Mitochondrial metabolic reprogramming by a glutamine antagonist radioenhancer conjugate drives pyroptosis and antitumor immunity

Abstract

Aberrant glutamine metabolism undermines the efficacy of cancer radiotherapy (RT) by promoting DNA repair, increasing oxygen consumption, and imposing metabolic stress that compromises the viability and function of cytotoxic T lymphocytes (CTLs). Here, we report a mitochondria-enriched, metabolism-regulating monolayered metal-organic framework (MOF) nanoplatform, Hf-Ir/DON, constructed by incorporating the glutamine antagonist 6-diazo-5-oxo-l-norleucine (DON) into a radiosensitizing hafnium-iridium (Hf-Ir) MOF to inhibit glutaminolysis. This design markedly sensitizes tumor cells to RT and, upon x-ray irradiation, enables mitochondria-targeted RT-radiodynamic therapy with substantially enhanced reactive oxygen species accumulation. The resulting oxidative stress induces extensive DNA damage, mitochondrial dysfunction, and immunogenic tumor cell pyroptosis. Concurrently, Hf-Ir/DON alleviates metabolic constraints on activated T cells, preserving CTL viability while sustaining effector and memory functions. In a murine colon cancer model, low-dose RT combined with Hf-Ir/DON achieved >95% tumor growth inhibition. Further combination with anti–PD-L1 therapy induced regression of bilateral triple-negative breast tumors and prevented lung metastasis. These results establish a radiometabolic MOF nanoplatform that potentiates RT while promoting durable antitumor immunity.

INTRODUCTION

Radiotherapy (RT) is widely used in cancer treatment and has increasingly been explored as a strategy to remodel the tumor immune microenvironment and improve responsiveness to immunotherapy (1–4). However, conventional RT primarily relies on direct DNA damage and often lacks sufficient immunogenicity to elicit robust systemic antitumor immunity (5, 6). Recently, metal-organic frameworks (MOFs) composed of high-Z elements and photosensitizing organic ligands have emerged as promising radiosensitizers capable of mediating RT–radiodynamic therapy (RDT) (7–10). By enhancing x-ray energy deposition and promoting the generation of multiple reactive oxygen species (ROS) (11–13), including hydroxyl radicals (·OH), singlet oxygen (¹O₂), and superoxide (14, 15), MOF-mediated RT-RDT induces pronounced oxidative stress and immunogenic cell death (ICD) in tumors (16, 17). This process leads to the release of damage-associated molecular patterns (18, 19), enhanced antigen presentation, and subsequent dendritic cell (DC) maturation and T cell priming (20, 21). Consequently, MOF-enabled RT-RDT provides a strong biological rationale for combining RT with immunotherapy by simultaneously enhancing tumor cytotoxicity and converting local irradiation into a systemic immune-activating process.

Despite these advances, clinical responses to RT-immunotherapy combinations remain highly heterogeneous (22, 23), largely due to tumor-intrinsic physiological and metabolic states (24–26). Aberrant tumor metabolism has emerged as a major determinant of radioresistance and an underappreciated barrier to effective antitumor immunity (27, 28). Among these metabolic alterations, dysregulated glutamine (Gln) metabolism is particularly prominent (29). Enhanced Gln utilization intensifies mitochondrial oxidative phosphorylation (OXPHOS)–associated oxygen consumption, exacerbating tumor hypoxia and limiting ROS accumulation during irradiation through increased production of redox-buffering molecules such as glutathione (GSH) and reduced form of NADP⁺ (nicotinamide adenine dinucleotide phosphate) (NADPH) (30–33). In parallel, Gln metabolism supplies key intermediates for nucleotide and amino acid biosynthesis, facilitating DNA repair and further reinforcing tumor radioresistance (34).

Beyond tumor cells, excessive Gln metabolism also disrupts mitochondrial homeostasis in immune cells. Its metabolic by-products, particularly ammonia, impair cytotoxic T lymphocyte (CTL) survival and effector function during immune activation (35), weaken effector-to-memory differentiation, and ultimately limit the persistence of antitumor immune responses (36, 37). Therefore, simply increasing radiation energy deposition may not generate durable immunity unless the metabolic constraints that limit ROS amplification and immune maintenance are simultaneously addressed.

Here, we report a mitochondria-targeted monolayered MOF radiosensitizing nanoplatform, Hf-Ir/DON, designed to integrate RT-RDT amplification with metabolic reprogramming. This system is constructed by functionalizing a radiosensitizing Hf-Ir MOF with the Gln antagonist 6-diazo-5-oxo-l-norleucine (DON) (26), enabling targeted modulation of aberrant mitochondrial metabolism. Upon x-ray irradiation, Hf-Ir/DON markedly amplifies intracellular ROS generation, enhances DNA damage, and disrupts mitochondrial homeostasis, ultimately inducing immunogenic tumor cell pyroptosis. The combined RT-RDT and metabolic effects promote efficient tumor antigen release and presentation, leading to robust CTL activation. Inhibition of Gln metabolism alleviates ammonia-mediated metabolic toxicity during immune activation, thereby preserving CTL viability and supporting memory T cell formation, which enables durable adaptive antitumor immunity. In a murine colorectal cancer model, Hf-Ir/DON combined with low-dose RT achieved >95% tumor growth inhibition. Furthermore, integration with anti–PD-L1 (αPD-L1) immune checkpoint blockade (ICB) markedly suppressed both primary and distant tumor growth in a bilateral triple-negative breast cancer model and nearly eliminated pulmonary metastasis. Together, these findings establish a MOF-based radiometabolic strategy that enhances RT efficacy while promoting durable systemic antitumor immunity.

RESULTS

Construction of a radiometabolic MOF platform, Hf-Ir/DON

The Hf-Ir monolayered MOF was synthesized via a modulated solvothermal reaction between HfCl₄ and a photosensitizing iridium complex, Ir(DBB)(ppy)₂⁺ [Ir-DBB; DBB = 4,4′-di(4-benzoato)-2,2′-bipyridine; ppy = 2-phenylpyridine] ligand (Ir-PS), in N,N-dimethylformamide at 80°C using dichloroacetic acid (DCA) and water as modulators (figs. S1 to S5). The resulting Hf-Ir MOF adopts a two-dimensional network with kgd topology, in which Hf₁₂ secondary building units (SBUs) are vertically capped by DCA ligands and laterally bridged by Ir-PS linkers, affording the molecular formula Hf₁₂(μ₃-O)₈(μ₃-OH)₈(μ₂-OH)₆(Ir-PS)₆(DCA)₆. Subsequently, DON was conjugated to the SBUs of Hf-Ir via carboxylate exchange following triflate activation, partially replacing the terminal DCA ligands and yielding Hf-Ir/DON with the formula Hf₁₂(μ₃-O)₈(μ₃-OH)₈(μ₂-OH)₆(Ir-L)₆(OTf)_3.6(DON)_2.4 (Fig. 1A and figs. S6 to S8).

Fig. 1. Synthesis, characterization, and RT-RDT performance of Hf-Ir/DON.

(A) Schematic illustration of the monolayered MOF structure and construction of the Hf-Ir/DON conjugate. (B) Transmission electron microscopy images and corresponding size distributions of Hf-Ir and Hf-Ir/DON. (C) Atomic force microscopy topography images and corresponding height profiles (insets) of Hf-Ir and Hf-Ir/DON. (D) Dynamic light scattering analysis showing number-average hydrodynamic diameter distributions of Hf-Ir and Hf-Ir/DON in PBS (n = 3). (E) X-ray photoelectron spectroscopy spectrum of Hf-Ir/DON. (F) Experimental Powder x-ray diffraction patterns of Hf-Ir and Hf-Ir/DON, together with the simulated pattern of Hf-Ir. (G) DON release profiles from Hf–Ir/DON in 0.1× and 1× PBS (n = 3). (H) Schematic illustration of the RT-RDT mechanism of Hf-Ir. (I to K) X-ray dose-dependent generation of total ROS (I), ¹O₂ (J), or ·OH (K) by PBS, Ir-PS, Hf-Ir, or Hf-Ir/DON (n = 3). Statistical significance was determined using two-way ANOVA, followed by Tukey’s post hoc test for multiple comparisons.

Transmission electron microscopy (Fig. 1B) and atomic force microscopy (Fig. 1C) revealed that both Hf-Ir and Hf-Ir/DON adopt monolayered nanosheet morphology with lateral dimensions ranging from 120 to 280 nm and an average thickness of ∼2 nm. Dynamic light scattering measurements showed hydrodynamic diameters of 202 ± 3 nm for Hf-Ir and 207 ± 2 nm for Hf-Ir/DON (Fig. 1D). X-ray photoelectron spectroscopy confirmed the presence of carbon, nitrogen, oxygen, hafnium, iridium, and fluorine elements in Hf-Ir/DON (Fig. 1E). Powder x-ray diffraction patterns of Hf-Ir/DON closely resembled those of Hf-Ir, indicating that the framework structure remained intact after DON conjugation (Fig. 1F). In aqueous solution, Hf-Ir exhibited a positive surface charge with a ζ-potential of 36.6 ± 0.36 mV, which slightly increased to 38.60 ± 0.56 mV after DON modification (fig. S9).

The release of DON from Hf-Ir/DON displayed phosphate concentration–dependent behavior under phosphate-buffered saline (PBS) conditions. Within 24 hours, Hf-Ir/DON released 22.9% of DON in 0.1× PBS, whereas DON release increased to 61.8% in 1× PBS, demonstrating sustained drug release under physiological conditions (Fig. 1G).

Under gradient x-ray irradiation, Hf-Ir exhibited efficient RT-RDT activity (Fig. 1H). The incorporation of Ir-PS into the Hf-based framework enhances x-ray absorption due to the high atomic number (Z) of Hf, leading to more efficient energy deposition under x-ray irradiation. This process promotes the generation of secondary electrons, thereby amplifying ROS production. Compared to the free Ir-PS ligand, Hf-Ir enhanced the generation of total ROS, ¹O₂, and ·OH by 2.3-, 1.7-, and 1.4-fold, respectively (Fig. 1, I to K). Notably, Hf-Ir/DON displayed ROS generation efficiency comparable to that of Hf-Ir, indicating that DON conjugation does not compromise the RT-RDT performance of the MOF platform.

Hf-Ir/DON–mediated Gln inhibition enhances RT-RDT-driven ROS accumulation

Further studies revealed that Hf-Ir/DON has intrinsic mitochondria-targeting properties and effectively disrupts Gln metabolism within tumor mitochondria, thereby suppressing the tricarboxylic acid (TCA) cycle and rendering tumor cells more susceptible to ROS accumulation (Fig. 2A). In vitro experiments demonstrated that Hf-Ir/DON was efficiently internalized by tumor cells in a time-dependent manner (fig. S10). Following cellular uptake, fluorescence imaging indicated that Hf-Ir/DON was predominantly localized in close proximity to mitochondria in CT26 cells (fig. S11). Quantitative mitochondrial colocalization analysis yielded a Pearson’s correlation coefficient of 0.86, indicating strong spatial correlation between Hf-Ir/DON and mitochondrial markers (Fig. 2B) (38). In the lysosomal colocalization analysis of Hf-Ir/DON, the Pearson correlation coefficient decreased to 0.54, with a colocalization rate of only 37.7%, indicating that a substantial proportion of Hf-Ir/DON had escaped from the lysosomal compartment. This lysosomal escape may be attributed to the combined effects of protonatable groups on the material surface and its positive surface potential. These features increase the osmotic pressure within endolysosomal vesicles and perturb membrane homeostasis, thereby promoting the release of Hf-Ir/DON into the cytosol (fig. S12). Ir-PS moieties confer both cationic and lipophilic properties, facilitating Hf-Ir accumulation in mitochondria through favorable interactions with the negatively charged mitochondrial membrane. Quantitative liquid chromatography–mass spectrometry (LC-MS) analysis showed that, in tumor cells treated with Hf-Ir/DON, 41.5% of intracellular DON was localized in the mitochondrial fraction, whereas 33.1% was detected in the cytosolic fraction. In contrast, following treatment with free DON, only 20.9% was localized in the mitochondrial fraction, whereas 55.4% of intracellular DON was detected in the cytosolic fraction (fig. S13).

Fig. 2. Mitochondria-targeted Hf-Ir/DON amplifies RT-RDT through inhibition of Gln metabolism.

(A) Schematic illustration of mitochondria-targeted Hf-Ir/DON disrupting the TCA cycle and inhibiting GSH synthesis through glutamate (Glu) blockade. (B) CLSM images and colocalization analysis of Hf-Ir/DON with mitochondria in CT26 cells. (C to G) Intracellular GSH levels (C), NADPH/NADP⁺ ratios (D), Gln levels (E), Glu levels (F), and Gln consumption from the culture medium (G) in CT26 cells after treatment with DON, Hf-Ir, or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours (n = 3). (H and I) CLSM images of Ru(dpp)₃Cl₂-stained CT26 cells after the indicated treatments (H) and quantification of fluorescence intensity using ImageJ (I). (J and K) CLSM images (J) and fluorescence intensity quantification (K) of DCFH-DA–stained CT26 cells after treatment with Hf-Ir or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours, followed by 2-Gy x-ray irradiation. (L) CLSM images of JC-1–stained cells treated with Hf-Ir or Hf-Ir/DON (5 μM Hf and 1 μM DON), followed by 2-Gy x-ray irradiation, JC-1 probe (green channel, λ_ex = 488 nm and λ_em = 520 ± 20 nm; red channel, λ_ex = 543 nm and λ_em = 600 ± 20 nm). (M) Monomer/J-aggregate ratios in the mitochondria of CT26 cells after the indicated treatments (n = 3). Statistical significance was determined using one-way ANOVA, followed by Tukey’s post hoc test for multiple comparisons.

Incubation of tumor cells with Hf-Ir/DON effectively blocked the mitochondrial conversion of Gln to glutamate (Glu), resulting in a marked reduction of intracellular GSH levels to approximately one-third of those in untreated cells, indicative of a substantially weakened antioxidant capacity (Fig. 2C). Moreover, Hf-Ir/DON treatment significantly reduced the intracellular NADPH/NADP⁺ ratio compared with Hf-Ir alone, indicating effective suppression of reductive metabolism and consequent attenuation of cellular antioxidant defenses (Fig. 2D). Intracellular Gln levels were markedly elevated in Hf-Ir/DON–treated cells (Fig. 2E), accompanied by a corresponding decrease in Glu levels (Fig. 2F). Consistently, during coincubation, Hf-Ir/DON significantly reduced Gln consumption by tumor cells from the culture medium and exhibited stronger inhibitory activity than free DON (Fig. 2G).

Mitochondrial oxygen levels were next assessed using the oxygen-quenchable probe Ru(dpp)₃Cl₂. Compared with PBS-treated cells, the mean fluorescence intensity decreased by 75.8% in Hf-Ir/DON treatment. This reduction was more pronounced than that observed for free DON, indicating substantial suppression of mitochondrial OXPHOS and consequent accumulation of intracellular oxygen (Fig. 2, H and I).

We next examined the impact of Gln inhibition on RT-RDT–mediated intracellular ROS accumulation. By depleting mitochondrial antioxidant reserves, including GSH and NADPH, Hf-Ir/DON markedly sensitized tumor cells to oxidative stress. Compared with Hf-Ir(+) [where (+) denotes x-ray irradiation], Hf-Ir/DON(+) increased intracellular accumulation of total ROS (Fig. 2, J and K), ¹O₂ (fig. S14), and ·OH (fig. S15) in CT26 cells by 2.26-, 1.65-, and 1.82-fold, respectively. Consistent with enhanced oxidative stress, JC-1 staining revealed that Hf-Ir/DON(+) induced a significantly greater loss of mitochondrial membrane potential than Hf-Ir(+), indicative of mitochondrial depolarization and severe functional impairment (Fig. 2, L and M).

Hf-Ir/DON–induced radiometabolic stress activates caspase-3/gasdermin E–dependent pyroptosis

As illustrated in Fig. 3A, we hypothesized that Hf-Ir/DON could induce tumor cell pyroptosis through mitochondria-targeted radiometabolic regulation under x-ray irradiation. Western blot analysis revealed that, compared with the PBS(+) and Hf-Ir(+) groups, Hf-Ir/DON(+) markedly promoted the release of cytochrome c (Cyt c) from mitochondria into the cytosol, whereas the mitochondrial marker protein COX IV remained unchanged, indicating a pronounced increase in mitochondrial outer membrane permeability (Fig. 3B and fig. S16).

Fig. 3. Hf-Ir/DON–enabled RT-RDT triggers pyroptotic death in tumor cells.

(A) Schematic illustration of the mechanism underlying Hf-Ir/DON–induced tumor cell pyroptosis upon x-ray irradiation. (B) Western blot analysis of mitochondrial COX IV and Cyt c protein levels in CT26 cells after the indicated treatments. (C and D) CLSM analysis of nuclear DNA DSBs (C) and γ-H2AX level (D) in CT26 cells treated with Hf-Ir or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours, followed by 2-Gy x-ray irradiation (n = 5). (E) Western blot analysis of full-length caspase-3 (caspase-3-F), cleaved caspase-3, full-length GSDME (GSDME-F), and cleaved GSDME in CT26 cells treated with Hf-Ir or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours, followed by 2-Gy x-ray irradiation. (F) Flow cytometry analysis of CT26 cells treated with Hf-Ir or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours, followed by 2 Gy x-ray irradiation before staining with annexin V–fluorescein isothiocyanate (FITC) and propidium iodide (PI). (G) Quantification of viable, apoptotic, and necrotic/pyroptotic CT26 cells following the indicated treatments (n = 3 biological replicates). (H) Live-cell imaging of cleaved GSDME expressing CT26 cells after Hf-Ir(+) or Hf-Ir/DON(+) treatment. Scale bar, 25 μm. Statistical significance was determined using one-way and two-way ANOVA, followed by Tukey’s post hoc test for multiple comparisons.

DNA damage induced by irradiation was further evaluated using γ-H2AX immunofluorescence staining. Hf-Ir/DON(+)–treated cells exhibited substantially enhanced γ-H2AX fluorescence signals, characterized by abundant nuclear DNA damage foci (Fig. 3C). Quantitative analysis demonstrated that the γ-H2AX level in the Hf-Ir/DON(+) group was 2.3- and 1.3-fold higher than those in the PBS(+) and Hf-Ir(+) groups, respectively (Fig. 3D), indicating that Hf-Ir/DON markedly amplified DNA double-strand breaks (DSBs) during RT-RDT. Consistent with this observation, clonogenic assays revealed that the survival fractions of cells treated with Hf-Ir/DON were significantly lower than those of the PBS and Hf-Ir groups at all tested irradiation doses (fig. S17), demonstrating that Hf-Ir/DON effectively enhances tumor cell sensitivity to x-ray irradiation.

We next examined proteins associated with apoptosis and pyroptosis. Hf-Ir/DON(+) treatment markedly increased caspase-3 activation in tumor cells, as evidenced by elevated levels of cleaved caspase-3. Concurrently, gasdermin E (GSDME) was efficiently cleaved, resulting in substantial accumulation of cleaved GSDME and indicating a shift in cell death mode from conventional apoptosis toward GSDME-dependent pyroptosis (Fig. 3E and fig. S18). In contrast, this effect was minimal in the PBS(+) and Hf-Ir(+) groups.

To further characterize cell death modes, we performed annexin V/propidium iodide (PI) double-staining flow cytometry analysis (Fig. 3F). Compared with PBS(+) and Hf-Ir(+) groups, Hf-Ir/DON(+) treatment significantly reduced the proportion of viable cells while markedly increasing the proportion of PI-positive necrotic/pyroptotic cells. Quantitative analysis confirmed that the Hf-Ir/DON(+) group exhibited the highest proportion of necrotic/pyroptotic cells, accompanied by a comparatively lower fraction of apoptotic cells (Fig. 3G), suggesting that Hf-Ir/DON predominantly induces a membrane-disruptive form of cell death under irradiation.

Immunofluorescence imaging of cleaved GSDME in tumor cells further supported these findings. As shown in Fig. 3H and fig. S19A, Hf-Ir/DON(+) treatment markedly enhanced cleaved GSDME fluorescence, which was predominantly localized in the plasma membrane, and induced characteristic pyroptotic morphological changes, including cellular swelling and plasma membrane rupture. In contrast, GSDME knockdown or pretreatment with the caspase-3 inhibitor Z-DEVD-FMK substantially reduced the plasma membrane–associated cleaved GSDME signal and attenuated these lytic morphological changes, indicating that Hf-Ir/DON(+)–induced pyroptosis depends on caspase-3–mediated GSDME cleavage. Consistent with these observations, flow cytometry analysis (fig. S19B) showed that the proportion of annexin V/PI double-positive cells increased to 35.4% following Hf-Ir/DON(+) treatment but decreased to 8.9 and 12.8% after GSDME knockdown and Z-DEVD-FMK pretreatment, respectively. These interventions were also accompanied by an increased proportion of cells exhibiting an early apoptotic phenotype, suggesting that inhibition of the caspase-3/GSDME pathway shifted the mode of cell death away from membrane-lytic pyroptosis. These results demonstrate that inhibition of GSDME expression or caspase-3 activity effectively preserves plasma membrane integrity during treatment and further supports the conclusion that Hf-Ir/DON(+) induces tumor cell pyroptosis through the caspase-3/GSDME signaling axis.

Hf-Ir/DON(+) promotes immunogenic priming and prevents CD8⁺ T cell dysfunction

We next systematically evaluated the ability of Hf-Ir/DON to elicit ICD under irradiation. Compared with PBS and Hf-Ir groups, Hf-Ir/DON treatment under low-dose x-ray irradiation markedly increased calreticulin (CRT) exposure on the tumor cell surface (Fig. 4A and figs. S20 and S21) and enhanced the release of high-mobility group box 1 (HMGB1) (Fig. 4B). In parallel, Hf-Ir/DON(+) significantly increased the release of lactate dehydrogenase (LDH) (Fig. 4C) and extracellular adenosine 5′-triphosphate (ATP) (fig. S22), consistent with compromised plasma membrane integrity, and elevated the secretion of the proinflammatory cytokine interleukin-1β (IL-1β) (Fig. 4D). These results show that Hf-Ir/DON(+) induces robust ICD, accompanied by pyroptotic features characterized by GSDME-dependent membrane pore formation and inflammatory cytokine release.

Fig. 4. Hf-Ir/DON(+) induces tumor cell ICD and preserves CD8⁺ T cell function.

(A) Flow cytometry analysis of CRT surface exposure on CT26 cells treated with Hf-Ir or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours, followed by a 2-Gy x-ray irradiation (n = 3). (B to D) Enzyme-linked immunosorbent assay (ELISA) analysis of HMGB1 release (B), LDH release (C), and IL-1β secretion (D) from CT26 cells after the indicated treatments (n = 3). (E) Flow cytometry analysis of mDCs after coincubation with CT26 cells that were subjected to the indicated treatments. (F) Schematic illustration of the in vitro tumor cell–BMDC–PBMC triple-cell coculture system using Hf-Ir/DON(+)–treated tumor cells. h, hours. (G and H) Flow cytometry analysis of CD8⁺ T cell generation (G) and the proportion of dead CD8⁺ T cells (H) in the triple-cell coculture (n = 3 biological replicates). (I) Flow cytometry analysis of CD8⁺ T cell subsets [T_eff, T_naive, T_cm, and T_DN cells (CD44, CD62L double-negative T cells)] in the triple-cell coculture system under different treatments (n = 3 biological replicates). (J and K) Quantification of ammonia (J) and lactate (K) levels in the triple-cell coculture system following the indicated treatments (n = 3 biological replicates). (L) Schematic illustration showing that Hf-Ir/DON(+) remodels tumor immunometabolism to promote controlled immune priming while preserving functional CD8⁺ T cell responses. Statistical significance was determined using one-way and two-way ANOVA, followed by Tukey’s post hoc test for multiple comparisons.

We then assessed DC maturation by flow cytometry (fig. S23). Coincubation with treated tumor cells revealed that the Hf-Ir/DON(+) group generated a significantly higher proportion of CD80⁺CD86⁺ mature DCs (mDCs) than PBS(+) and Hf-Ir(+) groups (Fig. 4E), along with increased secretion of tumor necrosis factor–α (TNF-α) and IL-6 (fig. S24). These findings indicate that Hf-Ir/DON(+) effectively promotes immunogenic DC activation. To further assess downstream regulation of T cell responses, we established a triple-cell coculture system in which tumor cells were treated with Hf-Ir/DON(+), followed by coincubation with bone marrow–derived DCs (BMDCs) and subsequent incubation with peripheral blood mononuclear cells (PBMCs) (Fig. 4F).

Within this system, we quantified CD8⁺ T cell abundance and viability (fig. S25). Hf-Ir/DON(+) markedly increased the proportion of CD8⁺ T cells among CD3⁺ cells to 21.43 ± 0.70% compared with 12.33 ± 0.38% for PBS(+) and 13.6 ± 1.45% for Hf-Ir(+) (Fig. 4G). Concurrently, Hf-Ir/DON(+) significantly improved CD8⁺ T cell survival (Fig. 4H), reducing the fraction of dead CD8⁺ T cells to 24.4 ± 2.1% compared with 55.5 ± 2.8% for PBS(+) and 52.2 ± 2.7% for Hf-Ir(+). These data indicate that Hf-Ir/DON(+) not only enhances CD8⁺ T cell expansion but also mitigates activation-associated cell death (fig. S26).

Phenotypic analysis further revealed that Hf-Ir/DON(+) decreased the proportion of terminally differentiated/dysfunctional T cells (T_DN cells), while increasing the fraction of central memory T cells (T_cm cells) and maintaining the naïve-like T cell (T_naive cell) population (Fig. 4I). Although the overall T_eff cell proportion was comparable across groups, Hf-Ir/DON(+) substantially enriched the KLRG1⁺ T_eff cell subset (57.7 ± 3.0%), markedly exceeding the corresponding levels in the control groups (∼18 to 31%), consistent with enhanced effector differentiation and cytotoxic programming (fig. S27). These findings suggest that Hf-Ir/DON–mediated radiometabolic modulation preserves CD8⁺ T cell functional integrity and supports memory- and effector-associated differentiation during ICD-associated immune activation, although tumor antigen specificity remains to be directly established.

To further characterize the metabolic microenvironment, we quantified key metabolites in the coculture supernatants. Relative to PBS(+) and Hf-Ir(+), Hf-Ir/DON(+) significantly reduced consumption of exogenous Gln, consistent with effective suppression of tumor cell–driven glutaminolysis (fig. S28). In addition, ammonia (NH₃/NH₄⁺) levels were significantly decreased (Fig. 4J), indicating attenuation of glutaminolysis-associated metabolic toxicity. Lactate levels were also markedly reduced (Fig. 4K), likely reflecting that mitochondrial dysfunction induced by Hf-Ir/DON(+) perturbs acid-generating metabolic flux. Together, these metabolic changes establish a more favorable microenvironment that supports enhanced CD8⁺ T cell viability and promotes the differentiation and enrichment of memory- and effector-associated phenotypes.

Hf-Ir/DON(+) reprograms antitumor immunity and synergizes with PD-L1 blockade to control primary and distant tumors

As we were unable to differentiate released and MOF-bound DON in tissues, we extracted and digested the tissues to quantify total DON amounts in tumors and the blood. Following intratumoral injection, Hf-Ir/DON markedly increased local drug exposure over free DON in tumors, with a 4.74-fold increase in AUC_0-∞ (Fig. 5A). The retention half-life (t_1/2) was extended to 42.2 hours for Hf-Ir/DON compared with 16.5 hours for free DON, demonstrating enhanced tumor retention and a prolonged therapeutic window. After intratumoral administration of Hf-Ir/DON at a dose of 10 mg/kg, plasma DON concentrations remained low (<0.35 μg/ml) and rapidly declined over time (fig. S29). In contrast, free DON showed higher plasma concentrations during the initial period, indicating that incorporation into the Hf-Ir framework slowed the release of DON from the tumor into the circulation.

Fig. 5. Hf-Ir/DON(+) induces robust antitumor immune responses to suppress subcutaneous CT26 tumors.

(A) LC-MS analysis of DON retention after intratumoral injection of DON and Hf-Ir/DON. (B) Schematic illustration of experimental timelines and the treatment protocol for subcutaneous CT26 tumor models. Mice were randomly divided into six groups (n = 5 biologically independent animals per group): PBS, PBS(+), Hf-Ir (Hf, 50 μmol/kg), Hf-Ir/DON (Hf, 50 μmol/kg; DON, 10 μmol/kg), Hf-Ir(+), and Hf-Ir/DON(+), where (+) indicates exposure to 2-Gy x-ray irradiation. (C) Tumor growth curves of subcutaneous CT26 tumors in mice receiving different treatments. Data are presented as means ± SD (n = 5 biological replicates). (D) Individual tumor growth curves of subcutaneous CT26 tumors in mice receiving different treatments. (E) t-SNE visualization of flow cytometry–derived immune cell populations in CT26 tumors after different treatments. (F to L) Flow cytometry analysis of T cell–related immune responses in CT26 tumors after different treatments. The frequencies of CD3⁺ T cells among CD45⁺ cells (F), CD8⁺ T cells among CD3⁺ cells (G), T_naive, T_em, and T_cm cells (H), GzmB⁺CD8⁺ T cells (I), and exhausted CD8⁺ T cells (J), natural killer T cells (NKT cells) (K), and B cells (L) were quantified. Data are shown as means ± SD (n = 5 biological replicates). Statistical significance was determined using one-way and two-way ANOVA, followed by Tukey’s post hoc test for multiple comparisons.

We next investigated the therapeutic efficacy of Hf-Ir/DON–mediated RT-RDT regimen in a CT26 tumor model. Hf-Ir/DON was administered intratumorally, followed by three fractions of low-dose x-ray irradiation (2 Gy) on every other day (Fig. 5B). A 2-Gy fraction size was selected to reflect conventional RT in the clinic (39). Fractionated irradiation was used to better mimic clinical practice than a single high-dose exposure (40). All groups exhibited comparable body-weight profiles throughout treatment (fig. S30), and hematoxylin and eosin (H&E) staining revealed no apparent systemic toxicity (fig. S31). Furthermore, no hematological toxicity was detected following intratumoral injection of Hf-Ir or Hf-Ir/DON (table S1), and liver and kidney functions remained normal in the treated mice (table S2).

Tumor growth analysis showed minimal inhibition in PBS and Hf-Ir groups, while Hf-Ir(+) modestly delayed tumor progression. In contrast, Hf-Ir/DON(+) achieved the most pronounced antitumor effect (Fig. 5C), with an average tumor growth inhibition rate of ∼95%. Individual tumor growth curves further demonstrated sustained tumor suppression in most mice treated with Hf-Ir/DON(+) up to day 19 (when PBS-treated mice reached humane end points), with reduced interanimal variability (Fig. 5D).

To delineate immune remodeling during therapy, we profiled tumor tissues by single-cell flow cytometry followed by t-SNE visualization (Fig. 5E and fig. S32). Compared with control groups, Hf-Ir/DON(+) induced a marked expansion and reorganization of immune cell clusters, characterized by increased CD45⁺ immune infiltration and more continuous distribution across t-SNE space. CD8⁺ T cell populations displayed a more favorable phenotypic landscape, with enrichment of memory-associated subsets, including T_cm cells and effector memory–like cells (T_em cells), indicating a shift toward a more immunologically active, T cell–dominant tumor microenvironment.

Quantitative analyses corroborated these observations. Hf-Ir/DON(+) increased intratumoral CD3⁺ T cells to 28.24 ± 11.1% of CD45⁺ cells (Fig. 5F and fig. S33) and elevated CD8⁺ T cells to 41.7 ± 7.4% of CD3⁺ cells (Fig. 5G and fig. S34). Within the CD8⁺ compartment, Hf-Ir/DON(+) markedly enriched T_cm cells while reducing T_naive cell populations, consistent with a redistribution toward memory-prone lineages (Fig. 5H and fig. S35). T_em cells showed modest but statistically significant increases relative to PBS. Cytotoxic function was enhanced, as evidenced increased GzmB⁺CD8⁺ T cells (Fig. 5I and fig. S36), alongside a reduction in exhausted CD8⁺ T cells (Fig. 5J and fig. S37). In addition, natural killer T cells (NKT cells) showed an increasing trend (Fig. 5K and fig. S38), whereas B cell populations remained largely unchanged (Fig. 5L and fig. S39). Collectively, these results indicate that Hf-Ir/DON(+) not only suppresses tumor growth but also reprograms the tumor immune microenvironment toward enhanced T cell infiltration, cytotoxicity, and memory formation. To assess durable antitumor immune memory, we rechallenged mice with nearly eradicated primary tumors with CT26 cells in the contralateral dorsal flanks without additional treatment. Both Hf-Ir/DON(+) alone and its combination with αPD-L1 suppressed rechallenged tumor growth, with combination treatment showing greater efficacy. These results indicate that Hf-Ir/DON(+)–based treatments induce durable systemic tumor–specific immune memory (fig. S40).

We next investigated whether Hf-Ir/DON–mediated radiometabolic therapy could synergize with ICB. An orthotopic bilateral 4T1 mammary fat pad tumor model was established, in which treatments were administered locally into the primary tumor, followed by unilateral irradiation (three fractions of 4 Gy per fraction on every other day). αPD-L1 antibody was administered with or without Hf-Ir/DON treatment, and tumor growth was subsequently monitored for both irradiated primary and contralateral nonirradiated distant tumors (Fig. 6A).

Fig. 6. Hf-Ir/DON(+) synergizes with αPD-L1 in bilateral orthotopic 4T1 tumors.

(A) Schematic illustration of the experimental timeline and treatment protocol in orthotopic bilateral 4T1 breast cancer models bearing primary and distant tumors. Mice were randomly assigned to five groups (n = 6 biologically independent animals per group): PBS, PBS(+), αPD-L1(+) (15 mg/kg), Hf-Ir/DON(+) (Hf, 50 μmol/kg; DON, 10 μmol/kg), and Hf-Ir/DON(+) + αPD-L1 (Hf, 50 μmol/kg; DON, 10 μmol/kg; αPD-L1, 15 mg/kg). ip, intraperitoneally. (B and C) Tumor growth curves of primary (B) and distant (C) tumors in orthotopic bilateral 4T1 tumor–bearing mice receiving the indicated treatments (n = 6). (D) Tumor weights at the treatment end point (n = 6 biological replicates). (E to I) Flow cytometry analysis of T cell–related immune responses in bilateral orthotopic 4T1 tumors after different treatments. The frequencies of CD3⁺ T cells among CD45⁺ cells (E), CD8⁺ T cells among CD3⁺ cells (F), IFN-γ⁺ T cells among CD8⁺ T cells (G), NKT cells (H), and regulatory T cells (T_reg cells) in CD4⁺ T cells (I) were quantified (n = 5 biological replicates). (J) Representative images of γ-H2AX immunohistochemistry (IHC) (top) and H&E staining (bottom) of tumor sections from mice under different treatments. (K) Representative H&E staining images of spleen and lung tissues at the treatment end point. Data are shown as means ± SD. Statistical significance was determined using two-way ANOVA, followed by Tukey’s post hoc test for multiple comparisons.

PBS and PBS(+) showed minimal control of either tumor, and αPD-L1(+) alone produced only modest effects. In contrast, Hf-Ir/DON(+) markedly suppressed primary tumor growth and also inhibited distant tumors. Notably, combining Hf-Ir/DON(+) with αPD-L1 resulted in the most pronounced suppression of both primary and distant tumors (Fig. 6, B and C). Individual tumor growth trajectories confirmed consistent and sustained control in the combination group across both tumor sites (fig. S41). Body weights remained stable across all groups (fig. S42), indicating good tolerability. At end point, combination treatment reduced primary and distant tumor weights to 19.4 and 17.3% of PBS controls, respectively (Fig. 6D and fig. S43).

Flow cytometry analysis of tumor-infiltrating immune cells further demonstrated that Hf-Ir/DON(+)–based treatment remodeled the immune microenvironment in both primary and distant tumors (fig. S44). Compared with PBS(+), Hf-Ir/DON(+) increased CD3⁺ T cell infiltration by approximately 2.6-fold in primary tumors and 1.7-fold in distant tumors. The addition of αPD-L1 resulted in approximately 3.5- and 2.1-fold increases over PBS(+) in primary and distant tumors, respectively (Fig. 6E and fig. S45). However, the increase in CD3⁺ T cell infiltration in the combination group relative to Hf-Ir/DON(+) treatment was not significant in primary tumors. Hf-Ir/DON(+) also increased the proportion of CD8⁺ T cells within the CD3⁺ T cell population by approximately 1.8-fold in primary tumors and 1.7-fold in distant tumors relative to PBS(+). Combination with αPD-L1 produced approximately 2.3- and 1.8-fold increases over PBS(+) in primary and distant tumors, respectively (Fig. 6F and fig. S46).

Compared with Hf-Ir/DON(+) monotherapy, combination treatment increased the proportion of interferon-γ–positive (IFN-γ⁺) cells among CD8⁺ T cells by approximately 2.9-fold in primary tumors and 2.0-fold in distant tumors, respectively (Fig. 6G and fig. S47), indicating substantially enhanced effector activity of tumor-infiltrating CTLs. Combination treatment also increased the abundance of NK1.1⁺CD3⁺ cells by approximately 2.5- and 2.8-fold over the PBS(+) group in primary and distant tumors (Fig. 6H and fig. S48). Conversely, Hf-Ir/DON(+) reduced the proportion of regulatory T cells within the CD4⁺ T cell population to approximately 0.45-fold in primary tumors and 0.50-fold in distant tumors relative to PBS(+), respectively (Fig. 6I and fig. S49). These findings indicate that Hf-Ir/DON(+)–based treatment, particularly in combination with αPD-L1, promoted T cell infiltration and activation while partially alleviating the immunosuppressive tumor microenvironment.

Immunohistochemical analysis showed that Hf-Ir/DON(+) markedly enhanced cleaved caspase-3 signals in primary orthotopic tumors, indicating potential activation of the caspase-3–dependent pyroptotic pathway under RT-RDT (fig. S50). The resulting release of tumor antigens and immunostimulatory damage-associated molecular patterns may convert the irradiated primary tumor into an in situ tumor vaccine, thereby generating a pronounced abscopal effect and a synergistic response to ICB.

Histological analysis further supported these findings. In primary tumors, Hf-Ir/DON(+) induced extensive DNA damage, and the combination with αPD-L1 resulted in widespread tumor necrosis (Fig. 6J). Both Hf-Ir/DON(+) treatment and combination treatment preserved splenic germinal center architecture, reduced spleen weight, and markedly decreased the number of pulmonary metastatic nodules (Fig. 6K and fig. S51), indicating alleviation of tumor-associated systemic immune dysregulation together with effective suppression of lung metastasis. Together, these results demonstrate that Hf-Ir/DON–mediated RT-RDT synergizes with PD-L1 blockade to achieve robust control of both irradiated primary tumors and nonirradiated distant tumors, highlighting its potential to elicit systemic antitumor immunity.

DISCUSSION

RT is cornerstone of cancer treatment and has been shown to remodel the tumor immune microenvironment, thereby enabling synergistic combinations with ICB in preclinical models. However, the clinical benefits of RT-ICB combination therapies have proven inconsistent (41–44), in part due to the limited immunogenicity of conventional RT (45, 46). MOF-mediated RT-RDT has recently emerged as a promising strategy that couples localized tumor cytotoxicity with immune modulation (47). By amplifying x-ray energy deposition and ROS generation, RT-RDT can effectively induce ICD and promote antigen release, presentation, and T cell priming (48–50). We previously reported GSH-depleting and chemotherapy-enhancing radiosensitizers that improved RT-RDT primarily by weakening tumor cell antioxidant defenses and impairing DNA repair (51, 52). However, these approaches generally do not address the tumor-intrinsic metabolic constraints that limit radiation-induced antitumor immune activation.

A major barrier to effective RT-ICB synergy arises from tumor-intrinsic metabolic adaptations that impose dual constraints on ROS amplification and immune maintenance (53–56). Among these, aberrantly elevated Gln metabolism is particularly prominent (57, 58). In tumor cells, Gln metabolism sustains mitochondrial OXPHOS and redox-buffering systems, such as GSH and NADPH (59), thereby limiting effective ROS accumulation during RT while simultaneously supporting DNA damage repair (60). In parallel, ammonia, a by-product of Gln catabolism, disrupts mitochondrial homeostasis in immune cells and impairs the survival and effector/memory maintenance of CTLs (37, 61). Consequently, RT-induced immune priming often fails to translate into durable systemic antitumor responses. These findings suggest that enhancing radiation energy deposition alone is insufficient and effective radioimmunotherapy requires concurrent alleviation of tumor-intrinsic metabolic resistance and mitigation of metabolic toxicity within the tumor microenvironment. Gln-based immunometabolic therapies can impair tumor metabolic fitness and reduce nutrient competition within the tumor microenvironment; however, systemic Gln blockade can cause substantial off-target toxicity.

In this study, we uncover an integrated radiometabolic nanoradioimmunotherapy strategy that addresses these limitations. Local delivery of DON via Hf-Ir/DON couples tumor-selective Gln inhibition with RT-RDT, thereby remodeling the immunometabolic microenvironment and widening the therapeutic window. We engineered a mitochondria-enriched monolayered MOF radiosensitizing platform by functionalizing the Gln antagonist DON onto the Hf-Ir MOF. This design enables coordinated suppression of tumor cell Gln metabolism while enhancing RT-RDT–mediated oxidative stress. Mechanistically, the Hf-Ir framework provides efficient x-ray energy deposition and ROS amplification, whereas DON disrupts metabolic defenses by attenuating redox buffering capacity and DNA repair. Together, these effects establish a synergistic radiosensitization axis that integrates RT-RDT amplification with metabolic vulnerability. As a result, mitochondrial OXPHOS and redox homeostasis are profoundly perturbed, antioxidant pools (GSH and NADPH) are depleted, and the TCA cycle flux is suppressed, rendering tumor cells highly susceptible to low-dose irradiation, leading to enhanced ROS accumulation.

This radiometabolic synergy not only enhances tumor cytotoxicity but also reprograms the mode of tumor cell death toward an immunostimulatory phenotype. Hf-Ir/DON(+) promotes GSDME-dependent pyroptosis and amplifies ICD-associated danger signals, thereby enhancing DC maturation and antigen presentation and facilitating robust priming of antitumor T cell responses.

Beyond immune initiation, effective radioimmunotherapy requires sustained cytotoxic T cell function. We found that Hf-Ir/DON(+) markedly reduced activation-induced CD8⁺ T cell death and promoted a shift toward a memory-associated phenotype, characterized by increased T_cm cells and reduced exhaustion signatures. This functional preservation is attributed to DON-mediated inhibition of Gln catabolism and consequent reduction of ammonia-associated metabolic toxicity. By improving the metabolic landscape of the tumor microenvironment, Hf-Ir/DON supports CTL survival, effector function, and differentiation into long-lived memory populations. These findings highlight metabolic regulation as a critical, yet underappreciated, determinant of durable RT-ICB synergy.

Therapeutically, a single intratumoral injection of Hf-Ir/DON combined with low-dose RT achieved ∼95% tumor growth inhibition in a colorectal cancer model. In an orthotopic bilateral triple-negative breast cancer model, the addition of αPD-L1 ICB further converted local RT into a robust systemic response, effectively suppressing both irradiated primary tumors and nonirradiated distant tumors. These results demonstrate that Hf-Ir/DON enables reliable conversion of RT-RDT into durable systemic immunity while enhancing responsiveness to ICB. Because the present study was limited to syngeneic murine tumor models, future studies should incorporate additional immunologically “cold” models, genetically engineered tumor models, and human-relevant systems to further evaluate the generalizability and translational potential of this strategy.

In summary, Hf-Ir/DON expands MOF-based RT-RDT by coupling enhanced radiation-induced oxidative stress with tumor-localized Gln metabolic inhibition and alleviation of metabolic constraints on antitumor immunity. By simultaneously targeting tumor redox defenses and metabolic toxicity that impairs CTL persistence, this strategy achieves more robust and durable radioimmunotherapeutic synergy. These findings highlight the potential of radiometabolic modulation to overcome radioresistance and improve clinical outcomes of cancer immunotherapy.

MATERIALS AND METHODS

Chemicals, cells, and animals

All starting reagents were obtained from commercial suppliers and used without further purification unless otherwise noted. Anhydrous solvents were purchased from Thermo Fisher Scientific (USA). DON was obtained from MedChemExpress (MCE; USA). αPD-L1 monoclonal antibody was purchased from Bio X Cell (USA). The murine colon carcinoma cell line CT26 and the murine triple-negative breast cancer cell line 4T1 were purchased from the American Type Culture Collection (USA). CT26 and 4T1 cells were maintained in Dulbecco’s modified Eagle’s medium or RPMI 1640, respectively, supplemented with 10% fetal bovine serum (FBS), penicillin (100 U/ml), and streptomycin (100 μg/ml). BMDCs were generated from bone marrow precursors isolated from 6-week-old BALB/c mice and cultured in RPMI 1640 containing 20% FBS, granulocyte-macrophage colony-stimulating factor (GM-CSF; 20 ng/ml), and IL-4 (10 ng/ml). PBMCs were isolated from BALB/c mouse blood and cultured in RPMI 1640 supplemented with 10% FBS. All cells were cultured at 37°C in a humidified incubator with 5% CO₂. Female C57BL/6 and BALB/c mice (6 weeks old, 18 to 22 g) were purchased from Harlan Laboratories Inc. (Indianapolis, USA). All animal experiments were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Chicago [PHS Assurance #D16-00322 (A3523-01)].

Total ROS detection

Total ROS generation in solution

Total ROS generation by Ir-PS, Hf-Ir, or Hf-Ir/DON under x-ray irradiation (denoted +) was quantified using the 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA; Thermo Fisher Scientific, USA) assay. Briefly, 100 μl of Ir-PS, Hf-Ir, or Hf-Ir/DON suspension (20 μM Hf in PBS) was dispensed into a 96-well plate and mixed with DCFH (final concentration, 5 μM). Samples were exposed to x-ray irradiation at 0, 2, 4, 8, and 10 Gy. Fluorescence signals were recorded on a Synergy HTX microplate reader (BioTek, USA).

Intracellular total ROS detection

Intracellular total ROS levels were assessed in CT26 cells using the DCFH-DA assay. CT26 cells were seeded in petri dishes (5 × 10⁵ cells per dish) and allowed to adhere overnight. Cells were incubated with PBS, Hf-Ir, or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours. Cells were then loaded with DCFH-DA (10 μM) for 1 hour, irradiated with 0- or 2-Gy x-rays, washed with PBS, and either imaged by confocal laser scanning microscopy (CLSM) or harvested for flow cytometry analysis.

Singlet oxygen detection

¹O₂ detection in solution

¹O₂ production by Ir-PS(+), Hf-Ir(+), or Hf-Ir/DON(+) was determined using Singlet Oxygen Sensor Green (SOSG; Life Technologies, USA). Briefly, 100 μl of Ir-PS, Hf-Ir, or Hf-Ir/DON suspension (20 μM Hf in PBS) was mixed with SOSG (final concentration, 5 μM) in a 96-well plate. Samples were irradiated with x-rays at 0, 2, 4, 8, and 10 Gy, and SOSG fluorescence was measured using a Synergy HTX microplate reader (BioTek, USA).

Intracellular ¹O₂ detection

Intracellular ¹O₂ generation was visualized by CLSM using SOSG. CT26 cells were seeded in petri dishes (5 × 10⁵ cells per dish) and cultured overnight. Cells were treated with DON, Hf-Ir, or Hf-Ir/DON (10 μM Hf and 2 μM DON, where applicable) for 8 hours, followed by incubation in fresh complete medium for 12 hours. Cells were then incubated with SOSG (10 μM) for 1 hour, exposed to 0- or 2-Gy x-rays, washed with PBS, and imaged by CLSM. SOSG fluorescence intensity was quantified using ImageJ.

Hydroxyl radical detection

·OH detection in solution

·OH generation by Ir-PS(+), Hf-Ir(+), or Hf-Ir/DON(+) was evaluated using hydroxyphenyl fluorescein (HPF; MCE, USA). Briefly, 100 μl of Ir-PS, Hf-Ir, or Hf-Ir/DON suspension (20 μM Hf in PBS) was mixed with HPF (final concentration, 2 μM) in a 96-well plate. The plate was irradiated with x-rays at 0, 2, 4, 8, and 10 Gy, and fluorescence signals were recorded using a Synergy HTX microplate reader (BioTek, USA).

Intracellular ·OH detection

Intracellular ·OH generation was visualized using HPF. CT26 cells were seeded in petri dishes (2 × 10⁵ cells per dish) and cultured overnight. Cells were treated with DON, Hf-Ir, or Hf-Ir/DON (10 μM Hf and 2 μM DON) for 8 hours, followed by incubation in fresh complete medium for 12 hours. Cells were then incubated with HPF (10 μM) for 1 hour, irradiated with 0- or 2-Gy x-rays, washed with PBS, and imaged by CLSM. HPF fluorescence intensity was quantified using ImageJ.

Cellular uptake

CT26 cells were seeded in six-well plates at a density of 5 × 10⁵ cells per well and cultured overnight. Cells were then treated with Hf-Ir or Hf-Ir/DON (10 μM Hf) and harvested at the indicated time points. After washing with PBS, the cells were scraped, collected by centrifugation, and digested with concentrated nitric acid. Samples were diluted to 1% HNO₃, filtered, and analyzed by inductively coupled plasma MS.

Mitochondrial and lysosome colocalization analysis

For fluorescence colocalization studies, CT26 cells were seeded in confocal dishes at a density of 5 × 10⁵ cells per dish and cultured overnight. Cells were then incubated with Hf-Ir/DON (10 μM Hf and 2 μM DON) for 6 hours. For mitochondrial colocalization analysis, cells were sequentially stained with MitoTracker Green (0.2 μM for 15 min) and Hoechst 33342 (2 μM for 10 min). For lysosomal colocalization analysis, cells were stained with LysoTracker Green DND-26 (Thermo Fisher Scientific; 0.2 μM for 30 min) and Hoechst 33342 (2 μM for 10 min). After staining, cells were washed with PBS and imaged using CLSM to examine the subcellular localization of Hf-Ir/DON. Colocalization analysis was performed using the Stellaris colocalization module, and Pearson’s correlation coefficients were calculated to quantitatively evaluate the extent of colocalization of Hf-Ir/DON with mitochondria or lysosomes.

Gln metabolism analysis

CT26 cells were seeded in six-well plates at 5 × 10⁵ cells per well and cultured overnight. Cells were treated with PBS, DON, Hf-Ir, or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours. The medium was then replaced with fresh complete medium, and cells were further incubated for 12 hours. Cells were washed with cold PBS and harvested on ice. Metabolites were extracted using prechilled 80% methanol in water, followed by vortexing and incubation at −20°C to precipitate proteins. After centrifugation, supernatants were collected, dried under vacuum, and reconstituted in appropriate LC-MS solvents before analysis. For culture medium analysis, aliquots were collected, deproteinized with 80% methanol in water, and centrifuged, and the clarified supernatants were subjected to LC-MS analysis.

In vitro NADPH/NADP⁺ and reduced GSH detection

CT26 cells were seeded in six-well plates at a density of 5 × 10⁵ cells per well and cultured overnight. Cells were treated with PBS, DON, Hf-Ir, or Hf-Ir/DON (5 μM Hf and 1 μM DON, where applicable) for 6 hours. The treatment medium was then replaced with fresh complete medium, and cells were incubated for an additional 12 hours. After treatment, cells were washed twice with ice-cold PBS, detached, and counted. For each biological replicate, 1 × 10⁶ viable cells were collected by centrifugation and lysed in 1 ml of ice-cold extraction buffer containing 50 mM MES (pH 6.0 to 7.0) and 1 mM EDTA by homogenization or sonication on ice. Cell lysates were centrifuged at 10,000g for 15 min at 4°C, and the clarified supernatants were collected for subsequent analyses.

Intracellular NADPH/NADP⁺ ratios were determined using a commercial NADP/NADPH assay kit (Sigma-Aldrich, USA). Separate aliquots of the clarified lysates were used to determine intracellular reduced GSH concentrations using a reduced GSH assay kit (Thermo Fisher Scientific, USA). Absorbance or fluorescence signals were measured using a microplate reader, and reduced GSH concentrations were calculated from a simultaneously generated GSH standard curve. GSH contents were corrected for the sample dilution factor and extraction volume and expressed as micromolars.

Intracellular O₂ sensing

CT26 cells were seeded in petri dishes at 5 × 10⁵ cells per dish and cultured overnight under hypoxic conditions (partial pressure of oxygen = 0.05). Cells were treated with PBS, DON, Hf-Ir, or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours, after which the medium was replaced with fresh complete medium and cells were incubated for an additional 12 hours. Cells were then stained with Ru(dpp)₃Cl₂ (10 μM) for 1 hour, washed with PBS, and imaged by CLSM using excitation at 455 nm and 580- to 630-nm emission collection. Fluorescence intensity was quantified using ImageJ.

Western blot

CT26 cells were seeded in six-well plates at 5 × 10⁵ cells per well and cultured overnight. Cells were treated with PBS, Hf-Ir, or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours. After replacing the medium with fresh complete medium, cells were exposed to 0- or 2-Gy x-ray irradiation and further incubated for 24 hours. Cells were washed with ice-cold PBS and lysed on ice in immunoprecipitation lysis buffer. Lysates were homogenized and centrifuged (14,000g for 15 min at 4°C), and supernatants were collected. For mitochondrial protein extraction, mitochondrial and cytosolic fractions were isolated using a Mitochondria/Cytosol Fractionation Kit (Abcam, ab65320) according to the manufacturer’s instructions. Equal amounts of protein from each fraction were subjected to Western blotting. Protein extracts were separated by SDS–polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membranes. Membranes were blocked with 5% (w/v) nonfat milk in TBST at 4°C for 1 hour, followed by overnight incubation at 4°C with primary antibodies against Cyt c (Cell Signaling Technology; 1:1000), CASP3 (Cell Signaling Technology; 1:1000), GSDME (Cell Signaling Technology, 1:1000), COX IV (Cell Signaling Technology; 1:1000), or β-actin (Cell Signaling Technology; 1:2000). After washing with TBST, membranes were incubated with horseradish peroxidase (HRP)–conjugated secondary antibodies at room temperature for 1 hour. Protein bands were visualized using the Pierce ECL Plus Western Blotting Substrate (Thermo Fisher Scientific, USA) and imaged with a ChemiDoc imaging system (Bio-Rad, USA).

Mitochondrial membrane potential assessment

CT26 cells were seeded in six-well plates at 5 × 10⁵ cells per well and cultured overnight. Cells were treated with PBS, Hf-Ir, or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours, followed by 2-Gy x-ray irradiation. After further incubation at 37°C for 24 hours, cells were stained with JC-1 (Abcam; 10 μM) for 20 min. After washing with PBS, cells were analyzed by flow cytometry. In parallel, JC-1 fluorescence was imaged by CLSM, and the ratio of JC-1 monomers to aggregates was quantified using ImageJ.

Colony formation assay

CT26 cells were seeded in six-well plates at 5 × 10⁵ cells per well and cultured overnight. Cells were treated with PBS, Hf-Ir, or Hf-Ir/DON (10 μM Hf and 2 μM DON) for 8 hours. After treatment, the medium was replaced with fresh complete medium, and cells were incubated for an additional 12 hours, followed by x-ray irradiation at 0, 2, 4, 6, and 8 Gy. Cells were trypsinized, counted, and reseeded into new six-well plates at 200 cells per well. Colonies were allowed to form for 7 days and then quantified. Plating efficiency and surviving fraction were calculated accordingly.

GSDME-dependent pyroptotic cell analysis

CT26 cells were seeded in petri dishes at 5 × 10⁵ cells per dish and cultured overnight. Cells were treated with PBS, Hf-Ir, or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours, followed by x-ray irradiation (2 Gy). After an additional 12-hour incubation at 37°C, cells were incubated with an anticleaved GSDME primary antibody (Cell Signaling Technology; 1:1000) at 37°C for 1 hour, followed by incubation with an Alexa Fluor 488–conjugated secondary antibody (Cell Signaling Technology; 1:1000) at 37°C for 1 hour. Cells were washed with PBS, counterstained with Hoechst 33342 (2 μM for 10 min) to label nuclei, and imaged by CLSM.

CT26 cells were seeded in confocal dishes at a density of 5 × 10⁵ cells per dish and cultured overnight. For GSDME knockdown, cells were transfected with Dfna5-specific small interfering RNA (si-Dfna5; 30 nM) for 24 hours before treatment. For caspase-3 inhibition, cells were preincubated with the caspase-3 inhibitor Z-DEVD-FMK (20 μM) for 2 hours before the addition of Hf-Ir/DON. Cells were subsequently treated with DON, Hf-Ir, or Hf-Ir/DON (10 μM Hf and 2 μM DON) for 8 hours. The culture medium was then replaced with fresh complete medium, and cells were further incubated for 12 hours before exposure to x-ray irradiation (2 Gy). After an additional 12-hour incubation at 37°C, the cells were incubated with an anticleaved GSDME primary antibody (Cell Signaling Technology; 1:1000) at 37°C for 1 hour, followed by an Alexa Fluor 488–conjugated secondary antibody (Cell Signaling Technology; 1:1000) at 37°C for 1 hour. After washing with PBS, the nuclei were counterstained with Hoechst 33342 (2 μM for 10 min), and the cells were imaged using CLSM.

For flow cytometry analysis, CT26 cells were transfected with si-Dfna5 (30 nM for 24 hours) or pretreated with Z-DEVD-FMK (20 μM for 2 hours), as described above. The cells were then treated with Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours, followed by x-ray irradiation at a dose of 2 Gy. Cells were subsequently harvested, washed with cold PBS, and resuspended in annexin V binding buffer. The cells were stained with annexin V–fluorescein isothiocyanate (FITC) and PI and immediately analyzed by flow cytometry. The distributions of annexin V⁻/PI⁻, annexin V⁺/PI⁻, annexin V⁺/PI⁺, and annexin V⁻/PI⁺ cell populations were quantified to evaluate the effects of GSDME knockdown and caspase-3 inhibition on Hf-Ir/DON(+)–induced cell death.

LDH and IL-1β release

CT26 cells were seeded in six-well plates at 5 × 10⁵ cells per well and cultured overnight. Cells were treated with PBS, Hf-Ir, or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours, followed by x-ray irradiation (2 Gy). After a further 12-hour incubation at 37°C, culture supernatants were collected, and LDH release and IL-1β levels were quantified using commercial assay kits (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions.

ICD analysis

CRT exposure

CT26 cells were seeded in six-well plates or petri dishes at 5 × 10⁵ cells per well (or per dish) and cultured overnight. Cells were treated with PBS, Hf-Ir, or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours, followed by x-ray irradiation (2 Gy). After treatment, the medium was replaced with fresh complete medium, and cells were incubated for 12 hours. Then cells were stained with Alexa Fluor 488–conjugated anti-CRT antibody (Novus; 1:100) for 30 min, washed with PBS, and analyzed by flow cytometry. In parallel, CRT exposure was visualized by CLSM.

HMGB1 release

CT26 cells were seeded in six-well plates at 5 × 10⁵ cells per well and cultured overnight. Cells were treated with PBS, Hf-Ir, or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours, followed by x-ray irradiation (2 Gy). After which the medium was replaced with fresh complete medium and further incubated for 12 hours. Culture supernatants were collected and centrifuged (14,000g for 10 min) to remove debris. HMGB1 levels in clarified supernatants were quantified using an enzyme-linked immunosorbent assay (ELISA) kit (Chondrex, USA) according to the manufacturer’s instructions.

Extracellular ATP release

Extracellular ATP in culture supernatants from identically treated CT26 cells was quantified using an ATP assay kit (Invitrogen, USA) following the supplier’s protocol.

BMDC maturation analysis

BMDCs were generated from mouse bone marrow. Briefly, bone marrow cells were flushed from femurs and tibias, passed through a 70-μm cell strainer, and subjected to red blood cell lysis. Cells were cultured in RPMI 1640 complete medium supplemented with GM-CSF (20 ng/ml) and IL-4 (10 ng/ml), with medium refreshed every 2 days. Immature BMDCs were harvested on day 6 or 7 for subsequent experiments.

For coculture experiments, CT26 cells were treated with PBS, Hf-Ir, or Hf-Ir/DON (5 μM Hf and 1 μM DON) for 6 hours. Cells were then exposed to x-ray irradiation (2 Gy) and further incubated at 37°C for 12 hours. BMDCs were then added to the treated CT26 cells and cocultured for 12 hours at 37°C. After coculture, BMDCs were subsequently collected and stained with fluorescently labeled antibodies: Brilliant Violet 421 anti-CD11c (BioLegend; 1:100), FITC anti-CD80 (BioLegend; 1:100), and PE anti-CD86 (eBioscience; 1:200). The proportion of mDCs (CD11c⁺CD80⁺CD86⁺) was quantified by flow cytometry. In parallel, IL-6 and TNF-α levels in culture supernatants were measured using ELISA kits (Invitrogen, USA) according to the manufacturers’ instructions.

PBMC proliferation and metabolite analysis

PBMC isolation

PBMCs were isolated from murine peripheral blood. Briefly, peripheral blood was collected from the orbital venous sinus and transferred into heparinized anticoagulant tubes. The blood was carefully layered onto Ficoll-Paque PREMIUM (1.084) and centrifuged at 800g to isolate PBMCs. The PBMC layer was collected, washed with PBS, counted, and used immediately.

Tumor cell–BMDC–PBMC triple coculture

CT26 cells were seeded in six-well plates at 5 × 10⁵ cells per well and cultured overnight. Cells were treated with PBS, Hf-Ir, or Hf-Ir/DON (10 μM Hf and 2 μM DON) for 8 hours, washed with PBS, and then exposed to x-ray irradiation (0 or 2 Gy). Irradiated cells were further incubated at 37°C for 12 hours. BMDCs (5 × 10⁴) were then added and cocultured with tumor cells for an additional 12 hours at 37°C. PBMCs (5 × 10⁵) were subsequently added to establish a triple coculture system, which was maintained for 3 days. At the end point, cells were gently resuspended and collected for flow cytometry analysis using Live/Dead Yellow (Thermo Fisher Scientific; 1:200), APC/Cyanine7 anti-mouse CD45 antibody (BioLegend; 1:100), Alexa Fluor 488 anti-mouse CD3 antibody (BioLegend; 1:100), PE/Cyanine5 anti-mouse CD4 antibody (BioLegend; 1:100), Alexa Fluor 700 anti-mouse CD8a antibody (BioLegend; 1:100), Brilliant Violet 421 anti-mouse/human CD44 antibody (BioLegend; 1:100), Brilliant Violet 650 anti-mouse CD127 (IL-7Rα) antibody (BioLegend; 1:100), PerCP/Cyanine5.5 anti-mouse/human KLRG1 (MAFA) antibody (BioLegend; 1:100), and APC anti-mouse CD62L antibody (BioLegend; 1:100).

Ammonia, lactate, and Gln assays

On days 3 and 5 of the triple coculture, culture medium was collected and centrifuged at 12,000g to remove debris. Supernatants were deproteinized by adding an equal volume of 0.5 M perchloric acid, incubated on ice for 5 min, and centrifuged at 10,000g at 4°C. The supernatant was then neutralized to pH ∼ 7 with 2.5 M K₂CO₃, followed by a second centrifugation at 10,000g at 4°C to remove precipitated salts. The resulting clarified supernatants were used for metabolite analysis. Ammonia levels were determined using an ammonia assay kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. Lactate concentrations in the supernatants were measured using an l-lactate assay kit (Abcam). For Gln analysis, culture medium was diluted with 80% methanol in water and centrifuged, and the clarified supernatants were analyzed by LC-MS.

Tumor retention of DON

A total of 2 × 10⁶ CT26 cells were injected subcutaneously into the right dorsal flank of BALB/c mice. When tumors reached ∼200 mm³, mice were randomized into two groups (free DON and Hf-Ir/DON) for intratumoral administration. Tumors were collected at 0 hours, 15 min, 1 hour, 2 hours, 8 hours, 24 hours, 48 hours, 96 hours, and 168 hours (n = 3 per time point). For the 0-hour time point, tumors were excised, transferred into preweighed 15-ml tubes, and intratumorally injected with the corresponding formulation directly into the tubes. At each time point, tumors were harvested and weighed before processing.

For calibration, eight additional 0-hour tumors of comparable sizes were excised and intratumorally injected with Hf-Ir/DON to yield theoretical DON amounts of 200, 100, 50, 25, 12.5, 6.25, 3.125, and 0 μg per tumor, and these samples were processed identically to generate a standard curve. Tumor tissues were homogenized by ultrasonication. Because DON is unstable under acidic conditions, extraction was performed under mildly basic conditions (pH 8 to 9) using 20 mM NH₄HCO₃ containing 2 mM EDTA and shaken at room temperature for 2 to 4 hours. Samples were centrifuged at 12,000 rpm to remove debris, and supernatants were mixed with methanol to a final concentration of 80% (v/v) to precipitate proteins and salts, followed by centrifugation at 12,000 rpm for 15 min. Clarified supernatants (200 μl) were transferred to LC-MS vials for analysis. DON was quantified by integrated LC-MS peak area, and a calibration curve was constructed by plotting peak area versus injected DON amount. DON levels in tumor samples were calculated from this calibration curve.

Intracellular distribution of DON

The intracellular distribution of DON was quantitatively analyzed by LC-MS. Mitochondrial, cytosolic, and other membrane-associated fractions were isolated using a mitochondrial isolation kit from Thermo Fisher Scientific. The collected fractions were treated with 20 mM NH₄HCO₃ containing 1 mM EDTA and shaken at room temperature for 2 hours to disrupt the coordination between DON and Hf-Ir. The samples were then freeze dried, and the resulting powders were resuspended in methanol and subjected to ultrasonication. Insoluble protein precipitates and salts were removed by centrifugation at 12,000 rpm. The clarified supernatants (150 μl) were transferred to LC-MS vials for analysis. DON standards were prepared by serial dilution in methanol to generate a concentration calibration curve. DON was quantified on the basis of its LC-MS peak area, and the calibration curve was constructed by plotting the peak area against the injected amount of DON. The actual DON concentration in each sample was subsequently calculated from the corresponding calibration curve.

Antitumor efficacy in a CT26 tumor model

CT26 cells (2 × 10⁶) were injected subcutaneously into the right dorsal flank of BALB/c mice. When tumor reached ∼100 mm³, mice were randomized into six groups (n = 5 per group): (i) PBS, (ii) PBS(+), (iii) Hf-Ir, (iv) Hf-Ir(+), (v) Hf-Ir/DON, and (vi) Hf-Ir/DON(+). Treatments were administered by intratumoral injection at doses of Hf (50 μmol/kg) and DON (10 μmol/kg) in a total volume of 50 μl per tumor. Twenty-four hours later, mice were anesthetized with 2.5% isoflurane, and tumors were exposed to x-ray irradiation (2 Gy; 225-kVp, 13-mA, 0.3-mm Cu filter) every other day for a total of three fractions. Tumor volume (V = length × width²/2) and body weight were recorded every 2 days. Animals were humanely euthanized when tumor volume exceeded 1500 mm³ or according to institutional animal care guidelines. All procedures were approved by the institutional animal care and use committee. To assess treatment-associated antitumor immune protection, mice that achieved complete tumor regression following Hf-Ir/DON(+) or Hf-Ir/DON(+) plus αPD-L1 treatment were selected for tumor rechallenge. In Hf-Ir/DON(+) and Hf-Ir/DON(+) + αPD-L1 groups, treatments were administered by intratumoral injection at doses of Hf (50 μmol/kg) and DON (10 μmol/kg) in a total volume of 50 μl per tumor. Twenty-four hours later, mice were anesthetized with 2.5% isoflurane, and tumors were exposed to x-ray irradiation (4 Gy; 225-kVp, 13-mA, 0.3-mm Cu filter) every day for a total of three fractions, and, at the same time, intraperitoneal injection of αPD-L1 (15 mg/kg) was administered three times. Twelve days after the last irradiation (day 20 after the initial tumor inoculation), mice were subcutaneously reinoculated with CT26 cells (2 × 10⁶ cells per mouse) into the contralateral dorsal flanks without further treatment. Age-matched naïve BALB/c mice inoculated with the same number of CT26 cells served as controls (n = 5 per group). Rechallenged tumor growth and body weight were monitored every 2 days until day 36. Tumor volume was calculated as (V = length × width²/2). Animals were humanely euthanized when tumor volume exceeded 1500 mm³ or when institutional humane end-point criteria were reached.

Histological assay

Tumors and major organs were collected and immediately fixed in 4% paraformaldehyde (pH 7.2 to 7.4) at 4°C for 24 hours. After PBS rinsing, tissues were dehydrated, cleared in xylene, and embedded in paraffin. Paraffin blocks were sectioned at 5 μm, mounted onto adhesive-coated glass slides, and baked at 60°C for 30 min. After deparaffinization and rehydration, sections were subjected to H&E staining or immunohistochemistry (IHC) for γ-H2AX. For γ-H2AX staining, antigen retrieval was performed in tris-EDTA buffer (pH 9.0) at 95°C for 15 min. Endogenous peroxidase activity was quenched with 3% H₂O₂ for 10 min at room temperature, and sections were blocked with 5% normal serum for 30 min. Slides were incubated overnight at 4°C with an anti–γ-H2AX primary antibody (Cell Signaling Technology; 1:500), followed by an HRP-conjugated secondary antibody for 60 min at room temperature. Signals were developed with DAB, and whole-slide images were acquired using a Panoramic MIDI II digital slide scanner at ×40 magnification.

In vivo antitumor immune response analysis

When subcutaneous CT26 tumors reached ∼100 mm³, BALB/c mice were randomized into six groups (n = 5 per group): (i) PBS, (ii) PBS(+), (iii) Hf-Ir, (iv) Hf-Ir(+), (v) Hf-Ir/DON, and (vi) Hf-Ir/DON(+). Formulations were administered by intratumoral injection at doses of Hf (50 μmol/kg) and DON (10 μmol/kg) in a total volume of 50 μl per tumor. Twenty-four hours later, mice were anesthetized with 2.5% (v/v) isoflurane, and tumors were irradiated with x-rays (2 Gy per fraction, three total fractions). Mice were euthanized on day 8 after the final irradiation, and tumors were harvested for downstream analyses.

For flow cytometry, tumors were enzymatically dissociated in digestion buffer containing collagenase I (1 mg/ml), collagenase IV (0.25 mg/ml), and deoxyribonuclease I (0.05 mg/ml), followed by gentle mechanical disruption to obtain single-cell suspensions. Red blood cells were lysed with ACK lysis buffer (Thermo Fisher Scientific). Cells were then stained with fluorophore-conjugated antibodies and analyzed by flow cytometry. For intracellular marker staining, cells were fixed with 4% paraformaldehyde for 15 min, permeabilized with BD Perm/Wash buffer for 10 min, and then stained.

Tumor-infiltrating immune cells in CT26 tumors were stained with Live/Dead Yellow (Thermo Fisher Scientific; 1:200), Brilliant Violet 421 anti-mouse CD45 (BioLegend; 1:100), PE/Cyanine7 anti-mouse CD3ε (BioLegend; 1:100), Alexa Fluor 700 anti-mouse NK-1.1 (BioLegend; 1:100), APC anti-mouse B220 (BioLegend; 1:100), APC/Cyanine7 anti-mouse CD4 (BioLegend; 1:100), PerCP/Cyanine5.5 anti-mouse CD8a (BioLegend; 1:100), PE/Dazzle 594 anti-mouse CD44 (BioLegend; 1:100), Brilliant Violet 650 anti-mouse CD62L (BioLegend; 1:100), PE/Cyanine5 anti-mouse CD25 (BioLegend; 1:100), PE anti-mouse PD-1, and FITC anti-mouse granzyme B (BioLegend; 1:100). After washing, cells were resuspended in fluorescence-activated cell sorting (FACS) buffer, acquired on an Aurora full-spectrum flow cytometer, and analyzed accordingly.

Tumor-infiltrating immune cells in bilateral 4T1 tumor were stained with Live/Dead Yellow (Thermo Fisher Scientific; 1:200), Brilliant Violet 421 anti-mouse CD45 (BioLegend; 1:100), PE/Cyanine7 anti-mouse CD3ε (BioLegend; 1:100), Alexa Fluor 700 anti-mouse NK-1.1 (BioLegend; 1:100), APC anti-mouse B220 (BioLegend; 1:100), APC/Cyanine7 anti-mouse CD4 (BioLegend; 1:100), PerCP/Cyanine5.5 anti-mouse CD8a (BioLegend; 1:100), PE/Cyanine5 anti-mouse CD25 (BioLegend; 1:100), PE anti-mouse IFN-γ, and Pacific Blue anti-mouse FoxP3 (BioLegend; 1:100). After washing, cells were resuspended in FACS buffer, acquired on an Aurora full-spectrum flow cytometer, and analyzed accordingly.

Antitumor efficacy in a bilateral 4T1 tumor model

BALB/c mice were orthotopically inoculated into the mammary fat pads with 4T1 cells (2 × 10⁶ cells in the right mammary gland and 5 × 10⁵ cells in the left mammary gland). When the right-sided primary tumor reached ∼100 mm³, mice were randomized into five groups (n = 6 per group): (i) PBS, (ii) PBS(+), (iii) αPD-L1(+), (iv) Hf-Ir/DON(+), and (v) Hf-Ir/DON(+) plus αPD-L1. Hf-Ir/DON was administered by intratumoral injection at doses of Hf (50 μmol/kg) and DON (10 μmol/kg) in a total volume of 50 μl per tumor, and αPD-L1 was administered intraperitoneally three times at 15 mg/kg. Twenty-four hours later, mice were anesthetized with 2.5% (v/v) isoflurane, and the primary tumors were irradiated with x-rays (4 Gy per fraction, three total fractions). Primary and contralateral tumors, as well as body weight, were monitored every 2 days. Mice were euthanized when the total tumor burden exceeded 1500 mm³, after which tumors were collected and spleens were harvested and weighed. Tumor and organ tissues were further processed for H&E staining and γ-H2AX IHC.

Statistical analysis

Statistical analysis was performed using GraphPad Prism (10.0). Data are shown as means ± SD. Statistical significance was determined using one-way or two-way analysis of variance (ANOVA), followed by Tukey’s post hoc test for multiple comparisons.

Supplementary Materials

This PDF file includes:

Supplementary Text

Figs. S1 to S51

Tables S1 and S2