pH-responsive in situ antigen capture and cytosolic delivery synergize to elicit robust systemic antitumor immunity Authors: Zhuoling Bi, Shucheng Zhang, Guanyu Jin, Zhaofan Yang, Shanyi Lin, Lanqing Wang, Bingzheng Yu, Luyao Wang, Yuanzhen Su, Sijun Xiang, Xinyu Xiang, Letian Lv, Tianlin Lao, Jianfeng Zhou, Shixian Lv, Xiaozhi Rong Published: 2026 ABSTRACT AbstractIn situ cancer vaccines hold strong potential for addressing tumor heterogeneity by using the patient’s own tumor as a personalized antigen source. However, their efficacy remains limited by insufficient antigen capture and inefficient cytosolic delivery. Here, we report an in situ antigen capture and delivery platform, DOX/PDiT@Gel, in which the pH-responsive function is provided by the polymer PDiT, while the hydrogel serves as a local retention matrix. In this system, doxorubicin (DOX) and a cationic polymer, PEG-DIPAMA-TAT (PDiT), are co-encapsulated within an oxidized dextran/carboxymethyl chitosan hydrogel. DOX induces immunogenic cell death, releasing diverse tumor antigens, while PDiT captures these antigens in situ via electrostatic interactions and promotes endo/lysosomal escape under acidic conditions. The hydrogel allows localized delivery of the therapeutic components at the postoperative tumor site. In vitro studies showed that PDiT markedly promoted antigen internalization, cross-presentation, and dendritic cell maturation. In murine models of postoperative recurrence and bilateral breast tumors, local treatment with DOX/PDiT@Gel inhibited the growth of both recurrent and distant tumors, together with stronger dendritic cell activation and enhanced memory T cell responses. Overall, this platform effectively amplifies antitumor immunity and offers a versatile strategy for personalized cancer immunotherapy.Graphical Abstract Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12951-026-04698-3.Keywords: Antigen capture, Endo/lysosomal escape, Immunotherapy, Polymer, In situ cancer vaccines Introduction Cancer vaccines are widely regarded as a valuable immunotherapeutic option, since they promote immune attack on malignant cells by introducing tumor-associated antigens (TAAs) [1, 2]. In conventional settings, these vaccines usually depend on predefined shared TAAs to stimulate specific T-cell responses [3]. However, this approach often shows limited clinical benefit because tumor cells can differ markedly in both genotype and phenotype [4, 5]. Because conventional vaccines target only a limited repertoire of pre-selected antigens, they often fail to adequately cover the antigenic diversity of the tumor. Consequently, antigen-negative or antigen-loss variant clones may evade immune recognition, survive therapeutic pressure, and ultimately contribute to tumor recurrence and metastasis [6]. This mismatch between a fixed antigen payload and the dynamic, heterogeneous landscape of an individual tumor underscores the need for vaccination strategies capable of engaging broader and more personalized antigen repertoires. To help address the challenge of antigenic heterogeneity, in situ cancer vaccines (ISCVs) have emerged as an attractive alternative [7]. Unlike conventional vaccines based on exogenous predefined antigens, ISCV strategies use the patient’s own tumor as an endogenous and individualized antigen source. By inducing tumor cell death locally, this approach promotes the release of a broad repertoire of patient-derived tumor-associated antigens directly within the tumor microenvironment [8]. As a result, the elicited immune response is not restricted to a limited predefined antigen set, but may instead better reflect the heterogeneous antigenic composition of the tumor. In this way, ISCVs offer a potential advantage over conventional antigen-defined vaccines in addressing tumor heterogeneity [7, 9]. Nevertheless, despite this conceptual advantage, the translation of ISCVs remains hampered by critical biological bottlenecks in antigen handling. Specifically, antigens released by tumor cells are scattered locally with low concentration and are readily degraded by proteases, resulting in a limited amount of available effective antigens [10–12]. In addition, the acidic tumor microenvironment suppresses DC membrane receptor activity and lowers endocytic efficiency [13, 14]. Even after DCs internalize tumor-derived antigens, a substantial fraction is trafficked to endo/lysosomal compartments, where proteolysis can limit antigen availability for cross-presentation [15, 16]. Consequently, antigen escape from these degradative compartments into the cytosol is often inefficient, limiting cytosolic cross-presentation and thereby constraining CD8⁺ cytotoxic T lymphocyte (CTL) priming [17]. Therefore, strategies that enhance both antigen capture and cytosolic delivery are critical for improving the efficacy of ISCVs. To address these bottlenecks, we developed an integrated ISCV platform, DOX/PDiT@Gel, designed to better accommodate tumor heterogeneity and improve antigen delivery efficiency. Doxorubicin (DOX) can induce immunogenic cell death in tumor cells, thereby promoting the release of tumor-derived protein antigens (TDPAs) [18, 19]. The polymer PEG-DIPAMA-TAT (PDiT) was engineered to enrich the capture of negatively charged antigens in the weakly acidic tumor microenvironment through electrostatic adsorption driven by protonated 2-diisopropylaminoethyl methacrylate (DIPAMA) units [20, 21]. In addition, PDiT facilitates the cytosolic access of internalized antigens by promoting endo/lysosomal escape, a process enabled by the membrane-active cell-penetrating peptide TAT [22–24]. Improved cytosolic accessibility is expected to favor proteasome-dependent MHC-I antigen processing and enhance DC cross-presentation [25]. PEGylation increases hydrophilicity and improves biocompatibility by forming a hydrated steric barrier [26, 27]. To enable local retention at the postoperative tumor site, DOX and PDiT were incorporated into an oxidized dextran/carboxymethyl chitosan hydrogel (hereafter referred to as “Gel” or “gel”) [28, 29]. Within this system, the gel acts chiefly as a local reservoir, whereas PDiT endows the platform with acid-triggered antigen-capturing and cytosolic-transport capabilities, thereby facilitating dendritic cell maturation, antigen cross-presentation, and subsequent T-cell activation. Collectively, this ISCV platform was designed to enhance local immune activation and support systemic antitumor immune responses (Scheme 1). Schematic representation of an ISCV strategy for antigen capture and enhanced cytosolic delivery to facilitate cross-presentation and boost anti-tumor immunity. (a) Preparation of DOX/PDiT@Gel. (b) Following DOX-induced tumor cell death, TDPAs are released. PDiT captures these antigens in the acidic tumor microenvironment. After endocytosis by DCs, PDiT promotes endosomal escape, enhancing antigen cross-presentation and thereby strengthening tumor immunotherapy Results PDiT synthesis and physicochemical characterization PDiT was synthesized by reversible addition-fragmentation chain transfer (RAFT) copolymerization of DIPAMA and 2-(2-pyridinyldithio) ethyl methacrylate (PDSMA) using PEG-CDTPA as the chain transfer agent, followed by disulfide exchange between pendant PDSE groups and thiolate TAT. The successful formation of the polymer was confirmed by nuclear magnetic resonance (NMR) and gel permeation chromatography (GPC). (Figure S1-S2; Table S1). PDiT underwent pH-triggered self-assembly during titration, forming a stable nanoparticle population when the pH was lowered into the mildly acidic/near-neutral range (Fig. 1a). Dynamic light scattering (DLS) showed that the prepared particles had a hydrodynamic diameter of approximately 140 nm (Fig. 1b). To probe antigen association under conditions relevant to the extracellular tumor microenvironment, PDiT nanoparticles were incubated with the model antigen ovalbumin (OVA) at pH 6.8 for 4 h. After incubation, both DLS and transmission electron microscopy (TEM) showed a marked increase in particle size (Fig. 1c, d), implying the formation of PDiT/OVA complexes. Taken together, these results reveal that PDiT effectively captures protein antigens. In addition, PDiT nanoparticles showed pronounced pH responsiveness in the range of pH 6–7, with an apparent pKₐ of 6.51 (Fig. 1e). At pH 6.8 (above its pKₐ), PDiT is partially protonated, which enhances its cationic character and antigen association. Further acidification (e.g., in endosomes) increases protonation and membrane-disruptive activity. Given the structural similarity between erythrocyte membranes and endosomal membranes, we employed a hemolysis assay to preliminarily assess the endosomal escape capability of PDiT under endosomal acidification-mimicking conditions [26]. Consistent with its pH-responsive behavior, PDiT induced almost no hemolysis at pH 7.4, whereas hemolysis increased markedly to approximately 24.4% at pH 6.8 and further escalated to nearly 96.5% at pH 5.5 (Fig. 1f; S3). These results indicate that PDiT remains inert under physiological pH but becomes strongly membrane-disruptive in acidic environments, supporting its potential to mediate efficient endosomal escape. Preparation and characterization of PDiT. (a) Chemical structure of PDiT and schematic illustration of pH-triggered nanoparticle formation via stepwise pH adjustment (final pH = 7.4). (b) Hydrodynamic diameters of self-assembled PDiT nanoparticles measured by DLS. (c) Hydrodynamic diameter and (d) TEM Images of PDiT nanoparticles before and after incubation with OVA at room temperature for 4 h (n = 3 independent replicates). (e) Changes in the protonation degree of PDiT induced by pH variations. (f) Hemolysis rate in the supernatant after incubation of PDiT with rat red blood cells under different pH conditions (7.4, 6.8, 5.5. n = 3 biological replicates). Data are expressed as mean ± standard deviation (SD). Comparisons between two groups were performed using Student’s t-test, whereas differences among multiple groups were analyzed by one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test. Statistical significance was assigned as: ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 In vitro TDPAs capture by PDiT Based on this design, we assessed the antigen-capturing capacity of PDiT in a cell-based in vitro model. Specifically, tumor cells were first exposed to DOX for 48 h to induce ICD and promote the release of TDPAs. Subsequently, the resulting cell-culture supernatants were incubated with PDiT for 4 h under pH 7.4 or pH 6.8, respectively. Following the removal of unbound proteins, PDiT-associated proteins were isolated and subjected to LC-MS analysis to evaluate the ability of PDiT to enrich tumor-derived antigens (Fig. 2a). Silver staining showed that, compared with the pH 7.4_PDiT group, the pH 6.8_PDiT group effectively bound and enriched TDPAs released upon DOX induction (Fig. 2b). To provide quantitative evidence for the protein-capturing capacity of PDiT, the total amount of proteins captured from DOX-treated tumor cell supernatants was quantified by BCA assay. The results showed that PDiT captured 78.0 ± 12.1 µg protein per mg polymer at pH 6.8, whereas only 5.1 ± 1.6 µg/mg was captured at pH 7.4, corresponding to an approximately 15.2-fold increase under weakly acidic conditions (Figure S4). These data quantitatively demonstrate that PDiT exhibits markedly enhanced protein-capturing capacity at pH 6.8, which is consistent with its pH-responsive protonation behavior. LC-MS analysis further confirmed that PDiT captured a greater number of proteins in the pH 6.8_PDiT group than in the pH 7.4_PDiT group, as illustrated by the Venn diagram (Fig. 2c). Additionally, among the captured danger-associated molecular patterns (DAMPs), the representative proteins heat shock protein HSP90, alarmin HMGB1, and calreticulin were markedly enriched, indicating the efficient tumor antigen-capturing ability of PDiT under pH 6.8 (Fig. 2d, e). To evaluate the breadth of the antigen spectrum, we performed cellular component annotation of 332 proteins specifically enriched in the pH 6.8_PDiT group (Fig. 2f). These proteins are widely distributed in various core structures of the cell, including the protein synthesis machinery (ribosome), RNA processing machinery (spliceosome), protein degradation system (proteasome), as well as the cytoplasm and ATPase complex. Together, these results indicate that under pH 6.8, PDiT effectively captures the tumor antigen pool released by ICD, that spans multiple subcellular compartments and displays substantial structural diversity, providing a solid foundation for subsequent studies on PDiT-based antigen delivery and immune activation. In vitro simulations were conducted to verify the antigen capture ability and protein identification of PDiT under different pH conditions. (a) Schematic illustration of the experimental design, created in https://BioRender.com. (b) Proteins eluted from PDiT were separated by SDS-PAGE and visualized by silver staining. Lanes 1–3: pH 7.4_PDiT group (PDiT incubated with supernatant under pH 7.4). Lanes 4–6: pH 6.8_PDiT group (PDiT incubated with supernatant under pH 6.8). Lane 7: Molecular weight marker. (c) Proteins identified in two groups are shown. Only proteins detected in ≥ 2 replicates per group were retained for analysis (n = 3 biological replicates). (d) Bar chart graphs displaying the relative abundance of four representative DAMPs captured by PDiT (n = 3 biological replicates). (e) Heatmap showing the normalized abundance (Z-score) of 9 key DAMPs in two groups. Protein abundances are represented by their corresponding gene symbols. (f) Bubble plot showing significantly enriched Gene Ontology Cellular Component terms for proteins specifically adsorbed in the pH 6.8_PDiT group. Bubble size represents the number of proteins annotated to each term, and color intensity reflects enrichment significance (-log10(p_value)). Data are expressed as mean ± SD. Comparisons between two groups were performed using Student’s t-test. Statistical significance was assigned as: ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 DC antigen uptake and endo/lysosomal escape DCs are the principal APCs that initiate adaptive immunity, and their antigen uptake and intracellular trafficking are crucial in determining cross-presentation efficiency and CTL activation [27, 30]. Nanocarrier properties, including size, surface charge, and ligands, distinctly influence DCs internalization and endosomal fate [31, 32]. Considering the pH-responsive cationization of PDiT and TAT-assisted membrane interaction, we evaluate that PDiT would enhance DCs uptake of antigen complexes at pH 6.8 and favor routing into cross-presentation pathways. Firstly, a CCK-8 assay identified 30 µg/mL PDiT as a non-cytotoxic concentration at pH 6.8 (Figure S5). PDiT was mixed with FITC-labeled OVA (OVA-FITC) in pH 6.8 medium and then incubated with DC2.4 cells for 1, 4, or 7 h. Antigen uptake was quantified by FCM based on the intracellular FITC signal (Fig. 3a). In parallel, carrier internalization was tracked using Cy5-labeled PDiT (PDiT-Cy5), which showed a time-dependent increase in intracellular signal as visualized and quantified by confocal laser scanning microscope (CLSM) imaging (Fig. 3b, c). Collectively, these data demonstrate efficient cellular internalization of both antigen (OVA-FITC) and carrier (PDiT-Cy5) under mildly acidic conditions, supporting that PDiT enhances antigen uptake by DCs. To gain mechanistic insight into the endocytic route responsible for PDiT-facilitated antigen uptake, DC2.4 cells were pretreated with pharmacological inhibitors of major endocytic routes, followed by co-incubation with OVA-FITC and PDiT, and the intracellular FITC signal was quantified. Chlorpromazine (CPZ), 5-(N-ethyl-N-isopropyl) amiloride (EIPA), genistein, and methyl-β-cyclodextrin (m-β-CD) were used to inhibit clathrin-mediated endocytosis, macropinocytosis, caveolin-associated endocytosis, and cholesterol-dependent membrane processes, respectively. Genistein and m-β-CD reduced the intracellular FITC signal by 39.2% and 17.4%, respectively, whereas CPZ and EIPA caused no obvious change under identical conditions (Fig. 3d, e). suggesting the involvement of cholesterol-dependent, caveolae-related pathways in PDiT-facilitated antigen uptake. In parallel, CLSM imaging of PDiT-Cy5 further showed reduced carrier internalization upon genistein treatment (Fig. 3f), indicating that PDiT entry into DCs is also sensitive to caveolae-related inhibition. PDiT enhances DCs uptake and antigen cytosolic escape to promote DCs maturation. (a) FCM analysis of FITC levels in DC2.4 cells (n = 3 biological replicates). (b) CLSM images of DC2.4 cells incubated with PDiT-Cy5 for 1, 4 and 7 h (red, PDiT-Cy5; green, phalloidin; blue, DAPI). (c) Quantitative analysis of PDiT-Cy5 mean fluorescence intensity in (b) (n = 5). (d, e) FCM analysis of FITC levels in DC2.4 cells treated with OVA-FITC, PDiT and different endocytic inhibitors for 4 h (CPZ, chlorpromazine; EIPA, ethyl-isopropyl amiloride; m-β-CD, methyl-β-cyclodextrin; n = 3). (f) CLSM images of DC2.4 cells incubated with PDiT-Cy5 with or without genistein for 4 h (red, PDiT-Cy5; green, phalloidin; blue: DAPI). (g) CLSM images of DC2.4 cells incubated with OVA-FITC and OVA-FITC + PDiT for 7 h (green, OVA-FITC; red, Lysosome; blue, DAPI). (h) Schematic illustration of the experimental design, created in https://BioRender.com. (i, j) FCM analysis of CD80+ CD86+ cells in CD11c+ BMDCs (n = 3 biological replicates). (k, l) FCM analysis of H-2Kb-SIINFEKL+ cells in CD11c+ BMDCs (n = 3 biological replicates). (m, n) FCM analysis of CD40+ cells in CD11c+ BMDCs (n = 3 biological replicates). Data are expressed as mean ± SD. Comparisons among multiple groups were analyzed by ANOVA followed by Tukey’s post hoc test. Statistical significance was assigned as: ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 Furthermore, lysosomes were labeled with a lysosomal probe to assess the subcellular trafficking of OVA. CLSM co-localization analysis showed that PDiT decreased the co-localization of OVA with lysosomes (Fig. 3g), indicating that PDiT facilitates antigen escape from endo/lysosomal compartments and promotes cytosolic access, thereby alleviating lysosome-mediated sequestration and degradation of antigens. Collectively, these results demonstrate that PDiT not only enhances antigen internalization by DCs, but also improves endo/ lysosomal escape efficiency, providing a basis for efficient cross-presentation and subsequent T cell activation. To further examine whether the reduced lysosomal sequestration was associated with enhanced cytosolic antigen processing, an additional inhibitor study was performed in DC2.4 cells using the proteasome inhibitor MG132. DC2.4 cells were treated with OVA alone or OVA combined with PDiT, with or without MG132, and antigen presentation was assessed by detecting H-2Kb-SIINFEKL complexes. As shown in Figure S6, MG132 treatment markedly attenuated the PDiT-induced increase in H-2Kb–SIINFEKL presentation, indicating that the enhanced antigen processing was at least partially dependent on proteasome activity. Together with the hemolysis assay and reduced lysosomal colocalization, these results provide stronger mechanistic support that PDiT facilitates endosomal/lysosomal escape and promotes cytosolic antigen access for MHC-I antigen processing. DC maturation and antigen cross-presentation in vitro DC maturation is essential for efficient antigen presentation and the initiation of T-cell priming, and this process is usually accompanied by increased expression of co-stimulatory and MHC molecules [33]. To evaluate the effect of PDiT on dendritic cell maturation, bone marrow–derived dendritic cells (BMDCs) were separated into three experimental groups: G1 (PBS), G2 (OVA), and G3 (OVA + PDiT), and cultured at pH 6.8 for 24 h (Fig. 3h, S7). FCM revealed that treatment in the G3 group markedly promoted BMDCs maturation. The proportion of CD11c⁺ cells co-expressing CD80⁺CD86⁺ increased from 17.5% in G1 to 31.2% in G3, corresponding to a 78% increase relative to the control group (Fig. 3i, j). Similarly, the fraction of CD40⁺ cells increased from 30.1% to 43.2%, representing a 44% increase compared with G1 (Fig. 3k, l). Importantly, beyond phenotypic maturation, PDiT significantly enhanced antigen cross-presentation, a functional readout that is closely linked to cytosolic antigen access. H-2Kᵇ–SIINFEKL staining reflects presentation of the OVA257–264 epitope on MHC-I, a hallmark of cross-presentation. The percentage of H-2Kᵇ–SIINFEKL⁺ CD11c⁺ cells increased from 3.9% to 16.6%, representing an approximately 4.35-fold increase over the control group (Fig. 3m, n). Overall, these data suggest that PDiT facilitates DC maturation and strengthens MHC-I cross-presentation. In vivo antitumor efficacy of DOX/PDiT@Gel Based on the above in vitro findings, PDiT was shown to possess both in situ antigen-capture and cytosolic-delivery capabilities. To evaluate its therapeutic potential under conditions closer to the physiological setting, we further constructed a composite system, DOX/PDiT@Gel, for local administration in vivo. In the final formulation, DOX and PDiT were co-loaded into an oxidized dextran/carboxymethyl chitosan hydrogel formed via reversible Schiff-base (imine) crosslinking, in which the hydrogel mainly serves as a local retention and release matrix, whereas PDiT provides the key bioactive functions of antigen capture and cytosolic delivery [28, 29]. Therefore, the preceding in vitro mechanistic studies were primarily performed using PDiT alone to define its intrinsic properties. Before in vivo evaluation, the formulation characteristics of DOX/PDiT@Gel were further assessed. The encapsulation efficiencies of DOX and PDiT in the hydrogel were determined by an indirect method and were found to be 95.6% and 97.2%, respectively. In addition, the release behavior of DOX and PDiT from DOX/PDiT@Gel was evaluated, and the particle size and stability of PDiT after release from the gel were further characterized (Figure S8a, b). These data indicate gradual release of both cargos and preservation of the nanoscale characteristics of released PDiT under the tested conditions. To further assess the cytotoxicity profile of the formulations, CCK-8 assays were performed in both 4T1 tumor cells and NIH-3T3 normal fibroblast cells using PDiT and extracts prepared from different hydrogel formulations. As shown in Figure S9, the DOX-containing hydrogel extract groups exhibited evident cytotoxicity in both cell lines, consistent with the known activity of DOX. By comparison, PDiT alone and the non-DOX hydrogel extract groups showed relatively limited cytotoxicity under the tested conditions. Together, these results further suggest that the principal cytotoxic activity of this system is largely driven by DOX, whereas PDiT and the hydrogel matrix themselves exhibit comparatively lower cytotoxicity. Given that postoperative recurrence and metastasis remain major determinants of poor prognosis in breast cancer, a murine model of postoperative breast cancer recurrence was established to evaluate the in vivo efficacy of this system (Fig. 4a). On day 0, BALB/c mice were subcutaneously inoculated with 1 × 10⁶ 4T1 cells on the back. On day 9, nearly 90% of the tumor tissue was surgically removed, leaving residual lesions to establish a postoperative recurrence model. The mice were then randomly divided into four groups and received local treatment at the resection site: G1 (PBS@Gel), G2 (DOX@Gel), G3 (DOX/AA@Gel), and G4 (DOX/PDiT@Gel). Here, AA (aluminum adjuvant) was included as a control due to its antigen adsorption and retention capability, which partially overlaps with the antigen-capturing function of PDiT. This comparison allows evaluation of whether additional cytosolic delivery and endosomal escape provided by PDiT contribute to enhanced immune responses [34]. DOX/PDiT@Gel suppresses postoperative tumor recurrence in vivo. (a) Schematic illustration of the experimental design, created in https://BioRender.com.BALB/c mice were randomly assigned to four treatment groups and received in situ hydrogel implantation at the tumor resection bed: G1, PBS@Gel; G2, DOX@Gel; G3, DOX/AA@Gel; G4, DOX/PDiT@Gel. (b) Image of excised tumors from each group (n = 5 mice per group). (c) Tumor growth curves (n = 5 mice per group). (d) Body weight change curve of mice (n = 5). (e–h) Individual tumor growth curves of mice in G1-G4, respectively (n = 5 per group). (i–l) Body weight changes of mice in G1-G4, respectively (n = 5 mice per group). Data are expressed as mean ± SD. Comparisons among multiple groups were analyzed by ANOVA followed by Tukey’s post hoc test. Statistical significance was assigned as: ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 After 18 days of treatment, tumor growth in the G4 group was markedly inhibited, and complete tumor regression was observed in some mice (Fig. 4b, c). In the surgical control group G1 and the G2, the mean tumor volume rapidly increased to ~ 1400 mm³ by day 18 after surgery (Fig. 4e, f). In contrast, implantation of G3 partially suppressed tumor growth, with a mean tumor volume of 555 mm³ (Fig. 4g). Notably, in the G4, tumor growth was strongly inhibited, with a mean tumor volume of only 126 mm³ (Fig. 4h). Moreover, body weight remained stable across all groups, with no significant differences observed (Fig. 4d, i-l). Consistently, Hematoxylin and eosin (H&E) staining of major organs, including the heart, liver, spleen, lungs, and kidneys, did not show obvious pathological alterations in any treatment group (Figure S10), indicating no significant systemic toxicity. Taken together, these results demonstrate that DOX/PDiT@Gel exhibits good biosafety while effectively suppressing tumor recurrence. To further examine whether DOX-containing formulations induced ICD-related changes in vivo, immunofluorescence staining of tumor sections was performed to assess CRT exposure and HMGB1 distribution. Compared with the PBS@Gel group, both the DOX@Gel group and the DOX/PDiT@Gel group showed increased CRT exposure and altered HMGB1 distribution, consistent with ICD-associated changes in tumor tissues (Figure S11). These findings support that DOX-containing hydrogel formulations can induce ICD-related responses in vivo. In vivo antitumor immune activation by DOX/PDiT@Gel Given that DOX/PDiT@Gel markedly suppressed tumor growth in vivo, we next examined the immunological mechanisms underlying its therapeutic effect. On day 18 of treatment (day 27 post-tumor inoculation), after euthanasia, tumors, spleens, and draining lymph nodes (dLN) were isolated from the mice for subsequent immune analysis (Fig. 5a). To assess the effect of each treatment on DCs activation within tumors, the expression of CD80, CD86, CD40, and MHC-II on CD11c⁺ cells was analyzed by FCM. Compared with the other groups, G4 showed the most robust DCs activation: compared with G1 and G3, the percentage of CD80⁺CD86⁺ cells among CD11c⁺ cells increased by 2.79-fold and 1.85-fold, respectively (Fig. 5b, S12). The frequencies of CD40⁺ and MHC-II⁺ cells reached 35.90% and 12.48%, respectively, representing the highest levels among all groups (Fig. 5c, d; S12). The dLN showed a similar tendency. Compared with the control group (G1), G4 exhibited 1.58-fold, 1.28-fold, and 2.04-fold increases in the proportions of CD80⁺CD86⁺, CD40⁺, and MHC-II⁺ cells among CD11c⁺ cells, respectively (Fig. 5e-g; S13). In both tumor tissues and dLNs, DOX + PDiT elevated the expression of markers associated with dendritic cell maturation, suggesting improved DC readiness to prime T-cell responses and a stronger immune activation along the tumor-draining lymph node axis. DOX/PDiT@Gel activates the immune system in vivo. (a). Schematic illustration of experimental design, created in https://BioRender.com.BALB/c mice were randomly assigned into four groups and subjected to different treatments: G1, PBS@Gel; G2, DOX@Gel, G3: DOX/AA@Gel; G4: DOX/PDiT@Gel. (b) FCM analysis of CD80+CD86+ cells in CD11c+ cells in tumor (n = 5 mice per group). (c) FCM analysis of CD40+ in CD11c+ in tumor. (d) FCM analysis of MHC-II+ cells in CD11c+ cells in tumor (n = 5 mice per group). (e) FCM analysis of CD80+CD86+ cells in CD11c+ cells in lymph node (n = 5 mice per group). (f) FCM analysis of CD40+ cells in CD11c+ cells in lymph node (n = 5 mice per group). (g) FCM analysis of MHC-II+ cells in CD11c+ cells in lymph node (n = 5 mice per group). (h) FCM analysis of CD4+ cells in lymph node (n = 5 mice per group). (i) FCM analysis of CD8+ cells in lymph node (n = 5 mice per group). (j-l) Representative FCM plots and quantification of Tcm and Tem in CD4⁺ T cells in spleen (n = 5 mice per group). (m-o) Representative FCM plots and quantification of Tcm and Tem in CD8⁺ T cells in spleen (n = 5 mice per group). Data are expressed as mean ± SD. Comparisons among multiple groups were analyzed by ANOVA followed by Tukey’s post hoc test. Statistical significance was assigned as: ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 We next examined T-cell responses in the dLN and spleen. Flow cytometric analysis revealed that, relative to G1, the proportions of CD4⁺ and CD8⁺ T cells in the dLN of G4 were increased by 15.97% and 15.84%, respectively (Fig. 5h, I; S14), indicating that this treatment promotes T-cell expansion and enhances effector responses. We further characterized splenic memory T-cell populations, focusing on CD44⁺CD62L⁻ effector memory T cells (Tem) and CD44⁺CD62L⁺ central memory T cells (Tcm) (Fig. 5j-o; S15). The proportion of CD4⁺ Tcm in G4 was 2.32-fold higher than in G1 and 1.63-fold higher than in G3 (Fig. 5k). Similarly, CD4⁺ Tem frequencies in G4 were 1.48-fold and 1.24-fold higher than those in G1 and G3, respectively (Fig. 5l). For CD8⁺ T cells, Tcm levels in G4 were 2.58-fold higher than in G1 and 1.65-fold higher than in G3 (Fig. 5n), while CD8⁺ Tem levels were 1.41-fold and 1.31-fold higher than those in G1 and G3, respectively (Fig. 5o). Collectively, these data indicate that DOX + PDiT not only enhances T-cell activation but also promotes the development of robust memory T-cell populations, thereby supporting long-term antitumor immunity. Overall, DOX/PDiT@Gel induced stronger immune activation both locally within tumors and in the dLN, accompanied by an increase in systemic memory T-cell subsets. These immunological changes were consistent with the pronounced in vivo tumor inhibition observed, supporting the potential of this platform to confer more durable antitumor immune protection. Systemic antitumor immunity against distant metastases after local DOX/PDiT@Gel Having confirmed that DOX/PDiT@Gel can suppress orthotopic tumors and enhance local immune responses in the postoperative recurrence model, we next established a bilateral 4T1 breast cancer model to assess whether it could elicit a systemic immune response. In BALB/c mice, the primary and secondary tumors were implanted on opposite flanks on days 0 and 6, respectively (Fig. 6a). On day 9, local intervention was performed at the site of the primary tumor, and the growth of the tumor on the opposite side was followed thereafter. The animals were randomly allocated into three groups: G1 (PBS@Gel), G2 (DOX@Gel), and G3 (DOX/PDiT@Gel). Compared with G1 and G2, the contralateral tumor volume in G3 was markedly reduced (Fig. 6b), suggesting that PDiT is capable of eliciting a systemic antitumor immune response. In line with the postoperative recurrence model, body weight remained comparable across all groups without significant differences (Fig. 6c), further supporting the good biosafety profile of PDiT. Moreover, survival analysis showed that overall survival was significantly prolonged in G3 (Fig. 6d). DOX/PDiT@Gel induces systemic antitumor immunity and suppresses distant tumors in vivo. (a) Schematic illustration of experimental design, created in https://BioRender.com.BALB/c mice were randomly assigned to three treatment groups: G1, PBS@Gel; G2, DOX@Gel; G3, DOX/PDiT@Gel. (b) Growth curves of distant tumors (n = 5 mice per group). (c) Body weight change curve of mice (n = 5 mice per group). (d) Survival curve of mice (n = 5 mice per group). (e) Representative FCM plot of CD80+ CD86+ cells in CD11c+ cells in lymph node (n = 5 mice per group). (f) Representative FCM plot of MHC-II+ cells in CD11c+ cells in lymph node (n = 5 mice per group). (g) FCM analysis of CD80+ CD86+ cells in CD11c+ cells in lymph node (n = 5 mice per group). (h) FCM analysis of MHC-II+ cells in CD11c+ cells in lymph node (n = 5 mice per group). (i) FCM analysis of CD40+ cells in CD11c+ cells in lymph node (n = 5 mice per group). (j) FCM analysis of CD69+ cells in CD3+ cells in spleen (n = 5 mice per group). (k) Representative FCM plots of CD44+ CD62L+ CD4+ T cells in spleen (n = 5 mice per group). (l) Representative FCM plots of CD44+ CD62L+ CD8+ T cells in spleen (n = 5 mice per group). (m) FCM analysis of Tem in CD4+ T cells in spleen (n = 5 mice per group). (n) FCM analysis of Tem in CD8+ T cells in spleen (n = 5 mice per group). (o) FCM analysis of TNF-α+ in CD8+ cells in spleen (n = 5 mice per group). (p) FCM analysis of IFN-γ+ in CD8+ T cells in spleen (n = 5 mice per group). Data are expressed as mean ± SD. Comparisons among multiple groups were analyzed by ANOVA followed by Tukey’s post hoc test. Survival was analyzed with Kaplan-Meier curves, and group differences were assessed using the log-rank (Mantel-Cox) test. Statistical significance was assigned as: ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 To further investigate whether PDiT promotes systemic immune activation, we analyzed lymph node DCs by flow cytometry. Compared with G1, the percentage of CD80⁺CD86⁺ cells among CD11c⁺ DCs in G3 rose by 3.91-fold, while MHC-II⁺ and CD40⁺ DC subsets showed 3.04-fold and 1.33-fold increases, respectively (Fig. 6e, j). These data indicate that PDiT-containing treatment promotes DCs maturation in dLN. In the spleen, immune profiling revealed that the proportion of CD69⁺ T cells in G3 was 45.6% greater than that in G1 (Fig. 6f, h; S16). Further analysis of memory T-cell subsets revealed that the frequencies of Tem within CD4⁺ and CD8⁺ T-cell populations in G3 were 102.96% and 173.00% higher than those in G1, respectively (Fig. 6k-n), indicating enhanced systemic T-cell activation and memory differentiation. Intracellular cytokine staining further showed that TNF-α and IFN-γ production by CD8⁺ T cells in G3 increased by 116.98% and 178.61%, respectively, compared with G1 (Fig. 6o, p; S17), consistent with markedly enhanced effector function. Collectively, the bilateral 4T1 model demonstrated that local treatment of the primary tumor with DOX + PDiT triggered a systemic antitumor immune response, which in turn suppressed the growth of the untreated tumor on the opposite side. This systemic effect was accompanied by enhanced maturation of lymph node DCs and broad activation of splenic T cells, including increased CD69 expression, expansion of effector-memory subsets, and elevated IFN-γ/TNF-α production by CD8⁺ T cells. Together, these findings support that PDiT amplifies antitumor immunity beyond the treated site, translating local therapy into durable, whole-body immune protection. Discussion Cancer vaccines are attractive because they can elicit tumor-specific T-cell responses. However, the effectiveness of conventional antigen-defined vaccines is often limited by their dependence on predefined targets, which may not adequately represent the heterogeneous and evolving antigenic landscape of real tumors [1, 2]. ISCVs provide an alternative by using the patient’s own tumor as a source of endogenous antigens [7, 8]. In this context, an important challenge is not simply to induce tumor cell death and antigen release, but to ensure that the released antigens are preserved, captured, internalized, and processed in a way that supports efficient cross-presentation. The present study was therefore aimed at this post-release step, focusing on how endogenous tumor antigens are handled after release rather than on further intensifying local tumor cell killing alone. From this perspective, the value of DOX/PDiT@Gel lies in integrating two coordinated functions within a single platform. DOX acts as an upstream trigger that promotes immunogenic tumor damage and antigen release, whereas PDiT acts downstream to improve antigen capture and subsequent intracellular processing. Although antigen capture-assisted ISCV strategies have been reported previously [7, 11], the current design places greater emphasis on linking extracellular antigen enrichment to intracellular trafficking events that favor MHC-I cross-presentation. This distinction is also reflected in the comparison between DOX/AA@Gel and DOX/PDiT@Gel. While both systems can interact with released proteins, the stronger antitumor and immune effects observed with DOX/PDiT@Gel indicate that simple adsorption or retention of extracellular antigens is unlikely to fully explain the therapeutic benefit [34]. Mechanistically, the current data indicate that PDiT is the principal bioactive component of this platform. The pH-dependent increase in protein capture, reduced lysosomal colocalization, enhanced antigen cross-presentation, and MG132 sensitivity collectively support the interpretation that PDiT facilitates endosomal/lysosomal escape and promotes access of internalized antigens to proteasome-dependent MHC-I processing. In parallel, DOX acts primarily as an upstream ICD-inducing trigger, as supported by CRT exposure and HMGB1 redistribution in tumor tissues. Accordingly, the platform appears to function through a coordinated sequence in which DOX induces ICD-related tumor cell changes and PDiT enhances subsequent antigen capture and intracellular processing, thereby contributing to improved dendritic-cell activation, T-cell priming, and systemic antitumor immunity in vivo. Conceptually, this design may help address an important limitation of conventional antigen-defined vaccination, namely the restricted coverage of the heterogeneous and evolving tumor antigen repertoire. By coupling endogenous antigen release with local antigen handling and more favorable intracellular processing for cross-priming, DOX/PDiT@Gel may provide a more adaptable strategy for personalized cancer immunotherapy. By contrast, the hydrogel component in this study should be interpreted primarily as an established local delivery matrix rather than as the main mechanistic focus. The oxidized dextran/carboxymethyl chitosan hydrogel used here has been characterized in previous studies, including its formation mechanism and general physicochemical properties [28, 29]. In the present work, it mainly serves as a local depot for formulation retention and release, thereby enabling localized administration of DOX and PDiT. Consistent with this role, the in vitro release results support gradual cargo release under the tested conditions; however, these data do not establish the actual in vivo release kinetics, retention time, or degradation behavior of the formulation after implantation. The present study is limited by the lack of a full understanding of how PDiT behaves in vivo. Although no overt acute systemic toxicity was observed under the present experimental conditions, the long-term biodistribution, degradation, metabolic fate, and clearance behavior of PDiT remain to be determined. Further studies are therefore needed to directly track PDiT in vivo and to more systematically evaluate its pharmacokinetic profile and long-term biosafety. Taken together, these results suggest that DOX/PDiT@Gel is a functionally integrated ISCV platform. Within this system, DOX promotes ICD-associated responses, PDiT strengthens antigen capture and intracellular processing, and the hydrogel enables localized administration, supporting its potential utility in personalized cancer immunotherapy. Materials and methods Materials and instruments 2-(Diisopropylamino) ethyl methacrylate (DIPAMA), TAT(CYGRKKRRQRRR) were purchased from Sigma-Aldrich (St. Louis, MO, USA). 4-cyano-4-(lauryl thioalkyl thiocarbonyl) thiocalkylvaleric acid (CDTPA), 2-Mercaptoethanol, Triethylamine (TEA), Methylene chloride (DCM), deuterated chloroform, deuterated heavy water, deuterated hydrochloric acid, and sodium deuterium oxide were purchased from Annegy Chemical (Shanghai, China). Ovalbumin (OVA) was purchased from Thermo Fisher Scientific (Massachusetts, USA). OVA-FITC and penicillin-streptomycin (PS) was purchased from Solarbio Technology Co., Ltd (Beijing, China). Cell Counting Kit-8 (CCK-8), lysosomal probe and Antifade Mounting Medium with DAPI were purchased from Beyotime Biotechnology Co., Ltd (Beijing, China). Phalloidin-FITC were purchased from Yeasen Biotechnology Co., Ltd (Shanghai, China). RPMI-1640 medium and Dulbecco’s Modified Eagle Medium (DMEM) were purchased from Beijing Lanjie Ke Technology Co., Ltd (Beijing, China). Fetal bovine serum (FBS) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Phosphate-buffered saline (PBS) was purchased from Meilun Biotechnology Co., Ltd (Dalian, China). Mouse granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin-4 (IL-4) were purchased from PeproTech (New Jersey, USA). The flow cytometry antibodies were purchased from BioLegend (San Diego, California, USA). Rehydragel@LV Alum Adjuvant was purchased from Beijing Biosynthesis Biotechnology Co., Ltd (Beijing, China). The antibodies used for flow cytometry analysis were purchased from BioLegend (San Diego, USA). 1H NMR was tested using Avance III 600 MHz NMR spectrometer (Bruker, Switzerland). Transmission electron microscopy (TEM) was performed using a JEM-F200 instrument (JEOL, Japan). Dynamic light scattering (DLS) measurements was conducted with a laser particle size analyzer (Malvern Instruments, UK). The absorbance was measured by microplate reader (Thermo, USA). Confocal laser scanning microscope (CLSM) images were shooting at LSM 880 NLO & DuoScan confocal microscopy system (ZEISS, Germany). FCM analysis was carried at Becton Dickinson flow cytometer (BD, USA). Cell lines and animals The 4T1, B16F10, and CT26 cell lines were purchased from Zhong Qiao Xin Zhou Biotechnology Co., Ltd (Shanghai, China) and were maintained in our laboratory. The culture medium for all cell lines contained 10% FBS and 1% PS, and the cells were incubated at 37 °C under humidified conditions with 5% CO2. Female BALB/c mice aged 6 weeks were supplied by SPF (Beijing) Biotechnology Co., Ltd. (Beijing, China). All mice were kept in a specific pathogen-free facility with a room temperature of 23 °C, relative humidity of 40%-60%, and a 12 h light/dark schedule. Synthesis of PEG-DIPAMA-TAT Synthesis of PEG-DIPAMA-PDSMA The synthesis of CDTPA, PDSEMA, and PEG-DIPAMA-PDSMA was carried out according to previously reported methods [35, 36]. Synthesis of PEG-DIPAMA-TAT The thiolated peptide TAT-SH was attached to PDSE residues via a disulfide exchange reaction carried out in methanol/water (5:1, v/v). After the reaction, purification by dialysis against deionized water afforded PEG-DIPAMA-TAT. Preparation and characterization of PDiT The PDiT nanoparticles were prepared using the pH titration method. For complex preparation, 1 mg of PDiT was first dissolved by ultrasonication in 1 mL of deionized water supplemented with 150 mM NaCl and 20 mM HCl. The solution pH was carefully brought to 7.4 through incremental addition of 0.5 M NaOH in 1 µL portions. PDiT was then incubated with OVA in deionized water at 1 mg/mL for 4 h at room temperature. After incubation, the sample was centrifuged in an Amicon® Ultra-0.5 centrifugal filter device (3k MWCO) at 3500 × g for 15 min at 4 °C and washed three times with PBS to remove free OVA, thereby obtaining the PDiT-OVA complex. Changes in particle size before and after OVA association were analyzed by DLS and TEM. PDiT pKa value determination The PDiT solution was prepared at 1 mg/mL in 10 mL deionized water supplemented with 150 mM NaCl and 20 mM HCl. The pH was first adjusted to 2.0 using 0.5 M NaOH. Next, 0.5 M NaOH was introduced in 1 µL portions, and the pH value was measured after each addition with a pH electrode until it increased to 11. The protonation level of PDiT was then analyzed according to the main inflection points in the titration profile. The lower inflection point was regarded as the fully protonated form, while the upper inflection point was considered to indicate complete deprotonation. The apparent pKa value of PDiT was determined by the pH at which the degree of protonation reached 50%. Hemolysis assay Anticoagulated peripheral blood was collected from healthy rats and centrifuged at 3000 rpm for 10 min at 4 °C to separate and remove plasma together with white blood cells. The packed red blood cells were then diluted with PBS to obtain a 2% erythrocyte suspension. In the hemolysis test, samples were exposed to PBS under different pH conditions, while pure water and PBS were used as the positive and negative controls, respectively. The erythrocyte suspension was mixed with an equal volume of PDiT nanoparticle solution and incubated for 30 min at 37 °C. Following centrifugation, the absorbance of the supernatant was measured at 540 nm, and the hemolysis rate was subsequently calculated. CCK-8 cell viability assay DC2.4, 4T1 and NIH-3T3 cells were seeded into 96-well plates at a density of 1 × 10⁴ cells per well and allowed to adhere overnight. The cells were then incubated for 24 h in RPMI-1640 complete medium containing different concentrations of PDiT or extracts prepared from different hydrogel formulations. Untreated cells cultured in medium alone served as controls. After incubation, the supernatant was removed, and CCK-8 solution was added according to the manufacturer’s instructions. Plates were incubated at 37 °C for 1 h, and absorbance at 450 nm was measured using the microplate reader. In vitro antigen capture assay Culture 4T1 cells in a 100 mm diameter dish until they reach approximately 80% confluence. After gentle washing with cold PBS, the cells were incubated for 48 h in DMEM supplemented with 10 µM DOX in the absence of FBS to facilitate antigen release. The supernatant was then harvested and clarified by centrifugation at 5000 × g for 10 min to eliminate insoluble cell fragments. Subsequently, 2 mg of PDiT nanoparticles were mixed with 10 ml of the collected supernatant and incubated for 4 h, with the pH adjusted to 7.4 or 6.8. Following incubation, the nanoparticles were rinsed with PBS at the same pH and processed with a centrifugal filter unit (Amicon® Ultra-15, 100k MWCO) at 5000 × g for 15 min at 4 °C. This washing procedure was performed no fewer than three times to eliminate free proteins, thereby obtaining PDiT bound with captured TDPAs. The samples were then mixed with a 10-fold volume of 50 mM ammonium bicarbonate (pH 8) and left at room temperature for 30 min. After centrifugation at 12,000 × g for 10 min at 4 °C, the supernatant was recovered for the eluted TDPAs. Identification by liquid chromatography mass spectroscopy (LC-MS) analysis of proteins captured by PDiT Sample preparation Protein solutions were first treated with 6 M urea and 10 mM DTT in 50 mM ammonium bicarbonate at 37 °C for 45 min. Alkylation was then carried out with 10 mM iodoacetamide in 50 mM ammonium bicarbonate for 1 h at room temperature under light-protected conditions. After dilution with 50 mM ammonium bicarbonate to lower the urea concentration to 2 M, the samples were digested overnight at 37 °C with trypsin at an enzyme-to-protein ratio of 1:40. Digestion was stopped by adding formic acid to 0.1% final concentration. The resulting samples were dried in a vacuum concentrator and redissolved in 0.1% formic acid aqueous solution for LC-MS analysis. Instrument methods Peptides were resolved on a chromatographic system consisting of a loading column (100 μm × 2 cm) and a C18 analytical capillary column (75 μm × 15 cm), which were packed in-house with 1.9 μm C18 resin (Dr. Maisch GmbH, Germany). The mobile phases were defined as follows: phase A, water with 0.1% formic acid; phase B, 80% acetonitrile containing 0.1% formic acid. Separation was performed on a Dionex Ultimate 3000 RPLC nano system (Thermo Fisher Scientific). During LC analysis, the gradient was kept at 2% B for the first 8 min, then raised to 10% B over 8–9 min, to 44% B over 9–63 min, and finally to 99% B over 63–68 min. Samples were analyzed on an Orbitrap Fusion LUMOS Tribrid Mass Spectrometer equipped with an EASY-Spray ionization source (Thermo Fisher Scientific). The spray voltage was set at + 2.0 kV relative to ground, and the inlet capillary temperature was maintained at 320 °C. Peptide precursor survey scans were recorded in the Orbitrap across an m/z range of 350–1600 with an AGC target of 400,000, a 50 ms maximum injection time, an RF lens of 30%, and a resolving power of 60,000 at m/z 200. Monoisotopic precursor selection was applied to peptide isotopic envelopes, and precursor ions with charge states from 2 to 7 were chosen for data-dependent MS/MS acquisition using a 3 s cycle. A dynamic exclusion duration of 15 s was used with a ± 10 ppm window around the monoisotopic precursor mass. During HCD acquisition, the scan range was determined automatically. Precursor selection was performed with the quadrupole using an isolation width of 1.6 Th. Fragment ions were analyzed in the Orbitrap with the first mass fixed at 110 Th, an AGC target value of 50,000, a maximum injection time of 30 ms, HCD energy set to 30%, and an Orbitrap resolution of 15,000. Data analysis Mass spectrometry data were analyzed with Thermo Proteome Discoverer 2.4 against the reviewed SwissProt mouse database (TaxID 10090, including subtaxonomy; 17,045 protein entries). Carbamidomethylation of cysteine was set as a fixed modification, whereas methionine oxidation was defined as a variable modification. The mass tolerance was set to 10 ppm for precursor ions and 0.02 Da for-fragment ions, with up to two missed cleavages allowed. Peptide spectral matches (PSMs) were filtered using the Percolator algorithm, and q-values were controlled at a false discovery rate (FDR) of 1%. Cell uptake capacity of PDiT For antigen uptake analysis, DC 2.4 cells were plated in 24-well plates at a density of 2 × 10⁵ cells per well, followed by different treatments including PBS, OVA-FITC (50 µg/mL), PDiT (30 µg/mL) + OVA-FITC (50 µg/mL). RPMI-1640 complete medium was maintained at pH 6.8. Cells were collected at 1, 4, and 7 h following incubation, rinsed three times with cold PBS, and then analyzed by FCM. To evaluate the internalization efficiency of PDiT by DCs, DC 2.4 cells were plated in 24-well plates at a density of 2 × 10⁵ cells per well on the cell culture slide. 30 µg/mL of Cy5-labeled PDiT was added to each well in the experimental group. RPMI-1640 complete medium was maintained at a slightly acidic pH 6.8. At 0, 1, and 4 h after incubation, cells were washed three times with cold PBS and fixed at room temperature for 15 min using fixation solution. The cells were subsequently rinsed with PBS, treated with 0.1% Triton X-100 for 5 min to permeabilize the membrane, and then washed three more times with PBS. Cells were then labeled with FITC-tagged phalloidin for 30 min, followed by three washes with PBS. Afterward, the cell-bearing slides were sealed with mounting medium containing DAPI, stored in the dark at 4 °C, and observed under a laser confocal microscope. Endocytosis inhibition assay DC2.4 cells were seeded into 24-well plates at a density of 2 × 105 cells per well using complete RPMI-1640 medium at pH 6.8. Before antigen exposure, the cells were pre-incubated for 30 min with pathway inhibitors, including chlorpromazine (CPZ, 20 µM) to block clathrin-dependent endocytosis, amiloride (25 µM) to suppress macropinocytosis, genistein (700 µM) to interfere with caveolae-mediated uptake, and methyl-β-cyclodextrin (m-β-CD, 5 mM) to disrupt caveolae-related internalization. After inhibitor treatment, OVA-FITC (50 µg/mL) and PDiT (30 µg/mL) were introduced into each well for a further 4 h incubation. The control groups consisted of the PBS group and the OVA-FITC group, which were used for FCM analysis. The same treatment was applied for laser confocal imaging, with the exception that OVA-FITC was not included, and only PDiT-Cy5 was used. Endosome escape assay DC2.4 cells were plated on cell culture slides placed in 24-well plates at a density of 2 × 105 cells per well. Each well received either OVA-FITC (50 µg/mL) alone or OVA-FITC (50 µg/mL) supplemented with PDiT (30 µg/mL), followed by incubation for 4 h. RPMI-1640 complete medium was maintained at a slightly acidic pH of 6.8. Once the incubation step was completed, the medium was removed and the cells were rinsed three times with PBS. A preheated red lysosomal probe was then added, and the cells were incubated at 37 °C for 1 h before being washed again with PBS three times. The cells were subsequently fixed at room temperature for 15 min using fixation solution and then washed three more times with PBS. At the end of the procedure, the cell-bearing slides were sealed with DAPI-containing antifade mounting medium and stored in the dark at 4 °C. Imaging was subsequently performed with a laser confocal microscope. Proteasome inhibition assay in DC2.4 cells To evaluate whether PDiT-enhanced antigen processing was dependent on proteasome activity, DC2.4 cells were seeded into 24-well plates at 2 × 105 cells per well and maintained overnight before further treatment. Cells were pre-incubated with MG132 (10 µM) for 30 min before treatment with either OVA (10 µg/mL) alone or OVA (10 µg/mL) plus PDiT (30 µg/mL) for 24 h. The treated cells were then collected, washed with ice-cold PBS, and labeled with anti-Zombie Aque-BV510, anti-CD11c-PE-Cy7, and anti-H-2Kb-SIINFEKL-PE antibodies, followed by FCM. In vitro bone marrow-derived dendritic cells (BMDCs) activation assay BMDCs were isolated from male C57BL/6 mice as previously described [37, 38]. Briefly, the tibias and femurs were dissected from the mice and rinsed with PBS. The isolated cells were cultured in RPMI-1640 medium supplemented with 20 ng/mL GM-CSF and 10 ng/mL IL-4. Fresh medium was replaced on days 3 and 5, and the cells differentiated into BMDCs by day 7. BMDCs were collected on day 7 and seeded into 24-well plates at a density of 2 × 10⁵ cells per well. The cells were exposed to 10 µg/mL OVA together with PDiT in pH 6.8 RPMI-1640 medium for 24 h. They were subsequently harvested and labeled with anti-Zombie Aque-BV510, anti-CD11c-PE-Cy7, anti-CD40-APC, anti-CD80-FITC, anti-CD86-BV421, and anti-H-2Kb-SIINFEKL-PE before FCM analysis. In vitro release study and post-release characterization of PDiT To avoid signal interference during quantification, DOX@Gel and PDiT-Cy5@Gel were prepared separately and incubated in 10-fold volume of PBS (pH 6.8) at 37 °C under static conditions. At each designated time point (days 0–7), the release liquid was withdrawn and replaced with an equivalent volume of fresh PBS. The released DOX and PDiT-Cy5 were quantified using a microplate reader, and cumulative release profiles were calculated accordingly. To further assess whether the hydrogel affected the physicochemical properties of PDiT, the PDiT released from DOX/PDiT@Gel was collected and analyzed by dynamic light scattering (DLS) to determine its hydrodynamic size and distribution, thereby evaluating its particle size and stability after release. Determination of encapsulation efficiency of DOX and PDiT The encapsulation efficiency (EE) of DOX and PDiT in the hydrogel was determined by an indirect method. Briefly, after hydrogel formation, the gel was washed with PBS to quantify the unencapsulated fraction. The EE was calculated as: EE (%) = [(W0 − Wfree) / W0] × 100%. where W0 is the initial feeding amount of DOX or PDiT during hydrogel preparation, and Wfree is the amount of unencapsulated cargo detected in the wash solution. In vivo postoperative tumor recurrence model On day 0, female BALB/c mice received a subcutaneous injection of 1.0 × 106 4T1 cells into the dorsal flank. Once the tumor volume reached around 100 mm3 on day 9, partial surgical excision was carried out, with roughly 90% of the tumor mass removed to establish a model of postoperative residual disease. Immediately after surgery, preformed hydrogels (200 µL) were implanted into the resection bed. Mice were randomly assigned to four groups: PBS@Gel, DOX@Gel (DOX, 5 mg/kg), DOX/AA@Gel (DOX, 5 mg/kg; aluminum adjuvant, 5 mg/kg), and DOX/PDiT@Gel (DOX, 5 mg/kg; PDiT, 5 mg/kg). Changes in tumor volume and body weight were recorded at 3-day intervals. At day 27, tumor tissues, dLN, and spleens were isolated for subsequent FCM and histological evaluations. In vivo bilateral tumor model To investigate systemic antitumor immune responses, female BALB/c mice were used to establish a bilateral 4T1 tumor model. On day 0, 1.0 × 106 4T1 cells were implanted subcutaneously on one side of the back, and a second inoculation of 5.0 × 105 cells was performed on the contralateral side on day 6. On day 9, the primary tumor was partially resected to remove approximately 90% of the tumor mass, and preformed hydrogels (200 µL) were implanted into the surgical bed. Mice were randomly assigned to three groups: PBS@Gel, DOX@Gel (DOX, 5 mg/kg), and DOX/PDiT@Gel (DOX, 5 mg/kg; PDiT, 5 mg/kg). Changes in tumor size and body weight were assessed every 3 days. At day 27, dLN and spleens were isolated for immunological analysis, whereas some animals were retained for survival studies. Histological analysis After fixation in 4% paraformaldehyde, major organs and tumor samples were embedded in paraffin, cut into sections, and processed for hematoxylin and eosin (H&E) staining following standard protocols. Immunofluorescence staining of tumor sections Tumor samples were fixed, embedded in paraffin, sectioned, and then subjected to immunostaining with antibodies against CRT and HMGB1. The nuclei were counterstained with DAPI. FCM analysis of immune cells Tumor tissues, dLN, and spleens were processed into single-cell suspensions by mechanical disruption, with red blood cell lysis performed when required. According to the experimental design, the cells were then stained with fluorophore-labeled antibodies targeting dendritic-cell, T-cell, and memory T-cell markers. For intracellular cytokine analysis, the cells were first stimulated with PMA/BFA, followed by surface staining, fixation, permeabilization, and intracellular staining. All samples were subsequently analyzed by FCM. Statistical analysis All statistical analyses were carried out using GraphPad Prism software (version 8.0). Data are shown as mean ± SD. Differences between two groups were analyzed using Student’s t-test, whereas comparisons involving more than two groups were conducted by one-way analysis of variance (ANOVA) followed by Tukey’s multiple-comparison test. Survival was analyzed with Kaplan–Meier curves, and group differences were assessed using the log-rank (Mantel–Cox) test. Statistical significance was assigned as follows: ns P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements The authors gratefully acknowledge the Analytical Instrumentation Center of Peking University for support with LC–MS analysis, as well as the Flow Cytometry Core of the National Center for Protein Sciences at Peking University for technical assistance. Abbreviations PEG-DIPAMA-TAT Doxorubicin Tumor-associated antigens In situ cancer vaccines Tumor-derived protein antigens Antigen-presenting cells Dendritic cells Cytotoxic T lymphocyte Immunogenic cell death 2-diisopropylaminoethyl methacrylate Oxidized dextran/carboxymethyl chitosan hydrogel Reversible addition-fragmentation chain transfer Nuclear magnetic resonance Gel permeation chromatography Dynamic light scattering Ovalbumin Transmission electron microscopy Liquid chromatography-tandem mass spectrometry Cy5-labeled PDiT FITC-labeled OVA Confocal laser scanning microscope Chlorpromazine 5-(N-ethyl-N-isopropyl) amiloride Methyl-β-cyclodextrin Bone marrow–derived dendritic cells Hematoxylin and eosin Draining lymph nodes Effector memory T cells Central memory T cells Fetal bovine serum Phosphate-buffered saline Granulocyte-macrophage colony-stimulating factor Interleukin-4 Standard deviation One-way analysis of variance Author contributions Author contributions: Z. B. and S. Z. contributed equally to this work.Conceptualization: Z. B., S. Z., S. L., X. R.Methodology: Z. B., S. Z., G. J.Investigation: Z. B., S. Z., G. J., Z. Y., S. L., L. W., B. Y., L. W., Z. S., S. X., X. X., L. L.Formal analysis: Z. B.Supervision: S. L., X. R.Writing—original draft: Z. B.Writing—review & editing: Z. B., S. Z., T. L., J. Z., S. L., X. R.All authors reviewed the manuscript. Funding This work is supported by the National Key R&D Program of China (2025YFE0114200), National Natural Science Foundation of China (52273114, 32571686), Key Science and Technology Project of Henan Province (242300421207), Open Project Program of Yaoshan laboratory (2024001), the Ministry of Science and Higher Education of the Russian Federation (grant no. 075-15-2025-639 from 20.08.2025). Data availability The data underpinning the findings of this study are accessible within the article and the supplementary information files. Declarations Ethics approval and consent to participate All animal studies described in this work were conducted in accordance with protocols reviewed and approved by the Peking University Laboratory Animals Welfare Ethics Committee (Approval ID: MSE-LSX-2). Declaration of generative AI and AI-assisted technologies in the manuscript preparation process ChatGPT was used during manuscript drafting to assist with language polishing. All output was carefully examined and further edited by the authors, and the authors remain fully accountable for the content of the published work Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Zhuoling Bi and Shucheng Zhang contributed equally to this work. Contributor Information Shixian Lv, Email: lvshixian@pku.edu.cn. Xiaozhi Rong, Email: rongxiaozhi@ouc.edu.cn. References Associated Data Data Availability Statement The data underpinning the findings of this study are accessible within the article and the supplementary information files.