The human breast cancer cell line SKBR-3 (HTB-30, ATCC) was cultured in RPMI-1640 medium (Invitrogen, USA) supplemented with 10% fetal bovine serum (FBS) and 1% (v/v) penicillin-streptomycin (P/S; Sigma-Aldrich, USA) at 37 °C in a humidified atmosphere containing 5% CO₂. The Her2⁺SKBR-3 cell line expressing luciferase (Her2⁺SKBR3-luc) was generated by stably transducing wild-type Her2⁺SKBR-3 cells with firefly luciferase.

The animal experimental protocols in this study were approved by the Institutional Animal Care and Use Committee (IACUC) of Changsha University and were conducted in strict compliance with its ethical guidelines and regulatory requirements. This study was performed under the Animal Ethics Approval No. CCSU-2023-014. For the euthanasia of the mice, mice were placed in an induction chamber for 2~3 minutes using a 30% volume per minute displacement rate of 100% CO₂ to terminate the experiment. Then the samples were collected as needs.

DNA fragments encoding the single-chain variable fragment (scFv) anti-Her2 chimeric receptors (Her2 CAR or Her2-IF1 CAR) were synthesized by Genescript (Nanjing, China) and cloned into the lentiviral vector plenti-EF1A-puro to generate the Her2 CAR and Her2-IF1 CAR constructs. For the construction of Her2-shIF1 CAR, a U6 promoter was inserted downstream of the Her2 CAR sequence, as illustrated in Figure 1A. The ATPIF1-targeting shRNA sequences were as follows:

ATPIF1-shRNA-F: CCGGCCATGAAGAAGAAATCGTTCACTCGAGTGAACGATTTCTTCTTCATGGTTTTTG.

ATPIF1-shRNA-R: AATTCAAAAACCATGAAGAAGAAATCGTTCACTCGAGTGAACGATTTCTTCTTCATGG.

Lentivirus production for CAR constructs was performed as described in our previous publication.¹⁰.

CD3+ T cells were isolated from the human Peripheral Blood Mononuclear Cell (PBMC) in preparation of the CAR-T cells. The experiment design was approved by the Medical Ethics Committee in Xinxiang Medical University, and the informed consent forms were signed by the donors before the blood collection.

PBMC was mixed with normal saline at a 1:1 ratio to dilute the blood. The diluted blood was then slowly layered onto lymphocyte separation medium (TBD, Tianjin, China) at a 1:3 ratio and centrifuged at 2000 rpm for 20 minutes. After centrifugation, a distinct white buffy coat layer was observed, which was carefully aspirated and transferred to a new centrifuge tube. An appropriate volume of PBS was added, followed by centrifugation at 2000 rpm for 8–10 minutes at room temperature. The supernatant was discarded, and this washing step was repeated twice. The resulting pellet contained PBMCs, which were resuspended in RPMI-1640 medium supplemented with 10% serum and thoroughly mixed by pipetting before cell counting.

T cells were isolated from the PBMCs using Human CD3 Positive Selection Kit (Cat. 8802-6830-74, Invitrogen). Purified T cells were stimulated with HuCD3/CD28/CD2 T cell Act (Cat. 10970, Stemcell) for 24 h. The T cells were then incubated in a complete RPMI 1640 medium containing hIL-2 (100 U/mL) and 10% FBS (Gibco) for expansion. On the following day, the T cells were infected with the lentivirus (Her2 CAR, Her2-IF1 CAR and Her2-shIF1 CAR) containing the target gene at a ratio of MOI = 5:1 of virus vs. cells. The transfected T lymphocytes were amplified and cultured in an incubator of 37 °C with 5% CO₂ to obtain CAR-T cells. Five days after lentivirus infection, CAR+ cells were analyzed using the biotin-labeled Her2 protein (Cat: 10004-H02H, SinoBiological, China) following the PE-conjugated anti-biotin antibody (Cat: 12-9895-82, eBioscience).

For the cultivation of CAR-T under different O₂ concentration, the CAR-T cells were placed in the conventional two gas incubator (21% O₂), or placed in the three gas incubator (Thermofisher Heracell™ VIOS 160i), and the concentration of nitrogen is automatic adjusted to maintain the oxygen level at 1% (1% O₂).

Total RNAs were extracted from the CAR-T cells using TRIzol reagent (Takara Bio) and converted into cDNAs. Real-time PCR analysis was performed using SYBR Green Master Mix (Takara Bio). The ^ΔΔCt method was used to calculate the gene expression levels after normalization to the GUSB. The primer information is listed in Supplementary Table 1.

The generation of ρ⁰cells has been described previously (12). To deplete mitochondrial DNA (mtDNA), CAR-T cells were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS), human IL-2 (100 U/mL), and ethidium bromide (EB, 25 μg/mL) for over 1 week. Subsequently, both EB-untreated control CAR-T cells and EB-treated CAR-T cells were harvested. Mitochondrial D-loop DNA was extracted and subjected to real-time RT-PCR analysis.

Different types of CAR-T cells were lysed in RIPA buffer containing phosphatase inhibitors. Protein concentrations of samples were determined using a BCA assay and normalized. Samples were then denatured at 100 °C for 10 minutes.

Proteins were separated by electrophoresis on 10% or 15% SDS-polyacrylamide gels. After sample loading, electrophoresis was performed at 90 V for 20 minutes initially, then adjusted to 120 V until completion. Subsequently, proteins were transferred to a PVDF membrane at 260 mA for 90 minutes. The membrane was blocked with 5% skim milk for 1 hour at room temperature, followed by four washes with TBST (5 minutes per wash). Primary antibodies were incubated overnight at 4 °C, after which the membrane was washed four times with TBST (5 minutes each). Secondary antibodies were applied for 1 hour at room temperature, and the membrane was washed again four times with TBST (5 minutes per wash). Finally, protein bands were visualized using a chemiluminescence detection kit.

An appropriate number of CAR-T cells, which had been stimulated with SKBR3-Luc tumor cell-conditioned medium for 48 hours, were transferred to a 1.5 mL EP tube and centrifuged at 1300 rpm for 5 minutes. The cells were washed with PBS and then stained using an apoptosis detection kit (Cat# C1065L, Beyotime, China). After staining, the cells were incubated at 37 °C in the dark for 30 minutes. Following an additional PBS wash, each sample was resuspended in 300 µL of assay buffer and analyzed by flow cytometry within 1 hour. Data were processed using FlowJo software.

The amount of mtDNA was quantified as previously described (13). Briefly, CAR-T cells (1 × 10⁶) were resuspended in 170 μl of digitonin buffer containing 150 mM NaCl, 50 mM HEPES, pH 7.4 and 25 μg/ml digitonin (HY-N4000, MedChemExpress). The homogenates were incubated on a rotator for 10 min at room temperature, followed by centrifugation at 15, 000g for 25 min at 4 °C. A 1:20 dilution of the supernatant (cmtDNA) was used for qPCR. The pellet was resuspended in 340 μl of lysis buffer containing 5 mM EDTA and proteinase K (HY-108717, MedChemExpress) and incubated at 55 °C overnight. The digested pellet was diluted with water (1:20 to 1:100) and heated at 95 °C for 20 min to inactivate proteinase K and the sample was used for qPCR with mtDNA-specific primers listed in Supplementary Table 1. The cmtDNA in the supernatant was normalized to the β-actin in the pellet for each sample.

For the persistence assay, tumor-free NCG mice were used. Her2 CAR-T and Her2-IF1 CAR-T cells were administered via tail vein injection, with simultaneous intraperitoneal injection of IL-2 (100, 000 IU per mouse, twice weekly). Two weeks later, mice were euthanized, and peripheral blood mononuclear cells (PBMCs) and splenocytes were collected. CAR-T cells were detected using flow cytometry.

For evaluation of antitumor activity, three independent antitumor experiments were performed. SKBR3 cells were resuspended in sterile saline to a density of 5×10⁶cells/mL and kept on ice for subsequent use. The cell suspension was gently mixed, and a disposable syringe was used to aspirate the suspension. Each NCG mouse (Cat# T001475, GemPharmatech, China) received a subcutaneous injection of 200 µL cell suspension in the axillary region. Approximately one month later, when axillary tumor volumes reached ~100 mm³, mice were randomly assigned to groups.

Throughout the experiments, mouse body weights were monitored, and tumor growth was recorded. Approximately 4 weeks later, NCG mice were euthanized, and axillary tumors were excised. Tumor size and weight in each group were measured and recorded.

The supernatant of CAR-T cells co-cultured with SKBR3-Luc cells was collected. Concentrations of IL-2 (Cat# 88-7025-88, ThermoFisher) and IFN-γ (Cat# 88-7316-86, ThermoFisher) in the supernatant were determined using ELISA kits following the manufacturer’s instructions.

CAR-T cells were loaded at 1 x 10⁶/well into XF24 plate, which was pre-coated with Cell-Tak for 20 min to increase the adhesion of CAR-T cells. The determination of OCR (Oxygen Consumption Rate) of CAR-T cells was performed as described in our previous study (10).

The CAR-T cells cultured under 21% O₂ and 1% O₂ concentration were collected, and were stained with biotin-labeled Her2 protein (Cat: 10004-H02H, SinoBiological, China) following the PE conjugated anti-biotin antibody (Cat: 12-9895-82, eBioscience), and with the addition of TMRE in the MMP assay kit (Cat# C2001S, Beyotime Biotechnology Limited Company, Shanghai, China) or the mPTP assay kit (Cat# C2009S, Beyotime Biotechnology Limited Company, Shanghai, China), the TMRM and mPTP was determined with flow cytometry (BD, FACSCalibur) and the results in CAR⁺ T cells were analyzed with the Flowjo software.

The CAR-T cells cultured under 21% O₂ and 1% O₂ concentration were collected, and were stained with biotin-labeled Her2 protein (Cat: 10004-H02H, SinoBiological, China) following the Alexa Fluor™ anti-biotin antibody (Cat: 53-9895-82, eBioscience), after that the addition of Mito-Tracker Red CMXros (Beyotime, China, Cat#C1049B-250 μg), the mitochondrial mass was determined with flow cytometry (BD, FACSCalibur).

Organizational Chip of Her2⁺ breast cancer was purchased from the Shanghai Outdo Biotech Co., LTD (Outdo Biotech) was used for mIHC analysis, the Formalin‐Fixed Paraffin‐Embedded sections labelled with DAPI (blue), ATPIF1 (yellow), CD3 (green), HIF-1α (Orange) were scanned using the Vectra imaging system.

CAR-T cells were stained with the Cell Tracker Red CMTPX live-cell tracking Dye (Cat# MX4109, Maokangbio, Shanghai, China). Different types of CAR-T cells were resuspended in RPMI-1640 complete medium and seeded into Transwell chambers at 200 µL per well. SKBR3-Luc conditioned medium (600 µL per well) was added to the lower chambers. After 4 hours of incubation, migrated CAR-T cells in the lower chambers were visualized using a fluorescence microscope. Five random fields of view were selected per well, imaged, and recorded for quantitative analysis.

Data are presented as mean ± standard deviation (SD). All in vitro experiments were repeated at least three times, and data were analyzed using an unpaired two-tailed t-test. For in vivo experiments, data were analyzed by post hoc tests following one-way analysis of variance (ANOVA) using the SPSS 20.0 software package. P value < 0.05 was considered statistically significant.

Given that ATPIF1 overexpression has been shown to enhance the antitumor activity of αCD19 CAR-T cells both in vitro and in vivo, we engineered Her2-targeting CAR-T cells to overexpress ATPIF1 with the aim of improving their antitumor efficacy against solid tumors. As illustrated in Figure 2A, ATPIF1-overexpressing Her2-targeted CAR-T cells (Her2-IF1 CAR-T) were successfully generated, and this was validated by flow cytometric analysis of GFP expression (Figure 2B) and Western blotting (WB; Supplementary Figure 1). For in vitro experiments, Her2 CAR-T cells were co-cultured with luciferase-expressing breast cancer SKBR3 cells (SKBR3-Luc⁺). As depicted in Figure 2C, Her2-IF1 CAR-T cells exhibited more potent tumor lysis activity compared to parental Her2 CAR-T cells. Additionally, the levels of IFN-γ and IL-2 in the co-culture supernatants were significantly higher in the Her2-IF1 CAR-T group than in the Her2 CAR-T group (Figures 2D, E). These results indicate that ATPIF1 overexpression enhances the in vitro antitumor efficacy of Her2 CAR-T cells. To explore the underlying mechanism, a Seahorse extracellular flux analyzer was used to assess oxidative phosphorylation (OXPHOS) in Her2 and Her2-IF1 CAR-T cells. As shown in Figures 2F–H, the oxygen consumption rate (OCR) was markedly elevated in Her2-IF1 CAR-T cells, suggesting that ATPIF1 overexpression improves the mitochondrial respiratory chain function of CAR-T cells.

The in vivo persistence of CAR-T cells is critical for their antitumor efficacy. Therefore, we assessed the impact of ATPIF1 overexpression on CAR-T cell persistence using flow cytometry. Nontumor-bearing mice were intravenously injected with equal numbers of CAR⁺Her2 CAR-T and Her2-IF1 CAR-T cells, and the number of CAR-T cells in peripheral blood mononuclear cells (PBMCs) and spleens was quantified two weeks post-injection. As shown in Figures 3A, B.

ATPIF1 overexpression significantly enhanced the in vivo persistence of CAR-T cells, the percentage of GFP⁺Her2-IF1 CAR-T cells was approximately 10-fold higher in both PBMCs and spleens compared to parental Her2 CAR-T cells. Surprisingly, however, the antitumor efficacy in mice xenografted with Her2⁺SKBR3 tumor cells was reversed (Figure 3C). Specifically, the tumor weight in the Her2-IF1 CAR-T treatment group was significantly greater than that in the Her2 CAR-T group, indicating that ATPIF1 overexpression impairs the in vivo antitumor efficacy of Her2 CAR-T cells. To elucidate the underlying mechanism, the memory phenotypes and exhaustion markers of CD8⁺ CAR-T cells were determined in vitro under 21% O₂ and 1% O₂ conditions, respectively. As shown in Figures 3D, F, there were no significant difference of Her2-IF1 CAR-T cells under different oxygen conditions, the ratio of different memory phenotype in CD8⁺ CAR-T was similar. However, the exhaustion markers of PD-1, LAG-3 and TIM-3 in Her2-IF1 CAR-T cells were significantly increased as compared to that of Her2 CAR-T cells under 1% O₂ condition in vitro, indicating that ATPIF1 overexpression might increase the exhaustion of CD8⁺ CAR-T cells under hypoxic condition (Figures 3E, G).

Given that ATPIF1 overexpression impaired the in vivo tumor-inhibitory efficacy of Her2 CAR-T cells, we further investigated the impact of ATPIF1 knockdown on Her2 CAR-T cell function. As illustrated in Figures 1A–E, both ATPIF1-overexpressing Her2-IF1 CAR-T cells and ATPIF1-knockdown Her2-shIF1 CAR-T cells were successfully constructed. Notably, Her2-shIF1 CAR-T cells exhibited the slowest proliferation rate among the three groups (Figure 1F). Consistent with our previous findings, Seahorse analysis revealed that the OCR was markedly elevated in Her2-IF1 CAR-T cells, whereas ATPIF1 knockdown reduced both the basal respiration and maximum respiratory capacity of CAR-T cells (Figures 1G–I). Meanwhile, Her2-IF1 CAR-T cells maintained enhanced tumor lysis activity compared to parental Her2 CAR-T cells, whereas ATPIF1 knockdown impaired the in vitro cytolytic capacity of CAR-T cells (Figure 1J). Also, Her2-IF1 CAR-T cells secreted significantly higher levels of IL-2 and IFN-γ than both Her2 CAR-T and Her2-shIF1 CAR-T cells (Figure 1K, L). In contrast, Her2-shIF1 CAR-T cells showed the lowest levels of IL-2 and IFN-γ secretion, as well as the weakest tumor lysis activity, compared to the other two groups (Figures 1J–L). Additionally, Her2-IF1 CAR-T cells still had the enhanced cytolytic lysis function and increased IL-2, IFN-γ secretion even under hypoxic condition (1% O₂) (Figures 1M–O). These results indicate that ATPIF1 knockdown diminishes the in vitro antitumor function of Her2 CAR-T cells.

To evaluate the in vivo antitumor activity of Her2 CAR-T, Her2-IF1 CAR-T, and Her2-shIF1 CAR-T cells, we conducted antitumor experiments using tumor-xenografted mice. As depicted in Figures 4A–C, tumor growth was significantly retarded following CAR-T cell administration. Surprisingly, Her2-shIF1 CAR-T cells exhibited the most potent inhibitory efficacy compared to that of untransfected T cells (UNT), Her2 CAR-T cells, and Her2-IF1 CAR-T cells. Notably, there were no significant differences in body weight or spleen weight among the various groups (Figures 4D, E). Consistent with the tumor growth curves, the tumor weight in the Her2-shIF1 CAR-T treatment group was the lowest (Figure 4B), indicating that ATPIF1 knockdown enhances the in vivo antitumor activity of Her2 CAR-T cells—despite its impairment of antitumor function in vitro. In contrast, ATPIF1 overexpression was found to diminish in vivo antitumor efficacy, which is consistent with the results presented in Figure 3C.

To elucidate the contradictory phenotypes of Her2-IF1 CAR-T and Her2-shIF1 CAR-T cells observed in vitro versus in vivo, we analyzed several key characteristics of these CAR-T cells. First, mitochondrial mass was quantified by flow cytometry. As shown in Figures 5A, B, ATPIF1 overexpression increased mitochondrial mass, whereas ATPIF1 knockdown reduced it—indicating a positive correlation between ATPIF1 expression levels and mitochondrial mass. Second, mitochondrial membrane potential (MMP) was assessed. We found that ATPIF1 overexpression decreased MMP, while ATPIF1 knockdown elevated it. Furthermore, MMP was further reduced under hypoxic conditions (1% O₂; Figures 5C, D). This regulatory effect of ATPIF1 on MMP is consistent with previous reports (15–17). To further unravel the dual role of ATPIF1 in modulating CAR-T cell activity in vitro and in vivo, we determined the apoptosis rate of CAR-T cells stimulated with SKBR3-conditioned medium under different oxygen concentrations. As shown in Figures 6A, B, ATPIF1 knockdown increased the apoptotic rate under normoxic conditions (21% O₂) but enhanced cell survival under hypoxic conditions (1% O₂), ATPIF1 overexpression exerted the opposite effects. Regarding reactive oxygen species (ROS) production, ATPIF1 overexpression increased ROS levels in CAR-T cells compared to parental Her2 CAR-T cells under 21% O₂ condition. Notably, ROS levels were universally reduced in Her2 CAR-T and Her2-IF1 CAR-T cells under hypoxic conditions, whereas Her2-shIF1 CAR-T cells exhibited the highest ROS content among the three groups under 1% O₂ condition (Figures 6C, D).

The aforementioned results indicated that ATPIF1 modulates the survival capacity of CAR-T cells under distinct oxygen conditions. Thus, the enhanced in vivo antitumor activity of Her2-shIF1 CAR-T cells might be attributed to their improved survival rate within solid tumors, which typically exhibit a hypoxic microenvironment. To test this hypothesis, we administered CAR-T cells to tumor-xenografted mice and employed immunofluorescence staining to evaluate CAR-T cell infiltration and survival within tumors. As illustrated in Figures 7A, B, Her2-shIF1 CAR-T cells showed the highest accumulation in solid tumors, as evidenced by the greatest number of CD3⁺-stained cells per high-power field (HPF). Concurrently, in vitro transwell assays were performed (Supplementary Figure 2), revealing that Her2-shIF1 CAR-T cells exhibited the strongest migratory capacity under hypoxic conditions (1% O₂). In contrast, Her2-IF1 CAR-T cells displayed optimal mobility under normoxic conditions (21% O₂), indicating that ATPIF1 exerts opposing effects on CAR-T cell migration depending on the oxygen concentration. To further explore the correlation between ATPIF1 expression and T cell migration in clinical tumors, multi-color immunofluorescence staining was conducted on human breast cancer tissues (Figure 7C). We found that ATPIF1 expression was negatively correlated with CD3⁺T cell infiltration, while HIF-1α expression was positively correlated with ATPIF1 levels (Figures 7D, E). These results suggest that higher ATPIF1 expression in breast cancer tissues is associated with reduced CD3⁺T cell infiltration and survival.

Mitochondrial permeability transition pore (mPTP) opening is suspected to promote cellular apoptosis, and Calcein-AM is a commonly used dye for detecting mPTP opening—with reduced intracellular Calcein-AM fluorescence indicating increased mPTP opening. As shown in Figures 8A, B, Her2-shIF1 CAR-T cells exhibited a significant decrease in Calcein-AM fluorescence intensity, suggesting enhanced mPTP opening, and this were might related with the ROS, as the elimination of ROS with N-Acetylcysteine (NAC) treatment could increase the Calcein-AM ratio. Western blot analysis revealed that Her2-shIF1 CAR-T cells had upregulated protein expression of phosphorylated STING (p-STING), total STING, and VDAC (Figures 8C–E), indicating activation of the STING signaling pathway. Previous studies have reported that mPTP opening can induce mitochondrial DNA (mtDNA) leakage, which further activates the STING pathway and subsequently enhances in vivo T cell infiltration (18, 19). To explore this mechanism, we extracted cytoplasmic mtDNA and detected the levels of ND-1 and D-loop (mtDNA-specific markers) using qPCR; we also quantified the mRNA expression of IFI44 and IFN-β (STING pathway downstream targets) via qPCR. As depicted in Figures 8F–H, the levels of ND-1, D-loop, IFI44, and IFN-β were all significantly upregulated in Her2-shIF1 CAR-T cells, suggesting that the increased STING phosphorylation in these cells may be associated with mtDNA leakage. To further validate this causal link, we depleted mtDNA from CAR-T cells using ethidium bromide (EB) to generate ρ⁰cells (Figure 8I). Following mtDNA depletion, there were no significant differences in the protein expression of p-STING and VDAC, or in the mRNA expression of IFI44, among Her2 CAR-T, Her2-IF1 CAR-T, and Her2-shIF1 CAR-T cells (Figure 8J; Supplementary Figure 3).

H151, a specific inhibitor of STING was used to investigate its impact on CAR-T cell migration. Consistent with our prior findings (Supplementary Figure 2), Her2-shIF1 CAR-T cells exhibited the strongest migratory capacity under hypoxic conditions (1% O₂), and this enhanced migration was significantly abrogated following H151 treatment (Figures 9A, B). These results indicate that activation of the STING signaling pathway is responsible for the improved in vitro migration of Her2-shIF1 CAR-T cells. To further validate this observation, we conducted in vivo animal experiments. As depicted in Figures 9C, D, NCG mice were xenografted with SKBR3 breast cancer cells and then administered CAR-T cells either alone or in combination with H151. Consistent with the results shown in Figure 4, Her2-IF1 CAR-T cells displayed reduced tumor-inhibitory efficacy compared to parental Her2 CAR-T cells, while Her2-shIF1 CAR-T cells exerted the most potent antitumor activity among the three groups. Notably, H151 treatment attenuated the antitumor efficacy of CAR-T cells, underscoring the critical role of the STING signaling pathway in regulating CAR-T cell antitumor function.