|Home | About | Journals | Submit | Contact Us | Français|
Cell-based therapeutic approaches are attractive for the restoration of the protective endothelial layer in arteries affected by atherosclerosis or following angioplasty and stenting. We have recently demonstrated a novel technique for the delivery of mesenchymal stem cells (MSCs) that are surface-coated with cationic lipid microbubbles (MBs) and displaced by acoustic radiation force (ARF) to a site of arterial injury. The objective of this study was to characterize ultrasound parameters for effective acoustic-based delivery of cell therapy. In vitro experiments were performed in a vascular flow phantom where MB-tagged MSCs were delivered towards the phantom wall using ARF generated with an intravascular ultrasound catheter. The translation motion velocity and adhesion of the MB-cell complexes were analyzed. Experimental data indicated that MSC radial velocity and adhesion to the vessel phantom increased with the time-averaged ultrasound intensity up to 1.65 W/cm2, after which no further significant adhesion was observed. Temperature increase from baseline near the catheter was 5.5 ±0.8°C with this setting. Using higher time-averaged ultrasound intensities may not significantly benefit the adhesion of MB-cell complexes to the target vessel wall (p=NS), but could cause undesirable biological effects such as heating to the MB-cell complexes and surrounding tissue. For the highest time-averaged ultrasound intensity of 6.60 W/cm2, the temperature increase was 11.6 ±1.3°C.
Ultrasound microbubble (MB) contrast agents have been widely used for numerous applications including enhancement of the blood pool for improved endocardial definition on echocardiographic imaging and gene and drug delivery (Unger et al. 2003; Klibanov 2002; Ferrara et al. 2007; Villanueva 2008; Hitchcock and Holland 2010, Qin et al 2009). It has been shown that it is possible to change the flow patterns of microbubbles if the ultrasound imaging parameters create significant acoustic radiation force (ARF) (Dayton et al. 1999; Zhao et al. 2004). Thus, microbubbles and their therapeutic payload can potentially be directed towards a specific target using ultrasound. ARF was first described by King (King 1934) and potential applications of this phenomenon have been suggested. Primary radiation force arises from the phase difference between the driving acoustic pressure and the oscillation of a particle or microbubble in a continuous medium. For axi-symmetric objects like MBs in a plane traveling wave field, the primary ARF is in the direction of sound propagation. The effects of ARF on bubbles in general (Crum 1975, Lee and Wang 1993, and Leighton 1994) and on contrast agents in particular (Dayton 1997, 1999, 2002), have been previously described and form the basis for the study described in this report.
Endothelial progenitor cells (EPCs) and mesenchymal stem cells (MSCs) may participate in effective restoration of functional endothelium following vascular injury (Yue et al. 2008; Forte et al. 2008). However, intra-arterially delivered cells would have to overcome significant hydrodynamic forces to adhere under flow, which is why experimental in vivo delivery of EPCs or MSCs to a specific vascular segment usually requires prolonged cessation of flow to that segment, an ineffective and impractical option from a clinical standpoint. We have previously reported a novel in vivo method for effective targeted delivery of stem cells to a site of arterial injury using ARF (Toma et al. 2011). By tagging MSCs with gas-filled lipid MBs, the MB-cell complexes were selectively pushed with acoustic energy to the endoluminal surface of the arterial wall.
Previous work demonstrated the use of targeted microbubbles for drug and gene delivery enhanced by ultrasound radiation force both in vivo and in vitro. In one of those earlier studies, Rychak et al. (2005) showed that acoustic pressure mechanism was suitable for achieving targeted microbubble delivery in high-flow vessels and quantified the dependence of microbubble adhesion on acoustic pressure and shear rate in vitro. Hallow et al (2007) evaluated the intracellular uptake and cell death when microbubbles and a drug were targeted into vascular endothelial cells (ECs) and smooth muscle cells (SMCs) in the coronary artery using ARF. They concluded that significant intracellular uptake of molecules can be targeted into ECs and SMCs by ultrasound-enhanced delivery. Phillips et al. (2008) suggested that intravascular ultrasound (IVUS) is essential to enhance gene transfection from microbubble carriers to the vessel wall in vivo (Phillips et al. 2008). Their group demonstrated that IVUS catheter mediated plasmid DNA transfection from microbubble carriers to the porcine coronary artery wall following balloon angioplasty. Patil et al. (2009) proposed an approach for simultaneously pushing and imaging microbubbles in a real-time environment. They concluded that for blood vessels of large size, and flow velocities in a certain range, radiation force generated using customized dual frequency sequences was able to translate microbubbles across the vessel resulting in accumulation of the microbubbles at the targeted wall. Another study by Gessner et al. (2009) presented a description of a prototype dual-frequency transducer which enabled simultaneous high-resolution imaging with the application of radiation force to enhance targeting and adhesion in vitro and in vivo. They observed a 7 fold increase in image signal intensity after radiation force enhanced targeting compared to standard rheologically-mediated targeting. Ultrasound based molecular imaging has also been used to image vascular pathology and to guide localized drug delivery in blood vessels (Patil et al. 2011).
Our study differs from previous drug and gene delivery techniques in that we attached the MBs on the surface of vascular cells such as stem cells to form MB-cell complexes before injection and the ARF on these complexes provided a means of delivering the cells to the treatment site. Therefore, the MBs themselves are not targeted, but they serve as carriers for MSCs. We have demonstrated previously that the MSCs could be successfully delivered to the endothelium in the abdominal aorta through in vivo experiments in a rabbit model (Toma et al, 2011).
To foster clinical translation of this novel cell delivery strategy, the objective of the current study was to further analyze and optimize ultrasound parameters for the delivery of biological cells such as stem cells using an in vitro flow model. We investigated how increasing the time-averaged ultrasound intensity delivered to the microbubbles through manipulation of acoustic pressure and duty cycle would modulate the extent of cell adhesion. We also explored whether there was an optimal combination of ultrasound parameters for this ultrasound and flow system that were sufficient to successfully push the majority of MB-labeled cells to the phantom wall. It is essential to define optimal ultrasound parameters since simply increasing time-averaged ultrasound intensities may not substantially improve adhesion of MB-cell complexes to the target vessel wall, but could cause undesirable biological effects such as heating to the MB-cell complexes and surrounding tissue.
The ARF on the MB can be calculated once the dynamic behavior of the bubble is determined. The short-term time-averaged force over one acoustic cycle can be expressed as: where is the energy density of the acoustic wave and the radiation force function is given by . This formulation can be considered as an alternative form of the equation (4.136) in Leighton (1994). Due to the presence of radiation force, the microbubble will experience translational motion (Crum 1975; and Watanabe and Kukita 1993). We are interested in the translational motion of a MB-cell complex when the ARF is used to delivery vascular cells for therapeutic purposes. Due to the much larger size of MSC compared with the MB used, the attainable translational velocity of the MB-cell complex is much smaller when compared with a single MB in the same ultrasound field. The ARF on the MSC can be obtained as a limiting case of the general theory presented by Chen and Apfel (1996). The force is in the direction of acoustic wave propagation and can be expressed as: where Rc is Yp radius of the cell, and is the radiation force function. For the frequency range of relevance, the stem cell can be considered as a Rayleigh scatter, and the radiation force function reduces to: Yp = 4[(γk/3 − γρ/9)2 + 2(γρ/9)2] (klRc)4 where γρ = 3(ρc − ρl)/(2ρc + ρl) is the density contrast between the cell and the host medium, and is the compressibility contrast between the cell and the host medium. For a stem cell with 10 μm radius driven at a 1.7 MHz and 100 kPa of acoustic pressure amplitude, and using the mechanical properties of 0.9% saline as the surrounding medium, the ARF is calculated to be 5.3 × 10−16 N, six orders of magnitude smaller than the ARF on a bubble of concern. Therefore, the ARF on the cell can be ignored when the translational motion of the MB-cell complex is evaluated. Once the ARF on the bubble is determined, either analytically or numerically, the translational motion of the MB-cell complex can be estimated. Strictly speaking, a coupled equation for the radial and translational motion of the MB needs to be considered. This treatment is outside the scope of work presented here.
The studies of radiation force in vitro were performed in a flow phantom that allowed simultaneous optical and acoustical observation of MB-cell complexes. The schematic of the experimental setup is illustrated in Figure 1. The phantom vessel was made of polyvinyl chloride (PVC) tubing with an internal diameter of 3.5 mm. The suspension of MB-cell complexes was prepared in a reservoir and withdrawn through the phantom vessel by a precision syringe pump (PHD 2000, Harvard Apparatus) with a 60 ml syringe (Becton Dickinson). A flow rate of 40 ml/min was chosen such that the wall shear stress (1.1 Pa) approximated that of average coronary blood flow (Doriot et al. 2000).
An intravascular ultrasound catheter with a 2-mm long 1.7 MHz transducer (MicroSonic SV, Ekos Corp) was placed along the center of the phantom. The ultrasound catheter was driven by a power amplifier (Model 100A250A, Amplifier Research) coupled to an arbitrary waveform generator (Model 33250A, Agilent). A calibrated hydrophone (HGL-0200, Onda Corp) was used for calibration of the acoustic system. The phantom vessel was placed under an Olympus IX81 inverted fluorescence microscope (Olympus Corp) equipped with a high sensitivity EMCCD camera (C9100-2, Hamamatsu Photonics K.K.). Digital fluorescence movies of MB-labeled fluorescent MSCs (MB-cell complexes, or MB-MSC complexes in this study) flowing through the phantom were captured at a frame rate of 50 frames/second. The optical focus of the microscope was co-localized with the ultrasound focus for continuous optical observation and recording. A long working distance low magnification Olympus 4x objective was used for visualization of the flow phantom, with a field of view of 2 mm × 2 mm, sufficient to cover one side of the flow phantom from the catheter to the vessel wall.
Fluorescent labeled MB-cell complexes were advanced through the vessel phantom while being visualized and continuously recorded for a total of 20 seconds: 5 seconds before ultrasound was turned on, followed by 10 seconds of ultrasound exposure, and another 5 seconds after ultrasound was turned off. After each experiment, the flow phantom was flushed with phosphate buffered saline (PBS) to remove all the adhered MB-cell complexes from the targeted focal area. Control experiments were performed with MSCs with no MBs.
Cationic lipid MBs were made from a mixture of lipids 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) and 1,2-distearoyl-sn-glycero-3-ethylphosphocholine (DSEPC) (Avanti Polar Lipids) and polyethylene glycerol stearate (Sigma-Aldrich). The DSEPC carries the quaternary ammonium ion, giving the bubble a cationic charge. The mixture dissolved in chloroform was evaporated with a steady Argon gas stream and degassed overnight under house vacuum. The dried lipids mixture was re-hydrolyzed in sterile saline and sonicated in the presence of perfluorocarbon gas. The MB suspension was then washed 3 times with sterile saline to eliminate unbound lipids and smaller microbubbles to reach a desired size distribution. The size distribution of the cationic lipid microbubbles was analyzed using a Multisizer 3 Coulter Counter (Beckman Coulter Inc.) and the charge on the microbubbles was quantified using a Zetasizer Nano-Zs90 (Malvern Instruments).
Low passage human MSCs (Lonza) were expanded in 10% hMSC culture media (Lonza). Prior to use, the cells were fluorescently labeled with 0.5 μM cytoplasmic dye (CellTracker Green, CMFDA, Invitrogen) for 30–60 minutes. The cells were then trypsinized using 1:4 dilution of trypsin–versene (Part 17–161, Lonza) mixture in PBS and washed in PBS.
The viscosity of the buffer solution for MB-MSCs was adjusted to that of blood at body temperature (4×10−3 Pa·s) using polyvinylpyrrolidone (PVP) (Sigma-Aldrich). The kinematic viscosity measurements of the solution were performed using a U-tube, glass capillary viscometer (Rheotek Inc.).
MSCs were mixed for 5 minutes with cationic MBs at a ratio of 1:40, previously experimentally shown by us to be optimal for achieving ARF-induced MB-MSC adhesion to a vascular surface (Toma et al. 2011). The MB-MSC suspension was then diluted to a concentration of 5 × 103 cells/ml in a sample reservoir, with a solution of PVP and PBS adjusted to include 1 mM of calcium chloride (CaCl2) and 0.5 mM of magnesium chloride (MgCl).
Flow cytometry was performed to quantify MB-MSC association (FACSCalibur, Becton Dickinson), based on the observation that MB attachment to the MSC surfaces increases their side scatter (SSC) due to the multiple reflective surfaces of the MBs present on the cells as described previously (Toma et al. 2011). Before each experiment, the presence of microbubbles on the surface of the MSCs was verified with flow cytometry and microscopy.
Digitally recorded fluorescence images and videos were transferred to ImageJ (Rasband 1997–2011) for further analysis. An algorithm was incorporated into the software to produce maximum intensity persistence of the MB-MSC images. Particle trajectory, radial velocity (ur), and amount of cell adhesion were measured. The radial velocity of the MB-cell complexes, which represents the translational motion conferred by ARF on the MB-cell complexes, was calculated by measuring the radial displacement in a known number of frames and the time elapsed to travel that distance. The number of cells adhering or intermittently adherent to the wall during 10 seconds of ultrasound exposure, as well as the number of cells persistently adherent 5 seconds after the ultrasound was turned off were counted by visually tracking and analyzing the images.
A total of 18 samples were used to measure the radial velocities of the MB-cell complexes. We investigated the effect of the following parameters on the radial velocity and adhesion of the MB-cell complexes: acoustic pressure (100 kPa, 250 kPa, 500 kPa, 750 kPa and 1 MPa), duty cycle (DC) (10%, 20% and 60%), pulse length (5 cycles, 10 cycles and 20 cycles) and corresponding pulse intervals (Table 1). Particle radial velocities derived from the experimental data and model predictions were compared. The axial velocities of the MB-cell complexes (in the direction of the fluid flow) were measured in the absence of ultrasound exposure.
We performed temperature measurements to determine heating effects at higher time-averaged ultrasound intensities. A thermocouple connected to a multimeter (114 True RMS Digital Multimeter, Fluke Corp.) was placed in the immediate vicinity of the ultrasound catheter tip within the phantom vessel under static physiological flow conditions (40 ml/min). The temperature increase from baseline temperature (approximately 22°C) was measured.
The experimental data was expressed as mean ± standard error of the mean. Experimental groups were compared using an analysis of variance on data by the Single Factor ANOVA model. Statistical significance was defined as p-values < 0.05.
The mean diameter of the MBs s was 2.54±0.91 μm and mean electrostatic charge was 51±9 mV. In addition to microscopic examination, flow cytometry verified the association of MBs with the negatively charged MSCs (Figure 2). MB attachment to the cell led to a significant increase in the amount of orthogonal scattered light (side-scatter) in comparison with MSCs alone to in flow cytometry.
Theoretically, the main flow pattern in the phantom is concentric flow. The experimental values for axial velocity were similar to the theoretical values. The maximum axial velocity (10.9±0.7 cm/s) occurred at 0.78 mm from the transducer tip and the minimum velocities were measured near the transducer tip and close to the vessel wall (Figure 3).
The experimentally measured radial velocity of the MB-cell complexes ur for various combinations of ultrasound parameters (n=3 experiments per condition) are shown in Figure 4, which groups the data by the time-averaged ultrasound intensity for a given combination of acoustic pressure and number of cycles (5, 10, or 20) per pulse. The time-averaged ultrasound intensity was the major determinant of the MB-MSC radial velocity, independent of the duty cycle/pulse length configurations (5 cycles, 10 cycles and 20 cycles and corresponding pulse intervals) (p=NS for multiple comparisons within the group with same acoustic pressure and duty cycles). This was further evidenced by comparing the second group (with 102 kPa and 60% DC) and third group (with 250 kPa and 10% DC) of parameters, both producing the same time-averaged ultrasound intensity of 0.21 W/cm2 and resulting in similar MB-MSC radial velocities (1.9±0.2 cm/s and 2.0±0.3 cm/s respectively, p=NS).
In Figure 5, the mean radial velocity of the MB-cell complexes for each of the time averaged ultrasound intensities represented in Figure 4 was plotted against the time-averaged ultrasound intensity. ur was nearly proportional to the time-averaged ultrasound intensity up to 0.83 W/cm2. The maximum ur of 6.6±0.4 cm/s was obtained with a time-averaged ultrasound intensity of 1.65 W/cm2. Note that this was near the velocity limit that the camera system could report at the achievable frame rate with the phantom used.
In the control group, MSCs with no MB association were not pushed in the radial direction.
The total numbers of MB-MSCs adhered on the vessel wall during 10 seconds of ultrasound exposure and 5 seconds after the ultrasound was turned off are shown in Figure 6. Following ultrasound mediated MSC displacement and adhesion to the wall (solid line), some of the adherent complexes were displaced downstream after ultrasound was turned off due to unopposed shear stress (dashed line). The number of persistently adherent MSCs increased with the time-averaged ultrasound intensity (dashed line), up to a time-averaged ultrasound intensity of 1.65 W/cm2 (Fig. 5), corresponding to the highest radial velocity measured. No appreciable increase in adhesion was observed with higher time-averaged ultrasound intensity under the flow conditions tested.
No adhesion was observed on the vessel wall for the control group, in which no MBs were attached to MSCs.
For a given MB-cell complex originating at a specific point within the vessel, its flow trajectory can be plotted as the net vector resulting from the measured radial velocity and theoretical axial flow profile at various ultrasound parameter combinations (Figure 7). For simplicity, only MB-cell complexes entering the field of view at 3 different locations (1.35 mm, 1 mm, and 0.6 mm away from the vessel wall) were plotted. For those complexes entering the ultrasound field near the center of the flow phantom (1.35 mm from the vessel wall or 0.4 mm from the center of the transducer), all reached the vessel wall with 500 kPa and 20% DC (1.65 W/cm2) or higher equivalent time-averaged intensity) but failed to reach the wall with other ultrasound parameter combinations. For those complexes entering the ultrasound field closer to the vessel wall (0.6 mm), all of the MB-cell complexes reached the wall at or above an acoustic pressure of 250 kPa and 10% DC (0.21 W/cm2).
Overall, all the MB-cell complexes entering the US field were expected to be delivered to the vessel wall with a time-averaged intensity of 1.65 W/cm2 or higher for the volume flow rate used. This intensity level depends on the axial flow velocity (governed by the volume flow rate) and the length of the vessel included within the ultrasound field.
The actual flow trajectories of the MB-cell complexes are shown in Figure 8 as images after processing for maximum intensity persistence, for various time-averaged ultrasound intensities.
At the physiological flow conditions (40 ml/min) the temperature near the tip of the ultrasound catheter increased 1.3±0.3°C for the time-averaged ultrasound intensity of 0.83 W/cm2 and 5.5 ±0.8°C for the time-averaged ultrasound intensity of 1.65 W/cm2. For the higher time-averaged ultrasound intensities of 4.13 and 6.60 W/cm2, the temperature increase from the baseline was recorded as 8.7 ±0.9°C and 11.6 ±1.3°C respectively.
Our group was the first to demonstrate the concept that therapeutic vascular cells can be delivered in vivo using ARF (Toma et al. 2011). In our prior work, we attached cationic uMBs to MSCs via electrostatic interaction, and found that ARF caused not only acute attachment of MSCs to injured rabbit aortic segments in vivo, but that the cells persisted and engrafted over time. In the particular scenario of endoluminal cell delivery for arterial repair, this technology appears to be more efficient compared with other techniques using EPCs loaded with supermagnetic iron oxide nanoparticles in arteries of similar caliber, which resulted in about 25 times less adhesion to the vessel wall (Kyrtatos et al. 2009; Pislaru et al. 2006).
In order to optimize this novel approach at directing the translation motion of MB-MSCs using ARF, we performed in vitro testing to optimize the ultrasound delivery parameters. Experimental data demonstrated that the time-averaged ultrasound intensity is the main determinant of the radial translational velocity and adhesion to the vessel wall. This indicates that we could use lower acoustic pressure with higher duty cycles to achieve similar translational motion of the MB-cell complex. Theoretically this pulse sequence could be less destructive to MB.
Experimental data indicated that the number of MSCs adhering to the wall increased with the radial velocity of the MB-cell complexes, as more and more complexes were able to reach the vessel wall during the finite time when the MB-cell complexes were within the finite size of the ultrasound field. Increasing the time-averaged ultrasound intensity further beyond an ultrasound intensity of 1.65 W/cm2 did not increase the extent of adhesion of MB-cell complexes to the wall, for the flow condition used. This is to be expected, however, as all of the MB-cell complexes reached the vessel wall due to the ARF for any time-averaged ultrasound intensity above 1.65 W/cm2. Therefore, a time-averaged intensity higher than this value should not be used, considering heating and other potential ultrasound related concerns such as bubble destruction at higher ultrasound intensities.
The conjugation of the MBs with the MSCs is essential to the success of ARF for vascular stem cell therapy. The ARF on the MSC alone is six orders of magnitude smaller than that on the MBs for identical ultrasound conditions. No significant influence on MSC trajectory nor adhesion was observed when no MBs were attached to the MSC surface.
One limitation of this study was the capability of the optical imaging system. It was not possible to measure the radial velocity higher than 6.8 cm/s as the MB-MSC would have traversed the region of interest in less than one frame. However, we could still quantify the amount of adhesion on the phantom vessel wall for all acoustic parameters considered. Another limitation is possible measurement error due to the cylindrical shape of the phantom vessel wall and difficulty in focusing consistently at the same region of interest.
In spite of these limitations, the in vitro analysis presented here identifies an optimal set of acoustic parameters for the ultrasound and flow system that has a potential for vascular cell-based therapies. Therapeutics based on progenitor cells can be potentially applied to a variety of applications, such as endothelial regeneration following angioplasty, increasing the patency of saphenous vein grafts used in coronary artery by-pass surgery, or the treatment of focal large vessel vasculidities. Optimization of the acoustic based cell delivery is a key element in progressing towards a cell delivery method that does not entail flow cessation, thus making it clinically attractive. While we only presented results with MSC, other cell types can be delivered for various disease models using similar technique developed in our study.
We would like to thank Frances Lutka and Emily Etling for their help with cell preparation and Dr. Marina Kameneva for her help with viscosity measurements. This work was funded by the Center for Ultrasound Molecular Imaging and Therapeutics at the University of Pittsburgh Medical Center. Drs. Kaya, Chen, Wang, Fu, Villanueva, and Toma were funded in part by DPI ODO003819, 5RC1EB010533, or R21HL102531 from the National Institutes of Health.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.