The presented experimental protocol can be used to perform real time measurements of cavitation activity in a cell culture device with the aim of enabling investigation of the conditions required for successful drug delivery and/or other bioeffects.
Interest in the therapeutic applications of ultrasound is significant and growing, with potential clinical targets ranging from cancer to Alzheimer’s disease. Cavitation – the formation and subsequent motion of bubbles within an ultrasound field – represents a key phenomenon underpinning many of these treatments. There remains, however, considerable uncertainty regarding the detailed mechanisms of action by which cavitation promotes therapeutic effects and there is a need to develop reliable monitoring techniques that can be implemented clinically. In particular, there is significant variation between studies in the exposure parameters reported as successfully delivering therapeutic effects and the corresponding acoustic emissions. The aim of this paper is to provide design guidelines and an experimental protocol using widely available components for performing studies of cavitation-mediated bioeffects, and include real-time acoustic monitoring. It is hoped that the protocol will enable more widespread incorporation of acoustic monitoring into therapeutic ultrasound experiments and facilitate easier comparison across studies of exposure conditions and their correlation to relevant bio-effects.
Ultrasound (US) has been used widely as a diagnostic imaging technique because of its safe and non-invasive nature, its ease of implementation at a patient’s bedside, and its cost effectiveness1. Next to its diagnostic and monitoring capabilities, US has considerable potential for therapeutic applications. Early work explored its use in thrombolysis, DNA transfection, and drug delivery2,3,4 and therapeutic US now represents a very active area of research, with applications including tumor treatment5,6,7, immunotherapy8,9, blood-brain barrier (BBB) disruption10,11,12, thrombolysis13,14,15, and bacterial infection treatment16,17. A key phenomenon underpinning these applications is cavitation: the nucleation, growth, and oscillation of gaseous cavities due to changes in fluid pressure18,19.
There is a range of mechanisms by which cavitation produces biological effects. For example, the highly nonlinear nature of bubble oscillations under the influence of an applied US field can generate microstreaming in the surrounding liquid that can both enhance drug convection20 and exert shear stresses on the tissue in the vicinity of the bubbles. This is particularly prevalent when bubbles are in the vicinity of a boundary, causing bubbles to oscillate non-spherically, and may potentially promote drug uptake through shear-induced permeabilization21,22,23,24. At higher pressures, larger amplitude oscillations and rapid bubble collapse are observed, imparting direct mechanical stress25 and frequently generating shock waves, and consequent large pressure gradients that can disrupt and permeabilize tissues26,27. The collapse of bubbles near a surface can also result in the formation of high-velocity liquid microjets28,29,30. These microjets can penetrate tissue, potentially creating pores or inducing secondary stress waves31,32. The permeabilization of biological membranes at both the tissue and cellular levels is variously referred to as sonophoresis, used primarily in the context of US-induced enhancement in skin permeability33,34, and sonoporation, used mainly to describe the reversible permeabilization of the cellular membrane due to formation of membrane pores35,36.
Viscous absorption in the liquid immediately surrounding the oscillating bubble can produce a substantial heating effect37. Moreover, the highly non-linear oscillations produce acoustic radiation at frequencies higher than the driving US field. This leads to increased absorption in the surrounding tissue and further heating38. Bubble collapse may also be accompanied by chemical effects due to the transient high temperatures and pressures in the bubble core, such as generation of highly reactive species and electromagnetic radiation, known as sonoluminescence32. These effects have been investigated to assess potential damage and/or activation of relevant cellular pathways for delivery39 and exploited in local activation of light-sensitive drugs in an approach known as sonodynamic therapy40,41,42,43.
Many US-mediated bioeffects may be initiated solely through control of US field parameters (pressure amplitude, frequency, pulse length and repetition frequency, and duration of exposure), but reliably generating cavitation in biological tissue often requires high input energies and hence carries an elevated risk of damage. Introduction of exogenous or artificial cavitation nuclei may substantially reduce the input energy required to produce the broad range of effects discussed above and further introduces additional effects that may not be possible with US alone. Cavitation nuclei include gas bubbles26,44, liquid droplets45,46,47 and solid particles48,49,50, with nanoscale cavitation nuclei being an emergent area of investigation for their benefits in terms of prolonged circulation time, improved extravasation and prolonged cavitation activity49,51,52,53.
The most commonly used nuclei are gas microbubbles (MBs), originally used as contrast agents in diagnostic imaging. They are typically 1-2 micrometers in diameter and contain a core of a high-molecular-weight gas with low aqueous solubility in the surrounding medium. The core is surrounded by a protective lipid, protein, or polymer shell most commonly consisting of phospholipids54. When exposed to a US field, the compressibility of the MBs causes them to undergo volumetric oscillations, consequently producing strong acoustic scattering, which is responsible for the success of MBs as a contrast agent. As mentioned, these oscillations also lead to the aforementioned mechanical, thermal, and chemical effects that can be harnessed in therapeutic applications. The MB coating process also offers a mechanism for encapsulating drugs within the MB structure and for attaching drugs and/or targeting species to the MB surface. This technique facilitates the triggered release of drugs to reduce systemic toxicity55. It has also recently been shown that material from the MB surface may be transferred to biological structures, enhancing drug delivery through so called “sonoprinting”56,57,58.
Monitoring of US-mediated cavitation activity can provide insights into the resultant biological effects both in vitro and in vivo and potentially allows for the tuning and optimization of these effects. The two most widely applied methods for monitoring cavitation activity are i) optical, which use ultra high speed video microscopy and are generally not feasible in vivo; and ii) acoustic, which record the re-radiated sound fields produced by oscillating and/or collapsing bubbles. Both the amplitude and frequency components of the acoustic signal contain information on bubble behavior. Low concentrations of bubbles at low incident US amplitudes have been shown to produce predominantly harmonic emissions (integer multiples of the driving frequency)59. As driving pressures increase, the bubble emission spectrum may also contain fractional components known as subharmonics and ultraharmonics60 that indicate stronger nonlinear behavior, as well as broadband noise, which is indicative of inertial cavitation. Integer harmonics are a primary indicator of bubble oscillation but can also be caused by non-linearities anywhere in an experimental system, e.g., due to non-linear propagation. By contrast, fractional harmonics and broadband noise are very strongly correlated with bubble dynamics.
The relationship between bubble behavior and the detected acoustic emissions may be complicated by factors including the incident US field, the nucleation environment, and the characteristics of the detection pathway60. Nevertheless, important information about bubble behavior and their interactions with cells can be gained by discerning trends in frequency and energy in the acoustic spectrum. These data can also provide valuable information that can be used to form the basis for clinical treatment monitoring techniques. To fully exploit this information, the development of robust, translatable, and reproducible experimental methods is required.
Currently there is substantial variation in reported protocols for designing systems and conducting studies to support development of cavitation-assisted therapies. In terms of the apparatus, a range of design approaches has been undertaken. Several groups have made use of parallel-plate chambers56,61,62,63, either custom built or commercially available (e.g., OptiCell, ThermoFisher Scientific). Hu et al. (2013) developed a cell chamber coupled with an US sonication module and real-time confocal imaging64, Carugo et al. (2015) used a system comprising a commercially available cell culture dish with a custom-made PDMS lid to allow for submersion in a water bath during US exposure65, and Pereno et al. (2018) used a device consisting of layered acoustofluidic resonators that allow for simultaneous optical and acoustic characterization of bubble dynamics and bubble-cell interactions66. The use of custom-fabricated and application-specific designs complicates characterization of the US field and other environmental exposure conditions, making cross study comparisons challenging. For example, there is considerable variation in the US parameters identified for achieving successful sonoporation, which include center frequencies ranging from 0.02 to 15 MHz, duty cycles varying from 1% to continuous wave, and rarefactional pressures ranging from 0.1 to 20 MPa23,64,67,68,69,70 (Table 1). There is similarly considerable variation in the spectral components (harmonics, sub-harmonics etc.) that have been identified as being associated with particular bioeffects.
The aim of this work, therefore, is to provide an easily reproducible system design and implementation framework for the in vitro study of cavitation-induced cellular bioeffects with the specific inclusion of a cavitation monitoring capability.
The critical steps for any acoustic measurement were encapsulated by Apfel in 198176 as “know thy liquid, know thy sound field, know when something happens.” In the context of this protocol, these encompass the transducer calibration and alignment and the water preparation and bubble handling steps. First, it is essential that the hydrophone used to calibrate the driving transducer and/or the PCD is itself accurately calibrated through regular external servicing or in-house comparison to a reference standard. Similarly, the response of both the driving transducer and PCD need to be regularly characterized to check for any change in output and/or loss of sensitivity. If the driving conditions and receive sensitivity of the system are unknown, then it will be impossible to infer any meaningful relationship between exposure conditions, bioeffects and acoustic emissions. Directly related to this, the alignment of the transducers to each other and the sample chamber needs to be carefully checked to ensure that the exposure conditions within the chamber are as expected and the sampling volume for the PCD corresponds to the region of interest. As indicated, the temperature and gas content of the suspending medium can affect the final results significantly and consistency is extremely important in this respect77,78. Similarly, the preparation, characterization, and handling of the cavitation agent suspension require very close attention to ensure that the expected size distribution and concentration of particles is present within the sample. For example, if the concentration of bubbles is too high, there will be effective shielding of the sample volume from the incident US field. MB agents are particularly susceptible to destruction and coalescence and further guidance on their handling may be found in Mulvana et. al. (2012)79.
A very common problem with detection of cavitation signals is achieving an adequate SNR. This is due partly to the nature of the signal itself, as described, but may also be due to sources of electrical noise within the experimental set up. Checking the connections between system components, in particular those involving co-axial cables, may help to eliminate some of these. Replacing or repairing co-axial cables may be necessary. Identifying and removing or deactivating other equipment in the laboratory such as pumps that may cause electrical noise can also help. Poor electrical impedance matching between system components can be a further cause of poor signal to noise ratio and also potentially of damage to equipment and should be carefully checked. The triggering settings on the signal generator and oscilloscope should similarly be checked to confirm that they are configured appropriately for the experiment and have not reverted to the manufacturer default settings. If there is significant destruction of bubbles during handling, in case of the SAT2, it may be helpful to attach a second syringe to the outlet port and use this to gently extract fluid from the chamber, thereby drawing in the suspension. This can also help in eliminating macrobubbles or enabling flow during US exposure, if desired.
It is not possible to completely eliminate acoustic reflections within the sample chamber and hence the incident field will not be completely uniform over the whole sample volume. As mentioned in steps 1.3.2 and 1.3.3, the transmissibility of acoustic windows will be frequency dependent and thus the desired bandwidth for acoustic emission measurements should be carefully considered. In particular, there may be significant multiple reflections of higher frequency components. This is another reason why calibration of the field within the fully assembled system is so important for minimizing the uncertainty in incident pressure. Appropriate gating of the recorded signals should also be considered to minimize the effects of multiple reflections. The use of commercial devices for convenience and the need for acoustic transparency means that some optical transparency must be sacrificed. This may impact the quality of subsequent imaging, e.g., to assess cell viability or drug uptake. Some of the membranes used in commercial devices are also porous and, thus, imperfect isolation occurs between the sample chamber and the surrounding water bath. As above, the corresponding risk of contamination can be mitigated by using a smaller sub-chamber, the contents of which can be regularly replaced. The cell culture devices indicated in the Table of Materials are suitable primarily for cell monolayers that may not be representative of tissues in terms of all US/cavitation-mediated bioeffects. The proximity of the cells to a solid surface will also affect MB dynamics in a way that may not be reflective of conditions in vivo, e.g., promoting microstreaming and microjetting as described in the introduction. These limitations can be addressed, however, through a simple substitution of alternative tissue models.
The aim in proposing the SATs is to provide a means of improving the reproducibility of acoustic exposure conditions and acoustic emissions between studies of US-mediated bioeffects, thus hopefully facilitating better understanding of the underlying mechanisms and the development of treatment monitoring techniques to improve safety and efficacy. The systems are designed to be compatible with commercially available cell culture devices, enabling a wide range of biological assays to be performed according to the application of interest and enabling the performance of high throughput experiments, removing the need for time consuming alignment procedures between runs. By standardizing protocols for the characterization of exposure conditions and the capture of acoustic emissions, the system dependent variability can hopefully be reduced. The range of parameters that should be explored for a particular experiment will depend upon the application (desired bio-effect, cell type, depth of target tissue if in vivo etc.) and the nature of any cavitation agent being used. Given the large number of variables (US frequency, pressure amplitude, pulse length, pulse repetition frequency etc.) fully exploring the whole parameter space is unlikely to be practicable. An advantage of the proposed protocol is that it enables some bounds on this parameter space to be quickly established. For example, it enables determination of the minimum pressure at which a cavitation signal is generated, the maximum pressure or pulse length that can be used before cell detachment/death occurs, and the pressure at which fractional harmonics or broadband noise are produced. It is recommended that such a set of scoping measurements be carried out as a first step in any study.
As presented, the SATs are designed for real-time monitoring of acoustic emissions, with biological assays being performed outside of the experiment. It would be relatively simple, however, to modify the SAT to enable direct optical observation of the sample chamber via a microscope objective. This could in turn be coupled to a fluorescence and/or high-speed microscopy system to enable observation of drug uptake and bubble dynamics, for example. The PCD output as currently presented in terms of voltage indicates: i) the types of cavitation behavior and their relative proportions; ii) how long these cavitation behaviors persist; iii) whether the observed time-cumulative exposure characteristics are correlated to a particular bioeffect; and iv) whether the relative levels and time-dependent behaviors are consistent with previous experiments in the exposure system. Whilst the receive sensitivity of the PCD can be quantified, in order to reliably characterize the acoustic emissions in terms of absolute energy, additional spatial information is required. This could be achieved by replacing the PCD with an array probe to implement passive acoustic mapping (PAM)80. This would, however increase the complexity of signal processing and the computational time and power required.
Other instrumentation for measurement of membrane electrical resistance or application of physical targeting methods, for example magnetic fields, could also be incorporated. It would also be possible to use three-dimensional tissue structures such as tumor spheroids, organoids, or even ex vivo tissue samples on acoustically “soft” gel substrates in place of the cell monolayers to study US and cavitation-mediated effects in more realistic tissue environments.
The authors have nothing to disclose.
The authors thank the Engineering and Physical Sciences Research Council for supporting this work through grant EP/L024012/1. VB is also supported by the Engineering and Physical Sciences Research Council (EPSRC) and Medical Research Council (MRC) (grant EP/L016052/1). VB and AV thank the Clarendon Foundation for Post Graduate Scholarships. AV also thanks Exeter College for a Santander scholarship. The authors are indebted to James Fisk and David Salisbury for their invaluable assistance in the manufacturing of the apparatus. They also gratefully acknowledge the contributions of Drs. Dario Carugo and Joshua Owen in the development of earlier prototype SATs.
Absorber | Precision Acoustics | APTFlex F28 panel | 1.0 cm standard thickness |
Amplifier (power) | E&I Ltd. | 1040L | 400W power amplifier to drive ultrasound source |
Amplifier (pre) | Stanford Research Systems | SR445A | Fixed gain multi-stage preamplifier for PCD signals |
Aquarium heater | Aquael | Ultra 50W | Different models for different tank sizes. |
Digitizer | TiePie Engineering | HS5-110-XM | Extended memory option: 32M points per channel |
Hydrophone | Precision Acoustics | FOH | 0.01 mm diameter sensitive area minimises directivity effects |
Microbubbles | Bracco | SonoVue | FDA approved microbubbles |
PCD mirror (SAT3) | Olympus NDT | F-102 | 90 degree beam reflection |
PCD transducer | Olympus NDT | V320-SU | Immersion transducer, 7.5MHz |
PCD waterproof cable | Olympus NDT | BCU-58-1 W | |
PDMS (SAT2 compartment lid) | Corning | Sylgard 184 | See Carugo et al. (2015) for preparation guidelines |
Polymer rod (SAT2 seal) | Zeus | PTFE monofilament | |
Rubber plug (SAT3 lid/seal) | VWR | 391-2101 | 6mm bottom dia., 8mm top dia., red |
Signal generator | Agilent | 33250 | Waveform generator for ultrasound source |
Substrate for cell exposure compartment, SAT2 | Ibidi | µ-Dish 35mm | |
Substrate for cell exposure compartment, SAT3 | Corning | Transwell 6.5mm | |
Ultrasound source (SAT3) | Sonic Concepts | H107 with central hole | Use of a HIFU-capable source allows pressures >1MPa to be generated both at the focus and pre-focally for expanded spatial coverage |