pH-triggered Mn-ZIF8 nanoreactor enhances tumor immunotherapy by amplifying the cGAS-STING pathway through a self-reinforcing gas-metal synergy strategy

Abstract

Although conventional immunotherapies have significantly transformed cancer treatment, their efficacy is still fundamentally limited by the immunosuppressive tumor microenvironment and systemic toxicity. In this study, we propose a Gas-Metal Synergy Strategy, which integrates immune activation and biosafety, by engineering a pH-responsive manganese-based zeolitic imidazolate framework (named MRPH) nanoplatform co-loaded with the nitric oxide (NO) donor RRX-001. MRPH selectively dissociates in the acidic TME, releasing Mn²⁺ and NO, which together amplify the activation of the cGAS-STING pathway through a synergistic mechanism: NO induces mitochondrial damage and the cytosolic release of mtDNA, while Mn²⁺ enhances the sensitivity of cGAS to mtDNA. Simultaneously, PEG-HA-modified nanoparticles enable tumor-targeted delivery, and this spatiotemporal coordination triggers immunogenic cell death, dendritic cell maturation, and CD8⁺ T cell infiltration, while inhibiting CD47-mediated immunosuppression and promoting tumor vascular normalization. Furthermore, our findings indicate that manganese-related genes are linked to immune modulation and tumor microenvironment remodeling in osteosarcoma patients. Both in vitro and in vivo studies demonstrate that MRPH significantly enhances gas-amplified metalloimmunotherapy. This work pioneers a low-toxicity paradigm that integrates gas therapy and metal-based immunotherapy, offering a transformative approach to solid tumor immunotherapy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-025-03950-6.

Introduction

Recent breakthroughs in cancer immunotherapy, including immune checkpoint inhibitors (e.g., PD-1/PD-L1 and CTLA-4 blockers), chimeric antigen receptor T-cell (CAR-T) therapy, and cancer vaccines, have significantly improved clinical outcomes [1, 2]. However, current immunostimulatory strategies face critical challenges: off-target toxicity, immune hyperactivation, immunosuppressive tumor microenvironment (TME) driven by CD47-mediated “don’t eat me” signaling, and the biophysical barriers present in solid tumors that limit drug penetration and immune efficacy [3, 4]. These limitations highlight an urgent demand for novel immunotherapeutic paradigms that harmonize efficacy with biosafety.

Metal ion-mediated immunity, leveraging unique physicochemical properties and biological functions, has emerged as a promising frontier [5]. Metal ions can participate in immune signal transduction by modulating pattern recognition receptor activity and influencing immune cell metabolic reprogramming [6]. Specifically, manganese ions (Mn²⁺) play dual roles in immune activation: (1) stabilizing the cyclic GMP-AMP synthase (cGAS)-double-stranded DNA (dsDNA) complex to enhance cytoplasmic DNA sensitivity, and (2) directly activating the stimulator of interferon genes (STING) pathway [7]. This cascade triggers type I interferon and pro-inflammatory cytokines production, thereby promoting immunogenic cell death (ICD) and enhancing anti-tumor immune responses [8]. Despite their potential, monotherapies with metal ions suffer from poor targeting efficiency, limited STING activation capacity, and systemic toxicity in solid tumors [9]. To overcome these challenges, recent strategies have focused on metal-based nanomaterials such as manganese oxide nanoparticles [10, 11] and metal-organic frameworks [10, 12], which enhance delivery efficiency and reduce off-target effects. Notably, studies integrating gas therapy with metal ions have demonstrated synergistic immunomodulatory effects [13, 14], though the precise coordination mechanisms underlying how gas molecules amplify Mn²⁺-mediated cGAS-STING signaling remain to be fully elucidated.

Gas therapy has gained traction in oncology for its excellent biosafety profile and unique therapeutic mechanisms [13, 15]. Nitric oxide (NO), the first discovered bioactive gas, serves not only as a crucial endogenous signaling molecule but also demonstrates multifaceted anti-tumor effects, including direct cytotoxicity and vascular normalization [16]. Literature reports indicate that NO can induce mitochondrial damage by reducing mitochondrial membrane potential (MMP), leading to cytoplasmic release of mitochondrial DNA (mtDNA) [17]. Since mtDNA is dsDNA, its release may be recognized by cGAS in the cytoplasm, suggesting that NO has the potential to activate the cGAS-STING pathway, thereby enhancing anti-tumor immune responses. However, the short half-life of NO (~ 2 s) [18] and limited diffusion distance (< 200 μm) [19] restricts its effective concentration in the tumor microenvironment, and systemic delivery may cause blood toxicity, limiting clinical application [20]. Therefore, utilizing a nanoplatform with NO to activate the cGAS-STING pathway may become a novel strategy for anti-tumor gas immunotherapy.

To address these challenges, we propose an innovative Gas-Metal Synergy strategy involving the construction of a manganese-based zeolitic imidazolate framework (Mn-ZIF8) nanoplatform. This synergy strategy, which incorporates tumor therapy with multiple interactions, is crucial for optimizing anti-tumor treatment effects [21–24]. Mn-ZIF8 is an metal-organic framework (MOF) composed of metal ions or metal clusters coordinated with organic ligands, featuring a high specific surface area, tunable pore structure, and good biocompatibility [25–27]. Importantly, Mn-ZIF8 exhibits unique pH-responsive properties [28, 29], remaining stable in physiological environments while selectively dissociating in the acidic TME to release Mn²⁺ and the encapsulated NO donor 1-bromoacetyl-3,3-dinitro-1,2,4-triazole (RRX-001). Compared to traditional NO donors, RRX-001 can release high concentrations of NO (> 150 µM) in the hypoxic tumor microenvironment [30] due to its unique dinitrotriazole structure, and it can enhance macrophage phagocytosis function by inhibiting CD47 expression and blocking the “don’t eat me” signal, thereby overcoming tumor cell immune evasion [31, 32]. The mechanism of action of this nanoplatform can be summarized as the NO-mtDNA-Mn²⁺ ternary cascade amplification effect. First, NO released from RRX-001 disrupts the integrity of tumor cell mitochondria, promoting the release of mtDNA into the cytoplasm. Second, Mn²⁺ enhances the recognition sensitivity of cGAS to cytoplasmic dsDNA. Finally, the synergistically activated STING pathway induces ICD, effectively enhancing tumor immunity.

The TME of osteosarcoma (OS) is characterized by an acidic, hypoxic environment, elevated CD47 expression on tumor cells, and limited sensitivity to immune therapies, resulting in a low objective response rate. Macrophages are the predominant immune cells infiltrating the tumor. These unique features make OS an ideal model for investigating manganese metabolism and its potential role in immune modulation. Based on this, this study constructs a LASSO regression model and identifies that manganese-related genes (MRGs) are linked to tumor microenvironment remodeling and immune regulation. Based on this finding, this study designs and synthesizes a Mn-ZIF8 nanoplatform with intelligent pH-responsive drug delivery, encapsulating the NO donor RRX-001 and surface-modified with PEG-HA for tumor-targeted delivery. After intravenous injection, this delivery system targets OS region, dissociating in the acidic microenvironment to release Mn²⁺ and RRX-001, the latter of which releases high concentrations of NO in hypoxic areas. This strategy minimizes systemic NO exposure and significantly amplifies immune activation of the cGAS-STING pathway via the NO-mtDNA-Mn²⁺ synergistic effect, promoting dendritic cell (DC) maturation and enhancing CD8⁺ T cell infiltration in tumor tissues, and promoting tumor vascular normalization (Scheme 1). This endogenous-driven Gas-Metal Synergy Strategy ingeniously combines the biological safety advantages of gas therapy with the immune-enhancing effects of metal ions, demonstrating potent anti-tumor effects and offering an innovative, low-toxicity therapeutic approach for solid tumor immunotherapy.

Scheme 1.

Schematic illustration of MPRH nanoplatform formation and MPRH for cGAS-STING pathway activation and anti-tumor immunity. (A) The constructed Mn-ZIF8 nanoparticles, loaded with RRX-001 and surface-modified with PEG-HA, form MRPH nanoparticles that can degrade in acidic environments to exert anti-tumor effects. (B) MRPH releases Mn²⁺ and RRX-001 in the tumor acidic microenvironment, with the latter releasing NO to amplify the cGAS-STING pathway activated by Mn²⁺, forming a NO-mtDNA-Mn²⁺ ternary cascade. This cascade triggers ICD, reduces CD47 expression, and enhances macrophage phagocytic function. Additionally, it promotes DC maturation, increases CD8⁺ T cell infiltration into tumor tissues, and contributes to tumor vascular normalization, thereby achieving the effect of gas-amplified metalloimmunotherapy

Results and discussion

Relationship between Mn and the immune status of OS patients

Increasing evidence indicates that metal ions play critical roles in the human immune response [33]. Of these ions, manganese (Mn) serves as a cofactor for various enzymes [34]. Historically, research has focused on how the body limits Mn uptake by pathogens to prevent infections [35]. Recent studies have revealed that Mn also functions as an alarmin, activating innate immunity by stimulating the cGAS-STING pathway, thereby contributing to metal-mediated cancer immunotherapy [7]. Therefore, Mn was incorporated into our Gas-Metal Synergy Strategy [36]. To explore the relationship between Mn and the prognosis of OS, we constructed a prognostic model using 108 manganese-related genes (MRGs) from the GESA dataset. Subsequently, univariate Cox regression analysis, combined with LASSO regression, developed an MRGs prognostic model consisting of 15 genes (Figure S1, Supporting Information). The MRGs score (MRGscore) was calculated for each patient in the Target-OS dataset, stratifying patients into high- and low-risk groups based on the median value cutoff. Kaplan–Meier survival analysis showed that patients with low MRGs scores had significantly shorter overall survival compared to those with high MRG scores (p < 0.05) (Fig. 1A). This was validated in the GSE21257 dataset (Figure S2, Supporting Information), supporting the link between high MRGs scores and favorable OS prognosis. The MRGscore may serve as a risk assessment tool for OS prognosis. To predict 5-year survival rates, a nomogram integrating MRGscore with clinical variables (age, gender, metastatic status) was constructed (Fig. 1B). Notably, the presence of metastasis and a low MRGscore significantly increased mortality risk. In addition, immune infiltration analysis in the Target-OS dataset revealed that high-MRGscore patients had higher Estimate scores, Immune Scores, and Stromal Scores, with reduced Purity scores (Figs. 1C-F). A similar outcome was obtained in the GSE21257 dataset (Figure S3, Supporting Information). Further analysis of 24 immune cell types in the tumor microenvironment revealed that DCs, T cells, macrophages, and NK cells were more abundant in the high-MRGscore group compared to the low-MRGscore group (Fig. 1G). Moreover, a positive correlation was observed between macrophage abundance and the MRGscore (p < 0.05) (Fig. 1H; Figure S4, Supporting Information). These findings suggest that the MRGscore is linked to tumor microenvironment remodeling and immune regulation. Additionally, the MRGs profile may influence immune cells, particularly macrophages. In this context, a high MRGscore correlates with a more immune-responsive tumor environment, making it more susceptible to immune therapies, particularly those targeting macrophages and immune evasion. However, it is important to note that the administration of Mn²⁺ or manganese chloride solutions alone often leads to central nervous system toxicity and hepatic and renal dysfunction [32], and the development of Mn-based nanomaterials with improved biosafety and controlled, targeted release could help mitigate these risks.

Fig. 1.

Relationship between Mn and the Immune Status and Prognosis of OS Patients. (A) Kaplan-Meier survival analysis shows a significant difference in survival between the high MRGscore and low MRGscore groups in the Target-OS set. (B) Nomogram model for predicting the survival probability at 1, 3, and 5 years. The ESTIMATE score (C), Immune score (D), tumor purity (E), and Stromal score (F) in different MRGscore groups were evaluated using ESTIMATE analysis. (G) The correlation between MRGscore and immune cell infiltration in the tumor microenvironment (TME). (H) Immune infiltration levels of various immune cell types in the high and low MRGscore groups. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001

Synthesis and characterization of MRPH

The Mn-ZIF8 nanoparticles were synthesized by dissolving zinc nitrate and 2-methylimidazole in methanol, followed by centrifugation and drying of the resulting solution. Manganese acetate was then added to the ZIF-8 solution, followed by washing and drying of the resultant product. The Mn-ZIF-8@RRX-001@PEG-HA (named MRPH) nanoparticles were obtained by loading the drug molecule RRX-001 into Mn-ZIF8, followed by surface modification with PEG-HA. The morphology of the materials was characterized using transmission electron microscopy (TEM). TEM images revealed nanoparticles with a rough, aggregated morphology and the structure appears moderately crystalline with a porous surface. (Fig. 2A). Powder X-ray diffraction (PXRD) analysis showed that both pristine ZIF-8 and Mn-ZIF-8 maintained crystallinity, with diffraction patterns consistent with established ZIF-8 crystal structures. Notably, PEG-HA surface modification caused minimal changes to the crystalline framework (Fig. 2B). Dynamic light scattering (DLS) measurements showed hydrodynamic diameters of 172 nm, 176 nm, and 210 nm for Mn-ZIF-8, Mn-ZIF-8@RRX-001, and MRPH with polydispersity indexes (PDI) at 0.10, 0.11, and 0.13, respectively (Fig. 2C and D). All particles exhibited narrow size distributions and excellent aqueous dispersibility. Zeta potential measurements indicated values of 19.1 ± 0.8 mV, 24.2 ± 1.0 mV, and 9.8 ± 0.7 mV for Mn-ZIF-8, Mn-ZIF-8@RRX-001, and MRPH, respectively (Fig. 2E). To evaluate the drug loading and release profile of MRPH, UV-vis spectroscopy confirmed concentration-dependent absorbance increases at 240 nm (Figure S5, Supporting Information). In vitro release studies at varying pH levels demonstrated pH-responsive behavior. At physiological pH (7.4), only 15.7% of the drug was released within 24 h, indicating its significant stability under normal conditions. In contrast, under acidic conditions, the framework underwent accelerated disintegration, leading to cumulative drug releases of 70.3% at pH 6.5 and 76.2% at pH 5.5. (Fig. 2F). Inductively coupled plasma optical emission spectroscopy (ICP-OES) analysis of Mn²⁺ release kinetics mirrored this pH dependence, with minimal ion leakage (< 5%) at neutral pH and 80.4 ± 3.7% Mn²⁺ release after 24 h at pH 5.5 (Fig. 2G). The endogenous NO generation capability was assessed using DAF-FM DA fluorescent probes. DAF-FM DA reacts with NO to generate fluorescein-benzotriazole, which emits strong green fluorescence, making it a probe for the quantitative detection of NO [37]. Flow cytometry (FCM) analysis showed that MRPH-treated tumor cells produced 4.2-fold higher NO levels than the MPH group (Figs. 2H, I). Additionally, cells were pretreated with Carboxy-PTIO (CP), a commonly used NO scavenger, prior to MRPH treatment [38] The results showed a reduction in NO generation. These findings collectively demonstrate that MRPH maintains excellent structural integrity during circulation while achieving targeted payload release in acidic tumor tissues, minimizing off-target effects and potential systemic toxicity.

Fig. 2.

Synthesis and Characterization of MRPH. (A) Representative TEM images of MRPH nanoparticles, Scale bar: 100 nm; (B) XRD patterns of ZIF8, Mn-ZIF8 and MnZIF8-PEG-HA; (C) Size distribution of DLS measured by Mn-ZIF8, Mn-ZIF8@RRX-001 and MRPH; (D) PDI of Mn-ZIF8, Mn-ZIF8@RRX-001 and MRPH by dynamic light scattering; (E) Zeta-potential of Mn-ZIF8, Mn-ZIF8@RRX-001 and MRPH; (F) pH-dependent release of MRPH at different pH values; (G) Mn²⁺ release at various time and under different pH values; (H) FCM analysis of NO release using DAF-FM DA in different treatment groups and quantification of mean fluorescence intensity (MFI). Data are shown as mean ± SD. ***p < 0.001

Cellular uptake and cytotoxicity evaluation

Efficient cellular internalization is essential for the biological activity of nanomaterials [39]. To investigate MRPH uptake by OS cells, MRPH was labeled with Cy5.5 (named MRPH@Cy5.5) and incubated with K7M2 cells. Confocal laser scanning microscopy (CLSM) was used to visualize cellular uptake. The results revealed that red fluorescence intensity within the cells gradually increased with prolonged incubation time (Fig. 3A). Quantitative analysis by FCM confirmed that the mean fluorescence intensity at 7 h was 28.9 times higher than that at 0 h (Fig. 3B, C). Subsequently, the cytotoxic effect of MRPH on tumor cells was evaluated using the CCK-8 assay. The results indicated that the half-maximal inhibitory concentration (IC₅₀) of RRX-001 was 9.59 µM (Fig. 3D), and cell viability progressively decreased with increasing RRX-001 concentrations (Figure S6, Supporting Information). Notably, at RRX-001 concentrations of 2 µM and 4 µM, the cell viabilities in the RRX-001-only group were 89.4 ± 1.3% and 72.2 ± 1.3%, respectively, whereas in the MRPH-treated group, the cell viabilities were 59.5 ± 3.6% and 34.6 ± 1.0%, respectively, indicating that MRPH exerts a stronger inhibitory effect on cell viability (Fig. 3E). Furthermore, tumor cells were treated with PBS, RRX-001 (R), Mn-ZIF8, MPH, and MRPH for 24 h. After treatment, cells were double-stained with Annexin V-FITC and propidium iodide (PI), and apoptosis was assessed by FCM. The analysis revealed that the total apoptosis rate in the MRPH group reached 40.5 ± 5.1%, 3.5 times higher than in the control group and 1.6 times higher than in the MPH group (Figs. 3F, G). Moreover, live/dead cell staining showed that cells in the PBS and R groups were predominantly stained green by Calcein AM (indicating live cells), whereas the MRPH-treated group exhibited a significantly higher proportion of red fluorescence (indicating cell death) compared to the MPH group (Fig. 3H). Collectively, these findings demonstrate that the MRPH formulation exhibits potent in vitro antitumor activity, enhancing tumor cell killing while minimizing the required dosage of RRX-001.

Fig. 3.

Cellular Uptake and Cytotoxicity Evaluation of MRPH. (A) CLSM detection of the endocytosis of MRPH@Cy5.5 (red) by K7M2 cells at different time points. Scale bar: 10 μm; (B, C) FCM results and quantification of MFI following co-incubation of MRPH with K7M2 cells for 0 h, 1 h, 4 h, and 7 h; (D) CCK-8 assay to determine the half-maximal inhibitory concentration (IC₅₀) of RRX-001 as a single agent; (E) CCK-8 assay to assess cell viability in different treatment groups; (F, G) FCM analysis of apoptosis rate in different treatment groups using Annexin-V/PI double staining and quantitative analysis; (H) Live/dead staining analysis of K7M2 cells after different treatments, where red fluorescence represents dead cells and green fluorescence represents live cells. Scale bar: 50 μm. Data are shown as mean ± SD. **p < 0.01, ***p < 0.001

Amplification of cGAS-STING pathway activation by MRPH via NO-mtDNA-Mn²⁺ cascade

NO is a critical biomedical gas that not only enhances the efficacy of conventional therapies, such as chemotherapy and radiotherapy, but also mitigates their adverse side effects [15, 40]. Research has demonstrated that NO can decrease MMP activity, inhibit ATP synthesis, and induce mitochondrial damage, which in turn leads to the release of mtDNA into the cytosol [41, 42]. Once released, mtDNA in the form of dsDNA is recognized by cGAS, and Mn²⁺ has been shown to enhance the sensitivity of cGAS to cytosolic dsDNA. Building on these findings, we propose that NO and Mn released from MRPH, in combination with mtDNA, trigger a ternary cascade that effectively activates the cGAS-STING pathway. This cascade forms the basis for the dual benefits of immune activation and enhanced biosafety inherent in the Gas-Metal Synergy Strategy. To validate this hypothesis, we examined the phosphorylation status of TANK-binding kinase 1 (TBK1), interferon regulatory factor 3 (IRF3), and STING. These proteins are key mediators in the cGAS-STING pathway, and their cascade phosphorylation is known to promote the expression of type I interferons (IFNs) [43]. First, CLSM was utilized to assess the intracellular expression of phosphorylated STING (p-STING) following various treatments. The results revealed that cells treated with Mn-ZIF8, MPH, and MRPH exhibited a marked increase in red fluorescence intensity corresponding to p-STING relative to the control group, with the MRPH-treated cells showing the strongest signal (Fig. 4A). In parallel, western blot analysis confirmed that the expressions of phosphorylated TBK1 (p-TBK1), IRF3 (p-IRF3), and p-STING were upregulated in cells treated with Mn-ZIF8, MPH, and MRPH compared to controls. Notably, the MRPH group exhibited the most pronounced upregulation, with p-TBK1, p-IRF3, and p-STING increased by 1.6-, 1.4-, and 2.0-fold, respectively, compared to the MPH group (Figs. 4B-E). Additionally, cells were pretreated with CP to reduce NO generation, aiming to demonstrate the important role of NO in this process. The experimental results demonstrated that, relative to the MRPH group, NO release in K7M2 cells decreased by 49.1% in the MRPH + CP group (Figs. 2H, I). Furthermore, CLSM images indicated that the red fluorescence intensity in the MRPH group was attenuated upon CP treatment (Fig. 4A). Western blot analysis showed that the expressions of p-TBK1, p-IRF3, and p-STING in the MRPH + CP group were reduced by 32.0%, 40.8%, and 51.1%, respectively, compared to the MRPH group (Figs. 4B-E). These findings confirm that MRPH strongly activates the cGAS-STING pathway and that this enhancement is mediated by NO production.

Fig. 4.

Amplification of cGAS-STING Pathway Activation by MRPH via NO-mtDNA-Mn²⁺ Cascade. (A) CLSM analysis of pSTING fluorescence expression in different treatment groups, with red fluorescence representing pSTING, blue fluorescence representing DAPI nuclear staining. Scale bar: 10 μm; (B) Western blot analysis of pTBK1, pIRF3, and pSTING expression in different treatment groups; (C-E) Quantification of relative expression of pTBK1, pIRF3, and pSTING proteins; (F, G) JC-1 staining to assess MMP depolarization in K7M2 cells after different treatments and quantitative analysis; (H) Representative TEM images of mitochondria in K7M2 cells treated with PBS, MPH, or MRPH. Scale bar: 500 nm; (I) CLSM detection of K7M2 cells after different treatments, with red fluorescence representing dsDNA, green fluorescence representing TOMM20, and blue fluorescence representing DAPI. Scale bar: 10 μm; (J) qRT-PCR analysis of relative expression of mitochondrial-encoded DNA mt-ND1 and mt-DLoop in the cytoplasm of K7M2 cells after different treatments. (K) Schematic illustration of MRPH for Metal-Gas Synergy therapy. Data are shown as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001

To further elucidate the mechanism by which MRPH activates the cGAS-STING pathway, we examined changes in MMP using JC-1 staining. This activation occurs via a NO-mtDNA-Mn ternary cascade amplification effect. Under normal conditions, JC-1 accumulates in healthy mitochondria as red fluorescent aggregates, whereas in damaged mitochondria, JC-1 exists as green fluorescent monomers in the cytosol [44]. The shift from red to green fluorescence indicates a loss of MMP. Quantification by FCM revealed that the MRPH-treated group exhibited the highest percentage of green fluorescence-positive cells (25.3 ± 0.8%), which was 1.4-fold and 2.6-fold higher than that of the MPH group (18.4 ± 0.7%) and the PBS group (9.8 ± 2.6%), respectively. In contrast, the MRPH + CP group (18.8 ± 1.3%) showed a 26.6% reduction compared to the MRPH group (Figs. 4F, G). These results indicate that MRPH treatment induces significant mitochondrial damage and dysfunction, while NO scavenging in the MRPH + CP group mitigates this damage. TEM further confirmed these findings by revealing pronounced morphological and structural alterations in the mitochondria of MRPH-treated cells. Compared to the PBS group, mitochondria in the MRPH group displayed indistinct cristae, with some membranes appearing swollen and ruptured (Fig. 4H), confirming the substantial mitochondrial damage induced by MRPH. The NO-induced mitochondrial damage leads to the release of mtDNA into the cytosol, which is a key trigger for cGAS-STING pathway activation. CLSM analysis demonstrated that in the PBS-treated group, cytosolic dsDNA (red) was confined within continuous membranes surrounding the mitochondrial marker TOM20 (green). In contrast, the MRPH-treated group showed pronounced mitochondrial membrane permeabilization, a reduction in the colocalization of mtDNA with mitochondria, and a substantial increase in the red fluorescence of cytosolic dsDNA outside the mitochondria. The MRPH + CP group showed partial recovery of these features (Fig. 4I). Moreover, genomic DNA and purified cytosolic DNA were isolated from K7M2 cells after various treatments. qRT-PCR was then performed to quantify the expression of mtDNA-encoded ND1 and DLoop, using nuclear DNA-encoded Tert as a reference [45]. The results revealed that, relative to the PBS group, the cytosolic expressions of ND1 and DLoop in the MRPH-treated group were 28.9-fold and 14.6-fold higher, respectively, which were 1.7-fold and 2.1-fold higher than those observed in the MPH group. In the MRPH + CP group, the levels of ND1 and DLoop decreased by 34.3% and 38.2%, respectively, compared to the MRPH group (Fig. 4J). These findings indicate that MRPH treatment significantly increases cytosolic mtDNA levels, an effect that can be partially reversed by NO scavenging. In summary, the data demonstrate that MRPH activates the cGAS-STING pathway in tumor cells by releasing NO, which decreases MMP, induces mitochondrial damage, and triggers the release of mtDNA into the cytosol. This cascade significantly enhances and amplifies the activation of the cGAS-STING pathway by the Mn-ZIF8 platform (Fig. 4K).

Activation of tumor immune response by MRPH in vitro

Previous studies have demonstrated that activation of the cGAS-STING pathway leads to the release of type I interferons and pro-inflammatory cytokines, thereby stimulating tumor immunity [46]. Moreover, ICD induced by cGAS-STING activation plays a crucial role in triggering robust immune responses [47–51] (Fig. 5A). To verify that MRPH induces the release of type I interferon and pro-inflammatory cytokines through cGAS-STING pathway activation, we employed ELISA kits to measure the levels of interferon-β (IFN-β) and interleukin-6 (IL-6) in the cell culture supernatant. The results showed that in the MPH-treated group, the levels of IFN-β and IL-6 were 55.2 ± 4.6 pg/mL and 28.5 ± 5.0 pg/mL, respectively, which were 4.8-fold and 4.0-fold higher than those in the PBS group. In the MRPH-treated group, IFN-β and IL-6 levels reached 96.0 ± 15.4 pg/mL and 53.5 ± 7.2 pg/mL, representing 1.7-fold and 1.9-fold increases over the MPH group. In addition, pre-treatment with CP (MRPH + CP group) resulted in IFN-β and IL-6 levels of 59.2 ± 6.6 pg/mL and 42.6 ± 1.6 pg/mL, corresponding to reductions of 38.4% and 20.3% relative to the MRPH group (Fig. 5B and C). These results clearly indicate that MRPH promotes the release of type I interferon and pro-inflammatory cytokines via NO-mediated activation of the cGAS-STING pathway in tumor cells.

Fig. 5.

Activation of Tumor Immune Response by MRPH in Vitro. (A) Schematic illustration of MRPH for ICD. (B, C) ELISA detection of IL-6 and IFN-β levels in the supernatant of K7M2 cells and quantitative analysis; (D, E) FCM analysis of CRT exposure in different treatment groups and quantification of MFI; (F) CLSM detection of HMGB1 expression in different treatment groups, with red fluorescence representing HMGB1 and blue fluorescence representing the cell nucleus. Scale bar: 10 μm; (G, H) FCM analysis of BMDC maturation after co-culturing K7M2 cells with BMDCs in different treatments and quantitative analysis; (I, J) Western blot analysis of CD47 expression in K7M2 cells under different treatments and relative quantification. GAPDH was used as the internal reference protein; (K) FCM quantification of macrophage phagocytosis. K7M2 cells were stained with 5 µM DSFE, and Raw264.7 cells were stained with F4/80-bv711. Data are shown as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001

The hallmark events of ICD include calreticulin (CRT) exposure, high-mobility group box 1 (HMGB1) release, and ATP secretion [52]. CLSM analysis revealed intense red fluorescence in MRPH-treated cells, indicating CRT exposure. In contrast, NO clearance in the MRPH + CP group markedly reduced red fluorescence intensity (Figure S7, Supporting Information). FCM analysis confirmed that CRT exposure in the MRPH group was 3.5-fold and 1.8-fold higher than in the PBS and MPH groups, respectively, with a 41.4% reduction in the MRPH + CP group (Fig. 5D, E). These findings suggest that MRPH induces CRT exposure and triggers a strong ICD response via NO release. HMGB1, typically retained in the nucleus, is released during ICD to attract immune cells. CLSM images showed reduced nuclear HMGB1 fluorescence in the MRPH group compared to PBS and MPH groups, with partial restoration in the MRPH + CP group (Fig. 5F). In addition, ELISA analysis of the culture supernatant demonstrated that HMGB1 levels in the MRPH group (321.9 ± 13.9 pg/mL) were 3.4-fold and 1.5-fold higher than those in the PBS group (93.7 ± 4.1 pg/mL) and the MPH group (218.3 ± 12.6 pg/mL), respectively, while HMGB1 secretion was reduced in the MRPH + CP group (235.3 ± 22.1 pg/mL) relative to the MRPH group (Figure S8A, Supporting Information). ATP release, crucial for dendritic cell (DC) maturation, was 1.4-fold higher in the MRPH group than in the MPH group, while the MRPH + CP group showed a 27.7% decrease (Figure S8B, Supporting Information). DCs are key antigen-presenting cells whose maturation is vital for the activation of both innate and adaptive immunity. To simulate ICD-induced DC maturation in vitro, K7M2 cells treated with different formulations were co-cultured with bone marrow-derived dendritic cells (BMDCs) isolated from Balb/c mice. FCM analysis revealed that the proportion of mature BMDCs (CD80⁺CD86⁺) in the MRPH group reached 31.8 ± 2.5%, which was significantly higher than that in the MRH group (16.0 ± 6.5%) and 5.3-fold greater than in the PBS group. In contrast, the MRPH + CP group showed a 47.2% decrease in mature BMDCs (16.8 ± 0.2%) (Fig. 5G, H). These results demonstrate that MRPH, through NO release, enhances the type I interferon response and ICD cascade induced by the Mn-ZIF8 platform, leading to CRT exposure, increased tumor antigen uptake, and the release of type I interferons, pro-inflammatory cytokines, HMGB1, and ATP, thereby promoting DC maturation and tumor immunity.

Furthermore, the NO donor RRX-001 is a well-known small-molecule inhibitor of CD47. Our prognostic model indicated that the MRG score was significantly positively correlated with macrophage levels. Given that the CD47-mediated “don’t eat me” signal allows tumor cells to evade phagocytosis by macrophages [53], a key mechanism of tumor immune evasion, inhibiting CD47 not only enhances the phagocytic function of tumor-associated macrophages but also synergizes with cGAS-STING pathway activation to efficiently kill tumor cells. Accordingly, we investigated the suppression of CD47 expression by MRPH nanoparticles loaded with RRX-001. Western blot analysis demonstrated that, compared with the MPH group, CD47 expression in cells treated with MRPH was reduced by 90%, indicating that MRPH effectively suppresses CD47 expression (Fig. 5I, J). In parallel, CFDASE-labeled K7M2 cells were treated with various formulations and co-cultured with RAW264.7 macrophages. The results showed that the percentage of CFDASE-positive K7M2 cells within F4/80⁺ macrophages reached 21.3 ± 2.0% in the MRPH group, which was 2.5-fold higher than in the MPH group (8.6 ± 1.9%) (Fig. 5K). These findings indicate that MRPH treatment enhances the phagocytosis activity of macrophages toward K7M2 tumor cells.

Biodistribution and antitumor effect of MRPH in vivo

Biosafety is a prerequisite for the in vivo application of nanomedicines. To assess the biosafety of our formulations, KM mice were intravenously injected via the tail vein with PBS, R, Mn-ZIF8, MPH, and MRPH. 48 h after the final injection, blood and major organs were collected for safety evaluation. Compared with the PBS group, mice treated with Mn-ZIF8, MPH, and MRPH did not exhibit any significant abnormalities in biochemical parameters, indicating that at this dosage, Mn-ZIF8, MPH, and MRPH have no adverse impact on the biochemical indices of mice (Figures S9 and S10, Supporting Information). Moreover, hematoxylin and eosin (H&E) staining of major organs-including the heart, liver, spleen, lung, and kidney-revealed no apparent pathological abnormalities compared to the PBS group (Figures S11, Supporting Information). These results demonstrate that MRPH exhibits excellent biocompatibility, supporting its potential for in vivo applications.

An orthotopic OS Balb/c mouse model was established to investigate the in vivo biodistribution of MRPH (Fig. 6A). Following tail vein injection of Cy7.5-labeled MRPH, the fluorescence intensity at the tumor site was dynamically monitored using an in vivo imaging system (IVIS Spectrum). Notably, a distinct fluorescence signal was detectable at the tumor site 3 h post-injection, and the intensity gradually increased over time, reaching a peak at 24 h. Although the fluorescence subsequently declined, a significant signal was still observed at 36 h, demonstrating that MRPH exhibits good stability and achieves targeted accumulation and prolonged retention at the tumor site (Figs. 6B and Figure S12, Supporting Information). Ex vivo imaging of major organs and tumors at 36 h further revealed strong fluorescence in the tumor, liver, and kidneys, indicating that these tissues are the primary sites of MRPH accumulation (Fig. 6C, D). In summary, MRPH rapidly targets tumor tissue, interacts with the tumor microenvironment, and accumulates with sustained retention.

Fig. 6.

Biodistribution and Antitumor Effect of MRPH in vivo. (A) Schematic illustration of the in vivo therapeutic study. Fluorescence bio-imaging was performed to visualize tumor progression in K7M2-bearing mice following intravenous injection of Cy7.5-labeled MRPH at various time points (B), and in major organs (C). The abbreviations T, H, Li, S, L, K, and I correspond to tumor, heart, liver, spleen, lung, kidney, and intestine, respectively. (D) The relative MFI of Cy7.5-labeled MRPH in the tumor and major organs at 36 h post-injection. (E) Luminescence images of Balb/c mice bearing orthotopic K7M2-luc OS tumors after different treatments on days 0, 5, 10, and 15 post-tumor implantation. (F) Quantified luminescence levels of mice using the Lumina IVIS system. (G) Tumor inhibition rate and (H) body weight changes of mice treated with different therapies. n = 5 mice per group. (I) H&E staining of tumor tissues in different groups (upper panel) and immunofluorescence images of CD31 (red) and α-SMA (green), showing endothelial cells and pericytes, respectively (lower panel), Scalar bar: 20 μm. Data are presented as mean ± SD. *p < 0.05, ***p < 0.001

Next, we established an orthotopic K7M2-Luc (luciferase-labeled) OS Balb/c mouse model. After intraperitoneal injection of D-luciferin potassium salt, IVIS was performed on days 0, 5, 10, and 15 to monitor tumor progression (Fig. 6E). Unsurprisingly, rapid tumor growth was observed in both the PBS and R groups. Semi-quantitative bioluminescence analysis of fluorescence intensity changes at different time points showed that the relative fluorescence intensity change in the Mn-ZIF8 group was 2.4-fold higher than that in the MPH group by day 15, despite similar tumor suppression phenotypes observed in vitro for both Mn-ZIF8 and MPH groups. (Fig. 6F). Furthermore, tumor inhibition rates were calculated for each treatment group. The results indicated that the tumor inhibition rate in the MPH group was 43.9%, significantly higher than the 12.2% observed in the Mn-ZIF8 group. Additionally, the MRPH group resulted in effective tumor clearance, with a tumor inhibition rate of 70.4%, consistent with our previous findings (Fig. 6G). Although the average body weight of mice in different treatment groups varied to some extent from baseline, no significant differences were observed among the groups, further corroborating the excellent biosafety of MRPH (Fig. 6H). H&E staining of excised tumor tissues revealed that tumors from the MRPH group underwent more severe necrosis and apoptosis (Fig. 6I, top panel), while Ki67 immunohistochemical staining demonstrated that the MRPH group had the lowest tumor cell proliferation rate (Figures S13, Supporting Information). These findings collectively suggest that MRPH exerts superior antitumor effects in vivo.

In addition, the normalization of tumor vasculature, which helps alleviate the hypoxic conditions within tumors, may contribute to the antitumor efficacy of MRPH [54]. To investigate this possibility, immunofluorescence staining was performed on excised tumor sections, using CD31 to label blood vessels and α-smooth muscle actin (α-SMA) as a marker for pericytes, to evaluate pericyte coverage (Fig. 6I, bottom panel). The results revealed that, compared to the MRH group, the MRPH-treated tumors exhibited a higher degree of vascular endothelial coverage by pericytes, suggesting that MRPH can induce vascular normalization. As oxygen deficiency is a sensitive marker of vascular function, we used hypoxia-inducible factor-1α (HIF-1α) as an indicator of intratumoral hypoxia [55]. We observed a significant reduction in HIF-1α-positive areas in the MRPH group relative to the MRH group (Figure S14A, Supporting Information). These findings collectively demonstrate that RRX-001-loaded MRPH not only enhances antitumor efficacy by directly targeting tumor cells but also improves the tumor oxygen microenvironment through vascular normalization.

Mechanistic studies on MRPH-Mediated antitumor immunity in vivo

MRPH has been found in vitro to activate the cGAS-STING pathway through NO release and inhibit CD47 expression to enhance tumor immunity. To validate this process in vivo, we first evaluated NO activity by monitoring the formation of 3-nitrotyrosine (3-NT), a nitrated tyrosine residue that serves as a marker of the therapeutic effect of NO in tumor tissues [56]. The results showed that the MRPH group exhibited the highest expression of 3-NT, which correlated with the tumor suppression trend. NO immunofluorescence staining confirmed these results, highlighting the important role of NO in tumor therapy (Fig. 7A, top panel and Figure S14B, Supporting Information). We then performed immunofluorescence staining for CD47, p-STING, and p-TBK1 in tumor sections, where p-STING and p-TBK1 are key proteins in the cGAS-STING pathway. The results showed that tumor sections from MRPH-treated mice showed the strongest red fluorescence for p-STING and p-TBK1, while CD47 exhibited the lowest red fluorescence, suggesting that MRPH increased the expression of cGAS-STING-related proteins and inhibited CD47 expression (Fig. 7A, bottom panel, Figures S14C and S15A, Supporting Information).

Fig. 7.

Mechanistic Studies on MRPH-Mediated Antitumor Immunity in vivo. (A) 3-NT staining and p-STING (red) immunofluorescence images of tumor tissues from different treatment groups. Scale bar: 20 μm. Serum levels of IL-6 (B) and IFN-β (C) measured by ELISA. (D) The percentage of mature DCs within the tumor. (E) The percentage of mature DCs within TLDN after different treatments. (F) Representative FCM analysis images of CD8⁺ and CD4⁺ T cells gated on CD45⁺ cells within tumor tissues after different treatments. (G) Representative FCM analysis images of M1 macrophages (F4/80⁺ CD80⁺) gated on CD45⁺CD11b⁺ cells within tumor tissues after different treatments. The percentage of CD8⁺ T cells within the tumor (H) and spleen (I) after different treatments. (J) Quantification of the M1/M2 macrophage ratio in tumor tissues. (K) The percentage of MDSCs within the tumor. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001

To further explore how MRPH-induced cGAS-STING activation alters the TME, we focused on its downstream effects. Activation of the cGAS-STING pathway is known to promote DCs maturation through the production of type I interferons, pro-inflammatory cytokines, and DAMPs, which in turn facilitate antigen presentation to T lymphocytes. Using ELISA, we measured the levels of IFN-β and IL-6 in the peripheral blood of mice from different treatment groups. The MRPH group exhibited an IFN-β level of 114.9 ± 2.7 pg/mL, which was 1.6-fold and 3.3-fold higher than the MRH and PBS groups, respectively (Fig. 7B). Similarly, IL-6 levels in the MRPH group were 32.7 ± 3.4 pg/mL, 1.5-fold and 4.0-fold higher than those in the MRH and PBS groups, respectively (Fig. 7C). In parallel, immunofluorescence staining of tumor sections for CRT, a DAMP that is exposed on the surface of dying tumor cells and triggers ICD, showed the strongest red fluorescence in the MRPH group (Figures S15B, Supporting Information). These data indicate that MRPH treatment activates the cGAS-STING pathway, leading to the release of factors that promote DCs maturation and initiate an adaptive immune response. FCM analysis further confirmed this effect in both tumor tissues and tumor-draining lymph nodes (TDLNs). DCs maturation, identified by the markers CD11c⁺CD80⁺CD86⁺, was significantly enhanced in the MRPH group. In tumors, the percentage of mature DCs in the MRPH group was 32.3 ± 2.2%, which was 1.8-fold and 4.3-fold higher than in the MPH and PBS groups, respectively (Figs. 7D and Figure S20, Supporting Information). Similarly, in TDLNs, the proportion of mature DCs in the MRPH group was 21.8 ± 2.4%, representing 2.0-fold and 6.1-fold increases over the MPH and PBS groups, respectively (Fig. 7E).

Furthermore, FCM analysis of CD8⁺ T lymphocyte infiltration in tumor tissues and spleens revealed that the MRPH group exhibited a marked increase in cytotoxic T cell presence. In tumor tissues, CD8⁺ T cells accounted for 37.8 ± 4.0% of the cell population, compared to 26.4 ± 0.6% in the MPH group and 10.1 ± 1.7% in the PBS group (Figs. 7F, H, and Figure S16, Supporting Information). In the spleen, the MRPH group exhibited a CD8⁺ T cell proportion of 42.4 ± 2.6%, significantly higher than the MPH group (19.0 ± 8.5%) and PBS group (1.2 ± 0.2%) (Figs. 7I, Figure S17, and Figure S19A, Supporting Information). CD8 immunofluorescence staining of tumor sections confirmed these findings, with the MRPH group displaying the strongest red fluorescence signal (Figures S15C, Supporting Information). These results suggest that MRPH significantly promotes the infiltration of cytotoxic T cells into both tumor and lymphoid tissues, enhancing the adaptive immune response. To overcome tumor immune resistance, solid tumors typically create an immunosuppressive microenvironment characterized by the infiltration of tumor-associated macrophages (TAMs) and the accumulation of myeloid-derived suppressor cells (MDSCs) [57]. We first evaluated whether MRPH could reprogram TAMs from a pro-tumor M2 phenotype to a tumoricidal M1 phenotype. FCM data showed that, in the MRPH-treated group, the proportion of M1 macrophages (CD11b⁺F4/80⁺CD80⁺) was 20.2 ± 3.0%, significantly higher than in other treatment groups (Figs. 7G, Figures S18 and S19B, Supporting Information). The proportion of M2 macrophages (CD11b⁺F4/80⁺CD206⁺) was 6.9 ± 3.7% (Figure S19C, Supporting Information). Consequently, the M1/M2 ratio in the MRPH group was 3.7 ± 1.4, which was 4.8-fold and 10.1-fold higher than in the MPH and PBS groups, respectively (Fig. 7J). Moreover, the percentage of MDSCs (CD11b⁺Gr-1⁺) in tumor tissues was significantly reduced to 5.8 ± 1.4% in the MRPH group, compared with 11.7 ± 0.8% in the MPH group and 18.0 ± 0.9% in the PBS group (Figs. 7K and Figure S21, Supporting Information).

Conclusion

In this study, we introduce a novel Gas-Metal Synergy Strategy utilizing a pH-responsive Mn-ZIF8 nanoplatform to enhance tumor immunotherapy through activation of the cGAS-STING pathway. This approach leverages the combined effects of metal ions and gas therapy to drive a robust immune response while promoting vascular normalization. The MRPH nanoplatform selectively releases Mn²⁺ and NO in the acidic TME, triggering a ternary cascade that amplifies cGAS-STING pathway activation. This cascade induces mitochondrial damage, mtDNA release, and immune activation, promoting ICD, DCs maturation, CD8⁺ T cell infiltration, and vascular normalization. By combining Mn²⁺, which enhances cGAS sensitivity to DNA, with the bioactive effects of NO, the platform not only strengthens immune responses but also minimizes systemic toxicity through targeted release. Both in vitro and in vivo studies confirm that MRPH effectively enhances tumor-specific immune activation, improves the TME, and promotes vascular normalization, presenting a promising low-toxicity alternative for solid tumor treatment. This work highlights the potential of integrating gas and metal-based therapies for immunotherapy. The Gas-Metal Synergy Strategy offers a cutting-edge platform for overcoming the immunosuppressive TME and promoting vascular normalization, with MRPH demonstrating superior biosafety and minimal toxicity, making it a strong candidate for clinical application. Leveraging its tumor selectivity, pH responsiveness, and PEG-HA targeting, MRPH is mechanistically designed to reduce off-target exposure. To advance clinical translation, we will implement a long-term safety assessment program centered on longitudinal biodistribution/total Mn²⁺ burden, NO-related monitoring, and focused hepatic and neurologic evaluations to define the safety margin and guide dose selection. Collectively, our findings position MRPH as an innovative therapeutic platform with broad implications for cancer treatment.

Experimental section

For all the experimental sections, please see the supporting information. All animal experimental protocols were approved by the Animal Ethics Committee of The Second Xiangya Hospital, Central South University (20231087).

Statistical analysis

All statistical analysis and statistical graph generation were performed using GraphPad Prism 9. Unless otherwise stated in the figure legend, data are presented as mean ± SD of at least 3 independent experiments of biological replicates. Statistical significance was analyzed using two-tailed Student’s t-test, one-way ANOVA or two-way ANOVA test. The difference was regarded as significant when the p value was less than or equal to 0.05. *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant.

Author contributions

XZ and TL designed the study and conducted the analysis. XZ and ZW drafted the manuscript. ZJ and SL were responsible for the synthesis and preparation of materials. ZW, GJ, YZ, YJ, LW, and PL were responsible for experimental operations and animal experiments. ZW, SL, XZ and TL contributed to the revision of the manuscript. All authors have read and approved the final manuscript.

Funding

The study was supported by the China Postdoctoral Fellowship Program of CPSF (GZB20240873), the China Postdoctoral Science Foundation (2024M763711), the Natural Science Foundation of Hunan Province, China (2023JJ60062), the Hunan Province Traditional Chinese Medicine Research Project (C2024028), the Hunan Provincial Innovation Foundation for Postgraduates (CX20230116), the Fundamental Research Funds for the Central Universities of Central South University (2023ZZTS0026), and the Central South University Research Program of Advanced Interdisciplinary Studies (2023QYJC026).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

All animal experimental protocols were approved by the Animal Ethics Committee of The Second Xiangya Hospital, Central South University (NO. 20231087). This article does not involve human participants; therefore, consent to participate was not required.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.