| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Radiation Oncology, Fox Chase Cancer Center, Philadelphia, Pennsylvania 19111, USA;
Department of Internal Medicine, Mayo Clinic, Rochester, Minnesota 55905, USA;
Department of Internal Medicine, University of South Alabama Hospitals, Mobile, Alabama 36617, USA; and
¶ Department of Radiology, University of Chicago Hospitals, Chicago, Illinois 60637, USA
1Correspondence: Fox Chase Cancer Center, Department of Radiation Oncology, 7701 Burholme Avenue, Philadelphia, PA 19111, USA. E-mail: JM_Kaminski{at}fccc.edu
ABSTRACT
Gene therapy in eukaryotes has met many obstacles. Research into the design of suitable nonviral vectors has been slow. To our knowledge, no nonviral vector has been proposed that allows for the possibility of highly efficient, site-selective integration into the genome of mammalian cells. On the basis of prior studies investigating the components necessary for transposon, retrovirus-like retrotransposon, and retroviral integration, we propose a nonviral system that would potentially allow for site-selective, efficient integration into the mammalian genome. Transposons have been developed that can transform a variety of cell lines. For example, the Sleeping Beauty transposon (SB) can transform a wide range of vertebrate cells from fish to human, and it mediates stable integration and long-term transgene expression in mice. However, the efficiency of transposition varies significantly among cell lines, suggesting the possible involvement of host factors in SB transposition. Here, we propose the use of a chimeric transposase (i.e., transposase-host DNA binding domain) to bypass the potential requirement of a host DNA-directing factor (or factors) for efficient, site-selective integration. We also discuss another potential method of docking the transposon-based vector adjacent to the host DNA, utilizing repetitive sequences for homologous recombination to promote efficient site-selective integration, as well as other site-selective nonviral approaches.—Kaminski, J. M., Huber, M. R., Summers, J. B., Ward, M. B. Design of a nonviral vector for site-selective, efficient integration into the human genome.
Key Words: gene therapy • transposon • recombination
BACKGROUND
Nonviral vectors are the future in gene transfer because of the limitations of viral vectors such as pathogenicity, expense of production, and systemic instability. We now know the three major components for efficient transport of viral and nonviral vectors through the cytoplasmic membrane and into the nucleus of eukaryotic cells. These include a specific ligand for receptor-mediated endocytosis, an endosomal disruption factor, and a nuclear localizing signal. These components have been employed successfully in nonviral vectors (1
2
3
4
5
6)
. In vectors that lack or fail to interact with a nuclear localizing signal, efficient transfection will occur only in those cells that are actively dividing. For integration to occur, an enzyme (e.g., transposase) is required to mediate the process.
Nonviral vectors (i.e., lipid based, polymer based, lipid-polymer based, and polylysine) are a synthetic means of encapsulating transgenic DNA until it reaches the cellular target. Compared with viral vectors, nonviral vectors are safer to prepare; the risk of pathogenic and immunological complications is diminished. Nonviral vectors have been designed by modifying the surface of the vector for targeted therapy (7
8
9
10
11
12)
. Liposomes are typically internalized into endosomes, which are then frequently directed to lysosomes, thus degrading the plasmid. Endosomal disruption factors and nuclear localizing signals have been employed in these vectors. DOPE (fusogenic lipid dioleylphosphatidylethanolamine) has been added to cationic liposome formulations. Electron microscopy observations have clearly shown the effect of DOPE on destabilization of the endosomal compartment containing the complex (3)
. The use of folate attached to ligands on the surface of the liposome (13
, 14)
changes the route of internalization, and the liposomes are preferentially taken up by the caveolae vesicles and their contents are more readily released into the cytoplasm (15)
. Liposomal DNA carriers known as stabilized plasmid-lipid particles (SPLPs) have recently been developed. SPLPs are small (70 nm) and consist of plasmid DNA encapsulated within a lipid bilayer composed of DOPE, a cationic lipid, and polyethylene glycol-ceramide. Plasmid encapsulation efficiencies of 50–70% are thereby achieved, and the plasmid/liposome ratio is approximately 1. SPLPs exhibit circulation half-lives of about 6 h following intravenous injection. However, the lipoplexes (plasmid DNA and liposome) are mainly limited to transfecting dividing cells unless a nuclear localizing factor is present or interacts with the vector (16)
. Furthermore, efficient host integration does not occur except in transposon-based plasmids (17
18
19
20)
. Thus far, liposomes have demonstrated their safety in human gene therapy trials (21
22
23
24)
.
The putative model of integration is similar in retroviruses, transposons, and retrovirus-like retrotransposons. For example, the catalytic domain is conserved in integrases and transposases. Furthermore, the terminal repeats are highly conserved typically in transposons, retrotransposons, and retroviruses. In vitro reactions have shown that integrase and transposase are the only enzymes necessary for integration (25
26
27
28)
Integrase and many transposases in bacteria and eukaryotes have been shown to bind specifically to the att site at the ends of the terminal repeats. They require the presence of CA at the 3'-end for both processing and cleavage-ligation (29
, 30)
. The mechanisms of transposon and retroviral integration are discussed in detail elsewhere (6
, 31
32
33
34)
.
Integrase and transposase depend on their own DNA binding domain or an interaction with a host DNA-directing factor to direct the DNA-enzyme complex (e.g., transposon-transposase) in juxtaposition to the host DNA for integration to occur (25
, 35
36
37)
. If the host does not have this directing factor or a specific host-DNA sequence recognized by the transposon-transposase complex, the efficiency of integration decreases substantially (25
, 38)
. For example, a specific human endogenous protein, integrase interacting 1, has been shown to affiliate with integrase and stimulate integration in vitro and possibly in vivo by binding and directing integrase to DNase 1 hypersensitive sites (25)
. Alternatively, the yeast retrovirus-like element Ty3 inserts at the transcription start sites of genes transcribed by RNA polymerase III because of its affiliation with this complex (37)
. Furthermore, some transposases or integrases require certain sites in the host DNA for catalytic activity even if the DNA-enzyme complex is brought into the vicinity of the host DNA. For example, Tc1/mariner transposon integrates into a TA dinucleotide (32)
.
DNA transposable elements for genetic manipulation have been available for more than 15 years. This technology has been applied in both bacteria and eukaryotes to verify whether a cloned DNA fragment contains the whole functional gene of interest. Rubin and Spradling first demonstrated this for P elements of Drosophila melanogaster (39)
. A fragment of DNA carrying the rosy gene was inserted within the terminal repeats of a P element and then cloned into a plasmid. This plasmid and another encoding the transposase were injected into the embryos of an M strain with a deletion in the rosy gene. About 50% of the flies derived from the injected embryos possessed the rosy phenotype, thereby suggesting that the rosy gene inserted into the chromosome and maintained its function at various sites within the genome. Furthermore, none of the flanking plasmid DNA was integrated in the host genome, suggesting that excision from the plasmid took place only at the terminal repeats (39)
. It must be noted that the terminal repeats of the transposon are the only region required for integration and can be constructed in a plasmid for amplification.
Integrase maintains its activity when fused to other proteins. This has been demonstrated by the use of the lambda repressor-integrase (40)
and maltose binding protein-integrase fusion proteins (41)
. In addition, chimeric recombinases, transcription factors, and oncogenes, among others, have maintained their activity when fused to other protein domains (42)
. However, attempts at in vivo targeting of site-selective retroviruses that included sequences encoding integrase fusion proteins have not yet been demonstrated (43
44
45)
. Recently, Holmes-Son and Chow used an in trans approach to deliver vpr-integrase-LexA to an integrase-defective virus. They were then able to demonstrate integration mediated by the retrovirus, containing integrase-LexA, in cells, although site selectivity was not determined (46)
. This method holds much promise but suffers from the limitations of retroviral vectors (47)
.
TRANSPOSON-BASED GENE THERAPY
The potential for transposon-based gene therapy is currently under consideration (35
, 48)
. The Tc1/mariner elements are promiscuous and have been successfully used as transgene vectors from one species to another in flies (49
50
51
52
53)
, mosquitoes (54)
, bacteria (55)
, protozoa (56)
, and vertebrates. Transgenic animals have been generated by microinjection of transposon-based plasmid vectors into eggs of chicken (57)
and zebrafish (58
, 59)
. Furthermore, successful transposition in cell lines has been obtained for mouse embryonic stem cells (60)
and human cells (48
, 61
62
63)
. The transposon Sleeping Beauty (SB) can efficiently insert transgenes into the mammalian chromosomes in vivo with long-term, possibly life-long, transgene expression in the livers of mice (48)
. However, depending on the turnover rate of the specific tissue, integrating vectors may require multiple administrations to maintain persistence of gene expression. Readministration of virus-based vectors can promote immune responses that can result in life-threatening systemic effects and can limit gene transfer efficacy (64
, 65)
. The cytological and immunological consequences of readministration of a transposon-based system are uncertain (35)
.
The plasmid-based vectors achieve stable chromosomal integration at levels comparable with the use of many virus-based technologies. As a rule, increasing the length of DNA between the terminal repeats of the transposon-based plasmid will decrease the efficiency of integration, whereas decreasing the length of DNA outside the terminal repeats increases the efficiency. Furthermore, DNA sequences around the transposon donor site can dramatically increase or decrease transpositional efficiencies (17)
. The efficiency and precision of transposition vary significantly among cell lines, suggesting a potential involvement of host factors in SB transposition. Yant et al. concluded that basic research into the mechanisms of DNA-mediated transposition may suggest new ways to further improve this novel integrating system (35)
. Transposon-based vectors (i.e., in the form of a plasmid construct) can easily be produced in mass quantity, purified, and maintained pathogen-free in any laboratory at minimal cost.
SITE-SELECTIVE TRANSPOSON GENE THERAPY
On the basis of the requirements for integration of the transposable elements, it appears a host DNA-directing factor is necessary for efficient integration by juxtaposing the transposon-transposase complex adjacent to the host DNA. Indeed, Tc1/mariner transposases do have DNA binding domains. However, these DNA binding domains are apparently not site selective (35)
, possibly lack strong recognition sites in certain host genomes, and may require other host proteins for efficient integration by docking the transposon-transposase to the host DNA. Two methods can be employed once the transposon-based plasmid is in the nucleus to potentially increase the site selectivity and the efficiency of integration. The first method would be to design a chimeric transposase (host DNA binding domain transposase); the second method would employ a recognition site (or sites) on the plasmid that would recognize an endogenous protein (or a newly introduced protein, e.g., produced from a gene located on the plasmid) that would then direct the complex into the vicinity of the host DNA for site-selective integration. A third method of incorporating repetitive elements (e.g., Alu-like sequences) in the transposon-based plasmid may enhance docking and at the same time allow for either homologous recombination (66
, 67)
or integration of the transgene into the host genome.
WHERE TO TARGET?
Some internal repeats (e.g., some short and long interspersed nuclear elements) are permissive for site-selective integration (68
, 69)
and would allow for transgene expression even without nuclear matrix attachment regions flanking the transgene (66
, 67)
. Proteins that selectively bind to interspersed repeat elements have been identified (70
71
72
73)
. Development of fusion proteins incorporating DNA binding domains to known transcription-permissive, repetitive DNA sequences is possible and would allow targeted integration as described earlier.
The Tc1/mariner transposase requires TA dinucleotides for integration (32)
. Many of these interspersed repeat elements contain TA dinucleotides (66
, 71
, 72
, 74)
. However, a mutant mariner/Tc1 transposase has recently been developed that allows insertion, although at reduced levels, at dinucleotides other than TA, including sequences with GC base pairs (75)
. By targeting repetitive sequences (i.e., more targets), the chimeric transposase would likely increase the efficiency of integration.
OTHER SITE-SELECTIVE NONVIRAL APPROACHES
Homologous recombination allows for selectivity, but the efficiency of recombination is far too low to permit its use in gene therapy in vivo even with targeted homologous recombination of repetitive sequences (66
, 67
, 76)
. Correction of point mutations can be accomplished with the use of RNA/DNA oligonucleotides, also termed chimeraplasts. The chimeraplast is designed to specifically bind to the target DNA sequence and create a mismatched base pair (or pairs). This mismatch then signals the endogenous cellular repair process to change the gene sequence by removing the mismatched base pairs and utilizing the chimeraplast template to synthesize the desired sequence. Although impressive results were initially reported (77)
, later results have shown persistent failures with this technology (78
79
80)
. Apparently, careful validation of previously reported experiments was not confirmed by others (80)
.
Site-specific recombinases allow recombination, and some do not require cofactors, thereby allowing activity outside their normal environment. For example, Cre recombinase, although derived from Escherichia coli phage P1, acts efficiently in plant, yeast, and mammalian cells (18)
. Site-selective recombinases such as FLP, Cre, and ß-recombinase perform both integration and excision efficiently with the same target sites; however, the net integration frequency is low (e.g., 0.03% for Cre) (18
19
20)
.
Calos and colleagues at Stanford University have demonstrated that the phage 
C31 integrase mediates facile integration in the human cell environment at the attB and attP phage attachment sites on extrachromosomal vectors through intraplasmid and interplasmid integration reactions of greater than 50% and 7.5%, respectively (81)
. This group then constructed human and mouse attP-containing cell lines. Even in these modified human and mouse cell lines, the integration frequencies were less than 1% and 5%, respectively. In nonmodified cell lines, integrase-mediated recombination at endogenous sites (i.e., "pseudo" att sites) made up about 90% of the integrating events; however, the frequency was only 5 to 10 times above the background of random integration (38)
. Although promising, this technology in its current state is not suitable for in vivo gene therapy.
CONCLUSION
From the homology among the different transposable elements, we have probably determined all the necessary exogenous components in a nonviral vector that allow for efficient, site-selective integration into the host genome. The terminal repeats, the enzyme that is required for integration, and a host DNA-directing factor appear to be the only components necessary for efficient integration if targeted to a transcription- permissive region. Therefore, we propose to use a construct containing the transgene flanked by terminal repeats of a transposable element, e.g., SB, and a required chimeric enzyme (host DNA binding domain transposase) in a nonviral packaging system for targeted integration into the host genome. This chimeric enzyme would substitute the DNA binding domain of the transposase for one that is host specific and site-selective, thereby bypassing the requirement of a host DNA-directing factor. The chimeric transposase cDNA could be located outside the terminal repeats on the transposon-based plasmid or introduced as a separate plasmid. The transposase could be under the control of an inducible or tissue-selective enhancer-promoter. Another method could employ recognition sites on the plasmid that would recognize an endogenous protein (or a newly introduced protein) that would then direct the complex to the vicinity of the host DNA. Finally, incorporating repetitive elements (e.g., Alu-like sequences) in the transposon-based plasmid may enhance docking and at the same time allow for either homologous recombination or integration of the transgene into the host genome.
ACKNOWLEDGMENTS
We thank David G. Perryman and Mark Westhafer, Ph.D., of the Law Offices of Needle and Rosenberg, P.C., George Wu, M.D., Ph.D., Nurul Sarkar, Ph.D., and Richard Katz, Ph.D., for our discussions with them and for reviewing portions of the manuscript.
Received for publication February 27, 2002. Revision received April 18, 2002. REFERENCES
C31 integrase. Mol. Cell. Biol. 21,3926-3934This article has been cited by other articles:
|
|
 |
|
  S. C.-Y. Wu, Y.-J. J. Meir, C. J. Coates, A. M. Handler, P. Pelczar, S. Moisyadi, and J. M. Kaminski From the Cover: piggyBac is a flexible and highly active transposon as compared to Sleeping Beauty, Tol2, and Mos1 in mammalian cells PNAS, October 10, 2006; 103(41): 15008 - 15013. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. J. Maragathavally, J. M. Kaminski, and C. J. Coates Chimeric Mos1 and piggyBac transposases result in site-directed integration FASEB J, September 1, 2006; 20(11): 1880 - 1882. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Vanderbyl, G. N. MacDonald, S. Sidhu, L. Gung, A. Telenius, C. Perez, and E. Perkins Transfer and Stable Transgene Expression of a Mammalian Artificial Chromosome into Bone Marrow-Derived Human Mesenchymal Stem Cells Stem Cells, May 1, 2004; 22(3): 324 - 333. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
