the help of TransITx2 (p values left to right: ***p = 0.0002, ns p = 0.604, ns p = 0.056, **p = 0.0012). Raw264.7 cells (n = 3) treated with SIINFEKL peptide‐fused STINGΔTM‐cGAMP complexes along with mutant controls E) with (p values left to right: **p = 0.0014, ns p = 0.896, **p = 0.001, ns p = 0.148) and F) without the help of TransITx2 (p values left to right: ***p < 0.0001, ns p = 0.804, *p = 0.0266, ***p < 0.0001). For DC2.4 and Raw264.7 cells, STING activation was determined by measuring secreted CXCL10 concentration in the culture media 48 h post treatment. Values are reported as means±SEM (*p < 0.05, **p < 0.01, ***p < 0.005). G) Confocal micrograph of DC2.4 and RAW264.7 cells 8 h post incubation with AF488‐SIINFEKL‐ STINGΔTM and cGAMP as a vehicle free treatment. Scale bar 20 µm.

We observed significantly stronger interferon activity upon the delivery of cGAMP‐STINGΔTM and cGAMP‐SIINFEKL‐STINGΔTM to HEK293T cells via commercial transfection reagent TransITx2 (Figure 2; Figure S2, Supporting Information). No significant differences were observed between the two groups. This is in contrast to the cGAMP‐only treatment group, which exhibits no STING signaling due to lack of endogenous STING protein. Treatment with functional mutants R237A/Y239A and S365A—which are unable to bind cGAMP and undergo TBK1 phosphorylation, respectively—likewise resulted in ineffective STING signaling. Ultimately, these mutant controls demonstrate the use of SIINFEKL‐STINGΔTM as a functional adaptor protein instead of an inert vehicle for cGAMP, underscoring its key ability to circumvent diminished STING signaling.

Delivery of the treatment groups to the two cell lines with endogenous STING—DC2.4 and RAW264.7—resulted in STING activity for cGAMP‐only, cGAMP‐SIINFEKL‐STINGΔTM, and cGAMP‐SIINFEKL‐STINGΔTM mutants (Figure 2C–G). This was evaluated via measuring the concentration of mouse C‐X‐C motif chemokine ligand 10 (CXCL10), a chemokine secreted further downstream of STING signaling.^{[ 16 ]} Once introduced to the cells via commercial transfection reagent, the cGAMP‐only control resulted in similar levels of activity to cGAMP‐SIINFEKL‐STINGΔTM and its cGAMP complexed mutants, with only the R237A/Y239A mutant resulting in significantly lower interferon activity in the RAW264.7 cell line. This suggests that SIINFEKL‐STINGΔTM may function primarily or partially as a carrier in the presence of endogenous STING, as opposed to playing a critical role in triggering STING signaling under STING‐depleted conditions.

2.3. Draining Lymph Node Trafficking and T Cell Priming

As briefly noted in the introduction, the low lymphatic trafficking of peptides presents an obstacle in the development of peptide vaccine candidates. Researchers in the field have demonstrated that fusion of the epitope peptide to a transport protein, such as mouse serum albumin (MSA), may ameliorate this problem.^{[ 5 ]} We thus sought to evaluate the efficacy of the cGAMP‐SIINFEKL‐STINGΔTM platform in targeting the lymphatic system against that of SIINFEKL‐MSA+ cGAMP and SIINFEKL‐only, and found that cGAMP‐SIINFEKL‐STINGΔTM and cGAMP‐STINGΔTM exhibited significantly higher accumulation in the draining lymph node 24 h post tail base injection (Figure  3A–C; Figure S3, Supporting Information). No significant differences were observed in lymph node trafficking and lymph weight between cGAMP‐SIINFEKL‐STINGΔTM and cGAMP‐STINGΔTM groups. Whole organ biodistribution confirmed accumulation of all proteins (SIINFEKL‐STINGΔTM, STINGΔTM, and MSA) in the lymph nodes, as well as significantly higher lymph node weight for the cGAMP‐SIINFEKL‐STINGΔTM and cGAMP‐STINGΔTM groups (Figure 3A–F). We also observed significantly higher adjuvant delivery to the lymph nodes through tail‐base administration of Cy7‐labeled cGAMP with unlabeled SIINFEKL‐STINGΔTM (Figure 3G–I), demonstrating sufficient stability of the tetrameric complex to result in effective lymphatic trafficking. The biocompatibility of this system was also evaluated via enzyme‐linked immunosorbent assay (ELISA) detection of TNF‐α and interleukin‐6 (IL‐6) levels in mouse serum 2 h post tail base injection (Figure S4, Supporting Information).

Figure 3.

Peptide‐fused STINGΔTM‐cGAMP complex effectively trafficks to draining lymph node. A) Groups of Balb/c mice (n = 3) were tail‐base injected with cGAMP plus Cy7_SIINFEKL_STINGΔTM, cGAMP plus Cy7_SIINFEKL_MSA, or cGAMP plus STINGΔTM. Imaging was performed 24 h post injection. B) Representative Cy7 fluorescence imaging of the inguinal lymph nodes (dLNs), spleen, kidneys, liver, lungs, and heart. C) Integrated fluorescent intensity of Cy7 labeled protein/peptide in draining lymph nodes, normalized to fluorescence over all organs (*p = 0.0104, ns p = 0.5379). D) Draining lymph node weight *p = 0.0177, ns p = 0.7589. E) Whole organ biodistribution (integrated fluorescent intensity normalized over all organs). F) Whole organ biodistribution normalized by organ or tissue mass. G) Groups of Balb/c mice (n = 3) were tail‐base injected with Cy7‐labeled cGAMP plus SIINFEKL_STINGΔTM, Cy7‐labeled cGAMP, and Cy7‐labeled cGAMP plus SIINFEKL_R237A/Y239A_STINGΔTM. H) Imaging of Cy7 fluorescence signal in inguinal lymph nodes. I) Integrated fluorescent intensity of Cy7‐cGAMP in draining lymph nodes (***p < 0.0001). Values are reported as means±SEM (*p < 0.05, **p < 0.01, ***p < 0.005).

We then performed an ex vivo antigen presentation assay with DC2.4 cells (Figure  4A) in order to evaluate peptide‐specific immune responses. DC2.4 cells were first treated for 24 h with either cGAMP‐SIINFEKL‐STINGΔTM complexes or controls such as OVA full protein, OVA + cGAMP, and cGAMP‐SIINFEKL‐STINGΔTM S365A mutants containing an equivalent amount of the SIINFEKL peptide. These cells were then co‐cultured for three days with lymphocytes stained with carboxyfluorescein succinimidyl ester (CFSE) intracellular protein dye that had been harvested from OT1 mice engineered to produce SIINFEKL‐specific CD8 T‐cells; cells were then stained with CD8 antibody and analyzed via flow cytometry for actively proliferating OT1 CD8 T‐cells in response to SIINFEKL presentation (represented by the CFSE ^low/CD8⁺ population). Treatment with cGAMP‐SIINFEKL‐STINGΔTM or the cGAMP‐SIINFEKL‐STINGΔTM S365A mutant resulted in significantly stronger T cell responses in comparison to all other treatment groups (Figure 4B; Figure S5, Supporting Information), demonstrating successful DC maturation and antigen presentation. This is in good agreement with our previous in vitro results, where no statistical difference between these two treatment groups were observed in DC2.4 cells with endogenous STING.

Figure 4.

Peptide‐fused STINGΔTM‐cGAMP complex activates antigen‐specific T cell response ex vivo and in vivo. A) Ex vivo antigen presentation: DC2.4 cells were first treated with cGAMP‐SIINFEKL_ΔTM along with controls, and co‐cultured with CFSE stained OT1 lymphocytes the following day. Cells were collected 3 days after co‐culture for flow cytometry analysis. B) Antigen specific T cells were gated as CFSE low CD8+ cells (*p = 0.0182, ns p = 0.3845). C) Groups of BL/6 mice were vaccinated with cGAMP‐SIINFEKL_ΔTM along with controls at week 0 and 2. D) At week 3 mice blood were collected for CD8 antibody and SIINFEKL tetramer staining followed by flow cytometry analysis for SIINFEKL‐specific CD8 T cells (p values left to right: ns p = 0.1349, ns p = 0.1216, ***p = 0.0005). Values are reported as means±SEM (*p < 0.05, **p < 0.01, ***p < 0.005).

Finally, we vaccinated groups of BL/6 mice with cGAMP‐SIINFEKL‐STINGΔTM and a varied panel of controls, including but not limited to various SIINFEKL‐STINGΔTM mutants, cGAMP + SIINFEKL peptide + STINGΔTM, and cGAMP + SIINFEKL‐MSA protein (Figure 4C,D; Figure S6, Supporting Information). Blood was collected at Week 3 following the initial prime dose and a boost at Week 2, after which peripheral blood mononuclear cells (PBMCs) were stained with CD8 antibody and SIINFEKL MHC Tetramer. Treatment with cGAMP‐SIINFEKL‐STINGΔTM resulted in the strongest average antigen‐specific T‐cell response amongst all groups, with no statistical difference between the two mutants and p < 0.0001 between all remaining groups. Notably, this comparison includes the SIINFEKL‐MSA + cGAMP and SIINFEKL‐MSA‐only treatment groups, both of which have been reported to increase trafficking to the draining lymph nodes.^{[ 5 ]} Our strategy produces a larger average antigen‐specific T cell response than this current standard, highlighting the special advantage of our delivery platform in enhancing T cell priming through its active signaling function and carrier nature.

2.4. Anti‐Tumoral Immunity

To evaluate the anti‐tumoral efficacy of this peptide delivery system, we tail‐base vaccinated groups of BL/6 mice with cGAMP‐SIINFEKL‐STINGΔTM alongside several controls. Mice were primed at Day 0, received a boost at Day 14, and challenged at Day 21 with a subcutaneous inoculation of 1 million SIINFEKL‐expressing MC38 colon carcinoma cells. Emergence of tumor as well as the subsequent growth of tumors were monitored over time (Figure 5 ). Some vaccination effect was observed in the SIINFEKL‐STINGΔTM R237A/Y239A group, corroborating trends observed in signaling activation in vitro and CD8⁺ T cell priming in vivo, though this was still significantly lower (*p = 0.0291) in comparison to the cGAMP‐SIINFEKL‐ STINGΔTM group. Overall, the cGAMP‐SIINFEKL‐STINGΔTM vaccinated group resulted in complete inhibition of tumor growth—all ten mice in the group remained tumor‐free, affirming the potential of this platform as a tool for peptide vaccine delivery.

Figure 5.

cGAMP‐SIINFEKL‐STINGΔTM promotes enhanced antitumoral immunity. A) Groups of BL/6 mice were vaccinated with cGAMP‐SIINFEKL‐STINGΔTM along with other controls via tail base injection on day 0 and day 14. On day 21, mice were challenged with 1 million SIINFEKL‐MC38 cells subcutaneously. B) Tumor growth and C) percentage tumor free mice were monitored over time. Values are reported as mean ± SEM. The p‐value reported (p < 0.0001) was obtained from a log‐rank test of all trial groups.

3. Conclusion

Despite their many benefits in cost and ease of synthesis, the effective delivery of peptide‐based vaccines remains a challenge due to poor immunogenicity and inefficient lymphatic accumulation. Here, we report a platform for a peptide‐based cancer vaccine that simultaneously traffics the peptide to draining lymph nodes and activates the STING signaling pathway, circumventing the aforementioned issues through cGAMP‐induced self‐assembly and enhanced adjuvanticity. Fusion of the SIINFEKL epitope peptide to the N‐terminus of STINGΔTM protein and subsequent complexation with cGAMP results in a well‐defined tetrameric structure of desirable size for draining lymph node trafficking; inclusion of the STING protein activates STING signaling in vitro and boosts antigen presentation ex vivo. Moreover, cGAMP‐SIINFEKL‐STINGΔTM is shown to induce a robust T cell response, as well as potent anti‐tumoral immunity. These results signal the promise in introducing peptide vaccines as part of a cGAMP‐peptide‐STINGΔTM complex, which provides an effective platform for the co‐localized delivery of peptide with a strong adjuvant. Ultimately, this strategy may not only benefit the field of cancer immunotherapy, but also find applications in other fields, such as vaccine research against infectious diseases.

4. Experimental Section

Protein Expression and Purification

Model peptide vaccine GLEQLESIINFEKLTEWTSS from chicken ovalbumin amino acids 252–272 was fused to the N‐terminus of mouse serum albumin (MSA) and both the N‐ and C‐terminus of STINGΔTM protein (amino acids 138–378 of mouse STING or 139–379 of human STING) and cloned into pSH200 backbone via NcoI and NotI restriction enzyme sites. A hexameric histidine tag was placed at the N‐terminus of all proteins for purification. Site‐specific mutagenesis was applied to generate mutant proteins such as SIINFEKL‐mSTINGΔTM R237A/Y239A and S365A. DE3 Escherichia coli (E. coli) was used to express the peptide‐fused STINGΔTM proteins,^{[ 16 ]} BL21 DE3 was used for mouse STINGΔTM and Rosetta DE3 was used for human STINGΔTM and MSA. 1 L of E. coli was cultured in Luria‐Bertani (LB) broth (with antibiotics 100 mg mL⁻¹ ampicillin for BL21 DE3, 100 mg mL⁻¹ ampicillin and 35 mg mL⁻¹ chloramphenicol for Rosetta DE3) at 37 °C, 220 rpm till OD600 reaches 0.4. Isopropyl‐β‐d‐thiogalactopyranoside (IPTG) was added to 0.5 × 10⁻³ m working concentration for induction at 20 °C, 220 rpm overnight. After induction, the bacteria culture was centrifuged to collect the pellet, which was then washed with phosphate buffer saline (PBS), re‐suspended in protein binding buffer (50 × 10⁻³ m sodium phosphate, 0.5 × 10⁻³ m NaCl, and 10 × 10⁻³ mimidazole) and lysed with 1 mg mL⁻¹ lysozyme, 1% Triton X‐100, and 1 × 10⁻³ m phenylmethylsulfonyl fluoride (PMSF) at room temperature for 20 min. A probe sonicator was then used to further disrupt the cells on ice water at 18 W with 3‐s on and 5 s off intervals for a total of 5 min. The cell lysate was then centrifuged, and the supernatant was incubated with cobalt beads for 1 h followed by two washing steps with 0.1% Triton X‐114 protein binding buffer for endotoxin removal. The cobalt beads were then loaded onto gravity flow columns (Poly‐Prep chromatography column, Bio‐Rad, 7 311 550) and eluted with 1.5 mL protein elution buffer (50 × 10⁻³ m sodium phosphate, 0.5 × 10⁻³ m NaCl, and 150 × 10⁻³ m imidazole). Protein elution was then loaded onto size exclusion desalting columns (Zeba Spin Desalting Columns 40k MWCO 10 mL, Thermo Fisher Scientific, 87 772) and buffer exchanged to protein storage buffer (20 × 10⁻³ m HEPES, 150 NaCl, 10% glycerol, and 1 × 10⁻³ m, 1,4‐Dithiothreitol). Protein concentration was determined with DC Protein Assay Kit I (Biorad 5 000 111) and protein purity was verified with SDS‐PAGE.

FPLC Characterization of cGAMP‐Binding Induced SIINFEKL‐STINGΔTM Tetramerization

AKTA pure fast protein liquid chromatography (FPLC) with Superdex 200 Increase 10/300 GL size exclusion column was used to analyze the interaction between cGAMP and SIINFEKL‐STINGΔTM proteins as described in literature.^{[ 17 ]} cGAMP concentration was confirmed using Nanodrop. For each run, 300 µg of protein with different molar ratios of cGAMP was first mixed in 500 µL PBS and equilibrated at room temperature for 30 min. The sample was then loaded onto the column followed by isocratic elution of PBS at 1 mL min⁻¹ flow rate. The protein concentration was monitored with OD280. A protein standard mix for size exclusion chromatography was used to calibrate FPLC elution time to molecular weight.

Cell Culture

HEK293T and RAW264.7 cells were obtained from the American Type Culture Collection (ATCC). DC2.4 cells were obtained from the Rock lab at University of Massachusetts Medical School, MA, USA. RAW‐Blue ISG cells were obtained from Invivogen. HEK293T and RAW264.7 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. For SEAP‐IFN assay, RAW‐Blue ISG cells were cultured in DMEM with 10% heat‐inactivated (56 °C, 30 min) FBS and 1% penicillin/streptomycin. DC2.4 cells were cultured in Roswell Park Memorial Institute (RPMI) medium with 10% FBS and 1% penicillin/streptomycin. All cells are cultured in a 37 °C, 5% CO₂ incubator, used at low passage number and tested negative for mycoplasma contamination.

Western Blotting

Cells were first washed with PBS and then collected in T‐PER tissue protein extraction reagent (30 µL per million cells) with Halt protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific, 78 442) and incubated at 4 °C for 30 min with gentle vortexing. The lysate was then centrifuged, and the supernatant was collected to measure the total protein concentration with DC Protein Assay Kit I (Biorad 5 000 111). Fifty microgram of total protein was loaded onto each lane of SDS‐PAGE and subsequently transferred to nitrocellulose membrane, which was then blocked with 5% w/v non‐fat milk (Cell signaling, 9999S) and incubated with primary antibodies mouse anti‐α‐tubulin (Cell signaling, 3873S), rabbit anti‐STING (Cell signaling, 13647S), rabbit anti‐IRF3 (Cell signaling, 4302S), rabbit anti‐mcGAS (Cell signaling, 31659S), rabbit anti‐hcGAS (Cell signaling, 83623S), rabbit anti‐TBK1 (Abcam, 235 253), and secondary antibodies anti‐rabbit HRP (Cell signaling, 7074S) and anti‐mouse HRP (Thermo Fisher Scientific, 62–6520).

Detection of In Vitro STING Activation with luc2p‐ISRE Reporter in HEK293T Cells

HEK293T‐luc2p/ISRE/Hygro cells were seeded in clear bottom flat white (or black) 96‐well plates 100 µL per well at a density of 3.5 × 10⁵ cells mL⁻¹. Following overnight incubation, each well of cells were treated with a mixture of 1 µg of SIINFEKL‐STINGΔTM protein, 0.025 µg of cGAMP, and 1 µL of TransIT‐X2 in a total volume of 20 µL OptiMEM media. After treated cells had been incubated for 24 h, a firefly luciferase assay kit (Biotium, 30075‐2) was used to determine the luciferase expression following the manufacturer's protocol. Briefly, the medium was aspirated and cells were lysed then mixed with luciferase assay buffer containing 0.2 mg mL⁻¹ freshly added D‐luciferin, followed by plate‐reading for bioluminescence.

Detection of In Vitro STING Activation with mCXCL10 ELISA in RAW264.7 and DC2.4 Cells

RAW264.7 or DC2.4 cells were seeded in 96‐well plates 100 µL per well at a density of 2 × 10⁵ cells mL⁻¹. Following overnight incubation, each well of cells was treated with a mixture of 1 µg of SIINFEKL‐STINGΔTM protein, 0.025 µg of cGAMP, and 1 µL of TransIT‐X2 in a total volume of 20 µL OptiMEM media. For vehicle‐free treatment, each well of cells were treated with 5 µg of SIINFEKL‐STINGΔTM protein and 0.125 µg of cGAMP in 20 µL OptiMEM media. Treated cells were incubated for 48 h. ELISA was performed according to the manufacturer's protocol: Mouse CXCL10 ELISA kit (R&D, DY466).

Immunocytochemistry

RAW264.7 and DC2.4 cells were cultured in Millicell EZ chamber slides (Millipore Sigma, Temecula, CA, USA) then treated with Alexa Fluor 488 NHS Ester (Thermo Fisher Scientific, A20000) labelled SIINFEKL‐STINGΔTM protein mixed with cGAMP for 8 h (10 µg of AF488‐SIINFEKL‐STINGΔTM protein and 0.25 µg of cGAMP in 50 µL OptiMEM per well). Cells were washed with PBS for three times, fixed by 4% formaldehyde in PBS for 15 min, permeabilized by 0.1% Triton X‐100 in PBS on ice for 10 min, then stained with Alexa Fluor 647 phalloidin (Thermo Fisher Scientific, A22287) in PBS with 0.05% Tween20 and 1% BSA for 30 min avoiding light followed by three washes in the same buffer. After staining, slides were immersed in DAPI mounting media (Thermo Fisher Scientific, 00‐4959‐52), covered with cover slips and then imaged by an Olympus FV1200 confocal microscope.

Mice

C57BL/6 (BL/6), Balb/c, and C57BL/6‐Tg(TcraTcrb)1100Mjb/J (OT1) mice were purchased from the Jackson laboratory and housed in the MIT Koch Institute animal facility. All mouse studies were performed according to the protocols (protocol numbers: 0815‐078‐18, 0818‐062‐21, and 0821‐052‐24) approved by the MIT Division of Comparative Medicine Committee on Animal Care (CAC). Immunizations were performed on female BL/6 mice 8–12 weeks old. Lymphocytes for ex vivo antigen presentation and flow cytometer SIINFEKL⁺ CD8⁺ T cell control were collected from OT1 mice 8–12 weeks old.

In Vivo Imaging

SIINFEKL‐STINGΔTM, STINGΔTM, and SIINFEKL‐MSA proteins were first labeled with Cy7‐NHS ester (Lumiprobe, 25 020), washed with PBS for six times through Amicon ultra‐4 centrifugal filter unit (Sigma, UFC8003, molecular weight cut off 3 kD), mixed with cGAMP then injected into the tail base of Balb/c mice. Twenty four hours post injection, inguinal lymph nodes, spleens, kidneys, livers, lungs, and hearts of the mice were collected, weighed, and imaged with Xenogen in vivo imaging system (IVIS). Acquisition and analysis of images were performed with Living Image software (Xenogen).

Ex Vivo Antigen Presentation

DC2.4 cells were seeded in 48‐well plates for 200 µL per well with a density of 10⁵ cells mL⁻¹. After overnight incubation, each well of cells were treated with 5 µg OVA, 4 µg SIINFEKL‐STINGΔTM, 0.1 µg cGAMP mixed in 20 µL OptiMEM. As a positive control, SIINFEKL peptides were added to the wells at a working concentration of 0.1–1 µg mL⁻¹. The following day, OT1 lymphocytes were extracted from OT1 mice inguinal lymph nodes and stained with 1 × 10⁻⁶ m CFSE in PBS at room temperature for 20 min. Fresh FBS was added to 10% to stop the reaction and wash once with PBS. Cells were re‐suspended in RPMI with 10 ng µL⁻¹ IL‐2, 50 × 10⁻⁶ m β‐mercaptoethanol, and 0.1 × 10⁻³ m non‐essential amino acids and incubated at 37 °C for 2 h. OT1 lymphocytes were added to each well to have ≈1:10 DC to OT1 cells. After 3 days of co‐culture, cells were washed, blocked with Fc‐blocker (anti‐mouse CD16/32), and stained with anti‐CD8‐APC antibody. Flow cytometric analysis was performed on a BD FACS Celesta flow cytometer.

Mice Immunization and Quantification of Antigen‐Specific T Cells with Intracellular Cytokine Staining and Tetramer Staining

Groups of female BL/6 mice were immunized via tail base injection with 40 µg SIINFEKL‐STINGΔTM or 100 µg SIINFEKL‐MSA mixed with 1 µg cGAMP along with other control groups on days 0 and 14. On day 21, mice blood was collected by cheek bleeding, followed by lysis of red blood cells (Millipore Sigma, R7757) to obtain peripheral blood mononuclear cells (PBMCs).

For intracellular cytokine staining, PBMCs were first stimulated 1 µg mL⁻¹ SIINFEKL peptide in RPMI media with 50 × 10⁻⁶ m β‐mercaptoethanol, 0.67 µL mL⁻¹ GolgiStop and 0.1 × 10⁻³ m non‐essential amino acids and incubated at 37 °C for 4 h. The PBMCs were then treated with Fc blocker followed by viability staining with LIVE/DEAD fixable aqua stain (Thermo Fisher Scientific, L34965) and staining with anti‐CD8 (BioLegend, 100707). The PBMCs were then fixed and permeabilized and stained with anti‐IFN‐γ (BioLegend, 505825) and anti‐TNF‐α (BioLegend, 506107) antibodies, then analyzed on flow cytometer.

For tetramer staining, PBMCs were similarly blocked with Fc blocker and stained with LIVE/DEAD fixable aqua stain, followed by surface staining with anti‐CD8 and H‐2Kb/SIINFEKL tetramer, and then analyzed on flow cytometer.

Quantification of Mouse Serum IL‐6 and TNF‐α level with ELISA

Mouse blood was collected via cheek bleeding 2 h post tail base injection of 100 µg SIINFEKL‐STINGΔTM, 100 µg SIINFEKL‐STINGΔTM + 2.5 µg cGAMP, or 2.5 µg cGAMP. Blood was then centrifuged at 500 × g for 3 min to separate the sera for ELISA detection of IL‐6 and TNF‐α according to the manufacturer's protocol: mouse IL‐6 ELISA kit (Abcam, ab46100) and mouse TNF alpha ELISA kit (ab208348).

Prophylactic Study with MC38‐SIINFEKL Colon Carcinoma Cell Line

MC38‐SIINFEKL cells were obtained from lentiviral transduction of MC38 cells to express SIINFEKL peptide. Groups of BL/6 mice were immunized via tail base injection with 100 µg SIINFEKL‐STINGΔTM mixed with 2.5 µg cGAMP as well as other control groups on days 0 and 14. On day 21, mice were challenged with 1 million MC38‐SIINFEKL cells inoculated subcutaneously in the right hind flank. Tumor volumes were measured every 2–3 days and calculated as (Length × Width²)/2.

Statistical Analysis

All data obtained from experiments were reported without preprocessing. The sample size (n) for all in vitro STING activation and ex vivo antigen presentation assays as well as in vivo and ex vivo imaging is 3, for in vivo vaccination assay followed by SIINFEKL tetramer staining is 5, and for in vivo prophylactic study is 10. Statistical analyses were carried out using GraphPad Prism 5. Data were analyzed with one‐way analysis of variance (ANOVA) followed by Student's t test for statistical significance, and presented as mean ± SEM.

Conflict of Interest

The authors declare no conflict of interest.

Author Contributions

Y.H. and C.H. contributed equally to this work. Y.H. and C.H. designed the experiments. Y.H., C.H., J.L., S.J.F., A.G.B, X.S., M.Y., and S.H. performed the experiments. P.T.H., J.L., D.J.I., and A.M.B. supervised the study. C.H., Y.H., and P.T.H wrote the manuscript.

Supporting information

Supporting Information

He Y., Hong C., Fletcher S. J., Berger A. G., Sun X., Yang M., Huang S., Belcher A. M., Irvine D. J., Li J., Hammond P. T., Peptide‐Based Cancer Vaccine Delivery via the STINGΔTM‐cGAMP Complex. Adv. Healthcare Mater. 2022, 11, 2200905. 10.1002/adhm.202200905

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.