Porphyrin oriented bimetallic sonodynamic MOF for enhanced ferroptosis and cGAS-STING pathway mediated antitumor immunotherapy

Abstract

Sonodynamic therapy (SDT) offers deep tissue penetration and noninvasive tumor treatment, yet its efficacy remains limited by inadequate reactive oxygen species (ROS) generation and an immunosuppressive tumor microenvironment (TME). In this work, we developed a porphyrin-based bimetallic nanoplatform by integrating Fe and Mn centers within a structurally ordered metal-organic framework (MOF), followed by surface modification with the tumor-homing peptide CRGDK to achieve active targeting. The rationally engineered Mn-Fe(TCPP) MOF exhibits a spatially confined configuration that minimizes π-π aggregation of porphyrins enhancing ROS production under ultrasound (US). Meanwhile, Fe³⁺ and Mn²⁺ are released in the mildly acidic TME. The Fe³⁺ catalyzes Fenton-like reactions to generate abundant •OH radicals, leading to glutathione (GSH) depletion, glutathione peroxidase-4 (GPX4) inhibition, and triggers ferroptosis. Simultaneously, Mn²⁺ and ROS inflict mitochondrial damage and cause cytosolic double-stranded DNA (dsDNA) leakage, thereby activating the cGAS-STING signaling pathway and amplifying type I interferon (IFN-I) responses. This dual immunometabolic modulation induces immunogenic cell death (ICD), promotes dendritic cells (DCs) maturation, and strengthens adaptive antitumor immunity. Overall, this study presents a concise strategy coupling ferroptosis induction with innate-immune activation through a porphyrinic bimetallic MOF, offering a promising direction for SDT-based immunotherapy against hard-to-treat tumors.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04302-8.

Highlights

A CRGDK-conjugated Mn–Fe porphyrinic MOF (MFMP) with ordered confinement is constructed to achieve active targeting of tumor tissues.

The ordered confinement suppresses ACQ and markedly enhances ultrasound-triggered ROS generation.

Synergistic ferroptosis–pyroptosis promotes cytosolic dsDNA leakage and robust activation of the cGAS–STING pathway.

MFMP achieves potent suppression of primary tumors, distant lesions, and pulmonary metastasis in the 4T1 triple-negative breast cancer model.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04302-8.

Introduction

Cancer immunotherapy has revolutionized clinical oncology, with adoptive cell therapy (ACT) and immune checkpoint blockade (ICB) serving as representative therapeutic strategies [1]. Unfortunately, these approaches are constrained by limited response rates, the complexity of the preparation process, unpredictable efficacy, specific immune response, and the risk of immune-related adverse events (irAEs) [2–4]. Growing research efforts in sonotherapy and nanotechnology have been directed toward modulating the immune system in cancer therapy, as they can induce the formation of a tumor-associated antigen (TAA) pool and facilitate DCs maturation [5, 6]. However, local SDT is usually insufficient to elicit robust and durable anti-tumor immunogenicity and effectively inhibit tumor growth and metastasis, largely due to the limited generation of reactive ROS and the barriers posed by the complex TME, which together restrict the therapeutic efficacy and clinical translation of SDT nanomedicines. Therefore, there is an urgent need to combine it with immune stimulation strategies to achieve a synergistic effect of sono-immunotherapy for triple-negative breast cancer (TNBC), which is characterized by poor immunogenicity [7].

US has emerged as a promising strategy due to its unique advantages, including low tissue scattering, deep penetration, and noninvasiveness, which distinguish it from conventional photodynamic therapy (PDT) [8, 9]. These properties enable US therapy to precisely and controllably catalyze the generation of abundant ROS within deep seated tumor lesions. ROS not only serve as key effectors of US mediated antitumor activity but also play an important role in the regulation of pyroptosis. As a pro-inflammatory form of programmed cell death, pyroptosis promotes the release of inflammatory cytokines and DAMPs, thereby reversing the immunosuppressive tumor microenvironment and enhancing antitumor immune responses. Compared with conventional approaches that induce pyroptosis using ions, small molecules, or chemotherapeutic agents, the induction of pyroptosis through exogenous energy triggered ROS generation offers superior spatiotemporal controllability and targeting specificity, highlighting its potential for precise cancer immunotherapy. Porphyrins, as a class of typical organic sonosensitizers, are well recognized for their excellent sonodynamic properties [10, 11]. However, their intrinsic planar conjugated structure often leads to an undesirable aggregation-caused quenching (ACQ) effect during application, resulting in insufficient ROS generation and consequently limited therapeutic efficacy [12]. To overcome ACQ-induced limitations and enhance immune activation, several strategies have been explored, including encapsulation within nanocarriers for spatial separation and targeted release, construction of metal-organic frameworks (MOFs) to restrict porphyrin aggregation, and integration with other therapeutic modalities to boost immune responses [13]. Among these, immobilizing porphyrins within structurally rigid and stable MOFs has been considered an effective approach to suppress ACQ and improve ROS production.

Despite these advances, the resulting immunogenicity remains suboptimal. Loss-of-function alterations in interferon (IFN) signaling pathway genes in tumor cells constitute a key mechanism contributing to their reduced immunogenicity [14]. Among the upstream regulators of IFN signaling, the STING pathway plays a central role by sensing cytosolic DNA and inducing IFN-I responses, thereby linking innate sensing to adaptive antitumor immunity [15, 16]. The pivotal role of metal ions in immune modulation has garnered increasing attention in recent years. Among them, Mn²⁺ has emerged as a potent activator of the cGAS-STING pathway, driving the production of IFN-I and proinflammatory cytokines, thereby facilitating DCs maturation and triggering antitumor immune responses [17]. Nevertheless, Mn²⁺ alone exhibits limited therapeutic efficacy and fail to induce a robust cellular immune response [18]. Ferroptosis, a distinct iron-dependent form of regulated cell death characterized by uncontrolled lipid peroxidation, provides a unique opportunity for the release of cytosolic dsDNA from damaged mitochondria and the generation of dsDNA fragments from dying tumor cells, thereby enabling dual activation of the cGAS-STING pathway. Therefore, combining ferroptosis with strategies that potentiate cGAS-STING pathway activation represents a promising direction for amplifying antitumor immunity.

Herein, we synthesized manganese-iron porphyrin nanoparticles (Mn-Fe(TCPP) MOF) via a one-step co-assembly. Iron porphyrin and manganese chloride were co-dissolved in PVP/DMF, and the resulting nanoparticles were collected by centrifugation and subsequently functionalized with CRGDK to afford MFMP (Scheme 1a). MFMP integrates dual immunometabolic functions that cooperate with SDT, driving ferroptosis and activating the cGAS-STING pathway (Scheme 1b). Specifically, MFMP releases abundant Fe³⁺ and Mn²⁺ ions in the mildly acidic TME and generates reactive ROS through the sonodynamic effect. Fe³⁺ further catalyzes Fenton-like reactions to produce •OH, depleting intracellular GSH, suppressing glutathione GPX4, and culminating in ferroptotic death. Meanwhile, MFMP under US irradiation induces immunogenic cell death (ICD) and promotes the release of damage-associated molecular patterns (DAMPs), thereby facilitating a robust adaptive immune response. In addition, elevated levels of Mn²⁺ and ROS inflict mitochondrial damage and trigger the release of cytosolic dsDNA, which contributes to the activation of the cGAS–STING pathway and enhances innate antitumor immunity (Scheme 1c). Collectively, this facile co-assembly yields MFMP with strong sonodynamic activity and coordinated activation of innate and adaptive immune responses, producing pronounced antitumor immunotherapeutic effects.

Scheme 1.

Synthesis and therapeutic mechanism of MFMP. (a) MFMP was synthesized by one-step co-assembly of FeTCPP and MnCl₂ followed by CRGDK modification. (b) Sonodynamic activation enhances MFMP-induced ferroptosis and pyroptosis, triggers the cGAS-STING pathway, and (c) consequently promotes DCs maturation, antigen presentation, and T cell mediated antitumor immunity

Results and discussion

Synthesis and characterizations of Mn-Fe(TCPP) MOF

During the synthesis of Mn-Fe(TCPP) MOF nanoparticles, polyvinylpyrrolidone (PVP) was employed as a morphology-directing agent. The carbonyl groups (C = O) in the pyrrolidone rings of PVP can coordinate with metal ions (Fe³⁺ and Mn²⁺), enabling preferential adsorption of PVP onto specific crystal facets [19]. This coordination suppresses nanoparticle aggregation, maintains good dispersion, and effectively confines the particle size within the nanoscale range. To clearly define the crystal structure of Mn-Fe(TCPP) MOF, an advanced continuous rotation electron diffraction (cRED) technique was employed. As shown in Figure S1, three-dimensional reciprocal-lattice reconstruction from the cRED dataset identifies Mn-Fe(TCPP) MOF as orthorhombic with space group Cmmm. The unit-cell parameters are a = 14.130 Å, b = 16.860 Å, c = 20.750 Å with α = β = γ = 90° (Table S1). To clarify the structural impact of Mn²⁺ incorporation, we compared the crystallographic parameters of Mn–Fe(TCPP) MOF with those of Fe(TCPP) MOF (Table S2) [17]. Both materials crystallize in the orthorhombic system with the same space group (Cmmm), indicating that Mn²⁺ introduction does not alter the overall symmetry of the framework. However, significant changes in lattice parameters and unit-cell volume were observed. The unit-cell volume increased from 3712.7 ų for Fe(TCPP) MOF to 4943.31 ų for Mn–Fe(TCPP) MOF, accompanied by expansion along the a and c axes. Meanwhile, the calculated density decreased from 0.834 to 0.700 g·cm⁻³, suggesting a more open and less compact framework after Mn²⁺ incorporation. These results indicate that Mn²⁺ participates in the coordination network and reconstructs the framework topology, leading to an expanded pore structure. All non-hydrogen atoms were precisely located using SHELXT. Figure S2 presents the crystal structure model, showing (a) the top view and (b, c) side views along different crystallographic axes. The structure exhibits a well-ordered two-dimensional coordination network. Each Fe(III) center within the porphyrin ring (TCPP) coordinates axially with Mn(II) nodes through carboxylate oxygen atoms, forming an extended Mn-O-Fe linkage. The porphyrin units are periodically aligned along the a-b plane, while the Mn-O-Fe connections propagate along the c axis, generating a highly ordered and spatially confined framework. The rigid “ordered-confinement” architecture effectively maintains an inter-porphyrin distance of 7.34 Å, which is much larger than the effective π-π stacking distance (3.3–4.5 Å), thereby significantly suppressing π-π interactions. The powder X-ray diffraction (PXRD) peaks of Mn-Fe(TCPP) MOF are in excellent agreement with the simulated pattern, confirming the high phase purity of the product. Minor additional reflections observed in the 30–35° region may arise from trace impurity phases or preferred crystal orientation effects during sample preparation, but they do not affect the overall phase assignment (Figure S3).

Scanning electron microscopy (SEM) images revealed that the as prepared Mn-Fe(TCPP) MOF nanoparticles were well dispersed, exhibiting a uniform cubic morphology with an average size of 100 nm (Fig. 1a, b). Elemental mapping images and X-ray photoelectron spectroscopy (XPS) further confirmed the presence of C, N, O, Fe, and Mn, which were homogeneously distributed throughout the Mn-Fe(TCPP) MOF nanoparticles (Fig. 1c-d). In particular, the high-resolution Mn 2p spectrum exhibits characteristic features of Mn²⁺, with the Mn 2p_3/2 peak centered at 642.1 eV and the Mn 2p_1/2 peak at 653.6 eV, separated by 11.5 eV. A distinct satellite peak located 5–6 eV above the Mn 2p_3/2 main peak further confirms the presence of divalent manganese species (Fig. 1e) [20]. The Fe 2p XPS spectrum reveals two characteristic peaks at 711.6 eV (Fe 2p_3/2) and 725.1 eV (Fe 2p_1/2), confirming the presence of Fe³⁺ species (Fig. 1f). Inductively coupled plasma (ICP) analysis revealed that the Fe³⁺ and Mn²⁺ contents in the Mn–Fe(TCPP) MOF were 9.61% and 12.15%, respectively. The UV/Vis spectra (Fig. 1g) demonstrate that Mn-Fe(TCPP) MOF shares a similar spectral profile with FeTCPP, characterized by a single Soret band and three Q bands, confirming that Mn(II) coordinates with the carboxyl groups rather than the central nitrogen atoms of FeTCPP. Moreover, the Soret band of Mn-Fe(TCPP) MOF is markedly broadened compared to that of FeTCPP, which can be ascribed to enhanced electron delocalization induced by Mn(II) incorporation, leading to extended conjugation and a modified electronic structure. MOFs often exhibit pH-responsive degradation and drug-release behavior, which is crucial for biomedical applications. As shown in Figure S4, Mn–Fe(TCPP) MOF exhibited minimal metal ion release under physiological conditions (pH 7.4, GSH = 0 mM), whereas both Mn²⁺ and Fe³⁺ were rapidly released in the tumor-mimicking microenvironment (pH 5.5, GSH = 10 mM), indicating degradability responsive to GSH and acidic conditions. Notably, Mn²⁺ were released more rapidly than Fe³⁺ under acidic conditions. This preferential Mn release is likely associated with the relatively lower coordination stability of Mn–O bonds compared to the highly stabilized Fe–N coordination within the porphyrin ring. Under acidic conditions, protonation weakens the Mn–O interactions, and the presence of GSH further accelerates ligand exchange, leading to faster Mn dissolution. In contrast, Fe remains more tightly confined within the porphyrin macrocycle, resulting in comparatively slower release kinetics. To demonstrate the superior performance of the sonodynamic Mn–Fe(TCPP) MOF, Fe(TCPP) MOF was synthesized following a previously reported method and subsequently conjugated with the CRGDK targeting ligand. As shown in Figure S5, the FT-IR spectrum of Fe(TCPP)-MOF exhibited markedly attenuated peaks corresponding to the C = O stretching vibration (1696 cm^{− 1}) and the C–O stretching vibration (1275 cm^{− 1}), compared with FeTCPP, confirming the successful synthesis of the material. Furthermore, the PXRD results corroborated the successful synthesis of Fe(TCPP)-MOF (Figure S6). However, the poor water solubility, limited dispersibility, and inadequate tumor targeting capability of Fe(TCPP) MOF restrict its efficacy. In vitro and in vivo studies have demonstrated that the CRGDK peptide can enhance tumor targeting and penetration of nanomedicines, thereby increasing intratumoral accumulation and therapeutic efficacy [21]. Through the EDC/sulfo-NHS coupling reaction, the amino groups of CRGDK were covalently linked to the carboxyl groups on the surface of Fe(TCPP)-MOF, forming stable amide bonds. The characteristic peaks at 1645 cm^{− 1} (Amide I, C = O stretching) and 1536 cm^{− 1} (Amide II, N–H bending and C–N stretching) confirmed the presence of amide bonds, indicating the successful conjugation of CRGDK to Fe(TCPP)-MOF. For simplicity, CRGDK-modified Fe(TCPP)-MOF and Mn-Fe(TCPP)-MOF are denoted as FMP and MFMP, respectively. Similarly, FT-IR spectroscopy was conducted for FeTCPP, Mn-Fe(TCPP) MOF, CRGDK, and MFMP. In MFMP, the characteristic peaks corresponding to the C = O stretching vibration at 1690 cm^{− 1} and the C-O stretching vibration at 1272 cm^{− 1} were significantly reduced, confirming the effective coordination between Mn²⁺ and the carboxyl groups, while a new absorption band at 1629 cm^{− 1} indicated the formation of –CONH– bonds between CRGDK and Mn-Fe(TCPP) MOF (Fig. 1h). Analysis of the TG-DSC curves indicated that the CRGDK content grafted onto Mn-Fe(TCPP) MOF was approximately 10.94 wt%, thereby confirming the successful surface modification (Fig. 1i). Dynamic light scattering (DLS) measurements revealed that Mn-Fe (TCPP) MOF nanoparticles exhibited a hydrodynamic diameter of 110 nm (Fig. 1j) with a zeta potential of -18.2 mV (Fig. 1k). Upon CRGDK modification, the particle size increased to 160 nm and the zeta potential shifted to -14.7 mV. Moreover, Mn-Fe(TCPP) MOF demonstrated excellent stability in aqueous environments, retaining its morphology and crystallinity even after 7 days of dispersion in water (Figure S7). N₂ adsorption analysis revealed that Mn-Fe(TCPP) MOF possesses a BET surface area of 426 m²/g (Fig. 1l). The high surface area and porous structure facilitate oxygen accessibility, enhance its adsorption and storage capacity, and thereby improve oxygen utilization, which is favorable for the subsequent generation of reactive ROS.

Fig. 1.

(a, b) SEM images of Mn–Fe(TCPP) MOF. (c) TEM image with corresponding elemental maps. (d) XPS survey spectrum and (e, f) high-resolution Mn 2p and Fe 2p spectra of Mn–Fe(TCPP) MOF. (g) UV–Vis absorption spectra of FeTCPP and Mn–Fe(TCPP) MOF. (h) FT-IR spectra of FeTCPP, Mn–Fe(TCPP) MOF, CRGDK, and MFMP. (i) TGA curves of FeTCPP, Mn–Fe(TCPP) MOF, and MFMP. (j) DLS size distributions of Mn–Fe(TCPP) MOF and MFMP. (k) Zeta potentials of FeTCPP, Mn–Fe(TCPP) MOF, and MFMP. (l) N₂ adsorption–desorption isotherm of Mn–Fe(TCPP) MOF

Mechanism and sonodynamic ROS generation of MFMP

The “ordered-confinement” architecture of the Mn-Fe(TCPP) MOF effectively suppresses ACQ that arises from π-π stacking and lowers the probability of nonradiative decay pathways, such as internal conversion and solvent coupled vibrational relaxation, thereby reducing thermally dominated energy loss. As illustrated in Figure. 2a, the introduction of Mn²⁺ strengthens stronger spin–orbit coupling, thereby promoting intersystem crossing (ISC) and suppressing non-radiative decay (NRD) pathways. In order to systematically elucidate the intrinsic correlation between structure and properties, we conducted a comparative analysis of the photoluminescence characteristics and ROS generation efficiency of FeTCPP and MFMP. As shown in Figure S8a, FeTCPP exhibited intense fluorescence emission when dissolved in DMF, while its fluorescence intensity markedly decreased as the water content increased to 90%, which was mainly attributed to the ACQ effect induced by π–π stacking. In contrast, MFMP showed relatively weak fluorescence in DMF, but its emission intensity significantly increased with the rise of the water fraction, indicating that the ACQ effect was effectively suppressed in this system (Figure S8b). Therefore, FeTCPP suffers from pronounced ACQ in the condensed state. In contrast to its quenched emission, the MFMP exhibits markedly boosted fluorescence intensity, corroborating that its unique ordered-confined architecture efficiently suppresses NRD pathways (Fig. 2b). To assess the sonodynamic performance of MFMP, 1,3-diphenylisobenzofuran (DPBF) and methylene blue (MB) were employed as probes to monitor and quantify the generation of ¹O₂ and •OH, respectively. For ¹O₂ detection, a solution containing MFMP and DPBF was subjected to US (1.0 MHz, 1.5 W cm^− 2, 50% duty cycle, 1 min). The characteristic absorption of DPBF at 410 nm decreased progressively with sonication time (Fig. 2c), indicating that US-activated MFMP produced ¹O₂, which oxidized DPBF to the colorless 1,2-dibenzoylbenzene. Compared with the other groups (US, FeTCPP, and FMP), MFMP induced the most pronounced DPBF bleaching (Fig. 2e). Subsequently, to detect •OH generation, a solution of MB and MFMP was irradiated under the same US conditions (1.0 MHz, 1.5 W cm^− 2, 50% duty cycle, 1 min). The MB absorption band at 668 nm gradually declined with increasing sonication time (Fig. 2d), and the decay became more pronounced upon addition of H₂O₂ (Fig. 2f). In aggregate, these results demonstrate that US-induced MFMP continuously generates •OH, reacts with MB, and progressively decolorizes the solution over time. To further verify US-triggered ROS generation, ESR spin trapping was employed with TEMP capturing ¹O₂ and DMPO capturing •OH. Under identical US conditions, the blank control showed virtually no signal, FeTCPP produced only a weak response, FMP displayed a markedly amplified signal, and the MFMP group exhibited the strongest signal with a characteristic 1:1:1 triplet, indicative of efficient ¹O₂ production (Fig. 2g). Upon the introduction of H₂O₂ with DMPO as the spin trapping agent, the blank control exhibited negligible signal intensity under identical US conditions, indicating minimal spontaneous •OH generation. Similarly, the FeTCPP + H₂O₂ + US group did not display a discernible quartet signal, suggesting limited hydroxyl radical formation in the absence of the ordered framework structure. In contrast, the FMP + H₂O₂ + US group generated a characteristic 1:2:2:1 quartet signal, while the MFMP + H₂O₂ + US group exhibited the strongest quartet signal among all groups (Fig. 2h), indicating markedly enhanced •OH production. These results indicate that the ordered and spatially confined coordination framework effectively mitigates π–π stacking induced ACQ, thereby promoting more efficient energy transfer and enhanced ROS generation. These synergistic effects collectively endowed MFMP with superior sonodynamic activity compared to FMP. To elucidate the mechanism underlying the enhanced ROS generation, we performed time-dependent density functional theory (TD-DFT) calculations on the geometry-optimized structures of FeTCPP and Mn-Fe(TCPP) MOF. As shown in Fig. 2i, Mn coordination and framework formation markedly depress the LUMO level from − 3.441 to -3.929 eV, while the HOMO remains essentially unchanged. Consequently, the band gap narrows to 2.03 eV for Mn-Fe(TCPP) MOF, compared with 2.52 eV for FeTCPP. Moreover, the optical excitation energy (S₁) of Mn-Fe(TCPP) MOF slightly decreases from 2.95 to 2.83 eV, whereas the triplet-state energy (T₁) increases from 2.13 to 2.28 eV. Accordingly, the singlet-triplet energy gap (ΔE_ST) is reduced from 0.82 to 0.55 eV, which facilitates ISC and promotes the population of triplet excitons. Such electronic modulation favors efficient charge delocalization and spin–orbit coupling, accelerating the conversion from the singlet excited state to the triplet state and thereby significantly enhancing sonodynamic ROS generation in Mn-Fe(TCPP) MOF.

Fig. 2.

(a) Schematic illustration of the sonodynamic mechanism of MFMP. (b) Photoluminescence spectra of FeTCPP and MFMP. (c, d) Time-dependent UV–Vis absorption spectra of DPBF and MB solutions containing MFMP under US irradiation. (e, f) Normalized absorbance decay curves of DPBF and MB in different groups. (g, h) ESR spectra using TEMP and DMPO as trapping agents for ¹O₂ and •OH, respectively (Mean ± SD, n = 3). (i) DFT calculated frontier molecular orbitals and energy level diagrams of FeTCPP and Mn-Fe(TCPP) MOF

In vitro therapeutic efficacy of MFMP mediated SDT

Given the excellent ROS generating capability of MFMP under US irradiation, we investigated the in vitro sonodynamic effects of MFMP, wherein both ROS mediated cytotoxicity and cavitation induced mechanical damage contribute to the therapeutic efficacy. To optimize the US irradiation conditions, intracellular ROS generation was evaluated under different exposure times and power densities. As shown in Figure S9a, ROS levels increased markedly when the irradiation time was extended from 1 to 5 min, while no substantial additional enhancement was observed at 10–15 min. Similarly, as illustrated in Figure S9b, ROS production significantly increased at 1.5 W·cm^-2 compared with 0.5 W·cm^-2, whereas further increasing the power density to 2.0 W·cm^-2 did not lead to a proportional rise in ROS levels. Considering both efficiency and potential safety, 5 min and 1.50 W·cm^-2 were selected as the optimized US parameters for subsequent experiments. As illustrated in Fig. 3a, US-irradiated MFMP induces inertial cavitation in the liquid medium, leading to bubble oscillation, collapse, and the release of mechanical forces, which cause severe disruption of the cell membrane. Concurrently, intracellular MFMP generates abundant cytotoxic ROS, exacerbating mitochondrial damage and ultimately triggering pyroptosis. Efficient intracellular internalization is a prerequisite for sonodynamic ROS generation and subsequent ferroptosis/STING activation. Therefore, we labeled the nanoparticles with FITC (FITC@MFMP) to quantify cellular uptake (Figure S10). Confocal laser scanning microscopy (CLSM) was employed to observe the uptake of FeTCPP, FMP, and MFMP by 4T1 cells after 12 h of incubation. As shown in Figure S11a, cells treated with MFMP exhibited markedly enhanced fluorescence intensity. Quantitative analysis of fluorescence signals revealed that the intensity in the FITC@MFMP group was 13.4-fold higher than that in the FITC@FeTCPP group, a finding further confirmed by flow cytometry analysis (Figure S11b). To further investigate the sonodynamic therapeutic effects of MFMP at the cellular level under US irradiation, 4T1 cells were co-incubated for 12 h with varying concentrations (0, 6.25, 12.5, 25, 50, 100, and 200 µg/mL) of FeTCPP, FMP, or MFMP. The cytotoxicity of these materials was subsequently evaluated using the standard cell counting kit-8 (CCK-8)assay. Within the concentration range of 0–200 µg mL^-1, FeTCPP exhibited negligible cytotoxicity toward 4T1 cells, irrespective of US exposure (Figure S12). In contrast, both FMP and MFMP displayed concentration-dependent cytotoxicity under US irradiation (1.0 MHz, 1.5 W/cm², 1 min, 50% duty cycle). Notably, at 100 µg/mL, MFMP induced significantly higher cytotoxicity than FMP, achieving a cell killing efficiency exceeding 80% (Fig. 3b). Subsequently, intracellular ROS generation was evaluated employing 2′,7′-Dichlorofluorescin diacetate (DCFH-DA) as a fluorescent probe. DCFH-DA itself is non-fluorescent, but upon oxidation by ROS within cells, it is converted into the highly fluorescent compound 2′,7′-dichlorofluorescein (DCF) [22]. As shown in Fig. 3c, the MFMP + US treatment group displayed markedly stronger green fluorescence compared with the other groups, with an intensity 2.34-fold higher than that of MFMP (Fig. 3d). In contrast, MFMP alone exhibited only a weak fluorescence signal, confirming that the combination of MFMP with US significantly enhances ¹O₂ generation and amplifies intracellular oxidative stress. Notably, MFMP + US treatment induced 4T1 cell swelling and large membrane bubbles, hallmarks of pyroptosis [23], whereas no obvious morphological changes were observed in the other groups (Fig. 3e). In addition, Figure S13 shows that cells treated with MFMP + US exhibit pronounced cell swelling accompanied by characteristic membrane bubbling. To verify that the ROS burst induced by MFMP under US irradiation triggers pyroptosis, Western blot analysis was performed to assess the expression of GSDM-related proteins, thereby elucidating the underlying mechanism (Fig. 3a). As shown in Fig. 3f, the MFMP + US group exhibited increased levels of cleaved caspase-1 and the N-terminal fragment of GSDMD (GSDMD-N), accompanied by a decrease in full-length GSDMD, indicating the activation of a caspase-1/GSDMD-dependent pyroptotic pathway. Subsequently, the expression of characteristic pyroptosis-related markers, including IL-1β, LDH, and IL-18, was quantitatively evaluated. As shown in Figure S14, the levels of IL-1β, LDH, and IL-18 in the supernatant of 4T1 cells treated with MFMP + US were the highest among all groups, exceeding those of the control + US group by approximately 2.3-fold, 2.1-fold, and 2.5-fold, respectively. Apoptosis and mechanical damage were assessed by flow cytometry using a standard Annexin V/PI co-staining protocol. As shown in Fig. 3g, MFMP + US treatment induced pronounced mechanical injury (19.6%) and significantly enhanced apoptosis (43.1%). The apoptosis rate substantially exceeded that of the Control + US (5.89%), FeTCPP + US (3.30%), and FMP + US (31.06%) groups. This demonstrates that MFMP can induce mechanical damage and pyroptosis in tumor cells under US.

Fig. 3.

Sonodynamic cytotoxicity and pyroptosis induction by MFMP. (a) Schematic illustration of the MFMP-induced sonodynamic process, including cavitation effect, mechanical injury, and oxidative damage leading to pyroptosis. (b) Cell viability of 4T1 cells treated with different formulations (I: Control, II: FeTCPP, III: FMP, IV: MFMP) with or without US irradiation (Mean ± SD, n=3, ***p<0.001). (c) Intracellular ROS generation detected by DCFH-DA staining. Scale bar = 20 μm. (d) Quantitative analysis of intracellular fluorescence intensity (Mean ± SD, n=3, **p<0.01, ***p<0.001). (e) Bright-field images of 4T1 cells showing cell morphological changes after various treatments. Scale bar = 20 μm. (f) Western blot analysis of GSDMD, GSDMD-N, and caspase-1 in different groups. (g) Annexin V–FITC/PI flow cytometry of live/dead cells

In vitro induction of ferroptosis by MFMP

Motivated by MFMP’s remarkable capacity to trigger pyroptosis under US, we hypothesize that US may potentiate its ferroptotic effect on tumor cells. Upon US irradiation, MFMP produces abundant ROS, while the released Fe³⁺ is reduced to Fe²⁺ by GSH in the acidic TME, thereby catalyzing hydrogen peroxide decomposition and amplifying ROS accumulation. The concomitant depletion of GSH further impairs GPX4 activity and disrupts phospholipid hydroperoxides (LPO) detoxification, ultimately promoting ferroptosis (Fig. 4a). As shown in Fig. 4b and Figure S15a, MFMP + US treatment resulted in an 86.3% reduction of GSH compared with the control group, which was further validated by flow cytometry (Figure S15b). Since GSH is a critical substrate of GPX4, we next examined GPX4 expression. As shown in Fig. 4c, d and Figure S16 MFMP + US treatment markedly reduced GPX4 levels compared with the control, which was further validated in Figure S17. The depletion of GSH together with GPX4 downregulation suggested impaired lipid peroxide detoxification. Consistently, malondialdehyde (MDA) assays revealed a significant increase in intracellular MDA levels under US irradiation (Fig. 4e), confirming excessive lipid peroxide accumulation and thereby amplifying ferroptosis. Ferroptosis and SDT-induced ICD promote the release of DAMPs, which act as tumor-associated antigens, including calreticulin (CRT) exposure on the cell surface, HMGB1 translocation, and ATP secretion. Western blot (Fig. 4f and Fig. S18) and Enzyme-linked immunosorbent assay (ELISA) results (Fig. 4g-h) demonstrated that MFMP + US treatment markedly reduced intracellular HMGB1 and ATP levels while enhancing extracellular ATP release. Flow cytometry further revealed a CRT-positive rate of 39.49% in the MFMP + US group, 17.3-fold higher than that of the control (Fig. 4i). Although US markedly enhances ROS generation and ferroptotic, the MFMP nanoplatform itself can induce a moderate level of ICD under intracellular. The acidic and reductive environment facilitates partial metal ion release, which may lead to basal redox activity and mild lipid peroxidation. Even in the absence of US stimulation, this process may trigger early ferroptotic events and mitochondrial stress, thereby promoting the exposure of DAMPs. These results establish that sonodynamic activation of MFMP potentiates ferroptosis, thereby driving a potent antitumor immune response.

Fig. 4.

Ferroptosis and ICD induced by MFMP under US irradiation. (a) Schematic illustration of MFMP-mediated ferroptosis and ICD processes. (b, c) Confocal fluorescence images of intracellular GSH and GPX4 in 4T1 cells treated with different formulations (I: Control, II: FeTCPP, III: FMP, IV: MFMP) with or without US irradiation. Scale bar = 40 μm. (d) Western blot analysis of GPX4 expression confirming ferroptosis activation. (e) Quantification of malondialdehyde (MDA) levels indicating enhanced lipid peroxidation upon MFMP + US treatment (ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.001). (f) Western blot analysis of HMGB1. (g, h) Quantification of extracellular and intracellular ATP levels (ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.001). (i) Flow-cytometric analysis of CRT exposure on the cell surface

In vitro activation of the cGAS-STING pathway by MFMP

DNA released from sonodynamic-induced mitochondrial damage, together with Mn²⁺, activates cGAS and drives the specific catalytic synthesis of 2′,3′-cGAMP. Acting as a second messenger, cGAMP binds to and activates the adaptor protein STING [9]. Upon activation, STING recruits kinases such as TANK-binding kinase 1 (TBK1) and IKB kinase (IKK), leading to the phosphorylation of nuclear factor κB (NF-κB) and IFN regulatory factor 3 (IRF3) which subsequently translocate into the nucleus to promote the transcription of type I interferons (IFN-I) and pro-inflammatory cytokines [24]. Phosphorylated IRF3 subsequently dimerizes and translocates into the nucleus to initiate transcription of interferon-β (IFN-β). Meanwhile, NF-κB enters the nucleus and cooperates with other transcription factors to amplify the expression of pro-inflammatory and immune-related genes, including TNF and IL-6 (Fig. 5a) [25]. To verify whether MFMP induces cytosolic dsDNA leakage, a quantitative ELISA assay was performed. As shown in Figure S19, compared with the control group (0.23 ± 0.06 ng/µL), the cytosolic dsDNA level in MFMP + US treated cells reached 5.13 ± 0.48 ng/µL, corresponding to an approximate 22.3-fold increase. In contrast, the dsDNA levels in the FeTCPP + US and FMP + US groups showed only moderate increases (0.59 ± 0.12 ng/µL and 2.13 ± 0.35 ng/µL, respectively), indicating that Mn²⁺ incorporation, together with ferroptosis and pyroptosis, synergistically promotes dsDNA leakage under sonodynamic activation. Given that dsDNA is a critical activator of the cGAS-STING signaling pathway, we examined the expression levels of phosphorylated IRF3 (p-IRF3), phosphorylated TBK1 (p-TBK1), phosphorylated STING (p-STING), phosphorylated NF-κB (p-NF-κB), IKK, and cGAS. Notably, both the Western blot results (Fig. 5b) and the corresponding quantitative analysis (Figure S20) revealed substantial upregulation of these phosphorylated proteins in cells treated with MFMP + US, indicating efficient nanoparticle internalization and robust activation of the cGAS–STING pathway, thereby further confirming its strong immunostimulatory potential. Consistent with the activation of the cGAS-STING signaling cascade, the secretion of IFN-β, TNF, and IL-6 was markedly elevated in the MFMP + US group compared with the control groups, as determined by ELISA (Fig. 5c-e). These results further confirmed that MFMP under US stimulation effectively triggered innate immune signaling and promoted proinflammatory cytokine release. Subsequently, we established a Transwell co-culture model to evaluate the effect of MFMP nanoparticles on DCs maturation via activation of the cGAS-STING pathway in vitro (Fig. 5f). As shown in Fig. 5g and h, CD80 and CD86 were selected as representative costimulatory markers of DCs maturation and activation. Flow cytometry analysis demonstrated that 38.2% of DCs in the MFMP-treated group reached a mature state, which was 3.99-fold higher than that observed in the control + US group.

Fig. 5.

Activation of the cGAS–STING pathway and DCs maturation induced by MFMP under US irradiation. (a) Schematic illustration of the mechanism by which MFMP activates the cGAS–STING signaling cascade. (b) Western blot analysis of cGAS–STING pathway related proteins in 4T1 cells after different treatments (I: Control, II: FeTCPP, III: FMP, IV: MFMP) with or without US irradiation. (c–e) ELISA quantification of IFN-β, TNF-α, and IL-6 expression levels (ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.001). (f) Schematic of the co-culture experiment: dying 4T1 cells after treatment were co-incubated with imDCs, followed by assessment of DCs maturation via CD80/CD86 expression. (g) Flow-cytometric analysis of CD80⁺ CD86⁺ DCs populations (***p < 0.001). (h) Quantitative statistics of CD80⁺ CD86⁺ DCs (***p < 0.001)

Transcriptome analysis of MFMP under US irradiation

In order to systematically elucidate the programmed cell death and immune activation mechanisms triggered by MFMP under US irradiation, we performed RNA-seq on tumor tissues from the MFMP + US treatment group. A total of 27,096 genes were analyzed in 4T1 cells. The volcano plot (Fig. 6a) shows that 196 differentially expressed genes (DEGs) were identified between the MFMP + US and control groups (| log2(fold change) | ≥ 1, p < 0.05). Among them, 82 genes were significantly upregulated, and 114 genes were significantly downregulated, indicating substantial transcriptional differences between the two groups. Gene Ontology (GO) enrichment analysis revealed that MFMP treatment significantly activated multiple innate immune pathways, particularly those associated with STING signaling, including the type I interferon response, NF-κB signaling, and Toll-like receptor mediated pathways (Fig. 6b). In addition, cellular responses related to interferon-β, cell killing, and tumor immune regulation were also enriched. These pathways are crucial for coordinating the activation, adhesion, and cytotoxic activity of immune cells within the tumor microenvironment, demonstrating that MFMP + US effectively enhances antitumor immunity [26]. Pathway enrichment analysis revealed that MFMP + US treatment significantly modulated several key immune and metabolic pathways (p < 0.05), particularly those associated with positive regulation of IL-1β production, glycolysis, ROS metabolism, leukocyte migration, and STAT signaling (Fig. 6c). The upregulation of glycolysis and pyruvate metabolism suggests that MFMP + US may enhance immune responses by boosting cellular energy metabolism. Furthermore, the enrichment of leukocyte migration pathways indicates that MFMP + US could facilitate the activation and infiltration of immune cells within the tumor microenvironment. Notably, the downregulation of pathways involved in the negative regulation of blood vessel endothelial cell migration suggests that MFMP + US may inhibit tumor angiogenesis and metastasis, further supporting its potential antitumor effects. As shown in Fig. 6d, the heatmap from transcriptome sequencing revealed a significant upregulation of several IFN-β genes, including Ifnb1, Ifna5, Ifit3, and Isg15, in the MFMP + US group compared to the control, indicating a strong induction of the IFN-I. Additionally, key genes associated with Toll-like receptors (such as Tlr3, Tlr9, and Tlr13) were also upregulated in the MFMP + US group, suggesting enhanced immune recognition and signaling. Furthermore, proinflammatory cytokines such as Tnf, Cxcl1, and Cxcl10 showed significant upregulation in response to MFMP + US treatment, reinforcing the activation of inflammatory pathways. These findings suggest that MFMP enhances the activation of STING, thereby amplifying the secretion of type I interferon and other proinflammatory cytokines. The mRNA expression levels of four key genes closely associated with the cGAS-STING pathway (Ifna4, Isg15, Tlr1 and Cxcl10) were validated using reverse transcription quantitative real-time PCR (RT-qPCR) (Fig. 6e). The results showed a significant upregulation of these immune-related genes in the MFMP + US group compared to the Control group.

Fig. 6.

Transcriptomic analysis of MFMP under US irradiation. (a) Volcano plot of differentially expressed genes in the MFMP + US group compared with the Control group. (b) GO analysis after different treatments. (c) Pathway analysis in the MFMP + US. (d) Heatmap of differentially expressed proteins. (e) The expression of the related genes (Ifna4, Isg15, Tlr1, and Cxcl10) in different treatment groups was analyzed by RT-qPCR (Mean ± SD, n = 3, ***p < 0.001)

In vivo antitumor effect

Encouraged by the excellent in vitro results, we next investigated the in vivo antitumor efficacy of MFMP. Prior to in vivo studies, we conducted comprehensive toxicological profiling of MFMP nanoparticles to establish biosafety. We first evaluated hemocompatibility. As shown in Figure S21, MFMP nanocatalysts induced < 5% hemolysis across 25–400 µg mL^-1, satisfying the threshold for intravenous administration. We then assessed acute and subacute toxicity by monitoring body weight and performing hematoxylin and eosin (H&E) histopathology of major organs. As illustrated in Figure S22 and Figure S23, tissue sections from MFMP-treated mice were unremarkable compared with untreated controls, and body-weight trajectories remained stable. Routine hematology further revealed no appreciable abnormalities, supporting the favorable biosafety profile of MFMP (Figure S24). To assess the tumor-targeting capability of CRGDK-functionalized nanoparticles (MFMP), we prepared FITC-labeled formulations and randomized mice into six groups (Control, FITC@FeTCPP, FITC@FeTCPP MOF, FITC@Mn–Fe(TCPP) MOF, FITC@FMP, and FITC@MFMP). Following tail-vein injection, real-time fluorescence imaging was performed at 0, 6, 12, 24, 36, 48, 60, and 72 h to monitor intratumoral accumulation. As shown in Figure S25, groups bearing the active targeting ligand (FMP and MFMP) exhibited discernible tumor signals by 6 h. The fluorescence intensity increased over time, reached a maximum at approximately 24 h, and persisted for more than 60 h. At all-time points, MFMP showed significantly higher tumor fluorescence than Mn-Fe(TCPP) MOF, indicating superior targeting efficiency and prolonged retention. We then readministered the formulation and performed ex vivo imaging at 24 h, which confirmed tumor accumulation for both FMP and MFMP (Figure S26). Based on these biodistribution data, 24 h post-injection was selected as the optimal time to initiate US treatment. Subsequently, pharmacokinetic analysis was performed. As shown in Figure S27, the blood concentration-time profiles of Fe and Mn ions released from MFMP followed a biexponential decay pattern, indicative of a two-compartment pharmacokinetic model. The Fe ion concentration decreased rapidly within the first few hours, with a calculated half-life (t_1/2) of 5.11 h and an excellent correlation (R² = 0.94), suggesting a fast distribution phase followed by slow elimination. Similarly, Mn ions exhibited a comparable kinetic trend with a slightly shorter half-life (t_1/2 = 4.22 h, R² = 0.93), reflecting their rapid clearance from systemic circulation. These results demonstrate that Mn–Fe(TCPP) MOF undergoes efficient degradation and metal ion release in vivo, supporting its biodegradability and biosafety for subsequent therapeutic applications.As illustrated in Fig. 7a, an orthotopic 4T1 model was established by implanting 1 × 10⁵ 4T1 cells into the right mammary fat pad of BALB/c mice. 48 tumor-bearing mice were then randomized into eight groups to receive Control (I), FeTCPP (II), FMP (III), or MFMP (IV) with or without US irradiation. Based on the tumor growth curves (Fig. 7b, Figure S28), tumor volumes in the Control, FeTCPP, and FMP groups increased rapidly irrespective of US exposure. By contrast, MFMP combined with US produced pronounced tumor suppression, highlighting a synergistic US activated effect. Terminal tumor weights and representative tumor images were consistent with these findings (Fig. 7c, d). Western blotting and immunofluorescence showed that MFMP, upon US activation, strongly induces ferroptosis and cGAS-STING pathway activation (Fig. 7e, Figure S29). TUNEL staining revealed sparse apoptotic nuclei in Control and FeTCPP irrespective of US, moderate positivity in FMP, and the strongest signals in MFMP, which were further enhanced by US (Fig. 7f). Consistently, H&E sections from MFMP + US exhibited reduced cellularity, marked nuclear pyknosis/karyorrhexis, and disrupted tumor architecture, whereas Control and FeTCPP tissues remained largely viable (Fig. 7g). Collectively, these findings indicate that US activated MFMP robustly promotes apoptosis and histopathological tumor damage.

Fig. 7.

In vivo antitumor efficacy of MFMP combined with US irradiation. (a) Schematic illustration of the treatment schedule for the 4T1 tumor-bearing mouse model. (b) Tumor growth curves of mice receiving different treatments. (c) Tumor weights of each group at Day 14, showing that MFMP + US treatment achieved the most significant tumor inhibition. (*p < 0.05; ***p < 0.001) (d) Photographs of tumors undergoing different treatments on day 14. (e) Western blot analysis of GPX4 and STING expression in tumor tissues. (f) TUNEL staining of tumor sections. (g) H&E staining of tumor slices. Scale bar = 50 μm

In vivo immune activation

Pyroptosis and ferroptosis elicit robust inflammatory responses within the TME, thereby promoting immune-cell recruitment and activation, converting immunologically “cold” tumors into “hot” tumors, and potentiating antitumor immune responses. Accordingly, we further interrogated the immune responses elicited by MFMP in combination with US. After 14 days of treatment, tumor-draining lymph nodes (TDLNs) and tumor tissues from each group were harvested, processed into single-cell suspensions, and subjected to flow-cytometric analysis (Figure S30). Given the central role of DCs maturation in initiating adaptive immunity, we first quantified DCs maturation in the TDLNs. In the absence of US, the proportion of mature DCs increased from 3.2% in Group I to 21.8% in Group IV, and with US Group IV rose further to 27.3%, approximately 8.5-fold higher than the control and 1.25-fold above IV(-US) (Fig. 8a, b). As shown in Fig. 8c and d, MFMP markedly increased intratumoral T-cell infiltration to 30.7% in Group IV (-US), approximately 3-fold higher than FMP, and US activation further potentiated this effect, consistent with US-mediated pyroptosis. In addition, splenocyte suspensions were collected to assess the induction of memory T cells. The frequency of effector memory CD8⁺ T cells (CD8⁺, CD44⁺, CD62L^-) exhibited a sustained and significant increase (Fig. 8e, f). Given that CD8⁺ T cells mediate cytotoxicity and produce pro-inflammatory mediators, we quantified cytokine secretion by ELISA, including IL-6, IL-12, IFN-γ, and TNF-α. As shown in Fig. 8g–j, nanoparticle (FMP and MFMP) treatment induced a progressive increase in these cytokines, which was further augmented by US, with IV(+ US) exhibiting the highest levels. These findings are consistent with tumoral Mn²⁺/Fe³⁺ accumulation and US-dependent DNA damage, which likely engage innate sensing pathways (e.g., cGAS–STING) and thereby amplify T cells derived pro-inflammatory cytokine production.

Fig. 8.

MFMP combined with US irradiation promotes DCs maturation and enhances antitumor immune activation. (a) Flow-cytometric analysis of CD80⁺CD86⁺ DCs in TDLNs after different treatments (I: Control, II: FeTCPP, III: FMP, IV: MFMP) with or without US irradiation. Flow-cytometric plots and statistical quantification of DCs (a, b), CD8⁺ T (c, d), CD44⁺ effector/memory T cells (e, f) in mice (***p < 0.001). IL-6 (g), IL-12 (h), IFN-γ (i), and TNF-α (j) levels in tumor tissues after different treatments (***p < 0.001)

In vivo antimetastatic efficacy

Beyond eliciting potent antitumor immunity, MFMP plays a pivotal role in establishing immune memory to prevent distant metastasis. Accordingly, we established pulmonary metastasis and contralateral distant-tumor models to evaluate its in vivo memory response. As illustrated in Fig. 9a, after the assigned treatments, the residual 4T1 tumor mass was surgically excised, followed by intravenous injection of 5 × 10⁵ 4T1 cells. As shown in Fig. 9b, the control and US-alone groups (I) exhibited multiple, poorly demarcated gray–white nodules with focal hemorrhage. The lesion burden decreased in the FMP group (III). In contrast, the MFMP + US group (IV + US) showed only scattered or minute foci, with largely preserved alveolar architecture and markedly attenuated interstitial reaction and inflammatory infiltration. H&E staining further confirmed that tumor cell nest–like proliferation was markedly suppressed in the IV + US group, with a sharply defined interface between tumor tissue and pulmonary parenchyma. Inflammatory exudation and perivascular cuffing were also substantially reduced (Fig. 9c). These findings indicate that MFMP synergized with US to effectively limit the formation of metastatic nodules and decrease lung metastatic load. Subsequently, we assessed MFMP’s capacity to suppress distant tumor in a bilateral orthotopic model. Female BALB/c mice were orthotopically inoculated with 1 × 10⁶ 4T1 cells in the right mammary fat pad (primary tumor). After 7 days, 1 × 10⁶ 4T1 cells were implanted into the left mammary fat pad to establish the contralateral distant tumor (Fig. 9d). The tumor growth curves showed that, relative to control group (I), the FeTCPP group (II) scarcely slowed tumor progression. Although FMP (III) modestly delayed growth of the primary tumor, its therapeutic effect on the distant tumor was minimal (Fig. 9e-f, Figure S31-32). Notably, MFMP (IV) suppressed distant tumor growth, consistent with the H&E staining results (Fig. 9g). Collectively, these findings indicate that, under US exposure, MFMP induces antitumor immunity and elicits a pronounced abscopal effect, leading to regression of distant metastatic lesions.

Fig. 9.

In vivo evaluation of MFMP-mediated sonodynamic immunotherapy in lung-metastasis and distant tumor models. (a) Treatment schedule for metastasis inhibition of the 4T1 breast tumor in BALB/c mice. (b) Representative photographs of lungs from mice after the indicated treatments (I: Control, II: FeTCPP, III: FMP, IV: MFMP) with or without US. (c) H&E staining of the lung tissues. (d) Schematic of the bilateral 4T1 model for distant tumors. (e) Growth curves of distant tumors under different treatments. (f) Representative photographs of excised distant tumors from each group. (g) H&E staining of distant tumor sections. Scale bar = 50 μm

Conclusion

In summary, we have developed a porphyrin-oriented bimetallic Mn-Fe(TCPP) MOF with an ordered–confinement architecture, further functionalized with the tumor targeting peptide CRGDK, to achieve synergistic ferroptosis and cGAS-STING mediated cancer immunotherapy. Specifically, the precisely ordered Mn-O-Fe coordination network effectively suppresses ACQ, enhances intersystem crossing, and markedly boosts ROS generation under US irradiation. MFMP releases Fe³⁺ and Mn²⁺ in the mildly acidic tumor microenvironment, triggering Fenton-like reactions and oxidative stress that induce GSH depletion, GPX4 inhibition, and ferroptotic cell death. Concurrently, Mn²⁺ mediated mitochondrial damage and dsDNA release activate the cGAS-STING signaling pathway, amplifying IFN-I responses and facilitating DCs maturation and T-cell activation. This dual immunometabolic regulation effectively converts immunologically “cold” tumors into “hot” ones, eliciting potent and durable antitumor immunity with significant inhibition of tumor growth, recurrence, and metastasis. Collectively, this work provides a generalizable strategy for integrating ferroptosis with innate immune activation through rationally engineered sonodynamic MOFs, offering new opportunities for precise and effective sono-immunotherapy against refractory cancers.

Experimental section

Materials

All reagents and chemicals were of analytical grade or higher and were used as received without further purification. Iron (III) meso-Tetra(4-carboxyphenyl) porphine chloride (FeTCPP), Tetrakis (4-carboxyphenyl) porphyrin (TCPP), N,N-Dimethylformamide (DMF) were purchased from MedChemExpress (New Jersey, USA). Manganese (II) chloride (MnCl₂) were bought from Aladdin (Shanghai, China). CendR peptide was obtained from Apeptide (Shanghai, China).

Cell lines and animals

Murine mononuclear macrophage cell line RAW264.7, Murine breast cancer cell line 4T1 were all purchased from the Fu Heng Biology (Shanghai, China). BALB/c mice (18–25 g) were brought from Yanbian University. Animal experimental protocols were reviewed and approved by Laboratory Animal the Ethics Committee Yanbian University (YD20250514003).

Characterization

Scanning electron microscopeas (SEM, GeminiSEM 300, ZEISS, Germany) utilized to detect the size and morphology of NPs. Transmission electron microscopy (TEM, JEM-F200, JEOL, Japan) was used to characterize the morphology of the sample. The X-ray photoelectron spectra (XPS, Krayos AXIS Ultra DLD, Shimadzu, Japan) were recorded using Mg Kα radiation (1200 eV) as the excitation source, with the samples mounted on silicon slices. The absorption spectrums were analyzed by UV-vis (U-3310, Hitachi, Japan). The automatic specific surface and pore size distribution analyzer (ASAP 2460, Micromeritics, USA) was used to measure nitrogen adsorption–desorption isotherms. Fourier transform infrared spectroscopy (FT-IR) spectra were obtained on a SHIMADZU IRTRACER-100. Dynamic light scattering (DLS) experiment was measured by Malvern Zeta Sizer-Nano ZS90 instrument at 25 °C. TG-DSC (STA200, HITACHI, Japan) measurements were performed on a simultaneous thermal analyzer.

Synthesis of Mn-Fe(TCPP) MOF

Mn–Fe(TCPP) MOF nanoparticles were synthesized through a one-step coordination self-assembly strategy. PVP (150 mg) was first dissolved in DMF (30 mL) to prepare a 5 mg mL^− 1 PVP solution under magnetic stirring for 30 min. Subsequently, 10 mL of MnCl₂·4H₂O solution in deionized water (10 mM) was introduced and stirred for another 5 min. Then, 10 mL of FeTCPPCl solution in DMF (1.3 mM) was slowly added dropwise (1 mL min^− 1) to the above mixture. The reaction mixture was allowed to stand at 60 ℃ for 12 h to complete the assembly. The resulting precipitates were collected by centrifugation (11000 rpm, 15 min) and washed three times with ethanol to remove unreacted species and residual PVP. The obtained Mn–Fe(TCPP) MOF was dried at 60 °C for 8 h to yield a brown powder.

Synthesis of the reported Fe(TCPP)-MOF nanoparticles

3 mg of TCPP was completely dissolved in 10 mL of DMF, followed by the addition of 5 mL of acetonitrile. Subsequently, 0.2 mL of an aqueous ferric chloride solution (50 µg mL^− 1) was introduced into the mixture. The reaction system was then maintained at 60 °C in an oven for 12 h. The obtained precipitate was harvested by centrifugation and washed three times with absolute ethanol, ultimately affording the Fe(TCPP)-MOF material.

Synthesis of CRGDK@Mn-Fe(TCPP) MOF (MFMP)

The CRGDK functionalized Mn–Fe(TCPP) MOF was synthesized via a carbodiimide mediated coupling reaction. Briefly, 30 mg of nanoscale Mn–Fe(TCPP) MOF was uniformly dispersed in 1 mL of 10 mg/mL 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) solution prepared in 0.1 M 2-(N-morpholino) ethanesulfonic acid (MES) buffer (pH 5.5) and stirred gently for 5 min. Subsequently, 10 mg/mL sulfo-N-hydroxysulfosuccinimide (sulfo-NHS) was added, and the mixture was shaken for 15 min to activate the carboxyl groups. Thereafter, 4.0 mL of 6 × 10^− 5 M CRGDK peptide in phosphate buffered saline (PBS, pH = 7.4) was introduced and allowed to react overnight under gentle stirring. The resulting CRGDK-functionalized Mn–Fe(TCPP) MOF was collected by centrifugation and washed three times with PBS to remove unbound peptide.

Synthesis of CRGDK@Fe(TCPP)-MOF (FMP)

The CRGDK functionalized Fe(TCPP)-MOF was synthesized through a carbodiimide mediated coupling reaction. Briefly, 30 mg of nanoscale Fe(TCPP)-MOF was uniformly dispersed in 1 mL of EDC solution prepared in 0.1 M MES buffer (pH = 5.5) and stirred gently for 5 min. Subsequently, 10 mg/mL sulfo-NHS was added, and the mixture was shaken for 15 min to activate the surface carboxyl groups. Thereafter, 4.0 mL of 6 × 10^− 5 M CRGDK peptide in PBS (pH = 7.4) was added and allowed to react overnight under gentle stirring at room temperature. The resulting CRGDK@Fe(TCPP)-MOF was collected by centrifugation and washed three times with PBS to remove unbound peptide.

Calculation of HOMO–LUMO Energy Levels

The frontier molecular orbitals (HOMO and LUMO) of the porphyrin-based structures were calculated using density functional theory (DFT). All geometry optimizations were performed with the Gaussian 16 software package. The initial molecular structures were constructed based on the crystallographic coordinates of Fe(TCPP)-MOF and Mn–Fe(TCPP)-MOF. The exchange–correlation effects were treated using the Becke’s three-parameter hybrid functional combined with the Lee–Yang–Parr correlation functional (B3LYP) with the 6-31G (d, p) basis set for H, C, N, and O atoms, and the LANL2DZ effective core potential for Fe and Mn centers. Frequency calculations were conducted at the same level of theory to confirm the absence of imaginary frequencies, ensuring that the optimized structures correspond to true minima on the potential energy surface. All visualization and orbital mapping were performed using GaussView 6.0 and Multiwfn 3.8.

Quantitative determination of singlet oxygen (¹O₂) generation

A DPBF solution (10 µg/mL in DMSO) was mixed with an MFMP solution (50 µg/mL in PBS) and exposed to US (1 W cm^− 2, 1.0 MHz, 50% duty cycle) for different time intervals (0, 2, 4, 6, 8, 10 min). The generation of ¹O₂ was monitored by recording the decrease in the characteristic absorption peak of DPBF at 420 nm.

Quantitative determination of hydroxyl radical (•OH) generation

50 µg/mL of MFMP solution was mixed with 5 µg/mL MB solution and exposed to US (1 W cm^− 2, 1.0 MHz, 50% duty cycle) for different time intervals (0, 2, 4, 6, 8, 10 min). The generation of •OH was monitored by recording the decrease in the characteristic absorption peak of MB at 668 nm.

ESR measurement of free radicals

FMP and MFMP (2 mg each) were separately dissolved in 1 mL of DMF, followed by the addition of H₂O₂ (0.1 mmol/L, 1000 µL) and the spin-trapping agents DMPO (0.02 mol/L, 100 µL) and TEMP (0.1 mol/L, 50 µL). The resulting mixtures were subjected to ultrasound irradiation (1 W cm^− 2, 1.0 MHz, 50% duty cycle) and incubated for 3 min, repeated three times in total.

Cell viability assays

Cell viability was assessed using a CCK-8 kit. RAW264.7 cells were seeded into 96-well plates at a density of 4 × 10⁴ cells per well and subjected to different treatments. After the different treatments, 10 µL of CCK-8 working solution was added to each well and incubated for 30 min, followed by measurement of the absorbance at 450 nm using a microplate reader.

4T1 cells were seeded in a 96-well plate at a density of 1 × 10⁴ cells per well in 100 µL of medium. After the different treatments, 10 µL of CCK-8 working solution was added to each well and incubated for 30 min, followed by measurement of the absorbance at 450 nm using a microplate reader.

Cellular uptake

4T1 cells were seeded in 24-well plates at a density of 1 × 10⁴ cells per well and allowed to adhere. After conjugation of FITC to the nanoformulations, FITC@FeTCPP, FITC@FMP, and FITC@ MFMP were individually added to the culture medium. The cellular uptake and distribution were then visualized using confocal laser scanning microscopy (CLSM, 2-PLUS, Nikon, Japan).

In vitro detection of ROS generation

Intracellular ROS levels after different treatments were evaluated using DCFH-DA. 4T1 cells were seeded into 24-well plates at a density of 1 × 10⁴ cells per well and cultured for 24 h. After incubation with the various formulations for 24 h, the cells were treated with DCFH-DA (10 µM) for 30 min, followed by ultrasound irradiation (1.5 W cm^− 2, 1.0 MHz, 50% duty cycle) for 3 min. The cells were then incubated for an additional 30 min, and ROS generation was analyzed using an inverted fluorescence microscope.

Detection of cGAS-STING pathway activation

4T1 cells were treated with different formulations under ultrasound and non-ultrasound conditions (Group I: Control; Group II: Fe(TCPP); Group III: FMP; Group IV: MFMP). Western blotting (WB) was performed to analyze the expression of signaling proteins, including p-IRF3, IRF3, p-TBK1, TBK1, p-STING, STING, p-NF-κB, NF-κB, IKK, and cGAS. In addition, ELISA kits were used to quantify the levels of IFN-β, TNF-α, and IL-6 in the cell culture supernatants.

Preparation of bone marrow–derived dendritic cells (BMDCs)

BMDCs were isolated from the femurs of male C57BL/6 mice. Bone marrow cells were flushed out with RPMI-1640 medium using a syringe, and the resulting suspension was passed through a 200-mesh cell strainer. Red blood cells were removed using Red Blood Cell Lysis Buffer, followed by washing and centrifugation. The remaining cells were resuspended at a density of 2 × 10⁶ cells mL^− 1 and seeded into 6-well plates. The culture medium was refreshed on days 3 and 5. Semi-adherent and loosely adherent cells were collected between days 7 and 9 for subsequent experiments.

In vitro DC stimulation transwell experiment

4T1 cancer cells were seeded into the upper chamber of a Transwell system and incubated with different formulations under ultrasound and non-ultrasound conditions (Group I: Control; Group II: FeTCPP; Group III: FMP; Group IV: MFMP) for 12 h. After treatment, the cells were washed twice with PBS. Immature BMDCs were then added to the lower chamber and cocultured for an additional 24 h. Subsequently, the harvested BMDCs were stained with FITC-conjugated anti-mouse CD80 and APC-conjugated anti-mouse CD86 antibodies for 30 min at 4 °C. The maturation status of DCs was analyzed using a BD FACSCelesta™ flow cytometer.

Detection of ferroptosis pathway activation

Pretreated cells were harvested and incubated overnight at 4 °C with primary antibodies against GSH and GPX4 (rabbit, 1:500). After washing, the cells were incubated with a Cy3-conjugated goat anti-rabbit IgG secondary antibody (1:1000) for 2 h at room temperature. Subsequently, nuclei were counterstained with DAPI, and fluorescence images were acquired using a confocal laser scanning microscope (CLSM).

In vitro evaluation of immunogenic cell death (ICD)

To assess ICD induction, the surface exposure of calreticulin (CRT), release of high-mobility group box 1 (HMGB1), and secretion of adenosine triphosphate (ATP) were evaluated in vitro. 4T1 cancer cells were treated with PBS, FeTCPP, FMP, and MFMP under conditions with or without US irradiation.

For flow cytometric analysis of surface CRT, pretreated cells were collected and incubated with anti-rabbit CRT antibody (1:100) for 30 min at 4 °C, followed by a Cy3-conjugated goat anti-rabbit IgG secondary antibody (1:1000) for another 30 min at 4 °C. The samples were then analyzed using a flow cytometer.

ATP release was quantified using an ATP Assay Kit. Briefly, 4T1 cells were seeded into 24-well plates and allowed to adhere for 24 h. After different treatments, the culture supernatants were collected to determine extracellular ATP levels. For intracellular ATP quantification, cells were lysed with 100 µL/well lysis buffer, and ATP content was measured according to the manufacturer’s instructions.

For Western blot analysis of HMGB1, pretreated cells were lysed using RIPA buffer, and total protein was quantified by BCA protein assay. Equal amounts of protein were subjected to SDS-PAGE and transferred onto PVDF membranes. The membranes were incubated overnight at 4 °C with primary antibodies against HMGB1 (1:500) and GAPDH (1:1000), followed by HRP-conjugated goat anti-rabbit IgG (1:1000) for 2 h at room temperature. Protein bands were visualized using an enhanced chemiluminescence detection system.

Hemolysis assay of MFMP

Blood samples were collected from healthy BALB/c mice and centrifuged at 1500 rpm for 15 min to separate red blood cells (RBCs). The RBCs were washed three times with normal saline and resuspended to prepare an RBC suspension. Equal volumes (1 mL) of RBC suspension and MFMP solutions at different concentrations (25, 50, 100, 200, and 400 µg/mL) were mixed and incubated at 37 °C for 2 h, followed by centrifugation at 5000 rpm for 10 min. Distilled water and 10 mM PBS (pH 7.4) were used as positive and negative controls, respectively. The absorbance of the supernatant was recorded at 540 nm using a UV spectrophotometer, and the hemolysis rate was calculated according to the following equation.

Blood routine analysis

Healthy female BALB/c mice were intravenously injected with MFMP (100 µL, 1 mg·mL^− 1). At day 30 post-injection, blood samples were collected for routine hematological examination and biochemical analysis.

In vivo biocompatibility of MFMP

Eighteen healthy female BALB/c mice were randomly divided into three groups and intravenously administered with MFMP (5 mg·kg^− 1). Body weights were recorded every three days to monitor systemic biosafety. On days 0, 7, and 28 post-administrations, the mice were euthanized, and major organs (heart, liver, spleen, lungs, and kidneys) were harvested, fixed in 10% paraformaldehyde, embedded in paraffin, sectioned, and subjected to H&E staining for histopathological examination.

In vivo biodistribution of MFMP

4T1 tumor-bearing mice (n = 3) were intravenously injected with MFMP (10 mg·mL^− 1). At predetermined time points (6 and 24 h), the mice were sacrificed, and tumor tissues as well as major organs (heart, liver, spleen, lung, and kidney) were collected, weighed, and digested in HNO₃ at 60 °C. The concentrations of Fe³⁺ and Mn²⁺ in tumor tissues and major organs were subsequently determined using inductively coupled plasma mass spectrometry (ICP-MS).

Tumor model establishment

A 4T1 breast tumor model was established by subcutaneous injection of 1 × 10⁶ 4T1 cells into the left mammary fat pad of female BALB/c mice. Mice were subjected to subsequent in vivo experiments after ten days of tumor growth.

Lung metastasis model

Following 14 days of drug treatment, the primary tumors in 4T1-bearing mice were surgically resected. Thereafter, the mice were intravenously injected with 4T1 cells to induce pulmonary metastasis. After 30 days, the mice were sacrificed, and lungs were harvested for metastasis evaluation. Visible metastatic nodules were photographed, and lung tissue sections were subjected to H&E staining for histopathological analysis.

In vivo antitumor therapy and immune response analysis

When tumor volumes reached 80–120 mm³, forty-eight tumor-bearing mice were randomly divided into eight groups and treated with PBS, PBS + US (1 MHz, 50% duty cycle, 1.5 W·cm^− 2, 5 min), Fe(TCPP), Fe(TCPP) + US, FMP, FMP + US, MFMP, or MFMP + US. MFMP was administered via intravenous injection on days 0, 3, and 6, followed by ultrasound irradiation on days 1, 4, and 7, with a 24 h interval between injection and US exposure. Tumor volume was calculated using the formula V = L × W² / 2, where L and W denote tumor length and width, respectively, and relative tumor growth was expressed as V/V₀ (V₀ = initial tumor volume prior to treatment). On day 14, mice were sacrificed, and major organs (heart, liver, spleen, lung, and kidney) as well as tumors were harvested for histological staining. In addition, single-cell suspensions from tumors and lymph nodes were prepared, and immune cell populations were analyzed by flow cytometry using fluorochrome-conjugated anti-mouse antibodies.

Statistical analysis

Data are presented as mean ± standard deviation (SD). Statistical comparisons between groups were performed using Student’s t-test. *p < 0.05 was considered statistically significant, while **p < 0.01 and ***p < 0.001 indicated highly significant and extremely significant differences, respectively.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Huan-Hui Wang: Methodology and Writing - Original Draft. Long-Yi Nan: Writing - Review & Editing. Yan Zheng: Software and Supervision. Jian-Pen Guo: Conceptualization.

Funding

This work was funded by the National Natural Science Foundation of China (82460701), the Natural Science Foundation of Jilin Province, China (YDZJ202501ZYTS188).

Data availability

Data will be made available on request.

Declarations

Competing interests

The authors declare no competing interests.