A “one-two punch” strategy to reverse immunosuppressive metabolism and activate T-cell immunity for enhanced cancer checkpoint immunotherapy

Abstract

Immune checkpoint blockade therapy, represented by CTLA-4 inhibitors like ipilimumab, faces limited clinical efficacy due to tumor immune evasion and the immunosuppressive tumor microenvironment. Particularly, metabolic reprogramming driven by hypoxia and the Warburg effect establishes an immunosuppressive microenvironment that inhibits T-cell-mediated antitumor immunity. To address this, we propose a “one-two punch” strategy designed to simultaneously enhance tumor immunogenicity and relieve T-cell suppression. This is achieved by a nanoplatform, ipilimumab&2-methoxyestradiol @PLGA (IM@PLGA), which co-delivers the CTLA-4 antibody ipilimumab and 2-methoxyestradiol (2-ME). 2-ME acts to reverse immunosuppression by inhibiting the HIF-1α/HK2 axis and induces immunogenic cell death (ICD) and autophagic cell death (ACD) for antigen exposure and dendritic cell (DC) activation. Concurrently, ipilimumab depletes regulatory T cells (Tregs), enabling robust activation of primed CD8⁺ T cells. In vitro and in vivo studies demonstrate that IM@PLGA effectively downregulates HIF-1α and HK2, induces ICD and ACD, promotes DC maturation, reduces intratumoral Tregs infiltration, and enhances CD8⁺ T cell recruitment. Furthermore, the treatment exhibits a potent abscopal effect in a metastatic tumor model. This work establishes a synergistic combination strategy that disrupts tumor metabolic defenses while boosting antitumor immunity, offering a promising approach to overcome resistance to cancer immune checkpoint blockade.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04436-9.

Introduction

Cancer immunotherapy, particularly immune checkpoint blockade (ICB) therapy, has profoundly revolutionized oncology [1, 2]. Among various immune checkpoints, cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) was the first validated target, and its inhibitor, ipilimumab, has achieved remarkable success in clinical practice, providing unprecedented survival benefits for cancer patients [3, 4]. Ipilimumab releases the “braking” of the immune system, depleting immunosuppressive regulatory T cells (Tregs) within the tumor microenvironment (TME) and effectively promoting the activation and infiltration of CD8⁺ T cells [5, 6]. However, despite its well-defined mechanism of action, the clinical response rate to ipilimumab remains low, with most patients facing the formidable challenge of primary or secondary resistance [7]. This indicates that the singular strategy of “releasing the immune brake” is still insufficient when confronting the complex tumor immune microenvironment. The root cause lies in the multi-layered defense system established by tumor cells. On the one hand, tumors are adept at “disguise and concealment” [8–12]. They downregulate the expression of major histocompatibility complex (MHC) molecules, preventing the effective presentation of tumor-associated antigens (TAAs) to T cells and thereby erasing their “identity tags” [13]. Additionally, they often assume a non-immunogenic character, failing to trigger effective immunogenic cell death (ICD) even during cell death [14]. On the other hand, and more critically, tumors undergo metabolic reprogramming by inducing hypoxia and the Warburg effect [15–17]. This shapes an acidic, nutrient-poor immunosuppressive microenvironment that directly suppresses the infiltration, function, and survival of cytotoxic T cells. The upregulation of key factors like hypoxia-inducible factor-1α (HIF-1α) further reinforces this immunosuppressive state, leading to the rapid exhaustion of T cells even after they have been activated by ipilimumab [16]. Therefore, how to effectively break through the “disguise” of tumors and the immunosuppressive microenvironment has become the core issue for improving the efficacy of ipilimumab.

To overcome the limitations of CTLA-4 inhibitors, various combination strategies have been explored. These range from coupling with chemotherapy to boost antigen exposure via ICD [18, 19], to dual checkpoint blockade (e.g., with nivolumab) [20, 21], and to combine with tumor metabolism metabolic inhibitors to disrupt the tumor metabolic microenvironment [22]. However, all these strategies are subject to certain limitations. The tumor hypoxic microenvironment stabilizes HIF-1α, induces chemotherapy resistance, and weakens the targeting of metabolic inhibitors [23, 24]. While combination with PD-1/PD-L1 blockade has become a clinical standard, it primarily intervenes at the downstream level of immunosuppressive pathways without addressing the upstream metabolic drivers. Therefore, a strategy capable of disrupting the tumor’s metabolic defenses while simultaneously enhancing tumor immunogenicity is critically needed. 2-Methoxyestradiol (2-ME), a natural estrogen metabolite, has attracted significant interest for its ability to alleviate tumor hypoxia and directly downregulates HIF-1α, a key transcriptional factor of the Warburg effect [25–27]. Importantly, our work further shows 2-ME can induce apoptosis and autophagic cell death (ACD), promoting the damage-associated molecular patterns (DAMPs) release, substantial antigen exposure, and antigen-presenting cell activation. We thus hypothesize that 2-ME can not only reverse immunosuppressive metabolism by inhibiting the HIF-1α/hexokinase 2 (HK2) axis but also induce potent tumor ICD, allowing ipilimumab to maximally relieve Treg-mediated immunosuppression, thereby fully activating immune response and achieving strong synergistic antitumor efficacy.

Herein, we propose a “one-two punch” strategy designed to concurrently remove the tumor’s “invisibility” and the “shackles” of T cells. To realize this, we engineered a nanoplatform, ipilimumab&2-ME@PLGA (IM@PLGA), which co-deliver the water-soluble ipilimumab and the lipid-soluble 2-ME by PLGA. As shown in Fig. 1, it employs the highly biocompatible polymer PLGA to co-deliver the water-soluble ipilimumab and the lipid-soluble 2-ME. In this strategy, 2-ME acts first to “expose” the tumor by inhibiting the HIF-1α/HK2 axis to reverse immunosuppressive metabolism and by inducing ICD, leading to antigen exposure and dendritic cell (DC) activation, thereby priming tumor-specific T cells. Concurrently, ipilimumab potently depletes Tregs, allowing the primed CD8⁺ T cells to become fully activated, resulting in a powerful synergistic antitumor immune response. To better evaluate this strategy, we selected sarcoma as our tumor model. Given that it typically shows a low response rate to ICB therapy but is highly sensitive to metabolism-induced ICD and ACD, this classic “cold” tumor serves as an ideal and challenging candidate to validate our synergistic immune sensitization approach. Both in vitro and in vivo experiments demonstrated that IM@PLGA effectively inhibits HIF-1α and HK2 and induces ICD. Furthermore, the nanoplatform significantly promoted DC maturation, reduced Treg infiltration at the tumor site, and enhanced CD8⁺ T cell recruitment. In the metastatic tumor model, the IM@PLGA treatment group exhibited a potent abscopal effect compared to groups treated with either agent alone. Notably, this study provides a potent combinatorial therapeutic strategy, offering a new approach to overcome resistance to cancer immune checkpoint inhibitors.

Fig. 1.

Schematic illustration of “one-two punch” strategy to reverse immunosuppressive metabolism and activate T-cell immunity for enhanced checkpoint immunotherapy

Results and discussion

Preparation and characterization of IM@PLGA

The schematic illustration of IM@PLGA preparation is presented in Fig. 1. Similarly, ipilimumab@PLGA (I@PLGA) and 2-ME@PLGA (M@PLGA) were also synthesized using the same method. The TEM images of IM@PLGA are shown in Fig. 2a, revealing that IM@PLGA exhibits a spherical structure with a diameter of approximately 100 nm. The hydrated particle size distributions of PLGA, M@PLGA, and IM@PLGA are shown in Fig. 2b, and the hydrodynamic diameter of the IM@PLGA was measured to be 132.5 ± 1.8 nm. Meanwhile, as shown in Fig. 2c, the Zeta potential of M@PLGA was − 18.8 ± 1.5 mV due to the negative charge of PLGA. Due to the positive charge of ipilimumab, the Zeta potential of I@PLGA and IM@PLGA increased to − 10.5 ± 0.8 mV and − 12.3 ± 1.1 mV, respectively, indicating the successful synthesis of IM@PLGA. Figure 2d shows the UV-Vis-NIR absorbance peaks of PLGA, 2-ME, ipilimumab, M@PLGA, I@PLGA, and IM@PLGA. Since the characteristic UV-Vis peak of both 2-ME and ipilimumab is at 280 nm, we characterized 2-ME and ipilimumab by High-Performance Liquid Chromatography (HPLC) and the BCA protein assay, respectively. Figure 2e and f confirm the successful loading of ipilimumab and 2-ME, respectively. Figure 2g and h show the hydrated particle size distributions of IM@PLGA in 10% FBS and PBS, respectively, over 7 days. The size distribution of IM@PLGA in PBS and FBS were approximately 150 nm, indicating that IM@PLGA had high stability in various media.

Fig. 2.

Characterizations of IM@PLGA. a) TEM image of IM@PLGA. b) Dynamic light scattering (DLS) analysis to characterize the size distribution of PLGA, M@PLGA, and IM@PLGA. c) DLS analysis to characterize the surface Zeta potential of M@PLGA, I@PLGA, and IM@PLGA. d) UV-Vis-NIR spectra of PLGA, 2-ME, ipilimumab, M@PLGA, I@PLGA, and IM@PLGA. e) BCA protein assay and f) HPLC assay for IM@PLGA. Size distributions of IM@PLGA in a 7-day stability test under g) 37 ℃ and h) 4 ℃, respectively. i) In vitro release profile of ipilimumab from free ipilimumab and IM@PLGA

The controlled release of ipilimumab and 2-ME from IM@PLGA

Figure S1 shows the standard curves of ipilimumab and 2-ME, which were determined by the BCA assay and UV spectrophotometry, respectively. After the successful synthesis of IM@PLGA, free ipilimumab and 2-ME was separated from the nanoplatform by dialysis, the drug encapsulation efficiency of ipilimumab and 2-ME was calculated to be 77.5% and 89.3%, respectively. The in vitro drug release curves of ipilimumab and 2-ME are shown in Fig. 2i and Figure S2, respectively. For ipilimumab, free ipilimumab released about 36% at 1 h and exceeded 85% after 6 h, which exhibited the fastest drug release rate, while IM@PLGA only released about 43% at 48 h. For 2-ME, free 2-ME released about 38% at 1 h and about 80% at 6 h, while IM@PLGA only released about 22% at 48 h. All these results demonstrate that PLGA enables the effective and controlled release of ipilimumab and 2-ME, thereby prolonging their in vivo half-life.

In vitro antitumor effects of IM@PLGA

The cell viability under normoxic and hypoxic environments after treatment with 2-ME, I@PLGA, and IM@PLGA was determined using the CCK-8 assay. As shown in Fig. 3a, within a specific concentration range, the cytotoxicity of 2-ME towards SW872 cells increased with higher concentrations. Figure 3b demonstrates that the cell viability of SW872 cells remained almost 100% as the concentration of I@PLGA increased. Meanwhile, Fig. 3c demonstrates that the cytotoxicity of IM@PLGA toward SW872 cells increased with higher concentrations. Furthermore, similar cytotoxicity trends were observed in the K7M2-WT cell line (Fig. 3d and e, and 3f). Figures 3g and S3 show the confocal laser scanning microscope (CLSM) images and quantitative data of IM@PLGA labeled with IR676 after internalization by cells at 1, 3, 6, and 12 h of incubation. The cell nuclei stained with DAPI emitted blue fluorescence, while the cytoplasm stained with IR676 emitted red fluorescence. With time, IR676 gradually accumulated in the cytoplasm, indicating that IM@PLGA could be effectively delivered into cells and accumulate in the cytoplasm.

Fig. 3.

In vitro cell culture studies. Cell viability of SW872 cells after incubation with a) 2-ME, b) I@PLGA, and c) IM@PLGA, respectively. Cell viability of K7M2-WT cells after incubation with d) 2-ME, e) I@PLGA, and f) IM@PLGA, respectively. g) Fluorescence images of cell internalization of IM@PLGA at 1, 3, 6, and 12 h, scale bars = 50 μm. h) Fluorescence microscopy images of Calcein AM/PI staining living/dead cells treated with PBS, 2-ME, I@PLGA, and IM@PLGA, scale bar = 100 μm. i) Apoptosis assays of SW872 cells treated with PBS, 2-ME, I@PLGA, and IM@PLGA. j) Quantification of apoptosis assays. Statistical significance was analyzed by one-way ANOVA, ***P < 0.001

Calcein AM/Propidium Iodide (PI) dual fluorescence staining and flow cytometry were used to verify the toxicity of the nanoplatform to tumor cells, as shown in Fig. 3h. Cells in the PBS and I@PLGA groups emitted predominantly green fluorescence, indicating that no cytotoxicity was detected. Cells in the 2-ME and IM@PLGA groups exhibited certain cytotoxicity, in which red fluorescence began to dominate, which is consistent with the cytotoxicity results. As shown in Fig. 3i and j, we further verified the cytotoxicity using flow cytometry. Slight apoptotic and necrotic cell death was demonstrated in the IM@PLGA group, whereas significantly greater cell death was observed in the 2-ME group compared to the PBS group (***P < 0.001). These results also confirmed the results of the CCK-8 assay and Calcein AM/PI dual fluorescence staining, indicating that IM@PLGA could induce tumor cell apoptosis in vitro.

In vitro ICD-induced by IM@PLGA

To clarify how IM@PLGA inhibits tumors, we conducted an in-depth investigation into its mechanism of action. As a classic HIF-1α inhibitor, 2-ME has been confirmed to effectively downregulate HIF-1α expression in tumor cells. Moreover, studies have shown that 2-ME can also induce ACD in tumors [28]. Therefore, we speculate that IM@PLGA can inhibit HIF-1α while inducing tumor cell apoptosis and ACD, thereby removing the tumor’s “invisibility” and enhancing ICD. Therefore, we first established an SW872-HRE-LUC cell line for the dynamic monitoring of tumor hypoxia. Figure 4a illustrates the schematic design of the hypoxia-reporting gene vector and the construction of the SW872-HRE-LUC cell line. As shown in Fig. 4b and c, I@PLGA had no effect on tumor HIF-1α expression. In contrast, both the M@PLGA and IM@PLGA treatment groups showed significant differences compared to the PBS group (**P < 0.01 and ***P < 0.001, respectively), indicating that IM@PLGA could significantly downregulate HIF-1α expression levels in tumor cells.

Fig. 4.

In vitro ICD and ACD induced by IM@PLGA. a) Schematic diagram of the construction of SW872-HRE-LUC cell line. b) Bioluminescence image (BLI) of SW872-HRE-LUC cells under varying treatment conditions. c) Quantitative BLI of SW872-HRE-LUC cells under varying treatment conditions. d) Western blot analysis of HIF-1α and HK2 expression in SW872 cells after treatment with PBS, M@PLGA, and IM@PLGA. e) Western blot analysis of SQSTM1/p62, LC3 I, and LC3 II expression in SW872 cells after treatment with PBS, M@PLGA, and IM@PLGA. f) Western blot analysis of Bcl-2 and Bax expression in SW872 cells after treatment with PBS, M@PLGA, and IM@PLGA. Quantification of intracellular g) glucose concentration, h) lactate concentration, and i) ATP level after treatment with virous groups. j) Schematic diagram of tumor ICD induction by IM@PLGA. k) Detection of surface MHC-I by flow cytometry in tumor cell. l) IFN-β, m) IL-6, and n) ATP levels in the supernatant of SW872 cells after treatment with PBS, M@PLGA, and IM@PLGA. Representative fluorescence images in SW872 cells after treatment with varying groups and then stained with o) anti-CRT antibodies and Multi-rAb™ CoraLite^® Plus 594-conjugated secondary antibodies and p) HMGB1 antibodies and CoraLite488-conjugated secondary antibodies. q) Mean fluorescence intensity of o). r) Cytosolic mean fluorescence intensity per cell of p). Groups are G1, PBS; G2, I@PLGA; G3, M@PLGA; G4, IM@PLGA. Statistical significance was analyzed by one-way ANOVA, **P < 0.01, ***P < 0.001

To further validate the inhibition of HIF-1α, western blot analysis was performed, as shown in Fig. 4d. The results revealed that M@PLGA and IM@PLGA downregulated HIF-1α as well as inhibited HK2 expression, resulting in the suppression of glycolysis in tumor cells. We hypothesize that this effect may be attributed to the downregulation of HIF-1α, which reduces the HIF-1 transcriptional complex activity, subsequently suppressing HK2 gene transcription and lowering its expression. Furthermore, autophagy and apoptosis-related proteins were examined by western blot. The results presented in Fig. 4e and f showed downregulated SQSTM1/p62 expression and an increased LC3 II/LC3 I ratio in the M@PLGA and IM@PLGA groups, indicating autophagosome formation and the induction of ACD in tumor cells. Concurrently, the expression level of Bcl-2 was significantly downregulated, accompanied by an upregulation of Bax expression in the M@PLGA and IM@PLGA groups, indicating that the nanoplatform could effectively promote tumor cell apoptosis. Given that the nanoplatform may inhibit the glycolytic level of tumor cells, we further measured the intracellular levels of glucose, lactate, and ATP. As shown in Fig. 4g and h, and 4i, compared to the PBS group, both the M@PLGA and IM@PLGA treatment groups significantly reduced intracellular lactate levels (***P < 0.001 and ***P < 0.001, respectively). Similarly, both groups exhibited a significant decrease in intracellular ATP levels (**P < 0.01 and **P < 0.01, respectively). These results indicate that IM@PLGA effectively inhibits HIF-1α and suppresses glycolysis, while simultaneously inducing ACD and apoptosis in tumor cells.

Subsequently, the induction of ICD by the IM@PLGA was assessed. Figure 4j presents a schematic diagram of tumor ICD induction by IM@PLGA. As shown in Fig. 4k and l, and 4m, compared to the PBS group, the expression of MHC-I on the tumor cell surface was significantly upregulated following treatment with M@PLGA and IM@PLGA (**P < 0.01 and **P < 0.01, respectively), accompanied by significantly elevated secretion levels of IFN-β (***P < 0.001 and ***P < 0.001, respectively) and IL-6 (***P < 0.001 and ***P < 0.001, respectively), indicating that the M@PLGA and IM@PLGA would contribute to immune activation. Meanwhile, ICD in tumor cells induces the release of ATP into the extracellular space, emitting a “find-me” signal to recruit anti-tumor immune effector cells [29]. As shown in Fig. 4n, ATP release was significantly elevated in the M@PLGA and IM@PLGA groups compared to the PBS group (***P < 0.001 and ***P < 0.001, respectively). In addition, Figs. 4o-r showed the calreticulin (CRT) expression on the cell surface and high mobility group box-1 (HMGB1) release. Under normal conditions, CRT is localized in the endoplasmic reticulum; however, under stress, CRT translocates to the cell membrane surface, emitting an “eat-me” signal to immune cells to activate immune responses [30]. As shown in Fig. 4o and q, evident CRT exposure was detected in the M@PLGA and IM@PLGA groups, confirming that M@PLGA and IM@PLGA significantly induce ICD. During ICD, HMGB1 typically translocates from the nucleus to the cytoplasm before being released extracellularly, activating downstream signaling pathways and promoting immune activation [31, 32]. As demonstrated in Fig. 4p and r, cells treated with M@PLGA and IM@PLGA exhibited higher cytosolic HMGB1 fluorescence intensity per cell compared to PBS group. This pronounced cytosolic translocation strongly reflects the active release process of HMGB1, further validating the potentiation of ICD by M@PLGA and IM@PLGA. These results indicate that IM@PLGA can significantly induce ICD in tumor cells.

In vitro DC maturation by IM@PLGA

The released or exposed DAMPs from dying tumor cells serve as adjuvant stimuli to DCs. Thus, we further investigated the effect of ICD induction on DCs. Figure 5a shows the schematic diagram of the experimental procedure. SW872 cells (1 × 10⁵) were seeded in the apical side of Transwells overnight and bone marrow-derived dendritic cells (BMDCs) (1 × 10⁵) were added to simulate resident DCs in the tumor. 100 µL PBS, I@PLGA, M@PLGA, or IM@PLGA were added. After co-culturing for 6 h, 9 × 10⁵ BMDCs in 800 µL RPMI-1640 medium was added to basolateral side of Transwells. After 24 h co-culturing, cells were washed with PBS and then analyzed for MHC-I, MHC-II expression, and DC maturation by flow cytometry. In addition, the supernatant of the media was collected for IFN-β and IL-6 analysis by ELISA kit. As shown in Fig. 5b and c, the elevated immunogenicity promoted the DC activation, as manifested by increased MHC molecules (MHC-I and MHC-II). Compared to the PBS group, more production of IFN-β and IL-6 was found in the supernatant of the media after coincubation with M@PLGA or IM@PLGA (Fig. 5d and e). Meanwhile, as shown in Fig. 5f and g, the elevated immunogenicity promoted DC maturation, as manifested by increased costimulatory molecules (CD80 and CD86). These results elucidated the dual role of IM@PLGA in direct cytotoxicity against tumor cell and ICD induction in DC maturation.

Fig. 5.

Activation of immune cells by IM@PLGA pretreated SW872 cells. a) Schematic diagram of immune cell activation by IM@PLGA pretreated SW872 cells. Detection of surface b) MHC-I and c) MHC-II by flow cytometry in BMDCs. d) IFN-β and e) IL-6 levels supernatant of the media after coincubation with PBS, I@PLGA, M@PLGA, or IM@PLGA. Representative f) flow cytometry plots and g) percentage of matured (CD80⁺CD86⁺) BMDCs after incubation with PBS, I@PLGA, M@PLGA, or IM@PLGA. Groups are G1, PBS; G2, I@PLGA; G3, M@PLGA; G4, IM@PLGA. Statistical significance was analyzed by one-way ANOVA, ***P < 0.001

Hemolysis and acute toxicity testing of IM@PLGA

Figures S4-S8 present the results of hemolysis and acute toxicity testing for IM@PLGA, demonstrating that the nanoplatform exhibits a high level of animal safety. As depicted in Figure S4, the Triton X-100 positive control group exhibited severe hemolysis. In contrast, the I@PLGA, M@PLGA, and IM@PLGA groups displayed negligible hemolysis, with rates of 0.88%, 0.84%, and 0.53%, respectively. Throughout the 14-day acute toxicity test, mice in all treatment groups exhibited normal behavior, maintained a 100% survival rate, and showed steady body weights within the normal range (Figures S5 and S6). To further assess potential systemic toxicity, we evaluated histopathological and whole-blood biochemical indices (Figures S7 and S8). Compared to the saline control group, the nanoplatform-treated mice showed no noticeable histopathological damage in major organs, nor any significant changes in liver function, renal function, or cardiac enzyme markers. These findings demonstrate the excellent in vivo biosafety and biocompatibility of the nanoplatforms.

In vivo anti-tumor efficacy of IM@PLGA

Before the anti-tumor efficacy experiments, we evaluated the in vivo biodistribution of the nanoplatform. Figure S9 displays the fluorescence images. Over time, the fluorescence intensities of tumor in the range progressively increased until they reached their highest point at 12 h, after which they gradually decreased, demonstrating the effective passive targeting capability of IM@PLGA to the tumor site.

To validate the anti-tumor efficacy of IM@PLGA, we evaluated their therapeutic efficacy in vivo using K7M2-WT-bearing mice, as illustrated schematically in Fig. 6a. Treatment was administered on a three-day cycle, with the entire treatment cycle lasting 12 days. The relative tumor volumes (V/V₀), final tumor volumes, and body weights presented in Figs. 6b-e, tumor growth was slightly suppressed in the I@PLGA and M@PLGA groups, with the most pronounced inhibition observed in the IM@PLGA treatment group. Notably, one mouse in the IM@PLGA group achieved complete tumor suppression after 12 days of treatment, indicating that 2-ME and ipilimumab can achieve effective synergistic antitumor effects. Additionally, as shown in Fig. 6d, the body weights of mice across all treatment groups remained stable, confirming the favorable biocompatibility of the nanoplatform.

Fig. 6.

In vivo anti-tumor efficacy of IM@PLGA. a) Schematic illustration of treatment schedules of the tumor-bearing mouse model. b) Relative tumor volumes (V/V₀) of the mice in each group during the 12 days. c) Final tumor volumes of the mice in each group following 12 days of treatment. d) Body weights of the mice in each group during the 12 days. e) Relative tumor volumes of the mice in each group during the 12 days. f) H&E staining, TUNEL staining, immunohistochemistry staining of the expression of HK2, and immunofluorescence staining of the expression of HIF-1α. Groups are G1, PBS; G2, I@PLGA; G3, M@PLGA; G4, IM@PLGA. Statistical significance was analyzed by one-way ANOVA, *P < 0.05

Mice in each group were euthanized after 12 days of treatment, and the tumor tissues were observed and analyzed via pathological sections. As shown in Fig. 6f, the results of the TUNEL fluorescence staining analysis show that there are almost no green fluorescence areas in the saline group, and no apoptosis was detected. In contrast, the I@PLGA, M@PLGA and IM@PLGA groups exhibited strong green fluorescence that covered a substantial portion of the tumor area, indicating effective tumor cell killing. Immunohistochemical and immunofluorescence analysis reveals the expression levels of HK2 and HIF-1α in tumor tissues. As observed, both the M@PLGA and IM@PLGA groups exhibited a significant reduction in HIF-1α and HK2 expression compared to the saline control.

In vivo anti-tumor immune response analysis

To demonstrate that IM@PLGA can not only induce ICD through 2-ME to remove the tumor’s “invisibility”, but also specifically block CTLA-4 on Tregs via ipilimumab, thereby alleviating the immunosuppressive tumor microenvironment and releasing the “shackles” on T cells. Thus, we further confirmed that IM@PLGA effectively induces anti-tumor immune responses. We first analyzed ICD induction using tissue immunofluorescence staining. As shown in Fig. 7a and b, the I@PLGA group exhibited slight CRT exposure and HMGB1 release, while the M@PLGA and IM@PLGA groups showed significant CRT exposure and HMGB1 release, consistent with in vitro results, confirming effective ICD induction in vivo.

Fig. 7.

In vivo anti-tumor immune response analysis. Immunofluorescence of a) CRT and b) HMGB1 in the tumor tissues after different treatments. Representative c) flow cytometry plots and d) percentage of matured (CD80⁺CD86⁺) DCs after different treatments. Representative e) flow cytometry plots and f) percentage of Tregs in CD4⁺ T cells after different treatments. Representative g) flow cytometry plots and h) percentage of CD8⁺ T cells in CD3⁺ T cells after different treatments. Groups are G1, PBS; G2, I@PLGA; G3, M@PLGA; G4, IM@PLGA. The ELISA assay of i) TNF-α and j) IFN-γ in tumor tissue. The ELISA test of k) TNF-α and l) IFN-γ in serum. Groups are G1, PBS; G2, I@PLGA; G3, M@PLGA; G4, IM@PLGA. Statistical significance was analyzed by one-way ANOVA, *P < 0.05, **P < 0.01, ***P < 0.001

The large number of TAAs generated by ICD effectively promotes DC maturation, and when combined with ipilimumab specifically blocking Tregs to alleviate immunosuppression, thereby increases the proportion of CD8⁺ T cells. Therefore, we used flow cytometry to assess the percentage of mature DCs in lymph nodes and the proportion of active CD8⁺ T cells. As shown in Fig. 7c and d, compared to the saline group (16.4%), the DC maturation rates in the and I@PLGA (24.9%, *P < 0.05) and IM@PLGA (30.1%, ***P < 0.001) treatment groups were significantly higher. Furthermore, DC maturation rates showed a significant difference between the I@PLGA group and the IM@PLGA group (**P < 0.01). Additionally, Treg infiltration in the treated tumor tissues was assessed by determining the proportion of Tregs among CD4⁺ T cells (Fig. 7e and f). Compared to the saline group (65.2%), Treg infiltration in the I@PLGA (38.0%, ***P < 0.001) and IM@PLGA (30.3%, ***P < 0.001) treatment groups were significantly lower, indicating that ipilimumab effectively inhibits Treg infiltration and alleviates immunosuppression in the tumor microenvironment. Moreover, we evaluated the activation efficiency of CD8⁺ T cells by analyzing the proportions of CD8⁺ T cells in CD3⁺ T cells (Fig. 7g and h). All treated groups had a higher proportion of CD8⁺ T cells than the saline group (21.8%), with the IM@PLGA group demonstrating the greatest enhancement (64.4%, ***P < 0.001). Moreover, enzyme-linked immunosorbent assay (ELISA) was used to measure tumor necrosis factor-α (TNF-α) and interferon-γ (IFN-γ) levels in serum (Fig. 7i and j) and tumor tissues (Fig. 7k and l). Compared to other groups, IM@PLGA significantly increased TNF-α and IFN-γ levels, triggering a strong anti-tumor immune response. Overall, the IM@PLGA nanoplatform delivers a sophisticated “one-two punch” through the combination of 2-ME and ipilimumab. It not only induces potent tumor ICD to strip away the tumor’s “invisibility” but also reduces Treg infiltration to relieve immunosuppression. Consequently, it significantly activates and enhances the infiltration of CD8⁺ T cells, thereby eliciting a robust antitumor immune response.

In vivo anti-tumor abscopal effect

Encouraged by the robust anti-tumor immune response elicited by IM@PLGA, we subsequently evaluated its therapeutic efficacy in a metastatic tumor model. First, an K7M2-WT metastatic tumor model was established, as illustrated in Fig. 8a. Treatment was administered in a three-day cycle, with the entire treatment cycle lasting 12 days. The body weights of mice across all groups remained within normal ranges during treatment (Figure S10). Figure 8b and c, and 8d present tumor volumes (V/V₀), final tumor volumes, and real-time tumor volumes for primary tumor, respectively. Primary tumor growth was slightly suppressed in the I@PLGA and M@PLGA groups, with the most pronounced inhibition observed in the IM@PLGA treatment group, and two mice in the IM@PLGA group achieved near-complete tumor suppression after 12 days of treatment. Furthermore, Figs. 8e-h shows the infiltration of Tregs and CD8⁺ T cells after 12 days of treatment. For Tregs, compared to the saline group (66.3%), Treg infiltration in the I@PLGA (33.9%, ***P < 0.001) and IM@PLGA (28.5%, ***P < 0.001) treatment groups were significantly lower. For CD8⁺ T cells, all treatment groups showed an increase in the proportion of CD8⁺ T cells relative to the saline control (20.3%), with the IM@PLGA group exhibiting the most substantial increase (62.2%, ***P < 0.001). Figure 8i and j show TNF-α and IFN-γ levels in primary tumor tissues. Compared to other groups, IM@PLGA significantly increased TNF-α and IFN-γ levels, triggering a strong anti-tumor immune response. Moreover, immunofluorescence staining of the primary tumor tissue revealed that the tumor site of IM@PLGA treatment group had the most abundant infiltration of CD8⁺ T cells (Fig. 8k), which corroborated the flow cytometry results. This result was attributed to both ICD-mediated immune activation and the enhancement of cytotoxic T lymphocyte infiltration and function via ipilimumab-induced Treg inhibition.

Fig. 8.

In vivo anti-tumor abscopal effect of IM@PLGA in bilateral tumor models. (a) Schematic illustration of treatment schedules of bilateral tumor models. (b) Relative primary tumor volumes (V/V₀) and (c) final primary tumor volumes of the mice in each group following 12 days of treatment. (d) Relative primary tumor volumes of the mice in each group during the 12 days. Representative flow cytometry plots of (e) Tregs in CD4⁺ T cells and (f) CD8⁺ T cells in CD3⁺ T cells in primary tumor after different treatments. Percentage of (g) Tregs in CD4⁺ T cells and (h) CD8⁺ T cells in CD3⁺ T cells in primary tumor after different treatments. (i) TNF-α and (j) IFN-γ levels in primary tumor tissue. k) Immunofluorescence of CD8⁺ T cells in primary tumor tissues after different treatments, scale bar = 50 μm. l) Relative distant tumor volumes (V/V₀) and m) final distant tumor volumes of the mice in each group following 12 days of treatment. n) Relative distant tumor volumes of the mice in each group during the 12 days. Representative flow cytometry plots of o) Tregs in CD4⁺ T cells and p) CD8⁺ T cells in CD3⁺ T cells in distant tumor after different treatments. Percentage of q) Tregs in CD4⁺ T cells and r) CD8⁺ T cells in CD3⁺ T cells in distant tumor after different treatments. s) TNF-α and t) IFN-γ levels in distant tumor tissue. u) Immunofluorescence of CD8⁺ T cells in distant tumor tissues after different treatments, scale bar = 50 μm. Representative v) flow cytometry plots and w) percentage of matured (CD80⁺CD86⁺) DCs after different treatments. x) TNF-α and y) IFN-γ levels in serum. Groups are G1, PBS; G2, I@PLGA; G3, M@PLGA; G4, IM@PLGA. Statistical significance was analyzed by one-way ANOVA, *P < 0.05, **P < 0.01, ***P < 0.001

For distant tumor, Fig. 8l and m, and 8n show the changes in the distant tumor during the treatment period. Notably, only the IM@PLGA treatment group showed significant inhibition of the distant tumor. Figures 8o-r show the infiltration of Tregs and CD8⁺ T cells in the distant tumor. In particular, the IM@PLGA treatment group exhibited significantly lower Treg infiltration than the other groups, while its CD8⁺ T cell infiltration was significantly higher. Figure 8s and t show TNF-α and IFN-γ levels in distant tumor tissues. Compared to other groups, IM@PLGA significantly increased TNF-α and IFN-γ levels. Moreover, immunofluorescence staining of the distant tumor tissue revealed that the tumor site of IM@PLGA treatment group had the most abundant infiltration of CD8⁺ T cells (Fig. 8u), demonstrating that IM@PLGA not only effectively kills the primary tumor but also activates the systemic immune system, generating a potent abscopal effect. Subsequently, to further evaluate the impact of IM@PLGA on overall immune activation, we assessed the maturation of DCs in the lymph nodes and measured the TNF-α and IFN-γ levels in serum. As shown in Fig. 8v and w, compared to the saline group (14.1%), the DC maturation rates in the and I@PLGA (23.7%, **P < 0.01) and IM@PLGA (36.2%, ***P < 0.001) treatment groups were significantly higher. Furthermore, IM@PLGA significantly increased TNF-α and IFN-γ levels in serum (Fig. 8x and y), effectively stimulating a systemic antitumor immune response, which corroborated the flow cytometry results.

Conclusion

In conclusion, we designed a PLGA-based nanoplatform (IM@PLGA) for the co‑delivery of 2‑ME and ipilimumab, implementing a “one‑two punch” immunotherapeutic strategy. In vitro experiments demonstrated that IM@PLGA effectively suppresses the HIF‑1α/HK2 axis in tumor cells, downregulates key glycolytic enzymes, and reduces lactate and ATP production, thereby reversing immunosuppressive metabolism. Simultaneously, the platform induces ICD and ACD, promotes CRT exposure, HMGB1 release, and ATP secretion, which significantly enhances DC maturation and antigen‑presenting function. In vivo studies further confirmed that IM@PLGA not only inhibits the growth of primary tumors but also effectively reduces the proportion of intratumoral Tregs while promoting the infiltration and activation of CD8⁺ T cells. Notably, in a metastatic tumor model, the treatment exhibited a potent abscopal effect, systemically suppressing distant tumor growth, accompanied by markedly enhanced DC maturation in lymph nodes and activation of systemic immune responses. This work not only verifies the synergistic mechanism of combining the metabolic modulator 2‑ME with the CTLA‑4 inhibitor ipilimumab but also provides a novel combinatorial therapeutic paradigm to overcome tumor metabolism‑mediated immunosuppression and resistance to immune checkpoint blockade.

Experimental section

Preparation of IM@PLGA

First, a 2% polyvinyl alcohol (PVA) solution was prepared by dissolving 2 g of PVA in 100 mL of deionized water. To prepare M@PLGA, 10 mg of PLGA and 1.2 mg of 2-ME were dissolved in 2 mL of CH₂Cl₂, and the solution was then emulsified via ultrasonication in an ice bath to form the M@PLGA emulsion. For the preparation of IM@PLGA, 1 mg of ipilimumab was first added to 10 mL of the 2% PVA solution. Subsequently, the M@PLGA emulsion was added to 10 mL of the 2% PVA solution containing 1 mg of ipilimumab. This mixture was then emulsified via ultrasonication in an ice bath to obtain the final IM@PLGA emulsion.

Characterization of IM@PLGA

The morphology of IM@PLGA was investigated using TEM. The hydrated particle size of PLGA, M@PLGA, and IM@PLGA of PLGA, M@PLGA, and IM@PLGA and the surface potential of M@PLGA, I@PLGA, and IM@PLGA were determined using a Malvern DLS particle size analyzer. The characteristic peaks of the UV-Vis spectra of PLGA, 2-ME, ipilimumab, M@PLGA, I@PLGA, and IM@PLGA were measured using a UV-Vis spectrophotometer. The successful loading of ipilimumab and 2-ME in IM@PLGA was determined using a BCA protein assay kit and HPLC, respectively. The in vitro stability of the IM@PLGA was assessed by measuring the hydrated particle size in PBS and 10% FBS solution.

Release of ipilimumab and 2-ME from IM@PLGA

The standard curve of ipilimumab was established using the BCA protein assay. Briefly, a 1 mg mL^-1 solution of ipilimumab was serially diluted to create standards with concentrations of 40, 80, 120, 160, and 200 µg mL^-1 for the assay. The standard curve of 2-ME was established by preparing a 1 mM solution of 2-ME, diluting it to 40, 80, 120, 160, and 200 µM with DMSO, and measuring the absorbance at 280 nm using a UV-Vis spectrophotometer. The in vitro release of ipilimumab and 2-ME were determined via dialysis. Specifically, 500 µL of free ipilimumab, free 2-ME, I@PLGA, and M@PLGA solution were added to the dialysis tube as the internal solution, and 10 mL of PBS (with 0.1% DMSO) was used as the external solution. The system was incubated at 37 °C. The concentration of ipilimumab and 2-ME in the external solution was quantified at 0, 1, 2, 4, 6, 12, 24, and 48 h to determine the cumulative release profile.

In vitro cell culture studies

Stable transfection

SW872-HRE-LUC, which can be induced to express luciferase, were generated by lentiviral infection with the recombinant plasmid HRE-hCMVmp-LUC. To produce high-titer lentiviral particles, this plasmid was co-transfected into 293T cells with the packaging plasmids pSPAX2 and pMD2.G. The transfected SW872 cells were then selected using 1.5 µg mL^-1 puromycin.

In vitro cell viability

SW872 and K7M2-WT cells were cultured in DMEM medium, respectively, supplemented with 10% FBS, and incubated at 37 ℃ in a 5% CO₂ humidified incubator. The cytotoxicity of 2-ME, I@PLGA, and IM@PLGA were studied under normoxic and hypoxic conditions (oxygen concentration: 0.6%). When the cells reached approximately 80% confluence in 96-well plates, the medium was replaced with fresh medium containing different concentrations of the sample, and the cells were incubated for 24 h. Then, 10 µL of CCK-8 reagent was added to each well, and the plates were incubated for 2 h. The absorbance of each well at 450 nm was measured using a microplate reader to calculate the cell viability.

Calcein AM/PI dual fluorescence staining was used to distinguish living and dead cells. When the cells reached approximately 80% confluence in the wells, the old medium was removed, and the cells were washed three times with PBS. Next, 1.0 mL of serum-free medium containing PBS, 2-ME, I@PLGA, and IM@PLGA was added, and the cells were incubated for 24 h. Living cells were stained with Calcein AM (green fluorescence, Ex/Em: 494/517 nm), and dead cells were stained with PI (red fluorescence, Ex/Em: 535/617 nm). After incubated with the dyes for 30 min, the cells were rinsed three times with PBS and visualized using a fluorescence microscope.

Cell uptake

When the cells reached approximately 70% confluence in the wells, the old medium was removed. The cells were washed three times with PBS, and 1.0 mL of serum-free medium containing ipilimumab&2-ME&IR676@PLGA (IMI@PLGA) was added. The cells were then incubated for 1, 3, 6, and 12 h. At each time point, the medium was removed, and the cells were washed three times with PBS. They were then fixed with 4% paraformaldehyde for 20 min and rewashed three times with PBS. After staining with 100 µL of DAPI solution (Ex/Em: 405/430–470 nm) for 15 min and three PBS washes, the cellular uptake of IMI@PLGA was observed using a CLSM. IR676 emitted red fluorescence (Ex/Em: 620–670 nm), while DAPI stained nuclei blue. The fluorescence intensity was quantified using ImageJ software.

Flow cytometry

When the cells reached approximately 80% confluence, the old medium was removed, and the cells were washed with PBS. Then, 100 µL of fresh medium containing PBS, 2-ME, I@PLGA, or IM@PLGA was added to each well, and the cells were incubated for 24 h. At the end of incubation, the cells were collected into centrifuge tubes, centrifuged at 1,000 × g for 5 min, washed three times with PBS, and then resuspended in PBS. Apoptosis was analyzed using the Annexin V-FITC/PI Apoptosis Detection Kit by flow cytometry.

In vitro bioluminescence imaging

The ability of IM@PLGA to alleviate hypoxia was evaluated using SW872-HRE-LUC cells. When the cells reached approximately 80% confluence, the old medium was removed, and the cells were washed with PBS. Then, 100 µL of fresh medium containing PBS, I@PLGA, M@PLGA, or IM@PLGA was added to each well. The cells were incubated under hypoxic (oxygen concentration: 0.6%) or normoxic conditions for 24 h. Following incubation, bioluminescence signals were captured using a small animal imaging system, and the intensity was quantified with Living Image (Version 4.3.0).

Western blot

SW872 cells were lysed with RIPA buffer, and proteins in supernatants (centrifuged at 14,000×g for 10 min) were quantified. Protein lysates were separated by SDS-PAGE, transferred to PVDF membranes, blocked, then incubated overnight at 4 °C with primary antibody. Membranes were washed, incubated with HRP-conjugated secondary antibody for 1 h, rinsed six times (5 min each), and visualized using ECL detection reagent.

In vitro metabolite detection

When the cells reached approximately 80% confluence, the old medium was removed, and the cells were washed with PBS. Then, 100 µL of fresh medium containing PBS, I@PLGA, M@PLGA, or IM@PLGA was added to each well, and the cells were incubated for 24 h. The culture medium was removed, and glucose assay kit, lactate assay kit, and ATP assay kit were used to quantify intracellular glucose, lactate, and ATP content, respectively.

In vitro induction of immunogenic cell death

To investigate the ICD, SW872 cells were treated with PBS, I@PLGA, M@PLGA, or IM@PLGA for 24 h. After treatment, the cells were washed with PBS and analyzed for the surface expression of MHC-I and CRT by flow cytometry, and for the release of HMGB1 by immunofluorescence. In addition, the culture supernatants were collected, and the levels of IFN-β and IL-6 were measured by ELISA, while ATP concentration was determined using ATP assay kit.

In vitro immune activation by IM@PLGA

SW872 cells (1 × 10⁵) were seeded in the apical side of transwells with 900 µL of DMEM overnight and 1 × 10⁵ BMDCs were then added to the same chamber to simulate resident DCs in the tumor. 100 µL of PBS, I@PLGA, M@PLGA, or IM@PLGA were added. After co-culturing for 6 h, 9 × 10⁵ BMDCs in 800 µL RPMI-1640 medium was introduced to basolateral side of transwells. Following a total co-culture period of 24 h, cells were washed with PBS and then analyzed for MHC-I, MHC-II expression, and DC maturation by flow cytometry. In addition, the culture supernatants were collected, and the concentrations of IFN-β and IL-6 were quantified using ELISA kits.

Hemolysis and acute toxicity tests

All animal experiments were performed in accordance with the animal ethics of the affiliated institutions (No. KYLL-2025-0828). 4% sheep blood erythrocytes were used for the hemolysis assay, 1% Triton X-100 solution was set as a positive control, and PBS was set as a negative control. 20 µL of each sample (1% Triton X-100, saline, I@PLGA, M@PLGA, and IM@PLGA), 480 µL of saline, and 500 µL of 4% sheep’s blood erythrocyte suspension were added to a 1.5 mL centrifuge tube, which were incubated in a 37 ℃ thermostat for 2 h. After the incubation, these tubes were centrifuged at 10,000 × g for 5 min, and then 200 µL of supernatant of each group was pipetted slowly and carefully into a 96-well plate, and the absorbance at 545 nm was measured using a microplate reader. The hemolysis rates of different samples were calculated according to the following equation:

OD_N indicates the absorbance at 545 nm of PBS, OD_P indicates the absorbance at 545 nm after treatment with 1% Triton X-100, and OD_S indicates the absorbance at 545 nm after treatment with different samples. Twelve BALB/c mice (15–17 g, approximately 4 weeks) were randomly divided into four groups, and each group was injected i.v. with 200 µL of saline, I@PLGA, M@PLGA, and IM@PLGA (dosages: 150 µg mL^-1 of ipilimumab, 150 µM of 2-ME), and the mice were observed to survive over 14 days and body weights were measured every two days. After euthanasia of the mice after 14 days, liver and kidney functions as well as cardiac enzymes were examined using blood biochemistry analysis, and in addition, the hearts, livers, spleens, lungs, and kidneys of the euthanized mice were collected for H&E staining.

In vivo fluorescence imaging of IM@PLGA

The mice were administered 50 µL of ipilimumab&2-ME&IR780@PLGA (IR780 concentration: 50 µg mL^-1 ) via the tail vein for fluorescence imaging. The IVIS imaging system was used to acquire fluorescence images at different time intervals (0, 1, 2, 4, 8, 12, 24, and 48 h) after injection. Then, the mice were euthanized, and the heart, liver, spleen, lung, kidney, and tumor tissues were removed for fluorescence imaging. The Ex/Em wavelengths used were 780 and 845 nm. The quantitative analysis of fluorescence intensity was conducted using Living Image.

In vivo antitumor efficacy of AMMDG

Subcutaneous K7M2-WT tumor models were established in immunocompetent BALB/c-hCTLA-4 mice to evaluate the tumor-suppressive efficacy of nanoplatforms. Mice were randomly divided into four groups (n = 5): (G1) saline, (G2) I@PLGA, (G3) M@PLGA, and (G4) IM@PLGA (dosages: 50 µg mL^-1 of ipilimumab, 50 µM of 2-ME). Treatments were administered on days 1, 4, 7, and 10. 100 µL of sample from each group was injected intravenously, and bioluminescence imaging data were acquired on days 0, 3, 6, 9, and 12. The entire treatment cycle lasted 12 days. The tumor volumes and body weights of the mice in each group were recorded during the treatment period.

In vivo antitumor immune response analysis

After the treatment, the mice were euthanized to collect the tumor and the inguinal lymph nodes near the tumors. The lymph node cell suspensions and tumor cell suspensions were then prepared by using the enzyme mixture solution (neutral protease, collagenase II, and hyaluronidase) in DMEM medium. The cells from the inguinal lymph nodes were stained with anti-CD11c-FITC, anti-CD86-APC, and anti-CD80-PE for the DCs maturation study. Tumor cells were analyzed by flow cytometry for Tregs study (anti-CD3-APC, anti-CD4-FITC, and anti-Foxp3-PE) and T-cell study (anti-CD3-APC, anti-CD4-FITC, and anti-CD8-PE). Furthermore, the TNF-α and IFN-β were measured by using an ELISA assay according to the vendor’s instructions.

In vivo antitumor abscopal effect

The bilateral tumor model was established to evaluate the abscopal effect of IM@PLGA. First, K7M2-WT cells (1 × 10⁷ cells per mouse) were subcutaneously injected into the right flank of each mouse. Four days later, K7M2-WT cells (1 × 10⁷ cells per mouse) were subcutaneously injected into the left flank of each mouse to establish the distant tumor. The tumor-bearing mice were divided into four groups (n = 5): (G1) saline, (G2) I@PLGA, (G3) M@PLGA, and (G4) IM@PLGA (dosages: 50 µg mL^-1 of ipilimumab, 50 µM of 2-ME). Specifically, the corresponding nanoplatforms or saline were administered directly into the primary (right) tumors via intratumoral injection. The distant (left) tumors were left untreated to monitor the abscopal effect. The tumor volumes and body weights of the mice in each group were recorded during the treatment period.

Staining of tissue sections

Paraformaldehyde (4%) was used to fix the tumor tissues, followed by paraffin-embedded sections to examine the morphological and histological features. Histopathological analysis was performed using H&E and TUNEL staining to assess the levels of tumor cell proliferation and cell death. Anti-HIF-1α, anti-HK2, anti-CRT, anti-HMGB1, anti-CD3, and anti-CD8 antibodies were used to observe the protein expression levels, ICD, and immune response analysis.

Statistical analysis

The data were represented at least three times as mean ± SD. Two or more groups were compared using one-way ANOVA. No significant difference was indicated as (n. s.), P < 0.05 (*), P < 0.01 (**), and P < 0.001 (***).

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Peng Chang: Writing – original draft, Visualization, Validation, Data curation, Conceptualization. Kuoye Tian: Data curation. Xiaoyu Yang: Data curation. Dehong Cheng: Data curation. Chenying Wang: Data curation. Ke Li: Funding acquisition. Dan Chen: Supervision, Project administration. Yangyang Feng: Funding acquisition. Yun Zeng: Writing – review & editing, Funding acquisition. Yonghua Zhan: Writing – review & editing, Project administration, Funding acquisition. Wenhua Zhan: Writing – review & editing, Project administration.

Funding

This work was supported, in part, by the National Natural Science Foundation of China (No. 82260473), the Natural Science Basic Research Key Program of Shaanxi Province (No. 2024JC-ZDXM-47), the Natural Science Basic Research Key Program of Ningxia Province of China (No. 2024AAC02079), the Natural Science Basic Research Program of Ningxia Province of China (No. 2025AAC030764), the State Key Laboratory of Pathogenesis, Prevention and Treatment of High Incidence Diseases in Central Asia Fund (No. SKL-HIDCA-2024-NX1), the Health Re-search Program on Ningxia (No. 2025-NWQP-B002), and the Xidian University Specially Funded Project for Interdisciplinary Exploration (Nos. TZJH2024029, TZJH2024030).

Data availability

All data supporting the findings of this study are available within the paper and its Supplementary Information.

Declarations

Competing interests

The authors declare no competing interests.