Supramolecular STING‐Hydrogel Spatiotemporally Boosts Tumor‐Derived Extracellular Vesicle‐Based Personalized Vaccine for Enhanced Cancer Immunotherapy

ABSTRACT

Personalized vaccines represent a promising approach for cancer treatment by eliciting tumor‐specific immune responses. However, their development faces persistent challenges, including the laborious and costly process of neoantigen identification, which substantially delays vaccine production. Moreover, conventional platforms suffer from rapid systemic clearance and inefficient delivery to antigen‐presenting cells (APCs), leading to only transient and weak immune activation. Here, we report a STING‐agonist‐integrated hydrogel system designed for localized delivery of tumor‐derived extracellular vesicles (TEVs) to achieve robust and durable antitumor immunity. Upon subcutaneous administration, the in situ formed hydrogel vaccine acts as a depot for antigen‐rich TEVs and STING agonists. We demonstrate that this vaccine promotes dendritic cells (DCs) recruitment and establishes an immune‐permissive niche through spatiotemporal control over antigen‐adjuvant distribution and prolonged APC‐antigen engagement. Within this niche, DCs efficiently internalize and process TEVs under STING‐mediated activation. These matured DCs subsequently migrate to draining lymph nodes, where they initiate potent and sustained tumor‐specific T‐cell responses. Our results show that rapidly producible personalized vaccines derived from melanoma TEVs effectively inhibit tumor growth. In a postoperative breast cancer model, patient‐tailored TEV vaccines also markedly prevent tumor recurrence and metastasis. This readily customizable platform represents a robust and translatable strategy for personalized cancer immunotherapy.

A supramolecular self‐adjuvant STING‐hydrogel serves as a delivery platform for antigen‐rich extracellular vesicles, constituting a personalized cancer vaccine. This hydrogel spatiotemporally controls vaccine biodistribution, promotes dendritic cell recruitment, and establishes an immune‐permissive niche, thereby initiating robust tumor‐specific T‐cell responses. The platform effectively controls tumor growth and prevents postoperative recurrence and metastasis, offering a promising strategy for personalized cancer immunotherapy.

1. Introduction

Cancer vaccines have emerged as a pivotal immunotherapeutic approach, demonstrating substantial potential in cancer treatment [1, 2, 3]. Personalized cancer vaccines, which are tailored to target patient‐specific neoantigens, are designed to elicit tumor‐specific immune responses, thereby enhancing treatment precision and efficacy [4, 5, 6, 7, 8]. However, despite their promise, the clinical translation of these vaccines remains constrained by suboptimal therapeutic outcomes and practical challenges [9, 10, 11]. A major obstacle lies in the laborious, time‐consuming, and costly process of neoantigen screening and identification, which severely delays vaccine production and frequently causes patients to miss the optimal therapeutic window [11, 12, 13]. Furthermore, conventional vaccine platforms suffer from rapid clearance by mononuclear phagocytes, poor lymph node targeting, and inefficient uptake by antigen‐presenting cells (APCs) [14, 15, 16]. These deficiencies collectively impair antigen processing and presentation, resulting in transient and weak immune activation that ultimately fails to restrain tumor progression [17, 18, 19]. Therefore, to fully realize the potential of personalized cancer vaccines, innovative strategies enabling rapid neoantigen acquisition combined with precisely engineered APC‐targeted delivery systems are essential for inducing robust and durable anti‐tumor immunity.

Extracellular vesicles are cell‐derived nanoscale structures that carry diverse biomolecules, including proteins and nucleic acids, and mediate essential intercellular communication [20, 21, 22]. Tumor‐derived extracellular vesicles (TEVs), enriched with tumor‐associated antigens and neoantigens, offer a promising antigen source for vaccine development [23, 24, 25, 26]. Their nanoscale dimension facilitates efficient uptake by APCs, leading to processing and presentation of TEV‐derived antigens via major histocompatibility complex molecules, thereby initiating antitumor immune responses [27, 28]. Clinically, autologous TEVs can be directly isolated from surgical tumor specimens, bypassing the need for laborious neoantigen identification and synthesis, thereby enabling rapid and efficient production of personalized vaccines [29]. Despite the potential as ideal antigen candidates, TEVs face challenges including rapid clearance, poor lymphatic trafficking, and low immunogenicity, which collectively compromise their capacity to elicit potent immune responses [30, 31, 32, 33]. In addition, TEVs inherently possess immunosuppressive properties, which facilitate immune evasion and tumor progression, limiting their direct therapeutic utility [34, 35]. Based on this, we hypothesized that spatiotemporal control over TEV‐APC interactions, combined with appropriate adjuvants, could establish robust and sustained antitumor immunity.

Hydrogels serve as an advanced delivery platform capable of precisely regulating the spatial distribution and temporal release of therapeutic agents, thereby maximizing their efficacy [36, 37]. Functioning as stable local reservoirs, they facilitate efficient co‐delivery of vaccine components while mitigating rapid systemic clearance and significantly prolonging therapeutic exposure [38, 39, 40]. Specifically, engineered hydrogel scaffolds can mimic lymph node immune niches through interconnected macroporous networks that promote immune cell infiltration and interaction [41, 42, 43]. Building on this premise, we hypothesize that the scaffolds establish specialized depots that co‐localize antigens and adjuvants while recruiting APCs, facilitating sustained exposure and enhancing antigen uptake and processing. Following activation and maturation, APCs migrate to draining lymph nodes to initiate antigen‐specific T cell responses. Our recent studies have developed an innovative approach through direct conjugation of therapeutic agents with self‐assembling peptides, creating a “self‐delivery” prodrug‐based hydrogelator. By employing a “drug‐delivering‐drug” strategy, this hydrogelator is well suited for localized delivery of immune modulators, enabling synergistic combined therapy [44, 45, 46, 47]. Based on these findings, we propose that adjuvant‐integrated prodrug hydrogels establish an immune niche via localized antigen delivery. By achieving spatiotemporal control over antigen/adjuvant distribution and prolonging APC–antigen engagement, this strategy induces robust and durable antitumor immunity.

In this context, we present a STING‐agonist‐integrated supramolecular hydrogel system for localized delivery of TEVs and validate its efficacy as a robust platform for personalized cancer immunotherapy (Figure 1). The personalized TEV/GC‐MF vaccine was fabricated by co‐encapsulating antigen‐rich TEVs—directly isolated from resected tumor specimens—and granulocyte‐macrophage colony‐stimulating factor (GM‐CSF) within self‐assembling MSA‐RGD filaments. Upon subcutaneous administration, the hydrogelator assembles into an in situ depot that serves as a sustained‐release reservoir for antigens and adjuvants. We show that the vaccine promotes dendritic cell recruitment and establishes an immune‐permissive niche, within which dendritic cells (DCs) efficiently internalize and process TEV‐derived antigens under STING‐mediated activation. These matured, antigen‐primed DCs subsequently migrate to draining lymph nodes, initiating a systemic and tumor‐specific cytotoxic T‐cell response. Collectively, our findings establish that this readily customizable TEV/GC‐MF vaccine platform effectively controls tumor growth, prevents postoperative recurrence, and suppresses metastasis, offering a promising strategy for personalized cancer immunotherapy.

FIGURE 1.

Design of the supramolecular STING‐hydrogelator and construction of the personalized TEV/GC‐MF vaccine for cancer immunotherapy. (a) Chemical structure of MSA‐RGD amphiphilic molecule. (b) Schematic illustration of the preparation process of the personalized TEV/GC‐MF vaccine and its proposed antitumor immunotherapeutic mechanism. Antigen‐rich TEVs, isolated directly from resected tumor tissue, are encapsulated within a STING‐agonist‐incorporated hydrogel. After subcutaneous injection, the in situ formed TEV/GC‐MF hydrogel recruits immature dendritic cells (DCs) and establishes an immune niche, wherein DCs efficiently internalize and process TEV‐derived antigens with STING adjuvant activation. These matured and antigen‐primed DCs then migrate to draining lymph nodes, initiating a robust tumor‐specific T‐cell response that suppresses tumor recurrence. Figure (b) was created with BioRender.

2. Results

2.1. Design and Characterization of TEV/GC‐MF Hydrogel Vaccine

To construct a bio‐responsive supramolecular adjuvant hydrogelator, we designed an amphiphilic prodrug through site‐specific conjugation. The hydrophobic STING agonist MSA‐2 was modified with 3‐(pyridin‐2‐yldisulfanyl) propan‐1‐ol to introduce a glutathione (GSH)‐cleavable disulfide bond. The resulting MSA‐2 prodrug was then conjugated to a hydrophilic CVVGRGD peptide, yielding the amphiphilic molecule MSA‐RGD (Figure 1a). Successful synthesis was confirmed by mass spectrometry and comprehensive characterization (Figure S1‐S5). The STING agonist incorporated amphiphile MSA‐RGD self‐assembled into filamentous structures (MSA‐RGD filaments, MF) in aqueous solution, as visualized by transmission electron microscopy (TEM) (Figure 2a).

FIGURE 2.

Characterization of TEV/GC‐MF hydrogel vaccine. (a) Representative TEM image of MSA‐RGD filaments (MF), Scale bar: 200 nm. (b) The solution‐to‐hydrogel transition of MF solution induced by the addition of PBS. (c) Storage modulus (G′) and loss modulus (G″) of MF solution transitioned to hydrogel with the addition of PBS. (d) Representative SEM image of MF‐gel, Scale bar: 20 µm. (e) Representative TEM image of TEVs, Scale bar: 100 nm. (f) Western blot analysis for extracellular vesicle markers Alix, CD63 and Gm130 expressed in TEV and cell lysate. (g) Zeta potential of MF solution and after loading of TEVs and GM‐CSF, n = 3. (h) Confocal image of TEVs distribution in MF‐gel, TEVs labeled with Dil. (i) G′ and G″ of TEV‐MF solution transitioned to hydrogel with the addition of PBS. (j) Cumulative release profiles of MSA‐RGD and GM‐CSF from GC/MF‐gel, n = 3. (k) Free MSA‐2 release profiles from MSA‐RGD solution incubated with 10 mM GSH, n = 3. (l) Images of gelation and degradation of MF hydrogel. (m) Quantification of the degradation profile of the MF hydrogel, n = 3. (n) Fluorescence IVIS images showing the in vivo retention of TEVs injected either in solution form or after loading into MF solutions, TEVs were labeled with DIR. (o) Quantification of the retention profile of TEVs, n = 3. All quantitative data are presented as mean ± SD, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Notably, the introduction of phosphate‐buffered saline (PBS) induced electrostatic screening, triggering an instantaneous sol‐to‐gel transition. This macroscopic phase change was visually confirmed by the inverted vial test (Figure 2b). Rheological analysis showed a pronounced increase in the storage modulus (G′), which exceeded the loss modulus (G″) upon PBS addition, confirming the formation of a hydrogel (Figure 2c). We propose that this supramolecular hydrogel offers a suitable platform for localized TEVs delivery and DCs recruitment. The resulting hydrogel was lyophilized and its microstructure examined by scanning electron microscopy (SEM). As shown in Figure 2d, the hydrogel exhibits a highly porous and interconnected architecture. This mesoporous 3D scaffold not only offers an ideal physical framework for efficient DCs infiltration and migration but also facilitates the establishment of a functional in situ immune niche, thereby serving as a critical structural foundation [41].

TEVs, serving as a key component of the hydrogel vaccine formulation, were isolated and validated in subsequent experiments. TEM imaging confirmed their characteristic vesicular morphology (Figure 2e), and dynamic light scattering analysis indicated a predominant particle size distribution ranging from 50 to 100 nm (Figure S6). Western blot analysis further identified the presence of positive markers Alix and CD63, along with the absence of the negative marker GM130 (Figure 2f), confirming the successful isolation of TEVs [48]. Given the measured negative zeta potential of TEVs (−16.07 ± 3.04 mV), we postulated that electrostatic interactions would drive their adsorption onto the positively charged MF component. This hypothesis was substantiated by zeta potential measurements of the MF solution following TEVs adsorption. A significant decrease in potential was observed, shifting from 19.97 ± 2.3 mV to 11.7 ± 0.97 mV (Figure 2g), providing direct evidence for successful TEVs loading onto the MF matrix via charge‐mediated interaction. To spatially visualize TEVs distribution within the hydrogel network, fluorescent labeling with Dil was performed. Confocal laser scanning microscopy (CLSM) imaging confirmed the homogeneous dispersion of TEVs throughout the MF‐gel structure (Figure 2h). Furthermore, rheological analysis demonstrated that the composite hydrogel containing TEVs and GM‐CSF maintained its gelation behavior (Figure 2i), supporting its suitability as a functional vaccine platform for subsequent applications.

2.2. TEV/GC‐MF Hydrogel Serves as Reservoir for Sustained Release of Vaccine Components

We evaluated the in vitro release profile of the hydrogel vaccine as a localized drug depot by monitoring the release of MSA‐RGD and GM‐CSF in PBS at 37°C. Over 30 days, the cumulative release of MSA‐RGD reached about 50.72%, while that of GM‐CSF was approximately 68.33% as the gel degraded (Figure 2j). These results confirm that the hydrogel vaccine acts as a stable reservoir for sustained drug release, enabling continuous immune activation and prolonged DCs recruitment to improve therapeutic outcomes. To investigate controlled release behavior, we assessed MSA‐2 release under 10 mM GSH conditions. In the presence of GSH, the disulfide bond was rapidly cleaved due to its strong reducibility, promoting release of the MSA‐2 prodrug from the MSA‐RGD molecular [45]. Within 4 h, 57.93% of the MSA‐2 prodrug was released under GSH conditions, compared to the control (Figure 2k). ESI spectra verified the successful release of the MSA‐2 prodrug (Figure S7). Once released, the MSA‐2 prodrug undergoes esterase‐mediated hydrolysis in vivo, cleaving its ester bonds to exert pharmacological activity [49].

After administering the MF solution to mice, a yellow‐white gel formed at the injection site within 30 min (Figure 2l), confirming the solution‐to‐gel transition under physiological conditions and the rapid formation of an in‐situ supramolecular hydrogel. Moreover, 14.91% of the hydrogel remained at the injection site on day 30 (Figure 2m), demonstrating its capacity for controlled and sustained drug release. To investigate TEV retention within the MF‐gel, we compared a TEV‐loaded hydrogel group with free TEV as a control. In vivo fluorescence imaging revealed that free TEVs were nearly undetectable by day 10, whereas the TEV/MF‐hydrogel group still exhibited strong signals on day 15 (Figure 2n,o). Additionally, biodistribution analysis of DiR‐labeled TEVs delivered via the MF hydrogel showed minimal detectable fluorescence in major organs (heart, liver, spleen, lung, and kidney) at all examined time points (Figure S8). These results demonstrate that the hydrogel effectively restricts TEVs to the local injection microenvironment, preventing systemic dissemination while ensuring sustained local retention. Additionally, TEVs isolated from the MF‐gel on day 10 retained well‐preserved structural integrity, as observed by TEM (Figure S9). Furthermore, we quantitatively analyzed the in vivo retention profiles of MSA‐RGD, TEV, and GM‐CSF (Figure S10). The results indicate that the system enables relatively synchronized and sustained release of each component while maintaining favorable local retention, thereby providing a basis for prolonged and stable immune activation.

2.3. TEV/GC‐MF Hydrogel Vaccine Constructs an Immune Niche Spatiotemporally Enhances Antigen Processing and Immune Activation

To evaluate the potential of TEVs as antigenic substitutes in vaccine development, we systematically assessed their antigenic properties. TEVs were first isolated from B16‐OVA cells and examined for key antigen content. Western blot analysis confirmed the presence of representative tumor‐associated antigen and neoantigens (GP100, Trp1, and OVA) in TEVs derived from B16‐OVA cells, with B16‐OVA whole‐cell lysates serving as biological controls (Figure 3a). This key finding established a foundation for their use as multicomponent antigen sources in vaccines. We next investigated their interaction with APCs by tracking the intracellular trafficking of Dil‐labeled TEVs in DC2.4 cells. Confocal imaging revealed time‐dependent colocalization of TEVs with lysosomal markers, suggesting lysosomal entry as part of the antigen processing and presentation pathway (Figure 3b) [48]. Quantitative colocalization analysis further confirmed a significant overlap between TEVs and lysosomal markers, providing robust evidence for lysosomal localization of internalized TEVs (Figure S11).

FIGURE 3.

TEV/GC‐MF hydrogel vaccine constructs an in‐situ immune niche for enhanced immune responses. (a) Western blot analysis of TEVs for the representative antigens GP100, Trp1 and OVA; B16‐OVA cell lysates were used as biological controls. (b) Confocal microscopy images showing the uptake of TEVs by DC2.4 cells at different time points. Nuclei were stained with DAPI, lysosomes with Lyso‐Tracker, and TEVs with Dil. Scale bar: 20 µm. (c) Quantification of CD80⁺/86⁺ and (d) SIINFEKL‐H2Kb⁺ DCs across the indicated treatment groups, n = 4. (e) Representative flow cytometry plots of CD11c⁺ DCs recruitment in hydrogel. (f) Quantification analysis of CD11c⁺ DCs recruitment in hydrogel across groups, n = 4. (g) Quantification of XCR1⁺ DCs in hydrogel across groups, n = 4. (h) Representative flow cytometry plots and (i) quantification analysis and of CD80⁺/86⁺ DCs in the indicated treatment groups, n = 4. (j) Lymph node‐homing CCR7⁺ DCs in hydrogel across groups, n = 4. (k) Immuno‐analysis in lymph nodes, quantification of activated CD80⁺/86⁺ DCs, (l) XCR1⁺ DCs, (m) CCR7⁺ DCs and (n) antigen‐specific SIINFEKL‐H2Kb⁺ CD11c⁺ DCs in lymph nodes across groups, n = 4. (o) Quantification of CD8⁺ T cells and (p) antigen‐specific SIINFEKL tetramer⁺ CD8⁺ T cells in lymph nodes at day 7 post‐immunization, n = 4. (q) Schematic illustration depicting the hydrogel vaccine as an in situ immune niche that recruits DCs, primes their activation, promotes migration to lymph nodes, and ultimately facilitates tumor‐specific T‐cell proliferation. Figure (q) was created with BioRender. All quantitative data are presented as mean ± SD, p values were assessed by one‐way ANOVA and marked with asterisks, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

In vitro activation assays using BMDCs revealed that both MSA‐2 and MSA‐RGD induced maturation, with CD80⁺/CD86⁺ expression rates of 25.65% and 23%, respectively. Notably, the combination of TEV and MSA‐RGD further enhanced this effect, increasing the proportion of CD80⁺/CD86⁺ cells to 34.15% (Figure 3c). MHC‐II expression was also significantly upregulated in groups containing MSA‐2, particularly under the TEV + MSA‐RGD condition (Figure S12). Moreover, the proportion of SIINFEKL‐H2Kb⁺ DCs was markedly increased in the TEV‐treated group, reaching 12.83%, and this effect was further amplified to 22.28% upon co‐treatment with MSA‐RGD (Figure 3d). Collectively, these results demonstrate robust activation of DCs, which was significantly enhanced by MSA‐RGD adjuvants.

The above results confirm that the antigenicity of TEVs effectively promotes DCs activation in the presence of MSA‐RGD, supporting the feasibility of an in situ personalized hydrogel vaccine. To achieve robust and long‐lasting immunity, we hypothesized that such TEV/GC‐MF hydrogel vaccine could establish an immune niche functioning as a local site for immune activation through DCs recruitment. We therefore evaluated the vaccine's capacity to recruit and activate DCs in a murine model. For comparison, we designed a control molecule, C6‐RGD, and its assembly was designated as EF, with characterization data provided in Figure S13. After 7 days of vaccination, we analyzed the infiltration of major immune cell types within the TEV/GC‐MF hydrogel. The results showed that incorporation of GM‐CSF significantly increased the proportion of CD11c⁺ DCs relative to other key immune populations, including B cells (CD19⁺), macrophages (F4/80⁺), and T cells (CD3⁺) (Figure S14). Notably, the TEV/GC‐MF hydrogel vaccine group exhibited a marked increase in the recruitment of CD11c⁺ cells (42.23%) compared to the EF group (22.63%) (Figure 3e,f). Furthermore, the incorporation of GM‐CSF into the TEV/GC‐MF system synergistically enhanced DCs recruitment, elevating the proportion of CD11c⁺ cells from 26.43% in the TEV‐MF group to 42.23%. Analysis of DC subtypes recruited by the vaccine revealed that the TEV/GC‐MF formulation significantly increased the population of antigen‐presenting cDC1 (XCR1⁺) DCs (Figure 3g), which was substantially higher than that in the control groups. Additionally, the TEV/GC‐MF hydrogel system led to elevated levels of cDC2 and monocyte‐derived DCs compared with the EF and TEV‐MF groups, whereas plasmacytoid DC populations showed no significant difference between groups (Figure S15).

To verify activation of the STING pathway, downstream signaling molecules, including phosphorylated TBK1 (p‐TBK1) and phosphorylated IRF3 (p‐IRF3) were examined. As shown in Figure S16a, the TEV/GC‐MF group exhibited markedly enhanced STING pathway activation compared with the TEV/GC‐EF group, as evidenced by the increased expression of p‐TBK1 and p‐IRF3. Consistently, the elevated secretion of type I interferon IFN‐β further confirmed effective STING activation in the TEV/GC‐MF group (Figure S16b), providing mechanistic support for the subsequent immunological analyses. Building on this finding, we next assessed whether STING activation translates into enhanced DC function. The TEV‐MF group showed approximately twofold higher levels of activated CD80⁺CD86⁺ DCs relative to the EF group (Figure 3h,i), an effect likely attributable to the synergistic DCs maturation induced by TEVs and MSA‐2. In the presence of GM‐CSF, the proportion of activated CD80⁺/CD86⁺ DCs in the TEV/GC‐MF group reached 43%. Meanwhile, MHC‐I expression also showed a marked upward trend (Figure S17), particularly in the TEV/GC‐MF group, indicating enhanced antigen‐presenting capacity of DCs. Lymph node‐homing CCR7⁺ DCs also showed a consistent increasing trend, reaching 31.2% in the TEV/GC‐MF group (Figure 3j). Taken together, these findings indicate that the STING agonist is essential for driving robust DC maturation in the presence of TEVs, thereby enhancing subsequent antigen presentation and migration.

After confirming the hydrogel vaccine's ability to recruit and activate DCs, we further analyzed immune cells in the inguinal lymph nodes to evaluate DC transport and T‐cell activation. Results indicated that the TEV/GC‐EF, TEV‐MF, and TEV/GC‐MF groups all elevated the proportions of CD80⁺CD86⁺ DCs, XCR1⁺ DCs, and CCR7⁺ DCs (Figure 3k–m), with the TEV/GC‐MF group exhibiting the most pronounced effects. Notably, antigen‐specific SIINFEKL‐H2Kb⁺DCs were also significantly increased in the TEV/GC‐MF group, attributable to OVA antigens carried by TEVs (Figure 3n). Subsequently, antigen‐laden DCs efficiently activated T cells in lymph nodes via antigen presentation. The proportion of CD8⁺ T cells in the TEV/GC‐MF group was approximately 37.1%, significantly higher than that in the TEV/GC‐EF group (27.08%) (Figure 3o), an outcome likely enhanced by MSA‐2–mediated augmentation of antigen presentation and T‐cell differentiation. Moreover, a marked increase in antigen‐specific SIINFEKL⁺ CD8⁺ T cells (7.05%) was observed, indicating the induction of a tumor‐specific immune response (Figure 3p). These data support and extend established evidence that TEVs can serve as potent tumor antigen carriers. Collectively, these findings demonstrate that the scaffold hydrogel vaccine effectively establishes an in situ immune niche that recruits DCs in the presence of GM‐CSF. In combination with TEVs and MSA‐2, it further enhances DC activation and migration to lymph nodes, ultimately driving potent T‐cell differentiation and the generation of tumor‐specific T cells, thereby establishing a robust immunotherapeutic response (Figure 3q).

To determine whether spatial colocalization is required for the formation of an in situ immune niche, we compared physically separated administration of TEV and GC‐MF gel (TEV + GC‐MF) with TEV loaded GC‐MF gel (TEV/GC‐MF). Although the separated TEV + GC‐MF treatment increased DC activation and the expression of CCR7, MHC‐II, and MHC‐I in hydrogel‐infiltrating immune cells and draining lymph nodes compared with EF, these responses were not significantly enhanced compared with GC‐MF alone. In contrast, TEV/GC‐MF treatment induced markedly stronger DC activation, migration‐associated phenotypes, and antigen‐presentation‐associated marker expression (Figure S18a–h). Consistently, TEV/GC‐MF significantly increased intratumoral CD8⁺ T‐cell infiltration compared with the separated TEV + GC‐MF treatment (Figure S18i,j), accompanied by stronger tumor growth inhibition (Figure S18k). These results indicate that spatial colocalization of TEVs and GC‐MF gel within the same local microenvironment is required for optimal immune potentiation and enhanced antitumor efficacy, supporting the functional role of a hydrogel‐defined local immune‐priming niche.

2.4. TEV/GC‐MF Hydrogel Vaccine Elicits Robust Tumor‐specific Immunity

To evaluate the antitumor efficacy of the hydrogel vaccine, we established a B16‐OVA tumor model in C57BL/6 mice. On day 10, mice were randomly assigned to five groups receiving saline, TEV/GC‐EF, Bolus vaccine, TEV/GC‐MF, or TEV/GC‐MF + aPD1. Groups containing TEVs and the MSA‐2 adjuvant showed varying degrees of early tumor suppression, outperforming both the saline and TEV/GC‐EF controls. As treatment progressed, the TEV/GC‐MF hydrogel vaccine exhibited progressively stronger inhibition, which was further enhanced by aPD1 blockade (Figure 4a). Survival analysis revealed significant intergroup differences, with the combination of hydrogel vaccine and aPD1 yielding the best outcomes (Figure 4b). Notably, 50% of mice in the aPD1 combination group survived, likely due to the blockade of PD‐1/PD‐L1 interaction and enhanced T‐cell infiltration in tumor. In addition, all groups maintained normal body weight gain (Figure S19). Histological and blood biochemical analyses revealed no significant abnormalities in major organs or systemic parameters following hydrogel vaccination (Figure S20 and S21), supporting its favorable biosafety profile.

FIGURE 4.

TEV/GC‐MF hydrogel vaccine elicits a robust antigen‐specific immune response and potent anti‐tumor efficacy. (a) Tumor growth curves in the indicated treatment groups. (b) Survival profiles of mice corresponding to the indicated treatment groups, n = 6. (c) Absolute CD8⁺ T cell counts normalized to tumor weight, n = 4. (d) Frequency of CD25⁺ FoxP3⁺ cells among CD4⁺ T cell, n = 4. (e) Proportion of Gr1⁺CD11b⁺ MDSCs in tumor tissue, n = 4. (f) Heatmap depicting cytokine expression profiles in tumor tissues, n = 4. (g) Proportion of SVYDFFVWL tetramer⁺ in CD8⁺ T cells from each group, n = 4. (h) Proportion of SIINFEKL tetramer⁺ in CD8⁺ T cells from each group, n = 4. (i) Representative ELISPOT images showing IFN‐γ spots from re‐stimulated splenocytes at day 7 post‐immunization. (j) Quantitative analysis of IFN‐γ spot‐forming units, n = 4. (k) Percentage of IFN‐γ⁺ cells among CD8⁺ T cells from re‐stimulated splenocytes at day 7, n = 4. (l) Flow cytometry validation of CD8⁺ T cell depletion following antibody administration. (m) Tumor growth kinetics in mice treated with hydrogel vaccine combined with CD8⁺ T cell depletion, n = 5. (n) Volcano plot illustrating differentially expressed genes between saline and TEV/GC‐MF groups. (o) Heatmap of differentially expressed genes in saline versus TEV/GC‐MF groups, n = 3. (p) GO enrichment analysis of up‐regulated differentially expressed genes. (q) Immune cell infiltration profiles in saline and TEV/GC‐MF treated tumors. All data are presented as mean ± SD, p values were assessed by one‐way ANOVA and marked with asterisks, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

As expected, the pronounced antitumor effect was associated with a higher density of CD8⁺ T cells induced by the hydrogel vaccine. The TEV/GC‐MF group showed CD8⁺ CD3⁺ T with density of 106 cells/mg tumor, compared to 73 cells/mg tumor in the bolus vaccine group and 61 cells/mg tumor in the TEV/GC‐EF group (Figure 4c). Furthermore, the TEV/GC‐MF vaccine significantly reduced the infiltration of Treg cells (CD25⁺ FoxP3⁺) (Figure 4d). The proportions of myeloid‐derived suppressor cells (MDSCs) were also suppressed across all treatment groups, with the TEV/GC‐MF hydrogel vaccine group exhibiting particularly lower levels (Figure 4e). These findings indicate a negative regulatory effect on immunosuppressive cells. We also assessed intratumoral levels of pro‐inflammatory cytokines. Compared with saline, both the bolus and hydrogel vaccines significantly elevated TNF‐α, IFN‐γ, IL‐2, and IL‐6, with more pronounced upregulation in the TEV/GC‐MF group, indicating stronger immune activation (Figure 4f; Figure S22).

Following therapeutic intervention with the vaccine platform, the proportions of endogenous TRP‐2‐specific (SVYDFFVWL tetramer⁺) and model OVA‐specific (SIINFEKL tetramer⁺) CD8⁺ T cells were also increased (Figure 4g,h). Notably, the TEV/GC‐MF group induced significantly higher frequencies of these antigen‐specific T cells compared with both the bolus vaccine and the TEV/GC‐EF groups. Upon re‐stimulation with TEVs, splenic mononuclear cells from the TEV/GC‐EF and bolus vaccine groups displayed moderate IFN‐γ spot formation in the ELISPOT assay. In contrast, the TEV/GC‐MF vaccine group showed a substantially higher number of spots, indicating the induction of a stronger antigen‐specific T‐cell response in vivo (Figure 4i,j). Flow cytometric analysis further confirmed a significant increase in the frequency of IFN‐γ⁺ cells among CD3⁺CD8⁺ T cells in the spleen after re‐stimulation (Figure 4k). Together, these results demonstrate the ability of the TEV/GC‐MF hydrogel vaccine to elicit a robust tumor‐specific immune response in mice. To validate the critical role of CD8⁺ T cells in mediating the observed antitumor effect, we performed a CD8⁺ T cell depletion assay. Efficient depletion of CD8⁺ T cells was confirmed by flow cytometry (Figure 4l). Subsequent tumor monitoring showed that CD8⁺ T cell ablation abolished the tumor growth inhibition conferred by the vaccine, underscoring their essential role in the therapeutic efficacy of the treatment (Figure 4m).

Transcriptomic profiling of tumor tissues from the TEV/GC‐MF vaccine and saline groups revealed substantial alterations in gene expression. A volcano plot highlighted numerous differentially expressed genes following vaccination (Figure 4n), including key regulators of T cell activation, differentiation, and proliferation (e.g., Ripk3, Adam8, and Il4ra), cytokine production (e.g., Slc11a1 and Tgfb1), and antigen processing and presentation (e.g., H2‐D1). Unsupervised hierarchical clustering further confirmed a clear separation between TEV/GC‐MF vaccine‐treated and control samples (Figure 4o), indicating that the vaccine remodels the tumor transcriptional landscape. Pathway enrichment analysis of up‐regulated genes identified significant involvement of immune‐related processes, particularly those governing T cell activation, differentiation, and proliferation (Figure 4p), supporting the role of TEV/GC‐MF as a potent immune activator. Consistent with these findings, immune infiltration analysis demonstrated a substantial increase in immune cells within vaccinated tumors, most notably activated DCs, activated CD8⁺ T cells, NK cells, and effector/central memory T cells (Figure 4q). Together, these data underscore the ability of the hydrogel vaccine to provoke a robust and multifaceted antitumor immune response.

2.5. Rapidly Producible Personalized TEV/GC‐MF Vaccine Effectively Suppresses Distant Tumors

Based on the above findings, the TEVs‐based hydrogel vaccine effectively elicits tumor‐specific immune responses, leading to potent suppression of tumor growth and improved survival in mice. Clinically, our proposed strategy enables the rapid preparation of a personalized vaccine by directly isolating TEVs from surgically resected patient tumor tissues, followed by immediate formulation and administration. To model this translational workflow, we subsequently developed a personalized hydrogel vaccine using TEVs isolated from excised murine tumors, which was applied directly via injection to evaluate its therapeutic efficacy (Figure 5a). In C57BL/6 mice, primary B16 tumor models were established, followed by the induction of distant tumors seven days later (Figure 5b). The primary tumor was completely resected and served as the source for TEV isolation and subsequent preparation of the personalized TEV/GC‐MF hydrogel vaccine.

FIGURE 5.

Personalized TEV/GC‐MF hydrogel vaccine effectively prevents distant tumor growth and establishes durable immune memory. (a) Schematic illustration of the preparation and therapeutic strategy of the personalized TEV/GC‐MF vaccine. (b) Timeline of tumor inoculation, personalized vaccine preparation, and treatment. Female C57BL/6 mice were subcutaneously inoculated with primary and distant B16 melanoma cells on day −7 and day 0, respectively. On day 10, the primary tumor was completely resected to extract TEVs for personalized vaccine preparation. Mice were then administered saline, bolus vaccine, TEV/GC‐MF, or TEV/GC‐MF combined with aPD1. (c) Tumor growth curves across experimental groups, n = 6. (d) Survival profiles of mice in different treatment groups. (e) Percentage of CD8⁺ T cells in peripheral blood post‐treatment, n = 4. (f) Tumor growth kinetics in the rechallenge experiments, n = 4 for saline, n = 3 for TEV/GC‐MF, and n = 5 for TEV/GC‐MF + aPD1. (g) Survival curves following tumor rechallenge. (h) Representative flow cytometry plots showing central memory (Tcm) and effector memory (Tem) T cells in naive, TEV/GC‐MF, and TEV/GC‐MF + aPD1 groups. (i) Quantification of CD8⁺ Tcm and (j) CD8⁺ Tem cells in splenocytes. Data are presented as mean ± SD, p values were assessed by one‐way ANOVA and marked with asterisks, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Despite sharing an identical formulation, the free bolus vaccine failed to elicit significant tumor suppression, whereas the TEV/GC‐MF vaccine group exhibited marked antitumor efficacy, particularly in combination with aPD1 therapy (Figure 5c). Notably, survival rates reached 50% (3/6) in the personalized hydrogel vaccine (TEV/GC‐MF) group and rose to 83% (5/6) with aPD1 co‐administration (Figure 5d). The potent tumor control achieved by a single injection of the hydrogel vaccine prompted further immune profiling. Analysis of peripheral blood revealed a substantial increase in CD8⁺ T cell levels among hydrogel‐treated mice: 23.75% in the bolus group, 32.7% in the TEV/GC‐MF group, and 39.1% in the TEV/GC‐MF + aPD1 group (Figure 5e). These results suggest that the personalized TEV/GC‐MF vaccine sustains robust immune activation, elevating CD8⁺ T cell frequencies, curbing tumor progression, and extending survival.

To assess whether the personalized hydrogel vaccine could elicit a memory immune response, we conducted tumor rechallenge experiments in mice that survived the initial treatment. Upon reinoculation with B16 tumor cells, unvaccinated control mice exhibited rapid tumor growth, whereas tumor progression was significantly suppressed in those that received the personalized hydrogel vaccine (Figure 5f). The rechallenge further revealed distinct protective efficacy, with survival rates of 66.7% (2/3) in the TEV/GC‐MF group and 80% (4/5) in the TEV/GC‐MF + aPD1 group (Figure 5g). This potent suppression led to investigate the underlying mechanism. Immunological analysis of mouse spleens demonstrated that the personalized hydrogel vaccine induced a robust memory response, characterized by significantly elevated populations of central memory T cells (Tcm) and effector memory T cells (Tem). Specifically, Tcm and Tem levels in the TEV/GC‐MF vaccine group were 3.02‐fold and 2.87‐fold higher, respectively, than those in the naïve group (Figure 5h–j). Collectively, these findings indicate that the personalized hydrogel vaccine not only exerts a strong antitumor effect but also establishes a durable immune memory.

2.6. Personalized TEV/GC‐MF Vaccine Prevents Postoperative Recurrence and Metastasis

Postoperative tumor recurrence and metastasis remain major clinical challenges, severely affecting patient prognosis and quality of life. These issues are primarily attributed to the incomplete surgical removal of the tumor, which leaves minimal residual disease capable of leading to local relapse and distant metastasis. To mimic this clinical scenario, we established an orthotopic breast tumor model by inoculating 4T1 cells into the mammary fat pads of BALB/c mice, along with a lung metastasis model via tail vein injection of 4T1‐Luc cells. Subsequently, ∼ 90% of the primary tumor was surgically resected to simulate postsurgical residual lesions. The excised tumor tissue was used to isolate TEVs for preparing personalized vaccines, which were then administered subcutaneously to monitor tumor recurrence and metastasis (Figure 6a).

FIGURE 6.

Personalized TEV/GC‐MF hydrogel vaccine mitigates post‐surgical tumor recurrence and metastasis. (a) Timeline illustrating the tumor inoculation, personalized vaccine preparation, and treatment. An orthotopic breast tumor model was established in female BALB/c mice by inoculating 4T1‐Luc cells on day 0. To induce metastasis, 4T1‐Luc cells were administered intravenously on day 15. Approximately 90% of the primary tumor was surgically resected on day 16 to serve as the source for TEV isolation and the subsequent preparation of the personalized hydrogel vaccine. Mice were then treated with saline, bolus vaccine, TEV/GC‐MF, or TEV/GC‐MF combined with aPD1. (b) In vivo bioluminescence imaging for monitoring 4T1‐Luc metastasis and recurrence. (c) Representative lung photographs and (d) corresponding H&E‐stained sections from each treatment group. Scale bar: 500 µm. (e) Quantification of lung metastatic foci across treatment groups, n = 5. (f) Tumor growth curves and (g) survival profiles for each treatment group, n = 5. (h) Proportion of CD8⁺ T cells in peripheral blood on day 23 and (j) day 40. (i, k) Statistical analysis of CD8⁺ T cell proportions at the corresponding time points, n = 4. All data are presented as mean ± SD, p values were assessed by one‐way ANOVA and marked with asterisks, **P ≤ 0.01, ****P ≤ 0.0001.

In vivo imaging revealed rapid local recurrence and lung metastasis in the saline control group following incomplete resection of the primary tumor. Notably, a bioluminescent signal was detected in the cervical region of some mice, which likely represents metastatic lesions attributable to the intrinsically high metastatic potential of 4T1 cells and their propensity for lymphatic spread (Figure 6b). In the bolus vaccine group, substantial recurrence was also observed. In contrast, the TEV/GC‐MF hydrogel vaccine significantly suppressed both recurrence and metastasis, an effect that was further strengthened by aPD1 blockade. Gross examination of lung tissues displayed numerous metastatic nodules in the saline group, and the bolus vaccine also failed to control metastasis. By comparison, almost no visible nodules were detected in the lungs of mice treated with the hydrogel vaccine alone or in combination with aPD1 (Figure 6c), a finding corroborated by H&E‐stained sections (Figure 6d). Quantitative analysis confirmed a marked reduction in lung nodule counts in the TEV/GC‐MF vaccine group (Figure 6e). Furthermore, tumor recurrence was poorly controlled in mice receiving the free bolus vaccine (Figure 6f), whereas the hydrogel vaccine significantly inhibited recurrence. This translated into a 60% survival rate in the TEV/GC‐MF vaccine‐only group, and the combination with aPD1 nearly completely suppressed recurrence, achieving 80% survival (Figure 6g).

To elucidate the mechanisms underlying the sustained antitumor immunity conferred by the hydrogel vaccine, we performed immunological profiling at short‐term (day 23) and long‐term (day 40) time points. Analysis of peripheral blood revealed that the pronounced suppression of recurrence and metastasis was primarily due to the vaccine's ability to persistently activate a robust immune response, characterized by a sustained elevation in CD8⁺ T cell levels. On day 23, the proportion of CD8⁺ T cells in the TEV/GC‐MF hydrogel vaccine group was approximately fourfold higher than that in the saline group (Figure 6h,i), and this difference remained pronounced at day 40, with a threefold increase still observed (Figure 6j,k). These results indicate that the hydrogel vaccine effectively controls tumor recurrence and metastasis by maintaining a durable and potent T‐cell–mediated immune response.

3. Conclusion

In this study, we developed a STING‐agonist‐incorporated hydrogel system for localized delivery of TEVs, enabling a facile and efficient platform for personalized cancer vaccination. By circumventing the need for complex antigen identification and screening, this strategy utilizes antigen‐rich TEVs directly isolated from resected tumor tissue, which are encapsulated in a STING‐activating hydrogel for subcutaneous administration. Upon injection, the hydrogel forms an in situ depot that ensures sustained release of antigens and adjuvants, promoting dendritic cell recruitment and establishing an immune‐permissive niche through spatiotemporal coordination of antigen and antigen‐presenting cell localization. This niche serves as potent hub for immune activation, where APCs efficiently internalize and process antigens, migrate to draining lymph nodes, and initiate a robust tumor‐specific T‐cell responses. In murine B16 melanoma models, a single administration of the rapidly fabricated personalized hydrogel vaccine significantly suppressed tumor growth, achieving a 50% survival rate that increased to 83% with aPD1 combination therapy. Moreover, in a breast cancer model mimicking incomplete tumor resection, tumor‐tailored TEV/GC‐MF vaccines potently prevented postoperative recurrence and metastasis, substantially extending overall survival. These results highlight the potential of this approach to address postsurgical residual disease, offering a practical strategy for long‐term tumor control and recurrence prevention. Collectively, our findings demonstrates that this readily customizable, TEV‐based hydrogel vaccine platform elicits potent and durable antitumor immunity with a single administration, representing a promising and clinically translatable strategy for personalized cancer immunotherapy.

The conceptual advance of this work lies in a system‐level integration strategy enabled by a bioactive self‐adjuvanting scaffold design. Unlike inert biomaterial matrices that serve merely for local retention and co‐delivery, our supramolecular hydrogel is constructed directly from a STING agonist prodrug. This allows the scaffold to function simultaneously as a structural depot and an intrinsic source of innate immune stimulation, distinguishing the platform from conventional scaffold‐assisted vaccines based on poly(lactide‐co‐glycolide), alginate cryogels, or mesoporous silica rods [50, 51, 52], in which immunostimulatory activity is typically introduced through separately loaded adjuvants or cytokines. In parallel, the use of TEV provides a personalized antigen source and reduces reliance on predefined epitope identification or antigen synthesis. Thus, the central novelty of this platform lies in the rational coordination of self‐adjuvanting scaffold formation, STING‐mediated innate immune activation, APC recruitment, and personalized antigen presentation within a single localized vaccine system. Notably, although TEV may carry immunosuppressive cargos, their local retention within the hydrogel depot limits systemic exposure and ensures that recruited APCs serve as the primary interacting cells. In this context, MSA‐2‐mediated STING activation enhances DC maturation and antigen presentation, thereby promoting downstream T‐cell activation. Therefore, the effect of the MF hydrogel is best interpreted as functional counterbalancing of TEV‐associated suppressive influences at the level of net immune output, rather than direct neutralization or elimination of specific suppressive cargo molecules.

Several factors must be addressed before clinical translation. In our murine melanoma model, the yield of TEV was approximately 0.3–0.4 µg per mg of tumor tissue, and a single therapeutic dose of 60 µg TEV was sufficient to elicit robust immune activation and antitumor efficacy. However, these data are derived from mouse models and cannot be directly extrapolated to human dosing without validation using patient‐derived specimens. Encouragingly, recent studies have reported TEV yields from human tumors, for example, 0.5‐0.6 µg/mg for hepatocellular carcinoma and 1‐1.5 µg/mg for melanoma [53, 54, 55], suggesting that clinically relevant TEV quantities may be achievable from routine surgical resections. Nevertheless, several open questions remain, including inter patient variability in TEV yield, composition, and immunogenicity; long‐term stability and storage conditions of the formulated TEV loaded hydrogel vaccine; scalable GMP compliant purification and quality control; cold chain logistics for clinical distribution; and the regulatory pathway for autologous EV‐based products. Addressing these challenges will be essential for advancing this personalized hydrogel vaccine platform toward clinical application. Future efforts will focus on establishing a scalable, GMP compliant manufacturing process that ensures sufficient TEV production within a clinically actionable timeframe.

4. Materials and Methods

4.1. Materials

2,2'‐Dipyridyl disulfide, 3‐mercapto‐1‐propanol, methanol, N‐(3‐Dimethylaminopropyl)‐N′‐ethylcarbodiimide hydrochloride (EDCI) 98%, 4‐DMAP (99%), Hexanoic acid, N,N‐Diisopropylethylamine, N,N‐Dimethylformamide, Glutathione (98%), 4‐(5,6‐dimethoxy‐1‐benzothiophen‐2‐yl)‐4‐oxobutanoic acid (MSA‐2) were purchased from Aladdin, Murine GM‐CSF was purchased from Peprotech, Lyso‐Tracker green was purchased from Beyotime. DAPI and DIL were purchased from Adamas. Fetal bovine serum (FBS) was purchased from Every Green. Exosome‐depleted FBS and IL4 was purchased from Absin. All of flow cytometry antibodies were purchased from Biolegend. Anti‐gp100, Anti‐TRP1, and Anti‐Ovalbumin antibodies were obtained from Abcam. All amino acids, HBTU, and Rink Amide AM resin were obtained from GL Biochem (Shanghai). Mouse IL‐2, IL‐6, IFN‐γ, IFN‐β, and TNF‐α ELISA kits were sourced from Shanghai Enzyme‐linked Biotechnology, Anti‐mouse PD1 was purchased from Bioxcell.

4.2. Cell Lines

B16‐OVA and 4T1‐Luc cell lines were cultured in RPMI‐1640 medium (Gibco) containing 10% fetal bovine serum and 1% penicillin‐streptomycin (Gibco), at 37°C in a humidified atmosphere with 5% CO₂.

4.3. Animals

Female C57BL/6 and BALB/c mice (6–8 weeks old; Shanghai SLAC Laboratory Animal) were housed under a 12 h:12 h light‐dark cycle with free access to food and water. All experimental procedures received approval from the Shanghai Jiao Tong University Animal Ethics Committee (A2025284‐001).

4.4. Synthesis of MSA‐RGD

The bio‐responsive linker SS‐py was first synthesized by dissolving 3.03 g of 2,2'‐Dithiodipyridine in methanol within a round‐bottom flask. Subsequently, 1 g of 3‐mercapto‐1‐propanol, dissolved in 10 mL of methanol, was added dropwise to the reaction mixture. After 6 h stirring, solution was processed to isolate SS‐py. Next, MSA‐2‐SS‐py was prepared by dissolving 100 mg of MSA‐2 in anhydrous dichloromethane, followed by 20 min of stirring. To this solution, 129.4 mg of EDCI and 8.25 mg of DMAP were added, after which 74.7 mg of SS‐py was introduced. The reaction proceeded under nitrogen protection for 48 h, and the product MSA‐2‐SS‐py was obtained through isolation and purification. The CVVGRGD peptide was synthesized via solid‐phase peptide synthesis and subjected to double acetylation. The acetylated peptide was then isolated, purified, and lyophilized for subsequent use. Finally, to construct the amphiphilic molecule, 100 mg of MSA‐2‐SS‐py was dissolved in 1 mL of DMSO, 300 mg of CVVGRGD peptide was added. The conjugation reaction was carried out under nitrogen atmosphere for 36 h, followed by purification, isolation, and lyophilization to yield MSA‐RGD. The product was confirmed by fourier transform ion cyclotron resonance mass spectrometry (SolariX 7.0 T).

4.5. TEVs Extraction

For cell‐derived TEVs, B16‐OVA cells at ∼70% density were washed with PBS, cultured in exosome‐free serum medium for 36 h. Subsequently, the culture medium was collected and pre‐treated. First, the supernatant was collected by centrifugation using 1000 g for 10 min, followed by 10 000 g for 20 min, the supernatant was collected and filtered using 0.22 µm sterile filter membrane. The filtered medium was ultracentrifuged through 1,20,000 g for 70 min at 4°C, and the supernatant was discarded, re‐suspended by adding PBS and washed by centrifugation to obtain TEVs. As for tissue TEVs extraction, fresh tumor tissue was finely minced and subjected to enzymatic digestion in serum‐free medium containing 1 mg/mL collagenase type IV and 0.02 mg/mL DNase I at 37°C for 30 min. The digestate was then filtered through a sterile 70‐µm filter membrane. Sequential low‐speed centrifugation steps (1,000 × g for 10 min, followed by 10 000 × g for 20 min) were performed to remove cells and large debris. The supernatant was subsequently filtered through a sterile 0.22‐µm filter membrane. TEVs were then pelleted by ultracentrifugation at 1,20,000 × g for 70 min at 4°C. The resulting TEVs was washed with sterile PBS and subjected to a second ultracentrifugation step under the same conditions. Finally, purified TEVs were resuspended in sterile PBS.

4.6. Preparation and Characterizations of MF and MF‐Gel

To form MSA‐RGD filaments (MF) assemblies, the MSA‐RGD lyophilized powder was dissolved directly in deionized water at concentrations exceeding 1 mM. The MF solution was then incubated at room temperature for 24 h to complete the assembly process. The morphology of the resulting MF was confirmed using transmission electron microscopy (TEM, Tecnai G2 Bio). To validate the gelation behavior of MF from solution to gel, MF gel was prepared by adding 20 µL of 10 × PBS to 180 µL of a 5 mM MF solution. Gel formation was verified using the inverted vial test and rheological measurements (TA, DHR‐20). The gel was subsequently flash‐frozen in liquid nitrogen, lyophilized, and its morphology was further characterized by scanning electron microscopy (SEM, VEGA 3‐XMU LaB6 type).

4.7. Construction of TEV/GC‐MF Vaccine

The TEV‐GM‐CSF/MF vaccine was prepared by mixing 60 µg of TEV with 2 µg of GM‐CSF in a prepared MF solution. After thorough mixing, the mixture was incubated in a 37°C water bath for 30 min to ensure complete integration of components, yielding the final vaccine formulation.

4.8. GSH Response Release Assay

We first prepared 20 mM GSH and 400 µM MSA‐RGD solution in aqueous solution, respectively. The same volume of the two solutions was mixed then to obtain a final solution containing 10 mM GSH and 200 µM MSA‐RGD solution. For the control group 200 µM MSA‐RGD solution was adopted directly. Samples were obtained by incubation at 37°C for 0, 0.5, 1, 2, 4, 8, 12, and 24 h. The sample solutions obtained were quantified by HPLC for quantitative analysis of prodrug release.

4.9. Hydrogel Release Studies

MF‐gel with GM‐CSF was pre‐prepared and incubated to 37°C. The supernatant solution was obtained on days 1, 3, 5, 7, 10, 15, 20, 25, and 30 for analysis. The release of MSA‐RGD was quantified by HPLC. The release of GM‐CSF was quantified by GM‐CSF Elisa kit.

4.10. In Vivo Hydrogel Degradation and TEVs Retention

Briefly, 6.8 mM MF solution was pre‐formulated. We injected equal volume (100 µL) of the solution subcutaneously into each C57BL/6 mouse. The state of the hydrogel was subsequently recorded by taking photographs at the 30 min, day 5, day 10, day 20, and day 30, respectively. For the retaining hydrogel, we dissolved and quantified the residues by HPLC. DIR‐labeled TEVs were incubated with the prepared MF solution. The mixture was then subcutaneously administered to C57BL/6 mice.The control group was injected with an equal number of free TEVs. The fluorescence intensity of TEVs was observed and quantitatively analyzed for TEVs retention by in vivo imaging on day 1, 3, 5, 10, and 15, respectively.

4.11. In Vivo Antitumour Assay

To establish B16‐OVA tumor model, 2 × 10⁶ B16‐OVA cells were subcutaneously injected to C57BL/6 female mice. At day 10, mice were divided randomly into five cohorts (n = 6). The mice received saline, TEV/GC‐EF, free MSA‐2+TEV+GM‐CSF, TEV/GC‐MF or TEV/GC‐MF+aPD1 (MSA‐2 10 mg/kg, TEV 60 µg, GM‐CSF 2 µg, 100 µL per mouse). For the combined immune checkpoint blockade group, aPD1 (100 µg each time) was injected intraperitoneally on day 10, 13 and 16. Tumor volume was continuously monitored after treatment. To construct TEV/GC‐MF personalized hydrogel vaccines for antitumor assay. On day ‐7, 2 × 10⁶ B16 cells were subcutaneously injected to C57BL/6 female mice to establish primary tumor model. On day 0, we establish distant tumor model. On day 10, we completely resected the primary tumors for personalized vaccine construction and treated with saline, free MSA‐2+TEV+GM‐CSF, TEV/GC‐MF, or TEV/GC‐MF+aPD1 (MSA‐2 10 mg/kg, TEV 60 µg, GM‐CSF 2 µg, 100 µL per mouse). For the combined immune checkpoint blockade group, aPD1 (100 µg each time) was injected intraperitoneally on day 10, 13, and 16. Subsequently, distal tumor volume was continuously monitored.

4.12. In Vivo Evaluation of Anti‐recurrence and Metastasis

We established orthotopic breast cancer tumor model by in situ injection of 1 × 10⁶ 4T1‐Luc cells at the breast site of BALB/c mice. On day 15, 2 × 10⁵ 4T1‐Luc cells was injected through the tail vein to establish a lung metastasis model. On day 16, the mice were randomly divided into four groups: saline, bolus vaccine, TEV/GC‐MF and TEV/GC‐MF+aPD1. We resected 90% of the in situ tumors of the mice for vaccine construction and treated with saline, free MSA‐2+TEV+GM‐CSF, TEV/GC‐MF or TEV/GC‐MF+aPD1 (MSA‐2 10 mg/kg, TEV 60 µg, GM‐CSF 2 µg, 100 µL per mouse). The mice were subsequently monitored for recurrence and lung metastasis. For the combined immune checkpoint blockade group, aPD1 (100 µg each time) was injected intraperitoneally on day 16, 19, and 22. Subsequently, the tumor recurrence was monitored by bioluminescence imaging system. The survival of the mice was monitored daily mice were continuously monitored.

4.13. Statistical Analysis

All values in the present study are expressed as mean ± SD. Statistical analysis was performed using GraphPad Prism software. The significance between two groups was analyzed by a two‐tailed unpaired t‐test. For multiple comparisons, a one‐way ANOVA test was used. Survival analysis was performed using Kaplan‐Meier methodology, with between‐group comparisons assessed via log‐rank testing. p‐value<0.05 was considered statistically significant, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Trial Registration

Not applicable, as this study did not involve a clinical trial.

Consent for Publication

No written consent has been obtained from the patients as there is no patient identifiable data included in this case report/series.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adma74204‐sup‐0001‐SuppMat.docx.

Data Availability Statement

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