Drug and Gene Delivery Across the Blood-Brain Barrier with Focused Ultrasound

1.1 The Blood-Brain Barrier

The BBB provides a formidable obstacle for drug delivery in the brain (Figure 1). Through a unique combination of transmembrane proteins and tightly regulated channels not seen elsewhere in the body, the BBB prevents nearly 100% of large molecule (>500 Da) drugs, including recombinant proteins and antibodies, and 98% of small molecule drugs, from passing into the brain [6]. Lipid soluble small molecule drugs may cross the BBB if they are capable of diffusing through the endothelial cell membrane itself [6], but few drugs fall into this category. The BBB’s remarkable exclusionary capability is attributed to tight junctions that join the endothelial cells lining the vasculature throughout the brain [7]. Tight junctions are comprised of several proteins, including various claudins, occludins, junctional adhesion molecules, and cadherins, which function to prevent molecules from passively diffusing between cells and out of the vasculature [7]. Rather, small molecules must pass through the endothelial cells themselves, either through diffusion, for lipid soluble molecules, or active transport, for most nutrients and other substances necessary for normal brain function [7]. Furthermore, if a certain molecule does manage to pass through the endothelial cell layer, the basement membrane provides an additional barrier to diffusion. Simply stated, nature’s best defense against infection significantly hinders our ability to treat diseases of the CNS by preventing drug delivery to the brain.

1.2 Conventional Approaches for Bypassing the Blood-Brain Barrier

Given the central role of the BBB in limiting drug and gene delivery to the CNS, numerous methods have been developed to bypass this barrier. For example, specific viruses or nanoparticles (NPs) with BBB-targeting ligands can cross the BBB after systemic administration[8]. However, in order to achieve effective concentrations in the brain, they must be administered in doses which are associated with adverse effects in peripheral organs[9]. For this reason, the majority of preclinical and clinical studies have used direct intracranial administration as a strategy to locally increase therapeutic concentration without off-target effects. Specific brain regions can be accessed with needles or catheters and more recent strategies have utilized fluid convection to enhance distribution of therapeutics in the brain[10]. By maintaining bulk flow with hydrostatic pressure differentials, convection enhanced delivery has demonstrated marked improvement over conventional direct intracranial injection methodologies[11,12]. Unfortunately, despite promising results for direct injection in several preclinical and clinical trials[13–17], these strategies are risky and surgical complications have hindered widespread adoption. Furthermore, macromolecular agents require long dissemination times and typically cannot spread beyond a few millimeters[11]. Indeed, the invasive nature of strategies like intracranial injections is not compatible with drugs that need to be dosed repeatedly.

In order to reduce risks associated with direct injection, less-invasive strategies to enhance therapeutic delivery across the BBB have been developed. These include intranasal administration and chemical disruption of the BBB by intra-arterial infusion of the osmotic agent mannitol[18] or vasodilators[19–21]. Intranasal administration permits transport to the brain through perineural or perivascular channels[22]. While intranasal drug delivery is non-invasive and obviates peripheral side effects associated with intravenous administration, it is limited by poor absorption across the nasal epithelium, inconsistent delivery efficiency and poor localization[22,23]. Similarly, mannitol infusions lead to global BBB disruption, causing non-specific uptake and potentiating adverse off-target effects. Infusion of mannitol into the carotid artery leads to an osmotic-driven movement of fluid out of endothelial cells[24], shrinking them and leading to fenestration of cerebral vessels. While disruption of the BBB with mannitol is reasonably safe, therapeutic delivery is inconsistent with up to 10-fold variations in drug concentrations[25].

1.3 Opening the Blood-Brain Barrier with Focused Ultrasound

FUS has the advantage of being the only modality capable of achieving non-invasive, safe, repeated, and targeted BBB disruption to enhance drug or gene delivery to the CNS. With the advent of MR-compatible transducers with sub-millimeter precision, it is now possible to apply image-guided transcranial FUS to the human brain [26–29] in an extremely localized manner, greatly reducing the risk of off-target effects. FUS treatments can be performed on awake patients, eliminating the need for general anesthetic and permitting real-time patient feedback. In the future, FUS may be capable of replacing traditional surgical techniques, eliminating invasive procedures and greatly increasing the feasibility of repeated treatments. Importantly, MR and integrated passive cavitation detection (PCD) facilitate real-time intraoperative treatment monitoring, while post-treatment MR imaging allows confirmation of treatment success[30–33] and safety[34,35]. The development of transcranial FUS has been a long process. Groundbreaking research by the Fry brothers performed over 50 years ago demonstrated that ultrasound could produce bioeffects in the human brain[36]. However, it wasn’t until recent technological advances were made in both ultrasound and MRI that the field experienced a surge in interest. In the past ten years, there has been an increase in the number of papers investigating the potential applications of ultrasound in the brain.

Ultrasound is, at its most basic, a pressure wave. As the wave passes through the tissue, the tissue experiences alternating periods of high pressure (compression) and low pressure (rarefaction). Ultrasound can be applied in a continuous fashion, common in treatments that require heat deposition, or in a pulsed manner, which is utilized for blood-brain barrier disruption (BBBD). Focusing the ultrasound beam (i.e. FUS) provides high spatial accuracy (less than 1 mm resolution in some cases) and localizes bioeffects. However, reflection and diffraction of the ultrasound wave at material interfaces (i.e. skull-tissue interface) can distort the focus and decrease the energy delivered at the target. While the favorable skull geometry of rats or mice allows the use of single-element transducers in pre-clinical trials (Figure 2), the far more complex topography of the human skull requires the use of a multi-element array with phase-correction software to re-focus the ultrasound beam as it passes through the skull. There are many combinations of FUS parameters (frequency, pressure, pulsing protocol) suitable for BBBD, but lower frequencies (≤1.0 MHz) experience less attenuation and distortion by the skull.

Transcranial FUS is typically applied in conjunction with intravenously administered MBs to effect BBBD. MBs are small (1–10 µm) lipid or protein shelled bubbles filled with an inert gas, most commonly a perfluorocarbon and are FDA-approved as a contrast agent during ultrasound imaging. Importantly, circulating MBs reduce the acoustic energy required to open the BBB by two orders of magnitude and confine mechanical effects to the vasculature[37]. This permits the use of low pressure FUS and virtually eliminates any concerns about skull heating during treatment. Extended off-time (low duty cycle) between FUS pulses allows MB reperfusion and thermal dissipation at the focus. At the lower ultrasonic pressures used for BBBD, MBs oscillate stably in the FUS field, expanding during rarefaction and contracting during compression, producing mechanical shear forces[38] and microstreaming[39] effects which act on the vessel wall. This behavior domain, called stable oscillation, is preferred for BBBD as its effects are more predictable. In contrast, at higher acoustic pressures, MBs experience unstable oscillations and eventually collapse inward, producing elevated local temperatures and high-pressure jet streams in a process termed inertial cavitation. While FUS-MB induced BBBD occurs in both regimes, inertial cavitation is more violent and is generally avoided for applications in healthy brain tissue. However, it may find use in diseased tissue, or for the delivery of very large (~100 nm) therapeutics, when the potential benefits outweigh the risks.

A collection of in vitro and ex vivo work[40–42] has demonstrated that the mechanical forces exerted by stably oscillating MBs cause vessel distension and invagination, as well as changes in the endothelial cells’ cytoskeletons and cell-cell interactions (Figure 3). Together, these effects produce BBBD via three mechanisms: disruption of tight junctions, induction of transcytosis, and sonoporation of the vascular endothelium. Work using transmission electron microscopy imaging[43] has demonstrated both a reduction and altered distribution of claudins 1 and 5, occludin, and ZO-1 after FUS exposure. Most notably, tight junction proteins were no longer clustered along the edges of endothelial cells, suggesting that they were no longer contributing to tight junction complexes. Furthermore, penetration of horse radish peroxidase between endothelial cells was evident, demonstrating that tight junctions were no longer sealing paracellular pathways from the vasculature to the brain parenchyma. In addition to this paracellular pathway, horseradish peroxidase was also taken up by the vascular endothelial cells after sonication. Later work[44–46] demonstrated increased expression of caveolin-1 in the vascular endothelium after sonication, identifying caveoli as the most likely transcytotic pathway. A unique approach using two-photon microscopy provided further support for both paracellular and transcellular pathways. Here, it was noted that dextrans crossed the BBB after FUS via either a fast or slow mechanism (i.e. less/greater than 10 min) and postulated that the fast and slow pathways were most likely paracellular (tight junction disruption) and transcellular (increased transcytosis and sonoporation), respectively[47]. Interestingly, 70kDa dextrans appeared to have a higher pressure threshold for BBB crossing than smaller 10kDa dextrans, in agreement with other studies indicating that the extent of BBBD (particularly the size of junctional clefts) is related to pressure[48]. Furthermore, smaller vessels (i.e. less than ~25–30 µm) were significantly more likely to be disrupted by FUS[49] than larger vessels, and fast leakage[47] (i.e. paracellular) was the dominant mode of transport in these vessels. This difference is attributed to the interactions between MBs and the vessel wall – in smaller vessels, oscillating MBs are more likely to come in close contact with the vessel wall, generating larger circumferential stresses than in larger vessels. Going forward, achieving a better understanding of the dynamics of these transport pathways will be critical for enabling more predictable BBBD, especially with the increased use of larger therapeutics such as antibodies, liposomes and polymer nanoparticles.

1.4 Safety and Monitoring of Blood-Brain Barrier Disruption

While it is well known that driving MBs into inertial cavitation with high acoustic pressures can lead to irreversible capillary damage and the leakage of red blood cells across the BBB [50], thresholds have been established wherein BBBD can be achieved without toxicity or damage. BBBD is transient and, depending on acoustic pressure[51], barrier function is typically restored within 4–6 hours after treatment (Figure 4A,B) [43,52,53]. Importantly, no motor or acuity deficits were found after repeated BBBD procedures with FUS at numerous targets in monkeys[54,55]. Interestingly, some of these monkeys had T2* hypointensities in post-FUS MR imaging, indicating minor red blood cell accumulation; however, these minor capillary leakages did not lead to any changes in visual acuity or motor skills. Furthermore, it is important to emphasize that, even in the rare occurrence of erythrocyte extravasation at the lower pressures used in these studies, no apoptotic bodies were found and cognitive function of the animals was not impaired[54]. These findings are in agreement with other studies suggesting that minor capillary damage and red blood cell extravasation is not expected to lead to long term effects[32,56]. Indeed, it is possible that such damage would be acceptable in treatment of debilitating or life-threatening neurological diseases. It is important to further emphasize that FUS-related safety issues would be minor compared to those of other treatment strategies like intracranial injection, which can lead to extensive damage along the needle tract, or even non-invasive treatments like gamma knife radiosurgery[57,58].

While FUS-mediated BBBD has been shown to be safe in numerous animal models, intraoperative monitoring with passive cavitation detection or MR imaging further reduces chances of aberrant FUS treatments. Passive cavitation detection (PCD) allows real-time assessment of MB cavitation[59]. While stable cavitation is most likely responsible for reversible BBBD[50,60], inertial cavitation has been linked with tissue damage[61]. Importantly, acquired PCD intensity is well correlated with BBBD[50,60]. PCD non-invasively detects the acoustic signatures resulting from MB oscillations, and can distinguish between stable oscillations and MB collapse.[62]. Stably oscillating MBs emit harmonic, subharmonic or ultraharmonic frequency acoustic emissions, whereas collapsing MBs emit broadband acoustic signals[63,64]. PCD has been used to ensure safe FUS settings in several large animal BBBD models[54,65] and systems are currently in development that will allow fully automatic feedback to control FUS sonications[66]. Once the FUS treatment is complete, MR imaging sequences including T2* and susceptibility weighted imaging (SWI) are sensitive tools that can be used to detect blood products present in tissue[34,52,67] and have been shown to be sensitive to measure even minor capillary damage[54].

1.5 Therapeutic Bioeffects of Focused Ultrasound

This review is focused on the delivery of drugs and genes across the BBB with FUS; however, it is also important to note that recent advances in our understanding of biological responses to FUS have potentiated novel approaches to treatments of brain disease, even without administration of pharmacological agents. For example, BBBD with FUS alone in a mouse model of Alzheimer’s disease led to a significant reduction of plaque burden four days after a single treatment[68]. The authors showed that plaque reduction was linked to significant enhancement of endogenous antibodies bound to the Aβ plaque as well as to activation of microglia and astrocytes in the FUS-treated region. Additionally, five successive treatments with ultrasound were shown to lead to further plaque clearance and improved subject performance on several memory tasks[69]. While these inflammatory responses bear further investigating, safety studies on repeated FUS treatments in primates have shown no deleterious effects [70] and a localized immune response may be beneficial in some applications. Non-thermal FUS has been used in other disease models to enhance the body’s anti-tumor immune response with great success [71], and it has been postulated that FUS-mediated BBBD may provide a similar benefit in brain tumors by increasing endogenous antibody delivery and reducing barriers to immune cell migration in the brain parenchyma [68,72]. In addition to alterations in immune function and cell behavior within the brain, FUS-mediated BBBD has been shown to enhance neurogenesis[73], which is attributed to the demonstrated upregulation of BDNF[74] and Akt [75] after FUS. This observation has led to suggestions that FUS could be used as a non-invasive alternative to deep brain stimulation for treatment of depression[76]. Importantly, low intensity FUS has also shown ability to transiently stimulate neurons in both animals[74,77–80] and humans[81,82], and has the ability to elicit acute sensory responses in the fingers and hands[82], potentially allowing non-invasive brain mapping. Ultimately, we believe that FUS will permit several therapeutic options in the CNS, beyond those dependent upon drug and/or gene delivery across the BBB.

2.1 Unencapsulated Drug

Currently, the most common application for FUS-mediated delivery of unencapsulated drugs is glioblastoma multiforme (GBM). While the tumor core can be resected, GBM is notorious for tumor recurrence due to the infiltrating tumor cells outside of the tumor core. Indeed, these regions are refractory to traditional systemic therapies due to the existence of an intact BBB. Thus, it has been hypothesized that the delivery of therapeutics via FUS-mediated BBBD may improve patient outcomes by providing drug delivery to these “protected” infiltrating cells [84,92]. Additionally, while the bulk of the tumor may exhibit BBB impairment, this impairment is often heterogeneous and accompanied by high interstitial pressures. FUS treatment has been shown to provide more extensive and more homogeneous drug delivery even in areas with impaired BBB function [83]. This application is particularly promising for patients with brain metastases, as many therapeutics which are efficacious against the primary tumor do not cross the BBB. FUS-mediated delivery across the BBB can significantly improve drug delivery and efficacy in the brain, and may permit a wider range of treatment options for patients with brain metastases, which are far more common than primary brain tumors.

Temozolomide (TMZ), a small molecule drug, is currently part of the recommended approach to the clinical treatment of GBM, and many clinical trials continue to test various TMZ dosing regimens as well as drug combinations [93–96]. In a rat model of GBM, BBBD in combination with orally administered medium dose TMZ significantly increased survival (IST_median=15% compared to controls) and controlled tumor growth as well as high dose TMZ alone [92]. A study using the U87 glioma model in mice further demonstrated that FUS treatment improves tumor growth control and survival over TMZ alone across a range of TMZ doses (IST_median=111% compared to control for highest dose TMZ + FUS), although the benefit is most pronounced for low dose regimens [84]. This effect appeared to be due to an increase in TMZ concentration and retention time (2.7- and 1.5-fold, respectively) in sonicated tissue [84]. These studies demonstrate that BBBD with FUS can enhance the efficacy of even “gold standard” drugs in GBM.

While able to cross the BBB, carmustine (bis-chloroethylnitrosourea, BCNU) is another small molecule chemotherapeutic drug whose effectiveness could be enhanced by improved local delivery, as it is highly toxic and degrades within 15 minutes. BCNU-loaded polifeprosan 20 Gliadel wafers were one of the first uses of biodegradable polymers for drug delivery in humans [97–100], representing a unique solution to the problems posed by BCNU. Disappointingly, Gliadel produced only mild improvements in patient survival (2.3 months compared to placebo), and is now only recommended for patients with fully resectable tumors[100]. Given these limitations, FUS-mediated delivery has been hypothesized to provide similar benefits. To this end, intravenous BCNU has been administered in conjunction with FUS-mediated BBBD, which doubled BCNU deposition in a C6 glioma model. It was shown that this combined treatment provides better tumor growth control and improved animal survival (IST_median=86% compared to control) [101]. This study also indicates that it may be possible to decrease the intravenous dose administered while maintaining therapeutically relevant drug concentrations in the brain, thus reducing systemic toxicity effects without the need for surgery. Ultimately, it is evident that BBBD with FUS can improve outcomes, even when used in conjunction with drugs that are able to cross the BBB, by increasing local drug concentrations and decreasing systemic toxicity.

PEGylated liposomal doxorubicin formulations are able to extravasate and collect in tumors, and they have been used in the treatment of glioma with some success (see Section 2.2). The liposomal formulation is necessary since free doxorubicin (DOX) exhibits systemic toxicity and is unable to cross the BBB [102]. FUS-mediated BBBD may facilitate the use of free DOX, generating high intratumor drug concentrations while preventing systemic toxicities associated with the liposomal formulation. While FUS was capable of delivering up to 17-fold increases in DOX concentration in healthy brain tissue, in the GL261 mouse model of GBM, treatment with FUS and free DOX increased DOX concentrations in the tumors by only 4-fold compared to contralateral controls, although this increase was significant. Animals treated with FUS + free DOX had improved survival times (IST_median = 68%) as well, and did not show effects of systemic drug toxicities [105]. While this work indicates that FUS can improve the delivery of free drug across the blood-tumor barrier, it is also apparent that FUS parameters may need to be optimized for tumor biology rather than healthy brain tissue.

Boron neutron capture therapy (BNCT) offers the ability to eradicate tumor cells without damaging healthy tissue, a characteristic which is particularly appealing for brain applications. It relies on the accumulation of a stable boron isotope in the tumor tissue, followed by irradiation with low-energy neutrons. The accumulated boron absorbs the neutrons and releases high energy particles, destroying the tumor cells [106]. BNCT has achieved some success in head-and-neck cancers [107], as well as GBM [108,109], but it is believed that FUS-mediated BBBD may improve BNCT efficacy by increasing the concentration of boron in the tumor tissue. Several rodent studies [91,110,111] have demonstrated that FUS significantly increases the concentration of BPA-f, a boron containing drug, in tumor tissue, as well as homogenizing distribution [91]. Nonetheless, it remains to be seen whether this increase correlates to an improvement in treatment efficacy.

Therapeutic antibodies, while currently showing promise in the treatment of numerous cancers, are too large to cross the BBB. Therefore, antibodies which have shown success against various cancers are not beneficial for patients with brain metastases [112], and antibodies designed to treat neurodegenerative diseases require a delivery system [113]. Early work indicated that FUS-mediated BBBD could be used to deliver endogenous IgG antibodies [114] as well as functionally intact D(4) receptor targeting antibody[115], opening the door for therapeutic applications. An exciting recent study in the TgCRND8 model of Alzheimer’s disease showed that FUS-mediated BBBD increased glial cell activation and the delivery of endogenous IgG and IgM antibodies, which led to a reduction in plaque load [68]. Further work in this model demonstrated the delivery of anti-Aβ across the BBB (Figure 5A), which then bound to the plaques (Figure 5B) and caused a significant 23% decrease in plaque surface area (Figure 5C). Plaque number and size were both decreased in the anti-Aβ + FUS group [86]. An earlier study in the APPswe/PSEN1dE9 Alzheimer’s model indicated that FUS-mediated BBBD produces a 3-fold increase in plaque-bound anti-Aβ compared to non-sonicated tissue [116]. These studies suggest the potential use of FUS-mediated antibody delivery for the treatment of neurodegenerative disorders. Indeed, compared to other transcranial delivery methods, FUS is particularly suited for the long term repeated treatments necessitated by the nature of these disorders due to its noninvasive application and highly localized effects.

Trastuzumab (Herceptin), a monoclonal antibody which binds to the Her2 receptor, has shown promise in the treatment of breast cancer [112], which frequently metastasizes to the brain. In healthy animals, FUS-mediated BBBD significantly increased the delivery of trastuzumab in sonicated tissue with no apparent toxicity [117]. In a rat model of breast cancer brain metastasis, animals receiving FUS + trastuzumab had significantly smaller tumors (4 of 10 tumors resolved completely) with an IST_median of 32% compared to untreated controls. Commonly, patients with brain metastases are omitted from clinical trials, as many therapeutics which work well against the primary tumor do not cross the BBB and have no efficacy against brain metastases. FUS-mediated delivery across the BBB can significantly improve drug delivery and efficacy in the brain, and may permit a wider range of treatment options for patients with brain metastases.

Neurotrophic factor administration has been shown to ameliorate a variety of CNS disorders, including schizophrenia [118], depression [119], autism [120], and Parkinson’s [121], and FUS delivery can improve delivery and distribution of neurotrophic factors across the BBB. BDNF, which shows promise as a neuroprotective agent [121], maintains its bioactivity after FUS-mediated delivery across the BBB and generates significant downstream signaling activity[122]. Neurturin (NTN), another factor that has been identified as a potential therapy for neurodegenerative diseases [121], has also been delivered successfully [122]. FUS-mediated delivery increased NTN bioavailability by 25-fold compared to direct injection, and activation of signaling downstream of NTN indicated retention of function [87]. Nonetheless, despite the success with BDNF and NTN, glial cell-line derived neurotrophic factor (GDNF) continues to pose problems for FUS-based delivery. One study [123] demonstrated a significant increase in the delivery of GDNF in FUS-treated regions; however, another was unable to detect GDNF delivery across the BBB due to rapid breakdown in the bloodstream [122].

Immunotherapy is especially intriguing for brain tumor applications, because toxicities associated with traditional drugs pose significant problems for healthy brain tissue[124,125]. However, the presence of the BBB confounds most traditional immunotherapeutic approaches. FUS-mediated delivery of immunostimulatory interleukin-12 (IL-12) significantly increased IL-12 deposition in intracranial C6 gliomas, improved tumor growth control and increased survival (IST_median=43%) [126]. This effect was attributed to a significant improvement in the cytotoxic T-lymphocyte/regulatory T-cell ratio in the FUS + IL-12 group, presumably due to a combination of increased IL-12 concentration and vascular permeability, which permitted enhanced cytotoxic T-lymphocyte infiltration [126]. With the success of recent immunotherapy trials, we speculate that FUS-mediated immunotherapy delivery may permit the inclusion of patients with brain metastases who would normally be denied treatment and ultimately represent a turning point in how brain metastases are treated.

2.2 Liposomes

Liposomal drug formulations are popular due to their versatility and biocompatibility [127]. Their structure, comprised of an aqueous core and a lipid shell, permits the loading of both hydrophilic and hydrophobic drugs [128], and the formulation of the lipid shell can be easily modulated for PEGylation [129], thermosensitivity, and/or targeting [130]. Furthermore, both the size and composition of the liposome can be altered to control circulation time and degradation rate [131]. Liposomes are particularly beneficial for packaging highly toxic drugs, since encapsulated drugs are not bioavailable. Conversely, their larger size makes them more difficult to deliver and FUS may trigger release of the drug payload. The liposomal formulation of doxorubicin, a potent anthracycline, was one of the first drug delivery systems used in combination with FUS[88,83,90,132–134]. Treat et al demonstrated that a single treatment combining FUS and liposomal DOX delayed tumor growth and improved survival time (IST_median=24% compared to 16% for liposomal DOX alone) in a rat gliosarcoma model. Later work by the same group showed that 3 weekly FUS + liposomal DOX treatments drastically improved survival (Figure 6B) compared to the liposomal DOX-only group (IST_median=100% and 16%, respectively), with complete tumor resolution (Figure 6A) in several animals in the FUS + liposomal DOX group [83]. Nonetheless, several animals did suffer from side effects, including skin toxicity, neural loss and intratumoral hemorrhage [83]. To verify that the combination of FUS and liposomal DOX was not causing additional toxicity, a safety study in healthy animals was conducted that demonstrated only minor damage at the focus in animals that received both liposomal DOX and FUS, believed to be due to high local concentrations of DOX deposited by aggressive FUS settings. Of note, the authors also demonstrated that administering liposomal DOX after treatment caused a 32% decrease in DOX delivery across the BBB, a finding we have substantiated with 60 nm polymeric NPs (unpublished studies). A study with animals bearing bilateral 9L gliosarcomas indicated that even late stage tumors benefit from FUS-mediated delivery, with treated tumors showing a two-fold increase in DOX concentration compared to unsonicated controls [88]. FUS treatment also significantly increased the delivery of tumor targeted liposomal DOX formulations in an intracranial mouse xenograft model [90,133], while decreasing some elements of DOX-related toxicity [133], presumably due to lower levels of drug in circulation post sonication. While it is still unclear whether intact liposomes cross the BBB, it is clear that the combination of liposomal encapsulation and FUS-mediated delivery provide excellent therapeutic results, increasing drug concentrations at the target while minimizing systemic toxicities.

2.3 Drug Loaded Microbubbles

Microbubbles can also be functionalized for use as drug delivery vehicles. Although drug loading is limited to the lipid or protein shell, the relatively large surface area (~50 µm²) permits conjugation for both targeting and therapeutics[135]. Because of its hydrophobicity, BCNU has been incorporated into the shell of lipid MBs with some success[89,136–138]. Encapsulation of BCNU within the MB’s lipid shell permitted simultaneous BBB opening and local drug delivery similar to that seen with unencapsulated drug, with the added benefit of increased tissue retention time at the target[138]. Treatment with BCNU-MBs and FUS showed excellent tumor control 30 days post inoculation and median survival time was increased by 12% compared to controls in a C6 glioma model [138]. The addition of VEGF-R2 to the BCNU-MBs provided antiangiogenic targeting capabilities and further improvements in tumor control and animal survival [89]. The same group also developed DOX-loaded MBs conjugated with superparamagnetic iron oxide (SPIO) NPs[139], which showed a two-fold increase in DOX deposition within a rat glioma compared to a non-sonicated control. FUS treatment followed by magnetic targeting also deposited SPIO NPs released from the MBs within the tumor tissue, permitting MR-based treatment monitoring. While drug-loaded MBs offer the benefit of highly localized delivery, they may also require higher pressure to release the drug and are limited to the circulation time of the MB itself.

3.1 Naked Plasmid Delivery

Anionic plasmid DNA can be electrostatically bound to cationic MBs to create MB-DNA carriers. As a result, DNA will be immediately exposed to the vasculature being disrupted by FUS-activated MBs, potentiating DNA extravasation and trans-BBB delivery. Several studies have shown that linking the plasmid DNA to the MB will enhance the transfection compared to free circulating plasmid delivered with FUS[135,164–166]. Interestingly, MRIgFUS exposure to MB-DNA carriers bearing a gene for eGFP led to a significant enhancement of transgene expression in neurons in a young mouse model[162]. Using a similar system, it was shown a MB-DNA carrier bearing a gene for BDNF led to a ~20-fold increase in BDNF protein content[163]. Unfortunately, very high doses of plasmid DNA were required due to susceptibility to degradation from nucleases in the blood and the cell, which reduce the efficiency of this vector system.

3.2 Adeno-Associated Virus

Adeno-associated virus (AAV), with its small ~20 nm size, transduction efficiency, and limited immunogenicity, is a well suited vector for delivery applications across the BBB. Indeed, some AAV vectors, like the self-complementary AAV9 (scAAV9) vector, are able to cross the BBB and mediate global transgene expression in the brain after intravenous injection. However, very high doses of scAAV9 are required, with up to 1×10¹¹ vg/g found to only transduce 19% of motor neurons in adult mice[167]. In contrast, FUS-mediated BBBD can yield transduction efficiencies of 80% in the brain [161] and 87% in the spinal cord (Figure 7) [168] at doses as low as 2.5×10⁹ or 2×10⁹ vg/g, respectively. This marks a robust enhancement of transgene expression in the CNS after intravenous administration of scAAV9. In each case, transgene expression was localized to the anatomical location targeted with FUS. In addition to scAAV9, other studies have shown the delivery of AAV1[169,170] or AAV2[171] across the BBB with FUS. Importantly, these studies showed that transgene expression can be limited to neurons through the use of the synapsin promoter[169]. Moreover, it was found that a transgene driven by the CMV promoter and packaged into the AAV2 capsid led to predominantly astrocytic expression after delivery with FUS[171], in contrast to intracranial injection of the same vector, which led to mostly neuronal expression[172,173]. To this end, it has been suggested that FUS could alter cellular receptor concentrations, including heparan sulfate proteoglycans. This receptor, in addition to being the cellular receptor for AAV2[174] is also known to have roles in the CNS injury response[175].

Despite its efficiency, AAV has shown significant limitations when considered in the context of CNS gene delivery applications. Concerns about safety, limited packaging capacity, difficulties in scale-up and high production costs limit its scope as a long-term solution to CNS gene delivery. Furthermore, repeated administration of AAV leads to production of neutralizing antibody immune responses that may ultimately reduce the efficiency of the vector[176,177]. Finally, scAAV vectors and AAV vectors have packaging capacities of just 2.4 kb or 4.8 kb [178], respectively, which hampers the versatility of this vector.