Hierarchically collapsible nanoactuator modulates mitochondrial ferroptosis-bioenergetic homeostasis cascade to decouple ischemic stroke

Summary

Ischemic stroke, a life-altering cerebrovascular emergency triggered by prolonged cerebral hypoperfusion, remains a therapeutic enigma. Current interventions struggle with ischemia-reperfusion injury; restoring blood flow unleashes reactive oxygen species (ROS), driving secondary neuronal damage and functional loss. Ischemia-induced mitochondrial dysfunction heightens oxidative stress and hastens neuronal death. We address oxidative-stress-driven neuronal injury by engineering a hierarchically collapsible nanoactuator suppressing mitochondrial ferroptosis and restoring cellular energy homeostasis. The nanoactuator integrates a diselenide-crosslinked shell conjugated with a mitochondrial-targeting peptide, enabling blood-brain barrier penetration and mitochondrial delivery. Its collapsible core, composed of an ATP-gadolinium coordination polymer encapsulating a ferroptosis inhibitor, enables MRI-guided tracking and ROS-responsive drug release. In damaged mitochondria, the nanoactuator replenishes ATP, restores membrane potential, reduces ROS levels, and alleviates ferroptosis. Intravenous administration in a transient middle cerebral artery occlusion (tMCAO) mouse model demonstrated robust multi-mechanistic neuroprotection. This hierarchical nanoactuator platform offers a strategy for ischemic stroke and related neurodegenerative diseases.

Graphical abstract

Highlights

• •Mit-CY@Nps sequentially target the mitochondrial ferroptosis-energy cascade
• •ATP-Gd core restores mitochondrial potential and neuronal bioenergetics
• •Mitochondrial targeting with ACSL4 inhibition synergistically blocks ferroptosis
• •Mit-CY@Nps provide strong neuroprotection with MRI-visible therapeutic response

Sun et al. report a hierarchically collapsible nanoactuator that restores mitochondrial function and blocks iron-driven oxidative damage, reducing neuronal injury after stroke. This work highlights a promising direction for improving outcomes in ischemic brain disease.

Introduction

Ischemic stroke, caused by acute and sustained reductions in cerebral perfusion due to ischemia and hypoxia, is characterized by rapid onset and high rates of mortality and long-term disability, often resulting in persistent neurological sequelae.^1,2 Upon reperfusion, the reintroduction of oxygen and nutrients triggers ROS production, initiating a cascade of oxidative damage, mitochondrial dysfunction, and neuronal loss—collectively referred to as cerebral ischemia-reperfusion (I/R) injury.^3,4 Ferroptosis represents a recently identified form of regulated cell death whose critical role in stroke is recognized.⁵ Oxygen and substrate deprivation disrupts mitochondrial function and adenosine triphosphate (ATP) production, destabilizing cellular homeostasis.⁶ The resulting bioenergetic failure increases AMP and inorganic phosphate (Pi), initially activating AMP-activated protein kinase (AMPK) but eventually sensitizing the mitochondrial permeability transition pore (mPTP).⁷ Pathological mPTP opening collapses membrane integrity, releases mitochondrial contents, amplifies oxidative stress, and promotes iron-dependent lipid peroxidation (LPO), leading to ferroptosis.^8,9 Therefore, the convergence of the ferroptosis-energetics axis, encompassing adenine nucleotide catabolism, phosphate accumulation, mPTP opening, and ferroptotic signaling, constitutes a self-amplifying pathological cascade that accelerates neuronal death and aggravates ischemic brain injury.¹⁰

The brain exhibits intrinsic susceptibility to ferroptosis owing to its high content of polyunsaturated fatty acids, elevated oxygen consumption, and substantial iron abundance.¹¹ During I/R, oxidative stress coupled with perturbed iron homeostasis establishes a biochemical milieu for ferroptotic execution. Excessive lipid ROS depletes glutathione (GSH) and inactivates glutathione peroxidase 4 (GPX4).^12,13 Moreover, acyl-CoA synthetase long-chain family member 4 (ACSL4), an enzyme involved in polyunsaturated fatty acid (PUFA) metabolism, has emerged as a key regulator of ferroptosis with a significant impact on stroke recovery.^14,15 For instance, galangin protects hippocampal neurons by activating the SLC7A11/GPX4 axis,¹² and vitamin E attenuates oxidative stress and inflammation by maintaining GSH levels and reducing the expression of genes like SLC7A11, tumor necrosis factor alpha (TNF-α), and interleukin (IL)-8.¹⁶ Carthamin yellow (CY), a flavonoid extracted from safflower, has shown efficacy beyond its antioxidant capacity; it exhibits anti-inflammatory properties and the ability to specifically inhibit ACSL4.^17,18 However, small-molecule ferroptosis inhibitors face limitations including insufficient targeting and poor barrier penetration,^19,20 low bioavailability,²¹ and the reliance on single therapeutic mechanisms that lack the precision and synergistic effects required for robust treatment outcomes.¹⁹

Mitochondria are central hubs for cellular energy metabolism, ATP synthesis, and various physiological processes.²² Mitochondria also act as primary intracellular iron reservoirs, where iron participates in essential biochemical processes such as cytochrome P450 biosynthesis, heme production, and electron transport chain activity.^23,24 Excessive iron accumulation triggers Fenton reactions, producing ROS that damages the mitochondrial inner membrane, impair mitochondrial function, and exacerbate ferroptosis.^25,26 Mitochondrial membrane rupture not only halts ATP synthesis but also increases intracellular calcium levels, further intensifying ferroptosis.²⁷ Additionally, ROS damages neuronal membranes, proteins, and DNA, amplifying oxidative stress and driving LPO, compounding cellular injury.^28,29 Given the centrality of mitochondria in orchestrating both metabolic dysfunction and ferroptotic signaling, precise intervention at this subcellular nexus is critical.³⁰

This study reports a hierarchically collapsible nanoactuator (Mit-CY@Nps) designed to modulate the mitochondrial ferroptosis-energetic homeostasis cascade to treat cerebral ischemia-reperfusion injury. The hierarchical architecture comprises a gadolinium-ATP co-assembled porous inner network loaded with CY and ROS-triggered ATP core collapse for localized mitochondrial energy replenishment. CY, embedded within the inner matrix, acts as the principal therapeutic agent by targeting ACSL4 to suppress PUFA peroxidation, thereby exerting antioxidant, anti-ferroptotic, and anti-inflammatory effects.¹⁷ The redox-collapse outer network consists of a selenium-based polymer and is functionalized with the mitochondrial-targeting peptide SS31, enabling precise subcellular delivery. Unlike conventional antioxidants, it accumulates within mitochondria, stabilizes cristae architecture, and enhances electron transport chain efficiency, thereby reducing mitochondrial ROS production at its biochemical origin.³¹ Acting in a temporally ordered manner, Mit-CY@Nps first restore mitochondrial membrane potential (Δψ_m) and structural integrity through targeted ATP delivery, preventing the pathological release of iron ions and ROS into the cytosol. The Mit-CY@Nps introduce a paradigm-shifting in stroke therapeutics by addressing the complex interplay between iron-induced oxidative damage and mitochondrial energy failure, ameliorate cerebral I/R injury, and facilitate neuronal repair by enabling a synergistic intervention targeting key pathological processes.

Results

Synthesis and characterization of Mit-CY@Nps

The Mit-CY@Nps feature a metal-coordination polymer inner core that forms a porous network encapsulating CY with antioxidant,³² anti-inflammatory,^33,34 and ferroptosis-inhibiting effects,^35,36 making it an potential therapeutic payload.¹⁷ This core is surrounded by a diselenide-containing polymer shell that enables ROS-responsive degradation, allowing sequential modulation of the mitochondrial ferroptosis-energetic homeostasis cascade in cerebral I/R injury (Figure 1A). The Gd-ATP porous inner scaffold (Gd-ATP NMs) was synthesized via in situ coordination of Gd³⁺ with ATP within DSPE-PEG-NH₂ templates. Transmission electron microscopy (TEM) revealed uniformly sized, well-dispersed, porous nanostructures (Figures S2A and S2B). Energy-dispersive X-ray spectroscopy (EDX) further verified the presence of phosphorus (P) and gadolinium (Gd), confirming successful coordination (Figures S2C–S2E). Nitrogen adsorption experiments confirmed the mesoporous nature of the structure (Figure S2F). A diselenide crosslinker (DSeDPA) was conjugated to DSPE-PEG-NH₂ on Gd-ATP NMs, followed by linkage of the SS-31, yielding the Mit-CY@Nps. TEM images revealed that the final material maintained uniform size and dispersibility with a smoother surface, confirming successful encapsulation by the diselenide polymer (Figure 1B). EDX analysis confirmed the presences of P, Gd, and Se, validating the formation of Mit-CY@Nps (Figures 1C–1E and S2G). Additionally, dynamic light scattering (DLS) measurements revealed an increase in particle size from 28 nm for the Gd-ATP NMs scaffold to 63 nm for Mit-CY@Nps, confirming the selenium outer network (Figure 1F). Zeta potential measurements highlighted a shift in surface charge from approximately −14 mV for the Gd-ATP NMs scaffold to −4 mV for Mit-CY@Nps, further confirming the encapsulation of the outer network and the conjugation of the positively charged peptide SS-31 (Figures 1G and S2H). Fourier transform infrared spectroscopy (FT-IR) analysis corroborated the incorporation of all key components in Mit-CY@Nps, with absorption peaks corresponding to ATP, SS-31, and DSeDPA-NHS (Figure S2I). X-ray photoelectron spectroscopy (XPS) provided further evidence of the element compositions of Mit-CY@Nps, detecting peaks associated with Gd, P, Se, and their respective bonding states (Figure 1H). The Gd_4d spectrum revealed characteristic peaks at 141.5 and 148.0 eV, and the Se_3d spectrum displayed peaks at 55.0 and 56.3 eV, corresponding to Se 3d_5/2 and Se 3d_3/2, respectively (Figure S2J). Stability studies using DLS (Figure S2K) indicated no significant changes in Mit-CY@Nps, confirming their stability under physiological conditions.

Figure 1.

Preparation, construction, and characterization of Mit-CY@Nps

(A) Schematic illustration of Gd-ATP NMs precursor formation and subsequent co-assembly into Mit-CY@Nps nanoparticles.

(B) TEM image of Mit-CY@Nps. Scale bars: 100 nm (main image) and 20 nm (inset).

(C–E) Elemental mapping of Mit-CY@Nps. Scale bars: 100 nm.

(F) DLS of Gd-ATP NMs and Mit-CY@Nps (n = 3).

(G) Zeta potentials of Gd-ATP NMs and Mit-CY@Nps (n = 3).

(H) X-ray diffraction pattern of Gd-ATP NMs and Mit-CY@Nps.

(I) UV-vis absorption spectra of CY, Gd-ATP NMs, and CY@Gd-ATP NMs.

(J) In vitro release profiles of CY from Mit-CY@Nps with or without incubation with 1 mM H₂O₂ at 37°C. (pH = 7.4, n = 3).

(K) Schematic representation of the ROS-triggered hierarchical collapse and drug release mechanism of Mit-CY@Nps.

(L) Particle size distributions of Mit-CY@Nps under oxidative conditions.

(M) TEM images of Mit-CY@Nps incubated with PBS or H₂O₂ for different durations. Scale bars: 20 nm.

(N) Schematic diagram illustrating MRI contrast enhancement via Gd³⁺-mediated T₁ relaxation.

(O) T₁-weighted MRI images of Magnevist, Gd-ATP NMs, and Mit-CY@Nps at pH 7.4 (n = 3).

(P) Longitudinal relaxivity (r₁) values of Magnevist, Gd-ATP NMs, and Mit-CY@Nps (n = 3). The numerical data in (G, J, and P) are presented as the mean ± SD. See also Figures S1–S3.

Hierarchically collapsible behavior and MRI functions

To verify the successful encapsulation of CY in Mit-CY@Nps, ultraviolet-visible spectroscopy (UV-Vis) measurement revealed a characteristic absorption peak at 410 nm for CY@Gd-ATP NMs, consistent with free CY (Figure 1I). Encapsulation efficiency was calculated to be 37.5% using free CY as standard, via UV-Vis measurements (Figures S2L–2N). Thermal gravimetric analysis (TGA) demonstrated weight loss in the range of 225°C–625°C for Mit-CY@Nps, indicative of CY thermal decomposition, further confirming successful encapsulation (Figure S2O). The release of CY was compared in PBS and 1 mM H₂O₂ to mimic the high inflammatory condition in ischemic stroke site. As illustrated (Figure 1J), in the presence of H₂O₂, the release rate of CY accelerated, with a rapid cumulative release reaching about 90% within the initial 4 h, eventually stabilizing at approximately 100%. CY release in PBS was slower, with a cumulative release increasing to around 40% over 24 h. Besides, the decomposition of Gd-ATP NMs was also evaluated in presence of H₂O₂ by monitoring the release of ATP using high-performance liquid chromatography (HPLC). In the H₂O₂, ATP release increased over time, with consistent retention times (∼2.45 min) confirming ATP identity (Figure S2P). To elucidate the nanostructure disassembly process, proton nuclear magnetic resonance (¹H NMR) spectroscopy was employed to assess the ROS-responsive cleavage of diselenide bonds at the molecular level (Figures 1K and S3). The diselenide bond structure was identified by comparison with DSPE-PEG-NH₂. Upon addition of H₂O₂, a clear shift in these peaks was observed, indicating oxidative cleavage of the diselenide bonds and the formation of selenium oxides. The structural disintegration of Mit-CY@Nps in ROS environments was further examined through DLS and TEM. DLS analysis (Figure 1L) revealed an increase in particle size and distribution breadth over time by 8 h in an H₂O₂ environment, reflecting extensive structural disintegration. Morphological changes of the nanoscale regulator were visualized using TEM (Figure 1M). In PBS, particles retained uniform spherical, smooth structures at 0 h with minimal degradation at 2, 4, and 8 h. In H₂O₂, initial degradation emerged at 2 h, marked structural collapse at 4 h, and complete disintegration into irregular fragments by 8 h.

In addition to degradability studies, Mit-CY@Nps were evaluated for its potential in MRI applications. The incorporated paramagnetic Gd³⁺ centers, commonly utilized as T₁-weighted contrast agents, can enhance signal intensity by modulating the local magnetic microenvironment and accelerating T₁ relaxation (Figure 1N).³⁷ The imaging performance of Mit-CY@Nps, Gd-ATP NMs, and Magnevist (a commercial Gd³⁺ contrast agent) was compared in vitro (Figure 1O). Increasing Gd³⁺ concentrations resulted in enhanced T₁ image brightness for all samples. Mit-CY@Nps and Gd-ATP NMs exhibited notable imaging performance compared to Magnevist, with Gd-ATP NMs achieving the greatest brightness enhancement. Quantitative analysis of T₁ relaxation rates (Figure 1P) demonstrated a linear relationship with Gd³⁺ concentration for all materials, with Gd-ATP NMs showing the highest relaxation enhancement.

Mitochondrial targeting and blood-brain barrier penetration of the Mit-CY@Nps in vitro

PC12 cells, a widely used chromaffin derived neuronal model,³⁸ are employed to establish oxygen-glucose deprivation/reperfusion (OGD/R) models, a classical in vitro paradigm of ischemic stroke.^39,40 Cytotoxicity assays demonstrated good cellular tolerance, maintaining exceeding 80% viability even at concentrations up to 800 μg/mL (Figure S4A). To determine the internalization process, fluorescein isothiocyanate (FITC)-labeled Mit-CY@Nps were incubated with PC12 cells and tracked via confocal laser scanning microscopy (CLSM). Quantitative fluorescence analysis revealed maximal cellular uptake at 6 h, followed by a gradual decline at 8 and 12 h (Figures 2A and 2B). Co-localization studies were conducted using FITC-labeled Mit-CY@Nps (green) and MitoTracker-labeled mitochondria (red). Merged fluorescence images revealed strong spatial overlap with mitochondrial signals (Figure 2C), and fluorescence intensity (Figure 2D) confirmed that the signal peaks of Mit-CY@Nps aligned closely with those of mitochondria, particularly at 4 and 6 h post-incubation.

Figure 2.

Mitochondrial-targeting and BBB-penetrating capability

(A) Time-dependent CLSM images of FITC-Mit-CY@Nps uptake in PC12 cells. Scale bars: 100 μm (main images) and 50 μm (insets).

(B) Quantified fluorescence intensity of cellular uptake (n = 3).

(C - D) CLSM images and co-localization analysis of Mit-CY@Nps (green) with mitochondria (red). Scale bars: 100 μm (C, left) and 20 μm (C, right).

(E) Schematic of the Transwell BBB co-culture model using bEnd.3 (upper) and PC12 cells (lower).

(F) Confocal images of Mit-CY@Nps internalization in PC12 cells after BBB translocation at various time points. Scale bars: 100 μm.

(G) Quantified fluorescence intensity in PC12 cells (n = 3).

(H) Fluorescence distribution profile showing intracellular localization of Mit-CY@Nps. The numerical data in (B and G) are presented as the mean ± SD. ∗p < 0.05.

A one-way ANOVA with Tukey’s multiple comparisons test (G) was used for statistical significance analysis. See also Figure S4.

The blood-brain barrier (BBB), formed by tight junctions between brain microvascular endothelial cells, restricts the entry of most exogenous substances and limits drug delivery for neurological diseases.⁴¹ Enhancing penetration across the BBB and accumulation at injury sites is therefore essential for effective stroke therapy.⁴² To simulate BBB penetration and neuronal targeting, a co-culture system was established using a Transwell model with bEnd.3 cells in the upper chamber and PC12 cells in the lower chamber (Figure 2E). CLSM images (Figure 2F) demonstrated the progressive accumulation of FITC-labeled Mit-CY@Nps in PC12 cells at 4, 6, 8, and 12 h, with maximum green fluorescence observed at 6 h. Quantitative fluorescence intensity analysis revealed that uptake at 6 h was 1.8-fold higher than at 4 h (Figure 2G). Co-localization studies further validated the intracellular delivery of Mit-CY@Nps within this time frame (Figure 2H). These results demonstrated that Mit-CY@Nps possess high BBB permeability and effective spatiotemporal targeting across tissue, cellular, and subcellular levels.

Modulation of the mitochondrial ferroptosis-energetic homeostasis cascade in vitro by Mit-CY@Nps

Under pathological conditions such as mitochondrial damage or ischemia-reperfusion, impaired electron transport chain function decreases Δψ_m and blocks ATP synthesis, leading to an energy crisis.²⁸ Electron leakage simultaneously increases ROS, which, with iron ions, induces LPO through the Fenton reaction.⁴³ GSH depletion and GPX4 inactivation exacerbate lipid oxidative damage, culminating in ferroptosis.⁴⁴ This highlights the bidirectional interplay between mitochondrial metabolic imbalance and ferroptotic cell death (Figure 3A). We employed the OGD/R oxidative stress model to simulate stroke-specific cellular damage.⁴⁵ A concentration of 200 μg/mL Mit-CY@NPs was selected to optimize therapeutic efficacy while minimizing cytotoxicity (Figures S4A and S4B). Intracellular ROS levels were assessed using the 2′,7′-dichlorofluorescin diacetate (DCFH-DA). Confocal imaging showed pronounced ROS accumulation in the PBS-treated OGD/R group, whereas Mit-CY@Nps markedly reduced ROS signals (Figure 3B), with quantitative analysis indicating a 79% decrease (Figure 3C). Mitochondrial ROS, evaluated by MitoSOX Red staining, were similarly elevated after OGD/R but attenuated by Mit-CY@Nps compared with other groups (Figures S4C and S5A). This enhanced antioxidant efficacy arises from the synergistic interplay between CY⁴⁶ and the diselenide network, which efficiently neutralizes ROS via reversible redox cycling,⁴⁷ while SS31 further amplifies therapeutic efficiency by promoting selective accumulation of Mit-CY@Nps within the mitochondrial membrane.³¹ To evaluate the anti-ferroptotic potential of Mit-CY@Nps, lipid peroxidation levels were assessed using the fluorescent probe BODIPY 581/591 C11, which exhibits a red-to-green emission shift upon oxidation of membrane lipids. In the OGD/R model (Figure 3D), PBS-treated cells displayed intense green fluorescence, indicative of lipid peroxidation and ferroptotic progression. Mit-Nps or CY alone partially reduced green fluorescence, reflecting moderate antioxidative protection. Cells treated with Mit-CY@Nps exhibited a dominant red emission with markedly diminished green fluorescence, approaching the level observed in the sham group. These results demonstrate that Mit-CY@Nps blocked OGD/R-induced lipid oxidation, thereby suppressing ferroptotic cell death. To further confirm the ferroptosis-inhibitory capacity of Mit-CY@Nps, a classical erastin-induced ferroptosis model was employed (Figure 3E). Treatment with either Mit-Nps or CY alleviated lipid oxidation, while Mit-CY@Nps almost completely suppressed green fluorescence, restoring red emission and indicating a significant reduction in ferroptotic activity. Biochemical quantifications further confirmed these findings. Figure S5B shows that Mit-CY@Nps reduced the intracellular level of malondialdehyde (MDA) compared with the PBS, Mit-Nps, and CY groups. Figure S5C demonstrates that Mit-CY@Nps treatment restored the intracellular GSH content, indicating an enhanced antioxidant defense system. These results demonstrate that Mit-CY@Nps attenuate erastin-induced ferroptosis in PC12 cells by reducing lipid peroxidation and reestablishing redox homeostasis.

Figure 3.

Mit-CY@Nps alleviates the mitochondrial ferroptosis-energetic homeostasis cascade in vitro

(A) Schematic of the ferroptosis-energetic homeostasis cascade.

(B) Fluorescence images of intracellular ROS (DCFH-DA) after OGD/R. Scale bars: 100 μm.

(C) Quantified ROS-scavenging ability of Mit-Nps, CY, and Mit-CY@Nps in OGD/R-treated PC12 cells (n = 3).

(D) BODIPY 581/591 C11 staining showing lipid peroxidation across treatment groups after OGD/R. Scale bars: 100 μm.

(E) LPO analysis under erastin stimulation using BODIPY 581/591 C11. Scale bars: 100 μm.

(F) Cell viability of ferroptosis-modeled M2-type BV2 cells after treatment for 6 h (n = 4).

(G) Flow cytometry plots of microglial activation under LPS stimulation.

(H) Intracellular ATP levels under different treatments (n = 3).

(I) JC-1 staining of mitochondrial membrane potential after OGD/R. Scale bars: 100 μm.

(J and K) Activities of respiratory chain complexes I and III in OGD/R-treated PC12 cells (n = 3).

(L) TEM images of mitochondrial morphology after treatments. Scale bars: 200 nm.

(M) Quantification of early/late apoptosis across treatment groups (n = 3).

(N) Flow cytometry analysis of neuronal apoptosis after treatments.

The numerical data in (C, F, H, J, K, and M) are presented as the mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, ns: not significant.

A one-way ANOVA with Tukey’s multiple comparisons test (C, F, H, J, K, and M) was used for statistical significance analysis. See also Figures S4–S6.

Given the high sensitivity of HT22 cells to ferroptotic stimuli, this model was further employed to validate the ferroptosis-modulatory effects. Mit-CY@Nps treatment alleviated erastin-induced lipid peroxidation, as evidenced by a diminished green fluorescence shift in BODIPY 581/591 C11 staining (Figure S5D) and a concomitant reduction in MDA levels (Figure S5E). GSH content was restored following Mit-CY@Nps treatment (Figure S5F). These findings confirm that Mit-CY@Nps exert potent, reproducible anti-ferroptotic effects in PC12 and HT22 neuronal cells. While PC12 cells were used to model neuronal ischemic injury, the consistent findings in HT22 cells further confirm the robustness and generalizability of Mit-CY@Nps-mediated ferroptosis inhibition across neuronal cell types, reinforcing its mechanistic link to neuroprotection against ischemia-induced oxidative damage.

Following I/R injury, a subset of microglia polarizes toward the anti-inflammatory M₂ phenotype, which facilitates tissue repair and suppresses inflammatory progression. Accumulating evidence indicates that M₂-type microglia are more susceptible to ferroptosis,^48,49 which may contribute to uncontrolled inflammatory responses during the later stages of ischemia-reperfusion. We evaluated the ability of Mit-Nps, CY, and Mit-CY@Nps to rescue ferroptosis in M₂-polarized BV2 microglia (Figure 3F). All three treatments exhibited varying degrees of efficacy in restoring the viability of M₂-type microglia, suggesting a potential anti-ferroptotic effect. To investigate the regulatory effects of Mit-CY@Nps on microglial polarization, flow cytometry analysis was performed using BV2 microglial cells. As shown in Figure 3G, LPS stimulation increased the proportion of CD86⁺ M₁-type cells while reducing the CD206+ M₂ population. Treatment with Mit-Nps or CY alleviated this imbalance, whereas Mit-CY@Nps suppressed the proportion of CD86⁺CD206− cells and increased the CD86⁻CD206+ M₂ population (Figures S6A and S6B). These results show that Mit-CY@Nps efficiently induce microglial polarization toward the anti-inflammatory M₂ phenotype. This effect stems from restored mitochondrial bioenergetics, which reprograms metabolism and inflammatory signaling to re-establish microglial homeostasis. Mit-CY@Nps not only protect M₂-type microglia from ferroptotic damage but also drive local microglial polarization toward an anti-inflammatory and tissue-repairing phenotype, thereby modulating the post-ischemic inflammatory microenvironment in a neuroprotective manner.

Mitochondrial dysfunction is a key pathological feature of I/R injury and a central driver of ferroptosis. Impaired electron transport increases ROS, collapses Δψ_m, and depletes ATP, forming a cycle of oxidative damage and cell death. To investigate whether Mit-CY@Nps could restore mitochondrial bioenergetic homeostasis, intracellular ATP levels were first quantified. As shown in Figure 3H, ATP content was reduced in the PBS-treated OGD/R group, indicating severe mitochondrial energy collapse. Mit-CY@Nps treatment significantly restored ATP production, demonstrating its ability to counteract hypoxia-induced bioenergetic failure. One of the fundamental causes of neuronal ferroptosis in brain ischemia-reperfusion injury is the disruption of mitochondrial membrane integrity, leading to the release of iron ions and ROS from the mitochondria into the cytoplasm.⁵⁰ Excessive iron accumulation generates ROS through the Fenton reaction, exacerbating LPO, damaging the cell membrane, and promoting the occurrence of ferroptosis.⁵¹ Restoring mitochondrial Δψ_m is essential for inhibiting ferroptosis. Δψ_m was assessed using the 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethyl-imidacarbocyanine (JC-1 probe), where red fluorescence (J-aggregates) indicates high Δψ_m and green fluorescence (monomers) reflects depolarization. CLSM analysis revealed substantial Δψ_m loss in the PBS-treated OGD/R group, indicative of severe membrane potential loss. However, Mit-CY@Nps treatment restored Δψ_m, as evidenced by increased red fluorescence and reduced green fluorescence (Figure 3I), confirming its ability to preserve mitochondrial bioenergetic function. Statistical analysis showed that treatment with Mit-NPs, CY, and Mit-CY@NPs enhanced mitochondrial depolarization in PC12 cells subjected to OGD/R (Figure S6C). This effect arises from the distinct yet complementary actions of each component: SS31 for mitochondrial targeting and stabilization by enhancing electron transport,³¹ and CY reduces oxidative damage by scavenging ROS and preserving Δψ_m. Moreover, Mit-CY@Nps treatment restored mitochondrial respiratory activity, as evidenced by increased complex I and complex III enzyme activities (Figures 3J and 3K) and enhanced oxygen consumption rate (OCR) (Figure S6D). These findings indicate that Mit-CY@Nps not only responds to ROS for controlled ATP/CY release but also alleviates mitochondrial dysfunction under ischemic conditions. Therefore, the nanoactuator contributes to preserving mitochondrial homeostasis, providing further mechanistic support for its neuroprotective effects. BioTEM revealed severe mitochondrial damage in the OGD/R group, including swelling, cristae disruption, and vacuolization (Figure 3L). Mitochondria in the Mit-CY@Nps-treated group displayed preserved mitochondrial morphology with intact membranes and well-organized cristae, further demonstrating the protective effect of Mit-CY@Nps on mitochondrial integrity.

To assess apoptosis, Annexin V/PI dual staining via flow cytometry was performed. The OGD/R and PBS group displayed an elevated apoptotic rate, whereas Mit-CY@Nps treatment reduced apoptosis to levels comparable to the normoxic control group, underscoring its neuroprotective efficacy (Figures 3M and 3N). Mit-CY@Nps exerts comprehensive neuroprotection against I/R injury via ferroptosis inhibition and mitochondrial restoration. These effects disrupted the pathological feedback between oxidative stress and iron-dependent lipid damage, preventing ferroptotic and apoptotic cell death. In addition, restoration of mitochondrial homeostasis by Mit-CY@Nps promoted microglial polarization toward the anti-inflammatory M₂ phenotype, creating a more favorable post-ischemic microenvironment.

Molecular mechanisms of Mit-CY@Nps in modulating the mitochondrial ferroptosis-energetic homeostasis cascade

To further investigate the therapeutic effects of Mit-CY@Nps on OGD/R-modulated mitochondrial ferroptosis-energetic homeostasis cascade, RNA sequencing was performed on PC12 cells treated under three conditions: Sham, OGD/R, and Mit-CY@Nps. RNA sequencing revealed distinct gene expression profiles across the experimental groups (Figure S7A). The Venn/UpSetR analysis identified 689 differentially expressed genes (DEGs) between the Sham and OGD/R groups, 495 DEGs between the Mit-CY@Nps and OGD/R groups, and only 30 DEGs between the Mit-CY@Nps and Sham groups, indicating that Mit-CY@Nps treatment significantly improved the expression profile of DEGs associated with OGD/R-induced cellular injury (Figure 4A). Six shared genes were identified as potential key therapeutic targets. According to the volcano plot, 427 genes were upregulated and 262 genes were downregulated in the Mit-CY@Nps group compared to the OGD/R group (Figure S7B). Construction of the PPI (protein-protein interaction) network between the Mit-CY@Nps and OGD/R groups demonstrated the interactions among these key genes. GPX4 and ACSL4, which are critical regulators of ferroptosis, were central in the network. Key genes related to inflammation (e.g., TNF and IL-6), oxidative stress (e.g., HIF1A), and mitochondrial function were implicated, suggesting that Mit-CY@Nps modulates OGD/R-induced ferroptosis through ACSL4 and GPX4 (Figure 4B). Western blot analysis was performed to assess the expression of ACSL4 and GPX4 in each group (Figures 4C, S8A, and S8B). OGD/R treatment reduced GPX4 expression and increased ACSL4 levels, consistent with ferroptosis activation. In contrast, Mit-CY@Nps treatment reversed these changes by suppressing ACSL4 and restoring GPX4 expression, indicating effective inhibition of lipid peroxidation and ferroptosis.

Figure 4.

Molecular mechanisms of Mit-CY@Nps in modulating the mitochondrial ferroptosis-energetic homeostasis cascade

(A) Venn plots of differential gene distribution among Sham, OGD/R, and Mit-CY@Nps groups.

(B) PPI network of DEGs.

(C) Western blot analysis of ferroptosis-related proteins Acsl4 and GPX4, with GAPDH as loading control.

(D) GO enrichment of biological processes associated with DEGs between OGD/R and Mit-CY@Nps groups.

(E) Core pathway network analysis of key genes in the Mit-CY@Nps group.

(F) Heatmap of selected key gene expression across groups.

(G–I) GSEA map of three significant pathways between the OGD/R- and Mit-CY@Nps-treated groups. All the p values were adjusted according to Benjamini-Hochberg multiple testing adjustment. p values <0.05 are considered statistically significant.

(J–M) ELISA quantification of TNF-α, IL-10, IL-6, and caspase-1 in ischemic brain tissues across treatment groups (n = 3). The numerical data in (J, K, L, and M) are presented as the mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗∗p < 0.0001.

A one-way ANOVA with Tukey’s multiple comparisons test (J, K, L, and M) was used for statistical significance analysis. See also Figures S7 and S8.

GO enrichment analysis revealed that multiple important biological processes related to lipid metabolism, mitochondrial function, and immune regulation are significantly enriched in each group of cells (Figure 4D). Pathways involved in iron ion transport, lipid metabolism, mitochondrial ATP synthesis, and LPO were enriched, underscoring their central roles in cellular energy homeostasis and antioxidant defense. These findings indicate that OGD/R induces profound metabolic reprogramming and mitochondrial dysfunction, which are corrected following Mit-CY@Nps treatment. Integrated pathway analysis further identified three principal mechanisms modulated by Mit-CY@Nps: mitochondrial dysfunction, arachidonic acid metabolism, and ferritin-mediated iron storage. Key regulatory targets such as ACSL4, ALOX, CYP, GPX4, and SOD were highlighted for their roles in modulating lipid oxidative stress and ferroptosis (Figure 4E). Heatmap analysis of representative gene expression profiles revealed upregulation of oxidative stress and pro-inflammatory markers (e.g., Acsl4, Slc7a11, Tnf, Ptgs2, S100a9, Myd88, and Mapk14) in the OGD/R group, whereas antioxidant and iron-regulating genes (GPX4, Fth1, Socs3, and Nfe2l2) were significantly elevated following Mit-CY@Nps treatment, suggesting enhanced redox balance and ferroptosis suppression (Figure 4F). KEGG-GSEA analysis was employed to examine changes in the TNF signaling pathway, apoptosis, and necroptosis between the OGD/R and Mit-CY@Nps groups. The TNF signaling pathway (Figure 4G), a pivotal mediator of inflammation,⁵² along with apoptosis (Figure 4H) and necroptosis pathways (Figure 4I), were activated under OGD/R conditions but attenuated upon Mit-CY@Nps intervention. Collectively, integrated transcriptomic profiling, protein-level quantification, and pathway enrichment analyses revealed that Mit-CY@Nps orchestrates the regulation of LPO, iron homeostasis, and mitochondrial function. To further elucidate the therapeutic mechanism of Mit-CY@Nps, enzyme-linked immunosorbent assay (ELISA) was conducted to quantify key inflammatory cytokines such as TNF-α, IL-6, caspase-1, and IL-10 in brain tissue from the tMCAO model. Compared to the sham-operated mice, tMCAO elevated pro-inflammatory cytokines (TNF-α, IL-6, and caspase-1). However, Mit-Nps, CY, and Mit-CY@Nps treatment reduced their levels while increasing the anti-inflammatory cytokine IL-10 (Figures 4J–4M). Collectively, integrated transcriptomic profiling, protein-level quantification, and pathway enrichment analyses revealed that Mit-CY@Nps orchestrates the regulation of LPO, iron homeostasis, and mitochondrial function.

Spatiotemporal tracking and tissue distribution of the Mit-CY@Nps

To validate BBB penetration and MRI-related capabilities in vivo, a tMCAO model was established to mimic I/R injury in the right hemisphere. The brain-targeting efficiency of the nanoactuator was assessed by comparing the biodistribution of IR780-labeled Mit-CY@NPs (IR780@Nps) and free IR780. The fluorescence signal of IR780@Nps in the ischemic brain region progressively increased over time, peaking at 6 h. Free IR780 exhibited relatively weak fluorescence signals with a non-specific distribution throughout the body. After 8 h, the fluorescence intensity of IR780@Nps began to decline, indicative of metabolism and gradual excretion (Figure 5A). Quantitative analysis showed that IR780@Nps exhibited a rapid increase in signal intensity from 2 h, peaked at 6 h, and then gradually declined, whereas free IR780 remained consistently low (Figure 5B). To further elucidate the biodistribution and clearance pathways of the Mit-CY@Nps, ex vivo fluorescence imaging of major organs was performed on tMCAO mice at 4 and 12 h post-injection. Strong fluorescence signals in the brain were observed in the brain for IR780@Nps at both time points, confirming efficient brain accumulation. Over time, fluorescence signals accumulated in the liver and kidneys, indicative of metabolic and excretory clearance. By 24 h, the fluorescence signal in the brain nearly disappeared, while the liver and kidneys exhibited high signal intensity (Figures 5C and 5D). To further distinguish physiological from pathological BBB permeability, control experiments were performed in healthy C57BL/6J mice. As shown in Figure S9A, in vivo IVIS imaging revealed weak fluorescence signals in the brains of healthy mice, indicating limited BBB permeability under normal physiological conditions. Quantitative analysis (Figure S9B) demonstrated that the total fluorescence intensity in the brains of tMCAO mice was higher than that in healthy controls, confirming that BBB disruption after ischemia-reperfusion enhanced nanoparticle penetration into the brain. Ex vivo imaging at 24 h (Figure S9C) further showed that fluorescence in healthy mice was distributed in the liver, spleen, and kidneys, with only negligible signals detected in the brain. The enhanced brain localization observed in the tMCAO model therefore originates from ischemia-induced BBB disruption rather than nonspecific leakage. These findings elucidate the in vivo biodistribution and clearance mechanisms of Mit-CY@Nps, confirming their brain-targeting efficiency and favorable biosafety profile.

Figure 5.

In vivo imaging of Mit-CY@Nps in tMCAO models

(A) Fluorescence imaging of IR780@Nps and free IR780 in tMCAO mice at different post-injection time points.

(B) Quantitative comparison of total radiation efficiency between IR780@Nps and free IR780 (n = 3).

(C) Ex vivo fluorescence images of major organs from IR780@Nps-treated tMCAO mice at 0, 4, 12, and 24 h.

(D) Quantified radiation efficiency of IR780@Nps in brain, liver, and kidney (n = 3).

(E) T₁-weighted MRI imaging showing signal enhancement in the ischemic brain region of tMCAO mice injected with Mit-CY@Nps or PBS at 0, 3, 6, and 12 h. The numerical data in (B and D) are presented as the mean ± SD. ∗∗p < 0.01.

A one-way ANOVA with Tukey’s multiple comparisons test (B and D) was used for statistical significance analysis. See also Figures S9 and S10.

Mit-CY@Nps, based on a Gd³⁺ metal coordination polymer, exhibit T₁-weighted MRI contrast due to the paramagnetic properties of Gd³⁺, which shorten T₁ relaxation times and enhance signal intensity.⁵³ To assess this functionality in vivo, tMCAO mice were injected with either PBS or Mit-CY@Nps, followed by T₁-weighted MRI at 0, 3, 6, and 12 h post-injection. Robust signal enhancement was observed in the ischemic brain region at 3 and 6 h post-injection of Mit-CY@Nps (Figure 5E), indicating efficient BBB penetration and targeted accumulation in infarcted tissue. These results highlighted the dual functionality of Mit-CY@Nps as both a therapeutic and diagnostic platform.

Therapeutic efficacy of Mit-CY@Nps in alleviating ischemic brain injury and neurological deficits

To determine the optimal therapeutic dosage of Mit-CY@Nps, a dose-response study was first conducted in tMCAO mice at gradient doses of 2.5, 5, 10, and 15 mg/kg. The neuroprotective efficacy was evaluated by 2,3,5-triphenyltetrazolium chloride (TTC) staining and quantitative analysis of infarct volume. As shown in Figures S9D and S9E, the infarct volume decreased as the Mit-CY@Nps dose increased from 2.5 to 10 mg/kg. Further increasing the dose to 15 mg/kg did not produce additional improvement, suggesting that the therapeutic efficacy reached a plateau at approximately 10 mg/kg. Following reperfusion, mice were administered PBS, Mit-Nps, CY, or Mit-CY@Nps (10 mg/kg). The doses of Mit-Nps and CY were designed to be equivalent to Mit-CY@Nps in terms of component content. After 24 h of treatment, brain tissues were collected for multimodal analysis (Figure 6A). TTC staining and brain water content measurements revealed that the PBS group exhibited a large infarct area (45.2%), while the Mit-Nps and CY groups reduced infarct volumes to 32.8% and 28.5%. Mit-CY@Nps treatment further decreased infarct volume to 17.6%, indicating a pronounced synergistic neuroprotective effect (Figures 6B and 6C). The brain water content decreased from 82.1% in the PBS group to 74.3% in the Mit-CY@Nps group (Figure 6D), consistent with significant alleviation of cerebral edema. These results were further corroborated by T₂-weighted MRI, which confirmed that Mit-CY@Nps achieved the most prominent structural protection against ischemic injury (Figure 6E). We further evaluated cerebral perfusion recovery using laser doppler flowmetry. The PBS group exhibited pronounced perfusion obstruction, while Mit-NPs showed limited improvement. In contrast, CY improved local vascular function through its antioxidant effects, enhancing blood flow in certain regions. Comparatively, the Mit-CY@NPs group improved cerebral perfusion across a broader region, reflecting its dual capacity to alleviate microvascular dysfunction driven by mitochondrial impairment and oxidative stress. (Figure 6F). To further explore the vascular architecture and integrity, three-dimensional time-of-flight (3D-TOF) MRI was utilized to visualize cerebrovascular networks. The PBS group displayed fragmented and poorly defined vasculature, whereas Mit-CY@NPs treatment restored a more continuous and intact vascular network, comparable to the Sham group (Figure 6G). This demonstrates the role of Mit-CY@NPs in promoting post-ischemic vascular repair.

Figure 6.

Improved brain damage and neurological dysfunction after ischemic brain injury by Mit-CY@Nps

(A) Schematic of the tMCAO experimental design.

(B) TTC-stained brain sections showing infarct regions across groups.

(C) Quantification of infarct volume relative to total brain volume (n = 3).

(D) Brain water content analysis to assess cerebral edema (n = 3).

(E) Representative T₂-weighted MRI images highlighting infarcted areas (dotted red lines) in Sham, PBS, Mit-Nps, CY, and Mit-CY@Nps groups.

(F) Heatmap of cerebral blood flow changes each group by laser speckle imaging.

(G) Cerebrovascular structural integrity by 3D-TOF illustrating cerebrovascular architecture and integrity.

(H) NeuN/DAPI staining to evaluate neuronal survival. Scale bars: 2 mm.

(I) Evans blue staining to assess BBB permeability.

(J and K) Double immunofluorescence of activated microglia (Iba1) with CD86 (M₁) and CD206 (M₂). Scale bars: 2 mm (left) and 200 μm (right).

(L and M) Immunofluorescence of GPX4 and 4-HNE indicating ferroptosis levels. Scale bars: 100 μm.

(N) TUNEL staining showing apoptotic cells in the ischemic penumbra. Scale bars: 200 μm. The numerical data in (C and D) are presented as the mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

A one-way ANOVA with Tukey’s multiple comparisons test (C and D) was used for statistical significance analysis. See also Figures S9–S12.

Neuronal survival was evaluated through immunofluorescence staining using the neuron-specific marker NeuN. The Mit-CY@Nps-treated group exhibited a higher density of NeuN-positive neurons compared to the PBS group, approaching levels observed in the Sham group (Figure 6H). This suggests that Mit-CY@Nps conferred neuroprotection by attenuating neuronal apoptosis and preserving structural integrity, thereby contributing to the maintenance of neurological function. We further assessed the integrity of the BBB using Evans blue extravasation.⁵⁴ Extensive dye leakage was observed in the PBS group, indicating severe BBB disruption. In contrast, Mit-CY@Nps treatment resulted in minimal Evans blue accumulation, demonstrating its capacity to preserve BBB integrity and restrict parenchymal exposure to circulating neurotoxic substances (Figure 6I). Quantitative analysis validated this protective effect, confirming that Mit-CY@Nps mitigates BBB permeability, consistent with its overall neurovascular protective profile in ischemic brain injury (Figure S9F).

To elucidate the inflammatory regulatory mechanisms, immunostaining of microglial phenotypes was performed using Iba1 with CD86 (M₁) or CD206 (M₂) markers. As shown in Figures 6J and 6K, PBS-treated mice exhibited extensive activation of CD86⁺ pro-inflammatory microglia, while Mit-CY@Nps treatment markedly decreased CD86⁺ cells and increased CD206+ anti-inflammatory microglia, indicating a phenotypic shift toward an inflammation-resolving M₂ state. Quantitative analysis (Figures S9G and S9H) revealed that the proportion of CD86+/Iba1+ cells reached approximately 38% in the PBS group, which was reduced to 25% following Mit-CY@Nps treatment. In contrast, the fraction of CD206+/Iba1+ cells increased from 10% in the PBS group to 40% in the Mit-CY@Nps group. These results confirm that Mit-CY@Nps reprogram microglial polarization in vivo, suppressing the pro-inflammatory M₁ response while promoting an M₂-dominant, inflammation-resolving phenotype, consistent with the flow cytometry results observed in vitro. Ferroptosis-related markers were analyzed to explore the underlying neuroprotective mechanisms. GPX4 immunofluorescence intensity was restored in the Mit-CY@Nps group compared with PBS-treated mice (Figures 6L and S10A), whereas lipid peroxidation marker 4-HNE was reduced (Figures 6M and S10B). TUNEL staining further demonstrated that Mit-CY@Nps reduced neuronal apoptosis (Figures 6N and S10C). Taken together, Mit-CY@Nps markedly improved both structural and functional recovery of the brain following ischemia-reperfusion injury through multiple synergistic mechanisms.

Furthermore, we explored whether this therapeutic strategy could modulate additional forms of programmed cell death beyond ferroptosis and apoptosis. Immunofluorescence staining (Figures S10D and S10E) revealed that, in the PBS-treated group, the expression levels of gasdermin D (GSDMD, a pyroptosis marker) and phosphorylated mixed lineage kinase domain-like protein (p-MLKL, a key effector of necroptosis) were upregulated following I/R injury. In contrast, Mit-CY@Nps treatment reduced the expression of both proteins, exhibiting superior inhibitory effects compared with the single-component and vehicle groups (Figures S10F and S10G). These findings indicate that Mit-CY@Nps not only suppress ferroptosis and apoptosis but also modulate multiple programmed cell death pathways and inflammatory processes, thereby providing broad-spectrum and integrated neuroprotection against cerebral I/R injury.

To obtain histopathological alterations after ischemic injury, hematoxylin-eosin (H&E) and Nissl staining were performed. H&E staining revealed cortical damage in the PBS group, characterized by disorganized cell architecture, indistinct neuronal morphology, and nuclear pyknosis (Figure S10H). Mit-NPs and CY treatment groups alleviated these structural impairments, showing preserved cell contours and intact cell bodies. In contrast, the Mit-CY@NPs group exhibited the intact tissue structure, with arranged cells and no observable pathological changes. Nissl staining was used to assess neuronal functional status (Figure S10I). Both Mit-NPs and CY provided moderate recovery, while Mit-CY@Nps restored Nissl body density and staining intensity, reflecting superior preservation of neuronal function. Furthermore, a 28-day safety assessment was conducted in healthy C57BL/6J mice, in which Mit-CY@Nps were administered at a dose of 10 mg/kg. Body weight monitoring revealed no significant difference between the PBS and Mit-CY@Nps groups throughout the 28-day observation period (Figures S11A and S11B). Serum biochemical analysis showed no changes in hepatic function indicators, including ALT, AST, ALP, LDH, ALB, and TP, or in renal function parameters such as UREA and CREA, between the two groups (Figures S11C–S11J). Histopathological examination of major organs, including the heart, liver, spleen, lung, kidney, and brain, also showed normal tissue architecture without signs of inflammation, necrosis, or other pathological alterations (Figure S11K). These results demonstrate good systemic biocompatibility of Mit-CY@Nps without evident hepatic or renal toxicity during long-term administration.

Therapeutic neurological outcomes and behavioral recovery after cerebral I/R injury by the Mit-CY@Nps

To assess the neuroprotective effects and functional recovery potential of Mit-CY@Nps, a series of behavioral evaluations were conducted in both normal and tMCAO mice (Figure 7A). To evaluate general physiological recovery, body weight trajectories were first analyzed. All groups exhibited postoperative weight loss following tMCAO surgery. However, while the PBS group displayed sustained weight decline and delayed recovery, mice treated with Mit-NPs or CY began to regain weight between days 3 and 7. The Mit-CY@NPs group demonstrated the most significant weight recovery, approaching Sham levels by day 7, indicating its efficacy in overall health improvement (Figure 7B). We further validated its resistance to lethality through survival rate analysis. The PBS group exhibited a marked decline in survival, falling to 50% by day 7. Treatment with Mit-NPs and CY improved survival rates to 60% and 65%, respectively, while Mit-CY@Nps enhanced survival to approximately 80%, highlighting its capacity to reduce mortality and improve post-stroke viability (Figure 7C). To assess neurological function recovery, the Longa scoring system was employed. The PBS group displayed elevated scores, indicating severe neurological impairment. Both the Mit-NPs and CY groups showed moderate improvements, while the Mit-CY@NPs group registered the lowest scores with the mildest neurological deficits, with statistically significant differences (Figure 7D).

Figure 7.

Promoting effects of Mit-CY@Nps on neural function and behavioral recovery in tMCAO mice

(A) Schematic of the tMCAO treatment timeline and behavioral testing schedule.

(B) Changes in body weight over 7 days across groups (n = 8).

(C) Survival curves across treatment groups (n = 8).

(D) Longa score after 7 days of different treatments (n = 8).

(E) Schematic diagram of the open field test setup.

(F) Representative locomotor trajectories on days 1, 3, 5, and 7.

(G) Quantification of total distance and average velocity (n = 3).

(H) Schematic of cylinder test setup.

(I) Forelimb use asymmetry across days 1, 3, 5, and 7 (n = 3).

(J) Schematic of adhesive test.

(K) Time to initial contact and removal in adhesive trials across groups (n = 3).

The numerical data in (B, D, G, I, and K) are presented as the mean ± SD. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

A one-way ANOVA with Tukey’s multiple comparisons test (D, G, I, and K) was used for statistical significance analysis. See also Figures S11 and S12.

To further validate its effects on motor function and neurobehavioral recovery, a series of behavioral assays was conducted, including the open field test, cylinder test, and adhesive removal test. In the open field test, PBS-treated mice exhibited restricted movement trajectories, with reduced travel distance and speed. Both Mit-NPs and CY treatments restored motor function by day 3, whereas the Mit-CY@NPs group showed continuous improvement in activity levels. By day 7, movement trajectories, travel distance, and speed in the Mit-CY@NPs group recovered compare to Sham group, demonstrating its capacity for restoring spontaneous motor behavior (Figures 7E–7G). After preliminary validation of motor behavior recovery, we employed the cylinder test to assess forelimb use asymmetry (Figures 7H and 7I). The PBS group exhibited significant lateralized behavior (asymmetry rate >70%), indicating that brain injury affected forelimb functional coordination. The Mit-NPs and CY groups alleviated lateralized behavior during the mid-postoperative period, whereas the Mit-CY@NPs group significantly reduced the asymmetry rate, demonstrating enhanced restoration of postural symmetry and forelimb coordination. To assess sensorimotor integration, the adhesive removal test was conducted, measuring both stimulus recognition (contact time) and motor response (removal time) (Figures 7J and 7K). The PBS group showed delayed responses, indicative of persistent deficits in cortical processing and motor execution. In contrast, the Mit-CY@NPs group showed rapid recovery in both metrics, restoring normal reaction speeds by days 5–7, suggesting its comprehensive promotion of cortical integration and sensorimotor recovery.

After confirming its short-term neurobehavioral recovery effects, we employed 3.0T T₂-weighted MRI on the 7^th postoperative day to reassess the brain tissue recovery status. Compared to 24 h post-injection, all treatment groups exhibited varying degrees of infarct resolution and edema attenuation. Notably, the Mit-CY@NPs treatment group demonstrated the most significant reduction in infarct volume and brain edema (Figure S12). These results highlight the significant role of Mit-CY@Nps in promoting the recovery of sensory and motor coordination functions. From systemic indicators such as body weight and survival rate to improvements in neurological deficit scores, locomotor function, and sensorimotor integration, Mit-CY@Nps outperformed monotherapies. These multifaceted protective effects likely arise from the Mit-CY@Nps’s ability to integrate neuroprotection, enhancement of neuronal survival, suppression of inflammation and ferroptosis, and restoration of mitochondrial function.

Discussion

This study established a mitochondria-targeted nanotherapeutic approach for the treatment of cerebral ischemia-reperfusion injury through the rational design of Mit-CY@Nps, which modulated the mitochondrial ferroptosis-bioenergetic homeostasis cascade. By integrating redox-triggered programable collapse, the nanoactuator enabled synchronized suppression of ferroptosis, attenuation of oxidative stress, and restoration of mitochondrial bioenergetic homeostasis. In a tMCAO-induced I/R mouse model, intravenously administered Mit-CY@Nps crossed the compromised BBB and accumulated in ischemic neuronal tissue. This intervention reduced pro-inflammatory cytokines, elevated the anti-inflammatory cytokine IL-10, and mitigated neuroinflammatory responses. Concurrently, preservation of mitochondrial function and inhibition of ferroptotic signaling promoted neuronal survival and functional recovery. The nanoactuator also reduced multiple forms of regulated cell death, including apoptosis, pyroptosis, and necroptosis, indicating that our strategy provides broad neuroprotection by attenuating primary insults such as oxidative stress and mitochondrial dysfunction. These findings highlighted the therapeutic potential of the engineered Mit-CY@Nps for the coordinated regulation of cell death pathways and metabolic dysfunction in ischemic stroke, offering a promising foundation for precision nanomedicine in neurovascular disease. Beyond stroke therapy, Mit-CY@Nps establishes a general and accessible platform for the targeted inhibition of regulated cell death pathways and the enhancement of bioactive molecule performance, with translational relevance across neurodegenerative and oxidative-stress-related disorders.

Limitations of the study

This study highlights the neuroprotective potential of Mit-CY@Nps in a tMCAO mouse model, though several limitations remain. The biodistribution, long-term biosafety, and immunogenicity in larger animal models or primates require further evaluation for clinical translation. Additionally, the molecular dynamics of the “hierarchical collapse” process in the in vivo environment need further clarification using advanced imaging techniques. Finally, interspecies differences in cerebrovascular anatomy, immune response, and drug metabolism may impact translational potential, emphasizing the need for validation in more clinically relevant models. Addressing these limitations is essential for advancing this therapeutic strategy.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Yu Chen (chenyuedu@shu.edu.cn).

Materials availability

This study did not generate new unique materials.

Data and code availability

• •The RNA-seq data reported in this paper have been deposited in the National Center for Biotechnology Information (NCBI) Short Read Archive under BioProject accession numbers PRJNA1393401. Accession numbers are listed in the key resources table.
• •This study did not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

This work is financially supported by the National Natural Science Foundation of China (82372029, 22205133), Discipline Construction of Pudong New Area Health Commission (PWZxk2022-03), Shanghai Pudong New District Health Committee Health Industry Special Project (PW2024E-02), and The Investigator-initiated Trial Program of Shanghai Pudong New Area Health Commission (the Cohort Study Program), 2025-PWDL-24.

Author contributions

Y.X., M.Y., and Y.C. conceived the idea for this project. G.S., Y.W., and Y.D. performed the experiments, analyzed the data, and produced the manuscript under the supervision of Y.X., M.Y., and Y.C. B.H. conceived the idea and manuscript development. Y.S., J.D., J.C., L.L, and X.C. offered advice on material design, self-assemble experiments, and data analysis. W.L. and K.C. contributed to the data analysis and manuscript development. All authors reviewed, edited, and approved the final version of the manuscript.

Declaration of interests

The authors declare no competing interests.

Declaration of generative ai and ai-assisted technologies in the writing process

During the preparation of this work, the authors utilized AI to enhance the language. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
FITC anti-mouse CD86 antibody	Biolegend	Cat#: 159219; RRID: AB_3106043
APC anti-mouse CD206 (MMR) antibody	Biolegend	Cat#: 141707; RRID: AB_10896057
ACSL4 antibody	Signalway	Cat#: 36176; RRID: AB_3717460
GPX4 antibody	Abcam	Cat#: ab231174; RRID: AB_3073732
GAPDH antibody	Abcam	Cat#: ab59164; RRID: AB_3676490
NeuN antibody	Abcam	Cat#: ab209898; RRID: AB_3695640
Iba1 antibody	Abcam	Cat#: ab220815; RRID: AB_3697452
4-HNE antibody	Abcam	Cat#: ab46545; RRID: AB_722490
p-MLKL antibody	Abcam	Cat#: CY5146; RRID: AB_3674778
GSDMD antibody	Abcam	Cat#: ab255603; RRID: AB_3718638
CD86 antibody	Abmart	Cat#: PU885801; RRID: AB_3717773
CD206 antibody	Abmart	Cat#: TU313804; RRID: AB_3712524
Chemicals, peptides, and recombinant proteins
Gadolinium (III) chloride (GdCl3)	Macklin	Cat#: G689764
Adenosine 5′-triphosphate disodium salt (ATP)	Macklin	Cat#: A800085
DSPE-PEG-NH2	Macklin	Cat#: N796614
Safflower yellow	MedChemExpress	Cat#: HY-N0938
Elamipretide (SS31)	Macklin	Cat#: E675352
Sodium borohydride	Sigma-Aldrich	Cat#: 452882
Selenium	Sigma-Aldrich	Cat#: 229865
3-Brompropanol	Sigma-Aldrich	Cat#: 167169
N-Hydroxysuccinimide	Sigma-Aldrich	Cat#: 130672
Mitotracker deep red FM	Beyotime	Cat#: C1032-50μg
Antifade mounting medium with DAPI	Beyotime	Cat#: P0131-25ml
Streptavidin-FITC (1mg/ml)	Beyotime	Cat#: A0316-5ml
IR-780	Beyotime	Cat#: HY-D1063
Lipopolysaccharide (LPS)	Beyotime	Cat#: S1732-5mg
Recombinant mouse IL-4	BioLegend	Cat#: 574302
Erastin	MedChemExpress	Cat#: HY-15763
TTC	Solarbio	Cat#: T8170
Evans blue	Sigma-Aldrich	Cat#: E2129
Critical commercial assays
Reactive oxygen species assay kit	Beyotime	Cat#: S0033S
Mitochondrial superoxide assay kit (MitoSOX™ Red)	Beyotime	Cat#: S0061S
JC-1 mitochondrial membrane potential kit	Beyotime	Cat#: C2006
Cell counting kit-8 (CCK-8)	Beyotime	Cat#: C0038
MDA assay kit	Beyotime	Cat#: S0131S
Total glutathione assay kit	Beyotime	Cat#: S0052
Enhanced ATP assay kit	Beyotime	Cat#: S0027
Annexin V-FITC/PI apoptosis kit	MedChemExpress	Cat#: HY-K1073
OCR fluorometric assay kit	Elabscience	Cat#: E-BC-F070
Lipid peroxidation assay kit with BDPY 581/591 C11	Beyotime	Cat#: S0043S
Mitochondrial respiratory complex I activity assay kit	Boxbio	Cat#: AKOP005U-1
Mitochondrial respiratory complex III activity assay kit	Boxbio	Cat#: AKOP007C-1
Mouse TNF-α ELISA kit	Beyotime	Cat#: PT512
Mouse IL-10 ELISA kit	Beyotime	Cat#: PI522
Mouse IL-6 ELISA kit	Beyotime	Cat#: PI326
Mouse Caspase-1 ELISA kit	Elabscience	Cat#: E-EL-M0201
Deposited data
mRNA sequencing	This paper	[NCBI]: [PRJNA1393401]
Experimental models: Cell lines
PC12 rat pheochromocytoma cells	Editgene	Cat#: EDJ-WQ0795
HT22 hippocampal neuronal cells	Editgene	Cat#: EDJ-WQ0809
BV2 mouse microglial cells	Editgene	Cat#: EDC00359
bEnd.3 mouse brain microvascular endothelial cells	Editgene	Cat#: EDJ-WQ0839
Experimental models: Organisms/strains
Mouse: C57BL/6, male	GemPharmatech	N/A
Software and algorithms
GraphPad Prism (version 9.5.1)	GraphPad Software	https://www.graphpad.com/
Adobe Illustrator	Adobe	https://www.adobe.com/products/illustrator/free-trial-download.html
FlowJo (version 10.8.1)	BD Biosciences	https://www.flowjo.com/
ImageJ	ImageJ Software Inc.	https://imagej.nih.gov/ij/index.html
FV31S software (FV31S-SW and FV31S-DT)	Olympus	https://www.olympus-lifescience.com/

Experimental model and study participant details

Cell line and cell culture

PC12 rat pheochromocytoma cells, HT22 hippocampal neuronal cells, BV2 mouse microglial cells and bEnd.3 Mouse brain microvascular endothelial cells were purchased from the Editgene (Guangzhou, China). Both PC12 and bEnd.3 cells were maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS), 100 mg/mL streptomycin, and 100 U/mL penicillin. HT22 cells and BV2 cells were cultured in high glucose DMEM medium supplemented with 10% FBS, 100 mg/mL streptomycin, and 100 U/mL penicillin. The cells were cultured at 37°C in a humidified atmosphere containing 5% CO₂. Cell line identity was validated using STR analysis. All cell lines were regularly tested negative for mycoplasma by PCR.

Animals

Male C57BL/6J mice, aged 6–8 weeks, were obtained from GemPharmatech Co., Ltd (Jiangsu China). All mice were housed under specific pathogen-free (SPF) conditions and provided with standard laboratory chow. All experimental procedures were conducted in accordance with the regulations of the Ministry of Health and were approved by the Animal Ethics Committee of Shanghai University (Approval ID: YS 2023-025).

Method details

Synthesis of 3′-diselenodipropionic acid (DSeDPA)

Under a nitrogen atmosphere, sodium borohydride (2.8 g, 74 mmol) and selenium (2.8 g, 35.45 mmol) were placed in a chilled three-necked flask equipped with a condenser, gas inlet adapter, and dropping funnel. Water (150 mL) was added while stirring. During the initial vigorous reaction (approximately 10 min), the ice bath was removed, and a second portion of selenium (2.8 g, 35.45 mmol) was added to the colorless mixture. The mixture was then briefly heated using a heat gun to dissolve the selenium (approximately 3 h). After cooling to room temperature, 3-bromopropanol (10.0 g, 72 mmol) was added to 100 mL of water. The resulting yellow solution was stirred for 18 h. The aqueous layer was extracted several times with ethyl acetate, and the combined organic extracts were dried over anhydrous sodium sulfate (Na₂SO₄). The solvent was concentrated under reduced pressure, yielding 5.19 g of the product (51%). ¹H NMR (600 MHz, Chloroform-d) δ 3.61 ppm (t, 4H), δ 2.86 ppm (t, 2H).

Synthesis of 3′-disselenodipropionic acid (orthosuccinimidyl ester) (DSeDPA-NHS)

DSeDPA powder (1.1 g, 3.6 mmol) was dissolved in 15 mL of anhydrous DCM in a round-bottom flask under a nitrogen atmosphere, followed by the addition of NHS (1 g, 8.64 mmol) with continued stirring. After 30 min, a solution of 1.65 g (8.64 mmol) EDC in 15 mL of anhydrous DCM was added to the mixture at 0°C, and the mixture was stirred overnight at room temperature. The reaction mixture was then filtered to remove impurities, and the DCM was evaporated using a rotary evaporator. After placing the product in a vacuum oven for 24 h, the pure DSeDPA-NHS powder was collected. ¹H NMR (600MHz, Chloroform-d) δ 3.09 ppm (q, 2H), δ 3.03 ppm (m, 2H), δ 2.89 ppm (s, 4H).

Synthesis of Gd-ATP coodination network (Gd-ATP NMs)

Gd-ATP NMs were synthesized based on a previously reported protocol with slight modifications.^55,56,57 Briefly, an aqueous solution of ATP disodium salt (10 mM, 10 mL) was slowly added to an aqueous solution of GdCl₃ (5 mM, 10 mL) under gentle stirring at room temperature. Subsequently, 4 mg of DSPE-PEG-NH₂ was introduced to the mixture, which was then heated to 80°C and stirred continuously for 2 h to promote coordination and nanoparticle formation. After cooling to room temperature, the resulting Gd-ATP NMs were collected via centrifugation (15 000 g, 10 min), washed, and re-dispersed in ultrapure water for subsequent use.

Formation of selenium network on the Gd-ATP based nanoparticles (Mit-Nps)

Mit-Nps were fabricated by forming a diselenide-crosslinked polymeric shell on the surface of Gd-ATP scaffold. The detailed synthetic procedure for the diselenide crosslinker (DSeDPA-NHS) is provided in the Figure S1. Briefly, 10 mL of SS31 peptide solution (4 mM) was added to the preformed Gd-ATP Mps under vigorous stirring and incubated at room temperature for 24 h. Subsequently, 20 mg of DSeDPA-NHS was introduced, and the reaction continued for an additional 24 h. The resulting Mit-Nps were collected by centrifugation at 15 000 g for 10 min, washed thoroughly with ultrapure water three times, and re-dispersed for further use and characterization.

Synthetic hierarchically collapsible nanoactuator containing CY (Mit-CY@Nps)

Mit-CY@Nps were synthesized following a similar procedure to Mit-Nps, with the incorporation of CY after the formation of the Gd-ATP core to enable in situ drug loading. Briefly, an aqueous solution of carthamin yellow (5 mM) was added to solution of Gd-ATP scaffold (10 mM, 30 mL) under vigorous stirring. The mixture was stirred at room temperature for 24 h to allow encapsulation of CY within the porous coordination matrix (CY@Gd-ATP NMs). Subsequently, 10 mL of SS31 peptide solution (4 mM) was introduced and stirred for another 24 h. Finally, 20 mg of DSeDPA-NHS was added to the reaction mixture and stirred for an additional 24 h to complete the formation of the diselenide-crosslinked polymeric shell. The resulting Mit-CY@Nps were collected by centrifugation (15 000 g, 10 min), washed three times with ultrapure water, and re-dispersed for further use.

General protocol of drug-carrying encapsulation rate of the nanoparticles

UV-vis spectra were used to calculate the encapsulation rate of CY. Encapsulation rate was calculated by comparing the drug concentrations in the original sample with the supernatant. EN was calculated using the formula EN % = (1-C_f/C_t)×100%. C_f is the amount of free drug; C_t is the total amount of drug in nanoparticles. The encapsulation rate of CY was 37.5%.

Transmission electron microscopy (TEM)

TEM images were acquired using a JEM-1400 Flash HC microscope (JEOL, Japan) at an accelerating voltage of 10–120 kV. TEM samples were prepared by depositing 10 μL of a 2 mg/mL sample onto copper grids coated with ultrathin carbon support films.

Dynamic light scattering (DLS) and zeta potential

DLS measurements were performed at 25°C using a Malvern Panalytical Zetasizer (Malvern Panalytical, UK), with each measurement repeated three times and 15 readings recorded per run. The zeta potential of the samples were measured using a Malvern Panalytical Zetasizer (Malvern Panalytical Ltd., UK). The D_h value was calculated according to the Stokes-Einstein equation, assuming the particles to be spherical.

X-ray photoelectron spectroscopy (XPS)

XPS measurements were conducted on a Thermo Fisher Scientific-Alpha XPS system (Thermo Fisher Scientific, USA).

Fourier transform infrared (FT-IR) spectra

FT-IR spectra were recorded using a Nicolet iS10 FT-IR spectrometer (Thermo Fisher Scientific, USA).

UV-visible (UV-Vis) absorption spectra

UV-Vis absorption spectra were recorded on a UV-2450 spectrophotometer (Shimadzu, Japan). Fluorescence spectra were measured using an FS5 integrated fluorescence spectrometer (Techcomp, UK).

Thermal gravimetric analysis (TG)

TG using the German Synchronous Thermal Analyzer STA449, with temperature ranging from 0 to 800°C, 10°C/min at N₂.

High-performance liquid chromatography (HPLC) analysis

HPLC analysis utilized a 5 mmol/L ammonium acetate aqueous solution as the mobile phase, with the column temperature maintained at 10°C and a flow rate of 0.3 mL/min. Calibration was conducted using adenosine 5′-triphosphate disodium salt (ATP) as the standard, and the data were analyzed using Agilent SEC software.

The proton nuclear magnetic resonance (¹H NMR) spectra

¹H NMR spectra were recorded using a Bruker AVANCE NEO 600M NMR spectrometer (Bruker, Switzerland) with CDCl₃ as the solvent. Chemical shifts (δ) are expressed in parts per million (ppm) relative to tetramethylsilane (TMS) as the internal standard at δ = 0 ppm. Column chromatography purification was performed using a SepaBean T automatic flash chromatography system (Santai-tech, Changzhou, China) with irregular silica gel columns (40–63 μm, 60 Å, 12 g) and eluted with hexane and dichloromethane.

Adsorption test (BET)

BET obtained under nitrogen conditions, using the US Mac ASAP2460.

Confocal laser scanning microscopy (CLSM)

CLSM images were acquired using a Leica STELLARIS 8 DIVE multiphoton confocal microscope (Leica Microsystems, Germany).

Flow cytometry

Flow cytometric analysis was performed using the BD LSRFortessa X-20 (USA).

Cell uptake and subcellular localization

To enable fluorescence tracking, Mit-CY@Nps were labeled with FITC via amine conjugation. Specifically, FITC was added to the Mit-CY@Nps suspension and stirred overnight at room temperature to facilitate covalent attachment through reaction with surface amino groups. The reaction mixture was subsequently dialyzed against deionized water for 48 h, with frequent water changes to remove unbound FITC. The resulting FITC-labeled Mit-CY@Nps were collected and stored at 4°C for further use. For cellular uptake studies, PC12 cells were seeded into standard culture dishes and allowed to adhere for 24 h. FITC-labeled Mit-CY@Nps were then added to the culture medium and incubated with the cells for 2, 4, 6, 8, and 12 h. After incubation, cells were washed thoroughly with PBS to remove excess nanoparticles, and nuclei were counterstained with DAPI for 5 min. Fluorescence images were acquired using a CLSM, and intracellular fluorescence intensity was quantified using ImageJ software. To investigate mitochondrial localization of Mit-CY@Nps, PC12 cells were seeded in confocal imaging dishes and incubated with FITC-labeled Mit-CY@Nps for 6 h. Following incubation, unbound nanoparticles were removed by triple PBS washes. Mitochondria were then stained with MitoTracker according to the manufacturer’s instructions. Fluorescence images were acquired using a laser scanning confocal microscope to visualize nanoparticle co-localization with mitochondrial structures.

In vitro cytotoxicity assay

To evaluate the potential impact of hierarchical collapse nanoactuator on cellular viability, CCK-8 assays were performed using PC12 neuronal cells. Briefly, the cells were seeded into 96-well plates and allowed to adhere for 24 h. Following co-culturing with hierarchical collapse nanoactuator at varying concentrations for 24 h, cell viability was assessed using the CCK-8 reagent.

Establishment of OGD/R model

To establish the OGD/R model, cells were initially seeded in culture dishes and allowed to adhere for 24 h. The culture medium was then replaced with glucose-free MEM, and the cells were subjected to OGD in an anaerobic chamber (Mitsubishi, Japan) for 4 h. Following this, the medium was replenished with normal culture medium, and the cells were transferred to a 5% CO₂ incubator for continued culture for an additional 24 h.

BBB assessment in vitro

An in vitro BBB model was established using bEnd.3 brain microvascular endothelial cells. Briefly, bEnd.3 cells were seeded at a density of 3 × 105 cells per well in the upper chamber of 12-well Transwell inserts (polyester membrane, 0.4 μm pore size, LABSELECT) and cultured until the transendothelial electrical resistance exceeded 300 Ω·cm², indicating the formation of a tight endothelial barrier. In parallel, PC12 neuronal cells were seeded in the lower chambers at the same density (3 × 105 cells per well) and incubated for 24 h to allow adhesion and stabilization. Subsequently, the test formulations were added to the upper chamber, and co-incubation was continued for an additional 24 h to allow translocation across the endothelial layer. Cellular internalization in the lower compartment was then assessed via fluorescence microscopy to evaluate BBB permeability and neuronal uptake.

Measurement of intracellular ROS

Intracellular ROS levels were quantified using the ROS fluorescent probe DCFH-DA, which is oxidized by ROS to generate the fluorescent compound 2′,7′-dichlorofluorescein (DCF). PC12 cells subjected to OGD/R were incubated with PBS or various treatment formulations for 6 h. Subsequently, cells were washed and stained with DCFH-DA according to the manufacturer’s instructions, followed by incubation at 37°C for 20 min to allow intracellular oxidation. After three additional PBS washes to remove unreacted probe, fluorescence images were acquired using a fluorescence microscope. Quantitative analysis of DCF fluorescence intensity was performed using ImageJ software to assess intracellular ROS levels.

Intracellular mitochondrial superoxide (MitoSOX) assay

MitoSOX Red Mitochondrial Superoxide Indicator was used to determine MitoSOX levels in PC12 cells. PC12 cells subjected to OGD/R were incubated with PBS or various treatment formulations for 6 h. Subsequently, cells were washed and stained with MitoSOX according to the manufacturer’s instructions, followed by incubation at 37°C for 30 min to allow intracellular oxidation. After three additional PBS washes to remove unreacted probe, fluorescence images were acquired using a fluorescence microscope. Quantitative analysis of MitoSOX fluorescence intensity was performed using ImageJ software to assess Intracellular mitochondrial superoxide levels.

Lipid peroxidation (LPO) detection

Lipid peroxidation levels were evaluated using the BODIPY 581/591 C11 fluorescent probe in conjunction with confocal microscopy. PC12 and cells subjected to OGD/R, were incubated with PBS or the designated treatment formulations for 6 h. After treatment, cells were washed with PBS and incubated with a working solution of BODIPY 581/591 C11 at 37°C in the dark for 30 min. Excess probe was removed by washing three times with PBS. Fluorescence signals were visualized using a laser scanning confocal microscope, where red fluorescence represents non-oxidized lipids and green fluorescence indicates oxidized lipids, reflecting the extent of lipid peroxidation.

Erastin-induced ferroptosis model and detection (HT22 and PC12 cells)

The cells were divided into five groups: Sham, PBS, Mit-Nps, CY, and Mit-CY@Nps. Except for the Sham group, all other groups were exposed to erastin (10 μM) for 24 h to induce ferroptosis. Before erastin treatment, Mit-Nps (200 μg/mL), CY (2 μM), or Mit-CY@Nps (2 μM CY-equivalent) were added 1 h in advance and then co-incubated with erastin for the remaining period. For lipid peroxidation imaging, cells were washed with PBS and incubated with 2 μM BODIPY™ 581/591 C11 in serum-free DMEM at 37°C for 30 min in the dark. After washing twice with PBS, fluorescence images were captured using a confocal microscope (E_x/E_m: 488/510 nm for oxidized form, 561/590 nm for reduced form). For biochemical assays, cells were lysed in cold extraction buffer and centrifuged (12,000 g, 10 min, 4°C). Supernatants were collected for measurement of malondialdehyde (MDA) using a TBARS assay (absorbance = 532 nm) and glutathione (GSH) content using a DTNB-based kit (absorbance = 412 nm), both normalized to protein concentration determined by BCA assay and expressed as μmol/g protein.

Microglial cell polarization and ferroptosis model establishment

To induce M₁ polarization, BV2 murine microglial cells were treated with LPS (1 μg/mL) for 24 h, while M₂ polarization was induced by IL-4 (20 ng/mL) for 24 h. To ferroptosis rescue assay of M₂-type BV2 cells, M₂-type BV2 cells were incubated with free medium containing 10 μM erastin for 24 h to construct the ferroptosis model. Mit-Nps (200 μg/mL), CY (2 μM), or Mit-CY@Nps (2 μM CY-equivalent) were added 1 h prior to erastin treatment, and cell viability was detected by CCK8. Flow cytometry was used to analyze microglial polarization by staining cells with anti-CD86 (FITC) for the M₁ phenotype and anti-CD206 (APC) for the M₂ phenotype. After treatment, cells were harvested and stained, followed by flow cytometry analysis to determine the percentage of M₁ and M₂ polarized cells.

Detection of mitochondria membrane potential

The mitochondrial membrane potential was assessed using JC-1. PC12 cells subjected to OGD/R, were subsequently incubated with PBS or different material groups for 6 h. After washing, the medium was supplemented with an equal volume of JC-1 staining solution, thoroughly mixed, and incubated at 37°C for 20 min. After washing twice with PBS, the cells were analyzed using laser confocal microscopy.

ATP measurement

Intracellular ATP levels were quantified using a luciferase-based ATP assay kit. Cells were harvested, and the ATP content was extracted using the provided cell lysis buffer. After the addition of the luciferase reagent, light emission was measured using a luminometer. ATP concentrations were calculated using a standard curve and normalized to protein content.

Mitochondrial respiratory complex activity (I and III) assays

Mitochondrial respiratory chain complex activities were measured using commercially available assay kits according to the manufacturer’s instructions. After treatment, cells were harvested, and mitochondria were isolated using a mitochondrial isolation kit following the manufacturer’s protocol. The enzyme activities of complex I and complex III were assessed by measuring the reduction of NADH and cytochrome c, respectively, in the presence of specific substrates for each complex. Complex I Activity: Measured by monitoring the reduction of NADH (450 nm) in the presence of substrates like malate and glutamate. Complex III Activity: Measured by monitoring the reduction of cytochrome c (550 nm) with the addition of succinate as a substrate. Both activities were normalized to protein content measured by the BCA assay and expressed as U/mg protein.

OCR measurement by fluorometric assay

OCR was measured using the OCR Fluorometric Assay Kit. PC12 cells were seeded in a 96-well plate at 5 × 10⁴ cells/well and cultured overnight. After replacing the medium with pre-warmed working solution (Reagent 1, 10×dilution), cells were treated with Mit-Nps, CY, or Mit-CY@Nps for 1 h. After treatment, Reagent 2 was added, and the plate was incubated at 37°C. OCR was measured continuously every 2 min for 90 min using a fluorescence plate reader (Ex: 405 nm, Em: 650 nm). The OCR was calculated as the change in fluorescence over time and normalized to protein content.

Cell apoptosis

PC12 cells were subjected to OGD/R treatment, followed by continued culture with the addition of PBS or different material groups for 24 h. After the treatment, the cells were washed three times with cold PBS and collected. The cells were then stained with Annexin V and propidium iodide (PI) using the Annexin V-FITC Apoptosis Detection Kit and analyzed by flow cytometry. Quantitative experiments were performed in triplicate.

RNA sequencing

Total RNA was extracted from PC12 cells, which were subjected to OGD/R treatment, followed by continued culture with the addition of PBS or Mit-CY@Nps for 24 h. RNA sequencing was subsequently performed by Shanghai Wei Huan Biological Technology Co., Ltd. All data analyses were conducted using the Dr. Tom online platform.

Western blot analysis

Western blot analysis was performed to evaluate protein expression levels in vitro. PC12 cells were subjected to OGD/R treatment, followed by continued culture with the addition of PBS or Mit-CY@NPs for 24 h. After treatment, cells were harvested and lysed on ice for 15 min using RIPA buffer supplemented with the protease inhibitor PMSF. The lysates were centrifuged to isolate total protein, and protein concentrations were determined using a BCA assay kit. Equal amounts of protein were separated by SDS-PAGE and subsequently transferred onto PVDF membranes. The membranes were blocked with 5% skim milk for 1 h at room temperature, followed by overnight incubation at 4°C with primary antibodies targeting all-spectrin and GAPDH (serving as a loading control). After washing, the membranes were probed with appropriate horseradish peroxidase (HRP)-conjugated secondary antibodies for 1 h at room temperature. Protein bands were visualized using enhanced chemiluminescence (ECL) and quantified using a Bio-Rad ChemiDoc Imaging System.

Establishment of tMCAO mouse model

The tMCAO/R model was established in C57BL/6 mouse using the standard intraluminal filament occlusion technique. Briefly, mice were anesthetized via intraperitoneal injection of 4% sodium pentobarbital. The left common carotid artery (CCA), external carotid artery (ECA), and internal carotid artery (ICA) were carefully isolated and exposed under a surgical microscope. The CCA was temporarily ligated, and a small incision was made at the distal end of the ECA. A nylon monofilament was gently inserted through the ECA incision and advanced along the ICA until it occluded the origin of the middle cerebral artery (MCA), thereby inducing focal cerebral ischemia. Occlusion was maintained for 1.5 h, after which the filament was withdrawn to allow reperfusion for 24 h.

Brain targeting and organ distribution

To assess the ischemic lesion-targeting capability of Mit-CY@NPs, the nanoparticles were labeled with the near-infrared fluorescent dye IR780. Specifically, Mit-CY@NPs were mixed with IR780 at a mass ratio of 50:1 and stirred in the dark for 24 h to allow for physical adsorption. The resulting IR780-labeled Mit-CY@NPs (IR780@Nps) were purified by three rounds of centrifugation and washing to remove unbound dye. Following reperfusion, tMCAO/R mice were intravenously injected with IR780@NPs via the tail vein. Realtime in vivo fluorescence imaging was conducted using an IVIS near-infrared imaging system at predetermined time points (0, 2, 4, 6, 8, 10, and 12 h post-injection; n = 3 per group). For quantitative analysis of biodistribution, major organs including brain, heart, liver, spleen, lungs, and kidneys were harvested at 0, 4, 12, and 24 h post-injection (n = 3 per time point). Fluorescence intensity in each tissue was measured ex vivo using IVIS imaging to determine the spatial and temporal distribution of the nanoparticles.

TTC staining

Following euthanasia, the whole brains were carefully extracted and rinsed with PBS to remove residual blood. The brain tissues were flash-frozen at −80°C for 3 min and subsequently sectioned coronally into five 1-mm-thick slices using a rodent brain matrix. The slices were placed in a Petri dish and stained with 1% TTC solution in a light-protected incubator at 37°C for 30 min. The stained sections were then imaged, and infarct volume quantification was performed using ImageJ software. The infarct volume percentage was calculated as follows: Infarct Volume (%) = [(Contralateral Hemisphere Volume - Non-infarcted Ipsilateral Hemisphere Volume)/Contralateral Hemisphere Volume] × 100.

Assessment of cerebral edema

To assess treatment effects on cerebral edema, brain samples were harvested from infarcted mice following experimental interventions. The wet weight of each brain was immediately measured post-collection. Samples were then lyophilized to constant mass, and dry weights were recorded. Cerebral water content was determined using the following equation: Water Content (%) = [(Wet Weight - Dry Weight)/Wet Weight] × 100.

Assessment of the imaging function of hierarchical collapse nanoactuator and therapeutic function in tMCAO mice using MRI techniques

The brains of tMCAO mice were imaged using a T₁-weighted imaging sequence on a 7.0T MRI scanner (Novila) with the following parameters: repetition time (TR) = 500 ms, echo time (TE) = 15 ms, flip angle (FA) = 180°, and acquisition (Acq) = 1. 3D-TOF magnetic resonance angiography was performed on the same 7.0T MRI scanner with imaging parameters: TR = 12 ms, TE = 2 ms, FA = 30°, and Acq = 1. Maximum intensity projection (MIP) was applied during post-processing to reconstruct the vascular architecture. Cerebral edema was assessed using T₂-weighted imaging sequences on a 3.0T Vantage Titan (Canon medical system, Japan) scanner with the following parameters: TR = 2500 ms, TE = 60 ms, number of acquisitions (NAQ) = 5, number of slices = 13, slice thickness = 1.2 mm, inter-slice gap = 0.3 mm, and field of view (FOV) = 5 × 5 cm².

Laser speckle imaging

The cranial crest skin of the mouse was carefully excised, and any residual hair was meticulously removed. A thin layer of coupling agent was then applied to the exposed skull surface. Subsequently, cerebral blood flow was assessed using a laser speckle contrast imaging system.

Tissue sections

Brain tissue was harvested from mice 7 days after reperfusion and fixed in 4% paraformaldehyde (PFA) for 24 h. Following fixation, the samples were dehydrated, embedded in paraffin, and sectioned into 4 μm slices. H&E staining, Nissl staining, and TUNEL staining were performed to assess histopathological changes in the tissue.

In vivo BBB permeability assessed by evans blue (EB) extravasation

To evaluate BBB integrity, EB dye was used. Mice were injected intravenously with 2% EB (4 mL/kg) and allowed to circulate for 1 h. Following anesthesia, brains were collected for imaging, washed with ice-cold PBS, and incubated in formamide at 55°C for 24 h to extract the EB. The extracted EB was quantified by spectrophotometry at 610 nm and normalized to tissue weight to assess BBB permeability. Increased EB extravasation indicates BBB disruption.

Immunofluorescence staining of brain sections

Mice were subjected to cerebral ischemia/reperfusion injury, and after perfusion with PBS followed by 4% paraformaldehyde (PFA), brains were harvested and post-fixed in PFA for 4 h. After cryoprotection in 30% sucrose, the brains were sectioned into 30 μm slices using a cryostat. For immunofluorescence staining, sections were blocked with 5% normal goat serum and 0.3% Triton X-100 in PBS for 1 h, then incubated overnight at 4°C with primary antibodies against p-MLKL (necrosis), GSDMD (pyroptosis), GPX4 and 4-HNE (ferroptosis), Iba1 (microglia), and NeuN (neuronal marker). Additionally, TUNEL staining was performed to assess apoptosis. After incubation with Alexa Fluor-conjugated secondary antibodies, the sections were stained with DAPI to label nuclei. Images were captured using a confocal microscope, and co-localization of Iba1 with CD86 (M₁) and CD206 (M₂) was quantified to evaluate microglial polarization. The expression levels of p-MLKL, GSDMD, GPX4, 4-HNE, and NeuN were measured in the peri-infarct region, while TUNEL staining was used to detect apoptotic cells.

Tissue inflammation damage assay

Drug-injected tMCAO mice were anesthetized for cardiac perfusion for brain extraction. A 10% brain tissue homogenate was prepared in PBS, then centrifuged at 12 000 rpm for 10 min, and the supernatant was collected. The levels of IL-6, IL-10, TNF-α, and Caspase-1 were measured using respective ELISA kits.

tMCAO model neurological function assessment

Neurological function was assessed using the Longa 5-point scoring system: a score of 0 indicates no neurological deficit; a score of 1 indicates inability to fully extend the contralateral forelimb; a score of 2 indicates turning toward the hemiplegic side while walking; a score of 3 indicates leaning toward the hemiplegic side while walking; a score of 4 indicates the inability to walk spontaneously, accompanied by impaired consciousness; and a score of 5 indicates death.

Behavioral tests

The following tests were conducted on days 1, 3, 5, and 7 post-stroke, with n = 3 mice per group. All behavioral tests were performed between 13:00 and 17:00.

Adhesive removal test

This test was used to assess tactile responsiveness and sensorimotor function in mice. A 3 × 3 mm² adhesive patch was placed on the forepaw contralateral to the infarct, and the time taken for the mouse to contact and successfully remove the patch was recorded.

Cylinder test

The cylinder test was used to assess forelimb asymmetry in mice. The mice were placed in a transparent cylinder (height: 15 cm; diameter: 9 cm) and recorded for 5 min using a digital camera. The number of contacts made by each forepaw with the cylinder (left, L; right, R; both, B) was recorded. The asymmetry score (%) was calculated using the formula: (L + R)/(L + R + B)×100.

Open field test

Mice were placed at the center of a 1 m × 1 m open field and allowed to move freely for 5 min while being recorded with a camera. Automatic analysis was performed using Tracker software (version 6.2) to calculate the total distance traveled (in meters) and the average speed (in meters per second).

In vivo toxicity evaluation

Healthy C57 mice were randomly assigned to two groups, with three mice in each group. The experimental group was injected with Mit-CY@Nps (10 mg/kg), while the control group received an equal volume of saline (n = 3). The mice were euthanized 28 days post-injection. H&E staining was performed to assess histological, morphological, and cellular structural changes in various organs. Hematological analysis was conducted on serum samples to evaluate liver and kidney function, including urea nitrogen (UR), creatinine (CR), alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), lactate dehydrogenase (LDH), total protein (TP), and albumin (ALB).

Quantification and statistical analysis

The data were shown as mean ± standard deviation (SD) and analyzed using GraphPad Prism 9.5.1. The significance of the data in this work was evaluated based on the one-way/two-way ANOVA. ns: not significant, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Published: February 24, 2026