| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
CONCISE REVIEW |
a Department of Medicine,
b Department of Genetics, Cell Biology, and Development, University of Minnesota Medical School, Minneapolis, Minnesota, USA;
c Institute of Liver Studies, Kings College Hospital, Denmark Hill, London, England
Key Words. Chimeraplasty • Gene repair • Chimeric RNA/DNA oligonucleotide • Single-stranded oligonucleotide • Triplex-forming oligonucleotide • Small fragment homologous replacement • Sleeping Beauty transposon system
Clifford J. Steer, M.D., Department of Medicine, Mayo Mail Code 36, Mayo Building, Room A536, 420 Delaware Street S.E., Minneapolis, Minnesota 55455, USA. Telephone: 612-624-6648; Fax: 612-625-5620; e-mail: steer001{at}tc.umn.edu
 |
ABSTRACT |
|---|
 Top Abstract IntroductionGene Correction StrategiesSleeping Beauty Transposon...Conclusions and PerspectivesReferences |
|---|
Site-specific genetic repair has many advantages over augmentation although it too is associated with significant limitations. This review outlines the advantages and disadvantages of gene correction. In particular, we discuss technologies based on chimeric RNA/DNA oligonucleotides, single-stranded and triplex-forming oligonucleotides, and small fragment homologous replacement. While each of these approaches is different, they all share a number of common characteristics, including the need for efficient delivery of nucleic acids to the nucleus. In addition, we review the potential application of a novel and exciting nonviral gene augmentation strategy—the Sleeping Beauty transposon system.
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionGene Correction StrategiesSleeping Beauty Transposon...Conclusions and PerspectivesReferences |
|---|
Traditional gene therapy has utilized both viral and nonviral delivery systems [4, 5]. Prolonged transgene expression is dependent on random genomic insertion. However, adenoviral vectors do not integrate, and actively dividing cells are required for retroviral integration, both major restrictions for somatic gene therapy. Production of the viral vectors is time consuming and expensive. In addition, despite the production of nonreplicative viral vehicles, the possibility of introducing replication competent viruses still exists. Nonviral delivery systems have the potential advantages of being less immunogenic and providing greater ease of production, purity, and standardization of molecules.
The limitations associated with gene augmentation have resulted in the pursuit of alternative strategies. Specifically, there is a growing interest in gene repair in which nucleic acids are designed to affect correction of precise genomic mutations [6]. Site-specific gene repair has a number of advantages over gene augmentation. First, in situ repair of the targeted mutation allows the gene to remain under the control of its natural regulatory elements. Second, precise repair has the potential to address both recessive and dominant mutations, whereas augmentation is generally restricted to recessive diseases. Third, the repair process usually involves the use of small synthetic molecules, which are typically less immunogenic. In fact, immunostimulatory CpG motifs may contribute significantly to the exaggerated immune states associated with some vectors/constructs [7]. Fourth, the smaller size of repair molecules may increase cell delivery and nuclear translocation compared with the larger plasmid systems. Finally, effective and permanent gene repair may be achieved with fewer treatments, and possibly even a single treatment.
While there are many advantages to gene correction, significant limitations do exist for the different approaches. For example, the exact sequence of each targeted mutation and surrounding genome is required, thus creating the notion of a "designer gene repair." Delivery remains a significant hurdle irrespective of cell type and/or repair mechanism. In addition, there are marked variations in the frequency of repair among the different strategies, as well as within individual ones. This article reviews the current gene repair strategies of chimeraplasty, single-stranded and triplex-forming oligonucleotides, as well as small fragment homologous replacement.
A major barrier in nonviral gene replacement therapy has been the lack of genomic integration of the relevant transgene, which obviously precludes long-term expression. The "reawakening" of Sleeping Beauty, a nonviral transposon system, is a major step in overcoming that barrier. The system utilizes a reconstructed transposase enzyme capable of excising a transposon by binding to specific sequence recognition sites within its DNA. The transposon/transposase complex inserts the transposon, containing the relevant transgene, into the genomic sequence at a TA dinucleotide site. The combination of transposase and transposon provides a vehicle for stable transgene integration and long-term expression, thus overcoming the Achilles' heel of nonviral gene augmentation. This article reviews Sleeping Beauty as a novel technology that combines the advantages of nonviral delivery with genomic integration and persistent transgene expression.
|
GENE CORRECTION STRATEGIES |
|---|
|
TopAbstractIntroductionGene Correction StrategiesSleeping Beauty Transposon...Conclusions and PerspectivesReferences |
|---|
The original chimeric design was a contiguous stretch of 68 nucleotides containing both RNA and DNA residues (Fig. 1
). One strand of the heteroduplex structure consisted of a central pentameric block of DNA bases flanked on either side by ten 2'-O-methylated RNA residues with increased resistance to RNaseH activity. The opposing strand consisted of all-DNA bases complementary to the RNA/DNA strand. Polythymidine hairpins were placed at both ends to maintain the secondary structure and a 3' 5-base GC clamp was introduced for increased exonuclease resistance. The 3' and 5' ends were juxtaposed, but not joined, to permit unwinding of the molecule and the potential interaction with recombinase proteins. The nick may also have provided a greater degree of flexibility to the chimeric molecule, thereby increasing its ability to interact with its homologous target [16]. Both upper and lower strands were made homologous to the intended genomic target, except for a single base mismatch in the center of the pentameric DNA stretch of the hybrid strand. The chimeric oligonucleotide was aligned in perfect register with its target gene except for the intended base mismatch. This produced a substrate that was recognized by the endogenous DNA repair system.
|
Recently, cell-free systems have been used to study and improve the chimeraplast structure. These studies found that repair frequency was markedly lower when the RNA/DNA strand alone contained the mismatch, whereas if the mismatch was only in the DNA strand, conversion rates were 50% greater when compared with standard chimeraplasts. These results suggested that the role of the RNA strand is to stabilize the D-loops and, in fact, is more efficient without a mismatch. In contrast, genomic conversion appears to be driven by the mismatch in the all-DNA strand [21–23].
Chimeraplasty was initially reported in an episomal target, followed by genomic repair of the sickle cell mutation in cultured lymphoblasts [24, 25]. The early studies demonstrated the vital role of transfection efficiency and nuclear delivery in maximizing conversion frequencies. In fact, specific targeting to the asialoglycoprotein receptor on hepatocytes resulted in significant increases in delivery using either lactosylated polyethyleneimine (L-PEI) or galactocerebroside-formulated liposomes as vehicles [16, 26, 27]. Chimeraplasty has been demonstrated in cell-free extracts [28–30], cell culture [29, 31], plant cells [32, 33], and more recently in yeast [34, 35], thus confirming the broad potential and application of the technique.
A major criticism of these results was that the high conversion frequencies could have resulted from polymerase chain reaction (PCR) artifacts and/or contamination by different cells [36, 37]. In an elegant study, melanocytes from albino mice were isolated and transfected with chimeraplasts to repair the defect in pigment production [29]. These cells contain a point mutation in the tyrosinase gene that is an essential enzyme in melanin synthesis. After transfection with chimeraplasts, single clones of pigmented cells were identified and isolated. Genomic correction was established by restriction fragment length polymorphism (RFLP), immunoblot analysis, and enzymatic activity and, moreover, the geno/phenotypic changes remained stable over numerous passages of the clones.
These findings prompted additional confirmatory in vivo studies. To this end, an RNA/DNA oligonucleotide was designed to induce a point mutation in the factor IX gene resulting in a phenotype that could be monitored by factor IX activity. The chimeraplast/vehicle complexes were designed to selectively bind the asialoglycoprotein receptor on hepatocytes and were then injected into the tail vein of rats. A dose-related genomic conversion rate of 15%-40% was observed together with a corresponding reduction in factor IX activity. Importantly, the altered clotting profiles and genomic change remained stable for nearly 2 years [38]. The same approach and delivery system was used to promote liver-specific delivery and to repair a single base deletion in the Gunn-rat model of the human disease Crigler-Najjar syndrome type 1. In this animal model, the mutation has been characterized as a single point deletion in the UDP-glucuronosyltransferase gene (UGT1A1) producing a frame shift and premature stop codon. Dose-dependent base insertions of up to 23% were achieved with long-term phenotypic change [39].
These two studies emphasized the potential of chimeraplasty in addressing many of the inherited disorders of the liver. Recently, the mutant isoform of apolipoprotein E (apoE2), associated with hyperlipidemia and premature atherosclerosis, was converted to the wild-type apoE3 [40]. Conversion rates as high as 54% were reported in stably transfected Chinese hamster ovary (CHO) cells, as well as 25% conversion in a transgenic animal model with a demonstrated phenotype. Interestingly, in cell culture there was an approximately 50% increase in conversion with chimeraplasts containing 30 modified RNA residues (i.e., an 88-mer) compared with the standard 68-mer molecule. This may reflect greater binding stability of the chimeraplast/genomic DNA duplex with the added RNA bases and/or an increased resistance to nuclease degradation. Surprisingly, the high levels of conversion were observed using unmodified PEI, and suggest that even higher frequencies could be achieved in vivo with hepatocyte-specific delivery.
Successful chimeraplasty has now been demonstrated in a variety of tissues, including skin melanocytes [41] and skeletal muscle in both dog and mouse models of Duchenne muscular dystrophy [42, 43]. Interestingly, there are little data in support of successful base conversion in stem cells. However, chimeraplasty has been used to introduce the sickle cell mutation in a CD34+-enriched cell population [44]. In fact, the normal ß-globin gene was mutated at a frequency of 5%-11% as determined by PCR and RFLP analyses. That study supports the notion that gene repair by chimeraplasty may be a viable approach in the treatment of a variety of genetic disorders, including sickle cell disease and other hemaglobinopathies. Finally, it was recently reported that mitochondrial extracts obtained from liver cells possess the necessary factors to perform chimeraplasty [45], suggesting a possible approach to treating these debilitating disorders [46].
Single-Stranded Oligonucleotides
Studies designed to optimize the chimeric structure suggested that the all-DNA strand directed the conversion event. Therefore, modified short 25-50 mer single-stranded oligonucleotides were tested for their ability to affect genomic alterations. The results suggested that a short single-stranded oligonucleotide with a central single base mismatch could, in fact, create a distortion with its target locus, resulting in DNA repair and base conversion. Like chimeraplasty, the precise mechanism of action remains unclear, but appears to involve mismatch repair proteins, polymerases, ligases, and the recombinase family of proteins.
In a recent study, the ends of the oligonucleotides were capped with phosphorothioate nucleotides to increase resistance to nucleases [21]. These modified molecules were four times more active than their unmodified counterparts. Surprisingly, all-phosphorothioate molecules showed no activity suggesting that unmodified nucleotides are required to stimulate recognition and repair of the mismatch. Mutant strains of E. coli were used to identify the essential proteins involved in the base conversion. The study demonstrated that, in contrast to chimeraplasty, nucleotide conversion by the single-stranded oligonucleotides was independent of the mismatch proteins MSH2 and MSH3. The broad application of single-stranded oligonucleotides was tested in diverse systems of both mammalian and plant cell-free extracts.
Previous studies have supported the potential of single-stranded molecules to correct point mutations. For example, a 40-nucleotide (nt) oligonucleotide cotransfected into human cells successfully repaired a plasmid containing a mutant copy of the neomycin phosphotransferase gene [47]. More recently, a mammalian cell-free extract system was used to determine that the optimal molecule length was 25 to 61 nt and that there was no significant difference in targeting the nontranscribed or transcribed strand [48]. A chimeric single-stranded oligonucleotide consisting of a pentameric DNA region flanked by ten 2'-O-methylated RNA residues showed significantly less activity. In the same study, CHO cells were then transiently transfected with a mutated ß-galactosidase plasmid. A single-stranded DNA oligonucleotide produced a significantly higher conversion frequency compared with the cell-free extract system. Surprisingly there was a 1,000-fold higher conversion when targeting the nontranscribed compared with the transcribed strand. When genomic DNA was targeted, nt conversion was higher than in the cell-free extract but an order of magnitude lower than the episomal studies. The optimal length of the molecule was 45 nt and, again, higher conversion frequencies were observed when targeting the nontranscribed strand.
The differences in activity observed among the three systems could be explained by a number of factors, including base pairing, the loss of essential proteins, accessibility of the target locus, as well as transcriptional activity. This may simply reflect structural interference or involve a dominant negative role for transcription factors such as RNA polymerases. In addition, it remains to be determined if targeting the nontranscribed strand is more likely to promote genomic DNA repair, while targeting the transcribed strand may lead to "repair" of the oligonucleotide. Most recently, single-stranded DNA molecules 39 nt in length were examined for their ability to bind the RecA protein [49]. The results indicated that pyrimidine-rich single-stranded molecules with little secondary structure and no base stacking bound RecA most strongly. In fact, preferential sequence-specific binding of oligonucleotides to mammalian recombinases may explain some of the variability in success rates observed at different genomic target sites. Taken together, these data underscore the requirement for in vivo studies to evaluate the feasibility and application of this approach.
Triplex-Forming Oligonucleotides
Repair by chimeraplasty and single-stranded oligonucleotides may be limited to several bases, thus restricting the potential clinical applications of these techniques. Modification of DNA by triplex-forming oligonucleotides (TFOs) derives from the observations that single-stranded DNA or RNA molecules can form stable and specific triple helical structures with homopurine-rich areas of the genome (Fig. 2A). The binding specificity and affinity results from hydrogen bonds formed in the major groove between the oligonucleotide and the purine-rich target sequence [50]. Although the requirement for a homopurine-rich (15-30 bp) region limits the number of potential sites, these regions occur about every kilobase of genome and make it probable that the majority of genes will contain TFO targets [51]. Replacement of the phosphodiester backbone with cationic phophoramidate linkages reduces the length of polypurine sequence required for effective activity and genomic DNA targeting in mouse cells [52].
|
More recently, TFOs containing mutagenic adjuncts, such as psoralen, have been used to target and induce site-specific chromosomal breaks within the genome (Fig. 2B
) [53]. The resultant DNA damage stimulates the genomic repair system and thereby increases the frequency of homologous recombination. This process has successfully produced specific mutations in a variety of somatic cells [54, 55]. Interestingly, TFOs without DNA damaging adducts have also been shown to facilitate homologous recombination, suggesting that TFO binding itself may be mutagenic and recognized as abnormal by repair systems [56, 57]. The mechanism for correction appears to involve at least the nucleotide excision repair (NER) and transcription-coupled repair pathways [58].
These limitations of TFOs for gene repair led to the development of a novel hybrid or bifunctional molecule (Fig. 2C). These molecules consisted of a triplex moiety linked to a DNA fragment homologous to a local area to which the TFO binds. They exploit the ability of the TFO to form stable complexes and to induce homologous recombination/NER processes in the vicinity of its target region. Combining this TFO sequence and a short homologous DNA fragment appears to create a heteroduplex with its intended target. That, together with the ability of the TFO to induce homologous recombination/NER has been used to repair an episomal target sequence [59]. Conversion was achieved with both single- and double-stranded tethered DNA and appeared to be dependent on the XPA protein, HsRad51, and NER system [60]. Surprisingly, this cell-free study also demonstrated that targeting could be achieved by co-mixture of binding and targeting domains that were not physically linked. In a separate study, transfection of a lymphoblast cell line with a mutant ADA gene and a human glioblastoma cell line mutant for p53 led to correction efficiencies of 1%-2% and 7.5%, respectively [61]. Bifunctional molecules may play a significant role in the future of gene repair.
Small Fragment Homologous Replacement
The generation of transgenic knockout mice via homologous recombination has greatly expanded our knowledge on the pathophysiology of human diseases [62, 63]. Classical homologous recombination is limited by the difficulty in the construction of targeting vectors, delivery techniques, the high incidence of nonhomologous recombination events, and the general low levels of success. Suboptimal targeting may, in part, be due to large areas of nonhomologous sequence contained in the vector, most commonly selectable marker genes such as for neomycin resistance. These regions of nonhomology markedly reduce the fidelity and stability of any resultant pairing event [6].
A modified approach has been developed to address the fundamental problems associated with homologous recombination as a gene targeting strategy. It is based on the use of single-stranded DNA fragments hundreds of bases in length that are homologous to the target sequence. Specific mismatches are designed into the oligonucleotides to specifically alter the genomic sequence by one or even several nucleotides (Fig. 3). These DNA fragments are typically single-stranded molecules synthesized by PCR but can also be double stranded. The mechanism of action, now referred to as small fragment homologous replacement (SFHR), is unclear but may involve similar pathways to that of homologous recombination [64]. It is likely that SFHR functions by targeting and replacement. In short, the mechanism involves the necessary proofreading and annealing to homologous target regions. The recombinase protein systems function to create intermediate structures, subsequent strand invasion, and exchange of genetic material. However, the precise mechanism of repair is unknown and may actually involve, heretofore, uncharacterized pathways for DNA repair. Early studies demonstrated the potential of this strategy with repair of genomic mutations at frequencies of 10-5 using small cDNA sequences that were comparable to dsDNA vectors [65, 66].
|
F508 deletion in the cystic fibrosis transmembrane conductance regulator gene (CFTR). In the first report, the mutation in transformed epithelial cells was corrected with a single-stranded molecule of 500 nt in length complementary to the wild-type sequence [67]. The repair frequency was about 1%, and a phenotypic effect of genomic correction was confirmed by detection of functional CFTR protein. In a subsequent study, a 488 base oligonucleotide containing a silent mutation (creating a novel Xho1 restriction site) 100 nt from the CFTR deletion was used to target normal and transformed epithelial airway cells [68]. The system was tested for the ability of the single-stranded DNA to replace the 3-bp deletion, despite the presence of an upstream mismatch. The transfected oligonucleotides produced site-specific change in both transformed and normal cells. The replacement was not affected by the upstream mismatch and was confirmed by mRNA transcript expression. However, the relatively low transfection efficiency resulted in poor delivery of the intact single-stranded DNA molecule to the nucleus. A combination of more efficient delivery and improved nuclease resistance could potentially increase genomic correction to therapeutic levels. A major problem with targeting the CFTR deletion is the absence of a selection mechanism to accurately define the genetic changes. To address this, a 4 bp deletion in the zeocin antibiotic resistance gene was utilized as a selectable marker after recovery and transformation of E. coli and culture on appropriate selection media [69]. Correction frequencies up to 4% were reported and confirmed by Southern analysis.
Recently, successful in vivo application of SFHR to an accurate animal model of a human disease has been reported. In the first study, the point mutation in the mdx mouse model of Duchenne muscular dystrophy was targeted [70] in both primary myoblast cultures and direct injection of affected muscle. Repair in cell culture was 15%-20% by PCR analysis, although there was no detection of normal dystrophin protein. Repair of 5 x 10-4 – 0.1% was detected with direct injection of the muscle, but again there was no evidence of gene expression at either the transcript or protein level. The genomic conversion frequencies were lower than those reported for chimeraplasty, which did result in protein expression [42, 43]. It was suggested that the disparity between the genomic repair and protein expression was possibly due to toxicity of the transfected agent on myoblasts, or a delay in protein expression.
Although no functional protein was detected, the results from the mdx mouse model were encouraging. In contrast, SFHR successfully mediated expression of the altered transcript for the CFTR 508 mutation in normal mouse lung [71]. This study used a 783-base fragment homologous to normal CFTR except for the 3-base deletion. The targeting fragment was again designed to introduce a novel restriction enzyme site upstream from the target zone. Interestingly, this study employed four different transfecting agents with varying results reconfirming that delivery is critical for successful gene therapy. The study also underscored the difficulties in extrapolating results from the in vitro to in vivo settings.
A potential concern with SFHR is that self-annealing of the single-stranded molecules in the host cell could form double-stranded DNA molecules that are able to randomly integrate into the genome [6]. Random insertion and the potential for insertional mutagenesis could be detected by inverse PCR or Southern genomic DNA studies. Additional confirmatory studies will be required to sustain the initial enthusiasm about the technology. Finally, it has been reported that adeno-associated viral vectors were used to introduce single base substitutions at relatively high frequencies in normal human cells, including the correction of point mutations responsible for genetic diseases [72, 73]. It remains to be determined whether the repair process overlaps or shares characteristics with that of SFHR.
|
SLEEPING BEAUTY TRANSPOSON SYSTEM |
|---|
|
TopAbstractIntroductionGene Correction StrategiesSleeping Beauty Transposon...Conclusions and PerspectivesReferences |
|---|
The use of the transposon as a molecular genetic tool has a rich history in plants, bacteria, fungi, and insects. Large-scale transposon mutagenesis is being applied to identify the function of genes in a number of organisms in which controlled transposition has been possible [78]. However, the application to vertebrate species is relatively recent. Reconstitution of an ancient transposon, Sleeping Beauty, from sequence alignment of nonfunctional remnants of members of the Tc1/mariner superfamily of transposons within the genomes of salmonids [79] provided the first functional class II transposon for use in vertebrate species (Fig. 4). Subsequent to the "reawakening" of Sleeping Beauty, functional Tc1/mariner superfamily members have been isolated from insects and observed to transpose following transfection into cells of fish, birds, and mammals ([80], and references therein). Sleeping Beauty requires no host factors for transposition and therefore transposes in cells from a wide range of organisms, including humans [77]. Whereas Tc1 transposons require one binding site for their transposase in each IR, Sleeping Beauty requires two direct repeat (DR) binding sites within each IR, and is therefore classified with Tc3 in an IR/DR subgroup of the Tc1/mariner superfamily [81, 82]. Sleeping Beauty transposes into TA dinucleotide sites and leaves the Tc1/mariner characteristic footprint, i.e., duplication of the TA, upon excision. The nonviral plasmid vector contains the transgene that is flanked by IR/DR sequences, which act as the binding sites for the transposase. The catalytically active transposase may be expressed from a separate (trans) or same (cis) plasmid system. The transposase binds to the IR/DRs, catalyzes the excision of the flanked transgene, and mediates its integration into the target host genome. Although increasing lengths of DNA between the IR/DRs decrease the efficiency of Sleeping Beauty transposition, decreasing the distance outside the region of the transposon within a plasmid tends to increase transposition frequency [82]. Therefore, Sleeping Beauty may be able to carry long DNA sequences into a host genome.
|
Sleeping Beauty has also been used to accomplish stable chromosomal integration of functioning genes in somatic cells of adult mice [86, 87]. Stable expression of therapeutic levels of factor IX in an adult mouse model of hemophilia has been achieved using a trans system consisting of a factor-IX-expressing transposon on the donor plasmid and Sleeping Beauty transposase on the helper plasmid [87]. Tail vein hydrodynamic injection of the naked plasmids resulted in factor IX plasma levels that significantly decreased clotting times, whereas no factor IX was detected when a mutant transposase helper plasmid was injected. The stable chromosomal integration of this transposon was demonstrated by the maintenance of factor IX expression following partial hepatectomy [87]. We are currently attempting to achieve stable expression of therapeutic genes in hematopoietic stem cells isolated from human umbilical cord blood, using a cis plasmid delivery system in which the Sleeping Beauty transposase expression cassette is outside the IR/DRs of the transposon. The use of transposons as gene therapy vectors equivocates a number of the problems encountered with viral vectors, including multiple copy integration, the requirement that cells be cycling for integration, strict limits on DNA size, insertion into hotspots where genetic expression is attenuated, antigenicity of coat proteins, and difficulties in production. Although a number of potential obstacles remain to be addressed, e.g., transposition efficiency, epigenetic effects of transposition, and efficient delivery of transposon vectors to the nuclei of cells, the use of transposons in gene therapy appears to be inevitable. In fact, its application to stem cell biology may be significant.
|
CONCLUSIONS AND PERSPECTIVES |
|---|
|
TopAbstractIntroductionGene Correction StrategiesSleeping Beauty Transposon...Conclusions and PerspectivesReferences |
|---|
The sources of these discrepancies will become more apparent as we begin to define the pathways and essential factors involved in gene targeting and repair. It is interesting that chimeraplasty would work at one site but not at another <100 nucleotides away [89]. However, the precise mismatch and location within the genome appear to impact the efficiency of recognition and alteration of the DNA by endogenous DNA repair pathways [19, 20, 49]. Despite the controversy, chimeraplasty has been highly successful for certain cell types and gene targets. Additional studies are necessary to elucidate the proteins and pathways involved in the process, and to evaluate what aspects can be manipulated in order to improve the success of gene repair.
A major concern with any in situ repair is the risk of creating random mutations within the genome. Currently, there are little data directly implicating the gene repair technologies in this process. In fact, efforts to detect mutations with chimeraplasty have been futile. Based on the reported rates of nonhomologous recombination with HR, it may be a much greater concern for the bifunctional and SFHR molecules [6]. Exhaustive genomic analyses using state-of-the-art technologies, including microarray, will be required to identify direct mutations as well as epigenetic changes.
To improve the efficiency of gene targeting, modified nucleotides, such as peptide or locked nucleic acids, may enhance resistance of molecules to nuclease degradation. Translocating peptides such as the HIV TAT peptide, herpes simplex VP22, and antenopedia peptides may improve delivery of molecules to the nucleus by bypassing endosomal compartments and avoiding nuclease degradation [91]. In addition, these peptides appear to mediate transport of proteins across the blood-brain barrier, thereby providing a potential delivery vehicle to the brain [92]. Proteins such as RecA or spermidine may provide relatively simple solutions to the daunting challenge of homologous targeting and strand invasion.
As we begin to dissect the pathways of repair, it may be possible to manipulate cells to upregulate or recruit the essential components required to affect gene correction. For example, alkylating agents have been shown to increase translocation of DNA repair proteins from the cytoplasm to the nucleus [93]. The short half-life of the targeting molecules could potentially be a disadvantage if a certain level of correction is required. Thus, transfection of plasmids expressing constant levels of the targeting molecule could be an advantage for gene repair by single-stranded oligonucleotides and SFHR [94]. The benefits of overexpression, however, must be considered against the possibility for random mutagenesis.
While the notion of a universal nonviral delivery system is still considered possible, delivery systems will most likely be designed for individual cell types. However, attempts are under way to synthesize such unique systems as virasomes, which combine viral proteins to aid cell entry, endosomal escape, and nuclear localization, with liposomes as delivery vehicles [95]. The virasomes may be more effective transfection agents and less immunogenic than viral vectors, with a broad potential application to stem cells.
The role of Sleeping Beauty in gene therapy may ultimately depend on improved levels of transfection as well as the ability to provide precise genomic integration avoiding insertional mutagenesis. A combination of TFOs with Sleeping Beauty technology may someday enable site-specific alterations, without potential mutagenesis. However, transposon plasmid systems may be limited by size thus preventing the incorporation of vital regulatory sequences. Thus, it remains to be determined if the transposon/transposase system is functional in larger constructs, such as cosmids or YACs, which could include these regulatory elements.
A great potential for gene alterations will undoubtedly be in stem cells. While precise mutations have been made in stem cells by homologous recombination, there is a paucity of data demonstrating high rates of repair with the novel techniques outlined in this review. Stem cells could be harvested and transduced ex vivo, and the corrected cells reintroduced into the host. Growth advantage of the corrected stem cells could easily translate into a significant improvement and possible cure for diseases [96]. With the identification of specific and unique receptors on the cell surface, it is conceivable that stem cells could be selectively transfected in vivo by receptor-mediated endocytosis. These autologous approaches would be applicable to the hemaglobinopathies, such as sickle cell disease and the thalassemias, and would have major advantages over allogeneic bone marrow transplants.
Gene conversion in stem cells might involve the alteration of single nucleotide polymorphisms for treatment of certain malignancies. For example, such alterations in hematopoetic stem cells could allow higher levels of chemotherapeutic agents to be used in the treatment of specific cancers. The plasticity of stem cells to transdifferentiate into cells of seemingly unrelated lineage [97, 98], coupled with efficient gene correction, offers an opportunity to treat many different disorders without the need to harvest cells for ex vivo repair. The possibility of correcting genetic defects in hematopoietic stem cells and then invoking their transdifferentiation into liver or nerve tissue may soon be a feasible approach in treating diseases.
Finally, targeting genes for mutation within embryonic stem cells by homologous recombination has been a powerful research tool in developing animal models of human disease. Homologous recombination is characterized by low mutational efficiency and random mutational events. Targeting with small molecules, such as chimeraplasts, may have an advantage over homologous recombination in reducing random mutations and generating subtle but precise changes to the genome sequence.
Despite the potential of stem cell gene repair, there are many diseases in which considerable damage occurs in utero, such as Gaucher's disease, Tay-Sachs, and some of the leukodystrophies. While bone marrow transplantation postnatally has been effective in reducing disease progression, it cannot reverse much of the damage already done to the nervous system. The advent of in utero gene therapy may offer substantial opportunities for prevention of damage and long-term cure of some of these disorders. In fact, with the increased permeability of the blood-brain barrier in utero, it may be possible to "correct" certain neurodegenerative disorders even before birth, such as Huntington's disease.
In utero therapy may involve the direct injection of the fetus or the ex vivo manipulation of fetal hematopoietic stem cells [99]. Gene targeting may have substantial advantages over the current protocols of retroviral transduction. In fact, insertional mutagenesis of germ cells by retroviral vectors may be a much greater concern with in utero therapy. In contrast, gene correction would target the specific area of mutation and avoid both somatic and germline mutagenesis. The site specificity of targeting would also prevent any potential harm to the mother, as one might expect with a transfer of retroviral vectors into the maternal circulation [100]. Random insertion of DNA by retrovirus may produce profound changes in DNA methylation and may significantly alter gene expression involved in fetal development.
With improvements in delivery, targeting, and manipulation of the processes involved in base conversion, it is not beyond our ability to achieve the potential of gene correction. The aims are clear, and the potential strategies are being developed and tested. Each step we take toward the goal leads to a better understanding of the challenge and the great reward of success.
|
ACKNOWLEDGMENT |
|---|
|
TopAbstractIntroductionGene Correction StrategiesSleeping Beauty Transposon...Conclusions and PerspectivesReferences |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionGene Correction StrategiesSleeping Beauty Transposon...Conclusions and PerspectivesReferences |
|---|
F508 CFTR in normal mouse lung after site-specific modification of CFTR sequences by SFHR. Gene Ther
2001;8:961–965.[CrossRef][Medline]
