10 May 2005 22:26
`
`AR
`
`AR261-BI74-24.tex
`
`XMLPublishSM(2004/02/24)
`P1: JRX
`10.1146/annurev.biochem.74.050304.091637
`
`Annu. Rev. Biochem. 2005. 74:711–38
`doi: 10.1146/annurev.biochem.74.050304.091637
`Copyright c(cid:1) 2005 by Annual Reviews. All rights reserved
`First published online as a Review in Advance on March 11, 2005
`
`GENE THERAPY: Twenty-First Century Medicine
`
`Inder M. Verma and Matthew D. Weitzman
`Laboratory of Genetics, The Salk Institute, La Jolla, California 92037;
`email: verma@salk.edu, weitzman@salk.edu
`
`Key Words
`viral vectors, retrovirus/lentivirus, adenovirus, AAV, clinical trials
`■ Abstract Broadly defined, the concept of gene therapy involves the transfer of
`genetic material into a cell, tissue, or whole organ, with the goal of curing a disease or at
`least improving the clinical status of a patient. A key factor in the success of gene ther-
`apy is the development of delivery systems that are capable of efficient gene transfer in
`a variety of tissues, without causing any associated pathogenic effects. Vectors based
`upon many different viral systems, including retroviruses, lentiviruses, adenoviruses,
`and adeno-associated viruses, currently offer the best choice for efficient gene delivery.
`Their performance and pathogenicity has been evaluated in animal models, and encour-
`aging results form the basis for clinical trials to treat genetic disorders and acquired
`diseases. Despite some initial success in these trials, vector development remains a
`seminal concern for improved gene therapy technologies.
`
`CONTENTS
`
`INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 711
`VIRAL VECTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 713
`RNA VIRUS VECTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 714
`Retroviruses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 714
`DNA VIRUS VECTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 719
`Adenovirus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 719
`Adeno-Associated Virus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 721
`Herpesvirus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 723
`VECTOR TROPISM AND THE SPECIFICITY OF TRANSDUCTION . . . . . . . . . . 725
`VECTOR RECOGNITION, PROCESSING, AND INTEGRATION . . . . . . . . . . . . . . 727
`Gene Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 728
`CLINICAL TRIALS: SUCCESSES AND SETBACKS . . . . . . . . . . . . . . . . . . . . . . . . 729
`PERSPECTIVES: WHAT IS NEXT? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 732
`
`INTRODUCTION
`
`Gene therapy is a form of molecular medicine that has the potential to influence
`significantly human health in this century. It promises to provide new treatments
`for a large number of inherited and acquired diseases (1). The basic concept of
`
`0066-4154/05/0707-0711$20.00
`
`711
`
`Annu. Rev. Biochem. 2005.74:711-738. Downloaded from www.annualreviews.org
`
`by Pohang University of Science and Technology (POSTECH) on 10/03/13. For personal use only.
`
`SKI Exhibit 2036
`Page 1 of 30
`
`

`

`10 May 2005 22:26
`
`AR
`
`AR261-BI74-24.tex
`
`XMLPublishSM(2004/02/24)
`
`P1: JRX
`
`712
`
`VERMA (cid:1) WEITZMAN
`
`gene therapy is simple: introduce into target cells a piece of genetic material that
`will result in either a cure for the disease or a slowdown in the progression of
`the disease. To achieve this goal, gene therapy requires technologies capable of
`gene transfer into a wide variety of cells, tissues, and organs. One of the biggest
`stumbling blocks to successful widespread application of such genetic treatments
`is the development of safe and effective vectors with which to ferry genetic material
`into a cell.
`The process of gene delivery and expression is known as transduction. Suc-
`cessful transduction requires overcoming a number of obstacles that are common
`to all vector systems (2). The first issue to be addressed is that of production. An
`ideal vector should be one that can be produced in a highly concentrated form,
`using a convenient and reproducible production scheme. This has been a chal-
`lenge for many of the currently used vector systems, but in many cases creative
`approaches have overcome this barrier. The vector must be capable of targeting
`the cell type most appropriate for the disease, whether it be dividing or nondi-
`viding cells. Understanding of the transduction process through studies of vector
`uptake, intracellular trafficking, and gene regulation has facilitated the develop-
`ment of efficient vehicles for gene delivery. In many cases it would be desirable
`to achieve stable, sustained gene expression, which requires either integration of
`the vector DNA into the host DNA or maintenance as an episome. When us-
`ing integrating vector systems, it is important to consider the potential hazards
`of insertional mutagenesis, and thus vectors capable of site-specific integration
`will be attractive. In many cases, expression of the therapeutic gene will require
`exquisite regulation, and thus the transcriptional unit must be capable of respond-
`ing to manipulations of its regulatory elements. Finally, no pathogenic or adverse
`effects should be elicited by vector transduction, including undesirable immune
`responses.
`Vectors that have been developed to overcome these obstacles fall into two
`broad categories: nonviral and viral vectors (1). The nonviral vectors consist of
`naked DNA delivered by injection, liposomes (cationic lipids mixed with nu-
`cleic acids), nanoparticles, and other means. Although nonviral vectors can be
`produced in relatively large amounts and are likely to present minimal toxic or
`immunological problems, presently they suffer from inefficient gene transfer. In
`addition, expression of the foreign gene tends to be transient, precluding their
`application to many diseased states in which sustained and high-level expression
`of the transgene is required. It is likely that future gene therapy protocols will use
`novel innovations to improve on the efficiency of nonviral vector systems, often
`building upon observations from viral vector transduction. Viral vectors are de-
`rived from viruses with either RNA or DNA genomes and are represented as both
`integrating and nonintegrating vectors. The former holds the promise of lifelong
`expression of the deficient gene product. Efficient gene transduction can also be
`achieved from vectors that are maintained as episomes, especially in nondividing
`cells.
`
`Annu. Rev. Biochem. 2005.74:711-738. Downloaded from www.annualreviews.org
`
`by Pohang University of Science and Technology (POSTECH) on 10/03/13. For personal use only.
`
`SKI Exhibit 2036
`Page 2 of 30
`
`

`

`10 May 2005 22:26
`
`AR
`
`AR261-BI74-24.tex
`
`XMLPublishSM(2004/02/24)
`
`P1: JRX
`
`GENE THERAPY
`
`713
`
`VIRAL VECTORS
`
`The basic concept of viral vectors is to harness the innate ability of viruses to de-
`liver genetic material into the infected cell. In general, the major preoccupation of
`viruses is to replicate and produce copious amounts of progeny. Most viruses gain
`little by killing the host, but unfortunately many viral infections lead to deleterious
`effects on the host, accompanied by destruction of infected host cells. Damag-
`ing effects can be caused by induction of genes whose products are hazardous
`to the host or by acquiring host genomic material that can lead to pathogenesis.
`The basic principle of turning these pathogens into delivery systems relies on the
`ability to separate the components needed for replication from those capable of
`causing disease (see Figure 1). The first step of viral vector design is, therefore,
`
`Figure 1 Principle of generating a viral vector. (a) Converting a virus into a recombinant
`vector. Schematic of a generic viral genome is shown with genes that are involved in replica-
`tion, production of the virion, and pathogenicity of the virus. The genome is flanked by cis-
`acting sequences that provide the viral origin of replication and the signal for encapsidation.
`The packaging construct contains only genes that encode functions required for replication
`and structural proteins. The vector construct contains the essential cis-acting sequences and
`the transgene cassette that contains the required transcriptional regulatory elements. (b) The
`packaging and vector constructs are introduced into the packaging cell by transfection, by
`infection with helper virus, or by generating stable cell lines. Proteins required for replica-
`tion and assembly of the virion are expressed from the packaging construct, and the repli-
`cated vector genomes are encapsidated into virus particles to generate the recombinant viral
`vector.
`
`Annu. Rev. Biochem. 2005.74:711-738. Downloaded from www.annualreviews.org
`
`by Pohang University of Science and Technology (POSTECH) on 10/03/13. For personal use only.
`
`SKI Exhibit 2036
`Page 3 of 30
`
`

`

`10 May 2005 22:26
`
`AR
`
`AR261-BI74-24.tex
`
`XMLPublishSM(2004/02/24)
`
`P1: JRX
`
`714
`
`VERMA (cid:1) WEITZMAN
`
`to identify the viral sequences required for replication, assembly of viral parti-
`cles, packaging of the viral genome, and delivery of the transgene into the target
`cells. Next, dispensable genes are deleted from the viral genome to reduce repli-
`cation and pathogenicity, as well as expression of immunogenic viral antigens.
`The gene of interest together with transcriptional regulatory elements (referred to
`as the transgene) are inserted into the vector construct, and a recombinant virus
`is generated by supplying the missing gene products required for replication and
`virion production. The more genes that are removed from the virus, the more repli-
`cation defective the vector will be, and there is less chance of recombination to
`generate the infectious parental virus. The nature of the virus biology will usu-
`ally determine the means of production. For example, retroviruses are produced
`in packaging cell lines, and vector particles accumulate in the culture medium.
`In contrast, adenovirus and adeno-associated virus (AAV) vectors are generally
`produced from transfections, and cells must be lysed to liberate the viral particles.
`In this review, we describe the salient features and applications of some of the most
`commonly used viral vectors. There are a number of other emerging vector systems
`that are still in their infancy and have been extensively discussed in other excellent
`reviews (1, 3).
`
`RNA VIRUS VECTORS
`
`The most commonly used RNA virus vectors are derived from retroviruses, and
`these were among the first viral delivery systems to be developed for gene therapy
`applications. Retroviruses are a large family of enveloped RNA viruses found in
`all vertebrates, and they can be classified into oncoretroviruses, lentiviruses, and
`spumaviruses.
`
`Retroviruses
`The enveloped virus particle contains two copies of the viral RNA genome, sur-
`rounded by a cone-shaped core (for an in-depth review of retrovirus biology, see
`Reference 4). The viral RNA contains three essential genes, gag, pol, and env, and
`is flanked by long terminal repeats (LTR). The gag gene encodes for the core pro-
`teins capsid, matrix, and nucleocapsid, which are generated by proteolytic cleavage
`of the gag precursor protein. The pol gene encodes for the viral enzymes protease,
`reverse transcriptase, and integrase, which are usually derived from the gag-pol
`precursor. The env gene encodes for the envelope glycoproteins, which mediate
`virus entry. Oncoretroviruses are simple viruses encoding only the structural genes
`gag, pol, and env, whereas lentiviruses and spumaviruses have a more complex
`organization and encode for additional viral proteins (see Figure 2).
`After binding to its receptor, the viral capsid containing the RNA genome enters
`the cell through membrane fusion. The viral RNA genome is subsequently con-
`verted into a double-stranded proviral DNA by the viral enzyme reverse
`
`Annu. Rev. Biochem. 2005.74:711-738. Downloaded from www.annualreviews.org
`
`by Pohang University of Science and Technology (POSTECH) on 10/03/13. For personal use only.
`
`SKI Exhibit 2036
`Page 4 of 30
`
`

`

`10 May 2005 22:26
`
`AR
`
`AR261-BI74-24.tex
`
`XMLPublishSM(2004/02/24)
`
`P1: JRX
`
`GENE THERAPY
`
`715
`
`ψ
`
`gag
`
`LTR
`
`Retrovirus
`
`pol
`
`Lentivirus
`
`env
`
`LTR
`
`ψ
`
`gag
`
`LTR
`
`pol
`
`vpu
`
`vif
`vpr
`tat
`
`env
`
`LTR
`
`nef
`
`rev
`
`L5
`E3
`
`Adenovirus
`
`L2
`
`L3
`
`L4
`
`E2
`
`L1
`VA
`
`IVa2
`
`ψ
`
`E1A E1B
`
`ITR
`
`Adeno-associated virus
`
`Rep
`
`Cap
`
`ITR
`
`ITR
`
`E4
`
`ITR
`
`X Essential elements retained in vectors
`
`X Genes supplied by packaging construct / cell line
`
`X Nonessential genes often deleted
`
`Figure 2 Schematics of the commonly used viruses that are converted to recombinant
`viral vectors. The colored boxes indicate genes or cis-acting elements that are either
`essential [and therefore retained in vectors (red ) orsupplied by packaging constructs
`or cell lines (green)] or that are nonessential and often deleted (blue). Only the major
`genetic elements are shown, and viruses are not drawn to scale. In second- and third-
`generation derivatives of each vector system, more viral genes are deleted (see text for
`details).
`
`Annu. Rev. Biochem. 2005.74:711-738. Downloaded from www.annualreviews.org
`
`by Pohang University of Science and Technology (POSTECH) on 10/03/13. For personal use only.
`
`SKI Exhibit 2036
`Page 5 of 30
`
`

`

`10 May 2005 22:26
`
`AR
`
`AR261-BI74-24.tex
`
`XMLPublishSM(2004/02/24)
`
`P1: JRX
`
`716
`
`VERMA (cid:1) WEITZMAN
`
`transcriptase. The proviral DNA is associated with viral proteins, including nu-
`cleocapsid, reverse transcriptase, and integrase, in a preintegration complex and
`translocates to the nucleus, where the integrase mediates integration of the provirus
`into the host cell genome. Host cell transcription factors initiate transcription from
`the LTR, and new viral particles are formed at the plasma membrane. Gag-pol and
`gag precursors assemble together with two copies of viral RNA, and env glyco-
`proteins are incorporated into the viral membrane during the budding process. In
`the newly formed virion, gag and gag-pol precursors are subjected to processing
`by the viral protease, which results in maturation of the virion.
`
`ONCORETROVIRAL VECTORS These vectors have been derived from a number of
`different oncoretroviruses, including murine leukemia virus (MLV), spleen necro-
`sis virus, Rous sarcoma virus, and avian leukosis virus. Replication-defective MLV
`vectors are generated by replacing all viral protein encoding sequences with the
`exogenous promoter-driven transgene of interest. In addition to the packaging sig-
`nal, the vectors retain viral LTRs and adjacent sequences, which are essential for
`reverse transcription and integration. Vector RNA production is either driven by
`the U3 region of the LTR or by a CMV/LTR hybrid with higher transcriptional ac-
`(cid:1)
`tivity. In these vectors, the 3
`U3 region of the LTR is intact and is copied over to the
`(cid:1)
`5
`LTR during reverse transcription, allowing efficient integration and LTR-driven
`transgene expression in the transduced cell.
`Packaging of retroviral vectors is achieved by providing the structural proteins
`in trans in the packaging cells. The first packaging cell lines expressed gag, pol, and
`env from a complete proviral DNA, lacking only the packaging signal (5). How-
`ever, sequence homology between the vector and packaging constructs facilitated
`recombination, allowing generation of a replication competent virus. To prevent
`recombination, packaging cells have been developed that express gag/pol and env
`from separate constructs. Furthermore, expression from the packaging constructs
`is no longer driven by the viral LTR but by constitutive promoters that allow a
`high level of virus production (5, 6). To avoid the laborious and time-consuming
`practice of generating cell lines, high-titer vectors can be produced by transient
`transfection (7). Viruses are recovered from the supernatants of actively growing
`producer cells.
`Major concerns in the use of retroviral vectors are the possibility of vector
`mobilization and recombination with defective (endogenous) retroviruses in the
`target cell. This prompted the development of self-inactivating vectors (8). In these
`vectors LTR-driven transcription is prevented in transduced cells by deletion in the
`U3, and transgene expression is driven instead by an internal promoter, allowing
`the use of regulated and tissue specific promoters.
`
`In addition to gag, pol, and env, lentiviruses encode three to
`LENTIVIRAL VECTORS
`six additional viral proteins, which contribute to virus replication and persistence
`of infection (for in-depth reviews of lentiviruses, see References 9 and 10). Two
`of the accessory proteins, tat and rev, are present in all lentiviruses and mediate
`
`Annu. Rev. Biochem. 2005.74:711-738. Downloaded from www.annualreviews.org
`
`by Pohang University of Science and Technology (POSTECH) on 10/03/13. For personal use only.
`
`SKI Exhibit 2036
`Page 6 of 30
`
`

`

`10 May 2005 22:26
`
`AR
`
`AR261-BI74-24.tex
`
`XMLPublishSM(2004/02/24)
`
`P1: JRX
`
`GENE THERAPY
`
`717
`
`transactivation of viral transcription and nuclear export of unspliced viral RNA,
`respectively. Although vectors based on Simian, Equine and Feline lentiviruses
`have been developed (11), we focus on human immunodeficiency virus type 1
`(HIV-1)-based vectors because they have been most extensively studied.
`In addition to gag, pol, and env, HIV-1 encodes six accessory proteins (tat, rev,
`vif, vpr, nef, and vpu). The retroviral vector design provided an excellent template
`for development of lentiviral vectors. The HIV-1-based lentiviral vectors are devoid
`of all viral sequences apart from essential cis-acting sequences, including the LTRs
`and the packaging signal (see Figure 2). The rev responsive element (RRE) is also
`included in the vector RNA. The viral rev protein is provided in trans to ensure
`efficient nuclear export of the full-length viral RNA genomes through binding to the
`RRE. Initially the endogenous LTR was used to drive vector RNA expression via
`transactivation by the tat protein (12), but further generations of the vector utilize a
`CMV/LTR hybrid promoter, which increases vector production and allows vector
`production to be independent of tat expression (13). The accessory genes, vif, vpr,
`nef, and vpu, are dispensable for lentiviral vector production and transduction, and
`they are deleted from the packaging construct (13). The central poly purine tract
`(cPPT), from the pol ORF, can be included in cis to improve nuclear import of
`the proviral DNA and hence accelerate transduction (14, 15). Furthermore, the
`biosafety of vectors is improved by the development of self-inactivating vectors,
`which are less likely to be mobilized following infection with HIV (16, 17).
`The HIV-1 env glycoprotein has a highly restricted host range in that it infects
`cells containing CD4 and coreceptors. To broaden the host range of lentiviral
`vectors, they can be pseudotyped with the vesicular stomatitis virus glycoprotein
`(VSV-G) env, which is provided in trans and imparts a wide tropism (12). Vectors
`are harvested from the supernatant, and those pseudotyped with VSV-G can be
`concentrated to produce high-titer preps. Titers can be determined using assays that
`measure the amount or activity of proteins incorporated in the vector particles, such
`as the p24gag ELISA assay. Stable packaging cell lines have now been developed,
`in which the producer cells express the structural proteins from minimal packaging
`constructs and expression is driven by an inducible promoter to minimize the
`toxicity of the VSV-G envelope protein (18, 19). Other viral glycoproteins have also
`been used to pseudotype lentiviral vectors and provide altered cell tropism (20, 21).
`
`FOAMY VIRAL VECTORS Foamy viral (FV) vectors have recently been developed
`and are quite similar to retroviral and lentiviral vectors (22). The FV genome
`contains three additional ORFs (tas/bel1, bel-2, and bel-3), with tas/bel1 being
`the coactivator of viral transcription (23). In addition to the packaging signal that
`(cid:1)
`(cid:1)
`consists of the 5
`-untranslated region and the 5
`portion of the gag ORF present in all
`(cid:1)
`retroviral vectors, FV vectors contain the 3
`region of the pol ORF, which is critical
`(cid:1)
`U3
`for efficient packaging of these vectors. Similar to other retroviral vectors, the 5
`region of the LTR in the vectors’ plasmid can been replaced by a CMV promoter,
`which increases vector expression and makes vector production independent of
`tas/bel1.
`
`Annu. Rev. Biochem. 2005.74:711-738. Downloaded from www.annualreviews.org
`
`by Pohang University of Science and Technology (POSTECH) on 10/03/13. For personal use only.
`
`SKI Exhibit 2036
`Page 7 of 30
`
`

`

`10 May 2005 22:26
`
`AR
`
`AR261-BI74-24.tex
`
`XMLPublishSM(2004/02/24)
`
`P1: JRX
`
`718
`
`VERMA (cid:1) WEITZMAN
`
`FV vectors are produced by transient transfection of the vector construct as well
`as the packaging constructs encoding for the structural proteins gag, pol, and env
`in 293T cells (22). Because the FV envelope has a broad cellular host range, it is
`used by the vector, and therefore, the env sequences are included in the packaging
`construct or expressed from a separate construct. The FV env contains, in contrast
`to other retroviruses, an ER sorting signal, which allows FV particles to bud from
`intracellular membranes, and therefore, the majority of the infectious virions is cell
`associated (24). Consequently, the infectious particles have to be released from the
`packaging cells by a freeze-thawing process.
`Because retroviral vectors were some of the first to be utilized for gene delivery,
`they have been extensively used in many applications (2, 25). Their ability to
`integrate and provide long-term gene expression has made them particularly useful
`for generating stable cell lines that express a transgene of choice and for marking
`studies of cell lineage. Their dependence upon cell division has restricted in vivo
`applications to gene delivery in actively dividing tissues, such as stem cells and
`cancer cells. In contrast, lentivirus vectors have been used for gene delivery in vivo
`by direct administration in many organs, including brain, eye, liver, and muscle
`(26). They have also been used for ex vivo transduction of hematopoietic cells,
`followed by bone marrow transplantation (27).
`The limited cellular tropism of the natural envelope of wild-type viruses is one
`of the barriers for retroviral transduction. However, retroviruses have the ability to
`incorporate env glycoproteins from related as well as unrelated viruses, thus allow-
`ing pseudotyping with alternative glycoproteins. A number of different envelopes
`have been used to generate pseudotyped retroviral vectors with broad host ranges,
`including the VSV-G glycoprotein or the amphotropic MLV envelope. Pseudo-
`typing also allows transfer of specific tropisms to the vector. Neurotropism and
`retrograde axonal transport was accomplished by the vector via pseudotyping with
`the G protein of Mokola lyssaviruses (28), and the filovirus (Ebola Zaire) envelope
`supported transduction of airway epithelia (21). Interestingly, the entry pathway
`of the retroviral vector has minimal effect on the transduction efficiency.
`Reverse transcription and nuclear translocation of the preintegration complex
`are thought to be limiting steps in retroviral transduction, especially in terminally
`differentiated postmitotic cells. Proviral DNA synthesis of all retroviruses depends
`strongly on cellular conditions, and low nucleoside pools or absence of cellular
`cofactors might explain the incomplete reverse transcription in quiescent or sta-
`tionary cells. In contrast to other retroviral vectors, FV vector particles can contain
`fully reverse-transcribed viral DNA, owing to activation of the process of reverse
`transcription before virus assembly. This suggests that FV vector gene transfer
`might be more efficient in certain postmitotic cells in which reverse transcription
`is limited.
`Because HIV-1 is a human pathogen, there are biosafety concerns about the
`use of HIV-1-based lentiviral vectors. The current HIV-1 lentiviral vector sys-
`tem is depleted of all accessory proteins, and viral sequences in the vectors have
`been minimized. Thus, the replication of these vectors is highly disabled, and the
`
`Annu. Rev. Biochem. 2005.74:711-738. Downloaded from www.annualreviews.org
`
`by Pohang University of Science and Technology (POSTECH) on 10/03/13. For personal use only.
`
`SKI Exhibit 2036
`Page 8 of 30
`
`

`

`10 May 2005 22:26
`
`AR
`
`AR261-BI74-24.tex
`
`XMLPublishSM(2004/02/24)
`
`P1: JRX
`
`GENE THERAPY
`
`719
`
`possibility of homologous recombination is minimized. In addition, codon op-
`timization of the packaging construct further decreases the risk of homologous
`recombination. The use of vectors based on other lentiviruses might eliminate
`some of the concerns (11); however, the risk associated with the introduction of
`nonhuman lentiviral vectors in human tissues is unknown.
`
`DNA VIRUS VECTORS
`
`Of the vectors derived from viruses with DNA genomes, the most prominent are
`those based on adenovirus (Ad) and the adeno-associated virus (AAV). Aden-
`oviruses contain a double-stranded DNA (dsDNA) genome of ∼36 kb, whereas
`AAVs consist of a single-stranded DNA molecule that is relatively small (∼4.7 kb).
`The basic principals of vector design (Figure 1) also apply to vectors derived from
`DNA viruses. However, there are important practical differences in terms of con-
`struction, production, and purification of these vectors.
`
`Adenovirus
`Ads have been isolated from a large number of species and tissue types (for an
`in-depth review of Ad biology see Reference 29). The human Ad family consists of
`more than 50 serotypes that can infect and replicate in a wide range of organs, such
`as the respiratory tract, the eye, urinary bladder, gastrointestinal tract, and liver.
`The Ad genome consists of a double-stranded linear DNA molecule (size: ∼36
`kb) with overlapping transcription units on both strands. Extensive splicing results
`in the production of over 50 proteins; 11 of which are structural virion proteins.
`The viral life cycle occurs in an early and a late phase, divided by the onset of viral
`DNA replication. Adenoviral genes fall into three major groups, depending on
`the time course of their expression during the viral replicative cycle: early (E1A,
`E1B, E2, E3, and E4), delayed (IX and IVa2), and the major late transcription unit
`(see Figure 2). The latter is processed into five mRNAs (L1–L5) that share the
`same carboxy terminus. These transcription units are transcribed by the cellular
`RNA polymerase II, whereas the viral-associated (VA) RNA is transcribed by
`RNA polymerase III. The viral genome contains two identical origins for DNA
`replication within each terminal repeat. The E2 region encodes proteins required
`for replication, including the viral polymerase, and proteins from the E1 and E4
`regions also contribute to efficient DNA replication. The gene products of the E3
`region are involved in immune surveillance and suppression but are nonessential
`for infection in vitro.
`The Ad genome is packaged in a nonenveloped icosahedral protein capsid. The
`fiber protein projects from the virion, and the carboxy-terminal knob domain forms
`a high-affinity complex with a host cell surface receptor protein. For the majority
`of Ad serotypes this receptor is the Coxsackie-adenovirus receptor (CAR) (30).
`In addition to attachment, efficient virus internalization requires an interaction
`
`Annu. Rev. Biochem. 2005.74:711-738. Downloaded from www.annualreviews.org
`
`by Pohang University of Science and Technology (POSTECH) on 10/03/13. For personal use only.
`
`SKI Exhibit 2036
`Page 9 of 30
`
`

`

`10 May 2005 22:26
`
`AR
`
`AR261-BI74-24.tex
`
`XMLPublishSM(2004/02/24)
`
`P1: JRX
`
`720
`
`VERMA (cid:1) WEITZMAN
`
`between the viral penton base and the cellular integrin αv receptor (31). After
`entry, the virus rapidly escapes from the endosome, and transport to the nucleus
`is accompanied by gradual disruption of the virus particle (32). The viral genome
`is imported through the nuclear pore and associates with the nuclear matrix to
`facilitate initiation of the primary transcription events (33). The Ad genome is
`transcribed and replicated at discrete replication centers in the nucleus of the
`infected cell (34), and the viral DNA does not normally integrate into the host
`genome.
`Adenoviral infection causes an initial nonspecific host response with synthesis
`of cytokines (tumor necrosis factor as well as interleukin 1 and 6), followed by a
`specific response of cytotoxic T lymphocytes directed against virus-infected cells
`that display viral peptide antigens (29). In addition, there is activation of B cells
`and the necessary CD4-positive T cells, leading to a humoral response. Serologic
`surveys found antibodies against Ad serotypes 1, 2, and 5 in 40% to 60% of
`children. The immune response of the host against adenoviral proteins is the major
`hurdle to the efficient and safe use of adenoviral vectors.
`Most adenoviral vectors are derived from Ad serotype 5; however, Ad vectors
`have also been generated from other serotypes, including human Ad2, Ad7, and
`Ad4 as well as nonhuman viruses. Replication-defective Ad vectors are designed
`by replacing crucial adenoviral coding regions (35). In the first generation of
`Ad vectors, the E1 gene was replaced with the transgene (36). Because E1A is
`the principal protein that activates the expression of other Ad transcription units’
`genes (37), and other E1 proteins play crucial roles in viral replication, these
`vectors are replication defective on most cell lines. E1-deleted vectors can be
`propagated in cell lines that provide the E1 gene products in trans, such as the
`human 293 cell line (38). Transgenes of up to 4.7–4.9 kb can be incorporated
`into E1-deleted adenoviral vectors (39). The cloning capacity of Ad vectors can
`be further increased by deletion of additional dispensable sequences from the
`Ad genome, such as the nonessential E3 region (40). Combining the E1 and E3
`deletions provides a total cloning capacity of 8.3 kb in one mutant virus. Newly
`developed E1-complementing cell lines with a minimal amount of viral sequences
`help reduce the chances of homologous recombination between the vector and
`the host genome, and thus, less replication-competent Ad is generated (41, 42).
`Vectors can be produced at titers of up to 1013 particles/ml and are purified by
`cesium chloride gradient ultracentrifugation or column chromatography.
`Although E1-deleted viruses are defective for replication, in some cell types
`they can produce virus proteins that serve as foreign antigens to induce a cellu-
`lar immune response. Many attempts have been made to reduce immunogenicity
`by engineering the second generation of Ad vectors that are additionally deleted
`in other viral transcription units, such as E2 and E4. Because the E2 region en-
`codes proteins that are essential for replication of the viral chromosome, it has
`to be provided in trans in the packaging cells. A number of deletions have been
`introduced into the E4 region, which encodes proteins required for efficient viral
`DNA replication and late protein synthesis. Partial deletions of the E4 region can
`
`Annu. Rev. Biochem. 2005.74:711-738. Downloaded from www.annualreviews.org
`
`by Pohang University of Science and Technology (POSTECH) on 10/03/13. For personal use only.
`
`SKI Exhibit 2036
`Page 10 of 30
`
`

`

`10 May 2005 22:26
`
`AR
`
`AR261-BI74-24.tex
`
`XMLPublishSM(2004/02/24)
`
`P1: JRX
`
`GENE THERAPY
`
`721
`
`produce viable vectors, and E4-complementing cell lines have been developed for
`more extensive deletions (43–46). Although deletions in the E4 region increase
`the cloning capacity of the Ad vector, some reports indicate that the E4 region may
`exert a positive effect on long-term expression (47–49).
`Theoretically it should be possible to reduce viral genes to a minimum and create
`“gutted” vectors that carry no viral sequences, apart from the inverted terminal
`repeats (ITRs) and the cis-acting packaging signal (50). These vectors require
`helper viruses for propagation, generating a problem in the purification of a helper-
`free virus. An important step toward a third generation of Ad vectors was the
`development of high-capacity, helper-dependent vectors based on the Cre/loxP-
`system of site-specific DNA excision (51–53). Using the CreloxP-system, 25 kb
`of adenoviral genome can be deleted from an Ad vect