COSAN-stabilised omega-3 oil-in-water nanoemulsions to prolong lung residence time for poorly water soluble drugs

Herein, we report on the capacity of the amphiphilic inorganic anion cobalt bis(dicarbollide) to stabilise oil-in-water nanoemulsions (NEs). The resulting NEs show long term stability in water and high drug-loading capacity, and can prolong the residence time of hydrophobic drugs in the lungs as determined by in vivo positron emission tomography imaging.

For sample preparation, freshly glow-discharged 200-mesh grid (R2.2 200 Mesh; QUANTIFOIL) was placed inside the chamber of a Vitrobot Mark III (FEI Company, USA), which is maintained at 8˚C temperature and relative humidity close to saturation (85% rH). Four microliters of non-diluted sample were dropped onto the grid for 15 seconds. After incubation, most of the liquid on the grid was removed by blotting (blot time was 3 seconds, number of blots was set to 1, drain time was zero and blot offset was 1 mm) with absorbent standard Vitrobot filter paper (Ø55/20mm, Grade 595, Thermo Fisher Scientific -FEI). After the blotting step, the grid was abruptly plunged into a liquid ethane bath, previously cooled with liquid nitrogen at approximately -170 ºC. Once the specimen was frozen, the vitrified grid was removed from the plunger and stored under liquid nitrogen inside a cryo-grid storage box.
During imaging, no-tilted zero-loss two-dimensional (2D) images were recorded under low-dose conditions, utilizing the 'Minimum Dose System (MDS)' of Jeol software, with a total dose on the order of 10-20 electrons/Å² per exposure, at defocus values ranging from 1.5 to 4.0 µm. The in-column Omega energy filter of the microscope helped us to record images with improved signal-to-noise ratio (SNR) by zero-loss filtering, using an energy selecting slit width of 30 eV centred at the zero-loss peak of the energy spectra. Digital images were recorded on a 4K × 4K (15 µm pixels) Ultrascan4000™ charge-coupled device (CCD) camera (Gatan Inc.) using DigitalMicrograph™ (Gatan Inc.) software, at different nominal magnifications from 4000× to 60000×. Number size distribution was achieved from several micrographs using an automatic image analyser (ImageJ). 500 droplets were selected for the analysis.

Absorbance analysis
Absorbance analyses were conducted using a Synergy™ HT Multi-Detection Microplate Reader (Bio-tek Instruments). All measurements were performed in disposable 96 wellplates at  = 405 nm.

Liquid-Liquid Interfacial Tension (IT) analysis by reverse pendant drop method
Interfacial Tension (IT) analyses were conducted by reverse pendant drop method using the Attension Theta Flex optical tensiometer from Biolin Scientific. The liquid with a lower density (i.e. DHA oil) was dispensed using a hooked needle (C205/C205A/C201, Biolin Scientific) for oil drop generation. The liquid with a higher density (i.e. aqueous phase) was placed in a GC10 glass cell (Dataphysics instrument GmbH). The drop was captured using a CCD camera at 3.5 fps. Data treatment was performed in One-Attension commercial software using Surface tension (Young-Laplace) analysis mode.

Effect of the amount of COSAN on the size distribution of DHA emulsion droplets
In order to determine the optimal amount of COSAN for the formation of the NEs, experiments using different amounts of this compound (1-8 mg) were carried out. In brief, the corresponding amount of COSAN (1-8 mg) was introduced in an 8 mL glass vial containing ultrapure water (3.6 g). DHA (oil phase; 400 mg, 425 µL) was added, and the emulsion was then formed by sonication (0°C, under stirring) using an UP400S (Hielscher) system at 100% of amplitude and pulse during 4 minutes (400 W) with a H3 sonotrode tip (3 mm diameter, 100 mm length).
A fraction of the final emulsion (500 µL) was purified by size exclusion chromatography (SEC) using an Illustra NAP TM-5 column (Fig. S1), and the collected fraction was diluted with ultrapure water to a final volume of 3 mL. A different fraction of the non-purified emulsion (500 µL) was directly diluted with ultrapure water to the same final volume (3 mL). The samples were freeze dried for water removal, diluted with dioxane/MeCN (80/20, 350 µL) and submitted to LD analysis (purified NEs; Fig. S2) and absorbance analysis (all samples; Fig. S3 and Table S1).

Quantification of the amount of COSAN in the NEs.
The amount of COSAN present in the NEs before and after purification was determined using absorbance analysis ( = 405 nm). With that aim, a calibration curve was first generated by dissolving increasing amounts of COSAN (0.156,0.313,0.625,1.25,2.5,5 and 10 mg/mL) in DHA/dioxane/MeCN (12.5/70/17.5) (Fig. S3). Quantification of the samples (Table S1) showed that the whole amount of COSAN remained inside the emulsion after SEC purification when 1-2 mg of COSAN were used for the preparation of the NEs, suggesting that all the COSAN is used to stabilize the oil droplet and located at the oil/water interface. As the solubility of COSAN in water is 0.5 mg/mL), all the emulsions produced with > 2 mg in the water phase represent an excess of COSAN in the aqueous phase. Still, quantification of the amount of COSAN in these NEs suggests that the majority of COSAN is incorporated in the NE.

Determination of COSAN solubility in DHA oil
COSAN (2 mg) was added to DHA (1 mL) and the mixture was heated at 50°C under continuous stirring. After 15 minutes, the presence of precipitate was confirmed by visual inspection, and the mixture was allowed to cool down to room temperature. After filtration, a small sample (50 µL) was diluted with Dioxane/MeCN (80/20, 350 µL) and submitted to absorbance analysis ( = 405 nm) using the calibration curve shown in Figure S3. COSAN solubility was determined as 0.77 ± 0.05 mg/mL.

Determination of interfacial tension
Interfacial tension analysis was carried out using the pendant drop method (see Figure S4 for experimental set up). Experiments were carried out in three different scenarios: (i) DHA

Determination of the amount of COSAN at the oil/water interface
Taking into account the hydrodynamic diameter of the NEs obtained via DLS (due to the high polydispersity of the oil droplets, this diameter was preferred compared to the number average diameter obtained by cryo-TEM), the amount of COSAN (m COSAN ) required to cover the total surface area of the droplet was calculated, assuming that the coverage of COSAN was around 2 nm 2 (ref 24 in the manuscript).
It is known that the total surface area of colloids in a dispersion can be calculated from the following equation: Where: -m is the weight of particles or droplets. In our case it is the weight of oil emulsified (400 mg).
-ρ is the density of the particles. In our case the density of the oil, i.e. 0.93 g cm -3 -D is the diameter of the oil droplet, here 170 nm Thus it was found that 6.5 mg of COSAN was required to cover the entire droplets, which is pretty similar to the amount of COSAN found after purification. So, these results suggest that most of the COSAN is located at the interface and also that the COSAN is adsorbed as monolayer at the oil/water interface.

Stability of the NEs
Long term stability in water at different storing conditions (accelerated test conditions) COSAN-stabilized DHA-in-water emulsion was freshly prepared as described above (using 8 mg of COSAN) and 1 mL fractions were stored at 4 different conditions (A: light at r.t., B: dark at r.t., C: dark at 5 °C and D: dark at 40 °C) for 3 months. Emulsion stability at 1, 2 and 3 months was subsequently evaluated by visual inspection; the hydrodynamic diameter was determined by DLS and volume average diameter by LD (Fig. S5).    (3x500 μL). The solvent was finally evaporated to dryness. Quality control was performed by radio-HPLC (Fig. S8).

In vivo Imaging experiments
Without recovering from sedation, animals were positioned in an eXploreVista-CT small animal PET-CT system (GE Healthcare, USA) to perform in vivo studies. During imaging, rats were kept normothermic using a heating blanket (Homeothermic Blanket Control Unit; Bruker). Anaesthesia was maintained by inhalation of 1.5-2% isoflurane in pure O 2 .
Dynamic PET images (energy window: 400-700 KeV) were acquired with the following frames: 4 x 15 s, 4 x 30 s, 3 x 60 s and 3 x 90 seconds. In all cases, four beds were defined to acquire whole body images (total acquisition time = 42 min). After each PET acquisition, a CT scan (X-Ray energy: 40 kV, intensity: 140 μA) was performed for a later attenuation correction in the image reconstruction and for unequivocal localization of the radioactivity. Random and scatter corrections were also applied to the reconstructed image (filtered back projection reconstruction algorithm), generating a 175x175x220 dimension image, with a 2 mm axial FWHM spatial resolution in the centre of the Field Of View (FOV). PET-CT images of the same animal were co-registered and analysed using PMOD image processing tool. First, Volumes of interest (VOIs) were manually delineated on the lungs to assess the amount of radioactivity deposited. Time-activity curves (decay corrected) were obtained as cps/cm 3 in each VOI and values were then normalised to the amount of radioactivity in the lungs in the first frame, and expressed in percentage (see Fig.   S9 for time activity curves obtained for the different labelled species).

Emulsion loading capacity
Increasing amounts of cold 17β-estradiol (20 mg, 30 mg and 100 mg) were solved overnight in 1 g of DHA oil. After dissolution, 400 mg of each mixture (2, 3 and 10 % wt.) and 50 µL of a trace solution of [ 18 F]FES in dichloromethane were poured into 3.6 mL of ultrapure water containing 8 mg of COSAN. Emulsions were produced as described above.
Once produced, 500 µL of each emulsion were purified by SEC using an Illustra NAP TM-5 column and diluted with ultrapure water to a final volume of 2 mL in plastic tubes. The same dilution was applied for 500 µL of non-purified emulsions. The radioactivity of the samples was measured using a Gamma Counter. The incorporation of [ 18 F]FES was subsequently calculated by relating the amount of radioactivity obtained for each nonpurified emulsion (equivalent to a theoretical 100% [ 18 F]FES incorporation) with its purified analogous (X% of [ 18 F]FES incorporation). Size distribution was also evaluated by DLS (Fig. S10).  Figure S10. NE size distribution after encapsulation of different amount of FES (0, 2, 3, and 10% wt.) as determined by DLS.