Ultrasound-mediated drug delivery for cardiovascular disease

Abstract

Introduction

Ultrasound (US) has been developed as both a valuable diagnostic tool and a potent promoter of beneficial tissue bioeffects for the treatment of cardiovascular disease. These effects can be mediated by mechanical oscillations of circulating microbubbles, or US contrast agents, which may also encapsulate and shield a therapeutic agent in the bloodstream. Oscillating microbubbles can create stresses directly on nearby tissue or induce fluid effects that effect drug penetration into vascular tissue, lyse thrombi or direct drugs to optimal locations for delivery.

Areas covered

The present review summarizes investigations that have provided evidence for US-mediated drug delivery as a potent method to deliver therapeutics to diseased tissue for cardiovascular treatment. In particular, the focus will be on investigations of specific aspects relating to US-mediated drug delivery, such as delivery vehicles, drug transport routes, biochemical mechanisms and molecular targeting strategies.

Expert opinion

These investigations have spurred continued research into alternative therapeutic applications, such as bioactive gas delivery and new US technologies. Successful implementation of US-mediated drug delivery has the potential to change the way many drugs are administered systemically, resulting in more effective and economical therapeutics, and less-invasive treatments.

1. Introduction

1.1 Drug delivery for cardiovascular disease

The barrier between the bloodstream and vascular tissue presents a significant challenge to extracorporeal drug delivery. Macromolecules can accumulate on the luminal side of this barrier if appropriate ligands exist to facilitate affinity. Once bound, transport into the local tissue bed occurs if the body’s natural mechanisms permit permeability of the endothelium and other vascular barriers [1,2]. Pharmacokinetic modeling, coupled with the development of targeted pharmaceuticals, has improved the specificity and efficacy of therapeutics. However, sensitizing the endothelium to increase local permeability remains an area of active research to improve cardiovascular drug delivery [3].

Methods to direct drug delivery to pathologic cardiovascular tissue are under development. Atherosclerosis, an intensively studied cardiovascular pathology, is characterized by the build-up of lipids, cholesterol and other substances in arterial walls and the formation of plaques [4]. These plaques elicit an immune response [5] and become unstable, shedding fatty deposits and calcified matter into the lumen [6]. The plaque fragments subsequently occlude downstream small-diameter vessels, or become nuclei for thrombi, leading to myocardial infarction, pulmonary thromboembolism or ischemic stroke [7,8].

The treatment of atherosclerosis and other cardiovascular diseases, such as congenital heart disease, rheumatic heart disease, cardiomyopathy and peripheral vascular disease, is limited by the inability to deliver therapeutics across the endothelium safely and effectively [9]. For example, peroxisome proliferator-activated receptor (PPAR) agonists, such as rosiglitazone, have been proposed as a pharmaceutical approach to arrest atheroma by reducing macrophage infiltration into atherosclerotic lesions [10]. However, despite effective in vitro activity, rosiglitazone is cytotoxic to healthy cardiovascular tissue and can induce fluid retention, heart failure and further endothelial dysfunction [11]. More effective targeting and delivery methods for rosiglitazone would simultaneously improve efficacy and specificity. Other promising strategies for treating cardiovascular disease include stem cells for repair of ischemic myometrium and valve damage [12]; macromolecular-assisted, thiomer-mediated approaches [13]; nanoparticle-assisted therapies [14,15] and microRNA-mediated gene expression modulation therapy [16]. This review will focus on strategies and mechanisms of ultrasound (US)-mediated cardiovascular drug delivery. Development of effective targeting and delivery methods for the treatment of cardiovascular disease would lead to a paradigm shift in the development of more effective, more economical and less-invasive treatments for many types of cardiovascular pathologies.

1.2 US technology for therapeutic delivery and monitoring

US has been developed as a diagnostic tool to interrogate the acoustic properties of tissue. More recently, therapeutic US has been developed to elicit beneficial bioeffects. When US interacts with stabilized gas bodies, such as echo contrast agents, a localized breach in the endothelial barrier can be triggered [17]. Specific mechanisms mediating this response are actively under study. In addition, theragnostic agents –containing both gas and drugs – are under development; US activates the gas bubble and causes the release of a therapeutic payload [18]. Interaction of US with the gas can also transduce mechanical energy to local tissue, which elicits a biochemical response [19].

The design of therapy-loaded microbubbles targeting cardiovascular tissue is provided in excellent reviews by Lindner [20], Liu et al. [21] and Laing and McPherson [14]. Laing and McPherson explored the use of liposomes as targeted therapeutic delivery vesicles [14]. Bekeredjian et al. outlined the role of US contrast agent destruction as a particular mechanism for gene and drug delivery [22]. Bull provided descriptions of microbubble designs and the fluid microdynamics induced by oscillating microbubbles [23]. Several relevant reviews on techniques for targeting drug-loaded vesicles to the endothelium have also been published [24–26]. This review focuses on US-mediated drug delivery for the treatment of cardiovascular disease, both updating and expanding the aforementioned reviews and focusing on treatment of cardiovascular diseases. The complex interaction of US with microbubbles, including the gently effervescent activity of stable cavitation and the more violent volume pulsations of inertial cavitation are discussed. This focus allows for a detailed discussion of tissue-specific targeting techniques, drug-loaded vesicle design and both physical and biochemical drug delivery pathways.

Acoustic cavitation is the formation and collapse of gaseous and vapor bubbles in a liquid due to an acoustic pressure field [27–41]. Cavitation is generally classified into two types: stable cavitation, which results in acoustic emissions at sub-harmonics of the fundamental frequency and associated ultra-harmonics, and inertial cavitation, which is characterized by broadband acoustic emissions. When a bubble oscillates non-linearly about its equilibrium radius, the radiated pressure wave may include subharmonic content [41–43]. Broadband emissions are generated when bubbles undergo large radial oscillations and collapse violently [34,37]. This type of bubble motion is dominated by the inertia of the surrounding fluid, hence the label ‘inertial’ cavitation. Stable cavitation can induce bubble-associated microstreaming [44,45] and inertial cavitation can cause microjetting and pitting on solid surfaces [46–48]. Exploiting the acoustic impedance mismatch between microbubbles and blood, acoustic radiation force on microbubbles [49,50] has been used to push drug-loaded agents in the bloodstream toward the endothelium for improved local deposition of the drug [51,52]. Similarly, acoustic streaming – a steady fluid flow caused by an acoustic field – can increase mass transport of a therapeutic [53–55].

Acoustic cavitation has been shown to mediate many therapeutic US applications, including drug and gene release and delivery [56–65], and sonothrombolysis [66–70]. Passive and active cavitation detection techniques have been developed to monitor acoustic cavitation [71,72]. Passive schemes employ a transducer that listens passively (i.e., no transmit) to emissions from acoustically activated microbubbles. More recently, multiple-element arrays have been developed to allow for spatial resolution of bubble activity over a large area [72–76]. Active subharmonic imaging techniques for bubble detection using US arrays have also been implemented [77–79].

2. Barriers to drug delivery

Many factors hinder the penetration of therapeutics from the vascular lumen into an arterial tissue bed, including blood flow, luminal concentration of drug, barrier permeability, pressure-induced convection and intracellular transport rates [80–82]. The specific contributions of each factor within large vessels have been investigated using analytical models specifically developed to describe the transport of tracers into vascular tissue. Penn et al. used one such model to describe the in vivo penetration of horseradish peroxidase into the tunica of hypertensive rats by treating the arterial bed as a system containing four compartments: the vascular lumen, the tunica intima, the tunica media and the tunica adventitia [82]. Though the endothelium was the dominant barrier to penetration, the internal elastic lamina (IEL) was responsible for up to 25% of penetration resistance. Another study by Penn et al. [83] showed that once the endothelial permeability exceeded a threshold, drug accumulated in the intima due to preferential pressure-driven convection. Analytical models of drug delivery are important due to their ability to describe the degree to which individual vascular properties contribute to vascular drug delivery.

The vascular endothelial barrier differs among organ types depending on the function and level of control required for optimal performance. Appreciable differences have been found in structure, signaling control, selectivity and permeability [84]. The vascular endothelial barrier function and integrity can be compromised in many ways, such as bacterial toxins, oxidative stress, ischemia, immune function and dysfunction and vasculopathy as well as cancerous changes [85–89]. Increased understanding of these disease processes not only helps develop theragnostic strategies, but also gives insight into the normal function of the vascular endothelium, which can be manipulated for drug delivery.

Most research effort directed toward drug delivery from the lumen into the tissue bed has been, and continues to be, focused on the blood–brain barrier (BBB). This bias is largely due to relatively recent understanding of the pathophysiology of many neurological diseases and the enormous potential benefit of effective therapeutic delivery. The BBB is a highly selective and tightly regulated endothelial barrier. A complex network of intercellular proteins forms a tight junction. Selective molecular passage is achieved by size exclusion. These proteins are coupled to intracellular structural and signaling proteins within the endothelial cells. Tight junction proteins are in contact with adjacent astrocytic endfeet, conferring a higher level of control and providing avenues for modulation. There is a large body of excellent literature describing this unique barrier [90], the challenges it presents [91], and the approaches to drug delivery currently under investigation [92].

Outside of the brain, transendothelial transport is less stringently regulated, but remains a significant barrier to drug delivery. Within the cardiovascular system, one may encounter endothelial barriers in conductance vessels, capillaries and neovascular tissue (e.g., tumors, ischemic damage) [93,94]. Aside from differences in interaction with and regulation of vascular smooth muscle [95], variations in barrier structure and function can be discerned [96]. Conductance vessels, such as aorta, pulmonary and coronary arteries, typically possess robust endothelial layers that regenerate quickly after trauma. Also, when healthy, these cells are resistant to shear stress and oxidative damage necessary in a high-pressure, high-flow elastic system. Conductance vessel endothelial function is closely coupled to vascular regulation, monitoring oxygen and substrate levels and providing feedback to the smooth muscle via intrinsic relaxation agents (nitric oxide, bradykinin) or contractile agents (adrenergic β₁-receptor agonists) to adjust bulk flow.

Cardiac microvascular endothelium is a unique barrier that also communicates with neuronal cells to synchronize and regulate heart rate and blood flow [97]. Accordingly, the microvascular endothelium is tightly controlled specifically with respect to electrochemical gradients and ion flow to preserve cardiac neurovascular coupling [98]. Disruption of this barrier could affect normal cardiac function and must be borne in mind when manipulating tight junction proteins in the heart.

Vessels formed during tumorigenesis, as well as in neovascularization of ischemic myometrium, tend to be tortuous and thin-walled. The endothelial layer is easily damaged, and not well regulated [89,94]. Thus, the vascularization of tumors is characterized as leaky, a condition often promulgated by tumor cells to foster further pathological angiogenesis serving the tumor [99]. This phenomenon is both a useful avenue for somewhat specific drug delivery to the tumor tissue, and a target for reduction – by drugs or radiation – to starve the tumor [100].

3. Vehicles for enhanced drug delivery

The prospect of controlling drug concentrations within the vascular system by developing responsive physiological systems was outlined in 1986 by Kost and Langer [101]. Since then, US has been established as an effective external trigger for enhanced drug delivery by acoustically agitating vesicles containing encapsulated gas. Several vascular drug delivery vehicles have been proposed to encapsulate therapeutics, such as nanogels [102,103], micelles [104,105] and perfluorocarbon droplet emulsions [106–108]. Drugs and bioactive gases have also been encapsulated in nanometer and micron-sized echogenic liposomes [109–114]. To be an effective vehicle, the agent must maintain stability in vivo while protecting the drug against endogenous agents. Second, the payload should be released at an optimal time and location. Table 1 highlights a few drug delivery vehicles being investigated for cardiovascular therapy, and their associated drug encapsulation characteristics.

Table 1.

Comparison of drug vehicles.

Authors	Vehicle	Drug	Vehicle concentration	Drug concentration during systemic therapy
Kee et al. [232]	Liposome	Papaverine	1.85 mg/ml	3 mg/ml [241]
Buchanan et al. [113]	Liposome	NF-κB	2 mM	10 mM [242]
Rothdiener et al. [234]	Liposome	siRNA	500 mmole/mole lipid	2.7 mM [243]
Fabiilli et al. [106]	PFP droplet	Chlorambucil	3.12 mg/ml	0.02 mg/ml [244]
Fabiilli et al. [107]	PFP droplet	Thrombin	3.1 IU/ml	1.3 IU/ml [245]
Huang et al. [235]	Liposome	Nitric oxide	0.045 mM	1 mM [246]
Britton et al. [227]	Liposome	Xenon	150 μl/ml	3.2 mM [247]
Smith et al. [236]	Liposome	rt-PA	40 μg/mg lipid	2.8 μg/ml [248]
Wang et al. [237]	Alginate	Diclofenac	7.5 mg/ml	2.05 mg/ml [249]
Wang et al. [238]	PFP droplet	Doxorubicin	304 mg/ml	1.8 mg/ml [250]
Jin et al. [239]	Nanogel	Urokinase	500 mg/ml	4 mg/ml [251]

Therapeutic concentrations calculated by assuming average human weight and blood volume.

NF-κB: Nuclear factor-kappaB; PFP: Perfluoropentane; rt-PA: Recombinant tissue-plasminogen activator; siRNA: Small interfering RNA.

Liposomal dispersions effectively encapsulate therapeutics [115] and have been granted the Food and Drug Administration (FDA) approval for clinical use to treat cancer and meningitis [111]. Liposomes either encapsulate the drug within the hydrophilic core or the hydrophobic bilayer [116], depending on the molecular structure of the drug. Amphiphilic macromolecules, such as proteins, intercalate within the lipid bilayer, exposing moieties to the extraliposomal space, while shielding other moieties internally. Gases stabilized by a lipid monolayer can also be encapsulated within these vesicles, rendering them echogenic and more physically responsive to US. Amphiphilic copolymer micelles, which encapsulate drugs within hydrophobic micellar cores or hydrophobic shells, have also been investigated as drug vesicles [117].

Coating the drug delivery agent with biochemically inert substances can effect additional stability against immunogenic or enzymatic degradation that narrow the window of time for drug delivery. Effective drug shielding and increased solubility have been achieved through the use of cyclodextrins [118], which are supramolecular structures capable of forming inclusion complexes with drug molecules. The hydrophilic outer surface of cyclodextrin produces sufficient solubility in aqueous media, yet its interior binding region is hydrophobic, allowing for efficient encapsulation of insoluble organic compounds [111]. Covalent coating with ethylene glycol, or ‘PEGylation’, decreases a drug’s kidney excretion rate, improves solubility and mitigates immunogenicity [119]. Together, these techniques permit the encapsulation of therapeutics that would otherwise be unavailable for use systemically, thus broadening the potential scope of US-mediated drug delivery for treatment of cardiovascular disease.

US-triggered drug release from the vesicle can occur via a variety of mechanisms. Acoustic cavitation is hypothesized to expel the contents of the vesicle within diseased portions of the vasculature, enhancing the therapeutic index and mitigating non-specific cytotoxicity. Gas-encapsulating liposomes are particularly suited for US-triggered release because the cavitation nucleation site is built into the vesicle [114]. Gas-free liposomes have also been shown to release their contents during acoustic cavitation [120]. Alternatively, thermally sensitive liposomes and copolymer micelles release their contents when exposed to hyperthermia in tumor tissue or thermal deposition of US energy [121,122].

4. Tissue targets for US-mediated cardiovascular drug delivery

US can exert direct effects on tissue to enhance cardiovascular drug uptake. These mechanical effects can promote beneficial bioeffects by augmenting the body’s natural mechanisms, as in the case of sonothrombolysis [72], or by creating new delivery pathways, as in the case of sonoporation where US induces pore formation in the cellular membrane [123,124]. US is capable of facilitating delivery of bioactive macromolecules, gases and cell nuclear material to a variety of vascular tissue and cell types, such as thrombus [69,125], endothelial cells [126], smooth muscle cells [112] and cellular nuclei [127]. Table 2 highlights some of these investigations.

Table 2.

Investigations of US-mediated drug delivery to cardiovascular tissue.

Authors	Tissue target	Vehicle	US exposure parameters	Bioeffect
Mehier-Humbert et al. [135]	In vitro; rat mammary	Perflutren liposomes	1.15 MHz, 0.42 MPa, 20% DC	Penetration of dextrans < 464 kDa
Liu et al. [240]	In vitro; bovine RBCs	None	24 kHz	Hemolysis correlated with f1/2 emissions
Skyba et al. [173]	In vivo; rat spinotrapezius	Optison™	Short pulse, 2.3 MHz, Pr = 0.61 MPar	Capillary rupture
Mukherjee et al. [153]	In vivo; rat myocardium	Dextrose-albumin, perfluorocarbon	1 MHz; 0.33 MPar; 15 min	8 × VEGF uptake by endothelium
Ward et al. [136]	In vitro; Jurkat lymphocytes	Optison™	2 MHz, Pp = 0.2 MPa, 10% DC	Lethal sonoporation when bubble near cell membrane
Chen et al. [155]	In vivo; rat myocardium	Lipid-shelled, perfluoropropane	1.3 – 12 MHz; Ml = 1.6; four cycles; ECG-gated	Increased delivery of viral and plasmid transgenes with US optimization
Lawrie et al. [57]	In vitro; porcine medial VSMCs	Optison™	956 kHz; Pr = 1.96 MPar	Free-radical independent luciferase expression
Bekeredjian et al. [22]	In vivo; rat heart	Lipid-shelled, octafluoropropane	1.3 MHz, Ml = 1.5; four cycles; ECG-gated	Increased luciferase expression in heart tissue
Culp et al. [200]	In vivo; porcine rete mirabele	Eptifibitide albumin MBs	1 MHz; 2 W/cm2	Recanalization in 6/8 occluded pigs
Datta et al. [69]	In vitro; porcine blood clot	Definity®	120 kHz, Ppk-pk = 0.32 MPa 80% DC	Penetration of fibrinolytics into clot; clot mass loss enhancement
Phillips et al. [151]	In vitro; porcine coronary	Plasmid MBs	5 MHz; 2 MPa	6.5 × increase in transgene expression with MB + US
Meijering et al. [143]	In vitro; bovine aortic endothelium	Sonovue®	1 MHz, Pr = 0.22 MPar	Endocytotic transport of dextrans with MB + US
Hitchcock et al. [150].	ex vivo; murine aorta	Anti-ICAM – ELIP	1 MHz, CW, 0.49 MPapk-pk	ELIP penetration into tunica media
Laing et al. [112]	In vivo; porcine carotid	Anti-Actin – ELIP	1 MHz, CW, 0.23 MPapk-pk	ELIP penetration into tunica media
Herbst et al. [149]	Ex vivo; porcine aorta	Anti-ICAM – ELIP	1 MHz, CW, 0.15 MPapk-pk	ELIP penetration into tunica intima
Fan et al. [133]	In vitro; rat cardiomyoblast	Definity®	1 MHz, Pr = 0.27 MPar	Dye delivery in cells with adjacent MBs
Phillips et al. [151]	In vivo; rat carotid	Rapamycin MBs	1.2 MHz, CW, 0.14 MPa 5 MHz, 1.5 MPa	10% dose required for neointima reduction with MB + US
Liu et al. [21,210,240]	In vitro; mouse endothelium	AntiCD34-Zhifuxian	Long pulse, 1.1 MHz, 0.07 MPa	Endothelial targeting

CW: Continuous wave; DC: Duty cycle; ELIP: Echogenic immunoliposomes; ICAM: Intracellular adhesion molecule; MB: Microbubble; MI: Mechanical index; Pp: Peak positive acoustic pressure; P_pk-pk: Peak-to-peak acoustic pressure; P_r: Peak rarefactional acoustic pressure; RBC: Red blood cell; US: Ultrasound; VSMC: Vascular smooth muscle cell.

US-mediated drug delivery to these different tissue targets could involve disparate mechanisms and transport routes, which depend on specific biochemical responses to US. For example, acoustically active, drug-loaded vesicles can become tethered to the endothelial membrane via a vascular ligand [128]. Thereupon, vesicles extravasate i) paracellularly, through the breakdown of tight junction proteins [129], ii) intracellularly by endocytosis [130] or iii) through sonoporation [124]. Recent attempts to manipulate the elastic properties of drug-loaded vesicles have rendered their adhesion to activated endothelium in a similar way to immunoreactive leukocytes [128]. Alternatively, drug delivery to intraluminal thrombi is accelerated by US in a similar fashion to that of blood flow around and through a clot. The fluid dynamics at the surface of a thrombus replenishes fibrinolytic proteins within the clot [131], while also removing fibrin degradation products [132]. US-mediated drug delivery seeks to promote these processes for the sake of delivering therapeutics to diseased tissue. These pathways, and the hypothesized mechanisms by which US could facilitate their manifestation are depicted in Figures 1, 2 and 3 and discussed sequentially below.

Figure 1. An overview of three penetration routes stimulated for ultrasound-mediated drug delivery.

Sonoporation refers to the localized, mechanical disruption of a plasma membrane, which allows drugs and ions to diffuse passively. Transcellular pathways, such as endocytosis, involve active transport of drug via cytosolic vesicles. The paracellular route occurs when endothelial cells spread apart, either due to desquamation or by tight junction breakdown from bubble-induced shear stress.

Figure 2. An overview of some proposed mechanisms for transcellular and paracellular ultrasound-mediated drug delivery.

Transcellular: as a result of bubble-induced shear stress along the cell membrane, extracellular drugs can undergo caveolin-1-mediated endocytosis. Additionally, sonoporation can create ‘holes’ in the cell membrane facilitating influx of ions or drugs. Paracellular: shear stress from cavitation-induced microstreaming can cause caveolin-1 to detach from endothelial nitric oxide synthase (eNOS). Here, it converts arginine to nitric oxide, stimulating vasodilation and possible paracellular permeability. Alternatively, this shear stress deforms the actin cytoskeleton, which can cause conformational changes and breakdown of tight junction proteins (ZO-1, occludin).

Figure 3. Propidium iodide (PI; left column) and intracellular [Ca²⁺]_i flux (right column) during microbubble oscillation near a cell membrane.

Cell outlines, indicated by white lines, demarcate PI and Ca²⁺ delivery, when a bubble (circle) oscillates near its cell membrane. The initial state of the cell is seen in the first row, followed by subsequent images during ultrasound exposure.

Reprinted from Fan et al. [133], with permission from Elsevier (2012).

4.1 US-mediated cellular delivery

Currently, a significant body of cardiovascular drug delivery work focuses on US-mediated drug delivery to individual cells. These in vitro investigations explore two important interactions vital to drug delivery: i) drug transfer dynamics from a delivery vesicle to a cell and ii) mechanical interactions between cells and oscillating bubbles. Acoustic pressure amplitude, bubble concentration and the proximity of bubbles to cells [133], all play a role in drug delivery efficiency.

Hamster ovary cells exposed to pulsed US exhibited increased uptake of fluorescein isothiocyanate (FITC)-dextran, a fluorescent macromolecule unable to diffuse across plasma membranes [134]. Furthermore, the cells promoted the expression of luciferase plasmids added to the extracellular serum. At higher acoustic pressures (> 0.3 MPa), cells were unable to repair pores in the local membrane caused by acoustic cavitation, resulting in lysis and loss of viability. Mehier-Humbert et al. further investigated the extent of pore formation in a study examining the dependence of tracer size on sonoporative transport. Particles of increasing size up to 37 nm in diameter were able to penetrate the plasma membrane in the presence of phospholipid microbubbles exposed to US (0.57 MPa; 1.15 MHz) [135].

The degree of sonoporation, whether ‘reparable’ or ‘lethal’, is also dependent on bubble-to-cell separation. In a flagship study on sonoporation, Ward et al. described a non-linear increase in lethal sonoporation as the theoretical bubble-to-cell spacing decreased from 60 μm to below 20 μm for Optison^™ microbubbles [136]. The authors posited that acoustic cavitation was responsible for this bioeffect. Based on the published acoustic parameters and experimental setup, recent cavitation simulations support this assertion, predicting that Optison^™ microbubbles would nucleate both stable and inertial cavitation [41].

4.2 Mechanisms of cellular delivery

Acoustic cavitation has been the prime emphasis of US-mediated drug delivery, especially for delivery of drugs into cells. When a microbubble experiences an acoustic pressure wave near a biological membrane, bubble oscillations provoke desired [124] or deleterious bioeffects [136]. The type of fluid dynamics that result depend on parameters such as the frequency, amplitude and duration of US [137]; the material properties of the microbubble, its shell and the nearby surface [138] and the degree of bubble proximity to that surface [139]. Several studies have investigated microbubble interactions experimentally [66,137] and theoretically [140].

The microjet formation from inertially collapsing microbubbles can create pores mechanically in the plasma membrane [141]. Endocytosis and subsequent transcytosis can be stimulated through this localized mechanical destruction of the endothelial membrane. Depicted in Figure 3, cavitating microbubbles near a cell membrane can create a pore for increased membrane permeability to macromolecules, genes and extracellular ions, which is repaired by cellular mechanisms within seconds [124,133,142]. Membrane permeability to extracellular ions upregulates endocytosis of perimembranous macromolecules due to the calcium flux into the cell [143]. Upon apical endocytosis, vesicular macromolecules can be expelled into the cytoplasm, or trafficked basolaterally for exocytosis into the intima, completing the transcytosis pathway [144].

The types of microbubble oscillations required to observe these effects have also been investigated. Inertially cavitating microbubbles near a membrane form 1 – 200 nm pores due to localized tensions [145]. These pores are capable of resealing bifunctionally by either rapid phospholipid rearrangement or gradual, exocytosis-mediated membrane tension changes, which restores membrane integrity [124]. Microbubble-induced tissue damage correlates with the mechanical index at high acoustic pressures, evidenced by linear correlations with cell viability and capillary rupture above 0.61 MPa_neg [146]. By definition [33], correlation of this bioeffect with the mechanical index further confirms that tissue damage is associated with inertial cavitation. Evidence of these effects at lower acoustic pressures exists as well. Juffermans et al. observed calcium influx into rat cardiomyoblast cells with Sonovue^® (Bracco Suisse SA; Geneva, Switzerland) microbubbles at a mechanical index of 0.1, which is likely not sufficient to cause inertial cavitation. Interestingly however, when intracellular hydrogen peroxide was not sequestered during the experiment, the extent of calcium influx increased, suggesting a biochemical mechanism at low acoustic pressures [147]. As the acoustic pressure amplitude is increased however, peroxide-dependent bioeffects are attenuated, as evidenced by Lawrie et al. (mechanical index 2.0) [57] and Lionetti et al. (mechanical index 1.2) [148].

4.3 Delivery to cardiovascular tissue

Investigations of drug delivery to tissue beds shed light on the extent of drug penetration and the risk of deleterious bio-effects. As illustrated in drug delivery studies to individual cells, a fine line exists between desirable cellular bioeffects and irreversible cell damage. Herbst et al. exposed atheromatous porcine carotid vessels to intercellular adhesion molecule (ICAM)-targeted echogenic liposomes conjugated to vascular stem cells in the presence of continuous-wave US [149]. Stem cell delivery to the arterial tunics increased significantly within the intima when exposed to US at a peak-to-peak pressure amplitude of 0.15 MPa, but was absent beyond the IEL. In ex vivo murine aortas, ICAM-targeted echogenic liposomes extravasated beyond the endothelium in the presence of 1-MHz continuous-wave US but did not extravasate in the absence of US exposure [150]. The spatial distribution of delivery was confined to the intimal layer, with scant penetration beyond the first few elastic laminae. At these modest acoustic pressures, no US-induced tissue damage was observed histologically.

To promote delivery within the tunica media, Laing et al. exposed porcine carotid arteries to α-actin-targeted echogenic immunoliposomes (ELIP) labeled with calcein during 1-MHz US exposure at a peak-to-peak pressure amplitude of 0.23 MPa [112]. Considerable calcein was observed in the tunica intima and adventitia, suggesting a strong contribution from transendothelial and adventitial delivery routes, respectively. Moderate calcein delivery was observed within the intact tunica media, depicting successful localization within smooth muscle cells (Figure 4).

Figure 4. Images of calcein fluorescence indicating penetration of liposomes targeting to smooth muscle actin, in the absence (left) and presence (right) of ultrasound.

Reprinted from Laing et al. [112], with permission from Informa Healthcare (2012).

Within descending porcine coronary arteries after balloon angioplasty, Phillips et al. observed increased transgene expression beyond the endothelium with intravascular US exposure in vivo [151]. The extent of delivery was again localized to the intimal layer, but endothelial desquamation that often accompanies balloon angioplasty was not investigated [152], opening the possibility of direct delivery to intimal cells aided by mechanical ‘pushing’ from acoustic radiation force. Together, these investigations of US-mediated drug delivery consistently demonstrate that small macromolecules and cells can be safely delivered to the artery wall when circulating microbubbles are exposed to moderate peak-to-peak pressure amplitude (0.15 – 0.5 MPa) US.

Investigations of US-mediated drug delivery to the myocardium consistently demonstrate robust delivery of growth factors and genetic material for therapeutic benefit [22,133,151,153–155]. Traditional techniques to revascularize diseased coronary artery tissue, such as percutaneous, transmyocardial administration of vascular endothelial growth factor (VEGF) can be improved up to eightfold by administering the growth factor with 1-MHz color-Doppler US and microbubbles, as demonstrated by Mukherjee et al. [153]. Chen et al. delivered plasmid DNA to the rat myocardium using a combination treatment of 1.3-MHz, intermittent US and plasmid-loaded perfluoropentane (PFP) microbubbles, and observed increased expression relative to untreated liver samples. Additionally, these investigators optimized their US exposures to deliver US bursts gated by the cardiac cycle, and demonstrated considerably higher therapeutic delivery compared with a continuously scanned US regime [154].

4.4 US-mediated mechanisms of vascular tissue delivery

The biochemical and mechanical mechanisms that facilitate drug delivery to vascular tissue are consistent with those that occur during sonoporation. At least three postulated pathways of US-induced passage of materials from the luminal to the adventitial side of vascular tissue have been proposed in the literature: i) paracellular widening of interendothelial clefts and tight junctions, ii) free passage through injured endothelial lining and iii) transcytosis via fenestration and channel formation [129]. These mechanisms have been proposed and modeled [156] for the BBB, but are also supported by studies in other tissue beds, including renal [157], prostate [158] and skeletal muscle [159] tissues.

4.4.1 Paracellular transport

Paracellular transport between functional endothelial cells is supported strongly in the drug delivery literature. Microbubble oscillations in small vessels are known to deform neighboring vessel walls mechanically by direct perturbation or fluid motion [160]. When a microbubble oscillates near a vessel wall, its shell is estimated to travel at speeds on the order of 250 m/s [161], either directly perturbing nearby structures or causing local fluid convection, dubbed as microstreaming (Figure 2) [162]. US exposure of intracranial vessels containing microbubbles showed reversible [163] enhancement of endothelial opening [129], with neither acute nor chronic tissue damage [164–166]. When microbubble oscillation occurs in the cerebral microcirculation, the BBB is disrupted, causing drug extravasation due to breakdown of transmembrane tight junction proteins [167]: claudins, connexins and occludins [168–170]. VEGF is expressed in vascular endothelial cells in response to shear stress, mechanical disruption and response to immune cytokines in the circulation, or adjacent extravascular tissue. Shang et al. [171] quantified three tight junction proteins (claudin-5, occludin and ZO-1) at the mRNA and protein level before, during and after a 3-h exposure to 1-MHz US, in the presence or absence of Optison^™ microbubbles [172]. The three tight junction components were significantly reduced (compared with control) by 1-MHz US at 3 and 6 h post-exposure. Twelve hours after exposure, however, the protein expression and presence at the blood–tumor barrier had returned to normal levels. Temporary opening of the endothelial junctions via VEGF signaling is known as an intercellular route for molecule transport [1].

4.4.2 Damage-induced transport

The most disruptive mechanism is the passage of molecules through a compromised endothelium. A recent ex vivo study [172] demonstrated that aortic smooth muscle exposed to pulsed color-Doppler and Optison^™ microbubbles induced loss of receptor-mediated contractility and endothelium-mediated relaxation. Histologically, loss of endothelium and evidence of endothelial cell apoptosis were apparent. These observations were not seen in tissue treated with US but without Optison^™. Chen et al. performed a series of experiments observing microbubbles oscillating in venules using ultrahigh-speed microscopy [64,160]. When Definity^® microbubbles were exposed to short US pulses (2 μs), small microjets were directed away from the nearest surface, causing considerable invagination of the vessel wall and partial endothelial desquamation. Capillary extravasation of cells [173,174], dyes [175] and other particles [176] due to these large-amplitude bubble oscillations has been observed by a number of investigators, supporting its therapeutic potential for triggering localized, reversible [163] endothelial permeability.

4.4.3 Transcytosis

Microbubble cavitation can also provoke drug transport through a vascular endothelial cell. Transcytosis, in the context of drug delivery, consists of the drug being endocytosed at the luminal surface of the endothelial cell membrane, either in a receptor-mediated fashion [177] or by invagination caused by cavitation with US [178]. Subsequent transportation through several intracellular compartments and secretion from the cell at the basolateral aspect [179,180] results in delivery to vascular tissue. Transcytosis is typically an alternative extension of the caveolar, endocytotic pathway [181]. This is a receptor-mediated process, and provides a pathway for targeted drug delivery via receptor-targeted drug vesicles and US-enhanced endocytosis [182].

4.5 Delivery to thrombus

Sonothrombolysis refers to the expedited dissolution of thrombi due to US exposure. This technique has been tested in clinical trials, but evidence to justify its universal use to treat ischemic stroke lacks consensus. The Combined Lysis of Thrombus in Brain Ischemia Using Transcranial Ultrasound and Systemic tPA (CLOTBUST) trial randomly exposed 126 patients with middle cerebral artery occlusions to intravenous recombinant tissue-plasminogen activator (rt-PA) and 2-MHz transcranial Doppler US and observed significant clinical improvement in 49% of patients, compared with 30% receiving rt-PA alone [183].

Molina et al. extended the clinical investigation of sonothrombolysis by implementing multiple IV microbubble infusions with US exposure [184]. Recanalization rates were significantly improved when Levovist^®, a contrast agent, was included with rt-PA and conventional transcranial Doppler US. However, these improvements were modest (54.5 vs 41% with rt-PA/US). Numerous investigators offer reasons for the absence of recanalization improvement in some patients, including inconsistent application of acoustic energy [185] and thrombus immunity to rt-PA lysis due to compositional factors [186,187]. Other investigators cite differences between recanalization and reperfusion, a metric that correlates better with clinical outcome [188].

4.6 Mechanisms of drug delivery to thrombus

Early sonothrombolysis investigations demonstrated efficacy in the presence of a fibrinolytic drug, such as rt-PA [189]. Rather than a direct mechanical or thermal mechanism, US affected the rate of plasminogen cleavage by rt-PA, hastening enzymatic fibrinolysis [190]. Datta et al. refined this hypothesis by promoting stable cavitation nucleated by an US contrast agent to enhance the penetration of both rt-PA and plasminogen into clots [69]. These results indicate that expedited fibrinolysis occurs due to increased availability of plasminogen binding sites for rt-PA.

5. Drug targeting

Proposed cavitation-based mechanisms of drug delivery rely on the physical forces exerted by oscillating microbubbles. Since the distance over which these forces are exerted – either by direct perturbation from bubble oscillation [123,124] or fluid streaming [191] – is on the order of a bubble diameter, the microbubbles must be in close proximity with their tissue target. In capillary beds, this requirement is generally met due to the small vessel diameters (< 10 μm) relative to micro-bubble size. However, in larger vessels such as the carotid, femoral or peripheral arteries, the relative fraction of microbubbles near the thrombus or endothelium may not be sufficient to provoke a therapeutic response. Thus, many groups have developed targeted delivery vehicles for efficient US-mediated drug delivery. The most common approach is to conjugate the microbubbles with vascular ligands, so that vesicle adherence to the diseased tissue is promoted. Molecularly targeted, acoustically active drug delivery vesicles have been recently investigated for therapeutic application and diagnosis of inflammation, atherosclerosis, angiogenesis, intravascular thrombosis and post-ischemic injury [192]. Several different antibodies conjugated to microbubbles are reported below.

Atherosclerosis and inflammation facilitate the adherence of both anti-ICAM-1 and anti-vascular cell adhesion protein (anti-VCAM-1) conjugated microbubbles to plated cells and vessels [193–195]. The amount of adherence has been found to vary directly with the degree of atherosclerotic plaque formation [193], the number of molecular receptors present or the activation state of endothelial cells [193]. The ability to target effectively depends on the disease state, as demonstrated with early indication of cardiac transplant rejection in rats [196,197], and initial inflammation of atherosclerosis in mice [194].

Another major approach to targeting endothelium is the use of anti-P-selectin conjugated microbubbles. Takalkar et al. showed that once anti-P-selectin conjugated microbubbles were adhered to P-selectin coated plates in a parallel plate flow chamber, shear stresses of 34 dyn/cm² were necessary to dislodge approximately half of the microbubbles [198]. To improve the probability of initial adherence, Rychak et al. [128] pressurized the microbubbles to cause a partial gas loss. This treatment resulted in an excess shell surface area of approximately 30%, allowing the microbubbles to deform from a spherical shape more readily. The deformable microbubbles adhered to the endothelium in the presence of higher sheer stresses than non-deformable microbubbles. Ferrante et al. demonstrated an alternative means of improving adherence by conjugating the microbubbles with both anti-P-selectin and anti-VCAM-1 antibodies [199]. They found that the dual-targeted microbubbles adhered almost twice as efficiently in flow chambers compared with microbubbles targeted with either anti-P-selectin or anti-VCAM-1 alone.

Targeting to improve drug delivery to intraluminal thrombi has also been developed. Platelets, which exist in high concentrations on the surface of thrombi, can be targeted by conjugating a glycoprotein (GP) IIb/IIIa receptor inhibitor to the delivery vesicle [200–204]. Investigations implementing this targeting strategy show enhanced thrombolysis in the presence of a thrombolytic and targeted microbubbles exposed to US. Culp et al. [200] and Xie et al. [201] both demonstrated that GP IIb/IIIa targeted microbubbles exposed to US expedite lysis more than non-targeted microbubbles in vivo. Hua et al. [202] demonstrated similar results in vitro using microbubbles conjugated to a tetrapeptide for targeting activated platelets. Fibrin can be targeted by conjugating an inactive portion of rt-PA to the drug delivery vesicle. Klegerman et al. accomplished targeting to thrombus by intercalating the D-phe-L-pro-L-arg-chloromethyl ketone (PPACK) moiety of rt-PA, which exhibits strong fibrin affinity, within the phospholipid bilayer of ELIP [205].

Anti-α-integrins have been used to target microbubbles to neovascularization in vivo [206,207]. Of the total microbubble infusion however, only a small fraction adhere to the diseased endothelium. This physical limitation will be shared by all targeted US theragnostic agents. Acoustic radiation force exerts enough force to push non-targeted microbubbles within close proximity of endothelium in vivo by causing particle motion along the axis of acoustic propagation [51,208]. When microbubbles were conjugated to targeting antibodies, the concentration of microbubbles along the vascular increased ex vivo [209] and in vivo [210]. Further, Liu et al. showed that acoustic radiation force can significantly increase adhesion in both arteries (using anti-CD34) and the microvasculature (using anti-ICAM-1) [51,210].

6. Conclusions

US-mediated drug delivery is a promising strategy to improve the way drugs are delivered to diseased cardiovascular tissue. Backed with sound experimental evidence, this method relies on the ability of US to focus mechanical energy on well-defined tissues to manipulate barrier properties, or actuate biochemical pathways leading to increased drug penetration into the cell or tissue of interest. Acoustic cavitation facilitates drug delivery through its mechanical effects and by releasing therapeutics from drug-loaded vesicles. Targeting methods, such as radiation force and biomolecular ligands, localize drug-loaded vesicles near the endothelium for efficient delivery. Sonoporation has been demonstrated as a means of cellular therapeutic delivery through US-induced pores. Delivery to the vascular wall and beyond relies on permeabilization of the endothelium, which can be enhanced by paracellular transport, transcytosis or frank breach of the endothelial layer. Drug delivery to thrombus can be augmented by sonothrombolysis, which has been established as an effective method to recanalize cerebral vessels in clinical trials. The mechanisms of US-mediated drug delivery outlined in this review remain debated. However, future conclusions drawn from ongoing investigations will both add to the understanding of US-mediated cardiovascular drug delivery and improve the efficacy of treatments based on these concepts. The ability to provide localized efficacious treatment to vascular beds to reverse endothelial dysfunction, treat cardiomyopathy, or accelerate thrombolysis would markedly change the clinical approach to the treatment of cardiovascular disease and stroke.

7. Expert opinion

7.1 Challenges and gaps in knowledge

In the US-mediated drug delivery literature, there is a paucity of studies demonstrating drug transport from the lumen into medial tissue. In their studies of calcein delivery to porcine carotid tissue, Laing et al. demonstrated strong calcein delivery to smooth muscle tissue, however, the route of transport could not be determined due to the considerable presence of vasa vasorum within porcine carotid medial tissue [112]. Within these studies, drug penetration seems to have occurred through and between the endothelial cells of the vasa vasorum into smooth muscle cells, and not via direct penetration through luminal endothelium. The IEL likely obstructed this route, which, along with endothelium, is a significant barrier to drug transport [82]. The mechanical forces required to manipulate this vascular layer – rich in dense elastin fibers – may be vastly different than those exerted on the endothelium. The IEL is relatively impermeable even to water, except for small, sparsely distributed fenestrated pores on the order of 1 μm in diameter [211]. Future studies should investigate the specific mechanical interactions necessary to manipulate this barrier to increase the efficacy of US-mediated drug delivery.

Nitric oxide, a short-lived but potent vasodilator, likely mediates paracellular permeability [212] via mechanotransduction of shear stress along the endothelium. Caveolin-1, a protein abundant within plasma membrane clefts, is known to bind endothelial nitric oxide synthase (eNOS) under normal shear conditions. Conformational changes in these proteins induced by high shear stress could liberate eNOS, leading to considerable activation of nitric oxide, vasodilation and tight junction opening [213]. Depicted in Figure 2, US-enhanced, caveolin-1-mediated paracellular permeability is achieved due to activation of mechanosensitive caveolins in the vascular endothelium [214,215]. Alternatively, paracellular permeability has been shown to diminish after periods of increased shear stress due to fluid flow in bovine brain microvascular endothelial cells [216]. These recent data on paracellular drug transport emphasize the need for continued research into the mechanisms of US-mediated drug delivery.

Currently, the efficiency of drug encapsulation techniques may limit the extent to which US-mediated drug delivery can elicit beneficial bioeffects in vivo. One promising strategy using a PFP emulsion formulation encapsulated up to 300 μg/ml of doxorubicin drug, with an approximate 50% release profile once exposed to US (Table 1) [217]. This compares with the systemic pharmacodynamic doxorubicin concentration near 17 μg/ml [218]. Targeting approaches will likely increase drug concentration near tissues of interest. Nevertheless, future improvements in encapsulation techniques will augment US-mediated drug delivery substantially.

Variability in the local cellular environment and the temporal dynamics of cavitation make individual responses to drug delivery therapy a challenge. In their investigation of cellular sonoporation in flow fields, Park et al. observed significant reduction in drug delivery in vitro when cells were cultured under shear stress [219]. Throughout the human vasculature, cells are exposed to varying amounts of shear stress depending on pathology, patient age and cardiovascular health. In vivo cavitation activity is also difficult to control, in part due to the use of polydisperse microbubble populations [220], variability of acoustic pressure in situ and vascular characteristics [221,222]. Understanding the manner in which these factors affect the cellular and mechanical mechanisms of US-mediated drug delivery will be imperative to successful treatment of cardiovascular disease.

The exact mechanisms that promote enhanced dissolution of blood clots in the presence of US and microbubbles have yet to be elucidated. While sonothrombolysis investigators have largely discounted thermal mechanisms [189], fibrin degradation product removal, fibrinolytic drug or microbubble penetration due to microstreaming and acoustic radiation force have been proposed as contributing factors (Figure 5). The mechanical and biochemical properties of the occlusive thrombus may also contribute to the degree of US enhancement of lysis [187,223,224]. Further research in this area will continue to refine the acoustic parameters necessary to develop efficient imaging and therapeutic US systems for sonothrombolysis.

Figure 5. Cavitational mechanisms in sonothrombolysis.

As a result of local fluid dynamics around a cavitating microbubbles, the fibrinolytic enzymes (recombinant tissue-plasminogen activator (rt-PA) and plasminogen) penetrate deeper into the fibrin matrix. Additionally, fluid microstreaming removes fibrin degradation products from the surface of the clot, expediting the fibrinolytic process.

7.2 Future directions

Non-invasive strategies for US-triggered local therapeutic gas delivery for cerebral ischemic injury are being developed. Bioactive gases such as xenon and nitric oxide are promising neuroprotective agents [225] with minimal adverse effects due to rapid vascular scavenging by hemoglobin. The results of preliminary studies [226,227] suggest that xenon, once released from a vesicle, provides neuroprotection for ischemic brain tissue quite efficiently.

Given the wide array of interdisciplinary studies that will inform future US-mediated drug delivery research, from vascular biology to computational simulation of bubble dynamics, we expect the development of savvy techniques to facilitate successful clinical translation. For example, determination of optimal drug delivery schemes will spur a new US modality for theragnostic applications. Molecularly targeted, theragnostic agents are a low-cost, functional strategy to image the extent of disease using US with high spatial sensitivity and specificity prior to clinical therapeutic intervention. The ability to confirm disease status and trigger delivery of a therapeutic with the same agent would help ensure that only potentially responsive patients would be treated, which in turn would streamline clinical trials of these agents. Patil et al. describe one such example to track flowing microbubbles accumulating along a vessel border with harmonic imaging [228]. In their study, three consecutive pulses (two inversion pulses, one radiation force pulse) were used to cancel linear tissue and bubble effects while ‘pushing’ microbubbles to the perimeter of the lumen. Their results revealed an optimum frequency range for microbubble radiation force, which can be harnessed for future investigations. Hitchcock et al. [72] and Goertz et al. [229] demonstrated that for a fixed treatment period, the ‘cavitation dose’ and thus beneficial bioeffect can be maximized by inserting quiescent periods in the US exposure. The quiescent periods were dependent on the vascular flow rate.

Real-time methods to monitor drug delivery are also under investigation. Magnetic resonance imaging (MRI) guidance has been critical to the clinical translation of US thermal ablation. It is expected that analogous advances will be made for US-mediated drug delivery. Passive cavitation imaging techniques have been developed to monitor microbubble-enhanced US thermal ablation [76,77,230]. These techniques have been applied to differentiate stable and inertial cavitation from contrast agents in flow [78,231]. Furthermore, the image quality has been shown to be independent of the US insonation parameters. The authors anticipate investigations to enhance and monitor drug delivery efficacy will increase in frequency as investigators continue to reveal the proficiency of US-mediated drug delivery in treating cardiovascular pathologies.

Article highlights.

• Developments in ultrasound (US) technology and theragnostic agents have fostered a new method to administer therapeutics within the cardiovasculature: US-mediated drug delivery.

• Acoustically responsive vesicles capable of encapsulating and shielding targeted therapeutics from vascular degradation are being developed to increase the efficacy and scope of US-mediated drug delivery.

• Through their interaction with tissue, acoustically activated microbubbles can manipulate the barrier properties of vascular tissue by i) mechanically forming pores within the membranes of cells, ii) increasing transcellular or paracellular permeability of the endothelium and iii) targeting encapsulated therapeutics to the endothelial surface via acoustic radiation force.

• Sonothrombolysis is an established method of US-mediated drug delivery in vitro and ex vivo, and is currently under clinical investigation.

• Modalities combining therapeutic and diagnostic US are currently under development, which couple the benefits of US-mediated drug delivery with the diagnostic capabilities of traditional US technology.

This box summarizes key points contained in the article.