Chemicals and materials
Bis(2,4,5-trichloro-6-carbopentoxyphenyl) oxalate (CPPO), Ficoll, DCFH-DAm and lipopolysaccharides (LPS) were obtained from Sigma-Aldrich (St. Louis, MO, USA). Alexa FlourTM-488-phalloidin and TAMRA were purchased from Thermo Fisher Scientific (Massachusetts, USA). DiR, DiO, DiD, propidium iodide, fluorescein isothiocyanate (FITC), 6-diamidino-2-phinylindolo dihydrochloride (DAPI)m and Cyanine 7 (Cy7) were obtained from Fanbo Biochemical (Beijing, China). AQ4N was purchased from MedChemExpress. Anti-CD9, anti-CD81, anti-ALIX, anti-TSG101, anti-Ki67, anti-iNOS, anti-CD163, anti-TMEM119, anti-GADPH, and anti-ZO-1 were purchased from Abcam (Cambridge, England). Anti-F4/80 was purchased from CST. GM-CSF was purchased from eBiosciences. Small EVs Spin Columns were obtained from Invitrogen Co. (California, USA). PEG-PLGA NPs was synthesized in our laboratory as previously described. Other chemicals were purchased from J&K (Beijing, China).
Cell culture
The mouse brain endothelial cells (bEnd.3) were supplied by the American Type Culture Collection (ATCC). The luciferase-transfected glioblastoma cell lines (U87MG-luc, GL261-luc, and G422-luc) were maintained in our laboratory (Shenzhen Second People’s Hospital, Shenzhen, China). Both types of cells were cultured in DMEM medium (Gibco BRL) containing 10% fetal calf serum (Gibco BRL), streptomycin (100 μg/mL, Sigma-Aldrich) and penicillin (100 U/mL, Sigma-Aldrich) at 37°C with 5% CO2, and all the cells were passaged at approximately 80% confluency.
Animals
Female BALB/c nude mice (6–8 weeks, 18–22 g), female Kunming mice (4–6 weeks), and male C57BL/6 mice (6–8 weeks, 18–22 g) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China), and all animals were kept in IVC mouse cages with standard conditions and free access to food and water. The animal protocol was approved by the Institutional Animal Care and Use Committees at the Institute of Process Engineering, Chinese Academy of Sciences (approval ID: IPEAECA2021103). All animal experiments was performed in accordance with the Guide for the Care and Use of Laboratory Animals (China, GB/T 35892-2018).
Patient samples, isolation of peripheral blood mononuclear cells (PBMCs), and construction of PDX model
Blood and brain tumor specimens were obtained from patients with informed consent and were reviewed by the pathologist and surgeon. Pathologist classified the type and grade of the tumors in accordance with the WHO histological grading of central nervous system tumors. Sixty-four cases of gliomas were selected from Shenzhen Second People’s Hospital. For IHC, tumor-adjacent tissue was taken as a control. We confirmed tumor-adjacent tissue through histopathology. Characteristics of glioma patients were shown in Supplementary Table S1. Healthy volunteers were recruited through the protocol at the Shenzhen Second People’s Hospital. Peripheral blood samples were acquired from healthy volunteers (age 25–30, male).
Isolation of PBMC: human whole blood was collected in heparin tubes and PBMCs were separated by Ficoll-hypaque density gradient centrifugation (BD, SanDiego, CA). Then PBMC were separated to CD14+ monocytes with corresponding magnetic beads. The CD14+ monocytes differentiated to macrophage under 50 ng/mL GM-CSF treatment and further polarized to M1-like macrophage under LPS (1 µg/mL) treatment.
Construction of PDX model:49 transport tumor sample from pathology to laboratory in HBSS at room temperature and record relevant patient information (e.g., age, sex, etc) (Supplementary Table S2). Then the tumor sample was subcutaneous transplanted into the axilla of the nude mice (female, 6–8 weeks old). After engraftment for three passages, tumor tissue was dissociated into single cells by treatment with trypsin-EDTA digestion. The cells (105 cells) were stereotactically injected into the brain parenchyma at a depth of 3 mm.50 At 2 weeks after the injection of the tumor cells, thin sections of the mouse brain (4 μm) were processed for H&E staining. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional committee. The study was approved by Ethics Committee of Shenzhen Second People’s Hospital Clinical trials (20200727004-FS01).
Orthotopic transplantation of glioma cells
After deep anesthesia, C57BL/6 mice (or Kunming) were positioned in a stereotactic frame (RWD, Shenzhen, China). Punch a small hole with a 25-gauge needle behind right bregma and 2.5–3 mm away from the midline. Then, the GL261-luc cells (or G422-luc) (105 cells) were stereotactically injected into the brain parenchyma at a depth of 3 mm.
Immunofluorescence assay and Immunohistochemistry
Four-μm-thick tissue sections (human and mice) were de-waxed in rehydrated through graded alcohols. Antigen retrieval was carried out using Dako PT link (Dako/Agilent Technologies, Santa Clara, CA). IHC staining of individual markers iNOS, CD163 or Ki67 was performed using EnVision™ G | 2 Doublestain System, rabbit/mouse (DAB/Permanent Red) kit (Dako/Agilent Technologies, Santa Clara, CA), according to the manufacturer’s instructions. IF staining of iNOS (or CD163) and TMEM119 was performed according to the manufacturer’s instructions. Slices were imaged using Vectra II Polaris Automated Quantitative Pathology Imaging System, and images were analyzed with inForm 2.4.
Database
cBioportal (URL: http://www.cbioportal.org/) was used to assess the dataset of LGG and HGG from TCGA.51 Freely accessible server, oncolnc tool (URL: www.oncol.nc.org) was used for analyzing the survival correlation of selected M1 and M2 macrophage marker.52 OncoLnc tool generated the Kaplan-Meier plots for the studied genes using the low and high-expressing M2/M1 ratio that are publically available in TCGA database.
Isolation and extraction of M0EVs, M1EVs, M2EVs, and EMVs
C57/BL6 mice were intraperitoneal injected with 6% starch broth to induce inflammatory responses and elicit large numbers of macrophages. We harvested starch broth-elicited peritoneal cells and cultured them in dishes (106 cells per pore of 12 microwell plate). Pure adherent macrophages would be separated from other types of cells in the peritoneal cavity. Macrophages were treated with or without 1 μg/mL LPS. Cell cultures were EV-depleted media prepared by ultracentrifugation of FBS for 3 h at 200,000 g. After 48 h treatments, cell culture supernatant was collected. EVs were prepared according to a typical protocol.53,54 Briefly, culture supernatant was centrifuged at 300 × g or 2000 × g for 10 min to remove cells and cell fragments, respectively. Then the obtained supernatant was centrifuged at 10,000 g for 30 min at 4 °C to remove debris. The final supernatant was then ultracentrifuged at 100,000 × g for 70 min twice to obtain a pellet containing M1EVs (mouse). The same method was used to obtain human PBMC M1EVs.
The whole blood of C57 mice was harvested by retro-orbital puncture and collected in heparinized mouse blood collection tubes. The blood was centrifuged at 3000 rpm, and the red blood cells were placed in a 3-fold amount of precooled isotonic phosphate buffer (pH 7.4). Then centrifugation at 5000 rpm×15 min, added 10 mmol/L low permeability Tris-HCl buffer solution, and placed in 4°C refrigerator for 1–2 h, 4°C 15 min at 9000 rpm. Red corpuscles (100 nm) were prepared by mini extruder (EMVs).55
In vivo and ex vivo imaging
M1EVs and EMVs were fluorescently labelled by incubation with 1 µM DiR solution, and excess dye was removed by either centrifugation or ultrafiltration for three times. The orthotopic glioma model using luc-U87MG cells were established in the study. The mice were anesthetized by intraperitoneal injection of pentobarbital sodium (1%). Then, cells (105 cells) were inoculated into the right striatum (2.5 mm from the midline, 3.0 mm anterior to the bregma, and 3.0 mm deep) of nude mice using a stereotactic fixation device with mouse adaptor (RWD Life Science, Shenzhen, China). Then the scalp was closed with a clip. In vivo fluorescence imaging was initially used to evaluate the biodistribution and targeting efficacy of nanoplatform on orthotopic GBM models. After seven days of feeding, the orthotopic GBM-bearing mice were i.v. injected with M1EVs, EMVs, and PEG NPs (labelled with DiR). The mice were observed at indicated time points (1, 3, 6, 9, 12, and 48 h) after injection using an IVIS imaging system (PerkinElmer, USA). After 48 h, the mice in the three groups were sacrificed, and the brain, heart, liver, spleen, lung, kidney, and intestine were taken for observation and analysis. Ex vivo images as described above were also recorded. Frozen sections of tumor at 48 h were prepared, and detected by automatic multispectral imaging system (PerkinElmer Vectra II) after DAPI staining.
The aggregation behavior of particles in vivo was also tested by a two-photon laser confocal scanning fluorescence microscope (labelled with DiO or FITC, tested by Leica TCS SP8, Germany). To fashion a cranial window, the skull was thinned away using a sterile stainless steel 2 mm diameter cylindrical drill bit attached to a high-speed hand drill until the underlying dura mater was exposed.
Synthesis of CCA-M1EVs
M1 macrophages were treated with LPS (1 μg/mL) and AQ4N (100 μM), after 48 h treatment, macrophages and cell culture supernatant were collected. At the same time, A-M1EVs were isolated from macrophages supernatant by ultracentrifugation. Then, we added 100 µg CPPO (dissolved in 5 µl THF) and 120 µg Ce6 (dissolved in 5 µl DMSO) in 200 µg A-M1EVs (dispersed in 1 ml PBS). After 1 h incubation, the resulting CCA-M1EVs were washed with PBS for three times to remove the free CPPO and Ce6. The unloaded drugs were removed by elution with a 100 kDa ultrafiltration tube (Merck Millipore Co., Darmstadt, Germany).
Characterization of CCA-M1EVs
Transmission electron microscope (TEM) samples of EVs were prepared according to a typical protocol,56 and imaged by the HITACHI HT7700 TEM. The expression of CD9/CD81/ALIX/TSG101 (EV marker), iNOS (M1 marker) and F4/80 (macrophages marker) in M1 macrophage and M1EVs were analyzed by ProteinSimple® WesTM capillary western blot analyzer (PS-MK15; ProteinSimple, USA). Briefly, total protein of EVs was quantified using the BCA assay kit. EVs extracted from same amounts of cells were diluted (1:2) with sample buffer (ProteinSimple) and the quantification was performed using a 12–230 kDa 25-lane plate (PS-MK15, ProteinSimple) in WES according to the manufacturer’s instructions. CLSM and flow cytometry were used to investigate the colocalization of AQ4N, Ce6, and DiO- labeled M1EVs. Owing to AQ4N and Ce6 share the same spectrum, Ce6 was replaced with TAMRA in the flow cytometry analysis. Drug loading efficiency of AQ4N/Ce6 on CCA-M1EVs was detected by the microplate reader (AQ4N A610 nm, ε=22.5 mL/mg/cm; Ce6 A404 nm, ε=161 mL/mg/cm) (Tecan Infinite M200). Drug loading efficiency of CPPO on CCA-M1EVs was detected by HPLC (Agilent, USA). An Agilent-C18 column (5 µm particles, 4.6×250 mm) was used, and acetonitrile was used as the mobile phase at a flow rate of 1 mL/min. The UV absorbance was determined at 220 nm, and the column temperature was 25 °C.
The particle size and zeta potential of M1EVs before and after drug loading was measured by nanoparticle tracking analysis (NTA; Zetaview, Particle Metrix) at 25 °C. The CCA-M1EVs were dispersed in water or PBS, and then, the zeta potential and diameter were measured every day over the following one week by NTA. The ability of Ce6, CPPO/Ce6, CC-M1EVs, and CCA-M1EVs to generate chemiexcited ROS was evaluated using ABDA as an indicator. The measurement of ROS production of different formulations by microplate assay, followed by addition of different concentrations of H2O2. In the presence of H2O2, drug release behavior of AQ4N from CCA-M1EVs was detected by the microplate reader. The measurement of O2 consumation of different formulations by multiparameter analyzer (Mettler Toledo, Shanghai, China). The measurement of AQ4/AQ4N ratio of different formulations under hypoxia conditions measured by HPLC. An Agilent-C18 column (5 µm particles, 4.6×250 mm) was used with mobile phase of acetonitrile-ammonium formate buffer (0.05 M) (22:78, v/v), with final pH adjusted to 3.6 with formic acid. The UV absorbance was determined at 242 nm, and the column temperature was at 25 °C.
Chemotactic migration across the BBB
The in vitro BBB model was constructed with bEnd.3 cells using a TranswellTM cell culture system. Briefly, bEnd.3 cells (1×104 cells/well) were seeded onto the upper chamber of the TranswellTM pre-coated with gelatine (2% w-v) in 24-well plates, and primary macrophages and GBM tumor cells (U87MG) (1×103 cells/well) were cultured in the lower chamber. After incubation for several days, the integrity of the cell monolayer was examined by measuring the tight junctional protein (ZO-1) using CLSM. Then, DiD-labeled EVs (~50 μg) in fresh culture media was added to the upper chamber. The penetration efficiency was determined by collecting samples from the lower chamber at the time points of 1, 2, 3, 4, 5, 6, and 8 h. The concentration of different formulations in the lower chamber was analyzed based on DiD fluorescence determined in a spectrofluorometer using excitation at 644 nm and emission at 665 nm.
Neutralizing antibodies against CCR2 (CXCR3, or CX3CR1) (Abcam, Cambridge, England) were used in antibody-blocking experiments. Then, DiD-labeled M1EVs (M0EVs) were pre-incubated with 10 µg/ml anti-CCR2 (anti-CXCR3, or anti-CX3CR1) antibodies for 30 min before added to the upper chamber. The penetration efficiency was determined by collecting samples from the lower chamber at 8 h. Based on DiD fluorescence to analyze the concentration of different formulations in the lower chamber.
The lower TranswellTM chamber of M2 (F4/80 + CD163 + )/M1 (F4/80+iNOS + ) ratio was analyzed by flow cytometry. Single-cell suspensisons were subsequently stained with fluorescent antibody. The cells were then washed and analyzed using CytoFLEX LX Flow Cytometry. The H2O2 concentrations in the TranswellTM lower chamber were measured using Hydrogen Peroxide Assay Kit according to the manufacturer’s instructions. The U87MG/MΦ cells were incubated with anti-CD11b magnetic beads and U87MG cells were obtained using the MACS cell sorting system protocol (Miltenyi Biotec., Germany). The intracellular ROS level of U87MG cells was detected by flow cytometry using the fluorescent probe DCFH-DA. The ROS were reacted with DCFH-DA for 20 min at 37 °C. Non-fluorescent DCFH can be converted to fluorescent DCF by ROS oxidation. The cellular ROS oxidized DCF can be used as indicator for ROS production (ex 488, em 510–555 nm). Cell apoptosis of CCA-M1EVs to U87MG cells under hypoxia conditions measured by flow cytometry using the Annexin V FITC/PI kit, according to the manufacturer’s instructions. Cell cytotoxicity of CCA-M1EVs to U87MG cells was measured by CCK8 (Beyotime Co., Shanghai, China). After 24 h incubation, CCK-8 solution was added and incubated for another 4 h. Percent viability was normalized according to the untreated cells.
Penetration and growth inhibiton of MCTSs in vitro
Tumor spheroids of U87MG and macrophage cells (3:1) were prepared using the liquid overlay methods. To evaluate drug penetration in MCTS, cell spheroids were incubated with M1EVs for 24 h, and then analyzed by CLSM. The MCTS were incubated with different treatments, including PBS, M1EVs, CC-M1EVs, A-M1EVs, and CCA-M1EVs (~50 μg). At the end of the culture, each supernatant was collected and cytokines content was measured by Luminex multiplex cytokine analysis platform. The M2/M1 ratio was analyzed by flow cytometry. The U87MG/MΦ cells were incubated with anti-CD11b magnetic beads and U87MG cells were obtained using the MACS cell sorting system protocol. The intracellular ROS level of U87MG cells was detected by flow cytometry using the DCFH-DA. Under hypoxic conditions, the apoptosis of U87MG cells was detected by Annexin V-FITC and PI.
To estimate the growth inhibition effect on multicellular tumor spheroids, growth inhibition of the tumor spheroids was monitored using an inverted phase microscope. The major (rmax) and minor (rmin) radii of each treated MCTS were determined, and the spheroid volume was calculated according to equation 1:
$${{{mathrm{V}}}} = left( {{{{mathrm{3}}}}/{{{mathrm{4}}}}} right) times pi times left( {{{{mathrm{r}}}}_{{{{mathrm{max}}}}}/{{{mathrm{2}}}} + {{{mathrm{r}}}}_{{{{mathrm{min}}}}}/{{{mathrm{2}}}}} right)^{{{mathrm{3}}}}$$
(1)
Therapeutic effect and toxicity in vivo
For investigating antitumor effect of different treatments, U87MG-derived tumor xenografts were generated as described above, and the growth of orthotopic GBM could be monitored by the assistance of bioluminescence imaging. The glioma-bearing mice were given i.p. injections of a D-luciferin potassium solution (3 mg/mouse). Photons emitted from the glioblastoma region were collected and quantified by using the living image software. All images were treated with the same conditions and color scale. U87MG glioma-bearing mice were randomly divided into five groups (for each group, n = 8): PBS, M1EVs, CC-M1EVs, A-M1EVs, and CCA-M1EVs (~200 μg). Then, 100 μL of different formulations were intravenously injected every three days for 5 times after 7 days. Mice were regularly measured for any signs of deterioration or weight loss, and the body-weight and survival time of each mouse were recorded daily during the whole period of treatment.
To estimate the ROS production in tumors in situ, DCFH-DA was intratumorally administered and further the tumor was imaged by two-photon confocal scanning microscopy. Hypoxia tendency in vivo in tumors was measured by PA imaging. Specifically, the tumor oxygenation status was detected by the ratios of oxygenated hemoglobin (λ = 850 nm) and deoxygenated hemoglobin (λ = 750 nm) after i.v. injection of different EV formulations. At predetermined time points, animals were sacrificed and tumors were harvested. Then, the tumors were embedded in OCT medium. Tumor slices were stained by M2 (green), M1 (red), and Ki67 markers, respectively slides were scanned with automatic multispectral imaging system (PE Vectra II) and images were analyzed with inForm 2.4. To further evaluate the safety of different formulations in vivo, body-weight changes were recorded.
Photoacoustic imaging and magnetic resonance in the PDX model
The orthotopic PDX-bearing mice were established as described above, and the tumor was determined using an in vivo imaging system (BioSpec 70/20 USR, Bruker, Germany). Then, the PDX mice were individually i.v. injected with CC-M1EVs and CCA-M1EVs (n = 5 per group) on days 7, 10, 13, 16, and 19. At 24 h post-injection, PA imaging was used to monitor the biodistribution of different formulations in the brain by Vevo LAZER (VisualSonics FujiFilm, Canada) due to the photoacoustic performance of Ce6.
Tumor progression was assessed by MRI at 7 day and 20 day. MRI imaging T1-weighted Gd contrast-enhanced (T1Gd) image regions allowed approximate delineation of tumor. The body weight and survival were measured every 3 days.
Statistics
All statistical calculations were conducted using GraphPad Prism 9.0.0. Data presentation, sample size, and probability values were indicated in figure legends. For comparison between groups, statistical significance was done using unpaired two-tailed Student’s t-test. One-way ANOVA with a Tukey post hoc test (or Kruskal-Wallis test) was used for multiple-group comparisons. Survival analysis was calculated by two-sided Log-rank Mantel-Cox tests. P values < 0.05 were considered as significant, and ns indicated no significant difference.
