Mucoadhesive oral nanovaccine delivery system with immunomodulatory capacity for synergistic colorectal cancer therapy Authors: Qian Ming, Junjie Lin, Yingying Zhou, Ziqi Song, Zihan Yu, Xinze Liu, Xiao Sun, Lianyu Zhao, Changyu Lu, Natong Liu, Yanan Qin, Jing Wu, Zhaowei Li, Fengying Shao Published: 2026 ABSTRACT Colorectal cancer (CRC) remains a prevalent malignancy, with current chemotherapies limited by drug resistance, poor bioavailability, and severe toxicity. Although cancer vaccines offer a promising alternative, their efficacy is constrained by inefficient tumor targeting and insufficient immune activation. Here, we present an oral nanovaccine delivery system (HML@HN) that integrates gastric protection, intestinal adhesion, tumor-specific delivery, and robust immune activation for enhanced CRC immunotherapy. The HML@HN nanovaccine is constructed by loading 5-FU-containing tumor cell lysate (TCL) into hollow manganese dioxide (HM) and co-coating with norepinephrine (NE) and hyaluronic acid (HA). The NE coating safeguards against gastric degradation while conferring mucoadhesive properties for prolonged intestinal retention, and HA enables active tumor targeting via CD44-mediated endocytosis. Upon internalization, the HM core responds to glutathione in the tumor microenvironment, triggering Mn2+ release and reactive oxygen species (ROS) generation for chemodynamic therapy (CDT). This process, together with 5-FU, induces damage-associated molecular patterns (DAMPs), while the loaded TCL augments immunogenicity and the released Mn2+ promotes dendritic cell maturation. In an orthotopic CRC mouse model, oral HML@HN significantly suppresses tumor growth through these synergistic mechanisms. Collectively, this work presents a multifunctional oral nanovaccine that uniquely combines gastric protection, mucosal adhesion, tumor-responsive drug release, and potent immune activation for CRC immunotherapy. Introduction Colorectal cancer (CRC) is one of the most prevalent malignancies worldwide, ranking third in incidence and second in cancer-related mortality [1]. In China, its incidence has risen to the second position, with a persistent upward trend and an alarming increase among younger populations [2]. Current treatment modalities for CRC primarily include surgical intervention, chemotherapy, and radiotherapy [3,4]. Among these, chemotherapy (exemplified by 5-fluorouracil, 5-FU) remains a cornerstone in both adjuvant and neoadjuvant settings [5,6]. However, its clinical efficacy is severely limited by significant challenges, including the occurrence of drug resistance, poor bioavailability, and pronounced systemic toxicity, underscoring the urgent need for novel therapeutic strategies with improved specificity and safety [7]. In recent years, cancer immunotherapy has emerged as a transformative approach, broadly categorized into immune checkpoint blockade, adoptive cell transfer, cancer vaccines and so on [[8], [9], [10], [11], [12]]. Among these, cancer vaccines offer distinct advantages by training the host immune system with tumor-specific antigens to precisely recognize and eliminate malignant cells [13,14]. Cancer vaccines are broadly classified into prophylactic vaccines, which are well-established for preventing virus-associated cancers, and therapeutic vaccines, most of which remain under clinical investigation [[15], [16], [17], [18]]. Therapeutic vaccines provide high specificity, durable immune memory, and a favorable safety profile, representing a promising strategy for establishing long-term antitumor immunity [[19], [20], [21]]. Notably, vaccines based on whole tumor cell lysates (TCL) have gained considerable attention due to their ability to present a broad spectrum of tumor-associated antigens, thereby circumventing the need for antigen identification and reducing the risk of immune evasion [22,23]. Despite these advantages, current cancer vaccines face critical hurdles in achieving efficient colorectal tumor targeting and eliciting potent, specific immune activation [[24], [25], [26]], highlighting the urgent need for advanced delivery strategies and effective immunoadjuvants. Given that CRC originates in the terminal portion of the digestive tract, oral administration represents one of the most direct and effective routes for drug delivery [27]. Localized drug release in the colon and rectum enables precise tumor targeting and can elicit robust immune responses within the colorectal region, thereby controlling tumor growth and improving therapeutic outcomes [[28], [29], [30], [31]]. Moreover, because the drug primarily acts locally, oral administration minimizes systemic drug exposure, reducing off-target toxicity [32]. Additionally, compared with intravenous therapy, oral regimens offer superior patient compliance, facilitating sustained and consistent treatment [33,34]. Beyond the delivery route, the design of the delivery carrier is equally critical. An ideal carrier must fulfill two key functions: firstly, it must protect the payload from degradation in the harsh gastrointestinal environment and enable targeted accumulation in colorectal tumors. For instance, certain delivery systems can withstand gastric acid and enzymatic digestion, reaching the intestinal tract intact, and can be engineered with surface properties (e.g., positive charge) to enhance retention and enrichment at tumor sites [35,36]. Secondly, beyond serving as a passive vehicle, the carrier should function as an immunoadjuvant, actively engaging innate immune pathways to amplify antitumor immunity [37,38]. In this regard, manganese-based materials have attracted considerable interest, as they can release Mn2+ ions that promote dendritic cells (DCs) maturation and enhance antigen presentation while simultaneously delivering tumor antigens [39,40]. Bearing this in mind, we developed an oral nanovaccine (HML@HN) based on hollow manganese dioxide (HM) loaded with TCL and sequentially coated with norepinephrine (NE) and hyaluronic acid (HA) to achieve efficient colorectal targeting and robust immunotherapy, as illustrated in Scheme 1. Upon oral administration, the nanovaccine adheres effectively to the colonic mucosa, facilitated by the mucoadhesive catechol groups of NE and the tumor-targeting properties of HA, enabling sustained retention and localized concentration. Following internalization by tumor cells, the hollow MnO2 structure responds to the high glutathione (GSH) concentration in the cytoplasm, releasing Mn2+ ions and catalyzing the conversion of endogenous H2O2 into cytotoxic ·OH radicals. Concurrently, the loaded TCL, which contains 5-FU derived from the lysate, exerts additional chemotherapeutic effects, collectively inducing robust tumor cell apoptosis. The resulting tumor debris is captured by DCs in the intestinal immune system, while the released Mn2+ promotes DCs activation, overcoming local immune tolerance, promoting DCs maturation, and ultimately driving a potent, antigen-specific antitumor immune response against colorectal cancer.Scheme 1Schematic illustration of the synthesis process and mechanism of HML@HN. Schematic illustration of the synthesis process and mechanism of HML@HN. Results and discussion Preparation and characterizations of HML@HN Initially, HM nanoparticles with a diameter of approximately 70 nm were synthesized following a previously reported method (Fig. 1a). After loading with TCL, the adhesive molecule NE and the tumor-targeting molecule HA were modified on the surface via ultrasonication and formed HML@HN. To confirm the successful construction, the nanoparticles were characterized from multiple perspectives. Firstly, we characterized the surface properties, mainly including surface morphology and electric potential. Transmission electron microscopy (TEM) images revealed that the prepared HML@HN nanoparticles exhibited uniform size distribution with an average diameter of around 75 nm (Fig. 1b and c). Further zeta potential analysis demonstrated that HML@HN featured a negative surface charge of −22.05 mV (Fig. 1d). Secondly, we investigated the elemental composition and valence state of the main elements in HML@HN. Energy-dispersive X-ray spectroscopy (EDS) results showed the presence of elements from HM (manganese and oxygen), TCL (carbon, oxygen, nitrogen, phosphorus and fluorine), NE and HA (carbon, oxygen, and nitrogen), confirming the successful construction of the HML@HN. This was also evident from the spectrum analysis, which consistently exhibited the elements of carbon, manganese, nitrogen, oxygen, fluorine and phosphorus (Fig. 1e, f and Table S1). Subsequently, we conducted X-ray photoelectron spectroscopy (XPS) analysis to evaluate the valence state of manganese. Results in Fig. 1g and h showed that the manganese element was present in the tetravalent state, highlighting the potential of generating reactive oxygen species (ROS) with H2O2 after being reduced by the reductive GSH in the tumor cells. To explore this property, we detected the ·OH production ability under different conditions using methylene blue (MB) as a probe to scavenge ·OH. As we can see from Fig. 1i, the groups of MB, MB + H2O2+ HM, and MB + GSH + HM shared similar characteristic peaks and absorbance, indicating the negligible ·OH generation ability. Notably, when GSH and H2O2 were present simultaneously, the UV-vis absorbance of MB was apparently decreased and the color faded (Fig. S1), suggesting the efficient ·OH production ability.Fig. 1Synthesis and characterizations of HML@HN. (a) TEM image of HM (scale bar: 50 nm). (b) TEM image of HML@HN (scale bar: 50 nm). (c) Size distribution analysis of HML@HN nanomedicine. (d) Zeta potential analysis of HM, HML, HML@NE, and HML@HN nanomedicine. (e) Elemental mapping analysis of HML@HN (scale bar: 50 nm). (f) EDS analysis of HML@HN nanomedicine. (g) XPS spectra of HM@HN. (h) High-resolution XPS spectrum of Mn 2p. (i) UV-vis spectra of MB after different treatments for ·OH generation analysis. (j) TEM images of HML@HN after incubation in simulated gastric fluid (scale bar: 50 nm). (k) TEM images of HML@HN after incubation in simulated intestinal fluid (scale bar: 50 nm). (l) Cumulative release of Mn ions from HML@HN during incubation in simulated gastric fluid (the inserted result represents the Mn ion release profile from 1 h to 2 h). (m) Cumulative release of Mn ions from HML@HN during incubation in artificial intestinal fluid. Data are expressed as the mean ± s.d. Synthesis and characterizations of HML@HN. (a) TEM image of HM (scale bar: 50 nm). (b) TEM image of HML@HN (scale bar: 50 nm). (c) Size distribution analysis of HML@HN nanomedicine. (d) Zeta potential analysis of HM, HML, HML@NE, and HML@HN nanomedicine. (e) Elemental mapping analysis of HML@HN (scale bar: 50 nm). (f) EDS analysis of HML@HN nanomedicine. (g) XPS spectra of HM@HN. (h) High-resolution XPS spectrum of Mn 2p. (i) UV-vis spectra of MB after different treatments for ·OH generation analysis. (j) TEM images of HML@HN after incubation in simulated gastric fluid (scale bar: 50 nm). (k) TEM images of HML@HN after incubation in simulated intestinal fluid (scale bar: 50 nm). (l) Cumulative release of Mn ions from HML@HN during incubation in simulated gastric fluid (the inserted result represents the Mn ion release profile from 1 h to 2 h). (m) Cumulative release of Mn ions from HML@HN during incubation in artificial intestinal fluid. Data are expressed as the mean ± s.d. The primary objective in designing HML@HN was to establish an effective vaccine depot within the colorectal region, with the key challenge being to maintain its stability throughout the gastrointestinal tract, particularly under the highly acidic conditions of the stomach, which is a major hurdle for most orally administered drugs. To this end, the stability of HML@HN in the digestive environment was evaluated by incubating the formulations in simulated gastric fluid (SGF) and simulated intestinal fluid (SIF). As shown in Fig. 1j and k, HML@HN retained its structural integrity even after 2 h of exposure to both SGF and SIF, confirming its robustness under gastric and intestinal conditions. For effective immunostimulation, Mn ions from the mucosal vaccine must be released in a sustained manner within the intestine while minimizing premature release in the stomach. Accordingly, the release profile of HML@HN was investigated in SGF and SIF. Given that the maximum gastric transit time in humans is approximately 2 h, only about 18% of Mn ions were released after 2 h in SGF (Fig. 1l), indicating the absence of a burst release in the stomach. In contrast, once HML@HN reached the intestinal environment, the remainingMn ions were gradually released, with over 80% released within 48 h (Fig. 1m). Collectively, this release profile ensures the vaccine's sustained immunostimulatory efficacy. In vitro endocytosis, cytotoxicity and immune activation analysis Following the successful construction of the HML@HN nanovaccine, we systematically evaluated its mode of action and therapeutic efficacy at the cellular level. To achieve efficient enrichment in both intestinal and tumor tissues, the nanoparticle surface was functionalized with NE and HA, which were designed to enhance targeting at both tissue and cellular levels. Specifically, CD44 receptors are known to be highly expressed on the surface of colorectal cancer cells, and HA can specifically bind to CD44 [41], thereby facilitating active cellular uptake and enabling targeted delivery. To this end, we first evaluated the endocytic efficiency of the nanoparticles in CT26 cells using confocal microscopy and flow cytometry. As shown in Fig. 2a–c, time-dependent increases in intracellular uptake were observed in both the HML@NE and HML@HN treatment groups. Notably, the HML@HN group exhibited significantly higher internalization compared to the HML@NE group, confirming the critical role of HA modification in enhancing the targeting capability toward colorectal cancer cells. To elucidate the underlying endocytic mechanism, HML@HN was co-incubated with CT26 cells in the presence of various endocytosis inhibitors, including chlorpromazine, cytochalasin D, nystatin, and genistein. The results demonstrated that internalization efficiency was markedly reduced upon treatment with chlorpromazine and 4 °C, indicating that the endocytic pathway of HML@HN mainly via clathrin-mediated endocytosis (Fig. 2d).Fig. 2In vitro endocytosis, cytotoxicity and immune activation analysis. (a) CLSM images of CT26 cells incubated with Cy3-labeled HML@NE and HML@HN at different time points (scale bar: 20 μm; red: nanoparticles, green: cytoskeleton, blue: nucleus). (b) Flow cytometry analysis of CT26 cells incubated with Cy3-labeled HML@NE or HML@HN at different time points. (c) Corresponding quantification of the mean fluorescence intensity after 6 h of co-incubation (n = 3). (d) Flow cytometry analysis of CT26 cells incubated with Cy3-labeled HML@HN after being treated with different types of endocytosis inhibitors, including nystatin, genistein, cytochalasin D, chlorpromazine, and 4 °C treatment (n = 3). (e) Fluorescence images of ROS generation in CT26 cells treated with PBS (Control), HM, HML, HML@NE, and HML@HN for 4 h (scale bar: 150 μm; green: ROS). (f) Cell viability of CT26 cells at 24 h after receiving different treatments (n = 3). (g) Apoptosis analysis of CT26 cells receiving different treatments for 12 h using flow cytometry. (h) Schematic illustration of the Transwell experiment. (i) Representative flow cytometry analysis of mature DCs in control, HM, HML, HML@NE, and HML@HN groups. Data are expressed as the mean ± s.d. ∗∗∗∗p < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) In vitro endocytosis, cytotoxicity and immune activation analysis. (a) CLSM images of CT26 cells incubated with Cy3-labeled HML@NE and HML@HN at different time points (scale bar: 20 μm; red: nanoparticles, green: cytoskeleton, blue: nucleus). (b) Flow cytometry analysis of CT26 cells incubated with Cy3-labeled HML@NE or HML@HN at different time points. (c) Corresponding quantification of the mean fluorescence intensity after 6 h of co-incubation (n = 3). (d) Flow cytometry analysis of CT26 cells incubated with Cy3-labeled HML@HN after being treated with different types of endocytosis inhibitors, including nystatin, genistein, cytochalasin D, chlorpromazine, and 4 °C treatment (n = 3). (e) Fluorescence images of ROS generation in CT26 cells treated with PBS (Control), HM, HML, HML@NE, and HML@HN for 4 h (scale bar: 150 μm; green: ROS). (f) Cell viability of CT26 cells at 24 h after receiving different treatments (n = 3). (g) Apoptosis analysis of CT26 cells receiving different treatments for 12 h using flow cytometry. (h) Schematic illustration of the Transwell experiment. (i) Representative flow cytometry analysis of mature DCs in control, HM, HML, HML@NE, and HML@HN groups. Data are expressed as the mean ± s.d. ∗∗∗∗p < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Following internalization, the capacity of HML@HN to generate ROS is a critical determinant of its cytotoxic activity. However, high intracellular concentrations of GSH can reduce cytotoxicity by scavenging ROS. To investigate the effect of HML@HN on GSH, the flow cytometry result demonstrated that HML@HN could effectively consume GSH after treating CT26 cells with different concentrations of HML@HN (Fig. S2). Furthermore, ROS production was assessed using a fluorescent probe following incubation of different formulations with CT26 cells. Fluorescence imaging revealed detectable ROS signals in the HM, HML, HML@NE, and HML@HN groups, confirming efficient ROS generation mediated by HM. Among these, the HML@HN group exhibited the most pronounced ROS production, underscoring its superior chemodynamic therapy (CDT) efficacy (Fig. 2e). Having confirmed efficient cellular uptake and ROS generation, we next evaluated the in vitro cytotoxicity of HM, HML, HML@NE, and HML@HN against CT26 cells using the CCK-8 assay. As illustrated in Fig. 2f, cell viability decreased in a concentration-dependent manner across all treatment groups, with HML@HN demonstrating the most potent cytotoxic effect. This finding was further corroborated by flow cytometric analysis of apoptosis using Annexin V-FITC/PI staining, which consistently confirmed that HML@HN induced the highest rate of tumor cell apoptosis (Fig. 2g and Fig. S3). Having established the potent cytotoxicity of HML@HN against CT26 cells, we next investigated its capacity to initiate downstream antitumor immunity. According to our initial design, the concerted action of Mn2+-mediated CDT and the chemotherapeutic agent 5-FU, which was derived from the loaded TCL, effectively kills tumor cells, leading to the release of damage-associated molecular patterns (DAMPs). For example, calreticulin (CRT) is an important member of DAMPs, which is exposed on the cell membrane surface during immunogenic cell death, sending an “eat-me” signal that promotes the maturation of DCs. As shown in Fig. S4, CT26 cells treated with HML@HN exhibited the strongest CRT fluorescence intensity, indicating that the HML@HN group was able to effectively induce DAMPs. Concurrently, Mn2+ released from the HM was capable of activating the cGAS-STING signaling pathway in DCs (Fig. S5), thereby orchestrating a potent immune response. Additionally, upon endosomal escape, the TCL can be released extracellularly, further contributing to DCs activation. Based on this rationale, we evaluated the ability of HML@HN to activate bone marrow-derived dendritic cells (BMDCs) using a Transwell co-culture system. As shown in Fig. 2h and i and Fig. S6, after 24 h of co-culture with CT26 cells pre-treated with various formulations, DCs maturation was progressively enhanced in the order of control, HM, HML, HML@NE, and HML@HN groups. These results collectively demonstrate that HML@HN exhibits superior capacity to activate DCs, underscoring its potential as a multifunctional oral nanovaccine capable of both direct tumor killing and robust immune activation. Biosafety and in vivo biodistribution of HML@HN After systematically evaluating the mode of action, cytotoxicity, and immune activation of HML@HN at the cellular level, we next performed a series of in vivo investigations. We first assessed the biosafety profile of the nanovaccine following oral administration. Specifically, blood samples were collected for routine hematological and serum biochemical analyses. As shown in Fig. 3a, no significant changes were observed in the levels of red blood cells (RBC), white blood cells (WBC), hemoglobin (HGB), or platelets (PLT), indicating favorable hematological compatibility. Moreover, markers reflecting liver and kidney function, including alanine aminotransferase (ALT), alkaline phosphatase (ALP), blood urea nitrogen (BUN), and creatinine (CR), all remained within normal ranges (Fig. 3b). In addition to hematological safety, histopathological analysis via H&E staining revealed no obvious damage to major organs in treated mice (Fig. 3c), collectively confirming the favorable biosafety profile of HML@HN.Fig. 3In vivo biosafety and biodistribution analysis of HML@HN. (a) Blood routine examinations including red blood cells (RBC), white blood cells (WBC), hemoglobin (HGB), and platelets (PLT) of BALB/c mice in control and HML@HN treatment groups (n = 3). (b) Blood biochemical indexes including aminotransferase (ALT), alkaline phosphatase (ALP), blood urea nitrogen (BUN), and creatinine (CR) of BALB/c mice in control and HML@HN treatment groups (n = 3). (c) H&E staining images of heart, liver, spleen, lung, kidney, stomach, small intestine, and cecum from mice in control or HML@HN treatment groups (scale bar: 200 μm). (d) In vivo fluorescence imaging of tumor-bearing mice at different time points after oral administration of DiR, DiR-labeled HML@NE, and DiR-labeled HML@HN. (e) Corresponding quantification of average radiant efficiency in different treatment groups (n = 3). (f) Ex vivo fluorescence imaging of major organs (heart, liver, spleen, lung, kidney, and tumor tissues). Nns represent no significance. Data are expressed as the mean ± s.d. p > 0.05 (no significance, ns). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) In vivo biosafety and biodistribution analysis of HML@HN. (a) Blood routine examinations including red blood cells (RBC), white blood cells (WBC), hemoglobin (HGB), and platelets (PLT) of BALB/c mice in control and HML@HN treatment groups (n = 3). (b) Blood biochemical indexes including aminotransferase (ALT), alkaline phosphatase (ALP), blood urea nitrogen (BUN), and creatinine (CR) of BALB/c mice in control and HML@HN treatment groups (n = 3). (c) H&E staining images of heart, liver, spleen, lung, kidney, stomach, small intestine, and cecum from mice in control or HML@HN treatment groups (scale bar: 200 μm). (d) In vivo fluorescence imaging of tumor-bearing mice at different time points after oral administration of DiR, DiR-labeled HML@NE, and DiR-labeled HML@HN. (e) Corresponding quantification of average radiant efficiency in different treatment groups (n = 3). (f) Ex vivo fluorescence imaging of major organs (heart, liver, spleen, lung, kidney, and tumor tissues). Nns represent no significance. Data are expressed as the mean ± s.d. p > 0.05 (no significance, ns). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Having established its biocompatibility, we next evaluated the in vivo biodistribution of HML@HN in a murine model of colorectal cancer using a small animal in vivo imaging system. To enhance intestinal retention, the nanoparticle surface was functionalized with NE, which undergoes autoxidation to form a protective poly-NE film that shields the nanocarrier from external environmental assaults. Furthermore, the catecholamine groups of NE, known for their strong adhesive properties in mussel foot proteins, endow the nanocarrier with robust mucosal adhesion [42,43], thereby prolonging intestinal retention without compromising its functional integrity. Compared with mice treated with free DiR dye, the HML@NE group exhibited significantly higher and more sustained intestinal retention. Notably, the HML@HN group, which incorporated HA modification, demonstrated even longer retention, with fluorescence signals detectable up to 24 h post-administration, underscoring the critical role of HA in enhancing colorectal tumor targeting. In accordance with these findings, ex vivo imaging of harvested tissues confirmed that HML@HN achieved the highest fluorescence accumulation in intestinal tumor tissues at 24 h post-oral administration, further validating its superior targeting capability (Fig. 3d–f). Collectively, these results demonstrate that HML@HN exhibits favorable biosafety and achieves sustained intestinal retention with enhanced tumor-specific accumulation through the synergistic effects of NE-mediated mucoadhesion and HA-mediated active targeting. In vivo antitumor efficacy of HML@HN in an orthotopic colorectal cancer model Having established the favorable biosafety profile and tumor-targeting capability of HML@HN, we next systematically evaluated its therapeutic potential as an oral cancer vaccine in an orthotopic colorectal cancer mouse model. To this end, a CT26-luc orthotopic colon cancer model was established, and mice were orally administered PBS, HM, HML, HML@NE, or HML@HN according to the treatment schedule illustrated in Fig. 4a. Tumor progression was dynamically monitored by bioluminescence imaging at the indicated time points (Fig. 4b). The control group exhibited rapid tumor growth, as evidenced by progressively increasing bioluminescence signals, confirming successful establishment of the orthotopic model. The HM group showed only marginal antitumor activity, reflecting the limited efficacy of the manganese dioxide carrier alone. In contrast, the HML@NE group demonstrated significantly enhanced therapeutic outcomes, with bioluminescence intensity being substantially suppressed by day 12, an effect primarily attributed to the prolonged intestinal retention conferred by NE modification. Notably, the HML@HN group achieved the most pronounced tumor inhibition among all treatment groups, underscoring the critical role of HA in facilitating tumor-specific targeting (Fig. 4c). Throughout the treatment period, body weight remained stable across all groups, providing strong evidence for the favorable in vivo biosafety profile of HML@HN (Fig. 4d). On day 15, mice were euthanized, and tumor tissues were harvested for direct assessment of antitumor efficacy. The tumor burden followed a descending order across the treatment groups, with the PBS group exhibiting the largest tumor size, followed sequentially by the HM, HML, HML@NE, and HML@HN groups (Fig. 4e). These results further corroborate the superior antitumor performance of HML@HN. To gain deeper insight into the underlying mechanism, tumor sections from each group were subjected to hematoxylin-eosin (H&E) staining and terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay. As shown in Fig. 4f and g, the HML@HN group exhibited the highest rate of tumor cell apoptosis, confirming its outstanding antitumor efficacy. Collectively, these results demonstrate that HML@HN achieves potent antitumor activity in an orthotopic colorectal cancer model through the synergistic effects, positioning it as a promising oral nanovaccine for colorectal cancer immunotherapy.Fig. 4In vivo antitumor efficacy analysis of HML@HN in an orthotopic colorectal cancer model. (a) Schematic illustration of the treatment schedule for CT26-luc tumor-bearing BALB/c mice. (b) Representative bioluminescence images of tumor-bearing mice after oral administration of different formulations. (c) Corresponding quantification of the average radiance (n = 3). (d) Body weight changes of mice in each group during the treatment period (n = 6). (e) Representative photographs of tumor tissues in different treatment groups. (f) H&E staining analysis of tumor tissues from mice receiving different treatments (scale bar: 100 μm). (g) TUNEL staining analysis of tumor tissues from mice receiving different treatments (scale bar: 100 μm). Data are expressed as the mean ± s.d. In vivo antitumor efficacy analysis of HML@HN in an orthotopic colorectal cancer model. (a) Schematic illustration of the treatment schedule for CT26-luc tumor-bearing BALB/c mice. (b) Representative bioluminescence images of tumor-bearing mice after oral administration of different formulations. (c) Corresponding quantification of the average radiance (n = 3). (d) Body weight changes of mice in each group during the treatment period (n = 6). (e) Representative photographs of tumor tissues in different treatment groups. (f) H&E staining analysis of tumor tissues from mice receiving different treatments (scale bar: 100 μm). (g) TUNEL staining analysis of tumor tissues from mice receiving different treatments (scale bar: 100 μm). Data are expressed as the mean ± s.d. Immune activation effects of HML@HN in vivo Based on the outstanding antitumor efficacy of HML@HN demonstrated in vivo, we next sought to elucidate its underlying immune activation mechanisms. The therapeutic effect of HML@HN is attributed to a synergistic combination of three contributing factors: Mn2+-mediated CDT, the chemotherapeutic action of 5-FU derived from the loaded TCL, and, importantly, the induction of a robust antitumor immune response. To investigate the immune mechanisms, an orthotopic colorectal cancer mouse model was established, and mice were subjected to various treatments, including PBS, HM, HML, HML@NE, and HML@HN, followed by flow cytometry and immunofluorescence staining analyses. We first examined the activation status of antigen-presenting cells, including DCs and macrophages. Consistent with the in vitro DCs activation results, HML@HN treatment induced the highest proportion of mature DCs, reaching 48.4%, demonstrating its potent capacity to stimulate antigen presentation (Fig. 5a). With respect to macrophage polarization, HML@HN treatment resulted in a marked increase in the proportion of M1-type macrophages (F4/80+CD86+) to 49.1%, approximately three times that observed in the control group (Fig. 5b). Correspondingly, the proportion of M2-type macrophages (F4/80+CD206+) decreased to 19.9%, representing a 2.5-fold reduction compared to the control group (Fig. 5c). Having established the effective activation of antigen-presenting cells, we next analyzed the impact on T lymphocyte populations. Flow cytometric analysis revealed that the HML@HN-treated group exhibited significantly higher proportions of both CD8+ T cells (CD8+CD3+) and CD4+ T cells (CD4+CD3+) compared to the other treatment groups (Fig. 5d and e), while showed lower proportions of Treg cells (FOXP3+CD25+CD45+CD4+CD3+), reflecting robust activation of adaptive immune responses (Fig. 5f). These findings were further corroborated by immunofluorescence staining, which confirmed a substantial increase in CD4+ and CD8+ T cell infiltration in the HML@HN group (Fig. 5g). Furthermore, the variation of CD8+ T cells and CD4+ T cells in the spleen after HML@HN treatment further demonstrated the enhancement of the adaptive immune system (Fig. S7). Collectively, these results demonstrate that HML@HN effectively orchestrates a multifaceted antitumor immune response by promoting DCs maturation, reprogramming macrophages toward the pro-inflammatory M1 phenotype, and enhancing the infiltration of effector T cells into the tumor microenvironment.Fig. 5In vivo immune activation analysis of HML@HN. (a) Representative flow cytometry analysis and corresponding quantification of mature DCs (CD80+CD86+ gated on CD11c+) in tumor tissues from mice receiving different treatments (n = 6). (b) Representative flow cytometry analysis and corresponding quantification of M1 macrophages (F4/80+CD86+) in tumor tissues from mice receiving different treatments (n = 6). (c) Representative flow cytometry analysis and corresponding quantification of M2 macrophages (F4/80+CD206+) in tumor tissues from mice receiving different treatments (n = 6). (d) Representative flow cytometry analysis and corresponding quantification of CD4+ T cells (CD4+CD3+ gated on CD45+) in tumor tissues from mice receiving different treatments (n = 6). (e) Representative flow cytometry analysis and corresponding quantification of CD8+ T cells (CD8+CD3+ gated on CD45+) in tumor tissues from mice receiving different treatments (n = 6). (f) Representative flow cytometry analysis and corresponding quantification of regulatory T cells (FOXP3+CD25+CD4+CD3+ gated on CD45+) in tumor tissues from mice receiving different treatments (n = 6). (g) Representative immunofluorescence staining images of tumor tissues receiving different treatments (blue: nucleus; red: CD4+ or CD8+ T cells). Data are expressed as the mean ± s.d.∗p< 0.05, ∗∗p< 0.01, ∗∗∗p< 0.001, ∗∗∗∗p < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) In vivo immune activation analysis of HML@HN. (a) Representative flow cytometry analysis and corresponding quantification of mature DCs (CD80+CD86+ gated on CD11c+) in tumor tissues from mice receiving different treatments (n = 6). (b) Representative flow cytometry analysis and corresponding quantification of M1 macrophages (F4/80+CD86+) in tumor tissues from mice receiving different treatments (n = 6). (c) Representative flow cytometry analysis and corresponding quantification of M2 macrophages (F4/80+CD206+) in tumor tissues from mice receiving different treatments (n = 6). (d) Representative flow cytometry analysis and corresponding quantification of CD4+ T cells (CD4+CD3+ gated on CD45+) in tumor tissues from mice receiving different treatments (n = 6). (e) Representative flow cytometry analysis and corresponding quantification of CD8+ T cells (CD8+CD3+ gated on CD45+) in tumor tissues from mice receiving different treatments (n = 6). (f) Representative flow cytometry analysis and corresponding quantification of regulatory T cells (FOXP3+CD25+CD4+CD3+ gated on CD45+) in tumor tissues from mice receiving different treatments (n = 6). (g) Representative immunofluorescence staining images of tumor tissues receiving different treatments (blue: nucleus; red: CD4+ or CD8+ T cells). Data are expressed as the mean ± s.d.∗p< 0.05, ∗∗p< 0.01, ∗∗∗p< 0.001, ∗∗∗∗p < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) Conclusions In summary, we have developed an oral nanovaccine delivery system (HML@HN) that integrates gastric protection, intestinal adhesion, tumor-specific delivery, and robust immune activation for colorectal cancer immunotherapy. The system was constructed by loading 5-FU-containing TCL into HM and co-coating with NE and HA. This rational design addresses key challenges in oral vaccine development as the NE coating provides dual functionality by protecting the nanovaccine from gastric degradation while conferring strong mucoadhesive properties for prolonged intestinal retention, the HA modification enables active tumor targeting via CD44-mediated endocytosis, and the TCL serves as a source of whole-cell tumor antigens to elicit potent antitumor immunity. Upon internalization by tumor cells, the HM core responds to the high glutathione concentration in the tumor microenvironment, triggering sustained Mn2+ release and ROS generation for CDT. This process, in combination with the chemotherapeutic effect of 5-FU derived from the loaded TCL, induces DAMPs and enhances tumor immunogenicity. Concurrently, the released Mn2+ promotes DCs maturation, bridging innate and adaptive immune responses. In an orthotopic CRC mouse model, oral administration of HML@HN significantly suppressed tumor growth, reshaped the immunosuppressive tumor microenvironment, and elicited robust antitumor immunity with excellent biosafety. Tumor nanovaccines represent a promising cutting-edge technology in the field of immunotherapy, offering innovative strategies to overcome traditional challenges in cancer treatment. Compared with existing clinical treatment strategies, tumor nanovaccines can induce specific immune responses against tumor antigens and exhibit potential efficacy in enhancing the antitumor capacity of the immune system. For the clinical application of the HML@HN nanovaccine developed in this study, the following aspects warrant further investigation. Firstly, the molecular mechanism by which Mn2+ mediates DCs activation in the intestinal microenvironment requires deeper exploration, particularly its interplay with other innate immune pathways. Secondly, combining HML@HN with immune checkpoint inhibitors represents a promising strategy to further enhance therapeutic efficacy. Next, from a translational medicine perspective, issues such as large-scale production, long-term storage stability, and batch-to-batch consistency need to be systematically evaluated. Furthermore, considering the oral administration route, the interaction between the nanovaccine and the gut microbiota is an intriguing research direction that may reveal novel mechanisms underlying its antitumor effects. With continued advances in immunology and materials science, as well as deepening collaboration between these two disciplines, the development of next-generation cancer nanovaccines is expected to be significantly accelerated. Materials and methods Reagent Hexanol, cyclohexane, ammonia solution, tetraethyl orthosilicate (TEOS), potassium permanganate (KMnO4), sodium carbonate (Na2CO3), and ethanol were purchased from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). 3-Aminopropyl triethoxysilane (APTES) was purchased from Sigma Aldrich (USA). Hyaluronic acid (HA) was purchased from Macklin Biochemical Co., Ltd (Shanghai, China). Norepinephrine (NE) was obtained from MedChemExpress LLC (USA). 5-Fluorouracil (5-FU) was purchased from Selleck Chemicals LLC (USA). Fetal bovine serum (FBS) was purchased from PAN-Biotech GmbH (Germany). RPMI 1640 medium was purchased from Gibco Life Sciences (USA). Penicillin-streptomycin-amphotericin B was purchased from M&C Gene Technology Ltd (Beijing, China). Annexin V-FITC/PI Apoptosis Kit and 4’,6-diamidino-2-phenylindole (DAPI) were purchased from Keygen Biotech Co., Ltd (Nanjing, China). Cell Counting Kit-8 (CCK-8) and D-Luciferin potassium salt were purchased from TargetMol Chemicals Inc (USA). Triton X-100 solution (10%, sterile), Actin-Tracker Green, ROS Assay Kit, and Matrix-Gel™ Basement Membrane Matrix were purchased from Beyotime Biotechnology Co., Ltd (Shanghai, China). Flow cytometry antibodies were purchased from BIOLEGEND Biotechnology Co., Ltd (USA). All reagents were analytical purity without further modification. Cell line and animals CT26 cells were obtained from the American Type Culture Collection and cultured in RPMI 1640 with 10% FBS. The morphology of the cells was observed regularly during culture at 37 °C with 5% CO2. BALB/c mice were purchased from the Vital River Laboratories (Beijing, China). In vivo analysis was performed in accordance with the relevant guidelines and was approved by the Ethics Committee of Shandong First Medical University and Shandong Academy of Medical Sciences (W202512290941). Preparation of HML@HN Synthesis of hollow manganese dioxide nanoparticles (HM): Amino-functionalized silica nanoparticles (ASiO2 NPs) were prepared according to a previously reported method [44]. Briefly, n-hexanol (5.4 mL), cyclohexane (22.5 mL), and Triton X-100 (5.3 mL) were added to a 100 mL beaker and stirred for 5 min. Subsequently, deionized water (1 mL) was introduced into the mixed solution. After reaction for 30 min, ammonia solution (0.75 mL) was added dropwise, and the reaction was allowed to proceed for another 30 min. Then, a mixture of APTES (100 μL) and TEOS (500 μL) was added dropwise, and stirring was continued for 24 h ASiO2 NPs were collected by centrifugation and washed with ethanol and water several times. Finally, the obtained ASiO2 NPs were dispersed in deionized water (5 mL), and the concentration was determined for further use. Thereafter, ASiO2 NPs (50 mg) were weighed and redispersed in deionized water (50 mL) containing potassium permanganate (350 mg), followed by stirring for 12 h to allow the reaction to occur. ASiO2 NPs coated with manganese dioxide (MnO2) (ASiO2@MnO2) were collected by centrifugation and washed three times with deionized water. Subsequently, the obtained ASiO2@MnO2 NPs were dispersed in sodium carbonate solution (50 mL, 2 M) under continuous stirring at 60 °C. After stirring for 12 h, hollow MnO2 (HM) was collected by centrifugation at 6000 rpm for 10 min and washed three times with deionized water. Finally, the HM was dispersed in deionized water, and the concentration was determined for subsequent use. Synthesis of HML: To prepare tumor cell lysate (TCL), CT26 cells were pretreated with 5-FU. After cell death was induced 24 h, the supernatant containing cytokines and tumor cell debris was collected, and the tumor cells were further disrupted using a cell ultrasonic homogenizer to obtain TCL containing 5-FU. HM was mixed with TCL, and the loading of cytokines and cell membrane fragments into HM was achieved by sonication. HML was then collected by centrifugation. Synthesis of HML@HN: NE (1 mg) and HA (1 mg) were dissolved in water (1 mL), and the resulting solution was added dropwise to the HML nanosuspension (4 mL, 3 mg HM) under stirring. After stirring for 6 h, HML@HN was collected by centrifugation at 6000 rpm for 10 min and washed twice with PBS. Characterizations The morphology and size of nanoparticles were determined by transmission electron microscopy (TEM) with an accelerating voltage of 120 kV (HITACHI, Japan). Elemental mapping of nanoparticles was obtained using a high-resolution TEM with an accelerating voltage of 200 kV instrument (Thermo Scientific, USA). Surface potential of nanoparticles was measured using a nanoparticle size and zeta potential analyzer of Zetasizer Pro (Malvern Panalytical, U.K.). Quantitative analysis of manganese release from nanoparticles was characterized by an ion chromatography-inductively coupled plasma mass spectrometry (ICP-MS) instrument (Thermo Fisher, USA). Atomic valence state and elemental information were measured by X-ray photoelectron spectrometer (XPS) instrument (Shimadzu, Japan). Cell viability was measured by a microplate reader (Thermo Fisher, USA). Fluorescence imaging of cells was performed via a laser confocal microscope (Olympus, Japan). Analysis of immune cells and apoptosis was evaluated using a flow cytometric analysis (Challenbio, China). Mouse imaging was achieved using a three-dimensional live imaging system for small animals (PerkinElmer, USA). Stability and release of HML@HN in the simulated gastrointestinal tract To evaluate stability, HML@HN was dispersed in simulated gastric fluid or simulated intestinal fluid, and the morphology of the nanomaterials was observed by TEM on day 7. Briefly, HML@HN (1 mL, 1 mg/mL) was placed into a dialysis bag (molecular weight cutoff of 3 kDa), which was then immersed in simulated gastric fluid (50 mL) or simulated intestinal fluid (50 mL) and placed on a shaker (37 °C, 100 rpm) with gentle agitation. At 0, 1, 2, 4, 6, 8, 10, 12, 24, 48, and 72 h, 1 mL of solution was withdrawn, and fresh simulated gastric or intestinal fluid (1 mL) was replenished to maintain a constant volume of the dissolution medium. The obtained samples were nitrified with nitric acid, and the concentration of Mn ions was determined using an ICP-MS instrument, allowing the calculation of Mn ion release. Cellular uptake analysis Cellular uptake of HML@HN by the mouse colon cancer cell line (CT26) was observed using confocal laser scanning microscopy (CLSM) and further quantitatively analyzed by flow cytometry. First, cells were seeded into each well of 12-well plates at a density of 8 × 104 cells per well and incubated overnight on sterile round glass coverslips to allow attachment. Then, Cy3-labeled HML@NE and HML@HN were added to each well, respectively, and incubated for 6 h. Subsequently, the cytoskeleton and nuclei were stained with Actin-Tracker Green-488 and DAPI for visualization. For quantitative analysis by flow cytometry, cells were seeded into 6-well plates at a density of 5 × 105 cells per well and incubated overnight. Cy3-labeled HML@NE and HML@HN were added to each well, respectively, and incubated for 1, 3, and 6 h. Cells without any treatment were used as a negative control. Then, cells from each treatment group were collected and quantified by flow cytometry analysis. Intracellular ROS detection CT26 cells were seeded in confocal dishes (1 × 105 cells per dish) for 24 h of incubation. The cells were treated with different groups for 6 h: PBS (control), HM, HML, HML@NE, and HML@HN (50 μg/mL). After discarding the treatment solution, the cells were incubated with DCFH-DA for 20 min at 37 °C. Then, the treated cells were gently rinsed with PBS three times. Finally, the ROS fluorescence was recorded by a fluorescence microscope. Cell viability assay CT26 cells were seeded into a 96-well plate (5 × 103 cells per well) and incubated for 12 h to allow cell adhesion. The supernatant was then discarded, and fresh culture medium containing different concentrations of various nanomedicines (HM, HML, HML@NE, and HML@HN) was added. The PBS treatment served as the control group. After 24 h of incubation, the supernatant was removed, and the cells were washed with PBS buffer. Subsequently, the diluted CCK-8 solution was added and thoroughly mixed, followed by incubation in a 37 °C incubator for 2 h. Finally, the absorbance at 450 nm was measured using a microplate reader to calculate cell viability. CRT detection CT26 cells were seeded into confocal dishes (1 × 105 cells per dish) and incubated for 12 h. Firstly, the cells were treated with PBS (Control), HM, HML, HML@NE, and HML@HN for 12 h, respectively. Subsequently, CT26 cells were fixed with paraformaldehyde for 30 min, permeabilized with 0.5% Triton X-100 for 10 min, and then blocked with 5% BSA (prepared in 0.5% Triton X-100) for primary antibody (anti-CRT) dilution, followed by incubation at 4 °C overnight. Then, CT26 cells were incubated with diluted secondary antibody at 37 °C for 1 h. The nuclei were stained with DAPI at room temperature for 30 min. After each step, the cells were washed twice with PBS. Finally, the CRT fluorescence was captured using a confocal imaging system. Cell apoptosis analysis CT26 cells were seeded in a 6-well plate (1 × 105 cells per well) for 24 h incubation. The cells were then treated for 12 h as follows: PBS (Control), HM, HML, HML@NE, and HML@HN (50 μg/mL). Subsequently, the cells were collected and stained with a mixture of PI and Annexin V-FITC for 10 min at room temperature in the dark. Finally, the cells were analyzed by flow cytometry. DCs maturation and activation Firstly, the femurs and tibias were aseptically isolated from the mouse, and then flushed the bone marrow cavity to obtain bone marrow cavity cells. Secondly, the bone marrow cavity cells were subjected to lysis of red blood cells, and a single-cell suspension was prepared. The cells were cultured in RPMI 1640 medium containing 10% FBS and were supplemented with 20 ng/mL GM-CSF and 10 ng/mL IL-4 for induction. The culture medium was performed with half-medium changes and replenished with cytokines on days 3 and 5-6 of culture. Subsequently, the immature DCswere collected on days 6-7 to build a Transwell model with CT26 cells. CT26 cells in the upper chambers were treated with different groups: PBS (control), HM, HML, HML@NE, and HML@HN (50 μg/mL). After 12 h, the DCs in the lower chamber were harvested and stained with antibodies labeled with fluorescence against CD11c (APC), CD80 (FITC), and CD86 (PE). Finally, the maturation and activation of DCs were analyzed using flow cytometry. In vivo safety study The in vivo safety of HML@HN was evaluated in healthy mice (n = 3). The formulation was administered orally every other day (5 mg/kg) , which was consistent with the doses in the anti-tumor efficacy analysis. After the final administration, whole blood was collected for hematological analysis, and serum samples were collected to assess liver function (alanine aminotransferase (ALT) and alkaline phosphatase (ALP)) and renal function (blood urea nitrogen (BUN) and creatinine (CR)). In addition, major organs, including the heart, liver, spleen, lungs, kidneys, stomach, small intestine, and cecum, were collected for hematoxylin and eosin (H&E) staining to evaluate potential tissue damage. Healthy mice served as the control group. In vivo biodistribution of HML@HN A CT26-luc orthotopic colon tumor mouse model was established to evaluate tumor progression using an IVIS Spectrum CT small animal in vivo three-dimensional imaging system. Briefly, female BALB/c mice (5 weeks old) were anesthetized with isoflurane and placed in a sterile environment. A small incision was made in the abdomen to expose the cecum, and a suspension of CT26-luc cells mixed with Matrigel at a 1:1 ratio (1 × 106 cells per mouse) was injected into the cecal wall using a 31-gauge microsyringe. Mice were maintained under isoflurane anesthesia throughout the procedure. Five days after CT26-luc cell inoculation, D-Luciferin potassium salt was administered intraperitoneally, and colorectal tumor growth was assessed using the IVIS imaging system. The intestinal accumulation ability of DiR, HML@NE, and HML@HN was observed using the IVIS imaging system. After successful tumor establishment was confirmed by in vivo imaging, free DiR, DiR-labeled HML@NE, and DiR-labeled HML@HN were orally administered to colon tumor-bearing mice at a dose of DiR(0.5 mg/kg). At various time points, mice were anesthetized with isoflurane, and the red fluorescent signal in each group was recorded using the IVIS imaging system. 24 h after oral administration, mice were dissected, and major organs, including the heart, liver, spleen, lungs, kidneys, and tumor, were collected for imaging. In vivo antitumor efficacy analysis Mice bearing luciferase-expressing CT26-luc orthotopic colon tumors were randomly divided into five groups (n = 6 per group). Groups 1 to 5 were orally administered PBS, HM, HML, HML@NE, or HML@HN (5 mg/kg), respectively, on days 1, 3, 5, 7, and 9, with a total treatment period of 15 days. Body weight was recorded every three days, and tumor growth was assessed every six days using an in vivo imaging system. D-Luciferin potassium salt (150 mg/kg) was intraperitoneally injected into the mice, and bioluminescence imaging was performed using the in vivo imaging system 10 min after injection to evaluate tumor progression. On day 15, the mice were euthanized, and the tumors were collected and rapidly stored in 4% paraformaldehyde solution for subsequent hematoxylin and eosin (H&E) staining and terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) assays. In vivo immune response evaluation Firstly, the tumor tissues were collected and processed through grinding, filtration, red blood cell lysis, and sieving to obtain a single-cell suspension. To detect mature DCs, the single-cell suspension was incubated with APC anti-mouse CD11c, PE anti-mouse CD86, and FITC anti-mouse CD80 for staining and then quantitatively analyzed by flow cytometry. Subsequently, for phenotypic analysis of macrophages, the single-cell suspension was stained with APC anti-mouse F4/80, PE anti-mouse CD86, and FITC anti-mouse CD206. M1 and M2 macrophages were analyzed by flow cytometry. T cells were stained with FITC anti-mouse CD45, Brilliant Violet 510™ anti-mouse CD3, Brilliant Violet 421™ anti-mouse CD4, and PE/Cyanine7 anti-mouse CD8a for 30 min and analyzed by flow cytometry. Similarly, CD4 and CD8 in the spleen were performed in the same manner as in tumors. Additionally, for Treg cell analysis, the single-cell suspension was fixed and permeabilized, followed by staining with FITC anti-mouse CD45, Brilliant Violet 510™ anti-mouse CD3, Brilliant Violet 421™ anti-mouse CD4, PE anti-mouse CD25, and APC FOXP3, and analyzed by flow cytometry. All the fluorescence staining was carried out under light-proof conditions. Furthermore, other mice were euthanized after the different treatments. Tumors were collected for immunohistochemical staining of CD4 and CD8. Statistical analysis Statistical analyses and graphical representations were carried out primarily using GraphPad Prism 8 (GraphPad, MA, USA). All statistical analyses were performed using a one-way ANOVA for comparison of multiple groups. Data are expressed as mean ± standard deviation (SD). A p-value of less than 0.05 was considered statistically significant. Statistical significance is denoted as ∗p< 0.05, ∗∗p< 0.01, ∗∗∗p< 0.001, and ∗∗∗∗p < 0.0001. CRediT authorship contribution statement Qian Ming: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – original draft. Junjie Lin: Data curation, Formal analysis, Methodology, Validation, Writing – original draft. Yingying Zhou: Methodology, Validation. Ziqi Song: Investigation, Software. Zihan Yu: Data curation, Methodology. Xinze Liu: Methodology, Visualization. Xiao Sun: Data curation, Software. Lianyu Zhao: Formal analysis, Validation. Changyu Lu: Methodology, Software. Natong Liu: Formal analysis, Software. Yanan Qin: Investigation, Methodology. Jing Wu: Formal analysis, Funding acquisition, Supervision, Writing – original draft. Zhaowei Li: Formal analysis, Supervision, Writing – review & editing. Fengying Shao: Conceptualization, Funding acquisition, Project administration, Supervision, Writing – review & editing. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.