Focused Ultrasound-Induced Blood Brain-Barrier Opening

Enhanced Vascular Permeability for GDNF Delivery in Huntington’s

Disease Mouse Model

Chung-Yin Lin,1,2 Chih-Hung Tsai,3 Li-Ying Feng,4 Wen-Yen Chai,3 Chia-Jung Lin,3

Chiung-Yin Huang,4 Kuo-Chen Wei,4 Chih-Kuang Yeh,5 Chiung-Mei Chen,6* and

Hao-Li Liu1,3,4*

1Medical Imaging Research Center, Institute for Radiological Research, Chang Gung

University/Chang Gung Memorial Hospital, Taoyuan 333, Taiwan2Department of Nephrology, Division of Clinical Toxicology, Chang Gung Memorial

Hospital, Lin-Kou Medical Center, Taoyuan 333, Taiwan3Department of Electrical Engineering, Chang Gung University, Taoyuan 333, Taiwan4Department of Neurosurgery, Chang Gung Memorial Hospital, Linkou Medical

Center and College of Medicine, Chang Gung University, Taoyuan 333, Taiwan5Department of Biomedical Engineering and Environmental Sciences, National Tsing

Hua University, Hsinchu 300, Taiwan6Department of Neurology, Chang Gung Memorial Hospital, College of Medicine,

Chang Gung University, Taoyuan 333, Taiwan

*Corresponding authors:

Chiung-Mei Chen, Neurology, Chang Gung Memorial Hospital, Taoyuan, Taiwan

333, Tel.:+886-3-328-1200 ext. 8729, email: cmchen@cgmh.org.tw; Hao-Li Liu,

Department of Electrical Engineering, Chang Gung University, Taoyuan, Taiwan 333,

Tel.:+886-3-211-8800 ext. 3778 (Office), email: haoliliu@mail.cgu.edu.tw.

1

Supplementary Data

1. Supplementary Materials and Methods

1.1. Animal model

HD transgenic mice on a R6/2 (B6CBA-Tg[HDexon1]62Gpb/1J) background were

purchased from the Jackson Laboratory under approval from the Institutional Animal

Care and Use Committee of Chang Gung University, and handled according to the

guidelines in The Handbook of the Laboratory Animal Center. For mating, male R6/2

mice were crossed with female control mice (B6CBAFI/J) to produce offspring,

which were then identified by genotyping of tail DNA. PCR genotyping was

performed using the following primers: 5’-CCG CTC AGG TTC TGC TTT TA-3’

and 5’-GGC TGA GGA AGC TGA GGA G-3’. Male R6/2 mice and the wild-type

littermates were used for investigation. R6/2 HD mice become increasingly

symptomatic from 6-8 weeks with a short lifespan of 18-22 weeks. All animals were

housed at Chang Gung University under a regular 12/12 h light/dark circadian cycle.

Body weights and blood sugar of the mice were recorded weekly.

1.2. Magnetic resonance imaging (MRI) equipment

Scans were performed using a 7-Tesla magnetic resonance scanner (Bruker ClinScan,

Germany) with a 4-channel surface coil for monitoring. Regions-of-interest (ROI)

2

consisted of: cortex, ventricle, and striatum area. ROIs were delineated to the

experimental groupings, and intra- and inter-rater reliability was consistently at a

confidence level ≥ 95%. Volumetric data were calculated and processed semi-

automatically using ImageJ (National Institutes of Health, USA).

1.3. Focused ultrasound (FUS) system

A single-element FUS transducer with frequency 500 kHz, diameter 64 mm,

curvature radius 63.2 mm, and electric-to-acoustic efficiency > 80% (Imasonics SAS,

France) was placed in an acrylic water tank filled with distilled and degassed water.

The focus of the ultrasonic field was used to guide the ultrasound beam to the desired

region. A function generator (33120A, Agilent, Palo Alto, CA) was connected to a

power amplifier (150A100B, Amplifier Research, Souderton, PA) and a power

meter/sensor module (Model 4421, Bird Electronics Corp., Cleveland, OH) to drive

the FUS transducer. The animal was placed directly under a 44 cm2 window of thin

polymer film at the bottom of the acrylic tank, with an acoustic connection established

using acoustic transmission gel (Pharmaceutical Innovations, Newark, NJ). The

output acoustic pressure at the focal point was 0.33 MPa, measured via a calibrated

polyvinylidene-difluoride-type (PVDF) hydrophone (Model HNP-0400, ONDA,

Sunnyvale, CA, USA). Ultrasound microbubbles (SonoVue; phospholipid-coated

3

microbubbles, mean diameter 2.5 m, and microbubble concentration: 2-5108

bubbles/mL) were administered intravenously at a dose of 0.1 mg/kg prior to FUS

exposure.

1.4. Preparation of plasmid DNA-encapsulating liposomes (pDNA-LPs)

GDNF plasmid DNA (GDNFp) was prepared according to a previously published

protocol [1]. In addition, two plasmids pcDNA3.1-Htt-(Q)25-hrGFP and pcDNA3.1-

Htt-(Q)109-hrGFP (a gift from Dr. Chern Y) [2] were used in the transient

transfection experiments. Loading of GDNF, pcDNA3.1-Htt-(Q)25-hrGFP and

pcDNA3.1-Htt-(Q)109-hrGFP into liposomes (LPs) was prepared according to the

protocol as GDNFp-LPs, L-Htt25Q, and L-Htt109Q. Briefly, the same liposome

suspension described above was then sonicated and extruded through 200-nm

polycarbonate filters using an Avanti Mini Extruder (Alabaster, AL), and then passed

through a spin chromatography column (GE Healthcare, IL) to remove the

unencapsulated pDNA. The pDNA concentration was measured using a Nanodrop

(ND-1000, Thermo Fisher Scientific Inc. Waltham, MA) at wavelengths of 260- and

280-nm. Encapsulation efficiency was calculated as the fraction of pDNA

incorporated into the LPs vesicles by spectrophotometer (Hitachi F-7000, Tokyo,

Japan) at a wavelength of 260 nm after filtration. The mean particle size and zeta

4

potential of pDNA and pDNA-LPs were respectively analyzed by dynamic light-

scattering (DLS) and zeta potential on a Nano-ZS90 particle analyzer (Malvern

Instruments, Malvern, Worcestershire, UK). The samples were prepared and imaged

by cryogenic-transmission electron microscopy (Cryo-TEM).

1.5. In vitro mutation induction by ultrasound sonication.

To examine the chromosome aberrations caused and gene transfer efficiencies by the

effect of ultrasound (US), Neuro2a (N2A) (3105 cells/well) were used to test in

vitro sonotransfection. The cells were cultured in the MEM medium media

supplemented with 10% fetal bovine serum and 1% pyruvate solution in a 37C, 5%

CO2 incubator for 48 h with complete cell attachment. L-HttQ25 or L-HttQ109 was

added to the wells in a pDNA concentration of 5 µg/mL aliquots. To study the

influence of the pDNA-encapsulating LPs with respect to the mutation induction, six

experimental groups were composed of the control (no plasmids, no US), US only, L-

HttQ25, L-HttQ25 with US, L-HttQ109, and L-HttQ109 with US. The cells were then

incubated at 37°C for 2 h before application of ultrasound sonication. A 20-kHz

ultrasonic probe (VCX750, Sonics and Materials INC, Newtown, CT) with a diameter

of 3 mm was inserted into culture wells with the tip approximately 1.5 cm above the

cell monolayer. The power generator was set to an intensity of about 1 W/cm2. US

5

was separately applied for 2 seconds for the groups of the US only and the pDNA-LPs

with US, respectively. After sonication, the media were removed from all the wells,

and the cells were washed with MEM three times to remove any un-incorporated

plasmids or empty vectors. Cells were then grown for 48 h after pDNA was delivered.

After rinsing with PBS, the cells were incubated with anti-1C2 antibody (1:1000;

Millipore, Temecula, CA) to determine whether the expanded polyQ had been

expressed and aggregated after plasmid transfection with US sonication. These cells

were then observed via a Leica TCS SP8X confocal microscope (Leica

Microsystems, Wetzlar, Germany) to visualize the intracellular polyQ expression and

aggregates in N2A cells, using an argon laser with a wavelength of 594 nm.

1.6. Imaging neuroanatomy

In vivo images of mouse brains demonstrated progressive regional atrophy of WT

normal and R6/2 HD mice. WT and HD male mice (6-week-old at the beginning of

the experiments, eight mice in each group) were subjected to weekly MRI scanning

until they were sacrificed. An un-biased whole-brain comparison of WT and R6/2 HD

mice at each imaging time-point was performed using an automated image processing

pipeline. Anatomical images were scanned by performing turbo spin-echo (TSE)

sequences to acquire T2-weighted images with the following imaging parameters:

6

pulse repetition time (TR)/echo time (TE) = 3000 ms/29 ms; FOV = 19.2 19.2

mm2; in-plane resolution = 320 320 pixels; slice thickness = 0.6 mm. Animals were

anesthetized with 2% vaporized isoflurane administered via a facemask. Collectively,

a set of three images was selected for each mouse at 6, 12, and 18-week-old in vivo

and the data were analyzed for the comparison of WT mice versus R6/2 HD mice at

each time point.

1.7. FUS-induced blood-brain barrier (BBB) opening

Induction of FUS-BBB opening was verified via MRI and HE staining without

histological damage. FUS-BBB opening was applied at the brain striatum (ST), with a

pressure of 0.33 MPa. Scanning was performed before and after FUS sonication using

a gradient echo FLASH sequence to acquire T1W1 images with the following

imaging parameters: pulse repetition time (TR)/echo time (TE) = 230 ms/3.51 ms;

FOV = 19.2 × 30.72 mm2; in-plane resolution = 160 × 256 pixels; slice thickness = 0.6

mm; flip angle = 70°. We applied pulsed FUS sonications with a 10-ms burst length, a

1% duty cycle, a 1-Hz pulse repetition frequency (RPF), and a 30-sec insonation at

the contralateral hemisphere and another 60-sec insonation at the ipsilateral

hemisphere. If not specified, the FUS exposure groups were followed with additional

FUS sonication at the ipsilateral hemisphere about 2 min later. Following FUS

7

sonication, an intravenous bolus (0.012 mmol/kg) of Gd-DTPA (gadolinium-

diethylenetriamine penta-acetic acid) MRI contrast agent (Magnevist, Berlex

Laboratories, Wayne, NJ, USA; molecular weight: 938 Da) was administered to

identify the BBB opening. Mice were sacrificed after experiments. HE staining was

used to assess the resulting histological damage in both the contralateral and

ipsilateral sonicated brain tissues.

1.8. Determination of FUS-BBB opening with TEM

Brain tissues were fixed with PBS containing 4% paraformaldehyde and 0.2%

glutaraldehyde for 8 h. After rinsing in PBS, sections were fixed again in 2%

glutaraldehyde in PBS for 1 h, osmicated in 1% OsO4 in PBS, and stained overnight

in 2% aqueous uranyl acetate. Sections used for electron microscopy were dehydrated

in ascending concentrations of ethanol solution. Ultrathin sections (60 nm) using a

vibratome were sectioned and examined using a Hitachi H-7500 electron microscope

to scan their morphologies.

1.9. Spectrophotometric quantification of EB dye

To study the effect of FUS-BBB opening on the vascular permeability, WT and R6/2

HD mice were collected in three age groups: 6, 12, and 18-week-old (n 15 in each

8

group). Mice of different ages were divided into four experimental groups: WT, R6/2

HD, WT with FUS exposure, and R6/2 HD with FUS exposure. MBs for both of the

WT+FUS and the HD+FUS groups, while the two other groups were treated with

neither FUS sonication nor MBs. The molecular size of EB is 960 Da in its free form

and 67 kDa when conjugated with albumin in plasma after IV administration. EB dye

(30 mg/kg in PBS) was IV injected into the tail vein and animals were sacrificed 2 h

later. Both brain hemispheres were weighed and placed in formamide (1 ml/100 mg)

at 60°C for 24 h. The samples were centrifuged for 20 min at 14,000 rpm. The

concentration of EB dye extracted from each brain was determined with

spectrophotometer at 620 nm by comparison with a linear regression standard curve

derived from seven concentrations of EB dye in 0.9% PBS solution.

1.10. Kinetic analysis of permeability changes

Data sets were acquired for 5 min. An IV bolus of Gd-DTPA was administered as an

MRI contrast agent for dynamic acquisition with an infusion rate of 0.24 mL/min, and

0.7 mL/kg of Gd-DTPA mixed with 1 mL of saline and heparin via tail vein injection.

Dynamic BBB permeability changes were measured by repeating the DCE (dynamic

contrast-enhanced) image sequence with or without FUS-BBB opening. After DCE

T1-weighted image acquisition, turbo-spin-echo (TSE) T2-weighted images were

9

acquired to confirm the absence of anatomical changes caused by sonication. The

following parameters were used: TR/TE = 2.38/0.8 ms; FOV = 25 × 31 mm2; in-plane

resolution = 0.16 × 0.16 mm2; and slice thickness = 0.6 mm. Susceptibility-weighted

imaging (SWI) sequences are known to be more sensitive than standard T2-weighted

images for detecting hemorrhages and were acquired to identify possible tissue

hemorrhage associated with FUS-BBB opening using the following parameters:

TR/TE = 30 ms/18 ms; flip angle = 40°; slice thickness = 0.6 mm; matrix size = 320 ×

512; and FOV= 26 × 41 mm2 [3, 4].

1.11. FUS-BBB opening for therapeutic efficacy

To study the effect of R6/2 HD induction on the neuroprotective response in the FUS-

BBB opening in the contralateral and ipsilateral rostral, four experimental groups

were composed of sham (no GDNFp-LPs, no FUS exposure and no MBs), GDNFp-

LPs only (pDNA at 60 μg), FUS exposure only (no GDNFp-LPs, no MBs), and

GDNFp-LPs (pDNA at 60 μg) injection followed by MBs injection with FUS

exposure. The experiment was run once weekly on each HD group from 6 to 14-

weeks of age. If not otherwise specified, the GDNFp-LPs were injected intravenously

through the tail vein. The MBs or the GDNFp-LPs were injected and FUS exposure

was immediately applied to the brains to open the BBB, resulting in enhanced

10

delivery of GDNFp-LPs across the BBB. Brain atrophy development was also

checked weekly to further study the therapeutic response via MRI scanning. ROIs

were delineated and obtained for the cortex, the ventricle, and the striatum area as

previous described.

1.12. Behavioral test

Body weight and general health condition were monitored. All motor skill learning

occurred during the light phase of the 12-h light/dark cycle. Motor performance was

assessed with a rotarod apparatus (RT-01, Singa Technology Corp., Taiwan) with

mouse walking on a rolling rod with a constant speed of 15 rpm. Each mouse was

tested for repeating three trials with a 5 min resting interval between two trials. The

mean latency to fall off the rotarod was recorded.

1.13. Western Blotting

Fresh frozen brain tissues from the WT and R6/2 HD groups were homogenized in

PRO-PREPTM protein extraction solution (iNtRON Biotechnology Inc., Summit, NJ),

kept overnight, and then centrifuged at 10,000×g for 30 min. After centrifugation to

remove tissue debris, the protein contents of the tissue homogenates were determined

according to the protein quantification. The supernatant with 25 μg protein was

11

dissolved in a sample buffer, separated by 10 % SDS-PAGE, and then transferred to a

polyvinylidene fluoride (PVDF) membrane. The PVDF membrane was further

incubated with a primary mouse anti-GDNF antibody (OriGene Technologies, Inc.,

Rockvile, MD) at a 1:500 dilution and secondary rabbit anti-mouse antibody

(Molecular Probes Inc., Grand Island, NY) at a 1:1000 dilution. To measure the

optical density of the positive bands, the film was scanned using the BioSpectrum

Imaging System (UVP LLC, Upland, CA).

1.14. Immunohistochemistry (IHC) staining

Brain sections for each antibody were placed in (pH 7.4) TRIS Buffered Saline

(TBS), and washed twice for 5 min prior IHC. Sections for EM48 (anti-Huntingtin

protein antibody), cleaved caspase 3, P-CREB (phosphorylated cyclic AMP-

responsive element binding protein), and MDA (Malondialdehyde) was performed

using an UltraVision Quanto detection system (ThermoFisher Sci., Waltham, MA)

following the manufacturer's instructions. Endogenous peroxidase activity was

inhibited by incubation in an UltraVision hydrogen peroxide block for 10 min. Non-

specific binding sites were blocked with UltraVision protein block for 5 min, and the

sections were incubated with EM48 (dilution 1:500, Chemicon Int., Temecula, CA),

cleaved caspase 3 (dilution 1:500, BD Biosci., San Jose, CA), P-CREB (dilution

12

1:200, Abcam, Cambridge, MA), and MDA (dilution 1:100, MyBioSource Inc, San

Diego, CA) antibodies overnight at 4 C. After several washes in PBS, the sections

were incubated with the Primary Antibody Amplifier Quanto (ThermoFisher Sci.,

Waltham, MA) for 10 min, and the sections were added with the secondary antibody

(HRP Polymer Quanto) for 10 min. Finally, the DAB chromogen was added to the

sections for 5 min, and counterstained with Mayers’ hematoxylin (DAKO), then the

slides were mounted with mounting medium (Vector Laboratories, Burlingame, CA).

The slides were then photographed with a phase-contrast microscope (TissueFAX

Plus, TissueGnostics, Austria). The method for EM48 staining was adapted from in

house-procedures and those previously described [5] for visualization of Htt

aggregation.

1.15. TUNEL assay

TUNEL assay was performed to examine neuronal apoptosis using a commercial kit

according to the instructions provided by the manufacturer (R & D Systems,

Minneapolis, MN). TUNEL-positive neurons with condensate nuclear were identified

as apoptotic neurons. Cell counting was performed on three selected area of cortex or

striatum region per slide and the mean of the three counts was used for each mouse.

13

1.16. Immunofluorescence (IF) staining

Brain tissues were prepared and sectioned according to standard procedures to

examine neuronal morphology and function. To identify the effect of therapeutic

response on neuronal function, tissue sections were stained overnight at 4C with the

following primary antibodies: anti--tubulin (dilution 1:500, Abcam, Cambridge,

MA), anti-GFAP (dilution 1:500, Abcam, Cambridge, MA), anti-IBA1 (dilution

1:500, OriGene Technologies, Inc., Rockvile, MD), and anti-MAP2 (dilution 1:1000,

Santa Cruz Biotechnology, Inc., Dallas, TX). After rinsing in PBS, the sections were

incubated in secondary antibody with goat anti-rabbit fluorescence 594 or donkey

anti-mouse fluorescence 594 (1:1000, for GFAP or IBA1) or with rabbit anti-mouse

fluorescence 488 (1:1000, for GFAP or IBA1) for 1 hr at room temperature. After

rinsing in PBS, coverslips were applied on slides with anti-fade reagent with the

nuclear marker DAPI (Molecular Probes Inc., Grand Island, NY). The sections

were then imaged using a Leica TCS SP8X confocal microscope (Leica

Microsystems, Wetzlar, Germany).

1.17. Statistical analysis

All the data are presented as mean standard deviation (SD). Statistical analysis was

performed on a personal computer using SPSS version 16.0 (SPSS Inc., Chicago, IL)

14

statistical software. Statistical differences were assessed using ANOVA combined

with post-hoc Tukey test or Mann-Whitney U test where appropriate. P < 0.05 was

considered to indicate a statistically significant difference.

2. Supplementary Results

2.1. Brain atrophy in R6/2 HD mice

Figure S1 shows R6/2 HD mice developed a progressive neurological phenotype.

Significant brain atrophy was shown by MR images. Volumetric changes of WT and

R6/2 HD mice in individual brain sections were identified and then quantitated via

T2-weighted MRI, including the cortex, ventricle, and striatum via individual image

slide processing and voxel accumulation (Fig. S1A, yellow: cortex; red: ventricle;

green: striatum). Significant increase in the ventricle was noted in the HD brain

starting from 6-weeks and persisted up to 20-weeks age. HD mice had approximately

a 30% increase in ventricular volume at 12-weeks age and about 130% increase at 20-

weeks when compared to the ventricles of WT controls. Similarly, cortex volume in

HD mice demonstrated approximately a 9.1% decrease (*P < 0.05) at 12-weeks and

20% decrease at 20-weeks (**P < 0.01), while striatum volume showed a 2.9%

decrease (*P < 0.05) at 12-weeks and 6.6% decrease at 20-weeks (**P < 0.01).

Apparent motor disability in HD animals was observed at 12-weeks (21.1% lower)

15

and continued worsening to 20-weeks, a pattern significantly different to the WT

controls (95.3% lower, ***P < 0.005). In addition, the weight of HD animals was

significantly lower than that of WT controls and continued to gradually worsen from

16 to 20-weeks (27.5% lighter, **P < 0.01).

Htt protein aggregates are a hallmark of HD and may play an important role in disease

progression. We therefore examined the numbers of polyQ aggregates in the cortex

and striatum from WT and R6/2 HD mice after immunostaining with EM48 antibody.

As shown in Fig. S1B, a significant increase in polyQ aggregates was observed in

R6/2 HD mice at 6-week-old (cortex: 26.0 7.5 versus 0.7 1.2 in WT; striatum:

17.0 5.6 versus 0.7 1.4, *P < 0.05), 12-week-old (cortex: 34.3 6.1 versus 0.3

0.6 in WT; striatum: 35.0 7.0 versus 1.0 1.0, *P < 0.05), and 20-week-old

(cortex: 38.3 6.1 versus 0.3 0.6 in WT; striatum: 52.7 5.0 versus 0.3 0.6, *P

< 0.05). At 20-weeks, the intranuclear aggregates in cortex and striatum significantly

increased than those at 6 or 12-weeks of age. These observations confirmed the

pathological change of the employed R6/2 HD animal can well mimic HD disease

progression.

16

Figure S1 (A) Representative MR images of volumetric changes in the cortex,

ventricle, and striatum; motor coordination; weight between WT and R6/2 HD mice

from 6- to 20-week-old. (B) Age-dependent accumulation of intraneuronal huntingtin

aggregates using EM48 antibody in both the cerebral cortex and striatum area of R6/2

HD mice (magnification, 400). R6/2 HD mice showed significantly increased

nuclear aggregates in the cortex and the striatum from 6- to 20-week-old. Error bars

indicate SD, n = 6; *, P < 0.05; **, P < 0.01; ***, P < 0.005 (Mann-Whitney U test).

2.2. FUS-gene delivery improved brain atrophy in HD transgenic mice

Here, we demonstrate the neuroprotective effect via FUS-BBB opening to facilitate

CNS GDNF-gene (GDNFp-LPs) delivery into the R6/2 HD mice. The mice were

grouped into sham, GDNFp-LPs only, FUS only, and GDNFp-LPs + FUS. Figure S2

shows R6/2 HD mice developed neuroanatomical changes in various experimental

17

conditions, with the volumetric change. Significant differences in cortex volume

preserving and striatum volume reduction were noted in the GDNFp-LPs + FUS

treated group when compared to the sham group. However, these improvements tend

to decrease after 18-weeks, related to end of treatment at 14-weeks.

Figure S2 Volumetric changes in the cortex, ventricle, and striatum; weight in treated

R6/2 HD mice. ANOVA combined with post-hoc Tukey test was performed for

comparisons. Compared with the HD sham group: single pound, P < 0.05. Values are

presented as means ± SD. N 6 mice/group.

Supplementary References

18

[1] Lin CY, Hsieh HY, Pitt WG, Huang CY, Tseng IC, Yeh CK, et al. Focused

ultrasound-induced blood-brain barrier opening for non-viral, non-invasive, and

targeted gene delivery. J Control Release 2015;212:1-9.

[2] Chiang MC, Huang CL, Chern Y. cAMP-response element-binding protein

contributes to suppression of the A2A adenosine receptor promoter by mutant

Huntingtin with expanded polyglutamine residues. J Biol Chem 2005;280:14331-

40.

[3] Chai WY, Chu PC, Tsai MY, Lin YC, Wang JJ, Wei KC, et al. Magnetic-

resonance imaging for kinetic analysis of permeability changes during focused

ultrasound-induced blood-brain barrier opening and brain drug delivery. J Control

Release 2014;192:1-9.

[4] Chu PC, Chai WY, Hsieh HY, Wang JJ, Wey SP, Huang CY, et al.

Pharmacodynamic analysis of magnetic resonance imaging-monitored focused

ultrasound-induced blood-brain barrier opening for drug delivery to brain tumors.

Biomed Res Int 2013;2013:627496.

[5] Li H, Li SH, Cheng AL. Mangiarini L, Bates GP, Li XJ. Ultrastructural

localization and progressive formation of neuropil aggregates in Huntington’s

disease transgenic mice. Hum Mol Gen 1999;8:1227-36.

19