Direct In Vivo Visualization of Intravascular Destruction of Microbubbles by Ultrasound and its Local Effects on Tissue
Background—Our aim was to observe ultrasound-induced intravascular microbubble destruction in vivo and to characterize any resultant bioeffects.
Methods and Results—Intravital microscopy was used to visualize the spinotrapezius muscle in 15 rats during ultrasound delivery. Microbubble destruction during ultrasound exposure caused rupture of ≤7-μm microvessels (mostly capillaries) and the production of nonviable cells in adjacent tissue. The number of microvessels ruptured and cells damaged correlated linearly (P<0.001) with the amount of ultrasound energy delivered.
Conclusions—Microbubbles can be destroyed by ultrasound, resulting in a bioeffect that could be used for local drug delivery, angiogenesis, and vascular remodeling, or for tumor destruction.
Microbubbles used as ultrasound contrast agents can be destroyed by ultrasound. Although this phenomenon has been described in vitro,1 2 it has not been directly visualized in vivo. The purpose of this study was to demonstrate intravascular microbubble destruction in vivo and to study any potential bioeffects of this phenomenon.
The study was approved by the Animal Research Committee at the University of Virginia and conformed to the American Heart Association Guidelines for Use of Animals in Research. Fifteen female Sprague-Dawley rats (Hilltop) were anesthetized by an injection (0.6 mL · kg−1 body wt IM) of a 1% α-chloralose and 13.3% urethane solution (Sigma Chemical Co). The left femoral vein was cannulated to allow microbubble infusion. The right spinotrapezius muscle was exteriorized3 and positioned in a custom-built chamber filled with Ringer’s solution (pH 7.4) presaturated with a mixture of 5% CO2 and 95% N2 and kept at a constant temperature of 37°C. To minimize the effects of differences in tissue flow between animals, maximal arteriolar vasodilation was achieved by addition of 10−4 mol · L−1 adenosine (Sigma) to the Ringer’s solution. Propidium iodide (PI) (Sigma), which exhibits red fluorescence when bound to DNA,4 was also added to the Ringer’s solution to visualize nonviable cells (final concentration, 2×10−6 mol · L−1).
The muscle was visualized by means of a ×20 water immersion objective attached to a microscope (ACM, Zeiss). Data were recorded on 1.25-cm videotape by means of a video recorder (model AG-1730, Matsushita) connected to a video camera (model CCD-72, Dage-MTI), which was mounted on the microscope. An image intensifier (GenIIsys, Dage-MTI) was used to improve the quality of the images, which were displayed on a high-resolution monochrome monitor (model PVM-137, Sony).
To facilitate their identification, microbubbles were labeled with 5-([4,6-dichlorotriazin-2-yl]amino)fluorescein hydrochloride (DTAF) (Sigma).3 In the first 2 rats, we used several types of microbubbles: Optison (Molecular Biosystems), Imagent (Alliance Pharmaceuticals), DMP-115 (ImaRx Pharmaceutical), and BR1 (Bracco Imaging). Optison was selected for the remaining 13 rats because it provided the best DTAF labeling. It consists of microbubbles containing a mixture of perfluoropropane and air,1 with a mean diameter of 3.7 μm and a mean concentration of 0.8×109 · mL−1.
A phased-array system (HDI 3000cv, Advanced Technologies Laboratories) was used to deliver ultrasound and to image the exteriorized rat spinotrapezius muscle. Harmonic imaging was performed with a mean transmit frequency of 2.3 MHz and a mean receive frequency of 4.6 MHz. The tip of the ultrasound transducer was lowered into the chamber containing the muscle, the distal edge of which was positioned at the single focal point (5.1 cm) of the transducer. The thinness of the muscle (≈0.25 mm) allowed it to be contained within the elevation of the ultrasound beam (≈5.0 mm).
Before microbubble infusion was started, 5 ultrasound frames were acquired as precontrast baseline images at a specified mechanical index (MI). Each frame consisted of 128 lines delivered over a period of 12.8 ms, forming a 90° sector. Each line was fired as a single burst of ultrasound with 4 cycles over 0.1 ms. The muscle was scanned to ensure that no microvessel ruptures were present, thereby demonstrating a control state. DTAF-labeled microbubbles (0.24 mL) were then infused into the femoral vein over 1 minute, followed by a single ultrasound frame applied at a specified MI. The different MIs used and their corresponding acoustic intensities and peak negative acoustic pressures are depicted in Table 1⇓. The MIs are the values displayed on the ultrasound system. The MIs actually delivered to the tissue were not measured. Each animal was subjected to 2 MIs varying from 0.4 to 1.0.
After 10 minutes was allowed for PI binding to nonviable cells, the muscle was optically scanned. Each field of view was examined under transillumination for microvessel rupture sites, which were identified by localized bleeding into the adjacent interstitium. These fields were then reexamined under epi-illumination with a dual red-green fluorescence filter. The total numbers of PI-positive nuclei (red fluorescence) and microbubble fragments (green fluorescence) were determined. To calculate the total numbers of microvessel ruptures and PI-positive nuclei at the higher MI, values from the previous lower MI were subtracted from the new total. After termination of each experiment, the wet weight of the muscle was determined.
Regions of interest encompassing ≈80% of the muscle (>2000 pixels) were placed over the digitally stored ultrasound images, and mean video intensity (VI) in these regions was measured in the precontrast and contrast-enhanced images obtained at various MIs.3 VI from averaged contrast-enhanced images was subtracted from that of the corresponding averaged precontrast images.
Comparisons between different MIs were made by 1-way repeated-measures ANOVA, and differences were considered significant at P<0.05 (2-sided). Correlations between MI and other measured variables were performed by use of the function f(x)=a×log10(x)+b.
Before infusion of microbubbles, no microvessel ruptures were seen after ultrasound exposure. Similarly, no ruptures were seen when microbubbles were infused in the absence of ultrasound. When microbubble infusion and ultrasound exposure were performed simultaneously, destruction of microbubbles was seen in microvessels ≤7 μm in diameter (mostly capillaries). Often, bubbles could be pulsed by ultrasound while in flux through the field of view, where their immediate destruction could be directly observed. Because of their rapid flux, it was practically impossible to capture an image of a bubble before its destruction. In fields of view not under direct observation, bubble destruction was evidenced by the presence of microvessel rupture and nonviable cells.
Figure 1A⇓ depicts a normal region of the spinotrapezius muscle under transillumination after microbubble infusion but before ultrasound exposure, where the microvessels are normal. Figure 1B⇓ illustrates a composite image created from transilluminated and epi-illuminated images after microbubble destruction by ultrasound. A capillary rupture site with extravasation of red blood cells (RBCs) into the interstitial space is noted. Microbubble fragments (green) are evident at the center of the rupture area (black arrow). Two PI-positive nuclei (red), indicating nonviable cells, are also seen (white arrows). The vascular damage was localized to short (5- to 10-μm) segments of the vessel length, usually to one side of the vessel wall. Because RBCs and microbubbles are anuclear, they are not labeled with PI. Therefore, their fragments are visible only in the green spectrum.
The number of both capillary rupture sites and PI-stained nonviable cells increased with an increase in the MI, and a close correlation was noted between the 2 (Figure 2⇓). Table 2⇓ shows the mean background-subtracted VI derived from the spinotrapezius muscle in all rats. A linear correlation was noted between MI and VI (y=51 log10(x)+27, P<0.0001, r=0.81, SEE=5.6).
This is the first report of direct visualization of ultrasound-induced intravascular microbubble destruction in vivo. When microbubbles are destroyed by ultrasound, they cause immediate rupture of the microvessel in which they are located, with RBC extravasation into the nearby interstitial space. Nonviable cells are also noted at the microvessel rupture sites. The number of microbubble destruction events and the magnitude of bioeffects is proportional to the MI applied.
The process by which ultrasound destroys shell-free microbubbles has been modeled previously.5 6 7 Microbubbles, being compressible, alternately contract and expand in a ultrasound field. At low acoustic pressure, this expansion and contraction are equal. At higher acoustic pressure, however, the expansion and contraction of the bubbles become unequal and also greatly exaggerated, leading to their destruction. Direct in vitro optical observations suggest that microbubbles with shells may have a similar fate on exposure to ultrasound.2 It is likely that oscillations produce defects within the shells, causing the gas to escape as soon as 5 ms after ultrasound exposure. The resultant microbubble fragments and escaped gas may further disintegrate and may not be detectable on imaging. The slow video frame rate precluded the in vivo confirmation of these events in our study.
Because liquid media have air entrapped in them, ultrasound exposure can result in the production of microbubbles, which can cavitate at high energy.5 6 7 Thus, the introduction of preformed bubbles is not even essential for the occurrence of ultrasound bioeffects. In vitro experiments have demonstrated the ability of ultrasound to cause cell lysis through the process of cavitation.8 In vivo experiments on organs that contain air, such as the lungs, have shown that ultrasound can cause tissue hemorrhage.9 Introduction of microbubbles has also been shown to result in hemolysis of blood within the cardiac chambers.10 Ours is the first study to report direct in vivo observation of bioeffects when tissue is exposed to ultrasound in the presence of preformed microbubbles. Ultrasound exposure before microbubble infusion did not result in microvessel rupture.
Several factors have to be considered before our findings can be extrapolated to the clinical setting. On the basis of our calculations, the maximal number of microvessels ruptured on each ultrasound exposure (formation of a single image frame) was 0.015% of all the ≤7-μm microvessels present in the rat spinotrapezius muscle.11 In our experiments, the transducer and tissue were separated by Ringer’s solution, so ultrasound attenuation was practically negligible. In the clinical setting, however, tissue attenuation is responsible for a decrease of 0.3 dB · cm−1 · MHz−1 in the amount of ultrasound energy that reaches the focal point of the transducer. The MI displayed on the ultrasound system is adjusted for this attenuation. Because the majority of the tissue in an image is not at the focal point of the transducer, the amount of ultrasound energy transmitted to the tissue decreases even further outside the focal region. We placed the tissue in a chamber from whose base ultrasound could be reflected, resulting in higher ultrasound energy delivered than indicated on the ultrasound system. Because we did not actually measure the MIs delivered to the muscle, the values at the site of microvascular injury are unknown.
The concentration of microbubbles in tissue will also determine the magnitude of bioeffects. We infused 0.24 mL of the contrast agent in a 0.2-kg rat over a period of 1 minute. In clinical practice, we would give the same dose to a 70-kg adult, which would significantly reduce the concentration of microbubbles in the blood pool and could result in a several hundred–fold reduction in microbubble destruction. The type of microbubble could also influence its destruction by ultrasound. Although our quantitative data are derived from Optison, the qualitative data were similar for the other agents tested. Agents with thick shells or very-high-molecular-weight gases may not be destroyed as easily by ultrasound.
The duration of ultrasound exposure could also determine the bioeffects. The bioeffects noted in our study occurred from 1 ultrasound sweep encompassing a single frame. If imaging were performed continuously (≥30 Hz, which is the general protocol in cardiology), the bioeffects could be greater for the same MI because more microbubbles would be destroyed during each additional sweep. Intermittent imaging, whereby ultrasound is transmitted periodically rather than continuously, reduces the duration of ultrasound exposure.1
Finally, the interspecies differences in terms of tissue vulnerability are very important. For instance, although lung hemorrhage has been observed with ultrasound in rodents,9 no such effect has been observed in humans exposed to similar ultrasound energies.12 Studies performed so far in thousands of patients with ultrasound contrast agents used in our study have failed to document any clinically detectable adverse effects.
A novel use of microvessel rupture by ultrasound could be in local drug delivery. Microbubble destruction could result in both microvessel rupture and release of the drug, which could enter the interstitial space through the rupture site. In this manner, effective drug concentrations could be achieved locally while its accumulation elsewhere in the body was limited, thus decreasing any side effects. This method may also offer a particularly efficient means of delivering genetic material directly into tissue, allowing its successful incorporation into cells.
There are other potentially beneficial uses of this bioeffect. Microvessel rupture could initiate angiogenesis and vascular remodeling. In addition, destruction of microvessels supplying a tumor could also result in its regression. Finally, destroying microbubbles containing thrombogenic material in tumors could result in vascular thrombosis and “choking” of tumors, with subsequent cell death and tumor eradication. Such applications will require specially designed transducers and will need to be tested in future studies.
This study was supported in part by grants from the National Institutes of Health, Bethesda, Md (R01-HL-48890 and R01-HL-52309) and from Molecular Biosystems, Inc, San Diego, Calif, and an equipment grant from Advanced Technologies Laboratories, Bothell, Wash. Dr Skyba is the recipient of a postdoctoral fellowship grant from the National Institutes of Health (F32-HL-09540), and Dr Price is supported by a Scientist Development Grant (9730025N) from the American Heart Association, Dallas, Tex. Dr Linka was supported by the CIBA-Geigy Jubiläums-Stiftung, Basel, Switzerland, and the Theodor und Ida Herzog-Egli Stiftung, Zurich, Switzerland. We would like to thank Jeff Powers, PhD, and Michalakis Averkiou, PhD, of Advanced Technologies Laboratories for the technical information they provided and for their valuable comments, and Katherine Ferrara, PhD, for a helpful critique of the manuscript.
Presented at the Young Investigator Award Competition of the 7th American Society of Echocardiography Annual Scientific Meeting, June 10, 1998, San Francisco, Calif.
- Received December 2, 1997.
- Revision received February 23, 1998.
- Accepted February 25, 1998.
- Copyright © 1998 by American Heart Association
Dayton P, Morgan K, Allietta M, Klibanov A, Brandenberg G, Ferrara K. Simultaneous optical and acoustical observations of contrast agents. IEEE Ultrasonics Symposium. 1997;1583–1591.
Skyba DM, Camarano G, Goodman NC, Price RJ, Skalak TC, Kaul S. Hemodynamic characteristics, myocardial kinetics and microvascular rheology of FS-069, a second-generation echocardiographic contrast agent capable of producing myocardial opacification from a venous injection. J Am Coll Cardiol. 1996;28:1292–1300.
Harris AG, Skalak TC. Effects of leukocyte plugging in skeletal muscle ischemia-reperfusion injury. Am J Physiol. 1996;271:H2653–H2660.
Apfel RE. Acoustic cavitation: a possible consequence of biomedical uses of ultrasound. Br J Cancer. 1982;45(suppl):140–146.
Flynn H. Physics of acoustic cavitation in liquids. In: Mason WP, ed. Physical Acoustics. New York: Academic Press; 1964:57.
Neppiras EA, Noltingk BE. Cavitation produced by ultrasonics: theoretical conditions for the onset of cavitation. Proc Phys Soc. 1951;B64:1032.
Dalecki D, Child SZ, Raeman CH, Cox C, Penney DP, Carstensen EL. Age-dependence of ultrasonically-induced lung hemorrhage in mice. Ultrasound Med Biol. 1996;23:917–925.
Hsiung HH, Skalak TC. Three-dimensional reconstruction of capillary networks in skeletal muscle. FASEB J. 1990;4:A1256. Abstract.
Meltzer RS, Adsumelli R, Risher WH, Hicks GL, Stern DH, Shah PM, Wojtczak JA, Lustik SJ, Gayeski TE, Shapiro JR, Carstensen EL. Lack of lung hemorrhage in humans after intraoperative transesophageal echocardiography with ultrasound exposure conditions similar to those causing lung hemorrhage in humans. J Am Soc Echocardiogr. 1998;11:57–60.