Fluorous-Phase Iron Oxide Nanoparticles as Enhancers of Acoustic Droplet Vaporization of Perfluorocarbons with Supra-Physiologic Boiling Point

Abstract

Perfluorocarbon emulsion nanodroplets containing iron oxide nanoparticles (IONPs) within their inner perfluorohexane (PFH) core were prepared to investigate potential use as an acoustically activatable ultrasound contrast agent, with the hypothesis that incorporation of IONPs into the fluorous phase of a liquid perfluorocarbon emulsion would potentiate acoustic vaporization. IONPs with an oleic acid (OA) hydrophilic coating were synthesized through chemical co-precipitation. To suspend IONP in PFH, OA was exchanged with perfluorononanoic acid (PFNA) via ligand exchange to yield fluorophilic PFNA-coated IONPs (PFNA-IONPs). Suspensions with various amounts of PFNA-IONPs (0-15% w/v) in PFH were emulsified in saline by sonication, using 5% (w/v) egg yolk phospholipid as an emulsifier. PFNA-IONPs were characterized with transmission electron microscopy (TEM), transmission electron cryomicroscopy (cryoTEM), and thermogravimetric analysis (TGA) with Fourier transform infrared spectroscopy (FTIR). IONP were between 5-10 nm in diameter as measured by electron microscopy, and hydrodynamic size of the PFH nanodroplets were 150 to 230 nm as measured by dynamic light scattering (DLS). Acoustic droplet vaporization of PFH nanodroplets (PFH-NDs) was induced using conversion pulses provided by a focused ultrasound transducer, and formed microbubbles were imaged using a clinical ultrasound scanner. The acoustic pressure threshold needed for PFH-NDs vaporization decreased with increasing temperature and IONP content. PFH-NDs containing 5% w/v IONP converted to microbubbles at physiological temperatures just above the exposure limits allowed by the FDA when using clinical ultrasound scanners. Furthermore, IONP-loaded PFH-NDs injected intravenously into melanoma-bearing mice at a dose of 120mg PFH/kg, converted into detectable microbubbles in vivo 5 hours, but not immediately after injection, indicating that this technique detects NDs accumulated in tumors.

GRAPHICAL ABSTRACT:

INTRODUCTION

Ultrasound is the most commonly performed cross-sectional imaging technique worldwide. It is portable, accessible, affordable, and can generate cross-sectional images at high resolution down to a 20-cm depth of field. Backscattered signal occurs at discontinuities based on differences in acoustic properties on either side of the interface. Except for fatty and non-fatty tissues, most tissues have nearly similar acoustic property resulting in low image contrast between normal and abnormal regions. Ultrasound contrast agents were introduced over four decades ago as a means to increase contrast between regions where they accumulate (e.g., liver, spleen and tumors), and background tissues improving tumor detection [1–3]. As ultrasound contrast agents, liquid perfluorocarbon (PFC) emulsions relied on differential tissue accumulation for detection [1], while the micron-sized gaseous microbubbles (MBs) that are limited to the intravascular space rely on differential wash-in and wash-out rates and fractional blood volume [2]. MBs have proven to be highly effective contrast agents because of their non-linear behavior when exposed to ultrasound. Since tissues respond linearly, tissue signal can be subtracted in real-time to produce MB only images. These non-linear imaging techniques have allowed the detection of a single MB in vitro and in vivo [4, 5]. Despite their clinical success, they have certain limitations, especially when used to detect molecular targets. Because of the extreme sensitivity of ultrasound to MBs, miniscule doses are used – between 10⁸ and 10⁹ MBs total dose, suspended in 0.1-2 mL depending on formulation. Because of their large diameter – 1 to 3 μm that limits their distribution to the intravascular space, and because they are short lived – 5 to 10 minutes [2], they have limited ability to find and interact with their receptors that can only be on endothelial cells [6, 7], circulating cells [8], and blood clots [9]. Direct cancer cell targeting with MBs is not possible, as these cells reside in the extravascular space.

Because of their small size (150-250 nm) that allows them to extravasate, their large particle count that is 3 to 5 orders of magnitude greater than MBs, and their long intravascular dwell time (hours) [10], PFC nanodroplets (ND) have an advantage when used for targeting, sparking renewed interest in recent years [11–17]. Unfortunately, PFC NDs have a weak backscatter signal [18], particularly when compared to MBs [19], requiring a larger dose ≥ 1.5 g of PFC/kg to be detected [20, 21]. Non-targeted PFC NDs accumulate in tumors in animals [1] and humans [21], because of the enhanced permeability and retention (EPR) effect. On histology, PFC NDs are found in macrophages in tumors [22], and inflammatory tissues [23] to serve as macrophage imaging agents. When NDs are used to target tumor receptors, the high particle load required for detection saturates all available receptors, where accumulation due to targeting is over-shadowed by non-specific accumulation. While a smaller particle load could achieve higher target-to-background ratio, tissue concentration will be too small to be detectable. If a more sensitive ultrasound imaging technique is developed to detect small concentrations of NDs, as with MBs, detection of targeted extravascular sites becomes possible.

Fortunately, PFC NDs can be vaporized into MBs in vitro and in vivo, which occurs spontaneously if a low boiling point PFC is used [24], or when exposed to ultrasound through a process called acoustic droplet vaporization (ADV) [25]. Resulting MBs can then be detected with non-linear imaging. ADV was first used to generate large MBs from perfluoropentane (PFP) (boiling point = 29 °C) droplets to embolize downstream vessels for therapy [26, 27]. However, undesired side-effects occurred due to spontaneous PFP droplet vaporization in animals [28, 29]. To avoid spontaneous vaporization PFC that boils above physiological temperatures, such as perfluorohexane (PFH) (boiling point = 58°C) could be advantageous. Kripfgans et. al. used 1 μm PFP droplets encapsulated in albumin, and observed that 4.5 MPa at 1.5 MHz (corresponding to a mechanical index (MI) of 3.7) was needed to induce ADV [25]. Because pressure required for ADV increases as droplet size decreases and PFC boiling point increases [30], it is expected that ADV of PFH NDs will be difficult at 37 °C, particularly if the desired acoustic pressure must remain at ≤ 1.9 MI limit set by the FDA [31].

We hypothesized that incorporating iron oxide nanoparticles (IONPs) within the PFH core of NDs as nucleation sites, as suggested in the literature [32, 33], would reduce ADV pressure requirement. While other nanoparticles have been proposed [33], we chose IONPs for their long clinical safety record, and the potential to heat the IONP containing PFC NDs using external alternating electromagnetic fields [34]. Therefore, the aim of this study was to manufacture NDs encapsulating supra-physiologic boiling point PFH to prevent spontaneous vaporization, and incorporate IONP in the fluorous-phase, to decrease the ADV pressure requirement. Potential applications of this work include targeting droplets to tumors and other biomarkers using small doses that do not saturate the receptor sites, but allow droplet vaporization to aid detection by ultrasound. Further, because ADV occurs in a preselected location using diagnostic pressures, detected resultant MBs would aid in diagnosis that can also be used for image-guided therapy. One potential therapeutic application is when used together with High Intensity Focused Ultrasound (HIFU). ADV generated at low energy can serve as a sound point source that enables time reversal techniques to focus the HIFU pulse to ablate the point where sound was generated [35, 36]. When used with HIFU, the formed MBs will increase local energy absorption lowering the HIFU energy required for ablation, minimizing energy deposition in adjacent tissues [37]. While this can also be accomplished following the IV injection of MBs, ADV of targeted NDs would produce MBs in only the region of interest, which will spare adjacent and intervening tissues since these tissues will not contain MBs.

Our study demonstrates that IONP is truly suspended within the fluorocarbon phase of PFH nanoemulsions, rather than being at the interface with the emulsifier. IONP also acted as nucleation sites decreasing the ADV pressure required vaporization allowing ND detection in vivo. More importantly, since the ultimate goal of this work, which is beyond the scope of this report, is to heat the PFC by exposing IONP to radio frequency (RF), we showed that heating the IONP-loaded PFH ND further reduced the ADV pressure required for ND vaporization.

MATERIALS & METHODS

Oleic acid (OA), iron (II) chloride tetrahydrate (FeCl₂•4H₂O), iron (III) chloride hexahydrate (FeCl₃•6H₂O), and perfluorononanoic acid (PFNA) were purchased from Sigma Aldrich. Perfluorohexane (PFH) was purchased from Synquest Laboratories. Egg yolk phospholipid (l-α-phosphatidylcholine) was purchased from Avanti Polar Lipids. These chemicals were used without additional purification. Degassed Milli-Q filtered water was used for the preparation of all aqueous solutions.

Synthesis of hydrophilic OA-coated IONPs (OA-IONPs)

OA-IONPs (Fe₃O₄) were synthesized by the chemical co-precipitation of Fe(II) and Fe(III) salts with ammonium hydroxide in an aqueous media [38–41]. In order to obtain 1 g of IONP, 0.86 g of FeCl₂•4H₂O and 2.35 g of FeCl₃•6H₂O were dissolved in 40 mL of deionized water with vigorous magnetic stirring under nitrogen gas. As the solution was heated to 80°C, a solution of OA (100 mg, 0.35 mmol) in acetone (5 mL) was added, followed by 28% (w/w) NH₄OH (5 mL) and black precipitates formed quickly. More OA was added to the suspension in five 200 mg increments over 5 min while maintaining the reaction mixture at 80°C for 30 min. The water-based stable suspension was then slowly cooled to room temperature. OA-IONPs were flocculated with a 50:50 v/v acetone and ethanol mixture, centrifuged for 5 min at 1000 rpm, and collected. This step was repeated at least five times to remove excess free OA. After the last wash, OA-IONPs were re-suspended in hexanes/acetone 1:1 v/v to form a stock concentration of 50 mg/mL.

Preparation of fluorophilic PFNA-coated IONPs (PFNA-IONPs)

A large excess of PFNA (0.25 g, 0.54 mmol) was added to the 50 mg/mL OA-IONPs suspension just prepared. The resulting suspension was sonicated for 20 minutes in a bath sonicator (Branson Ultrasonics) to facilitate the dispersion of OA-IONPs in the solvent and the ligand exchange between OA and PFNA. The suspension was left for 3 days at room temperature with slow continuous orbital shaking. The resulting PFNA-IONPs were then washed 5 times by flocculating PFNA-IONPs with a magnet while the solvent was decanted and replaced with 1 ml of fresh hexanes/acetone and sonicated for 10 minutes to remove free PFNA. After washing, PFH (1 mL) was added, forming an infranatant layer. The PFNA-IONPs dispersed homogenously within the PFH infranatant with no visible precipitation or flocculation while using a magnet. The hexanes/acetone supernatant was then carefully collected and discarded to yield PFNA-IONPs in PFH. Dry weight of IONP in PFH was calculated as the mass difference by evaporating the PFH from aliquots, and dilutions of varying weight basis were made from 1 to 15% w/v IONP in PFH from the stock suspension.

Characterization of PFNA-IONPs

Ligand exchange from OA to PFNA was confirmed using Fourier transform infrared spectroscopy (FTIR) (Nicolet 6700 FT-IR Spectrometer, Thermo Scientific) equipped with mercury cadmium telluride (MCT). In addition, thermogravimetric analysis (TGA) (Pyris 1 TGA, Perkin Elmer) was performed to measure the amount of coating used for the samples. For TGA measurements, the particles were precipitated after the ligand-exchange process and kept at 105°C for 24 hours to remove the solvents used during the synthesis. The mass loss of dried sample was monitored under nitrogen gas as the temperature was raised from 30 to 600°C at a rate of 50°C/min.

IONP size was determined by transmission electron microscopy (TEM) (Tecnai G2 Sphera, FEI) with 200 kV acceleration voltage and images were recorded on a 2K × 2K Gatan CCD camera. The samples were prepared at room temperature by applying 3.5 μL diluted suspensions of IONP on a carbon-coated copper grid (Ted Pella, Inc.) that had been glow discharged for 1.5 min at 20 mA using an Emitech K350 glow discharge unit. The sample was allowed to set on the grid for 2 minutes, then washed with 3 drops of glass distilled water. Excess liquid was blotted from the grid using Whatman filter paper and allowed to dry before imaging.

Formulation of IONP-loaded PFH-NDs (IONP-PFH-NDs)

Egg yolk phospholipid (5% w/v) was suspended in phosphate buffered saline (PBS 1X, 640 μL) at pH 7.4 and the desired PFNA-IONP suspension in PFH was added to reach a concentration of 60% w/v (weight of PFH to total volume) by diluting the stock suspension of PFNA-IONP in PFH to a total volume of 360 μL with PFH as needed. The resulting suspension was sonicated using a probe sonicator (Vibra-Cell VCX 130, Sonics & Materials Inc.) at 70% amplitude for 10 seconds while the container was immersed in an ice bath to reduce the heat generated by sonication. After this initial dispersion, sonication was continued in pulse mode at 50% amplitude in cycles of 2 seconds ON and 10 seconds OFF for 30 minutes, which decreased average droplet size.

Characterization of IONP-PFH-NDs

The size of IONP-PFH-NDs was measured using dynamic light scattering (Zetasizer Nano ZS, Malvern Instrument). The measurement was performed at a backscatter angle of 173°. Experiments were done in triplicate and the Z-average, PdI and derived count rate (mean ± standard deviation) were plotted as a function of temperature or time.

CryoTEM was performed with a FEI Tecnai G2 Sphera microscope (FEI, Hillsboro, OR) at 200 kV using low dose settings. Micrographs were collected using a 2k × 2k Gatan CCD camera. For sample preparation, emulsion was diluted to 10 mg/mL lipid in PBS 1X, and 3.5 μL of diluted emulsion was placed upon a Quantifoil^® holey carbon copper grid that had been glow discharged for 1.5 min at 20 mA using an Emitech K350 glow discharge unit. Excess liquid was blotted from the grid using Whatman filter paper, followed by immediate plunge freezing in liquid ethane cooled by liquid nitrogen. The sample was transferred to a Gatan 626 cryo-transfer holder under liquid nitrogen, maintaining the sample below the devitrification threshold throughout the imaging process.

Acoustic Droplet Vaporization (ADV)

To first evaluate the imaging appearance of IONP-PFH-NDs, a suspension of IONP-PFH-NDs containing 5% IONP w/v (approximately 1 × 10⁹ NDs/ml, 250 nm) in PBS 1× (4 mL) was placed in a plastic pipette bulb at 37°C. An arbitrary signal generator was set up to output 10 tone bursts at 0.8 MHz center frequency each containing 40 cycles with a 50% duty cycle. The signal was amplified by an RF amplifier (AG1006, T&C Power Conversion, Rochester NY) and used to drive a focused ultrasound transducer (H-102, Sonic Concepts, Bothell, WA), which will be referred thereafter as the ADV transducer. The transducer is spherically focused at 6 cm from its surface. The tip of the pipette bulb was then positioned at the focal point of the ADV transducer. A Siemens Sequoia^™ C512 scanner equipped with a 15L8 transducer and Cadence Contrast Pulse Sequencing Technology (CPS), a MB specific non-linear imaging technique (Siemens Medical Solutions USA, Mountain View, CA) was used to image the droplet-to-bubble vaporization. The ADV transducer, the pipette containing the nanoemulsion, and the imaging transducer were submerged in degassed water and positioned confocal within the pipette.

Next, as part of an experiment to determine the threshold of the peak negative pressure required for phase conversion, dilute solutions of NDs were acoustically phase converted in water baths at 22, 37, and 42 °C. At each temperature, plain PFH emulsion or PFH emulsion containing 1, 2, 5, 10, and 15% IONP were exposed to ADV pulses (100 cycle sinusoidal pulses at 1.1 MHz) as the pulse amplitude was increased to achieve peak negative pressures ranging from 0 to 4.5 MPa, while all other pulse parameters remained constant. The minimum peak negative pressure at which MBs were detected by imaging was recorded as the emulsion conversion threshold, which were measured three times independently for each emulsion preparation at each temperature.

Ultrasound imaging of IONP-PFH-NDs in vivo

Three nude mice were obtained from our institutional vivarium and implanted in the thigh with K1735 murine melanoma tumors. When tumors were approximately 1 cm in diameter, mice were anesthetized with isoflurane, and IONP-PFH-ND containing 5% w/v IONP was injected intravenously (IV) at a dose of 120 mg/kg PFH (200 μL/kg). Mice were then submerged in the 37°C water bath such that the center of the tumor was at the focal zone of both the imaging and ADV transducers. Immediately after infusion and again 5 hours later, ADV pulses – 40 cycles, 0.8 MHz, and 50% duty cycle, were delivered while imaging the tumor to assess for MB formation. Mice were allowed to recuperate between sessions.

RESULTS AND DISCUSSION

Formulation of PFNA-coated IONP

PFNA-IONPs were synthesized in a two-step process. OA-IONPs were first synthesized by chemical co-precipitation from an aqueous solution of Fe(II) and Fe(III) chloride at basic pH. Then OA molecules coating the IONPs were exchanged with PFNA. While the alkyl portion of OA is hydrophobic and is sufficient to suspend IONP in nonpolar solvents such as hexanes, IONP will flocculate in a fluorinated solvent such as PFH. Therefore, OA was replaced with PFNA, a perfluorocarboxylic acid molecule, to more easily suspend IONP in PFH. This substitution was accomplished by the ligand-exchange process [42, 43] (Figure 1A).

Figure 1.

Characterization of IONP coating with PFNA. A) Schematic representation of the PFNA ligand exchange. B) Photographs of PFNA-IONPs suspended in PBS 1× (left), 70% EtOH (center) and PFH (right). Note that PFNA-IONPs only suspend in PFH. C) TEM image of PFNA-IONPs with a 5× magnified inset showing an approximate 10 nm diameter particle size. D) Representative PFNA-IONP intensity-weighted size distribution as measured by DLS. E) FTIR spectrum and F) TGA heating curve for PFNA-IONPs.

The ligand exchange of OA by PFNA allowed the IONP to fully suspend within PFH, whereas PFNA-coated IONP flocculated and settled in PBS 1X and PBS 1X/Ethanol 30:70 (Figure 1B). Similarly, OA-coated IONP flocculated and settled in PFH. Direct visualization of PFNA-IONP by transmission electron microscopy (TEM) revealed that they were around 5 to 10 nm in diameter, as shown in the inset of Figure 1C. DLS intensity-weighted distribution of PFNA-IONPs in PFH showed a single peak centered at 20.7 ± 3.2 nm with a low polydispersity index (PdI = 0.02) (Figure 1D).

Coating of IONP with PFNA was also confirmed with FTIR and TGA (Figure 1E–F). FTIR spectrum of PFNA-IONP showed a very strong peak at 1054 cm⁻¹ characteristic of the Fe-O stretching vibration in the Fe-O-Fe group (Figure 1E). In addition, a peak was observed at 1208 cm⁻¹ – 1260 cm⁻¹ characteristic of the C-F stretching vibration in the -CF₂- group [44, 45]. Two weak peaks at 1407 cm⁻¹ and 1655 cm⁻¹ were attributed the C-F stretching vibration in the CF₃-CF₂- group and the symmetric stretching vibration of the carboxylate group in the PFNA molecule respectively. PFNA coating of IONPs was also confirmed with TGA (Figure 1F). The TGA curve is characterized by mass losses with distinct steps as the material is heated to 600°C to reveal the mass of the materials being lost [46]. The mass loss profile exhibited a gradual smooth decrease in mass between 220 to 430°C and then remained constant from 430°C to 600°C, which represented the weight percent of now uncoated IONP. Since the mass decrease was smooth starting at the boiling point of PFNA (218°C) without any additional deflections at the boiling point of OA (360°C). These results confirm that all OA molecules were exchanged with PFNA during the ligand exchange process [38, 47].

Formulation of IONP-PFH-NDs

IONP-PFH-NDs were formulated by emulsifying PFH containing various concentrations of PFNA-IONP in an aqueous solution of Egg yolk phospholipid using a tip sonicator (Figure 2A). Emulsions containing IONPs appeared black/brown without any precipitates, and increasing IONP content produced a darker emulsion (Figure 2B). The dynamic light scattering particle size distribution of emulsions containing no IONP and those with 2, 5, and 10% w/v IONP showed that the Z-average diameter was between 150 and 230 nm with a PdI below 0.3 (Figure 2C).

Figure 2.

Characterization of IONP-loaded PFH emulsions. A) Schematic representation of the emulsification of PFH containing IONPs in an aqueous solution of egg yolk phospholipid to yield IONP-PFH-NDs. B) Photograph of vials containing PFH-NDs without (left) and with 1, 2, 3, 7, 10 and 15% w/v IONP. Vials appear darker with increasing concentration of IONP (right). C) Representative particle size distribution of IONP-PFH-NDs without (0%) and with 2, 5, and 10% w/v IONP. D) Representative cryoTEM of a 10% IONP- PFH-NDs emulsion, with IONP particles indicated by white arrows within the fluorocarbon phase (indicated by black arrows) of PFH-NDs . Scale bar is 200 nm.

While the DLS provides a measure of the hydrodynamic diameter of the particles that takes into account the electric dipole layer around the particle, transmission electron cryomicroscopy (cryoTEM) allows for the direct visualization of particles based on their electron density that is represented as gray-scale intensities. As PFH is more electron-dense than water, PFH-encapsulating particles appear darker than the aqueous background or co-existing liposomes in the emulsion. Note, that the cryoTEM micrographs of a 10% w/v IONP-PFH-NDs emulsion demonstrated that PFNA-IONPs are suspended within the inner fluorocarbon phase (Figure 2D).

IONP-loaded PFH emulsions stability at 4°C

We examined IONP-PFH-ND shelf-life by measuring their size using DLS over 14 days of storage at 4°C. We did not observe any significant coalescence or spontaneous vaporization over the time period investigated, demonstrating the emulsion stability. Particle hydrodynamic diameter slightly increased from 200 nm on day 1 to reach a plateau around 250 nm after one week (Figure 3). This limited coalescence is in agreement with the literature, as the diffusion of highly hydrophobic liquid PFC is largely slowed down by the amphipathic phospholipid shell that creates a protective membrane and inhibits emulsion coarsening during storage [48].

Figure 3.

PFH-IONP-NDs hydrodynamic size over 14 days of storage at 4°C as measured by DLS.

IONP decreases the boiling point of PFH

The effect of PFNA-coated IONP concentration upon boiling point – spontaneous droplet vaporization, of encapsulated PFH was determined by measuring the temperature at which bubble formation was first observed (Figure 4A). The temperature required to vaporize PFH encapsulated in ~200 nm nanodroplets is higher (~63°C) than the reported boiling point of neat PFH (56°C) (Figure 4A, 0% IONP) as expected, since emulsifying PFH into NDs increases the interfacial surface tension. However, when incorporating IONP in the PFH core, and increasing its content from 1% to 15% w/v, spontaneous droplet vaporization – PFH boiling, decreased and reached 47°C when ND contained 15% w/v IONP. Since spontaneous droplet vaporization started to plateau at about 50°C with 5% w/v IONP, and since IONP-PFH-ND size decreased when 10% w/v IONP was suspended in PFH (Figure 2C), we chose to use PFH-NDs with 5% w/v IONP loading for the following experiments.

Figure 4.

A) Boiling point of encapsulated PFH as a function of the PFNA-IONP content in PFH-NDs (n = 3). B) Size distribution, Note that the 25°C and 35°C distribution have superimposable traces. C) Mean size and D) PdI of IONP-PFH-NDs at 25, 35, 45 and 55°C.

There was no coalescence – size increase, of the IONP-PFH-NDs as temperature was increased from 25 to 55 °C, which further demonstrates emulsion stability. Rather, there was a slight shift toward smaller sizes as temperature increased (Figure 4B,C), while the PdI remained unchanged (Figure 4D). This slight size decrease can be explained by the likely vaporization and loss of a subpopulation of larger NDs, as indicated in Figure 4C likely due to the lower Laplace pressure.

Ultrasound activation and imaging of IONP-PFH-NDs in vitro

Using the setup shown in Figure 5A, ADV of a suspension of a 5% IONP-PFH-ND in degassed water at 37°C was triggered by the ADV transducer and the resultant ADV imaged with the clinical scanner. The IONP-PFH-NDs suspension was anechoic when imaged before delivery of the converting ultrasound pulse (Figure 5B-before US). ADV was demonstrated after exposure to the first converting pulse, as a highly echogenic MB cloud formed at the focal zone of the converting ultrasound beam (Figure 5B-US pulse 1). The generation of echogenic MBs is further suggested by the acoustic shadowing of the far wall of the transfer pipette. Moreover, the contrast plume ascended against gravity parallel to the clinical transducer while also moving away from the transducer, creating a swirling appearance (Figure 6B-US pulse 2), behavior consistent with buoyancy of MBs and acoustic radiation force applied by the imaging ultrasound field upon MBs [49].

Figure 5.

A) Schematic representation of the experimental setup to trigger ADV and visualize NDs conversion to microbubbles. B) Representative ultrasound images acquired at low MI in CPS mode of a 5% w/v IONP-PFH-NDs at 37°C in degassed water before (Before US) and following two consecutive ADV ultrasound pulses (US pulse 1 and US pulse 2) at 0.8 MHz and 3.5 MPa.

Figure 6.

Effect of ambient temperature and IONP loading on the vaporization threshold. Peak Negative Pressure (PNP) to trigger ADV was measure for PFH-NDs and IONP-PFH-NDs with loadings of 0, 1, 2, 5, 10, and 15% w/v IONP (±standard deviation; n=3).

IONPs in PFH phase of ND provide nucleation sites that lower the peak negative pressure for ADV

Using the setup shown in Figure 5A, we measured the acoustic pressure required to trigger ADV of PFH-NDs containing 0, 1, 2, 5, 10, and 15 % w/v IONP at 22, 37 and 42°C. We gradually increased the peak negative pressure and recorded the pressure at which MBs were first observed by ultrasound imaging (at a low mechanical index so as not to contribute to vaporization events by imaging). As the IONP content and the ambient temperature increased, ADV occurred at lower peak negative pressures (Figure 6). The decreasing pressure threshold required to trigger ADV with increasing temperature was expected, as vapor pressure increases with temperature, making phase transition easier. Increasing IONP content within the ND core also decreased the ADV threshold, suggesting that IONP acted as cavitation nuclei. Specifically the peak negative pressure (PNP) required for ADV at 37°C decreased from 3.5 MPa for pure PFH-ND, to 3.02 MPa for 2% IONP PFH-ND, to 2.16 MPa for 15% IONP PFH-ND. At 42°C, a temperature that could be encountered in regions of high metabolic activity, 10% IONP PFH-ND and 15% IONP PFH-ND vaporized at thresholds of 1.96 and 1.30 MPa, respectively. It is important to note that 1.96 MPa at 1.1 MHz corresponds to a mechanical index of 1.86, which is less than the maximum mechanical index of 1.9 allowed by the FDA for diagnostic ultrasound imaging equipment [31], although a 100 cycle pulse was used in these experiments whereas most diagnostic imaging is accomplished with single cycle pulses.

Ultrasound imaging of IONP-PFH-NDs in vivo

A 60% w/v IONP-PFH-ND (5% w/v IONP) suspension in PBS 1X was injected intravenously at 120 mg/kg PFH in melanoma-bearing nude mice to assess whether ADV would occur in vivo (Figure 7). No MBs were detected before exposure to an ADV pulse either immediately post infusion (Figure 7C) or 5 hours later (Figure 7E). This confirms that no spontaneous vaporization occurs in vivo. While no MBs were observed following an ADV pulse immediately after injection (Figure 7D), MBs formed in the tumor at 5 hours after ND administration (Figure 7F). NDs shortly after infusion are limited to the vascular space and become diluted in the entire blood volume. Therefore, given that mice have ~80 mL of blood per kg, the 120 mg/kg PFH dose would result in a blood concentration of 1.5mg/mL of blood. Since tumors have a fractional blood volume of about 3-5%, we would expect PFH tumor concentration to be 45-75μg/g. This concentration is expected to increase by 5 hours via the EPR effect as we have previously shown using a different PFC in human subjects[21]. That ADV occurred at only 5 hours after infusion indicates that NDs were extravascular, and likely at higher concentration than immediately post infusion. Whether ADV occurred at 5 hours because of higher concentration or because NDs were stationary in the extravascular space is not clear. PFH tissue concentration was not measured in this pilot in vivo study.

Figure 7.

A) Schematic representation of the setup used to observe ADV in vivo. B) Representative picture of a melanoma-bearing mouse. C-F) US images of the melanoma tumor in a live mouse. In each panel, dual display shows the microbubble-specific imaging (CPS) mode (left image), with the conventional B-mode image reconstructed from the same backscatter dataset. Ultrasound images of the melanoma tumor acquired immediately after IONP-PFH-ND injection C) before and D) after exposure to ADV pulses (0.8MHz at 5 MPa nominally, likely 3-4 MPa at the transducer focal zone due to acoustic attenuation of tissue) shows no microbubble formation. US images taken 5 hours after IONP-PFH-ND injection E) before and F) after exposure to the same ADV pulse shows microbubbles that formed within the tumor after ADV (red circle in F). Vertical lines in D and F (white arrows) are radiofrequency interferences from the ADV pulse.

While this preliminary study did not assess for toxicity, animals were normal up to the 5-hour observation period and following the ADV experiments. Note, PFC emulsions have been used in human subjects for tumor imaging at much higher doses in clinical trials, where the only observed side-effect was a flu-like syndrome shown to be due to the large particle load [21].

CONCLUSION

We aimed to induce PFC phase transition using nanometer rather than micrometer sized NDs to allow for extravascular targeting. Since decreasing droplet size would require higher ultrasound pressure for phase transition, we incorporated IONP within the fluorous-phase to act as nucleation sites to decrease the ADV pressure requirement. We successfully synthesized IONP coated with PFNA to produce fluorophilic IONPs that suspended within PFH, which we subsequently emulsified into stable ~200 nm IONP-containing PFH-NDs. We showed that as temperature and IONP ND content increased, the pressure required to convert PFH-NDs to MBs decreased. Using PFH-NDs loaded with 5% w/v IONP, we were able to trigger ADV at physiological temperatures at the high end of the FDA-approved mechanical index of clinical ultrasound scanners. When IONP-PFH-NDs that passively accumulated in tumors by 5 hours after IV administration were exposed to an ADV pulse in vivo, they converted to MBs detectable by ultrasound. Mice tolerated the infusion and ADV with no obvious toxicity. This represents the first in vitro and in vivo demonstration of fluorophilic IONPs used as a doping material to lower the ADV threshold of supraphysiologic boiling point PFC emulsion droplets, allowing the detection of extravasated NDs. Our findings are in agreement with the known concept that particles within a liquid provides nucleation sites that lower the ultrasound cavitation threshold [32, 50, 51].

While there is recent precedent for including lipophilic nanoparticles in fluorocarbon emulsions, such inclusions sequester either into the hydrocarbon chains or at the interface of surfactants and the fluorous phase [52, 53]. In contrast, we showed that PFNA coating of IONP allowed them to reside within the fluorous phase itself. In addition, suspending IONP within the fluorous phase maximized the surface area of the IONP interaction with the PFC, which is expected to translate to more optimal thermal transfer if inductive heating were to be used to lower the ADV threshold.