Harnessing Small Extracellular Vesicles for Pro-Oxidant Delivery: Novel Approach for Drug-sensitive and resistant Cancer Therapy

Abstract

Multidrug resistance (MDR) is an inevitable clinical problem in chemotherapy due to the activation of abundant P-glycoprotein (P-gp) that can efflux drugs. Limitations of current cancer therapy highlight the need for the development of a comprehensive cancer treatment strategy, including drug-resistant cancers. Small extracellular vesicles (sEVs) possess significant potential in surmounting drug resistance as they can effectively evade the efflux mechanism and transport small molecules directly to MDR cancer cells. One mechanism mediating MDR in cancer cells is sustaining increased levels of reactive oxygen species (ROS) and maintenance of the redox balance with antioxidants, including glutathione (GSH). Herein, we developed GSH-depleting benzoyloxy dibenzyl carbonate (B2C)-encapsulated sEVs (BsEVs), which overcome the efflux system to exert highly potent anticancer activity against human MDR ovarian cancer cells (OVCAR-8/MDR) by depleting GSH to induce oxidative stress and, in turn, apoptotic cell death in both OVCAR-8/MDR and OVCAR-8 cancer cells. BsEVs restore drug responsiveness by inhibiting ATP production through the oxidation of nicotinamide adenine dinucleotide with hydrogen (NADH) and inducing mitochondrial dysfunction, leading to the dysfunction of efflux pumps responsible for drug resistance. In vivo studies showed that BsEV treatment significantly inhibited the growth of OVCAR-8/MDR and OVCAR-8 tumors. Additionally, OVCAR-8/MDR tumors showed a trend towards a greater sensitivity to BsEVs compared to OVCAR tumors. In summary, this study demonstrates that BsEVs hold tremendous potential for cancer treatment, especially against MDR cancer cells.

Graphical Abstract

1. Introduction

In 2019, the World Health Organization estimated cancer to be the first or second leading cause of death in those under 70 years of age in 112 of 183 countries [1]. One major impediment to successful cancer treatment is the pathophysiological phenomenon of the multi-drug resistance (MDR) [2–4]. Development of MDR and reduced drug sensitivity are responsible for both failure of traditional chemotherapy and over 90% of mortality in patients with metastatic cancer [5, 6]. One of the major MDR mechanisms is mediated via the expression of adenosine triphosphate (ATP)-binding cassette (ABC) transporters in the cellular membrane, which act as a drug efflux pump and are crucial for removing cytotoxic drugs from cancer cells [7]. The principal transporter involved in this MDR mechanism is P-glycoprotein (P-gp), also known as multidrug resistance protein 1 (MDR protein 1), overexpression of which in cancer cells puts most cancer patients at risk for developing MDR.

ABC transporters consume energy released from ATP produced in mitochondria to release various drugs from cells [8]. Inhibiting the synthesis of ATP, which requires mitochondrial membrane potential (MMP) and nicotinamide adenine dinucleotide with hydrogen (NADH), has garnered interest in overcoming cancer drug resistance [9, 10]. Mitochondria are considered the cellular power plant; they utilize oxygen for energy generation and are a major source of reactive oxygen species (ROS), which are byproducts of ATP production [11]. ROS contain an unpaired oxygen atom and, at basal levels, are essential in cell signaling and the cell cycle. However, when overexpressed, ROS induce cell death [12–14]. In the mitochondria, burst-generated ROS induce rapid depolarization of the mitochondrial membrane and obliterate the mitochondria resulting in reduced ATP supply to the efflux pumps [9, 15–17].

Cancer cells typically have higher ROS levels than do their normal counterparts due to the hypoxic environment, high metabolism rate, and gene mutations; moreover, adaptations to long-term and repeated chemotherapeutic treatments, including doxorubicin, buthionine, imexon, and cisplatin, induce MDR which, in turn, causes higher levels of ROS than seen in non-MDR cells [18–22]. The antioxidant system is a defense mechanism for cells with high levels of ROS, and maintaining a balanced redox status is essential in the carcinogenesis [23, 24]. L-γ-glutamyl-L-cysteinyl-glycine (GSH) is a tripeptide that is both a major antioxidant protecting against ROS-induced damage and a participant in many metabolic processes [25]. Elevated GSH protects cells against high free radical levels and contributes to drug resistance development [26]. Thus, GSH is a potent target for cancer therapy; normal cells are less sensitive to changes in GSH levels due to their well-controlled redox balance and low levels of ROS [27]. Due to the high oxidative stress environment of MDR cells, depleting the antioxidant system will further amplify oxidative stress in such cells [20]. In addition, the upregulation of proteins involved in intracellular GSH synthesis correlates with the expression of drug resistance and leads to high levels of intracellular GSH in MDR cells [28–30]. In terms of drug transport, GSH is a crucial component of the efflux function of the ABC transporter P-gp [31]. Thus, we hypothesized that GSH depletion would alter the redox balance in cancer cells and MDR cells by allowing ROS to accumulate while also inducing an efflux pump malfunction and a lack of energy for the ABC transporter to overcome drug resistance.

We previously reported GSH depletion caused by the pro-oxidant benzoyloxy dibenzyl carbonate (B2C). B2C undergoes esterase-triggered hydrolysis to release two quinone methides (QM), which deplete GSH [32]. B2C greatly induced cancer cell death and mitochondrial disruption by changing redox balance resulting in significantly reduced tumor growth without significant treatment-related toxicity. Despite the excellent anticancer activity of B2C, the presence of the efflux pump in resistant cells is a major hurdle to extending the application of B2C to MDR cancer therapy. Therefore, the availability of a drug carrier for efficiently delivering B2C to MDR cancer cells will restore drug sensitivity and produce antitumor activity.

Small extracellular vesicles (sEVs), which are naturally released from cells and vary in size from 50-200 nm [33], have a lipid bilayer structure that allows loading of both hydrophobic and hydrophilic drugs and are emerging as drug-delivery vehicles [34]. At the cellular level, sEVs participate in cell-to-cell communication and regulate biological processes [35, 36]. Since the membrane topology of sEVs reflects the lipid and protein composition, including ligands and receptors, of the cell of origin, sEVs exhibit a high degree of tropism towards the parental cells [37]. Thus, developing tumor-derived sEVs as drug carriers circumvents the need for surface modification of the sEVs to achieve tumor targeting [23]. Additional advantages of using sEVs as drug carriers include biocompatibility, low immunogenicity, and stability against enzymes in biological fluids [38], as well as greater adherence to and internalization by cancer cells compared to conventional liposomes (~10-fold greater) [39]. Additionally, sEVs, as naturally-derived nanoparticles, can bypass and evade the efflux system and release drugs directly into the cytoplasm via the endocytosis [7, 40].

Inspired by the ability of B2C to destabilize the mitochondria and induce apoptosis, combined with the advantages of sEVs as a drug carrier, we developed tumor-derived sEVs loaded with B2C (BsEVs) for treating MDR cancer cells. We hypothesized that BsEVs would target cancer cells, bypassing the efflux pump and delivering B2C to the cytoplasm, in turn increasing ROS levels, leading to starvation of ATP in the efflux pump and inducing cancer cell death (Scheme 1). Herein, we report the therapeutic potential of BsEVs as an oxidative stress modifier and drug resistance inhibitor against both drug-sensitive and MDR cancer cells.

Scheme 1.

BsEVs-mediated inhibition of P-gp and subsequent induction of apoptosis in both MDR and non-MDR cancer cells.

2. Materials and methods

2.1. Materials

4-Hydroxybenzyl alcohol, triethyl amine, benzoyl chloride, sodium carbonate, GSH, esterase, doxorubicin, 2’,7’-dichlorofluorescein-diacetate (DCFH-DA), BODIPY^™ TR-X NHS Ester, IR780 iodide, 5,5’-dithiobis-(2-nitrobenzoic acid), and poly(ethylene glycol) methyl ether-block-poly(lactide-co-glycolide) (PLGA-PEG) were purchased from Sigma-Aldrich (St. Louis, MO). Dichloromethane (DCM) was obtained from Samchun (Seoul, Korea). ROS Detection ROS Assay Kit was purchased from Dojindo (Kumamoto, Japan) The 5,5’,6,6’-tetrachloro-1,1’,3,3’-tetraethylbenzimidazolyl-carbocyanine iodide (JC-1) Mitochondrial Membrane Potential Detection Kit was obtained from Biotium (Fremont, CA). MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide), Dead Cell Apoptosis Kits with Annexin V for Flow Cytometry, and CellROX were purchased from Thermo Fisher Scientific (Waltham, MA). Exo-check kit was purchased from System Bioscience (Palo Alto, CA). The NAD/NADH Assay Kit, Luminescent ATP Detection Assay Kit, Creatine Kinase (CK) Activity Assay Kit, LDH Assay Kit / Lactate Dehydrogenase Assay Kit, and Alanine Transaminase (ALT) Activity Assay Kit were obtained from Abcam (Cambridge, MA). The DeadEnd^™ Fluorometric TUNEL System was purchased from Promega (Madison, WI). Human ovarian cancer OVCAR-8 and Kuramochi (National Cancer Institute [NCI], Bethesda, MD), SKOV-3 (ATCC, Manassas, VA), OVCAR-4 and A2780 cells (Millipore Sigma, Burlington, MA) and immortalized human ovarian epithelial cells (HOSE) (Applied Biological Materials, Inc., Richmond, Canada) were purchased from commercial vendors and used in the study. Methyl-beta-cyclodextrin was obtained from Alfa Aesar (Ward Hill, MA). Chlorpromazine Hydrochloride was purchased from TCI (Tokyo, Japan). EZ Standard Pack 1, Anti-Rabbit Detection Module, 12-230 kDa Separation Module, and RePlex^™ Module were obtained from Bio-techne (Minneapolis, MN).

2.2. Synthesis of B2C

B2C was synthesized as previously reported [32]. In brief, 4-hydroxybenzyl alcohol (40.0 mmol) and triethylamine (40.0 mmol) were completely dissolved in 125 mL of dichloromethane (DCM) and were vigorously stirred at 0 °C for 0.5 h. Benzoyl chloride, 40 mmol in 25 mL of DCM was added dropwise to the mixture and stirred for 5 h at room temperature. The solution was mixed with sodium carbonate solution and extracted. 4-(hydroxymethyl)phenyl benzoate (1) was obtained using silica gel chromatography. 1 (4.38 mmol) and 1,1’-carbonyldiimidazole (8.76 mmol) were dissolved in 15 mL of DCM and stirred at room temperature for 1 h. 4-(benzoyloxy)benzyl 1H-imidazole-1-carboxylate (2) was obtained by silica gel chromatography. To synthesize 4,4′-(carbonylbis(oxy)bis(methylene))bis-(4,1-phenylene) dibenzoate (B2C), 3.1 mmol of 1, 2, and 4-(dimethylamino)pyridine were completely dissolved in 15 mL of tetrahydrofuran, and the reaction was allowed to incubate at 40 °C for 10 h. B2C was obtained by silica gel chromatography and recrystallization. The chemical structure of B2C was verified by NMR (JEOL 500, JEOL Ltd., Tokyo, Japan), δ 8.20 (d, J = 7.2 Hz, 4H), 7.65 (t, J = 7.2 Hz, 2H), 7.52 (t, J = 7.2 Hz, H), 7.76 (d, J = 8.4 Hz, 4H), 7.23 (d, J = 8.4 Hz, 4H), 5.20 (s, 4H); 13C NMR: δ 165.21, 155.12, 151.25, 133.84, 132.96, 130.35, 129.90, 129.53, 128.75, 122.11, 69.29.

2.3. Cell lines and cell culture

OVCAR-8/MDR and Kuramochi MDR cell lines were generated by sequential treatment with an escalating dose of doxorubicin (0.1-1 μM). The expression of MDR/ABCB1 was evaluated using MDR/ABCB1 primary antibody (cat no. 13978, 1:1000, Cell Signaling Technology). OVCAR-8, OVCAR-8/MDR, OVCAR-4, Kuramochi, and A2780 human ovarian carcinoma cells were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium containing fetal bovine serum (FBS) (10%) with 2% penicillin. SKOV-3 ovarian adenocarcinoma cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) containing 10% FBS with 2% penicillin. HOSE cells were cultured in Prigrow I medium containing 10% FBS with 2% penicillin. Cell culture was performed in 5% CO₂ at 37 °C.

2.4. Formulation and characterization of BsEVs

To obtain sEVs, OVCAR-8 cells were cultured in FBS-free RPMI 1640 for 48 h at 37 °C in 5% CO₂, and the cell culture medium was replaced. After 48 h, the medium was collected and centrifuged at 2000 x g for 10 min. The supernatant was filtered through a 0.22-μm membrane (Nalgene syringe filter, Thermo Fisher Scientific, Rockford, IL). Ultracentrifugation was performed at 4000 x g for 30 min to collect sEVs from 15 mL of cell culture media using Amicon Ultra Centrifugal Filters (Cutoff MW 50000, MilliporeSigma, Burlington, MA) [41–43]. B2C was dissolved in DMSO at a concentration of 10 mg/mL and added to sEVs. Loading of B2C in purified sEVs was facilitated using a probe sonicator (VEVOR, Rancho Cucamonga, CA) with 0.25″ tip. The process was performed in six sets of alternating 30-second of on and off with 20% amplitude within a three-minute interval and allowed for a two-minute rest period among each set. Then, BsEVs were incubated at 37 °C for 30 min to restore the integrity of exosomal membranes [44]. B2C-loaded sEVs were separated using a size-exclusion chromatography column (Sephacry S-100, Sigma-Aldrich, St. Louis, MO). Zeta potential of BsEVs was measured using a ZetaPALS (Brookhaven Instrument Corp., Holtsville, NY). The size and number of sEVs were determined by nanoparticle analysis using a Nanosight NS300 (Malvern Instruments Ltd., Malvern, UK) – to obtain data on particles, 3 videos (60 sec each) were recorded for each diluted sample (camera level 9). BsEVs were stained with 2% uranyl acetate solution, and the shape of BsEVs was observed using a high-resolution transmission electron microscope (H7600, Hitachi, Tokyo, Japan) operated at 80 kV. The sEV marker proteins were detected using an Exo-Check^™ Exosome Antibody Array.

2.5. The release profile of BsEVs

Purified BsEVs were placed in a dialysis membrane tube (Cutoff MW 3500, Spectrum Laboratories, Inc., Rancho Dominguez, CA), which was submerged in 0.05% v/v Tween 80 in pH 7.4, 7.0, and 6.5 buffer then incubated at 37 °C with stirring. The samples were collected at specified times and lyophilized. B2C was extracted from lyophilized powders using DMSO, and the B2C content was determined using a microplate reader (Spectramax M3, Molecular Devices, San Jose, CA).

2.6. BsEVs-mediated GSH depletion

Mixtures of 100, 200, and 400 μM of BsEVs in 400 μM of GSH with or without (100 μg/mL) esterase in PBS were prepared and incubated for 24 h. Opacity changes over time were recorded. 400 μM of B2C, BsEVs, and sEVs (3.92 ± 0.16 x 10¹⁰/mL) were incubated with 400 μM GSH with esterase (100 μg/mL), and samples were collected at specified times. Cells were seeded on 6-well plates and treated with BsEVs. After 2 h, cells were rinsed twice using fresh PBS and then harvested. Cell lysates were collected with 100 μL of lysis buffer on ice and centrifuged to obtain the supernatant. To measure GSH levels, 10 μL of samples and 50 μL of Ellman’s reagent (0.5 mM, 5,5’-dithiobis-(2-nitrobenzoic acid)) were mixed, and the optical density was measured at 405 nm.

2.7. Cytotoxic activity of BsEVs

The anticancer activity of BsEVs was explored by MTT assay. OVCAR-8 and OVCAR-8/MDR cells were seeded on a cell culture-treated 96-well plate at 1 x 10⁴ cells/well. After 24 h, cells were treated with BsEVs for an additional 24 h. To measure cell viability, an MTT assay was performed, and the absorbance intensity at 570 nm was measured using a microplate reader. Cell viability was determined by comparison with control cells.

2.8. ROS level measurement

OVCAR-8 and OVCAR-8/MDR cells were seeded into 35-mm glass bottom dishes (VWR International, Radnor, PA). Following incubation for 24 h, B2C or BsEV was added to the cells, and cells were maintained for 24 h. Cells were washed with fresh PBS twice, stained with 20 μM of DCFH-DA for 45 min, and rinsed. Images were recorded using a confocal microscope (FV1000, Olympus, Inc., Center Valley, PA). OVCAR-8 and OVCAR-8/MDR cells (6 x 10⁵/well in a 6-well plate) were treated with BsEVs for 24 h. The cells were incubated with 10 μM DCFH-DA for 30 at 37 °C in the dark. The samples were analyzed using a flow cytometer (Stratedigm-3, Stratedigm, Inc., San Jose, CA). Kuramochi and Kuramochi-MDR cells (1 x 10⁴) were seeded into a well of 96 clear bottom black plate (Corning, Corning, NY) and incubated overnight at 37 °C. The cells were treated with BsEVs and ROS levels were measured using a ROS Detection ROS Assay Kit.

2.9. Measuring the levels of NAD⁺ and NADH in cells

2 x 10⁶ OVCAR-8/MDR cells were cultured in T25 flasks. The cells were incubated with various treatments with a fixed concentration of B2C (20 μM) for 12 h. Then, cells were washed three times with fresh PBS, and levels of NAD⁺ and NADH were measured using a NAD/NADH Assay Kit.

2.10. Cellular uptake study

1 x 10⁵ ovarian cancer (OVCAR-8, OVCAR-8/MDR, OVCAR-4, A2780, or SKOV-3) cells and HOSE cells were seeded into 35-mm glass bottom dishes and incubated overnight at 37 °C. Cells were treated with BODIPY-BsEVs and washed at specified times. Samples were fixed with 4% paraformaldehyde and stained with DAPI. The stained cells were imaged with a Leica SP8 confocal system (Leica Microsystems, Inc., Depew, NY). In a separate study, 3D spheroids of OVCAR-8 and OVCAR-8/MDR were grown on non-treated polystyrene plates and were treated with BODIPY-BsEVs. Spheroids treated with BODIPY-loaded OVCAR-4 sEVs, and BODIPY-loaded PLGA-PEG nanoparticles served as controls. At 24 h after treatment, the cells were washed three times and the cell uptake of BsEVs was imaged with Leica SP8 confocal system and quantitated. For determining endocytosis-mediated BsEVs uptake by cells, OVCAR-8 and OVCAR-8/MDR cells were cultured in 35-mm glass bottom dishes and incubated at 37 °C overnight. Then, cells were either not treated or treated with one of the endocytosis inhibitors, methyl-β-cyclodextrin (2.5 mM) or chlorpromazine (10 μg/mL). At 2 h after incubation, cells were washed and replenished with fresh cell culture medium containing BODIPY-loaded BsEVs and incubated for an additional 2 h at 37 °C. Cells treated with BsEVs and incubated at 4 °C were included as a positive control for measuring endocytosis. Cells receiving no treatment were included as a negative control. The cells were then fixed with 4% paraformaldehyde and stained with DAPI. OVCAR-8 and OVCAR-8/MDR (5 x 10⁴) cells were seeded on 12 well-clear-bottom plates and incubated overnight. The cells were treated with BODIPY-loaded BsEVs at a given time and incubated. After the incubation, the cells were washed with fresh PBS and replenished with a cell culture medium containing lysotracker green (Cell Signaling Technology, Danvers, MA). The cells were visualized with a Leica SP8 confocal system, and images were captured.

2.11. Flow cytometry to assess the anticancer activity of BsEVs

OVCAR-8 cells and OVCAR-8/MDR cells were treated with various formulations and concentrations of B2C or BsEVs. After treatment, cells were gently washed with fresh PBS twice and harvested using ACCUSTASE (Innovative Cell Technologies, San Diego, CA). The cells were resuspended in 1 X binding buffer at a concentration of 1 x 10⁵ cells/mL and stained with Annexin V-FITC (fluorescein isothiocyanate) and propidium iodide for 15 min. After incubation, 400 μL of 1 X binding buffer was added to the samples. Cells were washed twice and resuspended in 0.5 mL of PBS. The stained cells were analyzed using a flow cytometer. To measure mitochondrial membrane potential, B2C or BsEVs-treated cells were harvested, resuspended in 0.5 mL 1X JC-1 reagent working solution, and incubated at 37 °C for 15 min. Cells were subsequently washed, resuspended in 0.5 mL of PBS, and analyzed by flow cytometer.

2.12. Western blot for apoptosis analysis

OVCAR-8 cells and OVCAR-8/MDR cells were incubated in a 6-well plate. Following incubation for 24 h with B2C or BsEVs, cells were rinsed with PBS. Lysates were collected from cells using cell lysis buffer (Goldbio, Olivette, MO) in accordance with the manufacturer’s protocol. 12-230 kDa Separation Module was prepared with 10 μg of proteins, primary antibodies (caspase-3 [cat no. 9662, 1:20, Cell Signaling Technology], PARP [cat no. 9542, 1:20, Cell Signaling Technology], and Anti-GAPDH [6C5] [cat no. ab8245, 1:50, Abcam]) and Goat anti-Mouse IgG secondary antibody [cat no. 926-32210, 1:100, LI-COR]). Simple western blot was performed using JESS (ProteinSimple, San Jose, CA).

2.13. Measuring the levels of cellular ATP

1.5 x 10⁴ cells were seeded in 96-well clear flat bottom plates. After 24 h, the cells were incubated with various specified treatments for 12 h. An ATP assay was performed using a Luminescent ATP Detection Assay Kit following the manufacturer’s protocol.

2.14. Distribution of any changes in P-glycoprotein

OVCAR-8/MDR cells were cultured in 35-mm glass bottom dishes for 24 h. Cells were treated with various formulations of BsEVs for 12 h and fixed with 4% paraformaldehyde. The samples were blocked with 5% BSA and incubated with MDR1/ABCB1 primary antibody (cat no. 13978, 1:100, Cell Signaling Technology) in 1% BSA at 4 °C overnight. Then, cells were washed three times with fresh PBS and incubated with Alexa Fluor^® 488-conjugated secondary antibody (cat no. ab150077, 1:200, Abcam) in 1% BSA for 1 h at room temperature. Nuclei were stained with DAPI for 5 min and washed three times with PBS. The cells were visualized with a Leica SP8 confocal system.

2.15. BsEVs-induced drug uptake restoration

1 x 10⁵ OVCAR-8/MDR cells were seeded on a 12-well glass bottom plate and incubated for 24 h. Cells were treated with 20 μM B2C or BsEVs. After 12 h, cells were incubated with 1 μg of doxorubicin in culture medium (1mL) for 2 h and fixed with 4% paraformaldehyde solution. Then, cells were washed three times using PBS and stained with DAPI for 5 min. The cells were imaged using a Leica SP8 confocal system.

2.16. Therapeutic antitumor activities of BsEVs in a mouse tumor xenograft model

OVCAR-8 cells and OVCAR-8/MDR cells (5 x 10⁶) were inoculated subcutaneously into the right flank of six-week-old female athymic nude mice (Charles River Laboratories, Wilmington, MA). The mice were monitored until the tumor volumes reached ~50 mm³ and were randomly grouped (6 mice/group). Various formulations (1 or 5 mg/kg B2C, BsEVs at a B2C dosage of 1 or 5 mg/kg in sEVs [2.04 ± 0.08 x 10¹⁰/kg], or sEVs [2.04 ± 0.08 x 10¹⁰/kg]) were intravenously injected every three days for a total of four treatments. Mice that did not receive any treatment served as controls. The body weights and tumor volumes were observed for 46 days. The tumor volumes were calculated using the following formula: (width² x length)/2. At the end of the experiments, mice were euthanized, and tissues (tumor and major organs) were excised for histological analysis. The samples were sliced and stained with hematoxylin and eosin (H&E) and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL). Animal studies were conducted in accordance with the approved protocol (21-037-CHIL; The University of Oklahoma Health Sciences Center, Institutional Animal Care and Use Committee).

2.17. In vivo biodistribution

IR780-loaded BsEVs (2.04 ± 0.08 x 10¹⁰/kg) were injected intravenously in OVCAR-8 and OVCAR-8/MDR tumor-bearing mouse and were euthanized after 12 h. Mice receiving no treatment served as controls. The major organs and tumor tissue were collected and imaged using an IVIS Spectrum Imaging System (PerkinElmer, Waltham, MA).

2.18. Immunotoxicity and systemic toxicity

To study the potential systemic toxicity of BsEVs, female athymic nude mice (n=4) were given 10 mg/kg of BsEVs intravenously once a day for seven-days. 4 mice did not receive any treatment served as controls. Whole blood was collected and kept at room temperature for 20 min, and serum was collected after centrifugation at 2000 g for 10 min. ALT, CK, and LDH levels were measured by commercially available kits. For hematology analysis, BsEVs (10 mg/kg) was injected intravenously in female C57BL/6 mice (n=4). Four days later, blood samples were taken collected using 0.5 M ethylenediaminetetraacetic acid (EDTA) (pH 8.0)-coated syringes and kept in tubes containing EDTA [45, 46]. The population of immune cells was analyzed using a ProCyte Dx Hematology Analyzer (IDEXX, Westbrook, ME).

2.19. Statistical analysis

Values were expressed as mean ± standard deviation (SD). Differences between two groups were evaluated using a two-tailed unpaired t-test. One-way analysis of variance (ANOVA) was conducted using GraphPad Prism 9 (San Diego, CA) to make comparisons among multiple groups.

3. Results

3.1. Development of GSH-scavenger delivery system

B2C was designed to release two QM intermediates in an esterase-catalyzed manner. The chemical structure of B2C was confirmed by NMR spectroscopy (Figs. S1A–B). BsEVs were formulated by sonication to load B2C into cancer cell-derived sEVs. The hydrodynamic diameter of BsEVs was determined by NTA to be 205.2 ± 60.7 nm (Fig. 1A), and their spherical shape was confirmed by TEM (Fig. 1B). BsEVs are slightly larger than empty sEVs, which have diameters of 180.2 ± 72.3 nm (Figs. S2A–B). BsEVs had a zeta potential −25.8 ± 3.8 mV, lower than that of empty sEVs (−9.8 ± 1.5 mV; Fig. 1C); a difference that can be explained by the charge of the molecules encapsulated in the sEV [47, 48]. The membrane integrity of BsEVs was evaluated using an exosome array (Fig. 1D and Fig. S2C). BsEVs strongly expressed characteristic membrane proteins and inner lumen proteins (Tsg101 and ALIX) similar to empty sEVs, indicating that B2C loading did not affect the sEV membrane integrity [49]. Size exclusion chromatography (SEC) was performed to assess the elution profile of BsEVs and demonstrate the loading of the small molecule B2C in sEVs. Following SEC of BsEVs, both B2C and sEVs were found in the same fraction, i.e., fraction 10 in the elution profile graph, suggesting that B2C was loaded in sEVs (Fig. 1E). The loading content of B2C in BsEVs was calculated from the calibration curve of B2C (at 284 nm), and the concentration of B2C in sEVs was normalized by the number of sEVs; ~10¹⁰ sEVs incorporate 49.11 ± 2.04 μg of B2C (Fig. S3).

Fig. 1.

Development of BsEVs to deplete GSH. A, hydrodynamic diameter of BsEVs in pH 7.4 phosphate buffer. B, A representative transmission electron microscopy image of BsEVs. C, Zeta potential of OVCAR-8 sEVs and BsEVs. D, The expression of sEV marker proteins in BsEVs. E, Size exclusion chromatography elution profile of BsEVs. F, Release profile of B2C from BsEVs. Values are mean ± SD (n=4). G, BsEVs-triggered GSH depletion with or without esterase. Values are mean ± SD (n = 6. ****p < 0.001). H and I, GSH depletion-induced anticancer activity of BsEVs. GSH level after treatment in OVCAR-8 cells (H) and OVCAR-8/MDR cells (I). Values are mean ± SD (n = 4. **p < 0.05, ***p < 0.01 and ****p < 0.001).

We next studied the release kinetics of B2C from sEVs at the different pH. BsEVs in pH 7.4 released ~ 50% of B2C in the first 3.5 days, but BsEVs in cytoplasmic pH (7.0) and endosomal pH (6.5) released 50% of B2C in 2 days and then showed moderate sustained release through 18 days (Fig. 1F). The stability of BsEVs at various pH levels was verified by the size of BsEVs incubated for 24 hours in environments with pH 7.4, 7.0, and 6.5. In Fig. S4, samples incubated at each pH showed a similar size distribution as BsEVs before incubation. To confirm whether BsEVs deplete GSH through the esterase-catalyzed cleavage of ester linkages in B2C, BsEVs were incubated for 24 h in a solution of GSH. In the presence of esterase, BsEVs depleted GSH in a concentration and time-dependent manner (Fig. 1G and Fig. S5A). However, in the absence of esterase, BsEVs barely depleted GSH, while empty sEVs did not affect GSH levels. These observations indicate that loading B2C in sEVs improves its poor water solubility and enhances its dissolution rate [50]. The esterase-triggered cleavage of B2C in BsEVs was also indirectly evidenced by changes in the opacity of BsEVs over 24 h (Fig. S5B).

OVCAR-8/MDR cells were established by continuous long-term exposure of OVCAR-8 cells with doxorubicin which increases ROS generation in cells. The development of MDR was confirmed by measuring MDR pump activities and the level of MDR-1 protein (Figs. S6A–B). The loss of therapeutic response in OVCAR-8/MDR cells compared to OVCAR-8 cells was assessed by MTT assay using various concentrations of doxorubicin. OVCAR-8/MDR cells were resistant to doxorubicin, while OVCAR-8 cells were sensitive (Fig. S6C). Changes in ROS and GSH levels were also determined, showing that OVCAR-8/MDR cells had significantly higher ROS and GSH than OVCAR-8 cells (Figs. S6D–E, p < 0.001). These data demonstrate that long-term treatment with doxorubicin allows OVCAR-8 cells to acquire MDR characterized by overexpression of MDR proteins and MDR pump activation.

Since BsEVs were developed based on a GSH-depleting strategy, we first investigated their ability to scavenge intracellular GSH using Ellman’s reagent. As expected (Fig. S5A), BsEVs exhibited greater GSH depletion ability than equivalent levels of free B2C (20 μM and 40 μM) in both sensitive (OVCAR-8) and MDR (OVCAR-8/MDR) ovarian cancer cells. (Figs. 1H–I). In order to validate the strategy of depleting GSH in MDR cancer cells, we treated BsEVs to both Kuramochi and Kuramochi/MDR cells. This resulted in a decrease in GSH levels and an increase in ROS levels in Kuramochi/MDR cells (Fig. S7A–E).

3.2. Elevation of changes in redox status by GSH-depletion strategy regardless of MDR

To explore the GSH depletion-mediated cytotoxicity of BsEVs, an MTT assay was performed on OVCAR-8 cells and OVCAR-8/MDR cells. The anticancer activity of B2C was remarkably increased by sEV encapsulation; the effect of BsEVs was ~1.79-fold greater on OVCAR-8 cells and ~2.23-fold greater on OVCAR-8/MDR cells than free B2C at 30 μM (Figs. 2A–D). N-acetyl cysteine (NAC) treatment suppressed this cytotoxicity due to its antagonistic functions, i.e., inactivation of ROS and replenishment of the GSH [51]. These findings demonstrate that GSH inhibition is the basis for the anticancer activity of BsEVs. Interestingly, we observed that BsEVs were more efficacious against OVCAR-8/MDR cells than OVCAR-8 cells. Specifically, the lowest half-maximal inhibitory concentrations (IC₅₀) of BsEVs were 10.78 μM for OVCAR-8 cells and 6.53 μM for OVCAR-8/MDR cells, while in the case of B2C, they were 28.44 μM for OVCAR-8 cells and 43.63 μM for OVCAR-8/MDR cells. For BsEVs + NAC, the IC₅₀ was 17.35 μM for OVCAR-8 cells and 21.68 μM for OVCAR-8/MDR cells. These results demonstrate that sEV encapsulation enhances the therapeutic efficacy of B2C, likely due to instinct properties of sEVs, such as bypassing the efflux pump [52]. We evaluated cytotoxicity of BsEVs against non-cancerous HOSE cells. In Fig. S8, HOSE cells showed higher cell viability than cancer cells. This can be explained by the well-maintained low ROS levels and redox balance in normal cells [53].

Fig. 2.

Therapeutic effect of BsEVs-induced oxidative stress. A and B, Cytotoxicity of various concentrations of BsEVs in OVCAR-8 (A) and OVCAR-8/MDR cells (B). C and D, Cytotoxicity at 30 μM against OVCAR-8 (C) cells and OVCAR-8/MDR cells (D). Values are mean ± SD (n = 4. ****p < 0.001). E, Representative confocal images of ROS expression in cells treated with various formulations. F and G, Intracellular DCF fluorescence signals ratio to express the change of ROS level in OVCAR-8 (F) and OVCAR-8/MDR cells (G). Values are mean ± SD (n = 19-47. ****p < 0.001). H, The level of NADH and NAD⁺ in OVCAR-8/MDR cells. I, The ratio of NAD⁺/NADH in OVCAR-8/MDR after various treatments. Values are mean ± SD (n = 4. ****p < 0.001).

The elevation of the intracellular level of ROS by BsEVs-induced GSH depletion was confirmed by confocal microscopy. Cells were incubated with sEVs, B2C, BsEVs, or BsEVs plus NAC and then stained with DCFH-DA. BsEVs-treated cells exhibited strong green fluorescence signals indicating high ROS, while treatment with NAC suppressed ROS levels and thus fluorescence (Fig. 2E). BsEVs caused ~3.8-fold greater fluorescence signals in OVCAR-8 cells and ~6.8-fold stronger signals in OVCAR-8/MDR cells compared to the control (Figs. 2F–G). In addition, the intracellular ROS levels in the cells after BsEVs treatment were evaluated using flow cytometry (Figs. S9 A–D). BsEVs led to a rise in intracellular ROS levels, as demonstrated by the peaks shifting to the right, with DCF signals in OVCAR-8 being approximately 1.4 times higher, and in OVCAR-8/MDR cells being around 1.3 times higher than untreated cells. Since elevated ROS is known to oxidize NADH to nicotinamide adenine dinucleotide (NAD⁺) [9], we measured the levels of NADH and NAD⁺ after the various treatments. NADH content was decreased in cells treated with BsEVs, and the NAD⁺/NADH ratio increased (~2.3-fold; Figs. 2H–I), while the total NAD⁺ and NADH contents were similar across all groups. These results correspond to the findings of a previously reported strategy to overcome MDR, using ROS-induced NADH oxidation to reduce the generation of ATP [9, 18].

3.3. Origin-specific cellular uptake and cell entry mechanism of BsEVs.

To substantiate the benefits of tropism of tumor cells-derived sEVs, cellular uptake of BsEVs was investigated. Fig. 3A–B shows fluorescence images and quantified fluorescence signals of various cells (HOSE, OVCAR-8, OVCAR-8/MDR, OVCAR-4, A2780, and SKOV-3) incubated with fluorescent BODIPY-loaded BsEVs for 1 h. Strong and the greatest red fluorescence signal was observed only in the cytoplasm of OVCAR-8 and OVCAR-8/MDR cells, demonstrating that BsEVs, composed of sEVs derived from OVCAR-8 cells exhibit cell tropism and are readily internalized by the cell of origin. Three-dimensional (3D) tumor spheroids were also used to investigate the ability of BsEVs to penetrate tumor tissues by comparing BsEVs to equal numbers of PLGA-PEG nanoparticles and OVCAR-4 derived sEVs. BsEVs showed stronger fluorescence than either OVCAR-4 sEVs or PLGA-PEG nanoparticles in OVCAR-8 and OVCAR-8/MDR spheroids (Figs. 3C–E). Red fluorescence was seen both at the outer surface and distributed inside the spheroid structure. These findings suggest that BsEVs take advantage of the cancer cell-type tropism for potent tissue penetration.

Fig. 3.

Cellular uptake of BsEVs. A, Cellular uptake of BODIPY-labeled BsEVs in various cell lines. B, Quantification of fluorescence signals in the various cell lines. C, Comparison of cell-specific uptake of PLGA-PEG nanoparticles, OVCAR-4 sEVs, and BsEVs in OVCAR-8 and OVCAR-8/MDR tumor spheroids. D and E, Quantification of BsEVs uptake in OVCAR-8 spheroid (D) and OVCAR-8/MDR spheroid (E). Values are mean ± SD (n = 5. **p < 0.05, ***p < 0.01, and ****p < 0.001). F, Endocytosis mechanism of BsEVs in OVCAR-8 cells and OVCAR-8/MDR cells after incubation with BODIPY-labeled BsEVs in the presence of cellular uptake inhibitors or at 4 °C. G and H, Quantification of fluorescence signals of OVCAR-8 cells (G) and OVCAR-8/MDR cells (H). Values are mean ± SD (n = 7-18. **p < 0.05 and ****p < 0.001).

The route(s) by which BsEVs enter the recipient cell, including fusion and cellular uptake, were investigated. First, the fusion efficiency of BsEVs for OVCAR-8 cells or OVCAR-8/MDR cells was assessed. Cells were treated with lipophilic fluorescence dye (R18)-labeled BsEVs, and the fluorescence dequenching was monitored for three hours. The lipid mixing was observed and reached ~6% and ~8 % of the maximum fluorescence recorded after Triton X-100 treatment for OVCAR-8 and OVCAR-8/MDR, respectively (Figs. S10A–B). R-18-loaded BsEVs showed no spontaneous dequenching before Triton X-100 addition (Figs. S10C–D). Thus, less than 10% of BsEVs were internalized by the cell through fusion. Next, we performed flow cytometry to study the uptake route of BsEVs. BODIPY-FL-1,2-dihexadecanoyl-sn-glycerol-3-phosphoethanolamine (DHPE) was incorporated in BsEVs; on sEVs, BODIPY-FL has red fluorescence (BODIPY dimer), and the wavelength is maintained under endocytosis. However, membrane fusion dilutes BODIPY-FL and changes the spectrum from red to green [54]. Fig. S10E shows that the population of cells with red fluorescence (upper left) increased over time, demonstrating that cellular uptake of BsEVs is mainly via endocytosis. A cellular uptake study was performed to clarify the endocytosis mechanism of BsEVs using a caveolin-mediated endocytosis inhibitor (methyl-β-cyclodextrin) and a clathrin-mediated endocytosis inhibitor (chlorpromazine) at 37 °C or by incubating at 4 °C to suppress energy-dependent endocytosis. BsEVs internalization was strongly suppressed at 4 °C and in the methyl-β-cyclodextrin (caveolin-mediated endocytosis inhibitor) treatment group (Figs. 3F–H). These observations demonstrate that BsEVs are internalized mainly through caveolin-mediated endocytosis in an energy-dependent manner. Finally, the intracellular transport of BsEVs was evaluated using BODIPY-loaded BsEVs due to the similar release kinetics of BODIPY as B2C (Fig. S11A). As shown in Figs. S11B–D, we found that the internalized strong dot fluorescence signals of nanoparticles and light signals of released BODIPY in the cells within 1 h. Particle-shaped fluorescence signals were observed in greater numbers over a larger area than the lysosomal distribution in OVCAR-8 and OVCAR-8/MDR cells over 6 h (Figs. S11C–D). This might be attributed to endocytosis mediated by caveolin, leading to the formation of caveosomes, or multicaveola complexes. Since caveolins are unable to fuse with lysosomes, this allows BsEVs to escape degradation by lysosomes [55].

3.4. Therapeutic effect and overcoming MDR of GSH-scavenging system

To confirm BsEVs-mediated apoptosis, cells were stained with Annexin V-FITC (fluorescein isothiocyanate) as an apoptosis marker and propidium iodide (PI) as a cell viability marker using flow cytometry. Both B2C and BsEVs induced apoptotic death of OVCAR-8 cells and OVCAR-8/MDR cells in a concentration-dependent fashion, as evidenced by the increased populations of late apoptotic cells in the upper right quadrant (Figs. 4A–B). The apoptotic cell death was inhibited by antioxidant NAC, indicating that apoptosis is induced by the GSH-depleting B2C in BsEVs. Interestingly, we also found that BsEVs induced more apoptotic cell death than equivalent free B2C (Figs. 4C–D). Western blotting for apoptosis-related proteins was performed to provide more mechanistic insight. BsEVs resulted in the cleavage of caspase-3 and PARP, thereby increasing the ratio of cleaved caspase-3 to caspase-3 and cleaved PARP to PARP. (Figs. 4E–J), confirming BsEVs-mediated apoptosis.

Fig. 4.

BsEVs-induced apoptosis. A and B, Flow cytometric analysis of OVCAR-8 cells (A) and OVCAR-8/MDR cells (B) stained with Annexin V-FITC and propidium iodide. C and D, Quantification of BsEVs-induced apoptosis in OVCAR-8 (C) and OVCAR-8/MDR cells (D). Values are mean ± SD (n = 3. ****p < 0.001). E and F, Apoptosis-related protein expression of OVCAR-8 cells (E) and OVCAR-8/MDR cells (F). G and H, The ratio of cleaved caspase-3/caspase-3 (G) and cleaved PARP/PARP (H) of OVCAR-8 cells. I and J, The ratio of cleaved caspase-3/caspase-3 (I) and Cleaved PARP/PARP (J) of OVCAR-8/MDR cells after various treatments. Values are mean ± SD (n = 3. *** p < 0.01, and ****p < 0.001).

Reduction in mitochondrial membrane potential (MMP) is a key event of the apoptosis [56], and MMP disruption of the mitochondrial membrane reduces ATP production, leading to energy depletion of the efflux pump on the cellular membrane [8, 57]. To analyze changes in MMP, cells were stained with JC-1 to assess changes in membrane permeability (Figs. 5A–D). BsEVs induced significant loss of MMP, while NAC protected against BsEVs-mediated mitochondrial alteration. We also measured ATP levels to examine the effects of MMP alteration on the level of ATP. BsEVs caused a significant reduction in ATP to ~ 80% and ~55% of untreated cell levels in the 10 μM and 20 μM treatment groups, respectively (Fig. 5E). Since limited ATP is known to suppress the function and expression of P-gp [9, 58], the degree of P-gp expression in the cell membrane of OVCAR-8/MDR cells was evaluated. The green fluorescence signal for P-gp in the cellular membrane was significantly decreased by treatment with BsEVs (Fig. 5F). Motivated by the finding of P-gp suppression by BsEVs, we assessed the cellular uptake of doxorubicin, which has intrinsic fluorescence, in OVCAR-8/MDR cells. As shown in Fig. S12, red fluorescence of doxorubicin was observed after BsEVs treatment. These results suggest that BsEVs could overcome MDR in ovarian cancer cells by reducing MMP, ATP supply, and P-gp expression.

Fig. 5.

BsEVs-induced MMP disruption and efflux pump inhibition. A and B, Changes in mitochondria membrane potential in OVCAR-8 cells (A) or OVCAR-8/MDR cells (B) with BsEVs treatment. C and D, Quantitative analysis of the ratio of J1 monomer and aggregates in OVCAR-8 (C) and OVCAR-8 MDR cells (D). Values are mean ± SD (n = 3. ***p < 0.01, and ****p < 0.001). E, ATP level changes after various treatments in OVCAR-8/MDR cells. Values are mean ± SD (n = 4. ****p < 0.001). F, Representative confocal images of P-gp after various treatments in OVCAR-8/MDR cells.

3.5. In vivo therapeutic activity and biosafety evaluation

The anticancer efficacy of BsEVs was explored using a mouse tumor xenograft model. BsEVs were injected intravenously every three days for four total doses, and tumor growth and body weight were observed for 46 days. As shown in Figs. 6A–C and Fig. S13A, B2C alone at the two concentrations tested and sEV alone showed negligible effects on OVCAR-8 tumor growth. However, treatment with BsEVs at a dose of 5 mg/kg but not at 1 mg/kg dramatically suppressed tumor growth. Similar results were observed in OVCAR-8/MDR xenograft mice with BsEVs at 5mg/kg showing the greatest and most significant tumor growth inhibition (Figs. 6E–G and Fig. S13B, p < 0.001). During the 46 days of observation, treatment with BsEVs did not alter the body weight of OVCAR-8 and OVCAR-8/MDR mice (Figs. 6D and 6H), indicating the excellent biosafety of BsEVs. The results in xenograft mouse models indicate that BsEVs exert potent anticancer therapeutic efficacy in both MDR and non-MDR ovarian cancer mouse models with observable no treatment-related toxicity.

Fig. 6.

In vivo therapeutic anticancer activity of BsEVs. A, Representative images of tumor-bearing mice after treatment with various doses of BsEVs in OVCAR-8. B-D, Tumor weight measurement (B), changes in tumor volume (C), and body weights (D) in OVCAR-8 xenograft mice. E, Representative images of tumor-bearing mice after treatment with various doses of BsEVs in OVCAR-8/MDR. F-H, Determination of tumor weight (F), changes in tumor volume (G), and body weights (H) in OVCAR-8/MDR xenograft mice. Values are mean ± SD (n = 4-6. *p < 0.1, ***p < 0.01, and ****p < 0.001).

Histological studies were carried out to further evaluate the anticancer actions of BsEVs. While tumor tissues of untreated mice showed normal membrane structures and nuclei, dead tumor cells without nuclei were seen in the BsEVs treated 5 mg/kg group (Fig. 7A). Strong fluorescence signals of TUNEL-positive cells were found in the BsEVs 5 mg/kg treatment group (Fig. 7B), suggesting that BsEVs induce apoptotic tumor cell death. To confirm whether BsEVs induce oxidative stress in the tumor, tissues were stained with CellROX; mice treated with 5 mg/kg BsEVs showed the highest level of ROS in tumor tissues (Fig. 7C), consistent with the increased generation of ROS in both OVCAR-8 cells and OVCAR-8/MDR cells reported above (Fig. 2 and Fig. S9).

Fig. 7.

Histological evaluation of tumor tissues. A, Representative H&E-stained images. B, Representative TUNEL-stained images. C, Representative CellROX-stained images.

3.6. Systemic biodistribution and toxicity

To investigate the systemic biodistribution of BsEVs, IR780-encapsulated BsEVs were employed. BsEVs accumulated in the tumor tissues 12 h after tail vein injection in both OVCAR-8 and OVCAR-8/MDR xenograft models (Fig. 8A). BsEVs accumulated in tumor tissue specifically, which can be explained by the tropism of target cell-derived sEVs (Fig. 3A). Immune suppression is the most common side effect of chemotherapy [59]; thus, the impact of BsEVs administration on immune suppression was studied. The population of immune cells was not significantly changed in BsEVs-treated mice compared to the untreated group (Fig. 8B). BsEVs treatment elicited no significant effect on the number of WBCs or the proportion of neutrophils (Figs. 8C–D). Serum tests were performed to evaluate the hepatic function, muscular damage, and tissue damage following the administration of BsEVs. BsEVs-treated mice showed no significant changes in ALT, CK, or LDH levels compared with the untreated group (Figs. 8E–G). In addition, there was no substantial histological evidence of toxicities in major organs after BsEVs administration (Fig. 8H), consistent with the finding of minimal or no body weight change in the tumor-bearing mouse study (Figs. 6D and 6H).

Fig. 8.

Biodistribution and biosafety of BsEVs. A, Fluorescence images of dissected organs and tumor tissues after IR780-loaded BsEVs injections. B-D, Hematologic analysis after BsEVs administration. B, Dot plot analysis of blood. C, Total numbers of WBC. D, Percentage of neutrophil in WBC. E-G, In vivo toxicity tests of BsEVs. E, The level of creatine kinase. F, The level of alanine transaminase. G, The level of lactate dehydrogenase. Values are mean ± SD (n=4). H, H&E staining of organs from BsEVs-treated mice.

4. Discussion

ROS in biological systems play important roles depending on the balance of antioxidants and free radicals. The basal level of ROS is necessary for basic cellular processes in normal cells; however, cancer cells have elevated ROS levels along with overexpressed antioxidants due to metabolic alterations and proliferation, which induces vulnerabilities based on the imbalanced redox status [20]. Based on these phenomena, many chemotherapeutic agents have been developed to kill cancer cells by elevating intracellular ROS; this strategy induces MDR, which leads to the overproduction of both antioxidants and ROS in a cell [22]. Thus, the regulation of ROS is a potential strategy for controlling MDR in cancer therapy.

We previously developed B2C, which generates two QMs to deplete the antioxidant GSH in the presence of esterase. B2C exerted anticancer activities by amplifying oxidative stresses in various cancer cell lines. The therapeutic potential of B2C, which can disrupt redox balance, is greater against MDR cancer cells which express high levels of ROS and antioxidants; however, the elevated expression of the efflux pump P-gp in MDR cells is a major hurdle to the successful application of hydrophobic small molecular drugs, including B2C [60].

sEVs are naturally derived nanoparticles with low systemic immunogenicity and toxicity. sEV-mediated delivery provides multiple advantages, including stability and specific cell-targeting activity, due to the intrinsic role of extracellular vesicles in biological systems [23]. In cancer therapy, sEVs provide the advantages of nanoparticles with biological roles and can deliver payloads through the bypass route of P-gp in MDR cells, accompanied by long-term chemotherapy [44, 61]. To develop a more effective agent for the therapy of both normal and MDR tumors, we designed BsEVs, which overcome drug resistance and specifically target tumor sites. BsEVs were formulated to encapsulate B2C into OVCAR-8 cancer cell-derived sEVs and provide stability and sustained release of the small molecule without losing the intrinsic sEV characteristics.

A previous study showed that 20 mg/kg unencapsulated B2C suppresses tumor growth in a human colon adenocarcinoma mouse xenograft model [32]; herein, we show that BsEVs suppressed the growth of ovarian tumors at 5 mg/kg B2C equivalent without any additional chemotherapeutic agents. These findings indicate that sEVs enhance the therapeutic efficacy of B2C.

The tropism of cancer-derived sEVs allows BsEVs to achieve selective targeting of OVCAR-8 cells and OVCAR-8/MDR cells compared with other ovarian cancer cell lines and HOSE cells. The cellular uptake of BsEVs is dependent on endocytosis rather than fusion, thus bypassing the efflux systems. Cellular internalization depends on energy and the caveolin-mediated endocytosis pathway. Utilizing targetability and the endocytic route of uptake makes sEV encapsulation of B2C for cancer therapy a rational approach. As designed to utilize the benefits of sEVs for therapeutic purposes, BsEVs amplified ROS levels by reducing the GSH contents of cancer cells. BsEVs exhibited stronger cytotoxicity against OVCAR-8 cells and OVCAR-8/MDR cells than did free B2C, and BsEVs treatment induced mitochondrial disruption and upregulation of apoptotic protein expression. Administrating NAC alongside BsEVs suppressed the apoptosis pathway and cell death, suggesting that the anticancer effect of BsEVs is based on the imbalance of ROS levels. Increased oxidative stress can reduce ATP supply because it alters the ratio of NAD⁺/NADH and disrupts MMP. P-gp expression was downregulated in MDR cells, consistent with reduced ATP supply following BsEVs therapy. Thus, we believe that BsEVs represent an effective anticancer strategy that not only kills cancer cells by amplifying ROS but also overcomes MDR by suppressing the ATP supply to the efflux pump.

Drug resistance cancer therapy research using drug-loaded nanostructures has focused on overexpressed GSH as a target to overcome MDR, where the skeleton of nanoparticles includes GSH-sensitive linkages that trigger drug release. Examples include a pH and reduction-responsive polymer with doxorubicin [9], the Pt(IV) prodrug nano-assembly delivery system [62], and a redox-responsive micellar nanodrug carrier with doxorubicin [30]. These anticancer drug delivery systems exert anticancer effects against abundant GSH in MDR models. In the case of our BsEVs delivery system, even without additional conventional anticancer drugs, we saw notable suppression of tumor growth without significant damage to major organs or changes in enzyme levels. Many chemotherapeutic agents, including cyclophosphamide, 5-fluorouracil, and cytosine arabinoside induce neutropenia [63], for example, low-dose cyclophosphamide (150 mg/kg), which does not suppress tumor [64], induces severe neutropenia (≤10 neutrophils/mm³) [45]. However, a therapeutic dose of BsEVs did not significantly affect neutrophil numbers. Thus, manipulating GSH levels may be a potent therapeutic strategy for reducing chemotherapy-related side effects.

To combat MDR cancer, diverse nanomaterials have been developed using drug-resistant cell line-based models [18, 65, 66]. Herein, we assessed the anticancer activities of BsEVs in both MDR and sensitive cancer cell lines. The occurrence of drug resistance increases the levels of GSH and intracellular ROS, providing a forceful rationale for adopting GSH depletion for therapy. Our data show that more B2C is necessary to kill MDR cells than non-MDR cells due to the efflux pump; however, sEV encapsulation of B2C dramatically increased its efficacy (Figs. 2C–D). Interestingly, treatment with equivalent doses of BsEVs induced greater apoptosis and MMP alterations in MDR cells than in normal cells (Figs. 4A–F), probably due to the naturally high level of ROS in MDR cells and their vulnerability to oxidative stress. These findings corroborate our hypothesis that BsEVs, which take advantage of the relationship between drug tolerance and levels of GSH, can be a potent therapy for MDR cancer. To our knowledge, this study is the first evaluation of a GSH-depletion agent delivered via sEV as a treatment for both MDR and sensitive cancer cells; our data confirm that this strategy is more effective against MDR cancer cells than against sensitive cancer cells. Therefore, the GSH-depletion strategy is a potentially potent cancer therapy regardless of the resistance status of the target tumor.

Although BsEVs showed therapeutic potential in both sensitive and MDR cancer cells in in vitro and in vivo studies, further the fate of BsEVs after cellular uptake, pharmacological and toxicological studies in various MDR cancer cells are needed to support the potential of BsEVs as an anticancer therapy. Moreover, comparative studies using other cancer cell lines with high ROS levels are needed to expand the application of BsEVs as an option for long-term chemotherapy.

5. Conclusion

In summary, we developed BsEVs, a novel delivery system for B2C which amplifies oxidative stress by depleting GSH. BsEVs improved the therapeutic effect of free B2C and delivered B2C to the target by tropism. The redox balance of cancer cells was altered by BsEVs, and apoptosis was induced. High accumulation of ROS with GSH depletion induced by BsEVs led to reduced ATP supply mediated by MMP disruption and oxidation of NADH, resulting in an inability of efflux pumps to function. In tumor-bearing mouse models, BsEVs treatment suppressed tumor growth without notable side effects. Thus, BsEVs are expected to be a promising strategy for overall cancer treatment in both normal cancer cells and MDR cells, and holds great potential to facilitate clinical translation of a GSH-depletion strategy.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.