DNA polyplexes for gene therapy

Surg Oncol Clin N Am 11 (2002) 697–716

Protein/DNA polyplexes for gene therapy Richard J. Cristiano, PhD Department of Genitourinary Medical Oncology, The University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Boulevard, Box 427, Houston TX 77030, USA

Gene therapy is based on the ability to deliver and express a gene in a cell to either correct or alter cell function for therapeutic purposes. Although the choice of possible therapeutic genes is not limiting and will certainly increase based on the recent completion of sequencing the human genome, the ability to obtain efficient and specific gene delivery still remains one of the most important barriers to overcome for the successful use of gene therapy [1]. Although research on using viruses as delivery vectors continues to increase, limitations specific to viruses, such as toxicity and immunogenicity, may never be overcome. These limitations were recently brought to light when a patient with ornithine transcarbamylase deficiency died from an intravenous dose of recombinant adenovirus [2]. In contrast, nonviral based vectors have grown in popularity due to a lack of these limitations. The synthetic basis of nonviral vectors allows most to have, but not be limited to, the following characteristics: (1) the capability of targeting delivery to specific cells; (2) no current limitation on the size or type of nucleic acid that can be delivered; (3) no intact viral component and therefore safe for the recipient; (4) the ability to mediate delivery to a large number of cells regardless of cell division status; and (5) the potential to be completely synthetic. One promising vector demonstrating these characteristics is referred to as a Protein/ DNA complex or polyplex (poly ¼ polycations) that is based on the use of ‘‘molecular conjugates.’’ A molecular conjugate is defined as a delivery-specific vector component to which a nucleic acid or DNA-binding component has been attached. The combining of the conjugate with DNA produces the polyplex that can then mediate general or specific delivery, depending upon the components used. Since the initial use of a molecular conjugate for gene delivery over 14 years ago [3], we have learned that to obtain efficient, targeted gene delivery, the resulting polyplex must contain components capable of: (1) DNA-binding and compaction, so that the vector is sufficiently small

E-mail address: [email protected] (R.J. Cristiano). 1055-3207/02/$ - see front matter Ó 2002, Elsevier Science (USA). All rights reserved. PII: S 1 0 5 5 - 3 2 0 7 ( 0 2 ) 0 0 0 2 2 - 4

698

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

for cellular internalization; (2) targeting for cell-specific delivery of the vector; (3) endosome lysis to ensure the entry of functional DNA into the cell; (4) nuclear translocation to ensure efficient gene expression; and (5) DNA persistence or integration to obtain appropriate levels and duration of gene expression (Fig. 1). In this review, the components that have been used in these roles, the rationale behind their use, and the methods used to produce and analyze both molecular conjugates and Protein/DNA polyplexes are discussed. The advantages and disadvantages of using this vector in gene therapy in comparison with other nonviral and viral gene delivery vectors as well as present and future applications are also presented.

DNA-binding components Central to the performance of molecular conjugate based vectors is the DNA binding component. In addition to DNA binding, this component functions to compact DNA and to provide a point of attachment to the vector for other components. Typically, polycations have been used in this role based on cost and availability (Table 1). Poly-L-lysine (PLL), a synthetic,

Fig. 1. The general functions of a Protein/DNA polyplex required for efficient gene delivery and expression. 1) DNA-binding and compaction, 2) targeting, 3) endosome lysis, 4) nuclear translocation, and 5) DNA persistence or integration.

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

699

Table 1 Polycations used in molecular conjugate formation Polycation

Structure

Molecular weight

Endosomolytic properties

Cost

Spermine Spermidine Poly-L-Lysine Polyethylenimine

Linear Linear Linear Branched Linear Branched

150–250 200 200–300,000 25,000–800,000 22,000 3000

No No No Yes Yes Yes

25 g/$250 25 g/$250 1 g/$250 500 ml/$75 250 ll/$250 5 g/$60

Dendrimer

linear molecule consisting of repeating lysine residues, was first used in this role based on this molecule’s ability to efficiently bind and compact DNA and on it being a cost-effective reagent (see Table 1). Studies have demonstrated that the ability of polycations to bind and compact DNA depends on size. Long-chain versions of PLL have been shown to bind DNA with greater affinity and produce smaller particles than short-chain versions as small as 19 lysines (Fig. 2) [4]. These smaller particles are also much more stable in buffer-containing physiological salt than particles that result from using lower molecular weight PLLs (see Fig. 2). Although the resulting

Fig. 2. Analysis of particle size of PLL/DNA polyplexes made with PLL’s of various molecular weights. Vector was made in Hepes Buffer and then the NaCl concentration was increased to 0.15 M along with an overnight incubation at room temperature. Particle size was measured in nanometers (nm).

700

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

PLL/DNA polyplex can interact with cells through simple charge interaction with glycoproteins expressed at the cell surface [5], PLL has been most commonly used in molecular conjugate formation with a wide range of ligands to target vector delivery through many different receptor mediated endocytotic pathways [3,4,6–20]. Unfortunately, either alone or as part of a molecular conjugate, gene delivery is limited using PLL due to an inability to perform endosome lysis, which is required to ensure passage of the nucleic acid from the endosome compartment into the cytoplasm in a functional form (see Fig. 1). This has led to increased focus on developing synthetic components that can perform endosome lysis in addition to these other functions, but without increasing the complexity of vector composition. As a result, branched, synthetic polycations such as polyethylenimine (PEI) and dendrimers are now receiving greater attention, based on the ability of these molecules to mediate endosome lysis in addition to DNA binding and compaction (see Table 1) [21]. These molecules are also available in a wide range of molecular weights, but can vary in cost. Branched polyethylenimine has been used to a greater degree recently because it is very inexpensive, easily produced in large quantities, and highly efficient at mediating endosome lysis. Based on this last point, vector formulations involving PEI have now been shown to rival adenovirus in overall gene delivery efficiency, at least in vitro (Fig. 3). Although highly efficient gene delivery has been obtained using PEI/ DNA-based vectors on many different types of cultured cells, limited delivery has been obtained in vivo and is usually accompanied by toxicity, especially when PEI in molecular weights greater than 25,000 or high PEI/DNA ratios ([5/1) are used [22]. Attempts have been made to overcome this limitation through galactosylation or glucosylation of the amine groups, but delivery efficiency is reduced and must be compensated for by using more vector [23]. An alternative approach has been to develop improved types of PEI, such as a linear version that has reduced toxicity [24]. This form has demonstrated efficient gene delivery to the lung and to tumor cells in vivo with little or no toxicity [25,26]. The high cost of producing this form based on the complexity of synthesis reduces its utility, however (see Table 1). As a result, toxicity is still an important limitation that must be overcome if polycations such as PEI are going to be used in future gene therapy applications. In contrast to these synthetic components, other naturally occurring DNA-binding components have been used in Protein/DNA polyplex formation to decrease toxicity. The polycations spermine and spermidine have been used for DNA compaction, but lack endosomolytic activity and tend to have a lower binding affinity for DNA based on small size [27]. Additionally, other naturally occurring agents have been used, such as histone proteins. In one study, histone proteins were modified by galactosylation and used to target and deliver the attached DNA to the hepatocellular carcinoma cell line HepG2 [28]. Because these proteins are naturally

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

701

Fig. 3. Transduction comparison between DNA, PLL/DNA polyplex, PEI/DNA polyplex, and recombinant adenovirus (Adv/bgal). The analysis was performed with the cell line H1299 (nonsmall cell lung cancer cell line) and cells were either transduced with 2.5 lg of DNA (non-viral vectors) or 10 viral particles/cell (adenovirus). All non-viral and viral vectors carried the b-gal gene under the control of the cytomegalovirus enhancer/promoter.

occurring, this may result in better DNA-binding/compaction (due to a natural ability to compact DNA) and reduced toxicity [28]. Although the properties of these DNA-binding components vary, the process behind DNA-binding and compaction is typically based on simple self-assembly through ionic interactions between the positive charge of the polycation’s amine groups and the negative charge of the nucleic acid’s phosphate backbone. This contributes to a nondamaging, self-assembling charge neutralization and compaction of the DNA molecule, the extent of which can be viewed or analyzed by agarose gel electrophoresis (charge neutralization of DNA), electron microscopy, atomic force microscopy, or instruments using light scattering (particle size and charge analyzer) [3,4, 16,29]. Because charge neutralization is the basis for Protein/DNA polyplex vector formation, theoretically any nucleic acid of any size can be compacted and delivered. This is in sharp contrast to viruses that have packaging constraints on the size and type of nucleic acid that can be compacted and delivered. As a result, molecular conjugates have been used to deliver small oligonucleotides to large artificial chromosomes [30,31]. Once the nucleic acid is bound or compacted by the polycation, the resulting vector can range in size from as small as 10 nm to greater than 500 nm, depending upon the type and amount of polycation used. The resulting polyplex can also vary in structure from a doughnut-shaped toroid to a ball-shaped structure, again depending upon the type and amount of polycation used [27,32].

702

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

Most importantly, it is critical that the polyplexes are sufficiently small to be internalized by cells and to pass through liver fenestrations or between endothelial cells to gain access to target cells. This size constraint is usually in the 100 to 300 nm range, which is typically the size of an endosome compartment or a liver fenestration [33]. In addition to DNA binding, compacting, and endosome lysis, the polycation serves the important function of an attachment point for other vector components to the vector. Critical to attaching proteins or peptides to the DNA-binding component is the correct coupling of the two molecules to ensure that the functions of the component (ie, receptor binding, endosomal lysis, DNA binding, etc) and the polycation are maintained. A wide range of linkages can be generated using several different chemicals. The watersoluble carbodiimide 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC), has been used to link ASOR to PLL [3,16]. This carbodiimide activates carboxyl groups to form a nucleophile that targets amine groups, resulting in the formation of a covalent bond between the carboxyl groups of the protein and the amine groups of PLL. Unfortunately, this chemical is also known as a ‘‘zero-length cross-linker’’ that results in little or no spacing between the components and polycation. This can affect component function based on steric hindrance or other physical interactions [10]. The chemical 3(2-pyridyldithio)propionic acid N-hydroxysuccinimide ester (SPDP), converts terminal amino groups into sulfhydryl groups and has been used to link transferrin to PLL by the formation of a disulfide bond [6]. This bond, however, is very reactive with other sulfhydryl groups, which can result in instability of the disulfide bond and difficulty in conjugate formation. A third attachment is based on the interaction between biotin and streptavidin with its four biotin-binding sites. The binding between these molecules results in one of the highest dissociation constants in nature and has been used for coupling epidermal growth factor or adenovirus to PLL [10]. Biotinylation of proteins is a very mild method to modify components for coupling; however, the inclusion of streptavidin into the vector formulation for biotin/streptavidin bridge formation can increase vector size as well as vector immunogenicity. In contrast, it is possible to directly link components to DNA through the use of agents such as an ethidium homodimer [34]. More recently, peptide-nucleic acids or PNAs have been developed that bind a specific plasmid sequence (based on the nucleic acid portion), and at the same time introduce a peptide sequence that allows for the attachment of biotin, fluorescent labels, and other components to the plasmid DNA [35]. This provides the ability to easily attach vector components directly to the peptide component and as a result, the plasmid. This system was recently used to easily attach transferrin to plasmid DNA for targeting myogenic cells [36]. In each instance of direct coupling, however, a DNA compacting component is still required, which limits the utility of this approach. Overall, as long as the function of the components that are being added to the vector is maintained, then any coupling method can be used.

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

703

Cell targeting ligands The concept of using a molecular conjugate for gene delivery was originally based on the idea of using DNA-binding components to attach a ligand to DNA that would then target delivery to a specific population of cells for internalization and therapeutic gene expression. The ligand is usually chosen based on targeting a receptor that is expressed on a specific population of cells. One of the first examples of targeting with a molecular conjugate involved the use of asialoorosomucoid (ASOR) [3]. This glycoprotein is produced by treating orosomucoid with neuraminidase, which removes terminal scialic acids and exposes galactose groups. The ASOR is then recognized by the ASOR receptor, which is expressed almost exclusively by liver parenchymal cells (Table 2) [37]. In these studies, a plasmid expressing the chloramphenicol acetyl-transferase (CAT) reporter gene was delivered by an ASOR/PLL conjugate into HepG2 cells (a liver cell line that expresses 250,000 ASOR receptors/cell). The cells showed transient, lowlevel CAT expression that could be competed with free ASOR, indicating cell-specific targeting. Further analysis showed that tail vein administration of the ASOR/DNA polyplex resulted in liver specific expression of the CAT gene [38]. Since these initial experiments, many other ligands have been used as molecular conjugates for targeting (see Table 2). The malaria circumsporozite (CS) protein has been used in vitro to develop a liver-specific conjugate

Table 2 Ligands used as molecular conjugates for targeted DNA delivery and related disease applications Ligand

Target cell organ

Disease

Adenovirus fiber

Lung, liver, tumor cells, many others Hepatocyte/liver

Lung cancer phenylketonuria, hemophilia, many others Analbuminemia, phenylketonuria, hemophilia, other metabolic disorders Lung cancer, brain cancer, pancreatic cancer, other types of cancer Brain cancer, other types of cancer Ovarian cancer, other types of cancer Cervical cancer

Asialoorosomucoid

Epidermal growth factor Fibroblast growth factor Folate Human papilloma virus capsid Malaria CS protein Transferrin

Lung, breast, brain, prostate tumor cells, many other types of tumor cells Brain, prostate tumor cells, other types of tumor cells Many types of cancer cells Epithelial cells/cervix Hepatocyte/liver Liver, lung, tumor cells, many others

Hepatocellular carcinoma, diabetes, cirrhosis Many cancers, melanoma, many other applications

704

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

for vector delivery during conditions in which ASOR receptor expression on hepatocytes is low, such as in cirrhosis, diabetes, and hepatocellular carcinoma [11]. Transferrin, which has receptors that are expressed by many different cell types, has been used as a molecular conjugate to deliver DNA to erythroleukemic, lung, and liver cell lines [6,7]. Delivery has also been obtained in vivo to tumors as well as to the liver [39,40]. The use of the transferrin receptor for vector targeting also provides an example in which the status of a receptor can be modulated to increase delivery. Agents such as desferrioxamine, an iron chelator, can produce a four-fold increase in the number of receptors on responsive cells, resulting in increased transduction by a transferrin/DNA polyplex [7]. Molecular conjugates have also been used in receptor identification. The human papilloma virus (HPV) capsid was used to partially identify the HPV receptor on cervical cancer cells by using the capsid as a ligand for the attachment and delivery of a reporter gene to those cells [12]. In contrast to these high molecular weight ligands, vitamins and peptides have also been used. The vitamin folate has been used to promote delivery of DNA into cells that overexpress the folate receptor, such as ovarian carcinoma cells [9]. More recently, other ligands have been used to promote selective or specific uptake by certain cell types that vary in receptor expression, such as epidermal growth factor (EGF) or fibroblast growth factor receptor overexpression on tumor cells [10,41]. The overexpression of the EGF receptor on cancer cell lines has allowed for specific uptake of EGF/ DNA polyplexes by both non-small cell and small cell lung cancer cells [10]. The continued identification of ligands for targeted gene delivery has also resulted in the replacement of proteins with synthetic ligands. The protein ASOR, which binds to the ASOR receptor through terminal galactose groups, has now been replaced by a galactose conjugate for liver specific DNA delivery [13]. This conversion is an important finding because molecular conjugates can potentially be synthetically derived. Technological advances in other areas of cell biology will also provide a new source for synthetic targeting ligands. Phage display technology, which uses phage libraries that express a wide range of peptide sequences, is now being used to identify peptides that target surface receptors on different cells, such as tumor and endothelial cells [42]. Advances with this technology could provide a wide range of peptides to target many different tissues, however; it is extremely important that these ligands allow for internalization of vector as well as cellular binding. Further research into ligand identification in combination with the simplicity of producing conjugates should allow for the identification and use of many other ligands as conjugates for targeting polyplex vectors. Endosomal lysis agents: from viruses to polycations Once a suitable DNA-binding component has been identified and a molecular conjugate synthesized, the resulting vector must be tested for delivery efficiency. The incubation of just the PEI/DNA polyplex with cells

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

705

results in nonspecific binding to cells through interactions with cell surface glycoproteins, which is followed by nonspecific internalization [5]. In the context of the targeted polyplex, the binding between ligand and receptor results in ligand/receptor internalization or receptor-mediated endocytosis (see Fig. 1). One of the more critical barriers to overcome in either of these pathways follows endosome or vesicle fusion with the lysosome, where the contents of the endosome is degraded by lysosomal enzymes (see Fig. 1). As a result, DNA that is attached to either a polycation or a molecular conjugate will become degraded, resulting in little or no gene expression [15,16]. In the case of the folate receptor, the bound ligand is internalized through a process termed potocytosis, where the receptor binds the ligand, the surrounding membrane closes off from the cell surface, and the internalized material then passes through the vesicular membrane into the cytoplasm. As a result, the folate conjugate and attached DNA are not degraded, but essentially remain trapped inside larger vesicles in the cell as the Folate/ DNA polyplex is unable to pass through the membrane [9]. The initial vector formulations based on molecular conjugates did not contain components to enhance release of vector from either endosomes or vesicles. The success of these vector formulations was based on the leakiness of the uptake process because some molecules of DNA escaped degradation during the process of internalization. It quickly became apparent, however, based on the low level of gene expression, that some type of agent must be included in the vector formulation to enhance the efficiency of this process. As a result, several agents have been identified that can enhance the endosomal release of DNA into the cytoplasm of the cell, either through endosome disruption or by reducing DNA degradation. Chloroquine, which raises the pH of the endosome, has been used to decrease the degradation of endocytosed material by inhibiting lysosomal hydrolytic enzymes [7]. Physical procedures such as a partial hepatectomy, when performed along with DNA delivery by an ASOR/DNA polyplex, have resulted in increased persistence and expression of delivered DNA, from days to months [14]. These procedures, however, have been limited in use because of inefficiency and the lack of utility for enhancing endosomal release in other tissues. In contrast, human replication-defective adenovirus (ie, dl312, serotype 5 adenovirus with a deleted E1a gene) has been identified as one of the best agents thus far at performing endosome lysis [15]. The adenoviral particle, which is internalized by a receptor-mediated process, escapes degradation by fusion of the penton base in the viral capsid with the endosomal membrane, which leads to pore formation and endosome lysis [43]. When replication-defective adenovirus is coincubated with a Protein/DNA polyplex, both are internalized into the same endosome. The adenovirus then mediates endosomal lysis and the vector is released into the cytoplasm. The use of adenovirus in this role has resulted in as much as a 1000-fold increase in gene expression in various targeting approaches [15,16]. Unfortunately, viral titers of 1103 to 1104 viral particles per cell must be used, which can result in toxicity.

706

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

The specificity of delivery is still maintained by the ligand in the Protein/ DNA polyplex, as determined by competition with free protein [16]. The enhancement of endosomal release by adenovirus has been shown to work with many of the ligands listed in Table 2. This enhancement can be increased further by directly coupling the virus to the Protein/DNA polyplex [17]. This results in at least a one-order-of-magnitude drop in viral titer and increased specificity of endosomal lysis, because an endocytotic event will generally be accompanied by an adenovirus [17–19]. This type of polyplex has shown limited use in vivo, however, due to problems with nonspecific uptake through the adenovirus receptor, an increase in the size of the polyplex, and increased toxicity [17–20]. Attempts have been made to decrease toxicity from adenovirus by using ultraviolet light to inactivate the viral genome [44]. In contrast, other viruses such as chicken adenovirus (CELO virus) have been used to mediate similar levels of enhancement, but without the associated toxicity [45]. Unfortunately, the inclusion of a viral component in the vector formulation is still limiting based on toxicity and immunogenicity, and does not result in a true nonviral vector. To address this, many groups have now focused on developing completely nonviral endosome lysis components. One approach has been to incorporate peptides based on the membrane lysis portion of the influenza virus hemagglutinin HA2 into a transferrin/DNA polyplex to promote endosomal lysis [46]. As a result of endosome acidification before fusion with the lysosome, these short, 20-amino acid peptides mediate endosomal lysis by inserting into the endosomal membrane, causing pores to form, which leads to lysis [47]. A comparison of these peptides to adenovirus has shown that these molecules are less efficient at endosomal lysis than adenovirus [46]. More recently, branched chain polycations such as PEI have been used in this role. This synthetic polymer has a highly branched ratio of primary:secondary:tertiary amines (1:2:1) [21]. This allows for DNA binding, endosome release, and ligand attachment all in one molecule. Endosome release is based on PEI’s structure (terminal amines ionizable at pH 6.9, internal amines ionizable at pH 3.9), which buffers a pH change during endosome acidification, leading to vesicle swelling and lysis [21]. As a result, this multifunctional component is capable of mediating efficient gene delivery even in the absence of a targeting component (see Fig. 3). As with fusogenic peptides, however, the delivery efficiency of vectors containing this agent is still limited, especially in vivo. Some groups have hypothesized that this is due to the lack of a nuclear translocation component in the vector. Strengthening this thinking is the fact that adenovirus contains proteins that protect and translocate the linear viral DNA to the nucleus. The lack of these components in molecular conjugate based vectors may result in lower gene delivery to the nucleus. As a result, it is now becoming clear that components with nuclear translocation abilities and even agents capable of increasing the persistence and level of gene expression must be included in conjugate-based vectors to increase overall transduction efficiencies.

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

707

Nuclear translocation, episomal replication, and integration Although there has been much focus on developing potent nonviral delivery vectors, much less time has been spent on understanding issues related to the cellular fate of these vectors, such as nuclear translocation or factors related to the length and level of gene expression. For example, although recent studies have taken advantage of the high endosomolytic activity mediated by PEI, little is known about the intracellular fate of the molecule and the attached DNA. Some studies have suggested that PEI travels to the nucleus by interacting with the cytoskeleton, but this is still unclear [48]. Other groups have made attempts to enhance nuclear localization by using peptides based on viral protein sequences. Small peptides based on the nuclear translocation sequences of SV40 T-antigen have demonstrated increased nuclear delivery of DNA [49]. A more critical aspect of nuclear delivery, however, may be at the entry site to the nucleus, or the nuclear pore [50]. This structure is typically 20 nm in size and raises the question as to how a vector as big as 500 nm in size can translocate through this opening. Studies are now being directed towards this question and should reveal the mechanism involved and how nuclear translocation peptides affect this process. Once the DNA obtains entry into the nucleus, the delivered plasmid remains episomal and does not integrate into the host genome. Although initial studies indicated that a certain portion of the delivered DNA may integrate into the host genome, the lack of sequences in current plasmids does not support this outcome. To increase the duration of gene expression, studies using partial hepatectomy following Protein/DNA polyplex delivery resulted in an increase in the length of gene expression from days to months [14]. Unfortunately, this technique is not a viable approach for use in other tissues. Further analysis of the effect of this procedure on the delivered DNA has established that a greater amount of DNA is protected in vesicles in liver cells during the repair process, and that the protected DNA is slowly released with time. This was determined through the use of peroxisome stimulating agents that reproduced this effect in liver cells [51]. Overall, although expression does occur, the DNA does not persist and the resulting duration and level of expression is transient and currently requires repeated administration of vector. The limited immunogenicity of the vector may facilitate this type of administration, but it would be advantageous to develop plasmids that either promote plasmid replication or integration to allow for persistent, high-level expression of the therapeutic gene. Some studies have already demonstrated that plasmids modified to contain the EBV origin of replication and the EBNA-1 gene, persist for a longer period of time in replicating cells [52]. The expressed EBNA-1 protein binds to the EBV origin of replication, which leads to localization of the plasmid to the nuclear membrane and protection of the nucleic acid from degradation [52]. In contrast, some approaches are now focusing on the integration of the delivered DNA. Recent studies using retrotransposons rectors demonstrated insertion of DNA into the mouse

708

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

genome in 5% to 6% of transfected mouse liver cells [53]. These enhancements will be critical to the further use of nonviral vectors in clinical applications, particularly in cancer treatments where these vectors need to achieve persistent (for the life of the tumor cell), high-level gene expression. As a result, new plasmids must be developed to address this problem. Because Protein/DNA polyplexes can be generated very easily, the testing will be greatly simplified. As these questions are addressed, improved Protein/DNA polyplexes will be created, allowing for more efficient delivery and expression.

Nucleic acids: from oligonucleotides to plasmids Most research that has been done with Protein/DNA polyplexes has used plasmid DNA, based on the ability to easily manipulate and cost-effectively produce this molecule in large quantities. The initial work using plasmids focused on developing conditions that allowed for efficient formation of the Protein/DNA polyplex so that effective administration could be achieved. Original work by Wu et al demonstrated that self-assembly between the conjugate and DNA to form the Protein/DNA polyplex could be carried out under concentrated conditions ([DNA] ¼ 87 nm or 0.35 lg/ll). This was accomplished using a step dialysis protocol that allowed for a change in salt concentration from 2 M to physiological concentrations (0.15 M) [3]. Although polyplexes formed by this procedure are produced at concentrations suitable for in vivo administration, reproducibility and precipitation of the Protein/DNA polyplex has made this procedure difficult to use. In contrast, Wagner et al have shown that polyplexes can form efficiently in dilute DNA concentrations ([DNA] ¼ 3 nm or 0.012 lg/ll) in 0.15 M salt without precipitation [6]. Unfortunately, this formulation has shown limited in vivo application due to the low concentration of DNA. Although research continues towards developing improved conditions for efficient polyplex formation for in vivo delivery, these formulations have been suitable for in vitro studies. The simple self-assembling process by which Protein/DNA polyplexes can be generated has demonstrated that there are no current limits on the type and size of the nucleic acid that can be delivered. The primary type of nucleic acid that has been delivered thus far is DNA in plasmid form, ranging in size from several kilobases (kb) to over 48 kb in length [44]. This large plasmid was delivered by a transferrin/DNA polyplex without a loss of delivery efficiency, when compared with polyplexes made with a smaller plasmid carrying the same reporter gene. In contrast, small oligonucleotides have also been used [54]. In this instance, antisense oligonucleotides were used in combination with a liver-specific delivery system to allow for better targeting of the oligonucleotide. Because this vector is easily manipulated to use a plasmid of any size, anything from a reporter gene to a therapeutic gene can be delivered for

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

709

expression in a particular cell type without the labor and time required to generate new viral vectors. The E. coli b-galactosidase (b-gal) and luciferase genes have been used the most as reporter genes. Use of the b-gal gene allows for histochemical staining of cells to determine the number or percentage of cells transduced as well as quantitation of expression. In contrast, the luciferase gene allows for viewing of transduced cells by fluorescence analysis. Although the expression can also be quantitated, this protein also allows for continued viewing of the cells without fixation, which simplifies time-course studies. In either situation, depending upon the ligand used, the resulting Protein/DNA polyplex produces variable levels of transduction and expression in vitro. Transduction efficiencies as high as 80% to 99% have been achieved using an ASOR/DNA polyplex or EGF/DNA polyplex to deliver the b-gal gene to primary hepatocytes or tumor cells respectively [10,16]. A wide range of therapeutic genes have also been used in various vector formulations. The ASOR/DNA polyplex has been used to deliver the phenylalanine hydroxylase (PAH) gene, which encodes the PAH enzyme (deficient in phenylketonuria). These studies demonstrated that the complete replacement of PAH activity could be achieved with this vector in primary mouse hepatocytes deficient in PAH expression [16]. Other experiments have demonstrated that high-level factor IX expression can be obtained after factor IX gene delivery to primary hepatocytes, suggesting a basis for potentially correcting the disease phenotype associated with hemophilia [17]. A transferrin/DNA polyplex has been incorporated into a clinical protocol for the ex vivo transduction of melanoma cells with cytokine genes for the immunological rejection of melanoma cells [55]. More recently, an adenovirus/PLL polyplex has demonstrated efficient delivery of the tumor suppressor p53 gene in p53 deficient tumor cells [56]. The in vivo applications of Protein/DNA polyplexes have been limited thus far, with delivery occurring to the lung, liver, kidney, and more recently to solid tumors. Initial analysis of ASOR/DNA polyplex delivery to the liver showed that efficient gene delivery following intravenous administration could only be achieved when accompanied by a partial hepatectomy [57]. As an example, when the delivery of an ASOR/DNA polyplex was accompanied by partial hepatectomy, the levels of albumin after introduction of an albumin-expressing plasmid reached 34 lg/ml in the blood of analbuminemic rats [57]. More recently, Hanson et al have now shown that DNA delivery by a galactose conjugate to the liver can occur without partial hepatectomy, by generating Protein/DNA polyplexes that are very small (10– 12 nm in size) [58]. Attempts have also been made to use a transferrin/DNA polyplex coupled to adenovirus in which delivery to the lung epithelium by intratracheal administration resulted in less than 1% of the cells transduced [20]. More recently, an adenovirus/PLL conjugate was used to deliver the tumor suppressor p53 gene by intratumoral administration to non-small cell lung cancer tumors in mice [56]. This resulted in high-level p53 expression that induced apoptosis and produced at least a 50% reduction in tumor

710

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

size. Unfortunately though, each of these polyplex vectors required adenovirus as an endosome lysis agent. In contrast, much more recent formulations based on using PEI have now shown that gene delivery to the lung, brain, kidney and to tumors can be obtained with this polycation [21,24–26,59]. Perhaps the most impressive result was obtained when a linear form of PEI was used to deliver the bgal gene to the lung following intravenous administration [25]. This resulted in the vast majority of the lung turning blue after histochemical staining. Unfortunately, these formulations are not without limitations. The intratumoral administration of a PEI/DNA polyplex into a subcutaneous tumor produced little or no gene delivery [60]. Although delivery could be increased by using slow infusion of vector, procedures such as this increase the complexity of vector administration and future clinical uses. It has also been established that polycation based vectors typically elicit interaction with blood proteins such as complement [57]. Further analysis of this effect has shown that the charge of the vector contributes to this result. More recent studies have proven that agents such as polyethylene glycol must be included in the formulation to shield the vector from interaction with these blood proteins [61]. Overall, although these obstacles have been encountered, progress continues towards overcoming these limitations. It is clear that the limited in-vivo delivery efficiency of polycation and conjugate based vectors is now becoming less of a concern. Molecular conjugates versus viral and nonviral vectors A true comparison between all viral and nonviral vectors that are capable of gene delivery is beyond the scope of this review. In general, it is clearly understood that viruses have developed the most efficient methods of gaining entry into cells and methods to enhance expression of the delivered nucleic acid. The viral vectors that have been used in most gene therapy approaches have been based on recombinant versions of retrovirus (based on the moloney murine leukemia virus), adenovirus (based on serotype 5), adeno-associated virus, herpes simplex virus, and more recently, lentivirus. These viruses have developed unique characteristics that can be used in gene therapy applications. Recombinant retroviral vectors have the ability to integrate into the host cell genome, resulting in long-term gene expression; however, transduction occurs predominantly in actively dividing cells [62–64]. Recombinant adenoviral vectors have the ability to achieve much higher levels of transduction than retrovirus, especially in nondividing cells; however, gene expression is transient due to lack of integration and an immune response against infected cells because of low level adenoviral protein expression [65–67]. Recombinant adeno-associated virus has been engineered so that the complete viral genome has been removed, except for the terminal ITR that are involved in integration. This reduces immunogenicity but disrupts the integration pattern of the viral genome that usually occurs

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

711

at the same location in chromosome 19 [68]. Recombinant versions of herpes simplex virus have shown the best ability to infect neuronal cells because the natural tropism for this virus is the brain, but the complexity of this virus has made it difficult to use [69]. Lentiviral vectors have the ability to promote integration for long-term gene expression and can infect both dividing and nondividing cells, but the efficiency of this process and the type of cells this occurs in are still in question [70]. In addition, viral vectors as a group suffer from several major limitations that have hindered their use and efficiency: (1) all require packaging cell lines to allow for production of replication defective virus, which for retrovirus and adeno-associated virus has been problematic in producing high titer virus; (2) all have limits that affect the size and type of nucleic acid that can be packaged or delivered; (3) all still require many tests to ensure safety and lack of toxicity (both cellular and genotoxicity); (4) most lack the ability in their natural form to target the nucleic acid to a specific cell type; and (5) all lack the ability to become completely synthetic. In contrast, molecular conjugates and Protein/DNA polyplexes continue to become a much more versatile delivery vector for gene therapy applications, based on properties that cannot be matched by most other vectors and in particular viral vectors: (1) no need for packaging cell lines, allowing for the quick analysis of many plasmids or different types of nucleic acids; (2) the ability to target nucleic acids to specific types of cells, resulting in high level transduction of either dividing or nondividing cells; (3) the lack of a viral component, resulting in safety for the recipient; (4) no limitation on the size or type of the nucleic acid that can be delivered; and (5) the potential to become completely synthetic, allowing for simple and cost effective development of the delivery vector. In addition to Protein/DNA polyplexes, formulations involving liposomes [71] and the mechanical delivery of naked DNA [72,73] have been developed to deliver nucleic acids without the aide of viruses and the related limitations. Although these vectors have a range in ease of production, there are also recurrent problems with low in vivo gene delivery; naked DNA is simple to make but exhibits the lowest transduction, whereas liposomes are difficult to synthesize, expensive, and exhibit the highest transduction. In addition to these limitations, liposomes, which use lipid/nucleic acid complexes or lipoplexes (lipo ¼ lipid) to deliver the nucleic acid into the cytoplasm of the recipient cell, normally lack the ability to target specific cells. Different forms of lipids, such as glycolipids, can be used to target specific organs such as the liver, and more recently, ligands have been added directly to the lipids [74]. Unfortunately, the presence of the lipid component in the lipoplex can result in nonspecific uptake by the reticuloendothelial system, causing a loss of targeting specificity. The mechanical delivery of naked DNA can be done by either direct injection into the target organ or attachment of the DNA to gold particles, which are then delivered to tissues by high-velocity bombardment [72,73]. The injection of naked DNA into muscle has led to efficient DNA delivery and expression in vivo

712

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

[72,75]. This method of delivery, however, seems to work primarily with the muscle and results in only cells near the injection site acquiring the DNA. The delivery of DNA by particle bombardment has also shown expression in organs such as the liver, but suffers from the lack of targeting, the inability to transduce a large number of cells, and the need for a surgical procedure to allow access to the tissue [73]. In contrast, the self-assembling properties of conjugate based vectors (simple to use) and commercial availability (inexpensive/easy synthesis) of components gives greater potential for this vector in developing efficient, nonviral vectors. Although this comparison is not inclusive, it is also not meant to state that Protein/DNA polyplexes are not without problems and limitations. Some of the more important limitations are transient levels of gene expression, variability in transduction, limited safety testing in vivo, and thus far, limited use in vivo. It is clear, however, that the simplicity of the Protein/ DNA polyplex and its easy manipulation should allow these factors to be addressed much more easily than the problems associated with other vectors.

Future directions The development of molecular conjugates as components of Protein/ DNA polyplexes has resulted in the creation of a simple, nonviral vector for the targeted delivery of nucleic acids into specific cell types. This vector has many of the positive attributes of viruses, but without the limitations that continue to plague recombinant viruses. The simplicity of this vector allows for quick analysis of nucleic acids, expression vectors, and therapeutic genes in vitro and potentially in vivo, because the time that would be involved in the generation of recombinant viral vectors is not present. Essentially, the development of this delivery vector has resulted in the creation of a ‘‘synthetic virus’’ that has the capability of targeted delivery without the negative attributes of viruses. Future work will continue to use the advantages of this vector to address problems of transient expression by developing integration and episomal maintenance plasmids based on viral systems, will use peptides based on viral nuclear translocation signals for enhanced nuclear delivery and gene expression, and will also use other properties of viruses. The identification of new components will be crucial to the further development and use of this delivery vector, but at the same time will limit complexity. As this system matures it may also be possible to combine many of these components into one chimeric protein or peptide. The further manipulation of this system should also lead to tissue-specific and regulatable-expression systems, which should add another level of specificity to the vector. As a result, there will be no limit on the type of therapeutic gene that can be used, nor on the ligand that can be used for targeting. It is clear that any protein or peptide that can increase the utility of this vector can be easily incorporated, resulting in a much

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

713

greater use of molecular conjugates and Protein/DNA complexes not only in vitro, but more importantly for the in vivo applications of gene therapy.

References [1] Venter JC, Adams MD, Myers EW, et al. The sequence of the human genome. Sci 2001; 291(5507):1304–51. [2] Marshall E. Clinical trials—gene therapy death prompts review of adenovirus vector. Sci 1999;286(5448):2244–45. [3] Wu G, Wu C. Receptor-mediated in vitro gene transformation by a soluble DNA carrier system. J Biol Chem 1987;262:4429–32. [4] Xu B, Wiehle S, Roth JA, Cristiano RJ. The contribution of poly-L-lysine, streptavidin, and epidermal growth factor to EGF/PLL/DNA polyplex formation. Gene Ther 1998;5:1235–43. [5] Mislick KA, Baldeschwieler JD. Evidence for the role of proteoglycans in cation-mediated gene transfer. Proc Natl Acad Sci USA 1996;93:12349–54. [6] Wagner E, Zenke M, Cotten M, et al. Transferrin-polycation conjugates as carriers for DNA uptake into cells. Proc Natl Acad Sci USA 1990;87:3410–14. [7] Cotten M, Langle-Rouault F, Kirlappos H, et al. Transferrin-polycation-mediated introduction of DNA into human leukemic cells: stimulation by agents that affect the survival of transfected DNA or modulate transferrin receptor levels. Proc Natl Acad Sci USA 1990;87:4033–37. [8] Curiel DT, Wagner E, Cotten M, et al. High-efficiency gene transfer mediated by adenovirus coupled to DNA-polylysine complexes. Hum Gene Ther 1992;3:147–54. [9] Gottschalk S, Cristiano R, Smith L, et al. Folate-mediated gene delivery and expression in vitro. Gene Ther 1994;1:185–91. [10] Cristiano RJ, Roth JA. Epidermal growth factor mediated DNA delivery into lung cancer cells via the epidermal growth factor receptor. Cancer Gene Ther 1996;3:4–10. [11] Ding Z, Cristiano R, Roth J, et al. Malaria circumsporozite protein: a novel gene delivery vehicle to primary hepatocyte cultures and cultured cells. J Biol Chem 1995;270:3667–76. [12] Muller M, Gissmann L, Cristiano RJ, et al. Papillomavirus capsid binding and uptake by cells from different tissues and species. J Virology 1995;69:948–54. [13] Plank C, Zatloukal K, Cotten M, et al. Gene transfer into hepatocytes using asialoglycoprotein receptor mediated endocytosis of DNA complexed with an artificial tetraantennary galactose ligand. Bioconjugate Chem 1992;3:533–39. [14] Wu CH, Wilson JM, Wu GY. Targeting genes: delivery and persistent expression of a foreign gene driven by mammalian regulatory elements in vivo. J Biol Chem 1989;264:16985–87. [15] Curiel D, Agarwal S, Wagner E, et al. Adenovirus enhancement of transferrin-polylysinemediated gene delivery. Proc Natl Acad Sci USA 1991;88:8850–54. [16] Cristiano R, Smith L, Woo S. Hepatic gene therapy: adenovirus enhancement of receptormediated gene delivery and expression in primary hepatocytes. Proc Natl Acad Sci USA 1993;90:2122–26. [17] Cristiano R, Smith L, Woo S. Hepatic gene therapy: efficient gene delivery and expression in primary hepatocytes utilizing a conjugated adenovirus-DNA complex. Proc Natl Acad Sci USA 1993;90:11548–52. [18] Curiel D, Wagner E, Cotten M, et al. High-efficiency gene transfer mediated by adenovirus coupled to DNA-polylysine complexes. Hum Gene Ther 1992;3:147–54. [19] Wagner E, Zatloukal K, Cotten M, et al. Coupling of adenovirus to transferrin-polylysine/ DNA complexes greatly enhances receptor-mediated gene delivery and expression of transfected genes. Proc Natl Acad Sci USA 1992;89:6099–103. [20] Gao L, Wagner E, Cotten M, et al. Direct in vivo gene transfer to airway epithelium employing adenovirus-polylysine-DNA complexes. Hum Gene Ther 1993;4: 17–23.

714

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

[21] Bousiff O, Lezoualch F, Zanta MA, et al. Polyethylenimine: a versatile vector for gene and oligonucleotide transfer into cells in culture and in vivo. Proc Natl Acad Sci USA 1995;92:7297–301. [22] Godbey WT, Wu KK, Mikos AG. Poly(ethylenimine) and its role in gene delivery. J Control Release 1999;60(2–3):149–60. [23] Merlin JL, Dolivet G, Dubessy C, et al. Improvement of nonviral p53 gene transfer in human carcinoma cells using glucosylated polyethylenimine derivatives. Cancer Gene Ther 2001;8(3):203–10. [24] Wightman L, Kircheis R, Rossler V, et al. Different behavior of branched and linear polyethylenimine for gene delivery in vitro and in vivo. J Gene Med 2001;3(4):362–72. [25] Goula D, Benoist C, Mantero S, et al. Polyethylenimine-based intravenous delivery of transgenes to mouse lung. Gene Ther 1998;5(9):1291–5. [26] Coll JL, Chollet P, Brambilla E, et al. In vivo delivery to tumors of DNA complexed with linear polyethylenimine. Hum Gene Ther 1999;10(10):1659–66. [27] Wagner E, Cotten M, Foisner R, et al. Transferrin-polycation-DNA complexes: the effect of polycations on the structure of the complex and DNA delivery to cells. Proc Natl Acad Sci USA 1991;88:4255–59. [28] Chen J, Stickles R, Daichendt K. Galactosylated histone-mediated gene transfer and expression. Hum Gene Ther 1994;5:429–36. [29] Kneuer C, Sameti M, Bakowsky U, et al. A nonviral DNA delivery system based on surface modified silica-nanoparticles can efficiently transfect cells in vitro. Bioconjug Chem 2000;11(6):926–32. [30] Robaczewska M, Guerret S, Remy JS, et al. Inhibition of hepadnaviral replication by polyethylenimine-based intravenous delivery of antisense phosphodiester oligodeoxynucleotides to the liver. Gene Ther 2001;8(11):874–81. [31] Baker A, Cotten M. Delivery of bacterial artificial chromosomes into mammalian cells with psoralen-inactivated adenovirus carrier. Nucleic Acids Res 1997;25(10):1950–56. [32] Dunlap DD, Maggi AS, Marco R, Monaco L. 1997. Nanoscopic structure of DNA condensed for gene delivery. Nucl Acids Res 1997;25:3095–3101. [33] Parton RG, Schrotz P, Bucci C, et al. Plasticity of early endosomes. J Cell Sci 1992; 103(Pt 2):335–48. [34] Wagner E, Cotten M, Mechtler K, et al. DNA-binding transferrin conjugates as functional gene delivery agents: synthesis by linkage of polylysine or ethidium homodimer to the transferrin carbohydrate moiety. Bioconjugate Chem 1991;2:226–31. [35] Nielsen PE, Egholm M. An introduction to peptide nucleic acid. Curr Issues Mol Biol 1999;1(1–2):89–104. [36] Liang KW, Hoffman EP, Huang L. Targeted delivery of plasmid DNA to myogenic cells via transferrin-conjugated peptide nucleic acid. Mol Ther 2000;1(3):236–43. [37] Schwartz AL, Rup D, Lodish HF. Difficulties in the quantification of the asialoglycoprotein receptors on the rat hepatocyte. J Biol Chem 1980;255:9033–36. [38] Wu GY, Wu CH. Receptor-mediated Gene Delivery and expression in vivo. J Biol Chem 1988;263:14621–24. [39] Kircheis R, Wightman L, Schreiber A, et al. Polyethylenimine/DNA complexes shielded by transferrin target gene expression to tumors after systemic application. Gene Ther 2001; 8(1):28–40. [40] Kircheis R, Schuller S, Brunner S, Ogris M, Heider KH, Zauner W, Wagner E. Polycationbased DNA complexes for tumor-targeted gene delivery in vivo. J Gene Med 1999;1(2): 111–20. [41] Fisher KD, Ulbrich K, Subr V, et al. A versatile system for receptor-mediated gene delivery permits increased entry of DNA into target cells, enhanced delivery to the nucleus and elevated rates of transgene expression. Gene Ther 2000;7(15):1337–43. [42] Pasqualini R, Koivunen E, Kain R, et al. Aminopeptidase N is a receptor for tumorhoming peptides and a target for inhibiting angiogenesis. Cancer Res 2000;60:722–27.

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

715

[43] Greber U, Willetts M, Webster P, et al. Stepwise dismantling of adenovirus 2 during entry into cells. Cell 1993;75:477–86. [44] Cotten M, Wagner E, Zatloukal K, et al. High-efficiency receptor-mediated delivery of small and large (48 kilobase) gene constructs using the endosome-disruption activity of defective or chemically inactivated adenovirus particles. Proc Natl Acad Sci USA 1992;89:6094–98. [45] Cotten M, Wagner E, Zatloukal K, et al. Chicken adenovirus (CELO Virus) particles augment receptor-mediated DNA delivery to mammalian cells and yield exceptional levels of stable transformants. J Virol 1993;67:3777–85. [46] Wagner E, Plank C, Zatloukal K, et al. Influenza virus hemagglutinin HA-2 N-terminal fusogenic peptides augment gene transfer by transferrin-poly-lysine-DNA complexes: toward a synthetic virus-like gene-transfer vehicle. Proc Natl Acad Sci USA 1992;89:7934–38. [47] Rafalski M, Ortiz A, Rockwell A, et al. Membrane fusion activity of the influenza virus hemagglutinin: interaction of HA2 N-terminal peptides with phospholipid vesicles. Biochemistry 1991;30:10211–20. [48] Godbey WT, Wu KK, Mikos AG. Tracking the intracellular path of poly(ethylenimine)/ DNA complexes for gene delivery. Proc Nat Acad Sci USA 1999;96(9):5177–81. [49] Branden LJ, Mohamed AJ, Smith CI. A peptide nucleic acid-nuclear localization signal fusion that mediates nuclear transport of DNA. Nat Biotechnol 1999;17(8):784–7. [50] Talcott B, Moore MS. Getting across the nuclear pore complex. Trends Cell Biol 1999;9(8):312–8. [51] Chowdhury NR, Wu CH, Wu GY, et al. Fate of DNA targeted to the liver by asialoglycoprotein receptor-mediated endocytosis in vivo. Prolonged persistence in cytoplasmic vesicles after partial hepatectomy. J Biol Chem 1993;268(15):11265–71. [52] Oh SJ, Chittenden T, Levine AJ. Identification of cellular factors that bind specifically to the Epstein-Barr virus origin of DNA replication. J Virol 1991;65(1):514–19. [53] Yant SR, Meuse L, Chiu W, et al. Somatic integration and long-term transgene expression in normal and haemophilic mice using a DNA transposon system. Nat Genet 2000;25(1):35–41. [54] Lu X, Fischman A, Jyawook S, et al. Antisense DNA-delivery in vivo: liver targeting by receptor-mediated uptake. J Nucl Med 1994;35:269–75. [55] Zatloukal K, Schneeberger A, Berger M, et al. Elicitation of a systemic and protective antimelanoma immune response by an IL-2-based vaccine J Immunol 1995;154:3406–19. [56] Nguyen DM, Wiehle SA, Roth JA, Cristiano RJ. Delivery of the p53 tumor suppressor gene into lung cancer cells by an adenovirus/DNA complex. Cancer Gene Ther 1997;4:191–98. [57] Wu G, Wilson J, Shalaby F, et al. Receptor-mediated gene delivery in vivo. J Biol Chem 1991;266:14338–42. [58] Perales JC, Ferkol T, Beegen H, et al. Gene transfer in vivo: sustained expression and regulation of genes introduced into the liver by receptor-targeted uptake. Proc Natl Acad Sci USA 1994;91:4086–90. [59] Boletta A, Benigni A, Lutz J, et al. Nonviral gene delivery to the rat kidney with polyethylenimine. Hum Gene Ther 1997;8(10):1243–51. [60] Plank C, Mechtler K, Szoka Jr FC, et al. Activation of the complement system by synthetic DNA complexes: a potential barrier for intravenous gene delivery. Hum Gene Ther 1996; 7(12):1437–46. [61] Erbacher P, Bettinger T, Belguise-Valladier P, et al. Transfection and physical properties of various saccharide, poly(ethylene glycol), and antibody-derivatized polyethylenimines (PEI). J Gene Med 1999;1(3):210–22. [62] Bender M, Gelinas R, Miller A. A majority of mice show long-term expression of a human b-globin gene after retrovirus transfer into hematopoietic stem cells. Mol Cell Biol 1989;9:1426–34. [63] Kay M, Baley P, Rothenberg S, et al. Expression of human a1-antitrypsin in dogs after autologous transplantation of retroviral transduced hepatocytes. Proc Natl Acad Sci USA 1992;89:89–93.

716

R.J. Cristiano / Surg Oncol Clin N Am 11 (2002) 697–716

[64] Miller D, Adam M, Miller A. Gene transfer by retrovirus vectors only in cells that are actively replicating at the time of infection. Mol Cell Biol 1990;10:4239–42. [65] Li Q, Kay M, Woo S. Assessment of recombinant adenoviral vectors for hepatic gene therapy. Hum Gene Ther 1993;4:403–9. [66] Smith TAG, Mehaffey MG, Kayda DB, et al. Adenovirus mediated expression of therapeutic plasma levels of human factor IX in mice. Nat Genet 1993;5:397–402. [67] Engelhardt J, Litzky L, Wilson J. Prolonged transgene expression in cotton rat lung with recombinant adenovirus defective in E2a. Hum Gene Ther 1994;5:1217–29. [68] Flotte TR, Afione SA, Conrad C, et al. Stable in vivo expression of the cystic fibrosis transmembrane conductance regulator with an adeno-associated virus vector. Proc Natl Acad Sci USA 1993;90:10613–17. [69] Geller AI, Breakefield XO. A defective HSV-1 Vector expresses Eschericia coli bgalactosidase in cultured peripheral neurons. Science 1988;241:1667–68. [70] Park F, Kay MA. Modified HIV-1 based lentiviral vectors have an effect on viral transduction efficiency and gene expression in vitro and in vivo. Mol Ther 2001;4(3): 164–73. [71] Nabel E, Gordon D, Yang Z, et al. Gene transfer in vivo with DNA-liposome complexes: lack of autoimmunity and gonadal localization. Hum Gene Ther 1992;3:649–56. [72] Davis H, Whalen R, Demeneix B. Direct gene transfer into skeletal muscle in vivo: factors affecting efficiency of transfer and stability of expression. Hum Gene Ther 1993;4:151–59. [73] Cheng L, Ziegelhoffer P, Yang N. In vivo promoter activity and transgene expression in mammalian somatic tissues evaluated by using particle bombardment. Proc Natl Acad Sci USA 1993;90:4455–59. [74] Das P, Murray G, Zirzow G, et al. Lectin-specific targeting of b-glucocerebrosidase to different liver cells via glycosylated liposomes. Biochem Med 1985;33:124–31. [75] Acsadi G, Jiao S, Jani A, et al. Direct gene transfer and expression into rat heart in vivo. New Biol 1991;3:71–81.