118 (2007) 285–293www.elsevier.com/locate/jconrel

Journal of Controlled Release

Ultrasonically targeted delivery into endothelial and smooth musclecells in ex vivo arteries

Daniel M. Hallow, Anuj D. Mahajan, Mark R. Prausnitz ⁎

School of Chemical and Biomolecular Engineering, Georgia Institute of Technology, 311 Ferst Drive, Atlanta, GA 30332-0100, USA

Received 4 October 2006; accepted 28 December 2006Available online 12 January 2007

Abstract

This study tested the hypothesis that ultrasound can target intracellular uptake of drugs into vascular endothelial cells (ECs) at low tointermediate energy and into smooth muscle cells (SMCs) at high energy. Ultrasound-enhanced delivery has been shown to enhance and targetintracellular drug and gene delivery in the vasculature to treat cardiovascular disease, but quantitative studies of the delivery process are lacking.Viable ex vivo porcine carotid arteries were placed in a solution containing a model drug, TO-PRO®-1, and Optison® microbubbles. Arteries wereexposed to ultrasound at 1.1 MHz and acoustic energies of 5.0, 66, or 630 J/cm2. Using confocal microscopy and fluorescent labeling of cells, theartery endothelium and media were imaged to determine the localization and to quantify intracellular uptake and cell death. At low to intermediateultrasound energy, ultrasound was shown to target intracellular delivery into viable cells that represented 9–24% of exposed ECs. Theseconditions also typically caused 7–25% EC death. At high energy, intracellular delivery was targeted to SMCs, which was associated withdenuding or death of proximal ECs. This work represents the first known in-depth study to evaluate intracellular uptake into cells in tissue. Weconclude that significant intracellular uptake of molecules can be targeted into ECs and SMCs by ultrasound-enhanced delivery suggestingpossible applications for treatment of cardiovascular diseases and dysfunctions.© 2007 Elsevier B.V. All rights reserved.

Keywords: Ultrasound; Cavitation; Intracellular drug delivery; Endothelial cells; Targeted drug delivery

1. Introduction

Cardiovascular disease is one of the leading causes of deathworldwide, resulting in nearly 17 million deaths annually [1].Targeted drug and gene delivery to vascular endothelial cells(ECs) and smooth muscle cells (SMCs) has increasinglybecome a focus for treating cardiovascular diseases anddysfunctions, such as coronary artery disease, hypercholester-olemia, hypertension, and restenosis, because of the pivotal roleof these cells in controlling and maintaining vascular functions[2–4]. In addition, targeted drug and gene delivery to theendothelium is being used to treat cancerous tumors by anti-vascular therapy [5] and myocardial and peripheral ischemia bypromoting angiogenesis [6]. Many drug and gene deliverysystems are being developed to increase targeting to vascular

⁎ Corresponding author. Tel.: +1 404 894 5135; fax: +1 404 894 2291.E-mail address: prausnitz@gatech.edu (M.R. Prausnitz).

0168-3659/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.jconrel.2006.12.029

cells, such as drug-eluting stents [7], catheter-based systems [8],viral vectors [9,10], and targeted liposomes [11] and micro-bubbles [12]. However, most current techniques lack either theeffectiveness or specificity to adequately treat these disorderswhile safely administering the therapeutic agent and avoidingtoxic systemic effects or require significantly invasive inter-vention. A safe, effective, and non-invasive method to targetdelivery of drugs or genes to the specific disease site in thevasculature would greatly benefit cardiovascular treatments.

A novel approach to targeting drug and gene administrationis the method of ultrasound-enhanced delivery [13–15].Ultrasound-enhanced delivery often exploits cavitation bubbleactivity, which can be produced by the pressure oscillations ofultrasound [16]. Furthermore, ultrasound pressures above acertain threshold can cause oscillating bubbles to undergoviolent collapse known as inertial cavitation [17]. Inertialcavitation is believed to cause transient disruptions in cellmembranes, enabling transport of extracellular molecules (e.g.,

Fig. 1. (A) Equipment schematic. (B) artery segments were sutured to TPX®plastic disks exposing the endothelium for flat artery experiments. Confocalimages for quantification of bioeffects were captured in a 5×4 array, where eachimage was spaced by 1.5 mm.

286 D.M. Hallow et al. / Journal of Controlled Release 118 (2007) 285–293

drugs or genes) into viable cells [18–20]. Cavitation-mediatedcellular disruptions can allow uptake of small molecules,macromolecules (e.g., proteins), and genetic material (e.g.,plasmid DNA or siRNA). Furthermore, ultrasound-enhanceddelivery has been studied in a variety of in vitro and in vivoscenarios and has demonstrated promising therapeutic resultsafter intracellular uptake of drugs and gene expression[14,16,21]. Most importantly, ultrasound can be targeted inthe body by non-invasive extracorporeal focused ultrasound[22] or by minimally invasive catheter-based transducers [23].Greater efficacy and reduced side effects could be realized bythis targeted therapy.

The vascular endothelium is an attractive target forultrasound-enhanced delivery because cavitation can be readilyproduced in the vasculature, which currently can occur to a mildextent during diagnostic imaging [24]. Moreover, cavitation isexpected to have limited effects beyond cell layers directlyexperiencing cavitation activity, and the endothelium is the firstpoint of contact to cavitation activity [25]. A number ofresearchers are currently studying ultrasound-enhanced genetherapy for cardiovascular disorders in order to control intimalhyperplasia, restore vascular function, or promote angiogenesis[21]. These studies have shown expression of reporter plasmidsas well as plasmids with a therapeutic purpose. Clinicalpotential of this method has been shown by causing proteinexpression by plasmid DNA [26] or blocking specific proteinsby oligonucleotides [27].

A recognized limitation of ultrasound-enhanced delivery isthe need for more cells with drug uptake or gene expression[21]. Many studies have demonstrated gene transfection andtissue response in ex vivo and in vivo systems; however, fewstudies have been directed at quantifying the intracellularuptake of molecules and imaging bioeffects (i.e., intracellularuptake and loss of viability) caused by ultrasound-enhanceddelivery in viable tissue [16,28]. It is important to know theuptake efficiency at different ultrasound energies in order todesign and apply this technique for drug or gene deliveryapplications. In this study, we sought to determine whetherultrasound-enhanced delivery of a model drug can be targetedinto ECs and SMCs in ex vivo arteries. This work represents thefirst known in-depth study to quantify and image theintracellular uptake of molecules and loss of viability tovascular cells by ultrasound-enhanced delivery. We hypothesizethat ultrasound can target intracellular uptake of drugs into thevascular endothelium at low to intermediate energy and intoSMCs at high energy. To assess this hypothesis, our goals in thisstudy were to image the localization of (i) intracellular uptake ofa model drug and (ii) loss of viability to vascular cells and (iii) tospecifically quantify endothelial bioeffects in the targetedregion.

2. Methods

2.1. Porcine carotid artery isolation and preparation

Porcine carotid arteries were chosen for this study becausethey can be delicately excised without tissue damage, preserving

an intact and viable endothelium. Furthermore, carotid arteriesare relatively straight vessels without much branching, simpli-fying handling and imaging of the artery. Porcine carotid arterieswere harvested from freshly killed female swine at a localabattoir (Holifield Farms, Convington, GA). Excised arterieswere immediately rinsed with sterile phosphate-buffered saline(PBS, MediaTech, Herndon, VA) and placed in 15-ml conicaltubes filled with Dulbecco's Modified Eagle Medium (DMEM,MediaTech) supplemented with 100 μg/ml penicillin–strepto-mycin (MediaTech) and 10% (v/v) heat-inactivated fetal bovineserum (Atlanta Biologicals, Atlanta, GA). Arteries were thenplaced on ice for approximately 1 h during transport to thelaboratory. Upon return to the laboratory, arteries were placed infresh DMEM with 15 μM sodium nitroprusside (Sigma, St.Louis, MO) and incubated at 37 °C for approximately 30 minwhile the ultrasound apparatus was prepared. Sodium nitroprus-side, a nitric oxide donor, was added to the media to relax thearteries, improving the en face images of the artery (as describedbelow). Arteries were then cleaned of adventitial adipose andconnective tissue while submerged in DMEM.

2.2. Flat artery experiments

To prepare flat arteries for simplified exposure and analysis,the ends of the intact arteries were first discarded where theartery had been handled with forceps, which had likely causedtissue damage. Arteries were then cut cross-sectionally into 1-cm length segments and cut lengthwise to expose theendothelial surface. To approximate physiologic tension, arterysegments were stretched to ∼150% of their resting length [29],sutured to acoustically transparent TPX® plastic disks (25 mm

287D.M. Hallow et al. / Journal of Controlled Release 118 (2007) 285–293

dia., Ajedium Film, Newark, DE) with the endothelium exposed(Fig. 1), submerged in DMEM, and incubated at 37 °C. In selectexperiments, the artery endothelium was removed by rubbingthe endothelial surface with a plastic cell scraper (BectonDickinson, Franklin Lakes, NJ) denuding the artery completely,as verified by microscopy.

2.3. Intact artery experiments

To better mimic the in vivo geometry and environment, intactarteries were cut into 2.5-cm length segments and cannulated oneach end in custom-built chambers for ultrasound exposure.Chambers were rectangular-shaped boxes constructed of acrylicplastic with metal cannulas on each end to suspend the arterylengthwise and acoustically transparent windows constructed ofMylar® (DuPont, Wilmington, DE) plastic to allow ultrasoundexposure with minimal attenuation. Physiologic stretch andpressure were obtained by stretching the artery by ∼150% andpressurizing the artery with DMEM or blood to 100 mmHg. Thevolume outside the artery was filled with DMEM, and then thechamber was incubated at 37 °C.

2.4. Ultrasound apparatus

Ultrasound was produced using an immersible, focused,piezoceramic transducer (Sonic Concepts, Woodinville, WA,model no. H-101) at 1.1 MHz. The transducer had a 70-mmdiameter, 52-mm focal length, and 1.5-mm focal width at halfamplitude (−6 dB). Megahertz ultrasound was used in thisstudy because frequencies in the vicinity of 1.1 MHz have beenwell established in our lab and many others to stimulatecavitation activity and cause intracellular uptake of drugs andgenes [16]. Megahertz ultrasound is widely studied because itcan be highly focused and is considered safer than lowerfrequencies, since it is less likely to spontaneously producecavitation. Therefore, cavitation caused by megahertz ultra-sound is enhanced by microbubbles to nucleate cavitation andthereby adds another element of control by allowing one to alsoregulate cavitation by the microbubbles added to the system.

A sinusoidal waveform was produced by two functiongenerators (Standford Research Instruments, Sunnyvale, CA,model no. DS345 and Agilent, Austin, TX, model no. 33120A)used in conjunction to control pulse length, frequency, andpeak-to-peak voltage. The sinusoidal waveform was amplifiedby an RF broadband power amplifier (Electronic NavigationIndustries, Rochester, NY, model no. 3100LA) before passingthrough a matching impedance network and controlling theresponse of the transducer.

The transducer was housed in a polycarbonate tank(30.5×29×37 cm) containing approximately 26 L of deionizedand distilled water at 37 °C. A 5-cm thick acoustic absorber(SC-501 Acoustic Rubber, Sonic Concepts) was mountedopposite the transducer to minimize standing wave formation.A three-axis positioning system (10 μm resolution, Velmex,Bloomfield, NY) was mounted on top of the tank to positionsamples and a hydrophone at desired locations in the tank. APVDF membrane hydrophone (NTR Systems, Seattle, WA,

model no. MHA200A) was used to measure spatial-peak-temporal-peak negative pressure in order to map and calibratethe acoustic field produced by the transducer.

2.5. Experimental protocol

Prior to every experiment, the desired sample placement inthe acoustic field was found by using a needle hydrophone(NTR Systems, model no. TNU001A). The ultrasound beamfocus at 1.1 MHz had a focal-width at half-amplitude equal to1.5 mm, which would only allow a small region of the artery tobe exposed. To administer ultrasound to a larger portion of theartery, samples were placed in a location approximately 1 cmout of the ultrasound's focus towards the transducer, which hada focal-width at half-amplitude equal to 10.4 mm.

Prior to ultrasound exposure, a solution for each sample wasprepared with DMEM, 4 μM TO-PRO®-1 (Invitrogen, Molec-ular Probes, Eugene, OR), and 1.7% v/v (∼1.1×107 bubble/ml)Optison® (Mallinckrodt, St. Louis, MO). TO-PRO®-1, a greenfluorescent nucleic acid stain that is impermeable to intact cellmembranes, was used as a model drug and to quantifyintracellular uptake into viable cells. TO-PRO®-1 emits littlefluorescence until it binds to nucleic acids inside cells, whichfacilitates analysis by confocal microscopy by almost eliminat-ing background noise from extracellular TO-PRO®-1. In selectintact artery experiments, whole porcine blood, collected fromthe abattoir, with 5 USP units/ml of heparin (Baxter, Deerfield,IL) was used in place of DMEM to further imitate physiologicconditions.

Optison® is a suspension of perfluorocarbon gas bubblesstabilized with denatured human albumin that is an FDA-approved ultrasound imaging contrast agent but was used in thisstudy to nucleate cavitation activity [30]. Optison® bubbles ofmean diameter 3–5 μm are expected to have a resonantfrequency of ∼2 MHz and a peak negative pressure thresholdfor inertial cavitation of∼0.4 MPa at 1.1 MHz [17]. Optison® isconsidered unstable when exposed to air and can be difficult tokeep in a homogenous suspension due to buoyancy of thebubbles. Therefore, Optison® was only added immediatelybefore ultrasound exposure. During long ultrasound exposure,intense mixing caused by cavitation activity was expected tokeep the solution homogenous. It is envisioned that micro-bubbles of the relatively high concentrations used in this studycould be obtained in the body by local catheter release ofmicrobubbles [23] or by targeted microbubbles [14,31].

Flat arteries were placed in cylindrical sample chambers thathad the ends sealed with acoustically transparent TPX® plastic.Immediately before ultrasound exposure, sample chamberscontaining flat arteries were filled with a sample solution ofmedia, model drug, and Optison®. Intact arteries were filled withthe sample solution to physiologic pressure (∼100 mmHg).Sample chambers were then placed in the desired location inthe ultrasound apparatus using the three-axis positioningsystem to orient the flat artery surface and the intact arterysegment perpendicular to the acoustic beam. The endothe-lium was placed facing the transducer in the flat arteryexperiments.

Fig. 2. Confocal microscopy 2×2 image montages (10× magnification)displaying the localization of intracellular uptake enhanced by ultrasound.Images depict the endothelium of (A) a control sample and (B) a sample exposedto intermediate ultrasound energy. EC nuclei were labeled with Hoechst 33342(blue) to stain all cells, propidium iodide (red) to stain dead cells, and TO-PRO®-1 (green) to indicate intracellular uptake. A2 and B2 are shown withoutblue fluorescence to more clearly view intracellular uptake and cell death. Colorimages can be found in the web-based version of this article.

288 D.M. Hallow et al. / Journal of Controlled Release 118 (2007) 285–293

Ultrasound was applied to each sample at 1.1 MHz andacoustic energies of 5.0 (0.7 MPa–300 ms), 66 (1.4 MPa–1000 ms), or 630 J/cm2 (2.5 MPa–3000 ms), referred to as low,intermediate, and high ultrasound energy, respectively. To avoidheating, ultrasound was applied in 1-ms bursts followed by99 ms of off time, a 1% duty cycle. The experimental andacoustic parameters used for this study were based on previouswork with cells in suspension showing which conditions causedthe range of interesting effects [32].

Inertial cavitation production during ultrasound exposurewas verified using a passive cavitation detection system. Thissystem consisted of a focused hydrophone and a high-speeddigitizer to record cavitation emissions, as described previously[32]. Recordings of cavitation emissions were processed toderive frequency spectra and allowed monitoring of thebroadband noise, an established measure of inertial cavitation[33]. Cavitation detection further showed that cavitation activitywas sustained for a series of bursts equivalent to severalhundreds of milliseconds of ultrasound exposure. Afterultrasound exposure, samples were immediately incubated at37 °C for 15 min to allow arteries to “recover”, as membraneresealing is expected to occur on the order of a few minutes[19]. Arteries were then rinsed with DMEM and incubated at37 °C for b20 min until all samples had been exposed.

After ultrasound exposure, intact arteries were removed fromtheir chambers, cut open lengthwise, and sutured to rectangularTPX® plastic slides exposing the endothelium. Artery segmentswere then placed in DMEM with 0.07 mg/ml Hoechst 33342(Invitrogen) and 7 μg/ml propidium iodide (Invitrogen) for30 min in the incubator. Hoechst 33342 stains the nuclei of allcells with blue fluorescence, while propidium iodide only stainsthe nuclei of non-viable cells with red fluorescence.

2.6. Microscopy

Artery segments were observed using a Zeiss LSM 510confocal laser-scanning microscope or Zeiss LSM META/NLO510 multiphoton microscope (Zeiss, Thornwood, NY). Imagesof the endothelium and media were captured by directing a 40×magnification oil objective or a 10× magnification objective atthe endothelial surface of the artery. To grasp the extent andpattern of the bioeffects, a larger area of the endothelium wascaptured by taking four overlapping confocal images at 10×magnification and arranging them as a montage.

2.7. Quantification and statistical analysis

Quantification of endothelial bioeffects was objectivelydetermined by capturing 20 images at specific locations relativeto suturing holes that affixed the artery to the plastic disk.Twenty images of the artery endothelium were captured in a5×4 array at 40× magnification for each sample of the flatartery experiments (Fig. 1). Images were then processed inImage-Pro® Plus (Media Cybernetics, Silver Spring, MD) tocount all ECs (blue florescence), non-viable cells (redfluorescence), and viable cells with model drug uptake (greenfluorescence). The number of detached ECs from the artery

surface of each exposed sample was determined by subtractingthe number of remaining intact ECs from the total number ofintact ECs of a control sample, which was taken to represent thenumber of cells present prior to ultrasound exposure. Typicalcontrol samples would contain at least 4,000 ECs. The numberof cells for each population was averaged for each energy andnormalized to the control sample. Controls were samples thatunderwent a sham exposure, i.e., tissue treated identically toother samples except it was not exposed to ultrasound.

Statistical analysis was performed by use of paired Student'st-tests and analysis of variance (ANOVA) on quantifiedbioeffects from flat artery experiments. Values of Pb0.05were considered statistically significant.

3. Results

3.1. Endothelial bioeffects

To test the hypothesis that ultrasound-mediated cavitationcan cause intracellular uptake into ECs, we exposed ex vivoarterial segments to ultrasound at three different acousticenergies – termed low, intermediate, and high – and comparedthese tissues to control artery segments without ultrasoundexposure (i.e., sham exposure). Fig. 2 displays representativeen face images of the endothelium of a control tissue and atissue exposed to ultrasound at the intermediate energy. ECs arerecognized by their rounded nuclei morphology and orientationparallel to the direction of blood flow.

Fig. 2A1–A2 displays the endothelium of a representativecontrol sample. The control artery appears to have a complete

Fig. 3. Confocal microscopy images of the artery surface at 10× (A, C, E, G) and 40× magnification (B, D, G, H) showing the range of bioeffects at different ultrasoundenergies. Images depict (A–B) a control sample and samples exposed to ultrasound at (C–D) low, (E–F) intermediate, and (G–H) high energies. Figures A2–H2 areshown without blue fluorescence to more clearly view intracellular uptake and cell death. Color images can be found in the web-based version of this article.

289D.M. Hallow et al. / Journal of Controlled Release 118 (2007) 285–293

endothelium with high viability, as indicated by the continuumof ECs stained with blue fluorescent Hoechst 33342 (Fig. 2A1)and few ECs stained with red fluorescent propidium iodide(Fig. 2A2). None of the viable ECs appears to have greenfluorescent TO-PRO®-1 staining, verifying that TO-PRO®-1was membrane impermeable. These images also indicate thatarteries were isolated and handled in a manner that retained anintact endothelium and sustained the viability of the artery overthe course of the experiment.

We next investigated if ultrasound caused intracellularuptake in the endothelium. Fig. 2B1–B2 displays theendothelium of a representative artery exposed to ultrasoundat intermediate energy. Fluorescently green stained ECsobserved in the exposed sample (Fig. 2B2) clearly indicateintracellular uptake of the model drug, TO-PRO®-1. Since

propidium iodide stains any dead or permeable cells, the lack ofpropidium iodide staining in green ECs indicates that these cellswere reversibly permeabilized and remained viable. Thus, weconclude that ultrasound exposure significantly increasedintracellular uptake into viable ECs.

We also observed that intracellular uptake appeared in apattern of discreet regions or “patches” of fluorescently greennuclei on the surface of the artery (Fig. 2B2). This localizedpattern of intracellular uptake was regularly seen in tissuesexposed to ultrasound. We expect these regions of uptake can beexplained by considering the cavitation-mediated mechanismcausing bioeffects. It is commonly believed that bioeffectsobserved by this technique are due to discreet cavitation eventsand are heterogeneous within a sample due to the stochasticnature of cavitation. Similar patterns of discreet regions of

Fig. 4. Quantification of endothelial bioeffects following ultrasound exposure.Data represent the averages of n≥5 replicates with SEM shown.

290 D.M. Hallow et al. / Journal of Controlled Release 118 (2007) 285–293

bioeffects have been observed in in vitro monolayers of cellsexposed to ultrasound-mediated cavitation and attributed toindividual cavitation events occurring near the surface of themonolayer [34,35]. We therefore hypothesize that intracellularuptake is caused by many individual cavitation events occurringnear the surface of the artery and creates the observed scatteredpatches of intracellular uptake in the targeted region of the artery.

In addition to intracellular uptake, we also sought to assess theextent to which ultrasound exposure caused loss of cell viability.In Fig. 2B2, a decrease in viability of the exposed sample wasobserved, as shown by an increase in the number of fluorescentlyred nuclei compared to the control sample (Fig. 2A2). Thisindicates ultrasound exposure can cause EC death.

Guided by the general findings in Fig. 2, we wanted to assessendothelial bioeffects of intracellular uptake and loss of viabilityin greater detail at different ultrasound energies. In Fig. 3,representative confocal images of control samples and samplesexposed at low, intermediate, and high ultrasound energies areshown at two different magnifications. The extent of bioeffectsappears to sharply increase with increasing ultrasound energy.The bioeffects ranged from (i) little intracellular uptake or celldeath at the lowest energy (Fig. 3C–D) to (ii) widespreadintracellular uptake and reduced viabilities at the intermediateenergy (Fig. 3E–F) to (iii) widespread denuding of theendothelium and cell death at the highest ultrasound energy(Fig. 3G–H). In Fig. 3G1–H1, denuding of the endothelium isrecognized by the lack of ECs and, instead, the characteristicimage of the internal elastic lamina, which is highly autofluor-escent in the blue and green channels.

3.2. Quantification of endothelial bioeffects

We next wanted to quantify the extent of intracellular uptake,loss of viability, and denuding present after different ultrasoundexposures. We used multiple en face images at 40× magnifi-cation (e.g., Fig. 3B, D, F, and H) of each sample for analysisusing image processing software to quantify the number of ECsthat were (i) viable, (ii) viable with intracellular uptake, (iii)non-viable, and (iv) removed from the artery (presumed non-viable).

Fig. 4 displays the percentage of cells in each of these fourpopulations at low, intermediate, and high acoustic energy. Thenumber of viable cells with intracellular uptake increased withincreasing energy from the control (0% of cells) to low energy(9% of cells) to intermediate energy (24% of cells) (ANOVA,Pb0.002). The number of cells with intracellular uptake atthe highest acoustic energy (8% of cells) decreased due tosubstantial loss in viability.

Fig. 4 further indicates that the total number of cells presenton the artery surface and the number of viable ECs decreasedwith increasing acoustic energy (ANOVA, Pb0.001 andPb0.001, respectively). The difference in the total number ofECs compared to the control indicated the amount of denudingwas only statistically significant at the highest acoustic energy(∼67% of cells were removed, Pb0.001). The number of viableECs was only significantly different when comparing the highacoustic energy (∼25% of cells remained viable) to the control

(Pb0.004) and comparing the intermediate energy (∼82%) tothe control (Pb0.03). The viability at the low energy (∼93%)was not significantly different from the control (P=0.07).

Considering drug delivery applications, these results suggestthat low proportions of intracellular uptake with insignificantloss in viability can be obtained at low energies, whereas ahigher proportion of intracellular uptake with some loss ofviability can be achieved at intermediate energy. Finally, if alarge loss of viability and significant denuding is tolerable (ordesired), intracellular uptake can be achieved among remainingviable cells by applying high ultrasound energies.

3.3. Medial bioeffects

Having observed and quantified endothelial bioeffectscaused by ultrasound-mediated cavitation, we next determinedif cavitation had any effects deeper into arterial tissue to theSMCs present in the medial layer. In some drug deliveryapplications, it may be preferable to induce intracellular uptaketo the underlying SMCs, such as to inhibit SMC proliferation.

By changing the focus depth of the confocal microscope, itwas possible to optically slice the artery and capture imagesdeeper in the tissue to approximately 40 μm below the endo-thelial surface. In Fig. 5A, the surface of an artery exposed tothe highest acoustic energy was captured. Extensive denudingof the artery was evident, based on the lack of ECs and theexposure of the internal elastic lamina. Images captured deeperin the tissue, as shown in Fig. 5B–C, display SMCs that arerecognized by their long and slender nuclei and orientationperpendicular to the direction of original blood flow. Asindicated by propidium iodide and TO-PRO®-1 staining, SMCsin Fig. 5B–C exhibited intracellular uptake, as well as loss ofviability. In tissues exposed at lower acoustic energies, themedial layer did not show intracellular uptake or loss ofviability (data not shown).

Fig. 6. Confocal microscopy images of endothelium displaying bioeffectsmediated by ultrasound exposure to intact arteries at near physiologicconditions. Ultrasound was applied at (A, C) intermediate and (B, D) highenergy while the artery was filled with (A, B) DMEM or (C, D) blood. Colorimages can be found in the web-based version of this article.

291D.M. Hallow et al. / Journal of Controlled Release 118 (2007) 285–293

To determine if the presence of endothelial cells was a barrierto cavitation effects to smooth muscle cells, we pre-denudedarteries by rubbing the endothelial surface with a cell scraper.Confocal images from these studies (data not shown) indicatedthat SMC bioeffects (i.e., intracellular uptake and loss ofviability) were small at low to intermediate energy and wide-spread at high energy, similar to Fig. 5. This result suggests thatthe internal elastic lamina serves as the main barrier to cavi-tation, and high acoustic energies are necessary to mediateintracellular uptake to SMCs.

Overall these results demonstrate that bioeffects be candirected to SMCs in the medial layer of the artery; however, weexpect that effects are limited to the superficial layers of themedia considering the cavitation-mediated mechanism, which isexpected to have limited depth penetration [25]. This treatmentof SMCs may be applicable in cases where denuding of theartery is tolerable or has already occurred due to some pre-treatment, such as percutaneous coronary interventions.

3.4. Intact arteries

Lastly, we wanted to determine the role of physiologicalconditions such as geometry, pressure, tension, or presence ofblood (a higher viscosity fluid) on bioeffects induced by ultra-sound exposure. To better mimic physiologic geometry andconditions, intact arteries were stretched and pressurized toapproximately physiologic tension and pressure during expo-sure to ultrasound at the intermediate and high acoustic ener-gies. The arteries were also filled with DMEM or whole porcineblood containing Optison® and TO-PRO®-1.

Fig. 5. Confocal microscopy images at multiple depths in the artery displayingbioeffects to medial SMCs. Images were captured at (A) the artery surface and(B) 13 μm and (C) 21 μm below the artery surface. Color images can be found inthe web-based version of this article.

Fig. 6A–B displays representative confocal microcopyimages of an intact artery filled with DMEM and exposedat the intermediate and high acoustic energies. Similarly,Fig. 6C–D displays representative confocal microcopy imagesof the endothelium of an intact artery that was filled with wholeblood and exposed to ultrasound at intermediate and highacoustic energies. Intact arteries filled with either DMEM orwhole blood and exposed at either energy displayed similarbioeffects to the results found in the flat artery experiments(Fig. 2) and therefore show that the geometry and presence ofblood in a pressurized and stretched vessel did not significantlyinfluence ultrasound exposure to cause bioeffects. Bioeffectsappeared to be uniformly distributed about the circumferenceand not limited to the proximal or distal wall relative to thetransducer. These results also suggest that the bioeffectsobserved may be relevant for realistic drug delivery scenarios.

4. Discussion

Previous studies utilizing ultrasound-enhanced gene therapyhave shown promising results of positive gene transfection inex vivo [36] and in vivo [26] cardiovascular tissues. However,few studies have shown intracellular uptake of small moleculesor macromolecules (e.g., proteins) for drug delivery applicationsin ex vivo or in vivo cardiovascular tissues [16,21]. Moreover,there is a lack of knowledge on the efficiency of this technique interms of the number of cells affected and the relevantpopulations. In this study, viable ex vivo porcine carotid arterieswere exposed to ultrasound and demonstrated the ability of thismethod to cause intracellular uptake of small molecules into ECsand SMCs. Furthermore, quantitative data showed the approx-imate energies that are relevant for drug delivery applications.

292 D.M. Hallow et al. / Journal of Controlled Release 118 (2007) 285–293

Many cardiovascular diseases and dysfunctions wouldbenefit from a targeted means to deliver drugs or genes to theendothelium [3]. Quantitative results (Fig. 4) suggest that low tointermediate ultrasound energies would be appropriate for drugor gene delivery applications to cause intracellular uptake orpossibly transfect a large number of cells without a large loss ofviability. Extensive cell death could be a drawback for drugdelivery applications, since endothelial injury in the form ofdeath or denudation could initiate an inflammation response oreven thrombosis [37]. Intracellular uptake results suggest thatpotential applications might be limited to drug or gene deliveryto a minority of the cell population, yet some gene deliveryapplications may not require a high transfection efficiency if theexpressed protein is excreted and has a local effect. To causeintracellular uptake to the majority of the endothelium, repeatedapplications of ultrasound or other optimization approachesmay facilitate higher uptake percentages without a harmful lossof viability. Cardiovascular diseases that could benefit from thismethod of targeted delivery include diseased sites such asvulnerable atherosclerotic plaques and ischemic heart or pe-ripheral tissue. Intracellular therapeutic targets such as NADPHoxidases or gene expression of endothelial nitric oxide synthase(eNOS) in the endothelium could help control neointimalformation and restore vascular function [3,38]. Promotingangiogenesis by intracellular delivery or gene expression agentssuch as vascular endothelial growth factor (VEGF) or fibroblastgrowth factor (FGF) in the endothelium could also benefitischemic injuries in the myocardium or peripheral vasculature[10].

In other cases of cardiovascular disease, targeted drug or genedelivery to SMCs may be desired as is currently performedclinically with drug-eluting stents to control SMC proliferation[7] and is a goal of many gene therapy treatments [10].We foundthat ultrasound exposure could also cause medial bioeffects,which were only observed at high ultrasound energies. Resultssuggest that cavitation is obstructed by the internal elastic laminaand limited to superficial cell layers. Intracellular uptake toSMCs might be used in applications where denuding and loss ofviability is acceptable or has already occurred, such as treatmentfor restenosis post-angioplasty. In this potential application,extensive denuding and injury to the endothelium typicallyoccurs from balloon inflation during percutaneous coronaryinterventions [37]. At this point, it may be advantageous to useultrasound-enhanced cavitation to deliver drugs or genes toSMCs to inhibit smooth muscle proliferation and promote re-endothelialization for vascular recovery [10].

At higher ultrasound energies, extensive denuding of theendothelium was produced at the targeted site, as imaged inFig. 3G–H and quantified in Fig. 4. A targeted loss of viabilityor denuding may not be desirable for drug or gene deliveryapplications but may be beneficial for other therapeutic appli-cations. A number of vascular targeting agents, such as com-bretastatins, are currently being studied to cause EC death foranti-vascular targeting [39]. This therapy aims to selectivelydestroy endothelium of tumor blood vessels and subsequentlycause vascular shutdown, which limits sustaining nutrients andwaste removal required by the tumor [5]. High ultrasound

energies focused at a tumor may be able to create this desiredeffect by denuding the artery.

In conclusion, this work supported the hypothesis thatultrasound-mediated cavitation can target intracellular uptakeinto ECs. Furthermore, we have for the first time imaged thelocalized patterns of intracellular uptake and shown them to beheterogeneously scattered patches that may correspond toultrasound-mediated cavitation events. These regions of uptakeand cell death often appeared to be several cells in diameter, adimension consistent with estimates of “blast radii” from singlecavitation events [18,40] and related observations of denudingcells off cultured monolayers [35]. We have also quantified theendothelial bioeffects due to ultrasound-enhanced delivery andfound that intracellular uptake can occur in up to 24% of ECs atintermediate ultrasound energy and that widespread denudingof ECs from the arterial surface can occur at high energy. Usingintact arteries as an ex vivo drug delivery model, this studyshowed that the same bioeffects occur at near physiologicconditions. These results exhibit advantages and limitations ofultrasound-enhanced delivery that can further guide research onthis technique for drug or gene delivery applications to treatcardiovascular diseases and dysfunctions.

Acknowledgments

We thank Holifield Farms for their donation of porcine tissueand laboratory members Mangesh Despande, Vladimir Zarnit-syn, Robyn Schlicher, Joshua Hutcheson, Christina Rostad, andPrerona Chakravarty for their helpful discussions. This workwas supported in part by the U.S. National Institutes of Health,and a fellowship to DMH from the U.S. Department ofEducation GAANN program. DMH and MRP are members ofthe Center for Drug Design, Development and Delivery and theInstitute of Bioengineering and Biosciences at the GeorgiaInstitute of Technology.

References

[1] S.C. Smith Jr., R. Jackson, T.A. Pearson, V. Fuster, S. Yusuf, O. Faergeman,D.A. Wood, M. Alderman, J. Horgan, P. Home, M. Hunn, S.M. Grundy,Principles for national and regional guidelines on cardiovascular diseaseprevention: a scientific statement from the World Heart and Stroke Forum,Circulation 109 (25) (2004) 3112–3121.

[2] J.P. Cooke, The endothelium: a new target for therapy, Vasc. Med. 5 (1)(2000) 49–53.

[3] L.G. Melo, M. Gnecchi, A.S. Pachori, D. Kong, K. Wang, X. Liu, R.E.Pratt, V.J. Dzau, Endothelium-targeted gene and cell-based therapies forcardiovascular disease, Arterioscler. Thromb. Vasc. Biol. 24 (10) (2004)1761–1774.

[4] D.B. Cines, E.S. Pollak, C.A. Buck, J. Loscalzo, G.A. Zimmerman, R.P.McEver, J.S. Pober, T.M. Wick, B.A. Konkle, B.S. Schwartz, E.S.Barnathan, K.R. McCrae, B.A. Hug, A.M. Schmidt, D.M. Stern,Endothelial cells in physiology and in the pathophysiology of vasculardisorders, Blood 91 (10) (1998) 3527–3561.

[5] G.M. Tozer, C. Kanthou, B.C. Baguley, Disrupting tumour blood vessels,Nat. Rev., Cancer 5 (6) (2005) 423–435.

[6] M. Malecki, P. Kolsut, R. Proczka, Angiogenic and antiangiogenic genetherapy, Gene Ther. 12 (Suppl 1) (2005) S159–S169.

[7] F. Saia, A. Marzocchi, P.W. Serruys, Drug-eluting stents. The thirdrevolution in percutaneous coronary intervention, Ital. Heart J. 6 (4) (2005)289–303.

293D.M. Hallow et al. / Journal of Controlled Release 118 (2007) 285–293

[8] F. Sharif, K. Daly, J. Crowley, T. O'Brien, Current status of catheter- andstent-based gene therapy, Cardiovasc. Res. 64 (2) (2004) 208–216.

[9] A.H. Baker, A. Kritz, L.M. Work, S.A. Nicklin, Cell-selective viral genedelivery vectors for the vasculature, Exp. Physiol. 90 (1) (2005) 27–31.

[10] R. Quarck, P. Holvoet, Gene therapy approaches for cardiovasculardiseases, Curr. Gene Ther. 4 (2) (2004) 207–223.

[11] T. Niidome, L. Huang, Gene therapy progress and prospects: nonviralvectors, Gene Ther. 9 (24) (2002) 1647–1652.

[12] N.N. Kipshidze, T.R. Porter, G. Dangas, H. Yazdi, F. Tio, F. Xie, D.Hellinga, R. Wolfram, R. Seabron, R. Waksman, A. Abizaid, G. Roubin, S.Iyer, A. Colombo, M.B. Leon, J.W. Moses, P. Iversen, Novel site-specificsystemic delivery of Rapamycin with perfluorobutane gas microbubblecarrier reduced neointimal formation in a porcine coronary restenosismodel, Catheter. Cardiovasc. Interv. 64 (3) (2005) 389–394.

[13] E.C. Unger, E. Hersh, M. Vannan, T.O. Matsunaga, T. McCreery, Localdrug and gene delivery through microbubbles, Prog. Cardiovasc. Dis. 44(1) (2001) 45–54.

[14] Y. Liu, H. Miyoshi, M. Nakamura, Encapsulated ultrasound microbubbles:therapeutic application in drug/gene delivery, J. Control. Release 114 (1)(2006) 89–99.

[15] R. Suzuki, T. Takizawa, Y. Negishi, K. Hagisawa, K. Tanaka, K.Sawamura, N. Utoguchi, T. Nishioka, K. Maruyama, Gene delivery bycombination of novel liposomal bubbles with perfluoropropane andultrasound, J. Control. Release (2006), doi:10.1016/j.jconrel.2006.09.008.

[16] S.Mitragotri, Healing sound: the use of ultrasound in drug delivery and othertherapeutic applications, Nat. Rev. Drug Discov. 4 (3) (2005) 255–260.

[17] T.G. Leighton, The Acoustic Bubble, Academic Press, London, 1994.[18] J. Sundaram, B.R. Mellein, S. Mitragotri, An experimental and theoretical

analysis of ultrasound-induced permeabilization of cell membranes,Biophys. J. 84 (5) (2003) 3087–3101.

[19] R.K. Schlicher, H. Radhakrishna, T.P. Tolentino, R.P. Apkarian, V.Zarnitsyn, M.R. Prausnitz, Mechanism of intracellular delivery by acousticcavitation, Ultrasound Med. Biol. 32 (6) (2006) 915–924.

[20] S. Mehier-Humbert, T. Bettinger, F. Yan, R.H. Guy, Plasma membraneporation induced by ultrasound exposure: implication for drug delivery,J. Control. Release 104 (1) (2005) 213–222.

[21] R. Bekeredjian, P.A. Grayburn, R.V. Shohet, Use of ultrasound contrastagents for gene or drug delivery in cardiovascular medicine, J. Am. Coll.Cardiol. 45 (3) (2005) 329–335.

[22] P.E. Huber, M.J. Mann, L.G. Melo, A. Ehsan, D. Kong, L. Zhang, M.Rezvani, P. Peschke, F. Jolesz, V.J. Dzau, K. Hynynen, Focused ultrasound(HIFU) induces localized enhancement of reporter gene expression inrabbit carotid artery, Gene Ther. 10 (18) (2003) 1600–1607.

[23] P.G. Amabile, J.M. Waugh, T.N. Lewis, C.J. Elkins, W. Janas, M.D. Dake,High-efficiency endovascular gene delivery via therapeutic ultrasound,J. Am. Coll. Cardiol. 37 (7) (2001) 1975–1980.

[24] J.M. Correas, L. Bridal, A. Lesavre, A. Mejean, M. Claudon, O. Helenon,Ultrasound contrast agents: properties, principles of action, tolerance, andartifacts, Eur. Radiol. 11 (8) (2001) 1316–1328.

[25] J.H. Hwang, A.A. Brayman, M.A. Reidy, T.J. Matula, M.B. Kimmey, L.A.Crum, Vascular effects induced by combined 1-MHz ultrasound andmicrobubble contrast agent treatments in vivo, Ultrasound Med. Biol. 31(4) (2005) 553–564.

[26] Y. Taniyama, K. Tachibana, K. Hiraoka, T. Namba, K. Yamasaki, N.Hashiya, M. Aoki, T. Ogihara, K. Yasufumi, R. Morishita, Local deliveryof plasmid DNA into rat carotid artery using ultrasound, Circulation 105(10) (2002) 1233–1239.

[27] N. Hashiya, M. Aoki, K. Tachibana, Y. Taniyama, K. Yamasaki, K.Hiraoka, H. Makino, K. Yasufumi, T. Ogihara, R. Morishita, Localdelivery of E2F decoy oligodeoxynucleotides using ultrasound withmicrobubble agent (Optison) inhibits intimal hyperplasia after ballooninjury in rat carotid artery model, Biochem. Biophys. Res. Commun. 317(2) (2004) 508–514.

[28] D. Mukherjee, J. Wong, B. Griffin, S.G. Ellis, T. Porter, S. Sen, J.D.Thomas, Ten-fold augmentation of endothelial uptake of vascularendothelial growth factor with ultrasound after systemic administration,J. Am. Coll. Cardiol. 35 (6) (2000) 1678–1686.

[29] N.C. Chesler, B.S. Conklin, H.-C. Hand, D.N. Ku, Simplified ex vivoartery culture techniques for porcine arteries, J. Vasc. Investig. 4 (1998)123–127.

[30] P.P. Kamaev, J.D. Hutcheson, M.L. Wilson, M.R. Prausnitz, Quantificationof optison bubble size and lifetime during sonication dominant role ofsecondary cavitation bubbles causing acoustic bioeffects, J. Acoust. Soc.Am. 115 (4) (2004) 1818–1825.

[31] A.L. Klibanov, Targeted delivery of gas-filled microspheres, contrastagents for ultrasound imaging, Adv. Drug Deliv. Rev. 37 (1–3) (1999)139–157.

[32] D.M. Hallow, A.D. Mahajan, T.E. McCutchen, M.R. Prausnitz, Measure-ment and correlation of acoustic cavitation with cellular bioeffects,Ultrasound Med. Biol. 32 (7) (2006) 1111–1122.

[33] Acoustical Society of America, ANSI Technical Report: Bubble Detectionand Cavitation Monitoring, 2002.

[34] A.A. Brayman, L.M. Lizotte, M.W. Miller, Erosion of artificial endotheliain vitro by pulsed ultrasound: acoustic pressure, frequency, membraneorientation and microbubble contrast agent dependence, Ultrasound Med.Biol. 25 (8) (1999) 1305–1320.

[35] C.D. Ohl, B. Wolfrum, Detachment and sonoporation of adherent HeLa-cellsby shock wave-induced cavitation, Biochim. Biophys. Acta 1624 (1–3)(2003) 131–138.

[36] C. Teupe, S. Richter, B. Fisslthaler, V. Randriamboavonjy, C. Ihling, I.Fleming, R. Busse, A.M. Zeiher, S. Dimmeler, Vascular gene transfer ofphosphomimetic endothelial nitric oxide synthase (S1177D) usingultrasound-enhanced destruction of plasmid-loaded microbubblesimproves vasoreactivity, Circulation 105 (9) (2002) 1104–1109.

[37] C. Davis, J. Fischer, K. Ley, I.J. Sarembock, The role of inflammation invascular injury and repair, J. Thromb. Haemost. 1 (8) (2003) 1699–1709.

[38] R. Ray, A.M. Shah, NADPH oxidase and endothelial cell function, Clin.Sci. (Lond.) 109 (3) (2005) 217–226.

[39] P.E. Thorpe, Vascular targeting agents as cancer therapeutics, Clin. CancerRes. 10 (2) (2004) 415–427.

[40] H.R. Guzman, A.J. McNamara, D.X. Nguyen, M.R. Prausnitz, Bioeffectscaused by changes in acoustic cavitation bubble density and cellconcentration: a unified explanation based on cell-to-bubble ratio andblast radius, Ultrasound Med. Biol. 29 (8) (2003) 1211–1222.