Multifunctional nanocomposites utilizing ruthenium (II) complex/manganese (IV) dioxide nanoparticle for synergistic reinforcing radioimmunotherapy Authors: Jian Peng, Dong-Ling Quan, Guang Yang, Lin-Tao Wei, Zhuan Yang, Zhi-Ying Dong, Yi-Ming Zou, Ying-Ke Hou, Jin-Xiang Chen, Lin Lv, Bin Sun Published: 2024 ABSTRACT Radiotherapy (RT) stands as a frontline treatment modality in clinical breast oncology, yet challenges like ROS reduction, high toxicity, non-selectivity, and hypoxia hinder efficacy. Additionally, RT administered at different doses can induce varying degrees of radioimmunotherapy. High doses of radiation (>10 Gy) may result in immune suppression, while moderate doses (4–10 Gy), although capable of mitigating the immune suppression caused by high-dose radiation, are often insufficient in effectively killing tumor cells. Therefore, enhancing the generation of ROS and ameliorating the tumor hypoxic immune-suppressive microenvironment at moderate radiation doses could potentially drive radiation-induced immune responses, offering a fundamental solution to the limitations of RT. In this study, a novel multifunctional nanoplatform, RMLF, integrating a Ru (II) complex into folate-functionalized liposomes with BSA-MnO2 nanoparticles was proposed. Orthogonal experimental optimization enhances radiosensitization via increasing accumulation in cancer cells, elevating ROS, and contributing to a dual enhancement of the cGAS-STING-dependent type I IFN signaling pathway, aimed to overcome the insufficient DAMPs typically seen in the conventional RT at 4 Gy. Such a strategy effectively activated cytotoxic T lymphocytes for infiltration into tumor tissues and promoted the polarization of tumor-associated macrophages from the M2 to M1 phenotype, substantially bolstering immune memory responses. This pioneering approach represents the first use of a ruthenium complex in radioimmunotherapy, activating the cGAS-STING pathway to amplify immune responses, overcome RT resistance, and extend immunotherapeutic potential.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12951-024-03013-2. Introduction The latest data from the global cancer statistics for 2023 reveal that breast cancer has surpassed lung cancer as the most diagnosed cancer [1]. Radiotherapy (RT), one of the first-line treatments for breast cancer, directly damages double-stranded DNA (dsDNA) using external high-energy electrons, X-rays, or charged particle beams, and indirectly generates reactive oxygen species (ROS) through water decomposition. Furthermore, RT alone or in combination with other types of therapies such as surgery, chemotherapy, or immunotherapy are applied to an estimated 60% of all cancer patients at some point during their treatment [2, 3]. Despite the outstanding performance on RT, lower ROS production, higher radiation dose toxicity and radiation resistant hypoxia severely limit its application. With the increasing demand for RT treatment, massive efforts are striving to develop various radiosensitizers for the expansion of radiation oncology arsenal. Currently, two major tactics involve reoxygenating the tumors and improving ROS production, are areas of active research. Insufficient O2 content in TME induces the failure of dsDNA damage via ionizing radiation, consequently contributes to radio-resistance. Considering the overexpressed endogenic H2O2 in TME, designing and synthesizing catalase-like inorganic nanomaterials such as platinum nanoparticles (Pt NPs), manganese dioxide (MnO2) and cerium oxide (CeO2) to elevate the O2 level in situ is one approach to combat hypoxia [4–6]. On the other hand, maximize the cytotoxic ROS generation to oxidative damage cellular components (e.g.dsDNA, protein, lipid, and subcellular organelle) would promote the RT efficacy. Although higher radiation dose provides enough support for ROS enhancement, the associated inevitable side effects to healthy skin, tissues, and organs extremely limited RT advancement. Transition metal complexes have appeared to be an attractive series of cancer radiosensitizers, among them, cis-dichlorodiammineplatinum (CDDP) targeting dsDNA is recognized as a potent radiosensitizer. Besides platinum-containing complexes, ruthenium polypyridyl complexes (RPCs) have been heavily studied for their dsDNA-binding, potential anticancer activity, and adjustable spectroscope and photophysical properties [7–10]. Moreover, RPCs can sensitize RT by means of activating the p53 pathway, regulating cell cycle arrest, and promoting intracellular ROS levels [11]. However, most RPCs have limited their development in the field of RT sensitization due to their poor water solubility and lack of tumor targeting. To address this, Chen’s group has designed a PEGylated ruthenium polypyridyl complex capable of self-assembling into nanoparticles in aqueous solution by π-π stacking interaction, hydrophobic effect, hydrogen bond and van der Waals force, which promote its accumulation in tumor sites by enhanced permeability and retention (EPR) [12]. There is growing evidence that radioimmunotherapy has unprecedented potential in oncology treatment. As early as the 1950s, Mole and colleagues discovered that RT could induce regression of distal tumors not directly exposed to radiation, a phenomenon termed the “abscopal effect“ [13]. This effect suggests that RT possesses adjuvant immunomodulatory properties, eliciting systemic antitumor immune responses and inhibiting tumor metastasis. More recently, a study proposed a structurally controlled nanoheterojunction for direct destruction of tumor cells and shaping the tumor immune microenvironment to effectively promote RT, producing a dramatic “abscopal effect” in combination therapy mode to prevent tumor recurrence and metastasis [14]. Given the unique properties of RT, there is a compelling rationale for developing advanced radioimmunotherapy approaches that expand the abscopal response and elucidate the underlying mechanisms of action to enhance therapeutic gain. Previous studies have demonstrated that the abscopal response caused by RT mostly depends on the drive of DNA sensor cyclic GMP-AMP synthase (cGAS) and its downstream effector, stimulator of interferon genes (STING), which are triggered by the leakage of endogenous damaged dsDNA into the cytosol. Activation of the cGAS-STING pathway is important for priming antitumor immune responses, because it underlies the secretions of nuclear factor kappa-B (NF-κB), interferon regulatory factor 3 (IRF3), type I interferons (IFNs) and other immune response cytokines that contribute to the activation of dendritic cells (DCs), tumor-specific CD8+ T cells and natural killer (NK) cells. However, with histones protection and host DNase digestion, nuclear dsDNA seems to be a weak activator of cGAS, thereby hindering antitumor immunity [15, 16]. Mitochondria, as three prominent sources of ROS, ATP, and cell death (apoptosis or necrosis) within most mammalian cells, has been used extensively to enhance RT [17]. Compared to normal cells, tumor cells typically exhibit a higher mitochondrial membrane potential (-180 mV), allowing positively charged metal complexes to target mitochondria by modulating the lipid solubility of organic ligands or introducing mitochondria-targeted classical functional groups, such as triphenylphosphine [18, 19]. Mitochondrial matrix containing mitochondrial DNA (mtDNA), unlike nuclear dsDNA, mtDNA is characterized by a circular loop and located outside of the nucleus without histone protection and lack of nucleotide excision repair, which make it more susceptible to radiation-induced DNA damage [20]. Furthermore, mtDNA is considered an immunostimulatory agents capable of triggering the antitumor immunity. Once the mitochondrial integrity was compromised, cytoplasmic oxidized mtDNA leakage as damage-associated molecular patterns (DAMPs) to constantly stimulate the cGAS-STING pathway to media innate immunity with IFNs secretion [21, 22]. Emerging researches demonstrated that manganese ions (Mn2+) as a catalytic cofactor binding to cGAS, also participate in activating innate immunity via the imperative activation of the cGAS-STING signaling pathway in a dsDNA-independent mode [23–25]. Given the immunomodulatory effects of RT, the field of radiation-immunology has garnered significant attention in clinical anti-tumor therapy. Recognizing the crucial involvement of the cGAS-STING signaling pathway in antitumor immunotherapy, it is imperative to explore various strategies that harness the cGAS-STING pathway to expand the scope of radiation-immunology. These strategies aim to develop efficient and low-toxicity radiosensitizers capable of promoting ROS production, alleviating tumor hypoxia, and augmenting the effectiveness of radiation immunotherapy. However, the low efficiency of tumor killing at low radiation doses remains a critical factor limiting radiation immunotherapy. Therefore, the development of radiosensitizers that can achieve both effective tumor killing and immune activation at low radiation doses represents a new challenge in the field of RT. In this work, we have apply [Ru(DIP)2dppz]2+ (where DIP = 4,7-diphenyl-1,10-phenanthroline, dppz = Dipyridophenazine) and bovine serum albumin (BSA)-MnO2 nanoparticles to construct a [Ru(DIP)2dppz]2+/BSA-MnO2 based liposome-folate nanocomposite (RMLF) with cancer-targeted, TME modulation, ROS upsurge, mtDNA oxidative damage, immune regulation to optimize the radiosensitization. The nanocomposite was designed based on the strategies as follows: (1) [Ru (DIP)2dppz]2+ (RudZ) with more hydrophobic ancillary ligands DIP and classical DNA binding group dppz, which make it more easily to be entrapped in the lipid bilayer of liposome (Lipo), so as to significantly improve the solubility and bioavailability of RudZ [26]. (2) Using BSA as synthesis template to accomplish nano-scale BSA-MnO2 (MnB) for better evolved in the inner aqueous compartment of Lipo. The MnB nanoparticles could alleviate the oxygenous environment and the generated Mn2+ attempt to assist the activation of cGAS-STING pathway. (3) With the help of folate (FA) functionalized Lipo, it is desirable to facilitate the therapeutic efficacy by enhancing the tumor accumulation and penetration of RudZ and MnB, promoting radiation-induced DNA damage, modulating hypoxia atmosphere, adjusting redox balance of mitochondria with ROS outbreaks, prompting oxidative mtDNA leakage and activating the cGAS-STING innate immune signaling pathway to improve immunological function. The detailed RMLF construction and therapeutic route are both presented in Scheme 1. This study represents the first example of radiotherapy-immunology based on ruthenium complexes with dual activation of the cGAS-STING pathway for tumor-specific comprehensive therapy according to our knowledge. This strategy aims to enhance the efficacy of RT in killing tumors associated with low-dose radiation. Additionally, it seeks to improve the immunosuppressive hypoxic TME, thereby driving radiation-induced immune responses. It holds the potential to address the limitations of RT at a fundamental level. Scheme 1Synthesis process of RMLF (a) and its application in radioimmunotherapy of breast cancer (b) Synthesis process of RMLF (a) and its application in radioimmunotherapy of breast cancer (b) Results and discussion Preparation and characterizations of RMLF The vesicle size of liposome, the ratio of phospholipid to cholesterol, ultrasonic time, preparation temperature, drug-lipid ratio and other factors have direct effects on the long-term stability and bioactivity of the nanocomposites [27]. We optimized the preparation conditions of RMLF by consulting the literature and orthogonal experimental design (Table S1, S2) [28]. Under the optimal preparation process, the RudZ loading of RMLF is limited because of the fixed drug-fat ratio. Therefore, the activation effect of mtDNA-induced cGAS-STING pathway might be insufficient. In order to further optimize the radiation-immunology efficacy, we need to find a substance that can assist RudZ to achieve the aforementioned objectives. Interestingly, MnO2 nanoparticles can catalyze the production of oxygen from tumor endogenous H2O2 and regulate immunosuppressive anoxic TME. At the same time, the released Mn2+ could further amplify the activation of cGAS-STING pathway and significantly consolidate the immune effect mediated by RT. Details on the synthetic preparations and characterization assays of RudZ, MnB, Lipo, RudZ-Lipo (RL), RudZ + MnB@Lipo (RML), RudZ + MnB@Lipo-FA (RMLF) are in-depth described in the Supporting Information. The 1H-NMR of RudZ confirmed the presence of the dppz and ancillary ligands DIP, as reported by others (Figure S1) [29]. It is well known that dppz-derived ruthenium complexes can act as molecular light switches [30], therefore, we examined the properties of RudZ using Lipo as the assembly medium. After being wrapped by the phospholipid bilayer of Lipo to construct RL, the fluorescence intensity of RudZ increased significantly due to a hydrophobic environment (Fig. 1a). When RL was incubated with methanol solution for 30 min, the disrupted phospholipid bilayer of the Lipo contributed to the reduced fluorescence intensity (Fig. 1b). Inverted fluorescence microscopy (IFM) observed distinct red fluorescent spots corresponding to RL, RML and RMLF nanocomposites, whereas no fluorescent signal was observed for the RudZ or Lipo alone (Fig. 1c and Figure S2). This further confirmed that the Lipo provide a hydrophobic environment for enabling the RudZ to fluoresce consistently.The lyophilized MnB and RMLF were characterized by X-ray diffraction (XRD) (Fig. 1d), the experimental XRD profile of the MnB and RMLF share similar distinct peaks, which were in good match with the characteristic peak of simulated MnO2 [31],implying the presence of MnO2 in MnB and RMLF. The surface elemental composition of MnB and RMLF were investigated by X-ray photoelectron spectroscopy (XPS), two featured peaks at 641.48 eV and 653.98 eV were attributed to Mn (IV) 2p3/2 and Mn (IV) 2p1/2, this is lead to a spin-energy separation of 12.5 eV which suggests the Mn4+ ions are dominant in the MnB (Figure S3) [32]. For RMLF, the Ru 3d (284.48 eV), Ru 3p (485.17 eV), Mn 2p (641.48 eV), N 1s (396.1 eV) and C 1s (284.58 eV) signals were detected by full scanning mode. In addition, the presence of MnB in RMLF could be further confirmed by the O 1s peak at 531.8 eV, data above all proved the successful construction of RMLF (Fig. 1e). Then we observed the morphology of MnB, RL and RMLF by transmission electron microscopy (TEM) (Figure S4, Fig. 1f). From the TEM images of RMLF, the Lipo with defined characteristics spherical surfaces and the MnB is encapsulated inside the liposomes at nanoscale, energy-dispersive X-ray (EDX) analysis verify that Ru, Mn, C, O and N are the elemental components of RMLF, coinciding with the XRD results (Table S3).The sizes, polydispersity coefficient (PDI) and surface charges of MnB, Lipo, RL, RML, RMLF were measured by dynamic light scattering (DLS) and zeta potiential test (Fig. 1g-i). The average hydrated particle sizes of fresh prepared MnB, Lipo, RL, RML and RMLF were 40.81 ± 0.9 nm, 123.4 ± 5.1 nm, 95.31 ± 3.3 nm, 113.4 ± 6.1 nm and 137.7 ± 14.7 nm, respectively, with acceptable PDI in the range of 0.24–0.28. The zeta potential of Lipo is negative and increases upon the number of RudZ present in RL due to their positive charges, while the zeta potential of RML and RMLF were shown to decreased mainly depends on the negative charge of MnB. With these data, it is reasonable to believe that the RudZ and MnB were successfully encapsulated in the Lipo structure. Fig. 1Characterizations of RMLF. (a) The fluorescence spectra of RL (10 µM RudZ) and RudZ (10 µM) in Tris·HCl buffer (Ex 450 nm/Em 500–800 nm). And the difference of fluorescence intensity between RudZ (I0) and RL (I) at 630 nm (inset). (b) Fluorescent quenching of RL (10 µM RudZ) by methanol at various times points. And the graph of fluorescence intensity over time at 630 nm. I0, fluorescence intensity of the solution at 0 min; I, fluorescence intensity of the solution at 0–30 min (inset). (c) IFM images of RudZ, Lipo, and RMLF (10 µM RudZ). The red fluorescence is derived from RudZ in the lipid bilayer. The scale bar represents 100 μm. (d) XRD spectrum of simulated MnO2, MnB and RMLF. (e) XPS spectrum of RMLF. (f) TEM images of RMLF (scale bar: 200 nm,50 nm). (g, h, i) Size distribution, zeta potential and PDI analysis of Lipo, MnB, RudZ, RL, RML and RMLF (n = 3) Characterizations of RMLF. (a) The fluorescence spectra of RL (10 µM RudZ) and RudZ (10 µM) in Tris·HCl buffer (Ex 450 nm/Em 500–800 nm). And the difference of fluorescence intensity between RudZ (I0) and RL (I) at 630 nm (inset). (b) Fluorescent quenching of RL (10 µM RudZ) by methanol at various times points. And the graph of fluorescence intensity over time at 630 nm. I0, fluorescence intensity of the solution at 0 min; I, fluorescence intensity of the solution at 0–30 min (inset). (c) IFM images of RudZ, Lipo, and RMLF (10 µM RudZ). The red fluorescence is derived from RudZ in the lipid bilayer. The scale bar represents 100 μm. (d) XRD spectrum of simulated MnO2, MnB and RMLF. (e) XPS spectrum of RMLF. (f) TEM images of RMLF (scale bar: 200 nm,50 nm). (g, h, i) Size distribution, zeta potential and PDI analysis of Lipo, MnB, RudZ, RL, RML and RMLF (n = 3) The UV − Vis spectrum of RMLF exhibited the characteristic absorption peaks of RudZ, MnB and FA at 279, 290, 363 and 458 nm, respectively, also confirmed above conclusion [33, 34] (Figure S5). To examine the the storage and serum stability of RMLF under physiological conditions, we dispersed it in different physiological media, such as phosphate buffered solution (PBS), RPMI1640, RPMI1640 + 10% fetal bovine serum (FBS) and examined the changes in hydrated size over a week (Figure S6). The results showed that RMLF exhibits good in vitro stability under the above physiological conditions. The method for evaluating the encapsulation efficiency (EE) and drug loading content (LC) of RudZ in RL, RML and RMLF were detailed in the Supporting Information, the obtained results showed that EE of all above three nanocomposites were being in the range of 70–75%, and the LC were also around 3% (Figure S7). Furthermore, ICP-MS was performed to quantify the concentration of ruthenium (Ru) and manganese (Mn) ions of RL, RML, RMLF and MnB samples digested in aqua regia. The concentration of Ru ranged from 15.5 to 16.1 mg/L for RL, RML and RMLF per sample, and the concentration of Mn ranged from 87.2 to 90 mg/L for the MnB, RML and RMLF per sample (Table S4). Enveloping drug or nanomaterials with Lipo for mimicking cell structure is a unique way to increase biocompatibility and delivery efficiency. To determine a suitable drug release medium, we first examined the saturation solubility of RudZ in different media (PBS and 2% SDS + PBS) at 37 °C, a remarkable increase in solubility has been made in 2% SDS + PBS, which was used as the release medium for in vitro RudZ release assessment from RMLF. 28% of RudZ was released from the RMLF during the initial 12 h in 2% SDS + PBS (pH 5.5), suggesting that polyethylene glycol bilayers protect RMLF from biomolecular interactions, thereby improve the stability (Figure S8). As evidenced by all the above results, to sum up, the RMLF has been successfully prepared. Measurement of O2 production by RMLF MnB is capable of reacting endogenous tumor H2O2 under acidic conditions to produce O2, which regulates the tumor hypoxia microenvironment and promotes RT sensitivity. Moreover, if nanoparticles can self-degrade after entering tumor tissues, their damage to normal tissues will be greatly reduced. MnB could degrade into water-soluble Mn2+ under acidic conditions and is excreted through the renal site, which greatly improves the biosafety of RMLF and avoids the toxic effects of metal particle accumulation on normal tissues [35]. The relevant process can be represented by the following reaction equation [36].\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:H_2O_2\hspace{0.17em}+\hspace{0.17em}2H^++MnO_2\to\:2H_2O\hspace{0.17em}+\hspace{0.17em}O_2\hspace{0.17em}+\hspace{0.17em}Mn^{2+}$$\end{document} To explore whether the prepared MnB has good biocompatibility, we simulated the acidic environment of tumor tissues in vitro and investigated the degradation of MnB under different pH conditions. As shown in the Figure S9, under pH 5.5, the MnB degraded rapidly with a significant decrease in the absorption peak at 290 nm, along with the color of the solution became lighter and almost colorless after 24 h. In contrast, under pH 7.4, the degradation of MnB was minimal, and the solution color remained relatively unchanged. Next, we intended to determine whether RMLF can generate oxygen in simulated TME with 1 mM H2O2 [37]. RL, RML and RMLF all had better oxygen production capacity, while the releases of oxygen were slightly slower compared to MnB, suggesting that the Lipo wrapping in RML and RMLF played a role in decelerating oxygen release (Fig. 2a). Ru (DPP)3Cl2 (RDPP) is an oxygen-sensitive probe whose red fluorescence can be strongly quenched by oxygen. To visually demonstrate MnB-generated oxygen and its potential to enhance RT efficacy, we detected intracellular oxygen production using RDPP as a fluorescent probe [38]. As expected, when MnB was co-incubated with 4T1 cells for 24 h, the fluorescence of RDPP decreased significantly due to the oxygen production from endogenous H2O2 in the cells catalyzed by MnO2. When 100 µM H2O2 was added in advance for a 30 min treatment, more oxygen was generated in the cells, and the red fluorescence of the cultured cells was almost completely quenched (Fig. 2d). The quantitative analysis of fluorescence intensity of MnB and MnB + H2O2 (100 µM) is 2.26 and 0.06, respectively, consistent with above results (Fig. 2b). Moreover, a hypoxia inducible factor-1α (HIF-1α) immunofluorescence labeling assay was performed, HIF-1α has traditionally been thought to be a vital marker of tumor hypoxia [39]. As the Fig. 2c, e showed that there were significant down-regulation of HIF-1α in 4T1 cells after MnB + X-ray or RMLF + X-ray treatment compared with other groups. This further confirmed that MnB and RMLF could catalyze the endogenous tumor H2O2 decomposition to produce oxygen, effectively alleviating tumor hypoxia and in turn enhancing the efficacy of RT. The intracellular overexpressed glutathione (GSH) can reduce the RT efficiency to some extent via scavenging the ROS [40], the redox reaction between MnO2 and GSH can be applied in the GSH depletion strategy, with the Mn2+ product possessing catalytic properties that alleviate ROS clearance by GSH [41]. Methylene blue (MB) is a commonly used dye that can serve as an indicator for ROS, as depicted in Figure S10, when both H2O2 (1 mM) and GSH (0, 0.5 mM) were present in MnB or RMLF-treated samples ([Mn] = 10 µg/mL), MB degradation occurred first, likely due to the release of catalytic Mn2+ ions that produce ·OH. During the excessive GSH concentration process (1,10 mM), the MB degradation was weakened due to the scavenging effect of GSH on ·OH. The level of GSH reduction is relative to the concentration of MnB. Cellular uptake and active targeting of RMLF It is well recognized that cellular internalization of RPCs could be great enhanced when wrapped by Lipo. The cellular uptake behaviors of RMLF against 4T1 cells was detected by observing red luminescent of RudZ with IFM, which revealed that RMLF accumulated predominantly in 4T1 cells after 6 h co-incubation (Figure S11a). Fluorescence intensity quantification further validated this experimental result (Figure S11b). On the other hand, to elucidate the cellular uptake mechanism between RMLF and RudZ, we co-treat RMLF or RudZ with low temperature (4 °C), an energy metabolic inhibitor (carbonyl cyanide m-chlorophenyl hydrazine, CCCP) and an endocytosis inhibitor (chloroquine) respectively. As presented in Figure S12, 4T1 cells pretreated with 4 °C or CCCP showed a significant decrease in the efficiency of RudZ uptake by 4T1 cells, while treatment with chloroquine had no noticeable impact on cell uptake efficiency. This suggests that RudZ enter the cells through energy-dependent active transport. However, lower temperature, energy metabolism inhibitor as well as endocytosis inhibitor all substantially diminished the uptake efficiency of RMLF by 4T1 cells, indicating that RMLF entered the cells through endocytosis [42]. FA has been extensively applied to facilitate targeted delivery of nanocomposites by recognizing the FA receptor-α (FR-α) overexpressed in many human cancers [43, 44]. To further verify the effect of Lipo and FA functionalization on the endocytosis of RMLF by cells, we examined the cellular uptake of RudZ, RL, RML, RMLF and RMLF + free excessive FA in 4T1 cells for 6 h with fluorescence intensity quantification (Fig. 2f, g). The uptake of RL and RML were increased considerably (3.35 and 3.13 times) when wrapped with Lipo in contrast of RudZ. With Lipo further modified with FA, the uptake of RMLF increased further due to its active targeting effect, and quantitative determination by Image J indicated that nearly 4.3-fold higher mean fluorescence intensity of RudZ was discovered in comparison. Additionally, we used excessive free FA to seal the FA receptors on the surface of 4T1 cells and found that the fluorescence was much weaker than samples without free FA. The above results indicated that Lipo wrapping and FA modification in RMLF both play irreplaceable roles in specifically targeting and internalizing 4T1 cells. Fig. 2RMLF relieves hypoxic TME and actively targets 4T1 cells. (a)Typical oxygen generation kinetics curves of RL, RML, RMLF and MnB in the presence of H2O2 ([Mn] = 10 µg/mL, [H2O2] = 1 mM) (b, d) Quantitative analysis of the fluorescence intensity of MnB and MnB with 100 µM H2O2 by Image J (n = 3). IFM images of 4T1 cells treated with MnB ([Mn] = 10 µg/mL) for 24 h with or without H2O2 (100 µM) and stained with RDPP as an O2 probe, respectively. The scale bar represents 200 μm. (c, e) Quantitative analysis of the HIF-1α fluorescence intensity of more than 100 cells in each group by Image J (n = 3). Hypoxia-relieving efficacy of RMLF (2 µM RudZ, [Mn] = 10 µg/mL) assessed by HIF-1α immunofluorescence. The scale bar represents 50 μm. (f, g) In vitro uptake of RudZ and different liposome-based samples (2 µM RudZ) on 4T1 cells analyzed by IFM. The scale bar represents 200 μm. Quantitative analysis of the fluorescence intensity of RudZ and different liposome samples by Image J (n = 3). *** indicates p < 0.001; **** indicates p < 0.0001 RMLF relieves hypoxic TME and actively targets 4T1 cells. (a)Typical oxygen generation kinetics curves of RL, RML, RMLF and MnB in the presence of H2O2 ([Mn] = 10 µg/mL, [H2O2] = 1 mM) (b, d) Quantitative analysis of the fluorescence intensity of MnB and MnB with 100 µM H2O2 by Image J (n = 3). IFM images of 4T1 cells treated with MnB ([Mn] = 10 µg/mL) for 24 h with or without H2O2 (100 µM) and stained with RDPP as an O2 probe, respectively. The scale bar represents 200 μm. (c, e) Quantitative analysis of the HIF-1α fluorescence intensity of more than 100 cells in each group by Image J (n = 3). Hypoxia-relieving efficacy of RMLF (2 µM RudZ, [Mn] = 10 µg/mL) assessed by HIF-1α immunofluorescence. The scale bar represents 50 μm. (f, g) In vitro uptake of RudZ and different liposome-based samples (2 µM RudZ) on 4T1 cells analyzed by IFM. The scale bar represents 200 μm. Quantitative analysis of the fluorescence intensity of RudZ and different liposome samples by Image J (n = 3). *** indicates p < 0.001; **** indicates p < 0.0001 In vitro biocompatibility and therapeutic efficacy of RMLF + X-ray Encouraged by cellular uptake and target efficacy, we sequentially evaluated the cytotoxicity of RudZ, MnB, RL, RML, RMLF and CDDP prior to in vitro and in vivo experiments by MTT assay (Figure S13a-d). RudZ alone were not very toxic to 4T1 cells, and cell viability remained close to 80% at concentration of 8 µM. Meanwhile, even high concentrations of MnB ([Mn] = 20 µg/mL) did not cause significant effects on 4T1 cells and had good biocompatibility, in according with similar research [45]. For RL, RML and RMLF, the cell viability remained approximately 80% at 24 h when [RudZ] = 4 µM and [Mn] reached 20 µg/mL, showing good biosafety. In addition, to mimic the microenvironment of tumor tissue, we added a small amount of H2O2 to the cell experiments. Therefore, we also evaluated the cytotoxicity magnitude of different concentrations of H2O2 (Figure S13e). The results showed that when the concentration of H2O2 was 100 µM, there was minimal effect on 4T1 cells. Hence, the concentration of H2O2 added for subsequent in vitro cell experiments was chosen to be 100 µM. Clinical findings have indicated that intermediate-dose radiation (4–10 Gy) effectively triggers tumor cell apoptosis, without exacerbating hypoxic conditions or promoting immunosuppression, hence, we have selected 4 Gy as the radiation dosage for all subsequent experiments [46]. The in vitro antitumor properties of RudZ, MnB, RL, RML, RMLF and CDDP under irradiation of X-ray were then assessed on 4T1 cells. As shown in Fig. 3a, PBS and CDDP were used as controls, PBS + X-ray resulted in unconspicuous loss of cell viability. RudZ (2 µM) or CDDP (2 µM) + X-ray groups exhibited similar cytotoxity toward 4T1 cells due to promoting more DNA double-strand break. MnB ([Mn] 10 µg/mL) + X-ray groups showed moderate cytotoxity which is associated with relieving the deficiency of O2, leading to enhancement of radiation-induced DNA damage. In contrast, RML (2 µM RudZ, [Mn] = 10 µg/mL) + X-ray showed stronger activity against 4T1, more importantly, treatment with RMLF (2 µM RudZ, [Mn] = 10 µg/mL) + X-ray killed approximately 68% of the 4T1 cells, achieving synergistically-boosted antitumor efficacy. This is due to the synergistic sensitization of RudZ, MnB and the improved cancer targeting merit of Lipo-FA, which ultimately enhance the effect of radiotherapy. Therefore, according to the cell viability, the proportion of the therapeutic effects of radiotherapy, oxygen production, ruthenium complex-based chemotherapy and targeted therapy against 4T1 cells by RMLF could be calculated as about 8.27%, 40.59%, 39.02% and 12.12%, respectively (Figure S13f). Based on above result, we chose appropriate RMLF dose (2 µM RudZ, [Mn] = 10 µg/mL) for the subsequent studies. To evaluate the synergistic radiosensitization effect, visual clonogenic assays were conducted. As shown in Fig. 3b, e, the number of colonies in the control group and RMLF appeared almost similar, indicating that RMLF didn’t affect cell proliferation and exhibited outstanding biocompatibility. As expected, the lowest survival fraction can be found form RMLF + X-ray group, giving evidence that RMLF possess strong radiosensitizing effect to inhibit 4T1 cell proliferation. ROS amplification and dsDNA damage both play important roles in the process of RT and therefore it is crucial to assess the level of intracellular ROS and γ-H2AX after RT. It is widely acknowledged that dsDNA damage caused by direct exposing to ionization and indirect ROS oxidation. We evaluated the damage status by immunofluorescence probing with γ-H2AX expression which is correlate with the amount of unrepair dsDNA breaks [47]. As shown in Fig. 3c, f, few γ-H2AX foci was generated under PBS condition, both RudZ or MnB + X-ray treatment showed obvious increase in the number of γ-H2AX foci (green) compared with the groups treated with 4 Gy X-ray alone. The RMLF + X-ray treatment resulted in an even higher number of γ-H2AX foci, further indicating superior radiosensitizing. RT-induced dsDNA damage triggers the cellular responses such as cell-cycle checkpoints inhibition, apoptosis onset and so on. Therefore, flow cytometry was applied to quantitatively analyze the effect on cell cycle distribution of 4T1 cells, as revealed in Fig. 3d, RMLF + X-ray treatment led to the most significant accumulation of the radiosensitive G2/M phase of the cell cycle. Furthermore, the flow cytometry also revealed 4T1 cells treated with RMLF + X-ray showed a higher percentage of early and late apoptosis compared with other groups (Figure S14), further substantiating the fantastic RT enhancement of RMLF. Excess ROS will damage the nucleus, membrane structures, organelles and eventually lead to apoptosis. The production of ROS form RMLF under X-ray irradiation was monitored using the fluorescence indicator 2′,7′-dichlorofluorescin diacetate (DCFH-DA), whose fluorescence intensity is positively associated with intracellular ROS content. The PBS group without X-ray treatment served as a blank control, the IFM images indicated that 2 µM RudZ combined with 4 Gy X-ray produced significantly more ROS compared to the control and X-ray treatment alone. In contrast, MnB and RMLF demonstrated weak green fluorescence, whereas the fluorescence for RMLF + X-ray treatment was further enhanced compared to RudZ + X-ray, likely due to Lipo encapsulation significantly increasing the uptake of the RudZ and consequently boosting the ROS yield (Figure S15a).This experimental result was further validated by the quantitative analysis of fluorescence intensity, ROS was significantly increased in cells cultured with RMLF + X-ray compared to the other groups, suggesting that FA modification in Lipo is expected to further enhance the toxicity of 4T1 cells (Figure S15b). Taken together, we conclude that RMLF enhance therapeutic efficacy under X-ray exposure. Fig. 3Evaluation of radiotherapy sensitizing activity of RMLF in vitro. (a) Cell viability of 4T1 cells treated with the RudZ (2 µM), CDDP (2 µM), MnB ([Mn] = 10 µg/mL), RML (2 µM RudZ, [Mn] = 10 µg/mL) and RMLF (2 µM RudZ, [Mn] = 10 µg/mL) (n = 3). The dose of X-ray is 4 Gy. (b, e) The quantitative analysis of survival fraction (SF) from clone formation under different treatments (n = 3). And optical photographs of clonogenic survival assay under different treatments. (c, f) Quantitative analysis of the γ-H2AX fluorescence intensity of more than 100 cells in each group by Image J (n = 3). The DNA damage effect of RMLF (2 µM RudZ, [Mn] = 10 µg/mL) was assessed by γ-H2AX immunofluorescence. The scale bars represent 200 μm. (d) Flow cytometry was employed to perform cell cycle analysis of 4T1 cells subjected to different treatments. *** indicates p < 0.001; **** indicates p < 0.0001 Evaluation of radiotherapy sensitizing activity of RMLF in vitro. (a) Cell viability of 4T1 cells treated with the RudZ (2 µM), CDDP (2 µM), MnB ([Mn] = 10 µg/mL), RML (2 µM RudZ, [Mn] = 10 µg/mL) and RMLF (2 µM RudZ, [Mn] = 10 µg/mL) (n = 3). The dose of X-ray is 4 Gy. (b, e) The quantitative analysis of survival fraction (SF) from clone formation under different treatments (n = 3). And optical photographs of clonogenic survival assay under different treatments. (c, f) Quantitative analysis of the γ-H2AX fluorescence intensity of more than 100 cells in each group by Image J (n = 3). The DNA damage effect of RMLF (2 µM RudZ, [Mn] = 10 µg/mL) was assessed by γ-H2AX immunofluorescence. The scale bars represent 200 μm. (d) Flow cytometry was employed to perform cell cycle analysis of 4T1 cells subjected to different treatments. *** indicates p < 0.001; **** indicates p < 0.0001 Evaluation of mitochondria dysfunction caused by RMLF + X-ray Mitochondria play an irreplaceable role in tumorigenesis and development, making it as a vital target for cancer therapy. Given that a burst of ROS can lead to oxidative damage to mitochondrial proteins and DNA, we speculate that RMLF enters cells and subsequently acts on mitochondria, thereby inducing increased ROS production within mitochondria during RT. To investigate the subcellular localization of RMLF upon internalization, we employed the Mito Tracker Green kit to assess the subcellular distribution of RudZ, RL and RMLF in 4T1 cells by confocal laser scanning microscope (CLSM) (Fig. 4a). Co-localization experiments showed that RudZ could localize to mitochondria, although with a low Pearson coefficient of 0.338, possibly due to its weak entry capabilities into the cell. Intriguingly, after being wrapped by Lipo, the uptake capacity of RudZ increased and the Pearson coefficient of RL increased to 0.474. Upon further modification of Lipo with FA, the co-localization coefficient reached its highest value of 0.727, attributable to the active targeting of RMLF. The constructed radiosensitizer RMLF (red fluorescence) displayed a strong co-localization with Mito Tracker Green (green fluorescence), resulting in overlapping red and green signals and an overall yellow fluorescence within the cell. Consequently, we can infer that upon cellular internalization of RMLF, a high proportion of RudZ targets mitochondria, resulting in more ROS being produced by the RT process. Mitochondrial damage typically coincides with alterations in mitochondrial membrane permeability (MMP) [48], with MMP changes being the most common manifestation of mitochondrial dysfunction, frequently detected through the JC-1 assay. Therefore, we used the JC-1 probe with flow cytometry analysis to evaluate Δψm changes following administration of various radiosensitizers under X-ray irradiation. As depicted in Figure S16, 4T1 cells treating by RMLF alone had no significant effect on JC-1 signal compared to the control. However, when 4T1 cells were treated by RMLF with 4 Gy X-ray exposure, a substantial increase in the green-to-red fluorescence ratio was observed, showing maximum Δψm damage. These findings were in accordance with studies in ROS and apoptosis, suggest that elevated ROS generation lead to severe MMP disruption, subsequently initiating the apoptosis pathway for cell death. Encouraged by the positive response of the designed RMLF mitochondrial targeting capabilities, we anticipated mtDNA as a potential target due to the DNA affinity of RudZ. MtDNA, situated close to the respiratory strand of the inner mitochondrial membrane, lacks protective proteins, rendering it more susceptible to ROS-inducing oxidative damage. Moreover, the capacity to repair mtDNA damage is considerably lower than that of nuclear ds-DNA [20]. The mtDNA damage capacity of RMLF was investigated by a DNA intercalating dye (PicoGreen) staining technique (Fig. 4b) [49]. A number of punctate patterns surrounding the nucleus, representing stained mtDNA, were observed. Under the 4 Gy X-ray conditions, the luminescence signals outside the nucleus of RMLF and RudZ treated groups decreased significantly. In contrast, treatments of MnB or X-ray alone demonstrated no significant change in the punctate luminescence. Above PicoGreen experiment results indicated that the successful mitochondrial targeting of RMLF can cause oxidative stress in mitochondria and subsequently damage mtDNA. 8-hydroxydeoxyguanosine (8-OH-dG) as a major marker of oxidative mtDNA damage, is selected to assess the potential of RMLF + X-ray group to damage the mtDNA by immunofluorescence staining and quantitative fluorescence measurements [50]. As shown in Fig. 4c and Figure S17, there was a significant increase in the fluorescence expression of 8-OH-dG emitted from the RMLF + X-ray or RudZ + X-ray groups. By contrast, the enhancement of fluorescence intensity from irradiation alone or MnB + X-ray groups were lower. MtDNA damage in cytoplasmic matrix was also investigated by quantitative real-time polymerase chain reaction (qRT-PCR) analysis. The mitochondrial genomic segment was listed in Table S5, and the results were presented in Figure S18, RMLF + X-ray or RudZ + X-ray treatments markedly reduced the transcription levels of displacement loop (D-Loop), NADH-dehydrogenase 6 (ND6), ATP synthase complex subunits 6 (ATP6) and cytochrome c oxidase complex subunits I (CO I) compared to the other groups. These results all testified the mtDNA oxidative damage via effective mitochondrial targeting capacity of RMLF. Interestingly, the transcriptional level of D-Loop was increased in the RMLF and X-ray treatment alone groups compared to the control groups, this may due to the low ionizing radiation make changes in mitochondrial DNA copy number, with a tendency for an increase in mtDNA copy number. In the event of mitochondrial suffers mild damage, cells may attempt to restore mitochondrial function through an increase in mtDNA copy number. This phenomenon is especially prominent in the D-loop region, as it serves as the starting point for mtDNA replication, thus an increase in mtDNA copy number may lead to a corresponding increase in the copy number of the D-loop region. Nevertheless, some studies have found that under certain circumstances, mitochondrial severe damage may result in a decrease in mtDNA copy number. This could potentially be attributed to severe mitochondrial damage that renders the cell incapable of compensating for mitochondrial functional defects by increasing the mtDNA copy number. In such instances, the copy number in the D-loop region may also correspondingly decrease [51–53]. In vitro cGAS-STING activation by two-way enhancement of mitochondrial DNA leakage and Mn2+ assisting During the mitochondria damage process, mtDNA could leak into the cytoplasm as DAMPs, activating cGAS to contribute to the cyclic GMP-AMP (2´ 3´-cGAMP) production. Subsequently, 2´3´-cGAMP serves as a second messenger, responsible for STING-TBK1-IRF3 activation, which can promote the expression of IFNs to media antitumor immune response. Recently, Mn2+ is being exploited to activate the cGAS-STING mediated IFNs [54]. On this basis, we investigated the potential of a two-way enhancement tactic on cGAS-STING activation by using RMLF + X-ray. The impact of RMLF on the escape of mtDNA were evaluated via CLSM when the nucleus and mtDNA of 4T1 cells were co-stained by Hoechst (blue) and TFAM (green), respectively. As shown in Fig. 4e, the mtDNA with green fluorescence were close-packed and surrounding the nucleus in the control group, comparatively, the mtDNA of RMLF + X-ray treated cells were dividing and straying away from the nucleus, as illustrated by red box. To ensure the mtDNA leakage results on the RMLF + X-ray groups, we performed qRT-PCR analysis of mtDNA escape levels in cell supernatants, notably, a significantly higher number of escaped mtDNA fragments were detected in the supernatants of 4T1 cells treated with RMLF + X-ray compared to those in other control groups (Fig. 4g). The mechanism of mitochondrial release of mtDNA may be closely related to the BAX/BAK protein pathway [55]. Therefore, we measured the expression of BAK/BAX protein by Western blot (WB) analysis to understand the pathway of mtDNA release (Fig. 4f). These results suggesting that the oxidative mtDNA and its subsequent leakage from mitochondria depended on BAK/BAX oligomerization under the exposure of RMLF + X-ray. Next the proposed properties of RMLF to activate the cGAS-STING pathway under 4 Gy X-ray exposure was studied. WB analysis showed that cGAS and phosphorylated STING (p-STING) and phosphorylated NF-κB (p-NF-κB) were significantly increased in the RMLF + X-ray group, which further validated the dual activation of the cGAS-STING immune signaling pathway by RMLF (Fig. 4f). Similar results were also obtained in the qRT-PCR analysis (Fig. 4d). In addition, we used enzyme-linked immunosorbent assay (ELISA) to analyze 2´3´-cGAMP release in the lysates of 4T1 cells by different treatment modalities. Notably, the RMLF + X-ray group conferred higher activity to produce more 2´3´-cGAMP (Figure S19a). Additionally, we conjectured that the mtDNA, upon its release from mitochondria, might incur damage from ROS and subsequently be extruded into the extracellular milieu. We employed ELISA to quantify 8-OH-dG expression, a marker of DNA damage, in cell supernatants. Of all groups tested, the RMLF + X-ray group demonstrated the highest values (S19b). This provides evidence to suggest that our purpose-built RMLF, is capable of inflicting significant damage on mtDNA, and can also induce mtDNA leakage when used in conjunction with X-ray irradiation. Upon activation of the cGAS-STING intrinsic immune signaling pathway, NF-κB expression induces the expression of IFNs and other immune-related cytokines, thereby inducing the maturation of DCs and increasing the proportion of CD8+ T cells [56]. As cells expressing CD86 and CD80 are markers of DCs maturation [57], we used flow cytometry to analyze the effect of RMLF + X-ray treatment of 4T1 cells to induce DCs maturation. As shown in Figure S20, the expression of CD86 and CD80 was increased 5.16-fold in the RMLF + X-ray group compared with the control group. The results of this experiment indicated that RMLF + X-ray treatment could induce the maturation of DCs in 4T1 cells by activating the cGAS-STING pathway, thus triggering a strong anti-tumor immune response. Fig. 4RMLF induces oxidative damage to mtDNA and prompts its translocation to the cytoplasm, results in a dual activation of the cGAS-STING immune signaling pathway. (a) CLSM images of RudZ, RL, RMLF and Mito Tracker Green staining. The scale bar represents 40 μm. (b) PicoGreen staining was performed to analyze mtDNA damage. The scale bar represents 40 μm. (c) 8-OH-dG immunofluorescence staining was performed to analyze mtDNA damage. The scale bar represents 40 μm. (d) qRT-PCR analysis of the relative expression levels of cGAS-STING, NF-κB, IRF3 mRNA after different treatments of 4T1 cells (n = 3). (e) mtTFA antibody was used to distinguish mtDNA (TFAM-mtTFA Green) and Hoechst-stained nuclei, to detect the escape of mtDNA from mitochondria to the cytoplasm. The scale bar represents 40 μm and 10 μm. (g) qRT-PCR analysis of mtDNA escape levels in 4T1 cells supernatants after different treatments (n = 3). (f) Western blot analysis of the protein pathway associated with mtDNA release and activation of the cGAS-STING intrinsic immune signaling pathway in 4T1 cells. * indicates p < 0.05; ** indicates p < 0.01; *** indicates p < 0.001; **** indicates p < 0.0001 RMLF induces oxidative damage to mtDNA and prompts its translocation to the cytoplasm, results in a dual activation of the cGAS-STING immune signaling pathway. (a) CLSM images of RudZ, RL, RMLF and Mito Tracker Green staining. The scale bar represents 40 μm. (b) PicoGreen staining was performed to analyze mtDNA damage. The scale bar represents 40 μm. (c) 8-OH-dG immunofluorescence staining was performed to analyze mtDNA damage. The scale bar represents 40 μm. (d) qRT-PCR analysis of the relative expression levels of cGAS-STING, NF-κB, IRF3 mRNA after different treatments of 4T1 cells (n = 3). (e) mtTFA antibody was used to distinguish mtDNA (TFAM-mtTFA Green) and Hoechst-stained nuclei, to detect the escape of mtDNA from mitochondria to the cytoplasm. The scale bar represents 40 μm and 10 μm. (g) qRT-PCR analysis of mtDNA escape levels in 4T1 cells supernatants after different treatments (n = 3). (f) Western blot analysis of the protein pathway associated with mtDNA release and activation of the cGAS-STING intrinsic immune signaling pathway in 4T1 cells. * indicates p < 0.05; ** indicates p < 0.01; *** indicates p < 0.001; **** indicates p < 0.0001 The innate immune response can be largely influenced by immunosuppressive TME, for example, hypoxia can lead to low immune responsiveness. Furthermore, another major cause of low immune responsiveness is the suppression of immune cells in the tumor, such as tumor-associated macrophages (TAMs), regulatory T cells (Tregs), etc., which always lead to immunotherapy failure by limiting T cell infiltration and depleting cytotoxic T lymphocytes [58, 59]. And yet, the oxidized mtDNA derived from cancer cells as DAMPs has demonstrated potent immunostimulatory properties [60]. We used flow cytometry to verify whether escaped mtDNA can lead to the induction of TAMs polarization from the M2 phenotype to the M1 phenotype, leading to the regulation of TME and activation of innate immunity [61]. CD86 and CD206 are conventionally identified as characteristic markers for M1 and M2 immunocytes within TAMs, respectively. It is interesting to note that the proportion of M1 macrophages was significantly increased in the RMLF + X-ray group, while the proportion of M2 macrophages was significantly decreased (Figure S21). Taken together, these results suggested that RMLF + X-ray can reverse immunosuppressive TME by converting M2 phenotype macrophages to M1 phenotype macrophages. In vivo biosafety and imaging of RMLF Next, we evaluated the blood biocompatibility of RMLF, and the results of the study are shown in Figure S22. PBS was used as the negative control group and deionized (DI) water was used as the positive control group, respectively. Almost no significant hemolysis was observed to occur from the erythrocyte hemolysis photographs of mice. The highest hemolysis rate was only less than 3% after treatment with different concentrations of RMLF. Therefore, we can speculate that RMLF is less toxic to erythrocytes and has good biocompatibility. The real-time fluorescence imaging system was employed to measure the in vivo bio-distribution of RMLF at different time points after intravenous injection into tumor-bearing mice with RML as controls. As shown in Figure S23a, the tissue organs in the fluorescent images of the animals are arranged in the order of heart, liver, spleen, lung, kidney, and tumor. Due to the targeting behavior of RMLF, it began to accumulate at the tumor site within 6 h after tail vein injection and the fluorescence intensity reached a peak at 12 h, then gradually weakened in the next 24 h. In contrast, the mice treated with RML showed fewer signals at the tumor sites, which might be ascribed to the quick elimination by the immune system and non-targeted ability (Figure S23b, c). In vivo radioimmunotherapy effect and inducing cGAS-STING pathway via two-way synergistic strategy of RMLF The in vivo radioimmunotherapy effects of RMLF were assessed by tumor-bearing mice (Fig. 5a). When the cancer volume reached about 80 mm3, the mice were intravenously injected with PBS, RMLF, PBS + X-ray, RudZ + X-ray, MnB + X-ray, RMLF + X-ray or CDDP + X-ray treatments respectively. As depicted in Fig. 5b-c, the groups of RudZ or MnB + X-ray have a certain radioimmunotherapy effect compared with PBS + X-ray group. Specifically, the relative tumor volume and tumor weight of the mice conducted with RMLF + X-ray were greatly lower than all other groups, reflecting the significant importance of targeting ability and synergistic therapeutic effects in the practical management of cancer in vivo (Fig. 5b-d). The bodyweight of mice in each group were recorded and no obvious change were observed during the 12 days treatment (Fig. 5e). Moreover, no significant physiological morphology changes of the mice treated with different treatment modalities were observed in the main organs including heart, liver, spleen, lung, and kidney which were confirmed by hematoxylin-Eosin (H&E) staining, implying its good bio-safety in the in vivo circulation. Yet, a discernible degree of severe damage was present in the tumor tissue of the group treated with RMLF + X-ray (Figure S24), further attesting to the therapeutic efficacy of the devised radiosensitizer. Meanwhile, white blood cell (WBC), red blood cell (RBC), platelet count (PLT) are important parameters for the detection of routine blood flow. Alanine aminotransferase (ALT), aspartate aminotransferase (AST), blood urea nitrogen (BUN), creatinine (CREA) are important indicators of routine liver and kidney function. The results of the study showed no significant changes in the indicators, which indicates that RMLF has good biocompatibility (Figure S25). Furthermore, to evaluate the effect of RMLF + X-ray in causing DNA damage and alleviating hypoxia at tumor sites in 4T1 tumor-bearing mice, we performed γ-H2AX/HIF-1α immunofluorescence staining analysis on tumor tissues. The control group treated with PBS exhibited substantial tumor proliferation, virtually no γ-H2AX foci, and the most intense expression of HIF-1α fluorescence. In contrast, among all the treated groups, the RMLF + X-ray group displayed a substantial surge in γ-H2AX foci (as indicated by red fluorescence) and a concurrent decline in HIF-1α fluorescence (depicted by green fluorescence), signifying a marked difference (Figure S26). This indicates severe DNA damage in tumor tissue cells, which suggests that RMLF can effectively alleviate TME hypoxia and cause DNA damage in tumor cells, thus exerting its radiotherapy sensitizing effect. Subsequently, to verify whether the constructed radiotherapy sensitizer RMLF could induce the cGAS-STING pathway activation on the 4T1 tumor-bearing mice model, we performed immunohistochemical (IHC) section analysis of tumors treated with different treatment modalities (Figure S27). As expected, the RMLF + X-ray group greatly activated the cGAS-STING pathway by causing mtDNA damage and the adjuvant effect of Mn2+, which in turn promoted increased expression of nuclear factors (NF-κB). Encouraged by these results, we used flow cytometry to analyze the maturation of DCs and polarization of TAMs to assess the effect of anti-tumor immune response after RMLF combined with RT. Among all groups, the RMLF + X-ray group had the highest levels of CD80 and CD86, which is a marker of DCs maturation. Meanwhile, the proportion of M1 macrophages (CD86) was significantly increased in the RMLF + X-ray group, while the proportion of M2 macrophages (CD206) was significantly decreased (Fig. 5f. I-III). This was also consistent with the IHC section results (Figure S28). All these suggest that RMLF combined with RT can reverse immunosuppressive TME by converting M2 phenotype macrophages into M1 phenotype macrophages. Fig. 5Evaluation of in vivo radioimmunotherapy. (a) Schematic design of in vivo radioimmunotherapy. (b) Relative tumor volumes change curve during various treatment (n = 5). (c) Tumor photos at the end of various treatments. (d) Tumor weights at the end of various treatments (n = 5). (e) Body weight of mice during various treatments (n = 5). (f) Quantification of the proportion of cells expressing CD80 in tumor draining lymph nodes at the end of treatment by flow cytometry (I). Quantification of the proportion of cells expressing CD86 and CD206 in tumor tissue at the end of treatment by flow cytometry (II, III). (g-j) The levels of 2’3’-cGAMP (g), TNF-α, IFN-β, and IL-6 (h, i, j) in tumors of different treatment groups of mice were analyzed by ELISA (n = 3). * indicates p < 0.05; ** indicates p < 0.01; *** indicates p < 0.001; **** indicates p < 0.0001 Evaluation of in vivo radioimmunotherapy. (a) Schematic design of in vivo radioimmunotherapy. (b) Relative tumor volumes change curve during various treatment (n = 5). (c) Tumor photos at the end of various treatments. (d) Tumor weights at the end of various treatments (n = 5). (e) Body weight of mice during various treatments (n = 5). (f) Quantification of the proportion of cells expressing CD80 in tumor draining lymph nodes at the end of treatment by flow cytometry (I). Quantification of the proportion of cells expressing CD86 and CD206 in tumor tissue at the end of treatment by flow cytometry (II, III). (g-j) The levels of 2’3’-cGAMP (g), TNF-α, IFN-β, and IL-6 (h, i, j) in tumors of different treatment groups of mice were analyzed by ELISA (n = 3). * indicates p < 0.05; ** indicates p < 0.01; *** indicates p < 0.001; **** indicates p < 0.0001 In the cGAS-STING pathway, 2´3´-cGAMP, TNF-α, IFN-β, and IL-6 play pivotal roles, collaboratively participating in immune responses and anti-tumor actions. 2´3´-cGAMP, an intracytoplasmic second messenger synthesized after cGAS activation, transduces signals by binding to the C-terminus of STING protein. Phosphorylation of STING protein further activates IRF3 and NF-κB, which induces type I interferon (IFN-β) and other immune-related cytokines (TNF-α and IL-6) expression. Intriguingly, we observed a substantial elevation in the expression levels of 2´3´-cGAMP, TNF-α, IFN-β and IL-6 in the RMLF + X-ray treatment group. Conversely, the CDDP + X-ray treatment group did not exhibit a significant impact on the expression of 2´3´-cGAMP or the analyzed cytokines (Fig. 5g-j). These results suggest that RMLF + X-ray induces a strong anti-tumor immune response in 4T1 tumor-bearing mice through dual activation of the cGAS-STING intrinsic immune signaling pathway. To assess the level of apoptosis to tumor cells induced by the RMLF + X-ray group and its contribution to the immune response, we also performed TUNEL analysis, CD4+/CD8+ immunofluorescence staining analysis. In the TUNEL analysis of all groups, the RMLF + X-ray group exhibited the highest level of apoptosis, which indicates that RMLF has an excellent radiosensitizing effect and can effectively inhibit the growth of the primary tumor (Figure S29). The presence of cytotoxic T cells (CTL, CD8+ T cells) in tumors is associated with a good prognosis during cancer treatment [62]. Among all groups of immunofluorescence sections, CD8+ and CD4+ fluorescence intensity was the strongest in the RMLF + X-ray group (Figure S29), which indicates that the ratio of CD8+ T cells and CD4+ T cells was significantly higher compared to all the other group, activating the immune response and effectively inhibiting tumor recurrence and metastasis. The long-term protection effects of RMLF Although RT can effectively inhibit the growth of primary tumors, it exhibits limited efficacy against recently metastasized tumors. Thus, it is imperative not only to enhance the effectiveness of RT but also to prevent tumor recurrence and metastasis. To evaluate the long-term protection effects of the anti-tumor immune response triggered by RMLF in conjunction with RT, we established a recurrent metastasis model using 4T1 tumor-bearing mice (Fig. 6a). Remarkably, the RMLF + X-ray group had the best long-term immune memory effect. Upon secondary inoculation of tumor cells and subsequent extended observation, this group exhibited the slowest tumor growth rate, smallest tumor volume, lightest tumor weight, and most potent anti-tumor immune response (Fig. 6b-d). This outcome may be attributed to the dual activation of the cGAS-STING signaling pathway by RudZ and MnB present in the RMLF + X-ray group, coupled with the catalysis of endogenous H2O2 to generate oxygen, thereby remodeling the immunosuppressive TME. The experimental results of TUNEL analysis, CD4+/CD8+ immunofluorescence-stained sections, and H&E staining of tumor tissues further verified the above conclusions (Figure S30). Additionally, no significant alteration in mouse body weight was observed throughout the long-term immunological memory treatment experiment, underscoring the safety of our therapeutic strategies (Fig. 6e). Then, the content of effector memory T cells (Tem) and central memory T cells (Tcm) is a key indicator to evaluate the long-term immune memory effect, and their cell phenotypes may be defined as CD44hi CD62Llo and CD44hi CD62Lhi [61, 63]. We detected the expression of relevant memory T cells in the spleen of mice after the end of treatment by flow cytometry. The levels of Tem and Tcm were significantly increased in the RMLF + X-ray group compared with the other groups (Fig. 6f, g), indicating that a strong immune memory effect was generated through our treatment, which could provide effective protection against tumor recurrence and metastasis. In conclusion, RMLF demonstrates not only its potency as a radiotherapy sensitizer but also its capacity to effectively incite an anti-tumor immune response, resulting in a lasting memory effect. Fig. 6Evaluation of long-term immune memory effects in vivo. (a) Schematic design of long-term immunological memory effect evaluation. (b) Relative tumor volume change curve during treatment (n = 5). (c)Tumor weights at the end of treatment (n = 5). (d) Tumor photos at the end of treatment. (e) Body weight of mice during treatment (n = 5). (f, g) The determination of effector memory T cells (CD45+CD3+CD4+CD44+CD62− and CD45+CD3+CD8+CD44+CD62−) and central memory T cells (CD45+CD3+CD4+CD44+CD62+ and CD45+CD3+CD8+ CD44+CD62+) in CD8+ T cells and CD4+ T cells in spleens. * indicates p < 0.05; ** indicates p < 0.01, **** indicates p < 0.0001 Evaluation of long-term immune memory effects in vivo. (a) Schematic design of long-term immunological memory effect evaluation. (b) Relative tumor volume change curve during treatment (n = 5). (c)Tumor weights at the end of treatment (n = 5). (d) Tumor photos at the end of treatment. (e) Body weight of mice during treatment (n = 5). (f, g) The determination of effector memory T cells (CD45+CD3+CD4+CD44+CD62− and CD45+CD3+CD8+CD44+CD62−) and central memory T cells (CD45+CD3+CD4+CD44+CD62+ and CD45+CD3+CD8+ CD44+CD62+) in CD8+ T cells and CD4+ T cells in spleens. * indicates p < 0.05; ** indicates p < 0.01, **** indicates p < 0.0001 Conclusion In this work, the Ruthenium (II)/Manganese (IV) complexes-based liposome nanocomposite RMLF has been successfully fabricated. It exhibits active targeting towards 4T1 cells, H2O2-responsive oxygen production, effectively induces mitochondrial oxidative phosphorylation dysfunction and triggers sustained ROS production, leading to mtDNA damage and subsequent release into the cytoplasm. In combination with the release of Mn2+, a dual activation of the cGAS-STING innate immune signaling pathway occurred, inducing M1 type polarization of TAMs and ameliorating the TME of hypoxic immunosuppression, while inducing the maturation of DCs and effectively eliciting a strong anti-tumor immune response. This study provides a new design concept for the development of radiotherapy sensitizers in radiation-immunology and has the potential to be applied to various aspects of precision medicine in the future. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1