Nanomaterial synthesis and chemical functionalization
For the synthesis of Gd-doped, oleate-stabilized Zinc Oxide nanocrystals (Ol-ZnO-Gd NCs), 0.5268 g of zinc acetate dihydrate (Zn(CH3COO)2·2H2O, 2.4 mmol from Sigma-Aldrich), stored in dry conditions under vacuum, were dissolved in 40 ml of ethanol in a 100 ml round-bottom flask together with 0.0963 g of Gadolinium(III) acetate hydrate (Gd(CH3COO)3·H2O, 0.287 mmol from Sigma-Aldrich) to obtain the desired molar ratio of zinc precursor and the dopant, which was equal to 1:0.12. Then 140 µl of oleic acid (≥ 99%, Sigma-Aldrich) was added and the obtained solution was heated at 70°C on a stirring plate (VELP Scientifica ARE Hot plate stirrer) at 350 rpm using a heating silicone oil bath, after closing the flask with a refrigerating column. After complete precursors dissolution (after around 3 minutes), a solution made by 0.522 g of TMAH (Tetramethylammonium hydroxide, 98.5%, Sigma-Aldrich) dissolved in 1.052 ml of bidistilled water (obtained by a Direct Q3 system, Millipore) and 10 ml of ethanol was prepared and added to the flask, turning the colloidal solution from transparent to white. After 15 minutes, the reaction was stopped by placing the flask in an ice bath and rapidly adding 40 ml of ice-cold (0–4°C) ethanol. The nanocrystals suspension was then centrifuged for 5 minutes at 10000 RCF, the reaction supernatant discarded, and 30 ml of fresh ethanol added to redisperse the nanocrystals through a vortex stirrer and an ultrasound bath. Finally, two washing steps with ethanol were performed through the succession of centrifugation and resuspension steps by sonication and vortex.
As a control, pristine, oleate-stabilized Zinc Oxide nanocrystals (Ol-ZnO NCs) were synthesized following the same procedure without the addition of Gd acetate hydrate. In this case, the reaction was carried out for 5 minutes (instead of 15 as in the case of the doped NCs).
The synthetized NCs were also functionalized with amino-propyl groups. The functionalization process was implemented taking the volume corresponding to 40 mg of nanocrystals from the stock and putting it in the round bottom flask, where the amount of ethanol needed to reach a final concentration of 2.5 mg/ml was previously added. Then, the flask was immersed in a silicone oil bath at 70°C, closed with a refrigerating column with sealing grease and nitrogen gas was flowed to strip out humidity. When the condensation in the flask started, 12.5 µl of APTMS (3-Aminopropyl)trimethoxysilane, 0.0716 mmol, Sigma-Aldrich) was added, corresponding to 14.5% with respect to the moles of ZnO NCs. The whole procedure lasted 6 hours and when the final suspension cooled down, it was collected by centrifugation for 10 minutes at 10000 RCF and resuspended in 5 ml of fresh ethanol through sonication. Finally, two washing steps in ethanol were performed.
Field Emission Scanning Electron Microscopy (FESEM) measurements were carried out using a FESEM Merlin (from Zeiss) coupled with X-Ray detector for EDS Analysis. For the samples’ preparation, an ethanol-diluted aliquot of the nanocrystals’ solution was put dropwise on top of a silicon wafer.
A DC magnetometer (Lake Shore 7225, Lake Shore Cryotronics, Inc.) equipped with a cryogen-free magnet system was used to examine the magnetic properties of the Gd-doped NCs compared to the undoped ones. Measurements were performed at room temperature in quasistatic condition for 1 mg of both doped and undoped NCs.
Dynamic Light Scattering (DLS) and Zeta potential measurements (both in ethanol and water) were performed with the Zetasizer Nano ZS90 (Malvern Instruments). The samples were first sonicated for 10 minutes and 100 µg of NCs was used for each measurement. Each sample (final concentration of 100 µg/ml) was then sonicated, vortexed and put in a 1 ml cuvette (DTS0012), and the DLS Size measurements in both ethanol and water were performed. Then, 750 µl of the sample in water was put in another cuvette (DTS1070) for Zeta potential measurements.
Gemcitabine (Gemcitabine hydrochloride, Sigma-Aldrich) uptake analysis was performed with the UV-Vis spectrophotometer (Multiskan FC, Thermo Scientific), coupled with the Thermo Scientific Skan-it Software for data collection.
Gemcitabine calibration curve in water was derived through serial dilutions starting from a 1 mM concentrated drug stock (1 mM, 100 µM, 10 µM, 0.1 µM) and measuring the absorbance values at a wavelength of 272 nm. The next step was to identify the time necessary to obtain an optimal adsorption of the drug on the NCs’ surface. To this end, the volume corresponding to 200 µg of amine-functionalized NCs was put into different 1.5 ml tubes, one for each time step (2 h, 3 h, 4 h, 6 h), and centrifuged at 14000 RCF for 10 minutes. The pellets were resuspended in 400 µl of drug at 1 mM concentration (in bidistilled water) and samples were put on a magnetic stirring plate (200 rpm). Two additional samples were prepared as controls: the drug’s solvent (negative control, bidistilled water without drug), and the 1 mM drug solution (positive control). At each time step, the corresponding sample was centrifuged (14000 RCF, 10 min), and three replicates of 100 µl of the supernatant were put in a synthetic quartz glass (QG)96 wells plate (Hellma Analytics), together with the controls, and analyzed with the UV-Vis spectrophotometer. The absorbance value of the negative control was subtracted to the absorbance values of the other samples. Then, the drug’s concentration in the samples was derived using the equation of the calibration curve, and the concentration of drug adsorbed on the NCs’ surface was obtained by subtracting the molarity of the sample to the one related to the positive control.
Extracellular vesicles and lipid coating
The nanoconstruct coating was obtained using extracellular vesicles (EVs), derived from healthy B lymphocytes (IST-EBV-TW6B purchased from IRCCS AOU San Martino IST), and commercially available lipids to whom the targeting peptide CKAAKN was bound. Liposomes, EVs, and drug-loaded nanocrystals were separately prepared and the NCs’ encapsulation in the lipidic shell was obtained through the freeze-thaw technique in liquid nitrogen.
Targeted liposomes, henceforward called Lipo-pep, were formed by DOPC (1,2-dioleoyl-sn-glycero-3-phosphocholine from Avanti Polar Lipids Inc.) and DSPE-PEG(2000)Maleimide (Avanti Polar Lipids Inc.) coupled with CKAAKN synthetic peptide. First, a lipids’ stock was prepared: the volume corresponding to 2.5 mg of DOPC was dried to evaporate chloroform and then 400 µl of ethanol and 600 µl of bidistilled water were added obtaining a concentration of 2.5 mg/ml. Ethanol was added to prevent the formation of liposomes in the stock already.
Starting from this stock, Lipo-pep were prepared exploiting a solvent-exchange method(49, 60), diluting DOPC lipids 1:10 in bidistilled water. The complex of DSPE-PEG(2000)-maleimide and CKAAKN-peptide (molar ratio 3:1) was obtained dissolving the two components in dimethylsulfoxide (DMF, Sigma-Aldrich) at 37.5 and 50 mM, respectively. The peptide solution was diluted in 0.1 M sodium phosphate buffer (pH 7.4) and DSPE-PEG(2000)-maleimide was then added to the mixture obtaining a final reaction mixture of 1:1 DMF/(sodium phosphate buffer) with 5 mM peptide and 15 mM DSPE-PEG(2000)-maleimide. The reaction was allowed to proceed for 1 h at room temperature. Then, it was kept at -20°C and used as a stock. The so obtained DSPE-PEG(2000)-CKAAKN lipids, henceforward called functional lipids, were added to the already formed DOPC liposomes (Lipo) in the molar ratio of 99.5:0.5 mol% (Lipo:functional lipids). Considering 25 µl of DOPC lipids from the stock (2.5 mg/ml), the corresponding moles of functional lipids were calculated, equal to 0.0004 µmol. Accordingly, the needed volume of functional lipids (whose stock was at a concentration of 17.87 mg/ml) was calculated (0.08 µl).
Once prepared, the Lipo-pep solution was incubated into an orbital shaker (VWR Incubating orbital shaker, Professional 3500) for 8 h at 37°C and 200 rpm in order to stabilize the liposomes’ structure and allow the incorporation of functional lipid in the DOPC lipidic shell. Finally, a dialysis process was performed to remove the unbound peptide using a dialysis membrane (SnakeSkin dialysis Tubing 3.5K MWCO, 16 mm dry I.D., 35 feet by Thermofisher Scientific) against bidistilled water at the ratio of 1 l of water for 1 ml of Lipo-pep solution, at 80 rpm for 18 hours.
EVs were isolated from conditioned cell culture supernatants of B lymphocytes according to an ultracentrifugation protocol, as previously reported(61), and stored at -80°C in aliquots of 50 µl in 0.9% NaCl solution. The freeze-thaw technique, which consists in the alternation of a freezing of the sample in liquid nitrogen (-196°C) and a following thawing at room temperature, was exploited to coat the nanocrystal with the lipidic shell in a ratio of 1:0.125:0.5 (µg of protein content of EVs, µg of Lipo-pep, and µg of nanocrystals respectively).
Briefly, the Lipo-pep stock was sonicated for 1 minute and vortexed, before putting the required volume into a cryovial. Then, the pellet of the previously Gemcitabine-loaded NCs (stored at − 20°C overnight after the drug uptake) was redispersed in bidistilled water (concentration 10 µg/µl) through vigorous mixing, and the volume needed to obtain the desired ratio was added in the vial. The desired amount of thawed EVs (measured from the µg of proteins quantified by Bradford Assay) was concentrated through a 50 kDa Amicon filter (Sigma-Aldrich), at 14000 RCF for 10 minutes to remove the solution in excess. The final volumes of EVs in NaCl solution and drug-loaded NCs + Lipo-pep in bidistilled water was then adjusted to be in a ratio of 1:1 before mixing.
Then, 6 freeze-thaw cycles (3 minutes of freezing in liquid nitrogen followed by a thawing of 15 minutes) were performed, followed by 1 h of incubation in the orbital shaker (37°C, 200 rpm). At the end of the process, schematized in Fig. 1, the obtained final nanoconstruct was named as EV-Lipo-pep-NCs-Gem. The same nanoconstruct without loaded drug on the NCs surface (EV-Lipo-pep-NCs) was developed to characterize the preparation process and tested in terms of cellular uptake and cells viability, as control, to evaluate the nanoconstruct safety in absence of Gemcitabine.
DLS Size and Zeta potential measurements were performed as above on the nanoconstruct, as well as Nanoparticle Tracking Analysis (NTA), using the NanoSight NS300 from Malvern Panalytical. 50 µl of the nanoconstruct suspension was diluted 1:10 in a 1:1 solution of 0.1 µm filtered bidistilled water and 0.1 µm filtered 0.9% NaCl solution and filled in the instrument microfluidic chamber. The detection was carried out by a λ = 505 nm laser beam and 20x magnification objective. Three videos of 60 s were captured for each sample, each with an infusion rate of 50 and a camera level between 11 and 16. Videos were analyzed with the NTA 3.2 software, setting the detection threshold at 5.
Transmission Electron Microscopy (TEM) analyses were carried out using a Thermo Scientific Talos™ F200X G2 S(TEM) operating at 60 kV. For these measurements, performed without staining the sample, the freshly prepared EV-Lipo-pep-NCs nanoconstruct was diluted at a final concentration of 10 µg/ml in 1:1 bidistilled water and NaCl solution. Then, a single drop of the diluted sample was put on a copper holey carbon grid and let dry before being analyzed.
Fluorescence microscopy analyses were performed to assess the NCs co-localization with the lipidic shell; for these measurements, a wide-field fluorescence inverted microscope (Nikon Eclipse Ti-E), operating with the NIS-element software, equipped with a super bright wide-spectrum source (Shutter Lambda XL) and a high-resolution camera (Zyla 4.2 Plus, 4098×3264 pixels, Andor Technology) was exploited, using a 60x and a 100x immersion oil objective (Plan Apo 1.40, Nikon). The different components of the nanoconstruct were labeled as follows: the aminopropyl-functionalized NCs with Atto 550 NHS ester (excitation peak at 544 nm, emission peak at 576 nm, Sigma-Aldrich), the peptide-lipid conjugate with FITC (excitation peak at 495 nm, emission peak at 519 nm); EVs with WGA647 (Wheat Germ Agglutinin - Alexa FluorTM 647 Conjugate, Thermo Fisher, excitation peak at 650 nm, emission peak at 668 nm). For the NCs labeling, Atto 550 NHS ester (Sigma Aldrich) was added in the ratio of 1 µl per 500 µg of NCs and overnight stirred at 200 rpm. Then, the sample was centrifuged for 10 minutes at 14000 RCF and washed twice with ethanol. The pellet was dispersed in bidistilled water (concentration 10 µg/µl), and this new stock was used for the nanoconstruct preparation. Concerning the Lipo-pep labeling, the CKAAKN peptide already bound to FITC dye (BioFab) was ad-hoc prepared and bound to the functional lipid, obtaining DSPE-PEG-CKAAKN-FITC, then used to prepare the Lipo-Pep formulation. Finally, the stock 50 µl solution of EVs in the cryovials were double diluted with the addition of physiological solution, and 1 µl of WGA647 was added, orbitally shaken (37°C, 200 rpm) for 30 minutes, and then purified with 50 kDa Amicon filter to remove unbound dye.
Once all the nanoconstruct components were labeled, the coupling process was carried out by freeze-thaw method, as reported above. The fluorescence microscope analysis was carried out diluting the sample 1:20 in bidistilled water and 2 µl of the diluted solution was put onto a microscope slide, sheltered with a cover glass slip, and analyzed. The encapsulation’s efficacy was evaluated using the colocalization tool of the NIS software (Nikon): after setting a threshold between 0.1 and 1 µm (to exclude large aggregates), the spots in the 3 channels (the red one to visualize the NCs, the green one to visualize the liposomes, and the far-red one to visualize the EVs) and the overlapped spots in the merged images were counted. The successful coupling was measured on the ATTO 550-labeled NCs colocalized with the other channels, thus using the following equation: %colocalized NCs = (n◦ colocalized red spots)/(tot n◦ red spots).
To assess the Gemcitabine release, RPMI 1640 cell culture medium (ATCC) supplemented by 10%v of Fetal Bovine Serum (ATCC) was used as dispersant and the drug release kinetics was compared between the lipid-coated NCs with respect to the uncoated ones. The volume corresponding to 350 µg of aminopropyl functionalized Ol-ZnO-Gd NCs was taken in quadruplicates from the stock and centrifuged at 14000 RCF for 10 minutes into four different 1.5 ml tubes. Two of the pellets were resuspended in 700 µl of Gemcitabine 1 mM water solution, while the other two were dispersed in 700 µl of bidistilled water. Drug uptake was then carried out at 200 rpm for 2 h. Then, the samples were centrifuged at 14000 RCF for 10 minutes, and the supernatant was analyzed by UV-Vis spectroscopy to verify the successful adsorption of the drug. The four pellets were then put at -20°C. After 18 hours, two of the pellets (one drug-loaded NCs and one without drug as control) were thawed and coated only with liposomes in the ratio of 0.25:0.125:0.5 (µg of Lipo, µg of Lipo-pep, and µg of NCs respectively), resulting in Lipo-Lipo-pep-NCs nanoconstructs (see Supporting Information for the detailed preparation). Once prepared, these two samples were centrifuged at 10000 RCF for 5 minutes, and the supernatant was thrown away. Then, the other two samples (one drug-loaded NCs and one drug-unloaded, as control, both without lipid coating) were also thawed and all the four pellets were resuspended in 700 µl of cell culture medium. To obtain technical triplicates for each group, 600 µl from each sample were equally split (i.e. 200 µl) into three 1.5 ml tubes, filled with 800 µl of RPMI 1640, in order to reach the concentration of 100 µg/ml. Furthermore, 1 ml of medium was also prepared in triplicate as a blank control.
All the 15 samples were orbitally shaken at 37°C and 200 rpm, and the drug release was monitored after 0.5, 2, 24, 48, 72 and 96 hours. At each time step, samples were centrifuged at 10000 RCF for 5 minutes and the supernatant of each sample was analyzed in triplicate using a quartz microplate to measure UV-Vis absorbance. The pelleted particles were then dispersed, through vigorous mixing, in the remaining supernatant and the samples were put back in the orbital shaker. After the absorbance readings, the analyzed supernatants were put back in each tube to continue the drug release experiment.
To evaluate the amount of drug released after each time step a calibration curve was used and the absorbance value of the blank medium control was subtracted to the one of all the other samples; furthermore, the remaining absorbance value of the NCs and Lipo-Lipo-pep-NCs control samples was also subtracted to their respective drug-loaded counterpart in order to determine the sole concentration of the drug released in RPMI 1640 without background.
For in vitro cell biology assays, two different pancreatic cancer cell lines, namely BxPC-3 (a human pancreatic cancer cell line, CRL-1687 from ATCC) and AsPC-1 (a human pancreas adenocarcinoma ascites metastasis cell line, CRL-1682 from ATCC) were used. Cells were cultured in RPMI supplemented with 10% FBS and 1% penicillin & streptomycin antibiotic solution at 37°C and 5% CO2 atmosphere.
Nanoconstruct uptake in cells
Flow cytometry analyses were performed to evaluate the uptake of the nanoconstruct by the pancreatic cancer cells. These experiments were executed with a flow cytometer (Guava® EasyCyte 6-“L, Merck Millipore) equipped with 647 nm laser, to track internalization of the fluorophore-labeled nanoparticles into cells. For this purpose, cells were seeded (30000 cells per well) in a 24-well plate (Greiner Bio-One) for 24h before the treatment. NCs were labeled with ATTO 647 and coated with EV-Lipo shell (with and without target peptide), following the procedure reported in the “Extracellular vesicles and lipid coating” section. Then, the EV-Lipo and EV-Lipo-pep coated, ATTO 647 labeled NCs were administered to the cells at 30 µg/ml concentration of NCs for 20 hours. After that, cells were trypsinized and washed with PBS and subjected towards flow cytometric analysis, and the internalization rate was evaluated in terms of % of positive events, as reported in other works(56, 62).
Furthermore, for live-cell imaging optical fluorescence microscopy cells were seeded (5000 cells per well) in an 8-well chamber slide (Nunc™ Lab-Tek™ II CC2™ Chamber Slide System, Thermo Fisher Scientific™) and NCs were coated with EV-Lipo-pep shell having FITC bound to the peptide. Cell membranes were labeled with Wheat Germ Agglutinin, Alexa Fluor™ 647 Conjugate (WGA647, Thermo Fisher) and the cellular uptake of 10 µg/ml FITC-labeled nanoconstruct was assessed after 20 hours of incubation at 37°C and 5% CO2 atmosphere.
Cytotoxicity and cellular apoptosis
The cytotoxicity of EV-Lipo-pep-NCs-Gem nanoconstruct was assessed treating both pancreatic cancer cell lines with a concentration of 30 µg of NCs/ml of RPMI complete medium for 48 hours. The nanoconstruct without Gemcitabine was used as control. Furthermore, the cells were treated with the free drug in the same amount which was loaded on the NCs. Cells were seeded (2500 cells per well) in 96-well plate (Greiner Bio-One) for 24h before the treatment. To assess cells viability after the treatment, 10 µl of WST-1 (Cell Proliferation Reagent WST-1, Roche) was added to each well and after 2 h of incubation at standard conditions, the formazan absorbance was detected at 450 nm by the Multiskan Go microplate spectrophotometer (Thermo Fisher Scientific) using a 620-nm reference.
The evaluation of apoptosis processes involved in cell death was assessed by flow cytometry, seeding 30000 cells per well in a 24-well plate for 24h before the treatment and then administering the nanoconstruct to the cells at a concentration of 30 µg/ml for 24 hours. In parallel, cells were treated with the same amount of free Gemcitabine. After the treatments, cells were trypsinized, washed with PBS, then exposed to a specific reagent (Guava® Nexin Reagent containing Annexin V and 7-AAD, Luminex) for 30 min and evaluated through flow cytometry as recommended by manufacturers.
The statistical comparison between the treatment groups was performed using one-way or two-way analysis of variance (ANOVA) with Origin software. ***p < 0.001 and *p < 0.05 were considered significant. Independent experiments were performed three-times.