Ultrasound Microbubbles for Molecular Diagnosis, Therapy, and Theranostics

Microbubble Overview

Microbubbles for ultrasound contrast enhancement usually have a diameter of between 1 and 4 μm, thus restricting them to the vascular compartment. Microbubbles have evolved from free gas bubbles in solution over bubbles stabilized by surfactants to bubbles with a shell made from phospholipids, proteins, or polymers. Microbubbles are generally synthesized by mechanical agitation, sonication, or the use of microfluidic devices. A detailed description of the different kinds of microbubbles and their synthesis procedures is beyond the scope of this review (2). Here, we focus on the production of microbubbles for molecular imaging. Tissue and cell specificity can be achieved by using passively or actively targeted microbubbles. Passive targeting refers to the use of the intrinsic properties of the microbubble shell to increase its affinity for a particular cell type or tissue. Albumin-shelled microbubbles, for example, were shown to bind to adherent leukocytes via cell-surface integrins or complement-mediated opsonization solely due to their shell composition (2,4,5).

Active targeting refers to covalent or noncovalent attachment of specific targeting moieties to the microbubble shell, to allow for binding to specific receptors. This attachment can be achieved by a 1-step process, in which the targeting moieties are incorporated during microbubble synthesis. Such a 1-step process is suitable for molecules whose chemical properties are not affected by the harsh conditions encountered during microbubble synthesis. An important example is BR55, the first molecular ultrasound contrast agent to enter clinical trials, in which phospholipids preconjugated to a vascular endothelial growth factor receptor-2 (VEGFR2)–binding peptide are mixed with peptide-free phospholipids to form microbubbles (6).

Alternatively, targeting ligands can also be attached to preformed microbubbles. This is most often achieved by use of the strong noncovalent interaction between strept(avidin) and biotin, which are generally bound to the microbubble surface and to a specific antibody or peptide using either carbodiimide chemistry (7) or stable thioether bonds (8).

Unwanted side effects for targeted microbubbles are expected to be comparable to those for nontargeted ones. These side effects include hypersensitivity and allergic reactions that occurred occasionally in patients. Additionally, myocardial side effects, headaches, and abdominal pain have been reported for some contrast agents, and temporary microembolisms have been discussed. Vancraeynest et al. (9) reported on a reversible myocardial injury induced by the destruction of targeted microbubbles in rats. Damage to blood vessels at the microvascular level is possible after microbubble destruction, when using high microbubble concentrations and long insonation times. Nevertheless, microbubbles are usually administered at low doses, and the incidence of side effects is generally much lower than that for CT and MRI contrast agents.

Preclinical Applications and First Clinical Trial

Molecular ultrasound with targeted microbubbles is predominantly used to monitor molecular markers expressed on the activated endothelium. In this context, the potential of targeted microbubbles for assessing tumor angiogenesis and inflammation has been intensively investigated predominantly in preclinical models, which are summarized in Table 1.

The most prominent molecular markers for the assessment of tumor angiogenesis and therapy effects by ultrasound are VEGFR2 (2) and α_vβ₃ integrin (10). Both markers have been intensely addressed using various preclinical models. In angiogenic ovarian cancer xenografts, higher signal intensity was measured using dual-targeted microbubbles against VEGFR2 and α_vβ₃ integrin than using single-targeted ones (10). Furthermore, microbubbles targeting the VEGF–VEGFR complex, VEGFR2, and endoglin were used to analyze angiogenic activity in pancreatic carcinoma xenografts and to monitor antiangiogenic or cytotoxic responses (11). Similarly, in squamous cell carcinoma xenografts treated with the matrix-metalloproteinase inhibitor prinomastat, significantly lower binding of VEGFR2- and α_vβ₃ integrin-binding microbubbles was observed (12). However, histologic analyses demonstrated that the lower signal intensities recorded in the treated tumors were due to a general decrease in vessel density and not to decreased marker expression on the tumor endothelium (12). These results strongly recommend combining functional and molecular ultrasound to elucidate whether an altered microbubble accumulation during therapy is due to a change in the relative blood volume or to alterations in marker expression.

Furthermore, molecular ultrasound was tested for monitoring activated endothelium during inflammation (e.g., in postischemic injury or atherosclerosis). Here, commonly addressed targets are E-selectin, P-selectin, vascular cell adhesion molecule 1 (VCAM-1), and intercellular adhesion molecule 1. E- and P-selectin–specific microbubbles showed a significantly enhanced retention in postischemic myocardium (13,14), and the binding of VCAM-1–specific microbubbles to plaque-containing inflamed endothelium correlated well with the stage of atherosclerosis (Figs. 1A–1D) (15).

In preclinical research, ligands are predominantly coupled to microbubbles via (strept)avidin-biotin, which might be problematic because of potential immunogenicity issues in humans. The first formulation for clinical application is BR55, that is, (strept)avidin-biotin–free soft shell microbubbles that target mouse and human VEGFR2. BR55 was found to strongly bind to the tumor endothelium of breast (16) and prostate cancer xenografts in rodents (17). BR55 was also successfully applied, with high sensitivity, for monitoring antiangiogenic therapy effects in human colon cancer xenografts (6). In addition, BR55 sensitively depicted the angiogenic status in 2 breast carcinoma xenografts with different degrees of aggressiveness (18). BR55 is currently being evaluated in a phase 0 clinical study for its ability to identify prostate cancers on the basis of their increased VEGFR2 expression (using a visual score in comparison with histopathologic analysis).

Thrombolysis

In the 1970s, Truebestein et al. (19) for the first time demonstrated that high-frequency ultrasound can be used to dissolve thrombi. In 2004, Alexandrov et al. (20) confirmed the potential of ultrasound-mediated thrombolysis, showing in a cohort of over 100 stroke patients pretreated with tissue plasminogen activator (tPA) that the subsequent administration of transcranial Doppler ultrasound improves arterial recanalization. A decade before this landmark achievement, Tachibana et al. (21) had already demonstrated ex vivo that the coadministration of microbubbles can further enhance the thrombolytic activity of ultrasound in combination with pharmaceutical thrombolysis. This effect is largely the result of stable and/or inertial microbubble oscillation and cavitation, leading to clot destruction. Extending this strategy to patients, Molina et al. (22) showed in 2006 that the addition of microbubbles to ultrasound and to initial tPA-based pharmaceutical thrombolysis also works well clinically. In patients with middle cerebral artery (MCA) occlusion, they demonstrated that the triple combination of tPA, ultrasound, and microbubbles was significantly more effective than tPA plus ultrasound and than tPA alone (Figs. 2I–2J). Besides the findings for stroke, promising findings for microbubble-enhanced ultrasound-mediated fibrinolysis have also been reported for declotting thrombosed arteriovenous dialysis grafts in dogs, as well as for improving periinfarction microvascular flow in acute ST-segment elevation myocardial infarction in pigs. Together, these findings convincingly demonstrate that microbubbles hold significant potential for improving the efficacy of clinically relevant combinations of pharmaceutical and ultrasound-based thrombolysis.

Drug Delivery

Besides showing potential for thrombolysis, the application ultrasound plus microbubbles also holds significant potential for improving drug delivery across biologic barriers. Pioneering proof of this principle has been provided by Kinoshita et al. (23), who showed that microbubbles combined with focused ultrasound facilitate the delivery of trastuzumab across the blood–brain barrier (BBB). In this context, focused ultrasound refers to the concentration of acoustic energy on a focal spot a few millimeters in diameter. Depending on the energy used, biologic effects such as heat-induced coagulation can result. However, acoustic pressures induced by microbubble destruction with relatively low ultrasound energies can also temporarily increase blood vessel and/or cell membrane permeability in the insonated region, thereby enhancing local drug delivery. In mice, Kinoshita et al. initially demonstrated that the application of microbubbles together with focused ultrasound enhanced the delivery of the MRI contrast agent gadopentetate dimeglumine across the BBB (Fig. 2K), as well as that of trypan blue (Fig. 2L). Subsequently, they quantified the concentration of trastuzumab delivered across the BBB and observed that although the concentration was below the detection threshold in 8 of 9 cases without the simultaneous application of microbubbles and ultrasound, the concentration steadily increased when microbubbles plus ultrasound were used at pressures of 0.6 MPa and 0.8 MPa (Fig. 2M).

These findings confirm and extend previous findings on the ability of microbubbles to enhance the ultrasound-mediated delivery of therapeutic moieties across biologic barriers, including the BBB, extracranial endothelial layers, and cellular membranes (5). This mechanism, generally referred to as sonoporation, is expected to rely on several different physicochemical and physiological mechanisms, including, for example, on the generation of microjets (which might lead to transient pores in endothelial and cellular membranes), on the production of intracellular reactive oxygen species (which might also contribute to membrane permeability), and on the generation of heat (which might affect phospholipid and membrane fluidity) (24). Besides showing potential for facilitating the delivery of standard drugs to target cells, sonoporation has also been shown to hold significant potential for improving the delivery and the efficacy of genetic interventions based, for example, on plasmid DNA and on small interfering RNA (5,24).