The preparation process of DmTFH was illustrated in Fig. 1. First, DOX was loaded onto mBSA through electrostatic interactions. This mixture was then subjected to repeated sonication to generate DOX@mBSA MBs. Next, the MPNs, formed by Fe (III), TA, and HADA (Fig. S1), were assembled on the surface of these MBs. The Fe (III) ions interacted with both TA and HADA to form a stable MPN network around the MBs. As the air inside the MBs escaped naturally, DOX and mBSA with therapeutic and immunomodulatory properties were retained within the capsule. This resulted in the formation of DOX-loaded metal-phenolic capsules (DmTFH) that simultaneously possessed both drug delivery and immunomodulatory capabilities. Optical microscopy (OM), scanning electron microscopy (SEM), and transmission electron microscopy (TEM) images of DmTFH were shown in Fig. 2A to C, indicating the successful fabrication of DmTFH. High-angle annular dark-field (HAADF) microscopy and energy-dispersive x-ray (EDX) spectroscopy revealed that the capsule primarily contained C, O, N, and Fe elements (Fig. 2D). The elemental quantitative analysis further showed the presence of 69.70 at % C, 14.65 at % N, 15.09 at % O, and 0.57 at % Fe in DmTFH (Fig. S2). The atomic force microscopy (AFM) image showed that the bilayer thickness of the dried capsule was 19.876 nm. Therefore, the corresponding shell thickness of the capsule was approximately 10 nm. The hydrodynamic diameter of DmTFH was about 650 nm with a polydispersity index of 0.488 based on the dynamic light scattering (DLS) analysis (Fig. 2F). Compared to mTFH (capsules without DOX loading), the DLS curve of DmTFH exhibited a slightly broader peak shifted to a larger diameter, indicating an increase in capsule size due to DOX loading. Zeta potential analysis indicated that after loading positively charged DOX, the surface potential of DmTFH slightly decreased (Fig. S3). Given the inherent fluorescence properties of DOX, the confocal laser scanning microscopy (CLSM) image showed that DmTFH exhibited red fluorescence, indicating the presence of DOX within the DmTFH (Fig. 2G). The characteristic peak at ~484 nm in the ultraviolet–visible (UV–Vis) spectrum of DmTFH confirmed that DOX has been successfully loaded into the capsules (Fig. 2H). We observed that the drug loading capacity of DmTFH increased with the initial DOX feed concentration, but the drug loading efficiency decreased as the feed concentration further increased (Fig. S4). Therefore, we chose an initial DOX feed concentration of 1 mg/ml as the optimal concentration for subsequent studies. At this feed concentration, each BSA molecule was conjugated with approximately 4 DOX molecules [1]. Fourier transform infrared spectrum (FTIR) showed characteristic peaks for DOX at 1,730 cm⁻¹ (the vibration of aromatic aldehyde C=O) and 1,284 cm⁻¹ (C–O–C vibrations) [13]. These peaks were also detected in DmTFH, indicating the presence of DOX within the DmTFH structure (Fig. 2I). The secondary fitting curves of FTIR reflected the conformational changes in mBSA after DOX loading (Fig. 2J). The result suggested that the electrostatic interactions between DOX and mBSA increased the stability of the mBSA molecule, as indicated by the decreased content of random coil and increased content of β-sheet and α-helix [14,15]. The charge effects also influenced the β-extended part, showing a red shift but with a decrease in content. The turn and β-turn parts showed no significant changes, and the detailed quantitative analysis was summarized in Table S1. In addition, MPNs were widely reported to exhibit pH-responsive characteristics [16,17]. When pH > 7, they formed tris-complexes; when pH was between 3 and 6, they formed bis-complexes; and when pH < 2, they disassembled into mono-complex forms. Thus, the pH-responsive behavior of DmTFH was further investigated (Fig. 2K). At a pH of 5.0, the release profile showed an initial burst, with a subsequent sustained release extending over 24 h, resulting in about 60% of the drug being released. However, the drug release was much lower at pH 7.4, with only about 5% drug release observed over the same period. The inset image showed the color change of the DmTFH solution at pH 5.0 and pH 7.4, indicating the pH-disassembly properties of the DmTFH. Moreover, we found that after 24 h in phosphate-buffered saline (PBS) solution (pH 7.4), DmTFH released less drug compared to DmFH, which was formed by only coating HADA and Fe^III on DOX-loaded mBSA MBs (Fig. S5A). The UV–Vis spectrum demonstrated that DmTFH had a significantly higher drug loading capacity compared to DmFH (Fig. S5B and C). These results implied that the MPN structure in DmTFH, formed by Fe (III), TA, and HADA, was more stable, leading to less drug release and increased drug loading capacity. Given these findings, we further investigated the assembly mechanism of Fe (III) with TA and HADA on the surface of MBs. The secondary fitting curve of UV–Vis spectra of mTFH showed 5 distinct peaks (Fig. 2L). Peak 1 can be attributed to the broad absorption of BSA in aqueous solution. Peak 2 was associated with the broad absorption of BSA and TA. Notably, peak 3 represented the conjugation between Fe^III and deprotonated phenolic hydroxyl groups. Peak 4 may be attributed to the broad absorption of Fe^III in aqueous solution, and peak 5 indicated the presence of tris-complexes, characterized by the ligand-to-metal charge transfer band (LMCT) [16,18,19]. The secondary fitting curve of mFH lacked a clear LMCT band showing only the conjugation between Fe^III and deprotonated phenolic hydroxyl groups of HADA (Fig. 2M). This was likely due to the steric hindrance effect of HADA, which prevented the formation of local tris-type complexes [18]. These results confirmed that the stable MPNs formed by Fe^III, TA, and HADA on MBs were primarily composed of tris-complexes formed between TA and Fe^III, along with conjugations between Fe^III and deprotonated phenolic hydroxyl groups of both TA and HADA (Fig. 1).

After the successful loading of DOX, we further assessed the in vitro therapeutic effect of DmTFH. First, the standard cell counting kit-8 (CCK-8) assay result showed that drug treatment markedly decreased the survival rate of 4T1 cells. Notably, DmTFH exhibited higher cytotoxicity against 4T1 cells compared to the DOX group alone at 24 h with DOX concentrations ranging from 0.1 to 2.5 μg/ml (Fig. 3A and C). However, with the increase in concentration or incubation time, there was no marked variation in cell viability observed among the 2 groups (Fig. 3A and B). To further assess cell death under different drug concentrations, live/dead cell staining was performed. Consistent with the CCK-8 results, the staining results demonstrated that both free DOX and DmTFH significantly induced tumor cell death at a high concentration (10 μg/ml), as indicated by the high proportion of dead cells (Fig. 3E and F). At a lower concentration (2.5 μg/ml), the DmTFH group exhibited a higher number of dead cells (indicated by red fluorescence) compared to the DOX-only group after 24 h (Fig. 3D and F). Notably, at the same concentration, no substantial cell death was observed in the free DOX group at either 24 or 72 h, although the CCK-8 assay revealed a marked decrease in cell viability (Fig. 3D and Fig. S6). Since the cell viability measured by the CCK-8 assay reflected a combination of inhibited cell proliferation and cell death, we hypothesized that low-dose DOX mainly exerted its cytotoxic effect by suppressing cell proliferation. This was supported by the observation of a significant reduction in the number of living cells in the DOX-treated groups at 72 h (Fig. S6). These findings suggested that, at this concentration, free DOX primarily exerted its cytotoxic effects by inhibiting cell proliferation, which was consistent with previous reports [20]. The combination treatment with capsules can enhance the antitumor efficacy of low-concentration DOX by simultaneously inhibiting cell proliferation and promoting cell death. Studies have shown that breast cancer cells are highly sensitive to ROS, making these cells prone to lethal oxidative stress [21]. Previous research has demonstrated that DOX can enter mitochondria and capture an electron from the mitochondrial electron transport chain (mETC), resulting in the formation of a semiquinone intermediate. This intermediate subsequently transfers the electron to O₂ through redox cycling, ultimately leading to the production of ROS [2,22,23]. Our previous study and other studies have also reported that glycated proteins can elevate ROS levels by activating intracellular signaling pathways or impairing mitochondrial function [9,24,25]. Therefore, we speculated that the enhanced anticancer efficacy of low-concentration DmTFH might be attributed to elevated intracellular ROS levels. To test this, ROS accumulation in 4T1 cells was measured using the DCFH-DA (2′,7′-dichlorofluorescin diacetate) probe following different treatments. As illustrated in Fig. 3H and I, the DOX group exhibited only a slight increase in ROS production compared to the control group. This phenomenon has also been reported in other studies, primarily since the preferential accumulation of free DOX in the nucleus instead of in the mitochondria prevents DOX molecules from efficiently capturing electrons from the mETC, and the hypoxic tumor microenvironment limits the availability of sufficient O₂ for redox cycling-mediated ROS production [23]. In contrast, substantial ROS-positive cells were detected in groups treated with glycated protein-containing capsules (mTFH and DmTFH). To verify the essential role of the mannose-glycated BSA in ROS generation, the fluorescence from cells treated with the non-glycated capsule (TFH) was also detected. The results showed that only faint green fluorescence was observed, indicating TFH’s limited ability to induce ROS. This limitation could be attributed to the small amount of Fe³⁺ present on the capsule’s shell, which was insufficient for effective chemodynamic therapy to generate ROS (Fig. S2). These results indicated that the superior antitumor efficacy of low-concentration DmTFH was ascribed to the synergistic action of glycated capsules and DOX. On the one hand, DOX released from the capsule can directly reduce tumor cell viability. On the other hand, mannose-glycated BSA in the capsule induced ROS generation in 4T1 cells, thereby enhancing oxidative stress in tumor cells (Fig. 3G).

DOX has been widely reported to promote the exposure or release of DAMPs from tumor cells, such as CRT, ATP, and HMGB1. These danger signals can subsequently trigger the immune system to elicit tumor-specific cytotoxicity. In addition, ROS plays a pivotal role in anthracycline-induced ICD, such as facilitating the exposure of CRT on the cell surface [26]. Therefore, we evaluated whether DmTFH treatment may elicit ICD in tumor cells and enhance the release of DAMPs. First, we used immunofluorescence staining to evaluate the transfer of CRT protein from the endoplasmic reticulum to the cell surface after low-dose drug treatment. Immunofluorescence images revealed that low concentrations of DOX-treated cells showed only a small amount of CRT efflux, evident from the faint green fluorescence observed on the 4T1 cell surface (Fig. 4A). In contrast, a notable increase in CRT expression was detected on 4T1 cells following 24 h of exposure to DmTFH. Further quantitative assessment revealed that free DOX was also able to induce CRT expression compared to the control group, although the difference did not reach statistical significance (P = 0.0590). Notably, 4T1 cells treated with DmTFH displayed a mean fluorescence intensity (MFI) of CRT expression twice as high as that in the DOX group, indicating that DmTFH more effectively promoted CRT exposure, benefiting from the combined effects of DOX and ROS (Fig. 4B). Additionally, we evaluated the release of ATP and HMGB1. The results showed that ATP release from 4T1 cells was significantly increased following DmTFH treatment compared to the control group (Fig. 4C and Fig. S7). Moreover, both DOX- and DmTFH-treated cells exhibited a marked increase in HMGB1 release relative to the control and mTFH-treated groups (Fig. 4D and Fig. S7). However, no significant difference was observed between the DmTFH and free DOX groups, suggesting that HMGB1 release induced by DmTFH was primarily attributable to the effect of the loaded DOX. Collectively, these findings suggested that while free DOX alone can induce certain levels of ICD, its combination with the capsule significantly amplified CRT exposure and ATP release, thereby potentially enhancing the overall immune response. To further evaluate the immune response mediated by ICD, bone marrow-derived dendritic cells (BMDCs) were isolated (Fig. S8) and subsequently cocultured with the supernatants of 4T1 cells that had been treated with various groups. The supernatants contained DAMPs and debris from the 4T1 cells, which can promote DC maturation. Therefore, we conducted the flow cytometric analysis to assess DC maturation in each treatment group by labeling major histocompatibility complex class II (MHC II) and CD86. DmTFH was more effective in inducing DC maturation compared to other groups, with a significant increase to 16.5% of mature DCs (Fig. 4F and G). This indicated that the DmTFH-mediated ICD effect promoted the maturation of DCs. Notably, in addition to contacting tumor cells, drug carriers in the tumor microenvironment may directly act on immune cells. Therefore, low-concentration DmTFH (DOX concentration: 1 μg/ml) was incubated directly with immature BMDCs to study their direct immunostimulatory effects. We unexpectedly found that the proportion of mature BMDCs was approximately 13.8% in the mTFH group, representing a 1.8-fold increase compared to the group treated with free DOX (Fig. 4H and I). This finding may be attributed to mBSA in the capsule, as glycated BSA has been reported to induce DC maturation, as shown by heightened levels of CD1a, CD40, CD80, CD83, CD86, and MHC class II molecules [27]. Among the tested formulations, the highest proportion of CD86⁺MHC II⁺ BMDCs was detected in the DmTFH group, indicating its strong direct immunostimulatory effect on BMDCs due to the synergistic action of DOX and mBSA. Furthermore, numerous studies have shown that M1 macrophages facilitate DC maturation and boost antitumor immunity by secreting related paracrine signaling [28,29]. Our previous research found that glycated proteins can influence macrophage polarization toward the M1 phenotype via the activation of the nuclear factor κB (NF-κB) pathway [9]. Thus, we further investigated whether DmTFH could indirectly promote the maturation of DCs by modulating M1 polarization of macrophages. After treating RAW 264.7 cells (mouse macrophages) with interleukin-4 (IL-4) and IL-13 to induce an M2 phenotype (typical morphology shown in Fig. S9), we evaluated the ability of different treatment groups to induce M2-to-M1 repolarization in macrophages using flow cytometry. In contrast to the control and DOX groups, a higher number of CD86⁺ cells were detected in the mTFH group, suggesting that the capsule alone can induce macrophage polarization toward the M1 phenotype. Notably, the positive expression rate of CD86 in cells treated with DmTFH was approximately 24.1%, confirming that DmTFH significantly promoted a transition from the M2 to the M1 activation state in macrophages (Fig. 4J and K). Subsequently, CM from macrophages treated with different groups were collected to study their impact on BMDC maturation. It was observed that comparable levels of mature DCs were present in both the control and DOX groups, indicating that free DOX’s weak activation of M1 macrophages failed to significantly promote BMDC maturation (Fig. 4L). Conversely, CM from mTFH and DmTFH treatments effectively induced DC maturation, with proportions of 10.2% and 12.8%, respectively (Fig. 4L and M). This suggested that the paracrine signals from DmTFH-activated M1 macrophages can further promote DC maturation. Overall, these results revealed that DmTFH can mediate BMDC maturation through multiple pathways (Fig. 4E). First, DmTFH relied on DOX to cause ICD in cancer cells and synergized with ROS to enhance ICD effects. DAMPs triggered by ICD can increase the levels of mature DCs. Second, DmTFH directly promoted DC maturation due to the immunomodulatory effects of the carrier itself. Additionally, DmTFH can indirectly stimulate DC activation by regulating immunosuppressive immune cells (M2 macrophages). The combined action of these mechanisms made DmTFH exhibit promising immunostimulating effects, thus having potential in vivo antitumor applications.

Subsequently, the in vivo antitumor effects of DmTFH were explored. For this purpose, we established a bilateral tumor model by implanting 4T1 tumor cells on both sides of the dorsal region (Fig. 5A). Nine days later, all tumor-bearing mice were randomly divided into 4 groups, and different formulations were intratumorally injected into the primary tumor site. The body weight of the mice and the sizes of the bilateral tumors for each treatment group were monitored. The growth curves of bilateral tumors in each mouse were depicted in Fig. 5B and D, while the final weights and images of excised tumors post-treatment were shown in Fig. 5C and E and Fig. S10. Compared to other groups, DmTFH treatment resulted in a significant reduction in the development of the primary tumor, with average tumor sizes of 114.61 mm³ (Fig. 5B). The superior therapeutic effect of the DmTFH group in treating the primary tumor reflected the enhanced local chemotherapeutic effect of DOX. DmTFH enabled the release of DOX at the primary tumor site, inducing significant cytotoxicity and thereby more effectively suppressing the growth of the primary tumor. However, although DOX treatment inhibited the rapid development of primary tumors in mice due to its direct cytotoxic effects on tumor cells, the suppression of distant tumors was less effective, with an average tumor volume as high as 930.82 mm³, even exceeding that of the PBS group. The growth of distant tumors reflected the level of systemic immune activation. The volume of distant tumors in the DOX group was larger than that in the PBS group, suggesting that DOX treatment might induce systemic immunosuppression, which was confirmed by subsequent systemic immune analyses. The progression of both proximal and distal tumors in mice treated with mTFH was also suppressed, with average tumor volumes of 397.00 and 289.64 mm³, respectively. Subsequently, the bilateral tumors from each experimental group were recorded by weighing and photographing. The DmTFH group exhibited the smallest primary tumors, with an average weight of 0.196 g, revealing its potent antitumor capability (Fig. 5C and Fig. S10). Unexpectedly, the distant tumors from the mTFH group were the lightest, with an average weight of 0.41 g, indicating the significant impact of mTFH on strengthening immune-mediated antitumor effects (Fig. 5E and Fig. S10). The slightly larger volume of distant tumors in the DmTFH group compared to the mTFH group may be attributed to the mild immunosuppressive side effects caused by released DOX, which could weaken systemic immune activation. This observation was consistent with the finding that the distant tumor volume in the DOX group was greater than that in the PBS group, indicating that DOX may induce systemic immunosuppression in vivo. Furthermore, the histological and apoptotic changes in tumor tissues following different treatments were visualized by H&E (hematoxylin and eosin) and TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling) staining. As displayed in Fig. 5F and G, a significantly reduced tumor cell density was observed in bilateral tumors of both mTFH and DmTFH groups. Although some disruption of primary tumors was noted in the DOX group, the distant tumor tissue showed significant nuclear infiltration of tumor cells, with almost no apoptotic cells detected. The TUNEL staining combined with the quantification analysis further confirmed that DmTFH induced the highest level of apoptosis in primary tumor tissues, indicating its strong antitumor effect (Fig. 5G and Fig. S11). A substantial number of green fluorescence spots also appeared in the sections of distant tumors in the mTFH group, consistent with the results of the tumor growth curve. The stronger treatment effect for distant tumors than primary tumors in the mTFH group suggested that the distant tumor was more sensitive to the systemic immune response activated by mTFH. This could be attributed to a relatively weaker immunosuppressive microenvironment in the distant tumor, making it more susceptible to infiltration by immune effector cells. In the tumor model, primary and distant tumors were established by injecting different numbers of 4T1 cells, resulting in a larger tumor burden in the primary site and a smaller one in the distant site. Extensive preclinical and clinical evidence has demonstrated that, at both local and systemic levels, larger tumors exhibit stronger immunosuppressive properties than smaller ones, which can impair the immune responses elicited by immunotherapy [30]. In summary, these results highlighted the effectiveness of DmTFH in delaying tumor growth, underscoring its potential as a promising therapeutic strategy for cancer. Additionally, leveraging the immunomodulating properties of the capsule itself, DmTFH may improve the poor effects of free DOX on activating the immune system.

Given the promising inhibitory effects of DmTFH on bilateral tumors observed in the tumor treatment experiments, we proceeded to assess its impact on the immune system. Initially, we performed immunofluorescence staining for CRT protein in tumor tissues to assess the efficacy of ICD induced by different treatments. The results revealed a significant red fluorescence signal in the primary tumors of the DmTFH group, confirming DmTFH’s ability to trigger ICD in tumor cells in vivo (Fig. 6A). However, only partial red fluorescence was detected in the tumors of the DOX group, suggesting that the ICD effects mediated by DOX were limited in vivo. This observation aligned with previous findings reported in the literature [31,32]. Surprisingly, mTFH also exhibited some ability to induce ICD, as evidenced by the moderate red fluorescence observed in the tumor tissues. These findings largely aligned with the previously mentioned tumor suppression results. Next, we evaluated the infiltration of key immune cells within the tumor. The tumor microenvironment often contains M2-phenotype macrophages, which are immunosuppressive and can promote tumor growth while inhibiting the immune system’s ability to target and destroy tumor tissue. Extensive research has reported that triggering ICD in tumor cells and releasing DAMPs can prompt macrophage polarization within the tumor from an M2 to an M1 state [33,34]. Furthermore, our study found that glycation capsules can enhance interactions with macrophages, promoting the accumulation of intracellular inflammatory mediators to modulate the repolarization of macrophages from M2 to M1 [9]. These M1-type macrophages helped improve the tumor’s immunosuppressive microenvironment and activated the immune system to specifically target and destroy tumor cells. Therefore, we further investigated the M1 polarization status of macrophages within tumor tissues after different treatments. The percentage of CD86⁺ macrophages in the tumors of mice treated with mTFH and DmTFH increased significantly to 19.1% and 17.8%, respectively, which was more than a 4-fold increase over the control group (Fig. 6B and D). Moreover, the percentage of CD206⁺ cells in the DmTFH group markedly decreased (Fig. 6E). These findings indicated that DmTFH can effectively stimulate the polarization of tumor-associated macrophages from an M2 to an M1 phenotype, mainly benefiting from the role of mTFH. Furthermore, we assessed T cell infiltration in tumor tissues. CD4⁺ helper T cells and CD8⁺ cytotoxic T cells represent 2 crucial subtypes of T lymphocytes. Cytotoxic T cells are capable of identifying and eliminating tumor cells directly, while helper T cells assist in this process [35]. Immunofluorescence analysis of tumor tissue sections revealed that, relative to the control group, mice treated with DmTFH showed a marked enhancement in the intensity of green (CD4⁺) and red (CD8⁺) fluorescence signals. This indicated that DmTFH-induced immune activation can lead to an enhancement of both CD4⁺ and CD8⁺ T cell infiltration in tumors, ultimately contributing to tumor cell killing (Fig. 6C). In contrast, the DOX group showed only a slight increase in green fluorescence, suggesting that DOX had limited capacity to induce immune stimulation in vivo. The mTFH group also displayed a substantial increase in both green and red fluorescence signals, confirming that the capsule carrier itself possessed intrinsic immunostimulatory properties for T cell activation. Flow cytometry quantification indicated that, compared to treatments with PBS, DOX, and mTFH, DmTFH administration markedly enhanced the presence of both CD4⁺ and CD8⁺ T cells within the tumor microenvironment. Specifically, the proportions reached 35.30% for CD4⁺ T cells and 50.3% for CD8⁺ T cells (Fig. 6F). These findings demonstrated that DmTFH was highly effective in enhancing the immune response within tumors. We further observed differences in tumor metastasis in lung tissue sections from mice treated with different therapies. The DmTFH-treated group displayed the least amount of tumor infiltration and metastasis among all treatment groups (Fig. 6G). The high-magnification images provided a clearer view of the tumor cells within the lung tissues. These results suggested that the effective immune response triggered by DmTFH prominently delayed the progression of lung metastasis.

Motivated by the robust immune activation at the primary tumor site, we explored how DmTFH triggers systemic immune reactions. As key immune organs, the spleen and lymph nodes were specifically examined for signs of immune activation. Flow cytometry was employed to assess the T cell phenotype in the spleen and the maturation status of DCs in the lymph nodes. The strategy for gating to identify T cell populations and mature DCs was demonstrated in Figs. S12 and S13. As presented in Fig. 7A, a noticeable rise in CD8⁺ T cell frequency was observed in the spleens of mTFH group-treated mice, accounting for 34.1%. This observation aligned with our previous report showing that the carrier itself has the capacity to activate T cells through its intrinsic active component, mBSA [9]. Due to the absence of potential immunosuppressive side effects associated with DOX, the systemic immune response triggered by mTFH appeared to be stronger than DmTFH, showing better performance in T cell activation. This could be the crucial reason for significant distant tumor suppression and delayed lung metastasis observed in the mTFH group in Figs. 5E and 6G. Consequently, owing to the intrinsic immunomodulatory effect of the carrier, the DmTFH group showed markedly higher proportions of CD4⁺ and CD8⁺ T cells (69.0% and 28.1% in the spleen) compared to the free DOX group (Fig. 7B and C). These results demonstrated that intratumoral injection of DmTFH elicited a systemic immune response, which was pivotal in suppressing distant tumor growth and enhancing overall tumor treatment efficacy. Further observations revealed differences in the sizes of the inguinal lymph nodes following different treatments. Surprisingly, we observed that DOX-treated mice had smaller lymph node volumes compared to the PBS group, a finding consistent with a previous study reporting severe tissue damage and splenic contraction in DOX-treated mice [31]. This phenomenon was potentially linked to DOX-induced immunosuppression due to its undifferentiated toxicity toward immune cells. In comparison to the DOX group, both mTFH and DmTFH led to a significant enlargement of the inguinal lymph nodes, demonstrating a strong immune response (Fig. 7D). Analysis via flow cytometry of mature DCs in the lymph nodes revealed that mTFH and DmTFH treatments resulted in a notable increase in the percentage of CD86⁺MHC II⁺ DCs (11.3% and 18.1%) compared to the DOX-treated group (0%) (Fig. 7E and F). Additionally, the cytokine levels in mouse serum after various treatments were assessed using enzyme-linked immunosorbent assay (ELISA) tests to further characterize the systemic immunostimulating effects of DmTFH. The findings showed a marked rise in IL-6, tumor necrosis factor-α (TNF-α), and interferon-γ (IFN-γ) levels in the serum of DmTFH-treated mice relative to the control group (Fig. 7G to I). Especially, IFN-γ levels in the serum of the DOX group were lower than the PBS group, confirming that DOX treatment may cause systemic immunosuppression in vivo. Collectively, DmTFH demonstrated superior tumor suppression at the primary tumor site due to the chemotherapeutic effect of DOX, whereas mTFH exhibited better control of distant tumors and improved performance in certain immune indicators owing to its intrinsic immunostimulatory capability. These findings reflected the synergistic mechanisms between chemotherapy and immunotherapy and highlighted the local and systemic immune-activating effects of DmTFH compared to free DOX, indicating that DmTFH treatment can effectively reprogram the suppressive tumor microenvironment, thus boosting antitumor immunity.

Finally, we comprehensively evaluated the biological safety of DmTFH in vivo. Throughout the treatment period, no notable weight reduction was detected in the mice from the mTFH- and DmTFH-treated groups, indicating their high biosafety. In contrast, the mice in the DOX group experienced a reduction in weight over the course of the administration, suggesting that DOX may possess some toxicity that affects the body weight of the mice (Fig. 8A). Subsequently, we observed that most of the test indexes in the blood biochemical analysis of DmTFH-treated mice showed no significant difference when compared to the control group, indicating its good blood compatibility (Fig. 8B to I). It is worth noting that treatment with mTFH and DmTFH led to slight elevations in total bilirubin (TBIL) and creatinine (CREA), implying that mTFH and DmTFH treatments may cause mild alterations in liver and kidney function due to normal metabolism in the body. Therefore, we further performed H&E staining on histological sections of primary organs. The staining results showed no significant inflammatory cell infiltration or hepatocyte necrosis in the H&E-stained liver sections of the mTFH and DmTFH groups (Fig. 8J). However, severe inflammatory cell infiltration and disordered liver tissue structure were observed in the PBS and DOX groups, indicating lesions caused by tumor development. In addition, we found varying degrees of local cell density changes indicative of immune activation in the spleen tissues of the DOX, mTFH, and DmTFH groups (as indicated by white arrows), which were consistent with the results obtained from flow cytometry. Moreover, no evident abnormal changes were observed in the spleen, kidney, and heart tissues after treatment with mTFH and DmTFH, indicating that DmTFH was a safe carrier that could mitigate DOX-induced toxicity to some extent (Fig. 8J and Fig. S14). Meanwhile, compared to free DOX, DmTFH demonstrated an enhanced T cell-mediated immune response, attributed to its intrinsic ability to activate M1 macrophages and DCs (Figs. 7A and 8J). However, the mild systemic immunosuppression caused by DOX persisted (Figs. 5E and 7A). Future research could focus on developing more precise targeted delivery systems to reduce this systemic immunosuppression, such as modifying microcapsules with targeting ligands like folic acid or RGD (arginine–glycine–aspartic acid sequence) peptides to enhance their selectivity for cancer cells.

The main materials and instruments used in the study are detailed in the Supplementary Materials.

The mBSA solution was prepared using our previously reported wet-state heating method [9]. To load DOX in situ, varying concentrations of DOX (0.5 to 4 mg/ml) were introduced into an aqueous mBSA solution (100 mg/ml, 1 ml), and the combined solution was sonicated until the DOX was completely dissolved. Following this, the mBSA or DOX-loaded mBSA solution was subjected to heat at 75 °C for 5 min. At this point, the solution exhibited slight gelation. The solutions were then ultrasonicated for 2 min using a cell disruptor (Nanjing ATPIO Instruments Manufacture Co. Ltd.) with amplitude power set at 80%, an ultrasonic time of 3 s, and intervals of 3 s. Following centrifugation at 300g for 10 min, the top white foam layer was extracted, collected, and resuspended in 1 ml of MilliQ water. The centrifugation and washing processes were conducted twice (300g, 10 min) to eliminate any unbound DOX, and the MB dispersion was gathered for subsequent capsule preparation.

The preparation of the capsules followed a modified procedure, as detailed in a previous study [36]. Briefly, 25 μl of TA (2 mg/ml), 50 μl of HADA (17.7 mg/ml), and 25 μl of FeCl₃∙6H₂O (5.7 mg/ml) solutions were sequentially added to the prepared mBSA or DOX-loaded mBSA MB solution (390 μl). To adjust the pH, 10 μl of tris (hydroxymethyl) aminomethane solution (200 mg/ml) was added to the resulting solution. After each addition of TA, HADA, FeCl₃∙6H₂O, and tris solutions, the suspensions were vigorously mixed for 60 s using a vortex mixer. The mixture underwent 3 washes with MilliQ water to eliminate surplus TA, HADA, FeCl₃, and MBs. Washing steps involved centrifugation at 6,000g for 10 min, after which the pellet was collected. Capsules templated with mBSA MBs were named mTFH, whereas those templated with DOX-loaded mBSA MBs were named DmTFH. HADA was prepared using the previously described method, which involved coupling the carboxyl groups of hyaluronic acid with the amine groups of dopamine [9]. The DOX concentration in the final product was determined by measuring the absorbance at 484 nm and referencing it against a standard curve. Protein concentrations were assessed using a bicinchoninic acid protein assay kit, following the manufacturer’s instructions.

The DOX-loaded microcapsule solution was suspended in PBS buffer (pH 7.4 or 5.0) and shaken at 1,000 rpm in a thermostatic shaker at 37 °C. At specific time intervals, the capsules in the buffer were centrifuged, and the supernatant (1 ml) was collected for absorbance measurement using a UV–Vis spectrophotometer. The DOX concentration in the supernatant was determined using the established standard curve. Meanwhile, 1 ml of fresh buffer was introduced into the capsule suspension to maintain a total volume of 2 ml for subsequent release studies.