If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Ghent Research Group on Nanomedicines, Lab for General Biochemistry and Physical Pharmacy, Department of Pharmaceutical Sciences, Ghent University, Ghent, BelgiumLaboratory for Molecular and Cellular Therapy, Medical School of the Vrije Universiteit Brussel, Jette, BelgiumCancer Research Institute Ghent (CRIG), Ghent University Hospital, Ghent University, Ghent, Belgium
Department of Biomedical Engineering, Thoraxcenter, Erasmus MC University Medical Center Rotterdam, Rotterdam, The NetherlandsLaboratory of Acoustical Wavefield Imaging, Faculty of Applied Sciences, Delft University of Technology, Delft, The Netherlands
Physical Sciences Platform, Sunnybrook Research Institute, Toronto, Ontario, CanadaDepartment of Medical Biophysics, University of Toronto, Toronto, Ontario, CanadaInstitute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, Canada
Ghent Research Group on Nanomedicines, Lab for General Biochemistry and Physical Pharmacy, Department of Pharmaceutical Sciences, Ghent University, Ghent, BelgiumCancer Research Institute Ghent (CRIG), Ghent University Hospital, Ghent University, Ghent, Belgium
Department of Biomedical Engineering, College of Engineering and Applied Sciences, University of Cincinnati, Cincinnati, OH, USADepartment of Internal Medicine, Division of Cardiovascular Health and Disease, University of Cincinnati, Cincinnati, OH, USA
Therapeutic ultrasound strategies that harness the mechanical activity of cavitation nuclei for beneficial tissue bio-effects are actively under development. The mechanical oscillations of circulating microbubbles, the most widely investigated cavitation nuclei, which may also encapsulate or shield a therapeutic agent in the bloodstream, trigger and promote localized uptake. Oscillating microbubbles can create stresses either on nearby tissue or in surrounding fluid to enhance drug penetration and efficacy in the brain, spinal cord, vasculature, immune system, biofilm or tumors. This review summarizes recent investigations that have elucidated interactions of ultrasound and cavitation nuclei with cells, the treatment of tumors, immunotherapy, the blood–brain and blood–spinal cord barriers, sonothrombolysis, cardiovascular drug delivery and sonobactericide. In particular, an overview of salient ultrasound features, drug delivery vehicles, therapeutic transport routes and pre-clinical and clinical studies is provided. Successful implementation of ultrasound and cavitation nuclei-mediated drug delivery has the potential to change the way drugs are administered systemically, resulting in more effective therapeutics and less-invasive treatments.
Around the start of the European Symposium on Ultrasound Contrast Agents, ultrasound-responsive cavitation nuclei were reported to have therapeutic potential. Thrombolysis was reported to be accelerated in vitro (
). Since then, many research groups have investigated the use of cavitation nuclei for multiple forms of therapy, including tissue ablation and drug and gene delivery. In the early years, the most widely investigated cavitation nuclei were gas microbubbles, ∼1–10 µm in diameter and coated with a stabilizing shell, whereas today both solid and liquid nuclei, which can be as small as a few hundred nanometers, are also being investigated. Drugs can be co-administered with the cavitation nuclei or loaded in or on them (
). This review focuses on the latest insights into cavitation nuclei for therapy and drug delivery from the physical and biological mechanisms of bubble–cell interaction to pre-clinical (both in vitro and in vivo) and clinical (time span: 2014-2019) studies, with particular emphasis on the key clinical applications. The applications covered in this review are the treatment of tumors, immunotherapy, blood–brain barrier (BBB) and blood–spinal cord barrier, dissolution of clots, cardiovascular drug delivery and treatment of bacterial infections.
Cavitation nuclei for therapy
The most widely used cavitation nuclei are phospholipid-coated microbubbles with a gas core. For the 128 pre-clinical studies included in the treatment sections of this review, the commercially available and clinically approved Definity (Luminity in Europe; octafluoropropane gas core, phospholipid coating) (
) microbubbles were the most frequently used (in 22 studies). Definity was used for studies on all applications discussed here, mostly for opening the BBB (12 studies). SonoVue (Lumason in the United States) is commercially available and clinically approved as well (sulfur hexafluoride gas core, phospholipid coating) (
. Custom-made microbubbles are as diverse as their applications, with special characteristics tailored to enhance different therapeutic strategies. Different types of gasses were used as the core such as air (e.g.,
) for treatment of cardiovascular disease. The main phospholipid component of custom-made microbubbles is usually a phosphatidylcholine such as 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), used in 13 studies (e.g.,
). Both methods produce a population of microbubbles that is polydisperse in size. Monodispersed microbubbles produced by microfluidics have recently been developed, and are starting to gain attention for pre-clinical therapeutic studies.
. To improve drug delivery, therapeutics can be either co-administered with or loaded onto the microbubbles. One strategy for loading is to create microbubbles stabilized by drug-containing polymeric nanoparticles around a gas core (
). The cationic microbubbles utilized in the treatment sections of this review were used mostly for vascular drug delivery, with genetic material loaded on the microbubble surface by charge coupling (e.g.,
The physics of the interaction between bubbles or droplets and cells are described as these are the main cavitation nuclei used for drug delivery and therapy.
Physics of microbubble–cell interaction
Being filled with gas and/or vapor makes bubbles highly responsive to changes in pressure, and hence, exposure to ultrasound can cause rapid and dramatic changes in their volume. These volume changes in turn give rise to an array of mechanical, thermal and chemical phenomena that can significantly influence the bubbles’ immediate environment and mediate therapeutic effects. For the sake of simplicity, these phenomena are discussed in the context of a single bubble. It is important to note, however, that biological effects are typically produced by a population of bubbles and the influence of inter-bubble interactions should not be neglected.
A bubble in a liquid is subject to multiple competing influences: the driving pressure of the imposed ultrasound field; the hydrostatic pressure imposed by the surrounding liquid; the pressure of the gas and/or vapor inside the bubble; surface tension and the influence of any coating material; the inertia of the surrounding fluid; and damping caused by the viscosity of the surrounding fluid and/or coating, thermal conduction and/or acoustic radiation.
The motion of the bubble is determined primarily by the competition between the liquid inertia and the internal gas pressure. This competition can be characterized by using the Rayleigh–Plesset equation for bubble dynamics to compare the relative contributions of the terms describing inertia and pressure to the acceleration of the bubble wall (
where R is the time-dependent bubble radius with initial value Ro, pG is the pressure of the gas inside the bubble, p∞ is the combined hydrostatic and time-varying pressure in the liquid, σ is the surface tension at the gas–liquid interface, ρL is the liquid density, IF is inertia factor and PF the pressure factor.
) identified two scenarios: If the PF is dominant when the bubble approaches its minimum size, then the bubble will undergo sustained volume oscillations. If the inertia term is dominant (IF), then the bubble will undergo inertial collapse, similar to an empty cavity, after which it may rebound or it may disintegrate. Which of these scenarios occurs is dependent upon the bubble expansion ratio Rmax/Ro and, hence, the bubble size and the amplitude and frequency of the applied ultrasound field.
Both inertial and non-inertial bubble oscillations can give rise to multiple phenomena that affect the bubble's immediate environment and hence are important for therapy. These include:
Direct impingement: Even at moderate amplitudes of oscillation, the acceleration of the bubble wall may be sufficient to impose significant forces on nearby surfaces, easily deforming fragile structures such as biological cell membranes (
Ballistic motion: In addition to oscillating, the bubble may undergo translation as a result of the pressure gradient in the fluid generated by a propagating ultrasound wave (primary radiation force). Because of their high compressibility, bubbles may travel at significant velocities, sufficient to push them toward targets for improved local deposition of a drug (
Microstreaming: When a structure oscillates in a viscous fluid there will be a transfer of momentum as a result of interfacial friction. Any asymmetry in the oscillation will result in a net motion of that fluid in the immediate vicinity of the structure known as microstreaming (
). This motion will in turn impose shear stresses upon any nearby surfaces, as well as increase convection within the fluid. Because of the inherently non-linear nature of bubble oscillations (eqn ), both non-inertial and inertial cavitation can produce significant microstreaming, resulting in fluid velocities on the order of 1 mm/s (
Microjetting: Another phenomenon associated with non-spherical bubble oscillations near a surface is the generation of a liquid jet during bubble collapse. If there is sufficient asymmetry in the acceleration of the fluid on either side of the collapsing bubble, then the more rapidly moving fluid may deform the bubble into a toroidal shape, causing a high-velocity jet to be emitted on the opposite side. Microjetting has been reported to be capable of producing pitting even in highly resilient materials such as steel (
, in many cases a bubble will be sufficiently confined that microjetting will have an impact on surrounding structures regardless of jet direction.
Shock waves: An inertially collapsing cavity that results in supersonic bubble wall velocities creates a significant discontinuity in the pressure in the surrounding liquid leading to the emission of a shock wave, which may impose significant stresses on nearby structures.
Secondary radiation force: At smaller amplitudes of oscillation, a bubble will also generate a pressure wave in the surrounding fluid. If the bubble is adjacent to a surface, interaction between this wave and its reflection from the surface leads to a pressure gradient in the liquid and a secondary radiation force on the bubble. As with microjetting, the elastic properties of the boundary will determine the phase difference between the radiated and reflected waves and, hence, whether the bubbles move toward or away from the surface. Motion toward the surface may amplify the effects of phenomena 1, 3 and 6.
As described above, an oscillating microbubble will re-radiate energy from the incident ultrasound field in the form of a spherical pressure wave. In addition, the non-linear character of the microbubble oscillations will lead to the re-radiation of energy over a range of frequencies. At moderate driving pressures, the bubble spectrum will contain integer multiples (harmonics) of the driving frequency; and at higher pressures, also fractional components (sub- and ultraharmonics). In biological tissue, absorption of ultrasound increases with frequency and this non-linear behavior thus also increases the rate of heating (
). Bubbles will also dissipate energy as a result of viscous friction in the liquid and thermal conduction from the gas core, the temperature of which increases during compression. Which mechanism is dominant depends on the size of the bubble, the driving conditions and the viscosity of the medium. Thermal damping is, however, typically negligible in biomedical applications of ultrasound as the time constant associated with heat transfer is much longer than the period of the microbubble oscillations (
The temperature rise produced in the surrounding tissue will be negligible compared with that occurring inside the bubble, especially during inertial collapse when it may reach several thousand Kelvin (
). The gas pressure similarly increases significantly. Although only sustained for a very brief period, these extreme conditions can produce highly reactive chemical species, in particular reactive oxygen species (ROS), as well as the emission of electromagnetic radiation (sonoluminescence). ROS have been reported to play a significant role in multiple biological processes (
Droplets consist of an encapsulated quantity of a volatile liquid, such as perfluorobutane (boiling point: –1.7°C) or perfluoropentane (boiling point: 29°C), which is in a superheated state at body temperature. Superheated state means that although the volatile liquids have a boiling point below 37°C, these droplets remain in the liquid phase and do not exhibit spontaneous vaporization after injection. Vaporization can be achieved instead by exposure to ultrasound of significant amplitude via a process known as acoustic droplet vaporization (ADV) (
) by way of ADV. The mechanism behind this is that the emerging bubbles give rise to similar radiation forces and microstreaming as described earlier in the Physics of the Microbubble–Cell Interaction. It should be noted that oxygen is taken up during bubble growth (
), usually obtained by strong focusing of the generated beam, high frequency of the emitted wave and a relatively long distance between the transducer and the droplet. Another observation that has been made with micrometer-sized droplets is that vaporization often starts at a well-defined nucleation spot near the side of the droplet where the acoustic wave impinges (
). These facts can be explained by considering the two mechanisms that play a role in achieving a large peak negative pressure inside the droplet: acoustic focusing and non-linear ultrasound propagation (
). In the following, lengths and sizes are related to the wavelength, that is, the distance traveled by a wave in one oscillation (e.g., a 1-MHz ultrasound wave that is traveling in water with a wave speed, c, of 1500 m/s has a wavelength, w (m), of c/f = 1500/106 = 0.0015, that is, 1.5 mm.
Because the speed of sound in perfluorocarbon liquids is significantly lower than that in water or tissue, refraction of the incident wave will occur at the interface between these fluids, and the spherical shape of the droplet will give rise to focusing. The assessment of this focusing effect is not straightforward because the traditional way of describing these phenomena with rays that propagate along straight lines (the ray approach) holds only for objects that are much larger than the applied wavelength. In the current case, the frequency of a typical ultrasound wave used for insonification is in the order of 1–5 MHz, yielding wavelengths in the order of 1500–300 µm, while a droplet will be smaller by two to four orders of magnitude. In addition, using the ray approach, the lower speed of sound in perfluorocarbon would yield a focal spot near the backside of the droplet, which is in contradiction to observations. The correct way to treat the focusing effect is to solve the full diffraction problem by decomposing the incident wave, the wave reflected by the droplet and the wave transmitted into the droplet into a series of spherical waves. For each spherical wave, the spherical reflection and transmission coefficients can be derived. Superposition of all the spherical waves yields the pressure inside the droplet. Nevertheless, when this approach is only applied to an incident wave with the frequency that is emitted by the transducer, this will lead neither to the right nucleation spot nor to sufficient negative pressure for vaporization. Nanoscale droplets may be too small to make effective use of the focusing mechanism, and ADV is therefore less dependent on the frequency.
Non-linear ultrasound propagation
High pressure amplitudes, high frequencies and long propagation distances all promote non-linear propagation of an acoustic wave (
). In the time domain, non-linear propagation manifests as an increasing deformation of the shape of the ultrasound wave with distance traveled. In the frequency domain, this translates to increasing harmonic content, that is, frequencies that are multiples of the driving frequency. The total incident acoustic pressure p(t) at the position of a nanodroplet can therefore be written as
where n is the number of a harmonic, an and ϕn are the amplitude and phase of this harmonic and ω is the angular frequency of the emitted wave. The wavelength of a harmonic wave is a fraction of the emitted wavelength.
The aforementioned effects are both important in the case of ADV and should therefore be combined. This implies that first the amplitudes and phases of the incident non-linear ultrasound wave at the droplet location should be computed. Next, for each harmonic, the diffraction problem should be solved in terms of spherical harmonics. Adding the diffracted waves inside the droplet with the proper amplitude and phase will then yield the total pressure in the droplet. Figure 1 illustrates that the combined effects of non-linear propagation and diffraction can cause a dramatic amplification of the peak negative pressure in the micrometer-sized droplet, sufficient for triggering droplet vaporization (
). Moreover, the location of the negative pressure peak also agrees with the observed nucleation spot.
After vaporization has started, the growth of the emerging bubble is limited by inertia and heat transfer. In the absence of the heat transfer limitation, the inertia of the fluid that surrounds the bubble limits the rate of bubble growth, which is linearly proportional to time and inversely proportional to the square root of the density of the surrounding fluid. When inertia is neglected, thermal diffusion is the limiting factor in the transport of heat to drive the endothermic vaporization process of perfluorocarbon, causing the radius of the bubble to increase with the square root of time. In reality, both processes occur simultaneously, where the inertia effect is dominant at the early stage and the diffusion effect is dominant at the later stage of bubble growth. The final size that is reached by a bubble depends on the time that a bubble can expand, that is, on the duration of the negative cycle of the insonifying pressure wave. It is therefore expected that lower insonification frequencies give rise to larger maximum bubble size. Thus, irrespective of their influence on triggering ADV, lower frequencies would lead to more violent inertial cavitation effects and cause more biological damage, as experimentally observed for droplets with a radius in the order of 100 nm (
Biological mechanisms and bio-effects of ultrasound-activated cavitation nuclei
The biological phenomena of sonoporation (i.e., membrane pore formation), stimulated endocytosis and opening of cell–cell contacts and the bio-effects of intracellular calcium transients, ROS generation, cell membrane potential change and cytoskeleton changes have been observed for several years (
). However, other bio-effects induced by ultrasound-activated cavitation nuclei have recently been discovered. These include membrane blebbing as a recovery mechanism for reversible sonoporation (both for ultrasound-activated microbubbles [
). At the same time, more insight has been gained into the origin of the bio-effects, largely through the use of live cell microscopy. For sonoporation, real-time membrane pore opening and closure dynamics were revealed with pores <30 µm2 closing within 1 min, while pores >100 µm2 did not reseal (
revealed that enhanced clathrin-mediated endocytosis and fluid-phase endocytosis occur through distinct signaling mechanisms upon ultrasound and microbubble treatment. The majority of these bio-effects have been observed in in vitro models using largely non-endothelial cells and may therefore not be directly relevant to in vivo tissue, where intravascular micron-sized cavitation nuclei will only have contact with endothelial cells and circulating blood cells. On the other hand, the mechanistic studies by
do reveal translation from in vitro to in vivo. In these studies, ultrasound-activated microbubbles were found to induce a shear-dependent increase in intravascular adenosine triphosphate (ATP) from both endothelial cells and erythrocytes, an increase in intramuscular nitric oxide and downstream signaling through both nitric oxide and prostaglandins, which resulted in augmentation of muscle blood flow. Ultrasound settings were similar, namely, 1.3 MHz, mechanical index (MI) 1.3 for
, with MI defined as MI = P–/ , where P_ is the derated peak negative pressure of the ultrasound wave (in MPa) and f the center frequency of the ultrasound wave (in MHz).
Whether or not there is a direct relationship between the type of microbubble oscillation and specific bio-effects remains to be elucidated, although more insight has been gained through ultrahigh-speed imaging of the microbubble behavior in conjunction with live cell microscopy. For example, there seems to be a microbubble excursion threshold above which sonoporation occurs (
further found that displacement of targeted microbubbles enhanced reversible sonoporation and preserved cell viability, whilst microbubbles that did not displace were identified as the main contributors to cell death.
All of the aforementioned biological observations, mechanisms and effects relate to eukaryotic cells. Study of the biological effects of cavitation on, for example, bacteria is in its infancy, but studies suggest that sonoporation can be achieved in Gram-negative bacteria, with dextran uptake and gene transfection being reported in Fusobacterium nucleatum (
). The findings are conflicting because although they all reveal a reduction in expression of genes involved in biofilm formation and resistance to antibiotics, an increase in expression of genes involved with dispersion and detachment of biofilms was also found (
), less attention has been paid to the interactions between microbubbles and cells or their impact upon drug transport. Currently there are no models that describe the interactions between microbubbles, cells and drug molecules. Several models have been proposed for the microbubble–cell interaction in sonoporation focusing on different aspects: cell expansion and microbubble jet velocity (
propose that the microbubble-generated shear stress does not induce pore formation, but is instead due to microbubble fusion with the membrane and subsequent “pull out” of cell membrane lipid molecules by the oscillating microbubble. Models for pore formation (e.g.,
, that relates the uptake of the model drug propidium iodide (PI) to the size of the created membrane pore and the pore resealing time for a single cell in an in vitro setting. The model describes the intracellular fluorescence intensity of PI as a function of time, F(t), by
where α is the coefficient that relates the amount of PI molecules to the fluorescence intensity of PI-DNA and PI-RNA, D is the diffusion coefficient of PI, C0 is the extracellular PI concentration, r0 is the initial radius of the pore, β is the pore re-sealing coefficient and t is time. The coefficient α is determined by the sensitivity of the fluorescence imaging system, and if unknown, the equation can still be used because it is the pore size coefficient, α·πDC0·r0, that determines the initial slope of the PI uptake pattern and is the scaling factor for the exponential increase. A cell with a large pore will have a steep initial slope of PI uptake, and the maximum PI intensity quickly reaches the plateau value. A limitation of this model is that eqn (3) is based on 2-D free diffusion models, which holds for PI-RNA but not for PI-DNA because the latter is confined to the nucleus. The model is independent of cell type, as Fan et al. have reported agreement with experimental results in both kidney (
). Other researchers have also used this model for endothelial cell studies and also classified the distribution of both the pore size and pore resealing coefficients using principal component analysis (PCA) to determine whether cells were reversibly or irreversibly sonoporated. In the context of BBB opening,
have modeled the microbubble-generated shear and circumferential wall stress for 5-µm microvessels upon microbubble oscillation at a fixed MI of 0.134 for a range of frequencies (0.5, 1 and 1.5 MHz). The wall stresses were dependent upon microbubble size (range investigated: 2–18 µm in diameter) and ultrasound frequency.
have also modelled the wall shear stress generated by microbubble (2 µm in diameter) destruction at 3 MHz for larger microvessels (200 µm in diameter). The presence of red blood cells was included in the model and was found to cause confinement of pressure and shear gradients to the vicinity of the microbubble. Advances in methods for imaging microbubble–cell interactions will facilitate the development of more sophisticated mechanistic models.
Treatment of tumors (non-brain)
The structure of tumor tissue varies significantly from that of healthy tissue which has important implications for its treatment. To support the continuous expansion of neoplastic cells, the formation of new vessels (i.e., angiogenesis) is needed (
). As such, a rapidly developed, poorly organized vasculature with enlarged vascular openings arises. Between these vessels, large avascular regions exist, which are characterized by a dense extracellular matrix, high interstitial pressure, low pH and hypoxia. Moreover, a local immunosuppressive environment is formed, preventing possible anti-tumor activity by the immune system.
Notwithstanding the growing knowledge of the pathophysiology of tumors, treatment remains challenging. Chemotherapeutic drugs are typically administered to abolish the rapidly dividing cancer cells. Yet, their cytotoxic effects are not limited to cancer cells, causing dose-limiting off-target effects. To overcome this hurdle, chemotherapeutics are often encapsulated in nano-sized carriers, that is, nanoparticles, that are designed to specifically diffuse through the large openings of tumor vasculature, while being excluded from healthy tissue by normal blood vessels (
). In addition, drug penetration into the deeper layers of the tumor can be constrained by high interstitial pressure and a dense extracellular matrix in the tumor. Furthermore, acidic and hypoxic regions limit the efficacy of radiation- and chemotherapy-based treatments because of biochemical effects (
). Ultrasound-triggered microbubbles are able to alter the tumor environment locally, thereby improving drug delivery to tumors. These alterations are schematically represented in Figure 2 and include improving vascular permeability, modifying the tumor perfusion, reducing local hypoxia and overcoming the high interstitial pressure.
Several studies have found that ultrasound-driven microbubbles improved delivery of chemotherapeutic agents in tumors, which resulted in increased anti-tumor effects (
). Moreover, several gene products could be effectively delivered to tumor cells via ultrasound-driven microbubbles, resulting in a downregulation of tumor-specific pathways and an inhibition in tumor growth (
furthermore confirmed that nanoparticle accumulation can be achieved in tumors with low EPR effect. Drug transport and distribution through the dense tumor matrix and into regions with elevated interstitial pressure are often the limiting factors in peripheral tumors. As a result, several reports have indicated that drug penetration into the tumor remained limited after sonoporation, which may impede the eradication of the entire tumor tissue (
). These effects were seen at higher ultrasound intensities (>1.5 MPa) and are believed to result from inertial cavitation leading to violent microbubble collapses. As blood supply is needed to maintain tumor growth, vascular disruption might form a different approach to cease tumor development. Microbubble-induced microvascular damage was able to complement the direct effects of chemotherapeutics and antivascular drugs by secondary ischemia-mediated cytotoxicity, which led to tumor growth inhibition (
). In addition, a synergistic effect between radiation therapy and ultrasound-stimulated microbubble treatment was observed, as radiation therapy also induces secondary cell death by endothelial apoptosis and vascular damage (
Furthermore, oxygen-containing microbubbles can provide a local oxygen supply to hypoxic areas, rendering oxygen-dependent treatments more effective. This is of interest for sonodynamic therapy, which is based on the production of cytotoxic ROS by a sonosensitizing agent upon activation by ultrasound in the presence of oxygen (
). As ultrasound can be used to stimulate the release of oxygen from oxygen-carrying microbubbles while simultaneously activating a sonosensitizer, this approach has been reported to be particularly useful for the treatment of hypoxic tumor types (
). Additionally, low oxygenation promotes resistance to radiotherapy, which can be circumvented by a momentary supply of oxygen. Based on this notion, oxygen-carrying microbubbles were used to improve the outcome of radiotherapy in a rat fibrosarcoma model (
Apart from their ability to improve tumor uptake, microbubbles can be used as ultrasound-responsive drug carriers to reduce the off-target effects of chemotherapeutics. By loading the drugs or drug-containing nanoparticles directly into or onto the microbubbles, a spatial and temporal control of drug release can be obtained, thereby reducing exposure to other parts of the body (
). Additionally, tumor neovasculature expresses specific surface receptors that can be targeted by specific ligands. Adding such targeting moieties to the surface of (drug-loaded) microbubbles improves site-targeted delivery and has been found to potentiate this effect further (
Phase-shifting droplets and gas-stabilizing solid agents (e.g., nanocups) have the unique ability to benefit from both EPR-mediated accumulation in the “leaky” parts of the tumor vasculature because of their small sizes, as well as from ultrasound-induced permeabilization of the tissue structure (
). A different approach to the use of droplets for tumor treatment is ACT, which is based on microbubble-droplet clusters that upon ultrasound exposure, undergo a phase shift to create large bubbles that can transiently block capillaries (
Although microbubble-based drug delivery to solid tumors shows great promise, it also faces important challenges. The ultrasound parameters used in in vivo studies highly vary between research groups, and no consensus was found on the oscillation regime that is believed to be responsible for the observed effects (
). To elucidate these effects further, fundamental in vitro research remains important. Therefore, novel in vitro models that more accurately mimic the complexity of the in vivo tumor environment are currently being explored.
engineered a perfusable vessel-on-a-chip system and reported successful doxorubicin delivery to the endothelial cells lining this microvascular network. While such microfluidic chips could be extremely useful to study the interactions of microbubbles with the endothelial cell barrier, special care of the material of the chambers should be taken to avoid ultrasound reflections and standing waves (
). Apart from expanding the knowledge on microbubble–tissue interactions in detailed parametric studies in vitro, it will be crucial to obtain improved control over the microbubble behavior in vivo, and link this to the therapeutic effects. To this end, passive cavitation detection to monitor microbubble cavitation behavior in real time is currently under development, and could provide better insights in the future (
). Efforts are being committed to construction of custom-built delivery systems, which can be equipped with multiple transducers allowing drug delivery guided by ultrasound imaging and/or passive cavitation detection (
The tolerability and therapeutic potential of improved chemotherapeutic drug delivery using microbubbles and ultrasound were first investigated for the treatment of inoperable pancreatic ductal adenocarcinoma at Haukeland University Hospital, Norway (
). In this clinical trial, gemcitabine was administered by intravenous injection over 30 min. During the last 10 min of chemotherapy, an abdominal echography was performed to locate the position of pancreatic tumor. At the end of chemotherapy, 0.5 mL of SonoVue microbubbles followed by 5 mL saline was intravenously injected every 3.5 min to ensure their presence throughout the whole sonoporation treatment. Pancreatic tumors were exposed to ultrasound (1.9 MHz, MI 0.2, 1% DC) using a 4C curvilinear probe (GE Healthcare) connected to an LOGIQ 9 clinical ultrasound scanner. The cumulative ultrasound exposure was only 18.9 s. All clinical data indicated that microbubble-mediated gemcitabine delivery did not induce any serious adverse events in comparison to chemotherapy alone. At the same time, tumor size and development were characterized according to the Response Evaluation Criteria in Solid Tumors (RECIST) criteria. In addition, Eastern Cooperative Oncology Group performance status was used to monitor the therapeutic efficacy of microbubble-mediated gemcitabine delivery. All 10 patients tolerated an increased number of gemcitabine cycles compared with treatment with chemotherapy alone from historical controls (8.3 ± 6 vs. 13.8 ± 5.6 cycles, p < 0.008), thus reflecting an improved physical state. After 12 treatment cycles, one patient's tumor exhibited a twofold decrease in tumor size. This patient was excluded from this clinical trial to be treated with radiotherapy and then with pancreatectomy. In 5 of the 10 patients, the maximum tumor diameter was partially decreased from the first to last therapeutic treatment. Subsequently, a consolidative radiotherapy or a FOLFIRINOX treatment, a bolus and infusion of 5-fluorouracil, leucovorin, irinotecan and oxaliplatin, was offered to them. The median survival was significantly increased from 8.9 to 17.6 mo (p = 0.0001). Together, these results indicate that drug delivery using clinically approved microbubbles, chemotherapeutics and ultrasound is feasible and compatible with respect to clinical procedures. Nevertheless, the authors did not provide any evidence that the improved therapeutic efficacy of gemcitabine was related to an increase in intra-tumoral bioavailability of the drug. In addition, the effects of microbubble-assisted ultrasound treatment alone on tumor growth were not investigated, while recent publications describe that according to the ultrasound parameters, such treatment could induce a significant decrease in tumor volume through a reduction in tumor perfusion as described above.
Hepatic metastases from the digestive system
A tolerability study of chemotherapeutic delivery using microbubble-assisted ultrasound for the treatment of liver metastases from gastrointestinal tumors and pancreatic carcinoma was conducted at Beijing Cancer Hospital, China (
). Thirty minutes after intravenous infusion of chemotherapy (for both monotherapy and combination therapy), 1 mL of SonoVue microbubbles was intravenously administered and was repeated another five times in 20 min. An ultrasound probe (C1-5 abdominal convex probe; GE Healthcare, USA) was positioned on the tumor lesion, which was exposed to ultrasound at different MIs (0.4–1) in contrast mode using a LogiQ E9 scanner (GE Healthcare, USA). The primary aims of this clinical trial were to evaluate the tolerability of this therapeutic procedure and to explore the largest MI and ultrasound treatment time that cancer patients can tolerate. According to the clinical tolerability evaluation, all 12 patients exhibited no serious adverse events. The authors reported that the microbubble-mediated chemotherapy led to fever in 2 patients. However, there is no clear evidence this is related to the microbubble and ultrasound treatment. Indeed, in the absence of direct comparison of these results with a historical group of patients receiving the chemotherapy on its own, one cannot rule out a direct link between the fever and the chemotherapy alone. All adverse side effects were resolved with symptomatic medication. In addition, the severity of side effects did not worsen with increases in MI, suggesting that microbubble-mediated chemotherapy is a tolerable procedure. The secondary aims were to assess the efficacy of this therapeutic protocol using contrast-enhanced computed tomography (CT) and magnetic resonance imaging (MRI). Thus, tumor size and development were characterized according to the RECIST criteria. Half of the patients had stable disease, and one patient obtained a partial response after the first treatment cycle. The median progression-free survival was 91 d. However, comparison and interpretation of results are very difficult because none of the patients were treated with the same chemotherapeutics, MI and/or number of treatment cycles. The results of tolerability and efficacy evaluations should be compared with those for patients receiving the chemotherapy on its own to clearly identify the therapeutic benefit of combining therapy with ultrasound-driven microbubbles. Similar to the pancreatic clinical study, no direct evidence of enhanced therapeutic bioavailability of the chemotherapeutic drug after the treatment was provided. This investigation is all the more important as the ultrasound and microbubble treatment was applied 30 min after intravenous chemotherapy (for both monotherapy and combination therapy) independently of drug pharmacokinetics and metabolism.
Ongoing and upcoming clinical trials
Currently, two clinical trials are ongoing: (i) Professor F. Kiessling (RWTH Aachen University, Germany) proposes examining whether the exposure of early primary breast cancer to microbubble-assisted ultrasound during neoadjuvant chemotherapy results in increased tumor regression in comparison to that after ultrasound treatment alone (NCT03385200). (ii) Dr. J. Eisenbrey (Sidney Kimmel Cancer Center, Thomas Jefferson University, USA) is investigating the therapeutic potential of perflutren protein type A microspheres in combination with microbubble-assisted ultrasound in radioembolization therapy of liver cancer (NCT03199274).
A proof of concept study (NCT03458975) has been set in Tours Hospital, France, for treating non-resectable liver metastases. The aim of this trial is to perform a feasibility study with the development of a dedicated ultrasound imaging and delivery probe with a therapy protocol optimized for patients with hepatic metastases of colorectal cancer and who are eligible for monoclonal antibodies in combination with chemotherapy. A dedicated 1.5-D ultrasound probe has been developed and interconnected to a modified Aixplorer imaging platform (Supersonic Imagine, Aix-en-Provence, France). The primary objective of the study is to determine the rate of objective response at 2 mo for lesions receiving optimized and targeted delivery of systemic chemotherapy combining bevacizumab and FOLFIRI compared with those treated with only the systemic chemotherapy regimen. The secondary objective is to determine the tolerability of this local approach of optimized intra-tumoral drug delivery during the 3 mo of follow-up, by assessing tumor necrosis, tumor vascularity and pharmacokinetics of bevacizumab and by profiling cytokine expression spatially.
Cancer immunotherapy is considered to be one of the most promising strategies to eradicate cancer as it makes use of the patient's own immune system to selectively attack and destroy tumor cells. It is a common name that refers to a variety of strategies that aim to unleash the power of the immune system by either boosting antitumoral immune responses or flagging tumor cells to make them more visible to the immune system. The principle is that tumors express specific tumor antigens which are not expressed or expressed to a much lesser extent by normal somatic cells and hence can be used to initiate a cancer-specific immune response. In this section we aim to give insight into how microbubbles and ultrasound have been applied as useful tools to initiate or sustain different types of cancer immunotherapy, as illustrated in Figure 3.
) discovered the dendritic cell (DC) in 1973, its central role in the initiation of immunity made it an attractive target to evoke specific antitumoral immune responses. Indeed, these cells very efficiently capture antigens and present them to T lymphocytes in major histocompatibility complexes (MHCs), thereby bridging the innate and adaptive immune systems. More specifically, exogenous antigens engulfed via the endolysosomal pathway are largely presented to CD4+ T cells via MHC-II, whereas endogenous, cytoplasmic proteins are shuttled to MHC-I molecules for presentation to CD8+ cells. As such, either CD4+ helper T cells or CD8+ cytotoxic T-cell responses are induced. The understanding of this pivotal role played by DCs formed the basis for DC-based vaccination, where a patient's DCs are isolated, modified ex vivo to present tumor antigens and re-administered as a cellular vaccine. DC-based therapeutics, however, suffer from a number of challenges, of which the expensive and lengthy ex vivo procedure for antigen loading and activation of DCs is the most prominent (
reported that intact microbubbles are rapidly phagocytosed by both murine and human DCs, resulting in rapid and efficient uptake of surface-coupled antigens without the use of ultrasound. Subcutaneous injection of microbubbles loaded with the model antigen ovalbumin (OVA) resulted in the activation of both CD8+ and CD4+ T cells. Effectively, these T-cell responses could partially protect vaccinated mice against an OVA-expressing Listeria infection.
investigated a different approach, making use of messenger RNA (mRNA)-loaded microbubbles combined with ultrasound to transfect DCs. As such, they were able to deliver mRNA encoding both tumor antigens and immunomodulating molecules directly to the cytoplasm of the DCs. As a result, preferential presentation of antigen fragments in MHC-I complexes was ensured, favoring the induction of CD8+ cytotoxic T cells. In a therapeutic vaccination study in mice bearing OVA-expressing tumors, injection of mRNA-sonoporated DCs caused a pronounced slowdown of tumor growth and induced complete tumor regression in 30% of the vaccinated animals. Interestingly, in humans, intradermally injected microbubbles have been used as sentinel lymph node detectors as they can easily drain from peripheral sites to the afferent lymph nodes (
found that mRNA-loaded microbubbles were able to rapidly and efficiently migrate to the afferent lymph nodes after intradermal injection in healthy dogs. Unfortunately, further translation of this concept to an in vivo setting is not straightforward, as it prompts the use of less accessible large animal models (e.g., pigs, dogs). Indeed, conversely to what has been reported in humans, lymphatic drainage of subcutaneously injected microbubbles is very limited in the small animal models typically used in pre-clinical research (mice and rats), which is the result of substantial differences in lymphatic physiology.
Another strategy in cancer immunotherapy is adoptive cell therapy, in which ex vivo manipulated immune effector cells, mainly T cells and natural killer (NK) cells, are employed to generate a robust and selective anticancer immune response (
). These strategies have mainly led to successes in hematological malignancies, not only because of the availability of selective target antigens, but also because of the accessibility of the malignant cells (
). By contrast, in solid tumors, and especially in brain cancers, inadequate homing of cytotoxic T cells or NK cells to the tumor proved to be one of the main reasons for the low success rates, making the degree of tumor infiltration an important factor in disease prognosis (
, who used a xenograft HER-2-expressing breast cancer brain metastasis model to determine whether ultrasound and microbubbles could allow intravenously infused NK cells to cross the BBB. By loading the NK cells with superparamagnetic iron oxide nanoparticles, the accumulation of NK cells in the brain could be tracked and quantified via MRI. An enhanced accumulation of NK cells was found when the cells were injected immediately before BBB disruption. Importantly NK cells retained their activity and ultrasound treatment resulted in a sufficient NK-to-tumor cell ratio to allow effective tumor cell killing (
). In contrast, very few NK cells reached the tumor site when BBB disruption was absent or performed before NK cell infusion. Although it is not known for certain why timing had such a significant impact on NK extravasation, it is likely that the most effective transfer to the tissue occurs at the time of insonification, and that the barrier is most open during this time (
). Possible other explanations include the difference in size of the temporal BBB openings or a possible alternation in the expression of specific leukocyte adhesion molecules by the BBB disruption, thus facilitating the translocation of NK cells. Also, for tumors where BBB crossing is not an issue, ultrasound has been used to improve delivery of cellular therapeutics.
in a more recent publication where the homing of NK cells more than doubled after microbubble injection and ultrasound treatment of an ovarian tumor. Despite the enhanced accumulation, however, the authors did not observe an improved therapeutic effect, which might be owing to the limited number of treatments that were applied or the immunosuppressive tumor microenvironment that counteracts the cytotoxic action of the NK cells.
There is growing interest in exploring the effect of microbubbles and ultrasound on the tumor microenvironment, as recent work has indicated that BBB disruption with microbubbles and ultrasound may induce sterile inflammation. Although a strong inflammatory response may be detrimental in the case of drug delivery across the BBB, it might be interesting to further study this inflammatory response in solid tumors as it might induce the release of damage-associated molecular patterns (DAMPS) such as heat-shock proteins and inflammatory cytokines. This could shift the balance toward a more inflammatory microenvironment that could promote immunotherapeutic approaches. As reported by
. In their study, ultrasound treatment caused a complete shutdown of tumor vasculature followed by the expression of hypoxia-inducible factor 1α (HIF-1α), a marker of tumor ischemia and tumor necrosis, as well as increased infiltration of T cells. Similar responses have been reported after thermal and mechanical HIFU treatments of solid tumors (
). A detailed review of ablative ultrasound therapies is, however, out of the scope of this review.
At present, the most successful form of immunotherapy is the administration of monoclonal antibodies to inhibit regulatory immune checkpoints that block T-cell action. Examples are cytotoxic T lymphocyte-associated protein 4 (CTLA-4) and programmed cell death 1 (PD-1), which act as brakes on the immune system. Blocking the effect of these brakes can revive and support the function of immune effector cells. Despite the numerous successes achieved with checkpoint inhibitors, responses have been quite heterogeneous as the success of checkpoint inhibition therapy depends largely on the presence of intra-tumoral effector T cells (
to explore the synergy of microbubble and ultrasound treatment with PD-L1 checkpoint inhibition therapy in mice. Tumors in the treatment group that received the combination of microbubble and ultrasound treatment with checkpoint inhibition were significantly smaller than tumors in the monotherapy groups. One mouse exhibited complete tumor regression and remained tumor free upon rechallenge, indicative of an adaptive immune response.
Overall, the number of studies that have investigated the impact of microbubble and ultrasound treatment on immunotherapy is limited, making this a rather unexplored research area. It is obvious that more in-depth research is warranted to improve our understanding on how (various types of) immunotherapy might benefit from (various types of) ultrasound treatment.
BBB and blood–spinal cord barrier opening
The barriers of the central nervous system (CNS), the BBB and blood–spinal cord barrier (BSCB), greatly limit drug-based treatment of CNS disorders. These barriers help to regulate the specialized CNS environment by limiting the passage of most therapeutically relevant molecules (
). Although several methods have been proposed to circumvent the BBB and BSCB, including chemical disruption and the development of molecules engineered to capitalize on receptor-mediated transport (so-called Trojan horse molecules), the use of ultrasound in combination with microbubbles (
) to transiently modulate these barriers has come to the forefront in recent years because of the targeted nature of this approach and its ability to facilitate delivery of a wide range of currently available therapeutics. First demonstrated in 2001 (
), ultrasound-mediated BBB opening has been the topic of several hundred original research articles in the last two decades and, in recent years, has made headlines for groundbreaking clinical trials targeting brain tumors and Alzheimer's disease as described later under Clinical Studies.
Mechanisms, bio-effects and tolerability
Ultrasound in combination with microbubbles can produce permeability changes in the BBB via both enhanced paracellular and transcellular transport (
) have been reported. Although the exact physical mechanisms driving these changes are not known, there are several factors that are hypothesized to contribute to these effects, including direct tensile stresses caused by the expansion and contraction of the bubbles in the lumen, as well as shear stresses at the vessel wall arising from acoustic microstreaming. Recent studies have also investigated the suppression of efflux transporters after ultrasound exposure with microbubbles. A reduction in P-glycoprotein expression (
), and thus this mechanism requires further study.
Many studies have documented enhanced CNS tumor response after ultrasound and microbubble-mediated delivery of drugs across the blood–tumor barrier in rodent models. Improved survival has been observed in both primary (
Beyond simply enhancing drug accumulation in the CNS, several positive bio-effects of ultrasound and microbubble-induced BBB opening have been reported. In rodent models of Alzheimer's disease, numerous positive effects have been discovered in the absence of exogenous therapeutics. These effects include a reduction in amyloid-β plaque load (