Cyclometalated iridium(III) complex based on isoquinoline alkaloid synergistically elicits the ICD response and IDO inhibition <i>via</i> autophagy-dependent ferroptosis.

The development of anticancer drugs to treat triple-negative breast cancer (TNBC) is an ongoing challenge. Immunogenic cell death (ICD) has garnered considerable interest worldwide as a promising synergistic modality for cancer chemoimmunotherapy. However, only few drugs or treatment modalities can trigger an ICD response and none of them exert a considerable clinical effect against TNBC. Therefore, new agents with potentially effective chemoimmunotherapeutic response are required. In this study, five new cyclometalated Ir(III) complexes containing isoquinoline alkaloid CˆN ligands were designed and synthesized. Among them, Ir-1 exhibited the highest in vitro cytotoxicity. Mechanistically, Ir-1 could trigger autophagy-dependent ferroptosis and a subsequent ferroptosis-dependent ICD response as well as indoleamine 2,3-dioxygenase (IDO) inhibition via reactive oxygen species (ROS)-mediated endoplasmic reticulum (ER) stress in MDA-MB-231 cells. When immunocompetent BALB/c mice were vaccinated with Ir-1 -treated dying TNBC cells, antitumor CD8 + T-cell response and Foxp3 + T-cell depletion were induced, resulting in long-lasting antitumor immunity in TNBC cells. Moreover, combination therapy with Ir-1 and anti-PD1 could substantially augment in vivo therapeutic effects. Based on these results, Ir-1 is a promising candidate for chemoimmunotherapy against TNBC and its effects are mediated synergistically via ICD induction and IDO blockage. ## Introduction 1 Introduction Globally, breast cancer is the most prevalent cancer among women. It has the highest incidence among the commonly diagnosed cancers reported in GLOBOCAN 2020 1 . Although notable progress has been made in breast cancer treatment over recent decades, the survival rate of patients with triple-negative breast cancer (TNBC), accounting for 10%–30% of all breast cancers 2 , is relatively low because of its high metastatic capacity and limited pharmacological treatment options. Presently, immunotherapy is revolutionizing the treatment of TNBC, for example, the early administration of immune-checkpoint inhibitors (ICIs) during the disease course of TNBC has proven therapeutically efficacious 3 , 4 . Moreover, the use of ICIs alongside chemotherapy can considerably improve the overall survival of patients with TNBC 5 , 6 . Accumulating clinical studies have reported that conventional chemotherapeutic drugs not only induce direct cell inhibition/cytotoxicity against fast-growing cancer cells but also reactivate tumor-specific immune responses 7 , 8 . One of the several ways by which chemotherapeutic drugs activate tumor-specific immune responses is by inducing immunogenic cell death (ICD), wherein dying cancer cells function as “anticancer vaccines”, eliciting a strong adaptive immune response 9 , 10 . To date, several clinical chemotherapeutic agents have exhibited promising ICD activity, including doxorubicin, cyclophosphamide, and oxaliplatin (Oxa) 11 , 12 . However, the current ICD-inducing drugs are insufficient because the induction of ICD can depend on tumor types. For example, Oxa cannot trigger ICD in non-small cell lung cancer 13 . Therefore, novel anticancer complexes that can be employed as more efficient ICD inducers are urgently required. In addition to Oxa, some metal complexes based on Pt 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , Ru 27 , 28 , 29 , 30 , Ir 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , Cu 40 , Au 41 , 42 , 43 , 44 , 45 , 46 , 47 , Mn 48 , and Re 49 reportedly promote the ICD effect against tumor cells. For example, Sen and coworkers reported a rationally designed redox-active metal complex, Au(I) N-heterocyclic carbene (NHC), that could induce ICD 42 . Our group designed a Pt(II) complex containing an aminophosphonate ester ligand that selectively accumulated in the endoplasmic reticulum (ER) and triggered type II ICD 14 . Additionally, although the concept of ICD was initially proposed in chemotherapy-induced apoptosis, ICD can also be triggered in the other modes of programmed cell deaths caused by radiotherapy, chemotherapy, photodynamic therapy, or other anticancer treatments 50 . Organometallic complexes have unique properties, including structural diversity, ligand exchange, redox and catalytic properties, making them promising drug candidates for cancer therapy 51 . Cyclometalated complexes are specific types of organometallic complexes that exhibit attractive biological properties. The anticancer effects of these complexes can primarily be ascribed to the CˆN ligands that form the cyclometalated backbone 52 , 53 . Under physiological conditions, the chelating feature of the CˆN ligand and the robust metal–carbon bond in these complexes ensure structural integrity, thereby enabling the targeting of the cancer cells without ligand detachment 54 . Recently, several cyclometalated complexes were reported to possess ICD-inducing capacity. For example, Tham et al. identified two cyclometalated Pt(II)-based NHC complexes as a new category of type II ICD inducers 19 . Among non-platinum cyclometalated compounds, iridium complexes have attracted considerable attention as ICD inducers. Wang et al. reported a cyclometalated Ir(III) complex containing an N , N -bis(2-chloroethyl)-azane derivate that targeted the ER and acted as a type II ICD inducer in human non-small cell lung cancer cells 31 , and another cyclometalated Ir(III) complex containing 2-phenylbenzo[d]-thiazole ligand as a promising ICD-inducing photosensitive agent in melanoma cells for two-photon photodynamic immunotherapy 36 . Vigueras et al. developed a cyclometalated Ir(III) complex based on benzimidazole that induced a type II ICD effect against melanoma upon blue-light irradiation 33 . Tan and Mao's group developed ferrocene-containing cyclometalated Ir(III) complexes that induced ICD in cancer cells 34 , 35 and a phenylquinoline-containing cyclometalated Ir(III) photosensitizer that induced ICD under hypoxia 37 . In addition, our group reported a cyclometalated Ir(III)−bisNHC complex as a novel ICD inducer 38 . Overall, cyclometalated Ir(III) complexes are potential candidates for developing more effective ICD inducers against TNBC tumors. To develop more active metal-based therapeutic agents with lower side effects, natural active products or their derivatives used as ligands can produce novel metal-based agents 55 , 56 . Previously, our group reported a series of metal complexes bearing natural active (iso)quinoline alkaloids or their derivative ligands as promising anticancer candidates 57 , 58 . As our continuing research, we focused on the isoquinoline alkaloid derivative used as the CˆN structure of cyclometalated Ir(III) complexes, which may change the lipophilicity and electronic properties of the complexes, thereby enhancing their druggability. Herein, a series of cationic cyclometalated Ir(III) complexes including an isoquinoline alkaloid derivative known as cyclometalated CˆN ligand and different bidentate NˆN ancillary ligands were designed, synthesized, and characterized. The cationic cyclometalated Ir(III) complexes exhibited higher potency against TNBC cell lines than cisplatin. This study demonstrates that our complex has the following advantages: 1) novel approach for TNBC chemoimmunotherapy via a new cyclometalated Ir(III) complex using isoquinoline alkaloid as a CˆN ligand, 2) a new regulatory mechanism of ICD mediated by autophagy-dependent ferroptosis, and 3) a synergistically elicited chemoimmunotherapy response achieved by combining ICD and indoleamine 2,3-dioxygenase (IDO) inhibition. This study is expected to create new avenues for the development of novel iridium-based chemoimmunotherapy agents against TNBC. ## Results and discussion 2 Results and discussion 2.1 Synthesis and characterization To develop metal-based scaffolds with novel anticancer mechanisms, cyclometalated Ir(III) complexes ( Ir-1 – Ir-5 ) were designed bearing two identical CˆN isoquinoline derivative ligands and structurally diverse NˆN ligands ( Scheme 1 ). For the isoquinoline ligand 3,4-methylenedioxy-1-phenylisoquinoline (L a ) 59 , a phenyl group was introduced at the 1-position of the isoquinoline ring to facilitate the formation of the CˆN structure with iridium as the metal center. Furthermore, methoxy cyclopentane, a commonly used pharmacophore group, was designed to regulate the lipophilicity and bioactivity of the Ir(III) complexes. For the structurally diverse NˆN ligands, Ir-1 carries a substituted 2,2-dipyridinine amine scaffold (L 1 ), Ir-2 carries the 1,10-phenanthroline NˆN ligand (L 2 ), Ir-3 bears the 2-(2-pyridyl) benzimidazole ligand (L 3 ), Ir-4 contains the unsubstituted 2,2-bipyridine ligand (L 4 ), and Ir-5 contains the 2-aminomethylpyridine ligand (L 5 ). To the best of our knowledge, this study is the first to report these five complexes. The synthetic routes for these iridium(III) complexes are depicted in Scheme 1 . All new complexes were characterized by 1 H/ 13 C nuclear magnetic resonance (NMR) and electrospray ionization mass spectrometry (ESI-MS; see Supporting Information Experimental section and Figs. S1–S3 ). In subsequent studies, Ir-1 presented superior biological activity; hence, Ir-1 was further characterized via single crystal X-ray diffraction analysis ( Fig. 1 and Supporting Information Table S1 ). The Ir(III) center is coordinated by a distorted octahedral geometry with two carbon atoms from L a and two nitrogen atoms from auxiliary ligands L 1 in a cis orientation, which forms the equatorial coordination plane, while the axial positions are occupied by nitrogen atoms from two isoquinoline ligands L a . The stability of complexes Ir-1 – Ir-5 at room temperature was confirmed via HPLC analysis in TBS. The HPLC chromatograms at different times reveal no observable changes during the 24- and 48-h time courses, indicating that these cyclometalated Ir(III) complexes were stable in TBS for the tested time course ( Supporting Information Fig. S4 ). Unless otherwise specified, the metal complexes were dissolved in dimethylformamide (DMF) to prepare a 2 mmol/L stock solution, and then diluted into aqueous solution for subsequent experiments. Scheme 1 Synthetic route and structures of Ir-1 – Ir-5 . Reagent and conditions: (i) IrCl 3 ·3H 2 O, 2-ME/H 2 O, N 2 , 1 h, 120 °C; (ii) N^N ligand, (CH 2 OH) 2 , N 2 , 2 h, 120 °C, and then KPF 6 , for the synthesis of Ir-1 – Ir-4 ; N^N ligand, CH 3 OH, Na 2 CO 3 , N 2 , 65 °C, 30 min, for the synthesis of Ir-5 . Scheme 1 Figure 1 X-ray crystal structure of Ir-1 . All hydrogen atoms have been removed for clarity. CCDC number: 2289402. Figure 1 2.2 In vitro cytotoxic activity To investigate the in vitro anticancer activity of these cyclometalated Ir(III) complexes, the cytotoxicity of Ir-1 – Ir-5 and the isoquinoline ligand (L a ) was assessed against non-small cell lung carcinoma (NCl–H460, A549), human TNBC (MDA-MB-231), and human gastric adenocarcinoma (MGC-803) cells via 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays. Cisplatin was employed as a positive control. After 48 h of treatment with each compound, the IC 50 values were derived ( Table 1 ). The isoquinoline alkaloid ligand L a did not show considerable cytotoxicity (IC 50  > 20 μmol/L) against the tested cancer cell lines; however, all Ir(III) complexes ( Ir-1 – Ir-5 ) exhibited excellent cytotoxicity (IC 50  = 1.73–8.62 μmol/L), indicating a synergistic effect derived from the coordination between the Ir(III) ion and ligands. Additionally, Ir-1 – Ir-5 demonstrated considerably higher cytotoxic activity than cisplatin (IC 50  = 12.23–18.54 μmol/L) against the tested cell lines. Notably, Ir-1 and Ir-2 showed higher cytotoxic activity than the other three complexes, indicating that auxiliary ligands also play a substantial role in regulating the cytotoxicity of these iridium complexes bearing the isoquinoline alkaloid CˆN ligand. The differences in cytotoxicity among Ir-1 – Ir-5 may be attributed to the varying lipophilicity or electronic effects of their NˆN ligands. Furthermore, Ir-1 – Ir-5 exhibited more sensitive cytotoxicity against the MDA-MB-231 cells than that against the other tested cell lines overall. Consequently, the MDA-MB-231 cell line was selected as the representative cancer cell line for further investigation of anticancer mechanism. Table 1 IC 50 values (μmol/L) of Ir(III) complexes, cisplatin, and ligand L a toward four human tumor cell lines. Table 1 Complex NCl–H460 A549 MGC-803 MDA-MB-231 Ir-1 2.03 ± 0.15 2.40 ± 0.12 2.60 ± 0.14 1.82 ± 0.10 Ir-2 2.12 ± 0.11 2.75 ± 0.20 2.85 ± 0.17 1.73 ± 0.14 Ir-3 7.78 ± 0.18 8.45 ± 0.35 8.62 ± 0.26 7.14 ± 0.25 Ir-4 3.10 ± 0.12 4.20 ± 0.16 3.90 ± 0.16 2.50 ± 0.17 Ir-5 5.70 ± 0.10 6.27 ± 0.24 6.40 ± 0.33 3.13 ± 0.22 Cisplatin 18.29 ± 0.62 18.54 ± 0.68 12.33 ± 0.43 12.23 ± 0.46 L a >20 >20 >20 >20 Data are presented as mean ± SD ( n  = 3). 2.3 Induction of ICD activity by Ir-1 As Ir-1 – Ir-5 exhibited excellent in vitro cytotoxicity against MDA-MB-231 cells and cyclometalated Ir(III) complexes had demonstrated ICD characteristics, we subsequently sought to discover whether the newly synthesized cyclometalated complexes could induce ICD in MDA-MB-231 cells. A key endogenous signal that induces the immunogenicity of tumor cells is the extracellular secretion of high mobility group protein 1 (HMGB1) from the nucleus 60 . We investigated the levels of HMGB1 in MDA-MB-231 cell supernatants after treating the cells with Ir-1 – Ir-5 (4 μmol/L). Ir-1 and Ir-2 could significantly induce the release of HMGB1 from MDA-MB-231 cells ( Fig. 2 A), although Ir-1 induced a higher level of HMGB1 release than Ir-2 . Notably, after Ir-1 treatment, the extracellular level of HMGB1 showed a concentration-dependent increase ( Fig. 2 B), consistent with the concentration-dependent decreasing trend of intracellular HMGB1 levels ( Fig. 2 D). Overall, these results provide support for the suggestion that Ir-1 may be a potential ICD inducer. The Ir-1 -induced ICD activity and related mechanisms were subsequently investigated in MDA-MB-231 cells. Figure 2 ICD activity induced by Ir-1 . (A, B) Release of HMGB1 from MDA-MB-231 cells after treatment with (A) Ir-1 – Ir-5 (4 μmol/L) and (B) Ir-1 (1–4 μmol/L) for 12 h, respectively. (C) Representative confocal images showing CRT exposure after treating MDA-MB-231 cells with Ir-1 for 12 h. Scale bar = 20 μm. (D) Western blots of intracellular HMGB1 after treating MDA-MB-231 cells with Ir-1 for 12 h. (E) Release of ATP from MDA-MB-231 cells after treatment with Ir-1 for 12 h. Data are presented as mean ± SD ( n  = 3). ∗ P  < 0.05, ∗∗ P  < 0.01, ∗∗∗ P  < 0.001, as compared with the control group. Figure 2 Another dominant damage-associated molecular pattern (DAMP) signal of ICD is the exposure of calreticulin (CRT) on the outer cell surface 60 . To assess the Ir-1 -induced ICD activity, MDA-MB-231 cells were incubated with this complex and labeled with an anti-CRT antibody. Consequently, the immunofluorescence of CRT (green) was enhanced on the cell membrane surface (red) in a concentration-dependent manner ( Fig. 2 C), which indicated that potent CRT exposure had been induced in Ir-1 -treated MDA-MB-231 cells. As anticipated, the secretion of ATP, another DAMP signal of ICD, was also observed in a concentration-dependent manner up to 215 nmol/L, after treatment with 4 μmol/L Ir-1 ( Fig. 2 E). Overall, these results imply that Ir-1 is a potential candidate of ICD inducer, which encouraged us to investigate related mechanisms underlying the induction of ICD activity by Ir-1 in MDA-MB-231 cells. 2.4 Reactive oxygen species (ROS) production and ER stress response induced by Ir-1 ER stress is essential for initiating ICD 9 , 10 , 61 . In addition to factors including nutrient deficiency, infection, and hypoxia, ROS generation can contribute to ER stress 62 . Several cyclometalated Ir(III) complexes have also been implicated in the induction of considerable ROS production and ER stress response 31 , 63 , 64 , 65 . Consequently, we investigated whether Ir-1 could similarly cause ROS production and trigger an ER stress response. Initially, we sought to assess the ROS generation in Ir-1 -induced MDA-MB-231 cells using a commercial ROS fluorescent indicator DCFH-DA 66 . Confocal micrographs revealed a time-dependent increase in ROS levels ( Fig. 3 A), suggesting that ROS might stimulate an ER stress response. Subsequently, we investigated whether Ir-1 could induce ER stress. Alterations in expression levels of three ER-stress-related marker proteins, namely phosphorylated protein kinase RNA-like ER kinase (p-PERK), phosphorylated eukaryotic initiation factor 2 α (p-eIF2 α ), and C/EBP homologous protein (CHOP), were determined in Ir-1 -treated MDA-MB-231 cells. The expression levels of p-eIF2 α , p-PERK, and CHOP were elevated ( Fig. 3 B), which are typical signs of ER stress. To delve deeper into whether an ER stress response was induced by ROS generation, an ROS scavenger, N -acetyl- l -cysteine (NAC), was co-incubated, and Western blotting of ER-stress-related proteins was conducted. The results revealed that alterations in the expression levels of p-eIF2 α , p-PERK, and CHOP triggered by Ir-1 were inhibited after pretreating the MDA-MB-231 cells with NAC ( Fig. 3 C). This suggests that the ER stress induced by Ir-1 is dependent on ROS production. To explore the effect of ROS generation on Ir-1 -induced cytotoxicity in MDA-MB-231 cells, we conducted an MTT assay using NAC as an ROS scavenger. There was a dose-dependent increase in cell viability, which increased from ∼48% ( Ir-1 -treated group) to ∼90% (10 mmol/L NAC- and Ir-1 -treated group; Fig. 3 D). This implies that ROS production considerably contributes to the induction of cell death. Collectively, these findings suggest that Ir-1 triggers cell death through an ROS-mediated mechanism, accompanied by the induction of an ER stress response. Figure 3 Induction of ROS production and ER stress by Ir-1 . (A) Representative confocal micrographs of ROS production in MDA-MB-231 cells induced by Ir-1 (at IC 50 ). Scale bar = 20 μm. (B) Western blots of ER-stress-related proteins after incubating MDA-MB-231 cells with Ir-1 . (C) Western blots of ER-stress-related proteins after co-incubating MDA-MB-231 cells with NAC (10 mmol/L) and Ir-1 (at IC 50 ) for 48 h. (D) Viability of MDA-MB-231 cells treated with NAC and Ir-1 (at IC 50 ). Data are presented as mean ± SD ( n  = 3). ∗ P <  0.05, ∗∗ P  < 0.01, ∗∗∗ P  < 0.001, as compared with the indicated group. Figure 3 2.5 Modes of cell death triggered by Ir-1 ROS reportedly induces several types of cell death, including autophagy, ferroptosis, and apoptosis 67 . To determine the primary mode of cell death caused by the ROS effects induced by Ir-1 , we employed several inhibitors targeting cell death-associated mechanisms to modulate the viability of cells treated with Ir-1 . The commonly used apoptosis inhibitor Z-VAD-FMK did not significantly affect the viability of Ir-1 -treated cells ( Fig. 4 A). 3-Methyladenine (3-MA) functions as an autophagy inhibitor 68 . After pretreating cells with 3-MA (0.5, 1, and 2 mmol/L), Ir-1 -induced cell death was dose-dependently blocked ( Fig. 4 B). The viability of Ir-1 -treated cells increased by ∼20% upon pretreatment with 2 mmol/L 3-MA, suggesting that the mode of cell death induced by Ir-1 is related to autophagy. Furthermore, MDA-MB-231 cells pretreated with a ferroptosis inhibitor, ferrostatin-1 (Fer-1), showed that Fer-1 could significantly prevent Ir-1 -induced cell death ( Fig. 4 C), suggesting that Ir-1 induces ferroptosis as a mode of cell death as well. Thus, these results provide support for the suggestion that ferroptosis and autophagy play a primary role in regulating the cell death modes induced by Ir-1 . Figure 4 Induction of autophagy and ferroptosis by Ir-1 . (A–C) Viability of MDA-MB-231 cells co-treated with (A) Z-VAD-FMK, (B) 3-MA, or (C) Fer-1, respectively, and Ir-1 (at IC 50 ). (D–F) Western blots of (D) apoptosis, (E) autophagy, or (F) ferroptosis markers after incubating MDA-MB-231 cells with Ir-1 . (G) Representative confocal micrographs showing autophagic vesicles (green) in Ir-1 -treated MDA-MB-231 cells. Nuclei were stained with DAPI (blue). Scale bar = 20 μm. (H) Representative confocal micrographs showing lipid peroxidation in Ir-1 -treated MDA-MB-231 cells. The non-oxidized lipids (red) and oxidized lipids (green) were labeled with C11-BODIPY 581/591 . Scale bar = 20 μm. Data are presented as mean ± SD ( n  = 3). ∗ P  < 0.05, as compared with the indicated group. Figure 4 Subsequently, we conducted western blotting to assess the expression levels of cell death-associated marker proteins in Ir-1 -treated MDA-MB-231 cells. The cleavage of pro-caspase-3, commonly employed as an indicator of apoptosis, was examined. Treatment with Ir-1 did not considerably affect the expression level of pro-caspase-3, indicating that apoptosis was not associated with Ir-1 -induced tumor cell death ( Fig. 4 D). As a common molecular marker of autophagy, the microtubule-associated protein light chain 3 (LC3) undergoes conversion from type I to type II during autophagy 69 . The results reveal an increase in the LC3-II level and LC3-II/LC3-I ratio after Ir-1 treatment ( Fig. 4 E), suggesting that Ir-1 can induce autophagy. Additionally, autophagy was assessed through confocal microscopy analysis using a green dye that selectively labels autophagic vesicles. The formation of autophagic vesicles in Ir-1 -treated MDA-MB-231 cells increased, indicating the occurrence of autophagy ( Fig. 4 G). Solute carrier family 7 membrane 11 (SLC7A11) and glutathione peroxidase 4 (GPX4) are the key regulators of ferroptosis 70 . Inhibiting SLC7A11 and GPX4 activity can cause lipid peroxidation, ultimately leading to ferroptosis 71 . Western blotting revealed that the expression levels of SLC7A11 and GPX4 time-dependently decreased following Ir-1 treatment ( Fig. 4 F), suggesting that Ir-1 induced ferroptosis in MDA-MB-231 cells. As a key feature of ferroptosis, lipid peroxidation was assessed in MDA-MB-231 cells using a commercial dye C11-BODIPY 581/591 70 . Following incubation with Ir-1 , MDA-MB-231 cells exhibited enhanced green fluorescence signals ( Fig. 4 H), indicating the induction of lipid peroxidation by Ir-1 . As a prominent hallmark of lipid peroxidation, the production of malondialdehyde (MDA) in Ir-1 -treated MDA-MB-231 cells were also significantly increased compared with that in the control ( Supporting Information Fig. S5 ). These results provide support for the suggestion that Ir-1 could induce cell death via autophagy and ferroptosis. 2.6 Relation of ROS-mediated autophagy, ferroptosis, and ICD To confirm whether autophagy and ferroptosis were mediated by the ROS production induced by Ir-1 , NAC was employed to block ROS generation. After pretreating MDA-MB-231 cells with NAC, the Ir-1 -induced changes in the LC3-II level and LC3-II/LC3-I ratio were blocked ( Fig. 5 A). Simultaneously, the downregulated expression levels of SLC7A11 and GPX4 induced by Ir-1 were reversed ( Fig. 5 B). Overall, these results indicate that autophagy and ferroptosis in MDA-MB-231 are dependent on the ROS production induced by Ir-1 . Figure 5 Relation of ROS-mediated autophagy, ferroptosis, ICD, and IDO inhibition. (A, B) Western blots of (A) autophagy or (B) ferroptosis markers after co-incubating MDA-MB-231 cells with NAC (10 mmol/L) and Ir-1 (at IC 50 ) for 48 h. (C) Effects of inhibiting autophagy (by 2 mmol/L 3-MA) on the ferroptosis induced by Ir-1 (at IC 50 ) in MDA-MB-231 cells. (D) Effects of inhibiting ferroptosis (by 2 μmol/L Fer-1) on the autophagy induced by Ir-1 (at IC 50 ) in MDA-MB-231 cells. (E) Percentage of cells expressing surface-CRT after co-treating MDA-MB-231 cells with 3-MA (2 mmol/L) or Fer-1 (2 μmol/L) and Ir-1 (6 μmol/L) for 12 h, as determined by flow cytometry. (F, G) Secretion of (F) ATP and (G) HMGB1 from MDA-MB-231 cells after co-treatment with Ir-1 (at IC 50 ) and 3-MA (2 mmol/L) or Fer-1 (2 μmol/L) for 12 h. (H) Western blots of IDO after incubating MDA-MB-231 cells with Ir-1 . (I) Western blots of IDO after co-incubating MDA-MB-231 cells with NAC (10 mmol/L) and Ir-1 (at IC 50 ). Data are presented as mean ± SD ( n  = 3). ∗∗ P  < 0.01, ∗∗∗ P  < 0.001, as compared with the indicated group. Figure 5 Recent studies have emphasized ferroptosis as a form of cell death that is dependent on autophagy 59 , 72 , 73 . To explore the potential relation between the autophagy and ferroptosis induced by Ir-1 , we investigated the expression levels of related proteins in MDA-MB-231 cells in the presence of 3-MA and Fer-1, respectively. Ir-1 could reduce the expression levels of SLC7A11 and GPX4 ( Fig. 5 C); however, this effect was significantly reversed after pretreatment with the autophagy inhibitor 3-MA. Conversely, the LC3-II level and LC3-II/LC3-I ratio did not significantly change after treatment with the ferroptosis inhibitor Fer-1 ( Fig. 5 D). Overall, Ir-1 induces autophagy-dependent ferroptosis in MDA-MB-231 cells. Recently, it has been reported that ferroptotic cells release ATP and HMGB1, two key DAMP signals of ICD 74 , 75 . Therefore, we investigated whether Ir-1 -induced ICD is dependent on ferroptosis in MDA-MB-231 cells. Subsequently, we examined the changes in CRT exposure, ATP secretion, and HMGB1 release after pretreatment with ferroptosis inhibitor Fer-1. Immunofluorescence intensity of surface-CRT significantly decreased when cells were co-incubated with Fer-1 and Ir-1 ( Fig. 5 E and S6 ). Moreover, ATP and HMGB1 secretion induced by Ir-1 was significantly blocked by Fer-1 ( Fig. 5 F and G). These results suggest that inhibiting ferroptosis blocks the Ir-1 -induced ICD activity. As anticipated, the autophagy inhibitor 3-MA also reversed the exposure or release of three ICD biomarkers induced by Ir-1 ( Fig. 5 E–G and Supporting Information Fig. S6 ). Overall, Ir-1 can induce autophagy-dependent ferroptosis and ferroptosis-dependent ICD activity. 2.7 IDO inhibition induced by Ir-1 As a crucial negative feedback protein, IDO plays a pivotal role in shaping an immunosuppressive microenvironment conducive to tumor cell proliferation 76 . Notably, IDO overexpression in certain cancer cells causes the breakdown of tryptophan (TRP) and increase in TRP metabolites, causing the arrest of the cell cycle, demise of effector T cells, and proliferation of regulatory T cells 77 , 78 , 79 . A recent study indicated that IDO could inhibit ferroptosis 80 . IDO catalyzes the oxidation of TRP to kynurenine (KYN), which is subsequently converted to downstream metabolites that suppress ferroptosis. Thus, we investigated whether Ir-1 -induced ferroptosis is related to the downregulation of IDO. The expression level of IDO in Ir-1 -treated MDA-MB-231 cells was markedly downregulated ( Fig. 5 H). To explore the relation between ROS production and IDO expression downregulation, MDA-MB-231 cells were pretreated with NAC. The results revealed that the Ir-1 -induced downregulation of IDO expression was blocked ( Fig. 5 I), demonstrating that Ir-1 -induced ROS production could block IDO expression. These results provide support for the suggestion that Ir-1 could inhibit ROS-mediated IDO expression. 2.8 Antitumor vaccination in vivo MDA-MB-231 cells are a type of human-derived cancer cells; therefore, syngenetic TNBC 4T1 cells were selected for the subsequent in vivo experiments. As the gold standard of ICD evaluation, a vaccination model of 4T1 cells was established. Oxa was used as a positive control. In a first study, 2 × 10 6 4T1 cells were pretreated with different formulations (150 μmol/L Oxa, 20 μmol/L Ir-1 , or only solvent) for 6 h. Subsequently, these cells were injected (s.c.) as a tumor vaccine into the left flanks of BALB/c mice ( n  = 10 per group). To observe the differences in tumor development between different groups during the rapid proliferation of 4T1 cells, after 7 days (Day 0), 1 × 10 4 untreated 4T1 cells were rechallenged in the right flanks of the mice. For 36 days, tumor formation in the right flanks was closely observed daily ( Fig. 6 A); mice that exhibited no tumor formation in their right flanks were documented as tumor-free mice. In the control group, tumors were observed in 20% of mice by Day 8 after rechallenge, and 100% of mice in this group developed tumors in the right flanks by Day 20. By contrast, the Oxa and Ir-1 treatment groups exhibited a notable delay in tumor development after rechallenge ( Fig. 6 B). The first tumor in the Oxa group was detected on Day 10, whereas the first tumor in the Ir-1 group appeared on Day 14. After 22 days, Oxa treatment produced a stable population of 30% tumor-free mice, whereas Ir-1 treatment exhibited a slower progression of tumorigenesis, ultimately achieving a stable population of 50% tumor-free mice after 28 days ( Fig. 6 C and D). In contrast to the control group, mice that were administered the Ir-1 -treated vaccination cells exhibited significant inhibition of tumor growth. On the 36th day after inoculation, the average tumor volume in the Ir-1 group was 274 mm 3 , which was 2.31- and 3.45-fold smaller than those in the Oxa-treated (634 mm 3 ) and control (945 mm 3 ) groups, respectively. Furthermore, there were no significant differences observed in the body weights of mice among the different groups ( Fig. 6 E). These results demonstrate that Ir-1 is a bona fide ICD inducer. Figure 6 Anticancer vaccination. (A) ICD vaccine experiment of Ir-1 using the TNBC 4T1 cell line. (B) Percentage of tumor-free mice following rechallenge with 4T1 cells. (C) Image of representative tumors dissected from the right flanks of mice 36 days after rechallenge. (D) Curves of tumor volume during treatments as determined via vaccination assays. (E) Curves of body weight during treatments as determined via vaccination assays. Data are presented as mean ± SD ( n  = 10). Figure 6 To confirm that differences in tumor growth were caused by in vivo immune responses, we used time-of-flight mass cytometry (CyTOF) to conduct high-dimensional characterization of tumors, spleens, lymph nodes, and peripheral blood samples from each group of mice in the vaccination assay. Live immune cells were first sorted using the CD45 + marker antibody through a series of gating steps. Then, CD8 + T cells, CD4 + T cells, B cells, and macrophages/monocytes were identified based on classic markers ( Supporting Information Table S2 ) from the live immune cell clusters. The two-dimensional t -stochastic neighbor embedding ( t SNE) projections were constructed to compare the distinct landscapes of immune cells from the control and Oxa- and Ir-1 -treated groups ( Fig. 7 A–D). Figure 7 Analysis of immune cells using CyTOF. (A–D) tSNE maps showing distinct immune cell population in (A) tumors, (B) spleens, (C) lymph nodes, and (D) peripheral blood derived from each group of mice in the vaccination assay. (E–H) Quantitative data analysis of cell proportions in the corresponding mice tissues in (A–D). (I–L) Heatmaps showing the expression of distinct markers in (I) tumors, (J) spleens, (K) lymph nodes, and (L) peripheral blood. Data are presented as mean ± SD ( n  = 3). ∗ P <  0.05, ∗∗ P  < 0.01, ∗∗∗ P  < 0.001, as compared with the Ir-1 group. Figure 7 One of the key ICD responses in vivo is the activation of CD8 + T cells within tumor sites 81 . CD38 + is the marker protein of T cell activation 82 , whereas the expression level of Ki67 serves as an indicator of the extent of cell proliferation 83 . We observed an increase in the proportions of CD3 + CD8 + , CD3 + CD8 + CD38 + , and CD3 + CD8 + Ki67 + T cells in the four tested tissues after treatment with Ir-1 or Oxa compared with those in the tissues treated with the control solvent ( Fig. 7 A–L). Specifically, in tumor tissues, the proportion of CD3 + CD8 + T cell exhibited a significant increase, rising from 0.24% in the control group to 1.83% in the Ir-1 treatment group and from 9.83% to 42.00%, and 31.32%–42.65% for CD3 + CD8 + CD38 + and CD3 + CD8 + Ki67 + T cells, respectively, which were all higher than the proportions of these cells in the Oxa-treated group (0.88%, 18.95%, and 37.41%, respectively; Fig. 7 E and I). These results indicate that CD3 + CD8 + T cells were fully activated following vaccination with Ir-1 -treated cells. The proportions of cells expressing the cytokines tumor necrosis factor α (TNF- α ), interferon γ (IFN- γ ), and granzyme B (GZMB) reflect the killing capacity of CD8 + T cells against tumor cells 84 . Higher percentages of cells expressing TNF- α (6.95%), IFN- γ (26.90%), and GZMB (26.30%) in CD3 + CD8 + T cells were observed in the tumors treated with Ir-1 compared with those in the control (3.92%, 14.53%, and 11.04%, respectively); these levels were also higher than those in the Oxa-treated group (5.65%, 17.88%, and 14.64%, respectively; Fig. 7 I). CD69 + is a marker used to identify tissue resident memory T cells (T RM ), which play a crucial role in providing long-term surveillance of tumor recurrence 85 . Compared with the control group, a notable increasing trend was observed in the proportion of CD69 + cells in CD3 + CD8 + T cells of tumors, spleens, and lymph nodes from the Ir-1 group, which were also slightly higher than the proportions of cells from these tissues in the Oxa-treated group ( Fig. 7 I–K). These results further support that the ICD effect of cancer cells treated with Ir-1 can activate the immune system to combat cancer, exhibiting a greater activity than that elicited by Oxa. CD4 + T cells are also crucial for regulating the immune responses of tumors 86 . The proportion of CD3 + CD4 + T cells in tumors significantly increased from 1.31% in the control group to 5.97% in the Ir-1 -treated group, and the proportion of Th cells (CD4 + CD185 + ) also increased from 2.86% in the control group to 7.68% in the Ir-1 -treated group; these proportions were slightly higher than those in the Oxa-treated group (4.71% and 6.85%, respectively; Fig. 7 E and I). Treg cells (CD3 + CD4 + Foxp3 + T cells) can inhibit the function of cytotoxic T cells by secreting immunosuppressive cytokines 87 . Compared with that in control group, a significant decrease of Treg cells in Ir-1- treated group was observed ( Fig. 7 I–L). The proportion of Treg cells in tumor tissues significantly decreased from 10.74% in the control group to 4.49% in the Ir-1 treatment group ( Fig. 7 I). Overall, vaccination using cells treated with Ir-1 can promote immune responses by downregulating Treg cells. 2.9 In vivo therapeutic effect of Ir-1 After elucidating the mechanism by which Ir-1 activates anticancer immunity, the anticancer efficacy of Ir-1 was further validated in vivo using a murine model. First, the safety of Ir-1 was assessed in BALB/c mice, with four mice per group. Ir-1 was administered intravenously through the retro-orbital venous sinus once every 2 days at doses of 20, 30, and 40 mg/kg, respectively. While all mice survived, those administered with 40 mg/kg Ir-1 exhibited a degree of loss in body weight compared with that in the control group ( Supporting Information Fig. S7 ). Considering the changes in mice behavior and body weight, the 40 mg/kg dose was selected for the subsequent experiment. We explored the therapeutic efficacy of Ir-1 in a murine model of metastatic cancer in vivo . Studies have proven that ICD inducers have the potential to synergistically enhance the efficacy of PD-1/PD-L1 inhibitors 88 , 89 , 90 . Therefore, we wondered whether Ir-1 therapy would synergize with PD-1 blockade to enhance antitumor effect. When the 4T1 tumor volume reached ∼20 mm 3 (Day 0), the treatments, including anti-PD1 (10 mg/kg), Oxa (7 mg/kg), Ir-1 (40 mg/kg), or anti-PD1 (10 mg/kg) and Ir-1 (40 mg/kg) combination therapy, were injected (i.v.) into BALB/c mice bearing 4T1 tumors through the retro-orbital venous sinus once every 2 days. The selected doses and administration frequency were based on the condition of the mice and the safety profiles of the compounds. Under these conditions, while anti-PD1, Oxa, and Ir-1 demonstrated significant tumor growth inhibition compared with the vehicle, the in vivo antitumor effect of Ir-1 was superior to that of anti-PD1 or Oxa alone ( Fig. 8 A). Importantly, Ir-1 and anti-PD1 combination therapy exerted the most pronounced antitumor effect ( Fig. 8 A). A notable disparity in body weight was evident in the Ir-1 -treated group compared with that in the Oxa-treated group, with the Oxa-treated group exhibiting lower body weights than the Ir-1 -treated group ( Fig. 8 B). These findings further emphasize the minimal side effects of Ir-1 treatment. Upon completing the experiment, we collected and weighed the tumors, and determined the inhibition rate of tumor growth (IRT). Ir-1 exhibited an IRT of 42.1%, surpassing those of Oxa (29.6%) and anti-PD1 (26.0%). Notably, Ir-1 and anti-PD1 combination therapy demonstrated the highest IRT of 66.2% ( Fig. 8 C). These differences can also be observed from the image of tumors presented in Fig. 8 D. Thus, Ir-1 and anti-PD1 combination therapy exhibits significantly enhanced therapeutic effects in vivo while exhibiting reduced side effects compared with Oxa or anti-PD1 alone. Figure 8 Anticancer therapy. (A) Tumor growth profiles of BALB/c mice bearing 4T1 tumors. (B) Effects of the indicated compounds or vehicle treatment on the body weight of BALB/c mice bearing 4T1 tumors. (C) Tumor weights and IRT values in BALB/c mice bearing 4T1 tumors after treatments on Day 19. (D) Image of representative tumors collected from the in vivo antitumor experiment. Data are presented as mean ± SD ( n  = 6). Figure 8 ## Synthesis and characterization 2.1 Synthesis and characterization To develop metal-based scaffolds with novel anticancer mechanisms, cyclometalated Ir(III) complexes ( Ir-1 – Ir-5 ) were designed bearing two identical CˆN isoquinoline derivative ligands and structurally diverse NˆN ligands ( Scheme 1 ). For the isoquinoline ligand 3,4-methylenedioxy-1-phenylisoquinoline (L a ) 59 , a phenyl group was introduced at the 1-position of the isoquinoline ring to facilitate the formation of the CˆN structure with iridium as the metal center. Furthermore, methoxy cyclopentane, a commonly used pharmacophore group, was designed to regulate the lipophilicity and bioactivity of the Ir(III) complexes. For the structurally diverse NˆN ligands, Ir-1 carries a substituted 2,2-dipyridinine amine scaffold (L 1 ), Ir-2 carries the 1,10-phenanthroline NˆN ligand (L 2 ), Ir-3 bears the 2-(2-pyridyl) benzimidazole ligand (L 3 ), Ir-4 contains the unsubstituted 2,2-bipyridine ligand (L 4 ), and Ir-5 contains the 2-aminomethylpyridine ligand (L 5 ). To the best of our knowledge, this study is the first to report these five complexes. The synthetic routes for these iridium(III) complexes are depicted in Scheme 1 . All new complexes were characterized by 1 H/ 13 C nuclear magnetic resonance (NMR) and electrospray ionization mass spectrometry (ESI-MS; see Supporting Information Experimental section and Figs. S1–S3 ). In subsequent studies, Ir-1 presented superior biological activity; hence, Ir-1 was further characterized via single crystal X-ray diffraction analysis ( Fig. 1 and Supporting Information Table S1 ). The Ir(III) center is coordinated by a distorted octahedral geometry with two carbon atoms from L a and two nitrogen atoms from auxiliary ligands L 1 in a cis orientation, which forms the equatorial coordination plane, while the axial positions are occupied by nitrogen atoms from two isoquinoline ligands L a . The stability of complexes Ir-1 – Ir-5 at room temperature was confirmed via HPLC analysis in TBS. The HPLC chromatograms at different times reveal no observable changes during the 24- and 48-h time courses, indicating that these cyclometalated Ir(III) complexes were stable in TBS for the tested time course ( Supporting Information Fig. S4 ). Unless otherwise specified, the metal complexes were dissolved in dimethylformamide (DMF) to prepare a 2 mmol/L stock solution, and then diluted into aqueous solution for subsequent experiments. Scheme 1 Synthetic route and structures of Ir-1 – Ir-5 . Reagent and conditions: (i) IrCl 3 ·3H 2 O, 2-ME/H 2 O, N 2 , 1 h, 120 °C; (ii) N^N ligand, (CH 2 OH) 2 , N 2 , 2 h, 120 °C, and then KPF 6 , for the synthesis of Ir-1 – Ir-4 ; N^N ligand, CH 3 OH, Na 2 CO 3 , N 2 , 65 °C, 30 min, for the synthesis of Ir-5 . Scheme 1 Figure 1 X-ray crystal structure of Ir-1 . All hydrogen atoms have been removed for clarity. CCDC number: 2289402. Figure 1 2.2 In vitro cytotoxic activity To investigate the in vitro anticancer activity of these cyclometalated Ir(III) complexes, the cytotoxicity of Ir-1 – Ir-5 and the isoquinoline ligand (L a ) was assessed against non-small cell lung carcinoma (NCl–H460, A549), human TNBC (MDA-MB-231), and human gastric adenocarcinoma (MGC-803) cells via 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays. Cisplatin was employed as a positive control. After 48 h of treatment with each compound, the IC 50 values were derived ( Table 1 ). The isoquinoline alkaloid ligand L a did not show considerable cytotoxicity (IC 50  > 20 μmol/L) against the tested cancer cell lines; however, all Ir(III) complexes ( Ir-1 – Ir-5 ) exhibited excellent cytotoxicity (IC 50  = 1.73–8.62 μmol/L), indicating a synergistic effect derived from the coordination between the Ir(III) ion and ligands. Additionally, Ir-1 – Ir-5 demonstrated considerably higher cytotoxic activity than cisplatin (IC 50  = 12.23–18.54 μmol/L) against the tested cell lines. Notably, Ir-1 and Ir-2 showed higher cytotoxic activity than the other three complexes, indicating that auxiliary ligands also play a substantial role in regulating the cytotoxicity of these iridium complexes bearing the isoquinoline alkaloid CˆN ligand. The differences in cytotoxicity among Ir-1 – Ir-5 may be attributed to the varying lipophilicity or electronic effects of their NˆN ligands. Furthermore, Ir-1 – Ir-5 exhibited more sensitive cytotoxicity against the MDA-MB-231 cells than that against the other tested cell lines overall. Consequently, the MDA-MB-231 cell line was selected as the representative cancer cell line for further investigation of anticancer mechanism. Table 1 IC 50 values (μmol/L) of Ir(III) complexes, cisplatin, and ligand L a toward four human tumor cell lines. Table 1 Complex NCl–H460 A549 MGC-803 MDA-MB-231 Ir-1 2.03 ± 0.15 2.40 ± 0.12 2.60 ± 0.14 1.82 ± 0.10 Ir-2 2.12 ± 0.11 2.75 ± 0.20 2.85 ± 0.17 1.73 ± 0.14 Ir-3 7.78 ± 0.18 8.45 ± 0.35 8.62 ± 0.26 7.14 ± 0.25 Ir-4 3.10 ± 0.12 4.20 ± 0.16 3.90 ± 0.16 2.50 ± 0.17 Ir-5 5.70 ± 0.10 6.27 ± 0.24 6.40 ± 0.33 3.13 ± 0.22 Cisplatin 18.29 ± 0.62 18.54 ± 0.68 12.33 ± 0.43 12.23 ± 0.46 L a >20 >20 >20 >20 Data are presented as mean ± SD ( n  = 3). ## Induction of ICD activity by 2.3 Induction of ICD activity by Ir-1 As Ir-1 – Ir-5 exhibited excellent in vitro cytotoxicity against MDA-MB-231 cells and cyclometalated Ir(III) complexes had demonstrated ICD characteristics, we subsequently sought to discover whether the newly synthesized cyclometalated complexes could induce ICD in MDA-MB-231 cells. A key endogenous signal that induces the immunogenicity of tumor cells is the extracellular secretion of high mobility group protein 1 (HMGB1) from the nucleus 60 . We investigated the levels of HMGB1 in MDA-MB-231 cell supernatants after treating the cells with Ir-1 – Ir-5 (4 μmol/L). Ir-1 and Ir-2 could significantly induce the release of HMGB1 from MDA-MB-231 cells ( Fig. 2 A), although Ir-1 induced a higher level of HMGB1 release than Ir-2 . Notably, after Ir-1 treatment, the extracellular level of HMGB1 showed a concentration-dependent increase ( Fig. 2 B), consistent with the concentration-dependent decreasing trend of intracellular HMGB1 levels ( Fig. 2 D). Overall, these results provide support for the suggestion that Ir-1 may be a potential ICD inducer. The Ir-1 -induced ICD activity and related mechanisms were subsequently investigated in MDA-MB-231 cells. Figure 2 ICD activity induced by Ir-1 . (A, B) Release of HMGB1 from MDA-MB-231 cells after treatment with (A) Ir-1 – Ir-5 (4 μmol/L) and (B) Ir-1 (1–4 μmol/L) for 12 h, respectively. (C) Representative confocal images showing CRT exposure after treating MDA-MB-231 cells with Ir-1 for 12 h. Scale bar = 20 μm. (D) Western blots of intracellular HMGB1 after treating MDA-MB-231 cells with Ir-1 for 12 h. (E) Release of ATP from MDA-MB-231 cells after treatment with Ir-1 for 12 h. Data are presented as mean ± SD ( n  = 3). ∗ P  < 0.05, ∗∗ P  < 0.01, ∗∗∗ P  < 0.001, as compared with the control group. Figure 2 Another dominant damage-associated molecular pattern (DAMP) signal of ICD is the exposure of calreticulin (CRT) on the outer cell surface 60 . To assess the Ir-1 -induced ICD activity, MDA-MB-231 cells were incubated with this complex and labeled with an anti-CRT antibody. Consequently, the immunofluorescence of CRT (green) was enhanced on the cell membrane surface (red) in a concentration-dependent manner ( Fig. 2 C), which indicated that potent CRT exposure had been induced in Ir-1 -treated MDA-MB-231 cells. As anticipated, the secretion of ATP, another DAMP signal of ICD, was also observed in a concentration-dependent manner up to 215 nmol/L, after treatment with 4 μmol/L Ir-1 ( Fig. 2 E). Overall, these results imply that Ir-1 is a potential candidate of ICD inducer, which encouraged us to investigate related mechanisms underlying the induction of ICD activity by Ir-1 in MDA-MB-231 cells. ## Reactive oxygen species (ROS) production and ER stress response induced by 2.4 Reactive oxygen species (ROS) production and ER stress response induced by Ir-1 ER stress is essential for initiating ICD 9 , 10 , 61 . In addition to factors including nutrient deficiency, infection, and hypoxia, ROS generation can contribute to ER stress 62 . Several cyclometalated Ir(III) complexes have also been implicated in the induction of considerable ROS production and ER stress response 31 , 63 , 64 , 65 . Consequently, we investigated whether Ir-1 could similarly cause ROS production and trigger an ER stress response. Initially, we sought to assess the ROS generation in Ir-1 -induced MDA-MB-231 cells using a commercial ROS fluorescent indicator DCFH-DA 66 . Confocal micrographs revealed a time-dependent increase in ROS levels ( Fig. 3 A), suggesting that ROS might stimulate an ER stress response. Subsequently, we investigated whether Ir-1 could induce ER stress. Alterations in expression levels of three ER-stress-related marker proteins, namely phosphorylated protein kinase RNA-like ER kinase (p-PERK), phosphorylated eukaryotic initiation factor 2 α (p-eIF2 α ), and C/EBP homologous protein (CHOP), were determined in Ir-1 -treated MDA-MB-231 cells. The expression levels of p-eIF2 α , p-PERK, and CHOP were elevated ( Fig. 3 B), which are typical signs of ER stress. To delve deeper into whether an ER stress response was induced by ROS generation, an ROS scavenger, N -acetyl- l -cysteine (NAC), was co-incubated, and Western blotting of ER-stress-related proteins was conducted. The results revealed that alterations in the expression levels of p-eIF2 α , p-PERK, and CHOP triggered by Ir-1 were inhibited after pretreating the MDA-MB-231 cells with NAC ( Fig. 3 C). This suggests that the ER stress induced by Ir-1 is dependent on ROS production. To explore the effect of ROS generation on Ir-1 -induced cytotoxicity in MDA-MB-231 cells, we conducted an MTT assay using NAC as an ROS scavenger. There was a dose-dependent increase in cell viability, which increased from ∼48% ( Ir-1 -treated group) to ∼90% (10 mmol/L NAC- and Ir-1 -treated group; Fig. 3 D). This implies that ROS production considerably contributes to the induction of cell death. Collectively, these findings suggest that Ir-1 triggers cell death through an ROS-mediated mechanism, accompanied by the induction of an ER stress response. Figure 3 Induction of ROS production and ER stress by Ir-1 . (A) Representative confocal micrographs of ROS production in MDA-MB-231 cells induced by Ir-1 (at IC 50 ). Scale bar = 20 μm. (B) Western blots of ER-stress-related proteins after incubating MDA-MB-231 cells with Ir-1 . (C) Western blots of ER-stress-related proteins after co-incubating MDA-MB-231 cells with NAC (10 mmol/L) and Ir-1 (at IC 50 ) for 48 h. (D) Viability of MDA-MB-231 cells treated with NAC and Ir-1 (at IC 50 ). Data are presented as mean ± SD ( n  = 3). ∗ P <  0.05, ∗∗ P  < 0.01, ∗∗∗ P  < 0.001, as compared with the indicated group. Figure 3 ## Modes of cell death triggered by 2.5 Modes of cell death triggered by Ir-1 ROS reportedly induces several types of cell death, including autophagy, ferroptosis, and apoptosis 67 . To determine the primary mode of cell death caused by the ROS effects induced by Ir-1 , we employed several inhibitors targeting cell death-associated mechanisms to modulate the viability of cells treated with Ir-1 . The commonly used apoptosis inhibitor Z-VAD-FMK did not significantly affect the viability of Ir-1 -treated cells ( Fig. 4 A). 3-Methyladenine (3-MA) functions as an autophagy inhibitor 68 . After pretreating cells with 3-MA (0.5, 1, and 2 mmol/L), Ir-1 -induced cell death was dose-dependently blocked ( Fig. 4 B). The viability of Ir-1 -treated cells increased by ∼20% upon pretreatment with 2 mmol/L 3-MA, suggesting that the mode of cell death induced by Ir-1 is related to autophagy. Furthermore, MDA-MB-231 cells pretreated with a ferroptosis inhibitor, ferrostatin-1 (Fer-1), showed that Fer-1 could significantly prevent Ir-1 -induced cell death ( Fig. 4 C), suggesting that Ir-1 induces ferroptosis as a mode of cell death as well. Thus, these results provide support for the suggestion that ferroptosis and autophagy play a primary role in regulating the cell death modes induced by Ir-1 . Figure 4 Induction of autophagy and ferroptosis by Ir-1 . (A–C) Viability of MDA-MB-231 cells co-treated with (A) Z-VAD-FMK, (B) 3-MA, or (C) Fer-1, respectively, and Ir-1 (at IC 50 ). (D–F) Western blots of (D) apoptosis, (E) autophagy, or (F) ferroptosis markers after incubating MDA-MB-231 cells with Ir-1 . (G) Representative confocal micrographs showing autophagic vesicles (green) in Ir-1 -treated MDA-MB-231 cells. Nuclei were stained with DAPI (blue). Scale bar = 20 μm. (H) Representative confocal micrographs showing lipid peroxidation in Ir-1 -treated MDA-MB-231 cells. The non-oxidized lipids (red) and oxidized lipids (green) were labeled with C11-BODIPY 581/591 . Scale bar = 20 μm. Data are presented as mean ± SD ( n  = 3). ∗ P  < 0.05, as compared with the indicated group. Figure 4 Subsequently, we conducted western blotting to assess the expression levels of cell death-associated marker proteins in Ir-1 -treated MDA-MB-231 cells. The cleavage of pro-caspase-3, commonly employed as an indicator of apoptosis, was examined. Treatment with Ir-1 did not considerably affect the expression level of pro-caspase-3, indicating that apoptosis was not associated with Ir-1 -induced tumor cell death ( Fig. 4 D). As a common molecular marker of autophagy, the microtubule-associated protein light chain 3 (LC3) undergoes conversion from type I to type II during autophagy 69 . The results reveal an increase in the LC3-II level and LC3-II/LC3-I ratio after Ir-1 treatment ( Fig. 4 E), suggesting that Ir-1 can induce autophagy. Additionally, autophagy was assessed through confocal microscopy analysis using a green dye that selectively labels autophagic vesicles. The formation of autophagic vesicles in Ir-1 -treated MDA-MB-231 cells increased, indicating the occurrence of autophagy ( Fig. 4 G). Solute carrier family 7 membrane 11 (SLC7A11) and glutathione peroxidase 4 (GPX4) are the key regulators of ferroptosis 70 . Inhibiting SLC7A11 and GPX4 activity can cause lipid peroxidation, ultimately leading to ferroptosis 71 . Western blotting revealed that the expression levels of SLC7A11 and GPX4 time-dependently decreased following Ir-1 treatment ( Fig. 4 F), suggesting that Ir-1 induced ferroptosis in MDA-MB-231 cells. As a key feature of ferroptosis, lipid peroxidation was assessed in MDA-MB-231 cells using a commercial dye C11-BODIPY 581/591 70 . Following incubation with Ir-1 , MDA-MB-231 cells exhibited enhanced green fluorescence signals ( Fig. 4 H), indicating the induction of lipid peroxidation by Ir-1 . As a prominent hallmark of lipid peroxidation, the production of malondialdehyde (MDA) in Ir-1 -treated MDA-MB-231 cells were also significantly increased compared with that in the control ( Supporting Information Fig. S5 ). These results provide support for the suggestion that Ir-1 could induce cell death via autophagy and ferroptosis. ## Relation of ROS-mediated autophagy, ferroptosis, and ICD 2.6 Relation of ROS-mediated autophagy, ferroptosis, and ICD To confirm whether autophagy and ferroptosis were mediated by the ROS production induced by Ir-1 , NAC was employed to block ROS generation. After pretreating MDA-MB-231 cells with NAC, the Ir-1 -induced changes in the LC3-II level and LC3-II/LC3-I ratio were blocked ( Fig. 5 A). Simultaneously, the downregulated expression levels of SLC7A11 and GPX4 induced by Ir-1 were reversed ( Fig. 5 B). Overall, these results indicate that autophagy and ferroptosis in MDA-MB-231 are dependent on the ROS production induced by Ir-1 . Figure 5 Relation of ROS-mediated autophagy, ferroptosis, ICD, and IDO inhibition. (A, B) Western blots of (A) autophagy or (B) ferroptosis markers after co-incubating MDA-MB-231 cells with NAC (10 mmol/L) and Ir-1 (at IC 50 ) for 48 h. (C) Effects of inhibiting autophagy (by 2 mmol/L 3-MA) on the ferroptosis induced by Ir-1 (at IC 50 ) in MDA-MB-231 cells. (D) Effects of inhibiting ferroptosis (by 2 μmol/L Fer-1) on the autophagy induced by Ir-1 (at IC 50 ) in MDA-MB-231 cells. (E) Percentage of cells expressing surface-CRT after co-treating MDA-MB-231 cells with 3-MA (2 mmol/L) or Fer-1 (2 μmol/L) and Ir-1 (6 μmol/L) for 12 h, as determined by flow cytometry. (F, G) Secretion of (F) ATP and (G) HMGB1 from MDA-MB-231 cells after co-treatment with Ir-1 (at IC 50 ) and 3-MA (2 mmol/L) or Fer-1 (2 μmol/L) for 12 h. (H) Western blots of IDO after incubating MDA-MB-231 cells with Ir-1 . (I) Western blots of IDO after co-incubating MDA-MB-231 cells with NAC (10 mmol/L) and Ir-1 (at IC 50 ). Data are presented as mean ± SD ( n  = 3). ∗∗ P  < 0.01, ∗∗∗ P  < 0.001, as compared with the indicated group. Figure 5 Recent studies have emphasized ferroptosis as a form of cell death that is dependent on autophagy 59 , 72 , 73 . To explore the potential relation between the autophagy and ferroptosis induced by Ir-1 , we investigated the expression levels of related proteins in MDA-MB-231 cells in the presence of 3-MA and Fer-1, respectively. Ir-1 could reduce the expression levels of SLC7A11 and GPX4 ( Fig. 5 C); however, this effect was significantly reversed after pretreatment with the autophagy inhibitor 3-MA. Conversely, the LC3-II level and LC3-II/LC3-I ratio did not significantly change after treatment with the ferroptosis inhibitor Fer-1 ( Fig. 5 D). Overall, Ir-1 induces autophagy-dependent ferroptosis in MDA-MB-231 cells. Recently, it has been reported that ferroptotic cells release ATP and HMGB1, two key DAMP signals of ICD 74 , 75 . Therefore, we investigated whether Ir-1 -induced ICD is dependent on ferroptosis in MDA-MB-231 cells. Subsequently, we examined the changes in CRT exposure, ATP secretion, and HMGB1 release after pretreatment with ferroptosis inhibitor Fer-1. Immunofluorescence intensity of surface-CRT significantly decreased when cells were co-incubated with Fer-1 and Ir-1 ( Fig. 5 E and S6 ). Moreover, ATP and HMGB1 secretion induced by Ir-1 was significantly blocked by Fer-1 ( Fig. 5 F and G). These results suggest that inhibiting ferroptosis blocks the Ir-1 -induced ICD activity. As anticipated, the autophagy inhibitor 3-MA also reversed the exposure or release of three ICD biomarkers induced by Ir-1 ( Fig. 5 E–G and Supporting Information Fig. S6 ). Overall, Ir-1 can induce autophagy-dependent ferroptosis and ferroptosis-dependent ICD activity. ## IDO inhibition induced by 2.7 IDO inhibition induced by Ir-1 As a crucial negative feedback protein, IDO plays a pivotal role in shaping an immunosuppressive microenvironment conducive to tumor cell proliferation 76 . Notably, IDO overexpression in certain cancer cells causes the breakdown of tryptophan (TRP) and increase in TRP metabolites, causing the arrest of the cell cycle, demise of effector T cells, and proliferation of regulatory T cells 77 , 78 , 79 . A recent study indicated that IDO could inhibit ferroptosis 80 . IDO catalyzes the oxidation of TRP to kynurenine (KYN), which is subsequently converted to downstream metabolites that suppress ferroptosis. Thus, we investigated whether Ir-1 -induced ferroptosis is related to the downregulation of IDO. The expression level of IDO in Ir-1 -treated MDA-MB-231 cells was markedly downregulated ( Fig. 5 H). To explore the relation between ROS production and IDO expression downregulation, MDA-MB-231 cells were pretreated with NAC. The results revealed that the Ir-1 -induced downregulation of IDO expression was blocked ( Fig. 5 I), demonstrating that Ir-1 -induced ROS production could block IDO expression. These results provide support for the suggestion that Ir-1 could inhibit ROS-mediated IDO expression. ## Antitumor vaccination 2.8 Antitumor vaccination in vivo MDA-MB-231 cells are a type of human-derived cancer cells; therefore, syngenetic TNBC 4T1 cells were selected for the subsequent in vivo experiments. As the gold standard of ICD evaluation, a vaccination model of 4T1 cells was established. Oxa was used as a positive control. In a first study, 2 × 10 6 4T1 cells were pretreated with different formulations (150 μmol/L Oxa, 20 μmol/L Ir-1 , or only solvent) for 6 h. Subsequently, these cells were injected (s.c.) as a tumor vaccine into the left flanks of BALB/c mice ( n  = 10 per group). To observe the differences in tumor development between different groups during the rapid proliferation of 4T1 cells, after 7 days (Day 0), 1 × 10 4 untreated 4T1 cells were rechallenged in the right flanks of the mice. For 36 days, tumor formation in the right flanks was closely observed daily ( Fig. 6 A); mice that exhibited no tumor formation in their right flanks were documented as tumor-free mice. In the control group, tumors were observed in 20% of mice by Day 8 after rechallenge, and 100% of mice in this group developed tumors in the right flanks by Day 20. By contrast, the Oxa and Ir-1 treatment groups exhibited a notable delay in tumor development after rechallenge ( Fig. 6 B). The first tumor in the Oxa group was detected on Day 10, whereas the first tumor in the Ir-1 group appeared on Day 14. After 22 days, Oxa treatment produced a stable population of 30% tumor-free mice, whereas Ir-1 treatment exhibited a slower progression of tumorigenesis, ultimately achieving a stable population of 50% tumor-free mice after 28 days ( Fig. 6 C and D). In contrast to the control group, mice that were administered the Ir-1 -treated vaccination cells exhibited significant inhibition of tumor growth. On the 36th day after inoculation, the average tumor volume in the Ir-1 group was 274 mm 3 , which was 2.31- and 3.45-fold smaller than those in the Oxa-treated (634 mm 3 ) and control (945 mm 3 ) groups, respectively. Furthermore, there were no significant differences observed in the body weights of mice among the different groups ( Fig. 6 E). These results demonstrate that Ir-1 is a bona fide ICD inducer. Figure 6 Anticancer vaccination. (A) ICD vaccine experiment of Ir-1 using the TNBC 4T1 cell line. (B) Percentage of tumor-free mice following rechallenge with 4T1 cells. (C) Image of representative tumors dissected from the right flanks of mice 36 days after rechallenge. (D) Curves of tumor volume during treatments as determined via vaccination assays. (E) Curves of body weight during treatments as determined via vaccination assays. Data are presented as mean ± SD ( n  = 10). Figure 6 To confirm that differences in tumor growth were caused by in vivo immune responses, we used time-of-flight mass cytometry (CyTOF) to conduct high-dimensional characterization of tumors, spleens, lymph nodes, and peripheral blood samples from each group of mice in the vaccination assay. Live immune cells were first sorted using the CD45 + marker antibody through a series of gating steps. Then, CD8 + T cells, CD4 + T cells, B cells, and macrophages/monocytes were identified based on classic markers ( Supporting Information Table S2 ) from the live immune cell clusters. The two-dimensional t -stochastic neighbor embedding ( t SNE) projections were constructed to compare the distinct landscapes of immune cells from the control and Oxa- and Ir-1 -treated groups ( Fig. 7 A–D). Figure 7 Analysis of immune cells using CyTOF. (A–D) tSNE maps showing distinct immune cell population in (A) tumors, (B) spleens, (C) lymph nodes, and (D) peripheral blood derived from each group of mice in the vaccination assay. (E–H) Quantitative data analysis of cell proportions in the corresponding mice tissues in (A–D). (I–L) Heatmaps showing the expression of distinct markers in (I) tumors, (J) spleens, (K) lymph nodes, and (L) peripheral blood. Data are presented as mean ± SD ( n  = 3). ∗ P <  0.05, ∗∗ P  < 0.01, ∗∗∗ P  < 0.001, as compared with the Ir-1 group. Figure 7 One of the key ICD responses in vivo is the activation of CD8 + T cells within tumor sites 81 . CD38 + is the marker protein of T cell activation 82 , whereas the expression level of Ki67 serves as an indicator of the extent of cell proliferation 83 . We observed an increase in the proportions of CD3 + CD8 + , CD3 + CD8 + CD38 + , and CD3 + CD8 + Ki67 + T cells in the four tested tissues after treatment with Ir-1 or Oxa compared with those in the tissues treated with the control solvent ( Fig. 7 A–L). Specifically, in tumor tissues, the proportion of CD3 + CD8 + T cell exhibited a significant increase, rising from 0.24% in the control group to 1.83% in the Ir-1 treatment group and from 9.83% to 42.00%, and 31.32%–42.65% for CD3 + CD8 + CD38 + and CD3 + CD8 + Ki67 + T cells, respectively, which were all higher than the proportions of these cells in the Oxa-treated group (0.88%, 18.95%, and 37.41%, respectively; Fig. 7 E and I). These results indicate that CD3 + CD8 + T cells were fully activated following vaccination with Ir-1 -treated cells. The proportions of cells expressing the cytokines tumor necrosis factor α (TNF- α ), interferon γ (IFN- γ ), and granzyme B (GZMB) reflect the killing capacity of CD8 + T cells against tumor cells 84 . Higher percentages of cells expressing TNF- α (6.95%), IFN- γ (26.90%), and GZMB (26.30%) in CD3 + CD8 + T cells were observed in the tumors treated with Ir-1 compared with those in the control (3.92%, 14.53%, and 11.04%, respectively); these levels were also higher than those in the Oxa-treated group (5.65%, 17.88%, and 14.64%, respectively; Fig. 7 I). CD69 + is a marker used to identify tissue resident memory T cells (T RM ), which play a crucial role in providing long-term surveillance of tumor recurrence 85 . Compared with the control group, a notable increasing trend was observed in the proportion of CD69 + cells in CD3 + CD8 + T cells of tumors, spleens, and lymph nodes from the Ir-1 group, which were also slightly higher than the proportions of cells from these tissues in the Oxa-treated group ( Fig. 7 I–K). These results further support that the ICD effect of cancer cells treated with Ir-1 can activate the immune system to combat cancer, exhibiting a greater activity than that elicited by Oxa. CD4 + T cells are also crucial for regulating the immune responses of tumors 86 . The proportion of CD3 + CD4 + T cells in tumors significantly increased from 1.31% in the control group to 5.97% in the Ir-1 -treated group, and the proportion of Th cells (CD4 + CD185 + ) also increased from 2.86% in the control group to 7.68% in the Ir-1 -treated group; these proportions were slightly higher than those in the Oxa-treated group (4.71% and 6.85%, respectively; Fig. 7 E and I). Treg cells (CD3 + CD4 + Foxp3 + T cells) can inhibit the function of cytotoxic T cells by secreting immunosuppressive cytokines 87 . Compared with that in control group, a significant decrease of Treg cells in Ir-1- treated group was observed ( Fig. 7 I–L). The proportion of Treg cells in tumor tissues significantly decreased from 10.74% in the control group to 4.49% in the Ir-1 treatment group ( Fig. 7 I). Overall, vaccination using cells treated with Ir-1 can promote immune responses by downregulating Treg cells. 2.9 In vivo therapeutic effect of Ir-1 After elucidating the mechanism by which Ir-1 activates anticancer immunity, the anticancer efficacy of Ir-1 was further validated in vivo using a murine model. First, the safety of Ir-1 was assessed in BALB/c mice, with four mice per group. Ir-1 was administered intravenously through the retro-orbital venous sinus once every 2 days at doses of 20, 30, and 40 mg/kg, respectively. While all mice survived, those administered with 40 mg/kg Ir-1 exhibited a degree of loss in body weight compared with that in the control group ( Supporting Information Fig. S7 ). Considering the changes in mice behavior and body weight, the 40 mg/kg dose was selected for the subsequent experiment. We explored the therapeutic efficacy of Ir-1 in a murine model of metastatic cancer in vivo . Studies have proven that ICD inducers have the potential to synergistically enhance the efficacy of PD-1/PD-L1 inhibitors 88 , 89 , 90 . Therefore, we wondered whether Ir-1 therapy would synergize with PD-1 blockade to enhance antitumor effect. When the 4T1 tumor volume reached ∼20 mm 3 (Day 0), the treatments, including anti-PD1 (10 mg/kg), Oxa (7 mg/kg), Ir-1 (40 mg/kg), or anti-PD1 (10 mg/kg) and Ir-1 (40 mg/kg) combination therapy, were injected (i.v.) into BALB/c mice bearing 4T1 tumors through the retro-orbital venous sinus once every 2 days. The selected doses and administration frequency were based on the condition of the mice and the safety profiles of the compounds. Under these conditions, while anti-PD1, Oxa, and Ir-1 demonstrated significant tumor growth inhibition compared with the vehicle, the in vivo antitumor effect of Ir-1 was superior to that of anti-PD1 or Oxa alone ( Fig. 8 A). Importantly, Ir-1 and anti-PD1 combination therapy exerted the most pronounced antitumor effect ( Fig. 8 A). A notable disparity in body weight was evident in the Ir-1 -treated group compared with that in the Oxa-treated group, with the Oxa-treated group exhibiting lower body weights than the Ir-1 -treated group ( Fig. 8 B). These findings further emphasize the minimal side effects of Ir-1 treatment. Upon completing the experiment, we collected and weighed the tumors, and determined the inhibition rate of tumor growth (IRT). Ir-1 exhibited an IRT of 42.1%, surpassing those of Oxa (29.6%) and anti-PD1 (26.0%). Notably, Ir-1 and anti-PD1 combination therapy demonstrated the highest IRT of 66.2% ( Fig. 8 C). These differences can also be observed from the image of tumors presented in Fig. 8 D. Thus, Ir-1 and anti-PD1 combination therapy exhibits significantly enhanced therapeutic effects in vivo while exhibiting reduced side effects compared with Oxa or anti-PD1 alone. Figure 8 Anticancer therapy. (A) Tumor growth profiles of BALB/c mice bearing 4T1 tumors. (B) Effects of the indicated compounds or vehicle treatment on the body weight of BALB/c mice bearing 4T1 tumors. (C) Tumor weights and IRT values in BALB/c mice bearing 4T1 tumors after treatments on Day 19. (D) Image of representative tumors collected from the in vivo antitumor experiment. Data are presented as mean ± SD ( n  = 6). Figure 8 ## Conclusions 3 Conclusions We designed and synthesized five cyclometalated Ir(III) complexes containing two isoquinoline ligands and an N,N -heterocyclic ligand. All Ir(III) complexes exhibited markedly high cytotoxicity (at 1.73–8.62 μmol/L), which was higher than that exhibited by cisplatin against the tested human cancer cell lines. Notably, Ir-1 triggered an ICD effect in MDA-MB-231 cells by inducing ROS-mediated ER stress. The dominant mode of cell death contributing to the cytotoxicity induced by Ir-1 was identified as autophagy-dependent ferroptosis mediated by ROS-induced ER stress. The ferroptosis inhibitor Fer-1 mitigated Ir-1 -induced CRT exposure, ATP secretion, and HMGB1 release, indicating that Ir-1 induces ferroptosis-dependent ICD in MDA-MB-231 cells. Moreover, the IDO blocking induced by Ir-1 was mediated by ROS. In vivo results obtained from a tumor vaccination model of 4T1 cells further confirmed that Ir-1 was a more efficient ICD inducer than Oxa and could significantly induce a change in the in vivo immune microenvironment. In the in vivo tumor growth experiment, Ir-1 exhibited superior anticancer activity and lower toxicity compared with those exhibited by Oxa. Importantly, Ir-1 and anti-PD1 combination therapy significantly enhanced in vivo therapeutic effects. Our study findings indicate that Ir-1 is a promising chemoimmunotherapy candidate or model compound for developing anticancer drugs. ## Experimental 4 Experimental 4.1 Synthesis and characterization The ligand L a was firstly synthesized following a procedure described in our previous study 59 . Then, the general procedure for synthesizing compounds Ir-1 – Ir-5 is outlined as follows. The cyclometalated iridium(III) chloro-bridged dimer (c) was prepared according to a previously reported method 91 . IrCl 3 ·3H 2 O (5 mmol) and ligand L a (10 mmol) were added to a 6 mL mixture of 2-methoxylethanol (2-ME) and water (3:1, v / v ). The mixture underwent reflux at 120 °C under a nitrogen atmosphere for 1 h. After filtering the solution, the filtrate was dried at 60 °C to yield compound (c). Then, as previously described 92 , 0.14 mol of compound (c) and 0.3 mmol of NˆN ligand (L 1 , L 2 , L 3 or L 4 ) were added to 5 mL of ethylene glycol, and the mixture was refluxed for 2 h at 120 °C under a nitrogen atmosphere. Once the solution cooled to room temperature, an excess of KPF 6 solution was added, resulting in the formation of a precipitate. The solid precipitate was then filtered and left to dry at 60 °C overnight. The purification of the crude product was performed on a silica gel column (EA/MeOH, 9:1 v / v ), yielding a bright-orange product ( Ir-1 , Ir-2 , Ir-3 , or Ir-4 ). In a separate reaction, 0.14 mol of compound (c) and 0.3 mmol of NˆN ligand (L 5 ) were added to 10 mL of methanol. Subsequently, 0.6 mmol Na 2 CO 3 was added, and the mixture was refluxed at 65 °C under a nitrogen atmosphere for 30 min. The as-obtained solution was concentrated, and the crude product was purified on a silica gel column (EA/MeOH, 9:1 v / v ) to obtain the bright-orange product ( Ir-5 ). Data for Ir-1 : Yield 60%. 1 H NMR (400 MHz, DMSO- d 6 ) δ 10.64 (s, 1H), 8.21 (s, 2H), 8.14 (d, J  = 8.0 Hz, 2H), 8.04 (d, J  = 6.4 Hz, 2H), 7.82 (ddd, J  = 8.7, 7.2, 1.8 Hz, 2H), 7.64–7.53 (m, 4H), 7.41–7.30 (m, 4H), 7.09–6.94 (m, 2H), 6.78 (td, J  = 7.7, 2.1 Hz, 4H), 6.35 (d, J  = 3.2 Hz, 4H), 6.10 (dd, J  = 7.6, 1.3 Hz, 2H). 13 C NMR (101 MHz, DMSO- d 6 ) δ 165.15, 152.47, 151.79, 151.26, 150.14, 149.11, 145.56, 140.85, 139.70, 135.96, 131.52, 129.54, 129.48, 122.78, 121.95, 120.78, 119.02, 115.47, 103.35, 102.84, 102.05. ESI-MS: m / z 860.19 [M−PF 6 ] + . Data for Ir-2 : Yield 51%. 1 H NMR (400 MHz, DMSO- d 6 ) δ 8.89 (dd, J  = 7.8, 1.9 Hz, 2H), 8.40 (s, 2H), 8.28–8.22 (m, 4H), 8.06–7.99 (m, 4H), 7.42 (s, 2H), 7.24 (d, J  = 6.4 Hz, 2H), 7.14 (t, J  = 6.7 Hz, 4H), 6.94 (td, J  = 7.4, 1.2 Hz, 2H), 6.33 (d, J  = 3.0 Hz, 4H), 6.26 (dd, J  = 7.6, 1.3 Hz, 2H). 13 C NMR (101 MHz, DMSO- d 6 ) δ 165.21, 152.45, 151.81, 150.39, 150.27, 146.05, 145.74, 140.10, 138.78, 135.83, 131.68, 131.19, 129.88, 129.57, 128.39, 127.06, 122.72, 122.25, 121.32, 103.35, 102.89, 102.07. ESI-MS: m / z 869.17 [M−PF 6 ] + . Data for Ir-3 : Yield 56%. 1 H NMR (600 MHz, DMSO- d 6 ) δ 8.69 (d, J  = 8.0 Hz, 1H), 8.33 (t, J  = 7.8 Hz, 1H), 8.28–8.16 (m, 4H), 7.77 (d, J  = 8.3 Hz, 1H), 7.67 (dt, J  = 13.3, 5.6 Hz, 2H), 7.51 (d, J  = 6.3 Hz, 1H), 7.45 (d, J  = 7.1 Hz, 2H), 7.42–7.32 (m, 4H), 7.16 (t, J  = 7.7 Hz, 1H), 7.09 (t, J  = 7.7 Hz, 1H), 6.96 (t, J  = 7.9 Hz, 1H), 6.90 (q, J  = 7.2 Hz, 2H), 6.31 (dd, J  = 13.2, 5.1 Hz, 5H), 6.19 (d, J  = 7.6 Hz, 1H), 5.91 (d, J  = 8.4 Hz, 1H), 5.12 (s, 1H). 13 C NMR (151 MHz, DMSO- d 6 ) δ 165.73, 153.97, 153.68, 152.20, 152.13, 150.68, 150.64, 150.53, 150.28, 147.60, 146.44, 140.97, 140.75, 140.29, 140.22, 136.27, 136.09, 134.80, 132.82, 131.90, 130.29, 129.99, 129.58, 129.47, 129.00, 126.00, 124.78, 123.13, 122.96, 122.45, 122.30, 121.74, 121.67, 117.10, 114.22, 103.82, 103.79, 103.32, 103.27, 102.42, 102.37, 93.34. ESI-MS: m / z 884.19 [M−PF 6 ] + . Data for Ir-4 : Yield 48%. 1 H NMR (600 MHz, DMSO- d 6 ) δ 8.92 (d, J  = 8.2 Hz, 2H), 8.29–8.23 (m, 4H), 8.23–8.17 (m, 2H), 7.71 (t, J  = 5.0 Hz, 2H), 7.66 (ddd, J  = 6.9, 5.6, 1.2 Hz, 2H), 7.51–7.47 (m, 2H), 7.42 (d, J  = 6.4 Hz, 2H), 7.35 (dd, J  = 6.5, 3.4 Hz, 2H), 7.11–7.04 (m, 2H), 6.89 (t, J  = 7.3 Hz, 2H), 6.33 (q, J  = 4.4, 3.8 Hz, 4H), 6.17 (ddd, J  = 7.5, 5.0, 2.1 Hz, 2H). 13 C NMR (151 MHz, DMSO- d 6 ) δ 165.13, 155.33, 153.08, 151.85, 150.30, 149.62, 145.52, 139.92, 139.61, 135.87, 131.48, 129.92, 129.59, 128.59, 125.07, 122.77, 122.14, 121.40, 103.41, 102.92, 102.07. ESI-MS: m / z 845.18 [M−PF 6 ] + . Data for Ir-5 : Yield 53%. 1 H NMR (600 MHz, DMSO- d 6 ) δ 8.74 (d, J  = 6.4 Hz, 1H), 8.27 (s, 1H), 8.21 (s, 1H), 8.13 (ddd, J  = 8.0, 6.2, 1.3 Hz, 2H), 7.95 (td, J  = 7.8, 1.7 Hz, 1H), 7.77 (d, J  = 8.0 Hz, 1H), 7.64 (d, J  = 6.4 Hz, 1H), 7.60–7.54 (m, 3H), 7.47 (d, J  = 6.3 Hz, 1H), 7.42–7.39 (m, 1H), 7.28 (ddd, J  = 7.5, 5.6, 1.5 Hz, 1H), 7.00 (ddd, J  = 8.2, 7.1, 1.4 Hz, 1H), 6.94 (ddd, J  = 8.2, 7.2, 1.4 Hz, 1H), 6.77 (td, J  = 7.4, 1.2 Hz, 1H), 6.71 (td, J  = 7.4, 1.2 Hz, 1H), 6.40–6.33 (m, 4H), 6.20 (dd, J  = 7.7, 1.4 Hz, 1H), 6.13 (dd, J  = 7.7, 1.3 Hz, 1H), 5.57 (tt, J  = 8.0, 4.8 Hz, 1H), 4.98 (ddd, J  = 12.6, 8.1, 5.1 Hz, 1H), 4.61 (ddd, J  = 17.9, 7.8, 5.2 Hz, 1H), 4.34 (ddd, J  = 18.0, 8.1, 4.9 Hz, 1H). 13 C NMR (151 MHz, DMSO- d 6 ) δ 166.35, 164.23, 156.65, 152.04, 151.29, 150.40, 148.67, 146.45, 146.11, 142.22, 140.48, 138.61, 136.36, 136.19, 132.95, 131.87, 129.93, 129.89, 129.35, 129.14, 125.54, 123.43, 123.36, 122.98, 121.66, 121.42, 121.09, 103.84, 103.71, 103.24, 102.69, 102.59, 49.62. ESI-MS: m / z 797.18 [M−Cl] + . 4.2 Immunofluorescent detection of cell surface CRT MDA-MB-231 cells were incubated with Ir-1 at the specified concentration for 12 h in confocal dishes. After fixation and blocking, the cells were incubated with a primary anti-calreticulin antibody (Abcam, ab92516, Cambridge, UK) overnight at 4 °C. Then, the cells were washed and incubated with a secondary antibody (Abcam, ab150081, Cambridge, UK) for 2 h at room temperature. Upon removal of the antibody, the cells were stained with DAPI for 10 min. After washing again, the cells were stained with Wheat Germ Agglutinin (Invitrogen, W21404 , Burlington, Canada) for 30 min at 37 °C. The immunofluorescence of CRT was visualized using confocal fluorescence microscopy. 4.3 Extracellular HMGB1 detection assay HMGB1 release was assessed using an HMGB1 Detection Kit (Chondrex, 6010, Redmond, USA). MDA-MB-231 cells were seeded on a 10-cm dish at 1 × 10 6  cells/well overnight. The cells were then incubated with Ir-1 for the specified duration. Following this treatment, the HMGB1 assay kit was utilized to determine the HMGB1 levels following the manufacturer's instructions. Luminescence was measured using a BioTek plate reader (Agilent, Cytation5, Winooski, USA). 4.4 Extracellular ATP detection assay The secretion of ATP was quantified using an ATP Determination Kit (Invitrogen, A22066, Burlington, Canada). MDA-MB-231 cells were seeded on a 96-well plate at 1.5 × 10 4  cells/well overnight. Subsequently, cells were treated with Ir-1 at the indicated concentration for the specified duration. Following this treatment, the ATP assay kit was used to assess the ATP levels following the manufacturer's instructions. Luminescence was measured using a BioTek plate reader (Agilent, Cytation5, Winooski, USA). 4.5 Mouse vaccination assay Mice used in this study were purchased from Changsha's Animal Research Center (Hu Nan, China) and housed in the experimental animal facilities at the Guangxi Normal University under controlled temperature and humidity conditions with a 12-h light/12-h dark cycle. All mice had ad libitum access to commercial chow and water. All experimental procedures were executed according to the protocols approved by the Institutional Animal Care and Use Committee (IACUC) of Guangxi Normal University (Approval No: GXNU-202303-045). For the vaccination experiment, female BALB/c mice (5 weeks old, ∼20 g) were used. A total of 2 × 10 6 4T1 cells in the logarithmic growth phase underwent various treatments for 6 h: normal medium for the control group (subjected to repeated freeze-thaw) and 20 μmol/L Ir-1 or 150 μmol/L Oxa for each group ( n  = 10 per group). The treated 4T1 cells from each group were injected (s.c.) into the left flank of each mouse. After 7 days, the mice were rechallenged in the right flanks with untreated 4T1 cells (1 × 10 4  cells) and monitored daily for tumor formation on the right side. Mice without tumors in the right flanks were recorded as tumor-free. These observations continued until the end of the 36-day experiment, at which point the mice were euthanized. To uncover the underlying mechanism of the antitumor effect induced by Ir-1 -treated 4T1 cell vaccination, a comprehensive analysis of immune cells was performed using CyTOF following standard methods. On Day 36, the mice from the control, Oxa, and Ir-1 groups ( n  = 10) were euthanized. Tumors, spleens, and lymph nodes were dissected and grinded to obtain single-cell suspensions. Peripheral blood samples were collected via retro-orbital bleeding in heparin-coated tubes. Subsequently, cell suspensions were obtained via centrifugation at 1500 rpm for 5 min at 4 °C, followed by the removal of red blood cells at room temperature for 5 min and washing twice using 1 × PBS. ## Synthesis and characterization 4.1 Synthesis and characterization The ligand L a was firstly synthesized following a procedure described in our previous study 59 . Then, the general procedure for synthesizing compounds Ir-1 – Ir-5 is outlined as follows. The cyclometalated iridium(III) chloro-bridged dimer (c) was prepared according to a previously reported method 91 . IrCl 3 ·3H 2 O (5 mmol) and ligand L a (10 mmol) were added to a 6 mL mixture of 2-methoxylethanol (2-ME) and water (3:1, v / v ). The mixture underwent reflux at 120 °C under a nitrogen atmosphere for 1 h. After filtering the solution, the filtrate was dried at 60 °C to yield compound (c). Then, as previously described 92 , 0.14 mol of compound (c) and 0.3 mmol of NˆN ligand (L 1 , L 2 , L 3 or L 4 ) were added to 5 mL of ethylene glycol, and the mixture was refluxed for 2 h at 120 °C under a nitrogen atmosphere. Once the solution cooled to room temperature, an excess of KPF 6 solution was added, resulting in the formation of a precipitate. The solid precipitate was then filtered and left to dry at 60 °C overnight. The purification of the crude product was performed on a silica gel column (EA/MeOH, 9:1 v / v ), yielding a bright-orange product ( Ir-1 , Ir-2 , Ir-3 , or Ir-4 ). In a separate reaction, 0.14 mol of compound (c) and 0.3 mmol of NˆN ligand (L 5 ) were added to 10 mL of methanol. Subsequently, 0.6 mmol Na 2 CO 3 was added, and the mixture was refluxed at 65 °C under a nitrogen atmosphere for 30 min. The as-obtained solution was concentrated, and the crude product was purified on a silica gel column (EA/MeOH, 9:1 v / v ) to obtain the bright-orange product ( Ir-5 ). Data for Ir-1 : Yield 60%. 1 H NMR (400 MHz, DMSO- d 6 ) δ 10.64 (s, 1H), 8.21 (s, 2H), 8.14 (d, J  = 8.0 Hz, 2H), 8.04 (d, J  = 6.4 Hz, 2H), 7.82 (ddd, J  = 8.7, 7.2, 1.8 Hz, 2H), 7.64–7.53 (m, 4H), 7.41–7.30 (m, 4H), 7.09–6.94 (m, 2H), 6.78 (td, J  = 7.7, 2.1 Hz, 4H), 6.35 (d, J  = 3.2 Hz, 4H), 6.10 (dd, J  = 7.6, 1.3 Hz, 2H). 13 C NMR (101 MHz, DMSO- d 6 ) δ 165.15, 152.47, 151.79, 151.26, 150.14, 149.11, 145.56, 140.85, 139.70, 135.96, 131.52, 129.54, 129.48, 122.78, 121.95, 120.78, 119.02, 115.47, 103.35, 102.84, 102.05. ESI-MS: m / z 860.19 [M−PF 6 ] + . Data for Ir-2 : Yield 51%. 1 H NMR (400 MHz, DMSO- d 6 ) δ 8.89 (dd, J  = 7.8, 1.9 Hz, 2H), 8.40 (s, 2H), 8.28–8.22 (m, 4H), 8.06–7.99 (m, 4H), 7.42 (s, 2H), 7.24 (d, J  = 6.4 Hz, 2H), 7.14 (t, J  = 6.7 Hz, 4H), 6.94 (td, J  = 7.4, 1.2 Hz, 2H), 6.33 (d, J  = 3.0 Hz, 4H), 6.26 (dd, J  = 7.6, 1.3 Hz, 2H). 13 C NMR (101 MHz, DMSO- d 6 ) δ 165.21, 152.45, 151.81, 150.39, 150.27, 146.05, 145.74, 140.10, 138.78, 135.83, 131.68, 131.19, 129.88, 129.57, 128.39, 127.06, 122.72, 122.25, 121.32, 103.35, 102.89, 102.07. ESI-MS: m / z 869.17 [M−PF 6 ] + . Data for Ir-3 : Yield 56%. 1 H NMR (600 MHz, DMSO- d 6 ) δ 8.69 (d, J  = 8.0 Hz, 1H), 8.33 (t, J  = 7.8 Hz, 1H), 8.28–8.16 (m, 4H), 7.77 (d, J  = 8.3 Hz, 1H), 7.67 (dt, J  = 13.3, 5.6 Hz, 2H), 7.51 (d, J  = 6.3 Hz, 1H), 7.45 (d, J  = 7.1 Hz, 2H), 7.42–7.32 (m, 4H), 7.16 (t, J  = 7.7 Hz, 1H), 7.09 (t, J  = 7.7 Hz, 1H), 6.96 (t, J  = 7.9 Hz, 1H), 6.90 (q, J  = 7.2 Hz, 2H), 6.31 (dd, J  = 13.2, 5.1 Hz, 5H), 6.19 (d, J  = 7.6 Hz, 1H), 5.91 (d, J  = 8.4 Hz, 1H), 5.12 (s, 1H). 13 C NMR (151 MHz, DMSO- d 6 ) δ 165.73, 153.97, 153.68, 152.20, 152.13, 150.68, 150.64, 150.53, 150.28, 147.60, 146.44, 140.97, 140.75, 140.29, 140.22, 136.27, 136.09, 134.80, 132.82, 131.90, 130.29, 129.99, 129.58, 129.47, 129.00, 126.00, 124.78, 123.13, 122.96, 122.45, 122.30, 121.74, 121.67, 117.10, 114.22, 103.82, 103.79, 103.32, 103.27, 102.42, 102.37, 93.34. ESI-MS: m / z 884.19 [M−PF 6 ] + . Data for Ir-4 : Yield 48%. 1 H NMR (600 MHz, DMSO- d 6 ) δ 8.92 (d, J  = 8.2 Hz, 2H), 8.29–8.23 (m, 4H), 8.23–8.17 (m, 2H), 7.71 (t, J  = 5.0 Hz, 2H), 7.66 (ddd, J  = 6.9, 5.6, 1.2 Hz, 2H), 7.51–7.47 (m, 2H), 7.42 (d, J  = 6.4 Hz, 2H), 7.35 (dd, J  = 6.5, 3.4 Hz, 2H), 7.11–7.04 (m, 2H), 6.89 (t, J  = 7.3 Hz, 2H), 6.33 (q, J  = 4.4, 3.8 Hz, 4H), 6.17 (ddd, J  = 7.5, 5.0, 2.1 Hz, 2H). 13 C NMR (151 MHz, DMSO- d 6 ) δ 165.13, 155.33, 153.08, 151.85, 150.30, 149.62, 145.52, 139.92, 139.61, 135.87, 131.48, 129.92, 129.59, 128.59, 125.07, 122.77, 122.14, 121.40, 103.41, 102.92, 102.07. ESI-MS: m / z 845.18 [M−PF 6 ] + . Data for Ir-5 : Yield 53%. 1 H NMR (600 MHz, DMSO- d 6 ) δ 8.74 (d, J  = 6.4 Hz, 1H), 8.27 (s, 1H), 8.21 (s, 1H), 8.13 (ddd, J  = 8.0, 6.2, 1.3 Hz, 2H), 7.95 (td, J  = 7.8, 1.7 Hz, 1H), 7.77 (d, J  = 8.0 Hz, 1H), 7.64 (d, J  = 6.4 Hz, 1H), 7.60–7.54 (m, 3H), 7.47 (d, J  = 6.3 Hz, 1H), 7.42–7.39 (m, 1H), 7.28 (ddd, J  = 7.5, 5.6, 1.5 Hz, 1H), 7.00 (ddd, J  = 8.2, 7.1, 1.4 Hz, 1H), 6.94 (ddd, J  = 8.2, 7.2, 1.4 Hz, 1H), 6.77 (td, J  = 7.4, 1.2 Hz, 1H), 6.71 (td, J  = 7.4, 1.2 Hz, 1H), 6.40–6.33 (m, 4H), 6.20 (dd, J  = 7.7, 1.4 Hz, 1H), 6.13 (dd, J  = 7.7, 1.3 Hz, 1H), 5.57 (tt, J  = 8.0, 4.8 Hz, 1H), 4.98 (ddd, J  = 12.6, 8.1, 5.1 Hz, 1H), 4.61 (ddd, J  = 17.9, 7.8, 5.2 Hz, 1H), 4.34 (ddd, J  = 18.0, 8.1, 4.9 Hz, 1H). 13 C NMR (151 MHz, DMSO- d 6 ) δ 166.35, 164.23, 156.65, 152.04, 151.29, 150.40, 148.67, 146.45, 146.11, 142.22, 140.48, 138.61, 136.36, 136.19, 132.95, 131.87, 129.93, 129.89, 129.35, 129.14, 125.54, 123.43, 123.36, 122.98, 121.66, 121.42, 121.09, 103.84, 103.71, 103.24, 102.69, 102.59, 49.62. ESI-MS: m / z 797.18 [M−Cl] + . ## Immunofluorescent detection of cell surface CRT 4.2 Immunofluorescent detection of cell surface CRT MDA-MB-231 cells were incubated with Ir-1 at the specified concentration for 12 h in confocal dishes. After fixation and blocking, the cells were incubated with a primary anti-calreticulin antibody (Abcam, ab92516, Cambridge, UK) overnight at 4 °C. Then, the cells were washed and incubated with a secondary antibody (Abcam, ab150081, Cambridge, UK) for 2 h at room temperature. Upon removal of the antibody, the cells were stained with DAPI for 10 min. After washing again, the cells were stained with Wheat Germ Agglutinin (Invitrogen, W21404 , Burlington, Canada) for 30 min at 37 °C. The immunofluorescence of CRT was visualized using confocal fluorescence microscopy. ## Extracellular HMGB1 detection assay 4.3 Extracellular HMGB1 detection assay HMGB1 release was assessed using an HMGB1 Detection Kit (Chondrex, 6010, Redmond, USA). MDA-MB-231 cells were seeded on a 10-cm dish at 1 × 10 6  cells/well overnight. The cells were then incubated with Ir-1 for the specified duration. Following this treatment, the HMGB1 assay kit was utilized to determine the HMGB1 levels following the manufacturer's instructions. Luminescence was measured using a BioTek plate reader (Agilent, Cytation5, Winooski, USA). ## Extracellular ATP detection assay 4.4 Extracellular ATP detection assay The secretion of ATP was quantified using an ATP Determination Kit (Invitrogen, A22066, Burlington, Canada). MDA-MB-231 cells were seeded on a 96-well plate at 1.5 × 10 4  cells/well overnight. Subsequently, cells were treated with Ir-1 at the indicated concentration for the specified duration. Following this treatment, the ATP assay kit was used to assess the ATP levels following the manufacturer's instructions. Luminescence was measured using a BioTek plate reader (Agilent, Cytation5, Winooski, USA). ## Mouse vaccination assay 4.5 Mouse vaccination assay Mice used in this study were purchased from Changsha's Animal Research Center (Hu Nan, China) and housed in the experimental animal facilities at the Guangxi Normal University under controlled temperature and humidity conditions with a 12-h light/12-h dark cycle. All mice had ad libitum access to commercial chow and water. All experimental procedures were executed according to the protocols approved by the Institutional Animal Care and Use Committee (IACUC) of Guangxi Normal University (Approval No: GXNU-202303-045). For the vaccination experiment, female BALB/c mice (5 weeks old, ∼20 g) were used. A total of 2 × 10 6 4T1 cells in the logarithmic growth phase underwent various treatments for 6 h: normal medium for the control group (subjected to repeated freeze-thaw) and 20 μmol/L Ir-1 or 150 μmol/L Oxa for each group ( n  = 10 per group). The treated 4T1 cells from each group were injected (s.c.) into the left flank of each mouse. After 7 days, the mice were rechallenged in the right flanks with untreated 4T1 cells (1 × 10 4  cells) and monitored daily for tumor formation on the right side. Mice without tumors in the right flanks were recorded as tumor-free. These observations continued until the end of the 36-day experiment, at which point the mice were euthanized. To uncover the underlying mechanism of the antitumor effect induced by Ir-1 -treated 4T1 cell vaccination, a comprehensive analysis of immune cells was performed using CyTOF following standard methods. On Day 36, the mice from the control, Oxa, and Ir-1 groups ( n  = 10) were euthanized. Tumors, spleens, and lymph nodes were dissected and grinded to obtain single-cell suspensions. Peripheral blood samples were collected via retro-orbital bleeding in heparin-coated tubes. Subsequently, cell suspensions were obtained via centrifugation at 1500 rpm for 5 min at 4 °C, followed by the removal of red blood cells at room temperature for 5 min and washing twice using 1 × PBS. ## Acknowledgments Acknowledgments This study was supported by the 10.13039/501100001809 National Natural Science Foundation of China ( 22177022 , 22267005 ) and the 10.13039/501100004607 Natural Science Foundation of Guangxi Province of China ( AD22035193 ). ## Author contributions Author contributions Yuan Lu: Investigation, Visualization, Writing – original draft, Formal analysis. Shan-Shan Wang: Investigation, Methodology, Visualization. Meng-Ya Li: Investigation, Methodology, Visualization. Rong Liu: Methodology, Visualization. Meng-Fan Zhu: Methodology, Visualization. Liang-Mei Yang: Methodology, Visualization. Feng-Yang Wang: Funding acquisition, Methodology. Ke-Bin Huang: Conceptualization, Project administration, Writing – review & editing. Hong Liang: Funding acquisition, Project administration, Supervision. ## Conflicts of interest Conflicts of interest The authors declare no conflicts of interest.