EGFR-antibody-decorated cancer-cell-membrane-coated MOF nanoplatform enables targeted delivery of the STING agonist diABZI for pancreatic cancer immunotherapy

Abstract

Pancreatic ductal adenocarcinoma (PDAC) has limited immunotherapeutic effect because of poor drug penetration and tumor immunosuppression. Activation of the stimulator of interferon genes (STING) pathway may promote antitumor immunity, but systemic administration of STING agonists is constrained by inadequate tumor accumulation. Epidermal growth factor receptor (EGFR) is frequently overexpressed in PDAC and is closely associated with tumor growth, invasion, and treatment resistance. Herein, we developed an EGFR-antibody-decorated cancer-cell-membrane-coated metal–organic framework (MOF) nanoplatform loaded with the STING agonist diamidobenzimidazole (diABZI) (EM@MOF/diABZI) for targeted pancreatic cancer immunotherapy. This nanoplatform consists of an acid-responsive diABZI loaded MOF core and an EGFR-antibody-functionalized pancreatic cancer-cell-membrane shell. The nanoplatform displayed preserved membrane-associated proteins and EGFR-binding activity, pH-dependent diABZI release, significantly enhanced cellular uptake and preferential tumor accumulation. In a subcutaneous Pan02 pancreatic cancer mouse model, EM@MOF/diABZI produced greater tumor-growth inhibition and improved survival. Mechanistically, targeted diABZI delivery increased intratumoral p-STING expression and elevated the levels of IFN-β, CXCL10, TNF-α, IL-6, and IFN-γ, accompanied by enhanced infiltration of CD4⁺ and CD8⁺ T cells. No obvious histopathological abnormalities or substantial systemic toxicity were observed. Therefore, this nanoplatform provides a promising strategy for enhancing the therapeutic effect of STING agonists for pancreatic cancer immunotherapy.

Graphical abstract

Highlights

• •EGFR-antibody-decorated cancer-cell-membrane-coated MOF delivers diABZI to PDAC.
• •Biomimetic nanoplatform enables pH-responsive drug release and enhances tumor accumulation.
• •EM@MOF/diABZI activates STING signaling and cytokine production.
• •Targeted diABZI delivery increases CD4⁺ and CD8⁺ T-cell infiltration.
• •Nanoplatform suppresses tumor growth and improves survival with low toxicity.

1. Introduction

Pancreatic ductal adenocarcinoma (PDAC) is one of the most aggressive and therapeutically challenging malignancies (Thiel et al., 2025; Lin et al., 2024). PDAC is commonly diagnosed at an advanced stage and is characterized by rapid progression, metastasis, and therapeutic resistance (Park et al., 2021). Despite advances in surgery, chemotherapy, molecular targeting, and immunotherapy, clinical outcomes of PDAC remain poor (Lin et al., 2024; Sardar et al., 2022). The tumor microenvironment (TME), characterized by dense desmoplastic stroma, poor vascular perfusion, and extensive immunosuppression, is a major barrier to effective immunotherapy (Hilmi et al., 2023). These pathological features not only impair immune-cell infiltration but also markedly restrict the drug accumulation and cellular uptake by PDAC. Epidermal growth factor receptor (EGFR) is frequently overexpressed in PDAC and is closely associated with tumor growth, invasion, and treatment resistance (Nussinov et al., 2025; Zhang et al., 2026). Importantly, EGFR is a cell-surface receptor with an accessible extracellular domain, making it an attractive target for ligand-mediated drug delivery rather than solely for receptor-inhibition therapy (Fang et al., 2026). Therefore, a therapeutic strategy that simultaneously improves tumor-selective drug delivery by EGFR-targeting and reshapes the immunosuppressive microenvironment is urgently needed (Kuznetsova et al., 2023).

The stimulator of interferon genes (STING) pathway has emerged as an attractive target for cancer immunotherapy (Zhu et al., 2025; Kimmel et al., 2025). In PDAC, STING activation may be particularly valuable because it has the potential to overcome insufficient antigen presentation and limited cytotoxic T-cell infiltration (Luo et al., 2025; Chong et al., 2025). STING functions as a central innate immune adaptor that links cytosolic DNA sensing to the activation of TBK1 and IRF3, leading to the production of type I interferons and pro-inflammatory chemokines (Samson and Ablasser, 2022). These signals promote dendritic-cell maturation, antigen presentation, CD8⁺ T-cell priming, and recruitment of effector immune cells into the tumor microenvironment (Huang et al., 2025). Pharmacological activation of STING therefore offers a promising approach for converting immunologically “cold” tumors into inflamed tumors with improved immune responsiveness (Yang et al., 2024a). Among the available STING agonists, diamidobenzimidazole (diABZI) is a potent non-nucleotide small-molecule agonist with systemic activity (Ramanjulu et al., 2018). Unlike cyclic dinucleotide agonists, diABZI displays improved drug-like properties and has demonstrated the ability to activate STING signaling and induce antitumor immune responses (Zhang et al., 2023a). However, systemic administration of free STING agonists is still limited by rapid clearance, nonspecific biodistribution, and insufficient tumor accumulation. In particular, off-target activation of STING in peripheral immune cells may increase the risk of excessive cytokine production and systemic immune-related adverse effects. Therefore, the development of a tumor-targeted delivery system capable of increasing local diABZI exposure while reducing nonspecific systemic activation represents an important pharmaceutical strategy.

Nanomedicine has emerged as a promising strategy to improve drug circulation stability and tumor accumulation (Yang et al., 2021). Several nanodelivery strategies have been developed to improve the therapeutic performance of STING agonists. Luo et al. reported a STING-activating nanovaccine based on a polymeric nanoparticle physically mixed with antigen, which generated strong cytotoxic T-cell responses with relatively low systemic cytokine expression (Luo et al., 2017). More recently, Sheehy et al. developed a STING-activating polymer–drug conjugate platform. A diABZI prodrug was conjugated to hydrophilic polymer chains through a cathepsin B-responsive linker to prolong circulation and enable passive tumor accumulation (Sheehy et al., 2024). Zhang et al. prepared diABZI-encapsulated liposomes, and the liposomal formulation showed stronger tumor-growth inhibition than free diABZI (Zhang et al., 2023b). Among various nanocarriers, metal–organic frameworks (MOFs) have attracted increasing interest because of their high surface area, tunable pore structure, and adjustable particle size (Li et al., 2023). For example, Nie et al. reported MOF-coated MnO₂ nanosheets as a nanoplatform for anticancer drug delivery and combinatorial cancer treatment, highlighting the value of MOF-based architectures in improving therapeutic delivery and integrating multiple treatment modalities (Nie et al., 2020). In particular, acid-responsive MOFs can undergo structural degradation and drug release in the mildly acidic tumor microenvironment (Yao et al., 2022). However, bare MOF nanoparticles still exhibit limited tumor selectivity and rapid clearance by the mononuclear phagocyte system and nonspecific uptake by normal tissues.

Biomimetic and receptor-targeted nanoplatforms have been developed to improve tumor selectivity. For pancreatic cancer targeting, cetuximab/EGFR-mediated nanoparticle delivery has been investigated (Patra et al., 2008). Biomimetic cancer-cell-membrane coating provides a potential solution to these limitations (Yang et al., 2024b). Cancer-cell membranes preserve multiple membrane-associated proteins, adhesion molecules, lipids, and tumor-associated antigens derived from the source cells, which may facilitate homotypic recognition and preferential interaction with homologous tumor cells (Pereira-Silva et al., 2020). Fang et al. demonstrated that cancer-cell-membrane-coated PLGA nanoparticles could retain membrane-associated tumor antigens and enable homotypic binding for anticancer vaccination and drug delivery (Fang et al., 2014). Zhu et al. further showed that nanoparticles coated with homologous cancer-cell membranes exhibited preferential self-recognition and selective homing to homologous tumors (Zhu et al., 2016). To further enhance tumor-cell-specific recognition, antibody-mediated active targeting can be introduced onto the outer membrane surface. Monoclonal antibodies have emerged as important components of cancer immunotherapy owing to their high specificity and ability to recognize tumor-associated antigens. Nevertheless, their therapeutic application requires consideration of potential safety issues associated with systemic exposure and immune modulation. Recent studies have summarized both the therapeutic advantages and safety considerations of monoclonal antibody-based cancer therapies (Zhang et al., 2025). In the present study, anti-EGFR monoclonal antibody is not intended as an independent therapeutic agent but is used as a targeting ligand to enhance EGFR-mediated tumor-cell recognition and improve local delivery of the STING agonist diABZI. In this study, we designed an EGFR-antibody-decorated cancer-cell-membrane-coated MOF nanoplatform (EM@MOF/diABZI) for targeted delivery of diABZI in PDAC (Fig. 1A). The nanoplatform integrates two functional components: a diABZI loaded acid-responsive MOF core (MOF/diABZI) for drug loading and controlled release, and an EGFR-antibody-decorated pancreatic cancer-cell membrane shell (EM) for biomimetic tumor-associated interactions and active receptor-mediated targeting. The EM@MOF/diABZI nanoplatform could enhance the selective accumulation and cellular uptake of diABZI in EGFR-expressing pancreatic cancer cells through the combined effects of cancer-cell-membrane-mediated homotypic recognition and EGFR-antibody-mediated active targeting (Fig. 1B). This biomimetic delivery strategy may subsequently enhance STING activation, reshape the immunosuppressive tumor microenvironment, and improve antitumor immune responses in PDAC (Fig. 1C). Most reported systems emphasize one major delivery strategy, such as MOF-mediated immune activation, homotypic membrane coating, or antibody-mediated targeting. The present EM@MOF/diABZI platform integrates three delivery functions within a single pancreatic cancer immunotherapy system: an acid-responsive MOF core for diABZI loading and pH-enhanced release, a pancreatic cancer-cell-membrane shell for biomimetic tumor-associated recognition, and EGFR-antibody decoration for active targeting. This integrated design is intended to increase local diABZI accumulation and STING activation in pancreatic tumors.

Fig. 1.

Scheme for EGFR-antibody-decorated cancer-cell-membrane-coated MOF nanoplatform enabling targeted delivery of the STING agonist diABZI for pancreatic cancer immunotherapy.

2. Materials and methods

2.1. Synthesis of MOF and MOF/diABZI

2.64 g of Cu(NO₃)₂·3H₂O and 0.47 g of (CH₃COO)₂Mn·4H₂O were dissolved in separate beakers containing 30 mL of deionized water. In parallel, 1.26 g of 1,3,5-benzenetricarboxylic acid (H₃BTC) was dissolved in 30 mL of anhydrous ethanol to form the linker solution. The metal precursor solution and linker solution were then combined. After transferring into a Teflon-lined autoclave, the reaction was maintained at 120 °C for 12 h by hydrothermal reaction (Eldin et al., 2025). The products were washed 3 times with deionized water and anhydrous ethanol, and dried at 100 °C for 12 h. To obtain MOF/diABZI, the MOF was dispersed in the diABZI (HY-112921A, MedChemExpress, China) solution, and the mixture was stirred at room temperature for 24 h. By centrifugation, the product was collected with the supernatant carefully discarded.

2.2. Synthesis of EM@MOF and EM@MOF/diABZI

Cell membrane was isolated from tumor cell BxPC-3 (for in vitro cell experiments using human cell lines) and Pan02 (for in vivo mouse model), respectively. The isolation of cell membrane was performed, and collected by centrifuging the supernatants (20,000 g, 30 min) (Cai et al., 2021a). 1 mM 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide and N-hydroxysuccinimide were added to Amine-PEG-Thiol (SH-PEG-NH₂, MW 5000) solution and incubated for 30 min. To synthesize SH-PEG-anti-EGFR, the anti-EGFR (BE0466, Bio X Cell, NH, USA), was added and incubated for 30 min. BE0466 was selected because it recognizes domain III of both murine and human EGFR. Then, it was incubated with the cell membrane for 12 h at 4 °C to obtain EM. Finally, the prepared EM was mixed with MOF or MOF/diABZI by sonication and extruding through polycarbonate membranes. The resulting EM@MOF or EM@ MOF/diABZI were collected by centrifugation and washed with phosphate-buffered saline (PBS).

2.3. Characterization of nanoplatform

The particle morphology and size distribution were observed by transmission electron microscopy (TEM; Tecnai G220 S-TWIN, FEI, USA) and scanning electron microscopy (SEM; Tecnai F20 S-TWIN, FEI, USA). The size distribution was analyzed by SEM. Dynamic light scattering (DLS) was employed to further determine the hydrodynamic diameter and polydispersity index (PDI). X-ray diffraction (XRD) was used to analyze the crystalline structure of MOF. The colloidal stability of EM@ MOF/diABZI under different biological conditions was evaluated.

2.4. Characterization of biomimetic membrane functionalization

For protein analysis of the cancer-cell-membrane coating, sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) was performed using an electrophoresis system (Bio-Rad Laboratories, Hercules, USA). Preservation of the cancer-cell-membrane marker protein E-cadherin in EM@MOF/diABZI was confirmed by Western blot. The EGFR-binding activity of the anti-EGFR antibody displayed on EM@MOF/diABZI was evaluated using an EGFR competitive-binding enzyme-linked immunosorbent assay (ELISA). Recombinant human EGFR extracellular-domain protein (11302-ER, R&D Systems, Minneapolis, USA) was immobilized on high-binding 96-well plates and blocked with bovine serum albumin. Fluorescently labeled EM or EM@MOF/diABZI was then incubated in the EGFR-coated wells. After washing, the retained signal was quantified using a microplate reader.

2.5. Loading capacity and encapsulation efficiency of STING agonist

The amount of diABZI encapsulated in the EM@MOF/diABZI was quantified using UV–Vis spectroscopy (Lambda 25, PerkinElmer, USA). A standard calibration curve of diABZI was generated by measuring the characteristic wavelength absorbance at 325 nm. The loading capacity (LC) and encapsulation efficiency (EE) were calculated based on the following equation: LC (%) = (Amount of loaded diABZI) / (Total amount of nanoparticles) × 100%; EE (%) = (Amount of encapsulated diABZI) / (Total amount of diABZI) × 100%.

2.6. STING agonist release

The concentration of released diABZI was measured using UV–Vis spectroscopy at the characteristic wavelength of 325 nm. The released amount of diABZI was quantified using a standard calibration curve. To simulate the acidic tumor microenvironment, the release of diABZI from EM@MOF/diABZI was also evaluated at pH values ranging from 5.5, 6.5, and 7.4, simulating physiological and tumor microenvironment conditions.

2.7. Cell culture

Human pancreatic cancer BxPC-3 cells and murine pancreatic cancer Pan02 cells were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin solution. Human pancreatic duct epithelial HPDE6-C7 cells were cultured in keratinocyte serum-free medium supplemented with the recommended growth supplements. All cell lines were maintained at 37 °C in a humidified incubator containing 5% CO₂.

2.8. In vitro cell viability

The cytotoxic effects of the indicated formulations were assessed using a Cell Counting Kit-8 assay (CCK-8; Dojindo Molecular Technologies, Japan). Cells were plated in 96-well culture plates at 1 × 10⁴ cells/well and allowed to attach overnight. The culture medium was then replaced with fresh medium containing the respective formulations at matched diABZI-equivalent concentrations (the total amount of diABZI administered in each diABZI-containing group). Because MOF, EM@MOF, MOF/diABZI, and EM@MOF/diABZI differ in drug content and membrane composition, all comparisons were normalized. For free diABZI, the diABZI-equivalent concentration refers to the molar concentration of directly added diABZI. For MOF/diABZI and EM@MOF/diABZI, the diABZI-equivalent concentration refers to the total nanoparticle-associated diABZI concentration calculated from the experimentally determined drug-loading content. The nanoparticle mass concentration was adjusted to ensure that all diABZI-containing groups received the same total amount of diABZI. Blank MOF and EM@MOF groups were administered at carrier concentrations corresponding to the respective diABZI-loaded formulations. This normalization strategy ensured that the observed differences reflected the contribution of diABZI loading and membrane-mediated targeting rather than unequal formulation dosing. Following a 48-h exposure period, 10 μL of CCK-8 solution was added, and plates were maintained for 2 h. Absorbance at 450 nm was measured with a microplate reader (BioTek, USA). For concentration-dependent studies, viability data were fitted using nonlinear regression to establish dose–response relationships and determine the half-maximal inhibitory concentration (IC₅₀). On the basis of this analysis, a diABZI-equivalent concentration of 16.3 μM was selected for subsequent in vitro experiments. For free diABZI, the diABZI-equivalent concentration refers to the molar concentration of directly added diABZI. For MOF/diABZI and EM@MOF/diABZI, the diABZI-equivalent concentration refers to the total nanoparticle-associated diABZI concentration calculated from the experimentally determined drug-loading content. The nanoparticle mass concentration was adjusted to ensure that free diABZI, MOF/diABZI, and EM@MOF/diABZI groups contained the same total amount of diABZI. Blank MOF and EM@MOF groups were added at nanoparticle mass concentrations corresponding to those of the diABZI-loaded formulations.

2.9. Live/dead staining

BxPC-3 cells were seeded in 24-well plates and cultured overnight to allow cell attachment. The cells were then treated with PBS, MOF, EM@MOF, MOF/diABZI, or EM@MOF/diABZI. After 48 h of incubation, the culture medium was removed, and the cells were gently washed three times with PBS. Then, the cells were incubated with calcein-AM and propidium iodide (PI) staining. After 30 min, the cells were washed with PBS and observed using a fluorescence microscope. Representative fluorescence images were captured from randomly selected fields, and the percentage of dead cells was quantified using ImageJ software.

2.10. Mn²⁺ measurement and cellular uptake

For intracellular Mn²⁺ quantification, cells were treated with MOF/diABZI or EM@MOF/diABZI for 24 h. The Mn²⁺ content in the cell lysates was measured by inductively coupled plasma atomic emission spectroscopy (ICP-AES) (Li et al., 2024). To evaluate cellular uptake, MOF/diABZI or EM@MOF/diABZI were labeled with Nile Red (Thermo Fisher Scientific), and subsequently incubated with BxPC-3 cells and HPDE6-C7 cells for 4 h (Liu et al., 2025). After incubation, the cells were fixed with 4% paraformaldehyde for 15 min, washed with PBS, and observed using a fluorescence microscope (Nikon Eclipse Ti—S, Japan).

2.11. Antitumor efficacy in a mouse model

All animal procedures were approved by the Institutional Animal Care and Use Committee of Jilin university first hospital and were performed in accordance with relevant institutional guidelines. Female C57BL/6 mice, aged 8 weeks, were obtained from Charles River company, Beijing, China and maintained under specific-pathogen-free conditions. To establish a subcutaneous pancreatic cancer model, Pan02 cells were harvested and suspended in sterile PBS, and injected subcutaneously into the right flank of each mouse at a density of 1 × 10⁶ cells per mouse (Chen et al., 2023). When the tumor volume reached approximately 100 mm³, mice were randomly assigned to five groups (n = 5 per group): PBS, MOF, EM@MOF, MOF/diABZI, and EM@MOF/diABZI. The corresponding formulations were administered via intravenous injection with an equivalent delivered dose of diABZI (0.9 mg/kg) at a volume of 100 μL and treatment schedule. To evaluate tumor-targeting ability, Cy5.5-labeled EM@MOF/diABZI was intravenously injected into tumor-bearing mice, and fluorescence images were collected at predetermined time points. Tumor volume was calculated as: volume = (length × width²)/2. Body weight was recorded throughout the treatment period. At the experimental endpoint, mice were euthanized, and tumors and major organs were collected for tumor-weight determination, histological examination, biodistribution analysis, and immune-related assays. For survival analysis, a separate cohort of tumor-bearing mice received the same treatment regimen and was monitored for up to 60 days.

2.12. Hematoxylin and eosin staining

Tumor tissues and major organs, including the heart, liver, spleen, lungs, and kidneys, were collected from mice at the experimental endpoint and fixed in 4% paraformaldehyde. After fixation, tissues were dehydrated through a graded ethanol series, embedded in paraffin, and sectioned at a thickness of 5 μm. The sections were subsequently deparaffinized, rehydrated, and stained with hematoxylin and eosin (H&E) according to standard histological procedures (Cai et al., 2021b). Following staining, tissue sections were dehydrated, cleared, mounted, and examined under a light microscope.

2.13. Biodistribution

For quantitative determination of diABZI distribution, mice were sacrificed at 24 h after injection, and major organs and tumor tissues were collected and weighed. The concentration of diABZI in each tissue sample was quantified and normalized to tissue weight. To assess the time-dependent tumor accumulation of diABZI, additional tumor-bearing mice were treated with MOF/diABZI or EM@MOF/diABZI and sacrificed at 0.5, 1, 2, 4, 12, and 24 h after injection. Tumor tissues were harvested and processed for diABZI concentration quantification.

2.14. Immunofluorescence analysis

Immunofluorescence staining was conducted to assess phosphorylated STING (p-STING) expression in tumor tissues. Paraffin-embedded tumor sections were deparaffinized in xylene, rehydrated through a graded ethanol series, and rinsed with PBS. Heat-mediated antigen retrieval was performed in citrate buffer. Sections were blocked with 5% bovine serum albumin for 30 min at room temperature, and incubated overnight at 4 °C with an anti-p-STING primary antibody at 1:50 (62912, Cell Signaling Technology, Danvers, USA). After washing with PBS, sections were treated with secondary antibody. The stained sections were imaged using a fluorescence microscope (Nikon, Japan).

2.15. Cytokine quantification by ELISA

The concentrations of IFN-β, CXCL10, TNF-α, IL-6, and IFN-γ were determined by enzyme-linked immunosorbent assay (ELISA) using commercially available mouse ELISA kits (Invitrogen, USA), following the manufacturers' instructions. Briefly, standards and samples were added to capture-antibody-coated 96-well plates and incubated according to the kit protocol. After washing, detection antibody and enzyme conjugate were added sequentially. Color development was initiated by addition of the substrate solution and terminated using stop solution. Absorbance was measured at 450 nm with a microplate reader (BioTek, USA). Cytokine concentrations were calculated from standard curves generated for each assay.

2.16. Flow-cytometric analysis of tumor-infiltrating T cells

Fresh tumor tissues were mechanically dissociated and enzymatically digested to prepare single-cell suspensions. The cell suspensions were filtered through a 70-μm cell strainer, washed with PBS, and resuspended in staining buffer. Cells were stained with fluorophore-labeled antibodies, including anti-CD45-PC7, anti-CD3-PE, anti-CD4-FITC, and anti-CD8-APC antibodies (1200, BioLegend, USA) for 30 min at 4 °C in the dark. After staining, cells were washed with PBS, and analyzed using a flow cytometer with FlowJo software (Tree Star, USA) used for data processing. For analysis, debris was excluded on the basis of forward- and side-scatter characteristics, followed by removal of doublets using FSC-A versus FSC-H gating. Leukocytes were identified as CD45⁺ events, and CD3⁺ T cells were subsequently gated within the CD45⁺ population. CD4⁺ and CD8⁺ T-cell subsets were defined as CD45⁺CD3⁺CD4⁺ and CD45⁺CD3⁺CD8⁺ cells, respectively. Tumor-associated macrophages were identified as CD45⁺CD11b⁺F4/80⁺ cells, M1-like macrophages were defined as CD45⁺CD11b⁺F4/80⁺CD86⁺ cells, and M2-like macrophages were defined as CD45⁺CD11b⁺F4/80⁺CD206⁺ cells. Regulatory T cells (Tregs) were defined as CD45⁺CD3⁺CD4⁺CD25⁺Foxp3⁺ cells after surface staining, fixation, permeabilization, and intracellular Foxp3 staining. Gates were set using the appropriate control samples.

2.17. Statistical analysis

All quantitative data are expressed as the mean ± standard deviation based on at least three independent experiments. Statistical analysis was conducted using GraphPad Prism 8.0. For comparisons between two groups, an unpaired two-tailed Student's t-test was applied. Differences among multiple groups were analyzed using one-way analysis of variance (ANOVA), followed by Tukey's post hoc test. Survival outcomes were analyzed using the Kaplan–Meier method, and between-group differences were assessed with the log-rank (Mantel–Cox) test. A two-tailed P value less than 0.05 was considered statistically significant.

3. Results and discussion

3.1. Preparation and characterization of EM@MOF/diABZI

The preparation of the EGFR-antibody-decorated cancer-cell-membrane-coated MOF nanoplatform (EM@MOF/diABZI) is illustrated in Fig. 2A. As shown in Fig. 2B–C, the morphology of the nanoparticle was examined by transmission electron microscopy (TEM) and scanning electron microscopy (SEM) with a well-defined nanoscale structure, and an intact and uniform particulate morphology. The MOF core exhibited a non-spherical and partially irregular crystalline morphology, which is typical for this class of MOF particles. Size-distribution profile showed average diameter of 109.1 nm for EM@MOF/diABZI (Fig. 2D). The relatively uniform size distribution of EM@MOF/diABZI suggested that the membrane-coating process did not induce substantial nanoparticle aggregation. The powder X-ray diffraction (XRD) pattern was shown in Fig. S1. As shown in Fig. 2E, the particle size of EM@MOF/diABZI increased after membrane coating, due to the additional outer membrane layer on the MOF/diABZI surface. A nanoscale size range of 100–110 nm is favorable for cellular internalization and may support tumor accumulation in subsequent in vitro and in vivo studies. Surface-charge measurements provided additional evidence for membrane coating. The zeta potentials of EM, MOF/diABZI, and EM@MOF/diABZI were approximately −18 mV, −10 mV, and −15 mV, respectively (Fig. 2F). The zeta potential of EM@MOF/diABZI shifted from that of the MOF/diABZI core toward that of the membrane component, indicating that the outer surface characteristics of the final formulation were dominated by the cancer-cell-membrane shell. The moderately negative surface charge may also reduce nonspecific interactions with serum proteins and contribute to colloidal stability under physiological conditions. Elemental mapping analysis further characterized the composition of EM@MOF/diABZI. Carbon, nitrogen, and oxygen signals were broadly distributed throughout the nanoparticles, while Mn and Cu were also uniformly detected (Fig. 2G). The homogeneous distribution of Mn and Cu confirmed the preservation of the MOF framework after drug loading and membrane coating. The retention of membrane-associated proteins was assessed by SDS–PAGE. As shown in Fig. 2H, EM@MOF/diABZI exhibited protein-band patterns resembling those of the whole-cell lysate and isolated EM fraction, whereas negligible protein signals were observed for MOF/diABZI. EM@MOF/diABZI exhibited high coating efficiency of 89.3% (Fig. S2). These results indicated that the membrane-coating procedure preserved a substantial proportion of the source-cell membrane proteins. Western blotting further confirmed the presence of E-cadherin, a representative membrane-associated protein, in whole-cell lysate, isolated cell membrane, and EM@MOF/diABZI, while the MOF/diABZI group showed no detectable E-cadherin signal (Fig. 2I). The preservation of E-cadherin supported the successful transfer of membrane components onto the MOF surface and indicated that the biomimetic shell retained key cellular membrane features. Finally, the colloidal stability of EM@MOF/diABZI was evaluated by monitoring the polydispersity index (PDI) in PBS, serum-containing medium, and cell-culture medium (Fig. 2J). The PDI remained stable without an obvious increase under all tested conditions, indicating that EM@MOF/diABZI maintained satisfactory dispersion stability in physiologically relevant environments. Although the MOF core exhibits an irregular morphology, the flexible cancer-cell membrane vesicles can deform and conformally wrap around the nanoparticle surface during extrusion-based coating. Successful membrane functionalization was supported by increased particle size after coating, surface charge transition, preservation of membrane-associated proteins, quantitative membrane-coating efficiency, and maintained EGFR-binding activity. These results indicate that EM@MOF/diABZI possesses a stable and functional biomimetic membrane interface rather than requiring a geometrically spherical core structure. Collectively, these results demonstrate the successful construction of a stable EM coated MOF nanoplatform with preserved cancer-cell-membrane proteins and physicochemical properties suitable for subsequent targeted-delivery and antitumor-immunotherapy evaluations.

Fig. 2.

Preparation and characterization of EM@MOF/diABZI. (A) Schematic illustration of the preparation of EGFR-antibody-decorated cancer-cell-membrane-coated MOF nanoparticles loaded with diABZI. (B) TEM and (C) SEM image. Scale bar = 100 nm. (D) Size distribution by SEM. (E) Average diameters of MOF, MOF/diABZI, and EM@MOF/diABZI. (F) Zeta potentials of EM, MOF/diABZI, and EM@MOF/diABZI. (G) Elemental mapping of C, N, O, Mn, and Cu. (H) SDS–PAGE analysis of protein profiles from marker (I), whole cells (II), EM (III), MOF/diABZI (IV), and EM@MOF/diABZI (V). (I) Western blot analysis of E-cadherin in whole-cell lysate, cell membrane, EM@MOF/diABZI, and MOF/diABZI. (J) Colloidal stability of EM@MOF/diABZI in PBS, serum, and cell-culture medium assessed by PDI. Data are shown as mean ± S.D. of three independent experiments.

3.2. Drug-loading capacity, pH-responsive release, and cytotoxicity

The drug-loading, release, and pancreatic cancer cell-killing properties of EM@MOF/diABZI are illustrated in Fig. 3A. The loading capacity (LC) and encapsulation efficiency (EE) were evaluated at increasing initial diABZI concentrations. As shown in Fig. 3B, the LC increased from approximately 25% to 55% when the feeding concentration of diABZI increased from 1 to 5 mg/mL. In contrast, the EE gradually decreased from approximately 83%–59% with increasing diABZI input (Fig. 3C). These results indicate that a higher drug-feeding concentration enhanced the amount of diABZI incorporated into the MOF framework, although the proportion of the initial drug successfully encapsulated decreased at higher input concentrations. The influence of membrane coating on diABZI release was subsequently investigated. No significant difference in the release profiles was observed between MOF/diABZI and EM@MOF/diABZI (Fig. 3D), indicating that the cancer-cell-membrane coating did not substantially impair diABZI release from the MOF core. To evaluate whether diABZI remained stable after incorporation into the nanoplatform, the retained diABZI content in EM@MOF/diABZI was monitored during storage or incubation under different conditions (Fig. S3). diABZI exhibited good time-dependent stability in EM@MOF/diABZI, and the formulation stored at 4 °C showed better stability than that incubated at 37 °C, indicating that low-temperature storage was favorable for preserving the loaded STING agonist. The release behavior was strongly pH dependent, with the greatest diABZI release observed at pH 5.5, followed by pH 6.5 and pH 7.4 (Fig. 3E). Besides, acidic-stability analysis showed that EM@MOF/diABZI maintained stable particle size, acceptable PDI, substantial membrane-protein retention, and preserved EGFR-binding activity after pH 5.5 exposure, indicating that the biomimetic targeting layer remained functionally stable under drug-release conditions (Fig. S4). This result is consistent with the acid-responsive characteristics of the MOF carrier and suggests that EM@MOF/diABZI may preferentially release diABZI in the mildly acidic tumor microenvironment (Feng et al., 2023; Feng et al., 2026). Material stability is a critical prerequisite for MOF-based biomimetic nanoplatforms, because premature drug leakage, aggregation, membrane-protein loss, or impaired ligand-binding activity may reduce tumor selectivity and increase off-target immune activation. In particular, recent reviews on stimuli-responsive MOF drug-delivery systems have highlighted that pH-responsive MOFs should ideally remain relatively stable at physiological pH and undergo enhanced degradation or cargo release under acidic conditions, although long-term degradation, biological stability, and in vivo accumulation remain important translational concerns (Filiz et al., 2025; Guo et al., 2025a). The biocompatibility of EM@MOF/diABZI was evaluated in normal pancreatic duct epithelial HPDE6-C7 cells. Cell viability remained above 90% after treatment (Fig. 3F), indicating favorable cytocompatibility toward nonmalignant pancreatic cells. In contrast, free diABZI, MOF/diABZI, and EM@MOF/diABZI significantly reduced the viability of BxPC-3 pancreatic cancer cells (Fig. 3G). Based on the dose–response analysis, EM@MOF/diABZI induced approximately 50% cell death at a diABZI-equivalent concentration of 16.3 μM. Therefore, 16.3 μM was selected as the working concentration for subsequent in vitro experiments. Live/dead cell staining further confirmed the enhanced cytotoxic effect of EM@MOF/diABZI in BxPC-3 cells (Fig. 3H). Treatment with blank MOF or EM@MOF caused only a slight increase in dead cells, indicating limited intrinsic toxicity of the carrier materials. By comparison, MOF/diABZI and EM@MOF/diABZI markedly increased the number of dead cells. Quantitative analysis demonstrated that EM@MOF/diABZI induced significantly (P < 0.05) greater cell death than both EM@MOF and MOF/diABZI (Fig. 3I). BxPC-3 cells exhibited stronger p-STING induction and higher IFN-β/CXCL10 production than HPDE6-C7 cells after matched diABZI-equivalent treatment, indicating that the enhanced cancer-cell response resulted from both preferential nanoparticle uptake and greater STING-pathway responsiveness rather than targeting alone (Fig. S5).

Fig. 3.

Drug-loading capacity, pH-responsive release, and cytotoxicity. (A) Schematic illustration of diABZI loading, pH-responsive release, and pancreatic cancer cell killing. (B) Loading capacity (LC) and (C) encapsulation efficiency (EE) of diABZI at increasing initial drug concentrations. (D) In vitro diABZI release profiles of MOF/diABZI and EM@MOF/diABZI. (E) pH-dependent diABZI release from EM@MOF/diABZI at pH 5.5, 6.5, and 7.4. (F) Viability of HPDE6-C7 cells after treatment with the indicated formulations by CCK-8. (G) Viability of BxPC-3 cells after treatment with the indicated formulations by CCK-8. (H) Representative live/dead staining images of BxPC-3 cells after different treatments. (I) Quantification of dead cells. Scale bar = 100 μm. All diABZI-containing groups were normalized to the same total diABZI-equivalent concentration. For nanoparticle formulations, diABZI-equivalent concentration was calculated based on total loaded diABZI. Data are shown as mean ± S.D. of three independent experiments. NS, no significance. ^⁎P < 0.05, ^⁎⁎P < 0.01.

3.3. EM-mediated tumor targeting of EM@MOF/diABZI

The tumor-targeting mechanism of EM@MOF/diABZI is illustrated in Fig. 4A. The outer cancer-cell-membrane shell provides biomimetic interactions with pancreatic cancer cells, while the surface-decorated anti-EGFR antibody enables active recognition of EGFR-overexpressing tumor cells (Zhang et al., 2021). Through the combined effects of homotypic membrane-mediated recognition and EGFR-mediated binding, EM@MOF/diABZI is expected to preferentially accumulate in pancreatic tumors and promote intracellular delivery of the MOF-associated cargo. The anti-EGFR of EM@MOF/diABZI was measured to be 8.9 μM by plotting standard curve (Fig. S6). The EGFR-binding capability of different formulations was first evaluated. As shown in Fig. 4B, both EM and EM@MOF/diABZI exhibited EGFR inhibition rates above 90%, demonstrating that the membrane-associated anti-EGFR antibody retained strong EGFR-binding activity after nanoparticle assembly. In contrast, PBS and MOF/diABZI showed limited inhibitory effects, indicating that the enhanced EGFR interaction was dependent on the antibody-decorated membrane component rather than the MOF core itself. These results confirmed the successful functionalization of EM@MOF/diABZI with EGFR-targeting capability. To further evaluate the selective delivery of EM@MOF/diABZI, Mn²⁺ concentrations were measured in BxPC-3 pancreatic cancer cells and normal pancreatic duct epithelial HPDE6-C7 cells. In BxPC-3 cells, treatment with EM@MOF/diABZI resulted in a significantly (P < 0.05) higher Mn²⁺ concentration than treatment with MOF/diABZI (Fig. 4C). In contrast, no significant difference in intracellular Mn²⁺ concentration was observed between the EM@MOF/diABZI and MOF/diABZI groups in HPDE6-C7 cells (Fig. 4D), suggesting that the membrane-mediated targeting effect was selective for tumor cells rather than normal pancreatic epithelial cells. Cellular uptake was further visualized using fluorescence imaging. In BxPC-3 cells, the fluorescence intensity of EM@MOF/diABZI was significantly (P < 0.05) higher than that of MOF/diABZI, confirming the enhanced cellular uptake by EM in pancreatic cancer cells (Fig. 4E). However, in HPDE6-C7 cells, no significant difference in fluorescence intensity was observed between EM@MOF/diABZI and MOF/diABZI (Fig. 4F). Importantly, EM@MOF/diABZI-treated BxPC-3 cells showed significantly (P < 0.05) stronger fluorescence than EM@MOF/diABZI-treated HPDE6-C7 cells. Together, these findings demonstrate that EM@MOF/diABZI preferentially targets BxPC-3 pancreatic cancer cells, likely through EGFR-mediated active targeting combined with cancer-cell-membrane-associated homotypic recognition. The in vivo tumor-targeting ability of EM@MOF/diABZI was subsequently evaluated by ex vivo fluorescence imaging. As shown in Fig. 4G, the fluorescence signal in tumor tissue was markedly stronger than that in the major organs, indicating preferential accumulation of EM@MOF/diABZI in the tumor. A relatively weak fluorescence signal was also observed in the liver, which may be attributed to partial uptake by the mononuclear phagocyte system (Pattipeiluhu et al., 2022). Nevertheless, the predominant tumor-associated fluorescence suggested effective tumor localization of the nanoplatform (Zhao et al., 2022). Whole-body fluorescence imaging further revealed the tumor targeting of EM@MOF/diABZI in tumor-bearing mice (Fig. 4H). The tumor-associated fluorescence signal increased from 0 to 1 h and became further enhanced at 4 h after administration, indicating efficient nanoparticle accumulation at the tumor site. The signal gradually decreased at 12 and 24 h but remained detectable, suggesting prolonged tumor retention of EM@MOF/diABZI. Collectively, these results demonstrate that EGFR-antibody-decorated cancer-cell-membrane coating substantially improves the selective uptake and tumor accumulation, providing a favorable basis for subsequent STING activation and pancreatic cancer immunotherapy.

Fig. 4.

EM-mediated tumor targeting of EM@MOF/diABZI. (A) Schematic illustration of EGFR-antibody-mediated and cancer-cell-membrane-mediated tumor targeting of EM@MOF/diABZI. (B) EGFR-binding inhibition rates. (C) Mn²⁺ concentration in BxPC-3 cells after treatment with MOF/diABZI or EM@MOF/diABZI. (D) Mn²⁺ concentration in HPDE6-C7 cells after treatment with MOF/diABZI or EM@MOF/diABZI. (E) Quantitative fluorescence analysis of cellular uptake in BxPC-3 cells and HPDE6-C7 cells. (F) Representative cellular uptake fluorescence images in BxPC-3 cells and HPDE6-C7 cells. Scale bar = 50 μm. (G) Ex vivo fluorescence imaging of tumors and major organs after systemic administration of EM@MOF/diABZI. (H) Time-dependent whole-body fluorescence imaging of tumor-bearing mice after intravenous injection of EM@MOF/diABZI. Data are shown as mean ± S.D. of three independent experiments. ^⁎P < 0.05, ^⁎⁎P < 0.01.

3.4. In vivo antitumor efficacy of EM@MOF/diABZI in a pancreatic cancer mouse model

The establishment of the subcutaneous Pan02 pancreatic cancer mouse model and the treatment schedule are illustrated in Fig. 5A. After tumors reached the predetermined volume, mice were randomly assigned to receive PBS, MOF, EM@MOF, MOF/diABZI, or EM@MOF/diABZI. Tumor growth was monitored throughout the treatment period of the PBS, MOF, EM@MOF, MOF/diABZI, and EM@MOF/diABZI groups (Fig. 5B–F). As compared in Fig. 5G, the tumor volumes in the PBS and MOF groups were comparable, indicating that the blank MOF carrier alone did not exert a measurable antitumor effect. Treatment with EM@MOF produced only a slight reduction in tumor growth, suggesting that EM alone had limited therapeutic activity. In contrast, both diABZI-loaded formulations, MOF/diABZI and EM@MOF/diABZI, significantly inhibited tumor growth (P < 0.01). Notably, EM@MOF/diABZI produced a significantly greater reduction in tumor volume than MOF/diABZI (P < 0.05), demonstrating that the EGFR-antibody-decorated cancer-cell-membrane coating enhanced the in vivo antitumor efficacy of the diABZI-loaded MOF platform. Tumor-weight measurements at the experimental endpoint were consistent with the tumor-volume results (Fig. 5H). Moreover, additional control groups including membrane-coated MOF without anti-EGFR, free anti-EGFR, free diABZI, diABZI+anti-EGFR mixture were evaluated to further clarify the contribution of each therapeutic component in EM@MOF/diABZI (Fig. S7). The EM@MOF/diABZI group exhibited the lowest tumor weight among all treatment groups (P < 0.01), whereas MOF/diABZI also significantly reduced tumor burden relative to PBS (P < 0.05). These findings further support the superior therapeutic activity of EM@MOF/diABZI and indicate that tumor-targeted delivery improved the antitumor effect of diABZI in vivo. Notably, free anti-EGFR alone showed limited antitumor effects, and the free diABZI plus anti-EGFR mixture did not achieve the same effect as EM@MOF/diABZI, suggesting that simple co-administration of antibody and diABZI was insufficient to reproduce the targeting advantage of the surface-decorated nanoplatform. The present data indicate that the MOF and membrane components mainly serve as delivery-enhancing elements rather than intrinsically active immunotherapeutic agents. Blank MOF showed minimal antitumor activity, and EM@MOF produced only limited tumor-growth inhibition, indicating that the carrier itself does not substantially activate antitumor immunity. MOF core is interpreted as providing drug loading and pH-responsive release, whereas the EGFR-antibody-decorated cancer-cell membrane improves tumor-cell uptake and tumor accumulation. The superior efficacy of EM@MOF/diABZI over MOF/diABZI should thus be attributed mainly to delivery-mediated enhancement of local diABZI exposure, rather than to strong intrinsic immunostimulatory activity of the carrier. To distinguish the contribution of cancer-cell-membrane-mediated homotypic targeting from EGFR-antibody-mediated active targeting, M@MOF/diABZI without anti-EGFR decoration was introduced as an additional control (Fig. S8). Compared with MOF/diABZI, M@MOF/diABZI showed improved tumor accumulation and antitumor activity, supporting the role of cancer-cell-membrane-mediated homotypic recognition. Notably, EM@MOF/diABZI further enhanced tumor accumulation and tumor-growth inhibition, indicating that EGFR-antibody decoration provided an additional active-targeting advantage. These results demonstrate that the targeting performance of EM@MOF/diABZI is attributable to the combined contribution of cancer-cell-membrane homotypic recognition and EGFR-mediated targeting. Histological examination of tumor sections by hematoxylin and eosin staining provided further evidence (Fig. 5I). Tumors from the PBS, MOF, and EM@MOF groups retained relatively dense tumor-cell morphology, whereas the MOF/diABZI and EM@MOF/diABZI showed more pronounced pathological alterations. Mouse body weight was monitored throughout the study, and no treatment-associated body-weight loss was observed (Fig. 5J). Spleen weights were also assessed at the end of treatment as an indicator of systemic immune status (Fig. 5K). Compared with the PBS group, mice treated with MOF or EM@MOF showed a slight increase in spleen weight, while the MOF/diABZI and EM@MOF/diABZI groups exhibited significantly higher spleen weights than the PBS group (P < 0.05). Importantly, the observed spleen-weight profile should be interpreted together with the subsequent immune-cell and cytokine analyses to determine whether the antitumor response was accompanied by systemic immune activation. Survival was subsequently monitored in Pan02 tumor-bearing mice (Fig. 5L). Consistent with its pronounced tumor-growth inhibition, EM@MOF/diABZI showed the most favorable survival outcome (P = 0.0023). Together, these results demonstrate that targeted delivery of diABZI by the EGFR-antibody-decorated cancer-cell-membrane-coated MOF platform effectively suppresses tumor progression and improves therapeutic benefit in vivo. Nevertheless, several limitations remain. Although the subcutaneous Pan02 model is useful for evaluating preliminary antitumor activity in an immunocompetent setting, it does not fully recapitulate the complex pathological and immunological features of PDAC (Ohtani et al., 2024; Mei et al., 2025). The therapeutic benefit of EM@MOF/diABZI observed in the present subcutaneous model provides proof-of-concept evidence of in vivo antitumor efficacy rather than definitive validation for PDAC treatment. Future studies should consider the unique pancreatic tumor microenvironment, dense desmoplastic stroma, and metastatic behavior (Ye et al., 2025; Liu et al., 2022). Future studies using orthotopic Pan02 models, genetically engineered PDAC models, or patient-derived models will be required to more rigorously assess translational potential.

Fig. 5.

In vivo antitumor efficacy of EM@MOF/diABZI in a pancreatic cancer mouse model. (A) Schematic illustration of Pan02 tumor-model establishment and treatment schedule. (B–F) Tumor volume of mice treated with PBS, MOF, EM@MOF, MOF/diABZI, and EM@MOF/diABZI, respectively. (G) Tumor-growth curves comparison of all groups during treatment. (H) Tumor weights at the experimental endpoint. (I) Representative hematoxylin and eosin staining images of tumor sections. Scale bar = 100 μm. (J) Body-weight changes of mice during treatment. (K) Spleen weights at the experimental endpoint. (L) Kaplan–Meier survival curves of tumor-bearing mice during 60 days. Data are shown as mean ± S.D. (N = 5). ^⁎P < 0.05, ^⁎⁎P < 0.01.

3.5. In vivo biosafety and biodistribution evaluation

The in vivo biosafety of the different formulations was first assessed by histological examination of major organs, including the heart, liver, spleen, lungs, and kidneys. As shown in Fig. 6A, hematoxylin and eosin staining was performed to evaluate potential treatment-associated tissue injury. No obvious histopathological abnormalities were observed in the major organs of mice receiving the nanoparticle formulations, indicating acceptable systemic biocompatibility of the treatment regimen. Serum biochemical parameters, including alanine aminotransferase (ALT), aspartate aminotransferase (AST), albumin (ALB), and blood urea nitrogen (BUN), were subsequently measured to assess hepatic and renal function (Fig. 6B–E). In addition, white blood cell (WBC) and red blood cell (RBC) counts were evaluated to determine potential hematological effects of treatment (Fig. 6F–G). Together, these biochemical and hematological analyses provided a systemic evaluation of the safety profile of EM@MOF/diABZI and the control formulations. Although EM@MOF/diABZI enhanced antitumor immunity through STING activation, the long-term consequences of repeated or prolonged STING stimulation require further investigation. STING activation can promote type I interferon signaling and antitumor immune-cell recruitment. However, excessive or sustained activation may also cause systemic inflammation, immune-related toxicity, immune exhaustion, or compensatory immunosuppressive responses. In the present study, survival monitoring and short-term safety analyses showed no obvious acute toxicity, but these data do not fully exclude delayed immune-related adverse effects. Future studies should include longer follow-up periods, repeated-dose toxicology, longitudinal cytokine analysis, detailed immune-cell profiling, and evaluation of exhaustion or immunosuppressive markers to more comprehensively assess the therapeutic window and long-term safety of EM@MOF/diABZI. The tissue distribution of diABZI was further quantified in major organs and tumor tissues after administration of MOF/diABZI or EM@MOF/diABZI. As shown in Fig. 6H, the diABZI concentration in the collected tumor tissue was significantly higher in the EM@MOF/diABZI group than in the MOF/diABZI group. These findings indicate that cancer-cell-membrane coating and EGFR-antibody decoration altered the in vivo distribution of the diABZI-loaded MOF nanoplatform and enhanced the tumor-targeted delivery of diABZI. Importantly, time-dependent analysis of diABZI accumulation in tumor tissues further demonstrated the superior tumor-delivery performance of EM@MOF/diABZI. At 1, 2, 4, and 12 h after administration, the tumor diABZI concentration in the EM@MOF/diABZI group was significantly higher than that in the MOF/diABZI group (Fig. 6I). The sustained increase in intratumoral diABZI exposure is consistent with the enhanced tumor-targeting and retention observed by fluorescence imaging. These results demonstrate that EGFR-antibody-decorated cancer-cell-membrane coating improves the delivery and persistence of diABZI at the tumor site, thereby providing a pharmacokinetic basis for the superior antitumor efficacy of EM@MOF/diABZI. Premature release in circulation may reduce tumor selectivity and could potentially increase off-target STING activation in peripheral tissues. In the present study, no obvious acute systemic toxicity was detected based on body-weight monitoring, histological examination, serum biochemical parameters, and hematological analysis. Future studies should quantify circulating diABZI, serum inflammatory cytokines, and STING activation in normal organs after repeated dosing to more rigorously evaluate premature release and off-target immune stimulation. Recent advances in nanomedicine have demonstrated that biocompatible nanoplatforms can improve therapeutic efficacy by enhancing cargo stability, facilitating cellular uptake, and promoting intracellular delivery of bioactive molecules. Strategies enabling efficient intracellular and cytosolic delivery are particularly important for therapeutic agents (Wang et al., 2025). Therefore, rationally designed nanocarriers integrating controlled release and tumor-selective delivery represent promising approaches for improving the therapeutic performance of molecular immunomodulators.

Fig. 6.

In vivo biosafety and biodistribution evaluation. (A) Representative hematoxylin and eosin staining images of major organs, including heart, liver, spleen, lungs, and kidneys, from mice receiving the indicated treatments. Scale bar = 100 μm. (B-E) Serum biochemical analysis of ALT, AST, ALB, and BUN. (F-G) Hematological analysis of WBC and RBC counts. (H) diABZI concentrations in major organs and tumor tissues following treatment with MOF/diABZI or EM@MOF/diABZI. (I) Time-dependent diABZI concentration in tumor tissues at 0.5, 1, 2, 4, 12, and 24 h after administration. Data are shown as mean ± S.D. (N = 5). ^⁎P < 0.05, ^⁎⁎P < 0.01.

3.6. EM@MOF/diABZI-mediated STING activation for pancreatic cancer immunotherapy

The mechanism of EM@MOF/diABZI-mediated STING activation for pancreatic cancer immunotherapy is illustrated in Fig. 7A. After preferential accumulation in pancreatic tumors and internalization by tumor-associated cells, the MOF nanoplatform releases diABZI in the acidic tumor microenvironment. diABZI subsequently activates the STING signaling pathway, leading to STING phosphorylation and downstream induction of type I interferon signaling and pro-inflammatory cytokines. These immune signals promote chemokine-mediated immune-cell recruitment, and enhance T-cell activation. To assess intratumoral STING activation, p-STING expression was evaluated by immunofluorescence staining of tumor sections (Fig. 7B). Quantitative analysis showed that the PBS, MOF, and EM@MOF groups exhibited similarly low p-STING staining intensity (Fig. 7C). In contrast, treatment with MOF/diABZI or EM@MOF/diABZI markedly increased p-STING expression (P < 0.01). These results demonstrate that diABZI delivery effectively activated STING signaling in tumors, whereas the MOF core and EM alone showed minimal intrinsic STING-stimulatory activity. The activation of downstream inflammatory signaling was further confirmed by ELISA measurement of IFN-β, CXCL10, TNF-α, and IL-6 in tumor-associated samples (Fig. 7D–G). IFN-β is a major type I interferon produced following STING activation and is essential for cytotoxic T-cell priming (Tong et al., 2024). CXCL10 is an interferon-inducible chemokine that promotes the recruitment of activated T cells into the tumor microenvironment (Sen et al., 2019). In addition, TNF-α and IL-6 are pro-inflammatory cytokines associated with enhanced innate immune activation and local inflammatory remodeling (Peng et al., 2024; Guo et al., 2025b). The significantly increased production of these cytokines in the MOF/diABZI and EM@MOF/diABZI groups therefore further confirmed effective activation of the STING-associated immune response. The downstream adaptive immune response was subsequently evaluated by measuring IFN-γ levels and the abundance of tumor-infiltrating CD4⁺ and CD8⁺ T cells (Fig. 7H–J). Similar to the p-STING results, the PBS, MOF, and EM@MOF groups showed no significant differences in IFN-γ production or CD4⁺ and CD8⁺ T-cell levels. In contrast, both MOF/diABZI and EM@MOF/diABZI significantly increased IFN-γ expression and promoted the infiltration or activation of CD4⁺ and CD8⁺ T cells (P < 0.01). IFN-γ is a key effector cytokine produced by activated T cells and natural killer cells and reflects enhanced antitumor immune activity (Linder et al., 2025; Sun et al., 2024). Increased CD4⁺ T cells may support antigen-specific immune coordination and cytokine production, whereas elevated CD8⁺ T cells indicate enhanced cytotoxic immune-cell infiltration capable of directly eliminating tumor cells (Richter et al., 2023; Ni et al., 2022). To further evaluate immune-microenvironment remodeling after treatment, macrophage polarization and regulatory T cells were analyzed in tumor tissues (Fig. S9). MOF/diABZI and EM@MOF/diABZI increased the proportion of M1-like macrophages while reducing M2-like macrophages, indicating a shift toward a more pro-inflammatory and antitumor macrophage phenotype. In addition, MOF/diABZI and EM@MOF/diABZI reduced the level of CD4⁺CD25⁺Foxp3⁺ regulatory T cells, suggesting partial attenuation of Treg-associated immunosuppression. Further studies involving dendritic-cell activation, MDSC characterization, immune checkpoint regulation, and memory T-cell responses will provide additional understanding of long-term immune remodeling. Collectively, these results demonstrate that diABZI-loaded nanoparticles activate intratumoral STING signaling and induce a coordinated innate and adaptive immune response (Messaoud-Nacer et al., 2022). By enhancing p-STING expression, inflammatory cytokine production, IFN-γ secretion, and T-cell infiltration, EM@MOF/diABZI provides an immunological mechanism for its enhanced tumor-suppression effect in pancreatic cancer. Although EM@MOF/diABZI produced stronger tumor inhibition and immune activation than simpler diABZI-loaded formulations, the magnitude of improvement should be interpreted in the context of the platform complexity. The MOF core enables diABZI loading and pH-responsive release, the membrane coating improves biomimetic tumor-associated interaction and colloidal stability, and EGFR-antibody decoration provides an additional active-targeting effect. The rationale for the multicomponent formulation is to increase local diABZI exposure and reduce nonspecific delivery, rather than to introduce strong intrinsic immunotherapeutic activity from the carrier itself. Despite these encouraging mechanistic findings, several limitations should be considered. Although the blank MOF and EM@MOF groups showed limited antitumor and STING-stimulatory activity under the present conditions, Mn and Cu may influence redox balance, inflammatory signaling, immune-cell function, and long-term tissue toxicity (Wang et al., 2016). The anti-EGFR targeting strategy may have limited direct translatability because of PDAC heterogeneity (Fiorini et al., 2025). Recent advances in nanomedicine have expanded the potential of immunotherapy for pancreatic cancer by integrating therapeutic delivery with immune modulation. For example, Chen et al. developed MOF-derived hollow mesoporous carbon nanospheres for sono-immunotherapy against pancreatic cancer, demonstrating that MOF-derived nanomaterials can serve as multifunctional platforms to enhance antitumor immune responses (Chen et al., 2025). Nevertheless, pancreatic cancer remains challenging due to poor drug penetration, limited immune activation, and an immunosuppressive tumor microenvironment. While localized STING activation is desirable for promoting antitumor immunity, excessive exposure to diABZI may induce uncontrolled type I interferon and inflammatory cytokine responses (Hopfner and Hornung, 2020). Future long-term studies should validate clinically relevant anti-EGFR reagents in human clinical trial to test potential risks of systemic inflammation, immune-related toxicity, or off-target organ injury (Zhao et al., 2021). Although direct bioactivity comparison between freshly prepared diABZI and isolated released diABZI after long-term storage was not performed, the preserved STING activation and downstream immune responses observed after EM@MOF/diABZI treatment indicate that the loaded and delivered diABZI retained functional activity. Future studies will further evaluate the bioactivity of recovered released diABZI using direct in vitro STING activation assays. Although MOF-based nanocarriers, cancer-cell-membrane-coated nanoparticles, EGFR-targeted delivery systems, and STING agonist therapies have each been investigated, their combination for pancreatic cancer immunotherapy remains insufficiently explored. The present study develops an EGFR-antibody-decorated cancer-cell-membrane-coated MOF platform to integrate biomimetic tumor recognition, EGFR-mediated active targeting, acid-responsive drug release, and STING pathway activation within a single nanocarrier. The therapeutic advantage of this platform originates from the improvement of diABZI delivery and tumor immune activation rather than from any individual component alone. Nevertheless, increased structural complexity may introduce challenges related to manufacturing scalability, reproducibility, quality control, and clinical translation. Future studies should focus on formulation optimization, standardized production, and comprehensive translational evaluation. Advances in nanomedicine have also introduced diverse strategies to improve therapeutic delivery, including prodrug nanoassembly, carrier-free therapeutic assembly, and biomimetic nanoplatforms (Feng et al., 2024; Yang et al., 2025; Li et al., 2025; Zeng et al., 2025).

Fig. 7.

EM@MOF/diABZI-mediated STING activation for pancreatic cancer immunotherapy. (A) Schematic illustration of the mechanism by which tumor-targeted EM@MOF/diABZI releases diABZI and activates STING signaling, leading to cytokine secretion and immune activation. (B) Representative p-STING immunofluorescence staining images of tumor sections from the indicated treatment groups. Scale bar = 50 μm. (C) Quantitative analysis of p-STING staining intensity. (D-G) Level of IFN-β, CXCL10, TNF-α, and IL-6 by ELISA. (H) IFN-γ levels by ELISA. (I-J) Percentage of CD4⁺ and CD8⁺ T cells. Data are shown as mean ± S.D. (N = 5). ^⁎P < 0.05, ^⁎⁎P < 0.01.

4. Conclusions

In this study, we developed an EGFR-antibody-decorated cancer-cell-membrane-coated MOF nanoplatform (EM@MOF/diABZI) for targeted delivery of the STING agonist diABZI in pancreatic cancer. The nanoplatform exhibited preserved membrane-associated proteins, high diABZI-loading capacity, and pH-responsive drug release under acidic conditions. The cancer-cell-membrane shell and surface EGFR antibody enhanced cellular uptake and promoted tumor accumulation in vivo. EM@MOF/diABZI achieved stronger tumor-growth inhibition, reduced tumor burden, and improved survival in tumor-bearing mice with good biosafety. Mechanistically, targeted diABZI delivery enhanced STING activation, increased the production of IFN-β, CXCL10, TNF-α, IL-6, and IFN-γ, and promoted CD4⁺ and CD8⁺ T-cell infiltration. Overall, this biomimetic EGFR-targeted MOF platform represents a promising strategy for improving the therapeutic effect of STING agonists and immunotherapy for pancreatic cancer in a subcutaneous pancreatic cancer model. Membrane integrity and antibody function should be further examined under more complex and dynamic in vivo acidic microenvironments. Further validation in orthotopic and clinically relevant PDAC models, together with comprehensive safety, pharmacokinetic, and manufacturing studies, is required before broader conclusions regarding clinical translation can be established.

CRediT authorship contribution statement

Qingmin Chen: Investigation, Writing – original draft. Baifu Zhang: Formal analysis. Jintian Chen: Data curation. Zongyang Li: Software. Hongqiao Cai: Project administration, Funding acquisition, Conceptualization, Writing – review & editing.

Ethics declaration

Animal subject

This study was conducted in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines.

This study was approved by the the Institutional Animal Care and Use Committee of Jilin university first hospital.

Funding

This study is funded by Natural Science Foundation of Jilin province (No. YDZJ202401650ZYTS), Jilin Province Health Talent Special Program (JLSWSRCZX2025-101), Jilin Provincial Department of Finance (JLSRCZX2025-092), First Hospital of Jilin University (No. JDYY-TRP-2025003).

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.

Appendix A. Supplementary data

Supplementary material

XRD analysis, membrane-coating efficiency, stability of diABZI, acidic stability of the membrane coating, comparative STING-pathway activation, standard curve for measuring the anti-EGFR concentration, comparison of anti-tumor efficacy among different therapeutic component groups, contribution of EGFR-antibody decoration to the therapeutic performance and tumor targeting, analysis of macrophage polarization and regulatory T cells after different treatments.

Data availability

Data will be made available on request.