Microbubbles attached to drug-eluting beads enable ultrasound imaging and improved delivery of therapeutics

Materials

Distearoylphosphatidylcholine (DSPC), 1,2-distearoyl-3-trimethylammoniumpropane (TAP), and 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-(7-nitro-2-1,3-benzoxadiazol-4-yl) (ammonium salt) (NBD-DSPE) were purchased from Avanti Polar Lipids. Saline, chloroform, PEG-40 stearate, glycerol, and propylene glycol were purchased from Sigma Aldrich. Agarose and alginic acid were purchased from ThermoFisher Scientific. 3D printing material, thick black polyvinyl alcohol (PVA), for 3D printing was purchased from 3D Herndon. Polydimethylsiloxane (PDMS) and curing agent (Sylgard 184) were purchased from Dow Inc. LC Bead M1 Beads (70–150 µm) (M1) and LUMI M0 Beads (40–90 µm) (LUMI) (Biocompatibles Ltd, now Boston Scientific) were loaded with doxorubicin as previously reported.⁴¹.

Production of microbubbles

MB were prepared as previously described by Christiansen et al.⁴²Briefly, PEG-40 stearate, DSPC, and TAP in a molar ratio of 5:4:1 were sonicated in an aqueous dispersion to generate cationic MB under a positive pressure of sulfur hexafluoride. To demonstrate that positive charge is important for adsorption to drug-eluting beads, neutral MB was prepared using the same procedure as above without using TAP. To prepare fluorescent microbubbles, NBD-DSPE (1%) was added to the MB formulation and the same procedure was followed.

Microbubble size analysis

MB size was determined by optical microscopy (AxioImager M1, Zeiss, Thornwood, NY). 10 μL of MB was added to a hemocytometer and multiple images (n = 20) were taken at 10x magnification. These images were then analyzed in MATLAB (Mathworks, Inc., Natick, MA) using a previously published technique to determine microbubble concentration.⁴³.

Charge measurement

The zeta potential of neutral MB, cationic MB and M1 beads was measured using a Zetasizer Nano (Malvern, Westborough, MA). Approximately 1 × 10⁷ MBs were dispersed in 1 mM KCl at pH 7.4 and their charges were measured.

Binding of microbubbles to beads

LCM1 and LUMI M0 are both polyvinyl alcohol-based anionic microspheres. The particle size of LCM1 ranges from 70 to 150 μm, while that of LUMI M0 ranges from 40 to 90 μm. LUMI M0 beads have iodine covalently bound to them, allowing them to be imaged on X-ray or CT images. Both LC and LUMI beads are approved for clinical use. Both can be loaded with doxorubicin to create drug-eluting beads (DEBs) for transarterial embolization, although regulatory approval for drug loading varies by jurisdiction. To bind microbubbles to M1 beads (MB-M1), microbubbles were concentrated by centrifugation (1000 RCF, 10 min) in water (1 mL). Concentrated MB (50 µL) was then added to deionized water (200 µL), followed by the addition of M1 beads (120 µL) to the microbubble solution. The M1 would settle to the bottom of the vessel and could be separated from unbound MB. The solution could also be centrifuged (200 RCF, 1 min) to more rapidly separate MB-M1 from unbound MB.

Quantification of microbubble binding

M1 were loaded with 0%, 25%, 50%, 75% and 100% Dox, where 100% Dox loading corresponds to a clinical loading concentration of 37.5 mg Dox/mL DEB. MB were then added to the Dox-loaded M1 beads (MB-M1 DEB) using the same method as described above. MB-M1 DEB were then analyzed microscopically, taking images of at least 10 MB-M1 DEB using an upright microscope with 20x objective (AxioImager M1, Zeiss) equipped with a color CCD camera (AxioVision, Zeiss). MB on the bead surface were counted manually in FIJI (National Institutes of Health, Bethesda, Maryland, USA) by light microscopy and by fluorescence signal. Z-stack images were acquired using a Zeiss LSM 710 confocal microscope system (Zeiss, Thornwood, NY) using 467 nm excitation and 539 nm emission for NBD for bubble detection and 480 nm excitation and 590 nm emission for Dox. A Z-stack of the MB-DEB was acquired starting at the center of the bead and taking images every 10 µm to the tip of the bead (40–80 µm, depending on DEB size) using a 20× objective. The number of MB per M1 was then estimated for the entire DEB surface and recorded for each percentage of Dox loading of M1.

Release study

For the cavitation release study, HIFU was applied to the MB-M1 DEB at room temperature in saline in an ultrasound-transmissive PDMS device with a volume of approximately 800 µL MB-M1 DEBs. The HIFU conditions were 1 MHz, 1 MPa, 30% duty cycle, 10 ms pulse length, and a PRF of 100 Hz for 5 s. The MB-M1 DEB solution was then transferred to 15 mL of saline and placed in a rotating mixer (Ward's).^® Rotating mixers, VWR, Bridgeport, NJ) and stirred with an upright rocking motion at 37 °C at 70 rpm based on a method reported by Negussie et al.⁴⁴. This was designed to keep the saline well mixed and act as an infinite sink to optimize diffusion of Dox from the bead and allow comparison of release rates following different stimuli. The preparation was shaken and then 200 µL samples were taken from the sample at t = 0 and then hourly for 8 hours and then at 24 and 48 hours. After each sample was taken, an equivalent volume of saline, 200 µL, was then added to maintain the total volume throughout the release study. The application of HIFU and transfer to the saline solution took approximately 1 hour. As a control, MB-M1 DEB were exposed to the same temperature conditions and treatment environment but without HIFU exposure. The release study was performed for MB-M1 DEB loaded with 75% and 100% Dox. Dox was analyzed by high-performance liquid chromatography (HPLC) according to a standard protocol by Negussie et al. using an LC1290 (Agilent, Santa Clara, CA, USA)⁴⁴.

Delivery study

Ultrasound drug release studies were performed in a 1% agar and 0.5% alginic acid tissue-like material (TMM) vessel phantom with a tubular lumen to represent a vessel, which was created by casting the TMM into a 3D printed mold. The TMM was degassed in the mold during curing, creating a channel within the TMM with a diameter large enough to be visualized by magnetic resonance imaging (MRI) (diameter = 3 mm). For reference, the hepatic artery and the right and left hepatic arteries are approximately 2–5 mm in diameter⁴⁵. The ends of the phantom could also be sealed to accommodate the MB-M1. The vessel was targeted using MRI-guided HIFU. The HIFU transducer (FUS Instruments, Toronto, ON, Canada) had a center frequency of 1.125 MHz, an aperture of 75 mm, 60 mm diameter with a radius of curvature of F# 0.8. Calibrations for total power were performed using an acoustic radiation force balance (ultrasonic power meter UPM-DT-50SP, Ohmic Instruments Inc) with degassed water at 25 °C. A 3.0 T MRI system (Philips, Best, The Netherlands) was used and HIFU targets were planned using T2-weighted images with an in-plane resolution of 0.1 mm. The HIFU beam was registered using a phantom in the MR room. HIFU was applied from below at 1 MHz, 1 MPa, 30% duty cycle, 10 ms pulse length, and a PRF of 100 Hz for 5 s at set points approximately 0.2 mm apart along the length of the vessel (Fig. 2). Phantoms (n = 3 for each time point) were then cut in a plane perpendicular to the axis of the lumen 25 min and 100 min after HIFU application. The cut agar was then examined by fluorescence microscopy (Imager M1 Axio Fluorescence Microscope, Zeiss) at an excitation peak of 480 nm and emission of 590 nm and a composite image of the vessel wall was created. This vessel section was then analyzed using radial profiles in FIJI and the penetration depth of Dox was averaged over 3 vessel sections and plotted using Python (Centrum Wiskunde & Informatica, Amsterdam, the Netherlands). This was then compared to MB-M1 DEB without HIFU. MB-M1 DEB loaded with 75% and 100% Dox were studied to compare the amount of Dox to the amount of MB for drug release.

Ultrasound imaging

The same phantoms used to study Dox delivery were also used to study the imaging ability of MB-LUMI under ultrasound and CT. LUMI were chosen to allow confirmation between the two imaging modalities. After the vessel phantom was filled with saline, MB, MB-LUMI, or LUMI were injected into the lumen under ultrasound imaging (Sonosite Ultrasound, Fujifilm Sonosite, Inc., Bothell, WA) using an L25 x linear probe to ensure correct needle positioning. Flow was subsequently stopped to allow beads to sediment or MB to rise. To avoid contamination, a separate phantom was used for each phantom. After sample injection, US B-mode videos were recorded for one minute to allow sedimentation. After sedimentation, snapshots were then taken and a region of interest was selected for pixel analysis. The average pixel intensity within that region was then measured. After an ultrasound video was acquired, cone beam computed tomography images (Allura Xper FD20 X-ray system; Philips) were acquired at 80 kVp with a spatial resolution of 0.656 mm. The CT images were then examined using the Mimics Research software package (version 20.0; Materialise, Leuven, Belgium) to determine the position of MB-LUMI in the vessel.