STING Protein‐Based In Situ Vaccine Synergizes CD4⁺ T, CD8⁺ T, and NK Cells for Tumor Eradication

Abstract

Stimulator of interferon genes (STING) signaling is a promising target in cancer immunotherapy, with many ongoing clinical studies in combination with immune checkpoint blockade (ICB). Existing STING‐based therapies largely focus on activating CD8⁺ T cell or NK cell‐mediated cytotoxicity, while the role of CD4⁺ T cells in STING signaling has yet to be extensively studied in vivo. Here, a distinct CD4‐mediated, protein‐based combination therapy of STING and ICB as an in situ vaccine, is reported. The treatment eliminates subcutaneous MC38 and YUMM1.7 tumors in 70–100% of mice and protected all cured mice against rechallenge. Mechanistic studies reveal a robust T_H1 polarization and suppression of T_reg of CD4⁺ T cells, followed by an effective collaboration of CD4⁺ T, CD8⁺ T, and NK cells to eliminate tumors. Finally, the potential to overcome host STING deficiency by significantly decreasing MC38 tumor burden in STING KO mice is demonstrated, addressing the translational challenge for the 19% of human population with loss‐of‐function STING variants.

The stimulator of interferon genes protein is re‐purposed as an in situ vaccine that induces a robust T_H1 polarization and suppression of T_reg of CD4⁺ T cells, followed by an effective collaboration of CD4⁺ T, CD8⁺ T, and NK cells to eliminate tumors.

1. Introduction

Immune checkpoint blockade (ICB) has revolutionized the landscape of cancer immunotherapy over the past decades as one of the most promising treatment options for diseases such as advanced melanoma,^{[ 1} , ² , ^{3 ]} colorectal cancer,^{[ 4} , ^{5 ]} and non‐small cell lung cancer.^{[ 6} , ^{7 ]} However, response rates toward ICB treatment varies widely across individuals and cancer types, leaving only a minority of patients benefiting from significant tumor regression and long‐term protection from recurrence.^{[ 8} , ⁹ , ^{10 ]} While the exact mechanism of resistance is not fully understood, immunosuppression on multiple levels—genetic (e.g., impaired human leukocyte antigen HLA class I and mutation of the phosphage and tensin homolog PTEN gene), protein (e.g., vascular endothelial growth factor VEGF, interleukin IL‐10, and transforming growth factor TGFβ), and cellular (e.g., inadequate antigen presentation and tumor infiltration of lymphocytes)—has been reported to be closely correlated to clinical outcome,^{[ 10} , ¹¹ , ^{12 ]} suggesting the need for a more comprehensive approach in treatment.

Researchers have sought to address the above challenge through the development of combination therapies that engage a variety of innate and adaptive immune cells to orchestrate an effective immune response.^{[ 13} , ¹⁴ , ^{15 ]} These include the use of multiple checkpoint inhibitors such as Nivolumab and Ipilimumab,^{[ 15} , ¹⁶ , ^{17 ]} the combination of ICB with adoptive T cell therapy,^{[ 18} , ¹⁹ , ^{20 ]} as well as the combination of ICB with innate immune system agonists such as Toll‐like receptor 9 (TLR9) agonist.^{[ 21} , ^{22 ]} Of these combinations, the stimulator of interferon genes (STING) pathway has emerged as a highly promising immunostimulatory target, with numerous preclinical studies demonstrating efficacy when delivered alongside ICB^{[ 23} , ^{24 ]} and multiple ongoing clinical trials for the treatment of solid tumors via intratumoral injections.^{[ 25} , ^{26 ]}

STING signaling has been well established to drive dendritic cell maturation,^{[ 27 ]} M2 to M1 macrophage polarization,^{[ 28 ]} CD8⁺ T cell priming,^{[ 29} , ^{30 ]} and in recent years, natural killer (NK) cell activation.^{[ 31} , ³² , ^{33 ]} It has also been reported to polarize CD4⁺ T cells to the T_H1 phenotype in various prophylactic vaccination studies,^{[ 34} , ³⁵ , ^{36 ]} as well as the T_H9 phenotype when stimulated in vitro, driving anti‐tumoral activities with cytokines such as interferon γ (IFNγ).^{[ 37 ]} However, the potential role of CD4⁺ T cells in the context of STING‐based in situ cancer vaccines has remained comparatively unexplored, even though the aforementioned cell subsets are associated with potent antitumor effects,^{[ 38} , ^{39 ]} employed in clinical studies,^{[ 40 ]} and orchestrate a broad range of immune responses (ranging from the development of CD8⁺ T cell immunity^{[ 41} , ^{42 ]} to achieving direct tumor cytotoxicity). In light of these points, the effect of STING signaling in CD4⁺ T cell polarization and the ensuing activation of other cell types within the tumor microenvironment (TME) remains a highly interesting and clinically relevant question.

Another topic of interest lies in the limited clinical translation of STING therapies. To date, no FDA approval has been achieved for the widespread use of STING‐agonist–based cancer treatments,^{[ 25 ]} though an antibody‐drug conjugate has recently been granted orphan drug designation.^{[ 43 ]} This is primarily due to limitations in safety and efficacy, as prior studies have reported potential adverse effects such as dose‐limited toxicity and a potent increase in serum interferon (IFN) levels.^{[ 44} , ^{45 ]} While intratumoral injections may circumvent issues arising from systemic activation of STING,^{[ 46} , ^{47 ]} one challenge in terms of efficacy remains unresolved—the large human population with potentially defective STING due to genetic mutations. Studies have shown that the HAQ and H232 mutations of human STING, which account for ~30% East Asian and ~10% of European populations, are likely loss‐of‐function variants,^{[ 48} , ⁴⁹ , ^{50 ]} and that delivery of exogenous agonists to HAQ mice result in greatly impaired response.^{[ 51 ]} As existing STING‐targeting cancer immunotherapies rely on the intracellular delivery of STING agonists, the lack of fully functional endogenous STING could significantly abrogate signaling and treatment efficacy.

To this end, we previously developed an approach for cyclic guanosine monophosphate–adenosine monophosphate (cGAMP) delivery by adapting the cytosolic domain of STING protein (STINGΔTM) as a biomimetic carrier. The resulting STINGΔTM‐cGAMP complex was shown in vitro to restore STING signaling in cell lines without STING protein or expressing only HAQ mutant STING, demonstrating its use as a fully functional protein.^{[ 52} , ^{53 ]} In this work, we have engineered a novel fusion protein of ICB nanobody and STINGΔTM, which can then be complexed with cGAMP as an intratumorally‐injected therapeutic. Through depletion studies, we observed that CD4⁺ T cells played a critical role in achieving antitumor immunity, to an extent that no significant differences were observed when compared to the untreated control. Mechanistic studies revealed a remodeling of the TME with proinflammatory cytokines, as well as the polarization of CD4⁺ T cells toward the T_H1 phenotype, followed by NK and CD8⁺ T cells—mediated cytotoxicity toward tumor cells. The ICB‐STINGΔTM‐cGAMP complex was found to effectively eradicate tumors in the MC38 colon carcinoma and the YUMM1.7 melanoma model, inhibit metastasis in the 4T1 breast cancer model, and prevent cancer recurrence in cured mice. Finally, when treating tumors on STING‐knockout mice, the complex was able to achieve significant tumor regression via restored STING signaling with STINGΔTM. These results underscore the use of this combination as a promising and comprehensive mode of cancer immunotherapy, while simultaneously providing new insights regarding the potential benefits of STING‐activated CD4⁺ T cells for clinical translation.

2. Results

2.1. Design and Characterization of the ICB‐ΔTM‐cGAMP Complexes

In engineering the ICB‐STINGΔTM‐cGAMP complex, we fused two types of ICB nanobodies to the STINGΔTM protein (ΔTM)—an anti‐cytotoxic T‐lymphocyte associated protein 4 (CTLA4) nanobody (αCTLA4‐ΔTM), and an anti‐programmed death ligand 1 (PDL1) nanobody (αPDL1‐ΔTM) (Figure  1A). The function of the complexes in restoring STING signaling was then evaluated in a human embryonic kidney (HEK293T) IFN‐luciferase reporter cell line that is STING deficient (Figure 1B), where they were observed to achieve high levels of interferon activity relative to agonist‐only and controls of ΔTM (S365A mutant, which abrogates STING‐mediated IFN signaling). Nanobody activity on the complex was likewise confirmed through an ex vivo cancer cell killing assay, in which SIINFEKL peptide‐expressing MC38 cells were first stimulated with IFNγ to express PDL1 and then co‐cultured with SIINFEKL‐specific CD8⁺ T cells isolated from OT1 mice. Addition of the ICB‐ΔTM‐cGAMP complex resulted in a significant increase in the percentage of dead SIINFEKL‐MC38 cells relative to a control nanobody‐ΔTM targeting an irrelevant target 96G3M and untreated controls, as well as no significant difference when compared to the commercial anti‐PDL1 monoclonal antibody (mAb), demonstrating full functionality of the immune checkpoint blockade (Figure 1C,D). Subsequently, we recorded the biodistribution of the complexes through intratumoral delivery of cyanine 5 (Cy5)‐labeled ICB‐ΔTM‐cGAMP (Figure 1E,F). The complex remains almost exclusively within the tumor with some trafficking to the tumor draining lymph node (tdLN) 24 h post‐intratumoral injection, affirming the low probability of systemic toxicity through this route of administration.

Figure 1.

ICB‐STINGΔTM‐cGAMP in vitro function and in vivo biodistribution. A) Protein homology‐modeling structural prediction of dimerized anti‐PDL1‐STINGΔTM protein complexed with a cGAMP molecule with SWISS modeling. B) HEK293T interferon‐luciferase reporter cells (N = 3) treated with different ICB‐ΔTM‐cGAMP complexes, along with S365AΔTM mutant controls. Luciferase activity was determined 24 h post treatment. C) Ex vivo assay of OT1 CD8+ T cells’ cytotoxicity against SIINFEKL‐MC38 cancer cells, ICB‐ΔTM‐cGAMP complexes along with equal molar amount of anti‐PDL1 antibody were added during the co‐culture of CD8+ T cells and cancer cells, and the cytotoxicity was measured with flow cytometry with representative plot shown in (D). E) Representative IVIS images of F) biodistribution of Cy5‐labeled ICB‐ΔTM‐cGAMP complexes 24 h post intratumoral injections (tdLN for tumor draining lymph node and cLN for contralateral lymph node). Data were analyzed with one‐way ANOVA.

2.2. ICB‐ΔTM‐cGAMP Eradicates Subcutaneous MC38 and YUMM1.7 Tumors with Immune Memory

Next, we evaluated the therapeutic efficacy with two subcutaneous mouse tumor models, MC38 and YUMM1.7. We observed significant antitumor effect and sustained antitumor immune memory in both models following treatment with ICB‐ΔTM‐cGAMP (Figures  2A and  3A). Five groups were included in the MC38 study: the individual ICB‐ΔTM‐cGAMP complexes (αPDL1‐ΔTM‐cGAMP and αCTLA4‐ΔTM‐cGAMP), an equal molar mixture of both complexes (αPDL1&CTLA4‐ΔTM‐cGAMP), a STING signaling‐only control with nanobody targeting the green fluorescent protein (αGFP‐ΔTM‐cGAMP), and an ICB protein‐only control without cGAMP (αPDL1&CTLA4‐ΔTM). One additional group—STING with equal molar quantities of commercial αPDL1 and αCTLA4 mAbs (αGFP‐ΔTM‐cGAMP + αPDL1&CTLA4 mAb)—was included in the YUMM1.7 experiment.

Figure 2.

ICB‐STINGΔTM‐cGAMP effectively eliminates subcutaneous MC38 tumors. A) Groups of C57BL/6 mice (N = 10) were inoculated with 1 million MC38 cells subcutaneously on the hind right flank and then treated intratumorally with four doses of ICB‐ΔTM‐cGAMP complexes along with anti‐GFP‐nanobody and no cGAMP controls. B–E) Tumor growth (B) and survival (C) of mice were monitored, all surviving mice (grey blocks meaning the mouse has been euthanized after its tumor burden exceeded 1000 mm³) were tumor free as shown in (E) and successfully rejected a secondary rechallenge of 0.1 million MC38 cells subcutaneously injected on the hind left flank (D). F) A new experiment was performed to analyze the tumor microenvironment, groups of C57BL/6 mice (N = 5) were inoculated with 1 million MC38 cells subcutaneously then treated intratumorally with three doses of ICB‐ΔTM‐cGAMP complexes along with anti‐GFP‐ΔTM‐cGAMP, no cGAMP, and commercial anti‐PDL1&CTLA4 antibody controls. G) Percentage body weight change of mice since the start of treatment, same figure legend as (I). H) Cytokine/chemokine levels and I) immune cell compositions in the tumors were quantified on Day 17. Tumor volume and body weight change were analyzed with two‐way ANOVA, the survival curve was analyzed with Kaplan–Meier, the remaining plots were analyzed with one‐way ANOVA.

Figure 3.

ICB‐STINGΔTM‐cGAMP effectively eliminates subcutaneous YUMM1.7 tumors. A) Groups of C57BL/6 mice (N = 5 for no cGAMP and αGFP‐ΔTM‐cGAMP groups, N = 8 for αPDL1&CTLA4 mAb group, N = 10 for the remaining three groups) were inoculated with 1 million YUMM1.7 cells subcutaneously on the hind right flank and then treated intratumorally with five doses of ICB‐ΔTM‐cGAMP complexes along with controls. B–E) Tumor growth (B) and survival (C) of mice were monitored, all surviving mice (grey blocks meaning the mouse has been euthanized after its tumor burden exceeded 1000 mm³) were tumor free (E) and successfully rejected a secondary rechallenge of 0.1 million YUMM1.7 cells subcutaneously injected on the hind left flank except for one mouse from the αPDL1&CTLA4 mAb group. F) A new experiment was performed to analyze the tumor microenvironment, groups of C57BL/6 mice (N = 5) were inoculated with 1 million YUMM1.7 cells subcutaneously then treated with three doses of ICB‐ΔTM‐cGAMP complexes along with controls. G) Percentage body weight change of mice since the start of treatment, figure legend same as (I,J). H) Cytokine/chemokine levels, I) immune cell compositions, and J) intracellular cytokine staining of NK and NKT cells in the tumors were quantified on Day 15. Tumor volume and body weight change were analyzed with two‐way ANOVA, the survival curve was analyzed with Kaplan–Meier, the remaining plots were analyzed with one‐way ANOVA.

In both tumor models, all combinations of ICB‐ΔTM‐cGAMP were observed to significantly inhibit tumor growth and increase survival relative to STING‐only (αGFP‐ΔTM‐cGAMP) and ICB‐only controls (αPDL1&CTLA4‐ΔTM). Tumors decreased in size and formed black scabs following treatment, which sloughed off and healed without further intervention (Figure S1, Supporting Information). We also noted that therapeutic benefits from enhanced STING signaling typically occurred earlier than those from ICB; for instance, MC38 tumors treated with cGAMP exhibited tumor growth inhibition around Day 10 in comparison to the protein‐only group, while ICB‐ΔTM‐cGAMP formulations only began to outperform STING‐only treatments at Day 14. This effect was corroborated in the YUMM1.7 melanoma model, where significant differences between the ICB‐only group and ICB‐ΔTM‐cGAMP treatments could be observed at Day 12, whereas the benefit of checkpoint blockade was only evident at around Day 15. In particular, the dual blockade group αPDL1&CTLA4‐ΔTM‐cGAMP resulted in the most prominent survival benefit (7/10 and 10/10 for MC38 and YUMM1.7, respectively), more so than that of single blockade groups αPDL1‐ΔTM‐cGAMP and αCTLA4‐ΔTM‐cGAMP (Figures 2B,C and 3B,C; Figures S2 and S4, Supporting Information). All survivors from both studies treated with ICB‐ΔTM‐cGAMP combinations also exhibited 100% tumor rejection upon rechallenge on the opposite hind flank, while αGFP‐ΔTM‐cGAMP + αPDL1&CTLA4 mAb resulted in 75% tumor rejection (Figures 2D and 3D). All survivors remained tumor‐free until the end of the 6 months observation period, with representative images of healed skin provided in Figures 2E and 3E.

To characterize the TME, we performed separate treatment experiments for both MC38 and YUMM1.7 models, resecting the tumors from mice after the third dose for analysis (Figures 2F and 3F). Body weight was also monitored over the course of the treatments. ICB‐ΔTM‐cGAMP therapy led to a maximum weight loss of ≈5%, though this difference was insignificant compared to untreated groups (Figures 2G and 3G). Flow cytometry analysis of the excised tumors (Figures 2I and 3I; Figure S3, Supporting Information) revealed that ICB‐ΔTM‐cGAMP therapy led to a lower average percentage of tumor associated macrophages (TAMs), which could be beneficial to tumor treatment as they are generally associated with immunosuppression and poor prognosis.^{[ 54 ]} We also observed upregulated pro‐inflammatory cytokines in the TME: IL‐2, macrophage inflammatory protein‐2 (MIP‐2), IL‐15, and granulocyte‐colony stimulating factor (G‐CSF) for MC38, and IL‐2, IL‐12, IL‐15, G‐CSF, and tumor necrosis factor α (TNFα) for YUMM1.7 (Figures 2H and 3H). Finally, we characterized the NK and NKT cells in the YUMM1.7 model and found that ICB‐ΔTM‐cGAMP therapy could effectively activate the cytotoxicity of NK cells, as demonstrated by the upregulated levels of granzyme B and perforin in NK cells (Figure  3J and Figure S5, Supporting Information).

2.3. ICB‐ΔTM‐cGAMP Engages CD4⁺ T, NK, and CD8⁺ T Cells in a Multi‐Pronged Immune Response

The therapeutic mechanism of ICB‐ΔTM‐cGAMP was observed to rely significantly on CD4⁺ T, NK, and CD8⁺ T cell‐mediated pathways. We performed an immune cell depletion study with YUMM1.7 model (Figure  4A and Figure S6, Supporting Information), where the depletion of NK and NKT cells with αNK1.1 antibodies initially resulted in a steep increase in tumor growth. The trend persisted up to 2 weeks post‐tumor inoculation before plateauing, just as tumors on CD8⁺ T cell‐depleted mice began to drastically increase in size (Figure 4B and Figure S7, Supporting Information).

Figure 4.

Cellular pathways governing ICB‐STINGΔTM‐cGAMP therapeutic efficacy. A) Groups of C57BL/6 mice (N = 10) were each depleted of macrophages, CD4⁺ T cells, CD8⁺ T cells, NK cells with intraperitoneal antibody injections, along with two groups with no depletion (untreated N = 10, treated N = 8), then inoculated with YUMM1.7 tumors and treated intratumorally with two doses of αPDL1&CTLA4‐ΔTM‐cGAMP. B) Tumor growth and C) survival of mice were monitored. D) On day 23 post tumor inoculation, an IFNɣ ELISPOT assay was performed with mice PBMCs against X‐ray irradiated YUMM1.7 cells, with representative wells of each group showing in (E). SFU represents spot forming unit. F) Representative flow cytometry plots (group αCTLA4‐ΔTM + cGAMP) from cellular colocalization study: Cy5‐labeled ICB‐ΔTM proteins complexed with cGAMP were injected into YUMM1.7 tumors on C57BL/6 mice. G) 24 h post injection, tumors were harvested, digested, and analyzed with flow cytometry to show colocalization of cell types with ICB‐ΔTM‐cGAMP. H) Ex vivo C57BL/6 splenocyte stimulation with αPDL1&CTLA4‐ΔTM‐cGAMP complex followed by intracellular cytokine staining for IFNɣ in CD4⁺ T cells. I) Isolation of CD4⁺ T cells from splenocytes. J) Ex vivo isolated CD4⁺ T cell stimulation with αPDL1&CTLA4‐ΔTM‐cGAMP complex followed by intracellular cytokine staining for IFNɣ. K) Confocal micrographs of isolated CD4⁺ T cells incubated with Cy5‐labeled αPDL1&CTLA4‐ΔTM‐cGAMP, top and bottom images are representative cells interacting with ICB‐ΔTM‐cGAMP from the same treatment group. Scale bar = 10 µm. L) Proposed cellular mechanism of CD4⁺ T cells as first responders that interact with ICB‐ΔTM‐cGAMP complexes, followed by activation of NK‐mediated cytotoxicity and priming of CD8⁺ T cells with dendritic cells. Tumor antigens released from NK‐ and CD8‐mediated cytotoxicity are captured and presented by dendritic cells to sustain the immunity cycle. Tumor volume and body weight change were analyzed with two‐way ANOVA, the survival curve was analyzed with Kaplan–Meier, the remaining plots were analyzed with one‐way ANOVA.

Despite the initial anti‐tumoral effect of NK cells, the overall survival of the NK‐depleted group indicated that NK cells were not essential for eradication of the tumor. Both NK and macrophage‐depleted groups failed to exhibit any significant differences in overall survival compared to the no‐depletion control, with all groups achieving a survival rate of 25–30% (Figure 4C). In contrast, CD4⁺ and CD8⁺ T cells appeared to be indispensable for antitumor immunity, as their depletions fully abolished therapeutic efficacy from the treatment. Some tumor inhibition was observed for the CD8‐depleted group during Day 10–15 of treatment, coinciding with the period where NK‐depleted mice experienced the steepest increase in tumor growth, suggesting that NK cells were responsible for the initial suppression of tumor growth in CD8‐depleted mice. However, this effect was transient: after Day 15, we observed a drastic increase in tumor volume in the CD8‐depleted group due to lack of antitumoral immunity, coinciding with decreased tumor growth in NK‐depleted mice. Altogether, these results provide evidence for the temporal progression of first NK and then CD8⁺ T cell‐mediated antitumor immunity in response to treatment.

Most notably, the depletion of CD4⁺ T cells proved to be critical to the therapeutic effect of ICB‐ΔTM‐cGAMP at all stages of the study, resulting in tumor progression similar to that of untreated controls. In comparison with NK‐ and CD8‐depleted groups, CD4‐depleted group exhibited no therapeutic benefit from either NK or CD8⁺ T cells throughout the study. To investigate this phenomenon, we sought to quantify the amount of tumor antigen‐specific immune cells from peripheral blood mononuclear cells (PBMCs) via IFNγ enzyme‐linked immunosorbent spot (ELISPOT) and observed significantly compromised adaptive immunity in CD4‐depleted mice, worse even than that of the untreated group (Figure 4D,E).

The aforementioned results led us to speculate that the engagement of CD4⁺ T cells with ICB‐ΔTM‐cGAMP occurs at an early timepoint. To verify this hypothesis, we sought to characterize tumor immune cells 24 h post‐intratumoral injection of Cy5‐labeled ICB‐ΔTM‐cGAMP. Flow cytometry analysis of the excised tumors demonstrated the cellular targeting of ICB nanobodies—αCTLA4 toward CD4⁺ T cells and αPDL1 toward macrophages. Overall, the highest percentage Cy5⁺ cells were the CD4⁺ T cells (Figure 4F,G), indicating close interaction with ICB‐ΔTM‐cGAMP. We also observed CD4⁺ T cell T_H1 polarization as well as NK‐mediated cytotoxicity in treatment groups with cGAMP, demonstrating the early anti‐tumoral immunity from STING signaling (Figure S8, Supporting Information). To further explore this interaction, we treated the splenocytes of C57BL/6 mice with αPDL1&CTLA4‐ΔTM‐cGAMP ex vivo and analyzed the T_H1 polarization of CD4⁺ T cells 24 h post‐treatment. We observed that higher concentrations of αPDL1&CTLA4‐ΔTM‐cGAMP led to higher polarization toward the T_H1 phenotype (Figure 4H) and sought next to decouple this effect from other immune cells by repeating the experiment on CD4⁺ T cells isolated from splenocytes (Figure 4I,J). This was also found to result in enhanced T_H1 polarization, with confocal micrographs corroborating increased interaction (Figure 4K).

At this juncture, a therapeutic mechanism for the intratumoral delivery of ICB‐ΔTM‐cGAMP may be proposed. During the first 2 weeks, CD4⁺ T and NK cells are major contributors to tumor regression. The injected ICB‐ΔTM‐cGAMP primarily interacts with CD4⁺ T cells (Figure 4F), giving rise to T_H1 polarization (Figure S9, Supporting Information), a subtype known for its supporting role in inducing cellular immunity, including activating and sustaining NK cells via IL‐2 secretion.^{[ 55 ]} On the other hand, while NK cells did not directly interact with ICB‐ΔTM‐cGAMP, our intracellular staining of granzyme B and perforin suggested a robust activation of NK cytotoxicity (Figure S5, Supporting Information). We therefore hypothesize that CD4⁺ T cells play a role in stimulating NK‐mediated anti‐tumor immunity, though the specific mechanism remains to be uncovered. Concurrently, CD4⁺ T cells aid in antigen presentation and in priming CD8⁺ T cells, which leads to CD8⁺ T cell‐mediated antitumor immunity at 2–3 weeks post‐injection and CD8⁺ memory T cell–based protection against tumor recurrence (observed in YUMM1.7 and MC38 tumor models). The process has been illustrated in Figure 4L.

2.4. ICB‐ΔTM‐cGAMP Inhibits Metastasis in 4T1 Model

To assess the systemic immune protection afforded by the intratumoral delivery of ICB‐ΔTM‐cGAMP, we next treated the 4T1 metastatic breast cancer model (Figure  5A). When no significant decrease in tumor volume was observed following three intratumoral doses of αPDL1&CTLA4‐ΔTM‐cGAMP, the primary tumor was surgically resected. The rate of lung metastasis was then monitored indirectly via post‐op survival, where αPDL1&CTLA4‐ΔTM‐cGAMP–treated mice survived for significantly longer than those dosed with STING‐only or ICB‐only controls (Figure 5B,C). Analysis of resected tumors revealed a significant decrease in TAMs relative to untreated controls in the αPDL1&CTLA4‐ΔTM‐cGAMP treatment group. The resected tumors also exhibited significant T_H1 polarization of CD4⁺ T cells and suppression of T_reg (Figure 5G and Figure S9, Supporting Information), while cytokine analysis of the tumors suggested increased inflammation in the TME (as evidenced by higher levels of G‐CSF, granulocyte macrophage‐colony stimulating factor GM‐CSF, IL‐1α, and IL‐12). These results underscore the critical role of CD4⁺ T cells in the treatment, as well as its benefit in both metastatic and non‐metastatic tumor models.

Figure 5.

ICB‐STINGΔTM‐cGAMP controls metastasis in 4T1 model. A) Groups of Balb/c mice (N = 10) were inoculated with 1 million 4T1 cells in the mammary fat pad and then treated intratumorally with three doses of ICB‐ΔTM‐cGAMP complexes along with controls; on Day 15 the tumors were surgically removed. B) Tumor growth and C) survival of mice were monitored, the surviving mouse was tumor free. D) Percentage body weight change of mice since the start of treatment. E) Cytokine/chemokine levels, F) immune cell compositions, and G) CD4⁺ T cell phenotyping from excised tumors on Day 15. Tumor volume and body weight change were analyzed with two‐way ANOVA, the survival curve was analyzed with Kaplan–Meier, the remaining plots were analyzed with one‐way ANOVA.

2.5. ICB‐ΔTM‐cGAMP Restores STING Signaling in STING Deficient Host

Finally, we explored the possibility of restoring STING signaling with the functional STINGΔTM carrier by treating MC38 tumors on STING KO mice with αPDL1&CTLA4‐ΔTM‐cGAMP. To decouple immune responses against this protein complex not induced via the STING pathway, we included αPDL1&CTLA4‐S365AΔTM‐cGAMP as a control (Figure  6A) and increased the dosing frequency to account for lower therapeutic efficacy relative to immunocompetent mice. After four doses, we were able to observe a significant reduction in tumor volume relative to the S365A control, indicating that this result is due entirely to the STING‐activating functionality of the STINGΔTM carrier (Figure 6B). This was further confirmed through analysis of the cytokine profile in resected tumors, where we observed upregulated IL‐12, IL‐15, IL‐13, G‐CSF, TNFα, keratinocyte chemoattractant (KC), etc. This was consistent with the cytokine profiles of the MC38, YUMM1.7, and 4T1 models (Figure 6C), and likewise with the general profiles of STING activation in literature.^{[ 33 ]}

Figure 6.

ICB‐STINGΔTM‐cGAMP demonstrates the potential to overcome host STING deficiency. A) Groups of STING KO mice (N = 8) were inoculated with 1 million MC38 cells subcutaneously in the hind right flank and then treated intratumorally with four doses of αPDL1&CTLA4‐ΔTM‐cGAMP complexes along with controls. B) Shows the tumor growth curve. On Day 12 the tumors were removed for C) cytokine/chemokine analysis. Tumor volume change was analyzed with two‐way ANOVA.

3. Discussion

Potent immunostimulation as well as effective approaches to reverse immunosuppression and capture tumor cells that evade immunosurveillance remains a challenge in clinical cancer immunotherapy,^{[ 56 ]} as evidenced by the partial response observed in numerous STING and ICB‐based therapies evaluated in clinical trials (NCT03172936, NCT03010176, and NCT02675439).^{[ 57} , ⁵⁸ , ^{59 ]} As this process requires the simultaneous engagement of numerous cell types and cytokines, harnessing and synergizing the function of a broad range of immune cells are of paramount importance. For instance, NK cells guard the gaps in tumor elimination where CD8⁺ T cells fall short, targeting cells with low neoantigen burden or diminished MHC‐I expression that escape CD8⁺ T cells,^{[ 60 ]} while T_H1‐polarized CD4⁺ T cells aid in the priming of CD8⁺ T cells and collaborate with them at the tumor site to maintain an effective anti‐tumoral response.^{[ 61 ]}

At present, research on the mechanisms of various ICB therapies have been largely centered on dendritic cells and CD8⁺ T cells.^{[ 42} , ^{62 ]} While other work has explored the use of STING to induce NK‐mediated cytotoxicity for tumor cells with low antigen burden or diminished MHC‐I expression, CD4⁺ T cells have only recently been discovered to be important for the success of ICB treatment.^{[ 61} , ^{63 ]} Existing work on the effect of the STING pathway on CD4⁺ T cells has been restricted to either in vitro studies or prophylactic vaccination;^{[ 34} , ³⁷ , ⁶⁴ , ^{65 ]} moreover, few studies have characterized the role of STING signaling on CD4⁺ T cells in the specific context of ICB treatment.

The cancer immunity cycle has traditionally been depicted as occurring between dendritic cells, which prime CD8⁺ T cells with tumor antigens, resulting in the elimination of tumor cells and the generation of antigens in self‐sustaining loop.^{[ 66 ]} In recent years, NK cells have entered the picture, collaborating alongside CD8⁺ T cells.^{[ 67 ]} Here, we engineered a new strategy to effectively boost the cancer immunity cycle, elucidating the critical role of CD4⁺ T cells as an initiator for NK cell tumor elimination^{[ 68 ]} and an aide to CD8⁺ T cell priming and memory^{[ 69 ]} in STING‐mediated antitumor immunity. This is a highly salient observation, as the role of CD4⁺ T cells in cancer immunotherapy has been reported to be bidirectional. Despite evidence that CD4⁺ T cells are fundamentally important in supporting anti‐tumoral immunity,^{[ 70 ]} many studies over the past 2 decades have shown that CD4⁺ T cell depletion results in improved therapeutic outcomes, including a wide range of therapies such as adoptive T cell transfer,^{[ 71} , ^{72 ]} inflammatory cytokines,^{[ 73} , ^{74 ]} and ICB antibody treatments.^{[ 75 ]} There has even been a human clinical trial intravenously injecting anti‐CD4 depletion antibodies to treat solid tumors, where tumor shrinkage was observed in 5/11 patients over 3 months of observation.^{[ 76 ]} In these therapies, the presence of CD4⁺ T cells did not positively contribute to antitumor immunity, which was attributed to the immunosuppressive effects of T_reg and T_H2 outweighing that of T_H1.

While the positive role of CD4⁺ T cells has not been extensively investigated in many immunotherapies, recent studies have suggested that significant potential benefit could be exploited from CD4⁺ T cells if properly activated,^{[ 61 ]} with some therapies undergoing clinical studies (NCT04444622).^{[ 40 ]} In a recent Bacillus Calmette–Guérin (BCG) immunotherapy in mice, the authors reported a robust CD4‐dependent tumor‐specific immunity through IFNγ signaling, and that depletion of CD4⁺ T cells abolished the efficacy of the treatment.^{[ 77 ]} In our study, we observed similar results regarding the depletion of CD4⁺ T cells. Additionally, we found that CD4⁺ T cells could directly interact with our STING signaling complex ex vivo (Figure 4H–K). This resulted in effective suppression of the undesirable T_reg within the TME while achieving ≈30% T_H1 polarization of CD4⁺ T cells in vivo (Figure 5G and Figure S9, Supporting Information). While we recognize the reported toxicity and impairment of cellular function of T cells due to direct STING activation,^{[ 78} , ⁷⁹ , ^{80 ]} the overall role of CD4⁺ T cells in our treatment as well as in the BCG therapy^{[ 77 ]} has been demonstrated to be indispensably beneficial. Future investigations on the mechanism of CD4⁺ T cell STING signaling might decouple the undesirable side effects from the benefits of T_H1‐mediated anti‐tumoral effects and thus provide important insight for effective drug delivery strategies. Furthermore, this approach results in the suppression of TAMs, as well as the generation of multiple pro‐inflammatory cytokines. Altogether, our proposed therapy initiates a strong, collaborative response among various types of immune cells, with positive feedback loops to reverse immunosuppression and sustain antitumoral immunity.

Following initiation of the innate immune system via STING‐mediated CD4 T_H1‐polarization and NK‐mediated cytotoxicity, the ICB‐ΔTM‐cGAMP complex then acts upon the adaptive immune system through ICB. We sought to leverage the synergy of PDL1 and CTLA4 dual blockade and succeeded in eradicating both immunologically “cold” (YUMM 1.7^{[ 81 ]}) and “hot” (MC38^{[ 82 ]}) tumors, gaining long‐term protection against recurrence through memory T cells. When administrated as a neoadjuvant for the metastatic 4T1 model—a model recalcitrant to immunotherapy—our approach likewise resulted in prolonged post‐op survival. These results demonstrate this system's utility as a promising modular platform for future incorporation of other ICB or tumor‐targeting nanobodies, such as anti‐T cell immunoreceptor with immunoglobulin and ITIM domain (TIGIT) and anti‐natural killer group 2 member A (NGK2A) for NK cells, anti‐P‐selectin glycoprotein ligand‐1 (PSGL‐1) for CD4⁺ T cells, and anti‐fibronectin extra type III module B (EIIIB) for tumor extracellular matrices,^{[ 83 ]} allowing additional flexibility in treating a wide range of cancers.

One remaining challenge to clinical translation lies in host STING deficiency and the limited options to restore STING signaling in affected individuals. Previously, we demonstrated that STING signaling could be restored in vitro with the STING‐deficient HEK293T cell line^{[ 52 ]} and hypothesized that STING signaling could potentially be restored in vivo through a similar approach. Our present system provides a tangible solution by eliciting significant antitumoral immunity via STING signaling in STING KO mice, both in terms of tumor burden and in the production of pro‐inflammatory cytokines within the TME. Although the efficacy of this treatment falls short of its performance in STING‐competent hosts, there are several factors that may be tuned for future work, such as dosage, timing of administration, and other types of nanobodies for immune cell engagement. Ultimately, the aforementioned findings emphasize this platform's immense promise in CD4⁺ T cell activation and overcoming host STING deficiency, offering potential new solutions to improving the clinical outcomes of ICB treatments.

4. Experimental Section

Protein Purification

The expression plasmid was cloned based on the previous pSH200_STINGΔTM plasmid, where STINGΔTM stands for the segment of amino acids 138–378 of mouse STING.^{[ 52 ]} DNA sequences encoding for nanobodies of anti‐PDL1 (PDB ID: 5DXW_A), anti‐CTLA4 (PDB ID: 5E03_A), anti‐GFP (PDB ID: 3OGO_E), and anti‐96G3m (PDB ID: 6 × 07_B) were synthesized by gBlock (IDT) and inserted at the N‐terminus of the sequence encoding for STINGΔTM protein. Mutants such as S365A and R237A/Y239A of the STINGΔTM protein were created via site‐specific mutagenesis. Histidine₆‐tagged nanobody‐STINGΔTM proteins were expressed in Rosetta DE3 Escherichia coli (Rosetta E. coli, Millipore Sigma, Cat#: 70 954). The purification method is slightly modified based on the previous published protocol.^{[ 84 ]} Briefly, Rosetta E. coli transformed with the protein expression plasmid were cultured in 1 L volume of lysogeny broth (LB) at 37 °C until OD600 reaches 0.5–0.8. The culture was then induced with 0.5 mm isopropyl β‐d‐1‐thiogalactopyranoside (IPTG, Millipore Sigma, Cat#: I6758‐10G) at 18 °C for 20 h. Following induction, the cells were centrifuged, lysed, sonicated, and further centrifuged to obtain the cell lysate that contained the desired his‐tagged protein. The lysate was then incubated with cobalt resin (Thermo Fisher Scientific, Cat#: 89 964) followed by elution with 150 mm imidazole (Millipore Sigma, Cat#: I5513) and eventually buffer exchanged to 20 mm Hepes, 150 mm NaCl, 10% glycerol with 1 mm dithiothreitol (Millipore Sigma Cat#: 10 197 777 001), aliquoted, and stored in −80 °C freezer for future use.

In Vitro STING Signaling Activation in 293T Reporter Cells

All animal experiments were carried out with the approval of the MIT division of comparative medicine (No. 0222‐017‐25). The STING deficient HEK293T cell line was used to study the capability of restoring STING signaling in vitro. The assay was based on the previous published protocol.^{[ 84 ]} Briefly, a reporter cell line was generated from transfecting HEK293T cells with pGL4.45 [luc2p/ISRE/Hygro] plasmid (Promega), which was able to express luciferase driven by the interferon‐sensitive response element (ISRE) stimulated by STING signaling. 100 µL of 3 × 10⁵ cells mL⁻¹ HEK293T‐luc2p/ISRE/Hygro cells were seeded in 96‐well plate in Dulbecco's modified eagle's medium (Corning, Cat#: 10‐041‐CV) supplemented with 10% fetal bovine serum (FBS, Gibco, Cat#: 10437‐028), and 1% penicillin/streptomycin (Corning, Cat#: 30‐002‐CI). After an overnight incubation, 0.025 µg cGAMP (Invivogen, Cat#: tlrl‐nacga23‐02) mixed with 1.7 µg ICB‐ΔTM protein and 1.7 µL TransIT‐X2 commercial transfection reagent (Mirus, Cat#: MIR6004) were mixed in 20 µL Opti‐MEM medium (Gibco, Cat#: 31 985 062), incubated for 15 min for protein complex formation before added to the wells. 24 h post treatment, cells were lysed for firefly luciferase assay (Biotium, Cat#: 30075‐2) to quantify the interferon–luciferase activity.

Ex Vivo Cytotoxicity Assay with MC38‐SIINFEKL Cells

MC38‐SIINFEKL cells were created via lenti‐viral vector transduction. pLenti‐CMV‐GFP‐Puro plasmid was obtained from Addgene (Cat#: 17 448). The DNA sequence encoding for SIINFEKL peptide was inserted at the N‐terminus of GFP via mutagenesis. HEK293T cells were transfected with pLenti‐CMV‐SIINFEKL‐GFP‐Puro plasmid with the help of TransIT‐X2 transfection reagent. The culture medium containing lentivirus were collected, filtered through 0.45 µm filter and added to plated MC38 cells with 5 µg mL⁻¹ polybrene (Millipore Sigma, Cat#: H9268), followed by puromycin (Millipore Sigma, Cat#: P9620) selection in the concentration range of 5–20 µg mL⁻¹. The final transduction efficiency was confirmed with flow cytometry based on percentage GFP expression.

To perform the ex vivo cytotoxicity assay of CD8⁺ T cells against cancer cells, MC38‐SIINFEKL cells were first treated with 50 ng mL⁻¹ mouse IFNγ for 24 h to induce PDL1 upregulation, which is confirmed by qPCR to be approximately tenfold of untreated cells’ PDL1 expression level. Sequences of qPCR primers used were mPDL1_F: TGCGGACTACAAGCGAATCACG and mPDL1_R: CTCAGCTTCTGGATAACCCTCG. In parallel, splenocytes were harvested from C57BL/6‐Tg(TcraTcrb)1100Mjb/J (OT1) mice and activated with αCD3/CD28 Dynabeads (Gibco, Cat#: 11456D) at a density of 2 × 10⁶ cells mL⁻¹ Roswell Park Memorial Institute medium (RPMI, Corning, Cat#: 10‐013‐CV) according to the instruction manual. The IFNγ treated MC38‐SIINFEKL cells were then washed and plated at a density of 4 × 10⁵ cells mL⁻¹ with stimulated OT1 splenocytes at a ratio of 1:2 in 1 mL RPMI medium in 12‐well plates for 48 h, with the addition of 17 µg αPDL1‐ΔTM plus 0.25 µg cGAMP, 17 µg α96G3M‐ΔTM plus 0.25 µg cGAMP, 16 µg αPDL1 mono antibody (clone 10F.9G2, BioXCell, Cat#: BE0101), along with an untreated control. After 48 h of coculture, cells were stained with Zombie Aqua live–dead stain (Biolegend, Cat#: 423 101) and analyzed with flow cytometry. MC38 tumor cells were gated with FITC channel due to their GFP expression, and their percentage viability were used to assess the ex vivo PDL1 checkpoint blockade efficiency.

Tumor Inoculation and Intratumoral Treatment

MC38 and YUMM1.7 tumors were inoculated in C57BL/6 or B6(Cg)‐Sting1tm1.2Camb/J (STING knockout) mice by subcutaneously (s.c.) injecting 10⁶ cells suspended in 100 µL sterile phosphate buffered saline (PBS, Lonza, Cat#: 17–516F) into the shaved hind right flank. All C57BL/6 mice were female and purchased at the age of 8 weeks old, tumors were usually inoculated 1–2 weeks after mice arrival at the lab. STING knock‐out mice were bred in house, for tumor treatment studies, each group contained the equal number of male and female mice of the same age (8–12 weeks old). For rechallenge, 10⁵ cells in 100 µL sterile PBS were injected into the hind left flank. The tumor sizes were monitored every 2 days with caliper measurement, volumes calculated as length × width^{[ 2 ]} then divided by 2. Treatment by intratumoral (i.t.) injections started when the average tumor volume reached around 100 mm³. Each i.t. injection consisted of 170 µg of ICB‐ΔTM protein with or without 2.5 µg cGAMP. For ICB antibody controls, each i.t. injection consisted of 80 µg αPDL1 mAb (clone 10F.9G2, BioXCell, Cat#: BE0101) and 80 µg αCTLA4 mAb (clone 9D9, BioXCell, Cat#: BE0164). The humane endpoint tumor burden for euthanasia of all three mouse models was 1000 mm³.

4T1 cells were inoculated in Balb/c mice in the mammary fat pad. A small incision was made next to the fourth nipple, 10⁶ 4T1 cells suspended in 50 µL PBS were injected into the mammary fat pad through the incision, which is subsequently closed with suture (Ethicon, Cat#: R690G). Caliper measurement and intratumoral injections were performed similar to MC38 and YUMM1.7 models. The only caution was that intratumoral injections for 4T1 tumors were done very slowly to achieve a uniform distribution of drug inside the tumor. Fast injections would result in a pouch of liquid seeping outside the tumor.

Immune Cell Depletion in Mice

Depletion of mouse immune cell populations were carried out according to the literature^{[ 73} , ^{85 ]} by intraperitoneally (i.p.) injecting mice with antibodies against mouse CD8α (clone 2.43, BioXCell, Cat#: BP0061, 400 µg twice every week), CD4 (clone GK1.5, BioXCell, Cat#: BP0003‐1, 400 µg twice every week), NK1.1 (clone PK136, BioXCell, Cat#: BE0036, 400 µg twice every week), or CSF1R (clone AFS98, BioXCell, Cat#: BP0213, 300 µg every other day). 36 h post i.p. injection, mice PBMCs were sampled via cheek bleeding and stained for flow cytometry analysis to confirm the depletion efficiency.

Tumor Digestion for Immune Cell Analysis with Flow Cytometry

Approximately 100 mg mouse tumor was resected, weighed, and then minced to small pieces with diameters <1 mm in 500 µL PBS with 1 mg mL⁻¹ collagenase type IV (Gibco, Cat#: 17 104 019). The minced tumor in collagenase was then incubated at 37 °C under shaking for 20 min (MC38 tumors) or 30 min (YUMM1.7 and 4T1 tumors). After incubation, the mixture was diluted with 1 mL PBS, vortexed, and filtered through 70 µm nylon mesh cell strainer (Fisherbrand, Cat#: 22 363 548). In case there were a lot of red blood cells (RBCs), the filtrate was resuspended in 5 mL of RBC lysis buffer (Millipore Sigma, Cat#: R7767‐100 mL), incubated at room temperature for 5 min, then pelleted with centrifugation. The filtrate was then washed twice with PBS, resuspended in 10 mL PBS, and sampled for cell count. For intracellular cytokine staining samples, an additional 4 h 37 °C incubation was performed prior to staining in RPMI supplemented with non‐essential amino acid (NEAA, Gibco, Cat#: 11 140 050) and Golgi inhibitor (GolgiStop, BD Biosciences, Cat#: 554 724).

Afterward, the cells were collected and washed with PBS for FACS staining. 10⁶ cells from each tumor samples were loaded into 96 V‐bottom well plate in 50 µL PBS and first stained with Zombie Aqua live–dead dye, followed by 20 min incubation on ice avoiding light. The cells are then washed with FACS buffer (PBS with 1% BSA and 2 mm EDTA). After washing, each well of cells are stained with Fc‐blocker anti‐mouse CD16/32 (Thermo Fisher Scientific, Cat#: 14‐9161‐73) in 50 µL FACS buffer, followed by 15 min incubation on ice. When the blocking was complete, surface staining antibodies were added to the cells for 30 min incubation on ice, followed by two washes with FACS buffer. Intracellular cytokine staining was then performed with the BD Cytofix/Cytoperm kit (BD Biosciences, Cat #555 028). 100 µL fixation/permeabilization solution was first added to each well followed by 20 min incubation on ice. The cells were then washed twice with Perm/wash buffer and stained with intracellular staining antibodies in 50 µL Perm/wash buffer for 30 min on ice. After staining, the cells were washed twice with Perm/wash buffer, resuspended in FACS buffer, and analyzed with LSR‐II‐Fortessa flow cytometer (BD Biosciences). Data were analyzed with FlowJo software.

Flow cytometry antibodies (all from Biolegend) used were anti‐mouse CD4 (clone GK1.5, Cat#: 100 434), CD3 (clone 17A2, Cat#: 100 204 and 100 232), CD8α (clone 53–6.7, Cat#: 100 707), F4/80 (clone BM8, Cat#: 123 116 and 123 135), NK1.1 (clone S17016D, Cat#: 156 514), NKp46 (clone 29A1.4, Cat#: 137 618), Ly6C (clone HK1.4, Cat#: 128 031), Ly6G (clone 1A8, Cat#: 127 645), CD45 (clone 30‐F11, Cat#: 103 128), CD11b (clone M1/70, Cat#: 101 241), IFNγ (clone XMG1.2, Cat#: 505 813), IL4 (clone 11B11, Cat#: 504 133), PU.1 (clone 7C2C34, Cat#: 681 307), IL17 (clone TC11‐18H10.1, Cat#: 506 915), FoxP3 (clone MF‐14, Cat#: 126 419), granzyme B (clone GB11, Cat#: 515 403), and Perforin (clone S16009B, Cat#: 154 404).

Tumor Protein Extraction for Cytokine Analysis

Approximately 50 mg mouse tumor was resected and grinded with microcentrifuge tube pestles in T‐PER buffer (Thermo Fisher Scientific, Cat#: 78 510) with protease inhibitor (Thermo Fisher Scientific, Cat#: 78 425). The protein extraction solutions were then centrifuged at 14000 × g for 15 min. Supernatant protein concentrations were quantified with Nanodrop based on absorption at 280 nm and were all diluted to 5 mg mL⁻¹. Samples were then frozen and shipped to Eve Technology for multiplex assay for the cytokine concentrations. Heat maps of the protein levels were plotted as log₂ (fold change of untreated groups), with hierarchical clustering based on one minus Pearson correlation with complete linkage method performed with Morpheus software (Morpheus, https://software.broadinstitute.org/morpheus).

Splenocyte CD4⁺ T Cell Isolation and Phenotyping

Spleen from C57BL/6 mice was first grinded and filtered through 70 µm cell strainer and washed once by RBC lysis buffer. The cells were then washed and resuspended in FACS buffer. EasySep Mouse CD4⁺ T cell isolation kit (Stemcell Technology, Cat#: 19 852) was used to isolate CD4⁺ T cells from splenocytes. Splenocytes were re‐suspended in PBS containing 2% FBS and 1 mm EDTA at the concentration of 10⁸ cells mL⁻¹. Per mL of the splenocytes, 50 µL rat serum, and 50 µL isolation cocktail were added followed by a 10 min incubation at room temperature. 75 µL of vortexed magnetic beads Rapidspheres was added per mL of sample, mixed and incubated for 2.5 min. Then buffer was added to the sample to top it up to 2.5 mL, before placed in the magnet (EasySep) for 2.5 min. Afterward, the separated CD4⁺ T cells were poured out in one continuous motion. The separation effect was verified with flow cytometry with anti‐CD4 surface staining. For phenotyping, the isolated CD4⁺ T cells were then treated with ICB‐ΔTM protein mixed with cGAMP in RPMI media with NEAA for 8 h (GolgiStop added for the last 4 h) before staining for flow cytometry analysis as described in the tumor cell FACS section.

Immunocytochemistry

CD4⁺ T cells treated with Cy5‐labeled ICB‐ΔTM protein mixed with cGAMP were collected in microcentrifugation tubes along with untreated controls. The cells were washed three times with PBS, fixed with PBS containing 4% formaldehyde (Millipore Sigma, Cat#: 47608‐250ML‐F) for 15 min on ice, then permeabilized by PBS containing 0.1% Triton X‐100 (Millipore Sigma, Cat#: T8787) on ice for 10 min. Afterward, cells were washed by PBS with 0.05% Tween20 (Millipore Sigma, Cat#: P9416) and 1% BSA, and stained with Alexa Flour 488 phalloidin (Invitrogen, Cat#: A12379) and Hoechst dye (Thermo Fisher Scientific, Cat#: 62 249) in the same buffer for 30 min avoiding light. The cells were then washed three times with PBS containing 0.05% Tween20 and 1% BSA and loaded onto microscope slides with anti‐fade mounting media and covered with cover slips. Images were acquired with Olympus FV1200 confocal microscope and analyzed with ImageJ software.

Organ Biodistribution

C57BL/6 mice inoculated with YUMM1.7 tumor were injected i.t. with 170 µg Cy5‐labeled anti‐GFP/PDL1/CTLA4‐ΔTM protein mixed with 2.5 µg cGAMP in 25 µL PBS. 24 h post treatment, mice were sacrificed and organs plus tumors and both inguinal lymph nodes were collected for fluorescent imaging with the In Vivo Imaging System (IVIS, Xenogen). Data analysis was performed with the Living Image software (Xenogen).

IFNγ ELISPOT

PBMCs were obtained from cheek bleeding of the immune cell depleted mice collected with EDTA‐coated collection tubes (Greiner Bio‐one, K3EDTA Cat#: 450 530). Whole blood was washed twice with RBC lysis buffer to obtain PBMCs. ELISPOT was performed according to the kit instruction manual (BD ELISPOT reagent kit). Briefly, ELISPOT plate was first coated with anti‐IFNγ capture antibodies overnight. The following day, the plate was washed and blocked. Equal number of PBMCs from each mouse were co‐cultured with X‐ray irradiated YUMM1.7 cells in the plate overnight. On the third day, cell suspension was aspirated, wells were incubated with detection antibody, streptavidin‐HRP and substrate solution with five washes between every incubation. Spots were enumerated with an ELISPOT plate reader.

Statistical Analysis

Data were analyzed with Prism GraphPad: tumor volume and body weight change were analyzed with two‐way ANOVA, the survival curve was analyzed with Kaplan–Meier, and the remaining plots were analyzed with one‐way ANOVA. Outlier was analyzed and removed with the Grubb's test.

Conflict of Interest

The authors declare no conflict of interest.

Author Contributions

Y.H. and A.M.B. conceived the idea and designed the study. Y.H. and C.H. performed the experiments. S.H., J.A.K., G.C., I.S.P., and J.C.S. assisted with the experiments. C.H., Y.H. and A.M.B. analyzed the data and wrote the paper. A.M.B. and P.T.H. supervised the study.

Supporting information

Supporting Information

He Y., Hong C., Huang S., Kaskow J. A., Covarrubias G., Pires I. S., Sacane J. C., Hammond P. T., Belcher A. M., STING Protein‐Based In Situ Vaccine Synergizes CD4⁺ T, CD8⁺ T, and NK Cells for Tumor Eradication. Adv. Healthcare Mater. 2023, 12, 2300688. 10.1002/adhm.202300688

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.