Oxidation-responsive PEG-poly(α-lipoic acid) nanoparticles for coenzyme Q₁₀ delivery attenuate hepatic ischemia-reperfusion injury via ROS scavenging and ferroptosis inhibition

Abstract

Hepatic ischemia-reperfusion injury (IRI) is characterized by an acute surge of reactive oxygen species (ROS) upon reperfusion, leading to oxidative damage and cell death. Ferroptosis, a form of iron-dependent lipid peroxidation-driven cell death, has recently been implicated in hepatic IRI, compounding the injury. Here, we present an oxidation-responsive nanoparticle system designed to mitigate liver IRI by scavenging ROS and inhibiting ferroptosis. We synthesized a PEGylated poly(α-lipoic acid) (PEG-PαLA) copolymer that self-assembles into nanoparticles encapsulating the lipophilic antioxidant coenzyme Q₁₀ (CoQ₁₀). The PEG-PαLA/CoQ₁₀ nanoparticles have an average diameter of ∼100 nm and release CoQ₁₀ preferentially under oxidative conditions. In vitro, the nanoparticles efficiently neutralized free radicals and protected hepatocytes from oxidative injury. In a mouse model of partial hepatic IRI, treatment with PEG-PαLA/CoQ₁₀ nanoparticles significantly reduced liver injury markers, preserved liver histology, and abrogated lipid peroxidation, indicating suppression of ferroptosis. Our results demonstrated that the PEG-PαLA/CoQ₁₀ nanoparticle is a novel antioxidant nanomedicine that synergistically attenuates ROS-mediated damage and ferroptotic cell death in hepatic IRI. These findings highlight a promising strategy for protecting organs from ischemia-reperfusion damage by targeting oxidative stress and ferroptosis with responsive biomaterial carriers.

---

1. Introduction

Hepatic ischemia-reperfusion injury (IRI) is a critical complication in liver resection and transplantation, where temporary loss of blood flow (ischemia) followed by restoration of circulation (reperfusion) causes paradoxically exacerbated tissue damage.¹ The sudden influx of oxygenated blood during reperfusion triggers a burst of reactive oxygen species (ROS) that inflict oxidative damage on cellular components, leading to hepatocellular injury and dysfunction.² Among the mechanisms of IRI, excessive mitochondrial ROS generation is considered a hallmark event driving tissue damage and cell death.³ In addition to direct oxidative injury, a form of regulated necrotic cell death known as ferroptosis has been identified as a major contributor to hepatic IRI.^4,5 Ferroptosis is an iron-dependent mode of cell death distinguished by the accumulation of ferrous iron and overwhelming lipid peroxidation of cell membranes. Accumulating evidence links ferroptosis to the pathogenesis of liver IRI, and accordingly, preventing lipid peroxidation and ferroptotic cell death has emerged as a promising therapeutic approach.⁶

Despite the clear roles of ROS and ferroptosis in IRI, current therapeutic strategies have had limited success in translation. Numerous pharmacological agents, ranging from anti-inflammatory drugs to mitochondrial permeability transition pore inhibitors (e.g., cyclosporine A), have shown protective effects in experimental IRI models,⁷ yet none have translated into effective treatments in clinical practice.^8,9 One reason is that many interventions target downstream events (such as inflammation or cell death pathways) but do not adequately neutralize the initial oxidative burst at reperfusion. Consequently, unchecked ROS generation continues to inflict damage even in the presence of those agents. The direct use of small-molecule antioxidants to scavenge ROS has been explored, but such molecules often suffer from rapid clearance, poor accumulation in the liver, and low target specificity. Thus, there is a compelling need for a targeted delivery system that can concentrate antioxidant activity at the site of injury and concurrently interfere with ferroptotic processes to more effectively protect the liver from IRI.

Coenzyme Q₁₀ (CoQ₁₀) is a potent lipophilic antioxidant and a key component of the mitochondrial electron transport chain. In biological membranes, CoQ₁₀ acts as a radical-trapping antioxidant that terminates lipid peroxidation chain reactions, thereby preventing the membrane damage associated with ferroptosis.¹⁰ However, the clinical use of CoQ₁₀ is limited by its extremely poor water solubility and stability; unformulated CoQ₁₀ has low bioavailability. Similarly, α-lipoic acid (αLA) is a well-known antioxidant that can regenerate other antioxidants (such as glutathione and vitamins C/E), but it has suboptimal pharmacokinetics (oral αLA exhibits only ∼30% bioavailability and a short plasma half-life due to rapid metabolism).¹¹ Both CoQ₁₀ and αLA would benefit from a delivery vehicle to improve their solubility, stability, and tissue-targeted delivery. Nanoparticle-based delivery systems are especially attractive in this context, as they can prolong circulation time, reduce renal clearance, and preferentially accumulate in injured tissues, thereby enhancing the therapeutic index of antioxidant compounds. Zhao et al. developed a micellar nanocarrier for the co-encapsulation of doxorubicin and CoQ10, aiming to mitigate doxorubicin-induced cardiotoxicity and inhibit ferroptosis in cardiomyocytes.¹²

Indeed, nanocarriers have been investigated for IRI protection, either by shuttling drugs to injured cells or by providing intrinsic antioxidant activity. Poly(α-lipoic acid) (PαLA) is a biodegradable polymer derived from the naturally occurring αLA. PαLA contains multiple disulfide bonds in its backbone, rendering it sensitive to redox conditions.¹³ Under reducing environments (such as the cytoplasm), disulfides can cleave, and under oxidative stress, the polymer may undergo transformations that affect its stability and drug release profile. Importantly, nanoparticles formulated from PαLA have demonstrated inherent antioxidant properties: they can directly scavenge free radicals in cell-free assays and suppress intracellular ROS generation under oxidative challenge.¹⁴ This suggests that a PαLA-based nanocarrier can itself act as an ROS scavenger and cytoprotective agent. We hypothesized that encapsulating CoQ₁₀ into an oxidative stress-responsive PαLA nanoparticle would provide a synergistic defense against hepatic IRI: the polymer network would scavenge ROS at the onset of reperfusion, and the oxidative environment would trigger the release of CoQ₁₀ to reinforce antioxidant defenses and inhibit ferroptosis in hepatocytes.

In this study, we designed a PEGylated poly(α-lipoic acid) copolymer that self-assembles into CoQ₁₀-loaded nanoparticles (denoted PEG-PαLA/CoQ₁₀ NPs) and evaluated their efficacy in mitigating liver IRI. PEGylation of the PαLA was employed to improve the colloidal stability and biocompatibility of the nanoparticles, as the PEG corona can reduce opsonization and prolong circulation. We thoroughly characterized the nanoparticles’ physicochemical properties, including size, morphology, drug loading, and their responsiveness to oxidative stimuli (ROS). The antioxidant capacity of the PEG-PαLA/CoQ₁₀ NPs was examined in vitro through free radical scavenging assays and cellular protection experiments under oxidative stress. Finally, we tested the therapeutic impact of these NPs in a mouse partial liver ischemia-reperfusion model, focusing on outcomes related to oxidative damage and ferroptosis. Our results show that PEG-PαLA/CoQ₁₀ NPs significantly attenuate liver injury in IRI by reducing ROS accumulation and lipid peroxidation, ultimately preserving hepatocyte viability. This work represents an innovative approach to address the dual targets of oxidative stress and ferroptosis in IRI using a smart nanomedicine, and it highlights the potential of such biomaterial-based strategies for improving outcomes in liver transplantation and other settings of ischemia-reperfusion injury.

2. Materials and methods

2.1. Cell lines and animals

Murine hepatocyte cell line, AML 12 (SIBCB, SCSP-5036), and murine macrophage cell line, RAW 246.7 (SIBCB, SCSP-550) were cultured in Dulbecco's modified Eagle's medium (DMEM, Keygen, KGL1214-500, China) supplemented with 10% fetal bovine serum (FBS, FSS500, Excel, China), 100 U mL⁻¹ penicillin and 100 mg mL⁻¹ streptomycin. Cells were maintained at 37 °C in a humidified incubator with 5% CO₂.

Male C57BL/6 mice aged 6–8 weeks were purchased from GemPharmatech (Guangdong, China). All experimental procedures and animal care protocols were approved by the Ethics Committee for Animal Experimentation (South China Agricultural University, 2021d004). Animals were housed in groups under standard laboratory conditions, with a regular 12-hour light–dark cycle and were provided ad libitum access to food and water.

2.2. Preparation of PEG-PαLA/CoQ₁₀ nanoparticles

PEG-PαLA was synthesized via ring-opening polymerization of α-lipoic acid as previously described.¹³ CoQ₁₀-loaded PEG-PαLA nanoparticles were prepared by a nanoprecipitation/self-assembly method. Briefly, 1 mg of CoQ₁₀ (TargetMol, T2796, Shanghai, China) and 10 mg of PEG-PαLA were co-dissolved in 1.2 mL of N,N-dimethylformamide (DMF, Aladdin, D112004, Shanghai, China). Then the mixture was stirred at room temperature for 10 minutes and dropwise added into 10 mL of stirring deionized water to self-assemble into nanoparticles. The DMF was completely removed through dialysis against water. The concentrated PEG-PαLA/CoQ₁₀ nanoparticles was obtain after ultrafiltration for further studies.

2.3. Physicochemical characterization of PEG-PαLA/CoQ₁₀ Nanoparticles

The size and zeta potential of PEG-PαLA/CoQ₁₀ nanoparticles were measured using a Litesizer 500 particle analyzer (Anton Paar, Austria). Scanning electron microscopy (SEM) images of PEG-PαLA/CoQ₁₀ were captured using a Phenom ProX (Thermo Fisher Scientific, Netherlands). The ROS (the superoxide anion) scavenging capacity was evaluated through a 2-methyl-6-phenyl-3,7-dihydroimidazo[1,2-a]pyrazin-3-one (CLA, TCI, A5307, Japan) chemiluminescence assay.¹⁵ Briefly, superoxide anion was generated within a 5% glucose solution containing 0.2 mM xanthine and 0.23 μunits per mL xanthine oxidase (XO) by oxidation of xanthine. PEG-PαLA was added to the mixture to scavenging superoxide anion, shaking for 10 min at 37 °C. Then CLA was added to the mixture to give a final concentration of 16 μM. The light produced from oxidation of CLA by the superoxide anion was immediately measured using a microplate reader (BioTek, Synergy H1MF, USA). The H₂O₂ scavenging capacity was evaluated using the xylenol orange (Beyotime, S0038, China) assay. The test solution was first incubated with H₂O₂ (20 μM) in PBS buffer at 37 °C for 3 h. Subsequently, 50 μL of the resulting solution was mixed with the detecting solution according the manufactory introduction, producing a purple product with a specific absorption peak at 560 nm.

2.4. Live imaging of biodistribution of PEG-PαLA/DiR nanoparticles

For imaging of the biodistribution of nano-carrier PEG-PαLA in vivo, The DiR was used as a fluorescent probe encapsulated in PEG-PαLA. The synthesis of fluorescent nanoparticles PEG-PαLA/DiR followed a similar procedure as described above for the synthesis of PEG-PαLA/CoQ₁₀. The 200 μL of PEG-PαLA/DiR or free DiR solutions were injected into BALB/c mice via tail veins. The biodistribution of DiR in mice was observed by VISQUE InVivo Smart (Vieworks, Korea) using near-infrared fluorescence imaging for the indicating time. Twenty-four hours after injection, mice were euthanized, and the heart, lung, liver, spleen, and kidney of mice were collected for ex vivo fluorescence imaging using VISQUE InVivo Smart (Vieworks, Korea). The fluorescence of DiR in liver sections of mice at 6 hours after injection were observed by the confocal microscope (Leica TCS SP8, German).

2.5. Antioxidant activity of PEG-PαLA/CoQ₁₀ in CCCP or H₂O₂-challenged AML-12 Cells

Carbonyl cyanide m-chlorophenylhydrazone (CCCP, T7081, TargetMol, China) is an inhibitor of mitochondrial electron transport chain, which induces mitochondrial dysfunction and causes generation of ROS (superoxide).¹⁶ MitoSOX Red is a fluorescent probe that specifically targets mitochondria in live cells, with a great cellular membrane permeable ability. MitoSOX Red is exclusive oxidized by superoxide within mitochondria and reveal the ROS level induced by CCCP. JC-1 is a well-established fluorescent probe that is widely utilized for detecting mitochondrial membrane potential in live cells with high sensitivity and reliability. AML-12 were seeded into a 24-well plate overnight and preincubated with PEG-PαLA or PEG-PαLA/CoQ₁₀ for 3 h. Then, cells were exposed to CCCP (50 μM) for 30 min, or exposed to H₂O₂ stimulation (500 μM). Subsequently, cells were incubated with MitoSOX Red (50 μM) or JC-1 (10 μg mL⁻¹) in serum-free medium at 37 °C for 30 min. After washing with PBS buffer to remove excess probes, representative fluorescence images were captured using the fluorescence microscope (ECLIPSE Ni-U, Nikon, Japan).

2.6 Oxygen-glucose deprivation and reoxygenation (OGD/R) model

The standardized procedure for OGD/R was carried out as described.¹⁷ The culture medium for RAW 264.7 cells was replaced with glucose-free Dulbecco's Modified Eagle Medium (catalog no. 11966025, Thermo Fisher Scientific, MA, USA). Then cells were transferred to a hypoxic environment with a mixture of 94% N₂ and 5% CO₂ and 1% O₂ in a 37 °C incubator (Galaxy 48R, Eppendorf, Germany) for 9 hours. During reoxygenation, the hypoxic medium was replaced complete medium and cultured for 2 hours. A fluorescent probe, 2′,7′-dichlorofluorescin diacetate (DCFH-DA, D6883, Sigma-Aldrich, Germany), was employed to visualize the intracellular ROS in RAW 264.7 after OGD/R. Then Quantitative analysis of intracellular ROS levels was performed with a flow cytometry (LSRFortessa X-20, BD, USA) and the FlowJo software ver. 10.0.

2.7. Liver ischemia and reperfusion injury (LIRI) model

The LIRI model were established as previously described with modifications.⁵ Briefly, mice were fasted for 12 hours and free to water, then were anesthetized with intraperitoneal 1% pentobarbital sodium (50 mg kg⁻¹), and injected with 200 μL of normal saline, NAC, PEG-PαLA, or PEG-PαLA/CoQ₁₀ NPs via tail vein 1 h before surgery, followed by undergoing a median incision. After separation of hepatic portal region structure, the hepatic artery, bile duct, and portal vein of the left lateral and median lobes (approximately 70% of liver) were blocked with a non-invasive vascular clamp for 1 hour. Then mice received a 6 hours stage of reperfusion and euthanasia at the end of the experiment. The blood liver tissues of mice were collected for further experiment.

2.8. Serum biomarker measurement

Alanine transaminase (ALT), aspartate aminotransferase (AST), urea nitrogen (BUN), creatinine (Cr), lactate dehydrogenase (LDH), creatine kinase (CK), creatine kinase isoenzymes (CK-MB) level was analyzed by the automatic analyzer (Hitachi, 7600-020, Japan).

2.9. Liver ROS level detection

The frozen section of liver tissues of mice was prepared in and rewarmed to room temperature, then incubated with the fluorescence probe dihydroethidium (DHE, S0063, Beyotime, China) staining solution for 30 minutes at 37 °C. The images of liver were captured using the fluorescence microscope (ECLIPSE Ni-U, Nikon, Japan) and the relative fluorescent intensity were analyzed using Image J software (National Institutes of Health, USA).

2.10. Hematoxylin and eosin (H&E) staining

The heart, lung, liver, spleen, and kidney tissues were fixed with formalin, embedded with paraffin, and then were cut into 5-μm sections, deparaffinized in xylene, rehydrated through graded ethanol. Sections were subjected to hematoxylin (Jiu Zhou Bai Lin, Beijing, China) staining for 5 minutes and then eosin (ZSGB-BIO, Beijing, China) staining for 3 minutes. The pathological injury of liver was graded by Suzuki's Score.

2.11. Immunohistochemistry (IHC)

IHC was used to examine protein levels in liver tissues. Formalin-fixed, paraffin-embedded tissues were cut into 5-μm sections, deparaffinized in xylene, rehydrated through graded ethanol and applied to IHC staining using anti-MPO (Abcam, ab208670, UK), anti-4-HNE (Abcam, ab46545, UK), anti-8-OHdG (Abcam, ab48508, UK), Cyto C (CST), HO-1 (Zenbio, R24541, China), or COX-2 (Abcam, ab15191, UK). The slides were scanned by the Digital Slide Scanner (PANNORAMIC MIDI, 3DHISTECH, Hungary) and the intensity of MPO, 4-HNE, 8-OHdG, Cyto C, HO-1 and COX-2 staining was calculated based on per unit areas using Image J software (National Institutes of Health).

2.12. Mitochondrial morphological analysis

For the detection of mitochondrial morphological changes, liver tissues were fixed in 4% glutaraldehyde solution for 24 hours and stained with 1% osmium tetroxide. Then samples assessed to Transmission electron microscopy (TEM, HT7700, Hitachi, Japan) analysis. The axial ratio of mitochondria was calculated by dividing the major axis (length) by the minor axis (width) of individual mitochondria, and used to indicate the functional state of mitochondria.

2.13. Hemolysis test

1 mL of whole blood drawn from the abdominal aorta of mouse was mixed with 2 mL of saline in EDTA-anticoagulant tube. The mixture was centrifuged at 2500 × rpm for 5 minutes at 4 °C. Then the transparent supernatant was removed and the 2% erythrocyte suspension was prepared. And 1 mL of 2% erythrocyte suspension was mixed with l mL of PEG-PαLA/CoQ₁₀ NPs solution (0.5 mg mL⁻¹), 1 mL of normal saline, or 1 mL of deionized water. After incubation at 37 °C for 2 hours, the clarity of supernatant and remaining erythrocytes in tubes were observed and captured by camera.

2.14. Cell viability assay

The biocompatibility of PEG-PαLA/CoQ₁₀ NPs in RAW 264.7 and AML-12 cells was determined using a Cell Counting Kit-8 (CCK-8, KeyGen, China) assay. RAW 264.7 (1 × 10⁴) and AML 12 (8 × 10³) were treated with PEG-PαLA/CoQ₁₀ NPs in indicating concentration (0, 0.5, 1.0, 1.5, 2.0 mg mL⁻¹) for 24 hours. Then the cells were incubated with 10% Cell Counting Kit-8 (CCK-8, KeyGen, China) solution for another 1 hour. The cell viabilities were measured at an absorbance of 450 nm using a microplate reader (BioTek, USA).

2.15. Statistical analysis

Data were presented as the mean ± SD from at least three independent experiments. The differences between two groups were analyzed using one-way analysis of variance (ANOVA). A P-value < 0.05 was considered statistically significant, and all statistical tests were two-sided. All analyses were performed using GraphPad Prism software (GraphPad Software, Inc., San Diego, CA, USA).

3. Results and discussion

3.1. Synthesis and characterization of PEG-PαLA/CoQ₁₀ NPs

The synthesis process of PEG-PαLA/CoQ₁₀ nanoparticles are displayed in Scheme 1. Briefly, PEG-PαLA and CoQ₁₀ were prepared through self-assembly. As a hydrophobic drug, CoQ₁₀ can be effectively encapsulated by amphipathic PEG-PαLA into negatively charged NPs. Concerning the rationality of in vivo studies, the mass ratio of 1/10 (CoQ₁₀/PEG-PαLA) with smaller size and less electronegativity was chosen for the following studies. However, the relatively negative charged NPs tended to aggregate with increasing diameter after lyophilization. After concentration by lyophilizing, the NPs underwent a great physical change in the redispersed solution, comparing with those concentrated by ultrafiltration (170 nm vs. 110 nm, −30 mV vs. −18 mV) (Fig. 1a and b). These results indicate that ultrafiltration may slightly promote the reunion of NPs. The transmission electronic microscopy (TEM) image shown that NPs had uniform morphology and good monodispersity (Fig. 1c). Two typical ROS, namely, H₂O₂ and superoxide anion radicals were used to assess the multiple ROS-scavenging properties of PEG-PαLA in vitro. The superoxide anion radicals were generated by the xanthine/xanthine oxidase (XO) system. Intramolecular disulfide bonds of PEG-PαLA may be cleaved in oxidative microenvironment, achieving robust ROS-scavenging capabilities. As shown in Fig. 1d–f, PEG-PαLA effectively eliminate nearly 70% of superoxide anion radicals and H₂O₂ at a concentration of 0.25 mg mL⁻¹. These results demonstrated that PEG-PαLA have a great capacity for scavenging ROS in a concentration-dependent manner.

Scheme 1 Illustration depicting the production of PEG-PαLA/CoQ₁₀ nanoparticles and the treatment for HIRI. Created with https://Biorender.com.

Fig. 1 Characterization of PEG-PαLA/CoQ₁₀ NPs. (a, b) Average hydrodynamic sizes and zeta-potentials of PEG-PαLA/CoQ₁₀ NPs. NPs were prepared through self-assembly and dialysis. The sizes and zeta-potentials of nanoparticles were compared between samples prepared by ultrafiltration and lyophilization. (c) The SEM image of nanoparticles. (d) Schematic illustration of the ROS-scavenging capability of PEG-PαLA. (e) The superoxide anion or hydrogen peroxide was added to PEG-PαLA solutions and the extra component was detected. The luminescent intensity of CLA detecting extra superoxide anion in PEG-PαLA solutions (0, 0.025, 0.05, 0.1, 0.25, 0.5 mg mL⁻¹). (f) The relative content of remaining hydrogen peroxide in PEG-PαLA solutions.

3.2. PEG-PαLA/CoQ₁₀ NPs can scavenge intracellular ROS

As hepatocyte injury and oxidative stress caused by Kupffer cell activation is one of characteristics of liver IRI,¹⁸ the protection of PEG-PαLA/CoQ10 NPs against intracellular ROS was evaluated on mouse macrophages RAW 264.7 and hepatocytes AML-12, respectively.¹⁹ During hepatic I/R injury, macrophages can be activated and then release ROS, which creates oxidative microenvironment to further damage hepatocytes.⁵ Therefore, we hypothesized that pretreatment with PEG-PαLA/CoQ₁₀ NPs could protect hepatocytes by effectively scavenging intracellular ROS.

CCCP (carbonyl cyanide m-chlorophenylhydrazone) is an inhibitor of mitochondrial electron transport chain, which induces mitochondrial dysfunction and causes oxidative stress. CCCP treatment significantly induced the burst of mitochondrial superoxide which oxidized the exclusive probe, MitoSOX Red, to emit red fluorescence. PEG-PαLA (0.2 mg mL⁻¹) or PEG-PαLA/CoQ₁₀ treatment abolished the CCCP-induced increase of mitochondrial ROS level (Fig. 2a). H₂O₂ can attack organelle membrane and then impair integrity of mitochondria and induce depolarization of mitochondria. Fig. 2b showed that H₂O₂ (500 μM for 3 hours) treatment decreased the fluorescence ratio of red to green, indicating depolarization of the mitochondria. PEG-PαLA (0.2 mg mL⁻¹) or PEG-PαLA/CoQ₁₀ treatment attenuated H₂O₂-induced declination in red fluorescence, increase in green fluorescence and depolarization of mitochondria (Fig. 2b). Owing to the excellent antioxidant activity, PEG-PαLA/CoQ₁₀ NPs significantly decreased mitochondrial ROS levels and mitigated oxidative mitochondrial damage in AML-12 cells, indicating that PEG-PαLA/CoQ₁₀ NPs were effectively internalized by hepatocytes and that CoQ₁₀ played a crucial role in protecting mitochondria from oxidative stress.

Fig. 2 PEG-PαLA/CoQ₁₀ NPs scavenge intracellular ROS. (a, b) PEG-PαLA/CoQ₁₀ NPs treatment reduced intracellular oxidative stress. AML 12 cells incubated with PBS, PEG-PαLA, or PEG-PαLA/CoQ₁₀ for 24 hours, and then treated with or without CCCP or H₂O₂, and subjected to Mito-SOX (a) or JC-1 (b) staining. The cells were examined under a fluorescence microscopy. (c) PEG-PαLA/CoQ₁₀ NPs treatment reduced ROS level. RAW 264.7 cells were incubated with PEG-PαLA or PEG-PαLA/CoQ₁₀ NPs for 6 hours, followed by hypoxia for 9 hours and then reoxygenation for another 2 hours. The intercellular ROS levels were detected using DCFH-DA by flow cytometry (upper panel), and the mean fluorescence intensity was subsequently analyzed (lower panel). P values were assessed by one-way ANOVA (a)–(c). *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.

OGD/R is used as a universal cellular model to mimic ischemia of tissues. For OGD/R, RAW 264.7 cells were incubated in glucose-free and serum-free media under hypoxia for 9 hours and reoxygenation for 2 hours. Flow cytometry analysis confirmed the elevated ROS level in RAW 264.7, mimicking the process of Kupffer cell activation and bringing about injury after liver IR in vivo. We found that PEG-PαLA (0.2 mg mL⁻¹) or PEG-PαLA/CoQ₁₀ NPs treatment could weaken the OGD/R-induced increase in ROS level of RAW 264.7 (Fig. 2c).

However, the cellular experiments demonstrated that the effects of PEG-PαLA and PEG-PαLA/CoQ₁₀ treatments were comparable in RAW 264.7 cells, and the protective role of CoQ₁₀ was not prominently highlighted. This might be attributed to the use of a single-cell model and the limited scope of the experimental intervention, which may not fully recapitulate the in vivo physiological process at this stage.

3.3. PEG-PαLA/DiR NPs are passively enriched in liver

Ischemia-reperfusion or acute inflammation may cause liver injury and dysfunction,²⁰ those endothelial damages and activated platelets aggregation may impair the sinusoidal perfusion, further accelerating liver injury.² The NPs that are intravenously injected passing hepatic sinusoid can be captured by Kupffer cells and passively enriched in liver. And those nanomaterials with negative potentials are more likely to be phagocytosed by macrophages. To quantify the biodistribution of the NPs consist of PEG-PαLA, we prepared PEG-PαLA/DiR NPs and injected them into mice via tail veins. Lipid-soluble DiR can likewise be wrapped inside PEG-PαLA by self-assembly. The fluorescence of DiR can be visualized and captured by in vivo imaging system. In the initial 5 minutes, the radiant efficiency in the upper abdomen of mice injected with PEG-PαLA/DiR NPs was higher than that of mice injected with single DiR (Fig. 3a, b). DiR loaded in PEG-PαLA could rapidly accumulate in the liver, reaching a sufficient concentration to generate a detectable signal. In contrast, free DiR exhibited a much lower signal intensity in the liver, indicating that its accumulation in the liver was significantly less than that of PEG-PαLA/DiR. With time extension, both the free DiR-treated group and the PEG-PαLA/DiR-treated group exhibited fluorescent signal peaks and maintained stability since 1 h post-injection. The higher radiant efficiency observed in the upper abdominal region persisted for up to 24 hours in the PEG-PαLA/DiR group. Since the acute-phase inflammation alters the immune microenvironment during reperfusion stage, which may exacerbate liver injury,²¹ prolonged accumulation of anti-inflammatory drugs would be more effectively to alleviate damages.

Fig. 3 The biodistribution of the PEG-PαLA/DiR NPs in vivo. (a, b) The biodistribution of the PEG-PαLA/DiR NPs in vivo. The DiR or PEG-PαLA/DiR NPs were injected into mice via tail veins (n = 3/group), following by detecting the fluorescence of DiR in mice by animal in vivo imaging system in the indicating times (a). And the average fluorescent intensities of DiR in the upper abdominal region of mice were analyzed (b). (c) The fluorescence intensities of the PEG-PαLA/DiR NPs of ex vivo organs. The DiR or PEG-PαLA/DiR NPs were injected into mice via tail veins for 24 hours (n = 3/group), and the fluorescence of DiR in the indicating organs was detected by animal in vivo imaging system (left panel), and the average fluorescent intensities of DiR were analyzed correspondingly (right panel). (d) The PEG-PαLA/DiR NPs could be passively target to the liver of mice. The DiR or PEG-PαLA/DiR NPs were injected into mice via tail veins for 6 hours, and the fluorescence of DiR in mice liver sections were examined. P values were assessed by two-way ANOVA (b), or paired Student's t-test ((c), right panel). *P < 0.05; **P < 0.01; ***P < 0.001.

The radiant efficiency of major organs of mice, such as heart, lung, liver, spleen, kidney, was measured after injections in vitro. The results shown that both PEG-PαLA/DiR-injected group and DiR-injected group had highest aggregation of DiR in liver, and PEG-PαLA/DiR-injected group shown higher radiant efficiency than that single DiR-injected group (Fig. 3c). The passive liver targeting ability of PEG-PαLA/DiR NPs was further confirmed by fluorescence microscope of liver tissue sections (Fig. 3d). The results indicate that PEG-PαLA NPs can be passively enriched in liver in vivo.

3.4. PEG-PαLA/CoQ₁₀ NPs reduce oxidative stress and attenuate liver IRI

We further investigated the protective effects of PEG-PαLA/CoQ₁₀ NPs on liver IRI of mice. Based on the liver accumulation ability of PEG-PαLA/CoQ₁₀ NPs observed above, we selected 1 h as the pretreatment time point before liver ischemia. The prophylactic effect of PEG-PαLA/CoQ₁₀ on liver IRI was compared with a positive control, N-acetylcysteine (NAC), a clinically used hepato-protective agent that serves as a cysteine precursor.²² Recent studies have demonstrated that NAC can neutralize ROS and mitigate lipid peroxidation through restoring cellular glutathione (GSH) levels, thereby resisting ferroptosis.⁶ Our results demonstrated that ischemia-reperfusion (IR) induced significant histological damage to the liver, whereas treatment with PEG-PαLA or PEG-PαLA/CoQ₁₀ NPs (CoQ₁₀, 2 mg kg⁻¹) significantly alleviated IR-induced hepatocyte swelling and cell death, showing effects comparable to those of NAC (Fig. 4a and c). Consistently, IR markedly increased serum levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST), while treatment with NAC, PEG-PαLA, or PEG-PαLA/CoQ₁₀ NPs significantly mitigated the IR-induced upregulation of serum AST and ALT (Fig. 4e). Notably, PEG-PαLA/CoQ₁₀ treatment exhibited a more pronounced advantage in reducing histological damage and serum level of AST and ALT, compared to other treatments. Furthermore, the inflammation reactions of livers were measured by immunohistochemical staining of MPO (myeloperoxidase), which is a peroxidase enzyme and a biomarker of activated neutrophils. The result shown that neutrophils were accumulated in liver after IR, all treatments had comparable abilities in alleviating IR-induced accumulation of neutrophils (Fig. 4b and d). It may be attributed to adequate pretreatment and a lower level of inflammation during the early stage of reperfusion.

Fig. 4 PEG-PαLA/CoQ₁₀ NPs can attenuate liver IR injury. (a, c) PEG-PαLA/CoQ₁₀ NPs reduced histologically injury liver. The mice (n = 5/group) were treated with NS, NAC, PEG-PαLA, or PEG-PαLA/CoQ₁₀ NPs for 1 hour, followed by liver ischemia for 0 or 1 hour, and then reperfusion for 6 hours. The images of H&E-staining section of liver (a) and the Suzuki score of histologically injury (c) were shown. (b, d) PEG-PαLA/CoQ₁₀ NPs reduced inflammation level of liver. Immunochemistry staining of neutrophil markers MPO in liver (b) and average counting in the field (d). (e) The AST and ALT levels of serum. P values were assessed by one-way ANOVA (c–e). HPF, high power field. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.

Oxidative stress is widely recognized as a hallmark of tissue injury in liver IRI. The ROS levels in liver tissue were quantitatively assessed by DHE staining on liver sections. The results shown that ROS levels were significantly increased in liver IRI, and NAC, PEG-PαLA or PEG-PαLA/CoQ₁₀ NPs treatment reduced the IR-induced increase in ROS levels (Fig. 5a and b). The surge in ROS production results in oxidative stress and leads to oxidation of biomacromolecule like lipid and DNA, which causes membrane damage and amplification of inflammatory responses. Therefore, lipid peroxidation and DNA oxidation of liver were detected by immunohistochemistry of 4-HNE and 8-OHdG, respectively. Results revealed that lipid peroxidation and DNA oxidation of liver were increased in liver IRI, while NAC, PEG-PαLA or PEG-PαLA/CoQ₁₀ NPs treatment weaken the IR-induced increase in the level of lipid peroxidation and DNA oxidation (Fig. 5c–f). Our results showed that levels of ROS and 4-HNE were lower in PEG-PαLA/CoQ₁₀ treated group than in PEG-PαLA-treated group or NAC-treated group, indicating that PEG-PαLA/CoQ₁₀ NPs have a more desirable in down-regulating levels of cellular lipid peroxidation after hepatic ischemia-reperfusion. These results indicate that PEG-PαLA/CoQ₁₀ NPs may have good ability to attenuate liver IRI by reducing oxidative stress, even better than the NAC treatment. Although NAC has been used clinically, its therapeutic efficacy in HIRI is limited, possibly due to its relatively low bioavailability, rapid metabolism, and the need for high doses to achieve effective concentrations, which may lead to potential side effects. Moreover, NAC's ability to target specific tissues or cells involved in HIRI is also limited, resulting in suboptimal therapeutic outcomes in some cases. In contrast, PEG-PαLA/CoQ₁₀ NPs not only tend to be enriched in liver, but also ensure that CoQ10 is effectively utilized within the liver. This may explain why PEG-PαLA/CoQ₁₀ NPs have the potential to improve clinical outcomes for HIRI more effectively than NAC.

Fig. 5 PEG-PαLA/CoQ₁₀ NPs reduce oxidative stress in post-IR liver. (a–f) PEG-PαLA/CoQ₁₀ NPs could reduce ROS levels (a, b), lipid peroxidation levels (c, d), or DNA oxidation level (e, f) in post-IR liver. The mice were treated with NS, NAC, PEG-PαLA, or PEG-PαLA/CoQ₁₀ NPs for 1 hour, followed by liver ischemia for 0 or 1hour, and then reperfusion for 6 hours. Liver tissues were collected for DHE staining (a) or immunochemistry staining of 4-HNE (c) and 8-OHdG (e), and relative fluorescent intensity of DHE (b), relative level of 4-HNE (d) and 8-OHdG (f) were measured. P values were assessed by one-way ANOVA (b), (d) and (f). *P < 0.05; ***P < 0.001; ns, not significant.

3.5. PEG-PαLA/CoQ₁₀ NPs reduce mitochondrial damage by inhibiting cell ferroptosis

Accumulation of lipid peroxidation products triggers cell ferroptosis and induces biological membrane degeneration, including mitochondrial dysfunction.²³ Morphological alterations in mitochondrial damage are currently recognized as key indicators of ferroptosis initiation. When oxidative stress arises in the cytoplasm, the ensuing damage progressively propagates to the mitochondria, such as phospholipid oxidation within the mitochondrial membrane.²⁴ The rupture of the mitochondrial outer membrane signifies an irreversible stage in the progression of ferroptosis. Mitochondrial outer membrane disruption and cristae reduction occur following hepatic ischemia-reperfusion (IR). To compensate for these changes, mitochondria attempt to maintain a relatively healthy population and functionality through increased fission.²⁵ Transmission electron microscopy (TEM) analysis of liver tissue revealed that hepatic IR induced the formation of more rounded andshortened mitochondria, along with a significantly decreased axial ratio of mitochondria (Fig. 6a), revealing enhanced mitochondrial fission during ferroptosis. Notably, only PEG-PαLA/CoQ₁₀ NPs treatment effectively restored the axial ratio of mitochondria, whereas neither NAC nor PEG-PαLA treatment exhibited such an effect (Fig. 6e).

Fig. 6 PEG-PαLA/CoQ₁₀ NPs reduce mitochondrial damage by inhibiting cell ferroptosis. (a, b, e, and f) PEG-PαLA/CoQ₁₀ NPs reduced mitochondrial damage. (c, d, g, and h) PEG-PαLA/CoQ₁₀ NPs decreased the level of ferroptosis-related enzymes HO-1 and COX-2. The mice were treated with NS, NAC, PEG-PαLA, or PEG-PαLA/CoQ₁₀ NPs for 1 hour, followed by liver ischemia for 0 or 1hour, and then reperfusion for 6 hours. The mitochondrial morphology of liver tissues was detected by TEM images (a) and the axial ratio of mitochondria was calculated (e). The levels of Cyto C (b, f), HO-1 (c, g) and COX-2 (d, h) in liver were detected by immunochemistry staining. P values were assessed by one-way ANOVA. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.

Mitochondrial fission may cause release of cytochrome C (Cyto C) to cytoplasm. As shown in Fig. 6b, IR significantly increased the level of Cyto C of liver, while NAC, PEG-PαLA or PEG-PαLA/CoQ₁₀ NPs treatment impaired IR-induced increase in Cyto C level (Fig. 6f). In addition, PEG-PαLA/CoQ₁₀ NPs treatment reduced Cyto C level of liver more than NAC or PEG-PαLA treatment. These results suggest that PEG-PαLA/CoQ₁₀ NPs have excellent abilities in protecting mitochondrial damage caused by liver IRI. Furthermore, the expression of ferroptosis-related enzymes HO-1 and COX-2, which involved in generation of lipid peroxidase,^26,27 were detected by immunohistochemistry. The results shown that the expression of HO-1 (Fig. 6c and g) and COX-2 (Fig. 6d and h) were upregulated in IR liver, which was inhibited by NAC, PEG-PαLA or PEG-PαLA/CoQ₁₀ NPs treatment. Moreover, PEG-PαLA/CoQ₁₀ NPs treatment was likely more capable to decease the expression of HO-1 and COX-2 than NAC treatment. From the above evidence, PEG-PαLA/CoQ₁₀ NPs can inhibit upregulation of HO-1, lipid peroxidation and mitochondrial fission, suppressing cell ferroptosis. Previous study has reported that chiral polymer micelles effectively suppressed ferroptosis and repressed inflammatory cytokines in macrophage.²⁸ Our results highlight the multifunctional characteristics of the amphiphilic and self-assembled PEG-PαLA micelles, which not only effectively inhibit ferroptosis in hepatocytes but also suppress inflammation in macrophages.

3.6. PEG-PαLA/CoQ₁₀ NPs have good biocompatibility in vitro and in vivo

Next, we evaluated the biocompatibility of PEG-PαLA/CoQ₁₀ NPs both in vitro and in vivo. Co-incubation of PEG-PαLA/CoQ₁₀ NPs with erythrocytes did not induce significant hemolytic changes, suggesting the great biocompatibility with erythrocytes (Fig. 7a). RAW264.7 and AML12 cells were co-cultured with PEG-PαLA/CoQ₁₀ NPs at various concentrations, and cell viability was assessed using the CCK-8 assay. The results demonstrated that medium containing up to 2 mg mL⁻¹ of PEG-PαLA/CoQ₁₀ NPs did not significantly reduce the viability of RAW264.7 or AML12 cells (Fig. 7b). These results suggest that PEG-PαLA/CoQ₁₀ NPs exhibit good safety in vitro. Furthermore, we employed animal experiments to verify the biological safety of PEG-PαLA/CoQ₁₀ NPs in vivo. The serum and major organs of mice treated with PBS or PEG-PαLA/CoQ₁₀ NPs (1.1 g kg⁻¹) were collected on day 7 after injection for further evaluation. The hematoxylin–eosin (H&E) staining shown that PBS or PEG-PαLA/CoQ₁₀ NPs injection had similar effects on the major organs of mice (Fig. 7c). There had no significant differences in the levels of serum AST, ALT, blood urea nitrogen (BUN), creatinine (Cr), lactate dehydrogenase (LDH), and creatine kinase (CK, CK-MB), which reflected functions of major organs, between PBS-treated group and PEG-PαLA/CoQ₁₀ NPs-treated group (Fig. 7d). These results suggest that PEG-PαLA/CoQ₁₀ NPs exhibit excellent biosafety in animals.

Fig. 7 The great biocompatibility of the PEG-PαLA/CoQ₁₀ NPs. (a) PEG-PαLA/CoQ₁₀ NPs didn’t cause hemolysis of erythrocyte. Image of the hemolysis test. Erythrocyte were incubated in PEG-PαLA/CoQ₁₀ NPs solutions for 1 hour and then imaged by camera. Normal saline (NS) and deionized water (ddH₂O) were used as negative and positive controls, respectively. (b) PEG-PαLA/CoQ₁₀ NPs had no impair on cell viability. RAW264.7 and AML12 were incubated with PEG-PαLA/CoQ₁₀ NPs for 24 hour and subjected to cell viability analysis by CCK8 kit. (c, d) NPs exhibit excellent biosafety in mice. NPs or PBS were injected into mice via tail veins for 7 days, then major organs and serum of mice were collected (n = 3/group). Images of major organs section stained with H&E (c). Biomarkers level in serum of liver, kidney or heart (d) was detected. BUN; Cr; LDH; CK; CK-MB.

4. Conclusion

We have developed a versatile oxidative stress-responsive nanoparticle system that delivers CoQ₁₀ to sites of tissue injury and achieves potent protection against hepatic ischemia-reperfusion injury. The PEG-PαLA/CoQ₁₀ nanoparticles combine the intrinsic antioxidant capability of PEG-PαLA with the high potency of CoQ₁₀, providing a two-pronged defense against IRI. Under the oxidative conditions of reperfusion, the nanoparticles scavenge ROS and concurrently release CoQ₁₀, thereby significantly reducing oxidative damage and preventing ferroptotic cell death in the liver. Our in vivo studies in a mouse model of hepatic IRI demonstrated that PEG-PαLA/CoQ₁₀ NPs markedly scavenged intracellular ROS, reduced oxidative stress and lipid peroxidation, and alleviated mitochondrial damage. The present work highlights a promising strategy for addressing the long-standing challenge of mitigating ischemia-reperfusion injury. To our knowledge, this is the first demonstration of a polymeric nanomedicine that explicitly targets ferroptosis in the context of hepatic IRI by co-delivering a ROS-responsive matrix and a ferroptosis-suppressing drug. Targeting the dual pathways of oxidative stress and ferroptosis via an oxidation-responsive nanoparticle offers a compelling therapeutic avenue for preventing tissue damage in ischemia-reperfusion injuries. This work lays the foundation for future preclinical development and eventually clinical translation of antioxidant nanotherapeutics to improve outcomes in liver surgery and transplantation.

Author contributions

Yu Guan, Jing Yang, Sufang Chen and Shengyuan Gong: investigation, experiments, data curation, writing – original draft preparation. Jinyu Liu, Weifeng Yao: resources, writing – review and editing, project administration. Chunsheng Xiao, Mingqiang Li, Gangjian Luo and Ziqing Hei: funding acquisition, supervision, and writing – review and editing. All authors read and approved the final manuscript.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Joint Funds of the National Natural Science Foundation of China (U22A20276), Science and Technology Planning Project of Guangdong Province-Regional Innovation Capacity and Support System Construction (2023B110006), Science and Technology Program of Guangzhou, China (202201020429), the “Five and five” Project of the Third Affiliated Hospital of Sun Yat-Sen University (2023WW501), National Natural Science Foundation of China (82302469, 82402913), Guangdong Basic and Applied Basic Research Foundation (2022A1515012256, 2021A1515111133).

Notes and references

1. Y. Guan, W. Yao, K. Yi, C. Zheng, S. Lv, Y. Tao, Z. Hei and M. Li, Small, 2021, 17, e2007727.
2. H. Hirao, K. Nakamura and J. W. Kupiec-Weglinski, Nat. Rev. Gastroenterol. Hepatol., 2022, 19, 239–256.
3. Z. Liu, Z. Yao, H. Yang, S. Wu and Z. Tang, Nanoscale, 2025, 17, 3688–3697.
4. N. Yamada, T. Karasawa, T. Wakiya, A. Sadatomo, H. Ito, R. Kamata, S. Watanabe, T. Komada, H. Kimura, Y. Sanada, Y. Sakuma, K. Mizuta, N. Ohno, N. Sata and M. Takahashi, Am. J. Transplant., 2020, 20, 1606–1618.
5. S. Wu, J. Yang, G. Sun, J. Hu, Q. Zhang, J. Cai, D. Yuan, H. Li, Z. Hei and W. Yao, Br. J. Pharmacol., 2021, 178, 3783–3796.
6. S. Guo, Z. Li, Y. Liu, Y. Cheng and D. Jia, Free Radical Res., 2024, 58, 396–416.
7. N. Nighoghossian, M. Ovize, N. Mewton, E. Ong and T. H. Cho, Cerebrovasc. Dis., 2016, 42, 309–318.
8. C. Bellini, S. Antonucci, L. Morillas-Becerril, S. Scarpa, R. Tavano, F. Mancin, F. Di Lisa and E. Papini, Antioxidants, 2022, 11, 907.
9. J. Hefler, S. Hatami, A. Thiesen, M. J. Wagner, G. Mainardi, S. Himmat, C. J. Karvellas, D. L. Bigam, D. H. Freed and A. M. J. Shapiro, Ann. Transplant., 2024, 29, e941054.
10. Z. Li, M. Lange, S. J. Dixon and J. A. Olzmann, Annu. Rev. Biochem., 2024, 93, 499–528.
11. N. Mosallaei, A. Malaekeh-Nikouei, S. Sarraf Shirazi, J. Behmadi and B. Malaekeh-Nikouei, Bioimpacts, 2024, 14, 30136.
12. T. Chen, Y. Qin, Y. Li, Y. Li, J. Luo, L. Fan, M. Feng, Z. Wang and Y. Zhao, Adv. Funct. Mater., 2023, 33, 2300689.
13. H. Yang, W. Shen, W. Liu, L. Chen, P. Zhang, C. Xiao and X. Chen, Biomacromolecules, 2018, 19, 4492–4503.
14. R. Su, H. Wang, C. Xiao, Y. Tao, M. Li and Z. Chen, Biomater. Sci., 2021, 9, 4746–4754.
15. H. Katsumi, K. Fukui, K. Sato, S. Maruyama, S. Yamashita, E. Mizumoto, K. Kusamori, M. Oyama, M. Sano, T. Sakane and A. Yamamoto, Metallomics, 2014, 6, 1050–1056.
16. B. Xiao, X. Deng, G. G. Y. Lim, S. Xie, Z. D. Zhou, K. L. Lim and E. K. Tan, Cell Death Dis., 2017, 8, e3097.
17. G. Jin, N. Guo, Y. Liu, L. Zhang, L. Chen, T. Dong, W. Liu, X. Zhang, Y. Jiang, G. Lv, F. Zhao, W. Liu, Z. Hei, Y. Yang and J. Ou, Theranostics, 2023, 13, 4802–4820.
18. H. Wang, L. Wen, H. Wei, Y. Song, W. Song, M. Li, X. Lan, W. Cai and D. Jiang, J. Nanobiotechnol., 2024, 22, 394.
19. B. Bai, S. Qi, K. Yang, X. Yu, R. Jian, T. Zhang, D. Wang, H. Meng, Y. Zhao, Y. Xia, H. Xu, G. Yu and Z. Chen, Small, 2023, 19, e2300217.
20. Z. Yi, M. Deng, M. J. Scott, G. Fu, P. A. Loughran, Z. Lei, S. Li, P. Sun, C. Yang, W. Li, H. Xu, F. Huang and T. R. Billiar, Hepatology, 2020, 72, 1394–1411.
21. H. Liu and K. Man, Int. J. Mol. Sci., 2021, 22, 8210.
22. F. D'Amico, A. Vitale, D. Piovan, A. Bertacco, R. Ramirez Morales, A. Chiara Frigo, D. Bassi, P. Bonsignore, E. Gringeri, M. Valmasoni, G. Garbo, E. Lodo, F. E. D'Amico, M. Scopelliti, A. Carraro, M. Gambato, A. Brolese, G. Zanus, D. Neri and U. Cillo, Liver Transplant., 2013, 19, 135–144.
23. C. Mao, X. Liu, Y. Zhang, G. Lei, Y. Yan, H. Lee, P. Koppula, S. Wu, L. Zhuang, B. Fang, M. V. Poyurovsky, K. Olszewski and B. Gan, Nature, 2021, 593, 586–590.
24. B. Yan, Y. Ai, Q. Sun, Y. Ma, Y. Cao, J. Wang, Z. Zhang and X. Wang, Mol. Cell, 2021, 81, 355–369 e310.
25. S. Willenborg, D. E. Sanin, A. Jais, X. Ding, T. Ulas, J. Nuchel, M. Popovic, T. MacVicar, T. Langer, J. L. Schultze, A. Gerbaulet, A. Roers, E. J. Pearce, J. C. Bruning, A. Trifunovic and S. A. Eming, Cell Metab., 2021, 33, 2398–2414 e2399.
26. Z. M. Tang, Y. H. Ju, X. C. Dai, N. Ni, Y. Liu, D. D. Zhang, H. Q. Gao, H. Sun, J. Zhang and P. Gu, Redox. Biol., 2021, 43, 101971.
27. Y. Xu, Y. Liu, K. Li, D. Yuan, S. Yang, L. Zhou, Y. Zhao, S. Miao, C. Lv and J. Zhao, Mol. Neurobiol., 2022, 59, 1619–1631.
28. Y. Li, J. Wang, Y. Li, J. Luo, F. Liu, T. Chen, Y. Ji, H. Yang, Z. Wang and Y. Zhao, ACS Nano, 2023, 17, 12127–12139.

---

Footnote

† These authors contributed equally to this work.

---