1.

Introduction

Photoacoustic (PA) imaging has garnered attention in the last decade because it can make use of optical absorption contrast while providing high ultrasound resolution in deep tissues.^1–3 The principle of PA imaging is that local heat deposition following short laser irradiation pulses generates acoustic waves, and then the propagated waves are detected by conventional ultrasound (US) imaging scanners. PA imaging has been significantly investigated in cancers, brains, hearts, and eyes of small animals.^4–9 Additionally, thanks to the natural fusion of light excitation and US detection, PA imaging systems have been easily merged with existing US imaging systems following minor modifications (e.g., muting of US transmission and collection of raw radio frequency data).^10,11 Because integrated systems have shared acoustic detectors, they present the advantages of conventional US imaging systems, such as portability and real-time imaging capability. At the same time, contrast agents for both imaging modalities have been significantly explored to enhance detection sensitivities and specificities. For example, optically absorbing organic dyes, plasmonic gold nanostructures, and organic nanoparticles have been developed for PA imaging in various biological applications.^12–20 From a clinical point of view, biocompatibility and biodegradability of those nanoparticles for PA imaging have not been meaningfully studied, and thus safety is an ongoing issue as PA moves toward clinical application. So far, clinically approved dyes (i.e., methylene blue and indocyanine green) have the highest chance to be chosen as clinical PA contrast agents.^21,22 Methylene blue is currently being investigated as a PA lymph node tracer in breast cancer.^10,21 For US imaging, microbubbles filled with fluorinated gases are routinely used in clinical practices to map blood flow in hearts, livers, and kidneys.²³ In preclinical studies, microbubbles have been used for US imaging, US-guided drug delivery, and microbubble targeting.^24,25 Dual-functional contrast agents for simultaneous PA and US imaging have recently been reported, such as ink-encapsulated micro- or nanobubbles, encapsulated-gold nanorods human serum albumin shelled microbubbles, and liquid perfluorocarbon nanodroplets with encapsulated plasmonic nanoparticles.^26–28 Recently, we developed microbubbles filled with gases and shelled with porphyrin-lipid conjugates, termed porshe microbubbles.²⁹ Here, we report a straightforward formulation that gives rise to unique properties, composed of methylene blue microbubbles termed “${\mathrm{MB}}^2$,” as a dual-modality PA/US contrast agent formed from the simple combination of two clinically used materials. The PA contrast can be derived from the optical absorption of methylene blue followed by laser irradiation, while the US contrast can stem from the acoustic reflection of the transmitted US by gas-filled microbubbles.

2.

Methods and Materials

2.1.

Synthesis of the Methylene Blue Microbubbles (Referred to as $\mathit{M}{\mathit{B}}^{\mathit{2}}$)

Unless otherwise noted, chemicals were obtained from Sigma, St. Louis, Missouri. The synthesis process of $\mathrm{M}{\mathrm{B}}^2$ is shown in Fig. 1(a). $\mathrm{M}{\mathrm{B}}^2$ is composed of a lipid shell encapsulating octafluoropropane gas (Concorde Specialty Gases, Eatontown, New Jersey). The following lipids were obtained from Avanti Polar lipids: 1,2-dipalmitoyl-sn-glycero-3-phosphate (DPPA; Avanti # 830855); 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC; Avanti # 850355); and 1,2-dipalmitoyl-sn-glycero-3-phosphoethanolamine-N-[methoxy(polyethylene glycol)-5000] (MPEG5000; Avanti # 880200). A stock solution of 10 mL of 20 mM methylene blue was made in phosphate-buffered saline (PBS) solution ($\mathrm{pH}=7$) and diluted to desired concentrations in PBS. DPPA, DPPC, and MPEG5000 stock solutions were made of $20\mathrm{mg}/\mathrm{mL}$ in chloroform and kept at $-20°\mathrm{C}$. Lipid films were made at varying total lipid concentrations while maintaining a molar ratio of $\mathrm{DPPC}:\mathrm{DPPA}:\mathrm{MPEG}5000=10:1:1.2$. Chloroform was used to dissolve the lipids to form a lipid film and solvent was evaporated. The lipids used generated a $1\mathrm{mg}/\mathrm{mL}$ lipid solution unless otherwise noted. The film was a hydrated 750 μL of methylene blue-PBS solution combined with 100 μL of propylene glycol (Bioshop Canada #PRO888.1) and 100 μL of glycerol (Bioshop Canada # GLY001.1) Octafluoropropane gas was added to the vial headspace, the vial was sealed and sonicated to exchange gas within the solution, and then refilled with octafluoropropane. $\mathrm{M}{\mathrm{B}}^2$ was then formed using a vialmix activator (Lantheus Medical Imaging, North Billerica, Massachusetts) for 45 s. As shown in Fig. 1(a), $\mathrm{M}{\mathrm{B}}^2$ synthesis was straightforward and involved hydrating a lipid film with a solution of methylene blue, layering octafluoropropane in the vial, and mechanically agitating the vial to form microbubbles. Photographs of $\mathrm{M}{\mathrm{B}}^2$ and standard microbubbles (hydrated without methylene blue) before and after activation by mechanical agitation are shown in Fig. 1(b).

Fig. 1

Synthesis of methylene blue microbubbles ($\mathrm{M}{\mathrm{B}}^2$). (a) Schematic for generating $\mathrm{M}{\mathrm{B}}^2$. (b) Photographs of control microbubbles and $\mathrm{M}{\mathrm{B}}^2$ before and after activation.

2.3.

Photoacoustic and Ultrasound Imaging of $\mathit{M}{\mathit{B}}^{\mathit{2}}$

We used two types of combined PA and US imaging systems. The first one was operated with a single-element focused transducer with raster scanning, whereas the other one was modified from a clinical US array system. The system schematic and experimental setup for the first type is shown in Fig. 2(a). The details of the system have been previously reported.³¹ Laser pulsing was generated from a tunable laser (Surelite OPO PLUS; Continuum, Santa Clara, California; wavelength tuning range: 650 to 1064 nm) pumped by a Q-switched Nd:YAG laser (SLIII-10; Continuum; 532 nm). The pulse width and repetition rate were 5 ns and 10 Hz, respectively. An optical wavelength of 667 nm (corresponding to the $\mathrm{M}{\mathrm{B}}^2$ peak) was used for PA imaging. Light illuminated the samples via a concave lens, a conical lens, and an optical condenser. A home-made water tray was used for acoustic coupling. Induced PA waves were sensed by a single-element acoustic transducer (V308; Olympus NDT, Center Valley, Pennsylvania; 5 MHz center frequency). The PA signals were transferred to a low-noise amplifier (5072PR, Olympus NDT) and were recorded by a data acquisition system. In the US imaging mode, a low-noise amplifier was used as both an ultrasound pulse transmitter and receiver, and the same transducer was used. To form the volumetric data, we used mechanical raster scanning in two transverse directions along the $x\text{-}y$ directions. Figure 2(b) shows the sample holder, which is positioned under the water tank. The sample holder was 4.5 mm in diameter by 3.2 mm in depth filled with aqueous samples. To investigate the restoration of PA signals, we compared the PA and US signals of the $\mathrm{M}{\mathrm{B}}^2$ solution before and after sonication. Further, to confirm this restoration and investigate the clinical applicability of the mechanism, we employed a clinical PA imaging scanner. Figure 3(a) shows a schematic of the combined PA/US imaging system adapted with a clinical US array system (Samsung Medison, Seoul, Korea). 256-channel simultaneous analog-digital converters and external triggering capabilities enabled real-time PA/US imaging. Conventional US and PA images were obtained sequentially and displayed in the US imaging monitor to show both structural US and functional PA (i.e., optical absorption characteristics) images at the same time up to 10 Hz of a PA frame rate. A linear probe with a 7.5 MHz center frequency was used (Samsung Medison, Seoul, Korea). An optical parametric oscillator laser (Phocus HE, Opotek, California) was employed to provide laser pulses with an optical wavelength of 680 nm, a pulse width of 10 ns, and a pulse repetition rate of 10 Hz. A custom-made bifurcated optical fiber bundle was used to deliver light to the sample. For real-time image reconstruction, one-way (receiving mode only) conventional delay-sum beam forming method was employed. One side surface of a rectangular water container was cut opened, and the opened area was covered by a thin transparent window to prevent the leakage of aqueous solutions and enhance acoustic coupling. One optically transparent plastic vial with a diameter of 7 mm was filled with $\mathrm{M}{\mathrm{B}}^2$ (microbubble concentration $0.1\mathrm{mg}/\mathrm{ml}$; methylene blue concentration, 15 mM), and the other vial was filled with water as a control. Both vials were vertically positioned inside the container, which was filled with water. The PA/US probe was horizontally positioned with its surface directing the center of the vials as shown in Fig. 3(b). Before the microbubbles were destroyed in the solutions, we acquired the control PA image. Then, the US transmission voltage increased to 50 V (the typical voltage, 8 V), US was delivered to the vials for 60 s, and the PA image was again obtained. We repeated this process until the $\mathrm{M}{\mathrm{B}}^2$ was accumulatively exposed to high-voltage US for 10 min.

Fig. 2

Experimental setup of a photoacoustic and ultrasound imaging system with a single-element focused transducer and raster scanning. (a) Schematic of the dual modal system. (b) Photograph and dimension of the sample holder. OPO, optical parametric oscillator; PM, partial mirror; CL, collimating lens; BS, beam splitter; CCL, concave lens; CNL, conical lens; US, ultrasound transducer; PD, photo diode; AMP, amplifier; and DAQ, data acquisition.

Fig. 3

(a) Schematic of a photoacoustic/ultrasound imaging system adapted with a Samsung Medison US research platform. Experimental setup to pop microbubbles in $\mathrm{M}{\mathrm{B}}^2$ to enhance photoacoustic signals. (b) The vials were positioned vertically, and the ultrasound probe scanned them from the side. (c) The vials were positioned horizontally, and the ultrasound probe scanned them from the top.

3.

Results and Discussion

First, we have measured the physical, optical, and mechanical properties of $\mathrm{M}{\mathrm{B}}^2$. Using a highly concentrated solution (15 mM) of methylene blue to hydrate the film had minimal effect on the efficiency of microbubble formation compared to control microbubbles with $\sim 4.5\times {10}^9$ bubbles formed following activation of a $1\mathrm{mg}/\mathrm{mL}$ lipid solution [Fig. 4(a)]. The typical clinical concentration of methylene blue for sentinel lymph node biopsy is $\sim 30\mathrm{mM}$. The size of $\mathrm{M}{\mathrm{B}}^2$ was monodispersed with a peak size of just over 3 μm, which was also nearly identical to control microbubbles formed in the absence of methylene blue [Fig. 4(b)]. Due to the similar size distribution of $\mathrm{M}{\mathrm{B}}^2$ to commercial microbubbles, the US attenuation is dominant at the low frequencies (i.e., $<6\mathrm{MHz}$), which is in agreement with previous attenuation measurements using other lipid-encapsulated agents [Fig. 4(c)].^32,33 The near-infrared absorption generated by $\mathrm{M}{\mathrm{B}}^2$ was intense. Even a 1 in 500 dilution of the $\mathrm{M}{\mathrm{B}}^2$ solution yielded absorption $>1$ with spectral properties characteristic of methylene blue and thus unaffected by the microbubbles [Fig. 4(d)]. As a comparison, the optical spectra of oxy- and deoxy-hemoglobins are provided.

Fig. 4

Physical/optical properties of $\mathrm{M}{\mathrm{B}}^2$. (a) Yields of microbubbles and $\mathrm{M}{\mathrm{B}}^2$. (b) Sizes of control microbubbles and $\mathrm{M}{\mathrm{B}}^2$. (c) Ultrasonic frequency attenuation of $\mathrm{M}{\mathrm{B}}^2$. (d) Optical absorption spectra of $\mathrm{M}{\mathrm{B}}^2$ diluted in phosphate-buffered saline, methylene blue, oxy-hemoglobin, and deoxy-hemoglobin.

To investigate the dual-modal imaging capability of $\mathrm{M}{\mathrm{B}}^2$, we photoacoustically and ultrasonically imaged aqueous solutions of $\mathrm{M}{\mathrm{B}}^2$ by varying the concentration of either microbubbles or methylene blue using a single-element US transducer. As shown in Fig. 5(e), the concentration of microbbubles was varied from 0 to $0.25\mathrm{mg}/\mathrm{mL}$ by $0.05\mathrm{mg}/\mathrm{mL}$ increments, while the concentration of methylene blue was fixed at 15 mM. Figures 5(a) and 5(b) show the PA and US images of six samples, respectively. The quantified PA and US signals at various microbubbles concentrations are plotted in Figs. 5(c) and 5(d), respectively. Interestingly, we found a decrease in the PA signals with increasing microbubble concentration. With more than $0.15\mathrm{mg}/\mathrm{mL}$ lipid microbubble concentrations, PA signals were almost identical to the background PA signals. In contrast, ultrasound signals increased as the microbubble concentration mounted and reached a plateau after $0.15\mathrm{mg}/\mathrm{mL}$ lipid when the US signal became saturated. Typically, the amplitude of initial PA pressure can be expressed as $p_0=\mathrm{\Gamma }{\eta }_{\mathrm{th}}{A}_e$, where $\mathrm{\Gamma }$ is the Grueneisen parameter (dimensionless); $A_e$ is the specific optical absorption (energy deposition, $\mathrm{J}/{\mathrm{m}}^3$); and ${\eta }_{\mathrm{th}}$ is the percentage of $A_e$ that is converted into heat. Since the energy deposition ($A_e$) is equal to the product of the optical absorption coefficient of the target (${\mu }_a$) and the optical fluence ($F$), the PA amplitudes are directly proportional to optical absorption coefficients of the target. In our case, although none of these parameters were modulated, PA signals had interference attenuation. This phenomenon can be explained as follows. As the concentration of the microbubbles increase in the $\mathrm{M}{\mathrm{B}}^2$ solution, the intensity of backscattered US is, certainly, enhanced due to stronger US scattering. However, the initial PA waves are generated inside the medium and start propagating omnidirectionally through the medium. During propagation, the PA waves can be multiply scattered and absorbed (i.e., attenuated) by the surrounding microbubbles. Thus, the measured PA amplitudes decrease with the increase of the microbubbles’ concentration in spite of the constant methylene blue concentration. Therefore, by modulating the concentration of microbubbles in the medium, we can attenuate or restore PA signals, which presents a novel mechanism to modulate PA signals.

Fig. 5

(a) Photoacoustic and (b) ultrasound imaging of aqueous solutions of $\mathrm{M}{\mathrm{B}}^2$ with various concentrations of microbubbles at a fixed methylene blue concentration (15 mM). (c) Quantified, (a) photoacoustic, and (d) ultrasound signals versus microbubbles concentration. (e) Concentrations of microbubbles and methylene blue in six samples.

As shown in Fig. 6(e), the concentration of methylene blue was varied between 0, 1, 5, 10, 15, and 20 mM with the concentration of microbubbles fixed at $0.1\mathrm{mg}/\mathrm{mL}$. Figures 6(a) and 6(b) show the PA and US images of six samples, respectively. The quantified PA and US signals at various methylene blue concentrations are plotted in Figs. 6(c) and 6(d), respectively. As the concentration of methylene blue increased, consequently the PA signals increased due to greater optical absorption in the solutions. However, the US intensities remained constant because of the fixed bubble concentration. In this case, the PA signals are linearly proportional to the optical absorption coefficient, which is in line with conventional principles of PA wave generation.

Fig. 6

(a) Photoacoustic and (b) ultrasound imaging of aqueous solutions of $\mathrm{M}{\mathrm{B}}^2$ with various concentrations of methylene blue at a fixed microbubble concentration ($0.1\mathrm{mg}/\mathrm{mL}$). (c) Quantified, (a) photoacoustic, and (d) ultrasound signals versus methylene blue concentration. (e) Concentrations of methylene blue and microbubbles in six samples.

To further confirm our findings, we studied the switching of PA and US signals using sonication. We prepared $\mathrm{M}{\mathrm{B}}^2$ with $0.1\mathrm{mg}/\mathrm{mL}$ lipid microbubbles and 15 mM of methylene blue. Then we compared the PA and US signals of the $\mathrm{M}{\mathrm{B}}^2$ solution before and after sonication. Figures 7(a) and 7(b) show the PA and US images of the sample before and after sonication, respectively. The quantified signals are plotted in Fig. 7(c). It is clear that the PA signal was initially attenuated by microbubbles. However, it recovered after the bubbles were destroyed by sonication. The PA amplitude increased 2.5 times. Conversely, the US signals were initially strong, but decreased 2.5 times following sonication.

Fig. 7

(a) Photoacoustic and (b) ultrasound imaging of aqueous solutions of $\mathrm{M}{\mathrm{B}}^2$ before and after sonication. (c) Quantified photoacoustic and ultrasound signals before and after sonication.

Moreover, to prove this restoration and explore the practicability of this mechanism, we were able to disrupt the microbubbles in $\mathrm{M}{\mathrm{B}}^2$ and recover the PA signals using a clinically modified PA imaging scanner. As shown in Fig. 8(a), we acquired a control PA image of two vials (e.g., left filled with $\mathrm{M}{\mathrm{B}}^2$ and right filled with water) before the $\mathrm{M}{\mathrm{B}}^2$ was disturbed. The two white dotted circles show the locations of the vials in the medium. The PA probe detected the signals from the top in the image, which was indicated by a yellow dotted arrow [Figs. 3(b) and 8(a)]. The front surface of the left vial (i.e., filled with $\mathrm{M}{\mathrm{B}}^2$) was clearly visible, while the right vial (i.e., filled with water) was photoacoustically invisible. Within 3 min of applying high-voltage US (i.e., 100 Vp-p; corresponding mechanical index, 0.58; the safety limit for diagnostic US, 1.9), we were not capable of recovering PA signals [Fig. 8(b)], but the restoration was significantly enhanced after 10 min [Fig. 8(c)]. Figure 8(d) shows the PA signal enhancement versus high-voltage US application time. The PA signal increased 817 times at 10 min postapplication. Compared with the restoration enhancement obtained using our bench-top system, the improvement using the clinical system was dramatic. We found that the unwanted bulky bubbles in the vial floated up to the top surface over the time period. Thus, when we measured the PA signals from the side [i.e., Fig. 3(b)], our measurements did not have interference from any intact bulky bubbles. However, when we measured the signals from the top (i.e., bench-top experiments), the PA wave propagation was significantly disturbed, which accounts for the only 2.5-fold enhancement observed in our bench-top system. To confirm these findings, we changed the experimental geometry in the clinical system [Fig. 3(c)]. The vials were positioned horizontally, and the US probe scanned them from the top. In this case, the PA signal enhancement was limited to an $\sim 25$-fold increase.

Fig. 8

Photoacoustic images of aqueous solutions of $\mathrm{M}{\mathrm{B}}^2$ before (a) and after 1 min (b) and 10 min (c) of application of high-voltage ultrasound generated by a clinical ultrasound array. (d) Quantified photoacoustic signals versus ultrasound application times. The yellow dotted arrow indicates the position of the ultrasound imaging probe. The white dotted circles indicate the position of two plastic vials filled with $\mathrm{M}{\mathrm{B}}^2$ (left) and water (right). The vials were positioned vertically, and the ultrasound probe scanned them from the side.

4.

Conclusions

We successfully developed and showed the utility of $\mathrm{M}{\mathrm{B}}^2$ as a dual-modality contrast agent for US and activatable PA imaging. PA signals were significantly suppressed by the increase of microbubbles in the $\mathrm{M}{\mathrm{B}}^2$ solution (with fixed methylene blue concentration). When methylene blue concentration increased (with fixed microbubble concentration), no change in US intensity was observed. In addition, high powered US generated by a clinical US imaging scanner burst the microbubbles and drastically (817 times) enhanced PA signals, demonstrating a novel approach that can be used to modulate PA signal generation with spatiotemporal control. Conventionally, one or more parameters in the initial PA amplitude (e.g., Grueneisen coefficient, heat conversion efficiency, optical absorption coefficient, or optical fluence) within an object should be adjusted to control the PA signals. However, none of these parameters were involved in our approach. In terms of a molecular imaging perspective, both PA and US imaging typically suffer from strong background signals. To improve signal-to-noise ratios and clearly delineate the agent distribution, the background signals should be significantly suppressed, which can be easily done by switching between PA and US signals. Therefore, the dual-modality PA/US imaging system with a combination of the dual-modality agent, $\mathrm{M}{\mathrm{B}}^2$, can be applied to image cancers, monitor drug delivery, and visualize internal organs such as bladders and lymph nodes. From a clinical point of view, both methylene blue and microbubbles have been widely used in clinical practices. From an imaging system perspective, both custom-made bench-top and clinically feasible imaging scanners have been utilized in this study. Thus, there are few barriers impeding the clinical translation of $\mathrm{M}{\mathrm{B}}^2$ using clinical PA imaging systems.