Sonodynamic therapy of glioblastoma mediated by platelets with ultrasound-triggered drug release

Abstract

Sonodynamic therapy (SDT) has aroused great interest for its potential in the treatment of glioblastoma (GBM). SDT relies on tumor-selective accumulation of a sonosensitizer that is activated by ultrasound irradiation (UI) to generate cytotoxic actions. The efficacy of GBM-SDT depends on sufficient sonosensitizer buildup in the tumor, which is, however, seriously hampered by the anatomical and biochemical barriers of the GBM. To overcome this difficulty, we herein propose a delivery strategy of ‘platelets with ultrasound-triggered release property’, which takes advantage of 1) the platelets’ ability to carry cargo and release cargo upon activation, and 2) the ROS-generating property of SDT. To provide proof of concept for the strategy, we first stably loaded platelets with IOPD-Ce6, a nano-formed sonosensitizer consisting of iron oxide nanoparticles coated with polyglycerol and doxorubicin and loaded with chlorine e6. UI of the IOPD-Ce6-loaded platelets (IOPD-Ce6@Plt) elicited ROS generation in the IOPD-Ce6@Plt, which were immediately activated to release IOPD-Ce6 into GBM cells in co-culture which, when subjected to a second time of UI, exhibited pronounced ROS production, DNA injury, viability loss, and cell death in the GBM cells. In the in vivo experiments, mice bearing intracranial GBM grafts exhibited substantial tumor distribution of IOPD-Ce6 following intravenous injection of IOPD-Ce6@Plt and subsequent UI at the tumor site. The GBM grafts then exhibited pronounced cell injury and death after another round of UI of the tumors. Finally, the growth of intra-cranial GBM grafts was significantly slowed when an SDT protocol consisting of an intravenous IOPD-Ce6@Plt injection followed by multiple times of tumor UI had been applied twice to the mice. Our results are strong evidence for the idea that platelets are sound and amenable carriers to deliver sonosensitizers in the GBM in an ultrasound-triggered manner and thus to produce highly targeted and effective SDT of GBM.

Keywords:

• Glioblastoma
• platelets
• ultrasound-controlled release
• targeted delivery
• sonodynamic therapy

Introduction

Glioblastoma (GBM) is the most malicious type of primary cerebral tumor among adults. The current standard care of GBM is very aggressive and consists of gross total resection, adjuvant chemotherapy and radiotherapy (Katsevman et al., 2019; Xie et al., 2019; Fisher & Adamson 2021). However, most treated GBM patients still experience recurrence and progression, and second-line therapies are scarce and often ineffective (Zanders et al., 2019; Stadlbauer et al., 2021). New treatment modalities and strategies are urgently needed to bring meaningful improvement of the prognosis of GBM patients. Among the many novel treatment approaches under investigation is sonodynamic therapy (SDT), which has attracted immense interest and promised great potential in the treatment of GBM (Wu et al., 2019; Shono et al., 2021). SDT of GBM works by the synergetic interactions of a sonosensitizer that is accumulated in the tumor and subsequent ultrasound irradiation (UI) that activates the sonosensitizer to generate tumoricidal effects. Both the sensitizer and the ultrasound exposure per se are nontoxic but generate cytotoxicity when applied in tandem (Behzadpour et al., 2020). SDT represents an emerging treatment modality that affords the possibility of tumor eradication in a noninvasive and site-targeted manner. The toxic mechanism of SDT largely remains unelucidated. The most endorsed theories involve the cavitation effect and closely related generation of reactive oxygen species (ROS), and sensitizer-dependent mechanical destabilization of the cell membrane (An et al., 2020; Nguyen Cao et al., 2021). Large amounts of ROS lead to oxidative damage of vital organelles like the nucleus, mitochondria, and endoplasmic reticulum, while destabilized cell membrane renders a cell more susceptible to shear force. As mentioned earlier, the success of GBM-SDT depends on the accumulation of a sonosensitizer in the GBM. However, like other agents for GBM therapy, the distribution and enrichment of sonosensitizers are severely sabotaged by the blood brain barrier and other physiological and biochemical barriers unique to the tumor tissue, e.g. an acidic extracellular milieu with a high interstitial fluid pressure. Various strategies and tactics have been explored to overcome this conundrum (Wang et al., 2017), and platelets have emerged as a novel and effective tool for drug delivery in the GBM (Li et al., 2022).

Platelets are the smallest and anucleate blood cells with vital functions in the primary hemostasis and other physiological processes like thrombosis, tissue regeneration, and wound repair (Golebiewska & Poole 2015). Recent studies have also identified platelets as key players in the initiation and progression of malignant tumors, including the GBM (Bohm et al., 2020; Aldaz & Arozarena 2021). Platelets have close interactions with GBM. For instance, GBM promotes thrombocytosis and recruitment of platelets which promote tumor angiogenesis through secretion of angiogenic factors and help to stabilize tumor vasculature (Di Vito et al., 2017; Campanella et al., 2020). Resting platelets can be loaded up via endocytosis and have a long blood circulating time. Once activated, circulating platelets within minutes rapidly form aggregates at the vascular endothelium and quickly release their cargo granules (Huang et al., 2016; Li et al., 2022). Importantly, either exogenous or endogenous ROS are efficient inducers of platelet activation, rendering sonodynamic effects (SDE) a particularly useful means to trigger platelet activation (Qiao et al., 2018). Moreover, platelets preparation and storage are relatively easy to perform, and standard guidelines therefor are long in place in clinical practice. Based on the aforementioned justifications, we propose a strategy of ‘platelets with ultrasound-triggered release property’ for GBM-targeted sonosensitizer delivery and SDT. Working steps of this strategy are proposed as follows: Platelets loaded with a sonosensitizer distribute to the GBM via blood circulation. Irradiation of the GBM with ultrasonic waves generates ROS in the platelets which are immediately activated and form aggregates in the GBM blood vessels, with rapid and massive discharge of the cargo sonosensitizer. Disruption of the vascular endothelium facilitates sonosensitizer spreading in the GBM tissue. Once the sonosensitizer reaches into the tumor cells, a second time of UI of the GBM causes sonodynamic damage and ultimately death of the tumor cells. This strategy requires that the sonosensitizer (1) is not toxic to platelets, (2) does not activate the platelets in the absence of an ultrasound field, and (3) can be stably contained in resting platelets in quantity. We had conducted preliminary experiments with the sonosensitizer chlorin e6 (Ce6) and found that human platelets can easily take up Ce6 and maintain viability, and UI (1 MHz, 0.5 W/cm², 30 s) can activate the Ce6-containing platelets (Ce6@Plt). However, the platelet loading capacity is disappointedly low for Ce6 and the internalized Ce6 suffers significant spontaneous extrusion, which renders free Ce6 unworthy of the proposed strategy. The dilemma was solved when we used iron oxide-based nanoparticles (IO-NPs) to carry Ce6 and then loaded this nano-form of Ce6 into the platelets. The IO-NPs have a surface layer of polyglycerol (PG) that affords good aqueous solubility. Doxorubicin (DOX), an agent with high affinity for the cell membrane, was conjugated to the PG via the hydrazone linkage to promote cell uptake of the nanoparticles. Ce6 was loaded to the nanoparticles to yield the nano-formed sonosensitizer IO-PG-DOX-Ce6 (IOPD-Ce6) (Li TF et al., 2021).

The present work was conducted using IOPD-Ce6@Plt to prove the concept of ‘platelets with -controlled release property’ for targeted SDT of GBM. Key steps and working conditions of the strategy were first characterized in in vitro experiments. The feature of IOPD-Ce6@Plt to target GBM in an ultrasound-controlled manner was then demonstrated on mouse intracranial GBM models. Finally, an SDT regimen utilizing IOPD-Ce6@Plt was designed and its therapeutic efficacy was demonstrated on intracranial GBM models.

Materials and methods

Materials

Certified fetal bovine serum (04-001-1ACS, VivaCell, Shanghai, China), DMEM medium (SH30022.01B, HyClone, Beijing, China), CHLORIN E6 (C829662, MACKIN, Shanghai, China), Rhodamine123 (R8030, Solarbio, Beijing, China), ROS Assay Kit (S0033S, Beyotime, Shanghai, China), Annexin-V-iFluor488/PI (CM001-100D, CHAMOT, Shanghai, China), γ-H2AX (ET1705-97, HUABIO, Hangzhou, China), p-ATM (ET1705-50, HUABIO, Hangzhou, China), GADPH (PMK053C, BioPM, Wuhan, China), CD 41 (bs-2636R, Bioss, Beijing, China), CD62p (ab83583, Abcam, Cambridge, UK), Annexin-V (DF6580, Affinity, USA), and D-luciferin, potassium salt (luc001, Sciencelight, Shanghai, China) were used in our study.

Cell models

U87-MG and U87MG-luciferase (U87-MG-luc) cells were purchased from the Cell Bank of Shanghai Institutes for Biological Sciences (Shanghai, China), The genotyping findings of the STR and amelogenin loci of the cell line U87-MG are provided in Figure S7. DMEM media containing 10% fetal bovine serum (QmSuero/Tsingmu Biotechnology, Wuhan) was used to maintain the cells. All cells were maintained in a humidified incubator with 5% CO₂ at 37 °C.

Animals

Human SJA Laboratory Animal CO., LTD (Hunan, China) provided female Balb/c nude mice that were 4–5 weeks old and weighed between 15 and 20 g. Animal handling and experimental procedures were in line with protocols approved by the Ethics Committee of Experimental Animal Welfare of Zhongnan Hospital of Wuhan University (Approval Document No.: ZN2021233). In our preliminary experiments, the U87 tumor grafts were found to be less proliferative and disruptive than other GBM models (e.g. U251 and GL-261) in our hands, so that the U87 tumor-bearing mice generally displayed a rather long-life span. Yet, most tumor-bearing mice also presented with epileptic symptoms 5 weeks after the implantation of GBM cells. Hence, to reduce animal suffering, we limited the duration of therapy experiment to 5 weeks, calculated from GBM cell implantation to therapy completion, whereupon all animals alive were humanely sacrificed.

IOPD-Ce6 and preparation of IOPD-Ce6@Plt

IOPD-Ce6 was synthesized and fully characterized according to protocols detailed in a previous work (Li TF et al., 2021). Ce6 is a second-generation sonosensitizer with single-tissue targeting and a distinct chemical structure. It is nontoxic and produces cytotoxicity when triggered by ultrasound (Liao et al., 2023). IOPD-Ce6 is composed of iron oxide nanoparticles with a surface coating of polyglycerol (PG) to which doxorubicin (DOX) molecules are connected via the hydrazone linkage. The PG coating affords good aqueous solubility and the attached DOX, which is hydrophobic, provides affinity to the cell membrane and induces formation of nanoclusters with hydrophobic interparticular spaces, facilitating Ce6 loading through physical adsorption. IOPD-Ce6 has an aqueous hydrodynamic diameter of 236.1 ± 4.62 nm, a surface charge of −24.73 ± 1.002 mV, and a zeta potential of −24.73 ± 1.002 mv. IOPD-Ce6 has excellent dispersibility in PBS.

Human platelets (Plt) were isolated from peripheral venous blood in healthy volunteers with informed consent and the collection of blood samples conforms to the agreement approved by the Medical Ethics Committee of Zhongnan Hospital of Wuhan University (Approval Document No.: Kelun [2022042K]). Platelets were separated by gradient centrifugation. Briefly, 5 ml of whole blood in an anticoagulant tube was centrifuged at 100 g for 15 min at room temperature to remove red blood cells at first. The platelet-containing layer at the top was moved to another tube where it was concentrated through 800 g centrifugation for 15 min to produce a soft pellet. Platelet number was calculated using a hemocytometer under a microscope. Isolated platelets were characterized by scanning electron microscopy and fluorescent staining of surface CD62P assayed by flow cytometry (Figure S1). Platelet number was calculated using a hemocytometer under a microscope. Platelets loaded with IOPD-Ce6, IOPD or Ce6 (IOPD-Ce6@Plt, IOPD@Plt or Ce6@Plt) were prepared by incubating the platelets (2 × 10⁵) with 1 μg/mL of IOPD-Ce6, IOPD or Ce6 in 1 mL of PBS at room temperature for 2 hours. Drug-loaded platelets were then pelleted by centrifugation (800 g, 15 min) and re-suspended into PBS, observed with confocal microscopy and assayed by spectrophotometry using a microplate reader (Spark 10M, TECAN Group, Männedorf, Switzerland) for the determination of loaded DOX and Ce6. Drug-loaded platelets were maintained in PBS for 4 and 8 hours at room temperature before being assayed for drug contents by flow cytometry. For viability assay, Rodamine 123, a fluorescent probe for mitochondrial membrane potential, was incubated with drug-loaded platelets in PBS for 30 min at room temperature (300 nM, R8030, Solarbio, Beijing, China) and the platelets were assayed by flow cytometry (Nambiar et al., 2020).

Sono-activation of IOPD-Ce6@Plt in vitro

IOPD-Ce6@Plt or platelets (2 × 10⁵ in 1 mL of PBS) in a 3.5-cm petri dish were irradiated with ultrasound (1 MHz, 0.5 W/cm², 30 s) before being assayed by flow cytometry for DOX and Ce6 fluorescence, immunofluorescent staining of surface CD62P and ROS generation. Alternatively, scanning electron microscopy (SEM) was used to visualize IOPD-Ce6@Plt and platelets both before and after ultrasonic irradiation.

Drug release from sono-activated IOPD-Ce6@Plt to in vitro GBM cells

Drug-loaded platelets in suspension (2 × 10⁵ IOPD-Ce6@Plt in 1 mL of PBS) were added into a 3.5-cm petri dish with U87MG cells (2 × 10⁶) with 80% confluence and the mixture was immediately irradiated with ultrasound (1 MHz, 0.5 W/cm², 30 s) and then subjected to time-lapse fluorescence microscopy. Alternatively, the mixture was maintained for 60 min before taking out for flow cytometry analysis of drug fluorescence in the U87-MG cells. To evaluate IOPD-Ce6@Plt-mediated sonodynamic toxicity to the U87-MG cells, the mixture of IOPD-Ce6@Plt and U87-MG was ultrasound-irradiated (1 MHz, 0.5 W/cm², 30 s) and the culture media was replaced with fresh culture medium an hour later. A second time of UI (1 MHz, 1 W/cm², 30 s) was then applied to the U87MG cells. The U87MG cells were either immediately taken out for detection of intracellular ROS generation, or taken out 6 h later for evaluation of cell viability, DNA damage, and cell death. 2′,7′-dichlorofluorescin diacetate (DCFDA) staining and flow cytometry were used to measure ROS generation. Rhodamine 123 labeling and flow cytometry were used to measure cell viability. DNA damage was evaluated by western blotting analysis of phospho-ataxia telangiectasia mutated kinase (p-ATM) and H2A histone family member X (γ-H2AX). Annexin-v labeling and flow cytometry were used to measure cell death.

GBM-targeted drug delivery mediated by sono-activated IOPD-Ce6@Plt

Intracranial GBM models were replicated in mice for demonstration of tumor-targeted drug delivery by ultrasound-activated IOPD-Ce6@Plt. Briefly, balb/c nude mice with shaved heads under nembutal anesthesia were secured in a stereotactic head holder. 1 mm in front of the coronal suture and 2 mm to the right of the bregma, a burr hole was carried out in the mouse brain to a depth of 1.5 mm below the cortical surface. U87-MG-luc cells in log phase (1 × 10⁵) uspended in 6 μL of PBS were then slowly injected through the burr hole over 5 min. Tumor growth was monitored by bioluminescent imaging. Briefly, D-luciferin potassium salt (150 mg/kg, luc001, Sciencelight, Shanghai, China) was intraperitoneally injected into mice under nembutal anesthesia and bioluminescent imaging (Bruker In-Vivo Xtreme) was performed 10 min after the injection. Animals on day 20 post implantation were used in later drug delivery experiments. Animals bearing intracranial tumors were randomly assigned to five groups (6 mice per group) with designated treatments: ① UI-1, ② Ce6 + UI-1, ③ IOPD-Ce6 + UI-1, ④ IOPD-Ce6@Plt, and ⑤ IOPD-Ce6@Plt + UI-1. The animals, according to their designation, received intravenous injections of Ce6, IOPD-Ce6 or IOPD-Ce6@Plt each in 200 μL of PBS per mouse. A dosage of 0.66 g of Ce6 per mouse weighing 20 g (0.033 mg/kg b.w.) was determined and delivered by 2 × 10⁵ platelets in 200 µL of PBS. All animals but the ④ IOPD-Ce6@Plt group then received the first time of UI (UI-1: 1 MHz, 0.5 w/cm², 30 s) at the tumor location 3 hours post injection. Three mices from each group were sacrificed an hour after the UI-1 and the brains were taken for fluorescence imaging. The remaining 3 animals in all groups but the ④ IOPD-Ce6@Plt group received a second time of UI (UI-2: 1 MHz, 1 w/cm², 30 s) at 1 h after the UI-1. All animals were sacrificed 24 hours after the UI-2 and the brains were collected for fluorescence imaging, IHC staining of annexin-V (DF6580, Affinity, USA), p-ATM and γ-H2AX.

GBM-targeted SDT mediated by sono-activated IOPD-Ce6@Plt

Intracranial GBM grafts of U87-luc cells were produced as described above. The tumor-bearing mice were divided into 6 groups of 5 mice each on day 21 after implantation. Four groups of animals were intravenously injected with 200 μL per mouse of PBS containing ① platelets (Plt), ② Ce6, ③ IOPD-Ce6, and ⑥ IOPD-Ce6@Plt, respectively. The dosage was determined to be 0.66 g of Ce6 per moue weighing 20 g (0.033 mg/kg b.w.), and delivered by 2 × 10⁵ platelets in 200 µL of PBS. The animals were then subjected to a succession of 4 times of extracranial UI at intervals at the tumor site (Figure 7A). The initial UI (Ul-1: 1 MHz, 0.5 w/cm², 30 s) was performed at 3 h after drug injection in order to induce IOPD-Ce6@Plt activation and ensuing IOPD-Ce6 release in the intracranial GBM. One hour later when the released IOPD-Ce6 was supposed to have achieved adequate distribution in the GBM tissue, there more times of UI (UI-2, UI-3 and UI-4: 1 MHz, 1 w/cm², 30 s) were performed at intervals of 24 h to achieve SDT efficacy. The full protocol included one i.v. drug injection followed by 4 times of extracranial UI and took 3 days to complete (Figure 8A). The day that the first drug injection was performed was designated as day 1. The full protocol was repeated once from day 8 to day 10. For control, another group of animals (④) was injected with IOPD-Ce6@Plt but received no UI, and a third IOPD-Ce6@Plt group (⑤) only received the first time of UI for each treatment. Growth of brain tumors was monitored by bioluminescence imaging for the first weeks of the experiment period. For imaging, mice under nembutal anesthesia were administrated with luciferin (D-Luciferin potassium salt, 150 mg/kg, Sciencelight) via injection in the abdomen, and then imaged 10 minutes after the injection. On the 15th day post the first drug injection, all animals were sacrificed and brains were harvested for ultrasound imaging to evaluate the tumor volume and brain tissue sections were prepared for IHC staining of annexin-v, p-ATM and γ-H2AX, and fluorescence microscopy. The animals’ vital organs, including the heart, liver, spleen, lung, and kidney, were collected for H&E staining. Figure 1 shows a flow chart framework of the SDT experiments.

Figure 1. The whole experimental flow chart.

Statistical analysis

Data were analyzed by One-way analysis of variance (one-way ANOVA). Data were presented as mean ± standard deviation (SD). A two-tailed p < .05 was considered statistically significant.

Results

Preparation and characterization of IOPD-Ce6@Plt

Synthesis and characterization of IOPD-Ce6 are described in detail in our previous published work (Li TF et al., 2021). IOPD-Ce6@Plt were prepared by incubating human platelets (2 × 10⁵ platelets in 1 mL PBS) with IOPD-Ce6 (1 μg/mL) at room temperature for 2 hours. Efficiency and capacity of drug loading were 66% and 0.66 μg per 2 × 10⁵ platelets, respectively, calculated to Ce6. Confocal microscopy of IOPD-Ce6@Plt showed distinct Ce6 and DOX fluorescence from the platelets indicating uptake of IOPD-Ce6 (Figure 2A). Spontaneous release of internalized IOPD-Ce6 by IOPD-Ce6@Plt in fresh culture medium was insignificant as there was little change in DOX fluorescence over an incubation period of 8 hours (Figure 2B). There was an apparent elevation of Ce6 fluorescence in the IOPD-Ce6@Plt (Figure 2C), which might be owing to decreased self-quenching resulting from detachment and diffusion away of Ce6 from IOPD-Ce6. Finally, IOPD-Ce6@Plt showed uncompromised viability after 8 hours of maintenance in fresh culture medium, as indicated by rhodamine 123 staining (Figure 2D).

Figure 2. Characterization of platelets (Plt) loaded with IOPD-Ce6 (IOPD-Ce6@Plt). A: Fluorescent microscopic imaging of IOPD-Ce6@Plt. B, C: Drug fluorescence in platelets (Plt) loaded with IOPD-Ce6, IOPD, or Ce6 (IOPD-Ce6@Plt, IOPD@Plt, and Ce6@Plt) after incubation with PBS. D: Viability of platelets (Plt) loaded with IOPD-Ce6 (IOPD-Ce6@Plt) assayed by mitochondrial staining of rhodamine 123. Values are means ± standard deviation (SD). Representative flow cytometry histograms or dot plots for B, C & D are presented in Figure S2(B, C & D).

UI induced activation, aggregation and payload unloading of IOPD-Ce6@Plt

Then, we found that UI (1 MHz, 0.5 W/cm², 30 s) stimulated ROS generation in IOPD-Ce6@Plt (Figure 3B) and immediately triggered pronounced activation of IOPD-Ce6@Plt, as indicated by the dramatic morphological change and aggregation of the IOPD-Ce6@Plt and the increased surface expression of CD62P, which is a known marker of platelet activation (Figure 3A, C). Once activated, platelets experience morphological alterations (e.g. presenting pseudopodia), form aggregates, become highly adhesive, and secrete various pre-stored proteins and chemicals. CD62P is a protein whose presence in the platelet surface is drastically increased upon platelet activation (Gomes de Azevedo-Quintanilha et al., 2022). Note that UI and the loaded IOPD-Ce6 separately did not cause significant platelet activation (Figure 3A–C). Consistently, the IOPD-Ce6@Plt displayed marked unloading of IOPD-Ce6 upon UI, as indicated by the decreased fluorescence of both DOX and Ce6 (Figure 3D, E).

Figure 3. Sono-activation of platelets (Plt) loaded with IOPD-Ce6 (IOPD-Ce6@Plt). A: Scanning electron microscopy (SEM) of Plt, IOPD-Ce6@Plt, and ultrasound-irradiated Plt and IOPD-Ce6@Plt. B: Effect of ultrasound irradiation (UI) on ROS generation in Plt and IOPD-Ce6@Plt. C: Effect of UI on surface expression of CD62P in Plt and IOPD-Ce6@Plt. D, E: Effect of UI on DOX and Ce6 contents in IOPD-Ce6@Plt. Values are means ± standard deviation (SD). (n = 3, *p < .05, **p < .01). Representative flow cytometry histograms for B–E are presented in Figure S3(B–E).

Ultrasound-activated IOPD-Ce6@Plt released IOPD-Ce6 to co-cultured GBM cells which exhibited pronounced sonodynamic toxicity

We subsequently added IOPD-Ce6@Plt to cultured U87-MG human GBM cells and immediately applied UI (1 MHz, 0.5 W/cm², 30 s) to the co-culture. Time-lapse fluorescence microscopy (Figure 4A) and flow cytometry (Figure 4B, C) showed accelerated and higher uptake of IOPD-Ce6 in the U87-MG cells in the ultrasound-irradiated co-culture, indicating that UI promoted the IOPD-Ce6@Plt discharging payload into the U87-MG cells. Consequently, another round of UI of the co-culture (1 MHz, 1 W/cm², 30 s), 1 hour after the first round (Figure 4D), resulted in a spike in ROS production, mitochondrial membrane potential (MMP) collapse, DNA damage, and cell death (Figure 4E–H).

Figure 4. Sono-activated IOPD-Ce6@Plt release IOPD-Ce6 to co-cultured GBM (U87MG) cells. A: Still images of time-lapse observation of U87MG cells in co-culture with IOPD-Ce6@Plt (with or without UI) for 60 min, U87MG cells co-cultured with naïve platelets were used as control. B, C: DOX and Ce6 contents in the U87MG cells in co-culture with IOPD-Ce6@Plt (with or without UI). D: Experimental protocol for demonstrating IOPD-Ce6@Plt-mediatd sonodynamic toxicity to U87MG cells. Briefly, IOPD-Ce6@Plt were mixed with U87MG cells and the mixture received two times of UI at an interval of 1 h. The U87MG cells were taken out after the second time of UI for subsequent analysis. Mixtures with no UI or only with the 1^st or 2^nd time of UI were set up for control. Indicators analyzed are shown as follows. E: ROS levels in the U87MG cells. F: Viability of the U87MG cells assayed by rhodamine 123 staining. G: DNA damage in the U87MG cells assayed by western blotting analysis of p-ATM and γH2AX. H: Apoptosis of the U87MG cells assayed by annexin V fluorescent staining. Values are means ± standard deviation (SD). (n = 3, *p < .05, **p < .01). Representative flow cytometry histograms or dot plots for B, C, E, F, and H are presented in Figure S4(B, C, E, F, and H).

Ultrasound-activated IOPD-Ce6@Plt resulted in extensive distribution of Ce6 in GBM grafts which exhibited massive sonodynamic toxicity

In order to verify the in vitro findings, we replicated intracranial GBM grafts of U87-MG cells in mice. The animals were given IOPD-Ce6@Plt (0.66 μg of Ce6 per mouse of 20 g, delivered by 2 × 10⁵ platelets in 200 µL PBS) via tail vein injection and, 3 h later, received UI (1 MHz, 0.5 W/cm², 30 s) at the tumor sites (Figure 5A). This UI was meant for sonodynamic activation of IOPD-Ce6@Plt in the GBM vasculature to release IOPD-Ce6. At 1 hour after the UI, the GBM grafts that received IOPD-Ce6@Plt exhibited extensive presence of Ce6 and DOX in the tumor tissue (Figure 5B, C), indicating IOPD-Ce6 distribution in the GBM tissue. A second time of UI (1 MHz, 1 W/cm², 30 s), at 1 h after the first time of UI (Figure 6A) and meant for inducing sonodynamic toxicity in the GBM, was applied to the IOPD-Ce6@Plt-treated tumors. The tumors then displayed intensified fluorescence of Ce6 and DOX 24 h later (Figure 6B, C), indicating accumulation of IOPD-Ce6 in the GBM. Note that the GBM grafts that received Ce6 with UI, or IOPD-Ce6 with UI, or IOPD-Ce6@Plt without UI exhibited little sign of IOPD-Ce6 distribution and accumulation (Figures 5B, C and 6B, C). Importantly, the tumors that received IOPD-Ce6@Plt and two rounds of UI displayed pronounced cell (DNA) damage and death, as shown by the H&E staining (Figure 7D) and the enhanced expression of p-ATM, γ-H2AX, and annexin-V (Figure 7A–C). The above observations strongly suggest that IOPD-Ce6@Plt in tandem with site-directed UI could achieve highly GBM-targeted delivery of IOPD-Ce6 and thereby induce sonodynamic toxicity of GBM with efficiency and effectiveness.

Figure 5. GBM-targeted delivery of IOPD-Ce6 by sono-activated IOPD-Ce6@Plt. A: Experimental protocol. Briefly, intracranial GBM-bearing mice were intravenously injected with IOPD-Ce6@Plt and, 3 h later, received extracranial UI at the tumor site (⑤). the animals were sacrificed 1 h after the UI and brains were resected for subsequent analysis. Animals received PBS plus UI (①), Ce6 plus UI (②), IOPD-Ce6 plus UI (③) and IOPD-Ce6@Plt without UI (④) were used as controls. B: Fluorescent imaging of ex vivo brains showing Ce6-derived fluorescence. C: Confocal fluorescent microscopy of brain tissue slides showing DOX- and Ce6-derived fluorescence. Enlarged micrographs of group ⑤ are provided in Figure S8 showing the cytoplasmic distribution of DOX and Ce6 fluorescence around the nuclear staining by DAPI.

Figure 6. GBM-targeted delivery of IOPD-Ce6 by sono-activated IOPD-Ce6@Plt. A: Experimental protocol. Briefly, intracranial GBM-bearing mice were intravenously injected with IOPD-Ce6@Plt and, 3 h later, received 2 times of extracranial UI at the tumor site at an interval of 1 h (⑤). the animals were sacrificed 24 h after the UI and brains were resected for subsequent analysis. Animals received PBS plus UI (①), Ce6 plus UI (②), IOPD-Ce6 plus UI (③) and IOPD-Ce6@Plt without UI (④) were used as controls. B: Fluorescent imaging of ex vivo brains showing Ce6-derived fluorescence. C: Confocal fluorescent microscopy of brain tissue slides showing DOX- and Ce6-derived fluorescence.

Figure 7. Sonotoxicity of GBM mediated by sono-activated IOPD-Ce6@Plt. IHC staining of annexin-V (A) (marker of apoptosis) and γH2AX (B), p-ATM (C) (markers of DNA damage) and H&E staining (D) in GBM tumors at the end of therapy. Treatment protocol is shown in Figure 6(A).

SDT mediated by IOPD-Ce6@Plt exhibited anti-GBM efficacy

The therapeutic anti-GBM efficacy of IOPD-Ce6@Plt-mediated SDT was demonstrated on mice bearing intracranial U87-MG graft tumors. A SDT regimen was designed based on the findings presented in Figures 5–7. Four groups of GBM-bearing animals (①②③⑥) were intravenously injected with 200 μL per mouse of PBS containing platelets (Plt), Ce6, IOPD-Ce6, and IOPD-Ce6@Plt, respectively. The dosage was determined to be 0.66 g of Ce6 per mouse weighing 20 g, delivered by 2 × 10⁵ platelets in 200 µL of PBS. The animals were then subjected to a succession of 4 times of extracranial UI (UI-1:1 MHz, 0.5 W/cm², 30 s, UI-2, UI-3 & UI-4:1 MHz, 1 W/cm², 30 s) at the tumor site (Figure 8A). The first time of UI was performed at 3 h after drug injection in order to trigger IOPD-Ce6@Plt activation and ensuing IOPD-Ce6 discharge in the GBM. One hour later when the discharged IOPD-Ce6 had achieved distribution within the GBM, UI was performed for another 3 times at an interval of 24 h to obtain SDT efficacy. The treatment protocol took 3 days to complete and was repeated once (Figure 8A). For control, another group of GBM-bearing mice (④) received IOPD-Ce6@Plt injection but no UI, and a third group (⑤) received only the first time of UI in each treatment. As can be seen in Figure 8B–D and Figure S9, two rounds of IOPD-Ce6@Plt-mediated SDT significantly slowed the growth of intracranial GBM, leading to the smallest tumor volumes as compared with other control groups at the end of the experiment. Two IOPD-Ce6@Plt injections with only UI-1 achieved a similar effect. IHC staining confirmed significant cell damage and death in the tumors that received IOPD-Ce6@Plt-mediated SDT, as indicated by the enhanced expression of p-ATM, γ-H2AX and annexin-V (Figure 9). The above observations are solid evidence for the efficacy of IOPD-Ce6@Plt -mediated SDT.

Figure 8. Therapeutic efficacy of anti-GBM PDT mediated by sono-activated IOPD-Ce6@Plt. A: Experimental protocol. Briefly, four groups of GBM-bearing animals were intravenously injected with platelets (Plt) (①), Ce6 (②), IOPD-Ce6 (③), and IOPD-Ce6@Plt (⑥), respectively. The animals were then subjected to a succession of 4 times of extracranial UI at intervals at the tumor site. The treatment protocol took 3 days to complete and was repeated once. For control, another group of animals injected with IOPD-Ce6@Plt but with no UI (④) and a third group received only the first time of UI for each treatment (⑤). B: Fluorescent imaging of in vivo GBM tumors at day 1, day 7 and day 14 into therapy. C: Intensity of GBM tumor fluorescence at the end of therapy. D: Gross examination of GBM tumors and their host brains at the end of therapy. Values were means ± SD (n = 5).

Figure 9. Cell damage and death in GBM tumors at the end of therapy, as shown by IHC staining of γH2AX, p-ATM (markers of DNA damage) and annexin-V (marker of apoptosis). therapy protocol is shown in Figure 8(A).

It is worth noting that GBM tumors that received IOPD-Ce6 and IOPD-Ce6@Plt alone demonstrated appreciable drug retention at the end of experiment which was 8 days after the last time of drug administration (Figure 10A ③&④). But UI markedly increased drug retention in tumors that received IOPD-Ce6@Plt, as shown in Figure 10A ⑤&⑥. These observations are in good agreement with observations shown in Figures 5 and 6 and they combined are compellingly evidence that IOPD-Ce6@Plt per se are capable of GBM-targeting and retention, which can be dramatically enhanced by UI at the tumor site.

Figure 10. Drug retention in GBM tumors at the end of therapy. A: Confocal fluorescent microscopy of brain tissue slides showing DOX- and Ce6-derived fluorescence. B: HE staining of tumor tissues at the end of therapy. Therapy protocol is shown in Figure 8(A).

Discussion

SDT has emerged as a promising GBM therapeutic approach that harnesses the energy of US for tumor disruption or destruction via the mediation of a sonosensitizer. SDT cannot achieve ideal tumor elimination due to poor sonosensitizer distribution and accumulation in the GBM. Several approaches have been raised to solve this challenge. Liang et al. (2020) constructed a smart nanoplatform for SDT using holo-transferrin (holo-Tf) with in situ growth of MnO₂ nanocrystals prepared by a modified mild biomineralization process. In our present work, we proposed the ‘platelets with ultrasound-triggered release property’ for GBM-targeted drug delivery and SDT. For demonstrating the feasibility and effectiveness of this strategy, IOPD-Ce6, a nano-formed sonosensitizer was loaded in platelets to yield the device of IOPD-Ce6@Plt with ultrasound-controlled release property. We first delineated the key steps and conditions to realize IOPD-Ce6@Plt -mediated SDT of GBM. We showed that IOPD-Ce6@Plt have a vastly increased loading capacity and efficiency over Ce6@Plt and could maintain viability, loading stability and a resting state for 8 hours. Importantly, UI (1 MHz, 0.5 W/cm², 30 s) efficiently triggered ROS production in the IOPD-Ce6@Plt, resulting in immediate activation, aggregation, and discharge of IOPD-Ce6 into co-cultured mouse GBM cells. The GBM cells then displayed significant ROS production, DNA injury, viability loss, and cell death following a second round of UI. Next, we demonstrated the ultrasound-controlled GBM-targeting property of IOPD-Ce6@Plt in animal experiments. Intravenous injection of IOPD-Ce6@Plt with subsequent tumor-directed UI caused rapid and massive distribution of IOPD-Ce6 in intra-cranial GBM tumors which displayed substantial cell damage and death consequent after a second time of UI. Based on the above findings, we designed a SDT protocol consisting of one intravenous IOPD-Ce6@Plt injection and a subsequent succession of extracranial UI at the tumor site. A SDT regimen carrying out this protocol effected significant efficacy against intra-cranial GBM tumors in mouse. The SDT regimen appeared to be safe without causing damage to any vital organs.

Platelets as carriers have two critically amenable properties contributing to the GBM-targeting property of IOPD-Ce6@Plt. Firstly, platelets can take up and be stably loaded with IOPD-Ce6. Stability of drug loading i.e. the absence of spontaneous payload discharge is a feature that marks platelets out from other cell-based carriers we have tested, such as monocytes, macrophages, and dendritic cells, all of which when loaded have a strong propensity of spontaneous deloading (Li et al., 2017, 2018; Wang et al., 2018). Secondly, and most critically, platelets are capable of ROS-induced activation and ensuing aggregation and degranulation. On one hand, massive ultrasound-induced IOPD-Ce6@Plt aggregation in the tumor blood vessels provides a vast reservoir wherefrom large amounts of IOPD-Ce6 are quickly released into the tumor tissue as the platelets degranulate. One the other hand, aggregates of IOPD-Ce6@Plt and platelets clot the tumor blood vessels preventing the exodus of IOPD-Ce6 and Ce6, which also contributes to Ce6 build-up in the tumor. All these features, plus tumor-directed UI, afford the sono-controllability, GBM-targeting ability, speedy release, and massive intra-tumor accumulation of IOPD-Ce6@Plt-mediated drug delivery.

On the side of IOPD-Ce6, although IOPD worked very well as a vehicle of Ce6 in this study, it is neither unique nor superior to other vehicles for platelet-mediated drug delivery. We have also used boron nitride (BN)-based nanoparticles to carry Ce6 for platelet-mediated and GBM-targeted delivery (submitted elsewhere). The therapeutic efficacy of IOPD-Ce6@Plt-mediated SDT is believed to be mostly due to the sonodynamic toxicity produced from the IOPD-Ce6 delivered in the GBM, as demonstrated in the in vitro experiments. Still, other factors might have contributed to the in vivo anti-GBM efficacy. Firstly, DOX molecules can be slowly released from the IOPD and engender tumor toxicity. Secondly, GBM blood vessels clotted by aggregated IOPD-Ce6@Plt and platelets may lead to or exacerbate ischemic damage of the GBM tissue. Regarding the blood vessel clotting, we were initially concerned that IOPD-Ce6@Plt-mediated GBM-SDT might also cause ischemic injury to neighboring brain tissues. Nevertheless, peritumoral brain tissues of all treated mice presented no apparent ischemia damage (Figure S5). From a translational perspective, the risk of sonodynamic damage to peritumoral brain tissues can be minimized in patients by the precision delivery of ultrasonic waves in the GBM via an endoscopic ultrasound device. Vital organs in the periphery did not show any sign of injury, either (Figure S6).

The sonodynamic effect (SDE) and resultant generation of ROS are foundational to tumor SDT. SDE plays double, vital roles in the IOPD-Ce6@Plt-mediated SDT of GBM. On one hand, SDE generates ROS in the IOPD-Ce6@Plt which is thereby activated to discharge IOPD-Ce6, contributing to highly controlled and targeted delivery to the GBM. On the other hand, SDE produces toxic ROS levels in the cancer cells that have internalized IOPD-Ce6@Plt, causing organelle injury and cell death. These rationales were substantiated in both in vitro and in vivo experiments and a SDT protocol designed therefrom has proved efficacious in treating mouse intracranial GBM models. The protocol consisted of one intravenous injection of IOPD-Ce6@Plt shortly followed by one time of tumor-directed UI for drug delivery and a further three times of UI for inducing sonodynamic GBM toxicity. It should be noted that the GBM-directed UI was applied in an extracranial manner in this work. This approach was adopted mainly for demonstrating the potential of the SDT strategy exemplified by IOPD-Ce6@Plt as a stand-alone GBM-targeted therapy on mouse models. In clinical practice, it may be difficult for extracranially applied ultrasound to penetrate the human cranium. But an endoscopic ultrasound device may go through an opening in the cranium to deliver ultrasonic waves into the GBM under stereotactic guidance. This approach is supposed to have a high level of therapeutic precision, ensuring maximal anti-GBM efficacy and minimal brain tissue damage at the same time.

The present study innovatively combines ROS-activatable platelets and ROS-generating SDE for drug delivery in GBM. This strategy features GBM-targeting, controllability, speediness, and efficiency, and thus could vastly improve the efficacy and potency of anti-GBM SDT. It is worth noting that the dosage of Ce6 used in this work is 0.033 mg/kg (bw) which is far below the frequently reported 5 mg/kg (bw) dosage range (Qu et al., 2020). This strategy can also be extended to deliver other drugs, such as chemotherapeutic agents, to both central and peripheral solid tumors. There have been reports of platelets exploited as vehicles for drug delivery to peripheral tumors (Han et al., 2019; Li H et al., 2021). However, platelet-mediated passive delivery is much less GBM-targeted and less efficient than the active delivery mediated by SDE-activatable platelets, as demonstrated in the present work. Recently, photothermal effect (PTE) has been exploited to activate drug-loaded platelets for active drug delivery to peripheral tumors (Pei et al., 2020; Li L et al. 2021; Lv et al., 2021). However, SDE is advantageous over PTE in at least two aspects. On one hand, the irradiation time is much shorter for SDE than for PTE. Thirty seconds of UI at 1 MHz, 0.5 W/cm² in our work is enough to elicit effective SDE in platelets and the GBM cells as well, both in vitro and in vivo, whereas 10 min of laser irradiation is needed to induce PTE. On the other hand, SDE does not produce high temperature (>55 °C) seen with PTE, thus sparing neighboring tissues from thermal injury. We have also recently proposed a strategy of ‘platelets with light-controlled release property’ for GBM-targeted drug delivery and photodynamic therapy (PDT) (Xu et al., 2022). This strategy shares the same mechanism of ROS-triggered activation and deloading of loaded platelets, but uses the photodynamic effects (PDE) instead of the SDE to generate ROS. Poor tissue penetration of light poses a challenge to the utility of PDE while ultrasound travels much farther and more easily in most tissues than light. Intriguingly, our ongoing work also suggest that GBM cells are more vulnerable to SDE than their associated macrophages (data not shown).

Conclusion

To summarize, the strategy of ‘platelets with ultrasound-controlled release property’ exemplified by IOPD-Ce6@Plt is an innovative and efficacious tumor-targeted delivery approach with translational potential.

Author contributions statement

Conception and design: MYW and XC; Methodology: MYW, HZX, TFL and SYC; Analysis and interpretation of the data: KL, QZ and LZ; The drafting of the paper: MYW and XC; Revising it critically for intellectual content: MYW, JCC and XC; Experimental planning, data interpretation and manuscript review: JCC and XC; The final approval of the version to be published: ALL. All authors agree to be accountable for all aspects of the work. 

Ethical approval statement

Animal handling and experimental procedures were in line with protocols approved by the Ethics Committee of Experimental Animal Welfare of Zhongnan Hospital of Wuhan University (Approval Document No.: ZN2021233). To validate the in vitro results, we replicated intracranial glioblastoma grafts of U87-MG cells in mice. Female Balb/c nude mice at 4–5 weeks of age (15-20 g) were obtained from Hunan SJA Laboratory Animal CO., LTD (Hunan, China) and were fed on standard diets and water at a temperature of 18–25 °C and 60–65% humidity before experiments. Intracranial glioblastoma models were established in mice under nembutal anesthesia. The ARRIVE criteria were followed in all animal experiments. Human platelets (Plt) were isolated from Peripheral venous blood in healthy volunteers with informed consent and the collection of blood samples conforms to the agreement approved by the Medical Ethics Committee of Zhongnan Hospital of Wuhan University (Approval Document No.: Kelun [2022042K]).

Data availability statement

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

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China (No. 81771280) and China Scholarship Council.