A multifunctional PEI-based cationic polyplex for enhanced ... - Nature

Gene Therapy (2006) 13, 1512–1523 & 2006 Nature Publishing Group All rights reserved 0969-7128/06 $30.00 www.nature.com/gt

ORIGINAL ARTICLE

A multifunctional PEI-based cationic polyplex for enhanced systemic p53-mediated gene therapy S Moffatt, S Wiehle and RJ Cristiano Department of Genitourinary Medical Oncology, University of Texas MD Anderson Cancer Center, Houston, TX, USA

We recently reported a novel coupling strategy involving salicylhydroxamic acid and phenyl(di)boronic acid molecules to attach the CNGRC peptide to PEI/DNA for CD13 targeting in tumors. Here, we doubly coupled Simian Virus (SV) 40 peptide-(nuclear localization signal)) and oligonucleotidebased (DNA nuclear targeting signal) nuclear signals to the same vector using peptide nucleic acid chemistry. This vector, CNGRC/PEG/PEI/DNA-bgal/NLS/DNTS, was predominantly localized in the cell nucleus, yielding about 200-fold higher bgal gene expression in vitro, more than 20-fold increase in tumorspecific gene delivery, and a robust bgal gene expression as demonstrated in stained tumor sections. For gene therapy purposes, we further engineered a similar targeting polyplex, CNGRC/PEG/PEI/DNA-p53/NLS/DNTS, with EBV-based epi-

somal vector for sustained p53 gene expression. A distribution of vector DNA and apoptosis in p53-containing tumors was observed, yielding a significant tumor regression and 95% animal survival after 60 days. This multicomponent vector also co-targeted tumor and tumor-associated endothelial cells but not normal cells, and had more efficient therapeutic index than each vector administered as a single modality. The use of an efficient coupling strategy without compromising the vector’s integrity for DNA condensation and endosomal escape; nuclear import; tumor-specific and persistent p53 gene expression clearly provides a basis for developing a single combinatorial approach for non-viral gene therapy. Gene Therapy (2006) 13, 1512–1523. doi:10.1038/ sj.gt.3302773; published online 8 June 2006

Keywords: polyethylenimine (PEI); salicylhydroxamic acid (SHA); phenyl(di)boronic acid (PDBA); CNGRC peptide; nuclear localization signal (NLS); peptide nucleic acid (PNA); p53-mediated gene therapy

Introduction In light of the major obstacles plaguing non-viral therapeutic gene delivery, we recently developed an effective formulation between polyethylenimine (PEI) and plasmid DNA,1 carrying a tumor-specific CNGRC peptide for targeting aminopeptidase N/CD13 in the tumor vasculature.2 We modified the 25 kDa PEI by substituting 5% of the amine molecules with salicyldroxamic acid (SHA) to enable a self-assembly between the SHAmodified PEI and phenyl(di)boronic acid (PDBA)modified CNGRC peptide. Whereas gene expression with this targeting vector was impressive, we envisaged that using a p53-containing episomal vector to achieve a long-term in vivo gene expression,3 and further incorporating nuclear signals in the tumor-targeted PEI-based vector, would generate a more effective system for maximum and sustained therapeutic gene delivery. Non-viral gene transfer vectors, consisting mainly of plasmids, are generally delivered as a complex with chemical and/or biochemical vectors such as cationic lipids or polymers (polyplexes). Such complexes offer several advantages compared to viral vectors such as better safety profile, a lower immunogenicity, sizeCorrespondence: Dr S Moffatt, Department of Genitourinary Medical Oncology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030, USA. E-mail: [email protected] Received 14 December 2005; revised 6 March 2006; accepted 7 March 2006; published online 8 June 2006

independent transgene inclusion, and may also offer less stringent pharmaceutical regulations; however, they generally suffer from a relatively low transfection efficiency. The exact mechanism of non-viral gene transfer is not perfectly understood, but nuclear transport of oligonucleotides and plasmid DNA (pDNA) either by transfection of whole cells4–8 microinjection into the cytoplasm9–14 or transfection of plasma membranepermeabilized cells15–18 has been reported in recent years to highlight impediments to efficient delivery of nonviral vectors, including the passage of DNA across the cellular membrane, the escape from endosomal compartments, low metabolic stability of plasmids and polyplexes in the cytoplasm, and the inefficient trafficking of genes to the nucleus.19 By comparison, cellular and viral proteins enter the nucleus efficiently by means of nuclear localization sequences (NLS), which are stretches of amino acids that bind to intracellular transport receptors to facilitate transfer through the nuclear pore.20 Even though the application of NLS peptides for non-viral gene transfer has been somewhat investigated, some of the results are promising,4,6,17 whereas others are discouraging.14,21 There are, however, a couple of important findings from the literature regarding nuclear targeting. For example, it is now known that the nuclear transport capabilities of the Simian Virus (SV) 40 large T NLS is abrogated with a lysine to threonine mutation in the signal peptide.22 It has also become increasingly clear from studies above and numerous others that one of the major limitations in the use of NLS-conjugated DNA or peptides for

p53-mediated gene therapy with a multifunctional polyplex S Moffatt et al

increasing gene expression efficiencies lies in the coupling method and the type of chemistry used. For instance, coupling of streptavidin-NLS conjugates to biotinylated DNA enabled linear DNA fragments to enter the nuclear of digitonin-permealized cells by an active transport process.23 Increases in gene expression has also been observed following ligation of an oligonucleotide-NLS conjugate to one or both terminal to a linear DNA molecule.6 Conjugation of the NLS peptide to the N3-position of adenine bases in double-stranded DNA yielded no increase in the nuclear uptake of the modified DNA or transgene expression.17 In contrast, a follow-up report describing the synthesis a loopforming oligodeoxyribonucleotide crosslinked with the NLS peptide and this modified conjugate was enzymatically ligated to double-stranded DNA and transfected with PEI to enhance gene expression from 10- to a 100-fold.6 Clearly, an NLS-conjugate that is easy to synthesize, with a chemistry that does not interfere with gene expression and also extremely stable in biological systems is desired. With the above limiting factors in mind towards a multifunctional approach for gene therapy, we have utilized the peptide nucleic acid (PNA) to link the SV40 NLS to the plasmid DNA. Peptide nucleic acid is a DNA mimic with a pseudo-peptide backbone composed of N-(2-aminoethyl) glycine units to which the nucleobases are linked through methylene carbonyl linkers, resulting in an uncharged mimick.24 It is chemically stable, and thus not expected to be degraded inside the living cell. Peptide nucleic acids are capable of sequence-specific recognition and exhibit a capacity to hybridize with high affinity to complementary sequences of RNA and DNA, obeying the Watson–Crick hydrogen bonding scheme, with extraordinary thermal stability and unique ionic strength effects,24–26 exhibit high levels of solubility, and are resistant to proteinases and exonucleases.26–28 Peptide nucleic acids also recognize duplex homopurine sequences of DNA to which they bind by strand invasion, forming a stable PNA–DNA–PNA triplex with a looped out DNA strand.26 Polypyrimidine PNAs may also invade target DNA duplexes by forming a four-stranded complex consisting of a PNA–PNA–DNA triplex and a displaced DNA strand,24–26 and in this situation, subsequent experiments have revealed that strand invasion can be improved by use of bis-homopyrimidine PNAs.24,26,28 In this manuscript, we showed that the integrity of PNA–NLS binding to DNA encoding p53 protein could still be preserved even after further coupling to the CNGRC ligand using PEI as the cationic polymer. The ultimate goal of the studies presented here is to successfully generate and utilize a multifunctional PEI/ DNA vector using an efficient and stable coupling strategy that will significantly increase gene delivery to tumors while decreasing any background expression in other organs. Using a well-defined stoichiometry, we demonstrate that the incorporation of the tumor-targeting CNGRC peptide as well as the two SV40 NLS signals in a single PEI/DNA vector, by way of PDBA–SHA bridge and PNA, achieves this purpose by increasing vector specificity both in vitro and in vivo, and even more importantly, in the context of a p53-containing EBV-based episomal vector enhances long-term gene expression and significantly increases the overall therapeutic index and survivability of the animals. The successful design and

utilization of this multifunctional vector for p53-mediated gene therapy would certainly be an important step in the use of non-viral vectors for cancer gene therapy.

1513

Results Transfection of polyethylenimine/DNA-bgal with CNGRC- and/or nuclear localization signal-containing polyplexes To construct a multi-factorial vector capable of overcoming essential barriers to gene delivery, we first generated the pGeneGripbgal reporter plasmid that encodes bgal, as well as the SV40 DNA nuclear targeting signal (DNTS) (Figure 1a). The presence of a PNAbinding site in the plasmid facilitates the hybridization of the PNA-modified NLS. For targeted delivery to H1299 cells (since we have already shown previously that H1299 cells express CD13, the receptor for the CNGRC peptide), we further coupled the CNGRC peptide to the condensed PEI/DNA formulation by making use of the PDBA-SHA bridge as described previously.2 Significantly (Po0.03) higher levels of gene expression than the PEI/DNA-bgal alone was observed in vitro when the vector contained either NLS and/or DNTS (Figure 1b). This increased kinetically and seemed to act synergistically when both nuclear signals were in the same formulation. An even higher level of gene expression was observed when the DNTS- and NLS-containing vector was targeted with CNGRC peptide, resulting in approximately 500-fold higher gene expression than the vector without the CNGRC peptide, and 250-fold higher when the vector contained CNGRC peptide at 72 h after transfection. The DNTS-containing vectors seemed to yield higher expression levels than the NLS-containing vectors regardless of the vector components, indicating that perhaps the DNTS was a better targeting signal than the NLS peptide. On the whole, there was a general trend of increasing transfection levels towards the 72 h time point. The level of gene expression using bgal DNA alone or PEI/DNA-bgal as base references was negligible compared to the polyplexes with either the CNGRC or the NLS signals. To attribute the specificity of the vector to the presence of the CNGRC and NLS signals, competitive studies were done. The addition of a 100fold molar excess of the CNGRC or NLS peptide significantly reduced the level of transfection (Po0.02) up to 80 and 75%, respectively, in tumor cells as compared to transfection using just the targeting vector without competition (Figure 1c), but was not affected by competition with excess non-specific peptide (CARAC) or mutant NLS. However, a CD13-blocking monoclonal antibody (mAb) (WM15) significantly reduced (Po0.02) the level of transfection by approximately 70%, indicating its authenticity as the receptor for the CNGRC peptide. In contrast, transfection efficiencies under the same conditions with SHA-PEI/DNA-bgal were neither influenced by competition with excess CNGRC peptides, excess NLS peptides nor the CD13 antibody. Analysis of tumor targeting after intravenous administration of the polyplexes Based on our previous studies that the CNGRC peptide can mediate specific delivery to tumors in vivo, we Gene Therapy


1514

sought to determine if incorporation of the two NLS signals into the vector could significantly increase gene expression in the tumors after i.v. administration of the CNGRC/PEG/PEI/DNA-bgal/DNTS/NLS targeting

vector to mice bearing subcutaneous H1299 tumors. Analysis of bgal expression in tumors from these animals demonstrated more than 20 times higher gene expression over levels obtained in tumors from animals injected with just the CNGRC/PEG/PEI/DNA-bgal vector at 72 h (Figure 2a). Furthermore, whereas no significant difference in gene expression was detected between CNGRC/PEG/PEI/DNA-bgal/NLS- and CNGRC/PEG/ PEI/DNA-bgal/DNTS-injected mice, there were approximately 10- and 15-fold increases in gene expression, respectively, over the CNGRC/PEG/PEI/DNA-bgal vector. These results indicate that the two NLS signals can mediate more specific and efficient in vivo gene expression when attached to the CNGRC peptide in our novel vector formulation. Consequently, we proceeded to determine the ability of the vector to mediate targeted delivery in vivo by examining bgal transfection in tumor sections of injected mice. Even with this rather insensitive bgal staining assay (approximately 1000 molecules of bgal are required to turn one cell blue), we further corroborated our in vivo transfection results (Figure 2b). The data from the tumor sections further indicate that the vectors containing the nuclear signals were more efficient in transfecting distant tumors, resulting in robust gene expression when injected intravenously. There seemed to be no significant difference between the number of blue cells from CNGRC/PEG/PEI/DNAbgal/NLS- and CNGRC/PEG/PEI/DNA-bgal/DNTSinjected mice. Tumor sections from both mice however yielded less blue cells than sections from CNGRC/PEG/ PEI/DNA-bgal/NLS/DNTS-injected mice. Furthermore, the number of blue cells from the control PEI/DNA-bgal vector was negligible (data not shown). In addition to this, it was equally important to validate the targeting of CD13 using the CNGRC-containing vectors in the tumors. Evaluation of tumor sections by indirect immunohistochemistry using anti-CD13 mAb (WM15) showed that besides tumor cells, a significant number of stained cells in tumor stroma corresponds to endothelial cells of tumor-associated vessels (Figure 2c), as suggested by similar staining patterns obtained with an anti-CD31 mAb, a well-known marker of endothelial cells (data not shown). There was no clearly visible staining of endothelial cells on tumor sections from mice injected with PEI/DNA-bgal control vector. These results possibly suggest a co-targeting mechanism by the CNGRC-containing vectors.

Figure 1 (a) Map of plasmid pGeneGripbgal-DNA nuclear targeting signal (DNTS) showing Simian Virus (SV) 40 peptide DNTS and the peptide nucleic acid-binding site and hybridization to the modified SV40 nuclear localization signal peptide. (b). In vitro kinetic analysis of polyplex gene expression. H1299 cells were incubated with bgal DNA, PEI/DNA-bgal, and a combination of PEI/DNA-bgal with or without CNGRC peptide, DNTS, or NLS. bgal quantitation was performed 24 h (black), 48 h (medium gray), and 72 h (dark gray) later using a Tropix b-galactosidase detection kit and values were normalized per mg protein. Values are means7s.d. of three independent experiments. (c) Competitive inhibition assays. H1299 cells were incubated with either salicylhydroxamic acid-PEI/DNA-bgal or CNGRC/PEG/PEI/DNA-bgal/ NLS/DNTS plus or minus 100-fold excess free CNGRC peptide or a control non-specific peptide (CARAC), NLS or a mutant NLS peptide as well as a CD13-blocking mAb (WM15) and bgal analysis was analyzed 72 h later. Gene Therapy


In vitro and systemic transfection of polyethylenimine/ DNA-p53 vector with CNGRC and/or nuclear-specific signals We next tested the efficacy of the tumor- and nuclearspecific targeting vector in the context of p53-mediated gene therapy. For this purpose, we utilized an episomal vector to achieve even longer term gene expression

(Figure 3a). Using this p53-containing targeting vector, efficient cell killing was observed in H1299 cells after transfection (Figure 3b). Cell killing with PEI/DNA-bgal or PEI/DNA-p53 was comparatively insignificant and unlike the previous transfection results, there was no significant difference in cell killing between the NLS- and DNTS-containing vectors in vitro, with the most effective being CNGRC/PEG/PEI-DNA-p53/NLS/DNTS. To determine whether increase in gene delivery was due whole or in part to an increase in the transfer of an NLS-conjugated DNA to the nucleus, we examined in confocal studies the relationship between nuclear localization of CNGRC-coupled, p53-containing targeting vector and gene expression. Nuclear localization was evident in cells transfected with p53-containing vectors 72 h after transfection (Figure 4b–d) but not in cells transfected with the vector without NLS (Figure 4a and e). The seemingly differential increase in nuclear localization among the various polyplexes also corroborated our findings with bgal gene transfection in vitro, suggesting that higher nuclear accumulation of the vector was synonymous with increased bgal gene expression. There was no nuclear localization with mutant NLS peptide, which is nuclear transport deficient (data not shown).

1515

In vivo effect of p53-containing targeting vectors on tumor growth and mice survival To investigate the effect of p53 on tumor growth, we initially set up five groups of mice injected with different amounts of p53 DNA in the CNGRC/PEG/PEI-DNAp53/NLS/DNTS targeting vector. In this experiment, it was observed that 6 mg of p53 DNA was the most effective in reducing the already established tumor volumes in a 72-h time period followed by 10, 3, 1 mg, and empty vector (minus p53 DNA) in that order (Figure 5a), so this amount of p53 DNA was subsequently used in all therapeutic experiments. This targeting vector was still the most effective in significantly reducing established tumor volumes followed by

Figure 2 In vivo evaluation of gene expression. (a) Analysis of gene expression in tumors. Mice were injected i.v. with PEI/DNA-bgalor CNGRC-containing vectors with or without nuclear targeting signals and tumors were harvested after 72 h. bgal gene expression in tumors was quantitated with a Tropix b-galactosidase detection kit and values were normalized per mg protein. Values are means7s.d. of three independent experiments. (b) bgal gene expression in tumor sections. Tumor sections (12 mm) from mice (n ¼ 5) injected with CNGRC/PEG/PEI/DNA-bgal, CNGRC/PEG/ PEI/DNA-bgal/NLS, CNGRC/PEG/PEI/DNA-bgal/DNTS, and CNGRC/PEG/PEI/DNA-bgal/NLS/DNTS were fixed 72 h after the first injection in acetone:ethanol, stained for X-gal for 24 h at 371C and counterstained with Nuclear Fast Red solution. A total of four fields were examined for each vector with representative fields shown. (c) CD13 targeting by CNGRC-containing polyplexes in vivo. Paraffin-embedded tumor sections (8 mm) from mice (n ¼ 5) injected with PEI/DNA-bgal, CNGRC/PEG/PEI/DNA-bgal/NLS, CNGRC/PEG/PEI/DNA-bgal/DNTS, or CNGRC/PEG/PEI/ DNA-bgal/NLS/DNTS were incubated with WM15 CD13 mAb followed by biotinylated horse anti-mouse IgG. Slides were overlaid with 3,3-diamino-benzidine-tetrahydrochloride in deionized water containing 0.03% hydrogen peroxide, counterstained with Harris’ hematoxylin before microscopy. A total of four fields were examined for each vector with representative fields shown. Tumor endothelial cells (red arrows); tumor cells (black arrows). Gene Therapy


1516

Figure 4 In vitro nuclear localization with targeting vectors in H1299 cells. (a) PEI/DNA-p53, (b) CNGRC/PEG/PEI/DNA-p53/ NLS, (c) CNGRC/PEG/PEI/DNA-p53/DNTS, (d) CNGRC/PEG/ PEI/DNA-p53/NLS/DNTS, or (e) CNGRC/PEG/PEI/DNA-p53 were pre-labeled with SYTOX green-PNA conjugates for 30 min and allowed to incubate with methylene blue-stained H1299 cells for thirty more minutes at 371C before confocal microscopy. The lower panels (f), (g), (h), (i), and (j) are the corresponding differential interference contrast (DIC) for the top fields. A total of four fields were examined for each time point per conjugate with a representative field shown.

expressed p53 with the exception of the p53-minus vector (Figure 5d).

Figure 3 (a) Map of plasmid pCEP4/NTS/p53 episomal plasmid showing SV 40 DNA nuclear targeting signal (DNTS) and the peptide nucleic acid-binding site and hybridization to the modified SV40 nuclear localization signal peptide. (b) Inhibition of H1299 non-small-cell lung cancer cell growth by targeting vectors with or without p53. Cell number was measured spectrophotometrically at different time points after crystal violet staining. Values are means7s.d. of three independent experiments.

CNGRC/PEG/PEI-DNA-p53/DNTS, CNGRC/PEG/PEIDNA-p53/NLS, CNGRC/PEG/PEI-DNA-p53, PEI-DNAp53, and PEI/DNA-bgal in that order (Figure 5b) over an extended period of 28 days. Significant differences in the mean volume (7s.e., Po0.02) between the CNGRC/ PEG/PEI-DNA-p53/NLS/DNTS- and PEI/DNA-p53-injected mice was observed 14 days after tumors had ‘taken’ and the ratio of tumor volumes at this time was 8.2–14.1 (mean, 12.6). Terminal deoxynucleotidyl transferase-mediated uridine 50 -triphosphate nick-end labeling (TUNEL) assay in tumor sections of mice injected with these vectors revealed apoptotic foci in all vectors with, but not those without p53 (Figure 5c). To further validate these data as resulting from p53, we screened the tumors for the presence or absence of p53 DNA. All tumors from mice injected with p53-containing vectors Gene Therapy

Effect of p53-containing targeting episomal vectors on mice survival We next proceeded to ask if the dramatic decrease in tumor volumes as a result of the p53-containing targeting polyplexes would lead to an increase in animal survival. For this experiment, mice were randomized into six groups, 12 animals per group. At 14 days after initial cell injection, mice were systemically given single doses of the various episomal vector-based targeted formulations with or without CNGRC, NLS, DNTS, or p53 to initiate therapy. The experiment was allowed to run for an extended period of 60 days (tumors still expressed p53 with the episomal plasmid, data not shown) to better validate the full effect of the targeted polyplexes on survival. At the end of this experiment, there was a mean survival rate of 92% for animals injected with the CNGRC/PEG/PEI/DNA-p53/NLS/DNTS vector as opposed to only 20% with the untargeted PEI/DNA-bgal vector (Figure 6). The DNTS-containing targeting vector was again more efficient at prolonging the mean survival of the animals (9075%) than the NLS-containing vector (7474%).

Discussion With only 0.1% of naked DNA or 1% of polyplex DNA reaching the nucleus following microinjection of nonviral vectors into the cytoplasm,29 it has become imperative to address this issue in the context of gene therapy. Whereas the results of targeted delivery have shown some promise,30–32 most of them have been both


1517

150

b 600 6 µg 10 µg 3 µg 1 µg vehicle alone

CNGRC/PEG/PEI/DNA-p53/NLS/DNTS CNGRC/PEG/PEI/DNA-p53 -DNTS CNGRC/PEG/PEI/DNA-p53/NLS CNGRC/PEG/PEI/DNA-p53 PEI/DNA- p53 PEI/DNA-βgal

500

Tumor Volume (mm3)

Tumor Volume (mm3)

a 200

100

50

400 300 200

**

day 14

* * *

100 0

0 0

12

18

24

36

48

72

0

3 5 7

Hours after initial injection

c

9 11 13 15 17 19 21 28

Days after initial cell injection

PEI/DNA-βgal

PEI/DNA-p53

CNGRC/PEG/PEI/DNA-p53

CNGRC/PEG/PEI/DNA-p53/NLS

CNGRC/PEG/PEI/DNA-p53/DNTS

CNGRC/PEG/PEI/DNA-p53/NLS/DNTS

d

p53

S LS

LS

/D

/D

NT

NT

S /P

EG /

PE

I/D

NA -/N

3/ N GR C CN

CN

GR C/ PE

PE G/ PE GR C/

CN

I/D NA -p 5

I/D NA -

NA

p5

-p

53

3/D

/N

NT

LS

53 -p I/D

G/ PE

PE I/D NA PE GR C/ CN

S

β–actin

Figure 5 (a) Dose-dependent effect upon H1299 subcutaneous tumor volumes following injections with vectors containing 1, 3, 6, or 10 mg of p53 DNA or empty vector (vehicle alone) harvested at different time points. (b) Effect of targeting and control polyplexes on tumor volumes. Mice (n ¼ 5) were injected with polyplexes with or without CNGRC, p53, or nuclear targeting signals and H1299 subcutaneous tumor volumes were measured. Injections began 14 days (taken as day 0) after initial cell injection allowing for further establishment of primary tumor. *Po0.01, **Po0.005 versus untargeing controls. (c) TUNEL assay of tumor sections. Mice injected with or without p53 and/or nuclear targeting signals. Tumors were extirpated 72 h after injection. Many apoptotic cells were detected in tumor sections of mice injected with CNGRC/PEG/PEI/DNA-p53/NLS/DNTS (black arrows). (d) Transgene expression of p53 gene. Genomic DNA was extracted from xenografts and PCR was performed with p53- and b-actin-specific primers. b-actin is ubiquitously expressed, although p53 was detected only in tumors of mice injected with vectors containing p53 72 h after initial cell injection.

non-reproducible and conflicting. While we and others have made attempts to target exclusive cells using nonviral gene delivery vectors in the form of a molecular conjugate,1,2,33 it has become apparent that rather than addressing the current hurdles independently, one

attractive strategy would be a single vector having multiple factors for both tumor-targeted and nuclearspecific delivery. The goal of the present work was therefore to validate the hypothesis that p53 gene expression and therapeutic index could be significantly Gene Therapy


1518 120 100

% Survival

80 60 40 20 0 0

3

7 14 30 45 Days after tum or establishment

60

C NGRC/PEG/PEI/DNA-p53/NLS/DNTS C NGRC/PEG/PEI/DNA-p53/DNTS C NGRC/PEG/PEI/DNA-p53/NLS C NGRC/PEG/PEI/DNA-p53 PEI/DNA-p53 PEI/DNA-βgal

Figure 6 Survival of mice treated intravenous with single doses of vector with or without p53 and nuclear targeting signals. Xenografts were allowed to grow for 14 days. Untargeting controls were PEI/DNA-bgal or PEI/DNA-bgal-p53. Mice (n ¼ 12) in six different groups were followed for survival for 60 days. The remaining living mice in the CNGRC/PEG/PEI/DNA-bgal/NLS/DNTS were without any significantly gross or microscopic evidence of disease.

enhanced in a multifunctional setting over our previously reported tumor-specific vector in the context of an episomal vector-based peptide/PEI/DNA gene delivery system. The CNGRC peptide, which was previously isolated by in vivo phage display,34,35 was chosen in this study as the targeting ligand because of its high specificity for CD13 which is expressed at extremely high levels (41 million copies/cell) on tumor cells, as well as endothelial cells of angiogenic tumor vasculature but not normal vasculature.12 We and others2,34–36 have already demonstrated the efficiency and specificity of this peptide in mediating gene delivery in vivo. The SV40 large T antigen NLS peptide as well as its DNTS sequence (an enhancer element from SV40), which has also been shown to be absolutely essential for nuclear entry,13,32 were further incorporated in the delivery vector. Owing to the overwhelming advantages of PNA, including its high affinity binding to its complementary sequences, solubility, resistance to DNAses, and stability, we linked the NLS peptide to the PNA site in the plasmid DNA as a PNA–NLS hybrid. It must however be pointed out that, in order not to affect the functionality of the NLS peptide, the accurate quantitative determination of the sulfhydryl groups in the reduced peptide, together with the precise molar concentrations of both PNA and NLS, was important in determining the extent of saturation of the PNA-binding sites in the plasmid. One important consideration in the entire study was to fully characterize and purify the NLS peptide-PNA/DNA conjugates prior to in vitro and in vivo evaluation in order to avoid unconjugated NLSs competing for binding for transport receptors on the nuclear membrane. Nuclear localization signal peptides were therefore attached to PNA/DNA Gene Therapy

conjugates in the presence of high salt conditions to avoid non-specific interactions, and unreacted peptides removed by ethanol precipitation of the DNA. A prelabeled PNA assay to detect nuclear translocation of NLS peptide-PNA/DNA conjugates indicated that at least seven binding sites were required for efficient PNA binding (data not shown). First, in vitro evaluation of transfection efficiencies using our multi-factorial bgal targeting vector showed a significantly enhanced gene expression over the control PEI/DNA-bgal vector with efficiencies up to 80%. Competition studies confirmed the CNGRC- and NLS-specific mediation in the enhanced gene delivery. One added advantage in this system is the use of PEI in our vector. Unlike polylysine, PEI can facilitate endosomal release and mediate efficient cell transfection in vitro without the use of an endosome lytic agent.37 Next, we examined the relationship between the nuclear localization properties of our conjugates and gene expression in confocal studies. The basic tenet of this type of mechanistic confocal studies is that introduction of fluorescent labels does not substantially impair either the structural integrity or the transfection characteristics of the nonviral vector system so that conclusions about the intracellular trafficking can be extrapolated to the behavior of unlabelled components.38 The results from this assay demonstrated that plasmid DNA alone with the CNGRC targeting ligand does not confer nuclear localization functions, whereas nuclear transfer occurs in the presence of NLS peptide, DNTS, or both. In addition, both the NLS peptide and the DNTS sequence appear to act synergistically for nuclear transfer of plasmid DNA resulting in increased gene expression. The increase in gene expression as a result of nuclear localization corroborated our in vitro bgal quantitation results, suggesting that nuclear localization was synonymous with gene expression. We currently do not know the nuclear localization properties of the CNGRC ligand, but increased expression of the NLS conjugates in the presence of the ligand may possibly support the hypothesis that the ligand contributes somewhat to the nuclear targeting properties of the NLS conjugates. An explanation for this hypothesis may be that a substantial amount of NLS-conjugated DNA in the nucleus modifies the characteristics of the nucleus, which in turn favors the nuclear penetration of the CNGRC-coupled DNA. This hypothesis however remains to be proven. Integral to the nuclear transport studies, a more important study was to test the efficacy and specificity of gene expression when this vector is administered systemically, and compared with vectors with either the CNGRC, NLS, or DNTS alone. CNGRC/PEI/DNAbgal/NLS/DNTS-injected mice showed more than 20fold increase in bgal gene expression over just the CNGRC/PEI/DNA-bgal alone upon i.v. administration after 72 h. Furthermore, there was 10-fold increase in bgal gene expression of CNGRC/PEI/DNA-bgal/NLS/ DNTS over CNGRC/PEI/DNA-bgal/NLS-injected mice as opposed to approximately twofold over CNGRC/ PEI/DNA-bgal/DNTS-injected mice. The reason for this disparity in fold increases is not clear but it shows that the SV-40 DNTS enhancer sequence may be more efficient at nuclear targeting than its corresponding NLS peptide. One possible explanation may be that the oligonucleotide-based DNTS may be less refractory in


circulation and experience less aggregation with serum and plasma proteins, hence maintaining its functionality as opposed to the peptide-based NLS. Also taking into consideration the fact that approximately 1000 molecules of bgal are required to turn a single cell blue, which denotes a very insensitive staining assay, the demonstration of a robust bgal staining in tumor sections from mice injected with formulations containing both NLS motifs and the CNGRC peptide in the present study provides further empirical evidence that a single multi-factorial vector of this type would be a potentially suitable remedy to address the issue of gene delivery. In this study, the levels of in vivo gene expression with the complete multifunctional targeting vector was about 3 logs lower than that obtained in vitro. While this difference is not surprising due to the overwhelming barriers to gene delivery in animal models, by comparison, the levels of gene expression achieved in this study are up to 2 logs higher than yields obtained with other targeting vectors in vivo,30–32 making our strategy a very attractive potential for systemic gene delivery. We also noted that two serial administrations of the targeting vector resulted in similar levels of gene expression as observed after a single delivery (data not shown), suggesting that the formulations can be re-administered without apparent loss of efficacy. However, it is worth cautioning that this picture may be altered once non-viral vectors are conjugated with peptides or other molecules able to elicit an immune response. This has already been demonstrated in previous reports39,40 where DNA complexed with a molecular conjugate based on Fab fragments of antibodies to the Ig receptor on epithelial cells led to an escalating immune response after repeated administration to the lung in vivo. A possible remedy to this in the context of the multi-component vector described in this study may involve the use of immuno-suppressants or immunologically inert peptides/ligands; however, there are no reports to date of studies evaluating the efficacy of this type of repeated administrations. Even more importantly, we also demonstrated that the CNGRC/ PEG/PEI/DNA-bgal/NLS/DNTS vector not only targets tumor cells but also tumor-associated endothelial cells as seen by positive immunohistochemical staining of CD13 in contrast to only tumor cells in the control vector. Since both tumor and tumor endothelial cells are being targeted, the number of possible therapeutic genes that can be used is greatly increased. As a result, this would allow for the use of genes affecting either tumor cell growth or angiogenesis, or even both approaches to be used in developing anticancer gene therapies. Most tumor therapy studies with the retroviral system currently have inherent disadvantages such as low titers, viral instability, and the requirement for cell division for integration and expression.41 Also, although the adenovirus system can provide more efficient gene transfer and stability of the virus, difficulties in cell targeting and re-administration have become serious drawbacks.41 A plasmid-based molecular conjugate system in the form described here would therefore be more available and easy to construct and handle. Also, re-administration would be possible since plasmid DNA has little antigenicity.42,43 Using a p53-mediated PEI-based episomal delivery system, containing a tumor-specific ligand

1519 Ligand

Nuclear Translocation Signal

+ +- +- -

++ +

- - -

Episomal Plasmid DNA +- +-

Site-specific integrace

-

+-

Endosomal lysis or Bypass agent

- -+ ++

Therapeutic Gene Tissue Specific, Regulatable promoter

???

Figure 7 The general structure of a proposed non-viral targeting DNA delivery vector in the form of a ‘synthetic virus’ based on the use of molecular conjugates.

and nuclear localization signals to evaluate its therapeutic effect on human lung cancer xenografts in nude mice, we demonstrated that transfection of the therapeutic vector dramatically enhanced the cytotoxic effect of p53 as evidenced by both in vitro and in vivo analysis. As opposed to only 20% with the control vector, the significant reduction in tumor burden resulted in approximately 95% surviving animals when the CNGRC/PEG/PEI/DNA-p53/NLS/DNTS was administered. These mice were killed on day 60 and close to 90% were still positive for p53 in both groups by PCR. Also in this study, an optimum amount of 6 mg p53 DNA in the formulation was the most effective in reducing the tumor burden, suggesting the need for only a low threshold of DNA in the polyplex in achieving maximum effect. This report suggests that clinical application using this delivery method may be possible in the future. On the other hand, it is worth noting a few disadvantages with this method. In our study, although significant inhibition of tumor growth and prolonging of survival was observed, the targeted area was limited to local tumors. Therefore, the effect on growth of metastatic lesions cannot be empirically included in this therapy. We also did not evaluate repeated administrations more than twice, so the conclusive effect of re-administrations on tumor growth cannot really be deduced from this regimen. Nevertheless, based on these findings, we can conclude that a combination therapy involving the use of multiple genes or peptides in the same non-viral gene delivery vector is possible, opening up new options for gene therapeutics. We also envisage the incorporation of site-specific integrases in this novel PEI/DNA vector to add another level of specificity for cancer gene therapy. We hope that eventually after strict evaluation, testing, and optimization, the use of this coupling technique to develop a comprehensive multifunctional ‘synthetic virus’ (Figure 7) that has the advantages of a viral system combined with the advantages of a nonviral gene delivery vector could be of immense value for targeted therapy in the clinical setting.

Materials and methods Cell line, plasmids, and antibodies The human non-small-cell lung cancer cell line H1299 was cultured in RPMI with 5% FBS. Cells were cultured Gene Therapy


1520

in a 371C incubator with 5% CO2. The plasmid pCMVbgal contains the Escherichia coli bgal gene under the control of the cytomegalovirus enhancer and promoter. Plasmid pGeneGripbgal contains 10 tandem repeats of the PNA-binding sites and was generated by sub-cloning the the BamHI bgal fragment from pCMVbgal into the BamHI site of the original pGeneGripBlank plasmid (Gene Therapy Systems Inc., San Diego, CA, USA). Plasmid pGeneGripbgal-DNTS (Figure 1b), containing the SV 40 DNTS was generated by sub-cloning the 480 bp Eco47III–HpaI fragment from plasmid pGL3-Enhancer, containing the SV40 enhancer sequence, (Promega, Madison, WI, USA) into the unique XmnI site of plasmid pGeneGripbgal. Plasmid pCEP4/NTS/p53 was first generated by sub-cloning the 1.1 kb p53 gene into the BamH1 site of pCEP4 episomal expression plasmid (Invitrogen, Carlsbad, CA, USA), upstream of the CMV promoter resulting in pCEP4/p53, followed by cloning the 480 bp Eco47IIIHpaI DNTS fragment into the unique Nru1 site upstream of the Herpes Simplex Virus thymidine kinase promoter yielding plasmid pCEP4/NTS/p53, which also carries the Epstein–Barr virus replication origin (oriP) and nuclear antigen (encoded by the EBNA-1 gene) to permit extra-chromosomal replication in mammalian cells. The PNA site 1 (Gene Therapy Systems Inc., San Diego, CA, USA) containing 10 tandem repeats of the PNA site (101 bp) with 50 -GATC overhangs was cloned into the BglII site of pCEP4/NTS/p53 episomal plasmid downstream of the SV40 DNTS site. All plasmids were isolated by an alkaline-SDS lysis method, checked for the presence of endotoxin with a limulus amebocyte lysate endochrome-k kit (Charles River Endosafe, Charleston, SC, USA) and used for transfection studies both in vitro and in vivo. The activityblocking mAb to CD13, WM15 (Pharmingen, San Diego, CA, USA) was used at a concentration of 1 mg/ml by pretreating the cells for 30 min with the mAb in cell medium.

Polyplex vector formation and analyses The modification of the CNGRC peptide to yield CNGRC-PEG-PDBA and its subsequent linkage to SHA-PEI/DNA for the formation of the targeting polyplex has adequately been described and illustrated previously;2 The sizes and z potentials of PEI/DNA complexes were measured by photon correlation spectroscopy, using a Zeta-sizer 3000HS (Malvern, Southboro, MA, USA). The measurements were performed according to the manufacturer’s instructions with the following settings: temperature, 251C; scattering angle, 901; and analysis mode. Data were interpreted using the CONTIN software (Malvern Instruments, Malvern, UK). The results are presented as the mean7the standard deviation of at least three preparations. All conjugates contained 6 mg of bgal DNA with an N/P ratio of 2.7/1. The conjugates varied in both sizes and zeta potential as follows: SHAPEI/DNA-bgal (9876.2 nm, z +5.772.1); CNGRC/PEG/ PEI/DNA-bgal (21077.1 nm, z +4.671.5); CNGRC/ PEG/PEI/DNA-bgal/NLS (22179.1 nm, z +9.575.1); CNGRC/PEG/PEI/DNA-bgal/DNTS (20874.1 nm, z +3.871.1), and CNGRC/PEG/PEI/DNA-bgal/NLS/ DNTS (21877.1 nm, z +7.573.1). Gene Therapy

Peptide nucleic acid–nuclear localization signal coupling and binding to plasmid DNA The following peptides were utilized: peptide PKKKRKV containing wild-type SV40 large T NLS or peptide PKTKRKV (dNLS ) containing mutant SV40 large T NLS (the threonine mutant is known to be transport deficient) (Sigma, St Louis, MO, USA). To couple the SV40 NLS to PNA, we utilized the already synthesized maleimide-PNA (Gene Therapy Systems Inc., San Diego, CA, USA). First, 25 mM of the BondBreakert (Tris [2-Carboxyethylphosphine] hydrochloride) (TCEP) solution (Pierce, Rockford, IL, USA) was used to reduce the cysteine groups in the NLS peptide to sulfhydryls in order to facilitate the coupling of the activated sulfhydryls to the maleimide attached to the PNA. Briefly, 20 mg of the NLS peptide was mixed with 25 mM of TCEP, 50 mM of ethylenediaminetetraacetic acid (EDTA, Sigma) and 100 mM of NaPO4 in a total volume of 300 ml and let to stand at room temperature (RT) for 1 h. Ellman’s reagent (Pierce, Rockford, IL, USA) was then used to calculate the concentration of activated sulfhydryls in the reduced NLS sample according to the manufacturer’s instructions. The maleimide-PNA was then incubated with the reduced NLS sample for 2 h at RT. Based on the results from our preliminary in vitro coupling experiments (data not shown), we determined that a 1:1 molar ratio of maleimide-PNA to NLS was sufficient and optimum for coupling and efficient binding to the pGeneGripbgal, pGeneGripbgal-DNTS, and the pCEP4/NTS/p53 plasmids. This ratio was subsequently used in all our coupling experiments. We further determined that 0.14 pmol of NLS and 0.3 pmol of PNA which binds to seven and three PNA sites, respectively, within the pGeneGripbgal plasmid was optimum to saturate the PNA sites for effective hybridization. Coupling of the mutant NLS was carried out in a likewise manner. Unless otherwise stated, this ratio was used for all our PNA-binding experiments. In vitro transfection and bgal analysis Cells were plated in 12-well plates (Falcon, Becton Dickinson, Franklin Lakes, NJ, USA) in the appropriate medium 48 h prior to transfection. All incubations of the various polyplex formulations with cells were performed in 1 ml of serum-free medium for 3 h at 371C in a CO2 incubator. After incubation, the serum-free medium was replaced with 1 ml of growth medium and cells were then incubated for 24, 48, and 72 h at 371C. In vitro analysis of bgal expression in cells was analyzed using a Galacto-Lightt chemiluminescent reporter assay (Tropix) (Applied Biosystems, Framingham, MA, USA) and results were normalized per mg protein as previously described.2 Statistical analyses were then performed for all triplicate experiments. In vitro cell proliferation assays H1299 cells were plated at a density of 1 104 cells/ml in 12-well plates 24 h before transfection. Medium was replaced with fresh medium on the day of transfection and cells were transfected with polylexes in a total volume of 30 ml at 371C and cell growth was monitored at various time points. Cells were prepared for analysis from 6–72 h post-transfection. Media were aspirated from each replicate well at each time point and adherent


cells were stained with crystal violet and solubilized with acidified ethanol for spectrophotometric determination of cell number by optical density (OD) at 595 nm.

Detection of nuclear-positive cells Upon termination of the reaction for the vector formation, the polyplex was hybridized to 1 ml (1 mM) of NLS– PNA for an additional 30 min at RT. H1299 cells were then incubated with methylene blue for 15 min. To visualize nuclei, cells were stained live with SYTOX green (Molecular Probes,Eugene, OR, USA, 1:1000) in each well of a 12-well plate for an additional 15 min at 371C at the end of the incubation period and examined under an Olympus IX71 confocal laser scanning microscope equipped with a Zeiss axiocam (Zeiss) camera using Axivision 3.0 (Zeiss). For confocal microscopy in H1299 live cells, we found that this combination of dyes gave the best results for visualization. For each transfection, four different fields were examined independently and a representative field was selected for microscopy. Cell injections and tumor growth About 107 H1299 cells were thoroughly washed twice, resuspended in phosphate-buffered saline (PBS) (100 ml/ mouse), and injected subcutaneously into the dorsal flank of athymic nude mice (12 animals per experimental group, Harlan, Indianapolis, IN, USA). Intravenous administration of the various polyplex formulations began 14 days after initial tumor cell injection. The amounts and volumes of the polyplex are the same as described above. Tumor volumes were measured prior to each injection of the polyplex and calculated using the formula p/6 L W H.43 In vivo bgal and immunohistochemical studies Subcutaneous tumors were produced in 4–6 week-old female athymic nude mice (Harlan, Indianapolis, IN, USA) by injecting 107 H1299 cells into their dorsal subcutaneous space. Tumor formation was monitored until tumors reached 5 mm in size in one plane. The polyplexes (containing 6 mg bgal DNA) were injected in a volume of 150–180 ml via tail vein of mice by single injections and tumors examined after 72 h for bgal gene expression using a Galacto-Lightt chemiluminescent reporter assay (Tropix). For bgal staining, tissue sections (8 mm thick) of H1299 tumors on microslides were fixed in acetone:ethanol (1:1) at 201C for 20 min, washed in PBS for 5 min and stained for X-gal for 24 h at 371C in a humidified chamber. Slides were then washed with PBS and counterstained with Nuclear Fast Red solution for 5 min after which they were progressively washed in 70, 80, 90, and 100% ethanol for 3 min each followed by two changes of xylene for 10 min each. Finally, slides were air-dried and covered with a few drops of Cytoseal 60 Mounting Medium (Stephens Scientific, Kalamazoo, MI, USA) and examined under a Nikon Eclipse E400 Microscope (Nikon Instruments, Melville, NY, USA). Quantitation of bgal was determined in tumor and other tissues using a Galacto-light chemiluminescent reporter assay (Tropix), following the manufacturer’s protocol. Measurements were done using a Monolight 3010 luminometer and normalized for protein concentration using a BCA protein assay kit. For immunohistochemical analyses, 8 mM thick paraffin-embedded tissue sections were adsorbed on polylysine-coated slides before CD13

detection by avidin–biotin. Sections were rehydrated using two changes of xylenes in a graded alcohol series, boiled for 10 min in 1 mM EDTA, cooled, incubated in PBS containing 0.5% hydrogen peroxide for 10 min to quench endogenous peroxidase before a final incubation in 200 ml of PBS containing 2% BSA for 1 h at RT. This was followed by WM15 CD13 human mAb (IgG1) in PBS–BSA overnight at 41C. Slides were then washed twice with PBS and incubated with PBS–BSA containing 2% normal horse serum (PBS–BSA–normal horse serum, Vector Laboratories, Burlingame, CA, USA) for 5 min. The solution was then replaced with 3 mg/ml biotinylated horse anti-mouse IgG (Vector Laboratories) in PBS–BSA–normal horse serum and further incubated for 1 h at RT. Slides were washed again and incubated for 30 min with Vectastain Elite Reagent (Vector Laboratories) diluted 1:100 in PBS. A tablet of 3,3-diaminobenzidine-tetrahydrochloride (Merck, Darmstadt, Germany) was then dissolved in 10 ml of deionized water containing 0.03% hydrogen peroxide, filtered through a 0.2-mm membrane, and overlaid on tissue sections for 5 min. The slides were washed as above and counterstained with Harris’ hematoxylin before microscopy. All animal experiments were conducted under institutional guidelines established for the Animal Core Facility at MD Anderson Cancer Center.

1521

Detection of p53 gene expression Xenografts from mice intravenously injected with polyplex formulations with or without p53 were removed and immediately stored at 801C. Genomic DNA was extracted from homogenized tumors and exons 4–9 of the p53 gene were amplified using the following primers: forward 50 -TTCTGGTAAGGACAAGGGT-30 and reverse 50 -AGGCATTGAAGTCTCATGGA-30 . And for b-actin, forward primers 50 -TAATACGACTCACTATAGGGAGA GCGGGAAATCGTGCGTGACATT-30 and reverse primers 50 -GATGGAGTTGAAGGTAGTTTCGTG-30 . Histology and TUNEL assay The tumor tissue was fixed in 10% formalin and embedded in paraffin for histological examination. The paraffin sections were stained with hematoxylin and eosin (H&E). To access the incidence of apoptotic cell death, the sections were stained with an in situ cell death detection-Peroxidase kit (Boerhinger-Mannheim, Mannheim, Germany), according to the manufacturer’s recommendations. Briefly, after deparaffinization, dehydration and inactivation of intrinsic peroxidase activity, the sections were digested with 2 mg/ml proteinase K at 371C for 15 min, then with terminal transferase and biotin-16-UTP followed finally with a dilution of peroxidase-conjugated streptavidin. Labeling was detected after incubation with diaminobenzidine. Statistical analyses Unless noted otherwise, results from in vitro experiments represent at least two independent experiments, whereas in vivo experiments utilized five animals per group, except for the survival studies which utilized 12 animals per group. All results are expressed as means7s.d. A mean with Po0.05 was considered statistically significant. The statistical significance of differences between the means of at least two independent samples was Gene Therapy


1522

tested by t-test, and that for long-term data was tested with one-way analysis of variance (ANOVA).

Acknowledgements We are grateful to Dr Randall Evans (Department of Molecular Hematology and Therapy) of the Flow Cytometry/Cell Sorting and Confocal Microscope/Image Analysis Core of the University of Texas MD Anderson Cancer Center for his invaluable technical assistance in confocal microscopy. This work was supported in part by NIH grant #5P30CA016672–29 supporting the core facility.

References 1 Sweeney P, Karashima T, Ishikura H, Wiehle S, Yamashita M, Benedict W et al. Efficient therapeutic gene delivery after systemic administration of a novel polyethylenimine/DNA vector in an orthotopic bladder cancer model. Cancer Res 2003; 63: 4017–4020. 2 Moffatt S, Wiehle S, Cristiano RJ. Tumor-specific gene delivery mediated by a novel peptide/polyethylenimene/DNA polyplex targeting aminopeptidase N (CD13). Hum Gene Ther 2005; 16: 57–67. 3 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: 16985–16987. 4 Branden LJ, Mohamed AJ, Smith CIE. A peptide nucleic acid nuclear localization signal fusion that mediates nuclear transport of DNA. Nat Biotechnol 1999; 17: 784–7871. 5 Chan CK, Jans DA. Enhancement of polylysine-mediated transfer infection by nuclear localization sequences: polylysine does not function as a nuclear localization sequence. Hum Gene Ther 1999; 10: 1695–1702. 6 Zanta MA, Belguise-Valladier P, Behr JP. Gene delivery: a single nuclear localization signal peptide is sufficient to carry DNA to the cell nucleus. Proc Natl Acad Sci USA 1999; 96: 91–96. 7 Colin M, Maurice M, Trugnan G, Kornprobst M, Harbottle RP, Knight A et al. Cell delivery, intracellular trafficking and expression of an integrin-mediated gene transfer vector in tracheal epithelial cells. Gene Therapy 2000; 7: 139–152. 8 James MB, Giorgio TD. Nuclear-associated plasmid, but not cellassociated plasmid, is correlated with transgene expression in cultured mammalian cell. Mol Ther 2000; 1: 339–346. 9 Collas P, Alestrom P. Rapid targeting of plasmid DNA to zebrafish embryo nuclei by the nuclear localization signal of SV 40T antigen. Mol Marine Biol Biotechnol 1997; 6: 48–58. 10 Dean DA. Import of plasmid into the nucleus is sequence specific. Exp Cell Res 1997; 230: 293–302. 11 Escriou V, Ciolina C, Helbling-Leclerc A, Wils P, Scherman D. Cationic lipid-mediated gene transfer uptake and nuclear import of plasmid DNA. Cell Biol Toxicol 1998; 14: 95–104. 12 Pollard H, Remy JS, Loussouarn G, Demolombe S, Behr JP, Escande D et al. Polyethylenimine but not cationic lipids promotes transgene delivery to the nucleus in mammalian cells. J Biol Chem 1998; 27: 7507–7511. 13 Dean DA, Byrd JN, Dean BS. Nuclear targeting of plasmid DNA in human corneal cells. Cell Biol Toxicol 1999; 19: 66–75. 14 Neves C, Escriou V, Byk G, Scherman D, Wils P. Intracellular fate and nuclear targeting of plasmid DNA. Cell Biol Toxicol 1999; 15: 193–202. 15 Hagstrom JE, Ludtke JJ, Bassik MC, Sebestyen MG, Adam SA, Wolff JA. Nuclear import of DNA in digitonin-permeabilized cells. J Cell Sci 1997; 110: 2323–2331. Gene Therapy

16 Hartig R, Shoeman RL, Janetzko A, Grub S, Traub P. Active nuclear import of single-stranded oligonucleotides and their complexes with non-karyophilic macromolecules. Biol Cell 1998; 90: 407–426. 17 Sebestyen MG, Ludtke JJ, Bassik MC, Zhang G, Budker V, Lukhtanov EA et al. DNA vector chemistry: the covalent attachment of signal peptide to plasmid DNA. Nat Biotechnol 1998; 16: 80–85. 18 Vacik J, Dean BS, Zimmer WE, Dean DA. Cell-specific nuclear import of plasmid DNA. Gene Therapy 1999; 6: 1006–1014. 19 Brown MD, Schatzlein AG, Uchegbu IF. Gene delivery with synthetic (non-viral) carriers. Int J Pharm 2001; 229: 1–21. 20 Bremner KH, Seymour LW, Logan A, Read ML. Factors influencing the ability of nuclear localization sequence peptides to enhance non-viral gene delivery. Bioconjugate Chem 2004; 15: 152–161. 21 Ludtke JJ, Zhang G, Sebestyen MG, Wolff JA. A nuclear localization signal can enhance both the nuclear transport and expression of 1 Kb DNA. J Cell Sci 2003; 112: 2033–2041. 22 Kalderon D, Richardson WD, Markham AF, Smith AE. Sequence requirements for nuclear location of simian virus 40 large-T antigen. Nature 1984; 311: 33–38. 23 Tanimoto M, Kamiya N, Matsuda A. No enhancement of nuclear entry by direct conjugation of a nuclear localization signal peptide to linearized DNA. Bioconjugate Chem 2003; 14: 1197–1202. 24 Nielsen PE, Egholm M, Berg RH, Buchardt O. Sequence-selective recognition of DNA by strand displacement with a thyminesubstituted polyamide. Science 2001; 254: 1497–1500. 25 Egholm M, Buchardt O, Christensen L, Behrens C, Freier SM, Driver DA et al. PNA hybridizes to complementary oligonucleotides obeying the Watson–Crick hydrogen-binding rules. Nature 1993; 365: 566–568. 26 Haahr-Hansen MH, Sode LL, Hyldig-Nielsen JJ, Engberg J. Detection of PNA/DNA hybrid molecules by antibody Fab fragments isolated from a phage display library. J Immunol Methods 1997; 203: 199–207. 27 Misra HS, Pandey PK, Modak MJ, Vinayak R, Pandey VN. Polyamide nucleic acid-DNA chimera lacking the phosphate backbone are novel primers for polymerase reaction catalyzed by DNA polymerases. Biochemistry 1998; 37: 1917–1925. 28 Gambari R. Peptide nucleic acids: a tool for the development of gene expression modifiers. Curr Pharm Des 1991; 7: 1839–1862. 29 Roulon T, Helene C, Escude C. Coupling of a targeting peptide to plasmid DNA using a new type of padlock oligonucleotide. Bioconjugate Chem 2002; 13: 1134–1139. 30 Gottschalk S, Cristiano RJ, Smith L, Woo SLC. Folate receptormediated gene delivery into tumor cells: potosomal disruption results in enhanced gene expression. Gene Therapy 1994; 1: 185–191. 31 Wagner E, Zenke M, Cotton M, Beug H, Birnstiel ML. Transferrin-polycation conjugates as carriers for DNA uptake into cells. Proc Natl Acad Sci USA 1990; 87: 3410–3414. 32 Young JL, Benoit JN, Dean DA. Effect of DNA nuclear targeting sequence on gene transfer and expression of plasmids in the intact vasculature. Gene Therapy 2003; 10: 1465–1470. 33 Abdallah B, Hassan A, Benoist C, Goula D, Behr JP, Demeneix BA. A powerful non viral vector for in vivo gene transfer into the adult mammalian brain: polyethylenmine. Hum Gene Ther 1996; 7: 1947–1954. 34 Arap W, Pasqualini R, Ruoslahti E. Cancer treatment by targeted delivery to tumor vasculature in a mouse model. Science 1998; 279: 377–380. 35 Pasqualini R, Koivunen E, Kain R, Lahdenranta J, Sakamoto M, Stryhn A et al. Aminopeptidase N is a receptor for tumor-specific


1523 peptides and a target for inhibiting angiogenesis. Cancer Res 2000; 60: 722–727. 36 Ellerby HM, Arap W, Ellerby M, Kain R, Andrusiak R, Rio GD et al. Anti-cancer activities of targeted pro-apoptotic peptides. Nat Med 1999; 5: 1032–1038. 37 Boussif O, Lezoualc’h F, Zanta MA, Mergny MD, Scherman D, Demeneix B et al. A versatile vector for gene and oligonucleotide transfer into cells in culture and in vivo: polyethylenimine. Proc Natl Acad Sci USA 1995; 92: 7297–7301. 38 Keller M, Harbottle RP, Perouzel E, Colin M, Shah I, Rahim A et al. Nuclear localization sequence templated non-viral gene delivery vectors: investigation of intracellular trafficking events of LMD and LD vector systems. Chem Biochem 2003; 4: 286–298.

39 Ferkol T, Pellicena-Palle A, Eckman E, Perales JC, Trzaska T, Tosi M et al. Immunologic responses of gene transfer into mice via the polymeric immunoglobulin receptor. Gene Therapy 1996; 3: 669–678. 40 Mulligan RC. The basic science of gene therapy. Science 1993; 260: 926–932. 41 Kozarsky JF, Wilson JM. Gene therapy: adenovirus vectors. Curr Opin Genet Dev 1993; 3: 499–503. 42 Jiao S, Williams P, Berg RG, Hodgeman BA, Liu L, Repetto G et al. Direct gene transfer into nonhuman primate myofibers in vivo. Hum Gene Ther 1992; 3: 21–33. 43 Krygier S, Djakiew D. Neurotrophin receptor p75(NTR) suppresses growth and nerve growth factor-mediated metastasis of human prostate cancer cells. Int J Cancer 2002; 98: 1–7.

Gene Therapy