Biomimetic, ROS-detonable nanoclusters — A multimodal nanoplatform for anti-restenotic therapy

Abstract

The long-term success of endovascular intervention has long been overshadowed by vessel re-occlusion, also known as restenosis. Mainstream anti-restenotic devices, such as drug-eluting stent (DES) and drug-coated balloon (DCB), were recently shown with suboptimal performances and life-threatening complications, thereby underpinning the urgent need for alternative strategies with enhanced efficacy and safety profile. In our current study, we engineered a multimodal nanocluster formed by self-assembly of unimolecular nanoparticles and surface coated with platelet membrane, specifically tailored for precision drug delivery in endovascular applications. More specifically, it incorporates the combined merits of platelet membrane coating (lesion targetability and biocompatibility), reactive oxygen species (ROS)-detonable “cluster-bomb” chemistry (to trigger the large-to-small size transition at the target site, thereby achieving longer circulation time and higher tissue penetration), and sustained drug release. Using RVX-208 (an emerging anti-restenotic drug under clinical trials) as the model payload, we demonstrated the superior performances of our nanocluster over conventional poly (lactic-co-glycolic acid) (PLGA) nanoparticle. In cultured vascular smooth muscle cell (VSMC), the drug-loaded nanocluster induced effective inhibition of proliferation and protective gene expression (e.g., APOA-I) with a significantly reduced dosage of RVX-208 (1 μM). In a rat model of balloon angioplasty, intravenous injection of Cy5.5-tagged nanocluster led to greater lesion targetability, improved biodistribution, and deeper penetration into injured vessel walls featuring enriched ROS. Moreover, in contrast to either free drug solution or drug-loaded PLGA nanoparticle formulation, a single injection with the drug-loaded nanocluster (10 mg/kg of RVX-208) was sufficient to substantially mitigate restenosis. Additionally, this nanocluster also demonstrated biocompatibility according to in vitro cytotoxicity assay and in vivo histological and tissue qPCR analysis. Overall, our multimodal nanocluster offers improved targetability, tissue penetration, and ROS-responsive release over conventional nanoparticles, therefore making it a highly promising platform for development of next-generation endovascular therapies.

1. Introduction

Cardiovascular disease is the number one cause of death in developed countries [1]. Endovascular intervention strategies such as angioplasty and stenting remain the gold standard to reconstruct occluded vessels [2]. Unfortunately, these reconstructed vessels will re-occlude over time; this re-occlusion is referred to as restenosis [3]. Restenosis occurs primarily due to the formation of neointima in the vessel wall, a process termed intimal hyperplasia (IH) featuring the overproliferation and phenotypic transformation of vascular smooth muscle cells (VSMC) [4].

Drug-eluting stents (DES) and drug-coated balloons (DCB) are currently the standard of care for restenosis [5]. However, they do not completely prevent IH and in fact, worsen thrombogenic risks. In-stent restenosis occurs in ~10–23% of coronary DES applications, and up to 75% of those placed in the peripheral vasculature [6]. Moreover, stent thrombosis has become a chief concern, as it causes sudden death in up to 40% of those patients [7–9]. Increasing evidence reveals that the permanent presence of metal stents leads to sustained mechanical injury and hemodynamic disturbance, which ultimately nurtures a prorestenotic and prothrombogenic local endovascular milieu.⁷ Therefore, the ultimate outcome of DES implantation for restenosis prevention, paradoxically, can be the persistence of in-stent restenosis and exacerbated thrombogenic risks. To address the downsides of DES, alternative stent-free strategies such as drug-eluting balloons and absorbable stents have been clinically assessed. However, recent studies have demonstrated increased mortality and significant safety concerns associated with these alternatives, and Food and Drug Administration (FDA) has since issued multiple warnings [9,10]. Therefore, there is a clear and pressing clinical need for a stent-free anti-restenotic therapy.

In recent years, nanomedicine has been considered as a promising solution for targeted drug delivery in cardiovascular applications. In the case of anti-restenotic therapy, various types of drug nanocarriers have been documented, including inorganic [11,12], polymeric [13–17], liposomal nanoparticles [18,19], etc. Biointerfacing with cell membrane coatings is an emerging concept in precision medicine, which can facilitate drug delivery via integrating the biomimetic features of cell membrane with the versatile functions of nanocores [20]. Platelets are of particular relevance to endovascular interventions. Upon angioplasty or other endovascular injuries, platelet can autonomously home to the lesional vessel through the interaction between platelet membrane surface proteins (e.g., glycoprotein VI) and the exposed sub-endothelial matrix in the vessel wall [21,22]. Indeed, several studies, including ours, have demonstrated the feasibility of platelet-inspired biomimetic coating in improving nanoparticle performances, including lesion-targeting property after angioplasty [23,24].

In addition to lesion targetability, the ideal drug delivery system should also encompass enhanced tissue penetrability [25]. The tissue penetration capacity of existing nanoplatforms is often suboptimal for cardiovascular applications [26,27]. Unlike the “leaky”, thin-layered microvasculature in the case of tumor [28], the major arterial and aortic vessels are comprised of multi-layered smooth muscle, elastic laminae, and significant amount of extracellular matrix (ECM) — all of which impede the penetration of nanoparticles and hence the delivery of therapeutic payloads [27]. This is particularly concerning in diseased vasculatures, where IH and ECM over-accumulation mount extra physical barricade to nanoparticle infiltration. It is increasingly recognized that particle size plays a vital role in its biodistribution and tissue penetration, yet oftentimes these two features are not attainable simultaneously [29]. As summarized by Randa Zein et al. [30], nanoparticles with a size ranging between 100 and 200 nm typically possess longer circulation time due to reduced sequestration in spleen and liver, but their tissue penetration capacity is somewhat limited [31,32]. In contrast, smaller-sized nanoparticles offer much deeper penetration. Therefore, an ideal delivery system should have an initial size ranging from 100 to 200 nm to achieve longer circulation time, but it may be switched to smaller particle size once reaching the target lesions to facilitate tissue penetration for more efficient drug delivery.

To this end, we have developed a highly innovative, multimodal nanoplatform with the following unique features that are most desirable in anti-restenotic application: (1) The biomimetic surface coating with platelet membrane grants the vascular lesion-targeting property and biocompatibility; (2) The size-tunable feature (from 190 nm nanoclusters into 40 nm unimolecular nanoparticles) ensures the optimal balance between biodistribution and tissue penetration; (3) An reactive oxygen species (ROS)-detonable chemistry enables the lesion-specific de-clustering of the nanoclusters upon “landing” in the ROS-enriched vascular lesions, thus further enhancing the targetability and uptake; (4) Each unimolecular nanoparticle is equipped with high drug loading capacity as well as controlled release through its H40-polylactide (H40-PLA-OH) hydrophobic functional core.

In the current study, we sought to establish the pre-clinical utility of this multimodal nanocluster for anti-restenotic therapy. To highlight the unique properties of our nanoplatform, we sought to compare it to a conventional PLGA nanoparticle of the same overall size and platelet membrane coating but lacking the ROS-responsive size-tunable de-clustering feature. Utilizing a rat model of carotid balloon angioplasty, the platelet membrane-coated, ROS-detonable nanocluster (NC) demonstrated superior lesion-targeting specificity, biodistribution, and tissue penetrating capacity in injured arterial walls. RVX-208 (abbreviated as RVX), a small-molecule inhibitor that selectively blocks the bromo and extraterminal (BET) domain family of epigenetic reader proteins, was then chosen as the payload for therapeutic evaluation [33]. The role of BET as intervention targets in restenosis was first established in our prior studies. Futhermore, considering the ongoing phase III clinical trial of a daily RVX regimen in management of cardiovascular complications, it is therefore of considerable clinical implication to determine if our innovative nanocluster formulation could significantly improve its pharmacological and therapeutic performances [34,35]. Remarkably, while low dose RVX (1 μM in vitro, 10 mg/kg in vivo) in free solution or delivered through platelet membrane-coated poly(lactic-co-glycolic acid) nanoparticle (PLGA NP) failed to elicit any significant changes in VSMC behaviors and restenosis, the RVX-loaded nanocluster, in stark contrast, led to significant inhibition of VSMC proliferation and IH (hence decreasing restenosis).

2. Materials and methods

2.1. Materials

Boltorn^® H40 (a hyperbranched polyester with 64 hydroxyl terminal groups, molecular weight ~ 7.3 kDa) was kindly provided by Perstorp Polyols Inc., USA, and purified by fractional precipitation in acetone and tetrahydrofuran (THF). Tin(II) 2-ethylhexanoate (Sn(Oct)₂), triethylamine, tert-Butyl methacrylate, 2-(dimethylamino)ethyl methacrylate, copper(I) bromide (CuBr), and trifluoroacetic acid (TFA) were purchased from Sigma-Aldrich (St. Louis, MO, USA), respectively. D,l-lactide, pentamethyldiethylenetriamine, and 2-(4-(bromomethyl)phenyl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane were purchased from TCI (Tokyo, Japan), respectively. 2-bromoisobutyryl bromide and neutral aluminium oxide (Al₂O₃) for chromatography were purchased from Acros Organics (Pittsburgh, PA, USA). RVX-208 was purchased from ApexBio (Houston, TX). Poly(lactic-co-glycolic acid) (PLGA, 50: 50 M ratio of lactide: glycolide, the inherent viscosity of 0.15–0.25 dL/g, average molecular weight of 6.5 kDa) was obtained from LACTEL Absorbable Polymers (Birmingham, AL USA). Other reagents were purchased from ThermoFisher Scientific (Fitchburg, WI, USA) and used as received unless otherwise stated.

2.2. Cell culture

Human aortic VSMC and culture media (SmBm-2 basal medium for experimental purposes, and SmGm-2 complete medium for expansion) were purchased from Lonza (Walkersville, MD). Cells between passages 5 and 7 were maintained at 37 °C with 5% CO₂ and used for all experiments. As established in our previous studies, Accutase (Thermo Fisher Scientific, Waltham, MA) instead of Trypsin was used for cell detachment to ensure fast cell adhesion and optimal cell status [36].

2.3. Preparation of the RVX-loaded platelet membrane-coated nanoclusters

H40-PLA-P(MAA-co-(DMAEMA-PAPE)) (3 mg) and RVX (1 mg) were dissolved in 200 μL of methanol and then added dropwise into the deionized water under sonication. Methanol was evaporated at 37 °C under reduced pressure. The RVX-loaded nanoclusters were obtained after lyophilization using a Labconco lyophilizer (Labconco, Kansas, USA) at −80 °C and < 0.1 mbar for 48 h. All chemical solutions used in the preparation of the nanoclusters were sterilized by filtration through 0.2 μm PVDF filters. The platelet membrane vesicles were derived as previously reported [24]. Biomimetic nanoclusters were prepared by coating platelet membrane on the surface of the nanoclusters via the extrusion method [37,38]. Briefly, platelet membrane vesicles were mixed with the pre-sonicated RVX-loaded nanoclusters at a membrane protein to polymer weight ratio of 1:1 and then extruded through a 400 nm and a 200 nm polycarbonate porous membrane sequentially using an Avanti mini extruder under sterile condition. The empty biomimetic nanoclusters were prepared following a similar procedure by extruding empty nanoclusters with the platelet membrane vesicles. Cy5.5-loaded biomimetic nanoclusters were prepared in the same way. RVX-loaded PLGA NP were prepared by fusing platelet membrane vesicles with PLGA nanocores that fabricated by a previously described nano-precipitation method [37,38]. RVX was mixed with PLGA in acetone and added dropwise into 0.05% poloxamer 188 containing aqueous solution. The solvent was evaporated by rotary evaporation. Prior to in vitro and in vivo administrations, the biomimetic nanoparticle solutions were filtered through 0.45 μm PVDF syringe filters.

2.4. Characterization of the platelet membrane-coated, ROS-detonable nanoclusters

The hydrodynamic diameter and zeta potential of the biomimetic nanoclusters were obtained by dynamic light scattering (DLS) measurements by a ZetaSizer Nano ZS90 spectrometer (Malvern Instruments, USA) at a concentration of 0.1 mg/mL. The morphology of the biomimetic nanoclusters was characterized by transmission electron microscope (TEM, FEI Tecnai G2 F30 TWIN 300 KV, E.A. Fischione Instruments, Inc. USA). Phosphotungstic acid solution (1%, w/w) was applied for the negative staining of the samples. Detonation of the biomimetic nanoclusters triggered by hydrogen peroxide (H₂O₂) was studied by incubation with 100 μM H₂O₂ for 1 h. The change of particle size and morphology of the biomimetic nanoclusters were monitored by DLS and TEM, respectively. Moreover, the digital photos were taken for direct illustration.

An ultrafiltration method was used to separate the unencapsulated drug from the biomimetic nanoclusters for the evaluation of the drug loading efficiency and loading capacity [38]. Briefly, the biomimetic nanoclusters were added into the centrifugal filters (Millipore) with a molecular weight cutoff at 10 kDa, followed by centrifugation at 2800g for 20 min. The amount of drug in the filtrate was measured by reversed-phase high-performance liquid chromatography (HPLC) according to the manufacture’s instructions (ApexBio, Houston, TX, USA). Briefly, the HPLC system (Hitachi LaChrom Elite^®) equipped with an auto sampler (Hitachi LaChrom Elite L-2200), a HPLC pump (Hitachi LaChrom Elite L-2130), a column oven (Hitachi LaChrom Elite L-2350), and a UV-VIS detector (Hitachi LaChrom Elite L-2420) was used. Chromatographic separations were achieved using a LaChrom C18 reversed-phase analytical column (5 μm, 150 mm × 4.6 mm) at 30 °C. The mobile phase was comprised of solvent A (0.1% H₃PO₄ containing acetonitrile, HPLC-grade) and solvent B (0.1% H₃PO₄ containing deionized water) in gradient elution mode. The elution program was as follows: 0–6 min, 10%–95% A and 90%–5% B; 6–10 min, 95% A and 5% B. The samples were delivered at a flow rate of 0.5 mL/min and detected at 254 nm using UV detection. Loading efficiency and loading capacity in percentages were calculated according to the following equations: loading efficiency (%) = (1 − amount of unencapsulated drug/total amount of drug added) × 100%. Loading capacity (%) = total weight of encapsulated drug/total weight of nanoparticles × 100%.

In vitro release profiles of RVX from the biomimetic nanoclusters were studied via dialysis method. Briefly, 1.5 mL of the biomimetic nanoclusters (1 mg/mL of RVX) was sealed in a cellulose membrane dialysis bag with a molecular weight cutoff at 3.5 kDa. Then, the bag was placed in 50 mL of 1% Tween 80 containing PBS (0.1 M, pH 7.4) with or without 100 μM H₂O₂, which was then kept in a shaker (100 rpm) at 37 °C. At certain time points, 0.2 mL of the release medium was collected to measure the amount of released RVX by HPLC at a wavelength of 254 nm. To maintain a constant medium volume, equivalent volumes of pre-warmed fresh medium were added back to the release medium after each sampling. The cumulative release amounts of RVX were calculated based on the standard curve.

2.5. Cell viability assay of the nanoclusters

The cytotoxicity and anti-proliferation of the nanoclusters on VSMC was evaluated by a CellTiter-Glo Luminescent Cell Viability kit according to manufacturer’s instructions. VSMC were cultured in 96-well plates (2 × 10⁴ and 8 × 10⁴ cells/well for experiments of proliferation and cytotoxicity, respectively) and grown for overnight. For cytotoxicity study, fresh basal medium (supplemented with 0.5% fetal bovine serum) containing NC with different concentrations were added into the plates. For proliferation study, VSMC were treated with 20 nM PDGF-BB to stimulate ROS and proliferation. Different formulations of RVX (free solution, RVX-loaded PLGA NP, or RVX-loaded NC) and vehicle controls at 1 μM or 10 μM were applied to the VSMC. VSMC treated with fresh medium alone served as a control group. At each different time point (24 h, 48 h, 72 h, 96 h), 50 μL CellTiter-Glo reagent/50 μL PBS were added into each well upon removal of the culture medium. Following 10 min incubation at room temperature, the 96-well plates were analyzed in a FlexStation 3 Benchtop Multi-Mode Microplate Reader (Molecular Devices, San Jose, California) (250-ms integration).

2.6. Rat carotid artery balloon angioplasty model

All animal experiments conform to the protocols approved by the Animal Care and Use Committee (ACUC) of the University of Virginia. All adult male Sprague-Dawley rats (Charles River, weighing 300–350 g) were maintained in a temperature and humidity-controlled SPF level animal facility under a 12 h light-dark cycle. Food and water were available ad libitum. Animals were randomly assigned to each treatment group. The carotid artery balloon injury model was operated as previously described [21]. In brief, after anesthesia with isoflurane (5% for inducing and 2% for maintaining), a mid-line incision in the neck was made, and the carotid arteries and carotid bifurcation were exposed by blunt dissection. Proximal regions of left carotid artery, inner carotid artery, and external carotid artery were then temporarily ligated with 5–0 nylon suture to avoid excessive blood loss during the surgery. A 1 mm arteriotomy was created on the distal segment of the external carotid arteries. Subsequently, a 2-French arterial balloon catheter (Edwards Scientific, Irvine, CA) was introduced through the arteriotomy and advanced ~1.5 cm into the common carotid arteries. The balloon catheter was slowly inflated at 1.5 atm pressure followed by withdrawn to the carotid bifurcation for three times continuously with rotation to ensure complete denudation of the endothelium. The ligatures on the common carotid arteries were briefly lifted to check the status of back-bleeding (lack of back-bleeding indicates potential occlusion, and the animals will be excluded from the study). The external carotid arteries were then permanently ligated, and wounds were closed bioabsorbable sutures. Heating pad was provided as heat source during operation and the recovery phase. Analgesics and post-operative care were provided per ACUC policy.

2.7. Homing of the biomimetic nanoclusters to injured carotid arteries

Following angioplasty-induced arterial denudation, the Cy5.5-loaded NC were intravenously injected via the tail vein (n = 3 rats per group). 24 h later, the animals were euthanized by CO₂ inhalation, and the main organs or tissues including heart, liver, spleen, lungs, kidneys, intestine, and the balloon-injured and non-injured carotid arteries were carefully collected for ex vivo fluorescence imaging by an in vivo imaging system (IVIS) with fluorescence excitation/emission wavelengths of 676/705 nm. The Cy5.5-loaded PLGA NP with similar particle size was used for comparation.

2.8. Tissue penetration of the biomimetic nanoclusters into medial smooth muscle layer

To monitor the penetration of the biomimetic nanoclusters into the medial smooth muscle layer, the carotid arteries (n = 3 rats per group) that harvested for ex vivo imaging were frozen in optimal cutting temperature (OCT) compound (VWR International) and then cut into 8 μm sections using a Cyrotome cryostat machine (Leica). The tissue sections were placed on polylysine-treated glass slides. Subsequently, the sections on slides were first washed with PBS to remove residual OCT compound and then incubated with DAPI for 15 min in a light-protected humidified chamber to stain nuclei. Then, the stained carotid artery cross sections were observed using a confocal laser scanning microscopy (CLSM, Nikon, Japan).

2.9. Morphometric analysis of IH and restenosis

Sprague-Dawley rats after carotid artery balloon injury were randomly assigned into 5 groups, including saline, free RVX, empty NC, RVX-loaded PLGA NP, and RVX-loaded NC group (n = 4–6 rats). The dose of RVX in all drug-containing groups was constant at 10 mg/kg. RVX-loaded biomimetic nanoparticles or compound solutions were immediately intravenously administrated after balloon injury. 2 weeks later, the animals were euthanized and fixed via transcardial perfusion of saline at a pressure of 100 mmHg. The balloon-injured carotid arteries and relevant organs (e.g., heart, kidney, liver, spleen) were then carefully excised from the surrounding tissue. The collected specimens were sliced, with the majority portion of each for formalin-fixed, paraffin-embedded (FFPE) processing, and the rest for RNA extraction. Paraffin sections (5 μm per section) were obtained from the harvested arteries at equally spaced intervals. Haemotoxylin and Eosin (H&E) staining was performed for histopathological analysis. Area inside external elastic lamina (EEL area), area inside internal elastic lamina (IEL area), lumen area, intima area (=IEL area - lumen area), and media area (= EEL area - IEL area) were calculated using Image J. Intimal hyperplasis was defined as the ratio of intima area versus media area (Intima-to-media ratio, I/M ratio).

2.10. Biocompatibility

2.10.1. Histological examination

After euthanasia, the main organs or tissues including heart, liver, spleen, and kidney were collected, fixed with 4% paraformaldehyde, and embedded in paraffin in a tissue base mold (n = 4–6 rats). The blocks were sectioned at a microtome setting of 5 μm for H&E staining The images were captured by an optical microscope for morphometric analysis.

2.10.2. Real-time quantitative PCR (qRT-PCR) analysis of mRNA expression levels in cell culture and tissue homogenate

mRNA was isolated from cultured VSMC and collected carotid segments (n = 4–6 rats) using TRIzol following the manufacturer’s instructions. Purified mRNA (1 μg) was used for the first-strand cDNA synthesis and quantitative RT-PCR was performed using the QuantStudio3 (Applied Biosystems, Carlsbad, CA). Each cDNA template was amplified in triplicates using SYBR Green PCR Master Mix, with the following primer sets: Human APOA-I forward primer ACTGTGTACGTGGATGTGCTCAAAG, reverse primer CACGCTGTCCCAGTTGTCAAG; Human GAPDH forward primer ATTCCACCCATGGCAAATTCC, reverse primer GACTCCACGACGTACTCAGC; Rat MCP1 forward primer CTTCCAAGTGGCTAAGGGCA, reverse primer TCAAAGGGAGTCGGGGATCT; Rat TNFα forward primer GATCGGTCCCAACAAGGAGG, reverse primer TCCCTCAGGGGTGTCCTTAG; Rat BAX forward primer CACTAAAGTGCCCGAGCTGA, reverse primer TCCAGATGGTGAGTGAGGCA; Rat Caspase3 forward primer ACTGGAATGTCAGCTCGCAA, reverse primer TCAAATTCCGTGGCCACCTT; Rat GAPDH forward primer GACATGCCGCCTGGAGAAAC, reverse primer AGCCCAGGATGCCCTTTAGT.

2.11. Statistical analysis

All data are presented as mean ± standard error of the mean (SEM). Statistical analysis was performed using GraphPad Prism 9.1.0. For multiple group-wise comparison, one-way analysis of variance (ANOVA) was performed followed by Bonferroni post-hoc tests. For two-group comparison, Student’s t-test was performed. The value P < 0.05 was considered to be statistically significant.

3. Results

3.1. Design, synthesis and characterization of the platelet membrane-coated, ROS-detonable nanoclusters

Herein, we conceived a nanocluster that could enable multi-functionality ideal for anti-restenotic therapy. The nanocluster core was formed by self-assembly of unimolecular nanoparticles (H40-PLA-P (MAA-co-(DMAEMA-PAPE), Fig. 1). H40-PLA-OH (Polymer 1 in Fig. 1) was first synthesized by ring-opening polymerization of D,l-lactide monomer using hyperbranched polyester Boltorn^® H40-OH as the macromolecular initiator. Then, H40-PLA-OH was reacted with 2-bromoisobutyryl bromide via esterification to yield H40-PLA-Br (Polymer 2 in Fig. 1) [39]. Thereafter, H40-PLA-P(tBMA-co-DMAEMA) (Polymer 3 in Fig. 1) was synthesized by ATRP using H40-PLA-Br as the bromide initiator. The hydrophobic phenylboronic ester was conjugated to the tertiary amine group to obtain H40-PLA-P(tBMA-co-(DMAEMA-PAPE)). The final product used to prepare unimolecular nanoparticles (i.e., H40-PLA-P(MAA-co-(DMAEMA-PAPE)), Polymer 4 in Fig. 1) were obtained after deprotection of tert-butyl alcohol. The structure of all intermediate and final products was verified by ¹H NMR (Fig. S1-4). The nanoclusters were self-assembled from multiple small unimolecular nanoparticles due to the increased hydrophobicity after phenylboronic ester modification. Platelet membrane was coated onto the surface of the nanoclusters via an extrusion process to render the nanoclusters with biomimicry, stability, and solubility (hence injectability). DLS measurements showed that the hydrodynamic diameter of the biomimetic nanoclusters was 192.5 ± 2.1 nm, and the surface zeta potential was −29.7 ± 1.0 mV. Notably, without platelet membrane coating, the unimolecular nanoparticles were not stable and can readily form aggregates in an aqueous solution via hydrophobic interaction. The loading efficiency and loading capacity of the RVX in the biomimetic nanoclusters quantified by HPLC were 89% and 10.6%, respectively.

Fig. 1.

A schematic illustration for the synthesis of the unimolecular nanoparticles with ROS-detonable chemistry and the assembly of biomimetic nanoclusters. (A) A schematic description of the preparation and detonation of the nanocluster. First, multiple small unimolecular nanoparticles self-assemble to form the nanocluster structure in aqueous solution via the hydrophobic interaction between phenylboronic esters. Subsequently, platelet membrane is coated on the surface of the nanoclusters via an extrusion process to confer lesion-targetability and particle stability in aqueous solution. After homing to the restenotic vessels that are highly enriched with ROS, the ROS-detonable nanoclusters will de-cluster to smaller hydrophilic unimolecular nanoparticles and penetrate the multilayered vessel wall structure, thus allowing the effective delivery of anti-restenotic payloads to the “epicenter” of endovascular lesion. (B) Synthesis scheme of the unimolecular nanoparticles. H40-OH is used as a macromolecular initiator to synthesize H40-PLA-OH (Polymer 1) by ring-opening polymerization of D,l-lactide monomer. A bromide initiator (Polymer 2) for atom transfer radical polymerization (ATRP) can be synthesized through esterification. Thereafter, H40-PLA-P(tBMA-co-DMAEMA) (Polymer 3) can be synthesized via ATRP. The final product used to form the H₂O₂-responsive unimolecular nanoparticles (H40-PLA-P(MAA-co-(DMAEMA-PAPE)), Polymer 4) can be prepared through conjugation of hydrophobic phenylboronic ester followed by deprotection of tert-butyl alcohol. Upon sensing the elevated H₂O₂ at the restenosis site, the hydrophobic phenylboronic ester will be rapidly oxidized and cleaved to form hydrophilic unimolecular nanoparticles.

Excessive accumulation of ROS has been well documented in diseased vasculatures such as in the cases of atherosclerosis and angioplastied or stented vessels. Taking advantage of this pathophysiological feature, we designed a ROS-detonable “trigger” for our nanoclusters to enable “on-demand” size transition specifically at the lesion site (Fig. 2A). To showcase this feature, we transiently exposed the biomimetic nanoclusters to H₂O₂ at a physiologically relevant concentration as a surrogate of ROS. The changes of particle size and morphology were monitored by DLS and TEM, respectively, in the presence of 100 μM H₂O₂ for 1 h. As presented in Fig. 2B, the size distribution dramatically changed in response to 100 μM H₂O₂ as a characteristic peak appeared in around 40 nm, indicating that the biomimetic nanoclusters degraded into smaller nanoparticles. Moreover, turbidity of the nanoparticles solution remarkably changed from light blue opalescence to clear after the incubation with H₂O₂. TEM images showed the formation of the distinctive biomimetic nanoclusters and consistent membrane coatings over the nanocores. In addition, TEM visualization also confirmed the detonation of the biomimetic nanoclusters after H₂O₂ (100 μM) treatment (Fig. 2C). Furthermore, the nanoclusters disassembled into smaller nanoparticles after 40 min and 15 min exposure to 200 μM or 1 mM H₂O₂, respectively (Fig. S5, S6). The biomimetic nanoclusters were penetrated by H₂O₂, which quickly oxidizes and cleaves off the hydrophobic phenylboronic ester, converts the hydrophobic nanoparticle to hydrophilic, and generates two types of water-soluble small molecules, i.e., 4-(hydroxymethyl)phenol and boronic acid [40–42]. The drastically reduced hydrophobic interactions between small unimolecular nanoparticles resulting from the ROS-triggered hydrophobic to hydrophilic transition and the two water soluble byproducts led to an increased osmotic pressure within the platelet membrane-coated nanocluster and subsequently, the disassembly of the nanoclusters, thereby releasing the small-sized nanoparticles ideal for deep tissue penetration. Moreover, the influence of H₂O₂ on in vitro drug release profile of the nanoparticles was evaluated. The drug release behavior of the PLGA NP was not affected by H₂O₂, while an increased drug release was observed for the NC in the presence of H₂O₂, which further demonstrated the ROS responsiveness of the nanoclusters (Fig. 2D).

Fig. 2.

Characterization of the platelet membrane-coated, ROS-detonable nanoclusters. (A) Schematic illustration for the H₂O₂-triggered detonation of the nanocluster. (B) Size distribution measured by dynamic light scattering (DLS) and digital photos of the nanoclusters (Left) and de-clustered, unimolecular nanoparticles (Right) in response to stimulation with 100 μM H₂O₂ for 1 h. (C) Transmission electron microscope (TEM) images of the nanoclusters in response to 100 μM H₂O₂. (D) Drug release kinetics from the biomimetic nanoparticles in PBS (0.1 M, pH 7.4) with or without 100 μM H₂O₂. Data are presented as mean ± SEM (n = 3).

3.2. The multimodal nanocluster enables efficient delivery of anti-restenotic payload in vitro

To evaluate the cytotoxicity of our innovative nanocluster via CellTiter-Glo Luminescent cell viability assay, VSMC were incubated with various concentrations (varied from 0 to 1 mg/mL) of NC for 24, 48, or 72 h. In a separate set of experiments, NC was pre-exposed with H₂O₂, the stimulus that can cause de-clustering of the NC into small nanoparticles as described in Fig. 2, followed by administration to VSMC. As presented in Fig. 3A and Fig. S7, no obvious toxicity was observed with neither intact NC nor the de-clustered nanoparticles at concentrations less than 300 μg/mL.

Fig. 3.

In vitro cytotoxicity and anti-proliferation activities of the platelet membrane-coated, ROS-detonable nanoclusters in VSMC. (A) Determining the cytotoxicity profile of the platelet membrane-coated, ROS-detonable nanocluster. At different time points after exposing confluent VSMC to various concentrations (from to 10 ng/mL to 300 μg/mL) of the nanoclusters, cell viability was measured using CellTiter-Glo Luminescent Cell Viability assay. Data are presented as mean ± SEM (n = 3). (B and C) Evaluating the anti-proliferative and APOA-I-inducing performances of an anti-restenotic drug (RVX) delivered through the innovative nanoclusters. 20 ng/mL PDGF-BB was used to induce the proliferation of VSMC. In the RVX containing formulations-treated groups (i.e., free RVX, RVX-loaded NC, and RVX-loaded PLGA NP), PDGF-BB-stimulated cells were treated with various formulations at an equivalent amount of RVX (1 μM or 10 μM). At 24, 48, 72, and 96 h post-treatment, the cell proliferation activity was measured by a CellTiter-Glo Luminescent kit. Additionally, additional sets of VSMC were subject to mRNA isolation at 48 h post-treatment followed by qPCR evaluation of APOA-I mRNA levels. Data are presented as mean ± SEM (n = 4). *P < 0.05, **P < 0.01, ***P < 0.001. In B, statistical significances were marked between RVX-loaded NC and RVX-loaded PLGA NP. One-Way ANOVA followed by post-hoc Bonferroni analysis.

It is well known that the excessive proliferation of VSMC is the deciding factor behind the pathogenesis of IH and hence restenosis [43]. After endovascular injuries (e.g., angioplasty or stenting), VSMC are exposed to a milieu highly enriched in cytokines, growth factors, and ROS; and consequently, in the present study, platelet-derived growth factor BB (PDGF-BB), one of the most prominent growth factors in restenotic environment, was utilized to induce VSMC proliferation and ROS induction, as established in prior studies [44]. VSMC were then subjected to treatment with various formulations of RVX, an emerging anti-restenotic drug with a suboptimal pharmacological profile. These include RVX in free solution, or RVX-loaded NC, or RVX-loaded PLGA NP. Vehicle control groups were also included for comparison (e.g., saline, DMSO, and empty NC and PLGA NP). As shown in Fig. 3B, in the saline and DMSO treatment group, PDGF-BB robustly stimulated the proliferation of VSMC, whereas treatment with a low dose RVX (1 μM, either in its free solution or RVX-loaded PLGA NP formulation) failed to exert any notable effect. In contrast, incubation with RVX-loaded NC with the same RVX concentration effectively inhibited VSMC proliferation, with 35%, 40%, 52% reductions at 48 h, 72 h, and 96 h, respectively. Additionally, APOA-I, a vascular protective gene that can be specifically mobilized by RVX as shown in recent clinical trials [35,45,46], was upregulated exclusively in VSMC treated with our innovative nanocluster formulation (RVX-loaded NC, Fig. 3C). We reason this might be due to the improved intracellular drug release of RVX-loaded NC in response to PDGF-BB-induced intracellular ROS enrichment in VSMC [47]. Of note, at a dosage of 10 μM as commonly used in vitro, all formulations of RVX unanimously led to effective inhibition of VSMC proliferation (Fig. 3B).

3.3. The platelet membrane-coated, ROS-detonable nanocluster demonstrates superior targetability, biodistribution, and tissue penetration in vivo

To determine the targeting specificity and whole-body biodistribution, a balloon angioplasty-induced carotid artery injury model was established in rats. Cy5.5-loaded NC or control PLGA NP of similar particle size were injected via tail vein immediately after balloon injury. 24 h later, the main organs and injured and non-injured carotid arteries were excised for ex vivo imaging (Fig. 4). As demonstrated in Fig. 4A, the platelet membrane coating enabled highly enriched fluorescence signal in the injured carotid arteries over non-injured contralateral control for both nanoparticles, which is consistent with our previous study [21]. This active targeting ability is likely due to the specific interaction between glycoprotein VI on the platelet membrane with collagen exposed at the injury site. Moreover, the targeting specificity (shown as fluorescence signal ratio between injured and non-injured arteries in each rat) in the Cy5.5-loaded NC group was significantly higher than the Cy5.5-loaded PLGA NP group with an equivalent amount of Cy5.5 (Fig. 4). Further analysis in tissue homogenates revealed that 0.41 ± 0.29% of the intravenously administered Cy5.5-loaded NC could be found in the injured carotid arteries, while signals at mere background level were detectable in the non-injured ones (Fig. S8). A similar pattern could be observed in the control PLGA NP group (0.18 ± 0.22% in the injured arteries), albeit no statistical significance was detected between NC and PLGA NP groups. Additionally, PLGA NP displayed a suboptimal biodistribution profile with potential off-target issues, as evidenced by extensive accumulation in lung and liver (Fig. 4B). In contrast, our innovative NC demonstrated significantly enhanced targeting specificity, with a 4-fold superior performance (injured-to-un-injured signal ratio) over PLGA NP (Fig. 4C). Moreover, significantly less (~50%) lung accumulation was noted in animals administered with NC versus PLGA NP group.

Fig. 4.

The platelet membrane-coated, ROS-detonable nanocluster enabled improved lesion-targeting capacity and biodistribution pattern. (A) Ex vivo fluorescence images of the balloon-injured and non-injured carotid arteries. Cy5.5-loaded PLGA NP and Cy5.5-loaded NC were intravenously injected immediately after balloon angioplasty of the rat carotid artery. The balloon-injured and non-injured carotid arteries were collected 24 h later for ex vivo imaging using a IVIS system (Ex/Em: 676/705 nm). I and N represent the injured and non-injured carotid artery. (B) Ex vivo fluorescence images of the Cy5.5-loaded PLGA NP and Cy5.5-loaded NC in the major organs and injured and non-injured carotid arteries. H, L, S, Lu, K, In, I and N represent heart, liver, spleen, lung, kidney, intestine, injured artery, and non-injured artery, respectively. (C) Quantitative analysis of the mean fluorescence intensity per unit mass in each organ or tissue shown in the ex vivo images. Data are presented as mean ± SEM (n = 3 rats). *P < 0.05 and **P < 0.01. One-Way ANOVA followed by post-hoc Bonferroni analysis.

To compare the tissue penetration capabilities of the two nanoparticles, the injured carotid arteries excised from rats treated with Cy5.5-loaded NC or PLGA NP via intravenous injections were studied using a confocal microscopy. As shown in Fig. 5, green fluorescence indicated the autofluorescence of the internal elastic lamina, and cell nuclei were counterstained with DAPI as shown in blue signal. Cy5.5-loaded nanoaprticles displayed as red fluorescence signal. Consistent with the findings from the IVIS study, arterial cross-sectional images from the PLGA NP-treated rats showed minimal Cy5.5 fluorescence signal, exclusively located in the near-luminal intimal layer. The presence of NC, on the other hand, was readily observable in the distal layers of the vessel wall, shown as fluorescent puncta in medial layer. Collectively, these data support a superior tissue penetrating capacity of the ROS-detonable nanocluster over the conventional nanoplatform.

Fig. 5.

Fluorescence images of the injured carotid artery cross sections. Green channel: the autofluorescence of the internal elastic lamina. Blue channel: 4′,6-diamidino-2-phenylindole (DAPI) stained nuclei. Red channel: Cy5.5 signal. Scale bar: 50 μm. Cy5.5-loaded PLGA NP and Cy5.5-loaded NC were administered systemically following vascular injury. 24 h later, the injured carotid arteries were collected and sliced. Slides were stained with DAPI for nuclei and visualized by confocal laser scanning microscopy. Quantitative analysis of the fluorescence intensity of Cy5.5 using Image J. Data are presented as mean ± SEM (n = 3 rats). ***P < 0.001 with Student’s t-test. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.4. RVX-loaded ROS-responsive nanoclusters, but not conventional RVX-loaded PLGA NP, led to effective amelioration of IH

Prompted by the improved performances of the multimodal nanocluster, we then determined the anti-restenotic potency of RVX delivered through NC in comparison to the RVX-loaded PLGA NP. Immediately following angioplasty, rats received one-time intravenous injection of the nano-formulated RVX (10 mg/kg) or control treatment (saline, DMSO vehicle, empty NC). Two weeks following angioplasty, rats were euthanized, and the carotid arteries were collected for morphometric analysis. I/M ratio and lumen area were quantified to evaluate the vascular remodeling efficacy. As shown in Fig. 6, the empty NC did not exert any deleterious changes in the injured arteries, such as luminal restriction or immune cell infiltrations. Amongst the 3 formulations of RVX, only RVX-loaded NC elicited effective mitigation of IH. In contrast, neither free RVX solution nor that delivered through conventional PLGA NP led to any significant changes in intima-to-media (I/M) ratio. Moreover, a significant expansion of lumen area could be observed with treatment of RVX-loaded NC, but not other formulations of RVX. These evidence indicates that efficient delivery through our innovative nanoplatform with preferred targetability and tissue penetration can significantly reduce the effective dose required for RVX, an emerging cardiovascular drug with an undesirable pharmacological profile. Additionally, no signs of systemic toxicity could be observed, as demonstrated by the normal histological features of heart, lung, kidney, spleen, and liver (Fig. S9), as well as the minimal changes in gene transcription levels of inflammatory (TNFα and MCP1) and apoptotic markers (BAX and Caspase3) from spleen and liver tissue homogenates (Fig. S10).

Fig. 6.

Anti-restenotic effects of the biomimetic nanoclusters in a balloon angioplasty-induced carotid artery injury model. (A) H&E stained histological sections of balloon-injured carotid arteries after various treatments. Scale bars: 100 μm for 10× low magnification fields, and 50 μm for 40× high magnification fields. Saline, free RVX, empty NC, RVX-loaded PLGA NP, and RVX-loaded NC were injected via tail vein immediately after balloon angioplasty. Two-weeks later, the injured carotid arteries were excised, sliced, and stained with H&E for histopathological analysis. (B) Quantitative analysis of the I/M ratio and lumen area. Data are presented as mean ± SEM (n = 4–6 rats). *P < 0.05 and **P < 0.01. One-Way ANOVA followed by post-hoc Bonferroni analysis.

4. Discussion

Current anti-restenotic methods, including DES (standard-of-care), are plagued with significant drawbacks. They still leave a significant portion of patients unprotected from restenosis, particularly those with obesity and diabetes or with peripheral artery diseases (e.g. lower extremities, up to 75% incidence) [6,7,48]. In fact, stenting and angioplasty may not be applicable or suboptimal to small-diameter extremity vessels. Moreover, serious safety concerns, including thrombosis and increased mortality have been recently reported as resulting from both DES and DCB [9]. Additionally, once a DES or DCB is deployed, the anti-restenotic payload will be released within several weeks or days, therefore only affording a highly fixed regimen with no flexibility. Alternative stent-free strategies that could offer more effective yet safer targeted delivery of anti-restenotic agents will bring tremendous benefits to these patients.

The advent of nanotechnology ushers in a new era for targeted drug delivery. Currently, more than 60 nanomedicine have been approved by FDA or received the Conformité Européenne (CE) marking, mostly for cancer treatment [49]. Unfortunately, there remains a paucity of clinically approved or actively tested candidates for cardiovascular nanomedicine. Numerous types of nanoparticle-based therapies have been developed for anti-restenotic treatment due to their outstanding drug-loading and controlled-release properties, yet very few have entered human clinical trials. So far, the most successful ones primarily utilize a nanoparticle albumin bound drug (Nab) strategy, such as Nab-Paclitaxel (ABI-007). ABI-007 once entered phase I/II clinical trial (SNAPIST-I) for peripheral arterial diseases, but was later terminated [49,50]. LABR-312, a liposomal nanoproduct loaded with alendronate, recently showed inspiring result from Phase IIb trial, indicating early efficacy in diabetic patients who received angioplasty and DES [51]. Nevertheless, it is important to note that LABR-312 is mainly tested as an ancillary treatment, rather than an alternative to DES. In sum, despite the well documented safety and the potential advantages offered by nanomedicine, there has yet been any formal clinical evaluation and utility for anti-restenotic therapies.

Numerous nanoparticle designs have been reported with promising efficacy in pre-clinical models in recent years. To achieve targeted delivery to vascular lesion sites, various strategies have been adopted, ranging from surface ligand conjugation (e.g., collagen- or integrin-targeting peptides), to utilizing the shape-effect (e.g., cylindrical versus spherical nanostructures). Biomimetic surface functionalization represents a novel and highly promising strategy in targeted nanomedicine [52,53]. Our group and others recently developed a series of biomimetic (biomembrane-coated) nanoplatforms for cardiovascular applications [21,52,54]. Platelets are the natural “first responders” to intravascular injuries, enabled through their spontaneous homing and prompt accumulation to vascular lesion sites. Taking advantage of the plural targeting ligands endogenously expressed on the membrane surface of platelets, nanoparticles coated with platelet-derived membranes could display various biomimetic features, including the endovascular lesion-targeting capacity and certain level of biocompatibility. Additionally, due to the lack of intracellular contents essential to trigger the coagulation cascades, such platelet membrane coating strategies are free from concerns of increased thrombogenicity.

However, certain limitations exist in the earlier designs of nanoparticle-based therapies and biomimetic nanomedicine — most notably, the suboptimal tissue penetration and biodistribution. In the case of DES, these two issues were naturally solved by the physical presence of the drug-bearing devices deep inside the multi-layered vessel wall [25,27,28]. Therefore, there remain major obstacles for nanoparticle-based therapies to overcome in order to outcompete the current standard-of-care [26]. In the seminal work by Che-Ming J Hu et al. in which the biomimetic biointerfacing concept was first tested, the presence of platelet membrane-coated conventional PLGA nanoparticles and its fluorescent payload (i.e., docetaxel) was primarily seen in the luminal layer of the vessel wall, with minimal penetration into the medial and adventitial layers [24]. Moreover, despite the outstanding lesion selectivity between injured versus non-injured arteries and ultimately a promising therapeutic efficacy, a closer look at the biodistribution unveiled a liver/spleen-predominant pattern of retention over the arterial tissue, thereby posing a potential risk of significant “off-target” risk. This was similarly observed in our recent study with a prototypic design of platelet membrane-coated nanostructure for anti-restenotic application [21]. This suffices to say that major innovation is needed to improve the tissue penetration capability and biodistribution of nanoparticle-based anti-restenotic therapies.

Size matters — there is no such thing as one-size-fits-all in nanoparticle applications in complex biology. It is increasingly recognized that nanoparticle size plays a vital role in its biodistribution and tissue/cell penetration [29,30]. For longer circulation (less organ entrapment), a larger size is favorable [30], whereas a smaller size is desirable for better lesion penetration — it is impossible to have both in the same nanoparticle. To solve this dilemma, we designed a size-tunable nanocluster that could help attain optimal biodistribution and tissue penetration simultaneously. That is, a nanocluster (~190 nm) formed with small unimolecular nanoparticles circulates in the blood flow until it reaches the target restenotic lesion (guided through its platelet membrane coating). Once attached there, it is exposed to the excessive ROS, which triggers our innovative ROS-responsive detonation chemistry. The small molecule byproducts from this cleavage reaction (i.e., 4-(hydroxymethyl)phenol and boronic acid) together with the resulting hydrophilic nanoparticles then generate an osmotic pressure, which swells and bursts the membrane coating, and the small hydrophilic nanoparticles are released to penetrate through the multi-layered arterial wall into the restenotic lesion and provide sustained drug delivery.

On the translational front, our current study established the early pre-clinical evidence supporting the translational potential of our innovative nanoplatform capable of significantly improving the performance of RVX-208 (RVX, also known as Apabetalone), an orally available inhibitor of the BET protein family. Our group was the first to establish the role of BET as intervention target for IH and restenosis in both endovascular and open bypass interventions [21,44,55]. Indeed, RVX in its oral formulation has demonstrated safety and early efficacy in clinical trials for management of diseases ranging from coronary and pulmonary artery diseases to metabolic syndrome [56,57]. Despite its early promising results, the much anticipated phase III BETonMACE trial of RVX failed to meet its primary endpoint in reducing the major adverse cardiovascular events (MACE) after acute coronary syndrome [58,59]. Nevertheless, BET inhibiting therapy still represents a highly innovative and unique therapeutic avenue for a broad spectrum of diseases. Despite its suboptimal clinical efficacy, RVX was recently granted the breakthrough therapy designation by FDA and poised for broader therapeutic applications [60], thereby demonstrating the clinical necessity for further efforts to optimize its delivery formulation.

Our multi-modal nanoplatform offers substantial advantages over conventional drug formulations in improving the clinical utilities of RVX, which represent an emerging class of promising cardiovascular drugs with less desirable pharmacological properties. One major barrier facing first-generation BET inhibitor drugs (e.g., JQ1, RVX, ZEN-3694) is their short half-life (<2 h) upon systemic administration [61–63]. Consequently, the majority of previous studies resorted to a twice-daily (b.i.d) regimen (>50 mg/kg/d in pre-clinical studies; >2 mg/kg/d in clinical trials) [34,64]. A sustained drug release formulation will not only help improve the overall drug performance, but also mitigate the concerns of medication non-adherence. Herein, our rodent study demonstrated the potent anti-restenotic efficacy of a single-dose, nanocluster-formulated RVX over a 2-week course, highlighting a substantial improvement over the conventional b.i.d regimen in its conventional oral formulation. Another facet of the problem with RVX lies in its poor biodistribution pattern, severely limiting the local drug concentration at the lesioned vasculature. Indeed, the same issue has persistently plagued the clinical utilities of many existing anti-restenotic agents orally available (e.g., paclitaxel), which prompted the introduction of drug-releasing devices such as DES and DCB [65]. RVX is primarily localized in liver and small intestine instead of vasculature — a pattern ideal for lipid-lowering purposes but undesirable for anti-restenotic therapy [61]. Especially considering the recent safety concerns raised against these mainstream drug-releasing devices, our nanoparticle-based approach presents a promising new paradigm to revitalize RVX-based therapies.

Aside from RVX, other promising anti-restenotic drugs with established toxicities or suboptimal biodistribution, such as sirolimus, paclitaxel and dexamethasone [1,66], may broadly benefit from our nanoplatform. Additionally, our recent publication as well as ongoing study suggest that the platelet membrane coating in our nanoplatform could also enable specific homing to other vascular lesions such as thrombosis and aneurysm [67]. As highlighted in our recent review, drug delivery to these diseased vasculatures is oftentimes limited due to the inadequate tissue penetration [68]. Indeed, the presence of intraluminal thrombosis, together with the multi-layered wall matrix structure, present significant hurdles to the successful translation of promising therapeutic agents. The ROS-detonable “cluster-bomb” design, on the other hand, may offer a convenient yet promising solution for the inadequate drug penetration and facilitate future therapeutic developments toward a broader vascular application.

The scale-up potential constitutes a critical determinant of the future translational prospect of nanoformulations [69]. Our nanoplatform consists of two major component: drug-loaded nanoparticles (an ROS-responsive nanocluster), and an outer biomembrane coating derived from platelets (a liposomal-like system). For the former, the ingredient chemicals are all commercially available, and the synthesis process could be conveniently escalated to accommodate for larger-scale manufacturing. Regarding the liposomal-like component of our nanodesign, we contend its large-scale production is highly feasible, based on our assessment of raw material supply as well as established industry infrastructure and guidelines on liposomal formulations. In our current design, the biomimetic liposomal coating is directly derived from platelets. Unlike the most nucleate cells that require sophisticated methodologies and lengthy process [70,71], the anucleate cells (e.g., platelets, red blood cells) allow for efficient membrane isolation at high purities within 30 min, hence highly amenable for standardization and streamlined operation [21,24]. Given the high yield as well as the outstanding immunocompatibility of platelet membranes as previously established, the supply source of platelets can be both allogeneic (e.g., blood bank) and autologous [72]. Alternatively, platelet membrane vesicles can be replaced by a conventional liposome coating, which would enable surface functionalizations in a more well-defined, bottom-up approach. Biomimetic membrane coating of the pre-synthesized nanoclusters — the very last step of the manufacturing process — can follow the industry standard practice in liposomal/biomembrane drug production For instance, membrane production and coating at clinical-scale level could be readily achievable through and manufacturing production (GMP) grade membrane extrusion facilities (e.g., Avanti Polar Lipids, Avestin).

5. Conclusion

Restenosis is the main culprit behind the high failure rate of endovascular interventions. The mainstream anti-restenotic solutions, such as DES and DCB, are plagued with grave safety concerns. In search of an alternative strategy that can potentially replace DES, we developed a highly innovative nanostructure judiciously designed for stent-free applications in cardiovascular diseases. This multimodal nanocluster possess several key features that are most preferred for anti-restenotic therapy, including targetability (through platelet membrane coating), deep tissue penetration and optimal biodistribution (through an ROS-detonable size transition), and ROS-responsive drug release (through the H40-PLA-based unimolecular nanoparticle). Overall, the rationally engineered platelet membrane-coated, ROS-detonable nanocluster could provide new insight into the optimal design of multifunctional nanomedicine specifically catered for non-invasive management of cardiovascular diseases.

Supplementary Material

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.jconrel.2021.08.025.