Biomimetic “Trojan Horse” delivery of traditional Chinese medicine with immunogenic hybrid vesicles of bacterial outer membrane and tumor cell membrane for synergistic chemo-immunotherapy of liver cancers

Abstract

Nowadays it is a big challenge to treat liver cancer, owing to the lack of efficient therapeutic paradigms. Currently, the first-line chemotherapy and immunotherapy often suffer from drug resistance, and severe side effects. The emerging Chinese medicine of Glycyrrhizic acid (GA) shows promise for liver cancer therapy, but often exhibits substantial toxicity and side effects. To explore efficient liver cancer treatment paradigms, herein, a “Trojan Horse” strategy is developed using biomimetic nanoplatform to deliver GA for dual chemo-immunotherapy of liver cancer. The nanoplatform (GANPs@COP) is composed of PLGA-coated GANP (GANPs@P) as the core, and hybrid vesicle (CO) of tumor cell membranes (CMVs) and bacterial outer membranes (OMVs) as the shell. In the nanoplatform, CMVs and OMVs contribute to tumor targeting and immunogenicity, respectively, while GA executes chemotherapy to locally destruct primary tumor cells. The nanoplatform synergistically ablate tumors by modulating tumor immunosuppressive microenvironment, promoting secretion of inflammatory cytokines, repolarizing macrophage M2 to M1, maturating dendritic cells, and eliciting CD8⁺ T cell infiltration, with significantly boosted therapeutic outcomes. Importantly, the nanoplatform even constructs immune memory effects for long-term tumor inhibition. This study offers a novel perspective for the combined application of traditional Chinese medicine therapy and immunotherapy of liver cancer.

Graphical abstract

(A) Hybrid vesicles (CO) are composed of bacterial outer membrane vesicles (OMVs) and tumor membrane vesicles (CMVs), PLGA-coated glycyrrhizic acid nanoparticles (GANPs@P) were then loaded into the hybrid membrane as a core to create a biomimetic nanoparticle, designated GANPs@COP. (B) GANPs@COP can target the tumor, execute chemotherapy, modulate the tumor microenvironment, boost anti-tumor immunity, and contribute to tumor therapy. Additionally, biomimetic nanoparticle exhibits immune memory effects, which can aid in tumor prevention. This approach offers new insights into the combined application of traditional Chinese medicine with other treatment modalities in cancer therapy.

1. Introduction

Liver cancer is the sixth most commonplace cancer and the third leading cause of cancer-related deaths worldwide [[1], [2], [3]], with most cases diagnosed at advanced stages and a morbidity-to-mortality ratio close to 1 [4]. Due to its asymptomatic progression and late detection, the 5-year survival rate for liver cancer is roughly 18 % [5]. Traditionally, surgery is the radical treatment option, but it only suitable for the patients at the early stage and there is not complicated tumor infiltration and metastasis [6], and only a small percentage of patients—between 5 and 15 %—are candidates for surgical resection [7]. Local ablation therapy, which has been deployed intensively in recent years [[8], [9], [10]], has the advantages of producing less of an impact on liver function, causing less trauma, and having a specific curative effect [11]. As a result, certain patients who are not candidates for surgical resection of liver cancer can also have the opportunity for radical resection [12]. Local ablation does have several drawbacks, though, and the main drawback of percutaneous ablation is the high necessity for the tumor location [13]. In addition, local ablation needs to be applied in conjunction with additional therapies, since it is challenging to accomplish complete ablation by itself and has greater technical requirements for treating large liver cancer [14,15].

To combat liver cancer, scientists have dedicated themselves to developing novel chemotherapeutic drugs, advanced therapeutic approaches such as immunotherapy, and even combinatory therapies [16]. Among the numerous chemotherapeutic agents developed for liver cancer, traditional Chinese medicine has garnered increasing attention due to its broad applicability [17], particularly in addressing late-stage chemotherapy resistance, tumor recurrence and metastasis, and surgical sequelae, demonstrating notable efficacy [18,19]. Notably, glycyrrhizic acid (GA), the principal component of the Chinese herb licorice, has been extensively utilized in Japan for the treatment of chronic hepatitis B [20]. GA enhances liver function [21] and, in some cases, can lead to complete recovery from hepatitis [22]. Previously, a few of studies have primarily combined GA with other anticancer agents (such as triptolide and adriamycin) to treat liver cancer [23,24]. Therefore, GA is a promising anticancer candidate [25,26]. However, the systematic administration of GA often result in significant toxicity and side effects [27]. So far, there are rare studies which use nanosystem to deliver GA for liver cancer treatment, although it is foreseeable that the nano delivery platform would significantly reduce system toxicity, enhance drug solubility, improve bioavailability and realize drug release in a controlled manner, with foreseeably better treatment outcomes [28].

Immunotherapy has been extensively studied as a treatment for liver cancer [29]. However, the systemic administration of immune drugs—such as cytokines, antibodies, and inhibitors—often faces challenges related to limited efficacy and severe side effects. These issues arise from their short half-life in circulation, as well as the potential for lethal on-target and off-tumor side effects, compounded by the tumoral immunosuppressive microenvironment (TIM) [[30], [31], [32]]. As known, targeted nanodrug delivery systems have the potential to specifically target tumor lesions, enhance pharmacokinetics, increase drug concentration at the site of the tumor, and decrease the therapeutic threshold dosage, which could significantly boost efficacy of both chemotherapy and immunotherapy [33,34]. Among various targeted nanodelivery systems, biomimetic “Trojan Horse” strategies that utilize cellular membranes and vesicles (such as cancer cell membranes, bacterial outer membranes, extracellular vesicles, immune cell membranes, and even hybrid vesicles) have garnered increasing interest [[35], [36], [37]]. Tumor cell membranes have been investigated as alternative vaccine candidates, offering advantages such as homologous targeting and a plentiful supply of tumor antigens [38,39]. Tumor cell membrane vesicles (CMVs) are believed to retain all tumor membrane antigens, including neoantigens, which not only play a crucial role in drug transport as delivery vesicles but also hold promise for personalized immunotherapy [40,41]. Nevertheless, membrane antigens are typically hypoimmunogenic and often fail to elicit a robust anti-tumor immune response.

Bacteria that contain multiple pathogen-associated molecular patterns (PAMPs) are recognized as some of the most common immune stimulants, which may lead to side effects associated with bacterial infections, such as bacterial toxins and sepsis [42]. Most recently, it has been demonstrated that bacterial outer membrane vesicles (OMVs), which contain numerous natural adjuvant components derived from the parent bacteria, are promising immunogenic generators for cancer immunotherapy, while exhibiting limited side effects due to their acellular nature without replicating risk [43,44]. In various cancer models, OMVs have successfully stimulated immune maturation and triggered inflammation in a controlled manner, facilitating the removal of cancer cells [45]. In addition to their immune functions, OMVs possess a soft and stable bilayer structure that contributes to effective delivery of both lipophobic and lipophilic drugs. Importantly, they can be easily produced on a large scale through fermentation and purification procedures. These inherent properties make OMVs an attractive candidate for cancer therapeutic nanoplatform development [[46], [47], [48]].

Recent studies have demonstrated that combinatory chemotherapy and immunotherapy can synergistically enhance treatment efficacy by integrating the advantages of each therapy, lowering the therapeutic threshold dosage, and alleviating side effects [1,8,14]. As OMVs possess strong immunogenic properties and CMVs exhibit robust tumor-targeting capabilities, hybrid vesicles composed of both OMVs and CMVs present an excellent “Trojan Horse” nanoplatform for drug delivery [35,49]. Recently, Park et al. co-injected synthetic bacterial vesicles (SyBV) and tumor tissue-derived extracellular vesicles (tEV) alongside anti-PD-1 inhibitors to treat melanoma efficiently [50]. In year 2021, Zou et al. hybridized OMVs with tumor-derived cell membranes (mT) to create functional vesicles (mTOMV), which succeeded in treating of 4T1 tumors, CT26 tumors, and lung metastases [51]. For another example, Wang et al. fused breast cancer cell membranes with OMVs to form a hybrid membrane (HM) coating on the IR780-loaded PLGA nanocores, which is highly effective to suppress breast cancer bone metastasis [49]. In view of the previous studies on nanovesicles, in this contribution, biomimetic hybrid nanovesicles (CO) composed of tumor cell membrane vesicles (CMVs) and bacterial outer membrane vesicles (OMVs) were formulated as coating on the core of GA-containing poly (lactic-co-glycolic acid) (PLGA) nanoparticles (GANPs@P), yielding for targeted dual chemo- and immunotherapy of liver cancer (Scheme 1A). Specifically, GANPs@COP can target tumors, execute chemotherapy, and modulate the tumor microenvironment, by inducing the polarization of M2 macrophages to M1, enhancing the secretion of anti-tumor cytokines, maturing dendritic cells, and promoting CD8⁺ T cell infiltration. This process boosts anti-tumor immunity and ultimately improves therapeutic outcomes (Scheme 1B). Importantly, COP exhibits immune memory effects, which can aid in tumor prevention. In this nanoplatform, CMVs enhance tumor targeting, while OMVs contribute to immunogenicity; simultaneously, glycyrrhizic acid facilitates chemotherapy to locally eliminate primary tumor cells. This innovative approach offers new insights into the combined application of traditional Chinese medicine with other treatment modalities in cancer therapy, and presents a novel perspective on integrating traditional Chinese medicine therapy with immunotherapy for liver cancer.

Scheme 1.

Schematic illustration of biomimetic nanosystems based on bacterial outer membrane vesicles (OMVs) with cancer cell membrane vesicles (CMVs) and glycyrrhizic acid nanoparticles (GANPs@COP) for tumor treatment and prevention. (A) Hybrid vesicles (CO) are composed of bacterial outer membrane vesicles (OMVs) and tumor membrane vesicles (CMVs), PLGA-coated GANPs (GANPs@P) were then loaded into the hybrid membrane as a core to create a biomimetic nanoparticle with a nuclear membrane structure, designated GANPs@COP. (B) GANPs@COP can target the tumor, execute chemotherapy, modulate the tumor microenvironment by inducing the polarization of M2 macrophages to M1 and enhancing the secretion of anti-tumor cytokines, boost anti-tumor immunity, and contribute to tumor therapy. Additionally, COP exhibits immune memory effects, which can aid in tumor prevention.

2. Results and discussion

2.1. Preparation and characterization of hybrid vesicles

Hybrid vehicles were created by fusing OMVs from Pseudomonas aeruginosa (PA) with CMVs derived from mouse hepatoma cell line H22. OMVs were isolated from the culture supernatant of PA using ultracentrifugation [45,47]. H22 cells served as model cancer cell lines for isolating CMVs through grinding and superfreezing centrifugation [40,41]. The hybrid membrane-coated glycyrrhizic acid nanoparticles (GANPs@COP) were synthesized using OMVs, CMVs, and PLGA-loaded GANPs (GANPs@P) as nanocores. For consistency, individual CMVs or OMVs were also supplemented with a PLGA nanocore, designated as CMVP or OMVP, respectively. The transmission electron microscopy (TEM) image (Fig. 1A and Fig. S1) revealed that the nanoparticles coated by the vesicles were spherical, evenly distributed, and uniformly sized. The distinct core-shell structure was evident, consisting of an inner PLGA core and outer encapsulated membrane vesicles. The results of dynamic light scattering (DLS) analysis (Fig. 1B and C) indicated that the average hydrodynamic diameters of CMVP, OMVP, COP, GANPs@P, and GANPs@COP were 62.21 nm, 89.96 nm, 65.06 nm, 70.03 nm, and 84.55 nm, respectively. The average ζ potentials were −11.47 mV, −15.57 mV, −12.57 mV, −14.33 mV, and −18.53 mV, respectively. The small differences in hydrodynamic diameter and ζ potentials suggest that the core-shell structure exhibits high stability.

Fig. 1.

Preparation and characterization of COP. (A) The TEM images of CMVP, OMVP, COP, GANPs@P and GANPs@COP (scale bar: 200 nm). The hydrodynamic size (B) and ζ potential (C) of CMVP, OMVP, COP, GANPs@P and GANPs@COP. (D) SDS-PAGE protein analysis of CMV, OMV and CO. (E) Confocal microscopy colocalization in BMDC cells of membrane vesicles (scale bar: 20 μm).

To verify the formation of hybrid vesicles, the retention of membrane proteins on CMV, OMV, and CO was assessed using sodium dodecyl sulfate polyacrylamide gel electrophoresis [52] (SDS-PAGE) (Fig. 1D). Compared to single vesicles, the hybrid vesicle CO retained the surface membrane proteins of both CMV and OMV, indicating that CMV and OMV had successfully hybridized into vesicle CO. To further verify the successful fusion of hybrid vesicles, single vesicles were fluorescently labeled with membrane dyes. CMVP was labeled with DiI, resulting in red fluorescence, while OMVP was labeled with DiO, producing green fluorescence. Single vesicles were stained and subsequently fused to hybrid vesicles known as COP. Following this, CMVP, OMVP, and COP were incubated with bone marrow-derived dendritic cells (BMDCs) isolated from mice, and the fusion of the membrane vesicles was observed using confocal laser scanning microscopy [49] (CLSM). Fig. 1E illustrates that the fluorescence signals of red and green are distinct, indicating that the fluorescence signals of DiI and DiO are separated. After fusion, the two vesicles formed new lipid vesicles, which exhibited co-localization of red and green fluorescence signals in the cytoplasm of BMDCs, demonstrating the co-localization of DiI and DiO. We observed that BMDCs were more inclined to endocytose OMVs. This preference is attributed to the abundance of pattern recognition receptors (PRRs) on the surface of BMDCs, which facilitate the recognition of pathogen-associated molecular patterns (PAMPs) present on the surface of OMVs. The data presented above indicate that we have successfully prepared hybrid vesicles.

2.2. Hybrid vesicles could stimulate immune response in vitro

CO is a type of combined anti-tumor nano-system that can initiate an anti-tumor immune response. OMVs serve as adjuvants, inducing the production of cytokines such as IFN-γ and TNF-α, and activating DCs [45,49]. Additionally, the expression of specific antigens by CMVs can be processed by mature antigen-presenting cells (APCs) to trigger an anticancer immune response mediated by cytotoxic T lymphocytes [52]. To investigate this synergistic effect, CMVP, OMVP, and COP were co-incubated with BMDCs for 24 h. Subsequently, flow cytometry was employed to detect the expression of CD80, a marker of mature DCs [53]. As shown in Fig. 2A and B, vesicle-coated nanoparticles exhibit varying roles in inducing BMDC maturation. Notably, the proportion of mature BMDCs in the COP and OMVP groups was significantly higher than that in the CMVP group, indicating that the OMV component possesses strong adjuvant activity. This is attributed to the surface expression of microbial-associated molecular patterns (MAMPs) from OMVs, which can directly interact with PRRs on BMDCs to stimulate DC maturation. COP demonstrated the most pronounced effect in promoting DC maturation, suggesting that CMVs and OMVs can work synergistically.

Fig. 2.

Hybrid vesicles can stimulate immune response in vitro. Representative flow cytometry images (A) and quantification analysis (B) of CD80 on the surface of BMDCs to reflect the maturation of DCs. Representative flow cytometry images (C) and quantification analysis (D) of CFSE to reflect the proliferation of T cells. Relative RNA expression of IL-1β, IL-6, TNF-α (E) and IL-1α, IL-12, IFN-γ (F) in RAW264.7 cells measured by RT-qPCR. ∗p ≤ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001. ns, not significant.

Further, we evaluated the proliferation of T cells following treatment with BMDCs in vitro using flow cytometry [54]. T cells were isolated from the spleens of mice and incubated with CMVP, OMVP, or COP treated BMDCs for 24 h. As shown in Fig. 2C–D, T cell proliferation induced by CMVP and OMVP was significantly increased, with rates of 41.5 % and 75.4 %, respectively, compared to the untreated CTL control group (20.1 %). Notably, COP treatment resulted in an even higher proliferation rate of 96.4 %. Our results suggest that the hybrid vesicle COP has significant potential for activating the immune response from DCs to T cells. Subsequently, flow cytometry was employed to detect the two primary cytotoxic cytokines, granzyme B and IFN-γ, secreted by CD8⁺ T cells [55]. As illustrated in Fig. S2, both OMVP and COP induced a significant increase in the expression of CD8⁺ T cells secreting granzyme B and IFN-γ compared to the untreated CTL control. This finding further indicates that the hybrid vesicle COP plays a crucial role in the activation of CD8⁺ T cells. These results suggest that COP may be integral to the adaptive immune response.

Innate immune cells play a crucial role in anti-tumor immunity [56]. Macrophages express a variety of PRRs that recognize PAMPs on the surface of OMVs [45,57]. Upon activation, macrophages secrete a range of cytokines that contribute to immune regulation and target cell destruction. In this study, we co-incubated CMVP, OMVP, and COP with RAW264.7 cells for 24 h, followed by RT-qPCR to assess cytokine expression in RAW264.7 cells. As illustrated in Fig. 2E and F, the RNA expression levels of IL-1α, IL-1β, IL-6, IL-12, TNF-α and IFN-γ were significantly elevated following treatment with CMVP and OMVP. Notably, COP exhibited the most pronounced effect on the secretion of these cytokines. The synergistic effects of IL-1β, IL-12, and IFN-γ were particularly evident. These findings suggest that the hybrid vesicle COP can activate macrophages to elicit subsequent anti-tumor effects, indicating that COP may play a vital role in the innate immune response.

2.3. Hybrid vesicles combined with GANPs can effectively kill tumor cells in vitro

GANPs synthesized from GA demonstrate significant potential to inhibit tumor cell growth. We first evaluated the impact of GANPs on tumor cell proliferation in vitro. Here, we utilized two mouse hepatoma cell lines, H22 and Hepa1-6, along with MLE-12 cells (mouse alveolar epithelial cells) as normal controls. As illustrated in Fig. 3A, with a gradual increase in GANP concentration from 0.2 mg/mL to 1.2 mg/mL, the survival rate of normal MLE-12 cells remained nearly 100 %. In contrast, the survival rates of the tumor cells, H22 and Hepa1-6, decreased progressively with increasing GANP concentrations. At a concentration of 1.2 mg/mL, the survival rates of H22 and Hepa1-6 cells were approximately 58 % and 66 %, respectively. These results indicate that GANPs do not adversely affect the growth of normal cells but exhibit a notable inhibitory effect on the proliferation of hepatoma cells.

Fig. 3.

Hybrid vesicles combined with GANPs can kill tumor cells in vitro. (A) Cytotoxicity of GANPs on mouse alveolar epithelial cells (MLE-12) and mouse hepatocarcinoma cell line (H22 and Hepa1-6). (B, C) Cytotoxicity of CMVP, OMVP, COP and GANPs@COP on H22 cells (B) and MLE-12 cells (C). (D) Representative fluorescence images of the killing effect of GANPs at different concentrations on H22 and Hepa1-6 cells. (E) Representative fluorescent images of the killing effects of CMVP, OMVP, COP and GANPs@COP on H22 and Hepa1-6 cells. (Scale bar: 100 μm).

Subsequently, we further evaluated the effects of CMVP, OMVP, COP, and GANPs@COP on the growth of normal and tumor cells in mice. As shown in Fig. 3B-C and Figure S3, CMVP, OMVP and COP did not significantly affect the viability of MLE-12, H22 and Hepa1-6 cells at 100 μg/mL. The survival rates of H22 and Hepa1-6 cells after GANPs@COP treatment were approximately 60 % and 67 %, respectively, while GANPs@COP treatment did not affect the growth of MLE-12 cells. Consequently, GANPs and GANPs@COP exhibited nearly identical therapeutic effects on H22 tumor cells, while the effects of CMVP, OMVP, and COP began to diverge at 150 μg/mL. As the concentration of vesicles increased, the impacts of CMVP, OMVP, COP and GANPs@COP on H22 cell viability became increasingly pronounced. When the concentration of vesicles reached 300 μg/mL, CMVP reduced H22 cell viability to approximately 67 %, while OMVP and COP decreased viability to about 62 % and 53 %, respectively. In contrast, GANPs@COP reduced cell viability to around 32 %, indicating that the combination of CMV and OMV with GANPs exhibited potent tumor-killing effects. Hybrid vesicles demonstrated a significant advantage over OMVP and CMVP when used alone. Even when the vesicles were as high as 300 μg/mL, there was still no significant toxicity to MLE-12 cells, indicating that the vesicles are safe and non-toxic to normal cells.

Next, we utilized a fluorescence microscope to visually assess whether GANPs and GANPs@COP can induce cell death in tumor cells. Here, H22 and Hepa1-6 cells were stained using a calcein-AM/PI double staining kit, where green fluorescence indicated living cells and red fluorescence represented dead cells. As shown in Fig. 3D, with the gradual increase in GANPs concentration, the number of dead H22 and Hepa1-6 cells also increased. At a GANP concentration of 1.2 mg/mL, the number of dead cells for both H22 and Hepa1-6 reached its peak, indicating that GANPs exert a significant cytotoxic effect on mouse liver cancer cells. Fig. 3E shows that CMVP, OMVP, and COP do not exhibit any cytotoxic effects on H22 and Hepa1-6 cells, while GANPs@COP demonstrated a notable cytotoxic effect on these tumor cells, as it retains the tumoricidal properties of GANPs. The results obtained from flow cytometry analysis (Fig. S4) corroborated the findings presented in Fig. 3D and E. These results suggest that GANPs possess a significant cytotoxic effect on tumor cells, and GANPs@COP can inherit the tumoricidal properties of GANPs.

2.4. The biological safety of GANPs@COP in vivo

The results above confirmed the properties of COP in activating immune responses in vitro and demonstrated the excellent efficacy of GANPs@COP in tumor eradication. Furthermore, we continued to explore the effects of COP and GANPs@COP on tumor therapy in vivo. Initially, we systematically assessed the biocompatibility of COP and GANPs@COP in vivo. Thirty-six 6-week-old healthy female BALB/C mice were randomly divided into six groups: CTL, CMVP, OMVP, COP, GANPs@P, and GANPs@COP. Each group received an injection via the tail vein of either PBS, CMVP, OMVP, COP, GANPs@P, or GANPs@COP. The dosages administered were 30 mg/kg for GANPs and 2 mg/kg for vesicles. Seven days post-injection, the heart, liver, spleen, lung, and kidney of the mice were harvested, sectioned, and stained with H&E. Fig. 4 illustrates that there were no significant histological abnormalities or pathological changes in any of the organs, indicating that CMVP, OMVP, COP, GANPs@P, and GANPs@COP exhibited no apparent toxicity in vivo. The above data indicate that COP and GANPs@COP are safe and non-toxic to the major organs of mice, making them suitable for subsequent animal experiments, and further highlight the significant application potential of COP and GANPs@COP. The direct use of GA can result in obvious toxicity and side effects [27], while GA was formulated into nanoparticles and encapsulated by membranes to significantly improve the biocompatibility of GA.

Fig. 4.

Representative H&E staining images of the main organs from mice injected with PBS, CMVP, OMVP, COP, GANPs@P or GANPs@COP (n = 6). (Scale bar: 100 μm).

2.5. Tumor targeting and in vivo therapeutic effect

It has been reported that chemokines secreted by tumor regions exert chemotactic effects on OMV, while tumor cell-derived tumor membrane vesicles (CMV) exhibit homologous targeting to the same tumor [35]. Therefore, hybrid vesicles formed from the fusion of OMV and CMV may possess tumor-targeting capabilities. We utilized a subcutaneous hepatocellular carcinoma model to investigate the aggregation of hybrid vesicle COP at the tumor site. Here, we loaded indocyanine green [58] (ICG) into CMVP, OMVP, and COP, which carry fluorescence suitable for in vivo imaging. 6-week-old healthy BALB/c female mice were inoculated subcutaneously with H22 cells. When the tumor volume reached 60–100 mm³, CMVP, OMVP and COP were injected into the tail vein at a dosage of 2 mg/kg per mouse, with PBS administered as the control group. 12 h and 36 h post-injection, the Xenogen IVIS imaging system was employed to visualize tumors and major organs in different experimental groups. As shown in Fig. 5A, CMVP, OMVP, and COP aggregated at the tumor sites in the mice. As can be seen from Fig. 5B–C and Fig. S5, 12 h after intravenous injection, fluorescence was predominantly localized in the liver and tumor sites. In contrast, at 36 h post-injection, fluorescence was primarily observed in the liver, kidney, and tumor sites. This observation suggests that CMVP, OMVP, and COP are primarily metabolized by the liver and kidney. Additionally, OMVP was detected in small amounts in the spleen and lungs. This may be attributed to OMVP's tendency to accumulate in organs with a rich immune system, with its presence in the lungs potentially resulting from absorption by the reticuloendothelial system [49]. Statistical analysis of fluorescence at the tumor site revealed that OMVP exhibited some tumor-targeting capability, while CMVP demonstrated superior targeting compared to OMVP, with the fluorescence signal decreasing over time. Notably, the hybrid vesicle COP displayed the strongest tumor-targeting ability, indicating a synergistic effect between CMV and OMV. These results suggest that COP can effectively accumulate at the tumor site, providing a solid foundation for subsequent tumor treatment experiments.

Fig. 5.

Tumor targeting and systematic therapeutic effects. (A) In vivo imaging of mice to observe the aggregation of ICG fluorescence at the tumor site (n = 6). ICG fluorescence imaging (B) and relative quantitative analysis (C) of CMVP, OMVP and COP in major organs and tumors 12 h after intravenous injection (n = 6). (D) Images of tumors in different treatment groups (n = 8). (E) Tumor growth curve of mice under different treatments (n = 8). (F) Weight of tumors in different treatment groups (n = 8). (G) The body weight changes of mice in different treatment groups in H22 model (n = 8). (H) Survival curves of mice in different treatment groups. ∗p ≤ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001. ns, not significant.

Now that tumor targeting has been established, what about the efficacy of tumor therapy? To address this question, 48 healthy female BALB/c mice aged 6–8 weeks were randomly divided into six groups: CTL, GANPs@P, CMVP, OMVP, COP, and GANPs@COP. Tumor cells were subcutaneously inoculated using H22 cells. When the tumor volume reached 60–100 mm³, PBS, GANPs@P, CMVP, OMVP, COP, or GANPs@COP were injected via the tail vein, with GANPs administered at a dosage of 30 mg/kg and vesicles at 2 mg/kg for each mouse. The body weight and tumor size of the mice were measured the following day, and tumor volume was calculated. On day 19, when the maximum tumor diameter approached 1.5 cm, the tumor, spleen, and lymph nodes were dissected. As shown in Fig. 5D–H, compared to the CTL group, tumor growth was significantly inhibited following treatment with GANPs@P, CMVP, OMVP, COP, or GANPs@COP, resulting in smaller tumor sizes and lighter tumor weights. The survival rate of the mice improved, while their body weight remained unaffected. More importantly, the therapeutic effect of the COP treatment group was significantly superior to that of the CMVP and OMVP groups, and the GANPs@COP treatment group demonstrated significantly better outcomes than the GANPs@P and COP groups. Previous studies have indicated that GA combined with other drugs can effectively combat liver cancer [23,24], OMVs can possess strong immunogenic properties, CMVs can exhibit robust tumor-targeting capabilities [35,49], and we also advocated that GANPs, CMV and OMV can exert a synergistic targeted anti-tumor effect while remaining safe and non-harmful to the mice.

2.6. Regulation of tumor immunosuppressive microenvironment

The above results indicate that GANPs@COP can play a role in anti-tumor activity. However, what is the mechanism through which this anti-tumor effect is achieved? One of the challenges in tumor therapy is the immunosuppressive microenvironment [57,59]. The tumor microenvironment significantly influences tumor growth, metastasis, and drug resistance [60,61]. Therefore, we next investigated whether GANPs@COP could modulate the tumor immunosuppressive microenvironment. Tumor tissues from each group were homogenized and stained, followed by flow cytometry to assess immune cell subsets and cytokine secretion within the tumor microenvironment. The results demonstrated that, compared to the CTL group, the proportion of M1/M2 macrophages increased in the GANPs@P, CMVP, OMVP, COP, and GANPs@COP treatment groups (Fig. 6A and D), while the proportion of Tregs decreased (Fig. 6B and E). Additionally, the levels of IFN-γ, TNF-α, and granzyme B were elevated (Fig. 6G-I and Figure S6). Notably, the regulatory effect of the COP treatment group was significantly superior to that of the CMVP and OMVP groups, while the GANPs@COP group exhibited a markedly greater effect than the GANPs@P and COP groups. This suggests that CMV, OMV, and GANPs can work synergistically to regulate the tumor microenvironment. These findings indicate that GANPs@COP can modulate the tumor microenvironment by inducing the polarization of M2 macrophages to M1, reducing the proportion of Tregs, and enhancing the secretion of anti-tumor cytokines such as IFN-γ, TNF-α, and granzyme B.

Fig. 6.

GANPs@COP can regulate the immune microenvironment. Representative flow cytometry analysis images (A) and M1/M2 ratio in tumors (D) of M1-like macrophages (F4/80⁺Ly6c⁺ gating on CD11b⁺ cells) and M2-like macrophages (F4/80⁺CD206⁺ gating on CD11b⁺ cells). Representative flow cytometry images (B) and relevant quantitative analysis (E) of Tregs (CD4⁺CD25⁺Foxp3⁺ T cells) in tumors. Representative flow cytometry images (C) and relevant quantitative analysis (F) of CD80⁺CD86⁺ DC cells in lymph nodes. Flow cytometry quantitative analysis of IFN-γ (G), TNF-α (H) and granzyme B (I) in tumors. (n = 8) ∗p ≤ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001. ns, not significant.

In addition, we further investigated whether GANPs@COP could influence the changes in immune cell subsets within the spleen and lymph nodes. The results indicated that, compared to the CTL group, the proportion of CD8⁺ T cells in the spleen was increased across all treatment groups (Fig. S7), as was the proportion of DC cells in the lymph nodes (Fig. 6C and F). Notably, the enhancement observed in the COP treatment group was significantly greater than that in the CMVP and OMVP groups, while the effect of the GANPs@COP treatment group was markedly superior to that of the GANPs@P and COP groups. This suggests that CMV, OMV, and GANPs may also exhibit a synergistic effect, and GANPs@COP can exert an anti-tumor effect by activating CD8⁺ T cells in the spleen and promoting the maturation of DC cells in the lymph nodes. OMVs can successfully stimulated immune maturation and triggered inflammation to facilitate the removal of cancer cells [45], CMVs also have immunotherapy potential [40,41] and GA has anti-tumor effects [25,26]. We demonstrated here that GANPs@COP modulate tumor immunosuppressive microenvironment by promoting secretion of inflammatory cytokines, repolarizing M2 macrophage to M1 macrophage, maturating DCs, eliciting CD8⁺ T cell infiltration and reducing the proportion of Tregs, thereby synergistically ablating tumors.

2.7. Tumor prevention effect in vivo

The above results confirm that hybrid vesicles can activate anti-tumor immune response, thereby playing a significant role in tumor suppression. The long-term role of this hybrid vesicles in tumor surveillance and suppression remains to be explored. Since the migration of vaccines to lymph nodes and their absorption by APCs are prerequisites for eliciting immunity, we investigated the ability of COP loaded with ICG to target lymph nodes in vivo using fluorescence imaging. Four hours after the administration of CMVP, OMVP, or COP loaded with ICG into the left paw of mice, the lymph nodes were dissected, and the fluorescence signal of ICG was observed. As shown in Fig. 7A, the fluorescence intensity of the lymph nodes in the CMVP, OMVP, and COP groups was significantly increased compared to CTL group, which was injected only with PBS. Notably, the fluorescence intensity in the OMVP group was stronger than that in the CMVP group. This may be attributed to the presence of numerous microbial-associated molecules, bacteria-derived lipids, and cholesterol on the surface of OMV, which have been shown to facilitate the entry of OMV into DCs and thus promote greater accumulation in lymph nodes. Importantly, COP exhibited the strongest lymph node accumulation effect, suggesting that the hybrid vesicle COP can inherit and enhance the superior performance of CMV and OMV, achieving a beneficial synergistic effect.

Fig. 7.

Effect of tumor prevention in vivo. (A) Ex vivo ICG fluorescence image and quantitative analysis of isolated lymph nodes after injection of CMVP, OMVP and COP (n = 6). (B) Images of tumors in H22 tumor prevention model (n = 8). (C) Tumor growth curve of mice under different treatments in H22 tumor prevention model (n = 8). (D) Weight of tumors in H22 tumor prevention model (n = 8). (E) The body weight changes of mice in H22 tumor prevention model (n = 8). (F) Survival curves of mice under different treatments in H22 tumor prevention model. Representative flow cytometry images (G) and quantitative analysis (H) of proportions of effector memory T cells (CD3⁺CD8⁺CD44⁺CD62L⁻) and central memory T cells (CD3⁺CD8⁺CD44⁺CD62L⁺) expressed by splenic lymphocytes after immunizations (n = 8). Representative flow cytometry images (I) and relevant quantitative analysis (J) of CD80⁺CD86⁺ DC cells in lymph nodes after immunizations (n = 8). ∗p ≤ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001. ns, not significant.

Next, we further investigated the effect of COP on tumorigenesis in mice. 32 healthy BALB/c female mice aged 6–8 weeks were randomly divided into 4 groups, namely CTL, CMVP, OMVP and COP. Intradermal immunization was performed with PBS, CMVP, OMVP or COP, respectively, and the vesicle was 1.5 mg/kg for each mouse. Intradermal immunization test was conducted three times. At an interval of 7 days, 7 days after the end of the last intradermal immunization, H22 cells were inoculated subcutaneously to construct mouse liver cancer models. The body weight and tumor size of mice were measured the next day, and tumor volume was calculated. On day 21, when the maximum tumor diameter was close to 1.5 cm, the tumor, spleen and lymph nodes were dissected. As shown in Fig. 7B–F, compared with CTL group, single vesicle CMVP and OMVP had a certain inhibitory effect on tumor growth, and the survival rate of mice was improved to a certain extent, resulting in smaller tumor volume and lighter tumor weight. In contrast, COP hybrid vesicles can inhibit tumor growth more effectively. In the prevention model, the weight changes in the mice after these vesicles were small and negligible, suggesting a good safety profile.

To further clarify the mechanism of COP in preventing tumorigenesis, we analyzed the changes of immune memory T cell subsets in spleen and DC cell subsets in lymph nodes by flow cytometry. Central memory T cells (T_CM) and effector memory T cells (T_EM) are subsets of memory T cells found in peripheral circulation. T_CM cells exhibit lymphoid homing characteristics and possess a high capacity for proliferation. Upon re-stimulation by tumor antigens, T_CM cells can rapidly proliferate and differentiate into effector T cells, providing immune protection. T_EM cells are activated T_CM cells that, upon re-exposure to tumor antigens, can rapidly produce effector cytokines, which play a crucial role in tumor inhibition [32,[62], [63], [64]]. The results showed that compared with the CTL group, the proportion of T_EM (CD3⁺CD8⁺CD44⁺CD62L⁻) in spleen and DC cells in lymph nodes increased significantly after CMVP, OMVP and COP inoculation (Fig. 7G-J). Similarly, the promotion effect of COP treatment group was significantly better than that of CMVP and OMVP groups, indicating that OMV can be used as an adjuvant to enhance the anti-tumor immunity effect of CMV as tumor therapeutic vaccine. This suggests a possible mechanism for COP-mediated effective tumor prevention.

In addition, we further analyzed the cell subsets and cytokine secretion of tumor tissue. As shown in the results, compared to CTL group, the proportion of T_EM cells in CMVP, OMVP and COP groups was significantly increased (Fig. S8A–B). In addition, the proportion of anti-tumor cytokines IFN-γ and TNF-α increased (Fig. S8C–F). IFN-γ is a typical marker of cytotoxic T lymphocyte activation. Notably, OMV inoculations stimulated effector memory T cell differentiation and secretion of cytokines IFN-γ and TNF-α more strongly than CMV inoculations, which is consistent with previous reports that OMV can activate cellular immune responses and inflammation. The immunostimulant effect of hybrid vesicle COP may be due to the strong adjuvant activity of OMV, indicating that the treatment can effectively induce tumor-specific immunity. In summary, these data suggest that hybrid vesicle COP can play a role in tumor prevention by enhancing immune memory effect and anti-tumor immune response. CMVs, which are believed to retain all tumor membrane antigens, including neoantigens, have been investigated as alternative vaccine candidates [40,41], while OMVs contain numerous natural adjuvant components derived from the parent bacteria [43,44], and we here demonstrate the tremendous potential of CMVs combined with OMVs as tumor therapeutic nanoplatforms.

3. Conclusions

In this study, we demonstrate that biomimetic “Trojan Horse” delivery of glycyrrhizic acid is an effective method for achieving dual chemo-immunotherapy for liver cancer, utilizing bacterial outer membrane vesicles and cancer cell membranes. In the H22 subcutaneous liver cancer model, GANPs@COP can modulate the tumor immunosuppressive microenvironment and enhance the anti-tumor immune response, thereby inhibiting tumor growth. Additionally, COP contributes to tumor prevention by strengthening immune memory and the anti-tumor immune response. While GANPs have shown therapeutic effects on liver cancer, their efficacy in lung cancer, colorectal cancer, melanoma, breast cancer, and other tumors requires further experimental investigation. Moreover, other target molecules or tumor markers can be expressed on the surface of vesicles, allowing for expanded applications. The GANPs@COP platform exhibits high scalability, enabling the integration of glycyrrhizic acid-based traditional Chinese medicine therapy with immunotherapy. This approach presents a novel strategy for combining traditional Chinese medicine with other cancer treatment modalities.

CRediT authorship contribution statement

Zhaoyan Zhao: Writing – original draft, Validation, Methodology, Funding acquisition, Conceptualization. Lanxin Luo: Software, Methodology, Formal analysis, Data curation. Zhongting Wang: Software, Investigation. Bing Guo: Writing – review & editing, Visualization, Validation, Supervision, Funding acquisition. Changjun Gao: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition.

Notes

The authors declare no competing financial interest.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

ABBREVIATIONS

CMVs

cancer cell membrane vesicles

OMVs

bacterial outer membrane vesicles

GA

Glycyrrhizic acid

GANPs

glycyrrhizic acid nanoparticles

CO

hybrid vesicles of CMV and OMV

CMVP

CMVs loaded with PLGA

OMVP

OMVs loaded with PLGA

COP

CO loaded with PLGA

GANPs@P

PLGA-coated GANPs

GANPs@COP

CO loaded with GANPs@P

PAMPs

pathogen-associated molecular patterns

PA

Pseudomonas aeruginosa

DLS

dynamic light scattering

SDS-PAGE

sodium dodecyl sulfate polyacrylamide gel electrophoresis

PRRs

pattern recognition receptors

DCs

dendritic cells

BMDCs

bone marrow derived DCs

MAMPs

microbial-associated molecular patterns

ICG

indocyanine green

Tregs

regulatory T cells

Appendix A. Supplementary data

The following is the supplementary data to this article:

Data availability

Data will be made available on request.