Citation
Matrix attachment regions-containing plasmid DNA as facilitators in plasmid transfer

Material Information

Title:
Matrix attachment regions-containing plasmid DNA as facilitators in plasmid transfer
Creator:
Chancham, Pattravadee, 1974-
Publication Date:
Language:
English
Physical Description:
ix, 102 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
CHO cells ( jstor )
Chromatin ( jstor )
DNA ( jstor )
Gene expression ( jstor )
Gene therapy ( jstor )
Liposomes ( jstor )
Nuclear matrix ( jstor )
Plasmids ( jstor )
Transfection ( jstor )
Transgenes ( jstor )
DNA, Circular -- genetics ( mesh )
DNA, Superhelical -- genetics ( mesh )
Department of Pharmaceutics thesis Ph.D ( mesh )
Dissertations, Academic -- College of Pharmacy -- Department of Pharmaceutics -- UF ( mesh )
Gene Expression Regulation, Bacterial ( mesh )
Genetic Vectors -- therapeutic use ( mesh )
Interferon-beta -- metabolism ( mesh )
Nuclear Matrix -- genetics ( mesh )
Nuclear Matrix -- metabolism ( mesh )
Plasmids -- genetics ( mesh )
Plasmids -- pharmacokinetics ( mesh )
Plasmids -- therapeutic use ( mesh )
Transfection -- statistics and numerical data ( mesh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 2001.
Bibliography:
Bibliography: leaves 91-101.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Pattravadee Chancham.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
028449042 ( ALEPH )
54809251 ( OCLC )

Downloads

This item has the following downloads:


Full Text









MATRIX ATTACHMENT REGIONS-CONTAINING PLASMID DNA AS
FACILITATORS IN PLASMID TRANSFER

















By

PATTRAVADEE CHANCHAM












A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2001


























This work is dedicated to my parents, Yuwadee and Charun Chancham; and my grandparents, Kosum and Pramote Launpreeda.














ACKNOWLEDGMENTS


I would like to express my sincere appreciation to my mentor, Dr. Jeffrey Hughes, for his advice, patience, and understanding but most of all his brotherly generosity and fatherly protection. I would like to thank my other committee members, Dr. Gayle Brazeau, Dr. Guenther Hochhaus, Dr. Sean Sullivan, Dr. Edwin Meyer, and Dr. William Farmerie for their expert advice and kind support.

I also wish to acknowledge all the personnel in the Department of Pharmaceutics including, secretaries, graduate students, and particularly the Hughes group. I especially thank Adam Persky who helped me throughout my graduate life. I also would like to thank Wu Xiao and Yi Wen for advice on my experiments. My friends Oravaree, Nopadon, Jintana, Intira, and Ariya were around to continuously support me during good and bad times.

Lastly, I would like to express thanks to my parents and grandparents for their unselfish love; to my sisters for their encouragement; and to Jim Buranatrakul for being my inspiration.












111















TABLE OF CONTENTS
IRge

ACKNOWLEDGMENTS in1
LIST OF TABLES vi
LIST OF FIGURES vii
CHAPTERS

1 INTRODUCTION 1


2 BACKGROUND AND SIGNIFICANCE 5


Matrix Attchment Regions (MARs) Overview. 5
Characteristics 7
Functions 14

3 RELATIONSHIP BETWEEN PLASMID DNA TOPOLOGICAL FORMS AND IN VITRO TRANSFECTION 23

Introduction 23
Materials and Methods. 24 Results 31
Discussion and Conclusion. 35

4 USE OF PHARMACOKINETIC PARAMETERS TO INTERPRET GENE EXPRESSION 43

Introduction. 43
Materials and Methods. 45 Results 51
Discussion and Conclusion. 57

5 MATRIX ATTACHMENT REGIONS-CONTAINING PLASMID DNA INCREASES GENE EXPRESSION IN VITRO 60

Introduction. 60
Materials and Methods. 61
Results 70
Discussion and Conclusion 77
iv










6 CONCLUSION AND FUTURE PROSPECTS 87
Conclusion 87 Future Aims 89


REFERENCES 91
BIOGRAPHICAL SKETCH 102












































V















LIST OF TABLES

Table Page

3-1 Pharmacokinetics of plasmid DNAs in cytoplasmic solution. .37 4-1 Pharmacokinetic parameters of simulated plasmid DNAs. .53 4-2 Transgene expression of plasmid pGL3 and pGM in CHO cells. .54 4-3 Pharmacokinetic parameters of pGL3 and pGM. .56

5-1 Pharmacokinetic parameters of pGM and pGL3 in CHO, SKnSH, and neuronal cells.74 5-2 Plasmid DNA extracted from nucleus and cytoplasm of CHO) cells .82
































vi















LIST OF FIGURES

Figure Page

2-1 Nuclear matrix isolated by amine modification .21

3-1 Diagram of first-order kinetics .29

3-2 Agarose gel electrophoresis of plasmid DNA topoisoforms .32 3-3 Transfection efficiency of pDNA-liposome .33

3-4 Transfection efficiency of pDNAs by using electroporation.34 3-5 Agarose gel electrophoresis of plasmid DNAs in cytoplasm.36 3-6 Mean fluorescence of the labeled plasmid DNAs .38 4-1 Schematic representation of the plasmid DNAs .47 4-2 Simulation of transgene expression of plasmid DNAs 1, 2, 3, and 4.52 4-3 Transgene expression of plasmid DNAs, pGL3 and pGM, in CHO cells.55 5-1 Dose study of plasmid DNAs in CHO cells .64 5-2 Transgene expression of pGL3 and pGM in CHO and SKnSH .71 5-3 Transgene expression of pGL3 and pGM in hippocampal primary neuron, astroglia,
and microglia .72

5-4 Luciferase activity of pGL3 when contransfected with pEPI-1 in CHO and SKnSH
cells .78

5-5 Transgene expression of pGL3 and pGM in CHO cells when incubated with histone
deacetylase inhibitor, trichostatin A (TSA). .80 5-6 Plasmid DNA, pGM, and pGL3 extracted from nucleus and cytoplasm of CHO cells .81






vii














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

MATRIX ATTACHMENT REGIONS-CONTAINING PLASMID DNA AS FACILITATORS IN PLASMID TRANSFER By

Pattravadee Chancham

December 2001
Chairman: Dr. Jeffrey A. Hughes
Major Department: Pharmaceutics

Nonviral gene transfer is an alternative to viral vectors for gene transfer.

However, nonviral transgene expression remains undesirably low and transient. Matrix attachment regions (MARs) are DNA elements that are defined by their high affinity for the nuclear matrix. MARs may also be related to long-term transgene expression in vitro. The purpose of this research is to evaluate human interferon-P MARs element in various cell types. This was done by constructing MARs-containing pDNA and comparing their transgene expression with non-MARs-containing pDNA. We found that MARscontaining pDNA increased and prolonged the expression in Chinese hamster ovary (CHO), but not in human neuroblastoma cells (SKnSH) and neuronal cells (primary neuron, astroglia, and microglia). From cotransfection experiment, MARs-containing pDNA had trans effect on another pDNA. This vector also acted synergistically with histone deacetylase inhibitor, trichostatin A. Polymerase chain reaction was used to




viii








monitor intracellular distribution of pDNA. We found that MARs-containing pDNA has similar intracellular distribution as non-MARs-containing pDNA.

Because pDNA was a basic tool in nonviral gene transfer, the relationship

between DNA topoisoforms and in vitro transfection efficiency is discussed. The DNA topoisoforms did not affect the level of transgene expression even though the extent depends on cell types and pDNA promoter







































ix













CHAPTER 1
INTRODUCTION


Gene therapy is the ultimate method for delivering proteins into the body. The technique of introducing the genetic material into the target cells of a patient is a key component of every gene therapy protocol. A variety of gene delivery systems are currently used to insert therapeutic genes into somatic cells. They are divided into viral and nonviral gene transfer methods.

Conventional vectors currently used for gene delivery have a number of

limitations. Viral vectors may randomly integrate into the host genome. Integration of a therapeutic episome into the host chromatin has the distinct possibility of activating or inactivating important genetic loci, with the potential for deleterious consequences (Wendelburg and Vos, 1998). For example, insertional mutagenesis and transgene silencing result in such pathological changes in tissues as tumorigenesis, growth inhibition, or cell death. This problem of safety may be difficult to overcome. A nonviral or plasmid-based vector is an alternative. For instance plasmid DNA (pDNA) vectors do not have a size constraint imposed by viral packaging. Moreover, it has lower toxicity and immunogenicity compared to a viral vector (Wendelburg and Vos, 1998).

Even though naked DNA has used to successfully increase expression in many cell types, stability of unprotected DNA is a major concern. Plasmid DNA needs an efficient nonviral delivery method, such as receptor-mediated ligand targeting system (Wagner et al., 1992; Wilson et al., 1992), liposomes (Tang and Hughes, 1998),


1







2
hemagglutinating virus of Japan (HVJ)-liposome (Tsukamoto et al., 1999; Yanagihara et al., 1996) or physical methods, such as injection (Hengge et al., 1995) to deliver it to the target cell. Another important issue is the transient expression. Plasmid DNA appears to have a short lifetime in most tissues. One possible reason is the inherent susceptibility to nucleases of DNA. Vector DNA that is unable to replicate is rapidly diluted out in mitotic cells (Wells et al., 1998; Wells et al., 1997). In post-mitotic cells, it also appears that vector DNA is rapidly lost, perhaps by passage through nuclear pores and subsequent degradation (Wohlgemuth et al., 1996). Furthermore, most pDNAs do not integrate into the host genome, thus they tend to be targets for enzyme degradation. Plasmid DNA that could be replicated or could be retained in cells without integration into the host genome is required to increase the persistence of DNA in cells (Piechaczek et al., 1999). Increasing the persistence of pDNAs would lead to prolonged and enhanced transgene expression.

Matrix attachment regions (MARs) are DNA sequences that are identified through their high affinity to bind to the nuclear matrix or scaffolds in vitro (Gasser and Laemmli, 1986; Mirkovitch et al., 1984). The following evidence indicates that MARs-associtaed plasmid DNA might be a prime candidate for nonviral gene delivery:

MARs replicated and was retained in cells without integrating into host
chromosome (Piechaczek et al., 1999).


MARs inhibited methylation that would repress gene transcription (Dang et al.,
2000; Forrester et al., 1999).


MARs extended the histone acetylation domain and conferred chromatin
accessibility (Fernandez et al., 2001; Forrester et al., 1994; Jenuwein et al., 1997).






3
MARs protected DNA from neighboring chromatin effect (Kalos and Fournier,
1995; Klehr et al., 1992; Phi-Van et al., 1990; Poijak et al., 1994).


Proteins necessary for transcription were found in MARs (Boulikas, 1995).

Moreover, MARs do not express protein, thus oncogenicity, immunogenicity, and toxicity may be not the major concerns. This evidence may relate to increasing and maintaining expression.

Objective I

The main objective of the project was to develop pDNA vector for nonviral gene transfer by using DNA sequences called matric attachment regions.

Hypothesis

We hypothesized that MARs-containing pDNA prolongs or enhances transgene expression compared to non-MARs-containing pDNA. To test this hypothesis, MARscontaining pDNA are constructed and its transgene expression is compared with nonMARs-associtaed pDNA. The role of MARs-containing pDNA in various cell types is also investigated. Pharmacokinetic concepts can be applied to gene expression of pDNAs because transgene expression can be followed over time by treating pDNA as a prodrug and we can then measure pDNA active metabolites. A combination of multiplepoint measurements and pharmacokinetic parameters were used in data analysis.

Plasmid DNA is the basic tool for nonviral gene delivery. Different

conformational state of pDNA can assist in developing approaches to increase its transfection efficiency. The initial form of pDNA isolated from bacteria is supercoiled. However, exposure of pDNA to physical and chemical environments can lead to the their






4
degradation, which means that the conformation can change to open circular, linear, or even fragmented DNA.

Objective II

Another objective in this study is to investigate the relationship between

supercoiled, open circular, and linear pDNA topoisoforms and in vitro transfection efficiency.

Hypothesis

Plasmid DNA topoisoforms affect transfection efficiency. To test this hypothesis, open circular and linear pDNA are produced. These forms (including supercoiled) are transfected into cells by using either electroporation or liposomes. The half-life of pDNA topoisoform in cytoplasm is calculated.














CHAPTER 2
BACKGROUND AND SIGNIFICANCE



Matrix Attchment Regions (MARs) Overview


The architecture of the nuclear interior is composed of two nucleic acidcontaining structures: a DNA-containing structure called the chromatin and an RNAcontaining structure (He et al., 1990; Nickerson et al., 1998). Eukaryotic chromatin is organized into domains that may affect differential gene expression (Bode et al., 1995; Boulikas, 1995). This organization is brought about by the anchoring of specific DNA sequence landmarks to a network of protein crossties termed the nuclear matrix, at an interphase or chromosomal scaffold during mitosis. After a combination of nuclease digestion and extraction, the DNA sequences that tightly associate with the nuclear matrix or scaffold in vitro have been called matrix or scaffold attachment regions (MARs or SARs) (Bode et al., 1992). These MARs or SARs have an average size of 500 base pairs (bp), are spaced approximately every 30 kilobase pair (kbp) (Boulikas, 1995), and are control elements maintaining independent realms of gene activity. The DNA replication, transcription, repair, splicing, and recombination appear to take place on the nuclear matrix. The MARs have been experimentally defined for several gene loci, including the chicken lysozyme gene, human interferon-3 gene, humanp-globin gene, chicken o-globin gene, p53, human protamine gene cluster (Singh et al., 1997), and human serpin gene cluster (Rollini et al., 1999).

5






6
The sequences of MARs do not have clear-cut consensus sequences but share conformation characteristics (Benham et al., 1997; Yamamura and Nomura, 2001). MARs are typically 70% AT-rich sequences (Boulikas, 1993). These sequences are responsible for bending (Yamamura and Nomura, 2001), unwinding of DNA (Bode et al., 1992), and binding DNA with the cellular nuclear matrix, which is the prerequisite for MARs functions (Bode et al., 1992; Boulikas, 1993; Boulikas, 1995; Lechardeur et al., 1999). Binding of MARs-containing vector with nuclear matrix was believed to confer long-term expression. Many DNA-binding proteins are found in MARs such as Topoisomerase II (Razin et al., 1991), Histone HI (Izaurralde et al., 1989), lamin (Luderus et al., 1992), and SATBI (Dickinson et al., 1992). Topoisomerase II and Histone HI are major enzymes found in MARs. Topoisomerase II exhibits cooperative binding to MAR DNA and Histone H1 maintains a higher-order structure of chromatin (Boulikas, 1995). Much evidence supports MARs/protein interactions and their biological function (Alvarez et al., 2000; Liu et al., 1999; Ramakrishnan et al., 2000; Stratling and Yu, 1999; Sun et al., 2001), although the relationships are still unclear.

Matrix attachment regions are also boundary elements that define boundaries of the independent chromatin domain. They protect DNA from neighboring chromatin affecting transgene expression (position effect). Therefore, they enhance expressions (Kalos and Fournier, 1995; Klehr et al., 1992; Phi-Van et al., 1990; PoIjak et al., 1994). MARs-containing plasmid DNA confers position-independent and copy numberdependent expression (Phi-Van et al., 1990; Rollini et al., 1999; Stief et al., 1989). The studies showed that MARs enhances expression independent of orientation; forward or reverse orientation from its natural form (Klehr et al., 1992; Mielke et al., 1990).






7
However, MARs can block expression when placed between promoter and enhancer (Stief et al., 1989).

Matrix attachment regions are chromatin-remodeling elements with enhanced

chromatin accessibility. The combination of MARs and ji enhancer confers accessibility upon a distal promoter. This may be due to the generation of an extended domain of histone acetylation by MARs (Fernandez et al., 2001; Forrester et al., 1994; Jenuwein et al., 1997). MARs sequences also influence transgene methylation status (Dang et al., 2000; Forrester et al., 1999). MARs inhibit de novo methylation of retroviral 5'-LTR (long terminal repeat). Thus MARs inhibits promoter shutdown and leads to higher levels of expression (Dang et al., 2000).

Instead of the conserved sequence, MARs have a different characteristic

conformation (Yamamura and Nomura, 2001). Although MARs do not have consensus sequences (Singh et al., 1997), they do share similar characteristics with them.



Characteristics


MARs as Potential Origins of Replication

It is known that newly replicated DNA is specifically located on the nuclear matrix (Boulikas, 1993; Boulikas, 1995; Singh et al., 1997). Previous studies confirm that replication forks are associated with the nuclear matrix (Berezney and Coffey, 1975; Gerdes et al., 1994; Jackson and Cook, 1986; Ortega and DePamphilis, 1998; Vaughn et al., 1990) and that replication machinery is immobilized by attachment to the nuclear matrix (Cook, 1999). According to the model of Pardoll and coworkers, DNA is reeled through its nuclear matrix attachment site during replication. The replication origin is







8
transiently attached to the nuclear matrix. The replication origins associate with the nuclear matrix in the late G, phase and dissociate after initiation of DNA replication in S phase (Djeliova et al., 2001). MARs from yeast coincide with the putative origins of replication and Drosophila MARs can drive the autonomous replication of plasmids in yeast. In addition, the isolation and cloning of putative origins of replication from monkey cells in culture has shown them to possess sequence homology with MARs (Boulikas, 1993; Boulikas, 1995). It has been shown that nuclear matrix attachment sites, homeotic protein recognition and binding sites, and the origins of replication share the ATTA, ATTTA, and ATTTTA motifs (Boulikas, 1993). Enzymes necessary for replication, such as DNA polymerase, Topoisomerase II, and primase, were found in MARs (Boulikas, 1995; Lodish et al., 1998). The deformation of the nuclear matrix protein and DNA by chemotherapy agents such as alkylating agents and by ionizing radiation affected replication, which led to cell death (Muenchen and Pienta, 1999). This suggests that the differential activation of origins of replication may be regulated on the nuclear matrix. Studies show that both stimulation of replication by transcription factors and the presence of cis-acting elements distant from the origin of bidirectional replication are able to affect origin firing.

Matrix attachmet regions also affect expression and replication-timing patterns of translocated BCL2 oncogenes. The translocated allele replicates at the GUS boundary, while the wild-type allele continues to replicate as usual in mid-S phase. These differences are accompanied by allele-specific changes in BCL2 expression, since the major breakpoint region (mbr) of BCL2, which is implicated in 70% of translocation from Chromosome 18 to Chromosome 14 present in human follicular lymphoma, is a MARs






9
(Sun et al., 2001). The AT-rich region flanking the BCL2 mbr is a binding site for the MAR protein SATBI (Ramakrishnan et al., 2000). Major Classes of Matrix Attachment Sites Known to Be AT-Rich Sequences

These AT-rich sequences (approximately 100-1,000 base pairs in length) are

closely related to the binding and cleavage consensus of Topoisomerase II (Gasser and Laemmli, 1986; Mirkovitch et al., 1984) and the binding site of Histone HI, both of which are important for chromosome structure (van Drunen et al., 1999). Topoisomerase II can cleave double-stranded DNA, pass an uncut portion of the DNA between the cut ends, and then reseal the cut. This is critical for DNA replication (Lodish et al., 1998).

The most remarkable role of the AT-rich regions is that they facilitate unwinding of DNA previously catalyzed by helicase molecules, torsional strain, or proteins. The DNA unwinding is the first step for transcription, since it facilitates binding of RNA polymerase and other transcription factors to DNA. Under torsional strain, the DNA unwinding AATATATTT motif, present within the MARs of both IgH and P-interferon genes, becomes the site of nucleation (Bode et al., 1992). Mutation of this motif to ACTGCTTT voids both the MARs activities of the fragment as well as its DNA unpairing ability. Bode et al. (1992) showed that under various ionic conditions, the mutant form displays neither MAR activity nor unwinding capability. The unwinding property was shown to be important for binding to the nuclear matrix and for augmentation of gene expression in stable transformants. The discovery of Topoisomerase II as part of the chromatin remodelling CHRAC complex suggests that these elements may influence transcription regulation via nucleosome remodeling (van Drunen et al., 1999).






10
Many MARs contain significant stretches of AT-rich sequences. It has been

suggested that the simple occurrence of isolated AT-rich regions is not sufficient to cause matrix association. Rather several such regularly spaced motifs are required (Singh et al., 1997).

MARs May Represent Mass Binding Sites for Protein Transcription Factors

Studies show that the nuclear matrix is a microenvironment for transcription factors such as Myb, Myc, RFP, C/EBP, AP-1, Spl, and NMP-1 (ATF) (Bode et al., 1995). Protein transcription factors such as large T antigen, AP-1, and SpI that bind to nuclear sites were found to stimulate replication and transcription in viruses, metazoans, and yeast (Bode et al., 1995). In addition, DNA-binding proteins of noncoding DNA sequences, such as Topoisomerase II, Histone H1, lamin Bl, 120-kDa protein (SP120), scaffold attachment factor A (SAF-A), attachment region binding protein (ARBP) and a thymus specific MAR-binding protein (SATB 1), were found in the nuclear matrix (Boulikas, 1995; Luderus et al., 1994). These specific proteins exhibit cooperative binding and distinguish between MAR and non-MAR DNA (Boulikas, 1995; Luderus et al., 1994). The AT-tract exhibits a high affinity for these proteins (Boulikas, 1995). In the studies over the last 10 years, various types of proteins were found in the nuclear matrix. SAF-B interacts with RNA polymerase II and a subset of serine-/arginine-rich RNA processing factors (SR protein) serve as a molecular base to assemble a transcription complex (Nayler et al., 1998). Heterogenous nuclear ribonucleoprotein (hnRNP) is the major protein component of the nuclear matrix and serves as the transcription factor of the central nervous system (Stratling and Yu, 1999).






11
Transcription Enhancer May Be MARs

A significant number of studies have shown that MAR sequences are located near or at enhancer sites. In addition, MAR sequences were shown to act as transcriptional enhancers in experiments involving cells in culture and in transgenic animals (Boulikas, 1993; Mielke et al., 1990). Removal of the MARs decreased the abundance of mRNA by a factor of 35 to > 1000 in B lymphocytes of transgenic animals (Boulikas, 1995). The MARs model created by Boulikas (1995) explains looping out of DNA and juxtaposing enhancers by synergistic interaction of classical MAR proteins (composite MAR model) (Boulikas, 1995). The core enhancer is flanked by AT-rich sequences of about 300-500 bp able to sequester classical matrix proteins by cooperative interaction. This process of juxtaposing distant MAR elements and causing looping of DNA, brings together on the nuclear matrix two core enhancers or a core enhancer and a core origin of replication (100-200 bp) that cohabit with the AT-rich MAR (along with the transcription factors bound to them) to facilitate transcription and replication. MARs Harbor Intrinsically Curved DNA

Curve DNA has been identified at or near several matrix attachments sites (Singh et al., 1997). Bending of double-strand DNA is mainly caused by homopolymeric dA of at least 4 bp, called A-tract (Haran and Crothers, 1989; Koo et al., 1986).

Because of its A-tracts, MARs have a longer bent part and higher angle/helical turn than the other regions. The A-tract in MARs showed a nonrandom distribution but were clustered closely within MARs (Yamamura and Nomura, 2001). Intrinsically, curve DNA is important in nuclear processes involving specific protein-DNA interactions, such as recombination and transcription (Boulikas, 1993).






12

Curved DNA motifs reflect a 10.4 nucleotide periodicity from center to center of

(A)n stretches. Optimal curvature is expected for sequences with repeats of the motifs AAAAn6AAAAn7AAAA and TTTAAA (Singh et al., 1997). A Class of MARs May Harbor Kinked DNA

Kinked DNA has generally been associated with the presence of copies of the

dinucleotides TG, CA, or TA that are separated by 2-4 or 9-12 nucleotides. For example, kinked DNA is produced by the motif TAn3TGn3CA, with TA, TG, and CA occuring in any order. The CA, TA, and TG dinucleotides are overrepresented in DNA sequences that are protein recognition sites. Previously published studies show that MARs may display a usual richness of AT, TG, and CA (Boulikas, 1993; Singh et al., 1997). DNase I-Hypersensitive Sites May Be Diagnostic of MARs

Gene activation in eukaryotes has been proposed to consist of a multistep process that includes changes in chromatin structure, modifications of histones, and transcriptional activation of promoter (Blackwood and Kadonaga, 1998; Felsenfeld et al., 1996; Grosveld, 1999; Struhl, 1998). The relationship of these events is still under investigation, but MARs might be involved in these processes.

The decondensation of the chromatin structure reflected by increasing sensitivity to DNase I digestion is simply a consequence of transcriptional activity (Klehr et al., 1992). Within transcriptionally active regions of chromatin, some sites are nearly as sensitive to DNase I digestion as naked DNA. These DNase I-hypersensitive (DNase IHS) sites occur in regions where transcription factors abound. A number of studies showed that MARs could induce DNase I-HS sites in chromatin. The DNase I-HS sites in the Drosophila histone gene repeat, and coincide with, Topoisomerase II cleavage sites







13
and with the MAR sector in the H1-H3 intergenic region (Boulikas, 1995). The MAR/ORI of the 5' flanking region of the chicken oc-globin gene cluster harbors a constitutive DNase I-HS site that is detected in chromatin from many chicken tissues. The studies in transgenic animals showed that, whereas the 95-bp Iggt core enhancer was necessary and sufficient for the accessibility of TF to T7 promoter, it required the MAR region to induce DNase I hypersensitivity (Boulikas, 1995; Forrester et al., 1994). Furthermore, the ORI of SV40 minichromosomes is attached to the matrix and is hypersensitive to DNase I.

From the structural point of view, MARs are thought to be involved in chromatin condensation and chromosome formation. A synthetic AT-hook protein, which specifically binds to S/MARs, interferes with proper chromatin condensation in Xenopus laevis egg extracts (van Drunen et al., 1999). Moreover, alignment of S/MARs around the central core of mitotic and meiotic chromosomes may be required for correct condensation of DNA within these regions (van Drunen et al., 1999).

Transcription activation of the pt gene during normal lymphoid development

requires a synergistic collaboration between the enhancer and flanking MARs (Forrester et al., 1994). The t enhancer in combination with a flanking MAR can confer accessibility on a distal site in nuclear chromatin, whereas the enhancer alone mediates only local chromatin accessibility (Jenuwein et al., 1997). The immunoglobulin MARs antagonized methylation-dependent repression of long-range enhancer function (Forrester et al., 1999). The MARs allow the generation of an extended domain of histone acetylation, which could account for the long-range function of the [t enhancer in combination with MARs (Fernandez et al., 2001).






14
Histone HI, the most abundant repressor of gene activity (Boulikas, 1995), locks the two helical turns of the DNA around the nucleosome and maintains higher-order chromatin structures, which results in a more compact form of chromatin (van Drunen et al., 1999). Matrix attachment regions facilitated the displacement of Histone HI from chromatin through interactions with proteins with similar DNA binding motifs, such as HMG-I/Y (high motibility group). Competition between HI and these HMG proteins may contribute to determining the global distribution of active and inactive chromatin. Also, histone acetylation has been linked to transcriptional regulation via MARs. Hyperacetylation is a hallmark for active regions in the genome, while hypoacetylation is typical of regions that are transcriptionally inactive (van Drunen et al., 1999).



Functions


Matrix Attachment Regions Enhance and Prolonge Gene Expression

Matrix attachment regions are typically 70% AT, which has unwinding properties to enhance transgene expression (Bode et al., 1992). A single S/MAR construction of human IFN-P domain increases gene activity and the expression levels mirror the S/MAR activity in vitro (Bode et al., 1995). Many studies reported in vitro function of the MARs-containing vector that led to the following preliminary conclusions:

* MARs function is independent of S/MARs orientation (Klehr et al., 1992; Mielke et
al., 1990).

* MARs effects can be monitored if the marker and selector are physically coupled
(located in the same gene) (Bode et al., 1992). MARs can block the enhancer effects
if it is placed between the promoter and enhancer (Bode et al., 1995).

* Elevated expression levels in the presence of S/MARs are restricted to the stable
expression, not the transient expression (Bode and Maass, 1988; Kalos and Fournier,






15
1995; Klehr et al., 1992; Poijak et al., 1994; Wang et al., 1996). Therefore, MARs
have been used to detect of integrated transgenes in transgenic mice embryos
(Gutierrez-Adan and Pintado, 2000).

* MARs can form minidomains and, in the appropriate environment, MARs enhance
regulatory effects, thus resulting in increased level of gene expression (Bode et al.,
1995; Mielke et al., 1990; Stief et al., 1989).

* In stable transfection, the reporter genes flanked by certain MARs show positionindependent, copy number-dependent expression and augmentation of the
transcriptional activity (Bode et al., 1992; Bode et al., 1995; Kalos and Fournier,
1995; Stief et al., 1989).

* MARs replicate episomally in cells and have been stably maintained for more than
100 generations (Piechaczek et al., 1999).

* MARs influence transgene methylation status (Dang et al., 2000; Kirillov et al., 1996;
Lichtenstein et al., 1994). The hIFN-3 SAR inhibits de novo methylation of the retroviral 5'-LTR. The SAR element was able to alleviate methylation-mediated
transcriptional repression (Dang et al., 2000).

MARs can either be an enhancer or domain protector, the major function of which is to insulate transcription units from the regulatory influences of neighboring genes or chromatin domains.

Matrix attachment regions are prime candidates for domain borders. They are able to mediate an attachment to the nuclear matrix in vitro and are frequently found in the vicinity of the ends of a domain that are otherwise defined by constitutive DNase IHSsites or a decrease of the general DNase I sensitivity (Bode et al., 1995). Although not all MAR elements are boundaries of chromatin domains, a few elements with presumptive boundary function have been defined (Kalos and Fournier, 1995). The MAR-containing vector is a good candidate for long-term expression of viral/nonviral vector because the boundary effect may enhance and prolong the level of gene expression (Auten et al., 1999; Murray et al., 2000). These elements can be used for creating minidomains in cultured cells. The segments of chromatin with multiple points of






16
attachment may physically resist compaction into heterochromatin and gene silencing. The complete minidomains confer long-term stability to the enhanced expression level (Bode et al., 1995).

Elements that can insulate transgenes from position effects have stable integrated transgenes and those transgene are expressed irrespective of their sites of integration (position independent expression) (Kalos and Fournier, 1995). MARs, boundary element

(BE), and locus control region (LCR) have position independent effects. Thus, they are the candidates in the selection of stable cell lines with high-level expression characteristics. Among these entire elements, only MARs increase the proportion of high-expressing clones (Zahn-Zabal et al., 2001). Furthermore, MAR-containing vectors with a low copy transgene number are expressed at higher levels than high copy transgenes number (Bode et al., 1992; Kalos and Fournier, 1995). Cells harboring multiple-copy integration of the transgene are transcriptionally inactive (Kalos and Fournier, 1995). This is based on the fact that classical transfection techniques resulting in a multiple random integration may affect the properties of transgene (Bode et al., 1995). Furthermore, random integration may lead to insertional mutagenesis and to silencing of the transgene (Piechaczek et al., 1999). Transfection methods, such as electroporation are required to promote integration of one copy per cell (Bode et al., 1995).

The ideal vector for nonviral gene delivery should express the protein of interest and remain in the cell without integration for an extended time. Since the MAR element has a positive effect on transcriptional levels and is capable of stabilizing these high






17
levels over extended periods of time, they can be the instruments of construction for a new generation of expression vectors.

Binding to Nuclear Matrix Is a Prerequisite for MARs Functions

The physiological role of MARs is directly related to MARs binding to the

nuclear matrix (Luderus et al., 1992; Luderus et al., 1994). The increasing transcriptional efficiency of chicken lysozyme results from attachment of the MAR element to the nuclear scaffold material (Stief et al., 1989). Therefore binding to the nuclear matrix is a prerequisite for transcriptional enhancement. An excess of nonbound S/MARs constructs would lead to a transcriptional level not distinguishable from that of a non-S/MAR construct (Bode et al., 1995). Mutation of the tumor suppressor, p53, gene modulates gene expression. This mutation is reconstituted in human cancer. Murine and human mutation p53, but not wild type p53, specifically binds with high affinity to a variety of MARs DNA elements (in particular AATATATTT, the potential base-unpairing sequences). MARs associtaed with human artificial chromosomes were found to persist as episomes in long-term culture with stability per generation of approximately 80% and were containing with cell nuclear matrix (Cossons et al., 1997). Extrachromosomal retention of Epstein-Barr virus base vector depends on a tran-acting element, EpsteinBarr virus nuclear antigen-1 (EBNA-1), and a cis-acting element, latent origin of replication, oriP (Gahn and Schildkraut, 1989; Harrison et al., 1994; Wysokenski and Yates, 1989). The nuclear retention function is thought to reflect an interaction between EBNA-1, oriP-containing plasmid DNA, and the nuclear matrix chromosomal scaffold and appears to be sufficient to hold plasmid DNA in the nucleus (Jankelevich et al., 1992; Wensing et al., 2001). However, incorporation of nuclear matrix attachment regions into







18
the EBV Type 1 genome does not induce long-term expression of a foreign gene during latency (Makarova et al., 1996).

Binding of MARs is cell-type specific (Will et al., 1998). MARs have been

categorized as constitutive (permanent) or facultative (cell-type specific) (Singh et al., 1997). Constitutive MARs occur in all types of cells irrespective of the tissue in which they are found. In contrast, the presence of a facultative MAR is tissue-specific and that tissue governs its use. Moreover, some MAR binding transcription factors and classical matrix proteins are cell-type specific and bind strongly and selectively to MARs from different species (Boulikas, 1993; Boulikas, 1995). Plasmid DNA Topoisoforms and Transgene Expression

Topology of pDNA, whether supercoiled, open circular, also called singlestranded nicked, or linear form, is another aspect involved in the expression of pDNA. A number of studies showed that the topology of the transfected DNA molecules determines the level of gene expression (Kreiss et al., 1999; Ludtke et al., 1999; Pitard et al., 1997). The topological state of eukaryotic DNA seems to be important in gene expression, especially in the transcription process. For example, DNA supercoiling facilitates the formation of transcription preinitiation complex, which prevents subsequent assembly of promoter sequences into nucleosomes and allows transcription on the chromatin templates. In the fibroin gene, DNA superhelicity enhances transcription by accelerating formation of the complex (Hirose and Ohba, 1993; Hirose et al., 1985). A change in the superhelical density results in derepression of the yeast mating-type gene (Hirose et al., 1985).






19
Enzyme topoisomerase, especially Topoisomerase II (Topo II) is important in the topology of pDNA and also in the control of gene expression (Hirose and Ohba, 1993). Transcription of the mouse Hox-2.1 gene is inhibited by treatment of F9 embryonic carcinoma cells with a Topo II inhibitor etoposide (Hirose and Ohba, 1993). As an enzyme, Topo II can cleave double-stranded DNA, pass an uncut portion of the DNA between the cut ends, and then reseal the cut (Lodish et al., 1998), thus changing positive supercoiled DNA into negative supercoiled DNA. Transcription initiation of eukaryotic genes is thought to involve local unwinding of the DNA double helix within the promoter region (Hirose and Ohba, 1993).

Topo II is one of the major enzymes found in the nuclear matrix and scaffold

(Boulikas, 1995; Lodish et al., 1998) and is recognized by oligo (dA) tracts. The AT-rich region facilitates DNA unwinding when catalyzed by Topo II molecules (Bode et al., 1995). The AATATATTT motif present within the MARs of both IgH and P-interferon genes becomes the nucleation site of a DNA unwinding effect under torsional strain (Bode et al., 1992). Binding of negative supercoiled S/MAR associate pDNA or nonMAR pDNA with nuclear matrix overnight showed faint bands of nicked and linear form in the bound fragment, which were not present in the original mixture. These bands increased over time (Kay and Bode, 1994) thus demonstrating the action of Topo II in the nuclear matrix.

A binding assay study requires isolation techniques that highly preserve the underlying structure of the nuclear matrix while maintaining its functions. The separation methods of the nuclear matrix affect the quality of the structure and the binding site (Bode et al., 1995; Donev, 2000). After digesting DNA with nuclease,






20
various methods are used to elute the chromatin fragment. Harsh procedures such as very high to extremely low salt extractions (2 M NaCI, 0.4 M KC1, or 0.25 M NaCl) were responsible for the less than complete preservation of fine structure that led to some rearrangements in the attachment of DNA loops. In contrast, mild extraction procedures with 25 mM lithium 3,5-diiodosalicylate (LIS) or 0.65 M ammonium sulphate gave similar results for the type of MARs sequences investigated (strong, weak, non-MARs) (Donev, 2000). Utilization of amine modification, hydroxysulfosuccinimide acetate (sulfo-NHS), has been shown to preserve most of the nuclear matrix structure and its function (Wan et al., 1999).

The efficiency of nuclear matrix/pDNA binding was greatly affected by the

topological state of pDNA. Tsutsui and coworkers (1993) showed at least two classes of DNA-binding sites in the nuclear matrix: one is highly specific to supercoiled DNA in that it does not bind to relaxed or linear forms, whereas the other lacks this specificity (Tsutsui and Muller, 1988). The first class of binding sites in the nuclear matrix selectively binds supercoiled DNA without sequence specificity. This site was prepared by the mild extraction procedure of Mirkovitch et al. (1984) using LIS or nuclear halo with DNase I digestion. The scaffold associated DNA, after micrococcal nuclease digestion, has less DNA than that digested by DNase I. Therefore, it exposed additional DNA-binding sites in the nuclear scaffolds that are independent of ligand conformation. Consequently, it binds to relaxed, linear, and supercoild pDNA (Tsutsui and Muller, 1988). Like MAR associated pDNA, these topoisoform pDNA competed with single stranded pDNA (Kay and Bode, 1994; Tsutsui and Muller, 1988).











to 27


4 (1












X31400 0.90 pm X90200 0.20m

Figure 2-1: Nuclear matrix isolated by amine modification, N-hydroxysulfosuccinimide acetate (sulfo-NHS), method. Chinese hamster ovary cells were washed twice with cold PBS and extracted in cytoskeleton buffer (10 mM Pipes, pH 6.8/100 mM NaCl/300 mM sucrose/3 mM MaCl2/1 mM EGTA/1jag/ml leupeptin/1 jag/mL pepstatin/2 [tg/mL aprotinin/1 jag/mL antipain/1 mM aminoethyl benzensulfonyl fluoride/10 units/mL prime RNase inhibitor) containing 0.5 % Triton X-100 for 7 min at 4oC to remove the soluble proteins. After extraction, cells were treated with 600 units/mL each of the restriction enzymes (PstI and Haelll) in cytoskeletal buffer at 32 'C for 1 hr. Then extracted cells were exposed to 2 mg/mL freshly prepared sulfo-NHS-acetate (in cytoskeletal buffer, pH
7.0) for 20 min at room temperature. After being washed, cells were again treated with 2 mg/mL of sulfo-NHS for 20 min at room temperature. Cells were then washed with 10 mM glycine to quench the excess blocking reagent. The final pellet were resuspend in cytoskeleton buffer.






22
The binding of MARs-containing pDNA with nuclear matrix involves MAR

protein. To date, several matrix proteins that specifically interact with MAR in vitro have been identified. Some proteins, such as lamin A/B, bind with supercoiled and singlestranded MAR DNA, while some proteins, such as SATB 1, have a high affinity for supercoiled MAR DNA.

Furthermore, proteins such as Histone H1 have no preference when binding with single-stranded. Binding of the matrix proteins with DNA demonstrated the function of MAR, which is involved in active and inactive chromatin (Luderus et al., 1994).

The preliminary data showed that, in addition to the supercoiled form, the open circular and linear forms of non-MAR pDNA were also transcriptionally active. Studies relating to the stability of pDNA and the amount of pDNA associated with the cells were performed. As previously mentioned, the binding of pDNA to the nuclear matrix may be an important factor leading to the difference in transgene expression observed in DNA topoisoforms.













CHAPTER 3
RELATIONSHIP BETWEEN PLASMID DNA TOPOLOGICAL FORMS AND IN VITRO TRANSFECTION



Introduction


Plasmid DNA (pDNA) is the basic tool for nonviral gene delivery.

Understanding the conformational state of pDNA could assist in developing approaches to increasing its transfection efficiency. Important aspects for consideration include size and topology. The size limit for passive diffusion of linear pDNA into the nucleus was found to be between 200 and 310 base pair (bp); DNA of 310-1500 bp required energy to enter the nucleus (Ludtke et al., 1999). While smaller DNA molecules have higher transfection efficiency than the large one, the topology of the transfected DNA molecules determines the level of gene expression (Kreiss et al., 1999; Ludtke et al., 1999; Pitard et al., 1997). The supercoiled form of pDNA was shown to produce a higher level of transgene expression than nicked circular or linear DNA (Hirose et al., 1985). Moreover, the amount of supercoiled form in the preparation is indicative of the stability and activity of pDNA preparation. However, during DNA isolation or the formulation process, DNA can be altered by shear stress to convert the supercoiled plasmid form to open circular (single stranded nicked), linear, or even the small fragment (Adami et al., 1998). This conformation change may affect the transfection efficiency of pDNA.

Gel electrophoresis is one of the most direct and sensitive approaches for

examining plasmid DNA stability in cell lysate solution, rat plasma, and serum (Houk et 23






24
al., 1999; Lodish et al., 1995). This technique separates plasmid DNA on the basis of size and compactness; smaller and/or more compact molecules will migrate more rapidly through the matrix of the gel (Bates and Maxwell, 1993).

The objective of this study was to investigate the relationship between the topological form of pDNA and its transgene production. Since size was one of the factors affecting transfection efficiency, similarly sized pDNAs were utilized. Plasmid pCMV (6.7 kb) is driven by cytomegalovirus promoter, while pGL3 (5.2 kb) is driven by Simian virus (SV40) promoter. Because of its higher affinity to RNA polymerase, the cytomegalovirus promoter is stronger than the SV40 promoter. In order to account for differences in cellular uptake of pDNA, we utilized the flow cytometry method to measure the amount of pDNA associated with the cells. The stability of pDNA in cytoplasmic fraction was determined. These relationships are important for the future development of nonviral gene transfer.



Materials and Methods


Plasmid DNAs

In this study we used pDNA, pGL3 control, GeneBank accession number U47296 (Promega, Madison, WI) and pWiz or pCMV-luciferase (Gene Therapy System, San Diego, CA), which differ in promoter sequences and size. Supercoiled pGL3 (5256 bp) containing SV40 promoter was linearized with BamH I (Promega, Madison, WI). Supercoiled pCMV-luciferase (6732 bp) that contains cytomegalovirus (CMV) promoter was linearized with XmnI (Promega, Madison, WI). These enzymes cleaved the bacteria part, which is not necessary to the expression of the plasmid. One hundred units of






25
enzyme were used to cleave 88 [tg of pDNA at 370C for 4 h. The enzymatic reaction was then extracted with phenol/chloroform/isoamyl alcohol (25:24:1 v/v). Linear DNA was purified by ethanol precipitation at -200C followed by centrifugation at 13,000 xg for 20 min at 40C. The pDNA was analyzed using 0.8% agarose gel containing 0.5 mg/mL ethidium bromide. Open circular pDNA was prepared by creating single-strand nicks in supercoiled DNA. One mg of pDNA was incubated in 1 mL of Tris-EDTA buffer at 700C for 6 hours and was concentrated via ethanol precipitation as previously reported by Adami and coworker (Adami et al., 1998). Liposome Preparation

DOTAP/DOPE (1,2-dioleoyl-3-trimethylammonium-propane/L-dioleoyl

phosphatidyl-ethanolamine) liposomes were prepared as previously described by Tang and Hughes (1998). Briefly, DOTAP and DOPE were dissolved and mixed in 1:1 molar ratio in chloroform. The solution was evaporated in a round-bottomed flask using a rotary evaporator at room temperature. The lipid film was dried using nitrogen for an additional ten minutes. The lipid was then suspended in sterile water to make a concentration of 1 mg/mL. The mixtures were shaken for 30 minutes, followed by sonication by using Sonic Dismemebrator (Fisher Scienctific, Pittsburgh, PA) for 5 minutes at 5 watts to form homogenized liposomes. The particle size distribution of the liposomes was measured using a NICOMP 380 ZLS instrument (Santa Babara, CA) with the volume-weight distribution parameter. The approximate mean diameter of these liposomes was 200 20 nm. In order to study the interaction of pDNA and cationic liposomes, pDNA was extracted by phenol/chloroform immediately after the complexation with liposomes followed by the gel retardation study. The ratio of






26
pDNA:liposome was 1:2 w/w. There was no conformation change of pDNA after liposomes complexation (data not shown), indicating that complexation of pDNA with liposomes did not alter pDNA topoisoform. Transfection Experiment

Chinese Hamster Ovary (CHO) or human neuroblastoma (SKnSH ) cells were grown in cc-minimum essential and RPMI 1640 media (Gibco BRL), respectively, supplemented with 10% fetal bovine serum, penicillin (100 units/mL), and streptomycin (100 ptg/mL). All cells were maintained in humidified air at 370C and 5% CO2. Cell lines were cultured and seeded in 24-well plates (1 x 105 cell/well) and grown to 60 to 80 % confluence in 1 mL of media. Each topoisomer pDNA was complexed with cationic liposomes, DOTAP/DOPE, at a 1:2 w/w ratio of pDNA:1iposomes for 30 min in serumfree media. Before transfection, serum-containing media was changed to serum free media and the transfection mixtures added. After 4 hours, the media were changed to growth media containing serum and the cells were grown for another 48 hours. For electroporation, CHO and SKnSH cells were grown to 80% confluence in a 175-mL flask in 25 mL media and harvested using typsin/EDTA. Cells were washed once with phosphate buffer saline (PBS) and then twice with electroporation media (10 mM HEPES, pH 7.2, 150 mM NaCl, 5 mM CaCl2). The electroporation was performed using Gene Pulser II (Biorad, Hercules, CA) with 450 volts, 350 F, and 20 Q. The cell concentration for electroporation was 3x106 cells/pulse and pDNA wasl0 pg/pulse. Transgene production was determined by measuring luciferase expression after 48 hours of incubation for both methods as previously described (Tang and Hughes, 1998). Briefly, cells were rinsed twice with PBS then 100 pl of luciferase lysis buffer (0.1 M






27
potassium phosphate buffer, pH 7.8, 2 mM EDTA, 1% Triton X-100, 1 mM DTT) was added to the cells. Luciferase activity was quantified by using luciferase assay buffer (30 mM Tricine, 3 mM ATP, 15 mM MgSO4, 10 mM DDT, pH 7.8) and 1 mM D-luciferin (pH 6.3) (Molecular Probes, Eugene, OR) combined with cell lysate. The light emission over a 10-second reaction period was integrated using a luminometer (Monolight 2010, San Diego, CA). This experiment was done 3 times with 4 replicates each time. Luciferase expression was reported in RLU/well under the assumption of non-toxic protein.

Cytoplasmic Stability Study

Cytoplasmic fractions were obtained from CHO cells (1 x 106 cells) via treatment with a solution containing lysis buffer (150 mM NaC1, 10 mM Tris-HCl pH 7.9, 1.5 mM MgCl2, and 0.5% Nonident NP40) for 15 minutes (Piva et al., 1998). The ratio of the cell/solution was 1 x 106 cells/mL. After centrifuging at 10,000 xg for 10 min, a supernatant representing crude cytoplasmic fraction was obtained. This contained cytosol with different membranes and components originating from dissolution of such cellular vesicles as endosomes, lysosomes, and small-sized organelles. Forty Vtg of pDNA was incubated at 370C in 1 mL of the cytoplasmic fraction. After incubation for an appropriate time, all samples were phenol/chloroform extracted. The topological state of DNA was determined by gel electrophoresis. Contamination of the cytosolic extract with lysosomal enzymes was assessed by measuring the P-galactosidase activity of the extract. The enzyme was determined fluorometrically as previously described (Storrie and Madden, 1990). The assays showed that 11 5% of the luminal content of lysosomes was released into the extracellular medium, indicating minimal organelle contamination.







28
Cytoplasmic fractions of whole cells obtained from lysis buffer (0.1 M potassium phosphate buffer, pH. 7.8, 1% Triton X-100, 1 mM DTT, and 2 mM EDTA) showed that 24 2% of the luminal content of lysosomes was released into the extracellular medium. Therefore, the stability of pDNA in cytosolic fraction was mainly affected by the nuclease activity associated with the cytosol. Cellular Uptake Study of Plasmid DNA

Plasmid DNA was covalently labeled with fluorescein using an IT nucleic acid labeling kit (Pan Vera Corp., Madison, WI) according to the manufacturer's instructions. The labeled pDNAs were purified twice on microspin columns, then ethanol-precipitated. The beginning amount of labeled pDNA was standardized to the same fluorescence intensity. In the supercoiled preparation, a significant conformation change from supercoiled to open circular after fluorescein labeled was found. Therefore, this preparation was excluded from the experiment. The labeled pDNA:1iposome (DOTAP/DOPE with 1:2 w/w ratio of pDNA/liposome) complexes were transfected into CHO cells (2 x 105 cell/well) as described in the transfection section. After 48 hours of incubation, the cells were harvested using 0.05% trypsin/0.53 mM EDTA, then transferred to 1.5 mL tubes, centrifuged at 250 xg for 5 min, washed twice and resuspended in 800 pl of PBS. Flow cytometry was performed using a FACSort (Beckton, Dickinson, San Jose, CA). Fluorescein was monitored with a 530/30 bandpass filter, and photomultiplier tube pulses were amplified logarithmically. Ten thousand cells were counted at a flow rate between 100 and 200 cells per second. Cells were gated with their morphological properties, forward scatter, and side scatter, set on linear mode. The mean fluorescence intensity of the related populations of cells was calculated using






29
histograms and expressed in arbitary units corresponding to an intensity channel number ranging from 0 to 1023.

Data Analysis

Gel analysis was performed using Kodak 1D40 image analysis software version 3.0 (Rochester, NY). The data obtained from gel analysis was fitted for kinetic study to estimate the kinetic parameter of pDNA in cell lysate. Data were analyzed using the Micromath' Scientist program, version 2.0 (Salt Lake City, Utah). Statistical analysis was performed using StatView software version 4.5 (Abacus Software, Berkeley, CA). The ANOVA factorial test was used in this analysis with a 95% confidence interval. The data were considered to be significantly different when the p-value was < 0.05. Background

From the stability study of pDNA in a rat plasma model by Houk and coworker

(Houk et al., 1999), the degradation of supercoiled pDNA was assumed to follow pseudofirst-order kinetics, as shown in Figure 3-1.

The assumptions for this model are, first, pDNA degradation is considered to be a unidirectional process. Second, the degradation of linear pDNA is considered to yield fragments of heterogeneous lengths, thus the degradation products from the linear pDNA were excluded from the fitted model. Finally, no elimination from any of the compartments is assumed to occur through routes other than degradation to the subsequent topoisoform.


ks ko k
SC 1 OC 10 Ln a

Figure 3-1: Diagram of first-order kinetics (Houk et al., 1999).





30
Based on this model, the following differential equations were derived to describe the process in Figure 3-1 (Houk et al., 1999)

dSC
d C= -k.SC, (3-1)
dt

dOC k.SC-k.OC, (3-2)
dt

dL
= k OC-kL. (3-3)
dt

The amounts of supercoiled, open circular, and linear pDNA were then modeled using the integrated form of the following equations:

SC= SCo.e-kt, (3-4)


OC = kS.SCO. k .e + -k,,t +kk e-kt, (3-5)

(kkkk -kl)

L = koks.SCo .e -k "' + .-k t e- k
(ks koXki ko) (k -k,Xki -k,)* + (k, -k,Xks -k,) (3-6)

where SC, OC, and L are the amounts of supercoiled, open circular, and linear pDNA present at time t, respectively. SCo is the amount of supercoiled pDNA present at time (t) = 0. The constants ks ,ko, and k, represent the rate constants for the degradation of supercoiled, open circular, and linear pDNA, respectively. The constants represent the activity of all enzymes acting in the degradation process. Nonlinear curve-fitting and statistical analysis were carried out using Scientist (version 4.0, Micromath, Salt Lake City, UT).







31
Results


Figure 3-2 shows photographs of agarose gel electrophoresis of pDNA. Plasmid pCMV-luciferase and pGL3, heated at 70'C for 6 h, resulted in approximately 70% relaxed circular form for each. Supercoiled plasmid DNA treated with restriction enzyme demonstrated almost 100% linear form.

Plasmid DNA starting with aliquots of 70% supercoiled (SC) pCMV, 80% SC pGL3, 70% open circular (OC) pDNA, and 100% linear (Ln) pDNA, was delivered into CHO and SKnSH cells using either cationic liposomes or electroporation. The percentage of each topoisoform was calculated based on the binding affinity of each topoisoform to ethidium bromide in gel analysis. CHO and SKnSH cells gave similar results in transfection efficiency (TE) of the topoisoform, although the extent of the expression was cell-line dependent.

Plasmid DNA delivered by pDNA/liposomes, open circular pCMV showed

significantly higher transfection efficiency than the supercoiled or linear forms (Figure 33). There was a significant difference in TE of the topoisoform of pGL3 in CHO cell, whereas there was no significant difference in those of pGL3 in SKnSH cell (Figure 3-3). Plasmid DNA delivered by electroporation shows only a slightly difference in TE between the cell lines.

In CHO cells, there was no significant difference between TE of supercoiled and TE of open circular pCMV (Figure 3-4), whereas in SKnSH cells, TE was significantly higher in supercoiled pCMV than in open circular form. There was no significant difference in TE between the topoisoform of pGL3 for either cell line (Figure 3-4).







32


A. B.






Ln Ln
4 SC












STD 1 2 3 STD 1 2 3

Figure 3-2: Agarose gel electrophoresis of plasmid DNA. (A) topoisomer of pCMVluciferase: Lane 1, supercoiled preparation contained 70% supercoiled (SC) pCMVluciferase; Lane 2, open circular preparation contained 80% of open circular (OC) pCMV-luciferase; and Lane 3, linear preparation contained almost 100% linear (Ln) pCMV-luciferase. (B) topoisomer of pGL3: Lane 1, supercoiled preparation contained 80% of supercoiled pGL3; Lane 2, open circular preparation contained 65% of open circular pGL3; and Lane3, linear preparation contained almost 100% of linear pGL3. STD = Standard Lambda Hind III DNA marker.






33


A.



4 16
.X 14 > 12
10 T
EpCMV
8 6
4
.-2
.J0
SC 0C Ln


B.



25

0o 20

S_15j 5W

10- 0 pGL3

5

--0
SC OC Ln


Figure 3-3: Transfection efficiency of pDNA-liposome complex (1:2 w/w ratio) in (A) Chinese hamster ovary (CHO) cells and (B) Human neuroblastoma (SKnSH) cells. Transfection efficiency of liposome complex supercoiled (SC), complex open circular form (OC), and complex linear form (Ln) are presented. N = 4. The error bars represent mean S.D. (A) CHO: There was a significantly difference (p<0.05) in each topoisoform of pDNA. (B) SKnSH: represented significant difference in transfection efficiency
(TE) of OC and the other forms of pCMV-luciferase. There was no significant difference between the TE of each topoisoform of pGL3. Transfection efficiency is represented by the relative light unit (RLU)/well.







34
A.



6


4 -nCM1V
3 OpGL3



0
SC OC Ln




B.




10 *



6 PCMV
L~ 4EllEpGL3
4



0
SC OC Ln

Figure 3-4: Transfection efficiency of pDNAs by using electroporation (450 volt, 350 [tF, 20 Q) in (A) Chinese hamster ovary (CHO) cells and (B) Human neuroblastoma (SKnSH) cells in relative light unit (RLU)/well. Transfection efficiency of supercoiled
(SC), open circular (OC), and linear (Ln) pDNA is presented. N = 3. The error bars represent mean S.D. (A) CHO cell, represents the significant difference between OC and Ln pCMV-luciferase. No significant difference between transfection efficiency of SC and OC pCMV-luciferase and topology of pGL3 was observed. (B) SKnSH cell, represents the significant difference between SC and the other topoisoforms of pCMV.







35
For transfection methods, the ratio of pDNA:number of cells for liposomes

complex and electroporation was 1:0.3x1 05 and 1:3x105, respectively. The number of cells used for electroporation was 10 times higher than for liposomes transfection because the pulse electric field killed a large portion of cells.

Stability of plasmid DNA in the isolated cytoplasm was also addressed. Plasmid DNA incubated at 370C in cell lysate solution exhibited degradation of pDNA during a set period of time. Both pCMV and pGL3 topoisomers showed similar patterns of degradation (Figure 3-5). Supercoiled pDNA could be detected for up to 4 min, while open circular and linear form remained for 6 to 10 min. The kinetics of pDNA in cell lysate solution exhibited the pseudo-first-order of degradation at an initial amount of 40 jig pDNA. Table 3-1 gives the rate constants and half-lives of the topoisomers of pDNA.

In order to examine possible reasons for differences in expression, studies were conducted to determine whether similar amounts of DNA complexes were associated with the cells. CHO cells were incubated with fluorescein-labeled pDNA (OC or Ln) lipid complexes for 48 h at 370C, harvested with trypsin/EDTA and washed with PBS, as described previously above. Figure 3-6 shows the mean fluorescence intensity of the labeled pDNA in cells. Open circular pDNA showed higher intensity, indicating complexes made with this form of pDNA associated more with cells than linear forms of pDNA at 48 hours after transfection.


Discussion and Conclusion


CHO and SKnSH cells transfected with pDNA/liposome complexes demonstrated similarities in pattern but differed in levels of transgene produced (Figure 3-3).























Std. 0 5 10 15 20 25 30 35 40 45 min Std. 0 5 10 15 20 25 30 35 40 45 min Std. 0 5 10 15 20 25 30 35 40 45 min

Supercoiled pCMV Open circular pCMV Linear pCMV













Std 0 5 10 15 20 25 30 35 min Std 0 5 10 15 20 25 min Std 0 5 10 15 20 25 30 35 min

Supercoiled pGL3 Open circular pGL3 I Linear pGL3


Figure 3-5: Agarose gel electrophoresis of plasmid DNAs incubated in cytoplasmic solution at 370C in the period of time. (A) pCMV-luciferase topoisoform, (B) pGL3 topoisoform.






37



Table 3-1: Pharmacokinetics of plasmid DNAs in cytoplasmic solution.
Plasmid DNA Rate constant (min-) Half life (min)

pCMV Supercoiled 0.20 0.04 3.49 0.66

Open circular 0.11 0.02 6.52 1.65

Linear 0.11 0.02 6.19 2.12

pGL3 Supercoiled 0.19 0.03 3.57 0.65

Open circular 0.06 0.01 10.90 1.89 *

Linear 0.14 0.07 4.85 3.01

Kinetics of plasmid DNAs in cytoplasmic solution calculated using the Micromathe Scientist program. Represented the significant difference between the half-life of OC and other topoisoforms of pGL3.







38


700

600

500

400
K pCMV 30 pGL3 300-]

0 200

10:

0
cell 0C Ln

Figure 3-6: Mean fluorescence of the labeled plasmid DNAs (OC open circular and Ln linear forms) complex with DOTAP/DOPE liposome (1:2 w/w ratio) transfected into CHO cell. The cell was analyzed by flow cytometry, FACSort after 48 h of incubation at 37oC. The number of replication, N, is 4. The error bars represent mean S.D. represented the significantly different between OC and Ln forms of pCMV. No significantly different between mean fluorescence intensity of topology of pGL3.






39

This indicated the cell type was important in transfection efficiency (TE). SKnSH cells expressed lower amounts of luciferase activity than CHO cells, especially in SV40 promoter plasmid compared to the stronger CMV promoter plasmid. Researchers have previously shown that cells may have different levels and types of enzymes (Piva et al., 1998) involved in the pDNA process leading to alter transgene production. This might be one of the reasons for the similar patterns but varying levels of expression in the two cell lines.

Two important findings concerning pDNA/liposome complex were evident. First, open circular and linear pDNA were as active as supercoiled form, and open circular pCMV showed a higher transgene production than supercoiled form.

The import steps to getting pDNA into the nucleus include, 1) pDNA entering into cell, 2) movement through cytoplasm, and 3) pDNA entering into nucleus via nuclear pore complex. In the first step, pDNA delivered by liposome enters the cell by endocytosis (Gao and Huang, 1995; Xie et al., 1992), whereas pDNA delivered by electroporation enters by transient membrane pores (Courey and Wang, 1983; Weaver, 1993). Neither of these methods is thought to be topologically dependent (Chernomordik et al., 1990; Weaver, 1993; Xie and Tsong, 1993). Extrapolating from our data, 48 hours after transfection OC appeard to associate in cells more than does Ln pDNA (Figure 3-6). Plasmid DNA delivered by electroporation showed a pattern of expression among each topoisoform similar to pDNA delivered by liposomes, but the forms differed in the extent (Figures 3-4). Although the correction was made for the number of cells dying from electrical shock, it is possible that the initial amount of pDNA entering the cell might be different for each delivery methods. Moreover, the low transfection efficiency for







40
electroporation might be due to the rapidly shrinking of the transient pore size. The difference in amount might affect the level of expression, and the excess of pDNA in cytoplasm might saturate enzyme activity.

Stability of pDNA in cytoplasm is an important aspect in Step 2 of the pathway. Plasmid DNA has been reported to be destroyed mainly by cytosolic nucleases, including endonuclease and exonuclease (Lechardeur et al., 1999). The identification of which cytosolic nuclease(s) is or are involved in pDNA degradation needs to be determined in order to understand this mechanism. Studies by Lechardeur and coworker (1999), excluded DNasel and DNasell, which display an acidic pH optimum, from the degradation of pDNA in cytosol, since the maximum nuclease activity was attained between pH 7 and PH 8. Our data demonstrated that the OC form exhibited the longest half-life in the isolated cytoplasm of CHO cells, followed by Ln and then SC form (Table 3-1). The stability studies of pDNA in the rat plasma model showed OC form had the longest half-life (21 min), followed by Ln (11 min). The authors' data support previous results that indicated the SC form was the least stable form of pDNA in the rat plasma model (half-life of 1.2 min) (Houk et al., 1999). Most likely there is a plasma enzyme involved in the degradation of pDNA other than cytosolic endonuclease, and there might even be a special enzyme sequence in SC form that results in faster degradation. The movement of pDNA in cytoplasm through the cytoskeleton and the immobile object is another barrier for pDNA trafficking (Wilson et al., 1999). Fully condensed pDNA, the SC form, may have a higher binding affinity with these elements, thus increasing the chance of being destroyed by cytosolic enzyme. It is not easy to explain the differences in half-life between the OC forms found in pCMV and pGL3. These may be due to the






41
different attraction of nuclease enzymes or to the method of production of OC DNA, resulting in its being nicked in difference places. Because of its larger size, OC pCMV might be more susceptible to enzyme activity than OC pGL3. In pDNA/liposomes complexes, the patterns of expression were not relative to the stability of pDNA in cytosol (Figure 3-3), probably due to the protection provided by lipids to the pDNA (Lechardeur et al., 1999).

The final step is when pDNA enters the nucleus. Plasmid DNA nuclear import is facilitated by cell division. Therefore, nuclear import might not be a major barrier in transgene expression in our case, since both CHO and SKnSH cells are dividing during import. Although SV40 promoter has been reported previously as responsible for facilitating the specific sequence that enhanced nuclear import (Wilson et al., 1999), we noticed that CMV-based pDNA demonstrated higher expression than SV40-based DNA.

The superhelical (supercoiled) state of pDNA was considered to be important in transcription process in eukaryotes (Gellert, 1981; Hirose and Ohba, 1993; Hirose et al., 1985; Mizutani et al., 1991; Shlyakhtenko et al., 1998; Singhal and Huang, 1994; Wang, 1985). Our results demonstrated that compared to the SC form of each pDNA, OC pCMV is still active, while OC pGL3 was not active (Figure 3-3). The reason may be that the different promoter sequences lead to different levels of transcription initiation. Although the SC form facilitated the formation of the transcription preinitiation complex within the promoter area and prevented it from being included in heterochromatin, which causes the silencing of the gene (Singhal and Huang, 1994), in some genes, such as the Drosophila Hsp70 gene, closed circular DNA had the same rate of preinitiation complex formation as supercoiled DNA (Singhal and Huang, 1994), resulting in the same protein






42
expression. In our case, OC CMV might have facilitated the local unwinding of DNA without the help of the SC form, resulting in the same or higher amounts of luciferease activity. However the reason why OC pCMV delivered in the same amount as SC pDNA showed higher transgene expression than the SC form is still unclear. A possible reason might be the loss of the SC form in cytoplasm (Table 3-1). The second method of making OC DNA by random nicking may affect the efficiency of pDNA. If the singlestranded nick occured in a crucial place, the missing base pairs in OC pDNA might impair its function. This might be another reason to explain the low activity of OC pGL3 (Figure 3-3).

In this study we concluded that regardless of the nonviral delivery method, pDNA topology would not effect transgene expression, even though the pharmacokinetics of each isofroms are distinct. While, the delivery methods (electroporation and liposomes) facilitate the transport of pDNA across cell membranes and protect pDNA from cyotplasmic enzyme (liposomes), other factors for improving transfection efficiency need to be considered. These factors include the formulation for pDNA preservation (pH, salt effect, and ionic strength), the quality of pDNA (impurity, presence endotoxin, and amount of each topoisoform), and the cell type used for transfection. Supercoiled DNA is the most accepted form of pDNA for gene transfer because it is the initial form of DNA and it is the initiation form of DNA in the transcription process. However, the stability of SC pDNA in the cytoplasm may hinder the level of expression. The other forms of pDNA, OC and Ln forms are more stable in cytoplasm and are transcriptionally active.













CHAPTER 4
USE OF PHARMACOKINETIC PARAMETERS TO INTERPRET GENE EXPRESSION



Introduction


Plasmid DNA (pDNA), either naked or complexed with cationic molecules, is a frequently used vector for gene delivery since it does not exhibit the limitations, such as eliciting adverse immune responses (Hengge et al., 2001), insertional mutagenesis (Hengge et al., 2001), and the size limitation of the transgene (Kreiss et al., 1999) that characterize viral vectors. Despite certain advantages for gene delivery of naked pDNA, there are fundamental problems associated with the use of unprotected pDNA. In particular, the reduction of genome equivalents by enzyme endonucleases will translate into a loss of gene expression (Hengge et al., 2001). A more efficient nonviral method of delivery must be used to better deliver the vector to target cells. These methods can be characterized as (a) chemical methods such as liposome delivery, receptor-mediate endocytosis, and long-term expression vector, and (b) physical methods such as injection, or particle bombardment (Wohlgemuth et al., 1996). Despite such sophisticated techniques, the development of an efficient delivery system also depends on accurate analytical methods.

In vitro transfection is a procedure in which tissue culture cells are incubated with a plasmid and, after a certain time, the expression of the encoded gene is measured via the enzyme activity or concentration of the synthesized protein (Lasic, 1997). Gene 43






44
expression can be measured as a function of time, but is often reported as a single value of peak expression. However, the process of gene regulation, including transcription, translation, and post-modification, requires time. Moreover, it also depends on time for DNA to be initialized by cells, for DNA trafficking through the cytoplasm to the nucleus, to half-life (the time necessary for the concentration of the drug or pDNA in the plasma or cells to decrease by one-half) of intracellular DNA and mRNA. Time also determines that phase of the cell cycle and the cell turnover rate. Various cell processes and conditions result in a time variation in expression. Thus, a single-point measurement may not be the best way to compare delivery systems and would be similar to relating therapeutic levels of a drug in the blood by using such a parameter as peak drug concentration (CMAX)Pharmacokinetics are the study of the time course of drug absorption, distribution, metabolism, and excretion (Dipiro et al., 1997). Pharmacokinetics have been utilized for characterization of pDNA in vivo (Lew et al., 1995; Mahato et al., 1995; Osaka et al., 1996; Schubeler et al., 1996) and the fate of pDNA in vitro (Houk et al., 2001). Gene transfection and gene expression of pDNA can also utilize pharmacokinetic concepts since transgene expression can be followed over time by treating pDNA as a prodrug and measuring its active methabolites. From the plot of transgene expression against time, pharmacokinetic parameters for gene expression can be calculated that would be equivalent to the standard parameters of area under the curve (AUC) and mean residence time (MRT). We have used noncompartmental methods for estimation of certain pharmacokinetic parameters under the assumptions of linear pharmacokinetics with firstorder elimination and stable protein product. AUC, or, in the case of gene expression,







45
area under the expression curve (AUEC), can indicate the total amount of protein produced and mean expression time (MET) can indicate the average time of the expression. Thus, it can indicate not only the delivery of the genes into cells, the gene transfection, but also the amount of protein encoded by the delivered DNA, the gene expression (Lasic, 1997).

In this chapter, we address two cases of pDNAs-based delivery systems. The first is a simulation of transgene expression of four pDNAs with identical maximum expression (Cmax), but different times of maximum expression (tmax). The second is the experimental transgene expression of two plasmids coding for luciferase, but different in plasmid construct. Plasmid pGL3 contains SV40 promoter and an ampicillin-resistance gene. Plasmid pGM is pGL3 with an insertion of matrix attachment regions (MARs) that may enhance or prolong the expression. The ideal pDNA for gene delivery requires high and long-term expression. Thus, comparing plasmid transgene activity in both cases needs more than one point of measurement. In this paper, we utilized the multiple-point measurement together with pharmacokinetics to calculate gene expression for Case 1, (plasmids 1, 2, 3, and 4) and Case 2 (pGL3 and pGM). This method will allow comparing transfection efficiency of pDNA through the gene product and will provide more insight for data analysis.




Materials and Methods


Plasmid DNAs

Plasmid pGM was constructed from pGL3-Control vector (5.2 kb) (GeneBank

number U47296) (Promegao, Madison, WI), which contains a modified firefly luciferase







46
gene and SV 40 promoter. A 2.0 kb of MARs from 5'-region of hlFN-P (Piechaczek et al., 1999) was inserted at position 2442 downstream of the SV40 enhancer of pGL3 to create 7.2 kb MAR-containing vector (Figure 4-1). Plasmid DNA was obtained from E. coli (strain DH5cx) that was transformed with a pGL3 or pGM plasmid. Plasmid DNA was isolated using a Megawizard DNA purification kit (Promegao, Madison, WI). The concentration and purity of pDNAs were determined spectrophotometrically. The average concentration of pDNAs was 1.8 mg/ml and purity of pDNA, A260/A280 ratio, was 1.8.

Cells

Chinese hamster ovary (CHO) cells were obtained from American Type Culture Collection, MD. Cells were incubated at 370C in a humidified atmosphere containing 5% CO2 and maintained in Minimum Essential Medium-alpha (c-MEM) medium with 10% fetal bovine serum (FBS), 100 gg/mL of streptomycin, and 100 units/mL of penicillin (all from Gibco BRL, Grand Island, NY). At 24 hours before transfection, cells were seeded at 50,000 cells per well in 12-well plates (Coming Costar Corp., Cambridge, MA) in a final volume of 2 mL of medim.

Preparation of Cationic Liposomes and Transfection Experiment

DOTAP/DOPE (1,2-dioleoyl-3-trimethylammonium-propane/L-dioleoylphosphatidyl- ethanolamine) liposomes were prepared as previously described by Tang and Hughes and in Chapter 3 (Tang and Hughes, 1998). The approximate mean diameter of these liposomes was 200 80 nm. Equivalent molarx of pDNA, pGL3 (3ptg), or pGM (4.15 ptg) were complexed with cationic liposomes, DOTAP/DOPE, at a 1:2 w/w ratio of pDNA:1iposomes for 30 min in serum-free medim.










SV40 promoter SV40 promoter Luc
Arrp

Luc

pGL3 pGM
5256 bp 7217 bp



SV40 enhancer


SV40 enhancer M4Rs


Figure 4-1: Schematic representation of the plasmid DNAs vectors. (A) pGL3 5265 bp, (B) pGM 7217 bp. Abbreviations: SV40, Simian Virus 40; Luc, Luciferase coding region; MARs, matrix attachment regions; Amp, Ampicillin resistance gene.







48
Before transfection, serum-containing medium was changed to serum-free

medium and the transfection mixtures added. After 4 hours, the medium was changed to growth medium containing serum. Transgene production was determined by amount of luciferase expression as previously described (Tang and Hughes, 1998) from Day 1 to 5 posttranfection. Amount of 250 ptL of luciferase lysis buffer were utilized in this experiment. This experiment was done 3 times with 4 replicates each time. Luciferase expression was reported in RLU/well under assumption of non-toxic protein. Background

From a plot of luciferease expression in RLU/well and time, we used

noncompartmental methods to estimate the pharmacokinetic parameters, AUEC and MET. These methods were based on the estimation of the area under a plot of drug concentration versus time, while have been used previously for data analysis of biological systems (Gibaldi and Perrier, 1982). The statistical moments in classical pharmacokinetic are defined as follows:


AUC= Cdt, (4-1)
0


AUMC= ft-Cdt, (4-2)
0


ft.Cdt
MRT o AUMC (4-3)
AUC
JC dt
0

where AUC is area under the curve, which for transgene expression is equal to area under the expression curve (AUEC), and MRT is mean residence time and is equivalent to mean expression time (MET). The area under the curve of a plot of the product of






49
concentration and time (C-t) versus time from zero time to infinity is often referred to as the area under the first moment curve, AUMC (Gibaldi and Perrier, 1982).

Estimation of the AUC from zero time to infinity must be carried out in two steps. The AUC from zero time to some time, t (t*), is calculated by trapezoidal rule. To this partial area from the terminal portion to infinity, t** co, must be estimated as follows: X C*
AUC = C dt = (4-4)
t*

where 21 is the slope of the terminal exponential phase of a plot of natural log versus time and C* is the last measured point of expression. The sum of the two partial areas is AUC.

The same approach was used to estimate total AUMC. The area under the first moment curve from t* to infinity is estimated as follows in equation 4-5: t*-C* C*
AUMC= t.Cdt=- + (4-5)
2* 2

Calculations

The trapezoidal rule was utilized to determe the AUC from experimental data without integration. AUC can be calculated as the sum of its individual trapezoids:


AUCOst = o tI + 2 2 -t1)+ 2 (t3 -t2)+., (4-6)
2 2 2

AUCt- = (4-7)

where C is the luciferase activity in the cell and t is the time point of measurement. Co, CI, C2-. Ct is the expression of time points 0, 1, 2, .etc. and k is the terminal slope from the plot. The same approach has used for estimation AUMC:







50

AUMCo t = Co t *to tI + -I *j +C' t2(t2 ) 3 t2+C+
2 2 2
(4-8)

Ct-t Ct
AUMC t t (4-9)


Simulations

Simulations were performed using Scientist kinetic software package (Micromath, Salt Lake City, Utah). A one-compartment model with first-order expression degradation and first-order transgene expression was used:


E = I Ka (e-Ke -K 1t), (4-10)
(Ka Ke)

where E is the expression in RLU, I is concentration of plasmid in the inoculation medium, Ka is the rate constant for uptake and expression of the transgene, K, is the rate constant for expression degradation, and t is time.

The variable, I, was kept constant and the maximum expression,CMAX, were both kept constant by maintaining the Ka/Ke ratio (ratio = 2), but absolute values of Ka and K, varied. The parameter MET and AUEC were calculated as previously described in the calculations section. Time to maximal expression was calculated by

K
ln( a)
TMAX K (4-11)
(Ka Ke)

Statistical Analysis

The Student t-test with one-tail distribution and two-sample equal variance was used to compare pharmacokinetic parameters. Repeated-measures ANOVA was used to detect differences with respect to plasmid and time. Tukey's HSD was used for post-hoc






51
analysis when difference were detected. The data was considered to significantly different for p-value < 0.05.



Results


Figure 4-2 shows the simulation of transgene expression of four plasmids in the same cells and under the same conditions. The Cmax values were identical (2500 RLU/well). However tmax varied. The tmax of plasmids 1, 2, 3, and 4 were 0.46, 1.39,

2.77, and 6.93 days, respectively. Consequently, measurement of protein expression at one time point is not adequate for comparison of transfection efficiency of these plasmids. Kinetic parameters AUEC and MET (Table4-1) were distinct. The AUEC of plasmids 1, 2, 3, and 4 were 3256, 9974, 19985, and 49031 (RLU.day)/well, respectively, indicating various amounts of protein production. The MET of plasmids 1, 2, 3, and 4 are 1, 3, 6, 14.33, respectively, indicating a difference in average time to plasmid expression.

Table 4-2 shows raw data from luciferase expression of pGL3 and pGM from Day 1 to 10 posttranfection in Chinese Hamster Ovary (CHO) cells. Figure 4-3 is a plot of the expression based on the raw data in Table 4-2 against time point from Day 1 to 5 posttranfection. Peak expression is at Day 2, with transgene expression of 29.14 1.86 x 106 and 15.84 1.76 x 106 RLU/well of pGM and pGL3, respectively. Kinetic parameters AUEC and MET were calculated based on equations 6,7 and 8,9 respectively, described in the Material and Methods section of this chapter.

These parameters and the statistical analysis are shown in Table 4-3. The AUEC of pGM and pGL3 are 125.92 1.12 x 106 and 48.81 20.48 x 106 (RLU.day)/well,






52



3000 2500

2000 A

1500 1000 500



0 5 10 15 20 25
Day Post Transfeciton

Plasmid 1I Plasmid 2 Plasmid 3 P lasmid 4



Figure 4-2: Simulation of transgene expression of plasmid DNAs 1, 2, 3, and 4 in the same cells from Day 1 to 10 posttranfection. Transgene expression is represented in relatively light unit (RLU) per well.







53



Table 4-1: Pharmacokinetic parameters of simulated plasmid DNAs. Plasmid DNA AUEC MET Cmax tmax
RLU.day/well Day RLU/well Day

Plasmid 1 3256 1.00 2500 0.46

Plasmid 2 9974 3.00 2500 1.39

Plasmid 3 19985 6.00 2500 2.77

Plasmid 4 49031 14.33 2500 6.93


AUEC is area under the expression curve and MET is mean expression time. Cmax and tmax are maximum concentration and time at maximum concentration, respectively.







54



Table 4-2: Transgene expression of plasmid pGL3 and pGM in Chinese hamster ovary, CHO cells from Day 1 to 10 posttranfection.
Time (day) pGL3x106 pGMx106 p-value
RLU/well SEM RLU/well SEM

1 2.58 0.94 1.36 0.40 N/S

2 15.84 1.75 29.14 1.86 p<0.01

3 11.12 1.94 25.37 1.99 p<0.01

4 8.31 1.71 26.751.77 p<0.01

5 5.14 1.52 23.38 3.92 p<0.01

6 3.08 1.01 8.100.44 N/S

7 1.52 0.19 6.48 0.93 N/S

8 0.87 0.09 3.22 0.74 N/S

9 0.22 0.02 1.90 0.52 N/S

10 0.09 0.01 0.17 0.05 N/S

Transgene expression is represented by the relative light unit (RLU) per well.







55




35
30 G
25
20 pGL3
15 18pGM
10 5 0
1 2 3 4 5
Days Post Transfection

Figure 4-3: Transgene expression of plasmid DNAs, pGL3 and pGM, in Chinese Hamster Ovary, CHO cells from Day 1 to 5 posttranfection. N = 4. The error bars represent mean SEM (standard error of mean). represent significant difference with p<0.05. Transgene expression is represented by the relativley light unit (RLU) per well.






56



Table 4-3: Pharmacokinetic parameters of plasmid pGL3 and pGM. Plasmid DNA AUECx106 MET Cmaxx106 tmax
RLU.day/well Day RLU/well Day

pGL3 48.81 20.45 3.38 0.05 15.84 1.76 2

pGM 125.92 1.16 3.95 0.07 29.14 1.86 2

p-value 0.03 0.01
AUEC is area under the expression curve and MET is mean expression time. Cmax and tmax are maximum concentration and time at maximum concentration, respectively. Data were represented by mean SEM. (standard error of mean).






57
respectively, while MET of pGM and pGL3 are 3.95 0.07 and 3.38 0.05 days, respectively. From the statistical analysis, AUEC and MET for pGM are significantly higher than for pGL3.



Discussion and Conclusion


Pharmacokinetics have been utilized to characterize of pDNA in vivo and in vitro (Houk et al., 2001; Lew et al., 1995; Mahato et al., 1995; Osaka et al., 1996; Scheule and Cheng, 1996).

In this chapter, we demonstrate the utilization of the pharmacokinetic parameters, Area under the curve (AUC) and Mean residence time (MRT) to increase the accuracy of data analysis.

Generally, gene expression is determined by a single-point measurement of

synthesis protein at a peak of expression. The limit of a single-point measurement are time variation due to cellular uptake of DNA, DNA trafficking to nucleus, half-life of DNA and mRNA, phase of the cell cycle, and cell turnover rate. Moreover, at only one time point, the heterogeneous cellular population, the mixture of healthy and unhealthy populations at a particular time, results in an inaccurate reporter-protein measurement. Multiple-point measurements of the transgene expressed may be an alternative way.

Total amount of the expressed protein (AUEC) and average time of expression

(MET) were obtained from a plot of transgene expression and time. The AUEC, which is equal to AUC, and MET, which is equal to MRT, demonstrated total amount of protein production and average time of expression, respectively. Figure 4-2 is the simulation of four pDNAs that had maximum protein at various time points. Although they have the






58
same maximum concentration (Cmax), time at maximum concentration (tmax) is different. Comparing transgene activity of these plasmids requires multiple-point measurements. AUEC and MET are calculated by a plot of protein expression in RLU and time. Table 4-1 showed kinetic parameters obtained from the plot. Although plasmid 1 reached maximum concentration first (tmax = 0.46 day) because it expressed the lowest amount of protein (AUEC = 3256 RLU/well) and persisted in cells only 1 day, it was not the palsmid with the highest transgene activity. An ideal pDNA would require high and long-term expression. Plasmid 4 is close to the idea with an AUEC of 49031 RLU/well and MET of 14.33 days, followed by plasmid 2 (AUEC = 19985 RLU/well, MET = 6 days), and plasmid 3 (AUEC = 9974 RLU/well, MET = 3 days). Thus, due to having the lowest AUEC and MET, plasmid 1 is considered to be the lowest quality.

In another example, we transfected two plasmids, pGL3 and pGM, into CHO cells and measured luciferease expression at various time points. Plasmid pGM has the insertion sequences of matrix attachment regions (MARs) from human P-interferon, which may prolong or enhance expression compared to the non-MARs plasmid, pGL3. In such an experiment, single-point measurements of transgene expression may not be adequate to evaluate the hypothesis or to compare plasmid transgene activities since there might be a time-point variation in the expression. For example, both plasmids may have similar maximums in expression, but if the maintenance of this expression differs between plasmids, a single-time-point parameter would not describe this. Comparing gene products provides a better vision of gene activity. Therefore, the combination of multiple-point measurements of transgene expression and non-compartmental analysis is useful. From the raw data reported in Table 4-2, the plot of protein expression and time






59
point can be obtained (Figure 4-3). Kinetic parameters in Table 4-3 are calculated based on raw data and equation 6-9 in Materials and Methods. The AUEC is 2 to 3-fold higher for pGM than for pGL3 indicating that pGM expressed more protein than pGL3. Moreover, MET of pGM is significantly higher than for pGL3, indicating a longer time of expression. In conclusion, MAR-containing plasmid DNA, pGM, can enhance (from AUEC) and prolong (from MET) the transgene expression compared to non-MARs pDNA.

The multiple-point measurements combined with kinetic analysis are simple to perform, do not require complex experiments, and better describe the expression of the reporter gene. This novel application of pharmacokinetics can aid in screening therapeutic plasmids and vectors.













CHAPTER 5
MATRIX ATTACHMENT REGIONS-CONTAINING PLASMID DNA INCREASES GENE EXPRESSION IN VITRO



Introduction


As discussed in Chapter 2, matrix attachment regions (MARs) are associated with a variety of biological functions such as gene replication, regulation, and repairing of DNA (Agarwal et al., 1998; Alvarez et al., 2000; Malone et al., 2000; Whitelaw et al., 2000). MARs might also be related to long-term transgene expression in vitro (Baiker et al., 2000; Piechaczek et al., 1999).

MARs improve transgene expression in eukaryotic cells, but only when those

MARs are integrated into the genome. MARs sequences demonstrated a neutral or even negative effect on expression in transiently transfected cells (Bode and Maass, 1988; Kalos and Fournier, 1995; Klehr et al., 1992; Poljak et al., 1994; Wang et al., 1996). However, one study showed that human interferon-P (hIFN-P) MARs-containing pDNA in stable transfection replicated episomally and maintained extrachromosomally in nucleus of Chinese hamster ovary (CHO) albeit low copy number (Piechaczek et al., 1999).

This evidence raises the question whether this nucleic acid sequence would enhance and/or prolong expression in vitro using a transient expression system. Our working hypothesis is MARs-containing pDNA increases and prolongs expression in various cell types compared to non-MARs-containing pDNA. We chose CHO cells in 60






61
this study because it was previously reported to increase replication and retention of MARs-containing pDNA (Piechaczek et al., 1999). Human neuroblastoma (SKnSH) cells were another example of fast-dividng cell used in this experiment. We also investigated the effect of MARs-containing pDNA in primary neuron, astroglia, and microglia, because of the dysfunction of these neuronal cells in neurological diseases (e.g., Alzheimer's, Parkinson's). Increasing the persistence of a therapeutics gene is potentially another method to treat the dysfunction of neuronal cells and aid in the therapy for patients with neurological disease. To answer the question whether or not MARs-containing pDNA increases expression of another pDNA, a cotransfection experiment was performed.

MARs are chromatin-remodeling elements that enhance chromatin accessibility (Fernandez et al., 2001; Forrester et al., 1994; Jenuwein et al., 1997). Nucleosome alteration also could be induced by other chromatin remodeling agents, histone deacetylase inhibitors (HDAC), for trichostatin A (TSA), trapoxin, n-butyrate (Yoshida et al., 1990). MARs and HDAC influence chromatin remodeling, we investigated the effect of HDAC inhibitor, TSA, on the action of MARs-containing pDNA. Moreover, polymerase chain reaction (PCR) has been performed to monitor intracellular distribution of pDNA.



Materials and Methods


Plasmid DNAs

Plasmid pGM and pGL3 were used in this experiment. Plasmid DNAs sequences and preparations were previously reported in Chapter 4 in Materials and Methods.






62
Plasmid pEPI-1 was a generous gift from Dr. Lipps from Germany. The function elements of pEPI-1 are the pUC origin of replication, 2.0 kb MARs from 5'-region of the human interferon-3 (hIFN-P) gene, the promoter of the bacterial ampicillin resistence gene, SV 40 promoter, SV40 origin of replication, and the kanamycin resistance gene (Piechaczek et al., 1999).

Cell Culture and Conditions

All media in cell culture were obtained from Gibco BRL (Grand Island, NY). FBS is fetal bovine serum. PS is 5000 U/ml penicillin, 5 mg/ml streptomycin. Several cell types (50% confluence) were grown under the following conditions:

* CHO (Chinese hamster ovary): Minimum essential medium alpha medium, 10% FBS,
1% PS, at 5% CO2

* SKnSH (Human neuroblastoma): RPMI medium 1640, 10% FBS, 1% PS, at 5% CO2

* Rat hippocampal primary neuron: Neurobasal media, 2% 50X B-27, 0.25% 2 M Lglutamine, 1% PS, at 10%CO2

* Rat astroglia and microglia: Dulbecco's modified eagle medium, 10% FBS, 1% PS, at
10% CO2

Cationic Liposome Preparations

DOGSDSO (1', 2'-dioleoyl-sn-glycer-3'-succinyl-2-hydroxyethyl disulfide

ornithine conjugate), which contains a disulfide bond between the positively charged head group and lipophilic backbone, was synthesized using the method described by Tang and Hughes (1998). This previous study demonstrated that DOGSDSO, a disulfidecontaining lipid was efficient at transfecting neuronal, astroglia, and microglia cell cultures (Ajmani et al., 1999). Another cationic lipid, DOTAP (1,2-dioleoyl-3trimethylammonium-propane), was used in transfection of CHO and SKnSH cells.






63
DOTAP and DOPE (L-dioleoyl-phosphatidyl- ethanolamine) were purchased from Avanti Polar Lipids (Alabaster, AL). DOTAP/DOPE and DOGSDSO/DOPE liposomes were prepared as described in Chapter 3 by Tang and Hughes (1998). The final cationic liposome concentration was 1 mg/ml. The average diameter of each of the two liposomal preparations was 200 80 nm. In vitro transfection

CHO and SKnSH cells were seeded at 0.5x1 05 cells/well in a flat-bottomed, 12well plate (Costar, Cambridge, MA) and grown overnight in the media appropriate for each cell type at 370C. DOTAP/DOPE liposome was resuspended and added to DNA at a 2:1 w/w liposome:DNA ratio for 30 minutes (Tang and Hughes, 1998). The complete media was replaced with serum-free media. The dose of pGL3 was 3 pg and of pGM was 4.15 ptg to achieve equal molar ratio. These doses were chosen for maximal effect in the same molar ratio (Figure 5-1). The cells were transfected for 4 hours, then the transfection media was replaced with complete media.

Primary neuronal cultures were prepared from the hippocampus of newborn Sprague-Dawley rats as described by Ajmani et al. (1999). Astroglial and microglial cells were cultured from the cerebral cortices of newborn rats as previously described (Ajmani et al., 1999). The 3.5x10 mm culture disk was used for cell cultures. DOGSDSO/DOPE liposome was complexed with DNA at an 8:1 w/w liposome:DNA ratio in serum-free media for 30 min at room temperature. The liposome/DNA complexes were added to the cultures and incubated for 2 h (neuronal cultures) or 7 h (microglia and astroglia cultures), and then transfection media was replaced with complete media (Ajmani et al., 1999).







64


6000 5000 4000

--W-pGM
3000 -pGL pGL3

2000 1000


0
0 1 2 3 4 5 6 7
Plasmid DNA (fig) Figure 5-1: Dose study of plasmid DNAs in CHO cells. The ratio of DNA and DOTAP/DOPE liposome is 1:2 w/w. Luciferase activity was presented in RLU/well.







65
Cells were harvested every 24 hours from Dayl to 10 posttransfection. Each well was aspirated and rinsed once with 500 ptl of PBS (phosphate buffer saline, pH 7.4), and then 250 pl or 500 [tl of Ix lysis buffer (0.1 M postassium phosphate buffer, pH 7.8, 1% Triton X-100, 1 mM DTT, 2 mM EDTA) were added to the mammalian (CHO and SKnSH) cells or neuronal cells (primary neuron, astroglia, microglia), respectively. The amount of lysis buffer added depended on the size of the well and the growth area. After shaking for 15 minutes, luciferase activity was quantified by using a Monolight 2010 luminometer (San Diego, CA). Only in dose-study and HDAC inhibitor experiments, luciferase activity was quantified by using a plate luminometer (Microtiter, Chantilly, VA). One hundred pl of luciferase assay buffer (30 mM Tricine, 3 mM ATP, 15 mM MgSO4, 10 mM DDT, pH 7.8) and 20 pl of cell lysate were added to a 100 pl injection of 1 mM D-luciferin, pH 6.3 (Molecular Probes, Eugene, OR). The light emission over a 10 seconds reaction period was integrated. Luciferase activity was expressed as relative light unit (RLU) per well assuming that the delivery vectors and expressed protein are nontoxic to cells.

In the cotransfection experiment, CHO and SKnSH cells were seeded at 0.5x1 0 cells/well in a 24-well plate (Costar, Cambridge, MA) 24 hours before transfection. The amount of pEPI-1 and pGL3 were 0:3, 1:2, 2:3 3:0 and 3:3 [tg:pg. The ratio of pDNA:DOTAP/DOPE liposome was 1:2 w/w. Cells were transfected with pDNA:1iposome complex as previously described. Forty-eight hours after transfection, luciferase activity was measure as described above. Pharmacokinetics






66
AUEC (area under expression curve) and MET (mean expression time), were obtained from a plot of RLU/well and time period. AUEC indicated a total amount of protein produced by DNA and MET indicated a length of expression time. These parameters can be calculated from trapezoidal rules as mention in Chapter 4 in Background section (Gibaldi and Perrier, 1982). Total AUEC, AUMC and MET can be calculated base on equation in Chapter 4 in Calculation section. Histone Deacetylase Inhibitor

CHO cells were seeded at lx105 cells/well in 24-well plate (Costar, Cambridge,

MA) 24 hours before transfection. Trichostatin A was dissolved in serum free ca-MEM in a dilution series of 0.1-1 ptM. This medium was then mixed with pDNA and DOTAP/DOPE liposome (pDNA:1iposome = 1:2 w/w) for 30 min before transfection as described above. To investigate the effect of the exposure time of the cell to HDAC inhibitor on expression, cells were incubated with TSA I or 2 hours before transfection. The exposure time of the cells with TSA did not effect transgene expression of pDNAs (data not shown). Forty-eight hours after transfection, luciferase activity was measured as described above.

Plasmid DNA Extraction

To investigate the appearance of plasmid DNAs in nucleus and cytoplasm as a period of time, cell fractionation was performed followed by polymerase chain reaction (PCR) for detection of pDNAs. CHO and SKnSH cells were seeded at 0.5x105 cells/well 24 hours before transfection. The same amount of DNA and liposome was used as described above. At Day 1, 2, 4 and 6 posttransfection, cells were fractionated. Cells were washed with ice-cold PBS and trypsinized using 0.2 ml of 0.25% trpysin, 1 mM






67
EDTA (Gibco BRL, Grand Island, NY) per well. After adding 500 pil of PBS, cell suspensions were placed in a Model SFR13K [tSpeedfuges refrigerated microcentrifuge (Savant, Farmingdale, NY) and centrifuged for 5 min at 1000 xg twice. Cell pellets were resuspended by briefly vortexing in 200 [il of homogenization buffer (10 mM HEPES, pH 7.9, 10 mM KC1, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 0.5 mM PMSF) and incubated on ice for 15 min. After this incubation, 12.5 [d of a 10% Nonident P-40 solution was added to each tube, and the samples were vortexed vigorously for 10 seconds. This was followed by another centrifugation for 30 seconds at 10,000 xg. The supernatant was collected as the cytoplasmic fraction. The crude nuclear pellet was washed 2 additional times by repeating the procedure outlined above. The nuclear pellet was then resuspended in 100 [l of homogenization buffer (Sperinde and Nugent, 1998). Contamination of lysosome, acid phosphatase, in nuclear fraction was measured by using EnzChecke acid phosphatase assay kit (Molecular Probe, Eugene, OR). Less than I% of lysosomal contamination was found in the nuclear fraction using the procedure above.

One volume lysis buffer containing 0.6% sodium dodecyl sulphate (SDS), pH 7.5, and 0.01 M EDTA was added to the nuclear or cytoplasmic fractions and incubated for 20 min at room temperature. After incubation, 5 M NaCl was added to make a final concentration of 1 M and the sample was mixed by slowly inverting the tube 10 times. The sample was stored at 4oC for at least eight hours, then centrifuged at 13,000 xg for 30 min at the same temperature to remove the major portion of SDS and protein (Hirt, 1967). For the whole cell, after being incubated with homogenization buffer for 15 min, lysis buffer was added to the cell suspension following by NaCl.






68
All protein and SDS fractions were extracted by 1 volume of phonol/chloroform/ isoamyl alcohol (25:24:1 v/v/v) (Ameresco, Solon, Ohio). After centrifuging at 14,000 rpm for 5 min at room temperature, supernatant was removed to a new tube. Plasmid DNA was precipitated by adding 2 volumes of ethanol and NaCl to make a final volume of 0.2 M. Glycogen (Fisher Biotech, Fair Lawn, NJ) was used as a carrier for DNA precipite at 50 ptg/ml of final product. The mixture was stored at -800C for 20 min and centrifuged at 13,000 xg for 30 min. Ethanol was discarded. Pellet was air-dried for 10 min, then resuspended in TE buffer, Tris-EDTA, pH 7.4. PCR-Based Assay of Extracted pDNA

To identify DNA in the nuclear and cytoplasmic fractions, we used PCR.

Luciferase oligonucleotide was used as a primer for the PCR reaction. The upper primer (pGL3, position 280) was 5' GGA ATT GCT AGC TAC TGT TGG TAA AGC CAC 3' and the lower primer (pGL3, position 1929) was 5' GGA AGA TCT AAA GCA ATA GCA TAC CAA AT 3' (Gibco BRL, Grand Island, NY). The reaction contained a 5 pl sample, 0.5 pl 0.1 nmol upper/lower primer, 0.5 pl 10 mM dNTP (Promega, Madison, MI), 1 p1 100 mM MgSO4, 5 pl 1Ox PCR buffer, 0.5 pl Vento DNA polymerase (New England Biolabs, Beverly, MA), and sterile water to make a final volume of 50 pl. The conditions were 940C, 2 min primary denature. The cycle began with 94 'C, 30s denature, 52 'C, 30s annealing, 72 'C, 1.30 min elongation for an appropriate number of cycles and final elongation at 72 'C for 5 min. This reaction generated 1.6 kb of luciferase sequence.






69
Subsequently, aliquots of the PCR products were electrophoresed in a 1% agarose gel (Fisher Sciencetific, Fair Lawn, NJ) mixed with ethidium bromide (0.2 ptg/ml) and quantified using Kodak ID Image Analysis Software (Kodak, New Haven, CT). Competition Binding Study

Previous results indicated the nuclear matrix could bind to any DNA sequence. However, they had a high binding affinity to matrix attachment regions (Tsutsui et al., 1993). In this experiment, we attempted to distinguish between the binding affinities of MAR-containing pDNA, pGM, and non-MAR-containing pDNA, pGL3.

CHO and SKnSH cells were seed at 1x105 cells/well. One ptg of DNA was mixed with 1, 2, or 3 pig of competitive non-coding bacterial DNA to create Treatments 1, 2, or 3, respectively. Plasmid DNA from E. coli bacteria strain DH5cX was utilized as noncoding bacteria DNA. Plasmid DNA was isolated using a Megawizard DNA purification kit (Promegao, Madison, WI). The concentration and purity of pDNA were determined spectrophotometrically. The average concentration of pDNAs was 2.0 mg/ml and the purity of pDNA, A260/A280 ratio, was 1.8. DOTAP/DOPE was added to those mixture at 2:1 w/w lipid:DNA ratio. Two days after transfection, cells were fractionated, DNA was extracted and PCR was performed as described above. The quantity of PCR products was maintained within the linear range (increasing the concentration of the template or the number of cycles proportionally increases the signal) (Donev, 2000). This was achieved by using the standard curve. The ratio between the amount of plasmid DNA in each fraction and plasmid DNA in a whole cell was calculated. All PCR experiments were carried out in triplicate. Similar results were obtained and summarized graphically.







70
Statistical Analysis

A student t-test with one-tail distribution and two-sample equal variance was used to compare pharmacokinetic parameters and amount of pDNAs in the competition study. An one-way ANOVA was used to detect differences in treatment for cotranfection studies and repeated-measures ANOVA was used to detect differences with respect to plasmid and time of the plasmid expression and HDAC inhibitor studies. Tukey's HSD was used for post-hoc analysis when differences were detected. The data were considered significantly different for p-value < 0.05.



Results


Figure 5-1 illustrated a dose-study of pDNAs in CHO cells. The ratio of pDNA:1iposome is 1:2 w/w. The optimum dose for pGL3 and pGM is 3 and 5 ptg, respectively. The decrease in expression at 6 ptg of pGM is likely due to toxicity from the liposome. The amount of 4.15 [tg pGM was used in all experiment to achieve the same molar ratio as pGL3. MARs-containingpDNA, pGM, enhanced the transgene expression in Chinese hamster ovary (CHO) cells (Figure 5-2A), but not in SKnSH cells (Figure 5-2B) or neuronal cells (Figure 5-3). From a plot of expression in RLU/well against time, pharmacokinetic AUEC and MET were obtained (Table 5-1).

In CHO cells, AUEC of pGM and pGL3 was 1259.19 11.65 x 105 and 488.05 204.79 x 105 (RLU.day)/well while MET of pGM and pGL3 was 3.95 0.07 and

3.38 0.05 days, respectively. AUEC and MET of pGM were significantly (p < 0.05) higher than pGL3, indicated pGM has a higher level of transgene expression and higher average time of expression than pGL3. In SKnSH cells, pGM (AUEC = 67.35 28.74 x







71
A.



35 *
30
o 25

20 pGL3
15 1--pG M
10



1 2 3 4 5
days post transfection

B.


16
14
m 12 X 10
8 pGL3
-U-pGM

6 TT
27
0
1 2 3 4 5
days post transfection

Figure 5-2: Transgene expression of pGL3 and pGM in (A) Chinese hamster cells (CHO) and (B) Human neuroblastoma cells (SKnSH) in relative light unit (RLU) per well. represents significant difference of expression between plasmid DNAs at respective time with p-value < 0.05. Number of replicate, n = 4. The experiment was performed three times.







72




A.



4 pGL3
3
-.- DGM
2
1 4

0
2 4 6 8 10
days post transfection

B.



8
*pGL3
6
-- DGM

2 0
1 2 3 4 5 6 8 10
days post transfection

Figure 5-3: Transgene expression of pGL3 and pGM in (A) Hippocampal primary neuron, (B) Astroglia, and (C) Microglia in relative light unit (RLU) per well. represents significant difference of expression between plasmid DNAs at respective time with p-value < 0.05. Number of replicate, n = 3. The experiment was performed three times.







73




C.



10
8 pGL3
6
-- --- GM
4 2 0
1 2 3 4 5 6 8 10
days post transfection

Figure 5-3--Continued











Table 5-1: Pharmacokinetic parameters of pGM and pGL3 in CHO, SKnSH, and neuronal cells. Cell lines Plasmid Area Under Expression Mean Expression Time Cmax*10' tmax
Curve (AUEC) *105 SEM (MET) SEM
RLU.day/well day RLU/well day
CHO (Chinese pGL3 488.05 204.79 3.38 0.05 158.4 2
HamsterOvary) pG/M 1259.19 11.65 3.95 0.07 291.4 2


p-value 0.0321 0.0104

SKnSH (Human pGL3 48.09 11.60 4.10 0.66 7.09 2
neuroblastoma)
pG/M 67.35 28.74 5.13 0.68 9.67 2

p-value 0.2840 0.3026

Cmax represents maximum concentration or expression; tmax represents time of maximum expression.








Table 5-1--Continued
Cell lines Plasmid Area Under Expression Mean Expression Time Cmax* 105 tmax
Curve (AUEC) *105 SEM (MET) SEM (RLU.day)/well day RLU/well day
Primary Neuron pGL3 19.62 4.50 5.04 0.79 2.98 4
(0-10 day)
pG/M 21.36 3.69 5.40 0.96 3.25 4

p-value 0.3905 0.3910

Astroglia(O-Oday) pGL3 14.88 2.05 5.20 1.65 4.95 5

pG/M 24.80 15.14 5.53 1.27 5.88 3

p-value 0.2912 0.4448

Microglia(0-10 pGL3 1.77 0.43 7.08 0.66 3.92 10
day)
pG/M 2.80 0.30 6.94 0.71 6.82 8

p-value 0.0583 0.4476

Cmax represents maximum concentration or expression; tmax represents time of maximum expression.






76
105 (RLU.day)/well) expressed slightly more protein than pGL3 (AUEC = 48.09 11.60 x 105 (RLU.day)/well), but this was not significantly different. Mean expression times of pGM (5.13 0.68 days) and pGL3 (4.10 0.66 days) were similar. In primary neurons, AUEC of pGM and pGL3 was 21.36 3.69 x 105 and 19.62 4.50 x 105 (RLU.day)/well, respectively, while MET of pGM and pGL3 was 5.40 0.96 and 5.04 0.79 days, respectively. Plasmid pGM and pGL3 demonstrated similar protein expression and time of expression (Table 5-1).

In astroglia, AUEC of pGM and pGL3 was 24.80 15.14 x 105 and 14.88 2.05 x 105 (RLU.day)/well, respectively and MET of pGM and pGL3 was 5.53 1.27 and 5.20 1.65 days, respectively. In microglia, AUEC of pGM and pGL3 was 2.80 0.30 x 105 and 1.77 0.43 x 105 (RLU.day)/well, respectively and MET of pGM and pGL3 was 6.94

0.71 and 7.08 0.66 days, respectively. In astroglia and microglia, the average time of expression was identical for pGM or pGL3. According to MET, pGM did not prolong expression in the brain-derivatives cells. Time of the maximum expression (tmax) in fastdividing cells (CHO, SKnSH) was earlier than in non-dividing (primary neuron) or slowdividing cells (astroglia and microglia).

Figure 5-4 shows transgene expression of non-MAR-containing pDNA, pGL3, cotransfected with MARs-containing pDNA, pEPI-1. MARs has trans effect on other pDNAs. We found MARs-containing pDNA increase transgene expression of nonMAR-containing pDNA in CHO cells, but not in SKnSH cells. In SKnSH cells, luciferase activity depended on the amount of pGL3. Three jag of pGL3 showed higher level (p < 0.05) of luciferase activity than 2 jag or 1 [tg of pGL3. The expression was independent of the amount of pEPI-1. In CHO cells, transgene expression increased (p <






77
0.05) when MARs-containing pDNA, pEPI-1, was added from the ratio of 0:3 pig pEPI-1: pGL3 to 1:2 ptg pEPI-1:pGL3. In the last treatment that contained 3 [tg of each pGL3 and pEPI- 1, the expression was significantly higher than baseline (3 [tg of pGL3 and 0 tg of pEPI-1) in CHO cells. The expression was significantly decreased, however, in the last experiment, in SKnSH cells.

Figure 5-5 shows the effect of trichostatin A (TSA) on transgene expression of MAR and non-MAR-containing pDNA in CHO cells. At low concentration of trichostatin A (0.005-0.045 [tM), expression of pGM, but not pGL3, was significantly increased (Fig. 6) compared to baseline (no TSA added). Increasing the concentration of TSA resulted in a decrease of transgene expression, not different from no TSA added (data not shown).

Plasmid DNA in CHO cells was located in the cytoplasm at Day 1

posttransfection (Figure 5-6). Plasmid DNA was observed in the nucleus between Day 2 to 4. Plasmid DNA began to degrade at Day 6. In the competition study, Table 5-2 shows the ratio of pDNAs in the nucleus or cytoplasm of whole cells. The amounts of non-coding pDNA were from 1, 2, and 3 [.g in Treatments 1, 2, and 3, respectively. Ratio of pDNA:1iposome is 1:2 w/w. The more non-coding DNA added, the less DNA was found in both nucleus and cytoplasm.


Discussion and Conclusions


In our study, MARs increased and prolonged transgene expression in CHO, but not SKnSH or neuronal cells. It was apparently that cell lines are plays an important role in the extent of MARs function.







78 A.



50- *
45
40 *
35
0
-30
25 20 15 10
5
0-
0:3 1:2 2:1 3:0 3:3

pEPI-1:pGL3 (hg) B.



50
45
40 3530


S20 15
*
10
*
5
0

0:3 1:2 2:1 3:0 3:3

pEPI-1:pGL3 (hg) Figure 5-4: Luciferase activity of pGL3 when contransfected with MAR-assocaited plasmid DNA, pEPI-1, in (A) Chinese hamster ovary (CHO) cells and (B) human nueroblastoma (SKnSH) cells. represents significant difference of expression compared to 0:3 gg of pEPI-1:pGL3 with p-value < 0.05.






79

This may be due to the binding of MARs to the nuclear matrix because it is a prerequisite for MARs function (Luderus et al., 1992; Luderus et al., 1994; Stief et al., 1989; Tang and Hughes, 1998) and it is a cell type specific (Wilson et al., 1999). Baiker and coworkers (2000) found that hIFN-P MARs associated with host chromosome in CHO cells in non-covalent fashion.

This association did not change significantly between 20 and 50 generations after transfection in CHO cells (Baiker et al., 2000). The hIFN-3 MAR may have a strong interaction to a nuclear matrix of CHO cells, but has a moderate and weak interaction with the nuclear matrix of SKnSH cells and neuronal cells, respectively. The different cell lines may be linked to the differences in enzyme machinery, MARs-binding protein or MARs-binding transcription factors. The studies found that some classical matrix proteins are cell-type specific and bind strongly and selectively to MARs from different species (Boulikas, 1993; Boulikas, 1995).

The dynamic of cells may be another explanation. Non-dividing cells, such as

primary neuron or slow-dividing cells, such as astroglia and microglia, may lack of some components that are necessary for MARs-nuclear matrix binding, whereas there are plenty of these components in CHO cells. Moreover, it is more difficult for pDNA to enter non-dividing cells than dividing cells. Thus, less expression was observed. We hypothesized that pDNAs in non-dividing or slow-dividing cells were not diluted out as early as the fast-dividing cells, thus resulting in longer expression time (MET). However, the expression was poor and the peak of expression shifted from Day 2 in dividing cells to Day 4, or 10 in non-dividing or slow-dividing cells.







80





30000 25000

20000

-U-pGM 15000 10000 5000

0
0 0.01 0.02 0.03 0.04 0.05

Trichostatin A (pM)


Figure 5-5: Transgene expression of pGL3 and pGM in CHO cells when incubated with histone deacetylase inhibitor, trichostatin A (TSA), in low doses (TSA 0-0.045 pM). represents significant difference of expression between plasmid DNAs at respective concentration. # represents significant difference of expression compared to baseline (0 pM TSA added) of the repective pDNA.









Dayl Day2 Day4 Day6
Std pGL3 pGM pGL3 pGM pGL3 pGM pGL3 pGM



Nucleus






Cytoplasm




Figure 5-6: Plasmid DNA, pGM, and pGL3 extracted from nucleus and cytoplasm of Chinese hamster ovary (CHO) cells after 1, 2, 4, and 6 days of transfection. Polymerase chain reaction (PCR) was used to amplify luciferase sequences, which are the backbone of pGM and pGL3. Std is lambda DNA/Hind III marker.












Table 5-2: Plasmid DNA extracted from nucleus and cytoplasm of Chinese hamster ovary (CHO) cells % DNA in whole cell
Treatment DNA Non-coding Liposome
(pg) DNA (pg)
(g) Nucleus Cytoplasm

pGL3 pGM pGL3 pGM

1 1 1 4 13.54 5.75 7.33 0.21 59.04 44.65 71.44 18.24

2 1 2 6 6.54 0.71 12.83 4.61 80.65 11.87 62.21 6.29
00
3 1 3 8 4.10 1.77 5.93 5.91 50.08 7.17 55.69 7.64

Plasmid DNA was extracted after 2 days of trasnfection with one [tg of pDNA and 1, 2, and 3 ptg of non-coding pDNA from bacteria. Ratio between pDNA:1iposome is 1:2 w/w. Polymerase chain reaction (PCR) was used to amplify luciferase sequences, which are the backbone of pGM and pGL3. The amount of pDNA was normalized by the amount of pDNA in whole cells.






83
This might be due to slow DNA trafficking in non-dividing or slow-dividing nuclei. We observed multiple peak of expression in astroglia. This may be due to the heterogeneous cell population.

Another possible reason for differences might be the size of pDNAs. Large

pDNA might interfere with the binding of MAR sequences and nuclear matrix, chromatin remodeling, and domain forming. These procedures are reversible and important to MARs functions and might effect expression of MARs-containingpDNA in cells. Promoter effect is another explanation. SV 40 promoter might not be strong enough to initiate the expression in neuronal cells. It not may either bind tightly with RNA polymerase or not work with MARs to enhance expression. Developing plasmid DNA, which has a wide variety of cell types, is a future prospect. The smaller-sized plasmid DNA with stronger promoter might be an alternative. This could be done by cutting pDNA at bacterial part, but not at MARs part because decreasing in MARs size from the original results in decreasing in transgene expression (Bode et al., 1992).

We also observed trans-activity of MARs-containingpDNAs with another pDNA in CHO cells (Fig. 5A), but not in SKnSH cells (Fig. 5B). The data confirm a cell type specificity of the MARs vector. In SKnSH cells, we found that luciferase activity was dependent on the amount of pGL3, but not the amount of pEPI-1. However, this effect was not observed in CHO cells; that is transgene expression of pEPI-1:pGL3 ratio of 0:3 ptg and 2:1 ptg were similar. Moreover, when pEPI-1 was added to the pEPI-1:pGL3 ratio of 1:2 ptg, the transgene expression was significantly higher than pEPI-1:pGL3 ratio of 0:3 [tg. These data disagree with previous studies that indicated effects of MARs only when they located within reporter pDNA (Bode et al., 1992). In SKnSH cells, pEPI-1






84
seemed to decrease expression of pGL3 in the last treatment (3 Pg of pEPI-1 and pGL3) compared to baseline (0 pg of pEPI-1 and 3 pg pGL3). Because the amount of pDNAs in the last treatment is doubled that in other treatments (6 jig instead of 3 [tg), more liposome was added to keep the pDNA:liposome ratio (1:2 w/w) constant. Decreasing in expression might be due to the toxicity of the liposome to the SKnSH cells.

The effect of HDAC inhibitor, trichostatin A (TSA), on expression of MARscontaining pDNA was demonstrated. Since MARs enhance expression in CHO cells, we chose this cell type to study the effect of TSA, on transgene expression. MARscontaining pDNA acts synergistically with low concentration TSA (0.005-0.045 1IM). In high concentration of TSA (0.1-1.2 jiM) the expression of MARs and non-MARscontaining pDNA decreases to the baseline level (no TSA added). Trichostatin A is fungistatic antibiotic and a potent histone deacetylase (HDAC) inhibitor that leads to the reversible hyperacetylation of histone at nanomolar concentrations (Yoshida et al., 1990). This hyperacetylation results in an active gene in vitro and in vivo (Chen et al., 1997; Dion et al., 1997; Van Lint et al., 1996). Therefore, increasing transgene expression of MARs-containing pDNA may be due to hyperacetylation of gene by TSA together with chromatin alteration by MAR. The data showed a dose-dependent effect of TSA on transgene expression. This data agrees with a previous study by Yoshida (1990) that indicated the effect of TSA on histone deacetylase inhibition at nanomolar concentrations.

Plasmid DNAs were extracted at various times after transfection and PCR was

performed. Plasmid DNAs were detected mostly in cytoplasm at Day 1 posttransfection. Plasmid DNAs were detected mostly in nucleus at Day 2 to 4 and began to degrade at







85
Day 6. These data agree with transgene expression in CHO cells. Peak of expression is at Day 2 and mean expression time is around Day 4. DNA was then diluted out and degraded, leading to the decline in expression. MARs-containing pDNA (pGM) and nonMAR-containing pDNA (pGL3) have similar transfection efficiency, which is the ability to enter and degrade in cells. However, in the expression of pGM was higher than pGL3 in CHO cells. These data agree with previous studies showed that MAR- containing pDNA enhanced expression by increasing transcription level, which means the amount of mRNA, but did not depend on amount of DNA (Forrester et al., 1994). This enhancement might be due to their sequences. The competition study in Table 2 shows no difference between the ratio of pDNAs in nucleus and cytoplasm in whole cell in any treatment. However, non-coding pDNA effected the ability of pDNA to enter the nucleus and affected the complexing of DNA with liposome. This data confirmed that pGM and pGL3 has the similar transfection efficiency. Adding sequences, such as nuclear localization sequences (NLS) which facilitate the entrance of pDNA into the nucleus in MARs-containing pDNA, is suggested to improve the expression.

MARs containing pDNA can enhance and prolonge the expression in Chinese hamster ovary cells but not in human neuroblastoma cells or neuronal cells. This MAR vector showed trans-activity on another reporter pDNAs specifically in CHO cells. It has a synergistic effect with histone deacetylase inhibitor, trichostatin A, also in CHO cells. This evidence indicates the cell type specific character of this MARs-containing pDNA. This may be due to the high affinity of MARs for the nuclear matrix in the certain cell types. Despite the differences among cell lines, MARs-containing pDNA showed some advantages over non-MAR-containing pDNA because this vector combines the






86
advantages of non-viral vector with the retention activity of MARs. Although MAR vector has cells type limitations, this vector shows a potential for improvement and application to nonviral gene transfer.













CHAPTER 6
CONCLUSION AND FUTURE PROSPECTS



Conclusion


The goal of this research was to develop plasmid DNA for nonviral gene transfer. Conventional plasmid DNA suffers from various limitations, such as low and transient expression. This can lead to retreatment that is not desirable when used clinically. Plasmid DNA that remains in cells would result in enhanced and prolonged expression. MARs have been studied for application in nonviral gene transfer for the last decade. Previous studies indicated MARs replicated extrachromosomally and well maintained for more than 100 generation in cells. Therefore, our hypothesis was that MARs-containing pDNA has higher protein expression compared to non-MARs-containing plasmid DNA.

To test this hypothesis, we developed a data analysis method using multiple-point measurements of pharmacokinetic parameters, area under the curve (AUC) and mean residence time (MRT). Multiple-point measurements in gene expression have an advantage over single point measurements. Gene expression is a dynamic process that is takes time from DNA transcription to translation. For example, time for DNA to be initialized by cells and to move through the cytoplasm to the nucleus, stability of DNA and mRNA, phase of cell cycle, and cell turnover rate. Moreover, in heterogeneous populations, the mixing of healthy and unhealthy cells at certain times result in various expressions. Thus, multiple-point measurements provide insight into the whole picture, 87






88
not just one point of expression. Utilization of pharmacokinetic parameters allows the calculation of total protein production from a plot of transgene expression and time. Under the assumption of linear pharmacokinetics with first order elimination, AUEC and MET were obtained. AUEC indicates the total amount of protein produced and MET indicates the average time of the expression.

By using a combination of multiple-point measurements and pharmacokinetic parameters, we were able to compare transgene expression of MARs and non-MARs containing pDNA in various cell types. MARs-containing pDNA significantly enhanced and prolonged transgene expression in Chinese hamster cells (CHO), but not in human neuroblastoma cells (SKnSH). In cell types where transfection is difficult, such as primary neuron, astroglia, and microglia, MARs-containing pDNA showed similar transgene expression and time of expression compared to non MARs-containing plasmid DNA. MARs-containing plasmid DNA, like other pDNA, functions well in rapidly dividing cells. Once in the nucleus, MARs sequences trigger biological change leading to supenor gene expression. The variation of extent of expression may be due to the different binding affinities of MARs and nuclear matrix in different cell types. Physical properties of pDNA, such as size, promoter, reporter, or position of the inserted MARs fragment in pDNA may be another explaination for the difference extent of transgene expression in different cell lines.

We found that MARs-containing pDNA has the cell-type specific properties.

MARs-containing pDNA increased transgene expression of another reporter pDNA only in CHO cells. This pDNA acts synergistically with histone deacetylase inhibitor, trichostatin A (TSA). This effect was observed in low doses TSA and in CHO cells. As






89
previously discussed, different binding affinities of MARs and nuclear matrix in different cell types may be a main reason. However, MARs-containing pDNA exhibited a similar pattern of appearance in the nucleus as non-MARs-containing pDNA. Moreover, noncoding bacterial pDNA also effected the appearance of both MARs and non-MARs pDNA.

We also investigated the link between pDNA topoisoform (supercoiled, open

circular, and linear) and transfection efficiency. Plasmid DNA is a basic tool for nonviral gene transfer and supercoiled is the major form of pDNA when it is isolated from bacteria. When pDNA undergoes environmental change, such as physical or chemical changes, it results in degradation of pDNA, which can be open circular, linear, or fragmented. Our hypothesis was that different pDNA topoisoforms affect transfection efficiency differently.

To test this hypothesis, supercoiled, open circular, and linear forms were

transfected into cells by using either liposomes or electroporation. We observed that even though the open circular form had larger higher half-life in cytoplasm and a higher amount present in cell, its transfection efficiency was not significantly different from the supercoiled or linear forms. We concluded that different pDNA topoisomers do not affect transfection efficiency.



Future Aims


MARs-containing pDNA characteristics describes in this dissertation suggest that these nonviral vectors have potential for improvement. However, MARs biological functions are still unclear. Constructing MARs-containing pDNA that works in a wide







90
variety of cell types is required. This could be done by replacing the SV40 promoter with one that is more resilent and not shut down by cellular mechanism; reducing the size of pDNA but not the MARs sequences, since that would reduce their function; and adding sequences that facilitate the entrance of pDNA into nucleus, such as nuclear localization signal (NLS). This combination of MARs sequences and NLS might give superior results.















REFERENCES


Adami, R. C., Collard, W. T., Gupta, S. A., Kwok, K. Y., Bonadio, J. and Rice, K. G.
Stability of peptide-condensed plasmid DNA formulations. JPharm Sci 87, 678-83. (1998).

Agarwal, M., Austin, T. W., Morel, F., Chen, J., Bohnlein, E. and Plavec, I. Scaffold attachment region-mediated enhancement of retroviral vector expression in primary T cells. J Virol 72, 3720-8. (1998).

Ajmani, P. S., Tang, F., Krishnaswami, S., Meyer, E. M., Sumners, C. and Hughes, J. A. Enhanced transgene expression in rat brain cell cultures with a disulfide-containing cationic lipid. Neurosci Lett 277, 141-4. (1999).

Alvarez, J. D., Yasui, D. H., Niida, H., Joh, T., Loh, D. Y. and Kohwi-Shigematsu, T. The MAR-binding protein SATB 1 orchestrates temporal and spatial expression of multiple genes during T-cell development. Genes Dev 14, 521-35. (2000).

Auten, J., Agarwal, M., Chen, J., Sutton, R. and Plavec, I. Effect of scaffold attachment region on transgene expression in retrovirus vector-transduced primary T cells and macrophages. Hum Gene Ther 10, 1389-99. (1999).

Baiker, A., Maercker, C., Piechaczek, C., Schmidt, S. B., Bode, J., Benham, C. and Lipps, H. J. Mitotic stability of an episomal vector containing a human scaffold/matrixattached region is provided by association with nuclear matrix. Nat Cell Biol 2, 182-4. (2000).

Bates, A. and Maxwell, T. in DNA Topology (eds. Bates, A. & Maxwell, T.) 31 (Oxford University Press, Oxford, 1993).

Benham, C., Kohwi-Shigematsu, T. and Bode, J. Stress-induced duplex DNA destabilization in scaffold/matrix attachment regions. JMol Biol 274, 181-96. (1997).

Berezney, R. and Coffey, D. S. Nuclear protein matrix: association with newly synthesized DNA. Science 189, 291-3. (1975).

Blackwood, E. M. and Kadonaga, J. T. Going the distance: a current view of enhancer action. Science 281, 61-3. (1998).



91




Full Text

PAGE 1

MA TRIX ATTACHMENT REGIONS-CONTAINING PLASMID DNA AS FACILITATORS IN PLASMID TRANSFER By PATTRAVADEECHANCHAM A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2001

PAGE 2

This work is dedicated to my parents Yuwadee and Charun Chancham ; and my grandparents, Kosum and Pramote Launpreeda.

PAGE 3

ACKNOWLEDGMENTS I would like to express my sincere appreciation to my mentor Dr. Jeffrey Hughes for his advice, patience, and understanding but most of all his brotherly generosity and fatherly protection. I would like to thank my other committee members Dr. Gayle Brazeau, Dr. Guenther Hochhaus, Dr. Sean Sullivan, Dr Edwin Meyer and Dr. William Farmerie for their expert advice and kind support I also wish to acknowledge all the personnel in the Department of Pharmaceutics including, secretaries, graduate students, and particularly the Hughes group I especially thank Adam Persky who helped me throughout my graduate life. I also would like to thank Wu Xiao and Yi Wen for advice on my experiments. My friends Oravaree, Nopadon Jintana, Intira, and Ariya were around to continuously support me during good and bad times. Lastly, I would like to express thanks to my parents and grandparents for their unselfish love ; to my sisters for their encouragement; and to Jim Buranatrakul for being my inspiration lll

PAGE 4

ACKNOWLEDGMENTS LIST OF TABLES LIST OF FIGURES CHAPTERS 1 INTRODUCTION TABLE OF CONTENTS 2 BACKGROUND AND SIGNIFICANCE lll VI Vil 1 5 Matrix Attchment Regions (MARs) Overview ... .... ............. ...... ... ................ ......... .... 5 Characteristics ..................................................................................... ..... .... ................. 7 Functions ............... ......... ....... ....... ........ ................. ............. ................. .......... ..... 14 3 RELATIONSHIP BETWEEN PLASMID DNA TOPOLOGICAL FORMS AND IN VITRO TRANSFECTION 23 Introduction ...................................................................................... ...... ................ ...... 23 Materials and Methods ................................ ................... ..................... ............. ............ 24 Results ........................................................................................................................... 31 Discussion and Conclusion ........................................................................................... 35 4 USE OF PHARMACOKINETIC PARAMETERS TO INTERPRET GENE EXPRESSION 43 Introduction ............ .... ........... ... ..... ... ... ................................. ............................. ....... 43 Materials and Methods ........ ................................................................. ........ ........ .... .... 45 Results .... ....... ............................. ........................ ............... ....... .... ............. ........... 51 Discussion and Conclusion .............................................................. ................ . ........ 57 5 MATRIX ATTACHMENT REGIONS-CONTAINING PLASMID DNA INCREASES GENE EXPRESSION IN VITRO 60 Introduction ................................................................................................................... 60 Materials and Methods .................................................................................................. 61 Results ..................................... ....... ... ......... ............ ....... ...... ............... .... ............... ... 70 Discussion and Conclusion ........................................................................................... 77 lV

PAGE 5

6 CONCLUSION AND FUTURE PROSPECTS 87 Conclusion ... ............ ............ ..... ......... ..... . .... ................ .............. ........ .... ......... 87 Future Aims .......... ....... .... ....... . ............ . ...... ........ .......... ...... ................. .... .... 89 REFERENCES BIOGRAPHICAL SKETCH V 91 102

PAGE 6

LIST OF TABLES 3-1 Pharmacokinetics of plasmid DNAs in cytoplasmic solution ...... .. ............... .. ..... ..... ..... 37 4-1 Pharmacokinetic parameters of simulated plasmid DNAs .. ........... ........... ....... .......... .. 53 4-2 Transgene expression of plasmid pGL3 and pGM in CHO cells ....... .............. ... .. ......... 54 4-3 Pharmacokinetic parameters of pGL3 and pGM .. .... .... .. .. .. ....... ...... .. ........ .. .. ......... ....... 56 5-1 Pharmacokinetic parameters ofpGM and pGL3 in CHO, SKnSH, and neuronal cells .. 74 5-2 Plasmid DNA extracted from nucleus and cytoplasm of CHO) cells ...... ............... ..... ... 82 VI

PAGE 7

LIST OF FIGURES Figure 2-1 Nuclear matrix isolated by amine modification .......... ......... .. ....... .. ........ .... .. .......... ..... 21 3-1 Diagram of first-order kinetics .. ........ .... .............. ......... .. .... ......... ..... .. .. .. ........ ...... ........ 29 3-2 Agarose gel electrophoresis of plasmid DNA topoisoforms ............................... ..... ....... 32 3-3 Transfection effic i ency of pDNA-liposome .... ...... ............. ..... ................... .. ........... ...... 33 3-4 Transfection efficiency of pDNAs by using electroporation .... .................. .......... .... ...... 34 3-5 Agarose gel electrophoresis of plasmid DNAs in cytoplasm .... ........................ ...... ....... 3 6 3-6 Mean fluorescence of the labeled plasmid DNAs .... ............. ............ .............................. 38 4-1 Schematic representation of the plasmid DNAs .............................. ...................... ...... .. .4 7 4-2 Simulation oftransgene e x pression of plasmid DNAs 1 2 3, and 4 ...... ............. .......... 52 4-3 Trans gene expression of plasmid DNAs pGL3 and pGM in CHO cells ....................... 55 5-1 Dose study of plasmid DNAs in CHO cells ........... .................................... ..................... 64 5-2 Transgene expression ofpGL3 and pGM in CHO and SKnSH ......... .. ...... ........ ............ 71 5-3 Transgene expression of pGL3 and pGM in hippocampal primary neuron astroglia and microglia ............ ...... .. .... .... ................. ............... .................. ........................ .... ....... 7 2 5-4 Luciferase activity ofpGL3 when contransfected with pEPI-1 in CHO and SKnSH cells .... .. .......... .......... ............. .. .. .. .. .. .. .... ..... .. ...... .. ... .. ... .. ............... .... ..... ..... ...... .... .. 78 5-5 Transgene expression of pGL3 and pGM in CHO cells when incubated with histone deacetylase inhibitor, trichostatin A (TSA) ............. ..................................... ................... 80 5-6 Plasmid DNA pGM and pGL3 extracted from nucleus and cytoplasm of CHO cells .81 Vll

PAGE 8

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MA TRIX ATTACHMENT REGIONS-CONTAINING PLASMID DNA AS FACILITATORS IN PLASMID TRANSFER By Pattravadee Chancham December 2001 Chairman: Dr. Jeffrey A. Hughes Major Department: Pharmaceutics Nonviral gene transfer is an alternative to viral vectors for gene transfer. However, nonviral transgene expression remains undesirably low and transient. Matrix attachment regions (MARs) are DNA elements that are defined by their high affinity for the nuclear matrix. MARs may also be related to long-term transgene expression in vitro. The purpose of this research is to evaluate human interferon-~ MARs element in various cell types. This was done by constructing MARs-containing pDNA and comparing their transgene expression with non-MARs-containing pDNA. We found that MARs containing pDNA increased and prolonged the expression in Chinese hamster ovary (CHO), but not in human neuroblastoma cells (SKnSH) and neuronal cells (primary neuron, astroglia, and microglia). From cotransfection experiment, MARs-containing pDNA had trans effect on another pDNA. This vector also acted synergistically with histone deacetylase inhibitor, trichostatin A. Polymerase chain reaction was used to Vlll

PAGE 9

monitor intracellular distribution of pDNA. We found that MARs-containing pDNA has similar intracellular distribution as non-MARs-containing pDNA. Because pDNA was a basic tool in nonviral gene transfer, the relationship between DNA topoisoforms and in vitro transfection efficiency is discussed The DNA topoisoforms did not affect the level of trans gene expression even though the extent depends on cell types and pDNA promoter IX

PAGE 10

CHAPTER 1 INTRODUCTION Gene therapy is the ultimate method for delivering proteins into the body. The technique of introducing the genetic material into the target cells of a patient is a key component of every gene therapy protocol. A variety of gene delivery systems are currently used to insert therapeutic genes into somatic cells They are divided into viral and nonviral gene transfer methods. Conventional vectors currently used for gene delivery have a number of limitations. Viral vectors may randomly integrate into the host genome. Integration of a therapeutic episome into the host chromatin has the distinct possibility of activating or inactivating important genetic loci, with the potential for deleterious consequences (Wendelburg and Vos, 1998). For example, insertional mutagenesis and transgene silencing result in such pathological changes in tissues as tumorigenesis, growth inhibition, or cell death. This problem of safety may be difficult to overcome. A nonviral or plasmid-based vector is an alternative. For instance plasmid DNA (pDNA) vectors do not have a size constraint imposed by viral packaging. Moreover, it has lower toxicity and immunogenicity compared to a viral vector (Wendelburg and Vos, 1998). Even though naked DNA has used to successfully increase expression in many cell types, stability of unprotected DNA is a major concern. Plasmid DNA needs an efficient nonviral delivery method, such as receptor-mediated ligand targeting system (Wagner et al., 1992; Wilson et al., 1992), liposomes (Tang and Hughes, 1998), 1

PAGE 11

2 hemagglutinating virus ofJapan (HVJ)-liposome (Tsukamoto et al., 1999 ; Yanagihara et al., 1996) or physical methods, such as injection (Hengge et al., 1995) to deliver it to the target cell. Another important issue is the transient expression Plasmid DNA appears to have a short lifetime in most tissues. One possible reason is the inherent susceptibility to nucleases of DNA. Vector DNA that is unable to replicate is rapidly diluted out in mitotic cells (Wells et al., 1998 ; Wells et al., 1997). In post-mitotic cells it also appears that vector DNA is rapidly lost perhaps by passage through nuclear pores and subsequent degradation (Wohlgemuth et al., 1996) Furthermore most pDNAs do not integrate into th e host genome thus they tend to be targets for enzyme degradation Plasmid DNA that could be replicated or could be retained in cells without integration into the host genome is required to increase the persistence of DNA in cells (Piechaczek et al., 1999). Increasing the persistence of pDNAs would lead to prolonged and enhanced transgene expression. Matrix attachment regions (MARs) are DNA sequences that are identified through their high affinity to bind to the nuclear matrix or scaffolds in v itro (Gasser and Laemmli 1986 ; Mirkovitch et al., 1984). The following evidence indicates that MARs-associtaed plasmid DNA might be a prime candidate for nonviral gene delivery : MARs replicated and was retained in cells without integrating into host chromosome (Piechaczek et al., 1999). MARs inhibited methylation that would repress gene transcription (Dang et al., 2000; Forrester et al., 1999). MARs e x tended the histone acetylation domain and conferred chromatin accessibility (Fernandez et al., 2001; Forrester et al., 1994; Jenuwein et al., 1997)

PAGE 12

3 MARs protected DNA from neighboring chromatin effect (Kalos and Fournier, 1995; Klehr et al., 1992; Phi-Van et al., 1990; Poljak et al., 1994) Proteins necessary for transcription were found in MARs (Boulikas, 1995). Moreover, MARs do not express protein, thus oncogenicity, immunogenicity, and toxicity may be not the major concerns. This evidence may relate to increasing and maintaining expression. Objective I The main objective of the project was to develop pDNA vector for nonviral gene transfer by using DNA sequences called matric attachment regions. Hypothesis We hypothesized that MARs-containing pDNA prolongs or enhances transgene expression compared to non-MARs--containing pDNA. To test this hypothesis, MARs containing pDNA are constructed and its transgene expression is compared with nonMARs associtaed pDNA. The role of MARs-containing pDNA in various cell types is also investigated. Pharmacokinetic concepts can be applied to gene expression of pDNAs because transgene expression can be follwed over time by treating pDNA as a prodrug and we can then measure pDNA active metabolites. A combination of multiple point measurements and pharmacokinetic parameters were used in data analysis Plasmid DNA is the basic tool for nonviral gene delivery. Different conformational state of pDNA can assist in developing approaches to increase its transfection efficiency. The initial form of pDNA isolated from bacteria is supercoiled. However, exposure of pDNA to physical and chemical environments can lead to the their

PAGE 13

4 degradation which means that the conformation can change to open circular, linear or even fragmented DNA. Objective II Another objective in this study is to investigate the relationship between supercoiled, open circular and linear pDNA topoisoforms and in vitro transfection efficiency Hypothesis Plasmid DNA topoisoforms affect transfection efficiency To test this hypothesis open circular and linear pDNA are produced These forms (including supercoiled) are transfected into cells by using either electroporation or liposomes The half-life of pDNA topoisoform in cytoplasm is calculated

PAGE 14

CHAPTER2 BACKGROUND AND SIGNIFICANCE Matrix Attchment Regions (MARs) Overview The architecture of the nuclear interior is composed of two nucleic acid containing structures: a DNA-containing structure called the chromatin and an RNA containing structure (He et al., 1990; Nickerson et al., 1998) Eukaryotic chromatin is organized into domains that may affect differential gene expression (Bode et al., 1995; Boulikas, 1995). This organization is brought about by the anchoring of specific DNA sequence landmarks to a network of protein crossties termed the nuclear matrix, at an interphase or chromosomal scaffold during mitosis. After a combination of nuclease digestion and extraction, the DNA sequences that tightly associate with the nuclear matrix or scaffold in vitro have been called matrix or scaffold attachment regions (MARs or SARs) (Bode et al., 1992). These MARs or SARs have an average size of 500 base pairs (bp ), are spaced approximately every 30 kilobase pair (kbp) (Boulikas, 1995), and are control elements maintaining independent realms of gene activity The DNA replication, transcription, repair, splicing, and recombination appear to take place on the nuclear matrix The MARs have been experimentally defined for several gene loci including the chicken lysozyme gene, human interferon-~ gene, human~-globin gene, chicken a-globin gene, p53, human protamine gene cluster (Singh et al., 1997) and human serpin gene cluster (Rollini et al., 1999). 5

PAGE 15

6 The sequences of MARs do not have clear-cut consensus sequences but share conformation characteristics (Benham et al., 1997; Yamamura and Nomura 2001) MARs are typically 70% AT-rich sequences (Boulikas, 1993). These sequences are responsible for bending (Yamamura and Nomura 2001), unwinding of DNA (Bode et al., 1992) and binding DNA with the cellular nuclear matrix which is the prerequisite for MARs functions (Bode et al., 1992; Boulikas 1993; Boulikas 1995 ; Lechardeur et al., 1999). Binding of MARs-containing vector with nuclear matrix was belie v ed to confer long term expression Many DNA-binding proteins are found in MARs such as Topoisomerase II (Ra z in et al., 1991), Histone Hl (Izaurralde et al., 1989) lamin (Luderus et al., 1992) and SATB 1 (Dickinson et al., 1992) Topoisomerase II and Histone Hl are major enzymes found in MARs Topoisomerase II exhibits cooperative binding to MAR DNA and Histone Hl maintains a higher-order structure of chromatin (Boulikas 1995) Much evidence supports MARs / protein interactions and their biological function (Alvare z et al., 2000; Liu et al., 1999; Ramakrishnan et al., 2000; Stratling and Yu, 1999 ; Sun et al., 2001), although the relationships are still unclear. Matrix attachment regions are also boundary elements that define boundaries of the independent chromatin domain. They protect DNA from neighboring chromatin affecting transgene expression (position effect). Therefore, they enhance expressions (Kalos and Fournier, 1995; Klehr et al., 1992; Phi-Van et al., 1990 ; Poljak et al., 1994) MARs-containing plasmid DNA confers position-independent and copy number dependent expression (Phi-Van et al., 1990 ; Rollini et al., 1999 ; Stief et al., 1989). The studies showed that MARs enhances expression independent of orientation ; forward or reverse orientation from its natural form (Klehr et al., 1992; Mielke et al., 1990)

PAGE 16

7 However MARs can block expression when placed between promoter and enhancer (Stief et al., 1989). Matrix attachment regions are chromatin-remodeling elements with enhanced chromatin accessibility The combination of MARs and enhancer confers accessibility upon a distal promoter. This may be due to the generation of an extended domain of histone acetylation by MARs (Fernandez et al., 2001; Forrester et al., 1994; Jenuwein et al., 1997). MARs sequences also influence transgene methylation status (Dang et al., 2000; Forrester et al., 1999). MARs inhibit de novo methylation of retroviral 5'-LTR (long terminal repeat) Thus MARs inhibits promoter shutdown and leads to higher levels of expression (Dang et al., 2000). Instead of the conserved sequence, MARs have a different characteristic conformation (Yamamura and Nomura, 2001). Although MARs do not have consensus sequences (Singh et al., 1997), they do share similar characteristics with them Characteristics MARs as Potential Origins of Replication It is known that newly replicated DNA is specifically located on the nuclear matrix (Boulikas, 1993; Boulikas, 1995; Singh et al., 1997). Previous studies confirm that replication forks are associated with the nuclear matrix (Berezney and Coffey, 1975 ; Gerdes et al., 1994 ; Jackson and Cook, 1986; Ortega and DePamphilis 1998 ; Vaughn et al., 1990) and that replication machinery is immobilized by attachment to the nuclear matr i x (Cook, 1999). According to the model of Pardoll and coworkers, DNA is reeled through its nuclear matrix attachment site during replication. The replication origin is

PAGE 17

8 transiently attached to the nuclear matrix. The replication origins associate with the nuclear matrix in the late G1 phase and dissociate after initiation of DNA replication in S phase (Djeliova et al., 2001) MARs from yeast coincide with the putative origins of replication and Drosophila MARs can drive the autonomous replication of plasmids in yeast. In addition, the isolation and cloning of putative origins ofreplication from monkey cells in culture has shown them to possess sequence homology with MARs (Boulikas, 1993; Boulikas 1995). It has been shown that nuclear matrix attachment sites homeotic protein recognition and binding sites, and the origins of replication share the ATTA, ATTTA and ATTTTA motifs (Boulikas, 1993). Enzymes necessary for replication such as DNA polymerase, Topoisomerase II and primase, were found in MARs (Boulikas, 1995; Lodish et al., 1998). The deformation of the nuclear matrix protein and DNA by chemotherapy agents such as alkylating agents and by ionizing radiation affected replication, which led to cell death (Muenchen and Pienta 1999). This suggests that the differential activation of origins of replication may be regulated on the nuclear matrix. Studies show that both stimulation of replication by transcription factors and the presence of cis-acting elements distant from the origin of bidirectional replication are able to affect origin firing Matrix attachmet regions also affect expression and replication-timing patterns of translocated BCL2 oncogenes. The translocated allele replicates at the G 1/S boundary while the wild-type allele continues to replicate as usual in mid-S phase. These differences are accompanied by allele-specific changes in BCL2 expression, since the major breakpoint region (mbr) of BCL2, which is implicated in 70% oftranslocation from Chromosome 18 to Chromosome 14 present in human follicular lymphoma is a MARs

PAGE 18

9 (Sun et al., 2001). The AT-rich region flanking the BCL2 mbr is a binding site for the MAR protein SATB 1 (Ramakrishnan et al., 2000). Major Classes of Matrix Attachment Sites Known to Be AT-Rich Sequences These AT-rich sequences (approximately 100-1,000 base pairs in length) are closely related to the binding and cleavage consensus of Topoisomerase II (Gasser and Laemmli, 1986; Mirkovitch et al., 1984) and the binding site ofHistone Hl, both of which are important for chromosome structure (van Drunen et al., 1999). Topoisomerase II can cleave double-stranded DNA, pass an uncut portion of the DNA between the cut ends, and then reseal the cut. This is critical for DNA replication (Lodish et al., 1998). The most remarkable role of the AT-rich regions is that they facilitate unwinding of DNA previously catalyzed by helicase molecules, torsional strain, or proteins. The DNA unwinding is the first step for transcription, since it facilitates binding of RNA polymerase and other transcription factors to DNA. Under torsional strain, the DNA unwinding AA TAT ATTT motif, present within the MARs of both lgH and ~-interferon genes, becomes the site of nucleation (Bode et al., 1992). Mutation of this motif to ACTGCTTT voids both the MARs activities of the fragment as well as its DNA unpairing ability. Bode et al. (1992) showed that under various ionic conditions, the mutant form displays neither MAR activity nor unwinding capability The unwinding property was shown to be important for binding to the nuclear matrix and for augmentation of gene expression in stable transformants. The discovery of Topoisomerase II as part of the chromatin remodelling CHRAC complex suggests that these elements may influence transcription regulation via nucleosome remodeling ( van Drunen et al., 1999).

PAGE 19

10 Many MARs contain significant stretches of AT-rich sequences. It has been suggested that the simple occurrence of isolated AT-rich regions is not sufficient to cause matrix association. Rather several such regularly spaced motifs are required (Singh et al., 1997). MARs May Represent Mass Binding Sites for Protein Transcription Factors Studies show that the nuclear matrix is a microenvironment for transcription factors such as Myb, Myc, RFP, C/ EBP, AP-1, Spl, and NMP-1 (ATF) (Bode et al., 1995). Protein transcription factors such as large T antigen AP-1, and Sp 1 that bind to nuclear sites were found to stimulate replication and transcription in viruses, metazoans and yeast (Bode et al., 1995). In addition, DNA-binding proteins ofnoncoding DNA sequences, such as Topoisomerase II, Histone Hl, Jamin B 1 120-kDa protein (SP120) scaffold attachment factor A (SAF-A), attachment region binding protein (ARBP) and a thymus specific MAR-binding protein (SATBl), were found in the nuclear matrix (Boulikas 1995 ; Luderus et al., 1994) These specific proteins exhibit cooperative binding and distinguish between MAR and non-MAR DNA (Boulikas, 1995; Luderus et al., 1994) The AT-tract exhibits a high affinity for these proteins (Boulikas, 1995) In the studies over the last 10 years, various types of proteins were found in the nuclear matrix. SAF-B interacts with RNA polymerase II and a subset of serine/ arginine-rich RNA processing factors (SR protein) serve as a molecular base to assemble a transcription complex (Nayler et al., 1998). Heterogenous nuclear ribonucleoprotein (hnRNP) is the major protein component of the nuclear matrix and serves as the transcription factor of the central nervous system (Stratling and Yu, 1999)

PAGE 20

11 Transcription Enhancer May Be MARs A significant number of studies have shown that MAR sequences are located near or at enhancer sites. In addition, MAR sequences were shown to act as transcriptional enhancers in experiments involving cells in culture and in transgenic animals (Boulikas, 1993; Mielke et al., 1990). Removal of the MARs decreased the abundance of mRNA by a factor of 35 to> 1000 in B lymphocytes of transgenic animals (Boulikas, 1995). The MARs model created by Boulikas (1995) explains looping out of DNA and juxtaposing enhancers by synergistic interaction of classical MAR proteins ( composite MAR model) (Boulikas, 1995). The core enhancer is flanked by AT-rich sequences of about 300-500 bp able to sequester classical matrix proteins by cooperative interaction. This process of juxtaposing distant MAR elements and causing looping of DNA, brings together on the nuclear matrix two core enhancers or a core enhancer and a core origin ofreplication (100-200 bp) that cohabit with the AT-rich MAR (along with the transcription factors bound to them) to facilitate transcription and replication. MARs Harbor Intrinsically Curved DNA Curve DNA has been identified at or near several matrix attachments sites (Singh et al., 1997) Bending of double-strand DNA is mainly caused by homopolymeric dA of at least 4 bp, called A-tract (Haran and Crothers, 1989; Koo et al., 1986). Because of its A-tracts, MARs have a longer bent part and higher angle/helical tum than the other regions. The A-tract in MARs showed a nonrandom distribution but were clustered closely within MARs (Yamamura and Nomura, 2001). Intrinsically, curve DNA is important in nuclear processes involving specific protein-DNA interactions, such as recombination and transcription (Boulikas, 1993).

PAGE 21

12 Curved DNA motifs reflect a 10.4 nucleotide periodicity from center to center of (A)n stretches Optimal curvature is expected for sequences with repeats of the motifs AAAAn6AAAAn1AAAA and TTT AAA (Singh et al., 1997). A Class of MARs May Harbor Kinked DNA Kinked DNA has generally been associated with the presence of copies of the dinucleotides TG, CA, or TA that are separated by 2-4 or 9-12 nucleotides. For ex amp le, kinked DNA is produced by the motifTAn3TGn3CA, with TA, TG, and CA occuring in any order. The CA, TA, and TG dinucleotides are overrepresented in DNA sequences that are protein recognition sites Previously published studies show that MARs may display a usual richness of AT, TG, and CA (Boulikas, 1993; Singh et al., 1997) DNase I-Hypersensitive Sites May Be Diagnostic of MARs Gene activation in eukaryotes has been proposed to consist of a multistep process that includes changes in chromatin structure, modifications of hi stones, and transcriptional activation of promoter (Blackwood and Kadonaga, 1998; Felsenfeld et al., 1996; Grosveld 1999 ; Struhl 1998). The relationship of these events is still under investigation, but MARs might be involved in these processes The decondensation of the chromatin structure reflected by increasing sensitivity to DNase I digestion is simply a consequence of transcriptional activity (Klehr et al., 1992). Within transcriptionally active regions of chromatin, some sites are nearly as sensitive to DNase I digestion as naked DNA. These DNase I-hypersensitive (DNase IHS) sites occur in regions where transcription factors abound. A number of studies showed that MARs could induce DNase I-HS sites in chromatin. The DNase I-HS sites in the Drosophila histone gene repeat, and coincide with, Topoisomerase II cleavage sites

PAGE 22

13 and with the MAR sector in the Hl-H3 intergenic region (Boulikas, 1995). The MAR/ORI of the 5' flanking region of the chicken a-globin gene cluster harbors a constitutive DNase I-HS site that is detected in chromatin from many chicken tissues The studies in transgenic animals showed that, whereas the 95-bp lg core enhancer was necessary and sufficient for the accessibility of TF to T7 promoter, it required the MAR region to induce DNase I hypersensitivity (Boulikas, 1995; Forrester et al., 1994). Furthermore, the ORI of SV 40 minichromosomes is attached to the matrix and is hypersensitive to DNase I. From the structural point of view MARs are thought to be involved in chromatin condensation and chromosome formation. A synthetic AT-hook protein, which specifically binds to S / MARs, interferes with proper chromatin condensation in Xenopus laevis egg extracts (van Drunen et al., 1999). Moreover, alignment of S/MARs around the central core of mitotic and meiotic chromosomes may be required for correct condensation of DNA within these regions (van Drunen et al., 1999) Transcription activation of the gene during normal lymphoid development requires a synergistic collaboration between the enhancer and flanking MARs (Forrester et al., 1994). The enhancer in combination with a flanking MAR can confer accessibility on a distal site in nuclear chromatin, whereas the enhancer alone mediates only local chromatin accessibility (Jenuwein et al., 1997). The immunoglobulin MARs antagonized methylation-dependent repression oflong-range enhancer function (Forrester et al., 1999) The MARs allow the generation of an extended domain of hi stone acetylation, which could account for the long-range function of the enhancer in combination with MARs (Fernandez et al., 2001).

PAGE 23

14 Histone Hl, the most abundant repressor of gene activity (Boulikas, 1995) locks the two helical turns of the DNA around the nucleosome and maintains higher-order chromatin structures, which results in a more compact form of chromatin (van Drunen et al., 1999). Matrix attachment regions facilitated the displacement ofHistone Hl from chromatin through interactions with proteins with similar DNA binding motifs, such as HMG-I/Y (high motibility group) Competition between Hl and these HMG proteins may contribute to determining the global distribution of active and inactive chromatin. Also histone acetylation has been linked to transcriptional regulation via MARs. Hyperacetylation is a hallmark for active regions in the genome, while hypoacetylation is typical ofregions that are transcriptionally inactive (van Drunen et al., 1999) Functions Matrix Attachment Regions Enhance and Prolonge Gene Expression Matrix attachment regions are typically 70 % AT which has unwinding properties to enhance transgene expression (Bode et al., 1992). A single S IMAR construction of human IFN-P domain increases gene activity and the expression lev els mirror the S IMAR activity in vitro (Bode et al., 1995). Many studies reported in vitro function of the MARs-containing vector that led to the following preliminary conclusions: MARs function is independent of S/MARs orientation (Klehr et al., 1992 ; Mielke et al., 1990) MARs effects can be monitored if the marker and selector are physically coupled (located in the same gene) (Bode et al., 1992). MARs can block the enhancer effects if it is placed between the promoter and enhancer (Bode et al., 1995) Elevated expression levels in the presence of S I MARs are restricted to the stabl e expression, not the transient expression (Bode and Maass, 1988 ; K a los and Fournier

PAGE 24

15 1995; Klehr et al., 1992; Poljak et al., 1994; Wang et al., 1996). Therefore, MARs have been used to detect of integrated transgenes in transgenic mice embryos (Gutierrez-Adan and Pintado, 2000). MARs can form minidomains and, in the appropriate environment, MARs enhance regulatory effects, thus resulting in increased level of gene expression (Bode et al., 1995; Mielke et al., 1990; Stief et al., 1989). In stable transfection, the reporter genes flanked by certain MARs show position independent, copy number-dependent expression and augmentation of the transcriptional activity (Bode et al., 1992; Bode et al., 1995; Kalos and Fournier 1995; Stief et al., 1989). MARs replicate episomally in cells and have been stably maintained for more than 100 generations (Piechaczek et al., 1999). MARs influence transgene methylation status (Dang et al., 2000; Kirillov et al., 1996; Lichtenstein et al., 1994). The hIFN-P SAR inhibits de novo methylation of the retroviral 5'-LTR. The SAR element was able to alleviate methylation-mediated transcriptional repression (Dang et al., 2000). MARs can either be an enhancer or domain protector, the major function of which is to insulate transcription units from the regulatory influences of neighboring genes or chromatin domains. Matrix attachment regions are prime candidates for domain borders. They are able to mediate an attachment to the nuclear matrix in vitro and are frequently found in the vicinity of the ends of a domain that are otherwise defined by constitutive DNase IHSsites or a decrease of the general DNase I sensitivity (Bode et al., 1995) Although not all MAR elements are boundaries of chromatin domains, a few elements with presumptive boundary function have been defined (Kalos and Fournier, 1995). The MAR-containing vector is a good candidate for long-term expression of viral/nonviral vector because the boundary effect may enhance and prolong the level of gene expression (Auten et al., 1999 ; Murray et al., 2000). These elements can be used for creating minidomains in cultured cells. The segments of chromatin with multiple points of

PAGE 25

16 attachment may physically resist compaction into heterochromatin and gene silencing. The complete minidomains confer long-term stability to the enhanced expression level (Bode et al., 1995). Elements that can insulate transgenes from position effects have stable integrated transgenes and those transgene are expressed irrespective of their sites of integration (position independent expression) (Kalos and Fournier 1995). MARs, boundary element (BE), and locus control region (LCR) have position independent effects Thus, they are the candidates in the selection of stable cell lines with high-level expression characteristics. Among these entire elements, only MARs increase the proportion of high-expressing clones (Zahn-Zabal et al., 2001 ). Furthermore, MAR-containing vectors with a low copy transgene number are expressed at higher levels than high copy transgenes number (Bode et al., 1992; Kalos and Fournier, 1995). Cells harboring multiple-copy integration of the trans gene are transcriptionally inactive (Kalos and Fournier, 1995). This is based on the fact that classical transfection techniques resulting in a multiple random integration may affect the properties of transgene (Bode et al., 1995). Furthermore, random integration may lead to insertional mutagenesis and to silencing of the transgene (Piechaczek et al., 1999). Transfection methods, such as electroporation are required to promote integration of one copy per cell (Bode et al., 1995). The ideal vector for nonviral gene delivery should express the protein of interest and remain in the cell without integration for an extended time. Since the MAR element has a positive effect on transcriptional levels and is capable of stabilizing these high

PAGE 26

17 levels over extended periods ohime, they can be the instruments of construction for a new generation of expression vectors. Binding to Nuclear Matrix Is a Prerequisite for MARs Functions The physiological role of MARs is directly related to MARs binding to the nuclear matrix (Luderus et al., 1992; Luderus et al., 1994). The increasing transcriptional efficiency of chicken lysozyme results from attachment of the MAR element to the nuclear scaffold material (Stief et al., 1989) Therefore binding to the nuclear matrix is a prerequisite for transcriptional enhancement. An excess of nonbound S / MARs constructs would lead to a transcriptional level not distinguishable from that of a non-S I MAR construct (Bode et al., 1995) Mutation of the tumor suppressor, p53, gene modulates gene expression. This mutation is reconstituted in human cancer. Murine and human mutation p53, but not wild type p53, specifically binds with high affinity to a variety of MARs DNA elements (in particular AATATATTT, the potential base-unpairing sequences). MARs associtaed with human artificial chromosomes were found to persist as episomes in long-term culture with stability per generation of approximately 80% and were containing with cell nuclear matrix (Cossons et al., 1997). Extrachromosomal retention of Epstein-Barr virus base vector depends on a tran-acting element, Epstein Barr virus nuclear antigen-I (EBNA-1 ), and a cis-acting element, latent origin of replication, oriP (Gahn and Schildkraut, 1989; Harrison et al., 1994; Wysokenski and Yates, 1989). The nuclear retention function is thought to reflect an interaction between EBNA-1, oriP-containing plasmid DNA, and the nuclear matrix chromosomal scaffold and appears to be sufficient to hold plasmid DNA in the nucleus (Jankelevich et al., 1992; Wensing et al., 2001 ). However, incorporation of nuclear matrix attachment regions into

PAGE 27

18 the EBY Type 1 genome does not induce long-term expression of a foreign gene during latency (Makarova et al., 1996). Binding of MARs is cell-type specific (Will et al., 1998). MARs have been categorized as constitutive (permanent) or facultative (cell-type specific) (Singh et al., 1997) Constitutive MARs occur in all types of cells irrespective of the tissue in which they are found. In contrast, the presence of a facultative MAR is tissue-specific and that tissue governs its use. Moreover, some MAR binding transcription factors and classical matrix proteins are cell-type specific and bind strongly and selectively to MARs from different species (Boulikas, 1993; Boulikas, 1995). Plasmid DNA Topoisoforms and Transgene Expression Topology of pDNA, whether supercoiled, open circular, also called single stranded nicked, or linear form, is another aspect involved in the expression of pDNA. A number of studies showed that the topology of the transfected DNA molecules determines the level of gene expression (Kreiss et al., 1999; Ludtke et al., 1999; Pitard et al., 1997). The topological state of eukaryotic DNA seems to be important in gene expression, especially in the transcription process. For example, DNA supercoiling facilitates the formation of transcription preinitiation complex, which prevents subsequent assembly of promoter sequences into nucleosomes and allows transcription on the chromatin templates. In the fibroin gene, DNA superhelicity enhances transcription by accelerating formation of the complex (Hirose and Ohba, 1993; Hirose et al., 1985). A change in the superhelical density results in derepression of the yeast mating-type gene (Hirose et al., 1985)

PAGE 28

19 Enzyme topoisomerase, especially Topoisomerase II (Topo II) is important in the topology ofpDNA and also in the control of gene expression (Hirose and Ohba, 1993). Transcription of the mouse Hox-2.1 gene is inhibited by treatment of F9 embryonic carcinoma cells with a Topo II inhibitor etoposide (Hirose and Ohba, 1993). As an enzyme, Topo II can cleave double-stranded DNA, pass an uncut portion of the DNA between the cut ends, and then reseal the cut (Lodish et al., 1998), thus changing positive supercoiled DNA into negative supercoiled DNA. Transcription initiation of eukaryotic genes is thought to involve local unwinding of the DNA double helix within the promoter region (Hirose and Ohba, 1993). Topo II is one of the major enzymes found in the nuclear matrix and scaffold (Boulikas, 1995; Lodish et al., 1998) and is recognized by oligo (dA) tracts. The AT-rich region facilitates DNA unwinding when catalyzed by Topo II molecules (Bode et al., 1995). The AATATATTT motif present within the MARs of both IgH and ~-interferon genes becomes the nucleation site of a DNA unwinding effect under torsional strain (Bode et al., 1992). Binding of negative supercoiled SIMAR associate pDNA or non MAR pDNA with nuclear matrix overnight showed faint bands of nicked and linear form in the bound fragment, which were not present in the original mixture. These bands increased over time (Kay and Bode, 1994) thus demonstrating the action ofTopo II in the nuclear matrix. A binding assay study requires isolation techniques that highly preserve the underlying structure of the nuclear matrix while maintaining its functions. The separation methods of the nuclear matrix affect the quality of the structure and the binding site (Bode et al., 1995; Donev, 2000). After digesting DNA with nuclease,

PAGE 29

20 various methods are used to elute the chromatin fragment. Harsh procedures such as very high to extremely low salt extractions (2 M NaCl, 0.4 M KCl, or 0.25 M NaCl) were responsible for the less than complete preservation of fine structure that led to some rearrangements in the attachment of DNA loops. In contrast, mild extraction procedures with 25 mM lithium 3,5-diiodosalicylate (LIS) or 0.65 M ammonium sulphate gave similar results for the type ofMARs sequences investigated (strong, weak, non-MARs) (Donev, 2000) Utilization of amine modification, hydroxysulfosuccinimide acetate (sulfo-NHS), has been shown to preserve most of the nuclear matrix structure and its function (Wan et al., 1999). The efficiency of nuclear matrix / pDNA binding was greatly affected by the topological state of pDNA. Tsutsui and coworkers (1993) showed at least two classes of DNA-binding sites in the nuclear matrix: one is highly specific to supercoiled DNA in that it does not bind to relaxed or linear forms, whereas the other lacks this specificity (Tsutsui and Muller 1988). The first class of binding sites in the nuclear matrix selectively binds supercoiled DNA without sequence specificity This site was prepared by the mild extraction procedure of Mirkovitch et al. ( 1984) using LIS or nuclear halo with DNase I digestion. The scaffold associated DNA, after micrococcal nuclease digestion, has less DNA than that digested by DNase I. Therefore, it exposed additional DNA-binding sites in the nuclear scaffolds that are independent of ligand conformation. Consequently, it binds to relaxed, linear, and supercoild pDNA (Tsutsui and Muller, 1988). Like MAR associated pDNA, these topoisoform pDNA competed with single stranded pDNA (Kay and Bode, 1994; Tsutsui and Muller 1988)

PAGE 30

X31400 0 .90m X90200 0.20m Figure 2-1: Nuclear matrix isolated by amine modification, N-hydroxysulfosuccinimide acetate (sulfo-NHS) method. Chinese hamster ovary cells were washed twice with cold PBS and extracted in cytoskeleton buffer (10 mM Pipes pH 6.8 / 100 mM NaCl/300 mM sucrose / 3 mM MaCh/ 1 mM EGTA/lg/ml leupeptin/1 g/mL pepstatin/2 g /mL aprotinin/1 g/mL antipain/1 mM aminoethyl benzensulfonyl fluoride/10 units /mL prime RNase inhibitor) containing 0 5 % Triton X-100 for 7 min at 4C to remove the soluble proteins After extraction, cells were treated with 600 units /mL each of the restriction enzymes (Pstl and Hae III) in cytoskeletal buffer at 32 C for 1 hr. Then extracted cells were exposed to 2 mg/mL freshly prepared sulfo-NHS-acetate (in cytoskeletal buffer pH 7.0) for 20 min at room temperature. After being washed cells were again treated with 2 mg/mL of sulfo-NHS for 20 min at room temperature Cells were then washed with 10 mM glycine to quench the excess blocking reagent. The final pellet were resuspend in cytoskeleton buffer N ......

PAGE 31

22 The binding ofMARs-containing pDNA with nuclear matrix involves MAR protein. To date, several matrix proteins that specifically interact with MAR in vitro hav e been identified. Some proteins, such as lamin NB, bind with supercoiled and single stranded MAR DNA, while some proteins such as SATB 1, have a high affinity for supercoiled MAR DNA. Furthermore, proteins such as Histone Hl have no preference when binding with single-stranded Binding of the matrix proteins with DNA demonstrated the function of MAR, which is involved in active and inactive chromatin (Luderus et al., 1994 ) The preliminary data showed that, in addition to the supercoiled form, the open circular and linear forms of non-MAR pDNA were also transcriptionally active Studies relating to the stability of pDNA and the amount of pDNA associated with the cells were performed As previously mentioned the binding of pDNA to the nuclear matrix may be an important factor leading to the difference in transgene expression observed in DNA topoisoforms

PAGE 32

CHAPTER3 RELATIONSHIP BETWEEN PLASMID DNA TOPOLOGICAL FORMS AND IN VITRO TRANSFECTION Introduction Plasmid DNA (pDNA) is the basic tool for nonviral gene delivery. Understanding the conformational state of pDNA could assist in developing approaches to increasing its transfection efficiency. Important aspects for consideration include size and topology. The size limit for passive diffusion oflinear pDNA into the nucleus was found to be between 200 and 310 base pair (bp ); DNA of 310-1500 bp required energy to enter the nucleus (Ludtke et al., 1999). While smaller DNA molecules have higher transfection efficiency than the large one, the topology of the transfected DNA molecules determines the level of gene expression (Kreiss et al., 1999; Ludtke et al., 1999; Pitard et al., 1997). The supercoiled form ofpDNA was shown to produce a higher level of transgene expression than nicked circular or linear DNA (Hirose et al., 1985) Moreover, the amount of supercoiled form in the preparation is indicative of the stability and activity of pDNA preparation. However during DNA isolation or the formulation process DNA can be altered by shear stress to convert the supercoiled plasmid form to open circular (single stranded nicked), linear, or even the small fragment (Adami et al., 1998). This conformation change may affect the transfection efficiency of pDNA. Gel electrophoresis is one of the most direct and sensitive approaches for examining plasmid DNA stability in cell lysate solution, rat plasma, and serum (Houk et 23

PAGE 33

24 al., 1999; Lodish et al., 1995). This technique separates plasmid DNA on the basis of size and compactness; smaller and/or more compact molecules will migrate more rapidly through the matrix of the gel (Bates and Maxwell, 1993). The objective of this study was to investigate the relationship between the topological form of pDNA and its transgene production Since size was one of the factors affecting transfection efficiency, similarly sized pDNAs were utilized Plasmid pCMV (6.7 kb) is driven by cytomegalovirus promoter, while pGL3 (5.2 kb) is driven by Simian virus (SV 40) promoter. Because of its higher affinity to RNA polymerase, the cytomegalovirus promoter is stronger than the SY 40 promoter. In order to account for differences in cellular uptake ofpDNA, we utilized the flow cytometry method to measure the amount of pDNA associated with the cells. The stability of pDNA in cytoplasmic fraction was determined. These relationships are important for the future development of nonviral gene transfer. Materials and Methods Plasmid DNAs In this study we used pDNA, pGL3 control, GeneBank accession number U47296 (Promega, Madison, WI) and pWiz or pCMV-luciferase (Gene Therapy System, San Diego, CA), which differ in promoter sequences and size. Supercoiled pGL3 (5256 bp) containing SY 40 promoter was linearized with BamH I (Promega, Madison WI) Supercoiled pCMV-luciferase (6732 bp) that contains cytomegalovirus (CMV) promoter was linearized with XmnI (Promega, Madison, WI). These enzymes cleaved the bacteria part, which is not necessary to the expression of the plasmid. One hundred units of

PAGE 34

25 enzyme were used to cleave 88 g of pDNA at 37C for 4 h. The enzymatic reaction was then extracted with phenol/chloroforrn/isoamyl alcohol (25:24:1 v /v). Linear DNA was purified by ethanol precipitation at -20C followed by centrifugation at 13,000 xg for 20 min at 4C. The pDNA was analyzed using 0 .8% agarose gel containing 0.5 mg/mL ethidium bromide. Open circular pDNA was prepared by creating single-strand nicks in supercoiled DNA One mg ofpDNA was incubated in 1 mL ofTris-EDTA buffer at 70C for 6 hours and was concentrated via ethanol precipitation as previously reported by Adami and coworker (Adami et al., 1998). Liposome Preparation DOT AP/DOPE (l ,2-dioleoyl-3-trimethylammonium-propane / L-dioleoyl phosphatidyl-ethanolamine) liposomes were prepared as previously described by Tang and Hughes (1998). Briefly DOT AP and DOPE were dissolved and mixed in 1 : 1 molar ratio in chloroform The solution was evaporated in a round-bottomed flask using a rotary evaporator at room temperature. The lipid film was dried using nitrogen for an additional ten minutes. The lipid was then suspended in sterile water to make a concentration of 1 mg / mL. The mixtures were shaken for 30 minutes, followed by sonication by using Sonic Dismemebrator (Fisher Scienctific, Pittsburgh, PA) for 5 minutes at 5 watts to form homogenized liposomes The particle size distribution of the liposomes was measured using a NICOMP 380 ZLS instrument (Santa Babara CA) with the volume-weight distribution parameter. The approximate mean diameter of these liposomes was 200 20 nm. In order to study the interaction of pDNA and cationic liposomes, pDNA was extracted by phenol/chloroform immediately after the complexation with liposomes followed by the gel retardation study. The ratio of

PAGE 35

26 pDNA:liposome was 1 :2 w/w. There was no conformation change ofpDNA after liposomes complexation ( data not shown), indicating that complexation of pDNA with liposomes did not alter pDNA topoisoform. Transfection Experiment Chinese Hamster Ovary (CHO) or human neuroblastoma (SKnSH) cells were grown in a-minimum essential and RPMI 1640 media (Gibco BRL), respectively, supplemented with 10% fetal bovine serum, penicillin (100 units / mL), and streptomycin (100 g/mL). All cells were maintained in humidified air at 37C and 5% CO2 Cell lines were cultured and seeded in 24-well plates (1 x 105 cell/well) and grown to 60 to 80 % confluence in 1 mL of media. Each topoisomer pDNA was complexed with cationic liposomes, DOT AP/DOPE, at a 1 :2 w/w ratio of pDNA:liposomes for 30 min in serumfree media. Before transfection, serum-containing media was changed to serum free media and the transfection mixtures added. After 4 hours, the media were changed to growth media containing serum and the cells were grown for another 48 hours. For electroporation, CHO and SKnSH cells were grown to 80% confluence in a 175-mL flask in 25 mL media and harvested using typsin/EDT A. Cells were washed once with phosphate buffer saline (PBS) and then twice with electroporation media (10 mM HEPES, pH 7.2, 150 mM NaCl, 5 mM CaCh). The electroporation was performed using Gene Pulser II (Biorad, Hercules, CA) with 450 volts, 350 F, and 20 n. The cell concentration for electroporation was 3x I 06 cells / pulse and pDNA was IO g / pulse Transgene production was determined by measuring luciferase expression after 48 hours of incubation for both methods as previously described (Tang and Hughes, 1998). Briefly, cells were rinsed twice with PBS then 100 l of luciferase lysis buffer (0.1 M

PAGE 36

27 potassium phosphate buffer, pH 7.8, 2 mM EDTA, 1 % Triton X-100, 1 mM DTT) was added to the cells. Luciferase activity was quantified by using luciferase assay buffer (30 mM Tricine 3 mM ATP, 15 mM MgS04 10 mM DDT, pH 7 .8) and 1 mM D-luciferin (pH 6 3) (Molecular Probes, Eugene OR) combined with cell lysate. The light emission over a 10-second reaction period was integrated using a luminometer (Mono light 2010 San Diego, CA). This experiment was done 3 times with 4 replicates each time. Luciferase expression was reported in RLU / well under the assumption of non-toxic protein Cytoplasmic Stability Study Cytoplasmic fractions were obtained from CHO cells (1 x 106 cells) via treatment with a solution containing lysis buffer (150 mM NaCl, 10 mM Tris-HCl pH 7 .9, 1.5 mM MgCh, and 0.5% Nonident NP40) for 15 minutes (Piva et al., 1998). The ratio of the cell / solution was 1 x 106 cells / mL. After centrifuging at 10,000 xg for 10 min, a supernatant representing crude cytoplasmic fraction was obtained. This contained cytosol with different membranes and components originating from dissolution of such cellular vesicles as endosomes, lysosomes, and small-sized organelles. Forty g of pDNA was incubated at 37C in 1 mL of the cytoplasmic fraction. After incubation for an appropriate time, all samples were phenol/chloroform extracted. The topological state of DNA was determined by gel electrophoresis. Contamination of the cytosolic extract with lysosomal enzymes was assessed by measuring the P-galactosidase activity of the extract. The enzyme was determined fluorometrically as previously described (Storrie and Madden, 1990) The assays showed that 11 % of the luminal content of lysosomes was released into the extracellular medium, indicating minimal organelle contamination.

PAGE 37

28 Cytoplasmic fractions of whole cells obtained from lysis buffer (0 1 M potassium phosphate buffer, pH. 7.8, 1 % Triton X-100, 1 mM DTT, and 2 mM EDTA) showed that 24 % of the luminal content of lysosomes was released into the extracellular medium Therefore, the stability of pDNA in cytosolic fraction was mainly affected by the nuclease activity associated with the cytosol. Cellular Uptake Study of Plasmid DNA Plasmid DNA was covalently labeled with fluorescein using an IT nucleic acid labeling kit (Pan Vera Corp., Madison, WI) according to the manufacturer's instructions The labeled pDNAs were purified twice on microspin columns, then ethanol-precipitated. The beginning amount of labeled pDNA was standardized to the same fluorescence intensity. In the supercoiled preparation, a significant conformation change from supercoiled to open circular after fluorescein labeled was found. Therefore, this preparation was excluded from the experiment. The labeled pDNA:liposome (DOTAP /DOPE with 1 : 2 w / w ratio ofpDNA/liposome) complexes were transfected into CHO cells (2 x 105 cell / well) as described in the transfection section. After 48 hours of incubation, the cells were harvested using 0 05% trypsin/0 .53 mM EDTA, then transferred to 1.5 mL tubes, centrifuged at 250 xg for 5 min, washed twice and resuspended in 800 l of PBS. Flow cytometry was performed using a F AC Sort (Beckton, Dickinson, San Jose, CA) Fluorescein was monitored with a 530 / 30 bandpass filter and photomultiplier tube pulses were amplified logarithmically. Ten thousand cells were counted at a flow rate between 100 and 200 cells per second. Cells were gated with their morphological properties, forward scatter, and side scatter, set on linear mode. The mean fluorescence intensity of the related populations of cells was calculated using

PAGE 38

29 histograms and expressed in arbitary units corresponding to an intensity channel number ranging from Oto 1023 Data Analysis Gel analysis was performed using Kodak 1 D40 image analysis software version 3 0 (Rochester, NY). The data obtained from gel analysis was fitted for kinetic study to estimate the kinetic parameter of pDNA in cell lysate. Data were analyzed using the Micromath Scientist program, version 2.0 (Salt Lake City, Utah). Statistical analysis was performed using StatView software version 4 5 (Abacus Software Berkeley CA). The ANOV A factorial test was used in this analysis with a 95% confidence interval. The data were considered to be significantly different when the p-value was < 0 .05. Background From the stability study of pDNA in a rat plasma model by Houk and coworker (Houk et al. 1999) the degradation of supercoiled pDNA was assumed to follow pseudo first-order kinetics as shown in Figure 3-1. The assumptions for this model are, first, pDNA degradation is considered to be a unidirectional process Second, the degradation of linear pDNA is considered to yield fragments of heterogeneous lengths thus the degradation products from the linear pDNA were excluded from the fitted model. Finally, no elimination from any of the compartments is assumed to occur through routes other than degradation to the subsequent topoisoform. SC ... oc ... Ln Figure 3-1: Diagram of first-order kinetics (Houk et al., 1999)

PAGE 39

30 Based on this model, the following differential equations were derived to describe the process in Figure 3-1 (Houk et al., 1999) dSC = -ks SC dt (3-1) (32 ) (3-3) The amounts of supercoiled open circular and linear pDNA were then modeled using the integrated form of the following equations: SC = SC0 e k (3-4) (35 ) where SC, OC and L are the amounts of supercoiled, open circular and linear pDNA present at time t, respect i vely SC0 is the amount of supercoiled pDNA present at time (t) = 0. The constants k s ,ko, and k1 represent the rate constants for the degradation of supercoiled open circular, and linear pDNA, respectively The constants represent the activity of all enzymes acting in the degradation process Nonlinear curve-fitting and statistical analysis were carried out using Scientist (version 4 0 Micromath, Salt Lake City, UT)

PAGE 40

31 Results Figure 3-2 shows photographs of agarose gel electrophoresis of pDNA. Plasmid pCMV-luciferase and pGL3, heated at 70C for 6 h, resulted in approximately 70% relaxed circular form for each. Supercoiled plasmid DNA treated with restriction enzyme demonstrated almost 100% linear form Plasmid DNA starting with aliquots of 70% supercoiled (SC) pCMV, 80% SC pGL3, 70% open circular (OC) pDNA, and 100% linear (Ln) pDNA was delivered into CHO and SKnSH cells using either cationic liposomes or electroporation. The percentage of each topoisoform was calculated based on the binding affinity of each topoisoform to ethidium bromide in gel analysis. CHO and SKnSH cells gave similar results in transfection efficiency (TE) of the topoisoform, although the extent of the expression was cell-line dependent. Plasmid DNA delivered by pDNA/liposomes, open circular pCMV showed significantly higher transfection efficiency than the supercoiled or linear forms (Figure 33) There was a significant difference in TE of the topoisoform of pGL3 in CHO cell whereas there was no significant difference in those of pGL3 in SKnSH cell (Figure 3-3). Plasmid DNA delivered by electroporation shows only a slightly difference in TE between the cell lines. In CHO cells, there was no significant difference between TE of supercoiled and TE of open circular pCMV (Figure 3-4), whereas in SKnSH cells, TE was significantly higher in supercoiled pCMV than in open circular form. There was no significant difference in TE between the topoisoform of pGL3 for either cell line (Figure 3-4).

PAGE 41

A. STD 1 2 3 oc Ln SC 32 B. STD 1 2 3 Figure 3-2: Agarose gel electrophoresis of plasmid DNA. (A) topoisomer ofpCMVluciferase: Lane 1, supercoiled preparation contained 70% supercoiled (SC) pCMV luciferase; Lane 2, open circular preparation contained 80% of open circular (OC) pCMV-luciferase; and Lane 3, linear preparation contained almost 100% linear (Ln) pCMV-luciferase. (B) topoisomer ofpGL3: Lane 1, supercoiled preparation contained 80% of supercoiled pGL3; Lane 2, open circular preparation contained 65% of open circular pGL3; and Lane3, linear preparation contained almost 100% of linear pGL3. STD= Standard Lambda Hind III DNA marker. nr' Ln

PAGE 42

A. uS' 16 "'"" 14 12 -~ 10 8 6 4 0 2 .3 0 > () a. Lt) 0 .... B. 25 20 ...... M' .::;o ...1 15 s; (!) a. .. ,t~ 10 Q) >< en ca "'-0 ::::s ...J 5 SC oc SC oc 33 Ln Ln Figure 3-3 : Transfection efficiency of pDNA-liposome complex (1:2 w / w ratio) in (A) Chinese hamster ovary (CHO) cells and (B) Human neuroblastoma (SKnSH) cells Transfection efficiency of liposome complex supercoiled (SC) complex open circular form (OC) and complex linear form (Ln) are presented. N = 4. The error bars represent mean S D (A) CHO : There was a significantly difference (p < 0.05) in each topoisoforrn of pDNA. (B) SKnSH: represented significant difference in transfection efficiency (TE) of OC and the other forms ofpCMV-luciferase. There was no significant difference between the TE of each topoisoform of pGL3. Transfection efficiency is represented by the relative light unit (RLU) / well.

PAGE 43

6 b Es ~4 > .:. 3 Q) 2 c:; 1 ::::J ...J 0 >-(") -_J s; (9 Q. -0 ..; ns Cl) X (/) > ns ... Cl) (.) Q. 0 (0. :::I 0 _J rA. SC B. 10 8 6 4 2 0 SC 34 oc oc Ln Ln pCMV DpGL3 pCMV OpGL3 Figure 3-4: Transfection efficiency of pDNAs by using electroporation (450 volt, 350 F 20 Q) in (A) Chinese hamster ovary (CHO) cells and (B) Human neuroblastoma (SKnSH) cells in relative light unit (RLU) / well. Transfection efficiency of supercoiled (SC), open circular (OC), and linear (Ln) pDNA is presented. N = 3. The error bars represent mean S.D. (A) CHO cell,* represents the significant difference between QC and Ln pCMV-luciferase. No significant difference between transfection efficiency of SC and OC pCMV-luciferase and topology ofpGL3 was observed. (B) SKnSH cell,* represents the significant difference between SC and the other topoisoforms of pCMV.

PAGE 44

35 For transfection methods, the ratio of pDNA:number of cells for liposomes complex and electroporation was 1:0 3x105 and 1:3x105 respectively. The number of cells used for electroporation was 10 times higher than for liposomes transfection becaus e the pulse electric field killed a large portion of cells. Stability of plasmid DNA in the isolated cytoplasm was also addressed Plasmid DNA incubated at 37C in cell lysate solution exhibited degradation of pDNA during a set period of time Both pCMV and pGL3 topoisomers showed similar patterns of degradation (Figure 3-5) Supercoiled pDNA could be detected for up to 4 min while open circular and linear form remained for 6 to 10 min The kinetics of pDNA in cell lysate solution exhibited the pseudo-first-order of degradation at an initial amount of 40 g pDNA. Table 3-1 gives the rate constants and half-lives of the topoisomers ofpDNA. In order to examine possible reasons for differences in expression, studies wer e conducted to determine whether similar amounts of DNA complexes were associated w ith the cells CHO cells were incubated with fluorescein-labeled pDNA (OC or Ln) lipid complexes for 48 hat 37C harvested with trypsin/EDT A and washed with PBS as described previously above. Figure 3-6 shows the mean fluorescence intensity of the labeled pDNA in cells. Open circular pDNA showed higher intensity, indicating complexes made with this form of pDNA associated more with cells than linear forms of pDNA at 48 hours after transfection. Discussion and Conclusion CHO and SKnSH cells transfected with pDNNliposome complexes demonstrated similarities in pattern but differed in levels of transgene produced (Figure 3-3)

PAGE 45

Std 0 5 10 15 20 25 30 35 40 45 min Supercoiled pCMV Std O 5 10 15 20 25 30 35 min Supercoiled pGL3 Std. 0 5 10 15 20 25 30 35 40 45 min Open circular pCMV Std 0 5 10 15 20 25 min Open circular pGL3 Std O 5 10 15 20 25 30 35 40 45 min Linear pCMV Std O 5 10 15 20 25 30 35 min L i near pGL3 Fig u re 35 : Agarose ge l e l ectrop h ores i s of p lasmid DN As i n c ub ated in cytop l asm i c so lut io n at 3 7 C in t h e peri od of t ime. (A) pC MV-lu ciferase t o p o i sofor m ( B ) p GL 3 to p o i sofo rm.

PAGE 46

37 Table 3-1: Pharmacokinetics of plasmid DNAs in cytoplasmic solution. Plasmid DNA Rate constant (min-1 ) Half life (min) pCMV Supercoiled 0.20.04 3.49 66 Open circular 0.11.02 6 52.65 Linear 0.11.02 6.19.12 pGL3 Supercoiled 0.19.03 3.57.65 Open circular 0.06.01 10.90.89 Linear 0.14.07 4 85.01 Kinetics of plasmid DNAs in cytoplasmic solution calculated using the Micromath Scientist program. Represented the significant difference between the half-life of OC and other topoisoforms of pGL3.

PAGE 47

;;,., .-:: "' = Q,l ..... = Q,l (,J = Q,l (,J "' Q,l i.. 0 = rz 38 700 600 500 400 pCMV 300 DpGL3 200 100 0 cell oc Ln Figure 3-6: Mean fluorescence of the labeled plasmid DNAs (OC open circular a nd L n linear forms) complex with DOTAP/DOPE liposome (1 : 2 w / w rat i o) transf e cted into CHO cell. The cell was analyzed by flow cytom e try FACSort after 4 8 h of incubation at 3 7 C The number ofreplication, N, is 4 The error bars represent mean S D represented the s i gnificantly different between OC and Ln forms of pCMV No signific a ntly different between mean fluorescence intensity of topology of pGL3

PAGE 48

39 This indicated the cell type was important in transfection efficiency (TE) SKnSH cells expressed lower amounts of luciferase activity than CHO cells, especially in SY 40 promoter plasmid compared to the stronger CMV promoter plasmid. Researchers have previously shown that cells may have different levels and types of enzymes (Piva et al., 1998) involved in the pDNA process leading to alter transgene production. This might be one of the reasons for the similar patterns but varying levels of expression in the two cell lines. Two important findings concerning pDNA/liposome complex were evident. First open circular and linear pDNA were as active as supercoiled form, and open circular pCMV showed a higher transgene production than supercoiled form. The import steps to getting pDNA into the nucleus include, 1) pDNA entering into cell, 2) movement through cytoplasm, and 3) pDNA entering into nucleus via nuclear pore complex. In the first step, pDNA delivered by liposome enters the cell by endocytosis (Gao and Huang, 1995; Xie et al., 1992), whereas pDNA delivered by electroporation enters by transient membrane pores (Courey and Wang 1983; Weaver 1993) Neither of these methods is thought to be topologically dependent (Chernomordik et al., 1990; Weaver, 1993; Xie and Tsong, 1993). Extrapolating from our data, 48 hours after transfection OC appeard to associate in cells more than does Ln pDNA (Figure 3-6) Plasmid DNA delivered by electroporation showed a pattern of expression among each topoisoform similar to pDNA delivered by liposomes, but the forms differed in the extent (Figures 3-4) Although the correction was made for the number of cells dying from electrical shock, it is possible that the initial amount of pDNA entering the cell might be different for each delivery methods. Moreover, the low transfection efficiency for

PAGE 49

40 electroporation might be due to the rapidly shrinking of the transient pore size. The difference in amount might affect the level of expression, and the excess of pDNA in cytoplasm might saturate enzyme activity. Stability of pDNA in cytoplasm is an important aspect in Step 2 of the pathway. Plasmid DNA has been reported to be destroyed mainly by cytosolic nucleases, including endonuclease and exonuclease (Lechardeur et al., 1999). The identification of which cytosolic nuclease(s) is or are involved in pDNA degradation needs to be determined in order to understand this mechanism. Studies by Lechardeur and coworker ( 1999), excluded DNasel and DNasell, which display an acidic pH optimum, from the degradation of pDNA in cytosol, since the maximum nuclease activity was attained between pH 7 and PH 8. Our data demonstrated that the OC form exhibited the longest half-life in the isolated cytoplasm of CHO cells, followed by Ln and then SC form (Table 3-1 ). The stability studies of pDNA in the rat plasma model showed OC form had the longest half-life (21 min), followed by Ln (11 min) The authors' data support previous results that indicated the SC form was the least stable form of pDNA in the rat plasma model (half-life of 1.2 min) (Houk et al., 1999). Most likely there is a plasma enzyme involved in the degradation of pDNA other than cytosolic endonuclease, and there might even be a special enzyme sequence in SC form that results in faster degradation. The movement of pDNA in cytoplasm through the cytoskeleton and the immobile object is another barrier for pDNA trafficking (Wilson et al., 1999). Fully condensed pDNA, the SC form, may have a higher binding affinity with these elements, thus increasing the chance of being destroyed by cytosolic enzyme. It is not easy to explain the differences in half-life between the OC forms found in pCMV and pGL3. These may be due to the

PAGE 50

41 different attraction of nuclease enzymes or to the method of production of OC DNA resulting in its being nicked in difference places Because of its larger size OC pCMV might be more susceptible to enzyme activity than OC pGL3. In pDNA/liposomes complexes, the patterns of e x pression were not relative to the stability of pDNA in cytosol (Figure 3-3), probably due to the protection provided by lipids to the pDNA (Lechardeur et al. 1999). The final step is when pDNA enters the nucleus. Plasmid DNA nuclear import is facilitated by cell division. Therefore nuclear import might not be a major barrier in transgene expression in our case, since both CHO and SKnSH cells are div iding durin g import. Although SY 40 promoter has been reported pre v iousl y as responsible for facilitating the specific sequence that enhanced nuclear import (Wilson et al. 1999) we noticed that CMV-based pDNA demonstrated higher expression than SV40-based DNA The superhelical (supercoiled) state of pDNA was considered to be important in transcription process in eukaryotes (Gellert 1981; Hirose and Ohba 1993 ; Hirose et al., 1985; Mizutani et al. 1991; Shlyakhtenko et al., 1998; Singha! and Huang 1994 ; Wang, 1985) Our results demonstrated that compared to the SC form of each pDNA OC pCMV is still active while OC pGL3 was not active (Figure 3-3). The reason may be that the different promoter sequences lead to different levels of transcription initiation Although the SC form facilitated the formation of the transcription preinitiation comple x within the promoter area and prevented it from being included in heterochromatin which causes the silencing of the gene (Singha! and Huang, 1994 ), in some genes such as the Drosophila Hsp70 gene closed circular DNA had the same rate of preinitiation complex formation as supercoiled DNA (Singha! and Huang 1994), resulting in the sam e prote i n

PAGE 51

42 express10n. In our case OC CMV might have facilitated the local un w inding of DNA without the help of the SC form resulting in the same or higher amounts of luciferease activity. However the reason why OC pCMV delivered in the same amount as SC pDNA showed higher transgene expression than the SC form is still unclear. A possible reason might be the loss of the SC form in cytoplasm (Table 3-1). The second method of making OC DNA by random nicking may affect the efficiency of pDNA If the singl e stranded nick occured in a crucial place, the missing base pairs in OC pDNA might i mpair its function This might be another reason to explain the low activity of OC pGL3 ( Figure 3-3) In this study we concluded that regardless of the nonviral delivery method pDNA topology would not effect transgene expression, even though the pharmacokinetics of each isofroms are distinct. While, the delivery methods (electroporation and liposomes) facilitate the transport of pDNA across cell membranes and protect pDNA from cyotplasmic enzyme (liposomes), other factors for improving transfection efficiency need to be considered. These factors include the formulation for pDNA preservation (pH salt effect and ionic strength) the quality of pDNA (impurity, presence endotoxin and amount of each topo i soform) and the cell type used for transfection. Supercoiled DNA is the most accepted form of pDNA for gene transfer because it is the initial form of DNA and it is the initiation form of DNA in the transcription process. Howe v er, the stability of SC pDNA in the cytoplasm may hinder the level of expression. The other forms of pDNA, OC and Ln forms are more stable in cytoplasm and are transcriptionally active.

PAGE 52

CHAPTER4 USE OF PHARMACOKINETIC PARAMETERS TO INTERPRET GENE EXPRESSION Introduction Plasmid DNA (pDNA) either naked or complexed with cationic molecules is a frequently used vector for gene delivery since it does not exhibit the limitations, such as eliciting adverse immune responses (Hengge et al., 2001) insertional mutagenesis (Hengge et al. 2001) and the size limitation of the transgene (Kre i ss et al., 1999) that characterize viral vectors. Despite certain advantages for gene delivery of naked pDNA, there are fundamental problems associated with the use of unprotected pDNA. In particular, the reduction of genome equivalents by enzyme endonucleases will translate into a loss of gene expression (Hengge et al., 2001 ) A more efficient nonviral method of delivery must be used to better deliver the vector to target cells. These methods can be characterized as (a) chemical methods such as liposome delivery receptor-mediate endocytosis and long-term expression vector, and (b) physical methods such as injection or particle bombardment (Wohlgemuth et al. 1996). Despite such sophisticated techniques, the development of an efficient delivery system also depends on accurate analytical methods. In vitro transfection is a procedure in which tissue culture cells are incubated with a plasmid and, after a certain time, the expression of the encoded gene is measured v ia the enzyme activity or concentration of the synthesized protein (Lasic, 1997) Gene 43

PAGE 53

44 expression can be measured as a function of time but is often reported as a single value of peak expression. However the process of gene regulation, including transcription translation and post-modification, requires time. Moreover, i t also depends on time for DNA to be initialized by cells, for DNA trafficking through the cytoplasm to the nucleus to half-life (the time necessary for the concentration of the drug or pDNA in the plasma or cells to decrease by one-half) of intracellular DNA and mRNA. Time also determines that phase of the cell cycle and the cell turnover rate. Various cell processes and conditions result in a time variation in expression. Thus a single-point measurement may not be the best way to compare delivery systems and would be similar to relating therapeutic lev els of a drug in the blood by using such a parameter as peak drug concentration (CMAx). Pharmacokinetics are the study of the time course of drug absorption distribution metabolism, and excretion (Dipiro et al., 1997). Pharmacokinetics have been utilized for characterization ofpDNA in vivo (Lew et al., 1995 ; Mahato et al., 1995 ; Osaka et al., 1996; Schubeler et al., 1996) and the fate ofpDNA in vitro (Houk et al., 2001) Gene transfection and gene expression of pDNA can also utilize pharmacokinetic concepts since transgene expression can be followed over time by treating pDNA as a prodrug and measuring its active methabolites. From the plot of trans gene expression against time pharmacokinetic parameters for gene expression can be calculated that would be equivalent to the standard parameters of area under the curve (AUC) and mean residence time (MRT). We have used noncompartmental methods for estimation of certain pharmacokinetic parameters under the assumptions of linear pharmacokinetics with first order elimination and stable protein product. AUC or in the case of gene expression

PAGE 54

45 area under the expression curve (AUEC), can indicate the total amount of protein produced and mean expression time (MET) can indicate the average time of the expression. Thus, it can indicate not only the delivery of the genes into cells, the gene transfection, but also the amount of protein encoded by the delivered DNA, the gene expression (Lasic, 1997). In this chapter, we address two cases of pDNAs-based delivery systems. The first is a simulation oftransgene expression of four pDNAs with identical maximum expression (Cmax), but different times of maximum expression (tmax). The second is the experimental trans gene expression of two plasmids coding for luciferase, but different in plasmid construct. Plasmid pGL3 contains SV 40 promoter and an ampicillin-resistance gene. Plasmid pGM is pGL3 with an insertion of matrix attachment regions (MARs) that may enhance or prolong the expression. The ideal pDNA for gene delivery requires high and long-term expression. Thus, comparing plasmid transgene activity in both cases needs more than one point of measurement. In this paper, we utilized the multiple-point measurement together with pharmacokinetics to calculate gene expression for Case 1, (plasmids 1, 2, 3, and 4) and Case 2 (pGL3 and pGM). This method will allow comparing transfection efficiency of pDNA through the gene product and will provide more insight for data analysis. Materials and Methods Plasmid DNAs Plasmid pGM was constructed from pGL3-Control vector (5.2 kb) (GeneBank number U47296) (Promega Madison, WI), which contains a modified firefly luciferase

PAGE 55

46 gene and SY 40 promoter. A 2 0 kb of MARs from 5'-region of hIFN-~ (Piechaczek et al., 1999) was inserted at position 2442 downstream of the SY 40 enhancer of pGL3 to create 7.2 kb MAR-containing vector (Figure 4-1). Plasmid DNA was obtained from E. c oli (strain DH5a) that was transformed with a pGL3 or pGM plasmid Plasmid DNA was isolated using a Megawizard DNA purification kit (Promega, Madison WI). The concentration and purity of pDNAs were determined spectrophotometrically. The average concentration ofpDNAs was 1.8 mg/ml and purity ofpDNA, A260/ A280 ratio was 1 8 Cells Chinese hamster o v ary (CHO) cells were obtained from American Type Cultur e Collection, MD. Cells were incubated at 37C in a humidified atmosphere containing 5 % CO2 and maintained in Minimum Essential Medium-alpha (a-MEM) medium with 10% fetal bovine serum (FBS), 100 g/mL of streptomycin and 100 units / mL of penicillin ( all from Gibco BRL Grand Island, NY) At 24 hours before transfection, cells were seeded at 50 000 cells per well in 12-well plates (Coming Costar Corp., Cambridge MA) in a final volume of2 mL ofmedim. Preparation of Cationic Liposomes and Transfection Experiment DOTAP/DOPE (l, 2-dioleoyl-3-trimethylammonium-propane / L-dioleoyl phosphatidylethanolamine) liposomes were prepared as previously described by Tang and Hughes and in Chapter 3 (Tang and Hughes, 1998) The approximate mean diam e ter of these liposomes was 200 nm Equivalent molarx ofpDNA, pGL3 (3g), or pGM (4 .15 g) were complexed with cationic liposomes DOTAP / DOPE at a 1:2 w / w ratio of pDNA:liposomes for 30 min in serum-free medim.

PAGE 56

pGL3 5256 bp SV40 promoter Luc SV40 enhancer Arrp pGM 7217 bp SV 40 prorroter !\At\ Rs Luc SV40 enhancer Figure 4-1: Schematic representation of the plasmid DNAs vectors (A) pGL3 5265 bp (B) pGM 7217 bp Abbreviations: SY 40 Simian Virus 40; Luc Luciferase coding region; MARs, matrix attachment regions; Amp Ampicillin resistance gene.

PAGE 57

48 Before transfection, serum-containing medium was changed to serum-free medium and the transfection mixtures added. After 4 hours, the medium was changed to growth medium containing serum Transgene production was determined by amount of luciferase expression as previously described (Tang and Hughes 1998) from Day 1 to 5 posttranfection. Amount of 250 L ofluciferase lysis buffer were utilized in this experiment. This experiment was done 3 times with 4 replicates each time. Luciferase expression was reported in RLU/ well under assumption of non-toxic protein. Background From a plot of luciferease expression in RLU / well and time we used noncompartmental methods to estimate the pharmacokinetic parameters, AUEC and MET. These methods were based on the estimation of the area under a plot of drug concentration versus time, while have been used previously for data analysis of biological systems (Gibaldi and Perrier, 1982) The statistical moments in classical pharmacokinetic are defined as follows: C() AUC= fcdt, ( 4-1) 0 C() AUMC = ft C dt (4 2) 0 C() ft C dt MRT=..c.. o __ =AUMC 00 AUC' fcdt (4-3) 0 where AUC is area under the curve, which for transgene expression is equal to area under the expression curve (AUEC), and MRT is mean residence time and is equivalent to mean expression time (MET). The area under the curve of a plot of the product of

PAGE 58

49 concentration and time (Ct) versus time from zero time to infinity is often referred to as the area under the first moment curve, AUMC (Gibaldi and Perrier, 1982). Estimation of the AUC from zero time to infinity must be carried out in two steps The AUC from zero time to some time, t (t*), is calculated by trapezoidal rule To this partial area from the terminal portion to infinity, t*-;, oo, must be estimated as follows: 00 C AUC= Jcdt=-, t* A (4-4) where 'A is the slope of the terminal exponential phase of a plot of natural log versus time and C* is the last measured point of expression. The sum of the two partial areas is AUC The same approach was used to estimate total AUMC The area under the first moment curve from t* to infinity is estimated as follows in equation 4-5: OOJ t C C AUMC= t Cdt= +-2 1 .tt .tt (4-5) Calculations The trapezoidal rule was utilized to determe the AUC from experimental data without integration. AUC can be calculated as the sum of its individual trape z oids: C0 + C, C, + C 2 C 2 + C 3 AUCO->t = t, + (t2 -t,)+ (t3 -t2)+ ... (4-6) 2 2 2 AUCHoo = ~t (4-7) where C is the luciferase activity in the cell and t is the time point of measurement. C0 C1 C2 .. Ci is the expression of time points 0, 1, 2 .. etc and 'A is the terminal slope from the plot. The same approach has used for estimation AUMC:

PAGE 59

50 AUMC C0 t0 + C1 t0 C1 t1 + C2 t 2 ( ) C2 t 3 + C 3 t 2 ( ) 0-+t 2 t i + 2 t 2 ti + 2 t 3 t 2 + ... ct. t ct AUMCHaJ = -+ 2-. l l Simulations (4-8) (4-9) Simulations were performed using Scientist kinetic software package (Micromath, Salt Lake City Utah) A one-compartment model with first-order expression degradation and first-order transgene expression was used : E lK. ( K t -K. t) = e '-e (K. -K. ) (4-10) where Eis the expression in RLU, I is concentration of plasmid in the inoculation medium K a is the rate constant for uptake and expression of the trans gene, Ke is the rate constant for expression degradation, and t is time The variable, I, was kept constant and the maximum expression ,CMAX, were both kept constant b y maintaining the Ka/Ke ratio (ratio = 2), but absolute values of K a and Ke varied The parameter MET and AUEC were calculated as previously described in the calculations section. Time to maximal expression was calculated by ln(K ) T = K MAX (K. -K.). Statistical Analysis (4-11) The Student t-test with one-tail distribution and two-sample equal variance was used to compare pharmacokinetic parameters Repeated-measures ANOV A was used to detect differences with respect to plasmid and time. Tukey's HSD was used for post-hoc

PAGE 60

51 analysis when difference were detected. The data was considered to significantly different for p-value < 0 05. Results Figure 4-2 shows the simulation of trans gene expression of four plasmids in the same cells and under the same conditions The Cmax values were identical (2500 RLU / well). However tma x varied. The tmax of plasmids 1 2 3 and 4 were 0.46 1.39, 2.77, and 6.93 days, respectively. Consequently, measurement of protein expression at one time point is not adequate for comparison of transfection efficiency of these plasmids. Kinetic parameters AUEC and MET (Table4-l) were distinct. The AUEC of plasmids 1 2 3, and 4 were 3256, 9974, 19985, and 49031 (RLU.day) / well, respectively indicating various amounts of protein production. The MET of plasmids 1, 2 3, and 4 are 1, 3, 6, 14.33, respectively, indicating a difference in average time to plasmid express10n. Table 4-2 shows raw data from luciferase expression of pGL3 and pGM from Day 1 to 10 posttranfection in Chinese Hamster Ovary (CHO) cells. Figure 4-3 is a plot of the expression based on the raw data in Table 4-2 against time point from Day 1 to 5 posttranfection. Peak expression is at Day 2, with transgene expression of 29.14 .86 x 106 and 15.84 .76 x 106 RLU/well of pGM and pGL3 respectively. Kinetic parameters AUEC and MET were calculated based on equations 6 7 and 8 9 respectively described in the Material and Methods section of this chapter. These parameters and the statistical analysis are shown in Table 4-3 The AUEC of pGM and pGL3 are 125.92 .12 x 106 and 48.81 .48 x 106 (RLU.day) / well

PAGE 61

52 3000 2500 2000 I 1000 I 500 I 5 10 15 20 25 Day Post Transfeciton I-Plasmid I Plasmid 2 Plasmid 3 Plasmid4 I Figure 4-2 : Simulation oftransgene expression of plasmid DNAs 1, 2, 3, and 4 in the same cells from Day 1 to 10 posttranfection Trans gene expression is represented in relatively light unit (RLU) per well.

PAGE 62

53 T bl 4 1 Ph a e -f armaco metlc parameters o s1mu ate d 1 'dDNA p asm1 s Plasmid DNA AUEC MET Cmax tmax RLU.day / well Day RLU / well Day Plasmid 1 3256 1.00 2500 0.46 Plasmid 2 9974 3.00 2500 1.39 Plasmid 3 19985 6.00 2500 2.77 Plasmid 4 49031 14.33 2500 6 .93 AUEC is area under the expression curve and MET is mean expression time. Cmax and tmax are maximum concentration and time at maximum concentration, respectively

PAGE 63

54 Table 4-2: Transgene expression of plasmid pGL3 and pGM in Chinese hamster ovary CHO 11 fr D 1 10 fi ce s om ay to posttran ect10n. Time (day) pGL3xl06 pGMxl06 p-value RLU/wellSEM RLU/wellSEM 1 2 58.94 1.36.40 N / S 2 15.84.75 29.14.86 p < 0.01 3 11.12.94 25.37.99 p < 0.01 4 8.31.71 26.751.77 p < 0.01 5 5.14.52 23.38.92 p < 0.01 6 3.08.01 8.100.44 N / S 7 1.52.19 6.48.93 N / S 8 0 87 09 3.22.74 N / S 9 0.22.02 1.90 52 N / S 10 0.09.01 0.17.05 N / S Transgene expression is represented by the relative light unit (RLU) per well.

PAGE 64

35 ""o 30 .... 25 -20 a, :t 15 -10 5 / 0 1 55 * / --------------------.... -.. ---------------------------1 / 2 3 4 5 Days Post Transf ection -+-pGL3 111 pGM Figure 4-3 : Transgene expression of plasmid DNAs pGL3 and pGM, in Chinese Hamster Ovary CHO cells from Day 1 to 5 posttranfection N = 4 The error bars represent mean SEM (standard error of mean). represent significant difference with p < 0.05. Transgene expression is represented by the relativley light unit (RLU) per well.

PAGE 65

56 T bl 4 3 Ph a e -f 1 d GL3 d GM armaco metic parameters o p asm1 p an P' Plasmid DNA AUECxl06 MET CmaxX106 tmax RLU.day / well Day RLU / well Day pGL3 48.81.45 3 38.05 15.84.76 2 pGM 125.92.16 3 95.07 29.14.86 2 p-value 0 .03 0.01 AUEC is area under the expression curve and MET is mean expression time Cmax and tmax are maximum concentration and time at maximum concentration respectively. Data were represented by mean SEM. (standard error of mean).

PAGE 66

57 respectively while MET of pGM and pGL3 are 3.95 .07 and 3 .38 .05 days respectively. From the statistical analysis, AUEC and MET for pGM are significantly higher than for pGL3. Discussion and Conclusion Pharmacokinetics have been utilized to characterize of pDNA in vivo and in vitro (Houk et al., 2001 ; Lew et al., 1995; Mahato et al., 1995; Osaka et al., 1996; Scheule and Cheng, 1996). In this chapter we demonstrate the utilization of the pharmacokinetic parameters Area under the curve (AUC) and Mean residence time (MRT) to increase the accuracy of data analysis. Generally, gene expression is determined by a single-point measurement of synthesis protein at a peak of expression. The limit of a single-point measurement are time variation due to cellular uptake of DNA, DNA trafficking to nucleus, half-life of DNA and mRNA, phase of the cell cycle, and cell turnover rate. Moreover, at only one time point, the heterogeneous cellular population, the mixture of healthy and unhealthy populations at a particular time, results in an inaccurate reporter-protein measurement. Multiple-point measurements of the transgene expressed may be an alternative way Total amount of the expressed protein (AUEC) and average time of expression (MET) were obtained from a plot oftransgene expression and time. The AUEC, which is equal to AUC, and MET, which is equal to MRT, demonstrated total amount of protein production and average time of expression, respectively. Figure 4-2 is the simulation of four pDNAs that had maximum protein at various time points. Although they have the

PAGE 67

58 same maximum concentration (Cmax ), time at maximum concentration (tmax) is different. Comparing trans gene activity of these plasmids requires multiple-point measurements. AUEC and MET are calculated by a plot of protein expression in RLU and time Table 4-1 showed kinetic parameters obtained from the plot. Although plasmid 1 reached maximum concentration first (tma x = 0.46 day) because it expressed the lowest amount of protein (AUEC = 3256 RLU/ well) and persisted in cells only 1 day it was not the palsmid with the highest transgene activity An ideal pDNA would require high and long-term e x pression. Plasmid 4 is close to the idea with an AUEC of 49031 RLU / well and MET of 14.33 days followed by plasmid 2 (AUEC = 19985 RLU/ well MET = 6 days), and plasmid 3 (AUEC = 9974 RLU / well MET= 3 days) Thus, due to having the lowest AUEC and MET, plasmid 1 is considered to be the lowest quality. In another example we transfected two plasmids, pGL3 and pGM, into CHO cells and measured luciferease expression at various time points Plasmid pGM has the insertion sequences of matrix attachment regions (MARs) from human ~-interferon, which may prolong or enhance expression compared to the non-MARs plasmid pGL3 In such an experiment single-point measurements of trans gene expression may not be adequate to evaluate the hypothesis or to compare plasmid transgene activities since there might be a time-point variation in the expression. For example, both plasmids may ha v e similar maximums in expression, but if the maintenance of this expression differs between plasmids a single-time-point parameter would not describe this Comparing gene products provides a better vision of gene activity. Therefore the combination of multiple-point measurements of transgene expression and non-compartmental analysis is useful. From the raw data reported in Table 4-2 the plot of protein expression and tim e

PAGE 68

59 point can be obtained (Figure 4-3). Kinetic parameters in Table 4-3 are calculated based on raw data and equation 6-9 in Materials and Methods. The AUEC is 2 to 3-fold higher for pGM than for pGL3 indicating that pGM expressed more protein than pGL3. Moreover MET of pGM is significantly higher than for pGL3, indicating a longer time of expression. In conclusion, MAR-containing plasmid DNA, pGM, can enhance (from AUEC) and prolong (from MET) the transgene expression compared to non-MARs pDNA. The multiple-point measurements combined with kinetic analysis are simple to perform, do not require complex experiments, and better describe the expression of the reporter gene This novel application of pharmacokinetics can aid in screening therapeutic plasmids and vectors

PAGE 69

CHAPTERS MATRIX ATTACHMENT REGIONS-CONTAINING PLASMID DNA INCREASES GENE EXPRESSION IN VITRO Introduction As discussed in Chapter 2, matrix attachment regions (MARs) are associated with a variety of biological functions such as gene replication, regulation, and repairing of DNA (Agarwal et al., 1998; Alvarez et al., 2000; Malone et al., 2000; Whitelaw et al., 2000) MARs might also be related to long-term transgene expression in vitro (Baiker et al., 2000; Piechaczek et al., 1999). MARs improve transgene expression in eukaryotic cells, but only when those MARs are integrated into the genome MARs sequences demonstrated a neutral or even negative effect on expression in transiently transfected cells (Bode and Maass, 1988; Kalos and Fournier, 1995; Klehr et al., 1992; Poljak et al., 1994; Wang et al. 1996). However, one study showed that human interferon-P (hIFN-P) MARs-containing pDNA in stable transfection replicated episomally and maintained extrachromosomally in nucleus of Chinese hamster ovary (CHO) albeit low copy number (Piechaczek et al., 1999). This evidence raises the question whether this nucleic acid sequence would enhance and/or prolong expression in vitro using a transient expression system. Our working hypothesis is MARs-containing pDNA increases and prolongs expression in various cell types compared to non-MARs-containing pDNA. We chose CHO cells in 60

PAGE 70

61 this study because it was previously reported to increase replication and retention of MARs-containing pDNA (Piechaczek et al., 1999). Human neuroblastoma (SKnSH) cells were another example of fast-dividng cell used in this experiment. We also investigated the effect of MARs-containing pDNA in primary neuron, astroglia, and microglia, because of the dysfunction of these neuronal cells in neurological diseases (e.g., Alzheimer's, Parkinson's). Increasing the persistence of a therapeutics gene is potentially another method to treat the dysfunction of neuronal cells and aid in the therapy for patients with neurological disease. To answer the question whether or not MARs-containing pDNA increases expression of another pDNA, a cotransfection experiment was performed MARs are chromatin-remodeling elements that enhance chromatin accessibility (Fernandez et al., 2001; Forrester et al., 1994; Jenuwein et al., 1997). Nucleosome alteration also could be induced by other chromatin remodeling agents, histone deacetylase inhibitors (HDAC), for trichostatin A (TSA), trapoxin, n-butyrate (Yoshida et al., 1990). MARs and HDAC influence chromatin remodeling, we investigated the effect ofHDAC inhibitor, TSA, on the action ofMARs-containing pDNA. Moreover, polymerase chain reaction (PCR) has been performed to monitor intracellular distribution ofpDNA. Materials and Methods Plasmid DNAs Plasmid pGM and pGL3 were used in this experiment. Plasmid DNAs sequences and preparations were previously reported in Chapter 4 in Materials and Methods.

PAGE 71

62 Plasmid pEPI-1 was a generous gift from Dr. Lipps from Germany. The function elements of pEPI-1 are the pUC origin ofreplication, 2.0 kb MARs from 5'-region of the human interferon-P (hlFN-P) gene, the promoter of the bacterial ampicillin resistence gene, SY 40 promoter, SV 40 origin of replication, and the kanamycin resistance gene (Piechaczek et al., 1999). Cell Culture and Conditions All media in cell culture were obtained from Gibco BRL (Grand Island, NY) FBS is fetal bovine serum. PS is 5000 U / ml penicillin, 5 mg / ml streptomycin. Several cell types (50% confluence) were grown under the following conditions: CHO (Chinese hamster ovary): Minimum essential medium alpha medium, 10% FBS, 1% PS, at 5% CO 2 SKnSH (Human neuroblastoma): RPMI medium 1640, 10% FBS, 1 % PS, at 5% CO 2 Rat hippocampal primary neuron : Neurobasal media, 2% 50X B-27 0.25% 2 MLglutamine, 1 % PS at 10%C02 Rat astroglia and microglia: Dulbecco's modified eagle medium, 10% FBS, 1 % PS, at 10% CO2 Cationic Liposome Preparations DOGSDSO (l ', 2'-dioleoyl-sn-glycer-3'-succinyl-2-hydroxyethyl disulfide omithine conjugate), which contains a disulfide bond between the positively charged head group and lipophilic backbone, was synthesized using the method described by Tang and Hughes (1998). This previous study demonstrated that DOGSDSO, a disulfide containing lipid was efficient at transfecting neuronal, astroglia, and microglia cell cultures (Ajmani et al., 1999). Another cationic lipid, DOT AP (l ,2-dioleoyl-3trimethylammonium-propane), was used in transfection of CHO and SKnSH cells.

PAGE 72

63 DOT AP and DOPE (L-dioleoyl-phosphatidylethanolamine) were purchased from Avanti Polar Lipids (Alabaster, AL). DOTAP/DOPE and DOGSDSO / DOPE liposomes were prepared as described in Chapter 3 by Tang and Hughes (1998) The final cationic liposome concentration was 1 mg / ml. The average diameter of each of the two liposomal preparations was 200 80 nm In vitro transfection CHO and SKnSH cells were seeded at 0.5x105 cells / well in a flat-bottomed 12well plate (Costar, Cambridge MA) and grown overnight in the media appropriate for each cell type at 37C. DOTAP / DOPE liposome was resuspended and added to DNA at a 2 : 1 w / w liposome:DNA ratio for 30 minutes (Tang and Hughes 1998). The complete media was replaced with serum-free media The dose ofpGL3 was 3 g and ofpGM was 4 .15 g to achieve equal molar ratio. These doses were chosen for maximal effect in the same molar ratio (Figure 5-1 ) The cells were transfected for 4 hours, then the transfection media was replaced with complete media. Primary neuronal cultures were prepared from the hippocampus of newborn Sprague-Dawley rats as described by Ajmani et al. (1999) Astroglial and microglial cells were cultured from the cerebral cortices of newborn rats as previously described (Ajmani et al., 1999) The 3 5x10 mm culture disk was used for cell cultures DOGSDSO/DOPE liposome was complexed with DNA at an 8: 1 w / w liposome:DNA ratio in serum free media for 30 min at room temperature The liposome / DNA complexes were added to the cultures and incubated for 2 h (neuronal cultures) or 7 h (microglia and astroglia cultures), and then transfection media was replaced with complete media (Ajmani et al., 1999).

PAGE 73

a) --6000 5000 4000 3000 2000 1000 ___ ...,... 0 0 1 2 64 3 4 Plasmid DNA (g) 5 6 7 ---pGM +pGL3 Figure 5-1: Dose study of plasmid DNAs in CHO cells. The ratio of DNA and DOT AP/DOPE liposome is 1 : 2 w /w. Luciferase activity was presented in RLU / well.

PAGE 74

65 Cells were harvested every 24 hours from Day 1 to 10 posttransfection Each well was aspirated and rinsed once with 500 l of PBS (phosphate buffer saline, pH 7.4) and then 250 l or 500 l of lx lysis buffer (0.1 M postassium phosphate buffer, pH 7 8 1 % Triton X-100, 1 mM DTT, 2 mM EDTA) were added to the mammalian (CHO and SKnSH) cells or neuronal cells (primary neuron, astroglia, microglia), respectively. The amount of lysis buffer added depended on the size of the well and the growth area. After shaking for 15 minutes, luciferase activity was quantified by using a Mono light 2010 luminometer (San Diego, CA). Only in dose-study and HDAC inhibitor experiments luciferase activity was quantified by using a plate luminometer (Microtiter, Chantilly VA). One hundred l ofluciferase assay buffer (30 mM Tricine, 3 mM ATP, 15 mM MgS04 10 mM DDT, pH 7.8) and 20 of cell lysate were added to a 100 injection of 1 mM D-luciferin, pH 6.3 (Molecular Probes, Eugene, OR) The light emission over a 10 seconds reaction period was integrated Luciferase activity was expressed as relative light unit (RLU) per well assuming that the delivery vectors and expressed protein are nontoxic to cells In the cotransfection experiment, CHO and SKnSH cells were seeded at 0 5x 105 cells / well in a 24-well plate (Costar, Cambridge, MA) 24 hours before transfection The amount of pEPI-1 and pGL3 were 0 : 3, 1 :2, 2:3 3:0 and 3:3 g:g. The ratio of pDNA: DOT AP/DOPE liposome was 1 : 2 w/w. Cells were transfected with pDNA:liposome complex as previously described. Forty-eight hours after transfection luciferase activity was measure as described above. Pharmacokinetics

PAGE 75

66 AUEC (area under expression curve) and MET (mean expression time) were obtained from a plot of RLU/well and time period. AUEC indicated a total amount of protein produced by DNA and MET indicated a length of expression time These parameters can be calculated from trapezoidal rules as mention in Chapter 4 in Background section (Gibaldi and Perrier, 1982). Total AUEC, AUMC and MET can be calculated base on equation in Chapter 4 in Calculation section. Histone Deacetylase Inhibitor CHO cells were seeded at lxl05 cells / well in 24-well plate (Costar Cambridge, MA) 24 hours before transfection. Trichostatin A was dissolved in serum free a-MEM in a dilution series of 0.1-1 M. This medium was then mixed with pDNA ad DOT AP/DOPE liposome (pDNA:liposome = 1 :2 w / w) for 30 min before transfection as described above. To investigate the effect of the exposure time of the cell to HDAC inhibitor on expression, cells were incubated with TSA 1 or 2 hours before transfection The exposure time of the cells with TSA did not effect trans gene expression of pDNAs (data not shown) Forty-eight hours after transfection, luciferase activity was measured as described above. Plasmid DNA Extraction To investigate the appearance of plasmid DNAs in nucleus and cytoplasm as a period of time, cell fractionation was performed followed by polymerase chain reaction (PCR) for detection of pDNAs. CHO and SKnSH cells were seeded at 0 .5xl05 cells / well 24 hours before transfection. The same amount of DNA and liposome was used as described above. At Day 1, 2, 4 and 6 posttransfection, cells were fractionated. Cells were washed with ice-cold PBS and trypsinized using 0.2 ml of 0 25% trpysin, 1 mM

PAGE 76

67 EDTA (Gibco BRL, Grand Island, NY) per well. After adding 500 l of PBS cell suspensions were placed in a Model SFR13K Speedfuge refrigerated microcentrifuge (Savant Farmingdale, NY) and centrifuged for 5 min at 1000 xg twice. Cell pellets were resuspended by briefly vortexing in 200 l of homogenization buffer (10 mM HEPES, pH 7 .9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 0 5 mM PMSF) and incubated on ice for 15 min. After this incubation, 12. 5 l of a 10% Nonident P-40 solution was added to each tube, and the samples were vortexed vigorously for 10 seconds. This was followed by another centrifugation for 30 seconds at 10,000 xg. The supernatant was collected as the cytoplasmic fraction. The crude nuclear pellet was washed 2 additional times by repeating the procedure outlined above. The nuclear pellet was then resuspended in 100 l of homogenization buffer (Sperinde and Nugent, 1998). Contamination of lysosome acid phosphatase, in nuclear fraction was measured by using EnzCheck acid phosphatase assay kit (Molecular Probe, Eugene, OR). Less than 1 % of lysosomal contamination was found in the nuclear fraction using the procedure above. One volume lysis buffer containing 0.6% sodium dodecyl sulphate (SDS) pH 7 .5, and 0 .01 M EDTA was added to the nuclear or cytoplasmic fractions and incubated for 20 min at room temperature. After incubation 5 M NaCl was added to make a final concentration of 1 M and the sample was mixed by slowly inverting the tube 10 times. The sample was stored at 4C for at least eight hours, then centrifuged at 13, 000 xg for 30 min at the same temperature to remove the major portion of SDS and protein (Hirt, 1967). For the whole cell, after being incubated with homogenization buffer for 15 min lysis buffer was added to the cell suspension following by NaCl.

PAGE 77

68 All protein and SDS fractions were extracted by 1 volume of phonol / chloroform / isoamyl alcohol (25 : 24: 1 v / v / v) (Ameresco, Solon, Ohio) After centrifuging at 14, 000 rpm for 5 min at room temperature supernatant was removed to a new tube. Plasmid DNA was precipitated by adding 2 volumes of ethanol and NaCl to make a final volum e of 0.2 M. Glycogen (Fisher Biotech Fair Lawn NJ) was used as a carrier for DNA precipite at 50 g / ml of final product. The mixture was stored at 80 C for 20 min and centrifuged at 13, 000 xg for 30 min Ethanol was discarded. Pellet was air-dried for 10 min then resuspended in TE buffer Tris-EDTA, pH 7.4 PCR-Based Assay of Extracted pDNA To identify DNA in the nuclear and cytoplasmic fractions, we used PCR. Luciferase oligonucleotide was used as a primer for the PCR reaction The upper primer (pGL3, position 280) was 5' GGA ATT GCT AGC TAC TGT TGG TAA AGC CAC 3' and the lower primer (pGL3, position 1929) was 5' GGA AGA TCT AAA GCA AT A GCA TAC CAA AT 3' (Gibco BRL, Grand Island, NY). The reaction contained a 5 I sample 0 5 I 0 1 nmol upper / lower primer 0.5 l 10 mM dNTP (Promega Madison MI) 1 l 100 mM MgS04 5 l lOx PCR buffer 0 5 l Vent DNA polyn:ierase (Ne w England Bio labs Beverly MA), and sterile water to make a final volume of 50 I. The conditions were 94C 2 min primary denature. The cycle began with 94 C 30s denature, 52 C, 30s annealing, 72 C 1 30 min elongation for an appropriate number of cycles and final elongation at 72 C for 5 min. This reaction generated 1.6 kb of luciferase sequence

PAGE 78

69 Subsequently, aliquots of the PCR products were electrophoresed in a 1 % agarose gel (Fisher Sciencetific, Fair Lawn, NJ) mixed with ethidium bromide (0.2 g / ml) and quantified using Kodak lD Image Analysis Software (Kodak, New Haven, CT). Competition Binding Study Previous results indicated the nuclear matrix could bind to any DNA sequence. However, they had a high binding affinity to matrix attachment regions (Tsutsui et al., 1993). In this experiment, we attempted to distinguish between the binding affinities of MAR-containing pDNA, pGM, and non-MAR-containing pDNA, pGL3. CHO and SKnSH cells were seed at lxl05 cells / well. One g of DNA was mixed with 1, 2, or 3 g of competitive non-coding bacterial DNA to create Treatments 1, 2, or 3, respectively Plasmid DNA from E coli bacteria strain DH5a was utilized as non coding bacteria DNA. Plasmid DNA was isolated using a Megawizard DNA purification kit (Promega, Madison, WI). The concentration and purity of pDNA were determined spectrophotometrically. The average concentration of pDNAs was 2.0 mg/ ml and the purity ofpDNA, A260/ A280 ratio, was 1.8. DOTAP/DOPE was added to those mixture at 2 : 1 w / w lipid:DNA ratio. Two days after transfection, cells were fractionated, DNA was extracted and PCR was performed as described above. The quantity of PCR products was maintained within the linear range (increasing the concentration of the template or the number of cycles proportionally increases the signal) (Donev, 2000). This was achieved by using the standard curve The ratio between the amount of plasmid DNA in each fraction and plasmid DNA in a whole cell was calculated. All PCR experiments were carried out in triplicate. Similar results were obtained and summarized graphically

PAGE 79

70 Statistical Analysis A student t-test with one-tail distribution and two-sample equal variance was used to compare pharmacokinetic parameters and amount of pDNAs in the competition study. An one-way ANOVA was used to detect differences in treatment for cotranfection studies and repeated-measures ANOV A was used to detect differences with respect to plasmid and time of the plasmid expression and HDAC inhibitor studies Tukey's HSD was used for post-hoc analysis when differences were detected. The data were considered significantly different for p-value < 0.05 Results Figure 5-1 illustrated a dose-study of pDNAs in CHO cells The ratio of pDNA:liposome is 1 :2 w /w. The optimum dose for pGL3 and pGM is 3 and 5 g respectively The decrease in expression at 6 g of pGM is likely due to toxicity from the liposome. The amount of 4.15 g pGM was used in all experiment to achieve the same molar ratio as pGL3. MARs-containingpDNA, pGM enhanced the transgene expression in Chinese hamster ovary (CHO) cells (Figure 5-2A) but not in SKnSH cells (Figure 5-2B) or neuronal cells (Figure 5-3). From a plot of expression in RLU / well against time, pharmacokinetic AUEC and MET were obtained (Table 5-1 ). In CHO cells, AUEC ofpGM and pGL3 was 1259.19 .65 x 105and 488.05 .79 x 105 (RLU.day)/well while MET ofpGM and pGL3 was 3 95.07 and 3.38.05 days, respectively. AUEC and MET of pGM were significantly (p < 0 05) higher than pGL3, indicated pGM has a higher level of transgene expression and higher average time of expression than pGL3. In SKnSH cells, pGM (AUEC = 67 .35 .74 x

PAGE 80

71 A. 35 * 30 CD 0 25 'I'"" >< 20 +--pGL3 ai / ~-. 15 _.,_pGM / -.., :::::, / --~ ..... v ..J 10 / 0:: 5 -~-... ~"-~ ... 0 1 2 3 4 5 days post transfection B. 16 14 II') 12 0 'I'"" 10 >< Q) 8 -6 :::::, -..-pGL3 -pGM ..J 0:: 4 2 0 1 2 3 4 5 days post transfection Figure 5-2: Transgene expression of pGL3 and pGM in (A) Chinese hamster cells (CHO) and (B) Human neuroblastoma cells (SKnSH) in relative light unit (RLU) per welL represents significant difference of expression between plasmid DNAs at respective time with p-value < 0.05. Number of replicate n = 4 The experiment was performed three times.

PAGE 81

A. 4 "' ~3 1i 2 ~1 c2 0 2 B. 8 "' 0 6 .... 1i 4 2 c2 0 2 72 4 6 8 days post tramfection ."'--3 4 5 6 days post transfection 10 8 10 +-pGL3 I oGM pGL3 I oGM Figure 5-3: Transgene expression of pGL3 and pGM in (A) Hippocampal primary neuron, (B) Astroglia, and (C) Microglia in relative light unit (RLU) per well. represents significant difference of expression between plasmid DNAs at respective time with p-value < 0.05. Number ofreplicate, n = 3. The experiment was performed three times.

PAGE 82

C. 10 ... 0 8 >< 6 4 --;:;i ; 2 .,_. 0 2 Figure 5-3--Continued 73 3 4 5 6 days post transf ection 8 10 + pGL3 I oGM

PAGE 83

T bl 5 1 Ph k" f t f GM d GL3 CHO SKnSH d 11 a e -armaco me 1c parame ers o p an P' m an neurona ce s Cell lines Plasmid Area Under Expression Mean Expression Time Cmax *10:, tmax Curve (AUEC) 105 SEM (MET)SEM RLU.day/well day RLU / well day CHO (Chinese pGL3 488 05. 79 3.38.05 158.4 2 Hamster Ovary) pG/M 1259 19.65 3.95.07 291.4 2 p-value 0.0321 0 0104 SKnSH (Human pGL3 48 09.60 4 10.66 7 09 2 neuroblastoma) pG/M 67.35 74 5.13.68 9 67 2 p-value 0.2840 0.3026 Cmax represents maximum concentration or expression ; tmax represents time of maximum expression.

PAGE 84

Table 5-1--Continued Cell lines Plasmid Area Under Expression Mean Expression Time Cma x *105 tmax Curve (AUEC) 105 SEM (MET)SEM (RLU day)/well dav RLU / well day Primary Neuron pGL3 19. 62.50 5.04.79 2.98 4 (0-10 day) pG/M 21.36.69 5.40.96 3.25 4 p-value 0.3905 0.3910 Astroglia(0-1 Oday) pGL3 14.88.05 5 20.65 4 .95 5 pG/M 24.80.14 5.53.27 5.88 3 p-value 0.2912 0.4448 Microglia(0-10 pGL3 1.77.43 7.08.66 3 92 10 day) pG/M 2.80.30 6.94.71 6 82 8 p-value 0.0583 0.4476 Cmax represents maximum concentration or expression; tmax represents time of maximum expression.

PAGE 85

76 105 (RLU.day) / well) expressed slightly more protein than pGL3 (AUEC = 48.09 .60 x 105 (RLU.day)/well), but this was not significantly different. Mean expression times of pGM (5.13 .68 days) and pGL3 (4.10 .66 days) were similar. In primary neurons, AUEC ofpGM and pGL3 was 21.36 69 x 105 and 19.62 .50 x 105 (RLU.day) / well, respectively, while MET of pGM and pGL3 was 5.40 .96 and 5 04 .79 days, respectively Plasmid pGM and pGL3 demonstrated similar protein expression and time of expression (Table 5-1 ) In astroglia, AUEC ofpGM and pGL3 was 24.80 .14 x 105 and 14.88 .05 x 105 (RLU.day) / well, respectively and MET of pGM and pGL3 was 5.53 .27 and 5.20 .65 days, respectively. In microglia, AUEC of pGM and pGL3 was 2.80 .30 x 105 and 1.77 .43 x 105 (RLU.day) / well, respectively and MET ofpGM and pGL3 was 6 94 .71 and 7.08 .66 days, respectively In astroglia and microglia, the average time of expression was identical for pGM or pGL3. According to MET pGM did not prolong expression in the brain-derivatives cells. Time of the maximum expression (tmax ) in fast dividing cells (CHO, SKnSH) was earlier than in non-dividing (primary neuron) or slow dividing cells (astroglia and microglia). Figure 5-4 shows transgene expression of non-MAR-containing pDNA, pGL3, cotransfected with MARs-containing pDNA, pEPI-1. MARs has trans effect on other pDNAs. We found MARs-containing pDNA increase transgene expression of non MAR-containing pDNA in CHO cells, but not in SKnSH cells. In SKnSH cells, luciferase activity depended on the amount of pGL3. Three g of pGL3 showed higher level (p < 0 05) ofluciferase activity than 2 g or 1 g of pGL3. The expression was independent of the amount of pEPI-1. In CHO cells, trans gene expression increased (p <

PAGE 86

77 0 05) when MARs-containing pDNA, pEPI-1, was added from the ratio of 0:3 g pEPI-1 : pGL3 to 1 :2 g pEPI-1 : pGL3 In the last treatment that contained 3 g of each pGL3 and pEPI-1, the expression was significantly higher than baseline (3 g of pGL3 and O g of pEPI-1) in CHO cells. The expression was significantly decreased however in the last experiment, in SKnSH cells Figure 5-5 shows the effect of trichostatin A (TSA) on trans gene expression of MAR and non-MAR containing pDNA in CHO cells. At low concentration of trichostatin A (0 005-0.045 M), expression of pGM, but not pGL3, was significantly increased (Fig. 6) compared to baseline (no TSA added). Increasing the concentration of TSA resulted in a decrease of trans gene expression, not different from no TSA added (data not shown). Plasmid DNA in CHO cells was located in the cytoplasm at Day 1 posttransfection (Figure 5-6). Plasmid DNA was observed in the nucleus between Day 2 to 4. Plasmid DNA began to degrade at Day 6. In the competition study, Table 5-2 shows the ratio of pDNAs in the nucleus or cytoplasm of whole cells The amounts of non-coding pDNA were from 1, 2, and 3 gin Treatments 1, 2, and 3 respectively. Ratio of pDNA:liposome is 1 :2 w/w. The more non-coding DNA added, the less DNA was found in both nucleus and cytoplasm Discussion and Conclusions In our study, MARs increased and prolonged transgene expression in CHO but not SKnSH or neuronal cells. It was apparently that cell lines are plays an important role in the extent of MARs function.

PAGE 87

78 A. so 45 40 "' 35 Q .... 30 25 -=i 20 15 10 5 0 0:3 1:2 2:1 3:0 3:3 pEPI-1:pGL3 (g) B. 50 45 40 35 0 30 25 -;;i 20 c2 15 10 5 0 0:3 1 :2 2:1 3:0 3:3 pEPI-1:pGL3 (g) Figure 5-4: Luciferase activity of pGL3 when contransfected with MAR-assocaited plasmid DNA pEPI-1 in (A) Chinese hamster ovary (CHO) cells and (B) human nueroblastoma (SKnSH) cells represents significant difference of expression compared to 0 : 3 g ofpEPI-1:pGL3 with p-value < 0 05.

PAGE 88

79 This may be due to the binding ofMARs to the nuclear matrix because it is a prerequisite for MARs function (Luderus et al., 1992; Luderus et al., 1994; Stief et al., 1989; Tang and Hughes, 1998) and it is a cell type specific (Wilson et al., 1999). Baiker and coworkers (2000) found that hIFN-P MARs associated with host chromosome in CHO cells in non-covalent fashion. This association did not change significantly between 20 and 50 generations after transfection in CHO cells (Baiker et al., 2000) The hIFN-P MAR may have a strong interaction to a nuclear matrix of CHO cells, but has a moderate and weak interaction with the nuclear matrix of SKnSH cells and neuronal cells respectively. The different cell lines may be linked to the differences in enzyme machinery, MARs-binding protein or MARs-binding transcription factors. The studies found that some classical matrix proteins are cell-type specific and bind strongly and selectively to MARs from different species (Boulikas, 1993; Boulikas, 1995) The dynamic of cells may be another explanation. Non-dividing cells, such as primary neuron or slow-dividing cells such as astroglia and microglia, may lack of some components that are necessary for MARs-nuclear matrix binding, whereas there are plenty of these components in CHO cells. Moreover, it is more difficult for pDNA to enter non-dividing cells than dividing cells. Thus, less expression was observed. We hypothesized that pDNAs in non-dividing or slow-dividing cells were not diluted out as early as the fast-dividing cells, thus resulting in longer expression time (MET). However the expression was poor and the peak of expression shifted from Day 2 in dividing cells to Day 4, or 10 in non-dividing or slow-dividing cells

PAGE 89

30000 25000 *# 20000 1i -15000 ;:l 2 10000 5000 0 0 0.01 80 0 .02 0 .03 Trichostatin A (M) *# 0 .04 0.05 -pGM + pGL3 Figure 5-5: Trans gene expression of pGL3 and pGM in CHO cells when incubated with histone deacetylase inhibitor, trichostatin A {TSA) in low doses (TSA 0-0.045 M). represents significant difference of expression between plasmid DNAs at respective concentration. # represents significant difference of expression compared to baseline (0 M TSA added) of the repective pDNA.

PAGE 90

Dayl Std pGL3 pGM Nucleus Cytoplasm Day2 pGL3 pGM Day4 pGL3 pGM Day6 pGL3 pGM Figure 5-6 : Plasmid DNA, pGM, and pGL3 extracted from nucleus and cytoplasm of Chinese hamster ovary (CHO) cells after 1, 2, 4, and 6 days of transfection. Polymerase chain reaction (PCR) was used to amplify luciferase sequences, which are the backbone of pGM and pGL3. Std is lambda DNA/Hind III marker. 00 ......

PAGE 91

Table 5-2: Plasmid DNA extracted from nucleus and cyto:Jlasmof Chinese hamster ovary (CHO) cells .. % DNA in whole cell Treatment DNA Non-coding Liposome (g) DNA (g) (g) Nucleus Cytoplasm pGL3 pGM pGL3 pGM 1 1 1 4 13.54 .75 7.33.21 59.04 .65 71.44.24 2 1 2 6 6.54.71 12.83.61 80.65.87 62.21.29 3 1 3 8 4 10.77 5.93.91 50.08 .17 55.69.64 Plasmid DNA was extracted after 2 days oftrasnfection with one g of pDNA and 1 2, and 3 g of non-coding pDNA from bacteria. Ratio between pDNA : liposome is 1 :2 w/w. Polymerase chain reaction (PCR) was used to amplify luciferase sequences which are the backbone ofpGM and pGL3. The amount of pDNA was normalized by the amount ofpDNA in whole cells 00 N

PAGE 92

83 This might be due to slow DNA trafficking in non-dividing or slow-dividing nuclei We observed multiple peak of expression in astroglia. This may be due to the heterogeneous cell population. Another possible reason for differences might be the size of pDNAs. Large pDNA might interfere with the binding of MAR sequences and nuclear matrix, chromatin remodeling, and domain forming. These procedures are reversible and important to MARs functions and might effect expression ofMARs-containingpDNA in cells. Promoter effect is another explanation. SV 40 promoter might not be strong enough to initiate the expression in neuronal cells. It not may either bind tightly with RNA polymerase or not work with MARs to enhance expression. Developing plasmid DNA, which has a wide variety of cell types, is a future prospect. The smaller-sized plasmid DNA with stronger promoter might be an alternative. This could be done by cutting pDNA at bacterial part, but not at MARs part because decreasing in MARs size from the original results in decreasing in transgene expression (Bode et al., 1992). We also observed trans-activity ofMARs-containingpDNAs with another pDNA in CHO cells (Fig. 5A), but not in SKnSH cells (Fig. 58). The data confirm a cell type specificity of the MARs vector. In SKnSH cells, we found that luciferase activity was dependent on the amount ofpGL3, but not the amount ofpEPI-1. However, this effect was not observed in CHO cells; that is transgene expression of pEPI-1 :pGL3 ratio of 0:3 g and 2: 1 g were similar. Moreover, when pEPI-1 was added to the pEPI-1 :pGL3 ratio of 1 : 2 g, the transgene expression was significantly higher than pEPI-1 :pGL3 ratio of 0:3 g. These data disagree with previous studies that indicated effects ,of MARs only when they located within reporter pDNA (Bode et al., 1992). In SKnSH cells, pEPI-1

PAGE 93

84 seemed to decrease expression of pGL3 in the last treatment (3 g of pEPI-1 and pGL3) compared to baseline (0 g ofpEPI-1 and 3 g pGL3). Because the amount of pDNAs in the last treatment is doubled that in other treatments (6 g instead of 3 g), more liposome was added to keep the pDNA:liposome ratio (1 :2 w / w) constant. Decreasing in expression might be due to the toxicity of the liposome to the SKnSH cells. The effect of HDAC inhibitor, trichostatin A (TSA), on expression of MARs containing pDNA was demonstrated. Since MARs enhance expression in CHO cells we chose this cell type to study the effect of TSA, on trans gene expression. MARs containing pDNA acts synergistically with low concentration TSA (0 005-0.045 M). In high concentration ofTSA (0.1-1.2 M) the expression ofMARs and non-MARs containing pDNA decreases to the baseline level (no TSA added). Tricho~tatin A is fungistatic antibiotic and a potent histone deacetylase (HDAC) inhibitor that leads to the reversible hyperacetylation of histone at nanomolar concentrations (Yoshida et al., 1990). This hyperacetylation results in an active gene in vitro and in vivo (Chen et al., 1997; Dion et al., 1997; Van Lint et al., 1996). Therefore, increasing transgene expression of MARs-containing pDNA may be due to hyperacetylation of gene by TSA together with chromatin alteration by MAR. The data showed a dose-dependent effect of TSA on transgene expression. This data agrees with a previous study by Yoshida (1990) that indicated the effect of TSA on histone deacetylase inhibition at nanomolar concentrations. Plasmid DNAs were extracted at various times after transfection and PCR was performed. Plasmid DNAs were detected mostly in cytoplasm at Day 1 posttransfection. Plasmid DNAs were detected mostly in nucleus at Day 2 to 4 and began to degrade at

PAGE 94

85 Day 6. These data agree with transgene expression in CHO cells. Peak of expression is at Day 2 and mean expression time is around Day 4. DNA was then diluted out and degraded, leading to the decline in expression MARs-containing pDNA (pGM) and non MAR--containing pDNA (pGL3) have similar transfection efficiency, which is the ability to enter and degrade in cells. However, in the expression of pGM was higher than pGL3 in CHO cells. These data agree with previous studies showed that MARcontaining pDNA enhanced expression by increasing transcription level which means the amount of mRNA, but did not depend on amount of DNA (Forrester et al., 1994) This enhanceme.o.t might be due to their sequences. The competition study in Table 2 shows no difference between the ratio of pDNAs in nucleus and cytoplasm in whole cell in any treatment. However, non-coding pDNA effected the ability of pDNA to enter the nucleus and affected the complexing of DNA with liposome This data confirmed that pGM and pGL3 has the similar transfection efficiency Adding sequences such as nuclear localization sequences (NLS) which facilitate the entrance of pDNA into the nucleus in MARs-containing pDNA, is suggested to improve the expression. MARs containing pDNA can enhance and prolonge the expression in Chinese hamster ovary cells but not in human neuroblastoma cells or neuronal cells. This MAR vector showed trans-activity on another reporter pDNAs specifically in CHO cells It has a synergistic effect with histone deacetylase inhibitor, trichostatin A also in CHO cells. This evidence indicates the cell type specific character of this MARs-containing pDNA This may be due to the high affinity of MARs for the nuclear matrix in the certain cell types. Despite the differences among cell lines, MARs-containing pDNA showed some advantages over non-MAR-containing pDNA because this vector combines the

PAGE 95

86 advantages of non-viral vector with the retention activity ofMARs. Although MAR vector has cells type limitations, this vector shows a potential for improvement and application to nonviral gene transfer.

PAGE 96

CHAPTER6 CONCLUSION AND FUTURE PROSPECTS Conclusion The goal of this research was to develop plasmid DNA for nonviral gene transfer. Conventional plasmid DNA suffers from various limitations, such as low and transient expression. This can lead to retreatment that is not desirable when used clinically. Plasmid DNA that remains in cells would result in enhanced and prolongeq expression. MARs have been studied for application in nonviral gene transfer for the last decade. Previous studies indicated MARs replicated extrachromosomally and well maintained for more than 100 generation in cells. Therefore, our hypothesis was that MARs-containing pDNA has higher protein expression compared to non-MARs-containing plasmid DNA. To test this hypothesis, we developed a data analysis method using multiple-point measurements of pharmacokinetic parameters, area under the curve (AUC) and mean residence time (MRT) Multiple-point measurements in gene expression have an advantage over single point measurements. Gene expression is a dynamic process that is takes time from DNA transcription to translation. For example, time for DNA to be initialized by cells and to move through the cytoplasm to the nucleus, stability of DNA and mRNA, phase of cell cycle, and cell turnover rate. Moreover, in heterogeneous populations, the mixing of healthy and unhealthy cells at certain times result in various expressions Thus, multiple-point measurements provide insight into the whole picture, 87

PAGE 97

88 not just one point of expression. Utilization of pharmacokinetic parameters allows the calculation of total protein production from a plot of trans gene expression and time Under the assumption of linear pharmacokinetics with first order elimination, AUEC and MET were obtained. AUEC indicates the total amount of protein produced and MET indicates the average time of the expression. By using a combination of multiple-point measurements and pharmacokinetic parameters, we were able to compare transgene expression of MARs and non-MARs containing pDNA in various cell types MARs-containing pDNA significantly enhanced and prolonged transgene expression in Chinese hamster cells (CHO), but not in human neuroblastoma cells (SKnSH). In cell types where transfection is difficult, such as primary neuron, astroglia, and microglia, MARs-containing pDNA showed similar transgene expression and time of expression compared to non MARs-containing plasmid DNA. MARs-containing plasmid DNA, like other pDNA, functions well in rapidly dividing cells. Once in the nucleus, MARs sequences trigger biological change leading to superior gene expression. The variation of extent of expression may be due to the different binding affinities of MARs and nuclear matrix in different cell types. Physical properties of pDNA, such as size, promoter, reporter, or position of the inserted MARs fragment in pDNA may be another explaination for the difference extent oftransgene expression in different cell lines We found that MARs-containing pDNA has the cell-type specific properties MARs-containing pDNA increased transgene expression of another reporter pDNA only in CHO cells. This pDNA acts synergistically with histone deacetylase inhibitor, trichostatin A (TSA). This effect was observed in low doses TSA and in CHO cells. As

PAGE 98

89 previously discussed, different binding affinities of MARs and nuclear matrix in different cell types may be a main reason. However, MARs-containing pDNA exhibited a similar pattern of appearance in the nucleus as non-MARs---containing pDNA. Moreover, noncoding bacterial pDNA also effected the appearance of both MARs and non-MARs pDNA. We also investigated the link between pDNA topoisoform (supercoiled, open circular and linear) and transfection efficiency. Plasmid DNA is a basic tool for nonviral gene transfer and supercoiled is the major form of pDNA when it is isolated from bacteria. When pDNA undergoes environmental change, such as physical or chemical changes, it results in degradation of pDNA, which can be open circular, linear, or fragmented. Our hypothesis was that different pDNA topoisoforms affect transfection efficiency differently. To test this hypothesis, supercoiled, open circular, and linear forms were transfected into cells by using either liposomes or electroporation We observed that even though the open circular form had larger higher half-life in cytoplasm and a higher amount present in cell, its transfection efficiency was not significantly different from the supercoiled or linear forms. We concluded that different pDNA topoisomers do not affect transfection efficiency. Future Aims MARs-containing pDNA characteristics describes in this dissertation suggest that these nonviral vectors have potential for improvement. However, MARs biological functions are still unclear. Constructing MARs-containing pDNA that works in a wide

PAGE 99

90 variety of cell types is required. This could be done by replacing the SY 40 promoter with one that is more resilent and not shut down by cellular mechanism; reducing the size of pDNA but not the MARs sequences, since that would reduce their function ; and adding sequences that facilitate the entrance of pDNA into nucleus, such as nuclear locali z ation signal (NLS). This combination ofMARs sequences and NLS might give superior results.

PAGE 100

REFERENCES Adami, R. C., Collard, W. T., Gupta, S. A., Kwok, K. Y., Bonadio, J. and Rice, K. G Stability of peptide-condensed plasmid DNA formulations. J Pharm Sci 87, 678-83. (1998) Agarwal, M., Austin, T. W., Morel, F., Chen, J., Bohnlein, E. and Plavec, I. Scaffold attachment region-mediated enhancement ofretroviral vector expression in primary T cells. J Viral 72, 3720-8. (1998). Ajmani, P. S., Tang, F., Krishnaswami, S., Meyer, E. M., Sumners, C. and Hughes, J. A. Enhanced transgene expression in rat brain cell cultures with a disulfide-containing cationic lipid Neurosci Lett 277, 141-4. (1999). Alvarez, J. D., Yasui, D. H., Niida, H., Joh, T., Loh, D. Y. and Kohwi-Shigematsu, T. The MAR-binding protein SATB 1 orchestrates temporal and spatial expression of multiple genes during T-cell development. Genes Dev 14, 521-35. (2000). Auten, J., Agarwal, M., Chen, J., Sutton, R. and Plavec, I. Effect of scaffold attachment region on transgene expression in retrovirus vector-transduced primary T cells and macrophages. Hum Gene Ther 10, 1389-99. (1999). Baiker, A., Maercker, C., Piechaczek, C., Schmidt, S. B., Bode, J., Benham, C. and Lipps, H.J. Mitotic stability of an episomal vector containing a human scaffold/matrix attached region is provided by association with nuclear matrix. Nat Cell Biol 2, 182-4. (2000). Bates, A. and Maxwell, T in DNA Topology (eds. Bates, A. & Maxwell, T.) 31 (Oxford University Press, Oxford, 1993). Benham, C., Kohwi-Shigematsu, T. and Bode, J. Stress-induced duplex DNA destabilization in scaffold/matrix attachment regions. J Mal Biol 274, 181-96. (1997). Berezney, R. and Coffey, D. S. Nuclear protein matrix: association with newly synthesized DNA. Science 189, 291-3. (1975). Blackwood; E. M. and Kadonaga, J. T Going the distance: a current view of enhancer action. Science 281, 61-3. (1998). 91

PAGE 101

92 Bode, J., Kohwi, Y Dickinson, L., Joh, T., Klehr D. Mielke C. and Kohwi Shigematsu, T. Biological significance of unwinding capability of nuclear matrix associating DNAs. Science 255, 195-7. (1992) Bode, J. and Maass, K Chromatin domain surrounding the human interferon-beta gene as defined by scaffold-attached regions. Biochemistry 27, 4 706-11 ( 1988) Bode, J., Schlake, T., Rios-Ramirez, M., Mielke, C., Stengert, M. Kay, V. and Klehr Wirth, D Scaffold/matrix-attached regions : structural properties creating transcriptionally active loci International Review of C y tology 162A, 389-454 (1995). Boulikas, T Nature of DNA sequences at the attachment regions of genes to the nuclear matrix. J Cell Biochem 52, 14-22. (1993) Boulikas, T. Chromatin domains and prediction of MAR sequences. Int Rev C y tol 279388 (1995). Chen, W. Y Bailey, E. C., McCune, S. L., Dong, J. Y. and Townes, T M Reactivation of silenced, virally transduced genes by inhibitors ofhistone deacetylase. Proc Natl A cad Sci US A 94, 5798-803 (1997). Chemomordik, L. V., Sokolov, A. V. and Budker, V G. Electrostimulated uptake of DNA by liposomes. Biochim Biophys Acta 1024, 179-83 (1990) Cook, P R. The organization ofreplication and transcription Science 284, 1790-5. (1999). Cossons, N., Nielsen, T. 0., Dini, C Tomilin, N Young, D B., Riabowol, K. T., Rattner, J B., Johnston, R. N., Zannis-Hadjopoulos, M. and Price, G. B. Circular Y AC vectors containing a small mammalian origin sequence can associate with the nuclear matrix. J Cell Biochem 67, 439-50. (1997). Courey A. J. and Wang, J.C. Cruciform formation in a negatively supercoiled DNA may be kinetically forbidden under physiological conditions Cell 33, 817-29 ( 1983 ) Dang, Q Auten, J. and Plavec, I. Human beta interferon scaffold attachment region inhibits de novo methylation and confers long-term, copy number-dependent expression to a retroviral vector. J Virol 74, 2671-8. (2000) Dickinson, L.A., Joh, T., Kohwi, Y. and Kohwi-Shigematsu, T. A tissue-specific MAR/SAR DNA-binding protein with unusual binding site recognition C e ll 70, 631-645 (1992) Dion, L. D., Goldsmith, K. T., Tang, D. C., Engler, J A., Yoshida, M. and Garver, R. I., Jr. Amplification of recombinant adenoviral transgene products occurs by inhibition of histone deacetylase. Virology 231, 201-9. (1997).

PAGE 102

93 Dipiro, J. T Blouin, R. A., Pruemer, J. M. and Spruill, W. J. in Concepts in clinical pharmacokinetics: A self-instruction course (eds Dipiro, J. T., Blouin, R. A., Pruemer J. M & Spruill, W. J.) 3 (American Society of Health-System Pharmacists, Bethesda, 1997) Djeliova, V., Russev, G. and Anachkova, B. Dynamics of association of origins of DNA replication with the nuclear matrix during the cell cycle. Nucleic Acids Res 29, 3181-7 (2001). Donev, R. M The type of DNA attachment sites recovered from nuclear matrix depends on isolation procedure used. Mal Cell Biochem 214, 103-10. (2000) Felsenfeld, G., Boyes, J., Chung, J., Clark, D. and Studitsky, V. Chromatin structure and gene expression. Proc Natl Acad Sci US A 93, 9384-8 (1996). Fernandez, L.A., Winkler, M. and Grosschedl, R. Matrix attachment region-dependent function of the immunoglobulin mu enhancer involves histone acetylation at a distance without changes in enhancer occupancy Mal Cell Biol 21, 196-208. (2001). Forrester W C., Fernandez, L.A. and Grosschedl R. Nuclear matrix attachment regions antagonize methylation-dependent repression of long-range enhancer-promoter interactions. Genes Dev 13, 3003-14 (1999). Forrester, W. C., van Genderen, C., Jenuwein, T. and Grosschedl, R. Dependence of enhancer-mediated transcription of the immunoglobulin mu gene on nuclear matrix attachment regions. Science 265, 1221-5 (1994) Gahn, T. A. and Schildkraut, C. L. The Epstein-Barr virus origin of plasmid replication oriP, contains both the initiation and termination sites of DNA replication Cell 58, 52735. (1989). Gao, X. and Huang, L. Cationic liposome-mediated gene transfer. Gene Ther 2 710-22 (1995). Gasser, S M and Laemmli, U. K. Cohabitation of scaffold binding regions with upstream / enhancer elements of three developmentally regulated genes of D. melanogaster. Cell 46, 521-30 (1986). Gellert, M DNA topoisomerases. Annu Rev Biochem 50, 879-910 (1981) Gerdes, M G., Carter, K. C., Moen, P T., Jr. and Lawrence, J B. Dynamic changes in the higher-level chromatin organization of specific sequences revealed by in situ hybridization to nuclear halos. J Cell Biol 126, 289-304 (1994).

PAGE 103

94 Gibaldi, M and Perrier, D. in Pharmacokinetics (eds. Gibaldi, M. & Perrier, D.) 409-411 (Marcel Dekker, Inc, New York, 1982). Grosveld, F. Activation by locus control regions? Curr Opin Genet Dev 9, 152-7. ( 1999) Gutierrez-Adan, A. and Pintado, B. Effect of flanking matrix attachment regions on the expression of microinjected transgenes during preimplantation development of mouse embryos. Transgenic Res 9, 81-9. (2000). Haran, T E. and Crothers, D. M. Cooperativity in A-tract structure and bending properties of composite TnAn blocks Biochemistry 28, 2763-7. (1989). Harrison, S., Fisenne, K. and Hearing, J. Sequence requirements of the Epstein-Barr virus latent origin of DNA replication. J Viral 68, 1913-25 (1994). He D. C., Nickerson, J. A. and Penman, S. Core filaments of the nuclear matrix. J Cell Biol 110, 569-80. ( 1990). Hengge, U. R., Chan, E. F., Foster, R. A., Walker, P. S. and Vogel, J.C. Cytokine gene expression in epidermis with biological effects following injection of naked DNA. Nat Genet 10, 161-6. (1995). Hengge, U. R., Dexling, B., Udvardi, A., Vole-Platzer, B and Mirmohammdsadegh, A. in The skin and gene therapy (eds. Hengge, U R. & Vole-Platzer, B.) 67 (Springer Verlag, Berlin, 2001). Hirose, S and Ohba, R. in Molecular Biology of DNA Topoisomerases and its Application to chemotherapy (ed. Andoh, T., Ikeda, H., Oguro, M.) 87-94 (, 1993). Hirose, S., Tsuda, M. and Suzuki, Y. Enhanced transcription of fibroin gene in vitro on covalently closed circular templates. J Biol Chem 260, 10557-62. (1985). Hirt, B. Selective extraction of polyoma DNA from infected mouse cell cultures. J Mo! Biol 26, 365-9 (1967) Houk, B. E : Hochhaus, G. and Hughes, J. A. Kinetic modeling of plasmid DNA degradation in rat plasma. J Pharm Sci 1 (1999) Houk, B. E., Martin, R., Hochhaus, G. and Hughes, J. A. Pharmacokinetics of plasmid DNA in the rat. Pharm Res 18, 67-74. (2001). Izaurralde, E., Kas, E and Laemmli, U. K. Highly preferential nucleation of histone H 1 assembly o scaffold-associated regions. J Mal Biol 210, 573-85. (1989). Jackson, D. A. and Cook, P.R. Replication occurs at a nucleoskeleton. Embo J 5, 140310. (1986).

PAGE 104

95 Jankelevich, S., Kolman, J. L., Bodnar, J. W. and Miller, G. A nuclear matrix attachment region organizes the Epstein-Barr viral plasmid in Raji cells into a single DNA domain. EmboJll, 1165-76. (1992). Jenuwein, T. Forrester, W. C., Fernandez Herrero, L. A., Laible, G., Dull M. and Grosschedl, R. Extension of chromatin accessibility by nuclear matrix attachment regions Nature 385, 26972. ( 1997). Kalos, M and Fournier, R. E Position-independent transgene expression mediated by boundary elements from the apolipoprotein B chromatin domain. Mol Cell Biol 15, 198207. (1995). Kay, V. and Bode, J. Binding specificity of a nuclear scaffold: supercoiled, single stranded, and scaffold-attached-region DNA. Biochemistry 33, 367-74. (1994) Kirillov, A., Kistler, B., Mostoslavsky, R., Cedar, H., Wirth, T. and Bergman, Y A role for nuclear NF-kappaB in B-cell-specific demethylation of the Igkappa locus Nat Genet 13, 435-41. (1996). Klehr D Schlake, T., Maass, K. and Bode, J Scaffold-attached regions (SAR elements) mediate transcriptional effects due to butyrate. Biochemist ry 31, 3222-9. (1992) Koo, H. S., Wu, H. M and Crothers, D M DNA bending at adenine thymine tracts. Nature 320, 501-6. (1986). Kreiss, P., Cameron, B Rangara, R., Mailhe, P., Aguerre-Charriol, 0., Airiau, M Scherman, D., Crouzet, J. and Pitard, B. Plasmid DNA size does not affect the physicochemical properties of lipoplexes but modulates gene transfer efficiency Nucleic Acids Res 27, 3792-8 (1999). Lasic, D. D. in Liposome in gene delivery (ed. Lasic, D. D.) 45 (CRC press LLC, New York, 1997). Lasic, D. D. in Liposome in gene delivery (ed. Lasic, D D.) 145 (CRC press LLC, New York, 1997) Lechardeur, D., Sohn, K. J., Haardt, M Joshi, P. B., Monck, M., Graham, R W., Beatty, B., Squire, J., O'Brodovich, H. and Lukacs, G. L. Metabolic instability of plasmid DNA in the cytosol: a potential barrier to gene transfer. Gene Ther 6, 482-97 ( 1999) Lew, D Parker, S. E., Latimer, T., Abai A. M., Kuwahara-Rundell, A., Doh, S. G., Yang, Z. Y Laface, D., Gromkowski, S H Nabel, G J. and et al. Cancer gene therapy using plasmid DNA: pharmacokinetic study of DNA following injection in mice. Hum Gene Ther 6, 553-64. (1995).

PAGE 105

96 Lichtenstein, M., Keini, G., Cedar, H. and Bergman, Y. B cell-specific demethylation : a novel role for the intronic kappa chain enhancer sequence. Cell 76, 913-23 (1994) Liu, W. M., Guerra-Vladusic F. K., Kurakata, S., Lupu, R. and Kohwi-Shigematsu T. HMG-I(Y) recognizes base-unpairing regions of matrix attachment sequences and its increased expression is directly linked to metastatic breast cancer phenotype. Cancer Res 59, 5695-703. (1999). Lodish, H., Baltimore, D., Berk, A., Xipursky, S L., Matsudaira, P. and Darnell, J. in Molecular Cell Biology (ed J., D.) 381 (Scientific American Books, Inc., New York, 1995). Lodish, H., Baltimore, D., Berk, A., Zipursky, S. L., Matsudaira, P. and Darnell, J. in Molecular Cell Biology ( eds Lodish, H et al.) 350 (Scientific American Books, Inc., New York, 1998). Luderus, M. E., de Graaf, A., Mattia, E., den Blaauwen, J. L., Grande, M A., de Jong, L. and van Driel, R. Binding of matrix attachment regions to lamin B 1. Cell 70, 949-59. (1992). Luderus, M. E., den Blaauwen, J. L., de Smit, 0. J., Compton, D. A. and van Driel, R Binding of matrix attachment regions to lamin polymers involves single-stranded regions and the minor groove. Mol Cell Biol 14, 6297-305. (1994). Ludtke, J. J., Zhang, G., Sebestyen, M. G. and Wolff, J. A. A nuclear localization signal can enhance both the nuclear transport and expression of 1 kb DNA. J Cell Sci 112, 2033-41. (1999) Mahato, R. I., Kawabata, K., Nomura, T., Takakura, Y. and Rashida, M Physicochemical and pharrnacokinetic characteristics of plasmid DNNcationic liposome complexes. J Pharm Sci 84, 1267-71. (1995). Makarova, 0., Gorneva, G., Wu, F., Farutin, V., Villeponteau, B., Poliani, L., Fink D and Levine; M. Incorporation of nuclear matrix attachment regions into the herpes simplex virus type 1 genome does not induce long-term expression of a foreign gene during latency. Gene Ther 3, 829-33. (1996). Malone, C. S., Patrone, L., Buchanan, K. L., Webb, C. F. and Wall, R. An upstream Oct-1-and Oct-2-binding silencer governs B29 (lg beta) gene expression. J Immunol 164, 2550-6. (2000). Mielke, C., Kohwi, Y., Kohwi-Shigematsu, T. and Bode, J Hierarchical binding of DNA fragments derived from scaffold-attached regions: correlation of properties in vitro and function in vivo. Biochemistry 29, 7475-85 (1990)

PAGE 106

97 Mirkovitch, J., Mirault, M. E and Laemmli, U. K. Organization of the higher-order chromatin loop: specific DNA attachment sites on nuclear scaffold. Cell 39 223-32. (1984). Mizutani, M Ura, K. and Hirose, S. DNA superhelicity affects the formation of transcription preinitiation complex on eukaryotic genes differently. Nucleic Acids Res 19, 2907-11. (1991) Muenchen, H.J. and Pienta, K J. The role of the nuclear matrix in cancer chemotherapy Crit Rev Eukaryot Gene Expr 9, 337-43 (1999). Murray, L., Travis, M., Luens-Abitorabi, K., Olsson, K., Plavec, I., Forestell S., Hanania, E G and Hill, B. Addition of the human interferon beta scaffold attachment region to retroviral vector backbones increases the level of in vivo transgene expression among progeny of engrafted human hematopoietic stem cells Hum Gene Ther 11, 203950 (2000). Nayler, 0 Stratling, W., Bourquin, J.P., Stagljar, I., Lindemann L., Jasper H. Hartmann, A. M., Fackelmayer, F 0., Ullrich, A. and Stamm, S. SAF-B protein couples transcription and pre-mRNA splicing to SAR/MAR elements Nucleic Acids Res 26, 3542-9 (1998). Nickerson, J. A., Krockmalnic, G. and Penman, S in Cell Biology (ed. Celis, J.E.) 184192 (Academic Press, 1998). Ortega, J.M. and DePamphilis, M. L. Nucleoskeleton and initiation of DNA replication in metazoan cells J Cell Sci 111 3663-73. (1998). Osaka, G., Carey, K., Cuthbertson, A., Godowski, P., Patapoff, T Ryan, A., Gadek, T. and Mordenti, J. Pharmacokinetics, tissue distribution, and expression efficiency of plasmid [33P]DNA following intravenous administration of DNA/cationic lipid complexes in mice: use of a novel radionuclide approach J Pharm Sci 85, 612-8. (1996). Phi-Van, L., von Kries, J P., Ostertag, W. and Stratling, W. H. The chicken lysozyme 5' matrix attachment region increases transcription from a heterologous promoter in heterologous cells and dampens position effects on the expression of transfected genes Mo! Cell Biol 10, 2302-7. ( 1990) Piechaczek, C., Fetzer, C Baiker, A., Bode, J. and Lipps, H.J. A vector based on the SV 40 origin of replication and chromosomal S/MARs replicates episomally in CHO cells. Nucleic Acids Res 27, 426-8 (1999). Pitard, B., Aguerre, 0., Airiau, M., Lachages, A. M., Boukhnikachvili, T., Byk, G., Dubertret, C., Herviou, C., Scherman, D Mayaux, J. F. and Crouzet, J. Virus-sized self assembling lamellar complexes between plasmid DNA and cationic micelles promote gene transfer. Proc Natl Acad Sci US A 94, 14412-7. (1997).

PAGE 107

98 Piva, R., Lambertini, E Penolazzi, L., Facciolo, M. C., Lodi, A., Aguiari, G., Nastruzzi, C and del Senno, L. In vitro stability of polymerase chain reaction-generated DNA fragments in serum and cell extracts. Biochem Pharmacol 56, 703-8 (1998) Poljak, L., Seum, C., Mattioni T and Laemmli, U K. SARs stimulate but do not confer position independent gene expression. Nucleic Acids Res 22, 4386-94. (1994) Ramakrishnan, M ., Liu, W M DiCroce, P.A., Posner, A., Zheng, J., Kohwi Shigematsu, T. and Krontiris, T G Modulated binding of SATB 1, a matrix attachment region protein, to the AT-rich sequence flanking the major breakpoint region of BCL2. Mo! Cell Biol 20, 868-77. (2000). Razin, S. V ., Vassetzky, Y. S. and Hancock, R. Nuclear matrix attachment regions and topoisomerase II binding and reaction sites in the vicinity of a chicken DNA replication origin. Biochem Biophys Res Commun 177, 265-70. (1991). Rollini, P., Namciu, S. J., Marsden, M D and Fournier, R. E Identification and characterization of nuclear matrix-attachment regions in the human serpin gene cluster at 14q32.l. Nucleic Acids Res 27, 3779-91. (1999). Scheule, R. K. and Cheng, S H in Gene Therapy (eds. Lemoine, N. R. & Cooper, D N.) 92-112 (BIOS Scientific Publisher Ltd, Oxford, 1996). Schubeler, D., Mielke, C. Maass, K. and Bode, J. Scaffold/matrix-attached regions act upon transcription in a context-dependent manner. Biochemistry 35, 11160-9 (1996). Shlyakhtenko, L. S., Potaman, V. N., Sinden, R.R. and Lyubchenko, Y. L. Structure and dynamics of supercoil-stabilized DNA cruciforms J Mo! Biol 280, 6172. (1998) Singh, G. B. Kramer, J. A. and Krawetz, S. A. Mathematical model to predict regions of chromatin attachment to the nuclear matrix. Nucleic Acids Res 25, 1419-25 ( 1997) Singhal, A. and Huang, L. in Gene Therapeutics Methods and Applications of Direct GeneTransfer (ed. Wolff, J. A.) 118-142 (Birkhauser, Boston, 1994). Sperinde, G. V. and Nugent, M A. Heparan sulfate proteoglycans control intracellular processing ofbFGF in vascular smooth muscle cells. Biochemistry 37 13153-64. (1998) Stief, A., Winter, D. M Startling, W F. and Slippel, A. E. A nuclear DNA attachment element mediates elevated and position-independent gene activity Nature 341 343-345 (1989). Storrie, B. and Madden, E. A. Isolation of subcellular organelles. Methods Enzymol 182 203-25 (1990)

PAGE 108

99 Stratling, W. H. and Yu, F. Origin and roles of nuclear matrix proteins. Specific functions of the MAR-binding protein MeCP2 / ARBP Crit Rev Eukaryot Gene Expr 9, 311 8 (1999). Struhl, K. Histone acetylation and transcriptional regulatory mechanisms. Genes D e v 12, 599-606 (1998) Sun, Y., Wyatt, R. T., Bigley, A. and Krontiris, T. G. Expression and replication timing patterns of wild type and translocated BCL2 genes Genomics 73, 16170 (2001 ). Tang, F. an Hughes, J. A. Introduction of a disulfide bond into a cationic lipid enhances trans gene expression of plasmid DNA. Biochem Biophys Res Commun 242, 141-5. ( 1998). Tsukamoto, H Wells, D., Brown, S., Serpente, P., Strong, P., Drew, J., Inui, K., Okada, S. and Dickson, G. Enhanced expression ofrecombinant dystrophin following intramuscular injection of Epstein-Barr virus (EBV)-based mini-chromosome vectors in mdx mice. Gene Ther 6, 1331-5. (1999). Tsutsui K. and Muller, M. T. The nuclear scaffold exhibits DNA-binding sites selective for supercoiled DNA. J Biol Chem 263, 7235-41. (1988). Tsutsui, K., Okada, S., Watarai, S., Seki, S., Yasuda, T. and Shohmori, T. Identification and characterization of a nuclear scaffold protein that binds the matrix attachment region DNA. J Biol Chem 268, 12886-94. (1993) van Drunen, C. M ., Sewalt, R. G Oosterling, R. W., Weisbeek, P. J., Smeekens, S. C. and van Oriel, R. A bipartite sequence element associated with matrix / scaffold attachment regions. Nucleic Acids Res 27, 2924-30. (1999). Van Lint C., Emiliani, S Ott, M. and Verdin, E Transcriptional activation and chromatin remodeling of the HIV-1 promoter in response to histone acetylation. Embo J 15, 1112-20. (1996). Vaughn, J P., Dijkwel, P.A., Mullenders, L. H. and Hamlin, J. L. Replication forks are associated with the nuclear matrix. Nucleic Acids Res 18, 1965-9. (1990). Wagner, E., Zatloukal, K Cotten, M., Kirlappos, H., Mechtler, K., Curiel, D. T. and Bimstiel, M. L. Coupling of adenovirus to transferrin-polylysine /DNA complexes greatly enhances receptor-mediated gene delivery and expression of transfected genes. Proc Natl Acad Sci US A 89, 6099-103. (1992). Wan, K. M. Nickerson, J. A., Krockmalnic, G and Penman, S. The nuclear matrix prepared by amine modification. Proc Natl Acad Sci US A 96, 933-8. (1999).

PAGE 109

100 Wang, D M., Taylor, S and Le v y-Wilson, B Evaluation of the function of the human apolipoprotein B gene nuclear matrix association regions in transgenic mice J Lipid R es 37, 2117-24. (1996). Wang, J. C. DNA topoisomerases Annu Rev Biochem 54 665-97 (1985). Wea v er, J. C Electroporation : a general phenomenon for manipulating cells and tissues. J Cell Bio c h e m 51, 426-35. (1993) Wells, D. J., Maule J., McMahon J., Mitchell, R., Damien, E Poole, A. and Wells K. E. Evaluation of plasmid DNA for in vivo gene therapy: factors affecting the number of transfected fibers. J Pharm Sci 87 763-8. (1998) Wells, K. E., Maule, J., Kingston R., Foster, K., McMahon, J., Damien, E., Poole A. and Wells, D J. Immune responses not promoter inactivation, are responsible for decreased long-term expression following plasmid gene transfer into skeletal muscle. FEES Lett 407, 164-8. (1997) Wendelburg B. J. and Vos J.M. An enhanced EBNAl variant with reduced IR3 dom a in for long-term episomal maintenance and trans gene expression of oriP-based plasmids in human cells G e ne The r 5, 1389-99. (1998) Wensing, B., Stuhler, A Jenkins P., Hollyoake, M., Karstegl, C. E and Farrell P J. Variant chromatin structure of the oriP region of Epstein-Barr virus and regulation of EBERl expression by upstream sequences and oriP J Virol 15 6235-41. (2001) Whitelaw C. B. Grolli, S., Accornero, P., Donofrio, G., Farini E and Webster J. Matrix attachment region regulates basal beta-lactoglobulin transgene expression Gene 244 7380. (2000). Will, K., Warnecke, G Wiesmuller, L. and Deppert, W. Specific interaction of mutant p53 with regions of matrix attachment region DNA elements (MARs) with a high potential for base-unpairing. Proc Natl Acad Sci US A 95 13681-6. (1998). Wilson G L., Dean B. S., Wang G. and Dean D A. Nuclear import of plasmid DNA in digitonin-penneabilized cells requires both cytoplasmic factors and specific DNA sequences J Biol Chem 274, 22025-32. (1999). Wilson, J.M., Grossman, M., Wu C.H., Chowdhury, N. R., Wu, G. Y and Chowdhury, J. R. Hepatocyte-directed gene transfer in vivo leads to transient improvement of hypercholesterolemia in low density lipoprotein receptor-deficient rabbits. J Biol Che m 267 963-7. (1992) Wohlgemuth, J. G., Kang, S H Bulboaca, G. H Nawotka, K. A. and Cales M P Long-term gene expression from autonomously replicating vectors in mammalian cells Gene Ther 3, 503-12 (1996).

PAGE 110

101 Wysokenski, D. A. and Yates, J. L. Multiple EBNA I-binding sites are required to form an EBNAl-dependent enhancer and to activate a minimal replicative origin within oriP of Epstein-Barr virus. J Viral 63, 2657-66. (1989). Xie, T. D., Sun, L., Zhao, H. G., Fuchs, J. A. and Tsong, T. Y. Study of mechanisms of electric field-induced DNA transfection. IV. Effects of DNA topology on cell uptake and transfection efficiency. Biophys J 63, 1026-31. (1992) Xie, T D. and Tsong, T. Y. Study of mechanisms of electric field-induced DNA transfection. V. Effects of DNA topology on surface binding, cell uptake, expression, and integration into host chromosomes of DNA in the mammalian cell. Biophys J 65, 1684-9 (1993). Yamamura, J. and Nomura, K. Analysis of sequence-dependent curvature in matrix attachment regions. FEBS Lett 489, 166-70. (2001). Yanagihara, I., Inui, K., Dickson, G., Turner, G., Piper, T., Kaneda, Y. and Okada, S. Expression of full-length human dystrophin cDNA in mdx mouse muscle by HVJ liposome injection Gene Ther 3, 549-53. (1996). Yoshida, M., Kijima, M., Akita, M. and Beppu, T. Potent and specific inhibition of mammalian histone deacetylase both in vivo and in vitro by trichostatin A. J Biol Chem 265, 17174-9 (1990). Zahn-Zabal, M., Kohr, M., Girod, P.A., Imhof, M., Chatellard, P., de Jesus, M. Wurm, F. and Mermod, N. Development of stable cell lines for production or regulated expression using matrix attachment regions. J Biotechnol 87, 29-42. (2001).

PAGE 111

BIOGRAPHICAL SKETCH Pattravadee Chancham was born on May 16, 1974, in Bangkok, Thailand Pattravadee entered Mahidol University in June 1991 and obtained her Bachelor of Science degree in pharmacy in June 1996. She was then admitted to the Department of Pharmaceutics at the University of Florida in August 1997 under a full scholarship from the Thai government. She received her Doctor of Philosophy degree in pharmaceutical sciences in December 2001 under the supervision of Dr. Jeffrey A. Hughes. When not in the lab working hard for Dr. Hughes she enjoys flexing her muscles at the gym and traveling anywhere from Orlando, Florida, to Paris, Franc 102

PAGE 112

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality as a dissertation for the degree of Doctor of Philosophy. I I /'i ~/&~ k J \ighes, Chai~ Associate Professor of Pharmaceutics I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope a nd qualit y as a dissertation for the degree of Doctor of Philosophy. b." ~ -2t'c.~-s: Guenther Hochhaus Professor of Pharmaceutics I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Associate Professor of Pharmaceutics I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality, as a dissertation for the degree of Doctor of Philoyphy. /J.,, ( I / / ,/ ~,... ,...,(7t/'~ 17 ---Edwin M Meyer Associate Professor of Pharmacology and Therapeutics I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in-&eope and quality, as a dissert~tion for the degree of Doctor of Philosophy. ( ~ / (j_;d;f s l'--(/t..-~ ----William G Farrnerie Assistant Scientist and Associate Program Director This dissertation was submitted to the Graduate Faculty of the College of Agricultural and Life Sciences and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy.

PAGE 113

December 2001 Dean, Graduate School