Assembling Ruthenium Complexes to Form Ruthenosome Unleashing Ferritinophagy-Mediated Tumor Suppression.

We introduce ruthenosomes, a fusion of liposomal and reactive oxygen species (ROS)–generating properties meticulously engineered as potent ferroptosis inducers (FINs), marking a significant advancement in metallodrug design for cancer therapy. Formed through the self-assembly of oleate-conjugated ruthenium complexes, these ruthenosomes exhibit exceptional cellular uptake, selectively accumulating in mitochondria and causing substantial disruption. This targeted mitochondrial damage significantly elevates ROS levels, triggering autophagy and selectively activating ferritinophagy. Together, these processes sensitize cancer cells to ferroptosis. In vivo, ruthenosomes effectively suppress colorectal tumor growth, underscoring their therapeutic potential. Our study pioneers a design strategy that transforms ruthenium complexes into liposome-like structures capable of inducing ferroptosis independent of light activation. By leveraging ruthenosomes as multifunctional nanocarriers, this research offers a versatile and powerful platform for ROS-mediated, ferroptosis-driven cancer cell eradication. ## Results and Discussion Results and Discussion Ruthenium(II) complexes have been extensively utilized as ROS generators through mitochondrial targeting 18 or light irradiation, 19 inducing apoptosis in cancer cells. However, their potential as FINs is often limited by inadequate cellular uptake/accumulation and insufficient ROS generation. A recent study achieved a significant breakthrough by developing a Ru(II) polypyridine complex that exhibited enhanced cellular uptake and increased ROS production, ultimately leading to photoinduced ferroptosis in breast cancer cells, 20 providing empirical support for the use of Ru(II) complexes as potential FINs. Building on this insight, and inspired by the design of boron-containing liposomes (boronsomes), for improved intracellular delivery of boron molecules, 21 , 22 we designed a unique polypyridyl Ru(II)-based metallosurfactant 23 , 24 [Ru(bipy) 2 ( N 4 , N 4’ -bis(2-oleamidoethyl)-2,2′-bipyridine-4,4′-dicarboxamide)](PF 6 ) 2 ( RuOA2 , Scheme S1 ). This complex self-assembles into liposome-like structures, significantly enhancing cellular delivery and driving potent ferroptosis activation through ROS generation in cancer cells ( Figure 1 ). As presented in Figure 1 , the positively charged [Ru(bipy) 3 ] 2+ (Ru-trisBP) unit serves as the headgroup of the metallosurfactant. Unlike conventional designs that employ alkyl chains as tails, RuOA2 incorporates two oleate chains derived from oleic acid (OA)-an abundant component in cancer cell plasma membranes, thus enhancing cancer cell targeting. 25 The optimized structure of RuOA2 ( Figure 2 a) features an octahedral headgroup with a diameter of 12.5 Å and the flexible tails measuring 23.2 Å facilitating conformational adaptability. This large headgroup extends the overall length of RuOA2 to 35.8 Å, considerably longer than typical phospholipid designs. Molecular dynamics (MD) simulations reveal that RuOA2 self-assembles into a lipid-bilayer-like structure in aqueous solutions, with some OA tails oriented outward ( Figure 2 b). The bilayer’s thickness is approximately 78.25 Å–nearly double that of a traditional phospholipid bilayer. The spacing between adjacent headgroups ( Figure 2 c) indicates a densely packed assembly. Experimentally, UV–vis spectra of RuOA2 show hypochromicity and a redshift in the π–π* (∼280 nm) transitions of the bipyridine ligands, along with metal-to-ligand charge transfer transitions (∼480 nm) in 1XPBS compared to in methanol ( Figure S1 ), suggesting extended bipyridine J-aggregate stacking, consistent with the simulated tight packing of [Ru(bipy) 3 ] 2+ units along the bilayer surfaces. Upon excitation at 450 nm, RuOA2 emits bright red phosphorescence in water and shows no sensitivity to oxygen quenching ( Figure S2 ), further indicating that the self-assembled close packing effectively shields the excited states from oxygen. Figure 2 Dynamic self-assembly of RuOA2 in aqueous environment. (a) Spatial configuration and structural parameters of RuOA2 . Cross section (b) and top view (c) representations of bilayers formed from the self-assembly of RuOA2 , based on a 128-monomer MD simulation in 3200 water molecules. In panel (b), most RuOA2 molecules are depicted in wireframe mode, with one molecule shown in stick mode. In panel (c), intermolecular distances between Ru centers are highlighted in yellow, with number labels corresponding to distances 12.31, 9.81, 12.62, 9.87, 9.94, and 11.25 Å. (d) TEM images of the ruthenosome in PBS under weak electron beam and after electron beam irradiation for a few minutes. (e) Diameter distribution of RuOA2 self-assemblies in 1× PBS buffer as determined by DLS analysis. (f) Zeta potentials of Ru-trisBP, DOPC, OA, Ru-trisBP@DOPC, Ru-trisBP@OA, and RuOA2 in 1× PBS buffer at a concentration of 20 μM, alongside RuOA2 in water at the same concentration. Mean ± SD ( n = 3). (g) Schematic depiction of the dynamic self-assembly process of RuOA2 in an aqueous environment. Experimentally, RuOA2 self-assembled into particle clusters with an average size of ∼42.0 nm in water ( Figure S3a, S3b ). After negative staining with uranyl acetate (UAc), these self-assemblies transformed into evenly distributed vesicles with an average size of ∼32.0 nm, indicating that ionic conditions significantly influence the assembly process. Under physiological conditions simulated by 1xPBS, RuOA2 self-assembled into liposome-like vesicles ( Figure 2 d) with a predominant diameter of ∼423 nm ( Figure 2 e). This size is notably larger than the liposomes formed by 1,2-dioleoyl- sn -glycero-3-phosphocholine (DOPC) and OA encapsulating Ru-trisBP, which exhibit average diameters of ∼335 nm and ∼90 nm, respectively ( Figure S3a and S3b ). Similar to these liposomes, RuOA2 self-assemblies adopt a bilayer structure. However, unlike the typically negative surface charges seen in Ru-trisBP@DOPC and Ru-trisBP@OA liposomes, zeta potential measurements reveal that RuOA2 -assembled ruthenosomes possess a high positive surface charge (+41.58 mV) ( Figure 2 f), which offers a significant advantage for mitochondrial targeting. 26 To investigate the effect of ionic conditions, we examined RuOA2 assemblies in 0.01xPBS and 0.1xPBS, yielding average sizes of ∼75 nm and ∼159 nm, respectively. These results suggest that ion neutralization reduces electrostatic repulsion, stabilizing larger ruthenosome ( Figure S3c ). During TEM imaging, RuOA2 self-assemblies exhibited dynamic instability. Electron beam exposure triggered a fission event, forming smaller vesicles (∼25 nm) within the original ruthenosomes ( Figures S3d, S3e ). These newly formed vesicles remained adhered due to strong hydrophobic interactions between their outward-facing oleate chains, maintaining the integrity of the ruthenosome as a multivesicular, liposome-like structure ( Figure 2 g). Prolonged irradiation induced fusion, where small vesicles gradually coalesced into larger liposome-like structures ( Video S1 ). This reversible fission-fusion behavior highlights the dynamic adaptability of RuOA2 self-assemblies to environmental stimuli. Combined with their transformation from particle clusters in water to larger liposome-like structures under physiological conditions, these observations demonstrate that the size reduction is governed by both intrinsic self-assembly properties and dynamic environmental responsiveness. RuOA2 ruthenosomes exhibited broad cytotoxicity against a range of cancer cell lines commonly used as in vitro tumor models, including HeLa (cervical cancer), HepG2 (liver cancer), MCF7 (breast cancer), as well as cell lines from refractory cancers such as hypermutated CT26 and HCT116 (colorectal cancer, CRC), H1975 and A549 (lung cancer), and drug-resistant PANC-1 (pancreatic cancer). Notably, the IC50 values for RuOA2 were below 20 μM for all these cancer cells, with the greatest potency observed in CRC cells, achieving submicromolar IC50 values (0.33 ± 0.008 μM) ( Figure S4 ). CRC is the third leading cause of cancer-related deaths worldwide, characterized by a high frequency of mutations and significant heterogeneity in immune responses and metabolic profiles among patients. Given the pressing need to unravel the mechanisms regulating ferroptosis in CRC, as well as to harness these pathways for drug development and potential clinical translation, 27 we focused further investigation on HCT116 cells to explore RuOA2 ruthenosome’s role in ferroptosis regulation. Upon introducing RuOA2 to HCT116 cells, we observed rapid accumulation of RuOA2 at the plasma membrane within 10–20 min ( Figure 3 a). Fluorescence imaging further revealed transmembrane localization of RuOA2 after 30 min of treatment ( Figure 3 b). By 1 h, RuOA2 remained membrane-bound, but a significant portion had internalized into the cell, with some colocalizing with mitochondria ( Figure 3 c and 3 d). After an additional hour, the majority of RuOA2 was fully internalized and predominantly localized within mitochondria. By the 4-h mark, no RuOA2 was detected at the plasma membrane, and all had accumulated in mitochondria. Pearson coefficien analysis confirmed a time-dependent increase in mitochondrial accumulation upon RuOA2 treatment, with r values rising from 0.59 at 1 h to 0.89 at 4 h, indicating selective mitochondrial targeting. The shift in mitochondrial morphology, from tubular to blob-like structures, at 4 h further suggests mitochondrial stress, likely due to elevated mitochondrial ROS (mtROS) levels. 28 In comparison, Ru-trisBP@DOPC showed higher cellular delivery than Ru-trisBP@OA, but exhibited nonspecific localization in both mitochondria and lysosomes ( Figure S5 ). Figure 3 Intracellular delivery of RuOA2 to mitochondria in vitro. (a) Merged bright field and confocal fluorescence images of HCT116 cells upon the treatment of 5 μM RuOA2 (red) for 10, 15, and 20 min, costained with Hoechst (blue). The scale bar represents 20 μm. (b) Confocal fluorescence images of HCT116 cells upon the treatment of 5 μM RuOA2 (red) for 30 min, costained with Cell Mask (cyan) and DAPI (blue). The scale bars represent 10 μm. (c) Confocal fluorescence images of HCT116 cells treated with 5 μM RuOA2 (red) for 1, 2, and 4 h, costained with MitoTracker (green), CellMask (cyan), and DAPI (blue). Scale bars represent 10 μm. (d) Zoomed-in images of the white dash line squares in panel (c), along with the corresponding Pearson coefficient (Pearson’s r) for the colocalization between RuOA2 fluorescence and mitochondrial fluorescence. Scale bars represent 2 μm. (e) Schematic depiction of intracellular delivery of RuOA2 via transition of ruthenosome into vesicles targeting mitochondria. (f) Cellular uptake of Ru-complex in CT26, HCT116, and HEK293T cells upon the treatment of Ru-trisBP, Ru-trisBP@DOPC, Ru-trisBP@OA, and RuOA2 at a concentration of 1 μM for 4 h through ICP-MS analysis. Mean ± SD ( n = 3). (g) HCT116, CT26, and HEK 293T cell viabilities upon the treatment of RuOA2 at various concentrations for 24 h. Mean ± SD ( n = 8). (h) HCT116 cell viability upon the treatment of Ru-trisBP, Ru-trisBP@DOPC, Ru-trisBP@OA, and RuOA2 at a concentration of 400 nM for 24 h. Mean ± SD ( n = 8). As summarized in Figure 3 e, RuOA2 enters the cell through a membrane-fusion-like process, followed by the release of RuOA2 into the cytoplasma. These assemblies, driven by electrostatic attraction, accumulate at the mitochondria, triggering stress-induced morphological changes. The efficiency of cellular delivery was quantified via ICP-MS analysis ( Figure 3 f), showing that ruthenosomes enhanced the uptake of Ru-trisBP by nearly 2000-fold and increased uptake 6 times compared to Ru-trisBP@DOPC liposomes. The significant improvement in Ru-trisBP cellular delivery by ruthenosomes, more than 600-fold compared to Ru-trisBP@OA, highlights the advantages of RuOA2 ’s molecular design. Moreover, the high selectivity for mitochondrial targeting emphasizes the precision and effectiveness of this delivery system. In contrast, traditional liposome-mediated delivery of Ru-trisBP failed to induce cytotoxicity in HCT116 cells ( Figure 3 h). To further evaluate the impact of RuOA2 ’s molecular design on cancer cell selectivity, we measured the intracellular ruthenium content in CT26, HCT116, and HEK293T cells after RuOA2 treatment using ICP-MS. The results revealed a significantly higher uptake of RuOA2 in both CRC cell lines compared to HEK293T cells ( Figure 3 f, S6), confirming that the oleate modification enhanced cancer cell targeting. This was further supported by our cytotoxicity assays, which showed minimal toxicity to normal cells at nanomolar concentration of RuOA2 , while demonstrating potent cytotoxicity against CRC cell lines ( Figure 3 g). To access whether RuOA2 ruthenosomes effectively induce ferroptosis via mitochondrial targeting, we evaluated two hallmarks: increased levels of labile ferrous ion (Fe 2+ ) levels and lipid peroxidation. Intracellular Fe 2+ levels were measured using FeRhoNox-1, a nonfluorescent probe that specifically reacts with Fe 2+ to emit fluorescence. Compared to untreated HCT116 cells and those treated with Ru-trisBP@DOPC or Ru-trisBP@OA, which showed minimal fluorescence, RuOA2 -treated cells exhibited a marked increase in fluorescence, indicating elevated Fe 2+ levels ( Figure 4 a and Figure S7 ). Lipid peroxidation was detected using the C11-BODIPY probe, 29 where RuOA2 treatment caused a fluorescence shift from red to green, indicating lipid peroxidation ( Figure 4 b and Figure S8 ). To further investigate RuOA2 -induced ferroptosis, RNA sequencing (RNA-seq) was performed on HCT116 cells treated with 300 nM RuOA2 for 24 h. The analysis revealed insights into multiple cellular processes, including ferroptosis, endocytosis, and autophagy ( Figure S9 ). Figure 4 RuOA2 stimulates ferritinophagy-mediated cancer suppression. (a) Fluorescence images of HCT116 cells with and without the treatments of Ru-complexes (300 nM, 24 h) stained with FeRhoNox-1. Scale bars represent 20 μm. (b) Confocal fluorescence images of HCT116 cells incubated with C11-BODIPY after treatment with Ru-complexes at a concentration of 300 nM for 24 h. The fluorescence transition from red to green indicated significant lipid peroxidation. Scale bars represent 20 μm. (c) Cellular component (CC) of GO enrichment analysis of differentially expressed genes in HCT116 cells treated with RuOA2 (300 nM) for 24 h compared with the control. Bubble diagram was made based on the genes shown the top 30 CCs with high gene number. Bubble size represents the number of genes, while color represents the rich factor of gene expression. (d) Immunoblots for GPX4, NCOA4, FTH1, ATG5, ATG7, and LC3B I/II expression in HCT116 cells upon the treatment of Ru-complexes at a concentration of 300 nM for 24 h. β-Actin serves as the loading control. Immunoblots repeated in triplicate with similar results. (e) Relative intracellular GSH level in HCT116 cells upon treatment with Ru-trisBP@DOPC, Ru-trisBP@OA, and RuOA2 at a concentration of 300 nM for 24 h. Mean ± SD ( n = 3). (f) Ferrous ion content in HCT116 cells transfected with control vector or FTH1-overexpression encoding plasmids with or without the treatment of 300 nM RuOA2 for 24 h. Mean ± SD ( n = 3). (g) Viability of HCT116 cells transfected with control vector or FTH1-overexpression encoding plasmids with or without the treatment of 300 nM RuOA2 for 24 h. Mean ± SD ( n = 6). (h) TEM images (false color) of RuOA2 (1 μM, 4 h) treated HCT116 cells. The mitochondria are highlighted in pink, the phagophores are highlighted in light blue, the autophagosomes are highlighted in green, and the autolysosomes is highlighted in yellow. Scale bars represent 500 nm. (i) HCT116 cell viability upon the treatment of RuOA2 at a concentration of 400 nM for 24 h after the pretreatment of different PCD inhibitors. Mean ± SD ( n = 8). ns: no significant, *** p < 0.001, **** p < 0.0001. The Gene Ontology (GO) analysis of cellular components (CC) highlighted significant enrichment in mitochondria, peroxisome, lysosome, and autophagosome ( Figure 4 c), underscoring the critical roles of mitochondrial dysfunction and autophagic processes in the cellular response to RuOA2 . Specifically, gene expression analysis revealed distinct trends in the regulation of redox homeostasis, iron metabolism, lipid metabolism, and autophagy ( Figure S10 ), suggesting that RuOA2 treatment promotes iron-dependent lipid peroxidation, autophagy, and mitochondrial dysfunction. Notably, RuOA2 induced the downregulation of GPX4 ( 6 ) and NQO1, ( 30 ) two key regulators of redox balance. The reduction of GPX4 is particularly important, as its loss leads to excessive ROS accumulation, a hallmark of ferroptotic cell death. Additionally, the downregulation of CISD1 and NFS1 , 31 which regulate mitochondrial iron–sulfur cluster homeostasis, further supports the disruption of mitochondrial function and the induction of redox imbalance. The upregulation of the TFRC gene, 32 which governs cellular iron uptake via receptor-mediated endocytosis, indicates enhanced iron acquisition, contributing to ferroptosis initiation. Furthermore, ATF3 ( 33 ) and ACSL4 ( 34 )-key genes involved in lipid metabolism and the execution of ferroptosis-were upregulated, while SCD1, ( 35 ) a gene that regulates lipid homeostasis, was downregulated, highlighting an increase in lipid peroxidation and greater sensitivity to ferroptotic cell death. RNA-seq analysis also revealed significant upregulation of autophagy-related genes such as ATG5 and ATG7 , 36 indicating autophagic activation in RuOA2 -treated cells. The upregulation of HMOX1 further supports autophagy’s role in regulating cellular iron levels. Moreover, the downregulation of ferritin heavy chain 1 ( FTH1 ), 37 responsible for iron storage, combined with the upregulation of NCOA4 , 38 , 39 a key mediator of ferritin degradation via autophagy, suggests that RuOA2 induces ferritinophagy-a selective form of autophagy that releases stored iron. Collectively, these findings implicate ferritinophagy as a crucial mechanism in RuOA2 -induced ferroptosis ( Figure S11 ). The expression levels of key regulatory proteins involved in ferritinophagy were analyzed via Western blotting ( Figure 4 d). Compared to untreated HCT116 cells and those treated with Ru-trisBP@DOPC or Ru-trisBP@OA, RuOA2 treatment led to a 60% reduction in GPX4 expression, indicating elevated ROS production. To directly assess ROS levels, the DCFH-A assay was conducted, revealing that RuOA2 treatment significantly enhanced ROS production compared to both Ru-trisBP@DOPC and Ru-trisBP@OA ( Figure S12 and S13 ). Additionally, intracellular levels of the antioxidant GSH were also measured, and RuOA2 treatment resulted in a nearly 70% decrease in GSH levels, shifting the redox balance toward a more oxidative state ( Figure 4 e). To further evaluate mitochondrial dysfunction associated with oxidative stress, JC-1 staining was performed to access mitochondrial membrane potential (MMP). 26 The data revealed a significant reduction in MMP following RuOA2 treatment ( Figure S14 ). In comparison with cisplatin (CDDP), an established anticancer drug known to induce ROS-mediated cancer cell death, RuOA2 was markedly more effective at inducing mtROS ( Figure S15 ), underscoring its potency as a robust ROS inducer capable of triggering ferroptosis. Consistent with the RNA-seq analysis, treatment with RuOA2 resulted in a 60% increase in NCOA4 and a 30% decrease in FTH1, along with a 70% increase in ATG5 and a 60% increase in ATG7 expression. Additionally, immunoblotting revealed the conversion of LC3B from its LC3B–I to LC3B–II form. The increase in LC3B–II serves as a reliable indicator of autophagosome formation, 40 , 41 and we observed a 3.5-fold increase in the LC3B–II/LC3B–I ratio following RuOA2 treatment. Autophagosome formation was further confirmed using TEM imaging ( Figure 4 h). 42 , 43 RuOA2 treatment induced notable swelling of mitochondria with rounded cristae (highlighted in pink), indicative of reduced MMP due to increased calcium ion influx. The process of autophagy, a conserved transport pathway, involves the sequestration of targeted structures through the formation of phagophores, which mature into autophagosomes and are subsequently delivered to lysosomes for degradation. The TEM images presented in Figure 4 h show membrane expansion and sealing of phagophores (highlighted in blue), the formation of autophagosomes (highlighted in green), and the presence of autolysosomes (highlighted in yellow). Collectively, these findings suggest that RuOA2 induces ferritinophagy in HCT116 cells. To confirm that RuOA2 induces cell death through the ferroptosis pathway, we investigated the effect of overexpressing FTH1 ( Figure 4 f, 4 g and S16) and NCOA4 knockout ( Figure S17 ) in HCT116 cells treated with RuOA2 . In control cells, FTH1 overexpression had no significant impact on intracellular labile Fe 2+ levels or cell viability. However, in RuOA2 -treated cells, FTH1 overexpression reduced Fe 2+ levels by nearly 50%, leading to a restoration of cell viability. Additionally, NCOA4 knockout alleviated RuOA2 -induced cell death and Fe 2+ accumulation. These findings collectively confirmed the critical role of NOCA4-mediated ferritin degradation in regulating RuOA2 -induced ferroptosis. To further explore the involvement of autophagy in RuOA2 -induced ferroptosis, we evaluated the effects of various cell death inhibitors on RuOA2 -treated HCT116 cells. As shown in Figure 4 i, cell viability was 40.33% following RuOA2 treatment alone, remaining nearly unchanged with the pan-caspase and apoptosis inhibitor Z-VAD-FMK 44 (40.66%) or the lipid hydroperoxide scavenger Fer-1 (46.11%). In contrast, treatment with the autophagy inhibitor CQ, 45 which disrupts autophagosome-lysosome fusion, and the iron chelator DFO, 46 which suppresses ROS accumulation to inhibit ferroptosis, 47 significantly increased cell viability to 71.59% and 70.23%, respectively. These results confirm that ferritinophagy plays a critical role in RuOA2 -mediated cancer cell suppression. Furthermore, we examined HeLa, HepG2, MCF7, H1975, A549 and CT26 cells ( Figure S18 ), confirming that RuOA2 acts as a FIN and broadly induces ferritonophagy across various cancer cell types. To access the in vivo efficacy of RuOA2 ruthenosomes in suppressing CRC, a syngeneic model was established by subcutaneously implanting CT26 cells into C57BL/6 mice. The CT26 cell line, a murine colorectal cancer model, was selected for its compatibility with immunocompetent hosts, providing a more physiologically relevant context for assessing progression and therapeutic response. Prior to the syngeneic study, we confirmed that RuOA2 induced ferritinophagy-mediated cell death in CT26 cells ( Figure S19-27 ), consistent with findings from HCT116 human colorectal cancer cells. Mice were divided into four treatment groups: saline (control), Ru-trisBP@OA, CDDP, and RuOA2 , following the designated intratumoral injection regimen ( Figure 5 a). As shown in Figure 5 b, tumor growth progressed steadily in the saline and Ru-trisBP@OA groups, while RuOA2 treatment significantly inhibited tumor growth, outperforming CDDP in efficacy ( Figure 5 b, S28, ans S29). At the end of the 22-day study, tumors were excised and weighted ( Figure 5 c and S30), with tumor weights in the RuOA2 -treated group significantly lower than in the other groups, indicating effective tumor suppression. Histological analysis via H&E staining revealed extensive tumor cell death and loss of tissue integrity in the RuOA2 treatment group ( Figure S31 ), demonstrating pronounced anticancer activity. Figure 5 RuOA2 ruthenosome induces ferritinophagy-mediated CRC suppression in vivo . (a) Schematic representation of the injection regime in a CRC syngeneic model using C57BL/6 mice. (b) Tumor growth was monitored via caliber measurements before each injection, with tumor volumes calculated and plotted. Mean ± SD ( n = 6). (c) Tumor weight of mice across different treatment groups at the end of 22 days. Mean ± SD ( n = 6). * p < 0.05, *** p < 0.001. (d) Survival rate statistics of mice under various treatments. ( n = 6). (e) Representative images of immunofluorescence staining for Ki67, FTH1, and GPX4, along with FeRhoNox-1 staining for detecting Fe 2+ in tumor sections. Scale bars represent 100 μm. Additionally, all treatment formulations had minimal impact on body weight, indicating good tolerability ( Figure S32 ). Survival rates were monitored across treatment groups ( Figure 5 d), with the RuOA2 -treated group exhibiting a markedly higher survival rate of 80% at 40 days, compared to 30% in the CDDP-treated group. Furthermore, analysis of serum biochemical markers ( Figure S33 ) revealed that liver and kidney function remained within the healthy reference range, suggesting no detectable renal or hepatic toxicity in the RuOA2 -treated mice. We also evaluated the therapeutic potential of RuOA2 in the context of its pharmacokinetics and tissue distribution. To this end, we administered RuOA2 via intratumoral injection in C57BL/6 mice, which allowed for tumor-specific retention and demonstrated 33.5% of RuOA2 remained within the tumor after 48 h ( Figure S34 ). This confirmed efficient tumor targeting and minimal systemic exposure, addressing concerns of off-target delivery. To assess systemic clearance, we also performed tail vein injections at a dose of 5 mg/kg, followed by serum ruthenium level measurements using ICP-MS. These results demonstrated rapid clearance from the bloodstream, with a half-life of 0.84 h ( Figures S35 ), and initial accumulation in the liver and heart, which was substantially eliminated by 24 h ( Figures S36 ). Post-mortem analysis of major organs (heart, liver, spleen, lungs, and kidneys) via H&E staining ( Figure S37 ) revealed no signs of significant organ damage, further implying low systemic toxicity of RuOA2 . These findings, together with the rapid tumor suppression observed, highlight the potent antitumor efficacy and favorable biosafety profile of RuOA2. To further investigate the molecular mechanisms underlying tumor suppression, we performed immunofluorescence staining and Western blotting on tumor sections to assess the expression of the proliferation marker Ki67, the ferritin regulator FTH1, and the ferroptosis-related protein GPX4, alongside FeRhoNox-1 staining ( Figure 5 e, S38 and S39). The results revealed a significant reduction in Ki67 expression in the RuOA2 -treated tumors, indicating decreased tumor cell proliferation. Additionally, FTH1 and GPX4 levels were notably downregulated, accompanied by increased FeRhoNox-1 fluorescence, reflecting elevated levels of labile Fe 2+ . These findings support the conclusion that RuOA2 ruthenosomes effectively induce ferritinophagy-mediated tumor cell death in vivo , leading to more pronounced CRC suppression compared to CDDP. ## Conclusions Conclusions The development of novel liposome systems has garnered significant attention due to their efficiency in delivering therapeutic agents into cells. However, many current designs, such as functional liposomes (e.g., boronsomes and liposome-in-liposome systems), lack specificity for subcellular organelles. 48 In this study, we introduce a ruthenium-based liposomal system, termed ruthenosome, which represents the first ruthenium-based FIN that triggers ferroptotic tumor cell death without the need for light activation. Our design leverages the hierarchical self-assembly of RuOA2 into multiscale vesicle-to-vesicle structures due to strong hydrophobic interactions between the outward-facing oleate chains on their surfaces. This unique molecular architecture not only facilitates enhanced cellular uptake but also achieves precise mitochondrial targeting, a critical advantage over conventional liposomal systems. Ruthenosomes function as potent ROS generators, initiating ferroptosis and displaying broad applicability across various cancer types. The ability of ruthenosomes to selectively target mitochondria and induce ferroptosis without external stimuli marks a significant breakthrough in metallodrug design. The promising in vivo results, including superior tumor suppression and minimal toxicity, underscore the therapeutic potential of ruthenosomes in cancer treatment. This study highlights the versatility of ruthenosomes as multifunctional nanocarriers and paves the way for the development of the next generation of metallodrugs. Furthermore, while current formulations utilize direct tumor injection, we are actively exploring methods to optimize the size of ruthenosomes for potential intravenous administration. This future optimization could further enchance therapeutic efficacy, expand their application to a broader range of clinical settings, and minimize off-target effects, offering a transformative approach to targeted cancer therapies. By incorporating nanotechnology, future ruthenosome designs can continue to improve both therapeutic efficacy and safety profiles. ## Method Method Synthesis of RuOA2 To a suspension of oleic acid (1 mL) in anhydrous acetonitrile (10 mL), Ru-2 complex (103 mg, 0.1 mmol) was added, followed by HBTU (0.5 g) and DIC (0.2 mL). The reaction mixture was stirred and refluxed at 60 °C for 24 h, then allowed to cool to room temperature. After removing the precipitates by filtration, the solvent was evaporated under reduced pressure obtaining solid residue, which was resuspended in 10 mL of anhydrous acetonitrile. Any remaining insoluble material was filtered off. The crude product was purified by column chromatography on aluminum oxide to yield pure RuOA2 (92.2 mg) as a dark brown powder with a 59.0% yield. Details on the synthesis of Ru-2 are provided in the Supporting Information . Preparation of Ru-trisBP@DOPC and Ru-trisBP@OA Liposomes Ru-trisBP@DOPC liposomes were synthesized using a thin-film hydration method. Specifically, 27.64 mg of DOPC ( D68710 , Acmec, China) was dissolved in methanol, then evaporated at 40 °C for 30 min to form a uniform thin film. The dried lipid film was hydrated with 3 mL of a Ru-trisBP solution (0.5 mg/mL in water) and sonicated for 30 min. This mixture was then extruded through 0.45 μm pore-size polycarbonate membrane filters to produce Ru-trisBP@DOPC liposomes. Unencapsulated Ru-trisBP was removed by chromatography on a Sephadex G10 column. The final concentration of Ru-trisBP encapsulted in Ru-trisBP@DOPC liposomes was determined to be 1.98 mM, with a DOPC to Ru-trisBP molar ratio of 17.86:1, and a loading efficacy of 84%. Ru-trisBP@OA liposomes were prepared by mixing Ru-trisBP and oleic acid (OA) at a 1:2 molar ratio in deionized water to mimic the Ru-trisBP/OA ratio found in the structure of RuOA2 . The mixture was then subjected to continuous sonication for 30 min, facilitating coassembly and yielding Ru-trisBP@OA liposomes. Preparation of RuOA2 Ruthenosomes RuOA2 was first dissolved in dimethyl sulfoxide (DMSO) and then diluted at a 1:1000 ratio with deionized water or 1XPBS. The resulting solution was sonicated for 30 min to obtain ruthenosomes. Simulation The conformational optimization of RuOA2 was conducted using ORCA software with density functional theory (DFT), employing the B3LYP functional and the def2-SVP basis set. Molecular dynamics simulations were then carried out to model the self-assembly of RuOA2 . A supercell was constructed with 8 unit cells along both the x and y directions. Each unit cell contained two RuOA2 molecules arranged in an up-and-down mirror plane orientation, with the Ru atoms end facing outward. The initial dimensions of the supercell were 144 Å x 160 Å x 46 Å in the x, y, and z directions, respectively. Due to the limited availability of force field parameters for PF 6 – ions, Cl – ions were used as substitutes. The supercell was filled with 3200 water molecules and 256 Cl – ions using Packmol software. Molecular dynamics simulations were performed using LAMMPS software with the PVFF force field, applying Lennard-Jones potential with a cutoff distance of 10 Å. Bond, angle, dihedral, and improper interactions were defined using class2 force field types. The NVT ensembles was employed with a temperature of 300 K, allowing the simulation box to adjust its size dynamically. Over the course of the simulation, the box size decreased, ultimately reaching final dimensions of 70 Å x 80 Å x 47 Å in the x, y, and z directions, respectively. TEM Imaging of Liposomes and Ruthenosomes Following the preparation protocols detailed above, Ru-trisBP@DOPC liposomes (10 μM Ru-trisBP), Ru-trisBP@OA (10 μM), and RuOA2 ruthenosomes (10 μM) were prepared. After allowing the samples to stabilize at room temperature for 2 h, 10 μL aliquots of sample solution were deposited onto glow-discharge copper grids (400 mesh) coated with a thin carbon film. The samples were left on the grid for 30 s before excess solution was removed. The grids were then rinsed with deionized water up to three times. TEM images were acquired under high vacuum using a Talos-L120C transmission electron microscope (Thermo Fisher). The dynamic morphological transitions of ruthenosomes were captured with a multipurpose electron microscope (JEM-F200, JEOL, Japan). Cell Culture HCT116, CT26, and PANC-1 were obtained from the Cell Bank of the Chinese Academy of Sciences. HeLa, HepG2, MCF7, H1975 and A549 were generously provided by the Huadong Liu’s lab at Xi’an Jiaotong University. PANC-1, HeLa, HepG2, MCF7, and A549 cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM), while HCT116 cells were maintained in McCoy’s 5a medium. H1975 and CT26 cells were cultured in RPMI 1640 medium. All culture media were supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. Cells were incubated at 37 °C in a humidified atmosphere with 5% CO 2 and passaged every 3 days using a 0.25% trypsin-EDTA solution. Cell Viability Assay Cells in the exponential growth phase were seeded into a 96-well cell culture plate at a density of 5 × 10 3 cells per well and incubated for 24 h at 37 °C in 5% CO 2 atmosphere. Molecules, liposomes and ruthenosomes were added to the wells at the specific concentrations. After the desired incubation period, 10 μL of MTT solution (5 mg/mL) was added in each well. The plates were then incubated at 37 °C for 4 h. Subsequently, the MTT-containing supernatant was removed, and 100 μL of DMSO was added to dissolve the formazan crystals formed by viable cells. The resulting purple solution was measured for optical density at 490 nm using a SuPerMax 3100 plate reader (Shanpu, China). Intracellular Uptake Assessment CT26,HCT116 and HEK 293T cells were seeded in 6-well plates at a density of 5 × 10 4 cells per well and incubated overnight. Ru-trisBP@DOPC (1 μM Ru-trisBP), Ru-trisBP@OA (1 μM), or RuOA2 (1 μM) was added, and the cells were incubated at 37 °C for 4 h. After incubation, cells were washed with PBS, and then harvested, lyophilized and subsequently subjected to microwave digestion. The resulting residue was reconstituted in deionized water to a final volume of 10 mL, generating the test solution. A specified volume of this solution was then analyzed for Ru content using the Inductively Coupled Plasma Mass Spectrometry (ICP-MS) (PerkinElmer NexION 300X). RNA Sequencing and Bioinformatics Analysis HCT116 cells were seeded in 6-well plates at a density of 5 × 10 4 cells per well and treated with 300 nM RuOA2 for 24 h. Following treatment, cells were harvested, and total RNA was extracted using TRNzol Universal Reagent (DP424, Tiangen, China). RNA-seq was performed by Lc-Bio Technologies. Genes with false discovery rates (FDR) < 0.05 and lengths >200 bp were considered to show differential expression. Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) enrichment analyses were conducted using the DAVID online database. Gene Set Enrichment Analysis (GSEA) followed the standard procedures outlined in the GSEA User Guide ( http://www.broadinstitute.org/gsea/doc/GSEAUserGuideFrame.html ). Intracellular Glutathione (GSH) Quantification Assay HCT116 cells were seeded in 6-well plates at a density of 5 × 10 4 cells per well and incubation overnight. Cells were then treated with Ru-trisBP@DOPC (300 nM Ru-trisBP), Ru-trisBP@OA (300 nM) or RuOA2 ruthenosome (300 nM) for 24 h. The intracellular GSH levels were quantified using a GSH Assay Kit (A006–2–1, Njjcbio, China). For the calibration curve, 1 mM GSH standard solutions were diluted to prepare solutions at various concentrations (0, 5,10, 20, 50, and 100 μM). After treatment, cells were harvested, and colorimetric quantitative of GSH was performed at 405 nm. Confocal Microscopy and Intracellular Localization Analysis Cells in the exponential growth phase were seeded into 35 mm glass bottom dishes at a density of 2 × 10 4 cells per dish. Once fully adhered, the culture medium was replaced with fresh medium containing various concentrations of liposomes, ruthenosomes, or molecules. After the desired incubation time, cells were washed three times with PBS and subsequently stained with cell labeling dyes according to the manufacturer’s protocols. Fluorescent images were acquired using an Olympus FV3000 confocal laser scanning microscope. For intracellular localization studies, the following stains were used: Nuclear Staining Hoechst 33258 (1 μg/mL; H4047, UElandy, China) or DAPI (1 μg/mL; C1002, Beyotime, China) with excitation/emission at 405/460 ± 20 nm. Mitochondrial Staining MitoTracker Green (50 nM; A66441, Thermo Fisher Scientific, USA) with excitation/emission at 488/516 ± 10 nm. Cell Membrane Staining CellMask Deep Red stain (50 nM; A57245 , Thermo Fisher Scientific) with excitation/emission at 640/670 ± 20 nm. Lysosome Staining Lyso Tracker Green (50 nM, L7526, Thermo Fisher Scientific) with excitation/emission at 488/511 ± 10 nm. Mitochondrial Membrane Potential Assessment The mitochondrial membrane potential was measured using the JC-1 dye (C2006, Beyotime) following the manufacturer’s protocol. Briefly, cells were incubated with a 1 × JC-1 working solution at 37 °C for 30 min. Fluorescence settings were configured with excitation/emission at 488/535 ± 20 nm for JC-1 monomers and 488/570 ± 20 nm for JC-1 aggregates, to differentiate between polarized and depolarized mitochondrial states. ROS, Lipid Peroxidation (LPO), and Fe 2+ Quantification Assay To detect reactive oxygen species (ROS), 10 μM DCFH-DA (S0033S, Beyotime) was used with fluorescence settings of Ex/Em 488/525 ± 20 nm. For lipid peroxidation (LPO) detection, 5 μM C11 BODIPY (D3861, Thermo Fisher Scientific) was employed, with Ex/Em 561/591 ± 10 nm (red channel) and Ex/Em 488/510 ± 10 nm (green channel). Fe 2+ levels were assessed using 5 μM FeRhoNox-1 (MX4558, MKBio, China) with Ex/Em settings at 561/580 ± 10 nm. Images were processed and analyzed with ImageJ software. For quantitative analysis, background subtraction and normalization were applied, and the integrated density of fluorescence signals was measured using ImageJ. TEM Imaging of Cells HCT116 cells were cultured in 10 cm dishes for 24 h and treated with 300 nM RuOA2 for an additional 24 h. Cells were harvested by trypsinization and fixed in 2.5% glutaraldehyde, followed by postfixation with 1.0% osmium tetroxide. Samples were then embedded in epoxy resin and sectioned into 70 nm slices. TEM images were captured using a Hitachi H-7650 TEM at an acceleration voltage of 80 kV. Overexpression of FTH1 in HCT116 and CT26 Cells HCT116 and CT26 cells were seeded in 6-well plates and allowed to grow to 70–80% confluence. The FTH1-GFP expression plasmid was synthesized by Sangon Biotech. For each well, 2 μg of plasmid DNA was diluted in 125 μL of Opti-MEM I Reduced Serum Medium (Gibco). Separately, 5 μL of Lipo 6000 Transfection Reagent (C0526FT, Beyotime, China) was diluted in 125 μL of Opti-MEM I. After a 5 min incubation at room temperature, the diluted DNA and Lipo 6000 were mixed and incubated for 20 min to allow the formation of DNA-lipid complexes. The mixture was then added dropwise to each well containing cells in serum-free medium and incubated for 6 h at 37 °C in a CO 2 incubator. Following this, the transfection medium was replaced with fresh complete growth medium, and the cells were incubated for an additional 24–48 h prior to downstream analysis. Knockout of NCOA4 in HCT116 Cells NCOA4 CRISPR-Cas9 gene disruption in HCT116 cells was performed with the LentiCRISPR V2 vector (sg- NCOA4 ) using the following oligonucleotide sequence: 5′-ggaatgtcttagaagccgtg-3′. sg- NCOA4 and negative control (sg-NC) lentivirus (10 9 TU/ml) was obtained from Tsingke, China. 200 μL lentivirus transfection mixture (8 μL lentivirus, 8 μL HitransG P infection reagent, 184 μL DMEM) was added to 10 4 cells seeded in a 24-well plate. After coculturing for 24 h, replaced with fresh medium. Stably transduced cells were obtained by puromycin selection (2 μg/mL) Western Blot Analysis Cells and tumor tissues were lysed in cell lysis buffer with protease inhibitors (P0013, Beyotime), followed by centrifugation at 12000 rpm for 10 min at 4 °C. The supernatant was collected, and protein concentrations were determined using a BCA Quantitation Kit (23228, Thermo Fisher Scientific). Proteins were then mixed with SDS sample buffer containing 10% β-mercaptoethanol, separated by SDS-PAGE, and transferred to nitrocellulose membranes for immunoblotting. β-Actin was used as the loading control. Primary antibodies included: anti-β-Actin (66009–1-Ig, Proteintech, 1:5000), anti-LC3B I/II (ET1701–65, HuaBio, 1:1000), anti-ATG5 (ET1611–38, HuaBio, 1:1000), anti-ATG7 (ET1610–53, HuaBio, 1:1000), anti-NCOA4 (ER62707, HuaBio, 1:2000); anti-FTH1 ( T55648 , Abmart, 1:1000); anti-GPX4 (T56959S, Abmart, 1:1000). Blots were developed using an enhanced chemiluminescence (ECL) detection system (1705061, Bio-Rad, USA), and signal were visualized on a Chemiluminescence imaging analysis system (BG-gdsAUTO 720, Baygene, China). Band intensities were quantified using ImageJ software. Immunohistochemistry Tumor tissues were collected and embedded in Optimal Cutting Temperature (OCT) compound. OCT-embedded tissues were sectioned at a thickness of 5 μm and mounted onto glass slides. Sections were fixed in 4% paraformaldehyde for 20 min, permeabilized with 0.1% Triton X-100 for 1 h, and blocked with 5% bovine serum albumin (BSA) at room temperature for 30 min. The sections were then incubated overnight at 4 °C with primary antibodies against FTH1 (1:500), GPX4 (1:500), and Ki67 (PT0321R, Immunoway, 1:500), followed by incubation with a Cy3-conjugated goat antirabbit IgG secondary antibody (SA00009–2, Proteintech, 1:100) for 1 h at room temperature. Nuclear staining was performed using 4′,6-diamidino-2-phenylindole (DAPI, P0131, Beyotime) for 30 min. Confocal images were captured on an Olympus FV3000 laser scanning microscope. Animal Studies All animal experiments were approved by the Animal Ethics Committee of Xi’an Jiaotong University (No. 2021–1123). Female C57BL/6 mice (6–8 weeks) were obtained from the Experimental Animal Center of Xi’an Jiaotong University and housed under specific pathogen-free conditions with a 12-h light/dark cycle and free access to food and water. Six days after tumor cell inoculation, when tumor signals became detectable by in vivo imaging, mice were randomized into four groups and peritumorally injected every other day with a 5 mg/kg dose of liposome, ruthenosomes, or drug molecules. On day 22, 2 days after the last treatment, tumor tissues were collected for hematoxylin and eosin (H&E) staining to evaluate antitumor effects. Subcutaneous CRC Tumor Implantation To generate murine subcutaneous tumors, 5 × 10 6 CT26 cells stably expressing the GFP-Luc fusion gene (CT26-luciferase cells) were injected subcutaneously into female C57BL/6 mice aged 8 weeks. In Vivo Bioluminescence Imaging On day 6 postinoculation, detectable tumor signals were imaged using the in vivo imaging system. Mice were divided into four groups (Saline, Ru-trisBP@OA, CDDP, and RuOA2). Bioluminescence images were captured on days 6, 14, and 22 following an intraperitoneal injection of D-Luciferin potassium salt (150 mg/kg) using the VISQUE In VIVO Smart-LF imaging system. Therapeutic Evaluation Tumor growth was monitored by measuring tumor weight, length (L), and width (W). Tumor volume was calculated with the formula: 1/2(L × W 2 ). Mice body weights were recorded every 2 days to monitor health. At the end of the treatment period, mice were euthanized, and tumor tissues were stained with H&E for morphological evaluation. Survival time was defined as the duration from tumor implantation until sacrifice or when tumor volume reached 1000 mm 3 , with a maximum follow-up of 45 days. In Vivo Pharmacokinetics of Ruthenosomes RuOA2 was administered to the C57BL/6 mice via intravenous injection at a dose of 5 mg/kg. Blood samples were collected at 0.5, 1, 2, 4, 12, 24, 48 h postinjection. Collected blood samples were incubated at 37 °C for 30 min and then centrifuged at 3000 rpm for 15 min to harvest the serum. The serum was then transferred to clean polypropylene tubes for further analysis. Concentration of RuOA2 in serum samples was determined using ICP-MS. Statistical analysis of the pharmacokinetic parameters was performed by GraphPad Prism software. Data were expressed as mean ± standard deviation (SD) for each group. Ex Vivo Fluorescence Imaging Following intravenous injection of 5 mg/kg RuOA2 at 6 and 24 h, heart, lung, liver, spleen, and kidney were excised from C57BL/6 mice. The tissues were then rinsed with cold 1 x PBS to remove residual blood and debris. Fluorescent images were captured using the VISQUE In VIVO Smart-LF imaging system. Statistics and Reproducibility Data were analyzed using GraphPad Prism (version 9.0) and are presented as the mean ± standard deviation (SD) with n ≥ 3 unless specified otherwise. Unpaired Student’s t tests were used for comparisons the between two groups, and one-way analysis of variance (ANOVA) was used for multiple group. A p-value of less than 0.05 was considered statistically significant. Statistical significance is indicated by asterisks: p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001. ## Synthesis of Synthesis of RuOA2 To a suspension of oleic acid (1 mL) in anhydrous acetonitrile (10 mL), Ru-2 complex (103 mg, 0.1 mmol) was added, followed by HBTU (0.5 g) and DIC (0.2 mL). The reaction mixture was stirred and refluxed at 60 °C for 24 h, then allowed to cool to room temperature. After removing the precipitates by filtration, the solvent was evaporated under reduced pressure obtaining solid residue, which was resuspended in 10 mL of anhydrous acetonitrile. Any remaining insoluble material was filtered off. The crude product was purified by column chromatography on aluminum oxide to yield pure RuOA2 (92.2 mg) as a dark brown powder with a 59.0% yield. Details on the synthesis of Ru-2 are provided in the Supporting Information . ## Preparation of Ru-trisBP@DOPC and Ru-trisBP@OA Liposomes Preparation of Ru-trisBP@DOPC and Ru-trisBP@OA Liposomes Ru-trisBP@DOPC liposomes were synthesized using a thin-film hydration method. Specifically, 27.64 mg of DOPC ( D68710 , Acmec, China) was dissolved in methanol, then evaporated at 40 °C for 30 min to form a uniform thin film. The dried lipid film was hydrated with 3 mL of a Ru-trisBP solution (0.5 mg/mL in water) and sonicated for 30 min. This mixture was then extruded through 0.45 μm pore-size polycarbonate membrane filters to produce Ru-trisBP@DOPC liposomes. Unencapsulated Ru-trisBP was removed by chromatography on a Sephadex G10 column. The final concentration of Ru-trisBP encapsulted in Ru-trisBP@DOPC liposomes was determined to be 1.98 mM, with a DOPC to Ru-trisBP molar ratio of 17.86:1, and a loading efficacy of 84%. Ru-trisBP@OA liposomes were prepared by mixing Ru-trisBP and oleic acid (OA) at a 1:2 molar ratio in deionized water to mimic the Ru-trisBP/OA ratio found in the structure of RuOA2 . The mixture was then subjected to continuous sonication for 30 min, facilitating coassembly and yielding Ru-trisBP@OA liposomes. ## Preparation of Preparation of RuOA2 Ruthenosomes RuOA2 was first dissolved in dimethyl sulfoxide (DMSO) and then diluted at a 1:1000 ratio with deionized water or 1XPBS. The resulting solution was sonicated for 30 min to obtain ruthenosomes. ## Simulation Simulation The conformational optimization of RuOA2 was conducted using ORCA software with density functional theory (DFT), employing the B3LYP functional and the def2-SVP basis set. Molecular dynamics simulations were then carried out to model the self-assembly of RuOA2 . A supercell was constructed with 8 unit cells along both the x and y directions. Each unit cell contained two RuOA2 molecules arranged in an up-and-down mirror plane orientation, with the Ru atoms end facing outward. The initial dimensions of the supercell were 144 Å x 160 Å x 46 Å in the x, y, and z directions, respectively. Due to the limited availability of force field parameters for PF 6 – ions, Cl – ions were used as substitutes. The supercell was filled with 3200 water molecules and 256 Cl – ions using Packmol software. Molecular dynamics simulations were performed using LAMMPS software with the PVFF force field, applying Lennard-Jones potential with a cutoff distance of 10 Å. Bond, angle, dihedral, and improper interactions were defined using class2 force field types. The NVT ensembles was employed with a temperature of 300 K, allowing the simulation box to adjust its size dynamically. Over the course of the simulation, the box size decreased, ultimately reaching final dimensions of 70 Å x 80 Å x 47 Å in the x, y, and z directions, respectively. ## TEM Imaging of Liposomes and Ruthenosomes TEM Imaging of Liposomes and Ruthenosomes Following the preparation protocols detailed above, Ru-trisBP@DOPC liposomes (10 μM Ru-trisBP), Ru-trisBP@OA (10 μM), and RuOA2 ruthenosomes (10 μM) were prepared. After allowing the samples to stabilize at room temperature for 2 h, 10 μL aliquots of sample solution were deposited onto glow-discharge copper grids (400 mesh) coated with a thin carbon film. The samples were left on the grid for 30 s before excess solution was removed. The grids were then rinsed with deionized water up to three times. TEM images were acquired under high vacuum using a Talos-L120C transmission electron microscope (Thermo Fisher). The dynamic morphological transitions of ruthenosomes were captured with a multipurpose electron microscope (JEM-F200, JEOL, Japan). ## Cell Culture Cell Culture HCT116, CT26, and PANC-1 were obtained from the Cell Bank of the Chinese Academy of Sciences. HeLa, HepG2, MCF7, H1975 and A549 were generously provided by the Huadong Liu’s lab at Xi’an Jiaotong University. PANC-1, HeLa, HepG2, MCF7, and A549 cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM), while HCT116 cells were maintained in McCoy’s 5a medium. H1975 and CT26 cells were cultured in RPMI 1640 medium. All culture media were supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. Cells were incubated at 37 °C in a humidified atmosphere with 5% CO 2 and passaged every 3 days using a 0.25% trypsin-EDTA solution. ## Cell Viability Assay Cell Viability Assay Cells in the exponential growth phase were seeded into a 96-well cell culture plate at a density of 5 × 10 3 cells per well and incubated for 24 h at 37 °C in 5% CO 2 atmosphere. Molecules, liposomes and ruthenosomes were added to the wells at the specific concentrations. After the desired incubation period, 10 μL of MTT solution (5 mg/mL) was added in each well. The plates were then incubated at 37 °C for 4 h. Subsequently, the MTT-containing supernatant was removed, and 100 μL of DMSO was added to dissolve the formazan crystals formed by viable cells. The resulting purple solution was measured for optical density at 490 nm using a SuPerMax 3100 plate reader (Shanpu, China). ## Intracellular Uptake Assessment Intracellular Uptake Assessment CT26,HCT116 and HEK 293T cells were seeded in 6-well plates at a density of 5 × 10 4 cells per well and incubated overnight. Ru-trisBP@DOPC (1 μM Ru-trisBP), Ru-trisBP@OA (1 μM), or RuOA2 (1 μM) was added, and the cells were incubated at 37 °C for 4 h. After incubation, cells were washed with PBS, and then harvested, lyophilized and subsequently subjected to microwave digestion. The resulting residue was reconstituted in deionized water to a final volume of 10 mL, generating the test solution. A specified volume of this solution was then analyzed for Ru content using the Inductively Coupled Plasma Mass Spectrometry (ICP-MS) (PerkinElmer NexION 300X). ## RNA Sequencing and Bioinformatics Analysis RNA Sequencing and Bioinformatics Analysis HCT116 cells were seeded in 6-well plates at a density of 5 × 10 4 cells per well and treated with 300 nM RuOA2 for 24 h. Following treatment, cells were harvested, and total RNA was extracted using TRNzol Universal Reagent (DP424, Tiangen, China). RNA-seq was performed by Lc-Bio Technologies. Genes with false discovery rates (FDR) < 0.05 and lengths >200 bp were considered to show differential expression. Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) enrichment analyses were conducted using the DAVID online database. Gene Set Enrichment Analysis (GSEA) followed the standard procedures outlined in the GSEA User Guide ( http://www.broadinstitute.org/gsea/doc/GSEAUserGuideFrame.html ). ## Intracellular Glutathione (GSH) Quantification Assay Intracellular Glutathione (GSH) Quantification Assay HCT116 cells were seeded in 6-well plates at a density of 5 × 10 4 cells per well and incubation overnight. Cells were then treated with Ru-trisBP@DOPC (300 nM Ru-trisBP), Ru-trisBP@OA (300 nM) or RuOA2 ruthenosome (300 nM) for 24 h. The intracellular GSH levels were quantified using a GSH Assay Kit (A006–2–1, Njjcbio, China). For the calibration curve, 1 mM GSH standard solutions were diluted to prepare solutions at various concentrations (0, 5,10, 20, 50, and 100 μM). After treatment, cells were harvested, and colorimetric quantitative of GSH was performed at 405 nm. ## Confocal Microscopy and Intracellular Localization Analysis Confocal Microscopy and Intracellular Localization Analysis Cells in the exponential growth phase were seeded into 35 mm glass bottom dishes at a density of 2 × 10 4 cells per dish. Once fully adhered, the culture medium was replaced with fresh medium containing various concentrations of liposomes, ruthenosomes, or molecules. After the desired incubation time, cells were washed three times with PBS and subsequently stained with cell labeling dyes according to the manufacturer’s protocols. Fluorescent images were acquired using an Olympus FV3000 confocal laser scanning microscope. For intracellular localization studies, the following stains were used: Nuclear Staining Hoechst 33258 (1 μg/mL; H4047, UElandy, China) or DAPI (1 μg/mL; C1002, Beyotime, China) with excitation/emission at 405/460 ± 20 nm. Mitochondrial Staining MitoTracker Green (50 nM; A66441, Thermo Fisher Scientific, USA) with excitation/emission at 488/516 ± 10 nm. Cell Membrane Staining CellMask Deep Red stain (50 nM; A57245 , Thermo Fisher Scientific) with excitation/emission at 640/670 ± 20 nm. Lysosome Staining Lyso Tracker Green (50 nM, L7526, Thermo Fisher Scientific) with excitation/emission at 488/511 ± 10 nm. ## Nuclear Staining Nuclear Staining Hoechst 33258 (1 μg/mL; H4047, UElandy, China) or DAPI (1 μg/mL; C1002, Beyotime, China) with excitation/emission at 405/460 ± 20 nm. ## Mitochondrial Staining Mitochondrial Staining MitoTracker Green (50 nM; A66441, Thermo Fisher Scientific, USA) with excitation/emission at 488/516 ± 10 nm. ## Cell Membrane Staining Cell Membrane Staining CellMask Deep Red stain (50 nM; A57245 , Thermo Fisher Scientific) with excitation/emission at 640/670 ± 20 nm. ## Lysosome Staining Lysosome Staining Lyso Tracker Green (50 nM, L7526, Thermo Fisher Scientific) with excitation/emission at 488/511 ± 10 nm. ## Mitochondrial Membrane Potential Assessment Mitochondrial Membrane Potential Assessment The mitochondrial membrane potential was measured using the JC-1 dye (C2006, Beyotime) following the manufacturer’s protocol. Briefly, cells were incubated with a 1 × JC-1 working solution at 37 °C for 30 min. Fluorescence settings were configured with excitation/emission at 488/535 ± 20 nm for JC-1 monomers and 488/570 ± 20 nm for JC-1 aggregates, to differentiate between polarized and depolarized mitochondrial states. ## ROS, Lipid Peroxidation (LPO), and Fe ROS, Lipid Peroxidation (LPO), and Fe 2+ Quantification Assay To detect reactive oxygen species (ROS), 10 μM DCFH-DA (S0033S, Beyotime) was used with fluorescence settings of Ex/Em 488/525 ± 20 nm. For lipid peroxidation (LPO) detection, 5 μM C11 BODIPY (D3861, Thermo Fisher Scientific) was employed, with Ex/Em 561/591 ± 10 nm (red channel) and Ex/Em 488/510 ± 10 nm (green channel). Fe 2+ levels were assessed using 5 μM FeRhoNox-1 (MX4558, MKBio, China) with Ex/Em settings at 561/580 ± 10 nm. Images were processed and analyzed with ImageJ software. For quantitative analysis, background subtraction and normalization were applied, and the integrated density of fluorescence signals was measured using ImageJ. ## TEM Imaging of Cells TEM Imaging of Cells HCT116 cells were cultured in 10 cm dishes for 24 h and treated with 300 nM RuOA2 for an additional 24 h. Cells were harvested by trypsinization and fixed in 2.5% glutaraldehyde, followed by postfixation with 1.0% osmium tetroxide. Samples were then embedded in epoxy resin and sectioned into 70 nm slices. TEM images were captured using a Hitachi H-7650 TEM at an acceleration voltage of 80 kV. ## Overexpression of FTH1 in HCT116 and CT26 Cells Overexpression of FTH1 in HCT116 and CT26 Cells HCT116 and CT26 cells were seeded in 6-well plates and allowed to grow to 70–80% confluence. The FTH1-GFP expression plasmid was synthesized by Sangon Biotech. For each well, 2 μg of plasmid DNA was diluted in 125 μL of Opti-MEM I Reduced Serum Medium (Gibco). Separately, 5 μL of Lipo 6000 Transfection Reagent (C0526FT, Beyotime, China) was diluted in 125 μL of Opti-MEM I. After a 5 min incubation at room temperature, the diluted DNA and Lipo 6000 were mixed and incubated for 20 min to allow the formation of DNA-lipid complexes. The mixture was then added dropwise to each well containing cells in serum-free medium and incubated for 6 h at 37 °C in a CO 2 incubator. Following this, the transfection medium was replaced with fresh complete growth medium, and the cells were incubated for an additional 24–48 h prior to downstream analysis. ## Knockout of NCOA4 in HCT116 Cells Knockout of NCOA4 in HCT116 Cells NCOA4 CRISPR-Cas9 gene disruption in HCT116 cells was performed with the LentiCRISPR V2 vector (sg- NCOA4 ) using the following oligonucleotide sequence: 5′-ggaatgtcttagaagccgtg-3′. sg- NCOA4 and negative control (sg-NC) lentivirus (10 9 TU/ml) was obtained from Tsingke, China. 200 μL lentivirus transfection mixture (8 μL lentivirus, 8 μL HitransG P infection reagent, 184 μL DMEM) was added to 10 4 cells seeded in a 24-well plate. After coculturing for 24 h, replaced with fresh medium. Stably transduced cells were obtained by puromycin selection (2 μg/mL) ## Western Blot Analysis Western Blot Analysis Cells and tumor tissues were lysed in cell lysis buffer with protease inhibitors (P0013, Beyotime), followed by centrifugation at 12000 rpm for 10 min at 4 °C. The supernatant was collected, and protein concentrations were determined using a BCA Quantitation Kit (23228, Thermo Fisher Scientific). Proteins were then mixed with SDS sample buffer containing 10% β-mercaptoethanol, separated by SDS-PAGE, and transferred to nitrocellulose membranes for immunoblotting. β-Actin was used as the loading control. Primary antibodies included: anti-β-Actin (66009–1-Ig, Proteintech, 1:5000), anti-LC3B I/II (ET1701–65, HuaBio, 1:1000), anti-ATG5 (ET1611–38, HuaBio, 1:1000), anti-ATG7 (ET1610–53, HuaBio, 1:1000), anti-NCOA4 (ER62707, HuaBio, 1:2000); anti-FTH1 ( T55648 , Abmart, 1:1000); anti-GPX4 (T56959S, Abmart, 1:1000). Blots were developed using an enhanced chemiluminescence (ECL) detection system (1705061, Bio-Rad, USA), and signal were visualized on a Chemiluminescence imaging analysis system (BG-gdsAUTO 720, Baygene, China). Band intensities were quantified using ImageJ software. ## Immunohistochemistry Immunohistochemistry Tumor tissues were collected and embedded in Optimal Cutting Temperature (OCT) compound. OCT-embedded tissues were sectioned at a thickness of 5 μm and mounted onto glass slides. Sections were fixed in 4% paraformaldehyde for 20 min, permeabilized with 0.1% Triton X-100 for 1 h, and blocked with 5% bovine serum albumin (BSA) at room temperature for 30 min. The sections were then incubated overnight at 4 °C with primary antibodies against FTH1 (1:500), GPX4 (1:500), and Ki67 (PT0321R, Immunoway, 1:500), followed by incubation with a Cy3-conjugated goat antirabbit IgG secondary antibody (SA00009–2, Proteintech, 1:100) for 1 h at room temperature. Nuclear staining was performed using 4′,6-diamidino-2-phenylindole (DAPI, P0131, Beyotime) for 30 min. Confocal images were captured on an Olympus FV3000 laser scanning microscope. ## Animal Studies Animal Studies All animal experiments were approved by the Animal Ethics Committee of Xi’an Jiaotong University (No. 2021–1123). Female C57BL/6 mice (6–8 weeks) were obtained from the Experimental Animal Center of Xi’an Jiaotong University and housed under specific pathogen-free conditions with a 12-h light/dark cycle and free access to food and water. Six days after tumor cell inoculation, when tumor signals became detectable by in vivo imaging, mice were randomized into four groups and peritumorally injected every other day with a 5 mg/kg dose of liposome, ruthenosomes, or drug molecules. On day 22, 2 days after the last treatment, tumor tissues were collected for hematoxylin and eosin (H&E) staining to evaluate antitumor effects. ## Subcutaneous CRC Tumor Implantation Subcutaneous CRC Tumor Implantation To generate murine subcutaneous tumors, 5 × 10 6 CT26 cells stably expressing the GFP-Luc fusion gene (CT26-luciferase cells) were injected subcutaneously into female C57BL/6 mice aged 8 weeks. In Vivo Bioluminescence Imaging On day 6 postinoculation, detectable tumor signals were imaged using the in vivo imaging system. Mice were divided into four groups (Saline, Ru-trisBP@OA, CDDP, and RuOA2). Bioluminescence images were captured on days 6, 14, and 22 following an intraperitoneal injection of D-Luciferin potassium salt (150 mg/kg) using the VISQUE In VIVO Smart-LF imaging system. ## Therapeutic Evaluation Therapeutic Evaluation Tumor growth was monitored by measuring tumor weight, length (L), and width (W). Tumor volume was calculated with the formula: 1/2(L × W 2 ). Mice body weights were recorded every 2 days to monitor health. At the end of the treatment period, mice were euthanized, and tumor tissues were stained with H&E for morphological evaluation. Survival time was defined as the duration from tumor implantation until sacrifice or when tumor volume reached 1000 mm 3 , with a maximum follow-up of 45 days. In Vivo Pharmacokinetics of Ruthenosomes RuOA2 was administered to the C57BL/6 mice via intravenous injection at a dose of 5 mg/kg. Blood samples were collected at 0.5, 1, 2, 4, 12, 24, 48 h postinjection. Collected blood samples were incubated at 37 °C for 30 min and then centrifuged at 3000 rpm for 15 min to harvest the serum. The serum was then transferred to clean polypropylene tubes for further analysis. Concentration of RuOA2 in serum samples was determined using ICP-MS. Statistical analysis of the pharmacokinetic parameters was performed by GraphPad Prism software. Data were expressed as mean ± standard deviation (SD) for each group. Ex Vivo Fluorescence Imaging Following intravenous injection of 5 mg/kg RuOA2 at 6 and 24 h, heart, lung, liver, spleen, and kidney were excised from C57BL/6 mice. The tissues were then rinsed with cold 1 x PBS to remove residual blood and debris. Fluorescent images were captured using the VISQUE In VIVO Smart-LF imaging system. ## Statistics and Reproducibility Statistics and Reproducibility Data were analyzed using GraphPad Prism (version 9.0) and are presented as the mean ± standard deviation (SD) with n ≥ 3 unless specified otherwise. Unpaired Student’s t tests were used for comparisons the between two groups, and one-way analysis of variance (ANOVA) was used for multiple group. A p-value of less than 0.05 was considered statistically significant. Statistical significance is indicated by asterisks: p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.