# Nanotechnological engineering of extracellular vesicles for the development of actively targeted hybrid nanodevices – Cell & Bioscience

May 14, 2022

### Cell cultures

The three cell lines used were all cultured in conformity with the sterile technique and the standard mammalian cell culture protocols under a 5% CO2 atmosphere at 37 °C.

Lymphocyte cell line (IST-EBV-TW6B) was purchased from the cell bank IRCCS AOU San Martino IST (Italy). Cells were cultured in advanced RPMI 1640 culture medium (Gibco) with 20% of heat inactivated fetal bovine serum (FBS, Gibco), 1% penicillin/streptomycin (P/S, Sigma) and 1% of l-glutamine 200 mM (Lonza) in 75 cm2 not treated cell culture flasks (Corning) maintaining the cell density between 9 × 104–5 cells/mL.

Daudi cells (ATCC® CCL­213™), originating from a Burkitt’s lymphoma patient, were obtained from American Type Culture Collection (ATCC). Cells were grown in RPMI 1640 culture medium (ATCC) supplemented with 10% of heat inactivated FBS (ATCC), 1% P/S (Sigma) in 75 cm2 not treated cell culture flasks (Corning) with a cell density between 3 × 105–6 cells/mL.

HL60 cells (ATCC® CCL-240™), from an acute myeloid leukemia patient, were purchased from ATCC. They were maintained in Iscove’s Modified Dulbecco’s Medium (Sigma) with 20% heat inactivated FBS (Sigma), 1% Glutamine (Sigma), 1% P/S (Sigma) in 75 cm2 not treated cell culture flasks (Corning), adjusting cell density to 1 × 105–6 cells/mL.

### EVs isolation and characterization

EVs were isolated from the conditioned media of the lymphocytes cell line grown in RPMI supplemented with 20% EVs-depleted FBS, 1% glutamine and 1% P/S after 72 h of culture. The depleted FBS was the supernatant collected from the overnight ultracentrifugation at 100,000×g (Optima Max-XP Ultracentrifuge with MLA-50 rotor, Beckman Coulter) at 4 °C of FBS.

EVs were produced by plating 1.5 × 105 lymphocytes/mL in a total volume of 200 mL of medium complemented with depleted FBS in 75 cm2 untreated flasks and left in culture for 3 days at 37 °C under a 5% CO2 atmosphere. Just before the EVs extraction, lymphocytes viability was assessed via Trypan-blue (VWR) method using a TC20 TM automated cell counter (BiO-Rad Laboratories), and only samples with viability over 90% were processed to reduce the probability of apoptotic bodies’ recovery. The EVs isolation protocol is adapted from the sterile differential ultracentrifugation protocol described by Thery et al. [67]. Cell culture medium was collected in 50 mL tubes and centrifuged 10 min at 150×g at 4 °C to remove cells. Supernatants were collected and centrifuged 20 min at 2000×g at 4 °C to remove dead cells and cell debris. The supernatants collected were centrifuged again for 30 min at 10,000×g at 4 °C to discard aggregates of biopolymers, apoptotic bodies, and other structures with higher density than EVs. Supernatants were collected again, placed in ultracentrifuge polypropylene tubes (32 mL Optiseal tubes, Beckman Coulter) and ultracentrifuged at 100,000×g for 70 min at 4 °C. The obtained pellet was resuspended in sterile, cold, 0.1 μm filtered phosphate buffered saline (PBS) solution and ultracentrifuged for further 60 min at 100,000×g at 4 °C. The pellet, which contained EVs, was resuspended in 600 μL of sterile, cold, 0.1 μm filtered physiological solution (0.9% NaCl, NovaSelect), aliquoted in 50 μL cryovials and stored at − 80 °C for further uses.

The concentration and the size distribution of collected EVs were measured by nanoparticle tracking analysis (NTA) technique with a NanoSight NS300 (Malvern Panalytical) equipped with λ = 505 nm laser beam and a NanoSight syringe pump. Samples were diluted in a final volume of 500 μL of 0.1 μm-filtered physiological solution to meet the ideal particles per frame value (20–100 particles/frame). Different EVs aliquot were measured by capturing three videos of 60 s with an infusion rate of 50 a.u, and a camera level value between 14 and 16. The collected videos were then analyzed by the NTA 3.4 software (Malvern Panalytical), setting the detection threshold at 5.

The protein content of isolated EVs was measured by Bradford assay as described in literature [67]. Bradford reagent (Bio-Rad) was diluted 1:5 in bd water and added to EVs samples, diluted 1:2 in 0.1 μm-filtered PBS, and serially diluted bovine serum albumin (BSA, Sigma Aldrich) standards with known concentrations. The absorbance at 590 nm was then recorded using a microplate spectrophotometer (Multiskan GO, Thermo Fisher Scientific) and the protein concentration of EVs samples was extrapolated comparing their absorbance values with the calibration curve made on BSA standards. All samples were analyzed in triplicate.

The EVs morphology was analyzed trough Transmission Electron Microscopy (TEM) using a JEOL JEM-1400Plus TEM, with thermionic source (LaB6), operated at 120 kV. For TEM analyses, a drop of the sample solution was placed on a copper grid, 150 mesh, coated with amorphous carbon film; then, to highlight the EVs morphology, the EVs were stained before observation with a solution of 1% uranyl acetate in water. Energy Dispersive X-ray Spectroscopy was performed with the same instrument equipped with a JEOL-JED-2300 Energy Dispersive Spectroscopy (EDS) silicon drift type detector (area 30 mm2).

### ZnO NCs synthesis, functionalization and characterization

ZnO NCs were synthesized through a microwave-assisted solvothermal approach [40]. In details, the solution of zinc precursor, i.e., zinc acetate dihydrate (99.99% Sigma Aldrich) 0.1 M in methanol, was stirred directly in the microwave-reactor vessel. 0.48 mL of bidistilled (bd) water were added to initiate the nucleation and then a KOH solution (≥ 85% pellets, Sigma-Aldrich, 0.2 M in methanol) was rapidly added. The resulting solution, with an overall pH of 8, was put in the microwave oven (Milestone START-Synth, Milestone Inc) at 60 °C for 30 min, under temperature and pressure control and with a maximum microwave power of 150 W. Upon completion of the reaction, the obtained colloidal suspension was cooled down to room temperature (RT) and centrifuged (3500×g for 10 min) to remove the unreacted compounds. The as-obtained pellet was dispersed in fresh ethanol through sonication and the washing step was repeated two more times.

The ZnO NCs were further functionalized with amino groups (–NH2) [40]. Briefly, the synthesized ZnO NCs dispersed in ethanol were heated to 70 °C in a round glass flask under continuous stirring and nitrogen gas flow. After approximately 15 min, the functionalizing agent (3-amminopropyltrimethoxysilane, APTMS, 97% Sigma Aldrich), was added in a molar ratio of 10 mol% with respect to the total ZnO amount. The reaction was carried out in reflux condition under nitrogen atmosphere for 6 h and then washed twice, in to order to remove unbound APTMS molecules, by centrifuging (10,000×g for 5 min).

The morphology of ZnO NCs was evaluated through Transmission Electron Microscopy using a Jeol JEM-1011 transmission electron microscope operated at 100 kV of acceleration voltage. The crystalline structure of ZnO NCs was investigated by X-ray diffraction (XRD) measurements using a Panalytical X’Pert diffractometer (Malvern Panalytical) in configuration θ–2θ Bragg–Brentano equipped with a Cu-Kα radiation source operating at 40 kV and 30 mA. The sample in ethanol was deposited drop by drop on a silicon wafer and analyzed collecting the spectrum in the range of 20°–65° with a step size of 0.02° and an acquisition time per step of 100 s. The hydrodynamic size and the z-potential of the ZnO NCs were determined using the dynamic light scattering (DLS) technique with a Zetasizer Nano ZS90 (Malvern Instruments). The measurements were performed at RT on samples with concentration of 100 μg/mL and sonicated for 10 min before the acquisition.

### TNH assembly and characterization

The EVs:ZnO NCs ratio of 1:2 used during TNH assembly was calculated starting from a model which estimates the maximum number of nanocrystals that could be geometrically encapsulated within a single vesicle, as schematically represented in Additional file 1: Fig. S2a. Considering the EVs concentration as part/mL and μg/mL obtained from NTA and Bradford techniques respectively, the maximum theoretical number of ZnO NCs (indicated as n°ZnO NCs), corresponding to a fixed amount of μg of EVs, was calculated as follows:

$$({text{n}}^circ_{{{text{ZnO}};{text{NCs}}}} )_{{upmu {text{g}};{text{EVs}}}} = ({text{n}}^circ_{{{text{ZnO}};{text{NCs}}}} )_{{{text{EV}}}} cdot {text{n}}^circ_{{{text{EVs}};{text{per}};upmu {text{g}}}} cdotupmu {text{g}};{text{EVs}}quad {text{where}}quad {text{n}}^circ_{{{text{EVs}};{text{per}};upmu {text{g}}}} = frac{{{text{Conc}}_{{{text{EVs}}@{text{NTA}} left[ {{text{part}}/{text{mL}}} right]}} }}{{{text{Conc}}_{{{text{EVs}}@{text{Bradford}} left[ {upmu {text{g}}/{text{mL}}} right]}} }}.$$

Finally, the mass of a single particle was calculated considering its volume (ie the volume of a sphere with diameter equal to the ZnO NC diameter, dZnO) and the ZnO density (ρZnO = 5.606 g/cm3) and the obtained value was used to determine the corresponding NCs amount expressed as μg:

$$upmu {text{g}};{text{ZnO}};{text{NCs}} = ({text{n}}^circ_{{{text{ZnO}};{text{NCs}}}} )_{{upmu {text{gEVs}}}} cdot{text{mass}}_{{{text{ZnO}};{text{NC}}}} quad {text{where}}quad {text{mass}}_{{{text{ZnO}};{text{NC}}}} = left( {frac{uppi }{6}{text{d}}_{{{text{ZnO }};{text{NC}}}}^{3} cdot10^{ – 21} } right)cdotrho_{{{text{ZnO}}}} cdot10^{6} .$$

The model was then amended on the basis of experimental observations, as discussed in detail in “Results and discussion” section, and finally an excess of 10 μg of amino-functionalized ZnO NCs were combined with an amount of EVs corresponding to 5 μg of protein measured by Bradford assay. The encapsulation process was performed in a 1:1 (v/v) solution of 0.1 μm-filtered bd water and physiological solution, with a final concentration of 80 μg/mL for ZnO NCs and 40 μg/mL for EVs. As schematically represented in Additional file 1: Fig. S2b, opportunely labeled EVs dispersed in physiological solution were rapidly frozen in liquid nitrogen for 3 min and thawed at RT for 15 min. The freeze–thaw cycle was repeated twice and then the corresponding amount of ZnO NCs in bd water was added. The obtained mixture was incubated under continuous agitation (250 rpm) at 45 °C for 10 min, at 37 °C for 2 h and then overnight (O/N) at RT. In order to redisperse the obtained TNHs in media suitable for in vitro tests, a final step of centrifugation was performed. The samples were centrifuged at 5000×g for 5 min, suspended in the cell culture medium and redispersed by vortexing for 3 min.

The coupling efficiency was evaluated through fluorescence microscopy. The amino-functionalized ZnO NCs were labeled with Atto 550-NHS ester (λEx = 554 nm, ATTO-Tech), by adding 4 μg dye each mg of ZnO NCs suspension in ethanol; the solution was stirred in dark O/N and then washed twice. EVs, diluted 1:2 in physiological solution were labeled with Wheat Germ Agglutinin (WGA) conjugated with Alexa Fluor 488 (WGA488, λEx = 495 nm, Thermo Fisher) by adding 1 μL of dye (100 μg/mL in PBS) for each EVs aliquot containing approximately 1 × 1010 particles. The solution was kept under agitation (180 rpm) in dark at 37 °C for 30 min and then purified from unbound dye molecules with 50 kDa Amicon Ultra 0.5 centrifugal filter (Merck Millipore). The samples were analyzed using a wide-field fluorescence-inverted microscope (Eclipse Ti-E, Nikon) equipped with a super bright wide-spectrum source (Shutter Lambda XL), a high-resolution camera (Zyla 4.2 Plus, 4098 × 3264 pixels, Andor Technology) and an immersion oil 100× objective (Nikon). The collected images were analyzed with the colocalization tool of NIS-Element software (NIS-Element AR 4.5, Nikon). In brief, the spots in red and green channels (corresponding to ZnO NCs and EVs respectively) were counted and then a merge of the two images was performed, counting the spots in which the two fluorescence signals resulted superimposed. The percentage of colocalization with respect to the ZnO NCs (%co-ZnO) was then calculated doing the ratio between the number of colocalized spots and the total number of red spots. The analysis was performed on 9 regions of interest (ROIs) to evaluate the mean %co-ZnO and the results of 5 different samples were averaged to obtain the coupling efficiency at the end of TNHs assembly process. The same colocalization procedure was used to evaluate the maintenance of coupling efficiency after incubation in cell culture medium samples (advanced RPMI + 20% EVs-depleted FBS, Gibco). Four different samples were analyzed at t0 and after 1, 24 or 48 h of incubation and the results were expressed as percentage decrease of %co-ZnO with respect to t0.

To evaluate the TNHs morphology, a drop of the sample solution was placed on a copper grid, 150 mesh, coated with amorphous carbon film and the sample was stained with a solution of 1% uranyl acetate in water for 30 s. TEM analysis in Bright Field mode were performed using a JEOL 1011 operated at 100 kV. Annular Dark-Field (ADF) imaging in Scanning Transmission Electron Microscopy (STEM) mode and EDS analysis were performed using a TEM JEM-1400 Plus, with thermionic source, operated at 120 kV of accelerating voltage and equipped with a JEOL-JED-2300 EDS silicon drift type detector (detector area 30 mm2).

The size distribution of TNHs in bd water and physiological solution 1:1 (v/v) was assessed by NanoSight NS300 equipped with NanoSight syringe pump. The samples were diluted 1:5 and three videos of 60 s were recorder with camera level and detection threshold of 16 and 5, respectively. The size distribution of TNHs in advanced RPMI (Gibco) supplemented with 20% EVs-depleted FBS (Gibco) was measured in static conditions using the O-ring top plate cell with manual syringe connection. TNHs resuspended in cell culture medium (ZnO concentration 100 µg/mL) and medium alone as reference were diluted 1:10 in bd water and physiological solution 1:1 (v/v) and three videos of 30 s were acquired, advancing manually the samples between them. The camera level and detection threshold values were set at 15 and 6, respectively. At least two independent experiments were performed.

TNHs hemocompatibility was preliminarily evaluated through a simple turbidimetric assay as previously reported [47], using Na-citrate human recovered plasma (Zen Bio) and calcium chloride (CaCl2 0.025 M from HYPHEN BioMed) as clotting agent. Briefly, 75 µL of plasma for each sample were aliquoted in a 96 well plate and mixed with 75 µL of TNHs samples at concentration 75 µg/mL in bd water and physiological solution 1:1 (v/v). To monitor the dispersant influence, controls with the addition of 75 µL of physiological solution or bd water and physiological solution 1:1 (v/v) were also performed. Coagulation was started adding 75 µL of CaCl2, and the absorbance at 405 nm was measured through a microplate UV–VIS spectrophotometer every 30 s for 45 min at constant T = 37 °C. Three replicates per sample were averaged to obtain the mean absorbance at each time point and the coagulation time, was calculated as the time corresponding to the half maximal absorbance (t1/2). Four independent experiments were conducted and the results were expressed as mean ± S.E.

CD20 expression on TNH surface was evaluated by flow cytometry as described in “EVs isolation and characterization” section. Each experiment was repeated three times (n = 3).

### TNH functionalization with targeting antibodies

To obtain TNHCD20 samples, vesicle membranes were functionalized with anti-CD20 antibody as previously reported in [25] and schematically represented in Additional file 1: Fig. S2c. After the O/N co-incubation step, TNH samples corresponding to 5 μg of EVs proteins were mixed with functionalizing antibodies, added in three consecutive incubation steps (1 h at RT on a tube rotator with fixed speed of 20 min−1 each). In details, half of the EVs protein content measured by Bradford assay was considered equal to CD20 antigen and anti-CD20 antibody (Rituximab, Cat. n° TAB-016, Creative Biolabs, 5 mg/mL in PBS) was added in a molar ratio 4:1 with respect to the assumed antigen concentration, working in a large excess to favor antibody-antigen interaction. Then, anti-human secondary antibody (AffiniPure F(ab′)2 Fragment Goat Anti-Human IgG, Fcγ fragment specific, Jackson ImmunoResearch or AMCA AffiniPure F(ab′)2 Fragment Goat Anti-Human IgG, Fcγ fragment specific, Jackson ImmunoResearch, λEx = 450 nm) was added as cross-linker in molar ratio secondary Ab: anti-CD20 = 1:1. Finally, a second aliquot of anti-CD20 antibody was added in the same amount as the first incubation step. After the third hour of incubation, TNHCD20 samples were collected and centrifuged (5000×g for 5 min) to remove unbound antibodies and resuspended in cell culture medium suitable for in vitro tests.

### Cytotoxicity assay of TNH and TNHCD20

To evaluate the viability of lymphocytes, Daudi and HL60 cell lines treated with 5 μg/mL of TNH and TNHCD20 (considering the EVs protein content), after the centrifugation step at 5000×g, the two TNH samples were resuspended in the required volume of cell culture medium. Then, 2 × 105 cells for each mL of treatment were centrifuged at 130×g for 5 min for Daudi and HL60 and at 150×g for 5 min for lymphocytes, and the supernatants replaced with the treatment solutions of TNH and TNHCD20. A total volume of 100 μL was plated for each well in a 96-well flat-bottom plastic culture plate (Greiner Bio-one, 96 Well for suspension culture). After 20 and 44 h of incubation 10 μL of WST-1 reagent (CELLPRO-RO Roche) was added to each well and, after further 4 h of incubation, the formazan absorbance was detected at 450 nm by the microplate spectrophotometer using a 620-nm reference. All the experiments were carried out at least four times for each cell line and results were normalized to the untreated control.

### Cytofluorimetric analysis of TNH and TNHCD20 internalization

For the uptake evaluation of TNH and TNHCD20, the amino-functionalized ZnO NCs were labeled with Atto 647-NHS ester (λEx = 647 nm, ATTO-Tech) fluorescent probe as previously described, and the preparation of the TNH performed as described above. After the centrifugation steps, the two TNHs were resuspended in cell medium. 2 × 105 lymphocytes, Daudi and HL60 cells for each mL of treatment were centrifuged and the pellets were resuspended in the TNHs’ solutions. The experiment was carried out five times for TNH and in duplicate for TNHCD20. Data from untreated cells were used as reference.

Cells were cultured into not treated 96 well plates, 250 μl for each well. After 24 and 48 h of incubation, the contents of the different wells were collected and washed twice in PBS and resuspended in 350 μL of PBS for the 24 h and 500 μL for the 48 h cytofluorimetric analysis. 1 × 104 events were acquired with the flow cytometer with 0.59 μL/s flow rate, excluding cell debris. The analyses were performed using the red laser (λEx = 642 nm). Positive events were characterized by a shift of Red-R fluorescence intensity (emission filter 661/15) and the percentages of positive events were evaluated with respect to untreated cells using Guava InCyte Software (Merck Millipore).

### Fluorescence microscopy imaging of TNH and TNHCD20 internalization

For the fluorescence microscopy analysis, EVs were labelled with Wheat Germ Agglutinin (WGA) conjugated with Alexa Fluor 647 (WGA647, λEx = 650 nm, Thermo Fisher), ZnO NCs with Atto 550-NHS ester (λEx = 554 nm, ATTO-Tech), and the TNHCD20 nanoconstruct was assembled using the AMCA AffiniPure F(ab′)2 Fragment Goat Anti-Human IgG, Fcγ fragment specific as secondary antibody.

Samples were treated with the same protocol used for the cytofluorimetric analysis and plated in a volume of 100 μL. After 24 and 48 h of culturing at 37 °C, 5% CO2 in 96 well plates, the content of each well was collected, centrifuged, resuspended in 40 μL of the correspondent medium. The 40 μL drop was spotted in a 8-well chamber slide (Thermo Scientific™ Nunc™ Lab-Tek™ II CC2™ Chamber Slide System) and placed at 37 °C, 5% CO2 for 30 min to allow the attachment of the cells. After that, cells were fixed using 250 μL of Image-iT™ Fixative Solution (4% formaldehyde, methanol-free, Thermo Scientific) for 10 min, washed in PBS and cells’ membranes were labelled by incubating cells with 1.25 μL of WGA conjugated with Alexa Fluor 488 (WGA488, λex = 495 nm, Thermo Fisher) for 10 min and washed two other times in PBS. Images were acquired using a wide-field fluorescence-inverted microscope using an immersion oil 100× objective.

### Shock waves treatment

The cytotoxic effects of shock waves combined with both TNH and TNHCD20 treatments were evaluated on Daudi and lymphocytes cell lines.

After the centrifugation steps, the TNH and TNHCD20 were resuspended in cell culture medium. 2 × 105 lymphocytes and Daudi for each mL of treatment were centrifuged and the pellets were resuspended in the TNH and in the TNHCD20 solutions and seeded into 96 well plates, 100 µL for each well. Untreated cells were used as reference. After 24 h of incubation cells were treated with multiple SW (3 times/day, one treatment every 3 h). Each SW treatment was composed by 250 shots of 12.5 MPa, 4 shot/s, SW were generated by PW2 device from Richard Wolf. The cell viability was measured 24 h after the SW treatment with the WST-1 assay.

### Statistical analysis

Plotted data are mean ± S.E. The statistical analysis between the treatment groups was performed by using the two or the three-way analysis of variance (ANOVA) tools of the SIGMA Plot software’s data analysis package. **p < 0.001 and *p < 0.05 were considered significant. Independent experiments were performed at least two times.