Capsid modifications overcome low heterogeneous expression of heparan
sulfate proteoglycan that limits AAV2-mediated gene transfer and therapeutic
efficacy in human ovarian carcinoma
Wenfang Shia, Akseli Hemminkib, Jeffrey S. Bartletta,c,d,⁎
aDivision of Molecular Medicine, Department of Pediatrics, College of Medicine and Public Health,
The Ohio State University, Columbus, OH 43210-1393, USA
bCancer Gene Therapy Group, Rational Drug Design Program, University of Helsinki and Department of Oncology,
Helsinki University Central Hospital, Helsinki, Finland
cGene Therapy Center, Columbus Children’s Research Institute, Columbus Children's Hospital, Columbus, OH 43205, USA
dDepartment of Molecular Virology, Immunology, and Medical Genetics, College of Medicine and Public Health,
The Ohio State University, Columbus, OH 43210-1393, USA
Received 17 March 2006
Available online 25 July 2006
Objectives. Capsid-modified AAV vectors can mediate enhanced gene transfer to neoplasms characterized by low AAV receptor expression.
Here we sought to determine the therapeutic potential of a capsid-modified AAV vector for gene therapy of ovarian carcinoma (OvCa).
Methods. We tested a panel of OvCa cell lines for AAV2-mediated gene transduction and for sensitivity to ganciclovir (GCV) following
AAVHSVtk administration. Levels of AAV internalization and attachment receptor were assessed by flow cytometry and immunohistochemistry.
The role of receptors in AAV-mediated gene transfer was assessed by competition assays. Finally, we examined the ability of a modified vector
with an integrin-binding RGD motif inserted into the AAV capsid to improve gene delivery to OvCa and enhance AAVHSVtk/GCV-mediated
killing by cytotoxicity assay.
Results. All OvCa cell lines were poorly transduced with AAV2 vectors and showed variably sensitive to AAVHSVtk/GCV. While OvCa cell
lines expressed AAV2 internalization receptors (αvintegrins), expression of the AAV2 attachment receptor, HSPG, was variable and not detected on
many lines. Analysis of archived clinical specimens showed no detectable HSPG expression on approximately 45% of primary human tumors. Gene
transfer to OvCa was increased several fold using the RGD-modified vector. Gene transfer was independent of HSPG and specific to the targeted
receptor. Importantly, the RGD-modified capsid markedly increased the ability of the AAVHSVtk to kill OvCa cells in the presence of GCV.
Conclusions. The development of AAV vectors targeted to cell surface receptors other than HSPG will be critical to the advancement of AAV-
mediated gene therapy for treating OvCa.
© 2006 Elsevier Inc. All rights reserved.
Keywords: Gene therapy; Ovarian cancer; AAV vector targeting; Heparan sulfate proteoglycan (HSPG); Integrin
Ovarian cancer (OvCa) remains a leading cause of
women's death due to gynecologic malignancy. A women's
average lifetime risk for developing ovarian cancer is about 1
in 70. Despite the frequency of temporary remission
achieved with surgical cytoreduction and multiple-agent
chemotherapy, long-term survival has improved little in the
past 30 years. The majority of ovarian cancers are not
diagnosed until the disease has reached an advanced stage.
Late-stage disease 5-year survival rates are only 28%. The
annual U.S. mortality from ovarian cancer is nearly 15,000,
and worldwide nearly 150,000 women die from this disease
each year. More effective systemic therapies are clearly
Gynecologic Oncology 103 (2006) 1054–1062
⁎Corresponding author. Gene Therapy Center, WA3016, Columbus Chil-
dren's Research Institute, 700 Children's Drive, Columbus, OH 43205, USA.
Fax: +1 614 722 3273.
E-mail address: BartletJ@pediatrics.ohio-state.edu (J.S. Bartlett).
0090-8258/$ - see front matter © 2006 Elsevier Inc. All rights reserved.
The major promise of cancer gene therapy lies in the
potential to specifically target gene delivery and expression to
defined cell populations, thereby avoiding unwanted side
effects. Although a variety of gene transfer methods are
available AAV vectors have not been used very extensively,
for cancer gene therapy despite the fact that these vectors may
offer several advantages. First, due to a decade of continued
improvement in production strategies, AAV vectors can now be
generated at very high titers [1,2]. Furthermore, these vectors
are capable of transducing a wide variety of cell types [3–11],
transduce transformed cells more efficiently than non-trans-
formed cells [12,13], and promote stable gene expression
[14,15]. Additionally, the small size of AAV makes it attractive
as a vector for gene delivery to solid tumors where it is thought
that its size might allow more efficient tumor penetration.
There are at least eleven molecularly cloned serotypes of
AAV that have been adapted as gene transfer vectors. However,
vectors based on AAV type-2 (AAV2) are the most widely used
and best characterized. AAV2 attachment to cells is mediated by
a surface exposed loop within the AAV capsid protein that
projects outward from the particle surface and binds to a cell
surface heparan sulfate proteoglycan (HSPG) epitope [16,17].
Recently, through crystal structure and sequence mutational
analysis, capsid amino acid residues critical for binding HSPG
have been identified [18,19]. Insertion of foreign peptide
epitopes near these critical HSPG-binding residues can reduce
transduction of cells that were otherwise highly susceptible to
wild-type AAV2 gene transfer, while providing a mechanism
for transduction of target cells mediated by the inserted peptide
epitope [20–24]. The availability of such mutants should allow
both the determination of the role of HSPG for AAV gene
transfer in different cell types and the opportunity to divert AAV
transduction or infection away from being mediated by HSPG.
The observation that HSPG is expressed in many tissues may
in part explain the broad tropism of AAV2. Nevertheless, there
are many cell types that are poorly transduced by AAV2,
including some tumor cells. In these cases, transduction may be
limited by the requirement for AAV second-strand DNA
synthesis [25,26], aberrant intracellular trafficking of viral
particles [27–29], or failure to properly uncoat vector genomes
in the nucleus . In other cases, AAV gene transfer may be
limited solely due to the lack of receptor expression, or
expression of non-AAV-binding HSPG epitopes. Recently we
haveshownthat AAVparticles canbeengineeredinfectcellsvia
alternative cell surface receptors alleviating their dependence on
HSPG. In one instance, we successfully targeted AAV to
megakaryocytic leukemia cells deficient in HSPG expression by
conjugation with a bispecific antibody that linked AAV particles
to αIIbβ3integrins expressed on the surface of these cells.
transduction of cells that lack HSPG expression by providing a
foreign peptide ligand that binds to cell surface proteins other
than HSPG. The insertion of a peptide containing an arginine–
glycine–aspartate (RGD) integrin-binding site into a surface-
exposed loop of the AAV capsid markedly increased AAV
transduction of human lymphoblast and chronic myelogenous
leukemia cell lines . Gene transfer was also enhanced to
ovarian carcinoma cells in culture and in an in vivo xenograft
tumor model . Therefore, AAVs have the potential to be
genetically redirected in their delivery of therapeutic genes.
In prodrug-activated or “suicide” gene therapy, a gene
coding for an enzyme that converts a non-toxic prodrug into a
toxic compound is transferred into cancer cells . The most
widely used system involves the herpes simplex virus—1
thymidine kinase gene (HSVtk) and the prodrug ganciclovir
(GCV). HSVtk functions by selectively phosphorylating GCV
to a monophosphate form. Cellular kinases further phosphor-
ylate the monophosphate to di- and triphosphate metabolites;
the triphosphorylated form is the active compound. The
incorporation of triphosphorylated GCV into DNA results in
chain termination and consequently cell death . Cells
without the HSVtk gene cannot phosphorylate GCV and are
therefore insensitive to it. However, transfer of the suicide gene
into tumor cells results in tumor destruction following GCV
treatment . The major hurdle to date of this cancer gene
therapy strategy is poor gene transfer efficiency, which has
limited therapeutic efficacy, particularly in vivo. Although
progress has been made, no reliable method exists for achieving
efficient gene transfer to cancer cells in a tumor cell mass or the
majority of metastatic sites in vivo.
One OvCa cell line, SKOV3.ip1, has been previously shown
to lack HSPG expression and to be very poorly transduced by
AAV2-based vectors . In an effort to determine whether
AAV-mediated HSVtk gene expression in conjunction with
GCV therapy would be a feasible strategy by which to target
destruction of OvCa cells, we tested the effect of AAV2HSVtk/
GCV on SKOV3.ip1 and other OvCa cells in culture for
efficiency of both gene transfer and cell killing. We found that
OvCa cells were variably susceptible to AAV2HSVtk-mediated
killing. Although the majority of the OvCa cell lines we
examined expressed the AAV internalization receptors, αv
integrins, they either failed to express HSPG, the AAV2
attachment receptor, or expressed only low levels. Primary
human tumors also showed variable expression of HSPG by
immunocytochemistry. AAV2-mediated gene transfer to OvCa
celllineswasblockedby solubleheparinsulfate, suggesting that
HSPG expression mediated the low levels of AAV transduction
that were observed. AAVs comprised of modified capsids that
bypass HSPG were capable of higher levels of gene transfer for
OvCa cells and permitted gene delivery to previously non-
permissive OvCa cell lines. Finally, an AAV vector with a
AAV-588RGD-4C-HSVtk vector to kill OvCa cell in the
presence of GCV and provides a promising foundation onto
which to build future OvCa-directed gene therapies.
Materials and methods
Cell lines and recombinant AAV vectors
The human ovarian adenocarcinoma cell lines OVCAR-3, OVCAR-3N,
Hey, OV4, SKOV3.ip1, and OV3, and the teratocarcinoma cell line PA-1 have
been described previously [34,35]. Cells were maintained in Dulbecco's
modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated
fetal bovine serum (FBS), penicillin (100 U/ml) and streptomycin (100 U/ml),
1055W. Shi et al. / Gynecologic Oncology 103 (2006) 1054–1062
except for OVCAR-3 cells which were grown in RPMI 1640 medium with
2 mM L-glutamine adjusted to contain 1.5 g/l sodium bicarbonate, 4.5 g/l
glucose, 10 mM HEPES, and 1.0 mM sodium pyruvate and supplemented with
0.01 mg/ml bovine insulin and 20% fetal bovine serum. All media and
supplements were obtained from BRL Life-Technologies Invitrogen. Low
passage number (passage number 20–40) HEK 293 cells , HeLa cells, and
HeLa C12 cells  were grown in DMEM containing 10% fetal bovine serum,
penicillin (100 U/ml) and streptomycin (100 U/ml), at 37°C and 5% CO2.
Recombinant AAV vectors containing the enhanced green fluorescent
protein gene (eGFP), or herpes simplex virus thymidine kinase gene (HSVtk),
under control of the human CMV IE promoter/enhancer region were packaged
into either wild-type AAV2 capsids or A588-RGD-4C capsids as described
previously . A588-RGD-4C vectors contain a RGD peptide inserted in a
surface-exposed loop of the virus capsid protein enabling these vectors to infect
integrin-expressing cells independent of HSPG. Vectors were produced from
adenovirus infected packaging cell lines as previously described . Forty-
eight hours after adenovirus infection, cells were harvested by centrifugation at
500×g for 10 min, resuspended in PBS, and recombinant virus was released by
freezing and thawing three times. The crude lysate was clarified by
centrifugation at 500×g for 10 min and treated with benzonase at 250 U/ml
final concentration at 37°C for 30 min. Virus was further purified by iodixanol
step gradient and heparin sulfate affinity chromatography [2,39,40], and stored
at −20°C in PBS containing 20% glycerol. Particle titers, expressed as DRP
(DNase-resistant particles), were determined by ELISA and DNA dot blot as
previously described , and infectious titers were determined by gene
transduction assay on HeLa C12 cells in the presence of adenovirus type-5
(Ad5) at 2 IU/cell as described previously .
Flow cytometric analysis of receptor expression and gene transduction
Flow cytometry analysis of HSPG expression in tumor cell lines using the
HepSS-1 mouse monoclonal antibody (Seikagaku America, Falmouth, MA) has
been described previously . Similarly, analysis of integrin expression using
FITC-labeled LM609 and PIF6 antibodies (Chemicon International, Inc.,
Temecula, CA) has also been described . Anti-bacterial β-galactosidase
monoclonal antibody was used as an isotype-matched (IgG1) control and was
also purchased from Chemicon.
For analysis of gene transduction, cells were seeded in 24-well plates so that
they would reach about 75% confluence or about 5×105cells/ml at the time of
infection. Serial dilutions of wild-type AAV2eGFP and RGD-mutant AAVeGFP
preparationswere added to the cellsin the presence or absenceof Ad5 at an MOI
of 3 IU/cell. The cells and viruses were incubated at 37°C for 48 h, after which
time the media were removed and the cells were washed twice with PBS, fixed,
and analyzed for GFP gene expression by flow cytometry. Data have been
presented as the percentage of transduced cells in the cell population infected at
the indicated particle multiplicity.
Analysis of sensitivity to GCV/AAV-HSVtk in vitro
Cells transduced with AAV-HSVtk were tested for sensitivity to GCV
(InvivoGen, San Diego, CA) in a standard MTT-based cell proliferation assay
. Briefly, OvCa cells were seeded and cultured in 96-well plates at a density
of 2.5×103cells well for 24 h. The cells were then infected with either AAV2-
HSVtk, RGD-AAV-HSVtk, or control AAV2eGFP at multiplicity of 1000 DRP/
cell. After 24-h incubation, the medium was replaced with fresh medium
containing various concentrations of GCV. The cells were then cultured at 37°C
for another 3, 5, or 7 days and the number of viable cells was assessed by MTT
assay (CellTiter 96 Non-Radioactive Cell Proliferation Assay, Promega,
Madison, WI). Cytotoxicity results are expressed as a ratio of the number of
cells in wells containing drug as a percentage of that in corresponding drug-free
controls. Data in all experiments represent the mean of three to five samples for
Clinical samples of OvCa were obtained from the Cooperative Human
Tissue Network (CHTN) at Columbus Children's Hospital. Samples were
analyzed anonymously as approved by the Children's Hospital Institutional
Review Board. Sections of paraffin blocks were cut at 5 μm onto precoated
slides and heated at 65°C for 1 h. Paraffin was extracted with xylene and the
slides were then pre-treated with 3% hydrogen peroxide for 15 min at 20°C.
Anti-HSPG antibody (HepSS-1), anti-αvβ3 integrin monoclonal antibody
LM609, or anti-αvβ5 integrin monoclonal antibody P1F6 were applied to
parallel sections at various dilutions for 2 h. Sections were developed using
appropriate secondary antibodies, and signal was detected using streptavidin–
biotin and diaminobenzidine (DAB; Vector Laboratories, Burlingame, CA).
Sections were counterstained with hematoxylin, dehydrated with xylene,
mounted (Permount), and examined using a Nikon E600 microscope.
Ovarian cancer cell lines show low and variable transduction
To determine whether AAV2-mediated gene transfer would
be useful as a gene delivery strategy for the treatment of ovarian
cancer, Ad5 infected OvCa cell lines were exposed to
AAV2eGFP vector. HeLa C12 cells, which are fully permissive
for AAV2 infection in the presence of adenovirus, were used as
a positive control in these experiments. Adenovirus co-infection
was used to abrogate the requirement for AAV second-strand
DNA synthesis,thereby, allowing a true assessment of the initial
steps of viral infection independent of intracellular blocks. The
AAV vector used expresses the enhanced green fluorescent
protein (eGFP) gene under control of the strong immediate early
promoter from CMV. Therefore, 48 h after infection we could
harvest the OvCa cell lines and measure eGFP gene transduc-
tion easily by FACS. As shown in Fig. 1, the OvCa cell lines we
tested showed variable susceptibility to AAV2-mediated gene
transfer. Specifically, the cell lines Hey and OV3 were not
transduced at all, and the cell line, SKOV3.ip1, was transduced
only slightly above background. Even at 10-fold higher viral
load only a small percentage of these cells could be transduced
(data not shown). In contrast, the remaining four OvCa cell lines
Fig. 1. Susceptibility of OvCa cells to AAV vector-mediated gene transduction.
Adenovirus infected (3 infectious units (IU)/cell) OvCa cells were exposed to
AAV2eGFP vector at 500 DNase-resistant particles (DRP)/cell for 2 h and
harvested at 48 h. Gene transduction was assessed by flow cytometry. The
percentage of transduced cells in the population is shown as the average for
triplicate points (bars, SD).
1056W. Shi et al. / Gynecologic Oncology 103 (2006) 1054–1062
were susceptible to AAV2-mediated gene transfer. However,
transduction of these cell lines was still less efficient than that
observed for the HeLa C12 cells. These results demonstrate that
there is a wide range of AAV2 transduction potential among
OvCa cells in culture, with many OvCa cell lines being
completely non-permissive to AAV2 transduction and others
being very poorly transduced.
Ovarian cancer cells are variably susceptible to AAV2-HSVtk/
Because of the potential utility of AAV-mediated HSVtk
gene expression as an anticancer strategy , we sought to
determine whether AAV2-HSVtk-trasduced OvCa cell lines
were susceptible to killing when treated with GCV. Using cell
lines that showed different levels of AAV2-mediated eGFP gene
transduction (very low, SKOV3.ip1; intermediate OV4; and
high, PA-1), we assessed the therapeutic efficacy of AAV2-
HSVtk/GCV by examining cell killing. Viability was assessed
72 h, 5 days, and 1 week following infection and GCV
treatment by MTTassay. As expected, only those cell lines that
were permissive to AAV2eGFP transduction were sensitive to
the HSVtk-directed enzyme prodrug therapy (Fig. 2). SKOV3.
ip1 cells were resistant to killing, whereas PA-1 and OV4 cell
lines, which were transduced by AAV2 (Fig. 1), were
effectively killed by AAV2-HSVtk/GCV treatment. Never-
theless, killing of OvCa cell lines was significantly less efficient
than killing of control HeLa C12 cells. No cell killing was seen
following AAV2eGFP infection and GCV treatment with any
cell line. These results demonstrate that while AAV2-mediated
HSVtk gene expression might be an appropriate anti-cancer
strategy for some ovarian tumors, its applicability is limited due
to the poor efficiency of AAV2-mediated OvCa cell transduc-
tion and the inability of AAV2-based vectors to transduce a
wider variety of ovarian cancers.
Ovarian cancer cells express the AAV internalization receptor
but variably express the attachment receptor
Since the AAV2-mediated eGFP gene transduction studies
described above were carried out in the presence of adenovirus,
which should have abrogated any intracellular blocks to
efficient AAV2-mediated gene transduction (e.g., second-strand
DNA synthesis , or nuclear entry ), the most likely
cause of poor AAV gene transfer and therapeutic efficacy of
AAV2HSVtk/GCV was thought to be low expression of the
virus receptors. AAV2 binds cells primarily through the
interaction of its capsid with cell surface HSPG. Internalization
of the virus is then facilitated by the presence of cell surface
members of the αvintegrin family , however, AAV2 capsids
do not bind directly to these co-receptors . We tested the
expression of these proteins on the surface of our panel of OvCa
cells lines in an attempt to correlate their expression with
AAV2-mediated gene transduction and AAV2-HSVtk/GCV-
mediated killing. Flow cytometry profiles of the different cell
lines using an anti-HSPG antibody (HepSS-1) and anti-αvβ3
and αvβ5antibodies (LM609 and P1F6) are shown in Fig. 3.
HeLa cells served as positive controls and Raji cells served as a
negative control for HSPG expression . OvCa cell lines
expressed widely variable levels of HSPG relative to the control
cells. Expression of HSPG was extremely low on OV3 and Hey
cells and only slightly higher on SKOV3.ip1 cells. Not
surprisingly, the level of HSPG expression on the OvCa cell
lines correlated with susceptibly to gene transfer. With only
those cell lines that expressed appreciable levels of HSPG being
effectively transduced with AAV2 vectors. In contrast to HSPG,
αvintegrins (either αvβ3, αvβ5, or both) were expressed on the
majority of the ovarian cancer cell lines tested.
Variable HSPG expression is a feature of primary ovarian
We next sought to determine whether variable HSPG
expression is a feature of primary ovarian carcinomas or a
phenomenon of cells in culture. Therefore, we analyzed primary
clinical samples by immunohistochemistry with the anti-HSPG
antibody, HepSS-1. Archived ovarian cancer samples were first
reviewed to confirm the histological diagnosis and subtype. We
analyzed 40 different primary clinical samples of ovarian
carcinoma obtained from the CHTN. Approximately 50% of
these carcinomas showed membranous staining, indicating
expression of HSPG. Staining was variable among samples, but
no attempt was made to quantify staining intensity. About 5% of
samples showed very weak membranous or cytoplasmic
staining, which was thought to be a non-specific finding due
the staining of extracellular HSPG, or resulting from the high
levels of HepSS-1 antibody required for detection of HSPG in
some tissue samples. 45% of tumor samples were negative for
HSPG expression. Interestingly, nearly all (38 out of 40; 95%)
Fig. 2. In vitro GCV sensitivity assay for OvCa cells transducted with
AAV2HSVtk vector. Cells were seeded into 96-well plates at a density of
2.5×103cells well for 24 h and exposed to AAV2-HSVtk vector at multiplicity
of 1000 DRP/cell. After 24-h incubation, the medium was replaced with fresh
medium containing 20 μg/ml GCV. The cells were then cultured at 37°C for
another 3, 5, or 7 days and the number of viable cells was assessed by MTT
assay. Cytotoxicity results are expressed as the number of cells in wells
containing drug subtracted from the number of cells in corresponding drug-free
controls (bars, SD). Cytotoxicity results for cells transduced with AAV2eGFP
control vector and treated with GCV were less than 2.0% for all cell lines at all
1057W. Shi et al. / Gynecologic Oncology 103 (2006) 1054–1062
staining with isotype-matched (IgG1) control antibody. HeLa cells were used as a positive control, and Raji was a negative control for HSPG expression.
1058W. Shi et al. / Gynecologic Oncology 103 (2006) 1054–1062
of the primary clinical OvCa samples expressed high levels of
αvintegrins (either αvβ3, αvβ5, or both), in line with previous
reports [44,45]. Therefore variable HSPG expression, and
moderate αvintegrin expression, are features of both primary
ovarian carcinomas and established OvCa cell lines.
HS competition defines the role of receptors in AAV-mediated
gene transfer for OvCa
The variable expression of HSPG on OvCa cell lines was
thought to likely be responsible for the variable AAV2
transduction of these cell lines. However, it was also possible
that other intracellular blocks, not alleviated by adenovirus co-
infection (e.g., proteosomal degradation or altered intracellular
trafficking), could be limiting AAV2 transduction [27,28,46].
For this reason, we also carried at these experiments in the
presence of the proteosomal inhibitor LLnL (μM), which had
previously been shown to augment AAV2-mediated gene
transduction, but saw no effect (data not shown). We then
assessed gene transduction in the presence of soluble heparin
sulfate, a competitive inhibitor of the interaction between AAV2
and its primary attachment receptor, cell surface HSPG, to
determine whether the robust gene transfer seen in some OvCa
cells (Fig. 1) was in fact due to their residual HSPG expression
or other cell surface molecules.
As shown in Fig. 4, transduction of the OvCa cell lines
previously shown to be permissive for AAV2-mediated gene
transfer (PA-1, OVCAR-3, OVCAR-3N, and OV4) was
significantly inhibited in the presence of heparin sulfate. The
effect was equally dramatic in HeLa C12 cells which have
previously been shown to be transduced by AAV2-based
vectors in a HSPG-dependent manner. These studies support the
notion that residual HSPG plays a role in mediating AAV2 gene
transfer to OvCa cells, despite being expressed at low levels by
RGD capsid modification markedly improves transduction for
Our data suggest that lack of HSPG expression isa barrier for
AAV2 gene transfer to some of the OvCa cells. To further test
the role of HSPG in AAV2-mediated transduction of these cells,
we tested an AAV2 vector whose capsid had been modified to
contain an integrin-binding RGD peptide motif in a surface
exposed region of the virion. We have previously shown that
this vector is superior for gene transfer to other tumor cell lines
. Our hypothesis was that viral modifications that obviate
the requirement for HSPG by attaching to other surface
molecules might also enhance AAV2-mediated gene transfer
and AAV2HSVtk/GCV therapeutic efficacy. We thought it
likely that vectors modified to attach to cell surface molecules
more abundant than HSPG, such as RGD-binding integrins,
would provide more efficient gene delivery to HSPG-deficient
As shown in Fig. 5a, cell lines were transduced with the
modified vector (A588-4C-RGD capsid) or the control vector
(wild-type AAV2 capsid) at 500 DRP/cell (approximately 3 IU/
cell). For each cell line, the modified vector was able to
transduce a significantly greater number of cells than the
unmodified vector. In the OvCa cell lines, the fold increase was
highest for those cells that exhibited the worst transduction
efficiency with the unmodified vector. This is not surprising,
since these cells were shown to not express an attachment
receptor for the unmodified vector. Transduction of these cell
lines increased between 10- and 1000-fold with the RGD-
modified vector. Interestingly, transduction of all cell lines was
better with the modified vector than the wild-type vector
irrespective of their HSPG expression status suggesting either a
cooperative binding event, or a more efficient cell entry or
particle trafficking event coupled to the integrin receptor.
To confirm that infection of the OvCa cells by the RGD-
modified vector was proceeding via the targeted integrin
receptors, we utilized these vectors in an assay based on
competitive inhibition of AAV2-mediated gene delivery by
soluble heparin sulfate, a synthetic RGD peptide, and anti-
integrin antibody as described previously . As shown in Fig.
5b, eGFP expression in SKOV3.ip1 cells mediated by control
virus, AAV2eGFP with a wild-type capsid, was inefficient and
could be efficiently blocked by soluble heparin sulfate. In
marked contrast, A5884C-RGD capsid-mediated eGFP expres-
sion was significantly more efficient in these OvCa cells and
was only minimally affected by the competition with soluble
heparin sulfate, indicating that A5884C-RGD capsids are
capable of using an alternative, HSPG-independent cell entry
pathway to infect these cells. The contribution of that alternative
mechanism of cell entry was quite significant accounting for
entry of approximately 84% of total transducing vector
particles. To assess the specificity of this new interaction
infections were performed in the presence of either excess free
RGD peptide or anti-integrin antibody in conjunction with
Fig. 4. Effect of HS competition on AAV2-mediated gene transfer to OvCa cells.
Cellswereexposedfor 2h at4°Cto 500DRP/cellof AAV2eGFPinthe presence
or absenceof 500 μg/ml heparinsulfate (HS). Unbound viruswas then removed,
fresh medium was added, and the cells were analyzed for eGFP gene
transduction by FACS after 48 h. Data represent the means and standard
deviations of triplicate experiments.
1059 W. Shi et al. / Gynecologic Oncology 103 (2006) 1054–1062
soluble heparin sulfate. Both of these competitors were able to
significantly reduce A5884C-RGD-mediated gene expression
when utilized in conjunction with heparin sulfate, indicating the
specificity of the engineered interaction for RGD-binding
integrins expressed on the cell surface.
Low HSPG expression is the limitation in
AAVHSVtk/GCV-mediated killing of OvCa cells and is
significantly improved by RGD capsid modification
The test whether the improved transduction of OvCa cells
with the RGD-modified AAV2 capsid translated to improved
gene-directed enzyme prodrug therapy (GDEPT), we compared
AAV2HSVtk to a derivative containing the A588-4C-RGD
capsid modification, RGD-AAV2HSVtk. AAV2 HSVtk trans-
duced OvCa cell lines were variably resistant to GCV killing as
described above. However, RGD-AAV2HSVtk in combination
with GCV showed a marked increase in cytotoxic effect for all
three cell lines (Fig. 6). The effect was most dramatic in the
HSPG-deficient cell line (SKOV3.ip1), but was also significant
in the other cell lines, consistent with the higher gene transfer to
these cells. The enhanced killing of previously sensitive OvCa
cell lines is likely the result of either enhanced virus entry
(mediated by HSPG and RGD-binding integrins), or altered
intracellular trafficking of viral particles.
Here we show that human OvCa cell lines are variably
transduced by AAV2 vectors and are thereby variably
susceptible to killing mediated by AAV2 gene directed enzyme
pro-drug therapies (GDEPT). Variable AAV2 gene transduction
seems to be largely due to poor expression of the primary AAV2
attachment receptor, HSPG. There appears to be an absolute
requirement for HSPG expression for transduction of OvCa cell
lines with AAV2 vectors. However, other factors may also limit
AAV2 transduction in OvCa cells, since cell lines that express
high levels of HSPG are still less efficiently transduced than
non-OvCa cell lines. There is a direct correlation between
efficiency of transduction and cell killing mediated by
AAV2HSVtk/GCV and this appears to be directly related to
expression of cell surface HSPG which we show is variably
expressed on OvCa cell lines and primary ovarian tumors.
OvCa cell lines that expressed moderate levels of HSPG, could
be effectively transduced, and were effectively killed by
AAV2HSVtk/GCV. Whereas, OvCa cell lines that expressed
little,orno,HSPG failedtobetransduced andwereineffectively
Fig. 6. Effect of RGD capsid modification on AAVHSVtk/GCV-mediated
killing of OvCa cells. Cells were exposed to AAV2-HSVtk vector with either
RGD-modified (A588-4C-RGD capsid) capsid or unmodified (wild-type AAV2
capsid) capsid at multiplicity of 1000 DRP/cell. After 24-h incubation, the
medium was replaced with fresh medium containing 20 μg/ml GCV and the
number of viable cells was assessed at 7 days by MTT assay. Results are
expressed as described in the Fig. 2 legend (bars, SD).
Fig. 5. Effect of RGD capsid modification on AAV transduction of OvCa cells.
(a) Gene transfer to OvCa cell lines mediated by RGD-modified AAV vectors.
Cells were exposed for 2 h at 4°C to either RGD-modified (A588-4C-RGD
capsid) or standard AAVeGFP vector (wild-type AAV2 capsid) at an MOI of 500
DRP/cell. Unbound virus was then removed, fresh medium was added, and cells
were analyzed by FACS after 48 h. Data represent the percent of the cell
population that expressed the eGFP transgene and are shown as the means of
triplicate experiments (bars, SD). (b) Integrin-targeted, HSPG-independent,
transduction of OvCa cells mediated by RGD-modified AAV vectors. SKOV3.
ip1 cells were exposed for 2 h at 4°C to RGD-modified (A588-4C-RGD capsid)
or standard (wild-type AAV2 capsid) AAVeGFP vector at an MOI of 500 DRP/
cell. Unbound virus was then removed, fresh medium was added, and the cells
experiments, the viruses were bound to cells for 2 h at 4°C in the presence of
500 μg/ml heparin sulfate (HS), 200 μg/ml RGD-peptide (RGD), LM609
antibody (1:200 dilution), or combinations of HS and RGD, or HS and LM609
antibody. Isotype-matched control antibody failed to inhibit AAV-mediatedgene
transduction. Data represent the means of triplicate experiments (bars, SD).
1060W. Shi et al. / Gynecologic Oncology 103 (2006) 1054–1062
killed. Recently, the distribution and clinical significance of
HSPG in OvCa has been reported . While, the stromal
induction of syndecan-1 on OvCa was shown to contribute to its
pathology, the relevance of this finding to our study is unclear
since only the specific HSPG epitope recognized by the HepSS-
1 antibody appears to serve as the receptor for AAV2 .
AAV-mediated gene transfer is a treatment strategy that is
gaining attention for a number of different human cancers.
Whereas many cancer-derived cell lines in culture have been
shown to be susceptible to AAV infection, low expression of
AAVreceptors that attenuates AAV therapeutic efficacy may be
a common feature of primary tumors. Avariety of methods have
been devised to increase AAV-mediated gene transfer to cells
that are normally poorly transduced by AAV. Most notably,
these involve the genetic insertion of foreign peptide epitopes
into the AAV capsid which allow the viral particle to interact
with alternative cellular receptors [20–24]. We have found that
the modification of the AAV2 VP1, VP2, and VP3 capsid
proteins by the addition of an RGD-containing peptide
increased gene transfer to all OvCa cell lines tested. The fact
that a less dramatic increase was seen with cell lines that
expressed appreciable levels of HSPG (OV-4, OVCAR-3, and
PA-1) is consistent with low or absent HSPG being the limiting
factor in the other cell lines. Interestingly, the RGD capsid
modification was also able to increase gene transfer to
previously transducible OvCa cell lines, which suggests that a
second post-attachment block may also be overcome in these
cell lines through the use of this targeted vector, or that additive
binding to multiple cell surface receptors can lead to increased
The use of the RGD-modified capsid enabled AAV2HSVtk
in combination with GCV to kill OvCa cells that lacked HSPG
expression. Therapeutic efficacy was further increased by
endogenous HSPG expression in OvCa cells, suggesting that
the RGD modification might also be useful for enhancing AAV
efficacy for HSPG-positive tumors as well. At the concentra-
tions of virus tested, RGD-AAV2HSVtk/GCVappeared to be a
more potent therapeutic regimen than wild-type AAV2HSVtk/
GCVon these cell lines. The enhanced potency could be due in
part to an altered intracellular trafficking pathway imparted by
the RGD–integrin interaction, or avoidance of degradative
pathways (proteosomal or lysosomal) that may also limit
vector transduction. Interestingly, transduction of OV-3 cells
was significantly increased by the RGD-modified capsid in the
absence of appreciable cell surface αvβ3or αvβ5integrin on
these cells. The modified vectors could be interacting with a
different RGD-binding integrin on these cells . Alterna-
tively, transduction may be mediated by the small amount of
HSPG present on these cells and enhanced by the modified
capsid's ability to overcome a secondary post-attachment
block to effective gene transduction (e.g., intracellular
trafficking). Nevertheless, our data suggest that the increased
potency of RGD-AAV2HSVtk most likely results from
enhanced cell binding and/or entry because the enhanced
transduction of SKOV3.ip1 cells could be effectively blocked
via the addition of soluble RGD peptide or anti-integrin
The clinical utility of RGD-modified AAV vectors to
selectively treat OvCa has yet to be determined and in vivo
results may be different. However, previous work suggests that
these vectors will be superior to unmodified AAV vectors for
OvCa gene transfer . Although engineered to infect OvCa
cells through both HSPG and RGD-binding integrin interac-
tions, RGD-modified vectors could theoretically transduce non-
tumor tissue, e.g., normal epithelial cells or peritoneal
mesothelial cells. In an effort to address this issue, we have
investigated gene delivery to normal ovarian epithelial cells and
normal peritoneal mesothelial cells by AAV2 and RGD-
modified AAV2 vectors in culture. Although, these cells can
be transduced at a very low level with these vectors, there
appears to be at least a three order of magnitude difference in
gene transfer efficiencies between these cell lines and the least
transducible OvCa cell line we tested (data not shown). This
suggests that a substantial therapeutic opportunity exists for the
use of these agents to treat OvCa. Nevertheless, the develop-
ment of modified AAV vectors more specifically restricted for
OvCa should include modifications that restrict non-OvCa cell
We thank Amy Henning for the technical assistance. This
work was supported by a Young Investigator Award from the
Alliance for Cancer Gene Therapy (ACGT) to J.S.B, and grants
from the National Institutes of Health (R21 AI51388 and RO1
AI51388), and the Ohio Division of the American Cancer
Society to J.S.B.
 Liu X, Clark K, Johnson P. Production of recombinant adeno-associated
virus vectors using a packaging cell line and a hybrid recombinant
adenovirus. Gene Ther 1999;6 (Feb):293–9.
 Clark KR, Liu X, McGrath JP, Johnson PR. Highly purified recombinant
adeno-associated virus vectors are biologically active and free of
detectable helper and wild type viruses. Hum Gene Ther 1999;10(6):
 Hermonat PL, Muzyczka N. Use of adeno-associated virus as a
mammalian DNA cloning vector: transduction of neomycin resistance
into mammalian tissue culture cells. Proc Natl Acad Sci U S A 1984;81:
 Laughlin CA, Cardellichio CB, Coon HC. Latent infection of KB cells
with adeno-associated virus type 2. J Virol 1986;60:515–24.
 Lebkowski JS, McNally MM, Okarma TB, Lerch LB. Adeno-associated
virus: a vector system for efficient introduction of DNA into a variety of
mammalian cell types. Mol Cell Biol 1988;8:3988–96.
 McLaughlin SK, Collis P, Hermonat PL, Muzyczka N. Adeno-associated
virus general transduction vectors: analysis of proviral structures. J Virol
 Samulski RJ, Chang L-S, Shenk T. Helper-free stocks of recombinant
adeno-associated viruses: normal integration does not require viral gene
expression. J Virol 1989;63:3822–8.
 Teramoto S, Bartlett JS, McCarty D, Xiao X, Samulski RJ, Boucher RC.
Factors influencing adeno-associated virus-mediated gene transfer to
humancystic fibrosis airway epithelial cells: a comparison with adenovirus
vectors. J Virol 1998;44(11):8904–12.
 Tratschin J-D, Miller IL, Smith MG, Carter BJ. Adeno-associated virus
vector for high-frequency integration, expression, and rescue of genes in
mammalian cells. Mol Cell Biol 1985;5:3251–60.
1061 W. Shi et al. / Gynecologic Oncology 103 (2006) 1054–1062
 Walsh CE, Liu JM, Xiao X, Young NS, Nienhuis AW, Samulski RJ. Download full-text
Regulated high level expression of a human γ-globin gene introduced
into erythroid cells by an adeno-associated virus vector. Proc Natl Acad
Sci U S A 1992;89:7257–61.
 Walsh CE, Nienhuis AW, Samulski RJ, Brown MG, Miller JL, Young NS,
et al. Phenotypic correction of Fanconi anemia in human hematopoietic
cells with a recombinant adeno-associated virus vector. J Clin Invest
 Russell DW, Miller AD, Alexander IE. Adeno-associated virus vectors
preferentially transduce cells in S phase. Proc Natl Acad Sci USA
 Halbert CL, Alexander IA, Wolgamot GM, Miller AD. Adeno-associated
virus vectors transduce primary cells much less efficiently than
immortilized cells. J Virol 1995;69(3):1473–9.
 Song S, Morgan M, Ellis T, Poirier A, Chesnut K, Wang J, et al. Sustained
secretion of human alpha-1-antitrypsin from murine muscle transduced
with adeno-associated virus vectors. Proc Natl Acad Sci USA 1998;95
 Flotte TR, Afione SA, Conrad C, McGrath SA, Solow R, Oka H, et al.
Stable in vivo expression of the cystic fibrosis transmembrane regulator
with an adeno-associated virus vector. Proc Natl Acad Sci USA 1993;
 Kronenberg S, Kleinschmidt JA, Bottcher B. Electron cryo-microscopy
and image reconstruction of adeno-associated virus type 2 empty capsids.
EMBO Rep 2001;2(11):997–1002.
 Xie Q, Bu W, Bhatia S, Hare J, Somasundaram T, Azzi A, et al. The atomic
structure of adeno-associated virus (AAV-2), a vector for human gene
therapy. Proc Natl Acad Sci USA 2002;99(16):10405–10.
 Opie SR, Warrington Jr KH, Agbandje-McKenna M, Zolotukhin S,
MuzyczkaN. Identification of amino acid residues in the capsid proteins of
adeno-associated virus type 2 that contribute to heparan sulfate
proteoglycan binding. J Virol 2003;77(12):6995–7006.
 Kern A, Schmidt K, Leder C, Muller OJ, Wobus CE, Bettinger K, et al.
Identification of a heparin-binding motif on adeno-associated virus type 2
capsids. J Virol 2003;77(20):11072–81.
 Shi W, Arnold GS, Bartlett JS. Insertional mutagenesis of the adeno-
associated virus type 2 (AAV2) capsid gene and generation of AAV2
vectorstargeted toalternative cell-surface receptors. HumGeneTher2001;
 Shi W, Bartlett JS. RGD inclusion in VP3 provides adeo-associated virus
type 2 (AAV2)-based vectors with a heparan sulfate-independent cell entry
mechanism. Mol Ther 2003;7(4):515–25.
 Girod A, Ried M, Wobus C, Lahm H, Leike K, Kleinschmidt J, et al.
Genetic capsid modifications allow efficient re-targeting of adeno-
associated virus type 2 [published erratum appears in Nat. Med. 1999
Dec;5 (12):1438]. Nat Med 1999;5(9):1052–6.
 Grifman M, Trepel M, Speece P, Gilbert LB, Arap W, Pasqualini R, et al.
Incorporation of tumor-targeting peptides into recombinant adeno-
associated virus capsids. Mol Ther 2001;3(6):964–75.
 Wu P, Xiao W, Conlon T, Hughes J, M A-M, Ferkol T, et al. Mutational
analysis of the adeno-associated virus type 2 (AAV2) capsid gene and
construction of AAV2 vectors with altered tropism. J Virol 2000 (Sep);74:
 Fisher K, Gao G-P, Weitzmann MD, DeMatteo R, Burda JF, Wilson JM.
Transduction with recombinant adeno-associated virus for gene therapy is
limited by leading-strand synthesis. J Virol 1996;70(1):520–32.
 Ferrari FK, Samulski T, Shenk T, Samulski RJ. Second-strand synthesis is
a rate-limiting step for efficient transduction by recombinant adeno-
associated virus vectors. J Virol 1996;70(5):3227–34.
 Duan D, Yue Y, Yan Z, Yang J, Engelhardt JF. Endosomal processing
J Clin Invest 2000;105(11):1573–87.
 Douar AM, Poulard K, Stockholm D, Danos O. Intracellular trafficking of
adeno-associated virus vectors: routing to the late endosomal compartment
and proteasome degradation. J Virol 2001;75(4):1824–33.
 Hansen J, Qing K, Kwon HJ, Mah C, Srivastava A. Impaired intracellular
trafficking of adeno-associated virus type 2 vectors limits efficient
transduction of murine fibroblasts. J Virol 2000;74(2):992–6.
 Thomas CE, Storm TA, Huang Z, Kay MA. Rapid uncoating of vector
genomes is the key to efficient liver transduction with pseudotyped adeno-
associated virus vectors. J Virol 2004;78(6):3110–22.
 Moolten FL. Tumor chemosensitivity conferred by inserted herpes
thymidine kinase genes: paradigm for a prospective cancer control
strategy. Cancer Res 1986;46:5276–81.
 Faulds D, Heel RC. Ganciclovir. A review of its antiviral activity,
pharmacokinetic properties and therapeutic efficacy in cytomegalovirus
infections. Drugs 1990;39(4):597–638.
 Lal S, Lauer UM, Niethammer D, Beck JF, Schlegel PG. Suicide genes:
past, present and future perspectives. Immunol Today 2000;21(1):48–54.
 Hemminki A, Dmitriev I, Liu B, Desmond RA, Alemany R, Curiel DT.
Targeting oncolytic adenoviral agents to the epidermal growth factor
pathway with a secretory fusion molecule. Cancer Res 2001;61(17):
 Kanerva A, Zinn KR, Chaudhuri TR, Lam JT, Suzuki K, Uil TG, et al.
Enhanced therapeutic efficacy for ovarian cancer with a serotype 3
receptor-targeted oncolytic adenovirus. Mol Ther 2003;8(3):449–58.
 Graham FL, Smiley J, Russell WC, Nairn R. Characteristics of a human
cell line transformed by DNA from human adenovirus type 5. J Gen Virol
 Clark K, Voulgaropoulou F, Johnson P. A stable cell line carrying
adenovirus-inducible rep and cap genes allows for infectivity titration of
adeno-associated virus vectors. Gene Ther 1996 (Dec);3:1124–32.
 Clark K, Voulgaropoulou F, Fraley D, Johnson P. Cell lines for the
production of recombinant adeno-associated virus. Hum Gene Ther 1995
 Summerford C, Samulski RJ. Viral receptors and vector purification: new
approaches for generating clinical-grade reagents. Nat Med 1999;5(5).
 Zolotukhin S. Recombinant adeno-associated virus purification using
novel methods improves infectious titer and yield. Gene Ther 1999;6:
 van de Loosdrecht AA, Beelen RH, Ossenkoppele GJ, Broekhoven MG,
Langenhuijsen MM. A tetrazolium-based colorimetric MTT assay to
quantitate human monocyte mediated cytotoxicity against leukemic cells
from cell lines and patients with acute myeloid leukemia. J Immunol
 Xiao W, Warrington Jr KH, Hearing P, Hughes J, Muzyczka N.
Adenovirus-facilitated nuclear translocation of adeno-associated virus
type 2. J Virol 2002;76(22):11505–17.
 Summerford C, Bartlett JS, Samulski RJ. AlphaVbeta5 integrin: a co-
receptor for adeno-associated virus type 2 infection. Nat Med 1999;5(1):
 Carreiras F, Denoux Y, Staedel C, Lehmann M, Sichel F, Gauduchon P.
Expression and localizationof alphav integrins and theirligand vitronectin
in normal ovarian epithelium and in ovarian carcinoma. Gynecol Oncol
 Liapis H, Adler LM, Wick MR, Rader JS. Expression of alpha(v)beta3
integrin is less frequent in ovarian epithelial tumors of low malignant
 Yan Z, Zak R, Luxton GW, Ritchie TC, Bantel-Schaal U, Engelhardt JF.
Ubiquitination of both adeno-associated virus type 2 and 5 capsid proteins
affects the transduction efficiency of recombinant vectors. J Virol 2002;76
 Davies EJ, Blackhall FH, Shanks JH, David G, McGown AT, Swindell R,
et al. Distribution and clinical significance of heparan sulfate proteogly-
cans in ovarian cancer. Clin Cancer Res 2004;10(15):5178–86.
 Qiu J, Handa A, Kirby M, Brown KE. The interaction of heparin sulfate
and adeno-associated virus 2. Virology 2000;269(1):137–47.
 Li E, Brown SL, Stupack DG, Puente XS, Cheresh DA, Nemerow GR.
Integrin alpha(v)beta1 is an adenovirus coreceptor. J Virol 2001;75(11):
1062W. Shi et al. / Gynecologic Oncology 103 (2006) 1054–1062