Ultrasound-mediated disruption of the blood tumor barrier for improved therapeutic delivery

The blood-brain barrier (BBB) is a major anatomical and physiological barrier limiting the passage of drugs into brain. Central nervous system tumors can impair the BBB by changing the tumor microenvironment leading to the formation of a leaky barrier, known as the blood-tumor barrier (BTB). Despite the change in integrity, the BTB remains effective in preventing delivery of chemotherapy into brain tumors. Focused ultrasound is a unique noninvasive technique that can transiently disrupt the BBB and increase accumulation of drugs within targeted areas of the brain. Herein, we summarize the current understanding of different types of targeted ultrasound mediated BBB/BTB disruption techniques. We also discuss influence of the tumor microenvironment on BBB opening, as well as the role of immunological response following disruption. Lastly, we highlight the gaps between evaluation of the parameters governing opening of the BBB/BTB. A deeper understanding of physical opening of the BBB/BTB and the biological effects following disruption can potentially enhance treatment strategies for patients with brain tumors.


Introduction
Primary and metastatic malignant brain tumors are a leading cause of cancer related deaths in both men and women [ 1 ] conferring a 5-year survival rate of approximately 33 percent [ 1 ]. Despite extensive preclinical efforts in drug development and therapeutic strategies, treatment remains largely palliative. Lack of effective drug delivery and adequate concentration within brain lesions is a significant limitation in therapeutic efficacy. The blood-brain barrier (BBB) is a protective barrier that limits passage of most therapeutics into brain due to its unique anatomical and physiological properties. The BBB limits paracellular diffusion of therapies from blood to brain secondary to tight junction complexes that seal endothelial cells together. Further, molecules that can cross the endothelial membrane are often extruded back into the vascular compartment by richly expressed efflux transporters like P-glycoprotein (P-gp, ABCB1) and breast cancer resistance protein (BCRP, ABCG2). In case of brain tumorigenesis, an important early step is neoangiogenesis where the vessels are poorly formed and leakier (blood-tumor barrier; BTB) than the BBB. Despite being leaky, the BTB still inhibits drug permeability to tumors to the degree that they are largely ineffective. Compensatory mechanisms of the BTB like higher efflux, altered active transport mechanisms and refined fluid dynamics effectively reduce drug permeability across the BTB.
An emerging method to improve drug delivery is focused ultrasound to transiently increase BBB/BTB permeability by modulating the integrity of tight endothelial junctions. Although the early experiences are encouraging, current literature lacks consensus in key experimental conditions, limiting our understanding and wide translation of the technique. In this review, we focus on the gaps in current literature to understand drug distribution when BBB/BTB is transiently disrupted by focused ultrasound. We also discuss the role of tumor associated BTB dysfunction and its immunological influence on focused ultrasound mediated physical disruption. Successful disruption of the BBB/BTB by LIFU can potentially overcome the difficulties in drug delivery to brain tumors.

Blood-brain barrier
The BBB is a unique physiochemical barrier comprised of various cell types, which largely restricts solutes from entering brain from blood. Endothelial cells provide the first barrier between luminal blood flow and abluminal mural cells [ 2 ]. The endothelia connect themselves through tight junction protein complexes to form a contiguous barrier limiting paracellular diffusion of most molecules [ 3 ]. Adding to this physical barrier, ABC efflux transporters are highly expressed at the luminal and abluminal membrane and can remove a wide variety of lipid-soluble molecules through the numerous transporters including BCRP, P-gp, and Multi-drug Resistance Protein-1 (MRP1, ABCC1) [ 3 , 4 ]. Beyond the initial layer, astrocytes reside on the abluminal side of the BBB and support endothelia through end feet contact, maintaining barrier properties. Embedded in the basement membrane surrounding the capillaries are pericytes, which help regulate cerebral blood flow and contribute to the extracellular matrix [ 5 ]. Microglia are resident immune cells of the brain and act as part of the innate immune response. These cells release cytokines in response to a variety of pathological insults that can modify BBB properties [ 6 ]. These cells, collectively known as the neurovascular unit, provide a protective barrier, and allow for local and systemic response for the brain in healthy organisms.

Blood-tumor barrier
The BBB in primary and metastatic brain tumors is anatomically altered and disrupted or "leaky" and referred to as the BTB [ 7 ]. To develop a metastatic brain lesion, it is thought that cancerous cells extravasate from their primary site and invade the brain where they colonize and proliferate. After cells have accumulated, the lesion reaches a hypoxic state requiring neoangiogenesis for further progression [ 8 , 9 ]. During this process, tumor cells secrete VEGF within the hypoxic regions promoting formation and growth of new vessels. The resultant vessels are often abnormal, tortuous, poorly formed, and more permeable compared to the intact BBB [ 7 , 8 ].
Additionally, microvasculature within brain lesions is also disrupted in part by the lack of continuous tight junction proteins creating fenestrations that permit increased solute movement [ 10 ]. During lesion formation, the distribution of mural cells (pericyte, astrocytes, and microglia) around the BTB is often irregular, contributing to increased permeability [ 11 ].
While altered integrity of the BTB allows increased paracellular transport of molecules, efflux processes at both the BBB and BTB may also be increased by the presence of tumor cells. This may explain mechanistically why many chemotherapeutics fail in the treatment of CNS tumors [ 10 , 12 ]. The major facilitator superfamily domain containing 2A (Mfsd2a) is required for BBB formation and function [ 13 ]. In the healthy BBB, Mfsd2a limits transcytosis through modulation of lipid content. Formation and function of caveolae vesicles by brain endothelial cells is prevented by Mfsd2a through control of docosahexaenoic acid transport. However, in the BTB there is a downregulation or complete termination in expression of Mfd2a and other tight junction proteins like ZO-1, claudin-3, claudin-5, and occludin, which have been linked to a higher permeability of the BTB [ 14 , 15 ].
When the BBB is disrupted due to CNS lesions, the change in its integrity is similar to the disrupted BBB observed in CNS inflammatory pathologies, such as multiple sclerosis and neuropsychiatric systemic lupus erythematosus (NPSLE). A clinical study in 2019 showed that the volume transfer constant (K trans ) for contrast-enhancing lesions was nearly 6.5-fold higher as opposed to non-enhancing lesions [ 16 ]. Another study with six NPSLE patients showed significantly higher K trans in the hippocampus than all other regions averaged ( P < 0.001) compared to control patients [ 17 ]. These studies of local and systemic inflammatory disease pathologies are suggestive of the influence of the immune system in disrupting the BBB.

Heterogeneity of the BTB
Despite presence of leaky vessels in metastatic and primary brain tumors, chemotherapeutics only reach cytotoxic concentrations in less than 10% of brain lesions in preclinical models as well as in patients [ 7 , 18 ]. The reduced accumulation of chemotherapies within brain and tumor lesions can be attributed to a few reasons. First, expression of efflux transporters on the luminal membrane of the BBB/BTB and tumor cells markedly inhibits intracellular accumulation of numerous chemotherapeutics [ 19 ].
Secondly, heterogeneity between primary and metastatic brain tumors or different metastatic sites of the same tumor type display varying responses to chemotherapy. Primary and metastatic brain tumors have differential progression, uniquely influencing BTB within the tumor. For example, high grade primary tumors such as glioblastoma have a necrotic core with residual stem cells, a fast-growing central layer and a fully developed envelope which forms the leading edge of the tumor [ 20 ]. These layers have different degrees of hypoxia, proliferation rates, and extent of drug permeation.
Third, much of the heterogeneity is because of the unique environment of each tumor region, causing differential release of HIF1, HIF2, IL8, NF κB [ 21 ]. Cumulatively, these factors influence unequal drug distribution across the tumor mass resulting in significant challenges in treatment with systemically delivered chemotherapy.

Formation of a brain metastasis and it's environment
Metastatic cells from distant peripheral sites can disseminate into the vascular system, penetrate the BBB and ultimately invade the brain parenchyma. It is thought the chemokines Mmp3, Mmp9, TNF α, Cxcl12, IL6, IL10, TGF β promote the metastatic cells ability to infiltrate and proliferate within brain [ 22 ]. Further tumor progression and invasion depends on the cell's ability to interact and co-opt the endothelial cells and astrocytes of the BBB.
Different tumor types promote the formation of supporting vasculature that is variable in terms in the number and size of vascular defects. Gliomas have defects or pores that can be 10 times larger than those observed in brain metastases. This significantly alters the total amount of chemotherapy accumulation in the tumor, defining the upper limit of the size of a drug that can be effective in a CNS tumor. Mechanistically this may explain the ability of antibodies to produce an effect in glioblastoma, but why trastuzumab fails as a therapy for brain metastases of breast cancer [ 23 ].
Further, immune responses in metastases are unique as compared to those seen in primary gliomas. Brain metastases show decreased concentrations of T-lymphocytes with higher expression of PDL-2 and HLA-1, facilitating the

Disruption of BBB/ BTB
There is a significant need to design BBB/BTB disrupting techniques to overcome the challenges of delivery of therapeutic agents to target sites within the brain. Current approaches for BBB disruption include intracarotid injection of a hyperosmotic solution of mannitol, intraparenchymal injection of drug via catheters, radiation-mediated BBB disruption, and use of microbubbles in conjunction with transcranial ultrasound [ 19 ]. Among these, microbubble-enhanced focused ultrasound (FUS) is the least invasive, can focally target small brain structures and may have little toxicity on adjacent normal brain cells. Clinically, there appears to be minimal neurotoxicity, inflammation and stroke occurrences associated with the technique [ 24 ].
There are numerous studies showing FUS can open the BBB; however, they report variations in BBB opening parameters including power, energy related dose, duration, timing and cycles. Variability within literature limits the ability to provide consensus about optimal parameters needed to deliver a specific therapeutic predictably and repeatedly in targeted regions of the brain ( Table 1 ).

Ultrasound applications within brain
In the following sections, we will discuss the multiple forms of ultrasound that have been used for therapy and/or augmentation of therapy for tumors within the CNS. We will also highlight the physiologic and immunological response to the opening as well as the ability of the technique to improve drug distribution and effect in tumors. In general, ultrasound is specifically targeted to areas within the CNS using intra-treatment magnetic resonance imaging (MRI). Multiple wavelengths or "intensity" of the ultrasound wave, with and without vascular microbubbles are used for a variety of applications including ablation of small brain regions or opening of the BBB or BTB.

High intensity focused ultrasound
High intensity focused ultrasound (HIFU) utilizes a stereotactic device to distribute high intensity energy (100-10,000 W/cm 2 ) through the skull. This produces spatial ablation at target tumor sites by increasing the temperature to approximately 55 0 C. Cell death is induced by the thermal energy deposited, frictional vibration between cells or non-thermal pulsed changes in peak rarefaction pressure amplitude [ 25 ]. High intensity focused ultrasound is currently FDA approved for essential tremor and tremor dominant Parkinson's disease to create an ablation in thalamus to modulate the neural circuitry of the tremor [ 26 , 27 ]. The current HIFU system is limited in terms of brain tissue volume that can be ablated, which is an important consideration for tumor ablation [ 26 ]. Interstitial HIFU is an alternative to the traditional technique to circumvent this limitation. Here, single or multi-elemental catheters with cylindrical cooling elements deliver high ultrasound energy within the parenchyma of intracranial neoplasms. In a swine model this method was highly effective for tumor ablation [ 28 ]. A unique advantage of interstitial HIFU is that it can be used for theranostic purposes using a cannula and catheter for simultaneous biopsy and treatment. Further, it limits issues related to near-field heating or patient motion during longer therapy sessions by providing the ability to tailor heating patterns that conform to the tumor allowing precision in treatment margins [ 28 ].
While HIFU is clinically used for ablation, evidence from pre-clinical models suggests there are secondary immunomodulatory effects of the tumor microenvironment post tumor ablation [ 25 ]. The anti-tumor immunological response possibly arises from activation of the dendritic cells along with an increase in the CD4 + , CD3 + as well as the ratio of CD4 + /CD8 + cells in the blood [ 25 , 29 , 30 ]. Currently, the combination of HIFU with PDL-1 antibody blockade is being investigated clinically in solid tumors outside the CNS (NCT04116320). The effect of HIFU combined with immunotherapy in CNS tumors remains to be evaluated.

Low intensity focused ultrasound
Targeted disruption of the BBB or BTB can be achieved using low intensity focused ultrasound (LIFU) at lower frequencies. In this technique, ultrasound waves are co-exposed with intravenously administered gas-filled bubbles that are composed of perfluorocarbon encapsulated in phospholipid formulations. These microbubbles undergo stable oscillations to produce a transient vessel permeabilization [ 31 ]. Due to the mechanical effect, and its non-invasive nature, LIFU may eventually substitute other procedures such as transcranial magnetic stimulation or deep brain stimulation which potentially risk strong immune response or infection [ 32 ]. LIFU combined with advanced imaging modalities such as dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) has allowed additional insight and therapeutic applications to different pathologies including primary and metastatic brain tumors ( Table 2 ).

Mechanism of LIFU mediated BBB disruption
Disruption of the BBB at the vascular endothelia has been described by multiple mechanisms of interaction of LIFU and microbubbles, but none have been confirmed. A primary hypothesis suggests ultrasound waves force microbubbles to oscillate, resulting in increased vessel pressure, tight junction expansion and increased membrane permeability. A second hypothesis not exclusive of the first, suggests microbubble oscillation can activate and increase expression of cellular receptors or transcytoplasmic shuttling vesicles. This potentially increases transcellular permeability through a caveolin dependent mechanism where transport across arterioles and endothelia is increased by vesicular fusion and formation of transcytotic channels [ 33 , 34 ].
In both cases, 2 types of microbubble oscillations have been described. Under the influence of ultrasound, microbubbles can produce stable (noninertial), and inertial oscillations, which are termed as cavitations. Effect of the cavitations on the BBB can be defined by a mechanical index; the negative acoustic pressure over the square root of the frequency, or the cavitation index; the negative acoustic pressure over frequency [ 35 ]. The mechanical index defines biological effects produced mechanically by sonication, while cavitation index measures the scale of stable cavitation involved in FUS induced opening [ 35 ]. At low mechanical indices, microbubbles oscillate in a linear and uniform way and produce harmonic or sub-harmonic emissions. These oscillations are equivalent to the mechanical index applied ( Fig. 2 ) [ 36 ].
Stable cavitations are produced as a result of an equal amount of gas efflux and influx within the microbubbles causing their rhythmic expansion and contraction. When microbubble expansion occurs, there is a stretching of the vessel which may open cell-cell junctions transiently [ 37 ]. Oscillation of microbubbles produces micro-streams which induces shear stress on vascular endothelia resulting in increased rate of endocytosis ( Fig. 2 B) [ 38 ].
Stable cavitation can also cause acoustic radiation forces, where microbubbles are pushed towards endothelia resulting in a "kneading" or pounding effect leading to increased passive permeability [ 37 ].
Not all cavitation is well contained within the microbubbles. At times, higher mechanical indices will cause microbubbles to oscillate rapidly eventually resulting in bubble fragmentation. Bubble collapse causes microjet formation and small shock waves ( Fig. 2 B) directed at the endothelia, resulting in increased permeability [ 37 ]. Lastly, a more tangential or longer Table 1 Preclinical studies of CNS targeted ultrasound. Ultrasound-mediated disruption of the BTB for improved therapeutic delivery T.A. Arsiwala et al. Ultrasound-mediated disruption of the BTB for improved therapeutic delivery T.A. Arsiwala et al.  acting mechanism assumes that oscillating microbubbles increase local endothelial temperatures, which increases permeability through protein expression changes and not direct mechanical interactions [ 39 ]. While the modulatory effects of LIFU on endothelial cells is currently being investigated, effects on other cells of the neurovascular unit remains largely unknown. Preliminary studies indicate that mechanical disruption of the BBB leads to transient activation of microglia and astrocytes mediated by inflammatory mechanisms that can last upto 24 h postdisruption [ 40 ]. While increased clearance through astrocytic and microglial phagocytosis is expected post-LIFU, change in other homeostatic roles is yet to be investigated. For example: it is known that astrocytes may play an important role in cerebrospinal fluid clearance through AQP4 channels [ 41 ]. However, the effect of LIFU on these channels is unknown. There is also evidence of reduction in arterial blood flow post LIFU potentially mediated by neurovascular coupling, vasospasms, disrupted neurovascular signaling and suppression of neuronal response [ 41 ]. Change in clearance of CNS active drugs through these mechanisms post-LIFU needs to be elucidated.

Immune effects in BBB ultrasound LIFU disruption
The brain has been considered an immune-privileged site since the early 1900s through studies demonstrating tissue transplants into the brain parenchyma could occur without host rejection, despite a differential Neoplasia Vol. 23, No. 7, 2021 Ultrasound-mediated disruption of the BTB for improved therapeutic delivery T.A. Arsiwala et al.   immunological response in peripheral implants. It is widely accepted there is immunosurveillance within the brain, which can elicit strong immune responses [ 42 ]. Unfortunately, immunotherapies have largely been ineffective in treating CNS tumors due to poor penetration of therapeutics across the BBB and subsequently lack of activation of the CNS immune system. Further, often in high grade CNS tumors, such as glioblastoma multiforme (GBM) there is decreased effector T cells and increased T regulatory cells which shifts the tumor microenvironment to immunosuppressive and promotes tumor growth ( Fig. 3 ) [ 30 ].
There is a gap in understanding of the exact mechanism behind local and systemic post ultrasound immunomodulation. It is hypothesized that damage-associated molecular patterns (DAMPs) are produced by endothelial cells in response to the microbubbles cavitation to activate and recruit proinflammatory immune cells [ 43 ]. Changes in the local and systemic immunological environment significantly impacts BBB permeability [ 44 ]. Using LIFU to mechanically open the BBB results in influx of inflammatory cells and markers into the brain, potentiating further BBB disruption and promoting immune cell activation [42][43][44]. Ultrasound and microbubble mediated cavitations at tight-junctions induce changes in the expression of integral proteins, Ca 2 + influx and transient detachment of endothelia from the extracellular matrix [ 43 , 44 ]. A recent study by Hynynen et al. provided evidence of peripheral immune cell recruitment at the BBB immediately after LIFU mediated sonication [ 45 ]. Interaction between vascular endothelia and oscillating microbubbles causes an immediate but transient response by circulating neutrophils [ 41 ]. Further, the initial infiltration leads to an acute inflammatory cascade by release of chemokines and recruitment of more immune cells such as monocytes and phagocytes [ 45 ].
Initiation of immune responses due to physical changes within the vasculature can increase permeability across the barrier. Microbubble mediated disruption has also been implicated in slowing down blood perfusion by vasoconstricting the larger vessels, which may also contribute to increased BBB permeability [ 46 ]. The vasoconstricted vessels may mediate hypoxic stress responses through increased levels of heat shock protein 40, VEGF, erythropoietin, IL1 α, IL1 β and TNF α in the parenchyma [ 43 ] ( Fig. 3 ).
Recent studies show BBB opening using LIFU induces sterile inflammation for a minimum of 24 h [ 43 ]. Although not completely understood, the underlying mechanism may be through immediate triggered release of damage associated molecular patterns (DAMPs) from endothelia [ 43 ]. DAMPs like HMGB1 may induce sterile inflammation via the NF κB pathway and can be correlated with increased BBB permeability [ 47 ]. Elevation in levels of pro-inflammatory and tropic factors is also seen concurrently with sterile inflammation. Increased innate immunity responses up to 6 d post sonication is seen through infiltration and the continued presence of CD68 + macrophages. In a preclinical rat glioma model, LIFU increased intra and inter tumoral cytotoxic T cell populations [ 48 ]. When the T-cell activating cytokine IL-12 was administered with LIFU, an increase in cytotoxic T cell to T regulatory cell ratio was observed and appeared to correlate with an increase in overall survival [ 48 ]. Immunomodulation with LIFU may also cause antigen release into the bloodstream from the tumors in the CNS, resulting in an induction of a pro-inflammatory environment [ 49 ]. The antigens released are captured by peripheral antigen presenting cells (APCs) leading to T-cell activation. Primed T-cells then infiltrate the tumor by adhesion to tumor endothelia which results in apoptosis [ 49 ].
Similar immunological changes are observed with other techniques of BBB disruption like radiotherapy where a 10Gy dose in combination with chemotherapy/immune checkpoint inhibitors upregulated proinflammatory markers like CCL2, CCL11, and IL-6 [ 61 ]. The aggregate data suggest there are multiple underlying mechanisms by which the immunological milieu regulates BBB disruption. Further, studies are needed to evaluate balance between pro-inflammator y and anti-inflammator y responses.

Unfocused ultrasound
In contrast to application of LIFU for targeted disruption of the BBB, unfocused ultrasound with microbubble cavitation induces a broad opening of the BBB, which may be advantageous to deliver therapeutics for diffuse pathology. To accomplish this, a transducer is implanted within the skull, which provides a controlled distribution of the ultrasonic energy coupled with lower attenuation by the skull [ 50 ]. Device implantation allows for long-term, repetitive disruption without the need for MRI guidance. Similar to the studies described above, pre-clinical unfocused ultrasound has shown BBB disruption as evidenced by a 4-fold increase in cortical Evans blue concentration in the sonicated hemispheres of rabbits as well as a significant increase in MRI Gd enhancement [ 51 ]. Further studies have demonstrated increased tissue-plasma drug concentration ratio of temozolomide and irinotecan in between control and sonicated hemispheres. Clinical utility was demonstrated in a recent study where BBB disruption at higher acoustic pressures (1.1 MPa) resulted in cortical Gd enhancement without detectable adverse effects [ 52 ]. Despite its potential, unfocused ultrasound is primarily used for diagnostic and imaging purposes due to lack of available data regarding its use to target deep seated structures in the brain [ 53 ].

Factors influencing efficacy of LIFU mediated BBB disruption
Within this section, we will discuss the differences in instrumentation, the set-up and or parameters that influence the ability to produce BBB disruption.
The type of transducer used to emit the ultrasound directly affects the wave's propagation to brain and can profoundly influence efficacy of LIFU BBB disruption. One of the major obstacles in achieving targeted LIFU penetration is the skull's high impedance [ 37 ]. To overcome this, a geometrically archetypal transducer is required to prevent propagated wave distortion due to bone irregularity. Similarly, phase-array transducers reduce skull attenuation and undesirable heating due to improved focal volume dimensions which concentrate ultrasonic energy to focal regions [ 54 ]. Limited repeated application of multi-array transducers to superficial and deep-seated tumor lesions also gave rise to implanted cranial transducers like SonoCloud [ 54 ]. Efficacy of SonoCloud was noted in a recurrent glioblastoma with ultrasonication dose escalation prior to carboplatin administration [ 52 ]. This study revealed pressures up to 1.1 megapascal were well tolerated using pulsed ultrasound with an implanted transducer [ 52 , 55 ]. Other preclinical studies using small animals and primates also suggested higher safety margins and increased drug distribution with the same transducer [ 56 , 57 ]. A separate study suggested 8% yttria-stabilized-zirconia polycrystalline ceramics (8YSZ) as a biocompatible alternative to implantable transducers with an 81% maximum transmission efficiency [ 58 ].
Another factor affecting experimental outcome is the type, formulation and concentration of intravenously administered microbubbles. Multiple data sets have compared the effect of different commercially available, FDA approved microbubbles like SonoVue®, Optison and Definity® which differ in diameter, concentration, sizes, composition and pharmacokinetic parameters [59][60][61]. While some pre-clinical studies have indicated that no significant differences in BBB permeability occur due to microbubble concentrations; others suggest microbubble concentration and dose have differential delivered drug concentrations up to 6-fold and continue to affect BBB permeability up to 6 h post administration [  Optison and Definity depending on polymers and lipids composed within the microbubble shell [ 63 , 64 ]. Other factors influencing ultrasound mediated BBB disruption include pressure amplitude, frequency and power. Pressure amplitude affects resonance of the microbubbles; and appears to directly positively correlate to increasing BBB permeability. It should be noted that a pressure amplitude above of 0.47 MPa at a 300 second exposure time results in irreversible damage to brain endothelia [ 65 ]. Similarly, lower ultrasound frequency results in lower impedance through the skull which allows an increased BBB permeability [ 66 , 67 ]. However, removing the skull via craniotomy results in an opposite relationship: higher frequencies causes greater BBB permeability [ 63 ]. Lastly, increases in pulse repetition frequency from 0.1-20 Hz and pulse length between 0.1 and 20 ms produces enhanced BBB permeation [ 37 , 63 , 68 ].

Drug delivery using LIFU mediated BBB disruption
The extent of BBB opening is often correlated to the size and extent of drug permeation. In a HER-2 positive brain metastases of breast cancer preclinical model a seven-fold higher accumulation for the small molecule (doxorubicin) and a 5-fold increased accumulation of ADC (ado-trastuzumab emtansine T-DM1) was observed post LIFU [ 69 ]. Data from the work suggested small molecule penetration may be related to convective transport, higher diffusion and hydraulic conductivity. While increased penetration was observed for the larger T-DM1 molecule, it was only seen immediately following FUS. Another preclinical study showed 2.35-fold increase in doxorubicin accumulation in FUS treated tumor fractions of a GBM model, with a 3.3-fold higher AUC analyzed by intracerebral microdialysis [ 70 ]. BBB disruption with LIFU was able to improve paclitaxel loaded liposomes and PLGA nanoparticles distribution in a nude mouse model of glioblastoma with higher accumulation for the liposomes [ 71 ]. Increased FUS mediated permeability has been reported for other small molecules such as doxorubicin, irinotecan, cisplatin, temozolomide, cytarabine as well as antibodies such as trastuzumab, pertuzumab, bevacizumab, and their nanoparticulate formulations ( Table 1 ).

Clinical impact of FUS mediated BBB/BTB disruption
While many pre-clinical studies have individually demonstrated opening of the BBB/ BTB, very few trials have provided preliminary proof of clinical implementation of this technology. Clinical application of FUS is currently being tested in a few pathologies including primary tumors, Alzheimer's disease, essential tremor, Parkinson's disease and amyotrophic lateral sclerosis [72][73][74]. One of first clinical reports of HIFU used the ExAblate 3000 which could reduce acoustic backscatter and beam dispersion due to skull thickness variability [ 75 ]. This technology was coupled to an MRI to perform a guided surrogate surgical resection. However, there was insufficient power to generate the higher temperatures needed to produce thermal ablation. A later 2014 study overcame power related issues, but was limited by acoustic impedance through bone attenuation, time required for ablation and effects on healthy tissues [ 76 ]. Recently, many studies have attained pre-clinical success in achieving controlled ablation of the tumoral tissue by modification of acoustic specifications [ 77 , 78 ]. For example, when the transducer parameters including power, duty cycle and frequency are controlled a functional thermal dose can be obtained [ 77 ]. A recent study evaluated BTB disruption using FUS with adjuvant temozolomide and showed accurate and safe opening in patients for six sonication cycles at the same targeted sites [ 74 ]. Another recent investigation reported feasibility of BBB opening within peri-enhancing regions of the brain in six patients with recurrent GBM, using the NaviFUS system with concomitant Sonovue microbubbles [ 79 ]. This analysis suggests a dose-dependent BBB permeability effect of FUS, based on DCE kinetic parameter analysis (K trans and V e ) [ 79 ].
A distinct advantage of FUS is the precise, and reversible on demand BBB opening in the region of interest including the deep brain targets besides being noninvasive. Our group demonstrated safe and reversible opening of BBB in the complex and deep-seated structures of hippocampus and entorhinal cortex in patients with Alzheimer's disease [ 80 ]. The BBB opening was immediate and sizeable consisting of about 29% of hippocampus and the BBB closed within 24-h post sonication [ 80 ]. As evident from the basic science studies suggesting immunomodulation, MRI investigation revealed perivenous enhancement during acute BBB opening which persisted even after BBB closure, suggesting a downstream immunological response bloodmeningeal barrier in patients with Alzheimer's disease [ 73 ].
In the case of primary malignant tumors there have been reports of successful BBB opening and enhanced delivery of chemotherapeutics [ 55 , 81 ]. While initial clinical implementation of FUS has demonstrated safe, reproducible, and repeatable opening of the BBB, the long-term effects of this modality need to be delineated. Available clinical data encompasses various FUS devices and microbubble contrast agent as well as heterogeneous procedural and technical parameters. These preliminary studies have demonstrated the need to understand secondary effects accompanying BBB sonication under fixed procedural parameters. Despite these challenges, it is expected that the use of FUS to deliver therapeutics across the BBB for CNS malignancies and neurological conditions will increase in the coming years .

Conclusion
Magnetic resonance guided focused ultrasound is a non-invasive technique increasingly explored to treat various stages of cancers. An MRI provides detailed anatomical images with the capability of precise targeting a tumor region within the body [ 82 ]. The combination therapy of LIFU and MRI facilitates the localization, targeting, and real-time monitoring while simultaneously minimizing collateral damage to surrounding normal tissues [ 82 ]. The method of disrupting BBB by using LIFU to oscillate intravenously injected microbubbles may improve the distribution and efficacy of therapeutics to brain tumor sites. Additionally, the safety and reproducibility of this technique has been demonstrated by a few preclinical and clinical studies [ 55 , 83 ]. Nevertheless, there are limitations, which includes identifying optimal ultrasound parameters. Suboptimal parameters can induce hemorrhages, erythrocyte extravasation, and edema formation, while weak parameters may not have therapeutic effect [ 37 ]. Further work still needs to be done to understand the correlation between microbubble size (or concentration) and duration of BBB openings [ 37 ]. Other limitations, such as a lack of portability, long duration time for the treatment, an inability to monitor true acoustic cavitation on focused ultrasound therapies and selection of correct statistical methods to normalize data analysis gathered by multiple parameters needs to be overcome [ 83 ]. Despite these limitations, MRI guided focused ultrasound provides a novel way to increase drug distribution to brain through a reversible and on-demand opening of BBB. Further translational clinical studies are needed to explore the potential of this technology.