3.1.4. Inorganic Nanoparticles as ROS Scavengers

Studies have showed that metallic NPs can function as scavengers of ROS. ROS includes superoxide anion, hydrogen peroxide, and hydroxyl radical, and they are generated during ischemic stroke. ROS is involved with oxidative tissue damage that is the mechanism responsible for brain injury in ischemic stroke.

It is reported that ceria nanoparticles can scavenge free radicals by reversibly reacting with oxygen to Ce4<sup>+</sup> (oxidized) species from Ce3<sup>+</sup> (reduced) species [62]. Kim et al. firstly reported that ceria nanoparticles reduced ROS production and prevented cell apoptosis, thus protecting the brain from ischemic damage. Although ceria nanoparticles are effective to scavenge ROS, the integrity of BBB is the barrier to prevent their brain deposition. Bao and coworkers developed ceria nanoparticles (E-A/P-CeO2) modified with Angiopep-2 (ANG) and poly(ethylene glycol), which were able to cross BBB via receptor-mediated transcytosis (Figure 4A). In vitro transmigration assay (Figure 4B) showed that E-A/P-CeO2 could cross the brain capillary endothelial cells (BCECs) compared to P-CeO2 [63]. The BBB crossing capability of E-A/P-CeO2 was also confirmed on healthy rats by determining the ratio of ceria nanoparticles in brain tissue and injected (Figure 4C). Finally, results showed that E-A/P-CeO2 displayed ROS scavenging ability and decreased the infarct volume on rats (Figure 4D). The results showed that ceria nanoparticles might be promising in treating I/R injury.

Platinum nanoparticles (nPt) are novel ROS scavengers because the large surface area and the high electron density can potentiate their catalytic activity to quench ROS [75]. Takamiya et al. investigated whether nPt (2–3 nm) has the neuroprotection effect against I/R injury in the mouse model of transient middle cerebral artery occlusion (tMCAO) [64]. The results showed that treatment with nPt ameliorated the generation of superoxide via reduction of hydroethidine in the cerebral cortex and decreased the mouse infarct volume. The same group published another work to investigate whether nPt could ameliorate tissue plasminogen activator (tPA)-related ischemic injury since tPA treatment may upregulate the expression of MMP-9, which exacerbates cerebral infarction through low-density lipoprotein receptor-related protein [65]. They found that nPt decreased the MMP-9 activity and ameliorated the disrupted neurovascular unit (NVU) after tMCAO. Those results further strengthened that nPt could be combined with tPA reperfusion treatment to treat ischemic stroke patients.

**Figure 4.** Edaravone-loaded ceria nanoparticles cross the blood brain barrier and protect the brain from ischemic stroke. (**A**) The nanoplatform is comprised of ceria nanoparticle core, PEG shell, and ANG as targeting peptide. (**a**) Transition between cerium(III) and cerium(IV) species in ceria NPs; (**b**) main components of the nanoparticle system; (**c**) synthetic procedure for E-A/P-CeO2; (**d**) receptor-mediated (ANG-LRP) endocytosis of E-A/P-CeO2. (**B**) Transmigrated amount of P-CeO2 and E-A/P-CeO2 in in vitro transmigration model. Free ANG was used as the blocking agent in group 3. (**C**) Concentrations (μg Ce/g brain tissue) of ceria nanoparticles in normal brain tissue. The injection dose was 0.5 mg/kg. (**D**) Representative TTC (triphenyl tetrazolium chloride)-stained brain sections after different treatments within 24 h of stroke. Data represent mean ± standard deviation (SD). \* indicates *p* < 0.05. Reproduced with permission [63]. Copyright 2018, American Chemical Society.

#### *3.2. Delivery of Anti-inflammatory Reagents*

Inflammation plays a central role in ischemic stroke, thus delivering anti-inflammatory drugs may be a novel strategy to treat ischemic stroke. Exosomes are endogenous vesicles made from cell membrane with cellular targeting features, therefore, they may be drug delivery tools [76–78]. In addition, exosomes have other properties, such as innate stability, biodegradability, low immunogenicity, and ability to cross BBB. Tian et al. used mesenchymal stromal cell (MSC)-derived exosomes conjugated

with a c(RGDyK) peptide (called cRGD-Exo) to deliver curcumin to the ischemic brain [66]. Due to the high specific affinity of c(RGDyK) peptide to integrin αvβ3, highly expressed on cerebral vascular endothelial cells of ischemic tissues, cRGD-Exo was significantly accumulated in the stroke lesion. Administration of curcumin with exosomes (called cRGD-Exo-cur) suppressed the inflammatory response via the NF-κB pathway and protected the brain from ischemic injury.

Dexamethasone is an anti-inflammatory drug and can reduce proinflammatory cytokines, which prevent cellular damage in the ischemic brain [79]. Lee and co-workers incorporated dexamethasone to the hydrophobic component of a small peptide, R3V6 peptide (with a 3-arginine block and a 6-valine block), to form a stable micelle (named, R3V6-Dexa) used for gene delivery. R3V6-Dexa was loaded with a plasmid DNA, which expressed a heme oxygenase-1 (HO-1) as an antioxidant agent, and the complex of pSV-HO-1/R3V6-Dexa could target ischemic stroke tissues [67]. The authors stereotaxically injected the complex into rats and found that HO-1 expression was significantly increased. In addition, when mice were treated with the pSV-HO-1/R3V6-Dexa complex, the infarction size in the brain was significantly reduced compared to control groups.

Increased leukocyte infiltration is another factor to damage the brain in stroke. Within minutes after ischemia, adhesion molecules, such as P-selectin, are highly expressed on endothelial cells, and proinflammatory signals are rapidly generated [22]. Another strategy is to control the infiltration of neutrophils to alleviate the I/R injury in ischemic stroke. Targeting neutrophils may be a novel approach in treating ischemic stroke, but the methods in anti-neutrophil therapies and neutrophil depletion have failed in clinical studies. This failure may be associated with the complex signaling pathways and numerous receptors involved in neutrophil transmigration. For example, deactivating one receptor, such as anti-ICAM-1 therapy, may not be sufficient because neutrophil infiltration is involved with various types of surface proteins between neutrophils and endothelial cells [80,81]. Neutrophil depletion could be an effective approach, but systemic elimination of neutrophils can make the host vulnerable to bacterial and virus infections [82], and even impair other immune cells, such as natural killer cells [83].

Neutrophil infiltration is regulated by interactions between neutrophils and endothelial cells. Therefore, delivering drugs to the cerebral endothelium may effectively inhibit neutrophil infiltration without interfering neutrophils [77]. Dong et al. reported neutrophil cell membrane-derived nanovesicles (HVs), which can target inflamed endothelium at I/R injury sites and deliver therapeutics to treat the mouse I/R injury [68]. The nanovesicles were generated from differentiated HL-60 cells (neutrophil-like cells) using nitrogen cavitation. The nitrogen cavitation approach was used to disrupt cell membrane to eliminate nuclei and cytosols of cells. Figure 5A shows the scheme of experimental design. The TEM image showed the liposome-like structure of nanovesicles made from HL-60 cells and the size was 200 nm in diameter. In vivo imaging system (IVIS) and confocal microscopy (Figure 5B) showed that HVs (nanovesicles made from differentiated HL-60 cells) were specifically accumulated in the injured half of the brain rather than in the normal brain. To visualize how HVs interacted with brain vasculature, a cranial window was established and intravital microscopy was performed in live mice (Figure 5C). The real-time visualization images strongly indicated that HVs can specifically target ischemic vasculature. To examine the delivery of therapeutics with nanovesicles, Resolvin D2 (RvD2) was loaded into the membrane of nanovesicles because RvD2 is a new lipid mediator to resolve inflammatory responses [84]. After Resolvin D2-loaded nanovesicles (RvD2-HVs) were intravenously injected into mice, reduced neutrophil infiltration was observed using intravital microcopy in live mice, and brain homogenates also confirmed this observation (Figure 5D). Subsequently, the level cytokines, such as TNF-α (Figure 5E), were decreased after RvD2-HVs were administered. The data indicated that RvD2-HVs could alleviate inflammation responses in ischemic stroke. The diminished inflammatory responses reduced the infarction sizes and prevented neurological damage from ischemic stroke (Figure 5F).

**Figure 5.** Neutrophil membrane-derived nanovesicles target ischemic endothelium and deliver therapeutics for ischemic stroke treatment. (**A**) A concept of nanovesicles binding to endothelium at ischemic/reperfusion injury sites and alleviating the I/R injury. Resolvin D2-loaded nanovesicles were prepared by: (1) nitrogen cavitation to break the cells; (2) purification; (3) Resolvin D2 loading. (**B**) Nanovesicles (HVs) specifically accumulated in I/R damaged half of brain. (**C**) Intravital microscopy images of HVs (red) specifically bind to I/R endothelium (top panel), instead of normal endothelium (bottom panel). Scale bar = 20 μm. (**D**) Myeloperoxidase (MPO) activity in the injured half of brain after different treatment, which indicates the neutrophil infiltration condition. (**E**) Level of TNF-α in the injured half of brain after different treatment. (**F**) TTC staining of brain sections in different groups. Data represent mean ± standard deviation (SD). \* *p* < 0.05 and \*\*\* *p* < 0.005. Reproduced with permission [68]. Copyright 2019, American Chemical Society.
