**Ultrasound-Mediated Blood–Brain Barrier Disruption for Drug Delivery: A Systematic Review of Protocols, Efficacy, and Safety Outcomes from Preclinical and Clinical Studies**

**Kushan Gandhi 1,2 , Anita Barzegar-Fallah 1,2 , Ashik Banstola 1,2 , Shakila B. Rizwan 2,3 and John N. J. Reynolds 1,2,\***


**Abstract:** Ultrasound-mediated blood–brain barrier (BBB) disruption has garnered focus as a method of delivering normally impenetrable drugs into the brain. Numerous studies have investigated this approach, and a diverse set of ultrasound parameters appear to influence the efficacy and safety of this approach. An understanding of these findings is essential for safe and reproducible BBB disruption, as well as in identifying the limitations and gaps for further advancement of this drug delivery approach. We aimed to collate and summarise protocols and parameters for achieving ultrasound-mediated BBB disruption in animal and clinical studies, as well as the efficacy and safety methods and outcomes associated with each. A systematic search of electronic databases helped in identifying relevant, included studies. Reference lists of included studies were further screened to identify supplemental studies for inclusion. In total, 107 articles were included in this review, and the following parameters were identified as influencing efficacy and safety outcomes: microbubbles, transducer frequency, peak-negative pressure, pulse characteristics, and the dosing of ultrasound applications. Current protocols and parameters achieving ultrasound-mediated BBB disruption, as well as their associated efficacy and safety outcomes, are identified and summarised. Greater standardisation of protocols and parameters in future preclinical and clinical studies is required to inform robust clinical translation.

**Keywords:** focused ultrasound; blood–brain barrier opening; therapeutic agent delivery; ultrasound parameters; ultrasound safety; review

#### **1. Introduction**

#### *1.1. The Blood–Brain Barrier and Drug Delivery*

The blood–brain barrier (BBB) is a selectively permeable structure that restricts the passage of solutes from the brain's microvasculature into its extracellular space. Anatomically, the BBB is composed of the apical and basal membranes of the cerebrovascular endothelial cells (CECs), an associated basement membrane containing embedded pericytes, and perivascular foot-like processes of astrocytes ensheathing the abluminal capillary surface collectively referred to as the neurovascular unit [1] (see Figure 1). The CECs express a limited range of membrane carrier proteins and numerous membrane efflux pumps and have a dense number of tight junctions linking them to other CECs. This combined structural arrangement restricts the movement of most hydrophilic and high molecular weight molecules (exceeding 400 to 500 Da) [2], inflammatory cells, and pathogens, thereby providing a vital function in maintaining homeostasis and preventing the entry of harmful substances into the brain. The BBB, therefore, grants a significant survival advantage, but it

**Citation:** Gandhi, K.; Barzegar-Fallah, A.; Banstola, A.; Rizwan, S.B.; Reynolds, J.N.J. Ultrasound-Mediated Blood–Brain Barrier Disruption for Drug Delivery: A Systematic Review of Protocols, Efficacy, and Safety Outcomes from Preclinical and Clinical Studies. *Pharmaceutics* **2022**, *14*, 833. https://doi.org/10.3390/ pharmaceutics14040833

Academic Editors: Jingyuan Wen and Yuan Huang

Received: 13 March 2022 Accepted: 6 April 2022 Published: 11 April 2022 Corrected: 20 December 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

also poses a disadvantage in its inability to allow most therapeutic agents to penetrate it, rendering 98% of small-molecule agents and 100% of large-molecule agents unable to enter the brain parenchyma [3]. As a result, there is a significant limitation to the pharmacological agents available in the treatment of central nervous system (CNS) conditions, including brain malignancies, dementias, and other neurodegenerative conditions. to penetrate it, rendering 98% of small-molecule agents and 100% of large-molecule agents unable to enter the brain parenchyma [3]. As a result, there is a significant limitation to the pharmacological agents available in the treatment of central nervous system (CNS) conditions, including brain malignancies, dementias, and other neurodegenerative conditions.

of harmful substances into the brain. The BBB, therefore, grants a significant survival advantage, but it also poses a disadvantage in its inability to allow most therapeutic agents

*Pharmaceutics* **2022**, *14*, x FOR PEER REVIEW 2 of 34

**Figure 1.** Schematic representation of the anatomical structure of the blood–brain barrier (BBB) and the accompanying neurovascular unit. Note the abluminal CEC surface is ensheathed by a basement membrane embedded with pericytes. (*Created with Biorender.com.*) **Figure 1.** Schematic representation of the anatomical structure of the blood–brain barrier (BBB) and the accompanying neurovascular unit. Note the abluminal CEC surface is ensheathed by a basement membrane embedded with pericytes. (*Created with Biorender.com* (accessed on 12 March 2022)).

#### *1.2. Ultrasound-Mediated Drug Delivery and the BBB 1.2. Ultrasound-Mediated Drug Delivery and the BBB*

An approach to overcoming the challenge posed by the BBB is to temporarily induce BBB disruption, in a controlled and targeted manner, to enhance the uptake of therapeutic agents into desired target locations in the brain. A minimally invasive strategy that has been employed to achieve this is via the application of ultrasound. The use of transcranial ultrasound for enhancing drug delivery in the CNS also extends beyond BBB disruption, including in functioning as an external trigger to initiate drug release from nanoparticles [4] at targeted areas of the BBB in diseases such as epilepsy [5]. While the use of ultrasound in disrupting the BBB was first described in the 1950s [6], it is within the last 20 years [7] that this technique has garnered significant research interest in improving drug delivery to the brain. Subsequently, numerous preclinical animal studies, as well as several phase I/II clinical trials [8–11] (ClinicalTrials.gov identifier numbers: NCT03321487, NCT03322813, NCT02253212, NCT03608553, NCT03626896, NCT02343991, NCT02986932, NCT03712293), assessing the feasibility of this drug delivery approach have emerged. While the exact mechanisms underlying ultrasound-mediated BBB disruption are not yet well defined, the mechanical bioeffects of ultrasound exposure are thought to predominate. When exposed to ultrasound, dissolved gas bubbles within the vasculature experience a phenomenon known as cavitation, where they experience oscillatory changes in their volume, expanding in volume with rarefactions (low sonic pressure), and contracting with compressions (high sonic pressure) of the ultrasound waves [7]. Gas bubbles may also experience an acoustic radiation force, where they gain additional translational movement towards the direction of the ultrasound beam. These effects together are thought to contribute to the observed reduction in tight junction proteins between CECs of the BBB [12], the reduced expression of P-glycoprotein drug efflux pumps along the CEC membrane [13–15], and the increased formation of transcytotic vesicles across the CECs [16]. To enhance these mechanical effects ultrasound contrast agents or microbubbles (preformed gas-filled bubbles, typically of 1 to 6 µm in diameter) are co-administered, An approach to overcoming the challenge posed by the BBB is to temporarily induce BBB disruption, in a controlled and targeted manner, to enhance the uptake of therapeutic agents into desired target locations in the brain. A minimally invasive strategy that has been employed to achieve this is via the application of ultrasound. The use of transcranial ultrasound for enhancing drug delivery in the CNS also extends beyond BBB disruption, including in functioning as an external trigger to initiate drug release from nanoparticles [4] at targeted areas of the BBB in diseases such as epilepsy [5]. While the use of ultrasound in disrupting the BBB was first described in the 1950s [6], it is within the last 20 years [7] that this technique has garnered significant research interest in improving drug delivery to the brain. Subsequently, numerous preclinical animal studies, as well as several phase I/II clinical trials [8–11] (ClinicalTrials.gov identifier numbers: NCT03321487, NCT03322813, NCT02253212, NCT03608553, NCT03626896, NCT02343991, NCT02986932, NCT03712293), assessing the feasibility of this drug delivery approach have emerged. While the exact mechanisms underlying ultrasound-mediated BBB disruption are not yet well defined, the mechanical bioeffects of ultrasound exposure are thought to predominate. When exposed to ultrasound, dissolved gas bubbles within the vasculature experience a phenomenon known as cavitation, where they experience oscillatory changes in their volume, expanding in volume with rarefactions (low sonic pressure), and contracting with compressions (high sonic pressure) of the ultrasound waves [7]. Gas bubbles may also experience an acoustic radiation force, where they gain additional translational movement towards the direction of the ultrasound beam. These effects together are thought to contribute to the observed reduction in tight junction proteins between CECs of the BBB [12], the reduced expression of P-glycoprotein drug efflux pumps along the CEC membrane [13–15], and the increased formation of transcytotic vesicles across the CECs [16]. To enhance these mechanical effects ultrasound contrast agents or microbubbles (preformed gas-filled bubbles, typically of 1 to 6 µm in diameter) are co-administered, thereby increasing the number of available echogenic centres that can induce a mechanical effect within the cerebral vasculature, and thus reducing the threshold of ultrasound intensity required for BBB permeabilisation. Furthermore, magnetic resonance imaging (MRI) has been coupled with ultrasound and

microbubbles, as a way of ensuring precise targeting [17] of the ultrasound beam to specific regions within the brain, confining drug penetration to these foci only, e.g., tumour sites for glioma patients.

#### *1.3. Challenges with Ultrasound-Mediated BBB Disruption*

Though numerous studies have achieved variable levels of successful ultrasoundmediated BBB disruption, it has become clear that the extent of BBB disruption is greatly influenced by the choice of ultrasound parameters, as well as the type and dose of microbubbles used alongside each sonication [18]. Unfortunately, ultrasound-mediated BBB disruption has potential adverse effects, including haemorrhagic change (ranging from sparse erythrocyte extravasation to gross intracerebral haemorrhaging) [19], oedema [9,20–26], inflammation [26,27], neuronal ischaemia [28,29], and tissue apoptosis [23,27,28,30–34]. The occurrence of these adverse effects is theorised to be due to excessive mechanical activity, where sonicated bubbles within the vasculature experience inertial, non-stable cavitation, rapidly imploding to exert excessive endothelial force, causing the extravasation of fluid, erythrocytes, and leucocytes into the surrounding tissue. Additionally, these effects may contribute to direct neuronal and glial injury. This has precipitated a large body of both preclinical and clinical studies demonstrating ultrasound-mediated BBB disruption with a range of sonication parameters, accompanied by a diverse set of reported safety outcomes. Ultimately, this has made selecting an appropriate sonication protocol that provides both successful and safe BBB disruption a difficult task. Previous narrative reviews of this body of evidence have been conducted but only broadly summarise key ultrasound-related parameters and their associated effects on the efficacy and safety of ultrasound-mediated BBB disruption. A systematic review of such literature and a published database of individual sonication paradigms and their consequential safety outcomes has yet to be conducted. The results of this systematic review will aim to inform future researchers of methods, sonication protocols, and the parameters that influence BBB disruption, as well as the safety outcomes associated with each, across a variety of experimental models (rodents, rabbits, sheep, pigs, non-human primates (NHPs)) and humans.

#### **2. Materials and Methods**

The systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines [35,36], PRISMA's checklist or PRISMA Flow Diagram for systematic reviews has been completed and is available as Table S1 in Supplementary Materials. This systematic review was not registered.

#### *2.1. Eligibility Criteria*

All studies with (i) clearly outlined sonication protocols (containing the relevant terms below), where (ii) ultrasound was applied to the brain of in vivo animal or human participants and where (iii) successful BBB disruption was achieved and confirmed using a reliable method, were included within this systematic review. Additionally, included studies had to have conducted appropriate safety assessments and reported any adverse effects associated with each protocol achieving successful BBB disruption. Only sonication protocols achieving confirmed BBB disruption, with corresponding safety assessments, were included within the data collection process. Protocols where the primary aim was applying ultrasound for cellular, viral, or gene delivery; neuromodulation; stimulation; or tissue ablation were excluded. Studies not published in English and review papers were also excluded.

#### *2.2. Information Sources*

This systematic review was based on searches from the following online search databases: PubMed, Medline, and EMBASE. A complete search of databases was conducted on 22 September 2021 to include any subsequent articles published within this timeframe. Additional articles were identified from reviewing the reference lists of relevant

articles identified via the initial database search, as well as via Google Scholar alerts from the date of the initial search. A complete search of databases was conducted by K.G.

#### *2.3. Search Strategy*

The following MeSH terms were used to identify relevant articles:


#### *2.4. Study Selection*

Citations identified from the searches of the three databases were collated into a combined EndNote (X9) library and were de-duplicated as per a published protocol [37]. All unique citations were then screened for inclusion according to the eligibility criteria. Citation screening was conducted by K.G. in the following sequence: initially by their titles, then by their abstracts, and finally by reviewing the entire text. K.G., A.B.-F., and A.B. independently identified additional studies from Google Scholar alerts and from reviewing the reference list of relevant articles identified via the database search.

#### *2.5. Data Items and Collection Process*

From each included study, the following data items were extracted: (1) species of in vivo subject; (2) type of ultrasound transducer used; (3) methods for assessing BBB disruption; (4) methods for assessing safety; (5) microbubbles used (type, dose, administration protocol); (6) sonication parameters (frequency, peak negative pressure in situ, continuous or pulsated delivery, sonication duration, number of sonications, interstimulus interval, number of independent sessions, intersession interval). The extracted data items were then tabulated (Table S2—Supplementary Materials). Additional summary tables regarding methods used to assess BBB disruption efficacy Table 1) and safety, as well as microbubble and ultrasound parameters influencing ultrasound-mediated BBB disruption are also included in this systematic review. K.G. conducted all data extraction and collection.

#### **3. Results**

#### *3.1. Included Studies*

A total of 1480 citations were identified after the initial literature search from the three databases (Figure 2). After de-duplication, 882 unique citations were identified and were subsequently screened, in order, by their title and abstract, and then by a full-text review of the remaining articles (*n* = 100). After completing the screening process, a total of 76 studies were identified and included for data extraction and synthesis. An additional 31 studies were identified and included from Google Scholar alerts and after screening reference lists of included studies. Ultimately, a total of 107 studies were eligible for inclusion in our qualitative comparison and analysis.

**Figure 2***.* Flowchart highlighting the screening and selection process for studies included within this systematic review. (*Created with Biorender.com.*) **Figure 2.** Flowchart highlighting the screening and selection process for studies included within this systematic review. (*Created with Biorender.com* (accessed on 12 March 2022)).

#### *3.2. Ultrasound Devices 3.2. Ultrasound Devices*

A range of commercial (e.g., FUS Instruments, Imasonic, Riverside Research Institute, Sonic Concepts) and in-house manufactured ultrasound devices were utilised in the included study protocols, with the majority of these being single-element or single piezoelectric devices. In recent times, multi-element devices have been developed to overcome associated concerns around ultrasound attenuation and beam defocussing, providing better transcranial transmission. These devices have been tested in preclinical animal models as well as in many ongoing and published clinical trials. In addition, we identified a protocol utilising two single-element transducers in tandem [38], and two that even used diagnostic [39,40], imaging transducers to disrupt the BBB. A comparative figure highlighting therapeutic ultrasound devices identified amongst included literature is shown in Figure 3. A range of commercial (e.g., FUS Instruments, Imasonic, Riverside Research Institute, Sonic Concepts) and in-house manufactured ultrasound devices were utilised in the included study protocols, with the majority of these being single-element or single piezoelectric devices. In recent times, multi-element devices have been developed to overcome associated concerns around ultrasound attenuation and beam defocussing, providing better transcranial transmission. These devices have been tested in preclinical animal models as well as in many ongoing and published clinical trials. In addition, we identified a protocol utilising two single-element transducers in tandem [38], and two that even used diagnostic [39,40], imaging transducers to disrupt the BBB. A comparative figure highlighting therapeutic ultrasound devices identified amongst included literature is shown in Figure 3.

**Figure 3.** Comparison of therapeutic ultrasound devices used for BBB disruption. (*Created with Biorender.com.*) **Figure 3.** Comparison of therapeutic ultrasound devices used for BBB disruption. (*Created with Biorender.com* (accessed on 12 March 2022)).

#### 3.2.1. Single-Element Ultrasound Devices 3.2.1. Single-Element Ultrasound Devices

While single-element devices are smaller, and more accessible for testing in preclinical studies, their application in clinical studies has been more limited due to the attenuation and defocussing resulting from the application of a single ultrasound beam [41]. A strategy employed [9] to circumvent this issue has been to create a bony window in the calvaria and apply ultrasound directly through it, thereby reducing attenuation of the ultrasound beam at the bone interface [42]. Two animal studies directly compared this strategy with application through intact bone, with one highlighting no difference with the use of a 260 kHz transducer in rabbits [30] and the other highlighting significant improvements with the use of a 28 kHz transducer in pigs [23]. The SonoCloud-1 ® , a single-element device manufactured by CarThera, is one such example that requires implantation via a burr hole in the skull. Currently, this is the only implantable device we identified amongst all included protocols, and it has demonstrated efficacy in disrupting the BBB in both animal [20,43,44] and clinical studies [9]. Dual single-element transducers were employed in a couple of studies [38,45] to initiate BBB disruption using a lower frequency transducer, while a higher frequency transducer was employed in an attempt to stimulate the transport of a therapeutic agent into the brain parenchyma. While single-element devices are smaller, and more accessible for testing in preclinical studies, their application in clinical studies has been more limited due to the attenuation and defocussing resulting from the application of a single ultrasound beam [41]. A strategy employed [9] to circumvent this issue has been to create a bony window in the calvaria and apply ultrasound directly through it, thereby reducing attenuation of the ultrasound beam at the bone interface [42]. Two animal studies directly compared this strategy with application through intact bone, with one highlighting no difference with the use of a 260 kHz transducer in rabbits [30] and the other highlighting significant improvements with the use of a 28 kHz transducer in pigs [23]. The SonoCloud-1®, a single-element device manufactured by CarThera, is one such example that requires implantation via a burr hole in the skull. Currently, this is the only implantable device we identified amongst all included protocols, and it has demonstrated efficacy in disrupting the BBB in both animal [20,43,44] and clinical studies [9]. Dual single-element transducers were employed in a couple of studies [38,45] to initiate BBB disruption using a lower frequency transducer, while a higher frequency transducer was employed in an attempt to stimulate the transport of a therapeutic agent into the brain parenchyma.

#### 3.2.2. Multi-Element Ultrasound Devices 3.2.2. Multi-Element Ultrasound Devices

Multi-element array devices confer the benefit of being able to treat multiple, separate tissue foci simultaneously, as well as providing greater spatial coverage than singleelement devices [46]. Additionally, multi-element phased array devices confer the ability to alter the phase and amplitude of individual transducers, correcting for aberrations to the ultrasound beam as it surpasses more complex skull surfaces, such as the human calvaria [46,47]. An emerging leader in this category of devices is the ExAblate® system (manufactured by InSightec), a 1024-element phased array device, initially designed for the thermal ablation of tissue, prior to being applied in rat [48,49], NHP [50], and clinical studies [8,10,11,51–53] to disrupt and open the BBB. The ExAblate® system functions with a large, stereotactically positioned helmet that is coupled to an MRI system to help plan sonication targets and monitor the procedure. A second multi-element phased array system—NaviFUS® (produced by NaviFUS corporation)—is a 256-element phased array device that has also demonstrated its ability to disrupt the BBB in a recent clinical trial [54]. While this device and the ExAblate® share similarities in their appearance, NaviFUS® does not require stereotactic positioning and relies on traditional, neurosurgical, navigation to help plan and guide ultrasound beams to sonication targets. The SonoCloud-9 ® (manufactured by CarThera) is an iteration of the SonoCloud-1 ® and functions as an implantable Multi-element array devices confer the benefit of being able to treat multiple, separate tissue foci simultaneously, as well as providing greater spatial coverage than single-element devices [46]. Additionally, multi-element phased array devices confer the ability to alter the phase and amplitude of individual transducers, correcting for aberrations to the ultrasound beam as it surpasses more complex skull surfaces, such as the human calvaria [46,47]. An emerging leader in this category of devices is the ExAblate® system (manufactured by InSightec), a 1024-element phased array device, initially designed for the thermal ablation of tissue, prior to being applied in rat [48,49], NHP [50], and clinical studies [8,10,11,51–53] to disrupt and open the BBB. The ExAblate® system functions with a large, stereotactically positioned helmet that is coupled to an MRI system to help plan sonication targets and monitor the procedure. A second multi-element phased array system—NaviFUS® (produced by NaviFUS corporation)—is a 256-element phased array device that has also demonstrated its ability to disrupt the BBB in a recent clinical trial [54]. While this device and the ExAblate® share similarities in their appearance, NaviFUS® does not require stereotactic positioning and relies on traditional, neurosurgical, navigation to help plan and guide ultrasound beams to sonication targets. The SonoCloud-9® (manufactured by CarThera) is an iteration of the SonoCloud-1® and functions as an implantable grid of nine interconnected transducer elements. As yet, no published study has employed the

SonoCloud-9® for in vivo BBB disruption, but ongoing clinical trials exist (NCT03744026, NCT04614493, NCT04528680).

#### 3.2.3. Diagnostic Ultrasound Devices

Disruption of the BBB was also achieved using diagnostic ultrasound imaging devices in mice [39,40], albeit with much broader and less well-defined tissue coverage as compared to the prior mentioned therapeutic ultrasound devices. This is partly due to the higher central frequencies (2–10 MHz) that diagnostic devices tend to operate with, often resulting in greater aberration and defocussing of the penetrating ultrasound beam [39].

#### 3.2.4. Implantable Ultrasound Devices

Nearly all identified ultrasound devices were non-implantable and usually required precise (e.g., stereotactic, neuronavigation, MRI-guided) positioning over the cranium of each subject prior to sonication. Implantable devices (SonoCloud® devices) are highly portable and eliminate the need for repeat repositioning, at the cost of requiring more invasive, surgical placement. Non-implantable devices have the advantage of being surgically non-invasive and can be repositioned for targeting multiple sites, at the cost of longer ultrasound sessions, during which subjects are not ambulatory [55]. A recent review [55] suggested that non-implantable, extracranial devices were not appropriate for targeting superficial lesions, a claim not supported by findings from identified protocols that achieved successful and safe BBB disruption in superficial cortical regions (e.g., primary motor cortex [8,56], primary visual cortex [24], prefrontal cortex [10,57]) using both single-element and multi-element phased array devices.
