Microenvironment-responsive tetrahedral DNA nanoplatform for specific and enhanced tumor therapy through differential ROS and angiogenesis modulation

Abstract

Achieving spatiotemporally differential regulation of therapeutic activity between tumors and normal tissues remaines a paramount challenge in precision nanomedicine. To address this challenge, an intelligent nanostructure is engineered by assembling the DNAzyme-integrated tetrahedral DNA (TDN) and pyropheophorbide-a (PPa) on MnO nanoparticles (NPs), which is abbreviated as TDN-MPs. Specifically, TDN-MPs generate reactive oxygen species (ROS) under laser irradiation and exhibit excellent photodynamic therapy (PDT) activity to kill tumor cells. The Mn²⁺ ions released from TDN-MPs triggered by tumor microenvironment (TME) alter the geometric structure of TDN and release the activated DNAzyme for cleaving vascular endothelial growth factor receptor 2 (VEGFR2) mRNA, which inhibits the formation of new blood vessels at tumor site. Conversely, the structurally complete TDN-MPs in physiological environment not only maintain DNAzyme in a silent state but also act antioxidants to eliminate ROS, which enables the goal of reducing toxic effects of DNAzyme and ROS toward normal sites. This strategy takes advantage of the geometric variability of TDN to differentially regulate ROS and VEGFR2 levels between tumor and normal tissues, significantly enhancing the therapeutic effect while minimizing systemic toxicity. This study presents a robust approach for achieving high-specificity tumor therapy through the rational design of intelligent DNA-based nanostructures.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04408-z.

Introduction

Traditional anticancer drugs simultaneously lead to equivalent drug activity of tumor and normal tissue due to the indiscriminate cytotoxicity. These disadvantages result in severe systemic toxicity and compromised therapeutic outcomes to limit the application in clinic. Recently, activable nanomedicine, which exhibits activated therapeutic functions in response to either external stimuli or the tumor microenvironment (TME), has been developed to address this problem [1, 2]. In contrast to traditional anticancer drugs, these advanced nanoplatforms differentially regulate the level of drugs, reactive oxygen species (ROS), mRNA, and specific protein, suppressing therapeutic activity in healthy tissues while amplifying tumor-specific responses [3–5]. More importantly, the traditional anticancer drugs were found to induce systemic protein dysregulation, which often manifests as severe adverse effects and ultimately compromises treatment safety and long-term patient prognosis [6]. For instance, photodynamic therapy (PDT) generates singlet oxygen (¹O₂) and consumes intratumoral O₂, which promotes the compensatory upregulation of vascular endothelial growth factor (VEGF) in tumor cells [7–9]. VEGF family bind with their receptors to stimulates pathological angiogenesis and promote the survival and proliferation of tumor cells. Generally, VEGF exerts its angiogenic effects by activating vascular endothelial growth factor receptor 2 (VEGFR2) expressed on endothelial cells through its dominate isoforms, VEGFa [10, 11]. Reducing the expression of VEGFR2 in tumor have been discovered as an effective strategy to inhibit the angiogenesis and tumor growth. However, traditional anti-angiogenesis agents result in the insufficient nutrients and O₂ supply in normal tissue due to unspecific VEGFR2 level downregulation and anti-angiogenesis.

DNAzyme, as a catalytic nucleic acid with programmable target specificity, not only driven transformative advances in disease therapy through sequent-selective mRNA cleavage but also inhibits protein levels at the source by intercepting pathological mRNA transcripts prior to ribosomal translation [12–14]. Nevertheless, the DNAzyme-based gene therapy is still hampered by insufficient spatiotemporal regulation due to the nonspecificity of catalytic systems and biodistribution [15]. Specifically, the amount of endogenous divalent metal ions was too low to activate DNAzyme, while exogenous divalent metal ions may produce non-specific activation during transport before reaching the disease site. These unexpected disadvantages result in unexpected gene silencing in non-disease sites. Besides, the accidental release of DNAzyme during its circulation in vivo lead to irreversible toxic effects on non-target cells. Therefore, developing an effective strategy to ensure a spatiotemporal regulation of the activity of DNAzyme is desirable to achieve precise and efficient regulation of mRNA and protein levels at the lesion.

Tetrahedral DNA nanostructures (TDN) fold DNA strands into three-dimensional nanostructures with relatively long-term structural stability, which is an important feature to lengthen the intracellular structural integrity of DNAzyme in physiological environment [16, 17]. In addition, the TDN backbone has been discovered to exhibit unique antioxidant activities, which could efficiently scavenge ROS to reduce inflammation. Previous studies have indicated that TDN could upregulate antioxidant enzymes to neutralize ROS, while the guanine (G) bases with low oxidation potential could also adsorb or scavenge a portion of ROS [18–20]. Development of geometrically variable TDN based platform that responds to TME is a desirable strategy to specifically regulate the VEGFR2 and ROS levels in tumor and normal tissue, which is crucial important to effective tumor therapy. Herein, we engineered the VEGF aptamer and DNAzyme-integrated TDN onto MnO nanoparticle (NPs) loaded with pyropheophorbide-a (PPa, a typical photodynamic therapy agent) to obtain the intelligent platform TDN-MPs. The intelligent TDN-MPs could differentially regulate the ROS and VEGFR2 levels between tumor and normal tissues to achieve specific and enhanced tumor therapy. In tumor tissue, TDN-MPs generated ROS under laser irradiation to kill tumor cells with highly efficient. Besides, the Mn²⁺ ions released from MnO NPs acted as the trigger to alter the geometric structure of TDN and release the activated DNAzyme, which cleaved the VEGFR2 mRNA and effectively inhibiting the formation of new blood vessels at the tumor site. On contrary, DNAzyme maintained in the silent state at normal site, keeping the VEGFR2 mRNA level in normal levels and limit its anti-angiogenesis effect. Meanwhile, the complete TDN acts an antioxidant to eliminate the ROS generated by PPa, reducing the toxic effect of ROS and protecting normal cells. Compared with conventional drug delivery carriers, this ingeniously designed DNA nanostructure not only enhances the in vivo stability of DNAzyme delivery and reduces toxicity associated with off-target effects, but also innovatively utilizes DNAzyme as a switch to regulate the conformational and functional changes of the TDN through its activity modulation. This advanced nanoplatforms effectively treat tumor through differentially regulating the ROS and angiogenesis levels in normal and tumor tissue, offering a new avenue for specific and efficient tumor therapy.

Results and discussion

Synthesis and characterization of MPs, TDN and TDN-MPs

The construction of the TDN-MPs involved a multistep process as shown in Scheme 1A. PPa loaded MnO nanoparticles (MPs) were chose as the photosensitizer and Mn²⁺ ions donor, which acts as trigger to activate DNAzyme and generate ROS in tumor. Transmission electron microscopy (TEM) image indicated that the prepared MPs showed high uniformity morphology with diameters of approximately 30 nm (Fig. 1A). Atomic force microscopy (AFM) image indicated that MPs showed a characteristic topographic height of 4.8 nm (Fig. 1B). X-ray powder diffraction (XRD) patterns illustrated that the crystallographic structure of MPs was matched well with that of MnO (JCPDS no. 001-075-0275) (Figure S1). X-ray photoelectron spectroscopy (XPS) analyses indicated that MPs showed two distinct peaks at 641.67 eV and 653.48 eV, which could be attributed to Mn 2p_3/2 and 2p_1/2 peaks (Figure S2). These results confirmed that the Mn element exist as + 2 in MPs and could act as Mn²⁺ ions donor. The smooth M-H curve of MPs displayed that the magnetic moment increased with the applied magnetic field, which revealed their paramagnetic characteristics (Figure S3). To obtain DNAzyme-integrated TDN, we chose four single strands DNA (ssDNA) with VEGF-specific DNA aptamer to construct self-assemblied TDN by blocking DNAzyme into the edges. Polyacrylamide gel electrophoresis (PAGE) analyses diaplayed that the assembled TDN showed significantly higher molecular weight than all ssDNA (Fig. 1C). AFM images revealed uniformly distributed nanostructures with a topographic height of 1.8 nm (Fig. 1D), which was consistent with the theoretical height predicted for TDN based on Watson-Crick base-pairing geometry [21]. These results confirmed the successful assembly of TDN via base complementary pairing. Consequently, the obtained TDN was immobilized on MPs surface through the amidation reaction between amino group and carboxyl group. TEM image indicated that TDN-MPs showed almost indistinguishable morphology and size from that of MPs (Fig. 1E). It should note that it is difficult to directly observe DNA on MPs in TEM image due to the low atomic number (Z) composition and inherent limitations in electron scattering contrast of DNA [22]. AFM analyses clearly indicated that the topographic heights of TDN-MPs were obviously higher than MPs and TDN (Fig. 1F). Energy dispersive X-ray (EDX) mapping confirmed the presence of N and P elements, originating from the DNA constituent, in the TDN-MPs, which were absent in the control MPs (Figure S4). In addition, dynamic light scattering (DLS) and zeta potential measurement were conducted to investigate the hydrodynamic diameter and surface charges of MPs, TDN and TDN-MPs. It appears that the hydrodynamic diameter of TDN-MPs was 90 nm, which was larger than that of MPs (77 nm) (Fig. 1G). Since TDN show less negative zeta potential (-17 mV) compared to MPs (-22 mV), the zeta potential of TDN-MPs increased to -18 Mv (Fig. 1H). Besides, we observed characteristic absorption peaks of PPa (409 nm, 670 nm) and TDN (268 nm) in the UV-Vis spectra of TDN-MPs (Fig. 1I). These results clearly demonstrated the success of the bottom-up self-assembly of TDN-MPs. To further evaluate the loading capacity of TDN in TDN-MPs, we calculated the loading ratio of TDN-MPs based on UV-Vis spectra analyses. It appears that the loading ratio is as high as 78% based on the standard curve for different concentrations of TDN (Figure S5 and S6). Notably, both hydrodynamic diameter and UV-Vis spectra of TDN-MPs maintained in the similar level during 7 days incubation in serum (Figures S7 and 8). Moreover, polydispersity index (PDI) analysis showed that the TDN-MPs still maintained excellent dispersibility after being stored for 14 days (Figure S9). This robust stability feature highlights a key advantage of TDN-MPs over traditional carriers, which is the capability to ensure the stable encapsulation and delivery of DNAzymes in physiological environment for achieving high therapeutic efficacy and reducing off-target effects.

Scheme 1.

Schematic illustration of (A) the synthesis of TDN-MPs and (B) the spatiotemporally differential regulation of ROS and VEGFR2 levels for tumors therapy

Fig. 1.

Synthesis and characterization of TDN, MPs, TDN-MPs. (A) TEM and (B) AFM images of MPs. (C) PAGE image and (D) AFM images of TDN. (E) TEM and (F) AFM images of TDN-MPs. (G) DLS analyses of MPs and TDN-MPs in aqueous. (H) Zeta potential of MPs, TDN and TDN-MPs. (I) UV-vis spectra of MnO, MPs, TDN and TDN-MPs

The regulatory effects of TDN-MPs on VEGF mRNA and ROS under different pH values

The Mn²⁺ ions-mediated activation of DNAzyme and cleavage of VEGFR2 mRNA were systematically investigated. PAGE images showed that the intensity of VEGFR2 mRNA bands gradually reduced with the increase of Mn²⁺ ions concentration in the presence of DNAzyme (Fig. 2A). The results suggested that DNAzyme activity was positively correlated with Mn²⁺ ions concentration and could successfully cleave VEGFR2 mRNA. Fluorescence analysis was used to investigate the effect of Mn²⁺ ions concentration on the release of DNAzyme from TDN. We labeled the fluorophore (Cy5) and quencher (BHQ2) at 3’ and 5’ ends of DNAzyme, respectively. Theoretically, the loading of DNAzyme on TDN shortened the distance between Cy5 and BHQ2, which resulted in the static quenching through fluorescence resonance energy transfer (FRET) [23, 24]. As expected, the fluorescence intensity of DNAzyme-integrated TDN was obviously lower than that of single-stranded DNAzyme (Fig. 2B). In the presence of Mn²⁺ ions, the fluorescence intensity of TDN was recovered attributed to the cleavage activity of DNAzyme was activated to dissociate the intact TDN and release single-stranded DNAzyme, indicating the successful loading of DNAzyme on TDN and the Mn²⁺ ions mediate DNAzyme release. To visualize the disassembly process of TDN more directly, we characterized TDN structures using AFM after incubation with VEGF and VEGF + Mn²⁺ ions (Figure S10). AFM images revealed distinct morphological changes and a reduction in height after incubation with VEGF, consistent with the partial disassembly of TDN initiated by the specific binding of VEGF to its integrated aptamer. The dramatic decreased in height was observed when TDN was incubated with both VEGF and Mn²⁺ ions, indicating that the coexistence of the target and the auxiliary factor enabled the maximum structural decomposition of TDN. To evaluate concentration-dependent effect of Mn²⁺ ions on DNAzyme release from TDNs, we monitored fluorescence recovery of S-DNAzyme (with a corresponding TDN concentration of 0.25 µM) across a gradient of Mn²⁺ ions concentrations (Figure S11). Fluorescence recovery initiated at a threshold Mn²⁺ ions concentration of 0.0625 mM. Saturation kinetics were observed at 0.125 mM Mn²⁺ ions, achieving near-maximal fluorescence recovery (47.8%). The Mn²⁺ ions-mediated DNAzyme activation was further supported by analyzing the VEGFR2 mRNA level in the presence of TDN and Mn²⁺ ions. We noted a significant reduction on the intensity of VEGFR2 mRNA band upon Mn²⁺ ions supplementation, implying that Mn²⁺ ions-mediated activation of DNAzyme for structural disassembly of TDN and targeted cleavage of VEGFR2 mRNA (Fig. 2C). In order to verify the pH response capacity of TDN-MPs, we investigated the Mn²⁺ ions release kinetics of TDN-MPs at different pH values. It appears that about 40% and 77% Mn²⁺ ions released from TDN-MPs within 12 h at pH 6.5 and pH 5.5, which was significantly higher than that in physiological environment (Fig. 2D). In addition, the Mn²⁺ ions release process rapidly reached a plateau, indicating the rapid response of TDN-MPs at the tumor site. To determine whether the Mn²⁺ ions loaded in TDN-MPs was sufficient to activate the encapsulated DNAzyme, we performed an in vitro simulation under different pH conditions. The fluorescence intensity increased with DNAzyme release in acidic environments, indicating that the Mn²⁺ ions released from TDN-MPs was sufficient to activate DNAzyme (Figure S12).To investigate whether TDN-MPs showed acidic environment dependent DNAzyme activity, we exposed TDN-MPs in aqueous containing VEGFR2 mRNA with different pH values. PAGE image indicated evident VEGFR2 mRNA band at pH 7.4, which was comparable to the control sample. This result suggested that TDN-MPs would not affect the VEGFR2 mRNA in physiological environment, which was attributed to the intact TDN-MPs structure maintaining DNAzyme in an inactive conformation. Conversely, negligible VEGFR2 mRNA band could be noticed when the pH value decreased to 6.5 and 5.5 due to the complete VEGFR2 mRNA cleavage by TDN-MPs, demonstrating that TDN-MPs successfully released the DNAzyme and decreased the level of VEGFR2 mRNA in acidic environments (Fig. 2E). These results confirm that through ingenious design, the activity of the DNAzyme is directly coupled to TDN conformational changes, enabling the DNAzyme to switch from an inactive state under physiological conditions to an activated state in the acidic TME. To further evaluate the specific cleavage ability of the DNAzyme released from TDN-MPs, we investigated the effects of TDN-MPs at different pH conditions on EGR-1 and survivin mRNA. The PAGE images showed that both EGR-1 and survivin mRNA exhibited distinct bands under different pH conditions, indicating that the DNAzyme possessed excellent cleavage specificity (Figure S13).

Fig. 2.

The regulatory effects of TDN-MPs on VEGF mRNA and ROS under different pH values. (A) PAGE image of the effect of Mn²⁺ ions concentration on DNAzyme cleaving function. (B) Fluorescence analysis of S-DNAzyme, TDN, TDN + Mn²⁺ ions. (C) PAGE image of the TDN on DNAzyme cleaving function in the presence of Mn²⁺ ions. (D) The Mn²⁺ ions release profiles of TDN-MPs at different pH value. (E) PAGE image of the TDN-MPs on DNAzyme cleaving function under different pH conditions. (F) UV-vis spectra of the ROS regulatory efficacy of TDN-MPs at pH 6.5. (G) EPR spectra analyses of ¹O₂, ·O₂- and ·OH levels in in the solution with different treatments. (H) T₁-weighted MR images and (I) the corresponding linear fitting of TDN-MPs at different pH values

Subsequently, the ROS generation/scavenging capacity of TDN-MPs in acidic and physiological environment were assessed. Since ¹O₂ is the main product of PPa under laser irradiation, we chose MPs with the same concentration as control and analyzed ¹O₂ production using 1,3-diphenylisobenzofuran (DPBF) as probe [25]. We noticed the significantly decrease in absorption peak of DPBF in the presence of TDN-MPs with laser irradiation (660 nm, 50 mW/cm², 5 min) at pH 6.5, indicating the successful ¹O₂ generation by TDN-MPs in acidic environment (Fig. 2F). It should note that the absorption peak of DPBF in the presence of TDN-MPs was similar to that of MPs, demonstrating that the TDN showed negligible effect on the ¹O₂ generation of MPs in acidic environment. Interestingly, MPs remarkably decrease the absorption peak of DPBF with laser irradiation at pH 7.4, while no significant decrease in the DPBF absorption was noted in TDN-MPs treatment at the same condition (Figure S15). This result indicates that TDN-MPs could eliminate ¹O₂ in the physiological environment because TDN maintains the integrity of the structure. We further investigated its ROS scavenging activity by adding MPs, TDN, and TDN-MPs into methylene blue (MB)-stained FeSO₄/H₂O₂ system, which was widely used to generate ·OH. The obvious decrease in the absorbance of MB was noticed after mixing MB with FeSO₄/H₂O₂, clearly indicating the generation of ·OH [26]. Notably, we observed remarkable increase in absorbance of MB after adding TDN and TDN-MPs, while negligible change could be noted after adding MPs (Figure S16). These phenomenon confirmed that the inherent ROS-scavenging properties of structurally intact TDN in physiological environments could effectively eliminate the ROS generated by PPa under laser irradiation. To further investigate the ROS-scavenging capacity of TDN, electron paramagnetic resonance (EPR) spectra was applied to quantify the levels of ¹O₂, ·O₂- and ·OH under different treatment conditions. Consistent with the analyses observed using DPBF and MB as molecular probes, the EPR signals corresponding to ¹O₂, ·O₂- and ·OH decreased obviously after the addition of TDN (Fig. 2G). Interestingly, the TDN-MPs at equivalent concentrations demonstrated markedly enhanced ROS-scavenging efficacy compared to free TDN. This amplification is attributed to the nanoconfinement effect inherent to the spherical nucleic acids, which increased the local concentration of TDN and consequently augments its antioxidant activity [27]. Since Mn²⁺ ions has been considered as a typical T₁ contrast agent, we assessed the T₁ contrast capacity of TDN-MPs at different pH values [28]. As expected, the T₁ signal of TDN-MPs in a weakly acidic environment (pH 6.5) was significantly higher than that in a neutral environment, indicating that TDN-MPs possessed a pH-responsive activable MRI contrast agent capability (Fig. 2H and I). Further quantitative analyses indicated that the r₁ value of TDN-MPs was 2.3 mM⁻¹s^− 1 at pH 6.5, which was remarkably higher than that in neutral environment (0.6 mM⁻¹s^− 1). To validate the TME-responsive MRI capability of TDN-MPs at the cellular level, 4T1 and Hs 578Bst cells were incubated with MPs and TDN-MPs for varying durations (Figure S17). Comparative MRI of cells treated with TDN-MPs versus control MPs showed a specific signal enhancement, verifying that MRI could be achieved by Mn²⁺ ions released in response to the TME.

Tumor cytotoxicity and normal cytoprotection of TDN-MPs

In order to reinforce the tumor enrichment ratios of the TDN-MPs to achieve precisely gene cleavage, we investigated the tumor target capacity of TDN-MPs through the VEGF-specific DNA aptamer on TDN [29, 30]. To access the specifically target to tumor of TDN-MPs through binding to overexpressed VEGF on tumor cell membranes, we explored the cellular uptake and targeting ability of TDN-MPs to tumor cells (Scheme 1A). Fluorescence imaging and quantification indicated time-dependent accumulation of TDN-MPs in tumor cells, which reached near-saturation within 2 h, demonstrating effective internalization via aptamer-directed endocytosis (Fig. 3A and S18-S19). To verify the target capacity of TDN-MPs to VEGF overexpressed cells, we synthesized aptamer-free TDN-MPs as control to perform competitive binding assay. We observed that 4T1 cells treated by TDN-MPs for 2 h showed the evident red fluorescence from PPa, which was significantly stronger than those treated by aptamer-free TDN-MPs (Fig. 3B). Additionally, pre-incubation of 4T1 cell with free VEGF aptamer also resulted in a noticeable reduction of red fluorescence. Consistent with the fluorescence analyses, ICP-MS analyses indicated that the uptake amount of Mn²⁺ ions in 4T1 cells without free VEGF aptamer incubation was higher than that with free VEGF aptamer incubation (Figs. 3C and D). These results clearly indicated that the specific cellular binding of TDN-MPs to 4T1 cells was mediated by the interaction between VEGF aptamer and overexpressed VEGF protein on the membrane surface, effectively increasing the uptake amount of TDN-MPs by tumor cells. The high biocompatibility is essential for the biomedical application of TDN-MPs. To examine the cytotoxicity, the effects of MPs and TDN-MPs on the proliferation of 4T1 and Hs 578Bst cells were evaluated using the cell counting kit-8 (CCK8) assay. We observed that TDN-MPs exhibited slightly stronger cytotoxicity against 4T1 cells compared to MPs without laser irradiation, revealing that the TME-responsive release of DNAzyme mediated a gene therapeutic effect capable of inhibiting partial tumor cell proliferation (Fig. 3E). Interestingly, TDN-MPs significantly inhibited the growth of 4T1 cells under laser irradiation (660 nm, 50 mW/cm², 30s). The viability of 4T1 cells treated by TDN-MPs was about 18% when the concentration increased to 2 µg/mL, which was significantly lower than that of cells treated by MPs (59%). These results clearly demonstrated that under identical irradiation conditions, TDN-MPs exhibited a 41% greater antitumor efficacy than MPs, which was attributed to the activated DNAzyme-mediated gene therapy. More importantly, we did not observe obvious cytotoxicity of TDN-MPs to Hs 578Bst cells under laser irradiation (Fig. 3F). However, the MPs (2 µg/mL) results in almost 70% Hs 578Bst cell death under the same condition. These results demonstrated the certain cytoprotection effect of TDN-MPs on normal cells during treatment, which was attributed to the structural integrity preservation of TDN as an antioxidant in normal physiological microenvironments for protecting healthy cells from ROS damage. Notably, TDN showed negligible effect on the viability of 4T1 and Hs 578Bst cells, indicating that TDN could not achieve DNAzyme mediated gene therapy through release DNAzyme without the assistant of Mn²⁺ ions (Fig. 3G). Consistent with CCK-8 assessment, Calcein acetoxymethyl ester/propidium iodide (Calcein-AM/PI) double-staining analyses indicated that TDN-MPs were more effective than MPs in killing cancer cells under laser irradiation, which was due to the TME-triggered DNAzyme release and subsequent gene silencing of VEGFR2 mRNA for enhancing tumor efficacy (Fig. 3H). Whereas, the killing effect of laser-treated TDN-MPs on normal cells was significantly reduced, providing that TDN-MPS not only possessed excellent biocompatibility, but also could obviously reduce the side effects of light irradiation on normal tissues during treatment (Fig. 3I). Besides, TDN induced weak tumor cell death compared to negligible effects in normal cells, possibly because the rapid proliferation of tumor cells requires a large number of metal ions (such as Mn²⁺ ions and Fe²⁺ ions) as cofactors or signaling molecules, leading to partial leakage of DNAzyme (Fig. 3J) [31, 32]. These results clearly proved that TDN-MPs specifically induce the death of tumor cells while minimize off-target toxicity to normal cells, which was attributed to the ingenious design that utilized DNAzyme activity to regulate TDN conformational changes.

Fig. 3.

Tumor cytotoxicity and normal cytoprotection of TDN-MPs. (A) Fluorescence images for cellular uptake of TDN-MPs under different incubation time conditions. (B) Fluorescence images depicting the uptake of TDN-MPs with or without VEGF aptamers after incubation of 4T1 cells for 2 h. (C) Fluorescence images and (D) ICP-MS of Mn depicting the uptake of TDN-MPs in the presence or absence of free VEGF aptamer. Cell viability of (E) 4T1 cells and (F) Hs 578Bst cells treated with different treatments. (G) Cell viability of TDN at different concentrations in 4T1 cells and Hs 578Bst cells. Live/dead cell staining assays analysis of (H) 4T1 and (I) Hs 578Bst cells with different treatments. (J) Live/dead cell staining assays analysis of TDN in 4T1 cells and Hs 578Bst cells. Red, dead cells; green, live cells

Spatiotemporally differential regulation of ROS and VEGFR2 in Vitro

To verify whether the high cytotoxicity to tumor cells while low cytotoxicity to normal cells was dependent on the differential ROS and VEGFR2 regulation capacity of TDN-MPs (Fig. 4A), we evaluated the effect of TDN-MPs on cellular ROS and VEGFR2 level. DCFH-DA was used as the ROS fluorescent probe and SOSG as the ¹O₂ specific probe to detect the cellular ROS and ¹O₂ levels. Consistent with the CCK-8 analyses, 4T1 cells treated by TDN-MPs under laser irradiation showed evident ROS and ¹O₂ production, which was similar to that treated by MPs with laser irradiation (Fig. 4B and S21). Moreover, there was no significant green fluorescence could be noted in the Hs 578Bst cells treated by TDN-MPs after laser irradiation, suggesting that the levels of ROS and ¹O₂ in the normal cells treated with TDN-MPs were relatively lower than those in the MPs group (Fig. 4C and S22). These phenomenon could be attributed to the ROS scavenging capacity of TDN in normal cells. To further evaluate the ROS scavenging capacity of TDN-MPs, we chose Rosup as agent to increase the ROS level in cells. We noted significant green fluorescence in the 4T1 and Hs 578Bst cells treated by Rosup, indicating the successful increase of ROS level. It could be observed that the green fluorescence decreased significantly with the increase of TDN concentration. When the TDN concentration reached 125 nM, most of the ROS in the cells were eliminated (Figure S23). Notably, the fluorescence intensity was obviously reduced in Hs 578Bst cells while maintained at the high level in 4T1 cells in the present of TDN-MPs under laser irradiation (Figure S24). These results demonstrated that the geometrically variable TDN-MPs act as ROS scavengers through the ROS scavenging activity of TDN in normal cells for providing protective effects, while having no significant impact on ROS generation in tumor cells. Additionally, mitochondria as the energy supplier were sensitive to ROS attack and indicated a decreased membrane potential in response to ROS damage [33]. Therefore, mitochondrial membrane potential changes were detected by JC-1 to assess the mitochondrial dysfunction caused by TDN-MPs. As expected, the strongest green fluorescence was observed in the 4T1 cells treated by TDN-MPs with laser irradiation, indicating that TDN-MPs treatment with laser irradiation significantly decreased mitochondrial membrane potential and induced early apoptosis in a large number of tumor cells (Fig. 4D). However, the mitochondrial damage induced by TDN-MPs with laser irradiation was significantly limited compared with irradiation-treated MPs in Hs 578Bst cells, demonstrating that TDN-MPs was capable of protecting the mitochondrial function of normal cells from damage during treatment (Fig. 4E).

Fig. 4.

Spatiotemporally differential regulation of ROS and VEGFR2 in vitro. (A) Schematic representation of ROS and VEGFR2 mRNA regulation by TDN-MPs in tumor cells and normal cells. Fluorescence microscope observation of (B) 4T1 and (C) Hs 578Bst cells after different treatments to detect intracellular ROS and ¹O₂ levels by DCFH-DA and SOSG probes. Fluorescence images of (D) 4T1 cells and (E) Hs 578Bst cells after different treatments to detect mitochondrial membrane potential measurement by JC-1 probes. (F) qRT-PCR analysis of VEGFR2 mRNA treated with different samples in 4T1 and Hs 578Bst cells. (G) qRT-PCR analysis of EGR-1 and survivin mRNA treated with TDN-MPs in 4T1 cells. (H) ELISA, (I) immunofluorescence micrographsand and (J) western blot analysis of VEGFR2 protein expression treated with different samples in 4T1 and Hs 578Bst cells. Data are present as mean ± SD, *p < 0.05, **p < 0.01, and ***p < 0.001, ****p < 0.0001, compared to the Saline group

To confirm tumor-specific activity of TDN-MPs on cleaving VEGFR2 mRNA in tumor cells, we analyzed the mRNA and protein expression of VEGFR2 in 4T1 and Hs 578Bst cells. Quantitative real-time polymerase chain reaction (qRT-PCR) analyses showed that VEGFR2 mRNA expression in 4T1 cells treated by TDN-MPs was significantly reduced than that treated by MPs, owing to the release and activation of DNAzyme in response to TME for cleaving VEGFR2 mRNA. In contrast, there were no obvious changes of VEGFR2 mRNA level in Hs 578Bst cells treated by TDN-MPs because TDN-MPs maintained structural integrity in physiological environment resulting in silencing of DNAzyme function (Fig. 4F). To further assess the specificity of the DNAzyme released from TDN-MPs, we analyzed the mRNA expression levels of potential off-target genes (EGR-1 and survivin mRNA) in treated 4T1 cells. qRT-PCR analyses exhibited that TDN-MPs in 4T1 cells had no significant alteration in EGR-1 or survivin mRNA levels compared to controls (Fig. 4G), which indicated the absence of cleavage activity by DNAzyme for nonspecific mRNAs. This is because two flanked substrate binding arms with specific sequence of DNAzyme bind specifically to the VEGFR2 mRNA sequence through Watson–Crick base-pairing, thereby reducing off-target effects [34]. Enzyme-linked immunosorbent assay (ELISA) results exhibited that TDN-MPs treatment significantly reduced the expression of VEGFR2 protein, which attested the effectiveness of the proposed DNAzyme in the regulation of protein expression in tumor cells (Fig. 4H). It should note that laser irradiation resulted in the generation of ROS, which partially promoted the compensatory upregulation of VEGF in tumor cells by consuming oxygen during the treatment and reduced the downregulation effect of TDN-MPs on VEGFR2. Contrary to the inhibition effect on VEGFR2 in tumor cells, TDN-MPs treatment induced a slight but statistically significant elevation of VEGFR2 protein expression (P < 0.01) in Hs 578Bst cells, which may arise from compensatory mechanisms triggered by antioxidant intervention of TDN [35]. Similar results were also observed for the regulation of VEGFR2 protein levels in western blot (WB) technique and immunofluorescence staining assay (Figs. 4I-J and S27). These results confirmed that TDN-MPs could differentially regulate the mRNA and protein expression levels of VEGFR2 in healthy and tumor tissue. Based on the key role of VEGFR2 in tumor metastasis, we evaluated the anti-migratory effect of TDN-MPs via a scratch assay. The results demonstrated that TDN-MPs strongly inhibited tumor cell migration (Figure S28).

Imaging performance and therapeutic efficacy of TDN-MPs in Vivo

In vitro hemolysis tests revealed negligible hemolysis for both MPs and TDN-MPs (Figure S29). The main biochemistry indicators alanine aminotransferase (ALT), aspartate aminotransferase creatinine (AST), creatinine (CREA) and UREA were maintained at the normol levels in TDN-MPS-treated (4 mg/kg) mice (Figure S30). Further histological assessment analyses are highly consistent with the biochemistry indices. There was no obvious necrosis, in-flammation and pathological lesion in the main organs of mice treated by TDN-MPs (Figure S31). The comprehensive hematological analysis demonstrated that all key blood parameters, including Lymph%, Mon%,Gran%, red blood cell (RBC), hemoglobin (HGB), and platelet (PLT) counts, etc., remained within normal physiological ranges (Table S2). These results revealed that TDN-MPs with high biocompatibility are promising candidates for tumor therapy. To evaluated the capability of TDN-MPs as a MRI contrast agent with TME-responsive release of Mn²⁺ ions for tumor diagnosis, T₁ MRI was performed in tumor-bearing mice after intravenous injection of TDN-MPs. T₁-weighted MRI presented that the T₁ signal within the tumor area increased with time and reached maximum at 2 h post injection, which clearly highlighted the contrast between the tumor and normal tissue (Fig. 5A). Notably, the TDN-MPs-treated group displayed a stronger contrast than the MPs-treated group, which was attributed to the enhanced uptake ratios of the TDN-MPs at the tumor site by the VEGF aptamer. Furthermore, we explored the tumor accumulation of TDN-MPs in vivo by intravenous administration of DiD-labeled MPs and TDN-MPs (4 mg/kg) to 4T1-breast-tumor-bearing mice. Both DiD-labeled MPs and TDN-MPs gradually accumulated in the tumor tissue while the tumor enrichment ratios of TDN-MPs were significantly higher (Fig. 5B and D). Besides, the intratumoral fluorescence intensity of DiD-labeled MPs and TDN-MPs treated mice could be detected even at 24 h post-injection. It should note that the fluorescence intensity of normal tissues in DiD-labeled TDN-MPs treated group was obviously lower than that in DiD-labeled MPs treated group (Fig. 5C and E). These results suggested that TDN-MPs not only possessed high enrichment ratios and prolonged retention properties at tumor site, but also had a low uptake ratio at normal site, clearly showing the potential to maximize the therapeutic effect while minimizing the toxic damage in normal tissues. The excellent tumor accumulation of TDN-MPs prompted us to further explore the anticancer effect in vivo on a subcutaneous tumor model of 4T1 tumor-bearing female BALB/c mice (Fig. 5F). When the tumor reached to 60–100 mm³, different drug formulations were intravenously injected with an equivalent concentration of Mn²⁺ ions (4 mg/kg) every 7 days for two times. The mice were randomly divided into six groups: (1) saline; (2) MPs; (3) TDN-MPs; (4) MPs + L; (5) TDN-MPs + L. The tumor growth curves and tumor photos exhibited that the treatment with MPs had no obvious effect on tumor growth. The treatment of MPs under laser irradiation (660 nm, 200 mW/cm², 250 s) and TDN-MPs partially inhibited tumor growth due to the ROS-mediated PDT and the DNAzyme-mediated gene therapy. Notably, tumor growth of mice treated with TDN-MPs under laser irradiation (660 nm, 200 mW/cm², 250 s) displayed the best antitumor efficacy during the 14-day observation period, confirming the augmented therapeutic efficacy of PDT combined with gene therapy by DNAzyme for VEGFR2 mRNA cleavage (Figs. 5G and H). To further verify the synergistic antitumor effect between PDT and gene therapy, we performed normalized comparisons of VEGFR2 mRNA expression levels and antitumor growth curves (Figure S37). While MPs + L (PDT alone) showed negligible effect on VEGFR2 mRNA and inhibited tumor growth by 54%, the combination TDN-MPs + L (PDT+gene therapy) significantly knocked down VEGFR2 mRNA and achieved a vastly superior 91% inhibition. We speculated that this enhanced efficacy arised because the gene therapy component suppresses both the compensatory angiogenesis triggered by PDT and the intrinsically high levels of pathological angiogenesis in the TME, thereby amplifying the overall therapeutic outcome. Besides, we evaluated the in vivo biosafety, VEGFR2 regulatory capacity, and antitumor efficacy of the isolated TDN (Figure S38). The results showed that TDN exhibited excellent biocompatibility and induced no change in VEGFR2 expression or antitumor effect in the absence of exogenous Mn²⁺ ions activation. This result demonstrated that the enhanced efficacy of TDN-MPs + L was directly attributable to the DNAzyme-based gene therapy mechanism. In addition, the hematoxylin and eosin (H&E) staining and the terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining of tumor sections from various treatment groups were exhibited in Fig. 5I. Compared with the other treatment groups, the tumor sections in the TDN-MPs + L group showed obvious nuclear shrinkage, fragmentation, cytoplasmic condensation and structure loss, which induced the most tumor cell necrosis. The strongest green fluorescence of the TDN-MPs + L group revealed that there were mostly apoptotic cells in these groups, consistent with the tumor inhibition results.

Fig. 5.

Imaging performance and therapeutic efficacy of TDN-MPs in vivo. (A) T₁-weighted MRI of 4T1-tumor-bearing mice injected intravenously with MPs and TDN-MPs (4 mg/kg) at 0, 0.5, 1, 2 and 5 h. In vivo imaging of mice injected by (B) MPs and (D) TDN-MPs at 0,8 and 24 h. In vitro imaging of tumors and major organs 24 h after (C) MPs and (E) TDN-MPs administration. (F) Schematic diagram of the treatment procedure for subcutaneous tumor inhibition in vivo. (G) Tumor volume curves of mice in different treatment groups. (H) Photos of the excised tumors from each group. (I) H&E and TUNEL staining of the tumors section after different treatments

Spatiotemporally differential regulation of ROS and VEGFR2 of TDN-MPs in vivo

To investigate the ability of TDN-MPs to differentially regulate ROS levels in vivo, dihydroethidium (DHE) staining was used to detect ROS levels in tumor tissues and normal tissues surrounding the tumor. Compared with the saline group, ROS levels in tumor tissues of TDN-MPs + L group were dramatically increased, clearly revealing that TDN-MPs could efficiently generate ROS in tumor under laser irradiation (660 nm, 200 mW/cm², 250 s) (Fig. 6A and S39). In contrast, TDN-MPs + L treated mice showed no significant change in ROS level in normal tissues surrounding the tumor. Interestingly, we observed evident red fluorescence corresponding to ROS in normal tissue around tumor in MPs + L group (Fig. 6B and S39). In order to quantitatively assess the ROS-regulating capacity of TDN-MPs, we measured the ROS levels both within the tumor and normal tissues surrounding the tumor using a DCFH-DA probe-based assay. The results were consistent with those obtained from DHE fluorescence staining analysis (Figure S40). These results further confirmed that TDN-MPs acts as ROS generators in tumor to produce sufficient ROS for inhibiting tumor growth, whereas works as ROS scavenger in the physiological environment to protect normal tissues surrounding the tumor. To further provide molecular-level evidence for the differential regulation in tumor and normal tissues, we supplemented the analysis by examining key markers of the oxidative stress response (Figure S41). Specifically, we assessed the expression of nuclear factor erythroid 2-related factor 2 (Nrf2) as a master regulator of the antioxidant pathway and 8-hydroxy-2’-deoxyguanosine (8-OHdG) as a definitive marker of oxidative DNA damage in both tumor and peritumoral normal tissues. In tumor tissues under laser irradiation, both the MPs + L and TDN-MPs + L groups showed a significant upregulation of Nrf2 expression, owning to a cellular compensatory response to the severe oxidative stress induced by PDT. Concomitantly, a pronounced increase in 8-OHdG levels was detected in these groups, directly confirming the occurrence of substantial oxidative DNA damage within the tumor. These results demonstrated that the potent ROS burst effectively overwhelms the endogenous antioxidant defense of the tumor, leading to lethal oxidative damage. In contrast, TDN-MPs treatment led to a slight decrease in the expression of both Nrf2 and 8-OHdG in normal organs (kidney, liver, spleen), which indicated a mild and beneficial modulation of the redox state and pointed to a potential cytoprotective function in healthy tissues. Collectively, these findings provide mechanistic corroboration that TDN-MPs not only selectively generate a potent oxidative burst in tumors to override their adaptive responses but also maintain redox homeostasis in normal tissues, thereby achieving a true differential regulation of ROS with enhanced therapeutic specificity and safety. We further explored the spatiotemporally differential regulation of VEGFR2 by TDN-MPs at the tumor sites and normal sites in vivo. Immunohistochemical staining showed that MPs + L treatment dramatically induced the VEGFR2 expression in tumor tissues (Fig. 6C) owing to the compensatory upregulation of VEGFR2 by the depletion of O₂ within the tumor based on PDT. However, the TDN-MPs treatment significantly reduced the VEGFR2 expression, demonstrating that DNAzyme could be successfully released and cleaved VEGFR2 mRNA at the tumor site. Interestingly, the level of VEGFR2 in the TDN-MPs + L treatment group was still decreased compared with that of saline group. To comprehensively evaluate the anti-angiogenic mechanism, we analyzed the expression levels of the angiogenesis-related factors VEGF-A and Ang-2. The results showed that TDN-MPs downregulate the expression of both VEGF-A and Ang-2, revealing their potent anti-angiogenic activity (Figure S42). These results indicated that the DNAzyme could successful overcome the up-regulation effect of PDT on VEGFR2 and specifically decrease the angiogenesis-related factors level in tumor, which was beneficial to inhibit tumor angiogenesis and enhance the therapeutic effect. To further investigate TDN-MPs whether could inhibit tumor angiogenesis, tumors from different treatment groups were extracted for CD31 (marker of vascular endothelium) and α-smooth muscle actin (α-SMA, marker of pericytes) staining to identify vessel maturity [36]. Tumors in saline group showed abundant CD31 red fluorescence signals and α-SMA green fluorescence, which suggested that the tumors contained extensive immature microvessels (Fig. 6D). The MPs + L group exhibited significantly increased α-SMA green fluorescence while less colocalization with CD31 red fluorescence in tumor pericytes, showing the presence of structurally and functionally abnormal vessels. The abnormal and chaotic vascular system in the tumor could aggravate the acidic and hypoxic microenvironment, which would promote the growth and invasion of the tumor [37]. Notably, the TDN-MPs and TDN-MPs + L group obviously reduced CD31 red fluorescence and α-SMA green fluorescence, indicating that VEGFR2-specific DNAzyme activated by the released Mn²⁺ ions could inhibit angiogenesis in tumors. To verify the effect of TDN-MPs on VEGFR2 expression in normal tissues, VEGFR2 expression in kidney, liver and spleen was measured by immunofluorescence. It appears that there was no difference in VEGFR2 expression between TDN-MPs treatment group and saline group, suggesting TDN-MPs maintained the unactivated state of DNAzyme in the physiology environment (Fig. 6E). The similar results were obtained with the staining of microvessels (CD31, red) and pericytes (α-SMA, green) in different normal tissue sections after TDN-MPs treatment, further demonstrating that TDN-MPs had no effect on angiogenesis in normal tissue (Fig. 6F). This superior targeted gene cleavage originates from the ingenious integration of the DNAzyme into the TDN framework. In physiological environments, the structurally intact TDN acts as a silencer that occludes the active site of DNAzyme, thus inhibiting its binding and cleavage of the target gene. To further determine whether the observed vascular changes corresponded to functional normalization of the tumor vasculature, we assessed intratumoral hypoxia levels by fluorescence staining for hypoxia-inducible factor-1α (HIF-1α) (Figure S44). The strong and widespread HIF-1α staining was observed in tumors from the saline and MPs + L groups, confirming the presence of a hypoxic microenvironment resulting from chaotic and inefficient blood supply. In contrast, tumors treated with TDN-MPs and TDN-MPs + L exhibited a marked reduction in HIF-1α signal intensity and distribution. This decrease in hypoxia aligns with the earlier morphological evidence of reduced and more organized vessel density (CD31/α-SMA), supporting the conclusion that TDN-MPs not only inhibited angiogenesis but also promoted a shift toward a more normalized and functional vascular phenotype. TDN-MPs take advantage of the structural variability of TDN to respond to tumor microenvironment and induce effective and synergistic ROS/anti-angiogenesis tumor therapy while limited toxic effects on normal tissue through differential regulation ROS and VEGFR2 levels between tumor and normal tissues. This tumor-selective ROS generation and VEGFR2 pathway inhibition strategy provide new avenue for enhanced and specific tumor therapy.

Fig. 6.

Spatiotemporally differential regulation of ROS and VEGFR2 of TDN-MPs in vivo. Fluorescence images of (A) tumor and (B) normal tissues surrounding the tumor collected from different treatment groups stained by DHE (red fluorescence) and DAPI (blue fluorescence). (C) Immunohistochemistry of VEGFR2 in tumor section after different treatments. (D) Representative images of tumor microvasculature (CD31, red) and pericytes (α-SMA, green) after various treatments. (E) Immunohistochemistry of VEGFR2 in normal tissue section after TDN-MPS treatment. (F) Representative images of different normal tissue section microvasculature (CD31, red) and pericytes (α-SMA, green) after TDN-MPs treatment. (G) H&E staining peritumoral skin slices images of the mice with different treatments

Notably, elevated ROS level induced by PDT resulted in inevitable photodamage to the skin surrounding the tumor. To evaluate the protective effects of ROS produced by TDN-MPs on the skin surrounding tumor under laser irradiation, the peritumoral skin was taken for HE staining immediately after laser irradiation (660 nm, 200 mW/cm², 250 s). Under laser irradiation, the MPs group showed apparent nuclear fragmentation, necrosis and cytoplasmic condensation compared that of non-laser irradiation group, revealing obvious skin tissue damage. Notably, TDN-MPs group under laser irradiation exhibited normal and intact nuclear and cytoplasmic morphology in skin tissues, confirming TDN-MPs could counteract the generation of ROS by PDT for protecting peritumoral tissue from photodamage side effects (Fig. 6G). We further examined the molecular response in skin tissues by assessing the expression of the antioxidant regulator Nrf2 and the oxidative DNA damage marker 8-OHdG (Figure S45). Under laser irradiation, MPs caused skin tissues to exhibit intense nuclear accumulation of Nrf2 alongside pronounced 8-OHdG positivity, while MPs-TDN significantly reduces both the Nrf2 and 8-OHdG signals. These molecular-level findings corroborate the histological observations from H&E staining. The mechanism could be attributed to the inherent antioxidant capacity of the TDN scaffold within TDN-MPs, which acted as an efficient ROS scavenger to neutralize excess ROS that may diffuse into the surrounding normal tissues during PDT. Notably, the elevated expression of VEGF is a well-established driver of tumor metastasis. Based on this mechanism, we analyzed TDN-MPs as a potential inhibitor of VEGF-facilitated tumor metastasis. The abundant metastatic nodules were present in the lungs of the saline and MPs groups, while the TDN-MPs treatment suppressed pulmonary metastasis (Figure S46). This anti-metastatic effect could be attributed to the DNAzyme released from TDN-MPs, which specifically cleaved VEGFR2 mRNA and downregulated VEGF expression at the tumor site. Collectively, these findings demonstrated that TDN-MPs not only suppressed primary tumor growth but also held significant potential for preventing tumor metastasis. Therefore, TDN-MPs constitute a smart nanoplatform whose specificity fundamentally distinguishes it from conventional and non-discriminative antioxidant approaches. The core innovation lies in the coordinated yet differential regulation of two key hallmarks, angiogenesis and ROS, within the tumor microenvironment, achieving integrated diagnosis and therapy for a new generation of precise cancer treatment.

Conclusion

In summary, we developed the tumor microenvironment (TME)-triggered TDN-MPs with variable geometric structures, that differentially regulate the ROS and VEGFR2 level in tumor and normal tissue to achieve enhanced tumor treatment and minimize the toxic on normal tissues. This innovative design diverged from traditional carriers by integrating the VEGFR2-specific DNAzyme into the tetrahedral DNA nanostructure (TDN) framework through its intrinsic programmability, rather than relying on conventional apical or side-arm modifications. This integration strategy not only ensured the exceptional stability of the DNAzyme for in vivo delivery but also repurposed the DNAzyme as a molecular switch to precisely regulate the antioxidant capacity of the TDN. Furthermore, the TDN scaffolded itself functions as a silencer for the DNAzyme in normal cells, effectively minimizing off-target toxicity by suppressing unintended activation. Specifically, the 660 nm laser irradiation of PPa and the rapid released of Mn²⁺ ions from the acidic TME allowed TDN-MPs to increase the ROS level and activate DNAzyme, which resulted in the killing of tumor cells and limit its anti-angiogenesis effect. In addition, TDN-MPs maintained structural integrity in healthy tissues, not only silencing DNAzyme function, but also effectively scavenging ROS and reducing photosensitive disease around tumor sites due to the unique antioxidant activity of the TDN backbone. Collectively, TDN-MPs represent not merely a delivery vehicle but a dynamic therapeutic system that differentially regulate the levels of ROS and anti-angiogenesis in disese, thereby maximizing specificity and therapeutic outcome.

Experimental section

Preparation of MnO-PPa (MPs)

We dissolved 4.56 g of sodium oleate in 10 mL of H₂O and 15 mL of ethanol, and then sonicated and stirred the mixture until it was clear. After that, 0.99 g of MnCl₂·4H₂O was dissolved in 10 mL of H₂O and slowly added to the above mixed solution. A 4 h heating period followed in order to remove any low volatile impurities from the reaction solution. The resultant solution was then cooled to room temperature and washed three times with hexane and placed in a dish to volatilize the organic phase to obtain manganese oleate (Mn (OA)₂). Afterwards, 1.0274 g of Mn (OA)₂ was dissolved in 236.6 mL of oleic acid (OA) and 12 mL of 1-Octadecene (ODE), which was sonicated and stirred until it became clear. In order to remove oxygen at room temperature, the mixture was degassed in vacuum for 10 min and then backfilled with N₂. After the reaction solution was heated to 90 ℃, it was kept there for 50 min to remove any low volatile impurities. As a next step, heat the reaction solution to 300 ℃ and maintain it there for 20 min. To precipitate the nanoparticles, the solution was cooled to room temperature and mixed with 30 mL of isopropanol. The obtained MnO were dissolved in hexane and stored at 4 ℃ after being washed three times with ethanol. Subsequently, the MnO (1 mg), DSPE-PEG2000 carboxyl (4 mg), Pheophorbidae a (PPa) (100 µg) were mixed in 260 µL chloroform and the mixed solution was subjected to sonication for 15 min, then the mixed solution was evaporated slowly. The sample was re-dispersed in water by gentle sonication after the chloroform was removed completely under vacuum. We stored the final aqueous solution at 4 ℃ for long-term storage.

Preparation of TDN-MPs

First, EDC (0.6 mg) and NHS (0.3 mg) dissolved in 100 µL of PBS (pH = 7.4) were added to 30 µLof MPs (1730 µg/mL), activating carboxyl groups on the surface of MPs. Next, equal amounts (100µM, 50 µL) of S1, S2, S3, S-DNAzyme were dispersed in HB buffer solution and heated at 95 ℃ for10 min and progressively cooled to 4 ℃ to form tetrahedral DNA nanostructures (TDN). Then, the obtained TDN that modified with amino groups was mixed with the acquired MPs activated with carboxyl groups and stirred overnight at room temperature to obtain TDN-MPs. The water-dispersed sample was purified by ultrafiltration centrifuge tube (10KD, MWCO) to remove incompletely reacted DNA, PPa, and other ions.

Mn²⁺ ions release detection

We first pretreated TDN-MP in water with pH 7.4, pH 6.5, and pH 5.5. During the times specified (0, 0.5, 1, 2, 4, 6, 8, 10, 12 h), the supernatant was collected by centrifugation at 16,000 rpm for 10 min. Then, spectrophotometric measurements of formaldehyde oxime were also conducted to determine the Mn²⁺ ions content of the supernatant.

Cytotoxicity assay

In order to culture cells, we obtained 4T1 and Hs 578Bst cells from the cell bank of the Chinese Academy of Sciences and maintained them at 37 °C under 5% CO₂. The reagents related to cell culture were all purchased from Meilun. We seeded cells into 96-well tubes at 1 × 10⁴ cells/well in RPMI-1640 medium with 10% FBS and 1% penicillin/streptomycin at 37 °C under 5% CO₂ for 24 h. After 24 h, the cells were incubated with MPs, TDN and TDN-MP with different concentrations. Next, 660 nm laser radiation was applied for 1 min for 60 mW/cm² and the cells were then incubated for 2 h. Immediately following the washing with PBS, CCK-8 (10 µL) and fresh medium (100 µL) were added to each well, and the plate was incubated for 30 min. After measuring the absorbance of each cell using an enzyme-linked immunosorbent assay reader (Thermo Multiskan FC), the relative viability of each cell group was compared.

Intracellular ROS scavenging detection

In the present study, 4T1 cells and Hs 578Bst cells were seeded into 96-well cell culture plates at a density of 1 × 10⁴ per well. Following incubation for 24 h, the medium was removed and Rosup (0.25 mg/mL) was added to produce intracellular ROS. A serum-free medium was refreshed with PBS after 60 min, followed by MPs and TDN-MPs (2 µg/mL) for 24 h at 37 °C. Next, a 660 nm laser was used to irradiate the cells for 1 min at a power of 50 mW/cm² before incubation for 2 h. Afterward, DCFH-DA (5 μm) in serum-free medium was added for 30 min in darkness at 37 °C. Lastly, the fluorescence of cells was captured using a fluorescence microscope (Leica DM3000).

Intracellular ROS generating detection

4T1 tumor cells or Hs 578Bst cells were seeded in 96-well plates (1 × 10⁴ cells/well) and incubated for 24 h (37 °C, 5% CO₂). After PBS washing, Hs 578Bst cells or 4T1 cells were incubated with 2 µg/mL MPs or TDN-MPs for 24 h. Cells were then irradiated (660 nm laser, 50 mW/cm², 1 min) and incubated for 2 h. Subsequently, cells were loaded with DCFH-DA (5 µM in serum-free medium, 30 min, 37 °C, dark), washed thrice with PBS, and immediately imaged using a Leica DM3000 fluorescence microscope.

Calcein AM and PI staining

Hs 578Bst cells or 4T1 cells were inoculated into a 96-well cell culture plate at 1 × 10⁴ per well and incubated at 37 °C under 5% CO₂ for 24 h. After washing the cells with PBS, Hs 578Bst cells or 4T1 cells were incubated with a certain concentration MPs or the same content TDN-MPs for 24 h. After washing the cells with PBS, the cells were irradiated by the 660 nm (50 mW/cm²) laser for 1 min. After incubated at 37 °C under 5% CO₂ for 4 h, the live and dead cells were stained with calcein-AM/PI and the cell images were observed by using the fluorescence microscope (Leica DM3000).

qRT-PCR analysis

4T1 cells and Hs 5780Bst cells were seeded in a 6-well plate at a density of 2 × 10⁵ cells per well for 24 h. Following adding to MPs and TDN-MPs (equivalent Mn²⁺ ions concentration of 2 µg/mL) for 24 h at 37 °C, the cells were irradiated for 1 min by the 660 nm (50 mW/cm²) laser. After that, fresh RPMI 1640 was added to the supernatants for incubating 2 h. The RNA kit was then used to isolate total RNA. Transcripts of total mRNA were reverse transcribed into single-stranded cDNA with the help of a first-strand cDNA synthesis kit and gDNA erasing solution. Next, we investigated VEGR2 expression using an IQTM5 Multicolor Real-Time PCR Detection System (Bio-RAD, USA) and TransStart Green qPCR SuperMix. β-actin was used as the internal control. Primers for β-actin: forward (5′-3′)-CATTGCTGACAGGATGCAGAAGG, reverse (5′-3′)-TGCTGGAAGGTGGACAGTGAGG. Primers for VEGFR2: forward (5′−3′) TCCAGAATCCTCTTCCATGC; reverse (5′−3′) AAACCTCCTGCAAGCAAATG. The expression of VEGR2 was determined using the 2^−∆∆ CT method.

In vivo subcutaneous tumor MR imaging

Female BALB/c mice (7 weeks, Chongqing Medical University Laboratory Animal Center) were subcutaneously inoculated with 4T1 cells (2.5 × 10⁶) in the right flank. Institutional Animal Care guidelines (Approval #20230106). When tumors reached 60–100 mm³, mice were intravenously injected with MPs and TDN-MPs (equivalent Mn²⁺ ions concentration of 4 mg/kg) to in vivo MRI. The transverse plane 2D MR images were scanned using a sequence (TR/TE = 400/11 ms, 256 × 256 matrices, thickness = 1.25 mm, FOV = 30 × 30 mm², slice = 9, matrix = 256 × 256) on the T₁ microMRI scanner. The MR images were used to obtain T1-weighted MR images at preinjection, 0.5, 1, 2, 3, 5 h post-injection. To qualify the signal enhancement, the signal-to-noise ratio (SNR) was calculated by the equation: SNRtumor = SItumor / SDnoise, where SI represents signal intensity and SD represents standard deviation.

In vivo therapy

Female BALB/c mice (7 weeks) were subcutaneously inoculated with 4T1 cells (2.5 × 10⁶) in the right flank. When tumors reached 60–100 mm³, mice were randomly divided into five groups: (1) Saline, (2) MPs, (3) TDN-MPs, (4) MPs + L (200 mW/cm², 250s), (5) TDN-MPs + L (200 mW/cm², 250s) (equivalent Mn²⁺ ions concentration of 4 mg/kg) through the tail vein. Laser irradiation was performed 4 h and tumor dimensions were measured every other days as length × (width)² × 1/2 with a ruler. Mice were sacrificed at day 14. Tumors and major organs (heart, liver, spleen, lungs, kidneys) were harvested for histopathology.

Statistical analysis

Statistical analysis were attained by Graphpad Prism 8.0 and the statistical mode were analysis of one-/two-way analysis of variance (ANOVA) and Student’s t-test, the data are presented as mean + SD. p values of less than 0.05 was considered significant (*p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001).

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

**Meiling Liu** : Conceptualization, Investigation, Methodology, Data curation, Formal analysis, Writing – original draft. **Mengke Fan** : Methodology, Data curation. **Yu Guo** : Methodology, Investigation. **Mingya Tan** : Methodology, Data curation. **Taotao Chu** : Methodology; **Linlin Huo** : Formal analysis; **Jiayi Zhao** : Data curation. **Xianghua Yang** : Data curation. **Xiaojing He** : Conceptualization, Supervision, Methodology. **Zhenghuan Zhao** : Conceptualization, Supervision, Funding acquisition, Writing - review & editing.

Funding

This work was supported by the Natural Science Foundation of Chongqing (CSTB2023NSCQ-LZX0033), Program for Youth Innovation in Future Medicine, Chongqing Medical University (W0105), Chongqing Medical University Graduate Top Talent Program (BJRC202312), the Science and Health Joint Medical Research Project of Chongqing (2024ZDXM004), Chongqing Natural Science Foundation Project (CSTB2024NSCQ-MSX0616), Graduate Scientific Research and Innovation Project of Chongqing (CYB240192).

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study and included experimental procedures were approved by the institutional animal care and use committee of Chongqing Medical University (Approval No. IACUC-CQMU-2023-0106). All animal housing and experiments were conducted in strict accordance with the institutional guidelines for the care and useConsent for publication.

Consent for publication

All authors of this study agreed to publish.

Competing interests

The authors declare no competing interests.