Nature-inspired confined cascade enzyme nanoreactors for targeted atherosclerosis therapy

Abstract

Oxidative stress and excessive inflammatory responses are major drivers of atherosclerosis (AS) formation and progression. In this study, we report a nature-inspired nanoreactor (named USPB@SeDMSN@NM) with superoxide dismutase (SOD), catalase (CAT), and glutathione peroxidase (GPx) multienzymatic activities for targeted atherosclerosis therapy. The confined cascade nanocatalyst (USPB@SeDMSN) is designed by assembling ultrasmall Prussian blue nanoparticles (USPB NPs, SOD and CAT mimics) into the large pores of selenium (Se)-doped dendritic mesoporous silica nanoparticles (SeDMSNs, GPx mimics). The DMSN provides abundant immobilization sites for USPB NPs and Se to sequentially scavenge reactive oxygen species (ROS) in a cascade manner and forms confined reaction environments to significantly increase the local concentration of substrates and overall catalysis efficiency, which is inspired by multienzyme complexes (MECs) in nature. The neutrophil membrane was coated onto USPB@SeDMSN to endow the nanoplatforms with the ability to target atherosclerotic plaques. In vitro and in vivo results demonstrated that this nature-inspired enzyme cascade nanoreactor efficiently mitigated inflammation in macrophages and endothelial cells by scavenging various radicals and inhibited foam cell formation by reducing lipid accumulation in macrophages. Moreover, it has significant antiaging effects, protecting DNA from oxidative damage and slowing the onset of cell senescence. By conducting confined SOD-CAT/GPx cascade reactions for high-efficiency plaque microenvironment modulation, the USPB@SeDMSN@NM nanoreactor offers a powerful modality for targeted atherosclerosis therapy. This work highlights the potential of this biomimetic confined nanoreactor with cascaded multienzyme-like activities as an antioxidant and antisenescence agent for atherosclerosis treatment.

Introduction

Cardiovascular disease, of which atherosclerosis (AS) is the most prevalent form, has emerged as the leading cause of morbidity and disability worldwide.^1,2 Although its pathogenesis is not completely understood, atherosclerosis is generally considered a chronic inflammatory disease that is caused mainly by lipid metabolism disorders and vascular dysfunctions.³ In an atherosclerotic plaque microenvironment with a high concentration of reactive oxygen species (ROS),^2,4,5 oxidized low-density lipoprotein (oxLDL) accumulates in arterial walls, which can promote the recruitment of macrophages as well as the formation of extracellular lipid cores caused by excessive oxidative cholesterol.⁶ Senescent endothelial cells (ECs), smooth muscle cells (SMCs), and macrophages can secrete a variety of inflammatory cytokines, proteases, and ROS, which further accelerate atherosclerosis progression.^2,7 These findings highlight the critical role of reducing oxidative stress and cellular senescence in the treatment of atherosclerosis.^8,9 Notably, animal studies have suggested that both SOD and GPx play protective roles against atherogenesis.^10–12 However, the use of natural enzymes in antioxidant biomedical applications is limited because of their high cost and low stability.^13,14 Nanozymes have been exploited as alternatives for mimicking enzyme-like functions.^15–18 Among existing nanozymes, Prussian blue nanoparticles (PB NPs), especially USPB NPs with excellent multienzyme-like activities (SOD, CAT, and POD), are considered promising therapeutic agents,^19–22 as demonstrated in our previous studies.^22–28 Selenium (Se) plays essential roles in a range of biological processes as a unique essential trace element, and its crucial functions in antioxidant defense and redox regulation have received extensive attention.^29,30 Se-based nanomaterials with GPx-like activity are gaining importance in the field of medicine owing to their antioxidant and anticancer properties.³¹

In a healthy vascular microenvironment, a network of antioxidant defense enzymes, including SOD, CAT, and GPx, works cooperatively to eliminate excessive oxidative stress.^10,32 These natural enzymes with complementary functions are often densely organized within a close microenvironment, forming multienzyme complexes (MECs).³³ Many examples of such MECs can be found in nature. For example, the multienzyme α₂β₂ complex, composed of subunits α and β, was discovered to catalyze the last two steps in the biosynthesis of L-tryptophan.³⁴ These biocatalytic reactions are often confined within a single cellular compartment. The confined environment facilitates efficient interenzyme communication and cascaded catalytic activity, allowing the toxic intermediates generated by one enzyme to be directed to a complementary enzyme as a substrate, thereby enhancing catalytic efficiency and selectivity compared with those of nonorganized enzymes.^35–37 Therefore, to mimic the functionality of MECs, enabling the proximity of the active sites on different enzymes in colocalization is critical. Notably, a large amount of research has been conducted using mesoporous silica nanoparticles (MSNs) as excellent catalyst supports.³⁸ The porous structure of MSNs provides a separate confined microenvironment for the operation of two- and three-enzyme cascades.^39–41 In particular, dendritic mesoporous silica nanoparticles (DMSNs) are expected to become a new type of catalyst carrier with high loading capacity and excellent mass transfer efficiency because of their large pore size and uniform monodispersed nanoparticle structure, which is difficult to block and enhances catalytic performance, including activity, stability, and selectivity.^42–45

An ideal treatment for established atherosclerotic plaques should present the following characteristics.⁴⁶ This approach could enhance targeted drug delivery and achieve satisfactory therapeutic efficacy.⁴⁷ Recently, biomimetic cell membrane-derived drug delivery platforms, which possess highly complex biological functions while preserving the inherent physicochemical properties of the nanoparticle core, have been readily applicable in nanoscale biomedicine.^48,49 These cell membrane-camouflaged nanoparticles provide a unique means to overcome delivery barriers and target specific disease sites.^47,50 Notably, neutrophil membrane (NM)-derived nanovesicles or neutrophil membrane-coated NPs have been developed to treat acute or chronic inflammation (such as atherosclerosis).^51–58 Neutrophils and their mediators are implicated in the pathogenesis of atherosclerosis.⁵⁹ Inspired by the above, a delivery system derived from neutrophil membranes coated with confined cascade nanozymes might offer a therapeutic option for the efficient treatment of atherosclerosis.

Herein, we design a nature-inspired multienzyme nanoreactor with antioxidant and antisenescence abilities as a therapeutic agent for the treatment of atherosclerosis. Considering that GPx is a key antioxidant enzyme that plays a critical role in reducing cardiovascular risk, zero-valence Se was incorporated into DMSNs via a reduction-based method (named SeDMSN). After optimization, the USPB NPs were loaded within the large pores of SeDMSN (designated USPB@SeDMSN). This design was inspired by the catalytic mechanisms of MECs in biological systems.³⁶ Finally, the USPB@SeDMSN was coated with a neutrophil membrane (named USPB@SeDMSN@NM) to serve as a biomimetic drug delivery system for targeting inflammatory plaques (Scheme 1). The large-pore (~25 nm) DMSN architecture provides abundant immobilization sites for USPB NPs and Se, facilitating the sequential scavenging of ROS, including ^•OH, ^•O₂⁻ and H₂O_2, in a cascading manner. The structure forms a confined reaction environment that significantly enhances the local concentration of substrates and overall catalysis efficiency. Moreover, the NM-mimicking confined cascade nanoreactor can effectively target inflammatory atherosclerotic lesions, modulating the inflammatory microenvironment and inhibiting oxLDL-induced foam cell formation. Overall, the designed USPB@SeDMSN@NM demonstrates significant efficacy in halting atherosclerosis progression in animal models.

Scheme 1.

Illustration of NM-coated confined cascade nanozymes (USPB@SeDMSN@NM) for targeted atherosclerosis therapy. a Schematic illustration of the synthesis of the USPB@SeDMSN@NM nanoreactor. b The ROS-scavenging mechanism of the SOD-CAT/GPx synergistic system and the design of our bioinspired nanoreactor that mimics the natural antioxidant enzyme system. c Its application as an effective antioxidant and anti-senescence agent for the targeted treatment of atherosclerosis. (i) Neutrophil membrane coating facilitates specific accumulation in plaques via adhesion molecules (e.g., ICAM-1 and integrin β2). (ii) The nanoreactor exhibits integrated SOD/CAT/GPx-like activities, eliminating ROS (H₂O₂, •OH, •O₂⁻) and generating O₂. (iii) It promotes macrophage polarization from the M1 to the M2 phenotype, aiding in the clearance of inflammatory factors (IL-6, IL-1β, and TNF-α). (iv) The biomimetic coating improves internalization by target cells. (v) It protects against DNA damage and downregulates the expression of senescence-related genes (p53 and p21). Leveraging its spatially confined design, the nanoreactor exhibited superior multienzyme activities that effectively inhibited the progression of atherosclerosis. NM neutrophil membrane, ROS reactive oxygen species, EC endothelial cell, ICAM-1 intercellular adhesion molecule-1, LDL low-density lipoprotein, ox-LDL oxidized-low-density lipoprotein, CAT catalase, SOD superoxide dismutase, GPx glutathione peroxidase, ROS reactive oxygen species, O₂ oxygen, HIF-1α hypoxia-inducible factor 1-alpha, H₂O₂ hydrogen peroxide, •O₂⁻ superoxide anion radicals, •OH hydroxyl radical, IL-6 interleukin-6, IL-1β interleukin-1β, TNF-α tumor necrosis factor-alpha, p53 cellular tumor antigen p53, p21 cyclin-dependent kinase inhibitor 1

Results

Synthesis and characterization of USPB@SeDMSN@NM

The synthesis process of USPB@SeDMSN@NM is shown in Scheme 1. To confirm the structure of the USPB@SeDMSN@NM nanoreactor, transmission electron microscopy (TEM) images (Fig. 1a) of the as-synthesized DMSN and SeDMSN both revealed a typical dendritic structure and spherical morphology with a size of ~70 nm. Studies have demonstrated that nanoparticles within the optimal size range of ~50–70 nm exhibit the highest internalization efficiency.^60–62 Moreover, the dendritic porous structure provides large mesopores as reservoirs for the encapsulation of guest catalysts such as USPB nanozymes. The ~1 nm USPB NPs (Fig. 1a) were successively collected and accumulated into the mesopores of DMSN or SeDMSN, effectively forming USPB@DMSN (Fig. 1a) and USPB@SeDMSN (Fig. 1a) (the USPB indicated by white arrows), respectively. Prior to membrane coating, the Fe and Se contents (wt%) in USPB@SeDMSN, as determined via inductively coupled plasma‒optical emission spectrometry (ICP‒OES), were calculated to be 16.63% and 38.73%, respectively. To synthesize USPB@SeDMSN@NM, neutrophils were isolated from the bone marrow of C57BL/6 mice via the Percoll gradient method. Flow cytometric analysis revealed that the percentage of extracted cells coexpressing CD11b and Ly6G was ~89.0%, indicating that the optimal Percoll gradient consisted of 80% (v/v), 70% (v/v), and 55% (v/v) Percoll solution (Supplementary Fig. 1a, Supplementary Table 1). After being coated with NM, a USPB@SeDMSN@NM nanoreactor (Fig. 1a) was fabricated. Since the thickness of an NM is typically 5–10 nm,⁶³ the TEM image shown in Fig. 1b, USPB@SeDMSN@NM revealed that USPB@SeDMSN was coated with an outer shell of ~9 nm upon negative staining, suggesting that the neutrophil membrane was capped on the surface of the USPB@SeDMSN. USPB@SeDMSN coated with a liposome membrane (USPB@SeDMSN@Lip) was fabricated as the control nanoparticle for USPB@SeDMSN@NM (Supplementary Fig. 1b). The liposomes used for coating consisted of DOPC, cholesterol, and DSPE-PEG2000 at a mass ratio of 10:3:0.5.

Fig. 1.

Characterization of the nanoreactor. a TEM images of DMSN, SeDMSN, USPB, USPB@DMSN, USPB@SeDMSN, and USPB@SeDMSN@NM. b TEM images of USPB@SeDMSN, NM, and USPB@SeDMSN@NM stained with 1% phosphotungstic acid solution. c EDS line image showing the distributions of Si, O, Se, Fe, N, and P. d BF, HAADF, and element mapping images of USPB@SeDMSN@NM. Scale bar: 50 nm. e SDS‒PAGE image of proteins in neutrophils, NMs, USPB@SeDMSNs, and USPB@SeDMSN@NM. f XRD patterns of USPB, DMSN, SeDMSN, SeNPs, and USPB@SeDMSN. g UV−visible spectra of USPB and USPB@SeDMSN in deionized water. h Hydrodynamic diameter and i zeta potential of various nanomaterials. All the experiments were repeated three times (n = 3), and the data are presented as the means ± SD. Measurements of j POD-like, k CAT-like, l SOD-like, and m GPx-like activities of USPB, SeDMSN, and USPB@SeDMSN@NM. The GPx-like activities of the nanozymes were evaluated via a GPx detection kit. n TEM images of USPB@SeDMSN@NM after culture in neutral (pH = 7.4) and acidic (pH = 6.0) SBFs for various durations: 0, 3, 5, and 9 days. Scale bar: 50 nm. o The in vitro cumulative release curves of Si-based biodegradation products after USPB@SeDMSN@NM biodegraded into SBFs at different pH values were determined via ICP‒OES

To determine the composition of the encapsulated catalysts, we performed various characterization methods, including scanning transmission electron microscopy (STEM) (Fig. 1c, d and Supplementary Fig. 1c, d), SDS‒PAGE with Coomassie Brilliant Blue staining (Fig. 1e), X-ray powder diffraction (XRD) (Fig. 1f), scanning electron microscopy (SEM) (Supplementary Fig. 1e), confocal laser scanning microscopy (CLSM) (Supplementary Fig. 1f) and ultraviolet‒visible spectrophotometry (UV‒VIS) (Fig. 1g and Supplementary Fig. 1g). Elemental mapping of USPB@DMSNs revealed homogeneous distributions of Si, O, Fe and N, illustrating the successful and dispersive loading of USPB NPs in USPB@DMSNs (Supplementary Fig. 1c). The coexistence of Si, O, N, Se, Fe and P signals in the energy dispersive spectrometer (EDS) spectra of USPB@SeDMSN@NM (Fig. 1c) confirmed the dispersive loading of USPB NPs and Se. As shown in Fig. 1d, in the bright-field (BF) image, the high-Z USPB NPs correspond to relatively dark areas of USPB@SeDMSN@NM. Conversely, in the high-angle annular dark-field (HAADF) image, they correspond to bright areas. Furthermore, the corresponding elemental mappings revealed uniform distributions of Si, O, Se, Fe, N and P throughout the entire structure of USPB@SeDMSN@NM (Fig. 1d). Given that the cell membrane is composed of a phospholipid bilayer, EDS (Fig. 1c) and elemental mapping (Fig. 1d) demonstrated the distribution of phosphorus (P) on the surface of USPB@SeDMSN@NM, confirming the successful coating of the neutrophil membrane. Like USPB@SeDMSN@NM, the element map shown in Supplementary Fig. 1d reveals the distribution of P on the surface of USPB@SeDMSN@Lip, confirming the presence of the phospholipid coating. SEM‒EDS results confirmed the elemental composition of Se in SeDMSN and both Se and Fe in USPB@SeDMSN (Supplementary Fig. 1e). The SDS‒PAGE results suggested that our approach is able to translocate the majority of neutrophil membrane proteins to the surface of USPB@SeDMSN@NM, as shown in Fig. 1e. Further XRD measurements confirmed that the prepared USPB, SeDMSN, and USPB@SeDMSN had diffraction peaks corresponding to the PB crystal (JCPDS #73--0687), zero-valence Se crystal (JCPDS #06--0362), and both crystals, respectively, indicating their successful preparation (Fig. 1f). The fluorescence images are shown in Supplementary Fig. 1f. revealed that the FITC-labeled SeDMSN colocalized with the Dil-labeled NM, further confirming that the NM coating was on the surface of the USPB@SeDMSN. Prior to UV‒VIS measurement, the Fe concentration in both USPB and USPB@SeDMSN was adjusted to 10 µg/mL on the basis of inductively coupled plasma‒mass spectrometry (ICP‒MS) analysis. The UV‒VIS results revealed that USPB@SeDMSN retained near-infrared (NIR) absorption at 693 nm, which was identical to that of USPB (Fig. 1g), indicating that the SeDMSN coating did not interfere with the characteristic absorption peaks of USPB. Therefore, we quantified the Fe content in USPB@SeDMSNs via a standard curve established from UV‒VIS measurements of USPB NPs (Supplementary Fig. 1g). Taken together, these results demonstrated the successful synthesis of the USPB@SeDMSN@NM nanoreactor.

The USPB nanozymes were synthesized in an ethanol/water mixture in the presence of polyvinylpyrrolidone (PVP) to ensure colloidal stability and prevent aggregation of ultrasmall particles, as previously reported.²³ Dynamic light scattering (DLS) examination of the PVP-stabilized dispersion revealed an average hydrodynamic diameter of 32.58 ± 2.6 nm, with a polydispersity index (PDI) of 0.237 ± 0.02 (Fig. 1h and Supplementary Table 2). This value represents the apparent size of the dynamic nanocomplexes formed by the USPB cores and the stabilizing PVP molecules in solution. In contrast, transmission electron microscopy (TEM) imaging, which reveals the inorganic core under dry conditions, confirmed the ultrasmall size of the individual USPB crystallites at ~1 nm (Fig. 1a). Compared with that of the uncoated USPB@SeDMSN (140.67 ± 4.6 nm), the hydrodynamic size of USPB@SeDMSN@NM increased mildly to 158.53 ± 3.8 nm, as shown in Fig. 1h and Supplementary Table 2. The low PDI values for all formulations (Supplementary Table 2) indicate that they all had narrow size distributions and good colloidal stability. The USPB loading shifted the zeta potential of SeDMSN from a positive value (+27.67 ± 0.2 mV) to a negative value (−24.33 ± 0.7 mV). Furthermore, zeta potential measurements revealed that the surface charge of USPB@SeDMSN (−24.33 ± 0.7 mV) shifted to −20.90 ± 0.6 mV after NM coating. This value is closer to that of NM alone (−17.20 ± 0.2 mV), indicating effective shielding by the surface coating material (Fig. 1i, Supplementary Table 2). Similarly, after coating with phospholipids, the surface charge of USPB@SeDMSN (−24.33 ± 0.7 mV) changed to −18.80 ± 1.9 mV, suggesting the successful synthesis of USPB@SeDMSN@Lip (Supplementary Table 2). The N₂ adsorption‒desorption curve (Supplementary Fig. 1h) revealed that the DMSNs had a surface area of 768.13 m²/g, implying the presence of large mesopores. The Barrett–Joyner–Halenda (BJH) pore size calculated from the adsorption branch is 25.06 nm for the DMSNs, making them suitable supports for ultrasmall nanozymes. As shown in Supplementary Fig. 1i, the diameter of USPB@SeDMSN@NM mildly increased with decreasing pH after 24 h of storage in different media, indicating its colloidal stability under these conditions. The particle size and zeta potential of USPB@SeDMSN@NM were also stable for 7 days in water and 72 h in serum-containing medium (Supplementary Fig. 1j–m). These results indicated the stability of USPB@SeDMSN@NM.

Multienzymatic activity of USPB@SeDMSN@NM

Inspired by natural cascade reactions, our design seeks to implement a cascade multienzyme system encompassing SOD, CAT, and GPx activities by confining USPB NPs and Se within the pores of DMSN (Scheme 1). For a comprehensive investigation of catalytic activities, the multienzymatic activities of USPB@SeDMSN@NM (POD, SOD, CAT, and GPx mimic activities) were evaluated. First, the POD-like activity was evaluated (acid conditions) by recording the changes in the absorption of the substrate 3,3,5,5′-tetramethylbenzidine (TMB) at 650 nm in the presence of H₂O₂. The increase in absorbance at 650 nm over time suggested that the POD-like activity of USPB@SeDMSN@NM (Fig. 1j) or USPB@DMSN@NM (Supplementary Fig. 2a) was greater than that of USPB NPs. The typical Michaelis–Menten steady-state kinetics were investigated at room temperature, and the resulting curves followed the Michaelis−Menten equation (Supplementary Fig. 2b–e). On the basis of the calculated kinetic parameters (Supplementary Table 3), the K_cat value (0.128 s⁻¹) of USPB@DMSN@NM was slightly greater than that of USPB NPs (0.121 s⁻¹) with TMB as a substrate, indicating that USPB@DMSN@NM had greater POD-like activity. The K_m value of USPB@DMSN@NM (0.159 mM) when TMB was used as a substrate was less than that of USPB (0.165 mM), suggesting that USPB@DMSN@NM had a greater affinity for TMB than did USPB NPs. Notably, no change in absorbance at 650 nm was observed in the acidic TMB solution following the addition of SeDMSN or DMSN plus H₂O₂, indicating that both SeDMSN and DMSN lack POD-like enzymatic activity (Fig. 1j and Supplementary Fig. 2a). All of these results show that loading ultrasmall nanozymes (USPB NPs) into the confined architecture of nanoreactors (SeDMSN and DMSN) leads to enhanced POD-like catalytic activity. Second, CAT-like activity was evaluated by monitoring the O₂ generation rate. As shown in Fig. 1k and Supplementary Fig. 2f, g, both USPB NPs and USPB@SeDMSN@NM catalyze the decomposition of H₂O₂ into O₂ and H₂O under neutral conditions, whereas the behavior of SeDMSN or DMSN in catalyzing H₂O₂ was similar to that of the control group (water was used as a control). Additionally, USPB@SeDMSN@NM, USPB@DMSN, and USPB@DMSN@NM all presented better CAT-like catalytic activity than did USPB, indicating that the porous support can improve the surface availability and enzymatic catalytic activity of USPB nanozymes.⁶⁴ Due to the unique physicochemical properties and catalytic activities of these nanozymes, we applied a standardized protocol to define the active units of the USPB-based nanozymes and characterize their catalytic kinetics. These results are shown in Supplementary Table 4. Third, the SOD-like activity was investigated via electron spin resonance (ESR). As depicted in Fig. 1l and Supplementary Fig. 2h, all of these formulations, except the inert material DMSN, eliminate most of the •O₂⁻, indicating their outstanding SOD-mimicking activity. The ESR spectra suggested that the ROS-eradicating activity of the integrated cascade nanozymes (USPB@SeDMSN@NM) was greater than that of USPB or SeDMSN alone, indicating that the integrated cascade nanozymes have antioxidant properties for effective ROS clearance (Fig. 1l). The greater SOD-like activity of USPB@DMSN than USPB and its similarity to USPB@DMSN@NM indicate that the porous structure enhances the ROS-scavenging ability of USPB without being strongly affected by the subsequent NM coating (Supplementary Fig. 2h). Fourth, GPx-like activity was evaluated via the NADPH method. As shown in Fig. 1m and Supplementary Fig. 2i, the absorbance of NADPH at 340 nm indicated that only the nanozymes containing Se exhibited the GPx-like activities among these three nanozymes. This finding is consistent with the role of Se as the redox center of GPx mimics. Furthermore, surface coating with the NM of USPB@SeDMSN@NM did not adversely impact the intrinsic enzymatic activities of SeDMSN. To further measure the stability of the catalytic activities of the formulation during the course of systemic delivery, the POD and CAT-like activities of USPB@SeDMSN@NM were detected after 24 h of storage in different media. As shown in Supplementary Fig. 2j and k, no significant changes were observed in the enzyme-like activities (POD and CAT) of USPB@SeDMSN@NM after 24 h of storage in physiological (PBS, pH 7.4) or slightly acidic (PBS, pH 6.8) conditions compared with those after storage in water.

Biodegradation behavior and biosafety of USPB@SeDMSN@NM

To confirm the clinical translational potential of USPB@SeDMSN@NM, degradability and biosafety assessments were carried out. The dissolution experiment of USPB@SeDMSN@NMs was performed in neutral (pH = 7.4) and acidic (pH = 6.0) simulated body fluid (SBF) media, which were used to mimic in vivo neutral healthy body fluids and an inflamed, slightly acidic environment. The structural degradation of the silica framework was assessed via TEM (Fig. 1n), and the corresponding Si content in the degradation product was analyzed via ICP‒OES (Fig. 1o). TEM images revealed that the radial pore structure and the spherical particle shape were mildly altered after shaking at 37 °C in both neutral and acidic media. Throughout the initial 72-h period, the spherical morphology of USPB@SeDMSN@NM remained intact under both physiological conditions. Accordingly, the system has the ability to prevent premature payload release and avoid structural collapse in vivo within a single dosing interval. When the dissolution time period was prolonged to 9 days, the silica framework showed a high degree of biodegradation, and the accumulated degraded Si content reached 18.2% and 12.4% in neutral and acidic media, respectively. Notably, USPB@SeDMSN@NM exhibited a slower rate of biodegradation in acidic environments than in neutral environments over a 15-day period. These results indicated that the confined catalytic volume of USPB@SeDMSN@NM constructed from porous silica can be biodegraded slowly under both mildly acidic plaque inflammatory conditions and neutral conditions, and that the shrinkage and collapse of the Si-based framework may help the renal clearance of silica-based nanoparticles in vivo and reduce their biological toxicity.⁶⁵

Furthermore, the biosafety of USPB@SeDMSN@NM was investigated in vitro and in vivo. We evaluated the cytotoxicity of USPB@SeDMSN@NM in both endothelial cells and macrophages via a CCK-8 kit (Supplementary Fig. 3). The results showed that the USPB@SeDMSN and USPB@SeDMSN@NM groups presented better cell viability than did the USPB treatment group for both RAW264.7 cells and HUVECs, indicating that the support of porous DMSNs prevents the potential cytotoxicity of ultrasmall-sized nanozymes. A hemolysis assay was also carried out to evaluate the biocompatibility of USPB@SeDMSN@NM, and the results indicated no obvious hemolysis in any of the four groups when the RBCs were exposed, demonstrating the favorable biocompatibility of USPB@SeDMSN@NM (Supplementary Fig. 4). As shown in Supplementary Table 5, complete blood count (CBC) and blood chemistry analyses were further performed to evaluate the biosafety of various formulations at the same Fe concentration (0.05, 0.5 and 5 mg/kg Fe per mouse, equal to 0.3, 3 and 30 mg/kg SeDMSN) in vivo. All hematology parameters were within normal ranges, and no significant differences were detected between the low-dose groups (0.05 and 0.5 mg/kg) and the control groups. Notably, USPB@SeDMSN@NM had fewer adverse effects on CBC and blood chemistry than did its USPB counterpart when the Fe dose was increased to 5 mg/kg, particularly in terms of WBC counts, platelet counts, and AST levels. These findings suggest the enhancement of biosafety by the use of porous SeDMSNs to support ultrasmall-sized nanozymes. Collectively, these results demonstrated the excellent biocompatibility of USPB@SeDMSN@NM.

Neutrophil membrane-coated USPB@SeDMSNs enhance the cellular uptake efficiency of activated endothelial cells and macrophages

Inflammatory activation of endothelial cells orchestrates the recruitment of different subsets of circulating leukocytes, particularly monocytes and neutrophils, into the vascular wall.^59,66 Circulating neutrophils, once recruited, contribute to the progression of atherosclerosis and plaque rupture in patients.⁵⁴ Hence, to improve the targeting efficiency of USPB@SeDMSNs to atherosclerotic plaques, neutrophil membranes were adopted. To promote the expression of key adhesion molecules (integrin β2 and LFA-1) on neutrophils, the obtained cells were incubated with lipopolysaccharide (LPS) (a stimulant that activates neutrophils) before NM isolation. In addition, Western blot (WB) analyses confirmed that the key surface proteins integrin β2 and LFA-1 of USPB@SeDMSN@NM were both present and enriched, which could bestow inflammatory endothelium-targeting properties on the NPs (Fig. 2a). To investigate the ability of NM-coated nanoparticles to target inflammation in vitro, the red fluorescent dye Dil was applied as a fluorescence probe to assess cellular uptake under different circumstances. Confocal laser scanning microscopy (CLSM) revealed that, compared with nontargeting USPB@SeDMSN@Lip/Dil or free Dil, USPB@SeDMSN@NM/Dil nanoparticles exhibited increased cellular uptake efficacy for both TNF-α-stimulated HUVECs (Fig. 2b and Supplementary Fig. 5a) and LPS-activated RAW264.7 cells (Supplementary Fig. 5b). In addition, the corresponding flow cytometry analysis (FCA) was also used to quantitatively measure the cellular uptake efficiency, and the results (Supplementary Fig. 5c, d) were consistent with those obtained from the confocal fluorescence images shown in Fig. 2b and Supplementary Fig. 5a. As shown in Supplementary Fig. 6, after different incubation times with TNF-α-stimulated HUVECs (Supplementary Fig. 6a, b) or LPS-activated RAW264.7 cells (Supplementary Fig. 6c, d), Dil@SeDMSN@NM exhibited much stronger red Dil dye fluorescence than did the naked Dil@SeDMSN or free Dil, confirming that NM enhanced nanoparticle internalization in inflammatory cells. In contrast, in RAW264.7 cells (Supplementary Fig. 6e, f) without LPS treatment, the red fluorescence signal in the Dil@SeDMSN group was mildly stronger than that in the Dil@SeDMSN@NM group, indicating that the NM coating could protect the NPs from phagocytosis and clearance by macrophages. This is plausible because positively charged Dil@SeDMSN(with an amino group) enhances cellular uptake and internalization, whereas the Dil@SeDMSN@NM has a negative surface charge, which probably slightly reduces passive cellular uptake.^67,68 Similar results are also presented in Supplementary Fig. 7 and Supplementary Fig. 8. This enhanced uptake is likely due to specific interactions between NM-coated NPs and activated endothelial and macrophages through adhesion molecules, such as ICAM-1, which mediate adhesion and are prominently expressed on dysfunctional endothelial cells and LPS-stimulated macrophages.^58,69 Therefore, to clarify the mechanism of NPs targeting in an inflammatory environment, the increased expression of ICAM-1 in TNF-α-stimulated HUVECs (Fig. 2c, d) and LPS-activated RAW264.7 cells (Supplementary Fig. 9) was verified via WB. These results suggest that the neutrophil-membrane coating enhances the ability of NPs to target inflammation-activated endothelial cells and macrophages, which is consistent with previous reports.^58,66,69,70

Fig. 2.

Exploration of the cellular uptake efficiency, ROS scavenging ability, macrophage polarization, anti-inflammatory and antisenescence activities of USPB@SeDMSN@NM in vitro. a Western blots of integrin β2 and LFA-1 on USPB@SeDMSN, neutrophils, NMs, and USPB@SeDMSN@NM at the same protein amount loading. b In vitro binding to activated endothelial cells. Confocal microscopy images of USPB@SeDMSN@NM/Dil compared with those of USPB@SeDMSN@Lip/Dil and free Dil after 4 h of incubation. HUVECs were pretreated with TNF-α (25 ng/mL) for 6 h. Scale bar, 20 μm. c ICAM-1 expression in normal HUVECs and inflamed HUVECs induced by TNF-α. d Quantitative analysis of ICAM-1 protein expression from c normalized to that of GAPDH (n = 3). e Fluorescence microscopy was used to assess the attenuation of the inflammatory response in LPS-induced macrophages treated with various formulations, including USPB, SeDMSN, USPB@SeDMSN, and USPB@SeDMSN@NM. Green fluorescence indicates the presence of ROS, as detected by the DCFH-DA probe. f Quantitative analysis of the average optical density of DCF (green fluorescence) in RAW264.7 cells. The results are expressed as the means ± SD (n = 4). The related typical inflammatory factors g IL-6, h IL-1β, and i TNF-α were detected via an ELISA kit. The results are presented as the means ± SD (n = 3). j The mRNA levels of IL-10 after treatment with LPS alone or in combination with different formulations for 24 h (n = 3). All the data were normalized to that of β-actin. Quantitative analysis of k HIF-1α, l iNOS, m Arg1, n CD86, and o CD206 protein expression from panel (p), normalized to that of β-actin (n = 3). p Hypoxia-inducible factor (HIF-1α) and macrophage polarization-related proteins (M1: iNOS, CD86; M2: Arg1, CD206) were measured via western blotting in RAW264.7 cells treated with LPS or LPS plus different nanozymes for 24 h. q Representative optical microscopy images of the attenuation effects of USPB@SeDMSN@NM on the formation of foam cells induced by both LPS and oxLDL. Scale bar: 200 μm. r DNA damage in HUVECs was assessed by determining the colocalization of γ-H2AX (green) and DAPI (blue) across different treatment groups. Scale bar: 200 μm. s Quantification of Oil Red O-stained area size in RAW264.7 cells, as calculated from panel (q); n = 4. t Quantitative analysis of the average optical density of γ-H2AX, derived from (r). The data are presented as the means ± SD (n = 4). u Quantitative results of β-gal (blue) expression in HUVECs under different treatments, as shown in (v). The data are expressed as the means ± SD (n = 4). v Antisenescence activity of USPB@SeDMSN@NM verified by SA-β-gal staining. The model of cellular senescence was induced by 100 μM H₂O_2. Scale bar: 200 μm. Statistical significance was determined by one-way ANOVA and two-tailed Student’s t test: *p < 0.05, **p < 0.01, ***p < 0.001

ROS scavenging, macrophage polarization and anti-inflammatory effects of USPB@SeDMSN@NM

Despite strong evidence that oxidative stress contributes to atherogenesis and endothelial dysfunction, effective antioxidant therapies are lacking.^5,71 In light of this, we designed a cascade-based artificial multienzyme system (USPB@SeDMSN@NM) for targeted delivery of antioxidants to atherosclerotic plaques. First, we investigated the ability of USPB@SeDMSN@NMs to scavenge ROS in RAW264.7 macrophages and HUVECs. Macrophages are stimulated by LPS to induce an inflammatory phenotype, which is characterized by ROS overproduction and the release of inflammatory cytokines. HUVECs treated with TNF-α alone were used to mimic inflamed endothelial cells. HUVECs treated with TNF-α plus H₂O₂ were aimed at inducing a state of cellular senescence. We prestimulated HUVECs with TNF-α to induce a state of chronic inflammation and cellular activation, which upregulated the expression of adhesion molecules (such as ICAM-1) and other pathways relevant to nanozyme uptake. Following this priming step, we subsequently treated the cells with H₂O₂ to acutely induce a high level of oxidative stress, thereby creating a more pathologically relevant model that encompasses both inflammatory and oxidative elements.

The ROS probe DCFH-DA was used to evaluate ROS levels following different treatments. Fluorescence microscopy imaging confirmed a significant reduction in the green fluorescence intensity across all four nanozymes (USPB, SeDMSN, USPB@SeDMSN, USPB@SeDMSN@Lip, and USPB@SeDMSN@NM), whereas USPB@SeDMSN@NM had the most pronounced effect (Fig. 3e, f and Supplementary Fig. 10 b, e, g–j). Similar results were obtained when the ^•O₂⁻ scavenging ability of these four nanozymes in RAW264.7 cells was detected with the DHE probe (Supplementary Fig. 10a, f). Flow cytometry-based quantitative analysis revealed that USPB@SeDMSN@NM significantly reduced the fluorescence intensity in RAW264.7 cells treated with LPS (Supplementary Fig. 11a, b). Similar measurements were also conducted in HUVECs. As presented in Supplementary Fig. 10 b, c, g, h, and Supplementary Fig. 11, the results confirmed that TNF-α prestimulation alone (Supplementary Fig. 10b, g and Supplementary Fig. 11c, d) was sufficient to induce a significant increase in intracellular ROS, although the level was indeed lower than that achieved with the combined TNF-α and H₂O₂ (Supplementary Fig. 10 c, h, Supplementary Fig. 11e, f). As expected, USPB@SeDMSN@NM exhibited remarkable concentration-dependent ROS-scavenging effects in vitro. As presented in Supplementary Fig. 10e, j, the confined cascade nanozyme system exhibited significantly greater ROS-scavenging capacity than did the physically mixed nanozymes, highlighting the crucial role of nanoconfinement in improving the therapeutic efficacy of nanozyme-based platforms. These results demonstrated that, compared with other formulations, USPB@SeDMSN@NM significantly reduced ROS levels in inflammatory macrophages and activated endothelial cells, likely due to its cascaded multienzymatic antioxidant activities and enhanced cellular uptake efficiency.

Fig. 3.

Pharmacokinetics, biodistribution, and plaque-targeting efficacy of the multifunctional nanoplatform following intravenous injection in APOE^−/− mice. a Blood clearance curves of various formulations, as determined by Fe quantification via ICP‒MS analysis, reflecting their respective circulation persistence. Biodistribution of Fe (b) and Se (c) in mouse tissues following USPB@SeDMSN@NM administration, as measured by ICP–OES. The data are presented as the means ± SD (n = 3). d Representative ex vivo images and e quantitative analysis illustrating the accumulation of DiD fluorescent signals in the aorta. f Representative ex vivo fluorescence images of major organs (heart, liver, spleen, lung, and kidney) and g their corresponding quantitative data at 24 h postinjection of DiD, USPB@SeDMSN@Lip/DiD, or USPB@SeDMSN@NM/DiD. h Immunofluorescence analysis of the colocalization of DiD, USPB@SeDMSN@Lip/DiD, or USPB@SeDMSN@NM/DiD (DiD, red) with macrophages (CD68⁺, green) in cryosections of the aortic root. Scale bars: 100 μm

The accumulation of proinflammatory cytokines, such as interleukin-6 (IL-6), interleukin-1β (IL-1β), and tumor necrosis factor-α (TNF-α), at inflammatory sites is a key factor in plaque progression.¹ To further assess the anti-inflammatory ability of USPB@SeDMSN@NM, an enzyme-linked immunosorbent assay (ELISA) and quantitative reverse transcription polymerase chain reaction (qRT‒PCR) were performed to measure the levels of these three typical proinflammatory cytokines (IL-6, IL-1β and TNF-α) in LPS-stimulated RAW264.7 macrophages (Fig. 2g–i, Supplementary Fig. 12a–c). Treatment with various nanozyme formulations significantly reduced the release of these proinflammatory cytokines, with USPB@SeDMSN@NM showing the most pronounced effect. In addition, USPB@SeDMSN@NM markedly upregulated the mRNA levels of the anti-inflammatory cytokine interleukin-10 (IL-10) (Fig. 2j) while inhibiting the expression of transforming growth factor-beta (TGF-β) (Supplementary Fig. 12d). These findings indicate that USPB@SeDMSN@NM substantially decreases the expression of proinflammatory cytokines and alleviates inflammation.

To elucidate how USPB@SeDMSN@NM exerts its anti-inflammatory effects, we further investigated its impact on macrophage polarization. Thus, we performed multiple assays, including WB analysis (Fig. 2k–p, Supplementary Fig. 12i–n), qRT‒PCR (Supplementary Fig. 12e–h), and immunofluorescence (IF) staining (Supplementary Figs. 13 and 14), in LPS-stimulated RAW 264.7 cells following different formulation treatments. As shown in Fig. 2k–p and Supplementary Fig. 12i–n, USPB@SeDMSN@NM promoted macrophage polarization from the M1 phenotype to the M2 phenotype more effectively than the other formulations did, as evidenced by greater downregulation of iNOS and CD86 (M1 markers) and more pronounced upregulation of Arg1 and CD206 (M2 markers). As shown in Supplementary Fig. 12i–k, compared with the USPB@SeDMSN@Lip control, USPB@SeDMSN@NM induced a significantly greater ratio of M2 phenotype (anti-inflammatory, CD206) to M1 phenotype (proinflammatory, iNOS) macrophages. This result indicates that the neutrophil membrane coating not only serves as an inert cloak but also actively enhances the anti-inflammatory efficacy of the nanocatalytic core (USPB@SeDMSN) in vitro. We hypothesize that the neutrophil membrane confers improved targeting to inflammatory sites and facilitates more effective interactions with macrophages, leading to more potent reprogramming of the inflamed microenvironment of plaques. As shown in Supplementary Fig. 12l and m, the ability of the integrated nanozymes (USPB@SeDMSN and USPB@SeDMSN@NM) to polarize LPS-stimulated macrophages toward the M2 anti-inflammatory phenotype was dramatically greater than that of the physical mixture (USPB + SeDMSN), highlighting the definitive advantage of spatial confinement in nanozyme-based anti-inflammatory strategies. We also confirmed that, compared with other formulations, USPB@SeDMSN@NM more effectively reduced the expression of hypoxia-inducible factor 1-alpha (HIF-1α) (a key molecular sensor and mediator of the cellular hypoxia response), thereby sensing changes in intracellular oxygen levels and modulating the transcription of genes associated with macrophage polarization. Similar results were obtained from the corresponding qRT‒PCR analysis (Supplementary Fig. 12e–h) and IF staining of CLSM images (Supplementary Figs. 13 and 14).

USPB@SeDMSN@NM mitigates foam cell formation and endothelial senescence

Under an oxidative stress microenvironment at the site of atherosclerotic lesions, low-density lipoprotein (LDL) is oxidized into oxidized low-density lipoprotein (oxLDL) by ROS.^6,72 In addition, monocytes are recruited to the vessel wall and transform into macrophages, which are capable of recognizing and engulfing oxLDL.² Subsequently, these macrophages transform into foam cells, which are major hallmarks of atherosclerosis and initiate the pathology of atherogenesis. Mitigating foam cell formation through decreasing the cellular uptake of oxLDL is a profitable way to treat atherosclerosis. Moreover, recent studies revealed that a lack of functional GPxs accelerates atherosclerosis, increases oxLDL-induced foam cell formation, and leads to increased proliferation of peritoneal macrophages in ApoE^−/− mice.^72,73 Therefore, the ability of USPB@SeDMSN@NM to inhibit lipid deposition in macrophages was investigated. To this end, we employed a model group in which macrophages were stimulated with LPS and treated with oxLDL or Dil-labeled oxLDL (Dil-oxLDL) without further treatment. The cells treated with fresh medium only were used as the control group. Oil Red O staining revealed an attenuation of foam cell formation by treatment with USPB@SeDMSN@NM in RAW264.7 cells (Fig. 2q, s) compared with that in the model group. The cumulative USPB@SeDMSN or USPB@SeDMSN@NM can be clearly observed in the image (blue area), further demonstrating that loading USPB into a confined SeDMSN support enhances the cellular uptake efficiency. Flow cytometry analysis was used to quantify the Dil-oxLDL fluorescence intensity in RAW264.7 cells after various treatments (Supplementary Fig. 15a–c). Following treatment with USPB, SeDMSN, USPB@SeDMSN, or USPB@SeDMSN@NM, the red fluorescent signals of intracellular Dil-oxLDL decreased to 41.56%, 69.9%, 29.38% and 17.23% of those in the model group (100%) (Supplementary Fig. 15c), respectively. The fluorescence imaging results obtained by CLSM were consistent with those of flow cytometry (Supplementary Fig. 15d). Taken together, these results indicate that USPB@SeDMSN@NM treatment of macrophages enhances lipid scavenging and suppresses foam cell formation.

Several studies have indicated that senescent endothelial cells accumulate prior to the onset of atherosclerosis, which is associated with attenuated endothelial function.⁷⁴ Increased oxidative stress induces dysfunction of DNA repair mechanisms and contributes to the senescence of endothelial cells.⁷⁵ Namrata Singh et al. reported that multiple enzymes (CAT, SOD, and GPx) mimic Mn₃O₄ nanoparticles and have the ability to protect biomolecules from ROS-mediated DNA damage.⁷⁶ Bijun Zhu et al. prepared MIL-47(V)-F (MVF), a GPx-mimicking nanozyme that protects against DNA damage by attenuating the oxidative state of the cell.⁷⁷ Hence, the potential of USPB@SeDMSN@NM to promote DNA repair and alleviate cellular senescence in the context of atherosclerosis warrants investigation. DNA damage was assessed by phosphorylation of H2A histone family member X (γ-H2AX), a key marker of DNA damage and repair that occurs when double-stranded DNA breaks.⁷⁸ Next, we investigated whether USPB@SeDMSN@NM attenuates the senescence of HUVECs by modulating DNA repair mechanisms. Immunofluorescence staining revealed a similar phenomenon, with increased green fluorescence in the TNF-α plus H₂O₂-treated groups (the model for endothelial senescence), whereas the fluorescence in the USPB@SeDMSN@NM-treated groups decreased (Fig. 2r, t). Furthermore, senescence-associated β-galactosidase (SA-β-gal) staining revealed that, compared with the other groups, the USPB@SeDMSN@NM group exhibited the least amount of cell senescence. HUVECs treated with TNF-α plus H₂O₂ (Fig. 2u, v) or LPS (Supplementary Fig. 16) were used as models of endothelial senescence. To gain a deeper understanding of the mechanism by which USPB@SeDMSN@NM mitigates endothelial senescence, we assessed the mRNA levels of the senescence-associated genes p21 and p53 in senescent HUVECs via qRT‒PCR. All the treatment groups presented significant decreases in p53 and p21 mRNA expression (Supplementary Fig. 16c, d). Notably, senescent HUVECs treated with USPB@SeDMSN@NM presented significantly lower levels of both p53 (Supplementary Fig. 16c) and p21 mRNAs (Supplementary Fig. 16d) than those treated with USPB, SeDMSN, or USPB@SeDMSN. Interestingly, no significant difference in p53 mRNA expression was observed between the USPB-treated group and the model group (Supplementary Fig. 16c), suggesting that SeDMSNs contribute more substantially to the anti-senescence effects than does the USPB. This highlights the crucial role of selenium (Se) as the catalytic center of GPx mimicry in combating cellular senescence. Our results demonstrate that USPB@SeDMSN@NM effectively alleviates cellular senescence by suppressing p21 and p53 expression, with the cascade nanozyme design and membrane coating synergistically enhancing its protective function. These results suggest that nanozymes, particularly USPB@SeDMSN@NM, exhibit potent antioxidant activity that protects endothelial cells from ROS-induced DNA damage and cell senescence, which is consistent with findings from other studies.^75,79,80

Pharmacokinetics, biodistribution and plaque-targeting efficacy of neutrophil membrane-coated nanozymes

The pharmacokinetic profile of USPB@SeDMSN@NM was systematically evaluated in vivo via a validated ICP‒MS assay.²² As shown in Fig. 3a, the plasma half-life (t₁/₂) of USPB@SeDMSN@NM reached 14.69 h, which was significantly greater than that of USPB@SeDMSN (6.62 h) and USPB (1.7 h). This notable extension in circulation time can be attributed to the neutrophil membrane (NM) coating, which confers stealth properties by reducing immune recognition and macrophage uptake, thereby enhancing the blood residency of the nanozymes (NZs).⁵⁴ Furthermore, we investigated the biodistribution of iron (Fe) and selenium (Se), the fundamental structural elements of USPB@SeDMSN@NM, across major organs, including the heart, liver, spleen, lung, and kidney. Quantitative analysis via inductively coupled plasma‒optical emission spectrometry (ICP‒OES) indicated that both Fe and Se were predominantly distributed in the reticuloendothelial system organs, especially the spleen and liver (Fig. 3b, c). Importantly, the accumulation levels in these organs were not significantly elevated compared with those in the control groups, suggesting minimal nonspecific retention. This pattern implies that the nanozyme does not exhibit a strong tendency to accumulate in vital organs, which underscores its favorable biosafety profile.

Compelling evidence suggests that neutrophils can be recruited to the lesion site and release neutrophil extracellular traps that exacerbate atherosclerosis.⁸¹ To assess the in vivo targeting ability of neutrophil membrane-coated NZs, a synthetic dye, DiD, with near-infrared emission (excitation/emission of 645/665 nm), was employed to prepare USPB@SeDMSN@Lip/DiD and USPB@SeDMSN@NM/DiD. Six-week-old male ApoE^−/−mice fed a high-fat diet (HFD) for 12 weeks were used for plaque-targeting assessment. As shown in Fig. 3d, e, at 24 h after i.v. injection of different formulations (free DiD, USPB@SeDMSN@Lip/DiD and USPB@SeDMSN@NM/DiD) with the same dosage of DiD (2 mg/kg), significantly greater fluorescence was detected in the aorta tissues isolated from the mice injected with USPB@SeDMSN@NM/DiD than in those from the free DiD- and nontargeting USPB@SeDMSN@Lip/DiD-treated groups. Furthermore, the distributions of these formulations in the main organs (heart, liver, spleen, lung, and kidney) were also evaluated (Fig. 3f, g). Fluorescence imaging revealed that the majority of the fluorescence signals of all three formulations (free DiD, USPB@SeDMSN@Lip/DiD, and USPB@SeDMSN@NM/DiD) were detected in the liver at 24 h postinjection, whereas the USPB@SeDMSN@NM/DiD-treated groups presented significantly lower fluorescence intensities than did the USPB@SeDMSN@Lip/DiD NP-treated group. Additionally, immunofluorescence analysis of atherosclerotic plaques via anti-CD68 staining revealed greater accumulation of USPB@SeDMSN@NM/DiD in the plaques than USPB@SeDMSN@Lip/DiD or free DiD (Fig. 3h). This observation directly confirmed that the neutrophil membrane coating on the nanoparticle greatly enhances the targeting of the nanoparticles to atherosclerotic plaques in vivo. Similar experiments conducted in ApoE^−/− mice after the administration of free DiD, DiD@SeDMSN, or DiD@SeDMSN@NM yielded results consistent with those described above (Supplementary Fig. 17).

USPB@SeDMSN@NM attenuates atherosclerotic plaque progression in ApoE^−/− mice

After confirming the effectiveness of USPB@SeDMSN@NM in vitro, we next assessed its therapeutic potential in ApoE^−/− mice. ApoE^−/− mice were fed an HFD for 12 weeks and randomly divided into five groups (n = 5 per group). At the end of the 12-week HFD feeding period, each group was intravenously administered saline, USPB, SeDMSN, USPB@SeDMSN, or USPB@SeDMSN@NM at doses of 0.5 mg/kg Fe or 1.16 mg/kg Se every three days for an additional 7 weeks (Fig. 4a). At the end of the treatment, the entire aorta was collected and stained with Oil Red O (ORO). En face ORO staining analyses (Fig. 4b), along with quantification of the lesion area in the aorta (Fig. 4h) and the total lesion area (Fig. 4i), clearly demonstrated the therapeutic efficacy of the various formulations in inhibiting atherosclerosis progression. Notably, the control group (saline injection) presented the largest ORO-positive areas, accounting for approximately 40.8% of the total aortic tissue area. In contrast, treatment with USPB, SeDMSN, USPB@SeDMSN, or USPB@SeDMSN@NM resulted in varying degrees of plaque reduction, with average plaque areas of ~31.1%, 31.5%, 24.3%, and ~14.2% of the total aortic tissue area, respectively. Compared with the other groups, the USPB@SeDMSN@NM treatment group demonstrated the highest therapeutic efficacy, with significant differences, indicating effective inhibition of atherosclerosis progression.

Fig. 4.

Therapeutic efficacy of i.v.-delivered USPB@SeDMSN@NM in atherosclerotic mice. a Schematic of the experimental treatment protocol for ApoE^−/− mice. b Photographs of face-to-face ORO-stained aorta tissues from ApoE^−/− mice treated with saline or different formulations. c Representative in vivo images of the aortic arch and ex vivo sections of the aortic root. Sections were stained with d H&E, e anti-CD68, f Masson’s trichrome, and g anti-MMP-9. h Quantitative analysis of the percentage of total lesion area and i the total lesion area of whole aortas from (b). j–m Quantitative analysis of the stained aortic root sections shown in (d–g) for j total lesion area (H&E), k macrophage content (CD68), l collagen content (Masson’s trichrome), and m MMP-9 expression. The data are presented as the means ± SD (n = 5 for h, i; n = 4 for j–m). Scale bar, 500 μm (applies to d–g). Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001

In addition, en face analyses of the entire aortic arch (Fig. 4c) and histochemical examination (Fig. 4d–m) were conducted to further investigate the composition of atherosclerotic plaques. En face images of the aortic arch clearly demonstrated the effectiveness of the different treatments in reducing atherosclerotic plaque size, with USPB@SeDMSN@NM treatment showing the highest efficacy (Fig. 4c). Furthermore, hematoxylin/eosin (H&E) staining (Fig. 4d, j) of the aortic root revealed a significantly reduced necrotic core area in the USPB@SeDMSN@NM-treated group compared with the saline and other treatment groups. Immunofluorescence staining of CD68 (Fig. 4e, k) was employed to evaluate macrophage infiltration. The collagen content in the fibrous cap of atherosclerotic plaques plays a critical role in maintaining plaque stability. Reliable evidence suggests that atheromatous lesions are prone to disruption and subsequent thrombosis in the absence of sufficient collagen.⁸² Masson’s trichrome staining was conducted on sections of the aortic root to assess the degree of plaque stability (Fig. 4f, l). The significantly higher collagen (indicated by the blue color) concentration around the plaques in the mice indicated that USPB@SeDMSN@NM stabilized the plaques from further development and inhibited atherogenesis. Notably, matrix metalloproteinase-9 (MMP-9), a hydrolase secreted by foam cells derived from macrophages, is positively related to plaque vulnerability. Immunohistochemical analyses of matrix metalloproteinase-9 (MMP-9) revealed that USPB@SeDMSN@NM NPs dramatically decreased the number of macrophages and MMP-9 expression in plaques of the aortic root (Fig. 4g, m). Compared with the control group, which presented prominent MMP-9-positive staining (21.0 ± 3.5%), the USPB and SeDMSN groups presented similar reductions (13.9 ± 3.8%, P < 0.05, and 13.5 ± 1.1%, P < 0.01, respectively). The USPB@SeDMSN group presented a further reduction in MMP-9 staining (13.5 ± 1.1%, P < 0.01). MMP-9-positive staining in aortic sinus sections was notably low in the USPB@SeDMSN@NM group (6.4 ± 1.4%, P < 0.001), which effectively stabilized atherosclerotic plaques. These results, along with the decreased necrotic core and macrophage content in plaques, demonstrate that intravenous treatment with USPB@SeDMSN@NM effectively stabilizes atherosclerotic plaques.

USPB@SeDMSN@NM ameliorates oxidative stress and cell senescence in atherosclerotic plaques

To further investigate the underlying mechanism responsible for the in vivo anti-atherosclerotic effects, DHE staining was conducted on sections of the aortic root isolated from ApoE^-/- mice to evaluate their ROS levels (Fig. 5a, c). The intensity of the red fluorescence in the control group was the highest among all the groups (19.2 ± 1.3%), suggesting a high level of ROS production in these aorta tissues. Consistent with the in vitro studies, the treatment groups exhibited varying degrees of ROS scavenging. Specifically, the USPB, SeDMSN, and USPB@SeDMSN groups presented reduced fluorescence intensity (13.6 ± 0.7%, P < 0.001; 13.2 ± 1.1%, P < 0.001; and 9.5 ± 1.1%, P < 0.001, respectively). The USPB@SeDMSN@NM group presented a dramatic decrease in DHE-positive staining (7.9 ± 0.9%, P < 0.001).

Fig. 5.

Amelioration of plaque formation by USPB@SeDMSN@NM via ROS Scavenging and attenuation of cellular senescence. Representative images of aortic root sections stained with a a DHE fluorescent probe (red) and b an antibody against γ-H2AX (green) and DAPI for nuclei (blue). Quantitative analysis of c DHE- and d γ-H2AX-stained aortic root sections from an atherosclerotic mouse model after treatment with various formulations. The results are presented as the means ± SD (n = 4). Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001

In addition, senescent cells in vascular walls are causally linked to dysfunction of endothelia, contributing to the progression of atherosclerosis.² After the different formulations were administered, immunohistochemistry analyses of γ-H2AX were conducted to investigate their anti-senescence ability in atherosclerotic mice (Fig. 5b, d). Compared with the other groups, the USPB@SeDMSN@NM group presented significantly lower γ-H2AX staining, which was consistent with the reduced oxidative stress observed via DHE staining. Collectively, these findings demonstrate that USPB@SeDMSN@NM effectively attenuates systemic oxidative stress and reduces the number of senescent cells in plaques. To determine the biosafety of different formulations in ApoE^−/− mice after long-term treatment, the body weights of all the treated groups of mice were monitored during the treatment period. As shown in Supplementary Fig. 18a, b, none of the mice exhibited significant weight loss. Furthermore, no significant differences were observed in the net body weight change before and after treatment among the different formulation groups, except for the USPB@SeDMSN group (Supplementary Fig. 18b). These results indicate that the NM coating plays a critical role in enhancing biosafety. Additionally, H&E staining analysis of the major organs (heart, liver, spleen, lung, and kidney) of the mice in the various treatment groups was also performed. There was no significant difference in the histological morphology of these tissues following any of the treatments, which further confirmed their biocompatibility (Supplementary Fig. 18c). Additionally, the levels of total cholesterol (CHO), low-density lipoprotein (LDL), triglyceride (TG) and high-density lipoprotein (HDL) remained within normal ranges after long-term treatment with different formulations (Supplementary Fig. 19). Overall, these preliminary findings indicated that USPB@SeDMSN@NM displayed excellent biocompatibility during long-term treatment, demonstrating its promising potential as a therapeutic candidate for atherosclerosis.

Discussion

Atherosclerosis is fundamentally a chronic inflammatory disorder initiated by oxidative stress and propagated through lipid-driven pathways.^1,2 Consequently, drug development for atherosclerotic cardiovascular disease (ASCVD) has focused largely on two therapeutic strategies: lipid-lowering and anti-inflammatory interventions. Large-scale clinical trials involving statin therapy have demonstrated that even when low-density lipoprotein cholesterol (LDL-C) is effectively reduced to target levels, a considerable residual risk of cardiovascular events remains, with persistent inflammation identified as a key contributing factor.⁸³ The promising outcomes of anti-inflammatory agents observed in several major clinical trials have thus renewed interest in this therapeutic approach. Recent studies have indicated that low-dose small-molecule anti-inflammatory drugs, such as aspirin⁸⁴ and methotrexate⁸⁵ can inhibit inflammatory mediators and attenuate oxidative stress, thereby exerting anti-atherosclerotic effects. However, the clinical utility of these conventional agents is often limited by drawbacks, including poor solubility, low bioavailability, a narrow therapeutic window, high dosing requirements, and potential systemic side effects. In this context, nanomedicine has emerged as a promising platform to overcome these limitations. Although recent nanobased therapies have shown potential in targeting specific inflammatory pathways, the development of an integrated strategy capable of simultaneously addressing multiple proinflammatory mechanisms within atherosclerotic plaques remains an ongoing challenge.

Given the critical role of reactive oxygen species (ROS) in accelerating atherosclerosis within the complex inflammatory microenvironment, therapeutic strategies aimed at scavenging intraplaque ROS, thereby suppressing foam cell formation and mitigating endothelial cell senescence, represent a promising approach to attenuate disease progression. Currently, antioxidant therapeutics based on natural enzymes or small molecules have demonstrated considerable potential in preclinical studies; however, their clinical translation has been largely unsatisfactory. This challenge stems primarily from inherent pharmacokinetic limitations: natural enzymes are susceptible to proteolytic degradation, exhibit short half-lives, and struggle to accumulate at therapeutically effective concentrations in diseased tissues, whereas small molecules often suffer from poor bioavailability and rapid systemic clearance. Nanoplatforms offer a promising strategy to overcome these limitations by protecting therapeutic cargo from degradation during circulation and enabling targeted delivery to pathological sites through rational design. Nevertheless, significant hurdles remain: natural enzymes face challenges, including high production costs, nonreusability in catalytic reactions, and structural instability under physiological conditions, whereas many small-molecule antioxidants present substantial formulation difficulties owing to their poor aqueous solubility. Therefore, nanozymes with robust antioxidant capabilities constitute a novel therapeutic modality.^86,87 Prussian blue, an FDA-approved antidote for the clinical treatment of radioactive cesium and thallium poisoning, exhibits excellent biosafety. Prussian blue-based nanozymes have been demonstrated to possess potent antioxidant and anti-inflammatory activities. Importantly, we further demonstrated that the enzymatic activity and anti-inflammatory efficacy of Prussian blue nanoparticles are highly size dependent, with ultrasmall particles (USPB, ~3.4 nm) exhibiting superior performance compared with their larger counterparts (e.g., 60 and 170 nm).⁸⁸ Separately, selenium nanozymes (Se) have also shown promise in ameliorating atherosclerosis through their integrated antioxidant and anti-senescence functions.⁸⁹ However, the therapeutic efficacy of a single nanozyme system is often limited. To address this, we designed a spatially confined cascade nanoreactor (USPB@SeDMSN) through coencapsulation of USPB and Se nanozymes within a mesoporous silica framework, mimicking the architecture of natural enzyme complexes. Nanoconfinement has emerged as an innovative strategy for enhancing the performance and functionality of nanomaterials in diverse fields, including electrocatalysis, energy storage, and biomedicine. This approach involves the physical restriction of molecules, ions, or particles within nanoscale spaces. Such spatial confinement facilitates the penetration of external molecules into nanoconfined domains through well-defined pore channels, enabling unique modulations of chemical properties, such as concentrating reactive species, accelerating reaction kinetics, improving selectivity, and enhancing van der Waals interactions. Furthermore, nanoconfinement can alter the physicochemical properties of metal-based nanoparticles, influencing reaction thermodynamics and even modifying fundamental reaction mechanisms. These distinctive characteristics position nanoconfinement as a promising solution to longstanding challenges in reactivity, selectivity, stability, and mass transport encountered in both nanotechnology and conventional industrial processes.⁹⁰ On the basis of these principles, we hypothesized that the confined environment within USPB@SeDMSN could emulate natural catalytic systems and generate synergistic effects. Specifically, the ordered mesoporous structure of DMSN provides a high specific surface area and abundant active sites, creating an ideal microenvironment for catalytic reactions. Spatial confinement ensures uniform dispersion and stabilization of catalytic nanoparticles, preventing aggregation and enhancing durability. Moreover, this configuration can fine-tune the electronic structure of the active centers and their local chemical environments through coordination effects, thereby increasing the catalytic activity and enabling more efficient and multifaceted therapeutic outcomes.^91,92 In comparative evaluations, the confined cascade nanozyme system demonstrated superior anti-inflammatory activity and macrophage polarization capability relative to physically mixed nanozymes, underscoring the critical role of spatial confinement in enhancing the biological efficacy of nanozyme-based therapeutics. This approach for constructing artificial multienzyme complexes is gaining traction in nanomedicine, as it offers enhanced atom economy and reaction efficiency, which are crucial for effective in vivo applications.³⁷ The observed attenuation of atherosclerosis in our model suggests that this cooperative strategy successfully translates into improved biological efficacy.

Despite the considerable promise of nanomedicine in atherosclerosis therapy, its clinical advancement is significantly hampered by the unresolved challenge of targeted delivery. The field currently faces two major translational bottlenecks: the rapid clearance of nanoparticles by the mononuclear phagocyte system and their generally poor efficiency in crossing relevant biological barriers. These limitations collectively result in suboptimal drug accumulation at the atherosclerotic site, undermining therapeutic efficacy. In response, research efforts have increasingly shifted toward bioinspired nanocarrier designs. Among these methods, the cell membrane coating strategy has gained prominence as a powerful method to enhance biocompatibility and impart complex biological functions. By preserving the source cell surface antigenic profile, these biomimetic systems effectively evade immune detection, thereby addressing a key limitation of conventional nanocarriers.⁹³ A particularly compelling direction within this bioinspired paradigm involves leveraging the innate homing abilities of specific cell types. Successful nanomedicines should maintain a good balance between the stealth effect and interaction with diseased tissue.⁹⁴ In the context of atherosclerosis, which is fundamentally driven by inflammation, the roles of neutrophils and neutrophil extracellular traps (NETs) have been extensively implicated in disease progression and plaque destabilization.⁹⁵ In atherosclerotic plaques, activated macrophages release chemokines to recruit neutrophils to inflammatory sites and prolong neutrophil survival by secreting granulocyte‒macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF), and tumor necrosis factor-α (TNF-α). Neutrophils subsequently traverse the endothelial layer through enhanced endothelial contractility and the upregulation of intercellular adhesion molecule-1 (ICAM-1). Within atherosclerotic lesions, neutrophils are prone to apoptosis, releasing a range of peptides that promote the migration of monocytes/macrophages into the lesion area to phagocytose apoptotic cells.⁵⁹ Neutrophils influence monocyte differentiation and macrophage polarization processes. Under inflammatory conditions, the anti-inflammatory and reparative functions of macrophages that have phagocytosed neutrophils have garnered significant attention in recent years. Through exposure to annexin A1, TLR2/4-MyD88 signaling pathway-activated NADPH oxidase, and neutrophil gelatinase-associated lipocalin (NGAL), neutrophils can be phagocytosed by macrophages, acting as “messengers” that actively guide macrophages toward a reparative M2 phenotype, thereby facilitating inflammation resolution and tissue repair.⁵⁷ Very recently, Wu J. et al. elucidated a “phagocytosis-like” effector mechanism in a spinal cord injury (SCI) inflammation model, demonstrating that macrophage internalization of neutrophil membrane vesicles (NMVs) promotes glutamine metabolism to replenish the TCA cycle intermediate α-KG, thereby increasing oxidative phosphorylation, suppressing NF-κB signaling, and reprogramming inflammatory macrophages into a pro-regenerative phenotype.⁹⁶ This understanding has logically led to the hypothesis that neutrophil-mimetic nanocarriers could exploit inflammatory chemotaxis for site-specific targeting. The ability to concentrate therapeutic agents precisely within the inflammatory plaque microenvironment represents a major breakthrough, potentially reducing systemic side effects and lowering the requisite therapeutic dose. Furthermore, the anti-inflammatory effect mediated by neutrophil membrane-induced M2 polarization of macrophages may further promote inflammation resolution in atherosclerosis. Our in vitro and in vivo experiments demonstrated that neutrophil membrane-coated nanozymes can precisely target atherosclerotic plaques while extending systemic circulation through biomimetic “stealth” properties. Notably, we observed that compared with its liposome-coated counterpart (USPB@SeDMSN@Lip), USPB@SeDMSN@NM exhibited superior efficacy in promoting M2 polarization. This intriguing finding suggests that the neutrophil membrane coating may function beyond mere targeting and potentially actively participates in modulating inflammatory responses. We hypothesize that this enhanced immunomodulatory capability operates through dual mechanisms: first, macrophage recognition of neutrophil membrane components such as annexin A1 may initiate proresolving signaling through receptors such as formyl peptide receptor 2 (FPR2), directly promoting M2 polarization; second, the membrane-coated nanoparticles may mimic apoptotic cells, triggering efferocytosis-mediated anti-inflammatory programming in macrophages. This “active signaling” approach works synergistically with the ROS-scavenging capability of the nanoreactor core: while the neutrophil membrane provides positive signals for M2 polarization, ROS elimination negatively regulates proinflammatory signaling, which impedes M2 transition. This dual-pronged strategy enables optimal phenotypic remodeling of macrophages, representing a sophisticated biomimetic approach to inflammatory regulation. Therefore, the development of a biomimetic delivery platform that combines the inflammatory tropism of neutrophils with the therapeutic action of anti-inflammatory and anti-senescence agents constitutes a rational and promising strategy. This approach aligns with the growing emphasis in the field of creating targeted therapies that are not only effective but also exhibit improved safety profiles for clinical translation.

In summary, this study demonstrates the first application of large-pore (~25 nm) and small-sized (~70 nm) dendritic mesoporous silica nanoparticles (DMSNs) as a confined biomimetic space for the codelivery of ultrasmall Prussian blue (PB) and selenium (Se) nanozymes. These two nanozymes operate synergistically, enabling efficient cascade catalytic reactions. Compared with single nanozyme systems, the USPB@SeDMSN composite has enhanced enzyme-like activity, a wider range of catalytic functions, and overcomes the common limitations of individual nanozymes, such as single active sites, a low specific surface area, and moderate biocompatibility. The integrated nanozyme system not only contains spatially separated active sites that mimic the activities of peroxidase (POD), superoxide dismutase (SOD), catalase (CAT), and glutathione peroxidase (GPx) but also benefits from a nanoconfinement effect. This effect enhances the overall catalytic efficiency by promoting mass transfer and reducing the diffusion distances between active centers, as validated by enzyme kinetics analysis. In vitro assays confirmed that the neutrophil membrane-coated version (USPB@SeDMSN@NM) possesses outstanding catalytic performance, multienzyme mimicry, excellent biocompatibility, stability, and good dispersibility in aqueous solution. It effectively eliminates various reactive oxygen species (ROS), thereby mitigating ROS-induced inflammatory responses in macrophages and human umbilical vein endothelial cells (HUVECs). Additionally, it promotes macrophage polarization toward an anti-inflammatory phenotype and suppresses foam cell formation, indicating its potential to modulate key atherosclerotic processes. Moreover, USPB@SeDMSN@NM reduces ROS levels through a cascade mechanism and downregulates the expression of senescence-associated genes such as p53 and p21, thus protecting DNA from oxidative damage and delaying endothelial cell senescence. The neutrophil membrane coating not only facilitates active targeting to inflammatory sites without significantly compromising enzymatic activity but also functions as a biological activator to promote M2 macrophage polarization and anti-inflammatory responses. Using an ApoE^−/− mouse model of atherosclerosis, we verified that USPB@SeDMSN@NM avoids rapid clearance by the mononuclear phagocyte system, extends blood circulation time, and preferentially accumulates in atherosclerotic lesions. The in vivo results further demonstrated its high targeting specificity toward plaques and significant efficacy in slowing atherosclerosis progression. This bioinspired approach provides a viable strategy for designing nanozymes with multiantioxidant capacity and disease-targeting functionality, underscoring the therapeutic potential of cascade nanoreactors in the treatment of atherosclerosis and other inflammatory diseases.

Materials and methods

Materials

Potassium ferricyanide(K₃[Fe(CN)₆]), polyvinylpyrrolidone (PVP, K30), and triethanolamine (TEA) were obtained from Aladdin Chemical Reagent Company (China). Sodium Cetyltrimethylammonium bromide (CTAB), Hydrochloric acid (HCl), hydrogen peroxide (30 wt%), and anhydrous ethanol were purchased from Sinopharm Chemical Reagent Co., Ltd. (China). Tetraethyl orthosilicate (TEOS) and dimethyl sulfoxide (DMSO) were both purchased from Sigma-Aldrich (USA). Sodium heptafluorobutyrate (FC4), Dioleoylphosphatidylcholine (DOPC), and cholesterol (Chol) were purchased from Bide Pharmatech Co., Ltd. (Shanghai, China). 2′,7′-dichlorofluorescin diacetate (DCFH-DA) and Oil Red O were purchased from MedChemExpress. Sodium selenite (Na₂SeO₃, 99%) was acquired from Shanghai Titan Scientific Co., Ltd. DSPE-PEG2000 was purchased from Ponsure Biological Technology Co., Ltd. Lipopolysaccharide (LPS), glutathione (GSH), 3,3′-dioctadecyloxacarbocyanine perchlorate (DiO), 1,1’-dioctadecyl-3,3,3’,3’-tetramethylindocarbocyanine perchlorate(Dil),1’-Dioctadecyl-3,3,3’,3’-Tetramethylindodicarbocy-anine, 4-chlorobenzenesulfonate salt (DiD), cell counting kit-8 (CCK-8), the total GPx assay kit and senescence β-galactosidase staining kits (C0602) were purchased from Beyotime Chemical Reagent Co., Ltd. Human high-oxidized low-density lipoprotein (oxLDL) and Dil-labeled oxLDL (Dil-oxLDL) were purchased from Yiyuan Biotechnologies (China). Dihydroethidium (DHE), ascorbic acid (99%), and Dulbecco’s modified Eagle’s medium (DMEM) were purchased from Nanjing Keygen Biotech. Co., Ltd. (Nanjing, China). Trypsin–EDTA (0.25%) and heat-inactivated fetal bovine serum (FBS) were obtained from Gibco Laboratories (NY, USA). All aqueous solutions were prepared with deionized water (resistivity of 18.25 MΩ cm, Ulupure, China). All chemicals were of analytical grade. The RAW 264.7 and HUVECs cells were obtained from Nanjing Keygen Biotech Co., Ltd. (Nanjing, China).

Mice

Male C57BL/6 (8 weeks old) and male apolipoprotein E-deficient (ApoE^−/−) mice (6–8 weeks old) were both purchased from GemPharmatech Co., Ltd (China). After acclimatization for 7 days, mice were subjected to various experiments. All animal procedures were approved by the Institutional Animal Care and Use Committee of Southeast University (ICAUC-20240531005).

Methods

Synthesis of USPB@SeDMSN

USPB@SeDMSN was synthesized by adding the as-prepared USPB NPs solution (10 mg/mL Fe) into an aqueous suspension of SeDMSN (10 mg/mL). The mixture was stirred at 4 °C for 24 h. The final USPB@SeDMSN products were collected by centrifugation (10,000 rpm, 4 °C, 30 min). The contents of Fe (from USPB NPs) and Se (from SeDMSN) were determined by ICP-OES.

Isolation of mouse bone marrow-derived neutrophils and membrane

Neutrophils were purified from the bone marrow of the tibia and femur of C57BL/6 mice using the Percoll gradient method described by Boxio, R. et. al.⁹⁷ After the tibia and femur were separated, the bone marrow cavity was rinsed with RPMI 1640 medium without serum using a 1 mL syringe until the bone fragments appeared white. The cells were collected by centrifugation at 4000 rpm for 3 min, and the pellet was resuspended in RPMI 1640 medium. Different three-layer Percoll gradients were applied to purify neutrophils, including gradient 1 (80%, 65%, 55%), gradient 2 (90%, 75%, 65%), and gradient 3 (80%, 70%, 55%). As an example, the cell suspension was layered on top of a three-layer Percoll gradient 3 consisting of 80% (v/v), 70% (v/v), and 55% (v/v) Percoll solution and centrifuged at 750 g for 30 min at room temperature. Neutrophils were isolated from the Percoll layer between the 70% (v/v) and 80% (v/v) gradients. Then, cells were co-stained with fluorescein isothiocyanate (FITC) anti-mouse Ly6G antibody and APC anti-mouse CD11b antibody (Biolegend) to identify neutrophil cells. Flow cytometric analysis (Beckman Coulter, USA) was used to detect the neutrophil purity (Ly6G⁺, CD11b⁺). To obtain inflammation-activated neutrophils, the collected cells were cultured in medium supplemented with LPS (10 μg/mL) for 4 h. The activated cells were then washed with PBS and collected by centrifugation.

For isolation of the neutrophil membrane, the above-mentioned activated neutrophils were resuspended in ice-cold hypotonic lysis buffer supplemented with a protease inhibitor cocktail. After being disrupted using an ultrasound cell breaker (100 W for 5 min and rest for 2 min, 3 cycles) in an ice bath, the solution was centrifuged at 3000 rpm for 5 min at 4 °C. The supernatants were saved and centrifuged at 14,000 rpm for 30 min at 4 °C. The NM fragments were washed with PBS and water. After the protein concentration of the obtained NM was measured by bicinchoninic acid (BCA) assay, the NM solution was lyophilized and stored at −80 °C for subsequent studies.

Synthesis of USPB@SeDMSN@NM

To prepare USPB@SeDMSN@NM, 1 mg/mL (protein content) of the above NM solution was mixed with USPB@SeDMSN (1 mg/mL), and sonicated with an ultrasound cell breaker (100 W for 1 min and rest for 2 min, 3 cycles) in an ice bath. For removing any excess NM, the products were centrifuged at 8000 rpm at 4 °C for 30 min.

To coat the USPB@SeDMSN with liposomes (USPB@SeDMSN@Lip), a liposome solution was first prepared using the thin-film hydration method,⁹⁸ with a mass ratio of DOPC, cholesterol, and DSPE-PEG2000 at 10:3:0.5. Then, 200 µL of 1 mg/mL liposome solution was added directly to 200 µL of 1 mg/mL USPB@SeDMSN. Using a procedure similar to that used for NM-coating as described above, USPB@SeDMSN@Lip was fabricated to serve as the control nanoparticles for subsequent experiments.

Biodegradation behavior of USPB@SeDMSN@NM

The degradation experiment was performed on an ICP-OES (Thermo, USA) instrument according to the literature.^99,100 Briefly, 0.1 mg of pristine USPB@SeDMSN@NM samples were suspended in 1 mL of neutral (pH = 7.4) or acidic (pH = 6.0) simulated body fluid (SBF) medium. These aliquots were sealed into 1.5 mL Eppendorf tubes (Eppis). After incubation at 37 °C with shaking at 500 rpm for the specified durations, the samples were centrifuged at 8000 rpm for 30 min. To monitor the dissolution behavior of USPB@SeDMSN@NM in vitro, the solid residues and 0.5 mL of supernatant from each sample were analyzed by TEM and ICP-OES, respectively.

Assessment of POD, CAT, SOD, and GPx-like activity and steady-state kinetic assays

For the POD-like activity assays, various nanomaterials were measured using TMB as the substrate in the presence of H₂O₂ in 0.2 M HAc-NaAc buffer solution. Briefly, 1 mL aliquot of 0.2 M HAc-NaAc buffer (pH 3.6) was supplemented with 20 μL of 100 μg/mL Fe (Fe content) or 233 μg/mL (Se content) Se-based nanozymes in this study, 100 μL of 10 mg/mL TMB solution, and 200 μL of 30% H₂O₂.The absorbance of the color reaction at 650 nm was immediately measured using a UV spectrophotometer (UV-3600, Shimadzu, Japan) for 2 min. Steady-state kinetic experiments were conducted according to a previously reported method.²⁶ Briefly, each 200 μL reaction system consisted of 0.2 M HAc–NaAc buffer (pH 3.6), to which 10 μL of either USPB or USPB@SeDMSN@NM was added as the catalyst, along with H₂O₂ and TMB as substrates. For kinetic studies using TMB as the varying substrate, reactions contained nanozyme (2 μg/mL Fe) and 32 μL of 30% H₂O₂, with TMB concentrations adjusted by adding 0–10 μL of a 10 mg/mL TMB solution (in DMSO). Conversely, when H₂O₂ was the variable substrate, the system contained nanozyme (6 μg/mL Fe) and 10 μL of 10 mg/mL TMB, with H₂O₂ concentration modulated by adding 0–32 μL of 3% H₂O₂. Reaction progress was monitored via time-dependent absorbance measurements (SpectraMax Mini, China). Catalytic parameters were obtained by fitting the initial rate data to the Michaelis–Menten equation: v = v_max/{[S]K_m + [S]}, where v is the initial reaction rate, v_max denotes the maximum reaction rate, [S] represents the substrate concentration, and K_m stands for the Michaelis constant. The constant K_m indicates the substrate concentration at which the reaction velocity reaches half of v_max, and serves as a measure of enzyme–substrate affinity: a lower K_m value corresponds to a higher binding affinity.

The CAT-like activity of the nanomaterials was assessed using a multiparameter analyzer (Multi 3510 IDS, Germany) at room temperature. In general, 0.5 mL of 100 μg/mL Fe (Fe content) or 233 μg/mL (Se content) Se-based nanozymes in this study were added to 15 mL phosphate buffer solution (PBS, 0.2 M, pH 7.4) containing 0.5 mL of 30% H₂O₂. The oxygen solubility (unit: mg/L) generated was measured after different reaction times.

The SOD-like activity of USPB@SeDMSN@NM was measured using electron spin resonance (ESR) spectroscopy (EMX, Bruker, Germany). Typically, various nanomaterials (10 μg/mL Fe) were mixed with 1 mM xanthine, 0.2 U/mL xanthine oxidase, and 50 mM BMPO in PBS (pH 7.4). The mixture was then placed into a capillary tube, and the scavenging ability of •O₂⁻ was recorded using the ESR instrument. The total GPx-like activity of the nanomaterials at a concentration of 3.3 μg/mL (Fe content) Fe- or 7.7 μg/mL (Se content) Se-based nanozymes was determined using the total GPx assay kit (Beyotime, S0058, China).

To evaluate the enzymatic stability of USPB@SeDMSN@NM over a 24-h period, the nanozyme was incubated in PBS (pH 7.4 or 6.8) containing 1% FBS. Subsequently, USPB@SeDMSN@NM were collected by centrifugation at 8000 rpm for 30 min, and its peroxidase (POD)- and catalase (CAT)-like activities were assessed using the methods described above.

SDS-PAGE assay

To determine whether nanovesicles had similar membrane proteins to their parent cells. The cell membrane-associated proteins were evaluated using Sodium dodecyl sulfate–polyacrylamine gel electrophoresis (SDS–PAGE), accompanied by Coomassie Brilliant Blue staining analysis and Western blot. The LPS-stimulated neutrophils, NM, USPB@SeDMSN, and USPB@SeDMSN@NM were subjected to lysis using RIPA lysate (Beyotime, China) containing protease inhibitor cocktail on ice for 40 min. Thereafter, the total protein concentration was measured by the BCA Assay Kit (Thermo, USA). All the samples with equivalent protein (30 μg) were mixed with SDS–PAGE-loading buffer and incubated at 100 °C for 5 min. After short centrifugation, samples were loaded onto a 10% SDS–PAGE gel (Beyotime, China) and run in a Mini-PROTEAN Tetra Cell (BIO-RAD, USA) at 120 V for 2 h. To investigate whether ICAM-1, a ligand of integrin β2, was upregulated after HUVECs or RAW264.7 cells were activated. HUVECs with or without tumor necrosis factor (TNF-α, 25 ng/mL) (ABclonal, China) for 6 h and RAW264.7 cells treated with or without LPS (1 μg/mL) for 24 h were harvested and lysed on ice using RIPA lysate (Beyotime, China) containing protease inhibitor cocktail on ice for 40 min. The subsequent processing steps were similar to those described above. To evaluate the capacity of USPB@SeDMSN@NM for improving the polarization of pro-inflammatory M1 macrophages toward the anti-inflammatory M2 macrophages, RAW264.7 Cells treated with or without LPS (1 μg/mL) for 24 hours, following different treatments for 24 h were harvested and lysed on ice using RIPA lysate (Beyotime, China) containing protease inhibitor cocktail on ice for 40 min. The subsequent processing steps were similar to the above-described.

For Coomassie Brilliant Blue staining analysis, the resultant PAGE gel was stained with Coomassie Blue Fast Staining Solution (Beyotime, China) for 1 h and washed overnight for subsequent imaging. For Western blot analysis, the protein was transferred to Immobilon-P PVDF membrane (Merck Millipore, IPVH00010), which was then blocked with blocking buffer (Beyotime, P0252) for 1 h at room temperature. The blots were probed by antibodies against LFA-1 (Abcam, ab13219), integrin β2 (Cell Signaling Technology, 47598), HIF-1α (Cell Signaling Technology, 36169S), β-actin (Cell Signaling Technology, 4967), Arg1 (Santa Cruz, sc-271430), CD206 (Santa Cruz, sc-58986), ICAM-1 (Santa Cruz, sc-390483), iNOS (Proteintech, 22226-1-AP), CD86 (Proteintech, 83213-5-RR), and GAPDH (Proteintech, 10494-1-AP) overnight at 4 °C and then incubated with the corresponding horseradish peroxidase (HRP)-conjugated anti-rabbit (Proteintech, RGAR001) or anti-mouse IgG (Proteintech, RGAM001). Finally, the blots were incubated in chemiluminescent detection reagents (Tanon) before the final image capture, and the band intensities were analyzed using Image J software.

Cell uptake behavior

To quantify cellular uptake efficiency, RAW264.7 cells and HUVECs were seeded in six-well plates at a density of 2 × 10⁵ cells per well and incubated for 24 h. RAW264.7 cells were incubated with or without 1 μg/mL LPS for 24 h, while HUVEC cells were incubated with or without 25 ng/mL TNF-α for 6 h before incubation with Dil, USPB@SeDMSN@Lip/Dil, and USPB@SeDMSN@NM/Dil (or another group: Dil, Dil@SeDMSN, and Dil@SeDMSN@NM). The cells were incubated for 1, 2, and 4 h with various formulations at an equivalent dose of 0.5 μg/mL of Dil, respectively. Cells without any treatment were used as controls. After incubation, the cells were washed three times with PBS buffer and fixed with paraformaldehyde for 15 min. Following three additional washes with PBS buffer, the cells were collected and analyzed by flow cytometry (FACS, Beckman Coulter, Fullerton, CA, USA). In parallel, cellular uptake behavior was also examined using confocal laser scanning microscopy (LSCM, Olympus, Japan).

Evaluation of ROS scavenging ability and anti-inflammation of USPB@SeDMSN@NM

The intracellular ROS scavenging capacity of USPB@SeDMSN@NM was evaluated using the fluorescent probe DCFH-DA. Briefly, RAW264.7 cells or HUVECs were plated in glass-bottom dishes and cultured for 24 h. RAW264.7 cells were subsequently stimulated with 1 μg/mL LPS for 24 h, while HUVECs were treated with 25 ng/mL TNF-α for 6 h. After incubation with the respective nanozymes, the culture medium was removed, and the cells were rinsed three times with PBS. Additionally, HUVECs were exposed to 1 mM H₂O₂ for 6 h to induce ROS production. The medium was then aspirated, followed by three additional washes with PBS. Cells were incubated with DCFH-DA (1:1000 dilution in PBS) for 40 min at 37 °C. After removing the probe solution, the cells were washed three times with PBS and imaged under a fluorescence microscope (Olympus DX51). HUVECs treated with 25 ng/mL TNF-α only were also conducted. Quantitative assessment of ROS levels was carried out by flow cytometry (Attune NxT, Thermo Fisher Scientific). The same protocol was applied to HUVECs, except for the inclusion of the H₂O₂ stimulation step. Data acquired from flow cytometry were processed using FlowJo V10 software.

To measure the anti-inflammation ability of USPB@SeDMSN@NM, RAW264.7 cells were seeded into 12-well plates at a density of 1 × 10⁵ per well and cultured for 24 h at 37 °C. After being stimulated with 1 μg/mL of LPS for an additional 24 h, cells were incubated with different formulations (5 μg/mL Fe or 11.6 μg/mL Se) for 24 h. Subsequently, the cell supernatants were collected and centrifuged at 12,000 rpm for 10 min to remove redundant nanoparticles and cells. To determine the expression of inflammatory cytokines (TNF-α, IL-6, and IL-1β), the obtained clarified supernatants were analyzed by a commercial ELISA kit (Biosource, Camarillo, CA, USA) according to the manufacturer’s protocol.

Quantitative real-time polymerase chain reaction(qRT-PCR)

To evaluate the transcriptional activity of macrophage polarization-related genes, RAW264.7 cells were treated with LPS and different nanozyme formulations, including USPB, SeDMSN, USPB@SeDMSN, and USPB@SeDMSN@NM. Total RNA was isolated with the Universal RNA Extraction Kit (TaKaRa, 9767) following the manufacturer’s protocol. RNA purity and concentration were determined using a NanoPhotometer (NP80, Implen, Germany). Next, 1000 ng of total RNA was reverse transcribed into cDNA with the PrimeScript™ RT Reagent Kit (Takara, RR047A). Quantitative real-time PCR was carried out using TB Green® Premix Ex Taq™ (Takara, RR420a) on a LightCycler 96 instrument (Roche, Switzerland). Each sample was analyzed in triplicate to ensure statistical reliability, and gene expression levels were normalized to the endogenous control β-actin. Relative mRNA expression was calculated via the 2^−ΔΔCt method. The primer sequences employed in this study are provided in Supplementary Table 6. Together, these procedures enabled accurate quantification of treatment-induced alterations in gene expression within RAW264.7 cells.

Immunofluorescence assay

A standardized immunofluorescence protocol was employed for both assays. Briefly, cells were fixed with 4% formaldehyde for 15 min and permeabilized in 0.3% Triton X-100 for 10 min. After blocking with 5% bovine serum albumin (BSA) for 30 min at 37 °C, the cells were incubated with specific primary antibodies at 4 °C overnight. Following thorough washing, the samples were incubated with fluorophore-conjugated secondary antibodies (FITC or Cy3). Cell nuclei were counterstained with 4’,6-diamidino-2-phenylindole (DAPI, 1 μg/mL). Fluorescent images were captured using an inverted fluorescence microscope (IX73, Olympus, Japan) or a laser-scanning confocal microscope, and the fluorescence intensity was quantified using ImageJ software.

For the assessment of macrophage polarization, RAW264.7 cells were seeded onto glass coverslips in 24-well plates at a density of 1 × 10³ cells per well and cultured overnight. After attachment, the cells were stimulated with LPS (1 μg/mL) for 24 h. The cells were then treated with various nanozyme formulations, including USPB, SeDMSN, USPB@SeDMSN, and USPB@SeDMSN@NM, all standardized to equivalent concentrations of either 5 μg/mL Fe or 11.6 μg/mL Se. At designated time points, cells were processed for immunofluorescence as described above. The M1 polarization state was evaluated using antibodies against iNOS (Proteintech, 22226-1-AP; 1:200) and CD86 (Proteintech, 83213-5-RR; 1:100). The M2 polarization state was assessed using antibodies against Arg1 (Santa Cruz, sc-271430; 1:50) and CD206 (Santa Cruz, sc-58986; 1:50).

To assess the protective effect of USPB@SeDMSN@NM against ROS-induced oxidative stress in HUVECs, cells were seeded in 24-well plates and cultured for 12 h. They were then stimulated with TNF-α for 6 h prior to treatment with different formulations (USPB, SeDMSN, USPB@SeDMSN, and USPB@SeDMSN@NM). After 24 h of treatment, the cells were exposed to H₂O₂ (1 mM) for 6 h. Cells treated with fresh medium only served as the normal control, while those treated with H₂O₂ alone served as the oxidative stress model. DNA damage was evaluated via immunofluorescence staining using a primary antibody against γ-H2AX (AP0099, ABclonal; 1:200).

Inhibition of cellular uptake of oxLDL

To investigate the effect of different treatments on the cellular uptake of oxLDL by macrophages, RAW264.7 cells were cultured on 12-well glass plates at 37 °C with 5% CO₂ overnight. After co-culturing with LPS (1 μg/mL) for 24 h, RAW264.7 cells were incubated with different nanomaterials (5 μg/mL Fe or 11.6 μg/mL Se) for 24 h. The complete medium was then removed and replaced with fresh medium without FBS but containing 40 μg/mL Dil-oxLDL. Cells treated with fresh medium were used as the normal group, and cells cultured with Dil-oxLDL only were used as the model group. After 6 h, RAW264.7 cells were washed, fixed, and harvested. Intracellular fluorescence intensity was quantified by FACS. To obtain fluorescence images, cells were further stained with DiO (10 μM) and DAPI.

Intracellular oxLDL internalization stained by Oil red O

RAW 264.7 cells were incubated in DMEM medium containing 10% FBS for 12 h and stimulated with LPS (1 μg/mL) for 24 h. After treatment with different formulations (5 μg/mL Fe or 11.6 μg/mL Se) for 24 h at 37 °C, the cells were further incubated with fresh medium for 48 h in the presence of 1 μg/mL LPS and 50 μg/mL ox-LDL (Yiyuan Biotechnologies, YB-002) to induce foam cell formation. Cells treated with fresh medium were used as the normal group, while cells treated with LPS and ox-LDL only were used as the model group. The old medium was then removed and replaced with PBS. The cells were washed, fixed with 4% paraformaldehyde for 15 min at room temperature, and stained with filtered 0.3% Oil-Red O (ORO) solution. The cells were gently washed once with 60% isopropyl alcohol and then washed three times with PBS to remove excess ORO. Finally, the cells were observed under an optical microscope.

Effect of USPB@SeDMSN@NM on anti-senescence in HUVECs cells

Senescence β-galactosidase staining kit (C0602, Beyotime) was used to assess the anti-senescence activity of nanozymes. HUVECs were seeded in 24-well plates and incubated for 12 h. Followed by stimulation with TNF-α for 6 h, different formulations were added. To induce senescence, H₂O₂ (1 mM) or LPS (1 μg/mL) was added. After 1 hour, the medium was removed and replaced with fresh medium containing different formulations. Following a 24-h incubation, senescence-associated β-galactosidase (SA-β-gal) staining was performed according to the manufacturer’s instructions. Finally, the blue-colored cells were observed under a microscope.

In vivo pharmacokinetics, distribution and targeting affinity study

Following a previously established protocol,^101–104 the pharmacokinetic profiles of USPB, USPB@SeDMSN, and USPB@SeDMSN@NM were evaluated by measuring blood iron concentrations via inductively coupled plasma mass spectrometry (ICP-MS). Each formulation was administered intravenously to mice via tail vein injection at a standardized Fe dose of 0.5 mg/kg body weight. Blood samples (20 μL) were collected at predetermined time intervals (2, 4, 6, 8, 12, 24, 36, and 48 h), digested in 1 mL of aqua regia (HNO₃:HCl = 1:3, v/v) for 24 hours, and subsequently centrifuged at 6000 rpm for 10 min. The resulting supernatant was passed through a 0.22 μm membrane filter prior to Fe quantification by ICP-MS.

In order to detect the biodistribution of USPB@SeDMSN@NM, C57BL/6 male mice (8 weeks old) were randomly allocated into two experimental groups (n = 5 per group). The control group received intravenous injections of saline, whereas the treatment group was administered USPB@SeDMSN@NM at an iron-based dosage of 0.5 mg/kg every three days via the same route. Following four rounds of treatment, all animals were euthanized. Major organs, including the heart, liver, spleen, lungs, and kidneys, were harvested, weighed, and subjected to nitric acid digestion after boiling. Iron and selenium contents in the digested tissue samples were quantitatively analyzed using ICP‒OES.

To visually assess the in vivo targeting affinity of USPB@SeDMSN@NM, DiD-labeled nanoparticles (NPs) were employed. Male ApoE^−/− mice (6 weeks old) were maintained on a high-fat diet (HFD) containing 1.25% cholesterol and 40 kcal% fat for 12 weeks. Subsequently, the mice received intravenous injections via the tail vein of either DiD, USPB@SeDMSN@Lip/DiD, and USPB@SeDMSN@NM/DiD (or, in a separate experimental group: DiD, DiD@SeDMSN, and DiD@SeDMSN@NM) at an equivalent DiD dose of 2 mg/kg. Twenty-four hours post-injection, the mice were euthanized and perfused with cold phosphate-buffered saline (PBS) to eliminate circulating blood and unbound dyes. The entire aorta and major organs (including the lungs, heart, liver, spleen, kidneys, and brain) were harvested for ex vivo imaging and fluorescence quantification using an IVIS Spectrum imaging system (PerkinElmer, USA) with an excitation wavelength of 645 nm and an emission wavelength of 665 nm. Furthermore, the aortic root from each sample was fixed, dehydrated, and sectioned for immunofluorescence staining of CD68⁺ macrophages. For the group injected with DiD, USPB@SeDMSN@Lip/DiD, and USPB@SeDMSN@NM/DiD, images were captured using a confocal laser scanning microscope (CLSM, Olympus, Japan). For the other group (DiD, DiD@SeDMSN, and DiD@SeDMSN@NM), whole-slide imaging was performed using a Pannoramic MIDI system (3DHISTECH, Hungary), and the resulting images were processed using CaseViewer software.

In vivo antiatherogenic effects

Eight-week-old male ApoE^−/− mice were fed a high-fat diet (HFD) for 12 weeks. The mice were then randomly assigned to groups (5 mice per group): a model control group (Saline) and groups treated separately with USPB, SeDMSN, USPB@SeDMSN, and USPB@SeDMSN@NM, at a dosage of 0.5 mg/kg (Fe) via tail vein injection every 3 days for 12 injections. After the final injection, the mice were sacrificed, and their blood, whole aortas, and main organs (heart, liver, spleen, lungs, and kidneys) were collected for further analysis. Serum was separated from the whole blood for biochemical and hematological assays. Once the aortic arch was visible, images were captured using a Parallel Light Zoom Stereo Microscope (SZ6100, Novel, China). The aortas were then opened longitudinally and stained with ORO, while 3 μm-thick sections of the aortic root were analyzed for H&E staining, Masson staining, immunohistochemistry (CD68, MMP-9, and γ-H2AX), and DHE immunofluorescence. Main organs from each group were processed into paraffin sections for H&E staining. Quantitative analysis of atherosclerotic lesions, plaque areas, and positively stained areas was conducted using ImageJ.

Author contributions

Yuehuang Wu and Hongping Xia contributed equally to this work. Yuehuang Wu and Jingyi Sheng conceived the research project. Yuehuang Wu, Hongping Xia, He Ding, Mengmeng Long, and Miao Zhang performed and supported the experiments. Yuehuang Wu performed the data analysis. Yuehuang Wu and Jingyi Sheng primarily wrote the manuscript with input from all the authors. Hongping Xia helped with the data analysis. Jingyi Sheng, Ning Gu, and Hongping Xia supervised the research project. All the authors have approved the final version of the manuscript.

Data availability

All data generated or analyzed during this study have been included either in this article or in the supplementary information files.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41392-026-02598-4.