CS-1 induces pyroptosis-mediated cytotoxicity across diverse MDR cells
CS-1, the structure of which is depicted in Fig. 1A, demonstrated a wide-ranging anti-tumor activity [32]. By investigating the cytotoxicity of CS-1 on several MDR cell lines (MCF-7/ADR, A2780/DDP, and HepG2/DDP) as well as their respective parental counterparts, the three resistant cell lines demonstrated significant resistance to their corresponding chemotherapeutic agents (Fig. 1B-C), as previously reported [33,34,35]. On the contrary, CS-1 exhibited potent cytotoxicity in a dose-dependent manner against these resistant strains and their parental counterparts with IC50 values ranging from 1 to 5 µM (Fig. 1D). The resistance index (RI) of these cell lines was computed to be less than 5 to CS-1 (Fig. 1E), according to the formula of resistance index (RI) = IC50 of resistant cells/IC50 of parental cells [36]. Microscopic images demonstrated typical pyroptotic morphology in CS-1 treated cells, including cell swelling and the formation of large bubbles [37] (Fig. 1F). DCFH-DA probe staining assay visually demonstrated a significant increase in green fluorescence of three MDR cell lines after treatment with CS-1 for 12 h (Fig. 1G), reflecting the ability of CS-1 to induce high levels of ROS in MDR tumor cells.
Considering the higher RI and sensitivity to CS-1 (RI = 1.425 ± 0.024), we conducted RNA-seq assay to thoroughly investigate the effect of CS-1 on the gene transcription in MCF-7/ADR cells. The volcano plot showed 4520 upregulated genes and 2256 downregulated genes in MCF-7/ADR cells relative to the MCF-7. In contrast, CS-1 treatment caused the upregulation of 1386 genes and the downregulation of 2175 genes (p ≤ 0.05, fold change ≥ 2) (Fig. S1A-B). The Venn diagram also showed 1943 intersecting genes with common changes among MCF-7, MCF-7/ADR and CS-1 group (Fig.S1C). Among these 1943 genes, 978 genes upregulated in MCF-7/ADR group were inhibited by CS-1, which were similar to those of MCF-7 group. Conversely, 248 genes downregulated in MCF-7/ADR group were upregulated by CS-1, which were similar to that of the MCF-7 group. KEGG assay indicated the high enrichment of drug resistance protein regulatory pathway of PI3K-AKT and apoptotic pathways of MAPK, NF-κB and JAK-STAT in MCF-7/ADR cells (Fig. 1H). In contrast, CS-1 treatment induced high enrichment of pyroptosis-related inflammatory pathways such as MAPK, NF-κB and TNF (Fig. 1I). GSEA analysis also demonstrated that CS-1 treatment caused the activation of pyroptosis-related inflammatory pathways such as TNF, MAPK, and NF-κB in MCF-7/ADR cells (Fig. 1J-L). Heatmap analysis of DEGs showed downregulation of Multidrug resistance-associated protein ABC transporters (ABCG2) and Bcl-2, an antiapoptotic protein in MCF-7/ADR cells with CS-1 treatment (Fig. 1M). PPI analysis indicated that TNF, a master regulator of inflammation, connected NF-κB with Caspase-3 (Fig. 1N). Therefore, we further indicated the related protein expression level of pyroptosis by western blot combining with our reported result [38]. Then, we detected the changes of pyroptosis-associated proteins in MCF-7/ADR cells after treatment with different concentrations of CS-1. As we expected, CS-1 significantly activated Caspase-3 and GSDME-N levels (Fig. 1O). In conclusion, CS-1 can effectively induce TNF related Caspase-3-dependent pyroptosis in MCF-7/ADR cells (Fig. 1P), while may inhibiting ABCG2 to enhance drug concentration, which is urgent for effectively killing tumor cells.

CS-1 kill MDR tumor cells by inducing pyroptosis. (A) Chemical formula of CS-1. (B) Viability of A2780, A2780/DDP, HepG2, and HepG2/DDP cells following 24 h exposure to varying concentrations of cisplatin (DDP). (C) Viability of MCF-7 and MCF-7/ADR cells after 24 h treatment with different concentrations of doxorubicin. (D) Cytotoxicity of CS-1 across various cell lines (MCF-7, A2780, HepG2, MCF-7/ADR, A2780/DDP, HepG2/DDP) after 24 h treatment. (E) Drug resistance indexes (RI) to CS-1. (F) Morphological images of different cells. (G) Fluorescent images of different cells staining with ROS probe of DCFH-DA. (H) Bubble map of MCF-7 vs. MCF-7/ADR enrichment gene in KEGG pathway. (I) Bubble map of MCF-7/ADR vs. CS-1 enrichment gene in KEGG pathway. (J-L) GSEA analysis. (M) MCF-7/ADR vs. CS-1 heatmap. (N) MCF-7/ADR vs. CS-1 protein interaction network. (O) Western blot analysis of pyroptosis-associated proteins (GSDME, cleaved caspase-3) in MCF-7/ADR cells. (P) Schematic diagram of pyroptosis caused by CS-1
Preparation and characterization of HA@Lip-CS-1@PBCO NPs
Our study revealed that CS-1 has manifested significant anti-tumor potential in MDR tumors, however, its clinical implementation is limited by the cardiotoxic potential of high-dose administration [39]. To preserve the antitumor efficacy of CS-1 against MDR tumors while minimizing cardiotoxicity, we adopted a combinational strategy of CS-1/CO. MnCO was used as a precursor for generating CO. MTT assay demonstrated MCF-7/ADR cell viability of 48.2% after the combinational treatment of 1 µM CS-1 and 40 µM MnCO (Fig.S2A-B). However, the low targeting and bioavailability of CS-1 and MnCO limit their clinical application. Thus, we engineered a tumor-targeted “chemo-gas” nanocomplexe (HA@Lip-CS-1@PBCO NPs) to address these questions. Figure 2A demonstrated that subsequent to lipid encapsulation and hyaluronic acid modification, a 9.5 nm thick film covered the surface of PB NPs. DLS analysis disclosed an increment in particle size subsequent to the encapsulation of MnCO by PB NPs (resulting in a size of 102.7 nm), lipid encapsulation (yielding a size of 157.1 nm), and HA modification (resulting in a size of 160.1 nm), as illustrated in Fig. 2B. Furthermore, alterations in potential were observed. Starting from PB NPs with a value of −28.6 ± 1.6 mV, it changed to PBCO NPs at −25.1 ± 1.2 mV, then to Lip-CS-1@PBCO NPs at −36.9 ± 2.1 mV, and finally to HA@Lip-CS-1@PBCO NPs with a value of −37.56 ± 1.3 mV (Fig. 2C). Element mapping characterization results indicated that the manganese element in MnCO displays a high degree of overlap with the iron element in PB NPs, suggesting the effective loading of MnCO into PB NPs. Additionally, the detected phosphorus element signal further confirmed successful liposome encapsulation (Fig. 2D and S3A). The UV-vis absorption spectrum demonstrated characteristic peaks of MnCO and CS-1 at 340 nm and 290 nm, respectively, as shown in Fig. 2E. The FT-IR spectra of HA@Lip-CS-1@PBCO NPs exhibited characteristic absorption peaks corresponding to lipid components (C = O at 1740 cm⁻¹, P = O at 1230 cm⁻¹, and CH₂/CH₃ at 2850 cm⁻¹), MnCO (C ≡ O at ~ 2080 cm⁻¹) and the cyanide bridges of PB NPs (C ≡ N at 2080 cm⁻¹) [40], further confirming the successful integration of hyaluronic acid-modified Lip-CS-1 with PBCO NPs (Fig. 2F). It is noteworthy that ultrasonic treatment hardly affects the release of CO from PBCO NPs (Fig.S3B). UV spectroscopy assay data indicated that the encapsulation efficiency of CS-1 and MnCO were 21.45 ± 4.3% and 69.03 ± 1.13%, respectively (Fig. 2G). Furthermore, the material remained stable for approximately three days in water, PBS, and DMEM supplemented with 1% FBS, which is conducive to its in-vivo drug action (Fig. 2H). Collectively, these results suggest the successful establishment of HA@Lip-CS-1@PBCO NPs.

Synthesis and characterization of HA@Lip-CS-1@PBCO NPs. (A) TEM micrographs illustrating the sequential synthesis stages: PB NPs, PBCO NPs, Lip-CS-1@PBCO NPs, and HA@Lip-CS-1@PBCO NPs. (B) DLS and PDI values for the synthesized NPs at each stage. (C) Evolution of zeta potential throughout the synthesis stages. (D) Element mapping of HA@Lip-CS-1@PBCO NPs. (E) UV-vis absorption spectra of the nanocomplexes and their precursors (CS-1, PB NPs, PBCO NPs, Lip-CS-1@PBCO NPs, HA@Lip-CS-1@PBCO NPs), highlighting characteristic peaks. (F) FT-IR spectra of PB NPs, PBCO NPs, and HA@Lip-CS-1@PBCO NPs. (G) Encapsulation rates of CS-1 and MnCO at HA@Lip-CS-1@PBCO NPs. (H) The effect of different solvents on HA@Lip-CS-1@PBCO NPs stability
Functional characterization of HA@Lip-CS-1@PBCO NPs
According to the design features of HA@Lip-CS-1@PBCO NPs, we first investigated the release characteristics of CS-1 and CO (Fig. 3A). As shown in Fig. 3B, the release rate of CS-1 from HA@Lip-CS-1@PBCO NPs attained 88.28 ± 1% in 72 h in an acidic environment (pH6.8), which was attributed to the insertion of pH-responsive molecule DSPE-PEOz2K into the lipofilm. Then, using the FL-CO-1/PdCl2 probe (Fig. 3C) [28], we evaluated the CO release capability under microenvironment condition (high levels of H2O2). As shown in Fig. 3D, the CO release profiles of Lip-CS-1@PBCO NPs exhibited significant pH-dependent kinetics. Under physiological conditions (pH7.4 + 1 mM H2O2), the NPs demonstrated sustained but limited CO release, accumulating 18.5 ± 0.5 µM at 24 h with a near-linear progression. In stark contrast, an acidic microenvironment (pH5.4 + 1 mM H2O2) triggers rapid CO liberation, culminating in 70.5 ± 0.5 µM by 24 h – representing a 3.8-fold enhancement compared to neutral pH. Subsequently, the FL-CO-1 probe was utilized to further detect intracellular CO in cells subjected to treatment with HA@Lip-CS-1@PBCO NPs, with intracellular H2O2 serving as the substrate. While cells treated with PBS and CS-1 (1 µM) exhibited no increase in fluorescence, a significant enhancement of green fluorescence was observed in cells treated with HA@Lip-CS-1@PBCO NPs (1 µM CS-1, 20 µg/mL PBCO NPs). This enhancement was notably higher than that in cells treated with PBCO NPs (20 µg/mL), as illustrated in Fig. 3E-F. This result suggested that ROS induced by CS-1 can promote CO release behavior. In addition, the CO gas released by MnCO has the potential to form microbubbles within the organism. Owing to their relatively high acoustic impedance, these microbubbles can create a distinct contrast with the surrounding tissues and fluids [41], which provides the possibility for in vivo ultrasound imaging. As expected, an ultrasonic signal was detected in the sample containing 2 mM H2O2 and HA@Lip-CS-1@PBCO NPs solution (Fig. 3G). Conversely, no ultrasonic signal was detected in the H2O2 free sample. This result indicated that MnCO could be efficiently released in the high- H2O2 environment, which is similar to the tumor microenvironment (usually containing 100 µM-1 mM H2O2) [42].
Previous studies indicated that HA surface modification enhances nanodrug targeting efficiency through CD44 interaction, a receptor highly expressed on tumor cells [43, 44]. To evaluate cellular uptake, Ce6-labeled HA@Lip-Ce6@PBCO NPs were employed. Results revealed a progressive increase in characteristic red fluorescence within tumor cells over time, peaking at 6 h (Fig. 3H-I). Compared to free Ce6 or Lip-Ce6@PBCO NPs, HA@Lip-Ce6@PBCO-treated MCF-7/ADR cells exhibited stronger fluorescence intensity. Conversely, pretreatment with free HA markedly reduced this fluorescence signal (Fig. 3J), confirming CD44-mediated uptake inhibition. Using MCF-7/ADR multicellular spheroids to mimic solid tumors, the red fluorescence at depths of 0 ~ 60 μm was substantially higher in HA@Lip-Ce6@PBCO-treated samples than in free Ce6 or Lip-Ce6@PBCO groups (Fig. 3K). Consistent with previous reports [45], these results also demonstrated that HA could facilitate the entry of HA@Lip-Ce6@PBCO NPs into MCF-7/ADR cells via interaction with CD44. In conclusion, HA@Lip-Ce6@PBCO NPs demonstrate excellent targeting and penetration abilities, which hold great promise for tumor therapy.

Functional characterization of HA@Lip-CS-1@PBCO NPs. (A) Schematic diagram of CS-1 release and CO generation from HA@Lip-CS-1@PBCO NPs. (B) pH-dependent release profile of CS-1 from HA@Lip-CS-1@PBCO NPs. (C) The working principle of CO probe (FL-CO-1). (D) The amount of CO released from HA@Lip-CS-1@PBCO NPs (1 mg/mL PB NPs) at the presence of 1 mM H2O2 in PBS with different pH. (E-F) CLSM images (E) and corresponding fluorescence quantification (F) of MCF-7/ADR cells treated with PBS, CS-1, PBCO NPs, or HA@Lip-CS-1@PBCO NPs (1 µM CS-1, 20 µg/mL PB NPs). Green fluorescence indicates CO release detected by FL-CO-1. (G) Ultrasound images of HA@Lip-CS-1@PBCO NPs in the presence or absence of H2O2 in vitro. (H-I) CLSM images (H) and fluorescence quantification (I) of MCF-7/ADR cellular uptake of HA@Lip-Ce6@PBCO NPs over time (2, 4, 6 h). (J) CLSM images showing uptake in MCF-7/ADR cells incubated for 6 h with Ce6, Lip-Ce6@PBCO NPs, HA@Lip-Ce6@PBCO NPs, or HA@Lip-Ce6@PBCO NPs with free HA pre-treatment. (K) Fluorescence images demonstrating penetration into MCF-7/ADR 3D tumor spheroids after 24 h incubation with free Ce6, Lip-Ce6@PBCO NPs, or HA@Lip-Ce6@PBCO NPs. Bars are means ± SD (n = 3). **P
HA@Lip-CS-1@PBCO NPs effectively kill tumor cells in vitro
The cytotoxicity of HA@Lip-CS-1@PBCO NPs was initially assessed against MCF-7/ADR cells. As depicted in Fig. 4A, treatment with HA@Lip-CS-1@PBCO NPs reduced cell viability to 50.5 ± 10%, significantly lower than that achieved with PBCO NPs (95.5 ± 0.69%) or CS-1 alone (68.3 ± 5.5%). Meanwhile, this kind of NPs can significantly inhibit the formation of cell colony (Fig.S4A). Additionally, Live/dead staining similarly demonstrated the most intense red fluorescence in MCF-7/ADR cells treated with HA@Lip-CS-1@PBCO NPs, which reflected the high cell death rate. In contrast, tumor cells treated with PBCO NPs and CS-1 exhibited moderate red fluorescence compared to the group treated with HA@Lip-CS-1@PBCO NPs. As a control group, almost all cells showed green fluorescence after PBS treatment (Fig. 4B). Consistent with these findings, FACS analysis demonstrated that the cell death rate in HA@Lip-CS-1@PBCO NPs treated cells was 80.46%, which is significantly higher than that of CS-1 treated cells (60.92%) (Fig. 4C-D). To extend these observations to a more physiologically relevant model, 3D spheroids were employed. Following 5 days of treatment with HA@Lip-CS-1@PBCO NPs, spheroids exhibited loose and disintegrated morphology (Fig. 4E), and live/dead staining showed predominant red fluorescence, confirming low cell viability (Fig. 4F). Collectively, these results highlight the outstanding tumor-killing capacity of HA@Lip-CS-1@PBCO NPs in both two-dimensional and three-dimensional models.
Considering the improvement of efficacy of the nanodrug formulation, we further explored the changes in the genes of MCF-7/ADR before and after treatment. RNA-seq analysis revealed profound transcriptomic alterations in MCF-7/ADR cells following HA@Lip-CS-1@PBCO NPs treatment. Venn diagram in Fig.S4B identified changes of 12,208 mRNAs following treatment with HA@Lip-CS-1@PBCO NPs. According to the standard of fold change ≥ 2 fold (p ≤ 0.05), HA@Lip-CS-1@PBCO NPs caused 1689 gene upregulation and 2298 gene downregulation in MCF-7/ADR cells (Fig.S4C). Meanwhile, KEGG analysis of these genes with reverse changes revealed that HA@Lip-CS-1@PBCO NPs plays a significant role in the regulation of mitochondrial function-related PI3K-Akt pathway [46], the inflammation-related TNF, NF-κB, and MAPK pathways and oxidative stress-related HIF-1 and FoxO pathways, as well as the oxidative phosphorylation (OXPHOS) pathway (Fig. 4G). Among them, downregulation of OXPHOS-related genes (e.g., COX7C, ATP5F1D, MT-ND1, MT-ND4, and MT-ND6) strongly correlates with mitochondrial electron transport chain impairment, suggesting HA@Lip-CS-1@PBCO NPs induce oxidative stress via mitochondrial damage (Fig. 4H). Concurrent activation of HIF-1 (Fig. 4I) and FoxO pathways (Fig. 4J) further supports this hypothesis, as both pathways are redox-sensitive and amplify ROS accumulation under mitochondrial dysfunction. Heatmap analysis indicated the downregulation of antioxidant genes (NQO1, HSPA8, HSPD1, CAT) and mitochondrial fusion protein MFN1/MFN2, while upregulation of mitochondrial oxidative stress-related proteins (TNF, IL-18, IL-1β) in treated MCF-7/ADR cells (Fig. 4K). PPI analysis indicates the central role of HIF-1α, connecting MFN1/MFN2 with TNF, which is associated with Caspase-3-dependent pyroptosis (Fig. 4L). These findings collectively indicated that HA@Lip-CS-1@PBCO NPs can effectively kill MDR breast cancer cells by inducing mitochondrial oxidative stress.

HA@Lip-CS-1@PBCO NPs effectively kill tumor cells in vitro. (A) MTT assay and of MCF-7/ADR cells after 24 h treatment with different treatment. (B) Live/dead staining CLSM images of MCF-7/ADR cells with different treatment for 16 h. (C-D) Flow cytometry analysis of MCF-7/ADR cells with different treatment for 16 h. (E) Bright-field image of MCF-7/ADR 3D tumor spheres with different treatment. (F) Live/dead staining CLSM images of MCF-7/ADR 3D tumor spheres with different treatment for 24 h. (G) Bubble map of PBS vs. HA@Lip-CS-1@PBCO NPs in KEGG pathway. (H) Heatmap of genes in the oxidative phosphorylation pathway after HA@Lip-CS-1@PBCO NPs treatment. (I-J) GSEA analysis of HIF-1 and FoxO signaling pathway. (K) Heatmap of oxidative stress-related genes after HA@Lip-CS-1@PBCO NPs treatment. (L) Protein interaction network of oxidative stress-related proteins after HA@Lip-CS-1@PBCO NPs treatment. (Ⅰ: PBS, Ⅱ: CS-1 (1 µM), Ⅲ: PBCO NPs (20 µg/mL), and Ⅳ: HA@Lip-CS-1@PBCO NPs (1 µM CS-1 and 40 µM MnCO). Bars are means ± SD (n = 3). **P P P
HA@Lip-CS-1@PBCO NPs activate pyroptosis via oxidative stress-induced mitochondrial damage
Given that CS-1 induces pyroptosis in vitro, along with our RNA-seq data following HA@Lip-CS-1@PBCO NPs treatment, we investigated whether HA@Lip-CS-1@PBCO NPs could induce pyroptosis in MCF-7/ADR cells. The bright field images of cells treated with HA@Lip-CS-1@PBCO NPs revealed the cell swelling characteristic of pyroptotic cells (Fig. 5A). Furthermore, the staining result of T11 dyes, which can discriminate ruptured cell membrane from normal cell membrane, exhibited obvious membrane damage, cell swelling, and content leakage in MCF-7/ADR cells treated with HA@Lip-CS-1@PBCO NPs (Fig. 5A). TEM corroborated these findings, demonstrating significant plasma membrane damage in the HA@Lip-CS-1@PBCO NPs group (Fig. 5B). Consistent with membrane disruption, PI staining—which selectively labels cells with compromised membranes—showed markedly increased fluorescence intensity in HA@Lip-CS-1@PBCO NPs-treated cells compared to PBS group (Fig. 5C). As plasma membrane integrity loss facilitates the release of cytosolic contents, we quantified lactate dehydrogenase (LDH) release, a hallmark of pyroptotic cytotoxicity. HA@Lip-CS-1@PBCO NPs treatment resulted in a significant elevation of LDH release (10.3 ± 2.3-fold increase relative to PBS group), confirming successful pyroptosis induction (Fig. 5D). Subsequently, we examined whether HA@Lip-CS-1@PBCO NPs-induced pyroptosis was dependent on the activation of caspase-3/GSDME pathway. This result was similar to the Fig. 1 and RNA-seq data showing that HA@Lip-CS-1@PBCO NPs-activated caspase-3 triggers pyroptosis by cleaving GSDME (Fig. 5E).
Next, we endeavored to reveal the underlying process through which HA@Lip-CS-1@PBCO NPs induces GSDME-mediated pyroptosis. Combined with the RNA-seq results, we investigated whether pyroptosis triggered by HA@Lip-CS-1@PBCO NPs necessitates mitochondrial dysfunction. Notably, TEM images revealed that HA@Lip-CS-1@PBCO NPs significantly enhanced mitochondrial swelling in comparison to the control cells (Fig. 5F). Furthermore, mitochondrial fluorescence staining after treatment with HA@Lip-CS-1@PBCO NPs also showed severe mitochondrial damage in the cells (Fig. 5G). Mitochondrial membrane potential (ΔΨm) collapse, often associated with mitochondrial damage, triggers cytochrome c (Cyt C) release into the cytosol. Western blot analysis confirmed a significant increase in Cyt C release from mitochondria in HA@Lip-CS-1@PBCO NPs-treated cells (Fig. 5H). Assessment of ΔΨm using the JC-1 probe demonstrated the most intense green fluorescence (indicative of depolarization) in the HA@Lip-CS-1@PBCO NPs group compared to other treatments (Fig. 5I). Given the established role of mitochondrial dysfunction in ROS generation and the reported association between ROS and pyroptosis induction in cancer inhibition [47], we hypothesized that HA@Lip-CS-1@PBCO NPs might elevate cellular ROS levels. Indeed, both fluorescence imaging and flow cytometry analysis confirmed a significant upregulation of intracellular ROS in HA@Lip-CS-1@PBCO NPs-treated cells, exceeding levels observed with CS-1 or PBCO NPs alone (Fig. 5J).
Considering that mitochondria are a major source of ROS, we proceeded to investigate whether HA@Lip-CS-1@PBCO NPs induces the generation of mitochondrial ROS in MCF-7/ADR cells. As expected, HA@Lip-CS-1@PBCO NPs treatment triggered a significant upregulation of mitochondrial ROS (Fig. 5K). To assess the resulting oxidative stress and its impact on antioxidant defenses, we measured key biomarkers. First, cellular glutathione (GSH) levels, a critical antioxidant essential for maintaining mitochondrial structural/functional integrity and protecting mitochondrial DNA from oxidative damage, plummeted by 54.9 ± 0.9% in treated cells (Fig. 5L). This severe depletion indicates a profound disruption of the cellular antioxidant system in MCF-7/ADR cells. Further evidence of mitochondrial damage was observed: 1) Extracellular ATP release, a marker of mitochondrial permeability transition, was significantly elevated in HA@Lip-CS-1@PBCO NP-treated cells compared to PBCO NP-treated controls (Fig. 5M); 2) Malondialdehyde (MDA) levels, a well-established biomarker of lipid peroxidation and oxidative stress [48], increased dramatically (8.7 ± 1.8-fold) in the NPs-treated MCF-7/ADR cells (Fig. 5N), confirming extensive oxidative membrane damage. Given the established link between oxidative stress sensing and the p62/Nrf2 pathway [49], we investigated their involvement. Western blot analysis revealed that HA@Lip-CS-1@PBCO NPs significantly suppressed both p62 and Nrf2 protein levels in MCF-7/ADR cells (Fig. 5O), suggesting impaired activation of the endogenous antioxidant response. Importantly, and critically for overcoming MDR [50, 51], HA@Lip-CS-1@PBCO NPs significantly downregulated key MDR-associated efflux pumps, including P-glycoprotein (P-gp) and ABCG2 (Fig. 5P). This reduction functionally impairs drug efflux, a primary MDR mechanism. Collectively, these findings underscore the remarkable ability of HA@Lip-CS-1@PBCO NPs to induce pyroptosis and circumvent drug efflux, thereby facilitating the elimination of multidrug resistant tumor cells (Fig. 5Q).

HA@Lip-CS-1@PBCO NPs activate pyroptosis via oxidative stress-induced mitochondrial damage. (A) Confocal microscopy images of T11-stained MCF-7/ADR cells following treatment, assessing membrane integrity. (B) Representative bio-TEM micrographs of MCF-7/ADR cells after 8 h exposure to treatments. (C) PI fluorescence imaging of membrane integrity in treated MCF-7/ADR cells. (D) Quantification of LDH release in MCF-7/ADR cells. (E) Western blot analysis of GSDME and cleaved caspase-3 proteins in cells treated for 12 h. (F) Mitochondrial ultrastructure in MCF-7/ADR cells visualized by bio-TEM after 8 h treatment. (G) Mitochondrial morphology assessment using MitoTracker® Red CMXRos in treated cells after 8 h treatment. (H) Cytochrome c release analyzed by western blot after 12 h treatment. (I) JC-1 staining showing mitochondrial membrane potential (ΔΨm) changes after 8 h treatment. (J) Intracellular ROS detection via DCFH-DA fluorescence imaging and flow cytometry after 12 h treatment. (K) MitoSOX™ Red staining for mitochondrial superoxide production after 12 h treatment. (L) GSH levels measured after 24 h treatment. (M) extracellular ATP quantification post 24 h treatment. (N) MDA release in MCF-7/ADR cells after 24 h treatment. (O-P) Western blot analysis of p62, Nrf2, P-gp, and ABCG2 expression (12 h treatment). (Q) Schematic diagram of pyroptosis caused by HA@Lip-CS-1@PBCO NPs. Bars are means ± SD (n = 3). **P P P
Pharmacokinetics and biodistribution of HA@Lip-Ce6@PBCO NPs in vivo
The pharmacokinetic profile of HA@Lip-Ce6@PBCO NPs was evaluated by monitoring Ce6 fluorescence intensity in blood samples post-injection. As shown in Fig. 6A, blood samples showed a gradual decrease in fluorescence intensity following intravenous injection. The NPs exhibited significantly prolonged circulation, with a half-life (t₁/₂) of 3.4 h in BALB/c nude mice – 2.3-fold longer than free Ce6 (1.5 h) (Fig. 6B). Real-time fluorescence imaging revealed distinct biodistribution patterns. HA@Lip-Ce6@PBCO NPs accumulated preferentially at tumor sites, with fluorescence intensity of Ce6 and Lip-Ce6@PBCO NPs increasing progressively and plateauing at 6 h post-injection. Tumor fluorescence intensity in the HA-modified NP group consistently exceeded that of Lip-Ce6@PBCO NPs and free Ce6 at 6, 8, and 12 h (Fig. 6C), demonstrating HA-mediated active targeting. In contrast, free Ce6 exhibited weak, rapidly declining fluorescence due to non-specific distribution while the fluorescent intensity of Lip-Ce6@PBCO NPs was higher than that in the free Ce6 group due to the EPR effect, and HA@Lip-Ce6@PBCO NPs exhibited the strongest tumor fluorescent intensity. Ex vivo analysis at 12 h post-injection showed fluorescence signals primarily localized in the liver and lungs (Fig. 6D), attributable to hepatic metabolism and residual circulating NPs [52]. Notably, minimal cardiac fluorescence indicated reduced cardiotoxicity compared to free Ce6. Tumor fluorescence intensity in the HA@Lip-Ce6@PBCO NPs group was 2.6 ± 0.6-fold higher than in the free Ce6 group, confirming enhanced tumor targeting.
We next evaluated the ultrasound imaging potential of HA@Lip-Ce6@PBCO NPs in vivo. Following intravenous injection into tumor-bearing mice, ultrasound imaging revealed a detectable acoustic signal at the tumor site, which was due to the stimulation of endogenous H2O2 in the tumor microenvironment (TME) triggering the release of CO from the NPs. This gas generation significantly amplified the intratumoral acoustic signal compared to pre-injection levels (Fig. 6E). The pronounced signal enhancement demonstrates both successful tumor-targeted delivery of HA@Lip-CS-1@PBCO NPs and H2O2-responsive CO generation within the TME. Overall, these findings highlight the nanocomplexes’ strong targeting ability and prolonged blood half-life, essential for its therapeutic efficacy in animals.

In vivo biodistribution and pharmacokinetics of Ce6-labeled HA@Lip-Ce6@PBCO NPs. (A) Time-dependent blood fluorescence intensity (FI) profiles of free Ce6 vs. HA@Lip-Ce6@PBCO NPs. (B) Pharmacokinetic parameters of Ce6 and HA@Lip-Ce6@PBCO NPs. (C) Fluorescence images of MCF-7/ADR tumor-bearing mice at various time points following administration of free Ce6, Lip-Ce6@PBCO NPs, or HA@Lip-Ce6@PBCO NPs (2.5 mg/kg of equivalent Ce6). Tumor sites are denoted by yellow dashed line circles. (D) Fluorescence distribution of major organs and tumors after 12 h post-injection. (E) Ultrasound imaging of tumor in nude mice before and after with HA@Lip-CS-1@PBCO NPs injection for 10 min. Bars are means ± SD (n = 3)
HA@Lip-CS-1@PBCO NPs exhibit potent antitumor activity against multidrug-resistant cancer in vivo
Building on their in vitro cytotoxicity and in vivo tumor-targeting capabilities, we evaluated the therapeutic efficacy of HA@Lip-CS-1@PBCO NPs in nude mice bearing MCF-7/ADR xenografts according to the regimen in Fig. 7A. Compared to PBS group, all treatment groups (CS-1, PBCO NPs, HA@Lip-CS-1@PBCO NPs) showed significantly reduced tumor volumes (Fig. 7B). Ex vivo analysis confirmed striking tumor growth inhibition by HA@Lip-CS-1@PBCO NPs. Imaging and gravimetric quantification revealed a final tumor weight of 20.2 ± 9.4 mg in the HA@Lip-CS-1@PBCO NP group versus 112.8 ± 58.6 mg in PBS group (Fig. 7C-D), translating to a tumor inhibition rate (TIR) of 82.2% ± 5.4%. This efficacy significantly surpassed both free CS-1 (TIR: 54.3% ± 5.2%) and PBCO NPs (TIR: 49.5% ± 15.2%) (Fig. 7E). Critically, no significant body weight loss was observed across any group, supporting the biocompatibility of the nanomaterials (Fig. 7F). Moreover, H&E staining demonstrated extensive, demarcated tissue necrosis in all treatment groups (CS-1, PBCO NPs, HA@Lip-CS-1@PBCO NPs), contrasting sharply with the intact morphology and nuclear density of PBS-treated tumors (Fig. 7G). HA@Lip-CS-1@PBCO NPs treatment induced the most pronounced reduction in cellular proliferation, evidenced by the lowest Ki67 expression (Fig. 7G). Furthermore, significant upregulation of the pyroptosis executioner GSDME and the apoptosis-to-pyroptosis switch marker cleaved caspase-3 was detected via immunofluorescence in HA@Lip-CS-1@PBCO NP-treated tumors (Fig. 7H-I). These results demonstrate that HA@Lip-CS-1@PBCO NPs exhibit potent antitumor activity against MDR breast cancer in vivo.
We further assessed therapeutic activity in an A2780/DDP (cisplatin-resistant) xenograft model using the same regimen (Fig. 7J). HA@Lip-CS-1@PBCO NPs achieved a final tumor volume of 38.0 ± 24.5 mm³, markedly smaller than those of PBS group (236.8 ± 59.5 mm³; Fig. 7K). Ex vivo tumor imaging and weight quantification corroborated this potent efficacy: HA@Lip-CS-1@PBCO NP-treated tumors weighed 42.3 ± 12.5 mg, representing a 67.1% reduction compared to PBS group (128.7 ± 57.9 mg) (Fig. 7L-M). The corresponding TIR was 67.2% ± 6.0% (Fig. 7N). Body weight remained stable (Fig. 7O), and histological evaluation recapitulated the MCF-7/ADR findings: HA@Lip-CS-1@PBCO NP-treated A2780/DDP tumors exhibited significant necrosis (Fig. 7P), minimal Ki67 staining (Fig. 7P), and strong induction of cleaved caspase-3 and GSDME characteristic of pyroptosis (Fig. 7Q). These results robustly demonstrate that HA@Lip-CS-1@PBCO NPs efficiently inhibit the growth of MDR breast tumors and exhibit potent, broad-spectrum activity against diverse MDR cancers in vivo.

In vivo anti-MDR tumor efficacy of HA@Lip-CS-1@PBCO NPs. (A) Schematic illustration of MCF-7/ADR tumor implantation and the dosage regimen (I: PBS; II: CS-1; III: PBCO NPs; IV: HA@Lip-CS-1@PBCO NPs). (B) Tumor growth curves. (C) The morphological images of tumors in different group. (D) Tumor weights in different group. (E) Tumor inhibitory rates (TIR) in the mice with different treatment. (F) Body weight change of mice. (G) H&E and Ki67 staining of tumor sections. (H-I) Immunofluorescence staining of cleaved caspase-3 and GSDME in tumor tissues. (J) Diagrammatic representation of A2780/DDP tumor implantation and the dosing schedule. (I: PBS group; II: HA@Lip-CS-1@PBCO NPs group). (K) Tumor growth curves. (L) Excised tumor morphology. (M) Changes of tumor weight with different treatment. (N) TIR of HA@Lip-CS-1@PBCO NPs treatment. (O) Body weight monitoring. (P) H&E and Ki67-stained tumor sections. (Q) Immunofluorescence of cleaved caspase-3&GSDME in different tumor sections. Bars are means ± SD (n = 5). *P P P P
Biocompatibility and biosafety evaluation of HA@Lip-CS-1@PBCO NPs
Given the critical importance of biocompatibility for clinical translation, we rigorously evaluated the safety profile of HA@Lip-CS-1@PBCO NPs through multiple assays: hemolysis, platelet aggregation, cytotoxicity in normal cells, zebrafish embryo toxicity, and systemic toxicity in tumor-bearing mice. As shown in Fig.S5A-B, all tested formulations (PBS, CS-1, PBCO NPs, HA@Lip-CS-1@PBCO NPs) exhibited excellent blood compatibility, with hemolysis rates below the 5% safety threshold after 6 h incubation with red blood cells (RBCs). Furthermore, HA@Lip-CS-1@PBCO NPs significantly suppressed platelet aggregation, as evidenced by an optical density (OD650 nm) value of 94.8% compared to 48.6% for thrombin (Fig.S5C). MTT assays confirmed low cytotoxicity of HA@Lip-CS-1@PBCO NPs. Cell viability in VSMC, NIH/3T3, and H9c2 cells remained above 95% after 24 h exposure, significantly higher than free CS-1 and PBCO NPs in NIH/3T3 and H9c2 cells (Fig.S5D). In addition, zebrafish embryotoxicity testing revealed no significant differences in body length across treatment groups (Fig. 8A-B). Notably, HA@Lip-CS-1@PBCO NPs (122 beats per minute) mitigated the cardiotoxic effect observed with free CS-1 (130 beats per minute), demonstrating a significantly lower heart rate (Fig. 8C). This indicates the nanoformulation effectively reduces inherent cardiac adverse effects.
In vivo biosafety assessment was essential for evaluating the clinical feasibility of nanomedicines. We investigated the effect of HA@Lip-CS-1@PBCO NPs by performing whole blood cell count and liver and renal function analysis. In MCF-7/ADR-bearing mice, HA@Lip-CS-1@PBCO NPs showed good systemic safety profile. Complete blood counts revealed no adverse effects on RBC, PLT, or HGB levels. Importantly, treated mice exhibited a significant decrease in white blood cell (WBC) count compared with PBS, suggesting attenuation of the tumor-associated inflammatory response (Fig. 8D). Liver and kidney function markers (ALT, AST, CRE, URE) remained within normal ranges, confirming no hepatorenal toxicity (Fig. 8E). H&E staining of key organs (heart, liver, spleen, lung, kidney) showed no notable lesions or morphological changes (Fig. 8F). Collectively, these results demonstrate that HA@Lip-CS-1@PBCO NPs possess excellent biocompatibility, effectively reduced cardiotoxicity compared to the free drug, and low systemic toxicity, supporting their potential for further clinical development.

Biosafety evaluation of HA@Lip-CS-1@PBCO NPs. (A) The microscopic images of zebrafish were treated with PBS, CS-1 (1 µM), PBCO NPs (40 µM), and HA@Lip-CS-1@PBCO NPs (n = 5). (B-C) Body length and heart rate of zebrafish with different treatments. (D) Complete blood count of each MCF-7/ADR tumor bearing mice group, including WBC, RBC, PLT, and HGB (I: PBS; II: CS-1; III: PBCO NPs; IV: HA@Lip-CS-1@PBCO NPs). (E) Analysis of blood biochemistry in each group of MCF-7/ADR tumor-bearing mice, focusing on liver function indicators such as AST and ALT, as well as kidney function markers like CRE and UREA. (F) H&E stained images of major organs in each group. Bars are means ± SD (n = 5). *P P P

