Ultrasound-triggered carrier-free nanoprodrugs activate cGAS-STING pathway to enhance tumor-targeting chemo-immunotherapy.

1
Introduction
Triple-negative breast cancer (TNBC) is recognized for its aggressive behavior, high proliferation rate, and poor prognosis compared with other subtypes [1,2]. Recent advancements in immunotherapy, particularly immune checkpoint blockade (ICB) therapy, have shown promise in reducing tumor recurrence and metastasis [[3], [4], [5]]. However, the effectiveness of these therapies is often limited by the immunosuppressive tumor microenvironment and insufficient tumor immunogenicity, resulting in varied and disappointing outcomes. Fortunately, activating innate and adaptive immune responses is a new approach to tumor immunotherapy [[6], [7], [8], [9]].
The cyclic GMP-AMP synthase-stimulator of interferon genes (cGAS-STING) pathway is a crucial mechanism for sensing cytosolic DNA and plays a vital role in activating antitumor immunity [10,11]. Specifically, on detecting abnormal tumor DNA, the cGAS enzyme produces cyclic GMP-AMP (cGAMP), activating the STING protein and producing type I interferons (IFNs) and pro-inflammatory cytokines [12,13]. The cGAS-STING pathway has been extensively investigated as a promising therapeutic target for various tumors by stimulating immune cells, particularly dendritic cells (DCs) and CD8+ T cells [14]. Compared to the development of small-molecule immune agonists, such as cyclic dinucleotides, which have low metabolic stability and limited clinical application, the in situ activation of the cGAS-STING pathway through the accumulation of intrinsic double-stranded DNA (dsDNA) has emerged as a viable alternative [[15], [16], [17]]. However, challenges, including inadequate production and improper distribution of dsDNA fragments, can significantly impair the effectiveness of immune stimulation [18,19]. Thus, controllably enhancing endogenous dsDNA fragments in tumor lesions is essential for activating the cGAS-STING pathway and improving cancer immunotherapy.
Recent studies have shown that chemotherapy can damage DNA by binding to it and causing inter- or intra-strand cross-links, ultimately releasing dsDNA [[20], [21], [22]]. However, chemotherapeutic drugs often have high systemic toxicity and limited induction of dsDNA. Our previous study aimed to address the safety concerns associated with these drugs by developing a series of ultrasound (US)-activated carrier-free self-assembled biocompatible nanoprodrugs (PBSN38-OSs) to enhance the effectiveness of tumor-targeting chemotherapy. The PBSN38-OSs generate substantial numbers of reactive oxygen species (ROS) in situ on exposure to US irradiation, triggering the release of the chemotherapeutic drug SN38 and resulting in strong therapeutic effects against various solid tumor models in a spatially controlled and microenvironment-independent manner [23]. Compared with traditional chemotherapy alone, US-controlled release of the chemodrugs offers advantages such as increased effectiveness, especially for localized lesions, and reduced systemic side effects [[24], [25], [26]]. Moreover, US has been found to induce calcium (Ca2+) overload by elevating the intracellular Ca2+ concentration through Ca2+ influx from the extracellular fluid into the cytoplasm. This disruption of calcium homeostasis, combined with ROS generation, can lead to mitochondrial damage and an increase in dsDNA production [27,28]. Many intracellular Ca2+ nanomodulators can also generate dsDNA [29]. Considering that the cGAS-STING pathway is an inflammatory pathway, its abnormal activation may contribute to the development of certain inflammatory or degenerative disorders [17,30]. We believe that the use of US as an external stimulus to deliver ROS-sensitive prodrugs can overcome the poor specificity of endogenous methods and the challenges of tissue penetration associated with other stimuli, thus making it a promising tool for precisely controlling tumor dsDNA accumulation [31,32].
In this study, we developed a US-triggered carrier-free nanoprodrug PBSN38-CUR to enhance tumor-targeted chemo-immunotherapy by activating the cGAS-STING pathway in a microenvironment-independent manner. As shown in Fig. 1A, PBSN38-CUR was fabricated through the self-assembly of prodrug pinacol boronic ester-conjugated 7-ethyl-10-hydroxycamptothecin (PBSN38) and sonosensitizer curcumin (CUR). The resulting nanoprodrug demonstrated a high drug payload while maintaining favorable aqueous stability and a definite formulation. After intravenous administration, PBSN38-CUR exhibited negligible toxicity despite its inevitable distribution in normal tissues due to the nonresponsiveness of prodrug release. Once accumulated in tumor tissue and following US irradiation, the sonosensitizer CUR generated abundant ROS, which activated the PBSN38 prodrug. This facilitated the on-demand release of SN38 and the production of dsDNA, thereby improving chemotherapeutic efficacy while reducing side effects in the 4T1 tumor model. Moreover, CUR enhanced cellular Ca2+ concentration by inhibiting Ca2+ efflux from the cytoplasm to the extracellular space, leading to a severe disruption of oxidation-reduction homeostasis, mitochondrial dysfunction, and increased dsDNA accumulation. The released dsDNA broadly activated the cGAS-STING pathway, which helped reprogram the immunosuppressive tumor microenvironment by improving DC antigen presentation, elevating CD8+ T cell infiltration, and increasing pro-inflammatory cytokine production, ultimately leading to tumor inhibition (Fig. 1B). The growth of 4T1 tumors and tumor metastasis in mouse models was significantly inhibited following PBSN38-CUR treatment and US irradiation, indicating a substantial therapeutic benefit. Furthermore, the combination of PBSN38-CUR with the immune checkpoint inhibitor anti-PD-L1 antibody further enhanced antitumor performance. Collectively, this study presents a promising strategy for improved tumor-targeted chemo-immunotherapy through the cascade release of SN38 and Ca2+ overload by self-assembled carrier-free nanoprodrug to activate the cGAS-STING immune pathway, thus offering a potentially translatable clinical nanomedicine for US-mediated treatment of solid tumors.

2
Experimental section
2.1
Materials
The 7-ethyl-10-hydroxycamptothecin (SN38) and CUR were purchased from Shanghai Titan Scientific Co., Ltd. The RPMI-1640 medium, fetal bovine serum (FBS), and 0.25 % trypsin solution were purchased from Gibco. The Cell Counting Kit-8 (CCK-8) proliferation assay kit, ROS assay kit, calcium fluorescent probe Fluo-4 AM, Annexin V-FITC/PI apoptosis detection kit, calcein/PI cell live/dead assay kit, and mitochondrial membrane potential assay kit with JC-1 were purchased from Beyotime Biotechnology. Phospho-histone H2AX (Ser139) polyclonal antibody, TBK1 monoclonal antibody, STING polyclonal antibody, IRF3 monoclonal antibody, and phospho-IRF3 (Ser396) antibody were purchased from Proteintech Group, Inc. Phospho-TBK1/NAK (Ser172) (D52C2) XP rabbit mAb and phospho-STING (Ser365) (D8F4W) rabbit mAb were purchased from CST. Picogreen dsDNA quantitation reagent was purchased from Yeasen Biotechnology (Shanghai) Co., Ltd. InVivoMAb anti-mouse PD-L1 (B7-H1) antibody was purchased from Bioxcell. Mouse interferon beta (IFN-β) enzyme-linked immunosorbent assay (ELISA) kit, mouse interferon-gamma (IFN-γ) ELISA kit, mouse tumor necrosis factor-alpha (TNF-α) ELISA kit, mouse interferon-gamma–induced protein 10 kDa (IP-10)/CXCL10 ELISA kit, and TUNEL apoptosis assay kit were purchased from Elabscience. Zombie Aqua fixable viability kit, anti-mouse CD45, anti-mouse CD3, anti-mouse CD4, anti-mouse CD8a, anti-mouse Foxp3, anti-mouse CD11c, anti-mouse CD80, and anti-mouse CD86 were purchased from BioLegend.

2.2
Fabrication of PBSN38-CUR
PBSN38-CUR was fabricated through a self-assembly process using a one-step nanoprecipitation method. PBSN38 and CUR were first dissolved in dimethyl sulfoxide to prepare a stock solution. The two compounds were then mixed evenly in a molar ratio of 1:1 and injected dropwise into deionized water. This mixture was stirred at room temperature for 2 h and then dialyzed for 24 h using a dialysis bag with a molecular weight cutoff of 3500 Da to remove unassembled agents. Also, the assembly process of PBSN38-CUR was simulated using Hex 8.0.0. The results of the intermolecular interactions were visualized with PyMOL. The binding energy between the two molecules was calculated using AutoDock 4.2 software.

2.3
Characterization of PBSN38-CUR
The hydrodynamic diameter, polydispersity index (PDI), and zeta potential were measured using the Malvern zeta-sizer nano series. The stability of PBSN38-CUR in H2O, phosphate-buffered saline (PBS), and 10 % serum was determined by monitoring its hydrodynamic size. The nanoparticle morphology was measured using a transmission electron microscope (TEM) (JEM-1230, 200 kV). The content and cumulative release of SN38 under different conditions were measured using high-performance liquid chromatography (HPLC, Agilent 1260; Agilent Technologies, USA) with an eluent of methanol and water (80/20, v/v). The ultraviolet–visible (UV-vis) spectra of PBSN38, CUR, and PBSN38-CUR were recorded using a microplate spectrophotometer.

2.4
Cellular uptake of PBSN38-CUR
The PBSN38-CUR was first labeled with the fluorescent dye Cy5. Subsequently, the cellular uptake and the subcellular distribution of PBSN38-CUR in 4T1 cells were examined using a Leica confocal laser scanning microscope (CLSM). The data were quantitatively assessed by flow cytometry analysis at various time points for 24 h.
Intracellular ROS and Ca2+
assessments: The generation of ROS from BSN38-CUR under US irradiation was monitored using DCFHDA as an indicator. DCFHDA was pretreated with NaOH (0.01 mol L−1) to convert it into DCF. PBSN38-CUR was then added to the DCF solution and irradiated with US (1.5 W cm−2, 3 MHz, 50 % duty cycle) for 1 min. The absorption of generated ROS (Ex/Em = 488/525 nm) was measured with a fluorescence spectrometer. For flow cytometry, 4T1 cells were seeded into 12-well plates and incubated overnight. The cells were washed after treatment with PBSN38-CUR for 24 h, stained with 10 μM DCFHDA for 20 min, and subjected to US irradiation. The fluorescence intensity of intracellular DCF was quantified using a Beckman CytoFlex flow cytometer. The fluorescence images were also acquired with a Leica fluorescence microscope. 4T1 cells were cultured in a 24-well plate and incubated with PBSN38 and PBSN38-CUR at a 10 μM SN38-equivalent dose for 24 h before US irradiation. The intracellular ROS production and Ca2+ images were captured after washing with cold PBS.

2.5
Evaluation of DNA double-strand breaks
DNA double-strand breaks were analyzed using phosphorylated γ-H2AX detection and the Picogreen dsDNA detection agent. 4T1 cells (3 × 104 cells per well) were cultured overnight in a 24-well plate and then treated with PBSN38 and PBSN38-CUR at a 10 μM SN38-equivalent dose for 24 h. The cells were then irradiated with the US (3 MHz, 1.5 W cm−2, 50 % duty cycle) for 1 min. The cells were fixed with 4 % paraformaldehyde after 4 h of incubation, washed with PBS, blocked with 5 % FBS, and permeabilized with 0.3 % Triton-X in PBS. They were incubated overnight at 4 °C with the primary antibody (phosphohistone H2AX-Ser139-Ab, 1:500), followed by treatment with Alexa Fluor 488 anti-rabbit secondary antibodies (CST, 1:1000) for 1 h. After washing, the coverslips were mounted with DAPI and observed under a Leica fluorescence microscope. The dsDNA release was analyzed by staining the cells with the Picogreen dsDNA detection agent and characterizing them using a Leica CLSM after the PBSN38-CUR incubation and US treatment.

2.6
Mitochondrial membrane potential assay
The mitochondrial membrane potential was evaluated using the JC-1 assay kit. 4T1 cells were seeded in glass-bottom dishes at a density of 5 × 104 cells per well and treated with PBS, PBSN38, and PBSN38-CUR at a 10 μM SN38-equivalent dose for 24 h. The cells were then irradiated with US at 3 MHz and 1.5 W cm−2 for 1 min. The medium was replaced with 500 μL of JC-1 probe solution and washed as per the instructions after 1 h of incubation. JC-1 staining was detected using a CLSM with Ex490/Em530 nm for monomers and Ex525/Em590 nm for aggregates.

2.7
In vitro cytotoxicity and apoptosis tests
The cytotoxicity of PBSN38-CUR was assessed using the CCK-8 on 4T1 cell lines. The cells were seeded in 96-well plates at a density of 5000 cells per well and incubated with various concentrations of PBSN38-CUR for 48 h. For US treatment, the cells were irradiated for 1 min (3 MHz, 1.5 W cm−2, 50 % duty cycle) after 24 h of incubation, followed by an additional 24 h in fresh medium. The cell viability was measured at 450 nm and calculated as a percentage of absorbance from untreated cells. Apoptosis and cell death were evaluated using the calcein/PI, Annexin V-FITC/PI, and TUNEL assay kits. 4T1 cells were cultured overnight in 24-well plates and then incubated with PBSN38 and PBSN38-CUR (10 μM SN38) for 24 h. After US treatment, the cells were stained following the kit instructions and photographed using a Leica fluorescence microscope.

2.8
In vitro immune activation characterization
4T1 cells were seeded in six-well plates at a density of 2 × 105 cells per well and incubated overnight. The next day, they were treated with PBS, PBSN38, and PBSN38-CUR at a concentration of 10 μM SN38 for 24 h. They were then irradiated with US (3 MHz, 1.5 W cm−2, 50 % duty cycle) for 1 min. After 6 h of additional incubation, the cells were lysed for immunoblotting to detect proteins in the cGAS-STING pathway. The treated cells were then co-cultured with DC2.4 cells and stained with anti-mouse CD80 and CD86 antibodies for flow cytometry analysis.

2.9
Cell lines and animal models
The 4T1 cell line (murine breast cancer cells, RRID: CVCL_0125) was purchased from the American Type Culture Collection (ATCC) and is free of mycoplasma contamination. The cancer cells were cultured in RPMI-1640 medium that contained 10 % fetal bovine serum and 1 % penicillin/streptomycin (Invitrogen, Carlsbad, CA) in a humidified atmosphere containing 5 % CO2 and 37oC. Female BALB/c mice, aged 6-8 weeks, were obtained from the Laboratory Animal Center at Zhejiang Chinese Medical University. They were kept in colony cages at 25 °C with a 12-h light/dark cycle and 45 % humidity. All experiments (License No. IACUC-20240219-10) were approved by the Institutional Animal Ethics Committee and conducted following its guidelines.

2.10
In vivo biocompatibility assessment
The periorbital venous blood was collected from mice for routine blood examination and serum biochemical analysis. For the hemolysis test, fresh anticoagulated blood was centrifuged at 3000 rpm for 10 min to isolate red blood cells (RBCs), which were then washed thrice with PBS (pH 7.4). Different concentrations of PBSN38-CUR were added to the RBC suspension and incubated at 37 °C on a shaker at 60 rpm for 4 h. Negative controls (0 % lysis) and positive controls (100 % lysis) were incubated with PBS and 1 % Triton solution, respectively. After centrifugation, the absorbance of the hemoglobin supernatant was measured at 540 nm. The hemolysis ratio was calculated as follows:

2.11
Blood clearance and biodistribution of PBSN38-CUR
For pharmacokinetic analysis, Cy5-labeled PBSN38-CUR was administered intravenously to mice at a dosage of 4 mg kg−1 (n = 3). The blood samples (50 μL each) were collected from each mouse at various time points (1 min, 30 min, 1 h, 2 h, 4 h, and 6 h) and subsequently analyzed using the IVIS Spectrum in vivo imaging system. For the biodistribution studies, 4T1 tumor-bearing mice were intravenously injected with PBSN38-CUR. The mice were sacrificed 24 h after injection, and major organs (heart, liver, spleen, lung, and kidney), along with tumor tissues, were excised. These tissues were then imaged and analyzed using the IVIS Spectrum in vivo imaging system.

2.12
In vivo tumor inhibition by PBSN38-CUR and its combination with ICB therapy
All mice were inoculated subcutaneously with 4T1 cells (106 cells each) to evaluate the in vivo antitumor effects. When the volume of tumors reached 80 mm3, the mice were randomly divided into six groups (five mice per group) and treated with PBS, US, PBSN38-CUR, Onivyde (Irinotecan: 7.5 mg kg−1) + US, CUR + US, or PBSN38-CUR + US, receiving an SN38-equivalent dose of 4 mg kg−1
via the tail vein. Animals were randomly assigned to different treatment groups. Researchers involved in data collection and analysis were blinded to the group allocation during experiments and outcome assessment. The tumors were subjected to US (3 MHz, 1.5 W cm−2) after 4 h for 10 min, and the tumor volume and body weight were recorded every other day. On day 11, the mice were sacrificed, and tumors were photographed and weighed. In a subsequent study on the combination therapy of PBSN38-CUR and αPD-L1 antibody, the mice were inoculated with 4T1 cells (106 each) and divided into four groups (PBS, αPD-L1 antibody, PBSN38-CUR + US, and combination) once the volume of tumors was 80 mm3. Their tumor volume and body weight were recorded every other day. A lung metastasis model was established after 15 days of PBS and PBSN38-CUR treatment under US irradiation.

2.13
Histological examination and immunofluorescent analysis
The excised tumors and major organs were collected, fixed in a 4 % paraformaldehyde solution, paraffin-embedded, and sectioned into 5-μm slices. The tumor sections were stained with hematoxylin and eosin (H&E) and imaged using an inverted bright-field microscope for histological examination. TUNEL analysis was conducted using the TUNEL apoptosis assay kit from Elabscience, following the manufacturer's protocols.

2.14
Immuno-analysis of treated tumors
Tumors and lymph nodes were harvested to analyze immune cells after various treatments. The tissue samples were cut into small pieces and digested into single-cell suspensions using 1 mg mL−1 collagenase I, 0.1 mg mL−1 collagenase IV, and 0.1 mg mL−1 hyaluronidase. The suspensions were filtered through a 200-mesh nylon mesh, centrifuged at 1500 rpm for 5 min, and washed with PBS. The RBCs were lysed, and the suspension was incubated with anti-mouse CD16/CD32 for 15 min to block the Fc receptor. The immune cell detection was conducted using the Zombie Aqua fixable viability kit and antibodies against CD45, CD3, CD4, CD8a, Foxp3, CD11c, CD80, and CD86 to assess DC maturation and T cell infiltration. The serum levels of TNF-α, IFN-γ, CXCL-10, and IFN-β were measured using ELISA kits.

2.15
Statistical analysis
The data were presented as mean ± standard deviation. Graphing and statistical analyses were performed with GraphPad Prism 8. The two-tailed Student t-test was used to compare the two groups. One-way analysis of variance, followed by post-Tukey's multiple comparisons test, was used to compare various groups. P values were considered statistically significant as follows: ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001.

3
Results and discussion
3.1
Preparation and characterization of PBSN38-CUR
As shown in Fig. 2A and Fig. S1, PBSN38-CUR was prepared by the self-assembly of PBSN38 and CUR using a simple one-step nanoprecipitation method [33,34]. The formulated PBSN38-CUR particles were approximately 128 nm in diameter, as determined by dynamic light scattering (DLS) and TEM, with the zeta potential being −13 mV (Fig. 2B and C and Figure S2). The drug-loading content (DLC) for PBSN38 was measured using HPLC, yielding a calculated DLC value of 53.7 %. This confirmed that the ratio of PBSN38 to CUR in the nanoprodrugs was 1:1, highlighting their high drug-loading capability [35]. As shown in Fig. 2D, the characteristic absorption peaks of PBSN38 at 365 nm and CUR at 420 nm were evident in PBSN38-CUR, indicating successful nanostructure formation. To provide insight into the driving forces behind the self-assembly, we performed molecular docking simulations. Using Hex 8.0.0 and AutoDock 4.2, the interaction between PBSN38 and CUR was modeled, yielding a binding energy of −181.83 kcal mol−1, which indicates a strong driving force for assembly (Fig. 2A). Furthermore, the UV-vis spectra of PBSN38-CUR in the presence or absence of hydrophobic sodium dodecyl sulfate (SDS) and NaCl were recorded (Fig. 2E). The absorption peak shapes exhibited remarkable changes on adding SDS due to the disruption of the original assembly balance. As an ionic surfactant, SDS incorporates into the nanoprodrug structure, disrupts the hydrophobic packing via electrostatic repulsion and micelle formation, and disassembles the nanoparticles when exposed to aqueous environment. The increase in absorption is indeed a direct consequence of disassembly of nanoprodrugs. In the intact PBSN38-CUR nanoparticle, the aromatic chromophores of both CUR and PBSN38 are in close proximity due to dense packing, resulting in an aggregation-caused quenching effect. Upon SDS-mediated disassembly, these chromophores are separated and exposed to the solvent, leading to recovery and enhancement of their intrinsic absorbance. This result strongly affirmed that hydrophobic interactions were a primary driving force in the self-assembly process [36,37]. However, adding NaCl had a negligible influence on the spectral shape, implying that electrostatic interactions might not be responsible for the self-assembly. PBSN38-CUR also demonstrated excellent stability in aqueous H2O and PBS, although its size increased after incubation with a 10 % serum solution (Fig. 2F and S3). Additionally, the ROS generation of PBSN38-CUR following US irradiation with different parameters, along with the ROS-triggered release of SN38 from PBSN38-CUR, was measured (Fig. 2G and H), further confirming the mechanism of US-activated drug release [38]. The nanoparticles exhibit extensive dissociation upon exposure to high concentrations of H2O2, rendering their particle size measurement unreliabled (Fig. S4). As shown in Fig. S5, PBSN38-CUR demonstrated excellent stability in a 10 % serum solution over 3 days, with minimal changes in hydrodynamic diameter. In addition, as shown in Fig. S6, we screened different feeding ratios (PBSN38: CUR = 5:1, 3:1, 1:1, 1:3, 1:5). The 1:1 ratio yielded nanoparticles with the smallest size and PDI. These characteristics are crucial for ensuring colloidal stability and efficient tumor delivery. The one-step nanoprecipitation used to fabricate PBSN38-CUR is simple and affords high drug loading at the lab scale. For clinical translation, scaling up will require optimizing mixing, temperature, and solvent removal to ensure batch consistency and Good Manufacturing Practice compliance, which represents a key focus for future development.

3.2
In vitro therapeutic efficacy of PBSN38-CUR
Initially, we assessed the cellular uptake of PBSN38-CUR using confocal laser scanning microscopy and flow cytometry. The intracellular red fluorescence from the Cy5-labeled nanoprodrug was observed in 4T1 cells, localized within lysosomes, after 2 h of incubation (Fig. 3A). The flow cytometry results indicated increased fluorescence intensity with prolonged incubation, confirming effective internalization (Fig. 3B). Some nanoprodrugs showed signs of lysosomal escape after extending the incubation to 8 h. We proposed that lysosomal escape is driven by the low pH-triggered aggregation of nanoparticles. In the acidic lysosome, the phenolic hydroxyl groups of CUR remain protonated, rendering the molecule highly hydrophobic and insoluble, which was reported in previous research. This leads to rapid nanoparticle precipitation and phase transformation, which generates mechanical stress on the lysosomal membrane and compromises its integrity [39]. Next, we investigated US-triggered intracellular ROS production, Ca2+ concentration, and mitochondrial function using DCFH-DA, Fluo-AM4, and JC-1 staining probes. As illustrated in Fig. 3C and D, the cells treated with PBSN38-CUR + US exhibited enhanced green fluorescence signals of ROS. In contrast, untreated cells or those treated solely with PBSN38, CUR, or US demonstrated faint green fluorescence, indicating effective ROS production by sonosensitizer-based nanoprodrugs under US irradiation. The flow cytometric analysis showed that the ROS fluorescence intensity of cells in the PBSN38-CUR + US group was approximately 4.0-fold higher than that in the PBS group, 3.9-fold higher than that in the PBSN38 + US group, 1.5-fold higher than that in the CUR + US group, and 2.1-fold higher than that in the PBSN38-CUR group, which corroborated the aforementioned findings (Fig. S7). CUR promoted intracellular Ca2+ accumulation by inhibiting Ca2+ efflux from the cytoplasm to the extracellular environment, resulting in increased ROS levels, reduced GSH level and mitochondrial dysfunction, especially with the help of US irradiation (Fig. S8) [[40], [41], [42], [43]]. To investigate the pathway underlying US-induced Ca2+ influx, we pre-incubated 4T1 cells with the specific transient receptor potential ankyrin 1 (TRPA1) channel inhibitor HC-030031. In the presence of HC-030031, the US-triggered Ca2+ influx was significantly attenuated (Fig. S9), identifying TRPA1 as the primary entry channel. This is consistent with the role of TRPA1 as a ROS-sensitive channel that serves as a key Ca2+ entry pathway upon oxidative stress. Regarding the inhibition of Ca2+ extrusion, it has been reported that curcumin acts as a functional inhibitor of the plasma membrane Ca2+-ATPase (PMCA), the principal pump responsible for extruding Ca2+ from the cytosol. Consistently, our enzymatic assays confirmed that CUR significantly reduced PMCA activity in 4T1 cells (Fig. S10). Overall, the combination of TRPA1-mediated Ca2+ influx (triggered by US) and PMCA-inhibited Ca2+ efflux (mediated by CUR) creates a synergistic influx-inhibition mechanism. This dual action leads to severe and sustained intracellular Ca2+ overload, which we propose is central to the amplified therapeutic effect observed in our study. Our data also revealed increased Ca2+ accumulation in the PBSN38-CUR + US group compared with the other groups (Fig. 3E). To directly evaluate the contribution of each component to the overall cytotoxicity, we pre-treated cells with specific inhibitors prior to CUR + US therapy: BAPTA-AM (an intracellular Ca2+ chelator) to eliminate Ca2+ signals, and N-acetylcysteine (NAC, a potent ROS scavenger). Both BAPTA-AM and NAC treatment significantly rescued cell viability compared to the CUR + US group alone (Fig. S11). Critically, there was no statistically significant difference in the protective efficacy between these two inhibitors, indicating that the dramatic rise in intracellular Ca2+ is predominantly a consequence of ROS generation, rather than of CUR-mediated efflux inhibition. This interpretation is fully consistent with Fig. 3E, where US alone (without the sonosensitizer CUR to generate sufficient ROS) failed to elevate intracellular Ca2+. It confirms that substantial ROS production, enabled by CUR under US irradiation, is the necessary precursor to the observed Ca2+ overload. Furthermore, the JC-1 probe was used to assess the depolarization of the mitochondrial membrane potential. As depicted in Fig. 3F and S12, we observed a significantly higher abnormal JC-1 monomer (green) to a normal aggregate (red) ratio in the PBSN38-CUR + US group (∼14.1) compared with the PBSN38-CUR group (∼1.8). This indicated a notable decrease in mitochondrial membrane potential due to the release of nanoprodrugs responsive to US irradiation and calcium overloading [44].
The therapeutic effects of PBSN38-CUR + US on 4T1 murine breast cancer cells were explored (Fig. 4A–D). Initially, we evaluated the biocompatibility of PBSN38-CUR, both with and without US irradiation, using the CCK-8. The experimental findings indicated that PBSN38-CUR alone exhibited minimal cytotoxicity in 4T1 cells. In contrast, the relative viability of 4T1 cells significantly decreased following US treatment due to the ROS-triggered drug release, resulting in cancer cell death (Fig. 4A). The PBSN38-CUR + US group exhibited the highest proportion of late apoptotic cells at 82.9 % compared with the control groups (Fig. 4B and C). Meanwhile, no observable signs of cell death were noted, as characterized by the live/dead viability/cytotoxicity assay kit and TUNEL staining, after US irradiation or prodrug treatments alone. However, the combination of US irradiation with PBSN38-CUR resulted in significant DNA double-strand breaks, indicated by a 2.6-fold increase in phosphorylated γ-H2AX expression, increased red fluorescence from dead cell staining, and increased rate of apoptosis (green fluorescence of TUNEL-positive cells) compared with that in the PBSN38-CUR group, demonstrating the enhanced therapeutic effectiveness of US-activated PBSN38-CUR (Fig. 4D). The dose-dependent cell viability for PBSN38+US group (Fig. S13) was examined. US alone did not significantly alter SN38's cytotoxicity, confirming that the enhanced effect of PBSN38-CUR + US is due to the synergistic action of CUR-mediated drug release and Ca2+ overload. As shown in Fig. S14, the synergistic therapeutic outcome is predominantly driven by the combined action of SN38 chemotherapy and CUR-mediated sonodynamic therapy, with SN38 playing a major role in amplifying the overall cell death. Moreover, an increase in the release of dsDNA, alongside elevated ROS levels from US-activated PBSN38-CUR, was observed, confirming that the nanoprodrugs effectively induced simultaneous mitochondrial damage and dsDNA release, thus showing significant potential in enhancing the cGAS-STING pathway [45]. As illustrated in Fig. 4E, the damaged DNA in tumor cells was evaluated using a dsDNA assay. The results showed that US irradiation led to a rapid release of damaged dsDNA into the cytosol following PBSN38-CUR treatment, indicating activation of the innate cGAS-STING pathway. We further performed direct measurement of cytosolic dsDNA levels using quantitative immunofluorescence analysis of the images in Fig. 4E. The cytosolic dsDNA signal was specifically quantified using ImageJ by isolating the fluorescence signal outside the nuclei. Our analysis revealed that dsDNA remained primarily confined to the nuclei in the US-alone and PBSN38+US groups. In contrast, a significant fraction of dsDNA was observed in the cytosol of the CUR + US group, which is consistent with our model wherein CUR-mediated Ca2+ overload promotes dsDNA release. The PBSN38-CUR + US group showed the highest level of cytosolic dsDNA accumulation (approximately 2.5-fold higher than that in the CUR + US group), demonstrating a clear synergistic effect. These results demonstrated that the observed immune activation stems from the cooperation of SN38 and CUR, with SN38 contributing the major role of the immunogenic DNA substrate. Therefore, western blotting was conducted to analyze the levels of proteins related to the cGAS-STING pathway, including STING/p-STING, TBK1/p-TBK1, and IRF3/p-IRF3 (Fig. 4F). The cells treated with PBSN38-CUR and US demonstrated significantly increased phosphorylation of related proteins (p-TBK1, p-STING, and p-IRF3), correlating with ROS levels and cell death ratios [46]. The expression levels of p-TBK 1, p-IRF3, and p-STING in the PBSN38-CUR + US group were 2.3-, 1.8-, and 1.7- fold, 3.7-, 1.5-, and 2.7- fold, as well as 2.9-, 1.7-, and 2.8- fold higher than those in the PBS + US, CUR + US, and PBSN38-CUR groups, respectively. Furthermore, PBSN38-CUR + US treatment significantly enhanced the maturation of DCs (Fig. 4G). The flow cytometry analysis revealed that the percentage of CD80+ and CD86+ DCs (65.9 % and 40.9 %) was significantly higher in the PBSN38-CUR + US group than in the PBSN38-CUR group (22.3 % and 26.4 %). This was attributed to the US-activated prodrugs inducing mitochondrial stress and activating the cGAS-STING pathway (Fig. S15) [[47], [48], [49]]. To conclusively demonstrate that the IFN-β production are mediated by the cGAS-STING axis, we utilized two highly specific inhibitors: RU.521 (a potent and selective inhibitor of cGAS) and C-176 (a specific covalent inhibitor that targets the transmembrane protein STING). As shown in Fig. S16, the treatment of 4T1 cells with either RU.521 or C-176 to PBSN38-CUR + US therapy resulted in a significant and substantial reduction in the secretion of IFN-β, confirming pathway dependency. Generally, the ultrasound-sensitizing effect of PBSN38-CUR generates ROS and triggers Ca2+ influx via TRPA1. Meanwhile, cytosolic Ca2+ overload disrupts mitochondrial function, leading to amplified ROS production, which synergistically induces DNA damage. The accumulated cytosolic dsDNA is then sensed by cGAS, leading to STING pathway activation.

3.3
Biosafety evaluation of PBSN38-CUR
Next, the in vivo biosafety of PBSN38-CUR was examined. First, we performed the blood biochemistry and complete blood panel analysis of mice under various treatments. The hematological analyses revealed no significant alterations in the levels of key parameters, including hepatic function indexes such as alanine aminotransferase (ALT), aspartate aminotransferase (AST), and alkaline phosphatase (ALP), renal function parameters such as blood urea nitrogen (BUN), creatinine (CREA), and uric acid (UA), and blood routine analysis of white blood cells (WBCs), RBCs, and platelets (PLTs) (Fig. 5A–C). Moreover, the hemolysis assay results confirmed the excellent blood compatibility of PBSN38-CUR even at the concentration of 1000 μg mL−1 (Fig. 5D and S17). Subsequently, the blood circulation, biodistribution, and tumor-targeting ability of PBSN38-CUR after intravenous injection were investigated. The circulatory half-life (t1/2) of PBSN38-CUR was calculated as 0.39 h (Fig. 5E). As illustrated in Fig. 5F and S18, the fluorescence signals originating from tumor areas began to emerge 1 h after injection and gradually increased until reaching the peak in 4 h. Moreover, the in vivo and ex-vivo results both illustrated that PBSN38-CUR exhibited great tumor accumulation (Fig. 5G and H and S19), highlighting the exceptional tumor-retention capability and good biosafety of the nanoprodrug. The extended circulation of PBSN38-CUR leads to significantly enhanced tumor accumulation of the active drug (SN38) compared to CPT-11. The observed t1/2 is consistent with uptake by the reticuloendothelial system (RES), primarily in the liver, as corroborated by biodistribution data. However, the carrier-free design (lacking exogenous polymers or lipids) likely facilitates faster metabolic clearance upon RES uptake. This characteristic may reduce potential long-term organ burden compared to non-degradable conventional nanocarriers. Furthermore, the H&E staining images of the major organs revealed that histological abnormalities were not observed in any of the major organs in the PBSN38-CUR group with or without US irradiation (Fig. 5I), indicating the negligible systemic toxicity of PBSN38-CUR. Notebly, the tumor temperature remained stable at approximately 34.8 °C before and after US irradiation, demonstrating no significant tissue heating effect (Fig. S20). The H&E staining images also confirmed the safety of US (Fig. S21). Collectively, the results indicated that PBSN38-CUR can be safely used in vivo because it does not induce appreciable systemic toxicity.

3.4
In vivo therapeutic efficacy of PBSN38-CUR
Encouraged by the significant in vitro cytotoxic effects of US-activated PBSN38-CUR on 4T1 cells, we further evaluated its therapeutic efficacy in vivo using a mouse model using subcutaneous 4T1 tumors. The mice were randomly assigned to six groups, each comprising five mice: (1) PBS, (2) US, (3) PBSN38-CUR, (4) Onivyde (Irinotecan: 7.5 mg kg−1) + US, (5) CUR + US, and (6) PBSN38-CUR + US. After intravenous injections of PBSN38-CUR at a dose of 4 mg kg−1 SN38, US irradiation at a frequency of 3 MHz and an intensity of 1.5 W cm−2 was applied 4 h later. We monitored tumor volumes and body weights of the mice every 2 days (Fig. 6A). Compared with the control group, the tumor growth was moderately inhibited in the US-only and PBSN38-CUR-only groups compared with the control group, which was potentially due to the cavitation effect of the US and the endogenous ROS induced by the drug release from PBSN38-CUR (Fig. 6B-D). The PBSN38-CUR + US group exhibited the slowest tumor growth and the most significant tumor suppression, consistent with the final tumor weights. These findings suggested that PBSN38-CUR effectively responded to the exogenous US within tumors, resulting in excellent antitumor efficacy with a tumor inhibition rate of 78.4 %, compared with the clinical Onivyde, which has a tumor inhibition rate of 18.7 %. No significant fluctuations in body weight were observed throughout the treatment period, indicating that PBSN38-CUR did not adversely impact the overall health of the mice (Fig. S22). Additionally, the tumor sections from various treatment groups were subjected to H&E staining and TUNEL immunofluorescence to further evaluate treatment effectiveness (Fig. 6E). The H&E staining showed that the tumor cells in the control and US-only groups exhibited normal morphology with intact nuclei. In contrast, the PBSN38-CUR-only and the Onivyde + US groups displayed some degree of nuclear condensation. The PBSN38-CUR + US group showed severe nuclear condensation and an increase in TUNEL-positive tumor cells, indicating significant damage to the tumor tissues. These findings demonstrated that PBSN38-CUR could respond to the exogenous US to activate the toxic SN38 and lead to Ca2+ overload, thus significantly impeding tumor proliferation.
We then evaluated the antitumor immune responses elicited in mice with 4T1 tumors to clarify the underlying mechanism of the therapeutic effects of PBSN38-CUR-mediated therapy. After treatment, we collected tumor tissues and inguinal lymph nodes for analyzing immune cells, including DCs and T cells. The group receiving PBSN38-CUR + US irradiation demonstrated a significant enhancement in the maturation of both lymph node and tumor DCs, achieving an average maturity level of approximately 38.4 % and 25.1 %, respectively, which supported effective antigen presentation (Fig. 6F). Also, CD3+CD8+ T cells are crucial adaptive immune cells that can recognize tumor-specific antigens and kill cancer cells by secreting perforin and granzymes. While assessing T cell infiltration in tumor tissues, we found that the PBSN38-CUR + US group had a greater proportion (approximately 39.1 %) of CD3+CD8+ T cells compared with the PBS (approximately 21.2 %) and PBSN38-CUR groups (approximately 26.4 %). Meanwhile, the PBSN38-CUR + US group exhibited the lowest average proportion (approximately 9.2 %) of CD4+Foxp3+ T cells, indicating successful activation of the antitumor immune response (Fig. S23). Furthermore, the levels of various cytokines, such as pro-inflammatory IFN-γ and TNF-α, were found to be elevated in the PBSN38-CUR + US group using an ELISA. Additionally, the levels of IFN-β and CXCL-10, which are crucial downstream factors of the STING signaling pathway, were the highest in the serum of mice treated with PBSN38-CUR + US (Fig. 6G). These findings indicated the activation of the immune response in tumors facilitated by US-activated PBSN38-CUR drug release and dsDNA accumulation. The carrier-free PBSN38-CUR, with self-cascading STING activation, was enhanced further by responding to H2O2 within the tumor microenvironment (TME) and exogenous US stimulation. This approach helped reverse the immunosuppressive tumor microenvironment and enabled effective immunotherapy.
Given that PBSN38-CUR + US treatments can improve tumor immune response, we used H&E staining to further investigate the formation of lung metastatic nodules 15 days after PBS and PBSN38 + US treatments. Large lung metastases were observed in the PBS group, whereas the number and area of lung metastatic nodules in the PBSN38 + US group were significantly reduced (Fig. 6H and S24). In addition, the CD3+CD8+ T cell infiltration in lung tissues demonstrated enhanced immune surveillance of PBSN38-CUR + US treated group (Fig. S25). To provide direct molecular evidence of STING pathway activation in situ, we performed Western blot analysis on tumor tissue lysates for key signaling molecules. As shown in Fig. S26, the PBSN38-CUR + US treatment group exhibited markedly increased phosphorylation levels of STING, TBK1, and IRF3 compared to all control groups. This data quantitatively corroborates the serum cytokine findings and offers definitive proof of pathway activation specifically within the tumor. Moreover, the synergistic effects of combining PBSN38-CUR with an αPD-L1 ICB were evaluated (Fig. 6I–K and S27). The mice were randomly divided into four groups: (1) PBS, (2) αPD-L1, (3) PBSN38-CUR + US, and (4) PBSN38-CUR + US and αPD-L1 (Combo). The αPD-L1 was administered intraperitoneally to each mouse at a dose of 4 mg kg−1, with an interval of 2 days between each administration. As expected, the combination of PBSN38-CUR + US along with the αPD-L1 antibody exhibited superior therapeutic efficacy against 4T1 tumors, with a tumor inhibition rate of 79.5 %, demonstrating a significant benefit of the combined treatment strategy. These results indicated that the US-activated tumor-targeting chemo-immunotherapy, employing a STING initiator in conjunction with ICB, effectively enhanced the immune response to suppress solid tumor growth.
Based on above, our PBSN38-CUR platform presented distinct advantages over existing strategies. Compared to systemic STING agonists (e.g., cyclic dinucleotides), it enabled spatiotemporally controlled, in situ activation of immune pathway within the tumor, which is anticipated to minimize off-target immune activation and systemic toxicity. Compared to conventional sonosensitizer-based nanocarriers, our carrier-free design achieves ultra-high drug loading and integrates a unique Ca2+ overload-amplified STING activation mechanism, offering superior manufacturing simplicity and immunotherapeutic potency. We also thoughtfully consider the translational pathway. While ultrasound penetration for deep-seated tumors remains a clinical consideration, our strategy is immediately suitable for accessible tumors (e.g., breast, melanoma) and could be integrated with interventional or intraoperative US techniques in the future. Regarding safety, the inherent risk of STING overactivation is mitigated by our US-controlled approach, which acts as a built-in safety switch. Future work will focus on dosage and parameter optimization to fine-tune this balance between efficacy and safety.

4
Conclusion
A US-activated carrier-free self-assembled prodrug containing PBSN38 and CUR, with high DLC and significant biocompatibility, was constructed in this study to enhance tumor-targeted chemo-immunotherapy by amplifying cGAS-STING pathway signals. Besides avoiding the negative side effects of the distribution of off-target chemotherapeutic drugs, the resulting PBSN38-CUR also integrated DNA damage generation from activated SN38 and US-triggered Ca2+ overload from CUR, promoting in situ dsDNA accumulation for improved therapeutic effect against 4T1 tumors and innate cGAS-STING pathway activation. This led to potent local tumor inhibition and distant metastasis prevention. Overall, this study demonstrated a facile and robust US-activated carrier-free nanoplatform capable of stimulating the cGAS-STING pathway in a microenvironment-independent manner. This approach enhanced tumor targeting and improved treatment efficacy, thus holding immense potential for clinical translation.

Ethics approval statement
All experiments involving animals (License No. IACUC-20240219-10) were performed with the approval of and following the guidelines of Institutional Animal Ethics Committee of Laboratory Animal Center of Zhejiang Chinese Medical University.

Funding sources
This work was supported by the 10.13039/501100001809National Natural Science Foundation of China (Grant No. 82230069, 82371967, 82402264, and U25A20138), Zhejiang Provincial Natural Science Foundation of China (Grant No. LMRY26H180008 and LR26H180002), and Binjiang Institute of Zhejiang University (ZY202205SMKY005).

CRediT authorship contribution statement
Xiaodan Xu: Data curation, Funding acquisition, Investigation, Methodology, Project administration, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. Peile Jin: Data curation, Investigation, Methodology, Project administration, Validation. Yijie Chen: Data curation, Investigation, Methodology, Project administration, Validation. Bihan Wu: Data curation, Investigation, Methodology, Project administration, Validation. Xia Fang: Data curation, Software, Visualization. Yue Song: Data curation, Software, Visualization. Jieli Luo: Data curation, Software, Visualization. Guowei Wang: Conceptualization, Funding acquisition, Supervision, Validation, Writing – review & editing. Pintong Huang: Conceptualization, Funding acquisition, Supervision, Validation, 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.