Multifunctional miRNA delivery systems provide synergistic neuroprotection against cerebral ischemia reperfusion injury

Abstract

Effectively alleviating cerebral ischemia-reperfusion (I/R) injury is challenging despite the medical advances. Prompt restoration of blood flow after ischemic stroke causes secondary damage to the brain tissues, triggers neuroinflammation and overproduces reactive oxygen species (ROS). MiRNA regulates the genes involved in neuron apoptosis and neuroinflammation, thus exhibiting potential in ameliorating cerebral I/R injury. However, as miRNA is vulnerable to degradation, its effective delivery faces obstacles in clinical applications. Inspired by the therapeutic potential of miR-210 inhibitors in ischemic stroke-induced neuroinflammation, we constructed a multifunctional nanosystem composed of ceria nanozymes and zeolitic imidazolate framework-90 (ZIF-90) nanoparticles to deliver miR-210 inhibitors to treat cerebral I/R injury. Attributed to the proton sponge effect of ZIF-90, this nanosystem allows for the lysosome escape of miR-210 inhibitors to protect their intracellular bioactivity, while the integration of ZIF-90 and ceria nanozymes utilizes multi-enzyme cascade activities to constrain lipid peroxidation and reduce oxidative damage in brain tissues of mice with middle cerebral artery occlusion (MCAO). After crossing the blood–brain barrier, miR-210 inhibitors target TET2 to suppress pro-inflammatory cytokines, finally inhibiting neuroinflammation. More than the uncertain stability and efficacy of direct TET2 protein administration, the delivery of miR-210 inhibitors by multifunctional nanosystems engendered neuroprotection, indicating their potential for protein replacement therapy against cerebral I/R injury.

---

New concepts

This work introduces a novel combination of ceria nanozymes and one of the metal organic frameworks (MOFs), ZIF-90 nanoparticles, to deliver miR-210 inhibitors, addressing the challenges of miRNA degradation and its effective delivery in clinical gene therapy. Synergistic effects such as enhanced stability, targeted delivery, and multifunctional therapeutic functions are achieved in the therapy of cerebral ischemia-reperfusion (I/R) injury. Unlike traditional methods that struggle with miRNA stability and efficacy, this nanosystem provides a more stable and targeted miRNA delivery mechanism. The proton sponge effect of ZIF-90 and its ability to facilitate the lysosome escape of therapeutic agents provide new insights into the design of MOF-based nanosystems for miRNA delivery. The integration of ceria nanozymes and ZIF-90 overcomes the bottleneck of ceria nanozyme aggregation in vivo and proposes multienzyme-mimic cascade activities to alleviate oxidative damage. With effective miRNA delivery targeting TET2, it exceeds direct protein administration in stability and efficacy, exhibiting a superior antiinflammation therapy over protein-based therapy. Therefore, this work highlights the potential of integrating MOFs with nanozymes for biomedical applications. It also opens up new avenues for developing advanced nanomaterials that can address complex therapeutic challenges like neuroinflammation and oxidative damage in cerebral I/R injury.

1. Introduction

Ischemic stroke is the most common cerebrovascular disease, which causes deaths or severe disabilities with the loss or damage of neuronal function.^1,2 A cascade of biochemical disturbances, including energy deficit, cellular necrosis, inflammation and immune response, is triggered in the infarcted region of the brain, ultimately resulting in serious brain damage.³ For acute ischemic stroke patients, intravenous thrombolysis and mechanical thrombectomy within a narrow therapeutic window of 4.5 h are the representative interventions to reestablish cerebral perfusion.⁴ However, injured neurons undergo massive oxidative stress, mitochondrial dysfunction, protein oxidation and excessive inflammatory reactions following reperfusion, all of which could contribute to the secondary injury in cerebral vasculature and neuronal circuits termed as cerebral ischemia-reperfusion (I/R) injury.⁵ Current interventions are insufficient in improving neurological outcomes following a cerebral I/R injury.⁶

MicroRNAs (miRNAs) as endogenous RNAs regulate gene expression through post-transcriptional suppression by binding to the 3’UTR of their target mRNAs and also interact with other regions of the mRNA, such as the 5′UTR, coding sequences, and promoters.^7,8 They provide essential functions in maintaining cerebrovascular homeostasis and mediating pathological conditions. Neuroinflammation is a complex immune response and a key factor in determining prognosis after ischemic stroke, including the activation of microglia and subsequent release of inflammatory factors.⁹ Several types of miRNAs target several genes in cerebral I/R injury and participate in the regulation of neuroinflammation, which contributes to the necessity of miRNA as a “druggable” target in cerebral I/R therapy.¹⁰ MiR-32-3p plays a key role in nerve cell death and stimulates cerebral I/R injury by suppressing Cab39/AMPK.¹¹ Overexpression of miR-10b-3p can reduce neurological deficits, infarct volume and brain edema in cerebral I/R injury rats.¹² In comparison, miR-210 has been considered a novel biomarker in cerebral I/R injury.¹³ Elevated miR-210 expression levels potentiate neuronal apoptosis in rat models of cerebral ischemia via targeting the HIF-1α-VEGF signaling cascade.¹⁴ Inhibition of miR-210 expression decreased ischemic stroke-induced neuroinflammatory response via suppression of ten-eleven translocation methylcytosine dioxygenase 2 (TET2) in the adult mouse brain, wherein TET2 is responsible for regulating the expression of proinflammatory cytokines in macrophage/microglia.^15,16 Administration of complementary locked nucleic acid oligonucleotides targeting miR-210 demonstrates dual therapeutic efficacy in attenuating neuroinflammation and rescuing neuronal viability.^17,18 However, similar to the other miRNAs with unstable chemical structures, miR-210 inhibitors can be degraded easily by various nucleases in vivo.^19,20 Developing an effective miR-210 delivery system could improve their therapeutic efficacy in treating ischemic stroke.

Nanozymes are novel nanomaterials with enzyme-like activity such as those of superoxide dismutase (SOD), catalase (CAT), peroxidase (POD), and glutathione peroxidase (GPX).^21,22 With their reversible Ce³⁺/Ce⁴⁺ redox pair and surface oxygen vacancies, CeO₂ nanozymes exhibit SOD-, CAT- and POD-mimicking activities.^23,24 In multi-enzyme cascade catalysts, SOD catalyzes the dismutation of superoxide radicals (O₂⁻) into H₂O₂ and O₂, while H₂O₂ is transformed to innocuous H₂O and O₂ in the CAT- and POD-mimetic dismutation cycle, both of which endow CeO₂ nanozymes with high antioxidant activity and ROS scavenging ability.²⁵ However, due to their easy interparticle aggregation and unspecific catalytic reaction on exposed active sites, CeO₂ nanozymes have limited clinical application.^26,27 With their tunable pore size, large surface area, well-defined structure, and abundant single-metal sites, metal–organic frameworks (MOFs) composed of metallic centers and organic linkers have shown excellent performance in delivery.²⁸ One of the MOFs, ZIF-90, remains stable under physiological conditions and easily decomposes under acidic environments due to the protonation effect.²⁹ The degradation of ZIF-90 potentially stimulates the rupture of lysosomes by buffering the pH inside, which is called “proton sponge”.³⁰ It is beneficial for releasing cargos like nucleic acids and avoiding their inactivation in lysosomes. Loaded with CeO₂ nanozymes, the multifunctional ZIF-90-based nanoparticles as delivery systems of miR-210 inhibitors not only protects nucleic acids from degradation, but also serves as a robust, operationally stable and recyclable biocatalyst mimic.

During cerebral I/R, brain tissues experience a surge in ROS production, which can cause oxidative damage to neurons, cellular components and nucleic acids like miRNAs.³¹ However, the oxidation of miRNAs generally causes its degradation or functional impairment, thus compromising its stability and efficacy in brain tissues. This raises concerns about the feasibility of using miRNAs as therapeutic agents in cerebral I/R injury treatment without addressing the issue of ROS-induced damage. In this work, we designed ZIF-90-capped CeO₂ (ZIF-90/CeO₂) nanoparticles to deliver miR-210 inhibitors. ZIF-90/CeO₂ nanoparticles protect miR-210 inhibitors against premature degradation in lysosomes and oxidative damage in brain tissues following cerebral I/R, exhibiting a synergetic strategy to treat cerebral I/R injury. A schematic illustrating the overall work is shown in Fig. 1. ZIF-90/CeO₂ nanoparticles exhibited effective ROS scavenging capability, protecting against mitochondrial dysfunction and apoptosis of neurons subjected to oxygen-glucose deprivation/reoxygenation (OGD/R). In addition, miR-210 inhibitors can escape from ZIF-90/CeO₂ nanoparticles, while retaining their bioactivity due to the pH-buffering capability of ZIF-90. Collectively, this study sheds light on the combination of miRNA-based neuroprotective gene therapy and ROS clearance against cerebral I/R injury.

Fig. 1 Schematic for the synthesis of miR-210@ZIF-90/CeO₂ nanoparticles and their neuroprotective mechanism against reperfusion-induced injury in ischemic stroke.

2. Results and discussion

2.1. Synthesis and characterization of miR-210@ZIF-90/CeO₂ nanoparticles

ZIF-90 nanoparticles were prepared as reported, and ZIF-90/CeO₂ nanoparticles were synthesized by loading CeO₂ nanozymes within the nanoparticles and on its surface.^32,33 Transmission electron microscopy (TEM) images of the CeO₂ nanozymes, ZIF-90 nanoparticles and ZIF-90/CeO₂ nanoparticles are presented in Fig. S1. Then, we loaded the miR-210 inhibitors on the surface of the ZIF-90/CeO₂ nanoparticles via electrostatic interaction, and the morphology of miR-210@ZIF-90/CeO₂ nanoparticles was characterized by TEM (Fig. 2(a)). The images showed that the miR-210@ZIF-90/CeO₂ nanoparticles were larger than the ZIF-90/CeO₂ nanoparticles and ZIF-90 nanoparticles. This is due to the supplement of miR-210 inhibitors and CeO₂ nanozymes. Scanning electron microscopy (SEM) was employed to evaluate the morphology of the miR-210@ZIF-90/CeO₂ nanoparticles, which confirmed that the miR-210@ZIF-90/CeO₂ nanoparticles have uniform morphologies (Fig. 2(b)). The hydrodynamic size of the miR-210@ZIF-90/CeO₂ nanoparticles is about 236 ± 12.14 nm, as detected by dynamic light scattering (DLS) (Fig. 2(c)). The high-resolution transmission electron microscopy (HR-TEM) image and corresponding energy dispersive spectrum (EDS) elemental mapping in Fig. 2(d) and (e) show the spatial distribution of Ce elements in the miR-210@ZIF-90/CeO₂ nanoparticles, which reveal that the CeO₂ nanozymes are loaded by ZIF-90 nanoparticles both via encapsulation and surface absorbance. Moreover, although K₂PtCl₄ was used as a catalyst to prepare CeO₂ nanozymes, Pt did not exist in the miR-210@ZIF-90/CeO₂ nanoparticles, which avoids the side effects of residual Pt in vivo (Table S1 and Fig. S2). We compared the crystalline structure of the ZIF-90/CeO₂ nanoparticles with that of the calculated ZIF-90 using X-ray diffraction (XRD). The results indicate that the ZIF-90/CeO₂ nanoparticles still maintain the crystalline structure of ZIF-90 (Fig. 2(f)). After being loaded with miR-210 inhibitors and CeO₂ nanozymes, the miR-210@ZIF-90/CeO₂ nanoparticles had a decreased surface area and a slightly reduced pore volume, which also confirmed that the surface of the ZIF-90 nanoparticles is covered with CeO₂ and miR-210 inhibitors (Fig. 2(g) and (h)). We further used X-ray photoelectron spectroscopy (XPS) to assess the composition of ZIF-90/CeO₂. Results in Fig. 2(i) confirm the presence of Zn 2p and Ce 3d peaks. While Zn exhibits two electronic orbitals such as Zn 2p_1/2 and Zn 2p_3/2 in Fig. 2(j), Ce is constituted by Ce⁴⁺ (oxidized) and Ce³⁺ (reduced) in Fig. 2(k), whose shifting between the two forms contributes to the multienzyme-like capability. miR-210@ZIF-90/CeO₂ labeled with cy3 (denoted as miR-210-cy3@ZIF-90) nanoparticles exhibited an identical characteristic peak at 570 nm in the UV-vis spectra that is attributed to cy3-labeled miR-210 inhibitors, indicating the existence of cy3-labeled miR-210 inhibitors in the miR-210-cy3@ZIF-90/CeO₂ nanoparticles (Fig. 2(l)). In Fig. 2(m), miR-210@ZIF-90/CeO₂ and ZIF-90/CeO₂ nanoparticles present the same characteristic peaks in FTIR spectra as ZIF-90, which demonstrates that the ZIF-90 nanoparticles are their frameworks. MiRNAs are inherently negatively charged due to their phosphate backbone, while ZIF-90/CeO₂ nanoparticles are positively charged, which provides a favorable electrostatic environment for binding. The zeta potentials of miR-210@ZIF-90/CeO₂ nanoparticles detected by dynamic light scattering (DLS) were negative and similar to that of miRNAs (Fig. 2(n)). This further confirmed the loading of miR-210 inhibitors on ZIF-90/CeO₂. Stabilities of ZIF-90 nanoparticles in normal conditions were evaluated by monitoring their morphology via TEM. With the unchanged sizes and morphologies shown in Fig. S3, the ZIF-90 nanoparticles were verified to remain stable in PBS at a pH value of 7.4 for as long as 7 days. After being loaded with ZIF-90/CeO₂ nanoparticles, the miR-210 inhibitors also showed unchanged UV-vis spectra for up to 5 h (Fig. S4), indicating their stability when encapsulated into the ZIF-90/CeO₂ nanoparticles. The results above demonstrated that the miR-210@ZIF-90/CeO₂ nanoparticles are composed of ZIF-90 frameworks, CeO₂ nanozymes and miR-210 inhibitors, and they have an improved stability.

Fig. 2 Structural characterization of miR-210@ZIF-90/CeO₂ nanoparticles. (a) TEM image of miR-210@ZIF-90/CeO₂ nanoparticles, scale bar: 200 nm. (b) SEM image of miR-210@ZIF-90/CeO₂ nanoparticles, scale bar: 100 nm. (c) Hydrodynamic size distribution of miR-210@ZIF-90/CeO₂ nanoparticles. (d) High-resolution transmission electron microscopy (HR-TEM) image of miR-210@ZIF-90/CeO₂ nanoparticles, scale bar: 100 nm. (e) Energy dispersive X-ray spectroscopy (EDX) mapping of miR-210@ZIF-90/CeO₂ nanoparticles, scale bar: 100 nm. (f) X-ray diffraction (XRD) patterns of ZIF-90/CeO₂ nanoparticles. (g) N₂ adsorption–desorption isotherms of ZIF-90, ZIF-90/CeO₂ and miR-210@ZIF-90/CeO₂ nanoparticles. (h) Pore diameters of ZIF-90, ZIF-90/CeO₂ and miR-210@ZIF-90/CeO₂ nanoparticles. (i) XPS spectra of ZIF-90/CeO₂ nanoparticles. XPS analysis of Zn 2p (j) and Ce 3d (k) spectra of ZIF-90/CeO₂ nanomaterials. (l) UV-vis spectra of CeO₂, ZIF-90, ZIF-90/CeO₂, miR-210-cy3 inhibitors and miR-210-cy3@ZIF-90/CeO₂. (m) FTIR spectra of CeO₂, ZIF-90, ZIF-90/CeO₂, and miR-210@ZIF-90/CeO₂. (n) Zeta potentials of CeO₂, ZIF-90, ZIF-90/CeO₂, and miR-210@ZIF-90/CeO₂.

2.2. Cytotoxicity and cellular uptake of miR-210@ZIF-90/CeO₂ nanoparticles in vitro

Before the in vitro efficacy evaluation of the miR-210@ZIF-90/CeO₂ nanoparticles, we optimized the encapsulation and loading efficiency of miR-210 inhibitors in ZIF-90/CeO₂ nanoparticles. Mass ratios of cy3-labeled miR-210 inhibitors to ZIF-90/CeO₂, ranging from 1 : 5, 1 : 20, 1 : 40, 1 : 50, to 1 : 100, were analyzed for their encapsulation and loading efficiency. Based on a standard curve of cy3-labeled miR-210 inhibitors in UV-vis spectra (Fig. S5), we verified the miR-210 loading ratio in the aforementioned miR-210-cy3@ZIF-90/CeO₂ by detecting the amount of unloaded miR-210-cy3 in the supernatants of samples. We found that the encapsulation efficiency of the miR-210-cy3 inhibitors gradually increased as the amounts of ZIF-90/CeO₂ nanoparticles increased. However, there was an obvious decrease in the encapsulation efficiency when the mass ratio of the miR-210-cy3 inhibitors to ZIF-90/CeO₂ nanoparticles reached ∼1 : 100, which suggests a maximum miR-210 inhibitor loading efficiency at a mass ratio of 1 : 50 (Fig. 3(a)). As a result, the loading amount of miR-210 inhibitors has a similar tendency toward encapsulation efficiency, where a mass ratio of miR-210 inhibitors to ZIF-90/CeO₂ at 1 : 50 achieved the highest loading amount (≈1.66 μg) among all other mass ratios. However, since the loading efficiency of miR-210 inhibitors in the miR-210@ZIF-90/CeO₂ nanoparticles was related to the loading amount of miR-210 inhibitors and mass of ZIF-90/CeO₂ nanoparticles, it showed that the loading efficiency of the miR-210 inhibitors in the miR-210@ZIF-90/CeO₂ nanoparticles was decreased with an increasing amount of ZIF-90/CeO₂ nanoparticles, due to the growth of excessive ZIF-90/CeO₂ nanoparticles (Fig. 3(b)). Therefore, a mass ratio of 1 : 50 (miR-210-cy3 inhibitors to ZIF-90/CeO₂ nanoparticles) was selected to construct miR-210@ZIF-90/CeO₂ nanoparticles in the subsequent experiments.

Fig. 3 Cytotoxicity and cellular uptake of miR-210@ZIF-90/CeO₂ nanoparticles in vitro. (a) Encapsulation capacity of miR-210@ZIF-90/CeO₂ nanoparticles with mass ratios of miR-210-cy3 inhibitors to ZIF-90/CeO₂ nanoparticles of 1 : 5, 1 : 20, 1 : 40, 1 : 50, and 1 : 100 (w/w), n = 3. (b) Loading efficiency of miR-210@ZIF-90/CeO₂ nanoparticles with mass ratios of miR-210-cy3 inhibitors to ZIF-90/CeO₂ nanoparticles of 1 : 5, 1 : 20, 1 : 40, 1 : 50, and 1 : 100 (w/w), n = 3. (c) Stability of miR-210-cy3 inhibitors in miR-210@ZIF-90/CeO₂ nanoparticles by UV-vis spectra at 570 nm, n = 3. (d) Accumulative release of miR-210-cy3 inhibitors from miR-210@ZIF-90/CeO₂ nanoparticles in PBS of pH 5.5, n = 3. (e)–(g) MTT analysis for cytotoxicity of ZIF-90 (e), ZIF-90/CeO₂ (f) and CeO₂ (g) in SH-SY5Y with 24 h, n = 5. (h) Colocalization of miR-210@ZIF-90/CeO₂ nanoparticles (20 μg mL⁻¹) and endo/lysosomes in SH-SY5Y was studied at 4 h, 16 h and 24 h post-treatment. Scale bar: 10 μm. (i) Relative fluorescence intensity of miR-210-cy3 inhibitors, n = 6. (j) qPCR of miR-210 mRNA level in oxygen glucose deprivation/reoxygenation (OGD/R)-induced SH-SY5Y treated with miR-210 inhibitors (without transfection reagent, 20 μg mL⁻¹) or miR-210@ZIF-90/CeO₂ nanoparticles (20 μg mL⁻¹), n = 6. (k) Endocytosis of miR-210-cy3@ZIF-90/CeO₂ nanoparticles by SH-SY5Y cells pretreated with chlorpromazine, wortmannin and cytochalasin D for 6 h, scale bar: 10 μm. (l) Fluorescence intensity of miR-210-cy3@ZIF-90/CeO₂ nanoparticles in SH-SY5Y cells pretreated with chlorpromazine, wortmannin and cytochalasin D for 6 h, n = 6. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant.

Due to the strong electronegativity of N from the imidazole ester structure, ZIF-90 nanoparticles are vulnerable to electron withdrawal in acidic environments, contributing to the fast structural disruption of ZIF-90 nanoparticles and pH-sensitive release of cargos.^34,35 Firstly, we compared the UV-vis spectra of miR-210-cy3@ZIF-90/CeO₂ nanoparticles in PBS at pH values of 7.4 and 5.5 to reveal the release of miR-210 inhibitors in acidic conditions and their stability in normal conditions. As evidenced by the absorbance of cy3-labeld miR-210 inhibitors at 570 nm in the UV-vis spectra, the characteristic peak at 570 nm is present in miR-210-cy3@ZIF-90/CeO₂ nanoparticles immersed in PBS at pH 5.5, but absent in the miR-210-cy3@ZIF-90/CeO₂ nanoparticles immersed in PBS at pH 7.4, indicating the release of miR-210-cy3 under acidic conditions (Fig. S6). Then, we detected the changing quantity of miR-210-cy3 inhibitors in a PBS solution at pH 7.4 through UV-vis spectra after the miR-210-cy3@ZIF-90/CeO₂ nanoparticles were immersed for different intervals. With 40% of the miR-210 inhibitors remaining stable after 8 h, the stability of the miR-210 inhibitors loaded with ZIF-90/CeO₂ nanoparticles was revealed. This is stable enough to ensure their circulation in blood and endocytosis by cells (Fig. 3(c)). In addition, we detected the release profile of miR-210-cy3 inhibitors from miR-210-cy3@ZIF-90/CeO₂ nanoparticles in PBS at pH 5.5, which simulates the acidic environment of cellular lysosomes (Fig. 3(d)). The results indicated that almost all miR-210 inhibitors were released from the ZIF-90/CeO₂ nanoparticles within 45 min under lysosome-mimicking conditions. This confirms that the delivery of ZIF-90/CeO₂ maintains the stability of miR-210 inhibitors under normal conditions and promotes their release in lysosomes, thus benefiting the lysosome escape of miR-210 inhibitors. Next, the cytotoxicities of ZIF-90, CeO₂, and ZIF-90/CeO₂ nanoparticles to SH-SY5Y cells were evaluated using MTT assays. The results indicated that as the concentrations of ZIF-90, CeO₂ and the ZIF-90/CeO₂ nanoparticles gradually increased, the survival rate of the SH-SY5Y cells remained at almost 100%. Even at the maximum concentration of 20 μg mL⁻¹, all cell viabilities of ZIF-90, CeO₂, and ZIF-90/CeO₂ nanoparticles remained relatively high, suggesting that this miRNA delivery system exhibited negligible toxicity towards SH-SY5Y cells (Fig. 3(e)–(g)). Consequently, to optimize the therapeutic efficacy of miR-210-cy3@ZIF-90/CeO₂ nanoparticles, a concentration of 20 μg mL⁻¹ was selected as the working concentration of ZIF-90/CeO₂ nanoparticles in subsequent in vitro experiments.

Following the evaluation of release of miR-210 inhibitors under acidic conditions, we subsequently investigated the cellular internalization of miR-210-cy3@ZIF-90/CeO₂ nanoparticles, which is a prerequisite for satisfactory gene expression. After 4 h of incubation with miR-210-cy3@ZIF-90/CeO₂ nanoparticles, the SH-SY5Y cells internalized some miR-210-cy3@ZIF-90/CeO₂ nanoparticles, which were located within the lysosomes with overlapped red and green fluorescence under a confocal laser scanning microscope (CLSM). However, with increasing incubation time, the green fluorescence from the lysosomes and red fluorescence from the miR-210-cy3 inhibitors became separated, indicating that the miR-210-cy3 inhibitors escaped from the lysosomes (Fig. 3(h)). Moreover, the intensity of the red fluorescence from the miR-210 inhibitors in the SH-SY5Y cells gradually increased with increasing incubation time, suggesting a continuous internalization of the miR-210-cy3 inhibitors into the SH-SY5Y cells (Fig. 3(i)). The cellular internalization results validated the lysosomal escape of miR-210 inhibitors, thereby improving the bioactivity of miR-210 inhibitors when being delivered to lysosomes.

We analyzed the transfection efficiency of miR-210-cy3@ZIF-90/CeO₂ nanoparticles in oxygen glucose deprivation/reoxygenation (OGD/R)-induced SH-SY5Y cells by qPCR after 24 h of incubation. Overexpression of miR-210 in the OGD/R model of SH-SY5Y cells was significantly decreased by miR-210@ZIF-90/CeO₂ nanoparticles, compared to the treatment of naked miR-210 inhibitors (Fig. 3(j)), suggesting that miR-210 inhibitors were successfully transfected into SH-SY5Y cells. This was attributed to the bioactivity protection of the ZIF-90/CeO₂ nanoparticles.

Finally, to investigate the endocytic pathways of miR-210-cy3@ZIF-90/CeO₂ nanoparticles in SH-SY5Y cells, we used chlorpromazine (inhibitor of clathrin-mediated endocytosis), wortmannin (inhibitor of macropinocytosis) and cytochalasin D (inhibitor of phagocytosis) to pretreat SH-SY5Y cells for 30 min, and treated the cells with miR-210-cy3@ZIF-90/CeO₂ for 6 h.³⁶ The fluorescence intensity of the miR-210-cy3@ZIF-90/CeO₂ nanoparticles in all treated groups decreased in the presence of endocytic inhibitors compared with the control group cultured in 37 °C (Fig. 3(k)). Given that endocytosis was a major mechanism of internalization of the miR-210-cy3@ZIF-90/CeO₂ nanoparticles, clathrin-mediated endocytosis, micropinocytosis and phagocytosis simultaneously participate in the internalization process (Fig. 3(l)).

2.3. Antioxidant and anti-apoptosis of miR-210@ZIF-90/CeO₂ nanoparticles in vitro

CeO₂ nanozymes have been demonstrated to exhibit antioxidative properties. This is due to their enzyme-like activities in modulating ROS levels and the reversible conversion between the Ce³⁺ and Ce⁴⁺ valence states.³⁷ We examined the SOD-like enzymatic activity of ZIF-90 and ZIF-90/CeO₂ nanoparticles. It was found that the SOD enzymatic activity of the ZIF-90/CeO₂ nanoparticles was significantly higher than that of the ZIF-90 nanoparticles (Fig. 4(a)). Similar tendencies exist in CAT and POD enzymatic activity (Fig. 4(b) and (c)). The results above demonstrate that ZIF-90/CeO₂ could serve as an antioxidant for ROS clearance, which is attributed to CeO₂. We further analyzed the antioxidant activities of CeO₂, ZIF-90, and ZIF-90/CeO₂ nanoparticles by ABTS and DPPH free radical scavenging assays. They both indicated that the free radical scavenging capability of ZIF-90/CeO₂ nanoparticles was notably elevated compared to that of CeO₂ and ZIF-90 nanoparticles individually (Fig. 4(d) and (e)). This indicates that the ZIF-90/CeO₂ nanoparticles have superior antioxidative property compared to CeO₂ and ZIF-90, because of the inherent multienzymatic activity.

Fig. 4 Protection of miR-210@ZIF-90/CeO₂ to SH-SY5Y cells against oxygen glucose deprivation/reoxygenation (OGD/R)-induced oxidative damage. (a) SOD-like enzymatic activities of ZIF-90 and ZIF-90/CeO₂ nanoparticles, n = 3. (b) CAT-like enzymatic activities of ZIF-90 and ZIF-90/CeO₂ nanoparticles, n = 5. (c) POD-like enzymatic activities of ZIF-90 and ZIF-90/CeO₂ nanoparticles, n = 3. (d) and (e) In vitro antioxidant activity of CeO₂, ZIF-90, and ZIF-90/CeO₂ nanoparticles, as determined by ABTS (d) and DPPH (e) free radical scavenging assays, n = 3. (f) ROS levels of SH-SY5Y cells post-OGD/R detected by flow cytometric analysis. (g) Mitochondrial membrane potential (MMP) of SH-SY5Y cells post-OGD/R by flow cytometric analysis. (h) Apoptosis of SH-SY5Y cells post-OGD/R detected by flow cytometric analysis. (i) OGD/R-induced ROS levels of SH-SY5Y cells detected by flow cytometry, n = 6. (j) OGD/R-induced MMP of SH-SY5Y cells detected by flow cytometry, n = 6. (k) OGD/R-induced apoptosis ratio of SH-SY5Y cells detected by flow cytometry, n = 6. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant.

Next, the therapeutic efficacy of the miR-210@ZIF-90/CeO₂ nanoparticles in vitro was explored in SH-SY5Y cells under OGD/R conditions. This is the in vitro model for simulating the reperfusion injury of neurons after ischemia and hypoxia in vivo. For the induction of OGD/R, SH-SY5Y cells undergo 8 h of oxygen glucose deprivation and 12 h of reoxygenation, followed by 24 h of different treatments. Large amounts of ROS were produced during cerebral I/R injury, leading to oxidative stress and subsequent damage to neurons.³⁸ SH-SY5Y cells induced by OGD/R were stained with 2,7-dichlorofluorescein diacetate (DCFH-DA) and flow cytometry was used for the detection of ROS levels (Fig. 4(f)). Under OGD/R stress, cellular ROS levels significantly increased from 1.00 ± 0.03 to 4.59 ± 0.08 compared to the control group, demonstrating a boost of ROS levels by OGD/R. However, SH-SY5Y cells treated with naked miR-210 inhibitors still have a relatively higher ROS level (about 4.47 ± 0.05 of relative fluorescence intensity), indicating that naked miR-210 inhibitors have a limited efficacy in mitigating oxidative stress. In contrast, the cellular ROS levels significantly decreased after treatment of CeO₂, ZIF-90/CeO₂, and miR-210@ZIF-90/CeO₂ nanoparticles, among which miR-210@ZIF-90/CeO₂ eliminated ROS to the lowest levels (Fig. 4(i)). It not only underscores the intrinsic antioxidant stress properties of the miR-210@ZIF-90/CeO₂ nanoparticles, but also validates their efficacy in reducing cellular ROS levels.

The mitochondrial membrane potential (MMP), which refers to the electrochemical gradient across the mitochondrial inner membrane, serves as a prerequisite for maintaining mitochondrial oxidative phosphorylation and ATP production.³⁹ We then employed JC-1 to label the MMP alterations of SH-SY5Y cells under various intervention conditions by flow cytometry. JC-1 could visualize changes in MMP, which aggregates within the mitochondria, with red fluorescence at high MMP levels. Conversely, upon the decrease of MMP, JC-1 dissociates into monomers, resulting in the emission of green fluorescence. By quantifying the ratio of red to green fluorescence, we found that the mitochondria experienced notable changes in MMP under OGD/R conditions, wherein the ratio of JC-1 aggregates to monomers was markedly decreased from 23.81 ± 4.26 to 2.03 ± 0.17, compared to the control group. However, treatments of CeO₂, ZIF-90/CeO₂, and miR-210@ZIF-90/CeO₂ nanoparticles exert effective roles in decreasing the proportion of JC-1 monomers. Notably, the miR-210@ZIF-90/CeO₂ nanoparticles exhibited the most obvious efficacy, with the ratio of JC-1 aggregates to monomers at 8.21 ± 0.86 (Fig. 4(g) and (j)). This suggests the efficacy of the miR-210@ZIF-90/CeO₂ nanoparticles in alleviating MMP damage induced by OGD/R, which is largely due to the antioxidative function of ZIF-90/CeO₂.

Since high levels of ROS are associated with the induction of apoptosis, we further used flow cytometry to investigate the anti-apoptosis effects of miR-210@ZIF-90/CeO₂ nanoparticles following OGD/R. The results revealed that OGD/R pretreatment caused apoptosis of SH-SY5Y cells with 7.95% ± 1.64%. Naked miR-210 inhibitors did not exert any effects in anti-apoptosis, which is largely due to their inefficient internalization.⁴⁰ CeO₂ and ZIF-90/CeO₂ with antioxidative properties could decrease the apoptotic rate of SH-SY5Y cells to a certain extent because of their antioxidative activity. Integrated with anti-apoptosis of miR-210 inhibitors and antioxidative properties of ZIF-90/CeO₂, miR-210@ZIF-90/CeO₂ nanoparticles showed a synergetic antiapoptosis efficacy (Fig. 4(h) and (k)). These results suggest that the miR-210@ZIF-90/CeO₂ nanoparticles effectively mitigated oxidative stress-induced cellular apoptosis.

After being endocytosed by SH-SY5Y cells, miR-210@ZIF-90/CeO₂ nanoparticles could effectively release miR-210 inhibitors and CeO₂ nanozymes, which have been confirmed by the above results. With cy3-labeling, miR-210 inhibitors in SH-SY5Y cells were verified by red fluorescence (Fig. 3(g) and (h)), and almost all of miR-210 inhibitors were released within 45 min in PBS (Fig. 3(d)). As for the CeO₂ nanozymes, the antioxidant activity of the CeO₂ nanoparticles was evidenced by the decreased ROS levels of SH-SY5Y cells in Fig. 4(f) and (i), which is attributed to the higher multi-enzyme mimic activity of CeO₂. Collectively, the successful release of miR-210 inhibitors and ZIF-90/CeO₂ synergistically protects SH-SY5Y cells from OGD/R damage in vitro.

2.4. Therapeutic efficacy of miR-210@ZIF-90/CeO₂ nanoparticles in mice with cerebral ischemia-reperfusion injury

We established an acute middle cerebral artery occlusion (MCAO) mouse model to assess the therapeutic efficacy of miR-210@ZIF-90/CeO₂ nanoparticles on mice cerebral I/R injury (Fig. 5(a)). Mice were divided into five groups, including healthy mice in the sham group and MCAO mice administrated separately with miR-210@ZIF-90/CeO₂, ZIF-90/CeO₂, miR-210@ZIF-90, and miR-210@liposome. Firstly, miR-210 transfection efficiencies in brain tissues after different treatments were compared via qPCR, as shown in Fig. 5(b). It showed that miR-210 expression in the MCAO group was upregulated, compared to the sham group, which was reported to induce neuroinflammatory in the previous study.¹⁵ The expression of miR-210 was notably downregulated following intervention with miR-210@ZIF-90/CeO₂ nanoparticles. As we used liposomes in comparison with ZIF-90/CeO₂ nanoparticles, miR-210 inhibitors were loaded by liposomes, which is denoted as miR-210@liposome nanoparticles. The miR-210 levels in the brain tissue of mice treated with miR-210@ZIF-90/CeO₂ nanoparticles were much lower than those treated with miR-210@liposome nanoparticles. This confirms that ZIF-90/CeO₂ nanoparticles enable the high transfection efficiency of miR-210 inhibitors, which is due to its superior lysosomal escape capability compared to that of liposomes. Neuroinflammatory response and glial scar formation are accompanied by cerebral I/R injury, where astrocytes and microglia as major inflammatory cells can be activated by inflammation in the brain tissue. Immunofluorescence staining of the glial fibrillary acidic protein (GFAP; astrocyte-specific intermediate filament) and ionized calcium-binding adaptor molecule-1 (Iba-1; microglia-specific calcium-binding protein) in the brain sections of MCAO mice was employed to assess the activation of astrocytes and microglia in MCAO mice, respectively. We confirmed that microglia and astrocytes were activated by MACO, as the expression levels of Iba-1 and GFAP in the brain tissues of MCAO mice were significantly increased, which is the indicator for the successful construction of MCAO mice model (Fig. 5(c) and (d)). Following treatments with miR-210@ZIF-90/CeO₂, ZIF-90/CeO₂, miR-210@ZIF-90, and miR-210@liposome, the number of GFAP-positive and Iba-1-positive cells decreased to varying degrees, with the miR-210@ZIF-90/CeO₂ nanoparticles group exhibiting the lowest counts of both GFAP-positive and Iba-1-positive cells. This reveals that miR-210@ZIF-90/CeO₂ nanoparticles could effectively alleviate the activation of microglia and astrocytes in the hippocampus of MCAO mice (Fig. 5(e) and (f)). Additionally, we evaluated the Longa score of MCAO mice after intracerebral administration, and found that miR-210@ZIF-90/CeO₂ relieved stroke-induced neurological deficits (Fig. 5(g)). The results of 2,3,5-Triphenyl Tetrazolium Chloride (TTC) staining indicated the infarct areas of brain tissues in MCAO mice models after different treatments, which is used to evaluate the therapeutic efficacy. The results showed that mice in the MCAO group exhibited the largest cerebral infarcted sizes among all groups, accounting for 39.00% ± 5.55% of the brain volume. In contrast, treatments of miR-210@ZIF-90/CeO₂, ZIF-90/CeO₂, miR-210@ZIF-90, and miR-210@liposome decreased the infarct areas of brain tissue of MCAO mice models, which were 10.89% ± 1.06%, 21.99% ± 3.55%, 15.32% ± 1.90%, and 27.43% ± 4.09%, respectively. Notably, the brain of mice in the miR-210@ZIF-90/CeO₂ nanoparticles group exhibited the smallest infarct volume in brain tissue in all MCAO models (Fig. 5(h)). Although MCAO mice treated with ZIF-90/CeO₂ or miR-210@liposome have reduced infarct sizes and improved Longa scores, their therapeutic efficacy is relatively lower than that of miR-210@ZIF-90/CeO₂. This indicated that ZIF-90/CeO₂ with lysosome escape ability could maintain the activity of miR-210, thus increasing the efficacy of miR-210 inhibitors when compared to liposome (Fig. 5(i)). Furthermore, HE staining was performed on brain tissue sections from mice to further reveal the pathology. It showed that most of the pathological changes in the brain tissues from the miR-210@ZIF-90/CeO₂ group have been recovered, while mice in the other groups except the sham group still have a pathological injury in the brain (Fig. 5(j)). Collectively, these results demonstrate that the miR-210@ZIF-90/CeO₂ nanoparticle can restore cerebral injury in vivo.

Fig. 5 Therapeutic outcomes after different treatments in mice with cerebral I/RI. (a) Experimental flowchart and timeline for the in vivo study in mice with cerebral I/R injury. (b) miR-210 transfection efficiency in brain tissues using qPCR at 2 days after different treatments, n = 6. (c)–(f) Immunofluorescence staining analysis of hippocampal microglia activation (c), (e) and hippocampal astrocyte activation (d), (f) in MCAO mice after different treatments, n = 6. Scale bar: 20 μm. (g) Longa score analysis of neurological deficits from MCAO mice after different treatments, n = 6. (h) and (i) TTC staining analysis of the infarction area of brain tissues from MCAO mice after different treatments, n = 6. (j) Representative photomicrographs of H&E staining of brain tissues from different treatment groups, n = 6. Scale bar: 100 μm. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant.

2.5. Protective mechanism of miR-210@ZIF-90/CeO₂ nanoparticles on cerebral ischemia-reperfusion injury in vivo

Given that miR-210@ZIF-90/CeO₂ effectively alleviated cerebral I/R injury in vivo, we further studied their protective mechanism in mice with MCAO. We conducted dihydroethidium (DHE) staining on brain tissue to detect the ROS levels of the brain tissue. As depicted in Fig. 6(a), a notable increase in DHE fluorescence intensity was observed in the MCAO group, indicating ROS overproduction in the brain tissue after cerebral I/R injury. Among the miR-210@ZIF-90, ZIF-90/CeO₂, miR-210@liposome and miR-210@ZIF-90/CeO₂ groups, ROS levels of the brain tissues of all the groups decreased in comparison with those of the MCAO group, while the lowest ROS levels were observed in the brain tissue from mice in the miR-210@ZIF-90/CeO₂ group. Malondialdehyde (MDA) is generated during membrane lipid peroxidation, and is considered an indirect indicator of intracellular ROS levels.⁴¹ Compared to the sham group with an MDA level of 101.98 ± 30.82 nmol mg⁻¹, the MDA content in the brain tissues of mice from the MCAO group significantly increased to 484.53 ± 59.86 nmol mg⁻¹. We found that the MDA content in mice brain tissues was significantly reduced upon treatment with miR-210@ZIF-90, ZIF-90/CeO₂, miR-210@liposome and miR-210@ZIF-90/CeO₂, suggesting their capabilities in eliminating ROS (Fig. 6(b)). Reduced glutathione (GSH) and oxidized glutathione (GSSG) together are responsible for the redox balance within cells, while GSH is consumed and converted into GSSG at high ROS levels.⁴² We measured the GSH/GSSG ratio in the brain tissues of mice from various groups. The miR-210@ZIF-90/CeO₂ nanoparticles recovered the GSH/GSSG ratio as compared to the MCAO group, which is consistent with the MDA results (Fig. 6(c)). In addition, since oxidative stress triggers the activation of Nuclear Factor Erythroid 2-Related Factor 2 (NRF2), we analyzed the expression levels of NRF2 in the brain tissues to further validate the oxidative stress of ischemic brain tissue after treatment with miR-210@ZIF-90/CeO₂. As shown in Fig. 6(d) and (e), the expression level of NRF2 in the brain tissue of mice from the MCAO group is upregulated, which reveals the high oxidative stress after cerebral I/R injury. After miR-210@ZIF-90/CeO₂ treatment, the expression of NRF2 in brain tissues was significantly reduced, which is consistent with the results shown in Fig. 6(a)–(c). The results above suggest that although the brain tissues after I/R have relatively high oxidative stress, our ZIF-90/CeO₂-based nanoplatform effectively eliminated ROS, which ensured the bioactivity of the miR-210 inhibitors in the brain tissues. Together, it concluded that the miR-210@ZIF-90/CeO₂ nanoparticles could alleviate oxidative stress in the brain tissue after cerebral I/R injury.

Fig. 6 Brain protective mechanism of miR-210@ZIF-90/CeO₂ nanoparticles. (a) DHE staining of brain tissues after different treatments, n = 6. (b) MDA contents of brain tissues after different treatments, n = 6. (c) Total GSH content and GSH/GSSG ratio in brain tissues after different treatments, n = 6. (d) Western blot analysis of NRF2 in the brain tissues of MCAO mice. (e) Relative protein levels of NRF2 in the brain tissues of MCAO mice, n = 4. (f) TUNEL staining of the brain sections of MCAO mice after different treatments, n = 6. Scale bar: 20 μm. (g) Western blot analysis of Bax, Bcl2, cleaved-caspase-3 and caspase 3 levels in the brain tissues of MCAO mice after different treatments. (h) and (i) Relative protein levels of the Bax/Bcl2 ratio (h) and cleaved-caspase-3/caspase 3 ratio (i) in brain tissues, n = 4. (j)–(k) ELISA analysis of IL (interleukin)-6 (j) and IL-1β (k) in the brain, n = 6. (l) Western blot analysis of TET2 in brain tissues. (m) Relative protein levels of TET2 in brain tissues, n = 4. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant.

Moreover, we analyzed the apoptosis and neuroinflammation of brain tissues after treatment. As shown via TUNEL staining on the mice brain tissue sections, a TUNEL-positive rate of nerve cells (38.65 ± 1.93%) in the brain tissue of the MCAO group exhibited the apoptosis of nerve cells in ischemic areas. Administration of ZIF-90/CeO₂ exhibited a slightly lower apoptotic rate of brain tissues, which is attributed to the antioxidative property of ZIF-90/CeO₂. Compared with miR-210@liposome, the miR-210@ZIF-90 nanoparticles could largely decrease the number of TUNEL-positive cells, indicating the reduced apoptosis rate of neurons (Fig. 6(f)). This reveals that ZIF-90 helps the liposomes in improving the therapeutic efficacy of miR-210 inhibitors. In addition to CeO₂, miR-210@ZIF-90/CeO₂ nanoparticles further inhibited apoptosis of the nerve cells in brain tissues with the lowest TUNEL-positive rate of 6.43 ± 1.30%. We examined the expression of apoptosis-related proteins (Bax, Bcl2, caspase 3 and cleaved-caspase 3) in the brain tissues of mice. An increase in both the Bax/Bcl2 ratio and the cleaved-caspase3/caspase3 ratio suggested the induction of apoptosis in the brain tissue of the mice MCAO model (Fig. 6(g)–(i)). Consistent with the results shown in Fig. 6(f), the miR-210@ZIF-90/CeO₂ nanoparticles exerted anti-apoptotic effects in protecting nerve cells from cerebral I/R injury.

As neuroinflammation and brain injury are two sequelae following ischemic stroke, we quantified the levels of the key proinflammatory cytokines IL-6 and IL-1β in the brain tissues 24 h after MCAO. MiR-210@ZIF-90/CeO₂ nanoparticles diminished the MCAO-mediated IL-6 and IL-1β levels, which were the lowest among all groups except the sham group (Fig. 6(j) and (k)). This suggested that miR-210@ZIF-90/CeO₂ nanoparticles exerted anti-neuroinflammatory effects via downregulation of pro-inflammatory cytokines (IL-6 and IL-1β). Based on the previous findings that miR-210 regulates proinflammatory cytokines via targeting TET2,^43,44 we examined the expression levels of TET2 in the brain tissues of mice by western blot analysis. The expression level of TET2 in the brain tissues from the MCAO group was significantly reduced when compared to the sham group, which is associated with the production of IL-6 and IL-1β. As indicated in Fig. 6(l) and (m), both miR-210@ZIF-90/CeO₂ and miR-210@ZIF-90 nanoparticles decreased the expression of TET2, showing their capability of suppressing proinflammatory cytokines. These results revealed that miR-210@ZIF-90/CeO₂ not only decreased the ROS and apoptosis levels of ischemic brain tissues, but also inhibited the inflammation response after cerebral stroke.

2.6. Biocompatibility of miR-210@ZIF-90/CeO₂ nanoparticles

Before clinical application, all nanomedicines should be assessed for their biocompatibility in vivo. In this work, we evaluated the toxic side effects of miR-210@ZIF-90/CeO₂ during the therapy of cerebral I/R injury. We collected the serum from mice in different groups after 24 h of cerebral I/R to analyze alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), uric acid (UA), creatinine (Cr) and blood urea nitrogen (BUN). No significant difference in these indicators was found among the MCAO + miR-210@ZIF-90/CeO₂ group, sham group and MCAO group, revealing that ZIF-90/CeO₂ nanoparticles did not lead to acute injury in the liver and kidney (Fig. 7(a)). Next, we examined pathological sections of the lung, heart, kidney, spleen, and liver by H&E staining, which showed no cell death and inflammatory cell infiltration in these tissues after administration, suggesting their biosafety to other organs (Fig. 7(b)). Therefore, miR-210@ZIF-90/CeO₂ nanoparticles exhibited satisfying biocompatibility.

Fig. 7 Biocompatibility of miR-210@ZIF-90/CeO₂ in mice with cerebral I/RI. (a) Activities of typical biochemical markers of the liver, and renal functions for mice after treatment, n = 6. (b) H&E-stained tissue slices from mice after treatment, scale bar: 100 μm. ns, not significant.

3. Conclusion

Since clinical approaches for cerebral I/R injury have not achieved satisfactory efficacy, our work rationally applied engineered nanozymes to deliver miRNA in the therapy of ischemic stroke. With the loaded ceria nanozymes, the integrated nanozymes (ZIF-90/CeO₂) possessed multienzyme-like catalytic and antioxidative activities, which effectively eliminated ROS in brain tissues after cerebral I/R injury. The easy degradation of miRNA in vivo is one of the most important problems toward clinical translation. Attributed to the inherent proton sponge effect, ZIF-90/CeO₂ protects the miRNA from degradation in lysosomes and maintains their intracellular bioactivity. As a result, the released miR-210 inhibitors target TET2 and protect against neuroinflammation by suppressing the production of key proinflammatory cytokines. Moreover, miR-210@ZIF-90/CeO₂ exhibited excellent biocompatibility. Collectively, miR-210@ZIF-90/CeO₂ hindered the lipid peroxidation in brain tissues of MCAO model mice, thereby alleviating the oxidative damage and apoptosis of neurons in brain tissue without causing any side effects in other normal tissues. This work validates the protective role of ZIF-90/CeO₂ for the delivery of miR-210 inhibitors and antioxidative roles in damaged ischemic areas, revealing the neuroprotective application mechanisms against I/R injury in ischemic stroke.

4. Materials and methods

4.1. Reagents

Imidazole-2-formaldehyde was purchased from Aladdin (Shanghai, China). Zinc acetate and cerous acetate (Ce(AC)₃) were purchased from MACKLIN (Shanghai, China). Trypsin, Dulbecco's modified Eagle's medium (DMEM), and fetal bovine serum (FBS) were purchased from Gino Biological Pharmaceutical Technology Co., Ltd (Hangzhou, China). Dimethylformamide (DMF), ethanol, and acids were provided by Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). MiR-210 inhibitors were purchased from Ribo Bio Co., Ltd (Guangzhou, China).

4.2. Characterization

TEM (HT7800, Hitachi, Japan), SEM (Sigma 300, ZEISS, German), HR-TEM (JEM-F200, JEOL, Japan), XRD (Rigaku Smart Lab SE, Japan), FTIR spectroscopy (VERTEX, Bruker Hongkong Limited, German), UV photoelectron spectroscopy (VARIAN 100UV-VIS, Varian Australia Pty Ltd, Australia), DLS (Zeta sizer nano ZSE, Malvern, UK), and BET (Micromeritics ASAP 2460, U.S.A) were used for characterization of the nanoparticles.

4.3. Synthesis of CeO₂ nanozymes

25 mL of CTAB (0.1 M) was added to a mixture of 1 mL of K₂PtCl₄ solution (2 mM), 2.5 mL of Ce(AC)₃ (0.2 M) and 21.5 mL of D.I. H₂O, where K₂PtCl₄ was used as a catalyst. This mixture was sealed and heated in an oil bath at 100 °C for 1 h. Finally, the CeO₂ nanozymes were collected and washed with D.I. H₂O twice, followed by lyophilization.

4.4. Synthesis of miR-210@ZIF-90/CeO₂ nanoparticles

100 mg of polyvinylpyrrolidone (PVP) and 3.5 mg of CeO₂ were fully dissolved in D.I. H₂O, and stirred at room temperature for 24 h. After being washed with D.I. H₂O three times, the collected samples were combined with 38.4 mg of imidazole-2 formaldehyde, which was dissolved in 2 mL of DMF, and then stirred for 15 min. Subsequently, 36.696 mg of Zn(CH₃COOH)₂ was dissolved in 2 mL of DMF, which was then added to the above solution and stirred for 5 min. Then, 10 mL of DMF was added to the mixture and stirred at 70 °C for 20 min. The product was collected via centrifugation, washed once with DMF and twice with absolute ethanol, followed by lyophilization.

The miR-210 inhibitors and ZIF-90/CeO₂ were prepared at a concentration of 1 μg μL⁻¹ in DEPC-treated water. These solutions were then mixed at various mass ratios (1 : 5, 1 : 20, 1 : 40, 1 : 50, 1 : 100) and placed on a shaker at 250 rpm for 30 min. The miR-210-cy3 inhibitors were used to prepare miR-210-cy3@ZIF-90/CeO₂ nanoparticles. The synthesis method was the same as that used for the miR-210@ZIF-90/CeO₂ nanoparticles.

4.5. Loading efficacy and release profile of miR-210@ZIF-90/CeO₂ nanoparticles

2 μg of miR-210-cy3 inhibitors and ZIF-90/CeO₂ nanoparticles were mixed at different mass ratio (1 : 5, 1 : 20, 1 : 40, 1 : 50, and 1 : 100). After centrifugation, the unloaded miR-210-cy3 inhibitors in the supernatant were detected by UV-vis spectroscopy. Their concentrations were calculated by the standard absorbance curve shown in Fig. S5 and the absorbance at 570 nm of cy3-miR-210 inhibitors.

The miR-210-cy3@ZIF-90/CeO₂ nanoparticles were individually dispersed in PBS solutions with a pH 5.5. The release amount of miR-210-cy3 inhibitors was detected at different intervals by UV-vis spectroscopy.

4.6. Multienzyme activity of ZIF-90/CeO₂ nanoparticles

The SOD Activity Assay Kit (Dojindo, Japan) was used to detect the SOD-like enzymatic activity of ZIF-90, and ZIF-90/CeO₂ at equal doses. The CAT-like enzymatic activities of ZIF-90 and ZIF-90/CeO₂ were assessed at equivalent doses with a Micro Catalase Activity Assay Kit (Abbkine, USA). The POD-like enzymatic activity of ZIF-90 and ZIF-90/CeO₂ at equivalent doses was measured using a Peroxidase Activity Assay Kit (Nanjing Jiancheng Bioengineering Institute).

4.7. ABTS free radical scavenging assay

The ABTS scavenging assay was performed utilizing a total antioxidant capacity assay kit with the ABTS method (Beyotime, Shanghai) to evaluate the antioxidant capacity of CeO₂, ZIF-90, and ZIF-90/CeO₂ at equal doses.

4.8. DPPH free radical scavenging assay

The DPPH scavenging assay was performed to evaluate the antioxidant capacity of CeO₂, ZIF-90, and ZIF-90/CeO₂ nanoparticles with a DPPH kit (Solarbio, Beijing).

4.9. Cell culture

SH-SY5Y cell cultures were maintained in DMEM containing 20% FBS and 1% penicillin–streptomycin (100 U mL⁻¹ penicillin and 100 μg mL⁻¹ streptomycin), and incubated at 37 °C with 5% CO₂ in a humidified incubator.

4.10. Cellular uptake of miR-210@ZIF-90/CeO₂ nanoparticles

The miR-210-cy3 inhibitors were used for the preparation of miR-210-cy3@ZIF-90/CeO₂ nanoparticles. A total of 8 × 10³ SH-SY5Y cells per well were seeded in an 8-well plate (Millipore). Then, DMEM was replaced with fresh medium containing miR-210-cy3 inhibitor@ZIF-90/CeO₂ nanoparticles (20 μg mL⁻¹, prepared previously at a mass ratio of 1 : 50 of miR-210-cy3 inhibitors to ZIF-90/CeO₂) in each well. The cultured cells were washed twice sequentially with PBS after 4, 16, or 24 h. Subsequently, fluorescent dyes were added to stain the lysosomes for 20 min, followed by visualization under a confocal laser scanning microscope (CLSM, FV3000, Olympus, Germany).

For the other samples, a total of 8 × 10³ SH-SY5Y cells per well were seeded in 8-well plates (Millipore). They were then treated with chlorpromazine (inhibitor of clathrin-mediated endocytosis), wortmannin (inhibitor of clathrin-independent endocytic pathways) and cytochalasin D (inhibitor of phagocytosis and macropinocytosis) for 30 min, followed by 6 h of incubation of miR-210-cy3 inhibitor@ZIF-90/CeO₂ nanoparticles (20 μg mL⁻¹, prepared previously at the mass ratios of 1 : 50 of miR-210-cy3 inhibitors to ZIF-90/CeO₂). Finally, the cultured cells were washed twice sequentially with PBS. Fluorescent dyes were then added to stain the nuclei for 20 min, followed by visualization under a confocal laser scanning microscope (CLSM, FV3000, Olympus, Germany).

4.11. MTT assays

A total of 1.0 × 10⁴ SH-SY5Y cells per well were seeded in 96-well Petri dishes. They were treated with a gradient of ZIF-90, CeO₂, and ZIF-90/CeO₂ nanoparticles for 24 h. Then, absorbance measurements were performed at a wavelength of 490 nm using a microplate reader (Spectra Max iD3, Molecular Devices, China).

4.12. miRNA expression Analysis by qPCR

The miRNA expression was analyzed using qPCR, following the manufacturer's protocol (RiboBio, Guangzhou, China). Total RNA was extracted by the Trizol method from cells and tissue for qPCR. The miRNA quantification was initiated through first-strand cDNA synthesis employing reverse transcriptase and stem-loop structured miRNA-specific primers (RiboBio, Guangzhou, China). The synthesized cDNA was then amplified using SYBR Green PCR Master Mix. qPCR reactions were performed in a 10 μL system, containing 1 μL of cDNA, 5 μL of SYBR Green PCR Master Mix, 0.5 μL of forward and reverse primers, and RNase-free water. The cycling conditions were as follows: 95 °C for 10 min for initial denaturation, followed by 40 cycles of 95 °C for 2 s and 60 °C for 30 s. The relative expression of miRNAs was calculated using the comparative cycle threshold (Ct) method, with 5S ribosomal RNA serving as an internal reference for data normalization.

4.13. Apoptosis analysis by flow cytometry

To establish an in vitro model of OGD/R, SH-SY5Y cells were cultured in 6-well culture plates. The cell apoptosis was measured by using the Annexin V-FITC apoptosis detection kit (Dojindo, Japan), in accordance with the manufacturer's instructions. The flow cytometer (cytoFlex, Beckman, USA) was then used to detect and quantify the apoptosis rate of the cells.

4.14. ROS levels by flow cytometry

To establish an in vitro model of OGD/R, SH-SY5Y cells were cultured in 6-well culture plates. After different treatments for 24 h, the fluorescent ROS probe 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) was incubated with the cells for 30 min. The intracellular ROS levels were ultimately determined through flow cytometric analysis (cytoFlex, Beckman, USA).

4.15. Mitochondrial membrane potential (MMP) by flow cytometry

To establish an in vitro model of OGD/R, SH-SY5Y cells were cultured in 6-well plates. After different treatments, the cells were stained with JC-1 for 20 min. Finally, cellular MMP was analyzed through flow cytometric analysis (cytoFlex, Beckman, USA).

4.16. Mice experiments

Male C57BL/6J mice (8 weeks old) were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd (Beijing, China). All animal experiments were performed following “Regulations on the Administration of Laboratory Animals”, and were approved by the Institutional Animal Care and Use Committee of Shanghai University.

4.17. Mice models with cerebral ischemia/reperfusion (I/R) injury

The mice model with cerebral I/R injury was established by 2 h of middle cerebral artery occlusion (MCAO) using the intraluminal filament method. Briefly, mice were subcutaneously anesthetized with pentobarbital (200 mg kg⁻¹, intraperitoneal injection).

Following careful surgical exposure of the right common carotid artery (CCA), right external carotid artery (ECA), and internal carotid artery (ICA), the ECA was ligated. A silicone-coated nylon filament (diameter: 0.20 ± 0.02 mm) with a rounded tip was introduced into the ICA via the ECA to induce middle cerebral artery (MCA) occlusion. After 2 h of ischemia, the nylon filament was carefully removed to allow reperfusion. Mice in the sham group underwent identical procedures excluding filament insertion.⁴⁴ The left lateral ventricle was targeted for injection, positioned at coordinates 1.1 mm posterior to the bregma, 1.5 mm to the left of the midline, and 2.5 mm below the dura mater. Then, 10 μL of drugs including miR-210@ZIF-90 nanoparticles, ZIF-90/CeO₂ nanoparticles, miR-210@ZIF-90/CeO₂ nanoparticles and miR-210@liposome nanoparticles at a dosage of 0.2 mg mL⁻¹ were injected into the lateral ventricle using a brain stereotaxic injector (L0107-1B, Changzhou, China) and micro syringe pump (TJ-2A, Baoding, China) at a rate of 1 μL min⁻¹, respectively. The sham group received no administration of drugs. We divided the mice into six groups (n = 6 for each group) as follows: sham operation (denoted as sham group), cerebral I/R injury without any treatment (denoted as MCAO group), cerebral I/R injury with miR-210@ZIF-90 treatment (denoted as MCAO + miR-210@ZIF-90 group), cerebral I/R injury with ZIF-90/CeO₂ treatment (denoted as MCAO + ZIF-90/CeO₂ group), cerebral I/R injury with miR-210@ZIF-90/CeO₂ treatment (denoted as MCAO + miR-210@ZIF-90/CeO₂ group), and cerebral I/R injury with miR-210@liposome treatment (denoted as MCAO + miR-210@liposome group).

4.18. 2,3,5-Triphenyl tetrazolium chloride (TTC) staining for brain infarct volume

The whole brain tissue of mice was rapidly removed, and the tissue was cut into 6 slices of 2 mm of coronal brain sections after 24 h of reperfusion. The slices were incubated in 1% TTC (Sigma Aldrich, Germany) in a water bath at 37 °C for 30 min under dark conditions. Subsequently, the stained tissue samples were fixed in 4% paraformaldehyde overnight. Image-Pro Plus 6.0 software was used to analyze the infarcted area, and the infarct volume was calculated based on the section thickness.

4.19. TUNEL staining for apoptosis analysis

The brain tissue slices were fixed in 4% paraformaldehyde, permeabilization using 0.1% Triton X-100, and blocked with 5% bovine serum albumin (BSA) for 1 h at ambient temperature, followed by incubation with Cy3-AffiniPure goat anti-mouse IgG (H + L) the next day. Subsequently, the TUNEL staining kit (Cat No. C1090, Beyotime, China) was used to detect apoptotic cells. Images of the brain slice were acquired using CLSM to visualize TUNEL-positive cells (green) and Hoechst-positive cells (blue).

4.20. Immunofluorescence staining

The brain tissue slices were fixed in 4% paraformaldehyde, permeabilized using 0.1% Triton X-100, and blocked with 5% BSA for 1 h at ambient temperature. Primary antibodies against Iba-1 (A19776, Abclonal) and GFAP (A19058, Abclonal) were incubated overnight at 4 °C, followed by incubation with Cy3-AffiniPure goat anti-mouse IgG (H + L) (Jackson Immunoresearch) the following day. Images of the brain slice were acquired using CLSM to visualize Iba-1-positive cells (red), GFAP-positive cells (red), and Hoechst-positive cells (blue).

4.21. Assessment of neurological deficits

Neurological deficits in mice were assessed by using a modified Longa method. Briefly, the modified Longa scale used the following criteria: 0, no obvious signs of neurological deficits (similar to the sham mice); 1, contralateral flexion of the forelimb during tail lifting, suggesting a slight neurological deficit; 2, significant tail collision during free walking, with the mouse turning to the left, suggesting a neurological deficit; 3, difficulty walking and hemiplegic behavior, showing severe neurological deficits; 4, inability to walk freely and significantly reduced level of consciousness. Behavioral assessments were conducted by investigators who had prior knowledge of the experimental group allocation.

4.22. Dihydroethidium (DHE) staining for ROS levels

The DHE working solution was added dropwise to brain tissue cryosections obtained from mice after different treatments. These sections were incubated for 30 min at 4 °C under dark conditions, allowing DHE to fully react with the intracellular ROS. Subsequently, the stained brain tissue sections were observed under a fluorescence microscope and the intensity of the fluorescent signals was analyzed to assess the ROS levels within the brain tissue.

4.23. Western blot analysis

Proteins were extracted from the infarcted brain sites and then denatured by boiling for 5 min, followed by separation by SDS-PAGE and transfer to PVDF membranes. Following blocking with 5% skim milk, the membranes were incubated with primary antibodies as follows: Bax (A12009, Abclonal), Bcl2 (A19693, Abclonal), Caspase3 (9662, Cell Signaling Technology), β-actin (AC006, Abclonal), TET2 Rabbit pAb (A5682, Abclonal), and NRF2 Rabbit mAb (A21176, Abclonal).

4.24. Detection of antioxidant enzymes and inflammatory factors in brain tissue

The infarct brain sites were obtained as a homogenate and their protein was extracted. The contents of MDA and antioxidant enzymes of GSH/GSSG in the brain tissue of the MCAO mice were examined using the lipid peroxidation MDA assay kit and GSH and GSSG assay kit (Beyotime, China), in accordance with the specification method. The concentration of inflammatory factors including IL-6, and IL-1β was also examined using the ELISA kit (COIBO BIO, Shanghai), in accordance with the specifications.

4.25. Blood biochemical test

The serum samples from mice after different treatments were assayed for aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), creatinine (Cr), uric acid (UA) and blood urea nitrogen (BUN) using a microplate reader (SpectraMax iD3, Molecular Devices, China) by commercially available kits (Nanjing Jiancheng Bioengineering Institute).

4.26. Hematoxylin & Eosin (HE) staining

HE staining was conducted for the morphological structure evaluation of the brain, heart, liver, lung, spleen, and kidney in mice after different treatments.

4.27. Statistical analysis

All data were analyzed by GraphPad Prism, and presented as the means ± standard deviation. An unpaired, two-tailed Student's t-test was used to compare the two groups. One-way ANOVA test followed by Bonferroni or Dunnett T3 test, or two-way ANOVA test followed by Tukey post hoc test was performed as appropriate to compare differences among multiple groups. The statistical significance was set as p < 0.05.

Author contributions

Linlin Zhao conducted most of the experiments and data analysis, and wrote the manuscript. Meiyu Hu contributed to the refinement of the research concept and performed the animal experiments. Xiaohang Yin and Hao Yang constructed the mice models, and were responsible for the imaging studies. Liyun Zhu was responsible for the HR-TEM and TEM characterization. Pingyuan Sun performed SEM measurements and image analysis. Xiya Wang and Songwei Ai contributed to the improvement of the figure preparation. Yonghui Zheng and Genjie Li provided revisions to the manuscript. Tingting Yang, Xuerui Chen and Jingyu Zhang supervised the research, provided critical feedback and revisions to the manuscript, and secured funding. All authors participated in the discussion and revision of the manuscript.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Additional data related to this paper may be requested from the corresponding author.

Experimental details, characterization data and performance comparison are provided. See DOI: https://doi.org/10.1039/d5mh00875a

Acknowledgements

This work was supported by the National Natural Science Foundation of China (22373064, 62272138, 22003038, 82401851), Horizontal Project Funding (HX2020-12), China Postdoctoral Science Foundation (2023M742205), and Postdoctoral Fellowship Program of CPSF (GZB2023039).

References

1. J. H. DeLong, S. N. Ohashi, K. C. O'Connor and L. H. Sansing, Semin. Immunopathol., 2022, 44, 625–648.
2. Y. Zhao, X. Zhang, X. Chen and Y. Wei, Int. J. Mol. Med., 2022, 49, 15.
3. Y. Zhang, H. Zhang, F. Zhao, Z. Jiang, Y. Cui, M. Ou, L. Mei and Q. Wang, Acta Pharm. Sin. B, 2023, 13, 5107–5120.
4. C. M. Stinear, C. E. Lang, S. Zeiler and W. D. Byblow, Lancet Neurol., 2020, 19, 348–360.
5. Z. Mao, L. Tian, J. Liu, Q. Wu, N. Wang, G. Wang, Y. Wang and S. Seto, Phytomedicin, 2022, 101, 154111.
6. H. Lu, H. Zeng, W. Wei, Y. Chen, Z. Zhou, X. Ning and P. Hu, Theranostics, 2024, 14, 7589–7603.
7. D. Kiltschewskij and M. J. Cairns, Nucleic Acids Res., 2019, 47, 533–545.
8. Q. Zheng and W. Hou, Mol. Med. Rep., 2021, 24, 583.
9. S. E. Khoshnam, W. Winlow and M. Farzaneh, J. Neuropathol. Exp. Neurol., 2017, 76, 548–561.
10. W. Xu, L. Gao, J. Zheng, T. Li, A. Shao, C. Reis, S. Chen and J. Zhang, Cell Transplant., 2018, 27, 1778–1788.
11. H. Tao, L. Dong, X. Shan, L. Li and H. Chen, Int. Immunopharmacol., 2023, 121, 110504.
12. K. Sun, J. Zhang, Q. Yang, J. Zhu, X. Zhang, K. Wu, Z. Li, W. Xie and X. Luo, Pfluegers Arch., 2022, 474, 343–353.
13. J. Wang, Y. Zhang and F. Xu, Exp. Ther. Med., 2018, 15, 1263–1268.
14. J. J. Sun, X. Y. Zhang, X. D. Qin, J. Zhang, M. X. Wang and J. B. Yang, Eur. Rev. Med. Pharmacol. Sci., 2019, 23, 2548–2554.
15. Y. Li, R. Song, G. Shen, L. Huang, D. Xiao, Q. Ma and L. Zhang, Stroke, 2023, 54, 857–867.
16. Q. Ma, C. Dasgupta, G. Shen, Y. Li and L. Zhang, J. Neuroinflam, 2021, 18, 6.
17. B. Li, C. Dasgupta, L. Huang, X. Meng and L. Zhang, Cell. Mol. Immunol., 2020, 17, 976–991.
18. S. M. Yerrapragada, H. Sawant, S. Chen, T. Bihl, J. Wang and J. C. Bihl, Exp. Neurol., 2022, 358, 114211.
19. X. Chen, H. Chen, L. Zhu, M. Zeng, T. Wang, C. Su, G. Vulugundam, P. Gokulnath, G. Li, X. Wang, J. Yao, J. Li, D. Cretoiu, Z. Chen and Y. Bei, ACS Nano, 2024, 18, 19470.
20. X. J. Deng, J. H. Wang, S. S. Yu, S. Y. Tan, T. T. Yu, Q. X. Xu, N. H. Chen, S. Q. Zhang, M. R. Zhang, K. Hu and Z. Y. Xiao, Exploration, 2024, 4, 20230090.
21. L. Gao, H. Wei, S. Dong and X. Yan, Adv. Mater., 2024, 36, e2305249.
22. Y. Yang, Z. Li, X. Fan, C. Jiang, J. Wang, Y. Rastegar-Kashkooli, T. J. Wang, J. Wang, M. Wang, N. Cheng, X. Yuan, X. Chen, B. Jiang and J. Wang, ACS Nano, 2024, 18, 16450–16467.
23. Q. Bao, P. Hu, Y. Xu, T. Cheng, C. Wei, L. Pan and J. Shi, ACS Nano, 2018, 12, 6794–6805.
24. H. Yang, P. Lin, B. Zhang, F. Li and D. Ling, Adv. Mater., 2024, 36, e2410031.
25. X. Chen, H. Chen, L. Zhu, Q. Li, P. Sun, M. Spanos, C. Su, X. Wang, L. Zhao, R. Gui, T. Wang, X. Wang, X. Zhou and Z. Chen, Small, 2025, 2502778.
26. E. Casals, M. Zeng, M. Parra-Robert, G. Fernández-Varo, M. Morales-Ruiz, W. Jiménez, V. Puntes and G. Casals, Smal, 2020, 16, e1907322.
27. Y. G. Kim, Y. Lee, N. Lee, M. Soh, D. Kim and T. Hyeon, Adv. Mater., 2024, 36, e2210819.
28. Y. Liu, H. Zhao, R. Niu, B. Zhang, B. T. G. Lim, S. Song, Y. Wang, H. Zhang and Y. Zhao, Chem, 2024, 10, 1183–1203.
29. M. Li, Y. Tao, Y. Shu, J. R. LaRochelle, A. Steinauer, D. Thompson, A. Schepartz, Z. Y. Chen and D. R. Liu, J. Am. Chem. Soc., 2015, 137, 14084–14093.
30. C. Fan, Y. Tang, H. Wang, Y. Huang, F. Xu, Y. Yang, Y. Huang, W. Rong and Y. Lin, Nanoscale, 2022, 14, 7985–7990.
31. I. O. Zakharova, L. V. Bayunova and N. F. Avrova, Curr. Issues Mol. Biol., 2025, 47, 462.
32. X. Chen, X. Yin, L. Zhan, J. Zhang, Y. Zhang, Y. Wu, J. Ju, Y. Li, Q. Xue, X. Wang, C. Li, R. L. Reis and Y. Wang, Adv. Funct. Mater., 2021, 32, 2108603.
33. L. He, G. Huang, H. Liu, C. Sang, X. Liu and T. Chen, Sci. Adv., 2020, 6, eaay9751.
34. S. Mishra, J. D. Heidel, P. Webster and M. E. Davis, J. Controlled Release, 2006, 116, 179–191.
35. S. K. Alsaiari, S. Patil, M. Alyami, K. O. Alamoudi, F. A. Aleisa, J. S. Merzaban, M. Li and N. M. Khashab, J. Am. Chem. Soc., 2018, 140, 143–146.
36. P. Chen, X. Liu, C. Gu, P. Zhong, N. Song, M. Li, Z. Dai, X. Fang, Z. Liu, J. Zhang, R. Tang, S. Fan and X. Lin, Nature, 2022, 612, 546–554.
37. J. Liao, Y. Li, L. Fan, Y. Sun, Z. Gu, Q. Q. Xu, Y. Wang, L. Xiong, K. Xiao, Z. S. Chen, Z. Ma, C. Zhang, T. Wang and Y. Lu, ACS Nano, 2024, 18, 8979–8995.
38. S. Orellana-Urzúa, I. Rojas, L. Líbano and R. Rodrigo, Curr. Pharm. Des., 2020, 26, 4246–4260.
39. N. M. Alpert, M. Pelletier-Galarneau, Y. Petibon, M. D. Normandin and G. El Fakhri, Nature, 2020, 583, E17–E18.
40. C. C. Chio, J. W. Lin, H. A. Cheng, W. T. Chiu, Y. H. Wang, J. J. Wang, C. H. Hsing and R. M. Chen, Arch. Toxicol., 2013, 87, 459–468.
41. D. Tsikas, Anal. Biochem., 2017, 524, 13–30.
42. S. Sentellas, O. Morales-Ibanez, M. Zanuy and J. J. Albertí, Toxicol. In Vitro, 2014, 28, 1006–1015.
43. Z. Miao, Y. He, N. Xin, M. Sun, L. Chen, L. Lin, J. Li, J. Kong, P. Jin and X. Xu, Hum. Mol. Genet., 2015, 24, 5855–5866.
44. A. Carrillo-Jimenez, Ö. Deniz, M. V. Niklison-Chirou, R. Ruiz, K. Bezerra-Salomão, V. Stratoulias, R. Amouroux, P. K. Yip, A. Vilalta, M. Cheray, A. M. Scott-Egerton, E. Rivas, K. Tayara, I. García-Domínguez, J. Garcia-Revilla, J. C. Fernandez-Martin, A. M. Espinosa-Oliva, X. Shen, P. St George-Hyslop, G. C. Brown, P. Hajkova, B. Joseph, J. L. Venero, M. R. Branco and M. A. Burguillos, Cell Rep., 2019, 29, 697–713.

---

Footnote

† These authors contributed equally to this work.

---