In vivo acoustic manipulation of microparticles in zebrafish embryos

Abstract

In vivo micromanipulation using ultrasound is an exciting technology with promises for cancer research, brain research, vasculature biology, diseases, and treatment development. In the present work, we demonstrate in vivo manipulation of gas-filled microparticles using zebrafish embryos as a vertebrate model system. Micromanipulation methods often are conducted in vitro, and they do not fully reflect the complex environment associated in vivo. Four piezoelectric actuators were positioned orthogonally to each other around an off-centered fluidic channel that allowed for two-dimensional manipulation of intravenously injected microbubbles. Selective manipulation of microbubbles inside a blood vessel with micrometer precision was achieved without interfering with circulating blood cells. Last, we studied the viability of zebrafish embryos subjected to the acoustic field. This successful high-precision, in vivo acoustic manipulation of intravenously injected microbubbles offers potentially promising therapeutic options.

Fig. 1.. Experimental design and concept of the in vivo acoustic manipulation system.

(A) Schematic of the in vivo acoustofluidic device, which comprises a transparent PDMS chamber for the ZFE and four identical piezo transducers. (B) Micrograph of the in vivo acoustic manipulation chamber. Red dye was infused to highlight the channel. (C) Schematic of an acoustic wave penetrating a ZFE. A standing acoustic wave is established across a ZFE that is fixed in agar gel. When the acoustic wave encounters the PDMS/agar and agar/ZFE interfaces, its energy is partially reflected and partially transmitted (see also table S1). (D) The 2D lattice-like acoustic pattern is produced inside the microchannel by the acoustic system. (E) The 1D patterning demonstrates microbubbles that have resonance frequencies larger than the excitation frequency positioned at the pressure antinodes, while microbubbles having smaller resonance frequency than the excitation frequency are trapped at the nodes. An ultrasound induced 1D patterning of gas-filled microbubbles. The images illustrate the distribution of microbubbles before and after applying ultrasound. In the 1D pattern, microbubbles with resonance frequencies higher than the excitation frequency are shown positioned at the pressure antinodes, while microbubbles with resonance frequencies lower than the excitation frequency are seen trapped at the nodes. (F) Fluorescence image of a tg(kdrl:eGFP) ZFE, expressing green fluorescent protein (GFP) in endothelial cells, highlighting its vasculature network. Red arrows indicate direction of blood flow, e.g., dorsal artery (DA), intersegmental vessels (ISV), and posterior cardinal vein (PCV).

Fig. 2.. Microbubble manipulation inside the DA of ZFEs.

(A) Image sequences demonstrating a gas-filled microbubble migrating under right-to-left flow that became trapped in the DA when ultrasound was activated (at ~2.3 s). The microbubble traveled against flow to reach the antinode, indicated by the dotted red circle. (B) A microbubble traveling under left-to-right flow in the DA of another fish similarly became trapped. Note that the microbubbles are traveling much faster upstream than downstream, which can be attributed to acoustic radiation force. The ZFE was observed to drift vertically 7.6 μm (smaller than the wavelength of 365 μm). (C) The plots illustrating the location and speed (inset) of a microbubble tracked for the image sequence in (A). A magenta dotted line indicates the location of the pressure node. (D) Plots of the x position and the x velocity (inset) of the microbubble in (B). The pressure nodal position is indicated by magenta dotted lines.

Fig. 3.. 2D control of a microbubble using frequency modulation control.

(A) A schematic demonstrates the zebrafish chamber positioned 2 mm off-center with respect to the orthogonally positioned transducer pairs enabling a wider manipulation range. (B) A microbubble is made to move in the x direction against the flow as the transducers placed in line with the fish have their frequencies changed from f_x1 = 4.0 MHz to f_x2 = 4.25 MHz, 15 V_PP, while f_y = 4.1 MHz is kept constant at 12.5 V_PP. (C) Equivalently, altering the excitation frequency of the piezo transducers alongside the ZFE from f_y1 = 4.0 MHz to f_y2 = 4.25 MHz at 12.5 V_PP and back to f_y3 = 4.1 MHz at 12.5 V_PP thus results in movement of the microbubble in the y direction, i.e., perpendicular to the blood flow. In the process, the particle remains stable in the x direction as f_x = 4.1 MHz is kept constant.

Fig. 4.. Acoustic manipulation in vivo at different locations within the zebrafish vasculature.

To demonstrate full directional control, a bubble was directed along the vasculature and back to its initial position multiple times by shifting the frequency. (A) The microbubble was reversibly moved horizontally 22 μm left and right in the 21-μm-wide DA by varying the frequency controlling the horizontal position f_x between 4.1 and 4.15 MHz at a constant amplitude of 10 V_PP. (B) The bubble likewise traveled and returned 29.85 μm in the y direction without stiction in the ISV, which has a diameter close to that of contrast agents and passing RBCs by varying the frequency controlling the vertical position f_y between 4.15 and 4.25 MHz at a constant amplitude of 12.5 V_PP. (C) Last, the same reversible behavior was demonstrated with a traveled distance of 9 μm in a cerebral blood vessel, close to the otoliths by varying the frequency controlling the vertical position f_y between 4.1 and 4.25 MHz at a constant amplitude of 10 V_PP.

Fig. 5.. Viability of ZFE in the presence of ultrasound.

To demonstrate the viability of the acoustic actuation on cardiac function, untreated ZFEs were placed into the acoustic channel and exposed to acoustic excitation across a range of voltages in pulses of 30 s with a subsequent recovery time of 1 min. Up to a voltage of 17.5 V_PP, fibrillation was reversible and did not impair ZFE viability. The error bar represents the SD for a minimum of five data points.