A Thermosensitive Bi-Adjuvant Hydrogel Triggers Epitope Spreading to Promote the Anti-Tumor Efficacy of Frameshift Neoantigens

Abstract

Tumor-specific frameshift mutations encoding peptides (FSPs) are highly immunogenic neoantigens for personalized cancer immunotherapy, while their clinical efficacy is limited by immunosuppressive tumor microenvironment (TME) and self-tolerance. Here, a thermosensitive hydrogel (FSP-RZ-BPH) delivering dual adjuvants R848 (TLR7/8 agonist) + Zn²⁺ (cGAS-STING agonist) is designed to promote the efficacy of FSPs on murine forestomach cancer (MFC). After peritumoral injection, FSP-RZ-BPH behaves as pH-responsive sustained drug release at sites near the tumor to effectively transform the immunosuppressive TME into an inflammatory type. FSP-RZ-BPH orchestrates innate and adaptive immunity to activate dendritic cells in tumor-draining lymph nodes and increase the number of FSPs-reactive effector memory T cells (T_EM) in tumor by 2.9 folds. More importantly, these T_EM also exhibit memory responses to nonvaccinated neoantigens on MFC. This epitope spreading effect contributes to reduce self-tolerance to maintain long-lasting anti-tumor immunity. In MFC suppressive model, FSP-RZ-BPH achieves 84.8% tumor inhibition rate and prolongs the survival of tumor-bearing mice with 57.1% complete response rate. As a preventive tumor vaccine, FSP-RZ-BPH can also significantly delay tumor growth. Overall, the work identifies frameshift MFC neoantigens for the first time and demonstrates the thermosensitive bi-adjuvant hydrogel as an effective strategy to boost bystander anti-tumor responses of frameshift neoantigens.

Keywords: TLR7/8 agonist; cGAS‐STING agonist; frameshift neoantigens; thermosensitive hydrogel; tumor immunotherapy.

Scheme 1

Schematic illustration of the thermosensitive bi‐adjuvant hydrogel (FSP‐RZ‐BPH). a) Formation of FSP‐RZ‐BPH. The schematic diagram shows that R848, Zn²⁺ and FSPs are separately loaded on the BSA‐PEG to form a hydrogel skeleton, and its aqueous solution can form a hydrogel at 37 °C. b) Mechanism of synergistic proinflammatory effect of R848 + Zn²⁺ (RZ). R848 is a TLR7/8 agonist, which inhibits the degradation of STING by activating NF‐κB molecules, thereby promoting the Zn²⁺‐activated cGAS‐STING pathway. c) Schematic diagram to illustrate the immune response induced by FSP‐RZ‐BPH in vivo. Combined application of RZ can not only directly activate intratumoral reactive oxygen species (ROS), but also have pro‐inflammatory effects on DCs, T cells, NK cells and macrophages. Activation of DCs promotes the presentation of FSPs, generating a large number of effector T cells in the lymph nodes, which eventually leads to the increase of tumor‐infiltrating NRTs to kill tumor cells. In particular, sufficient tumor neoantigens released by ROS‐induced tumor cell death and the sustained immune activation of FSP‐RZ‐BPH endows T cells epitope spreading to maintain long‐term anti‐tumor immune responses.

Figure 1

Identification of FSPs and the pro‐inflammatory effect of R848 + Zn²⁺ (RZ). a) Amino acid sequence of FSPs and their affinity to H‐2K^K. b) Concentration of IFN‐γ secreted by splenocytes of 615 mice stimulated by FSPs or wild‐type peptides (WTPs). NS represented normal saline, and conA represented concanavalin A. c) Representative flow cytometry images of mature DCs (mDCs, CD11c⁺CD80⁺CD86⁺) after co‐incubation with RZ in vitro for 48 h. d) The percentage of mDCs (n = 3). e) Representative flow cytometry images of MHC I expression on DCs after co‐incubation with RZ in vitro for 48 h. f) The levels of MHC I expression on DCs (n = 3). g, Representative flow cytometry images of MHC II expression on DCs after co‐incubation with RZ in vitro for 48 h. h) The levels of MHC II expression on DCs (n = 3). i) Confocal images of DCs after co‐incubation with FSP‐Cy5 and RZ in vitro for 48 h (The scale bar is 10 µm). j) The levels of FSP‐Cy5 expression on DCs after co‐incubation with FSP‐Cy5 and RZ in vitro for 48 h (n = 3). k) Representative flow cytometry images of CD8⁺CD25⁺ T cells and CD8⁺CD69⁺ T cells after co‐incubation with RZ in vitro for 48 h. l, The percentage CD8⁺CD25⁺ T cells and CD8⁺CD69⁺ T cells (n = 3). The error bars represented mean ± SEM. P‐values were calculated by two‐tailed unpaired Student's t‐tests. ns represented p > 0.05.

Figure 2

Construction and characterization of FSP‐RZ‐BPH. a) Schematic diagram to show the construction of FSP‐RZ‐BPH. b) HPLC spectra of FSP, R848, BSA‐PEG and FSP‐RZ‐BPH. c) Gelation time of BSA‐PEG hydrogel in different concentration of BSA‐PEG and Zn²⁺ (n = 3). d) Formation of 10% BSA‐PEG hydrogel (BPH) and its scanning electron microscope (SEM) images. e) Formation of FSP‐RZ‐BPH and its scanning SEM images. f) The C, N, O and Zn element mapping in FSP‐RZ‐BPH by SEM with energy‐dispersive X‐ray spectroscopy (SEM‐EDS). g) EDS spectrum of SEM‐EDS mapping for FSP‐RZ‐BPH. h) Computed tomography (CT) images of 615 mice on D0, D30, D60 and D90 after injecting 16% BSA‐PEG hydrogel (16% BPH) or FSP‐RZP‐BPH into the lower right abdomen. The hydrogel was circled in red.

Figure 3

RZ‐BPH activates local immune responses with pH‐responsive drug release. a) Near‐infrared (NIR) imaging of 615 mice on D1, D2, D3, D4, D7and D10 after subcutaneous injection with FSP, FSP/BPH or FSP‐RZ‐BPH. FSPs were dyed with Cy5‐NHS. b) Total radiant efficiency in vivo of 615 mice after injection with FSP, FSP/BPH or FSP‐RZ‐BPH over time (n = 3). c) Cumulative release of Zn²⁺ from FSP‐RZ‐BPH in different pH in vitro (n = 3). d) Cumulative release of R848 from FSP‐RZ‐BPH in different pH in vitro (n = 3). e) Cumulative release of FSP from FSP‐RZ‐BPH in different pH in vitro (n = 3). f) Schematic diagram of different administrations of RZ in MFC tumor suppression experiment. g) Average tumor‐growth curves of 615 mice bearing MFC gastric cancer with different treatments as indicated in 20 days (n = 5). h) Average body weight of mice in different groups in 20 days (n = 5). i) Levels of the eight inflammatory cytokines (IL‐5, IL‐13, IL‐2, IL‐6, IL‐10, IFN‐γ, TNF‐α, IL‐4) in the serum of mice in different groups detected on D15 (2 days post last administration) and D20 (7 days post last administration) (n = 3). j) Percentage of CD80⁺CD86⁺ DCs in tumor‐draining lymph nodes (TDLNs) in different groups detected on D15 and D20 (n = 5). k) Percentage of T_EM in tumor in different groups detected on D15 and D20 (n = 5). The error bars represented mean ± SEM. P‐values were calculated by two‐way ANOVA and Tukey post‐test and correction g,h) or two‐tailed unpaired Student's t‐tests j,k). ns represented p > 0.05.

Figure 4

In vivo anti‐tumor effect of FSP‐RZ‐BPH. a) Schematic diagram of administration route of FSP‐RZ‐BPH in MFC tumor suppression experiment. b) Average tumor‐growth curves of 615 mice bearing MFC gastric cancer with different treatments as indicated in 27 days (n = 7). c) Photos of tumors harvested from mice in all groups on D23 (10 days after last administration) (n = 7). d) Tumor‐growth curves of each mouse in different groups (n = 7). CR represented complete response. e) Survival curves of 615 mice in different groups for 60 days (n = 7). f) Average body weight of 615 mice in different groups in 27 days (n = 7). The error bars represented mean ± SEM. P‐values were calculated by two‐way ANOVA and Tukey post‐test and correction b, f) or log‐rank (Mantel–Cox) test e). ns represented p > 0.05.

Figure 5

Immune response induced by FSP‐RZ‐BPH. a) Percentage of mature DCs (CD80⁺CD86⁺/CD11c⁺) in tumor‐draining lymph nodes (TDLNs) detected on D23 (n = 6). b) Percentage of CD8⁺CD11c⁺ DC (CD8⁺CD11c⁺/CD11c⁺) in TDLNs detected on D23 (n = 6). c) Percentage of central memory T cells (T_CM, CD44⁺CD62L⁺/CD3⁺CD8⁺) in TDLNs detected on D23 (n = 6). d) Percentage of effector memory T cells (T_EM, CD44⁺CD62L⁻/CD3⁺CD8⁺) in TDLNs detected on D23 (n = 6). e) Percentage of M1 tumor‐associated macrophage (M1‐TAM, CD86⁺/F4/80⁺CD11b⁺) detected on D23 (n = 6). f) Percentage of M2 tumor‐associated macrophage (M2‐TAM, CD206⁺/F4/80⁺CD11b⁺) detected on D23 (n = 6). g) Percentage of CD4⁺ T cells (CD4⁺/live cells) in tumor detected on D23 (n = 6). h) Percentage of CD8⁺ T cells (CD8⁺/live cells) in tumor detected on D23 (n = 6). i) Immunofluorescence staining of CD8⁺ cells in tumors of mice in NS and FSP‐RZ‐BPH groups (The scale bar is 100 µm in merge images and 200 µm in the other images). j) Representative flow cytometry images of CD8⁺IFN‐γ⁺ T cells in tumors in different groups detected on D23. k) Percentage of CD8⁺IFN‐γ⁺ T cells (CD8⁺IFN‐γ⁺/CD8⁺) in tumors in different groups detected on D23 (n = 5). l) Representative flow cytometry images of T_EM in tumors of mice in different groups detected on D23. m) Percentage of T_EM (CD44⁺CD62L⁻/CD3⁺CD8⁺) in tumors of mice in different groups detected on D23 (n = 6). n) Representative flow cytometry images of T cells in spleen of mice in different groups detected on D23. o) Percentage of T cells in spleen of mice in different groups detected on D23 (n = 6). The error bars represented mean ± SEM. P‐values were calculated by two‐tailed unpaired Student's t‐tests. ns represented p > 0.05.

Figure 6

RNA‐seq analysis of tumors after FSP‐RZ‐BPH administration. a) Volcano map of differentially expressed genes (DEGs) between the NS and FSP‐RZ‐BPH groups (n = 3). b) Gene Set Enrichment Analysis (GSEA) analysis of cell proliferation and immune response (n = 3). c) Gene Ontology (GO) analysis of upregulated genes (n = 3). d) GO network analysis of significantly upregulated genes in tumors of mice in FSP‐RZ‐BPH compared with the NS‐treated mice. Color scales represented the values of log2‐transformed fold changes and the circle size represented gene enrichment (n = 3). e) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis (n = 3). f) Schematic diagram to demonstrate the mechanism by which R848 and Zn²⁺ exert synergistic pro‐inflammatory effects as TLR7/8 agonists and cGAS‐STING agonists, respectively. g) Differential genes related to cGAS‐STING, Toll‐like receptor and NF‐κB signaling pathways. The high‐expression genes and low‐expression genes were clustered by log 10 (FPKM+1) (n = 3).

Figure 7

FSP‐RZ‐BPH induced MFC‐specific immune memory responses. a) Schematic diagram to show the administration route of pre‐immunized FSP‐RZ‐BPH in MFC tumor prevention experiment and splenocytes analysis. b) Average tumor‐growth curves of 615 mice bearing MFC gastric cancer with different treatments as indicated in 28 days (n = 6). c) Survival curves of 615 mice in different groups for 60 days (n = 6). d) Representative flow cytometry images of T_CM and T_EM in spleen of mice in different groups detected on D0. e) Percentage of T_CM and T_EM in spleen of mice in different groups detected on D0 detected on D23 (n = 5). f) Representative flow cytometry images of CFSE⁺PI⁺ MFC tumor cells after co‐incubation with splenocytes of mice after different treatments. g) Killing ability of splenocytes to MFC tumors after different treatments (n = 3). h) Representative flow cytometry images of CFSE⁺PI⁺ cells of four different tumor cells (B16F10, 4T1, CT26 and MFC cells) after co‐incubation with splenocytes of mice after FSP‐RZ‐BPH treatment. i) Killing ability of splenocytes to different tumor cells after NS or FSP‐RZ‐BPH treatment (n = 3). j) Concentration of IFN‐γ secreted by splenocytes of 615 mice stimulated by WTP peptides (abscissa coordinates 1–6 represented WTP‐1 to WTP‐6), MFC neoantigens (abscissa coordinates 1–6 represented MFC‐1 to MFC‐6) or FSP neoantigens (abscissa coordinates 1–6 represented FSP‐1 to FSP‐6). Before testing, the mice were inoculated with MFC tumor cells and treated with FSP‐RZ‐BPH twice as the schematic diagram showed (n = 3). The error bars represented mean ± SEM. P‐values were calculated by two‐way ANOVA and Tukey post‐test and correction b), log‐rank (Mantel–Cox) test c) or two‐tailed unpaired Student's t‐tests e,g,i). ns represented p > 0.05.