lable at ScienceDirect

Biomaterials 32 (2011) 8654e8662

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomateria ls

Magnetically enhanced adeno-associated viral vector delivery for human neuralstem cell infection

Eunmi Kim a, Ji-Seon Oh a, Ik-Sung Ahn a, Kook In Park b, Jae-Hyung Jang a,*

aDepartment of Chemical and Biomolecular Engineering, Yonsei University, Seoul 120-749, Republic of KoreabDepartment of Pediatrics and BK21 Project for Medical Science, Yonsei University College of Medicine, Seoul, Republic of Korea

a r t i c l e i n f o

Article history:Received 28 June 2011Accepted 24 July 2011Available online 15 August 2011

Keywords:Adeno-associated virusSuperparamagnetic iron oxide nanoparticleHexa-histidineGene delivery

* Corresponding author. Tel.: þ82 2 2123 2756; faxE-mail address: j-jang@yonsei.ac.kr (J.-H. Jang).

0142-9612/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.biomaterials.2011.07.075

a b s t r a c t

Gene therapy technology is a powerful tool to elucidate the molecular cues that precisely regulate stemcell fates, but developing safe vehicles or mechanisms that are capable of delivering genes to stem cellswith high efficiency remains a challenge. In this study, we developed a magnetically guided adeno-associated virus (AAV) delivery system for gene delivery to human neural stem cells (hNSCs). Magnet-ically guided AAV delivery resulted in rapid accumulation of vectors on target cells followed by forcedpenetration of the vectors across the plasma membrane, ultimately leading to fast and efficient cellulartransduction. To combine AAV vectors with the magnetically guided delivery, AAV was geneticallymodified to display hexa-histidine (6xHis) on the physically exposed loop of the AAV2 capsid (6xHisAAV), which interacted with nickel ions chelated on NTA-biotin conjugated to streptavidin-coatedsuperparamagnetic iron oxide nanoparticles (NiStNPs). NiStNP-mediated 6xHis AAV delivery undermagnetic fields led to significantly enhanced cellular transduction in a non-permissive cell type (i.e.,hNSCs). In addition, this delivery method reduced the viral exposure times required to induce a highlevel of transduction by as much as to 2e10 min of hNSC infection, thus demonstrating the greatpotential of magnetically guided AAV delivery for numerous gene therapy and stem cell applications.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

The advancement of stem cell technologies requires the eluci-dation of molecular or environmental cues that can regulate theproperties of self-renewal, the capacity to expand stem cells, anddifferentiation into specialized cell lineages. However, developingtools that can mediate the precise control of stem cells remainsa challenge. Delivering genes that encode transcription factors orregulatory molecules can be a powerful strategy to circumvent thechallenges associated with current stem cell technologies becauseit can serve as an efficient tool to modulate stem cell fates by theup- or down-regulation of specific genes [1e4]. Critical issues,however, still remain in the development of efficient and safe genedelivery vectors that can be utilized in stem cell applications.

Current approaches to mediate gene delivery to stem cells canbe classified as plasmid- or virus-based gene delivery. Althoughplasmid-based gene delivery has great advantages in terms ofsafety and the ability to carry large genes, this method may havelimitations in the stimulation of stem cells due to its low stem celltransfection efficiencies [5]. Alternatively, viral vectors, which

: þ82 2 312 6401.

All rights reserved.

typically offer high infectivity and stable gene expression, havebeen extensively utilized to manipulate stem cells [6,7]. However,there exist major concerns about the use of viral vectors that areprimarily based on retroviruses and lentiviruses for a variety ofstem cell applications; these concerns are primarily based oninsertional mutagenesis [8,9]. Therefore, the transient expression ofregulatory signals tomanipulate stem cells may be desirable, whichcan relieve safety issues associated with insertional mutagenesis.Adenovirus-mediated vectors have demonstrated efficient genedelivery and transient gene expression in stem cells, but virallyencoded factors that can cause immune responses may hinder thewidespread use of adenoviral vectors for stem cell applications[10,11].

Alternatively, adeno-associated virus (AAV), which is beingbroadly explored in clinical trials, may satisfy the safety and effi-ciency criteria for use as a gene delivering vector in stem cellresearch. AAV is a parvovirus with a 4.7 kb single-stranded DNAgenome containing two genes (rep and cap), and AAV has the abilityto induce high delivery efficiencies in a broad range of both dividingand non-dividing cells [12e14]. Additionally, one of the greatadvantages of AAV for stem cell applications is that AAV vectorgenomes can mediate homologous recombination with targetsequences in a cellular genome at efficiencies that are 103e104

E. Kim et al. / Biomaterials 32 (2011) 8654e8662 8655

higher than plasmid constructs [15,16], thus increasing its potentialutility in numerous stem cell applications, such as gene correction,precise stem cell regulation, and the creation of disease models.Unfortunately, AAV has intrinsic properties that prevent it fromefficiently delivering genes to themajority of stem cell types [17,18].Therefore, the development of AAV vector-based delivery systemsthat are capable of efficient stem cell infection is a prerequisite toinvestigate stem cells using this vector.

In this study, the enhancement of AAV-mediated gene deliveryto stem cells by employing magnetic fields that act on viral vectorsassociated with superparamagnetic nanoparticles was attempted.Streptavidin-coated superparamagnetic iron oxide nanoparticles(SPIONs) were used to form complexes with AAV vectors, and theSPIONs were additionally conjugated with biotin-nitrilotriaceticacid (NTA) to allow nickel ion chelation (Fig. 1). To associate AAVvectors with the NiStNPs, a hexa-histidine (6xHis) tag was insertedinto a physically exposed loop of the wild type AAV2 capsid (i.e.,amino acid position 587). Magnetic nanoparticle-mediated AAVdelivery was demonstrated to be an effective tool to enhance genedelivery efficiency in human neural stem cells (hNSCs), a cell typethat is non-permissive to AAV, thus demonstrating the greatadvantages of magnetically guided AAV delivery for numerouspotential stem cell applications, such as gene correction, stem cellregulation, or the creation of disease models.

2. Materials and methods

2.1. Superparamagnetic nanoparticles conjugated with Ni-NTA-biotin

Streptavidin-coated superparamagnetic iron oxide nanoparticles (SPIONs;fluidMAG-Streptavidin) were purchased from Chemicell (Berlin, Germany). Thehydrodynamic diameter of a SPION was approximately 100 nm, and the bindingcapacity of 1 mg of SPIONs was approximately 80 pmol of biotinylated protein,according to the manufacturer’s specifications. The SPIONs were diluted in aqueoussolution containing 40 mM NaCl and subsequently agitated gently with the excessamount of biotin-nitrilotriacetic acid (biotin-NTA, 240 pmol; Biotium, Hayward, CA)and nickel chloride (100 mM; SigmaeAldrich, St. Louis, MO) for 2 h to generate Ni-NTA-biotin-streptavidin-SPION complexes (NiStNPs). The presence of Ni-NTA-biotinon the streptavidin SPIONs was confirmed using Fourier-Transform Infrared (FT-IR)spectroscopy (Spectrum 100, Perkin Elmer, USA).

2.2. Cell culture

AAV 293 cells (Stratagene, La Jolla, CA) were utilized for viral packaging andwere cultured in Dulbecco’s modified Eagle medium (DMEM; Invitrogen, Carlsbad,CA) with 10% fetal bovine serum (FBS) (Invitrogen) and 1% penicillin/streptomycin(Invitrogen) at 37 �C and 5% CO2. HEK293T cells and human fetal neural stem cells(hNSC, 13 weeks of gestational age) were cultured to analyze the transductionefficiencies. HEK293T cells were grown in DMEM with 10% FBS and 1% penicillin/

Fig. 1. Schematic illustration of magnetically guided 6xHis AAV/NiStNP complexes: (A)conjugation onto the streptavidin-coated SPIONs, (C) 6xHis AAV bound NiStNPs.

streptomycin at 37 �C and 5% CO2. Human NSCs, which were derived from thetelencephalon (HFT13) as previously described [19], were cultured in DMEM/F12(Invitrogen) containing N-2 supplement, 20 ng/mL of fibroblast growth factor 2,8 mg/mL of heparin (SigmaeAldrich, St. Louis, MO), and 10 ng/mL of leukemiainhibitory factor (LIF; Chemicon, Temecula, CA).

2.3. Production and purification of 6xHis-AAV vectors

Two AAV vectors were produced in this study: i) 6xHis AAV vectors for immo-bilization on the NiStNPs and ii) wild-type AAV2 vectors as a control. Hexa-histidine(6xHis) residues were utilized as moieties to allow the AAV vectors to interact withthe nickel ions that were chelated to the NTA-biotin-streptavidin-coated SPIONs(NiStNPs). The hexa-histidine-coding plasmid (pXX2 His6) and pXX2 were con-structed as described in a previous study [20]. Hexa-histidine residues were insertedat position 587 in the amino acid sequence, corresponding to physically exposedloop domain: - LQRGNLGHHHHHHSRQA- (wt AAV2: 583LQRGNRQA-) [20].Recombinant 6xHis-AAV vectors carrying cDNA encoding green fluorescent protein(GFP) driven by a cytomegalovirus (CMV) promoter were packaged using the tran-sient transfectionmethod. Briefly, an equal mass (17 mg) of three plasmids, includingan AAV helper plasmid (i.e., pXX2 His6), the CMV GFP vector plasmid containinginverted terminal repeats (ITR) (pAAV CMV GFP SN), and an adenoviral helperplasmid (Stratagene), were transfected via calcium phosphate into AAV 293 cells[21]. Because the heparan sulfate proteoglycan (HSPG) binding domain waspreserved upon hexa-histidine insertion, the resulting viral vectors were harvestedas previously described [22] and purified using heparin column chromatography[23]. For purification using a heparin column, the crude lysate of 6xHis-AAV vectorswas loaded onto a 1 mL of HiTrap heparin column (GE Healthcare, Pittsburgh, PA)that had been previously equilibrated with Tris buffer (50 mM, pH 7.5) containing150 mM NaCl. The columnwas washed with 3 mL of 150 mM NaCl, and elutions wereperformed with 3 mL of 1 M NaCl. The viral fractions were desalted and bufferexchanged into phosphate-buffered saline (PBS)/0.01% Tween 20 using AmiconUltra-15 Centrifugal Filter Units (Millipore, Billerica, MA) according to the manu-facturer’s instructions. DNase-resistant genomic titers were determined by quanti-tative PCR (QPCR; Mini Opticon, Bio-Rad, Hercules, CA).

2.4. Preparation of AAV/NiStNP complexes

The resulting 6xHis-AAV vectors (2 � 109 viral genomic particles) were dilutedin 50 mL of serum-free DMEM andmixed gently with the Ni-NTA-biotin-streptavidinFluidMAG nanoparticles suspended in 50 mL of serum-free DMEM with differentquantities of nanoparticles (i.e., 0.02, 0.2, 2, or 4 mg). The mixture was additionallyagitated for 120 min at room temperature, and an additional 200 mL of serum-freeDMEM was added to reduce the aggregation of 6xHis-AAV/NiStNP complexesprior to addition to the cell culture medium for infection. To examine the associationof AAV vectors with the NiStNPs, after magnetic sedimentation (NeoDeltaMagnet1170e1250 mT; IBS magnet, Berlin, Germany), the number of genomic particles ofunbound AAV vectors in the supernatant was determined by quantitative PCR(QPCR).

2.5. Characterization of AAV/NiStNP complexes

The presence of Ni-NTA-biotin on the streptavidin-coated SPIONs wasconfirmed by Fourier-Transform Infrared (FT-IR) spectroscopic measurements(Spectrum 100, Perkin Elmer, USA). Hydrodynamic particle sizes were determined

streptavidin-coated superparamagnetic iron oxide nanoparticles, (B) Ni-NTA-biotin

Fig. 2. Characterization of NiStNPs and 6xHis AAV/NiStNP complexes: (A) FT-IRanalysis of streptavidin-coated SPIONs (bottom line) and Ni-NTA-biotin-conjugatedstreptavidin-SPIONs (upper line), (B) NiStNP-bound quantity of AAV2 and 6xHis AAVas a function of the amount of NiStNPs, (C) hydrodynamic particle sizes determined bydynamic light scattering, and transmission electron microscopic (TEM) images ofNiStNPs only (D) and 6xHis AAV (arrow)-bound NiStNPs (dotted line) (E). The scale barindicates 20 nm.

E. Kim et al. / Biomaterials 32 (2011) 8654e86628656

by dynamic light scattering using a Zetasizer (Zen3600, Malvern Instruments Ltd.,Worcestershire, United Kingdom). Additionally, the conjugation of Ni-NTA-biotinand the immobilization of 6xHis AAV onto the streptavidin-coated SPIONs werevisualized using transmission electron microscopy (TEM) (JEM-2011, JEOL Ltd.,Tokyo, Japan) at the Yonsei Center for Research Facilities. Each sample (NiStNP only,6xHis AAV/NiStNPs) was prepared by dipping a carbon-coated grid into the samplefollowed by drying overnight at room temperature, and TEM images were takenunder 200 kV with 400000� magnification.

2.6. Transduction assay

To test the infectivity of SPION-mediated AAV delivery, the AAV/NiStNPcomplexes (2 � 109 viral genomic particles) carrying cDNA encoding GFP driven bythe CMV promoter were added on top of the medium containing cells (HEK293T:4�104 cells/well and hNSCs: 1�105 cells/well) grown on 24-well plates (TCP) in thepresence or absence of a magnetic field (MagnetoFACTOR plate 24, Chemicell, Berlin,Germany). Each well was exposed to the magnetic field for 2, 10, or 120 min, and cellculture media containing AAV (AAV2, 6xHis AAV)/NiStNP or AAV vectors only (AAV2,6xHis AAV) was replaced with fresh media at the same time. The percentage of GFP-expressing cells out of the total number of cells was determined at 48 h post-infection using flow cytometry at the Yonsei University College of MedicineMedical Research Center (Becton Dickinson FACS Caliber).

2.7. Internalization and inhibition of AAV-nanoparticles in the cytoplasm

Fluorescently labeled streptavidin-SPIONs (nano-screenMAG-Streptavidin:excitation at 578 nm, Chemicell, Berlin, Germany) were utilized to visualize theinternalization of AAV/NiStNP complexes within the cytoplasm. To image the fluo-rescence, cells were trypsinized immediately after exposure to AAV/NiStNPcomplexes for 2, 10, or 120 min, re-plated onto new 24-well TCPs, and grown foradditional 24 h prior to fluorescence imaging (Eclipse Ti-S, Nikon, Tokyo, Japan) andfor 48 h prior to flow cytometry analysis. The number of GFP-expressing cells wasquantified using flow cytometry to assess the capability of SPION-mediated AAVdelivery under magnetic guidance for rapid internalization into the cytoplasmcompared with non-SPION-mediated delivery (i.e., virus only) or SPION-mediateddelivery without magnetic forces. Additionally, to test the inhibition of theenergy-dependent endocytosis pathway for magnetically guided AAV delivery,hNSCs were exposed to the 6xHis AAV/NiStNPs for 2, 10, or 120 min at 4 �C, and thepercentage of GFP-expressing cells was compared with that at 37 �C. After viralexposure at 4 �C, the supernatant was removed at designated time points (2, 10, and120 min), and the cells were vigorously washed twice and cultured for additional48 h at 37 �C prior to flow cytometry analysis.

2.8. Cell viability

To analyze the viability of the cells infected by the 6xHis AAV/NiStNP complexes,metabolically active cells were stained with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; SigmaeAldrich), and colorimetric changeswere detected using a spectrophotometer (NanoDrop 2000; Thermo Scientific, WestPalm Beach, FL) relative to the positive control (i.e., 6xHis AAV only delivery of5 � 104 genomic MOI for HEK293T and 2 � 104 MOI for hNSCs). Each cell (HEK293T,hNSC) was infected by AAV only (AAV2, 6xHis AAV), NiStNP only, or AAV/NiStNPcomplexes and harvested after 48 h. Briefly, 0.1 volume of MTT solution relative tothe volume of the medium containing cells (20 mL of 2.5 mg/mL in 200 mL medium)was dispensed into each well, and the cultures were incubated for 4 h at 37 �C.Supernatants were collected and mixed with dimethyl sulfoxide (DMSO; EMDBiosciences, Gibbstown, NJ), and spectrophotometric absorbances (550 nm) weremeasured using a NanoDrop 2000 (Thermo Scientific, West Palm Beach, FL). Finally,the absorbance values were normalized to the absorbance value of the positivecontrol, in which cells were cultured with AAV only for the same durations.

2.9. Statistical analysis

All the conditions were performed in triplicate, and experimental data wereexpressed as mean � standard deviation (SD). Statistical analyses between condi-tions (i.e., Student’s t test) were performed using the JMP software package (SASInstitute, Inc., Cary, NC, USA).

3. Results and discussion

3.1. Characterization of AAV/NiStNP complexes

We developed a strategy to immobilize genetically modifiedAAV vectors (i.e., 6xHis tagged AAV) onto superparamagnetic ironoxide nanoparticles (SPIONs), as schematically illustrated in Fig. 1.Specifically, we inserted a 6xHis tag into a physically exposed loopof the AAV2 capsid (i.e., amino acid position 587), as previously

E. Kim et al. / Biomaterials 32 (2011) 8654e8662 8657

described [20], which enabled the immobilization of the AAV ona surface displaying nickel ions chelated by biotin-nitrilotriaceticacid (biotin-NTA) moieties bound to a streptavidin-coated surface[24]. Similarly, the same 6xHis AAV vectors were bound tostreptavidin-coated SPIONs conjugated to biotin-NTA-Ni (NiStNPs).Genomic titers of 6xHis AAV vectors, produced by the transienttransfectionmethod followed by heparin column purification, wereslightly lower (2.2 � 1012 vg (viral genomes)/mL) than those of thewild type AAV2 (3.8 � 1012 vg/mL) generated by the same method.The 6xHis AAV vectors, including crystal structures, heparinaffinity, and infectious units, compared with wild-type AAV2 havebeen previously characterized in detail [20].

Prior to immobilizing the 6xHis AAV vectors on the NiStNPs,successful conjugation of biotin-NTA onto the streptavidin-coatednanoparticles for the subsequent chelation of Ni ions on the NTA

Fig. 3. Metabolic activities of HEK293T cells (A) and hNSCs (B) upon exposure to virus only,metabolic activities (i.e., cell viabilities) of cells treated with different quantities of NiStNPsstatistically significant differences compared with those that were treated with virus (6xHi

was confirmed using Fourier-Transform Infrared spectroscopy (FT-IR) analysis (Fig. 2A). Characteristic peaks appearing upon theaddition of Ni-NTA-biotin to the streptavidin-coated iron oxidenanoparticles were observed in the regions of 3025e2918, 2078,and 1600e1370 cm�1. The stretching vibration of the C]O occur-ring at approximately 1600 cm�1 confirmed the presence of the C]O bond within the tetrahydroimidizalone ring fused with the tet-rahydrothiophene ring of biotin [25]. Additionally, the peaks at1495 and 1449 cm�1 corresponded to the NeH bending and thering vibrations of the tetrahydroimidizalone-tetrahydrothiophenering of biotin, respectively [25]. Regions 1370e1600 and2900e3000 cm�1 of the spectrum, assigned to the NTA-amine [26],confirmed the presence of NTA. Because the presence of nickel ioncannot be identified in the mid-infrared region, successful chela-tion of nickel ions on the NTA was verified by its specific

magnetically guided 6xHis AAV vectors. The MTT assay was performed to measure theunder various viral exposure times (2, 10, and 120 min). The symbols * and ** indicates AAV) only for 24 h (P < 0.05 and P < 0.01, respectively).

E. Kim et al. / Biomaterials 32 (2011) 8654e86628658

interactions with the 6xHis-tagged proteins (i.e., AAV vectors) onthe nanoparticles.

The 6x-histidine displayed on the viral capsid led to the bindingof the 6xHis AAV vectors to the NiStNPs, with high efficienciesranging from 98.1% to 99.5%. In contrast, the absence of the histi-dine residues (i.e., AAV2 vectors) resulted in binding capacities lessthan 30%, revealing the specific interactions between the 6x-HisAAV vectors and the nickels ions chelated on the nanoparticles(Fig. 2B). Because the majority of the 6xHis AAV vectors (>98%)successfully bound to the NiStNPs, the separation of the unbound-6xHis AAV vectors prior to infectionwas not necessary. The averageamount of NiStNP-bound AAV2 vectors slightly increased withincreased quantities of NiStNP, possibly indicating non-specificinteractions between the vectors and the nanoparticles.

An increased quantity of nanoparticles led to an increase in thehydrodynamic particle size of the 6xHis AAV/NiStNP complexes,ranging from 104.60 nm to 161.89 nm (Fig. 2C). Similarly, thehydrodynamic diameters of the nanoparticles alone increased inconjunction with an increased quantity of nanoparticles (i.e.,85.75 nme126.30 nm), implying that the larger diameter of thevirus/NiStNP complexes may simply be a result of the agglomera-tion of nanoparticles. Interestingly, formulations of complexes

Fig. 4. Cellular transduction of HEK293T cells by AAV2/NiStNP complexes vs. that by 6xHis Aimages showing GFP-expressing HEK293T cells (with 2 mg NiStNPs). The scale bar indicates 1at 2 days post-infection. The genomic MOI (i.e., genomic viral particles/cell) was 5 � 104. Noinfection by 6xHis AAV/NiStNPs without magnetic forces and compared with infections btransduction as compared with 24 h-exposure to 6xHis AAV only was observed in HEK293

using AAV2 resulted in larger nanoparticles ranging from135.98 nm to 188.67 nm, presumably due to loose binding of AAV2to the NiStNPs or agglomerations of the AAV2/NiStNP complexescaused by non-specific interactions. Transmission electronmicroscopy was used to visualize the conjugation of Ni-NTA-biotinand the immobilization of 6xHis AAV vectors on the streptavidin-coated SPIONs (Fig. 2D and E). The NiStNPs displayed a sphericalmorphology with small agglomerated particles on the surface,which were expected to be Ni-NTA-biotin-streptavidin tags on theiron oxide nanoparticles (Fig. 2D). The particle sizes were approx-imately 100e120 nm, which is consistent with the dynamic lightscattering analysis shown in Fig. 2C. After immobilizing the 6xHisAAV vectors, the icosahedral shape of the 6xHis AAV attached to theNiStNPs (the arrow in Fig. 2E) was observed, confirming thesuccessful formulation of the 6xHis AAV/NiStNP complexes prior tocellular transduction.

3.2. The effect of SPION-mediated gene delivery on cell viability

Next, we assessed the influence of SPION-mediated AAVdelivery under magnetic forces on the metabolic functions of cellsby comparing cells exposed to the 6xHis AAV/NiStNP complexes

AV/NiStNP complexes with or without magnetic forces: (A) representative fluorescence00 mm. (B) The percentage of GFP-expressing HEK293T cells out of total number of cellste that the symbols * and ** indicate statistically significant differences compared withy 24 h-exposure to 6xHis AAV only, respectively (P < 0.05). However, no increasedT infection.

E. Kim et al. / Biomaterials 32 (2011) 8654e8662 8659

with those that were treated only with the virus for 24 h (Fig. 3Aand B). Human neural stem cells mostly maintained metabolicactivity upon exposure to a small quantity (<2 mg) of magneticallyguided 6xHis AAV/NiStNP complexes (Fig. 3B), but exposure ofHEK293T cells to even a small quantity (0.02e0.2 mg) of nano-particles yielded significantly reduced cellular metabolic activitiescompared with positive controls (i.e., virus only) (Fig. 3A). At thehighest quantity of nanoparticles (4 mg), substantial reductions inthe viability of hNSCs were observed, possibly due to the cytotox-icity of the nanoparticle itself or the presence of a greater numberof internalized nanoparticles within the cytoplasm [27]. The pres-ence of magnetic forces did not influence the cell viability of eithercell type.

3.3. Enhanced infection with 6xHis AAV/NiStNP complexes

SPION-mediated 6xHis AAV delivery under magnetic forcessignificantly enhanced the cellular transduction of HEK293T cellscompared with SPION-bound 6xHis AAV delivery withoutmagnetic forces (Fig. 4A and B). However, due to the intrinsicability of AAV2 vectors to induce high transduction efficiencies inHEK293T cells, magnetically guided 6xHis AAV vectors demon-strated only slightly improved or comparable transduction effi-ciencies compared with AAV2 vectors alone or with AAV2/NiStNPcomplexes. Additionally, NiStNP-mediated 6xHis AAV delivery

Fig. 5. Cellular transduction of hNSCs by AAV2/NiStNP complexes vs. that by 6xHis AAV/NiStshowing GFP-expressing hNSCs (with 2 mg NiStNPs). The scale bar indicates 100 mm. (B) Thinfection. The genomic MOI (i.e., genomic viral particles/cell) was 2 � 104. The symbols * andNiStNPs without magnetic forces and compared with infection by 24 h-exposure to 6xHis

induced a lower transduction efficiency than conventional trans-duction with AAV2 vectors, that is, a 24-hour exposure to viralvectors prior to medium replacement. The genomic multiplicity ofinfection (MOI) employed for infecting the HEK293T cells was5 � 104. The magnetically guided 6xHis AAV vector deliveryenhanced cellular transduction in a NiStNP dose-dependentmanner. Upon infection with AAV2 vectors, magnetically guideddelivery did not exhibit any trends of cellular transduction withvarious amounts of nanoparticles, but infectivities graduallyincreased as the exposure times increased, as expected. Non-specific interactions between AAV2 and the NiStNPs might resultin limited binding to or easy detachment from the nanoparticles,consequently leading these two species (i.e., AAV2 vs. nano-particles) to act independently within the magnetic fields.

The effect of magnetic guidance of 6xHis AAV/NiStNP complexes(genomic MOI: 2 � 104) on cellular transduction was maximizedwhen infecting non-permissive cell types, such as hNSCs (Fig. 5),which is consistent with previous studies [28e30]. Compared withnon-magnetically guided viral infection and 24-hour viral exposureprior to medium replacement (i.e., conventional transduction),magnetically guided 6xHis AAV vectors yielded 2e5 fold andapproximately 2 fold enhanced infectivities, respectively (Fig. 5Aand B). The levels of cellular transductionmediated bymagneticallyguided 6xHis AAV delivery (w50%) are comparable to thoseobtained by a previous study in which a novel AAV vector (i.e., AAV

NP complexes with or without magnetic forces: (A) representative fluorescence imagese percentage of GFP-expressing hNSCs out of the total number of cells at 2 days post-** indicate statistically significant differences compared with infection by 6xHis AAV/

AAV only, respectively (P < 0.05).

E. Kim et al. / Biomaterials 32 (2011) 8654e86628660

r3.45) with enhanced properties for NSC infection was developedby directed evolution [16]. Importantly, the magnetically guidedAAV delivery system induced similar cellular transduction in hNSCseven with a 4-fold decreased viral genomic MOI (2 � 104 MOI vs.1 � 105 MOI). Significantly improved gene correction frequencies

Fig. 6. Internalization of 6xHis AAV/NiStNPs into the cytoplasm and GFP expression exclusiand quantifying GFP expression, cells exposed at different viral exposures (2, 10, and 120 midishes. Representative images of the presence of fluorescently labeled NiStNPs (with 2 mg) wfield images taken to show cell morphologies were merged with the fluorescence images of tHEK293T cells (C) and hNSCs (D) after the internalization of viral vectors. The symbols * andNiStNPs (with the same quantity of NiStNPs) without magnetic forces (P < 0.05 and P < 0

within the NSCs were achieved with comparable transductionefficiencies (w50%) in the previous study; therefore, we can expecta future in potential stem cell applications of the magneticallyguided 6xHis AAV delivery system, such as in gene targeting, stemcell regulation, and creation of disease models.

vely produced by the internalized viral vectors. Prior to visualizing the internalizationn) were vigorously rinsed with PBS, trypsinized, and finally replated onto fresh 24-wellithin the HEK293T (A) and hNSC (B) cytoplasm. To visualize the cells effectively, brighthe NiStNPs (red color). The scale bar indicates 50 mm. The percentage of GFP-expressing** indicate statistically significant differences compared with infections by 6xHis AAV/.01, respectively).

Fig. 7. Gene expression of hNSCs after viral exposure at different temperatures (4 �Cvs. 37 �C). To investigate the energy-independent cell entry of magnetically guidedviral vectors at 4 �C, hNSCs were exposed to viral vectors for 2, 10, or 120 min at 4 �C,the cell culture medium containing 6xHis AAV/NiStNPs (MOI 2 � 104, 2 mg NPs) wasreplaced with fresh medium, and hNSCs were cultured for additional 48 h at 37 �Cprior to flow cytometry analysis. The percentage of GFP-expressing cells was analyzedby flow cytometry. The data presenting GFP expression at 37 �C, which are shown inFig. 5B, are also presented here to effectively demonstrate the comparison with theexpression at 4 �C.

E. Kim et al. / Biomaterials 32 (2011) 8654e8662 8661

While AAV2 infection with or without NiStNP complexationresulted in low cellular transduction (<13%) without demon-strating improvement or reduction in transduction, NiStNP-guided6xHis AAV infection induced gradually enhanced transductionunder the influence of the magnetic field as the viral exposure timeincreased up to 10 min. However, as the quantity of NiStNPsincreased, cellular transduction decreased, presumably due to thecytotoxicity of the higher amounts of NiStNPs, as shown in Fig. 3.Interestingly, upon infection with 6xHis AAV/NiStNP complexes,even in the absence of magnetic forces, the longest exposure (i.e.,2 h) of 6xHis AAV/NiStNPs to the hNSCs led to a comparableenhancement of cellular transduction with magnetically guided6xHis AAV delivery approaches.

Importantly, the magnetically enhanced viral gene deliveryapproach significantly reduced the viral exposure times required fora high level of cellular transduction, especially in a non-permissivecell type. Conventional transduction using viral vectors typicallyrequires a 24-hour viral exposure tomaximize the efficiency of genedelivery. However, for NiStNP-mediated 6xHis AAV delivery undera magnetic field for hNSC infection, exposure times of only a fewminutes (2e10 min) were sufficient to yield 2e3.5 fold improvedtransduction comparedwith the conventional transductionmethod(i.e., 24-hourexposure). Aftera10minviral exposureof hNSCsunderthe magnetic field, further enhancement of cellular transductionwas not observed (Fig. 5B), indicating that a 10 min viral exposureunder the magnetic field was fully sufficient for hNSC infection. Theenhanced transduction of a non-permissive cell type even whenusing a short viral exposure time may be due to the rapid sedi-mentation of the full vector dose on the target cells followed by theforced penetration of the vector across the plasma membrane withthe aid of the NiStNPs acted upon by the magnetic field [31e33].

3.4. Internalization of 6xHis AAV/NiStNP

Substantial barriers to AAV infection of hNSCs may includesurface binding, intracellular trafficking, and/or nuclear entry[34,35]. AAV2 utilizes heparan sulfate proteoglycan as its primaryreceptor [36], and this molecule is present on every cell type inorganisms with tissue organization, possibly including neural stemcells [37,38]. Hexa-histidine AAV vectors, which still exhibit heparinaffinity [20], also demonstrated low delivery efficiencies in hNSCtransduction, as did wild-type AAV2 (Fig. 5). A recent studydemonstrated that surface modification of AAV2 vectors bydirected evolution significantly improved transduction efficienciesfor hNSCs, implying that key blocking steps against AAV delivery inhNSCs may be related to the external surface of the AAV vectors[16]. Therefore, the low transduction efficiency of AAV2 and 6xHisAAV for hNSCs may be caused by the reduced level of HSPG on thehNSCs or by the requirement of an alternate secondary receptor tobe internalized, both of which are related to the entry of viralvectors across the plasma membrane. Consequently, the applica-tion of a method to cross the cellular membrane is expected to bea key strategy in enhancing cellular transduction in hNSCs.Consistently, the rapid accumulation of viral vectors on the hNSCsfollowed by direct penetration across the cellular membrane by themagnetically guided gene delivery approach resulted in signifi-cantly improved cellular transduction.

Magnetically guided 6xHis AAV/NiStNP complexes were inter-nalized rapidly compared with non-magnetically deliveredcomplexes, ultimately resulting in a high level of transduction. Toverify the rapid internalization of 6xHis AAV/NiStNP complexesunder magnetic fields, each cell type (HEK293T, hNSC) was exposedto the complexes for designated time points (i.e., 2, 10, and 120min)and then trypsinizedandreplatedonto a freshdish. Trypsinizing cellsprovides a typical way to remove the majority of proteins, including

surface-bound viral vectors, on the cellular membrane. Trypsiniza-tion can be used to investigate the effect of the entry of 6xHis AAV/NiStNP complexes via magnetic forces on the enhanced cellulartransduction. The presence of fluorescently labeled nanoparticleswithin cells was visualized at 24 h post-internalization. Additionally,gene expression exclusively produced by the internalized viralvectors was quantified using flow cytometry at 48 h post-infection.Consequently, large amounts of fluorescently labeled nanoparticles,whichweredeliveredunder themagneticfield,weredetectedwithinthe cytoplasm as the exposure time increased (Fig. 6A and B). Theadvantageouseffectof rapid internalizationbymagnetic guidanceoncellular transduction was not substantial in a permissive cell type(HEK293T), but it was maximized in non-permissive hNSCs (Fig. 6CandD). Additionally, a 10minviral exposure under themagneticfieldduring hNSC infection caused a dramatic increase in internalizationfollowed by transduction compared with non-magnetically guided6xHis AAV delivery (Fig. 6D). A large quantity of 6xHis AAV vectorswas rapidly internalized into the hNSCs as the result of themagneticguidance of NiStNPs under magnetic fields, with approximately5e100 foldenhancedGFPexpressioncomparedwith the internalizedAAV2 vectors (Fig. 6B and D). An approximate 2 fold increase in GFPexpression forhNSCs shown in Fig. 5 comparedwith gene expressiondemonstrated in Fig. 6 may have resulted from the removal ofsurface-associated viral vectors, specifically non-internalized viralvectors, upon trypsinization for replating onto a new tissue culturedish.

The entry of magnetically guided 6xHis AAV particles wasindependent of the energy-dependent endocytosis pathway, asdemonstrated by the fact that cellular transduction bymagneticallyguided 6xHis AAV vectors at 4 �C was comparable with that at 37 �C(Fig. 7). Therefore, these results suggest that direct penetration ofviral vectors by magnetic forces can be a key factor in enhancingcellular transduction in non-permissive cell types, consistent withthe results of previous studies [28,30,39,40].

4. Conclusion

We reported a highly efficient AAV-mediated gene deliverysystem for the transduction of non-permissive cell types, such as

E. Kim et al. / Biomaterials 32 (2011) 8654e86628662

hNSCs, by employing magnetic forces. Magnetically guided 6xHisAAV delivery induced rapid internalization for the subsequentsignificantly enhanced cellular transduction. The successful estab-lishment of an AAV delivery system with enhanced abilities toefficiently infect stem cells will provide a powerful means toregulate stem cell functions. Thus, this technique will havenumerous applications that range from basic stem cell biology tostem cell therapy or the development of disease models.

Conflict of interestThe authors confirm that there are no known conflicts of

interest associated with this publication.

Acknowledgments

This work was supported by Advanced Biomass R&D Center(ABC) of Korea (Grant No. 2010-0029734) and a National ResearchFoundation (NRF) grant funded by the Korea government (MEST)through the Active Polymer Center for Pattern Integration (No. R11-2007-050-00000-0). K.I.P. was supported by National ResearchFoundation grant (2010-0020289) by Korean Government.

References

[1] Maitra A, Arking DE, Shivapurkar N, Ikeda M, Stastny V, Kassauei K, et al.Genomic alterations in cultured human embryonic stem cells. Nat Genet2005;37:1099e103.

[2] Kim JH, Auerbach JM, Rodriguez-Gomez JA, Velasco I, Gavin D, Lumelsky N,et al. Dopamine neurons derived from embryonic stem cells function in ananimal model of Parkinson’s disease. Nature 2002;418:50e6.

[3] Lai K, Kaspar BK, Gage FH, Schaffer DV. Sonic hedgehog regulates adult neuralprogenitor proliferation in vitro and in vivo. Nat Neurosci 2003;6:21e7.

[4] Young RA. Control of the embryonic stem cell state. Cell 2011;144:940e54.[5] Jo J, Tabata Y. Non-viral gene transfection technologies for genetic engineering

of stem cells. Eur J Pharm Biopharm 2008;68:90e104.[6] Singer O, Tiscornia G, Ikawa M, Verma IM. Rapid generation of knockdown

transgenic mice by silencing lentiviral vectors. Nat Protoc 2006;1:286e92.[7] Marumoto T, Tashiro A, Friedmann-Morvinski D, Scadeng M, Soda Y, Gage FH,

et al. Development of a novel mouse glioma model using lentiviral vectors.Nat Med 2009;15:110e6.

[8] Pfeifer A, Ikawa M, Dayn Y, Verma IM. Transgenesis by lentiviral vectors: lackof gene silencing in mammalian embryonic stem cells and preimplantationembryos. Proc Natl Acad Sci U S A 2002;99:2140e5.

[9] Gropp M, Itsykson P, Singer O, Ben-Hur T, Reinhartz E, Galun E, et al. Stablegenetic modification of human embryonic stem cells by lentiviral vectors. MolTher 2003;7:281e7.

[10] Suzuki K, Mitsui K, Aizawa E, Hasegawa K, Kawase E, Yamagishi T, et al. Highlyefficient transient gene expression and gene targeting in primate embryonicstem cells with helper-dependent adenoviral vectors. Proc Natl Acad Sci U S A2008;105:13781e6.

[11] Thomas CE, Ehrhardt A, Kay MA. Progress and problems with the use of viralvectors for gene therapy. Nat Rev Genet 2003;4:346e58.

[12] Kaplitt MG, Leone P, Samulski RJ, Xiao X, Pfaff DW, O’Malley KL, et al. Long-term gene expression and phenotypic correction using adeno-associated virusvectors in the mammalian brain. Nat Genet 1994;8:148e54.

[13] Flannery JG, Zolotukhin S, VaqueroMI, LaVailMM,MuzyczkaN,HauswirthWW.Efficient photoreceptor-targeted gene expression in vivo by recombinantadeno-associated virus. Proc Natl Acad Sci U S A 1997;94:6916e21.

[14] Fisher KJ, Jooss K, Alston J, Yang Y,Haecker SE, HighK, et al. Recombinant adeno-associated virus for muscle directed gene therapy. Nat Med 1997;3:306e12.

[15] Russell DW, Hirata RK. Human gene targeting by viral vectors. Nat Genet1998;18:325e30.

[16] Jang JH, Koerber JT, Kim JS, Asuri P, Vazin T, Bartel M, et al. An evolved adeno-associated viral variant enhances gene delivery and gene targeting in neuralstem cells. Mol Ther 2011;19:667e75.

[17] Hughes SM, Moussavi-Harami F, Sauter SL, Davidson BL. Viral-mediated genetransfer to mouse primary neural progenitor cells. Mol Ther 2002;5:16e24.

[18] Smith-Arica JR, Thomson AJ, Ansell R, Chiorini J, Davidson B, McWhir J.Infection efficiency of human and mouse embryonic stem cells using adeno-viral and adeno-associated viral vectors. Cloning Stem Cells 2003;5:51e62.

[19] Kim HT, Kim IS, Lee IS, Lee JP, Snyder EY, Park KI. Human neurospheresderived from the fetal central nervous system are regionally and temporallyspecified but are not committed. Exp Neurol 2006;199:222e35.

[20] Koerber JT, Jang JH, Yu JH, Kane RS, Schaffer DV. Engineering adeno-associatedvirus for one-step purification via immobilized metal affinity chromatog-raphy. Hum Gene Ther 2007;18:367e78.

[21] Maheshri N, Koerber JT, Kaspar BK, Schaffer DV. Directed evolution of adeno-associated virus yields enhanced gene delivery vectors. Nat Biotechnol 2006;24:198e204.

[22] Koerber JT, Maheshri N, Kaspar BK, Schaffer DV. Construction of diverseadeno-associated viral libraries for directed evolution of enhanced genedelivery vehicles. Nat Protoc 2006;1:701e6.

[23] Zolotukhin S, Byrne BJ, Mason E, Zolotukhin I, Potter M, Chesnut K, et al.Recombinant adeno-associated virus purification using novel methodsimproves infectious titer and yield. Gene Ther 1999;6:973e85.

[24] Jang JH, Koerber JT, Gujraty K, Bethi SR, Kane RS, Schaffer DV. Surfaceimmobilization of hexa-histidine-tagged adeno-associated viral vectors forlocalized gene delivery. Gene Ther 2010;17:1384e9.

[25] Jiang X, Zuber A, Heberle J, Ataka K. In situ monitoring of the orientatedassembly of strep-tagged membrane proteins on the gold surface by surfaceenhanced infrared absorption spectroscopy. Phys Chem Chem Phys 2008;10:6381e7.

[26] Smith BM, Lappi SE, Brewer SH, Dembowy S, Belyea J, Franzen S. Covalentattachment of a nickel nitrilotriacetic acid group to a germanium attenuatedtotal reflectance element. Langmuir 2004;20:1184e8.

[27] Song HP, Yang JY, Lo SL, Wang Y, Fan WM, Tang XS, et al. Gene transfer usingself-assembled ternary complexes of cationic magnetic nanoparticles, plasmidDNA and cell-penetrating tat peptide. Biomaterials 2010;31:769e78.

[28] Orlando C, Castellani S, Mykhaylyk O, Copreni E, Zelphati O, Plank C, et al.Magnetically guided lentiviral-mediated transduction of airway epithelialcells. J Gene Med 2010;12:747e54.

[29] Ito A, Takahashi T, Kameyama Y, Kawabe Y, Kamihira M. Magnetic concen-tration of a retroviral vector using magnetite cationic liposomes. Tissue EngPart C Methods 2009;15:57e64.

[30] Haim H, Steiner I, Panet A. Synchronized infection of cell cultures bymagnetically controlled virus. J Virol 2005;79:622e5.

[31] Dobson J. Gene therapy progress and prospects: magnetic nanoparticle-basedgene delivery. Gene Ther 2006;13:283e7.

[32] Plank C, Schillinger U, Scherer F, Bergemann C, Remy JS, Krotz F, et al. Themagnetofection method: using magnetic force to enhance gene delivery. BiolChem 2003;384:737e47.

[33] Chorny M, Polyak B, Alferiev IS, Walsh K, Friedman G, Levy RJ. Magneticallydriven plasmid DNA delivery with biodegradable polymeric nanoparticles.FASEB J 2007;21:2510e9.

[34] Girod A, Ried M, Wobus C, Lahm H, Leike K, Kleinschmidt J, et al. Geneticcapsid modifications allow efficient re-targeting of adeno-associated virustype 2. Nat Med 1999;5:1052e6.

[35] Hansen J, Qing K, Kwon HJ, Mah C, Srivastava A. Impaired intracellular traf-ficking of adeno-associated virus type 2 vectors limits efficient transduction ofmurine fibroblasts. J Virol 2000;74:992e6.

[36] Opie SR, Warrington Jr KH, Agbandje-McKenna M, Zolotukhin S, Muzyczka N.Identification of amino acid residues in the capsid proteins of adeno-associated virus type 2 that contribute to heparan sulfate proteoglycanbinding. J Virol 2003;77:6995e7006.

[37] Dietrich CP, Tersariol IL, Toma L, Moraes CT, Porcionatto MA, Oliveira FW, et al.Structure of heparan sulfate: identification of variable and constant oligo-saccharide domains in eight heparan sulfates of different origins. Cell Mol Biol(Noisy-le-grand) 1998;44:417e29.

[38] Araujo AP, Ribeiro ME, Ricci R, Torquato RJ, Toma L, Porcionatto MA. Glial cellsmodulate heparan sulfate proteoglycan (HSPG) expression by neuronalprecursors during early postnatal cerebellar development. Int J Dev Neurosci2010;28:611e20.

[39] Kamei K, Mukai Y, Kojima H, Yoshikawa T, Yoshikawa M, Kiyohara G, et al.Direct cell entry of gold/iron-oxide magnetic nanoparticles in adenovirusmediated gene delivery. Biomaterials 2009;30:1809e14.

[40] Tresilwised N, Pithayanukul P, Mykhaylyk O, Holm PS, Holzmuller R, Anton M,et al. Boosting oncolytic adenovirus potency with magnetic nanoparticles andmagnetic force. Mol Pharm 2010;7:1069e89.