original article
© The American Society of Gene Therapy
Improved Retinal Transduction In Vivo and
Photoreceptor-specific Transgene Expression Using
Adenovirus Vectors With Modified Penton Base
Siobhan M Cashman1, Laura McCullough1 and Rajendra Kumar-Singh1
Department of Ophthalmology, Tufts University School of Medicine, Boston, Massachusetts, USA
1
Adenovirus (Ad) vectors can be injected into human
ocular tissues without producing adverse events and
are therefore a promising means of gene transfer to the
retina. However, when administered subretinally, Ad vectors primarily transduce the retinal pigment epithelium
(RPE), whereas the majority of mutant gene products that
cause photoreceptor (PR) degeneration are expressed
exclusively in the PR cells. While it has been shown previously that pseudotyping of Ad can partially overcome the
limited PR transduction by Ad5, we found that pseudotyping of Ad is not necessary for transduction of PR cells.
We determined that, in the context of Ad, the cytomegalovirus (CMV) promoter is not significantly active in PRs.
We compared expression levels from CMV and chicken
beta actin (CBA) promoters in neural retina and found
that CBA has a 173-fold greater potency than CMV. We
also investigated the nature of the Ad–RPE interaction
in murine retina and determined that the RGD domain
in Ad penton plays a key role in RPE tropism. Deletion
of the RGD domain coupled with use of the CBA promoter permitted transgene expression in neural retina
approximately 667 times more efficiently than with Ad5
vectors. The use of these vectors in combination with a
4.7 kilobase (kb) rhodopsin promoter enabled transgene
expression exclusively in PR cells in vivo.
Received 13 June 2006; accepted 3 April 2007; published online
15 May 2007. doi:10.1038/sj.mt.6300203
INTRODUCTION
The use of adenovirus (Ad) vectors has led to substantial success in demonstrating rescue of retinal degeneration in animal models of human disease.1,2 In fact, the only two human
ocular gene therapy trials to date have selectively used Ad for
gene delivery.3,4 These studies strongly suggest that Ad may be
an ideal vector for gene transfer to the human retina. Some
advantages of Ad over other vector systems include a packaging
capacity of 36 kilobase (kb),5,6 and ease of scaleable production.
Also,episomal persistence and transgene expression extending
over years have been demonstrated in non-human primates,7
and lifetime correction of disease in mice.8
One of the most common diseases of the retina leading to
blindness is retinitis pigmentosa, affecting approximately 1 in
3,000 individuals worldwide.9 The most common serotype of Ad
used in gene therapy studies is Ad5, which almost exclusively
infects the retinal pigment epithelium (RPE) when administered
subretinally.10–12 Although the RPE provides the essential components for the proper functioning and survival of the adjacent
PR cells, the majority of the genes associated with retinitis pigmentosa are expressed exclusively in the PRs.13 Therefore it would
initially appear that Ad would not be useful in the treatment of
retinitis pigmentosa.
Other than pseudotyped adeno-associated virus (AAV), no
previously described gene delivery vectors were able to transduce
PRs efficiently in the post-mitotic retina. For example, while lentivirus vectors can transduce14 and rescue15 PR degeneration when
administered to neonatal mice, they do not transduce the fully
developed PRs found in the post-mitotic adult murine retina.16–18
Hence, lentivirus vectors have limited use in the treatment of diseases affecting human PRs which, in contrast to murine PRs, are
almost fully developed in utero.19 Indeed, it has been postulated
that the most common time period for intervention in retinal
gene therapy experiments in mice (3–7 days after birth) is likely
equivalent to the second trimester of pregnancy in humans.20
While AAV vectors can overcome this limitation, they have a
limited cloning capacity of approximately 5 kb and therefore cannot include the large upstream regulatory elements necessary for
regulated expression of transgenes. Regulation of rhodopsin, for
example, has been shown to be essential in order to prevent the
retinal degeneration that may be induced by as little as 23% overexpression of wild-type rhodopsin in PRs.21 Because of the limitations in current vector technology, it would be prudent to develop
Ad vectors specifically for retinal gene therapy.
In this study, we examined the utility of Ad5 pseudotyped with
Ad17 fiber (Ad5/F17), a serotype that has been previously shown to
infect neurons in cell culture.22 We found that Ad5/F17 vectors perform better than Ad5 vectors in their ability to transduce neural retina but do not express high levels of transgene product in the PRs.
All previous studies, including those carried out by us earlier, have apparently made the assumption that, in the context of
Ad5 in PR cells, the cytomegalovirus (CMV) promoter is strongly
active, and that the absence of reporter gene expression (green
Correspondence: Rajendra Kumar-Singh, Department of Ophthalmology, Tufts University School of Medicine, 136 Harrison Avenue, Boston,
Massachusetts 02111, USA. E-mail: rajendra.kumar-singh@tufts.edu
1640
www.moleculartherapy.org vol. 15 no. 9, 1640–1646 sept. 2007
© The American Society of Gene Therapy
5
Ad
Molecular Therapy vol. 15 no. 9 sept. 2007
31
9
08
{N
,
5
Ad
31
de
el
Nd
Ad5
Fiber
EGFP
Poly-A E4 (Ad5)
BGH pA
MLT
E1
pA
Poly-A L5
(Ad17)
Ad17 Fiber
ITR E3
ITR
E2
E4
CMV
5/
5
Ad
Ad
EG
PN
PN
F
F
EG
denatured
7
F1
5/
5
Ad
PN
F
EG
Ad
7
F1
PN
F
EG
native
184.8 kd
(Trimer)
(Trimer)
121.2 kd
61.6 kd
(Monomer)
RESULTS
Ad5 pseudotyped with Ad17 fiber improves
transduction of neural retina in vivo
Previous studies have shown that Ad2 pseudotyped with Ad17
fiber can infect primary neuronal cells in culture.22 Because the
retina contains a variety of specialized neurons including PRs, we
wished to test the hypothesis that Ad5 pseudotyped with Ad17
fiber may infect retinal neurons more efficiently than Ad5. In order
to construct such a vector, we genetically modified the fiber gene in
Ad5. In brief, we made modifications such that the resultant vector
would express a hybrid fiber comprised of the N terminal 10 amino
acids unique to Ad5 fiber followed by the amino acids FNPVYPY
that are common to both Ad5 and Ad17 fiber, followed by amino
acids 18–366 of Ad17, resulting in a total fiber length of 366 amino
acids and identical in length to wild-type Ad17 fiber (Figure 1a).
The modified fiber was followed by 62 base pairs of Ad17 sequence
derived from the 3a end of Ad17 fiber and contained a bovine
growth hormone pA signal further downstream in order to ensure
appropriate processing of the modified transcript. In order to allow
quantitation of potentially transduced neurons in vivo, we cloned
a CMV-GFP expression cassette in the E1 region of this modified
virus backbone (Figure 1b). In order to confirm that the modified
monomeric Ad5/F17 fiber trimerized and was incorporated into
mature capsids, we probed the hybrid fiber by Western blot analysis using protein prepared from purified virions. Our results generally indicated that the monomeric fiber is expressed, trimerized,
and incorporated into virions and that denaturation by boiling the
samples prior to loading allows all of the Ad5/F17 fibers to resolve
into monomeric fibers (Figure 1c). The molecular weights of the
various fibers were roughly consistent with the predicted sizes, as
determined by amino acid sequence analysis software (MacVector).
However, we did note that levels of fiber from Ad5/F17 were significantly lower than those observed for Ad5 (Figure 1c).
l}
2
4
,0
Ad
5
3
A 2,7
TG d5 44
A 32,
Ad 78
5 7
Fi
be
r
TGA
fluorescent protein (GFP), LacZ etc.) in PRs is an indirect indicator of the failure of Ad5 to transduce PRs. Hence, all of these
studies have reached a similar conclusion, i.e., that Ad5 has almost
exclusive tropism for the RPE when administered subretinally and
that modifications such as pseudotyping are necessary in order
to make Ad5 transduce PR cells. In this study we demonstrate
that pseudotyping of Ad is in fact not necessary for achieving
PR transduction, because a simple exchange of the transgeneassociated promoter reveals that Ad5 can transduce PR cells to
previously unprecedented levels. We expand these observations
by probing the structural components of the Ad5 capsid that may
be responsible for the substantially higher levels of transduction
of RPE at the expense of the PR cells. Given the robust integrinmediated phagocytic activity of the RPE for rod outer segment
discs, we postulated that the RGD domain in the Ad penton base
may play a significant role in Ad5 uptake by the RPE. We corroborated this hypothesis by demonstrating that deletion of the
RGD domain in the Ad penton base allows redirecting of Ad5 tropism from its preferential transduction of RPE cells to a now more
equitable transduction of both RPE and PRs, and at levels significantly greater than those achieved by pseudotyping or promoter
exchange alone. These observations further allow for the design of
Ad vectors that can express transgenes exclusively in the PR cells.
Gene Transfer to Photoreceptors
(Monomer)
40.4 kd
EGFPNAd5/F17
RPE
BF10x
GFP10x
MERGE
RPE
BF40x
GFP40x
MERGE
Figure 1 Structure and characterization of EGFPNAd5/F17 virus.
(a) N terminal amino acid sequences of wild-type Ad5, Ad17, and the
Ad5F17 fusion fiber. The hybrid fiber is composed of the first 10 amino
acids of Ad5 followed by amino acids 11–366 of wild-type Ad17.
(b) Structure of EGFPNAd5/F17. (c) Western blot analysis of protein prepared from purified virions indicating the presence of both the monomeric and the trimerized fiber. (d) GFP expression in the RPE cells (arrow)
and occasional Müller cells (arrowhead) in murine retina infected with
EGFPNAd5/F17. The s40 images are the boxed areas in the s10 images.
ITR, adenovirus inverted terminal repeat; 9, Ad packaging signal;
BGHpA, bovine growth hormone polyadenylation signal; BF, bright
field; CMV, cytomegalovirus promoter/enhancer; E, early region; EGFP,
enhanced green fluorescent protein; GFP, green fluorescent protein;
MLT, major late transcription unit; RPE, retinal pigment epithelium.
In order to determine whether Ad5/F17 (EGFPNAd5/F17)
could transduce retinal neurons, we injected EGFPNAd5/F17
into the subretinal space of adult C57BL/6J mice and measured
GFP expression qualitatively by direct fluorescence and quantitatively by quantitative real-time polymerase chain reaction of
neural retina (i.e., with RPE removed). Qualitatively, it appeared
that Ad5/F17 virions transduced the same cell types as the
parental Ad5 virus, i.e., we observed primarily RPE-specific GFP
1641
© The American Society of Gene Therapy
Gene Transfer to Photoreceptors
expression and the occasional Müller cell and cells of the inner
nuclear layer. However, quantitative real-time polymerase chain
reaction revealed that Ad5/F17 virus transduced murine neural
retina 24 (±4) times more efficiently than Ad5 did, the greater
part of the signal being possibly due to slightly enhanced Müller
cell transduction. However, Ad5/F17 did not transduce PRs very
effectively (Figure 1d).
Expression of GFP in PRs transduced by Ad5
The assumption that the CMV promoter is sufficient to determine which cell types are transduced by Ad in the retina has
previously led several investigators including ourselves to conclude that PRs are not transduced by Ad5, and that alternative
approaches such as pseudotyping are necessary.23,24 In order to
examine the role of the CMV promoter in the transduction of
PRs by Ad, we constructed Ad5 vectors expressing GFP from
a chicken beta actin (CBA) promoter/CMV enhancer/rabbit globin intron25 (Ad5CAGGFP) and from a CMV promoter
(Ad5CMVGFP), and injected one or other of these into the subretinal space of adult C57BL/6J mice. An examination of frozen
retinal sections (Figure 2) 6 days following subretinal injection
yielded results in agreement with previous studies, showing that
GFP expression from the CMV promoter occurs almost exclusively in the RPE (Figure 2e). In contrast, in addition to the RPE,
cells in the outer nuclear layer are also strongly GFP-positive
when the CBA promoter is utilized to regulate transgene expression (Figure 2h). A close examination of these retinal sections
leads to the identification of these GFP-positive cells as PR cell
bodies and a very large number of GFP-positive inner and outer
segments (Figure 2k), thereby confirming the presence of GFPpositive PRs. Areas in the same retinal section at the border of
GFP-positive RPE, presumably the border of the injection site
(Figure 2n), do not reveal any GFP-positive PRs at the exact
same exposure. We hence conclude that Ad5 vectors can indeed
transduce PR cells, and that the CBA promoter allows robust
transgene expression in PRs whereas CMV does not. Robust PR
transduction was not limited to the site of injection; a single subretinal injection led to the transduction of approximately 20% of
the neural retina and approximately 35% of the RPE (Figure 2h).
Transduction was also observed in the corneal endothelium
(data not shown), presumably due to leakage or transport of the
vector from the subretinal space into the vitreous. Quantitative
real-time polymerase chain reaction of these transduced neural
retinas (with RPE removed) indicated that GFP expression from
Ad5CAGGFP virus containing a CBA promoter was approximately 173 (±55) times more efficient than expression from the
CMV promoter using the same Ad5 backbone.
The role of RGD in the Ad penton base
in Ad5-RPE tropism
Although PRs are in fact transduced by Ad5, the RPE is still substantially better transduced than the PRs. Beginning with the
unproven assumption that the CBA promoter is equally active in
both cell types, we sought to investigate the molecular basis for
the strong Ad–RPE interaction in order to potentially enhance
PR transduction by ameliorating or reducing such interaction.
We took into consideration the fact that RPE cells are phagocytic,
1642
Ad5CMVGFP
P
P
BF10x
GFP10x
MERGE
P
RPE
BF40x, P
GFP40x, P
MERGE, P
Ad5CAGGFP
R
Q
R
Q
BF10x
GFP10x
BF40x, Q
R
Q
MERGE
RPE
ONL
MERGE, Q
BF40x, R
GCL
GFP40x, Q
GFP40x, R
MERGE, R
Figure 2 Comparision of cytomegalovirus (CMV) and chicken beta actin
(CBA) promoters in Ad5 background in vivo. Low (a–c and g–i, s10)
and high magnification (d–f and j–o, s40) images from frozen sections
of murine retina injected with a virus expressing green fluorescent protein (GFP) from a CMV promoter (Ad5CMVGFP) or a CBA promoter
(Ad5CAGGFP). The s40 images encompass boxed areas (P,Q,R) depicted
in the s10 images. GFP-positive photoreceptors (PRs) are present (k) in
areas with GFP-positive RPE but absent at the edge of the injection site
(n). Sporadic GFP-positive ganglion cells in (k) demarcate the inner edge
of the retina and ganglion cell layer, more clearly seen in the merged
image (l). BF, bright field; ONL, outer nuclear layer; RPE, retinal pigment
epithelium. Exposure k and n = 2 seconds).
whose function include binding to and engulfing large amounts
of rod outer segments discs shed by the PR cells.26 These spent
discs are thought to bind the RPE through an RGD-αvβ5 integrin interaction.27 Coincidentally, Ad5 also binds αvβ5 integrin
through an RGD domain in the Ad5 penton base,28 and this
interaction is crucial for Ad5 endocytosis following the initial
Ad5 fiber knob-coxsackie and adenovirus receptor interaction.29
We hypothesized that this well-characterized RGD-αvβ5-integrin
interaction could be the cause of the very efficient uptake of Ad5
vectors into the RPE at the expense of the adjacent PRs. In order
to investigate this hypothesis, we injected a tetramethylrhodamine isothiocyanate-labeled RGD-containing peptide into the
subretinal space of adult C57BL/6J mice. Mouse eyes harvested
90 minutes later indicated that the RGD-containing peptide
bound primarily to the RPE and the blood vessels present in the
inner retina (Figure 3a) but not to the PR cells, a pattern similar
in part to Ad5 transduction of the retina. In contrast, a tetramethylrhodamine isothiocyanate-labeled peptide containing RGE did
www.moleculartherapy.org vol. 15 no. 9 sept. 2007
© The American Society of Gene Therapy
RPE
BF20x
Gene Transfer to Photoreceptors
RGD-TRITC
Ad5CAGGFPRGD
a
RPE
TRITC20x
RPE
RPE
RPE
BF20x
TRITC20x
MERGE
Figure 3 Binding of RGD-TRITC or RGE-TRITC to C57BL/6J mouse retina in vivo 90 minutes after subretinal injection. (a) RGD-TRITC peptide
primarily binds the RPE (arrow) and blood vessels (arrowheads) and does
not substantially bind the photoreceptors (PRs). (b) Control RGE-TRITC
peptide does not bind RPE (arrow) but does bind some blood vessels
(arrowheads); also, it does not bind PR cells. BF, bright field; RPE, retinal
pigment epithelium; TRITC, tetramethylrhodamine isothiocyanate.
not significantly bind the RPE but did bind the blood vessels in
the inner retina, albeit at a lower efficiency relative to the RGDcontaining peptide (Figure 3b).
If indeed RGD were a critical player in the Ad–RPE interaction, one would have expected to transduce the RPE as well as
the blood vessels with RGD-containing Ad5 viruses expressing
GFP, from a promoter active in both these tissues. However, in
this study, high levels of Ad5 transduction were observed only
in the RPE and PRs (Figure 2). This could be because: (i) the
spaces between the neurons in the retina do not allow easy access
of Ad to the blood vessels in the inner retina following subretinal administration, or (ii) insufficient virus is available to reach
the inner blood vessels because the potentially strong RGD–RPE
interaction functions as a sink for Ad vectors in the subretinal
space. Nonetheless, these data are consistent with the hypothesis
that RGD in the Ad5 penton base may play an important role in
the substantial affinity of Ad5 for RPE.
If this hypothesis is correct, deletion of the RGD from the Ad5
vector should reduce the Ad–RPE interaction and hence potentially make available more of the Ad for uptake by adjacent cells
such as the PRs. In order to test this hypothesis, we rescued Ad5
vector expressing GFP from a CBA promoter on an Ad5 backbone
with an RGD deletion in Ad penton base (Ad5CAGGFP∆RGD).
These constructs were injected into the subretinal space of adult
C57BL/6J mice, and frozen retinal sections were examined 6
days later. We noticed key differences between Ad5CAGGFPand Ad5CAGGFP∆RGD-injected mouse retina. First, the
region of the neural retina that was transduced was substantially
larger when Ad5CAGGFP∆RGD was used, with as much as
50% of the neural retina being transduced by a single injection
(Figure 4b and c), and the outer nuclear layer was significantly
GFP-positive within this transduced area (Figure 4e). Second, a
very large number of the PR cell bodies were GFP-positive in the
area of subretinal injection (Figure 4h), as well as blood vessels
(arrowheads, Figure 4h) and select ganglion cells (Figure 4h).
Longer exposures of these sections (data not shown) indicated
that cells in the inner layers of the retina were also significantly
transduced. Indeed, a “blanket” of GFP-positive inner segments
(Figure 4h, inset) could be easily discerned. In contrast, similar
exposures of relatively untransduced areas from the same retina
N
ONH
LENS
BF4x
RPE
Molecular Therapy vol. 15 no. 9 sept. 2007
M
N
MERGE
RGE-TRITC
c
b
M
N
ONH
M
ONH
LENS
LENS
GFP4x
MERGE
d
e
f
BF20x, M
GFP20x, M
MERGE, M
g
h
RPE
ONL
BF40x, M
i
IS
GCL
MERGE, M
GFP40x, M
Ad5CAGGFPRGD
(Relatively nontransduced area of same retina, N)
j
k
BF40x, N
GFP40x, N
l
MERGE
Figure 4 Transduction of photoreceptors (PRs) by Ad5CAGGFP∆RGD.
Low power (a–c, s4) and higher power (d–f, s20; g–l, s40) images of
mouse retina injected with Ad5CAGGFP∆RGD indicating a large portion
of the neural retina to be green fluorescent protein (GFP)-positive (b,c).
d–i represent boxed areas M in a–c. j–l represent relatively untransduced area N boxed in a. Exposure for k = 2 seconds, similar to e and
h, indicates a lack of autofluorescence as a contributor to GFP-positive
cells in e and h. Inset in h indicates a very large number of GFP-positive
inner segments (IS). Because of some variability in subretinal injections
and sizes of retinal detachment, we observed a variation in the total area
of retina transduced by the virus. This image depicts a result found in
approximately 50–60% of injected mice. However, PR transduction was
observed in almost all injected mice. BF, bright field; GFP, green fluorescent protein; ONH, optic nerve head; ONL, outer nuclear layer; PR, photoreceptor; RPE, retinal pigment epithelium; GCL, ganglion cell layer.
did not reveal any GFP-positive PRs (Figure 4k). Quantitation
of these data by real-time polymerase chain reaction indicated
that Ad5CAGGFP∆RGD drove GFP expression in neural retina
667 (±19) times higher than Ad5 with a CMV promoter and 4
(±1.24) times higher than Ad5CAGGFP. Whether the level of
transgene expression achieved in PR cells in these studies surpasses that which can be achieved by a single administration of
AAV remains to be determined.
PR-specific transgene expression
Having demonstrated improved transgene expression in neural
retina and specifically in PRs using Ad vectors with deletion of
RGD penton base, we sought to examine whether such vectors
could be utilized for expressing transgenes in a PR-specific manner. As discussed earlier, the control of transgene expression is
critically important in retinal gene therapy and, in theory, the
36 kb cloning capacity of helper-dependent Ad vectors can be
utilized for delivering the entire human rhodopsin gene or other
1643
© The American Society of Gene Therapy
Gene Transfer to Photoreceptors
Ad5RhoGFPRGD
a
MERGE
GFP20x
BF20x
e
d
RPE
RPE
ONL
ONL
BF40x
c
b
GCL
GFP40x
Outer segments
f
RPE
ONL
GCL
MERGE
GCL
Figure 5 Photoreceptor (PR)-specific transgene expression from
Ad5RhoGFP∆RGD. Frozen retinal sections 3 weeks after injection from
mice injected subretinally with Ad5RhoGFP∆RGD demonstrating a
spread of GFP (b) that localized exclusively to the outer nuclear layer
(e, f), a region that contains the PR cell bodies. s40 images (d–f) are
from a different area of the same retina shown in (a–c). Individual GFPpositive PR outer segments are visible in e. BF, bright field; GCL, ganglion cell layer; GFP, green fluorescent protein; ONL, outer nuclear layer;
PR, photoreceptor; RPE, retinal pigment epithelium.
genes involved in retinal regeneration including introns and
large upstream regulatory elements. Appropriately regulated PRspecific transgene expression would thus provide a further level
of safety in human gene therapy protocols. As a step towards that
goal, we generated an Ad vector (Ad5RhoGFP∆RGD) expressing
GFP regulated by a 4.7 kb murine opsin promoter. Frozen retinal
sections prepared from mice 3 weeks after subretinal injection
revealed GFP expression in the area of the injection (Figure 5b).
GFP expression was exclusively localized to the outer nuclear
layer that contains the PR cell bodies (Figure 5e and f) and
the inner and outer segments of the PR cells. Individual GFPpositive outer segments of PR cells could be easily discerned in
these sections (Figure 5e). Significantly, there was absolutely no
transgene expression in any other retinal cell type examined,
including the RPE cell which is the one that is most efficiently
transduced by all viral vector groups tested to date.
DISCUSSION
The use of ad vectors has met with substantial success in the only
two human ocular gene therapy trials to date. In both these trials,
a dose escalation study revealed no significant adverse events at
the highest doses tested—109.5 (ref. 3) or 1011 (ref. 4) virion particle
units per patient. Despite this success, the overwhelming majority
of ocular gene therapy studies in animals have not involved the
use of Ad as the gene transfer vector.30 The abandonment of Ad
in animal ocular gene therapy has come about, in part, as a result
of the substantial immune response observed in animals following administration of Ad into ocular tissues.31,32 However, in some
murine ocular gene transfer studies it has been demonstrated
that long-term transgene expression is indeed possible, exceeding 6 months, the longest time period of examination.24,33 Given
the promising data from human ocular gene therapy studies using
Ad, its exclusion in favor of alternative vectors may have been premature, especially given some of the advantages it has over other
vector systems.
In this study we have examined the current drawbacks of
Ad vector technology in the treatment of diseases involving PR
degeneration. As is the case with lentivirus, earlier generations
1644
of Ad5 vectors can transduce PRs only during the early phase of
retinal development.34 Hence, while such vectors can be used for
rescuing degenerating PRs in murine neonates,1,2 they do not have
much potential for application in similar diseases in humans. This
is probably because of key differences between mice and humans
in the timing of postnatal PR development. This limitation in tropism is not valid for in utero gene transfer but technical and ethical hurdles are likely to slow such applications.
In order to address these deficiencies, we first examined the
potential extent of PR transduction by Ad5 vectors pseudotyped
with Ad17 fiber, a pseudotype known to infect post-mitotic neuronal cells in culture.22 Previous efforts to redirect Ad tropism
from RPE to PRs include similar approaches, e.g., Ad5 pseudotyped with Ad37 (Ad5/f37)23 or Ad35 (Ad5/f35)24 fiber. We have
previously shown35 that Ad5/f37 can be redirected from binding
its native receptor to binding sialic acid, an amino carbohydrate
abundantly present in the retina. In those former studies, Ad vectors expressing GFP from a CMV promoter were utilized to demonstrate PR transduction. GFP expression was directly observable
in PRs when the Ad5/f37 vector was used,23 whereas, in the case
of the Ad5/f35 vector, indirect and significantly more sensitive
methods of detecting GFP, i.e., antibodies, were found to be necessary.24 Although we did find an improvement in transduction of
neural retina with Ad5/f17 vectors, it appeared to be on account
of a slightly increased transduction of Müller cells and other cells
in the inner nuclear layers rather than the PRs.
Although the CMV promoter has been shown to be active in
PRs when CMV-regulated expression cassettes are delivered to PRs
by alternative vectors such as AAV,36 it is surprising that there is
almost no transgene expression in PRs in the context of Ad. One
possible explanation for this observation is that Ad may generate a
greater immune response than AAV does in murine ocular tissues,
and this may initiate a cascade of events that rapidly shuts off, at
least partially, viral promoters such as CMV. For example, interferon-γ has been shown to bind and shut down expression from
the CMV promoter as part of the natural host immune response.37
We have tested PR expression from CMV promoters in the context
of Ad with both sense and antisense orientations with respect to
the E1 enhancer and seen similar results (data not shown). This
implies that the location of the CMV promoter and its proximity
to Ad E1 enhancers may not be relevant. Surprisingly, we were able
to demonstrate substantial transduction of PRs by use of the CBA
promoter, from which it appears that the obstacle to PR transduction is at least partly at the level of transcription and not exclusively
at the infection stage as had been documented previously.
Previous studies have examined the role that RGD in the Ad
penton base plays in mediating entry into non-ocular cells in vivo.
In those studies it was determined that RGD deletions do not
substantially change the levels of Ad uptake by, for example, the
liver following intravenous injection.38 A variety of studies have
reported enhanced Ad uptake into cells, often neoplastic cells that
express high levels of integrin, by incorporation of RGD in the
H1 loop of the fiber knob.39 Our study is hence atypical, in that
we achieved enhanced targeting of the cell of interest by reducing rather than increasing the RGD–integrin interaction. While
we did not demonstrate conclusively that the RGD–RPE interaction is exclusively responsible for the efficient uptake of Ad at the
www.moleculartherapy.org vol. 15 no. 9 sept. 2007
© The American Society of Gene Therapy
expense of the PRs, we did nonetheless find that deletions in RGD
allowed substantial improvements in PR transduction. Further
studies are necessary in order to characterize in detail the interactions between RPE and Ad-uptake.
Finally, we demonstrated that a 4.7 kb murine opsin promoter
may be utilized to drive GFP expression exclusively in the PR cells.
To the best of our knowledge, this is the largest PR-specific promoter delivered ectopically to PR cells thus far. While the RGD
deletion in Ad penton base (Ad5∆RGD) allowed for improved PR
transduction, the use of the 4.7 kb opsin promoter on such viral
backbones allowed expression exclusively in PR cells. We did not
examine whether Ad5 vectors expressing transgenes regulated
from this same promoter are also capable of PR-specific transgene
expression. However, our data support the idea that transgene levels from Ad5 vectors are likely to be lower than from Ad5∆RGD
vectors in PRs, because of an a priori difference in rates of PR
transduction.
With a greater understanding of Ad tropism in ocular tissues,
the vectors described in this study can be applied in gene therapy
studies of a variety of ocular diseases involving the degeneration
of the retina. Cumulatively, the data and vectors described in this
study should be useful for the rescue of PRs in animal models of
retinal degeneration and patients with retinitis pigmentosa and
allied retinal disorders.
MATERIALS AND METHODS
Construction of plasmids. pAd5/F17 and pShEGFPN were generated essen-
tially as previously described for pAd5/F37.35 Wild-type Ad17, strain Ch.22,
was obtained from the American Type Tissue Culture Collection. pShCAG
GFP was constructed by cloning an SpeI/HindIII fragment of pCAGGFP (a
generous gift from C. Cepko and T. Matsuda) into XbaI/HindIII-digested
pShuttle.40 pAdAdEasy1∆RGD was constructed by cloning an FseI fragment from AdHM4∆RGD (generously provided by A. Lieber) into FseIdigested pAdEasy1.40 The 4.7 kb murine opsin promoter was cloned as a
BamHI fragment from pSB6.25 (generously provided by W. Baehr) and
used for generating a GFP-expressing shuttle as earlier described for pSh
CAGGFP. Further cloning details are available upon request.
Production of adenoviruses. Viruses were produced as previously des-
cribed.12,40 In brief, adenoviruses were produced by recombination between
the respective shuttle plasmid and modified Ad backbone in BJ5183 cells.
Recombined plasmids were digested with PacI and transfected into the
human embryonic retinoblast (911) cell line,41 and the resultant viruses
were purified using the Adenopure kit (Puresyn).
Western blot analysis of fiber. 3.2 × 1010 virus particles were resus-
pended in 50 mmol/l Tris–HCl, pH 8.0/150 mmol/l NaCl/0.1% Sodium
dodecyl sulfate/1% Triton X-100 containing leupeptin (10 µg/ml), aprotinin (10 µg/ml), and phenylmethanesulphonylfluoride (0.1 mmol/l).
Half of the lysate was pre-incubated at 100 oC. These were then loaded on
a 12% denaturing gel (BMA, Rockland, ME) and probed for fiber using
the monoclonal antibody Ab-4 (Clone 4D2; NeoMarkers), followed by a
horseradish peroxidase-conjugated goat anti-mouse antibody (Jackson
ImmunoResearch).
Subretinal injections. C57Bl/6J mice were bred and maintained in a 12-
hour light–dark cycle and cared for in accordance with federal, state and
local regulations. The use of animals in this work was in accordance with the
Association for Research in Vision and Ophthalmology (ARVO) Statement
for the Use of Animals in Ophthalmic and Vision Research. Mice were
anesthetized by intraperitoneal injection of xylazine (10 mg/ml)/ketamine
Molecular Therapy vol. 15 no. 9 sept. 2007
Gene Transfer to Photoreceptors
(1 mg/ml). Subretinal injections were performed using the transcleral
approach with a 32G needle attached to a 5 μl glass syringe (Hamilton).
3 × 109 virus particles were injected into C57Bl/6J mice. 17.5 nmol of either
the tetramethylrhodamine isothiocyanate-labeled 6-amino acid GRGDSP
peptide or the GRGESP peptide was injected into the sub-retinal space of
C57Bl6/J mice. Eyes were harvested 90 minutes post-injection and imaged
as described in the following text.
Quantitative RT-PCR. Retinas (n = 4 per construct) were dissected
and processed for total RNA using RNA STAT-60, according to the
manufacturer’s instructions (Tel-Test). DNAse-treated RNA was reversetranscribed using oligod(T)16 and TaqMan (Applied Biosystems) reverse
transcription kit according to the manufacturer’s instructions. Single
PCRs were performed on cDNA using TaqMan Universal PCR Master
Mix (Applied Biosystems) and analyzed using the ABI PRISM 7900HT
Sequence Detection System. GFP messenger RNA was detected using
the primer/probe combination EGFP-F 5a-GAGCGCACCATCTTCTT
CAAG-3a, EGFP-R 5a-TGTCGCCCTCGAACTTCAC-3a and EGFPM1 ACGACGGCAACTACA, an assay custom-designed by Applied
Biosystems. Normalization was performed using a TaqMan endogenous
mouse β-actin control primer/probe combination (Applied Biosystems
part no. 4352663). All probes contained the covalently-linked reporter
FAM dye at their 5a ends and a TAMRA quencher dye at their 3a ends.
Image analysis. Six days after injection (except for Ad5RhoGFP∆RGD
which was analyzed 3 weeks after injection), mice were sacrificed by CO2
inhalation. Eyes were harvested and fixed with 4% paraformaldehyde
prior to sectioning. Tissues were visualized with a Nikon Eclipse TS100
or Olympus IX51 microscope with a 120 W metal halide lamp and a GFP
filter (excitation/emission maxima 474 nm/509 nm). Images were captured
using Image Pro or CoolSnap software.
ACKNOWLEDGMENTS
This study was supported by grants to R.K.-S. from the NIH/NEI
(EY013887 & EY014991), The Foundation Fighting Blindness, The Ellison
Foundation, Lions Eye Foundation and Research To Prevent Blindness.
REFERENCES
1.
Bennett, J, Tanabe, T, Sun, D, Zeng, Y, Kjeldbye, H, Gouras, P et al. (1996).
Photoreceptor cell rescue in retinal degeneration (rd) mice by in vivo gene therapy.
Nat Med 2: 649–654.
2. Kumar-Singh, R and Farber, DB (1998). Encapsidated adenovirus
mini-chromosome-mediated delivery of genes to the retina: application to the
rescue of photoreceptor degeneration. Hum Mol Genet 7: 1893–1900.
3. Campochiaro, PA, Nguyen, QD, Shah, SM, Klein, ML, Holz, E, Frank, RN et al. (2006).
Adenoviral vector-delivered pigment epithelium-derived factor for neovascular
age-related macular degeneration: results of a phase I clinical trial. Hum Gene Ther
17: 167–176.
4. Chevez-Barrios, P, Chintagumpala, M, Mieler, W, Paysse, E, Boniuk, M, Kozinetz, C
et al. (2005). Response of retinoblastoma with vitreous tumor seeding to
adenovirus-mediated delivery of thymidine kinase followed by ganciclovir.
J Clin Oncol 23: 7927–7935.
5. Kumar-Singh, R and Chamberlain, JS (1996). Encapsidated adenovirus
minichromosomes allow delivery and expression of a 14 kb dystrophin cDNA to
muscle cells. Hum Mol Genet 5: 913–921.
6. Parks, RJ, Chen, L, Anton, M, Sankar, U, Rudnicki, MA and Graham, FL (1996).
A helper-dependent adenovirus vector system: removal of helper virus by
Cre-mediated excision of the viral packaging signal. Proc Natl Acad Sci USA
93: 13565–13570.
7. Morral, N, O’Neal, W, Rice, K, Leland, M, Kaplan, J, Piedra, PA et al. (1999).
Administration of helper-dependent adenoviral vectors and sequential delivery of
different vector serotype for long-term liver-directed gene transfer in baboons.
Proc Natl Acad Sci USA 96: 12816–12821.
8. Kim, IH, Jozkowicz, A, Piedra, PA, Oka, K and Chan, L (2001). Lifetime correction of
genetic deficiency in mice with a single injection of helper-dependent adenoviral
vector. Proc Natl Acad Sci USA 98: 13282–13287.
9. Berson, EL (1993). Retinitis pigmentosa. The Friedenwald Lecture. Invest Ophthalmol
Vis Sci 34: 1659–1676.
10. Li, T, Adamian, M, Roof, DJ, Berson, EL, Dryja, TP, Roessler, BJ et al. (1994). In vivo
transfer of a reporter gene to the retina mediated by an adenoviral vector.
Invest Ophthalmol Vis Sci 35: 2543–2549.
11. Bennett, J, Wilson, J, Sun, D, Forbes, B and Maguire, A (1994). Adenovirus
vector-mediated in vivo gene transfer into adult murine retina. Invest Ophthalmol
Vis Sci 35: 2535–2542.
1645
Gene Transfer to Photoreceptors
12. Cashman, SM, Sadowski, SL, Morris, DJ, Frederick, J and Kumar-Singh, R (2002).
Intercellular trafficking of adenovirus-delivered HSV VP22 from the retinal pigment
epithelium to the photoreceptors-implications for gene therapy. Mol Ther 6: 813–823.
13. Hartong, DT, Berson, EL and Dryja, TP (2006). Retinitis pigmentosa. Lancet
368: 1795–1809.
14. Miyoshi, H, Takahashi, M, Gage, FH and Verma, IM (1997). Stable and efficient gene
transfer into the retina using an HIV-based lentiviral vector. Proc Natl Acad Sci USA
94: 10319–10323.
15. Takahashi, M, Miyoshi, H, Verma, IM and Gage, FH (1999). Rescue from
photoreceptor degeneration in the rd mouse by human immunodeficiency virus
vector-mediated gene transfer. J Virol 73: 7812–7816.
16. Bainbridge, JW, Stephens, C, Parsley, K, Demaison, C, Halfyard, A, Thrasher, AJ
et al. (2001). In vivo gene transfer to the mouse eye using an HIV-based lentiviral
vector; efficient long-term transduction of corneal endothelium and retinal pigment
epithelium. Gene Ther 8: 1665–1668.
17. Gruter, O, Kostic, C, Crippa, SV, Perez, MT, Zografos, L, Schorderet, DF et al. (2005).
Lentiviral vector-mediated gene transfer in adult mouse photoreceptors is impaired by
the presence of a physical barrier. Gene Ther 12: 942–947.
18. Pang, J, Cheng, M, Haire, SE, Barker, E, Planelles, V and Blanks, JC (2006). Efficiency
of lentiviral transduction during development in normal and rd mice. Mol Vis
12: 756–767.
19. Narayanan, K and Wadhwa, S (1998). Photoreceptor morphogenesis in the human
retina: a scanning electron microscopic study. Anat Rec 252: 133–139.
20. Reh, TA (2006). Neurobiology: right timing for retina repair. Nature 444: 156–157.
21. Tan, E, Wang, Q, Quiambao, AB, Xu, X, Qtaishat, NM, Peachey, NS et al. (2001). The
relationship between opsin overexpression and photoreceptor degeneration. Invest
Ophthalmol Vis Sci 42: 589–600.
22. Chillon, M, Bosch, A, Zabner, J, Law, L, Armentano, D, Welsh, MJ et al. (1999). Group
D adenoviruses infect primary central nervous system cells more efficiently than those
from group C. J Virol 73: 2537–2540.
23. Von Seggern, DJ, Aguilar, E, Kinder, K, Fleck, SK, Gonzalez Armas, JC, Stevenson, SC
et al. (2003). In vivo transduction of photoreceptors or ciliary body by intravitreal
injection of pseudotyped adenoviral vectors. Mol Ther 7: 27–34.
24. Mallam, JN, Hurwitz, MY, Mahoney, T, Chevez-Barrios, P and Hurwitz, RL (2004).
Efficient gene transfer into retinal cells using adenoviral vectors: dependence on
receptor expression. Invest Ophthalmol Vis Sci 45: 1680–1687.
25. Niwa, H, Yamamura, K and Miyazaki, J (1991). Efficient selection for high-expression
transfectants with a novel eukaryotic vector. Gene 108: 193–199.
26. Young, RW and Bok, D (1969). Participation of the retinal pigment epithelium in the
rod outer segment renewal process. J Cell Biol 42: 392–403.
27. Finnemann, SC, Bonilha, VL, Marmorstein, AD and Rodriguez-Boulan, E (1997).
Phagocytosis of rod outer segments by retinal pigment epithelial cells requires
alpha(v)beta5 integrin for binding but not for internalization. Proc Natl Acad Sci USA
94: 12932–12937.
1646
© The American Society of Gene Therapy
28. Wickham, TJ, Mathias, P, Cheresh, DA and Nemerow, GR (1993). Integrins alpha
v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus
attachment. Cell 73: 309–319.
29. Bergelson, JM, Cunningham, JA, Droguett, G, Kurt-Jones, EA, Krithivas, A, Hong, JS
et al. (1997). Isolation of a common receptor for Coxsackie B viruses and adenoviruses
2 and 5. Science 275: 1320–1323.
30. Rolling, F (2004). Recombinant AAV-mediated gene transfer to the retina: gene
therapy perspectives. Gene Ther 11 (Suppl 1): S26–S32.
31. Bennett, J, Pakola, S, Zeng, Y and Maguire, A (1996). Humoral response after
administration of E1-deleted adenoviruses: immune privilege of the subretinal space.
Hum Gene Ther 7: 1763–1769.
32. Hoffman, LM, Maguire, AM and Bennett, J (1997). Cell-mediated immune response
and stability of intraocular transgene expression after adenovirus-mediated delivery.
Invest Ophthalmol. Vis Sci 38: 2224–2233.
33. Kreppel, F, Luther, TT, Semkova, I, Schraermeyer, U and Kochanek, S (2002).
Long-term transgene expression in the RPE after gene transfer with a high-capacity
adenoviral vector. Invest Ophthalmol Vis Sci 43: 1965–1970.
34. Pang, J, Cheng, M, Stevenson, D, Trousdale, MD, Dorey, CK and Blanks, JC (2004).
Adenoviral-mediated gene transfer to retinal explants during development and
degeneration. Exp Eye Res 79: 189–201.
35. Cashman, SM, Morris, DJ and Kumar-Singh, R (2004). Adenovirus type 5
pseudotyped with adenovirus type 37 fiber uses sialic acid as a cellular receptor.
Virology 324: 129–139.
36. Auricchio, A, Kobinger, G, Anand, V, Hildinger, M, O’Connor, E, Maguire, AM et al.
(2001). Exchange of surface proteins impacts on viral vector cellular specificity and
transduction characteristics: the retina as a model. Hum Mol Genet 10: 3075–3081.
37. Harms, JS and Splitter, GA (1995). Interferon-gamma inhibits transgene expression
driven by SV40 or CMV promoters but augments expression driven by the
mammalian MHC I promoter. Hum Gene Ther 6: 1291–1297.
38. Mizuguchi, H, Koizumi, N, Hosono, T, Ishii-Watabe, A, Uchida, E, Utoguchi, N
et al. (2002). CAR- or alphav integrin-binding ablated adenovirus vectors, but not
fiber-modified vectors containing RGD peptide, do not change the systemic gene
transfer properties in mice. Gene Ther 9: 769–776.
39. Mizuguchi, H, Koizumi, N, Hosono, T, Utoguchi, N, Watanabe, Y, Kay, MA et al.
(2001). A simplified system for constructing recombinant adenoviral vectors
containing heterologous peptides in the HI loop of their fiber knob. Gene Ther
8: 730–735.
40. He, TC, Zhou, S, da Costa, LT, Yu, J, Kinzler, KW and Vogelstein, B (1998). A simplified
system for generating recombinant adenoviruses. Proc Natl Acad Sci USA
95: 2509–2514.
41. Fallaux, FJ, Kranenburg, O, Cramer, SJ, Houweling, A, Van Ormondt, H, Hoeben, RC
et al. (1996). Characterization of 911: a new helper cell line for the titration and
propagation of early region 1-deleted adenoviral vectors. Hum Gene Ther
7: 215–222.
www.moleculartherapy.org vol. 15 no. 9 sept. 2007