Introduction

Crucial requirements in gene therapy are a successful delivery of the therapeutic gene into the right target cells and a controlled expression in the transfected cell, at the appropriate levels and over the required period. The current gene delivery systems are not optimal in these aspects. During the last decade increased attention has been given to nonviral gene delivery vehicles, such as naked plasmid DNA, cationic lipo-some-DNA complexes (lipoplexes), polymer/DNA-based systems (polyplexes), and combinations thereof (1-2). These delivery systems are attractive because of their simplicity (they can be generated from few, defined components), and the flexibility in synthesis and assembly of the gene transfer complex. If desired, they can be designed protein-free/nonim-munogenic and can be very flexible regarding the size of DNA to be transported. However, their efficiency in gene delivery does not meet the requirements for most therapeutic applications.

Evolution has resulted in several mechanisms that enable the transport of molecules into cells that include not only simple diffusion, active transport, phagocytosis, receptor-mediated endocytosis, but also other mechanisms such as those used by various viruses to infect cells, bacterial conjugation, and natural transformation procedures seen in some microorganisms such as Helicobacter pylori (3). These natural pathways can be used in a new, artificial setting as means of a transfer route to deliver nucleic acid into cells. This article describes the nonviral delivery system, which exploits receptor-mediated endocytosis to target and transfer nucleic acids into cells.

The concept of receptor-mediated gene transfer draws its attention from the natural delivery mechanism based on receptor-mediated endocytosis (4,5). There are basically 2 mechanisms of endocytosis: (1) clathrin-dependent receptor-mediated endocytosis (coated pit endocytosis) (6), and (2) clathrin-independent endocytosis (7). Clathrin-dependent receptor-mediated endocytosis involves the binding of a ligand to a specific cell surface receptor, resulting in the clustering of the ligand-receptor complexes in clathrin-coatedpits, invagination into the cell and budding off of the coated pits from the cell surface membrane to form intracellular coated vesicles, and maturation (uncoating, fusion of vesicles) into endo-somes. Within these endosomes, ligands and receptors are each sorted to their appropriate (intra)cellular destination (e.g., lysosome, golgi apparatus, nucleus, or cell surface membrane). The clathrin-independent mechanisms resulting in un-coated pits include phagocytosis, pinocytosis, and potocytosis. Phagocytosis is a mechanism of internalizing large particles and microorganisms (>0.5 ^m). This mode of internalization is initiated by receptors on the phagocyte recognizing the particle either directly or indirectly, via opsonization of the particle. Internalization is primarily mediated via pseudopod action rather than pits (invaginations) on the cell surface. The engulfed particle, initially situated in the early phagosome, is eventually destroyed along the endocytosed pathway. Thus, cell capability to recognize particles via receptors and to form pseudopods seems to be a major characteristic mediating phagocytic internalization. Other forms of internalization that do not use clathrin, however, do seem to rely on pits (invaginations) on the cell surface. Such internalization systems include macropinocytosis, pinocytosis, potocytosis, and transcytosis. Potocytosis and transcytosis may use caveolae as routes for internalization (5,8).

Many receptors contain motifs in their cytoplasmic domains that act as recognition sequences for initiating the process of enhanced intracellular uptake of macromolecules. With ligand interaction, the rate of receptor internalization is increased. Whether the intracellular pathways of the internalized molecule and biochemical consequences is the same between the different forms of endocytosis and ligands remains unclear (9-12).

To describe and elucidate the science of receptor-targeted polyplexes used in gene transfer, special attention is given to the many barriers that need to be overcome for gene transfer to be efficient. Factors such as DNA condensation, particle size of the DNA complex, route of administration, stability of the transferred gene in vivo, physical barriers that need to be overcome in order to reach target sites, and other in vivo confrontations, are discussed. The current concepts on binding of DNA complexes to the cell surface receptor and internalization and intracellular trafficking are reviewed, discussing also strategies for enhanced intracellular vesicular release, cytoplasmic trafficking, and nuclear targeting.

II. POLYPLEXES, CELLULAR TARGETING, AND RECEPTOR-MEDIATED ENDOCYTOSIS

A. DNA-condensing Carrier Molecules and Polyplexes

To transfer DNA safely and efficiently into the cell, it needs to be primarily stabilized. This is achieved by using carrier molecules capable of binding DNA. Such DNA-binding carrier molecules are also used not only to condense and compact the DNA into a size preferred by the target cell for being internalized, but also to neutralize the negative charges of the DNA, hence preventing repulsive forces between the DNA and the cell plasma membrane. In addition, carrier molecules may also protect the DNA fromcertain undesired affects of thephysiological environment. DNA-binding carrier molecules should interact with the genetic material in a reversible, noncovalent, nondam-aging manner. Cationic lipid-based systems, formulating DNA into ''lipoplexes,'' and cationic polymers, formulating DNA into ''polyplexes,'' have been used (1). For the majority of the cases, the cationic polymers—polylysine and polyethyleni-mine (PEI)—have been applied for binding and condensing DNA into polyplexes with sizes of 50 nmup to several hundred nm (13,14), which can be taken up by cells (15-19). A series of other DNA-binding polycations have also been used in pol-yplexes: polyarginine (20), protamine (13), or nonhistone nuclear proteins such as high-mobility group (HMG) proteins 1

(21), as well as nuclear proteins such as histones H1, H3, or H4

The cationic portion of the DNA complex can enhance binding to the cell (in addition to ligand-mediatedreceptor binding), and may also facilitate and mediate the transfer of the DNA to the cytoplasm, such as by disruption of the vesicular mem branes. Thus, DNA condensation is essential for efficient transfer into cells; however, not any molecule that binds DNA results in an appropriate condensation.

B. Receptor Ligands for Targeting Specific Cells

Gene transfer vehicles have to fulfill several major delivery tasks: to transfer the therapeutic gene from the site of administration to the surface of the target cells, and to facilitate the internalization of the gene into the cells, trafficking to the nucleus. Incorporation of receptor ligands has been considered to support these tasks [i.e., targeting and enhanced cellular internalization of the transferred (therapeutic) gene, and in some cases, trafficking to and targeting to the nucleus, possibly including intranuclear trafficking]. Furthermore, nonspecific cellular internalization of a DNA-polycation complex via a net positive charge could result in cytotoxic side effects, such as disruption of cell membranes and aggregation of erythrocytes (23). Coupling ligands to the DNA-binding carrier (see Fig. 1) may overcome these obstacles by reducing positive surface charges of DNA complexes, preventing erythro-cyte aggregation, and by enabling the complex to target specific cells, binding them strongly by ligand-receptor interactions, and thus minimizing interactions with nontarget cells (23).

Specific cell binding and enhanced cellular uptake can be mediated by 1 ligand, but could also be regarded as 2 separate processes because some ligands might target and bind a specific cell surface receptor, but not necessarily result in enhanced in-ternalization, whereas other ligands have no cell-type specificity, but efficiently mediate uptake. The importance of these 2 points is very clear when working in vitro as well as in vivo. In vitro, targeting the gene of interest to specific cells is not a problem as such because the delivered gene gets in direct contact with the target cells only, and hence, need not seek out specific cells. However, it is necessary to enhance cellular uptake of the transferred gene. This can be achieved by using ligands as intracellular delivery-enhancing elements. For in vivo application, both cellular targeting and enhanced intracellular delivery of the polyplex are crucial for successful gene therapy.

Ligands, internalized via receptor-mediated endocytosis, represent a wide variety of macromolecules with varying physiological activities, including nutrient provision (e.g., low-density lipoprotein, transferrin); modified molecules from the circulation (e.g., ASGP, plasminogen activator inhibitor complexes); growth factors and hormones (e.g., insulin, VEGF, EGF), and some lysosomal enzymes. Some of these ligands may be coupled to DNA complexes (see Fig. 1), targeting them to specific cells. Ligands can be (1) proteins such as transferrin and asialoglycoproteins, and (2) small natural or synthetic molecules such as folic acid, peptides, or sugar derivatives. Examples of ligands that have been used as conjugates with polycationic carriers for targeted gene transfer are shown in Table 1 and references (9-10,13, and 24-87).

A general aspect to be considered in the selection of a ligand to be coupled to the carrier is that certain ligands are very spe

Figure 1 Assembly of DNA complexes (polyplexes).

cific in targeting certain cells or tissues in the body (e.g., asialoglycoproteins/hepatocytes), whereas others are not [e.g., transferrin/iron supply to many cell types]. Some ligands are internalized very efficiently (e.g., transferrin, anti-CD3 antibody bound to the T cell receptor-associated surface molecule CD3), whereas some others may be internalized either very slowly or not at all. Thus, the choice of ligand for efficient gene transfer is fundamental. Regarding this, one may also take advantage of the ligand binding and internalization enhancement as 2 separate processes, using 2 different ligands, 1 for targeting and the other for internalization. Hypothetically, 2 different lig-ands could also be used, where 1 ligand serves cell binding function only, whereas the other ligand enhances cellular inter-nalization. The target cells need to contain cognate receptors, enabling the ligands to work in concert. Efficient internaliza-tion will only take place with cells containing both receptor types, enabling the 2 ligands to work in concert. Furthermore, the biology of specific cell types may also play an important role for efficient targeting. For example, cellular targeting li-gand coupled to the DNA complex can be used as an element

Table 1 Ligands Used in Receptor-mediated Gene Transfer

Ligands Refs.

Alpha2 macroglobulin 24,25

Anti-CD3 26,27

Anti-CD5 28

Anti-CD117 29

Anti-EGF 30

Anti-HER2 31

Anti-IgG 32,33

Antisecretory component Fab 34-36

Anti-Tn 37

Antithrombomodulin 38

Antibody ChCE7 39

Asialoglycoproteins 40-49

EGF 50-52

Fibroblast growth factor 2(FGF2) 9,53

Folate 54-56

Glycosylated synthetic ligands 57-69

IgG (FcR ligand) 32,70

Insulin 10,71

Invasin 72

Lectins 73-75

Malarial circumsporozoite protein 76

RGD-motif (integrin binding) 77

Steel factor (CD117 ligand) 78

Surfactant proteins A and B 79,80

Transferrin 13,81-87

to only target the corresponding cell, whereas internalization can be achieved via phagocytosis. Thus, cells containing the corresponding receptor will be targeted, but internalization will only take place if the cells contain competent phagocytic apparatus.

1. Ligand Conjugates

Cell-targeting ligands can be coupled to the DNA-binding elements (for examples, see Table 1), forming a conjugate capable of interacting with and condensing the DNA, and targeting it to specific cells (Fig. 1). The DNA-binding element (DNA carrier molecule) should be bound to the ligand, without its DNA binding, condensing, and protective functions being affected, and without affecting the cell targeting property of the ligand. Such a molecular conjugate containing both DNA-binding and cell targeting properties can be combined with DNA, forming the conjugate-DNA complex (''polyplex''). This polyplex (1) should contain the DNA in a highly condensed state, with the ligands positioned in a manner, free to interact with the target cell receptors.

Conjugates are commonly synthesized by covalently coupling the ligand to the DNA-binding polymer. Twenty years ago, reports were published describing the concept of exploiting natural endocytosis pathways of ligands for the delivery of DNA

macromolecules. A method was published for the covalent coupling ofDNA to protein ligands, such as alpha2-macroglob-ulin, formulating the concept of receptor-mediated endocytosis (24). Alternative approaches use compositions containing DNA associated with the ligand in noncovalent mode, using the ligand linked to liposomes (88). This concept was expanded, with approaches such as modifying proteins (transferrin and asialoglycoprotein) with positively charged N-acylurea groups that enable electrostatic binding to DNA, to generate DNA-binding ligands for receptor-mediated gene transfer (89). A chemically more defined approach involves conjugates of asi-alo-orosomucoid, and the polycation poly(L)lysine (40). Complexes of these DNA-binding conjugates with DNA plasmids encoding CAT marker genes or therapeutically relevant genes were shown to result in gene expression both in vitro in cultured HepG2 hepatoma cells, and in vivo in the liver of rats or rabbits (40-47). This was the beginning of the era ''receptor-mediated gene transfer'', where ligand-polycation conjugates were complexed with DNA, to condense, target, and transfer the gene into specific cells. Since then, many successful attempts have been performed in synthesizing ligand-conjugates, for example, conjugation of transferrin, folic acid, anti-CD3 antibody, to the DNA-binding elements polylysine, protamines, histones, intercalators, and complexing them with DNA for achieving targeted gene transfer (see Table I).

Taking transferrin as an example, conjugate synthesis involves the modification of transferrin and DNA-binding poly-cation with bifunctional reagents such as succinimidyl 3-(2-pyridyldithio) propionate (40,81). In the first steps, the activated (succinimidyl) esters can separately react with some amino groups of transferrin and the polycation. Subsequent steps result in disufide bonds (reducible) between the polyca-tion and transferrin. Alternatively, a bifunctional reagent containing a maleimido group has been used (85), resulting in a (nonreducible) thioether linkage. Such an approach is not necessarily specific because the actual site of ligation between transferrin and the polycation is unknown. Conjugate synthesis can also be achieved in a more specific manner (e.g., via ligation through the transferrin carbohydrate moiety) (83). Transferrin contains 2 carbohydrate chains that are attached by N-glycosyl-ationto Asn-413 and Asn-611. The glycan chains have abian-tennary structure, composed of a core bearing 2 N-acetylneur-aminyl-N-acetyllactosamine units. The glycosylation on the transferrin has no known influence on receptor binding or any other biological function (apart from the clearance of asialo-transferin fromtheplasma). Thus, this site of the transferrin carbohydrate moieties is a good choice for attachment of polyly-sine and other nucleic acid-binding elements, without disturbing the cell receptor targeting characteristics of trans-ferrin. To couple the transferrin to the amino groups of polyly-sine, the transferrin carbohydrate groups need to be activated. The terminal point of the transferrin carbohydrate chains consists of sialic acids. The 2 terminal exocyclic carbon atoms of the sialic acids can be selectively removed by periodate oxidation, resulting in the formation of aldehyde groups at the end of the carbohydrate chains.

The concept has been applied to other targeting ligands (Table I). Ideally, the chosen ligand must be recognized by specific cell surface receptors, bound with high affinity and internalized. Most research in this field has been performed by targeting the liver-specific asialoglycoprotein receptor and the ubiquitous transferrin receptor. Other possibilities, such as using antibodies to target specific cells, have been investigated successfully. Anti-CD3 antibody, coupled to carrier-DNA complex has been shown to be very efficient in targeting T cells via binding the CD3 T cell receptor complex (26). Malignant B cells have also been successfully targeted by aids of anti-idio-type antibodies (33). Others have achieved selective targeting of cells by using the ligand folic acid, to target cells that target the folic acid receptor (54-56). It has been demonstrated that the growth factor receptors HER-2 (31), EGF-R (30,51,50), and FGF2-R (9), also highly overexpressed in many human tumors, serve as good targets. Transferring genes into macrophages by mannose/fucose or galactose-specific membrane lectin is another example demonstrating that the receptor characteristics of cells can be exploited to target the cell via an appropriate ligand (57-69).

More recently, the DNA delivery activity of the cationic polymer PEI has been combined with the concept of receptor-mediated gene delivery by incorporating cell-binding ligands (transferrin, anti-CD3 antibody, lactose) by covalent linkage to PEI. Incorporation of cell-binding ligand results in an up to 1000-fold increased transfection efficiency (69,87). This activity was obtained with electroneutral particles, but depends on ligand-receptor interaction, resulting in enhanced cellular internalization. It is important to note that coupling of the ligand did not disturb the DNA condensing and transfer capability of PEI, which is a required aspect in formation of a DNA/conjugate polyplex.

In conclusion, enhancing cellular internalization by incorporating ligands to the carrier, not only enhances cellular inter-nalization, but also avoids problems such as cytotoxicity of positively charged complexes, as well as undesired interactions with other cells and molecules, which would otherwise arise due to the need of a net positive charge of the DNA complex required for efficient (nonligand-mediated) transfec-tion.

Alternative approaches include those that employ DNA-binding domains derived from transcription factors to attach a protein component to plasmid DNA (90). For example, plasmid DNA containing GAL4 recognition motifs can interact with GAL4-containing carrier molecules. Cell-specific ligands can be combined with the GAL4 domain to form a chimeric fusion protein to allow gene delivery to specific cells. For example, GAL4/invasin fusion protein has been shown to transfect target cells in an invasin receptor-dependent manner (91). However, for complex formation, and hence condensation, an additional condensing agent such as polylysine is required.

Was this article helpful?

0 0

Post a comment