Engineered Akkermansia muciniphila vesicles for targeted pyroptosis and trained immunity to enhance immunotherapy in hepatocellular carcinoma

Lanxiang Huang; Yuan Rong; Minghui Guo; Min Liu; Fei Long; Wei Zhong; Yue Hu; Xin He; Jiurong He; Diwei Zheng; Chunhui Yuan; Fubing Wang

doi:10.1016/j.xcrm.2025.102524

. 2026 Jan 20;7(1):102524. doi: 10.1016/j.xcrm.2025.102524

Engineered Akkermansia muciniphila vesicles for targeted pyroptosis and trained immunity to enhance immunotherapy in hepatocellular carcinoma

Lanxiang Huang ^1,^2,^3,⁸, Yuan Rong ^1,⁸, Minghui Guo ^2,^4,⁸, Min Liu ^2,^4,⁸, Fei Long ^2,⁴, Wei Zhong ^2,⁴, Yue Hu ^2,⁴, Xin He ^2,⁴, Jiurong He ^2,⁴, Diwei Zheng ^5,^∗, Chunhui Yuan ^6,^∗∗, Fubing Wang ^1,^2,^7,^9,^∗∗∗

PMCID: PMC12866108 PMID: 41564859

Summary

Hepatocellular carcinoma (HCC) features a tumor immunosuppressive microenvironment (TIME) and limited response to immune checkpoint inhibitors (ICIs). To address this, we develop ultrasound-responsive nanoparticles by encapsulating PD-L1-targeting small interfering RNA (siRNA) and sonodynamic metal-organic frameworks (MOFs) into bacterial membrane vesicles (BMVs) derived from Akkermansia muciniphila. The siRNA-MOF@BMV (SMB) demonstrates HCC-specific accumulation via N-acetylgalactosamine (GalNAc) and induces pyroptosis through NLRP3/Caspase-1/GSDMD pathway activation under ultrasound, releasing tumor antigens. Simultaneously, SMB further induces trained immunity in tumor-associated macrophages (TAMs), promoting CXCL9⁺ phenotypes that enhance antigen presentation and chemotaxis capacity. This increases cytotoxic CD8⁺ T cell infiltration and reduces exhausted T cells, reshaping the TIME. Furthermore, SMB exhibits superior tumor suppression compared to clinical ICIs through systematic evaluations in orthotopic HCC mouse models, primary HCC models, patient-derived xenograft (PDX), and organoid models. SMB presents a multifunctional immunotherapeutic strategy integrating targeted pyroptosis induction, innate immune training, and ICI delivery, representing a potent immunotherapeutic agent for HCC.

Keywords: immune checkpoint inhibitors, trained immunity, tumor-associated macrophages, bacterial membrane vesicles, metal-organic frameworks, pyroptosis

Graphical abstract

Highlights

•
SMB nanoparticles integrate pyroptosis induction and immune checkpoint silencing
•
SMB nanoparticles trigger ultrasound-responsive pyroptosis and remodel TIME
•
SMB nanoparticles reprogram TAMs toward CXCL9⁺ macrophages and enhance CTLs
•
SMB nanoparticles show potent antitumor efficacy in preclinical HCC models

Wang et al. engineer A. muciniphila-derived vesicles to co-deliver PD-L1 siRNA and sonodynamic MOF. The nanoparticles trigger pyroptosis, induce trained immunity in tumor-associated macrophages, enhance cytotoxic T cell infiltration, and achieve potent antitumor efficacy in HCC models.

Introduction

Hepatocellular carcinoma (HCC) remains a major global health burden, with incidence projected to be over 1 million new cases by 2025.¹ The advent of immune checkpoint inhibitors (ICIs) has revolutionized HCC management and represents a new first-line standard of care, as evidenced by landmark clinical trials such as IMbrave150 and HIMALAYA.² Specifically, the combination of Durvalumab (anti-PD-L1) and Tremelimumab (anti-CTLA4) yielded a 4-year survival rate of 25.2% in patients with advanced HCC.³ However, the resistance mediated by tumor immunosuppressive microenvironment (TIME) limits durable responses in HCC, with objective response rates remaining below 30% across ICI-based regimens.⁴^,⁵

Antitumor T cell responses are largely blunted by the lack of tumor-associated antigens (TAAs) or tumor-associated neoantigens (TNAs), deficient antigen presentation, and infiltration of immunosuppressive innate immune cells such as tumor-associated macrophages (TAMs) and myeloid-derived suppressor cells (MDSCs).⁶^,⁷^,⁸ In response to these challenges, we propose an interventional strategy designed to synergistically target three key stages of the antitumor immune response: (1) induction of trained immunity to remodel the immune microenvironment and enhance innate immune readiness, (2) activation of pyroptosis to promote immunogenic antigen release, and (3) small interfering RNA (siRNA)-mediated blockade of immune checkpoints to prevent T cell exhaustion and sustain cytotoxic activity.

To overcome the barriers posed by the TIME, recent strategies have focused on dual modulation of both innate and adaptive immunity. Notably, trained immunity, termed as a functional reprogramming of innate immune cells that enhances responsiveness to secondary stimuli, has emerged as a promising approach to re-educate immunosuppressive innate immune cells.⁹ Bacterial membrane vesicles (BMVs) carry multiple pathogen-associated molecular patterns, while circumventing infection risks associated with live bacteria.¹⁰^,¹¹^,¹²^,¹³ These properties make BMVs an ideal candidate for inducing trained immunity. Among potential BMV sources, Akkermansia muciniphila (A. muciniphila), a gut commensal with established immunomodulatory properties, holds particular promise. Preclinical studies have demonstrated the A. muciniphila’s effective capacity to promote antitumor immune response in the process of tumorigenesis and potentiate ICI efficacy through gut microbiota modulation.¹⁴^,¹⁵^,¹⁶ Specifically, A. muciniphila have alleviated non-alcoholic fatty liver disease (NAFLD),¹⁷^,¹⁸ a major HCC risk factor, and highly expressed N-acetylgalactosamine (GalNAc).¹⁹ GalNAc serves as a ligand for asialoglycoprotein receptor (ASGPR), enabling hepatocyte-specific targeted delivery of therapeutic oligonucleotides.²⁰ We, therefore, hypothesize that A. muciniphila-derived BMVs (A. muciniphila BMVs) could serve as multifunctional carriers capable of both targeted delivery and TIME remodeling via trained immunity induction.

The antitumor immune response relies on three key phases: sufficient antigen release, effective antigen presentation, and the maintenance of T cell function. However, in HCC, both antigen exposure and antigen presentation are fundamentally constrained by the pronounced spatial and molecular heterogeneity. To circumvent the challenges of identifying pan-HCC immunodominant antigens and the inherently low frequency of TNAs,⁶^,²¹ emerging strategies are increasingly focused on endogenous modulation of tumor cell death to bypass antigen-selection barriers. Apoptosis, ferroptosis, pyroptosis, and necroptosis are distinct forms of regulated cell death, each with characteristics and effects on the cellular immune response. Ferroptosis, for instance, influences the immune microenvironment by reducing the number of MDSCs, promoting polarization of TAMs, and exhibiting limited immunogenicity. In contrast, pyroptosis is characterized by a robust inflammatory response and high immunogenicity, which activates both innate and adaptive immune pathways.²² This activation can convert immunologically “cold” tumors into more treatment-responsive “hot” tumors, a transformation of particular importance in liver cancer, which is typically considered an immunologically “cold” tumor. Mechanistically, this property originates from the lytic nature of pyroptotic cell death, which leads to the simultaneous release of TAAs and potent immunoadjuvant signals, primarily damage-associated molecular patterns (DAMPs) and pro-inflammatory cytokines.²³

In the present study, we engineered a multifunctional therapeutic nanoparticle that synergistically integrates pyroptosis induction with PD-L1 blockade, using A. muciniphila BMVs as targeted delivery vesicles. We synthesized a metal-organic framework (MOF) that acts as a sonosensitizer and an siRNA carrier. The high surface area and tunable pore structure of MOF allow efficient reactive oxygen species (ROS) generation under ultrasound (US) and coordination loading of siRNA through metal-phosphate interactions.²⁴ Moreover, US-triggered ROS oxidize lysosomal membranes, promoting endosomal escape of siRNA.²⁵ Afterward, the sonodynamic MOF material adsorbed with PD-L1 siRNA was loaded into A. muciniphila BMVs. This strategy facilitated the development of an US-responsive PD-L1 siRNA-MOF@BMV (SMB) nanoparticles. We systematically evaluated the therapeutic efficacy of SMB nanoparticles across multiple platforms, including in vitro assays, orthotopic mouse models, oncogenic plasmid-induced primary HCC models, patient-derived xenograft (PDX), and HCC organoid models. Our results demonstrate that these engineered SMB nanoparticles effectively induce pyroptosis in HCC cells through the NLRP3/Caspase-1/GSDMD pathway, eliciting a robust anti-tumor immune response under US stimulation. This immune response is characterized by the enhanced enrichment of CXCL9⁺ macrophages and activation of intracellular signatures associated with trained immunity, as well as increased infiltration of cytotoxic T lymphocytes (CTLs) within the HCC microenvironment. These findings highlight the immunomodulatory role of A. muciniphila-engineered SMB nanoparticles and reveal a promising immunotherapeutic strategy for HCC.

Results

SMB nanoparticles target tumor cells and specifically accumulate in HCC tissues in vivo

Preclinical studies have shown that specific strains, including A. muciniphila,²⁶ Bifidobacterium longum,²⁷ Lactobacillus reuteri,²⁸ and Lactobacillus rhamnosus,²⁹ can enhance ICIs responses via distinct mechanisms. Given their favorable immunoregulatory properties, we screened BMVs derived from these bacterial strains to identify candidates exhibiting high specificity for HCC cells. Under anaerobic culture, we isolated BMVs from A. muciniphila, B. longum, L. reuteri, and L. rhamnosus, and characterized them using transmission electron microscopy (TEM), which revealed nanoscale vesicles with strain-specific size distributions (Figures S1A and S1B). Flow cytometry analysis indicated that BMVs derived from A. muciniphila exhibited superior targeting efficiency to HCC cells compared to those from other strains, a result consistent across both murine and human HCC cell lines (Figures 1A, S1C, and S1D). GalNAc is a key ligand involved in ASGPR-mediated endocytosis in hepatocytes.³⁰ qPCR and western blot analyses revealed significantly higher ASGPR expression in murine Hepa 1-6 and H22 cells and human HepG2 cells compared to normal hepatocytes. Single-cell transcriptomics further showed elevated ASGPR levels in liver cancer tissues, particularly HCC, with uniform manifold approximation and projection (UMAP) visualization indicating predominant localization in tumor cells (Figures 1B–1D, S1E, and S1F). Quantitative analysis of GalNAc revealed that A. muciniphila BMVs had a significantly higher concentration of GalNAc (0.1263 ± 0.0153 μg/mL) compared to BMVs from other strains (Figure 1E), potentially explaining their enhanced uptake by HCC cells. Based on these findings, A. muciniphila BMVs were selected for subsequent studies.

SMB nanoparticles target tumor cells and selectively accumulate in HCC tissues *in vivo*

(A) Flow cytometric quantification of targeted uptake of gut-probiotic-derived BMVs by murine HCC (Hepa 1-6) cells.

(B) Dot plot of ASGR1 expression across adjacent liver (AL), HCC, intrahepatic cholangiocarcinoma (ICC), combined HCC-cholangiocarcinoma (CHC), and secondary liver cancer (SLC) tissues. Circle size denotes ASGR1⁺ cell fraction; color intensity indicates mean expression. Data source: CNCB: PRJCA007744.

(C) UMAP plot of ASGR1 expression across the five tissue categories at single-cell resolution.

(D) UMAP plot displaying the density distribution of ASGR1 expression.

(E) LC-MS/MS quantification of N-acetylgalactosamine in BMVs isolated from *A. muciniphila, B. longum, L. reuteri,* and *L. rhamnosus*. Statistical significance was determined by t test (∗∗p < 0.01, n = 3, mean ± SEM).

(F) TEM images of BMVs derived from *A. muciniphila*, MOF, and engineered probiotic membrane vesicles (SMB nanoformulations). Scale bars, 200 nm.

(G) Time-dependent relative absorbance of DBPF solution under US irradiation in the presence of MOF, SM, BMVs, MB, and siRNA-MOF@BMV (SMB) nanoformulations compared to the control group (n = 3, mean ± SEM).

(H) Confocal laser scanning microscopy time course showing Hepa 1-6 cells incubated with Cy3-labeled SMB nanoformulations. Scale bars, 36.3 μm.

(I) *In vivo* fluorescence imaging of orthotopic H22 liver tumor-bearing mice at indicated time points after intravenous injection of Cy5.5-NHS-labeled SMB nanoformulations.

(J) Representative H&E-stained kidney sections from H22 orthotopic HCC mice following SMB nanoparticle injection. Scale bars, 50 μm.

(K) Serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), albumin (ALB), and alkaline phosphatase (ALP) levels in healthy, non-tumor-bearing mice after SMB nanoparticle administration; statistical analysis by t test, ns: p > 0.05 (n = 6, mean ± SEM).

To integrate therapeutic functions, we engineered a core-shell nanoparticle comprising (1) a sonodynamic MOF built from zinc ions and porphyrin linkers, synthesized as ∼180-nm rhombic dodecahedra—the porphyrin stabilizes the framework and acts as a sonosensitizer,³¹ while Zn²⁺ overload promotes ROS-driven pyroptosis³²; (2) PD-L1-targeting siRNA, loaded via metal-phosphate coordination; and (3) A. muciniphila BMVs, encapsulating the siRNA-MOF (SM) cores (siRNA-MOF@BMV, hereafter “SMB”). TEM imaging confirmed the core-shell morphology of the SMB nanoparticles (Figure 1F), with an average diameter of ∼200 nm, which was corroborated by dynamic light scattering analysis (Figure S1G). Structural integrity assessment through X-ray diffraction and UV-visible spectroscopy revealed successful siRNA incorporation, while maintaining the crystalline framework of MOF (Figures S1H and S1I). Energy-dispersive X-ray spectroscopy mapping confirmed the homogeneous spatial distribution of zinc (Zn), carbon (C), and oxygen (O) throughout the nanoparticles (Figure S1J). The SMB nanoparticles exhibited a characteristic negative surface charge, contributing to its enhanced biocompatibility and cellular internalization efficiency, while demonstrating remarkable stability in both phosphate-buffered saline (PBS) and serum for up to 7 days (Figures S1K and S1L). Notably, under US irradiation, the SMB nanoparticles maintained robust singlet oxygen (¹O₂) generation capacity, comparable to pristine MOF, confirming that its sonodynamic function remained intact after siRNA loading and BMV encapsulation (Figure 1G).

To systematically assess cellular internalization, we incubated Cy3-labeled SMB nanoparticles with Hepa 1-6 cells for varying durations. Confocal microscopy revealed a time-dependent increase in intracellular Cy3 fluorescence, indicating efficient nanoparticle uptake and sustained retention of the nanoparticles within the cells (Figure 1H). Flow cytometry analysis showed that siRNA@BMV (SB) and SMB nanoparticles coated with the A. muciniphila membrane exhibited comparable uptake efficiency in Hepa 1-6 cells, both of which were significantly higher than that of uncoated SM nanoparticles. In addition, the uptake of SMB by 4T1, RAW264.7 cells, and M1/M2 bone marrow-derived macrophages (BMDMs) was lower than that observed in the hepatic cancer cell line, indicating that the A. muciniphila membrane facilitates cellular internalization and that SMB exhibits enhanced targeting specificity toward liver cancer cells (Figures S2A and S2B). For in vivo biodistribution assessment, an orthotopic H22 liver cancer model was employed and Cy5.5-labeled SMB nanoparticles were administered via intravenous injection. Real-time in vivo imaging revealed substantial hepatic accumulation within 2 h post-administration, with remarkable retention efficiency maintained at 24 h (Figure 1I). Ex vivo organ imaging further validated tumor-specific accumulation, demonstrating preferential distribution in hepatic tumor tissues with minimal off-target accumulation, except for expected renal clearance (Figure S2C). Collectively, these comprehensive characterizations demonstrate the successful development of multifunctional SMB nanoparticles with optimized tumor-targeting capabilities and therapeutic potential.

To comprehensively assess the biosafety profile of the SMB nanoparticles, the inflammatory cytokine expression patterns were systematically evaluated in both tumor microenvironments and systemic circulation. Quantitative analysis revealed that SMB nanoparticles induced a significant upregulation of pro-inflammatory cytokines, including tumor necrosis factor (TNF)-α, interleukin (IL)-6, and interferon (IFN)-γ, specifically within the tumor microenvironment, while maintaining systemic cytokine homeostasis (Figures S3A and S3B). Comprehensive hematological profiling demonstrated that all serum cytokine levels and complete blood count parameters remained within physiological ranges (Figure S3C). Although renal clearance of SMB nanoparticles was observed, no significant renal impairment was detected, and analysis of histopathological examination of major organs (heart, spleen) showed no significant pathological alterations (Figures 1J and S3D). Furthermore, biochemical analysis of renal function markers, such as uric acid (UA) and creatinine (CREA), and liver function markers alanine aminotransferase (ALT), aspartate aminotransferase (AST), albumin (ALB), and alkaline phosphatase (ALP) in healthy, non-tumor-bearing mice following SMB nanoparticle injection showed no evidence of significant toxicity (Figures 1K and S3E). Overall, these collective findings demonstrate that the SMB nanoparticles effectively induce localized tumor-specific inflammatory responses, while maintaining systemic immune homeostasis.

SMB nanoparticles specifically kill tumor cells and suppress tumor growth in vivo

After confirming the targeted uptake capability of the SMB nanoformulation, we further evaluated its gene silencing efficiency. Hepa 1-6 cells were co-incubated with various formulations (BMV, SB, MOF@BMV [MB], and SMB) under US irradiation. Quantitative analysis revealed that SMB nanoparticles induced a significant downregulation of PD-L1 expression compared to the control, SB, and SM groups (Figure 2A). Moreover, SMB-mediated silencing of the PD-L1 gene demonstrated broad efficacy across multiple HCC cell lines (Figure S4A). This enhanced therapeutic effect can be mechanistically attributed to two synergistic factors: (1) BMV-mediated enhancement of cellular internalization and (2) MOF-induced generation of ROS, which promoted the endosomal escape of siRNA—a critical bottleneck in RNA interference therapy. ROS generation was quantitatively confirmed using a fluorescent probe, demonstrating that MOF-containing formulations produced a substantial amount of intracellular ROS (Figure 2B). These results demonstrate the capability of SMB nanoparticles to enhance siRNA-mediated gene silencing through ROS-facilitated endosomal disruption. Furthermore, the strong induction of ROS suggests that SMB nanoparticles possess significant potential for tumor cell eradication through ROS-mediated sonodynamic effects.

SMB nanoparticles specifically induce tumor cell death and suppress tumor growth *in vivo*

(A–D) Hepa 1-6 cells were treated with BMV, SB, SM, and SMB nanoformulations, followed by US irradiation. (A) Western blot of intracellular PD-L1 protein levels (GAPDH as a loading control). (B) Flow cytometric analysis of intracellular ROS levels. (C) Assessment of cell viability using SYTOX Red dead cell staining (red indicates dead cells). Scale bars, 50 μm. (D) Quantitative analysis of cell viability. Statistical analysis was performed using one-way ANOVA (∗∗∗∗p < 0.0001, n = 3, mean ± SEM).

(E and F) Flow cytometry analysis of PI-stained cell death rates in H22, RAW 264.7, 4T1, and Huh-7 cells after incubation with SMB nanoformulations and US irradiation (n = 3, mean ± SEM).

(G) Representative US images (left) and dissected liver photographs (right) of H22 orthotopic liver tumor-bearing mice from control and treatment groups. Arrows indicate the tumor region.

(H) Body weight change curves of mice after initiation of treatment (n = 6, mean ± SD).

(I) Representative H&E-stained images of tumor tissues, lung, ascites, and lymph nodes from different treatment groups. Scale bars, 100 μm.

(J) Representative Ki67 and ROS staining of tumor tissues from different treatment groups. Scale bars, 100 μm.

To evaluate the therapeutic specificity of SMB nanoparticles, their cytotoxicity was assessed under US irradiation at different intensities. Flow cytometry analysis revealed that the cell death rate increased with rising US intensity and plateaued at 1.5–2.5 W/cm², indicating that 1.5 W/cm² represents an optimal US intensity (Figure S4B). SYTOX Red staining revealed minimal cell death in Hepa 1-6 cells following treatment with BMV or PD-L1 siRNA formulations, confirming both the biocompatibility of these components and the non-cytotoxic nature of PD-L1 silencing (Figures 2C and 2D). Importantly, SMB nanoparticles selectively and robustly induced cytotoxicity in Hepa 1-6 liver cancer cells, while exhibiting minimal cytotoxic effects on non-target cell lines, including RAW264.7 macrophages and 4T1 breast cancer cells. This selectivity is likely due to the limited internalization of SMB nanoparticles by these non-target cells (Figures 2E and 2F). In an orthotopic H22 HCC model, longitudinal US imaging revealed significant tumor regression in mice treated with SMB nanoparticles combined with US irradiation, demonstrating superior therapeutic efficiency compared to control, BMV, SB, and MB treatment groups (Figures 2G, 2H, and S4C). Histopathological examination of tumor sections from SMB-treated mice exhibited extensive necrotic regions, as indicated by nuclear pyknosis and chromatin margination (Figure 2I). These tumors also exhibited a significantly reduced Ki67 proliferation index, elevated ROS levels, and markedly downregulated PD-L1 expression compared to other treatment groups (Figures 2J, S4D, and S4E). In addition, SMB nanoparticles further significantly attenuated malignant ascites formation (Figure S4F), accompanied by decreased incidence of both lymph node and pulmonary metastases (Figure 2I).

We then induced the development of HCC via hydrodynamic injection with c-MYC oncogenic plasmids and initiated SMB nanoparticle treatment in week 2. Compared to untreated controls, SMB-treated mice exhibited markedly reduced tumor growth, demonstrating that these nanoparticles effectively impede early-stage HCC progression (Figures S5A–S5C). This finding not only corroborates the antitumor potency of SMB but also highlights its promise as a prophylactic intervention for individuals at high risk of developing HCC. Collectively, these comprehensive results substantiate that SMB nanoparticles represent a promising multimodal therapeutic strategy for HCC, effectively inhibiting both primary tumor growth and metastatic progression through its precisely engineered multifunctional design.

SMB nanoparticles trigger tumor cell pyroptosis via the NLRP3/Caspase-1/GSDMD pathway

To elucidate the cell death pathways underlying the tumor-suppressive effects of engineered SMB nanoparticles, RNA sequencing (RNA-seq) analysis was performed on tumor tissues derived from control and SMB-treated groups in H22 orthotopic HCC model. Single-sample gene set enrichment analysis (GSEA) revealed a significant upregulation of gene sets associated with pyroptosis, ferroptosis, and necroptosis in the SMB-treated group, suggesting that SMB treatment may induce different subtypes of tumor cell death (Figure 3A). However, inhibition assays using specific inhibitors for ferroptosis, necroptosis, and pyroptosis demonstrated that only the pyroptosis inhibitor significantly attenuated SMB-induced cell death (Figure 3B). Consistent with this, markers associated with ferroptosis (ACSL4 and GPX4) and necroptosis (pMLKL and pRIPK3) remained unaltered (Figure S6A), further supporting the predominant role of pyroptosis in SMB-mediated tumor cell death.

Characteristic pyroptotic morphological features, including large membrane protrusions and cytoplasmic leakage, were observed in Hepa 1-6 cells treated with SMB nanoparticles, as visualized by bright-field microscopy and TEM (Figures 3C, S6B, and S6C). ELISA assays confirmed significantly elevated levels of IL-18 in the SMB-treated group compared to the control group (Figure 3D), consistent with the induction of pyroptosis. Pyroptosis is characterized by the formation of cell membrane pores, a process mediated by the gasdermin protein family.³³ SMB nanoparticles specifically induced the cleavage of GSDMD, but not other gasdermin family members such as GSDMA, GSDMB, GSDMC, or GSDME (Figure 3E). Moreover, knockdown of GSDMD significantly reduced SMB-induced cell death (Figures 3F and S6D), demonstrating that SMB nanoparticles primarily induce pyroptosis in a GSDMD-dependent manner in HCC.

Caspase-1, Caspase-4, Caspase-5, and Caspase-11 are known to cleave GSDMD in the canonical pyroptosis pathway.³⁴ In response to SMB treatment, only Caspase-1 was activated, while Caspase-4, Caspase-5, and Caspase-11 remained inactive (Figure 3G). Caspase-1 knockdown inhibited GSDMD cleavage and cell death, confirming that pyroptosis is Caspase-1-dependent (Figures 3H, S6E, and S6F). Activation of AIM2 and NLR family members, including NALP1, NLRP3, and NLRC4, is known to induce pyroptosis via Caspase-1.³⁵ In SMB-treated cells, NLRP3 protein levels were significantly upregulated, whereas AIM2, NALP1, and NLRC4 remained unchanged (Figure 3I). Knockdown of NLRP3 attenuated Caspase-1 activation and cell death, further confirming the critical role of NLRP3 in SMB-induced pyroptosis (Figures 3J, 3K, and S6G). The NLRP3 inflammasome is a cytosolic multi-molecular complex composed of the sensor molecule NLRP3, the adaptor protein ASC, and the effector protein Caspase-1. Upon activation of the inflammasome sensor molecule, ASC oligomerizes to form ASC specks, which are essential for downstream signaling (Figure 3L). Collectively, these findings demonstrate that SMB nanoparticles induce tumor cell pyroptosis through the NLRP3/Caspase-1/GSDMD pathway in vitro.

SMB nanoparticles activate the NLRP3 inflammasome may through dual mechanisms. First, SMB induces the production of excessive ROS within HCC cells. Elevated ROS levels disrupt mitochondrial function, a key event associated with NLRP3 inflammasome activation.³⁶ Mitochondrial ROS can react with mitochondrial DNA to generate oxidized mitochondrial DNA, which subsequently binds to cytosolic NLRP3, triggering inflammasome assembly and activation. Additionally, mitochondrial ROS may exacerbate pyroptosis by promoting GSDMD oligomerization and pore formation.³⁷ Furthermore, the SMB nanoparticles contain bacterial components, such as nucleic acids, proteins, and polysaccharides derived from A. muciniphila, which are known to activate the NLRP3 inflammasome.³⁸ Consistent with these mechanisms, GSEA of RNA-seq data from the in vivo H22 tumor model revealed significant enrichment of the NLRP3 signaling pathway in the SMB-treated group. Immunohistochemical analysis demonstrated increased NLRP3 protein levels, and immunoblotting confirmed the activation of the NLRP3/Caspase-1/GSDMD axis (Figures 3M, 3N, S6H, and S6I). In conclusion, these results confirmed that SMB nanoparticles specifically induce HCC cell pyroptosis via the NLRP3/Caspase-1/GSDMD pathway both in vitro and in vivo.

SMB nanoparticles remodel the tumor immunosuppressive microenvironment in HCC

To investigate whether the anti-tumor effects of SMB nanoparticles involve remodeling of the TIME in HCC, RNA-seq data from H22 tumor tissues in control and SMB nanoparticle-treated groups were analyzed. Differentially expressed genes were enriched in pathways associated with immune activation, including leukocyte-mediated immunity, lymphocyte activation, innate immune regulation, and positive regulation of immune responses (Figure 4A). Deconvolution of single-cell RNA-seq (scRNA-seq) data further revealed increased infiltration of M1 macrophages and a concomitant reduction in M2 macrophages, along with elevated proportions of activated and memory CD8⁺ T cells. Cluster analysis showed a strong correlation between CD8⁺ T cells and M1 macrophages, suggesting a significant association in their abundance (Figures 4B and S7A). Flow cytometry confirmed that SMB nanoparticles induced significant alterations in immune cell infiltration, reducing M2 macrophages, while increasing activated dendritic cells (DCs) and CD8⁺ T cells (Figures 4C and S7B–S7D), and the increased infiltration of CD8⁺ T cells within the tumor microenvironment is tumor specific (Figure S7E). Further analysis demonstrated that SMB nanoparticles enhanced the infiltration of granzyme B⁺ cytotoxic CD8⁺ T cells, while reducing the proportion of TIM-3⁺-exhausted CD8⁺ T cells (Figures 4D, 4E, S7F, and S7G). Long-term immune memory was confirmed by increased infiltration of memory precursor, central memory, and tissue-resident memory CD8⁺ T cells (Figures 4F–4H and S7H–S7J). In addition, the data revealed a significant increase in the proportion of PD1⁺ CD8⁺ T cells and PD1⁺ CD4⁺ T cells in tumors treated with SMB nanoparticles compared to control groups. This enhanced infiltration is likely due to the combination of PD-L1 suppression and the pro-inflammatory tumor microenvironment created by the induction of pyroptosis (Figures S7K and S7L). Together with the RNA-seq deconvolution data, these findings demonstrate that SMB nanoparticles promote the establishment of an immunologically reactive microenvironment in HCC.

SMB nanoparticles stimulate macrophage trained immunity in tumor microenvironment

As previously demonstrated, the type I IFN (IFN-I signaling pathway was activated in the SMB-treated group (Figure 4A). This is consistent with the established role of pyroptosis in triggering IFN-I pathway,³⁹ which was further validated in our SMB-treated group (Figure 5A). GSEA analysis revealed significant enrichment of pathways associated with innate immune receptor signaling, pattern recognition receptor pathways, and IFN-α/β responses in the SMB group compared to controls, corroborating the activation of the IFN-I response (Figure 5B). Notably, IFN-I response activation is a hallmark of trained immunity, supporting our hypothesis that SMB induces trained immunity.⁴⁰ Moreover, the cGAS-STING pathway, a key regulator of the IFN-I response,⁴¹ was significantly enriched in SMB-treated tumors. Immunoblotting demonstrated increased levels of cGAS, STING, p-IRF3, and p-STAT1 in SMB-treated tissues, further validating the activation of the cGAS-STING pathway (Figure 5C). Immunofluorescence analysis of tumor tissue sections revealed robust STING protein colocalization with F4/80⁺ macrophages but not with CD11c⁺ DCs, across different treatment groups (Figures 5D and S7M). Moreover, the number of IL-1β⁺F4/80⁺ and TNF-α⁺F4/80⁺ double-positive cells was significantly increased in the SMB-treated group. Given that IL-1β and TNF-α are canonical markers of trained immunity, this finding suggests that SMB promotes a trained immunophenotype in macrophages (Figures S7N and S7O). Based on these results, we propose that SMB induces tumor cell pyroptosis, releasing cellular contents and inflammatory mediators that subsequently activate the cGAS-STING pathway in TAMs. This activation triggers downstream IFN-I responses and promotes trained immunity of TAMs.

To validate this speculation, mouse bone marrow-derived macrophages were extracted and polarized to the tumor-promoting M2 phenotype to mimic TAMs. These M2 macrophages were co-cultured with tumor cells in Transwell chambers and treated with different formulations. Supernatants and cell pellets from the co-cultures were collected. Upon lipopolysaccharide (LPS) stimulation, macrophages from the co-cultures rapidly secreted high levels of IFN-β (Figures 5E and S8A), a direct indicator of trained immunity. RNA-seq analysis of the collected cell pellets revealed upregulation of trained immunity-related pathways and genes in SMB-treated macrophages (Figures 5F and 5G). Thus, these findings demonstrated that SMB nanoparticles remodel and induce a trained immunity signature in TAMs.

To identify the specific macrophage subtypes induced by SMB, publicly available single-cell sequencing data from mouse HCC tumors⁴² was analyzed and annotated five major cell types (Figure 5H). Quantitative assessment using UCell scoring revealed that macrophages in the SMB-treated group showed the highest enrichment score for upregulated genes (Figure 5I). Subsequent subtyping of macrophages identified six distinct subtypes. The SMB-treated group demonstrated pronounced enrichment of upregulated genes specifically within the CXCL9⁺ macrophage subtype (Mph_CXCL9) (Figures 5J–5M). Notably, the proportion of the CXCL9⁺ macrophages significantly increased in the anti-PD-1-treated group in the referenced study, indicating a positive response to immunotherapy (Figure 5N). Consistent with this observation, co-culture experiments showed that macrophages incubated with SMB-treated HCC cells significantly increased CXCL9 secretion in the supernatant (Figure 5O). These findings collectively indicate that SBM drives the transformation of TAMs into the immunoresponsive CXCL9⁺ macrophage subtype. Furthermore, macrophage depletion via gadolinium chloride treatment substantially attenuated SMB-mediated tumor suppression (Figures 5P, S8B, and S8C), highlighting the essential role of macrophages in mediating the anti-tumor activity of SMB.

SMB nanoparticles exhibit comparable tumor suppression and immune activation in HCC PDX and PDO models

To further evaluate the clinical relevance of SMB-induced macrophage subtype changes in human HCC, scRNA-seq data from 124 HCC patients (160 samples) from the China National Center for Bioinformation (CNCB: PRJCA007744)⁴³ was retrieved. Cell subpopulations were identified using the FindNeighbors and FindClusters algorithms, followed by dimensionality reduction and visualization with UMAP. To identify macrophage subtypes associated with SMB treatment in human samples, mouse homologous genes were mapped to human orthologs and gene set scores were calculated using the UCell algorithm (Figures 6A–6D). UCell scores of SMB-enriched genes were favorably correlated with patient survival (Figure 6E). Notably, consistent with the findings in mouse single-cell data, analysis of the human dataset revealed that CXCL9⁺ macrophages represented the high-scoring subpopulation. This subset exhibited significant enrichment in pathways related to interferon induction, M1 macrophage polarization, chemokine signaling, T cell migration and activation, as well as antigen presentation, all of which are beneficial for anti-tumor immune responses (Figures 6F–6J). Kaplan-Meier analysis of the TCGA-LIHC cohort demonstrated that patients with high CXCL9⁺ macrophage scores had significantly longer overall survival (Figure 6K). Analysis of an immunotherapy-treated HCC cohort revealed that a higher abundance CXCL9⁺ macrophage population was associated with improved prognosis and enhanced response to immunotherapy (Figures S9A and S9B). These results highlight the potential role of CXCL9⁺ macrophages in mediating the therapeutic effects of HCC treatment, suggesting that targeting these macrophage subsets could enhance anti-tumor immunity.

SMB nanoparticles reprogrammed tumor-associated macrophages into CXCL9⁺ phenotypes

(A) The UMAP plot shows the distribution of 15 major cell types annotated based on human single-cell sequencing data.

(B) The dot plot displays the expression of cell type-specific markers for 15 cell types, where the size of the circles represents the percentage of expression and the color intensity indicates the average expression level.

(C) The UMAP plot shows the distribution density of the upregulated gene scores across all cells in the drug-treated group, calculated by UCell.

(D) The violin plot displays the expression of upregulated gene scores, calculated by UCell, across 15 major cell types in the drug-treated group.

(E) The Kaplan-Meier (KM) curve shows that samples with high UCell scores have longer overall survival in the single-cell cohort.

(F) The UMAP plot shows the distribution of 10 macrophage subtypes annotated based on human single-cell sequencing data.

(G) The UMAP plot shows the expression of the selected cell-specific marker in macrophages.

(H) UMAP projection illustrates the density distribution of UCell-computed up-regulated gene-signature scores within macrophages from the drug-treated group.

(I) Violin plots show UCell-calculated up-regulated gene-signature scores across 10 macrophage subtypes in the drug-treated group.

(J) The heatmap shows the scores of key functional gene sets in different macrophage subtypes from the collected macrophages.

(K) The KM curve shows that samples with high Mph_CXCL9 scores have longer overall survival in the TCGA-LIHC cohort.

To further validate these findings, PDX (Figure S10A) and patient-derived organoid (PDO) models were established. In the PDX model, liver anatomical imaging demonstrated a significant reduction in tumor volume in the SMB-treated group compared to the Durvalumab (anti-PD-L1) group at day 22 post-treatment (Figures 7A, S10B, and S10C). Western blot analysis revealed robust activation of the NLRP3/Caspase-1/GSDMD pathway in SMB-treated PDXs, indicating pyroptosis induction within tumor tissues (Figure S10D). Notably, enhanced expression of CD86, a marker linked to macrophage activation and antigen presentation, was observed, consistent with trained immunity induction. Flow cytometry quantified increased infiltration of CD86⁺ macrophages and granzyme B⁺ CD8⁺ T cells in the SMB-treated group, alongside reduced TIM3⁺ exhausted CD8⁺ T cells (Figures 7B–7D and S10E–S10H). Immunohistochemistry confirmed elevated CXCL9 levels and reduced PD-L1 level in SMB-treated PDXs (Figures 7E and S10I). In PDOs, immunofluorescence staining showed increased infiltration of CD86⁺ macrophages and CD8⁺ T cells in the SMB group compared to both the control and anti-PD-L1 groups, reflecting macrophage activation and T cell stimulation (Figures 7F and S10J). Bright-field imaging revealed suppressed growth and structural disruption in SMB-treated PDOs, while Calcein-AM/propidium iodide (PI) staining validated substantial cell death (Figure 7G). Western blot analysis further confirmed activation of both NLRP3/Caspase-1/GSDMD and cGAS-STING pathways in SMB-treated PDOs (Figures S10K and S10L). Collectively, these results demonstrate that SMB nanoparticles induce the infiltration of CXCL9⁺ macrophages, enhance T cell functionality, and trigger a potent anti-tumor immune response, effectively inhibiting tumor growth in humanized PDX and PDO models.

SMB nanoparticles exhibit comparable tumor suppression and immune activation in HCC PDX and PDO models

(A) Images of dissected livers from the control and treatment groups to evaluate the tumor growth inhibition effect of SMB nanoformulations derived from engineered probiotic membrane vesicles in a humanized orthotopic PDX model of liver cancer.

(B–D) Flow cytometry analysis of single-cell suspensions from liver tumor tissues of different treatment groups in the humanized orthotopic liver cancer PDX model. (B) Proportion of CD86⁺ macrophages within total macrophages (CD68⁺CD11b⁺). (C) Proportion of granzyme B⁺ cells within CD8⁺ T cells. (D) Proportion of TIM-3⁺ cells within CD8⁺ T cells. Representative plots are shown for each analysis.

(E) Representative immunohistochemistry images of CXCL9 expression in tumor tissues from different treatment groups. Scale bars, 20 μm.

(F) Immunofluorescence staining of liver cancer PDOs treated with durvalumab, SMB nanoformulations, and US irradiation. Organoids were stained for CD8/CD68 (green), CD3/CD86 (red), and DAPI (blue). Representative images are shown. Scale bars, 50 μm.

(G) Bright-field and live/dead cell staining images of liver cancer PDOs treated with Durvalumab, SMB nanoformulations, and US irradiation. Live cells: Calcein-AM (green); dead cells: propidium iodide (PI, red). Representative images are shown. Scale bar, 50 μm.

In summary, SMB nanoparticles induce tumor cell pyroptosis via the NLRP3/Caspase-1/GSDMD pathway and trigger trained immunity, initiating a potent anti-tumor immune response in both humanized HCC-PDX and PDO models. These findings highlight the potential of SMB nanoparticles as an innovative immunotherapy for HCC, offering a promising strategy for clinical translation.

Discussion

Our study presents two major highlights. First, SMB nanoparticles induce pyroptosis in HCC cells via the NLRP3/Caspase-1/GSDMD axis under US-mediated spatiotemporal control and precise targeting by A. muciniphila BMVs, thereby triggering a potent anti-tumor immune response and significant tumor growth suppression. Second, we validate this mechanism across multiple experimental models, including PDX and PDO models, which closely recapitulate human tumor complexity.

As a primary highlight, our study achieves precise induction of pyroptosis, evidenced by characteristic cell swelling, plasma membrane rupture, and the massive release of intracellular contents. These contents include various inflammatory cytokines that activate the immune system, along with the release of a large number of TAAs.⁴⁴

Inducing pyroptosis holds particular promise for treating HCC, which is characterized by extreme spatial and molecular heterogeneity of tumor neoantigens.⁴⁵ This heterogeneity impedes recognition of shared immunodominant antigens and leads to weak T cell responses—a major limitation of immunotherapeutic strategies for HCC.⁴⁶ Although neoantigen-based vaccines can improve immunogenicity, their reliance on patient-specific antigen panels limits scalability.²¹ By contrast, in situ induction of pyroptosis releases endogenous tumor antigens without prior antigen identification, offering a more practical, cost-effective approach. To date, the arsenal of drugs that can effectively induce pyroptosis and activate adaptive immunity remains limited. Chemotherapeutic drugs, such as doxorubicin, require high-dose administration to induce pyroptosis, a strategy often associated with substantial toxicity to healthy tissues. Alternative approaches for inducing tumor cell pyroptosis—including gene therapy, electrostimulation therapy, and radiotherapy—face significant limitations.⁴⁷ These methods are costly and uncontrollable in the treatment process. Moreover, excessive and non-tumor-cell-specific pyroptosis can lead to cytokine release syndrome, which may be life-threatening. These limitations have considerably impeded the further development of pyroptosis-mediated antitumor immunotherapy. In our study, the BMVs derived from A. muciniphila demonstrated efficient and specific targeting toward liver cancer cells. Combined with the US-triggered release of MOF material, this approach enabled tumor cell-specific pyroptosis, significantly reducing the safety risks associated with excessive and off-target effects.

SDT has emerged as a promising tumor treatment modality, offering advantages over traditional photodynamic therapy through its non-invasive nature, low toxicity, and superior US-mediated tissue penetration for deep-seated tumors.⁴⁸ Despite these advantages, the development of SDT still faces certain challenges, particularly in the selection of sonosensitizers. MOF, as emerging sonosensitizers, offer high surface area, tunable pore structure, and good biocompatibility, making them promising for SDT. Notably, MOF absorb photons through organic linkers to activate metal ions or clusters, acting as alternative semiconductors in photocatalytic reactions via the ligand-to-metal charge transfer (LMCT) mechanism. This concept of LMCT, along with the phenomenon of sonoluminescence, suggests that MOF materials with photocatalytic properties may have substantial potential in SDT applications, although mechanistic details require further study.⁴⁹ The extensive research on photocatalytic MOFs in other fields provides a rich candidate pool for developing sonosensitizer materials.⁵⁰

This study is designed to overcome key challenges in HCC immunotherapy, including insufficient TAA release, TIME, poor antigen uptake and presentation, and irreversible T cell exhaustion. To address these challenges, we strategically target three critical nodes of the antitumor immune response. First, we induce pyroptosis in tumor cells to liberate TAAs into the microenvironment. Second, we reprogram innate immunity through trained immunity to boost antigen uptake and presentation, facilitating the priming and activating tumor-specific effector T cells. Third, we deliver PD-L1 siRNA to block the corresponding checkpoint, rescue exhausted T cells, and sustain their infiltration and cytotoxic function within the HCC niche. This end-to-end therapeutic cascade establishes a self-reinforcing immune-activation circuit: pyroptosis-driven antigen exposure, trained immunity-enhanced antigen presentation, and siRNA-mediated checkpoint blockade. Such interventions could effectively circumvent the compensatory resistance seen in single-target therapies. By orchestrating a three-stage cascade of interventions, we establish a fully integrated platform to drive a positive cycle of antitumor immunity against HCC.

Strikingly, SMB selectively reprograms TAMs into CXCL9⁺ macrophages. CXCL9 exhibits significantly high expression in both macrophage polarization and trained immunity. Previous studies have revealed that CXCL9⁺ macrophage subsets emerge during polarization, and these subsets participate in immune regulation by secreting cytotoxic factors such as CXCL9.⁵¹^,⁵²^,⁵³^,⁵⁴ Similarly, upregulation of CXCL9 is commonly observed in trained immune responses.⁵⁵ Macrophage polarization and trained immunity share overlapping regulatory mechanisms. In terms of driving factors, CD8⁺ T cell-derived IFN-γ not only induces classical M1 polarization to generate CXCL9⁺ macrophages but also mediates trained immunity through epigenetic reprogramming.⁵⁶^,⁵⁷^,⁵⁸^,⁵⁹ At the metabolic level, the coordinated activation of the mTOR-HIF signaling pathway serves as a central regulatory node—this pathway not only regulates proinflammatory phenotypic switching in M1 polarization but also supports trained immunity by enhancing aerobic glycolysis, thereby providing the metabolic basis for sustained immune activation. This indicates that these two immune processes may share core signaling pathways and energy metabolic networks.⁶⁰^,⁶¹ This suggests that CXCL9⁺ macrophages induced by SMB serve as a critical bridge linking macrophage polarization and trained immunity. The CXCL9⁺ macrophages are essential for CD8⁺ T cell recruitment and exhibit a pivotal role in determining ICI efficacy,⁶² suggesting that the engineered AKK-derived SMB nanoparticles represent a trained immunity inducer with translational potential for HCC immunotherapy. Beyond its effects on macrophages, SBM treatment also enriches monocytes and DCs—key cellular components of trained immunity.⁶³ The potential immunomodulatory effects of these cell subtypes to SMB-mediated immunomodulation remain to be investigated.

As a second highlight, we established and employed PDX and PDO models to evaluate SMB efficacy, which recapitulate tumor heterogeneity and drug response dynamics observed in HCC. PDX models preserve the histopathological, genetic, and epigenetic features of original tumors while mimicking clinical therapeutic responses, making them valuable for predicting patient-specific drug sensitivity and elucidating resistance mechanisms.⁶⁴ Humanized mouse PDX models establish a fully human tumor-immune system where diverse myeloid and lymphoid cells infiltrate the tumor, overcoming previous limitations in reconstructing key immune populations like macrophages, DCs, mast cells, and granulocytes.⁶⁵ This dual-humanized platform supports preclinical efficacy testing of next-generation immunotherapies and offers developers a broader array of therapeutic targets and strategies, accelerating the discovery of differentiated cancer immunotherapies.

More recently, organoids have emerged as innovative preclinical models. PDOs are three-dimensional, self-organizing cultures that faithfully reproduce the architecture of the source tumor. Large-scale studies confirm that PDO drug sensitivity assays strongly correlate with clinical chemotherapy outcomes, establishing them as valuable research tools.⁶⁶ Crucially, advances in organoid culture techniques now enable the retention of immune components, creating immune-competent tumor organoids for preclinical evaluation of immunotherapies.⁶⁷ Our team has developed a rapid, 1-week protocol to generate immune-preserving tumor organoids from patient samples.⁶⁸ Applying this technology to HCC, we aim to establish liver cancer PDOs that closely mimic the native tumor microenvironment, thereby providing an additional preclinical platform for assessing the antitumor efficacy of SMB nanoparticle.

Limitations of the study

Our study has certain limitations that warrant further investigation. First, the mechanisms by which SMB nanoparticles activate NLRP3 inflammasome assembly warrant deeper investigation. scRNA-seq could better resolve nanoparticle-driven reprogramming across immune subsets. Second, the epigenetic basis of macrophage training remains unexplored. Investigating dynamic epigenetic changes and their interplay with metabolic rewiring would clarify the mechanisms of trained immunity. Finally, while promising in initial organoid models (n = 3), future studies should include diverse HCC etiologies (e.g., hepatitis B virus, hepatitis C virus, and NAFLD) and molecular subtypes (e.g., CTNNB1/TP53-mutant) to confirm broad applicability. Addressing these gaps will enhance the translational potential of SMB nanoparticles for HCC immunotherapy.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Fubing Wang (wfb20042002@sina.com).

Materials availability

This study did not generate new unique reagents.

Data and code availability

•
RNA-seq data have been deposited at Genome Sequence Archive in National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences, and publicly available as of the date of publication (no. PRJCA041597; https://ngdc.cncb.ac.cn/bioproject/browse/PRJCA041597).
•
This paper does not report original code.
•
Any additional information required to reanalyze the data reported in this work paper is available from the lead contact upon request.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (no. 12375349 and 32422042); the Basic and Clinical Medical Research Joint Fund of Zhongnan Hospital, Wuhan University (no. ZNLH202209); the Medical Science and Technology Innovation Platform Support Project of Zhongnan Hospital, Wuhan University (no. PTXM2025033); and the Natural Science Foundation of Hubei Province (no. 2025AFB689).

Author contributions

L.H. and Y.R. conceptualized the study and experimental design. L.H. performed probiotic membrane vesicle extraction, engineered siRNA-MOF@BMV nanocomplexes, and conducted in vivo therapeutic evaluations. L.H. and M.G. developed the LC-MS/MS quantification method for N-acetylgalactosamine and executed Cy3-NHS vesicle labeling. L.H. and M.L. established all cell culture systems, performed gene knockdown experiments, and validated nanoparticle cytotoxicity. L.H. and F.L. conducted single-cell RNA-seq and bioinformatics analyses. L.H. and W.Z. executed histopathological evaluations (H&E and TSA staining) and western blot validations. L.H. and Y.H. processed clinical samples and established patient-derived organoids. L.H. and X.H. provided surgical specimens and clinical data interpretation. L.H. and J.H. developed the humanized mouse system. D.Z., C.Y., and F.W. provided strategic mentorship and project oversight. All authors participated in manuscript revision and approved the final version.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies

BD Horizon™ BV421 Hamster Anti-Mouse CD3e	BD Biosciences	Cat# 562600; RRID:AB_11153670
BD Horizon™ BV480 Hamster Anti-Mouse CD11c	BD Biosciences	Cat# 565627; RRID:AB_2739309
BD Horizon™ BV510 Hamster Anti-Mouse CD69	BD Biosciences	Cat# 563030; RRID:AB_2737963
BD Horizon™ BV605 Rat Anti-Mouse CD62L	BD Biosciences	Cat# 563252; RRID:AB_2738098
BD OptiBuild™ BV650 Mouse Anti-Mouse CD366	BD Biosciences	Cat# 747623; RRID:AB_2744189
BD OptiBuild™ RY743 Rat Anti-Mouse CD44	BD Biosciences	Cat# 772976
BD Pharmingen™ FITC Rat Anti-CD11b	BD Biosciences	Cat# 561688; RRID:AB_10898180
BD Pharmingen™ PE Rat Anti-Mouse CD103	BD Biosciences	Cat# 561043; RRID:AB_396732
BD Horizon™ PE-CF594 Hamster Anti-Mouse CD279 (PD-1)	BD Biosciences	Cat# 562523; RRID:AB_2737634
BD Pharmingen™ PE-Cy™7 Rat Anti-Mouse CD127	BD Biosciences	Cat# 560733; RRID:AB_1727424
BD Pharmingen™ APC Rat Anti-Mouse CD8a	BD Biosciences	Cat# 561093; RRID:AB_398527
BD Pharmingen™ Alexa Fluor® 647 Rat Anti-Mouse CD206	BD Biosciences	Cat# 565250; RRID:AB_2739133
BD Horizon™ APC-R700 Rat Anti-Mouse F4/80	BD Biosciences	Cat# 565787; RRID:AB_2869711
BD Pharmingen™ APC-Cy™7 Rat Anti-Mouse CD45	BD Biosciences	Cat# 561037; RRID:AB_396774
Fixation/Permeablization Kit	BD Biosciences	Cat# 554715; RRID:AB_2869009
BD Pharmingen™ Purified Rat Anti-Mouse CD16/CD32	BD Biosciences	Cat# 567021; RRID:AB_2870011
Brilliant Violet 711™ anti-mouse/human KLRG1 (MAFA) Antibody	BioLegend	Cat# 138427; RRID:AB_2629721
PerCP/Cyanine5.5 anti-human/mouse Granzyme B Recombinant Antibody	BioLegend	Cat# 372211; RRID:AB_2728378
Anti-TMS1/ASC antibody	Abcam	Cat# ab175449; RRID:AB_3096354
Anti-Caspase-4 antibody	Abcam	Cat# ab22687; RRID:AB_2070060
Anti-Caspase-11 antibody	Abcam	Cat# ab22684; RRID:AB_447254
Anti-NLRP3 antibody	Abcam	Cat# ab4207; RRID:AB_955792
Caspase 5 antibody	GeneTex	Cat# GTX31701
Caspase-1 Rabbit pAb	ABclonal	Cat# A0964; RRID:AB_2757485
AIM2 Rabbit pAb	ABclonal	Cat# A3356; RRID:AB_2765071
NLRC4 Rabbit pAb	ABclonal	Cat# A7382; RRID:AB_2767914
GAPDH Rabbit mAb	ABclonal	Cat# A19056; RRID:AB_2862549
GSDME Polyclonal antibody	Proteintech	Cat# 13075-1-AP; RRID:AB_2093053
GSDMA Polyclonal antibody	Proteintech	Cat# 30354-1-AP; RRID:AB_3086299
GSDMD Monoclonal antibody	Proteintech	Cat# 66387-1-Ig; RRID:AB_2881763
Gasdermin B (GSDMB) Antibody	Abbexa	Cat# abx136072
GSDMC Rabbit pAb	ABclonal	Cat# A14550; RRID:AB_2769694
Anti-STING antibody	Abcam	Cat# ab181125; RRID:AB_2916053
Anti-F4/80 antibody	Abcam	Cat# ab111101; RRID:AB_10859466
Anti-CD11c antibody	Abcam	Cat# ab33483; RRID:AB_726084
Anti-PD-L1 antibody	Abcam	Cat# ab205921; RRID:AB_2687878

Bacterial and virus strains

Akkermansia muciniphila	American Type Culture Collection	N/A
Bifidobacterium longum	American Type Culture Collection	N/A
Lactobacillus reuteri	American Type Culture Collection	N/A
Lactobacillus rhamnosus	American Type Culture Collection	N/A

Biological samples

HCC patient samples	This paper	N/A
Mouse liver	This paper	N/A
Mouse spleen and heart	This paper	N/A
Mouse lymph node	This paper	N/A

Chemicals, peptides, and recombinant proteins

H₂TCPP	Xi’an Qiyue Biotechnology Co.	N/A
Zn(NO₃)₂·6H₂O	Sigma-Aldrich	CAS 10196-18-6
Benzoic acid	Sigma-Aldrich	CAS 65-85-0
N,N-dimethylformamide	Sigma-Aldrich	CAS 68-12-2
Acetonitrile	Sigma-Aldrich	CAS 75-05-8
Methanol	Sigma-Aldrich	CAS 67-56-1
Dimethyl sulfoxid	Sigma-Aldrich	CAS 67-68-5
Gadolinium(III) chloride hexahydrate	Sigma-Aldrich	CAS 13450-84-5
Glutaraldehyde	Sigma-Aldrich	CAS 111-30-8
Paraformaldehyde	Sigma-Aldrich	CAS 30525-89-4
Cyanine3 NHS	ChemicalBook	CAS 146368-16-3
Cy5.5 NHS ester	Macklin	CAS 1469277-96-0
Human EGF Recombinant protein	Sangon	D145467
Advanced DMEM F12 medium	Gibco™	12634010
RPMI 1640 medium	Gibco™	11875093
B27 Supplement (50×)	Gibco™	17504044
GlutaMAX (100×)	Gibco™	35050061
N-acetylcysteine	Sigma-Aldrich	CAS 38520-57-9
Nicotinamide	Sigma-Aldrich	CAS 98-92-0
Prostaglandin E2	Sigma-Aldrich	CAS 363-24-6
HEPES (1M)	Gibco™	15630080
siRNA	Sangon	N/A

Critical commercial assays

Mouse CXCL9 enzyme-linked immunoassay kit	HUABlO	Cat# EM0015
Pierce™ Dilution-Free™ Rapid Gold BCA Protein Quantification Kit	Thermo	Cat# A55863
Mouse IFN-β enzyme-linked immunoassay kit	HUABlO	Cat# EM0040

Deposited data

Raw RNA-seq	This paper	National Center for Biological Information:PRJCA041597; https://ngdc.cncb.ac.cn/bioproject/browse/PRJCA041597

Experimental models: Cell lines

Hepa 1–6 cells	China Center for Type Culture	N/A
H22 cells	China Center for Type Culture	N/A
RAW 264.7 cells	China Center for Type Culture	N/A
4T1 cells	China Center for Type Culture	N/A
Huh-7 cells	China Center for Type Culture	N/A

Experimental models: Organisms/strains

C57BL/6 mice	GemPharmatech Co., Ltd	N/A
NOG-EXL mice	Shanghai Model Organisms	N/A
NOD/SCID mice	GemPharmatech Co., Ltd	N/A

Oligonucleotides

siRNA	Thermo Fisher Scientific	GTAGTAGATGTTACAATTT

Software and algorithms

GraphPad Prism 10.0	Graphpad software	https://www.graphpad.com
ImageJ	N/A	https://imagej.net/ij/.
FlowJo	FlowJo 10.4	https://www.flowjo.com/solutions/flowjo
R-4.3.1	N/A	https://www.r-project.org

Open in a new tab

Experimental model and study participant details

Human specimens

Fresh hepatocellular carcinoma tissue specimens were collected 8 patients (6 male and 2 female) who underwent surgery at the Department of Hepatobiliary Surgery, Zhongnan Hospital of Wuhan University (Wuhan, China) between January 2023 and December 2024. The study protocol was approved by the Medical Ethics Committee of Zhongnan Hospital of Wuhan University (approval no. 2023073K), and all human-tissue-related procedures were performed in accordance with the 2013 revision of the Declaration of Helsinki. Written informed consent was obtained from each patient prior to tissue collection.

Mice

All animal experiments were approved by the Experimental Animal Welfare Ethics Committee of Zhongnan Hospital of Wuhan University (approval no. 2023185), and conducted in accordance with the ARRIVE 2.0 guidelines and relevant institutional policies. C57BL/6 and NOD/SCID mice were procured from GemPharmatech Co., Ltd. NOG-EXL mice were obtained from the Shanghai Model Organisms Center and subsequently maintained within the institutional barrier facility. All mice were housed in a specific pathogen-free (SPF) environment and and provided ad libitum access to standard chow. In the transplantation model, 6- to 8-week-old female mice were utilized and co-injected with tumor cells. The mice were enrolled at 15 weeks of age and maintained on a standard diet. When the tumor diameter reached 2 cm, the mice were euthanised for tumor analysis or survival assessment.

Cell lines

The Hepa 1–6, H22, RAW 264.7, 4T1 and Huh-7 cell lines were purchased from the China Center for Type Culture Collection (CCTCC, Wuhan, China). The identity of all cell lines was authenticated by short tandem repeat profiling, and mycoplasma contamination was not detected. The Hepa 1–6, RAW 264.7 and Huh-7 cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM), supplemented with 10% heat-inactivated fetal bovine serum (FBS), 100 U/ml penicillin and 100 μg/mL streptomycin, at 37°C in a 5% CO₂ atmosphere. The H22 and 4T1 cells were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium, supplemented with 10% FBS, 100 U/ml penicillin and 100 μg/mL streptomycin, under identical conditions. All cell lines were authenticated by morphology and growth rate and were mycoplasma-free.

The syngeneic H22 orthotopic model

Female C57BL/6 mice (6–8 weeks old, SPF) were subjected to a single intrahepatic injection of 1 × 10⁶ H22 cells suspended in 30 μL of phosphate-buffered saline (PBS). Tumors were permitted to develop for 14 days, after which the mice were randomly assigned to treatment groups. Beginning on day 0, bacterial-membrane-vesicle nano-formulations were administered via the tail vein every four days, followed by ultrasound (US) irradiation 24–48 h after each injection. On day 22, tumour-bearing livers were excised for tumor volume quantification, metastasis assessment, histological evaluation, RNA sequencing, and cytokine analysis.

Patient-derived xenograft (PDX) model in humanized mice

Female NOG-EXL mice (4–8 weeks old) were subjected to 1.5 Gy total-body irradiation, followed by the intravenous transplantation of 2.5 × 10⁵ human CD34⁺ cells into the tail vein. Eight weeks later, humanization was confirmed (hCD45⁺ > 25%), and fragments (∼1 mm³) of F3-generation human HCC-PDX were surgically implanted into the right hepatic lobe. Tumor growth was monitored via ultrasound imaging. Treatment was initiated when tumors reached 1–2 cm³ and consisted of the same nano-formulation plus ultrasound cycles employed in the syngeneic model. At the experimental endpoint, livers were excised for tumor burden assessment, human immune infiltration analysis, and multi-omics profiling; small intestinal and colorectal tissues were additionally procured as non-tumour controls.

Method details

Extraction of intestinal probiotic membrane vesicles

A.muciniphila was cultured anaerobically at 37°C for 48 h in brain-heart infusion (BHI) medium supplemented with 0.5% mucin and 0.05% L-crystallites. After centrifugation, the cells were resuspended in anaerobic PBS and the concentration was determined by measuring the optical density at 600 nm. B. longum was cultured anaerobically at 37°C in BHI medium supplemented with 5 μg/mL chloroferrioxetin and 2.5 μg/mL vitamin K. L. reuteri was cultured in BHI medium containing 0.05% L-crystallites and 0.5% mucin. L. reuteri and L. rhamnosus were cultured anaerobically at 37°C in MRS medium.

Cultures were centrifuged at 10,000 × g for 30 min. The supernatant was sequentially filtered through 0.45 μm and 0.22 μm membranes and concentrated with a 100 kDa Centricon Plus-70 unit. Vesicles were pelleted by ultracentrifugation at 150,000 × g for 2 h at 4°C and subsequently washed in sterile PBS. After removal of the supernatant, the vesicles were resuspended in sterile PBS, filtered through a 0.22 μm membrane, and stored at −80°C. Vesicle morphology was analyzed using transmission electron microscopy (TEM), and concentration and size were determined by nanoparticle tracking analysis (NTA).

Bacterial external vesicle staining and cell-targeted uptake assay

Bacterial vesicles were labeled with Cy3-NHS at 4°C for 4 h in a total volume of 2 mL. The labeled vesicles were concentrated with a 100 kDa ultrafiltration device, centrifuged at 3,000 × g for 5 min, and washed twice with PBS. Vesicles were filtered through a 0.22 μm membrane and stored at 4°C. For uptake assays, cells were seeded to 60% confluence and incubated with the labeled vesicles or nanoparticles. Following incubation, cells were analyzed by flow cytometry or fluorescence microscopy.

Quantification of N-acetylgalactosamine by triple quadrupole mass spectrometry-liquid chromatography

Aliquots (100 μL) of the probiotic vesicle suspension were transferred into 1.5 mL tubes and combined with an equal volume of acetonitrile:methanol (1:1, v/v). The samples were vortex-mixed, sonicated on ice for 10 min, and stored at −20°C for 1 h prior to centrifugation at 12,000 × g for 15 min at 4°C. The supernatants were dried under N₂ in a vacuum freeze-dryer, reconstituted in acetonitrile:water (1:1, v/v), and re-centrifuged at 12,000 × g for 15 min at 4°C. The resulting supernatants were transferred into LC-MS/MS autosampler vials. N-acetylgalactosamine standards (0.001–2 ppm in acetonitrile:water, 1:1, v/v) were prepared and analyzed by LC-MS/MS.

Synthesis of engineered probiotic membrane vesicles

H₂ TCPP (100 mg), Zn(NO₃)₂·6H₂O (300 mg), and benzoic acid (2.8 g) were dissolved in 200 mL of DMF at 90°C for 5 h with continuous stirring at 300 rpm. The MOF was precipitated, collected by centrifugation at 13,000 × g for 30 min, and washed with DMF. siRNA was loaded onto the MOF at concentrations ranging from 50 to 500 nM prior to MOF addition, and loading efficiency was verified by UV-Vis spectrophotometry. Engineered vesicles were prepared by sequential extrusion of bacterial vesicles suspended in saline through a 400 nm membrane (10 cycles), followed by addition of the siRNA-MOF complex and extrusion through a 200 nm membrane (15 cycles). The samples were centrifuged at 10,000 × g for 10 min at 4°C, washed with saline, and resuspended in saline for subsequent use.

Characterization of engineered probiotic membrane vesicles

Nanoparticle morphology and elemental distribution were evaluated by depositing samples onto carbon-coated copper grids, staining with 1% uranyl acetate, and imaging via TEM. Crystal structure was determined by powder X-ray diffraction. Particle size and zeta potential were quantified by dynamic light scattering (DLS). Colloidal stability was evaluated by incubating samples in PBS or DMEM supplemented with 10% FBS at ambient temperature, with particle size monitored by DLS over time. UV-visible absorption spectra were acquired with a UV-Vis spectrophotometer.

Cell culture

Mouse hepatocellular carcinoma cell lines Hepa 1–6 and H22, murine monocyte-macrophage leukemia cell line RAW 264.7, murine breast carcinoma cell line 4T1, and human hepatocellular carcinoma cell line Huh-7 were acquired from the CCTCC. Hepa 1–6 and Huh-7 cells were maintained in DMEM supplemented with 10% FBS, 100 U mL⁻¹ penicillin, and 100 μg mL⁻¹ streptomycin; 4T1 cells were cultured in RPMI-1640 medium containing the same additives; H22 and RAW 264.7 cells were cultured in DMEM supplemented with 10% FBS, 100 U mL⁻¹ penicillin, and 100 μg mL⁻¹ streptomycin. Huh-7 cells were additionally maintained in pyruvate-free DMEM. H22 and RAW 264.7 cells, which grow in suspension, were handled accordingly. All cell lines were incubated at 37°C under 5% CO₂, sub-cultured at a 1:3-1:6 ratio, and cryopreserved in basal medium supplemented with 10% dimethyl sulfoxide (DMSO) and 20% FBS.

Cells were seeded at 50% confluence in antibiotic-free medium 24 h prior to transfection. To prepare the Lipo3000-siRNA complex, 2 μL of Lipofectamine 3000 was gently vortexed and diluted in 100 μL serum-free medium, followed by incubation at ambient temperature for 5 min siRNA (provided by Shenggong Biotech) was diluted to 40 pM in 100 μL of serum-free medium. After the initial incubation, the diluted Lipofectamine 3000 was gently combined with the siRNA solution and incubated for an additional 20 min at ambient temperature to allow complex formation. The resulting complex was added to the cell culture plates, gently agitated, and incubated at 37°C for 6 h. Subsequently, the transfection medium was aspirated, replaced with complete growth medium supplemented with serum and antibiotics, and incubation was continued until analysis. Knockdown efficiency was evaluated 72h post-transfection, and cells were harvested 48h post-transfection for downstream analyses.

Uptake efficiency of nanoparticles in various cell types

To address the uptake efficiency of nanoparticles in various cell types and the promoting effect of BMVs on Hepa 1–6 cell targeting, we conducted the following experiments. First, we incubated Cy3 dye with SM nanoparticles under stirring conditions to label the nanoparticles. Simultaneously, we labeled SB and SMB nanoparticles with Cy3-NHS to prepare Cy3-labeled SM, SB, and SMB nanoparticles. These labeled nanoparticles were then co-incubated with Hepa 1–6 cells for 2 h at 37°C in 5% CO₂. Cells without nanoparticles served as the control group. After incubation, the cells were harvested, washed with PBS, and resuspended in 1 mL of PBS. The uptake of different nanoparticles by Hepa 1–6 cells was detected using flow cytometry by measuring the fluorescence intensity of Cy3. In a separate experiment, we co-incubated Cy3-labeled SMB nanoparticles with Hepa 1–6 cells, 4T1 cells, RAW264.7 macrophages, and M1-and M2-phenotype bone marrow-derived macrophages that were induced and differentiated from bone marrow-derived macrophages. Cells without nanoparticles were used as the control group. The uptake of SMB nanoparticles by these different cell types was also detected using flow cytometry by measuring the fluorescence intensity of Cy3.

Experimental steps for ultrasound intensity and release efficiency of SMB

To evaluate the ultrasound-triggered release efficiency of SMB nanoparticles at varying acoustic intensities, Hepa 1–6 cells were seeded in 6-well plates at 1 × 10⁶ cells per well and incubated at 37°C under 5% CO₂ for 24 h to enable adherence and proliferation. SMB nanoparticles were dispersed in DMEM to 500 μg mL⁻¹, after which the culture medium was aspirated and replaced with 2 mL of the nanoparticle suspension. Cells were incubated with the nanoparticle suspension for 2 h at 37°C under 5% CO₂. The ultrasound device was calibrated to deliver intensities of 0.5, 1.0, 1.5, 2.0, and 2.5 W cm⁻², and cells were sonicated for 5 min at each intensity. Control wells were left unsonicated. Following sonication, the nanoparticle-containing medium was exchanged for fresh DMEM, and cells were incubated for a further 24 h at 37°C under 5% CO₂ to permit cytotoxic effects to manifest. Cells were subsequently detached with 0.25% Trypsin-EDTA, washed twice with PBS, and resuspended in 1 mL of PBS. SYTOX Deep Red staining solution (1 μL) was added to each suspension, followed by incubation for 15 min at ambient temperature in the dark. Stained cells were analyzed by flow cytometry (excitation/emission: 640/660 nm) to quantify mortality, and the percentage of dead cells was recorded.

Cytotoxicity assessment of SMB

Wild-type or knockdown cells were seeded into 96-well plates at 5 × 10⁵ cells mL⁻¹ and subsequently exposed to nanoparticles. After 2h, cells were sonicated (1.0 MHz, 1.5 W/cm2, 50% duty cycle, 5 min); where indicated, inhibitors of ferroptosis, necroptosis, and pyroptosis were added. Cells were incubated for an additional 24 h, after which 200 nM SYTOX Deep Red was added to the culture medium. Imaging was performed immediately, and the number of red-fluorescent cells was quantified by automated image-analysis software.

Observation of pyroptosis-induced cell morphology by engineered probiotic membrane vesicles

Cells were seeded at 5 × 10⁵ cells mL⁻¹ and subsequently exposed to nanoparticles. Following 2 h of ultrasonic exposure, cultures were maintained for an additional 24 h. After incubation, cells were washed twice with PBS and bright-field images were acquired using a light microscope. For ultrastructural analysis, cells were fixed in 0.2 M cacodylate buffer (pH 7.4) containing 2.5% glutaraldehyde, 2.5% paraformaldehyde, and 0.05% picric acid for ≥2 h, washed twice with deionized water and once with 0.1 M maleate buffer (MB), and then incubated for 1 h in MB supplemented with 1% uranyl acetate. After two additional washes in deionized water, samples were dehydrated in a graded ethanol series (50%, 70%, 90%, and two changes of 100%, 10 min each). Samples were subsequently treated with propylene oxide for 1 h and infiltrated with a 1:1 mixture of propylene oxide and TAAB Epon. The following day, samples were embedded in TAAB Epon and polymerized at 60°C for 48 h. Finally, 60 nm ultrathin sections were cut with an ultramicrotome, mounted on copper grids, stained with lead citrate, and imaged by TEM.

Enzyme-linked immunosorbent assay

Concentrations of interleukin-18 (IL-18), interferon-β (IFN-β), and C-X-C motif chemokine ligand 9 (CXCL9) were quantified by enzyme-linked immunosorbent assay. Per the manufacturer’s protocol, cell supernatants were collected by centrifugation. All samples and standards were assayed in duplicate. Required reagents-wash buffer, detection buffer, detection antibody, HRP-conjugated secondary antibody, and standards-were prepared according to the kit insert. Standards, samples, or controls were loaded into designated wells; blank wells received dilution buffer only. Detection antibody, HRP-conjugated secondary antibody, 3,3′,5,5′-tetramethylbenzidine substrate, and stop solution were sequentially added. Absorbance was measured at 450 nm on a microplate reader, and cytokine concentrations were interpolated from the standard curve.

Enzyme-linked immunosorbent spot assay

To assess tumor antigen-specific T cell responses, IFN-γ enzyme-linked immunospot (ELISpot) assays were performed with T cells isolated from H22 orthotopic tumors of mice treated with control or SMB nanoparticles. Tumor-infiltrating T cells were enriched by CD3-positive magnetic-activated cell sorting. ELISpot assays were carried out using a commercial mouse IFN-γ ELISpot kit in accordance with the manufacturer’s protocol. 96-well ELISpot plates were pre-coated overnight with anti-mouse IFN-γ monoclonal antibody, then blocked with complete culture medium. T cells were pre-stimulated for 4 h and plated at 2 × 10⁵ cells per well. Wells were configured as follows: experimental (H22 cell lysate), negative (culture medium), positive (PHA), and background (NIH/3T3 + RAW264.7 cell lysate). Following 24h of incubation, IFN-γ-producing cells were enumerated by ELISpot according to the manufacturer’s instructions.

Establishment of mouse hepatocellular carcinoma model and drug administration strategy

A murine orthotopic hepatocellular carcinoma model was established by intrahepatic injection of 1 × 10⁶ H22 cells, and tumor formation was confirmed two weeks later. For in vivo imaging, Cy5.5-NHS-labeled SMB nanoformulations were administered intravenously to tumor-bearing mice; real-time fluorescence was monitored at designated intervals, and ex vivo organ fluorescence was quantified 24h post-injection. Antitumor studies were initiated on day 0, and mice were randomly allocated to five groups (n = 6): Control, BMV, siRNA@BMV (SB), siRNA-MOF (SM), and siRNA-MOF@BMV (SMB). Nanoformulations were administered intravenously every 4 days and immediately followed by ultrasound exposure 24–48 h post-injection (1.0 MHz, 2.5 W/cm2, 50% duty cycle, 5 min); animal health was continuously monitored. On day 22, tumor volumes were measured by ultrasonography, and postmortem examinations were performed to assess ascites, pulmonary, and lymphatic metastases. For macrophage depletion, a 2 mg mL⁻¹ GdCl₃ solution was prepared, sterile-filtered (0.22 μm), aliquoted, and stored at 4°C for up to one week; administrations were performed twice weekly at 10 mg kg⁻¹ via the tail vein. All animal procedures were approved by the IACUC of Zhongnan Hospital, Wuhan University.

Transcriptome RNA-Seq analysis of mouse tumor tissues

A reference transcriptome of murine tumor tissues was generated by RNA-seq, encompassing RNA extraction, library preparation, sequencing, gene set enrichment analysis (GSEA), functional enrichment, CIBERSORTx deconvolution, and single-sample GSEA (ssGSEA). Total RNA was extracted with TRIzol reagent, and RNA integrity and concentration were assessed with a NanoDrop 2000 spectrophotometer and an Agilent 2100 Bioanalyzer. Libraries were prepared with the VAHTS mRNA-seq kit and sequenced on an Illumina NovaSeq 6000 platform by Shanghai Ouyi Biotechnology Co. Raw reads were processed with fastp, aligned to the mouse reference genome using HISAT2, and quantified with HTSeq-count. Principal component analysis was conducted to evaluate sample heterogeneity. GSEA was performed with the limma package and clusterProfiler; pathways were considered significantly enriched at false discovery rate <0.05 and |NES| > 1. Functional enrichment was further assessed using limma and clusterProfiler against the org.Mm.e.g.,.db background. Immune cell fractions were estimated by CIBERSORTx using a murine signature matrix with 100 permutations for robustness. ssGSEA was executed via the GSVA package against MSigDB gene sets; human gene symbols were converted to mouse orthologs with Homologene when required. The immune infiltration landscape was characterized by CIBERSORTx (https://cibersortx.stanford.edu/) deconvolution employing a murine signature matrix (https://www.nature.com/articles/srep40508) with 1,000 permutations.

Procedure for HE staining of mouse tumor tissue

Haematoxylin and eosin (H&E) staining of murine tumor sections was performed by sequential dewaxing, rehydration, haematoxylin staining, eosin staining, dehydration, clearing, mounting, and microscopic examination. Sections were dewaxed in xylene (two changes, 5 min each) and rehydrated through descending ethanol gradients (100%, 95%, 80%, 70%, 50%, each 2 min) followed by a brief rinse in distilled water. Sections were stained with Harris haematoxylin for 5 min, differentiated in 1% acid alcohol (10 s), and blued in 0.1% ammonia water (30 s). Sections were counterstained with eosin Y for 2 min. Sections were dehydrated through ascending ethanol gradients (70%, 80%, 95%, 2 × 100%, each 2 min), cleared in xylene (2 × 5 min), and mounted with neutral balsam. Mounted sections were examined under a light microscope, and digital images were acquired for downstream analysis.

Flow cytometry of single cell suspensions of mouse tumor tissue

Single-cell suspensions were prepared from murine tumor tissue or cultured cells. Tissue fragments were dissociated in 15 mL of cell-staining buffer and centrifuged at 350 × g for 5 min. Supernatants were aspirated, and pellets were resuspended in 3 mL of 1× erythrocyte-lysing buffer and incubated on ice for 5 min. After addition of 10 mL of cell-staining buffer, suspensions were centrifuged at 350 × g for 5 min, and supernatants were aspirated to complete one wash cycle. Cells were resuspended in 0.5 mL of cell-staining buffer containing 5 μL (0.25 μg) of viability dye per 10⁶ cells and incubated on ice for 3–5 min in the dark. Suspensions were blocked with 100 μL Fc-receptor-blocking solution per 10⁶ cells for 5–10 min on ice. Surface staining was performed by incubating cells with optimally diluted fluorochrome-conjugated primary antibodies on ice for 20–30 min in the dark. Cells were washed twice with ≥2 mL of cell-staining buffer, fixed in 0.5 mL of fixation buffer for 30 min at ambient temperature in the dark, then centrifuged at 350 × g for 5 min and supernatants were aspirated. Fixed cells were permeabilised in 1× permeabilisation buffer and centrifuged at 350 × g for 5–10 min. Intracellular staining was carried out by resuspending fixed, permeabilised cells in residual buffer containing fluorochrome-conjugated antibodies or matched isotype controls and incubating for 30 min at ambient temperature in the dark. Cells were washed twice with 2 mL of intracellular-staining wash buffer, resuspended in 0.5 mL of cell-staining buffer, and analyzed by flow cytometry using appropriate controls.

Western blotting

Protein immunoblotting was performed via sequential sample preparation, SDS-PAGE, electrotransfer, immunodetection, and image acquisition. Samples were prepared as cell or tissue lysates and quantified by bicinchoninic acid assay. Adherent cells were lysed in ice-cold extraction buffer, sonicated, and clarified by centrifugation at 4°C; supernatants were aliquoted and stored at −80°C. Tissues were snap-frozen, homogenised in lysis buffer, and incubated at 4°C for 2 h with gentle agitation. Homogenates were centrifuged at 12,000 × g for 15 min at 4°C. Lysates were quantified, reduced with 5% β-mercaptoethanol, denatured at 95°C for 5 min, and stored at −20°C. Twenty to thirty micrograms of protein were resolved on 10% SDS-PAGE gels at 100 V for 90 min, electrotransferred onto polyvinylidene difluoride membranes, blocked with 5% non-fat milk, and probed sequentially with primary and HRP-conjugated secondary antibodies. Signals were visualised by enhanced chemiluminescence.

Immunofluorescence staining and on immunohistochemical staining tumor tissue

Tumor tissues were fixed in 4% paraformaldehyde, paraffin-embedded, and sectioned at 4 μm. Sections were deparaffinised in xylene and rehydrated through graded ethanol. Antigen retrieval was carried out by microwave heating in 10 mM citrate buffer (pH 6.0). For immunofluorescence, sections were blocked with 5% (w/v) bovine serum albumin at ambient temperature for 1 h, incubated with primary antibodies at 4°C overnight, and then with fluorophore-conjugated secondary antibodies for 1 h at ambient temperature. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole and sections were mounted in anti-fade medium. Images were acquired using a laser-scanning confocal microscope. Endogenous peroxidase activity was quenched by incubation with 3% (v/v) H₂O₂ for 10 min at ambient temperature. Blocking and primary antibody incubation were performed as described for immunofluorescence. Sections were incubated with horseradish peroxidase-conjugated secondary antibodies for 1h at ambient temperature and developed with 3,3′-diaminobenzidine. Sections were counterstained with haematoxylin, dehydrated through graded ethanol, cleared in xylene, and mounted with neutral balsam. Images were acquired using a bright-field microscope.

Construction of the PDX model in humanised mice

Humanised mice were generated by total-body irradiation of NOG-EXL mice (1.5 Gy), followed by intravenous transplantation of 2.5 × 10⁵ CD34⁺ cells resuspended in 0.2 mL RPMI-1640 medium. Reconstitution was validated at week 8 when ≥25% of peripheral blood leukocytes expressed human CD45.

A preservation medium consisting of 10% FBS and 1% penicillin-streptomycin was prepared for tumor handling. Tumors were trimmed of necrotic and non-malignant tissues under sterile conditions, minced into 1–2 mm³ fragments, dissociated into single-cell suspensions, and washed. Suspensions were injected subcutaneously into the dorsal flanks of 4-8-week-old athymic nude mice. Engraftment was deemed successful upon tumor volumes of 1–2 cm³ within 2–4 months; mice lacking tumors at six months were considered engraftment failures. Successfully engrafted tumors were serially passaged (F1-F3); F3-generation xenografts were employed for therapeutic studies. When tumors attained 1–2 cm³, they were excised and maintained in preservation medium. Under sodium pentobarbital anesthesia, tumor fragments were orthotopically implanted into the right hepatic lobe via a left-upper-quadrant transverse incision. The incision was closed in layers to complete the procedure.

Organoids culture

Organoid medium was prepared in Advanced DMEM/F12 supplemented with 1% (v/v) penicillin-streptomycin, 10 mM HEPES, 1x GlutaMAX, 1x B27, 1 μM N-acetylcysteine, 50 ng mL⁻¹ human recombinant EGF, 5 μM A83-01, 3 μM SB202190, 1 μM nicotinamide, 10 nM prostaglandin E2, and 10% (v/v) Noggin- and R-spondin1-conditioned media. Digested cells were resuspended in 4 mg mL⁻¹ Matrigel and printed onto Petri dish lids by acoustic droplet ejection; droplets contained 3-4 x 10⁴ cells each. Humidity was maintained by adding PBS to the dish before inversion of the lid. Plates were incubated at 4°C for 30 min to allow droplet settling, then transferred to 37°C, 5% CO₂. After 48 h, pre-formed organoids were transferred to 96-well plates and cultured for further observation.

Quantification and statistical analysis

Bioinformatics analyses were executed in R (v4.3.0); data are expressed as mean ± standard error of the mean. Statistical analyses and visualisations were performed with GraphPad Prism (v8.0). Confocal images were taken on a Zeiss LSM 880 Airyscan confocal microscope (Zeiss). Tumor regression was assessed using a ‘change to the baseline’ analysis. Between-group differences were evaluated by unpaired two-tailed t-tests; multiple comparisons were assessed by one-way analysis of variance with Tukey’s post-hoc test where appropriate. Significance levels were denoted as ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001. Results with p < 0.05 were considered statistically significant.

Published: January 20, 2026

Footnotes

Supplemental information can be found online at https://doi.org/10.1016/j.xcrm.2025.102524.

Contributor Information

Diwei Zheng, Email: dwzheng@ipe.ac.cn.

Chunhui Yuan, Email: chunhuii.yuen@whu.edu.cn.

Fubing Wang, Email: wfb20042002@sina.com.

Supplemental information

Document S1. Figures S1–S10

mmc1.pdf^{(4.9MB, pdf)}

Data S1. Clinical information of source patients for PDX and organoid model generation in this study, related to STAR Methods

mmc2.xlsx^{(10.6KB, xlsx)}

Document S2. Article plus supplemental information

mmc3.pdf^{(56.1MB, pdf)}

References

1.Llovet J.M., Kelley R.K., Villanueva A., Singal A.G., Pikarsky E., Roayaie S., Lencioni R., Koike K., Zucman-Rossi J., Finn R.S. Hepatocellular carcinoma. Nat. Rev. Dis. Primers. 2021;7:6. doi: 10.1038/s41572-020-00240-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Cheng A.-L., Qin S., Ikeda M., Galle P.R., Ducreux M., Kim T.Y., Lim H.Y., Kudo M., Breder V., Merle P., et al. Updated efficacy and safety data from IMbrave150: Atezolizumab plus bevacizumab vs. sorafenib for unresectable hepatocellular carcinoma. J. Hepatol. 2022;76:862–873. doi: 10.1016/j.jhep.2021.11.030. [DOI] [PubMed] [Google Scholar]
3.Sangro B., Galle P.R., Kelley R.K., Charoentum C., De Toni E.N., Ostapenko Y., Heo J., Cheng A.L., Wilson Woods A., Gupta C., et al. Patient-Reported Outcomes From the Phase III HIMALAYA Study of Tremelimumab Plus Durvalumab in Unresectable Hepatocellular Carcinoma. J. Clin. Oncol. 2024;42:2790–2799. doi: 10.1200/jco.23.01462. [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Sharma P., Goswami S., Raychaudhuri D., Siddiqui B.A., Singh P., Nagarajan A., Liu J., Subudhi S.K., Poon C., Gant K.L., et al. Immune checkpoint therapy-current perspectives and future directions. Cell. 2023;186:1652–1669. doi: 10.1016/j.cell.2023.03.006. [DOI] [PubMed] [Google Scholar]
5.Lau G., Abou-Alfa G.K., Cheng A.L., Sukeepaisarnjaroen W., Van Dao T., Kang Y.K., Thungappa S.C., Kudo M., Sangro B., Kelley R.K., et al. Outcomes in the Asian subgroup of the phase III randomised HIMALAYA study of tremelimumab plus durvalumab in unresectable hepatocellular carcinoma. J. Hepatol. 2025;82:258–267. doi: 10.1016/j.jhep.2024.07.017. [DOI] [PubMed] [Google Scholar]
6.Yarchoan M., Gane E.J., Marron T.U., Perales-Linares R., Yan J., Cooch N., Shu D.H., Fertig E.J., Kagohara L.T., Bartha G., et al. Personalized neoantigen vaccine and pembrolizumab in advanced hepatocellular carcinoma: a phase 1/2 trial. Nat. Med. 2024;30:1044–1053. doi: 10.1038/s41591-024-02894-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Shen K.Y., Zhu Y., Xie S.Z., Qin L.X. Immunosuppressive tumor microenvironment and immunotherapy of hepatocellular carcinoma: current status and prospectives. J. Hematol. Oncol. 2024;17:25. doi: 10.1186/s13045-024-01549-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Donne R., Lujambio A. The liver cancer immune microenvironment: Therapeutic implications for hepatocellular carcinoma. Hepatology. 2023;77:1773–1796. doi: 10.1002/hep.32740. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Hajishengallis G., Netea M.G., Chavakis T. Trained immunity in chronic inflammatory diseases and cancer. Nat. Rev. Immunol. 2025;25:497–514. doi: 10.1038/s41577-025-01132-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Liu G., Ma N., Cheng K., Feng Q., Ma X., Yue Y., Li Y., Zhang T., Gao X., Liang J., et al. Bacteria-derived nanovesicles enhance tumour vaccination by trained immunity. Nat. Nanotechnol. 2024;19:387–398. doi: 10.1038/s41565-023-01553-6. [DOI] [PubMed] [Google Scholar]
11.Ding C., Shrestha R., Zhu X., Geller A.E., Wu S., Woeste M.R., Li W., Wang H., Yuan F., Xu R., et al. Inducing trained immunity in pro-metastatic macrophages to control tumor metastasis. Nat. Immunol. 2023;24:239–254. doi: 10.1038/s41590-022-01388-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Chen Z., Yong T., Wei Z., Zhang X., Li X., Qin J., Li J., Hu J., Yang X., Gan L. Engineered Probiotic-Based Personalized Cancer Vaccine Potentiates Antitumor Immunity through Initiating Trained Immunity. Adv. Sci. 2024;11 doi: 10.1002/advs.202305081. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Singh A.K., Praharaj M., Lombardo K.A., Yoshida T., Matoso A., Baras A.S., Zhao L., Srikrishna G., Huang J., Prasad P., et al. Re-engineered BCG overexpressing cyclic di-AMP augments trained immunity and exhibits improved efficacy against bladder cancer. Nat. Commun. 2022;13:878. doi: 10.1038/s41467-022-28509-z. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
14.Derosa L., Routy B., Thomas A.M., Iebba V., Zalcman G., Friard S., Mazieres J., Audigier-Valette C., Moro-Sibilot D., Goldwasser F., et al. Intestinal Akkermansia muciniphila predicts clinical response to PD-1 blockade in patients with advanced non-small-cell lung cancer. Nat. Med. 2022;28:315–324. doi: 10.1038/s41591-021-01655-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Jiang Y., Xu Y., Zheng C., Ye L., Jiang P., Malik S., Xu G., Zhou Q., Zhang M. Acetyltransferase from Akkermansia muciniphila blunts colorectal tumourigenesis by reprogramming tumour microenvironment. Gut. 2023;72:1308–1318. doi: 10.1136/gutjnl-2022-327853. [DOI] [PubMed] [Google Scholar]
16.Bae M., Cassilly C.D., Liu X., Park S.M., Tusi B.K., Chen X., Kwon J., Filipčík P., Bolze A.S., Liu Z., et al. Akkermansia muciniphila phospholipid induces homeostatic immune responses. Nature. 2022;608:168–173. doi: 10.1038/s41586-022-04985-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Rao Y., Kuang Z., Li C., Guo S., Xu Y., Zhao D., Hu Y., Song B., Jiang Z., Ge Z., et al. Gut Akkermansia muciniphila ameliorates metabolic dysfunction-associated fatty liver disease by regulating the metabolism of L-aspartate via gut-liver axis. Gut Microbes. 2021;13:1–19. doi: 10.1080/19490976.2021.1927633. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Nie Q., Luo X., Wang K., Ding Y., Jia S., Zhao Q., Li M., Zhang J., Zhuo Y., Lin J., et al. Gut symbionts alleviate MASH through a secondary bile acid biosynthetic pathway. Cell. 2024;187:2717–2734.e33. doi: 10.1016/j.cell.2024.03.034. [DOI] [PubMed] [Google Scholar]
19.Ioannou A., Berkhout M.D., Geerlings S.Y., Belzer C. Akkermansia muciniphila: biology, microbial ecology, host interactions and therapeutic potential. Nat. Rev. Microbiol. 2025;23:162–177. doi: 10.1038/s41579-024-01106-1. [DOI] [PubMed] [Google Scholar]
20.Debacker A.J., Voutila J., Catley M., Blakey D., Habib N. Delivery of Oligonucleotides to the Liver with GalNAc: From Research to Registered Therapeutic Drug. Mol. Ther. 2020;28:1759–1771. doi: 10.1016/j.ymthe.2020.06.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Shi D., Shi Y., Kaseb A.O., Qi X., Zhang Y., Chi J., Lu Q., Gao H., Jiang H., Wang H., et al. Chimeric Antigen Receptor-Glypican-3 T-Cell Therapy for Advanced Hepatocellular Carcinoma: Results of Phase I Trials. Clin. Cancer Res. 2020;26:3979–3989. doi: 10.1158/1078-0432.Ccr-19-3259. [DOI] [PubMed] [Google Scholar]
22.Tang R., Xu J., Zhang B., Liu J., Liang C., Hua J., Meng Q., Yu X., Shi S. Ferroptosis, necroptosis, and pyroptosis in anticancer immunity. J. Hematol. Oncol. 2020;13:110. doi: 10.1186/s13045-020-00946-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Huang C., Li J., Wu R., Li Y., Zhang C. Targeting pyroptosis for cancer immunotherapy: mechanistic insights and clinical perspectives. Mol. Cancer. 2025;24:131. doi: 10.1186/s12943-025-02344-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Li J., Bhatt P.M., Li J., Eddaoudi M., Liu Y. Recent Progress on Microfine Design of Metal–Organic Frameworks: Structure Regulation and Gas Sorption and Separation. Adv. Mater. 2020;32 doi: 10.1002/adma.202002563. [DOI] [PubMed] [Google Scholar]
25.Zhang M., Shao W., Yang T., Liu H., Guo S., Zhao D., Weng Y., Liang X.J., Huang Y. Conscription of Immune Cells by Light-Activatable Silencing NK-Derived Exosome (LASNEO) for Synergetic Tumor Eradication. Adv. Sci. 2022;9 doi: 10.1002/advs.202201135. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Routy B., Le Chatelier E., Derosa L., Duong C.P.M., Alou M.T., Daillère R., Fluckiger A., Messaoudene M., Rauber C., Roberti M.P., et al. Gut microbiome influences efficacy of PD-1-based immunotherapy against epithelial tumors. Science (New York, N.Y.) 2018;359:91–97. doi: 10.1126/science.aan3706. [DOI] [PubMed] [Google Scholar]
27.Sivan A., Corrales L., Hubert N., Williams J.B., Aquino-Michaels K., Earley Z.M., Benyamin F.W., Lei Y.M., Jabri B., Alegre M.L., et al. Commensal Bifidobacterium promotes antitumor immunity and facilitates anti-PD-L1 efficacy. Science (New York, N.Y.) 2015;350:1084–1089. doi: 10.1126/science.aac4255. [DOI] [PMC free article] [PubMed] [Google Scholar]
28.Bender M.J., McPherson A.C., Phelps C.M., Pandey S.P., Laughlin C.R., Shapira J.H., Medina Sanchez L., Rana M., Richie T.G., Mims T.S., et al. Dietary tryptophan metabolite released by intratumoral Lactobacillus reuteri facilitates immune checkpoint inhibitor treatment. Cell. 2023;186:1846–1862.e26. doi: 10.1016/j.cell.2023.03.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Gao G., Shen S., Zhang T., Zhang J., Huang S., Sun Z., Zhang H. Lacticaseibacillus rhamnosus Probio-M9 enhanced the antitumor response to anti-PD-1 therapy by modulating intestinal metabolites. EBioMedicine. 2023;91 doi: 10.1016/j.ebiom.2023.104533. [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Brown C.R., Gupta S., Qin J., Racie T., He G., Lentini S., Malone R., Yu M., Matsuda S., Shulga-Morskaya S., et al. Investigating the pharmacodynamic durability of GalNAc-siRNA conjugates. Nucleic Acids Res. 2020;48:11827–11844. doi: 10.1093/nar/gkaa670. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Qu H., Hang L., Diao Y., Wang H., Fang L., Liu W., Liu J., Sun H., Wang J., Meng X., et al. Mn-doped MOF nanoparticles mitigating hypoxia via in-situ substitution strategy for dual-imaging guided combination treatment of microwave dynamic therapy and chemotherapy. J. Colloid Interface Sci. 2025;685:912–926. doi: 10.1016/j.jcis.2025.01.202. [DOI] [PubMed] [Google Scholar]
32.Yang Y., Fan H., Xu X., Yao S., Yu W., Guo Z. Zinc Ion-Induced Immune Responses in Antitumor Immunotherapy. CCS Chem. 2024;6:2210–2229. doi: 10.31635/ccschem.023.202303514. [DOI] [Google Scholar]
33.Lawlor K.E., Murphy J.M., Vince J.E. Gasdermin and MLKL necrotic cell death effectors: Signaling and diseases. Immunity. 2024;57:429–445. doi: 10.1016/j.immuni.2024.02.011. [DOI] [PubMed] [Google Scholar]
34.Tong X., Tang R., Xiao M., Xu J., Wang W., Zhang B., Liu J., Yu X., Shi S. Targeting cell death pathways for cancer therapy: recent developments in necroptosis, pyroptosis, ferroptosis, and cuproptosis research. J. Hematol. Oncol. 2022;15:174. doi: 10.1186/s13045-022-01392-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Sanders M.G., Parsons M.J., Howard A.G.A., Liu J., Fassio S.R., Martinez J.A., Bouchier-Hayes L. Single-cell imaging of inflammatory caspase dimerization reveals differential recruitment to inflammasomes. Cell Death Dis. 2015;6:e1813. doi: 10.1038/cddis.2015.186. [DOI] [PMC free article] [PubMed] [Google Scholar]
36.Akbal A., Dernst A., Lovotti M., Mangan M.S.J., McManus R.M., Latz E. How location and cellular signaling combine to activate the NLRP3 inflammasome. Cell. Mol. Immunol. 2022;19:1201–1214. doi: 10.1038/s41423-022-00922-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Miao R., Jiang C., Chang W.Y., Zhang H., An J., Ho F., Chen P., Zhang H., Junqueira C., Amgalan D., et al. Gasdermin D permeabilization of mitochondrial inner and outer membranes accelerates and enhances pyroptosis. Immunity. 2023;56:2523–2541.e8. doi: 10.1016/j.immuni.2023.10.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Liu J., Kang R., Tang D. Lipopolysaccharide delivery systems in innate immunity. Trends Immunol. 2024;45:274–287. doi: 10.1016/j.it.2024.02.003. [DOI] [PubMed] [Google Scholar]
39.Wang J.B., Gao Y.X., Ye Y.H., Zheng Q.L., Luo H.Y., Wang S.H., Zhang T., Jin Q.W., Zheng C.H., Li P., et al. Comprehensive multi-omics analysis of pyroptosis for optimizing neoadjuvant immunotherapy in patients with gastric cancer. Theranostics. 2024;14:2915–2933. doi: 10.7150/thno.93124. [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Khan N., Downey J., Sanz J., Kaufmann E., Blankenhaus B., Pacis A., Pernet E., Ahmed E., Cardoso S., Nijnik A., Mazer B. M. tuberculosis Reprograms Hematopoietic Stem Cells to Limit Myelopoiesis and Impair Trained Immunity. Cell. 2020;183:752–770. doi: 10.1016/j.cell.2020.09.062. [DOI] [PMC free article] [PubMed] [Google Scholar]
41.Chin E.N., Sulpizio A., Lairson L.L. Targeting STING to promote antitumor immunity. Trends Cell Biol. 2023;33:189–203. doi: 10.1016/j.tcb.2022.06.010. [DOI] [PubMed] [Google Scholar]
42.Xiang X., Wang K., Zhang H., Mou H., Shi Z., Tao Y., Song H., Lian Z., Wang S., Lu D., et al. Blocking CX3CR1+ Tumor-Associated Macrophages Enhances the Efficacy of Anti-PD1 Therapy in Hepatocellular Carcinoma. Cancer Immunol. Res. 2024;12:1603–1620. doi: 10.1158/2326-6066.Cir-23-0627. [DOI] [PubMed] [Google Scholar]
43.Xue R., Zhang Q., Cao Q., Kong R., Xiang X., Liu H., Feng M., Wang F., Cheng J., Li Z., et al. Liver tumour immune microenvironment subtypes and neutrophil heterogeneity. Nature. 2022;612:141–147. doi: 10.1038/s41586-022-05400-x. [DOI] [PubMed] [Google Scholar]
44.Gao W., Wang X., Zhou Y., Wang X., Yu Y. Autophagy, ferroptosis, pyroptosis, and necroptosis in tumor immunotherapy. Signal Transduct. Target. Ther. 2022;7:196. doi: 10.1038/s41392-022-01046-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Sun Y., Wu P., Zhang Z., Wang Z., Zhou K., Song M., Ji Y., Zang F., Lou L., Rao K., et al. Integrated multi-omics profiling to dissect the spatiotemporal evolution of metastatic hepatocellular carcinoma. Cancer Cell. 2024;42:135–156.e17. doi: 10.1016/j.ccell.2023.11.010. [DOI] [PubMed] [Google Scholar]
46.Dong L.Q., Peng L.H., Ma L.J., Liu D.B., Zhang S., Luo S.Z., Rao J.H., Zhu H.W., Yang S.X., Xi S.J., et al. Heterogeneous immunogenomic features and distinct escape mechanisms in multifocal hepatocellular carcinoma. J. Hepatol. 2020;72:896–908. doi: 10.1016/j.jhep.2019.12.014. [DOI] [PubMed] [Google Scholar]
47.Zeng S., Wang J., Kang H., Li H., Peng X., Yoon J. Photon-Driven Dye Induction Pyroptosis: An Emerging Anti-Tumor Immunotherapy Paradigm. Angew. Chem. Int. Ed. Engl. 2025;64 doi: 10.1002/anie.202417899. [DOI] [PubMed] [Google Scholar]
48.Zhang Y., Zhang X., Yang H., Yu L., Xu Y., Sharma A., Yin P., Li X., Kim J.S., Sun Y. Advanced biotechnology-assisted precise sonodynamic therapy. Chem. Soc. Rev. 2021;50:11227–11248. doi: 10.1039/d1cs00403d. [DOI] [PubMed] [Google Scholar]
49.Yang F., Dong J., Li Z., Wang Z. Metal–Organic Frameworks (MOF)-Assisted Sonodynamic Therapy in Anticancer Applications. ACS Nano. 2023;17:4102–4133. doi: 10.1021/acsnano.2c10251. [DOI] [PubMed] [Google Scholar]
50.Liu D., Wan J., Pang G., Tang Z. Hollow Metal–Organic-Framework Micro/Nanostructures and their Derivatives: Emerging Multifunctional Materials. Adv. Mater. 2019;31 doi: 10.1002/adma.201803291. [DOI] [PubMed] [Google Scholar]
51.Bill R., Wirapati P., Messemaker M., Roh W., Zitti B., Duval F., Kiss M., Park J.C., Saal T.M., Hoelzl J., et al. CXCL9:SPP1 macrophage polarity identifies a network of cellular programs that control human cancers. Science. 2023;381:515–524. doi: 10.1126/science.ade2292. [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Luan X., Lei T., Fang J., Liu X., Fu H., Li Y., Chu W., Jiang P., Tong C., Qi H., Fu Y. Blockade of C5a receptor unleashes tumor-associated macrophage antitumor response and enhances CXCL9-dependent CD8(+) T cell activity. Mol. Ther. 2024;32:469–489. doi: 10.1016/j.ymthe.2023.12.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
53.de Masson A., Darbord D., Dobos G., Boisson M., Roelens M., Ram-Wolff C., Cassius C., Le Buanec H., de la Grange P., Jouenne F., et al. Macrophage-derived CXCL9 and CXCL11, T-cell skin homing, and disease control in mogamulizumab-treated CTCL patients. Blood. 2022;139:1820–1832. doi: 10.1182/blood.2021013341. [DOI] [PubMed] [Google Scholar]
54.Wu J., Liu W., Qiu X., Li J., Song K., Shen S., Huo L., Chen L., Xu M., Wang H., et al. A Noninvasive Approach to Evaluate Tumor Immune Microenvironment and Predict Outcomes in Hepatocellular Carcinoma. Phenomics. 2023;3:549–564. doi: 10.1007/s43657-023-00136-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Mishra B., Ivashkiv L.B. Interferons and epigenetic mechanisms in training, priming and tolerance of monocytes and hematopoietic progenitors. Immunol. Rev. 2024;323:257–275. doi: 10.1111/imr.13330. [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Yao Y., Jeyanathan M., Haddadi S., Barra N.G., Vaseghi-Shanjani M., Damjanovic D., Lai R., Afkhami S., Chen Y., Dvorkin-Gheva A., et al. Induction of Autonomous Memory Alveolar Macrophages Requires T Cell Help and Is Critical to Trained Immunity. Cell. 2018;175:1634–1650.e17. doi: 10.1016/j.cell.2018.09.042. [DOI] [PubMed] [Google Scholar]
57.Dangaj D., Bruand M., Grimm A.J., Ronet C., Barras D., Duttagupta P.A., Lanitis E., Duraiswamy J., Tanyi J.L., Benencia F., et al. Cooperation between Constitutive and Inducible Chemokines Enables T Cell Engraftment and Immune Attack in Solid Tumors. Cancer Cell. 2019;35:885–900.e10. doi: 10.1016/j.ccell.2019.05.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Zhao C., Zheng T., Wang R., Lin X., Hu Z., Zhao Z., Dai Z., Sun D. Synergistically Augmenting Cancer Immunotherapy by Physical Manipulation of Pyroptosis Induction. Phenomics. 2024;4:298–312. doi: 10.1007/s43657-023-00140-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Wang J., Lu Y., Zhang R., Cai Z., Fan Z., Xu Y., Liu Z., Zhang Z. Modulating and Imaging Macrophage Reprogramming for Cancer Immunotherapy. Phenomics. 2024;4:401–414. doi: 10.1007/s43657-023-00154-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
60.Cheng S.C., Quintin J., Cramer R.A., Shepardson K.M., Saeed S., Kumar V., Giamarellos-Bourboulis E.J., Martens J.H.A., Rao N.A., Aghajanirefah A., et al. mTOR- and HIF-1α-mediated aerobic glycolysis as metabolic basis for trained immunity. Science. 2014;345 doi: 10.1126/science.1250684. [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Wang S., Liu R., Yu Q., Dong L., Bi Y., Liu G. Metabolic reprogramming of macrophages during infections and cancer. Cancer Lett. 2019;452:14–22. doi: 10.1016/j.canlet.2019.03.015. [DOI] [PubMed] [Google Scholar]
62.House I.G., Savas P., Lai J., Chen A.X.Y., Oliver A.J., Teo Z.L., Todd K.L., Henderson M.A., Giuffrida L., Petley E.V., et al. Macrophage-Derived CXCL9 and CXCL10 Are Required for Antitumor Immune Responses Following Immune Checkpoint Blockade. Clin. Cancer Res. 2020;26:487–504. doi: 10.1158/1078-0432.Ccr-19-1868. [DOI] [PubMed] [Google Scholar]
63.Zhang B., Moorlag S.J., Dominguez-Andres J., Bulut Ö., Kilic G., Liu Z., van Crevel R., Xu C.J., Joosten L.A., Netea M.G., Li Y. Single-cell RNA sequencing reveals induction of distinct trained-immunity programs in human monocytes. J. Clin. Investig. 2022;132 doi: 10.1172/jci147719. [DOI] [PMC free article] [PubMed] [Google Scholar]
64.Dao V., Yuki K., Lo Y.H., Nakano M., Kuo C.J. Immune organoids: from tumor modeling to precision oncology. Trends Cancer. 2022;8:870–880. doi: 10.1016/j.trecan.2022.06.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
65.Shultz L.D., Brehm M.A., Garcia-Martinez J.V., Greiner D.L. Humanized mice for immune system investigation: progress, promise and challenges. Nat. Rev. Immunol. 2012;12:786–798. doi: 10.1038/nri3311. [DOI] [PMC free article] [PubMed] [Google Scholar]
66.Llovet J.M., Kelley R.K., Villanueva A., Singal A.G., Pikarsky E., Roayaie S., Lencioni R., Koike K., Zucman-Rossi J., Finn R.S. Hepatocellular carcinoma. Nat. Rev. Dis. Primers. 2021;7:6. doi: 10.1038/s41572-021-00245-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
67.Tuveson D., Clevers H. Cancer modeling meets human organoid technology. Science. 2019;364:952–955. doi: 10.1126/science.aaw6985. [DOI] [PubMed] [Google Scholar]
68.Gong Z., Huang L., Tang X., Chen K., Wu Z., Zhang L., Sun Y., Xia Y., Chen H., Wei Y., et al. Acoustic Droplet Printing Tumor Organoids for Modeling Bladder Tumor Immune Microenvironment within a Week. Adv. Healthc. Mater. 2021;10 doi: 10.1002/adhm.202101312. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Figures S1–S10

mmc1.pdf^{(4.9MB, pdf)}

Data S1. Clinical information of source patients for PDX and organoid model generation in this study, related to STAR Methods

mmc2.xlsx^{(10.6KB, xlsx)}

Document S2. Article plus supplemental information

mmc3.pdf^{(56.1MB, pdf)}

Data Availability Statement

•
RNA-seq data have been deposited at Genome Sequence Archive in National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences, and publicly available as of the date of publication (no. PRJCA041597; https://ngdc.cncb.ac.cn/bioproject/browse/PRJCA041597).
•
This paper does not report original code.
•
Any additional information required to reanalyze the data reported in this work paper is available from the lead contact upon request.

[bib1] 1.Llovet J.M., Kelley R.K., Villanueva A., Singal A.G., Pikarsky E., Roayaie S., Lencioni R., Koike K., Zucman-Rossi J., Finn R.S. Hepatocellular carcinoma. Nat. Rev. Dis. Primers. 2021;7:6. doi: 10.1038/s41572-020-00240-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib2] 2.Cheng A.-L., Qin S., Ikeda M., Galle P.R., Ducreux M., Kim T.Y., Lim H.Y., Kudo M., Breder V., Merle P., et al. Updated efficacy and safety data from IMbrave150: Atezolizumab plus bevacizumab vs. sorafenib for unresectable hepatocellular carcinoma. J. Hepatol. 2022;76:862–873. doi: 10.1016/j.jhep.2021.11.030. [DOI] [PubMed] [Google Scholar]

[bib3] 3.Sangro B., Galle P.R., Kelley R.K., Charoentum C., De Toni E.N., Ostapenko Y., Heo J., Cheng A.L., Wilson Woods A., Gupta C., et al. Patient-Reported Outcomes From the Phase III HIMALAYA Study of Tremelimumab Plus Durvalumab in Unresectable Hepatocellular Carcinoma. J. Clin. Oncol. 2024;42:2790–2799. doi: 10.1200/jco.23.01462. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib4] 4.Sharma P., Goswami S., Raychaudhuri D., Siddiqui B.A., Singh P., Nagarajan A., Liu J., Subudhi S.K., Poon C., Gant K.L., et al. Immune checkpoint therapy-current perspectives and future directions. Cell. 2023;186:1652–1669. doi: 10.1016/j.cell.2023.03.006. [DOI] [PubMed] [Google Scholar]

[bib5] 5.Lau G., Abou-Alfa G.K., Cheng A.L., Sukeepaisarnjaroen W., Van Dao T., Kang Y.K., Thungappa S.C., Kudo M., Sangro B., Kelley R.K., et al. Outcomes in the Asian subgroup of the phase III randomised HIMALAYA study of tremelimumab plus durvalumab in unresectable hepatocellular carcinoma. J. Hepatol. 2025;82:258–267. doi: 10.1016/j.jhep.2024.07.017. [DOI] [PubMed] [Google Scholar]

[bib6] 6.Yarchoan M., Gane E.J., Marron T.U., Perales-Linares R., Yan J., Cooch N., Shu D.H., Fertig E.J., Kagohara L.T., Bartha G., et al. Personalized neoantigen vaccine and pembrolizumab in advanced hepatocellular carcinoma: a phase 1/2 trial. Nat. Med. 2024;30:1044–1053. doi: 10.1038/s41591-024-02894-y. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib7] 7.Shen K.Y., Zhu Y., Xie S.Z., Qin L.X. Immunosuppressive tumor microenvironment and immunotherapy of hepatocellular carcinoma: current status and prospectives. J. Hematol. Oncol. 2024;17:25. doi: 10.1186/s13045-024-01549-2. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib8] 8.Donne R., Lujambio A. The liver cancer immune microenvironment: Therapeutic implications for hepatocellular carcinoma. Hepatology. 2023;77:1773–1796. doi: 10.1002/hep.32740. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib9] 9.Hajishengallis G., Netea M.G., Chavakis T. Trained immunity in chronic inflammatory diseases and cancer. Nat. Rev. Immunol. 2025;25:497–514. doi: 10.1038/s41577-025-01132-x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib10] 10.Liu G., Ma N., Cheng K., Feng Q., Ma X., Yue Y., Li Y., Zhang T., Gao X., Liang J., et al. Bacteria-derived nanovesicles enhance tumour vaccination by trained immunity. Nat. Nanotechnol. 2024;19:387–398. doi: 10.1038/s41565-023-01553-6. [DOI] [PubMed] [Google Scholar]

[bib11] 11.Ding C., Shrestha R., Zhu X., Geller A.E., Wu S., Woeste M.R., Li W., Wang H., Yuan F., Xu R., et al. Inducing trained immunity in pro-metastatic macrophages to control tumor metastasis. Nat. Immunol. 2023;24:239–254. doi: 10.1038/s41590-022-01388-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib12] 12.Chen Z., Yong T., Wei Z., Zhang X., Li X., Qin J., Li J., Hu J., Yang X., Gan L. Engineered Probiotic-Based Personalized Cancer Vaccine Potentiates Antitumor Immunity through Initiating Trained Immunity. Adv. Sci. 2024;11 doi: 10.1002/advs.202305081. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib13] 13.Singh A.K., Praharaj M., Lombardo K.A., Yoshida T., Matoso A., Baras A.S., Zhao L., Srikrishna G., Huang J., Prasad P., et al. Re-engineered BCG overexpressing cyclic di-AMP augments trained immunity and exhibits improved efficacy against bladder cancer. Nat. Commun. 2022;13:878. doi: 10.1038/s41467-022-28509-z. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]

[bib14] 14.Derosa L., Routy B., Thomas A.M., Iebba V., Zalcman G., Friard S., Mazieres J., Audigier-Valette C., Moro-Sibilot D., Goldwasser F., et al. Intestinal Akkermansia muciniphila predicts clinical response to PD-1 blockade in patients with advanced non-small-cell lung cancer. Nat. Med. 2022;28:315–324. doi: 10.1038/s41591-021-01655-5. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib15] 15.Jiang Y., Xu Y., Zheng C., Ye L., Jiang P., Malik S., Xu G., Zhou Q., Zhang M. Acetyltransferase from Akkermansia muciniphila blunts colorectal tumourigenesis by reprogramming tumour microenvironment. Gut. 2023;72:1308–1318. doi: 10.1136/gutjnl-2022-327853. [DOI] [PubMed] [Google Scholar]

[bib16] 16.Bae M., Cassilly C.D., Liu X., Park S.M., Tusi B.K., Chen X., Kwon J., Filipčík P., Bolze A.S., Liu Z., et al. Akkermansia muciniphila phospholipid induces homeostatic immune responses. Nature. 2022;608:168–173. doi: 10.1038/s41586-022-04985-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib17] 17.Rao Y., Kuang Z., Li C., Guo S., Xu Y., Zhao D., Hu Y., Song B., Jiang Z., Ge Z., et al. Gut Akkermansia muciniphila ameliorates metabolic dysfunction-associated fatty liver disease by regulating the metabolism of L-aspartate via gut-liver axis. Gut Microbes. 2021;13:1–19. doi: 10.1080/19490976.2021.1927633. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib18] 18.Nie Q., Luo X., Wang K., Ding Y., Jia S., Zhao Q., Li M., Zhang J., Zhuo Y., Lin J., et al. Gut symbionts alleviate MASH through a secondary bile acid biosynthetic pathway. Cell. 2024;187:2717–2734.e33. doi: 10.1016/j.cell.2024.03.034. [DOI] [PubMed] [Google Scholar]

[bib19] 19.Ioannou A., Berkhout M.D., Geerlings S.Y., Belzer C. Akkermansia muciniphila: biology, microbial ecology, host interactions and therapeutic potential. Nat. Rev. Microbiol. 2025;23:162–177. doi: 10.1038/s41579-024-01106-1. [DOI] [PubMed] [Google Scholar]

[bib20] 20.Debacker A.J., Voutila J., Catley M., Blakey D., Habib N. Delivery of Oligonucleotides to the Liver with GalNAc: From Research to Registered Therapeutic Drug. Mol. Ther. 2020;28:1759–1771. doi: 10.1016/j.ymthe.2020.06.015. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib21] 21.Shi D., Shi Y., Kaseb A.O., Qi X., Zhang Y., Chi J., Lu Q., Gao H., Jiang H., Wang H., et al. Chimeric Antigen Receptor-Glypican-3 T-Cell Therapy for Advanced Hepatocellular Carcinoma: Results of Phase I Trials. Clin. Cancer Res. 2020;26:3979–3989. doi: 10.1158/1078-0432.Ccr-19-3259. [DOI] [PubMed] [Google Scholar]

[bib22] 22.Tang R., Xu J., Zhang B., Liu J., Liang C., Hua J., Meng Q., Yu X., Shi S. Ferroptosis, necroptosis, and pyroptosis in anticancer immunity. J. Hematol. Oncol. 2020;13:110. doi: 10.1186/s13045-020-00946-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib23] 23.Huang C., Li J., Wu R., Li Y., Zhang C. Targeting pyroptosis for cancer immunotherapy: mechanistic insights and clinical perspectives. Mol. Cancer. 2025;24:131. doi: 10.1186/s12943-025-02344-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib24] 24.Li J., Bhatt P.M., Li J., Eddaoudi M., Liu Y. Recent Progress on Microfine Design of Metal–Organic Frameworks: Structure Regulation and Gas Sorption and Separation. Adv. Mater. 2020;32 doi: 10.1002/adma.202002563. [DOI] [PubMed] [Google Scholar]

[bib25] 25.Zhang M., Shao W., Yang T., Liu H., Guo S., Zhao D., Weng Y., Liang X.J., Huang Y. Conscription of Immune Cells by Light-Activatable Silencing NK-Derived Exosome (LASNEO) for Synergetic Tumor Eradication. Adv. Sci. 2022;9 doi: 10.1002/advs.202201135. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib26] 26.Routy B., Le Chatelier E., Derosa L., Duong C.P.M., Alou M.T., Daillère R., Fluckiger A., Messaoudene M., Rauber C., Roberti M.P., et al. Gut microbiome influences efficacy of PD-1-based immunotherapy against epithelial tumors. Science (New York, N.Y.) 2018;359:91–97. doi: 10.1126/science.aan3706. [DOI] [PubMed] [Google Scholar]

[bib27] 27.Sivan A., Corrales L., Hubert N., Williams J.B., Aquino-Michaels K., Earley Z.M., Benyamin F.W., Lei Y.M., Jabri B., Alegre M.L., et al. Commensal Bifidobacterium promotes antitumor immunity and facilitates anti-PD-L1 efficacy. Science (New York, N.Y.) 2015;350:1084–1089. doi: 10.1126/science.aac4255. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib28] 28.Bender M.J., McPherson A.C., Phelps C.M., Pandey S.P., Laughlin C.R., Shapira J.H., Medina Sanchez L., Rana M., Richie T.G., Mims T.S., et al. Dietary tryptophan metabolite released by intratumoral Lactobacillus reuteri facilitates immune checkpoint inhibitor treatment. Cell. 2023;186:1846–1862.e26. doi: 10.1016/j.cell.2023.03.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib29] 29.Gao G., Shen S., Zhang T., Zhang J., Huang S., Sun Z., Zhang H. Lacticaseibacillus rhamnosus Probio-M9 enhanced the antitumor response to anti-PD-1 therapy by modulating intestinal metabolites. EBioMedicine. 2023;91 doi: 10.1016/j.ebiom.2023.104533. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib30] 30.Brown C.R., Gupta S., Qin J., Racie T., He G., Lentini S., Malone R., Yu M., Matsuda S., Shulga-Morskaya S., et al. Investigating the pharmacodynamic durability of GalNAc-siRNA conjugates. Nucleic Acids Res. 2020;48:11827–11844. doi: 10.1093/nar/gkaa670. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib31] 31.Qu H., Hang L., Diao Y., Wang H., Fang L., Liu W., Liu J., Sun H., Wang J., Meng X., et al. Mn-doped MOF nanoparticles mitigating hypoxia via in-situ substitution strategy for dual-imaging guided combination treatment of microwave dynamic therapy and chemotherapy. J. Colloid Interface Sci. 2025;685:912–926. doi: 10.1016/j.jcis.2025.01.202. [DOI] [PubMed] [Google Scholar]

[bib32] 32.Yang Y., Fan H., Xu X., Yao S., Yu W., Guo Z. Zinc Ion-Induced Immune Responses in Antitumor Immunotherapy. CCS Chem. 2024;6:2210–2229. doi: 10.31635/ccschem.023.202303514. [DOI] [Google Scholar]

[bib33] 33.Lawlor K.E., Murphy J.M., Vince J.E. Gasdermin and MLKL necrotic cell death effectors: Signaling and diseases. Immunity. 2024;57:429–445. doi: 10.1016/j.immuni.2024.02.011. [DOI] [PubMed] [Google Scholar]

[bib34] 34.Tong X., Tang R., Xiao M., Xu J., Wang W., Zhang B., Liu J., Yu X., Shi S. Targeting cell death pathways for cancer therapy: recent developments in necroptosis, pyroptosis, ferroptosis, and cuproptosis research. J. Hematol. Oncol. 2022;15:174. doi: 10.1186/s13045-022-01392-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib35] 35.Sanders M.G., Parsons M.J., Howard A.G.A., Liu J., Fassio S.R., Martinez J.A., Bouchier-Hayes L. Single-cell imaging of inflammatory caspase dimerization reveals differential recruitment to inflammasomes. Cell Death Dis. 2015;6:e1813. doi: 10.1038/cddis.2015.186. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib36] 36.Akbal A., Dernst A., Lovotti M., Mangan M.S.J., McManus R.M., Latz E. How location and cellular signaling combine to activate the NLRP3 inflammasome. Cell. Mol. Immunol. 2022;19:1201–1214. doi: 10.1038/s41423-022-00922-w. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib37] 37.Miao R., Jiang C., Chang W.Y., Zhang H., An J., Ho F., Chen P., Zhang H., Junqueira C., Amgalan D., et al. Gasdermin D permeabilization of mitochondrial inner and outer membranes accelerates and enhances pyroptosis. Immunity. 2023;56:2523–2541.e8. doi: 10.1016/j.immuni.2023.10.004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib38] 38.Liu J., Kang R., Tang D. Lipopolysaccharide delivery systems in innate immunity. Trends Immunol. 2024;45:274–287. doi: 10.1016/j.it.2024.02.003. [DOI] [PubMed] [Google Scholar]

[bib39] 39.Wang J.B., Gao Y.X., Ye Y.H., Zheng Q.L., Luo H.Y., Wang S.H., Zhang T., Jin Q.W., Zheng C.H., Li P., et al. Comprehensive multi-omics analysis of pyroptosis for optimizing neoadjuvant immunotherapy in patients with gastric cancer. Theranostics. 2024;14:2915–2933. doi: 10.7150/thno.93124. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib40] 40.Khan N., Downey J., Sanz J., Kaufmann E., Blankenhaus B., Pacis A., Pernet E., Ahmed E., Cardoso S., Nijnik A., Mazer B. M. tuberculosis Reprograms Hematopoietic Stem Cells to Limit Myelopoiesis and Impair Trained Immunity. Cell. 2020;183:752–770. doi: 10.1016/j.cell.2020.09.062. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib41] 41.Chin E.N., Sulpizio A., Lairson L.L. Targeting STING to promote antitumor immunity. Trends Cell Biol. 2023;33:189–203. doi: 10.1016/j.tcb.2022.06.010. [DOI] [PubMed] [Google Scholar]

[bib42] 42.Xiang X., Wang K., Zhang H., Mou H., Shi Z., Tao Y., Song H., Lian Z., Wang S., Lu D., et al. Blocking CX3CR1+ Tumor-Associated Macrophages Enhances the Efficacy of Anti-PD1 Therapy in Hepatocellular Carcinoma. Cancer Immunol. Res. 2024;12:1603–1620. doi: 10.1158/2326-6066.Cir-23-0627. [DOI] [PubMed] [Google Scholar]

[bib43] 43.Xue R., Zhang Q., Cao Q., Kong R., Xiang X., Liu H., Feng M., Wang F., Cheng J., Li Z., et al. Liver tumour immune microenvironment subtypes and neutrophil heterogeneity. Nature. 2022;612:141–147. doi: 10.1038/s41586-022-05400-x. [DOI] [PubMed] [Google Scholar]

[bib44] 44.Gao W., Wang X., Zhou Y., Wang X., Yu Y. Autophagy, ferroptosis, pyroptosis, and necroptosis in tumor immunotherapy. Signal Transduct. Target. Ther. 2022;7:196. doi: 10.1038/s41392-022-01046-3. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib45] 45.Sun Y., Wu P., Zhang Z., Wang Z., Zhou K., Song M., Ji Y., Zang F., Lou L., Rao K., et al. Integrated multi-omics profiling to dissect the spatiotemporal evolution of metastatic hepatocellular carcinoma. Cancer Cell. 2024;42:135–156.e17. doi: 10.1016/j.ccell.2023.11.010. [DOI] [PubMed] [Google Scholar]

[bib46] 46.Dong L.Q., Peng L.H., Ma L.J., Liu D.B., Zhang S., Luo S.Z., Rao J.H., Zhu H.W., Yang S.X., Xi S.J., et al. Heterogeneous immunogenomic features and distinct escape mechanisms in multifocal hepatocellular carcinoma. J. Hepatol. 2020;72:896–908. doi: 10.1016/j.jhep.2019.12.014. [DOI] [PubMed] [Google Scholar]

[bib47] 47.Zeng S., Wang J., Kang H., Li H., Peng X., Yoon J. Photon-Driven Dye Induction Pyroptosis: An Emerging Anti-Tumor Immunotherapy Paradigm. Angew. Chem. Int. Ed. Engl. 2025;64 doi: 10.1002/anie.202417899. [DOI] [PubMed] [Google Scholar]

[bib48] 48.Zhang Y., Zhang X., Yang H., Yu L., Xu Y., Sharma A., Yin P., Li X., Kim J.S., Sun Y. Advanced biotechnology-assisted precise sonodynamic therapy. Chem. Soc. Rev. 2021;50:11227–11248. doi: 10.1039/d1cs00403d. [DOI] [PubMed] [Google Scholar]

[bib49] 49.Yang F., Dong J., Li Z., Wang Z. Metal–Organic Frameworks (MOF)-Assisted Sonodynamic Therapy in Anticancer Applications. ACS Nano. 2023;17:4102–4133. doi: 10.1021/acsnano.2c10251. [DOI] [PubMed] [Google Scholar]

[bib50] 50.Liu D., Wan J., Pang G., Tang Z. Hollow Metal–Organic-Framework Micro/Nanostructures and their Derivatives: Emerging Multifunctional Materials. Adv. Mater. 2019;31 doi: 10.1002/adma.201803291. [DOI] [PubMed] [Google Scholar]

[bib51] 51.Bill R., Wirapati P., Messemaker M., Roh W., Zitti B., Duval F., Kiss M., Park J.C., Saal T.M., Hoelzl J., et al. CXCL9:SPP1 macrophage polarity identifies a network of cellular programs that control human cancers. Science. 2023;381:515–524. doi: 10.1126/science.ade2292. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib52] 52.Luan X., Lei T., Fang J., Liu X., Fu H., Li Y., Chu W., Jiang P., Tong C., Qi H., Fu Y. Blockade of C5a receptor unleashes tumor-associated macrophage antitumor response and enhances CXCL9-dependent CD8(+) T cell activity. Mol. Ther. 2024;32:469–489. doi: 10.1016/j.ymthe.2023.12.010. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib53] 53.de Masson A., Darbord D., Dobos G., Boisson M., Roelens M., Ram-Wolff C., Cassius C., Le Buanec H., de la Grange P., Jouenne F., et al. Macrophage-derived CXCL9 and CXCL11, T-cell skin homing, and disease control in mogamulizumab-treated CTCL patients. Blood. 2022;139:1820–1832. doi: 10.1182/blood.2021013341. [DOI] [PubMed] [Google Scholar]

[bib54] 54.Wu J., Liu W., Qiu X., Li J., Song K., Shen S., Huo L., Chen L., Xu M., Wang H., et al. A Noninvasive Approach to Evaluate Tumor Immune Microenvironment and Predict Outcomes in Hepatocellular Carcinoma. Phenomics. 2023;3:549–564. doi: 10.1007/s43657-023-00136-8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib55] 55.Mishra B., Ivashkiv L.B. Interferons and epigenetic mechanisms in training, priming and tolerance of monocytes and hematopoietic progenitors. Immunol. Rev. 2024;323:257–275. doi: 10.1111/imr.13330. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib56] 56.Yao Y., Jeyanathan M., Haddadi S., Barra N.G., Vaseghi-Shanjani M., Damjanovic D., Lai R., Afkhami S., Chen Y., Dvorkin-Gheva A., et al. Induction of Autonomous Memory Alveolar Macrophages Requires T Cell Help and Is Critical to Trained Immunity. Cell. 2018;175:1634–1650.e17. doi: 10.1016/j.cell.2018.09.042. [DOI] [PubMed] [Google Scholar]

[bib57] 57.Dangaj D., Bruand M., Grimm A.J., Ronet C., Barras D., Duttagupta P.A., Lanitis E., Duraiswamy J., Tanyi J.L., Benencia F., et al. Cooperation between Constitutive and Inducible Chemokines Enables T Cell Engraftment and Immune Attack in Solid Tumors. Cancer Cell. 2019;35:885–900.e10. doi: 10.1016/j.ccell.2019.05.004. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib58] 58.Zhao C., Zheng T., Wang R., Lin X., Hu Z., Zhao Z., Dai Z., Sun D. Synergistically Augmenting Cancer Immunotherapy by Physical Manipulation of Pyroptosis Induction. Phenomics. 2024;4:298–312. doi: 10.1007/s43657-023-00140-y. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib59] 59.Wang J., Lu Y., Zhang R., Cai Z., Fan Z., Xu Y., Liu Z., Zhang Z. Modulating and Imaging Macrophage Reprogramming for Cancer Immunotherapy. Phenomics. 2024;4:401–414. doi: 10.1007/s43657-023-00154-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib60] 60.Cheng S.C., Quintin J., Cramer R.A., Shepardson K.M., Saeed S., Kumar V., Giamarellos-Bourboulis E.J., Martens J.H.A., Rao N.A., Aghajanirefah A., et al. mTOR- and HIF-1α-mediated aerobic glycolysis as metabolic basis for trained immunity. Science. 2014;345 doi: 10.1126/science.1250684. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib61] 61.Wang S., Liu R., Yu Q., Dong L., Bi Y., Liu G. Metabolic reprogramming of macrophages during infections and cancer. Cancer Lett. 2019;452:14–22. doi: 10.1016/j.canlet.2019.03.015. [DOI] [PubMed] [Google Scholar]

[bib62] 62.House I.G., Savas P., Lai J., Chen A.X.Y., Oliver A.J., Teo Z.L., Todd K.L., Henderson M.A., Giuffrida L., Petley E.V., et al. Macrophage-Derived CXCL9 and CXCL10 Are Required for Antitumor Immune Responses Following Immune Checkpoint Blockade. Clin. Cancer Res. 2020;26:487–504. doi: 10.1158/1078-0432.Ccr-19-1868. [DOI] [PubMed] [Google Scholar]

[bib63] 63.Zhang B., Moorlag S.J., Dominguez-Andres J., Bulut Ö., Kilic G., Liu Z., van Crevel R., Xu C.J., Joosten L.A., Netea M.G., Li Y. Single-cell RNA sequencing reveals induction of distinct trained-immunity programs in human monocytes. J. Clin. Investig. 2022;132 doi: 10.1172/jci147719. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib64] 64.Dao V., Yuki K., Lo Y.H., Nakano M., Kuo C.J. Immune organoids: from tumor modeling to precision oncology. Trends Cancer. 2022;8:870–880. doi: 10.1016/j.trecan.2022.06.001. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib65] 65.Shultz L.D., Brehm M.A., Garcia-Martinez J.V., Greiner D.L. Humanized mice for immune system investigation: progress, promise and challenges. Nat. Rev. Immunol. 2012;12:786–798. doi: 10.1038/nri3311. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib66] 66.Llovet J.M., Kelley R.K., Villanueva A., Singal A.G., Pikarsky E., Roayaie S., Lencioni R., Koike K., Zucman-Rossi J., Finn R.S. Hepatocellular carcinoma. Nat. Rev. Dis. Primers. 2021;7:6. doi: 10.1038/s41572-021-00245-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib67] 67.Tuveson D., Clevers H. Cancer modeling meets human organoid technology. Science. 2019;364:952–955. doi: 10.1126/science.aaw6985. [DOI] [PubMed] [Google Scholar]

[bib68] 68.Gong Z., Huang L., Tang X., Chen K., Wu Z., Zhang L., Sun Y., Xia Y., Chen H., Wei Y., et al. Acoustic Droplet Printing Tumor Organoids for Modeling Bladder Tumor Immune Microenvironment within a Week. Adv. Healthc. Mater. 2021;10 doi: 10.1002/adhm.202101312. [DOI] [PubMed] [Google Scholar]

PERMALINK

Engineered Akkermansia muciniphila vesicles for targeted pyroptosis and trained immunity to enhance immunotherapy in hepatocellular carcinoma

Lanxiang Huang

Yuan Rong

Minghui Guo

Min Liu

Fei Long

Wei Zhong

Yue Hu

Xin He

Jiurong He

Diwei Zheng

Chunhui Yuan

Fubing Wang

Summary

Graphical abstract

Highlights

Introduction

Results

SMB nanoparticles target tumor cells and specifically accumulate in HCC tissues in vivo

Figure 1.

SMB nanoparticles specifically kill tumor cells and suppress tumor growth in vivo

Figure 2.

SMB nanoparticles trigger tumor cell pyroptosis via the NLRP3/Caspase-1/GSDMD pathway

Figure 3.

SMB nanoparticles remodel the tumor immunosuppressive microenvironment in HCC

Figure 4.

SMB nanoparticles stimulate macrophage trained immunity in tumor microenvironment

Figure 5.

SMB nanoparticles exhibit comparable tumor suppression and immune activation in HCC PDX and PDO models

Figure 6.

Figure 7.

Discussion

Limitations of the study

Resource availability

Lead contact

Materials availability

Data and code availability

Acknowledgments

Author contributions

Declaration of interests

STAR★Methods

Key resources table

Experimental model and study participant details

Human specimens

Mice

Cell lines

The syngeneic H22 orthotopic model

Patient-derived xenograft (PDX) model in humanized mice

Method details

Extraction of intestinal probiotic membrane vesicles

Bacterial external vesicle staining and cell-targeted uptake assay

Quantification of N-acetylgalactosamine by triple quadrupole mass spectrometry-liquid chromatography

Synthesis of engineered probiotic membrane vesicles

Characterization of engineered probiotic membrane vesicles

Cell culture

Uptake efficiency of nanoparticles in various cell types

Experimental steps for ultrasound intensity and release efficiency of SMB

Cytotoxicity assessment of SMB

Observation of pyroptosis-induced cell morphology by engineered probiotic membrane vesicles

Enzyme-linked immunosorbent assay

Enzyme-linked immunosorbent spot assay

Establishment of mouse hepatocellular carcinoma model and drug administration strategy

Transcriptome RNA-Seq analysis of mouse tumor tissues

Procedure for HE staining of mouse tumor tissue

Flow cytometry of single cell suspensions of mouse tumor tissue

Western blotting

Immunofluorescence staining and on immunohistochemical staining tumor tissue

Construction of the PDX model in humanised mice

Organoids culture

Quantification and statistical analysis

Footnotes

Contributor Information

Supplemental information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK