Ultrasound molecular imaging of ovarian cancer with CA-125 targeted nanobubble contrast agents.

To formulate Pluronic PEG-NB, the lipids DPPC (1,2-dipalmitoyl- sn -glycero-3-phosphocholine), DPPA (1,2-dipalmitoyl- sn -glycero-3-phosphate), DPPE (1,2-dipalmitoyl- sn -glycero-3-phosphoethanolamine) (Avanti Polar Lipids, Pelham, AL), and DSPE-PEG-COOH (1,2-distearoyl- sn -glycero-3-phosphoethanolamine- N -[carboxy(polyethylene glycol)-2000]) (Nanocs, New York, NY) were dissolved in chloroform in a 4:1:1:1 mass ratio. The solvent was then removed by evaporation, leaving a lipid film. The film was hydrated by adding 1 mL of 0.6 mg/ml Pluronic L10 solution in PBS (Sigma Aldrich, Milwaukee, WI) and 50 μL glycerol. Hydration of the lipid films took place at 75 °C in a water bath for 60 min. The air was removed from the sealed vials and replaced with octafluoropropane (C 3 F 8 ) until the vial pressure equalized. Finally, the vial was shaken on a VialMix shaker (Bristol-Myers Squibb Medical Imaging, Inc., N. Billerica, MA) for 45 s to form the NBs. Bubble samples were stored at 4 °C until use. DiI (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate) (Thermo Fisher Scientific, Waltham, MA) labeled bubbles were prepared by adding 50 μg of DiI in the above mentioned mixed lipid film. Alexa-488 (Thermo Fisher Scientific, Waltham, MA) fluorophore-encapsulated bubbles were prepared using the same procedure, with the addition of 0.3% mol DSPE-Alexa 488 (Thermo Fisher Scientific, Waltham, MA) in the above mixture lipids, dissolved in chloroform in a mass ratio of 4:1:0.7:1:0.3 respectively. The bubble samples were centrifuged (50 g , 5 min) one time to remove excess microbubbles. EDC (1-ethyl-3-dimethylaminopropyl carbodiimide hydrochloride)–NHS ( N -hydroxysuccinamide) chemistry was employed to bind the carboxylic acid (COOH) group of the DSPE-PEG lipid to the amine (NH 2 ) functional groups of the antibody. EDC (Sigma Aldrich, Milwaukee, WI) was added to the NB samples in a 20:1 mole ratio with the carboxylic acid groups, and in a 40:1 ratio with respect to the antibody (625 nmol). NHS (Sigma Aldrich, Milwaukee, WI) was added immediately after in a 5:2 ratio with EDC. After 15 min of rotation at room temperature, CA-125 antibody MUC16 (Abcam, Cambridge, MA) was added. The NB-EDC-NHS-MUC16 mixture was dialyzed in a 14,000 MWCO dialysis tube, rotating in PH 7.4 PBS buffer for 2 h at 4 °C. Within these 2 h, the PBS buffer was changed every 40 min to remove the excess free EDC and NHS. The mean diameter and polydispersity of the unconjugated NBs and NBs conjugated with CA-125 antibody were measured using dynamic light scattering (DLS) (90 Plus, Brookhaven Instruments Corp). Measurements were performed at 25 °C, with a laser wavelength of 660 nm at an angle of 90°. Bubble size was measured by diluting a sample 1:1000 with PBS at pH 7.4 ( n = 3). Bubble size is reported as a number average. A custom tissue mimicking phantom (ATS Laboratories, Bridgeport, CT) was immersed in a 1 L glass beaker containing 700 mL of PBS (Fisher Scientific; Pittsburgh, PA) at 37 °C. NB or AB-NB solution (700 μL) was injected into the PBS solution separately, and the contents of the beaker were continuously stirred at 150 rpm to agitate the bubbles. The change in contrast as a function of time was measured using an Aplio XG SSA-790A clinical ultrasound imaging system (Toshiba Medical Imaging Systems, Otawara-Shi, Japan) equipped with a 12 MHz linear array transducer. System acquisition parameters were set to Contrast Harmonic Imaging (CHI) with 8.0 MHz harmonic frequency, 0.1 mechanical index (MI), 13 Hz imaging frame rate, 65 dB dynamic range, and 80 dB gain. Regions of interest (ROI) of the same size were drawn in the tissue phantom area and in the contrast agent vicinity. The data were exported to Excel (Microsoft, Redmond, WA) and normalized by the contrast of tissue phantom. The signal intensity (logarithmic scale of the normalized data) as a function of time was plotted to obtain the decay rate of NBs. To acquire the rate of NB dissolution over 4 h period, at different time points ( t = 0, 1, 2, 4 h) of post-injection of bubbles, linear raw data images were stored for analysis. The experiment was repeated three times. A secondary antibody (SAB) (Thermo Fisher Scientific, Waltham, MA) tagged with the fluorophore Alexa-488 was added to the CA-125 antibody-targeted non-fluorescent NB samples and dialyzed in the same dialysis tube for 30 min at 4 °C. Centrifuge wash (805 g , 15 min, 4 °C) was performed twice at the end to purify the NBs. Fluorescent microscopy was used to confirm the presence of bright green dots from the Alexa-488 fluorophore. Quantitative analysis of total image fluorescence was carried out in Image J. OVCAR-3 cells and SKOV-3 cells were cultured in complete RPMI-1640 medium (10% fetal bovine serum, 1% penicillin–streptomycin; Invitrogen, Carlsbad, CA) and placed in a humidified atmosphere at 37 °C and 5% CO 2 . At 90% confluence, cells were detached using 0.25% trypsin–EDTA (Invitrogen, Carlsbad, CA) for passaging. To confirm binding specificity, cells were cultured on 35 mm glass bottom dishes at 100,000 cells per dish until they reached 70–80% confluence. Then cells were fixed, stained with Hoechst 33342 (Thermo Fisher Scientific, Waltham, MA) and stored at 4 °C until use. Alexa-488 encapsulated NB or NB-AB was incubated with the fixed OVCAR-3 cells in 5 mL of PBS buffer for 3 h on a rotary shaker at room temperature. The cell-NB solution was washed three separate times with PBS buffer to remove the unbound NBs. Each time, 5 mL of solution was removed from the dish and replaced with 5 mL of PBS buffer. The dishes were then incubated on the rotary shaker for another 5 min before being washed again. Fluorescent microscopy was used to check for the presence of green dots from the Alexa-488 fluorophore localized on the cells and the fluorescence intensity was measured using Image J (National Institutes of Health, USA). Total fluorescence intensity in 5 high powered (×200) fields was calculated. SKOV-3 cells were used as a negative control in the same manner. Data were recorded as fluorescence ratio (total bubbles fluorescence/cells fluorescence per field). All experiments were repeated in triplicate. Animals were handled according to a protocol approved by the Institutional Animal Care and Use Committee and were in accordance with all applicable protocols and guidelines in regard to animal use. BALB/C nude female mice were purchased from Animal Research Center of Case Western Reserve University. Before all procedures, the animals were anesthetized with 3% isoflurane with 1 L/min oxygen. OVCAR-3 cells (3 × 10 6 cells) or SKOV-3 cells (5 × 10 6 cells) were suspended in 100 μL of growth incomplete medium and 100 μL Matrigel (BD Biosciences, San Jose, CA, USA). 200 μL of the tumor cell–Matrigel mixture was injected subcutaneously into the flank of mice. Tumors were allowed to reach a diameter of 0.5 cm before imaging. To compare contrast enhanced ultrasound images with the same tumor in the same mouse, first, 200 μL (4.26 × 10 10 bubbles) non-fluorescent AB-NB was injected through tail vein in tumor-bearing mice that had been inoculated with either OVCAR-3 cells ( n = 5) or SKOV-3 cells ( n = 5). Second, two days after injecting AB-NB, the same amount with the same ultrasound intensity of non-fluorescent NB was injected to the same mice. To acquire ultrasound images, each animal was placed in the left lateral position, and the same ultrasound probe (PLT-1204BT) used in vitro was placed longitudinally to the axis of the animal body. The probe was immobilized using a clamp and was adjusted so that the field of view included kidney and tumor. Contrast harmonic imaging (CHI) was used to image changes in tissue contrast density as a function of time using the following imaging settings: harmonic frequency 8.0 MHz; MI 0.08; dynamic range 65 dB; gain 80 dB; and imaging frame rate 0.1 frame/s. The images were acquired in the raw data format. Fifteen seconds after raw data acquisition started, NB solution was administrated and continuous image acquisition continued for 10 min. The acquired linear raw data images were processed with CHI-Q quantification software (Toshiba Medical Imaging Systems, Otawara-Shi, Japan). ROIs were drawn outlining the kidney and subcutaneous tumor. The mean echopower value in each ROI as a function of time, or time–intensity curve (TIC), was calculated and exported into an Excel-compatible format for statistical analysis. To analyze the decay of ultrasound contrast, the baseline was first subtracted from the TIC. Next, the peak value of the TIC was calculated, and data were normalized to the peak value of kidney to obtain a normalized TIC. The intensity ratios (AB-NB/NB normalized intensity in the same tumor) at 0, 1, 2, and 5 min after peak contrast enhancement were compared with the same tumor between NB and AB-NB in each group. The normalized TICs were averaged over five animals in each group in a time analysis window of 1.5 min starting after the peak value. Finally, data on a logarithmic scale were fit using simple linear regression to derive the decay slope. One week after undergoing ultrasound imaging, OVCAR-3 ( n = 5) and SKOV-3 ( n = 5) tumor-bearing mice were all injected with 200 μL of Alexa-488 encapsulated AB-NB contrast material through the tail vein. Fifteen minutes after bubble injection, tumors were excised and mounted in optimal cutting temperature compound (OCT Sakura Finetek USA, Inc., Torrance, CA) and frozen on dry ice. The tissues were cut into 8 μm slices using Leica CM1850 cryostat (Leica, Germany), stained with DAPI (Thermo Fisher Scientific, Waltham, MA) using standard techniques. The fluorescence images were obtained and analyzed using AxioVision V 4.8.1.0, Cals Zeiss software (Thornwood, NY). All data are presented as mean ± SD (standard deviation) unless otherwise noted. Statistical significance of differences between experimental groups was derived using paired-sample t tests or univariate analysis of variance as appropriate. All statistical analyses were performed using SPSS software (Version 19; SPSS Inc. Chicago, IL, USA). P < 0.05 was considered significant.

DLS measurements showed NB and AB-NB diameters of 66.6 ± 22.0 nm and 74.6 ± 16.7 nm ( n = 5), respectively ( Figure 1 ). In vitro ultrasound stability was analyzed by imaging the nanobubbles at 37 °C under constant agitation and calculating the relative signal loss over time. Figure 2 , A shows the regions of interest (ROI) of the same size in the tissue phantom area and in the contrast agent vicinity. The ratios of mean echo-power values of contrast agents (both NB and AB-NB) and the tissue were plotted over time period of 60 min ( Figure 2 , B ). The initial contrast at time t = 0 was the same with the two types of bubbles. The AB-NB showed the same decay rate compared to the decay rate of NB over 4 h. Two different samples were used: an (Anti-MUC16)-(SAB-Alexa488)-nanobubble (AB-NB-SAB) and an unmodified nanobubble (NB) control. The results qualitatively indicate that the NBs that had both the Anti-MUC16 and SAB-Alexa488 yielded the higher green fluorescent signal ( P < 0.05) ( Figure 3 ). This signal indicates the presence of SAB-Alexa488 on the NBs, which suggests that anti-MUC16 itself was bound to the NBs. AB-NB incubated with OVCAR-3 cells exhibited the greatest fluorescence signal ( P < 0.05) which likely indicates a binding interaction ( Figure 4 ). The untargeted NB showed little to no attachment to OVCAR-3 cells and SKOV-3 cells. AB-NB also adhered to SKOV-3 cells at very low levels. In vivo US distribution of NB and AB-NB was examined using a clinical imaging ultrasound system. In the 10 mice from both groups, under the same ultrasound conditions, AB-NB did not result in any recognizable symptoms of toxicity, and none of the animals exhibited apparent signs of distress after examination. Contrast-enhanced images of the tumors continuously exposed to ultrasound at t = 0 (before bubble injection) t = peak intensity, and t = 2 min after the peak are shown in Figure 5 . The mean TIC is shown in Figure 6 , A – B for the two tumor types and both bubble types. In the OVCAR-3 tumors (CA-125 positive), the average peak echo power of AB-Nbs is over 2-fold higher than for untargeted NB, while in the SKOV-3 tumors (CA-125 negative), there is no difference seen between AB-NB and NB. Figure 6 , D shows the ratio of the normalized echo intensity of AB-NB to that of NB in OVCAR-3 tumors compared to SKOV-3 tumors. Here again at 0, 1, 2, and 5 min after peak, the ratios were significantly greater than those in SKOV-3 tumors ( P 5 times contrast enhancement than NB in transplanted OVCAR-3 tumors. Bubble dissipation included an initial phase and a terminal phase. The terminal phase commonly started 1.5 min after peak enhancement. The mean normalized TICs after peak enhancement and the mean normalized TICs of the initial phase are plotted in Figure 7 . Although the mean decay slopes of AB-NB in OVCAR-3 tumor were lower than that of NB in the initial phase (0.3919 ± 0.0628 compared to 0.9858 ± 0.2020), the difference was not statistically significant ( P = 0.05). There were also no significant differences observed in the decay slope comparing AB-NB with NB in SKOV-3 tumor during the initial phase ( P = 0.05). No significant differences were noted in the terminal phase comparing AB-NB with NB neither in OVCAR-3 nor in SKOV-3 tumor (P > 0.05). The NB fluorescence ratio (total bubble fluorescence/cells fluorescence per field) in OVCAR-3 tumor tissue was significantly higher ( Figure 8 , C , P < 0.05, n = 5) compared to that in SKOV-3 tumor signal.

The goal of this study was to formulate a novel targeted, sub-micron ultrasound contrast agent to detect CA-125 positive ovarian cancer. The resulting bubbles, had a relatively polydisperse distribution, with a mean hydrodynamic diameter below 200 nm, and with a size range reaching to 500 or 600 nm. The bubbles were clearly visible at 12 MHz both in vitro and in vivo . While potentially exciting and substantiated in this study by biodistribution, kinetic and histological analysis, the mechanism behind such strong acoustic response remains unclear. Other reports of echogenic, sub-micron particles have suggested that stabilizing membrane properties, as well as shell events such as buckling in response to the ultrasound field could be a factor, 24 , 25 , 32 – 34 but more in-depth investigation is needed to determine the precise reasons for the observed behavior. To demonstrate functionality of functionalized bubbles, we investigated the specificity and targeting capability of the developed NBs to ovarian cancer cells in vitro . The results showed that the targeted NBs were stable and biologically specific. The binding rate between CA-125 antibody and the NBs was efficient. They were found to aggregate selectively around OVCAR-3 cells but not SKOV-3 cells in an in vitro cell binding experiment, suggesting that the adhesion of NB to CA125-positive ovarian cancer cells is due to antigen–antibody interactions. This confirmed that AB-NB could be effectively targeted. Additionally, there was no adhesion of unconjugated NBs to both cell types, suggesting that the CA-125 monoclonal antibodies that coupled with the surface of the targeted NBs were able to specifically recognize and bind to ovarian cancer cells with high CA-125 expression. Because there was no CA-125 monoclonal antibody such as anti-MUC16 on their surface, unconjugated NBs were not capable of recognizing the CA-125 and thus did not adhere to the membrane of ovarian cancer cells. The lack of a basement membrane and smooth muscle and the expansion of the intercellular space in cancer vasculature result in a maximum pore size of approximately 380-780 nm. 35 This is ample space for NB to pass through the tumor neovasculature to allow extravascular imaging. Based on the above results, we established OVCAR-3 cell tumor and SKOV-3 cell tumor model in nude mice and performed in vivo enhanced-contrast imaging experiments with unconjugated NBs and targeted NBs. The results showed that, the contrast enhancement intensity ratios (AB-NB/NB normalized intensity in the same tumor) at 0, 1, 2, and 5 min after peak were significantly higher in transplanted OVCAR-3 tumors. The peak intensity ratio of AB-NB to NB was 5 times, compared with the same tumor in transplanted OVCAR-3 tumors, which suggested that AB-NB strongly accumulated in ovarian cancer tissue and produced longer-lasting contrast enhancement in vivo than NB. As expected, there were no significant differences between NBs and AB-NBs in transplanted SKOV-3 tumors. A possible reason for this is that the antigen–antibody coupling occurred in the tumor tissue after AB-NB penetrated the vessel wall, resulting in a greater number of NBs adhering to target cells that highly express the corresponding antigen, particularly in transplanted OVCAR-3 tumors. The adherent AB-NB remained visible for a long time, resulting in a higher peak intensity and longer persistence of enhanced contrast compared to NB for CA-125 positive ovarian cancer tumors. This result further suggests that peak intensity and duration of contrast enhancement may be used as indicators for the investigation of targeted NB enhanced imaging. There were no significant differences observed in the decay slope comparing AB-NB with NB in SKOV-3 tumor during the initial phase. Although the mean decay slopes of AB-NB in OVCAR-3 tumor were lower than that of NB in the initial phase, the difference was not statistically significant. The reasons for this are not clear. However, they could be attributed to elevated interstitial fluid pressure which could reduce penetration and cell uptake. 36 Furthermore, CA-125 may create reservoirs of shed antigen in high CA-125 regions around the tumor vessels because of its shedding properties. Upon histological analysis, AB-NB showed more extensive extravasation beyond and inside the CA-125 positive tumor area, while it showed limited extravasation in the CA-125 negative tumor. However, as typically seen with nanoparticle distribution, the localization of NBs within the tumor is not homogeneous. In addition, solid tumors which have heterogeneous vasculature and high interstitial fluid pressure could restrict convective flow and antibody extravasation, except in large necrosis and hypoxia areas. 37 Due to the complex nature of tumor targeting, it is difficult to understand or predict antibody distribution. Some studies have shown that, after systemic delivery of antibodies, distribution in tumors is highly heterogeneous, with regions of no detectable antibody located near regions of high concentration. 38 In general, the accumulation of nanoparticles in tumors may be affected by many factors: size, characteristics of the nanoparticle surface, duration of the nanoparticle and the tumor characteristics, such as degree of tumor angiogenesis, necrotic. 39 In general, clearance and antibody degradation may reduce extravasation. Our study had some limitations. In vivo , both in the ultrasound imaging and in the histological results the localization of nanobubbles within the tumor is not homogeneous. To understand antibody distribution, additional studies using immunohistochemistry will be necessary for analysis. Second, our studies were limited by the relatively small number of animals. In addition, because the tumor microenvironment is often critical in influencing tumor physiology, in future studies, orthotopic tumors in the ovary instead of flank ones described above may lead to acquisition of more directly clinically relevant information. These future experiments will determine the specificity and sensitivity of the targeted NBs in detecting ovarian tumors ovary and establish a timeline of detection corresponding to their staging. We have demonstrated the development of a low-cost, noninvasive approach using targeted NBs to directly image CA-125 positive ovarian tumors. This work demonstrates the proof of concept for ultrasound molecular imaging beyond the vasculature in this model. Antibody-functionalized nanobubbles showed enhanced tumor uptake, retention and extravasation compared to untargeted nanobubbles. Despite their sub-100 nm diameter, these constructs were echogenic at 12 MHz and were clearly visible in tumors over the entire duration of the imaging exam. With additional optimization, targeted NBs have the potential to become a more sensitive detection tool in diagnosis of early stage ovarian cancer.