inhibition of hepatic fibrosis with artificial microrna...

ORIGINAL ARTICLE

Inhibition of hepatic fibrosis with artificial microRNA usingultrasound and cationic liposome-bearing microbubblesD Yang1,2,4, Y-H Gao1,4, K-B Tan1, Z-X Zuo1, W-X Yang1, X Hua3, P-J Li1, Y Zhang1 and G Wang1

We sought to investigate the antifibrotic effects of an artificial microRNA (miRNA) targeting connective tissue growth factor (CTGF)using the ultrasound-targeted cationic liposome-bearing microbubble destruction gene delivery system. Cationic liposomes wereconjugated with microbubbles using a biotinavidin system. Plasmids carrying the most effective artificial miRNA sequences weredelivered by ultrasound-targeted cationic liposome-bearing microbubble destruction gene delivery system to rats with hepaticfibrosis. The results show that this method of gene delivery effectively transported the plasmids to the rat liver. The artificial miRNAreduced hepatic fibrosis pathological alterations as well as the protein and mRNA expressions of CTGF and transforming growthfactor b1. Furthermore, the CTGF gene silencing decreased the levels of type I collagen and a-smooth muscle actin (Po0.01). Thesedata suggest that delivery of an artificial miRNA targeted against CTGF using ultrasound-targeted cationic liposome-bearingmicrobubble destruction may be an efficacious therapeutic method to ameliorate hepatic fibrosis.

Gene Therapy (2013) 20, 11401148; doi:10.1038/gt.2013.41; published online 22 August 2013

Keywords: artificial microRNA; connective tissue growth factor; gene delivery; hepatic fibrosis; ultrasound; microbubble

INTRODUCTIONHepatic fibrosis is a common pathological change in the chronicliver injury characterised by the accumulation of extracellularmatrix (ECM).1 Following liver injury, hepatic stellate cells (HSCs)become activated and transform into myofibroblast-like cells thatexpress alpha-smooth muscle actin (a-SMA) and produce ECM.2,3

Transforming growth factor b1 (TGF-b1) has been considered oneof the most potent fibrogenic mediators to fibrosis.4 Because of itsmultiple actions, the complete blockage of TGF-b1 may haveserious side effects. Connective tissue growth factor (CTGF) hasreceived significant attention as a downstream effector of theTGF-b1 fibrogenic effect.5,6 CTGF not only can lead to theactivation, proliferation and migration of primary HSCs, but canalso promote the accumulation of ECM components and inhibittheir degradation.7 Disruption of CTGF expression has beenshown to effectively suppress the activation of HSCs and theaccumulation of ECM.8 CTGF expression is low or absent in thenormal liver tissue; however, its expression progressively increasesin the fibrotic liver. CTGF expression is significantly correlated withthe development of hepatic fibrosis.9 Strategies to block CTGF bysmall interfering RNAs have achieved favourable anti-fibroticeffects for treating hepatic fibrosis.10,11

On the basis of recent studies, microRNAs (miRNAs) have gainedsignificant attention for the treatment of hepatic fibrosis.12,13

Artificial miRNAs (amiRNAs) exploit the backbone of natural miRNAsto generate designed miRNAs that can efficiently silence the geneof interest.14 This strategy has produced highly efficient and specificgene silencing for therapeutic applications.15,16 However, thechallenge is to effectively deliver amiRNAs to the target organ.

Ultrasound-targeted microbubble destruction (UTMD) hasemerged as a novel non-viral gene delivery technique because

of its safety, high efficiency and local gene transfer.17,18 Thismethod involves the attachment of genes to microbubbles, whichare then injected and circulated through blood vessels anddestroyed at the target site by ultrasound insonation.19 Thedestruction increases capillary permeability and generates tran-sient holes in the cell membrane and releases the payload, whichis incorporated intracellularly. Microbubbles, which consist of agas core and a stabilizing shell. The size of microbubbles is usuallyless than 10 mm. Nevertheless, the gene-loading capacity of themicrobubbles remains a consideration. Fortunately, studies haveshown that the loading capacity can be increased by attachingcationic liposomes to the lipid shell of microbubbles.20,21

In this study, we utilised novel cationic liposome-bearing micro-bubbles combined with ultrasound to transfect a plasmid-basedamiRNA designed against CTGF mRNA, which showed high gene-silencing efficacy in vitro, to determine whether this techniquecould efficiently suppress the expression of CTGF and exertantifibrogenic effects on hepatic fibrosis in vivo.

RESULTSThe effect of artificial miRNA targeted against CTGF on CTGFexpression in vitroWestern blotting showed that there was no obvious decline inCTGF protein expression in the scrambled amiRNA group(amiScramble), compared with the untreated HSC-T6 cells group(blank). Cells treated with four different artificial miRNAs targetedagainst CTGF showed varying degrees of inhibition, and CTGF-4exerted the strongest effect. The same effects were observed onCTGF mRNA expression levels (Figure 1). These data suggest thatthe CTGF-4 vector provided the most effective inhibition of CTGF

1Department of Ultrasound, Xinqiao Hospital, The Third Military Medical University, Chongqing, China; 2Department of Ultrasound, 324th Military Hospital, Chongqing, China and3Department of Ultrasound, Southwest Hospital, The Third Military Medical University, Chongqing, China. Correspondence: Professor K-B Tan, Department of Ultrasound, XinqiaoHospital, The Third Military Medical University, No.183 Xinqiao Street, Chongqing 400037, China.E-mail: [email protected] authors contributed equally to this work.Received 6 November 2012; revised 10 June 2013; accepted 12 June 2013; published online 22 August 2013

Gene Therapy (2013) 20, 11401148& 2013 Macmillan Publishers Limited All rights reserved 0969-7128/13

www.nature.com/gt

http://dx.doi.org/10.1038/gt.2013.41mailto:[email protected]://www.nature.com/gtJoseph GeorgeRectangle

expression in HSC-T6 cells. Thus, the CTGF-4 vector was selectedfor subsequent experiments.

The characterisation of cationic liposomesThe average particle size and the zeta potential (z) was 120 nmand 50 Mv, respectively, for the biotinylated cationic liposomesand 200 nm and 20 Mv, respectively, for the DNA-loadingliposomes. Similar size distributions of cationic liposomes werefound in transmission electron microscope image shown inFigure 2a. The biotinylated cationic liposomes appeared asspherical or nearly spherical nanoparticles. 3b-[N-(N,N-dimethy-laminoethane)-carbamoyl]-cholesterol (DC-Chol) is a chargedlipid that gives the liposomes a positive charge that can efficientlycombine with a negatively charged DNA plasmid. A gelretardation assay showed that the biotinylated cationicliposomes/DNA volume ratio required to completely retain DNAin the complex was 6:1 (Figure 2b).

Properties of cationic liposome-bearing microbubblesThe concentration and size distribution of microbubbles carryingDNA-loading liposomes were determined to be 15 109 ml 1and 17 mm, respectively, which allowed them to circulatedthrough the microvessels and capillaries. As shown in Figure 3a,

the surface of microbubble carrying DNA-loading liposomesappeared rough and was surrounded by a number of DNA-loading liposomes with red fluorescence.

The cationic liposome-bearing microbubbles exhibited acousticactivity. Similar to previous observations of common microbub-bles,22 the echo of the liver was low without the presence ofcationic liposome-bearing microbubbles. The echo intensity ofthe liver significantly increased after the injection of cationicliposome-bearing microbubbles. Furthermore, this effect wasclearly abated after the cationic liposome-bearing microbubbleswere destroyed by ultrasound (Figure 3bd).

Delivery of amiRNA to the rat liver by cationic liposome-bearingmicrobubbles combined with ultrasoundTo explore whether the amiRNA expression plasmid is deliveredinto the rat liver by ultrasound-targeted novel cationic liposome-bearing microbubbles destruction (UTMD) in vivo, we transferredit, including emerald green fluorescent protein (EmGFP) gene, as areporter for the tracking of amiRNA-mediated expression. Noexpression was observed in the kidney, lung or spleen (data notshown). Results from fluorescence microscope (Figure 4af) andquantified fluorescence intensity (Figure 4g) indicated that nofluorescence was observed in the untreated normal liver tissue,

Figure 1. The effect of amiRNA targeted against CTGF on CTGF expression in HSC-T6 cells. HSC-T6 cells were treated with media (blank),scrambled amiRNA (amiScramble) or four artificial microRNA (miRNA) that targeted CTGF (CTGF-1, CTGF-2, CTGF-3 and CTGF-4). (a) Theexpression of CTGF protein was determined by western blot. (b) The expression of CTGF protein was normalised to the level of GAPDH usingdensitometric scanning. Changes in CTGF protein expression were expressed as a percentage of the blank group. (c) The ratio of CTGF mRNAexpression to GAPDH mRNA was detected by quantitative real-time PCR. Changes in mRNA expression were expressed as a percentage of theblank group. Data are presented as the means.d. (n 6). *Po0.01 when compared with the blank control group.

Figure 2. The characterisation of the cationic liposomes. (a) A transmission electron microscope image of the biotinylated cationic liposomes.They appeared as spherical or nearly spherical nanoparticles. (b) A gel retardation assay was used to detect the biotinylated cationicliposomes with the proportion of plasmid. Lanes 18 contained a cationic liposome/DNA volume ratio of 0:1, 1:1, 2:1, 3:1, 4:1, 5:1, 6:1 and 7:1,respectively. The biotinylated cationic liposome/DNA volume ratio that completely retained DNA in the complex was 6:1.

Inhibition of hepatic fibrosis with artificial microRNAD Yang et al

1141

& 2013 Macmillan Publishers Limited Gene Therapy (2013) 1140 1148

but the fluorescence intensity in the treated liver was enhancedfollowing UTMD at multiple time points when compared with theuntreated normal group (Po0.01). Fluorescence intensity reacheda maximum at 4 days, still presented at 7 days and weakened at10 days when compared with the 24 h post-transfection group.The efficient delivery of EmGFP-labelled plasmid suggested thatthe amiRNA could be delivered to the liver in vivo and this effectremained at least 7 days post-transfection.

The effects of CTGF amiRNA on hepatic fibrosis pathologyThe results of haematoxylin and eosin staining and picrosirius redstaining are shown in Figure 5. The control liver had a normallobular architecture, and red collagen fibres were deposited onlyaround the blood vessels. The untreated hepatic fibrosis modelgroup and the amiScramble-treated model group exhibitedpseudo-lobular formations and abundant mature collagen fibres.Only mild fibrosis and small collagen staining areas were observedin the model group treated with amiCTGF (Figure 5a). Thecollagen staining areas in the amiCTGF-treated model group weremarkedly lower than in the model group and the amiScramble-treated group (Po0.01) (Figure 5b).

Immunohistochemical staining of CTGF, TGF-b1, type I collagenand a-SMAThe results of the immunohistochemical detection of CTGF, TGF-b1,type I collagen and a-SMA in each experimental group areshown in Figure 6. The expression of CTGF, TGF-b1, type I collagenand a-SMA was significantly increased in the model group andthe amiScramble-treated group, compared with the normalcontrol group (Po0.01). However, the staining of these proteinswas markedly decreased in the group treated with amiCTGF,

compared with the model and the amiScramble-treated groups(Po0.01).

The inhibitory effects of CTGF and TGF-b1 in the rat liver asdetermined by western blottingThe expression levels of CTGF and TGF-b1 protein were very low inthe normal control group. There was a dramatic increase in thelevels of CTGF and TGF-b1 protein expression in the model andamiScramble-treated groups, compared with the normal controlgroup (Po0.01). However, the expressions of these proteins weremarkedly decreased in the amiCTGF-treated group, comparedwith the model and the amiScramble-treated groups (Po0.01)(Figure 7a and b).

Quantitative real-time PCR analysis of CTGF and TGF-b1 mRNAexpression in the rat liverThe expression levels of CTGF and TGF-b1 mRNA were very low inthe normal control group. The CTGF and TGF-b1 mRNA levels weresignificantly increased in the model and amiScramble-treatedgroups, compared with the normal control group (Po0.01). How-ever, the mRNA levels were markedly decreased after amiCTGFtreatment, compared with the model and amiScramble-treatedgroups (Po0.01) (Figure 7c).

DISCUSSIONA major obstacle for amiRNAs in vivo is the development of anefficient delivery system into the target organ. Although viralvector systems are efficient, their safety and the difficulty oftargeting specific sites limit their applications.23 Plasmid deliverymethods are safer, but their transfection efficiencies are low. Theuse of hydrodynamic injections to effectively deliver genes to the

Figure 3. Properties of cationic liposome-bearing microbubbles. (a) Confocal laser scanning microscopy images of avidinylated microbubblesincubated with DNA-loading liposomes coated with propidium iodide (red fluorescence) via biotin-avidin bridge. A number of visible little red,round pitted DNA-loading liposomes surrounded the microbubble surface. The acoustic activity of the cationic liposome-bearingmicrobubbles (bd). The echo of the liver was low before the microbubble injection (b). The echo increased significantly after the injection ofcationic liposome-bearing microbubbles (c). This effect was clearly abated after the cationic liposome-bearing microbubbles were destroyedby ultrasound (d).


1142

Gene Therapy (2013) 1140 1148 & 2013 Macmillan Publishers Limited

rodent liver has become common,24 but it presents a risk for acuteheart failure.25 Recently, UTMD has been successfully used as agene delivery tool in vivo experiments.2628 Microbubbles havebeen recognized as targeted carriers when combined withultrasound for gene delivery. Many experimental UTMD studieshave been performed by incubating gene vectors with micro-bubbles to coat their surface. The self-made microbubbles with asize range from 1 to 8 mm are long circulation ultrasound contrastagents, which consist of a lipid shell encapsulating perfluoropro-pane gas.29 They have a significant role both in diagnosis and intherapy, which has been confirmed by numerous previousresearches.3032 Unfortunately, the gas core and the thin shell ofmicrobubbles limit its gene-carrying capacity. In contrast tomicrobubbles, the gene loading efficacy of liposomes has beenconfirmed, especially the cationic liposomes, which carry genesefficiently by binding negatively charged DNA to positively

charged liposomes through electrostatic interaction.33 DC-Chol isa charged lipid that gives the liposomes a positive charge that canefficiently combine with a negatively charged DNA plasmid. Theability of the biotinylated cationic liposomes to bind DNA wasanalysed using a gel retardation assay. When cationic liposomesand plasmid DNA volume ratio is greater than or equal to 6:1,no free DNA could be detected. This implies that all DNA iscomplexed with the liposomes. However, the fluid structuremakes them no acoustic properties and can not be used to targettissue with ultrasound. Avidinbiotin interactions are strong andhave been used successfully.34 Therefore, we developed the novelcationic liposome-bearing microbubbles by mounting cationicliposomes on microbubbles shells through the biotinavidinsystem to increase gene-loading capacity. We used those novelliposome-bearing microbubbles to deliver plasmid expressingamiCTGF into the liver tissue that was exposed to ultrasound.

Figure 4. The amiRNA expression plasmid was labelled with EmGFP and delivered by ultrasound-targeted novel cationic liposome-bearingmicrobubbles destruction (UTMD) in vivo. Fluorescence micrograph indicated that no EmGFP expression was observed in the untreatednormal liver tissue (a), but appeared in the the liver as early as 24 h (b), peaked at 4 days (c), lasted at least a week (d) and weakened at 10 days(e and f ) after UTMD transfection. (g) Quantified fluorescence intensity (integrated optical density). *Po0.01 when compared with the normalgroup. #Po0.01 when compared with the 24 h post-transfection group.


1143


The novel cationic liposome-bearing coating did not alter theacoustic characteristics of the microbubbles. Furthermore, theresults showed that the EmGFP signal appeared in the rat livertissue as early as 24 h after transfection and was maintained for atleast a week. Given these results, we delivered the plasmid intothe rat liver tissue every 7 days to achieve stable expression ofamiCTGF in vivo.

Dimethylnitrosamine (DMN)-induced liver fibrosis in rats is avaluable method to study early hepatic cirrhosis in humans andthe corresponding pathological changes, including inflammatorycell infiltration, hepatocyte necrosis, distinctive fibrous septa andcirrhosis.35 Hepatic fibrosis is the final pathological outcome forthe majority of chronic liver injuries regardless of the aetiology,and it progresses to liver cirrhosis, which has high morbidity andmortality. If treated properly, hepatic fibrosis is a reversibleprocess, avoiding the development of cirrhosis.36,37 In our study,short fibrotic septa extending into the liver lobules and some thinfibrotic septa between portal tracts were present 14 days afterDMN administration. These findings are similar to observations byLi Ch et al.38 Therefore, we chose to begin the therapy 14 daysafter DMN injection. Our study shows that repeated delivery ofamiRNAs to treat hepatic fibrosis by UTMD is safe and convenient.In addition, our data demonstrate that the progression of hepaticfibrosis is effectively ameliorated, and cirrhosis is successfullyprevented by downregulating CTGF expression through thedelivery of amiCTGF by UTMD. However, the liver injury was notreversed after amiCTGF treatment, likely due to persistent DMNdamage.

To investigate the role of amiCTGF in the treatment of hepaticfibrosis, we examined the expression of TGF-b1, a-SMA and type Icollagen. It is generally accepted that CTGF is a downstreameffector for TGF-b1. CTGF upregulation was positively correlated

with the enhanced expression of TGF-b1 in fibrotic livers.10,11

Our present study found that silencing CTGF markedly reducedthe upregulation of TGF-b1 expression. a-SMA is a marker ofactivated myofibroblasts. In the present study, CTGF inhibitiondownregulated a-SMA expression, indicating that amiCTGF canreduce HSC activation. Hepatic fibrosis is characterised by theaccumulation of ECM, including type I collagen as the maincomponent. Our study demonstrated that amiCTGF downregula-ted type I collagen expression and attenuated the development ofhepatic fibrosis in rats.

In conclusion, the present study shows that the combination ofnovel liposome-bearing microbubbles and ultrasound is aneffective way to deliver amiRNAs to the liver. We have successfullyshown that amiCTGF delivered via microbubbles combined withultrasound effectively silenced CTGF gene expression in thefibrotic liver tissue. Furthermore, amiCTGF significantly down-regulated TGF-b1 expression, inhibited HSC activation, reducedthe deposition of collagen and significantly regressed liverfibrogenesis. Our results suggest that gene silencing of CTGFusing amiRNAs delivered via liposome-bearing microbubblescombined with ultrasound may be a valuable specific approachfor the treatment of hepatic fibrosis.

MATERIALS AND METHODSPlasmid DNA vector constructionFour different sequences targeting CTGF genes were designed using theBLOCK-iT RNAi Designer at http://rnaidesigner. invitrogen. com/rnaiexpressand verified using a BLAST search. The amiRNA and control scramblesequences (Table 1) were annealed and ligated into a pcDNA6.2-GW/EmGFP-miR plasmid (Invitrogen, Carlsbad, CA, USA) that drives theexpression of amiRNA with polymerase II, and contains a spectinomycin

Figure 5. Artificial miRNA targeting CTGF ameliorated hepatic fibrosis. (a) Hepatic fibrosis was assessed by HE staining ( 100) and picrosiriusred staining ( 50). (b) Semiquantitative analysis of the area of picrosirius red staining. Data are representative of six randomly selectedmicroscopic fields from eight samples. The results showed pseudo-lobular formations and abundant mature collagen fibres in the model andamiScramble groups. The hepatic fibrosis was ameliorated by amiCTGF treatment. *Po0.01 when compared with the normal group. #Po0.01when compared with the model and amiScramble treatment groups.


1144


http://rnaidesigner. invitrogen. com/rnaiexpress

Figure 6. Artificial miRNA targeting CTGF suppressed the expression of CTGF, TGF-b1, type I collagen and a-SMA in the liver as determined byimmunohistochemical staining. (a) Representative microphotographs of CTGF, TGF-b1, type I collagen and a-SMA immunohistochemicalstaining in the liver ( 400). (b) Quantitative evaluation of immunohistochemical staining using Image-Pro Plus software. Data arerepresentative of six randomly selected microscopic fields from eight samples. An increase in CTGF, TGF-b1, type I collagen and a-SMAexpressions was observed in the model and amiScramble-treated groups when compared with the normal control group. In contrast, CTGF,TGF-b1, type I collagen and a-SMA expressions were significantly decreased in the group treated with amiCTGF when compared with themodel and the amiScramble-treated groups. *Po0.01 when compared with the normal group. #Po0.01 when compared with the model andamiScramble treatment groups.

Figure 7. The effects of artificial miRNAtargeting CTGF on CTGF and TGF-b1 expression. The levels of CTGF and TGF-b1 protein and mRNAwere analysed using western blot and quantitative real-time PCR, respectively. (a) The expression of CTGF and TGF-b1 protein was determinedusing western blot analysis. (b) Band intensities were measured and protein signals were normalised to GAPDH levels. (c) The mRNAexpressions of CTGF and TGF-b1 were detected by quantitative real-time PCR. The expression data were normalised to GAPDH expression. Alldata are reported as the means.d. of eight samples. *Po0.01 when compared with the normal group. #Po0.01 when compared with themodel and amiScramble treatment groups.


1145


resistance gene. The modified pre miRNA sequence structure was derivedfrom murine miR-155. The plasmid was purified using an Endo-freePlasmid mega or giga Kit (Qiagen Inc, Valencia, CA, USA), according to themanufacturers manual. The concentration of the plasmid was measuredusing spectrophotometry (1mgml 1).

Cell culture and transfectionA rat hepatic stellate cell strain (HSC-T6) was purchased from the Instituteof Tumour Research of the Chinese Academy of Medical Sciences, Beijing,China. HSC-T6 cells were cultured in Roswell Park Memorial Institutemedium-1640 medium (Gibco, Grand Island, NY, USA) containing 10%foetal bovine serum (HyClone, Logan, UT, USA) and then placed in anincubator with 5% CO2 at 37 1C. HSC-T6 cells were seeded in six-well plates1 day before transfection. Cells were transfected when the confluenceof the cells reached B80%. CTGF amiRNA expression vectors (CTGF-1,CTGF-2, CTGF-3 and CTGF-4) were individually transfected into the HSC-T6cell line with lipofectamine TM 2000 (Invitrogen). Untreated HSC-T6 cellsand cells transfected with a scrambled sequence amiRNA (amiScramble)were used as negative controls. Transfection was performed according tothe manufacturers instructions. Total RNA was extracted and the CTGFmRNA level was determined by real-time quantitative PCR after incubationfor 48 h. The CTGF protein was extracted and determined by the westernblot analysis after incubation for 72 h.

Preparation of cationic liposome-bearing microbubbles3b-[N-(N,N-dimethylaminoethane)-carbamoyl]-cholesterol (DC-Chol), 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC), 1,2 -distearoyl-sn-gly-cero-3-phosphoethanolamine-N-[biotinyl(polyethylene glycol)2000](DSPE-PEG2000-Biotin), polyethylene glycol 4000 (PEG4000) and 1,2distearoyl-sn-glycero-3-phosphatidylcholine (DSPC) were purchased fromAvanti Polar Lipids Inc. (Alabaster, AL, USA). Avidin from egg whites wasobtained from Sigma-Aldrich (St Louis, MO, USA), and neutravidin waspurchased from Pierce Biotechnology (Rockford, IL, USA). All otherreagents were analytical grade and used as received.

Preparation of biotinylated cationic liposomes and DNA-loadingliposomesBiotinylated cationic liposomes containing DPPC, PEG4000-positive lipids(DC-Chol) and DSPE-PEG2000-Biotin were mixed well with ultrapure waterabove the phase transition temperature of 60 1C. The suspension wasdivided into 1 ml aliquots in 3.5 ml vials and then lyophilised with a freezedrier overnight. Next, the freeze-dried powder was dissolved in 1 mlultrapure water. The sealed vial was then oscillated for 60 s with amechanical amalgamator (AT & M, Beijing, China) at a frequency of 6000.After sonication, the lipid suspension was washed three times with 10 mMphosphate-buffered saline (pH 7.4), followed by 5 min of centrifugation at300 g. The suspension was then lyophilised again with the freeze drier. Thefreeze-dried biotinylated cationic liposomes were divided into 200 mgaliquots in 3.5 ml vials. Next, 1 ml of ultrapure water was added to the vialto completely dissolve the powder. Then, the biotinylated cationicliposome solution was extruded through 0.22-mm pore size filters, andthe biotinylated cationic liposome solution was prepared. The sampleswere observed with a transmission electron microscope (Hitachi Inc, Tokyo,Japan).

A gel retardation assay was used to determine the volume ratio for thebiotinylated cationic liposomes that exhibited complete complex forma-tion with DNA. DNA (0.5mg, 0.25mg ml 1) was complexed with thebiotinylated cationic liposomes (5 mg ml 1) at different volume ratios andincubated at room temperature for 20 min. Bromophenol blue was addedto the samples. Next, the samples were electrophoresed in a 1% agarosegel containing ethidium bromide in a TAE buffer solution at 90 V for40 min. The average particle size and the z potential of the liposomes andDNA-loading liposomes were measured by a Coulter particle counter(Omec Technology Co, Ltd, Zhuhai, China) and a Malvern zeta potentialmeasuring instrument (Malvern instrument Co, Ltd, Malvern, England),respectively.

Preparation of biotinylated microbubblesAn appropriate amount of DPPG, DSPC and PEG-4000 supplemented withDSPE-PEG2000-Biotin was mixed well with ultrapure water in a 60 1Cthermostatic water bath. The resulting suspension was dispersed into 2 mlvials (1 ml per vial) and then lyophilised with a freeze drier. The preparedpowder from lyophilisation was dispersed in 1 ml of a solution containing50% glucose, propylene glycol and glycerine with a volume ratio of 8:1:1.The air in the vial was substituted by perfluoropropane. The vial was closedtightly and oscillated with a mechanical shaker (AT & M) with a frequencyof 4000 for 60 s.29

Conjugation of biotinylated cationic liposomes with biotinylatedmicrobubbles via avidinThe biotinylated microbubbles were washed three times with phosphate-buffered saline at 300 g for 5 min. An excess of avidin was added to thewashed microbubbles. Next, the microbubbles were vortexed shortly andincubated with gentle shaking for 30 min at room temperature. Afterincubation, the avidin-bound biotinylated microbubbles were washed withthree times with phosphate-buffered saline for 5 min centrifugations at300 g to remove the free avidin, and then resuspended in 10 ml ofphosphate-buffered saline. The DNA-loading cationic liposomes were thenadded to this suspension at a ratio of 6:1 (v/v) and incubated for 30 min atroom temperature with gentle shaking. The DNA-loading cationicliposomes-bearing microbubbles were washed three times to removethe free biotinylated cationic liposomes. The concentration and averageparticle size of the microbubbles carrying DNA-loading liposomes weremeasured by light microscopy (Leica, Wetzlar, Germany) and a Coulterparticle counter (Omec Technology Co, Ltd), respectively.

Subsequently, to test the ability of the DNA-loading liposomes to bind tothe surface of the biotinylated microbubbles, the procedure is as describedpreviously except for the DNA-loading liposomes-containing propidiumiodide-labelled plasmid DNA. Then, the samples were observed withconfocal laser scanning microscopy (Leica).

Acoustic response and delivery of DNA-loading cationic liposome-bearing microbubbles in vivoAll animal experiments were approved by the Local Ethical Committee ofthe Third Military Medical University. All animals (SpragueDawley (s.d.)rats) weighed B200 g and were housed in cages under standard animallaboratory conditions in the experimental animal centre of the ThirdMilitary Medical University. The rats were anaesthetised with 40 mg ml 1

pentobarbital throughout each procedure.

Table 1. Sequences of artificial miRNA against different target regions of CTGF

Name Sequence

CTGF-1 Top oligo strand 50-TGCTGTTCAGCTGCGCACTGACACTGGTTTTGGCCACTGACTGACCAGTGTCAGCGCAGCTGAA-30

Bottom oligo strand 50-CCTGTTCAGCTGCGCTGACACTGGTCAGTCAGTGGCCAAAACCAGTGTCAGTGCGCAGCTGAAC-30

CTGF-2 Top oligo strand 50-TGCTGTACACGGACCCACCGAAGACAGTTTTGGCCACTGACTGACTGTCTTCGGGGTCCGTGTA-30

Bottom oligo strand 50-TGCTGTACACGGACCCACCGAAGACAGTTTTGGCCACTGACTGACTGTCTTCGGGGTCCGTGTA-30

CTGF-3 Top oligo strand 50-TGCTGTTCCCGGGCAGCTTGACCCTTGTTTTGGCCACTGACTGACAAGGGTCACTGCCCGGGAA-30

Bottom oligo strand 50-CCTGTTCCCGGGCAGTGACCCTTGTCAGTCAGTGGCCAAAACAAGGGTCAAGCTGCCCGGGAAC-30

CTGF-4 Top oligo strand 50-TGCTGAGAACAGGCGCTCCACTCTGTGTTTTGGCCACTGACTGACACAGAGTGGCGCCTGTTCT-30

Bottom oligo strand 50-CCTGAGAACAGGCGCCACTCTGTGTCAGTCAGTGGCCAAAACACAGAGTGGAGCGCCTGTTCTC-30

Scramble Top oligo strand 50-TGCTGAAATGTACTGCGCGTGGAGACGTTTTGGCCACTGACTGACGTCTCCACGCAGTACATTTBottom oligo strand 50-CCTGAAATGTACTGCGTGGAGACGTCAGTCAGTGGCCAAAACGTCTCCACGCGCAGTACATTTC-30

Abbreviation: CTGF, connective tissue growth factor.


1146


Twenty-five rats were used. The GE Vivid 7 Dimension (General ElectricCo, Fairfield, CT, USA), a commercial clinical apparatus, was used at theoperating second-harmonic mode (transmit 1.5 MHz/receive 3.2 MHz) with anS4 transducer. The parameters used were described previously.22 The DNA-loading cationic liposomes-bearing microbubbles were created as describedabove. The liver was located using two-dimensional echographic images.Next, the microbubbles were injected using a scan depth of 3.0 cm and atotal of 0.5 ml kg 1 of the DNA-loading cationic liposomes-bearingmicrobubbles with 0.2 mg kg 1 plasmid were infused slowly via the caudalvein. The real time scanning mode was used with a low mechanical index(MI 0.08) to visualise the liver. The scanning mode was rapidly switched toFLASH mode (MI 1.0) to destroy the microbubbles. To allow for a widerfield of delivery, the transducer was positioned transversely to the liver, andultrasound was transmitted using a slow sweep along the length of the liver.All rats survived the experiment. The rats were randomly divided into fourgroups. The right liver lobes were collected after 24 h for six rats, 4 days forseven rats, 7 days for six rats and 10 days for six rats. The kidney, lung andspleen tissues were collected on day 4. The right liver lobes of six normaluntreated-rats were collected as control under general anaesthesia andimmediately frozen in liquid nitrogen. The frozen sections were preparedrapidly and then observed under a fluorescence microscope (Leica). Fivefields were selected randomly from each of two sections. The fluorescenceintensity of EmGFP was quantified using IMAGE-PRO PLUS software (MediaCybernetics Inc, Bethesda, MD, USA).

Animal models and protocolThe rats were randomly divided into four groups, including the normalcontrol group (n 8), the dimethylnitrosamine (DMN, Sigma)-inducedmodel group (n 8), the amiScramble group (n 8) and the amiCTGFgroup (n 8). In the normal control group, rats received intraperitonealsaline injections for 3 consecutive days per week for up to 4 weeks. Inthe DMN model, amiScramble and amiCTGF groups, rats received anintraperitoneal injection of 1% DMN (1 ml kg 1 body weight) for 3consecutive days per week for up to 4 weeks. At the end of the second andthird weeks, the DMN model group received a slow bolus injection of0.5 ml kg 1 body weight of microbubbles, followed by 1 ml of saline towash the tube via the caudal vein. Next, ultrasound was applied asdescribed. At the end of the second and third weeks, the amiScramble andthe amiCTGF groups were injected with 0.5 ml kg 1 of cationic liposomes-bearing microbubbles coupled with amiScramble plasmid or amiCTGFplasmid, respectively. Injections were followed by 1 ml of saline to washthe tube via the caudal vein, and ultrasound was applied as described.

All rats were euthanised 4 weeks after the DMN or saline administrationunder general anaesthesia. The liver tissue was quickly removed and aportion was instantly frozen in liquid nitrogen for western blot and real-time reverse transcriptase-PCR analyses. Another portion of the liver wasfixed in 10% phosphate-buffered formalin for histopathological studies.

Histology and immunohistochemistryAll paraffin-embedded liver tissues were stained with haematoxylin andeosin for histopathological examination, and picrosirius red for collagendetermination. Collagen densitometry was quantified using IMAGE-PROPLUS software. Immunohistochemical examinations were completed todetect the expression of CTGF, TGF-b1 (Abcam, Cambridge, MA, USA),a-SMA and type I collagen (Boster, Wuhan, China) in the liver. The stainingintensity of CTGF, TGF-b1, a-SMA and type I collagen was quantified usingIMAGE-PRO PLUS software. Three fields were selected randomly from eachof two sections, and eight rats from each group were examined.

Western blot analysisCells or liver tissue proteins were extracted using a Total Protein ExtractionKit (ProMab Biotechnology Inc, Richmond, CA, USA). The protein contentwas determined using the bicinchoninic acid method. The samples weremixed with 5 loading buffer. The lysates were boiled for 3 min and runthrough a 12% SDS-polyacrylamide gel at 300 mA for 70 min and thentransferred to an immobilon polyvinylidene fluoride membrane. Afterblocking with 5% skim milk blocking buffer for 2 h at room temperature,the membranes were incubated with CTGF antibody, TGF-b1 primaryantibody or GAPDH antibody (Santa Cruz, Santa Cruz, CA, USA) overnightat 4 1C. After washing, the membranes were hybridised with a horseradishperoxidase-conjugated secondary antibody at room temperature for 1 hand developed with an enhanced chemiluminescence detection kit (GE

Healthcare Life Sciences, Pittsburgh, PA, USA). The protein signals werenormalised to GAPDH levels.

Quantitative real-time PCR assayTotal RNA was isolated from the frozen liver tissue or HSC-T6 cells using aTrizol reagent according to the manufacturers instructions. The followingprimers were used: CTGF (no.NM_022266; forward 50-AAGCAGTTGCAAATACCAGT-30 , reverse 50-ATGTGTCTTCCAGTCGGTAG-30), TGF- b1 (no.NM_021578; forward 50-AACTACTGCTTCAGCTCCAC-30 , reverse 50-AGGACCTTGCTGTACTGTGT-30), GADPH (no.NM_017008; forward 50-CTCATGACCACAGTCCATGC-30 , reverse 50-TTCAGCTCTGG GATGACCTT-30 . Total RNA (2mg)was extracted and converted into complementary DNA using a Revert Aid HMinus First Strand complementary DNA Synthesis Kit (Fermentas Inc,Hanover, MD, USA). Quantitative real-time PCR was performed using a SYBRgreen PCR Kit (Applied Biosystems, Inc, Foster City, CA, USA) on an ABI7900real-time PCR instrument (Applied Biosystems). The PCR protocol consistedof 5 min at 95 1C, and 40 cycles of 20 s at 94 1C, 20 s at 59 1C, 20 s at 72 1C and5 min at 72 1C, 10 s at 55 1C. The housekeeping gene GAPDH was used as aninternal control for standardisation in triplicate. The specificity of theamplification products was tested using a melting-curve analysis, and thecomparative threshold (Ct) method was used to calculate the relativeamount of mRNA of the treated sample to the control samples.

Statistical analysisThe results are presented as the mean and s.d. In the semiquantitativeanalysis, nonparametric test (Wilcoxons test) was used; other statisticalanalysis of values was performed using the unpaired Students t-test. Allcalculations were performed using SPSS 17.0 software (SPSS Inc, Chicago,IL, USA). Differences were considered to be highly significant at Po0.01.

CONFLICT OF INTERESTThe authors declare no conflict of interest.

ACKNOWLEDGEMENTSThis work was supported by a grant from the National Nature Science Foundation ofChina (No. 30970831 and No.30901389).

REFERENCES1 Hernandez-Gea V, Friedman SL. Pathogenesis of liver fibrosis. Annu Rev Pathol

2011; 6: 425456.2 Lee UE, Friedman SL. Mechanisms of hepatic fibrogenesis. Best Pract Res Clin

Gastroenterol 2011; 25: 195196.3 Jiao J, Friedman SL, Aloman C. Hepatic fibrosis. Curr Opin Gastroenterol 2009; 25:

223229.4 Cheng K, Yang N, Mahato RI. TGF-beta1 gene silencing for treating liver fibrosis.

Mol Pharm 2009; 6: 772779.5 Sisco M, Kryger ZB, OShaughnessy KD, Kim PS, Schultz GS, Ding XZ et al.

Antisense inhibition of connective tissue growth factor (CTGF/CCN2) mRNA limitshypertrophic scarring without affecting wound healing in vivo. Wound RepairRegen 2008; 16: 661673.

6 Gao RP, Brigstock DR. Connective tissue growth factor hammerhead ribozymeattenuates human hepatic stellate cell function. World J Gastroenterol 2009; 15:38073813.

7 Huang G, Brigstock DR. Regulation of hepatic stellate cells by connective tissuegrowth factor. Front Biosci 2012; 17: 24952497.

8 Li G, Li D, Xie Q, Shi Y, Jiang S, Jin Y. RNA interfering connective tissue growthfactor prevents rat hepatic stellate cell activation and extracellular matrixproduction. J Gene Med 2008; 10: 10391047.

9 Zhang D, Wang NY, Yang CB, Fang GX, Liu W, Wen J et al. The clinical value ofserum connective tissue growth factor in the assessment of liver fibrosis. Dig DisSci 2010; 55: 767774.

10 George J, Tsutsumi M. siRNA-mediated knockdown of connective tissue growthfactor prevents N-nitrosodimethylamineinduced hepatic fibrosis in rats. GeneTher 2007; 14: 790803.

11 Li G, Xie Q, Shi Y, Li D, Zhang M, Jiang S et al. Inhibition of connective tissuegrowth factor by siRNA prevents liver fibrosis in rats. J Gene Med 2006; 8: 889900.

12 Roderburg C, Urban GW, Bettermann K, Vucur M, Zimmermann H, Schmidt S et al.Micro-RNA profiling reveals a role for miR-29 in human and murine liver fibrosis.Hepatology 2011; 53: 209218.


1147


Joseph GeorgeRectangle

13 Maubach G, Lim MC, Chen J, Yang H, Zhuo L. miRNA studies in vitro and in vivoactivated hepatic stellate cells. World J Gastroenterol 2011; 17: 27482773.

14 Hu T, Fu Q, Chen P, Ma L, Sin O, Guo D. Construction of an artificial MicroRNAexpression vector forsimultaneous inhibition of multiple genes in mammaliancells. Int J Mol Sci 2009; 10: 21582168.

15 Baek MN, Jung KH, Halder D, Choi MR, Lee BH, Jung MH et al. Artificial microRNA-based neurokinin-1 receptor gene silencing reduces alcohol consumption inmice. Neurosci Lett 2010; 475: 124128.

16 Fan ZD, Zhang L, Gan XB, Gao XY, Zhu GQ. Artificial microRNA interferencetargeting AT(1a) receptors in paraventricular nucleus attenuates hypertension inrats. Gene Therapy 2011; 19: 810817.

17 Suzuki R, Oda Y, Utoguchi N, Maruyama K. Progress in the development ofultrasound-mediated gene delivery systems utilizing nano- and microbubbles.J Control Release 2011; 149: 3641.

18 Mayer CR, Geis NA, Katus HA, Bekeredjian R. Ultrasound targeted microbubbledestruction for drug and gene delivery. Expert Opin Drug Deliv 2008; 5:11211138.

19 Bekeredjian R, Chen S, Frenkel PA, Grayburn PA, Shohet RV. Ultrasound-targetedmicrobubble destruction can repeatedly direct highly specific plasmid expressionto the heart. Circulation 2003; 108: 10221026.

20 Kheirolomoom A, Dayton PA, Lum AF, Little E, Paoli EE, Zheng H et al.Acoustically-active microbubbles conjugated to liposomes: characterization of aproposed drug delivery vehicle. J Control Release 2007; 118: 275284.

21 Vandenbroucke RE, Lentacker I, Demeester J, De Smedt SC, Sanders NN. Ultra-sound assisted siRNA delivery using PEG-siPlex loaded microbubbles. J ControlRelease 2008; 126: 265273.

22 Yang D, Tan KB, Gao YH, Liu H, Yang WX. Effects of diagnostic ultrasound-targetedmicrobubble destruction on permeability of normal liver in rats. Ultrasonics 2012;52: 10651071.

23 Mingozzi F, High KA. Immune responses to AAV in clinical trials. Curr Gene Ther2011; 11: 321330.

24 Chen SW, Zhang XR, Wang CZ, Chen WZ, Xie WF, Chen YX. RNA interferencetargeting the platelet-derived growth factor receptor beta subunit amelioratesexperimental hepatic fibrosis in rats. Liver Int 2008; 28: 14461457.

25 Sawyer GJ, Rela M, Davenport M, Whitehorne M, Zhang X, Fabre JW.Hydrodynamic gene delivery to the liver: theoretical and practical issues forclinical application. Curr Gene Ther 2009; 9: 128135.

26 Carson AR, McTiernan CF, Lavery L, Hodnick A, Grata M, Leng X et al. Gene therapyof carcinoma using ultrasound- targeted microbubble destruction. UltrasoundMed Biol 2011; 37: 393402.

27 Vancraeynest D, Havaux X, Pasquet A, Gerber B, Beauloye C, Rafter P et al.Myocardial injury induced by ultrasound-targeted microbubble destruction:evidence for the contribution of myocardial ischemia. Ultrasound Med Biol 2009;35: 672679.

28 Qiu L, Zhang L, Wang L, Jiang Y, Luo Y, Peng Y et al. Ultrasound-targetedmicrobubble destruction enhances naked plasmid DNA transfection in rabbitAchilles tendons in vivo. Gene Therapy 2011; 19: 703710.

29 Liu P, Wang X, Zhou YS, Hua X, Liu Z, Gao Y. Effects of a novel ultrasound contrastagent with long persistence on right ventricular pressure: comparison withSonovue. Ultrasonics 2011; 51: 210214.

30 Liu H, Wang X, Tan KB, Liu P, Zhuo ZX, Liu Z et al. Molecular imaging of vulnerableplaques in rabbits using contrast-enhanced ultrasound targeting to vascularendothelial growth factor receptor-2. J Clin Ultrasound 2011; 39: 8390.

31 Xu YL, Gao YH, Liu Z, Tan KB, Hua X, Fang ZQ et al. Myocardium-targetedtransplantation of mesenchymal stem cells by diagnostic ultrasound-mediatedmicrobubble destruction improves cardiac function in myocardial infarction ofNew Zealand rabbits. Int J Cardiol 2010; 138: 182195.

32 Gao Y, Gao S, Zhao B, Zhao Y, Hua X, Tan K et al. Vascular effects of microbubble-enhanced, pulsed, focused ultrasound on liver blood perfusion. Ultrasound MedBiol 2012; 38: 9198.

33 Tros de Ilarduya C, Sun Y, Duzgunes N. Gene delivery by lipoplexes andpolyplexes. Eur J Pharm Sci 2010; 40: 159170.

34 Wang W, Liu GJ, Xie XY, Xu ZF, Chen LD, Huang GL et al. Development andevaluation of lipid microbubbles targeted to alpha(v)beta(3)-integrin via biotin-avidin bridge. J Microencapsul 2012; 29: 177184.

35 Lee ES, Shin MO, Yoon S, Moon JO. Resveratrol inhibits dimethylnitrosamine-induced hepatic fibrosis in rats. Arch Pharm Res 2010; 33: 925932.

36 Gieling RG, Burt AD, Mann DA. Fibrosis and cirrhosis reversibilitymolecularmechanisms. Clin Liver Dis 2008; 12: 915937.

37 Ismail MH, Pinzani M. Reversal of liver fibrosis. Saudi J Gastroenterol 2009; 15:7279.

38 Li CH, Piao DM, Xu WX, Yin ZR, Jin JS, Shen ZS. Morphological and serumhyaluronic acid, laminin and type IV collagen changes in dimethylnitrosamine-induced hepatic fibrosis of rats. World J Gastroenterol 2005; 11: 76207624.


1148


title_linkIntroductionResultsThe effect of artificial miRNA targeted against CTGF on CTGF expression invitroThe characterisation of cationic liposomesProperties of cationic liposome-bearing microbubblesDelivery of amiRNA to the rat liver by cationic liposome-bearing microbubbles combined with ultrasound

Figure1The effect of amiRNA targeted against CTGF on CTGF expression in HSC-T6 cells. HSC-T6 cells were treated with media (blank), scrambled amiRNA (amiScramble) or four artificial microRNA (miRNA) that targeted CTGF (CTGF-1, CTGF-2, CTGF-3 and CTGF-4)Figure2The characterisation of the cationic liposomes. (a) A transmission electron microscope image of the biotinylated cationic liposomes. They appeared as spherical or nearly spherical nanoparticles. (b) A gel retardation assay was used to detect the bThe effects of CTGF amiRNA on hepatic fibrosis pathologyImmunohistochemical staining of CTGF, TGF-beta1, type I collagen and agr-SMAThe inhibitory effects of CTGF and TGF-beta1 in the rat liver as determined by western blottingQuantitative real-time PCR analysis of CTGF and TGF-beta1 mRNA expression in the rat liver

DiscussionFigure3Properties of cationic liposome-bearing microbubbles. (a) Confocal laser scanning microscopy images of avidinylated microbubbles incubated with DNA-loading liposomes coated with propidium iodide (red fluorescence) via biotin-avidin bridge. A numbeFigure4The amiRNA expression plasmid was labelled with EmGFP and delivered by ultrasound-targeted novel cationic liposome-bearing microbubbles destruction (UTMD) invivo. Fluorescence micrograph indicated that no EmGFP expression was observed in the untrMaterials and methodsPlasmid DNA vector construction

Figure5Artificial miRNA targeting CTGF ameliorated hepatic fibrosis. (a) Hepatic fibrosis was assessed by HE staining (times100) and picrosirius red staining (times50). (b) Semiquantitative analysis of the area of picrosirius red staining. Data are repreFigure6Artificial miRNA targeting CTGF suppressed the expression of CTGF, TGF-beta1, type I collagen and agr-SMA in the liver as determined by immunohistochemical staining. (a) Representative microphotographs of CTGF, TGF-beta1, type I collagen and agr-SFigure7The effects of artificial miRNAtargeting CTGF on CTGF and TGF-beta1 expression. The levels of CTGF and TGF-beta1 protein and mRNA were analysed using western blot and quantitative real-time PCR, respectively. (a) The expression of CTGF and TGF-betCell culture and transfectionPreparation of cationic liposome-bearing microbubblesPreparation of biotinylated cationic liposomes and DNA-loading liposomesPreparation of biotinylated microbubblesConjugation of biotinylated cationic liposomes with biotinylated microbubbles via avidinAcoustic response and delivery of DNA-loading cationic liposome-bearing microbubbles invivo

Table 1 Animal models and protocolHistology and immunohistochemistryWestern blot analysisQuantitative real-time PCR assayStatistical analysis

A5A6ACKNOWLEDGEMENTSA7

inhibition of hepatic fibrosis with artificial microrna...

Documents