Skip to main content
NCBI home page
Scientific Reports logoLink to Scientific Reports
. 2026 Jun 18;16:27892. doi: 10.1038/s41598-026-58290-8

Alkaline loading of extracellular vesicles produced from human neural stem cell-derived neurospheres enables CNS drug delivery

Amar M Singh 1,✉, Charles M White 2, Adeline Phillips 2, Logan P Crowe 2, Robert Marti 2, Morgan C Finnerty 3, Martonio Ponte Viana 2, William Antoniades 2, Michael G Bartlett 3, Viviana Martinez 2, Raymond Swetenburg 2, Steven L Stice 1,2,✉
PMCID: PMC13547375  PMID: 42315580

Abstract

The blood brain barrier and blood tumor barrier (BBB and BTB, respectively) represent significant obstacles for the delivery of drugs to treat diseases of the central nervous system, such as brain cancers and neurodegenerative diseases. Extracellular vesicles (EVs) or exosomes have emerged as a new drug delivery vehicle for CNS diseases as they may penetrate the BBB/BTB and are less immunogenic than liposomal carriers. EVs derived from human neural stem cells (hNSC) provide additional benefits over other EV sources due to their increased homing capability to neural cells and demonstrated efficacy for treating stroke and traumatic brain injury in rodent models. However, the utilization of EVs from hNSC for drug delivery remains largely unexplored, due in part to difficulties in manufacturing capacity compared to traditional cell lines. Here, we report the development of a hNSC suspension neurosphere system for EV production and drug delivery. As proof of concept, doxorubicin was loaded into hNSC-EV, using a novel, high-efficiency alkaline passive loading method, and shown to be effective at inducing cytotoxicity in glioma cells in vitro and exhibiting higher BBB penetrance than doxorubicin-alone in vivo. These studies demonstrate the potential for hNSC-EV loaded doxorubicin as a therapeutic treatment for brain cancers such as glioblastoma, while also establishing hNSC-EVs as a drug-delivery vehicle for CNS diseases.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-58290-8.

Keywords: Extracellular vesicles, Exosomes, Doxorubicin, Neural stem cells, Neural progenitor cells, Blood–brain barrier, Glioblastoma, Drug delivery, Alkaline loading

Subject terms: Biotechnology, Cancer, Drug discovery, Neuroscience

Introduction

The blood–brain barrier (BBB) serves a critical function in protecting the brain and spinal cord from toxins and other chemical insults, while maintaining the necessary homeostatic balance to maintain oxygen and nutrient levels1. However, due to this strict control between the vasculature and the central nervous system (CNS), the delivery of therapeutics to treat CNS diseases is severely hampered2. Furthermore, upon the development of a primary brain tumor or brain metastasis from a non-CNS tumor, a blood-tumor barrier (BTB) is established. Unlike the BBB, the BTB is heterogeneous and often “leaky” due to the lack of tight junctions and reduced supporting cells3. However, similarly to the BBB, the BTB serves a restrictive role in preventing the delivery of therapeutics at optimal concentrations, which could otherwise be used to treat and prevent further tumor progression4.

Over the past decade, multiple approaches, including receptor-mediated transcytosis, nanoparticles and extracellular vesicles (EVs), have been evaluated to enable the delivery of drugs through the BBB/BTB to the CNS with varying levels of success2,5. The method that has been most intensely investigated is the use of the Transferrin receptor (TfR), which relies on receptor-mediated transcytosis to transport a cargo through the BBB5. Several caveats to this approach exist however, such as lack of brain specificity6, TfR recycling partially preventing exocytosis in the brain7,8, and lysosomal degradation9. Several other receptors are still being evaluated, with most still in the pre-clinical stage2,5. Nanoparticle delivery offers some utility for CNS drug delivery, as drugs can be loaded directly in a scalable manner and the nanoparticles can be chemically-modified to facilitate CNS-targeting10. Weaknesses with nanoparticles, however, such as rapid clearance, immune-tolerability, and the lack of an efficient, validated brain targeting method limit their current uses for CNS diseases. EVs, on the other hand, may offer several key advantages over nanoparticles for drug delivery through the BBB/BTB11.

EVs, such as exosomes and microvesicles, have been reported to traverse the BBB through mechanisms that are still yet to be fully elucidated12,13. Irrespective to the mechanism, EVs provide a significant opportunity for the delivery of cargo, including small molecules, RNA and protein, to the CNS for disease treatment. Indeed, numerous studies have described methods for loading EVs with various cargo using exogenous (e.g. sonication, electroporation, passive incubation) or endogenous (genetic modification of the EV-donor cell) methods14. There are several clinical trials that have been initiated or are currently ongoing evaluating the efficacy of drug-loaded EVs in non-CNS indications, especially for cancer treatments15.

An important aspect of EV production for drug-delivery is the choice of donor cells. To date, most of the therapeutic development around EVs has focused on utilizing HEK293 cells or mesenchymal stromal cells (MSC) as the donor cell type, as these cells are easy to grow in a high-production capacity. However, for treating CNS-diseases, these cells may not serve as an ideal choice as several studies have indicated that EVs possess natural homing capability to their tissue and/or cell-type of origin16,17. For example, it has been demonstrated that EVs from lung, liver and brain tumor cells exhibit preferential uptake into their organ of origin18. Potentially, EVs from neural cell types, such as human neural stem cells (hNSC), may provide increased uptake into CNS cells compared to MSC as they more closely align with the cell of origin. In support of this, the therapeutic efficacy of hNSC-derived EVs outperformed MSC-EV in a murine embolic stroke model19. EVs from other neural cell types, such as astrocytes, neurons and microglia have also been described20,21. One weakness with utilizing these cell types, either from primary or hPSC sources, versus hNSCs for EV production is that astrocytes, neurons and microglia are unable to self-renew and be cultured long-term22,23, and therefore have significant scalability constraints limiting their use for therapeutic development.

Human neural stem cells (hNSC) serve as a valuable donor cell type for EV production due to their self-renewal capability with moderate-to-high proliferation rate, karyotypic stability upon long-term passaging and neural targeting capabilities24–29. These neural EVs have been successfully validated as a therapeutic in animal models of stroke and traumatic brain injury19,30–33, and may therefore provide added benefits for some CNS diseases. Further, cell aggregates in suspension have been scaled in large cell bioreactors34–36. Therefore, we expanded the potential of hNSC-derived EVs by establishing a hNSC neurosphere culture system for EV collection and demonstrated their drug-delivery potential by loading the EVs with doxorubicin to evaluate brain uptake levels in vivo. To this end, we successfully developed a novel alkaline passive loading methodology that proved to be highly effective at Dox incorporation into EVs. Overall, the technology described in this study demonstrates novel hNSC suspension scale-up and drug loading technologies that enables delivery to the brain for treating CNS diseases.

Materials and methods

hNSC culture and EV collection

The hNSCs, derived from human pluripotent stem cells, were previously described and are capable of long term growth (> 6 months)27 and cultured in proliferation media consisting of 1X ANS™ Neural Supplement in AB2™ medium (Aruna Bio), supplemented with 2 mM L-glutamine (Gibco), 20 ng/ml bFGF (R&D Systems) and 10 ng/ml leukemia inhibitory factor (LIF) (Millipore). To form neurospheres, hNSCs were seeded at 1 × 106 cells/ml in proliferation media in 1L spinner flasks (Corning) on magnetic stirrer plates at 37° C in 5% CO2. hNSCs and neurospheres were passaged with trypsin–EDTA with cell counts and viability evaluation. Media was collected daily from the spinner flasks for EV isolation for 30 days. Approximately 100 L of media was processed for EV collection and concentrated by utilizing tangential flow filtration (TFF) and anion exchange chromatography. Briefly, pooled frozen media was thawed at 2 to 8 °C. The pooled media was buffer exchanged using a prepared TFF assembly (Repligen) into tangential flow and diafiltration buffer 2 (TFD2, Aruna Bio, Inc.). Following diafiltration, the EV-containing solution was further concentrated by TFF and treated with benzonase (Sigma Adrich) to remove extravesicular nucleic acid material following supplier recommended protocols (25–50 units/ml for 1 h at 37 °C). After benzonase treatment, the EV-containing solution was again concentrated by TFF and impurities were removed by anion exchange chromatography (Akta, Bia Separations). Fractions containing EVs were combined and a final buffer exchange was performed into Final Formulation Buffer (FFB1; Aruna Bio, Inc.), filtered through a 0.22 μm filter, and stored at -20 ± 5 °C. DNA measurements were performed on a Qubit 4 Fluorometer with the Qubit dsDNA HS Assay Kit (ThermoFisher) and Agilent Bioanalyzer. Benzonase levels were evaluated by ELISA with the NucA Nuclease DuoSet ELISA Kit (R&D Systems), according to manufacturer instructions, on a Molecular Devices SpectraMax iD5.

Nanoparticle tracking analysis, leprechaun analysis and nanoanalyzer analysis

EV size and concentration were determined using a Nanosight NS300 (Malvern Panalytical), according to manufacturer instructions. Briefly, samples were diluted to a range between 1 × 108 – 1 × 109 particles/ml, such that the particles/frame were in the range of 20–60. Five captures were taken over 60 s with an infusion rate of 25. From the merged data, the particle concentration (particles/ml) and mean and mode diameter (nm) were acquired.

CD63 levels following EV collection were determined using the Leprechaun (Unchained Labs) Exosome Human Tetraspanin Kit, according to manufacturer instructions. Briefly, 5 × 107 particles/ml were suspended in solution A and pipetted into the center of the Luni, sealed and incubated overnight. The following day, the Luni was washed in a CW-100 plate washer and fluorescent antibody solution (containing Anti-CD63) was prepared and incubated with sample for 1.5 h. Following additional washing steps, the chips were scanned on the Leprechaun.

The Flow NanoAnalyzer (NanoFCM) was used to determine the levels of Dox incorporation into EVs. Following Dox loading, EVs were diluted to 1 × 108/ml in PBS and analyzed on the NanoAnalyzer and compared to unloaded EVs. Excitation/emission for Dox was evaluated using the PE channel and triggered from the SS.

Cryogenic Transmission Electron Microscopy

Cryogenic TEM was performed by Alpha Nano Tech, LLC. The grids were glow discharged immediately prior to sample prep. The Vitrobot sample chamber was equilibrated at 96% humidity at 4 °C. 3 µL of the EV sample was loaded on the sample excess was blotted away using Whatman paper. A total of four grids were prepared per sample using 3- or 4-s blots. The grids were plunge-frozen in the liquid ethane/propane mixture (-186 °C). The samples were imaged using Talos Arctica 200 kV cryogenic electron microscope equipped Gatan K3 5,760 × 4,092-pixel direct electron detector.

EV-Dox loading and release

Doxourubicin (Dox) was loaded into hNSC-EVs by passive incubation under alkaline conditions. The large-scale loading by passive incubation was performed as follows: 6 × 1013 EVs were mixed with Dox to a final concentration 100 µM in 200 ml (3 × 1011 EVs/ml). The pH of the suspension was adjusted to 9.5 and was shaken for 1 h at room temperature and then subsequently neutralized with HCl to return the sample to a pH of 7.0. The EVs were next subjected to ultracentrifugation (100,000 × g) for 1 h. EV-Dox was washed in PBS and re-concentrated by ultracentrifugation (100,000 × g) for 1 h. The loading levels of Dox were determined by fluorescence quantification (ex/em: 470/595) on a plate reader (Molecular Devices) and by mass spectrometry by comparing EV-Dox levels to a Dox standard curve.

EV-Dox release assays were performed by performing dialysis using 20 K MWCO slide-a-lyzer MINI dialysis devices (Thermo Scientific). EV-Dox samples (100 µL) were dialyzed against 1L of PBS and 10 µL samples were collected at indicated time points. Dox levels were measured by fluorescence quantification on a plate reader, as described above.

Mass spectrometry

Chemicals and reagents: The stable deuterated isotope labeled internal standard (IS), paclitaxel-d5, was purchased from Cayman Chemical Co. (Ann Arbor, Michigan, USA). Paclixtaxel-d5 was not loaded into EVs and used only as an internal instrument standard as previously described37. Ammonium formate, formic acid, acetonitrile (ACN), methanol (MeOH) and water were all liquid chromatography-mass spectrometry (LC–MS) grade reagents from Sigma Aldrich Inc. (St. Louis, MO, USA).

LC–MS/MS conditions: The LC–MS/MS analysis was performed with an ACQUITY ultra-performance liquid chromatography (UPLC) H-Class PLUS system (Waters, Milford, MA, USA) interfaced to a Waters Xevo Micro TQS mass spectrometer with an electrospray ionization (ESI) source (Waters, Milford, MA, USA). Waters Masslynx 4.2 software (Milford, MA, USA) was used for instrumentation and quantitative analysis. The separation was performed on a BDS Hypersil C8 column (50 × 2.1 mm, 5 μm; Thermo Scientific, West Palm Beach, FL, USA) coupled with a SecurityGuard C8 guard column (4 × 3.0 mm; Phenomenex, Torrance, CA, USA).

Mobile phase A was 10 mM ammonium formate aqueous buffer with 0.1% formic acid and mobile phase B was MeOH. To separate the analytes, an isocratic elution with a washing gradient was used (time/minute, % mobile phase B): (0.0, 30), (1.5, 30), (4.0, 95), (5.0, 95), (5.1, 30), (6.0, 30). The flow rate was 0.8 mL/min, the column temperature was 40 °C, and the autosampler temperature was 20 °C. The injection volume was 10 μL and the injection needle was washed with 70% ACN/water (v/v) after each injection. The mass spectrometer was operated in positive electrospray ionization (ESI +) mode. Nitrogen was used as the desolvation gas at a flow rate of 800 L/h and a temperature of 120 °C. The cone gas (N2) flow rate was 50 L/h and the source temperature was 120 °C. The capillary voltage and the cone voltage were set at 0.5 kV and 50 V, respectively. Argon was used as the collision gas at a collision cell pressure of 3.5 × 10–3 mbar. Multiple reaction monitoring (MRM) functions were used to detect and quantify the analyte and IS. The collision energy for doxorubicin and paclitaxel-d5 were at 12 eV and 27 eV, respectively. With the dwell time at 0.054 s, the transitions m/z 544.1 → 397.0 and 881.3 → 313.0 of the [M + Na] + ions were monitored for doxorubicin and paclitaxel-d5, respectively, while the transitions of m/z 544.1 → 130.0 and 881.3 → 591.1 were used for confirmation.

Preparation of stock, standard solutions and samples: The stock solution of paclitaxel-d5 was prepared by dissolving 1.0 mg of solid in 1.0 mL of ACN to yield a concentration of 1.0 mg/mL. Calibration working solutions of doxorubicin were also prepared for the doxorubicin-EV stocks at concentrations of 1, 2, 4, 8, 10, 12, 16, 20, 24, and 30 μg/mL by dilutions from the stock solution. The IS working solution was 1.0 μg/mL paclitaxel-d5 in ACN. Stock solutions were kept at -20 °C when not in use.

A volume of 10 μL of the calibration working solution was spiked into 190 μL of brain homogenate or PBS to generate the corresponding calibration standards. The final concentrations of the calibration standards were 0.1, 0.2, 0.4, 1.0, 3.0, 5.0, 8.0, 10.0, 12.0, and 15.0 ng/mL in brain homogenate. The final concentrations of the calibration standards were 0.05, 0.1, 0.2, 0.4, 0.5, 0.6, 0.8, 1.0, 1.2, and 1.5 μg/mL in PBS. Calibration samples were then processed as described below. All biological samples were stored at -80 °C when not in use.

A protein precipitation (PPT) method was used for sample preparation. To each 100 μL of biological sample or calibration standard, 1 mL of chilled ACN containing the IS (4.3 ng/mL) were added. The mixture was vortexed with a VX-2500 Multi Tube Vortexer (VWR, Radnor, PA, USA) for 10 min and then centrifuged with an Eppendorf Centrifuge 5427 R (Hamburg, Germany) at 25,001 g and 18 °C for 12 min to extract doxorubicin and the IS. To a MS vial, 90 μL of the supernatant and 210 μL water were added to resemble the initial mobile phase composition (30% MeOH/water (v/v)). The samples were vortexed for 10 min before LC–MS/MS analysis.

Caspase and MTS analysis

CT-2A mouse glioma cells were seeded at 20,000 cells/well in a 96-well plate for caspase and MTS assays. After 24 h, cells were treated with EV-Dox or control conditions for 24 h and assayed for cytoxicity (MTS assay, Abcam) or caspase 3/7 activation using the Caspase-Glo 3/7 Assay kit (Promega), according to manufacturer guidelines.

Animal studies

C57BL/6 mice were intravenously injected by tail-vein with hNSC-EV-Dox or Dox only at 1 mg/kg. Prior to injections mice were weighed and randomly separated into two groups, average weight of EV-Dox mice and Dox-only mice were 26.53 + /- 0.33 g and 25.78 + /- 0.95 g, respectively. All animals were euthanized 1 h after dox or dox-EV treatment by carbon dioxide inhalation followed by transcardial perfusion with ice-cold phosphate buffer saline. Brains were snap frozen in liquid nitrogen and stored at -80 °C. The left hemisphere was resuspended in molecular grade water and homogenized using a beadmill homogenizer (25 Hz for 45 s per cycle for 10 cycles). Homogenates were re-frozen and later processed for mass spectrometry analysis. All animal protocols and experiments were fully approved and conducted in full accordance with the Institutional Animal Care and Use Committee (IACUC) guidelines at the University of Georgia. All methods are reported in accordance with the ARRIVE guidelines (https://arriveguidelines.org).

Statistical analyses

All statistical analyses were performed using Graphpad Prism using either a Student T-test or one-way ANOVA with p-values, as indicated. Statistical analyses were performed on independent replicates.

Results

Suspension culture of hNSCs results in increased EV production and purity

To establish a scalable platform amenable to large bioreactors for EV production from adherent hNSCs, the hNSCs were grown in proliferation media in suspension culture in spinner flasks. The hNSCs were able to be grown for more than 30 days without any significant loss in cell viability as determined by sphere morphology (Fig. 1A). Interestingly, the transition from adherent to suspension culture resulted in > tenfold increase in the EV output (Fig. 1B), as also indicated for other cell types38. The EV mean diameter, as determined by nanoparticle tracking analysis (NTA), peaked at 140 nM, consistent with our analysis from adherent cultures (Fig. 1C)19,30–33. To evaluate day-to-day levels in EV production from the suspension culture, CD63 Tetraspanin levels were evaluated on the Leprachaun over time (Fig. 1D). CD63 levels were largely stable throughout the time course, indicating consistent EV output from the spheres. To further characterize the hNSC-EVs, transmission electron microscopy was performed (Fig. 1E, Supplementary Fig. 1). The EVs produced exhibited the typical EV morphology consisting of a lipid membrane surrounding a lumen. Multiple purification steps were employed during the EV preparation to improve concentration and purity, including TFF, nuclease treatment and anion exchange chromatography. The removal of contaminating DNA by nuclease treatment was verified on an a bioanalyzer, while with the subsequent removal of the nuclease was confirmed by ELISA (Supplementary Fig. 2). Overall, these data demonstrate the hNSCs can be grown in suspension culture for scale-up requirements with improved concentration facilitating their use as a therapeutic.

Fig. 1.

Fig. 1

Open in a new tab

Scale-up culture of hNSC-derived neurospheres and EV production characterization. (A) Representative images of hNSCs following one week of neurosphere development, bar: 500 µm. (B) Quantification of hNSC-EV production from 100 L of biologically pooled adherent or suspension cultures using Nanoparticle Tracking Analysis. (C) Size analysis from pooled 100 L cultures by NTA of the hNSC-EVs. (D) CD63 exhibited consistent levels from 3 L daily pooled suspension hNSC-EVs over time. (E) Cryo-TEM analysis of hNSC-EVs, bar: 20 nm.

Loading and quantification of doxorubicin into hNSC-EVs

As a scalable system for producing EVs from hNSCs was successfully established and characterized (Figs. 1), the potential function of the EVs as a drug delivery vehicle was evaluated. As a proof of concept, doxorubicin (Dox) was chosen as the drug of choice to evaluate for hNSC-EV drug loading and delivery due to several attributes. First, Dox has been demonstrated both as a free-drug or following encapsulation into liposomes (Doxil or Myocet) to have little to no ability to penetrate the blood–brain barrier39. Second, Dox is fluorescent and therefore enables easy visualization for both drug loading and cell uptake. Third, Dox has potential to serve as a therapeutic for brain cancers if sufficient brain delivery can be achieved40–42. As such, clinical trials are currently ongoing that evaluate methods of focused ultrasound to disrupt the BBB with Dox treatment (clinical trial: NCT05615623).

To load Dox into the hNSC-EVs, passive incubation was performed for 1 h at room temperature under alkaline conditions at a pH of 9.5. This pH enables free Dox to be deprotonated, as the isolectric point is 8.4, and was hypothesized to enable Dox to freely pass across EV membranes. After the incubation, the EV-Dox sample was neutralized and returned to a physiological pH, promoting Dox entrapment into EVs. To accurately measure Dox in the EVs and then later in brain tissue, a mass spectrometry protocol was developed using previously established protocols43. Following the analysis by mass spectrometry, a single peak was observed pertaining to Dox in the EVs (Fig. 2A). Using a standard curve of free Dox, the number of molecules per EV was calculated to be approximately 20,000 with EV concentration determined by NTA. (Fig. 2B and 2C). No molecules were detected in unloaded EVs and no significant differences were observed in loaded or non-loaded EVs (Fig. 1C and Fig. 2C). To further evaluate the Dox loading levels of hNSC-EVs, nano-flow cytometry was performed (Fig. 2D). EV sizes from nano flow cytometry were equivalent with or without loading (Fig. 2E). More than 90% of the hNSC-EV were successfully loaded with Dox, indicating the high efficacy of alkaline-based passive loading. Next, to evaluate how quickly Dox is released from the EVs, dialysis of EV-Dox was performed, and Dox levels were measured by fluorescence over time (Fig. 2F). Approximately 35% of Dox was released in 1 h and 70% of the Dox was released from the EVs following 24 h. These data demonstrate that Dox can be efficiently loaded into hNSC-EVs and suggest that it will be released upon uptake into cells.

Fig. 2.

Fig. 2

Open in a new tab

Loading and drug-release evaluation of hNSC-EV-Dox. (A). Mass spectrometry analysis of hNSC-EV-Dox indicating main peak for Dox presence. (B) Concentration levels expressed as Dox molecules per EV based on mass spec analysis compared to a standard curve, mean + /- standard deviation from 3 biologically independent replicates, analyzed by Student’s t-test. (C) A representative NTA analysis following Dox loading. (D). Single particle analysis of hNSC-EV samples indicating loading levels of Dox. (E) Size analysis based on nanoanalyzer. (F) Dox-release assay performed by dialysis and sampling after 1, 6 and 24 h, with concentrations determined by fluorescence measurements (ex/em: 470 nm/595 nm) compared to starting concentrations. Data represented as mean + /- standard deviation from 3 biologically independent replicates, using one-way ANOVA with Tukey’s post-hoc analysis. *P < 0.05, **P < 0.01, ***P < 0.001.

Dox-loaded hNSC-EVs promote cell death in glioma cells

To evaluate if EV-Dox will promote cytotoxicity in glioma cells, CT-2A cells were treated with Dox-loaded EV and compared to EVs alone (Fig. 3). After 24 h of treatment, Dox uptake could be visualized by light microscopy in the cells (Fig. 3A), along with a clear reduction in cell confluency. To evaluate the cytotoxic effects of hNSC-EV-Dox in mouse glioma cells, MTS assays were performed with increasing concentrations of hNSC-EVs (Fig. 3B). A dose-dependent reduction in cell viability and proliferation was observed from the hNSC-EV-Dox treatment (Fig. 3B), which was equivalent to the Dox-only control. Next, apoptotic induction was measured using caspase 3/7 activity assays following the treatment of hNSC-EVs or hNSC-EV-Dox samples for 24 h (Fig. 3C). A ~ twofold increase in caspase activity was observed from the hNSC-EV-Dox treatment, compared to the hNSC-EV alone, indicating that the hNSC-EV-Dox samples activated apoptotic cascade signaling within the cells with comparable efficacy to a Dox-only control. Altogether, these data demonstrate that Dox-loaded hNSC-EVs are readily taken up by glioma cells, and this subsequently results in their loss of viability and promotion of apoptotic pathways.

Fig. 3.

Fig. 3

Open in a new tab

hNSC-EV-Dox promotes cell death in glioma cells. (A) Phase-contrast images of CT-2A mouse glioma cells treated with equal concentrations of hNSC-EVs or hNSC-EV-Dox for 24 h, bar: 200 µm. (B) Cytotoxicity levels performed by MTS assay in CT-2A cells after 48 h of treatment with hNSC-EV, hNSC-EV-Dox, or Dox-only at indicated concentrations, mean + /- standard deviation, n = 3 per group, ***p < 0.001, using one-way ANOVA with Tukey’s post-hoc analysis, comparing hNSC-EV to hNSC-EV-Dox at dose-matched concentrations. (C) Caspase 3/7 activity assay following 24 h treatment of CT-2A cells with hNSC-EV, hNSC-EV-Dox, or Dox-only at 3.3 µM, mean + /- standard deviation n = 3 per group, ***p < 0.001 one-way ANOVA with Tukey’s post-hoc analysis.

Dox-loaded hNSC-EVs are transported to the brain

Chemotherapeutic drugs, such as Dox, often fail to penetrate the BBB and therefore have limited utility for treating brain cancers like glioblastoma39. The hNSC-EVs were previously demonstrated to penetrate the BBB and have therapeutic efficacy for treating stroke and traumatic brain injury19,30–33. As the hNSC-EVs were successfully loaded with Dox and had efficacy at inhibiting glioma cell proliferation and promoting apoptosis in vitro, we next sought to determine if hNSC-EVs could efficiently deliver Dox through the BBB to the brain. Our previous studies demonstrated that hNSC-EVs can be detected in the rodent brain at 1 h, but not 24 h, post injection by single photon emission computed tomography, so we evaluated Dox levels in the brain at 1 h19. Mice were injected intravenously with hNSC-EV-Dox or Dox-only at equivalent concentrations as determined by mass spectrometry. At 1 h post EV injection, the mice were sacrificed, and brain tissues were perfused to remove any vascular EV-Dox biodistribution and homogenized for mass spectrometry analysis of Dox to determine brain tissue drug concentrations (Fig. 4). Importantly, mice treated with hNSC-EV-Dox had an approximate 2.5-fold increase over mice treated with Dox alone. Dox was not detectable in the brain at 24 h or later (data not shown). These data indicate that the hNSC-EVs deliver Dox through the BBB and may serve as a useful therapeutic for brain cancers. Moreover, these studies demonstrate the utility of hNSC-EVs for drug delivery in CNS diseases.

Fig. 4.

Fig. 4

Open in a new tab

hNSC-EV-Dox have increased brain uptake compared to Dox alone. (A) Method for evaluating brain uptake by hNSC-EV-Dox. Following hNSC-EV-Dox loading, mice were intravenously (I.V.) injected with 1 mg/kg of Dox alone or hNSC-EV-dox for 1 h and brains were isolated and homogenized for mass spectrometry, n = 3 mice per group. (B) Quantification of Dox levels by LC–MS/MS in mouse brain tissue of Dox alone or hNSC-EV-Dox. Data represented as mean + /- standard deviation and the percent injected dose/gram (%ID/g) is indicated above the graphs. **p < 0.01 using Student’s t-test.

Discussion

Drug delivery across the BBB for CNS diseases remains a significant challenge in drug development2. In this study, hNSC-EVs were evaluated as a drug-delivery vehicle for drugs that have little-to-no BBB penetration. As a proof of concept, Dox was chosen as the drug of choice to evaluate due to its low BBB penetrance39, previous studies indicating EV-Dox loading ability44–49, and the potential utility of doxorubicin for brain cancer treatment if BBB penetrance is overcome40–42.

A major hurdle in the EV field that hampers their therapeutic use is scale-up requirements. To overcome this issue, a suspension culture system using neurospheres from the hNSCs was developed. This system provided several benefits over the traditional adherent culture system, which included (1) cost-reduction due to the removal of basement membrane protein (i.e. laminin) vessel coating, (2) increased yield based on total particles collected. This suspension system enhances the use of native hNSC-EVs for therapeutic treatment in stroke and TBI19,31–33, and for further development in drug delivery of small molecules, RNAs or proteins30.

Scale-up approaches for other stem cell types, such as MSC or induced pluripotent stem cells (iPSC) that make use of bioreactors have been previously described34–36. While EV production from iPSCs grown in bioreactors has not been described, EVs produced from MSCs grown in bioreactors with the aid of microcarriers has been achieved50. Neurosphere culture from hNSC for the purpose of generating EVs has also been described51–53. Here we expanded upon these findings to show that scale-up culture of hNSC-derived neurospheres can facilitate EV-based drug delivery for CNS diseases.

Following the establishment of a scalable system for hNSC-EV production, Dox was loaded into hNSC-EVs at a large scale (~ 6.0 × 1013 hNSC-EVs). While previous studies have utilized sonication or electroporation for Dox-loading of EVs and demonstration of therapeutic efficacy for GBM in vivo45–48, these loading approaches are often not scalable for therapeutic applications beyond proof of concept and could potentially damage EV integrity. As such, we focused on optimizing co-incubation under alkaline conditions as a passive loading method, which was found to be equivalent or exceed the membrane-based disruption approaches, as previously described44. This approach is akin to liposome loading via pH gradients, where the Dox is entrapped within the liposomes by heavy metals54,55. It should be noted that Dox is also known to bind plasma proteins at 74–76%, independent of the plasma concentration of doxorubicin up to 1.1 µg/mL due to its general physiochemical properties56,57, and further studies will be needed to understand how Dox becomes entrapped within EV. By single EV assessment, we found that Dox-loading levels could be achieved above 90%, with approximately 20,000 molecules/EV, which demonstrates the utility of the alkaline loading approach. Importantly, these levels provide a clinically-feasible dose concentration for the treatment of brain cancer. To our knowledge, this is the first example of using a mild alkaline pH to load doxorubicin into EVs. The 1-h alkaline treatment had no discernible effects based on two independent methods (NTA and nanoflow cytometry). While these methods differ in how they determine particle size (Brownian motion and light scattering for NTA and high-resolution imaging for nanoflow) and result in different overall size estimates58,59, the untreated and Dox-loaded EVs showed similar size profiles, suggesting there was little-to-no impact on membrane integrity. However, previous studies have utilized stronger alkaline conditions (pH > 11), in combination with sonication, on cells to disrupt cell membranes and generate EV-mimetics for the purpose of loading drugs, such as dexamethasone60. The approach described is considerably different, as it is used with native EVs, not on cells to generate EV-mimetics, and does not rely on disrupting membranes.

We evaluated Dox release from EVs by dialysis and found that after 1 h, > 60% was still retained within the EVs. This suggests that some Dox may be retained within the EVs as it is transported through the BBB to target cells of interest. We next confirmed the in vitro efficacy of hNSC-EV-Dox at promoting cell death using a glioma cell line. Importantly, hNSC-EV-Dox showed identical efficacy to Dox-alone, indicating that binding to EVs does not compromise its activity. Lastly, we demonstrated that hNSC-EV-Dox enhances the BBB penetrance of Dox into the brain when injected intravenously when measured at 1 h, but not 24 h. Additional studies with time points at 6 and 12 h may provide additional durability data. An overall 2.5-fold increase in brain delivery was observed at 1 h with EV-Dox versus Dox-alone, but the overall levels in the brain were still low. This could potentially be overcome by evaluating additional routes of administration. Recent studies have indicated that intranasal drug delivery may provide a more effective approach for brain targeting than intravenous delivery61–63. However, challenges remain with this approach, such as the nasal epithelial barrier and lower volume dose constraints. Altogether, these findings establish hNSC-EV-Dox as a potential treatment for brain cancers, such as glioblastoma.

As a therapeutic, Dox suffers from several drawbacks beyond BBB penetrance, such as off-target toxicity. The formulation of Dox into pro-drug conjugates, such as liposomes or other modifications, have been suggested to reduce these drawbacks and improve pharmacokinetics, reduce Dox resistance in tumors and provide opportunities for oral route of administrations64,65. Interestingly, Dox-loaded EVs have also been shown to have reduced cardiotoxicity compared to Dox alone47. The loading of pro-drug formulations of Dox into EVs may therefore provide additional benefits, such as reducing tumor resistance, for cancer treatments.

The development of scalable, Dox-loaded hNSC-derived EVs demonstrates an enabling technology for hNSC-derived EV drug delivery. The use of hNSCs as an EV producer cell could potentially be expanded for the delivery of other cargo, including proteins or siRNA through endogenous genetic engineering approaches as we have demonstrated previously66 or by exogenous loading into EVs directly15. Additional studies are warranted to expand the drug delivery utility of hNSC-EVs.

Conclusion

The present study focused on developing a scalable neurosphere suspension system for producing hNSC-EVs and demonstrating the capability for delivery of the CNS impermeable drug, doxorubicin, to the brain. This proof-of-concept study establishes hNSC-EVs as a drug-delivery vehicle, which can be broadly applied to most CNS diseases. Furthermore, the novel alkaline passive EV-loading method described here may have broad utility for basic compounds to facilitate entrapment for drug delivery purposes. Additional studies will expand the drug-delivery capabilities of hNSC-EVs for other BBB impermeable chemotherapeutics such as Paclitaxel and Gleevec. Future evaluation of hNSC-EV loaded chemotherapeutics in GBM animal models will enable these drugs for future clinical development.

Supplementary Information

Supplementary Information. (370.9KB, docx)

Acknowledgements

The authors wish to thank Julie Nelson for providing assistance at the UGA flow cytometry core facility and Wided Najahi-Missaoui for consulting with Dox-loading strategies. We also thank Ravi Koripella and Ricardo Guerrero at the Emory electron microscopy core facility. Aruna Bio was supported by NIH grant 1R41NS122704, awarded to S.L.S.

Author contributions

Conceptualization, writing, editing: A.M.S., R.S., S.L.S. Experimentation and Data analysis: A.M.S., C.M.W., A.P., L.P.C., R.M., M.C.F., M.P.V., W.A. Project management, supervision and oversight: A.M.S., M.G.B., V.M., R.S. and S.L.S.

Data availability

The datasets used and/or analysed during the current study available from the corresponding authors on reasonable request.

Declarations

Competing interests

A.M.S, C.M.W., A.P., L.P.C., R.M., M.P.V, W.A., and V.M., R.S. and S.L.S are former or current employees of Aruna Bio, Inc.

Footnotes

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Amar M. Singh, Email: singha@uga.edu

Steven L. Stice, Email: sstice@arunabio.com

References

  • 1.Correale, J. & Villa, A. Cellular elements of the blood-brain barrier. Neurochem Res34(12), 2067–2077. 10.1007/s11064-009-0081-y (2009). [DOI] [PubMed] [Google Scholar]
  • 2.Terstappen, G. C. et al. Strategies for delivering therapeutics across the blood-brain barrier. Nat Rev Drug Discov20(5), 362–383. 10.1038/s41573-021-00139-y (2021). [DOI] [PubMed] [Google Scholar]
  • 3.Jimenez Macias, J. L. et al. Decoding the biology of the blood-brain tumor barrier in brain cancer. Mol Cancer Res 10.1158/1541-7786.MCR-25-0908 (2026). [DOI] [PubMed] [Google Scholar]
  • 4.Steeg, P. S. The blood-tumour barrier in cancer biology and therapy. Nat Rev Clin Oncol18(11), 696–714. 10.1038/s41571-021-00529-6 (2021). [DOI] [PubMed] [Google Scholar]
  • 5.Pulgar, V. M. Transcytosis to Cross the Blood Brain Barrier. New Adv Challenges. Front Neurosci12, 1019. 10.3389/fnins.2018.01019 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Johnsen, K. B. et al. Modulating the antibody density changes the uptake and transport at the blood-brain barrier of both transferrin receptor-targeted gold nanoparticles and liposomal cargo. J Control Release295, 237–249. 10.1016/j.jconrel.2019.01.005 (2019). [DOI] [PubMed] [Google Scholar]
  • 7.Roberts, R. L., Fine, R. E. & Sandra, A. Receptor-mediated endocytosis of transferrin at the blood-brain barrier. J Cell Sci104(Pt 2), 521–532. 10.1242/jcs.104.2.521 (1993). [DOI] [PubMed] [Google Scholar]
  • 8.Sheff, D. R. et al. The receptor recycling pathway contains two distinct populations of early endosomes with different sorting functions. J Cell Biol145(1), 123–139. 10.1083/jcb.145.1.123 (1999). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Yu, Y. J. et al. Boosting brain uptake of a therapeutic antibody by reducing its affinity for a transcytosis target. Sci Transl Med3(84), 84ra44. 10.1126/scitranslmed.3002230 (2011). [DOI] [PubMed] [Google Scholar]
  • 10.Pinheiro, R. G. R. et al. Nanoparticles for targeted brain drug delivery: What do we know?. Int J Mol Sci 10.3390/ijms222111654 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Witwer, K. W. & Wolfram, J. Extracellular vesicles versus synthetic nanoparticles for drug delivery. Nat Rev Mater6(2), 103–106. 10.1038/s41578-020-00277-6 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Banks, W. A. et al. Transport of extracellular vesicles across the blood-brain barrier: Brain pharmacokinetics and effects of inflammation. Int J Mol Sci 10.3390/ijms21124407 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Ramos-Zaldivar, H. M. et al. Extracellular vesicles through the blood-brain barrier: a review. Fluids Barriers CNS19(1), 60. 10.1186/s12987-022-00359-3 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.de Jong, O. G. et al. Drug Delivery with Extracellular Vesicles: From Imagination to Innovation. Acc Chem Res52(7), 1761–1770. 10.1021/acs.accounts.9b00109 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Herrmann, I. K., Wood, M. J. A. & Fuhrmann, G. Extracellular vesicles as a next-generation drug delivery platform. Nat Nanotechnol16(7), 748–759. 10.1038/s41565-021-00931-2 (2021). [DOI] [PubMed] [Google Scholar]
  • 16.Edelmann, M.J. and P.E. Kima, Current understanding of extracellular vesicle homing/tropism. Zoonoses (Burlingt), 2022. 2. 10.15212/zoonoses-2022-0004. [DOI] [PMC free article] [PubMed]
  • 17.Wiklander, O. P. et al. Extracellular vesicle in vivo biodistribution is determined by cell source, route of administration and targeting. J Extracell Vesicles4, 26316. 10.3402/jev.v4.26316 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Hoshino, A. et al. Tumour exosome integrins determine organotropic metastasis. Nature527(7578), 329–335. 10.1038/nature15756 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Webb, R. L. et al. Human Neural Stem Cell Extracellular Vesicles Improve Tissue and Functional Recovery in the Murine Thromboembolic Stroke Model. Transl Stroke Res9(5), 530–539. 10.1007/s12975-017-0599-2 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Zhao, S. et al. Astrocyte-derived extracellular vesicles: A double-edged sword in central nervous system disorders. Neurosci Biobehav Rev125, 148–159. 10.1016/j.neubiorev.2021.02.027 (2021). [DOI] [PubMed] [Google Scholar]
  • 21.Haney, M.J., et al., Extracellular Vesicles as Drug Delivery System for Treatment of Neurodegenerative Disorders: Optimization of the Cell Source. Adv Nanobiomed Res, 2021. 1(12). 10.1002/anbr.202100064. [DOI] [PMC free article] [PubMed]
  • 22.Lendemeijer, B. et al. Human Pluripotent Stem Cell-Derived Astrocyte Functionality Compares Favorably with Primary Rat Astrocytes. Eneuro11(9), 1–21. 10.1523/Eneuro.0148-24.2024 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Zhang, Y. et al. Purification and Characterization of Progenitor and Mature Human Astrocytes Reveals Transcriptional and Functional Differences with Mouse. Neuron89(1), 37–53. 10.1016/j.neuron.2015.11.013 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Dhara, S. K. et al. Human neural progenitor cells derived from embryonic stem cells in feeder-free cultures. Differentiation76(5), 454–464. 10.1111/j.1432-0436.2007.00256.x (2008). [DOI] [PubMed] [Google Scholar]
  • 25.Dhara, S. K. & Stice, S. L. Neural differentiation of human embryonic stem cells. J Cell Biochem105(3), 633–640. 10.1002/jcb.21891 (2008). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Shin, S., Dalton, S. & Stice, S. L. Human motor neuron differentiation from human embryonic stem cells. Stem Cells Dev14(3), 266–269. 10.1089/scd.2005.14.266 (2005). [DOI] [PubMed] [Google Scholar]
  • 27.Shin, S. et al. Long-term proliferation of human embryonic stem cell-derived neuroepithelial cells using defined adherent culture conditions. Stem Cells24(1), 125–138. 10.1634/stemcells.2004-0150 (2006). [DOI] [PubMed] [Google Scholar]
  • 28.Shin, S. et al. Whole genome analysis of human neural stem cells derived from embryonic stem cells and stem and progenitor cells isolated from fetal tissue. Stem Cells25(5), 1298–1306. 10.1634/stemcells.2006-0660 (2007). [DOI] [PubMed] [Google Scholar]
  • 29.Wilson, P. G. & Stice, S. S. Development and differentiation of neural rosettes derived from human embryonic stem cells. Stem Cell Rev2(1), 67–77. 10.1007/s12015-006-0011-1 (2006). [DOI] [PubMed] [Google Scholar]
  • 30.Jurgielewicz, B., Stice, S. & Yao, Y. Therapeutic Potential of Nucleic Acids when Combined with Extracellular Vesicles. Aging Dis12(6), 1476–1493. 10.14336/AD.2021.0708 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Spellicy, S. E. et al. Neural Stem Cell Extracellular Vesicles Disrupt Midline Shift Predictive Outcomes in Porcine Ischemic Stroke Model. Transl Stroke Res11(4), 776–788. 10.1007/s12975-019-00753-4 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Sun, M. K. et al. Extracellular Vesicles Mediate Neuroprotection and Functional Recovery after Traumatic Brain Injury. J Neurotrauma37(11), 1358–1369. 10.1089/neu.2019.6443 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Webb, R. L. et al. Human Neural Stem Cell Extracellular Vesicles Improve Recovery in a Porcine Model of Ischemic Stroke. Stroke49(5), 1248–1256. 10.1161/STROKEAHA.117.020353 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.King, J. A. & Miller, W. M. Bioreactor development for stem cell expansion and controlled differentiation. Curr Opin Chem Biol11(4), 394–398. 10.1016/j.cbpa.2007.05.034 (2007). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Rohani, L. et al. Stirred suspension bioreactors maintain naive pluripotency of human pluripotent stem cells. Commun Biol3(1), 492. 10.1038/s42003-020-01218-3 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Stephenson, M. & Grayson, W. Recent advances in bioreactors for cell-based therapies. F1000Res 10.12688/f1000research.12533.1 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Finnerty, M. C. et al. Identification of blood lipid markers of docetaxel treatment in prostate cancer patients. Sci Rep14(1), 22069. 10.1038/s41598-024-73074-8 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Wu, J. et al. Scale-out production of extracellular vesicles derived from natural killer cells via mechanical stimulation in a seesaw-motion bioreactor for cancer therapy. Biofabrication 10.1088/1758-5090/ac7eeb (2022). [DOI] [PubMed] [Google Scholar]
  • 39.Luo, R. et al. Distinct biodistribution of doxorubicin and the altered dispositions mediated by different liposomal formulations. Int J Pharm519(1–2), 1–10. 10.1016/j.ijpharm.2017.01.002 (2017). [DOI] [PubMed] [Google Scholar]
  • 40.Aryal, M. et al. Enhancement in blood-tumor barrier permeability and delivery of liposomal doxorubicin using focused ultrasound and microbubbles: evaluation during tumor progression in a rat glioma model. Phys Med Biol60(6), 2511–2527. 10.1088/0031-9155/60/6/2511 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Kovacs, Z. et al. Prolonged survival upon ultrasound-enhanced doxorubicin delivery in two syngenic glioblastoma mouse models. J Control Release187, 74–82. 10.1016/j.jconrel.2014.05.033 (2014). [DOI] [PubMed] [Google Scholar]
  • 42.Yang, F. Y. Treating glioblastoma multiforme with selective high-dose liposomal doxorubicin chemotherapy induced by repeated focused ultrasound. Int J Nanomedicine7, 965–974. 10.2147/IJN.S29229 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Li, P. et al. A rapid analytical method for the quantification of paclitaxel in rat plasma and brain tissue by high-performance liquid chromatography and tandem mass spectrometry. Rapid Commun Mass Spectrom27(19), 2127–2134. 10.1002/rcm.6671 (2013). [DOI] [PubMed] [Google Scholar]
  • 44.Chen, C. et al. Single-particle assessment of six different drug-loading strategies for incorporating doxorubicin into small extracellular vesicles. Anal Bioanal Chem415(7), 1287–1298. 10.1007/s00216-022-04248-4 (2023). [DOI] [PubMed] [Google Scholar]
  • 45.Haney, M. J. et al. Macrophage-Derived Extracellular Vesicles as Drug Delivery Systems for Triple Negative Breast Cancer (TNBC) Therapy. J Neuroimmune Pharmacol15(3), 487–500. 10.1007/s11481-019-09884-9 (2020). [DOI] [PubMed] [Google Scholar]
  • 46.Schindler, C. et al. Exosomal delivery of doxorubicin enables rapid cell entry and enhanced in vitro potency. PLoS ONE14(3), e0214545. 10.1371/journal.pone.0214545 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Wu, J. Y. et al. Exosomes and biomimetic nanovesicles-mediated anti-glioblastoma therapy: A head-to-head comparison. J Control Release336, 510–521. 10.1016/j.jconrel.2021.07.004 (2021). [DOI] [PubMed] [Google Scholar]
  • 48.Wu, Y. W. et al. Platelet extracellular vesicles are efficient delivery vehicles of doxorubicin, an anti-cancer drug: Preparation and in vitro characterization. Platelets34(1), 2237134. 10.1080/09537104.2023.2237134 (2023). [DOI] [PubMed] [Google Scholar]
  • 49.Zhou, Y. et al. Glioblastoma cell-derived exosomes functionalized with peptides as efficient nanocarriers for synergistic chemotherapy of glioblastoma with improved biosafety. Nano Res.16(12), 13283–13293. 10.1007/s12274-023-5921-6 (2023). [DOI] [Google Scholar]
  • 50.de Almeida Fuzeta, M. et al. Scalable production of human mesenchymal stromal cell-derived extracellular vesicles under serum-/xeno-free conditions in a microcarrier-based bioreactor culture system. Front Cell Dev Biol8, 553444. 10.3389/fcell.2020.553444 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Abedi, M. et al. Exosome derived from human neural stem cells improves motor activity and neurogenesis in a traumatic brain injury model. BioMed Res. Int.2022, 6409346. 10.1155/2022/6409346 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Branscome, H. et al. Retroviral infection of human neurospheres and use of stem cell EVs to repair cellular damage. Sci. Rep.12(1), 2019. 10.1038/s41598-022-05848-x (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Koopaei, N. N. et al. Method for Isolating Extracellular Vesicles from Human Neural Stem Cells Expanded Under Neurosphere Culture. Methods Mol Biol2389, 87–94. 10.1007/978-1-0716-1783-0_8 (2022). [DOI] [PubMed] [Google Scholar]
  • 54.Abraham, S. A. et al. Formation of transition metal-doxorubicin complexes inside liposomes. Biochim. Biophys. Acta1565(1), 41–54. 10.1016/s0005-2736(02)00507-2 (2002). [DOI] [PubMed] [Google Scholar]
  • 55.Fritze, A. et al. Remote loading of doxorubicin into liposomes driven by a transmembrane phosphate gradient. Biochim. Biophys. Acta1758(10), 1633–1640. 10.1016/j.bbamem.2006.05.028 (2006). [DOI] [PubMed] [Google Scholar]
  • 56.DrugBank. Doxorubicin. [cited 2026; Available from: https://go.drugbank.com/drugs/DB00997.
  • 57.Pubchem, N.L.o.M., National Institutes of Health. Compound: Doxorubicin. [cited 2026; Available from: https://pubchem.ncbi.nlm.nih.gov/compound/Doxorubicin.
  • 58.Mladenovic, D. et al. Quantitative fluorescent nanoparticle tracking analysis and nano-flow cytometry enable advanced characterization of single extracellular vesicles. J. Extracellular Biol.4(1), e70031. 10.1002/jex2.70031 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Xu, S. et al. Comparison of nanoimaging and nanoflow based detection of extracellular vesicles at a single particle resolution. J. Extracellular Biol.3(10), e70016. 10.1002/jex2.70016 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Go, G. et al. Extracellular vesicle-mimetic ghost nanovesicles for delivering anti-inflammatory drugs to mitigate gram-negative bacterial outer membrane vesicle-induced systemic inflammatory response syndrome. Adv Healthcare Mater8(4), e1801082. 10.1002/adhm.201801082 (2019). [DOI] [PubMed] [Google Scholar]
  • 61.Chen, Y. et al. Intranasal drug delivery: The interaction between nanoparticles and the nose-to-brain pathway. Adv Drug Deliv Rev207, 115196. 10.1016/j.addr.2024.115196 (2024). [DOI] [PubMed] [Google Scholar]
  • 62.Crowe, T. P. & Hsu, W. H. Evaluation of recent intranasal drug delivery systems to the central nervous system. Pharmaceutics 10.3390/pharmaceutics14030629 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63.Zhao, J. et al. Biomedical applications of artificial exosomes for intranasal drug delivery. Front Bioeng Biotechnol11, 1271489. 10.3389/fbioe.2023.1271489 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64.Lee, J., Choi, M. K. & Song, I. S. Recent advances in doxorubicin formulation to enhance pharmacokinetics and tumor targeting. Pharmaceuticals (Basel) 10.3390/ph16060802 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 65.Zhao, H. et al. Doxorubicin prodrug-based nanomedicines for the treatment of cancer. Eur J Med Chem258, 115612. 10.1016/j.ejmech.2023.115612 (2023). [DOI] [PubMed] [Google Scholar]
  • 66.Dhara, S. K. et al. Genetic manipulation of neural progenitors derived from human embryonic stem cells. Tissue Eng Part A15(11), 3621–3634. 10.1089/ten.tea.2009.0155 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information. (370.9KB, docx)

Data Availability Statement

The datasets used and/or analysed during the current study available from the corresponding authors on reasonable request.

Articles from Scientific Reports are provided here courtesy of Nature Publishing Group

RESOURCES