Local and reversible blood–brain barrier disruption by noninvasive focused ultrasound at frequencies suitable for trans-skull sonications

www.elsevier.com/locate/ynimg

NeuroImage 24 (2005) 12–20

Local and reversible blood–brain barrier disruption by noninvasive

focused ultrasound at frequencies suitable for trans-skull sonications

Kullervo Hynynen,* Nathan McDannold, Nickolai A. Sheikov,

Ferenc A. Jolesz, and Natalia Vykhodtseva

Department of Radiology, Brigham and Women’s Hospital, and Harvard Medical School, Boston, MA 02115, USA

Received 20 November 2003; revised 4 June 2004; accepted 11 June 2004

The purpose of this study was to test the hypothesis that burst

ultrasound in the presence of an ultrasound contrast agent can disrupt

the blood–brain barrier (BBB) with acoustic parameters suitable for

completely noninvasive exposure through the skull. The 10-ms

exposures were targeted in the brains of 22 rabbits with a frequency

of 690 kHz, a repetition frequency of 1 Hz, and peak rarefactional

pressure amplitudes up to 3.1 MPa. The total exposure (sonication)

time was 20 s. Prior to each sonication, a bolus of ultrasound contrast

agent was injected intravenously. Contrast-enhanced MR images were

obtained after the sonications to detect localized BBB disruption via

local enhancement in the brain. Brain sections were stained with H&E,

TUNEL, and vanadium acid fuchsin (VAF)–toluidine blue staining. In

addition, horseradish peroxidase (HRP) was injected into four rabbits

prior to sonications and transmission electron microscopy was

performed. The MRI contrast enhancement demonstrated BBB

disruption at pressure amplitudes starting at 0.4 MPa with approx-

imately 50%; at 0.8 MPa, 90%; and at 1.4 MPa, 100% of the sonicated

locations showed enhancement. The histology findings following 4 h

survival indicated that brain tissue necrosis was induced in approx-

imately 70–80% of the sonicated locations at a pressure amplitude level

of 2.3 MPa or higher. At lower pressure amplitudes, however, small

areas of erythrocyte extravasation were seen. The electron microscopy

findings demonstrated HRP passage through vessel walls via both

transendothelial and paraendothelial routes. These results demonstrate

that completely noninvasive focal disruption of the BBB is possible.

D 2004 Elsevier Inc. All rights reserved.

Keywords: Ultrasound; Bioeffects; Cavitation; Blood–brain barrier; Apop-

tosis; Ischemia

1053-8119/$ - see front matter D 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.neuroimage.2004.06.046

* Corresponding author. Department of Radiology, Brigham and

Women’s Hospital, and Harvard Medical School, 75 Francis Street, Boston,

MA 02115. Fax: +1 617 278 0610.

E-mail address: [email protected] (K. Hynynen).

Available online on ScienceDirect (www.sciencedirect.com.)

Introduction

Advances in neuroscience have resulted in the development of

new diagnostic and therapeutic agents and genes that may be used

to study and treat many central nervous system (CNS) diseases

(Pardridge, 2002a). However, the use of these agents is often

limited by their access to the CNS via the blood supply because the

blood–brain barrier (BBB) protects the brain from foreign

molecules (Abbott and Romero, 1996; Kroll and Neuwelt, 1998;

Nag, 2003b; Pardridge, 2002a).

Several methods have been tested to circumvent the BBB

including chemical modification of drugs to make them lipophilic

or the use of carriers to aid propagation through the barrier

(Pardridge, 2002a,b, 2003). Another option is to utilize an

intraarterial infusion of hypertonic solution that initiates a BBB

disruption lasting from a few minutes to a few hours (Doolittle et

al., 2000). Each of the methods described above requires intra-

arterial catheterization and produces diffuse, nonfocal BBB

disruption within the entire tissue volume supplied by the injected

artery branch (Abbott and Romero, 1996; Kroll and Neuwelt,

1998). The BBB disruption over a large volume, however, may

prove a disadvantage for many applications since the agents may

have undesired, often dose-limiting side-effects due to their spread

within the CNS. To prevent this, a localized and reversible image-

guided disruption of the BBB would provide anatomically or

functionally targeted drug delivery while preserving an intact BBB

to protect the nontargeted regions. Today, such localized drug

delivery can only be accomplished by a direct injection of the agent

into the targeted brain volume (Kroll and Neuwelt, 1998). Such

direct injections are limited by the slow diffusion of molecules in

the brain, the need to open the skull and penetrate nontargeted

brain tissue, and the attendant risk of neurological damage,

bleeding, and infection.

There is experimental evidence that focused ultrasound can

selectively disrupt the BBB locally in the brain (Mesiwala et

al., 2002; Patrick et al., 1990; Vykhodtseva et al., 1995).

However, the disruption is often associated with damage to the

exposed brain tissue. Recently, a method has been developed

wherein the ultrasound exposures are performed in the presence

Table 1

Summary of the experiments

Experiment No. of

rabbits

No. of

locations

Pressure

amplitude

(MPa)

Average

time between

sonication and

contrast injection

Determine

thresholds

for the BBB

disruption/

tissue damage

14/13a 52/48a 0.4–3.1 24 min (n = 14)

81 min (n = 4)

Verify closing

of the BBB

4 7 0.8, 1.0 6 min (n = 4)

3 h (n = 4)

5 h (n = 1)

Electron

microscopy

4 11 0.8 0 (n = 4)

a One brain was lost before histology preparation.

K. Hynynen et al. / NeuroImage 24 (2005) 12–20 13

of a circulating agent consisting of preformed gas bubbles

(Hynynen et al., 2001). The MRI and histology results have

shown reproducible and reversible BBB disruption at the

frequency of 1.63 MHz. Since the bubbles are restricted to

the vasculature, the injury to the surrounding brain tissue has

proven minimal.

One major shortcoming of ultrasound is that it is strongly

attenuated by bone; hence, for brain applications (according to

conventional practice), an acoustic window must be constructed

by a craniotomy. Recent experimental and theoretical papers

(Hynynen and Jolesz, 1998; Sun and Hynynen, 1998; Thomas

and Fink, 1996) have shown that focal, trans-skull ultrasound

exposure of brain tissue may be accomplished by using large

surface area phased arrays and information derived from modern

imaging methods to correct the ultrasound wave distortion

produced by the skull (Clement and Hynynen, 2002; Sun and

Hynynen, 1998). These studies indicate, moreover, that optimal

transcranial focusing can be achieved at frequencies below 1

MHz. However, since the behavior of gas bubbles in an

ultrasound field is highly dependant on the frequency; it is not

known whether the BBB disruption can be safely induced at

frequencies sufficiently low to allow propagation through the

skull bone.

The purpose of this paper is to explore the disruption of the

BBB in a frequency range feasible for trans-skull sonications and

to determine the biological route for the material transport into the

brain tissue.

Materials and methods

Ultrasound equipment

The ultrasound fields were generated by an in-house manufac-

tured, focused, piezoelectric transducer with a 100-mm diameter,

an 80-mm radius of curvature, and a resonant frequency of 0.69

MHz. The half-maximum pressure amplitude diameter and length

of the produced focal spot were 2.3 and 14 mm, respectively. The

transducer driving equipment were similar to those described in

(Hynynen et al., 2001).

Animals

Twenty-two (22) New Zealand white rabbits (approximately

3–4 kg, males) were anesthetized by a mix of 40 mg/kg

ketamine (Aveco Co., Inc., Fort Dodge, IA) and 10 mg/kg of

Fig. 1. A diagram showing the sonication arrangement.

xylazine (Lloyd Laboratories, Shenandoah, IA). An ultrasound

window was created by removing a piece of skull (approx-

imately 20 � 20 mm) and replacing the skin over the bone

window. This opening allowed for a more accurate estimation of

the peak focal pressure amplitude in the tissue than is achievable

with sonications through the skull. The experiments were

performed no sooner than 10 days postsurgery. The head was

shaved and any residual hair was removed with hair remover

lotion. The head was fixed in the treatment position by a plastic

holder on top of the water tank (Fig. 1). The sonications were

performed through the skull opening such that each of the focal

spots were aimed into a different location of the brain (at least 3

mm apart) so that the focal spots were not overlapping. All of

these experiments were approved by our institutional animal

committee.

Sonications

The animals were prepared for the experiments as described

above and placed on the system. T2-weighted images were

obtained to determine the extent of the bone window and to aid

in the selection of target locations in the brain. The sonications

were delivered to between one and six locations in the gray matter

with the center of the focal spot at approximately 10 mm deep in

each brain. The sonications were pulsed with a burst length of 10

ms and a repetition frequency of 1 Hz. The duration of the whole

sonication was 20 s. The peak acoustic power levels were kept

constant over the duration of each sonication but ranged from 0.08

to 8 W depending on the sonication. The acoustic power output

and the focal pressure amplitude as a function of applied

radiofrequency power was measured as described earlier (Hynynen

et al., 1997).

Approximately 10 s before each sonication, a bolus (0.05

ml/kg) of ultrasound contrast agent (Optison, Mallinckrodt Inc.,

St. Louis, MO) that contains albumin coated microbubbles

(mean diameter = 2.0–4.5 � 10�6 m; concentration = 5–8 �108 bubbles/ml) was injected into the ear vein (Hynynen et al.,

2001). Five- to ten-minute intervals between sonications allowed

the bubbles to mostly clear from the circulation before the next

sonication.

Three sets of experiments were performed (Table 1). In the first,

14 rabbits were sonicated over the full range of pressure amplitudes

(0.4– 3.1 MPa). These locations were examined for signal intensity

Table 2

The MRI parameters used in the study

Sequence Purpose Repetition

time (ms)

Echo

time

(ms)

Matrix Flip

angle

(8)

Bandwidth

(kHz)

Number of

acquisitions

FSE T1-W Contrast Enhancement 500 15 256 � 256 90 16 4

FSE T2-W Tissue Anatomy Targeting 2000 75 256 � 256 90 16 2

FSE T1-W = fast spin echo T1-weighted sequence; FSE T2-W = fast spin echo T2-weighted sequence.

K. Hynynen et al. / NeuroImage 24 (2005) 12–2014

enhancement in MRI and with light microscopy to establish the

thresholds for BBB disruption and tissue damage. In these

experiments, sonications were performed at locations in both

hemispheres to maximize the number of locations.

In the second set of experiments, an additional four rabbits were

sonicated at pressure amplitudes of 0.8 and 1.0 MPa, which is

below the threshold for tissue damage established in the first set of

experiments (see below). In these animals, one to two locations

were sonicated in one hemisphere; the contralateral hemisphere

served as a control. These rabbits were followed with MR imaging

for several hours to establish the amount of time needed for repair

of the BBB disruption. The MR contrast agent injection was

repeated at desired time points after the sonication. These animals

were not part of the light microscopy study.

In the third set of experiments, another four rabbits were

sonicated at pressure amplitudes of 0.8 MPa and were used for the

electron microscopy study described below. These animals were

not examined in light microscopy.

MRI

The MRI scanner was a standard 1.5 T Signa system

(General Electric Medical Systems, Milwaukee, WI). A 7.5-cm

diameter surface coil was placed under the head. The sonications

were performed through the hole of the coil that was filled with

the bag containing degassed water (Fig. 1). A T2-weighted fast

spin echo (FSE) sequence was used to aim the beam at the

brain. Following the sonications, T1-weighted FSE images were

obtained and repeated after injection of an intravenous (iv) bolus

(0.125 mmol/kg) of gadopentetate dimeglumine MR contrast

agent (Magnevist, Berlex Laboratories Inc., Wayne, NJ) to detect

and evaluate the opening of the blood–brain barrier (BBB). The

parameters for the MRI scans are given in Table 2. The time

between the sonications and the contrast agent are listed in

Table 1.

Fig. 2. A contrast-enhanced FSE T1-weighted image of a rabbit brain. (A) Image p

and 1.4 MPa (subtraction: before � after injection) and (B) image oriented along

Signal analysis

The MRI contrast enhancement was evaluated at each target

location by averaging the signal intensity over a 3 � 3 voxel (1.2 �1.2 mm) region of interest (ROI). The signal was normalized to the

baseline value in the ROI before the contrast injection. ROIs of

nonsonicated brain tissue proximal to sonicated locations served as

controls. The average, normalized contrast enhancement observed

in these control locations was subtracted from that in the targeted

locations to estimate the impact of the sonications. In addition, a

larger 6 � 6 voxel control area was selected outside of the

sonicated regions, and the average standard deviation of the signal

intensities of the voxels in that area was calculated over the entire

time series of images. If the contrast enhancement at the focal area

was higher than the average standard deviation value in the control

area, then the sonication was judged to have induced disruption of

the BBB.

In the four rabbits sonicated in one hemisphere only, the

signal enhancement at the focal coordinate was compared to that

in the same structure in the contralateral hemisphere. This

method allowed for an internal control that reduced any

uncertainties that might be associated with selecting the control

regions in different structures when targets were sonicated in

both hemispheres.

Light microscopy

The animals were sacrificed approximately 4 h after the last

sonication (in one case, at 48 h). The brains were immediately

removed and fixed in 10% buffered neutral formalin. They were

embedded in paraffin and serially sectioned at 6 Am (across the

beam direction; parallel to the MRI slices). Every 50th section

(interval of 0.3 mm) was stained with hematoxylin and eosin

(H&E) for histologic examination. The sonication locations in the

MR images were correlated with the histology slides by measuring

lane across the focus after sonications at pressure amplitudes of 0.4, 0.5, 0.8,

the ultrasound beam of two of the locations.

Fig. 4. The average normalized signal intensity change measured in the T1-

weighted FSE images after sonication as a function of the time between the

sonication and the MRI contrast agent injection. This value was obtained

for each sonication location from the enhancement curves such as is

presented in Fig. 3 by averaging the signal intensity changes in the six to

eight images after the injection of the contrast agent. Thus, it is

approximately the maximum enhancement induced by the contrast agent

injection. Open circles: four animals from the threshold study (group 1 in

Table 1); closed circles: animals from the follow up study (group 2 in Table

1) (mean F SD). The pressure amplitude of these sonications was 0.8 MPa.

K. Hynynen et al. / NeuroImage 24 (2005) 12–20 15

their distance from anatomical structures and then identifying the

same structures in the histology.

Stains to detect apoptotic cells and ischemic neurons were used

on the neighboring sections of the focal plane. These stains were

employed because a possible vascular injury can induce a

microischemic impact and result in ischemic neuronal injury. The

apoptotic staining method utilized in this study was first described

by Gavrieli et al. (1992). TUNEL staining (ApoptTag kit, Intergen

Company, Purhase, NY) was used for the detection of DNA

fragmentation and apoptotic bodies in the cells. The sections were

counter stained with 0.5% Methyl Green.

To visualize ischemic neurons in the sonicated areas, vanadium

acid fuchsin (VAF) staining and toluidine blue counter staining

were employed (Victorov et al., 2000). The total number of

ischemic (acidophilic and hyperchromatic) and the total number of

normal neurons were counted in 10 nonoverlapping microscopic

fields (�1000) chosen in the regions exhibiting the most

pronounced morphological changes in each sonicated area. Some

ischemic neurons represent postmortem histological artifacts due to

the mechanical damage that may occur during removal and

dissection of the brain (Cammermeyer, 1978); hence, as a control,

the total number of ischemic and normal neurons was also counted

in 10 nonoverlapping microscopic fields outside of each sonicated

region close to its boundary. The author who performed the

histology evaluation was blinded to the ultrasound parameters but

had knowledge of the targeted locations.

Electron microscopy

Horseradish peroxidase (Molecular weight 40,000 Da and

diameter of 5 nm), used extensively as a marker for transmission

electron microscopy (Nag, 2003a), was selected to evaluate the

Fig. 3. The normalized signal intensity measured in FSE T1-weighed

images as a function of time after injection of contrast agent for a sonicated

location (solid circle) and a nonsonicated location in the contralateral

hemisphere (open circle) (mean F SD). The data are from one animal from

the second group in Table 1 sonicated at 0.8 MPa.

ultrastructural basis of BBB disruption. Four rabbits were

anesthetized as described above and given 1% promethazine (1

ml/kg ip) (Sigma) to block histamine release induced by horse-

radish peroxidase (HRP) administration. Trypan blue (1.5 ml/kg of

Fig. 5. Top: The normalized signal intensity change measured in FSE T1-

weighed images after injection of contrast agent as a function of the applied

pressure amplitude (mean F SD). Bottom: The percentage of sonicated

locations that demonstrated BBB disruption and tissue necrosis (bgrade 3Qeffects) as a function of the applied pressure amplitude.

K. Hynynen et al. / NeuroImage 24 (2005) 12–2016

2% solution in saline) was administered (through the ear vein) to

macroscopically indicate the degree of BBB disruption (Nag,

2003a). HRP type VI (Sigma) dissolved in saline (300 mg/kg) was

injected immediately after the trypan blue through the same vein in

two rabbits. In the other two animals, HRP type II (Sigma) was

administered in the same doses. Two to three locations in each

brain were then sonicated; the animals were sacrificed at

deliberately selected intervals of approximately 1, 5, 20, and 60

min following each sonication. The animals were euthanized

(EuthasolR, Delmara Laboratories Inc., Midlothian, VA, 1 ml iv),

and the brains were fixed either by perfusion through the aorta (one

animal) or by immersion fixation using 2.5% paraformaldehyde +

1.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.2). The two

methods were used to eliminate potential problems associated with

each method. Pieces of about 0.5 mm3 from the sonicated areas

(blue spots) and from nonsonicated areas (controls) were fixed for

2 h in the same fixative. Distance measurements from the MR

Fig. 6. Examples of the histology observations in VAF–toluidine sections. (A) Gra

extravasated erythrocytes were seen within the focal region. (C) Grade 2: microsco

of them associated with evident damage to the brain parenchyma (insert magnifica

lesions (infarct).

images were used to correlate the sample locations and the MR

image analysis. They were then washed in 0.1 M TRIS buffer (pH

7.4) and transferred to an incubation medium containing 3,3V-diaminobenzidine and hydrogen peroxide for 45 min at room

temperature. After postfixation in 2% osmium tetraoxide in PBS

for 2 h, the pieces were dehydrated in ethanol, passed through

propylene oxide, and embedded in Epon-Araldite. Ultrathin

sections unstained or stained with uranyl acetate and lead citrate

were observed with a JEM-1200EX electron microscope at 80 kV.

Results

MRI observations

The BBB opening was observed in the sonicated locations in

the T1-weighted contrast-enhanced scans (Fig. 2). Fig. 3 shows

de 0: no damage was observed in the sonicated location. (B) Grade 1: a few

pic areas of perivascular extravasations over the whole sonicated area, some

tion of an ischemic cell). (D) Grade 3: extensive extravasation; hemorrhagic

Table 3

The histology grading of the sonicated tissues

Grade Description

0 No detected damage

1 One to a few tiny red blood cell extravasations

2 Petechial hemorrhages; mild damage to the brain parenchyma

3 Hemorrhagic or nonhemorrhagic local lesions

Table 4

The number of ischemic and apoptotic cells in the sonicated areas

Pressure

amplitude

(MPa)

Number of

apoptotic cells

Number of

apoptotic

cells in

vessels

Number

of ischemic

cells

Number of

sonicated

locations

0 0.5 F 0.7 0.1 F 0.5 1.9 F 3.3 45

0.4 0.5 F 0.6 0.2 F 0.5 0.7 F 0.5 4

0.5 0.6 F 1.1 0.4 F 0.5 0.6 F 0.8 7

0.8 1.4 F 1.2 1.1 F 1.1 5.3 F 5.2 10

1.1 1.8 F 2.3 1.4 F 1.8 3.6 F 6.5 9

1.4 3.0 F 1.8 2.5 F 1.0 6.7 F 6.4 4

2.3 Many Many 14 F 16 4

3.1 Many Many 12 F 7.2 7

The number of cells was counted in 10 nonoverlapping microscopic fields

chosen in each sonicated area. These counts were averaged for each

pressure amplitude value (mean F SD shown). The pressure amplitude

value 0 MPa represents the average value for all control areas in all of the

brains. When the sonicated areas were compared with the control areas (t

test), the number of apoptotic cells was statistically higher ( P b 0.05) at

pressure amplitude values at or above 0.8 MPa. The number of cells on the

vessel walls was statistically higher than in the control locations in all

sonicated areas at or above 0.5 MPa. The number of ischemic cells was

statistically higher than the control locations at or above 0.8 MPa.

K. Hynynen et al. / NeuroImage 24 (2005) 12–20 17

the signal intensity in T1-weighted imaging in one sonicated

location and likewise in the same anatomical location on the

opposite side of the brain. Following sonication, the signal

intensity at the targeted location increased by 15–20% compared

to the control location. BBB disruption was not detected after

an additional MRI contrast injection 5 h following the

sonication. The signal intensity change was greatest immediately

after the sonications and decreased as a function of time (Fig.

4). Three hours following the sonications, the signal intensity

increase was only approximately 10–20% of that measured

initially.

The signal intensity enhancement due to BBB disruption

increased as a function of the pressure amplitude and was evident

at the lowest value tested (0.4 MPa) (Fig. 5, top). Sixty percent of

the locations had focal contrast enhancement greater than the

standard deviation of the signal in the normal brain at 0.4 MPa.

By 1.4 MPa, all locations showed evidence of BBB disruption

(Fig. 5, bottom).

Histology

Forty-eight locations from among 13 brains were histolog-

ically evaluated in H&E-stained serial sections. In many cases,

the sonicated areas differed little from the nonsonicated regions.

The most noticeable effect observed in other cases was the

areas of extravasation (42/48 locations), ranging from a few

scattered erythrocytes in the mildest cases to extensive

extravasation at the highest pressure amplitude levels (Fig. 6).

These extravasations were limited to the focal volume of the

ultrasound beam. These histological observations were divided

into four grades ranging from no damage (grade 0) to tissue

necrosis (grade 3) (Table 3). The degree of these vascular

effects increased as a function of pressure amplitude when

necrosis was induced at or above 2.3 MPa (Fig. 5B). At

pressure amplitudes below 2.3 MPa, the extravasations were

observed in the absence of obvious parenchymal damage. In the

mildest cases (grade 1), the effects were limited to one or a few

small erythrocyte extravasations, which was the dominant effect

at pressure amplitudes up to approximately 0.5 MPa. In more

severe cases (grade 2), there were multiple areas of erythrocyte

extravasations, light vacuolation of the neuropil adjacent to

some of affected microvessels and alteration in the shape and

stainability of a few neurons and astrocytes. Mild neutrophil

infiltration was seen in the affected regions 48 h following

sonication. Grade 2 damage dominated at exposure levels of

between 0.5 and 1.4 MPa.

In the VAF–toluidine-blue-stained sections with the largest

apparent effect, the entire sonicated area contained on average

five or fewer injured neurons when the pressure amplitude was

below 1.4 MPa. Most of these neurons appeared dark stained

(hyperchromatic) and thus might represent reversibly injured

neurons. At a pressure amplitude of 2.3 MPa, the number of

ischemic neurons increased (Table 4). For tissue effects less than

grade 3, the only cells that were positive for ischemia were

neurons.

TUNEL-stained sections showed a few apoptotic cells in both

in the vessel walls and at the brain parenchyma at pressure

amplitude values lower than 2.3 MPa (Fig. 7). With TUNEL

technique, using counterstaining, some cell types could be clearly

identified according to their size, shape, and specific locations as

endothelial cells, intravascular erythrocytes, or leukocytes. Overall,

the TUNEL technique showed very few positive-stained cells, and

these cells were primarily located in blood vessels and glia.

However, it is possible that some individual cells were small

neurons and could not be easily distinguished from glial cells and

extravascular leukocytes by light microscopy. The average number

of apoptotic cells increased with pressure amplitude. At pressure

amplitudes of 2.3 MPa and higher, the tissue damage was always

associated with multiple apoptotic cells. A few scattered apoptotic

cells were also found outside the sonicated areas (Table 4).

Electron microscopy

In the samples obtained from the nonsonicated areas, HRP

filled the vessel’s lumina in brains fixed by immersion, while in the

cases of the perfusion fixation, the vessels appeared to be empty.

HRP did not penetrate the interendothelial clefts. Vesicles

containing HRP were present in only a few endothelial cells

(ECs). No HRP was seen in the subendothelial space: the basement

membrane, pericytes, or neuropil.

In all of the sonicated areas, passage of HRP through the vessel

walls could be seen. Capillaries, arterioles, and venules were

involved in this process, which showed a dependence on the

circulation time of the tracer. For animals sacrificed approximately

1 min after sonication, the HRP reaction product was seen in small

vacuoles and vesicles located at the luminal front of the ECs. Some

of the vesicles were in the process of formation and taking HRP

from the lumen (Fig. 8). This process was more apparent in the

samples obtained from the immersion-fixed brains. No HRP was

Fig. 7. Examples of apoptotic cells after sonications at 1.4 MPa (A) and 3.1 MPa (B). In (A), only a single small cell, probably a glia cell (arrow), next to

intact neurons was seen to be undergoing apoptosis. In (B), many of the cells within this lesion appeared apoptotic (left: low magnification; right: high

magnification).

K. Hynynen et al. / NeuroImage 24 (2005) 12–2018

present in the interendothelial clefts, basement membrane, peri-

cytes, or neuropil (Fig. 8).

For animals sacrificed approximately 5 min after sonication,

HRP-positive reaction was seen in caveolae, vacuoles, and

lysosome-like profiles in the ECs cytoplasm. The tight junctions

in some of the vessels appeared to be open and permitted HRP to

Fig. 8. Photomicrograph showing a cross section of a capillary 1 min after

sonication in the presence of circulating HRP. HRP fills the lumen (L) but

is not present in the basement membrane (b), pericyte (P), or in the

neuropil (NP). Formation of a caveola (long arrow) incorporating HRP is

seen at the luminal surface of the endothelial cell, while the caveolae at the

abluminal front of the cell are free from HRP (arrowheads). An

interendothelial cleft shows the penetration of HRP into its luminal

opening, but the tracer seems to be stopped at the tight junction (short

arrows). RBC—red blood cell in the lumen; NEC—nucleus of the

endothelial cell (immersion fixation of the brain).

pass through the intercellular cleft, penetrating into the basement

membrane and the perivascular space.

In the samples acquired in animals sacrificed approximately 20

min after sonication, a positive reaction for HRP was observed in

the abluminal zones of the EC, in the basal membrane, and in the

interstitium of the neuropil (Figs. 9A and B).

For a sacrifice time of 1 h following sonication, the

distribution of the HRP was similar. In some sonicated locations,

mild to moderate perivascular and interstitial edema was found.

Small groups of extravasated red blood cells were seen in three of

eleven locations, and a few ruptures of the microvessel walls

were found.

Discussion

The results showed that BBB disruption is possible at a

frequency of 0.69 MHz with minimal damage to the exposed brain

parenchyma cells. The frequency used is suitable for highly

localized sonication through human skull as has been reported

earlier (Clement and Hynynen, 2002). This noninvasive method

may be useful in delivering therapeutic or diagnostic agents in the

central nervous system (Pardridge, 2002a).

The electron microscopy findings confirm the effective

disruption of the BBB by ultrasound under the conditions applied.

The cellular mechanisms of HRP passage through vessel wall

included both transendothelial (via caveolae and cytoplasmic

vacuolar structures) and paraendothelial (via intercellular clefts)

routes. These mechanisms are similar to those observed in other

conditions resulting in BBB disruption (such as hypertension,

injection of bradykinin, hyperosmotic agents, etc.) (Nag, 2003c).

The BBB disruption took place in capillaries, arterioles, and

venules, demonstrating similar mechanisms. The data obtained

show that blood-circulating macromolecules (at least with molec-

ular weight 40,000) invade the brain interstitial space and thus

demonstrate that many therapeutic and diagnostic agents may be

used in the brain.

Fig. 9. (A and B) Vessel and perivascular neuropil 19 min after sonication

in the presence of HRP (perfusion fixation of the brain). (A) Numerous

caveolae containing peroxidase (arrowheads) have moved to abluminal

front of the endothelial cell (EC) and into the pericyte (P). The tracer has

infiltrated the basement membrane (b) and the interstitium of the neuropil

(arrows). (B) The passage of HRP through interendothelial clefts with

evidently opened tight junctions is shown with arrows. Peroxidase has

reached the middle and abluminal part of the cleft and has penetrated the

basement membrane (b). Staining for HRP also is seen outside the vessel’s

wall in the neuropil (asterisks). L—lumen; E—red blood cell in the lumen;

NP—neuropil.

K. Hynynen et al. / NeuroImage 24 (2005) 12–20 19

The BBB disruption occurred in this study at a lower pressure

amplitude (0.4 MPa) than was found in the earlier study using the

same animal model and the same ultrasound contrast agent (0.7

MPa in that case) (Hynynen et al., 2001). This is consistent with

the threshold of inertial cavitation that have been shown to occur at

a lower pressure amplitude level when the frequency is decreased

(Apfel and Holland, 1991).

In this study, we used more sensitive histological techniques

than in the earlier study (Hynynen et al., 2001) to investigate the

role of apoptosis or ischemia following ultrasound disruption of the

BBB. In addition, the focal spot size was larger and increased the

likelihood of detecting any adverse events associated with the

sonications. Thus, these results increase the confidence that

ultrasound can be used to disrupt the BBB for delivering agents

to the brain. Although found relatively rarely (3 of 11 cases), the

endothelial damage observed with electron microscopy suggests

the need for further investigations of endothelial and nerve cell

ultrastructure. More work is needed to verify that there are no long-

term effects and that the exposed brain tissue remains fully

functional. Similarly, additional experiments must demonstrate that

an adequate amount of a specific agent can be delivered through

the BBB to achieve the desired therapeutic or diagnostic goal. In

addition, there can be a difference in sensitivity of the cell to

ultrasound exposure between white and gray matter. We did not

investigate this matter in the present study; all locations were

targeted in gray matter.

The parameters used in this study are suitable for noninvasive

sonications through human skull. The maximum acoustic power

over the 20-s duration averaged 80 mW distributed over the focal

spot area, which however was not strong enough to induce

measurable temperature elevation in the tissue (Hynynen et al.,

1997) and should therefore be safe for trans-skull sonications. At

pressure amplitudes up to approximately 0.5 MPa, the BBB was

disrupted in most sonicated locations with only a few extravasated

erythrocytes and zero to two ischemic neurons or apoptotic cells in

the entire area. These results suggest that these exposures can be

used to deliver therapeutic or imaging agents with an acceptable

impact on the brain tissue. At higher pressure amplitudes up to 1.4

MPa, there were more extravasated erythrocytes, and the total

number of the ischemic and apoptotic cells was increased on

average to five to six in the section. The importance of these effects

is not clear, though they appear minimal and would most likely be

acceptable for therapeutic interventions such as the delivery of

genes or chemotherapy agents for tumor or other serious disease

treatments.

In conclusion, this study has demonstrated that for the first

time, focused ultrasound can produce focal BBB disruption deep

within the brain at a frequency that can be focused through human

skull bone while sustaining little damage to the brain tissue. This

finding may have considerable impact on future brain research,

diagnostics, and therapy.

Acknowledgments

This research was supported by NIH research grant no.

CA76550 and CA089017 and a grant from CIMIT.

References

Abbott, N.J., Romero, I.A., 1996. Transporting therapeutics across the

blood–brain barrier. Mol. Med. Today 2, 106–113.

Apfel, R.E., Holland, C.K., 1991. Gauging the likelihood of cavitation from

short-pulse, low duty cycle diagnostic ultrasound. Ultrasound Med.

Biol. 17, 179–185.

Cammermeyer, J., 1978. Is the solitary dark neuron a manifestation of

postmortem trauma to the brain inadequately fixed by perfusion?

Histochemistry 56, 97–115.

Clement, G.T., Hynynen, K., 2002. A noninvasive method for fo-

cusing ultrasound through the human skull. Phys. Med. Biol. 47,

1219–1236.

Doolittle, N.D., Miner, M.E., Hall, W.A., Siegal, T., Jerome, E., Osztie, E.,

McAllister, L.D., Bubalo, J.S., Kraemer, D.F., Fortin, D., Nixon, R.,

Muldoon, L.L., Neuwelt, E.A., 2000 (Feb. 1). Safety and efficacy of a

multicenter study using intraarterial chemotherapy in conjunction with

osmotic opening of the blood–brain barrier for the treatment of patients

with malignant brain tumors. Cancer 88 (3), 637–647.

Gavrieli, Y., Sherman, Y., Ben-Sasson, S.A., 1992. Identification of

programmed cell death in situ via specific labeling of nuclear DNA

fragmentation. J. Cell Biol. 119, 493–501.

Hynynen, K., Jolesz, F.A., 1998. Demonstration of potential noninvasive

ultrasound brain therapy through intact skull. Ultrasound Med. Biol. 24,

275–283.

Hynynen, K., Vykhodtseva, N.I., Chung, A., Sorrentino, V., Colucci, V.,

Jolesz, F.A., 1997. Thermal effects of focused ultrasound on the brain:

Determination with MR Imaging. Radiology 204, 247–253.

Hynynen, K., McDannold, N., Vykhodtseva, N., Jolesz, F.A., 2001.

K. Hynynen et al. / NeuroImage 24 (2005) 12–2020

Noninvasive MR imaging-guided focal opening of the blood–brain

barrier in rabbits. Radiology 220, 640–646.

Kroll, R.A., Neuwelt, E.A., 1998. Outwitting the blood–brain barrier for

therapeutic purposes: osmotic opening and other means. Neurosurgery

42, 1083–1099.

Mesiwala, A.H., Farrell, L., Wenzel, H.J., Silbergeld, D.L., Crum, L.A.,

Winn, H.R., Mourad, P.D., 2002. High-intensity focused ultrasound

selectively disrupts the blood–brain barrier in vivo. Ultrasound Med.

Biol. 28, 389–400.

Nag, S., 2003a. Blood–brain barrier permeability using tracers and

immunohistochemistry. Methods Mol. Med. 89, 133–144.

Nag, S., 2003b. Morphology and molecular properties of cellular

components of normal cerebral vessels. Methods Mol. Med. 89, 3–36.

Nag, S., 2003c. Pathophysiology of blood–brain barrier breakdown.

Methods Mol. Med. 89, 97–119.

Pardridge, W.M., 2002a. Drug and gene delivery to the brain: the vascular

route. Neuron 36, 555–558.

Pardridge, W.M., 2002b. Drug and gene targeting to the brain with

molecular Trojan horses. Nat. Rev., Drug Discov. 1, 131–139.

Pardridge, W.M., 2003. Blood–brain barrier genomics and the use of

endogenous transporters to cause drug penetration into the brain. Curr.

Opin. Drug Discov. Dev. 6, 683–691.

Patrick, J.T., Nolting, M.N., Goss, S.A., Dines, K.A., Clendenon, J.L., Rea,

M.A., Heimburger, R.F., 1990. Ultrasound and the blood–brain barrier.

Adv. Exp. Med. Biol. 267, 369–381.

Sun, J., Hynynen, K., 1998. Focusing of ultrasound through a human skull:

a numerical study. J. Acoust. Soc. Am. 104, 1705–1715.

Thomas, J.-L., Fink, M.A., 1996. Ultrasonic beam focusing through

tissue inhomogeneities with a time reversal mirror: application to

transskull therapy. IEEE Trans. Ultrason. Ferroelectr. Freq. Control.

43, 1122–1129.

Victorov, I.V., Prass, K., Dirnagl, U., 2000. Improved selective, simple, and

contrast staining of acidophilic neurons with vanadium acid fuchsin.

Brain Res. Brain Res. Protoc. 5, 135–139.

Vykhodtseva, N.I., Hynynen, K., Damianou, C., 1995. Histologic

effects of high intensity pulsed ultrasound exposure with sub-

harmonic emission in rabbit brain in vivo. Ultrasound Med. Biol.

21, 969–979.