AGuIX® NPs characteristics
The Gd-based nanoparticles are made of a polysiloxane network surrounding by cyclic chelates of gadolinium [35]. Their main characteristics are their ultra-small size with a hydrodynamic diameter of 3 ± 2 nm and a mass of approximately 10 kDa, a mean Gd/Si ratio of 10/40, and a strong complexation constant (log β110) for Gd (i.e. 24.78). The particles were developed for medical theranostic approaches, combining the contrast properties for MRI acquisition in T1 mode, and the radiosensitizing action of the high-Z element gadolinium (Z = 64). Different fluorescent dyes (FITC, Cy5.5, and Rhodamine-B (RhoB)) have been conjugated to the AGuIX® for the biological investigations as previously reported [36], with a fluorescent conjugation yield of 1/600 to 1/250. A short characterization of the NPs can be found in Additional file 1: Figure S1. The AGuIX® are freeze dried for long-term storage and can be solubilized in water and biocompatible solvent (NaCl 0.9%, as example) before administration.
Culture of MucilAir/OncoCilAir™ tissues
MucilAir™ and OncoCilAir™ were produced by Epithelix and OncoTheis Companies respectively (see Additional file 1: Figure S2). A complete description of the cultures can be found on the websites (Epithelix, Geneva, Switzerland, http://www.epithelix.com and Oncotheis http://www.oncotheis.com). Upon receipt, bronchial MucilAir™ pool of donors cultures or OncoCilAir™ cultures (KRAS mutated) containing A549-GFP cells were transferred into 24-well plates filled with 700 µL of specific pre-warmed MucilAir™ or OncoCilAir™ culture medium. The tissues were routinely cultured at 37 °C and 5% CO2 in a saturated humidity environment (≥ 99%) in air–liquid interface (ALI) culture condition. Medium renewal was performed every two–three days. Trans-epithelial electrical resistance (TEER) was assessed as a standard indicator of tissues integrity. Measurements were conducted using an EVOM® resistance meter and STX 2 electrodes (World Precision Instruments, Sarasota, USA) at least once a week, and before/after NPs exposure. The TEER values (Ω) were established by using the following formula: TEER (Ω·cm2) = (resistance value (Ω) − 100 (Ω)) × 0.33 (cm2), where 100 is the resistance of the insert membrane and 0.33 cm2 is the total surface of the 24-well plates culture insert. The TEER values described for well-preserved MucilAir™ cultures could be > 200 Ω·cm2 (well-preserved 200–600 Ω·cm2 [24] with optimal mean TEER 300–400 Ω·cm2 [25]), and for OncoCilAir™ around 100–200 Ω·cm2, the presence of tumor A549 areas decreasing the tightness of the epithelium, as reported by the company.
AGuIX® acute toxicity evaluation on MucilAir™
MucilAir™ AGuIX® exposure has consisted of a liquid apical (air/liquid interface) or basal (liquid/liquid interface) 24-hour exposure, for final concentrations of 1 and 10 mM ([Gd3+]) NPs according to the experiments in their native or conjugated (FITC, RhoB, or Cy5.5) form; with kinetic monitoring up to 72-hour after apical and basal washing with PBS and addition of fresh medium in the basal compartment. AGuIX® solutions have been prepared with specific MucilAir™ culture medium for basal exposure (Vf = 700 µL), or for apical exposure advisable vehicle saline solution of 10 mM Hepes, NaCl 0.9% and 1.25 mM CaCl2 (Vf = 30 µL). For apical exposure without mucus, the mucus was gently removed using a PBS washing step. All the bio-toxicity markers used for the analysis were assessed after the 24-h AGuIX® exposure.
Morphological observations of cultures before and after AGuIX® exposure were performed to evidence potentially figure modifications as decrease of ciliary frequency, detachment of cells, or apparition of cavity/hole into the insert using an inverted optical microscope (MOTIC AE2000 Trino).
Cell mortality was then evaluated by quantifying the lactate dehydrogenase (LDH) released from cells into the basal compartment with damaged membranes using the LDH Cytotoxicity kit plus (Roche Diagnostics, Mannheim, Germany), according to the manufacturer’s instructions. Optical density detection was performed using a microplate reader (Multiskan GO, Thermo Fisher Scientific Inc., Wyman Street, Waltham, USA) at 490 nm. The amount of the released LDH was reported as a percentage of cytotoxicity according to the following equation:
$${{\text{Cytotoxicity }}}\left( {\%} \right) = \frac{{\left( {{\text{OD\; of\; AGuIX\; exposed\; samples}} - {{\text{ OD\; of\; unexposed\; samples}}}} \right)}}{{\left( {{{\text{OD\; of\; positive\; controls }}} - {{\text{ OD\; of\; unexposed\; samples}}}} \right)}} \times 100$$
(1)
after subtraction for all values of blank control (OD of culture medium). The positive control corresponds to OD values after the complete lysis of control cells, with adequate buffer provided in the kit.
IL-8 pro-inflammatory response was assessed using a commercial enzyme-linked immunosorbent assay kit (Quantikine® Human IL-8 Immunoassay, R&D systems Inc., McKinley Place NE, Minneapolis, USA). The optical density was determined according to the manufacturer’s instructions, using a microplate reader (Multiskan GO, Thermo Scientific) at 450 nm. A standard curve was established, and results were expressed as pg mL−1 of IL-8. A specific IL-8 positive control (cytomix stimulation: TNF-alpha at 500 ng/mL and LPS at 200 μg/mL incubated for 24-h in the basal culture medium) was used, according to the manufacturer recommendations [37]. It allowed the production of 4400 ± 510 pg/mL IL-8 (p = 0.028 compared to unexposed cells). Possible interaction of the NPs and the assays was checked to ensure the absence of false negative results.
AGuIX® permeability assessment on MucilAir™ barrier
To assess the apparent permeability (Papp) of a compound across upper-airway epithelium of a respiratory tract, we specifically investigate the apical-to-basal transport of AGuIX-FITC in MucilAir™ tissues, using the same NP’s preparation as described in “AGuIX® acute toxicity evaluation on MucilAirTM” section. For this experiment, only inserts displaying a TEER around 300 Ω·cm2 were used. Atenolol (10 mM) and salicylic acid (100 μM) were respectively used as low and high chemical permeability molecules, and the values were compared with the one of AGuIX® exposure [24, 38]. All compounds were diluted in HBSS transport buffer. Prior to incubation, all solutions were pre-warmed to 37 °C and the pH was adjusted to 7.4. Each condition was evaluated in triplicate. Papp was determined as follow. A pre-warmed HBSS donor solution (200 μL) was added to the apical compartment. A fraction of the donor solution was immediately withdrawn to determinate the effective initial concentration or fluorescence (AGuIX®-FITC). At the same time, 500 μL HBSS solution were filled into the basal acceptor compartment. Transport experiments were performed during 2 h at 37 °C. The fluorescent intensity of AGuIX®-FITC into apical and basal AGuIX®-FITC after this permeability experiment were measured with a fluorometer (Fluoskan Ascent, Thermo Fisher Scientific Inc., Wyman Street, Waltham, USA). The amount of atenolol and salicylic acid that permeated through the culture insert in both compartments was quantified by liquid chromatography-mass spectrometry (Aquity UPLC system) coupled with a Xevo TQ-D triple quadrupole mass spectrometer (Waters, Saint-Quentin-en-Yvelines, France). The permeability was finally calculated using the following equation:
$$Papp = \frac{V}{Ci \times A} \times \frac{Cf}{\Delta t}$$
(2)
where V is the volume of the donor compartment (cm3), Ci is the initial concentration of the compound (in g/L or M) or fluorescence intensity, A is the area of the insert (0.33 cm2), Cf is the final concentration of the compound in the acceptor compartment (in g/L or M) or fluorescence intensity, and ∆t is the duration of the experiment (in s). Papp was expressed as cm·s−1.
Functional monitoring of MucilAir™ cultures
Measurement of the Cilia Beating Frequency (CBF) was performed at room temperature (RT) by a dedicated set-up made of three parts: a camera (Sony XCD V60 Firewire), a PCI card, and a specific package of software. The CBF was calculated using CiliaX software (Epithelix, Geneva, Switzerland), and expressed as Hz.
The mucociliary clearance was monitored using a high-speed acquisition camera (Sony) connected to an Axiovert 200 M microscope (Zeiss, Jena, Germany). Microbeads (30 µm of diameter) were added onto the apical surface of the MucilAir™. Then, 30 s’ movies (3 movies/insert) showing the movement of the small beads will be taken and analyzed using the imaging software Image Pro Plus (Mediacy). The movement of the beads was tracked, and velocity of each particle was calculated to determine the speed of the mucociliary clearance.
MucilAir™ phenotypic analysis by Flow Cytometry (FCM)
According to Epithelix company, MucilAir™ tissues were dissociated to surface insert with a trypsinization protocol to obtain a cellular suspension containing approximately 300 000 cells/insert. For each immuno-phenotyping, 100,000 cells were used. To avoid fixation/permeabilization steps inducing a significant loss of the intracellular AGuIX® fluorescent signal, the cell type discrimination was performed with a vital CLCA1 staining (anti-CLCA1.PE-Cy5.5 human polyclonal goat antibody, #AC21-1575-16, Abcore, CA). CLCA1 is a membrane marker expressed in mediating calcium-activated chloride conductance for mucus production by goblet cells, whose specificity could be assessed in a previous MucilAir™ characterization study [25, 26]. This immunolabeling allowed a discriminant gating between the two differentiated cell types of CLCA1+ goblet population vs. CLCA1− population, that included mostly ciliated cells and at last basal cells. Finally, for each sample, 50,000 cells were analyzed by flow cytometry (FCM), and the fluorescent signal was considered as quantitative. Intact cells were selected by a sequential gating using the FSC vs. SSC cytogram before selection according to specific CLCA1+ vs. CLCA1− cell populations. Gain setting was established by running unstained and isotype controls cells (Lightning-link PE-Cy5.5 tandem conjugation kit to IgG isotype control antibody, Innova Biosciences). AGuIX®-FITC or GFP, green fluorescence, excited by 488 nm laser line was acquired through a BP 530/30 nm filter, PE-Cy5.5 excited by 488 nm laser line, through a 695/40 nm BP filter, and far red fluorescence of Cy5.5, excited by 633 nm laser line, through a 712/21 nm BP filter. Compensations were calculated by acquisition of fluorescent beads of each used fluorophore (Calibrat™ BD Biosciences, CA). Both percentage and geometric mean of fluorescence were collected. The results were expressed as the percentage of cells of interest, and geometric mean fluorescence intensity (MFI). FACS DiVa (BD Biosciences, CA) equipped with an argon ion and He–Ne lasers (Coherent, CA) was used. Data were analyzed with DiVa 5.03 software.
Quantitative NPs’ uptake evaluation by FCM
Tumor cell population on OncocilAir™ model was ranked by FCM according to the GFP fluorescent signal intensity. The total cell population was therefore split into 3 sub-populations, i.e., normal healthy cells (GFP−), GFP+ and GFP++ tumor cells, expressing respectively low and high level of GFP protein. For each sub-population, the level of AGuIX® uptake was determined and expressed as the % of positive cells and MFI values. Finally, a relative uptake ratio values between tumor and normal areas was calculated as follow, after deduction of autofluorescence values:
$$AGuIX \; relative\; uptake\; ratio = \frac{{\left( {MFI_{GFP + , + + } } \right)}}{{\left( {MFI_{GFP - } } \right)}}$$
(3)
Cell cycle analyzed by FCM
The cell cycle profile was determined according the double staining of DNA A-T specific vital dye Hoechst combined with propidium iodide (PI), a passive DNA intercalating dye. Hoechst fluorescence was collected after a UV excitation, while emission was collected through a 424/44 nm BP filter, and PI was excited at 488 nm, and its emission collected through a 695/40 nm BP filter. Dead cells and doublets of cells were excluded based upon PI positive staining, and specific gating drawing based upon Hoechst width signals versus Hoechst area signals respectively. Quiescent (G0/G1) and proliferative (S and G2 + M) cells were determined on the basis of linear quantification of Hoechst fluorescence.
Microscopic observations of flat mounted MucilAir™/OncoCilAir™ tissues
Tissue preparations and observations
For observations of living tissues, the whole living MucilAir™/OncoCilAir™ tissues was removed from its cell culture insert and placed on a microscopic slide. A Viscoelastic System (Viscot, DuoVisc, Alcon) containing hyaluronate de sodium, chondroitin sulfate was added on the tissue for minimizing the cell stress during coverslip mounting. The tissue was then gently flattened using a large glass coverslip retained by adhesive tape.
Epifluorescence microscopy
By using an epifluorescence inverted microscope IX81 (Olympus, Tokyo, Japan) equipped with the CellSens imaging software (Olympus, Munster, Germany), the cells were visualized by phase-contrast, and tumor cells were highlighted by the GFP-fluorescence.
Confocal microscopy
The internalization of AGuIX®-Cy5.5 was relatively evaluated for A549-GFP tumor cells and healthy surrounding epi-airway epithelium using a FLUOVIEW FV1200 laser scanning confocal microscope (Olympus, Tokyo, Japan) equipped with the FV10-ASW4.1 imaging software. A 635 nm laser was used to excite the Cy5.5 dye, while the signal was collected using a long-pass filter λ > 650 nm.
Tumor-nodule area monitoring on OncoCilAir™
The growth of A549 tumor nodules tagged by GFP was observed on whole surface of OncoCilAir™ without any sample preparation, using a fluorescence macroscope MVX10 (Olympus, Tokyo, Japan) equipped with the software CellSens imaging systems (Olympus, Hamburg, Germany) or the macroscope system Leica equipped with the software Andor SOLIS (Oxford Instruments).
Measurement of tumor area surface on OncoCilAir™
From images acquired with Olympus fluorescence macroscope at the time culture reception (i.e. 20–25 days after A549 cell implantation) and during their follow-up, an image processing using ImageJ platform has been performed (open-source program of the National Institutes of Health, http://rsb.info.nih.gov/ij/). Eight [8] bits raw images were first treated to remove the background and tumor edges were detected (edge filtering). Raw and edges images were combined to increase automatic detection of tumor areas. Detection of tumor areas was performed by applying a Reny threshold. Binary areas were sent to ROI manager to extract individually settings: area, perimeter, mean fluorescent intensity of all detected tumors. The mean fluorescent intensity signal of 100 over 255 was chosen to discriminate GFP+ and GFP++ populations. Mean signals under 100 were identified as GFP+, and GFP++ was attributed for values > 100. Individual temporal tracking of tumors areas and intensities were exported to GraphPad Prism7 for figures and statistical analysis.
Additional nanoparticle observations by two-photon confocal microscopy
AGuIX®-RhoB (10 mM) were added to the apical phase of the cells during the indicated time. Two-photon microscopy was performed as described in Sancey et al. [39], using a LSM 7 MP (Zeiss, Germany) equipped with a 20 × water-immersion objective (NA 1.0; Zeiss) and ZEN 2010 software for detection of the NPs. Laser excitation was done at 800 nm with a Ti:Sapphire laser (Chameleon vision II; Coherent, UK). Fluorescence emissions were detected simultaneously by three non-descanned photomultiplier tubes with a 492/SP25 nm filter (Semrock, US) for blue autofluorescence and Hoechst emission, a 542/50 nm filter (Semrock, US) for green autofluorescence emission, and a 617/73 nm filter (Semrock, US) for AGuIX®-RhoB fluorescence emission. Autofluorescence and second harmonic generation of biological structures could also be collected in the 3 channels due to the presence of collagen, lectin and elastin as example. Confocal microscopy was performed using an LSM 510 (Zeiss Germany) equipped with a 40 × oil-immersion objective (NA 1.2; Zeiss). Laser excitations/emissions were 760 nm biphotonic/400–450 nm for Hoechst, 488 nm/500–550 nm for FITC/GFP, 543 nm/550–600 nm for Rhodamine-B, 633 nm/650–705 nm for Cy-5, respectively.
Immuno-labeling for confocal microcopy
To discriminate the cellular organization of tumor areas from healthy tissue in OncoCilAir™ model, the sample was fixed in 4% paraformaldehyde (PFA) for 30 min at RT, then permealized in 0.5% Triton X-100 for 15 min at RT. The cells were stained with either ethidium bromide (FT-25810A, Molecular Probes) or phalloidin (Actin-stain™ 555, PHDH1, Cytoskeleton) for 30 min at RT, for nuclear and F-actin staining respectively. Phalloidin stained the cell outline of the respiratory epithelium. Before microscopic observations, the tissues were placed on a glass slide, covered with an anti-fading mounting medium (H1000, VECTASHIELD®, Vector) and gently flattened using a large glass coverslip retained by adhesive tape.
The inserts were incubated with AGuIX® solution at 10 mM for 60 min to 72 h, in air/CO2 5%, 37 °C environment. The inserts were cryofixed using liquid nitrogen vapor, and sliced at 8 µm. A solution of 5% goat serum in PBS was used to block any unspecific sites. To discriminate ciliated and goblet cells, a beta IV tubulin antibody was used (#T6793, 1/500, Sigma-Aldrich, France) and anti-mouse IgG-Alexa488 (#A11029, 1/2000, Life Technology, France). Beta IV tubulin is a major constituent of microtubules in motile ciliary axonemes. The sections were incubated with 1 mM Hoechst solution before mounting with the coverslip.
Quantification of Gd content by inductively coupled plasma—optical emission spectrometry (ICP-OES)
Determination of the Gd content in the samples (apical and basal fractions, collected first after 24 h-exposure and secondly at 72 h end kinetic) was performed by ICP-OES analyses (Varian 710-ES) with a detection limit of 0.5 µg/L. The apical fraction was collected by gentle washing using saline solution, while basal fraction was sampling during medium renewal. Before measuring Gd concentration, 3 mL of aqua regia (mixture of acids: nitric and hydrochloric) were added to each fraction before heating for 3 h at 80 °C (SCP Science DigiPREP MS). After complete mineralization, the samples were diluted with HNO3 (5%, w/w) to reach a 50 or 10 mL-volume (respectively), and finally filtered at 0.2 µm for the measurements (i.e. a dilution factor of 500). The results were expressed as the total mass of Gd in the sample or its percentage relative to the administrated dose.
Cultures radiation exposure and γH2AX labeling
The cultures were incubated or not with the AGuIX® NPs (10 mM apical) 24 h before the radiation exposure, in presence of mucus. The x-ray exposure was performed at 4 Gy at 250 kVp; this dose was selected based on previous experiments. Similar experiments were performed on A549 2D cultures and the radiation exposure was conducted at 4 and 6 Gy. γH2AX labeling was performed on fixed cells as described in Kotb et al. [4], using anti-phospho-histone H2AX (Merck Millipore). Colony assay was also performed as described in Kotb et al. [4].
Statistics
The Mann–Whitney test was used for the statistical analysis using Prism software for bio-toxicity and uptake experiments. Temporal tracking of tumors areas was compared using an ordinary one-way ANOVA with a Tukey test corrected for multiple comparisons as post hoc tests using a Graph Pad Prism7 suite.