Spermine oxidase-DOX conjugates reshape tumor microenvironment via carbonyl stress to potentiate bladder cancer chemotherapy

Abstract

The tumor's antioxidant defenses restrict the efficacy of standard therapies, resulting in compromised treatment responses and systemic tolerance. Conventional redox-disrupting strategies, designed to generate reactive oxygen species (ROS), are limited by both their reliance on single generation methods and the rapid clearance. Herein, we report an innovative strategy that employs spermine oxidase (SMOX) to catalyze the production of endogenous and highly cytotoxic acrolein, thereby creating a destructive carbonyl stress microenvironment to potentiate antineoplastic drug's efficacy. As an example, SMOX and doxorubicin (DOX) were conjugated and co-encapsulated within mesoporous silica nanoparticles to form an integrated delivery platform (PMD). Notably, in vitro and in vivo results demonstrated that PMD sensitizes bladder and subcutaneous tumor cells to DOX by locally inducing carbonyl stress and oxidative damage, allowing for a reduced dosage of the chemotherapeutic agent. The underlying mechanisms involved enhanced ACR-mediated mitochondrial dysfunction and lipid peroxidation, P53 upregulation via GAPDH binding to induce apoptosis, and inhibition of the Warburg effect via suppression of glycolytic enzymes and Ras. Importantly, the therapeutic efficacy is achieved via either local instillation or systemic administration, highlighting the translational promise of this platform for bladder cancer and other malignancies.

Graphical abstract

To overcome tumor redox defenses, an integrated SMOX/DOX nanoplatform is developed. It promotes the in-situ production of highly cytotoxic acrolein via remodeling polyamine metabolism, inducing local carbonyl stress to potentiate chemotherapy and reduce required drug dosage, offering a novel strategy for treating bladder cancer and other malignancies.

Highlights

• •SMOX/DOX nanoplatform potently disrupts the redox homeostasis of TME.
• •PMD triggers in-situ acrolein production via remodeling polyamine metabolism.
• •PMD-induced carbonyl stress potentiates chemo, offering a novel cancer therapy.

1. Introduction

Distinct from normal tissues, the tumor microenvironment (TME) is characterized by a profound “redox imbalance” due to its highly heterogeneous and dynamic nature [1]. Elevated reactive oxygen species (ROS) levels act as crucial signaling molecules in tumor cells, underpinning multiple hallmarks of malignancy such as proliferation, survival under stress, and adaptation to hypoxia, while also facilitating invasion, metastasis, and the maintenance of genomic instability [[2], [3], [4]]. To counteract the resulting oxidative stress and prevent cellular damage, tumors paradoxically strengthen their antioxidant defense system. This adaptation includes the upregulation of non-enzymatic systems, such as glutathione (GSH) and NADPH, and enzymatic systems like Nrf2 and superoxide dismutase (SOD) [5,6]. This heightened redox homeostasis is not only connected with the occurrence and development of cancer progression, it also creates a formidable barrier to clinical treatment [7]. Many frontline interventions, including chemotherapy, radiotherapy, and immunotherapy, rely on disrupting the delicate redox balance to induce lethality [[8], [9], [10], [11]]. However, the robust antioxidant shield of the TME often neutralizes these therapeutic insults, leading to limited efficacy and systemic tolerance [12,13]. Targeting this intracellular redox equilibrium thus emerges as a critical strategy to enhance therapeutic sensitivity.

Substantial research efforts have focused on remodeling the TME redox landscape to circumvent therapeutic recalcitrance, strategies include leveraging nano-delivery systems to enhance intra-tumoral drug accumulation [14], and directly perturbing the tumor redox balance through hydrogen peroxide (H₂O₂) prodrugs [15], GSH depletion [16], or pro-oxidants to augment oxidative damage [[16], [17], [18], [19]]. Despite their promise, the clinical application of these approaches remains hampered by intricate synthesis and limited functional modularity [20,21]. Furthermore, the reliance on a singular mode of ROS generation is often insufficient, as the inherently short half-life and limited spatial diffusion of ROS, coupled with efficient endogenous clearance, inevitably curtail the sustained impact and overall reach of these interventions [12,22].

Emerging evidence indicates that polyamine (PA) metabolism is frequently dysregulated in cancer, the proliferation and survival of tumor cells are highly dependent on elevated PA levels [[23], [24], [25]]. Fortunately, PA can be catalyzed by plasma amine oxidase to produce a large amount of H₂O₂ and the highly reactive acrolein (ACR), which is a common reactive carbonyl specie (RCS). ACR induces carbonyl stress through electrophilic adduct formation, causing DNA damage and suppressing glutathione peroxidase 4 (GPX4) and DNA repair proteins, making up the drawback of oxidative damage therapy perfectly [26,27]. In our previous studies, we also have shown that PA-induced ACR accumulation is more cytotoxic than ROS and exacerbates oxidative damage via this pathway [28,29]. Thus, ACR is more conducive to creating stable oxidative stress TME, yet the exploitation of carbonyl stress to sustain chemotherapy effectiveness is rarely explored.

Herein, to address the drawbacks of oxidative damage therapy, we aim to establish a carbonyl stress-rich TME. First, we demonstrated that ACR is a far more potent cytotoxic agent than formaldehyde (Met). Given the high reactivity of ACR that precludes its direct delivery, harnessing the endogenous, spermine oxidase (SMOX) catalyzed generation of ACR from spermine is used to create a vulnerable redox environment for antineoplastic drugs. We coupled SMOX with doxorubicin (DOX) because of that this common anthracycline drug's efficacy is notably constrained by cellular antioxidant defenses. SMOX-DOX conjugates are delivered by mesoporous silica nanoparticles (PMD) to treat subcutaneous or in situ bladder cancer (BC) in vivo. PMD significantly increases carbonyl accumulation and disrupts intracellular redox homeostasis in multiple ways (Scheme 1). It is noted that this drug delivery harnesses the tumor's own spermine to induce ACR carbonyl stress, thereby amplifying oxidative damage and reducing the dosage of chemotherapy drugs. The PMD delivery system demonstrates excellent anti-tumor efficacy in both local and systemic treatments by depleting GSH and causing LPO. Furthermore, its ACR component binds to GAPDH to induce nuclear translocation, which in turn promotes apoptosis and inhibits glycolysis-related enzymes. This integrated mechanism shows broad effectiveness across diverse tumor cell types.

Scheme 1.

(A) Schematic illustration of PMD nanoparticle preparation and treatment. (B) Mechanisms of PMD-induced redox imbalance potentiating chemotherapy.

2. Results and discussions

2.1. Lethal effects of ACR and Met on BC cells

Aldehydes readily form adducts with macromolecules, especially protein and nucleic acid, leading to cell apoptosis [30,31]. ACR and Met, as the most prevalent reactive aldehydes in cells, were used to assess the specificity of carbonyl stress. MTT assays revealed that ACR exhibited potent cytotoxicity against SW780 and T24 cells, the IC₅₀ of ACR was one-tenth that of Met (Figs. S1A, B, E). Given that aldehydes are easily scavenged by GSH in vivo [32], we further investigate their impact on cellular GSH levels. Flow cytometry analysis demonstrated that ACR (50 μM) significantly depleted intracellular GSH in BC cells, whereas Met at the same concentration showed minimal consumption (Fig. 1A and B). The GSH/GPX4 axis is a crucial cellular antioxidant defense system that maintains redox homeostasis [34]. We observed that ACR more markedly reduced GPX4 protein expression compared to Met (Fig. 1C–E). Since GPX4 is a key regulator of ferroptosis [35], we subsequently examined ferritin-1 (Fer-1) expression and found it was downregulated by ACR treatment (Fig. 1D–F). These results suggest ACR induces ferroptosis through GSH depletion and concomitant suppression of related proteins. Ferroptosis is a known contributor to mitochondrial dysfunction [36], we further compared the effects of ACR and Met on mitochondria and found that ACR altered mitochondrial morphology within 24 h (Fig. 1G and H). Furthermore, ACR upregulated the pro-apoptotic protein BAX and downregulated the anti-apoptotic protein Bcl-2 in mitochondria (Fig. 1I–L). Cell cycle analysis by flow cytometry indicated that ACR increased the proportion of cells in G1 phase (Fig. 1M). These results suggest that ACR induces cell death through multiple synergistic mechanisms—including GSH depletion, ferroptosis promotion, mitochondrial disruption, and apoptosis induction with efficacy superior to that of Met (Fig. 1N), supporting its prospective application in BC treatment.

Fig. 1.

Effects of ACR and Met on BC cells. (A) Flow cytometry analysis GSH content in cells with ACR or Met treatment. (B) Fluorescence intensity of GSH in A. Western blot the expression of GPX4 (C) and Ferritin (D) after ACR and Met treatment. Quantitative analysis of GPX4 (E) and Ferritin (F) in C and D. (G) Confocal laser scanning microscopy (CLSM) images showing the morphology of mitochondrial after treatment with ACR or Met (Scale bar: 50 μm). (H) Fluorescence intensity of mito-tracker in G. Western blot of ACR and Met on expression of Bax (I) and Bcl-2 (J). Quantitative analysis of Bax (K) and Bcl-2 (L) in I and J. M) Analysis of the influence of ACR and Met on BC cell cycle by flow cytometry. N) Diagram of comparison between ACR and Met on cell viability. SMOX -triggered ACR to induce cell apoptosis. Data are shown as mean ± s.d. ∗p < 0.05; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, not significant by one-way ANOVA test. The schematic diagram was created with BioGDP.com [33].

2.2. Combination application of ACR and DOX on BC treatment

DOX is a commonly chemotherapy drug used in BC treatment and has multiple mechanisms, including DNA damage, ROS production, apoptosis and even ferroptosis [37,38]. To verify whether ACR-induced carbonyl stress could boost the chemotherapeutic efficacy of DOX, we conducted combination treatment experiments with ACR and DOX (termed AD). The IC₅₀ value of DOX on SW780 cells and T24 cells were 2.00 ± 0.18 μM and 2.47 ± 0.16 μM, respectively. When DOX was combined with 20 μM ACR and applied to BC cells, the IC 50 value of DOX decreased 10 times (Fig. S1C–E). The combination index of DOX and ACR on SW 780 cells and T24 cells were 0.58 and 0.60 respectively, indicating ACR and DOX had a synergistic effect. Next, the combined effect of ACR and DOX on oxidative damage was conducted. Bodipy staining revealed that both ACR and DOX slightly induced ROS production, but combination application significantly increased the content of ROS (Fig. 2A and B). The result was similar as shown by MFI of DCFH through flow cytometry (Fig. 2C and D).

Fig. 2.

Combined application of ACR and DOX on cell viability. (A) CLSM shows the bodipy staining after different treatments (Scale bar: 50 μm). (B) Quantitative of MFI of bodipy in A. (C) Flow cytometry of DCFH after different treatments. (D) Quantitative of MFI of DCFH in B. (E) Diagram of SMOX -triggered ACR to induce cell apoptosis. (F) Western blot of SMOX expression after transfected pCDNA3.1-SMOX for different times. (G) Cell viability of cells transfected with pCDNA3.1-SMOX for 48 and 72 h. (H) CLSM shows the ACR expression after different treatments (Scale bar: 50 μm). (I) Quantitative of MFI of ACR expression in H. (J) Relative intensity of H₂O₂ expression after different treatments. (K) Relative intensity of MDA after different treatments. (L) Flow cytometry of GSH after different treatments. (M) Quantitative of MFI of GSH in K. Data are shown as mean ± s.d. ∗p < 0.05; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, not significant by one-way ANOVA test.

Directly delivery ACR into cells to exert carbonyl stress is an ideal strategy, but the significant toxicity of its payload (especially severe off-target toxicity and irreversible Michael reaction) poses a huge challenge to the design of delivery systems [26]. Therefore, utilizing self-generated ACR from tumor cells for BC treatment is an effective strategy. Studies have shown that ACR is a by-product of lipid peroxidation (LPO) and PA metabolism processes, and primarily derived from spermine oxidation by SMOX accompanied by the production of H₂O₂. So, we carried out the potential of SMOX-triggered ACR for BC treatment (Fig. 2E). The pCDNA3.1-SMOX plasmid was transfected into BC cells, the cells expressed SMOX after 48 h, and their expression levels slightly increased after 72 h (Fig. 2F). Cells were transfected with pCDNA3.1-SMOX for 48 h, and then were incubated for another 24 h. The MTT assay showed cell proliferation was significantly inhibited (Fig. 2G). CLSM images showed that cells with transfection accumulated ACR (Fig. 2H and I), and the H₂O₂ and malonaldehyde (MDA) content was significantly increased (Fig. 2J and K). Flow cytometry showed cells transfected with pCDNA3.1-SMOX decreased in GSH content, which was similar with the effect of GSH synthetase inhibitor L-buthionine-(S, R)-sulfoximine (BSO) (Fig. 2L and M). These results suggest that it is feasible to self-generate ACR using SMOX, and the combination with chemotherapy drugs can achieve an amplifying lethal effect.

2.3. Preparation and characterization of PMD delivery system

To achieve co-delivery of SMOX and DOX, as shown in Fig. 3A, SMOX and DOX were linked through glutaraldehyde (termed SD) via Schiff base bonds. The UV-Vis absorption spectra showed that SMOX-DOX conjugates at molar ratios of 1:1, 1:5, 1:10 exhibited the prominent absorption enhancement at ∼280 nm (the characteristic peak of SMOX) and ∼480 nm (the characteristic peak of DOX) (Fig. 3B), confirming the successful conjugation reaction. And then SMOX-DOX conjugates were co-packaged into the mesoporous silica nanoparticles (MSN), forming PMD. Despite PMD_1:10 exhibiting a nominally lower IC₅₀ of 1.74 ± 0.32 μg/mL compared to the 1.99 ± 0.13 μg/mL of PMD_1:5 and the 3.41 ± 0.33 μg/mL of PMD_1:1, it showed no statistically significant advantage in cytotoxicity, therefore, PMD_1:5 was selected for subsequent experiments (Fig. S2A and B). SMOX was encapsulated into MSN (termed SM). The average particle size of SM and PMD were ∼105.45 nm and ∼110.65 nm, and their zeta potential were −10.65 mV and −15.65 mV, respectively (Fig. 3C–E, Fig. S2C). Transmission electron micrographs (TEM) revealed that PMD had a spherical vesicular structure with a diameter of ∼110.08 nm (Fig. 3F).

Fig. 3.

Preparation of PMD delivery system. (A) Schematic illustration of PMD preparation. (B) Full-wave ultraviolet scanning of SD. (C) Zeta potential of each NPs (n = 3). (D) Size distributions of each NP (n = 3). (E) DLS analysis of NPs (n = 3). (F) TEM of PMD (Scale bar: 100 nm). (G) Release rate of SMOX in pH 5.5 PBS, PBS and medium, respectively. (H) LSCM images of co-localization of PMD and lysosomes for incubation of different times. (I, J) Flow cytometry analysis and quantification of cellular uptake of NPs in MB 49 cells. (K) LSCM showing cellular uptake of DOX and PMD by BC cells MB49 (Scale bar: 20 μm). (L) Relative intensity of fluorescence intensity of different bladder cell lines by flow cytometry. Data are shown as mean ± s.d. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗∗p < 0.0001; ns, not significant by one-way ANOVA test.

PMD demonstrated pH-responsive release of both SMOX and DOX due to the formation of Schiff base bonds (Fig. 3G, Fig. S3). Consistent with the subcellular location, PMD exhibited colocalization with lysosomes in MB49 cells after 30 min incubation, while the majority had escaped from the lysosomes by 2 h (Fig. 3H). Flow cytometry showed that the drug uptake by cells increased significantly within a short period, reaching approximately twice the uptake of DOX alone (Fig. 3I and J). Further CLSM visualization indicated that the majority of PMD were internalized by the cells within 30 min (Fig. 3K). Besides, flow cytometry showed that the drug uptake levels in BC lines T24 and SW 780 cells were 3-fold and 4-fold higher, respectively, than that in the human ureteral epithelium cell lines SV-HUC-1 after 10 min of incubation (Fig. 3L, Fig. S4). These results suggest that PMD was successfully constructed and efficiently internalized by cells, thereby having the potential to induce ACR accumulation and disrupt the redox regulation in the TME.

2.4. Mechanism of PMD-induced carbonyl stress

After successfully constructing PMD, the multiple mechanisms of PMD induced were verified. PMD effectively inhibited cell viability to approximately one-tenth of that achieved by DOX alone (∼5 μM) (Fig. 4A). Besides, PMD exhibited a relatively weak cytotoxic effect on SV-HUC-1 cells (an SV40-immortalized human urothelial cell line). At a PMD concentration of 4 μg/mL, nearly all MB49 cells underwent cell death, whereas approximately 20% of SV-HUC-1 cells remained viable. Notably, the IC₅₀ value of PMD for SV-HUC-1 cells was 1.5-fold higher than that for MB49 cells (Fig. S5). Calcein-AM/PI staining showed an approximately 100-fold increase in the population of red-fluorescence dead cells in the PMD groups than in the DOX group (Fig. 4B and C). PMD treatment significantly reduced the colony formation of MB49 cells (Fig. S6), as well as cell migration evaluated by Transwell assay (Fig. S7) and scratch wound healing assay (Fig. S8) Following PMD delivery, it significantly increased ACR accumulation (Fig. S9). The effect of PMD on the redox TME of tumor cells was further assessed. Flow cytometric analysis revealed a significant reduction in GSH levels following PMD treatment (Fig. S10A). Quantitative analysis indicated that GSH content in the PMD-treated group was approximately 8-fold lower than that in the control group (Fig. S10B). Correspondingly, ThiolTracker Violet staining revealed a substantial decline in fluorescence intensity, indicative of GSH depletion after PMD treatment (Fig. S10C). Interestingly, GSH fluorescence signal in control SV-HUC-1 cells was markedly lower than that in MB49 cells, and PMD treatment further decreased the GSH signal in SV-HUC-1 cells (Fig. S10D). Quantitative analysis of GSH revealed that the relative GSH level in PMD-treated SV-HUC-1 cells was 5-fold lower than that in the control group, and approximately 8-fold lower than that in the bladder cancer cell line (Fig. S10E). These results suggest that cancer cell lines with high GSH expression are more sensitive to PMD. As expected, the DOX groups exhibited an approximately two-fold increase in ROS, while the PMD groups showed a dramatic 10-fold increase (Fig. 4D and E). Consistently, the levels of H₂O₂ (Fig. S11A) and MDA (Fig. S11B) were also sharply elevated in the PMD groups. Since ACR reduced Fer-1 and GPX4 expression, using an Fe²⁺ probe, we observed significant Fe²⁺ accumulation specifically in the PMD groups (Fig. 4F). Ferrostatin-1 (Fer) and deferoxamine mesylate (DFOM) are classic ferroptosis inhibitors, which function by suppressing the chelation of intracellular iron ions and LPO, respectively [39,40]. Fer and DFOM rescued cell viability by 60% and 40% (Fig. S12), respectively, thereby confirming that the combined oxidative damage from ACR and DOX induces ferroptosis via two parallel pathways: LPO and iron-dependent toxicity.

Fig. 4.

Effect of PMD on cell viability. (A) MTT assay of MB49 cells under various treatments. (B) Viability assessment by Calcein-AM/PI staining (Scale bar: 20 μm). (C) Quantitative analysis of live/dead cells from B. (D) CLSM images of MB49 cells stained with DCFH-DA for ROS detection (Scale bar: 50 μm). (E) MFI of DCFH from D. (F) CLSM images of MB49 cells stained with an Fe²⁺ probe (Scale bar: 50 μm). (G) Western blot analysis of P53, Bcl-2 and BAX expression. (H) JC-1 staining of MB49 cells for mitochondrial membrane potential (Scale bar: 50 μm). (I) Flow cytometry analysis of DCFH-DA-stained MB49 cells for ROS. (J) Western blot analysis of Ras expression. (K) Scheme illustrating carbonyl stress-enhanced oxidative stress. Data are shown as mean ± s.d. Asterisks indicate significant difference compared to control group (∗∗p < 0.01; ∗∗∗∗p < 0.0001); Hash signs indicate significant difference between DOX and PMD groups (^####p < 0.0001); ns, not significant by one-way ANOVA test.

ACR binding inactivates GAPDH, which then interacts with the E3 ubiquitin ligase Siah and undergoes nuclear translocation. In the nucleus, this complex promotes P300/CBP expression, leading to the subsequent activation of P53 expression [[41], [42], [43]]. After treatment with PMD, the ACR (red) and GAPDH (green) exhibited strong colocalization in the cytoplasm of SW780 cells, indicating that ACR binds to GAPDH (Fig. S13A). Next, we investigated the essential role of the P53 protein in the PMD therapy. Although PMD significantly increased P53 expression, treatment with the P53 inhibitor pifithrin-α hydrobromide (PH) partially reversed this effect, leading to an approximately 60% increase in cell proliferation (Fig. 4G, Fig. S13B). Besides, P53 protein induces the expression of downstream apoptotic proteins, including BAX/Bcl-2 [44,45]. Western blot indicated that PH suppressed pro-apoptotic BAX whereas it enhanced the anti-apoptotic Bcl-2 (Fig. 4G). Based on this role of BAX/Bcl-2 in regulating mitochondrial membrane integrity [46,47], we postulated that PMD induces mitochondrial damage. This was confirmed by TMRE and JC-1 staining, which specifically showed that PMD collapsed the mitochondrial membrane potential, and that this collapse was rescued by PH (Fig. 4H, Fig. S14). Notably, PH failed to reverse the PMD-induced ROS accumulation (Fig. 4I), suggesting that the elevated ROS levels were primarily attributable to the direct action of the metabolic product, ACR.

GAPDH is also a key glycolytic enzyme that catalyzes ATP and pyruvate production [48,49]. PMD was found to downregulate the expression of GAPDH and other glycolytic enzymes, such as PGK, PFK and PK (Fig. S15A and B). To further verify whether the reduced expression of glycolysis-related enzymes was caused by PMD-induced ACR, we treated SW 780 cells with hydralazine. The results showed that 5 μM hydralazine almost completely abolished ACR signals (Fig. S15C). Moreover, co-treatment with PMD and hydralazine significantly restored the expression levels of GAPDH, PFK, PGK and PK (Fig. S15D), indicating that PMD-induced ACR suppressed glycolysis-related enzymes expressions and thereby impaired glycolysis. Concurrently, PMD significantly suppressed Ras expression (Fig. 4J), a central driver of the Warburg effect, indicating that PA metabolism may impair glycolytic activity through this oncogenic signaling pathway as well. Above all, PMD strategy promotes endogenous ACR generation, induces carbonyl stress, and thereby reshapes the tumor redox TME by enhancing oxidative stress and inhibiting glycolysis (Fig. 4K). This remodeling sensitizes tumor cells, allowing effective killing with reduced chemotherapy doses.

2.5. Antitumor activity of PMD in subcutaneous BC models

Tumor-bearing mice were divided into four groups and administered different treatments via intravenous injection weekly over four weeks (Fig. 5A). To track biodistribution, FITC-labeled PMD was injected into tumor-bearing mice, which were euthanized at various time points for ex vivo imaging of organs and bladder tissue. Analysis showed tumor-selective accumulation of PMD, with about 50% of the fluorescence signal retained within tumors at 24 h post-injection (Fig. 5B–D). To prove that the PMD delivery system effectively reached and entered bladder tumor cells in vivo, we measured the resultant increase in intratumoral SMOX protein levels following its intravenous administration (Fig. 5E). Hemolysis tests revealed no detectable hemolytic activity for PMD, indicating excellent hemocompatibility (Fig. S16). Throughout the treatment period, body weight in the Control group declined after day 24, whereas it remained stable in all other groups (Fig. 5F). Following treatment, tumor growth was significantly suppressed in the PMD group, with the final tumor volume being only one-fifth of that in the DOX used alone (Fig. 5G, Fig. S17). Upon euthanasia, the average tumor weight in the PMD group was approximately half and one-fifth of that in the SM and DOX groups, respectively (Fig. 5H and I). Consistent with this, H&E staining of excised tumors revealed extensive necrotic areas in both the SM and PMD groups (Fig. S15A). ACR staining showed significantly higher levels in the SM and PMD groups than in the DOX group (Fig. 5J). Treatment with PMD induced extensive tumor necrosis without signs of systemic toxicity (Fig. S18A), as blood tests and histology of internal organs revealed no significant abnormalities or damage (Fig. S18B). These results indicate that the delivery system not only maintains its tumor-targeting capability via EPR within the complex systemic circulation but also holds promise for application beyond the current model, potentially extending to other diseases.

Fig. 5.

Anti-tumor activity and biosafety evaluation of PMD in subcutaneous tumor-bearing mice. (A) Schematic of the subcutaneous BC model and the intravenous treatment protocol. (B) In vivo fluorescence imaging of PMD accumulation in tumors at indicated time points (Scale bar: 1 cm). (C) Quantification of fluorescence intensity in tumors and major organs from B. (D) Quantification of tumor-specific fluorescence intensity from C. (E) IHC staining for SMOX in tumor sections (Scale bar: 500 μm). (F) Body weight changes of mice across treatment groups (n = 5). (G) Tumor growth curves monitored over 28 days (n = 5). (H) Tumor weights corresponding to samples shown in I. (I) Representative photographs of excised tumors from each group. (J) IF staining for ACR in tumor tissues (Scale bar: 200 μm). Data are shown as mean ± s.d. ∗∗∗∗p < 0.0001; ns, not significant by one-way ANOVA test.

To perform a long-term and systematic toxicity study, we established a subcutaneous tumor model using MB49 cells, with mice divided into Control, DOX (4 mg/kg) and PMD (20 mg/kg) group, and drugs were administered twice a week. First, PBS, DOX, and PMD were injected via the tail vein, and blood samples were collected at 2 h, 4 h, and 6 h to determine serum DOX residual concentrations by fluorescence scanning. The results showed that serum DOX levels in the DOX group were significantly higher than those in the PMD group at all time points, with approximately 20% DOX still remaining at 6 h (Fig. S19A). Further monitoring during treatment revealed that mice in the control and DOX groups began to die on 27 d and 30 d, respectively, and DOX groups exhibited body weight loss on 21 d, whereas no significant changes were observed in the PMD group (Fig. S19B–E). After treatment, the tumor weight in the PMD group was only one-eighth of that in the DOX group (Fig. S19F and G). Serum analysis on 24 d and 33 d demonstrated that TNF-α and IL-6 levels in the DOX group were 1.5-fold and 3-fold higher than those in the PMD group, respectively (Fig. S19H and I). These findings indicate that the markedly reduced DOX accumulation induced by PMD significantly alleviates systemic toxicity and confers favorable biosafety.

Furthermore, at the end of treatment, tumor tissues from each group were sectioned via cryostat and stained with 4-hydroxynonenal (4-HNE). Results revealed that, despite the high dose of DOX administered, the green fluorescence signal indicative of LPO was still significantly weaker in the DOX group than in the PMD group (Fig. S19J), suggesting that PMD exerted its antitumor activity in vivo via a LPO mechanism.

2.6. Antitumor activity of PMD in orthotopic BC mouse model

Benefiting from the bladder's luminal structure, local intravesical instillation is applicable, whereas intravenous injection is the standard systemic route for other tumors, an orthotopic BC mouse model was established. Tumor-bearing mice were then randomly assigned to four treatment groups for therapy via intravesical instillation (Fig. 6A and S20). First, to assess the permeability of PMD, Cy5.5-labeled PMD was used for intravesical instillation of both normal and tumor-bearing mice. Results showed preferential accumulation of PMD in tumor tissue over adjacent normal bladder tissue, with an approximately 10-fold higher fluorescence intensity in the tumor region (Fig. 6B–D). At 36 h post intravesical instillation, 10% of the initial fluorescence signal was still detectable in the tumors (Fig. 6E and F). Tumor growth was markedly inhibited in the PMD group, with the corresponding bioluminescence signal reduced to 10% and 20% of the levels observed in the SM and DOX groups at the end of treatment, respectively (Fig. 6G, H, Fig. S21). There was no significant difference in the body weight of the four groups (Fig. 6I). The PMD group exhibited a significantly higher survival rate than that of the DOX-alone group, with a 10-fold increase., However, this superior efficacy was achieved with a DOX dosage of only 0.4 mg/kg, which is one-tenth of the common dose of ∼4 mg/kg (Fig. 6J). These results demonstrate that co-delivering SMOX and DOX to tumor cells is significantly more effective than using either agent alone, indicating a clear synergistic therapeutic effect.

Fig. 6.

In vivo targeting, biodistribution, and therapeutic efficacy of the PMD system in an orthotopic BC model. (A) Schematic of the orthotopic BC model establishment and the intravesical PMD treatment regimen. (B) Images of PMD targeting bladder tumor in vivo. (C) Ex vivo fluorescent imaging of PMD distribution in bladder tissues and major organs (Scale bar: 1 cm). (D) Quantification of fluorescence intensity in bladder tissues from C. (E) Time-course in vivo imaging of PMD retention in bladder tumors. (F) Quantitative analysis of tumor-associated fluorescence intensity from E. (G) Quantification of tumor burden by bioluminescence imaging across treatment groups (n = 5). (H) Representative bioluminescence images of mice from each treatment group. (I) Body weight of BC tumor-bearing mice during treatment (n = 5). (J) Kaplan-Meier survival analysis of mice in different treatment groups (n = 10). Data are shown as mean ± s.d. ∗∗∗∗p < 0.0001; ns, not significant by one-way ANOVA test.

2.7. PMD-induced redox imbalance in vivo

To investigate the multiple mechanisms underlying PMD anti-tumor efficacy in vivo, tumors were harvested from the four treatment groups for molecular detection, IF and IHC. The GSH content in tumors from the SM and PMD groups was significantly decreased, with 0.92 ± 0.45 mg/kg and 0.12 ± 0.03 mg/kg, respectively (Fig. 7A), demonstrating that SMOX-triggered ACR effectively depleted the antioxidant pool, thereby exacerbating oxidative damage. As a key marker of LPO and oxidative stress, the level of MDA was elevated approximately 5-fold and 8-fold in the SM and PMD groups, respectively, compared to the control (Fig. 7B). To assess glycolytic activity, we analyzed the mRNA expression of key glycolytic enzymes (GAPDH, PFK, PGK, and PK) in tumor lysates by RT-qPCR. The results showed that the transcript levels of all four enzymes were significantly suppressed in the PMD group (Fig. 7C). This directly led to a substantial decrease in lactate production in the PMD group, with just 1 ± 0.03 mM (Fig. 7D). Strong SMOX signals were detected in both the PMD and SM groups (Fig. 7E), confirming successful NP delivery and suggesting their functional activity within tumor cells. SMOX and DOX acted synergistically to downregulate GPX4 expression (Fig. 7F). Moreover, Perls’ Prussian blue staining revealed a substantial accumulation of free iron in PMD-treated tumors (Fig. 7G). Together, these results indicate that PMD not only triggers LPO but also promotes ferroptosis in vivo. Immunohistochemistry (IHC) revealed P53 overexpression in both the PMD and SM groups (Fig. 7H), consistent with ACR-induced P53 upregulation thereby likely promoting apoptosis.

Fig. 7.

Therapeutic efficacy and mechanistic studies of the PMD system in orthotopic BC models. (A) Quantitative analysis of GSH concentration in tumor tissues. (B) Measurement of MDA levels in tumors. (C) Relative mRNA expression of glycolytic enzymes in tumors. (D) Lactate concentration in tumor tissues across treatment groups. IHC for SMOX expression (E) and the ferroptosis marker GPX4 (F). (G) Perls' Prussian blue staining for free iron in tumors. (H) IHC for the tumor suppressor P53. (I) H&E staining of tumor tissues from different treatment groups. (J) TUNEL staining of tumor sections. (K) IHC for the proliferation marker Ki-67. Scale bar: 500 μm. Data are shown as mean ± s.d. ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, not significant by one-way ANOVA test.

Next, to evaluate the safety of PMD, tumors were fixed in 4% paraformaldehyde, paraffin-embedded, and sectioned for histological staining. H&E staining revealed an extensive area of necrosis in the PMD-treated group, whereas tumors treated with DOX or SM alone showed only focal or minimal necrotic regions (Fig. 7I). Consistent with this, TUNEL assay demonstrated a markedly higher number of apoptotic cells in the PMD group compared to the single-agent treatment groups (Fig. 7J). Ki-67 expression was high in the Control group but significantly suppressed in the PMD group (Fig. 7K), indicating effective inhibition of tumor cell proliferation, which is associated with better clinical prognosis. Following the completion of treatment and euthanasia, major organs were collected and subjected to H&E staining. The results showed no significant pathological abnormalities in the lung, liver, spleen, kidney, or heart (Fig. S22).

Collectively, these results demonstrate that PMD exerts potent anti-tumor activity via multiple mechanisms, including the enhancement of LPO, inhibition of glycolysis, and induction of apoptosis. Owing to its effective tumor-targeting capability and favorable biocompatibility, PMD also exhibits a reduced risk of systemic side effects.

3. Conclusion

In summary, SMOX-induced endogenous ACR acts as a central orchestrator of therapy by amplifying oxidative damage to drive LPO, apoptosis, and ferroptosis. Its direct binding to GAPDH further induces apoptotic signaling while suppressing the Warburg effect through inhibition of glycolytic enzymes, thereby crippling tumor metabolism. ACR-induced carbonyl stress disrupted the tumor redox TEM, enabling potent anti-tumor activity with significantly reduced doses of chemotherapeutic agents. Given its significant therapeutic effect in subcutaneous models, the PMD delivery system may offer a promising treatment paradigm for other tumor types, especially cancers arising in tissues with high PA metabolic activity (e.g., intestinal, prostate, and cardiovascular cancers). Altogether, our findings demonstrate that exploiting the metabolic characteristics of tumor cells for self-destruction offers promising therapeutic prospects for BC treatment.

4. Experimental section

Animals: The 4∼6 weeks C57 mice weighting 18.0 to 20.0 g were raised in comfortable environment, fed standard food, and allowed to drink freely. All animal experiments were approved by the Animal Care and Experiment Committee of Qilu Hospital affiliated to Shandong University (approval No. DWLL-2024-150).

Synthesis of the SMOX-DOX conjugates: SMOX was conjugated to DOX using glutaraldehyde as linker [50]. Briefly, SMOX (5 mg) and DOX (5 mg/mL) were mixed at molar ratios of 1:1, 1:5, and 1:10 (SMOX: DOX) in 1 mL of PBS. Glutaraldehyde was added at a 1:1 M ratio to DOX. The reaction mixtures were incubated at 4 °C for 12 h with gentle agitation, after which the reaction was quenched by adding 0.1 M glycine (pH 5.0) to a final concentration of 50 mM. The conjugates were purified from unreacted components using an Amicon Ultra 0.5 mL centrifugal filter (10 kDa MWCO, Millipore) and washed three times with PBS. The coupling yield of SMOX was quantified using a BCA protein assay kit. The concentration of conjugated DOX was measured spectrophotometrically at 480 nm and calculated based on a standard curve of free DOX (0–10 μg/mL).

Antitumor activity of PMD in orthotopic BC model: The mice were anesthetized, and luciferase-expressing MB49 cells were inoculated into the bladder wall using an insulin syringe. Following tumor inoculation, once the fluorescence intensity reached 20,000 ± 5000, mice were randomly allocated into four treatment groups (n = 5) for intravesical instillation once a week: Control, DOX (0.4 mg/kg), SM (20 mg/kg, containing 12 mg/kg SMOX), and PMD (20 mg/kg, containing 0.4 mg/kg DOX and 12 mg/kg SMOX). Tumor growth and body weight were monitored throughout the 21-day treatment period. For bioluminescence imaging, mice were injected intraperitoneally with D-luciferin, anesthetized, and then imaged using an IVIS Spectrum system. The images were analyzed by the Living Image software. At the experimental endpoint, mice were euthanized. Tumors and major organs were harvested for subsequent histological and molecular analyses.

Anti-tumor activity of PMD in subcutaneous tumor models: Subcutaneous BC mouse model was constructed as previously reported [51]. Briefly, MB49 cells were harvested and inoculated subcutaneously into mice at a density of 1 × 10⁶ cells per mouse. When the tumor volume reached approximately 100 mm³, mice were randomly assigned to four groups (n = 5) and treated twice weekly via intravenous injection with: PBS (Control), DOX (0.4 mg/kg), SM (20 mg/kg, containing 12 mg/kg SMOX), or PMD (20 mg/kg, containing 0.4 mg/kg DOX and 12 mg/kg SMOX). Tumor volume (calculated as length × width² × 0.5) and body weight were monitored regularly over a 21-day treatment period. At the experimental endpoint, mice were euthanized. Tumors were excised, weighed, and photographed. Key organs (e.g., heart, liver, spleen, lung, kidney) were harvested for subsequent histopathological analysis.

Statistical analysis: Statistical analyses were performed using GraphPad Prism9.3.0 and SPSS version 23. Each experiment was done at least three times by triplicates. For comparison the statistical differences between tween groups, unless otherwise stated, Student's t-test (two-tailed) was carried out. Asterisks indicate significant difference (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001), p < 0.05 was considered statistically significant.

CRediT authorship contribution statement

Yingying Xu: Conceptualization, Data curation, Methodology, Software, Writing – original draft, Writing – review & editing. Chengyang Zhao: Formal analysis, Investigation, Project administration, Software. Shuai Fu: Investigation, Methodology, Project administration, Resources. Yaohai Wu: Methodology, Project administration, Resources, Software. Yingzhao Shao: Formal analysis, Project administration, Resources. Dongchen Li: Project administration, Software, Supervision. Yingwen Wang: Supervision, Validation. Lingchang Zhang: Methodology, Resources, Supervision. Aijing Zhang: Investigation, Resources, Software. Yuxiang Meng: Methodology, Resources, Software. Zhongwei Zhao: Methodology, Software. Jiaqi Li: Resources, Supervision. Zhao Zhang: Methodology, Resources, Supervision, Visualization. Huimin Geng: Funding acquisition, Project administration, Visualization, Writing – review & editing. Nengwang Yu: Conceptualization, Funding acquisition, Resources, Writing – original draft, Writing – review & editing.

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.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.