Ultrasound-sensitive nanodroplets achieve targeted neuromodulation

Abstract

Focused ultrasound (FUS) offers an attractive tool for non-invasive neuromodulation, addressing a clinical need to develop more minimally invasive approaches that are safer, more tolerable and versatile. In combination with a cavitation agent, the effects of ultrasound can be amplified and localized for therapy. Using c-Fos expression mapping, we show how ultrasound-sensitive nanodroplets can be used to induce either neurosuppression or neurostimulation, without disrupting the blood-brain barrier in rats. By repurposing a commercial ultrasound contrast agent, Definity, lipid-shell decafluorobutane-core nanodroplets of 212.5 ± 2.0 nm were fabricated and loaded with or without pentobarbital. FUS was delivered with an atlas-based targeting system at 1.66 MHz to the motor cortex of rats, using a feedback-controller to detect successful nanodroplet vaporization and drug release. Neuromodulation was quantified through changes in sensorimotor function and c-Fos expression. Following FUS-triggered delivery, sham nanodroplets induced a 22.6 ± 21% increase in local c-Fos expression, whereas pentobarbital-loaded nanodroplets induced a 21.7 ± 13% decrease (n=6). Nanodroplets, combined with FUS, offer an adaptable tool for neuromodulation, through local delivery of small molecule anesthetics or targeted mechanical effects.

Graphical Abstract

1. Introduction

Neuromodulation describes the process of altering neuronal behaviour by exciting or suppressing nerve activity and can be used for both investigating and treating the nervous system. Clinically, common forms of neuromodulation include implantable devices like deep brain stimulation (DBS) and spinal cord stimulators, and non-invasive devices that exploit electromagnetic energy through external electric or magnetic fields, known as transcranial electric stimulation (tES) (1) and transcranial magnetic stimulation (TMS) (2). However, these current techniques suffer from invasiveness (such as implanting electrodes (3)), poor accuracy (such as diffuse external electrical and magnetic fields (4)) and limited ability to penetrate deep targets (such as decaying magnetic field gradients (5)). In contrast, focused ultrasound (FUS) has the potential to achieve neuromodulation with precise non-invasive targeting to regions deep within the brain (6). Furthermore, utilizing a FUS system for targeted delivery of psychoactive drugs, with high spatial and temporal control, could offer unique advantages over current neuromodulation techniques.

Over the last decade the treatment of neurological disorders using FUS-mediated neuromodulation has been explored, including epilepsy in both mouse (7) and monkey (8) models, Alzheimer’s (9) and chronic pain (10) in patients. Using microbubbles in conjunction with FUS to open the blood-brain barrier (BBB) has also been shown to coincide with neuromodulation, in rat models (11,12), in non-human primates (13) and in human subjects (14). Furthermore, cavitation-mediated neurostimulation without compromising the BBB has also been shown in a mouse model (15,16). However, an adequate understanding of the underlying mechanisms necessary for consistent application remains elusive.

Ultrasound-triggered pentobarbital release from intravenously injected nanodroplets combines the non-invasiveness and spatial advantages of FUS with the reliability of psychoactive pharmaceuticals. Like gaseous microbubbles, liquid-filled nanodroplets also respond to ultrasound. Ultrasound-sensitive nanodroplets, also known as nanoemulsions, have grown in popularity as a tool for diagnostic ultrasound (17) and ultrasound-based therapies (18). Nanodroplets offer an attractive alternative to microbubbles because of their sub-micron size (19), high stability in circulation (20) and drug-loading capabilities (21). Previously, the delivery of propofol was achieved using ultrasound-responsive nanodroplets (22), followed by our study demonstrating pentobarbital delivery using nanodroplets fabricated from a commercially-available ultrasound contrast agent (23), inducing localized neuromodulation without disrupting the BBB.

Pentobarbital is a GABA_A receptor agonist, potentiating inhibitory tone as well as suppressing glutamatergic neurotransmission. It is a small molecule drug and therefore can pass through the intact BBB. Since it has numerous undesirable side-effects in systemic administration, such as respiratory and widespread central nervous system (CNS) depression, nanodroplets as drug-carriers protect the systemic circulatory system from pentobarbital until triggered release following FUS.

Here we show it is feasible to deliver pentobarbital with anatomic precision without an MRI system, which can be cumbersome and costly in clinical applications. This is achieved using a brain atlas-based targeting system, real-time acoustic monitoring, and a transducer with millimetre-scale precision. We conduct brain activity mapping based on c-Fos immediate early gene (IEG) expression to study the effects of drug-loaded and sham nanodroplets on neural activation. Behavioural assays and c-Fos mapping are used to validate neuromodulation. This work will bring the system closer to clinical translation.

2. Methodology

2.1. Ultrasound-responsive nanodroplets

Materials

Decafluorobutane (DFB) (C₄F₁₀) was purchased from Synquest Labs, USA. Pentobarbital solution in methanol was purchased from Sigma Aldrich (Millipore Sigma, US). Definity, a commercially available FDA-approved ultrasound contrast agent, was purchased from Lantheus Medical Imaging, USA. In its native form, Definity has a gaseous octafluoropropane (OFP) (C₃F₈) core that is contained by a lipid shell composed of (R)-4-hydroxy-N,N,N-trimethyl-10-oxo-7-[(1oxohexadecyl)oxy]-3,4,9-trioxa-4-phosphapentacosan-1-aminium, 4-oxide (DPPC); (R)-hexadecanoic acid, 1-[(phosphonoxy)methyl]-1,2-ethanediyl ester (DPPA); and (R)-∝-[6-hydroxy-6-oxido-9-[(1-oxohexadecyl)oxy]-5,7,11-trioxa-2aza-6-phosphahexacos-1-yl]- ω-methoxypoly(ox-1,2-ethanediyl) (MPEG5000 DPPE), at a weight ratio of 82:10:8, giving a lipid concentration of 0.75 mg/mL. Milli-Q ultrapure water (Millipore Sigma, US) was used throughout the fabrication.

Fabrication

Definity-based nanodroplets with lipid shell (DPPC:DPPA:MPEG5000 DPPE) and liquid DFB core, loaded with pentobarbital (25 μg/mL), were prepared using a modified condensation technique, as previously described (23), based on the protocol developed by Sheeran et al. (24). In brief, 200 μL of pentobarbital solution in methanol was transferred into ultrapure water (1mg/mL) and combined with 1.5 mL of Definity lipid solution (0.75mg/mL) using a tip sonicator (S-450D, Branson Ultrasonics, USA) with 3 mm tip at 10% power, 20 s of sonication pulsed at 1 s on, 1 s off. The solution was crimp-sealed in a 3 mL headspace vial (Wheaton, DWK Life Sciences, US) and connected to a vacuum pump, thoroughly degassed to remove the native OFP and then the headspace filled with DFB.

Precursor microbubbles were formed by agitating the vial for 45 s using a VialMix agitator (Lantheus Medical Imaging, USA). A bath of isopropanol was cooled to −10°C with dry ice and the vial submerged and swirled gently for 2 minutes. An additional 1 mL of air was injected into the headspace to increase internal vial pressure and ensure complete condensation of precursor microbubbles into nanodroplets. Remaining free pentobarbital and microbubbles were removed through centrifugation at 200 G for 5 minutes at 4°C, repeated three times. The final nanodroplet solution was passed slowly through a cold 0.8 μm sterile syringe filter (Minisart Syringe Filter, Sartorius, Germany) to remove erroneous large droplets.

Characterization

The size distributions of three independent batches of nanodroplets were measured using nanoparticle tracking analysis (NanoSight, Malvern Panalytical, UK), where three movies of 30 s duration were acquired of the sample under Brownian motion at 1:10 dilution. Resultant microbubbles were sized using a Coulter Counter (Multisizer 3, Beckman Coulter, USA) with 30 μm aperture, where three independent batches of nanodroplets were left to spontaneously vaporize in room-temperature, filtered, phosphate-buffered saline solution (Beckman Coulter™ ISOTON™ II Diluent). The same measurement procedure was applied to a vial of Definity microbubbles to compare microbubble size distributions.

Acoustic emissions and drug release from pentobarbital-loaded nanodroplets were measured in vitro using the same FUS system utilized for in vivo experiments (described in the following section). Three ultrasonic frequencies (0.58 MHz, 1.66 MHz and 1.78 MHz) were compared in vitro to assess the influence of sonication frequency on the pressure required to induce droplet vaporization. The onset of vaporization was detected by monitoring subharmonic emissions. 0.58 MHz and 1.66 MHz were compared in vivo to assess the vaporization threshold when flowing in the cerebral vasculature.

To quantify drug release in vitro, the sample of drug-loaded nanodroplets was pushed through a tube with internal diameter of 1.1 mm, sitting in a bath of degassed deionized water aligned below the transducer. The aqueous suspension of nanodroplets was sonicated at 1.66 MHz with bursts of length 1 ms, 2 ms, 5 ms or 10 ms at 2.0 MPa. The released drug was extracted from the effluent using an organic solvent sink of hexane and ethyl acetate (1:9 volume ratio) and transferred into ethanol for UV-vis measurements (NanoDrop 2000c, ThermoFisher Scientific).

2.2. Brain atlas-guided focused ultrasound

Animals

Male Sprague Dawley rats (n = 30) were purchased from Taconic Biosciences (Germantown, NY, USA) and had a mean weight of 351 ± 83 g on the day of experimentation. Rats were divided at random between five experimental groups, with 2 additional animals purchased later for histological safety assessment (table 1). Animals were housed in the Sunnybrook Research Institute animal facility (Toronto, ON, Canada) on a reverse light cycle and had access to food and water ad libitum. All animal procedures were approved by the Animal Care Committee at Sunnybrook Research Institute and are in accordance with the Canadian Council on Animal Care and ARRIVE guidelines.

Table 1.

Distribution of animals between treatment groups.

Treatment group	n
Behavioural assessment (pentobarbital nanodroplets with FUS)	6
Behavioural assessment (sham nanodroplets with FUS)	6
c-Fos expression (pentobarbital nanodroplets with FUS)	6
c-Fos expression (sham nanodroplets with FUS)	6
c-Fos expression (FUS only)	4
Histological safety assessment	2

Animal preparation

General anesthesia was induced using 5% isoflurane (ISO), then maintained at 2% for the duration of the FUS procedure. The level of general anesthesia was assessed using the pedal reflex (firm toe pinch). The scalp of the rat was shaved, and the remaining hair was removed with depilatory cream to avoid air bubbles and ensure sufficient coupling with a thin layer of ultrasound gel (Wavelength CL, Ontario, Canada). A 22-gauge tail vein catheter was placed, and the animal was positioned prone underneath the transducer, breathing into a nose cone secured with a bite bar. The animal’s skull was aligned and fixed in place using the stereotactic frame that is integrated in the brain atlas-guided focused ultrasound system described below. A warm saline bag was placed on the animal torso to maintain body temperature during sonication.

Brain atlas-guided focused ultrasound

Focused ultrasound (FUS) was delivered using the pre-clinical prototype system (an early prototype for the RK50, FUS Instruments Inc., Canada), which uses the Waxholm Space Atlas of the Sprague Dawley rat brain, co-registered with the focus of a single element focused transducer (centre frequency 1.66 MHz, element diameter 43 mm, focal length 35 mm) and narrowband PZT (lead zirconate titanate) hydrophone (centre frequency 830 kHz), using a stereotactic positioning frame. The stereotactic frame is integrated in the pre-clinical prototype FUS system and consists of adjustable ear bars, bite bar and nose cone (similar in design to the Kopf stereotaxic frame). The alignment of the transducer focus and motorised positioning system are calibrated based on the geometry and orientation of the integrated frame. The transmit transducer was characterized using a fibre-optic hydrophone (Precision Acoustics, UK) and the pressure field was mapped with 0.1 mm resolution. The full width at half maximum (FWHM) of the ultrasound focus had a volume of 0.9 × 0.9 × 4.5 mm (shown with the white dashed contour in figure 1B), allowing higher spatial specificity than in the previous study (23).

Figure 1. Brain atlas-guided focused ultrasound system.

(A) Schematic of the prototype RK50 atlas-based FUS-targeting system, (B) lateral and axial ultrasound pressure field maps with FWHM contour shown in white (0.9 × 0.9 × 4.5 mm), (C, D, E) screenshots from the positioning software showing three sonications with five target spots in each.

Ultrasound was delivered at 1.66 MHz in 10 ms pulses with a pulse repetition frequency of 1 Hz. Sonication pressure was determined using a modified acoustic feedback algorithm detecting droplet activation based on the detection of subharmonic emissions by the narrowband PZT hydrophone, as used previously in monitoring safe microbubble-mediated BBB opening (25). The threshold for droplet activation was defined as the pressure required for subharmonic emissions to exceed 3.5 times the baseline subharmonic signal. A pressure ramp was used to detect the onset of droplet activation, starting at peak negative pressure of 1.1 MPa where 10 s of baseline measurements were recorded prior to droplet injection. After injecting nanodroplets, ultrasound pressure was increased by 9.6 kPa each second until droplet vaporization was detected, when the pressure was fixed for the remainder of the 180 s sonication. For the FUS only group, a fixed pressure of 1.74 MPa was used based on the average pressure required for droplet activation as recorded from the 12 sham and treatment rats. All reported in vivo pressure values are derated, assuming 62% transmission through the skull of a 400 G rat at 1.66 MHz (26).

Treatment scheme

Three sequential injections of pentobarbital-loaded nanodroplets (500 μL, approximately 4×10¹⁰ droplets containing 12 μg of pentobarbital) were delivered via the tail vein catheter in a slow bolus injection lasting approximately 60 s, followed by a 250 μL saline flush each time. FUS exposure lasted 180 s, covering five focal spots during each of the three sonications. Injections were staggered by 10 minutes to ensure clearance of the previous injectate, based on acoustic emissions indicating nanodroplet circulation half-life of 8.2 ± 0.5 minutes (23). Sonication targets were shifted for each exposure to cover a larger region of the frontal cortex, including motor and sensory regions (coordinates relative to bregma are given in table 2 and illustrated in figure 1C–E).

Table 2. Sonication target coordinates.

Coordinates from bregma of five target locations for three sonications.

	Sonication 1			Sonication 2			Sonication 3
Target	L/R (mm)	A/P (mm)	D/V (mm)	L/R (mm)	A/P (mm)	D/V (mm)	L/R (mm)	A/P (mm)	D/V (mm)
1	−2.5	−1.0	2.0	−2.0	1.0	2.0	−1.5	3.0	2.0
2	−3.5	−1.0	2.0	−3.0	1.0	2.0	−2.5	3.0	2.0
3	−4.5	−1.0	2.0	−4.0	1.0	2.0	−3.5	3.0	2.0
4	−4.5	−2.0	2.0	−4.0	0.0	2.0	−3.5	2.0	2.0
5	−3.5	−2.0	2.0	−3.0	0.0	2.0	−2.5	2.0	2.0

Safety assessment

MR imaging with a 7-tesla small bore MRI scanner (BioSpec 70/30 USR, Bruker, Billerica, MA, USA) was carried out on a subset of 6 behavioural assessment animals (3 pentobarbital nanodroplets with FUS, 3 sham nanodroplets with FUS) to assess the safety of the FUS treatment and to monitor potential bioeffects. MR imaging was carried out immediately following treatment, prior to behavioural testing. Anaesthesia was maintained following treatment at 2% ISO whilst in the MRI scanner.

Blood-brain barrier (BBB) integrity was assessed using T1-weighted contrast-enhanced MR imaging (4000 ms repetition time (TR), 14 ms echo time (TE), 256 × 256 matrix, 1.5 mm slice thickness), immediately following an injection of a gadolinium-based contrast agent (0.2 mL/kg, Gadovist, Schering AG, Germany). T2-weighted imaging (1500 ms TR, 6 ms TE, 256 × 256 matrix, 1.5 mm slice thickness) and T2*-weighted imaging (800 ms TR, 3 ms TE, 256 × 256 matrix, 1.5 mm slice thickness) were used to identify edema and haemorrhage respectively.

2.3. Histology

Following sonication, rats were kept in their home cages in a quiet room for 90 minutes. Rats were then intracardially perfused with ice cold normal saline, followed by 4% paraformaldehyde (PFA). Brains were removed, post-fixed in 4% PFA for 24 hours, and saturated in 30% sucrose 0.1 M phosphate buffer. Brains were cryosectioned axially at 40 μm and stored in cryoprotectant (30% ethylene glycol and 25% glycerol in phosphate buffer) at −20°C.

From each of the c-Fos expression experimental and control groups, a series of sections (1 in 4) from AP 3 mm to −2 mm was selected for immunofluorescent staining. Sections were washed in PBS with 0.3% Triton-X and quenched in 0.5% w/v sodium borohydride for 10 minutes at room temperature (RT). Sections were treated with blocking 10% goat serum for 1 hour at RT and incubated in primary antibody (anti-cfos rabbit anti-mouse IgG, 1:1000, Abcam, #ab190289) for 24 hours at 4°C. Then, sections were placed in secondary antibody (Cy™3-conjugated goat anti-rabbit IgG, 1:200, Jackson ImmunoResearch Laboratories, #711–166-152) for 2 hours at RT. Counterstaining with DAPI (1:10 000) was also performed. Sections were washed with phosphate buffer and mounted on Superfrost Plus slides (Fisher Scientific, Canada) and dried overnight.

c-Fos expression of the right (untreated) and left (treated) motor cortex was quantified by spinning disk confocal microscopy (Zeiss Axio Observer Z1; Carl Zeiss, Germany) under 10x objective and the same exposure settings. Nuclear positive cells within the motor cortex were counted semi-automatically using ImageJ thresholding and particle analysis. The basic ImageJ software was used without any additional plugin. The threshold function was applied to the same percentage of the histogram, followed by particle analysis with the same size and circularity parameters applied across all images. Wilcoxon signed rank test was used to identify differences between hemispheres with an alpha of 0.05. A representative microscope image overlayed with the motor cortex region of interest (yellow) and location of sonication targets (red) is shown in figure 5A. The region of interested was selected to cover the m1 and m2 regions of the motor cortex, as shown in the corresponding brain atlas (figure 5A left panel).

Figure 5. Quantifying c-Fos expression following FUS treatment.

(A) Diagram of c-Fos expression quantification within the motor cortex (yellow contour), overlaid with microscopy and sonication targets (red contours). Representative immunofluorescent stained sections acquired with spinning disk confocal microscopy under 10x objective, (B) no focused ultrasound (FUS), (C) FUS only, (D) FUS with sham nanodroplets and (E) FUS with pentobarbital-loaded nanodroplets (scale bar 200 μm). (F) Absolute c-Fos expression data showing number of expressing cells in each microscope tile (sonicated cortex shown with solid colour bar, contralateral cortex shown with stripped bar), errors bars show one standard deviation (*ANOVA p<0.05). Mean second harmonic emissions detected from vaporizing nanodroplets correlated with the difference in c-Fos expression between cortices, when treated with (G) sham nanodroplets and FUS (n=6), and (H) pentobarbital-loaded nanodroplets and FUS (n=6). (I) Difference in c-Fos expression comparing sonicated and non-sonicated regions 90 minutes after treatment with FUS only (n=4), sham nanodroplets with FUS (n=6), or pentobarbital-loaded nanodroplets with FUS (n=6), errors bars show one standard deviation (*Wilcoxon signed rank test p<0.05). Peak subharmonic emissions correlated with difference in c-Fos expression when treated with (J) sham nanodroplets and FUS (n=6), and (K) pentobarbital-loaded nanodroplets and FUS (n=6).

For histological assessment of treatment safety, 2 rats were treated with FUS and sham nanodroplets, or FUS and pentobarbital-loaded nanodroplets, and intracardially perfused 90 minutes after sonication as previously described. Brains were excised, immersed in 10% neutral buffered formalin for 24 hours, then transferred into 70% ethanol for 48 hours before being embedded in paraffin. 5 μm thick axial sections were taken at 500 μm separation. Hematoxylin and eosin (H&E) staining and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining (DeadEnd™ Colorimetric TUNEL System, Promega) were performed on tissue sections from one sham nanodroplet and one pentobarbital-loaded nanodroplet animal to identify signs of hemorrhage, edema, or apoptotic cells.

2.4. Sensorimotor deficit

Sensorimotor deficit was assessed through simple gait analysis comparing before and 90 minutes after treatment with pentobarbital-loaded or sham nanodroplets. The hind paws of the rat were dipped in paint (Crayola Washable Fingerpaint), before walking along a channel (1 m x 9 cm x 15 cm) lined with paper, with a dark enclosure at the far end. The footprints were used to measure stride length, base width, paw angle and gait abnormalities. Changes in paw angle following sonication were assessed with a paired t-test. Similar methods have been used in the assessment of asymmetric motor function in the context of traumatic brain injury (27).

3. Results

3.1. Ultrasound-responsive nanodroplets

Nanodroplets were found to have an average diameter of 212.5 ± 2.0 nm, as determined by NanoSight measurements of three independent batches (figure 2B). Resultant microbubbles produced by the nanodroplets were found to have a mean diameter of 1.13 ± 0.1 μm (figure 2C), where 90% of resultant bubbles were less than 1.46 ± 0.1 μm in diameter. This resultant bubble size gives an average diameter expansion ratio of 5.3, comparable to that recorded by Sheeran et al. (28) using ultra-high-speed video measurements of vaporizing DFB nanodroplets.

Figure 2. Characterization of ultrasound-responsive nanodroplets.

(A) Schematic of decafluorobutane (DFB) nanodroplet with lipid coating, loaded with pentobarbital (PB). Size distribution measurements of (B) nanodroplets (n=3), and (C) resultant microbubbles (MBs) produced by vaporizing nanodroplets (n=3) compared to Definity microbubbles (MBs). Acoustic response of nanodroplets (D) in vitro and in vivo at three sonication frequencies, (E) vaporization threshold (Vp) determined by the onset of subharmonic emissions (3.5-times baseline) with varying pulse length at 1.66 MHz and (F) corresponding percentage of pentobarbital (PB) released at 2.0 MPa at 1.66 MHz. All error bars represent one standard deviation.

When compared to native Definity microbubbles, measured within 30 minutes of activation, the mean bubble diameter was similar to the resultant microbubbles from vaporizing nanodroplets. However, the distribution of bubble size was significantly different, as shown in figure 2C and quantified in table 3. The fresh population of Definity contained a much wider distribution of bubble diameters, skewed towards larger diameters, with a maximum measured bubble diameter of 11.05 ± 1.6 μm compared to 2.29 ± 0.5 μm for nanodroplet-derived bubbles.

Table 3. Distribution of mean bubble diameters.

Average Coulter Counter size measurements from three samples of Definity microbubbles (Definity MBs) and nanodroplet-derived microbubbles (ND-derived MBs) following vaporization, showing mean diameter and one standard deviation.

	Definity MBs (μm)	ND-derived MBs (μm)
Mean diameter	1.4 ± 0.2	1.1 ± 0.1
90th percentile	2.1 ± 0.4	1.5 ± 0.1
95th percentile	2.7 ± 0.5	1.6 ± 0.2
99th percentile	4.5 ± 0.8	1.8 ± 0.2
Maximum diameter	11.1 ± 1.6	2.3 ± 0.5

The extent of nanodroplet vaporization is modulated by the magnitude of ultrasound pressure. The pressure required to induce vaporization – known as the vaporization threshold – was found to increase with increased sonication frequency and increase due to the in vivo environment (even when pressures were derated to account for skull attenuation) (figure 2D). Increasing burst length fractionally decreased the vaporization threshold in vitro, as determined by the onset of subharmonic emissions with increasing pressure (figure 2E). Increasing burst length from 1 ms to 10 ms was found to significantly increase the quantity of pentobarbital released but came to a plateau at 10 ms (figure 2F).

3.2. Safety assessment

To assess treatment safety and BBB integrity following sonication with nanodroplets, a series of MR images were acquired (figure 3). T1-weighted contrast-enhanced images using a gadolinium-based MRI contrast agent showed no sign of BBB permeability (figure 3B, 3F). Furthermore, there was no evidence of edema or haemorrhage on T2-weighted (figure 3C, 3G) and T2*-weighted (figure 3D, 3H) images, respectively, following sonication with sham or pentobarbital-loaded nanodroplets. Furthermore, H&E staining revealed no signs of edema or hemorrhage following sonication with sham or pentobarbital-loaded nanodroplets (figure 4A–D), and TUNEL staining revealed no apoptotic cells in any of the sonicated regions (figure 4E–H). These findings are in agreement with our previous work (23), and have here been shown to hold true for a higher nanodroplet dose and with sequential sonications.

Figure 3. MRI evidence of safe FUS-mediated drug delivery.

(A, E) FUS target locations covering the left cortex (bregma shown with a cross), (B, F) T1-weighted contrast-enhanced images showing intact BBB and (C, G) T2-weighted images showing no indication of edema or (D, H) haemorrhage on T2*-weighted images following focused ultrasound with sham and pentobarbital-loaded nanodroplets, respectively (scale bars 5 mm).

Figure 4. Histological evidence of safe FUS-mediated drug delivery.

(A-D) Representative H&E-stained sections acquired under 20x magnification comparing control (non-sonicated) (A, C) and sonicated (B, D) regions following focused ultrasound with sham and pentobarbital-loaded nanodroplets, respectively, showing no signs of RBC extravasation or edema. (E-H) Representative TUNEL-stained sections acquired under 20x magnification comparing control (non-sonicated) (E, G) and sonicated (F, H) regions following focused ultrasound with sham and pentobarbital-loaded nanodroplets, respectively, showing no visible apoptotic cells (scale bars 100 μm).

3.3. c-Fos expression

c-Fos is an IEG that is expressed rapidly and transiently in response to cellular stimulation, such as depolarization. Modulation of neuronal firing was assessed post-sonication by quantifying the difference in c-Fos expression comparing the sonicated and non-sonicated motor cortex, following FUS treatment without nanodroplets, with sham nanodroplets or with pentobarbital-loaded nanodroplets (figure 5A–K). In the presence of pentobarbital-loaded nanodroplets, we found a 21.7 ± 13% reduction in c-Fos positive cells (Wilcoxon signed rank test p<0.001) in the sonicated hemisphere relative to the contralateral, suggesting an overall inhibitory effect (figure 5E, 5I). The magnitude of neuronal suppression was found to correlate with peak subharmonic emissions (figure 5K), as indicated by linear regression analysis (Pearson linear regression R²=0.75), following treatment with pentobarbital-loaded nanodroplets, but did not correlate with second harmonic emissions (R²=0.23) (figure 5J). This correlation suggests drug release and corresponding biological effects may be predicted and monitored with acoustic emissions from vaporizing nanodroplets, as supported by in vitro findings (23).

In contrast, sonication with sham nanodroplets increased c-Fos expression by 22.6 ± 21% (Wilcoxon signed rank test p<0.001) (figure 5D, 5I). The magnitude of increase was found to correlate most significantly with second harmonic emissions (figure 5G), as indicated by linear regression analysis (Pearson linear regression R²=0.85), but did not correlate with subharmonic emissions (R²<0.01) (figure 5H).

Sonication following an injection of saline (FUS only) produced a 4.1 ± 2% increase in c-Fos positive cells in the sonicated hemisphere relative to the contralateral. This increase was not found to be statistically significant (Wilcoxon signed rank test p<0.43) (figure 5C, 5I).

When comparing absolute number of c-Fos expressing cells in the contralateral (non-sonicated) hemispheres from animals that received FUS alone, FUS with sham, or FUS with pentobarbital-loaded nanodroplets, no significant differences were found between groups (ANOVA, p>0.05) (figure 5F). This finding suggests that nanodroplets and pentobarbital had no effect outside of the sonicated region. However, when comparing absolute number of c-Fos expressing cells in the sonicated hemispheres, FUS with pentobarbital-loaded nanodroplets significantly decreased expression compared to FUS alone (ANOVA, p<0.01).

3.4. Sensorimotor deficit

A distinct gait pattern was seen 90 minutes following FUS with pentobarbital-loaded nanodroplets, indicating asymmetric anesthesia of the motor cortex (figure 6). All six rats treated with pentobarbital-loaded nanodroplets showed the same characteristic gait pattern following sonication, namely stumbling onto the ipsilateral side shown by overlapping paw prints, increased paw angle, and faint prints on the contralateral side (figures 6A, 6B). In contrast, rats treated with sham nanodroplets showed no measurable change in gait. Stride length was not significantly influenced (figure 6C), but paw angle was found to be significantly altered following treatment with pentobarbital-loaded nanodroplets and FUS (figure 6D) (one-way ANOVA p<0.05). Similarly, the difference between sonicated and contralateral paws following treatment with pentobarbital-loaded nanodroplets was also found to be significant (figure 6E) (one-way ANOVA p<0.05), but this was not the case for the sham nanodroplet group. All animals returned to baseline behaviour when assessed at 24 hr following treatment.

Figure 6. Asymmetric gait change showing a sensorimotor deficit.

Following sonication of the cortex, deficit is shown through paw prints from rats (A) pre- and (B) post-treatment with pentobarbital-loaded nanodroplets and FUS (PB ND). Gait pattern is summarized as (C) stride length, (D) paw angle and (E) change in paw angle following sonication with pentobarbital-loaded nanodroplets (PB ND) or sham nanodroplets (sham ND), for the sonicated (black) and contralateral sides (white), n=6. Error bars show one standard deviation (*ANOVA p<0.05).

4. Discussion

In this work we have shown for the first time that nanodroplets can be used for both neurosuppression – through the delivery of an anesthetic – and neurostimulation – through cavitation activity, without disrupting the BBB in rats. Definity microbubbles were repurposed to form a versatile nano-scale tool with theranostic capabilities, achieving localized drug delivery or targeted mechanical effects, whilst being monitored in real-time through acoustic emissions.

Atlas-based targeting establishes the feasibility of using pre-existing MRI scans. This, together with the demonstration of safe and successful delivery via acoustic monitoring, obviates the need for MRI coupling, which will be expensive and cumbersome in the clinic. Without MRI coupling, this new platform of FUS neuromodulation is scalable to more clinical applications and settings.

Previously, ultrasound-responsive nanodroplets have been used for targeted delivery of propofol (22,29). This approach to FUS neuromodulation takes advantage of the highly reliable and reproducible effect of pharmacologic agents. Pentobarbital, in comparison, has a longer duration of action – hours as opposed to minutes for propofol. The offline effect of a pentobarbital system is critical for practical use for neuromodulation in the clinic. Furthermore, pentobarbital is known to have a neuroprotective effect. It has historically been used to treat severe, resistant focal cerebral ischemia and increased intracranial pressure. The mechanism is not clear but thought to be reduced cerebral metabolic rate, mitigating calcium accumulation and glutamatergic excitotoxicity (30).

Furthermore, we chose to repurpose a commercial microbubble formulation, Definity – used clinically for diagnostic ultrasound and in clinical research for therapy – to enhance clinical relevance of this system. Previous studies have used agents formed from polymers (21,22), that not only require greater ultrasound pressures for vaporization, but also use experimental materials that have not been used clinically.

Overall, we have developed a flexible system for non-invasive neuromodulation with high readiness for clinical translation.

4.1. Ultrasound-responsive nanodroplets

Previous studies have illustrated how vaporization threshold – the pressure required to convert a nanodroplet into a microbubble – is influenced by sonication frequency (31), pulse length (32) and the in vivo environment (33). Our results are in agreement with these studies, showing that the lowest vaporization threshold is found when the sonication frequency is low and the pulse length is long. However, low sonication frequencies produce larger ultrasound foci, limiting targeting precision (particularly on the scale of the rat brain), and long pulse lengths are associated with unwanted thermal effects. Therefore, this study utilized a 1.66 MHz centre frequency transducer and a 10 ms pulse length (1 Hz pulse repetition frequency). Benchtop studies illustrated that drug release began to plateau at this pulse length, suggesting increasing the pulse length would not significantly increase the therapeutic effect, but may increase the likelihood of thermal effects.

Benchtop findings indicated that subharmonic emissions from vaporizing droplets could be used as a signal of drug release. This result supported in vivo findings where subharmonic emissions correlate with a reduction in c-Fos expression induced by the local delivery of pentobarbital. The sensorimotor deficit seen in gait analysis verified the localization of anesthesia and showed that the magnitude of neurosuppression was robust enough to induce gait abnormalities.

In contrast, using sham nanodroplets with FUS was found to induce an increase in c-Fos positive cells, indicative of neurostimulation. The level of increase correlated with second harmonic emissions, which are typical of non-inertial bubble behaviour. Stable bubble activity is known to be responsible for significant and persistent shearing forces on vessel walls, it is possible that vaporizing nanodroplets and pulsating nanodroplet-derived bubbles activate mechanosensitive ion channels, resulting in a neurostimulatory effect (34).

A second potential mechanism for this cavitation-mediated increase in c-Fos expression is sonoporation. Neurostimulation in the presence of microbubbles has been observed following pore formation in cells produced by collapsing bubbles (known as sonoporation), causing jetting flow and generating intracellular calcium ion waves (35). Large lipid-shell droplets have been shown to permeabilize nearby cell membranes, visualized with live-cell microscopy imaging (36). An increase in c-Fos expression has also been reported previously in a mouse model, whilst the BBB remained intact (15,16). However this is the first time vaporizing nanodroplets have been reported to induce an increase in c-Fos expression.

4.2. Localization and spatial specificity

15 targets in total were used to cover the motor cortex, and the effect of pentobarbital delivery was mapped specifically to this region through c-Fos expression. No measurable effects of pentobarbital were recorded in the non-sonicated motor cortex (figure 5F), suggesting the effects of nanodroplets and pentobarbital leakage outside of the sonicated region were non-existent.

In treatment planning, skull aberrations were deemed to be negligible, assuming the pressure profile in the rat brain closely resembled the free-field focus (figure 1B). O’Reilly et al. (37) showed minimal distortion of the pressure profile caused by a rat skull at 1.4 MHz. We would expect similar effects, including negligible change in focal location with slight broadening of the focal volume. Furthermore, using the acoustic feedback controller, we were able to ensure sufficient and safe cavitation activity in the brain through real-time monitoring of subharmonic emissions.

4.3. Safety assessment

As in our previous work, the BBB remained intact following nanodroplet vaporization (23). We hypothesize this is due to the small resultant bubble diameter and associated low gas volume, which has been shown previously to prevent BBB disruption (38). To investigate this further we compared the resultant bubble size distribution from nanodroplets with native Definity (figure 2C, table 3). While the mean diameter was comparable, a significant proportion of large microbubbles were present in native Definity, where the largest measured diameter was almost five times that of nanodroplet-derived microbubbles, containing over 100 times the gas volume.

Potential adverse bioeffects following exposure to Definity and ultrasound have been extensively studied in the context of BBB opening in pre-clinical and clinical trials. Hemorrhage, ischemia, edema and inflammation have been recorded following excessive exposure to microbubbles and focused ultrasound (39,40). However, effects such as microhemorrhage are resolved by 1–4 days following treatment, seen in both pre-clinical models (41) and in patients (42).

Microbubble dose has been found to play an important role in adverse ultrasound-mediated bioeffects (41), but by using a feedback controller sensitive to microbubble activity, inflammation can be minimised, and edema and hemorrhage can be avoided, even with high microbubble doses. Similarly in the current study, even though significantly higher acoustic pressures are required to activate nanodroplets compared to microbubbles (approximately 1.4 MPa compared to 0.4 MPa (43) at 1.78 MHz respectively), by limiting the maximum sonication pressure based on the detected subharmonic bubble emissions, droplet vaporisation and drug delivery can be achieved without any detectable hemorrhage, edema or cell death (figure 4).

Finally, a key safety advantage of localized drug delivery strategies is the reduced systemic drug dose. For example, the dose of pentobarbital used in this study (12.8 ± 3.1 ug/kg) is 3 orders of magnitude smaller than typical doses required to induce general anaesthesia in a rat (30 – 60 mg/kg) (44). By delivering the drug locally, less drug is required, and the likelihood of systemic toxicity is reduced.

4.4. Limitations

c-Fos is one of the most commonly employed IEGs for brain activity mapping. We were able to show differential c-Fos expression between the experimental and control groups that is consistent with the expected outcome from pentobarbital (45). Yet the non-specificity of c-Fos as a crude marker of cellular activation to stimuli limits our ability to discern the underlying mechanism to cavitation-induced neurostimulation by vaporizing nanodroplets. This will require further investigation potentially by more responsive indicators such as calcium imaging, electrophysiology or metabolic imaging, although these are technically challenging when combined with FUS.

It has been well documented that many factors, including mechanical stress, can increase c-Fos expression (46). This is one possible explanation for the increased c-Fos expression following exposure to FUS and vaporizing sham nanodroplets. Although it should be noted that FUS alone showed little increase in c-Fos expression. Furthermore, knowing that mechanical stress can increase c-Fos expression, highlights the potency of local pentobarbital delivery in supressing c-Fos expression in the presence of vaporizing nanodroplets. The combination of FUS and pentobarbital-loaded nanodroplets was able to overcome the c-Fos expression associated with vaporising nanodroplets and still induce significant levels of suppression. Future work will look to use NeuN as a co-stain alongside c-Fos mapping to investigate neuronal expression further, specifically to identify neuron type.

Although gait analysis was sufficient to detect a sensorimotor deficit following local anesthesia, gait analysis was not able to detect neurostimulation following treatment with sham droplets. Other behavioural markers, such as performance in tactile discrimination (47) or visual-motor (13) tasks, could be used in future studies. Furthermore, we may be able to improve the monitoring of droplet activation and drug delivery by exploiting the unique acoustic emissions from vaporization found below the fundamental frequency (28), or even through super resolution imaging (48). This may also help to elucidate the neurostimulatory mechanism.

5. Conclusions

Nanodroplets, formed from a repurposed microbubble contract agent, were shown to be a tool for neurosuppression and neurostimulation, without disrupting the BBB in rats. Treatment was targeted using an atlas-based FUS system with high spatial precision, and treatment was found to be safe and reproducible. FUS-triggered sham nanodroplets produced a significant increase in c-Fos expression, whilst the FUS-triggered delivery of pentobarbital significantly reduced c-Fos expressed and was sufficient to impair sensorimotor function.