Implantable artificial lymph node enables rapid in vivo neoantigen-specific T-cell generation and expansion for improving antitumor immunity Authors: Zhixiong Cai, Kongying Lin, Xiuqing Dong, Zhenli Li, Geng Chen, Luobin Guo, Pan Mou, Peizhe Chen, Junjing Huang, Ling Li, Xiaolong Liu, Ruijing Tang, Yongyi Zeng Published: 2025 ABSTRACT AbstractBackgroundTumor neoantigen vaccines typically require multiple prime-boost immunizations over several weeks or even months to elicit sufficient neoantigen-specific T-cell responses, which greatly limits their effectiveness against rapidly progressing cancers, such as hepatocellular carcinoma.MethodsHere, we developed a functionalized alginate scaffold for T-cell activation as an artificial lymph node (FAST-LN) to rapidly and efficiently generate tumor neoantigen-specific T cells in vivo. The FAST-LN consists of an alginate scaffold crosslinked with calcium ions and hyaluronic acids conjugated with anti-TIM-3 (T cell immunoglobulin and mucin domain-containing protein-3) antibodies, then incorporates tumor neoantigen-encoding adenoviral vaccines, dendritic cells (DCs), and various cytokines (C-C motif chemokine ligand 21 (CCL21), interleukin (IL)-2, IL-4, and Granulocyte-Macrophage Colony Stimulating Factor (GM-CSF)) that maintain DC maturation and recruit T cells for antigen presentation.ResultsThis design facilitates efficient DC-T cell interactions within FAST-LN, enabling rapid generation of neoantigen-specific T cells for antitumor responses. In vivo antitumor studies demonstrated superior therapeutic efficacy of FAST-LN over traditional vaccines in mouse tumor models. Mechanistically, FAST-LN significantly enhanced tumor infiltration of CD8+ T cells, Th1 cells and macrophages, orchestrating robust antitumor responses.ConclusionsOur results demonstrate that implantable FAST-LN rapidly induces tumor neoantigen-specific T cells in vivo, offering a promising strategy for neoantigen-based cancer immunotherapy.Keywords: Immune Checkpoint Inhibitor, Immunotherapy, Solid tumor, T cell, Vaccine WHAT IS ALREADY KNOWN ON THIS TOPIC WHAT THIS STUDY ADDS Functionalized alginate scaffold for T-cell activation as an artificial lymph node (FAST-LN) significantly accelerates the induction of neoantigen-specific T cells compared with traditional DC vaccines, achieving detectable responses within 2 days versus 6 days. The localized delivery of cytokines (IL-2, IL-4, GM-CSF) and CCL21 in FAST-LN promotes DC maturation, T-cell recruitment, and sustained immune activation, leading to robust antitumor responses. In preclinical models of hepatocellular carcinoma and metastatic lung cancer, FAST-LN demonstrated superior tumor suppression, long-term immune memory, and 100% survival rates. Functionalized alginate scaffold for T-cell activation as an artificial lymph node (FAST-LN) significantly accelerates the induction of neoantigen-specific T cells compared with traditional DC vaccines, achieving detectable responses within 2 days versus 6 days. The localized delivery of cytokines (IL-2, IL-4, GM-CSF) and CCL21 in FAST-LN promotes DC maturation, T-cell recruitment, and sustained immune activation, leading to robust antitumor responses. In preclinical models of hepatocellular carcinoma and metastatic lung cancer, FAST-LN demonstrated superior tumor suppression, long-term immune memory, and 100% survival rates. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY Our study introduces FAST-LN, an implantable artificial lymph node scaffold that rapidly generates tumor neoantigen-specific T cells in vivo, overcoming the limitations of conventional neoantigen vaccines that require prolonged immunization schedules. This discovery introduces a promising strategy for improving the management of immunotherapy in liver cancer. Our study introduces FAST-LN, an implantable artificial lymph node scaffold that rapidly generates tumor neoantigen-specific T cells in vivo, overcoming the limitations of conventional neoantigen vaccines that require prolonged immunization schedules. This discovery introduces a promising strategy for improving the management of immunotherapy in liver cancer. Introduction Therapeutic tumor vaccines have shown promise in cancer immunotherapy by activating adaptive immunity against tumor neoantigens-unique peptides arising from genetic alterations or post-translational modifications. These vaccines, delivered through peptide, DNA, or messenger RNA (mRNA) platforms, are processed by dendritic cells (DCs) to prime neoantigen-specific T cells.1,5 However, the efficacy of endogenous DCs in vaccine uptake and presentation is often limited, requiring repeated administrations to achieve therapeutic responses; furthermore, the slow generation of antitumor responses poses significant challenges, particularly for aggressive cancers like hepatocellular carcinoma (HCC) where rapid disease progression may outpace immune activation.6 7 While the administration of a classic DCs stimulated with neoantigens ex vivo could accelerate the induction of neoantigen-specific T-cell responses compared with traditional vaccination methods, this approach also faces several critical limitations: suboptimal DC activation compromises T-cell induction; poor DC survival and migration through conventional delivery routes (intravenous, subcutaneous, or intradermal) results in over 90% failing to reach lymph nodes or tumor sites; and DC immune checkpoint overexpression (eg, TIM-3) impairs antigen presentation.8,11 These challenges highlight the need for innovative strategies that create supportive niches for DC-T cell interactions, enabling rapid, localized T-cell priming while circumventing DC migration requirements. Biomaterials, particularly functionalized alginate scaffolds, offer unique advantages including excellent biocompatibility, biodegradability, and three-dimensional (3D) porous structures conducive to cell adhesion and culture.12,14 When modified with immune-stimulatory molecules, these scaffolds can create a conducive microenvironment for immune cell recruitment, activation, and expansion.15,17 Given the natural 3D structure and excellent biocompatibility of alginate materials, constructing a lymph node-like scaffold for DC vaccine delivery represents a promising solution for enhancing DC-T cell co-localization and antigen presentation. Among neoantigen-loading methods for DC vaccines, adenoviral vectors offer superior advantages over conventional approaches like peptide pulsing or mRNA electroporation, combining high transfection efficiency with specific targeting. Unlike retroviral vectors, adenoviruses maintain DC functionality without genomic integration while enhancing DC maturation through increased major histocompatibility complex (MHC) and co-stimulatory molecule expression. Their large packaging capacity (up to 37 kb) enables simultaneous delivery of neoantigens and immunomodulatory molecules.18 For example, adenovirus-mediated CD40L (a ligand for CD40 that activates DCs to enhance their antigen-presenting capacity) expression enhances DC immunomodulatory function via increased interleukin (IL)-12 secretion and co-stimulatory molecule expression, improving cytotoxic responses against various tumors.19 Importantly, alginate scaffolds enhance viral transfection in 3D culture, suggesting an efficient approach for constructing lymph node-like structures through adenovirus-mediated delivery of neoantigens and immunomodulators. The migration and homing of mature DCs to lymph nodes are orchestrated by specific chemokines, particularly CCL21 (secondary lymphoid tissue chemokine), which is predominantly expressed by high endothelial venules and within T cell-rich zones of lymphoid tissues.20,22 CCL21 interaction with its receptor CCR7 plays an indispensable role in lymphoid tissue organization by mediating the coordinated co-localization of mature DCs, naive T cells, and memory T cells within lymph nodes, thereby facilitating efficient immune response initiation. Moreover, recent studies have highlighted the significance of TIM-3, a novel immune checkpoint molecule, as a critical exhaustion marker on both DCs and T cells.10 23 Notably, TIM-3 blockade has been shown to enhance DC function through increased expression of CXCL9 and IL-12. The upregulation of CXCL9 not only promotes spatial proximity between cDC1 and CD8+ T cells but also enhances T-cell antitumor activity. These insights suggest that engineering a 3D scaffold system incorporating DCs, CCL21, and anti-TIM-3 antibody could recreate a lymph node-like microenvironment, potentially providing an optimal setting for neoantigen-specific DC-T cell interactions and subsequent immune activation. Here, we present the all-in-one functionalized alginate scaffold for T-cell activation as an artificial lymph node (FAST-LN), an innovative biomaterial-based platform designed to mimic lymph node in vivo for the neoantigen-specific T-cell generation and expansion. FAST-LN integrates unmatured DCs and adenovirus-encoded neoantigens into a scaffold enriched with immune-stimulatory components, including anti-TIM-3 antibody, CCL21, IL-2, IL-4, and GM-CSF to promote DC maturation, enhance antigen presentation, and facilitate the recruitment and expansion of naïve T cells. By providing a localized, supportive environment for DC-T cell interactions, FAST-LN aims to overcome the limitations of conventional vaccine delivery and immunosuppressive tumor microenvironment, offering a promising avenue for controlling tumor growth and improving therapeutic outcomes in cancer immunotherapy. Materials and methods Cell lines Hepa1-6 and HEK293A cells were obtained from ATCC. DC2.4 cells were purchased from Sigma-Aldrich. The mouse Lewis lung cancer (LLC) cell line was obtained from BNCC (BeNa Culture Collection, China). All these cells were cultured in DMEM (Dulbecco's Modified Eagle Medium) containing 10% FBS (Fetal Bovine Serum) and 1% penicillin/streptomycin at 37°C in a humidified environment with 5% CO2. Luciferase-expressing Hepa1-6/LLC (Hepa1-6-Luc/LLC-Luc) cells were established through lentivirus transfection expressing luciferase reporter gene, and then the positive cells were selected by puromycin and maintained in low concentration of puromycin (Shanghai Genechem). In vivo tumor models and evaluation of antitumor efficacy The tumor challenges were performed by injecting tumor cells (Hepa1-6-Luc and LLC-Luc cells) into 6–8 weeks old female C57BL/6 mice purchased from Shanghai SLAC Laboratory Animal (License number: SCXK(Hu) 2022–0004, Shanghai, China). HLA-A*0201/Kb transgenic mice were purchased from National Human Disease Animal Model Resource Center (Beijing, China), which are a well-established model for the study of HLA-A*0201 restricted epitopes. The orthotopic HCC mouse model and lung cancer liver metastasis mouse model were established as follows: (1) the mice were intraperitoneally anesthetized with 50 mg/kg of pentobarbital; (2) the skin was prepared aseptically, and the midline laparotomy was performed after the abdomen shaving; (3) liver subcapsular inoculation with tumor cells (Hepa1-6-Luc cells for the HCC model, LLC-Luc cells for the lung cancer liver metastasis model) mixed with the matrigel plugs; (4) the abdominal muscle and skin were fully-layer sutured. The mice were randomly divided into four groups (n=5) and treated with phosphate-buffered saline (PBS) (as Ctrl), scaffold (only alginate), DC vaccine and FAST-LN on days 0 and 14, respectively. The experimental procedures for each group were as follows: (1) Ctrl: 200 µL of PBS solution was subcutaneously injected near the axillary lymph nodes of the mice. (2) Scaffold: scaffolds were subcutaneously implanted into the back of the mice. (3) DC vaccine: 2×106 bone marrow-derived dendritic cells (BMDCs) were infected with adenovirus for 48 hours to prepare the DC vaccine, and then subcutaneously injected near the axillary lymph nodes of the mice. (4) FAST-LN: 2×106 BMDCs were mixed with adenovirus and added to FAST-LN, and then placed in a 37°C, 5% CO2 incubator for 1 hour to allow the BMDCs and adenovirus mixture to permeate FAST-LN. Subsequently, FAST-LN was subcutaneously implanted into the back (or armpit as indicated) of the mice. Tumor burdens of the models were monitored by using the IVIS Spectrum Animal Imaging System (PerkinElmer, USA). At the end of treatment, all mice were sacrificed by cervical dislocation, and the number of tumor nodules was recorded. For the subcutaneous HCC model constructed by HLA-A*0201/Kb transgenic mice, the Hepa1-6 cells were infected by adenovirus to express HBV Core 18–27 (HBc18-27) epitope and HLA-A*0201 before subcutaneous injection. Tumor monitoring of subcutaneous HCC model was performed by a single operator every 2 days in two dimensions using calipers. Tumor volume was calculated from the measured data as follows: V=AB2/2 (A is the long diameter and B is the short diameter). When the tumor volume grew to 50–80 mm3, it was recorded as day 0, and then the mice were randomly divided into four groups to start treatment by PBS, scaffold, DC vaccine, or FAST-LN, respectively. Data availability All data generated or analyzed during this study are included in this article. Inquiries about further material documents can be directed to the corresponding authors. The single-cell transcriptome sequencing data from this publication have been deposited to the Genome Sequencing Archive (GSA) database (GSA number: CRA024359). Ethics approval All animal procedures were conducted according to the “National animal management regulations of China” and approved by the Animal Ethics Committee of Mengchao Hepatobiliary Hospital of Fujian Medical University (MCHH-AEC-2023–16). Statistical analysis Statistical analysis of data was analyzed by one-way or two-way analysis of variance. *p<0.05 was set as statistically significant, also including **p<0.01, ***p<0.001, ****p<0.0001. Results are presented as mean±SEM at least three experiments by the software of GraphPad Prism (San Diego, USA). Results Alginate-based functional scaffold design as the artificial lymph node To overcome the inefficient neoantigen-specific T-cell induction caused by limited DC vaccine homing and migration, we engineered a 3D, porous scaffold that enables efficient preparation of neoantigen-loaded DCs and simultaneously promotion of DC-T cell interactions for rapid antigen-specific T-cell generation. As illustrated in figure 1A, alginate was chosen as the foundational material for the cell culture scaffold due to its exceptional biocompatibility, biodegradability, and ability to form a 3D porous structure that supports cell adhesion and growth. To transform this alginate scaffold into a FAST-LN capable of rapidly generating neoantigen-specific T cells, we incorporated key immune-modulatory components, including the anti-TIM-3 antibody (αTIM-3), cytokines (IL-2, IL-4, and GM-CSF), and the chemokine CCL21. These elements were strategically embedded within the scaffold to create a microenvironment that mimics the natural lymph node, facilitating efficient immune cell interaction and activation. Design and characterization of the FAST-LN scaffold. (A) Schematic illustration of the FAST-LN scaffold, composed of an alginate-based matrix crosslinked with calcium ions and hyaluronic acid conjugated with anti-TIM-3 antibodies. The scaffold was also embedded with cytokines (IL-2, IL-4 and GM-CSF) and chemokine (CCL21) to promote DC-T cell interaction. (B) Photograph of the FAST-LN scaffold after lyophilization in 48-well plate. (C) TEM image of the scaffold exhibiting the porous structure with a diameter of~100 µm. (D) Confocal microscopy image showing DC2.4 cells adhering and clustering within the scaffold. DC2.4 cells were stained with Dil for 2 hours under 37°C before loading to the scaffold. (E–F) Apoptosis levels of DC2.4 cells cultured in FAST-LN under different time points (48, 72 and 96 hours). Ctrl represents normal liquid culture medium. (G–H) Apoptosis levels of BMDCs cultured in FAST-LN under different time points (48, 72 and 96 hours). Ctrl represents normal liquid culture medium. (I–J) Flow cytometry analysis of CD11c+CD80+CD86+ mature DCs in FAST-LN and in liquid medium (Ctrl). Data are presented as the mean±SEM. BMDCs, bone marrow-derived dendritic cells; DCs, dendritic cells; FAST-LN, functionalized alginate scaffold for T-cell activation as an artificial lymph node; IL, interleukin; TEM, transmission electron microscopy. TIM-3, a crucial immune checkpoint molecule, not only serves as a marker of T-cell exhaustion, but also plays a regulatory role in DC function by inhibiting inflammasome activation and suppressing the stimulator of interferon genes pathway in DCs.10 23 By incorporating αTIM-3 into the alginate matrix, we aimed to block the TIM-3 pathway, thereby enhancing the antigen presentation capabilities of DCs and maintaining T cells in a less exhausted, more functional state. This dual action significantly improves the tumor-killing potential of the activated T cells. To ensure the sustained and localized activity of αTIM-3 within the scaffold, we conjugated it to hyaluronic acid (HA), a naturally occurring polysaccharide known for its high biocompatibility and safety as a delivery vehicle. The conjugation process used NHS/EDC chemistry, which allowed for the covalent attachment of αTIM-3 to HA. To verify the conjugation efficiency, αTIM-3 was first labeled with indocyanine green (ICG), and a standard curve for ICG-αTIM-3 was established (online supplemental figure S1A). Subsequent conjugation of ICG-αTIM-3 with HA was confirmed by measuring the fluorescence intensity of the conjugate (online supplemental figure S1B). The calculated conjugation efficiency of 33.3% confirmed the successful integration of αTIM-3 into the HA matrix, ensuring its stable presence within the FAST-LN scaffold. Additionally, to simulate a lymph node-like environment, IL-4, GM-CSF, CCL21, and IL-2 were added into the liquid mixture of alginate and HA-αTIM-3 for the preparation of FAST-LN, aiming to promote DC maturation, T-cell recruitment and proliferation. Following calcium ion cross-linking of alginate and lyophilization, the liquid mixture was transformed into a solid scaffold in 48-well plate with the diameter of ~6 mm and the height of ~2 mm (figure 1B). Transmission electron microscopy (TEM) analysis further revealed that the scaffold of FAST-LN exhibited a well-defined porous structure, with diameter size of ~100 µm and surface folding inside (figure 1C). This architecture facilitated cellular attachment and provided microenvironments conducive to cell-material interactions. When DC2.4 cells were introduced, successful adhesion and clustering were observed, suggesting a supportive microenvironment for cell growth (figure 1D). Further examination with 3D scanning by confocal microscopy confirmed that the cells could penetrate deeply and uniformly throughout FAST-LN (figure 1D), indicating its capacity to support cell migration and interaction. To evaluate the biocompatibility of the scaffold, DC2.4 or BMDCs cultured within FAST-LN were extracted by enzymatic digestion at the time point of 48, 72 and 96 hours, and the apoptosis levels were assessed (figure 1E–H). The results showed that the number of viable cells was comparable to the controls cultured in liquid medium, with all viability rates above 80%, indicating that FAST-LN did not exhibit cytotoxicity during in vitro cell culture. Moreover, after 96 hours of culture, BMDCs were collected from FAST-LN to assess their maturation status according to the percentage change of CD11c+CD80+CD86+ cells (figure 1I,J). Flow cytometry analysis revealed that DCs cultured within FAST-LN exhibited sustained maturation, with 47.5±1.193% of cells expressing high levels of maturation markers (CD80/CD86). This level of phenotypic maturation, achieved in the presence of the embedded IL-4 and GM-CSF, was comparable to the 45.3±1.258% observed in DCs cultured in liquid medium. These results demonstrate that FAST-LN supports not only DC viability but also their phenotypic maturation in culture. Preparation of neoantigen-loaded DCs to activate T cell in FAST-LN Although FAST-LN exhibits good cell permeability and can maintain BMDC viability during culture through embedded IL-4 and GM-CSF, it is unclear whether the 3D structure of FAST-LN can be effectively used for the preparation of neoantigen-loaded DCs. Based on the ability of alginate scaffolds to enhance viral infection,12 we selected adenovirus to infect DCs in FAST-LN for the preparation of neoantigen-loaded DCs. Compared with liposomes and electroporation, adenovirus not only exhibits higher infection efficiency but also acts as an adjuvant by activating DC maturation via TLR (Toll like receptor) signaling, thereby enhancing the antigen-presenting function of DCs.24 Therefore, CD40L, mCherry, and seven highly immunogenic neoantigens derived from the Hepa1-6 HCC cell line previously identified in our study for developing tumor vaccines25 26 were first packaged into adenovirus (termed AdvNeo), which were then used to infect DC2.4 and BMDCs in 24-well plates. After 48 hours, the fluorescence intensity of mCherry in DC2.4 and BMDCs increased significantly with higher multiplicity of infection (MOI), demonstrating that the AdvNeo successfully infected both DC2.4 and BMDCs (online supplemental figure S2A, B). Flow cytometry analysis was further used to validate CD40L expression in DC2.4 and BMDCs. As shown in online supplemental figure S2C, F, the proportion of CD40L-positive cells in DC2.4 increased from 20% to 90%, and in BMDCs, it increased from 5% to 50%, following MOI changing from 50 to 800. Given that CD40L enhances DC maturation and antigen presentation,27 we examined the proportion of CD11c+CD80+CD86+ cells under varying MOI. The results showed that higher CD40L expression correlated with an increased percentage of mature DCs (online supplemental figure S3A, B). When the MOI increased from 400 to 800, although the intensity of CD40L expression on DCs increased, the proportion of mature DCs did not further increase. Moreover, at an MOI of 800, the viability of DCs significantly decreased to 55.92±0.8454%, which was notably lower than 83.69±0.1986% observed at an MOI of 400 (online supplemental figure S3C). Therefore, MOI=400 was chosen for the preparation of the neoantigen-loaded DCs in the subsequent experiments. In general, these findings demonstrate that the AdvNeo effectively infected DCs, induced CD40L expression, and promoted DC maturation to increase antigen-presenting capability. To further analyze the preparation efficiency of neoantigen-loaded DCs in FAST-LN, BMDCs were mixed with AdvNeo (MOI=400) and added to FAST-LN or 24-well plate (figure 2A). After 48 hours, DCs were extracted by enzymatic digestion, and CD40L expression level from DCs was measured. Compared with neoantigen-loaded DCs prepared in a 24-well plate, the preparation efficiency in FAST-LN improved ~20% (FAST-LN vs 24-well plate: 48.03±0.4978% vs 30.4±0.1732%) (figure 2B,C). Meanwhile, the percentage of mCherry+ cells in FAST-LN was significantly higher than that in the 24-well plate (FAST-LN vs 24-well plate: 68.13±2.682% vs 50.53±1.084%, online supplemental figure S4). This finding indicates that FAST-LN can enhance the transduction efficiency of the AdvNeo, leading to a greater number of successfully transduced DCs and consequently, a higher overall CD40L expression. Furthermore, mouse spleen-derived T cells were added to FAST-LN and 24-well plate, and CD69 expression on T cells was assessed after 48 hours (figure 2D,E). The results demonstrate that T-cell activation efficiency was significantly higher in FAST-LN compared with the 24-well plate (FAST-LN vs 24-well plate: 27.50±1.10% vs 13.53±1.40%). Consistent with this enhanced activation, the key effector cytokines, including interferon (IFN)-γ and tumor necrosis factor (TNF)-α, were also markedly elevated in the T cells co-cultured in FAST-LN compared with the T cells in the 24-well plate (p=0.0004, online supplemental figure S5). These data collectively suggest that the 3D reticular structure of FAST-LN better facilitated DC-T cell interaction, leading to enhanced T-cell activation and functional polarization. In addition, this beneficial effect was further enhanced when αTIM-3 was incorporated into the scaffold, as it not only increased the expression of MHC molecules on DCs to strengthen antigen presentation, but also significantly reduced the upregulation of exhaustion markers (including PD-1 and TIM-3) on the T cells, thereby maintaining their functional state (online supplemental figure S6). Preparation and functional evaluation of neoantigen-loaded DCs in FAST-LN. (A) Schematic of the neoantigen-loaded DC preparation process in FAST-LN. (B–C) Flow cytometry analysis of CD40L expression in BMDCs infected by AdvNeo for 48 hours (n=3; one-way ANOVA). Ctrl: cells were cultured in a 24-well plate in the presence of IL-4, GM-CSF and IL-2; in plate: cells were co-cultured with AdvNeo in a 24-well plate with IL-4, GM-CSF and IL-2; FAST-LN: cells were co-cultured with AdvNeo in FAST-LN with IL-4, GM-CSF and IL-2. (D–E) Flow cytometry analysis of CD69 expression on T cells after culture in FAST-LN for 48 hours (n=3; one-way ANOVA). Ctrl: cells were cultured in a 24-well plate in the presence of IL-4, GM-CSF and IL-2; in plate: cells were co-cultured with AdvNeo in a 24-well plate with IL-4, GM-CSF and IL-2; FAST-LN: cells were co-cultured with AdvNeo in FAST-LN with IL-4, GM-CSF and IL-2. (F–G) T-cell chemotaxis assay about the recruitment of T cells from the upper chamber to FAST-LN containing CCL21/IL-2 or not (Ctrl) (n=3; one-way ANOVA). (H) Sustained release profile of CCL21 from FAST-LN over 120 hours. (I–J) Cytotoxicity assay of T cells activated by FAST-LN in vitro. The supernatant LDH (Lactate Dehydrogenase) from the co-culture system was detected to represent the tumor-killing efficiency of CD8+ T cells against Hepa1-6 or LLC cells (n=6; one-way ANOVA). Scale bar=50 µm. Data are presented as the mean±SEM. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; BMDCs, bone marrow-derived dendritic cells; DC, dendritic cell; FAST-LN, functionalized alginate scaffold for T-cell activation as an artificial lymph node; IL, interleukin; LLC, Lewis lung cancer. In addition to providing an effective site for DC-T interaction, FAST-LN should also possess T-cell chemotactic functions, enabling the recruitment and activation of T cells in vivo to exert effective antitumor effects. To achieve a chemotactic effect on T cells, CCL21 was incorporated during the preparation of FAST-LN, along with IL-2, which further promotes the proliferation of activated T cells. When the CCL21/IL-2 containing FAST-LN were placed below the co-culture chamber, T cells located above the chamber were effectively recruited to FAST-LN after 24 hours, ~8-folds higher than that in CCL21/IL-2 deficient FAST-LN (figure 2F,G). Additionally, the release efficiency of CCL21 in FAST-LN showed that 40% of CCL21 was released rapidly within 12 hours, followed by a sustained release over 5 days (figure 2H). This sustained release profile ensured sufficient recruitment of T cells in vivo and efficiently enhanced the utilization of the neoantigen-loaded DCs. To further assess the functional integrity of FAST-LN, we loaded it with neoantigen-pulsed DCs to evaluate its capacity for recruiting spleen-derived CD3+CD8+ T cells and activating their cytotoxic function. When T cells were recruited to FAST-LN and activated by neoantigen-loaded DCs, they exhibited significantly enhanced cytotoxicity after co-culture with Hepa1-6 cells, achieving a killing efficiency of 85.28±2.164%. In contrast, cytotoxicity against the LLC cell line, which lacks the specific neoantigens, was minimal (12.7±0.9932%), confirming the antigen-specific nature of the response. In addition, the AdvNeo-unloaded group exhibited markedly reduced efficiency (33.98±1.267%), substantially impaired in the CCL21/IL-2-deficient group (14.42±1.349%, figure 2I,J). These results demonstrate that in vitro FAST-LN, embedded with IL-4/GM-CSF/CCL21/IL-2, can effectively support neoantigen-loaded DCs, recruit T cells, and activate them to achieve enhanced antigen-specific tumor-killing efficiency in vitro. Subcutaneous implanted FAST-LN for in vivo neoantigen-specific T-cell generation Before investigating the antitumor efficacy of FAST-LN implanted into a mouse tumor model, we first assessed whether the adenovirus could leak from FAST-LN and then infect host cells. Thus, we mixed BMDCs with AdvNeo and incubated the mixture within FAST-LN for 2 hours. FAST-LN was then placed in the upper chamber of the transwell, with 293T cells seeded in the lower chamber (online supplemental figure S7A). When FAST-LN was seeded with AdvNeo alone, mCherry expression was observed in 293T cells, indicating leakage of AdvNeo from FAST-LN in the absence of BMDCs (online supplemental figure S7B). In contrast, when FAST-LN was co-seeded with both BMDCs and AdvNeo, no transduction was observed in 293T cells, suggesting that the adenovirus interacts with and transduces cells in its immediate vicinity. This provided indirect evidence that FAST-LN will not transduce host cells on implantation. In vivo, we further evaluated the release profile of cytokines from FAST-LN after subcutaneous implantation in mice. Quantification by ELISA showed that the initial loading doses of IL-2 (20 ng) and CCL21 (70 ng) translated into sustained systemic concentrations. The higher initial load of CCL21 corresponded to a prolonged release profile, with detectable serum levels persisting for up to 72 hours, compared with 48 hours for IL-2 (online supplemental figure S8). Critically, this release kinetics ensured that systemic cytokine exposure remained within a safe window, demonstrating a favorable safety profile. Subsequently, we implanted FAST-LN containing BMDCs and AdvNeo subcutaneously into the back of mice. After 5 days, major organs of the mice were harvested for H&E staining analysis, which showed no significant inflammation or organ damage (online supplemental figure S9A). Additionally, blood biochemical analysis revealed no deleterious effect (online supplemental figure S9B), suggesting that the biocompatibility and safety of FAST-LN in mice are excellent. On subcutaneous implantation of FAST-LN into the back of a mouse, AdvNeo initially infected BMDCs, leading to the expression of CD40L in BMDCs for self-activation and the presentation of neoantigens. Simultaneously, FAST-LN released CCL21 to recruit endogenous T cells from the mice (figure 3A). To examine whether FAST-LN could efficiently recruit T cells in vivo, FAST-LN were extracted 5 days post-implantation, degraded with alginate lyase, and the cells inside FAST-LN were analyzed using flow cytometry. The results showed that FAST-LN containing CCL21 significantly recruited CD8+ T cells, which was ~4 times higher than that without CCL21 (CCL21+ vs CCL21−: 13.53±0.3480% vs 3.983±0.1770%) (figure 3B,C). Subsequently, once T cells were recruited to FAST-LN, they were further activated by the neoantigen-loaded mature DCs. Flow cytometry analysis revealed a 15% increase of CD69 expression in CD8+ T cells. Meanwhile, it is noteworthy that when αTIM-3 was incorporated into the FAST-LN, the activation efficiency of CD8+ T cells reached 79.2%, clearly demonstrating the ability of FAST-LN to effectively activate recruited T cells in vivo (figure 3D,E). Theoretically, in FAST-LN, neoantigen-loaded DCs can not only activate the recruited T cells but also drive the generation of neoantigen-specific T cells through MHC/neoantigen complexes. These neoantigen-specific T cells then progressively migrate from FAST-LN to other regions of the body, ultimately reaching the tumor site via the bloodstream, where they unleash powerful antitumor responses. Here, neoantigen-specific T-cell frequencies in blood and tumor were detected using a custom-synthesized tetramer antibody based on PTPN2, one of seven specific neoantigens identified from Hepa1-6 cells. Remarkably, our analysis of CD8+ T cells in mouse blood revealed a significant surge in the proportion of neoantigen-specific T cells, rising from 0.79% to an impressive 5.63% under FAST-LN induction. Even more striking, this proportion increased to 11.3% with the addition of αTIM3 (figure 3F,G). Furthermore, we employed an IFN-γ ELISpot (enyzme-linked immunospot) assay to comprehensively evaluate the immunogenicity of all neoantigens. T cells from FAST-LN-treated mice exhibited significantly enhanced responses to each neoantigen peptide compared with the PBS control (online supplemental figure S10), with the strongest reactivity against Ptpn2_I383T, consistent with prior reports.25 These results confirm that FAST-LN effectively elicits immune responses against all loaded neoantigens. Given its dominant immunogenicity, Ptpn2_I383T was selected as a representative marker for subsequent tetramer-based T-cell quantification. In vivo generation of neoantigen-specific T cells by FAST-LN. (A) Schematic illustrating the in vivo process of T-cell recruitment and activation following subcutaneous implantation of FAST-LN. (B–C) Flow cytometry analysis of CD8+ T-cell recruitment to FAST-LN in the presence of CCL21 (n=3; one-way ANOVA). Group 1: FAST-LN without CCL-21 and anti-TIM-3 antibody; Group 2: FAST-LN without anti-TIM-3 antibody; Group 3: the complete FAST-LN system. (D–E) Flow cytometry analysis of CD69 expression on CD8+ T cells in the presence of anti-TIM-3 antibody (n=3; one-way ANOVA). (F–G) Flow cytometry analysis of the neoantigen-specific T cells in mouse blood after FAST-LN implantation (n=3; one-way ANOVA). (H) Schematic of the mice implanted with FAST-LN or injected with conventional DC vaccine on day 0 for the analysis of new antigen-specific T-cell induction. (I–J) Flow cytometry analysis of the neoantigen-specific T-cell induction efficiency between FAST-LN and conventional DC vaccine (n=3; one-way ANOVA). T cells were collected at different time points (0, 2, 4, 6 and 8 days). (K) Schematic of the orthotopic liver cancer mouse model construction for evaluating the migration of neoantigen-specific T cells to tumor sites. (L–M) Flow cytometry analysis of the neoantigen-specific T cells in tumor tissues from FAST-LN and DC vaccine-treated mice. Data are presented as the mean±SEM. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; DC, dendritic cell; FAST-LN, functionalized alginate scaffold for T-cell activation as an artificial lymph node; PBS, phosphate-buffered saline. Conventional DC vaccines rely on the migration of DCs to lymph nodes or tumor sites for the induction of neoantigen-specific T cells. However, less than 10% of the DCs are able to reach the designated sites and exert their function, resulting in delayed T-cell activation and limited antitumor responses. Our above results have demonstrated that FAST-LN not only serves as a site for the generation of neoantigen-loaded DCs but also facilitates DC-T cell interactions, which, theoretically, could enable DCs to activate T cells in FAST-LN as soon as they express tumor neoantigens, resulting in a faster and more efficient generation of neoantigen-specific T cells. To confirm this, we further compared the efficiency of FAST-LN and conventional DC vaccines in inducing neoantigen-specific T cells. BMDCs were first mixed with AdvNeo and then incorporated into FAST-LN, which was implanted subcutaneously into the dorsal region of mice, designated as day 0. Peripheral blood was then collected on days 2, 4, 6, and 8 to assess the levels of neoantigen-specific T cells (figure 3H). In comparison, a conventional DC vaccine was prepared by infecting BMDCs with AdvNeo for 48 hours, followed by direct subcutaneous injection into mice. Flow cytometry analysis revealed that the neoantigen-specific T cells were detectable on day 6 following a single administration of the DC vaccine (1.427±0.1629%) (figure 3I,J). Notably, in the FAST-LN group, the neoantigen-specific T cells were detected as early as day 2 (1.730±0.2884%), with their levels further increasing to 9.773±1.975% by day 6 and reaching 14.50±0.3786% by day 8. These findings clearly demonstrate that FAST-LN not only induces a robust expansion of the neoantigen-specific T cells but also exhibits a much faster rate compared with the conventional DC vaccine. Although FAST-LN implantation on the dorsal side of mice can rapidly and efficiently induce the neoantigen-specific T cells, these T cells only exert targeted tumor cell killing activity once they reach the tumor site. To evaluate whether the neoantigen-specific T cells induced by FAST-LN can migrate to the tumor site, we established an orthotopic liver cancer mouse model in C57BL/6 mice using Hepa1-6 cells (figure 3K). 7 days after tumor cell implantation into the mouse liver, we treated the tumor-bearing mice with either FAST-LN or DC vaccine recorded as day 0. Tumor tissues were then harvested on day 6 and day 8 for the detection of neoantigen-specific T cells. Compared with the DC vaccine group, the proportion of neoantigen-specific T cells in the FAST-LN group exhibited a threefold increase on day 6 (FAST-LN vs DC vaccine: 1.527±0.09062% vs 0.4567±0.09244%). By day 8, the proportion of neoantigen-specific T cells in the FAST-LN group increased to 2.806±0.3755%, significantly higher than the 0.9767±0.06489% observed in the DC vaccine group (p=0.0086). Overall, these findings strongly demonstrate that the rapid generation of neoantigen-specific T cells achieved by FAST-LN highlights its potential as an effective therapeutic strategy, particularly for rapidly progressing malignancies such as HCC. FAST-LN exerts effective antitumor effects in vivo To further verify whether FAST-LN has antitumor functionality, an orthotopic liver cancer mouse model was reconstructed by Hepa1-6 cells to truly simulate the tumor microenvironment. As shown in figure 4A, the constructed orthotopic liver cancer models were randomly divided into four groups (n=5), and further respectively received treatment with PBS (Ctrl), FAST-LN without BMDCs and AdvNeo (scaffold), conventional DC vaccine (DC vaccine), and FAST-LN loaded with BMDCs and AdvNeo (FAST-LN) (figure 4A). Compared with the Ctrl and scaffold group, the mice receiving the DC vaccine showed good tumor growth inhibition, but still, 2/5 mice died within 60 days (figure 4B,C). Impressively, the tumor burden in mice with the FAST-LN treatment was completely inhibited, and the 70-day survival rate (100%) was consistently much higher compared with the other three groups (p=0.0369) (figure 4D). To further evaluate the immune response under different treatments, the orthotopic HCC mouse model was reconstructed, and the treatment by PBS, scaffold, DC vaccine and FAST-LN was re-performed, and the liver samples were harvested and analyzed on day 20. As illustrated in figure 4E, the remarkable therapeutic efficacy of FAST-LN was further corroborated by visual examination of the liver, which showed significant reduction in tumor size and improved tissue integrity. Additionally, histological analysis using H&E staining provided further confirmation of the treatment effectiveness, revealing a marked decrease in tumor cell proliferation and enhanced preservation of normal liver architecture. These findings collectively underscore the potent antitumor activity of FAST-LN and its potential as a promising therapeutic strategy for liver cancer. Antitumor efficacy of FAST-LN in orthotopic HCC models. (A) Schematic of the experimental design for evaluating the antitumor efficacy of FAST-LN in orthotopic HCC models. (B–D) Tumor growth curves and survival rates of mice treated with PBS, scaffold alone, DC vaccine, or FAST-LN. (E) H&E staining of liver tissues (T: tumor, L: normal liver). (F) t-SNE (t-distributed stochastic neighbor embedding) plot showing cells merged from all groups. (G) Histogram indicating the proportion of clusters within each group as indicated. (H) Tumor cells identified by copy number analysis and distribution in each group as indicated. (I) t-SNE plot showing the number of neoantigen-corresponding somatic mutations detected in all cell clusters. (J) Proportion of tumor cells harboring mutant alleles of neoantigen mutations across treatment groups, analyzed exclusively in cells with sequencing reads covering at least one neoantigen mutation site. (K and L) The difference in expression frequency of different neoantigens in each group as indicated. (M) Expression levels of MHC-I and MHC-II in tumor cells of each group. Data are presented as the mean±SEM. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. DC, dendritic cell; FAST-LN, functionalized alginate scaffold for T-cell activation as an artificial lymph node; HCC, hepatocellular carcinoma; MHC, major histocompatibility complex; PBS, phosphate-buffered saline. Meanwhile, in light of the remarkable therapeutic results, we further investigated whether FAST-LN therapy could induce long-term immune memory to prevent tumor recurrence. To simulate tumor recurrence, five mice cured in the FAST-LN treatment group of the Hepa1-6 HCC model from figure 4B were re-inoculated with Hepa1-6 cells for recurrence re-challenge, and five new untreated mice served as the Ctrl group (online supplemental figure S11A). Importantly, no significant tumor growth was observed in all five cured mice after reimplantation of the tumors, whereas tumors continued to grow in the Ctrl group (online supplemental figure S11B). Furthermore, a significantly higher proportion of effector memory CD8+ T cells (Tem, CD44+ and CD62L−) in the spleen were observed in the five cured mice compared with the Ctrl group (FAST-LN vs Ctrl: 39.68±1.367% vs 24.54±1.067%, p<0.0001) (online supplemental figure S11C). To functionally validate the antigen-specific recall response, we performed an IFN-γ ELISpot assay on splenocytes from these mice, which showed significantly enhanced reactivity to neoantigen peptide stimulation in FAST-LN group (online supplemental figure S11D). In general, these results suggest that FAST-LN treatment strategy induces long-term neoantigen-specific immune memory and effectively prevents the recurrence of HCC. To comprehensively explore the changes of immune microenvironment under FAST-LN, single-cell RNA sequencing together with target sequencing for seven neoantigen mutations (Lbr_A341P, Dtnb_K40T, Traf7_C403W, Lmf1_F523V, Ptpn2_I383T, Samd9l_K752M, Makp3_S284F) was performed for the tumor tissues collected from mice treated by PBS or FAST-LN on days 14 (figure 4A). After data processing, 8,000 cells were randomly selected from the sequencing data and cell clustering was performed (figure 4F). As shown in figure 4G, the cell ratio analysis chart shows that the number of immune cells such as macrophages (marked by Adgre1, CD68, Csf1r, online supplemental figure S12A) and T cells (marked by Klrb1c, CD3e, CD4, CD8a) were significantly upregulated in FAST-LN treatment group, while hepatocytes (marked by Alb, Afp) were notably reduced. Further analysis of this expanded macrophage population revealed a significantly polarized functional state, characterized by a highly enriched M1-like gene signature (eg, Nos2, Il12b, Tnf) and no significant change in the M2-like signature (eg, Mrc1, Arg1, Cd163), indicating a shift toward a pro-inflammatory phenotype in the FAST-LN group (online supplemental figure S12B). To further identify the tumor cell components in hepatocytes, we used SCEVAN by evaluating cells’ copy number profiles, which could discriminate tumor cells from normal cells (figure 4H). Consequently, tumor cells were significantly decreased in the FAST-LN treatment group. Since FAST-LN induces neoantigen-specific T cells that target and kill tumor cells expressing seven different neoantigens, we aim to observe the dynamic changes in the mutation frequency of these seven targets at the single-cell level. As shown in figure 4I–K, the mutation frequency of 5/7 neoantigens (Mapk3_S284F, Lmf1_F523V, Smad91_K752M, Traf7_C403W, and Ptpn2_I383T) in tumor cells was significantly downregulated in the FAST-LN treatment group compared with the control group. By further analyzing the number of neoantigens expressed in one tumor cell, it was found that when tumor cells expressed two or more neoantigens, they were more easily killed by FAST-LN-induced neoantigen-specific T cells (figure 4L). Further characterization of the changes in the expression levels of MHC-I and MHC-II in tumor cells found that the FAST-LN treatment can significantly increase the expression of MHC-I and MHC-II in tumor cells, thus significantly enhancing the immunogenicity of the tumor (figure 4M). Overall, these results demonstrate that FAST-LN can effectively induce neoantigen-specific T cells to specifically kill tumor cells expressing neoantigen mutation sites. The dynamic changes of immune cells in tumor during FAST-LN treatment The above results have demonstrated that FAST-LN exerts antitumor effects; however, the mechanisms by which FAST-LN mobilized the immune system in the orthotopic HCC tumor model require further investigation. Neoantigen vaccines primarily rely on T-cell immunity to exert their potent antitumor effects. To explore this, we further analyzed the intratumoral T-cell infiltration in the tumor model of figure 4A using flow cytometry. The results showed that after DC vaccine or FAST-LN treatment, the infiltration of CD4+ and CD8+ T cells was upregulated compared with the other treatment groups (scaffold and Ctrl groups, figure 5A,B). In addition, both CD4+ and CD8+ T cells in the FAST-LN treatment group were significantly higher than those in the DC vaccine treatment group (CD4+ T cell: FAST-LN 5.524±0.5781% vs DC vaccine 3.298±0.2721%, p=0.0012; CD8+ T cell: FAST-LN 15.78±2.185% vs DC vaccine 8.784±0.3485%, p=0.0014). These results are consistent with figure 4G, providing strong evidence that FAST-LN treatment enhances the infiltration of CD4+ and CD8+ T cells in tumors. To gain deeper insights into the functional status changes of different T-cell subpopulations under different treatment conditions, T cells were further extracted and re-clustered based on the single-cell RNA sequencing results. According to the expression of common-acknowledged T-cell functional markers, the re-clustered T cells were identified as five CD8+ T cell clusters (CD8_CCL5 (0), CD8_CD69 (1), CD8_CCNA2 (3), CD8_PRF1 (4), CD8_DUT (5)), and one CD4+ T-cell cluster (CD4_IFN-γ T cells (2)) (figure 5C,D). Notably, compared with the Ctrl group, all T-cell subsets demonstrated significantly enhanced infiltration levels in the FAST-LN treatment group, with particularly pronounced enrichment of CD4_IFN-γ T cells (T-helper 1 cells, Th1) and CD8_CCL5 T cells (figure 5E). CD4_IFN-γ T cells typically facilitate antitumor immunity through secretion of IFN-γ and other cytokines, activating and augmenting CD8+ T-cell function while modulating the tumor microenvironment. Cell communication analysis revealed that CD4_IFN-γ T cells can recruit CD8_CD69 T cells, CD8_PRF1 T cells, and CD8_DUT T cells via the CCL3/4/5-CCR5 signaling axis, while also exhibiting antigen presentation capability to all CD8 T cell subsets (eg, H2-d1-Cd8a) (figure 5F). Importantly, CD8_CD69 T cells, previously identified as primary neoantigen-specific T cells,25 28 demonstrated high expression of cytotoxic markers including Gzmb, while not expressing exhaustion markers such as Pdcd1, Lag3, and Tigit, suggesting potent antitumor cytotoxic capacity (figure 5G). Furthermore, we also observed that CD8_CCL5 T cells can recruit macrophages through the secretion of chemokines including CCL5 (online supplemental figure S13). Macrophages, which were significantly enriched in the FAST-LN treatment group (figure 4G), have been demonstrated to synergize with T cells in vaccine-based therapeutic approaches to enhance antitumor effects.29 In summary, FAST-LN significantly enhanced tumor infiltration and interaction of CD8+ T cells, Th1 cells and macrophages, orchestrating robust antitumor responses. The dynamic changes of immune cells in tumor during FAST-LN treatment. (A and B) Flow cytometry analysis of the proportion of CD3+CD4+ T cells (A) and CD3+CD8+ T cells (B) in mouse tumor tissues (n=5; one-way ANOVA). (C) Violin plot showing the expression of CD3, CD8 and CD4 in T cells. (D) Heatmap showing the marker genes for all T-cell subtypes. (E) t-SNE plots of the enrichment of different T-cell clusters in each group. (F) Ligand-receptor interactions between Th1 cells and CD8+ T cells calculated using Cellchat. (G) Violin plot showing the expression of T-cell function markers. Data are presented as the mean±SEM. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; DC, dendritic cell; FAST-LN, functionalized alginate scaffold for T-cell activation as an artificial lymph node; IFN, interferon. Antitumor efficacy of FAST-LN across various tumor models Considering that the different location of FAST-LN implantation may influence tumor treatment, we further compared the antitumor effects of implantation in the back and axilla. Then, Hepa1-6 cells were implanted into the liver of C57BL/6 mice, and tumor formation was assessed 7 days later. Mice with tumors were randomly divided into three groups. The control group was treated with PBS, while the other two groups received FAST-LN loaded with AdvNeo and BMDCs, implanted in the back or armpit, respectively, on day 0 (figure 6A). After 55 days of tumor monitoring, we observed significant suppression of liver tumor growth in the FAST-LN treatment groups compared with the Ctrl group, with no fatalities recorded (figure 6B–D). Moreover, the antitumor efficacy of FAST-LN was comparable between back and armpit implantation sites. This suggests that FAST-LN can effectively recruit and activate neoantigen-specific T cells and exert potent antitumor effects regardless of the subcutaneous implantation site. In addition, to further validate the antitumor ability and confirm immune response mechanism by FAST-LN, a lung cancer liver metastasis model was constructed by LLC cells, and then subcutaneously implanted with FAST-LN as indicated (online supplemental figure S14A). As expected, similar results as the orthotopic HCC mouse model were observed in the LLC liver metastasis model. Compared with the Ctrl, scaffold and DC vaccine-treated groups, the FAST-LN treatment group exhibited significantly reduced tumor burden (online supplemental figure S14B, C), and the tumor-infiltrated T cells, including Th1 cells and CD8+CD69+ T cells, were also significantly increased (online supplemental figure S14D–G), demonstrating its potential as a powerful strategy for treating metastatic tumors. Antitumor efficacy of FAST-LN across various tumor models. (A) Schematic diagram of different therapy in orthotopic HCC model (n=5 mice). The mice were sacrificed on the day of 55 after initiating treatment. (B) Tumor burden monitoring of PBS (Ctrl), back implant and armpit implant treated mice by bioluminescence imaging. (C) The photographs of tumor-bearing livers collected from mice after different treatments as indicated. (D) Kaplan-Meier survival curves for each treatment group (n=5). (E) Schematic diagram of different therapy in subcutaneous HCC model constructed by HLA-A*0201 transgenic mice (n=6 mice). (F and G) Growth curves of the average tumor volumes in the indicated groups (n=6 mice; two-way ANOVA). (H) Representative images of tumors harvested from tumor-bearing mice on day 20. (I) Schematic diagram of different therapy in subcutaneous HCC model constructed by HLA-A*0201 transgenic mice (n=5 mice). WT cells: the mice inoculated with WT Hepa1-6 cells; Hepa1-6HLA-A*0201/HBc18-27 cells: the mice inoculated with Hepa1-6 cells expressing HLA-A*0201 and the HBc18-27 epitope. (J) Growth curves of the average tumor volumes in the indicated groups (n=5 mice; one-way ANOVA). (K) Representative images of tumors harvested from tumor-bearing mice on day 20. Data are presented as the mean±SEM. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ANOVA, analysis of variance; DC, dendritic cell; FAST-LN, functionalized alginate scaffold for T-cell activation as an artificial lymph node; HCC, hepatocellular carcinoma; PBS, phosphate-buffered saline. To further confirm the function of T-cell stimulation and antitumor effect of FAST-LN in the human immune system, we incorporated an HBV-derived antigen (HBc18-27, with high binding affinity with HLA-A*0201) instead of neoantigens from Hepa1-6 cells to adenovirus to prepare FAST-LN and DC vaccine. Subsequently, the therapeutic efficacy of FAST-LN was evaluated on HLA-A*0201 transgenic mice bearing Hepa1-6 tumor cells. Hepa1-6 cells were infected with adenovirus to express HLA-A*0201 and the HBc18-27 epitope prior to seeding in mice as our previous report.26 After inoculating Hepa1-6 cellsHLA-A*0201/HBc18-27, the mice with tumors of approximately 60 mm3 were randomly divided into four groups and the time was recorded as day 0. Meanwhile, these mice were treated by PBS, scaffold, DC vaccine and FAST-LN on day 0, respectively (figure 6E). The mice inoculated with PBS and scaffold suffered from continuous and rapid tumor growth, while the tumors in the DC vaccine group showed a slow progression. Notably, under FAST-LN treatment, significant tumor regression was observed in mice (figure 6F–H). This regression was also very consistent across multiple time points, indicating a sustained and potent antitumor effect. Furthermore, this potent antitumor effect was strictly dependent on the expression of the HBc18-27 epitope by the tumor cells, as FAST-LN showed no significant efficacy against wild-type Hepa1-6 tumors lacking this specific antigen (figure 6I–K). Collectively, our results suggest that FAST-LN, as an artificial lymph node for DC-T cell interaction, can elicit strong immune responses in humanized mice with HLA-A*0201 restriction, highlighting its potential as an effective immunotherapeutic strategy to induce antitumor immunity in human system. Discussion The development of effective cancer immunotherapies demands innovative strategies to overcome the limitations of conventional tumor vaccines, such as their delayed and inefficient induction of neoantigen-specific T cells.13 30 31 To address these challenges, we introduce FAST-LN, an implantable artificial lymph node designed to provide a biomimetic microenvironment for the rapid in vivo generation and expansion of tumor neoantigen-specific T cells. Unlike traditional DC vaccines, which rely on the migration of DCs to lymph nodes or tumor sites to activate T cells, FAST-LN enables direct T-cell activation within its engineered microenvironment. This approach significantly enhances the efficiency of neoantigen-specific T-cell induction. The FAST-LN platform leverages a biocompatible alginate scaffold enriched with cytokines, which not only preserves the functional integrity of DCs and T cells but also prolongs their interaction time, resulting in more robust and durable antitumor responses. Furthermore, at the same injection frequency, FAST-LN exhibits faster and stronger efficacy compared with conventional tumor vaccines, suggesting its potential to reduce the need for repeated administrations while achieving superior therapeutic outcomes. By integrating efficient neoantigen-specific T-cell induction with sustained antitumor activity, FAST-LN represents a transformative advancement in cancer immunotherapy, offering a promising solution to improve therapeutic outcomes. Clinically, to rapidly generate antigen-specific T cells, adoptive T-cell therapies (ACTs), including CAR-T (Chimeric Antigen Receptor T) and TCR-T (T Cell Receptor Engineered-T), are commonly employed.32,35 These approaches involve the ex vivo preparation of antigen-specific T cells via lentiviral infection, followed by reinfusion into patients. While ACTs have demonstrated significant antitumor efficacy, particularly in B-cell malignancies, their broader clinical application is hindered by several challenges. The complex manufacturing process, high cost of generating clinical-grade T cells, poor T-cell persistence, and the onset of T-cell exhaustion limit their overall therapeutic benefits.36 These challenges are even more pronounced in solid tumor treatment, where antigen-specific T cells struggle to efficiently expand, persist, and infiltrate the tumor microenvironment, thereby reducing the effectiveness of ACTs in cancer therapy.37 Unlike conventional CAR-T or TCR-T cell therapies that require ex vivo preparation, FAST-LN enables the in vivo generation of antigen-specific T cells, significantly simplifying the process. In this system, DCs play a central role in T-cell induction through MHC/neoantigen complexes. This in vivo activation process is more physiologically relevant and occurs under milder conditions, helping to maintain high T-cell stemness while minimizing exhaustion. Additionally, co-stimulatory molecules such as CD28-CD80/CD86 further enhance the activation and proliferation of these antigen-specific T cells. Therefore, in preclinical studies using orthotopic HCC and metastatic lung cancer models, FAST-LN treatment resulted in complete tumor suppression and a 100% survival rate in tumor-bearing mice. Importantly, the sustained release of cytokines and chemokines within the scaffold ensures prolonged immune activation, while its biodegradability minimizes the risks associated with long-term implantation. Furthermore, the ability of FAST-LN to induce effector memory T cells and protect against tumor rechallenge highlights its capacity to establish durable immune memory, addressing a critical unmet need in cancer therapy. In conclusion, FAST-LN represents a paradigm shift in neoantigen-based immunotherapy by creating an in vivo platform for rapid, localized T-cell priming. Its ability to overcome the spatial and temporal limitations of conventional vaccines positions it as a promising strategy for treating rapidly progressing malignancies. Supplementary material Footnotes Funding: This work was supported by the National Natural Science Foundation of China (Grant No.82573538, 82572108 and 82202982); the Major Research Projects for Young and Middle-aged Talent of Fujian Provincial Health Commission (Grant No.2022ZQNZD014); Fujian Province Joint Funding Project for Scientific and Technological Innovation (2024Y9420); Clinical Research Center for Infectious Diseases of Fujian Province (Grant No.2022Y2018); Fuzhou Science and technology planning project (Grant No. 2023-S-wt1). Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Not applicable. Ethics approval: Not applicable. Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information. References Associated Data Supplementary Materials Data Availability Statement All data generated or analyzed during this study are included in this article. Inquiries about further material documents can be directed to the corresponding authors. The single-cell transcriptome sequencing data from this publication have been deposited to the Genome Sequencing Archive (GSA) database (GSA number: CRA024359). All data relevant to the study are included in the article or uploaded as supplementary information.