Sonogenic malate depleting modulator for tumor metabolic reprogramming and antitumor immune activation.

1
Introduction
Malate plays a key role in tumor genesis and development. As a key metabolic intermediate, malate is involved in cellular energy metabolism and acid-base homeostasis [[1], [2], [3]]. Studies have shown that malate not only affects cell growth and apoptosis, but also regulates the cell cycle and plays a facilitating role in the energy metabolism of tumor cells [[4], [5], [6]]. In the tumor microenvironment, changes in malate concentration may affect the metabolic status and proliferative capacity of tumor cells, further promoting their survival and metastasis [[7], [8], [9]]. Malate plays a central role in the maintenance of intracellular mitochondrial membrane potential (MMP) by supplying sufficient H+ to mitochondria through the malate/aspartate shuttle pathway [10,11]. The maintenance of MMP is critical for the survival and energy supply of tumor cells, and this potential difference is not only a direct power source for ATP production, but also the energy basis for continued tumor cell growth and division [[12], [13], [14], [15]]. Therefore, an in-depth study of the specific mechanism of malate in the process of tumor genesis and progression, especially targeting the malate/aspartate shuttle pathway for intervention, inducing depolarization of the mitochondrial membrane potential and prompting apoptosis of tumor cells, will provide an important rationale for the development of innovative anticancer strategies and the improvement of the efficiency of existing therapeutic regimens.
Malate dehydrogenase is a key enzyme involved in cellular metabolism and plays an important role especially in certain tumor cells [[16], [17], [18]]. In recent years, several researchers have identified a variety of inhibitors, such as specific phenolic compounds and small molecule drugs, that effectively inhibit malate dehydrogenase activity [[19], [20], [21]]. Among them, resveratrol, a compound of natural origin, is able to inhibit the proliferation of tumor cells due to its antioxidant and anti-inflammatory properties, and has shown inhibitory effects on malate synthesis in some studies [22,23]; Quercetin is a widely available flavonoid with good antioxidant capacity and can effectively inhibit the growth of many types of cancer cells, and it has been suggested that quercetin may affect malate synthesis by inhibiting malate dehydrogenase activity [24,25]; N-acetylcysteine (NAC) has high safety and antioxidant properties and may slow down tumor growth by indirectly inhibiting malate synthesis [[26], [27], [28]]. However, the low bioavailability and rapid metabolism of these compounds in the body make it difficult to maintain their effective concentration in the body, and their use may cause side effects such as muscle pain and liver function abnormalities [[29], [30], [31]]. Therefore, more effective and safer malate regulation tools need to be explored.
Piezoelectric nanomaterials, which convert ultrasound-induced mechanical energy into electrical charges, have recently gained attention for their ability to catalyze redox reactions, thereby offering a noninvasive and controllable strategy for tumor therapy [[32], [33], [34]]. Traditional piezoelectric materials such as BaTiO3 (BTO) and PbTiO3:Zr (PZT) have shown promise in tumor inhibition through ROS generation and redox regulation, but their effects are largely limited to direct oxidative damage or immune stimulation, and their biocompatibility is often restricted due to lead toxicity or low efficiency [[35], [36], [37]]. Recent reviews have highlighted the pivotal roles of tumor metabolic reprogramming and mitochondrial targeting in cancer therapy, suggesting that combining metabolic intervention with immune modulation could achieve superior therapeutic outcomes [[38], [39], [40]].
Here, we have developed a tumor malate depletion modulator (MDM) called GO/BCT:Mn. In contrast to traditional piezoelectric nanomaterials, GO/BCT:Mn incorporates three critical design features: Ca/Mn co-doping of BTO to enhance charge separation and narrow the band gap, integration of conductive graphene oxide (GO) to facilitate electron transfer, and the capacity to simultaneously regulate tumor metabolism while activating anti-tumor immunity. This combination establishes a distinctive dual “metabolism–immunity” regulatory mechanism. Upon ultrasound stimulation, charge carriers generated on the nanoparticle surface exhibit pronounced reducing and oxidizing activities. Specifically, the electrons drive the reduction of H+ to H2, thereby lowering the MMP while concurrently providing a gas therapy effect, whereas the holes oxidize NADH to NAD+, suppressing malate synthesis and disrupting the malate/aspartate shuttle pathway. This significantly reduces tumor MMP, causing excessive mitochondrial autophagy and apoptosis in tumor cells (Fig. 1). When applying GO/BCT:Mn MDM for tumor treatment, we observed a significant decrease in malate levels and MMP in tumors, which effectively interfered with the metabolic processes of the tumor cells and consequently inhibited their growth. There was a significant reduction in the expression of cell proliferation marker Ki67, angiogenesis markers VEGF and CD31, and tumor hypoxia marker HIF-1α. Meanwhile, in terms of apoptosis markers, we observed enhanced expression of TUNEL and Caspase-3. These results suggest that the piezoelectric catalysis of GO/BCT:Mn MDM not only showed effects in inhibiting the growth and metabolism of tumor cells, but also activated the apoptotic pathway, thus demonstrating efficient efficacy in CT26 colon cancer treatment. In the Balb/C mouse model, treatment markedly suppressed tumor growth, and all treated animals remained alive throughout the 40-day observation period. However, further studies are needed to assess long-term survival and potential toxicity. Transcriptomic analyses further revealed activation of immune-related pathways, including antigen presentation and T cell activation. Flow cytometry confirmed increased infiltration of CD4+/CD8+ T cells and mature dendritic cells, highlighting the immunostimulatory effects of metabolic disruption. These findings collectively emphasize the superior capabilities of GO/BCT:Mn over traditional piezoelectric nanoparticles, enabling concurrent metabolic disruption and immune activation, and offering a promising strategy for the design of advanced metabolism-regulating therapies with improved efficacy and safety.

2
Results and discussions
2.1
Synthesis and characterization of Graphene Oxide/Manganese-Doped Barium Calcium Titanate
Graphene oxide (GO)/Ba0.85Ca0.15Ti0.9Mn0.1O3 (BCT:Mn) malate depletion modulator (MDM) were prepared through a two-step process, as illustrated in Fig. 2a. The introduction of 25 % anhydrous ethanol into the reaction system facilitated the generation of smaller nanoparticles and synthesized BCT:Mn with a size of 75.48 ± 9.20 nm (Fig. S1a). Subsequently, a second hydrothermal synthesis involved a BCT:Mn to GO mass ratio of 9:1, with polyvinyl pyrrolidone (PVP) serving as a binder to enhance the stability of the composite. GO/BCT:Mn, with a size of 83.43 ± 14.14 nm, displayed a crystal plane spacing of 4.20 Å, aligning with the (001) crystal plane of BaTiO3 (BTO) (Fig. 2b–d). The increase in nanoparticle size was attributed to the larger dimensions of GO (0.2–10 μm), each encasing several nanoparticles. Throughout the hydrothermal reaction, smaller nanoparticles decomposed and regrew on the surface of larger ones, ultimately leading to an enlargement in the synthesized nanoparticles' size. Additionally, BTO, GO/BTO and GO/Ba0.85Ca0.15TiO3 (GO/BCT) with similar dimensions were also successfully synthesized (Fig. S1b-d and Fig. S2). High-resolution transmission electron microscope (HRTEM) images confirmed the effective binding of BCT:Mn with GO (Fig. 2d). Further characterization through high-angle circular dark field (HAADF) imaging, elemental mapping and energy dispersive spectrometer (EDS) of GO/BCT:Mn revealed a well-distributed presence of Ba, Ca, Ti, Mn, and O within the composite structure (Fig. 2e, f and Fig. S3).
The Single-crystal electron diffraction pattern and X-ray diffraction (XRD) pattern of GO/BCT:Mn provides a clear depiction of the crystal structure of the synthesized nanoparticles, and the positions of the peaks align with the standard card (JCPDS: 05–0626) for the tetragonal phase of BaTiO3 (Fig. 2g and Fig. S4). Notably, distinct diffraction peaks at 44.855° and 45.377° correspond respectively to the (002) and (200) crystal planes of the tetragonal phase BTO. Throughout the hydrothermal synthesis, there is a phase transition in the crystalline structure of the nanoparticles, shifting from a cubic phase to a tetragonal phase. To prevent oversizing, the reaction temperature and duration are intentionally lowered, leading to a controlled adjustment and resulting in a mixed phase in the crystalline structure of the nanoparticles. Zooming in on the region of 44–46° in the XRD pattern, following Bragg's law of diffraction (2dsinθ = nλ), the peak positions of GO/BCT exhibit a slight angular shift towards larger angles compared to GO/BTO due to the smaller atomic radius of Ca compared to Ba. Conversely, with Mn having a larger atomic radius than Ti, the peak positions of GO/BCT:Mn show a slight angular shift towards smaller angles compared to GO/BCT. This shift provides evidence for the effective doping of Ca and Mn elements into the BTO. The chemical composition and oxidation states of GO/BCT:Mn were explored using X-ray photoelectron spectroscopy (XPS) (Fig. 2h). The peaks of Ba 3 d, Ca 2p, Ti 2p, Mn 2p, O1s and C1s can be clearly visible from from the survey spectrum and detailed XPS spectrum. The presence of GO is indicated by the presence of characteristic C=O peaks in the O 1s and C 1s spectra (Fig. S5). Additionally, Mn 2p spectra exhibit peaks at 641.17 eV (641.8 eV) and 653.24 eV (653.7 eV), corresponding to Mn 2p1/2 and Mn 2p3/2, with an observed difference of approximately 12.07 eV (11.9 eV), indicative of Mn4+ (Fig. 2i) [41].
Fourier-transform infrared (FTIR) spectroscopy offers further evidence of the binding between BCT:Mn and GO (Fig. 2j). The distinctive absorption peaks at 532 cm−1 and 683 cm−1 correspond to Ti-O and Ba-O bonds, signifying the successful synthesis of BTO-based nanoparticles. The peaks around 1700 cm−1 indicate C=O stretching and bending vibrations, confirming the presence of graphene oxide. Zooming in on the 1500–1800 cm−1 range reveals strong C=O stretching and bending vibration peaks in GO-combined nanoparticles, exhibiting a redshift in peak positions. This could be due to the formation of a heterojunction, influencing their vibrations. In the Raman spectrum, the characteristic peaks of BTO at 307 cm−1 and 713 cm−1 represent the tetragonal phase of BTO (Fig. 2k). Moreover, peaks at 1348 cm−1 and 1610 cm−1 align with the characteristic peaks of GO, reaffirming the binding of nanoparticles with GO. Thermal gravimetric (TG) analysis further confirmed the composition of hydrothermally synthesized GO/BCT:Mn (Fig. 2l). At 800 °C, residue percentages were 97.78 % for BCT:Mn, 89.83 % for GO/BCT:Mn, and 0.47 % for GO. The noticeable decrease in GO mass around 190 °C is attributed to released bound water overcoming van der Waals forces, transitioning to water vapor. However, in hydrothermally synthesized nanoparticles, the absence of bound water results in no sudden mass loss for GO/BCT:Mn at 190 °C. Around 500 °C, both GO and GO/BCT:Mn display reduced mass, indicating the presence of GO in GO/BCT:Mn. The preceding findings also underscore that the mass ratio of GO to BCT:Mn in GO/BCT:Mn stands at merely 1:12.

2.2
First-principles calculations and piezoelectric property analysis of GO/BCT:Mn
In our research, density-functional theory (DFT) computations were conducted using the VASP software to investigate the impacts of doping with elements like Ca and Mn on nanoparticle characteristics. Fig. 3a shows a schematic of the internal polarization of GO/BCT:Mn under ultrasonic stimulation. Under ultrasonic stimulation, BCT:Mn generates electrons and holes on its surface, and GO can be used as a conductive filler to combine with BCT:Mn to form conductive paths while increasing its dispersion, thus enhancing the piezoelectric properties of the material. The focus of our study was the alterations in differential charge density and their influence on the electronic attributes of these nanoparticles. Our findings reveal that the formation of heterojunctions with graphene oxide led to an increase in the amount of electron cross-transfer within the structure, which also implies an enhanced overlap of electrons between elements. Meanwhile, the disappearance of the strong spin polarization of Ti d orbitals and the enhancement of the spin polarization of O p orbitals after the formation of the Go/BaTiO3 heterojunction may be attributed to the formation of covalent bonds (Fig. 3b–e). Whereas doping with Ca notably escalates electron transfer within the nanoparticle, indicating a more pronounced overlap of electrons among different elements (Fig. 3f and g). This enhanced overlap is typically indicative of stronger covalent bonds, as it suggests a heightened sharing of electrons among atoms, thereby reinforcing the nanoparticle's covalent nature. Specifically noteworthy is the interaction between the Ti 3 d orbitals and oxygen's p orbitals around the Fermi level, demonstrating a significant overlap and hybridization. Such orbital intermingling is fundamental to the formation of covalent bonds, facilitating electron mobility across orbitals and expanding the distribution of electrons spatially, which in turn strengthens the covalent interactions within the nanoparticle. Additionally, our examination of the overall density of states unveils that the incorporation of Ca and Mn, along with the creation of heterostructures, enriches the electronic state density at the Fermi level, implying a bolstered conductive and covalent capacity of the nanoparticles (Fig. 3h and i). Collectively, these insights suggest that the doping of Ca and Mn, along with the resultant electronic structural modifications, are crucial in enhancing both the conductivity and the strength of covalent bonding in these nanoparticles.
The band gaps of BTO, GO/BTO, GO/BCT and GO/BCT:Mn were determined to be 3.00 eV, 2.74 eV, 2.66 eV and 2.19 eV (Fig. 3j and Fig. S6), using ultraviolet–visible (UV–Vis) diffuse reflectance spectroscopy (DRS). Encapsulation of BTO-based nanoparticles with conductive graphene oxide (GO) forms conductive pathways and heterojunctions, reducing the band gap, while Ca and Mn co-doping further narrows the band gap, enhancing charge carrier generation and separation. Electrochemical impedance spectroscopy (EIS) measurements revealed that GO/BCT:Mn exhibited lower charge-transfer resistance than GO/BTO, further supporting improved electron mobility and catalytic activity (Fig. S7a). The valence band (VB) of GO/BTO and GO/BCT:Mn, calculated via ultraviolet photoelectron spectroscopy (UPS), is measured at 3.86 eV and 3.64 eV, further corroborating enhanced charge separation and catalytic performance of GO/BCT:Mn when considered alongside DRS and EIS results (Fig. 3k). Moreover, current–time responses of GO/BCT:Mn under varying ultrasonic power stimulations demonstrate the robustness of its enhanced charge separation and transfer properties (Fig. S7b).
Piezoelectric performance was assessed using piezoresponse force microscopy (PFM) for GO/BCT:Mn. At an applied voltage of 10 V, both GO/BTO and GO/BCT:Mn pressed-pellet samples exhibited significant piezoelectric response in the amplitude images (Fig. S8a and b). Surface roughness 3D images revealed heights of 298.2 nm and 115.2 nm for GO/BTO and GO/BCT:Mn, with deviations from TEM images likely due to nanoparticle stacking (Fig. S8c and d). Analysis of typical butterfly-shaped voltage curves indicated that, as the DC voltage ranged from −10 to +10 V, the maximum amplitude voltages for GO/BTO and GO/BCT:Mn were 103.4 mV and 448.8 mV. These curves exhibited distinct hysteresis shapes with voltage differences of 0.31 V and 0.39 V, highlighting the superior piezoelectric response of GO/BCT:Mn (Fig. S9). Additionally, the phase maps for both GO/BTO and GO/BCT:Mn displayed a 180° phase transition as the DC voltage varied from −10 to +10 V, indicating the ferroelectric polarization switching process. Single-particle PFM measurements further verified these results. 3D surface roughness images of individual nanoparticles confirmed uniform morphology (Fig. S10), while voltage–amplitude and phase curves of single nanoparticles showed consistent hysteresis behavior (Fig. S11). The piezoelectric coefficient (d33), a parameter directly reflecting the piezoelectric performance of nanoparticles, was measured as 1.79 mV/V for GO/BTO and 7.50 mV/V for GO/BCT:Mn (Fig. S12), despite more severe stacking in the pressed-pellet samples of GO/BCT:Mn. Single-particle measurements, after conversion from the applied driving voltage, yielded d33 values of 9.59 p.m./V for GO/BTO and 13.67 p.m./V for GO/BCT:Mn (Fig. 3l), confirming that GO/BCT:Mn exhibits superior piezoelectric performance even at the individual nanoparticle level. These results collectively demonstrate the excellent piezoelectric properties of the synthesized GO/BCT:Mn nanoparticles.
To evaluate the stability of GO/BCT:Mn under physiological conditions, the nanoparticles were incubated in demineralized water, 1640 medium, serum, and 1640 medium supplemented with 10 % serum, and their hydrodynamic diameters were monitored over 7 days using dynamic light scattering (DLS) (Fig. S13). The nanoparticles maintained consistent size distributions in all media throughout the observation period, demonstrating excellent colloidal stability and minimal aggregation under long-term physiological conditions. The average particle size in demineralized water was 248.23 nm, slightly larger than TEM measurements, which is attributable to the hydration layer and minor aggregation in suspension. These results indicate that GO/BCT:Mn nanoparticles are stable in biologically relevant media, supporting their potential for reliable in vitro and in vivo applications. Subsequently, it was examined for H+ reduction and NADH oxidation. Fig. 3m illustrates the augmented electron transport process within the GO/BCT:Mn system. Ultrasound stimulation induces a bending of the energy bands in GO/BCT:Mn, causing electrons to migrate towards lower energy levels and holes towards higher energy levels, thereby creating polarization within the material. Specifically, the presence of Mn4+ on the surface increases the abundance of oxygen holes, amplifying the electron-hole separation process and intensifying polarization. Simultaneously, electron holes, generated through polarization, combine with H+ and NADH on the nanoparticle surface to produce H2 and NAD+. In Fig. 3n, an experimental setup for hydrogen testing is schematically depicted. Hydrogen yields from five experimental groups were evaluated under ultrasound stimulation, revealing that the GO/BTC:Mn group exhibited significantly higher hydrogen yields compared to the control group (Fig. 3o). Fig. S14 illustrates the process through which holes, generated by the polarization of GO/BCT:Mn, oxidize nicotinamide adenine dinucleotide hydride (NADH) to nicotinamide adenine dinucleotide (NAD+). This principle guided the assessment of changes in the UV–Vis absorption peaks of NADH and NAD+ in six groups under ultrasound stimulation (Fig. S15). The NAD+ yield, calculated from relative peak intensities, increased from 0 % in the control to 2 %, 33 %, 36 %, 8 %, and 69 % in the BTO, GO/BTO, GO/BCT, BCT:Mn, and GO/BCT:Mn groups, demonstrating the superior catalytic oxidation of NADH by GO/BCT:Mn (Fig. S16). This observation indicates that the piezoelectric nature of GO/BCT:Mn facilitates the substantial oxidation of NADH to NAD+ under ultrasound stimulation.

2.3
GO/BCT:Mn MDM interaction with CT26 cells
In our study, we assessed the cells' survival and death in the different experimental groups (Fig. 4a). This was achieved by capturing fluorescence microscopy images of calcein-AM/PI stained cells. The experimental setup consisted of six different groups: control, BTO + US, GO/BTO + US, GO/BCT + US, GO/BCT:Mn and GO/BCT:Mn + US. Our results showed that the GO/BCT:Mn + US combination had a significant effect on reducing cell viability compared to the other groups (Fig. 4b). To further validate the cytotoxic effect of the GO/BCT:Mn nanoplatform upon ultrasound activation, we conducted Annexin V/PI-based flow cytometry analysis to quantitatively assess apoptosis in CT26 cells (Fig. S17). Compared with the control and single-treatment groups, the GO/BCT:Mn + US group exhibited a significantly higher proportion of Annexin V+/PI− (early apoptotic) and Annexin V+/PI+ (late apoptotic) cells, indicating robust induction of programmed cell death upon ultrasound-triggered piezoelectric activation. In contrast, minimal apoptosis was observed in the GO/BCT:Mn alone or US alone groups, confirming that the therapeutic efficacy is dependent on the combined action of the nanomaterial and ultrasound stimulation. These results further support the potent and controllable cytotoxicity of our sonodynamically activated modulator system.
The mitochondrial proton pump is an important component located in the inner mitochondrial membrane. The presence of the proton pump leads to the accumulation of protons in the membrane interstitial space, creating a proton gradient and leading to a significant electrical change in the mitochondrial membrane: the membrane interstitial space is positively charged and the substrate is negatively charged. This generates the mitochondrial transmembrane potential (MTP), which is essential for maintaining mitochondrial functions including oxidative phosphorylation and ATP production. On this basis, we measured the mitochondrial membrane potential (MMP) in different groups (Fig. 4c). It was found that mitochondrial membrane potential depolarization was most pronounced in cells treated with GO/BCT:Mn + US (Fig. 4d). In addition, we conducted a series of experiments to illustrate the effects of this treatment on cells. Autophagy is an important cellular mechanism that converts excess or unnecessary components into valuable nutrients. Triggered by starvation or chemical stimuli, autophagy begins to form isolation membranes that encapsulate the target material, eventually forming autophagosomes. These segregation membranes then fuse with lysosomes, allowing the acidic hydrolases in them to degrade the inclusions, thus completing the autophagy process. Studies have shown that tumor cells depend on the mitochondrial autophagy process to maintain the cellular energy supply, and disrupting the balance of autophagy within tumor cells will reduce the energy supply of mitochondria to tumor cells. In view of this, we assessed the level of treatment-induced mitochondrial autophagy using a specialized kit containing Mtphagy dye for staining autophagic mitochondria and Lyso dye for staining lysosomes. The Mtphagy dye chemically binds to mitochondria and under normal conditions emits a weak fluorescence. However, during mitochondrial autophagy, fusion with lysosomes and the resulting acidic environment enhances this fluorescence. Double staining with lysosomal dyes allows a comprehensive view of the mitochondrial autophagy process. As shown in Fig. 4e–g, the results showed that the GO/BCT:Mn + US group had the strongest induction of mitochondrial autophagy compared to the other groups.
Typical mitochondrial autophagy was detected by TEM beginning 12 h after ultrasound stimulation, deformed and damaged mitochondria encircled by forming autophagosomes were observed (Fig. 4h). The GO/BCT:Mn + US group exhibited the highest formation of autophagosomes (Fig. 4i). The results once again showed that the GO/BCT:Mn + US group had the strongest induction of mitochondrial autophagy compared to the other groups. This observation is particularly important because it suggests that the treatment has a profound effect on mitochondrial function, which may be a key factor in the mechanism of action of the treatment on tumor cells.

2.4
Interaction of GO/BCT:Mn MDM with tumor spheres
In this investigation, we explored the cellular internalization of GO/BCT:Mn MDM, marked with rhodamine, over several time intervals - specifically at 0, 1, 2, 4, 8, and 12 h. Using confocal microscopy, we observed that cell uptake of the material commenced within 1 h and progressively increased with time. To better replicate the conditions within a tumor, we developed a tumor spheres model (Fig. 5a). This approach enabled us to recreate aspects of tumor spheres, such as hypoxic conditions, which are unachievable in standard two-dimensional cell cultures. The cultivation of CT26 cancer cell spheres was conducted in tailored media and under specific conditions to facilitate the formation and growth of the spheres, which were closely monitored using the Invitrogen EVOS imaging system. Further, we assessed the incorporation of the GO/BCT:Mn MDM into tumor spheres. Confocal microscopy was employed to capture images at the largest cross-sectional areas at intervals of 1, 2, 4, 6, 8, and 12 h (Fig. 5b). Notably, material uptake was evident after the initial hour and continued to increase over time. To determine the presence of hypoxic conditions within the organoids, we utilized Hypoxyprobe and Pimonidazole as a hypoxic cell marker (Fig. 5c). The findings confirmed the presence of hypoxia within the tumor spheres. After an 8-h co-incubation of GO/BCT:Mn with the tumor spheres followed by ultrasound stimulation, we evaluated metabolic perturbations. ESR measurements confirmed that, under ultrasound, GO/BCT:Mn generates a significant amount of hydroxyl radicals (·OH) but only minimal superoxide radicals (·O2−) in vitro (Fig. S18), indicating that ROS-mediated effects are largely dominated by ·OH. The schematic in Fig. 5e illustrates the proposed mechanism of malate depletion in tumor metabolism. Consistently, NAD+/NADH, malate, oxaloacetate, and ATP levels in the tumor spheres (Fig. 5f–i and Fig. S19) showed substantial metabolic disruption, supporting the effect of ultrasound-triggered charge separation on mitochondrial function. Compared to the control group, the levels of malate, glucose and ATP in the cell cytosol were significantly decreased, whereas the levels of NAD+/NADH were significantly increased. These observations suggest that the piezoelectric material disrupts the intracellular equilibrium between NADH and NAD+ under ultrasonic stimulation and affects the conversion of oxaloacetate to malate. Given the key role of malate in the malate/aspartate shuttle pathway, the ATP produced through this process is essential for maintaining tumor cell viability. We went on to examine the effect of GO/BCT:Mn MDM on the activity of the mitochondrial respiratory chain complex I-V in tumor spheres after ultrasound stimulation (Fig. 5j). Compared with the control group, the activities of mitochondrial respiratory chain complexes I-V were all significantly reduced in tumor spheres after ultrasound stimulation, especially the activity of mitochondrial respiratory chain complex I was substantially reduced. Thus, our study demonstrates that under the influence of ultrasound, piezoelectric materials enhance the conversion of intracellular NADH to NAD+, thereby affecting malate production and the subsequent malate/aspartate shuttle pathway.
To comprehensively assess the safety and mechanistic effects of GO/BCT:Mn, cytotoxicity was evaluated across multiple cell lines, including tumor cells (B16, 4T1) and normal cells (3T3 fibroblasts, BV2 microglia), using MTT assays (Fig. S20). Under ultrasound stimulation, GO/BCT:Mn exhibited effective cytotoxicity against B16 and 4T1 tumor cells, whereas in the absence of ultrasound, even at 50 μg/mL, the nanoparticles showed negligible toxicity toward 3T3 and BV2 cells. ICP-MS analysis of CT26 culture supernatants further quantified Mn4+ release, which reached 76.74 μg/L after 24 h (Fig. S21), indicating controlled ion leaching under physiological conditions. These results demonstrate that GO/BCT:Mn exerts ultrasound-triggered tumor-selective cytotoxicity while maintaining a favorable safety profile for normal cells.

2.5
Analysis of therapeutic effect of GO/BCT:Mn MDM in tumor model mice
In our investigation, a CT26 tumor model was established in female Balb/C mice by subcutaneously administering 2 million CT26 cells into the dorsal flank (Fig. 6a). Tumors were allowed to develop for seven days post-inoculation. We then administered the GO/BCT:Mn MDM intravenously at 50 μg/mL, and utilized rhodamine-tagged materials for tracking purposes during live animal imaging conducted 12 h following administration (Fig. 6b–f). The majority of the administered material localized to the tumor site as evidenced by the EPR effect, a finding corroborated by ex vivo organ fluorescence imaging. Further in vivo distribution studies of the GO/BCT:Mn MDM post-intravenous administration were performed using ICP-MS (Fig. 6g). Notably, elevated levels of barium were observed in organs associated with the reticuloendothelial system, such as the liver, spleen, and lungs, as well as the kidneys. Barium ion concentrations within the tumor tissues reached a maximum 12 h post-injection (9.84 % ID/g), which, although modest, together with localized ultrasound activation, is sufficient to induce therapeutic effects. Concentrations remained appreciable at 48 h (3.95 % ID/g), indicating sustained presence at the tumor site. The pharmacokinetic profile of the GO/BCT:Mn MDM revealed half-lives of 4.73 h for the alpha phase and 15.31 h for the beta phase, suggesting an advantageous circulation duration for tumor accumulation (Fig. S22). Despite high uptake in liver and spleen due to reticuloendothelial clearance, significant tumor growth inhibition and immune activation were observed, demonstrating that therapeutic efficacy is achieved through sonodynamic metabolic disruption at the tumor site rather than solely relying on passive accumulation. In therapeutic studies, 42 tumor-bearing mice were stratified into six treatment cohorts: (i) Control (PBS), (ii) Contriol + US, (iii) BTO + US, (iv) GO/BTO + US, (v) GO/BCT + US, (vi) GO/BCT:Mn, and (vii) GO/BCT:Mn + US, with treatments administered post nanoparticle injection (Fig. 6h–i and S23a-e). Notably, the combination of GO/BCT:Mn MDM with ultrasound (GO/BCT:Mn + US group) led to an outstanding tumor growth inhibition rate of 98.28 %, a stark contrast to the negligible impact seen in the control group. The suppression of tumor growth by the GO/BCT:Mn + US treatment is ascribed to the ultrasound-induced polarization of the piezoelectric nanoparticles. This polarization resulted in the formation of electron-rich and hole-rich regions, promoting the reduction of protons to hydrogen gas and the oxidation of NADH to NAD+. This mechanism disrupts the conversion of oxaloacetate to malate in tumor cells and disrupts the malate/aspartate shuttle pathway, thereby curtailing tumor cell proliferation and migration, while also ultimately leading to disruption of nutrient support to the tumor. These concerted actions culminate in the suppression of tumor proliferation, induction of apoptosis, and inhibition of migration. During the detailed 7-day and 14-day monitoring phases, daily weight tracking of all experimental mouse groups was diligently conducted (Fig. 6j and S23f). The absence of notable weight fluctuations across these groups implies the non-toxic systemic nature of the intravenously delivered GO/BCT:Mn MDM. Mice receiving the combined GO/BCT:Mn + US therapy demonstrated enhanced longevity with no tumor re-emergence, signifying the combined treatment's efficacy in elevating the survival rates of tumor-bearing mice (Fig. 6k).

2.6
Immune response remodeling in tumor tissues treated with GO/BCT:Mn MDM
GO/BCT:Mn MDM modulates tumor metabolism and may influence antitumor immune responses. To systematically evaluate these effects, we analyzed immune cell infiltration, activation status, metabolic profiles, and gene expression in tumor tissues following treatment. Quantitative flow cytometry revealed that GO/BCT:Mn + US–treated tumors exhibited markedly increased leukocyte (CD45+) infiltration (43.6 %) compared with controls (Fig. 7a and b). Notably, neutrophils (42.8 %) and dendritic cells (36.1 %) were significantly enriched (Fig. S25a and b), with mature dendritic cells (CD45+CD11c+I-A/I-E+) increasing to 34.1 % and cross-presenting dendritic cells reaching 90.7 % (Fig. 7c and d; Fig. S25c), indicating enhanced antigen presentation and potential CD8+ T cell priming. Regulatory T cells (CD3+CD4+Foxp3+) were markedly reduced (25.9 %), reflecting a shift toward an immunostimulatory tumor microenvironment (Fig. 7e and f). Concurrently, intratumoral CD4+ (37.1 %) and CD8+ (32.6 %) T cell infiltration increased significantly (Fig. 7g–j), consistent with potentiated adaptive immune responses. These immunophenotypic changes were further corroborated by increased macrophage activation (46.2 %) and memory T cell activation (28.9 %) (Fig. S25d and e).
Metabolic profiling of tumor tissues revealed that GO/BCT:Mn + US treatment substantially disrupted key metabolites involved in energy production and anabolic processes. Levels of pyruvic acid, glycerol 3-phosphate, succinic acid, D-ribulose 5-phosphate, L-asparagine, L-lactate, AMP, and IMP were significantly altered compared with controls (Fig. 7k), indicating interference with glycolysis, the TCA cycle, and nucleotide metabolism. These metabolic perturbations likely contribute to immune reprogramming by modulating the availability of metabolites that influence immune cell activation, proliferation, and effector functions, linking metabolic disruption directly to enhanced antitumor immunity. To elucidate the molecular mechanisms underlying these immunological changes, we conducted transcriptomic analysis. Principal component analysis (PCA) revealed distinct transcriptional signatures in GO/BCT:Mn + US treated tumors compared to controls (Fig. 7l), indicating substantial gene expression reprogramming. Differential expression analysis (Fig. 7m) identified numerous upregulated and downregulated genes, while Gene Ontology (GO) enrichment analysis (Fig. 7n) highlighted key immune-related biological processes, including adaptive immune response, T cell activation, leukocyte proliferation, antigen processing and presentation, response to type II interferon, and cell killing. Among these, we observed a significant downregulation of Cd163, a hallmark of immunosuppressive M2-like tumor-associated macrophages (TAM), suggesting a reduction in pro-tumor macrophage populations and a transition toward a pro-inflammatory microenvironment. Simultaneously, genes associated with T cell activation, including Cd2, Cd4, Cd5, Cd6, Cd8a, Cd27, Cd28 and Cd40, were robustly upregulated, reflecting enhanced activation and proliferation of both helper and cytotoxic T cells. Genes involved in antigen processing and presentation, including MHC class II molecules (H2-Aa, H2-Ab1, H2-Eb1) and associated processing machinery (Tap1, Tap2, Tapbp, Tapbpl, Psmb8, Psmb9, Psmb10), were also significantly upregulated, supporting improved tumor antigen visibility. Chemokines such as Cxcl9, Cxcl10 and Cxcl16 and their corresponding receptors were markedly induced, facilitating effective immune cell trafficking to the tumor site. In addition, increased expression of Cd83 and Cd274 indicates APC maturation and modulation of immune checkpoints, respectively. Upregulation of cytokine-related genes, including Il27, Il27ra, Il12rb1, Il10ra and Il18r1, further reflects a pro-inflammatory cytokine milieu, consistent with enhanced type I/II interferon signaling and a Th1-skewed immune response. Notably, elevated Klrk1 expression suggests increased cytotoxic potential of both NK and CD8+ T cells. These findings are further supported by the DEG heatmaps (Fig. 7o, Fig. S26), which show broad upregulation of immune-activating genes across multiple pathways, confirming that GO/BCT:Mn + US treatment effectively reprograms the tumor immune microenvironment to favor antitumor immunity.

2.7
Response of mouse tumor tissues to GO/BCT:Mn MDM treatment
We further examined the therapeutic effects of GO/BCT:Mn MDM by performing histological staining and apoptosis assays on tumor tissue sections, including hematoxylin and eosin (H&E) staining and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining. These were followed by immunofluorescence analysis to quantify the expression levels of Ki67, VEGF, Caspase3, CD31, and HIF-1α. H&E results revealed extensive apoptosis and tissue damage in the tumor tissues of the GO/BCT:Mn + US treatment group, with disrupted tumor cell organization and condensed nuclei, indicating significant damage inflicted on the tumor tissues (Fig. 8a). The GO/BCT:Mn + US treatment group exhibited markedly increased TUNEL and Caspase3 activity, along with reduced levels of Ki67, VEGF, CD31 and HIF-1α, consistent with the observed tumor inhibition (Fig. 8b–h). In this group, the tumor tissue suffered extensive damage with an elevated count of apoptotic cells, the highest among all the study groups. Meanwhile, Caspase3 results indicated that the treatment activated the Caspase3 pathway, promoting apoptosis. Furthermore, the treatment significantly suppressed tumor growth and angiogenesis by inhibiting cell proliferation (Ki67), downregulating the expression of vascular endothelial growth factor (VEGF) and hypoxia-inducible factor (HIF-1α), and reducing the activity of vascular endothelial cells (CD31).

2.8
Response of mouse organs to GO/BCT:Mn MDM treatment
We further investigated the biosafety of GO/BCT:Mn MDM by performing histological staining and inflammatory cytokine assays on mouse isolated tissue sections. Tissue sections from major organs, including liver and kidney, were subjected to hematoxylin and eosin (H&E) staining, and high-power microscopic images were obtained. No signs of tumor metastasis or pathological alterations were observed, indicating minimal organ toxicity and confirming the biosafety of GO/BCT:Mn MDM (Fig. 9a). In parallel, inflammatory cytokine levels (TNF-α, IFN-γ, IL-6, IL-8, and IL-1β) in serum were quantified using enzyme-linked immunosorbent assays (ELISA) (Fig. 9b–f). No significant alterations in cytokine levels were observed in the GO/BCT:Mn + US treatment group, indicating minimal inflammatory response and high biosafety.There were no significant changes in TNF-α, IFN-γ, IL-6, IL-8, and IL-1β activities in the GO/BCT:Mn + US treatment group, reflecting the high biosafety of GO/BCT:Mn MDM. For an evaluation of GO/BCT:Mn MDM's biocompatibility in real-world medical applications, we performed full blood panel and biochemical tests on healthy Balb/C mice post nanoparticle administration (Fig. 9g and h). Observations revealed that the liver and kidney functions remained stable, with no discernible deviations post-injection. Comparative analysis of hematological markers over the assessment period showed consistency with the control group, suggesting the nanoparticles had negligible hematological side effects. Furthermore, after a 15-day therapeutic period, we noted no evident inflammatory responses or damage in the mice's vital organs across all treatment groups. These comprehensive findings underscore the biocompatibility and safety profile of the GO/BCT:Mn MDM, reinforcing their suitability for therapeutic use within living organisms.

3
Conclusion
In this study, we demonstrate a breakthrough approach to effectively achieve tumor metabolic remodeling by creating a nano-formulation called GO/BCT:Mn malate depletion modulator (MDM). This unique nano-formulation combines the advantages of piezoelectric effect and ultrasound technology, introducing a novel, dual catalytic mechanism through the co-doping of calcium and manganese elements and the encapsulation of GO. This represents an innovative advancement in nanotechnology and tumor therapy and provides us with a new means to inhibit tumor growth and proliferation by directly interfering with the metabolic pathways in tumor cells. GO/BCT:Mn MDM effectively disrupts important metabolic pathways in tumor cells through the high reduction potential and high oxidation potential generated on its surface. In particular, GO/BCT:Mn MDM effectively lowers the mitochondrial membrane potential through the reduction of H+ to H2, as well as blocks malate dehydrogenation by oxidizing NADH to NAD+, thus blocking malate dehydrogenase-mediated malate production and opening a new pathway for apoptosis in tumor cells. This unique mechanism of action not only limits the nutrient uptake and proliferative capacity of tumors but also proves its high efficiency and safety in tumor therapy by significantly decreasing the malate content in tumors, inhibiting tumor cell proliferation, and activating apoptotic pathways. In vivo studies in Balb/c mice demonstrated that GO/BCT:Mn MDM not only effectively inhibited tumor growth and was well tolerated over the 40-day observation period, but also promoted infiltration and activation of dendritic cells and CD4+/CD8+ T cells while reducing immunosuppressive populations, collectively eliciting a robust antitumor immune response. These findings underscore the therapeutic promise of GO/BCT:Mn MDM as a multifunctional nanoplatform that integrates metabolic disruption and immune activation for synergistic tumor suppression.

4
Method
4.1
Synthesis of BaTiO3 nanoparticles (BTO NPs)
BaTiO3 nanoparticles were synthesized using a typical hydrothermal reaction. For the synthesis procedure, 1.897 g (10 mmol) TiCl4 was dissolved in 10 mL anhydrous ethanol. Then, 3.664 g (15 mmol) BaCl2·2H2O was dissolved in 30 mL deionized water. Further, the two clear solutions were mixed before adding 3.2 g (80 mmol) NaOH and 1.0 g PVP. Finally, the suspension was transferred to a 50 mL stainless steel autoclave lined with tetrafluoroethylene and reacted at 220 °C for 24 h. At the end of the reaction, the obtained precipitates were repeatedly washed with deionized water and anhydrous ethanol until pH of 7, and then dried in an oven at 60 °C for 24 h.

4.2
Synthesis of Ba0.85Ca0.15TiO3 nanoparticles (BCT NPs)
Synthesis is similar to BTO NPs. The only difference is that 3.114 g (12.75 mmol) BaCl2·2H2O and 0.166 g (2.25 mmol) CaCl2 were dissolved in 30 mL deionized water.

4.3
Synthesis of Ba0.85Ca0.15Ti0.9Mn0.1O3 nanoparticles (BCT:Mn NPs)
Synthesis is similar to BTO NPs. The only difference is that 1.707 g (9 mmol) TiCl4 was dissolved in 10 mL anhydrous ethanol. Then, 3.114 g (12.75 mmol) BaCl2·2H2O, 0.166 g (2.25 mmol) CaCl2 and 0.126 g (1 mmol) MnCl2 were dissolved in 30 mL deionized water.

4.4
Synthesis of graphene oxide (GO)/BaTiO3 nanoparticles (GO/BTO NPs)
GO/BTO NPs were synthesized using a typical hydrothermal reaction. For the synthesis procedure, 500 mg BTO NPs, 50 mg GO and 250 mg PVP were dissolved in 35 mL deionized water. Then, the suspension was transferred to a 50 mL stainless steel autoclave lined with tetrafluoroethylene and reacted at 220 °C for 24 h. At the end of the reaction, the obtained precipitates were repeatedly washed with deionized water and anhydrous ethanol, and then dried in an oven at 60 °C for 24 h.

4.5
Synthesis of GO/Ba0.85Ca0.15TiO3 and GO/Ba0.85Ca0.15Ti0.9Mn0.1O3 nanoparticles (GO/BCT and GO/BCT:Mn NPs)
Synthesis is similar to GO/BTO NPs.

4.6
Characterization of nanoparticles
The morphology and size of NPs were characterized by SEM (APREO, FEI) and TEM (JEM-2100F). The hydrodynamic particle size of the NPs was measured with a DLS analyzer (Malvern Zetasizer Nano ZS90). The composition of the NPs was determined by Fourier transform infrared spectroscopy (FTIR) on a Thermo-Nicolet Nexus 670 ATR-IR spectrometer. UV absorption spectra were recorded on a Genesys 10S UV–Vis spectrophotometer.

5
Description of the calculation method
In this research, computational analyses were carried out using the VASP (Vienna Ab initio Simulation Package) framework. The core of our computational approach was anchored in Density Functional Theory (DFT), utilizing the GGA-PBE (Generalized Gradient Approximation-Perdew-Burke-Ernzerhof) scheme for the exchange-correlation functional. We employed the Projector Augmented-Wave (PAW) approach to facilitate the interactions between ion cores and valence electr + ons. The cutoff energy for the plane-wave basis set was established at 500 eV. Structural optimization was meticulously conducted, ensuring that Hellmann–Feynman forces were minimized to a threshold of less than 0.02 eV/Å and energy variations were confined below 10∧-5 eV. To effectively minimize the interactions between adjacent layers, a vacuum separation of 25 Å was consistently applied across all models. For the Brillouin zone integration, we adopted a 3 × 3 × 3 Γ-centered k-point mesh in the relaxation phase. Moreover, to capture dispersion forces in our adsorption models, Grimme's DFT-D3 approach was implemented.
5.1
Measurement of hydrogen gas using the unisense hydrogen microelectrode
Initially, an acidic solution with a pH level of 6 was prepared, forming the foundation for all the experimental groups. Various materials, namely GO/BTO, GO/BCT, BCT:Mn, and GO/BCT:Mn, were then added to different portions of this solution, creating five unique groups for the experiment: CON (control), GO/BTO, GO/BCT, BCT:Mn, and GO/BCT:Mn. Each solution, after being mixed with the respective materials, was then poured into individual small beakers. To ensure the retention of hydrogen gas during ultrasonic stimulation, each beaker was tightly sealed at the top with a parafilm or a similar type of sealing film. The sealed solutions were then exposed to ultrasonic waves, under the conditions of 1.5 W/cm2 intensity and a frequency of 1 MHz, for a period of 2 min per sample. Post-ultrasonication, the task of measuring hydrogen gas concentrations was conducted using the Unisense hydrogen microelectrode. This involved the delicate insertion of the microelectrode through the seal of each beaker, taking care not to disrupt the experimental setup. The maximum hydrogen gas concentration detected by the microelectrode for each sample was duly noted as its respective hydrogen level measurement.

5.2
Ultraviolet–visible photometric detection of NADH solutions before and after sonication
Initially, a microacidic solution with a pH value of 6 was formulated, serving as the fundamental reaction medium for this experiment. Into this medium, five distinct groups of nanomaterials – namely BaTiO3 (BTO), Graphene Oxide/BaTiO3 (GO/BTO), Graphene Oxide/Barium Calcium Titanate (GO/BCT), Barium Calcium Titanate doped with Cerium (BCT:Mn), and Graphene Oxide/Barium Calcium Titanate doped with Cerium (GO/BCT:Mn) – were introduced. This was done to maintain uniform nanomaterial concentration across all groups. Subsequently, to each of these groups, NADH was added in a controlled manner (10 mM concentration, 10 μl volume) to preserve experimental homogeneity. Utilizing a UV–visible spectrophotometer, these solutions underwent an initial spectral analysis within the 240 nm–400 nm wavelength range. The focus was primarily on two spectral features: the NAD+ peak at 260 nm and the NADH peak at 340 nm. Following this, the solutions were subjected to ultrasonic stimulation, employing a device set at a power density of 1.5 w/cm2 and a frequency of 1 MHz, for 2 min. Post-ultrasonic stimulation, a second round of spectral analysis was conducted on all groups using the UV–visible spectrophotometer, again concentrating on the alterations observed at the 260 nm NAD+ peak and the 340 nm NADH peak.

5.3
CT26 cancer cell culture
In this study, CT26 mouse colon cancer cells were cultured under sterile conditions. Initially, cryopreserved cells were rapidly thawed at 37 °C and transferred to a 15 mL centrifuge tube containing 10 mL of pre-warmed 1640 medium supplemented with 10 % fetal bovine serum (FBS) and 1 % penicillin-streptomycin to dilute and remove cryoprotectants. The cell suspension was then centrifuged at 1000 rpm for 5 min, and the supernatant was discarded. The cell pellet was re-suspended in fresh 1640 medium with supplements and seeded into sterile Petri dishes at the appropriate density for the planned experiment. The cells were incubated at 37 °C in a 5 % CO2 humidified environment, with the medium replaced every two days. Once the cells reached 70–80 % confluence, they were washed with phosphate-buffered saline (PBS), detached using a 0.25 % trypsin-ethylenediaminetetraacetic acid (EDTA) solution, and centrifuged at 1000 rpm for 5 min. After discarding the supernatant, the cells were resuspended in fresh medium and re-seeded for subsequent experiments. Cell morphology and potential signs of contamination were monitored periodically throughout the culture process.
For the culture of CT26 cancer cell spheres, CT26 cells were maintained in Thermo Scientific Nunc EasYFlasks. To initiate sphere culture, the cells were seeded in Gibco 1640 medium containing GlutaMAX, 10 % FBS, 1X minimum essential medium (MEM) non-essential amino acids, 100 U/mL penicillin-streptomycin, and 25 mM 4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES). The cells were seeded into Nunclon Sphera 6-well U-bottom plates at a density of 100 to 5000 cells per well (200 μL per well). After centrifuging at 250×g for 5 min, the plates were incubated at 37 °C in a 5 % CO2 atmosphere. The medium was changed every 72 h by removing 100 μL per well and replenishing it with fresh medium. Spheroid formation and growth were monitored using the Leica TCS SP8 Confocal Laser Scanning Microscope System.

5.4
Measurement of cellular uptake of nanoparticle
In this experiment, to enable the monitoring of cellular internalization through fluorescence microscopy, the positively charged rhodamine b dye was selected for its ability to physically adhere to negatively charged surfaces, thereby facilitating the labeling of GO/BCT:Mn. CT26 cells (1 × 10∧5) were seeded in confocal microscopy dishes and treated with GO/BCT:Mn tagged with rhodamine b dye (concentration of 50 μg/mL) for a period of 4 h. The uptake of these particles by Hepa 1–6 cells was then examined using a Thermo Scientific DXR3xi Raman Imaging Microscope. Subsequent to this, the lysosomes within the CT26 cells were stained using Lyso-Tracker Green. Following the washing steps to remove any unbound dye and nanoparticles, the CT26 cells were visualized and analyzed using a Leica confocal laser scanning microscope.

5.5
Assessment of antitumor effectiveness in vitro
The effectiveness of GO/BCT:Mn MDM against CT26 cells was determined using the conventional MTT assay. For this purpose, CT26 cells underwent treatment with varying levels of CVS nanoparticles, ranging from 0 to 1000 μg/ml, at incremental concentrations (6.25, 12.5, 25, 50, 100, 200, and 1000 μg/ml) over a 24-h period. The MTT assay's standard methodology was employed to evaluate cytotoxic effects. Additionally, the therapeutic impact of the GO/BCT:Mn MDM was examined in more detail. In brief, the treatments involved for CT26 cells were (i) untreated control, (ii) ultrasound-irradiated BTO (BTO + US), (iii) GO/BTO combined with ultrasound (GO/BTO + US), (iv) GO/BCT with ultrasound exposure (GO/BCT + US), (v) GO/BCT:Mn alone, and (vi) GO/BCT:Mn accompanied by ultrasound (GO/BCT:Mn + US). Each of these groups (i-vi) received nanoparticles at a concentration of 50 μg/ml. For the groups (ii), (iii), (iv), and (vi), ultrasound irradiation was applied at an intensity of 1.5 W/cm2 for a duration of 3 min.

5.6
NAD+/NADH assay
The following protocol details the steps involved in the extraction of NAD+ and NADH from cells and their subsequent analysis. To begin with, cells were treated with NAD+/NADH extraction buffer and gently pipetted for lysis. On the other hand, suspension cells underwent centrifugation at 600g for 5 min followed by the addition of NAD+/NADH extraction buffer and gentle pipetting to facilitate lysis. The lysing process was conducted at room temperature or on ice. The lysates were then centrifuged at 12,000 g at 4 °C for 5–10 min, and the supernatant was collected for further analysis. For the preparation of NADH standards, 5 mg of NADH provided in the kit was dissolved in 655 μl of NADH preparation solution to obtain a 10 mM NADH standard solution. The solution was aliquoted and stored at −80 °C away from light. A standard curve of NADH was established by diluting the 10 mM NADH standard with NAD+/NADH extraction buffer to obtain concentrations ranging from 0 to 10 μM. For each assay in a 96-well plate, 20 μl of these standards corresponding to 0, 25, 50, 100, 150, and 200 pmol of NADH were added per well. The zero concentration served as the blank control. The standards were prepared freshly before use due to the instability of NADH. For the preparation of the alcohol dehydrogenase working solution, the enzyme was diluted 45 times with the reaction buffer. Specifically, 2 μl of alcohol dehydrogenase was added to 88 μl of reaction buffer to obtain 90 μl of working solution. Each standard or sample assay required a fresh 90 μl of this working solution. To determine the levels of NAD+ and NADH or their ratio in the samples, 50–100 μl of the test sample was heated at 60 °C for 30 min in a water bath or PCR instrument to decompose NAD+. If insoluble material formed post-heating, the mixture was centrifuged at 10,000 g at room temperature or 4 °C for 5 min, and 20 μl of the supernatant was used for the assay.

5.7
ATP assay
For the assessment of ATP content in cell samples, our protocol began with the collection of cells in centrifuge tubes, followed by the removal of the supernatant. We used an extraction buffer to cell ratio of 500–1000 μL per 10∧4 cells, recommending 1 mL buffer for every 5 million cells. The cells underwent ultrasonic disruption for 1 min in an ice bath, at an intensity of 20 % or 200W, employing an ultrasonic cycle of 2 s on and 1 s off. This was followed by centrifugation at 10,000 g at 4 °C for 10 min. The supernatant was then carefully transferred to a fresh Eppendorf tube. To this, 500 μL chloroform was added, followed by rigorous shaking to ensure thorough mixing. A subsequent centrifugation at 10,000 g at 4 °C for 3 min separated the phases, and the aqueous layer was retained on ice for immediate analysis. In the assay procedure, initial absorbance (A1) at 340 nm was measured for 10 s after complete mixing of the sample. The cuvettes containing the reaction mixture were then incubated at 37 °C for mammalian cells or at 25 °C for other species, lasting 3 min. Post incubation, the cuvettes were promptly cleaned, and a second absorbance reading (A2) was taken at the 3-min and 10-s mark. The differential absorbance for the test (ΔA test) and standard (ΔA standard) was computed by subtracting A1 from A2. The ATP concentration was calculated on a per-cell basis using the formula: ATP concentration (μmol/10∧6 cells) = ΔA test ÷ (ΔA standard ÷ C standard) × V extraction ÷ 5 = 0.125 × ΔA test ÷ ΔA standard, providing a quantifiable measure of the ATP levels in the cell samples.

5.8
Malate assay
To determine the malate levels in cell populations, we utilized a protocol optimized for precise quantification. Cell suspensions were harmonized at a ratio of 5–10 μL of the first extraction buffer per million cells, with the optimal volume being 1 mL for a batch of 5 million cells. The cells were disrupted by sonication on ice, using a 300W power setting in a pulsing mode of 3 s on and 7 s pause, for a total duration of 3 min. The lysates were then centrifuged at 4 °C with a force of 12,000 g for 10 min, after which 0.8 mL of the clear supernatant was extracted. To this, we slowly added 0.15 mL of a second extraction buffer, mixing with care to avoid bubble formation, followed by an additional centrifugation at 4 °C and 12,000 g for 10 min to clarify the sample further. Dilution of a 100 μmol/mL stock standard solution was conducted with distilled water to achieve a series of standards ranging from 0 to 0.4 μmol/mL for the assay calibration. The absorbance of the resulting solutions was measured at 570 nm for various tubes—test (test), control (control), standard (standard), and blank (blank)—and the net absorbance change (ΔA) was computed for both the test (ΔA test = A test - A control) and standard (ΔA standard = A standard - A blank) samples. Malate content per million cells was calculated with the formula: malate content (μmol/10∧6 cells) = x × (Volume of the supernatant + Volume of Extraction Buffer II) ÷ (Cell count × Volume of the supernatant ÷ Volume of Extraction Buffer I) ÷ Cell count = 1.1875 × x ÷ Cell count.

5.9
Glucose assay
To quantify glucose levels within cultured cells, we initiated our procedure by removing the growth medium from each well on a 6-well plate. The cells were rinsed twice with phosphate-buffered saline (PBS) to eliminate any residual medium. Lysis of the cells was achieved by adding a proprietary or in-house prepared lysis buffer, with quantities adjusted to ensure optimal cell disruption. Pipetting was performed to mix the lysis buffer with the cell layer thoroughly, ensuring complete cell dissolution. Following lysis, the cell debris was pelleted by centrifugation at 12,000×g for a 5-min interval, after which the supernatant was carefully extracted for use in glucose analysis. For the calibration of glucose measurements, a high-concentration glucose stock solution of 200 mg/mL was serially diluted using distilled water, saline, or PBS to generate a range of standards suitable for plotting a glucose calibration curve, extending from 0 to 2000 mg/dL. To carry out the assay, we allocated 5 μL of either standard or test sample into an Eppendorf tube and introduced 185 μL of Glucose Assay Reagent, culminating in a final assay mixture volume of 190 μL. This mixture was homogenized by vortexing and then quickly centrifuged at 5000×g, ensuring the solution collected at the base of the tube. The samples were then subjected to a temperature of 95 °C for 8 min within a metallic heating block and subsequently cooled to 4 °C. Upon reaching the target low temperature, the Eppendorf tubes were retrieved, and 180 μL of the assay mixture from each tube was pipetted into a pristine 96-well plate. The absorbance of this final preparation was assessed at 630 nm to determine the glucose concentration.

5.10
Animal experiment
In this study, we utilized female Balb/C mice, aged 5 weeks, obtained from Beijing Huafukang Co. Ltd (Beijing, China). The experimental protocols involving these animals received approval from the ethical committee (approval number: TJUE-2024-526). The mice were accommodated in an environment ensuring freedom from specific pathogens, with controlled temperature settings of 25 ± 2 °C and humidity levels maintained at 55 ± 10 %. Additionally, they were subjected to a consistent cycle of 12 h of light followed by 12 h of darkness. To establish a colon cancer model, the right abdomen of each Balb/C mouse receives a subcutaneous injection of 2 × 10∧6 CT26 cells suspended in 100 μl of phosphate buffer solution (PBS).

5.11
Small animal live imaging
Initially, the nanomaterials were prepared through a series of cleansing and centrifugal steps to eliminate any contaminants. These cleaned nanomaterials were then combined with rhodamine dye. This mixture was maintained at ambient temperature for an extended period to facilitate the coupling reaction. After this reaction period, any unbound dye was meticulously removed through multiple PBS washes, followed by centrifugation, effectively isolating the rhodamine-tagged nanomaterials. These tagged nanomaterials were then administrated intravenously into the tail vein of tumor-bearing small animal models. 4 h subsequent to the injection, the animals underwent imaging with an in vivo imaging system designed for small animals. This was done to observe the in vivo distribution and targeting of the rhodamine-conjugated nanomaterials to the tumor sites. Following the in vivo imaging session, the animals were euthanized, and vital organs, including tumors, liver, and kidneys, were harvested for further examination. The harvested organs were then subjected to imaging using the same small animal live imaging system.

6
Immunofluorescence protocol for frozen tissue sections
Initially, the tissue was sectioned into slices of 10 μm thickness using a cryotome and mounted on glass slides. The slices were then fixed in a 4 % solution of paraformaldehyde for a duration of 10 min at ambient temperature. Subsequent to fixation, the slices were rinsed thrice in PBS, each rinse lasting 5 min. To prevent unspecific binding of antibodies, the slices were incubated for 1 h at room temperature in a PBS solution supplemented with 5 % bovine serum albumin (BSA). Afterwards, the slices were covered with primary antibodies targeting CD31, caspase-3, Ki-67, HIF-1α, and VEGF, which were diluted in the BSA-containing PBS solution, and left to incubate at 4 °C through the night. Post incubation, the slices underwent three PBS washes, each for 5 min. The slices were then overlaid with a fluorescent secondary antibody specific to the primary antibodies' host species and incubated for 1 h at room temperature, shielded from light. Following this, the slices were washed three times in PBS, each wash lasting 5 min. Finally, the slices were mounted using a mounting medium designed to inhibit fluorescence fading. Examination and imaging of the slices were performed using a fluorescence microscope.
6.1
Hematoxylin and eosin (H&E) staining procedure
Initially, tissues such as heart, liver, spleen, lungs, kidneys, and tumor are extracted from a mouse specimen. These samples are then subjected to fixation using glutaraldehyde, followed by a dehydration process. Post-dehydration, the samples undergo a clearing step with xylene and are subsequently embedded in paraffin wax. Using a microtome, the embedded tissues are sliced into sections with a thickness of 10 μm. These sections are then sequentially subjected to a dehydration protocol using ethanol solutions of varying concentrations (for instance, 50 %, 70 %, 80 %, 95 %, and 100 %), with each stage lasting approximately 3 min. The sections are rendered transparent using xylene or a similar substitute for about 3 min. The sections are immersed in a hematoxylin staining solution for a duration of 10 min. A brief wash in 95 % ethanol is performed to eliminate any superfluous hematoxylin. The sections are then stained with eosin for 2 min. To discard non-specific staining, the sections are dipped briefly (for around 5 s) in 1 % hydrochloric alcohol. A swift rinse in Scott's Tap Water, lasting about 5 s, is employed to intensify the blue hue of the nuclei. The sections are then re-dehydrated using 95 % and 100 % ethanol solutions, followed by a second clearing in xylene. Finally, the sections are mounted with a sealing agent such as dimethyl ether hydroxymethylbenzoate. The prepared slides are then examined and documented using a microscope.

6.2
Assessment of hematological and biochemical parameters post nanoparticle administration
The administration of GO/BCT:Mn malate depletion modulator (MDM) was conducted intravenously through the mouse tail vein in a sterile environment. At specific time points - day 1, day 7, and day 30 post-injection - blood specimens were procured either from the tail vein or the retro-orbital plexus. To inhibit coagulation, the collected blood was transferred into EDTA-containing tubes. A comprehensive blood count including metrics like the count of erythrocytes, leukocytes, and thrombocytes was carried out with an automated blood analyzer. Concurrently, at the same time intervals following the nanoparticle treatment, a separate blood sample was acquired for serum analysis. This sample was placed into coagulant-free tubes and subsequently centrifuged for serum extraction. The extracted serum underwent a detailed examination for various biochemical indices, encompassing hepatic markers (ALT, AST), renal markers (creatinine, blood urea nitrogen), and electrolyte profiles, utilizing a biochemical analyzer.

6.3
Statistical methods for analysis
The two groups' statistical significance was determined using the two-tailed Student's t-test without pairing. For comparisons involving more than two groups, we applied one-way ANOVA followed by Tukey's multiple comparison post hoc test. The mean and standard deviation (SD) represent the quantitative data. Asterisks indicated significance, where 'ns' denotes insignificant, ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001. The GraphPad Prism version 8.0.2 software facilitated all the statistical analysis processes.

CRediT authorship contribution statement
Run Yang: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Bowen Li: Writing – review & editing, Writing – original draft, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation. Yun Fu: Investigation. Chenxu Shang: Investigation, Writing – review & editing. Guoqing Feng: Investigation. Xuheng Chen: Investigation, Writing – review & editing. Huining He: Project administration. Zhengmian Zhang: Project administration. Yang Bai: Project administration. Bin Zheng: Writing – review & editing, Writing – original draft, Project administration, Funding acquisition.

Graphical illustrations
Graphical illustrations were made with Biorender.com.

Ethics approval and consent to participate
The experimental protocols involving these animals received approval from the ethical committee (approval number: TJUE-2024-526).

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.