STING synergizes with TOX suppressing HO-1 expression to trigger ferroptosis in tumor-infiltrating CD8+ T cell and immunotherapy resistance

Zhu, Qian; Zhang, Jun-bao; Nie, Cai-ping; Liu, Xiu-feng; Zhan, Liang-ping; Li, Ming; Zeng, Xi-liang; He, Jia; Huang, He; Xia, Xiao-jun; Gao, Song; Zhang, Xiao-shi; Li, Jiang

doi:10.1038/s41467-026-69350-y

Download PDF

Article
Open access
Published: 09 February 2026

STING synergizes with TOX suppressing HO-1 expression to trigger ferroptosis in tumor-infiltrating CD8+ T cell and immunotherapy resistance

Nature Communications volume 17, Article number: 2543 (2026) Cite this article

5696 Accesses
1 Citations
1 Altmetric
Metrics details

Subjects

Abstract

CD8⁺ T cell abundance within the tumor microenvironment is a critical determinant of immunotherapy efficacy. Here we show that CD8⁺ T cells lacking STING or TOX display markedly improved antitumor activity, with enhanced tumor infiltration and elevated IFN-γ and granzyme B production. These STING or TOX deficient cells exhibit a stem-like transcriptional state and resist ferroptosis by suppressing lipid peroxidation pathways while promoting mitochondrial biogenesis. Mechanistically, STING and TOX form a positive regulatory loop that represses HO-1 expression, leading to iron accumulation, mitochondrial oxidative stress, and ferroptosis in tumor-infiltrating CD8⁺ T cells. We further identify lactate as a microenvironmental trigger of STING–TOX–HO-1-mediated CD8⁺ T-cell ferroptosis. In mouse tumor models, engineered STING/TOX-deficient CD8⁺ T cells synergize with immune checkpoint blockade, chemotherapy, or STING agonist to enhance tumor control. These findings reveal a central pathway governing CD8⁺ T-cell ferroptosis in tumors and suggest therapeutic strategies to overcome immunotherapy resistance.

Synergistic reprogramming of the tumor immune microenvironment by Senecavirus A and STING agonist

Article 11 February 2026

Universal STING mimic boosts antitumour immunity via preferential activation of tumour control signalling pathways

Article 13 March 2024

AND logic nanoparticle for precision immunotherapy of metastatic cancers

Article Open access 09 March 2026

Introduction

Dysfunction of CD8⁺ cytotoxic T cells within the tumor microenvironment (TME) is a key driver for tumor immune escape and a significant barrier to the clinical efficacy of immunotherapies, such as immune checkpoint blockade (ICB) and adoptive cell therapy (ACT). Anti–PD-1/PD-L1 antibodies have shown substantial objective response rates (ORR) in several malignancies, including non–small cell lung cancer (NSCLC)^1,2,3. However, due to the complexity and heterogeneity of the TME, blockade of alternative inhibitory receptors such as TIM-3 has yielded limited clinical benefit, and more than 60% of initial responders eventually develop acquired resistance^4,5,6. Consequently, enhancing the efficacy and durability of ICB while mitigating resistance has become a major focus in cancer immunotherapy. Current efforts often involve combining immune checkpoint inhibitors (ICIs) with chemoradiotherapy, targeted therapies, or cancer vaccines^7,8. Despite these advances, there remains an urgent need to design combination strategies that are both broadly effective and minimally toxic.

STING (stimulator of interferon genes), a sensor of cytosolic DNA activated via cyclic dinucleotides, regulates innate immune responses by inducing production of interferons and cytokines⁹. Beyond its role in innate immunity, STING also influences CD8⁺ T cell-mediated responses to DNA damage¹⁰. Although STING is broadly expressed in a wide variety of cell types throughout normal tissues, its activation produces tissue-specific effects that can either support protective immunity or contribute to pathological outcomes^11,12. Current evidence indicates that STING activation in tumor cells enhances type I interferon signaling, thereby stimulating the activation of antigen-presenting cells and anti-tumor T cell responses¹³. However, clinical trials evaluating the STING agonist MIW815 (ADU-S100) in advanced cancers have demonstrated limited clinical efficiency^14,15. Notably, STING expression is critical for immune cell development and functionality, human peripheral T cells exhibit low basal levels of STING under physiological conditions in contrast to their murine counterparts; However, STING expression is significantly upregulated upon T cell activation in contexts such as infection and cancer^16,17. Recently, emerging research shows that STING activation within human and murine T cells suppresses proliferation and triggers their death by apoptosis¹⁸. Moreover, STING signaling has been implicated in mitochondrial-dependent ferroptosis in tumor cells¹⁹.

Tumor-infiltrating CD8⁺ T cells undergo functional remodeling within the TME in response to both intercellular and intracellular signals²⁰. The nuclear factor thymocyte selection-associated high mobility group box (TOX) has emerged as a master regulator of tumor-specific T cell differentiation²¹. Under conditions of chronic antigen stimulation, such as in cancer or persistent viral infections, TOX expression is upregulated, promoting T cell exhaustion and dysfunction. TIM-3 and PD-1 double-positive exhausted T cells usually are co-expressed with TOX within TME²². Murine studies demonstrate that TOX or TOX2 knockout enhances tumor control by augmenting interferon-gamma (IFN-γ) and tumor necrosis factor-alpha (TNF-α) secretion in tumor-specific T cells²³. Despite these advances, the regulatory mechanisms linking TOX to CD8⁺ T cell dysfunction within the TME remain poorly understood.

Here, we show that the mice harboring CD8⁺ T cells deficient in either STING or TOX exhibit markedly enhanced antitumor activity. STING signaling synergizes with TOX to suppress heme oxygenase-1 (HO-1) expression, thereby promoting ferrous-ion-accumulation-induced mitochondrial ferroptosis in CD8⁺ T cells. Moreover, the lactate accumulation in the TME serves as a critical upstream driver of STING-TOX-HO-1-mediated ferroptosis. These findings provide mechanistic insight into the regulation of CD8⁺ T-cell ferroptosis in the TME and highlight potential therapeutic strategies to overcome immunotherapy resistance.

Results

STING- or TOX-deficient CD8⁺ T cells display enhanced tumor clearance

To investigate the regulatory role of intracellular STING and TOX signaling in CD8⁺ T cells during antitumor immunity, the Cd8^creSting1^flox/wt (Sting1-cKD), Cd8^creSting1^flox/flox (Sting1-cKO), Cd8^creTox^flox/flox (Tox-cKO) and double-knockdown (Sting1/Tox-D-cKD) mice were generated by crossing Sting1^flox/flox and/or Tox^flox/flox CD45.2⁺ C57BL/6 J mice with Cd8^cre CD45.2 C57BL/6 J mice, and Sting1^flox/flox, Tox^flox/flox and wild-type (WT, CD45.1⁺ and CD45.2⁺) C57BL/6 J mice were included as controls. The Cd8^creSting1^flox/wt (Sting1-cKD), Cd8^creSting1^flox/flox (Sting1-cKO), Cd8^creTox^flox/flox (Tox-cKO) and double-knockdown (Sting1/Tox-D-cKD) mice exhibited phenotypes comparable to Sting1^flox/flox, Tox^flox/flox, Cd8^cre and WT mice (Supplementary Fig. S1a–c). Notably, Sting1-cKD, Tox-cKO, and Sting1/Tox-D-cKD mice displayed enhanced tumor clearness across multiple models (TC-1, MC38, and B16F10) compared with WT mice (Fig. 1a and Supplementary Fig. S1d–e). These cKO mice displayed increased frequencies of CD3⁺CD8⁺ tumor-infiltrating lymphocytes (TILs), CD8⁺PD1⁺TIM3^-CXCR5⁺ stem-like and CD8⁺TCF-1⁺Ly108⁺CD69⁺ precursor of exhausted TILs, as well as CD8⁺ TILs producing interferon-gamma (IFN-γ) and Granzyme B (GrB); additionally, an increased CD8⁺Ki67⁺ proliferating TIL subset was observed in tumor-bearing Tox-cKO and Sting1/Tox-D-cKD mice but not in Sting1-cKD mice related to WT mice. In contrast, the proportion of CD8⁺PD-1⁺TIM-3⁺CXCR5^- exhausted TILs was reduced in these cKO mice; and the number of CD3⁺CD4⁺ TILs remained equal in these three cKO and WT mice (Supplementary Fig. S2a and Fig. 1b–d). Accordingly, peripheral T cell homeostasis remained unaffected in tumor-bearing Sting1-cKD and Tox-cKO mice compared with WT controls (Supplementary Fig. S2b). In ACT models, TC-1-bearing NSG mice receiving preactivated CD45.1⁺ WT-CD3⁺ T cells combined with preactivated CD45.2⁺CD3⁺ T cells from WT, Sting1-cKD, Tox-cKO, or Sting1/Tox-D-cKD donors exhibited greater tumor regression than PBS-treated controls (Fig. 1e, f). Donor-derived CD45.2⁺CD8⁺ TILs from Sting1-cKD, Tox-cKO, and Sting1/Tox-D-cKD groups showed higher infiltration and greater production of IFN-γ and GrB compared with CD45.1⁺ WT counterparts. In contrast, CD45.2⁺ WT-derived CD8⁺ TILs did not significantly differ from CD45.1⁺ WT TILs in terms of infiltration or cytokine production (Supplementary Fig. S2c and Fig. 1g, h). Furthermore, in TC-1-bearing CD45.1⁺ WT mice, adoptive transfer of pre-activated CD45.2⁺ WT CD8⁺ T cells treated with the STING inhibitor H151 led to enhanced tumor clearance compared with untreated CD45.2⁺ WT CD8⁺ T cell and PBS control groups, accompanied by increased lymphocyte infiltration and elevated IFN-γ production (Supplementary Fig. S3).

Fig. 1: STING- or TOX-deficiency in CD8+ T cells augments anti-tumor immunity. — **Fig. 1: STING- or TOX-deficiency in CD8⁺ T cells augments anti-tumor immunity.**

Transcriptomic features of tumor-infiltrating CD8⁺ T cells deficient in STING or TOX show resistance to ferroptosis

We performed bulk RNA sequencing of tumor-infiltrating CD8⁺ T cells from WT, Sting1-cKO, and Tox-cKO tumor-bearing mice. Differential gene expression analysis revealed 1,860 downregulated and 723 upregulated genes in Sting1-cKO CD8⁺ TILs, and 362 downregulated and 286 upregulated genes in Tox-cKO CD8⁺ TILs, compared to WT CD8⁺ TILs ( | log2FC | > 1, FDR < 0.05; Supplementary Fig. S4). Notably, transcripts encoding Sting1, the immune inhibitory receptors (Tox, Havcr2, Pdcd1, Lag3, Ctla4, and Entpd1), chemokine receptors (Cx3cr1, Ccr5 and Cxcr6) related to CD8⁺ T cell homing and ferroptosis drivers (Trim21, Pparg, Taz, Ptpn6, Flt3, and Ncoa4) were significantly downregulated. In contrast, genes associated with proliferative and stem-like properties (Mki67, Pcna, Top2a, Sell, Tcf7, and Cxcr5), along with the ferroptosis suppressor genes (Hmox1, Slc7a11, Acsl3, Trib2, Slc40a1, and Lcn2), were upregulated in both Sting1-cKO- and Tox-cKO-CD8⁺ TILs relative to WT CD8⁺ TILs (Supplementary Fig. S4 and Fig. 2a, b). We further demonstrated reductions in STING, phosphorylated TBK1 (p-TBK1) and IRF3 (p-IRF3), TOX, TIM-3 and PD-1, and ACSL4 and cathepsin B (CTSB). Conversely, expression of ferroptosis suppressors HO-1 and GPX4 was elevated in Sting1-cKO and Tox-cKO CD8⁺ TILs compared to WT CD8⁺ TILs (Fig. 2c). Multicolor immunofluorescence (IF) staining of tumor tissues revealed reduced frequencies of CD8⁺TOX⁺, CD8⁺PD-1⁺, CD8⁺TIM-3⁺, and CD8⁺CTSB⁺ cells, alongside increased CD8⁺HO-1⁺ and CD8⁺GPX4⁺ populations in Sting1-cKO and Tox-cKO mice (Fig. 2d, e). Pathway enrichment analysis identified shared alterations in Sting1-cKO- and Tox-cKO-CD8⁺ TILs, including pathways involved in iron ion transport regulation, lipid metabolic process and granzyme-mediated cell death (Fig. 2f). Specifically, intracellular iron homeostasis pathway was upregulated in both Sting1-cKO- and Tox-cKO-CD8⁺ TILs, while fatty acid metabolism pathway was downregulated in Sting1-cKO-CD8⁺ TILs; and ferroptosis-associated gene pathway was downregulated in both Sting1-cKO and Tox-cKO CD8⁺ TILs compared to WT CD8⁺ TILs (Fig. 2g). Consistently, Sting1-cKO- and Tox-cKO- CD8⁺ TILs exhibited decreased total cellular iron levels and reduced lipid peroxidation relative to WT-CD8⁺ TILs (Figs. 2h, i).

Fig. 2: Transcriptomic profiling reveals ferroptosis resistance of tumor-infiltrating CD8+ T cells lacking STING or TOX. — **Fig. 2: Transcriptomic profiling reveals ferroptosis resistance of tumor-infiltrating CD8⁺ T cells lacking STING or TOX.**

STING- and TOX-deficient tumor-infiltrating CD8⁺ T cells exhibit reduced susceptibility to mitochondria-dependent ferroptosis

To investigate enhanced tumor infiltration observed in CD8⁺ T cell in Sting1-cKO and Tox-cKO mice, we evaluated the proliferation and cell death rates of STING- and TOX-deficient CD8⁺ T cells compared to WT CD8⁺ T cells in vitro and found out that the proliferation remained comparable across genotypes (Supplementary Fig. S5a), but the STING- and TOX-deficient CD8⁺ T cells exhibited significantly reduced mortality following co-culture with TC-1 tumor cells and the STING agonist DMXAA (Supplementary Fig. S5b). Notably, the treatment of ferroptosis inhibitor Ferrostatin-1 (Fer-1), but not necrosis inhibitor Necrostatin-1 (Nec-1) or apoptosis inhibitor Z-VAD-FMK (Z-VAD), extremely significant rescued cell death of WT CD8⁺ T cells upon TC-1-stimulation (Supplementary Fig. S5c). Furthermore, the STING- and TOX-deficient CD8⁺ T cells displayed dose-dependent resistance to ferroptosis induced by Erastin, an SLC7A11 inhibitor (Supplementary Fig. S5d). Consistent with these results, WT CD8⁺ T cells showed increased nuclear CTSB expression and lipid peroxidation following TC-1/DMXAA stimulation, relative to their STING- and TOX-deficient counterparts (Supplementary Figs. S5e, f). Single-cell transcriptomic analysis revealed that ferroptosis is the predominant mode of cell death in tumor-infiltrating CD8⁺ and CD4⁺ T cells in contrast to their peripheral counterparts from patients with cervical cancer (CC, n = 2) or nasopharyngeal carcinoma (NPC, n = 10), accompanied by upregulation of STING and lipid metabolic pathways (Supplementary Fig. S6a, b). Consistently, pharmacological inhibition with Fer-1, but not with Nec-1 or Z-VAD, prevented tumor-induced human CD8⁺ T cell death in vitro (Supplementary Fig. S6c).

Subcellular localization analysis revealed mitochondrial enrichment of STING and ferroptosis-associated proteins, including STING, IRF3, HO-1, and ACSL4 in all CD8⁺ T cells upon TC-1 stimulation; whereas the CD8⁺ T cells deficient in STING or TOX exhibited reduced STING and ACSL4 expression but elevated HO-1 expression following TC-1 stimulation (Fig. 3a). Transcriptomic clustering using MSigDB (https://www.gsea-msigdb.org/gsea/msigdb) identified upregulated survival pathways (Cluster1) and downregulated death/lipid metabolism genes (Cluster3) in both STING- and TOX-deficient CD8⁺ TILs. Notably, TOX-deficient CD8⁺ TILs uniquely showed enrichment in iron ion homeostasis and cytokine production pathways (Cluster 2) (Fig. 3b). Functional assessments demonstrated preserved mitochondrial integrity in STING- and TOX-deficient CD8⁺ T cells, evidenced by reduced mitochondrial ROS (Mito-ROS) and mitochondrial lipid peroxidation (Mito-lipid peroxidation) along with increased mitochondrial mass (Mito-mass), and sustained ATP production and mitochondrial (mt)DNA copy number following TC-1 stimulation (Fig. 3c–g). Using a transwell co-culture system with murine tumor cell lines (CT26, B16F10 and TC-1), human tumor cells (SiHa and C666), and viral antigens peptide (HPV-E6/E7 and EBV-LMP1), we observed activation of STING signaling in murine and human CD8⁺ T cells. This was characterized by elevated levels of Ifit1, Isg15, and Isg56, as well as increased p-TBK1, p-IRF3, alongside TOX upregulation and HO-1 downregulation (Supplementary Fig. S7a, b), as well as the metabolic collapse in WT murine and human CD8⁺ T cells following tumor cell or viral antigen stimulation in vitro (Supplementary Fig. S7c–g). Collectively, these data suggest that STING/TOX signaling promotes ferroptosis of tumor-derived CD8⁺ T cells through mitochondrial metabolic reprogramming within the TME.

Fig. 3: STING- and TOX-deficient tumor-infiltrating CD8+ T cells exhibit resistant to mitochondrial metabolism-dependent ferroptosis. — **Fig. 3: STING- and TOX-deficient tumor-infiltrating CD8⁺ T cells exhibit resistant to mitochondrial metabolism-dependent ferroptosis.**

STING-TOX-HO-1 axis drives tumor-primed ferroptosis in CD8⁺ T cells through iron-overloaded mitochondrial metabolic reprogramming

Transcriptional analysis revealed reciprocal regulation between STING and TOX in CD8⁺ T cells, with HO-1 upregulated in both STING- and TOX-deficient CD8⁺ T cells (Fig. 2a, b). Knockout of TOX or STING reduced the sensitivity of SiHa-stimulated human CD8⁺ T cells to ferroptosis, as indicated by decreased cell death, diminished ferrous ion accumulation, and reduced lipid peroxidation (Supplementary Fig. S8a–c). These deficiencies also restored mitochondrial metabolism, reflected by increased Mito-mass and reduced Mito-ROS (Supplementary Fig. S8d, e), and reversed the induction of TOX, STING, p-TBK1, p-IRF3, and HO-1 (Supplementary Fig. S8f). Knockdown of IRF3, a downstream mediator of the STING signaling pathway, in WT CD8⁺ T cells led to decreased TOX expression and increased HO-1 expression, accompanied by decreased total iron accumulation and increased Mito-mass (Supplementary Fig. S9a–c). Ectopic expression of TOX or STING in TC-1-stimulated STING- and TOX-deficient CD8⁺ T cells, respectively, restored sensitivity to ferroptosis, as evidenced by elevated cell death, increased ferrous ion accumulation, and enhanced lipid peroxidation (Fig. 4a–c). Simultaneously, ectopic expression of TOX or STING induced mitochondrial depolarization, evidenced by reduced Mito-mass and elevated Mito-ROS (Fig. 4d, e), while reactivating the STING-TOX-HO-1 signaling axis in the corresponding deficient cells (Fig. 4f).

Fig. 4: STING-TOX-HO-1 axis promotes tumor-induced CD8+ T cell ferroptosis through iron-overloaded mitochondrial reprogramming. — **Fig. 4: STING-TOX-HO-1 axis promotes tumor-induced CD8⁺ T cell ferroptosis through iron-overloaded mitochondrial reprogramming.**

Previous studies have shown that HO-1 suppression can induce ferrous ion overload^24,25. However, HO-1 knockdown paradoxically increased ferroptosis susceptibility in TC-1-stimulated STING- and TOX-deficient CD8⁺ T cells, as evidenced by significantly elevated cell death and ferrous ion levels without affecting lipid peroxidation. However, this intervention also exacerbated mitochondrial depolarization, as indicated by reduced Mito-mass and increased Mito-ROS (Fig. 4g–k). Iron chelation experiments using Deferoxamine (DFO) persevered mitochondrial integrity, reflected by increased Mito-mass and modulated mitochondrial dynamics by reducing DRP1 expression while increasing levels of MFN1 and HO-1 (Supplementary Fig. S9d–f). HO-1 knockdown in TOX-deficient CD8⁺ T cells impaired tumor clearance in TC-1-bearing CD45.1 WT mice, which was associated with reduced lymphocyte infiltration and diminished IFN-γ production (Supplementary Fig. S10). Collectively, these data identify the STING-TOX-HO-1 axis as a key ferroptosis checkpoint that regulates mitochondrial iron homeostasis in CD8⁺ T cells (Fig. 4l).

Tumor microenvironment induces STING-TOX-HO-1-mediated ferroptosis of tumor-infiltrating CD8⁺ T cells through lactate accumulation

In vitro and in vivo data demonstrated that ferroptosis was the predominant form of cell death of in both murine and human tumor-infiltrating CD8⁺ T cells (Supplementary Figs. S5, S6). Given prior evidence linking lactate to mitochondrial metabolic dysfunction and ferroptosis²⁶, we observed that the lactate release in human (HNE-1, EBV⁺-TW03, SiHa) and murine (TC-1, CT26, B16F10) tumor cells were largely exceeded those in human and murine peripheral blood mononuclear cells (PBMCs, Supplementary Fig. S11a). Notably, WT CD8⁺ T cells exhibited greater sensitivity to lactate-induced cell death in a dose-dependent manner compared to WT CD4⁺ T cells and TC-1 cells (Supplementary Fig. S11b). Lactate treatment also induced iron accumulation, reduced Mito-mass, and decreased mtDNA copy number in WT CD8⁺ T cells (Supplementary Fig. S11c–1e). Further analysis revealed dose-dependent increases in the mitochondrial fission regulator DRP1, reductions in the mitochondrial fusion protein MFN1 and ferroptosis suppressor HO-1, along with elevated STING signaling (p-TBK1 and p-IRF3) and TOX expression in lactate-stimulated WT CD8⁺ T cells (Fig. S11f, g). Notably, STING- or TOX-deficient CD8⁺ T cells were resistant to lactate-induced cell death compared to their WT counterparts (Fig. 5a). Lactate specifically induced nuclear CTSB expression, lipid peroxidation, and mtDNA loss in WT-CD8⁺ T cells but not in STING- or TOX-deficient CD8⁺ T cells (Figs. 5b–f). Strikingly, treatment with lactate, TC-1 cells, or STING agonist DMXAA increased p-TBK1, p-IRF3 and TOX levels while suppressing HO-1 expression in WT CD8⁺ T cells while these effects were absent in STING- or TOX-deficient CD8⁺ T cells (Fig. 5g). However, these lactate-induced alterations in WT CD8⁺ T cells were reversed by the MCT1 inhibitor AZD3965, and this effect was markedly attenuated in Sting1-cKO or Tox-cKO CD8⁺ T cells, consistent with impaired activation of STING signaling and ferroptosis pathways (Fig. 5h). The Transcriptional profiling further revealed a pronounced reduction in several solute carrier family 16 (SLC16) members which was involved in lactate transport²⁷, including Slc16a2, Slc16a3, Slc16a4, Slc16a6, Slc16a7, Slc16a11, and Slc16a13, in CD8⁺ TILs from Sting1-cKO or Tox-cKO mice compared to WT controls in vivo. In vitro, co-culture with TC-1 tumor cells resulted in reduced MCT11 protein abundance in Sting1-cKO CD8⁺ T cells, but not changed in Tox-cKO CD8⁺ T cells, whereas MCT1 levels remained unchanged (Fig. S11h, i). Collectively, these findings identify that the SLC16 transporters may be involved in lactate-driven ferroptosis in T cells mediated by STING-TOX-HO-1 signals. Consistent with this, AZD3965 inhibition of MCT1 markedly impaired B16F10 tumor growth in WT mice compared with PBS-treated controls, yielding a phenotype comparable to that observed in Sting1-cKO mice (Fig. S12a b). AZD3965 treatment increased the frequencies of CD8⁺ TILs, CD4⁺ TILs, TCF-1⁺Ly108⁺CD69⁺ precursor of exhausted CD8⁺ T cells, and IFN-γ–producing CD8⁺ TILs, but not Ki67⁺CD8⁺ proliferating TILs in tumor-bearing WT mice (Fig. S12c–f). Furthermore, AZD3965 reduced apoptosis and lipid peroxidation in CD8⁺ TILs compared to WT control (Fig. S12g, h). These results suggest that lactate enrichment in the TME promotes CD8⁺ T cell ferroptosis by inducing mitochondrial damage and activating the STING-TOX-HO-1 signaling axis.

**Fig. 5: Lactate accumulation in the tumor microenvironment promotes STING–TOX–HO-1–mediated ferroptosis in CD8⁺ T cells.**

Neoadjuvant combined therapy based on ACT using STING- or TOX-deficient engineered T cells together with ICIs, STING agonist or cisplatin enhances tumor clearance

ACT using Sting1-cKD or Tox-cKO CD8⁺ T cells, in combination with low-dose PD-1 or TIM-3 neutralizing antibodies, significantly enhanced antitumor efficacy compared to monotherapy with low-dose ICIs alone or with WT CD8⁺ T cells combined with ICIs (Figs. 6a–c and Supplementary Fig. S13). To further explore combinatorial strategies, we evaluated DDP, a widely used broad-spectrum chemotherapeutic agent that induced DNA damage, and the STING agonist ADU-S100^14,28. We found that DDP effectively suppressed TC-1 tumor growth in NSG mice, but ADU-S100 alone showed no significant antitumor activity in the murine tumor model (Fig. 6d, e). To assess potential synergy, we reconstituted CD8⁺ T cell-depleted CD45.1 WT mice (via CD8-neutralizing antibody) with CD45.2 Sting1-cKD, Tox-cKO, or WT CD8⁺ T cells and established TC-1 xenografts. Mice reconstituted with Sting1-cKD or Tox-cKO CD8⁺ T cells exhibited significantly reduced tumor growth rate and tumor weight compared to those receiving WT CD8⁺ T cells, regardless of DDP or ADU-S100 treatment (Figs. 6f–h). Strikingly, combining Sting1-cKD- or Tox-cKO CD8⁺ T cells with DDP or ADU-S100 further enhanced tumor suppression, underscoring the therapeutic potential of ACT using genetic engineering T cells by targeting T cell-intrinsic STING/TOX axis in combination with conventional chemotherapy or immunomodulatory agents.

**Fig. 6: Combination therapy with engineered STING- or TOX-deficient CD8⁺ T cells and checkpoint blockade, cisplatin, or a STING agonist enhances tumor control.**

Prognostic significance of TOX and HO-1 in tumor specimens from patients with cervical cancer

Single-cell RNA-seq of peripheral blood and tumor-infiltrating immune cells from patients with CC (n = 2), together with flow cytometry analysis (n = 7), revealed preferential TOX expression in tumor-derived CD8⁺ and CD4⁺ T-cell clusters relative to peripheral counterparts, STING transcripts were broadly expressed across CD4⁺ and CD8⁺ T-cell subsets, with relative enrichment in intratumoral CD4⁺ T cells, and HO-1 mRNA was lower in peripheral and tumor-associated T-cell clusters relative to peripheral and tumor-associated B-cell clusters (Fig. 7a, b). Furthermore, transcriptional levels of TOX and STING exhibited a positively correlated in tumor-infiltrating CD8⁺ T cells (Fig. 7c). In our previous study, we demonstrated that a high abundance of STING-positive lymphocytes within tumor tissues was associated with poor prognosis in patients with cervical cancer²⁹. Here, immunohistochemical (IHC) analysis of tumor specimens from 123 patients with CC revealed markedly higher levels of TOX and HO-1 in tumor-associated lymphocytes compared with adjacent non-tumor tissues (n = 14 paired samples; Fig. 7d). Strikingly, increased TOX expression or decreased HO-1 expression in tumor-infiltrating lymphocytes correlated with shorter disease-free survival (DFS) and overall survival (OS) in the CC cohort (n = 123; Fig. 7e, f). Together, these findings establish the STING-TOX–HO-1 axis as a prognostic determinant in cervical cancer progression.

**Fig. 7: Clinical relevance of TOX and HO-1 in cervical cancer.**

Discussion

The abundance of cytotoxic CD8⁺ T cells within the TME is a critical determinant of clinical outcomes in cancer patients undergoing immunotherapy, including ICB and ACT³. While activation-induced cell death (AICD) plays a key role in regulating T cell homeostasis during responses to allogeneic antigens, such as those encountered in viral infections or tumors, we demonstrate that STING- or TOX-deficient CD8⁺ T cells are resistance to ferroptosis, exhibiting enhanced tumor clearance, elevated intratumoral infiltration, increased cytotoxic cytokine production, a greater proportion of stem-like phenotypes, and reduced expression of exhaustion markers. These functional improvements are associated with upregulation of ferroptosis-inhibitory pathways and downregulation of genes involved in lipid peroxidation. Mechanistically, STING signaling acts in concert with TOX to repress HO-1 expression, leading to iron-overloaded mitochondrial dysfunction, characterized by reduced mitochondrial mass, diminished mtDNA copy numbers, impaired ATP synthesis, and elevated mitochondrial ROS and lipid peroxidation, culminating in CD8⁺ T cell ferroptosis. Tumor-derived lactate further amplifies this process through activation of STING and TOX pathways. Therapeutically, combining adoptive transfer of STING- or TOX-knockout engineered T cells with immune checkpoint inhibitors (anti-TIM-3/PD-1), STING agonizts (ADU-S100), or DDP significantly improved tumor clearance in murine models. Clinically, elevated TOX and reduced HO-1 levels in intratumoral T cells correlated with poor survival in patients with cervical cancer.

As an endoplasmic reticulum (ER)-resident adapter, STING is best known for mediating cytosolic DNA sensing through the cGAS-STING axis, thereby coordinating antimicrobial immunity and inducing type I interferons and proinflammatory cytokines³⁰. Beyond its canonical role, accumulating evidence implicates STING in diverse contexts, including age-associated cellular senescence and ROS/DNA damage–driven tumor sensitization^31,32. In tumor cells and myeloid populations, STING activation promotes antitumor immunity through type I IFN production^19,29,30. By contrast, TCR engagement in T cells can also trigger STING signaling via ER stress and the unfolded protein response (UPR), leading to apoptosis and impaired memory T cell differentiation through induction of C/EBP homologous protein (CHOP) and indoleamine 2,3-dioxygenase (IDO) activity^18,33. Moreover, T cell–intrinsic STING signaling has been shown to paradoxically promote immune tolerance by enhancing regulatory T cell expansion and driving calcium/ER stress–dependent apoptosis^18,29. Similarly, TOX, a key regulator of thymic development, has been shown to promote exhaustion in tumor-specific CD8⁺ T cell and serves as a predictive biomarker for immunotherapy response^21,34. In this study, we found that conditional knockout of STING or TOX in CD8⁺ T cells resulted in enhanced antitumor immunity, characterized by improved tumor clearance, increased intratumoral infiltration, a stem-like phenotype, and robust cytotoxic cytokine production. Furthermore, transcriptomic profiling of tumor-infiltrating CD8⁺ T cells revealed that STING or TOX deficiency reprograms pathways involved in iron transport, lipid metabolism, and granzyme-mediated apoptosis. These cells exhibited decreased expression of PD-1, TIM-3, ACSL4, and CTSB, but increased levels of HO-1 and GPX4. Functional assays confirmed reduced ferric ion accumulation and lipid peroxidation, alongside preserved mitochondrial integrity and viability under tumor challenge. This resistance to ferroptosis contrasts with established paradigms in which cGAS-STING activation promotes mtDNA release and ferroptotic cell death in tumor and myeloid cells, underscoring the cell type-specific regulation of this death pathway. STING activation has been reported to promote the formation of mitochondria-associated endoplasmic reticulum membranes (MAMs), thereby facilitating its trafficking from the ER to mitochondria^35,36. Intriguingly, in both in vivo and in vitro settings, we observed mitochondrial localization of key regulators of type I interferon signaling and ferroptosis, including STING, phosphorylated TBK1, IRF3, HO-1, and ACSL4, in CD8⁺ T cells activated by tumor cells. A comparable mechanism, whereby STING activation drives mitochondrial ferroptosis, has also been reported in cardiomyocytes subjected to ischemia–reperfusion injury³⁷. While prior studies have shown that viral activation of STING recruits the E3 ubiquitin ligase RNF5 for mitochondrial ubiquitination and degradation in myeloid cells³⁸, our data reveal that tumor-induced STING signaling instead upregulates TOX expression and elevates Mito-ROS and lipid peroxidation in murine and human CD8⁺ T cells (Fig. S7). In addition, we identified a STING-TOX transcriptional positive feedback loop at the mRNA level, with IRF3 knockdown (shIrf3) in human and murine CD8⁺ T cells markedly reduced TOX protein expression (Supplementary Figs. S8f and S9a). Mechanistically, CD8⁺ T cells exhibited reduced HO-1 expression, increased ferric ion accumulation and low mitochondrial mass upon tumor cell or viral antigen stimulation, alongside the phenotypes reversible upon inhibition of STING signaling or iron chelation. HO-1, a rate-limiting heme oxygenase, liberates redox-active iron, a recognized ferroptosis driver in sickle cell disease³⁹. However, HO-1 has been reported as a ferroptosis suppressor in liver cancer and renal proximal tubule cells^40,41. Our findings are consistent with this suppressive role: HO-1 attenuated ferroptosis by mitigating iron overload, Mito-ROS, and mitochondrial lipid peroxidation. Importantly, ectopic expression of TOX or STING, or pharmacological inhibition of HO-1, restored both iron accumulation and ferroptosis susceptibility in TC-1-stimulated STING- and TOX-deficient CD8⁺ T cells in vitro. Therapeutically, adoptive transfer of Lenti-shHmox1-transduced TOX-deficient CD8⁺ T cells abrogated the enhanced tumor control conferred by TOX deficiency in vivo (Supplementary Fig. S10). Collectively, these results establish the STING-TOX-HO-1 axis as a critical regulator of ferroptosis in CD8⁺ T cells within TME.

Ferroptosis, an iron-dependent form of metabolic cell death, plays a dual role in tumor biology⁴²; while its induction in tumor cells enhances antigen presentation and T cell cytotoxicity to suppress malignancies, it paradoxically undermines antitumor immunity by targeting infiltrating immune cells in the TME⁴³. In the present study, we further demonstrated that tumor-infiltrating CD8⁺ T cells from patients with CC and NPC displayed heightened ferroptotic signatures, including increased STING activation and lipid peroxidation, relative to circulating counterparts (Supplementary Fig. S6). In vitro, tumor cell engagement induced STING and TOX signaling, suppressed HO-1 expression, and provoked mitochondrial dysfunction in CD8⁺ T cells (Supplementary Fig. S7); While specific inhibition of HO-1 in T cells declined the enhanced anti-tumor immunity of TOX-deficient T cells in the murine tumor model with ACT (SupplementaryFig. S10). Effector T cell functionality in the TME is regulated not only by canonical TCR (signal 1), co-stimulatory (signal 2), and cytokine (signal 3) pathways but also by metabolic cues (signal 4), including nutrient availability, ion homeostasis, and local pH⁴⁴. Notably, TME-enriched factors such as labile iron, cholesterol, and fatty acids have been shown to drive ferroptosis in immune cells^43,45. For instance, elevated fatty acids promote CD8⁺ T cell ferroptosis via CD36-mediated lipid uptake, thereby impairing their antitumor capacity⁴³ Furthermore, the Warburg effect in tumor cells exacerbates lactate accumulation in the TME, a metabolite linked to metastatic progression and poor clinical outcomes⁴⁶. Lactate has been shown to suppresses T and NK cell survival, facilitating immune evasion^47,48,49. Lactate, abundantly secreted by tumors, selectively sensitizes murine CD8⁺ T cells but not tumor cells to mitochondrial reprogramming and ferroptosis (Supplementary Fig. S11a–g). Strikingly, STING- or TOX-deficient CD8⁺ T cells exhibited resistance to lactate-driven ferroptosis and mitochondrial dysregulation, a phenotype potentially linked to altered expression of Slc16a family genes (SupplementaryFig. S11h, i). WT-CD8⁺ T cells showed upregulation of STING and TOX signals and suppression of HO-1 expression under lactate stress, all of which were reversed by the MCT1 inhibitor AZD3965. In contrast, this reversal was markedly diminished in Sting1-cKO or Tox-cKO CD8⁺ T cells, consistent with defective activation of STING signaling and ferroptosis pathways. These findings identify lactate as a TME-derived mediator of CD8⁺ T cell ferroptosis through activation of the STING-TOX-HO-1 axis.

The clinical efficacy of ICB (e.g., anti-TIM-3), T cell-based adoptive cell therapy, and targeted agents such as STING agonizts in advanced cancers is substantially limited by the heterogeneity, complexity, and spatiotemporal dynamics of the tumor immune microenvironment^4,14,50. This limitation persists despite the robust antitumor immunity elicited by STING agonizts in preclinical tumor models^51,52,53. Drug resistance remains a major therapeutic barrier, promoting a shift toward neoadjuvant combinatorial strategies aimed at mitigating toxicity and limited efficacy associated with monotherapies⁵⁴. Preclinical studies have demonstrated that combining TIL-based ACT with PD-1 blockade can overcome ICB resistance in solid tumors, while synergizing ADU-S100 or CAR-T-based ACT with PD-1 inhibitors enhances therapeutic outcomes^15,55,56. In this study, we show that STING- or TOX-deficient engineered T cells not only augment the efficacy of ACT but also improve tumor clearance in neoadjuvant regimens combining ACT with TIM-3/PD-1 blockade, DDP, or ADU-S100 in murine models. Notably, ADU-S100 monotherapy lacked efficacy in NSG mice, underscoring its dependent on immune activation rather than direct tumor cytotoxicity. In contrast, DDP, a DNA-damaging agent, amplifies STING signaling across multiple cellular TME compartments⁵⁷. Our findings suggest that inhibition of intracellular STING/TOX pathways enhances CD8⁺ T cell survival and effector functions (e.g., cytokine secretion), thereby potentiating ICB, chemotherapy, and targeted therapies in neoadjuvant settings. Clinically, elevated STING, TOX and reduced HO-1 expression in CC patients were associated with poor prognosis, reinforcing their respective roles in promoting immune tolerance. Collectively, we demonstrate that lactate accumulation in the TME induces iron-overloaded mitochondrial reprogramming and ferroptosis in CD8⁺ T cells via the STING-TOX-HO-1 axis, contributing to immune escape and adverse clinical outcomes. Importantly, the use of STING- or TOX-deficient engineered T cells significantly enhances tumor clearance in preclinical models when combined with TIM-3/PD-1 blockade, cisplatin, or ADU-S100 in neoadjuvant ACT regimens (Fig. 7f). These findings reveal a potential strategy to enhance anti-tumor immunity through modulation of T cell-intrinsic STING–TOX signaling and provide a rationale for optimizing neoadjuvant combination therapies integrating ICB, chemotherapy, targeted agents, and engineered adoptive T-cell therapy.

Methods

Generation of CD8 conditional STING or TOX knockout mice

Tox^flox/flox and Cd8^cre CD45.2 C57BL/6 J mice were purchased from Gempharmatech Co., Ltd (JiangSu, China). Sting1^flox/flox CD45.2 C57BL/6 J mice were a gift from XJ. Xia (Sun Yat-sen University). Sting1^flox/flox or/and Tox^flox/flox CD45.2 C57BL/6 J mice were crossed with Cd8^cre CD45.2 C57BL/6 J mice to generate STING or TOX conditional knockout (cKO) or knockdown (cKD) mice: Cd8^creSting1^flox/wt (Sting1-cKD), Cd8^creSting1^flox/flox (Sting1-cKO), Cd8^creTox^flox/flox (Tox-cKO) and Cd8^creSting1^flox/wtTox^flox/wt (Sting1/Tox-D-cKD) mice. Wild type (WT) CD45.2 and CD45.1 C57BL/6 J mice, Sting1^flox/flox and Tox^flox/flox CD45.2 C57BL/6 J mice were included as controls. Mice were bred and maintained under specific pathogen-free conditions in the Animal Facility of the Sun Yat-sen University Cancer Center, and all experiments were performed according to the regulations of Sun Yat-sen University (L102012022001Z). For all experiments, 6 to 8-week-old mice were used unless otherwise indicated, with a randomized gender. Mice were kept under standard housing conditions with a 12:12 h light-dark cycle, an ambient temperature between 20–24 °C and humidity between 45–65%.

Human and murine cell lines culture

Human HEK-293T (ATCC, #CRL-3216), human cervical cancer cell line SiHa (ATCC, #HTB-35), human NPC cell lines HNE-1, C666 (provided from Dr. Mu-sheng Zeng, Sun Yat-sen University Cancer Center), and EBV⁺-TW03 (provided from Maria G Masucci, Karolinska Institutet), murine TC-1 (Jining BIO, #JN26363), B16F10 (ATCC, #CRL-6475), MC38 (Procell, #CL--0972) and CT26 (ATCC, #CRL-2638) cell lines were cultured in Dulbecco’s Modified Eagle Medium (Gibco) or RPMI-1640 (Gibco) supplemented with 10% fetal bovine serum (ExcellBio, China). All of the cell lines were subjected to the mycoplasma test and were found to be free of contamination.

Human samples

Paraffin-embedded tumor tissues (n = 123) were collected from patients diagnosed with primary CC patients at the Sun Yat-Sen University Cancer (Guangzhou, China) from 2011 to 2015. The detailed clinical data are described in Supplementary Table 1. Peripheral blood from healthy donors was also included in this study.

Isogenous xenograft tumor models and therapy models

Isogenous xenograft tumor models

Sting1-cKO, Sting1-cKD, Tox-cKO, Sting1/Tox-D-cKD and WT mice (4 mice/group) were injected subcutaneously in the right rear flank with 4 × 10⁵ TC-1 cells, 1 × 10⁵ MC38 cells, and 2 × 10⁵ B16F10 cells. In ACT models, TC-1-bearing NSG mice (5 mice/group) were administered preactivated 5 × 10⁵ CD45.1⁺ WT-CD3⁺ T cells in combination with 5 × 10⁵ CD45.2⁺ CD3⁺ T cells derived from WT, Sting1-cKD, Tox-cKO, or Sting1/Tox-D-cKD mice. Tumor growth was monitored every 3 days, and the tumor volume was calculated using the following formula: V = 0.5 × L × W², where L is the length, and W is the width (in mm). The mice were sacrificed on the indicated days. The spleen, lymph nodes, and partial tumor were acquired from tumor-bearing mice and used for T cell isolation, and the partial tumor was fixed by formalin (Biosharp, China) for IHC or IF staining analysis.

Adoptive cell therapy using T cells with STING or HO-1 inhibition

For H151 (STING inhibitor)-treated-CD8⁺ T cell adoptive therapy, CD8⁺ T cells isolated from TC-1-preimmune CD45.2 WT mice were stimulated with 5 μg/mL H151 (MCE, USA) for 2 days in vitro. For Hmox1-knockdown (Lentivirus containing shHmox1 vector)-CD8⁺ T cells adoptive therapy, CD8⁺ T cells were isolated from the spleens and lymph nodes of TC-1-preimmune WT and Tox-cKO mice; TC-1-activated-WT-CD8⁺ T cells were treated with Lenti-pLKO (pLKO, empty vector), and TC-1-activated-Tox-cKO CD8⁺ T cells were transfected with Lenti-shHmox1 for 48 hours. CD45.1 WT mice (4 mice/group) were injected subcutaneously in the right rear flank with 4 × 10⁵ TC-1 cells. TC-1-bearing CD45.1 WT mice were intravenously injected with 1 × 10⁶ H151-treated-WT CD8⁺ T cells or shHmox1-transfected-Tox-cKO CD8⁺ T cells. After the mice were sacrificed, the tumor tissues were excised and weighed, and the T cells were isolated from the tumor tissues and used for subsequent experiments.

Combined therapy with STING- or TOX-deficient CD8⁺ T cells and immune checkpoint inhibitors, STING agonist or cisplatin in murine tumor model

For the ICB therapy group, WT mice (5 mice/group) were injected subcutaneously in the right rear flank with 2 × 10⁵ TC-1 cells, and the TC-1-bearing WT mice were injected with anti-PD1 and anti-TIM3 antibodies (200 μg per mouse) on days 19, 21, 24, and 27. After these mice were sacrificed on day 37, the tumor tissues were excised and weighed. For combination therapy groups receiving ACT with engineered STING-, TOX-deficient or WT CD8⁺ T cells and low-dose PD-1 or TIM-3 neutralizing antibodies (50 μg per mouse), the TC-1-bearing CD45.2 WT mice were intravenously injected with 1 × 10⁶ TC-1-preactivated CD8⁺ T cells from Sting1-cKD-, Tox-cKO-, or WT mice (4 mice/group) and intraperitoneal injected with anti-PD-1 and anti-TIM-3 antibodies (50 μg per mouse, BIOXCELL, USA) on days 4, 7, 10, and 13. After the mice were sacrificed on day 22, the tumor tissues were excised and weighed.

For the ADU-S100 and DDP monotherapy group: NSG mice (5 mice/group) were injected subcutaneously in the right rear flank with 4 × 10⁵ TC-1 cells. TC-1-bearing NSG mice were intraperitoneal injected with DDP (50 μg per mouse) or intratumoral injected with ADU-S100 (25 μg per mouse) on days 6, 9, 12, and 15. The tumor was measured every 3 days, and weighted after the mice were sacrificed. Immune reconstitution: the endogenous CD8⁺ T cells of CD45.1 WT mice were depleted by intraperitoneal injected with mouse CD8-neutralizing antibody (200 μg per mouse, ‌BIOXCELL, USA), and then intravenously injected with 1 × 10⁶ CD45.2 Sting1-cKD-, Tox-cKO-, or WT-CD8⁺ T cells. The immune reconstituted CD45.1 mice were injected subcutaneously in the right rear flank with 4 × 10⁵ TC-1 cells (4 per group). TC-1-bearing immune reconstituted CD45.1 WT mice were treated with DDP (50 μg per mouse, intraperitoneal injection, i.v.) on days 8, 11, 14, 17, and 20 or ADU-S100 (25 μg per mouse, intratumoral administration, i.t.) on days 9, 12, 15, and 18. The tumor was measured every 3 days, and weighted after the mice were sacrificed.

T cell isolation, transwell co-culture and transfection system

Lymphocytes were separated from health human donors or mice using human or murine Ficoll (Dakewe, China) and isodensity centrifugation, and cultured in OKT-3-coated 24-well plates (Corning, USA) in X-VIVO medium (Lonza, USA) with IL-2. CD8⁺ T cells isolated from lymphocytes using the CD4-positive selection kit and cocultured with tumor cells (Transwell, 0.4 μm pore size), antigens peptide (0.05 mg/mL E6/E7 or 0.05 mg/mL LMP1), cell death inhibitors, indicated concentrations of ferroptosis-inducer Erastin, 1 μg/mL DMXAA (STING agonist), TC-1 cells in the presence of increasing concentrations of the iron chelator DFO, indicated concentrations of lactate, or 500 nM AZD3965 (MCT1 inhibitor) for 48 h in OKT-3-coated 24-well plates in vitro.

Lentivirus was produced containing mouse TOX and STING full-length protein expression vector PCDH-TOX (TOX-OE) and PCDH-STING (STING-OE) or mouse Hmox1 and Irf3 knockdown vector pLKO-shHmox1 (shHmox1) and pLKO-shIrf3 (shIrf3), or human STING1- and TOX-sgRNAs recombinant plasmid. Lentivirus was added to 1 × 10⁶ enriched CD8⁺ T cells supplemented with polybrene (10 μg/mL, Sigma-Aldrich, USA) and centrifuged at 800 × g at 37 °C for 1 h. The lentivirus-treated cells were cocultured with TC-1 in OKT-3-coated 24-well plates with a transwell co-culture system for 48 h in vitro. The primers of TOX and STING overexpression and shHmox1 and shIrf3 sequences are listed in Supplementary Dataset 1.

Bulk and single-cell transcriptome profile and quantitative RT-PCR analysis

Bulk RNA sequencing

CD8⁺ T cells from tumor tissues from tumor-bearing Sting1-cKO, Tox-cKO and WT mice were sorted using the mouse CD8a-positive selection kit (Stemcell, Canada) and used for bulk RNA-seq. RNA-seq libraries were sequenced by Illumina (Illumina, USA) with 150-bp paired-ends at Ribobio Co., Ltd (Guangzhou, China). Differential expression analysis was performed using the edgeR R package⁵⁸ with the read counts as input. Gene set enrichment analysis (GSEA) was performed by the clusterProfiler R package⁵⁹. The raw sequence data were deposited in the GSA database (Accession number: CRA011572) and publicly accessible at https://ngdc.cncb.ac.cn/.

Single-cell RNA sequencing

A single-cell library was constructed using Chromium Single Cell 3′ Reagent Kits following the manufacturer’s instructions (10 × Genomics) at Guangzhou Saliai Stem Cell Science and Technology Co., Ltd (Guangzhou, China). The gene expression data were aligned to the human genome reference sequence (GRCh38) using the Cellranger pipelines (version 3.0.2, USA). The DoubletFinder package was used to remove doublets, with an expected doublet rate of 0.04⁴⁶. The data were analyzed by the Seurat R package⁶⁰. DEGs were identified using the ‘FindMarkers’ function in the Seurat R package, with a significance threshold of p < 0.05 and |log fold change | > 0.25. Gene signatures were calculated using the ‘AddModuleScore’ function in the Seurat R package. A list of cell death-related genes can be found in Supplementary Dataset 1. GSEA was performed by the clusterProfiler R package⁶¹. The raw sequence data have been deposited in the Genome Sequence Archive⁵⁸ in the National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA-Human: HRA003358) and publicly accessible at https://ngdc.cncb.ac.cn/gsa-human. Single-cell RNA-seq data for nasopharyngeal carcinoma (GEO: GSE162025) was publicly accessible at the GEO database (http://www.ncbi.nlm.nih.gov/geo/).

Real-time quantitative PCR analysis

According to the manufacturer’s instructions, total RNA was extracted from cells using the EZ-press RNA Purification Kit (EZBioscience, USA). Quantitative real-time polymerase chain reaction (qPCR) was performed using a RevertAid First-Strand cDNA Synthesis Kit (Invitrogen, USA) and qPCR SYBR Master Mix (Vazyme, China). The primers are synthesized by Rui Biotech (Beijing, China) and primer sequences are shown in Supplementary Dataset 1.

Flow cytometry, immunofluorescence, immunoblot and Immunohistochemistry

Flow cytometry

Cells were washed twice using phosphate buffered saline (PBS) (Biosharp, China), labeled with fixable viability dye (eBioscience, USA), and stained for biomarkers of CD3, CD8, CD4, CD19, CD16, Ki67, PD1, TIM-3, CXCR5, TCF-1, Ly108, CD69, PI, and Annexin Ⅴ using fluorochrome-conjugated antibodies (see Supplementary Dataset 1) according to the manufacturer’s instructions. Intracellular cytokine IFN-γ and GrB staining was performed using a fixation/permeabilization solution kit (BD, USA) following the manufacturer’s instructions. For the cell proliferation assay, cells were labeled with CellTrace CFSE (CFDA-SE, Invitrogen, USA) according to the manufacturer’s instructions. Data were acquired by a Beckman Colter flow cytometer (CytoFLEX LX, Beckman, Germany) and analyzed by FlowJo software (BD, USA).

Immunofluorescence

Multiplex immunofluorescence staining was performed using the Opal Polaris 7-Color Manual IHC Kit (Akoya Biosciences, USA) according to the manufacturer’s instructions. The following primary antibodies were used with designated fluorophores: anti-mouse CD8 (white), TOX (yellow), PD-1 (red), TIM-3 (green), HO-1 (pink), GPX4 (green), and CTSB (orange) (see Supplementary Dataset 1 for details). Nuclei were counterstained with DAPI. Images were acquired using the Vectra 3.0 imaging system and analyzed with Vectra software (Akoya Biosciences, USA). Quantification of CD8⁺, TOX⁺, PD-1⁺, TIM-3⁺, HO-1⁺, GPX4⁺, and CTSB⁺ cells was performed using HALO software (Indica Labs). The frequencies of CD8⁺TOX⁺, CD8⁺PD-1⁺, CD8⁺TIM-3⁺, CD8⁺HO-1⁺, CD8⁺GPX4⁺, and CD8⁺CTSB⁺ subsets were calculated as a proportion of total CD8⁺ T cells across the entire tissue section.

Immunoblot

Murine or human CD8⁺ T cells were subjected to lysis in ice-cold lysis buffer (Beyotime, China) supplemented with 5 mg/mL protease inhibitor cocktail (Coolaber, China). Extracts (15–20 μL) were subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis‌ (SDS-PAGE). The antibodies used are available in Supplementary Dataset 1.

Immunohistochemistry

Tumor tissue sections from 123 CC patients were incubated with anti-TOX and anti-HO-1 at 4 °C overnight (see Supplementary Table 1 and 2), and operated using an IHC kit (Zhongshan Jinqiao Biotechnology, China) according to the manufacturer’s instructions. The secondary antibody (Zhongshan Jinqiao Biotechnology, China) was incubated for 30 minutes at room temperature, and 3,3’-diaminobenzidine (DAB, Zhongshan Jinqiao Biotechnology, China) was used for visualization. Finally, the pathological sections were counterstained with hematoxylin (Zhongshan Jinqiao Biotechnology, China) and then dehydrated and sealed with neutral gum (Jieli Biotechnology, China) for optical microscopy (ECLIPSE Ni, NIKON, Japan).

Ferroptosis and metabolism assay

Cell death and viability assay

A total of 2 × 10⁵ pre-treated-CD8⁺ T cells were seeded into 96-well plates (Corning, USA) and 10 μL of CCK-8 (Mien, China) mixed into each well. The absorbance at 450 nm was measured after 3 h using a microplate reader (Bio-Tek EPOCH2, USA).

Metabolism measurement

Experiments were performed according to the manufacturer’s protocol. Briefly, cells were washed with phosphate-buffered saline (PBS) twice and stained with 10 nM of Phen Green SK (Glpbio, USA) for 15 min at 37 °C, Lipid Peroxidation Sensor (1 μM, Thermo Fisher, USA) for 30 minutes at 37 °C, MitoTracker (25 nM, ThermoFisher, USA) for mitochondrial mass and MitoSOX Red (1 μM, Beyotime, China) for Mito-ROS for 30 minutes at 37 °C, and MitoPePP (0.2 μM, DOJINDO, Japan) for mitochondrial lipid peroxidation for 15 min at 37 °C. Cells were washed with PBS and examined by flow cytometry.

Nuclear cathepsin B and total cellular oxidized lipid peroxides measurement

Cells were stained by CTSB primary antibody by an Opal Polaris 7 color manual immunohistochemistry kit (Akoya Biosciences, USA). For cytosolic lipid peroxidation staining, cells incubated in 1640 cell culture medium with Lipid Peroxidation Sensor (1 μM, Thermo Fisher, USA) in a humidified chamber at 37 °C with 5% CO₂ for 30 min. All stained cells were smeared to polylysine-coated slide and confocal laser microscopy LSM880 (Zeiss, Germany), and ImageJ plugin StarDist (NIH, USA) were used to analyze the images.

Quantification of ATP production and mtDNA content

The ATP was determined by use of the ATP determination assay (Beyotime Biotechnology, China). Mitochondrial DNA (mtDNA) was quantified with real-time PCR. Total DNA was isolated from CD8⁺ T cells using the Phenol-chloroform extraction method, and mtDNA was quantified using primers specific for the mouse mtCo2 gene and normalized using primers specific for the murine reference gene App. The primer sequences are shown in Supplementary Dataset 1.

Statistical analysis

All statistical analyses were performed using R software (version 4.0.3), GraphPad Prism 9 (La Jolla, CA, USA), and SPSS software (version 18.0; Chicago, IL, USA). Continuous variables are presented as the mean ± standard deviation (SD) and were compared using the unpaired Student’s t test or one-way analysis of variance (ANOVA), as appropriate. Nonparametric data were analyzed using the two-sided Wilcoxon test. Patient survival was assessed using the Kaplan–Meier method, and differences between groups were evaluated with the log-rank test. Optimal cut-off values were determined using the survival and survminer packages in R. A P-value < 0.05 was considered statistically significant; ns indicates no statistical significance (P > 0.05).

Ethical approval

The study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board (IRB) of Sun Yat-sen University Cancer Center (G2022-133-01), and all patients and healthy donors signed a consent form. All experiments involved mice were performed according to the regulations of Sun Yat-sen University (L102012022001Z), and the maximum tumor volume is limited to 2000 mm³, or when the tumor metastasizes or grows rapidly to ulcerate, causing infection or necrosis.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The authenticity of this article has been validated by uploading the key raw data to the Research Data Deposit (RDD) public platform [https://www.researchdata.org.cn/] under RDD approval number RDDB2025700905. The single-cell RNA sequencing data generated in this study have been deposited in the GSA-Human database under accession code HRA003358 [https://ngdc.cncb.ac.cn/gsa-human/]. The bulk RNA sequencing data generated in this study have been deposited in the GSA database under accession code CRA011572 [https://ngdc.cncb.ac.cn/]. Single-cell RNA-seq data for nasopharyngeal carcinoma (GEO: GSE162025) was publicly accessible at the GEO database [http://www.ncbi.nlm.nih.gov/geo/]. All other data are available in the article and Supplementary files or from the corresponding author, Jiang Li (lijiang2@mail.sysu.edu.cn) upon request. Source data are provided in this paper.

References

Zebley, C. C., Zehn, D., Gottschalk, S. & Chi, H. T cell dysfunction and therapeutic intervention in cancer. Nat. Immunol. 25, 1344–1354 (2024).
Article CAS PubMed PubMed Central Google Scholar
Wang, T. W. et al. Blocking PD-L1-PD-1 improves senescence surveillance and ageing phenotypes. Nature 611, 358–364 (2022).
Article ADS CAS PubMed Google Scholar
Philip, M. & Schietinger, A. CD8(+) T cell differentiation and dysfunction in cancer. Nat. Rev. Immunol. 22, 209–223 (2022).
Article CAS PubMed Google Scholar
Memon, D. et al. Clinical and molecular features of acquired resistance to immunotherapy in non-small cell lung cancer. Cancer Cell 42, 209–224 (2024).
Article CAS PubMed PubMed Central Google Scholar
Harding, J. J. et al. Blocking TIM-3 in treatment-refractory advanced solid tumors: a phase Ia/b study of LY3321367 with or without an anti-PD-L1 antibody. Clin. Cancer Res. 27, 2168–2178 (2021).
Article CAS PubMed Google Scholar
Schoenfeld, A. J. & Hellmann, M. D. Acquired Resistance to Immune Checkpoint Inhibitors. Cancer Cell 37, 443–455 (2020).
Article CAS PubMed PubMed Central Google Scholar
Oladejo, M., Paulishak, W. & Wood, L. Synergistic potential of immune checkpoint inhibitors and therapeutic cancer vaccines. Semin. Cancer Biol. 88, 81–95 (2023).
Article CAS PubMed Google Scholar
Lorusso, D. et al. Pembrolizumab or placebo with chemoradiotherapy followed by pembrolizumab or placebo for newly diagnosed, high-risk, locally advanced cervical cancer (ENGOT-cx11/GOG-3047/KEYNOTE-A18): a randomised, double-blind, phase 3 clinical trial. Lancet 403, 1341–1350 (2024).
Article CAS PubMed Google Scholar
Chen, C. & Xu, P. Cellular functions of cGAS-STING signaling. Trends Cell Biol. 33, 630–648 (2023).
Article CAS PubMed Google Scholar
Li, W. et al. cGAS-STING-mediated DNA sensing maintains CD8(+) T cell stemness and promotes antitumor T cell therapy. Sci. Transl. Med. 12, eaay9013(2020).
Rai, P. et al. IRGM1 links mitochondrial quality control to autoimmunity. Nat. Immunol. 22, 312–321 (2021).
Article CAS PubMed PubMed Central Google Scholar
Wang, Q. et al. The E3 ubiquitin ligase AMFR and INSIG1 bridge the activation of TBK1 kinase by modifying the adaptor STING. Immunity 41, 919–933 (2014).
Article CAS PubMed Google Scholar
Somatilaka, B. N. et al. STING activation reprograms the microenvironment to sensitize NF1-related malignant peripheral nerve sheath tumors for immunotherapy. J. Clin. Invest. 134, e176748 (2024).
Meric-Bernstam, F. et al. Phase I dose-escalation trial of MIW815 (ADU-S100), an intratumoral STING agonist, in patients with advanced/metastatic solid tumors or lymphomas. Clin. Cancer Res. 28, 677–688 (2022).
Article CAS PubMed Google Scholar
Meric-Bernstam, F. et al. Combination of the STING agonist MIW815 (ADU-S100) and PD-1 inhibitor spartalizumab in advanced/metastatic solid tumors or lymphomas: an open-label, multicenter, phase Ib study. Clin. Cancer Res. 29, 110–121 (2023).
Article CAS PubMed PubMed Central Google Scholar
Knox, K. et al. Dynamic STING repression orchestrates immune cell development and function. Sci. Immunol. 10, eado9933 (2025).
Article CAS PubMed Google Scholar
Larkin, B. et al. Cutting edge: activation of STING in T cells induces type I IFN responses and cell death. J. Immunol. 199, 397–402 (2017).
Article CAS PubMed PubMed Central Google Scholar
Wu, J. et al. STING-mediated disruption of calcium homeostasis chronically activates ER stress and primes T cell death. J. Exp. Med. 216, 867–883 (2019).
Article CAS PubMed PubMed Central Google Scholar
Zhang, S. et al. Manganese induces tumor cell ferroptosis through type-I IFN dependent inhibition of mitochondrial dihydroorotate dehydrogenase. Free Radic. Biol. Med. 193, 202–212 (2022).
Article CAS PubMed Google Scholar
Galassi, C., Chan, T. A., Vitale, I. & Galluzzi, L. The hallmarks of cancer immune evasion. Cancer Cell 42, 1825–1863 (2024).
Article CAS PubMed Google Scholar
Scott, A. C. et al. TOX is a critical regulator of tumour-specific T cell differentiation. Nature 571, 270–274 (2019).
Article CAS PubMed PubMed Central Google Scholar
Khan, O. et al. TOX transcriptionally and epigenetically programs CD8(+) T cell exhaustion. Nature 571, 211–218 (2019).
Article CAS PubMed PubMed Central Google Scholar
Seo, H. et al. TOX and TOX2 transcription factors cooperate with NR4A transcription factors to impose CD8(+) T cell exhaustion. Proc. Natl. Acad. Sci. USA 116, 12410–12415 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Poss, K. D. & Tonegawa, S. Heme oxygenase 1 is required for mammalian iron reutilization. Proc. Natl. Acad. Sci. USA 94, 10919–10924 (1997).
Article ADS CAS PubMed PubMed Central Google Scholar
Kovtunovych, G., Eckhaus, M. A., Ghosh, M. C., Ollivierre-Wilson, H. & Rouault, T. A. Dysfunction of the heme recycling system in heme oxygenase 1-deficient mice: effects on macrophage viability and tissue iron distribution. Blood 116, 6054–6062 (2010).
Article CAS PubMed PubMed Central Google Scholar
Deng, H. et al. Ubiquinol-mediated suppression of mitochondria-associated ferroptosis is a targetable function of lactate dehydrogenase B in cancer. Nat. Commun. 16, 2597 (2025).
Article ADS CAS PubMed PubMed Central Google Scholar
Felmlee, M. A., Jones, R. S., Rodriguez-Cruz, V., Follman, K. E. & Morris, M. E. Monocarboxylate transporters (SLC16): function, regulation, and role in health and disease. Pharmacol. Rev. 72, 466–485 (2020).
Article CAS PubMed PubMed Central Google Scholar
Kowalczyk, J. T. et al. Exploiting the immunogenic potential of standard of care radiation or cisplatin therapy in preclinical models of HPV-associated malignancies. J. Immunother. Cancer 10, 142503 https://doi.org/10.1136/jitc-2022-005752 (2022).
Ni, H. et al. T cell-intrinsic STING signaling promotes regulatory T cell induction and immunosuppression by upregulating FOXP3 transcription in cervical cancer. J. Immunother. Cancer 10, https://doi.org/10.1136/jitc-2022-005151 (2022).
Sun, L., Wu, J., Du, F., Chen, X. & Chen, Z. J. Cyclic GMP-AMP synthase is a cytosolic DNA sensor that activates the type I interferon pathway. Science 339, 786–791 (2013).
Article ADS CAS PubMed Google Scholar
Glück, S. et al. Innate immune sensing of cytosolic chromatin fragments through cGAS promotes senescence. Nat. Cell Biol. 19, 1061 (2017).
Article PubMed PubMed Central Google Scholar
Dou, Z. X. et al. Cytoplasmic chromatin triggers inflammation in senescence and cancer. Nature 550, 402–406 (2017).
Article ADS CAS PubMed PubMed Central Google Scholar
Quaney, M. J. et al. STING controls T cell memory fitness during infection through T cell-intrinsic and IDO-dependent mechanisms. Proc. Natl. Acad. Sci. USA 120, e2205049120 (2023).
Article CAS PubMed PubMed Central Google Scholar
Sekine, T. et al. TOX is expressed by exhausted and polyfunctional human effector memory CD8(+) T cells. Sci. Immunol. 5, https://doi.org/10.1126/sciimmunol.aba7918 (2020).
Padhiar, A. A. et al. MAM-STAT3-driven mitochondrial Ca(+2) upregulation contributes to immunosenescence in type a mandibuloacral dysplasia patients. Adv. Sci. 12, e2407398 (2025).
Article Google Scholar
Liu, J. et al. cGAS-STING, inflammasomes and pyroptosis: an overview of crosstalk mechanism of activation and regulation. Cell Commun. Signal 22, 22 (2024).
Article CAS PubMed PubMed Central Google Scholar
Wang, X. et al. STING aggravates ferroptosis-dependent myocardial ischemia-reperfusion injury by targeting GPX4 for autophagic degradation. Signal Transduct. Target. Ther. 10, 136 (2025).
Article PubMed PubMed Central Google Scholar
Zhong, B. et al. The ubiquitin ligase RNF5 regulates antiviral responses by mediating degradation of the adaptor protein MITA. Immunity 30, 397–407 (2009).
Article CAS PubMed Google Scholar
Menon, A. V. et al. Excess heme upregulates heme oxygenase 1 and promotes cardiac ferroptosis in mice with sickle cell disease. Blood 139, 936–941 (2022).
Article CAS PubMed PubMed Central Google Scholar
Sun, X. et al. Activation of the p62-Keap1-NRF2 pathway protects against ferroptosis in hepatocellular carcinoma cells. Hepatology 63, 173–184 (2016).
Article CAS PubMed Google Scholar
Adedoyin, O. et al. Heme oxygenase-1 mitigates ferroptosis in renal proximal tubule cells. Am. J. Physiol. Renal. Physiol. 314, F702–F714 (2018).
Article CAS PubMed Google Scholar
Jiang, X., Stockwell, B. R. & Conrad, M. Ferroptosis: mechanisms, biology and role in disease. Nat. Rev. Mol. Cell Biol. 22, 266–282 (2021).
Article PubMed PubMed Central Google Scholar
Ma, X. et al. CD36-mediated ferroptosis dampens intratumoral CD8(+) T cell effector function and impairs their antitumor ability. Cell Metab. 33, 1001–1012.e1005 (2021).
Article CAS PubMed PubMed Central Google Scholar
Raynor, J. L. & Chi, H. B. Nutrients: Signal 4 in T cell immunity. J. Exp. Med. 221, https://doi.org/10.1084/jem.20221839 (2024).
Xu, S. et al. Uptake of oxidized lipids by the scavenger receptor CD36 promotes lipid peroxidation and dysfunction in CD8(+) T cells in tumors. Immunity 54, 1561–1577 (2021).
Article CAS PubMed PubMed Central Google Scholar
Dai, E., Wang, W., Li, Y., Ye, D. & Li, Y. Lactate and lactylation: Behind the development of tumors. Cancer Lett. 591, 216896 (2024).
Article CAS PubMed Google Scholar
Quinn, W. J. et al. Lactate Limits T Cell Proliferation via the NAD(H) Redox State. Cell Rep. 33, https://doi.org/10.1016/j.celrep.2020.108500 (2020).
Lopez, E. et al. Inhibition of lactate transport by MCT-1 blockade improves chimeric antigen receptor T-cell therapy against B-cell malignancies. J. Immunother. Cancer 11, https://doi.org/10.1136/jitc-2022-006287 (2023).
Brand, A. et al. LDHA-Associated lactic acid production blunts tumor immunosurveillance by T and NK cells. Cell Metab. 24, 657–671 (2016).
Article CAS PubMed Google Scholar
Hsu, J. M. et al. MTAP deficiency confers resistance to cytosolic nucleic acid sensing and STING agonists. Science 390, eadl4089 (2025).
Article CAS PubMed Google Scholar
Zhang, H. et al. Endothelial STING-JAK1 interaction promotes tumor vasculature normalization and antitumor immunity. J. Clin. Invest. 135, https://doi.org/10.1172/jci180622 (2025).
Girard, M. et al. STING agonists drive recruitment and intrinsic type I interferon responses in monocytic lineage cells for optimal anti-tumor immunity. J. Immunol. 214, 3634–3646 (2025).
Ramanjulu, J. M. et al. Design of amidobenzimidazole STING receptor agonists with systemic activity. Nature 564, 439–443 (2018).
Article ADS CAS PubMed Google Scholar
Chi, H., Pepper, M. & Thomas, P. G. Principles and therapeutic applications of adaptive immunity. Cell 187, 2052–2078 (2024).
Article CAS PubMed PubMed Central Google Scholar
Hall, M. S. et al. Combination nivolumab, CD137 agonism, and adoptive cell therapy with tumor-infiltrating lymphocytes for patients with metastatic melanoma. Clin. Cancer Res. 28, 5317–5329 (2022).
Article CAS PubMed PubMed Central Google Scholar
L’Orphelin, J. M. et al. NIVO-TIL: combination anti-PD-1 therapy and adoptive T-cell transfer in untreated metastatic melanoma: an exploratory open-label phase I trial. Acta Oncol. 63, 867–877 (2024).
Article PubMed PubMed Central Google Scholar
Cohen, S. M. & Lippard, S. J. Cisplatin: from DNA damage to cancer chemotherapy. Prog. Nucleic Acid Res. Mol. Biol. 67, 93–130 (2001).
Article CAS PubMed Google Scholar
Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010).
Article CAS PubMed Google Scholar
Wu, T. et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innovation 2, 100141 (2021).
CAS PubMed PubMed Central Google Scholar
Stuart, T. et al. Comprehensive Integration of Single-Cell Data. Cell 177, 1888–1902 (2019).
Article CAS PubMed PubMed Central Google Scholar
Xu, C. et al. Comprehensive multi-omics single-cell data integration reveals greater heterogeneity in the human immune system. iScience 25, 105123 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

This work was supported by the National Key Research and Development Program (2025YFC3409800) and the National Natural Science Foundation of China (grant no. 82272780 and 81972874 to J.L., and 8216110322 to H.H.).

Author information

These authors contributed equally: Qian Zhu, Jun-bao Zhang, Cai-ping Nie, Xiu-feng Liu, Liang-ping Zhan.

Authors and Affiliations

State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Guangdong Provincial Clinical Research Center for Cancer, Guangzhou, China
Qian Zhu, Jun-bao Zhang, Cai-ping Nie, Xiu-feng Liu, Liang-ping Zhan, Ming Li, Xi-liang Zeng, Jia He, He Huang, Xiao-jun Xia, Song Gao, Xiao-shi Zhang & Jiang Li
Department of Biotherapy, Sun Yat-sen University Cancer Center, Guangzhou, China
Qian Zhu, Jun-bao Zhang, Cai-ping Nie, Xiu-feng Liu, Liang-ping Zhan, Ming Li, Jia He, Xiao-shi Zhang & Jiang Li
Department of Gynecological Oncology, Sun Yat-sen University Cancer Center, Guangzhou, China
He Huang

Authors

Qian Zhu
View author publications
Search author on:PubMed Google Scholar
Jun-bao Zhang
View author publications
Search author on:PubMed Google Scholar
Cai-ping Nie
View author publications
Search author on:PubMed Google Scholar
Xiu-feng Liu
View author publications
Search author on:PubMed Google Scholar
Liang-ping Zhan
View author publications
Search author on:PubMed Google Scholar
Ming Li
View author publications
Search author on:PubMed Google Scholar
Xi-liang Zeng
View author publications
Search author on:PubMed Google Scholar
Jia He
View author publications
Search author on:PubMed Google Scholar
He Huang
View author publications
Search author on:PubMed Google Scholar
Xiao-jun Xia
View author publications
Search author on:PubMed Google Scholar
Song Gao
View author publications
Search author on:PubMed Google Scholar
Xiao-shi Zhang
View author publications
Search author on:PubMed Google Scholar
Jiang Li
View author publications
Search author on:PubMed Google Scholar

Contributions

J.L., Q.Z., J.B.Z., C.P.N., and X.F.L. were responsible for the study conceptualization. Q.Z., J.B.Z., C.P.N., L.P.Z., M.L., X.L.Z., and J.H. conducted experiments. X.F.L. contributed to the bioinformatics analysis. H.H., X.J.X, S.G., and X.S.Z. provided gene-edited mice and reagents. J.L., Q.Z., and X.F.L. were involved in the writing and revision of the manuscript. All authors have read and approved the manuscript. The order of the co-first authors was determined on the basis of their efforts and contributions to the manuscript.

Corresponding author

Correspondence to Jiang Li.

Ethics declarations

Competing interests

The authors declare that they have no competing interests.

Peer review

Peer review information

Nature Communications thanks anonymous reviewers for their contribution to the peer review of this work. [A peer review file is available].

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information (download PDF )

Description of Additional Supplementary Files (download PDF )

Supplementary Dataset 1 (download XLSX )

Reporting Summary (download PDF )

Transparent Peer Review file (download PDF )

Source data

Source Data (download XLSX )

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Zhu, Q., Zhang, Jb., Nie, Cp. et al. STING synergizes with TOX suppressing HO-1 expression to trigger ferroptosis in tumor-infiltrating CD8+ T cell and immunotherapy resistance. Nat Commun 17, 2543 (2026). https://doi.org/10.1038/s41467-026-69350-y

Download citation

Received: 27 May 2025
Accepted: 27 January 2026
Published: 09 February 2026
Version of record: 18 March 2026
DOI: https://doi.org/10.1038/s41467-026-69350-y